Molecular Characterization of Hereditary Spherocytosis Mutants of the Cytoplasmic Domain of Anion Exchanger 1
and their Interaction with Protein 4.2
by
Susan Pilar Bustos
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Graduate Department of Biochemistry University of Toronto
© Copyright by Susan Pilar Bustos 2011 Year of Convocation
ii
Molecular Characterization of Hereditary Spherocytosis Mutants of the
Cytoplasmic Domain of Anion Exchanger 1 and their Interaction with
Protein 4.2
Susan Pilar Bustos
Doctor of Philosophy
Graduate Department of Biochemistry
University of Toronto
2011
Abstract
Anion exchanger 1 (AE1) is a red cell membrane glycoprotein that associates with cytoskeletal
protein 4.2 in a complex bridging the cell membrane to the cytoskeleton. Disruption of this
linkage results in unstable erythrocytes and hereditary spherocytosis (HS). Three HS mutations
(E40K, G130R and P327R) in the cytoplasmic domain of AE1 (cdAE1) result in a decreased
level of protein 4.2 in the red cell yet maintain normal amounts of AE1. Biophysical analyses
showed the HS mutations had little effect on the structure and conformational stability of the
isolated domain. However, the conformation of the cytoplasmic domain of the kidney anion
exchanger, lacking the first 65 amino acids including a central -strand, was thermally
destabilized relative to cdAE1 and had a more open structure. In transfected human embryonic
kidney (HEK)-293 cells the HS mutants had similar expression levels as wild-type AE1, and
protein 4.2 expression level was not dependent on the presence of AE1. Protein 4.2 localized to
the plasma membrane with wild-type AE1, the HS mutants of AE1, the membrane domain of
AE1 and kidney AE1, and to the ER with Southeast Asian ovalocytosis AE1. A fatty acylation
mutant of protein 4.2, G2A/C173A, could not localize to the plasma membrane in the absence of
AE1. Subcellular fractionation showed wild-type and G2A/C173A protein 4.2 were mostly
iii
associated with the cytoskeleton. Co-immunoprecipitation and Ni-NTA pull-down assays
revealed impaired binding of protein 4.2 to HS mutants compared to AE1, while the membrane
domain of AE1 was unable to bind protein 4.2. These studies show that HS mutations in cdAE1
cause impaired binding of protein 4.2, without causing gross structural changes in the domain.
The mutations change the binding surface on cdAE1 by the introduction of positive charges into
an otherwise acidic domain. This binding impairment may render protein 4.2 more susceptible to
degradation or loss during red cell development.
iv
Acknowledgements
I would like to thank Jing Li for her assistance in cDNA vector construction and in performing
replicates of some of the experiments in Chapter 3. I thank Dr. Walid Houry for the use of his
spectropolarimeter and fluorescence spectrophotometer and Dr. David Clarke for the use of his
differential scanning calorimeter. I would like to thank Dr. Avi Chakrabartty and Sylvia Ho for
the use of the analytical ultracentrifuge and collection of data. I would also like to thank Dr.
Lewis Kay and Dr. Ranjith Muhandiram for the use of the NMR facility and collection of NMR
spectra. I would like to thank Dr. Michael L. Jennings from the University of Arkansas for
Medical Sciences for the mouse anti-AE1 antibody. I would like to thank the Canadian Institutes
of Health Research for supporting this research in the form of a grant, and the Canadian Blood
Services for their support in the form of a Graduate Student Fellowship.
Lastly, and certainly not least, I would like to emphatically thank my supervisor, Dr.
Reinhart Reithmeier. As a supervisor, your encouragement and guidance during the research
lows were as crucial for my success as your praise and kudos during the research highs, which
were always marked with a cake or a celebratory lunch. As a mentor, your quiet confidence,
determination and integrity are things to which I aspire. You’ve sent me around the world to
present my work as a scientist, yet you’ve offered me unconditional support in the
unconventional path I have chosen to take with this degree. Your enthusiasm for your own work
has been an inspiration for my search for meaningful work I can be passionate about. Thank you,
Reinhart, for making my time in your lab a fun, challenging and rewarding experience. I leave
with fond memories of these last several years and anticipation for what comes next.
v
Dedications
This degree is dedicated to my husband, Jim, and to my daughter, Sidney. Thank you, Jim, for
giving me the time and space to study all night when I needed to, and letting me sleep in the next
day. Thanks for all your understanding during the stressful times and for making sure I
celebrated all the milestones. You’ve been my biggest cheerleader all the way through, and I
share this success with you. Thank you, Sidney, for bringing me into your world every day
through your wild imagination and joyful, playful nature. Playing with you was, and continues to
be, the highlight of my day. Thank you for waiting so patiently while I went to the lab, or while I
studied in my ‘office’. Knowing I’d see your smiling face after I finished my work drove me to
work faster, and hoping to be a good model for you drove me to work better.
This work is also dedicated to my parents, Luis and Amelia, and to my mother-in-law,
Nuala. Thank you, Mom and Dad (Yaya and Dado), for always supporting whatever it was that I
wanted to study or pursue, finding ways to make it happen, yet letting me navigate my own
course. Thanks for your optimism, and for believing in me, even when I felt like an ‘imposter’.
Muchos besos y abrazos con ruido. Nuala (Nana), in supporting my whole, little family, you’ve
made it possible for me to get through some challenging times, both academically and
emotionally. Your dedication to us has helped us stay balanced and strong. Thanks for playing
such a crucial role in this journey.
I also dedicate this work to all my other cheerleaders; my siblings and siblings-in-law,
extended family (Bustos, Munroe, Calzato and Fell), and all of my awesome friends. Even when
you weren’t sure what I was doing, you were still there shaking the pom-poms and
congratulating me at the finish line.
vi
Table of contents
Abstract ........................................................................................................................................... ii
Acknowledgements ........................................................................................................................ iv Dedications ......................................................................................................................................v
Table of contents ............................................................................................................................ vi
List of tables ................................................................................................................................... ix List of figures ...................................................................................................................................x
List of appendices .......................................................................................................................... xi
List of abbreviations ..................................................................................................................... xii 1 Chapter 1: Introduction ...............................................................................................................1
1.1 Preamble ..............................................................................................................................1 1.2 Hereditary spherocytosis (HS) .............................................................................................2
1.2.1 Background ..............................................................................................................2 1.2.2 Red cell membrane and cytoskeleton ......................................................................5
1.2.3 Molecular basis of HS ..............................................................................................7 1.3 Anion exchanger 1 (AE1) ..................................................................................................10
1.3.1 Anion exchanger gene family ................................................................................10
1.3.2 AE1 in various species ...........................................................................................14
1.3.3 Structure and function ............................................................................................14 1.3.4 Oligomeric state .....................................................................................................15 1.3.5 Kidney AE1 (kAE1) ..............................................................................................15
1.3.6 Distal renal tubular acidosis (dRTA) .....................................................................16 1.3.7 Southeast Asian ovalocytosis AE1 (AE1SAO) .....................................................17
1.3.8 AE1 knock-outs .....................................................................................................18 1.3.9 HS mutations in AE1 .............................................................................................19
1.4 Membrane domain of AE1 (mdAE1) ................................................................................20
1.4.1 Physiological function ...........................................................................................20 1.4.2 Topology ................................................................................................................24
1.5 Cytoplasmic domain of AE1 (cdAE1) ...............................................................................27 1.5.1 Structure .................................................................................................................27 1.5.2 Physiological function ...........................................................................................30 1.5.3 HS mutations in cdAE1 affecting levels of protein 4.2 .........................................31
1.6 Protein 4.2 ..........................................................................................................................33 1.6.1 Background ............................................................................................................33 1.6.2 Transglutaminase family of enzymes ....................................................................33 1.6.3 Protein 4.2 isoforms ...............................................................................................36 1.6.4 Tissue distribution and expression in different species .........................................37
1.6.5 Structure and properties .........................................................................................38
1.6.6 Interactions with AE1, ankyrin and spectrin..........................................................40
1.6.7 Interactions with the Rh complex and CD47 .........................................................43 1.6.8 HS mutations in protein 4.2 ...................................................................................47
1.7 Thesis focus .......................................................................................................................49 1.7.1 Effect of HS mutations on the structure and stability of the cytoplasmic
domain of AE1 (Chapter 2) ....................................................................................49
vii
1.7.2 Protein 4.2 localization and interaction with wild-type and HS mutants of AE1
in HEK-293 cells (Chapter 3) ................................................................................49 2 Chapter 2: Structure and stability of hereditary spherocytosis mutants of the cytoplasmic
domain of the erythrocyte anion exchanger 1 protein ...............................................................51
2.1 Abstract ..............................................................................................................................51 2.2 Introduction ........................................................................................................................51 2.3 Materials and methods .......................................................................................................52
2.3.1 Materials ................................................................................................................52 2.3.2 Plasmid construction and mutagenesis ..................................................................53
2.3.3 Protein expression and purification .......................................................................53 2.3.4 Analytical ultracentrifugation ................................................................................54
2.3.5 Circular dichroism .................................................................................................54 2.3.6 pH dependence of intrinsic fluorescence ...............................................................55 2.3.7 Calorimetry ............................................................................................................55 2.3.8 Urea denaturation measured by intrinsic fluorescence ..........................................55 2.3.9 Limited tryptic digestion ........................................................................................56
2.4 Results ................................................................................................................................56
2.4.1 Expression and purification of cdAE1 and cdAE1 HS variants in Escherichia
coli..........................................................................................................................56 2.4.2 Analytical ultracentrifugation of wild-type and HS mutant cdAE1 proteins ........56
2.4.3 Secondary structure analysis of wild-type and HS mutant cdAE1 proteins ..........57 2.4.4 Effect of pH on the intrinsic fluorescence of the wild-type and HS mutant
cdAE1 proteins.......................................................................................................60 2.4.5 Thermal denaturation of wild-type and HS mutant cdAE1 proteins by circular
dichroism................................................................................................................60 2.4.6 Thermal denaturation of wild-type and HS mutant cdAE1 proteins by
calorimetry .............................................................................................................63 2.4.7 Urea denaturation of wild-type and HS mutant cdAE1 proteins ...........................66 2.4.8 Limited tryptic digestion of WT and HS mutant cdAE1 proteins .........................66
2.5 Discussion ..........................................................................................................................67 3 Chapter 3: Protein 4.2 interaction with hereditary spherocytosis mutants of the
cytoplasmic domain of human anion exchanger 1 ....................................................................72
3.1 Abstract ..............................................................................................................................72 3.2 Introduction ........................................................................................................................73 3.3 Materials and methods .......................................................................................................73
3.3.1 Materials ................................................................................................................73 3.3.2 Site-directed mutagenesis ......................................................................................74 3.3.3 Transient transfection and expression of AE1 and protein 4.2 in HEK-293
cells ........................................................................................................................75 3.3.4 SDS-PAGE and immunoblotting ...........................................................................75
3.3.5 Immunofluorescence and confocal microscopy.....................................................75 3.3.6 Ni-NTA pull-down.................................................................................................76
3.3.7 Co-immunoprecipitation ........................................................................................76 3.3.8 Subcellular fractionation of HEK-293 cells expressing wild-type or
G2A/C173A protein 4.2 .........................................................................................77 3.4 Results ................................................................................................................................78
3.4.1 Expression of protein 4.2 and AE1 proteins in transfected HEK-293 cells ...........78
3.4.2 Localization of AE1 proteins in HEK-293 cells ....................................................78 3.4.3 Co-localization of protein 4.2 and AE1 in HEK-293 cells ....................................80
viii
3.4.4 Interaction of protein 4.2 with AE1 proteins in HEK-293 cells ............................83
3.4.5 Co-localization of wild-type and G2A/C173A protein 4.2 and AE1 in HEK-
293 cells .................................................................................................................86 3.4.6 Co-localization of wild-type and G2A/C173A protein 4.2 and cell surface
glycans in the absence or presence of AE1 in HEK-293 cells ...............................88 3.4.7 Subcellular fractionation of HEK-293 cells expressing wild-type or
G2A/C173A protein 4.2 .........................................................................................88 3.5 Discussion ..........................................................................................................................90
4 Discussion and future directions ...............................................................................................95
4.1 Structure and conformational stability of the cytoplasmic domain of AE1.......................98 4.2 Protein 4.2 interaction with HS mutants of cdAE1..........................................................100
4.2.1 Quantitative in vitro binding analysis ..................................................................105 4.2.2 Structure determination ........................................................................................106 4.2.3 Interaction of protein 4.2 and AE1 during red cell development ........................108
4.3 Conclusions ......................................................................................................................108 References ....................................................................................................................................111
ix
List of tables
Table 1.1 SLC4 bicarbonate transporter family ..............................................................................11 Table 1.2 HS mutations in the AE1 protein ....................................................................................21 Table 1.3 HS mutations in protein 4.2 ............................................................................................48
Table 2.1 Effects of HS mutations on structure and stability of cdAE1 protein .............................62 Table 4.1 Mutant protein DNA constructs ..........................................................................................95
Table A1 Summary of biophysical properties of cdAE1, cdkAE1, and cdΔ54AE1 at pH 7.5 .....142
x
List of figures
Figure 1.1 Healthy red blood cells and spherocytes ........................................................................3 Figure 1.2 Red cell membrane proteins separated by SDS-PAGE ..................................................4 Figure 1.3 Schematic model of the red cell membrane ...................................................................6 Figure 1.4 Diagrams of proteins expressed by the genes encoding human AE1, AE2 and AE3 ..12 Figure 1.5 Map of the human AE1 gene ........................................................................................13 Figure 1.6 Topographical model of human erythroid AE1 ............................................................26 Figure 1.7 Crystal structure of the cytoplasmic domain of AE1 (cdAE1) .....................................28 Figure 1.8 Crystal structure of the cytoplasmic domain of AE1 (cdAE1) and homology model
of protein 4.2 .............................................................................................................................41 Figure 1.9 Diagram of the protein 4.2 polypeptide with important regions mapped.....................44 Figure 1.10 Schematic model of AE1 and Rh protein complexes at the red cell membrane ........46 Figure 2.1 Analytical ultracentrifugation of wild-type cdAE1 ......................................................58 Figure 2.2 CD spectra of wild-type cdAE1 and HS mutants .........................................................59 Figure 2.3 Intrinsic fluorescence intensity of wild-type cdAE1 and HS mutants as a function
of pH ..........................................................................................................................................61 Figure 2.4 Thermal denaturation of wild-type cdAE1 and HS mutants monitored by CD ...........64 Figure 2.5 Thermal denaturation of wild-type cdAE1 and HS mutants by DSC ..........................65 Figure 2.6 Urea denaturation of wild-type and HS mutant cdAE1 proteins ..................................68 Figure 2.7 Trypsin digestion of wild-type and HS mutant cdAE1 proteins ..................................69 Figure 3.1 Expression of AE1 proteins and protein 4.2 in HEK-293 cells ....................................79 Figure 3.2 Immunofluorescence images of wild-type and mutant AE1 in HEK-293 cells ...........81 Figure 3.3 Immunofluorescence images of protein 4.2 and AE1 proteins in HEK-293 cells .......82
Figure 3.4 Ni-NTA pull-down of wild-type and mutant His6-tagged AE1 with protein 4.2 in
HEK-293 cells ...........................................................................................................................84 Figure 3.5 Co-immunoprecipitation (co-ip) of wild-type and mutant AE1 with protein 4.2 in
HEK-293 cells ...........................................................................................................................85 Figure 3.6 Immunofluorescence images of wild-type and G2A/C173A protein 4.2 and AE1
proteins in HEK-293 cells .........................................................................................................87 Figure 3.7 Immunofluorescence images of cell surface glycans using PNA and wild-type and
G2A/C173A protein 4.2 expressed in the absence or presence of AE1 in HEK-293 cells .......89 Figure 3.8 Subcellular fractionation of HEK-293 cells expressing wild-type or G2A/C173A
protein 4.2 ..................................................................................................................................91 Figure A1 Crystal structure of human cdAE1 .............................................................................135 Figure A2 Domain structure of AE1 isoforms and gel-separated AE1 constructs ......................136 Figure A3 Analytical ultracentrifugation of cdAE1 ....................................................................142 Figure A4 CD spectra of purified cdAE1, cdkAE1, and cdΔ54AE1 at pH 7.5 ...........................143 Figure A5 Thermal denaturation of purified cdAE1, cdkAE1, and cdΔ54AE1 by DSC at pH
7.5 ............................................................................................................................................145 Figure A6 Intrinsic tryptophan fluorescence emission spectra of (A) cdAE1, (B) cdkAE1, and
(C) cdΔ54AE1 at various pH values .......................................................................................147 Figure A7 Average emission wavelength of purified cdAE1, cdkAE1, and cdΔ54AE1 ............148 Figure A8 Urea denaturation of cdAE1 and cdkAE1 monitored by intrinsic tryptophan
fluorescence .............................................................................................................................149
xi
List of appendices
Appendix ......................................................................................................................................133
xii
List of abbreviations
2,3-DPG 2,3-diphosphoglycerate
3D three-dimensional
AE anion exchanger
AE1 anion exchanger 1
AE1SAO Southeast Asian ovalocytosis AE1
AE1HS HS mutants of AE1
AE2 anion exchanger 2
AE3 anion exchanger 3
AE4 anion exchanger 4
βME β-mercaptoethanol
BSA bovine serum albumin
BSSS bis(sulfosuccinimidyl)-suberate
CAII carbonic anhydrase II
CD circular dichroism
cdAE1 cytoplasmic domain of AE1
cdΔ54AE1 cytoplasmic domain of AE1 missing the first 54 amino acids
cdkAE1 cytoplasmic domain of kidney AE1
Cm apparent midpoint of the urea unfolding transition
CNX calnexin
co-ip co-immunoprecipitation
Cp excess heat capacity
CtAE1 carboxy-terminal AE1
C-terminal carboxy-terminal
C-terminus carboxy-terminus
DEER double electron-electron resonance
DMEM Dulbecco's modified Eagle's medium
DTT dithiothreitol
dRTA distal renal tubular acidosis
DSC differential scanning calorimeter/calorimetry
EE R34E/R35E protein 4.2
EPR electron paramagnetic resonance
ER endoplasmic reticulum
ESI-TOF electrospray ionization time-of-flight
GAPDH glyceraldehyde 3-phosphate dehydrogenase
GC G2A/C173A protein 4.2 mutant
GPA glycophorin A
GPB glycophorin B
GPC glycophorin C
GST glutathione S-transferase
H2DIDS 4,4'-diisothiocyanodihydrostilbene-2,2'-disulphonate
HA hemagglutinin
HEK-293 human embryonic kidney-293
HPLC high-performance liquid chromatography
HS hereditary spherocytosis
IOVs inside-out vesicles
IPTG isopropyl-β-D-thiogalactoside
xiii
ITC isothermal calorimetry
kAE1 kidney AE1
LB Luria Bertani
LW intracellular adherence molecule 4
mdAE1 membrane domain of AE1
MWapp apparent molecular weight
MWseq sequence molecular weight
NBC Na+-coupled HCO3
- transporter
Ni-NTA nickel-nitrilotriacetic acid
NMR nuclear magnetic resonance
NMT N-myristoyl transferase
N-terminal amino-terminal
N-terminus amino-terminus
Ni-NTA nickel-nitrilotriacetic acid
PAT palmitoyl-acyl transferase
PBS phosphate buffered saline
PLC phospholipase C
PMSF phenylmethanesulfonyl fluoride
PNA peanut agglutinin
Rh Rhesus
RhAG Rh-associated glycoprotein
SAO Southeast Asian ovalocytosis
SCAM scanning cysteine accessibility mutagenesis
S.D. standard deviation
SDSL site-directed spin labeling
SDS-PAGE sodium dodecyl sulfate-polyacrylamide gel electrophoresis
SE sedimentation equilibrium
SLC4 solute carrier 4
SV sedimentation velocity
tAE1 trout AE1
TG transglutaminase
Tm apparent midpoint of the thermal denaturation transition
TM transmembrane
Wrb Wright b
1
1 Chapter 1: Introduction
1.1 Preamble
Protein-protein interactions and protein networks are fundamentally important in maintaining the
structural integrity and functionality of cells. The red blood cell is essentially a sac of
hemoglobin whose primary function is to deliver oxygen from the lungs to tissues and remove
carbon dioxide, a metabolic waste product, from the tissues to the lungs. The strength and
flexibility of the red cell allows it to travel through the circulatory system, including capillaries
half its diameter, without rupturing for its 120 day lifespan. The proteins that link the red cell
membrane to its cytoskeleton are responsible for these dynamic properties. Integral membrane
glycoprotein anion exchanger 1 (AE1) mediates the electroneutral exchange of chloride for
bicarbonate across the plasma membrane, while its N-terminal cytoplasmic domain (cdAE1)
binds a number of red cell proteins including ankyrin, a cytoskeletal protein. Ankyrin in turn,
binds to spectrin which, along with actin, creates the scaffolding of the red cell cytoskeleton.
Protein 4.2 binds to both AE1 and ankyrin, strengthening their interaction. Perturbation of this
membrane-cytoskeleton linkage caused by mutations in any of these four proteins (AE1, ankyrin,
spectrin and protein 4.2) can lead to membrane destabilization and hereditary spherocytosis
(HS), a hemolytic anemia.
While HS mutations in the membrane domain of AE1 typically cause misfolding of AE1 and
ER retention of the protein, three HS mutations in the cytoplasmic domain (E40K, G130R, and
P327R) are associated with a deficiency of protein 4.2 in the red cell, while maintaining normal
levels of AE1 at the membrane. The focus of this thesis is to characterize the effect of these three
HS mutations on the structure and conformational stability of cdAE1, and on the interaction of
AE1 with protein 4.2 in transfected cells. In a related study, the characterization of the
cytoplasmic domain of kidney AE1 (cdkAE1), a truncated AE1 isoform missing residues 1-65, is
included in the Appendix.
2
1.2 Hereditary spherocytosis (HS)
1.2.1 Background
HS is a common hemolytic anemia and the most common human inherited red cell membrane
disorder. It occurs in about 1 to 2000 Caucasians, but is also common in the Japanese population
(Delaunay 2007). HS is characterized by osmotically fragile spherocytes (spherically-shaped
erythrocytes) that are selectively trapped and destroyed in the spleen (Iolascon et al. 1998). The
accumulation of defective erythrocytes causes enlargement of the spleen, or splenomegaly, and
the removal of the cells from circulation leads to anemia. The clinical severity of the disease
ranges from mild (asymptomatic) compensated hemolysis, where erythrocyte production and
destruction are balanced (Perrotta et al. 2008), to severe hemolytic anemia requiring frequent
blood transfusions and splenectomy. The variation in severity is due to the different molecular
defects that lead to HS and to different levels of bone marrow compensation.
HS symptoms can appear as early as one to two weeks following birth (Delaunay 2007). An
increase in hyperdense red cells with higher hemoglobin concentration is found. The osmotic
fragility occurs partly because of the spheroid shape of the cells, which are less deformable than
the normal discoid shape. HS has been described as resulting from defects of proteins connecting
the red cell membrane to the cytoskeleton (Eber and Lux 2004). Due to the weakened
interactions between the membrane and cytoskeleton, fragments of the membrane bleb off, the
surface area-to-volume ratio of the cell decreases, and the cell becomes spherical and unstable.
Blood smears from HS patients show variable numbers of spherocytes, which lack the central
pale area that is seen in healthy red cells. The images in Figure 1.1 show a comparison between
healthy red cells and spherocytes from cattle, but the effect is the same in humans.
When the red cell membrane proteins of HS patients are subjected to sodium dodecyl sulfate-
polyacrylamide gel electrophoresis (SDS-PAGE), the responsible protein can often be detected
by a decrease in its intensity on the gel. However, the deficiency observed for a certain protein
can also be caused indirectly by a defect in another protein. Figure 1.2 shows red cell membrane
proteins separated by SDS-PAGE following centrifugation of red cell ghosts. The numbering of
proteins based on their order on the gel and their common names are shown to the left.
3
Figure 1.1: Healthy red blood cells and spherocytes.
Electron micrographs of wild-type bovine erythrocytes on the left and HS bovine spherocytes on
the right. Modified from Inaba et al. (1996).
HS
4
Figure 1.2: Red cell membrane proteins separated by SDS-PAGE.
Red cell ghosts incubated in 5 mM sodium phosphate buffer, pH 8, were centrifuged, solubilized
with SDS sample buffer, run on SDS-PAGE and stained with Coomassie blue. Modified from
Steck (1974).
AE1, Band 3
Protein 4.2
GAPDH
Ankyrin
α-spectrin
β-spectrin
Actin
Protein 4.1
5
1.2.2 Red cell membrane and cytoskeleton
When speaking of the red cell membrane, included are the bilayer surrounding the cell,
transmembrane proteins within the bilayer, and the underlying cytoskeleton (Tse and Lux 1999).
The interaction between proteins of the cytoskeleton and integral membrane proteins gives the
red cell its strength and flexibility. These components are shown in a diagram of the red cell
membrane in Figure 1.3. The cytoskeleton is a mesh-like network made up primarily of the
proteins spectrin and actin (Ursitti et al. 1991). Monomers of α-spectrin and β-spectrin intertwine
to form heterodimers (Bennett 1985). These dimers connect head-to-head to form long spectrin
heterotetramers. There are two main points of contact between the cytoskeleton and the
membrane. At one point of contact the tail end of the spectrin tetramer interacts with a junctional
complex, consisting of short actin filaments, tropomyosin, tropomodulin and adducin (Tse and
Lux 1999). Protein 4.1 interacts with the N-terminus of β-spectrin at the actin-binding domain in
the junctional complex and with integral membrane protein glycophorin C (GPC), thereby
linking the cytoskeleton to the membrane. Protein 4.1 also enhances spectrin-actin binding.
At the other major point of contact, the C-terminus of β-spectrin binds to ankyrin, which
binds to the cytoplasmic domain of the integral membrane protein anion exchanger 1 (AE1), or
Band 3. Protein 4.2 stabilizes the interaction between AE1 and ankyrin (Rybicki et al. 1996).
Only the tetrameric form of AE1 binds to ankyrin (Van Dort et al. 1998). Dimeric AE1 is found
associated with the protein 4.1-GPC junctional complex, or as a freely diffusing complex (van
den Akker et al. 2010b). These membrane-cytoskeletal interactions give the red cell the strength
and flexibility that allows it to travel through the circulatory system without rupturing for its 120
day lifespan. In a study using human red blood cell precursors (bone marrow erythroblasts) AE1,
spectrin and ankyrin were found to be expressed at the same time during erythropoiesis (Nehls et
al. 1993). The proteins appeared at the proerythroblast stage, which is the earliest erythroblast
stage. Another study using human erythroid precursors from peripheral blood showed the early
expression of spectrin followed by AE1. Protein 4.2 was not observed until the late erythroblast
stage (orthochromatic) even though protein 4.2 mRNA was seen in the early erythroblasts (Wada
et al. 1999). However, a more recent study of human erythroblast differentiation showed that
AE1 and protein 4.2 are not only expressed at the same time in basophilic erythroblasts, but they
also interact as soon as they appear (van den Akker et al. 2010a).
6
Figure 1.3: Schematic model of the red cell membrane.
The interactions of the integral membrane proteins and cytoskeletal proteins are shown.
Modified from Tse and Lux (1999).
7
1.2.3 Molecular basis of HS
HS is caused by defects in erythrocyte proteins that are involved in the major interaction between
the erythrocyte membrane and the cytoskeleton: spectrin, ankyrin, protein 4.2 and AE1.
Approximately 45 % of HS cases are due to mutations in ankyrin and about 30 % are due to
mutations in β-spectrin (Tse and Lux 1999). Mutations in AE1 account for about 20 % of HS
cases, while approximately 5 % are due to protein 4.2 mutations.
Most of the HS ankyrin mutations are nonsense or frameshift mutations that result in unstable
mRNA transcripts leading to ankyrin deficiencies or truncated peptides causing binding defects.
Several of the latter are truncated at the AE1-binding domain of ankyrin and others are affected
at the spectrin-binding domain (Tse and Lux 1999). The inability of ankyrin to properly bind to
either AE1 or spectrin would disrupt the important linkage between the membrane and
cytoskeleton. Ankyrin defects have been found in both dominant and recessive HS and the
clinical picture ranges from mild hemolysis to transfusion-dependent anemia. Defects lead to
deficiencies in ankyrin, as seen by SDS-PAGE, as well as secondary deficiencies in spectrin and
protein 4.2 (Delaunay 2007). Ankyrin defects can also have a greater effect on the levels of
interacting proteins than on ankyrin itself. Ankyrin from an HS patient with a missense mutation
in the AE1-binding domain (V463I) has decreased affinity for AE1 (Eber et al. 1996) and this
mutant results in a higher deficiency of AE1 than of ankyrin. Even though there is no AE1
molecular defect, the levels of AE1 are affected more than those of the defective ankyrin. A
group recently studied protein sorting in enucleating erythroblasts from ankyrin-deficient mice
(Salomao et al. 2010). They found that while AE1 sorted to the reticulocytes in wild-type mice,
AE1 sorted to both reticulocytes and expelling nuclei in the ankyrin-deficient mice. This shows
that one mechanism of deficiency of AE1 in HS caused by defects in ankyrin may be abnormal
sorting during nuclear extrusion. The same may be true for other proteins where a defect in one
protein results in the deficiency of another.
β-spectrin is the limiting component in the formation of the red cell cytoskeleton (Hanspal
and Palek 1987), while α-spectrin is made in excess. For this reason, HS associated with spectrin
is mostly caused by mutations in β-spectrin. Several null mutations are known to cause dominant
HS where only one allele is expressed (Tse and Lux 1999). No recessive mutations have been
found, nor any compound heterozygous states. Mutations in the β-spectrin gene often appear de
8
novo and the associated HS is moderate to severe, but in general does not require transfusions
after the first year of life (Delaunay 2007). The prominent β-spectrin mutation Promissão results
in removal of the initiation codon since Met1 is mutated to Val. Mutations such as this one, and
others causing amino acid substitutions, result in β-chain deficiency. Because the β-chain is the
limiting step its deficiency results in a deficiency in spectrin assembly. Mutations in α-spectrin
causing HS show a recessive pattern and result in severe spectrin deficiency (Delaunay 2007).
However, no homozygous null alleles have been discovered since this would probably be lethal.
Since α-spectrin is made in three- to four-fold excess of what is used in the cytoskeleton,
defective α-spectrin production from only one allele does not cause HS (Tse and Lux 1999). In
this case, protein expression from the remaining wild-type allele is sufficient for normal
cytoskeletal construction.
HS associated with mutations in AE1 is for the most part dominantly inherited (Tse and Lux
1999), however there are some cases of recessive inheritance (Inoue et al. 1998, Ribeiro et al.
2000, Rybicki et al. 1993). Mutations are found throughout the gene, including regions coding
for the membrane domain (mdAE1) and the cytoplasmic domain (cdAE1). Details about the
specific domains of AE1 will be discussed in the AE1 section to follow. The clinical picture
ranges from mild to severe HS requiring splenectomy and transfusions. SDS-PAGE usually
shows a reduction of AE1 with a secondary reduction in protein 4.2. A homozygous null
mutation (Band 3 Coimbra), which is the equivalent of a human AE1 knock-out, is caused by a
point mutation, V488M (Ribeiro et al. 2000). In the patient with this mutation, both AE1 and
protein 4.2 were completely absent, and spectrin, ankyrin, glyceraldehyde 3-phosphate
dehydrogenase (GAPDH) and glycophorin A were significantly reduced. HS caused by this
mutation was severe, with transfusions required immediately after birth, splenectomy performed
at 3.5 years, and a regimen of regular blood transfusions thereafter.
Some nonsense mutations result in mRNA instability, resulting in mRNA degradation and
AE1 deficiency (Tse and Lux 1999). This results in fewer points of contact for the cytoskeleton
at the membrane. Missense mutations located in the transmembrane (TM) domain can cause
protein conformational changes leading to protein degradation and a decreased level of AE1 at
the membrane. More details about how these mutations in mdAE1 are associated with HS will be
discussed later. Two recessive HS cases caused by AE1 mutations, E40K and G130R, in the
cytoplasmic domain have normal AE1 content in red cells, but a deficiency in protein 4.2 (Inoue
9
et al. 1998, Rybicki et al. 1993). A decrease in protein 4.2 with normal amounts of AE1 is also
seen in one dominant mutation, P327R, located in the cytoplasmic domain (Jarolim et al. 1992a).
Since these three mutations are in the cytoplasmic domain and only result in protein 4.2
deficiency, it is believed they result in impaired protein 4.2 binding. In fact, when the G130R
mutant exists in trans with another AE1 mutant (Band 3 Okinawa) that prevents its stable
incorporation into the membrane, there is a near total deficiency of protein 4.2 (Kanzaki et al.
1997). In the patient with these AE1 mutations, only the G130R AE1 protein is seen in the
mature erythrocyte. The authors reasoned that in red cell precursors virtually all of the protein
4.2 is bound to the Okinawa mutant and because of its inability to stably incorporate into the
membrane, the Okinawa-protein 4.2 complex became degraded, leaving only the G130R mutant
at the membrane and no protein 4.2 in the mature cell. In the homozygous G130R case, it may be
that protein 4.2 binds sufficiently to both copies of mutant AE1 to retain about half of the protein
4.2 content in the red cell.
Several mutations in the protein 4.2 gene result in red cells with no detectable protein 4.2, but
the resulting HS is not as severe as it is in patients with a complete absence of AE1. The clinical
picture resulting from protein 4.2 mutations in the homozygous state ranges from mild HS to
moderate, uncompensated hemolytic anemia, while heterozygotes are normal (Iolascon et al.
1998). These mutations result in decreased levels of protein 4.2 in the red cell, and some affect
binding to AE1 (Toye et al. 2005) and ankyrin (Su et al. 2007). Deficiency of protein 4.2 occurs
with a secondary reduction in CD47, an Rh complex protein believed to help anchor the
membrane to the cytoskeleton (Bruce et al. 2002). Two novel protein 4.2 mutations in an HS
patient, Chartres-1 (Tyr435STOP) and Chartres-2 (out of frame deletion in exon 9), resulted in
complete protein 4.2 deficiency (van den Akker et al. 2010a). In the protein 4.2-deficient
erythrocytes of this patient, the AE1-ankyrin interaction was weakened, supporting the role of
protein 4.2 as a stabilizer of the membrane-cytoskeleton interaction. As mentioned above, some
HS mutations in cdAE1 result in protein 4.2 deficiency with a normal AE1 content. The
mechanism of disease with these cdAE1 mutations may be indirect, whereby impaired protein
4.2 binding to cdAE1 causes a deficiency in protein 4.2, which then results in a weakened AE1-
ankyrin interaction.
10
1.3 Anion exchanger 1 (AE1)
1.3.1 Anion exchanger gene family
Anion exchangers (AE) are integral membrane proteins encoded by members of the solute
carrier 4 (SLC4) gene family, also known as the bicarbonate transporter family (Romero et al.
2004). The human family of ten members includes three AE proteins (AE1-3) and five Na+-
coupled HCO3- transporters (NBC). The last two members, AE4 and BTR1, have unknown
functions, but based on sequence homology lie between the two major groups of exchangers and
co-transporters. Table 1.1 shows the names, functions and tissue distribution of the family
members, SLC4A1 (AE1, Band 3) being the subject of this thesis.
The AE proteins can be divided into three structural regions: a hydrophilic amino-terminal
(N-terminal) extension, a hydrophilic cytoplasmic core domain, and a carboxyl-terminal (C-
terminal) hydrophobic TM domain. The N-terminal extensions of AE2 and AE3 are homologous
and are about 300 and 200 residues longer, respectively, than the AE1 extension which is only 58
residues long. The conserved cytoplasmic core of AE1 begins at Val59 (Reithmeier et al. 1996).
The membrane domain of the AE proteins, which in AE1 begins at Gly361 (Tanner 1997) and
extends to the C-terminus, catalyzes the electroneutral exchange of Cl- for HCO3
- across the
plasma membrane. The membrane domains in the AE family have high sequence identity,
indicating that the anion exchange function has been conserved and operates by a similar
mechanism. AE1 (Band 3) is expressed abundantly in erythrocytes and a truncated kidney
isoform (kAE1) missing the first 65 amino acids, including the N-terminal extension, is
expressed in the α-intercalated cells of the collecting tubule of kidney. AE2 is ubiquitously
expressed and AE3 is expressed mainly in the heart and the brain. Figure 1.4 shows
diagrammatic comparisons of the protein domains of AE1, kAE1, AE2 and AE3.
The gene encoding human AE1 was localized to chromosome 17q21-qter (Showe et al.
1987). The complete protein-coding sequence of human AE1 was obtained from a fetal liver
library cDNA clone and was predicted to code for a protein 911 amino acids in length (Tanner et
al. 1988). It was later discovered that this AE1 sequence was not the most common sequence,
but contained the Memphis I variant, K56E (Jarolim et al. 1992b, Yannoukakos et al. 1991), a
widespread polymorphism with no effect on AE1 function. The only known molecular
characteristic of this variation is that it causes proteolytic fragments derived from the amino-
11
terminus of the protein to run 3 kDa slower on SDS-PAGE (Mueller and Morrison 1977). The
wild-type sequence of AE1 was also cloned from a fetal liver cDNA library (Lux et al. 1989).
The AE1 gene is about 18 kb in length and contains 20 exons (Sahr et al. 1994, Schofield et al.
1994). Erythroid AE1 is transcribed from a promoter region immediately upstream of exon 1,
while the kAE1 mRNA lacks exons 1 to 3, and uses an initiator codon within exon 5. This gives
rise to a truncated protein product missing the first 65 residues from the N-terminus of the
erythroid version and begins at Met66 (Kollert-Jons et al. 1993). Figure 1.5 compares the exon
structures and transcriptional start sites of AE1 and kAE1.
Table 1.1: SLC4 bicarbonate transporter family
Gene Common name Function Tissue distribution
SLC4A1 AE1, Band 3 Cl-/HCO3
- exchange,
erythrocyte stability
Erythrocyte, kidney
SLC4A2 AE2 Cl-/HCO3
- exchange Ubiquitous distribution
SLC4A3 AE3 Cl-/HCO3
- exchange Brain, heart, retina,
kidney, GI tract
SLC4A4 NBCe1, NBC1 Na+/HCO3
- cotransport NBCe1-A: kidney, eye
NBCe1-B: ubiquitous
NBCe1-C: brain
SLC4A5 NBCe2, NBC4 Na+/HCO3
- cotransport Liver, testes, spleen
SLC4A7 NBCn1, NBC2, NBC3 Na+/HCO3
- cotransport Ubiquitous distribution
SLC4A8 NDCBE Na+-dependent Cl
-/
HCO3- exchange
Brain, testes, kidney,
ovary
SLC4A9 AE4 Cl-/HCO3
- exchange? Kidney
SLC4A10 NCBE Na+/HCO3
- cotransport? Brain
SLC4A11 BTR1 Unknown Kidney
12
Figure 1.4: Diagrams of proteins expressed by the genes encoding human AE1, AE2 and
AE3. Predicted TM domains are blue, while cytoplasmic domains, including the N-terminal
extensions, are white. First and last amino acids are indicated. Modified from Alper (2009).
13
Figure 1.5: Map of the human AE1 gene.
The positions of exons 1-20 are indicated. Protein coding regions are shown as black boxes, and
the 5’ and 3’ noncoding regions shown as white boxes. Erythroid (ATGE) and kidney (ATGK)
initiator codons are indicated, and exon K1 that forms the 5’ end of kidney AE1 mRNA is shown
as a blue box. Modified from Sahr et al. (1994).
14
1.3.2 AE1 in various species
A wide range of vertebrate cells display electroneutral Cl-/HCO3
- exchange function. In
erythrocytes and specialized kidney cells, this function is assigned to AE1. AE1 was first cloned
from the mouse erythrocyte in 1985 (Kopito and Lodish 1985). The human erythroid AE1
sequence was deduced from fetal liver cDNA three years later (Tanner et al. 1988). The mouse
AE1 protein is 18 residues longer than human AE1 and their sequences are highly similar with
an 81 % sequence identity. Other mammalian AE1 sequences have been determined for bovine
(GenBank AF163826) and rat erythrocyte AE1 (NCBI NP_036783.2). Chicken is the only bird
whose erythroid AE1 sequence has been deduced (Cox and Lazarides 1988). AE1 has also been
sequenced from fish species, including trout (Hubner et al. 1992) and zebrafish (Paw et al.
2003). The transport activity of trout AE1 (tAE1) can be regulated by cell swelling, whereby
tAE1 releases taurine in order to regulate cell volume (Motais et al. 1997).
An anion transporter homologue was identified in the yeast Saccharomyces cerevisiae
(Philippsen et al. 1997) by sequence similarity to mammalian anion exchangers and Na+/HCO3
-
transporters. Expression of its cDNA in yeast produced a protein localized to the plasma
membrane that could bind anions including Cl- and HCO3
- (Zhao and Reithmeier 2001). Yeast
YNL275w, the AE1 homologue, is related to Bor1 of plants, a borate transporter (Jennings et al.
2007, Takano et al. 2002). AE gene family members are expressed in Caenorhabditis elegans
(Sherman et al. 2005), but not in bacteria. However, a 4,4'-diisothiocyanodihydrostilbene-2,2'-
disulphonate (H2DIDS)-sensitive membrane protein involved in bicarbonate transport into the
photosynthesizing cells of the marine alga sea lettuce (Ulva sp.) was discovered (Drechsler et al.
