disulfide-linked protein folding pathwayssites.lsa.umich.edu/webbkeane/wp-content/uploads/... ·...

ANRV356-CB24-09 ARI 3 September 2008 15:27

Disulfide-Linked ProteinFolding PathwaysBharath S. Mamathambika1,3

and James C. Bardwell2,3,∗

1Biophysics Graduate Program, 2Department of Molecular, Cellular, and DevelopmentalBiology, 3Howard Hughes Medical Institute, University of Michigan, Ann Arbor,Michigan 48109; email: [email protected], [email protected]

Annu. Rev. Cell Dev. Biol. 2008. 24:211–35

First published online as a Review in Advance onJune 26, 2008

The Annual Review of Cell and DevelopmentalBiology is online at cellbio.annualreviews.org

This article’s doi:10.1146/annurev.cellbio.24.110707.175333

Copyright c© 2008 by Annual Reviews.All rights reserved

1081-0706/08/1110-0211$20.00

∗Corresponding author.

Key Words

thiol-disulfide exchange, oxidative folding, RNAse A, BPTI, hirudin,oxidoreductase

AbstractDetermining the mechanism by which proteins attain their native struc-ture is an important but difficult problem in basic biology. The studyof protein folding is difficult because it involves the identification andcharacterization of folding intermediates that are only very transientlypresent. Disulfide bond formation is thermodynamically linked to pro-tein folding. The availability of thiol trapping reagents and the relativelyslow kinetics of disulfide bond formation have facilitated the isolation,purification, and characterization of disulfide-linked folding intermedi-ates. As a result, the folding pathways of several disulfide-rich proteinsare among the best known of any protein. This review discusses disul-fide bond formation and its relationship to protein folding in vitro andin vivo.

211

Click here for quick links to

Annual Reviews content online,

including:

• Other articles in this volume

• Top cited articles

• Top downloaded articles

• Our comprehensive search

FurtherANNUALREVIEWS

Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


Contents

INTRODUCTION . . . . . . . . . . . . . . . . . . 212DISULFIDE BONDS

AS A MEASURE OFPROTEIN FOLDING . . . . . . . . . . . . 213Factors That Influence Disulfide

Bond Formation in Proteins . . . . . 213Disulfide Bonds and Stability

of Proteins . . . . . . . . . . . . . . . . . . . . . 214OXIDATIVE PROTEIN

FOLDING TECHNIQUES. . . . . . . 214Choice of Redox Reagent . . . . . . . . . . 215Quenching Methods . . . . . . . . . . . . . . . 216Isolation of Disulfide

Intermediates . . . . . . . . . . . . . . . . . . . 217Characterization of

Disulfide Intermediates . . . . . . . . . 217OXIDATIVE FOLDING OF

MODEL DISULFIDE-BONDED PROTEINS . . . . . . . . . . . 218Bovine Pancreatic Ribonuclease A . . 218Bovine Pancreatic

Trypsin Inhibitor . . . . . . . . . . . . . . . 220Hirudin . . . . . . . . . . . . . . . . . . . . . . . . . . . 222

ROLE OF PROTEIN FOLDINGFACTORS IN CATALYZINGPROTEIN FOLDING . . . . . . . . . . . . 224Disulfide Bond Formation

in Bacteria . . . . . . . . . . . . . . . . . . . . . . 224Disulfide Bond Formation

in Eukaryotes . . . . . . . . . . . . . . . . . . . 226CATALYSIS OF OXIDATIVE

PROTEIN FOLDING BYPROTEIN FOLDINGFACTORS IN VITRO . . . . . . . . . . . . 227Catalysis of Oxidative Folding

by PDI . . . . . . . . . . . . . . . . . . . . . . . . . 228Catalysis of Oxidative Folding

by DsbA and DsbC . . . . . . . . . . . . . 229CONCLUSION AND FUTURE

PERSPECTIVES . . . . . . . . . . . . . . . . . 230

INTRODUCTION

Proteins are synthesized as linear polypeptidechains. Following synthesis on ribosomes, the

polypeptide chains are rapidly folded into theirunique three-dimensional structures. Properfolding is necessary for the biological function-ing of all proteins. Most purified proteins canspontaneously fold in vitro under suitable con-ditions. Thus, the information needed to spec-ify the three-dimensional structure is containedwithin the protein’s primary structure. The ki-netic processes or pathways by which proteinsadopt the native structure have been extensivelyinvestigated over the past few decades. To thisend, the focus of protein folding has been theidentification and characterization of the initial,final, and intermediate conformational states aswell as the determination of the steps by whichthey are interconverted. Most protein foldingintermediates are only transiently present, mak-ing difficult their isolation and characterizationby commonly used spectroscopic techniques.However, most secretory proteins have an im-portant covalent modification: disulfide bonds.Disulfide bonds are one of the few posttrans-lational covalent modifications that occur dur-ing protein folding. Disulfide bond formationin proteins is required not only for folding butalso for stability and function. Failure to formthe correct disulfide bonds is likely to cause pro-tein aggregation and subsequent degradation bycellular proteases.

Disulfide bonds are formed because of thereduction-oxidation chemistry of the covalentinteraction between two thiol groups. The rel-atively slow kinetics of formation of the disul-fide bond and the availability of thiol trappingreagents that rapidly quench disulfide bondformation have facilitated the isolation, pu-rification, and characterization of folding in-termediates. These trapped intermediates havebeen used to determine the pathways of severaldisulfide rich proteins in vitro. Knowledge ofthese disulfide-linked folding pathways has fur-thered our understanding of protein structure-function relationships.

Disulfide bonds can be formed sponta-neously by molecular oxygen. For instance, un-der aerobic conditions, a thin layer of cysteineis generated at the air-liquid interface when acysteine solution is left exposed to air. However,

212 Mamathambika · Bardwell

Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


this type of spontaneous, random air-oxidationreaction is very slow and cannot account for therapid rates of disulfide bond formation neededby the cell. This discrepancy led Anfinsen to thediscovery of the first catalyst for disulfide bondformation, the eukaryotic protein disulfide iso-merase (PDI). PDI is a part of the complex ma-chinery responsible for the formation and iso-merization of disulfide bonds in the eukaryoticendoplasmic reticulum (ER) (Sevier & Kaiser2002, 2006). In the prokaryotic periplasm, thesame function is carried out by the Dsb familyof proteins (Kadokura et al. 2003, Nakamoto &Bardwell 2004).

In this review, we discuss the central role ofdisulfide bonds in protein folding. The char-acterization of in vitro disulfide-linked proteinfolding pathways is studied with the help ofsmall disulfide-linked proteins. In addition, weattempt to point out the role of important pro-tein folding catalysts in catalyzing the in vitroprotein folding of these model proteins. A dis-cussion of the methodology of oxidative foldingis also included.

