Date post: | 11-Nov-2023 |
Category: |
Documents |
Upload: | independent |
View: | 0 times |
Download: | 0 times |
www.elsevier.com/locate/yexcr
Experimental Cell Research 297 (2004) 11–26
Keratin 8/18 breakdown and reorganization during apoptosis
Bert Schutte,a,* Mieke Henfling,a Wendy Kolgen,a Maartje Bouman,a Stephan Meex,a
Mathie P.G. Leers,b Marius Nap,b Viveka Bjorklund,c Peter Bjorklund,c
Bertil Bjorklund,d E. Birgitte Lane,e M. Bishr Omary,f
Hans Jornvall,d and Frans C.S. Ramaekersa
aDepartment of Molecular Cell Biology (Box 17), Research Institute Growth and Development (GROW), University of Maastricht, The NetherlandsbDepartment of Pathology, Atrium Medical Centre, Heerlen, The Netherlands
cPeviva AB, Stromkarlsvagen 82, SE-167 62 Bromma, SwedendDepartment of Medical Biochemistry and Biophysics, Karolinska Institutet, SE-171 77, Stockholm, SwedeneCancer Research UK, Cell Structure Research Group, School of Life Sciences, University of Dundee, UK
fPalo Alto VA Medical Center and Stanford University School of Medicine, Stanford University, Palo Alto, CA 94304, USA
Received 28 October 2003, revised version received 19 February 2004
Available online 27 March 2004
Abstract
Monoclonal antibodies that specifically recognize caspase cleaved K18 fragments or specific (phospho)epitopes on intact K8 and K18
were used for a detailed investigation of the temporal and causal relationship of proteolysis and phosphorylation in the collapse of the keratin
cytoskeleton during apoptosis. Caspases involved in the specific proteolysis of keratins were analyzed biochemically using recombinant
caspases and specific caspase inhibitors. Finally, the fate of the keratin aggregates was analyzed using the M30-ApoptoSensek Elisa kit to
measure shedding of caspase cleaved fragments into the supernatant of apoptotic cell cultures. From our studies, we conclude that C-terminal
K18 cleavage at the 393DALD/S site is an early event during apoptosis for which caspase 9 is responsible, both directly and indirectly by
activating downstream caspases 3 and 7. Cleavage of the L1-2 linker region of the central a-helical rod domain is responsible for the final
collapse of the keratin scaffold into large aggregates. Phosphorylation facilitates formation of these aggregates, but is not crucial. K8 and K18
remain associated in heteropolymeric aggregates during apoptosis. At later stages of the apoptotic process, that is, when the integrity of the
cytoplasmic membrane becomes compromised, keratin aggregates are shed from the cells.
D 2004 Elsevier Inc. All rights reserved.
Keywords: M30-CytoDeath; Caspase; Keratin; Phosphorylation; Cytoskeleton
Introduction helical NH -terminal (head) and COOH terminal (tail)
Keratins, the epithelium-specific intermediate filament
proteins, compose of 20 cytoplasmic members in man,
excluding the trichocytic keratins [1]. They are subdivided
into type I and type II keratins [2], of which non-covalent
1:1 heteropolymers are expressed in a cell type-specific
manner. For example, K8 and 18 are co-expressed and form
heterodimers in glandular epithelia. Intermediate filament
proteins contain a central a-helical rod domain and non-a-
0014-4827/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.yexcr.2004.02.019
* Corresponding author. Department of Molecular Cell Biology (Box
17), Maastricht University, P.O. Box 616, 6200 MD Maastricht, The
Netherlands. Fax: +31-43-3884151.
E-mail address: [email protected] (B. Schutte).
2
domains. The head and tail domains contain the sites for
several posttranslational modifications, including phosphor-
ylation and glycosylation [3].
Keratins are highly dynamic and become reorganized
during various cellular events such as differentiation, mito-
sis, and apoptosis. Although many of their functions remain
to be established, it is evident that keratins provide structural
support to the cell and help cells to cope with stress [4,5].
Periodic filament remodeling is essential to keratin function.
Potential mechanisms involved in filament reorganization
include phosphorylation [3], proteolysis [6–8], and interac-
tion with non-keratin proteins [9].
During apoptosis, a cell undergoes dramatic changes in
morphology due to a complete reorganization of its cyto-
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–2612
plasmic and nuclear skeleton [10,11]. These alterations
probably facilitate the rapid, but securely orchestrated break-
down of the cell into apoptotic bodies, ensuring the mainte-
nance of an intact cellular membrane and efficient clearance
by phagocytes [12–14].
Many components of the cytoplasmic and nuclear cyto-
skeleton are targeted for proteolysis, including members of
the intermediate filament proteins, and proteins associated
with the cytoskeleton such as plectin [15], gelsolin [16],
Gas2 and h-catenin [17–20]. Recently, it was shown that
procaspases 3 and 9 are specifically targeted to the K8/18
intermediate filament network by the death effector domain
containing DNA binding protein (DEDD) [21,22]. Interfer-
ence with this apoptosis-dependent targeted breakdown
results in severe impairment of the cytoplasmic and nuclear
cytoskeleton reorganization and accompanying processes
such as chromatin condensation [11]. Aside from proteoly-
sis, targeted phosphorylation of cytoplasmic and nuclear
cytoskeleton components also potentiates their disassembly
during cell death [23,24].
A key feature during the process of apoptosis is the
activation of the proteolytic caspase cascade that is respon-
sible for the stepwise cleavage of many critical cellular
proteins. These cleavage processes result in activation of
enzymes including kinases and procaspases, proteolytic
inactivation of enzymes, such as poly-ADP-ribose-polymer-
ase (PARP), or degradation of structural proteins including
intermediate filaments. Many intermediate filament proteins
contain a caspase consensus site in the conserved L1-2
linker region of the central a-helical rod domain [25]. For
instance, nuclear lamins and type I keratins possess a
VEVD/A or VEID/A site that is targeted by caspase 6. In
contrast, type II keratins are apparently resistant to proteol-
ysis by caspases and remain associated with fragments of
their partner keratins during apoptosis [6]. A second caspase
cleavage site was identified in the COOH terminal (tail)
domain of K18 [7], which has been further characterized as393DALD/S. Cleavage at this site generates a neo-epitope
that is specifically recognized by the M30 CytoDeath
monoclonal antibody [8]. It has been suggested that caspase
cleavage of the K18 COOH terminal domain is an early
event in the apoptotic cascade, preceding loss of membrane
asymmetry, DNA fragmentation as determined by DNA
nick end labeling, and cleavage of the L1-2 linker region
of the central a-helical rod domain [8].
During apoptosis, K8/18 intermediate filaments reorga-
nize into granular structures enriched for K18 that has been
phosphorylated on serine 52. In fact most of the known K8/
18 phosphorylation sites are phosphorylated during apopto-
sis [6,26]. Whether hyperphosphorylation is an early or a late
event during apoptosis is still unclear. Some authors suggest
that hyperphosphorylation of keratins may represent an early
physiologic stress marker for simple epithelia [27,28]. Al-
ternatively, keratins might be phosphorylated during the
execution phase of the apoptotic process, facilitating the
rapid collapse of the cytoskeletal architecture.
The proteolytic fragments of human type I keratins are
stable [6] and persist as large aggregates in the apoptotic
bodies [8,10]. Indeed, it has been suggested that these
keratin aggregates are shed from the cells and escape the
clearance by phagocytes, since keratins can be detected in
sera of cancer patients [29–33].
In this study, a detailed investigation of temporal and
mechanistic aspects of keratin breakdown and reorganiza-
tion during apoptosis has been undertaken. The fate of the
cleaved keratin fragments was studied in relation to each
other and to known indicators of the apoptotic process using
monoclonal antibodies directed against different epitopes on
K8 and K18. Caspases involved in the specific proteolysis
of keratins were analyzed biochemically using recombinant
caspases and specific caspase inhibitors. The role of hyper-
phosphorylation during pre-apoptotic reorganization of the
cytoskeleton was studied using phosphokeratin-specific
antibodies and specific kinase inhibitors. Finally, the fate
of the keratin aggregates was analyzed using the M30-
ApoptoSensek Elisa kit to measure shedding of caspase
cleaved fragments into the supernatants of apoptotic cell
cultures. From our studies, we conclude that C-terminal
K18 cleavage is an early event during apoptosis for which
caspase 9 is responsible directly and indirectly by activating
downstream caspases 3 and 7. Cleavage of the L1-2 linker
region of the central a-helical rod domain is responsible for
the final collapse of the keratin scaffold into large aggre-
gates. Phosphorylation facilitates formation of these aggre-
gates, but is not crucial. K18 and K8 remain associated in
heteropolymeric aggregates during apoptosis. At later stages
of the apoptotic process, that is, when the integrity of the
cytoplasmic membrane becomes compromised, keratin
aggregates are shed into the culture medium.
Materials and methods
Cell lines
The human squamous cell lung cancer cell line MR65
was cultured in Eagle’s modified minimal essential medi-
um (Cat. no. 12565-024; GIBCO, Paisley, Scotland, Unit-
ed Kingdom), supplemented with 1% non-essential amino
acids (GIBCO), 1% HEPES (GIBCO), 2 mM L-glutamine
(Cat. no. 22942; Serva, Heidelberg, Germany), 10% heat-
inactivated newborn calf serum (Cat. no. 021-6010M;
GIBCO) and 50 Ag/ml gentamycin (AUV, Cuyck, The
Netherlands). In some experiments, the cells were cultured
on eight-well glass slides (Nutacon, Schiphol, The Nether-
lands) pre-coated with freshly isolated rat-tail collagen I
(10 mg/ml).