1993). Shortly after, a 95 kDa membrane protein in the same organism was recognized by
antibodies against human AE1 (Sharkia et al. 1994), indicating that a similar protein may have
evolved in marine algae for the purpose of bicarbonate transport, likely involved in CO2 fixation.
1.3.3 Structure and function
The human AE1 protein consists of two structurally and functionally distinct domains. The 43
kDa N-terminal cdAE1, from Met1 to Lys360 (Tanner 1997) binds to cytoskeletal proteins,
glycolytic enzymes and deoxyhemoglobin (Low 1986, Willardson et al. 1989). This region,
through its interaction with the cytoskeleton, is responsible for helping maintain the structural
integrity of the red cell. This domain can be cleaved from the red cell ghost membrane using
15
mild trypsin or chymotrypsin treatment (Steck et al. 1976) and has been shown to maintain its
protein-binding function independent of the membrane domain (Low 1986, Wang et al. 1992).
The 52 kDa mdAE1, from Gly361 to Val911 (Lepke and Passow 1976, Tanner 1997) spans the
plasma membrane about 12 times and is responsible for the exchange of Cl- for HCO3
- (Jennings
1989b). mdAE1 has been shown to maintain its transport function in the absence of cdAE1
(Grinstein et al. 1978). The extreme C-terminal 33 residues of AE1 are cytoplasmic and contain
a region that binds carbonic anhydrase II (CAII) (Vince and Reithmeier 1998), which produces
the HCO3- that is transported out of the cell by AE1. The association of CAII with AE1 increases
the anion exchange efficiency of AE1 and is therefore believed to form a transport metabolon,
linking transport (AE1) and metabolism (CAII) (Sterling et al. 2001).
1.3.4 Oligomeric state
AE1 exists as a mixture of dimers and tetramers in the erythrocyte membrane and in detergent
solutions (Jennings 1989b). When examined by high-performance liquid chromatography
(HPLC) in solutions containing the nonionic detergent C12E8, the predominant species (70 %)
were dimers with the remainder being tetramers (Casey and Reithmeier 1991). Dissociation of
AE1 to monomers requires the use of denaturing detergents such as SDS (Salhany et al. 1997).
Ankyrin binds and stabilizes the tetrameric form of AE1 (Pinder et al. 1995, Thevenin and Low
1990). AE1 tetramers are dimers of dimers based on the cytoplasmic domain crystal structure
where only one subunit from each dimer is in contact within the tetramer (Zhang et al. 2000).
After the combined cross-linking of the membrane and cytoplasmic domains of AE1 in intact red
cells with the membrane impermeant active ester bis(sulfosuccinimidyl)-suberate (BSSS) and
then Cu2+
/o-phenanthroline to cross-link cytoplasmic sulfhydryls, respectively, a study showed
that mostly dimers were formed. These results indicate that membrane domains and cytoplasmic
domains of the same pair of subunits become cross-linked to each other within a stable dimer
(Jennings and Nicknish 1985).
1.3.5 Kidney AE1 (kAE1)
A kidney isoform of AE1 (kAE1) is expressed in the basolateral membrane of α-intercalated
acid-secreting cells in the collecting ducts of the distal nephron (Wagner et al. 1987). The kAE1
protein exchanges Cl- for HCO3
- across the basolateral membrane resulting in bicarbonate
reabsorption into the blood allowing acid excretion into the urine by a H+-ATPase (Karet 2002).
16
The kAE1 protein lacks the first 65 residues of AE1 (Kollert-Jons et al. 1993) and is unable to
bind glycolytic enzymes and ankyrin (Ding et al. 1994, Wang et al. 1995). The acidic extreme
N-terminus of AE1 is missing in the kidney isoform. Since this region has been implicated as the
binding region for glycolytic enzymes and one of the regions of ankyrin binding in AE1, it is not
surprising that kAE1 is unable to bind to these proteins. Another region missing in kAE1 is the
first β-strand that runs through the centre of the globular protein-binding domain of cdAE1.
Biophysical studies carried out by Biochemistry project student Allison Pang under my
supervision (see Appendix) have revealed that the cytoplasmic domain of kAE1 (cdkAE1) is less
thermally stable than erythroid cdAE1 and exists in a more open structure (Pang et al. 2008).
Tryptophan residues that are buried in a hydrophobic environment in cdAE1 are more solvent-
exposed in the folded cdkAE1 protein.
A novel mutant we constructed that is missing the acidic N-terminus but retains the central
β-strand (cdΔ54AE1) had a similar folded structure and thermal stability as cdAE1, indicating
that the differently-folded structure of cdkAE1 could be attributed to the missing β-strand (Pang
et al. 2008). This altered structure of cdkAE1 may account for the impaired binding of kAE1 to
AE1-binding partners. Its ability to bind protein 4.2 has not been determined, but even with its
altered structure, cdkAE1 may have retained its protein 4.2-binding site. As well, protein 4.2 is
present in kidney cells, so it is possible it plays a similar role in kidney as it does in erythrocytes,
stabilizing the interaction between the membrane and the underlying cytoskeleton.
1.3.6 Distal renal tubular acidosis (dRTA)
Mutations in kAE1 can result in distal renal tubular acidosis (dRTA), a kidney disease
characterized by impaired acid secretion into the urine. This leads to metabolic acidosis,
hypokalaemia, bone disease, and nephrocalcinosis (Batlle et al. 2001, Rodriguez-Soriano 2000).
Several mutations in the membrane domain of kAE1 cause dRTA either because of mistargeting
of the protein to the apical membrane, or because of impaired exchanger function. Interestingly,
there are no reports of dRTA mutations in the cytoplasmic domain of kAE1. Normally,
mutations in AE1 cause either HS or dRTA, but not both. For example, the R589H mutation is
associated with dRTA, but not HS. Not surprisingly, when R589H AE1 and R589H kAE1 were
transiently expressed in HEK-293 cells, the mutation caused intracellular retention of kAE1, but
not of AE1 (Quilty et al. 2002). There are two known cases where a patient presents with both
diseases. The V488M mutation (Band 3 Coimbra) (Ribeiro et al. 2000), essentially results in a
17
knock-out of AE1 in erythrocytes and kAE1 in kidney cells, as discussed above. It is not
surprising that a total absence of AE1 and kAE1 can cause HS and dRTA. The S667F mutation
(Band 3 Courcouronnes) (Toye et al. 2008) causes an erythrocyte AE1 reduction to about 35 %
and results in both HS and dRTA.
As with Southeast Asian ovalocytosis AE1 (AE1SAO), which is described below,
glycophorin A (GPA) can facilitate the cell surface expression in Xenopus oocytes of the dRTA
mutant G701D that is otherwise retained intracellularly (Tanphaichitr et al. 1998). Kidney cells
do not express GPA, nor do they express an equivalent protein behaving in the same way
(Kittanakom et al. 2004). Thus, no rescue of the dRTA trafficking mutants to the kidney cell
surface occurs as it does in red cell precursors expressing GPA. This accounts for the lack of
effect of dRTA mutations on red cell AE1 expression and trafficking.
1.3.7 Southeast Asian ovalocytosis AE1 (AE1SAO)
Southeast Asian ovalocytosis (SAO) is a condition whereby red blood cells become rigid and
oval-shaped (Amato and Booth 1977). As the name implies, SAO is found almost exclusively in
Southeast Asia, including the Malay archipelago, the Philippines, Indonesia, Thailand, and Papua
New Guinea (Nurse et al. 1992). This hematological condition is caused by a nine amino acid
deletion (Δ400-408) at the junction between the cytoplasmic and transmembrane domains of
AE1 (Jarolim et al. 1991). The SAO deletion mutant of AE1 (AE1SAO) causes the cell to be
extraordinarily rigid (Mohandas et al. 1992), conferring some protection against malaria since
the red cells become resistant to entry of the parasite (Kidson et al. 1981). This clinical
protection has allowed SAO to persist in the Southeast Asian population which has a high
incidence of Plasmodium falciparum malaria (Allen et al. 1999). AE1SAO is asymptomatic in
the heterozygous state, but the homozygous state is thought to be embryonic lethal since this
state has never been observed and because of a higher incidence of miscarriage when both
parents carry the deletion (Liu et al. 1994).
In ovalocytes the AE1SAO mutant is able to traffic to the plasma membrane to about 48 %
of the total (Sarabia et al. 1993) but has no anion exchange function (Schofield et al. 1992,
Tanner 1997) as a result of a misfolded membrane domain (Moriyama et al. 1992). The mutant
has an increased association with the cytoskeleton, decreased lateral (Liu et al. 1990) and
rotational (Liu et al. 1995) movement, and decreased extractability from red cell ghost
18
membranes (Sarabia et al. 1993), which contribute to the rigidity of the SAO red cell. In HEK-
293 cells the AE1SAO mutant becomes retained in the ER during biosynthesis and is not seen at
the plasma membrane (Cheung et al. 2005a). The misfolded AE1SAO protein in HEK-293 cells
is most likely retained in the endoplasmic reticulum (ER) by the cellular quality control
machinery, whereas in the red cell erythroid-specific factors may allow for cell-surface
expression. In HEK cells the deletion disrupts the proper membrane integration of
transmembrane region 1, likely causing the polar amino acid region directly N-terminal to the
deletion to get pulled into the membrane (Cheung and Reithmeier 2005). This would account for
a less flexible link between the cytoplasmic and membrane domains and the rigidity seen in SAO
cells. Red cells express glycophorin A (GPA), which is known to associate with AE1 at the
membrane. The association of AE1 and GPA at the plasma membrane creates the Wright b (Wrb)
blood group antigen at the surface. Arg61 of GPA interacts with Glu658 in TM 8 of AE1 to
form the Wrb antigen (Bruce et al. 1995). In Xenopus oocytes, GPA can facilitate the cell surface
expression of wild-type AE1 (Groves and Tanner 1992) as well as that of the SAO deletion
mutant that is otherwise retained intracellularly (Groves et al. 1993). This shows that the region
of AE1 around TM 8 that associates with GPA remains intact in AE1SAO.
AE1SAO is associated with hemolytic anemia at birth, but disappears within the first three
years of life (Laosombat et al. 2010). The lack of hemolytic anemia in SAO individuals after
three years of age implies that the cytoskeletal protein interactions of its cytoplasmic domain are
intact. In fact, AE1SAO interacts more strongly with ankyrin than does wild-type AE1 (Liu et al.
1990). Circular dichroism (CD) and calorimetric data showed structural similarities between the
cytoplasmic domains of wild-type AE1 and AE1SAO (Moriyama et al. 1992, Sarabia et al.
1993).
1.3.8 AE1 knock-outs
As mentioned above, Band 3 Coimbra (V488M) in the homozygous state is the equivalent of an
AE1 knock-out in humans (Ribeiro et al. 2000). The mutation resulted in a membrane trafficking
defect and the absence of AE1 causing severe HS requiring splenectomy and regular blood
transfusions. Protein 4.2 was also absent, and significant reductions in spectrin, ankyrin, GAPDH
and glycophorin A were observed. The mutation also caused a lack of kAE1 in kidney, which
resulted in dRTA. In addition to regular transfusions to treat the HS, the patient required daily
bicarbonate supplements to treat the metabolic acidosis associated with dRTA. In the
19
heterozygous state Band 3 Coimbra caused typical HS and was associated with partial deficiency
in AE1 and protein 4.2.
Studies in AE1 knock-out mice revealed that most AE1-/-
mice die within two weeks of birth
(Southgate et al. 1996). These mice have severe hemolytic anemia, no protein 4.2 or glycophorin
A in their red cells, but an assembled red cell cytoskeleton detached from the membrane.
Heterozygote mice are identical to normal mice in appearance and growth. A natural AE1
mutation in cattle, Arg646Stop, resulted in red cells with no AE1 in homozygous animals
(Inaba et al. 1996). This mutation created the equivalent of an AE1 knock-out in cattle and
occurred with deficiencies in red cell spectrin, ankyrin, actin, and protein 4.2. The red cells were
spherocytic and extremely unstable and the cattle had moderate uncompensated anemia with HS.
This indicated that AE1 provides membrane structural stability, but is not essential to the
survival of cows. Zebrafish with AE1 mutations (Retsina) equivalent to an AE1 knock-out
developed chronic anemia (Paw et al. 2003). In addition, the absence of AE1 caused a
cytokinetic defect at the erythroblast stage. This cytokinetic defect was also seen in mouse AE1
knock-outs, indicating a conserved role for AE1 in normal erythroid cytokinesis.
1.3.9 HS mutations in AE1
Since AE1 is expressed in erythrocytes and kidney cells, both of these tissues may be affected by
mutations in AE1. The affected tissue depends on the site of the mutation in the AE1 gene. The
kidney defect, dRTA, was discussed above in the kAE1 section. In erythrocytes, defective AE1
may lead to hemolytic HS. HS-associated mutations can occur in mdAE1 and cdAE1. Mutations
in mdAE1 have been shown to cause misfolding of the protein and retention in the ER in studies
using transfected mammalian cells (Quilty and Reithmeier 2000). ER retention of AE1 in
differentiating red cell precursors would result in a deficiency of AE1 at the plasma membrane of
mature cells, and fewer sites of membrane-cytoskeletal linkage since the ER is removed during
erythropoiesis. The decrease in AE1 at the plasma membrane accounts for the membrane
structural instability observed in HS patients with these mdAE1 mutations. Some mutations in
the cdAE1 create a premature stop codon in the gene, which codes for a truncated protein
product. This type of mutation also results in less AE1 at the plasma membrane of red cells, and
accounts for the membrane structural instability of these cells.
20
Interestingly, three HS mutations in cdAE1 (E40K, G130R, and P327R) do not result in a
misfolded or truncated protein product (Inoue et al. 1998, Jarolim et al. 1992a, Rybicki et al.
1993). These mutant proteins are present at the red cell membrane in normal amounts, yet result
in HS nonetheless. The only known phenotype of these mutants is a deficiency of protein 4.2, a
peripheral membrane protein that binds to AE1 and ankyrin. Table 1.2 lists all known mutations
of AE1 associated with HS.
1.4 Membrane domain of AE1 (mdAE1)
1.4.1 Physiological function
The 52 kDa membrane domain of AE1 encompassing residues 361 to 911 is responsible for the
electroneutral exchange of Cl- for HCO3
- across the red cell membrane. AE1 works by an
electroneutral ping-pong mechanism (Furuya et al. 1984). The protein has only one anion
binding site but exists in two main conformations allowing the binding site to be exposed to
either the cytoplasmic side or the extracellular side (Pal et al. 2006). Interconversion between the
two conformational states occurs with anion binding, but this conformational change is very slow
in the absence of substrates. AE1 is a very fast transporter operating at 100 000 chloride ions per
second per molecule. AE1 is able to transport other anions, such as sulfate, but at a much slower
rate (Milanick and Gunn 1984). mdAE1 is able to carry out exchange function in the absence of
the cytoplasmic domain. A study using AE1 in red cell inside-out vesicles (IOVs) found that
trypsinization to release the cytoplasmic region had no effect on sulfate efflux showing that the
transport activity of the membrane domain was intact (Grinstein et al. 1978).
Some studies have uncovered clues regarding important residues involved in anion
transport. Anion transport can be inhibited by a variety of anionic compounds, most notably
stilbene disulfonates such as H2DIDS, which bind to one site per AE1 molecule accessible from
the cell exterior (Lepke et al. 1976). At low temperature (0 ºC) and short times, this binding is
reversible. At higher temperature (37 ºC) H2DIDS can react covalently with Lys539 of AE1, and
can react more slowly with Lys851 of the same AE1 protein, cross-linking two parts of AE1
(Okubo et al. 1994). Mutation of homologous lysine residues (Lys558Asn and Lys869Met)
in mouse AE1 prevented irreversible inhibition by H2DIDS but did not affect the transport
function of AE1 when expressed in Xenopus oocytes (Wood et al. 1992), showing that this Lys
was not essential for transport function.
21
Table 1.2: HS mutations in the AE1 protein
AE1 variant Mutation Abnormal allele Reference
Cytoplasmic domain defects: protein 4.2 deficiency
Montefiore Glu40Lys Homozygous (Rybicki et al. 1993)
Fukuoka Gly130Arg Homozygous (Inoue et al. 1998)
Tuscaloosa Pro327Arg (Memphis II:
Lys56Glu in cis) Heterozygous (Jarolim et al. 1992a)
Cytoplasmic domain defects: affects ankyrin binding
Nachod (Hradec
Kralove II) Δcodons 117-121 Heterozygous (Jarolim et al. 1996)
Cytoplasmic domain defects: mRNA instability
Neapolis Δ of 1st 11 amino acids Heterozygous (Perrotta et al. 2005)
Genas G89A in exon 2 Heterozygous (Alloisio et al. 1996)
Foggia ΔACCCACACCAC in codon
54 or 55 Heterozygous
(Miraglia del Giudice
et al. 1997)
Bohain ΔT in codon 81 Heterozygous (Dhermy et al. 1997)
Hodinin (Prague IV) Trp81Stop Heterozygous (Jarolim et al. 1996)
Napoli I Insertion TCTTTCT in codon
100 Heterozygous
(Miraglia del Giudice
et al. 1997)
Osnabruck I (Lyon) Arg150Stop Heterozygous (Alloisio et al. 1996)
Worcester Insertion G into codons 170-172 Heterozygous (Jarolim et al. 1996)
Campinas Stop codon (+13) after exon 7 Heterozygous (Lima et al. 1997)
Princeton Insertion C into codons 273-275 Heterozygous (Jarolim et al. 1996)
Noirterre Gln330Stop Heterozygous (Jenkins et al. 1996)
22
Table 1.2: HS mutations in the AE1 protein (continued)
AE1 variant Mutation Abnormal allele Reference
Cytoplasmic domain defects: AE1 deficiency
Cape Town Glu90Lys (Arg870Trp in trans)
Compound
heterozygous with
Band 3 Prague III
(Bracher et al. 2001)
Mondego Pro147Ser (E40K in cis; V488M in
trans)
Compound
heterozygous with
Band 3 Coimbra
(Alloisio et al. 1997)
Boston Ala285Asp Heterozygous (Jarolim et al. 1996)
Membrane domain defects: mRNA instability
Bruggen ΔC in codon 419 Heterozygous (Eber et al. 1996)
Bicêtre II ΔG in codons 454-456 Heterozygous (Dhermy et al. 1997)
Pribram (Prague VI) Stop codon (+7) after exon 12 Heterozygous (Jarolim et al. 1996)
Evry ΔT in codon 496 Heterozygous (Dhermy et al. 1997)
Smichov (Prague VII) ΔC in codon 616 Heterozygous (Jarolim et al. 1996)
Trutnov Tyr628Stop Heterozygous (Jarolim et al. 1996)
Hobart ΔG in codons 646-647 Heterozygous (Jarolim et al. 1996)
Osnabruck II Δ of codon 663 or 664 Heterozygous (Eber et al. 1996)
Membrane domain defects: AE1 deficiency
Benesov (Prague V) Gly455Glu Heterozygous (Jarolim et al. 1996)
Edmonton Cys479Trp (Gly701Asp in trans) Compound
heterozygous with
Gly701Asp
(Chu et al. 2010)
Coimbra Val488Met Heterozygous and
homozygous
(Alloisio et al. 1997)
Bicêtre I Arg490Cys Heterozygous (Dhermy et al. 1997)
Pinhal Arg490His Heterozygous (Lima et al. 1999)
Milano Insertion of 23 amino acids in codon 498 Heterozygous (Bianchi et al. 1997)
23
Table 1.2: HS mutations in the AE1 protein (continued)
AE1 variant Mutation Abnormal allele Reference
Membrane domain defects: AE1 deficiency
Dresden Arg518Cys Heterozygous (Eber et al. 1996)
Tambaú Met663Lys Heterozygous (Lima et al. 2005)
Courcouronnes Ser667Phe Heterozygous (Toye et al. 2008)
Most (Prague VIII) Leu707Pro Heterozygous (Eber et al. 1996)
Okinawa Gly714Arg (Gly130Arg in trans) Compound
heterozygous with
Band 3 Fukuoka
(Kanzaki et al. 1997)
Kumamoto (Prague II) Arg760Gln Heterozygous (Jarolim et al. 1995)
Hradec Kralove Arg760Trp Heterozygous (Jarolim et al. 1995)
Chur Gly771Asp Heterozygous (Maillet et al. 1995)
Napoli II Ile783Asn Heterozygous (Miraglia del Giudice
et al. 1997)
Jablonec Arg808Cys Heterozygous (Jarolim et al. 1995)
Prague Duplication of nucleotides 2455-2464 Heterozygous (Jarolim et al. 1994)
Birmingham His834Pro Heterozygous (Jarolim et al. 1996)
Philadelphia Thr837Met Heterozygous (Jarolim et al. 1996)
Tokyo Thr837Ala Heterozygous (Iwase et al. 1998)
Prague III Arg870Trp Heterozygous (Jarolim et al. 1995)
Vesuvio ΔACCAC in codon 894 Heterozygous (Perrotta et al. 1999)
Conversion of Glu681 to an alcohol by reduction using Woodward’s reagent K and BH4-,
thereby removing its negative charge, inhibits chloride anion exchange, yet activates chloride-
sulfate exchange (Jennings 1995). The same thing happens when a low pH outside the cell
protonates a titratable carboxyl group, thought to be the same Glu681. This glutamate is believed
to be the binding site for the proton that is co-transported with sulfate during net chloride-sulfate
exchange (Jennings and Smith 1992). This residue is located at the end of TM 8 on the
24
cytoplasmic side of the membrane, yet treatment yielding the alcohol is accessible only to the
extracellular side. This is a small region of the protein that is alternately accessible to the outside
and inside of the cell as a result of a conformational change.
The mutation of His721, His837, and His852 to glutamine, or of His752 to serine, inhibited
chloride transport mediated by mouse AE1 expressed in Xenopus oocytes (Muller-Berger et al.
1995). Mutation of Lys558 to asparagine (Lys558Asn) in combination with any of the
HisGln mutations restored chloride flux. This indicates a possible interaction between Lys558
and these histidine residues. The authors reasoned that the transmembrane helices carrying these
histidines, Lys558 and Glu699 (the residue necessary for chloride flux and homologous to
human Glu681) could form an access channel lined with histidine and glutamate residues. Even
with evidence for involvement of specific residues in the anion transport mechanism, an active
anion binding site is yet to be identified within mdAE1.
1.4.2 Topology
mdAE1 spans the red cell membrane up to 12 times (Popov et al. 1997) and is N-glycosylated at
Asn642 in extracellular loop 4 (Jay 1986, Tanner et al. 1988). Proteolysis studies have
determined protease-sensitive sites which are exposed to either the cytoplasmic or extracellular
side of the cell membrane (Kang et al. 1992). For example, cleavage at Lys360 helped determine
its intracellular location and boundary between the cytoplasmic and transmembrane domains.
Tyr553 was cleaved by chymotrypsin on the outside of the cell, positioning this residue on the
extracellular surface. Proteolysis studies also helped determine which residues were located in
surface-exposed loops and others which were part of TM domains (Hamasaki et al. 1997).
Chemical labeling studies with non-permeant reagents have also helped determine
localization of specific residues within the folded structure of AE1. The anion transport inhibitor
eosin-5-maleimide was found to react with Lys430 (Cobb and Beth 1990) which, according to
hydropathy plots, was localized to the first extracellular loop connecting TM domains 1 and 2.
As mentioned above, the membrane-impermeable inhibitor H2DIDS reacted with Lys539,
positioning this residue on the extracellular surface (Okubo et al. 1994). One consideration in
these studies is that the reagents are often quite small and could enter an access channel, reacting
with a residue close to the cytoplasmic side of the membrane. Glu681 is an example of such a
residue. The crystal structures of numerous polytopic membrane proteins (White 2010) have
25
revealed that many channels and transporters contain long tilted helices, kinked helices and re-
entrant loops beyond the typical 20 amino acid hydrophobic span seen in proteins such as
rhodopsin. While valuable, topology models do not provide the necessary detail to illuminate the
mechanism of action of membrane proteins. Scanning N-glycosylation mutagenesis has been
used to accurately map the ends of the TM segments of AE1 (Popov et al. 1997). In this
technique, N-glycosylation acceptor sites are introduced throughout the protein followed by
detection of glycosylation. For an extracellular loop to become efficiently glycosylated, the loop
had to be larger than 25 residues in size. Acceptor sites had to be located more than 12 residues
away from the preceding TM segment and more than 14 residues away from the next TM
segment. This 12 + 14 rule was used to localize the ends of TM segments in combination with
hydropathy analysis. This technique has been used to define the extracellular limits of TMs 7 and
8 that limit the loop containing the endogenous N-glycosylation site (Groves and Tanner 1999),
the re-entrant loop between TMs 9 and 10 (Kanki et al. 2002) and the limits of TM 1 and 2 in
normal AE1 and AE1SAO (Cheung et al. 2005b).
Scanning cysteine accessibility mutagenesis (SCAM) has also been used to establish
membrane topology (Zhu and Casey 2007). In this technique, individual cysteine residues are
introduced by mutagenesis into a cysteineless mutant of AE1. The intracellular or extracellular
location of each new cysteine is determined based on its reactivity to sulfhydryl-directed
reagents, such as biotin maleimide. Using this technique, 80 cysteine residues were introduced
between Phe806 and Cys885 (Zhu et al. 2003). These residues were systematically localized to
extracellular loops, intracellular loops, and to the TM domains.
Natural variations or polymorphisms in the AE1 sequence create blood group antigens that
are exposed to the outside of the red cell (Tanner 1997) which have also assisted in establishing
membrane topology. For example, the Diego blood group is created by a P854L polymorphism
located in the extracellular loop connecting the last two TM segments (Salhany et al. 1996). The
results of the above studies to determine AE1 topology are displayed in the folding model of
AE1 in Figure 1.6.
26
Figure 1.6: Topographical model of human erythroid AE1.
The kidney AE1 start site at Met66 is indicated with an arrow. Extracellular blood group
antigens are shown in blue. Mutations associated with HS and ovalocytosis (Δ400-408) are
shown in orange. Mutations associated with hereditary stomatocytosis and xerocytosis are shown
in red. Mutations associated with dRTA are shown in green and are only found in the C-terminal
domain. AE1 is N-glycosylated at Asn642. The probable site of limited trypsin digestion, K174,
is indicated. Modified from Alper (2009). HS: hereditary spherocytosis; HSt: hereditary
stomatocytosis; dRTA: distal renal tubular acidosis.
N-glycosylation
Diego antigen
Wright b antigen
Probable site of
limited trypsin digestion
27
Electron microscopic analysis of two-dimensional crystals has shown that mdAE1 (residues
361 to 911) reconstituted with lipids exists as a dimer of two monomers related by two-fold
symmetry (Wang et al. 1993). This structure shows that the membrane domain alone maintains
the native dimeric structure of the intact protein and the cytoplasmic domain is not necessary for
dimer formation, confirming earlier biochemical studies (Casey and Reithmeier 1991). Three-
dimensional (3D) crystals of the mdAE1 have been grown that diffract to low resolution (14 Å)
(Lemieux et al. 2002). Deglycosylation increased homogeneity and inhibitor binding increased
the conformational stability of the protein. The types of detergents used and the amount of
phospholipids co-purifying with mdAE1 were critical to the formation of 3D crystals. More
recent work using electron microscopy has revealed some details of the angles of the α-helical
segments in mdAE1 (Yamaguchi et al. 2010). The three-dimensional structure of the outward-
open conformation of mdAE1 was solved at 7.5 Å resolution and several long and tilted helices
were recognized. V-shaped densities near the centre of the dimer were observed. Even with this
higher resolution mdAE1 structure, further investigation is needed to determine the mechanism
of anion exchange.
1.5 Cytoplasmic domain of AE1 (cdAE1)
1.5.1 Structure
The crystal structure of cdAE1 encompassing residues 1 to 379 was solved to 2.6 Å resolution at
pH 4.8 in the laboratory of Dr. P.S. Low (Zhang et al. 2000a) and is shown in Figure 1.7. A tight
symmetric dimer was formed by cdAE1 and was stabilized by interlocked dimerization arms,
which encompassed the C-terminal part of the domain comprising residues 304 to 357. Within
the dimerization domain, eight backbone-to-backbone hydrogen bonds formed between two
strands of the intermonomeric antiparallel β-sheet. As well, a hydrophobic core of nine
interacting leucine residues stabilized the dimer. A globular domain spanning residues 55 to 290,
which is the peripheral protein-binding domain, extended away from the dimerization arm of
each subunit. A short helix-and-loop segment, residues 291-303, connected the two domains.
Residues 1 to 54 (extreme N-terminus), 202 to 211, and 357 to 379 (C-terminus which links to
the membrane domain) were not observed in the crystal structure, presumably due to segmental
flexibility. A cysteine cluster, made up of Cys201 and Cys317 from each dimer subunit, was
28
Figure 1.7: Crystal structure of the cytoplasmic domain of AE1 (cdAE1). The symmetric dimer of cdAE1 was crystallized and its structure determined by X-ray
diffraction (Zhang et al. 2000). The two subunits are coloured in grey and green. The locations
of the HS mutations G130R and P327R are indicated. Residues 1-54 (red dotted line) were not
resolved in the structure. The E40K HS mutation is located here and its approximate location is
indicated on the structure. Residues 1-65 (red) are missing in kAE1. Nt, amino-terminus; Ct,
carboxy-terminus.
G130R
P327R
E40K
Nt Ct
29
located near the dimer interface. Interdomain and intradomain disulfide bond formation is
possible under experimental oxidizing conditions, but the cysteines in cdAE1 normally exist as
free thiols in red cells (Low 1986). The subunits of the dimer appear to swap dimerization arm
domains, such that the arm from one subunit becomes more associated with the globular domain
of the other subunit, and vice versa. This domain swapping may create protein-binding surfaces
that can only exist in the dimer, thereby necessitating dimer formation for proper cdAE1
function.
The cytoplasmic domain of AE1 purified from erythrocytes at physiological pH was found
to exist as a dimer in solution (Appell and Low 1981). This was determined after unmodified
cdAE1 and disulfide cross-linked cdAE1, both proteolytically released from erythrocyte
membrane vesicles, eluted in the same peak fraction from a gel filtration column. As well, the
molecular weight of the native fragment as determined by sedimentation velocity (SV)
experiments was approximately the same as that of the cdAE1 dimer. Recombinant cdAE1
expressed in, and purified from, E. coli was also shown to exist as a dimer in solution as
determined by calibrated gel filtration (Wang et al. 1992a), which supports the validity of the
dimeric form observed in the crystal structure solved at low pH. In fact, cdAE1 remains a dimer
over a large pH range, but exhibits a pH-dependent conformational change (Zhou and Low
2001). This is seen by a more-than-doubling of its intrinsic fluorescence as the pH is raised from
pH 6 to 11 and an increase in Stokes radius from 51 Å to 62 Å without a change in secondary
structure. Four tryptophan residues (Trp75, Trp81, Trp94 and Trp105) are located relatively
close to each other in the primary sequence of each subunit. A hydrogen bond between Trp105
of one subunit and Asp316 of the other subunit that is present in the lower pH conformation is
broken in a higher pH conformation of cdAE1 as the peripheral protein-binding domain moves
away from the dimerization arm. The crystal structure represents the low pH conformation and is
more compact, while increasing the pH opens up the structure.
Site-directed spin labeling (SDSL) in combination with electron paramagnetic resonance
(EPR) and double electron-electron resonance (DEER) spectroscopies was used to study the
solution structure of cdAE1 at neutral pH (Zhou et al. 2005). Single cysteine mutants of cdAE1
were constructed from a cysteineless mutant followed by SDSL of the single cysteines. Solvent
accessibility of labeled residues was performed using a paramagnetic broadening agent. This
30
technique confirmed the α-helical structure and solvent accessibility of regions in the peripheral
protein-binding domain and in the dimer interface that were consistent with the crystal structure.
As well, intersubunit distances between spin labels on various residues in the dimerization arm
agreed with distances found between these residues in the crystal structure. These results indicate
that the compact crystal structure solved at low pH may, in fact, represent the physiological form
of cdAE1.
1.5.2 Physiological function
The cytoplasmic domain of AE1 functions as a site of protein interaction at the internal surface
of the red cell membrane. Protein binding functions to regulate glycolysis and oxygen binding to
hemoglobin, and to maintain the structure and stability of the red cell. The extreme N-terminus
of cdAE1 is extraordinarily acidic and the N-terminal methionine is N-acetylated (Drickamer
1978, Kaul et al. 1983). This region interacts directly with glycolytic enzymes,
deoxyhemoglobin, hemichromes, protein 4.1 and ankyrin (Low 1986, Willardson et al. 1989).
Two tyrosine residues present in this region (Tyr8 and Tyr21) become phosphorylated by
tyrosine kinases in red cells, and ankyrin binding to cdAE1 blocks the phosphorylation of Tyr8
(Willardson et al. 1989). Phosphorylation of Tyr8 regulates the binding and activity of glycolytic
enzymes, and consequently, glycolysis in the red cell (Harrison et al. 1991). This region of AE1
binds to the active sites of glycolytic enzymes like GAPDH, inhibiting their activity.
Phosphorylation of Tyr8 by the combined action of Syk and Lyn kinases releases glycolytic
enzymes, allowing glycolysis to proceed. Indeed, recent evidence has suggested that glycolytic
enzymes that bind to cdAE1 exist as a complex on the cytoplasmic surface of the red cell
membrane, anchored there by the association with AE1 (Campanella et al. 2005).
The extreme N-terminal 11 residues of cdAE1 bind to the central cavity between the β-
chains of deoxyhemoglobin (Walder et al. 1984). The binding site extends deep into the cavity
and includes most of the basic residues within the 2,3-diphosphoglycerate (2,3-DPG) binding
site. Binding of cdAE1 to deoxyhemoglobin lowers its oxygen affinity and increases the Bohr
effect. This has the same effect that binding of 2,3-DPG has on hemoglobin, which is to lower
the affinity for oxygen and thereby transfer it from red blood cells to respiring tissues. In the
lungs, the opposite occurs where cdAE1 affinity for hemoglobin decreases and oxygen inhaled
from the air must bind to hemoglobin to be delivered to the tissues for oxidative metabolism.
31
An important function of cdAE1 is to help stabilize and strengthen the red cell membrane. It
does this through its interaction with the cytoskeleton. cdAE1 binds directly to ankyrin and
protein 4.2, which in turn bind to spectrin, creating a physical link between the red cell
membrane and the cytoskeleton. Details of the protein 4.2 interaction with cdAE1 will be
discussed in Section 1.6. The ankyrin binding site on AE1 has been localized to two regions of
cdAE1: the extreme N-terminus and a β-hairpin loop in a hinge region made up of residues 175
to 185 (Chang and Low 2003, Willardson et al. 1989). Binding of ankyrin to AE1 induces the
oligomerization of AE1 to a tetramer. The AE1 tetramer’s rotational mobility is restricted as a
result of its association with the cytoskeleton (Van Dort et al. 1998). The association between
ankyrin and AE1 has been reported to occur in the ER or in the first compartment of the Golgi
apparatus. In addition to the role of ankyrin in anchoring AE1 to the cytoskeleton, it also appears
to be involved in the exit of AE1 out of the ER (Gomez and Morgans 1993). An interesting
model has been proposed for the association of chicken AE1 with the cytoskeleton using
circulating erythroid cells from chicken embryos (Ghosh et al. 1999). AE1 dimers assemble in
the ER and traffic out to the plasma membrane. Then they are endocytosed and traffic back to the
Golgi where they form tetramers, bind ankyrin and return to the plasma membrane to lock into
the cytoskeleton.
1.5.3 HS mutations in cdAE1 affecting levels of protein 4.2
As mentioned in Section 1.2, three mutations in cdAE1 are associated with HS and a decrease in
protein 4.2 in the red cell while maintaining a normal amount of AE1. These mutations are E40K
(Band 3 Montefiore), G130R (Band 3 Fukuoka) and P327R (Band 3 Tuscaloosa). The mutations
are associated with 88 % (homozygous state), 55 % (homozygous state), and 29 % (heterozygous
state) decreases in the amount of protein 4.2 in the red cell, respectively, as determined from one
HS patient each (Inoue et al. 1998, Jarolim et al. 1992a, Rybicki et al. 1993). The Band 3
Tuscaloosa mutation has been found to occur in conjunction with the Memphis I mutation
(K56E) in one individual with HS (Jarolim et al. 1992a). Since Band 3 Memphis I is an
asymptomatic variant that represents a widespread polymorphism found in hematologically
normal subjects (Jarolim et al. 1992b) it is unlikely that this mutation contributes to the
phenotype seen in this patient. Studies where these three mutations were first characterized
revealed that AE1 in IOVs prepared from patient red cells containing these mutations have a
lower binding capacity for protein 4.2. However, denaturing conditions were used to purify
32
protein 4.2 and to strip peripheral proteins from IOVs. As well, there were large variations in the
binding results of some of the experiments, perhaps due to the harsh purification conditions. For
these reasons, a true comparison of protein 4.2 binding by these HS mutants and wild-type AE1
has yet to be performed.
The three HS mutations are included in Table 1.2 and are indicated in the crystal structure of
cdAE1 in Figure 1.7. Interestingly, the three mutations do not cluster into a small region of
cdAE1, but rather are quite far apart in the structure. The E40K mutation is located in the
unstructured N-terminal region of the protein so its approximate location is indicated. The
G130R mutation is located in a prominent α-helix in the globular protein-binding domain. The
P327R mutation is located at the N-terminus of the α-helix of the dimerization arm.
As mentioned earlier in this section, the X-ray structure of cdAE1 reveals a domain swap
between monomer subunits, where the dimerization arms intertwine and become associated with
the globular domain from the other subunit. In fact, Pro327 (the site of the HS mutation P327R)
from one subunit occurs on the same face of the dimer as Glu40 and Gly130 (sites of HS
mutations E40K and G130R) from the other subunit. It may be that these three sites make up part
of the large binding surface for protein 4.2, which would explain why mutations at these sites
cause deficiency of protein 4.2 in the red cell.
AE1 species sequence alignment shows that Glu40 is conserved in mouse and rat while the
conservative substitution of aspartate, retaining the negative charge, occurs in chicken, trout and
zebrafish. Another conservative substitution of glutamine occurs in cattle, but with a loss of the
negative charge. The high conservation of this negative charge probably reflects the importance
of this acidic region in the binding of several red cell proteins, including glycolytic enzymes,
hemoglobin and protein 4.2. Gly130 is conserved in mouse and rat but is replaced with alanine in
chicken and trout, and by serine in zebrafish, thereby retaining a small amino acid side chain in
this position. Pro327 is perfectly conserved in mouse, rat, cow and chicken, but a glutamine
occurs in this position for trout and zebrafish. The conservation in mammals and birds reflects
this residue’s importance as well. The protein 4.2 binding site on cdAE1 has not been
determined, but the deficiency of protein 4.2 in red cells that have these mutations in AE1
suggests these three residues may be directly involved in the interaction. Alternatively, these
mutations could cause a conformational change in cdAE1 resulting in loss of protein 4.2 binding.
33
1.6 Protein 4.2
1.6.1 Background
Protein 4.2 is a 72 kDa cytoskeletal protein that resides on the inner surface of the plasma
membrane of red cells. It makes up 5 % of the total membrane protein of the red cell and is
present at about 200 000 copies per cell (Branton et al. 1981). The main role of protein 4.2 is in
maintaining the structural integrity of the red cell membrane through interactions with plasma
membrane and cytoskeletal proteins (Tanner 2002). The gene for mouse protein 4.2 was found to
co-localize on chromosome 2 with the gene for the mouse pallid mutation, which serves as a
mouse model for human platelet storage pool deficiencies (White et al. 1992). Pallid was
believed to be a mutation in the mouse protein 4.2 gene, and for this reason protein 4.2 was given
the name pallidin. The mouse pallid mutation and the gene for mouse protein 4.2 have since been
discovered to reside at distinct loci (Gwynn et al. 1997).
Full-length human protein 4.2 cDNA was cloned and sequenced from a human reticulocyte
expression library (Korsgren et al. 1990, Sung et al. 1990). The complete amino acid sequence
of protein 4.2 was derived from the nucleotide sequence and was found to be homologous to
transglutaminase (TG) protein sequences (Korsgren et al. 1990, Sung et al. 1990).
1.6.2 Transglutaminase family of enzymes
Transglutaminases are calcium-dependent cross-linking enzymes that carry out various
biological functions, including blood coagulation, skin-barrier formation and extracellular-matrix
assembly (Lorand and Graham 2003). TGs are found in many different types of species from
humans and other mammals, chicken and fish, all the way down to insects, invertebrates and
slime mould. TGs catalyze the formation of covalent γ-glutamyl-ε-lysine cross-links between
glutamine acyl-donors and lysine acyl-acceptors. These cross-links result in either the
polymerization of proteins or the covalent linking of proteins which were already reversibly
bound by non-covalent bonds. TGs are also involved in other reactions, such as the mediation of
post-translational modifications by deamidation and amine incorporation, and protein
esterification.
There are eight catalytically active members of the human TG family (TG1 – TG7, and
factor XIIIa) and protein 4.2 is considered to be the catalytically inactive ninth member. Protein
34
4.2 shares sequence identities with TG family members from 23 % (TG1) to 33 % (TG2). A
critical cysteine in the active site of TGs, necessary for catalysis, is substituted by alanine in
protein 4.2 (Korsgren et al. 1990). When TG binds calcium, a large conformational change takes
place exposing the active site (Bergamini 1988). GTP binding by TG negatively regulates its
cross-linking activity (Murthy et al. 1999). ATP has a similar, but less pronounced, effect.