DISULFIDE BONDSAS A MEASURE OFPROTEIN FOLDING

During in vitro oxidative folding, disulfidebonds in proteins are formed by two thiol-disulfide exchange reactions with a redoxreagent. During the first reaction, a mixeddisulfide is formed between the protein andthe redox reagent (Figure 1a). This reactionis followed by an intramolecular attack on themixed disulfide bond in which a second cysteinethiol displaces the mixed disulfide (Figure 1b).Disulfide bonds can also be formed intramolec-ularly wherein the thiolate of a cysteine may at-tack a disulfide bond of the same protein. Thisprocess that leads to the rearrangement of disul-fide bonds within the protein is called disulfidereshuffling.

Factors That Influence Disulfide BondFormation in Proteins

Disulfide bond formation is influenced by fourmajor factors: the concentration of thiolate an-

R1

SH S–a

S S

R2

+Reducedprotein

S S R1

SH

Mixed disulfide

b

S S R1

SH

R2S–+Mixed

disulfide

Oxidizedprotein

S S

R1SH+

Figure 1Thiol-disulfide exchange reaction between a protein and a redox reagent.Disulfide bonds in proteins are formed by two thiol-disulfide exchangereactions. (a) In the first step, a thiolate anion (S−), which is formed bydeprotonation of a free thiol, displaces a sulfur atom of the redox reagent. Thisleads to the formation of a mixed disulfide bond between the protein and redoxreagent. (b) In the second step, the remaining thiol anion attacks the mixeddisulfide, leading to the formation of the oxidized protein.

ions and the accessibility, proximity, and reac-tivities of the thiol groups and disulfide bonds.During thiol-disulfide exchange, a nucleophilicthiolate group (S−) attacks a sulfur atom of adisulfide bond (-S-S-). The original disulfidebond is broken, and a new disulfide bond formsbetween the attacking thiolate and the originalsulfur atom. The ionized, thiolate form (S−) iscapable of forming the disulfide bond, whereasthe protonated thiol (SH) form is unreactive.Thiol-disulfide exchange is dependent on theconcentration of the reactive thiolate anion rel-ative to that of the unreactive thiol group, bothof which in turn are strongly dependent on so-lution pH. Thus, one of the most importantconditions that influences oxidative folding isthe pH of the refolding buffer (Scheraga &Wedemeyer 2001). The pKa of a residue is thepH at which it is 50% ionized. The amountof reactive thiolate ion decreases tenfold foreach pH unit below the pKa of the thiol. ThepKa of cysteines in denatured proteins is usuallyapproximately 8.7. Therefore, rapid oxidativefolding reactions tend to occur when the pHis above 9, and oxidative folding becomes pro-gressively slower at solution pH values belowthe pKa of the thiols involved. In vivo folding

www.annualreviews.org • Disulfide-Linked Protein Folding Pathways 213

Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


environments are not as flexible in terms of pHas in vitro reactions. One way in which organ-isms get around this limitation is by decreasingthe pKa of the reactive thiol groups in disulfideoxidoreductases so that these thiol groups re-main ionized and reactive at physiological pH.Because thiol-disulfide exchange is affected bythe reactivities of the thiolate groups and thedisulfide bond, any change in the electrostaticenvironment of the reactive group (i.e., in thepH or pKa of the thiol group) will influencedisulfide bond formation (Arolas et al. 2006,Wedemeyer et al. 2000).

Thiol-disulfide exchange reactions can oc-cur only when a thiolate and a disulfide bondcome in contact. Therefore, burial of reactivegroups in the protective tertiary structure ofa protein will inhibit disulfide bond reactions.The rate of disulfide bond formation is also in-fluenced by the proximity of the two reactivethiol groups. Thus, disulfide bond formation inproteins is influenced by multiple factors and isnot a simple chemical reaction.

Disulfide Bonds and Stabilityof Proteins

Most disulfide bonds serve to stabilize proteinstructure. It is generally accepted that proteindisulfide bonds stabilize the native conforma-tion of a protein by destabilizing the denaturedform; i.e., they decrease the entropy of the un-folded form, making it less favorable comparedwith the folded form (Thornton 1981). Accord-ing to theoretical studies, the increase in thestability of the native structure due to the for-mation of a particular disulfide bond is directlyproportional to the number of residues betweenthe linked cysteines: the larger the number ofresidues between the disulfide, the greater isthe stability imparted to the native structure(Flory 1953, Pace et al. 1988). The kinetics ofprotein folding are greatly affected by the loca-tion of the disulfide bond relative to the fold-ing nucleus. Disulfide bonds introduced in ornear the folding nucleus accelerate protein fold-ing, whereas disulfide bonds introduced else-where can decelerate folding by up to three or-

ders of magnitude (Abkevich & Shakhnovich2000).

The overall equilibrium constant Keq for athiol-disulfide exchange reaction is a measureof the stability of the protein disulfide bond.The Keq value can be as high as 105 in foldedproteins and as low as 10−3 in unfolded proteins(Darby & Creighton 1993). The formation ofdisulfide bonds is thermodynamically coupledto the process of protein folding. The foldedconformation stabilizes the disulfide bond tothe same extent that the conformation is stabi-lized by the formation of that particular disul-fide bond (Figure 2) (Creighton 1990).

In general, most disulfide bonds stabilizeproteins and affect the rate of protein fold-ing. However, a minor population of the disul-fide bonds also serves a functional role. Func-tional disulfides can be further classified intocatalytic disulfides and allosteric disulfides.Catalytic disulfides are typically found at theactive site of enzymes that mediate thiol-disulfide exchange (oxidoreductases). Thesedithiols/disulfides are transferred to a proteinsubstrate, resulting in the formation, reduction,or isomerization of disulfide bonds. Allostericdisulfides regulate function in a nonenzymaticway by mediating changes in the protein struc-ture (Hogg 2003, Schmidt et al. 2006).

OXIDATIVE PROTEINFOLDING TECHNIQUES

Oxidative protein folding is a composite pro-cess in which a reduced, unfolded protein notonly forms its native set of disulfide bonds butalso undergoes conformational folding, lead-ing to the formation of a native and biolog-ically active form (Narayan et al. 2000). In atypical oxidative folding study, proteins are ini-tially fully reduced and denatured (Creighton1986). The reducing and denaturing agents areremoved, and the protein is refolded in the pres-ence of suitable buffers containing redox agents.At various intervals, a reagent is added; thisreagent quenches disulfide bond formation andthus serves to halt the oxidative folding process.The trapped folding intermediates are then


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


separated, identified, and quantified. We nowdiscuss the various factors that have to be con-sidered for the oxidative folding of proteins.

