MOVING PROTEINS INTO MEMBRANES AND …kbp-srmc.yolasite.com/resources/Chapter 16.pdf · targeting...

16

A live bovine endothelial cell stained to reveal

different intracellular compartments. The lacelike

membranes of the endoplasmic reticulum were

stained with a green fluorescent dye, and the

wormlike mitochondria were stained with an

orange fluorescent dye. [Molecular Probes, Inc.]

MOVING PROTEINSINTO MEMBRANESAND ORGANELLES

Atypical mammalian cell contains up to 10,000 differ-ent kinds of proteins; a yeast cell, about 5000. Thevast majority of these proteins are synthesized by cyto-

solic ribosomes, and many remain within the cytosol. How-ever, as many as half the different kinds of proteins producedin a typical cell are delivered to a particular cell membrane,an aqueous compartment other than the cytosol, or to thecell surface for secretion. For example, many hormone re-ceptor proteins and transporter proteins must be delivered tothe plasma membrane, some water-soluble enzymes such asRNA and DNA polymerases must be targeted to the nucleus,and components of the extracellular matrix as well aspolypeptide signaling molecules must be directed to the cellsurface for secretion from the cell. These and all the otherproteins produced by a cell must reach their correct locationsfor the cell to function properly.

The delivery of newly synthesized proteins to their propercellular destinations, usually referred to as protein targetingor protein sorting, encompasses two very different kinds ofprocesses. The first general process involves targeting of aprotein to the membrane of an intracellular organelle andcan occur either during or soon after synthesis of the pro-

tein by translation at the ribosome. For membrane proteins,targeting leads to insertion of the protein into the lipid bi-layer of the membrane, whereas for water-soluble proteins,targeting leads to translocation of the entire protein acrossthe membrane into the aqueous interior of the organelle.

657

O U T L I N E

16.1 Translocation of Secretory Proteins Across the ER Membrane

16.2 Insertion of Proteins into the ER Membrane

16.3 Protein Modifications, Folding, and QualityControl in the ER

16.4 Export of Bacterial Proteins

16.5 Sorting of Proteins to Mitochondria and Chloroplasts

16.6 Sorting of Peroxisomal Proteins

the secretory pathway. These proteins include not only solu-ble and membrane proteins that reside in the ER itself butalso proteins that are secreted from the cell, enzymes andother resident proteins in the lumen of the Golgi complexand lysosomes, and integral proteins in the membranes of

Proteins are sorted to the endoplasmic reticulum (ER), mito-chondria, chloroplasts, peroxisomes, and the nucleus by thisgeneral process (Figure 16-1).

A second general sorting process applies to proteins thatinitially are targeted to the ER membrane, thereby entering

658 CHAPTER 16 • Moving Proteins into Membranes and Organelles

Cytosol

RibosomesmRNA

Golgicomplex

Lysosome

SECRETORY PATHWAY

Rough endoplasmicreticulum

Plasmamembrane

mRNA

ER signalsequence

Intermembrane spaceOuter membrane

Innermembrane

Matrix

Mitochondrion

Chloroplast

Thylakoids

Outermembrane

Innermembrane

Stroma

Peroxisome

Matrix

Membrane

Nucleus

Nuclearpore

Inner nuclearmembrane

Outer nuclearmembrane

3

4a 4b

1

2 2

3

4

5

6

Cytosolicprotein

1

Targetingsequence

ME

DIA

C

ON

NE

CT

IO

NS

Ove

rvie

w A

nim

atio

n: P

rote

in S

ortin

g

(step ). Those proteins that contain no targeting sequence arereleased into the cytosol and remain there (step ). Proteinswith an organelle-specific targeting sequence (pink) first arereleased into the cytosol (step ) but then are imported intomitochondria, chloroplasts, peroxisomes, or the nucleus (steps

– ). Mitochondrial and chloroplast proteins typically passthrough the outer and inner membranes to enter the matrix orstromal space, respectively. Other proteins are sorted to othersubcompartments of these organelles by additional sorting steps.Nuclear proteins enter through visible nuclear pores byprocesses discussed in Chapter 12.

63

2

21▲ FIGURE 16-1 Overview of major protein-sorting

pathways in eukaryotic cells. All nuclear-encoded mRNAsare translated on cytosolic ribosomes. Left (secretorypathway): Ribosomes synthesizing nascent proteins in thesecretory pathway are directed to the rough endoplasmicreticulum (ER) by an ER signal sequence (pink; steps , ).After translation is completed on the ER, these proteins canmove via transport vesicles to the Golgi complex (step ).Further sorting delivers proteins either to the plasmamembrane or to lysosomes (steps , ). Right(nonsecretory pathways): Synthesis of proteins lacking an ER signal sequence is completed on free ribosomes

4b4a

3

21

2. What is the receptor for the signal sequence?

3. What is the structure of the translocation channel thatallows transfer of proteins across the membrane bilayer? Inparticular, is the channel so narrow that proteins can passthrough only in an unfolded state, or will it accommodatefolded protein domains?

4. What is the source of energy that drives unidirectionaltransfer across the membrane?

In the first part of the chapter, we cover targeting of pro-teins to the ER, including the post-translational modifica-tions that occur to proteins as they enter the secretorypathway. We then look at several mechanisms for exportingproteins from bacteria, some of which are similar to proteinsorting in eukaryotic cells. The last two sections describe tar-geting of proteins to mitochondria, chloroplasts, and perox-isomes. We cover the transport of proteins in and out of thenucleus through nuclear pores in Chapter 12 because nucleartransport is intimately related to post-transcriptional eventsand involves several nucleus-specific variations on the mech-anisms discussed in this chapter.

Translocation of SecretoryProteins Across the ER MembraneAll eukaryotic cells use essentially the same secretory path-way for synthesizing and sorting secreted proteins and solu-ble luminal proteins in the ER, Golgi, and lysosomes (seeFigure 16-1, left). For simplicity, we refer to these proteinscollectively as secretory proteins. Although all cells secrete avariety of proteins (e.g., extracellular matrix proteins), cer-tain types of cells are specialized for secretion of large

16.1

these organelles and the plasma membrane. Targeting to theER generally involves nascent proteins still in the process ofbeing synthesized. Proteins whose final destination is theGolgi, lysosome, or cell surface are transported along the se-cretory pathway by small vesicles that bud from the mem-brane of one organelle and then fuse with the membrane ofthe next organelle in the pathway (see Figure 16-1, left). Wediscuss vesicle-based protein sorting in the next chapter be-cause mechanistically it differs significantly from protein tar-geting to the membranes of intracellular organelles.

In this chapter, we examine how proteins are targeted tothe membrane of intracellular organelles and subsequentlyinserted into the organelle membrane or moved into the in-terior. Two features of this protein-sorting process initiallywere quite baffling: how a given protein could be targetedto only one specific membrane, and how relatively large pro-tein molecules could be translocated across a membranewithout disrupting the bilayer as a barrier to ions and smallmolecules. Using a combination of biochemical purificationmethods and genetic screens for identifying mutants unableto execute particular translocation steps, cell biologists haveidentified many of the cellular components required fortranslocation across each of the different intracellular mem-branes. In addition, many of the major translocationprocesses in the cell have been reconstituted using in vitrosystems, which can be freely manipulated experimentally.

These studies have shown that despite some variations,the same basic mechanisms govern protein sorting to all thevarious intracellular organelles. We now know, for instance,that the information to target a protein to a particular or-ganelle destination is encoded within the amino acid se-quence of the protein itself, usually within sequences of20–50 amino acids, known generically as signal sequences,or uptake-targeting sequences (see Figure 16-1). Each or-ganelle carries a set of receptor proteins that bind only tospecific kinds of signal sequences, thus assuring that the in-formation encoded in a signal sequence governs the speci-ficity of targeting. Once a protein containing a signalsequence has interacted with the corresponding receptor, theprotein chain is transferred to some kind of translocationchannel that allows the protein to pass through the mem-brane bilayer. The unidirectional transfer of a protein into anorganelle, without sliding back out into the cytoplasm, isusually achieved by coupling translocation to an ener-getically favorable process such as hydrolysis of ATP. Someproteins are subsequently sorted further to reach a subcom-partment within the target organelle; such sorting dependson yet other signal sequences and other receptor proteins. Finally, signal sequences often are removed from the matureprotein by specific proteases once translocation across themembrane is completed.

For each of the protein-targeting events discussed in thischapter, we will seek to answer four fundamental questions:

1. What is the nature of the signal sequence, and what dis-tinguishes it from other types of signal sequences?

16.1 • Translocation of Secretory Proteins Across the ER Membrane 659

0.5 �mAttached ribosomes

Free ribosomes

ER membraneCytosol ER lumen

▲ FIGURE 16-2 Electron micrograph of ribosomes attached

to the rough ER in a pancreatic acinar cell. Most of theproteins synthesized by this type of cell are to be secreted andare formed on membrane-attached ribosomes. A few membrane-unattached (free) ribosomes are evident; presumably, these aresynthesizing cytosolic or other nonsecretory proteins. [Courtesyof G. Palade.]

amounts of specific proteins. Pancreatic acinar cells, for in-stance, synthesize large quantities of several digestive en-zymes that are secreted into ductules that lead to the intestine.Because such secretory cells contain the organelles of the se-cretory pathway (e.g., ER and Golgi) in great abundance,they have been widely used in studying this pathway.

Early pulse-labeling experiments with pancreatic acinarcells showed that radioactively labeled amino acids are in-corporated primarily into newly synthesized secretory pro-teins. The ribosomes synthesizing these proteins are actuallybound to the surface of the ER. As a consequence, the por-tion of the ER that receives proteins entering the secretorypathway is known as the rough ER because these mem-branes are densely studded with ribosomes (Figure 16-2).When cells are homogenized, the rough ER breaks up intosmall closed vesicles, termed rough microsomes, with thesame orientation (ribosomes on the outside) as that foundin the intact cell. The experiments depicted in Figure 16-3, in


▲ EXPERIMENTAL FIGURE 16-3 Labeling experiments

demonstrate that secretory proteins are localized to the

ER lumen shortly after synthesis. Cells are incubated for a brief time with radiolabeled amino acids, so that only newlysynthesized proteins become labeled. The cells then arehomogenized, fracturing the plasma membrane and shearing therough ER into small vesicles called microsomes. Because theyhave bound ribosomes, microsomes have a much greaterbuoyant density than other membranous organelles and can be separated from them by a combination of differential andsucrose density-gradient centrifugation (Chapter 5). The purifiedmicrosomes are treated with a protease in the presence orabsence of a detergent. The labeled secretory proteinsassociated with the microsomes are digested by addedproteases only if the permeability barrier of the microsomalmembrane is first destroyed by treatment with detergent. Thisfinding indicates that the newly made proteins are inside themicrosomes, equivalent to the lumen of the rough ER.

Rough ER

Microsomeswith attachedribosomes

Labeledsecretoryprotein

Digestion ofsecretory protein

Add protease

Add protease

No digestion ofsecretory protein

Treat withdetergent

Homogenization

mRNA ▲which microsomes isolated from pulse-labeled cells aretreated with a protease, demonstrate that although secretoryproteins are synthesized on ribosomes bound to the cytosolicface of the ER membrane, they become localized in thelumen of ER vesicles during their synthesis.

A Hydrophobic N-Terminal Signal SequenceTargets Nascent Secretory Proteins to the ERAfter synthesis of a secretory protein begins on free ribo-somes in the cytosol, a 16- to 30-residue ER signal sequencein the nascent protein directs the ribosome to the ER mem-brane and initiates translocation of the growing polypeptideacross the ER membrane (see Figure 16-1, left). An ER signalsequence typically is located at the N-terminus of the protein,the first part of the protein to be synthesized. The signal se-quences of different secretory proteins contain one or morepositively charged amino acids adjacent to a continuousstretch of 6–12 hydrophobic residues (the core), but other-wise they have little in common. For most secretory proteins,the signal sequence is cleaved from the protein while it is stillgrowing on the ribosome; thus, signal sequences are usuallynot present in the “mature” proteins found in cells.

The hydrophobic core of ER signal sequences is essentialfor their function. For instance, the specific deletion of sev-eral of the hydrophobic amino acids from a signal sequence,or the introduction of charged amino acids into the hy-drophobic core by mutation, can abolish the ability of the N-terminus of a protein to function as a signal sequence. As aconsequence, the modified protein remains in the cytosol,unable to cross the ER membrane into the lumen. Using re-combinant DNA techniques, researchers have produced cy-tosolic proteins with added N-terminal amino acidsequences. Provided the added sequence is sufficiently longand hydrophobic, such a modified cytosolic protein istranslocated to the ER lumen. Thus the hydrophobic residues

in the core of ER signal sequences form a binding site that iscritical for the interaction of signal sequences with receptorproteins on the ER membrane.

Biochemical studies utilizing a cell-free protein-synthesizingsystem, mRNA encoding a secretory protein, and micro-somes stripped of their own bound ribosomes have clarifiedthe function and fate of ER signal sequences. Initial experi-ments with this system demonstrated that a typical secretoryprotein is incorporated into microsomes and has its signal se-quence removed only if the microsomes are present duringprotein synthesis (Figure 16-4). Subsequent experimentswere designed to determine the precise stage of protein syn-thesis at which microsomes must be present in order fortranslocation to occur. In these experiments, a drug that prevents initiation of translation was added to protein-synthesizing reactions at different times after protein synthe-sis had begun, and then stripped microsomes were added tothe reaction mixtures. These experiments showed that mi-crosomes must be added before the first 70 or so amino acidsare linked together in order for the completed secretory pro-tein to be localized in the microsomal lumen. At this point,the first 40 amino acids or so protrude from the ribosome,including the signal sequence that later will be cleaved off,and the next 30 or so amino acids are still buried within achannel in the ribosome. Thus the transport of most secre-tory proteins into the ER lumen occurs while the nascentprotein is still bound to the ribosome and being elongated,a process referred to as cotranslational translocation.

Cotranslational Translocation Is Initiated by TwoGTP-Hydrolyzing ProteinsSince secretory proteins are synthesized in association withthe ER membrane but not with any other cellular membrane,a signal-sequence recognition mechanism must target themthere. The two key components in this targeting are the signal-recognition particle (SRP) and its receptor located inthe ER membrane. The SRP is a cytosolic ribonucleoproteinparticle that transiently binds simultaneously to the ER sig-nal sequence in a nascent protein, to the large ribosomalunit, and to the SRP receptor.

Six discrete polypeptides and a 300-nucleotide RNAcompose the SRP (Figure 16-5a). One of the SRP proteins(P54) can be chemically cross-linked to ER signal sequences,evidence that this particular protein is the subunit that bindsto the signal sequence in a nascent secretory protein. A re-gion of P54 containing many amino acid residues with hy-drophobic side chains is homologous to a bacterial proteinknown as Ffh, which performs an analogous function to P54in the translocation of proteins across the inner membrane ofbacterial cells. The structure of Ffh contains a cleft whoseinner surface is lined by hydrophobic side chains (Figure 16-5b). The hydrophobic region of P54 is thought to containan analogous cleft that interacts with the hydrophobic N-termini of nascent secretory proteins and selectively targetsthem to the ER membrane. Two of the SRP proteins, P9 andP14, interact with the ribosome, while P68 and P72 are re-quired for protein translocation.

In the cell-free translation system described previously,the presence of SRP slows elongation of a secretory proteinwhen microsomes are absent, thereby inhibiting synthesis ofthe complete protein (see Figure 16-4). This finding suggeststhat interaction of the SRP with both the nascent chain of asecretory protein and with the free ribosome prevents thenascent chain from becoming too long for translocation intothe ER. Only after the SRP/nascent chain/ribosome complexhas bound to the SRP receptor in the ER membrane doesSRP release the nascent chain, allowing elongation at thenormal rate.

Figure 16-6 summarizes our current understanding of secretory protein synthesis and the role of the SRP and its


No incorporationinto microsomes;no removal ofsignal sequence

Cotranslational transport of protein into microsome and removal of signalsequence

Mature proteinchain without signal sequence

(b) Cell-free protein synthesis; microsomes present

N-terminalsignal sequence

(a) Cell-free protein synthesis; no microsomes present

Add microsomemembranes

Completed proteinswith signal sequences

▲ EXPERIMENTAL FIGURE 16-4 Cell-free experiments

demonstrate that translocation of secretory proteins into

microsomes is coupled to translation. Treatment ofmicrosomes with EDTA, which chelates Mg�2 ions, strips themof associated ribosomes, allowing isolation of ribosome-freemicrosomes, which are equivalent to ER membranes (see Figure16-3). Synthesis is carried out in a cell-free system containingfunctional ribosomes, tRNAs, ATP, GTP, and cytosolic enzymes towhich mRNA encoding a secretory protein is added. Thesecretory protein is synthesized in the absence of microsomes(a), but is translocated across the vesicle membrane and loses itssignal sequence only if microsomes are present during proteinsynthesis (b).

▲

nascent secretory protein with the ER membrane but also acttogether to permit elongation and synthesis of complete pro-teins only when ER membranes are present.

Ultimately, the SRP and SRP receptor function to bringribosomes that are synthesizing secretory proteins to the ERmembrane. The coupling of GTP hydrolysis to this target-ing process is thought to contribute to the fidelity by whichsignal sequences are recognized. Probably the energy fromGTP hydrolysis is used to release proteins lacking proper sig-nal sequences from the SRP and SRP receptor complex,thereby preventing their mistargeting to the ER membrane.(A similar coupling of GTP hydrolysis with binding of trans-lation elongation factors to ribosomes increases the fidelityof translation by ejecting aminoacyl-tRNA molecules thatcannot form correct base pairs with the codons in mRNA.)Interaction of the SRP/nascent chain/ribosome complex withthe SRP receptor is promoted when GTP is bound by boththe P54 subunit of SRP and the � subunit of the SRP recep-tor (see Figure 16-6). Subsequent transfer of the nascentchain and ribosome to a site on the ER membrane wheretranslocation can take place allows hydrolysis of the boundGTP. After dissociating, SRP and its receptor release thebound GDP and recycle to the cytosol ready to initiate an-other round of interaction between ribosomes synthesizingnascent secretory proteins with the ER membrane.

Passage of Growing Polypeptides Through the Translocon Is Driven by Energy Released During TranslationOnce the SRP and its receptor have targeted a ribosome syn-thesizing a secretory protein to the ER membrane, the ribo-some and nascent chain are rapidly transferred to thetranslocon, a protein-lined channel within the membrane. Astranslation continues, the elongating chain passes directlyfrom the large ribosomal subunit into the central pore of thetranslocon. The 60S ribosomal subunit is aligned with thepore of the translocon in such a way that the growing chainis never exposed to the cytoplasm and does not fold until itreaches the ER lumen (see Figure 16-6).

The translocon was first identified by mutations in theyeast gene encoding Sec61�, which caused a block in the


Interact with ribosomes

P9/P14

P68/P72

P19

P54

Required forprotein translocation

Binds ERsignalsequence

(a) Signal-recognition particle (SRP)

(b) Ffh signal sequence–binding domain (related to P54 subunit of SRP)

Hydrophobicbinding groove

RNA

▲ FIGURE 16-5 Structure of the signal-recognition particle

(SRP). (a) The SRP comprises one 300-nucleotide RNA and sixproteins designated P9, P14, P19, P54, P68, and P72. (The numeral indicates the molecular weight � 103.) All proteins except P54 bind directly to the RNA. (b) The bacterial Ffh proteinis homologous to the portion of P54 that binds ER signal sequences. This surface model shows the binding domain in Ffh, which contains a large cleft lined with hydrophobic aminoacids (purple) whose side chains interact with signal sequences. [Part (a) see K. Strub et al., 1991, Mol. Cell Biol. 11:3949; and S. High andB. Dobberstein, 1991, J. Cell Biol. 113:229. Part (b) adapted from R. J. Keenan et al., 1998, Cell 94:181.]

receptor in this process. The SRP receptor is an integralmembrane protein made up of two subunits: an � subunitand a smaller � subunit. Treatment of microsomes with verysmall amounts of protease cleaves the � subunit very nearits site of attachment to the membrane, releasing a solubleform of the SRP receptor. Protease-treated microsomes areunable to bind the SRP/nascent chain/ribosome complex orsupport cotranslational translocation. The soluble SRP re-ceptor fragment, however, retains its ability to interact withthe SRP/nascent chain/ribosome complex, causing release ofSRP and allowing chain elongation to proceed. Thus the SRPand SRP receptor not only help mediate interaction of a

▲

secretory protein was translocated from its SRP/ribosomecomplex into the vesicles. This finding indicates that the SRPreceptor and the Sec61 complex are the only ER-membraneproteins absolutely required for translocation. Thus the en-ergy derived from chain elongation at the ribosome appearsto be sufficient to push the polypeptide chain across themembrane in one direction.