Protein 4.2 was shown to lack cross-linking activity in red blood cell ghosts and IOVs in the
absence or presence of calcium (Korsgren et al. 1990). Also, no calcium-stimulated cross-linking
activity was detected by protein 4.2 in solution. Protein 4.2 has two highly conserved glutamate
residues (Glu439, Glu444) that are known to bind calcium in other TG family proteins
(Satchwell et al. 2009). The spectrin-binding tryptic peptide of protein 4.2 encompassing
residues 380-462, which contains the conserved glutamate residues, has been shown to bind
calcium in vitro (Korsgren et al. 2010), but the effect of calcium on full-length protein 4.2 is not
known.
Protein 4.2 shares over 30 % sequence identity with the majority of the TG family members,
which suggests a clear evolutionary relationship (Murzin et al. 1995). In fact, the genes for
protein 4.2, TG5 and TG7 are arranged in tandem on the 15q15.2 region of human chromosome
15 (Grenard et al. 2001). The genes encoding TG2, TG3 and TG6 are found on the 20q11
segment of human chromosome 20. In mouse these six genes are all found on distal chromosome
2. These findings suggest that these genes were duplicated from a single gene followed by
redistribution to two distinct chromosomes in the human genome.
Interestingly, human protein 4.2 shares the highest amino acid sequence identity with TG2
(33 %) which is found, among many other tissues, in human erythrocytes (Chen and Mehta 1999,
Lee et al. 1986). The role of TG2 in red blood cells is still not clear, but there is evidence that it
may be involved in programmed cell death (Sarang et al. 2007). Calcium-activated TG2 in
erythrocytes has been found to catalyze the formation of high-molecular weight polymers which
contain AE1, ankyrin, spectrin, protein 4.1, catalase and hemoglobin (Lorand and Graham 2003).
The main TG2-reactive site on AE1 is Gln30 in the cytoplasmic domain (Murthy et al. 1994).
These polymers have been isolated from the red blood cells of sickle cell anemia (Lorand et al.
1980) and Hb-Köln (Lorand 2007, Lorand et al. 1987) patients. It is believed that TG2 may
contribute to the decreased life span of these erythrocytes by exposing cell surface epitopes that
are recognized by macrophages (Lorand and Graham 2003).
35
The presence of protein 4.2 in these polymers was not examined, but a study using the
purified recombinant mouse proteins cdAE1, protein 4.2 and reticulocyte TG2 found a
relationship between the three proteins (Gutierrez and Sung 2007). In vitro assays revealed that
as calcium concentration increased, TG2-mediated cross-linking of cdAE1 increased, but TG2
binding to cdAE1 decreased. This suggests that the conformational change associated with
calcium-binding by TG2, and necessary for cross-linking activity, causes it to dissociate from
cdAE1. Protein 4.2 stabilized binding of TG2 to cdAE1 and inhibited TG2-mediated cross-
linking. It is not known what the reason or mechanism for this inhibition could be in
erythrocytes. As mentioned above, it is not known whether or not full-length protein 4.2 binds
calcium in vivo or undergoes a conformational change in its presence. Protein 4.2 may prevent
TG2 calcium binding or dissociation from cdAE1, thereby inhibiting cross-linking, by some
other mechanism. The observations offer the intriguing possibility that protein 4.2 could regulate
the mechanical properties of the red blood cell membrane, partly through regulation of TG2-
mediated cross-linking of AE1.
Another suspected role of TG2, independent of its cross-linking activity, is in signaling
(Lorand and Graham 2003). TG2 is able to bind, hydrolyze, and be inhibited by GTP (Lee et al.
1993). TG2 is also able to bind and hydrolyze ATP (Iismaa et al. 1997). Protein 4.2 has a
nucleotide-binding P-loop that binds ATP, but not GTP (Azim et al. 1996), but whether or not it
hydrolyses ATP is not known. It has been proposed that protein 4.2 may assist in creating a
membrane reservoir of ATP needed for membrane transport proteins in the erythrocyte.
TG2 from rat liver was discovered to be identical to a high-molecular weight G-protein,
Ghα, which was shown to mediate the activation of phospholipase C (PLC) by the α1B-
adrenergic receptor in transfected COS-1 cells (Nakaoka et al. 1994). TG2 (Ghα) exists in a
heterodimeric complex with Ghβ (Mhaouty-Kodja 2004), which was shown to be calreticulin
when isolated from rat liver (Feng et al. 1999). Ghβ purified from human heart was also
discovered to be calreticulin by amino acid sequencing. TG2 maintains its signaling function
even when its cross-linking activity has been inactivated by mutation (Lorand and Graham
2003). However, binding of calreticulin to TG2 inhibits both the transglutaminase and signaling
functions of the enzyme (Feng et al. 1999).
36
Calreticulin is a calcium-binding protein that plays an important role in calcium storage in
the ER (Nash et al. 1994). The protein has also been found in the nuclear envelope, the nucleus
and the cytoplasm (Rojiani et al. 1991), where it was found to regulate integrins by interacting
with the cytoplasmic tails of their α-subunits in Jurkat cells (human T-lymphoblastoid cells)
(Coppolino et al. 1995). Since mature red blood cells no longer have the ER, which contains the
majority of the cellular calreticulin, it is unclear as to the existence or purpose of a calreticulin-
TG2 interaction in these cells. Evidence of cytoplasmic calreticulin in the mature erythrocyte
was shown in a study of chaperones in differentiating human CD34+ erythroid progenitor cells.
Terminally differentiated cells with expelled nuclei maintained levels of calreticulin, while its
paralog calnexin, a resident ER protein, was lost (Patterson et al. 2009). TG2 has also been found
to accelerate the differentiation of K562 cells (an erythroleukemia cell line) through activation of
Akt phosphorylation and inactivation of ERK1/2 phosphorylation (Kang et al. 2004), suggesting
a possible role for TG2 in erythrocyte differentiation.
1.6.3 Protein 4.2 isoforms
The human protein 4.2 gene consists of 13 exons and 12 introns (Korsgren and Cohen 1991).
Two cDNA sequences were obtained from human reticulocytes which were 2.4 and 2.5 kilobases
(kb) in length (Sung et al. 1990). The minor, longer isoform (Type I) contains a 90-base pair (bp)
in-frame insertion encoding an extra 30 amino acids near the N-terminus (Sung et al. 1992) and
codes for a 74 kDa protein (Korsgren and Cohen 1991). The major, shorter isoform (Type II) is
created by alternative splicing within exon 1, where 90 nucleotides are removed, resulting in
translation of a 72 kDa protein. The first exon contains the 5′-untranslated cDNA sequence and
the first 10 nucleotides of the coding region, which code for the first three amino acids of both
isoforms (Sung et al. 1992). The 90-bp insert in the longer isoform is adjacent to the first exon
and is followed by the rest of intron 1. The splicing donor site for the short isoform matches the
splice site consensus sequence (Ohshima and Gotoh 1987) better than that of the longer isoform,
and may help explain the greater abundance of short isoform mRNA in reticulocytes and
expressed protein in erythrocytes. The function of the 30 amino acid insert in the long protein
isoform is not known, but there is some speculation that it may affect phosphorylation levels of
membrane skeleton proteins (Sung et al. 1992).
Two other protein 4.2 mRNA splicing isoforms have been described where nucleotides
comprising exon 3 are removed from the Type I and Type II isoforms (Bouhassira et al. 1992,
37
Korsgren and Cohen 1991). These additional isoforms, designated Type III and IV, have not
been detected at the protein level. Interestingly, only one protein 4.2 mRNA and one protein
isoform have been detected in mouse reticulocytes, which correspond to the Type II isoform in
human (Rybicki et al. 1994). Amino acid numbering of protein 4.2 in this thesis is according to
the Type II isoform unless otherwise indicated.
1.6.4 Tissue distribution and expression in different species
Immunoreactive forms of protein 4.2 have been detected in association with the plasma
membranes of non-erythroid cells and tissues in human and pig, including platelets, brain and
kidney (Friedrichs et al. 1989). In these non-erythroid human tissues, the short 72 kDa isoform,
which is the most common in erythrocytes, was detected. These cells also contain isoforms of the
erythrocyte cytoskeleton proteins spectrin, protein 4.1 and ankyrin (Bennett 1979, Bennett 1985,
Moon and McMahon 1987). This indicates that protein 4.2 may have similar roles in membrane
stabilization in non-erythroid cells.
In another study, immunoreactive protein 4.2 was detected in human lymphocytes, platelets,
spinal cord and bovine brain tissue (Schwartz et al. 1987). In lymphocytes and platelets, protein
4.2 was detected as the shorter isoform and was present mostly in cell membranes. In spinal cord
and brain, protein 4.2 was found in the cell cytosol and membrane fractions, and brain protein
4.2 was detected as a doublet of the short and long isoforms. Protein 4.2 was immunodetected in
bovine and chicken eye lenses and erythrocytes (Sung and Lo 1997). The protein 4.2 detected
corresponded to the minor, long isoform found in human erythrocytes. The lack of detection of
the major, short isoform was attributed to the species-specific nature of the anti-protein 4.2
antibody used. Other major proteins involved in the erythrocyte membrane-cytoskeletal linkage –
AE1, ankyrin, spectrin, actin and protein 4.1 – have been immunodetected in human eye lens
tissue (Allen et al. 1987). Because these proteins appear and disappear together during lens
maturation, it is suggested that an erythrocyte-like membrane structural organization may exist in
these cells.
In contrast to studies where protein 4.2 was detected in various tissues of different species,
another study suggested that protein 4.2 in the developing mouse is restricted to erythroid cells
(Zhu et al. 1998). The authors detected protein 4.2 mRNA in the erythroid cell-producing organs
38
and circulating erythrocytes during embryonic development and in adult mice, but not in other
tissues.
While TGs are found in a wide range of species, from humans to invertebrates, protein 4.2
appears to be the younger, catalytically inactive cousin detected only in mammals and birds. A
Brazilian study collected and lysed red cells from 44 different mammalian species (Guerra-
Shinohara and Barretto 1999). Red cell ghosts were solubilized and run on SDS-PAGE gels
followed by Coomassie Blue staining. Protein 4.2 was detected in human, monkey, gorilla, seal,
members of the cat and dog families, raccoon, mouse, rat, hamster, rabbit, sloth, elephant, camel,
giraffe, deer, sheep, cattle, goat, tapir, dolphin, bat, manatee and opossum. Protein 4.2 was not
detected in guinea pig, swamp rat, or horse. However, this staining method may not have been
sensitive enough to detect low levels of protein 4.2, which a more sensitive immunodetection
technique might have done.
In addition to the human and mouse cDNA sequences discussed previously, the bovine
erythrocyte cDNA sequence for protein 4.2 has been determined (NCBI Reference Sequence:
NP_776737.1). As well, cDNA sequences similar to erythrocyte protein 4.2 have been
determined for chimpanzee (XP_001156126.1), Norway rat (XP_342504), dog (XP_851181),
horse (XP_001500600), opossum (XP_001364606), and chicken (XP_417393).
Studies in protein 4.2 knock-out mice showed the mice to have mild HS, about 30 %
decrease in AE1 (probably due to loss of membrane by vesiculation), normal spectrin and
ankyrin content, and an intact cytoskeleton (Peters et al. 1999). AE1-mediated ion transport was
decreased in these mouse erythrocytes by about 40 %, a value close to the decrease in AE1
protein, which indicates that the decrease in ion transport may be a function of fewer AE1
proteins.
1.6.5 Structure and properties
Human protein 4.2 is a 691 amino acid protein with unusual solubility and membrane-binding
properties. It’s a peripheral membrane protein, versus an integral membrane protein, yet requires
harsh conditions for removal from IOVs, such as pH of 11 or higher (Steck and Yu 1973). It is
also conformationally unstable in physiological salt solutions and requires low concentrations of
detergent to stay in solution (Dotimas et al. 1993). Protein 4.2 has an N-terminal glycine
39
following the initiating methionine, the latter being removed co-translationally (Korsgren et al.
1990, Sung et al. 1990). The N-terminal glycine has been shown to be N-myristoylated (Risinger
et al. 1992). This modification may partially account for the protein’s unusual solubility. N-
myristoylation is an irreversible, co-translational modification important for protein association
with cell membranes, which is required for proper protein localization or function (Wright et al.
2009).
N-myristoylated protein 4.2 (Type II) was found to localize to the plasma membrane when
expressed in Sf9 insect cells (Risinger et al. 1996). No myristoylation was detected in the G2A
mutant of Type II protein 4.2, Gly2 being the site of the modification. The G2A mutant was
detected in an intracellular compartment, indicating its inability to associate with the plasma
membrane. Myristoylation of the minor, unspliced protein 4.2 isoform (Type I) was barely
detectable and this protein was also localized intracellularly. The fifth residue of a protein
substrate helps determine the likelihood of myristoylation by N-myristoyl transferase (NMT),
with small, uncharged residues being more favorable. Type II protein 4.2 has a glycine in this
position, while Type I has a proline here, which results in it being a poor substrate (Towler et al.
1988). Mouse protein 4.2, which corresponds to the shorter human isoform, was myristoylated to
similar levels as the human Type II protein 4.2, and was also found at the plasma membrane. The
mouse protein has a serine at the fifth position, making myristoylation favorable. Similar results
were seen in COS 7 cells, except for the G2A mutant, which was not able to be expressed in
these cells. In subcellular fractionation experiments, homogenized Sf9 cells were spun to
separate the soluble and particulate fractions. Human protein 4.2 (Type II) and the G2A mutant
were both found to be associated with the particulate fraction, which contains the cytoskeleton.
This indicates that the myristoylation defect in the G2A mutant does not affect its ability to
interact with proteins present in the particulate fraction.
Protein 4.2 is also palmitoylated at Cys173 (Das et al. 1994). S-palmitoylation is a reversible
post-translational modification where a palmitoyl group is added to a cytoplasmic cysteine
residue by palmitoyl-acyl transferases (PATs) (Bijlmakers and Marsh 2003). Other proteins that
are palmitoylated in human erythrocytes include AE1, ankyrin, protein 4.1, and a subpopulation
of spectrin (Das et al. 1994). Palmitoylation, like myristoylation, adds an extra hydrophobic
moiety to a protein, important for hydrophobic protein-protein or protein-membrane interactions
(Bijlmakers and Marsh 2003). There is no known consensus sequence for palmitoylation, since a
40
variety of amino acids can influence the modification. However, it may be the recognition of a
common structural feature on a protein that allows it to be palmitoylated. In peripheral
membrane proteins, such as protein 4.2, cysteines that are close to membrane-associated domains
seem to be preferred for palmitoylation, possibly because of their proximity to membrane-
localized PATs. Protein 4.2 could be depalmitoylated using hydroxylamine, which indicated a
thioester linkage (Das et al. 1994). Depalmitoylated protein 4.2 displayed decreased binding to
protein 4.2-stripped IOVs compared to native protein 4.2.
The three-dimensional structures of several TGs have been solved and can be used as
templates in creating a homology model of protein 4.2. A homology model of protein 4.2 is seen
in Figure 1.8 next to the crystal structure of cdAE1 to show their relative sizes. This homology
model of protein 4.2 was created using human transglutaminase 2 (TG2) as a template instead of
sea bream TG, as was previously done (Toye et al. 2005), since it is a human homologue, is
found in erythrocytes, has the greatest sequence similarity to protein 4.2 and the mouse TG2
interacts with mouse cdAE1 and protein 4.2 (Gutierrez and Sung 2007). The orientation of our
model is turned 180° relative to the previous one so that the palmitoyl group at Cys173 points
upwards towards the conventional placement of the plasma membrane. The published crystal
structure of cdAE1 was also oriented this way so our protein 4.2 model was made to match this
orientation. The orientation of the model is such that the palmitoyl group that attaches to Cys173
would point upwards towards the plasma membrane. Cys173 is located at the top of a predicted
β-hairpin loop that has been found to be important for binding to cdAE1.
1.6.6 Interactions with AE1, ankyrin and spectrin
Protein 4.2’s main point of interaction at the membrane is cdAE1 (Korsgren and Cohen 1986). A
fraction of AE1 extracted from red cells is associated with protein 4.2 (Yu and Steck 1975).
Protein 4.2 purified from red cells bound to protein 4.2-depleted IOVs, and purified cdAE1
competed for this binding. As well, when the cytoplasmic domain was cleaved from AE1 in
these IOVs by mild trypsin treatment, protein 4.2 binding was almost completely abolished.
Purified protein 4.2 was also found to bind strongly to cdAE1 in vitro (Korsgren and Cohen
1988). The presence of AE1 is crucial for the maintenance of protein 4.2 in the red cell, since
AE1-deficient red cells in human (Ribeiro et al. 2000), mouse (Peters et al. 1996), and cow
41
Figure 1.8: Crystal structure of the cytoplasmic domain of AE1 (cdAE1) and homology
model of protein 4.2.
The symmetric dimer of cdAE1 (left) was crystallized and its structure determined by X-ray
diffraction (Zhang et al. 2000a). The sites of the HS mutations are indicated on the structure and
the residues missing in kAE1 are shown in red. Residues 1-54 (red dotted line) were not resolved
in the structure. The structure of cdAE1 is shown alongside the homology model of protein 4.2
(right). The protein 4.2 structure was created with SWISS-MODEL (Arnold et al. 2006) using
human TG2 (Liu et al. 2002) as the template. The regions in blue and green of the protein 4.2
model represent the 23 kDa N-terminal cdAE1-binding region. Residues 33-45 are coloured in
light green and are predicted to form a β-strand. Residues 157-181 are coloured in dark green,
and encompass the predicted β-hairpin loop containing the palmitoylatable Cys173 at the top.
Structures of cdAE1 and protein 4.2 are on the same scale. aa: amino acids; Nt: amino-terminus;
Ct: carboxy-terminus.
cdAE1 protein 4.2 homology
model
42
(Inaba et al. 1996) are also deficient in protein 4.2. This relationship is not interdependent, since
the absence of protein 4.2, either in mouse knock-outs or HS patients with protein 4.2 mutations,
does not always cause a deficiency in AE1. However, in protein 4.2-deficient red cells, AE1 has
a weaker interaction with the cytoskeleton since it is more easily extracted from cells and has
increased lateral diffusion (Rybicki et al. 1996). This observation points to the role of protein 4.2
as a strengthener of membrane-cytoskeleton linkage, via AE1 interaction.
In a study to determine its cdAE1-binding region, partial digestion of protein 4.2 with
staphylococcal V8 protease and blot-overlay using biotinylated cdAE1 was performed
(Bhattacharyya et al. 1999). The N-terminal region of protein 4.2 encompassing residues 1-238
was found to represent the cdAE1-binding domain. The interaction of protein 4.2 glutathione S-
transferase (GST)-fusion proteins with purified cdAE1 was then performed by blot-overlay.
Residues 157-181 of protein 4.2, which contain the palmitoylatable Cys173, were found to be
critical for the interaction. This binding region is predicted to form a β-hairpin in the centre of
folded protein 4.2 with Cys173 located at the tip of the hairpin shown in dark green in Figure
1.8. Another study used biotinylated protein 4.2 peptides in binding assays with cdAE1 isolated
from red blood cells (Rybicki et al. 1995). Amino acids 33-45 of protein 4.2 were found to
contain a binding site. Arg34 and Arg35 were essential for the interaction and the authors
reasoned that this basic Arg-Arg motif may bind to an acidic region, such as the extreme acidic
N-terminus of cdAE1. Since this region of cdAE1 contains the HS mutation E40K, it is possible
that going from an acidic glutamate residue to a basic lysine residue is enough to repel the basic
Arg-Arg motif of protein 4.2 that may bind there. This Arg-Arg motif is also perfectly conserved
in all nine human TGs, as well as across species from human to mouse, dog and cow. The
conservation of this motif in protein 4.2 suggests it may have an important role, possibly
influencing the protein 4.2-AE1 interaction.
There is also evidence that protein 4.2 associates with ankyrin in red cells. Protein 4.2,
ankyrin and AE1 can be co-immunoprecipitated in a complex from red cells (Bennett and
Stenbuck 1980). As with AE1 deficiencies, ankyrin deficiencies in red blood cells also cause a
decrease in protein 4.2 (Yawata 1994). Su et al. (2006) used far-western blots and pull-down
assays using GST-fusion proteins derived from protein 4.2 to reveal an ankyrin-binding site
located within residues 157-170 of protein 4.2 (Su et al. 2006), which is the same region found to
43
bind to cdAE1 in the blot-overlay experiments performed by Bhattacharyya et al. (1999) above,
and the one predicted to form a β-hairpin loop. However, the study by Su et al. (2006) did not
see an obvious interaction between the fusion protein containing these residues of protein 4.2 and
cdAE1. Korsgren and Cohen (1988) also showed an interaction between protein 4.2 and ankyrin,
both purified from red blood cells (Korsgren and Cohen 1988).
Another protein involved in the main linkage of the red cell membrane to the cytoskeleton is
spectrin. Protein 4.2 was found to bind spectrin in solution and to promote the binding of spectrin
to ankyrin-stripped IOVs (Golan et al. 1996), which suggests an additional mode of membrane-
cytoskeleton stabilization by this protein. In another study using blot overlays, Mandal et al..
(2002) showed that biotinylated spectrin was able to bind to full-length protein 4.2 and a 30 kDa
fragment of proteolysed protein 4.2 whose N-terminus began at Gly239 (Mandal et al. 2002).
Spectrin was found to interact with a protein 4.2 peptide comprising residues 440-462 in
solution, and this peptide was able to inhibit the interaction between spectrin and full-length
protein 4.2. This region of protein 4.2 is a highly charged stretch which is predicted to form an α-
helix.
The AE1-ankyrin-spectrin-protein 4.2 complex has been long accepted to comprise the
major link between the membrane and cytoskeleton, with the complex located at one end of the
spectrin tetramer. However, it was recently found that protein 4.2 was able to bind to the other
end of the spectrin tetramer, the end associated with actin. Actin binds to the N-terminus of β-
spectrin, and the adjacent C-terminal end of α-spectrin has a calmodulin-like domain called the
EF-domain. Korsgren et al.. (2010) found that the EF-domain of spectrin bound native and
recombinant protein 4.2 in pull-down assays (Korsgren et al. 2010). This study also showed
inhibition of full-length protein 4.2 binding to the EF-domain of spectrin using a protein 4.2
peptide encompassing residues 380-462. Figure 1.9 shows a diagram of the protein 4.2
polypeptide with important regions of known locations mapped.
1.6.7 Interactions with the Rh complex and CD47
The Rh (Rhesus) complex is a group of membrane proteins that interact with red cell cytoskeletal
proteins and provide another linkage point between the membrane and cytoskeleton (Van Kim et
al. 2006). This linkage helps maintain the structural integrity of the cell. The RH system is a
highly immunogenic and polymorphic blood group system. The complex is composed of the Rh
44
Figure 1.9: Diagram of the protein 4.2 polypeptide with important regions mapped. Protein 4.2 is shown from Met1 to A691, with the dotted line denoting a break in the sequence
between residue 462 and the C-terminal amino acids. Important regions with known sequence
locations are indicated. The region of protein 4.2 homologous to the TG active site (266-270) is
inactive in protein 4.2. The region of protein 4.2 homologous to the TG Ca2+
-binding domain
(423-432) is shown, but Ca2+
-binding by full-length protein 4.2 has not been demonstrated.
45
protein (D, CcEe), Rh-associated glycoprotein (RhAG), LW (also known as Intracellular
Adherence Molecule 4, or ICAM-4), CD47 and glycophorin B (GPB). Functional studies in
yeast (Marini et al. 2000) and Xenopus oocytes (Westhoff et al. 2002) have implicated RhAG as
an ammonium transporter. Studies in human red blood cells indicate that RhAG mediates
transfer across the membrane of ammonia (Ripoche et al. 2004) and possibly CO2 (Bruce 2008).
CD47 (Integrin Associated Protein, IAP) is a highly glycosylated, five-TM domain protein with
an IgV-like domain at its N-terminus (Van Kim et al. 2006). The integrin-CD47 complex
couples to G proteins to form a signaling complex (Brown 2001). However, integrins are not
expressed in mature erythrocytes (Lindberg et al. 1994), indicating other roles for CD47 in red
cells. One proposed role has been as a marker for self in mature red blood cells, preventing their
clearance by macrophages (Oldenborg et al. 2000).
Evidence for an interaction between protein 4.2 and CD47 comes from studies where red
cells from HS patients deficient in protein 4.2, due to protein 4.2 frameshift deletions, also have a
deficiency in CD47 (Bruce et al. 2002, Mouro-Chanteloup et al. 2003). As well, in red cells with
the AE1 mutation S667F (Band 3 Courcouronnes), both AE1 and protein 4.2 are reduced to
about 35 % of normal levels, while CD47, Rh polypeptides and RhAG are reduced to about 60 %
(Toye et al. 2008). A strong association at the membrane between the Rh complex and the AE1
complex, through CD47 interaction with protein 4.2, creates a strong antigenic entity at the red
cell surface. This may allow for the prolonged survival of red cells by allowing them to evade
macrophages in the immune system by serving as a marker of self. Direct evidence of association
between the AE1 and Rh protein complexes comes from a study where Rh proteins co-
immunoprecipitated with AE1 from red blood cells (Bruce et al. 2003). The authors speculated
that there is a functional role for the association of these complexes into a macrocomplex, such
as a CO2/O2 gas exchange unit (metabolon) in the red cell. Figure 1.10 shows a model of the
AE1 and Rh protein complexes linking the plasma membrane and cytoskeleton. Protein 4.2 is
shown as the protein linking the two complexes through its interactions with cdAE1 and CD47.
46
Figure 1.10: Schematic model of AE1 and Rh protein complexes at the red cell membrane. The integral membrane and cytoskeletal proteins of the AE1 and Rh protein complexes are
shown, as well as those proteins involved in linking the two complexes. Modified from Bruce et
al. (2003).
(AE1) (AE1)
P cdAE1 cdAE1
47
1.6.8 HS mutations in protein 4.2
Eleven protein 4.2 mutations associated with HS have been discovered so far in humans
(Satchwell et al. 2009, van den Akker et al. 2010a). These mutations are summarized in Table
1.3. All of these mutations, in the homozygous state or compound heterozygous state, result in a
complete, or near complete, absence of protein 4.2 in red cells. As mentioned above, deficiency
of protein 4.2 occurs with a concomitant deficiency in CD47. Most of the protein 4.2 mutations
result in premature termination of translation leading to truncated protein products. The
deficiency of protein 4.2 seen is either due to degradation of these protein products or to mRNA
instability, since a reduced number of transcripts is often observed. The rest of the defects are
due to missense mutations resulting in protein products with an amino acid substitution. Protein
4.2 Nippon (see Table 1.3) was the first variant discovered (Bouhassira et al. 1992). This
mutation results in less than 1 % of normal protein 4.2 expression in red blood cells. However,
when this mutant was co-expressed with AE1 in Xenopus oocytes it displayed similar binding to
AE1 compared to wild-type protein 4.2 (Toye et al. 2005). Protein 4.2 Nippon was also able to
localize to the plasma membrane in the absence of AE1 in these cells, similar to wild-type
protein 4.2.
The mechanism of loss of Protein 4.2 Nippon in the red cell, for example, by mRNA or
protein degradation, is unknown. It may be that with successful expression and localization in
red cells, Protein 4.2 Nippon would be functional in that context. In four out of five compound
heterozygotes, Protein 4.2 Nippon is one of the defective alleles. Protein 4.2 Komatsu
(homozygous) and Protein 4.2 Tozeur (compound heterozygous with p4.2 Nippon) (see Table
1.3) displayed impaired binding to AE1 and did not localize to the plasma membrane when
expressed alone or when co-expressed with AE1 in Xenopus oocytes (Toye et al. 2005). This
shows that impaired protein 4.2 binding to AE1 can be associated with HS.
48
Table 1.3: HS mutations in protein 4.2
Protein 4.2
variant
Mutation Abnormal allele Reference
Lisboa ΔG at base 174 or 175 in exon 2
introduces Stop codon
Homozygous (Hayette et al. 1995)
Fukuoka Trp119Stop (Type I) Compound heterozygous with
p4.2 Nippon
(Takaoka et al. 1994)
Nippon Ala142Thr (Type I) Homozygous (Bouhassira et al. 1992)
Komatsu Asp175Tyr (Type I) Homozygous (Kanzaki et al. 1995b)
Notame GA in intron 6 splice donor
site introduces Stop codon
Compound heterozygous with
p4.2 Nippon
(Matsuda et al. 1995)
Tozeur Arg310Gln (Type I) Compound heterozygous with
p4.2 Nippon
(Hayette et al. 1995)
Shiga Arg317Cys (Type I) Compound heterozygous with
p4.2 Nippon
(Kanzaki et al. 1995a)
Chartres I Tyr435 Stop Compound heterozygous with
p4.2 Chartres II
(van den Akker et al.
2010a)
Chartres II ΔA1176-T1177 in exon 9
introduces Stop codon
Compound heterozygous with
p4.2 Chartres I
(van den Akker et al.
2010a)
Hammersmith Arg593Stop (Type I) Homozygous (Bruce et al. 2002)
Nancy ΔG in codon 287 in exon 7
introduces Stop codon
Homozygous (Beauchamp-Nicoud et
al. 2000)
49
1.7 Thesis focus
The hypothesis of this thesis is that HS mutations in the cytoplasmic domain of AE1 cause
impaired protein 4.2 binding. This occurs either because of changes in cdAE1 conformation
resulting in misfolded binding regions, or because of changes in the cdAE1 interaction surface of
the properly folded protein.
1.7.1 Effect of HS mutations on the structure and stability of the cytoplasmic
domain of AE1 (Chapter 2)
The cause of the deficiency of protein 4.2 in red cells carrying the three HS mutations, E40K,
G130R and P327R located in cdAE1 is unknown. I hypothesize that these cdAE1 mutations
cause either gross conformational changes in the domain or changes in the direct binding surface
of cdAE1. Either would result in impaired binding of protein 4.2, leading to the loss of protein
4.2 during red cell differentiation. In Chapter 2, the structure and conformational stability of
purified cdAE1 with these mutations was studied using a variety of biophysical methods. The
results of a similar study, carried out by Biochemistry project student Allison Pang under my
supervision, comparing the structure and conformational stability of cdAE1 with kidney cdAE1
are presented in the Appendix.
1.7.2 Protein 4.2 localization and interaction with wild-type and HS mutants of
AE1 in HEK-293 cells (Chapter 3)
Mutations in protein 4.2 and three mutations in cdAE1 have been shown to cause protein 4.2
deficiency and HS. In the case of Protein 4.2 Tozeur and Komatsu, part of the molecular basis of
HS can be attributed to impaired interaction with cdAE1. I hypothesize that the three cdAE1 HS
mutations result in impaired binding to protein 4.2 leading to HS. This binding impairment may
be caused by conformational changes of cdAE1 with these HS mutations, as addressed in
Chapter 2, or by a change in the interaction surface of properly folded cdAE1. In Chapter 3, the
interaction of protein 4.2 with AE1 and the three HS mutants (E40K, G130R and P327R) was
studied in HEK-293 cells using pull-down assays. Protein 4.2 interaction with kAE1 and
AE1SAO was also examined. The localization of protein 4.2 in the absence or presence of wild-
type and HS mutant AE1 proteins was also studied in transfected HEK-293 cells using
immunofluorescence and confocal microscopy. The plasma membrane localization of wild-type
protein 4.2 and an acylation mutant of protein 4.2, G2A/C173A (GC), were compared in the
50
absence or presence of AE1. The degree of cytoskeletal attachment of wild-type and GC protein
4.2 were compared by subcellular fractionation studies.
51
2 Chapter 2: Structure and stability of hereditary
spherocytosis mutants of the cytoplasmic domain of the
erythrocyte anion exchanger 1 protein
Adapted with permission from Bustos, S.P. & Reithmeier, R.A.F. Structure and stability of
hereditary spherocytosis mutants of the cytosolic domain of the erythrocyte anion exchanger 1
protein. Biochemistry, 45, 1026-1034. Copyright 2006 American Chemical Society
(http://pubs.acs.org/doi/abs/10.1021/bi051692c) (Bustos and Reithmeier, 2006). Jing Li
contributed to this work by the generation of the cdAE1 HS mutants. Jing Li also assisted in the
design of the purification protocol of the His6-tagged cdAE1 proteins. I designed and conducted
the experiments and analyzed all of the results.
2.1 Abstract
The N-terminal cytoplasmic domain of AE1 anchors the cytoskeleton to the membrane. Several
proteins bind to cdAE1, including protein 4.2, a cytoskeletal protein. Three mutations in cdAE1
are associated with HS and decreased levels of protein 4.2 in erythrocytes. In this study these
cdAE1 mutants (E40K, G130R and P327R) were expressed and purified from Escherichia coli.
Sedimentation equilibrium (SE) experiments using the analytical ultracentrifuge showed that the
wild-type and mutant proteins are dimers. No difference in secondary structure between mutant
and wild-type proteins was detected using CD analysis. The wild-type and mutant proteins
underwent similar pH-dependent conformational changes when monitored by intrinsic
tryptophan fluorescence. Urea denaturation of proteins monitored by intrinsic fluorescence
showed no significant differences in the sensitivity of the proteins to this chemical denaturant.
Thermal denaturation monitored by CD and by calorimetry revealed that only the P327R mutant
had a significantly lower midpoint of transition (~ 5 °C) than the wild-type protein, suggesting a
modest decrease in thermal stability. The results show that the HS mutant cdAE1 proteins do not
differ to any great extent in structure from the wild-type protein, suggesting that the HS
mutations may directly affect protein 4.2 binding.
2.2 Introduction
Three cytoplasmic domain mutations associated with HS occur with a normal amount of AE1 at
the red cell membrane. These mutations are E40K (Band 3 Montefiore), G130R (Band 3
52
Fukuoka), and P327R (Band 3 Tuscaloosa). These mutations are associated with an 88 %
(homozygous state), 55 % (homozygous state) and 29 % (heterozygous state) decrease in the
amount of protein 4.2 in the red cell, respectively (Inoue et al. 1998, Jarolim et al. 1992a,
Rybicki et al. 1993). Maintaining normal levels of protein 4.2 in the red cell may depend on the
ability of protein 4.2 to assemble with AE1 during biosynthesis since these proteins have been
shown to associate early in erythroblast differentiation (van den Akker et al. 2010a). All three of
these HS mutations result in the incorporation of a positively charged amino acid into a domain
with predominantly negative surface potential. While introduction of a positive charge at these
sites may perturb direct protein 4.2 interactions, these HS mutations may also cause
conformational changes in the protein disrupting the binding site altogether.
The goal of this work was to express and purify the three HS mutant cdAE1 proteins and to
compare their structure and conformational stability to the wild-type cdAE1 protein. The
Memphis I cdAE1 protein was included in the study as an asymptomatic mutant control. The
conformational stability of the Band 3 Tuscaloosa/Memphis I double mutant (P327R/K56E) was
also examined in order to determine the role of the Memphis I background on this HS mutant. I
hypothesize that the three HS cytoplasmic domain mutations in AE1 affect either the folding and
conformational stability of cdAE1, or the direct binding surface of cdAE1. Either may result in
impaired binding of protein 4.2 and lead to its deficiency in the red blood cells of patients with
these mutations.
2.3 Materials and methods
2.3.1 Materials
The following is a list of materials used and their suppliers: pcDNA3 vector (Invitrogen, San
Diego, CA, USA); QuikChange® site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA);
mutagenic primers (ACGT Corp., Toronto, ON, Canada); pETBlue-1 vector and
Tuner(DE3)pLacI Escherichia coli competent cells (Novagen, Madison, WI, USA); growth
media for Escherichia coli (BD, Sparks, MD, USA); chloramphenicol and carbenicillin (Sigma,
St. Louis, MO, USA); isopropyl-β-D-thiogalactopyranoside (IPTG) (Bioshop, Burlington, ON,
Canada); nickel-nitrilotriacetic acid (Ni-NTA) agarose resin (QIAGEN, Germantown, MD,
USA); sequanal grade urea (Pierce, Rockford, IL, USA).
53
2.3.2 Plasmid construction and mutagenesis
cdAE1 from Met1 to Ser356 was amplified by PCR from full-length human AE1 located on the
pcDNA3 vector and cloned into the expression vector pETBlue-1, which contains an IPTG-
inducible T7lacO promoter (Giordano et al. 1989). The reverse primer included DNA encoding
six histidine residues which are located at the C-terminus of the protein to provide a
hexahistidine tag for purification. The HS mutants, the K56E variant and the P327R/K56E
double mutant were constructed using the QuikChange® system with complementary mutagenic
primers and the mutations were confirmed by sequencing by the ACGT Corp.
2.3.3 Protein expression and purification
All protein expression was performed in the Escherichia coli strain Tuner(DE3)pLacI. This
strain of E. coli is a derivative of the BL21 strain, but has a mutation in the lac permease gene
which allows uniform entry of IPTG into all cells and reduces basal level protein expression.
pETBlue-1 vectors were expressed in Tuner(DE3)pLacI which contains the gene for T7 RNA
polymerase under the control of the IPTG-inducible lacUV5 promoter (Studier and Moffatt
1986). Cells were grown at 37 °C in Luria Bertani (LB) media containing both carbenicillin (50
µg/ml) and chloramphenicol (34 μg/ml) until the cell density reached A600 of 0.6. Protein
expression was induced with 1 mM IPTG, then cells were grown for an additional 4 h at 37 °C
and cell density reached A600 of 1.3. Cells were harvested by centrifugation (4,400 ×g, 30 min)
and solubilized in 80 ml of lysis buffer per litre of cell culture (lysis buffer: 50 mM sodium
phosphate, 300 mM sodium chloride, 5 mM imidazole, 0.2 % β-mercaptoethanol (βME), 0.2 %
Triton X-100, pH 8.0) containing the following protease inhibitors: 0.70 µg/ml pepstatin, 2.0
µg/ml aprotinin, 4.3 µg/ml leupeptin, and 0.28 µg/ml phenylmethanesulfonyl fluoride (PMSF).
Solubilized cells were allowed to sit on ice for 30 min after addition of lysozyme to 1 mg/ml,
followed by sonication at 40 % duty cycle for 2 min on ice. Purification was carried out by a
batch procedure at 4 °C using 1 ml of Ni-NTA agarose resin (QIAGEN) per 80 ml of cell lysate.
Resin was washed twice with 10 ml of wash buffer (50 mM sodium phosphate, 300 mM sodium
chloride, 20 mM imidazole, 0.2 % βME, pH 8.0). Proteins were eluted three times with 1 ml of
elution buffer (50 mM sodium phosphate, 300 mM sodium chloride, 250 mM imidazole, 0.2 %
βME, pH 8.0). Protein solutions were applied to pre-equilibrated PD-10 gel filtration columns
(Amersham Biosciences) for the purpose of buffer exchange into 10 mM ammonium bicarbonate
and subsequent lyophilization. Proteins were lyophilized overnight and stored at -20 °C. Protein
54
purity was determined to be > 95 % by Coomassie staining of samples following SDS-PAGE.
Protein concentrations were determined by the Bio-Rad Bradford total protein assay (Bradford
1976). Protein samples were subjected to mass spectroscopy on an electrospray ionization time-
of-flight (ESI-TOF) instrument for molecular weight analysis to verify the identity of the
proteins. A cysteineless protein where both cysteines of cdAE1 were mutated to alanine
(C201A/C317A) was also submitted for analysis.
2.3.4 Analytical ultracentrifugation
Proteins were freshly purified using Ni-NTA resin and buffer-exchanged into sedimentation
buffer using PD-10 columns without freeze-drying. SE experiments were performed at 20 °C on
an Optima XL-A / XL-I Analytical Ultracentrifuge (Beckman Instruments, Palo Alto, CA) using
an AN50-Ti rotor, quartz windows, and standard six-sector charcoal-filled Epon centerpieces.
Samples were centrifuged at 18,000 × g, 32,000 × g and 50,000 × g for 27 h at each speed to
ensure equilibrium was reached before absorbance measurements were taken. Global analysis of
the data was performed using XL-A / XL-I data analysis software (Origin version 4.1) from
Beckman Instruments. SE experiments were performed on three different concentrations of each
protein (0.32 mg/ml, 0.64 mg/ml and 1.29 mg/ml) in 10 mM sodium phosphate, 50 mM sodium
chloride, pH 7.5.
2.3.5 Circular dichroism
Freeze-dried cdAE1 proteins were dissolved in 10 mM sodium phosphate, 50 mM sodium
fluoride, 1 mM dithiothreitol (DTT), pH 7.0 (Johnson 1990). CD spectra from 190 to 260 nm and
1 nm data pitch were recorded on a Jasco J-810 spectropolarimeter using a final protein
concentration of 0.3 mg/ml cdAE1 in a 1 mm path-length cell at 24 °C. Deconvolution of spectra
was done using the CDPro software package (Sreerama and Woody 2000) for the determination
of secondary structure percentages. For temperature denaturation, samples were heated from 30
°C to 86 °C with a 2 °C data pitch at a scan rate of 2 °C /min and ellipticity was measured at 208
nm. Thermal denaturation data were fit to a standard equation by nonlinear least-squares
regression (using SigmaPlot 2004 version 9.0) assuming a two-state transition for a dimeric
species. Tm is the temperature at the transition midpoint of thermal unfolding. Experiments were
repeated using at least three different preparations of purified protein.