Choice of Redox Reagent

Typically, oxidizing agents such as glutathionedisulfide (GSSG) and oxidized dithiothreitol(DTTox) are used to form disulfide bonds,and reducing agents such as the reduced formof these reagents (GSH and DTTred, respec-tively) are used to reduce and reshuffle disul-fide bonds. Normally, a redox buffer com-posed of a mixture of oxidized and reducedreagents is used during in vitro folding experi-ments. Two classes of redox reagents are com-monly used in oxidative folding: cyclic redoxreagents (e.g., DTTox/DTTred) and linear re-dox reagents (e.g., GSSG/GSH). Cyclic redoxreagents are powerful reducing agents in partbecause the close proximity of the two thiolsin these reagents leads to rapid resolution ofthe mixed disulfides between the reagents andthe proteins. This greatly simplifies interpreta-tion of oxidative folding experiments (Scheraga& Wedemeyer 2001, Wedemeyer et al. 2000).The vast majority of thiol-disulfide oxidoreduc-tases present in nature also contain two thiolsin a CXXC motif and thus function as cyclicredox reagents. Linear redox reagents such asGSSG are better oxidizing agents than aresmall-molecule cyclic redox agents owing to themore oxidizing redox potential of the former.However, with linear disulfide reagents, themixed disulfide species are much more stablethan with cyclic disulfide reagents and can beproblematic because they can accumulate sig-nificantly during folding reactions. High con-centrations of linear redox reagents can evenblock all the free thiols, interrupting oxidativefolding.

The composition of the redox buffer candrastically affect the rate of in vitro oxidativeprotein folding. For example, hirudin, a pro-tein with three disulfides, can be refolded withinseconds in a buffer of optimal redox composi-tion, whereas refolding can take as long as 24 hin a buffer containing only catalytic amounts

Disulfides

SH

SH

Unfolded

Folded

Thiols

Thiol disulfide equilibria (KSS

)

Co

nfo

rma

tio

na

le

qu

ilib

ria

(K F

)

S

S

SH SH S S

KSS

U

KF

SH KF

SS

KSS

F

=

ΔΔGF = ΔΔGSS

KFSS KSS

F

KFSH KSS

U

Equation 1: relation between disulfide bond formation and conformational folding

Net free-energy change in the process is zero:

Figure 2Relationship between disulfide bond formation and conformational folding of aprotein. A protein with two cysteine residues is shown in its unfolded andfolded conformations. The stability conferred by the disulfide bond and theconformation is given by the equilibrium constants KSS and KF, respectively.The equilibrium constants for folding with and without the disulfide bond aregiven by K SS

F and K SHF , respectively. The stability of the disulfide bond in the

folded and unfolded states is given by K FSS and K U

SS, respectively. The netfree-energy change around the cycle should be zero, and therefore theequilibrium constants are linked (Equation 1). Thus, the folded conformationaffects the stability of the disulfide bond to the same extent to which thedisulfide bond affects the stability of the conformation (Creighton 1990).

GSSG: glutathionedisulfide/oxidizedglutathione

DTT: dithiothreitol

GSH: reducedglutathione

of a thiol (Chang 1994, Chatrenet & Chang1993).

Efforts have been made to develop novelredox reagents. Some of them include redox-active cyclic bis(cysteinyl)peptides (Cabreleet al. 2002); ortho- and meta-substituted aro-matic thiols (Gough et al. 2006); selenoglu-tathione (GSeSeG), an analog of glutathionethat contains a diselenide bond in place ofthe natural disulfide (Beld et al. 2007); and


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


BMC: dithiol( ± )-trans-1,2-bis(2-mercaptoacetamido)cyclohexane

RNase A: bovinepancreaticribonuclease A

AEMTS:aminomethylthiosul-fonate

NEM:N-ethylmaleimide

AMS:4-acetamido-4′-maleimidylstilbene-2,2′-disulfonic acid

the dithiol ( ± )-trans-1,2-bis(2-mercaptoaceta-mido)cyclohexane (BMC), which acts as asmall-molecule mimic of PDI (Woycechowskyet al. 1999).

Oxidative folding through nonredox chem-istry has been explored recently. Curcumin,an antioxidant and an anticancer chemother-apeutic, was successfully used to accelerateoxidative folding of bovine pancreatic ribonu-clease A (RNase A) (Gomez et al. 2007). Ox-idized BMC, when used in combination withnonredox-active molecules [trimethylamine-N-oxide (TMAO) and trifluoroethanol (TFE)],accelerated the oxidative folding rate of RNaseA as compared with that achieved by oxidizedBMC alone (Fink et al. 2008). In both cases,the increase in the oxidative folding rate wasattributed to the ability of nonredox-activemolecules (curcumin, TMAO, TFE) to inducenative-like elements in the reduced protein andstabilize key folding intermediates that haveproperties of the native protein.

The optimal redox buffer for in vitro proteinfolding is empirical for each protein and may re-flect unknown differences between the in vivoand in vitro conditions, different in vivo fold-ing environments, or the fact that every proteinhas a unique folding pathway. Not surprisingly,optimal folding is generally observed under re-dox conditions in which the conformational andoxidative kinetic traps (such as mixed disulfidesleading to the formation of dead-end interme-diates) are kept to a minimum (Kibria & Lees2008).

Quenching Methods

By the use of appropriate thiol quenchingreagents, disulfide bond formation can bestopped, and disulfide intermediates can then beisolated, enabling the detailed study of proteinfolding pathways. An ideal quenching method,according to Wedemeyer et al. (2000), should(a) completely quench free thiols, (b) impose along-term block on disulfide reshuffling to al-low for efficient fractionation, (c) aid in the sep-aration and isolation of disulfide intermediates,(d ) be reversible so that refolding of interme-

diates can be resumed following their isolation,and (e) prevent unfolding of structured disulfideintermediates.

Quenching is usually achieved by oneof four methods: acid quenching, blockingwith maleimides, blocking with alkyl halides,or blocking with aminomethylthiosulfonate(AEMTS). Acid quenching is fast and re-versible, and the protein does not undergoany covalent modification. This method alsohas the advantage that it denatures proteins,which simultaneously allows access to all thiolgroups and also destroys any special reactivi-ties by bringing all thiol pKas into the nor-mal range. However, this first method justgreatly slows thiol reactions, by tenfold foreach pH unit of the buffer below that of thepKas of the reactive thiols; it does not com-pletely stop thiol-disulfide exchange. Block-ing with maleimides such as N-ethylmaleimide(NEM) or 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonic acid (AMS) is rapid and very spe-cific for cysteine residues (Zander et al. 1998).Maleimides, however, block irreversibly and donot always gain access to all free thiols. Iodoac-etate and iodoacetamide are the commonly usedalkyl halides. When used at low-millimolar con-centrations, alkyl halides often block too slowlyto completely prevent disulfide reshuffling. Athigher concentrations, alkyl halides can mod-ify other residues of the protein (Scheraga &Wedemeyer 2001). Blocking by AEMTS is re-versible and rapid [five orders of magnitudefaster than iodoacetate (Rothwarf & Scheraga1991)], and this compound adds a positivelycharged cysteamine group for every blockedthiol. Acid quenching and AEMTS blockinghave the disadvantage that they usually perturbthe conformational structure of folding inter-mediates. Thus, no single quenching methodmeets all of Wedemeyer et al.’s (2000) criteria.The use of different reagents can generate verydifferent results. For instance, the slow reac-tion rate of alkyl halides is thought to be one ofthe chief reasons behind fundamental disagree-ments between Creighton and Kim (Creighton1988, 1990; Weissman & Kim 1991) on thenature of the intermediates in the bovine


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


pancreatic trypsin inhibitor (BPTI) foldingpathway. One approach is to use a combina-tion quench methodology, an initial quench us-ing acid, to exploit its ability to quench rapidlyand denature proteins, followed by dilution intoa buffer containing a chemical quench reagentsuch as a maleimide (Zander et al. 1998).