Multiple copies of the Sec61 complex assemble to forma translocon channel, which can be visualized by electron microscopy. Images of the channel have been generated bycomputer averaging of electron micrographs of purifiedSec61 channels bound to ribosomes (Figure 16-8). Theseshow the channel as a cylinder, 5–6 nm high and 8.5 nm in

translocation of secretory proteins into the lumen of the ER. Subsequently, three proteins called the Sec61 complexwere found to form the mammalian translocon: Sec61�, an integral membrane protein with 10 membrane-spanning �helices, and two smaller proteins, termed Sec61� and Sec61�. Chemical cross-linking experiments demonstrated that the translocating polypeptide chain comes into con-tact with the Sec61� protein in both yeast and mammalian cells, confirming its identity as a translocon component (Figure 16-7).

When microsomes in the cell-free translocation systemwere replaced with reconstituted phospholipid vesicles con-taining only the SRP receptor and Sec61 complex, nascent


GTP

GTP

NH3+

Signal sequence

5'

mRNA

Cleavedsignalsequence

ER lumen

Cytosol

3'

SRP

SRP receptor

α

β

654

3

2

Signalpeptidase

Translocon(closed)

Foldedprotein

8

1

7

Translocon(open)

GDP + Pi

GDP + Pi

ERmembrane

ME

DIA

C

ON

NE

CT

IO

NS

Focus Anim

ation: Synthesis ofSecreted and M

embrane-B

ound Proteins

▲ FIGURE 16-6 Synthesis of secretory proteins and their

cotranslational translocation across the ER membrane. Steps, : Once the ER signal sequence emerges from the

ribosome, it is bound by a signal-recognition particle (SRP). Step : The SRP delivers the ribosome/nascent polypeptidecomplex to the SRP receptor in the ER membrane. Thisinteraction is strengthened by binding of GTP to both the SRPand its receptor. Step : Transfer of the ribosome/nascentpolypeptide to the translocon leads to opening of thistranslocation channel and insertion of the signal sequence andadjacent segment of the growing polypeptide into the centralpore. Both the SRP and SRP receptor, once dissociated from the translocon, hydrolyze their bound GTP and then are ready to

4

3

21

initiate the insertion of another polypeptide chain. Step :As the polypeptide chain elongates, it passes through thetranslocon channel into the ER lumen, where the signalsequence is cleaved by signal peptidase and is rapidlydegraded. Step : The peptide chain continues toelongate as the mRNA is translated toward the 3� end.Because the ribosome is attached to the translocon, thegrowing chain is extruded through the translocon into theER lumen. Steps , : Once translation is complete, theribosome is released, the remainder of the protein isdrawn into the ER lumen, the translocon closes, and theprotein assumes its native folded conformation.

87

6

5

sequence and approximately 30 adjacent amino acids, caninsert into the translocon pore (see Figure 16-6).

The mechanism by which the translocon channel opensand closes is controversial at this time. Some evidence sug-gests that a protein within the ER lumen blocks the translo-con pore when a ribosome is not bound to the cytosolic sideof the translocon. Other observations, however, indicate thatSec61 complexes may normally reside in the ER membranein an unassembled state and that the gating process involvesthe assembly of a translocon channel at the site where theribosome and nascent chain are brought to the membrane bythe SRP and SRP receptor.

As the growing polypeptide chain enters the lumen of theER, the signal sequence is cleaved by signal peptidase, whichis a transmembrane ER protein associated with the translo-con (see Figure 16-6). This protease recognizes a sequence onthe C-terminal side of the hydrophobic core of the signalpeptide and cleaves the chain specifically at this sequenceonce it has emerged into the luminal space of the ER. Afterthe signal sequence has been cleaved, the growing polypep-tide moves through the translocon into the ER lumen. Thetranslocon remains open until translation is completed andthe entire polypeptide chain has moved into the ER lumen.

diameter, with a central pore, roughly 2 nm in diameter, per-pendicular to the plane of the membrane.

If the translocon were always open in the ER membrane,especially in the absence of attached ribosomes and a translo-cating polypeptide, small molecules such as ATP and aminoacids would be able to diffuse freely through the centralpore. To maintain the permeability barrier of the ER mem-brane, the translocon is regulated so that it is open onlywhen a ribosome–nascent chain complex is bound. Thus thetranslocon is a gated channel analogous to the gated ionchannels described in Chapter 7. When the translocon firstopens, a loop of the nascent chain, containing the signal


Cross-linkingagent

tRNA

Nascent protein

Cytosol

Microsomalmembrane

Microsomallumen

NH3+

5'

Artificial mRNA

Sec61α

40S

60S

Ribosome

▲ EXPERIMENTAL FIGURE 16-7 Cross-linking experiments

show that Sec61� is a translocon component that contacts

nascent secretory proteins as they pass into the ER lumen.

An mRNA encoding the N-terminal 70 amino acids of thesecreted protein prolactin was translated in a cell-free systemcontaining microsomes (see Figure 16-4b). The mRNA lacked achain-termination codon and contained one lysine codon, nearthe middle of the sequence. The reactions contained a chemicallymodified lysyl-tRNA in which a light-activated cross-linkingreagent was attached to the lysine side chain. Although theentire mRNA was translated, the completed polypeptide couldnot be released from the ribosome and thus became “stuck”crossing the ER membrane. The reaction mixtures then wereexposed to an intense light, causing the nascent chain tobecome covalently bound to whatever proteins were near it inthe translocon. When the experiment was performed usingmicrosomes from mammalian cells, the nascent chain becamecovalently linked to Sec61�. Different versions of the prolactinmRNA were used to place the modified lysine residue atdifferent distances from the ribosome; cross-linking to Sec61�

was observed only when the modified lysine was positionedwithin the translocation channel. [Adapted from T. A. Rapoport, 1992,Science 258:931, and D. Görlich and T. A. Rapoport, 1993, Cell 75:615.]

40Ssubunit

60Ssubunit

tRNA

Translocon10 nm

▲ EXPERIMENTAL FIGURE 16-8 Electron microscopy

reconstruction reveals that a translocon associates closely

with a ribosome. Purified Sec61 complexes were solubilized bytreatment of ER membranes with detergents. When ribosomeswere added, translocons (blue) reassembled in artificial phospholipidbilayers. The resulting particles were frozen, and electronmicrographs of a large number of particles were generated,stored in a computer, and then averaged to produce a singleimage. A representation of the approximate size and position ofthe ER lipid bilayer has been added. Note that although theribosome is firmly attached to the translocon, there is a gapbetween the two structures. The fingerlike appendage below thetranslocon channel is thought to be formed from a proteincomplex that associates with the translocon. [Courtesy Dr. ChristopherAkey and Jean-Francois Menetret, Boston University School of Medicine.]

a peptide-binding domain and an ATPase domain. Thesechaperones bind and stabilize unfolded or partially foldedproteins (see Figure 3-11).

The current model for post-translational translocationof a protein into the ER is outlined in Figure 16-9. Once theN-terminal segment of the protein enters the ER lumen, sig-nal peptidase cleaves the signal sequence just as in cotrans-lational translocation (step 1). Interaction of BiP·ATP withthe luminal portion of the Sec63 complex causes hydrolysisof the bound ATP, producing a conformational change in BiPthat promotes its binding to an exposed polypeptide chain(step 2). Since the Sec63 complex is located near the translo-con, BiP is thus activated at sites where nascent polypeptidescan enter the ER. Certain experiments suggest that in theabsence of binding to BiP, an unfolded polypeptide slidesback and forth within the translocon channel. Such randomsliding motions rarely result in the entire polypeptide’scrossing the ER membrane. Binding of a molecule ofBiP·ADP to the luminal portion of the polypeptide preventsbacksliding of the polypeptide out of the ER. As further in-ward random sliding exposes more of the polypeptide onthe luminal side of the ER membrane, successive binding of

ATP Hydrolysis Powers Post-translationalTranslocation of Some Secretory Proteins in Yeast

In most eukaryotes, secretory proteins enter the ER by co-translational translocation, using energy derived from trans-lation to pass through the membrane, as we’ve justdescribed. In yeast, however, some secretory proteins enterthe ER lumen after translation has been completed. In suchpost-translational translocation, the translocating proteinpasses through the same Sec61 translocon that is used in co-translational translocation. However, the SRP and SRP re-ceptor are not involved in post-translational translocation,and in such cases a direct interaction between the translo-con and the signal sequence of the completed protein appearsto be sufficient for targeting to the ER membrane. In addi-tion, the driving force for unidirectional translocation acrossthe ER membrane is provided by an additional protein com-plex known as the Sec63 complex and a member of theHsc70 family of molecular chaperones known as BiP. Thetetrameric Sec63 complex is embedded in the ER membranein the vicinity of the translocon, while BiP is localized to theER lumen. Like other members of the Hsc70 family, BiP has


ATP

ATP ADP ATP ATPADP

ADP

ADP

ADP

ADP

ADP

ADP

ADP

ATP

ADP

ER lumen

Cytosol


Translocatingpolypeptidechain

1

Pi

4

5

6

3

Pi

Sec63 complex

BiP(bound to ATP)

2

Pi

Translocon

NH3+

▲ FIGURE 16-9 Post-translational translocation across ER

membrane.This mechanism is fairly common in yeast andprobably occurs occasionally in higher eukaryotes. Small arrowsinside the translocon represent random sliding of thetranslocating polypeptide inward and outward. Successive bindingof BiP·ADP to entering segments of the polypeptide prevents thechain from sliding out toward the cytosol. See the text fordiscussion. [See K. E. Matlack et al., 1997, Science 277:938.]

Insertion of Proteins into the ER MembraneIn previous chapters we have encountered many of the vastarray of integral (transmembrane) proteins that are presentin the plasma membrane and other cellular membranes.Each such protein has a unique orientation with respectto the membrane’s phospholipid bilayer. Integral proteinslocated in ER, Golgi, and lysosomal membranes and in theplasma membrane, which are synthesized on the roughER, remain embedded in the membrane as they move totheir final destinations along the same pathway followedby soluble secretory proteins (see Figure 16-1). During thistransport, the orientation of a membrane protein is pre-served; that is, the same segments of the protein alwaysface the cytosol, while other segments always face in theopposite direction. Thus the final orientation of thesemembrane proteins is established during their biosynthesison the ER membrane. In this section, we first see how in-tegral proteins can interact with membranes. Then we ex-amine how several types of sequences, collectively knownas topogenic sequences, direct the insertion and orienta-tion of various classes of integral proteins into the mem-brane. These processes build on and adapt the basicmechanism used to translocate soluble secretory proteinsacross the ER membrane.

Several Topological Classes of Integral MembraneProteins Are Synthesized on the ERThe topology of a membrane protein refers to the numberof times that its polypeptide chain spans the membrane andthe orientation of these membrane-spanning segments withinthe membrane. The key elements of a protein that determineits topology are membrane-spanning segments themselves,which usually contain 20–25 hydrophobic amino acids. Eachsuch segment forms an � helix that spans the membrane,with the hydrophobic amino acid residues anchored to thehydrophobic interior of the phospholipid bilayer.

Scientists have found it useful to categorize integral mem-brane proteins into the four topological classes illustrated inFigure 16-10. Topological classes I, II, and III comprise single-pass proteins, which have only one membrane-spanning �-helical segment. Type I proteins have a cleavedN-terminal signal sequence and are anchored in the mem-brane with their hydrophilic N-terminal region on the luminal face (also known as the exoplasmic face) and theirhydrophilic C-terminal region on the cytosolic face. Type IIproteins do not contain a cleavable signal sequence and areoriented with their hydrophilic N-terminal region on the cy-tosolic face and their hydrophilic C-terminal region on theexoplasmic face (i.e., opposite to type I proteins). Type IIIproteins have the same orientation as type I proteins, but donot contain a cleavable signal sequence. These differenttopologies reflect distinct mechanisms used by the cell to

16.2BiP·ADP molecules to the polypeptide chain acts as aratchet, ultimately drawing the entire polypeptide into theER within a few seconds (steps 3 and 4). On a slower timescale, the BiP molecules spontaneously exchange theirbound ADP for ATP, leading to release of the polypeptide,which can then fold into its native conformation (steps 5and 6). The recycled BiP·ATP then is ready for another in-teraction with Sec63.

The overall reaction carried out by BiP is an importantexample of how the chemical energy released by the hydrol-ysis of ATP can power the mechanical movement of a proteinacross a membrane. Some bacterial cells also use an ATP-driven process for translocating completed proteins acrossthe plasma membrane. However, the mechanism of post-translational translocation in bacteria differs somewhat fromthat in yeast, as we describe in Section 16.4.

KEY CONCEPTS OF SECTION 16.1

Translocation of Secretory Proteins Across the ER Membrane

■ Synthesis of secreted proteins; enzymes destined for theER, Golgi complex, or lysosome; and integral plasma-membrane proteins begins on cytosolic ribosomes, whichbecome attached to the membrane of the ER, forming therough ER (see Figure 16-1, left).

■ The ER signal sequence on a nascent secretory proteinconsists of a segment of hydrophobic amino acids, gener-ally located at the N-terminus.

■ In cotranslational translocation, the signal-recognitionparticle (SRP) first recognizes and binds the ER signal se-quence on a nascent secretory protein and in turn is boundby an SRP receptor on the ER membrane, thereby target-ing the ribosome/nascent chain complex to the ER.

■ The SRP and SRP receptor then mediate insertion of thenascent secretory protein into the translocon. Hydrolysisof GTP by the SRP and its receptor drive this dockingprocess (see Figure 16-6). As the ribosome attached to thetranslocon continues translation, the unfolded proteinchain is extruded into the ER lumen. No additional energyis required for translocation.

■ In post-translational translocation, a completed secre-tory protein is targeted to the ER membrane by interac-tion of the signal sequence with the translocon. Thepolypeptide chain is then pulled into the ER by a ratchet-ing mechanism that requires ATP hydrolysis by the chap-erone BiP, which stabilizes the entering polypeptide (seeFigure 16-9).

■ In both cotranslational and post-translational translo-cation, a signal peptidase in the ER membrane cleaves theER signal sequence from a secretory protein soon after theN-terminus enters the lumen.


initiates cotranslational translocation of the protein throughthe combined action of the SRP and SRP receptor. Once theN-terminus of the growing polypeptide enters the lumen ofthe ER, the signal sequence is cleaved, and the growing chaincontinues to be extruded across the ER membrane. However,unlike the case with secretory proteins, a sequence of about22 hydrophobic amino acids in the middle of a type I pro-tein stops transfer of the nascent chain through the translo-con (Figure 16-11). This internal sequence, because of its hydrophobicity, can move laterally between the protein sub-units that form the wall of the translocon and become an-chored in the phospholipid bilayer of the membrane, whereit remains. Because of its dual function, this sequence iscalled a stop-transfer anchor sequence.

Once translocation is interrupted, translation continuesat the ribosome, which is still anchored to the now unoccu-pied and closed translocon. As the C-terminus of the proteinchain is synthesized, it loops out on the cytosolic side of themembrane. When translation is completed, the ribosome isreleased from the translocon and the C-terminus of the newlysynthesized type I protein remains in the cytosol.

Support for this model, depicted in Figure 16-11, hascome from studies in which cDNAs encoding various mutantreceptors for human growth hormone (HGH) are expressedin cultured mammalian cells. The wild-type HGH receptor, atypical type I protein, is transported normally to the plasmamembrane. However, a mutant receptor that has chargedresidues inserted into the �-helical membrane-spanning seg-ment, or that is missing most of this segment, is translocated

establish the membrane orientation of transmembrane seg-ments, as discussed in the next section.

The proteins forming topological class IV contain multi-ple membrane-spanning segments (see Figure 16-10). Manyof the membrane transport proteins discussed in Chapter 7and the numerous G protein–coupled receptors covered inChapter 13 belong to this class, sometimes called multipassproteins. A final type of membrane protein lacks a hy-drophobic membrane-spanning segment altogether; instead,these proteins are linked to an amphipathic phospholipid anchor that is embedded in the membrane.

Internal Stop-Transfer and Signal-AnchorSequences Determine Topology of Single-Pass ProteinsWe begin our discussion with the membrane insertion of integral proteins that contain a single, hydrophobic membrane-spanning segment. Two sequences are involved intargeting and orienting type I proteins in the ER membrane,whereas type II and type III proteins contain a single, inter-nal topogenic sequence.

Type I Proteins All type I transmembrane proteins possessan N-terminal signal sequence that targets them to the ERand an internal hydrophobic sequence that becomes themembrane-spanning � helix. The N-terminal signal sequenceon a nascent type I protein, like that of a secretory protein,

16.2 • Insertion of Proteins into the ER Membrane 667

Cytochrome P450

NH3+

COO−

Asialoglycoprotein receptor

Transferrin receptor

Sucrase–isomaltase precursor

Golgi galactosyltransferase

Golgi sialyltransferase

Influenza HN protein

COO−

NH3+

Type II

Type III

G protein–coupled receptors

(e.g., β-adrenergic receptor)

Glucose transporters (e.g., GLUT1)

Voltage-gated Ca2+ channels

ABC small molecule pumps

CFTR (Cl−) channel

Sec61

Connexin

COO−

NH3+

Type IV

Glycophorin

LDL receptor

Influenza HA protein

Insulin receptor

Growth hormone receptor

NH3+

COO−

Type I

Cleavedsignal sequence

Exoplasmic

space

(ER or Golgi

lumen;

cell exterior)

Cytosol

▲ FIGURE 16-10 Major topological classes of integral

membrane proteins synthesized on the rough ER. Thehydrophobic segments of the protein chain form � helicesembedded in the membrane bilayer; the regions outside themembrane are hydrophilic and fold into various conformations. All type IV proteins have multiple transmembrane � helices. The type IV topology depicted here corresponds to that of G

protein–coupled receptors: seven � helices, the N-terminus onthe exoplasmic side of the membrane, and the C-terminus onthe cytosolic side. Other type IV proteins may have a differentnumber of helices and various orientations of the N-terminus and C-terminus. [See E. Hartmann et al., 1989, Proc. Nat’l. Acad. Sci.USA 86:5786, and C. A. Brown and S. D. Black, 1989, J. Biol. Chem.264:4442.]

units forming the translocon wall into the phospholipid bilayer, where it functions as a membrane anchor. Thus this function is similar to the anchoring function of the stop-transfer anchor sequence in type I proteins.

In the case of type III proteins, the signal-anchor se-quence, which is located near the N-terminus, inserts the nas-cent chain into the ER membrane with its N-terminus facingthe lumen, just the opposite of type II proteins. The signal-anchor sequence of type III proteins also prevents further ex-trusion of the nascent chain into the ER lumen, functioningas a stop-transfer sequence. Continued elongation of thechain C-terminal to the signal-anchor/stop-transfer sequenceproceeds as it does for type I proteins, with the hydropho-bic sequence moving laterally between the translocon sub-units to anchor the polypeptide in the ER membrane (seeFigure 16-11).

One of the features of signal-anchor sequences that ap-pears to determine their insertion orientation is a high den-sity of positively charged amino acids adjacent to one endof the hydrophobic segment. For reasons that are not wellunderstood these positively charged residues tend to remainon the cytosolic side of the membrane, thereby dictating theorientation of the signal-anchor sequence within the translo-con. Thus type II proteins tend to have positively charged

entirely into the ER lumen and is eventually secreted fromthe cell. These findings establish that the hydrophobic membrane-spanning � helix of the HGH receptor and ofother type I proteins functions both as a stop-transfer se-quence and a membrane anchor that prevents the C-terminusof the protein from crossing the ER membrane.

Type II and Type III Proteins Unlike type I proteins, type IIand type III proteins lack a cleavable N-terminal ER signalsequence. Instead, both possess a single internal hydrophobicsignal-anchor sequence that functions as both an ER signalsequence and membrane-anchor sequence. Recall that type IIand type III proteins have opposite orientations in the mem-brane (see Figure 16-10); this difference depends on the ori-entation that their respective signal-anchor sequences assumewithin the translocon.