55
2.3.6 pH dependence of intrinsic fluorescence
Stock solutions of cdAE1 proteins were made by dissolving proteins in 50 mM sodium
phosphate, 50 mM sodium borate, 70 mM sodium chloride, 1 mM DTT, pH 7.0 (Appell and Low
1981). Stock protein solution was diluted 50 times into the same buffer pre-adjusted to the
desired pH to a final protein concentration of 0.17 mg/ml. Samples were equilibrated for at least
two h at room temperature prior to measurement. In all fluorescence experiments the intrinsic
fluorescence of the proteins was monitored using a Fluorolog FL3-22 fluorescence
spectrophotometer at 24 °C. The excitation wavelength was 290 nm and the fluorescence
emission was measured from 300 to 420 nm for each sample at each pH. Experiments were
repeated using three different preparations of purified protein.
2.3.7 Calorimetry
cdAE1 proteins were dissolved in 10 mM sodium phosphate, 50 mM sodium chloride, pH 7.5 to
a final concentration of 1.3 mg/ml. Heat capacity measurements were obtained on a MicroCal
VP-DSC differential scanning calorimeter. Samples were heated from 25 °C to 90 °C at a rate of
1.5 °C /min. Temperature denaturation data were fit to a two-state transition model using the
Origin 7.0 data analysis software which employs the Marquardt-Levenberg algorithm for least
squares regressions. Experiments were repeated using at least three different preparations of
purified protein.
2.3.8 Urea denaturation measured by intrinsic fluorescence
Stock solutions of cdAE1 proteins were made by dissolving proteins in 50 mM sodium
phosphate, 50 mM sodium borate, 70 mM sodium chloride, 1 mM DTT, pH 7.0. Stock protein
solution was diluted 50 times into the same buffer pre-adjusted to the desired urea concentration
to a final concentration of 0.17 mg/ml. Samples were equilibrated in denaturant at room
temperature for two h prior to measurement. The excitation wavelength was 290 nm and the
fluorescence emission was measured from 300 to 420 nm for each protein at each urea
concentration at 24 °C. Urea denaturation data were fit to a standard equation by nonlinear least-
squares regression (using SigmaPlot 2004 version 9.0) assuming a two-state transition of a
dimeric species. Cm is the urea concentration at the midpoint of the unfolding transition.
Experiments were repeated using six different preparations of purified protein, except for
experiments on the K56E mutant which were repeated using three protein preparations.
56
2.3.9 Limited tryptic digestion
Proteins were dissolved in 50 mM sodium phosphate, 50 mM sodium borate, 70 mM sodium
chloride, pH 7.0. Trypsin dissolved in the same buffer was added to each sample to a final
concentration of 6.5 μg/ml (final protein concentration was 0.3 mg/ml). The digestion reaction
was allowed to proceed for five min at room temperature and was stopped by the addition of an
equal volume of 2 × sample buffer containing SDS and βME. Samples were run on SDS-PAGE
and detected by Coomassie blue staining. The relative sizes of the protein bands were determined
by the loading of Protein Molecular Weight Markers (Fermentas, Burlington, ON) on the gel.
2.4 Results
2.4.1 Expression and purification of cdAE1 and cdAE1 HS variants in
Escherichia coli
The region of cdAE1 encompassing amino acids 1-356 of the protein was subcloned from full-
length wild-type AE1 cDNA on the pcDNA3 vector into a pETBlue-1 expression vector and
expressed in E. coli Tuner(DE3)pLacI cells as described in the Methods. These residues extend
from the N-terminus of AE1 to the last residue visible in the crystal structure, followed directly
by a His6 tag. The purification of the His6-tagged wild-type cdAE1 and cdAE1 carrying the
E40K, G130R, and P327R mutations was carried out by Ni-NTA affinity chromatography. This
purification method yielded over 20 mg of protein per litre of cell culture of more than 95 %
purity as determined by SDS-PAGE. The cdAE1 proteins ran as monomers of approximately 41
kDa on SDS-PAGE (data not shown). Mass spectroscopy analysis revealed expected molecular
masses for all proteins with additional molecular masses of either 76 or 152 Da on all of the
proteins (data not shown). These additions were not seen with the cysteineless (C201A/C317A)
mutant. The extra molecular masses were likely due to a βME adduct on one (+76 Da) or both
(+152 Da) of the cysteines.
2.4.2 Analytical ultracentrifugation of wild-type and HS mutant cdAE1
proteins
SE experiments using the analytical ultracentrifuge were carried out to determine whether or not
the HS mutations affected the oligomeric structure of the cdAE1 protein. Purified wild-type
cdAE1 has been shown to exist as a dimer (Appell and Low 1981, Colfen et al. 1996, Wang et
al. 1992a, Zhang et al. 2000a). The predicted sequence molecular weight (MWseq) of the wild-
57
type cdAE1 with a His6 tag is 40 866. SE experiments were performed on the wild-type, HS and
K56E mutant cdAE1 proteins using different protein concentrations and rotor speeds, and the
data were fit to a single ideal species model. The analysis gave apparent molecular weights
(MWapp) that were twice that of the MWseq which indicated that the wild-type, HS and K56E
mutant proteins existed as stable dimers in solution. No evidence of concentration dependence of
the molecular weight values was observed over a range from 0.32 – 1.29 mg/ml. The ratios of the
MWapp to the MWseq determined by SE experiments are listed in Table 2.1, and a
representative plot from the SE experiments on wild-type cdAE1 run at 32 000 × g is shown in
Figure 2.1.
2.4.3 Secondary structure analysis of wild-type and HS mutant cdAE1 proteins
CD analysis was carried out to determine whether the HS mutations affect the secondary
structure of the cdAE1 protein. The CD spectrum of the wild-type cdAE1 has been shown to
exhibit a negative extreme at 208 nm and a shoulder at 223 nm typical of an α-helical structure-
containing protein (Appell and Low 1981). Figure 2.2 shows the CD spectra obtained from the
wild-type and HS mutant cdAE1 proteins. The spectra from this study display the same
characteristics as those obtained from native cdAE1, and those of the mutant cdAE1 proteins
overlap with the spectrum of the wild-type protein. Similar spectra were obtained for the
Memphis I mutant protein (data not shown). The helical content of the cdAE1 protein obtained
from the crystal structure was 26 % (Zhang et al. 2000a). The helical content of the wild-type,
HS and K56E mutant cdAE1 proteins obtained by deconvolution of the CD spectra shown in
Table 2.1 are similar to that found in the crystal structure. The HS mutations did not result in any
major change in the secondary structure of cdAE1.
58
Figure 2.1: Analytical ultracentrifugation of wild-type cdAE1. Absorbance is plotted as a function of radius, and the residuals from fitting the data to a single
ideal species model are shown. The SE experiment was performed at 20 °C with a rotor speed of
32000 × g for 27 h. The protein concentration was 0.64 mg/ml and the protein was in 10 mM
sodium phosphate, 50 mM sodium chloride, pH 7.5.
59
Figure 2.2: CD spectra of wild-type cdAE1 and HS mutants.
Purified proteins were dissolved in 10 mM sodium phosphate, 50 mM sodium fluoride, 1 mM
DTT, pH 7.0 and scanned at 24 °C in a 1 mm cell in a Jasco J-810 spectropolarimeter. The final
concentration of all proteins was 0.3 mg/ml. Spectra are expressed as mean residue ellipticity.
Wild-type cdAE1 (closed circles); E40K (open circles); G130R (closed triangles); P327R (open
triangles).
60
2.4.4 Effect of pH on the intrinsic fluorescence of the wild-type and HS mutant
cdAE1 proteins
cdAE1 has been shown to undergo a reversible pH-dependent conformational change that is
characterized by a dramatic increase in the intrinsic tryptophan fluorescence and an increase in
Stokes radius without a change in secondary structure at alkaline pH (Wang et al. 1992a). A
hydrogen bond between W105 of one subunit and D316 of the other subunit, that is present in
the lower pH conformation, is broken at neutral pH (Zhou and Low 2001). At alkaline pH the
conformation of cdAE1 changes extensively, as the peripheral protein binding domain moves
away from the dimerization arm. In the present study, intrinsic tryptophan fluorescence was
measured over a pH range to determine whether or not the HS mutations caused a difference in
the pH-dependent conformational change of the cdAE1 protein. Figure 2.3 shows the intrinsic
fluorescence emission intensity at 347 nm of the wild-type and HS mutant cdAE1 proteins as a
function of pH. The proteins exhibited a similar increase in fluorescence intensity, representing a
dequenching of tryptophans, and an increase in peak wavelength (red-shift) at alkaline pH (data
not shown), indicating that the tryptophans were exposed to a more polar environment. There
was a moderate increase in fluorescence emission intensity between pH 5 and 8, followed by a
more dramatic increase between pH 8 and 10 for all samples. The results indicate that the wild-
type and mutant proteins undergo similar two stage pH-dependent conformational changes.
2.4.5 Thermal denaturation of wild-type and HS mutant cdAE1 proteins by
circular dichroism
The HS mutations did not appear to affect the folded structure or oligomeric state of the cdAE1
protein, but the mutations may affect the thermal stability of this domain. Thermal denaturation
using CD was performed in order to determine if the HS mutants lowered the thermal stability of
the cdAE1 protein. Figure 2.4 shows the results of thermal denaturation as monitored by CD at
208 nm at pH 7.0. The thermal denaturation of all the proteins was irreversible as determined by
rescanning of cooled samples after heating to 86 °C. The data were fit to a two-state transition
model for the purpose of obtaining the midpoints of the thermal denaturation transitions (Tm) so
that they could be compared to one another. Because the transitions were irreversible, these Tms
were relative and were considered to be the apparent Tms of the transitions. The Tm for each
protein is listed in Table 2.1. The wild-type protein had a Tm of 69.0 °C. The P327R cdAE1
protein had a Tm that was 5 °C lower than wild-type.
61
Figure 2.3: Intrinsic fluorescence intensity of wild-type cdAE1 and HS mutants as a
function of pH.
Purified proteins were dissolved in 50 mM sodium phosphate, 50 mM sodium borate, 70 mM
sodium chloride, 1 mM DTT, pH 7.0 to a stock concentration of 10 mg/ml. All samples were
filtered through a 0.22 µm syringe filter and 6 μl of each stock protein solution was added to 294
μl (50 × dilution) of the same buffer preadjusted to the desired pH. The final concentration was
0.17 mg/ml for all proteins, and the actual pH was measured in each reaction tube using a pH
meter. The intrinsic fluorescence intensity at 347 nm of wild-type cdAE1 (closed circles), E40K
(open circles), G130R (closed triangles) and P327R (open triangles) is plotted as a function of
pH.
62
Table 2.1: Effects of HS mutations on structure and stability of cdAE1 protein
cdAE1
protein
MWapp/
MWseq α-helix (%) Tm (°C) (CD) Tm (°C) (calorimetry) Cm (M)
WT 1.981a 28.5
b ± 3.3 69.0
c ± 2.2 66.2
e ± 1.5 4.79
d ± 0.31
E40K 2.026 28.3 ± 1.7 68.4 b ± 0.3 64.9
b ± 0.3 4.73 ± 0.13
G130R 1.950 27.9 ± 1.5 68.7 b ± 0.4 65.5
b ± 0.3 4.85 ± 0.09
P327R 1.921 28.3 ± 2.6 64.0 b ± 0.4 61.5
b ± 0.2 4.56 ± 0.20
K56E 2.011 26.7 ± 4.4 67.7 d ± 1.5 66.5
f ± 2.0 4.60
b ± 0.15
a Data from nine SE experiments (three different protein concentrations run at three different speeds) were
globally fit to a single ideal species model to obtain the MWapp. b
Values are averages of measurements
obtained from three different preparations of purified protein. c
Values are averages of measurements obtained
from nine different preparations of purified protein. d Values are averages of measurements obtained from six
different preparations of purified protein. e Values are averages of measurements obtained from twelve different
preparations of purified protein. f
Values are averages of measurements obtained from seven different
preparations of purified protein. For CD and calorimetry measurements Tm is the apparent temperature midpoint
of the thermal denaturation transition. Cm is the concentration of urea at the apparent midpoint of the urea-
induced unfolding transition.
63
The E40K, G130R and Memphis I (data not shown) cdAE1 proteins did not have significantly
lower melting temperatures. The P327R mutant appears to be less thermally stable than the other
proteins, while the E40K and G130R mutants have similar thermal stabilities to the wild-type
protein.
2.4.6 Thermal denaturation of wild-type and HS mutant cdAE1 proteins by
calorimetry
Thermal denaturation was also performed using differential scanning calorimetry (DSC). The
melting temperature of the wild-type cdAE1 protein measured by DSC has been shown to be pH
dependent and was reported to be 67 °C at pH 7.51 (Appell and Low 1981). Figure 2.5 shows the
results of the baseline-subtracted DSC scans. The thermal denaturation of all the proteins was
irreversible since rescanning cooled samples after heating to 90 °C yielded scans with no
transition. The data were fit to a two-state transition model for the purpose of obtaining
temperature midpoints of transition so that they could be compared to one another. As with the
CD analysis, because the transitions were irreversible, the Tms were relative and were considered
to be the apparent Tms of the transitions. These Tms agree with the temperature corresponding to
the maximum excess heat capacity (Cp), which was used in other studies of proteins with
irreversible thermal denaturation profiles (Idakieva et al. 2005, Nielsen et al. 2003). The Tm
values for each protein at pH 7.5 obtained using DSC are listed in Table 2.1. The wild-type
protein had a Tm of 66.2 °C, which is in agreement with previous studies. The P327R cdAE1
protein had a Tm that was 5 °C lower than wild-type. The E40K, G130R and K56E cdAE1
proteins had similar melting temperatures to wild-type. The P327R/K56E double mutant protein
had a Tm (62.2 °C) that was not significantly different than that of the P327R single mutant (61.5
°C). Although the transition midpoints differ slightly from those obtained from the CD thermal
denaturation experiments, the relative values are similar. The thermal denaturation studies show
that only the P327R mutation results in very modest thermal destabilization of the cdAE1.
64
Figure 2.4: Thermal denaturation of wild-type cdAE1 and HS mutants monitored by CD.
Purified wild-type and HS mutant cdAE1 proteins were dissolved in 10 mM sodium phosphate,
50 mM sodium fluoride, 1 mM DTT, pH 7.0 for CD measurements. The final concentration of
the proteins was 0.3 mg/ml. Ellipticity was measured at 208 nm as the temperature was increased
from 30 °C to 86 °C in 2 °C increments. Wild-type cdAE1 (closed circles); E40K (open circles);
G130R (closed triangles); P327R (open triangles).
65
Figure 2.5: Thermal denaturation of wild-type cdAE1 and HS mutants by DSC.
The proteins were dissolved in 10 mM sodium phosphate, 50 mM sodium chloride, pH 7.5 at a
final protein concentration of 1.3 mg/ml. The heat capacity was measured as the temperature was
increased from 35 °C to 80 °C and the baseline-subtracted values are presented. No transition
was observed after cooling the samples and reheating. Wild-type cdAE1 (closed circles); E40K
(open circles); G130R (closed triangles); P327R (open triangles).
66
2.4.7 Urea denaturation of wild-type and HS mutant cdAE1 proteins
Urea denaturation was also used to study the conformational stability of the cdAE1 proteins.
Urea is a chaotropic agent that causes proteins to denature, and proteins with decreased
conformational stabilities will unfold at lower concentrations of urea (Pace 1986). The
concentration dependence of urea denaturation of the wild-type cdAE1 protein at pH 6.0 has
been shown to cause an increase in intrinsic fluorescence intensity as urea concentration
increased, with the midpoint of the transition at about 5 M urea (Zhou and Low 2001). In the
present study urea denaturation of the wild-type and HS mutant cdAE1 proteins at pH 7.0
resulted in an increase of intrinsic fluorescence intensity as tryptophans became dequenched, and
an increase in peak wavelength as protein unfolding exposed tryptophans to a more polar
environment. Figure 2.6 shows emission intensity spectra for the wild-type cdAE1 at different
urea concentrations and the increase in average emission wavelength as a function of urea
concentration for wild-type and HS mutant proteins. The apparent transition midpoints (Cm) for
each protein are listed in Table 2.1. The Cm values for the mutant proteins were similar to the
wild-type protein and their differences were not statistically significant. Dilution of cdAE1
proteins in 8 M urea into buffer to give a final urea concentration of less than 1 M did not result
in refolding of cdAE1 (data not shown), indicating that the unfolding was irreversible.
2.4.8 Limited tryptic digestion of WT and HS mutant cdAE1 proteins
Limited tryptic digestion of cdAE1 proteins was used to measure the folded structures of the
proteins. Trypsin cleaves after lysine and arginine residues in areas of a folded protein that are
accessible to the protease. Regions of the protein where cleavage sites are buried within the
folded structure of the protein or where the backbone is well-ordered will not be readily cleaved.
Wild-type and HS mutant cdAE1 proteins (0.3 mg/ml) were incubated with trypsin (6.5 µg/ml)
for five min at room temperature and the reaction was stopped by addition of an equal volume of
sample buffer containing SDS and βME. Samples were run on SDS-PAGE and visualized with
Coomassie blue staining as seen in Figure 2.7. Cleavage patterns of the G130R and P327R
mutants were similar to the wild-type pattern. The full-length 41 kDa cdAE1 protein was present
along with the appearance of two smaller bands of approximately 21 kDa and 20 kDa, perhaps
representing the two fragments obtained from cleavage at one site of the protein by trypsin.
Undigested wild-type and HS cdAE1 proteins each appear as a single 41 kDa band.
67
The E40K protein gave a different cleavage pattern, with a prominent band of about 36 kDa
and the 21 kDa band. The E40K mutant introduces a new trypsin cleavage site at lysine 40 and
cleavage at this new site would account for the appearance of the 36 kDa band, which is
approximately 40 residues smaller than the full-length protein. A western blot using an antibody
against the N-terminus of cdAE1 confirmed that the 20 kDa band was derived from the N-
terminal region of the protein (data not shown). The 36 kDa band from the E40K digest was not
detected since its first 40 residues had been cleaved off. The 21 kDa band for all of the proteins
was not detected by the N-terminal antibody indicating that this fragment was derived from the
C-terminal region of the protein. Judging from the amino acid sequence of cdAE1, the possible
single site of cleavage that would result in fragment sizes of 20 kDa and 21 kDa is at K174. This
lysine residue is located in an unstructured region of the peripheral protein-binding domain and
is exposed to solvent. The shorter fragment produced from cleavage at this lysine would be from
the N-terminus. K174 is indicated on the topology model of AE1 in Figure 1.6.
The Memphis I mutant protein removes a potential trypsin cleavage site (K56E) in the
protein, but the wild-type and HS mutant proteins examined still have a lysine at position 56. No
fragment corresponding to cleavage at this site was observed for any of the proteins. Trypsin did
not cleave at this position, indicating that either it was not accessible to the protease or that the
backbone is well-ordered in this region. In fact, according to the crystal structure of cdAE1, K56
is located at the N-terminus of the first β-strand of the protein. The protease digestion results
indicate that the HS mutations did not induce gross misfolding of cdAE1 and that the region
around residue 40 is accessible and disordered, while that of residue 56 is not.
2.5 Discussion
HS mutations in the cytoplasmic domain of AE1 were found to cause no major changes in the
structure of the domain. The HS mutants of cdAE1 retained the dimeric structure of the protein
as seen in analytical ultracentrifugation experiments. Even the P327R mutation which is located
at the N-terminus of an α-helix in the dimerization arm did not interfere with the formation of a
dimer by the protein. The mutants also retained the normal secondary structure of the protein as
seen in CD experiments. The mutations may not be expected to cause a major disruption in the
secondary structure since the E40K mutation is located in a structurally unresolved region and
the G130R residue points away from the protein and into the solvent.
68
Figure 2.6: Urea denaturation of wild-type and HS mutant cdAE1 proteins.
Purified wild-type and HS mutant cdAE1 proteins were dissolved in 50 mM sodium phosphate,
50 mM sodium borate, 70 mM sodium chloride, 1 mM DTT, pH 7.0 to a stock concentration of
10 mg/ml. All samples were filtered through a 0.22 µm syringe filter and 6 μl of each stock
protein solution was added to 294 μl (50 × dilution) of the same buffer preadjusted to the desired
urea concentration. The final concentration was about 0.17 mg/ml for all proteins. Samples were
equilibrated for 2 h before conducting measurements. The intrinsic fluorescence emission was
measured from 300 - 420 nm (λex, 290 nm) for each protein at each urea concentration.
(A) Intrinsic fluorescence emission spectra of wild-type cdAE1 are plotted for three urea
concentrations: 0 M (black line), 5 M (dark gray line), 8 M (light gray line). (B) The average
emission wavelength of wild-type cdAE1 (closed circles), E40K (open circles), G130R (closed
triangles) and P327R (open triangles) was calculated from each spectrum and plotted as a
function of urea concentration.
69
Figure 2.7: Trypsin digestion of wild-type and HS mutant cdAE1 proteins.
Purified wild-type and HS mutant cdAE1 proteins were dissolved in 50 mM sodium phosphate,
50 mM sodium borate, 70 mM sodium chloride, pH 7.0 and incubated with trypsin dissolved in
the same buffer. Digestion proceeded for 5 min at room temperature and the reaction was
stopped by addition of an equal volume of 2 × sample buffer. The final cdAE1 protein and
trypsin concentrations were 0.3 mg/ml and 6.5 μg/ml, respectively. Digested proteins were run
on SDS-PAGE and detected by Coomassie blue staining.
70
Arginine at position 327, instead of proline, could allow continuation of the helix that
follows it and no detectable change in the helical content of this mutant was observed. All three
mutants underwent similar pH-dependent conformational changes as the wild-type protein, as
monitored by intrinsic tryptophan fluorescence, indicating that their folded structures are similar.
This pH-dependent behavior is important physiologically since binding of glycolytic enzymes,
deoxyhemoglobin and cytoskeletal proteins to AE1 in the red cell membrane is very sensitive to
pH.
The E40K mutant displayed a different digestion pattern by limited tryptic digestion since it
introduced a new trypsin cleavage site at residue 40. This indicates that the residue at position 40
is in a disordered region accessible to trypsin. The position of the Memphis I mutation, lysine 56,
at the N-terminus of the first β-strand of cdAE1 prevented cleavage at this site. Even though
residues 40 and 56 are only sixteen residues apart in the primary sequence of the protein, their
environments in the tertiary structure of the protein are quite different. Residue E40’s lack of
resolution in the crystal structure also provides evidence of its surface accessibility and disorder
while K56 is one of the first structured residues. Since trypsin was unable to cleave at lysine 56
in the wild-type and HS mutant proteins, this provides evidence for the similarity in their folded
structures.
The P327R mutation caused a modest thermal destabilization of cdAE1 as seen in thermal
denaturation experiments. Interestingly, the thermal stability of the double mutant, P327R/K56E,
was not significantly different from that of the single P327R mutant, and the thermal stability of
the K56E mutant was similar to that of the wild-type protein. This provides evidence for the first
time that the underlying cause for the HS phenotype in the patient with this double mutant is the
P327R mutation, and not K56E.
Conformational destabilization of HS mutant proteins was not seen in the experiments
where urea was used as the denaturant, not even with the P327R mutant. The lack of
conformational destabilization seen with the P327R mutant could be due to different modes of
unfolding by thermal and chemical denaturation. Intrinsic fluorescence spectroscopy, the
technique used to probe for unfolding in the urea experiments, is a measure of the environment
of the four tryptophans in the protein. From these experiments there does not appear to be much
difference between the HS mutant proteins and the wild-type. The results show that protein 4.2
71
deficiency in red cells is not the result of a major structural change in cdAE1, but may be the
result of changes at the binding interface due to the introduction of positive charges, which
occurs with each of these HS mutations. The asymptomatic polymorphism K56E also retains the
major structural features of the protein and has similar thermal stability to the wild-type protein.
This is the first study to show that the molecular basis of HS associated with these three HS
mutations is not caused by a structural deformity in the cytoplasmic domain of AE1, which is in
keeping with the normal amount of these mutant proteins at the red cell membrane of these
patients. The phenotype is likely caused by perturbation at the binding interface between cdAE1
and protein 4.2 at the specific sites of mutation. The HS mutation sites are not clustered together
in the folded structure of cdAE1 as seen in the crystal structure, but are located far apart from
each other, which indicates that the cdAE1-protein 4.2 binding interface may be large. The loss
of protein 4.2 binding to the cytoplasmic domain of AE1 may result in degradation or loss of
protein 4.2.
72
3 Chapter 3: Protein 4.2 interaction with hereditary
spherocytosis mutants of the cytoplasmic domain of human
anion exchanger 1
A version of this research was originally published in Biochemical Journal. Bustos, S.P. &
Reithmeier, R.A. Protein 4.2 interaction with hereditary spherocytosis mutants of the
cytoplasmic domain of human anion exchanger 1. Biochemical Journal. 2010; 433: 313-322 ©
the Biochemical Society (http://www.biochemj.org/bj/433/bj4330313.htm) (Bustos and
Reithmeier 2010). Jing Li contributed to this work by the generation of the AE1HS mutants and
mdAE1, and in carrying out some replicate experiments of the co-immunoprecipitation and
immunofluorescence assays. Jing Li carried out the Ni-NTA pull-down and subcellular
fractionation experiments. I performed the His6-tagged AE1HS mutant construction, protein 4.2
subcloning and hemagglutinin (HA)-tag engineering, and G2A/C173A protein 4.2 mutant
construction. I designed all of the experiments, analyzed all of the results and conducted replicate
experiments of the co-immunoprecipitation and immunofluorescence assays.
3.1 Abstract
Anion exchanger 1 and protein 4.2 associate in a protein complex bridging the erythrocyte
membrane and cytoskeleton; disruption of the complex results in unstable erythrocytes and HS.
Three HS mutations (E40K, G130R and P327R) in cdAE1 occur with deficiencies of protein 4.2.
The interaction of wild-type AE1, AE1HS mutants, mdAE1, kAE1 and AE1SAO with protein
4.2 was examined in transfected HEK-293 cells. The HS mutants had wild-type expression levels
and plasma membrane localization, and protein 4.2 expression was not dependent on the
presence of AE1. Protein 4.2 was localized throughout the cytoplasm, and co-localized at the
plasma membrane with the HS mutants, mdAE1, and kAE1, but at the ER with AE1SAO. Pull-
down assays revealed diminished levels of protein 4.2 associated with the HS mutants relative to
AE1. mdAE1 did not bind protein 4.2, while kAE1 and AE1SAO bound wild-type amounts of
protein 4.2. A protein 4.2 fatty acylation mutant, G2A/C173A, had decreased plasma membrane
localization compared to wild-type protein 4.2, and co-expression with AE1 enhanced its plasma
membrane localization. Subcellular fractionation showed that the majority of wild-type and
G2A/C173A protein 4.2 was associated with the cytoskeleton of HEK-293 cells. This study
73
shows that HS mutations in cdAE1 cause impaired binding of protein 4.2 to AE1, leaving protein
4.2 susceptible to degradation or loss during red cell development.
3.2 Introduction
Since the E40K, G130R and P327R AE1 mutations all affect the levels of protein 4.2 in red
blood cells, I hypothesize that these mutations impair the binding of protein 4.2 to AE1. To test
this hypothesis, the interaction of protein 4.2 with wild-type AE1 and the three cytoplasmic
AE1HS mutants was studied in transfected HEK-293 cells. mdAE1, which is unable to bind
protein 4.2 (Korsgren and Cohen 1986) was included in these studies as a negative control.
As discussed later in the Appendix, the cytoplasmic domain of kAE1 (cdkAE1) is less
thermally stable than erythroid cdAE1 and exists in a more open structure (Pang et al. 2008). As
well, it does not bind ankyrin or glycolytic enzymes (Ding et al. 1994b, Wang et al. 1995b) most
likely due to the absence of the acidic N-terminal tail. We included kAE1 in these studies to see
what effect the missing tail and altered structure would have on protein 4.2 binding. Also, protein
4.2 is expressed in kidney cells (Friedrichs et al. 1989) where it may perform a similar function
as in erythrocytes. We also included AE1SAO in our studies to see if this mutant’s increased
cytoskeletal attachment (Liu et al. 1995, Liu et al. 1990, Sarabia et al. 1993) translated into an
increased interaction with protein 4.2.
The localization and interaction of protein 4.2 with these AE1 proteins was examined in
HEK-293 cells, as was the role of fatty acid modifications on protein 4.2 localization. Protein
4.2 had a broad distribution in transfected cells, was predominantly associated with the
cytoskeletal fraction, and co-localized with AE1 at the plasma membrane. The AE1HS mutants,
but not kAE1 or AE1SAO, had impaired protein 4.2 binding; this weakened interaction may
account for the loss of protein 4.2 during red cell development.
3.3 Materials and methods
3.3.1 Materials
The following is a list of materials used (Suppliers): pcDNA3 vector (Invitrogen); mutagenic
primers (ACGT Corp.); QuikChangeTM
site-directed mutagenesis kit (Stratagene); HEK-293
cells (ATCC); Dulbecco's modified Eagle's medium (DMEM), calf serum, penicillin,
streptomycin (Gibco BRL); LipofectamineTM
2000 (Invitrogen); C12E8 (Nikko Chemical Co.);
74
Protein G-Sepharose (Amersham Biosciences); poly-L-lysine (Sigma-Aldrich); Ni-NTA agarose
resin (QIAGEN); mouse anti-AE1 (BRIC 6) antibody, which recognizes an extracellular epitope
at the C-terminus of AE1 (Bristol Institute for Transfusion Sciences); mouse anti-AE1 antibody,
which recognizes an intracellular epitope of AE1 (a gift from Dr Michael L. Jennings, University
of Arkansas for Medical Sciences, Little Rock, AR, USA); rabbit anti-CtAE1 antibody raised
against the last 16 amino acids of AE1 (SynPep Corporation); mouse anti-HA antibody
(Covance); rat anti-HA antibody (Roche); rabbit anti-calnexin (CNX) antibody (Stressgen
Biotech); lectin peanut agglutinin (PNA) Alexa Fluor® 488 conjugate (Molecular Probes);
mouse anti-actin antibody (Chemicon); mouse anti-GAPDH antibody (Millipore); goat
peroxidase-conjugated anti-rabbit IgG and anti-mouse IgG (New England Biolabs); Cy3-
conjugated goat anti-mouse, anti-rat and anti-rabbit antibodies (Jackson ImmunoResearch);
Alexa Fluor® 488-conjugated goat anti-mouse and anti-rat antibodies (Molecular Probes);
Chemiluminescence Kit (Roche).
3.3.2 Site-directed mutagenesis
The coding sequence for wild-type human AE1 was inserted into the XhoI and BamHI sites of
the pcDNA3 vector. The construction of AE1SAO and kAE1 are described in Cheung and
Reithmeier (2005) and Quilty et al. (2002), respectively. mdAE1 was constructed by PCR on the
full-length AE1 protein with a methionine engineered at the start of the DNA sequence coding
for amino acids Asp369-Val911, followed by subcloning into the pcDNA3 vector. Asp369 is the
first perfectly conserved residue in the human AE1 family (AE1, AE2 and AE3) and defines the
beginning of the membrane domain. The AE1HS mutants were created using the QuikChangeTM
mutagenesis kit using complementary mutagenic primers, with wild-type AE1 as template. The
coding sequence for wild-type human protein 4.2, Type II, was subcloned from the pGEM-
7Zf(+/-) vector into the EcoRI site of the pcDNA3 vector. The C-terminal HA tag on protein 4.2
was created by PCR, inserting the tag after position Ala691. The G2A/C173A fatty acylation
mutant of protein 4.2 was created using the QuikChangeTM
mutagenesis kit using complementary
mutagenic primers, with wild-type HA-tagged protein 4.2 as template. HA-tagged protein 4.2
and the HA-tagged G2A/C173A mutant were used in all experiments, but were referred to as
protein 4.2 and G2A/C173A in the text, respectively. Sequencing of constructs was performed by
ACGT Corp.
75
3.3.3 Transient transfection and expression of AE1 and protein 4.2 in HEK-
293 cells
HEK-293 cells were grown in DMEM supplemented with 10 % (v/v) calf serum, 0.5 %
penicillin and 0.5 % streptomycin under 5 % CO2 at 37 °C as described (Popov et al. 1999).
Cells were transfected by the Lipofectamine method (Sells et al. 1995) with 2 μg of plasmid
DNA per well of a six-well plate.
3.3.4 SDS-PAGE and immunoblotting
Proteins were resolved by SDS-PAGE (8 % or 10 % gels) (Laemmli 1970) and transferred to
nitrocellulose membrane (Towbin et al. 1979). AE1 was detected with a mouse monoclonal anti-
AE1 (Jennings) antibody (Jennings et al. 1986). HA-tagged protein 4.2 was detected using
mouse monoclonal anti-HA antibody. Actin was detected using mouse anti-actin antibody and
GAPDH was detected using mouse anti-GAPDH antibody. Goat peroxidase-conjugated anti-
mouse IgG was then added followed by detection by chemiluminescence and film exposure or by
a VersaDoc Imaging System Model 5000. Band intensities of immunoblots in the linear range of
intensity were determined using the ImageJ software (version 1.41o).
3.3.5 Immunofluorescence and confocal microscopy
HEK-293 cells transfected with pcDNA3 plasmids were grown on glass cover slips. In some
cases, cover slips were coated with poly-L-lysine, but this made no difference in the growth or
adherence of the cells. Cells were fixed with 3.8 % (w/v) paraformaldehyde for 15 min and
washed once with 100 mM glycine. Cells were either non-permeabilized or permeabilized and
incubated with antibodies. Non-permeabilized cells were blocked with 0.2 % (w/v) bovine serum
albumin (BSA) for 30 min, followed by incubation with 1:100 diluted mouse anti-AE1 (BRIC 6)
or 1:100 diluted PNA Alexa Fluor® 488 conjugate in 0.2 % BSA for 30 min. These cells were
then permeabilized with 0.2 % (v/v) Triton X-100 for 5 min and blocked with 0.2 % (w/v) BSA
for 30 min. Next, 1:250 diluted rat anti-HA antibody was added in 0.2 % BSA for 30 min. Cells
that were permeabilized at the beginning were permeabilized with 0.2 % (v/v) Triton X-100 for 5
min and blocked with 0.2 % (w/v) BSA for 30 min. These cells were then incubated with 1:500
diluted mouse anti-AE1 (Jennings) antibody, 1:100 diluted mouse anti-AE1 (BRIC 6) antibody,
1:250 diluted rat anti-HA antibody, or 1:250 diluted rabbit anti-CNX antibody in 0.2 % BSA for
30 min. Following several washes, samples were incubated with 1:1000 dilution of Alexa
76
Fluor® 488-conjugated goat anti-mouse antibody, Alexa Fluor® 488-conjugated goat anti-rat
antibody, Cy3-conjugated goat anti-mouse antibody, Cy3-conjugated donkey anti-rat antibody,
or Cy3-conjugated donkey anti-rabbit antibody for 30 min. A Zeiss laser confocal microscope
LSM 510 was used to observe the samples.
3.3.6 Ni-NTA pull-down
HEK-293 cells transiently transfected with His6-tagged AE1 and HA-tagged protein 4.2
constructs, were harvested with lysis buffer (1 % C12E8, 300 mM NaCl, and 10 mM imidazole
with protease inhibitors in PBS). Cell lysates were centrifuged at 14 000 g for 30 min at 4 °C to
remove insoluble material. The supernatants were added to 50 μl of a 50 % slurry of Ni-NTA
agarose in binding buffer (0.1 % C12E8, 300 mM NaCl and 10 mM imidazole with protease
inhibitors in PBS) and incubated for 2 h at 4 °C. Resin was washed with 0.3 ml of wash buffer
(0.2 % C12E8, 300 mM NaCl and 30 mM imidazole with protease inhibitors in PBS) three times.
Bound proteins were eluted with elution buffer (0.5 % C12E8, 300 mM NaCl and 500 mM
imidazole in PBS) and solubilized in 2 × SDS sample buffer. Samples were analyzed by SDS-
PAGE (8 % gels) and immunoblotting was performed as described above. Band intensities were
determined by ImageJ 1.41o software. The amount of protein 4.2 associated with the amount of
AE1 eluted from the resin was calculated from immunoblots from eight separate transfection
experiments and normalized to the total amount of protein 4.2 expressed in the cells. Each value
was reported as relative to that of AE1, which was set to 100 %, to give a value of protein 4.2
relative binding. Results for protein 4.2 binding are given as means ± S.D. Mean values were
considered to be significantly different (p < 0.05) when the Student’s t test was used.
3.3.7 Co-immunoprecipitation
HEK-293 cells transiently transfected with AE1 and HA-tagged protein 4.2 constructs were
harvested with lysis buffer (1 % C12E8 with protease inhibitors in PBS). Cell lysates were
centrifuged at 14 000 × g for 30 min at 4 °C to remove insoluble material. AE1 was
immunoprecipitated from supernatants with 4 μl of rabbit anti-CtAE1 antibody followed by 100
μl of Protein G-Sepharose. Proteins were eluted with 25 μl 0.1 M glycine, pH 2.5, on ice for 20
min. Two μl of 1 M Tris, pH 9.0, was added to the eluate and proteins were solubilized in 2 ×
SDS sample buffer. Samples were analyzed by SDS-PAGE (8 % gels) and immunoblotting was
performed as described above. Band intensities were determined by ImageJ software (version
77
1.41o) from various blot exposures ensuring that the band intensities were in the linear range.
Linear dilutions of protein were included on the blots to ensure linearity. The amount of protein
4.2 associated with the amount of AE1 eluted from the resin was calculated from immunoblots
from eight separate transfection experiments and normalized to the total amount of protein 4.2
expressed in the cells. Each value was reported as relative to that of AE1, which was set to 100
%, to give a value of protein 4.2 relative binding. Results for protein 4.2 binding are given as
means ± S.D. Mean values were considered to be significantly different (p < 0.05) when the
Student’s t test was used.
3.3.8 Subcellular fractionation of HEK-293 cells expressing wild-type or
G2A/C173A protein 4.2
HEK-293 cells transiently transfected with HA-tagged wild-type protein 4.2 or the G2A/C173A
protein 4.2 mutant were suspended in PBS with protease inhibitors, followed by sonication at 50
% duty cycle for 20 pulses on ice to lyse cells. Cell lysates were centrifuged at 1500 × g for 10
min at 4 °C to remove cell debris and unbroken cells. The supernatant, containing the membrane,
cytoskeletal and cytoplasmic fractions (labeled the Total fraction) was collected and centrifuged
at 100 000 × g for 1 h at 4 °C. The resulting supernatant contained the soluble cytoplasmic
fraction and was labeled S1. The pellet was washed with 1 ml of PBS with protease inhibitors
and centrifuged at 100 000 × g for 1 h at 4 °C and the wash fraction labeled Sw. The pellet,
containing the membrane and cytoskeletal fractions, was resuspended in lysis buffer (1 % C12E8
with protease inhibitors in PBS) and was labeled P1. Samples were slowly rotated for 1 h at 4 °C
to solubilize the membrane proteins in the detergent. Samples were then centrifuged at 100 000
× g for 1 h at 4 °C. The resulting supernatant contained detergent-soluble membrane protein and
was labeled S2. The pellet contained the cytoskeletal fraction and was labeled P2. Samples were
solubilized in 2 × SDS sample buffer and analyzed by SDS-PAGE (10 % gels). Immunoblotting
was performed as described above and band intensities were determined by ImageJ software
(version 1.41o). Detection of GAPDH was used as a marker for soluble cytoplasmic proteins.
Detection of actin was used as marker for soluble (G-actin) and cytoskeletal (F-actin) proteins.
78
3.4 Results
3.4.1 Expression of protein 4.2 and AE1 proteins in transfected HEK-293 cells
Protein 4.2 was expressed alone or co-expressed with AE1, the HS mutants (E40K, G130R,
P327R), as well as mdAE1, kAE1 and AE1SAO in transiently transfected HEK-293 cells. Figure
3.1 shows a representative immunoblot of total cell extracts. AE1 and AE1HS proteins were
detected as approximately 95 kDa bands (A: lanes 2, 4, 5, 6). No immunoreactive band was seen
in the empty vector-transfected control lane (A: lane 1). All three AE1HS proteins had similar
expression levels compared to the wild-type protein. The immunodetection of mdAE1 (A: lane
3) was variable and often lower than AE1 due to the presence of multiple bands. Expression
(data not shown) of kAE1 was comparable to AE1, while AE1SAO was lower as reported
previously (Cheung et al. 2005a). HA-tagged protein 4.2 expressed in HEK-293 cells in the
absence or presence of AE1 proteins was readily detected as an approximately 72 kDa
immunoreactive band (B: lanes 1-6). No HA-immunoreactive band was seen in the non-
transfected control cells (data not shown). In multiple transfection experiments, there was no
statistically significant difference in the amount of protein 4.2 expressed in the absence or
presence of AE1, AE1HS mutants, mdAE1, kAE1 or AE1SAO. These results show that the
AE1HS mutants had similar stabilities compared to AE1 and that co-expression with AE1 was
not necessary for the expression of protein 4.2 in transfected HEK-293 cells.
3.4.2 Localization of AE1 proteins in HEK-293 cells
The localization of wild-type and mutant AE1 proteins expressed in HEK-293 cells was
determined by immunofluorescence and confocal microscopy. The first column in Figure 3.2
shows immunofluorescence of AE1 proteins in non-permeabilized cells. AE1 was readily
detected at the cell surface indicating that it had trafficked to the plasma membrane, as seen in
previous studies (Tang et al. 1998). The three AE1HS mutants were also detected at the cell
surface at similar intensities, indicating their ability to traffic to the plasma membrane. The
mdAE1 lacking the entire cytoplasmic domain was also detected at the cell surface, indicating
that the TM domain alone is sufficient for plasma membrane localization. The kAE1 isoform
was also able to traffic to the plasma membrane. In contrast, AE1SAO was not detected at the
cell surface, indicating its inability to traffic from the ER to the plasma membrane in transfected
HEK-293 cells, as was previously reported (Cheung et al. 2005a).