Induction of apoptosis
Apoptosis was induced by incubation of cell cultures for
various periods of time with 50 AM roscovitine (final con-
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–26 13
centration of DMSO was 2.5%), a kind gift from Dr. Laurent
Meyer (Station Biologique, CNRS, Roscoff, France) or with
50 AM etoposide (final concentration of DMSO was 0.3%).
Cell harvesting and fixation
Cells growing in monolayers were harvested quantita-
tively by standard trypsinization. The cells were rinsed
once in phosphate-buffered saline (PBS) before fixation
for 5 min in absolute methanol at �20jC. For double
staining with M30 CytoDeath monoclonal antibody and
CMXRos (Molecular Probes, Leiden, The Netherlands),
or anti-active caspase 3, cells were fixed in 4% parafor-
maldehyde for 15 min at RT, followed by permeabiliza-
tion for 10 min at room temperature (RT) in PBS
containing 0.005% SDS.
Kinase inhibitors
For inhibition of keratin hyperphosphorylation, stauro-
sporine (Sigma) was added to 100 nM final concentration
0.5 h before induction of apoptosis in exponentially growing
MR65 cells.
Caspase inhibitors
The cell-permeable caspase 3 inhibitor (Ac-DEVD-CHO)
and z-VAD.fmk were purchased from Biomol (Sanvertech,
Heerhugowaard, The Netherlands). The cell-permeable cas-
pase 9 inhibitor (Ac-LEHD-CHO) was purchased from
Calbiochem. Caspase 6 inhibitor (z-VEID-fmk) was obtained
from Alexis (Kordia, Leiden, The Netherlands). Protease
inhibitors were dissolved in DMSO and added to the cell
Fig. 1. Schematic representation of the K8/18 dimer. The different head, tail,
interconnecting solid lines. Phosphorylation sites, caspase cleavage sites, as well as
this study are indicated.
cultures at the following concentrations unless stated differ-
ently: Ac-DEVD-CHO or Ac-LEHD, 20 AM; z-VEID-fmk or
z-VAD-fmk, 100 AM.
Antibodies
The following primary antibodies were used: (1)
Monoclonal antibody IIB5 (diluted 1:50) to bromodesox-
yuridine (BrdU) from Euro-Diagnostics BV (Arnhem, The
Netherlands) as a negative control; (2) rabbit antiserum
directed against active caspase 3, a gift from IDUN
Pharmaceuticals (San Diego, CA); (3) monoclonal anti-
caspase 3 from BD Transduction Laboratories Cat. No.
610322); (4) monoclonal anti-cytochrome C antibody from
BD Pharmingen Cat. No. 556432 (San Diego, CA); (5)
monoclonal antibody RCK106 directed against an epitope
in the NH2-terminus of K18 (see also Fig. 1). The
antibody is available from MuBio Products BV, Maas-
tricht, The Netherlands; (6) monoclonal antibody LE61,
recognizing a shared epitope on K8/18 [34]. (7) Polyclon-
al rabbit antiserum CA1, raised against the carboxy-
terminal peptide of K18; (8) monoclonal antibody M30
CytoDeath (diluted 1:40, Peviva AB, Bromma, Sweden)
recognizing the neo-epitope formed by caspase cleavage at
the 393DALD/S site of K18. Isotyping of the monoclonal
antibody M30 was performed with the mouse monoclonal
antibody isotyping kit (GIBCO BRL, Cat. no. 19663020,
Cell Biology Products). The isotype of the M30 mono-
clonal antibody was IgG2b; (9) monoclonal antibodies
RCK102 and M20 recognizing K8 (MuBio Products
BV) [35]; (10) the phosphokeratin antibodies mAb LJ4
(K8 pSer73), mAb 5B3 (K8 pSer431), mAb IB4 (K18
pSer33) [28]; (11) monoclonal antibody 41CC4 directed
and rod domains are indicated by cylinders and linker region domain as
the location of the different epitopes of K8 and 18 of the antibodies used in
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–2614
against Lamin A/C was kindly provided by Dr. G Warren
(Heidelberg, Germany).
As secondary antibodies, R-Phycoerythrin (RPE)-conju-
gated F(ab’)2 fragment of affinity-purified goat anti-mouse
immunoglobulins (R0480, DAKO A/S, Glostrup, Den-
mark; diluted 1:10), fluorescein isothiocyanate (FITC)-
conjugated F(ab’)2 fragments of rabbit anti-mouse IgG
(F313; DAKO A/S; diluted 1:10), FITC or Texas red-
conjugated goat anti-rabbit immunoglobulins (Southern
Biotechnology Associates (SBA), Birmingham, USA; di-
luted 1:80), FITC- and Texas red-conjugated goat anti-
mouse IgG1 and goat anti-mouse IgG2b (SBA; diluted
1:80), FITC-conjugated streptavidin (F0422; DAKO A/S;
diluted 1:100) were used.
Annexin V labeling
To quantify the apoptotic frequency based on the expo-
sure of phosphatidylserine at the outer leaflet of the mem-
brane, annexin V biotin (APOPTEST-BIOTIN, Nexins Re-
search BV, Hoeven, The Netherlands) was added to cells.
Adherent cells were harvested by mechanical scraping with a
disposable cell scraper and stained and analyzed as described
previously [36].
Staining for mitochondrial membrane potential
MR65 cells were loaded with 500 nM CMXRos (Mo-
lecular Probes) at 37jC, 30 min before harvesting. As a
control, cells were loaded with CMXRos in the presence
of 75 mM of the uncoupling reagent carbonyl cyanide m-
chlorophenylhydrazon (CCCP, Sigma, The Netherlands).
In vitro keratin cleavage
Cytoskeletal keratin extracts were prepared by resuspend-
ing approximately 5 � 107 MR65 cells for 10 min in 10 ml
cold RSB/PMSF/TritonX100 buffer (1.5 mMMgCl2, 10mM
Tris–HCl, 0.5 mM PMSF, 0.5% Triton X100, pH 7.4). The
suspension was centrifuged for 5 min at 1500� g at 4jC and
the pellet was resuspended in 200 Al DNAse/RNAse mix (1.5
mM MgCl2, 110 mM NaCl, 10 mM Tris–HCl pH 7.4 to
which 1.1 mg/ml DNAse I, 50 ng/ml RNAse, and 0.5 mM
PMSF have been added). The resuspended pellet was incu-
bated for 20min at RT. Five milliliters RSB/PMSFwas added
to the suspension and a pellet fraction was recovered after
centrifugation for 5 min at 1500 � g at 4jC. The pellet wasrinsed once in 1 ml RSB/PMSF buffer, and finally resus-
pended in 500 Al caspase working buffer (20 mMHepes, 100
mMNaCl, 0.1%CHAPS, 10mMdithiothreitol (DTT), 1 mM
EDTA, pH 7.4), and stored at �70jC.For in vitro cleavage studies, recombinant caspases
were added to 40 Al keratin extract, diluted four times
in caspase working buffer. The final concentration of the
recombinant caspases was 0.8 Ag/ml or 1 unit in the case
of caspase 9. Recombinant caspases 3, 6, 7, and 8 were
kind gifts from Dr. G. Salvesen (Burnham Institute, San
Diego, USA). Recombinant caspase 9 was purchased
from BioVision Inc., Mountain View, CA.
Immunocytochemistry on glass slides
Apoptosis was induced by replacing the culture medi-
um with fresh medium to which etoposide or roscovitine
was added. At various time intervals after induction of
apoptosis, the cells were washed twice with PBS and
fixed in methanol (�20jC). The cells were rinsed with
PBS and incubated with the primary antibody for 1 h in a
humidified chamber at RT. The cells were then rinsed
three times with PBS/BSA (1 mg/ml BSA). Antibody
binding was visualized by incubation for 1 h at RT in the
dark with an appropriate FITC-labeled secondary anti-
body. The cells were counterstained with 5 Ag/ml propi-
dium iodide (PI, Sigma) in PBS containing RNAse (1
mg/ml; Serva) for 7 min and mounted in glycerol/
DABCO (9:1 glycerol, 1 part 0.2 M Tris–HCl pH 8.0,
2% DABCO pH 8.0, 0.02% NaN3). In double staining
experiments, the cells were incubated simultaneously with
the two primary antibodies, either combinations of anti-
mouse and anti-rabbit Ig or monoclonal antibodies of
different isotype, which were detected by anti-isotype-
specific secondary antibodies. Antibody binding was vi-
sualized by incubating the cells with FITC- and TxRed-
conjugated antibodies simultaneously. Annexin V biotin
was added to the cells before fixation of the cells.
Annexin V binding was visualized using FITC-labeled
streptavidin. The cells were counterstained with DAPI (5
Ag/ml; Sigma, St. Louis, MO) for 7 min and mounted in
glycerol/DABCO (9:1 glycerol, 1 part 0.2 M Tris–HCl
pH 8.0, 2% DABCO pH 8.0, 0.02% NaN3). The glass
slides were covered and sealed before examination by
confocal scanning laser microscopy. As a negative con-
trol, mouse anti-BrdU (clone IIB5) was used as primary
antibody.