The internal signal-anchor sequence in type II proteins di-rects insertion of the nascent chain into the ER membraneso that the N-terminus of the chain faces the cytosol (Figure16-12). The internal signal-anchor sequence is not cleavedand remains in the translocon while the C-terminal regionof the growing chain is extruded into the ER lumen by co-translational translocation. During synthesis, the signal-anchor sequence moves laterally between the protein sub-


Signalpeptidase

Opentranslocon

mRNA

NH3+

Stop-transferanchorsequence

Nascentpolypeptidechain

Cytosol

ER lumen

5'


1 2 5436 3'

NH3+

NH3+

NH3+

NH3+

NH3+

COO−

▲ FIGURE 16-11 Synthesis and insertion into the ER

membrane of type I single-pass proteins. Step : After theribosome/nascent chain complex becomes associated with atranslocon in the ER membrane, the N-terminal signal sequenceis cleaved. This process occurs by the same mechanism as theone for soluble secretory proteins (see Figure 16-6). Steps , :The chain is elongated until the hydrophobic stop-transfer anchor sequence is synthesized and enters the translocon, where itprevents the nascent chain from extruding farther into the ER

lumen. Step : The stop-transfer anchor sequence moves laterallybetween the translocon subunits and becomes anchored in thephospholipid bilayer. At this time, the translocon probably closes.Step : As synthesis continues, the elongating chain may loopout into the cytosol through the small space between theribosome and translocon (see Figure 16-8). Step : Whensynthesis is complete, the ribosomal subunits are released intothe cytosol, leaving the protein free to diffuse in the membrane.[See H. Do et al., 1996, Cell 85:369, and W. Mothes et al., 1997, Cell 89:523.]

1

2 3

4

5

6

Multipass Proteins Have Multiple Internal Topogenic SequencesFigure 16-13 summarizes the arrangements of topogenicsequences in single-pass and multipass transmembrane pro-teins. In multipass (type IV) proteins, each of the membrane-spanning � helices acts as a topogenic sequence in the waysthat we have already discussed. Multipass proteins fall intoone of two types depending on whether the N-terminus ex-tends into the cytosol or the exoplasmic space (i.e., the ERlumen, cell exterior). This N-terminal topology usually isdetermined by the hydrophobic segment closest to the N-terminus and the charge of the sequences flanking it. If atype IV protein has an even number of transmembrane �helices, both its N-terminus and C-terminus will be ori-ented toward the same side of the membrane (see Figure16-13d). Conversely, if a type IV protein has an odd num-ber of � helices, its two ends will have opposite orientations(see Figure 16-13e).

Proteins with N-Terminus in Cytosol (Type IV-A) Amongthe multipass proteins whose N-terminus extends into the cy-tosol are the various glucose transporters (GLUTs) and mostion-channel proteins discussed in Chapter 7. In these pro-teins, the hydrophobic segment closest to the N-terminus ini-tiates insertion of the nascent chain into the ER membranewith the N-terminus oriented toward the cytosol; thus this �-helical segment functions like the internal signal-anchor sequence of a type II protein (see Figure 16-12). As the nas-cent chain following the first � helix elongates, it movesthrough the translocon until the second hydrophobic � helixis formed. This helix prevents further extrusion of the na-scent chain through the translocon; thus its function is simi-lar to that of the stop-transfer anchor sequence in a type Iprotein (see Figure 16-11).

After synthesis of the first two transmembrane � helices,both ends of the nascent chain face the cytosol and the loopbetween them extends into the ER lumen. The C-terminusof the nascent chain then continues to grow into the cytosol,as it does in synthesis of type I and type III proteins. Accord-ing to this mechanism, the third � helix acts as another typeII signal-anchor sequence, and the fourth as another stop-transfer anchor sequence (see Figure 16-13d). Apparently,once the first topogenic sequence of a multipass polypeptideinitiates association with the translocon, the ribosome re-mains attached to the translocon, and topogenic sequencesthat subsequently emerge from the ribosome are threadedinto the translocon without the need for the SRP and the SRPreceptor.

Experiments that use recombinant DNA techniques to ex-change hydrophobic � helices have provided insight into thefunctioning of the topogenic sequences in type IV-A multipassproteins. These experiments indicate that the order of the hy-drophobic � helices relative to each other in the growingchain largely determines whether a given helix functions as asignal-anchor sequence or stop-transfer anchor sequence.

residues on the N-terminal side of their signal-anchor se-quence, whereas type III proteins tend to have positivelycharged residues on the C-terminal side of their signal-anchor sequence.

A striking experimental demonstration of the importanceof the flanking charge in determining membrane orientationis provided by neuraminidase, a type II protein in the sur-face coat of influenza virus. Three arginine residues are lo-cated just N-terminal to the internal signal-anchor sequencein neuraminidase. Mutation of these three positively chargedresidues to negatively charged glutamate residues causes neu-raminidase to acquire the reverse orientation. Similar exper-iments have shown that other proteins, with either type II ortype III orientation, can be made to “flip” their orientation inthe ER membrane by mutating charged residues that flankthe internal signal-anchor segment.


Nascentpolypeptidechain

Signal-anchorsequence

21

Translocon

ER lumen

Cytosol

3'

NH3+

NH3+

COO−

NH3+

3

5'

mRNA

3

+++ +

++

+++

▲ FIGURE 16-12 Synthesis and insertion into the ER

membrane of type II single-pass proteins. Step : After theinternal signal-anchor sequence is synthesized on a cytosolicribosome, it is bound by an SRP (not shown), which directs theribosome/nascent chain complex to the ER membrane. This issimilar to targeting of soluble secretory proteins except that thehydrophobic signal sequence is not located at the N-terminus andis not subsequently cleaved. The nascent chain becomes orientedin the translocon with its N-terminal portion toward the cytosol.This orientation is believed to be mediated by the positivelycharged residues shown N-terminal to the signal-anchorsequence.Step : As the chain is elongated and extruded intothe lumen, the internal signal-anchor moves laterally out of thetranslocon and anchors the chain in the phospholipid bilayer. Step : Once protein synthesis is completed, the C-terminus ofthe polypeptide is released into the lumen, and the ribosomalsubunits are released into the cytosol. [See M. Spiess and H. F.Lodish, 1986, Cell 44:177, and H. Do et al., 1996, Cell 85:369.]

1

2

3

▲

by alternating type II signal-anchor sequences and stop-transfer sequences, as just described for type IV-A proteins.

A Phospholipid Anchor Tethers Some Cell-Surface Proteins to the MembraneSome cell-surface proteins are anchored to the phospholipidbilayer not by a sequence of hydrophobic amino acids but by a covalently attached amphipathic molecule, glyco-sylphosphatidylinositol (GPI) (Figure 16-14a). These pro-teins are synthesized and initially anchored to the ERmembrane exactly like type I transmembrane proteins, witha cleaved N-terminal signal sequence and internal stop-transfer anchor sequence directing the process (see Figure 16-11). However, a short sequence of amino acids in the lu-minal domain, adjacent to the membrane-spanning domain,is recognized by a transamidase located within the ER mem-brane. This enzyme simultaneously cleaves off the original

Other than its hydrophobicity, the specific amino acid se-quence of a particular helix has little bearing on its function.Thus the first N-terminal � helix and the subsequent odd-numbered ones function as signal-anchor sequences, whereasthe intervening even-numbered helices function as stop-transfer anchor sequences.

Proteins with N-Terminus in the Exoplasmic Space (TypeIV-B) The large family of G protein–coupled receptors, all ofwhich contain seven transmembrane � helices, constitute themost numerous type IV-B proteins, whose N-terminus ex-tends into the exoplasmic space. In these proteins, the hy-drophobic � helix closest to the N-terminus often is followedby a cluster of positively charged amino acids, similar to atype III signal-anchor sequence. As a result, the first � helixinserts the nascent chain into the translocon with the N-terminus extending into the lumen (see Figure 16-13e). Asthe chain is elongated, it is inserted into the ER membrane


Lumen

Cytosol

STA = Internal stop-transfer anchor sequenceSA-II = Internal signal-anchor sequenceSA-III = Internal signal-anchor sequence

NH3+

NH3+

NH3+

NH3+

NH3+

COO−

COO−

COO−

COO−

COO−

STA

+++

Lumen

LumenCytosol

SA-II

Lumen

SA-III

Cytosol

LumenLumen

LumenLumenLumen

Cytosol

CytosolCytosolCytosolCytosol

SA-IISA-IISA-III STA STA SA-II STA

Cytosol Cytosol

STASTA SA-IISA-II

Signal sequence

(a) Type I

(b) Type II

(c) Type III

(d) Type IV-A

(e) Type IV-B

+++

+++ +++

+++ +++ ++++++

▲ FIGURE 16-13 Arrangement of topogenic sequences in

single-pass and multipass membrane proteins inserted into

the ER membrane. Topogenic sequences are shown in red;soluble, hydrophilic portions, in blue. The internal topogenicsequences form transmembrane � helices that anchor theproteins or segments of proteins in the membrane. (a) Type Iproteins contain a cleaved signal sequence and a single internalstop-transfer anchor (STA). (b, c) Type II and type III proteinscontain a single internal signal-anchor (SA) sequence. Thedifference in the orientation of these proteins depends largely on

whether there is a high density of positively charged amino acids(��) on the N-terminal side of the SA sequence (type II) or onthe C-terminal side of the SA sequence (type III). (d, e) Nearly allmultipass proteins lack a cleavable signal sequence, as depictedin the examples shown here. Type IV-A proteins, whose N-terminus faces the cytosol, contain alternating SA-II sequencesand STA sequences. Type IV-B proteins, whose N-terminus facesthe lumen, begin with a SA-III sequence followed by alternatingSA-II and STA sequences. Proteins of each type with differentnumbers of � helices (odd or even) are known.

the cytoskeleton. In addition, the GPI anchor targets the at-tached protein to the apical domain of the plasma mem-brane in certain polarized epithelial cells, as we discuss inChapter 17.

Some membrane proteins are tethered to the cytosolicface of the plasma membrane by other types of lipid anchors(see Figure 5-15). An important example is Ras protein,which plays a key role in intracellular signaling pathwaysdiscussed in Chapter 14.

The Topology of a Membrane Protein Often Can Be Deduced from Its SequenceAs we have seen, various topogenic sequences in integralmembrane proteins synthesized on the ER govern interactionof the nascent chain with the translocon. When scientistsbegin to study a protein of unknown function, the identifi-cation of topogenic sequences within the corresponding genesequence can provide important clues about the protein’stopological class and function. Suppose, for example, thatthe gene for a protein known to be required for a cell-to-cellsignaling pathway contains nucleotide sequences that encodean apparent N-terminal signal sequence and an internal hy-drophobic sequence. These findings would suggest that theprotein is a type I integral membrane protein and thereforemay be a cell-surface receptor for an extracellular ligand.

Identification of topogenic sequences requires a way toscan sequence databases for segments that are sufficiently hy-drophobic to be either a signal sequence or a transmembraneanchor sequence. Topogenic sequences can often be identifiedwith the aid of computer programs that generate a hydropa-thy profile for the protein of interest. The first step is to assigna value known as the hydropathic index to each amino acid inthe protein. By convention, hydrophobic amino acids are as-signed positive values, and hydrophilic amino acids negativevalues. Although different scales for the hydropathic indexexist, all assign the most positive values to amino acids withside chains made up of mostly hydrocarbon residues (e.g.,phenylalanine and methionine) and the most negative valuesto charged amino acids (e.g., arginine and aspartate). Thesecond step is to identify longer segments of sufficient over-all hydrophobicity to be N-terminal signal sequences or in-ternal stop-transfer sequences and signal-anchor sequences.To accomplish this, the total hydropathic index for each suc-cessive sliding “window” of 20 consecutive amino acids iscalculated along the entire length of the protein. Plots of thesecalculated values against position in the amino acid sequenceyield a hydropathy profile.

Figure 16-15 shows the hydropathy profiles for three dif-ferent membrane proteins. The prominent peaks in suchplots identify probable topogenic sequences, as well as theirposition and approximate length. For example, the hydropa-thy profile of the human growth hormone receptor revealsthe presence of both a hydrophobic signal sequence at the ex-treme N-terminus of the protein and an internal hydropho-bic stop-transfer sequence (see Figure 16-15a). On the basis

stop-transfer anchor sequence and transfers the remainderof the protein to a preformed GPI anchor in the membrane(Figure 16-14b).

Why change one type of membrane anchor for another?Attachment of the GPI anchor, which results in removal ofthe cytosol-facing hydrophilic domain from the protein, canhave several consequences. Proteins with GPI anchors, forexample, can diffuse in the plane of the phospholipid bilayermembrane. In contrast, many proteins anchored by membrane-spanning � helices are immobilized in the mem-brane because their cytosol-facing segments interact with


PreformedGPI anchor

COO− COO−

ER lumen

Cytosol

NH3+

Hydrophobic Polar NH3+

NH3+

PO4 PO4 NH3+

PO4 NH2

NH3+

Precursorprotein NH3

+

Mature GPI-linkedprotein

= Inositol

= Glucosamine

= Mannose

= Phosphoethanolamine

(a)

(b)GPItransamidase

Fatty acyl chains

▲ FIGURE 16-14 GPI-anchored proteins. (a) Structure of aglycosylphosphatidylinositol (GPI) from yeast. The hydrophobicportion of the molecule is composed of fatty acyl chains,whereas the polar (hydrophilic) portion of the molecule iscomposed of carbohydrate residues and phosphate groups. Inother organisms, both the length of the acyl chains and thecarbohydrate moieties may vary somewhat from the structureshown. (b) Formation of GPI-anchored proteins in the ERmembrane. The protein is synthesized and initially inserted intothe ER membrane as shown in Figure 16-11. A specifictransamidase simultaneously cleaves the precursor protein withinthe exoplasmic-facing domain, near the stop-transfer anchorsequence (red), and transfers the carboxyl group of the new C-terminus to the terminal amino group of a preformed GPI anchor.[See C. Abeijon and C. B. Hirschberg, 1992, Trends Biochem. Sci. 17:32,and K. Kodukula et al., 1992, Proc. Nat’l. Acad. Sci. USA 89:4982.]

▲

Finally, sequence homology to a known protein may per-mit accurate prediction of the topology of a newly discoveredmultipass protein. For example, the genomes of multicellularorganisms encode a very large number of multipass proteinswith seven transmembrane � helices. The similarities be-tween the sequences of these proteins strongly suggest thatall have the same topology as the well-studied G protein–coupled receptors, which have the N-terminus oriented tothe exoplasmic side and the C-terminus oriented to the cy-tosolic side of the membrane.


Insertion of Proteins into the ER Membrane

■ Integral membrane proteins synthesized on the rough ERfall into four topological classes (see Figure 16-10).

■ Topogenic sequences—N-terminal signal sequences, in-ternal stop-transfer anchor sequences, and internal signal-anchor sequences—direct the insertion and orientation ofnascent proteins within the ER membrane. This orienta-tion is retained during transport of the completed mem-brane protein to its final destination.

■ Single-pass membrane proteins contain one or two topo-genic sequences. In multipass membrane proteins, each �-helical segment can function as an internal topogenic se-quence, depending on its location in the polypeptide chain

of this profile, we can deduce, correctly, that the humangrowth hormone receptor is a type I integral membrane pro-tein. The hydropathy profile of the asialoglycoprotein recep-tor reveals a prominent internal hydrophobic signal-anchorsequence but gives no indication of a hydrophobic N-terminal signal sequence (see Figure 16-15b). Thus we canpredict that the asialoglycoprotein receptor is a type II ortype III membrane protein. The distribution of chargedresidues on either side of the signal-anchor sequence oftencan distinguish between these possibilities, as positivelycharged amino acids flanking a membrane-spanning segmentusually are oriented toward the cytosolic face of the mem-brane. For instance, in the case of the asialoglycoprotein re-ceptor, a type II protein, the residues on the N-terminal sideof the hydrophobic peak carry a net positive charge.

The hydropathy profile of the GLUT1 glucose trans-porter, a multipass membrane protein, reveals the presence ofmany segments that are sufficiently hydrophobic to be mem-brane-spanning helices (Figure 16-15c). The complexity ofthis profile illustrates the difficulty both in unambiguouslyidentifying all the membrane-spanning segments in a multi-pass protein and in predicting the topology of individual signal-anchor and stop-transfer sequences. More sophisti-cated computer algorithms have been developed that takeinto account the presence of positively charged amino acidsadjacent to hydrophobic segments, as well as the length ofand spacing between segments. Using all this information,the best algorithms can predict the complex topology of mul-tipass proteins with an accuracy greater than 75 percent.


(a) Human growth hormone receptor (type I)

(b) Asialoglycoprotein receptor (type II) (c) GLUT1 (type IV)

N-terminus 100

−3−2−1

01234

200 300 400 500

Signal sequence Stop-transfer sequence

Signal-anchor sequence Transmembrane sequences

C-terminus

100

−3−2−1

01234

200 300 400100

−3−2−1

01234

200

▲ EXPERIMENTAL FIGURE 16-15 Hydropathy profiles can

identify likely topogenic sequences in integral membrane

proteins. Hydropathy profiles are generated by plotting the totalhydrophobicity of each segment of 20 contiguous amino acidsalong the length of a protein. Positive values indicate relativelyhydrophobic portions; negative values, relatively polar portions of

the protein. Probable topogenic sequences are marked. Thecomplex profiles for multipass (type IV) proteins, such as GLUT1in part (c), often must be supplemented with other analyses todetermine the topology of these proteins. See the text fordiscussion.

A Preformed N-Linked Oligosaccharide Is Addedto Many Proteins in the Rough ERBiosynthesis of all N-linked oligosaccharides begins in therough ER with addition of a preformed oligosaccharide pre-cursor containing 14 residues (Figure 16-16). The structureof this precursor is the same in plants, animals, and single-celled eukaryotes—a branched oligosaccharide, containingthree glucose (Glc), nine mannose (Man), and two N-acetylglucosamine (GlcNAc) molecules, which can be writtenas Glc3Man9(GlcNAc)2. This branched carbohydrate struc-ture is modified in the ER and Golgi compartments, but 5of the 14 residues are conserved in the structures of all N-linked oligosaccharides on secretory and membrane proteins.

The precursor oligosaccharide is linked by a pyrophos-phoryl residue to dolichol, a long-chain polyisoprenoid lipidthat is firmly embedded in the ER membrane and acts as acarrier for the oligosaccharide. The dolichol pyrophospho-ryl oligosaccharide is formed on the ER membrane in a com-plex set of reactions catalyzed by enzymes attached to thecytosolic or luminal faces of the rough ER membrane

and the presence of adjacent positively charged residues(see Figure 16-13).

■ Some cell-surface proteins are initially synthesized as typeI proteins on the ER and then are cleaved with their lumi-nal domain transferred to a GPI anchor (see Figure 16-14).

■ The topology of membrane proteins can often be correctlypredicted by computer programs that identify hydrophobictopogenic segments within the amino acid sequence and gen-erate hydropathy profiles (see Figure 16-15).

Protein Modifications, Folding,and Quality Control in the ERMembrane and soluble secretory proteins synthesized on therough ER undergo four principal modifications before theyreach their final destinations: (1) addition and processing ofcarbohydrates (glycosylation) in the ER and Golgi, (2) for-mation of disulfide bonds in the ER, (3) proper folding ofpolypeptide chains and assembly of multisubunit proteins inthe ER, and (4) specific proteolytic cleavages in the ER,Golgi, and secretory vesicles.

One or more carbohydrate chains are added to the vastmajority of proteins that are synthesized on the rough ER;indeed, glycosylation is the principal chemical modificationto most of these proteins. Carbohydrate chains in glycopro-teins may be attached to the hydroxyl group in serine andthreonine residues or to the amide nitrogen of asparagine.These are referred to as O-linked and N-linked oligosaccha-rides, respectively. O-linked oligosaccharides, such as thosefound in collagen and glycophorin, often contain only one tofour sugar residues. The more common N-linked oligosac-charides are larger and more complex, containing severalbranches in mammalian cells. In this section we focus on N-linked oligosaccharides, whose initial synthesis occurs in theER. After the initial glycosylation of a protein in the ER, theoligosaccharide chain is modified in the ER and commonlyin the Golgi, as well.

Disulfide bond formation, protein folding, and assemblyof multimeric proteins, which take place exclusively in therough ER, also are discussed in this section. Only properlyfolded and assembled proteins are transported from therough ER to the Golgi complex and ultimately to the cell sur-face or other final destination. Unfolded, misfolded, or partlyfolded and assembled proteins are selectively retained in therough ER. We consider several features of such “quality con-trol” in the latter part of this section.

As discussed previously, N-terminal ER signal sequencesare cleaved from secretory proteins and type I membraneproteins in the ER. Some proteins also undergo other specificproteolytic cleavages in the Golgi complex or forming secre-tory vesicles. We cover these cleavages, as well as carbohy-drate modifications that occur primarily or exclusively in theGolgi complex, in the next chapter.