79
A
1 2 3 4 5 6
M
empty
vector
AE1
md
AE1
E40K
G130R
P327R
B
Protein 4.2 expressed with: 1 2 3 4 5 6
M
empty
vector
AE1
md
AE1
E40K
G130R
P327R
Figure 3.1: Expression of AE1 proteins and protein 4.2 in HEK-293 cells.
Immunoblot analysis was performed on whole-cell detergent extracts from HEK-293 cells
expressing AE1 proteins and protein 4.2. A mouse monoclonal anti-AE1 antibody was used to
detect AE1 proteins (A) and a mouse monoclonal anti-HA antibody was used to detect HA-
tagged protein 4.2 (B). Incubation with peroxidase-conjugated anti-mouse IgG and detection by
chemiluminescence was then performed (see Methods).
total AE1
total mdAE1
total p4.2
80
The second column of Figure 3.2 shows immunofluorescence of AE1 proteins in cells that
have been subsequently permeabilized. In these cells AE1, AE1HS, mdAE1 and kAE1 proteins
showed strong staining at the plasma membrane, indicating that most of the AE1 is at the cell
surface. This was seen by the yellow staining pattern in the merged images showing the strong
overlap of fluorescent signals from intact and permeabilized cells. In contrast, AE1SAO was not
detected at the cell surface, but only in permeabilized cells due to its ER localization (Cheung et
al. 2005a).
3.4.3 Co-localization of protein 4.2 and AE1 in HEK-293 cells
HA-tagged protein 4.2 was expressed alone or with AE1, AE1HS, mdAE1, kAE1 and
AE1SAO proteins, and visualized by immunofluorescence and confocal microscopy to
determine its intracellular localization in HEK cells. The first row of Figure 3.3 shows
immunofluorescence of protein 4.2 expressed alone in permeabilized cells and CNX, an ER-
resident protein. Protein 4.2 (green) displayed a wide distribution throughout the cell and also co-
localized with CNX (blue) as seen by the cyan colour in the merged image. This suggests that
some protein 4.2 localizes at the ER. In the first column of the remaining rows, protein 4.2
displayed a wide cellular distribution when expressed with all AE1 proteins, similar to when
expressed alone. The second column shows the immunofluorescence of AE1 proteins in
permeabilized cells. The proteins were predominantly localized at the cell surface with the
exception of AE1SAO, which was intracellular. In the merged images in the third column, the
yellow colour indicates that a fraction of protein 4.2 co-localized to the plasma membrane with
AE1, AE1HS, mdAE1 and kAE1 proteins. Its co-localization with mdAE1 is interesting since
mdAE1 lacks the cytoplasmic domain, which is needed for protein 4.2 interaction (Korsgren and
Cohen 1986). Protein 4.2 co-localized with AE1SAO, indicating its partial localization at the ER
similar to its co-localization with CNX above. The proteins do not co-localize completely, as
seen by distinct green and red staining in the merged image, indicating localization of protein 4.2
in other cellular compartments. Thus, protein 4.2 has a wide intracellular distribution in
transfected HEK-293 cells localizing to the ER and other intracellular compartments, as well as
to the plasma membrane.
81
non-perm. perm.
AE1 AE1 merge
Figure 3.2: Immunofluorescence
images of wild-type and mutant AE1 in
HEK-293 cells.
Non-permeabilized HEK-293 cells
expressing wild-type or mutant AE1
proteins were incubated with mouse anti-
AE1 antibody (BRIC 6) against an
external epitope, followed by incubation
with Alexa Fluor® 488-conjugated goat
anti-mouse antibody to detect cell surface
AE1 (green). Cells were then washed,
permeabilized and incubated with the
same primary antibody followed by
incubation with Cy3-conjugated goat anti-
mouse antibody to detect total AE1 (red).
Confocal microscopy was then performed.
An enlarged region of the cell is shown in
the inset for each image. In the merged
image, yellow indicates the co-localization
of cell surface and total AE1. non-perm.,
non-permeabilized; perm., permeabilized.
AE1
E40K
G130R
P327R
mdAE1
kAE1
AE1SAO
82
perm. perm.
protein 4.2 CNX merge
protein 4.2 AE1 merge
Figure 3.3: Immunofluorescence images of
protein 4.2 and AE1 proteins in HEK-293
cells.
Permeabilized cells expressing HA-tagged
protein 4.2 alone or co-expressed with wild-
type or mutant AE1 proteins were incubated
with rat anti-HA and rabbit anti-CNX or mouse
monoclonal anti-AE1 (Jennings) antibodies.
Incubation of samples with fluorescently-
labeled secondary antibodies and confocal
microscopy was performed (see Methods). In
the merged image of the top row, cyan
indicates co-localization of protein 4.2 (green)
and CNX (blue). In the remaining merged
images, yellow indicates co-localization of
protein 4.2 (green) and AE1 (red). perm.,
permeabilized.
p4.2
E40K
G130R
P327R
AE1
p4.2
p4.2
p4.2
p4.2
p4.2
kAE1
AE1SAO
p4.2 mdAE1
p4.2
83
3.4.4 Interaction of protein 4.2 with AE1 proteins in HEK-293 cells
To examine the interaction of protein 4.2 with the AE1 proteins we conducted Ni-NTA pull-
down assays on HEK-293 cells transiently expressing His6-tagged AE1 proteins and HA-tagged
protein 4.2. Figure 3.4 shows representative immunoblots of protein 4.2 associated with AE1
(top panel), AE1 pulled down (middle panel), and total protein 4.2 in the HEK-293 lysate
(bottom panel). The amount of protein 4.2 bound to AE1 proteins was calculated as described in
the Methods. The control experiment with protein 4.2 expressed in the absence of AE1 (lane 1)
showed background amounts of bound protein 4.2. The AE1HS mutants consistently bound less
protein 4.2 at levels of 54 ± 14 % (E40K), 43 ± 14 % (G130R), and 65 ± 29 % (P327R), n = 8,
relative to wild-type AE1. Taking into account the dilution factor of the total protein 4.2 fraction
compared to the bound protein 4.2 fraction, it was determined from their band densities that
wild-type AE1 bound approximately 9 % of the total cellular protein 4.2.
As expected, the mdAE1 was unable to pull down protein 4.2 above background levels (lane
6). The kAE1 and AE1SAO proteins bound similar amounts of protein 4.2 at levels of 110 ± 21
% and 99 ± 33 %, n = 8, relative to wild-type AE1, respectively. These small differences were
not statistically significant indicating that the loss of the first 65 amino acids for kAE1 and
deletion of nine amino acids at the first TM in AE1SAO do not diminish protein 4.2 binding.
To confirm the protein 4.2 binding impairment by the AE1HS mutants seen in the Ni-NTA
pull-down assays, we conducted co-immunoprecipitation (co-ip) experiments on HEK-293 cells
transiently expressing AE1HS proteins and protein 4.2. Figure 3.5A shows representative
immunoblots of protein 4.2 associated with AE1 (top panel), AE1 bound to the resin (middle
panel), and total protein 4.2 in the HEK-293 lysate (bottom panel). The amount of protein 4.2
bound to AE1HS proteins relative to wild-type AE1 was calculated as described in the Methods.
The control experiment with protein 4.2 expressed in the absence of AE1 (lane 1) showed no
bound protein 4.2. The AE1HS proteins (lanes 3-5) consistently co-purified less protein 4.2
relative to AE1 (lane 2). In Figure 3.5B we show the results for AE1 (lane 1) and mdAE1 (lane
2) where the amount of immunoreactive mdAE1 expressed was comparable to AE1. As seen in
lane 2, there was no bound protein 4.2 in agreement with the Ni-NTA results.
84
Protein 4.2 expressed with: 1 2 3 4 5 6 7 8
M
empty
vector
AE1
E40K
G130R
P327R
mdAE1
kAE1
AE1SAO
Figure 3.4: Ni-NTA pull-down of wild-type and mutant His6-tagged AE1 with protein 4.2 in
HEK-293 cells.
His6-tagged AE1 in whole-cell extracts was prepared with the detergent C12E8 and pulled-down
using Ni-NTA agarose. Total and bound fractions were resolved by SDS-PAGE and
immunoblots were probed with mouse anti-AE1 (Jennings) antibody to detect AE1 and mouse
anti-HA antibody to detect HA-tagged protein 4.2 (see Methods). Immunoblot band intensities
were measured from eight independent experiments. The amount of protein 4.2 bound to the
P327R HS mutant appears high in this blot. However, after correcting for bound AE1 and total
protein 4.2, the amount of protein 4.2 bound to P327R HS mutant was less than that bound to
AE1.
bound p4.2
bound AE1
bound mdAE1
total p4.2
85
A
Protein 4.2 expressed with: 1 2 3 4 5
M
empty
vector
AE1
E40K
G130R
P327R
B
Protein 4.2 expressed with: 1 2
M
AE1
md
AE1
Figure 3.5: Co-immunoprecipitation (co-
ip) of wild-type and mutant AE1 with
protein 4.2 in HEK-293 cells.
(A) AE1 in whole-cell extracts prepared
with the detergent C12E8 was
immunoprecipitated with rabbit anti-CtAE1
antibody followed by Protein G Sepharose
binding. Total and bound fractions were
resolved by SDS-PAGE and immunoblots
probed with mouse anti-AE1 (Jennings)
antibody to detect AE1 and mouse anti-HA
antibody to detect any co-purifying HA-
tagged protein 4.2 (see Methods).
Immunoblot band intensities were measured
from eight independent experiments. The
mdAE1 lanes were excluded from these
blots due to variable and low expression
level, and the remaining lanes on the same
blots were spliced together for clarity
(splice points are denoted by vertical black
lines).
(B) Immunoblots following co-ip of AE1
and mdAE1 and associated protein 4.2 from
an experiment with comparable amounts of
bound AE1 and mdAE1 proteins.
bound AE1
bound mdAE1
bound p4.2
bound AE1
total p4.2
bound p4.2
86
Also in agreement with the Ni-NTA pull-down results, the three AE1HS mutants consistently
bound less protein 4.2 at levels of 75 ± 16 % (E40K), 68 ± 16 % (G130R), and 79 ± 17 %
(P327R), n = 8, relative to wild-type AE1. Clearly, protein 4.2 could still interact with all three
AE1HS mutants, however at a statistically significant lower level compared to wild-type AE1.
Thus, the HS mutations in the cdAE1 result in impaired binding of protein 4.2 in HEK-293 cells.
3.4.5 Co-localization of wild-type and G2A/C173A protein 4.2 and AE1 in
HEK-293 cells
The ability of protein 4.2 to co-localize with mdAE1 at the plasma membrane of HEK-293
cells despite a lack of interaction was puzzling. Protein 4.2’s requirement of Gly2 myristoylation
for plasma membrane localization in Sf9 cells led us to believe its acylation state may allow for
similar localization in HEK-293 cells in the absence of cdAE1. To confirm this, we constructed
a double mutant removing sites of myristoylation and palmitoylation (G2A/C173A) and
observed its co-localization with AE1 proteins. Figure 3.6 shows the immunofluorescence of
HA-tagged protein 4.2 and the G2A/C173A (GC) mutant expressed with AE1, mdAE1 and
AE1SAO. In the first row, wild-type protein 4.2 is widely distributed throughout the cell and co-
localizes with AE1 at the plasma membrane, as seen by the yellow colour in the merged image.
In the second row, the GC mutant has a similar cellular distribution and co-localizes to the
plasma membrane with AE1. In the third row, the co-localization of protein 4.2 and mdAE1 is
again seen by the yellow colour at the plasma membrane. However, in the fourth row when the
GC mutant is expressed with mdAE1 there is very little yellow staining at the plasma membrane,
indicating poor plasma membrane localization. This suggests that the fatty acid moieties promote
protein 4.2 targeting to the plasma membrane. This also suggests that the presence of wild-type
AE1 allows the GC mutant to localize to the plasma membrane as seen in the second row,
consistent with an interaction between these proteins. The fifth row shows, as before, that wild-
type protein 4.2 also co-localizes with AE1SAO at the ER. In the sixth row, the GC mutant is
also seen to co-localize with AE1SAO at the ER. Co-localization of protein 4.2 and the GC
mutant with AE1SAO is not complete as seen by distinct green and red staining in the merged
image. This indicates that both wild-type and GC protein 4.2 are associated with other
intracellular compartments in addition to the ER.
87
perm. non-perm.
protein 4.2 AE1 merge
perm. perm.
protein 4.2 AE1 merge
Figure 3.6: Immunofluorescence
images of wild-type and
G2A/C173A protein 4.2 and AE1
proteins in HEK-293 cells.
Non-permeabilized cells expressing
wild-type or mutant protein 4.2 and
wild-type or mdAE1 proteins (first
four rows) were incubated with
mouse monoclonal anti-AE1
antibody (BRIC 6) against an
external epitope. Cells were then
permeabilized and incubated with
rat anti-HA antibody. In the last two
rows, permeabilized cells
expressing wild-type or mutant
protein 4.2 and AE1SAO were
incubated with mouse anti-AE1 and
rat anti-HA antibodies. Incubation
of samples with fluorescently-
labeled secondary antibodies and
confocal microscopy was
performed (see Methods). In the
merged images, yellow indicates
co-localization of protein 4.2 and
AE1. perm., permeabilized; non-
perm., non-permeabilized.
p4.2
GC
GC
AE1
AE1SAO
AE1SAO
p4.2 mdAE1
mdAE1 GC
p4.2 AE1
88
3.4.6 Co-localization of wild-type and G2A/C173A protein 4.2 and cell surface
glycans in the absence or presence of AE1 in HEK-293 cells
In order to further support the idea that the GC mutant displays impaired plasma membrane
targeting, HA-tagged protein 4.2 and the GC mutant were expressed either alone or with AE1,
and their localization to the plasma membrane using PNA binding was determined. PNA is a
lectin that binds to the disaccharide galactose β1,3 N-acetylgalactosamine found in
oligosaccharides of glycoproteins (Slupsky et al. 1993), serving as a marker of the outer surface
of the plasma membrane. In the first row of Figure 3.7, protein 4.2 expressed alone is widely
distributed within the cell (green) and the red stain, representing bound PNA, marks the outside
surface of the cell. Even though PNA is on the outside of the bilayer and protein 4.2 on the inner
side, we see an overlap of fluorescence seen by the yellow colour in the merged image since the
confocal microscope cannot resolve the thickness of the membrane. A distinctive green-yellow-
red pattern (going from the inside of the cell outward) is seen in the merged image, indicating
that protein 4.2 associates with the plasma membrane in the absence of AE1. The same pattern is
seen in the second row when protein 4.2 is expressed with AE1. In the third row when the
acylation mutant is expressed alone there is no yellow colour in the merged image, indicating its
inability to associate with the plasma membrane. In the fourth row when the acylation mutant is
co-expressed with AE1 the yellow colour is seen in the merged image, indicating that AE1 can
localize the acylation mutant of protein 4.2 to the plasma membrane.
3.4.7 Subcellular fractionation of HEK-293 cells expressing wild-type or
G2A/C173A protein 4.2
In order to confirm that impaired plasma membrane localization was not indirectly caused by
impaired cytoskeletal attachment by the double mutant we performed subcellular fractionation of
HEK-293 cells expressing HA-tagged wild-type and G2A/C173A mutant protein 4.2. The bulk
of wild-type and GC mutant protein 4.2 was associated with the detergent-insoluble
cytoskeleton. The top two panels in Figure 3.8 are representative immunoblots showing wild-
type and GC protein 4.2 in various fractions. After centrifugation at 100 000 × g the majority of
protein 4.2 (73 ± 3 %, n = 4) was associated with the membrane/cytoskeleton fraction (P1)
versus the cytoplasmic fraction (S1). Similar results (71 ± 6 %, n = 4) were seen with the
acylation mutant.
89
perm. non-perm.
protein 4.2 PNA merge
Figure 3.7: Immunofluorescence images of cell surface glycans using PNA and wild-type
and G2A/C173A protein 4.2 expressed in the absence or presence of AE1 in HEK-293 cells.
Non-permeabilized HEK-293 cells expressing wild-type or mutant protein 4.2 proteins in the
absence or presence of AE1 were incubated with PNA Alexa Fluor® 488 conjugate to detect the
outer membrane. Cells were then permeabilized and incubated with rat anti-HA antibody to
detect HA-tagged protein 4.2 or G2A/C173A. Incubation of samples with fluorescently-labeled
anti-rat secondary antibody and confocal microscopy was performed as described in the
Methods. In the merged image, yellow indicates the co-localization of protein 4.2 (intracellular)
and PNA (extracellular) at the level of the plasma membrane. perm., permeabilized; non-perm.,
non-permeabilized.
PNA
GC+AE1
p4.2+AE1
PNA
p4.2 alone PNA
GC alone PNA
90
Detergent extraction of the membrane/cytoskeletal fraction revealed that most (87 %) of the
membrane-associated protein 4.2 was in the cytoskeletal fraction (P2) with a small amount in the
membrane fraction (S2). Similar results (92 %) were seen with the acylation mutant, although the
amount solubilized by detergent was decreased relative to wild-type protein 4.2 (8 % versus 13
%). These observations support a previous report (Risinger et al. 1996) where myristoylated and
non-myristoylated (G2A) protein 4.2 expressed in Sf9 cells were both associated with the
particulate fraction. These results indicate that fatty acylation does not affect protein 4.2
cytoskeleton association.
The bottom two panels of Figure 3.8 show the detection of actin and GAPDH in the various
fractions of HEK-293 cells expressing wild-type or GC mutant protein 4.2. In HEK-293 cells
most of the actin (~ 75 %) is in the soluble versus the filamentous form (Oprea et al. 2008). In
this study, the majority of the actin was seen in the cytoplasmic fraction (S1) as opposed to the
membrane/cytoskeletal fraction (P1). Of the amount present in the membrane/cytoskeletal
fraction, most of it was in the cytoskeletal fraction (P2). Virtually all of the GAPDH, a soluble
protein, was detected in the cytosol, with a negligible amount detected in the
membrane/cytoskeleton fraction. These results support our assignment of cellular compartments
to the various fractions, allowing us to localize the majority of wild-type and GC protein 4.2 to
the cytoskeletal compartment with confidence.
3.5 Discussion
In the current study, I was able to express full-length HA-tagged protein 4.2 with AE1 proteins in
HEK-293 cells. The AE1HS proteins were expressed at similar levels to AE1 and could traffic to
the plasma membrane of the cells. These findings are consistent with the observation that the
levels of these HS mutants in patient red cells are normal (Inoue et al. 1998, Jarolim et al. 1992a,
Rybicki et al. 1993). I also demonstrated that protein 4.2 was expressed at similar levels in HEK-
293 cells in the absence or presence of AE1 proteins. Thus, protein 4.2 association with AE1 is
not required for its expression in HEK cells.
91
1 2 3 4 5 6 M Total S1 Sw P1 S2 P2
Figure 3.8: Subcellular fractionation of HEK-293 cells expressing wild-type or G2A/C173A
protein 4.2.
Following subcellular fractionation of HEK-293 cells expressing HA-tagged wild-type and GC
mutant protein 4.2, fractions were run on SDS-PAGE and transferred to nitrocellulose. Blots
were probed with mouse anti-HA antibody, mouse anti-actin antibody and mouse anti-GAPDH
antibody to detect protein 4.2, actin and GAPDH, respectively (see Methods). Immunoblot band
intensities were measured from four independent experiments. S1, cytoplasmic fraction; Sw,
wash fraction; P1, membrane/cytoskeletal fraction; S2, detergent-solubilized membrane fraction;
P2, cytoskeletal fraction.
WT p4.2
G2A/C173A p4.2
Actin with WT p4.2
GAPDH with WT p4.2
Actin with GC p4.2
GAPDH with GC p4.2
92
When co-expressed in HEK-293 cells, protein 4.2 and AE1 co-localized at the plasma
membrane. Similar observations were made with the AE1HS mutants and kAE1, but also with
mdAE1, which does not interact with protein 4.2. Furthermore, protein 4.2 could target to the
plasma membrane in the absence of AE1, as seen by its co-localization with PNA. This result
agrees with previous reports of protein 4.2 localization at the plasma membrane of Xenopus
oocytes (Toye et al. 2005) and Sf9 cells (Risinger et al. 1996) in the absence of AE1. We also
found protein 4.2 widely distributed among intracellular compartments, including the ER, as
seen by its co-localization with AE1SAO and CNX. These results show that protein 4.2
associates with various membranes within the cell in the absence of AE1, including the plasma
membrane, the ER membrane, and possibly other intracellular membranes in transfected HEK-
293 cells.
Ni-NTA pull-down experiments revealed that protein 4.2 can associate with AE1 proteins in
whole-cell detergent extracts. Protein 4.2 co-purified with AE1, AE1HS, kAE1 and AE1SAO,
but not mdAE1. Protein 4.2 binding to AE1HS mutants was diminished relative to AE1,
indicating an impaired interaction. Co-immunoprecipitation assays supported the Ni-NTA pull-
down results. We were not expecting binding of protein 4.2 to AE1HS mutants to be completely
abolished since there is only a partial deficiency of protein 4.2 in HS patient red cells.
Introduction of a positive charge by the E40K, G130R or P327R mutants may serve to disrupt
the acidic protein 4.2 binding surface on cdAE1. As shown in the crystal structure of cdAE1
(Figure 1.8), these mutations are not localized to a “hot spot”, but are widely distributed on one
surface of the protein. Residues 40 and 130 on one monomer occur on the same side of the
domain as residue 327 from the other monomer, perhaps requiring dimer formation to create the
binding surface for protein 4.2. The cdAE1-binding regions shown in the protein 4.2 homology
model (Figure 1.8) are predicted to form a binding surface that could be large enough to interact
with all three AE1HS mutation sites. It is not surprising then that alterations of residues at these
distant sites in cdAE1 affect protein 4.2 binding.
The binding of protein 4.2 to kAE1 as measured in Ni-NTA pull-downs was not significantly
different from that of AE1. This result may seem puzzling when we consider that the kidney
isoform is missing the first 65 residues of the N-terminus, while the E40K HS mutant which has
a single point mutation in this region shows impaired binding to protein 4.2. However, this
93
mutation replaces an acidic with a basic residue. This introduction of a positive charge into a
potential binding spot may compromise protein 4.2 binding in a way that absence of the region
does not. As was mentioned earlier, it is believed that the Arg-Arg motif of protein 4.2 (residues
34 and 35) is necessary for cdAE1 interaction and that this motif has an electrostatic interaction
with an acidic region of cdAE1, possibly at the N-terminus. If this acidic region were to become
more basic, the electrostatic repulsion could result in the binding impairment we saw with E40K.
The cytoplasmic domain of kAE1 has been shown to be a more open structure than AE1 (Pang et
al. 2008), as discussed in the Appendix, and it is not known how this would affect protein
binding. It is possible that this open structure compensates for the missing central β-strand. This
kAE1 structural difference and its possible effect on protein 4.2 binding would have to be
explored using structural techniques, such as nuclear magnetic resonance (NMR). Wild-type
levels of AE1SAO binding to protein 4.2 were not surprising as the cytoplasmic domain of this
mutant is unchanged, deficiencies of protein 4.2 in AE1SAO red cells have not been reported,
and AE1SAO exhibits increased binding to the cytoskeleton in red blood cells (Liu et al. 1995).
The co-localization of protein 4.2 with mdAE1 at the plasma membrane, despite a lack of
interaction, turned our attention to the possible role of fatty acid modifications in plasma
membrane localization. We found that G2A/C173A protein 4.2 had diminished plasma
membrane localization when expressed alone or with mdAE1. This suggests that protein 4.2 is
able to associate with the plasma membrane via its lipid anchors, as suggested in previous reports
(Das et al. 1994, Risinger et al. 1996). This is at least the case in HEK-293 cells where AE1 is
not needed for protein 4.2 expression or plasma membrane localization. Subcellular
fractionation studies showed that the majority of wild-type and GC protein 4.2 expressed in
HEK-293 cells was associated with the cytoskeleton. This demonstrates that the GC mutation
does not affect cytoskeletal binding. A portion (< 30 %) of wild-type and GC protein 4.2 was
seen in the cytoplasmic fraction, which agrees with the immunofluorescence localization results
where the protein was broadly localized throughout the cell.
The current study shows that AE1 is not required for the expression and plasma membrane
localization of protein 4.2 in HEK-293 cells. As well, we have shown that the three AE1HS
mutants, E40K, G130R and P327R, have impaired binding to protein 4.2. How this impaired
binding translates into protein 4.2 deficiency in erythrocytes is yet to be determined. Protein 4.2
and AE1 have been found to appear simultaneously during erythropoiesis, and interact with each
94
other as soon as they are expressed (van den Akker et al. 2010a). Indeed, protein 4.2 may
associate with AE1 at the ER during their biosynthesis and they may traffic together to the cell
surface. The fate of protein 4.2 that gets expressed but is unable to fully associate with AE1HS
mutants during differentiation is not known. Protein association with the cytoskeleton can
determine protein sorting during enucleation of the differentiating red cell (Lee et al. 2004),
where an increased cytoskeletal association causes retention in the reticulocyte. It is possible that
the portion of protein 4.2 unable to bind to AE1HS never makes it to the membrane, or does not
localize to a region of the membrane where protein 4.2 can bind to the cytoskeleton. This would
prevent its retention in the reticulocyte during enucleation and explain its deficiency in
erythrocytes of patients with these three cytoplasmic AE1HS mutations. Studies of protein 4.2
sorting during enucleation of red cell precursors from healthy individuals and HS patients, or
relevant HS mouse models, would help to determine the mechanism of protein 4.2 loss in
hereditary spherocytosis.
95
4 Discussion and future directions
In this thesis, I have made two significant findings. The first finding is that three HS mutations,
E40K, G130R and P327R, in cdAE1 do not cause gross conformational changes in the structure
of the domain. However, the P327R mutant displayed a slight thermal destabilization with a Tm
that was 5 ºC lower than that of wild-type cdAE1, as determined by calorimetry and CD thermal
melts. The second finding is that these mutations in full-length AE1 cause impaired binding to
protein 4.2 as determined in HEK-293 cells. In addition to these results, I have created several
other variants of AE1 and protein 4.2 to be used as reagents in future studies. Table 4.1 lists all
the AE1 and protein 4.2 mutant DNA constructs that I have designed and made, their research
purpose and the results of their expression and characterization, if applicable. Following the table
is a more detailed analysis of the major results of this thesis and discussion of future research,
where I will refer back to the table when discussing possible uses for these constructs.
Table 4.1: Mutant protein DNA constructs
Mutation/
Common
Name
Protein/
Vector/
Expression System
Predicted
Product
Comments / Research Purpose / Results
E90K
Cape Town
cdAE1 + His6 tag
pETBlue1
E. coli
HS mutant - causes HS in trans with Prague III
P147S
Mondego
cdAE1 + His6 tag
pETBlue1
E. coli
HS mutant - causes HS in cis with E40K and in trans with Coimbra
- constructed by Allison Pang under my supervision
W75F cdAE1 + His6 tag
pETBlue1
E. coli
Trp mutant - to study specific sites of folding by intrinsic fluorescence
W81F cdAE1 + His6 tag
pETBlue1
E. coli
Trp mutant - to study specific sites of folding by intrinsic fluorescence
W94F cdAE1 + His6 tag
pETBlue1
E. coli
Trp mutant - to study specific sites of folding by intrinsic fluorescence
W105F cdAE1 + His6 tag
pETBlue1
E. coli
Trp mutant - to study specific sites of folding by intrinsic fluorescence
E40A cdAE1 + His6 tag
pETBlue1
E. coli
Ala mutant - to test site of HS mutation using Ala
- wild-type CD spectrum
G130A cdAE1 + His6 tag
pETBlue1
E. coli
Ala mutant - to test site of HS mutation using Ala
- wild-type CD spectrum
P327A cdAE1 + His6 tag
pETBlue1
E .coli
Ala mutant - to test site of HS mutation using Ala
- wild-type CD spectrum
96
Table 4.1: Mutant protein DNA constructs (continued)
Mutation/
Common
Name
Protein/
Vector/
Expression System
Predicted
Product
Comments / Research Purpose / Results
C201A cdAE1 + His6 tag
pETBlue1
E. coli
single-Cys - to make the Cys-less mutant
- can use for Cys-scanning methods
C317A cdAE1 + His6 tag
pETBlue1
E. coli
single-Cys - to make the Cys-less mutant
- can use for Cys-scanning methods
C201/317A cdAE1 + His6 tag
pETBlue1
E. coli
Cys-less - same CD spectrum as WT, maybe slightly higher thermal
stability by calorimetry
- allows use of Cys-scanning methods
- helped determine extra mass on cdAE1 was β-
mercaptoethanol adduct by mass spectrometry
kC201A cdAE1 + His6 tag
pETBlue1
E. coli
kidney
single-Cys
- to make the Cys-less mutant
- can use for Cys-scanning methods
kC317A cdAE1 + His6 tag
pETBlue1
E. coli
kidney
single-Cys
- to make the Cys-less mutant
- can use for Cys-scanning methods
L319K cdAE1 + His6 tag
pETBlue1
E. coli
monomer - puts positive charge in hydrophobic pocket and β-strand
of dimer interface to disrupt dimer
- mixture of monomers and dimers by SE
L319E cdAE1 + His6 tag
pETBlue1
E. coli
monomer - puts negative charge in hydrophobic pocket and β-strand
of dimer interface to disrupt dimer
- not pure enough for SE analysis
L321K cdAE1 + His6 tag
pETBlue1
E. coli
monomer - puts positive charge in hydrophobic pocket and β-strand
of dimer interface to disrupt dimer
- not pure enough for SE analysis
L321E cdAE1 + His6 tag
pETBlue1
E. coli
monomer - puts negative charge in hydrophobic pocket and β-strand
of dimer interface to disrupt dimer
- not pure enough for SE analysis
L319K/L321K cdAE1 + His6 tag
pETBlue1
E. coli
monomer - puts two positive charges in hydrophobic pocket and β-
strand of dimer interface to disrupt dimer
- dimer by SE
L319E/L321E cdAE1 + His6 tag
pETBlue1
E. coli
monomer - puts two negative charges in hydrophobic pocket and β-
strand of dimer interface to disrupt dimer
- monomer by SE
Δh10 cdAE1 + His6 tag
pETBlue1
E. coli
monomer - missing helix 10 of dimerization arm
- mixture of monomers and dimers
Δh10,s11 cdAE1 + His6 tag
pETBlue1
E. coli
monomer - missing helix 10 and strand 11
- no expression
Δh9,10,s11 cdAE1 + His6 tag
pETBlue1
E. coli
monomer - missing helix 9 and 10, and strand 11
- low expression, poor purification, poor solubility
Δh8, 9,10,s11 cdAE1 + His6 tag
pETBlue1
E. coli
monomer - missing helix 8, 9, and 10 and strand 11
- low expression, poor purification, poor solubility
97
Table 4.1: Mutant protein DNA constructs (continued)
Mutation/
Common
Name
Protein/
Vector/
Expression System
Predicted
Product
Comments / Research Purpose / Results
E40K
Montefiore
AE1-His6
pcDNA3
HEK-293 cells
HS mutant
(AE1HS)
- used in Ni-NTA pull-down assays
- good expression
G130R
Fukuoka
AE1-His6
pcDNA3
HEK-293 cells
HS mutant
(AE1HS)
- used in Ni-NTA pull-down assays
- good expression
P327R
Tuscaloosa
AE1-His6
pcDNA3
HEK-293 cells
HS mutant
(AE1HS)
- used in Ni-NTA pull-down assays
- good expression
Wild-type protein 4.2-GST
pGEX5-1
E. coli
full-length
wild-type
- for in vitro binding studies with cdAE1
- no expression
1-238 aa protein 4.2-GST
pGEX5-1
E .coli
23kDa N-
terminus
- for in vitro binding studies with cdAE1
- poor expression and purification
1-144 aa protein 4.2-GST
pGEX5-1
E. coli
cdAE1-
binding site
- for in vitro binding studies with cdAE1
- contains Arg34-Arg35 binding motif
- poor expression and purification
145-238 aa protein 4.2-GST
pGEX5-1
E. coli
cdAE1-
binding site
- for in vitro binding studies with cdAE1
- contains palmitoylatable Cys173
- good expression, fair purification with degradation
145-203 aa protein 4.2-GST
pGEX5-1
E. coli
β-hairpin - for in vitro binding studies with cdAE1
- contains palmitoylatable Cys173
- good expression and purification
Wild-type protein 4.2-HA
pcDNA3
HEK-293 cells
full-length
wild-type
- for immunofluorescent localization and pull-down assays
- good expression and plasma membrane localization
- good binding to AE1, impaired binding to AE1HS
C173A protein 4.2-HA
pcDNA3
HEK-293 cells
acylation
mutant
- for immunofluorescence localization: trafficking effects
of palmitoylation
G2A/C173A protein 4.2-HA
pcDNA3
HEK-293 cells
double
acylation
(GC) mutant
- for immunofluorescence localization: trafficking effects
of palmitoylation and myristoylation
- good expression and poor plasma membrane localization
Arg34-Arg35
Glu34-Glu35
protein 4.2-HA
pcDNA3
HEK-293 cells
binding
rescue
mutant
- for pull-down assays to measure interaction with AE1
and E40K: possible interaction rescue mutant
- good expression and preliminary co-ip assay showing
rescue of binding to E40K AE1
E40K
Montefiore
AE1-myc
pFB-Neo
K562 cells
HS mutant
(AE1HS)
- used in immunofluorescent localization
- good expression and localization
- unable to make stable transfectants
G130R
Fukuoka
AE1-myc
pFB-Neo
K562 cells
HS mutant
(AE1HS)
- used in immunofluorescent localization
- good expression and localization
- unable to make stable transfectants
P327R
Tuscaloosa
AE1-myc
pFB-Neo
K562 cells
HS mutant
(AE1HS)
- used in immunofluorescent localization
- good expression and localization
- unable to make stable transfectants
Wild-type protein 4.2-HA
pFB-Neo
K562 cells
full-length
wild-type
- used in immunofluorescent localization
- variable expression and inconsistent localization
- unable to make stable transfectants
98
4.1 Structure and conformational stability of the cytoplasmic domain
of AE1
The E40K, G130R and P327R mutations in cdAE1 are associated with HS and a lower amount
of protein 4.2 in the red cell, while maintaining a normal amount of AE1 at the plasma
membrane. I hypothesized that these mutations would cause either a change in the structure of
the domain or a change in the interaction surface, thereby preventing proper binding of protein
4.2. I found that these mutations do not cause gross changes in the structure, folding or
conformational stability of the domain. The secondary structures of these mutant proteins were
similar to wild-type as determined by CD. The mutant proteins retained the dimeric structure of
the wild-type protein, as determined by SE experiments using the analytical ultracentrifuge. The
P327R mutant was expected to affect the dimeric state since this mutation is located at the N-
terminus of an α-helix in the dimerization domain, but this and another study (Zhou et al. 2007)
confirmed that it remained a dimer. This mutation, however, caused a slight thermal
destabilization of the domain with a Tm that was 5 ºC lower than for wild-type cdAE1.
The mutant proteins underwent similar pH-dependent conformational changes when
monitored by intrinsic tryptophan fluorescence. In this experiment, the fluorescence intensity
increased in a similar manner for all the proteins as pH was raised between 5 and 10,
representing a dequenching of tryptophan residues. Tryptophan residues located within the
interior of globular proteins are most likely hydrogen-bonded to other groups (Voet and Voet
2004). The cdAE1 protein has four tryptophans per subunit: W75, W81, W94 and W105 (see
Figure A1). In fact, a hydrogen bond between W105 of one subunit and D316 of the other
subunit that is present in the lower pH conformation is broken in a higher pH conformation of
cdAE1 as the peripheral protein binding domain moves away from the dimerization arm (Zhou
and Low 2001). Nearby groups can quench tryptophan fluorescence, hence the increase in
fluorescence intensity as the protein begins to open up with breaking of hydrogen bonds caused
by increasing pH. There was also a similar increase in peak wavelength (red-shift) at alkaline
pH, indicating that the tryptophans were exposed to a more polar environment. These results
indicate that the proteins have similarly folded structures. Urea denaturation of proteins
monitored by intrinsic fluorescence showed no significant differences in the sensitivity of the
proteins to this chemical denaturant, again indicating the proteins had similarly folded structures.
I have constructed cdAE1 variants where each of the four tryptophan residues in cdAE1 has been
99
mutated to phenylalanine: W75F, W81F, W94F and W105F. These cdAE1 Trp mutants, listed in
Table 4.1, can be used to study their role in stabilizing the folded structure of cdAE1, and further
pinpoint specific regions of the protein affected by various denaturing conditions. As well, these
mutants could potentially be used as templates to make cdAE1 variants with single tryptophan
residues. Replacement of tryptophan with phenylalanine is a conservative substitution in terms of
retaining hydrophobicity, but its hydrogen-bonding capability will be removed which may have
an impact on protein folding.
The thermal stabilities of the mutant proteins were measured by CD during thermal melts
and by DSC. By both methods the E40K and G130R mutants were found to have similar
stabilities to the wild-type cdAE1. The P327R mutant had an apparent midpoint transition that
was 5 ºC lower than wild-type indicating a slight thermal destabilization of the domain. Another
research group using SDSL in combination with EPR and DEER spectroscopies found that this
mutant had no global effect on the structure, but did affect the packing of the helix in the
dimerization arm where it resides (Zhou et al. 2007). The P327 residue from one subunit occurs
on the same side of the dimer as E40 and G130 from the other subunit. This made the creation of
a monomer mutant of cdAE1 desirable, both for structural purposes and for binding assays. It is
likely that formation of the dimer is necessary for protein 4.2 binding since the three sites of HS
mutation may form part of a large binding surface.
The dimer interface is composed of a β-sheet formed by one β-strand from each monomer.
As well, nine leucine residues contributed from both subunits interact in a hydrophobic pocket.
Two of these leucine residues, Leu319 and Leu321, also reside on the β-strands used to form the
β-sheet. Given the dual role of these leucine residues, I decided they were good candidates to
mutate for dimer disruption. As listed in Table 4.1, I mutated one or both of Leu319 and Leu321
to lysine or glutamate in hopes that introduction of one or two charged residues into the
hydrophobic pocket would disrupt the hydrophobic interactions and weaken the bonding
between β-strands. I also made more extreme mutations where I deleted successive helices and
strands from the C-terminus of cdAE1, which includes the dimerization arm. Deletion of the
final helix, but retention of the β-strand, resulted in formation of a mixture of dimers and
monomers as determined by SE experiments using the analytical ultracentrifuge. Deletion of the
β-strand in the other three deletion mutants resulted in low to no protein expression. Of the single
point mutants, only L319K yielded pure enough protein for analysis, but resulted in a dimer-
100
monomer mix. The L319/321K double mutant resulted in dimer formation, but the L319/321E
double mutant resulted in a monomer and was further characterized. As determined by
calorimetry, it was not surprising that the melting temperature was lower for the L319/321E
monomer than for wild-type cdAE1 since dimer dissociation was not a part of the thermal
unfolding pathway. This monomer could be used in future binding studies with protein 4.2 to
determine whether dimer formation is in fact necessary for conformational stability and protein
interactions.
The cdAE1 mutant A285D (Band 3 Boston) represents a point mutation in the cytoplasmic
domain resulting in defective protein expression leading to HS (Jarolim et al. 1996). The mutant
mRNA is present in normal amounts but AE1 protein level is decreased suggesting the mutation
results in a protein product that gets degraded. Deficiency of AE1 at the cell surface removes
points of contact between the membrane and cytoskeleton, leading to HS. It would certainly be
worthwhile to study AE1 A285D in transfected HEK-293 cells to compare its expression level
and protein half-life to those of wild-type AE1. However, in cdAE1 with the three HS mutations
E40K, G130R and P327R, protein degradation of AE1 is not the cause of HS since normal
amounts of AE1 are detected at the red cell membrane and almost all proteins known to associate
with cdAE1 are seen in normal amounts in patient red cells. The only deficient protein in the
whole AE1 complex is protein 4.2. This led to the hypothesis that these three HS mutations,
through conformational changes or interaction surface changes, cause impaired binding of
protein 4.2 to cdAE1 resulting in protein 4.2 deficiency. This was the basis for the next part of
the present study.
4.2 Protein 4.2 interaction with HS mutants of cdAE1
Full-length protein 4.2 HA-tagged at its C-terminus was successfully expressed in the HEK-293
human cell line. Previous unsuccessful attempts prompted researchers to instead express only the
23 kDa N-terminal region with a FLAG tag (Mandal et al. 2003). The expression of full-length
protein 4.2 in this thesis was possible in the presence or absence of AE1, indicating that AE1 was
not needed for protein 4.2 expression in these cells, as it may be in red blood cells. Protein 4.2
was found throughout the HEK-293 cells, including the cytoplasm, ER and plasma membrane. It
was predominantly associated with the cytoskeletal fraction. It was able to reach the plasma
membrane in the absence of AE1, showing its lack of dependence on AE1 for membrane
101
localization. When co-expressed, protein 4.2 co-localized with AE1HS mutants at the plasma
membrane, as well as with kAE1 and mdAE1. Since protein 4.2 is unable to interact with the
membrane domain (Korsgren and Cohen 1986) this co-localization does not suggest any
interaction between protein 4.2 and AE1, but simply indicates that protein 4.2 is able to localize
to the plasma membrane. Protein 4.2 co-localized with AE1SAO at the ER indicating its ability
to localize to that internal membrane. Again, interaction cannot be inferred from this result since
protein 4.2 also co-localized with CNX at the ER in the absence of AE1SAO. The broad
distribution of protein 4.2 in HEK-293 cells may be a consequence of over-expression of this
protein.