Immunocytochemistry for flow cytometric analysis
After harvesting and fixation, the cells were subse-
quently rinsed twice with PBS. The resuspended pellet of
approximately 106 cells was incubated for 1 h in the dark
with the monoclonal antibody M30. Antibody binding
was visualized with FITC-conjugated Fab2 fragments of
anti-mouse IgG (DAKO F0313). After incubation for 1 h,
the cells were rinsed twice with PBS and finally resus-
pended in PBS, containing 20 Ag/ml propidium iodide
(Calbiochem) and 100 Ag/ml RNase (Serva). For double
labeling experiments with M30 and anti-keratin antibod-
ies, both primary antibodies were added to the cells
simultaneously. M30 immunoreactivity was visualized
by incubation with RPE-conjugated goat anti-mouse
IgG2b (SBA), and keratin immunoreactivity was visual-
ized using FITC-conjugated anti-mouse IgG1 (SBA).
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–26 15
Both secondary antibodies were added simultaneously
and the cells were incubated for 1 h at RT. The cells
were rinsed twice in PBS/BSA, finally resuspended in
PBS/PI/RNase, and kept on ice for at least 15 min before
flow cytometric analysis.
Measurement of caspase activity
MR65 cells were washed three times with PBS and
harvested in cold 50 mM Tris–HCl, pH 7.5, with a
disposable cell scraper. Subsequently cells (1 � 107/ml)
were sonicated for 5 s using a Soniprep 150 (MSE Scientific
Instruments, Beun-De Ronde) sonicator (amplitude 12 Am).
Cell lysates were centrifuged at 10,000 � g at 4jC in an
Eppendorf centrifuge and supernatants were either immedi-
ately assayed or frozen at �20jC. Caspase 3 activity was
measured in 50 Al aliquots of cell lysate, which is the
equivalent of 500,000 cells.
Fig. 2. Temporal order of K18 cleavage during apoptosis in MR65 cells. Apop
monoclonal antibody alone (a, b) or in combination with other known markers of
microscopy. (a) A shadow projection of a confocal stack of images of M30 reacti
(green) combined with a fragmented and condensed chromatin (red); (b) a shadow
cell. This early apoptotic cell is characterized by a fine filamentous M30 sta
fragmentation; (c) linear projection of a confocal stack of images of a triple labeled
counterstained with DAPI (blue). Note that PS exposure does not occur in the cell
cell, which shows a membranous annexin V staining pattern. (d) Double labeling
antibodies. Note that in one cell showing a filamentous M30 staining pattern har
pattern is observed in very few M30-positive (green) cells after loading of the cel
showing a typical mitochondrial labeling pattern with CMXRos were always devo
accumulation of CMXRos in the mitochondria was observed. (f) Double labeling ex
M30-reactive cells, including those with a filamentous type of staining, showed a
apoptotic cells always showed a discrete mitochondrial labeling with the anti-cyt
Fifty microliters of substrate mix (50 mM Tris/HCl,
pH7.5, containing 2.5 mM DTT and 40 AM Ac-DEVD-
AMC (Biomol), Ac-WEHD, Ac-IETD, AcVEID, Ac-
VDAD or Ac-LEHD (Alexis, Kordia)) was added to 50
Al of cell lysate. The substrate mixture was incubated for
60 min at 37jC and then diluted with 1.8 ml PBS.
Fluorescence was measured using a SPF-500C (SLM-
Aminco) spectrofluorometer.
Western blotting
Apoptosis was induced by replacement of the culture
mediumwith fresh medium to which roscovitine or etoposide
was added. At various time intervals after induction of
apoptosis, cells were harvested. Detached and adherent cells
were combined, washed with PBS and resuspended in lysis
buffer (62.5 mM Tris–HCl pH 6.8, 12.5% glycerol, 2%
Nonidet P40 (NP40), 2.5 mM phenylmethylsulphonyl fluo-
tosis was induced using roscovitine and cells were stained with the M30
apoptosis (c– f). Stained cells were analyzed using confocal scanning laser
vity in a typical late apoptotic cell. Note the granular M30 staining pattern
projection of a confocal stack of images of a M30-reactive early apoptotic
ining pattern without any obvious signs of chromatin condensation or
apoptotic cells with annexin V (green) and M30 (red) antibody. Nuclei are
showing a filamentous M30 staining pattern, in contrast to the late apoptotic
of early apoptotic cells with anti-active caspase 3 (green) and M30 (red)
dly any active caspase 3 is detected. (e) A discrete mitochondrial staining
ls with the mitochondrial-specific dye CMXRos (red); the majority of cells
id of M30 staining; in apoptotic cells showing a dot-like M30 reactivity no
periment with the M30 (red) and an anti-cytochrome c (green) antibody. All
diffuse distribution of cytochrome c in the cytoplasm. M30-negative, non-
ochrome c antibody.
Fig. 3. Quantification of the immunoreactivity of anti-active caspase 3
antibody in M30-positive MR65 cells. M30 staining patterns were scored
negative, filamentous or dot like. Anti-active caspase 3 immunoreactivity
was measured using the public domain software Image J in linear
projections of stacks of confocal images. Note that some cells with a
filamentous staining pattern showed background levels of anti-active
caspase 3.
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–2616
ride (PMSF), 1.25 mM EDTA, 12.5 Ag/ml leupeptin and 10
Ag/ml aprotinin) to a concentration of 2 � 107 cells/ml. As a
control, non-induced, exponentially growing MR65 cells
were rinsed with PBS, scraped, collected, and resuspended
in lysis buffer. The cell lysates were kept on ice for 30 min
before freezing at �20jC. Before gel electrophoresis, the
samples were supplemented with sample buffer (1:1), con-
taining 2.3% SDS, 62.5 mMTris–HCl pH 6.8, 10% glycerol,
5% h-mercaptoethanol, and 0.05% bromophenol blue. Sam-
ples were boiled for 5 min.
One-dimensional sodium dodecylsulphate polyacryl-
amide gel electrophoresis (SDS-PAGE) was performed
according to Laemmli [37] using gels containing 10%,
13%, or 15% polyacrylamide (Bio-Rad Laboratories, Her-
cules, CA) and 0.1% SDS (Merck, Darmstadt, Germany). As
molecular weight markers, a low molecular weight marker
kit (97, 66, 45, 31, 21, and 14 kDa; BioRad Laboratories)
was used. Gels were run on the Mini-Protean II system (Bio-
Rad Laboratories) for 45 min at 200 V. For Western blotting,
the method described by Towbin et al. [38] was used.
Detection was performed using enhanced chemilumines-
cence (ECL; Cat. no. RRN 2106; ECL-kit, Amersham Life
Science, Amersham, UK).
Confocal scanning laser microscopy
Cells grown on glass slides were analyzed using the
MRC600 confocal scanning laser microscope (Bio-Rad,
Hemel Hempstead, United Kingdom), equipped with an
air-cooled Argon-Krypton mixed-gas laser and mounted
onto an Axiophot microscope (Zeiss, Oberkochen, Ger-
many). The laser-scanning microscope was used in the
dual-parameter setup, according to manufacturer’s specifi-
cation, using dual wavelength excitation at 488 and 568 nm.
Emission spectra were separated by the standard set of
dichroic mirrors and barrier filters. All scans were recorded
in the Kalman filtering mode. For recording of DAPI
staining, mercury arc illumination was used. To this end,
the laser pathway was blocked and the filters were removed.
Recordings were made, using the Kalman filtering mode.
Image deconvolution was performed using Huygens
software (Scientific Volume Imaging, The Netherlands)
and three-dimensional reconstruction using Imaris software
(Bitplane AG, Switzerland). Both software packages were
run on a Silicon Graphics workstation. For microscopic
quantification of anti-active caspase 3 immunofluorescence,
the public domain software Image J was used (Wayne
Rasband, National Institutes of Health, USA) on a linear
projection of a stack of confocal images.
Flow cytometry
For flow cytometric analysis, a FACSort (Becton Dick-
inson, Sunnyvale, CA) equipped with an Argon ion laser and
a diode laser was used. Excitation was performed at 488 nm,
and the emission filters used were 515–545 BP (green;
FITC), 572–588 BP (orange; PE), and 600 LP (red; PI). A
minimum of 10,000 cells per sample was analyzed. FITC and
PE signals were recorded as logarithmic amplified data, while
the PI signals were recorded as linear amplified data. Elec-
tronic compensation was used to eliminate any bleed-through
of fluorescence. For bivariate analysis, no compensation was
used. Data analysis was performed using the CellQuest 3.1
software (Becton Dickinson, San Jose, CA). As a standard
procedure for all analyses, data were gated on pulse-pro-
cessed PI signals to exclude doublets and larger aggregates.
Detection of shedded K18 fragments
Keratin fragments in culture supernatants was determined
using the M30-ApoptoSensek ELISA kit (Peviva AB). The
assay was performed according to the manufacturer’s
instructions.
Results
In a previous study, we showed that the M30 antibody
recognizes a neo-epitope at the C-terminus of K18 [8]. This
neo-epitope is generated during the apoptotic process by
caspase cleavage at 393DALD/S. After induction of apopto-
sis, the majority of immunoreactive cells show a dot-like
staining pattern with M30, as well as chromatin fragmenta-
tion (Fig. 2a). However, in a minority of the cells, all still
attached to the glass surface, a filamentous M30 staining
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–26 17
pattern was observed without obvious chromatin condensa-
tion or nuclear disintegration (Fig. 2b).
These staining patterns suggest that caspase cleavage
plays an important role in the reorganization of the keratin
cytoskeleton during apoptosis. Therefore, experiments were
designed to determine the temporal order of the different
caspase-mediated cleavage processes of keratins using
known hallmarks of apoptosis as key reference points.