16.3

16.3 • Protein Modifications, Folding, and Quality Control in the ER 673

X XAsn

GlcNAc

Man

Man

Man

GlcNAc

GlcNAc = N-Acetylglucosamine

Man = Mannose

Glc = Glucose

(Ser/Thr) . . . COO−. . .NH3+

ManManMan

Man ManMan

Glc

Glc

Glc

= Conserved

= Variable

▲ FIGURE 16-16 Common 14-residue precursor of N-linked

oligosaccharides that is added to nascent proteins in the

rough ER. Subsequent removal and in some cases addition ofspecific sugar residues occur in the ER and Golgi complex. Thecore region, composed of five residues highlighted in purple, isretained in all N-linked oligosaccharides. The precursor can belinked only to asparagine (Asn) residues that are separated byone amino acid (X) from a serine (Ser) or threonine (Thr) on thecarboxyl side.

nal that the oligosaccharide is complete and ready to betransferred to a protein.

Oligosaccharide Side Chains May PromoteFolding and Stability of GlycoproteinsThe oligosaccharides attached to glycoproteins serve vari-ous functions. For example, some proteins require N-linkedoligosaccharides in order to fold properly in the ER. Thisfunction has been demonstrated in studies with the antibiotictunicamycin, which blocks the first step in formation of thedolichol-linked precursor of N-linked oligosaccharides (seeFigure 16-17). In the presence of tunicamycin, for instance,the hemagglutinin precursor polypeptide (HA0) is synthe-sized, but it cannot fold properly and form a normal trimer;in this case, the protein remains, misfolded, in the rough ER.Moreover, mutation in the HA sequence of just one asparagine that normally is glycosylated to a glutamineresidue, thereby preventing addition of an N-linked oligosac-charide to that site, causes the protein to accumulate in theER in an unfolded state.

In addition to promoting proper folding, N-linkedoligosaccharides also confer stability on many secreted gly-

(Figure 16-17). The final dolichol pyrophosphoryl oligosac-charide is oriented so that the oligosaccharide portion facesthe ER lumen.

The entire 14-residue precursor is transferred from thedolichol carrier to an asparagine residue on a nascentpolypeptide as it emerges into the ER lumen (Figure 16-18,step 1). Only asparagine residues in the tripeptide sequencesAsn-X-Ser and Asn-X-Thr (where X is any amino acid ex-cept proline) are substrates for oligosaccharyl transferase, theenzyme that catalyzes this reaction. Two of the three sub-units of this enzyme are ER membrane proteins whose cytosol-facing domains bind to the ribosome, localizing athird subunit of the transferase, the catalytic subunit, nearthe growing polypeptide chain in the ER lumen. Not all Asn-X-Ser/Thr sequences become glycosylated; for instance, rapidfolding of a segment of a protein containing an Asn-X-Ser/Thr sequence may prevent transfer of the oligosaccharideprecursor to it.

Immediately after the entire precursor, Glc3Man9(Glc-NAc)2, is transferred to a nascent polypeptide, three differentenzymes remove all three glucose residues and one particularmannose residue (Figure 16-18, steps 2–4). The three glucoseresidues, which are the last residues added during synthesisof the precursor on the dolichol carrier, appear to act as a sig-


Dolicholphosphate

= N-Acetylglucosamine

= Mannose

= GlucoseER lumen

Cytosol

Blockedby tunicamycin

"flip"

P1 2 3 4

5 6

P

P

Completedprecursor

PP

"flip" "flip"

UDP

UMP

UDP

UDP

5 GDP

5 GDP4 GDP

4 GDP3 UDP

3 UDP

P

P P P P P

PPPP P P P P

▲ FIGURE 16-17 Biosynthesis of the dolichol pyrophos-

phoryl oligosaccharide precursor of N-linked oligosaccharides.

Dolichol phosphate is a strongly hydrophobic lipid, containing75–95 carbon atoms, that is embedded in the ER membrane.Two N-acetylglucosamine (GlcNAc) and five mannose residuesare added one at a time to a dolichol phosphate on the cytosolicface of the ER membrane (steps – ). The nucleotide-sugardonors in these and later reactions are synthesized in thecytosol. Note that the first sugar residue is attached to dolicholby a high-energy pyrophosphate linkage. Tunicamycin, whichblocks the first enzyme in this pathway, inhibits the synthesis of

all N-linked oligosaccharides in cells. After the seven-residuedolichol pyrophosphoryl intermediate is flipped to the luminalface (step ), the remaining four mannose and all three glucoseresidues are added one at a time (steps , ). In the laterreactions, the sugar to be added is first transferred from anucleotide-sugar to a carrier dolichol phosphate on the cytosolicface of the ER; the carrier is then flipped to the luminal face,where the sugar is transferred to the growing oligosaccharide,after which the “empty” carrier is flipped back to the cytosolicface. [After C. Abeijon and C. B. Hirschberg, 1992, Trends Biochem. Sci. 17:32.]

1 3

65

4


(Glc)3(Man)9(GlcNAc)2

(Man)8(GlcNAc)2

Dol = Dolichol

= N-Acetylglucosamine

= Mannose

= Glucose

Dol

Tocis-Golgi

P

P

1 2 3

3a

4

ER lumen

3

▲ FIGURE 16-18 Addition and initial processing of N-linked

oligosaccharides in the rough ER of vertebrate cells. TheGlc3Man9(GlcNAc)2 precursor is transferred from the dolicholcarrier to a susceptible asparagine residue on a nascent proteinas soon as the asparagine crosses to the luminal side of the ER(step ). In three separate reactions, first one glucose residue

(step ), then two glucose residues (step ), and finally onemannose residue (step ) are removed. Re-addition of one glucose residue (step ) plays a role in the correct folding ofmany proteins in the ER, as discussed later. [See R. Kornfeld and S. Kornfeld, 1985, Ann. Rev. Biochem. 45:631, and M. Sousa and A. J.Parodi, 1995, EMBO J. 14:4196.]

coproteins. Many secretory proteins fold properly and aretransported to their final destination even if the addition ofall N-linked oligosaccharides is blocked, for example, by tunicamycin. However, such nonglycosylated proteins havebeen shown to be less stable than their glycosylated forms.For instance, glycosylated fibronectin, a normal componentof the extracellular matrix, is degraded much more slowly bytissue proteases than is nonglycosylated fibronectin.

Oligosaccharides on certain cell-surface glycoproteinsalso play a role in cell-cell adhesion. For example, the plasmamembrane of white blood cells (leukocytes) contains cell-adhesion molecules (CAMs) that are extensively glycosylated.The oligosaccharides in these molecules interact with asugar-binding domain in certain CAMs found on endothelialcells lining blood vessels. This interaction tethers the leuko-cytes to the endothelium and assists in their movement intotissues during an inflammatory response to infection (seeFigure 6-30). Other cell-surface glycoproteins possessoligosaccharide side chains that can induce an immune re-sponse. A common example is the A, B, O blood-group anti-gens, which are O-linked oligosaccharides attached toglycoproteins and glycolipids on the surface of erythrocytesand other cell types (see Figure 5-16).

Disulfide Bonds Are Formed and Rearranged by Proteins in the ER LumenIn Chapter 3 we learned that both intramolecular and inter-molecular disulfide bonds (–S–S–) help stabilize the tertiaryand quaternary structure of many proteins. These covalent

bonds form by the oxidative linkage of sulfhydryl groups(–SH), also known as thiol groups, on two cysteine residuesin the same or different polypeptide chains. This reaction canproceed spontaneously only when a suitable oxidant is pres-ent. In eukaryotic cells, disulfide bonds are formed only inthe lumen of the rough ER; in bacterial cells, disulfide bondsare formed in the periplasmic space between the inner andouter membranes. Thus disulfide bonds are found only insecretory proteins and in the exoplasmic domains of mem-brane proteins. Cytosolic proteins and organelle proteinssynthesized on free ribosomes lack disulfide bonds and de-pend on other interactions to stabilize their structures.

The efficient formation of disulfide bonds in the lumen ofthe ER depends on the enzyme protein disulfide isomerase(PDI), which is present in all eukaryotic cells. This enzymeis especially abundant in the ER of secretory cells in such or-gans as the liver and pancreas, where large quantities of pro-teins that contain disulfide bonds are produced. As shownin Figure 16-19a, the disulfide bond in the active site of PDIcan be readily transferred to a protein by two sequentialthiol-disulfide transfer reactions. The reduced PDI generatedby this reaction is returned to an oxidized form by the ac-tion of an ER-resident protein, called Ero1, which carries adisulfide bond that can be transferred to PDI. It is not yet un-derstood how Ero1 itself becomes oxidized. Figure 16-20 de-picts the organization of the pathway for proteindisulfide-bond formation in the ER lumen and the analogouspathway in bacteria.

In proteins that contain more than one disulfide bond,the proper pairing of cysteine residues is essential for nor-mal structure and activity. Disulfide bonds commonly are

1

2 3

4

3a

Most proteins used for therapeutic purposes in hu-mans or animals are secreted glycoproteins stabi-lized by disulfide bonds. When researchers first

tried to synthesize such proteins using plasmid expressionvectors in bacterial cells, the results were disappointing. Inmost cases, the proteins were not secreted (even when a bac-terial signal sequence replaced the normal one); instead, theyaccumulated in the cytosol and often in a denatured state,in part owing to the lack of disulfide bonds. After it becameclear that disulfide-bond formation occurs spontaneouslyonly in the ER lumen, biotechnologists eventually developedexpression vectors that can be used in animal cells (Chapter9). Nowadays, such vectors and cultured animal cells arepreferred for large-scale production of therapeutic proteinssuch as tissue plasminogen activator (an anticlotting agent)and erythropoietin, a hormone that stimulates productionof red blood cells. ❚

formed between cysteines that occur sequentially in theamino acid sequence while a polypeptide is still growing onthe ribosome. Such sequential formation, however, some-times yields disulfide bonds between the wrong cysteines. Forexample, proinsulin has three disulfide bonds that link cys-teines 1 and 4, 2 and 6, and 3 and 5. In this case, disulfidebonds initially formed sequentially (e.g., between cysteines1 and 2) have to be rearranged for the protein to achieve itsproper folded conformation. In cells, the rearrangement ofdisulfide bonds also is accelerated by PDI, which acts on abroad range of protein substrates, allowing them to reachtheir thermodynamically most stable conformations (Figure16-19b). Disulfide bonds generally form in a specific order,first stabilizing small domains of a polypeptide, then stabi-lizing the interactions of more distant segments; this phe-nomenon is illustrated by the folding of influenza HA proteindiscussed in the next section.


Formation of a disulfide bond(a)

SH

OxidizedPDI

S−

SH

S S

Oxidizedsubstrateprotein

ReducedPDI

Rearrangement of disulfide bonds(b)

Protein withincorrect disulfide bonds Protein with

correct disulfide bonds

Reducedsubstrateprotein

SH

S

S

S

S

SH

SH

S S

S S

SH

S−

S

S

S

S

S

S

S

S

S−

SHReduced

PDI SH

SH

12

2

1

3

2

1

ReducedPDI

▲ FIGURE 16-19 Formation and rearrangement of disulfide

bonds by protein disulfide isomerase (PDI). PDI contains anactive site with two closely spaced cysteine residues that areeasily interconverted between the reduced dithiol form and theoxidized disulfide form. Numbered red arrows indicate thesequence of electron transfers. Yellow bars represent disulfidebonds. (a) In the formation of disulfide bonds, the ionized (–S�)form of a cysteine thiol in the substrate protein reacts with thedisulfide (SXS) bond in oxidized PDI to form a disulfide-bonded

PDI–substrate protein intermediate. A second ionized thiol in thesubstrate then reacts with this intermediate, forming a disulfidebond within the substrate protein and releasing reduced PDI. (b) Reduced PDI can catalyze rearrangement of improperlyformed disulfide bonds by similar thiol-disulfide transfer reactions.In this case, reduced PDI both initiates and is regenerated in thereaction pathway. These reactions are repeated until the moststable conformation of the protein is achieved. [See M. M. Lylesand H. F. Gilbert, 1991, Biochemistry 30:619.]

that are unfolded or misfolded. Binding of calnexin and cal-reticulin to unfolded nascent chains prevents aggregation ofadjacent segments of a protein as it is being made on the ER.Thus calnexin and calreticulin, like BiP, help prevent prema-ture, incorrect folding of segments of a newly made protein.

Other important protein-folding catalysts in the ERlumen are peptidyl-prolyl isomerases, a family of enzymesthat accelerate the rotation about peptidyl-prolyl bonds inunfolded segments of a polypeptide:

Chaperones and Other ER Proteins FacilitateFolding and Assembly of ProteinsAlthough many reduced, denatured proteins can sponta-neously refold into their native state in vitro, such refoldingusually requires hours to reach completion. Yet new solubleand membrane proteins produced in the ER generally foldinto their proper conformation within minutes after theirsynthesis. The rapid folding of these newly synthesized pro-teins in cells depends on the sequential action of several pro-teins present within the ER lumen.

We have already seen how the chaperone BiP can drivepost-translational translocation in yeast by binding fully syn-thesized polypeptides as they enter the ER (see Figure 16-9).BiP can also bind transiently to nascent chains as they enterthe ER during cotranslational translocation. Bound BiP isthought to prevent segments of a nascent chain from mis-folding or forming aggregates, thereby promoting folding ofthe entire polypeptide into the proper conformation. Proteindisulfide isomerase (PDI) also contributes to proper folding,which is stabilized by disulfide bonds in many proteins.

As illustrated in Figure 16-21, two other ER proteins, thehomologous lectins (carbohydrate-binding proteins) calnexinand calreticulin, bind selectively to certain N-linked oligosac-charides on growing nascent chains. The ligand for these twolectins, which contains a single glucose residue, is generatedby a specific glucosyltransferase in the ER lumen (see Figure16-18, step 3a). This enzyme acts only on polypeptide chains


12

3

1

3

4

S

S

S S

Periplasmicspace


Outer membrane

Ubiquinone

S−

DsbA

DsbBInner membrane

S−

(b) Bacteria


PDI

S− S−

S

S

S S

Ero1

ER lumen

ER membrane

(a) Eukaryotes

2

▲ FIGURE 16-20 Pathways for the formation of disulfide

bonds in eukaryotes and bacteria. Numbered red arrowsindicate the sequence of electron flow. A disulfide bond (yellowbar) is formed by loss of a pair of electrons from cysteine thiol(–SH) groups. (a) In the ER lumen of eukaryotic cells, electronsfrom an ionized thiol in a newly synthesized substrate protein aretransferred to a disulfide bond in the active site of PDI (seeFigure 16-19). PDI, in turn, transfers electrons to a disulfide bond

in the luminal protein Ero1, thereby regenerating the oxidizedform of PDI. The mechanism of reoxidation of Ero1 is not known.(b) In the periplasmic space of bacterial cells, the soluble proteinDsbA functions like eukaryotic PDI. Reduced DsbA is reoxidizedby DsbB, an inner-membrane protein. Finally, reduced DsbB isreoxidized by transferring electrons to oxidized ubiquinone, a lipidcofactor of the electron-transport chain in the inner membrane(see Chapter 8). [See A. R. Frand et al., 2000, Trends Cell Biol. 10:203.]

OC

CO NH

CH2N

CH2HC2

CH

C

OC CH2N

CH2HC

CH2

O

trans

cis

Prolyl

ProlylRotation about

peptide bond

NH

Such isomerizations sometimes are the rate-limiting stepin the folding of protein domains. Many peptidyl-prolyl iso-merases can catalyze the rotation of exposed peptidyl-prolylbonds indiscriminately in numerous proteins, but some havevery specific protein substrates.

Many important secretory and membrane proteins syn-thesized on the ER are built of two or more polypeptide sub-units. In all cases, the assembly of subunits constituting these

chase period, membranes are solubilized by detergent andexposed to monoclonal antibodies specific for either HA0

monomer or trimer. Immediately after the pulse, themonomer-specific antibody is able to immunoprecipitate allthe radioactive HA0 protein. During the chase period, in-creasing proportions of the total radioactive HA0 protein in-stead react with the trimer-specific monoclonal antibody.Such experiments have shown that each newly made HA0

polypeptide requires approximately 10 minutes to fold andbe incorporated into a trimer in living cells.

Improperly Folded Proteins in the ER InduceExpression of Protein-Folding CatalystsWild-type proteins that are synthesized on the rough ER can-not exit this compartment until they achieve their completelyfolded conformation. Likewise, almost any mutation thatprevents proper folding of a protein in the ER also blocksmovement of the polypeptide from the ER lumen or mem-brane to the Golgi complex. The mechanisms for retainingunfolded or incompletely folded proteins within the ERprobably increase the overall efficiency of folding by keepingintermediate forms in proximity to folding catalysts, whichare most abundant in the ER. Improperly folded proteins re-tained within the ER generally are found permanently bound


multisubunit (multimeric) proteins occurs in the ER. An im-portant class of multimeric secreted proteins is the im-munoglobulins, which contain two heavy (H) and two light(L) chains, all linked by intrachain disulfide bonds. Hemag-glutinin (HA) is another multimeric protein that provides agood illustration of folding and subunit assembly (see Fig-ure 16-21). This trimeric protein forms the spikes that pro-trude from the surface of an influenza virus particle. HA isformed within the ER of an infected host cell from threecopies of a precursor protein termed HA0, which has a singlemembrane-spanning � helix. In the Golgi complex, each ofthe three HA0 proteins is cleaved to form two polypeptides,HA1 and HA2; thus each HA molecule that eventually re-sides on the viral surface contains three copies of HA1 andthree of HA2 (see Figure 3-7). The trimer is stabilized by in-teractions between the large exoplasmic domains of the con-stituent polypeptides, which extend into the ER lumen; afterHA is transported to the cell surface, these domains extendinto the extracellular space. Interactions between the smallercytosolic and membrane-spanning portions of the HA sub-units also help stabilize the trimeric protein.

The time course of HA0 folding and assembly in vivo canbe determined by pulse-labeling experiments. In a typical ex-periment, virus-infected cells are pulse-labeled with a ra-dioactive amino acid; at various times during the subsequent

Membrane-spanningα helix

Luminalα helix

CompletedHA0 monomer

1a2 3

HA0 trimer

1b

Calnexin

Calreticulin

ER lumen

Oligosaccharyl transferase

Dolichololigosaccharide

BiP

S SPDI

SH

Cytosol

▲ FIGURE 16-21 Folding and assembly of hemagglutinin

(HA0) trimer in the ER. Transient binding of the chaperone BiP(step ) to the nascent chain and of two lectins, calnexin andcalreticulin, to certain oligosaccharide chains (step ) promotesproper folding of adjacent segments. A total of seven N-linkedoligosaccharide chains are added to the luminal portion of thenascent chain during cotranslational translocation, and PDIcatalyzes the formation of six disulfide bonds per monomer.Completed HA0 monomers are anchored in the membrane by a

single membrane-spanning � helix with their N-terminus in thelumen (step ). Interaction of three HA0 chains with one another, initially via their transmembrane � helices, apparently triggersformation of a long stem containing one � helix from the luminalpart of each HA0 polypeptide. Finally, interactions between the three globular heads occur, generating a stable HA0 trimer(step ). [See U. Tatu et al., 1995, EMBO J. 14:1340, and D. Hebert etal., 1997, J. Cell Biol. 139:613.]

1a

1b

2

3

tion of a transcription factor by such regulated intramem-brane proteolysis also occurs in the Notch signaling pathway(see Figure 14-29). We will encounter another example ofthis phenomenon, the cholesterol-responsive transcriptionfactor SREBP, in Chapter 18.

The hereditary form of emphysema illustrates thedetrimental effects that can result from misfoldingof proteins in the ER. This disease is caused by a

point mutation in �1-antitrypsin, which normally is secretedby hepatocytes and macrophages. The wild-type protein bindsto and inhibits trypsin and also the blood protease elastase.In the absence of �1-antitrypsin, elastase degrades the fine tis-sue in the lung that participates in the absorption of oxygen,eventually producing the symptoms of emphysema. Althoughthe mutant �1-antitrypsin is synthesized in the rough ER, itdoes not fold properly, forming an almost crystalline aggregatethat is not exported from the ER. In hepatocytes, the secre-tion of other proteins also becomes impaired, as the rough ERis filled with aggregated �1-antitrypsin. ❚

Unassembled or Misfolded Proteins in the ER AreOften Transported to the Cytosol for DegradationMisfolded secretory and membrane proteins, as well as theunassembled subunits of multimeric proteins, often are de-graded within an hour or two after their synthesis in therough ER. For many years, researchers thought that prote-olytic enzymes within the ER catalyzed degradation of mis-folded or unassembled polypeptides, but such proteases werenever found. More recent studies have shown that misfoldedmembrane and secretory proteins are transported from theER lumen, “backwards” through the translocon, into the cy-tosol, where they are degraded by the ubiquitin-mediatedproteolytic pathway (see Figure 3-13).