I showed that protein 4.2 was able to interact with AE1 using co-ip and Ni-NTA pull-down
assays. This interaction has been shown in the past between red cell-purified protein 4.2 and red
cell IOVs (Korsgren and Cohen 1986) as well as with chymotrypsin-released cdAE1 from red
cells (Korsgren and Cohen 1988). The interaction has also been shown in pull-downs from
Xenopus oocytes (Toye et al. 2005), and now I have demonstrated it in human HEK-293 cells.
This expression system allows for an efficient and economical analysis of protein 4.2 binding
with AE1 and its mutants in a human cell line. AE1HS proteins were also able to bind protein
4.2, but the binding was impaired. The binding was not abolished by these mutants, and was not
expected to be, since even in patient red cells there is some protein 4.2 present. It is possible that
the binding interface between protein 4.2 and cdAE1 is so large that one point mutation will not
cause complete binding impairment since other parts of the binding surface remain intact. The
introduction of a positive charge by each of the three HS mutations may cause electrostatic
repulsion, enough to repel the region of protein 4.2 that binds there.
Surprisingly, kAE1 was able to bind similar amounts of protein 4.2 as erythroid AE1. The
kidney isoform is missing the first 65 residues, which includes the site of the E40K mutation.
Since a positively-charged Arg-Arg motif in protein 4.2 has been found to be essential for
interaction with cdAE1 (Rybicki et al. 1995), it may be that introduction of a positive charge into
a very acidic binding region of cdAE1 compromises the binding in a way that absence of this
region does not. The kidney isoform, with its more open structure (Pang et al. 2008), may
compensate structurally for the missing central β-strand in such as way as to accommodate the
binding of protein 4.2. Wild-type levels of AE1SAO binding to protein 4.2 were not surprising as
102
the cytoplasmic domain of this mutant is unchanged and deficiencies of protein 4.2 in AE1SAO
red cells have not been reported.
The fatty acylation mutant of protein 4.2, G2A/C173A (GC), showed reduced plasma
membrane localization in the absence of AE1 and in the presence of mdAE1. When co-expressed
with wild-type AE1, it localized to the plasma membrane indicating a protein interaction. Both
wild-type and GC protein 4.2 were found to be mostly associated with the cytoskeleton of HEK-
293 cells. This would put protein 4.2 in a good position for binding to plasma membrane
proteins, such as AE1. The fatty acyl moieties on protein 4.2 would also help it interact with the
plasma membrane. With GC protein 4.2 lacking these fatty acid modifications, even though it is
present in the same amounts as wild-type protein 4.2 at the cytoskeleton, it is unable to interact
with the plasma membrane on its own. It is only when co-expressed with AE1 that it becomes
localized to the plasma membrane, probably through protein interaction.
Future studies would focus on characterizing the interaction between cdAE1 and protein 4.2
in vitro by using purified full-length protein 4.2 and purified cdAE1 proteins. In the absence of
pure protein 4.2, a detergent-free cell lysate containing the GC mutant form of protein 4.2 could
be used in semi-quantitative binding experiments with pure cdAE1. Lack of detergent may allow
for more native interactions to occur between protein 4.2 and cdAE1 since these proteins interact
in the cytoplasm, not the membrane. I have grown HEK-293 cells expressing protein 4.2 and the
GC mutant and lysed them using detergent, and by needle aspiration followed by sonication,
respectively. The latter method yielded more protein in the soluble versus particulate fraction.
His6-tagged cdAE1 and protein 4.2-expressing HEK-293 lysate could be incubated together
followed by pull-down on Ni-NTA resin. Alternatively, the His6-tagged cdAE1 could be first
immobilized on Ni-NTA resin, to mimic the situation in red cells where cdAE1 is immobilized at
a high density at the plasma membrane as part of AE1. Then HEK-293 lysate expressing protein
4.2 would be added for binding to cdAE1. The advantage of using HEK-293 cells for protein
expression is that full-length, and presumably properly folded, protein 4.2 can be obtained, as
opposed to expression in E. coli or cell-free translation. The disadvantage of this method is that
by using a cell lysate, many other proteins are present and we cannot say with certainty that an
interaction reflects direct binding between proteins. Ni-coated microplates could also be used for
this purpose allowing measurement of protein 4.2 binding to several cdAE1 variants and
103
concentrations at one time. However, the same advantages and disadvantage apply to this
method as with the Ni-NTA pull-down.
A variation on this pull-down assay would be to use a rabbit reticulocyte cell-free system to
express protein 4.2 using [35
S]-methionine for detection. This would provide radiolabeled protein
4.2 to be used in pull-down assays. This system would allow for protein 4.2 to be enriched in the
lysate, although it would contain a large amount of hemoglobin which also binds to cdAE1. An
advantage would be that radiographic detection would be much more sensitive and direct, since
[35
S]-Met would be incorporated into protein 4.2, thereby removing the primary and secondary
antibody steps involved in immunodetection. I have expressed full-length protein 4.2 in this
system and preliminary results have shown its ability to bind cdAE1 and HS mutants above
background. However, the yield for this expression system is low and obtaining higher yields
would be very costly.
There has been much difficulty in expressing and purifying full-length protein 4.2 in
bacteria by myself and others (Bhattacharyya et al. 1999, Korsgren et al. 2010). For this reason I
have created several protein 4.2 GST-fusion proteins in the hope that at least the cdAE1-binding
regions could be expressed in E. coli and purified using glutathione resin for use in binding
assays (see Table 4.1). The boundaries of these fusion proteins were determined based on natural
protease cleavage sites (Bhattacharyya et al. 1999), regions known to bind cdAE1
(Bhattacharyya et al. 1999, Rybicki et al. 1995) and predicted surface-exposed residues based on
the protein 4.2 homology model (Figure 1.8). I designed protein 4.2 GST fusion proteins to
express full-length protein 4.2, the 23 kDa N-terminal domain (residues 1-238), the cdAE1-
binding domain encompassing residues 1-144, the cdAE1-binding region encompassing residues
145-238 and the region predicted to form a β-hairpin (residues 145-203). Following vector
transformation into E. coli, only the shortest fusion proteins containing the predicted β-hairpin
(Figure 1.8) encompassing residues 145-238 and 145-203 were successfully expressed and
purified. Since this hairpin region appears to be important for cdAE1 binding (Bhattacharyya et
al. 1999) these GST-fusion proteins could be used in binding assays with wild-type and HS
cdAE1 to determine the hairpin-binding region on cdAE1. These fusion proteins could be used in
Ni-NTA pull-down assays where His6-tagged cdAE1 proteins are first immobilized on resin
followed by fusion protein incubation. The reciprocal experiment could also be performed where
GST-fusion proteins are first immobilized on glutathione resin followed by cdAE1 incubation. If
104
one of the sites of HS mutation were a region of interaction for any particular protein 4.2
fragment, this would become obvious in the binding studies. The hairpin fragment could also be
used in competition binding assays with full-length protein 4.2.
These fusion proteins could also be used in far western, or blot overlay, assays. In this case
fusion proteins would be run on SDS-PAGE and transferred to nitrocellulose. Wild-type and HS
mutant cdAE1 proteins would then be added to separate protein 4.2-GST blots, followed by a
western blot procedure using anti-cdAE1 antibodies to detect and compare binding. I have
performed the reciprocal version of this experiment where cdAE1 proteins are run on the gel and
transferred to the blot, followed by incubation with β-hairpin-GST fusion protein. Results were
inconclusive, partly owing to the fact that cdAE1 probably needs to dimerize for protein 4.2
binding, but SDS-PAGE is denaturing and most likely does not allow proper refolding of the
cdAE1 domain. The same problem could occur when the protein 4.2 GST fusion proteins are run
on the SDS-PAGE gel.
It may be preferable to use an expression system that allows protein 4.2 to be myristoylated
and palmitoylated, as it is in nature, since these modifications may have an effect on cdAE1
binding. Because of the need for these co- and post-translational modifications and protein
aggregation problems, bacterial systems are not ideal. Using purification tags, such as a His6 tag
or GST tag, would allow purification of protein 4.2 from HEK-293 cells on Ni-NTA or
glutathione columns, respectively. Alternatively, Protein G-Sepharose and an anti-protein 4.2
antibody could be used to purify protein 4.2, or an anti-HA antibody to purify HA-tagged protein
4.2. But this would be more costly, and protein 4.2 would still have to be separated from the
antibody used to purify it. In any case, the expression of protein 4.2 in HEK-293 cells would
have to be scaled up quite a bit to achieve high enough yields for structural and in vitro binding
studies.
I have also expressed the AE1HS mutants and protein 4.2 in K562 cells, an erythroleukemia
cell line. The advantage of these cells is that they are from the erythroid lineage and may contain
many of the same proteins in erythroid precursors that make for a more native environment for
AE1 and protein 4.2. I had hoped that by using a viral transfection method with these cells,
including viral expression vector pFB-Neo (see Table 4.1), I could create stable transfectants, but
this proved to be quite difficult. As well, while expression of AE1 proteins and their plasma
105
membrane localization were consistent, this was not the case with protein 4.2. These cells gave
variable protein 4.2 expression and localization, where protein 4.2 was sometimes seen at the
plasma membrane and other times throughout the cell. Successful expression was heavily
dependent on the health and age of the cells and their use became increasingly cumbersome. For
this reason, HEK-293 cells were used for interaction and localization studies and gave more
reliable and reproducible results.
4.2.1 Quantitative in vitro binding analysis
Binding of purified protein 4.2 to HS mutants of cdAE1 could be performed, as well as to other
cdAE1 variants where selected residues have been mutated to determine specific binding
regions. By using quantitative methods such as isothermal calorimetry (ITC), surface plasmon
resonance (Biacore) or analytical ultracentrifugation, we could assign importance to each
mutation site depending on its ability to disrupt the protein interaction. The major advantage of
these quantitative methods is that binding affinities can be determined for each interaction.
Another advantage is that direct interactions between two proteins can be measured. I would be
able to compare direct binding between cdAE1 and the HS mutants. I have also created alanine
mutants in cdAE1 where alanine has replaced each of the HS mutations: E40A, G130A and
P327A (Table 4.1). These mutant proteins have similar CD spectra to wild-type cdAE1. Using
these Ala mutants in binding assays with protein 4.2 will help determine whether the binding
defect is caused by the introduction of a positive charge in cdAE1 or because of the absence of
specific residues, namely glutamate, glycine and proline. Perhaps these residues have specific
roles in the binding that can only be fulfilled by the native residue or those with similar
properties. These Ala mutations can also be made in full-length AE1 for expression in HEK-293
cells and for use in co-ip assays. The monomer mutant of cdAE1 that I have constructed could
also be used in binding assays to determine whether dimer formation is necessary for protein 4.2
binding. Double and triple HS mutations could also be made in cdAE1, for example, the double
mutant E40K/G130R and the triple mutant E40K/G130R/P327R. An additive effect by these
mutants would be seen as a more pronounced protein 4.2 binding impairment than was seen for
any of the single HS mutants. This would strengthen the idea that these three HS mutation sites
contribute to a large protein 4.2-binding surface.
Mutations would also be made in protein 4.2 in order to characterize important residues
involved in the interaction. I have already created a protein 4.2 mutant where the important
106
Arg34-Arg35 motif has been mutated to Glu34-Glu35 (EE protein 4.2) and I have done
preliminary co-ip experiments in HEK-293 cells with wild-type and E40K AE1. If the Arg-Arg
motif does indeed interact with E40 in cdAE1 as proposed (Rybicki et al. 1995), and the lysine at
position 40 does electrostatically repel the Arg-Arg motif, then it is reasonable to expect the EE
mutant to restore E40K binding to wild-type levels. In this case, the salt bridge would be
restored. In fact, in my preliminary co-ip experiments I saw that EE protein 4.2 was able to bind
E40K AE1 to a similar level as wild-type protein 4.2 binding to wild-type AE1. More trials
would have to be done to confirm these results. As well, purification of the EE protein 4.2
mutant would enable more quantitative in vitro binding studies with wild-type and E40K cdAE1.
It would be interesting to study the AE1-protein 4.2-ankyrin complex, since protein 4.2
binds to both cdAE1 and ankyrin (Korsgren and Cohen 1988) and strengthens the AE1-ankyrin
interaction (van den Akker et al. 2010a). Even though these three proteins exist in a complex, a
fraction of AE1 extracted from red cells is associated only with protein 4.2 (Yu and Steck 1975)
showing that these two proteins can interact on their own without ankyrin. Study of a three-
protein complex is much more complicated than a two-protein complex, hence our focus at this
point only on the AE1-protein 4.2 interaction. In the future, addition of ankyrin into the mix will
help to further characterize the interaction of these three proteins and the effect of AE1HS
mutations. Ankyrin is, however, a very large cytoskeletal protein consisting of 1880 residues,
with many binding partners (Lux et al. 1990). The N-terminal region of ankyrin is the
membrane-binding domain and contains 24 ankyrin repeats. The D34 region contains repeats 13-
24 (residues 402-827) and is known to bind to cdAE1 and protein 4.2 (Su et al. 2006). The D34
domain is easily expressed in E. coli and purified using affinity tags engineered into the
recombinant protein. The use of purified cdAE1, protein 4.2 and D34 ankyrin would make
possible the characterization of this three-protein complex by quantitative binding studies.
4.2.2 Structure determination
Determining the binding interface between cdAE1 and protein 4.2 by introducing point
mutations and measuring protein affinities would help in the identification of important residues
involved in binding, and the effect of various natural mutations. However, a structure of the
cdAE1-protein 4.2 complex would be far more informative. A crystal structure of cdAE1 at low
pH has been solved (Zhang et al. 2000a), but no structure of protein 4.2 exists as of yet, let alone
in complex with cdAE1. The problem of protein 4.2 structure determination lies with expression
107
and purification of high amounts of protein. It may be that expression of protein 4.2
simultaneously with cdAE1 would help stabilize it, allowing for purification of the complex.
Expressing the proteins together may also negate the need for detergents to solubilize protein 4.2.
Since protein 4.2 and AE1 appear together and interact as soon as they are expressed in red blood
cell precursors (van den Akker et al. 2010a) this may provide a more suitable environment for
the production of a complex. In this case the proteins would be purified together. Since a
complex is what we seek for structure determination, this may be the ideal situation. cdAE1
could only be crystallized at low pH and this structure may not completely represent the native
form. However, structural features determined by SDSL in combination with EPR and DEER
spectroscopies at neutral pH were in agreement with those in the crystal structure (Zhou et al.
2005b). Co-expression and co-purification of cdAE1 with protein 4.2 may also allow structure
determination of both proteins at neutral pH if it is found that the proteins stabilize each other in
solution. Another way to obtain high levels of a cdAE1-protein 4.2 complex would be to
trypsinize inside-out red cell ghosts which would release cdAE1 from the membrane along with
interacting protein 4.2. Even if only 10 % of AE1 is associated with protein 4.2 at the membrane
(Yu and Steck 1975) that is still a large number of cdAE1-protein 4.2 complexes that can be
collected, since there are about 1.2 million AE1 molecules per red cell. Removal of protein 4.2
from IOVs required harsh alkaline conditions (Steck and Yu 1973) which are needed to disrupt
the cdAE1-protein 4.2 interaction. Such harsh conditions may not be required if protein 4.2 is to
be isolated in a complex with cdAE1, since the latter may be proteolytically cleaved from the
membrane. From there, the complex may be purified using traditional protein purification
techniques, such as anion exchange chromatography and gel filtration chromatography.
Determining the structure of the monomer mutant of cdAE1 would be necessary to
determine whether the monomer is folded in the same way as the individual subunits within the
dimer, or becomes misfolded when expressed alone. Binding experiments could proceed with
confidence if it is discovered that the monomer truly represents a single subunit from the dimer
that is properly folded. Following purification of large amounts of monomer cdAE1,
crystallization screens would be performed to determine the best conditions for crystal
formation. Alternatively, NMR could be used for structure determination of the cdAE1 monomer
as well as wild-type cdAE1 at neutral pH. A structure for the kidney isoform of cdAE1 would
also be desirable, as well as that of the variant missing the N-terminal 54 residues from cdAE1
(cdΔ54AE1). I have obtained preliminary 2D NMR spectra for all of these proteins, some at
108
varied pH, temperature, time and magnetic field strength. Thus far, the monomer cdAE1 gives
the best spectra with the most resolved peaks, while cdAE1, cdkAE1 and cdΔ54AE1 appear to
be aggregated. The poor spectra may be due to the size of the proteins since cdAE1 forms a
dimer of 43 kDa subunits. Conditions would have to be optimized for all constructs in order to
obtain spectra that can be reasonably compared.
4.2.3 Interaction of protein 4.2 and AE1 during red cell development
Great interest lies in studying the dynamic interactions that occur during red cell development.
Protein 4.2 and AE1 have been found to appear simultaneously during early erythropoiesis, and
interact with each other as soon as they are expressed (van den Akker et al. 2010a) as determined
from differentiating erythroblasts. CD34+ cells are early hematopoietic progenitors that can be
stimulated to differentiate into erythroblasts, and later into erythrocytes. CD34+ cells mainly
reside in adult bone marrow, but can also be isolated from peripheral blood (Lataillade et al.
2005). The use of CD34+ cells from peripheral blood of healthy donors and AE1HS patients
would allow the study of protein 4.2 interactions with AE1 and AE1HS from the moment they
are expressed and interact, and onward in the differentiation process. Co-immunoprecipitation
assays using anti-AE1 or anti-protein 4.2 antibodies could be employed at various stages of
differentiation. We would be able to study the trafficking of the AE1-protein 4.2 complex as it
moves through the ER, Golgi complex, and finally to the plasma membrane. This would also
allow us to determine the fate of protein 4.2 in red cells of HS patients where impaired AE1-
protein 4.2 interaction leads to protein 4.2 deficiency. CD34+ cells would be superior to
transformed erythroid cell lines, such as K562 erythroleukemia cells, since they represent the
most native erythrocyte environment.
4.3 Conclusions
In my thesis I have investigated the structure and conformational stability of three HS mutants of
cdAE1 and their interaction with protein 4.2.
I have expressed and purified the three HS mutants of cdAE1, E40K, G130R and P327R, in
E. coli. Through various biophysical methods I discovered that the mutations do not cause any
gross conformational changes in the protein other than a slight thermal destabilization caused by
P327R, which had a Tm 5 ºC lower than that of wild-type cdAE1. This supports the finding that
normal amounts of these mutants occur at the plasma membrane of red blood cells. Many
109
disease-causing mutations result in misfolded and degradation-prone proteins, if they get
expressed at all. The three HS mutations at the centre of this thesis are part of a minority of
disease-causing missense mutations that are stably expressed in the cell, yet affect the functional
properties of the protein (Gregersen et al. 2000). In the case of these three HS mutations, the
disease mechanism is quite different from the straightforward mutation resulting in degradation
of protein.
I have expressed full-length HS mutants of AE1 in HEK-293 cells and have shown that their
expression levels are similar to that of wild-type AE1 and they localize to the plasma membrane
of these cells. I have also expressed full-length protein 4.2 in these cells and have shown that its
expression levels are similar in the absence or presence of AE1. I demonstrated the broad
localization pattern of protein 4.2 in HEK cells, including at the plasma membrane even in the
absence of AE1. I showed co-localization of protein 4.2 at the membrane with AE1, AE1HS
mutants, kAE1 and mdAE1 and its co-localization with AE1SAO at the ER. I showed that the
three AE1HS mutants had impaired binding to protein 4.2 as was expected. The partial, versus
complete, deficiency of protein 4.2 in HS patient red cells supports the moderate, but significant,
binding impairment seen in HEK cells. I also showed that kAE1 and AE1SAO bind to protein
4.2 as well as wild-type AE1. Fatty acylation of protein 4.2 was necessary for plasma membrane
localization in the absence of AE1, as demonstrated by the fatty acyl mutant G2A/C173A (GC)
protein 4.2. By subcellular fractionation analysis, fatty acylation was shown to have no effect on
the ability of protein 4.2 to associate with the cytoskeleton.
Based on the crystal structure of cdAE1, the three HS mutations are located far apart from
each other in the folded protein. Since all three mutations affect protein 4.2 binding, each most
likely contributes to a large binding surface. Protein 4.2 is a fairly large protein at 72 kDa and,
based on its homology model, is large enough to interact with each of the HS mutation sites on
cdAE1 at once. The deficiency of protein 4.2 in HS red cells caused by these mutations is, at
least in part, a result of this impaired interaction.
Protein-protein interactions have become intriguing drug targets in recent years, not for
inhibiting interactions, as has been the standard methodology, but for promoting them (Corson et
al. 2008). This can be done using bifunctional ligands that bind two different protein targets
together. These ligands are composed of two different protein-binding moieties that are joined
110
together by linkers of appropriate length. The protein-binding moieties can be natural or
synthetic protein-binding small molecules, and the linkers can be polymers such as polyethylene
glycol chains of varying length. In a disease such as HS, which is caused by impaired protein-
protein interactions, patients may benefit from such a strategy. This type of therapy requires
specific knowledge of the binding surfaces of the proteins that are to be brought together. It is
my hope that my work has contributed to elucidating the mechanism of the cdAE1-protein 4.2
interaction. As well, I hope I have contributed to the understanding of the mechanism of
defective protein interactions in disease in general, and in HS in particular.
111
References
Allen, D.P., Low, P.S., Dola, A. & Maisel, H. (1987) Band 3 and ankyrin homologues are
present in eye lens: evidence for all major erythrocyte membrane components in same
non-erythroid cell. Biochem Biophys Res Commun, 149, 266-275.
Allen, S.J., O'Donnell, A., Alexander, N.D., Mgone, C.S., Peto, T.E., Clegg, J.B., Alpers, M.P.
& Weatherall, D.J. (1999) Prevention of cerebral malaria in children in Papua New
Guinea by southeast Asian ovalocytosis band 3. Am J Trop Med Hyg, 60, 1056-1060.
Alloisio, N., Maillet, P., Carre, G., Texier, P., Vallier, A., Baklouti, F., Philippe, N. & Delaunay,
J. (1996) Hereditary spherocytosis with band 3 deficiency. Association with a nonsense
mutation of the band 3 gene (allele Lyon), and aggravation by a low-expression allele
occurring in trans (allele Genas). Blood, 88, 1062-1069.
Alloisio, N., Texier, P., Vallier, A., Ribeiro, M.L., Morle, L., Bozon, M., Bursaux, E., Maillet,
P., Goncalves, P., Tanner, M.J., Tamagnini, G. & Delaunay, J. (1997) Modulation of
clinical expression and band 3 deficiency in hereditary spherocytosis. Blood, 90, 414-
420.
Alper, S.L. (2009) Molecular physiology and genetics of Na+-independent SLC4 anion
exchangers. J Exp Biol, 212, 1672-1683.
Amato, D. & Booth, P.B. (1977) Hereditary ovalocytosis in Melanesians. P N G Med J, 20, 26-
32.
Appell, K.C. & Low, P.S. (1981) Partial structural characterization of the cytoplasmic domain of
the erythrocyte membrane protein, band 3. J Biol Chem, 256, 11104-11111.
Arnold, K., Bordoli, L., Kopp, J. & Schwede, T. (2006) The SWISS-MODEL workspace: a web-
based environment for protein structure homology modelling. Bioinformatics, 22, 195-
201.
Azim, A.C., Marfatia, S.M., Korsgren, C., Dotimas, E., Cohen, C.M. & Chishti, A.H. (1996)
Human erythrocyte dematin and protein 4.2 (pallidin) are ATP binding proteins.
Biochemistry, 35, 3001-3006.
Batlle, D., Ghanekar, H., Jain, S. & Mitra, A. (2001) Hereditary distal renal tubular acidosis: new
understandings. Annu Rev Med, 52, 471-484.
Beauchamp-Nicoud, A., Morle, L., Lutz, H.U., Stammler, P., Agulles, O., Petermann-Khder, R.,
Iolascon, A., Perrotta, S., Cynober, T., Tchernia, G., Delaunay, J. & Baudin-Creuza, V.
(2000) Heavy transfusions and presence of an anti-protein 4.2 antibody in 4.2(-)
hereditary spherocytosis (949delG). Haematologica, 85, 19-24.
Bennett, V. (1979) Immunoreactive forms of human erythrocyte ankyrin are present in diverse
cells and tissues. Nature, 281, 597-599.
112
Bennett, V. (1985) The membrane skeleton of human erythrocytes and its implications for more
complex cells. Annu Rev Biochem, 54, 273-304.
Bennett, V. & Stenbuck, P.J. (1980) Association between ankyrin and the cytoplasmic domain of
band 3 isolated from the human erythrocyte membrane. J Biol Chem, 255, 6424-6432.
Bergamini, C.M. (1988) GTP modulates calcium binding and cation-induced conformational
changes in erythrocyte transglutaminase. FEBS Lett, 239, 255-258.
Bhattacharyya, R., Das, A.K., Moitra, P.K., Pal, B., Mandal, I. & Basu, J. (1999) Mapping of a
palmitoylatable band 3-binding domain of human erythrocyte membrane protein 4.2.
Biochem J, 340 ( Pt 2), 505-512.
Bianchi, P., Zanella, A., Alloisio, N., Barosi, G., Bredi, E., Pelissero, G., Zappa, M., Vercellati,
C., Baronciani, L., Delaunay, J. & Sirchia, G. (1997) A variant of the EPB3 gene of the
anti-Lepore type in hereditary spherocytosis. Br J Haematol, 98, 283-288.
Bijlmakers, M.J. & Marsh, M. (2003) The on-off story of protein palmitoylation. Trends Cell
Biol, 13, 32-42.
Bouhassira, E.E., Schwartz, R.S., Yawata, Y., Ata, K., Kanzaki, A., Qiu, J.J., Nagel, R.L. &
Rybicki, A.C. (1992) An alanine-to-threonine substitution in protein 4.2 cDNA is
associated with a Japanese form of hereditary hemolytic anemia (protein 4.2NIPPON).
Blood, 79, 1846-1854.
Bracher, N.A., Lyons, C.A., Wessels, G., Mansvelt, E. & Coetzer, T.L. (2001) Band 3 Cape
Town (E90K) causes severe hereditary spherocytosis in combination with band 3 Prague
III. Br J Haematol, 113, 689-693.
Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal Biochem, 72,
248-254.
Branton, D., Cohen, C.M. & Tyler, J. (1981) Interaction of cytoskeletal proteins on the human
erythrocyte membrane. Cell, 24, 24-32.
Brown, E. (2001) Integrin-associated protein (CD47): an unusual activator of G protein
signaling. J Clin Invest, 107, 1499-1500.
Bruce, L.J. (2008) Red cell membrane transport abnormalities. Curr Opin Hematol, 15, 184-190.
Bruce, L.J., Beckmann, R., Ribeiro, M.L., Peters, L.L., Chasis, J.A., Delaunay, J., Mohandas, N.,
Anstee, D.J. & Tanner, M.J. (2003) A band 3-based macrocomplex of integral and
peripheral proteins in the RBC membrane. Blood, 101, 4180-4188.
Bruce, L.J., Ghosh, S., King, M.J., Layton, D.M., Mawby, W.J., Stewart, G.W., Oldenborg, P.A.,
Delaunay, J. & Tanner, M.J. (2002) Absence of CD47 in protein 4.2-deficient hereditary
spherocytosis in man: an interaction between the Rh complex and the band 3 complex.
Blood, 100, 1878-1885.
113
Bruce, L.J., Ring, S.M., Anstee, D.J., Reid, M.E., Wilkinson, S. & Tanner, M.J. (1995) Changes
in the blood group Wright antigens are associated with a mutation at amino acid 658 in
human erythrocyte band 3: a site of interaction between band 3 and glycophorin A under
certain conditions. Blood, 85, 541-547.
Bustos, S.P. & Reithmeier, R.A. (2010) Protein 4.2 interaction with hereditary spherocytosis
mutants of the cytoplasmic domain of human anion exchanger 1. Biochem J, 433, 313-
322.
Bustos, S.P. & Reithmeier, R.A.F. (2006) Structure and stability of hereditary spherocytosis
mutants of the cytosolic domain of the erythrocyte anion exchanger 1 protein.
Biochemistry, 45, 1026-1034.
Campanella, M.E., Chu, H. & Low, P.S. (2005) Assembly and regulation of a glycolytic enzyme
complex on the human erythrocyte membrane. Proc Natl Acad Sci U S A, 102, 2402-
2407.
Casey, J.R. & Reithmeier, R.A. (1991) Analysis of the oligomeric state of Band 3, the anion
transport protein of the human erythrocyte membrane, by size exclusion high
performance liquid chromatography. Oligomeric stability and origin of heterogeneity. J
Biol Chem, 266, 15726-15737.
Chang, S.H. & Low, P.S. (2003) Identification of a critical ankyrin-binding loop on the
cytoplasmic domain of erythrocyte membrane band 3 by crystal structure analysis and
site-directed mutagenesis. J Biol Chem, 278, 6879-6884.
Chen, J.S. & Mehta, K. (1999) Tissue transglutaminase: an enzyme with a split personality. Int J
Biochem Cell Biol, 31, 817-836.
Cheung, J.C., Cordat, E. & Reithmeier, R.A. (2005a) Trafficking defects of the Southeast Asian
ovalocytosis deletion mutant of anion exchanger 1 membrane proteins. Biochem J, 392,
425-434.
Cheung, J.C., Li, J. & Reithmeier, R.A. (2005b) Topology of transmembrane segments 1-4 in the
human chloride/bicarbonate anion exchanger 1 (AE1) by scanning N-glycosylation
mutagenesis. Biochem J, 390, 137-144.
Cheung, J.C. & Reithmeier, R.A. (2005) Membrane integration and topology of the first
transmembrane segment in normal and Southeast Asian ovalocytosis human erythrocyte
anion exchanger 1. Mol Membr Biol, 22, 203-214.
Chu, C., Woods, N., Sawasdee, N., Guizouarn, H., Pellissier, B., Borgese, F., Yenchitsomanus,
P.T., Gowrishankar, M. & Cordat, E. (2010) Band 3 Edmonton I, a novel mutant of the
anion exchanger 1 causing spherocytosis and distal renal tubular acidosis. Biochem J,
426, 379-388.
Cobb, C.E. & Beth, A.H. (1990) Identification of the eosinyl-5-maleimide reaction site on the
human erythrocyte anion-exchange protein: overlap with the reaction sites of other
chemical probes. Biochemistry, 29, 8283-8290.
114
Colfen, H., Harding, S.E., Boulter, J.M. & Watts, A. (1996) Hydrodynamic examination of the
dimeric cytoplasmic domain of the human erythrocyte anion transporter, band 3. Biophys
J, 71, 1611-1615.
Coppolino, M., Leung-Hagesteijn, C., Dedhar, S. & Wilkins, J. (1995) Inducible interaction of
integrin α2β1 with calreticulin. Dependence on the activation state of the integrin. J Biol
Chem, 270, 23132-23138.
Corson, T.W., Aberle, N. & Crews, C.M. (2008) Design and applications of bifunctional small
molecules: Why two heads are better than one. ACS Chem Biol, 3, 677-692.
Cox, J.V. & Lazarides, E. (1988) Alternative primary structures in the transmembrane domain of
the chicken erythroid anion transporter. Mol Cell Biol, 8, 1327-1335.
Das, A.K., Bhattacharya, R., Kundu, M., Chakrabarti, P. & Basu, J. (1994) Human erythrocyte
membrane protein 4.2 is palmitoylated. Eur J Biochem, 224, 575-580.
Delaunay, J. (2007) The molecular basis of hereditary red cell membrane disorders. Blood Rev,
21, 1-20.
Dhermy, D., Galand, C., Bournier, O., Boulanger, L., Cynober, T., Schismanoff, P.O., Bursaux,
E., Tchernia, G., Boivin, P. & Garbarz, M. (1997) Heterogenous band 3 deficiency in
hereditary spherocytosis related to different band 3 gene defects. Br J Haematol, 98, 32-
40.
Ding, Y., Casey, J.R. & Kopito, R.R. (1994) The major kidney AE1 isoform does not bind
ankyrin (ANK1) in vitro. An essential role for the 79 NH2-terminal amino acid residues
of band 3. J Biol Chem, 269, 32201-32208.
Dotimas, E., Speicher, D.W., GuptaRoy, B. & Cohen, C.M. (1993) Structural domain mapping
and phosphorylation of human erythrocyte pallidin (band 4.2). Biochim Biophys Acta,
1148, 19-29.
Drechsler, Z., Sharkia, R., Cabantchik, Z.I. & Beer, S. (1993) Bicarbonate uptake in the marine
macroalga Ulva sp. is inhibited by classical probes of anion exchange by red blood cells.
Planta, 191, 34-40.
Drickamer, L.K. (1978) Orientation of the band 3 polypeptide from human erythrocyte
membranes. Identification of NH2-terminal sequence and site of carbohydrate attachment.
J Biol Chem, 253, 7242-7248.
Eber, S. & Lux, S.E. (2004) Hereditary spherocytosis-defects in proteins that connect the
membrane skeleton to the lipid bilayer. Semin Hematol, 41, 118-141.
Eber, S.W., Gonzalez, J.M., Lux, M.L., Scarpa, A.L., Tse, W.T., Dornwell, M., Herbers, J.,
Kugler, W., Ozcan, R., Pekrun, A., Gallagher, P.G., Schroter, W., Forget, B.G. & Lux,
S.E. (1996) Ankyrin-1 mutations are a major cause of dominant and recessive hereditary
spherocytosis. Nat Genet, 13, 214-218.
115
Feng, J.F., Readon, M., Yadav, S.P. & Im, M.J. (1999) Calreticulin down-regulates both GTP
binding and transglutaminase activities of transglutaminase II. Biochemistry, 38, 10743-
10749.
Friedrichs, B., Koob, R., Kraemer, D. & Drenckhahn, D. (1989) Demonstration of
immunoreactive forms of erythrocyte protein 4.2 in nonerythroid cells and tissues. Eur J
Cell Biol, 48, 121-127.
Fujinaga, J., Tang, X.-B. & Casey, J.R. (1999) Topology of the membrane domain of human
erythrocyte anion exchange protein, AE1. J Biol Chem, 274, 6626-6633.
Furuya, W., Tarshis, T., Law, F.Y. & Knauf, P.A. (1984) Transmembrane effects of intracellular
chloride on the inhibitory potency of extracellular H2DIDS. Evidence for two
conformations of the transport site of the human erythrocyte anion exchange protein. J
Gen Physiol, 83, 657-681.
Ghosh, S., Cox, K.H. & Cox, J.V. (1999) Chicken erythroid AE1 anion exchangers associate
with the cytoskeleton during recycling to the Golgi. Mol Biol Cell, 10, 455-469.
Giordano, T.J., Deuschle, U., Bujard, H. & McAllister, W.T. (1989) Regulation of coliphage T3
and T7 RNA polymerases by the lac repressor-operator system. Gene, 84, 209-219.
Golan, D.E., Corbett, J.D., Korsgren, C., Thatte, H.S., Hayette, S., Yawata, Y. & Cohen, C.M.
(1996) Control of band 3 lateral and rotational mobility by band 4.2 in intact
erythrocytes: release of band 3 oligomers from low-affinity binding sites. Biophys J, 70,
1534-1542.
Gomez, S. & Morgans, C. (1993) Interaction between band 3 and ankyrin begins in early
compartments of the secretory pathway and is essential for band 3 processing. J Biol
Chem, 268, 19593-19597.
Gregersen, N., Bross, P., Jorgensen, M.M., Corydon, T.J. & Andresen, B.S. (2000) Defective
folding and rapid degradation of mutant proteins is a common disease mechanism in
genetic disorders. J Inherit Metab Dis, 23, 441-447.
Grenard, P., Bates, M.K. & Aeschlimann, D. (2001) Evolution of transglutaminase genes:
identification of a transglutaminase gene cluster on human chromosome 15q15. Structure
of the gene encoding transglutaminase X and a novel gene family member,
transglutaminase Z. J Biol Chem, 276, 33066-33078.
Grinstein, S., Ship, S. & Rothstein, A. (1978) Anion transport in relation to proteolytic dissection
of band 3 protein. Biochim Biophys Acta, 507, 294-304.
Groves, J.D., Ring, S.M., Schofield, A.E. & Tanner, M.J. (1993) The expression of the abnormal
human red cell anion transporter from South-East Asian ovalocytes (band 3 SAO) in
Xenopus oocytes. FEBS Lett, 330, 186-190.
Groves, J.D. & Tanner, M.J. (1992) Glycophorin A facilitates the expression of human band 3-
mediated anion transport in Xenopus oocytes. J Biol Chem, 267, 22163-22170.
116
Groves, J.D. & Tanner, M.J. (1999) Topology studies with biosynthetic fragments identify
interacting transmembrane regions of the human red-cell anion exchanger (band 3; AE1).
Biochem J, 344 Pt 3, 687-697.
Guerra-Shinohara, E.M. & Barretto, O.C. (1999) The erythrocyte cytoskeleton protein 4.2 is not
demonstrable in several mammalian species. Braz J Med Biol Res, 32, 683-687.
Gutierrez, E. & Sung, L.A. (2007) Interactions of recombinant mouse erythrocyte
transglutaminase with membrane skeletal proteins. J Membr Biol, 219, 93-104.
Gwynn, B., Korsgren, C., Cohen, C.M., Ciciotte, S.L. & Peters, L.L. (1997) The gene encoding
protein 4.2 is distinct from the mouse platelet storage pool deficiency mutation pallid.
Genomics, 42, 532-535.
Hamasaki, N., Okubo, K., Kuma, H., Kang, D. & Yae, Y. (1997) Proteolytic cleavage sites of
band 3 protein in alkali-treated membranes: fidelity of hydropathy prediction for band 3
protein. J Biochem, 122, 577-585.
Hanspal, M. & Palek, J. (1987) Synthesis and assembly of membrane skeletal proteins in
mammalian red cell precursors. J Cell Biol, 105, 1417-1424.
Hargreaves, W.R., Giedd, K.N. & Branton, D. (1980) Reassociation of ankyrin with band 3 in
erythrocyte membranes and in lipid vesicles. J Biol Chem, 255, 11965-11972.
Harrison, M.L., Rathinavelu, P., Arese, P., Geahlen, R.L. & Low, P.S. (1991) Role of band 3
tyrosine phosphorylation in the regulation of erythrocyte glycolysis. J Biol Chem, 266,
4106-4111.
Hayette, S., Morle, L., Bozon, M., Ghanem, A., Risinger, M., Korsgren, C., Tanner, M.J.,
Fattoum, S., Cohen, C.M. & Delaunay, J. (1995) A point mutation in the protein 4.2 gene
(allele 4.2 Tozeur) associated with hereditary haemolytic anaemia. Br J Haematol, 89,
762-770.
Hubner, S., Michel, F., Rudloff, V. & Appelhans, H. (1992) Amino acid sequence of band-3
protein from rainbow trout erythrocytes derived from cDNA. Biochem J, 285 ( Pt 1), 17-
23.
Idakieva, K., Parvanova, K. & Todinova, S. (2005) Differential scanning calorimetry of the
irreversible denaturation of Rapana thomasiana (marine snail, Gastropod) hemocyanin.
Biochim Biophys Acta, 1748, 50-56.
Iismaa, S.E., Chung, L., Wu, M.J., Teller, D.C., Yee, V.C. & Graham, R.M. (1997) The core
domain of the tissue transglutaminase Gh hydrolyzes GTP and ATP. Biochemistry, 36,
11655-11664.
117
Inaba, M., Yawata, A., Koshino, I., Sato, K., Takeuchi, M., Takakuwa, Y., Manno, S., Yawata,
Y., Kanzaki, A., Sakai, J., Ban, A., Ono, K. & Maede, Y. (1996) Defective anion
transport and marked spherocytosis with membrane instability caused by hereditary total
deficiency of red cell band 3 in cattle due to a nonsense mutation. J Clin Invest, 97, 1804-
1817.
Inoue, T., Kanzaki, A., Kaku, M., Yawata, A., Takezono, M., Okamoto, N., Wada, H., Sugihara,
T., Yamada, O., Katayama, Y., Nagata, N. & Yawata, Y. (1998) Homozygous missense
mutation (band 3 Fukuoka: G130R): a mild form of hereditary spherocytosis with near-
normal band 3 content and minimal changes of membrane ultrastructure despite moderate
protein 4.2 deficiency. Br J Haematol, 102, 932-939.
Iolascon, A., Miraglia del Giudice, E., Perrotta, S., Alloisio, N., Morle, L. & Delaunay, J. (1998)
Hereditary spherocytosis: from clinical to molecular defects. Haematologica, 83, 240-
257.
Iwase, S., Ideguchi, H., Takao, M., Horiguchi-Yamada, J., Iwasaki, M., Takahara, S., Sekikawa,
T., Mochizuki, S. & Yamada, H. (1998) Band 3 Tokyo: Thr837→Ala837 substitution in
erythrocyte band 3 protein associated with spherocytic hemolysis. Acta Haematol, 100,
200-203.