Isolation of Disulfide Intermediates

Disulfide intermediates trapped during oxida-tive folding can be fractionated and character-ized to determine the number of free thiols anddisulfides (Weissman & Kim 1991). This is thefirst step in determining disulfide connectivityfor each of these folding intermediates, whichcan lead to a detailed study of the folding path-way. Separation of intermediates becomes anincreasing problem as the number of disulfidebonds increases. A protein with three disulfidesin its mature form has 15 possible one-disulfideintermediates, a protein with four disulfides has28, and a protein with five has 45. This explo-sion in the number of intermediates as the num-ber of disulfides increases has limited the de-tailed analysis of folding pathways for proteinswith more than three disulfide bonds.

The intermediates of acid quenching mustbe analyzed at low pH, which eliminates manyseparation techniques. Disulfide-linked fold-ing intermediates trapped by acid quenchingcan in some cases be separated by reversed-phase high-performance liquid chromatogra-phy at low pH. However, not all proteins canbe acid quenched owing to stability issues, andfor many proteins, not all intermediates canbe separated from each other. Intermediatesblocked with AMS or polyethylene-maleimidecan be separated by SDS-PAGE because of thelarge mass of the modification group added.Intermediates blocked with biotin polyethy-lene glycol-maleimide can be isolated on avidincolumns. Intermediates blocked with iodoac-etate have the thiols in the form of S-carboxymethylcysteine and the disulfides in theform of cysteine. These intermediates can beseparated by chromatography, and the fractionof the free thiols and disulfides can be deter-

BPTI: bovinepancreatic trypsininhibitor

ESMS: electrospraymass spectrometry

MALDIMS: matrix-assisted laserdesorption andionization massspectrometry

mined by amino acid analysis (Chang & Knecht1991). Intermediates blocked with AEMTShave a positively charged 2-aminoethanethiolgroup for every free thiol present on the pro-tein, and therefore ion exchange chromatog-raphy can be used to separate these interme-diates. In general, a wide variety of analyticaltechniques, such as gel filtration, capillary elec-trophoresis, and 2-D gel electrophoresis, can beapplied to separate the quenched intermediates.

Characterization ofDisulfide Intermediates

The ensemble of intermediates and the disul-fide bonds formed at various steps along thefolding pathways of bovine pancreatic RNaseA and three-fingered toxins have been char-acterized by mass spectrometry (Ruoppoloet al. 1996a,b). Electrospray mass spectrome-try (ESMS) can be used to determine the num-ber of disulfides on the basis of the increase inmolecular weight by the addition of a block-ing reagent for each of the thiol groups. Therelative abundance of different intermediatescan also be determined, facilitating kinetic anal-ysis of the intermediates. To assign the cor-rect disulfide pairings, ESMS analysis can befollowed by disulfide mapping of proteolyti-cally digested intermediates through the useof matrix-assisted laser desorption and ioniza-tion mass spectrometry (MALDIMS) or liquidchromatography ESMS (Ruoppolo et al. 2005).

A multitude of other biophysical techniquescan be used to characterize the structure, activ-ity, and role of the intermediates in the fold-ing pathway. For example, fluorescent reso-nance energy transfer and nuclear Overhausereffects (NOEs) of nuclear magnetic resonance(NMR) spectroscopy can be used to measureinter-residue distances to understand how thestructure of the protein changes as it folds to itsnative structure (Elisha 2005, Roques et al.1980). Structural characterization of disulfideintermediates has been used to understand theoxidative folding pathway of some disulfide-rich proteins such as RNase A and BPTI(Weissman & Kim 1991). The high-resolution


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


structure of those rare intermediates that arekinetically stable has occasionally been de-termined with X-ray crystallography (Pearsonet al. 1998) and NMR (Laity et al. 1997, vanMierlo et al. 1994).

The role of particular disulfides in the fold-ing and stability of particular proteins canbe studied by site-specific mutagenesis experi-ments that replace the corresponding cysteineswith serines or alanines. In some cases, these ex-periments can generate stable analogs of foldingintermediates, which can then be structurallycharacterized.

Kinetic analysis of the regeneration pro-cess can be carried out by modeling the con-centrations of various disulfide intermediates.Detailed kinetic analysis has been carried outfor the regeneration of RNase A (Rothwarfet al. 1998a) and hirudin variant 1 (rHV1)(Thannhauser et al. 1997) by the method devel-oped by Konishi and colleagues (Konishi et al.1982, Rothwarf & Scheraga 1993a, Scheragaet al. 1987). In this method, the relative concen-trations of the various intermediate species aredetermined as a function of time and redox con-ditions. These data are then modeled to a fold-ing pathway consistent with the regenerationkinetics (Konishi et al. 1982). This approachcan be used to determine the rate constant forthe regeneration of native protein as well as therate constants for the formation and reductionof disulfide bonds at different stages of regener-ation (Rothwarf et al. 1998a, Thannhauser et al.1997).

OXIDATIVE FOLDING OF MODELDISULFIDE-BONDED PROTEINS

The oxidative folding pathway of several smalldisulfide-rich proteins has been determined.Here, we discuss the pathway of folding ofthree of the most extensively characterized pro-teins: RNase A, BPTI, and hirudin. The knownpathways of oxidative folding exhibit a highdegree of diversity, as revealed by the disul-fide heterogeneity of folding intermediates, thepredominance of native disulfide bonds in inter-mediates, and the level of accumulation of fully

oxidized but scrambled isomers as intermedi-ates (Chang 2004). Nevertheless, the pathwaysof oxidative folding can be classified broadlyinto two major types on the basis of the nature ofthe folding intermediates (Narayan et al. 2000).First, an oxidative folding pathway in whichthe intermediates are unstructured is called adesU pathway (des refers to the disulfide speciesor ensemble that possess all but one of thenative disulfide bonds). The rate-determiningstep in a desU pathway is the formation of thenative protein N by oxidation and conforma-tional folding of the disulfide intermediates.Second, a pathway in which the des species arestructured is called a desN pathway. The na-tive disulfide bonds in desN species are buriedin the stable tertiary structure, but the freethiol groups are kept readily accessible to formthe native protein. The rate-determining stepin such desN pathways is the formation of thestructured des species by disulfide reshufflingand conformational folding of the unstructuredprecursors.