Ubiquitinylating enzymes localized to the cytosolic faceof the ER add ubiquitin to misfolded ER proteins as they exitthe ER. This reaction, which is coupled to hydrolysis of ATP,


to the ER chaperones BiP and calnexin. Thus these luminalfolding catalysts perform two related functions: assisting inthe folding of normal proteins by preventing their aggrega-tion and binding to irreversibly misfolded proteins.

Both mammalian cells and yeasts respond to the presenceof unfolded proteins in the rough ER by increasing tran-scription of several genes encoding ER chaperones and otherfolding catalysts. A key participant in this unfolded-proteinresponse is Ire1, an ER membrane protein that exists as botha monomer and a dimer. The dimeric form, but not themonomeric form, promotes formation of Hac1, a transcrip-tion factor in yeast that activates expression of the genes in-duced in the unfolded-protein response. As depicted inFigure 16-22, binding of BiP to the luminal domain ofmonomeric Ire1 prevents formation of the Ire1 dimer. Thusthe quantity of free BiP in the ER lumen probably determinesthe relative proportion of monomeric and dimeric Ire1. Ac-cumulation of unfolded proteins within the ER lumen se-questers BiP molecules, making them unavailable for bindingto Ire1. As a result the level of dimeric Ire1 increases, lead-ing to an increase in the level of Hac1 and production of pro-teins that assist in protein folding.

Mammalian cells contain an additional regulatory path-way that operates in response to unfolded proteins in the ER.In this pathway, accumulation of unfolded proteins in the ERtriggers proteolysis of ATF6, a transmembrane protein in theER membrane. The cytosolic domain of ATF6 released byproteolysis then moves to the nucleus, where it stimulatestranscription of the genes encoding ER chaperones. Activa-

UnsplicedHac1 mRNA

SplicedHac1 mRNA

Endonuclease-cut Hac1 mRNA

ERlumen

Endo-nuclease

Ire1monomer

Unfolded proteinswith BiP bound

Hac1 transcriptionfactor

Ire1 dimer

BiP

2

1

3

4

Unfolded proteins

Translation

Cytosol

▲ FIGURE 16-22 The unfolded-protein response. Ire1, atransmembrane protein in the ER membrane, has a binding site forBiP on its luminal domain; the cytosolic domain contains a specificRNA endonuclease. Step : Accumulating unfolded proteins in theER lumen bind BiP molecules, releasing them from monomericIre1. Dimerization of Ire1 then activates its endonuclease activity.Steps , : The unspliced mRNA precursor encoding thetranscription factor Hac1 is cleaved by dimeric Ire1, and the two exons are joined to form functional Hac1 mRNA. Current evidenceindicates that this processing occurs in the cytosol, although pre-mRNA processing generally occurs in the nucleus. Step :Hac1 is translated into Hac1 protein, which then moves back intothe nucleus and activates transcription of genes encoding severalprotein-folding catalysts. [See U. Ruegsegger et al., 2001, Cell 107:103;A. Bertolotti et al., 2000, Nature Cell Biol. 2:326; and C. Sidrauski and P. Walter, 1997, Cell 90:1031.]

▲

1

32

4

Export of Bacterial ProteinsThe cell wall surrounding gram-negative bacteria comprisesan inner membrane, which is the main permeability barrierfor the cytoplasm, a periplasmic space containing variousproteins and a layer of peptidoglycan, and an outer mem-brane, which is permeable to small molecules but not pro-teins. The peptidoglycan layer gives the cell wall its strength,while the periplasmic proteins function in sensing and im-porting extracellular molecules and in assembling and main-taining the structural integrity of the cell wall. These proteins(like all bacterial proteins) are synthesized on cytosolic ribo-somes and then are translocated in the unfolded state acrossthe inner membrane (also called the cytoplasmic membrane).Most proteins translocated across the inner membrane re-main associated with the bacterial cell, either as a membraneprotein inserted into the outer or inner membrane or trappedwithin the periplasmic space. Some bacterial species also pos-sess specialized translocation systems that enable proteins tomove through both membranes of the cell wall into the ex-tracellular space. In this section, we describe both types ofprotein secretion in gram-negative bacteria.

Cytosolic SecA ATPase Pushes BacterialPolypeptides Through Translocons into the Periplasmic SpaceThe mechanism for translocating bacterial proteins acrossthe inner membrane shares several key features with thetranslocation of proteins into the ER of eukaryotic cells.First, translocated proteins usually contain an N-terminalhydrophobic signal sequence, which is cleaved by a signalpeptidase. Second, bacterial proteins pass through the innermembrane in a channel, or translocon, composed of proteinsthat are structurally similar to the eukaryotic Sec61 complex.Third, bacterial cells express two proteins, Ffh and its recep-tor (FtsY), that are homologs of the SRP and SRP receptor,respectively. In bacteria, however, these latter proteins ap-pear to function mainly in the insertion of hydrophobicmembrane proteins into the inner membrane. Indeed, all bac-terial proteins that are translocated across the inner membranedo so only after their synthesis in the cytosol is completed butbefore they are folded into their final conformation.

The post-translational translocation of bacterial pro-teins across the inner membrane cannot involve a ratchetmechanism similar to that mediated by BiP in the ER lumen(see Figure 16-9) because the ATP required for such amechanism would be lost by diffusion through the outermembrane. Rather, the driving force for translocation ofbacterial proteins is generated by SecA, which binds to thecytosolic side of the translocon and hydrolyzes cytosolicATP. In the model depicted in Figure 16-23, SecA binds tothe unfolded translocating polypeptide, and then a confor-mational change in SecA, driven by the energy released

16.4


may provide some of the energy required to drag these pro-teins back to the cytosol. The resulting polyubiquitinylatedpolypeptides are quickly degraded in proteasomes. Exactlyhow misfolded soluble and membrane proteins in the ER arerecognized and targeted to the translocon for export to thecytosol is not known.


Protein Modifications, Folding, and Quality Control in the ER

■ All N-linked oligosaccharides, which are bound to as-paragine residues, contain a core of three mannose and twoN-acetylglucosamine residues and usually have severalbranches (see Figure 16-16).

■ O-linked oligosaccharides, which are bound to serine orthreonine residues, are generally short, often containingonly one to four sugar residues.

■ Formation of all N-linked oligosaccharides begins withassembly of a ubiquitous 14-residue high-mannose pre-cursor on dolichol, a lipid in the membrane of the roughER (see Figure 16-17). After this preformed oligosaccha-ride is transferred to specific asparagine residues of na-scent polypeptide chains in the ER lumen, three glucoseresidues and one mannose residue are removed (see Figure16-18).

■ Oligosaccharide side chains may assist in the properfolding of glycoproteins, help protect the mature proteinsfrom proteolysis, participate in cell-cell adhesion, and func-tion as antigens.

■ Disulfide bonds are added to many secretory proteinsand the exoplasmic domain of membrane proteins in theER. Protein disulfide isomerase (PDI), present in the ERlumen, catalyzes both the formation and the rearrangementof disulfide bonds (see Figure 16-19).

■ The chaperone BiP, the lectins calnexin and calreticulin,and peptidyl-prolyl isomerases work together to assureproper folding of newly made secretory and membraneproteins in the ER. The subunits of multimeric proteinsalso assemble in the ER.

■ Only properly folded proteins and assembled subunitsare transported from the rough ER to the Golgi complexin vesicles.

■ The accumulation of abnormally folded proteins andunassembled subunits in the ER can induce increased ex-pression of ER protein-folding catalysts via the unfolded-protein response (see Figure 16-22).

■ Unassembled or misfolded proteins in the ER often aretransported back through the translocon to the cytosol,where they are degraded in the ubiquitin/proteasomepathway.

from ATP hydrolysis, acts to push the bound polypeptidesegment through the translocon pore toward the periplas-mic side of the membrane. Repetition of this cycle eventu-ally pushes the entire polypeptide chain through thetranslocon into the periplasmic space, where disulfidebonds are formed and the polypeptide folds into its properconformation.

Several Mechanisms Translocate BacterialProteins into the Extracellular SpaceQuite different mechanisms from the one shown in Figure16-23 are used to translocate bacterial proteins from the cy-tosol across both the inner and outer bacterial membranes tothe extracellular space. These secretion mechanisms are par-ticularly important for pathogenic bacteria, which com-monly use secreted extracellular proteins to colonize specifictissues within the host and to evade host defense mecha-nisms. Well-known examples of extracellular proteins thatpromote the growth and dissemination of pathogenic bacte-ria include protein toxins (e.g., cholera toxin and tetanustoxin) and pili, which are proteinaceous fibers that projectfrom the outer membrane and assist enteric bacteria in ad-hering to the epithelium of the gut.

The numerous specialized bacterial secretion systems that have been identified can be classified into four generaltypes based on their mechanism of operation. Both the type Iand the type II secretion systems involve two steps. First, substrate proteins are translocated across the inner mem-brane into the periplasmic space, where they fold and oftenacquire disulfide bonds. Second, the folded proteins aretranslocated from the periplasmic space across the outer

membrane by complexes of periplasmic proteins that spanthe inner and outer membranes. The energy for this translo-cation comes from hydrolysis of ATP in the cytosol, but themechanisms that couple ATP hydrolysis and translocationacross the outer membrane are not well understood.

Translocation by the type III and type IV secretion sys-tems, on the other hand, entails a single step. These systemsconsist of large protein complexes that span both mem-branes, allowing proteins to be translocated directly from thecytosol to the extracellular environment. The type III systemis adapted not only for secreting proteins but also for inject-ing them into target cells, a very useful property for patho-genic bacteria.

Pathogenic Bacteria Can Inject Proteins intoAnimal Cells via Type III Secretion Apparatus

Yersinia pestis is the bacterial species responsiblefor the bubonic plague, one of the deadliest dis-eases in human history. One reason Yersinia is such

a virulent pathogen lies in its ability to disable hostmacrophage cells that might otherwise engulf and destroythe invading bacterial cells. The incapacitating effect ofYersinia is mediated primarily by a small set of proteins thatthe bacterial cells inject into macrophage cells.

Various pathogenic bacteria inject proteins into host cellsvia a complicated syringe-like machine composed of morethan 20 different proteins. This type III secretion apparatus,shown in Figure 16-24, has ringlike components embeddedin both the inner and outer membranes of the bacterial cellwall and a hollow needlelike structure (pilus) that projects

16.4 • Export of Bacterial Proteins 681

Cytosol

Periplasmic

space Translocon(SecY, SecE, SecG)

SecA

121

ATP ADP + PiATP ADP + Pi ATP

1 2

Inner

membrane

ATPATP ATP

2

▲ FIGURE 16-23 Post-translational translocation across

inner membrane in gram-negative bacteria. The bacterial innermembrane contains a translocon channel composed of threesubunits that are homologous to the components of theeukaryotic Sec61 complex. Translocation of polypeptides from thecytosol to the periplasmic space is powered by SecA, a cytosolicATPase that binds to the translocon and to the translocatingpolypeptide. In the model shown here, binding and hydrolysis ofATP cause conformational changes in SecA that push the bound

polypeptide segment through the channel (steps , ).Repetition of this cycle results in movement of the polypeptidethrough the channel in one direction. Current evidence indicatesthat the N-terminal signal sequence moves from the channel intothe bilayer but at some point is cleaved by a signal peptidase, sothat the mature polypeptide enters the periplasmic space. [See A. Economou and W. Wickner, 1994, Cell 78:835, and J. Eichler and W. Wickner, 1998, J. Bacteriol. 180:5776.]

1 2

out from the outer membrane. Proteins at the outer (distal)end of the pilus can penetrate the plasma membrane of cer-tain mammalian cells, thus creating a conduit that spansthree membranes, providing a connection between the bac-terial cytoplasm and that of the target host cell.

A key insight into how the type III secretion apparatusmay operate came from the observation that many of thecomponents of the secretion apparatus are homologous toproteins in the base of the bacterial flagellum. The flagellarbase is located in the inner membrane and functions as amotor to drive rotation of the attached flagellum, which ex-tends outward from the cell surface. The flagellum, which isa hollow helical tube made up entirely of a repeating poly-mer of the protein flagellin, grows by addition of new fla-gellin subunits at the distal tip. Several proteins in theflagellar base are thought to use the energy from ATP hy-drolysis to push new flagellin subunits through the centralchannel of the flagellum toward the distal end. Quite possi-bly, the type III secretion apparatus uses a similar ATP-drivenmechanism to push proteins through the central channel ofthe pilus into target cells.

Recent experiments have identified the signal sequencesthat target bacterial proteins for transport through the typeIII secretion apparatus. For instance, recombinant DNAmethods were used to express in Yersinia cells chimeric pro-teins containing adenylate cyclase attached to different por-tions of YopE, which normally is secreted via the type IIIapparatus. An amphipathic sequence at the N-terminus ofYopE was capable of directing adenylate cyclase, which nor-mally resides in the cytosol, to the type III secretion appara-tus for injection into mammalian cells. Other experimentsshowed that YopE proteins in which this amphipathic tar-


geting sequence is mutated still are secreted normally. Thissurprising finding eventually lead to discovery of a second,independent targeting signal that allows YopE to bind to asmall chaperone molecule. The chaperone-YopE complexcan be successfully secreted even in the absence of an N-terminal amphipathic sequence. Each of the proteins secretedby the type III apparatus is thought to interact with one of aset of small chaperone proteins (see Figure 16-24). Thesechaperones may act to keep secreted proteins in a partiallyunfolded state as they pass through the central channel of thetype III apparatus. Once the transported proteins have beenreleased into the target cell, folding can be completed by cy-tosolic target-cell chaperones. ❚


Export of Bacterial Proteins

■ Gram-negative bacteria translocate completed proteinsacross the inner membrane through a translocon related tothe ER translocon of eukaryotic cells.

■ The driving force for post-translational translocationacross the inner membrane of bacteria comes from SecAprotein, which uses energy derived from hydrolysis of cy-tosolic ATP to push polypeptides through the transloconchannel (see Figure 16-23).

■ Four bacterial secretion systems exist for translocatingproteins from the cytosol across both the inner and outermembranes. Hydrolysis of cytosolic ATP drives secretionin all the systems.

Secretedprotein

Needle/pilus

Eukaryotic

cytosol

(a)Targetingsequence

Chaperone(b)

Bacterial

cytosol

Plasma membrane

Inner membrane

Periplasmic space

Outer membrane

▲ FIGURE 16-24 Type III secretion

apparatus for injecting bacterial proteins

into eukaryotic cells. (a) Schematicdiagram of the type III secretion apparatus,which is similar in size and morphology tothe bacterial flagellum. Bacterial proteinswith targeting sequences (red) that allowinteraction with specialized chaperones(orange) enter the cytosol-facing portion ofthe type III secretion apparatus, travel downthe hollow core of the pilus in an ATP-dependent process, and ultimately aredelivered to the cytoplasm of the targeteukaryotic cell. (b) Electron micrograph ofisolated type III secretion apparatuses. Longneedlelike pili can be seen extending fromthe widened basal portions, which areembedded in the outer and innermembranes. See the text for discussion.[Part (a) adapted from D. G. Thanassi and S. J.Hultgren, 2000, Curr. Opin. Cell Biol. 12:420. Part(b) from T. Kubori et al., 1998, Science 280:602.]

▲

■ The type III secretion apparatus, used by pathogenic bac-teria to inject proteins into eukaryotic cells, consists of abasal portion that spans both membranes and an extra-cellular needlelike structure that can penetrate the plasmamembrane of a target cell (see Figure 16-24).

Sorting of Proteins toMitochondria and ChloroplastsIn the remainder of this chapter, we examine how proteinssynthesized on cytosolic ribosomes are sorted to mitochon-dria, chloroplasts, and peroxisomes (see Figure 16-1). Bothmitochondria and chloroplasts are surrounded by a doublemembrane and have internal subcompartments, whereasperoxisomes are bounded by a single membrane and have asingle luminal compartment known as the matrix. Because ofthese and other differences, we consider peroxisomes sepa-rately in the last section.

Besides being bounded by two membranes, both mito-chondria and chloroplasts also contain similar types of electron-transport proteins and use an F-class ATPase to synthesize ATP (see Figure 8-2). Remarkably, gram-negativebacteria also exhibit these characteristics. Also like bacterialcells, mitochondria and chloroplasts contain their ownDNA, which encodes organelle rRNAs, tRNAs, and someproteins (Chapter 10). Moreover, growth and division of mi-tochondria and chloroplasts are not coupled to nuclear divi-sion. Rather, these organelles grow by the incorporation ofcellular proteins and lipids, and new organelles form by di-vision of preexisting organelles—both processes occurring

16.5

continuously during the interphase period of the cell cycle.The numerous similarities of free-living bacterial cells withmitochondria and chloroplasts have led scientists to hypoth-esize that these organelles arose by the incorporation of bac-teria into ancestral eukaryotic cells, forming endosymbioticorganelles. Striking evidence for this ancient evolutionary re-lationship can be found in the many proteins of similar se-quences shared by mitochondria, chloroplasts, and bacteria,including some of the proteins involved in membranetranslocation described in this section.

Proteins encoded by mitochondrial DNA or chloroplastDNA are synthesized on ribosomes within the organelles anddirected to the correct subcompartment immediately aftersynthesis. The majority of proteins located in mitochondriaand chloroplasts, however, are encoded by genes in the nu-cleus and are imported into the organelles after their synthe-sis in the cytosol. Apparently over eons of evolution muchof the genetic information from the ancestral bacterial DNAin these endosymbiotic organelles moved, by an unknownmechanism, to the nucleus. Precursor proteins synthesizedin the cytosol that are destined for the matrix of mitochon-dria or the equivalent space, the stroma, of chloroplasts usu-ally contain specific N-terminal uptake-targeting sequencesthat specify binding to receptor proteins on the organelle sur-face. Generally, this sequence is cleaved once it reaches thematrix or stroma. Clearly, these uptake-targeting sequencesare similar in their location and general function to the signalsequences that direct nascent proteins to the ER lumen. Al-though the three types of signals share some common se-quence features, their specific sequences differ considerably,as summarized in Table 16-1.

16.5 • Sorting of Proteins to Mitochondria and Chloroplasts 683

TABLE 16-1 Uptake-Targeting Sequences That Direct Proteins from the Cytosol to Organelles*

Location of Sequence Removal of Target Organelle Within Protein Sequence Nature of Sequence

Endoplasmic N-terminus Yes Core of 6–12 hydrophobic amino reticulum (lumen) acids, often preceded by one or

more basic amino acids (Arg, Lys)

Mitochondrion N-terminus Yes Amphipathic helix, 20–50 residues(matrix) in length, with Arg and Lys

residues on one side and hydrophobic residues on the other

Chloroplast N-terminus Yes No common motifs; generally rich(stroma) in Ser, Thr, and small hydrophobic

residues and poor in Glu and Asp

Peroxisome (matrix) C-terminus (most No PTS1 signal (Ser-Lys-Leu) at proteins); N-terminus extreme C-terminus; PTS2 (few proteins) signal at N-terminus

*Different or additional sequences target proteins to organelle membranes and subcompartments. See Chapter 12 for targeting sequences required for uptake of proteins into the nucleus.


1 Yeast mitochondrialproteins made bycytoplasmic ribosomesin a cell-free system

Uptake-targetingsequence andmitochondrial

protein degraded

Protein taken upinto mitochondria;uptake-targetingsequence removedand degraded

Proteins seques-tered withinmitochondriaare resistant totrypsin

Uptake-targetingsequence

Mitochondrialprotein

Add energizedyeastmitochondria

Trypsin

Trypsin

2 3a

3b

EXPERIMENTAL FIGURE 16-25 The

post-translational uptake of precursor

proteins into mitochondria can be

assayed in a cell-free system. Theimported protein must contain anappropriate uptake-targeting sequence.Uptake also requires ATP and a cytosolicextract containing chaperone proteins thatmaintain the precursor proteins in anunfolded conformation. Protein uptakeoccurs only with energized (respiring)mitochondria, which have a protonelectrochemical gradient (proton-motiveforce) across the inner membrane. Thisassay has been used to study targetingsequences and other features of thetranslocation process.