Furthermore, we addressed the relationship between keratin
hyperphosphorylation and these cleavage processes during
apoptosis. Finally, we determined the fate of the keratin
Fig. 4. (a, b) Western blotting analysis of in vitro caspase cleaved cytoskeletal pre
RCK106 antibody (a) and M30 (b). Note that with RCK106 (a) a specific 30 kDa c
seen in caspase 6-treated samples. Caspases 3-, 7-, or 9-treated samples show a
untreated sample. The M30 antibody (b) shows a 45 kDa reactive only in caspases 3
immunoreactivity. When samples were simultaneously treated with caspases 3 and
control lysate, are observed. These cleavage products correspond to the K18 molec
(c) Western blotting analysis of an apoptotic cell lysate (lane 1) and a crude cyt
RCK106 and anti procaspase 3 antibodies. Note that the cytoskeletal preparation i
roscovitine in the presence or absence of different caspase inhibitors, probed with
and the caspase 6 inhibitor prevent the appearance of the 30-kDa RCK106-reacti
could not completely prevent the appearance of a 45-kDa M30-reactive fragm
Apoptosis in the presence of caspases 3 or 9 inhibitor resulted in the accumulatio
cleavage products and keratin aggregates generated during
apoptosis.
Temporal order of K18 cleavage events versus known
apoptosis hallmarks
C-terminal K18 cleavage precedes phosphatidylserine
exposure
We demonstrated earlier that C-terminal caspase cleav-
age of K18 at the 393DALD/S site precedes loss of mem-
brane phospholipid asymmetry [8]. Early after induction of
parations of MR65 cells treated with different caspases and probed with the
leavage product, identical to the product present in the control lysate, is only
n additional fragment with slightly lower Mr (arrowhead) than that of the
-, 7- and 9-treated samples. Untreated or caspase 6-treated samples show no
6, two characteristic cleavage products of 45 and 20 kDa, also found in the
ule lacking either the tail or the tail and the rod 2 domain of K18 (see Fig. 1).
oskeletal preparation (lane 2) of MR65 cells. The blots were probed with
s devoid of procaspase 3. (d, e) Western blotting of MR65 cells treated with
RCK106 (d) or the M30 antibody (e). The general caspase inhibitor z-VAD
ve cleavage product, in contrast to the caspases 3 and 9 inhibitors. z-VAD
ent. Caspase 6 inhibitor failed to inhibit this cleavage event completely.
n of the 20-kDa M30 reactive fragment.
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–2618
apoptosis the few M30-positive cells, which are observed
are all devoid of annexin-V binding. The M30 staining
pattern in these few cells is largely filamentous, indicating
the keratin network did not yet collapsed (Fig. 2c).
C-terminal K18 cleavage precedes activation of caspase 3
The temporal relationship between caspase 3 activation
and the C-terminal cleavage of K18 was studied by
double labeling of MR65 cells with M30 and an antise-
rum directed against the activated caspase 3. Early after
induction of apoptosis by roscovitine, many M30-positive
cells, showing a filamentous type of staining, were
virtually devoid of active caspase 3 (Fig. 2d). In M30-
positive cells showing an aggregated type of staining,
active caspase 3 was clearly present. Active caspase 3
immunoreactivity was quantified in M30-negative cells,
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–26 19
M30-positive cells with filamentous type of staining and
M30-positive cells with an dot-like type of staining. As
shown in Fig. 3, active caspase 3 immunoreactivity of
M30-positive cells with filamentous type of staining over-
lapped with that of M30-negative non-apoptotic cells.
These experiments show that a caspase upstream of
caspase 3 must be responsible for the initial C-terminal
K18 cleavage at the 393DALD/S site.
C-terminal K18 cleavage follows loss of mitochondrial
membrane potential and cytochrome c release
To compare the appearance of M30 immunoreactivity
with loss of mitochondrial membrane potential, apoptotic
cells were harvested at various time intervals after induction
with roscovitine. Before harvesting, roscovitine-treated cells
were loaded for 30 min with the membrane potential-
sensitive probe CMXRos. The majority of M30-reactive
cells showed a loss of discrete mitochondrial accumulation
of CMXRos fluorescence. Very few cells with a filamentous
type of M30-staining pattern showed discrete labeling of
mitochondria (Fig. 2e). Similarly, when cells were double
labeled for M30 and cytochrome c, all M30-positive cells
showed a diffuse cytoplasmic staining for cytochrome c. No
M30-positive cells were found with cytochrome c discretely
distributed over the mitochondria (Fig. 2f). From these
experiments, it is clear that C-terminal cleavage of K18
occurs (immediately) after the mitochondria loose their
membrane potential and release cytochrome c into the
cytoplasm.
The different caspases involved in K18 cleavage
To identify the different caspases that are potentially
responsible for the generation of the M30 epitope, cytoskel-
eton preparations were incubated with recombinant initiator
caspases 8, 9 and executioner caspases 3, 6, and 7. Western
blots were probed with the anti-K18 antibody RCK106 or
Fig. 5. (a–g) Keratin phosphorylation and cleavage. Non-treated MR65 cells (a–c
LJ4 (K8 pSer73, c) 5B3 (K8 pSer431). All cells at all stages were phosphorylated
Ser73 to a much lower extent (b), but massive hyperphosphorylation of K8 on Se
the latter two antibodies. Treatment of MR65 cells with roscovitine (d–g) resulted
seen by comparison of figures d (control) and e (roscovitine treated). To analyze
apoptosis was induced in MR65 cells using roscovitine in the presence (f) or absen
labeled with M30 (red) and LJ4 (green), display a filamentous M30 staining pat
apoptotic cells that show K18 aggregates (f). Induction of apoptosis in the presenc
prevent the collapse of the K8/18 network into aggregates (g). (h–n) Dependenc
labeling with M30 (green) and propidium iodide (red) in roscovitine treated MR65
caspase 6 resulted in an almost complete absence of keratin aggregates, but also ha
condensed but not fragmented chromatin. As shown in figure j, the C-terminal K18
these roscovitine and caspase 6 inhibitor-treated cells. Inhibition of caspase 6 activ
for lamin A/C (k, l) and lamin B2 (m, n). In the absence of caspase 6 inhibitor (k
cells, whereas a complete nuclear lamina in combination with a filamentous M30
observation is made for lamin A/C, except for the fact that in the absence of caspas
Fate of the different K8/18 cleavage products. (o) The C-terminal K18 fragment
cytoplasm, in contrast to the M30 reactive fragments (red) which accumulate as
recognized by the RCK106 antibody (red) in relation to the diffuse distribution o
RCK106 (green) and M30 (red) reactivity colocalize in the large cytoplasmic gran
(red) reactive fragments, indicating that caspase cleaved K18 still forms stable di
reactivity (green), loss of M30-reactive K8/18 granules (red) is observed.
M30. As shown in Fig. 4a, K18 is cleaved in the L1-2 linker
region only by caspase 6, resulting in a cleavage fragment of
approximately 30 kDa. Fig. 4b shows that caspases 3, 7, and 9
are able to generate an M30-reactive fragment of approxi-
mately 45 kDa, corresponding to K18 with the C-terminus
removed. When extracts were treated with combinations of
caspases 3 and 6, an additional cleavage product of approx-
imately 20 kDa was observed, corresponding to the coil2
region, resulting from cleavage at both the L1-2 linker region
and at the C-terminus (Fig. 4b, lane 7). When blots were
probed with RCK106, a cleavage fragment with slightly
lower Mr than intact K18 was detected in caspases 3-, 7-,
or 9-treated samples (arrowhead in Fig. 4a). Unlike the other
caspases, recombinant caspase 8 was not able to cleave K18
(data not shown). To exclude the possibility that not recom-
binant caspase 9 but residual procaspase 3 activity in the
cytoskeleton preparation is indirectly responsible for C-
terminal K18 cleavage, blots of cytoskeletal preparations
were probed with a monoclonal antibody against procaspase
3. Fig. 4c shows that cytoskeleton preparations were devoid
of procaspase 3, while the protein could be detected in the
total cell extract.
To determine which caspases are responsible for cleavage
at the DALD site in vivo, cells were treated with roscovitine
in the presence or absence of various caspase inhibitors. In the
presence of caspase 6 inhibitor or the general caspase
inhibitor zVAD-fmk, cleavage at the L1-2 linker region was
inhibited (Fig. 4d). The caspases 3 and 9 inhibitors did
prevent cleavage of K18 at the 393DALD/S site to some
extent, but not completely (Fig. 4e). Also, cleavage at the393DALD/S site was only partially inhibited in the presence
of zVAD-fmk, although cleavage of the L1-2 linker region
was completely inhibited (Figs. 4d and e).
These experiments show that caspases 3, 7, and 9 are
able to cleave K18 at the 393DALD/S site and that caspase 6
is responsible for cleavage at the L1-2 linker region.