Bovine Pancreatic Ribonuclease A

RNase A is a single-domain protein with fournative disulfide bonds (26-84, 40-95, 58-110,and 65-72). Over the years, a multitude of phys-ical and chemical methods have been used togain insight into the folding of RNase A. Boththe NMR and crystal structures of the fullyfolded protein have been determined (Santoroet al. 1993, Wlodaver et al. 1988). One struc-tural complication is that Pro93 and Pro114 inRNase A adopt the less common cis conforma-tion in the folded state, making cis-trans prolineisomerization an important part of the foldingprocess. This problem, along with difficulties inexpressing RNase A at high levels in bacterialhosts, has complicated detailed analysis of itsfolding pathway. Nevertheless, RNase A is oneof the best-studied models for disulfide-linkedfolding; for reviews, see Narayan et al. 2000,Scheraga & Wedemeyer 2001, and Wedemeyeret al. 2000.

Scheraga and coworkers have exten-sively characterized the process of sequential


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


formation of its disulfide bonds and the acquisi-tion of native structure in fully reduced RNaseA. Reoxidation in the presence of a redoxagent (DTTox) resulted in the formation ofheterogeneous disulfide-bonded intermediates(scrambled RNase A). According to this model(Figure 3a) (Iwaoka et al. 1998; Rothwarfet al. 1998a,b; Scheraga & Wedemeyer 2001;Wedemeyer et al. 2000; Xu & Scheraga1998), RNase A folds through two stages at25◦C.

During the first prefolding stage, reducedRNase A undergoes sequential oxidation, re-sulting in the formation of four unstructureddisulfide ensembles containing 1–4 disulfidebonds (ensembles 1S, 2S, 3S, and 4S). The com-position of the disulfide bonds among 1S and2S intermediates is nonrandom (Volles et al.1999, Wedemeyer et al. 2002, Xu et al. 1996).The intermediate that contains a disulfide be-tween cysteines 65 and 72 (called here the [65-72] intermediate) is the most populated species,accounting for 40% of the 1S ensemble. Therepresentation of all other individual 1S inter-mediates is less than 10%. The preference forthe [65-72] intermediate is due not only to en-tropic stabilization but also presumably to thesignificant enthalpic stabilization offered by theformation of a β-turn-like structure in residues65–68. This β-turn is also observed in the na-tive protein (Laity et al. 1997, Wlodaver et al.1988) and in the NMR structure of the mu-tant (C40A, C95A) of RNase A (Laity et al.1997). Because this β-turn structure is presentin the fully folded protein, it may serve as achain-folding initiation site. The [65-72] disul-fide intermediate is also predominant amongthe 2S intermediates. Thus, the [65-72] disul-fide bond may act to accelerate protein foldingby decreasing the conformational space that hasto be scanned for RNase A to attain native struc-ture (Volles et al. 1999).

In the second stage, structured disulfide in-termediates form and are subsequently con-verted to the native form. In RNase A, therate-determining step in the regeneration of na-tive protein is the formation of two disulfidespecies with native-like structure, des[40-95]

R 1S 3S 4S2S

des[40-95] des[65-72]

N

des[26-84]

des[58-110]

R 1S 3S2S

N

a

b

Figure 3The folding pathways of wild-type RNAse A and its three-disulfide mutants. Rrepresents the reduced protein, and nS represents the ensemble of species withn disulfide bonds. des[ ] represents a disulfide species with native disulfidebonds but lacking the disulfide bond in the brackets, and N represents thenative protein. (a) Wild-type RNAse A follows a desN type of pathway. Therate-limiting step in the formation of RNAse A at 25◦C is the formation of twodes species, [40-95] and [65-72], from the 3S ensemble. At 15◦C, two other desspecies, [26-84] and [58-110], are formed; these reshuffle slowly to des[40-95]and des[65-72] via the 3S ensemble (Iwaoka et al. 1998; Rothwarf et al.1998a,b; Scheraga & Wedemeyer 2001). (b) The three-disulfide mutants ofRNase A follow a desU pathway. The rate-limiting step is the formation of thetwo native proteins by oxidation and conformational folding of theunstructured 2S ensemble (Scheraga & Wedemeyer 2001).

and des[65-72]. These two species are formedfrom the 3S ensemble by disulfide reshuffling(Figure 3a). Upon the formation of thesespecies, RNase A mainly attains a locked-inconformation wherein the three native disul-fide bonds are protected from further reduc-tion/reshuffling. However, the two remainingthiol groups still remain largely accessible to thesolvent, thus allowing them to undergo rapidoxidation to the native protein.


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


Conformational stability of the des species iscritical for oxidative folding in RNase A. Con-ditions that destabilize the structure of the desspecies (e.g., high temperatures) greatly hin-der the regeneration of RNase A (Rothwarf &Scheraga 1993b, Rothwarf et al. 1998b). How-ever, the rate of regeneration can be restoredby anions that restabilize the conformationalstructure (e.g., phosphate, fluoride) (Lawrenceet al. 2000, Low et al. 2002). Thus, the native-like conformation conferred by the des speciesis important for the oxidative folding of RNaseA, in part because the conformational struc-ture of the intermediates protects the nativedisulfides from rearrangement (Narayan et al.2000).

Oxidative folding studies of the mutants ofRNase A having only three disulfides, [C40A,C95A] and [C65S, C72S], reveal that nativeprotein is formed through direct oxidation andconformational folding of the 2S ensemble.The des species (the 2S ensemble) have very lit-tle structure, and therefore these three-disulfidemutants of RNase A follow the desU pathway(Figure 3b).

Native RNase A can also be generatedthrough the oxidation of two other des species,des[26-84] and des[58-110] (Figure 3a). Butthese two species have a stable conforma-tion only in the presence of stabilizing salts[phosphate (Low et al. 2002)] or at low tem-peratures [≤15◦C (Welker et al. 1999)]. Evenat low temperatures, des[26-84] and des[58-110] are metastable intermediates that reshuf-fle preferentially to the 3S ensemble ratherthan directly oxidize to form the native pro-tein (Welker et al. 2001). The burial of theirfree thiol groups in their hydrophobic coresof a native-like structure presumably inhibitsany redox reactions of these thiols in theseintermediates.

Bovine Pancreatic Trypsin Inhibitor

BPTI is a member of the serine protease familyof inhibitors; it is a very small globular pro-tein, 58 amino acid residues in length in its ma-ture form. BPTI adopts a tertiary fold compris-

ing two strands of antiparallel β-sheet and twoshort segments of α-helix. It contains three sta-bilizing disulfide bonds in its structure: betweencysteines 5 and 55, 14 and 38, and 30 and 51.

The oxidative folding of BPTI was one ofthe first protein folding pathways to be studiedand remains among the best characterized. His-torically, there have been two models of BPTIfolding, which differ greatly in the role playedby nonnative intermediates in the foldingpathway.