In both mitochondria and chloroplasts, protein importrequires energy and occurs at points where the outer andinner organelle membranes are in close contact. Because mi-tochondria and chloroplasts contain multiple membranesand membrane-limited spaces, sorting of many proteins totheir correct location often requires the sequential action oftwo targeting sequences and two membrane-bound translo-cation systems: one to direct the protein into the organelle,and the other to direct it into the correct organellar com-partment or membrane. As we will see, the mechanisms forsorting various proteins to mitochondria and chloroplastsare related to some of the mechanisms discussed previously.

Amphipathic N-Terminal Signal Sequences DirectProteins to the Mitochondrial MatrixAll proteins that travel from the cytosol to the same mito-chondrial destination have targeting signals that share com-mon motifs, although the signal sequences are generally notidentical. Thus the receptors that recognize such signals areable to bind to a number of different but related sequences.The most extensively studied sequences for localizing pro-teins to mitochondria are the matrix-targeting sequences.These sequences, located at the N-terminus, are usually20–50 amino acids in length. They are rich in hydrophobicamino acids, positively charged basic amino acids (arginineand lysine), and hydroxylated ones (serine and threonine),but tend to lack negatively charged acidic residues (aspar-tate and glutamate).

Mitochondrial matrix-targeting sequences are thought toassume an �-helical conformation in which positively

charged amino acids predominate on one side of the helixand hydrophobic amino acids predominate on the other side;thus these sequences are amphipathic. Mutations that dis-rupt this amphipathic character usually disrupt targeting tothe matrix, although many other amino acid substitutions donot. These findings indicate that the amphipathicity of ma-trix-targeting sequences is critical to their function.

The cell-free assay outlined in Figure 16-25 has beenwidely used in studies on the import of mitochondrial precur-sor proteins. In this system, respiring (energized) mitochondriaextracted from cells can incorporate mitochondrial precursorproteins carrying appropriate uptake-targeting sequences thathave been separately synthesized in the absence of mitochon-dria. Successful incorporation of the precursor into the or-ganelle can be assayed either by resistance to digestion by anexogenously added protease or, in most cases, by cleavage ofthe N-terminal targeting sequences by specific proteases. Theuptake of completely synthesized mitochondrial precursorproteins by the organelle in this system contrasts with the cell-free cotranslational translocation of secretory proteins, whichgenerally occurs only when microsomal membranes are pres-ent during synthesis (see Figure 16-4).

Mitochondrial Protein Import Requires Outer-Membrane Receptors and Translocons in Both MembranesFigure 16-26 presents an overview of protein import fromthe cytosol into the mitochondrial matrix, the route into themitochondrion followed by most imported proteins. We willdiscuss in detail each step in protein transport into the matrix

▲

and then consider how some proteins subsequently are tar-geted to other compartments of the mitochondrion.

After synthesis in the cytosol, the soluble precursors ofmitochondrial proteins (including hydrophobic integralmembrane proteins) interact directly with the mitochondrialmembrane. In general, only unfolded proteins can be im-ported into the mitochondrion. Chaperone proteins such ascytosolic Hsc70 keep nascent and newly made proteins in anunfolded state, so that they can be taken up by mitochon-dria. Import of an unfolded mitochondrial precursor is ini-tiated by the binding of a mitochondrial targeting sequenceto an import receptor in the outer mitochondrial membrane.These receptors were first identified by experiments inwhich antibodies to specific proteins of the outer mitochon-drial membrane were shown to inhibit protein import into

isolated mitochondria. Subsequent genetic experiments, inwhich the genes for specific mitochondrial outer-membraneproteins were mutated, showed that specific receptor pro-teins were responsible for the import of different classes ofmitochondrial proteins. For example, N-terminal matrix-targeting sequences are recognized by Tom20 and Tom22.(Proteins in the outer mitochondrial membrane involved intargeting and import are designated Tom proteins fortranslocon of the outer membrane.)

The import receptors subsequently transfer the precur-sor proteins to an import channel in the outer membrane.This channel, composed mainly of the Tom40 protein, isknown as the general import pore because all known mito-chondrial precursor proteins gain access to the interior com-partments of the mitochondrion through this channel. When


Cytosol

Outer membrane

Intermembrane

space

Inner membrane

Precursorprotein

Cytosolic Hsc70

ADP + Pi

ATP

Matrix-targeting sequence

NH3+

COO−

Active protein

1

2

MatrixHsc70

ADP + Pi

ATP

Import receptor

General import pore(Tom40)

3

4

Contact siteTim23/17

Cleaved targetingsequence

Matrix processingprotease

Tim44

ADP + Pi

ATP

7

5

6

Mitochondrial matrix

FIGURE 16-26 Protein import into

the mitochondrial matrix. Precursorproteins synthesized on cytosolicribosomes are maintained in an unfolded orpartially folded state by bound chaperones,such as Hsc70 (step ). After a precursorprotein binds to an import receptor near asite of contact with the inner membrane(step ), it is transferred into the generalimport pore (step ). The translocatingprotein then moves through this channeland an adjacent channel in the innermembrane (steps , ). Note thattranslocation occurs at rare “contact sites”at which the inner and outer membranesappear to touch. Binding of thetranslocating protein by the matrixchaperone Hsc70 and subsequent ATPhydrolysis by Hsc70 helps drive import intothe matrix. Once the uptake-targetingsequence is removed by a matrix proteaseand Hsc70 is released from the newlyimported protein (step ), it folds into themature, active conformation within thematrix (step ). Folding of some proteinsdepends on matrix chaperonins. See thetext for discussion. [See G. Schatz, 1996, J. Biol. Chem. 271:31763, and N. Pfanner et al.,1997, Ann. Rev. Cell Devel. Biol. 13:25.]

54

3

2

1

7

6

▲

purified and incorporated into liposomes, Tom40 forms atransmembrane channel with a pore wide enough to accom-modate an unfolded polypeptide chain. The general importpore forms a largely passive channel through the outer mi-tochondrial membrane, and the driving force for unidirec-tional transport into mitochondria comes from within themitochondrion. In the case of precursors destined for themitochondrial matrix, transfer through the outer membraneoccurs simultaneously with transfer through an inner-membrane channel composed of the Tim23 and Tim17 pro-teins. (Tim stands for translocon of the inner membrane.)Translocation into the matrix thus occurs at “contact sites”where the outer and inner membranes are in close proximity.

Soon after the N-terminal matrix-targeting sequence ofa protein enters the mitochondrial matrix, it is removed bya protease that resides within the matrix. The emerging pro-tein also is bound by matrix Hsc70, a chaperone that is lo-calized to the translocation channels in the innermitochondrial membrane by interacting with Tim44. This in-teraction stimulates ATP hydrolysis by matrix Hsc70, andtogether these two proteins are thought to power transloca-tion of proteins into the matrix.

Some imported proteins can fold into their final, activeconformation without further assistance. Final folding ofmany matrix proteins, however, requires a chaperonin. Asdiscussed in Chapter 3, chaperonin proteins actively facilitateprotein folding in a process that depends on ATP. For in-stance, yeast mutants defective in Hsc60, a chaperonin in themitochondrial matrix, can import matrix proteins and cleavetheir uptake-targeting sequence normally, but the importedpolypeptides fail to fold and assemble into the native terti-ary and quaternary structures.

Studies with Chimeric Proteins DemonstrateImportant Features of Mitochondrial ImportDramatic evidence for the ability of mitochondrial matrix-targeting sequences to direct import was obtained withchimeric proteins produced by recombinant DNA tech-niques. For example, the matrix-targeting sequence of alco-hol dehydrogenase can be fused to the N-terminus ofdihydrofolate reductase (DHFR), which normally resides inthe cytosol. In the presence of chaperones, which prevent theC-terminal DHFR segment from folding in the cytosol, cell-free translocation assays show that the chimeric protein istransported into the matrix (Figure 16-27a). The inhibitormethotrexate, which binds tightly to the active site of DHFRand greatly stabilizes its folded conformation, renders thechimeric protein resistant to unfolding by cytosolic chaper-ones. When translocation assays are performed in the pres-ence of methotrexate, the chimeric protein does notcompletely enter the matrix. This finding demonstrates thata precursor must be unfolded in order to traverse the importpores in the mitochondrial membranes.

Additional studies revealed that if a sufficiently longspacer sequence separates the N-terminal matrix-targeting se-


quence and DHFR portion of the chimeric protein, then a sta-ble translocation intermediate forms in the presence ofmethotrexate (Figure 16-27b). In order for such a stabletranslocation intermediate to form, the spacer sequence mustbe long enough to span both membranes; a spacer of 50amino acids extended to its maximum possible length is ade-quate to do so. If the chimera contains a shorter spacer—say,35 amino acids—no stable translocation intermediate is ob-tained because the spacer cannot span both membranes.These observations provide further evidence that translocatedproteins can span both inner and outer mitochondrial mem-branes and traverse these membranes in an unfolded state.

Microscopic studies of stable translocation intermediatesshow that they accumulate at sites where the inner and outermitochondrial membranes are close together, evidence thatprecursor proteins enter only at such sites (Figure 16-27c).The distance from the cytosolic face of the outer membraneto the matrix face of the inner membrane at these contactsites is consistent with the length of an unfolded spacer se-quence required for formation of a stable translocation in-termediate. Moreover, stable translocation intermediates canbe chemically cross-linked to the protein subunits that com-prise the translocation channels of both the outer and innermembranes. This finding demonstrates that imported pro-teins can simultaneously engage channels in both the outerand inner mitochondrial membrane, as depicted in Figure 16-26. Since roughly 1000 stuck translocation intermediatescan be observed in a typical yeast mitochondrion, it isthought that mitochondria have approximately 1000 generalimport pores for the uptake of mitochondrial proteins.

Three Energy Inputs Are Needed to ImportProteins into MitochondriaAs noted previously and indicated in Figure 16-26, ATP hy-drolysis by Hsc70 chaperone proteins in both the cytosol andthe mitochondrial matrix is required for import of mitochon-drial proteins. Cytosolic Hsc70 expends energy to maintainbound precursor proteins in an unfolded state that is compe-tent for translocation into the matrix. The importance of ATPto this function was demonstrated in studies in which a mito-chondrial precursor protein was purified and then denatured(unfolded) by urea. When tested in the cell-free mitochon-drial translocation system, the denatured protein was incor-porated into the matrix in the absence of ATP. In contrast,import of the native, undenatured precursor required ATP forthe normal unfolding function of cytosolic chaperones.

The sequential binding and ATP-driven release of multiplematrix Hsc70 molecules to a translocating protein may sim-ply trap the unfolded protein in the matrix. Alternatively, thematrix Hsc70, anchored to the membrane by the Tim44 pro-tein, may act as a molecular motor to pull the protein into thematrix (see Figure 16-26). In this case, the functions of matrixHsc70 and Tim44 would be analogous to the chaperone BiPand Sec63 complex, respectively, in post-translational translo-cation into the ER lumen (see Figure 16-9).

The third energy input required for mitochondrial pro-tein import is a H� electrochemical gradient, or proton-motive force, across the inner membrane (Chapter 8). In general, only mitochondria that are actively undergoing res-piration, and therefore have generated a proton-motive forceacross the inner membrane, are able to translocate precur-sor proteins from the cytosol into the mitochondrial matrix.Treatment of mitochondria with inhibitors or uncouplers ofoxidative phosphorylation, such as cyanide or dinitrophenol,dissipates this proton-motive force. Although precursor pro-teins still can bind tightly to receptors on such poisoned mi-tochondria, the proteins cannot be imported, either in intactcells or in cell-free systems, even in the presence of ATP andchaperone proteins. Scientists do not fully understand howthe proton-motive force is used to facilitate entry of a pre-cursor protein into the matrix. Once a protein is partially inserted into the inner membrane, it is subjected to a trans-membrane potential of 200 mV (matrix space negative),which is equivalent to an electric gradient of about 400,000

V/cm. One hypothesis is that the positive charges in the am-phipathic matrix-targeting sequence could simply be “elec-trophoresed,” or pulled, into the matrix space by theinside-negative membrane electric potential.

Multiple Signals and Pathways Target Proteins to Submitochondrial CompartmentsUnlike targeting to the matrix, targeting of proteins to the in-termembrane space, inner membrane, and outer membraneof mitochondria generally requires more than one targetingsequence and occurs via one of several pathways. Figure 16-28 summarizes the organization of targeting sequencesin proteins sorted to different mitochondrial locations.

Inner-Membrane Proteins Three separate pathways areknown to target proteins to the inner mitochondrial mem-brane. One pathway makes use of the same machinery that is


Mitochondrialmatrix

Outermembrane

Inner membrane

Intermembranespace

Inner mem

brane

Cleavedtargetingsequence

Cleavedtargetingsequence

Cytosol

Intermembranespace

Mitochondrialmatrix

COO−

COO−

FoldedDHFR

Translocationintermediate

NH3+

UnfoldedDHFR

FoldedDHFR

Outer membrane

Spacer sequence

Boundmethotrexateinhibitor

(a) (b)

NH3+

(c)

0.2 m�

▲ EXPERIMENTAL FIGURE 16-27 Experiments with

chimeric proteins show that a matrix-targeting sequence

alone directs proteins to the mitochondrial matrix and that

only unfolded proteins are translocated across both

membranes. The chimeric protein in these experimentscontained a matrix-targeting signal at its N-terminus (red),followed by a spacer sequence of no particular function (black),and then by dihydrofolate reductase (DHFR), an enzyme normallypresent only in the cytosol. (a) When the DHFR segment isunfolded, the chimeric protein moves across both membranes tothe matrix of energized mitochondria and the matrix-targetingsignal then is removed. (b) When the C-terminus of the chimericprotein is locked in the folded state by binding of methotrexate,translocation is blocked. If the spacer sequence is long enough

to extend across the transport channels, a stable translocationintermediate, with the targeting sequence cleaved off, isgenerated in the presence of methotrexate, as shown here. (c) The C-terminus of the translocation intermediate in (b) can bedetected by incubating the mitochondria with antibodies that bindto the DHFR segment, followed by gold particles coated withbacterial protein A, which binds nonspecifically to antibodymolecules (see Figure 5-51). An electron micrograph of asectioned sample reveals gold particles (red arrowhead) bound tothe translocation intermediate at a contact site between theinner and outer membranes. Other contact site (black arrows)also are evident. [Parts (a) and (b) adapted from J. Rassow et al., 1990,FEBS Letters 275:190. Part (c) from M. Schweiger et al., 1987, J. Cell Biol.105:235, courtesy of W. Neupert.]

used for targeting of matrix proteins (Figure 16-29, path A).A cytochrome oxidase subunit called CoxVa is a typical pro-tein transported by this pathway. The precursor form ofCoxVa, which contains an N-terminal matrix-targeting se-quence recognized by the Tom20/22 import receptor, istransferred through the general import pore of the outermembrane and the inner-membrane Tim23/17 translocationcomplex. In addition to the matrix-targeting sequence, whichis cleaved during import, CoxVa contains a hydrophobicstop-transfer sequence. As the protein passes through theTim23/17 channel, the stop-transfer sequence blocks translo-cation of the C-terminus across the inner membrane. The


membrane-anchored intermediate is then transferred later-ally into the bilayer of the inner membrane much as type I in-tegral membrane proteins are incorporated into the ERmembrane (see Figure 16-11).

A second pathway to the inner membrane is followed byproteins (e.g., ATP synthase subunit 9) whose precursorscontain both a matrix-targeting sequence and internal hydrophobic domains recognized by an inner-membraneprotein termed Oxa1. This pathway is thought to involvetranslocation of at least a portion of the precursor into thematrix via the Tom20/22 and Tim23/17 channels. Aftercleavage of the matrix-targeting sequence, the protein is in-

Mature protein

Cleavage bymatrix protease

Matrix-targeting sequence

Locations of targeting sequencesin preprotein

Importedprotein

Alcoholdehydro-genase III

Cytochrome oxidase subunit CoxVa

Porin(P70)

Cytochrome b2

Location of imported protein

Matrix

Innermembrane(path A)

Intermembranespace(path A)

Outermembrane

ATP synthasesubunit 9

Innermembrane(path B)

Cytochrome c heme lyase

Intermembranespace(path B)

ADP/ATPantiporter

Innermembrane(path C)

Targeting sequence forthe general import pore

Internal sequences recognized by Tom70 receptor and Tim22 complex

Stop-transfer and outer-membrane localization sequence

Hydrophobic stop-transfer sequence

First cleavage bymatrix protease

Second cleavage by proteasein intermembrane space

Intermembranespace–targeting sequence

Internal sequences recognized by Oxa1



FIGURE 16-28 Arrangement of

targeting sequences in imported

mitochondrial proteins. Most mitochondrialproteins have an N-terminal matrix-targetingsequence (pink) that is similar but notidentical in different proteins. Proteinsdestined for the inner membrane, theintermembrane space, or the outermembrane have one or more additionaltargeting sequences that function to directthe proteins to these locations by severaldifferent pathways. The designated pathwaysin parentheses correspond to thoseillustrated in Figures 16-29 and 16-30). [SeeW. Neupert, 1997, Ann. Rev. Biochem. 66:863.]

▲

serted into the inner membrane by a process that requiresinteraction with Oxa1 and perhaps other inner-membraneproteins (Figure 16-29, path B). Oxa1 is related to a bacter-ial protein involved in inserting some inner-membrane pro-teins in bacteria. This relatedness suggests that Oxa1 mayhave descended from the translocation machinery in theendosymbiotic bacterium that eventually became the mito-chondrion. However, the proteins forming the inner-

membrane channels in mitochondria are not related to theSecY protein in bacterial translocons. Oxa1 also participatesin the inner-membrane insertion of certain proteins (e.g.,subunit II of cytochrome oxidase) that are encoded by mito-chondrial DNA and synthesized in the matrix by mitochon-drial ribosomes.

The final pathway for insertion in the inner mitochon-drial membrane is followed by multipass proteins that con-


Tom40

Tim9/10

Tom20 Tom22 Tom40

Intermembranespace

Cytosol

Preprotein

Cleavedmatrix-targetingsequences

Tom70

COO−

NH3+

2

3

NH3+COO−

Tim44

Tim23/17

Mitochondrialmatrix

Innermembrane

Oxa1

Hsc70Hsc70

Assembled protein

Tim23/17Tim22 Tim54

NH3+

COO−

1 11

22

Tom40

NH3+

COO−

Outermembrane

Preprotein Protein

Stop-transfersequence

Matrix-targetingsequence

Matrix-targetingsequence

Oxa1-targetingsequence

Internal targetingsequences

Path A Path B Path C

▲ FIGURE 16-29 Three pathways for transporting proteins

from the cytosol to the inner mitochondrial membrane.

Proteins with different targeting sequences are directed to theinner membrane via different pathways. In all three pathways,proteins cross the outer membrane via the Tom40 general importpore. Proteins delivered by pathways A and B contain an N-terminal matrix-targeting sequence that is recognized by theTom20/22 import receptor in the outer membrane. Although boththese pathways use the Tim23/17 inner-membrane channel, theydiffer in that the entire precursor protein enters the matrix and

then is redirected to the inner membrane in pathway B. MatrixHsc70 plays a role similar its role in the import of soluble matrixproteins (see Figure 16-26). Proteins delivered by pathway Ccontain internal sequences that are recognized by the Tom70import receptor. A different inner-membrane translocation channel(Tim22/54) is used in this pathway. Two intermembrane proteins(Tim9 and Tim10) facilitate transfer between the outer and innerchannels. See the text for discussion. [See R. E. Dalbey and A. Kuhn,2000, Ann. Rev. Cell Devel. Biol. 16:51, and N. Pfanner and A. Geissler,2001, Nature Rev. Mol. Cell Biol. 2:339.]

tain six membrane-spanning domains, such as the ADP/ATPantiporter. These proteins, which lack the usual N-terminalmatrix-targeting sequence, contain multiple internal mito-chondrial targeting sequences. After the internal sequencesare recognized by Tom70, a second import receptor locatedin the outer membrane, the imported protein passes throughthe outer membrane through the general import pore (Figure16-29, path C). The protein then is transferred to a secondtranslocation complex in the inner membrane composed ofthe Tim22 and Tim54 proteins. Transfer to the Tim22/54complex depends on a multimeric complex of two small pro-teins, Tim9 and Tim10, that reside in the intermembranespace. These may act as chaperones to guide imported pro-teins from the general import pore to the Tim22/54 complexin the inner membrane. Ultimately the Tim22/54 complex isresponsible for incorporating the multiple hydrophobic seg-ments of the imported protein into the inner membrane.