Jarolim, P., Murray, J.L., Rubin, H.L., Taylor, W.M., Prchal, J.T., Ballas, S.K., Snyder, L.M.,
Chrobak, L., Melrose, W.D., Brabec, V. & Palek, J. (1996) Characterization of 13 novel
band 3 gene defects in hereditary spherocytosis with band 3 deficiency. Blood, 88, 4366-
4374.
Jarolim, P., Palek, J., Amato, D., Hassan, K., Sapak, P., Nurse, G.T., Rubin, H.L., Zhai, S., Sahr,
K.E. & Liu, S.C. (1991) Deletion in erythrocyte band 3 gene in malaria-resistant
Southeast Asian ovalocytosis. Proc Natl Acad Sci U S A, 88, 11022-11026.
Jarolim, P., Palek, J., Rubin, H.L., Prchal, J.T., Korsgren, C. & Cohen, C.M. (1992a) Band 3
Tuscaloosa: Pro327→Arg327 substitution in the cytoplasmic domain of erythrocyte band
3 protein associated with spherocytic hemolytic anemia and partial deficiency of protein
4.2. Blood, 80, 523-529.
Jarolim, P., Rubin, H.L., Brabec, V., Chrobak, L., Zolotarev, A.S., Alper, S.L., Brugnara, C.,
Wichterle, H. & Palek, J. (1995) Mutations of conserved arginines in the membrane
domain of erythroid band 3 lead to a decrease in membrane-associated band 3 and to the
phenotype of hereditary spherocytosis. Blood, 85, 634-640.
Jarolim, P., Rubin, H.L., Liu, S.C., Cho, M.R., Brabec, V., Derick, L.H., Yi, S.J., Saad, S.T.,
Alper, S., Brugnara, C. et al. (1994) Duplication of 10 nucleotides in the erythroid band 3
(AE1) gene in a kindred with hereditary spherocytosis and band 3 protein deficiency
(band 3PRAGUE). J Clin Invest, 93, 121-130.
Jarolim, P., Rubin, H.L., Zhai, S., Sahr, K.E., Liu, S.C., Mueller, T.J. & Palek, J. (1992b) Band 3
Memphis: a widespread polymorphism with abnormal electrophoretic mobility of
erythrocyte band 3 protein caused by substitution AAG→GAG (Lys→Glu) in codon 56.
Blood, 80, 1592-1598.
118
Jay, D.G. (1986) Glycosylation site of band 3, the human erythrocyte anion-exchange protein.
Biochemistry, 25, 554-556.
Jenkins, P.B., Abou-Alfa, G.K., Dhermy, D., Bursaux, E., Feo, C., Scarpa, A.L., Lux, S.E.,
Garbarz, M., Forget, B.G. & Gallagher, P.G. (1996) A nonsense mutation in the
erythrocyte band 3 gene associated with decreased mRNA accumulation in a kindred
with dominant hereditary spherocytosis. J Clin Invest, 97, 373-380.
Jennings, M.L. (1989a) Structure and function of the red blood cell anion exchange protein. Prog
Clin Biol Res., 292, 327-338.
Jennings, M.L. (1989b) Structure and function of the red blood cell anion transport protein. Annu
Rev Biophys Biophys Chem, 18, 397-430.
Jennings, M.L. (1995) Rapid electrogenic sulfate-chloride exchange mediated by chemically
modified band 3 in human erythrocytes. J Gen Physiol, 105, 21-47.
Jennings, M.L., Anderson, M.P. & Monaghan, R. (1986) Monoclonal antibodies against human
erythrocyte band 3 protein. Localization of proteolytic cleavage sites and
stilbenedisulfonate-binding lysine residues. J Biol Chem, 261, 9002-9010.
Jennings, M.L., Howren, T.R., Cui, J., Winters, M. & Hannigan, R. (2007) Transport and
regulatory characteristics of the yeast bicarbonate transporter homolog Bor1p. Am J
Physiol Cell Physiol, 293, C468-476.
Jennings, M.L. & Nicknish, J.S. (1985) Localization of a site of intermolecular cross-linking in
human red blood cell band 3 protein. J Biol Chem, 260, 5472-5479.
Jennings, M.L. & Smith, J.S. (1992) Anion-proton cotransport through the human red blood cell
band 3 protein. Role of glutamate 681. J Biol Chem, 267, 13964-13971.
Johnson, W.C., Jr. (1990) Protein secondary structure and circular dichroism: a practical guide.
Proteins, 7, 205-214.
Kang, D., Okubo, K., Hamasaki, N., Kuroda, N. & Shiraki, H. (1992) A structural study of the
membrane domain of band 3 by tryptic digestion. Conformational change of band 3 in
situ induced by alkali treatment. J Biol Chem, 267, 19211-19217.
Kang, S.K., Lee, J.Y., Chung, T.W. & Kim, C.H. (2004) Overexpression of transglutaminase 2
accelerates the erythroid differentiation of human chronic myelogenous leukemia K562
cell line through PI3K/Akt signaling pathway. FEBS Lett, 577, 361-366.
Kanki, T., Sakaguchi, M., Kitamura, A., Sato, T., Mihara, K. & Hamasaki, N. (2002) The tenth
membrane region of band 3 is initially exposed to the luminal side of the endoplasmic
reticulum and then integrated into a partially folded band 3 intermediate. Biochemistry,
41, 13973-13981.
Kanzaki, A., Hayette, S., Morle, L., Inoue, F., Matsuyama, R., Inoue, T., Yawata, A., Wada, H.,
Vallier, A., Alloisio, N., Yawata, Y. & Delaunay, J. (1997) Total absence of protein 4.2
and partial deficiency of band 3 in hereditary spherocytosis. Br J Haematol, 99, 522-530.
119
Kanzaki, A., Yasunaga, M., Okamoto, N., Inoue, T., Yawata, A., Wada, H., Andoh, A.,
Hodohara, K., Fujiyama, Y., Bamba, T. et al. (1995a) Band 4.2 Shiga: 317 CGC→TGC
in compound heterozygotes with 142 GCT→ACT results in band 4.2 deficiency and
microspherocytosis. Br J Haematol, 91, 333-340.
Kanzaki, A., Yawata, Y., Yawata, A., Inoue, T., Okamoto, N., Wada, H., Harano, T., Harano, K.,
Wilmotte, R., Hayette, S. et al. (1995b) Band 4.2 Komatsu: 523 GAT→TAT (175
Asp→Tyr) in exon 4 of the band 4.2 gene associated with total deficiency of band 4.2,
hemolytic anemia with ovalostomatocytosis and marked disruption of the cytoskeletal
network. Int J Hematol, 61, 165-178.
Karet, F.E. (2002) Inherited distal renal tubular acidosis. J Am Soc Nephrol, 13, 2178-2184.
Kaul, R.K., Murthy, S.N., Reddy, A.G., Steck, T.L. & Kohler, H. (1983) Amino acid sequence of
the N α-terminal 201 residues of human erythrocyte membrane band 3. J Biol Chem, 258,
7981-7990.
Keskanokwong, T., Shandro, H.J., Johnson, D.E., Kittanakom, S., Vilas, G.L., Thorner, P.,
Reithmeier, R.A.F., Akkarapatumwong, V., Yenchitsomanus, P. & Casey, J.R. (2007)
Interaction of integrin-linked kinase with the kidney chloride/bicarbonate exchanger,
kAE1. J Biol Chem, 282, 23205-23218.
Kidson, C., Lamont, G., Saul, A. & Nurse, G.T. (1981) Ovalocytic erythrocytes from
Melanesians are resistant to invasion by malaria parasites in culture. Proc Natl Acad Sci
U S A, 78, 5829-5832.
Kittanakom, S., Keskanokwong, T., Akkarapatumwong, V., Yenchitsomanus, P.T. &
Reithmeier, R.A. (2004) Human kanadaptin and kidney anion exchanger 1 (kAE1) do not
interact in transfected HEK-293 cells. Mol Membr Biol, 21, 395-402.
Kollert-Jons, A., Wagner, S., Hubner, S., Appelhans, H. & Drenckhahn, D. (1993) Anion
exchanger 1 in human kidney and oncocytoma differs from erythroid AE1 in its NH2
terminus. Am J Physiol, 265, F813-821.
Kopito, R.R. & Lodish, H.F. (1985) Primary structure and transmembrane orientation of the
murine anion exchange protein. Nature, 316, 234-238.
Korsgren, C. & Cohen, C.M. (1986) Purification and properties of human erythrocyte band 4.2.
Association with the cytoplasmic domain of band 3. J Biol Chem, 261, 5536-5543.
Korsgren, C. & Cohen, C.M. (1988) Associations of human erythrocyte band 4.2. Binding to
ankyrin and to the cytoplasmic domain of band 3. J Biol Chem, 263, 10212-10218.
Korsgren, C. & Cohen, C.M. (1991) Organization of the gene for human erythrocyte membrane
protein 4.2: structural similarities with the gene for the a subunit of factor XIII. Proc Natl
Acad Sci U S A, 88, 4840-4844.
120
Korsgren, C., Lawler, J., Lambert, S., Speicher, D. & Cohen, C.M. (1990) Complete amino acid
sequence and homologies of human erythrocyte membrane protein band 4.2. Proc Natl
Acad Sci U S A, 87, 613-617.
Korsgren, C., Peters, L.L. & Lux, S.E. (2010) Protein 4.2 binds to the carboxyl-terminal EF-
hands of erythroid α-spectrin in a calcium- and calmodulin-dependent manner. J Biol
Chem, 285, 4757-4770.
Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of
bacteriophage T4. Nature, 227, 680-685.
Laosombat, V., Viprakasit, V., Dissaneevate, S., Leetanaporn, R., Chotsampancharoen, T.,
Wongchanchailert, M., Kodchawan, S., Thongnoppakun, W. & Duangchu, S. (2010)
Natural history of Southeast Asian Ovalocytosis during the first 3 years of life. Blood
Cells Mol Dis, 45, 29-32.
Lataillade, J.J., Clay, D., David, C., Boutin, L., Guerton, B., Drouet, M., Herodin, F. & Le
Bousse-Kerdiles, M.C. (2005) Phenotypic and functional characteristics of CD34+ cells
are related to their anatomical environment: is their versatility a prerequisite for their bio-
availability? J Leukoc Biol, 77, 634-643.
Lee, J.C., Gimm, J.A., Lo, A.J., Koury, M.J., Krauss, S.W., Mohandas, N. & Chasis, J.A. (2004)
Mechanism of protein sorting during erythroblast enucleation: role of cytoskeletal
connectivity. Blood, 103, 1912-1919.
Lee, K.N., Arnold, S.A., Birckbichler, P.J., Patterson, M.K., Jr., Fraij, B.M., Takeuchi, Y. &
Carter, H.A. (1993) Site-directed mutagenesis of human tissue transglutaminase: Cys-277
is essential for transglutaminase activity but not for GTPase activity. Biochim Biophys
Acta, 1202, 1-6.
Lee, K.N., Birckbichler, P.J. & Fesus, L. (1986) Purification of human erythrocyte
transglutaminase by immunoaffinity chromatography. Prep Biochem, 16, 321-335.
Lemieux, M.J., Reithmeier, R.A. & Wang, D.N. (2002) Importance of detergent and
phospholipid in the crystallization of the human erythrocyte anion-exchanger membrane
domain. J Struct Biol, 137, 322-332.
Lepke, S., Fasold, H., Pring, M. & Passow, H. (1976) A study of the relationship between
inhibition of anion exchange and binding to the red blood cell membrane of 4,4'-
diisothiocyano stilbene-2,2'-disulfonic acid (DIDS) and its dihydro derivative (H2DIDS).
J Membr Biol, 29, 147-177.
Lepke, S. & Passow, H. (1976) Effects of incorporated trypsin on anion exchange and membrane
proteins in human red blood cell ghosts. Biochim Biophys Acta, 455, 353-370.
Lima, P.R., Baratti, M.O., Chiattone, M.L., Costa, F.F. & Saad, S.T. (2005) Band 3Tambau: a de
novo mutation in the AE1 gene associated with hereditary spherocytosis. Implications for
anion exchange and insertion into the red blood cell membrane. Eur J Haematol, 74, 396-
401.
121
Lima, P.R., Gontijo, J.A., Lopes de Faria, J.B., Costa, F.F. & Saad, S.T. (1997) Band 3
Campinas: a novel splicing mutation in the band 3 gene (AE1) associated with hereditary
spherocytosis, hyperactivity of Na+/Li
+ countertransport and an abnormal renal
bicarbonate handling. Blood, 90, 2810-2818.
Lima, P.R., Sales, T.S., Costa, F.F. & Saad, S.T. (1999) Arginine 490 is a hot spot for mutation
in the band 3 gene in hereditary spherocytosis. Eur J Haematol, 63, 360-361.
Lindberg, F.P., Lublin, D.M., Telen, M.J., Veile, R.A., Miller, Y.E., Donis-Keller, H. & Brown,
E.J. (1994) Rh-related antigen CD47 is the signal-transducer integrin-associated protein.
J Biol Chem, 269, 1567-1570.
Liu, S., Cerione, R.A. & Clardy, J. (2002) Structural basis for the guanine nucleotide-binding
activity of tissue transglutaminase and its regulation of transamidation activity. Proc Natl
Acad Sci U S A, 99, 2743-2747.
Liu, S.C., Jarolim, P., Rubin, H.L., Palek, J., Amato, D., Hassan, K., Zaik, M. & Sapak, P.
(1994) The homozygous state for the band 3 protein mutation in Southeast Asian
Ovalocytosis may be lethal. Blood, 84, 3590-3591.
Liu, S.C., Palek, J., Yi, S.J., Nichols, P.E., Derick, L.H., Chiou, S.S., Amato, D., Corbett, J.D.,
Cho, M.R. & Golan, D.E. (1995) Molecular basis of altered red blood cell membrane
properties in Southeast Asian ovalocytosis: role of the mutant band 3 protein in band 3
oligomerization and retention by the membrane skeleton. Blood, 86, 349-358.
Liu, S.C., Zhai, S., Palek, J., Golan, D.E., Amato, D., Hassan, K., Nurse, G.T., Babona, D.,
Coetzer, T., Jarolim, P. et al. (1990) Molecular defect of the band 3 protein in southeast
Asian ovalocytosis. N Engl J Med, 323, 1530-1538.
Lorand, L. (2007) Crosslinks in blood: transglutaminase and beyond. FASEB J, 21, 1627-1632.
Lorand, L. & Graham, R.M. (2003) Transglutaminases: crosslinking enzymes with pleiotropic
functions. Nat Rev Mol Cell Biol, 4, 140-156.
Lorand, L., Michalska, M., Murthy, S.N., Shohet, S.B. & Wilson, J. (1987) Cross-linked
polymers in the red cell membranes of a patient with Hb-Köln disease. Biochem Biophys
Res Commun, 147, 602-607.
Lorand, L., Siefring, G.E., Jr. & Lowe-Krentz, L. (1980) Ca2+
-triggered and enzyme-mediated
cross-linking of membrane proteins in intact human erythrocytes. In: Membrane
Transport in Erythrocytes (ed. by U.V. Lassen, H.H. Ussing & J.O. Wieth), Vol. 14, pp.
285-299. Munksgaard, Copenhagen.
Low, P.S. (1986) Structure and function of the cytoplasmic domain of band 3: center of
erythrocyte membrane-peripheral protein interactions. Biochim Biophys Acta, 864, 145-
167.
122
Lux, S.E., John, K.M. & Bennett, V. (1990) Analysis of cDNA for human erythrocyte ankyrin
indicates a repeated structure with homology to tissue-differentiation and cell-cycle
control proteins. Nature, 344, 36-42.
Lux, S.E., John, K.M., Kopito, R.R. & Lodish, H.F. (1989) Cloning and characterization of band
3, the human erythrocyte anion-exchange protein (AE1). Proc Natl Acad Sci U S A, 86,
9089-9093.
Lux, S.E. & Palek, J. (1995) Disorders of the red cell membrane. In: Blood: Principles and
Practice in Hematology (ed. by R.I. Handin, S.E. Lux & T.P. Stossel), p. 1701.
Lippincott, Philadephia.
Maillet, P., Vallier, A., Reinhart, W.H., Wyss, E.J., Ott, P., Texier, P., Baklouti, F., Tanner, M.J.,
Delaunay, J. & Alloisio, N. (1995) Band 3 Chur: a variant associated with band 3-
deficient hereditary spherocytosis and substitution in a highly conserved position of
transmembrane segment 11. Br J Haematol, 91, 804-810.
Mandal, D., Baudin-Creuza, V., Bhattacharyya, A., Pathak, S., Delaunay, J., Kundu, M. & Basu,
J. (2003) Caspase 3-mediated proteolysis of the N-terminal cytoplasmic domain of the
human erythroid anion exchanger 1 (band 3). J Biol Chem, 278, 52551-52558.
Mandal, D., Moitra, P.K. & Basu, J. (2002) Mapping of a spectrin-binding domain of human
erythrocyte membrane protein 4.2. Biochem J, 364, 841-847.
Marini, A.M., Matassi, G., Raynal, V., Andre, B., Cartron, J.P. & Cherif-Zahar, B. (2000) The
human Rhesus-associated RhAG protein and a kidney homologue promote ammonium
transport in yeast. Nat Genet, 26, 341-344.
Matsuda, M., Hatano, N., Ideguchi, H., Takahira, H. & Fukumaki, Y. (1995) A novel mutation
causing an aberrant splicing in the protein 4.2 gene associated with hereditary
spherocytosis (protein 4.2Notame). Hum Mol Genet, 4, 1187-1191.
Mhaouty-Kodja, S. (2004) Ghα/tissue transglutaminase 2: an emerging G protein in signal
transduction. Biol Cell, 96, 363-367.
Milanick, M.A. & Gunn, R.B. (1984) Proton-sulfate cotransport: external proton activation of
sulfate influx into human red blood cells. Am J Physiol, 247, C247-259.
Miraglia del Giudice, E., Vallier, A., Maillet, P., Perrotta, S., Cutillo, S., Iolascon, A., Tanner,
M.J., Delaunay, J. & Alloisio, N. (1997) Novel band 3 variants (bands 3 Foggia, Napoli I
and Napoli II) associated with hereditary spherocytosis and band 3 deficiency: status of
the D38A polymorphism within the EPB3 locus. Br J Haematol, 96, 70-76.
Mohandas, N., Winardi, R., Knowles, D., Leung, A., Parra, M., George, E., Conboy, J. & Chasis,
J. (1992) Molecular basis for membrane rigidity of hereditary ovalocytosis. A novel
mechanism involving the cytoplasmic domain of band 3. J Clin Invest, 89, 686-692.
Moon, R.T. & McMahon, A.P. (1987) Composition and expression of spectrin-based membrane
skeletons in non-erythroid cells. Bioessays, 7, 159-164.
123
Moriyama, R., Ideguchi, H., Lombardo, C.R., Van Dort, H.M. & Low, P.S. (1992) Structural and
functional characterization of band 3 from Southeast Asian ovalocytes. J Biol Chem, 267,
25792-25797.
Motais, R., Fievet, B., Borgese, F. & Garcia-Romeu, F. (1997) Association of the band 3 protein
with a volume-activated, anion and amino acid channel: a molecular approach. J Exp
Biol, 200, 361-367.
Mouro-Chanteloup, I., Delaunay, J., Gane, P., Nicolas, V., Johansen, M., Brown, E.J., Peters,
L.L., Van Kim, C.L., Cartron, J.P. & Colin, Y. (2003) Evidence that the red cell skeleton
protein 4.2 interacts with the Rh membrane complex member CD47. Blood, 101, 338-
344.
Mueller, T.J. & Morrison, M. (1977) Detection of a variant of protein 3, the major
transmembrane protein of the human erythrocyte. J Biol Chem, 252, 6573-6576.
Muller-Berger, S., Karbach, D., Konig, J., Lepke, S., Wood, P.G., Appelhans, H. & Passow, H.
(1995) Inhibition of mouse erythroid band 3-mediated chloride transport by site-directed
mutagenesis of histidine residues and its reversal by second site mutation of Lys 558, the
locus of covalent H2DIDS binding. Biochemistry, 34, 9315-9324.
Murthy, S.N., Lomasney, J.W., Mak, E.C. & Lorand, L. (1999) Interactions of
G(h)/transglutaminase with phospholipase CΔ1 and with GTP. Proc Natl Acad Sci U S A,
96, 11815-11819.
Murthy, S.N., Wilson, J., Zhang, Y. & Lorand, L. (1994) Residue Gln-30 of human erythrocyte
anion transporter is a prime site for reaction with intrinsic transglutaminase. J Biol Chem,
269, 22907-22911.
Murzin, A.G., Brenner, S.E., Hubbard, T. & Chothia, C. (1995) SCOP: a structural classification
of proteins database for the investigation of sequences and structures. J Mol Biol, 247,
536-540.
Nakaoka, H., Perez, D.M., Baek, K.J., Das, T., Husain, A., Misono, K., Im, M.J. & Graham,
R.M. (1994) Gh: a GTP-binding protein with transglutaminase activity and receptor
signaling function. Science, 264, 1593-1596.
Nash, P.D., Opas, M. & Michalak, M. (1994) Calreticulin: not just another calcium-binding
protein. Mol Cell Biochem, 135, 71-78.
Nehls, V., Zeitler-Zapf, P. & Drenckhahn, D. (1993) Different sequences of expression of band
3, spectrin, and ankyrin during normal erythropoiesis and erythroleukemia. Am J Pathol,
142, 1565-1573.
Nielsen, A.D., Fuglsang, C.C. & Westh, P. (2003) Effect of calcium ions on the irreversible
denaturation of a recombinant Bacillus halmapalus α-amylase: a calorimetric
investigation. Biochem J, 373, 337-343.
124
Nurse, G.T., Coetzer, T.L. & Palek, J. (1992) The elliptocytoses, ovalocytosis and related
disorders. Baillieres Clin Haematol, 5, 187-207.
Ohshima, Y. & Gotoh, Y. (1987) Signals for the selection of a splice site in pre-mRNA.
Computer analysis of splice junction sequences and like sequences. J Mol Biol, 195, 247-
259.
Okubo, K., Kang, D., Hamasaki, N. & Jennings, M.L. (1994) Red blood cell band 3. Lysine 539
and lysine 851 react with the same H2DIDS (4,4'-diisothiocyanodihydrostilbene-2,2'-
disulfonic acid) molecule. J Biol Chem, 269, 1918-1926.
Oldenborg, P.A., Zheleznyak, A., Fang, Y.F., Lagenaur, C.F., Gresham, H.D. & Lindberg, F.P.
(2000) Role of CD47 as a marker of self on red blood cells. Science, 288, 2051-2054.
Oprea, G.E., Krober, S., McWhorter, M.L., Rossoll, W., Muller, S., Krawczak, M., Bassell, G.J.,
Beattie, C.E. & Wirth, B. (2008) Plastin 3 is a protective modifier of autosomal recessive
spinal muscular atrophy. Science, 320, 524-527.
Pace, C.N. (1986) Determination and analysis of urea and guanidine hydrochloride denaturation
curves. Methods Enzymol, 131, 266-280.
Pal, P., Lebedev, D., Salim, S. & Knauf, P.A. (2006) Substrates induce conformational changes
in human anion exchanger 1 (hAE1) as observed by fluorescence resonance energy
transfer. Biochemistry, 45, 6279-6295.
Pang, A.J., Bustos, S.P. & Reithmeier, R.A. (2008) Structural characterization of the cytosolic
domain of kidney chloride/bicarbonate anion exchanger 1 (kAE1). Biochemistry, 47,
4510-4517.
Patterson, S.T., Li, J., Kang, J.A., Wickrema, A., Williams, D.B. & Reithmeier, R.A. (2009)
Loss of specific chaperones involved in membrane glycoprotein biosynthesis during the
maturation of human erythroid progenitor cells. J Biol Chem, 284, 14547-14557.
Paw, B.H., Davidson, A.J., Zhou, Y., Li, R., Pratt, S.J., Lee, C., Trede, N.S., Brownlie, A.,
Donovan, A., Liao, E.C., Ziai, J.M., Drejer, A.H., Guo, W., Kim, C.H., Gwynn, B.,
Peters, L.L., Chernova, M.N., Alper, S.L., Zapata, A., Wickramasinghe, S.N., Lee, M.J.,
Lux, S.E., Fritz, A., Postlethwait, J.H. & Zon, L.I. (2003) Cell-specific mitotic defect and
dyserythropoiesis associated with erythroid band 3 deficiency. Nat Genet, 34, 59-64.
Perrotta, S., Borriello, A., Scaloni, A., De Franceschi, L., Brunati, A.M., Turrini, F., Nigro, V.,
del Giudice, E.M., Nobili, B., Conte, M.L., Rossi, F., Iolascon, A., Donella-Deana, A.,
Zappia, V., Poggi, V., Anong, W., Low, P., Mohandas, N. & Della Ragione, F. (2005)
The N-terminal 11 amino acids of human erythrocyte band 3 are critical for aldolase
binding and protein phosphorylation: implications for band 3 function. Blood, 106, 4359-
4366.
Perrotta, S., Gallagher, P.G. & Mohandas, N. (2008) Hereditary spherocytosis. Lancet, 372,
1411-1426.
125
Perrotta, S., Polito, F., Cone, M.L., Nobili, B., Cutillo, S., Nigro, V., Iolascon, A. & Amendola,
G. (1999) Hereditary spherocytosis due to a novel frameshift mutation in AE1
cytoplasmic COOH terminal tail: band 3 Vesuvio. Blood, 93, 2131-2132.
Peters, L.L., Jindel, H.K., Gwynn, B., Korsgren, C., John, K.M., Lux, S.E., Mohandas, N.,
Cohen, C.M., Cho, M.R., Golan, D.E. & Brugnara, C. (1999) Mild spherocytosis and
altered red cell ion transport in protein 4.2-null mice. J Clin Invest, 103, 1527-1537.
Peters, L.L., Shivdasani, R.A., Liu, S.C., Hanspal, M., John, K.M., Gonzalez, J.M., Brugnara, C.,
Gwynn, B., Mohandas, N., Alper, S.L., Orkin, S.H. & Lux, S.E. (1996) Anion exchanger
1 (band 3) is required to prevent erythrocyte membrane surface loss but not to form the
membrane skeleton. Cell, 86, 917-927.
Philippsen, P., Kleine, K., Pohlmann, R., Dusterhoft, A., Hamberg, K., Hegemann, J.H.,
Obermaier, B., Urrestarazu, L.A., Aert, R., Albermann, K., Altmann, R., Andre, B.,
Baladron, V., Ballesta, J.P., Becam, A.M., Beinhauer, J., Boskovic, J., Buitrago, M.J.,
Bussereau, F., Coster, F., Crouzet, M., D'Angelo, M., Dal Pero, F., De Antoni, A., Hani,
J. et al. (1997) The nucleotide sequence of Saccharomyces cerevisiae chromosome XIV
and its evolutionary implications. Nature, 387, 93-98.
Pinder, J.C., Pekrun, A., Maggs, A.M., Brain, A.P. & Gratzer, W.B. (1995) Association state of
human red blood cell band 3 and its interaction with ankyrin. Blood, 85, 2951-2961.
Popov, M., Li, J. & Reithmeier, R.A. (1999) Transmembrane folding of the human erythrocyte
anion exchanger (AE1, Band 3) determined by scanning and insertional N-glycosylation
mutagenesis. Biochem J, 339 ( Pt 2), 269-279.
Popov, M., Tam, L.Y., Li, J. & Reithmeier, R.A. (1997) Mapping the ends of transmembrane
segments in a polytopic membrane protein. Scanning N-glycosylation mutagenesis of
extracytosolic loops in the anion exchanger, band 3. J Biol Chem, 272, 18325-18332.
Quilty, J.A., Li, J. & Reithmeier, R.A. (2002) Impaired trafficking of distal renal tubular acidosis
mutants of the human kidney anion exchanger kAE1. Am J Physiol Renal Physiol, 282,
F810-820.
Quilty, J.A. & Reithmeier, R.A. (2000) Trafficking and folding defects in hereditary
spherocytosis mutants of the human red cell anion exchanger. Traffic, 1, 987-998.
Reithmeier, R.A.F., Chan, S.L. & Popov, M. (1996) Structure of the Erythrocyte Band 3 Anion
Exchanger. In: Transport Processes in Eukaryotic and Prokaryotic Organisms (ed. by
W.N. Konings, H.R. Kaback & J.S. Lolkema), Vol. 2, pp. 281-309. Elsevier Science,
Amsterdam.
Ribeiro, M.L., Alloisio, N., Almeida, H., Gomes, C., Texier, P., Lemos, C., Mimoso, G., Morle,
L., Bey-Cabet, F., Rudigoz, R.C., Delaunay, J. & Tamagnini, G. (2000) Severe hereditary
spherocytosis and distal renal tubular acidosis associated with the total absence of band 3.
Blood, 96, 1602-1604.
126
Ripoche, P., Bertrand, O., Gane, P., Birkenmeier, C., Colin, Y. & Cartron, J.P. (2004) Human
Rhesus-associated glycoprotein mediates facilitated transport of NH3 into red blood cells.
Proc Natl Acad Sci U S A, 101, 17222-17227.
Risinger, M.A., Dotimas, E.M. & Cohen, C.M. (1992) Human erythrocyte protein 4.2, a high
copy number membrane protein, is N-myristylated. J Biol Chem, 267, 5680-5685.
Risinger, M.A., Korsgren, C. & Cohen, C.M. (1996) Role of N-myristylation in targeting of band
4.2 (pallidin) in nonerythroid cells. Exp Cell Res, 229, 421-431.
Rodriguez-Soriano, J. (2000) New insights into the pathogenesis of renal tubular acidosis-from
functional to molecular studies. Pediatr Nephrol, 14, 1121-1136.
Rojiani, M.V., Finlay, B.B., Gray, V. & Dedhar, S. (1991) In vitro interaction of a polypeptide
homologous to human Ro/SS-A antigen (calreticulin) with a highly conserved amino acid
sequence in the cytoplasmic domain of integrin α subunits. Biochemistry, 30, 9859-9866.
Romero, M.F., Fulton, C.M. & Boron, W.F. (2004) The SLC4 family of HCO3- transporters.
Pflugers Arch, 447, 495-509.
Rybicki, A.C., Musto, S. & Schwartz, R.S. (1995) Identification of a band-3 binding site near the
N-terminus of erythrocyte membrane protein 4.2. Biochem J, 309 ( Pt 2), 677-681.
Rybicki, A.C., Qiu, J.J., Musto, S., Rosen, N.L., Nagel, R.L. & Schwartz, R.S. (1993) Human
erythrocyte protein 4.2 deficiency associated with hemolytic anemia and a homozygous
40glutamic acid→lysine substitution in the cytoplasmic domain of band 3 (band
3Montefiore). Blood, 81, 2155-2165.
Rybicki, A.C., Schwartz, R.S., Hustedt, E.J. & Cobb, C.E. (1996) Increased rotational mobility
and extractability of band 3 from protein 4.2-deficient erythrocyte membranes: evidence
of a role for protein 4.2 in strengthening the band 3-cytoskeleton linkage. Blood, 88,
2745-2753.
Rybicki, A.C., Schwartz, R.S., Qiu, J.J. & Gilman, J.G. (1994) Molecular cloning of mouse
erythrocyte protein 4.2: a membrane protein with strong homology with the
transglutaminase supergene family. Mamm Genome, 5, 438-445.
Sahr, K.E., Taylor, W.M., Daniels, B.P., Rubin, H.L. & Jarolim, P. (1994) The structure and
organization of the human erythroid anion exchanger (AE1) gene. Genomics, 24, 491-
501.
Salhany, J.M., Cordes, K.A. & Sloan, R.L. (1997) Gel filtration chromatographic studies of the
isolated membrane domain of band 3. Mol Membr Biol, 14, 71-79.
Salhany, J.M., Sloan, R.L. & Schopfer, L.M. (1996) Characterization of the stilbenedisulphonate
binding site on band 3 Memphis variant II (Pro-854→Leu). Biochem J, 317 ( Pt 2), 509-
514.
127
Salomao, M., Chen, K., Villalobos, J., Mohandas, N., An, X. & Chasis, J.A. (2010) Hereditary
spherocytosis and hereditary elliptocytosis: aberrant protein sorting during erythroblast
enucleation. Blood.
Sarabia, V.E., Casey, J.R. & Reithmeier, R.A. (1993) Molecular characterization of the band 3
protein from Southeast Asian ovalocytes. J Biol Chem, 268, 10676-10680.
Sarang, Z., Madi, A., Koy, C., Varga, S., Glocker, M.O., Ucker, D.S., Kuchay, S., Chishti, A.H.,
Melino, G., Fesus, L. & Szondy, Z. (2007) Tissue transglutaminase (TG2) facilitates
phosphatidylserine exposure and calpain activity in calcium-induced death of
erythrocytes. Cell Death Differ, 14, 1842-1844.
Satchwell, T.J., Shoemark, D.K., Sessions, R.B. & Toye, A.M. (2009) Protein 4.2: a complex
linker. Blood Cells Mol Dis, 42, 201-210.
Schofield, A.E., Martin, P.G., Spillett, D. & Tanner, M.J. (1994) The structure of the human red
blood cell anion exchanger (EPB3, AE1, band 3) gene. Blood, 84, 2000-2012.
Schofield, A.E., Reardon, D.M. & Tanner, M.J. (1992) Defective anion transport activity of the
abnormal band 3 in hereditary ovalocytic red blood cells. Nature, 355, 836-838.
Schwartz, R.S., Rybicki, A.C., Heath, R., Shew, R. & Lubin, B. (1987) Immunoreactive forms of
human erythrocyte protein 4.2 are present in non-erythroid cells and tissues. Blood, 70,
Suppl. 1, 42a (abstr.).
Sells, M.A., Li, J. & Chernoff, J. (1995) Delivery of protein into cells using polycationic
liposomes. Biotechniques, 19, 72-76, 78.
Sharkia, R., Beer, S. & Cabantchik, Z.I. (1994) A membrane-located polypeptide of Ulva sp.
which may be involved in HCO3- uptake is recognized by antibodies raised against the
human red-blood-cell anion-exchange protein. Planta, 194, 247-249.
Sherman, T., Chernova, M.N., Clark, J.S., Jiang, L., Alper, S.L. & Nehrke, K. (2005) The abts
and sulp families of anion transporters from Caenorhabditis elegans. Am J Physiol Cell
Physiol, 289, C341-351.
Showe, L.C., Ballantine, M. & Huebner, K. (1987) Localization of the gene for the erythroid
anion exchange protein, band 3 (EMPB3), to human chromosome 17. Genomics, 1, 71-
76.
Slupsky, J.R., Duggan-Keen, M., Booth, L.A., Karpas, A., Rhodes, E.G., Cawley, J.C. & Zuzel,
M. (1993) The peanut-agglutinin (PNA)-binding surface components of malignant
plasma cells. Br J Haematol, 83, 567-573.
Southgate, C.D., Chishti, A.H., Mitchell, B., Yi, S.J. & Palek, J. (1996) Targeted disruption of
the murine erythroid band 3 gene results in spherocytosis and severe haemolytic anaemia
despite a normal membrane skeleton. Nat Genet, 14, 227-230.
128
Sreerama, N. & Woody, R.W. (2000) Estimation of protein secondary structure from circular
dichroism spectra: comparison of CONTIN, SELCON, and CDSSTR methods with an
expanded reference set. Anal Biochem, 287, 252-260.
Steck, T.L. (1974) The organization of proteins in the human red blood cell membrane. A
review. J Cell Biol, 62, 1-19.
Steck, T.L., Ramos, B. & Strapazon, E. (1976) Proteolytic dissection of band 3, the predominant
transmembrane polypeptide of the human erythrocyte membrane. Biochemistry, 15,
1153-1161.
Steck, T.L. & Yu, J. (1973) Selective solubilization of proteins from red blood cell membranes
by protein perturbants. J Supramol Struct, 1, 220-232.
Sterling, D., Reithmeier, R.A. & Casey, J.R. (2001) A transport metabolon. Functional
interaction of carbonic anhydrase II and chloride/bicarbonate exchangers. J Biol Chem,
276, 47886-47894.
Studier, F.W. & Moffatt, B.A. (1986) Use of bacteriophage T7 RNA polymerase to direct
selective high-level expression of cloned genes. J Mol Biol, 189, 113-130.
Su, Y., Ding, Y., Jiang, M., Hu, X. & Zhang, Z. (2007) Protein 4.2 Komatsu (D175Y) associated
with the lack of interaction with ankyrin in human red blood cells. Blood Cells Mol Dis,
38, 221-228.
Su, Y., Ding, Y., Jiang, M., Jiang, W., Hu, X. & Zhang, Z. (2006) Associations of protein 4.2
with band 3 and ankyrin. Mol Cell Biochem, 289, 159-166.
Sung, L.A., Chien, S., Chang, L.S., Lambert, K., Bliss, S.A., Bouhassira, E.E., Nagel, R.L.,
Schwartz, R.S. & Rybicki, A.C. (1990) Molecular cloning of human protein 4.2: a major
component of the erythrocyte membrane. Proc Natl Acad Sci U S A, 87, 955-959.
Sung, L.A., Chien, S., Fan, Y.S., Lin, C.C., Lambert, K., Zhu, L., Lam, J.S. & Chang, L.S.
(1992) Human erythrocyte protein 4.2: isoform expression, differential splicing, and
chromosomal assignment. Blood, 79, 2763-2770.
Sung, L.A. & Lo, W.K. (1997) Immunodetection of membrane skeletal protein 4.2 in bovine and
chicken eye lenses and erythrocytes. Curr Eye Res, 16, 1127-1133.
Takano, J., Noguchi, K., Yasumori, M., Kobayashi, M., Gajdos, Z., Miwa, K., Hayashi, H.,
Yoneyama, T. & Fujiwara, T. (2002) Arabidopsis boron transporter for xylem loading.
Nature, 420, 337-340.
Takaoka, Y., Ideguchi, H., Matsuda, M., Sakamoto, N., Takeuchi, T. & Fukumaki, Y. (1994) A
novel mutation in the erythrocyte protein 4.2 gene of Japanese patients with hereditary
spherocytosis (protein 4.2 Fukuoka). Br J Haematol, 88, 527-533.
Tang, X.B., Fujinaga, J., Kopito, R. & Casey, J.R. (1998) Topology of the region surrounding
Glu681 of human AE1 protein, the erythrocyte anion exchanger. J Biol Chem, 273,
22545-22553.
129
Tanner, M.J. (1997) The structure and function of band 3 (AE1): recent developments (review).
Mol Membr Biol, 14, 155-165.
Tanner, M.J. (2002) Band 3 anion exchanger and its involvement in erythrocyte and kidney
disorders. Curr Opin Hematol, 9, 133-139.
Tanner, M.J., Martin, P.G. & High, S. (1988) The complete amino acid sequence of the human
erythrocyte membrane anion-transport protein deduced from the cDNA sequence.
Biochem J, 256, 703-712.
Tanphaichitr, V.S., Sumboonnanonda, A., Ideguchi, H., Shayakul, C., Brugnara, C., Takao, M.,
Veerakul, G. & Alper, S.L. (1998) Novel AE1 mutations in recessive distal renal tubular
acidosis. Loss-of-function is rescued by glycophorin A. J Clin Invest, 102, 2173-2179.
Thevenin, B.J. & Low, P.S. (1990) Kinetics and regulation of the ankyrin-band 3 interaction of
the human red blood cell membrane. J Biol Chem, 265, 16166-16172.
Towbin, H., Staehelin, T. & Gordon, J. (1979) Electrophoretic transfer of proteins from
polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl
Acad Sci U S A, 76, 4350-4354.
Towler, D.A., Gordon, J.I., Adams, S.P. & Glaser, L. (1988) The biology and enzymology of
eukaryotic protein acylation. Annu Rev Biochem, 57, 69-99.
Toye, A.M., Ghosh, S., Young, M.T., Jones, G.K., Sessions, R.B., Ramauge, M., Leclerc, P.,
Basu, J., Delaunay, J. & Tanner, M.J. (2005) Protein-4.2 association with band 3 (AE1,
SLCA4) in Xenopus oocytes: effects of three natural protein-4.2 mutations associated
with hemolytic anemia. Blood, 105, 4088-4095.
Toye, A.M., Williamson, R.C., Khanfar, M., Bader-Meunier, B., Cynober, T., Thibault, M.,
Tchernia, G., Dechaux, M., Delaunay, J. & Bruce, L.J. (2008) Band 3 Courcouronnes
(Ser667Phe): a trafficking mutant differentially rescued by wild-type band 3 and
glycophorin A. Blood, 111, 5380-5389.
Tse, W.T. & Lux, S.E. (1999) Red blood cell membrane disorders. Br J Haematol, 104, 2-13.
Ursitti, J.A., Pumplin, D.W., Wade, J.B. & Bloch, R.J. (1991) Ultrastructure of the human
erythrocyte cytoskeleton and its attachment to the membrane. Cell Motil Cytoskeleton,
19, 227-243.
van den Akker, E., Satchwell, T.J., Pellegrin, S., Flatt, J.F., Maigre, M., Daniels, G., Delaunay,
J., Bruce, L.J. & Toye, A.M. (2010a) Investigating the key membrane protein changes
during in vitro erythropoiesis of protein 4.2 (-) cells (mutations Chartres 1 and 2).
Haematologica, 95, 1278-1286.
van den Akker, E., Satchwell, T.J., Williamson, R.C. & Toye, A.M. (2010b) Band 3 multiprotein
complexes in the red cell membrane; of mice and men. Blood Cells Mol Dis, 45, 1-8.