) stained with phosphokeratin-specific antibodies: (a) IB4 (K18 pSer33, (b)
at K18 Ser33 (a), whereas mitotic cells showed phosphorylation of K8 on
r431 (c). Spontaneously occurring apoptotic cells were clearly positive with
in rapid hyperphosphorylation of K8 Ser431 in virtually all cells as can be
the relationship between phosphorylation and caspase cleavage of keratins,
ce (g) of the kinase inhibitor staurosporine. Early apoptotic cells (f), double
tern and show hardly any phosphorylation of K8 Ser73, in contrast to late
e of staurosporine completely inhibited phosphorylation of K8, but did not
e of K8/18 network collapse on cleavage and/or phosphorylation. Double
cells in the absence (h) or presence (i) of caspase 6 inhibitor. Inhibition of
d a profound effect on nuclear morphology. Affected nuclei showed heavily
fragment was diffusely distributed in the cytoplasm and the nucleoplasm of
ity (l, n) also resulted in the absence of nuclear lamina breakdown as shown
, m), lamin B2 (green) is completely absent in apoptotic M30-reactive (red)
staining is observed in the presence of caspase 6 inhibitor (l, n). A similar
e 6 inhibition some residual lamin A/C staining can be observed (m). (o–s)
, recognized by the CA-1 antiserum (green) is diffusely distributed in the
large granules; (p) aggregated distribution of N-terminal fragments of K18
f C-terminal K18 fragments recognized by the CA1 antiserum (green); (q)
ules of apoptotic MR65 cells. (r) Colocalization of LE61 (green) and M30
mers with K8; (s) in some late apoptotic cells, identified by anti-caspase 3
Table 1
K8/18 phosphorylation of apoptotic MR65 cells
Antibody Staining intensity Roscovitine Roscovitine +
Staurosporine
K18 pSer 33, n.s. weakly positive 18 19
strongly positive 30 31
K8 pSer431, n.s. weakly positive 17 21
strongly positive 27 21
K8 pSer73, P < 0.01 weakly positive 8 32
strongly positive 35 13
The number of M30-positive cells showing (either weak or strong) staining
with the different phosphokeratin-specific antibodies were scored. Only
M30-positive cells showing an aggregated keratin staining pattern were
included.
l Cell Research 297 (2004) 11–26
Role of keratin phosphorylation in apoptosis
Having established the role of caspases in the reorga-
nization of the keratin cytoskeleton during apoptosis, we
embarked on a series of experiments to determine to what
extent (hyper)phosphorylation contributes to this reorga-
nization and to the final collapse of the cytoskeleton.
Some studies suggest that hyperphosphorylation might be
a very early physiological stress signal preceding prote-
olysis. However, since caspases activate certain protein
kinases, (hyper)phosphorylation of keratins during the
execution phase of apoptosis could facilitate the rapid
collapse of the cytoskeletal architecture. To investigate the
role of phosphorylation during apoptosis in more detail,
we used phosphokeratin-specific antibodies. First, we
determined the staining patterns of these antibodies in
exponentially growing cell cultures. Staining was carried
out 1 day after seeding of the cells, since we found that
trypsinization of cells causes rapid phosphorylation of
some serine residues (especially keratin pSer431, data
not shown). After 1 day, cells had recovered from the
stress induced by trypsinization as concluded from the
absence of pSer431 staining in the majority of the cells.
K18 is phosphorylated on pSer33 in all cells during
exponential growth (Fig. 5a), whereas phosphorylation
on pSer73 and pSer431 on K8 was absent in the majority
of interphase cells (Figs. 5b, c). Mitotic cells and some
cells undergoing spontaneous apoptosis showed phosphor-
ylation of K8 on pSer73 and pSer431.
Treatment of cells with roscovitine induced rapid phos-
phorylation of pSer431 in K8 in all cells (Figs. 5d, e),
whereas phosphorylation of pSer73 on K8 remained at
very low or undetectable levels. Similar results were
obtained using etoposide, a toposiomerase inhibitor and
inducer of apoptosis (data not shown). In apoptotic cells,
the keratin aggregates stained brightly for all phosphoepi-
tope-specific antisera tested. To test whether phosphoryla-
tion of pSer73 either preceded or resulted from caspase
activation, cells were double labeled with M30 and the
phosphokeratin-specific antibodies. As shown in Fig. 5f,
cells showing a filamentous staining pattern with M30
were often negative for LJ4, while apoptotic cells showing
the aggregated keratin phenotype were positive for this
antibody recognizing phosphorylation of Ser73.
Role of caspase cleavage and phosphorylation in keratin
collapse
Effect of kinase inhibition on keratin reorganization
When apoptosis was induced in the presence of the
kinase inhibitor staurosporine, we noticed a decrease in
pSer431-immunoreactivity in all interphase cells (Table
1). Most strikingly, a complete absence of K8 pSer73
phosphorylation was observed only in apoptotic cells.
Despite loss of phosphorylation, the aggregated pheno-
type persisted (Fig. 5g).
B. Schutte et al. / Experimenta20
Phosphorylation of pSer33 and pSer431 on K18 and
8, respectively, was not sensitive to staurosporine treat-
ment. Apoptosis was not inhibited but, more importantly,
neither the level of phosphorylation in both non-apopto-
tic and apoptotic cells nor the filamentous versus aggre-
gated staining pattern was affected after staurosporine
treatment. Apoptotic cells still showed the characteristic
aggregated keratin staining with both antibodies. In
contrast, phosphorylation of pSer73 on K8 was sensitive
to staurosporine treatment. However, a phosphorylation
of this serine residue seemed not to be responsible for
the collapse of the keratin network, since double labeling
with M30 showed apoptotic cells with an aggregate-like
staining pattern, devoid of K8 pSer73 immunoreactivity
(Fig. 5g).
Effect of caspase inhibition on keratin reorganization
We tested whether caspase cleavage of K18 is an
important trigger for the final collapse of the keratin
cytoskeleton. To this end, apoptosis was induced in the
presence of various caspase inhibitors. Immunoblots of
the cultures show that neither caspase 9 nor caspase 3/7
inhibitors could completely prevent the C-terminal cleav-
age of K18 (Figs. 4d, e). In contrast, caspase 6 inhibitor
was very effective in preventing the L1-2 linker region
cleavage. When treated cells were stained for M30, it
became evident that caspase 9, and caspase 3/7 inhibitors
could not prevent the appearance of the aggregated
phenotype. In contrast, cells that were induced to undergo
apoptosis in the presence of the caspase 6 inhibitor
showed a predominantly filamentous staining pattern
(Figs. 5h, i; Table 2). Furthermore, nuclear morphology
was also affected. Chromatin fragmentation was absent,
but the nuclei of M30-positive apoptotic cells showed
prominent perinuclear condensation. The C-terminal K18
fragment was also in the nucleoplasm in between the
condensed chromatin (Fig. 5j). Since nuclear lamins
belong to the intermediate filament family and also
contain a caspase 6 consensus site, we double-labeled
these cells with M30 and antibodies directed against
lamin A/C and lamin B1, respectively. As expected, in
caspase 6 inhibitor-treated cells, the nuclear lamina was
Fig. 6. Shedding of K8/18 by apoptotic MR65 cells. Cells were treated with
roscovitine and cell and culture supernatants samples were taken at various
periods of time. The increase in the number of M30-positive cells (solid
circle) precedes the increase in annexin V-positive/PI-negative apoptotic
cells (solid square), indicating that exposure of the M30 epitope precedes
loss of membrane phospholipid asymmetry. Concomitant with the decrease
of annexin V-positive and M30-positive cells, an increase in the number of
cells with a compromised plasma membrane (annexin V-positive/PI-
positive secondary necrotic cells, solid triangle) is observed. The
appearance of K18 fragments in the culture medium (solid diamonds),
measured in U/l, follows similar kinetics as those of the secondary necrotic
cells. Three separate experiments were performed showing similar results.
Data of a single representative experiment are shown.
Table 2
The keratin staining pattern of apoptotic (M30-positive) MR65 cells were
analyzed in the presence or absence of caspase 6 inhibitor
Antibody Phenotype Roscovitine Roscovitine +
caspase 6 inhibitor
M30, P < 0.001 filamentous 13 95
dot-like 5 87
Lamin A/C, P < 0.001 intact lamina 2 29
irregular lamina 34 20
loss of lamina 14 1
Lamin B2, P < 0.001 intact lamina 13 47
loss of lamina 37 3
The number of cells with either a filamentous or dot-like M30 staining was
scored. In double-labeled cells, the percentage of M30-positive cells with
different organizations of nuclear lamin A/C or B2 lamin were determined.
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–26 21
still intact in contrast to cells induced to undergo apoptosis
in the absence of the caspase 6 inhibitor (Figs. 5k–n;
Table 2).
Fate of the keratin cleavage fragments
Intracellular distribution of keratin cleavage fragments
K8 does not contain any of the identified caspase
cleavage sites. In contrast K18 is cleaved at two sites,
as shown in previous experiments [7,8,39]. K18 forms
stable dimers with K8. Therefore, we investigated the fate
of the K18 cleavage fragments and their association with
K8 during cytoskeletal reorganization. For this purpose,
antibodies against known epitopes on the K18 molecule
were used (see Fig. 1). Both in early and late apoptotic
cells, the cleaved C-terminal fragments were diffusely
distributed in the cytoplasm and nucleoplasm. At this
stage, caspase cleavage of the keratin cytoskeleton at
the 393DALD/S site was complete, since no filamen-
tous staining was seen with the CA1 antiserum. CA-1
immunoreactivity was never associated with the keratin
aggregates (Figs. 5o, p). These aggregates showed immu-
noreactivity with keratin antibodies that recognize both the
cleaved coil2 region and the N-terminal fragments of K18
(Fig. 5q). Immunostaining with K8 antibodies revealed
similar aggregates. Interestingly, at least part of the
cleaved coil2 region was probably still found to be
associated with K8, since these aggregates were still
immunoreactive with the LE61 antibody (Fig. 5r). At late
stages of apoptosis, when the chromatin was heavily
fragmented, immunoreactivity with all antibodies dimin-
ished and some late apoptotic cells were completely
negative (Fig. 5s).