Intermembrane-Space Proteins Two pathways deliver cy-tosolic proteins to the space between the inner and outer mi-tochondrial membranes. The major pathway is followed by


proteins, such as cytochrome b2, whose precursors carry twodifferent N-terminal targeting sequences, both of which ul-timately are cleaved. The most N-terminal of the two se-quences is a matrix-targeting sequence, which is removed bythe matrix protease. The second targeting sequence is a hy-drophobic segment that blocks complete translocation of theprotein across the inner membrane (Figure 16-30, path A).After the resulting membrane-embedded intermediate dif-fuses laterally away from the Tim23/17 translocation chan-nel, a protease in the membrane cleaves the protein near thehydrophobic transmembrane segment, releasing the matureprotein in a soluble form into the intermembrane space. Ex-cept for the second proteolytic cleavage, this pathway is sim-ilar to that of inner-membrane proteins such as CoxVa (seeFigure 16-29, path A).

Cytochrome c heme lyase, the enzyme responsible for thecovalent attachment of heme to cytochrome c, illustrates asecond pathway for targeting to the intermembrane space. Inthis pathway, the imported protein is delivered directly to theintermembrane space via the general import pore withoutinvolvement of any inner-membrane translocation factors

Protease

Path A Path B

Inner membrane

Cleavedmatrix-targetingsequence

Mitochondrialmatrix

2

Tim23/17

Tim44Heme

Tom20Tom22

Preprotein

Tom40

Intermembrane space

Outer membrane

Cytosol

3

Intermembrane space–targeting sequence

Matrix-targetingsequenceCOO−

NH3+

COO−NH3

+

1

Intermembrane space–targeting sequence

Tom40

Protein

▲ FIGURE 16-30 Two pathways for transporting proteins

from the cytosol to the mitochondrial intermembrane space.

Pathway A, the major one for delivery to the intermembranespace, is similar to pathway A for delivery to the inner membrane(see Figure 16-29). The major difference is that the internaltargeting sequence in proteins such as cytochrome b2 destinedfor the intermembrane space is recognized by an inner-membrane protease, which cleaves the protein on the

intermembrane-space side of the membrane. The releasedprotein then folds and binds to its heme cofactor within theintermembrane space. Pathway B involves direct delivery to theintermembrane space through the Tom40 general import pore inthe outer membrane. See the text for further discussion. [See R. E. Dalbey and A. Kuhn, 2000, Ann. Rev. Cell Devel. Biol. 16:51; N. Pfanner and A. Geissler, 2001, Nature Rev. Mol. Cell Biol. 2:339; and K. Diekert et al., 1999, Proc. Nat’l. Acad. Sci. USA 96:11752.]

(Figure 16-30, path B). Since translocation through theTom40 general import pore does not seem to be coupled toany energetically favorable process such as hydrolysis of ATPor GTP, the mechanism that drives unidirectional transloca-tion through the outer membrane is unclear. One possibilityis that cytochrome c heme lyase passively diffuses throughthe outer membrane and then is trapped within the inter-membrane space by binding to another protein that is deliv-ered to that location by one of the translocation mechanismsdiscussed previously.

Outer-Membrane Proteins Experiments with mitochon-drial porin (P70) provide clues about how proteins are tar-geted to the outer mitochondrial membrane. A short ma-trix-targeting sequence at the N-terminus of P70 isfollowed by a long stretch of hydrophobic amino acids (seeFigure 16-28). If the hydrophobic sequence is experimen-tally deleted from P70, the protein accumulates in the ma-trix space with its matrix-targeting sequence still attached.This finding suggests that the long hydrophobic sequencefunctions as a stop-transfer sequence that both preventstransfer of the protein into the matrix and anchors it as anintegral protein in the outer membrane. Normally, neitherthe matrix-targeting nor stop-transfer sequence is cleavedfrom the anchored protein. The source of energy to driveouter membrane proteins through the general import porehas not yet been identified.

Targeting of Chloroplast Stromal Proteins IsSimilar to Import of Mitochondrial Matrix ProteinsAmong the proteins found in the chloroplast stroma are theenzymes of the Calvin cycle, which functions in fixing carbondioxide into carbohydrates during photosynthesis (Chapter8). The large (L) subunit of ribulose 1,5-bisphosphate car-boxylase (rubisco) is encoded by chloroplast DNA and syn-thesized on chloroplast ribosomes in the stromal space. Thesmall (S) subunit of rubisco and all the other Calvin cycle en-zymes are encoded by nuclear genes and transported tochloroplasts after their synthesis in the cytosol. The precur-sor forms of these stromal proteins contain an N-terminalstromal-import sequence (see Table 16-1).

Experiments with isolated chloroplasts, similar to thosewith mitochondria illustrated in Figure 16-25, have shownthat they can import the S-subunit precursor after its syn-thesis. After the unfolded precursor enters the stromal space,it binds transiently to a stromal Hsc70 chaperone and the N-terminal sequence is cleaved. In reactions facilitated byHsc60 chaperonins that reside within the stromal space,eight S subunits combine with the eight L subunits to yieldthe active rubisco enzyme.

The general process of stromal import appears to be verysimilar to that for importing proteins into the mitochon-

drial matrix (see Figure 16-26). At least three chloroplastouter-membrane proteins, including a receptor that binds thestromal-import sequence and a translocation channel pro-tein, and five inner-membrane proteins are known to be es-sential for directing proteins to the stroma. Although theseproteins are functionally analogous to the receptor and chan-nel proteins in the mitochondrial membrane, they are notstructurally homologous. The lack of homology betweenthese chloroplast and mitochondrial proteins suggests thatthey may have arisen independently during evolution.

The available evidence suggests that chloroplast stromalproteins, like mitochondrial matrix proteins, are imported inthe unfolded state. Import into the stroma depends on ATPhydrolysis catalyzed by a stromal Hsc70 chaperone whosefunction is similar to Hsc70 in the mitochondrial matrix andBiP in the ER lumen. Unlike mitochondria, chloroplasts can-not generate an electrochemical gradient (proton-motiveforce) across their inner membrane. Thus protein import intothe chloroplast stroma appears to be powered solely by ATPhydrolysis.

Proteins Are Targeted to Thylakoids by Mechanisms Related to Translocation Across the Bacterial Inner MembraneIn addition to the double membrane that surrounds them,chloroplasts contain a series of internal interconnectedmembranous sacs, the thylakoids (see Figure 8-30). Pro-teins localized to the thylakoid membrane or lumen carryout photosynthesis. Many of these proteins are synthesizedin the cytosol as precursors containing multiple targetingsequences. For example, plastocyanin and other proteinsdestined for the thylakoid lumen require the successive ac-tion of two uptake-targeting sequences. The first is anN-terminal stromal-import sequence that directs the pro-tein to the stroma by the same pathway that imports the ru-bisco S subunit. The second sequence targets the proteinfrom the stroma to the thylakoid lumen. The role of thesetargeting sequences has been shown in cell-free experimentsmeasuring the uptake into chloroplasts of mutant proteinsgenerated by recombinant DNA techniques. For instance,mutant plastocyanin that lacks the thylakoid-targeting se-quence but contains an intact stromal-import sequence ac-cumulates in the stroma and is not transported into thethylakoid lumen.

Four separate pathways for transporting proteins fromthe stroma into the thylakoid have been identified. All fourpathways have been found to be closely related to analogoustransport mechanisms in bacteria, illustrating the close evo-lutionary relationship between the stromal membrane andthe bacterial inner membrane. Transport of plastocyanin and related proteins into the thylakoid lumen occurs by anSRP-dependent pathway (Figure 16-31, path A). A secondpathway for transporting proteins into the thylakoid lumen


utilizes a protein related to bacterial SecA and is thought toutilize a mechanism similar to that depicted in Figure 16-23.A third pathway, which targets proteins to the thylakoidmembrane, depends on a protein related to the mitochon-drial Oxa1 protein and the homologous bacterial protein(see Figure 16-29, path B). Some proteins encoded by chloro-plast DNA and synthesized in the stroma or transported intothe stroma from the cytosol are inserted into the thylakoidmembrane via this pathway.

Finally, thylakoid proteins that bind metal-containing co-factors follow another pathway into the thylakoid lumen (Fig-ure 16-31, pH pathway). The unfolded precursors of theseproteins are first targeted to the stroma, where the N-


terminal stromal-import sequence is cleaved off and the proteinthen folds and binds its cofactor. A set of thylakoid-membraneproteins assists in translocating the folded protein and boundcofactor into the thylakoid lumen, a process powered by thepH gradient normally maintained across the thylakoid mem-brane. The thylakoid-targeting sequence that directs a proteinto this pH-dependent pathway includes two closely spacedarginine residues that are crucial for recognition. Bacterial cellsalso have a mechanism for translocating folded proteins witha similar arginine-containing sequence across the inner mem-brane. The molecular mechanism whereby these large foldedglobular proteins are transported across the thylakoid mem-brane is currently under intense study.

Cleaved importsequence

4

RR

3

2

COO−

COO−

NH3+

NH3+

RR

3

2ChloroplastSRP

Matureplastocyanin

ChloroplastSRP receptor

Maturemetal-bindingprotein

1

Cleaved importsequence

Plastocyaninprecursor

Metal-bindingprecursor

SRP-dependentpathway

∆pH pathway

Bound metal ions

Ticcomplex

Toccomplex

RR

Stromal-importsequence

Thylakoid-targetingsequence

1

Ticcomplex

Toccomplex

Cytosol

Intermembranespace

Stroma

Outer membrane

Inner membrane

Thylakoidlumen

Thylakoid membrane

Metal-bindingproteinPlastocyanin

FIGURE 16-31 Two of the four

pathways for transporting proteins

from the cytosol to the thylakoid

lumen. In these pathways, unfoldedprecursors are delivered to the stroma viathe same outer-membrane proteins thatimport stromal-localized proteins. Cleavageof the N-terminal stromal-import sequenceby a stromal protease then reveals thethylakoid-targeting sequence. At this pointthe two pathways diverge. In the SRP-dependent pathway (left), plastocyanin andsimilar proteins are kept unfolded in thestromal space by a set of chaperones (notshown) and, directed by the thylakoid-targeting sequence, bind to proteins thatare closely related to the bacterial SRP,SRP receptor, and SecY translocon, whichmediate movement into the lumen. Afterthe thylakoid-targeting sequence isremoved in the thylakoid lumen by aseparate endoprotease, the protein foldsinto its mature conformation. In the pH-dependent pathway (right), metal-bindingproteins fold in the stroma, and complexredox cofactors are added. Two arginineresidues (RR) at the N-terminus of thethylakoid-targeting sequence and a pHgradient across the inner membrane arerequired for transport of the folded proteininto the thylakoid lumen. The translocon inthe thylakoid membrane is composed ofat least four proteins related to proteins inthe bacterial inner membrane. [See R. Dalbey and C. Robinson, 1999, TrendsBiochem. Sci. 24:17; R. E. Dalbey and A. Kuhn,2000, Ann. Rev. Cell Devel. Biol. 16:51; and C. Robinson and A. Bolhuis, 2001, Nature Rev.Mol. Cell Biol. 2:350.]

▲

16.6 • Sorting of Peroxisomal Proteins 693


Sorting of Proteins to Mitochondria and Chloroplasts

■ Most mitochondrial and chloroplast proteins are encodedby nuclear genes, synthesized on cytosolic ribosomes, andimported post-translationally into the organelles.

■ All the information required to target a precursor pro-tein from the cytosol to the mitochondrial matrix orchloroplast stroma is contained within its N-terminal uptake-targeting sequence. After protein import, the uptake-targeting sequence is removed by proteases withinthe matrix or stroma.

■ Cytosolic chaperones maintain the precursors of mito-chondrial and chloroplast proteins in an unfolded state.Only unfolded proteins can be imported into the or-ganelles. Translocation occurs at sites where the outer andinner membranes of the organelles are close together.

■ Proteins destined to the mitochondrial matrix bind to re-ceptors on the outer mitochondrial membrane, and then aretransferred to the general import pore (Tom40) in the outermembrane. Translocation occurs concurrently through theouter and inner membranes, driven by the proton-motiveforce across the inner membrane and ATP hydrolysis by theHsc70 ATPase in the matrix (see Figure 16-26).

■ Proteins sorted to mitochondrial destinations other thanthe matrix usually contain two or more targeting se-quences, one of which may be an N-terminal matrix-targeting sequence (see Figure 16-28).

■ Some mitochondrial proteins destined for the inter-membrane space or inner membrane are first imported intothe matrix and then redirected; others never enter the ma-trix but go directly to their final location.

■ Protein import into the chloroplast stroma occursthrough inner-membrane and outer-membrane transloca-tion channels that are analogous in function to mitochon-drial channels but composed of proteins unrelated insequence to the corresponding mitochondrial proteins.

■ Proteins destined for the thylakoid have secondary tar-geting sequences. After entry of these proteins into thestroma, cleavage of the stromal-targeting sequences revealsthe thylakoid-targeting sequences.

■ The three known pathways for moving proteins from thechloroplast stroma to the thylakoid closely resemble translo-cation across the bacterial inner membrane (see Figure 16-31). One of these systems can translocate folded proteins.

Sorting of Peroxisomal ProteinsPeroxisomes are small organelles bounded by a single mem-brane. Unlike mitochondria and chloroplasts, peroxisomeslack DNA and ribosomes. Thus all peroxisomal proteins are

16.6

encoded by nuclear genes, synthesized on ribosomes free inthe cytosol, and then incorporated into preexisting or newlygenerated peroxisomes. As peroxisomes are enlarged by ad-dition of protein (and lipid), they eventually divide, formingnew ones, as is the case with mitochondria and chloroplasts.

The size and enzyme composition of peroxisomes varyconsiderably in different kinds of cells. However, all peroxi-somes contain enzymes that use molecular oxygen to oxi-dize various substrates, forming hydrogen peroxide (H2O2).Catalase, a peroxisome-localized enzyme, efficiently decom-poses H2O2 into H2O. Peroxisomes are most abundant inliver cells, where they constitute about 1 to 2 percent of thecell volume.

Cytosolic Receptor Targets Proteins with an SKL Sequence at the C-Terminus into the Peroxisomal MatrixThe import of catalase and other proteins into rat liver per-oxisomes can be assayed in a cell-free system similar to thatused for monitoring mitochondrial protein import (see Figure 16-25). By testing various mutant catalase proteins inthis system, researchers discovered that the sequence Ser-Lys-Leu (SKL in one-letter code) or a related sequence at theC-terminus was necessary for peroxisomal targeting. Fur-ther, addition of the SKL sequence to the C-terminus of anormally cytosolic protein leads to uptake of the alteredprotein by peroxisomes in cultured cells. All but a few of themany different peroxisomal matrix proteins bear a sequenceof this type, known as peroxisomal-targeting sequence 1,or simply PTS1.

The pathway for import of catalase and other PTS1-bearing proteins into the peroxisomal matrix is depicted inFigure 16-32. The PTS1 binds to a soluble receptor proteinin the cytosol (Pex5), which in turn binds to a receptor in the peroxisome membrane (Pex14). The soluble and membrane-associated peroxisomal import receptors appearto have a function analogous to that of the SRP and SRP re-ceptor in targeting proteins to the ER lumen. Still bound toPex5, the imported protein then moves through a multimerictranslocation channel, a feature that differs from protein im-port into the ER lumen. At some stage either during or afterentry into the matrix, Pex5 dissociates from the peroxiso-mal matrix protein and is recycled back to the cytoplasm. Incontrast to the N-terminal uptake-targeting sequences onproteins destined for the ER lumen, mitochondrial matrix,and chloroplast stroma, the PTS1 sequence is not cleavedfrom proteins after their entry into a peroxisome. Proteinimport into peroxisomes requires ATP hydrolysis, but it isnot known how the energy released from ATP is used topower unidirectional translocation across the peroxisomalmembrane.

The peroxisome import machinery, unlike most systemsthat mediate protein import into the ER, mitochondria, andchloroplast, can translocate folded proteins across the

membrane. For example, catalase assumes a folded confor-mation and binds to heme in the cytoplasm before travers-ing the peroxisomal membrane. Cell-free studies have shownthat the peroxisome import machinery can transport a widevariety of molecules, including very large ones. To explainthis unusual ability, scientists have speculated that a translo-cation channel of variable size may assemble by an unknownmechanism to fit exactly the diameter of the PTS1-bearingsubstrate molecule and then disassemble once translocationhas been completed.

A few peroxisomal matrix proteins such as thiolase are synthesized as precursors with an N-terminal uptake-targeting sequence known as PTS2. These proteins bind to a different cytosolic receptor protein, but otherwise import is thought to occur by the same mechanism as forPTS1-containing proteins.

Peroxisomal Membrane and Matrix Proteins AreIncorporated by Different Pathways

Autosomal recessive mutations that cause defectiveperoxisome assembly occur naturally in the humanpopulation. Such defects can lead to severe impair-

ment of many organs and to death. In Zellweger syndromeand related disorders, for example, the transport of many orall proteins into the peroxisomal matrix is impaired; newlysynthesized peroxisomal enzymes remain in the cytosol and


are eventually degraded. Genetic analyses of cultured cellsfrom different Zellweger patients and of yeast cells carryingsimilar mutations have identified more than 20 genes that arerequired for peroxisome biogenesis. ❚

Studies with peroxisome-assembly mutants have shownthat different pathways are used for importing peroxisomalmatrix proteins versus inserting proteins into the peroxiso-mal membrane. For example, analysis of cells from someZellweger patients led to identification of genes encoding theputative translocation channel proteins Pex10, Pex12, andPex2. Mutant cells defective in any one of these proteins can-not incorporate matrix proteins such as catalase into perox-isomes; nonetheless, the cells contain empty peroxisomesthat have a normal complement of peroxisomal membraneproteins (Figure 16-33b). Mutations in any one of threeother genes were found to block insertion of peroxisomalmembrane proteins as well as import of matrix proteins (Figure 16-33c). These findings demonstrate that one set ofproteins translocates soluble proteins into the peroxisomalmatrix but a different set is required for insertion of proteinsinto the peroxisomal membrane. This situation differsmarkedly from that of the ER, mitochondrion, and chloro-plast, for which, as we have seen, membrane proteins andsoluble proteins share many of the same components fortheir insertion into these organelles.

Although most peroxisomes are generated by division ofpreexisting organelles, these organelles also can arise de novoby the two-stage process depicted in Figure 16-34. In thiscase, peroxisomal membrane proteins first are targeted toprecursor membranes by sequences that differ from bothPTS1 and PTS2. Analysis of mutant cells revealed thatPex19 is the receptor protein responsible for targeting ofperoxisomal membrane proteins, while Pex3 and Pex16 arenecessary for their proper insertion into the membrane. Theinsertion of peroxisomal membrane proteins generatesmembranes that have all the components necessary for im-port of matrix proteins, leading to the formation of mature,

COO−

NH3+

PTS1peroxisomal-targeting sequence

Pex5 receptor

Peroxisomalmatrix protein

Pex12

Pex10

1

2

3

4Pex14

Peroxisomalmatrix

Cytosol

Pex2

Peroxisomal membrane

FIGURE 16-32 Import of peroxisomal matrix proteins

directed by PTS1 targeting sequence. Step : Catalase andmost other peroxisomal matrix proteins contain a C-terminalPTS1 uptake-targeting sequence (red) that binds to the cytosolic receptor Pex5. Step : Pex5 with the bound matrix proteininteracts with the Pex14 receptor located on the peroxisomemembrane. Step : The matrix protein–Pex5 complex is thentransferred to a set of membrane proteins (Pex10, Pex12, andPex2) that are necessary for translocation into the peroxisomalmatrix by an unknown mechanism. Step : At some point, eitherduring translocation or in the lumen, Pex5 dissociates from thematrix protein and returns to the cytosol, a process that involvesthe Pex2/10/12 complex and additional membrane and cytosolicproteins not shown. Note that folded proteins can be importedinto peroxisomes and that the targeting sequence is not removedin the matrix. [See P. E. Purdue and P. B. Lazarow, 2001, Ann. Rev. CellDevel. Biol. 17:701; S. Subramani et al., 2000, Ann. Rev. Biochem.69:399; and V. Dammai and S. Subramani, 2001, Cell 105:187.]

1

2

3

4

▲

functional peroxisomes. Division of mature peroxisomes,which largely determines the number of peroxisomeswithin a cell, depends on still another protein, Pex11. Over-expression of the Pex11 protein causes a large increase inthe number of peroxisomes, suggesting that this protein

controls the extent of peroxisome division. The small per-oxisomes generated by division can be enlarged by incor-poration of additional matrix and membrane proteins viathe same pathways described previously.


Sorting of Peroxisomal Proteins

■ All peroxisomal proteins are synthesized on cytosolic ribosomes and incorporated into the organelle post-translationally.