Early studies on the oxidative folding ofBPTI were carried out by trapping the disul-fide bond folding intermediates with the use ofalkyl halides (Creighton 1988, 1990; Creightonet al. 1996). Using this approach, Creighton andcolleagues found a heterogeneous population ofunfolded molecules that fold by a distinct path-way via disulfide reshuffling of one-disulfideintermediates (Figure 4a). The native-like[30-51] intermediate composed 60% of theone-disulfide molecules, the nonnative [5-30]intermediate composed another 30% of theone-disulfide molecules, and the remainingpossible 13 disulfide species accounted for therest of the one-disulfide molecules (BPTI canform 15 unique 1S intermediates). The one-disulfide intermediates appear to be in an equi-librium state prior to folding. The [30-51] inter-mediate is kinetically significant for the foldingof BPTI, and all further intermediates retainthis disulfide bond. In native BPTI, the [30-51]disulfide links a major α-helix to a β-sheet, andthe interaction between these secondary struc-ture elements can stabilize the protein. Theother native disulfide, [5-55], appeared in only3% of the one-disulfide intermediates, whereasthe [14-38] disulfide was not present at de-tectable levels in the 1S and 2S ensembles. Ac-cording to this model, the rate-limiting step inBPTI refolding is the formation of the quasi-native two-disulfide intermediate [30-51, 5-55].Once this native-like intermediate is formed,the third native disulfide, [14-38], is rapidlyformed; its thiols are held in proximity on themolecule surface.

In a subsequent study, using acid trap-ping followed by high-performance liquid


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


a

R

[30-51, 5-14]

des[30-51] des[5-55]

des[14-38] N

[30-51, 5-38]

++

1S

[30-51][5-55]

[30-51]

[5-55]

des[5-55]

des[30-51]

des[14-38] N

N'

N*

NSH

SH

b

R

Figure 4Pathways of folding of BPTI. nS represents the ensemble of species with n disulfide bonds. (a) The reducedprotein (R) folds by forming both native and nonnative disulfide bonds. The major intermediates areindicated by the brackets. According to the model, the rate-limiting step in the regeneration of native (N)BPTI is the formation of the quasi-native species des[14-38]. The one-disulfide species, [5-55], leads to theformation of a dead-end intermediate, des[30-51] (Creighton 1990). (b) As per this model, native BPTI isformed mainly by intermediates that contain native disulfide bonds. The rate-limiting step is the formationof the des[14-38] species from two kinetic traps, N′ and N∗ (Weissman & Kim 1991).

chromatography separation, Weissman & Kim(1991) reexamined the folding pathway of BPTIand proposed a revised model of folding. Unlikethe earlier model of Creighton and colleagues,Weissman & Kim found that nonnative inter-mediates are not populated significantly duringfolding. Only six well-populated native disul-fide species were isolated in their experiments(Figure 4b). Upon oxidation, reduced BPTIrapidly forms one-disulfide intermediates thatthen rearrange rapidly to form either of twonative intermediates, [30-51] or [5-55]. Bothof these intermediates presumably have sub-stantial native-like structure. The thiols of cys-teines 14 and 38 are solvent exposed and read-

ily form the [14-38] disulfide bond to gener-ate two two-disulfide species: [30-51; 14-38](designated N′) and [5-55; 14-38] (designatedN∗). The two-disulfide species serve as kinetictraps for BPTI folding. The N∗ intermediateis very stable and does not undergo significantrearrangement. The other intermediate, N′, ishighly stable and rearranges slowly to form ei-ther NSH

SH or N∗. NMR spectra of native BPTIand the N′ intermediate indicate that N′ is verynative-like in configuration, burying the thiolsof cysteines 5 and 55, thereby inhibiting thedirect oxidation to native protein and explain-ing the observed slow disulfide rearrangementwithin N′. The rearrangement of N′ to either


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


NSHSH or N∗ occurs via the formation of non-

native intermediates [30-51; 5-14] and [30-51;5-38]. Upon the formation of NSH

SH, the pro-tein is rapidly oxidized to form the third nativedisulfide bond, [14-38].

During oxidative folding of BPTI, only halfof the molecules form native BPTI. The re-maining molecules stay kinetically trapped inthe form of the stable intermediate N∗. Al-though N∗ lacks the native [30-51] disulfide,NMR and crystallographic studies reveal thatit has a structure similar to that of native BPTI.This native structure buries cysteines 30 and51, rendering them inaccessible to oxidizingagents. Consequently, N∗ is a dead-end inter-mediate that can persist for weeks.

Weissman & Kim (1995) explained the pref-erence of N′ to undergo rearrangement to formNSH

SH and then N over direct oxidation to formthe third native disulfide [5-55] by measuringthe rate of direct oxidation of N′. The thiolsin N′ are buried and constrained by the struc-ture of this intermediate, preventing any directoxidation. As a result, any formation of nativeprotein from N′ requires substantial unfold-ing, and N′ unfolds and rearranges its disul-fide bonds before forming NSH

SH. Unlike N∗,which leads to a nonproductive pathway, therearrangement of N′ to NSH

SH leads to the pro-ductive folding of BPTI. Very recently, Kibria& Lees (2008) reexamined the folding path-way of BPTI, using optimized concentrationsof glutathione (5 mM GSSG and 5 mM GSH).Under these conditions, the N∗ intermedi-ate was decreased in concentration, leadingto substantially faster BPTI folding (Kibria &Lees 2008).

BPTI is initially synthesized as a pre-proform containing a signal sequence required forthe secretion of BPTI into the ER. This se-quence is cleaved following secretion. BPTIalso has a pro sequence that greatly influencesthe folding of BPTI in part because it containsa cysteine residue that is involved in disulfideisomerization reactions, with cysteines presenton the mature portion of BPTI (Weissman &Kim 1992). This pro region is cleaved off ofBPTI following maturation. Unfortunately, the

vast majority of the work on BPTI folding hasbeen with the less physiologically relevant ma-ture protein.

Hirudin

Hirudins are a family of thrombin-specificprotease inhibitors isolated from the medici-nal leech Hirudo medicinalis. Members of thisfamily of proteins contain approximately 65residues and share three highly conserved disul-fide bonds: C6-C14, C16-C28, and C22-C39.Hirudin contains an N-terminal globular do-main (residues 1–49) that binds to the catalyticsite of thrombin and a disordered, acidic C-terminal domain (residues 50–65) that interactswith the fibrinogen recognition site of the en-zyme (Chang 1983, Grutter et al. 1990, Rydelet al. 1990).

Investigators have determined the in vitrooxidative folding pathway of a recombinantvariant of hirudin (rHV1) in both the presenceand the absence of a redox reagent (Chatrenet& Chang 1992, 1993; Thannhauser et al. 1997).Two similar models of oxidative folding forhirudin have been proposed on the basis of theseexperiments.