Determination of K18 fragments in the culture supernatant
Since anti-keratin immunoreactivity is completely lost
in late apoptotic cells, in the absence of additional prote-
olysis, we investigated whether K18 fragments were shed
into the culture medium. To correlate loss of K18/M30
immunoreactivity in the cells with appearance of keratin
fragments in the culture supernatant, cells were harvested
by trypsinization and split into two aliquots. One aliquot
was labeled with annexin V-FITC and used for bivariate
annexin V/PI analysis. The second aliquot was fixed and
analyzed for M30 immunoreactivity. Culture supernatants
were analyzed for the presence of caspase cleaved K18
fragments using the ApotoSensek ELISA. As shown in
Fig. 6, an increase in M30-positive cells preceded an
increase in annexin V-positive/PI-negative cells. The max-
imum level of apoptosis was reached after approximately
8 h, at which time the number of annexin V-positive/PI-
positive (secondary necrotic) cells started to increase.
Parallel to this, increase in secondary necrotic cells, the
levels of M30-positive K18 cleavage fragments in the
culture supernatants increased dramatically.
To investigate whether K8 and other K18 fragments
were also lost from the dying cells, dual staining experi-
ments were performed using M30 and anti-K8 (M20),
anti-N terminal fragment-specific K18 (RCK106) and
with the antibody recognizing the shared epitope on K8
and K18 (LE61). As shown in Fig. 7, immunoreactivity
of both M30 and the anti-keratin antibodies decreased in
Fig. 7. Bivariate keratin/DNA flow cytometric analysis of control MR65 cells (a) and MR65 cells treated with roscovitine for 7 h (b–d) and double labeled
for M30 ( y-axis) and different anti-K8/18 antibodies (x-axis). a) Exponentially growing MR65 control cells double labeled with the M30 and M20 antibody.
(b–d) The decrease in M30 staining intensity is paralleled by a similar decrease in staining intensity in roscovitine treated cells for the K8-specific M20 (b),
the K18-specific RCK106 (c) and the K8/18-shared epitope LE61 (d) antibodies.
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–2622
parallel, indicating simultaneous loss of the different
types of K8/K18 fragments from the dying cells.
Discussion
During apoptosis, a cell undergoes dramatic changes in
morphology, partly due to a complete reorganization of its
cytoplasmic and nuclear cytoskeletal structures [10,11]. This
breakdown process involves the caspases and is well orches-
trated, since some of the cytoskeletal components retain a
function during apoptosis. For example, actin seems essential
for membrane integrity and seems to play a role in the
formation of apoptotic blebs [10]. Other components of the
cytoplasmic and nuclear cytoskeleton, such as the interme-
diate filaments, are targeted for rapid proteolysis during
apoptosis. These cleavage events are pivotal for the rapid
fragmentation of the cell into apoptotic bodies, ensuring rapid
clearance by phagocytes [12–14], as particle size seems to be
an important parameter in this process [40].
Monoclonal antibodies that specifically recognize either
caspase cleaved keratin fragments or specific phosphoe-
pitopes on K8 and K18 allow a detailed investigation of
the temporal and causal relationship of proteolysis and
phosphorylation in the collapse of the keratin cytoskeleton
during apoptosis. In this study, we have shown that K18
is initially cleaved in the carboxy-terminal domain at the393DALD/S site, resulting in the exposure of a neo-
epitope recognized by the M30-CytoDeath antibody. The
appearance of M30 reactivity is a very early event during
the caspase cascade, since the first M30-immunoreactive
cells exhibit hardly any detectable active caspase 3
immunoreactivity. On the other hand, M30 reactivity is
always observed in cells showing loss of cytochrome c
from the depolarized mitochondria. However, in a small
fraction of the apoptotic cells, polarized mitochondria
were found in M30-reactive cells. A likely explanation
for this phenomenon is the lag-time between CMXRos
loading and fixation of the cells. CMXRos accumulates in
polarized mitochondria and is retained by covalent bind-
ing with mitochondrial proteins through SH moieties.
When cells become apoptotic during labeling with
CMXRos, a discrete mitochondrial staining can be
expected.
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–26 23
In experiments using K8/K18-containing cytoskeleton
extracts, recombinant caspases 9, 3, and 7, but not caspase
6, could generate the M30 neo-epitope. Indirect activation
of procaspase 3 activity by apical caspases could be ruled
out by the fact that no cleavage of K18 was observed in the
presence of recombinant caspase 8, which is an activator of
caspase 3. Furthermore, no procaspase 3 could be detected
in our cytoskeleton preparations. In vitro assays using
caspase inhibitors were less conclusive. None of the indi-
vidual caspase inhibitors could prevent the C-terminal K18
cleavage completely. The most probable explanation for this
is the redundancy of the caspase cascade or the presence of
additional caspase(-like) activities, which are less sensitive
to inhibition by the general caspase inhibitor zVAD-fmk
[41–46].
Based on these observations, it is most likely that K18
is first cleaved at the 393DALD/S site by caspase 9 or
another apical caspase [45,46], and at a later stage by
caspases 3/7 in a kind of amplification loop. The recent
finding that procaspases 3 and 9 are specifically targeted
to the K8/18 intermediate filament network [21,22] sup-
ports this notion. As a consequence of caspase 3 activa-
tion, caspase 6 is also activated, resulting in a second
cleavage event in the L12 linker region of K18. This
event is responsible for the final collapse of the keratin
skeleton into large aggregates (see below) and coincides
with loss of intracellular contacts and detachment of cells
from their substrates. In a previous report, we could show
that in apoptotic cells that were detached from the culture
flasks a 20-kDa M30-reactive cleavage product could be
detected due to cleavage of K18 at the two caspase
consensus sites. On the other hand, in early apoptotic
cells, still attached to the culture flasks, an approximately
45-kDa M30 reactive fragment was seen. The biological
significance of the C-terminal cleavage step remains
obscure. The 393DALD/S site is only present in K18
and is conserved during evolution, since it is also present
in murine K18. Obviously, C-terminal cleavage does not
seem to affect the keratin organization to a great extend,
since cells still show a normal filamentous structure and
remain attached to their substrate. We noticed that early
during apoptosis, the C-terminal K18 peptide is distributed
diffusely throughout the cyto- and nucleoplasm. Whether or
not this C-terminal keratin fragment plays a role in changes
in nuclear architecture during apoptosis remains to be
established.
L12 linker region cleavage in K18
Cleavage of the L12 linker region in K18 has a dramatic
effect on the organization of the keratin cytoskeleton. In the
absence of this cleavage event, for example, after addition of
a caspase 6 inhibitor, the apoptotic cells tend to partially
round up but still remain attached to their substrates.
Breakdown of the keratin filamentous structure into large
granules is prevented, although large condensed filament
bundles are seen. Caspase 6 seems to be the caspase solely
responsible for this cleavage event, since only recombinant
caspase 6 was able to cleave keratin in this L12 linker
region, while caspases 3, 7, and 9 failed to do so. Secondly,
pretreatment of cells with a caspase 6 inhibitor was suffi-
cient to prevent this L12 linker cleavage in vitro. It should
be kept in mind that under all these circumstances, K8
remains associated with K18.
Phosphorylation of K8/K18
Although all keratin aggregates are highly phosphorylat-
ed on serine residues, this hyperphosphorylation seems not
to be important for the final collapse of the keratin network
as no obvious change in phosphorylation state of K18 Ser33
was noticed during the process of apoptosis. This serine
residue on K18 seems to be constantly phosphorylated. It
has been shown that K18 Ser33 phosphorylation is essential
for the association of K18 with 14-3-3 proteins, and plays a
role in keratin organization and distribution [47]. In our
study, we could not show a role for K18 Ser33 phosphor-
ylation during apoptosis.
In contrast, phosphorylation of either the head or the tail
domain of K8 was prominent during apoptosis. It has been
suggested that accumulation of hyperphosphorylated K8
might represent a physiologic stress marker for simple
epithelia [27]. Our results support the interpretation of K8
phosphorylation at Ser 431 as stress marker, since K8
Ser431 phosphorylation occurred in all cells immediately
after treatment with the CDK inhibitor roscovitine, although
only some of the cells became apoptotic. Our findings that
trypsinization and treatment with other apoptosis-inducing
compounds also showed a rapid phosphorylation of K8
Ser431 were in line with this reasoning. Ku and Omary
[26], however, showed that K8 phosphorylation at Ser431
increased dramatically upon stimulation of cells with epi-
dermal growth factor (EGF) or after mitotic arrest. These
authors also showed that the phosphorylation at this site was
likely to be mediated by mitogen-activated protein and cdc2
kinases. In our experiments, stress-induced phosphorylation
of K8 Ser431 is not likely to be mediated by cdc2 kinases,
since phosphorylation took place in the presence of the
general CDK inhibitor roscovitine. However, the stress
response was to some extent inhibited in the presence of
staurosporine, a kinase inhibitor with high affinity towards
PKCs. This finding is supported by previous reports that
showed a role for PKC in the phosphorylation of epithelial
keratins [48,49]. Many such kinases are specifically acti-
vated by caspase-dependent cleavage. They are generally
involved in the execution of apoptosis and contribute to the
cell death response. Among these kinases are MEKK1
[23,50–52], p21-activated kinase 2, and focal adhesion
kinase [53], as well as protein kinase C (PKC) family
members [24,54,55].
Phosphorylation at K8 Ser73 seems to be a late event
during apoptosis. We were able to completely inhibit
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–2624
phosphorylation at this site by induction of apoptosis in the
presence of staurosporine. These results indicate that phos-
phorylation of K8 Ser73 is initiated during the execution
phase of apoptosis, by kinases sensitive to staurosporine.