■ Most peroxisomal matrix proteins contain a C-terminalPTS1 targeting sequence; a few have an N-terminal PTS2targeting sequence. Neither targeting sequence is cleavedafter import.

■ All proteins destined for the peroxisomal matrix bind toa cytosolic receptor, which differs for PTS1- and PTS2-

16.6 • Sorting of Peroxisomal Proteins 695

Stained forPMP70

Stained forcatalase(a) Wild-type cells

(b) Pex1 mutants (deficientin matrix-protein import)

(c) Pex3 mutants (deficientin membrane-protein insertion)

CatalasePMP70

Peroxisome

EXPERIMENTAL FIGURE 16-33 Fluorescent-antibody

staining of peroxisomal biogenesis mutants reveals different

pathways for incorporation of membrane and matrix

proteins. Cells were stained with antibodies to PMP70, aperoxisomal membrane protein, or with antibodies to catalase, aperoxisomal matrix protein, then viewed in a fluorescentmicroscope. (a) In wild-type cells, both peroxisomal membraneand matrix proteins are visible as bright foci in numerousperoxisomal bodies. (b) In cells from a Pex12-deficient patient,catalase is distributed uniformly throughout the cytosol, whereasPMP70 is localized normally to peroxisomal bodies. (c) In cellsfrom a Pex3-deficient patient, peroxisomal membranes cannotassemble, and as a consequence peroxisomal bodies do notform. Thus both catalase and PMP70 are mis-localized to thecytosol. [Courtesy of Stephen Gould, Johns Hopkins University.]

Pex3

Pex16

Pex19 Pex5

Pex14

Pex10

Pex12

Pex2

Pex7

Pex11

Peroxisomalmembraneproteins

Peroxisomal

ghost PTS1-bearingmatrix protein

PTS2-bearingmatrix protein

Mature peroxisome

Precursormembrane

PMP70 Catalase

▲ FIGURE 16-34 Model of peroxisomal biogenesis and

division. The first stage in the de novo formation of peroxisomesis the incorporation of peroxisomal membrane proteins intoprecursor membranes. Pex19 acts as the receptor for membrane-targeting sequences. Pex3 and Pex16 are required for the properinsertion of proteins into the forming peroxisomal membrane.Insertion of all peroxisomal membrane proteins produces aperoxisomal ghost, which is capable of importing proteins

targeted to the matrix. The pathways for importing PTS1- andPTS2-bearing matrix proteins differ only in the identity of thecytosolic receptor (Pex5 and Pex7, respectively) that binds thetargeting sequence (see Figure 16-32). Complete incorporation ofmatrix proteins yields a mature peroxisome. The proliferation ofperoxisomes requires division of mature peroxisomes, a processthat depends on the Pex11 protein.

▲

KEY TERMS

chaperones 665cotranslational

translocation 661general import pore 685hydropathy profile 671N- and O-linked

oligosaccharides 673post-translational

translocation 665signal-anchor

sequence 668

bearing proteins, and then are directed to common importreceptor and translocation machinery on the peroxisomalmembrane (see Figure 16-32).

■ Translocation of matrix proteins across the peroxisomalmembrane depends on ATP hydrolysis. Many peroxisomalmatrix proteins fold in the cytosol and traverse the mem-brane in a folded conformation.

■ Proteins destined for the peroxisomal membrane containdifferent targeting sequences than peroxisomal matrix pro-teins and are imported by a different pathway.

■ Unlike mitochondria and chloroplasts, peroxisomes canarise de novo from precursor membranes, as well as by di-vision of preexisting organelles (see Figure 16-34).

P E R S P E C T I V E S F O R T H E F U T U R E

As we have seen in this chapter, we now understand many as-pects of the basic processes responsible for selectively trans-porting proteins into the endoplasmic reticulum (ER),mitochondrion, chloroplast, and peroxisome. Biochemicaland genetic studies, for instance, have identified cis-actingsignal sequences responsible for targeting proteins to the cor-rect organelle membrane and the membrane receptors thatrecognize these signal sequences. We also have learned muchabout the underlying mechanisms that translocate proteinsacross organelle membranes, and have determined whetherenergy is used to push or pull proteins across the membranein one direction, the type of channel through which proteinspass, and whether proteins are translocated in a folded oran unfolded state. Nonetheless, many fundamental questionsremain unanswered, including how fully folded proteinsmove across a membrane and how the topology of multipassmembrane proteins is determined.

The peroxisomal import machinery provides one exam-ple of the translocation of folded proteins. It not only is ca-pable of translocating fully folded proteins with boundcofactors into the peroxisomal matrix but can even direct theimport of a large gold particle decorated with a (PTS1) per-oxisomal targeting peptide. Some researchers have specu-lated that the mechanism of peroxisomal import may berelated to that of nuclear import, the best-understood exam-ple of post-translational translocation of folded proteins(Chapter 12). Both the peroxisomal and nuclear import ma-chinery can transport folded molecules of very divergentsizes, and both appear to involve a component that cycles be-tween the cytosol and the organelle interior—the Pex5 PTS1receptor in the case of peroxisomal import and the Ran-importin complex in the case of nuclear import. However,there also appear to be crucial differences between the twotranslocation processes. For example, nuclear pores repre-sent large stable macromolecular assemblies readily observedby electron microscopy, whereas analogous porelike struc-tures have not been observed in the peroxisomal membrane.


Moreover, small molecules can readily pass through nuclearpores, whereas peroxisomal membranes maintain a perma-nent barrier to the diffusion of small hydrophilic molecules.Taken together, these observations suggest that peroxisomalimport may require an entirely new type of translocationmechanism. The evolutionarily conserved mechanisms fortranslocating folded proteins across the cytoplasmic mem-brane of bacterial cells and across the thylakoid membraneof chloroplasts also are poorly understood. A better under-standing of all of these processes for translocating foldedproteins across a membrane will likely hinge on future de-velopment of in vitro translocations systems that allow in-vestigators to define the biochemical mechanisms drivingtranslocation and to identify the structures of trappedtranslocation intermediates.

Compared with our understanding of how soluble pro-teins are translocated into the ER lumen and mitochondrialmatrix, our understanding of how cis-acting sequences spec-ify the topology of multipass membrane proteins is quite el-ementary. For instance, we do not know how the transloconchannel accommodates polypeptides that are oriented differ-ently with respect to the membrane, nor do we understandhow local polypeptide sequences interact with the transloconchannel both to set the orientation of transmembrane spansand to signal for lateral passage into the membrane bilayer.A better understanding of how the amino acid sequences ofmembrane proteins can specify membrane topology will becrucial for decoding the vast amount of structural informa-tion for membrane proteins contained within databases ofgenomic sequences.

A more detailed understanding of all translocationprocesses should continue to emerge from genetic and bio-chemical studies, both in yeasts and in mammals. These stud-ies will undoubtedly reveal additional key proteins involved inthe recognition of targeting sequences and in the translocationof proteins across lipid bilayers. Finally, the now mostly rudi-mentary structural studies of translocon channels will likely beextended in the future to reveal the structures and conforma-tional states for the channels at resolutions on the atomic scale.

signal-recognition particle (SRP) 661

signal (uptake-targeting) sequences 659

stop-transfer anchor sequence 667

topogenic sequences 666topology of membrane

proteins 666translocon 662unfolded-protein response 679

P E R S P E C T I V E S F O R T H E F U T U R E

8. Describe how you might use recombinant DNA to en-gineer a strain of Yersinia such that it would be capable of in-serting a protein of interest into the cytosol of mammalianmacrophage cells.

9. Describe what would happen to the precursor of a mi-tochondrial matrix protein in the following types of mito-chondrial mutants: (a) a mutation in the Tom22 signalreceptor; (b) a mutation in the Tom70 signal receptor; (c) amutation in the matrix Hsc70; and (d) a mutation in the ma-trix signal peptidase.

10. Describe the similarities and differences between themechanism of import into the mitochondrial matrix and thechloroplast stroma.

11. Design a set of experiments using chimeric proteins,composed of a mitochondrial precursor protein fused to di-hydrofolate reductase (DHFR), that could be used to deter-mine how much of the precursor protein must protrude intothe mitochondrial matrix in order for the matrix-targetingsequence to be cleaved by the matrix-processing protease (seeFigure 16-27).

12. Protein targeting to both mitochondria and chloroplastsinvolves the sorting of proteins to multiple sites within the respective organelle. Briefly list these sites. Taking the mito-chondrion as an example and the proteins ADP/ATP anti-porter and cytochrome b2 as the specific cases, compare andcontrast the extent to which a common mechanism is usedfor the site-specific targeting of these two proteins.

13. Suppose that you have identified a new mutant cell linethat lacks functional peroxisomes. Describe how you coulddetermine experimentally whether the mutant is primarilydefective for insertion/assembly of peroxisomal membraneproteins or matrix proteins.

ANALYZE THE DATA

Imagine that you are evaluating the early steps in transloca-tion and processing of the secretory protein prolactin. Byusing an experimental approach similar to that shown in Figure 16-7, you can use truncated prolactin mRNAs to control the length of nascent prolactin polypeptides that aresynthesized. When prolactin mRNA that lacks a chain-termination (stop) codon is translated in vitro, the newly syn-thesized polypeptide ending with the last codon included onthe mRNA will remain attached to the ribosome, thus al-lowing a polypeptide of defined length to extend from theribosome. You have generated a set of mRNAs that encodesegments of the N-terminus of prolactin of increasing length,and each mRNA can be translated in vitro by a cytosolictranslation extract containing ribosomes, tRNAs, aminoacyl-tRNA synthetases, GTP, and translation initiation and elon-gation factors. When radio-labeled amino acids are included

Analyze the Data 697

REVIEW THE CONCEPTS

1. Describe the source or sources of energy needed for uni-directional translocation across the membrane in (a) co-translational translocation into the endoplasmic reticulum(ER); (b) post-translational translocation into the ER; (c) translocation across the bacterial cytoplasmic membrane;and (d) translocation into the mitochondrial matrix.

2. Translocation into most organelles usually requires theactivity of one or more cytosolic proteins. Describe the basicfunction of three different cytosolic factors required fortranslocation into the ER, mitochondria, and peroxisomes,respectively.

3. Describe the typical principles used to identify topogenicsequences within proteins and how these can be used to de-velop computer algorithms. How does the identification oftopogenic sequences lead to prediction of the membranearrangement of a multipass protein? What is the importanceof the arrangement of positive charges relative to the mem-brane orientation of a signal anchor sequence?

4. The endoplasmic reticulum (ER) is an important site of“quality control” for newly synthesized proteins. What ismeant by quality control in this context? What accessory pro-teins are typically involved in the processing of newly synthe-sized proteins within the ER? Cells generally degradeER-exit-incompetent proteins. Where within the cell does suchdegradation occur and what is the relationship of Sec61p pro-tein translocon to the degradation process?

5. Temperature-sensitive yeast mutants have been isolatedthat block each of the enzymatic steps in the synthesis of thedolichol-oligosaccharide precursor for N-linked glycosyla-tion (see Figure 16-17). Propose an explanation for why mu-tations that block synthesis of the intermediate with thestructure dolichol-PP-(GlcNAc)2Man5 completely preventaddition of N-linked oligosaccharide chains to secretoryproteins, whereas mutations that block conversion of thisintermediate into the completed precursor—dolichol-PP-(GlcNAc)2Man9Glc3—allow the addition of N-linkedoligosaccharide chains to secretory glycoproteins.

6. Name four different proteins that facilitate the modifi-cation and/or folding of secretory proteins within the lumenof the ER. Indicate which of these proteins covalently modi-fies substrate proteins and which brings about only confor-mational changes in substrate proteins.

7. Because you are interested in studying how a particularsecretory protein folds within the ER, you wish to determinewhether BiP binds to the newly synthesized protein in ER ex-tracts. You find that you can isolate some of the newly syn-thesized secretory protein bound to BiP when ADP is addedto the cell extract but not when ATP is added to the extract.Explain this result based on the mechanism for BiP bindingto substrate proteins.

in the translation mixture, only the polypeptide encoded bythe added mRNA will be labeled. After completion of trans-lation, each reaction mixture was resolved by SDS poly-acrylamide gel electrophoresis, and the labeled polypeptideswere identified by autoradiography.

a. The autoradiogram depicted below shows the results of anexperiment in which each translation reaction was carried outeither in the presence (�) or the absence (�) of microsomalmembranes. Based on the gel mobility of peptides synthesizedin the presence or absence of microsomes, deduce how longthe prolactin nascent chain must be in order for the prolactinsignal peptide to enter the ER lumen and to be cleaved by sig-nal peptidase. (Note that microsomes carry significant quan-tities of SRP weakly bound to the membranes.)


translation to engage the SRP and thereby become boundto microsomal membranes.

− + − + − + − + − + − +

Siz

e o

f la

bel

edp

oly

pep

tid

e

Size of mRNA (in codons)50 70 110 15090 130

Size of mRNA (in codons)50

T M

70

T M

110

T M

150

T M

90

T M

130

T M

Siz

e o

f la

bel

edp

oly

pep

tid

e

b. Given this length, what can you conclude about the con-formational state of the nascent prolactin polypeptide whenit is cleaved by signal peptidase? The following lengths willbe useful for your calculation: the prolactin signal sequenceis cleaved after amino acid 31; the channel within the ribo-some occupied by a nascent polypeptide is about 150 Å long;a membrane bilayer is about 50 Å thick; in polypeptides withan �-helical conformation, one residue extends 1.5 Å,whereas in fully extended polypeptides, one residue extendsabout 3.5 Å.

c. The experiment described in part (a) is carried out in anidentical manner except that microsomal membranes are notpresent during translation but are added after translation iscomplete. In this case none of the samples shows a differencein mobility in the presence or absence of microsomes. Whatcan you conclude about whether prolactin can be translo-cated into isolated microsomes post-translationally?

d. In another experiment, each translation reaction wascarried out in the presence of microsomes, and then themicrosomal membranes and bound ribosomes were sepa-rated from free ribosomes and soluble proteins by cen-trifugation. For each translation reaction, both the totalreaction (T) and the membrane fraction (M) were resolvedin neighboring gel lanes. Based on the amounts of labeledpolypeptide in the membrane fractions in the autoradi-ogram depicted below, deduce how long the prolactin nas-cent chain must be in order for ribosomes engaged in

REFERENCES

Translocation of Proteins Across the ER MembraneDalbey, R. E., and G. Von Heijne. 1992. Signal peptidases in

prokaryotes and eukaryotes: a new protease family. Trends Biochem.Sci. 17:474–478.

Johnson, A. E., and M. A. van Waes. 1999. The translocon: adynamic gateway at the ER membrane. Ann. Rev. Cell Devel. Biol.15:799–842.

Keenan, R. J., D. M. Freymann, R. M. Stroud, and P. Walter.2001. The signal recognition particle. Ann. Rev. Biochem.70:755–775.

Menetret, J. F., A. Neuhof, D. G. Morgan, K. Plath, M. Rader-macher, T. A. Rapoport, and C. W. Akey. 2000. The structure of ribosome-channel complexes engaged in protein translocation. Mol.Cell 6:1219–1232.

Rapoport, T. A., K. E. Matlack, K. Plath, B. Misselwitz, and O.Staeck. 1999. Post-translational protein translocation across themembrane of the endoplasmic reticulum. Biol. Chem. 380:1143–1150.

Insertion of Proteins into the ER MembraneEnglund, P. T. 1993. The structure and biosynthesis of glyco-

sylphosphatidylinositol protein anchors. Ann. Rev. Biochem.62:121–138.

Goder, V., and M. Spiess. 2001. Topogenesis of membrane pro-teins: determinants and dynamics. FEBS Lett. 504:87–93.

Mothes, W., et al. 1997. Molecular mechanism of membraneprotein integration into the endoplasmic reticulum. Cell 89:523–533.

von Heijne, G. 1999. Recent advances in the understanding ofmembrane protein assembly and structure. Q. Rev. Biophys.32:285–307.

Protein Modifications, Folding, and Quality Control in the ER

Bertolotti, A., Y. Zhang, L. M. Hendershot, H. P. Harding, andD. Ron. 2000. Dynamic interaction of BiP and ER stress transduc-ers in the unfolded-protein response. Nature Cell Biol. 2:326–332.

Ellgaard, L., M. Molinari, and A. Helenius. 1999. Setting thestandards: quality control in the secretory pathway. Science286:1882–1888.

Helenius, A., and M. Aebi. 2001. Intracellular functions of N-linked glycans. Science 291:2364–2369.

Kornfeld, R., and S. Kornfeld. 1985. Assembly of asparagine-linked oligosaccharides. Ann. Rev. Biochem. 45:631–664.

Parodi, A. J. 2000. Protein glucosylation and its role in proteinfolding. Ann. Rev. Biochem. 69:69–93.

Patil, C., and P. Walter. 2001. Intracellular signaling from theendoplasmic reticulum to the nucleus: the unfolded protein responsein yeast and mammals. Curr. Opin. Cell Biol. 13:349–355.

Plemper, R. K., and D. H. Wolf. 1999. Retrograde proteintranslocation: eradication of secretory proteins in health and disease.Trends Biochem. Sci. 24:266–270.

Sevier, C. S., and C. A. Kaiser. 2002. Formation and transfer ofdisulphide bonds in living cells. Nature Rev. Mol. Cell Biol.3:836–847.

Silberstein, S., and R. Gilmore. 1996. Biochemistry, molecularbiology, and genetics of the oligosaccharyltransferase. FASEB J.10:849–858.

Tsai, B., Y. Ye, and T. A. Rapoport. 2002. Retro-translocationof proteins from the endoplasmic reticulum into the cytosol. NatureRev. Mol. Cell Biol. 3:246–255.

Export of Bacterial ProteinsDanese, P. N., and T. J. Silhavy. 1998. Targeting and assembly

of periplasmic and outer-membrane proteins in Escherichia coli. Ann.Rev. Genet. 32:59–94.

Galan, J. E. 2001. Salmonella interactions with host cells: typeIII secretion at work. Ann. Rev. Cell Devel. Biol. 17:53–86.

Thanassi, D. G., and S. J. Hultgren. 2000. Multiple pathwaysallow protein secretion across the bacterial outer membrane. Curr.Opin. Cell Biol. 12:420–430.

Wickner, W., and M. R. Leonard. 1996. Escherichia coli pre-protein translocase. J. Biol. Chem. 271:29514–29516.

Sorting of Proteins to Mitochondria and ChloroplastsChen, K., X. Chen, and D. J. Schnell. 2000. Mechanism of pro-

tein import across the chloroplast envelope. Biochem. Soc. Trans.28:485–491.

Dalbey, R. E., and A. Kuhn. 2000. Evolutionarily related inser-tion pathways of bacterial, mitochondrial, and thylakoid membraneproteins. Ann. Rev. Cell Devel. Biol. 16:51–87.

Matouschek, A., N. Pfanner, and W. Voos. 2000. Protein un-folding by mitochondria: the Hsp70 import motor. EMBO Rept.1:404–410.

Neupert, W., and M. Brunner. 2002. The protein import motorof mitochondria. Nature Rev. Mol. Cell Biol. 3:555–565.

Pfanner, N., and A. Geissler. 2001. Versatility of the mitochon-drial protein import machinery. Nature Rev. Mol. Cell Biol.2:339–349.

Robinson, C., and A. Bolhuis. 2001. Protein targeting by thetwin-arginine translocation pathway. Nature Rev. Mol. Cell Biol.2:350–356.

Sorting of Peroxisomal ProteinsDammai, V., and S. Subramani. 2001. The human peroxisomal

targeting signal receptor, Pex5p, is translocated into the peroxisomalmatrix and recycled to the cytosol. Cell 105:187–196.

Gould, S. J., and C. S. Collins. 2002. Opinion: peroxisomal-protein import: is it really that complex? Nature Rev. Mol. Cell Biol.3:382–389.

Gould, S. J., and D. Valle. 2000. Peroxisome biogenesis disor-ders: genetics and cell biology. Trends Genet. 16: 340–345.

Purdue, P. E., and P. B. Lazarow. 2001. Peroxisome biogenesis.Ann. Rev. Cell Devel. Biol. 17:701–752.

Subramani, S., A. Koller, and W. B. Snyder. 2000. Import of per-oxisomal matrix and membrane proteins. Ann. Rev. Biochem.69:399–418.

References 699

Date post:	11-May-2018
Category:	Documents
Upload:	ngodang
View:	214 times
Download:	1 times

MOVING PROTEINS INTO MEMBRANES AND …kbp-srmc.yolasite.com/resources/Chapter 16.pdf · targeting...

Documents