130
Van Dort, H.M., Moriyama, R. & Low, P.S. (1998) Effect of band 3 subunit equilibrium on the
kinetics and affinity of ankyrin binding to erythrocyte membrane vesicles. J Biol Chem,
273, 14819-14826.
Van Kim, C.L., Colin, Y. & Cartron, J.P. (2006) Rh proteins: key structural and functional
components of the red cell membrane. Blood Rev, 20, 93-110.
Vince, J.W. & Reithmeier, R.A. (1998) Carbonic anhydrase II binds to the carboxyl terminus of
human band 3, the erythrocyte C1-/HCO3
- exchanger. J Biol Chem, 273, 28430-28437.
Voet, D. & Voet, J.G. (2004) Biochemistry. John Wiley & Sons, Inc., Hobeken.
Wada, H., Kanzaki, A., Yawata, A., Inoue, T., Kaku, M., Takezono, M., Sugihara, T., Yamada,
O. & Yawata, Y. (1999) Late expression of red cell membrane protein 4.2 in normal
human erythroid maturation with seven isoforms of the protein 4.2 gene. Exp Hematol,
27, 54-62.
Wagner, S., Vogel, R., Lietzke, R., Koob, R. & Drenckhahn, D. (1987) Immunochemical
characterization of a band 3-like anion exchanger in collecting duct of human kidney. Am
J Physiol, 253, F213-221.
Walder, J.A., Chatterjee, R., Steck, T.L., Low, P.S., Musso, G.F., Kaiser, E.T., Rogers, P.H. &
Arnone, A. (1984) The interaction of hemoglobin with the cytoplasmic domain of band 3
of the human erythrocyte membrane. J Biol Chem, 259, 10238-10246.
Wang, C.C., Badylak, J.A., Lux, S.E., Moriyama, R., Dixon, J.E. & Low, P.S. (1992)
Expression, purification, and characterization of the functional dimeric cytoplasmic
domain of human erythrocyte band 3 in Escherichia coli. Prot Sci, 1, 1206-1214.
Wang, C.C., Moriyama, R., Lombardo, C.R. & Low, P.S. (1995) Partial characterization of the
cytoplasmic domain of human kidney band 3. J Biol Chem, 270, 17892-17897.
Wang, D., Kühlbrandt, W., Sarabia, V.E. & Reithmeier, R.A.F. (1993) Two-dimensional
structure of the membrane domain of human Band 3, the anion transport protein of the
erythrocyte membrane. The EMBO Journal, 12, 2233-2239.
Westhoff, C.M., Ferreri-Jacobia, M., Mak, D.O. & Foskett, J.K. (2002) Identification of the
erythrocyte Rh blood group glycoprotein as a mammalian ammonium transporter. J Biol
Chem, 277, 12499-12502.
White, R.A., Peters, L.L., Adkison, L.R., Korsgren, C., Cohen, C.M. & Lux, S.E. (1992) The
murine pallid mutation is a platelet storage pool disease associated with the protein 4.2
(pallidin) gene. Nat Genet, 2, 80-83.
White, S. (2010) Membrane proteins of known 3D structure. University of California at Irvine.
Willardson, B.M., Thevenin, B.J., Harrison, M.L., Kuster, W.M., Benson, M.D. & Low, P.S.
(1989) Localization of the ankyrin-binding site on erythrocyte membrane protein, band 3.
J Biol Chem, 264, 15893-15899.
131
Wood, P.G., Muller, H., Sovak, M. & Passow, H. (1992) Role of Lys 558 and Lys 869 in
substrate and inhibitor binding to the murine band 3 protein: a study of the effects of site-
directed mutagenesis of the band 3 protein expressed in the oocytes of Xenopus laevis. J
Membr Biol, 127, 139-148.
Wright, M.H., Heal, W.P., Mann, D.J. & Tate, E.W. (2009) Protein myristoylation in health and
disease. J Chem Biol.
Yamaguchi, T., Ikeda, Y., Abe, Y., Kuma, H., Kang, D., Hamasaki, N. & Hirai, T. (2010)
Structure of the membrane domain of human erythrocyte anion exchanger 1 revealed by
electron crystallography. J Mol Biol, 397, 179-189.
Yannoukakos, D., Vasseur, C., Driancourt, C., Blouquit, Y., Delaunay, J., Wajcman, H. &
Bursaux, E. (1991) Human erythrocyte band 3 polymorphism (band 3 Memphis):
characterization of the structural modification (Lys 56→Glu) by protein chemistry
methods. Blood, 78, 1117-1120.
Yawata, Y. (1994) Band 4.2 abnormalities in human red cells. Am J Med Sci, 307, 190-203.
Yu, J. & Steck, T.L. (1975) Isolation and characterization of band 3, the predominant
polypeptide of the human erythrocyte membrane. J Biol Chem, 250, 9170-9175.
Zhang, D., Kiyatkin, A., Bolin, J.T. & Low, P.S. (2000) Crystallographic structure and
functional interpretation of the cytoplasmic domain of erythrocyte membrane band 3.
Blood, 96, 2925-2933.
Zhao, R. & Reithmeier, R.A. (2001) Expression and characterization of the anion transporter
homologue YNL275w in Saccharomyces cerevisiae. Am J Physiol Cell Physiol, 281,
C33-45.
Zhou, J. & Low, P.S. (2001) Characterization of the reversible conformational equilibrium in the
cytoplasmic domain of human erythrocyte membrane band 3. J Biol Chem, 276, 38147-
38151.
Zhou, Z., DeSensi, S.C., Stein, R.A., Brandon, S., Dixit, M., McArdle, E.J., Warren, E.M., Kroh,
H.K., Song, L., Cobb, C.E., Hustedt, E.J. & Beth, A.H. (2005) Solution structure of the
cytoplasmic domain of erythrocyte membrane band 3 determined by site-directed spin
labeling. Biochemistry, 44, 15115-15128.
Zhou, Z., DeSensi, S.C., Stein, R.A., Brandon, S., Song, L., Cobb, C.E., Hustedt, E.J. & Beth,
A.H. (2007) Structure of the cytoplasmic domain of erythrocyte band 3 hereditary
spherocytosis variant P327R: band 3 Tuscaloosa. Biochemistry, 46, 10248-10257.
Zhu, L., Kahwash, S.B. & Chang, L.S. (1998) Developmental expression of mouse erythrocyte
protein 4.2 mRNA: evidence for specific expression in erythroid cells. Blood, 91, 695-
705.
Zhu, Q. & Casey, J.R. (2007) Topology of transmembrane proteins by scanning cysteine
accessibility mutagenesis methodology. Methods, 41, 439-450.
132
Zhu, Q., Lee, D.W. & Casey, J.R. (2003) Novel topology in C-terminal region of the human
plasma membrane anion exchanger, AE1. J Biol Chem, 278, 3112-3120.
133
Appendix: Structural characterization of the cytoplasmic domain of the
kidney chloride/bicarbonate anion exchanger 1 (kAE1)
Adapted with permission from Pang, A.J., Bustos, S.P. & Reithmeier, R.A. Structural
characterization of the cytosolic domain of kidney chloride/bicarbonate anion exchanger 1
(kAE1). Biochemistry, 47, 4510-4517. Copyright 2008 American Chemical Society
(http://pubs.acs.org/doi/abs/10.1021/bi702149b) (Pang et al. 2008). Jing Li contributed to this
work by generation of the cdAE1, cdkAE1, and cdΔ54AE1 cDNA vectors. I supervised and
trained summer student Allison Pang, who performed all experiments and assisted in the
experimental design and analysis of results. I designed all experiments and assisted in the
analysis of results.
Abstract
kAE1 is a membrane glycoprotein expressed in α-intercalated cells in the collecting ducts of the
kidney where it mediates electroneutral chloride/bicarbonate exchange. Human kAE1 is a
truncated form of erythroid AE1 that is missing the first 65 residues of the N-terminal
cytoplasmic domain, which includes a disordered acidic region (residues 1-54) and the first β-
strand (residues 55-65) of the folded region. Unlike erythroid AE1, kAE1 does not bind
deoxyhemoglobin, glycolytic enzymes, or ankyrin. To understand the effect of the N-terminal
deletion on the structure of the cytoplasmic domain, we performed an extensive biophysical
analysis on His6-tagged cdAE1, the cytoplasmic domain of kAE1 (cdkAE1), and a novel
truncation mutant (cdΔ54AE1) missing the first 54 residues, but retaining the β-strand. CD did
not reveal any major differences in secondary structure, and sedimentation equilibrium
experiments showed that all three proteins were dimeric. DSC revealed that cdAE1 and
cdΔ54AE1 had similar thermal stabilities, with apparent midpoints of transition higher than
cdkAE1. cdAE1 and cdΔ54AE1 underwent similar pH-dependent fluorescence changes, while
cdkAE1 exhibited a higher fluorescence at neutral and acidic pH. Urea denaturation resulted in
dequenching of tryptophan fluorescence in cdAE1, while tryptophans in cdkAE1 were already
dequenched in the native state. We conclude that the absence of the central β-strand in cdkAE1
results in a less conformationally stable and more open structure than cdAE1. This structural
change, in addition to the loss of the acidic N-terminal region, may account for the altered
protein binding properties of kAE1.
134
Introduction
AE1, also known as Band 3, is a major integral membrane protein of the human erythrocyte
membrane. It is a 911 amino acid glycoprotein that is responsible for the electroneutral exchange
of bicarbonate for chloride (Lepke and Passow 1976). Human AE1 has a monomer molecular
weight of 95 kDa and exists as a dimer and a tetramer in the red blood cell membrane (Casey and
Reithmeier 1991b, Wang et al. 1993). Mild proteolytic cleavage of this protein in erythrocyte
membranes yields two functional domains. The 52 kDa C-terminal TM domain (Gly361-Val911)
spans the membrane up to 12 times (Fujinaga et al. 1999, Popov 1999, Reithmeier et al. 1996,
Zhu et al. 2003) and mediates the anion transport function (Jennings 1989a, Lepke and Passow
1976). The 43 kDa N-terminal cytoplasmic domain includes Met1-Lys360 and provides binding
sites for various red cell cytoskeletal and cytoplasmic proteins (Lepke and Passow 1976). The
cytoplasmic domain of AE1 (cdAE1) acts as an anchoring site for ankyrin and protein 4.2, both
of which are found in the membrane’s cytoskeleton (Hargreaves et al. 1980, Perrotta et al. 2005).
AE1 is therefore thought to play an important role in maintaining the shape, stability, and
flexibility of the red blood cell (Lux and Palek 1995, Peters et al. 1996). AE1 also binds and
regulates the function of deoxyhemoglobin and various glycolytic enzymes such as GAPDH,
aldolase, and phosphofructokinase (Low 1986, Perrotta et al. 2005, Walder et al. 1984). The
cytoplasmic domain is characterized by a highly acidic N-terminal region with an N-acetylated
terminal methionine (Wang et al. 1992). This region of the domain is involved in binding
associated proteins (Perrotta et al. 2005), although other parts of the cytoplasmic domain also
provide protein binding sites, such as the ankyrin-binding loop (Chang and Low 2003).
In erythrocytes, AE1 transports bicarbonate, which is produced by CAII, into the plasma in
exchange for chloride, thereby increasing the CO2-carrying capacity of the blood (Jennings
1989a). In the human kidney, a truncated form of AE1 (kAE1) catalyzes the exchange of
bicarbonate and chloride across the basolateral membrane of acid-secreting cells in the collecting
ducts of the kidney, resulting in bicarbonate reabsorption into the blood and promoting acid
excretion into the urine. Human kAE1 is missing residues 1-65 of the erythroid form (Wang et
al. 1995). Since the acidic residues of the N-terminal region, which electrostatically bind
cytoskeletal and cytoplasmic proteins, are absent in this protein, the cytoplasmic domain of
135
kidney AE1 (cdkAE1) is thought to have a different function than the erythroid isoform,
potentially mediating the binding of specific kidney proteins. Indeed, kAE1 has been shown not
to bind to glycolytic enzymes, deoxyhemoglobin, protein 4.1, or ankyrin (Ding et al. 1994, Wang
et al. 1995a), but has been found to bind to integrin-linked kinase, a protein involved in actin
cytoskeletal interactions (Keskanokwong et al. 2007).
The crystal structure of erythroid cdAE1 has been elucidated (Figure A1) at low pH by X-
ray diffraction, revealing a tight, symmetric dimer stabilized by interlocking dimerization arms
contributed by each subunit (Zhang et al. 2000). The purified recombinant protein has been
shown to have the same secondary structure, Stokes radius, and display similar pH-dependent
conformational changes as the native cytoplasmic domain prepared from red blood cells (Wang
et al. 1992b). Except for a missing N-acetylated N-terminus, no significant differences were
observed between the recombinant and native proteins. The first 54 residues of AE1 were
unresolved in the crystal structure due to the strongly anionic and disordered nature of the region
(Zhang et al. 2000, Zhou et al. 2005). The crystal structure also shows that residues 55-65 in
AE1 form a β-strand that is present in the core of the protein. Since the kidney isoform is missing
this β-strand, it is possible that the kidney cytoplasmic domain has a significantly different
structure from the full-length erythroid form.
Figure A1: Crystal structure of human cdAE1.
The structure of the cdAE1 dimer was solved at a resolution of 2.90 Å at pH 4.8 (Zhang et al.
2000). One monomer is coloured dark gray and the other is coloured light gray. The β-strand
insert, which is absent in the kidney isoform, is shown in black. The N-terminal residues 1–54
are not visible in the crystal structure. The positions of the four tryptophan residues in cdAE1 are
shown in black.
136
The goal of this work was to structurally analyze cdAE1 and cdkAE1, as well as a novel
truncation mutant, cdΔ54AE1, which lacks the first 54 residues but retains the core β-strand. A
schematic of these three constructs is illustrated in Figure A2.
Figure A2: Domain structure of AE1 isoforms and gel-separated AE1 constructs.
(A) Domain structure of AE1, kAE1, and the cdAE1, cdkAE1, and cdΔ54AE1 constructs.
Residues 1–54 are not visible in the crystal structure, and residue 356 is the last residue that was
visible in the crystal structure of cdAE1. Residues 55–65 encompass the first β-strand in cdAE1.
(B) 12 % SDS-PAGE gel of the purified constructs stained with Coomassie blue. Lane 1:
molecular weight markers. Lane 2: eluate of purified cdkAE1. Lane 3: eluate of purified cdAE1.
Lane 4: eluate of purified cdΔ54AE1.
137
Materials and methods
Materials
The following is a list of materials used and their suppliers: pcDNA3 vector (Invitrogen, San
Diego, CA); mutagenic primers (ACGT Corp., Toronto, ON); pETBlue-1 vector and Tuner
BL21(DE3)pLacI E .coli competent cells (Novagen, Madison, WI); growth media for E. coli
(BD, Sparks, MD); chloramphenicol and carbenicillin (Sigma, St.Louis, MO); IPTG (Bioshop,
Burlington, ON); Ni-NTA agarose resin (QIAGEN, Germantown, MD); PD-10 gel filtration
columns (Amersham Biosciences); lysozyme from chicken egg white (Sigma, St.Louis, MO);
deoxyribonuclease from bovine pancreas (Sigma, St. Louis, MO); and Sequanal grade urea
(Pierce, Rockford, IL).
Plasmid construction and mutagenesis
The cdAE1 construct was amplified by PCR from full-length human AE1 on a pcDNA3 vector,
which was then cloned into a pETBlue-1 expression vector. This expression vector contained an
IPTG-inducible T7lacO promoter (Giordano et al. 1989). The reverse primer encoded six
histidine residues that were used as a C-terminal tag for purification purposes. cdkAE1 and
cdΔ54AE1 were also amplified by PCR from full-length human AE1 using primers that
corresponded to their respective N- and C-termini. The primary sequence of all three proteins
extended to the last visible residue in the crystal structure of cdAE1, which is serine 356.
Constructs were confirmed by sequencing by ACGT Corp.
Protein expression and purification
All constructs were expressed in E. coli Tuner BL21(DE3)pLacI competent cells at 37 °C. Large
cultures of cells were grown in one liter of LB medium containing 50 μg/ml carbenicillin and 34
μg/ml chloramphenicol until an A600 of 0.5-0.6. Expression of the constructs was induced by the
addition of 1 mM IPTG (Studier and Moffatt 1986). Cells were grown for an additional five h
and then harvested by centrifugation at 4 000 rpm for 30 min. Cell pellets were solubilized in 80
ml of lysis buffer (50 mM sodium phosphate, 300 mM sodium chloride, 10 mM imidazole, 0.2 %
βME, and 0.2 % Triton X-100 (pH 8.0)). The following protease inhibitors were used: 2 μg/ml
aprotinin, 1.6 mM PMSF, 0.7 μg/ml pepstatin A, and 10 μM leupeptin. DNase and lysozyme
138
were added to the cell pellets. The solubilized cell pellets were sonicated at 40 % duty cycle for
1.5 min on ice. Purification of the protein constructs was carried out using 1 ml of Ni-NTA
agarose beads (QIAGEN) per 80 ml lysate at 4 ºC. Resin was equilibrated using the lysis buffer
and, following protein binding, was washed twice with 10 ml of wash buffer (50 mM sodium
phosphate, 300 mM sodium chloride, 20 mM imidazole, and 0.2 % βME, pH 8.0). Protein was
eluted three times using 1 ml of elution buffer (50 mM sodium phosphate, 300 mM sodium
chloride, 250 mM imidazole, and 0.2 % βME pH 8.0). The three elution fractions were
combined, filtered, and applied to a pre-equilibrated PD-10 gel filtration column in order to
exchange the buffer with 10 mM NH4HCO3. Proteins were lyophilized overnight and stored at -
20 °C. Protein purity was determined to be >95 % by SDS-PAGE and Coomassie blue staining.
The final concentration of the purified proteins was measured by the BioRad protein assay,
which is based on the Bradford assay.
Analytical ultracentrifugation
SE experiments were performed on an Optima XL-A/XL-I analytical ultracentrifuge (Beckman
Instruments, Palo Alto, CA) at 10 000, 13 000, 16 000 and 19 000 rpm at 20 °C. Samples were
prepared by dissolving lyophilized protein in 10 mM sodium phosphate and 50 mM sodium
chloride (pH 7.5). SE experiments were carried out on cdAE1, cdkAE1, and cdΔ54AE1, each at
three different concentrations with the corresponding A280 values of 0.3, 0.6, and 1.0. Data
analysis was performed using XL-A/XL-I software (Origin version 4.1) from Beckman
Instruments.
SV experiments were also performed on an Optima XL-A/XL-I analytical ultracentrifuge
(Beckman Instruments, Palo Alto, CA) at 30 000 rpm at 4 ºC. Samples were prepared by
dissolving lyophilized protein in 50 mM sodium phosphate and 100 mM sodium chloride (pH
7.5). SV experiments were carried out on cdAE1, cdkAE1, and cdΔ54AE1, each at three
different concentrations with the corresponding A280 values of 0.3, 0.6, and 1.0. Data analysis
was performed using XL-A/XL-I software (SedFit and Ultrascan) from Beckman Instruments.
139
Circular dichroism
Lyophilized protein was dissolved in buffer containing 50 mM sodium phosphate and 100 mM
sodium chloride, and adjusted to either pH 5.0, 7.5, or 10.5. Samples were filtered through a 0.22
µm syringe filter. The final concentrations of the protein solutions are indicated in the figure
captions. CD was performed on a Jasco J-810 spectropolarimeter. All spectra were measured at
24 °C from 200 to 260 nm with a 1 nm data pitch in a 1 mm path length cell. The mean residue
ellipticity was measured as a function of wavelength. The secondary structure content of the
spectra was determined using the SELCON3 program from the CDPro software package.
Differential scanning calorimetry
Lyophilized protein was dissolved in buffer containing 50 mM sodium phosphate and 100 mM
sodium chloride, and pre-adjusted to either pH 5.0, 7.5, or 10.5. Samples were filtered through a
0.22 µm syringe filter. The final concentrations of the protein solutions are indicated in the figure
captions. Heat capacity measurements from 25 to 90 °C were obtained on a Microcal VP-DSC
differential scanning calorimeter. Samples were heated at a rate of 1.5 °C/min. The heat capacity
was plotted as a function of temperature. Origin 7.0 data analysis software was used to baseline
correct and to analyze the temperature denaturation data by fitting to a two-state transition
model.
pH dependence of intrinsic fluorescence
Stock solutions of each protein were made by dissolving lyophilized samples in 50 mM sodium
phosphate and 100 mM sodium chloride (pH7.5) and by filtering them through a 0.22 µm
syringe filter. The stock protein was then diluted 50 × into the same buffer preadjusted to the
desired pH, ranging from pH 5.0 to pH 10.5 in 0.5 increments. The final concentration of the
protein samples was ~0.1-0.2 mg/ml. Samples were equilibrated overnight at 4 °C and were
allowed to reach room temperature prior to fluorescence measurement. The excitation
wavelength was 290 nm and the fluorescence emission was measured from 300 to 420 nm for
each protein sample at each pH value. Measurements were taken at 24 °C by a Fluorolog-3 FL3-
22 spectrofluorometer. The average emission wavelength for each sample, which is equal to
140
Σ(λ*intensity)/Σ(intensity), was plotted as a function of pH. Fluorescence intensity was also
plotted against wavelength for samples at pH 5.0, 7.5, and 10.5.
Urea denaturation measured by intrinsic fluorescence
Stock solutions of cdAE1 and cdkAE1 were made by dissolving lyophilized samples in 50 mM
sodium phosphate and 100 mM sodium chloride (pH7.5) and filtering them through a 0.22 µm
syringe filter. The stock protein was diluted 50 × into the same buffer preadjusted to the desired
urea concentration ranging from 0 – 8 M. Samples were equilibrated overnight at 4 °C and were
allowed to reach room temperature prior to fluorescence measurement. The excitation
wavelength was 290 nm and the fluorescence emission was measured from 300 to 420 nm for
each protein sample at each urea concentration. Measurements were taken at 24 °C by a
Fluorolog-3 FL3-22 spectrofluorometer. The maximum peak fluorescence intensity was plotted
against urea concentration. The average emission wavelength was also plotted as a function of
urea concentration.
Results
Expression and purification of cdAE1, cdkAE1, and cdΔ54AE1 proteins
The purification of His6-tagged cdAE1, cdkAE1, and cdΔ54AE1 was performed by Ni-NTA
affinity chromatography as described in Methods. This method of purification yielded
approximately 15-20 mg of protein per liter of cell culture, the kidney isoform routinely having
the highest expression. All three proteins were more than 95 % pure as determined by SDS-
PAGE, and they ran as monomers with molecular masses of roughly 43 kDa (cdAE1), 32 kDa
(cdkAE1), and 35 kDa (cdΔ54AE1). The predicted molecular weights of cdAE1, cdkAE1, and
cdΔ54AE1, including their His6 tag, are 40 866 Da, 33 269 Da, and 34 738.5 Da, respectively.
The identity of the constructs were confirmed by mass spectrometry analysis which gave
molecular weights of 40 942.301 Da (cdAE1), 33 345.699 Da (cdkAE1), and 34 814.801 Da
(cdΔ54AE1), all of which were ~76 Da higher than the predicted values. This extra molecular
mass was most likely due to the formation of a βME adduct (+76 Da) on one of the cysteine
residues during purification.
141
Analytical ultracentrifugation: sedimentation equilibrium and sedimentation
velocity experiments on cdAE1, cdkAE1, and cdΔ54AE1 proteins
SE experiments were carried out to determine the oligomeric state of the proteins and to see if
the missing 65 (cdkAE1) or 54 residues (cdΔ54AE1) affects the dimerization of the cytoplasmic
domain of AE1. It was previously shown by gel filtration combined with disulfide bond cross-
linking experiments (Appell and Low 1981), and SE experiments (Bustos and Reithmeier 2006)
that cdAE1 exists as a dimer. It was predicted that the missing N-terminal residues of cdkAE1
and cdΔ54AE1 would not distort the native structure of the C-terminal dimerization arm. SE
experiments indicated MWapp values that were approximately twice that of their MWseq. The
ratios of MWapp versus MWseq are listed in Table A1. The plots of ln(Abs) versus the radius
squared were linear for all samples, indicating that the samples were primarily composed of one
major species. A representative plot from the SE experiments on cdAE1 is shown in Figure A3,
as well as the residuals from fitting the data to a single-ideal species model. The results show that
the three proteins existed mainly as stable dimers in solution.
SV experiments were carried out to compare the relative conformations of cdAE1, cdkAE1,
and cdΔ54AE1 based on their S20,w values. The sedimentation coefficient of the erythroid
cytoplasmic fragment has been previously estimated to be ~4.1 S (Appell and Low 1981), in
agreement with our own result for this construct. cdAE1 and cdΔ54AE1 had sedimentation
values of 3.8 S and 3.9 S, respectively. Smaller S20,w values indicated that the truncated proteins
moved more slowly in response to the centrifugal force, likely due to their lower molecular
weights and perhaps more extended structures.
142
Table A1: Summary of biophysical properties of cdAE1, cdkAE1, and cdΔ54AE1 at pH 7.5
Property cdAE1 cdΔ54AE1 cdkAE1
MWapp/MWseq
oligomeric state (SE)
1.83 dimer 1.77 dimer 1.93 dimer
S20,W (SV) 4.1 S 3.9 S 3.7 S
Secondary structure (CD) % α-helix = 30.5 ± 1.2;
n = 3
% α-helix = 34.7 ± 1.8;
n = 3
% α-helix = 32.0 ± 0.5;
n = 3
Apparent midpoint of
transition (Tm (°C)) (DSC)
pH 5 = 73.7 ± 1.7; n = 4
pH 7.5 = 65.3 ± 1.2; n = 12
pH 10.5 = 56.1 ± 2.7; n = 6
pH 5 = 73.1 ± 2.9; n = 5
pH 7.5 = 65.8 ± 1.2; n = 11
pH 10.5 = 56.0 ± 2.0; n = 6
pH 5 = 62.4 ± 5.3; n = 5
pH 7.5 = 60.1 ± 2.2; n = 11
pH 10.5 = 50.7 ± 3.0; n = 5
Figure A3: Analytical ultracentrifugation of cdAE1.
Absorbance at 280 nm is plotted as a function of centrifuge cell radius. The residuals from fitting
the data to a single ideal species model are shown. The SE experiment was done at 20 °C at a
speed of 13 000 rpm. The protein was dissolved in 50 mM sodium chloride and 10 mM sodium
phosphate at pH 7.5.
143
Secondary structure content of cdAE1, cdkAE1, and cdΔ54AE1: analysis by
circular dichroism spectroscopy
The secondary structures of cdAE1, cdkAE1, and cdΔ54AE1 were examined by CD
spectroscopy. A representative spectrum of each protein at pH 7.5 is shown in Figure A4. A
comparison of the CD spectra, however, revealed only minor variations in secondary structure
content between the constructs. Each spectrum exhibited a negative maximum at 208 nm and a
shoulder at 222 nm, which are dominant features of an α-helical protein.
Figure A4: CD spectra of purified cdAE1, cdkAE1, and cdΔ54AE1 at pH 7.5. Lyophilized
proteins were dissolved in 100 mM sodium chloride and 50 mM sodium phosphate (pH 7.5) and
scanned at 24 °C in a 1 mm cell in a Jasco J-810 spectropolarimeter. The final concentrations of
cdAE1 ( — ), cdkAE1 (– – –), and cdΔ54AE1 (−×−), were 0.36 mg/ml, 0.38 mg/ml, and 0.34
mg/ml, respectively. The spectra are expressed as the mean residue ellipticity (θ) as a function of
wavelength.
The helical content of cdAE1 based on its crystal structure was 26 %. The deconvolution of
the CD spectra revealed similar helical content between cdAE1, cdkAE1, and cdΔ54AE1 of
approximately 30-35 % α-helix (Table A1). Since the crystal structure of cdAE1 was elucidated
at pH 4.8 (Zhang et al. 2000b), it was important to determine if the secondary structure differs at
physiological and acidic conditions. For this reason, pH values of 5.0, 7.5, and 10.5 were used to
detect possible pH-dependent structural changes. Based on previous fluorescence and gel
144
filtration experiments, there is evidence that the global conformation of cdAE1 elongates as pH
is increased (Appell and Low 1981). However, there were no changes detected in the CD spectra
of the three constructs at the different pH values tested, indicating that the secondary structural
features were maintained.
Thermal denaturation of cdAE1, cdkAE1, and cdΔ54AE1 by differential scanning
calorimetry
Calorimetry can detect thermal stability changes in proteins, which appear as perturbations in the
midpoint of thermal denaturation transitions (Tm). Thermal denaturation by DSC was used to
compare the thermal stabilities of the cdAE1, cdkAE1, and cdΔ54AE1 proteins. Representative
plots of the thermal denaturation of cdAE1, cdkAE1, and cdΔ54AE1 at pH 7.5, as performed by
DSC, are shown in Figure A5. Cooling and reheating of the samples revealed that the thermal
denaturation was irreversible. The data were fit to a two-state transition model for the purpose of
obtaining the midpoints of the thermal denaturation transitions (Tm) so that they could be
compared to one another. Because the transitions were irreversible, these Tms were relative and
were considered to be the apparent Tms of the transitions. These Tms agree with the temperature
corresponding to the maximum excess heat capacity (Cp), which was used in other studies of
proteins with irreversible thermal denaturation profiles (Idakieva et al. 2005, Nielsen et al.
2003). It was found that cdkAE1 had the lowest Tm (60 ºC) and that cdAE1 (65 ºC) and
cdΔ54AE1 (66 ºC) had similar Tms. The results indicate that the thermal stability of cdAE1 was
unaffected by removal of the disordered N-terminal extension (residues 1-54) but was thermally
destabilized by removal of the core β-strand.
It was also observed that the thermal stability of the three proteins was pH-dependent; the
proteins are the least thermally stable at alkaline pH and the most thermally stable at acidic pH.
The effect of pH on the Tm values of cdAE1, cdkAE1, and cdΔ54AE1 is recorded in Table A1.
cdAE1 and cdΔ54AE1 behaved similarly, while cdkAE1 had a lower Tm at all pH values. These
observations corresponded with other studies that have reported the Tm of cdAE1 to be pH-
dependent with a value of 67 ºC at pH 7.51 (Appell and Low 1981). It was also noticed that the
peaks at low pH were broader, most likely due to aggregation of the samples. This may be
145
Figure A5: Thermal denaturation of purified cdAE1, cdkAE1, and cdΔ54AE1 by DSC at
pH 7.5.
All DSC scans were baseline-subtracted. The proteins were dissolved in 100 mM sodium
chloride and 50 mM sodium phosphate (pH 7.5) to final protein concentrations of 0.9 mg/ml
(cdAE1 —– ), 1.2 mg/ml (cdkAE1 – – –), and 1.0 mg/ml (cdΔ54AE1 −×−). The specific heat
capacity (Cp) was measured as the temperature was increased from 40 to 80 °C. No transition
was observed once the proteins were thermally denatured, cooled, and reheated.
explained by the fact that at pH 5.0, the proteins approach their isoelectric points, the pI of
cdAE1, cdkAE1, and cdΔ54AE1 being 4.70, 5.29, and 5.30, respectively. These thermal
denaturation studies show that the β-strand plays a structural role in the thermal stabilization of
the cytoplasmic domain. Furthermore, the cytoplasmic domains containing the β-strand are the
most thermally stable at low pH. cdkAE1 has only a slightly higher midpoint of transition at pH
5 than at neutral pH suggesting that it is less well-packed under acidic conditions compared to
cdAE1.
Effect of pH on the intrinsic fluorescence of cdAE1, cdkAE1, and cdΔ54AE1
Differences in tertiary structure between cdAE1, cdkAE1, and cdΔ54AE1 were further studied
by carrying out intrinsic tryptophan fluorescence studies. It has been previously shown that
cdAE1 undergoes a pH-dependent conformational change, characterized by a dramatic increase
146
in tryptophan fluorescence at alkaline pH (Appell and Low 1981) and an increase in its Stokes
radius (Wang et al. 1992b). Since a cluster of four tryptophan residues is located C-terminal to
the 65 residue amino extension, the local environment of these tryptophans in cdkAE1 may
differ from that of cdAE1, and the tryptophans would therefore exhibit different fluorescence
properties and sensitivities towards pH titration. As the pH increased, cdAE1, cdkAE1, and
cdΔ54AE1 experienced an increase in fluorescence intensity due to dequenching of their
tryptophan residues (Figure A6). cdAE1 and the novel mutant cdΔ54AE1 both showed a similar
increase in intensity, doubling in value from pH 5 to 10.5. At acidic and neutral pH, however,
cdkAE1 exhibited a higher intrinsic fluorescence than cdAE1, which suggests that it has a more
open structure under these conditions compared to the cdAE1 protein.
Figure A7 shows representative plots of the average emission wavelength of all three
proteins as a function of pH. A red-shift in peak wavelength at alkaline pH was observed,
indicating that the tryptophans were becoming exposed to a more polar environment. Even at
neutral pH the fluorescence spectrum of cdkAE1 was more red-shifted, indicating a higher
degree of exposure of tryptophans to a more polar environment. At alkaline pH, the three
constructs exhibited similar fluorescence properties, consistent with a similar exposure of
tryptophans to solvent.
Urea denaturation of cdAE1 and cdkAE1 by intrinsic tryptophan fluorescence
Urea denaturation was used to examine the conformational stability of cdAE1 and cdkAE1, and
to further understand the consequence of truncating the N-terminus of AE1’s cytoplasmic
domain. The observed fluorescence emission intensity of cdAE1 peaked at 4 M urea due to the
dequenching of its tryptophans, as seen in Figure A8 A. This increase was then followed by a
decrease in intensity at higher urea concentrations due to further exposure of the Trp residues to
the aqueous solvent. In contrast, cdkAE1displayed a higher intrinsic fluorescence under native
conditions and a continual decrease in fluorescence as urea increased. Both constructs, however,
eventually denatured to a similar unfolded state as seen by the convergence of their fluorescence
intensities at 6 M urea.
147
Figure A6: Intrinsic tryptophan fluorescence emission spectra of (A) cdAE1, (B) cdkAE1,
and (C) cdΔ54AE1 at various pH values. The intensity is plotted as a function of wavelength.
Lyophilized proteins were dissolved in 100 mM sodium chloride and 50 mM sodium phosphate
preadjusted to the desired pH. The intrinsic fluorescence emission was measured from 300 to
420 nm following excitation at 290 nm. pH 5.31(—), pH 7.48 (− − − ), and pH 10.04 (−×−).
148
The average emission wavelength red-shifted as the amount of urea increased (Figure A8 B).
This observed red-shift indicated that the tryptophans were progressively becoming exposed to a
more polar environment. As shown in Figure A8 B, cdAE1 and cdkAE1 underwent similar
conformational changes; however, cdkAE1 was less resistant to urea at lower concentrations of
the denaturant.
Figure A7: Average emission wavelength of purified cdAE1, cdkAE1, and cdΔ54AE1.
The average emission wavelength is plotted as a function of pH. Lyophilized proteins were
dissolved in 100 mM sodium chloride and 50 mM sodium phosphate preadjusted to the desired
pH. The intrinsic fluorescence emission was measured from 300 to 420 nm following excitation
at 290 nm. cdAE1 (—), cdkAE1 (− − −), and cdΔ54AE1 (−×−).
Discussion
In this study, the biophysical properties of cdAE1, cdkAE1, and the novel mutant cdΔ54AE1
were compared to determine the effect of loss of the central β-strand on the structure and
conformational stability of the cytoplasmic domain. As shown by SE experiments, removal of
residues 1-65 had no effect on the dimerization of the cytoplasmic domain of AE1, indicating
that cdkAE1 was still able to retain a dimeric structure in solution. CD analysis showed that the
amino-truncations (either Δ1-65 or Δ1-54) did not have a significant effect on the secondary
structure content of the cytoplasmic domain. The loss of the short β-strand alone (residues 55-65)
would not be expected to produce dramatic changes in the CD spectrum.
149
Figure A8: Urea denaturation of cdAE1 and cdkAE1 monitored by intrinsic tryptophan
fluorescence.
Lyophilized proteins were dissolved in 100 mM sodium chloride and 50 mM sodium phosphate
preadjusted to the desired urea concentration. The intrinsic fluorescence intensity was measured
from 300 to 420 nm following excitation at 290 nm. The vertical bars represent the standard
deviation; n = 4. (A) Intrinsic fluorescence emission intensity of the proteins at 341 nm is plotted
as a function of urea concentration. (B) Average emission wavelength was calculated from each
spectrum and plotted as a function of urea concentration. cdAE1 (—), cdkAE1 (− − −).
DSC compared the thermal stability of cdkAE1 to that of cdAE1. cdkAE1 consistently had
lower Tm values regardless of the pH, indicating a less thermally stable structure. The
calorimetric values of the novel mutant cdΔ54AE1 were similar to those of cdAE1,
strengthening the contention that the β-strand (residues 55-65) helps stabilize the cytoplasmic
domain of AE1. These results were predicted because the crystal structure of AE1 shows that the
β-strand is located within the core of the domain. When the β-strand was removed, as in cdkAE1,
150
there was a significant effect on the Tm. As expected, removal of residues 1-54 did not alter the
thermal stability of the domain as these residues form a flexible extension that was not observed
by X-ray crystallography. Varying the pH had a noticeable effect on the thermal stability of all
three proteins. At acidic pH, the Tm value increased compared to neutral and alkaline pH. This
indicated that the proteins acquired a greater thermal stability at low pH, perhaps adopting a
more compact conformation as revealed by the crystal structure. cdkAE1 was always the least
thermally stable at all pH values, however, the difference in Tm was greatest (~10 °C) at low pH.
The intrinsic fluorescence of a folded protein is primarily dependent on the local
environment of its tryptophan residues. Normally, when tryptophan is exposed to a hydrophobic
environment, the fluorescence emission is of high intensity and is blue-shifted. Upon exposure to
a more polar environment, tryptophan fluorescence decreases and its emission spectrum
experiences a red-shift. However, this is often not the case for multi-tryptophan proteins, such as
cdAE1, where there are four tryptophan residues found at position 75, 81, 94, and 105. At
alkaline pH, the fluorescence intensity of cdAE1 increased and there was a red-shift in average
emission wavelength. These observations indicated that an opening of the structure had occurred,
thereby dequenching fluorescence and exposing the tryptophan residues to a more polar
environment. Both cdAE1 and cdΔ54AE1 underwent similar pH-dependent conformational
changes, indicating structural similarity. cdkAE1 was not as structurally sensitive to pH titration
as were the other two protein constructs because of its more open structure at low pH. In
comparison, cdAE1 and the novel mutant had a more compact and folded structure at low pH,
and gradually adopted a more open conformation as pH increased. Although the secondary
structure content between proteins is comparable, the intrinsic fluorescence data suggest that
cdkAE1 has a less compact structure, especially at low pH.
Intrinsic fluorescence was also used as a probe for the urea denaturation experiments, once
again revealing information about the local environment of the N-terminal tryptophan cluster.
The tryptophan residues of cdAE1 first become dequenched, and then quenched at higher urea
concentrations. During denaturation, the tryptophan residues in cdAE1 move away from
quenching residues and become dequenched, but as the protein continues to denature, the
tryptophan residues become further exposed to the more polar, aqueous environment, which in
return causes a decrease in fluorescence. Under native conditions, the intrinsic fluorescence of
cdkAE1 was high, indicating that the tryptophan residues were dequenched. The tryptophans in
151
this construct would already be more apart from quenching residues so that denaturation would
only cause the tryptophans to become even more exposed to the solvent, hence the decrease in
fluorescence that was observed. At 6 M urea, the fluorescence intensity of both cdAE1 and
cdkAE1 converge and plateau, indicating a similarly unfolded state.
In conclusion, the biophysical techniques employed in this study were useful in structurally
characterizing cdkAE1. It was shown through SE experiments and CD experiments that cdkAE1
retains a dimeric structure and has similar secondary structure content compared to the erythroid
isoform. The analyses made by SV experiments, DSC, and intrinsic tryptophan fluorescence
experiments established that removal of the acidic N-terminal extension (residues 1-54 only) had
little effect on the folding and conformational stability of cdAE1, whereas deletion of the β-
strand (residues 55-65), as in cdkAE1, produced a less thermally stable protein with a more open
structure.
The crystal structure of the cytoplasmic domain of AE1 has been elucidated for the erythroid
isoform in acidic conditions. In addition, structural features of specific regions of cdAE1 at
neutral pH, as investigated by SDSL in combination with EPR and DEER spectroscopies, were
in agreement with those from the crystal structure (Zhou et al. 2005a). We plan to further
characterize the structure of cdkAE1 by X-ray crystallography or NMR analysis. Based on the
results presented in this study, NMR would perhaps be the technique most amenable to solving
the structure of cdkAE1, which seems to adopt a less compact and a more open structure. NMR
would also be an ideal technique to study the protein dynamics of cdkAE1 and cdAE1.
The deletion of residues 1-65 has two effects on the cytoplasmic domain: 1) the loss of the
acidic N-terminal extension that is involved in protein binding, and 2) the loss of the β-strand,
which causes changes to the structure and thermal stability of the domain. Both of these effects
could account for the inability of kAE1 to bind cytoskeletal proteins and glycolytic enzymes.
The changes in the structure of cdkAE1 may also allow kAE1 to interact with different proteins
than AE1. In order to fully understand the function and regulation of kAE1 it would be important
to identify its interacting protein partners in kidney cells.