Chatrenet & Chang (1992) proposed thetrial-and-error mechanism of folding forhirudin on the basis of refolding experimentsin the absence of a redox couple such asGSSG/GSH or DTTox/DTTred. Oxidationwas achieved by dissolved atmospheric O2,and a reductant (β-mercaptoethanol/GSH)was used to achieve full regeneration of nativehirudin from the mixture of nonnative three-disulfide hirudin intermediates (scrambledhirudin). According to this model, all cysteinesof hirudin participate in disulfide shufflingthroughout the folding process, and the foldingproceeds via a mechanism of trial and errorwithout preferred pathways (Chatrenet &Chang 1992).

In a similar study, using a fragment ofrHV1 (residues 1–49), Chatrenet & Chang(1993) proposed a sequential biphasic pathwayfor the folding of hirudin. Unlike BPTI andRNase A, the folding intermediates revealed an


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


exceedingly high heterogeneity among thedisulfide isomers. Of the 60 theoretically pos-sible one- and two-disulfide intermediates, atleast 30 fractions have been identified. Amongthe 14 possible three-disulfide scrambled iso-mers, 11 are present as folding intermediatesand have been characterized. Hirudin spon-taneously and sequentially flows from fullyreduced protein to one-disulfide isomers totwo-disulfide isomers to three-disulfide iso-mers (Figure 5a). The first stage of foldinginvolves the packing of the polypeptide chaindriven mostly by hydrophobic collapse. Start-ing with the reduced and unfolded hirudin,disulfides are randomly paired sequentiallyand irreversibly to form one-, two-, and fi-nally three-disulfide intermediates (scrambledhirudin). During the second stage, the nativeprotein is formed by disulfide reshuffling ofthe scrambled three-disulfide intermediates (β-mercaptoethanol was used as a reductant). Thisprocess is driven by noncovalent specific inter-actions that stabilize the native protein. Takentogether, the proposal suggests that refoldingof hirudin is dependent on the consolidation(disulfide rearrangement) of a heterogeneousscrambled population of three-disulfide specieswithout any preferred pathway.

Thannhauser et al. (1997) studied the kinet-ics of folding of rHV1 under anaerobic condi-tions in the presence of DTTox and DTTred.According to their kinetic model (Figure 5b),the reduced protein and the disulfide bond in-termediates (1S, 2S, and 3S) rapidly approacha pre-equilibrium steady state during refold-ing. Unlike the earlier model of Chatrenet &Chang (1992, 1993), in which the 3S scrambledspecies is directly converted to native species,the 2S ensemble undergoes oxidation to forma 3S∗ ensemble that presumably has the samedisulfide bonds as does the native protein, butmay possess a different conformation. The 3S∗

ensemble folds rapidly to the native state. Therate of regeneration of the native protein wasdependent on both the concentration of the2S ensemble and the DTTox concentration.Therefore, Thannhauser et al. (1997) suggestedthat the rate-determining step in the regen-

R 1S 3S2S N

a

b

R 1S 3S2S

3S* N

Figure 5Oxidative folding of two variants of recombinant hirudin (rHV1). (a) Thepathway of folding of a fragment of rHV1 (1–49 residues) follows a desU type ofpathway. The reduced protein (R) undergoes successive oxidation to form amixture of highly heterogeneous intermediates (1S, 2S, and 3S, where nSrepresents the ensemble of species with n disulfide bonds). This process offormation of intermediates is driven by hydrophobic collapse. The rate-limiting step is the reshuffling of the scrambled 3S species to form the nativeprotein (N) (Chatrenet & Chang 1992, 1993). (b) The pathway of folding ofrHV1 determined by kinetic fitting. The reduced protein (R) undergoessuccessive oxidation to form the intermediates (1S, 2S, and 3S). Therate-limiting step is the irreversible formation of a 3S∗ species by the oxidationof the 2S ensemble and is dependent on the concentrations of the 2Sintermediate and the redox reagent (Thannhauser et al. 1997).

eration of rHV1 is the oxidation of one ormore species of the 2S ensemble to form 3S∗.Moreover, these researchers postulated that the2S species should contain at least two nativedisulfide bonds ([6-14; 16-28], [6-14; 22-39], or[16-28; 22-39]), raising the possibility of the ex-istence of three distinct regeneration pathways.However, this analysis could not determine thespecific composition of the 2S disulfide ensem-ble that is involved in the rate-limiting step.

The different model proteins and tech-niques used to derive data led to differencesin these two models of hirudin folding. Un-like Thannhauser et al. (1997), who used rHV1and a redox couple to oxidize rHV1, Chatrenet& Chang (1992) used a truncated rHV1(1–49 residues) and dissolved atmospheric O2

in their experiments. rHV1(1–49) regeneratesnative structures at a different rate than doesrHV1. The importance of 3S scrambled species


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


in the earlier model may be because of the use ofhighly oxidizing agents that caused the reduc-tion of the 3S ensemble to become rate-limitingin the regeneration process.

Taken together, the data suggest that ox-idative folding of hirudin is characterized bythe presence of heterogeneous folding inter-mediates, indicating that the protein follows adesU -type folding pathway. The prominence ofscrambled isomers along the folding pathwaycontradicts the conventional wisdom that theyare off-pathway or dead-end intermediates.

ROLE OF PROTEIN FOLDINGFACTORS IN CATALYZINGPROTEIN FOLDING

Disulfide-bonded proteins can be folded invitro with the help of redox reagents that as-sist the process of thiol-disulfide exchange. Invivo, proteins are aided by molecular chaper-ones that protect them from forming insolubleaggregates. Disulfide bond formation is one ofthe major rate-limiting steps in protein foldingin vivo. Therefore, in addition to the molecu-lar chaperones, catalysts that can catalyze thethiol-disulfide exchange at rates comparable toin vivo protein folding rates are required. Thissection reviews the disulfide folding catalystsfound in bacteria and eukaryotes and the roleof these catalysts in catalyzing protein foldingand directing folding pathways in vitro.

Disulfide Bond Formation in Bacteria

The process of disulfide bond formation in thebacterial periplasm has been extensively studiedover the past decade. In Escherichia coli, disul-fide bonds are introduced in the periplasm bythe Dsb (disulfide bond formation) family ofproteins. The family includes DsbA and DsbB,which are involved in forming disulfide bonds,and DsbC and DsbD, which are involved in iso-merizing disulfide bonds.

Disulfide bond formation: DsbA and DsbB.DsbA, a 21-kDa soluble protein, is the immedi-ate donor of disulfide bonds to proteins secreted

into the E. coli periplasm (Bardwell et al. 1991).In the absence of DsbA, most of the periplas-mic proteins have their thiols in the reducedform. DsbA consists of a thioredoxin-like foldand a CXXC motif as its active site. For DsbAto be active, its two cysteines must be in theoxidized state. In general, disulfide bonds sta-bilize proteins. However, oxidized DsbA is un-stable compared with reduced DsbA and re-acts rapidly with unfolded proteins (Zapun et al.1993). This instability of the oxidized protein,along with the instability of the mixed disulfideintermediate with a target protein, provides athermodynamic driving force for the transfer ofdisulfide bonds from DsbA to the target protein(Figure 6). The transfer of disulfides by DsbAcan be thought of as a thiol-disulfide exchangereaction of a protein with a redox reagent. Thistransfer reaction has two steps. First, an un-stable mixed disulfide is formed between DsbAand a target protein. Then, another thiol groupin the target protein attacks the mixed disul-fide. This results in the formation of a disul-fide bond in the target protein and the even-tual reduction of the active-site cysteines ofDsbA.