Inhibition of K8 Ser73 phosphorylation did not prevent the
formation of keratin aggregates, therefore, phosphorylation
at this site may facilitate but is not a prerequisite for
cytoskeletal collapse. In line with this reasoning, the obser-
vation of Ku et al. [6] showed that keratin hyperphosphor-
ylation does not render the keratins to be better substrates
for caspases, but can protect from caspase-mediated degra-
dation at the L12 site but not the 393DALD/S site [39].
From our experiments, we conclude that different kinases
are responsible for phosphorylation of keratins at the sep-
arate serine residues. Firstly, K18 Ser33 is phosphorylated
in all cells during exponential growth of cells, whereas K8
Ser 431 and K8 Ser73 are not. Secondly, treatment of cells
with the CDK inhibitor roscovitine resulted in phosphory-
lation of all cells at K8 Ser431, whereas phosphorylation of
K8 Ser73 was absent in the majority of the cells. This kinase
activity was to a small extent sensitive to inhibition by
staurosporine. Thirdly, phosphorylation during the execu-
tion phase of apoptosis was prominent only at K8 Ser73 and
was almost completely inhibited by staurosporine. This
latter kinase activity might be ascribed to PKCs since these
kinases are efficiently inhibited by staurosporine.
K8/18 fragment shedding
The cleaved keratin fragments are apparently stable [6]
since no signs of additional proteolysis is observed in the
immunoblotting studies. The cleaved fragments are retained
in the apoptotic cells as long as the cell membrane remains
intact. Loss of membrane phospholipid asymmetry does not
affect membrane permeability for these fragments, since
phosphatidylserine exposure proceeds shedding of K18
fragments by several hours. Once membrane integrity is
lost during secondary necrosis, K18 fragments are shed into
the culture medium. In patients, the clearance of apoptotic
cells seems not to be very efficient since keratin fragments
can be observed in sera of cancer patients. In support of this
finding, we observed M30-reactive material in small blood
vessels of different tissues using immunohistochemistry
[56]. The recently developed M30-ApoptoSensek Elisa
assay might therefore be useful in quantitating cell death
using sera of patients. M30 serum levels might be informa-
tive on growth rate of tumors, since a direct correlation
between proliferative activity and apoptosis was found in a
series of 300 colorectal tumors [57]. Furthermore, an M30-
ELISA assay could be used to monitor response to therapy,
since most cancer treatments result in rapid induction of
apoptosis in tumor cells.
In conclusion, we propose the following sequence of
events for the rapid reorganization of the K8/18 cytoskele-
ton during apoptosis. Firstly, a cell responds to an apopto-
genic trigger by rapid phosphorylation of K8 Ser431. This
might represent a stress response aimed at protecting the cell
from the insult. Secondly, once the apoptotic program is
initiated and procaspases 9 and 3 are specifically targeted to
the K8/18 intermediate filament network possibly through
DEDD, caspase 9 is activated, resulting in C-terminal
cleavage of K18 at the 393DALD/S site. During the execu-
tion phase of the apoptotic program executioner caspases,
such as caspases 3, 7 and 6, are activated, resulting in (i)
more efficient C-terminal cleavage of K18 by caspases 3
and 7, (ii) activation of kinases resulting in hyperphosphor-
ylation of K8 Ser73, and (iii) L12 linker cleavage of K18 by
caspase 6. Hyperphosphorylation of keratins probably
results in the formation of condensed fibers but is not a
prerequisite for the final collapse of the filament network. It
may even slow down degradation at the L12 site [39] to
allow coordination of the execution phase of apoptosis.
Cleavage in the L12 linker region of K18 by caspase 6
seems to be solely responsible for the formation of keratin
aggregates. In these aggregates, K8 and K18 fragments form
stable interactions. At later stages of the apoptotic process,
these aggregates can be shed into the surrounding medium
or serum.
Acknowledgments
The authors acknowledge the Dutch Science Foundation
(NWO; project nr. 901-28-134) for the financial support for
the Imaris/Huygens software running on a Silicon Graphics
workstation. Support by National Institute of Health grant
DK47918, Department of Veterans Affairs (to M.B.O.) and
Cancer Research, UK (to E.B.L) is also acknowledged.
References
[1] E. Fuchs, K. Weber, Intermediate filaments: structure, dynamics,
function, and disease, Annu. Rev. Biochem. 63 (1994) 345–382.
[2] R. Moll, W.W. Franke, D.L. Schiller, B. Geiger, R. Krepler, The
catalog of human cytokeratins: patterns of expression in normal ep-
ithelia, tumors and cultured cells, Cell 31 (1982) 11–24.
[3] M.B. Omary, N.O. Ku, J. Liao, D. Price, Keratin modifications and
solubility properties in epithelial cells and in vitro, Subcell. Biochem.
31 (1998) 105–140.
[4] E. Fuchs, D.W. Cleveland, A structural scaffolding of intermediate
filaments in health and disease, Science 279 (1998) 514–519.
[5] N.O. Ku, X. Zhou, D.M. Toivola, M.B. Omary, The cytoskeleton of
digestive epithelia in health and disease, Am. J. Physiol. 277 (1999)
G1108–G1137.
[6] N.O. Ku, J. Liao, M.B. Omary, Apoptosis generates stable fragments
of human type I keratins, J. Biol. Chem. 272 (1997) 33197–33203.
[7] C. Caulin, G.S. Salvesen, R.G. Oshima, Caspase cleavage of keratin
18 and reorganization of intermediate filaments during epithelial cell
apoptosis, J. Cell Biol. 138 (1997) 1379–1394.
[8] M.P. Leers, W. Kolgen, V. Bjorklund, T. Bergman, G. Tribbick, B.
Persson, P. Bjorklund, F.C. Ramaekers, B. Bjorklund, M. Nap, H.
Jornvall, B. Schutte, Immunocytochemical detection and mapping
of a cytokeratin 18 neo-epitope exposed during early apoptosis, J.
Pathol. 187 (1999) 567–572.
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–26 25
[9] J. Liao, M.B. Omary, 14-3-3 proteins associate with phosphorylated
simple epithelial keratins during cell cycle progression and act as a
solubility cofactor, J. Cell Biol. 133 (1996) 345–357.
[10] M. van Engeland, H.J. Kuijpers, F.C. Ramaekers, C.P. Reuteling-
sperger, B. Schutte, Plasma membrane alterations and cytoskeletal
changes in apoptosis, Exp. Cell Res. 235 (1997) 421–430.
[11] F.A. Oberhammer, K. Hochegger, G. Froschl, R. Tiefenbacher, M.
Pavelka, Chromatin condensation during apoptosis is accompanied
by degradation of lamin A + B, without enhanced activation of cdc2
kinase, J. Cell Biol. 126 (1994) 827–837.
[12] V.A. Fadok, P.M. Henson, Apoptosis: getting rid of the bodies, Curr.
Biol. 8 (1998) R693–R695.
[13] V.A. Fadok, Clearance: the last and often forgotten stage of apoptosis,
J. Mammary Gland Biol. Neoplasia 4 (1999) 203–211.
[14] J. Savill, V. Fadok, Corpse clearance defines the meaning of cell
death, Nature 407 (2000) 784–788.
[15] A.H. Stegh, H. Herrmann, S. Lampel, D. Weisenberger, K. Andra, M.
Seper, G. Wiche, P.H. Krammer, M.E. Peter, Identification of the
cytolinker plectin as a major early in vivo substrate for caspase 8 dur-
ing CD95- and tumor necrosis factor receptor-mediated apoptosis,
Mol. Cell. Biol. 20 (2000) 5665–5679.
[16] S. Kothakota, T. Azuma, C. Reinhard, A. Klippel, J. Tang, K. Chu,
T.J. McGarry, M.W. Kirschner, K. Koths, D.J. Kwiatkowski, L.T.
Williams, Caspase-3-generated fragment of gelsolin: effector of mor-
phological change in apoptosis, Science 278 (1997) 294–298.
[17] C. Brancolini, M. Benedetti, C. Schneider, Microfilament reorganiza-
tion during apoptosis: the role of Gas2, a possible substrate for ICE-
like proteases, EMBO J. 14 (1995) 5179–5190.
[18] C. Brancolini, A. Sgorbissa, C. Schneider, Proteolytic processing
of the adherens junctions components beta-catenin and gamma-
catenin/plakoglobin during apoptosis, Cell Death Differ. 5 (1998)
1042–1050.
[19] A. Sgorbissa, R. Benetti, S. Marzinotto, C. Schneider, C. Brancolini,
Caspase-3 and caspase-7 but not caspase-6 cleave Gas2 in vitro: impli-
cations for microfilament reorganization during apoptosis, J. Cell Sci.
112 (1999) 4475–4482.
[20] C. Brancolini, P. Edomi, S. Marzinotto, C. Schneider, Exposure at
the cell surface is required for gas3/PMP22 To regulate both cell
death and cell spreading: implication for the Charcot-Marie-Tooth
type 1A and Dejerine-Sottas diseases, Mol. Biol. Cell 11 (2000)
2901–2914.
[21] J.C. Lee, O. Schickling, A.H. Stegh, R.G. Oshima, D. Dinsdale, G.M.
Cohen, M.E. Peter, DEDD regulates degradation of intermediate fil-
aments during apoptosis, J. Cell Biol. 158 (2002) 1051–1066.
[22] D. Dinsdale, J.C. Lee, G. Dewson, G.M. Cohen, M.E. Peter, Inter-
mediate filaments control the intracellular distribution of caspases
during apoptosis, Am. J. Pathol. 164 (2004) 395–407.
[23] C. Widmann, P. Gerwins, N.L. Johnson, M.B. Jarpe, G.L. Johnson,
MEK kinase 1, a substrate for DEVD-directed caspases, is involved in
genotoxin-induced apoptosis, Mol. Cell. Biol. 18 (1998) 2416–2429.
[24] T. Cross, G. Griffiths, E. Deacon, R. Sallis, M. Gough, D. Watters,
J.M. Lord, PKC-delta is an apoptotic lamin kinase, Oncogene 19
(2000) 2331–2337.
[25] R.G. Oshima, Apoptosis and keratin intermediate filaments, Cell
Death Differ. 9 (2002) 486–492.
[26] N.O. Ku, M.B. Omary, Phosphorylation of human keratin 8 in
vivo at conserved head domain serine 23 and at epidermal growth
factor-stimulated tail domain serine 431, J. Biol. Chem. 272
(1997) 7556–7564.
[27] J. Liao, N.O. Ku, M.B. Omary, Stress, apoptosis, and mitosis induce
phosphorylation of human keratin 8 at Ser-73 in tissues and cultured
cells, J. Biol. Chem. 272 (1997) 17565–17573.
[28] C. Stumptner, M.B. Omary, P. Fickert, H. Denk, K. Zatloukal, Hepa-
tocyte cytokeratins are hyperphosphorylated at multiple sites in human
alcoholic hepatitis and in a Mallory body mouse model, Am. J. Pathol.
156 (2000) 77–90.
[29] K.N. Gaarenstroom, G.G. Kenter, J.M. Bonfrer, C.M. Korse, M.J. Van
de Vijver, G.J. Fleuren, J.B. Trimbos, Can initial serum cyfra 21-1,
SCC antigen, and TPA levels in squamous cell cervical cancer predict
lymph node metastases or prognosis? Gynecol. Oncol. 77 (2000)
164–170.
[30] J.D. Maass, A.M. Niemann, B.M. Lippert, S. Gottschlich, B.J. Folz,
J.A. Werner, CYFRA 8/18 in head and neck cancer, Anticancer Res.
19 (1999) 2699–2701.
[31] J. Pujol, J. Boher, J. Grenier, X. Quantin, Cyfra 21-1, neuron specific
enolase and prognosis of non-small cell lung cancer: prospective
study in 621 patients, Lung Cancer 31 (2001) 221–231.
[32] S.S. Sun, J.F. Hsieh, S.C. Tsai, Y.J. Ho, J.K. Lee, C.H. Kao, Cytoker-
atin fragment 19 and squamous cell carcinoma antigen for early pre-
diction of recurrence of squamous cell lung carcinoma, Am. J. Clin.
Oncol. 23 (2000) 241–243.
[33] J.M. Wolff, H. Borchers, B. Brehmer, A. Brauers, G. Jakse, Cytoker-
atin markers in patients with prostatic diseases, Anticancer Res. 19
(1999) 2649–2652.
[34] A. Waseem, E.B. Lane, D. Harrison, N. Waseem, A keratin antibody
recognizing a heterotypic complex: epitope mapping to complemen-
tary locations on both components of the complex, Exp. Cell Res. 223
(1996) 203–214.
[35] A. Waseem, U. Karsten, I.M. Leigh, P. Purkis, N.H. Waseem, E.B.
Lane, Conformational changes in the rod domain of human keratin
8 following heterotypic association with keratin 18 and its implication
for filament stability, Biochemistry 43 (2004) 1283–1295.
[36] M. van Engeland, F.C. Ramaekers, B. Schutte, C.P. Reuteling-
sperger, A novel assay to measure loss of plasma membrane asym-
metry during apoptosis of adherent cells in culture, Cytometry 24
(1996) 131–139.
[37] U.K. Laemmli, Cleavage of structural proteins during the assembly of
the head of bacteriophage T4, Nature 227 (1970) 680–685.
[38] H. Towbin, T. Staehelin, J. Gordon, Electrophoretic transfer of
proteins from polyacrylamide gels to nitrocellulose sheets: proce-
dure and some applications, Proc. Natl. Acad. Sci. U. S. A. 76
(1979) 4350–4354.
[39] N.O. Ku, M.B. Omary, Effect of mutation and phosphorylation of
type I keratins on their caspase-mediated degradation, J. Biol. Chem.
276 (2001) 26792–26798.
[40] I. Catelas, O.L. Huk, A. Petit, D.J. Zukor, R. Marchand, L. Yahia,
Flow cytometric analysis of macrophage response to ceramic and
polyethylene particles: effects of size, concentration, and composition,
J. Biomed. Mater. Res. 41 (1998) 600–607.
[41] M. Garcia-Calvo, E.P. Peterson, D.M. Rasper, J.P. Vaillancourt, R.
Zamboni, D.W. Nicholson, N.A. Thornberry, Purification and catalytic
properties of human caspase family members, Cell Death Differ. 6
(1999) 362–369.
[42] M. Garcia-Calvo, E.P. Peterson, B. Leiting, R. Ruel, D.W. Nicholson,
N.A. Thornberry, Inhibition of human caspases by peptide-based and
macromolecular inhibitors, J. Biol. Chem. 273 (1998) 32608–32613.
[43] N.A. Thornberry, T.A. Rano, E.P. Peterson, D.M. Rasper, T. Timkey,
M. Garcia-Calvo, V.M. Houtzager, P.A. Nordstrom, S. Roy, J.P.
Vaillancourt, K.T. Chapman, D.W. Nicholson, A combinatorial ap-
proach defines specificities of members of the caspase family and
granzyme B. Functional relationships established for key mediators
of apoptosis, J. Biol. Chem. 272 (1997) 17907–17911.
[44] P.G. Ekert, J. Silke, D.L. Vaux, Caspase inhibitors, Cell Death
Differ. 6 (1999) 1081–1086.
[45] Y. Guo, S.M. Srinivasula, A. Druilhe, T. Fernandes-Alnemri, E.S.
Alnemri, Caspase-2 induces apoptosis by releasing proapoptotic pro-
teins from mitochondria, J. Biol. Chem. 277 (2002) 13430–13437.
[46] J.D. Robertson, M. Enoksson, M. Suomela, B. Zhivotovsky, S. Orre-
nius, Caspase-2 acts upstream of mitochondria to promote cyto-
chrome c release during etoposide-induced apoptosis, J. Biol.
Chem. 277 (2002) 29803–29809.
[47] N.O. Ku, J. Liao, M.B. Omary, Phosphorylation of human keratin 18
serine 33 regulates binding to 14-3-3 proteins, EMBO J. 17 (1998)
1892–1906.
B. Schutte et al. / Experimental Cell Research 297 (2004) 11–2626
[48] C.F. Chou, M.B. Omary, Phorbol acetate enhances the phosphoryla-
tion of cytokeratins 8 and 18 in human colonic epithelial cells, FEBS
Lett. 282 (1991) 200–204.
[49] M.B. Omary, G.T. Baxter, C.F. Chou, C.L. Riopel, W.Y. Lin, B.
Strulovici, PKC epsilon-related kinase associates with and phosphor-
ylates cytokeratin 8 and 18, J. Cell Biol. 117 (1992) 583–593.
[50] M.H. Cardone, G.S. Salvesen, C. Widmann, G. Johnson, S.M. Frisch,
The regulation of anoikis: MEKK-1 activation requires cleavage by
caspases, Cell 90 (1997) 315–323.
[51] S.G. Shiah, S.E. Chuang, M.L. Kuo, Involvement of Asp-Glu-Val-
Asp-Directed, Caspase-Mediated Mitogen-Activated Protein Kinase
Kinase 1 Cleavage, c-Jun N-Terminal Kinase Activation, and Subse-
quent Bcl-2 Phosphorylation for Paclitaxel-Induced Apoptosis in HL-
60 Cells, Mol. Pharmacol. 59 (2001) 254–262.
[52] C. Widmann, N.L. Johnson, A.M. Gardner, R.J. Smith, G.L. Johnson,
Potentiation of apoptosis by low dose stress stimuli in cells expressing
activated MEK kinase 1, Oncogene 15 (1997) 2439–2447.
[53] S. Gibson, C. Widmann, G.L. Johnson, Differential involvement of
MEK kinase 1 (MEKK1) in the induction of apoptosis in response to
microtubule-targeted drugs versus DNA damaging agents, J. Biol.
Chem. 274 (1999) 10916–10922.
[54] H. Koriyama, Z. Kouchi, T. Umeda, T.C. Saido, T. Momoi, S. Ishiura,
K. Suzuki, Proteolytic activation of protein kinase C delta and epsilon
by caspase-3 in U937 cells during chemotherapeutic agent-induced
apoptosis, Cell Signalling 11 (1999) 831–838.
[55] L. Smith, L. Chen, M.E. Reyland, T.A. DeVries, R.V. Talanian, S.
Omura, J.B. Smith, Activation of atypical protein kinase C zeta by
caspase processing and degradation by the ubiquitin-proteasome sys-
tem, J. Biol. Chem. 275 (2000) 40620–40627.
[56] M.P.G. Leers, V. Bjorklund, B. Bjorklund, H. Jornvall, M. Nap, An
immunohistochemical study of the clearance of apoptotic cellular
fragments, Cell. Mol. Life Sci. 59 (2002) 1358–1365.
[57] J.D. Rupa, A.P. de Bruine, A.J. Gerbers, M.P. Leers, M. Nap,
A.G. Kessels, B. Schutte, J.W. Arends, Simultaneous detection of
apoptosis and proliferation in colorectal carcinoma by multipa-
rameter flow cytometry allows separation of high and low-turn-
over tumors with distinct clinical outcome, Cancer 97 (2003)
2404–2411.