DsbA is one of the most oxidizing thiol-disulfide oxidoreductases known, with a redoxpotential of −120 mV. The oxidizing power isattributed to unusual electrostatic properties ofthe CXXC motif, particularly the unusually lowpKa of the most N-terminal cysteine in this ac-tive site, Cys30. Cys30 has a pKa of ≈ 3, whereasthe pKa of most cysteine residues is between 8and 9 (Grauschopf et al. 1995). Owing to itslow pKa, Cys30 is almost entirely in the thio-late anion state at physiological pH. The thio-late anion is also stabilized by hydrogen bonds,electrostatic interactions, and helix dipole in-teractions; an electrostatic interaction betweenHis32 and Cys30 is thought to be the mostimportant stabilizing influence (Guddat et al.1997, Martin et al. 1993). Mutations that alterHis32 greatly decrease the oxidizing power ofDsbA by altering the pKa of Cys30 (Grauschopfet al. 1995). Other residues that are locatednear the active disulfide, such as Pro31, alsocontribute to the oxidizing power of DsbA but


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


DsbA DsbA

Substrate Substrate

UQ MQDsbB

e–

Aerobic conditions Anaerobic conditons

Misfoldedsubstrate

a b

e–e–

Foldedsubstrate

SH SH

DsbC

SHS SS SSH

Membrane

Periplasm

CytoplasmSH SH

TrxA

DsbD

e–

e–

S S SH SH

SH SH S S

Figure 6Disulfide bond formation and isomerization in the Escherichia coli periplasm (Pan & Bardwell 2006). (a) DsbAintroduces disulfides into newly secreted proteins. The inner-membrane protein DsbB reoxidizes the activesite of DsbA. Under aerobic conditions, the electrons from DsbB are passed to ubiquinone (UQ), and underanaerobic conditions they are passed to menaquinone (MQ). The electrons eventually flow to molecularoxygen via the electron transport chain. (b) Nonnative disulfide bonds are reshuffled by DsbC. DsbC is keptreduced by another inner-membrane protein, DsbD, which receives its electrons from the periplasmicthioredoxin (TrxA) system.

are less important. DsbA, in addition to beingstrongly oxidizing thermodynamically, also ox-idizes proteins with very fast kinetics. Investi-gators have solved the three-dimensional struc-ture of DsbA both in the oxidized and reducedstates (Guddat et al. 1998, Martin et al. 1993),revealing a helical domain embedded into athioredoxin domain. A rotational motion oc-curs between the two domains upon substratebinding.

Following transfer of its disulfide bond tothe target protein, DsbA is reduced. To regainits ability to transfer disulfide bonds, DsbA mustbe reoxidized by DsbB (Figure 6). DsbB re-moves electrons from DsbA and transfers themto the respiratory chain. In aerobic conditions,the electrons are passed on to ubiquinone and

ultimately to molecular oxygen. However, inanaerobic conditions, electrons are accepted bymenaquinone, which transfers them to nitrate,nitrite, and fumarate (Bader et al. 1999, 2000;Kobayashi et al. 1997; Xie et al. 2002). There-fore, DsbA is reactivated via thiol-disulfide ex-change reactions between the active sites ofDsbA and DsbB (Inaba et al. 2006).

Disulfide bond reshuffling: DsbC andDsbD. DsbA is a very strong oxidase and in-troduces disulfide bonds into proteins relativelynonspecifically. Therefore, DsbA can introducenonnative disulfides into proteins with multi-ple cysteines (Rietsch et al. 1996). Formation ofthe native set of disulfide bonds is essential forattaining the proper folded conformation. To


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


accomplish this, an enzyme or redox reagentthat catalyzes the reduction and reshuffling ofdisulfide bonds is required.

In E. coli, a protein called DsbC acts to iso-merize incorrectly oxidized proteins. DsbC isa two-domain 23.5-kDa protein. It has a C-terminal thioredoxin-like domain and an N-terminal dimerization domain. The dimericprotein is V-shaped, with an uncharged cleftthat is thought to be involved in peptide bind-ing (Collet et al. 2002, McCarthy et al. 2000).Each arm of the V consists of an N-terminal do-main and a C-terminal catalytic domain. TheN-terminal domains of each monomer formthe dimer interface at the base of the V. TheC-terminal catalytic domain has a thioredoxin-like fold. The sulfur of the first cysteine of theactive-site CXXC motif (Cys98) is partially sol-vent exposed and is therefore able to form amixed disulfide bond with a substrate. DsbCpresumably functions by detecting hydropho-

+ +

+

A

B

Misfoldedsubstrate

Foldedsubstrate

SH SH

DsbC

SS SS SSH S SSH

S SH

DsbC

SH SH

DsbC

S S

DsbCReducedsubstrate

SH SHSH

Figure 7DsbC isomerizes wrongly formed disulfide bonds (Nakamoto & Bardwell2004). Reduced DsbC attacks an incorrect disulfide bond of a substrate protein,forming an intermolecular disulfide intermediate. The intermolecular disulfideis exchanged for the correct intramolecular disulfide in the substrate protein,releasing DsbC in a reduced state (reaction A). Alternatively, the intermoleculardisulfide between the substrate protein and DsbC is exchanged for anintramolecular disulfide within DsbC, releasing the substrate protein in areduced state and DsbC in an oxidized state (reaction B). The reduced proteincan then be oxidized by DsbA, and DsbC is regenerated by DsbD (not shown).

bic patches on misfolded proteins via its un-charged cleft. Upon binding proteins, DsbC’sreduced cysteines probe for disulfides in themisfolded protein. Cys98, the nucleophilic cys-teine in DsbC, attacks a substrate disulfide andforms a mixed disulfide with the substrate pro-tein (Figure 7). The mixed disulfide can be re-solved when the substrate protein’s reduced cys-teine attacks the mixed disulfide, creating a newdisulfide bond in the target protein and return-ing DsbC back to its reduced state (Figure 7,reaction A). Alternatively, the second active-sitecysteine from DsbC attacks the mixed disul-fide, forming an oxidized DsbC and a reducedprotein via an intramolecular thiol-disulfide ex-change reaction (Figure 7, reaction B). Thisprocess causes the target substrate protein tobecome reduced and allows DsbA to reoxi-dize that protein to the correct conformation(Figure 6).

DsbC must be kept reduced in the periplasmto stay active as an isomerase. The inner-membrane protein DsbD carries out the reduc-tion of DsbC (Figure 6) and another protein,DsbG, whose function is yet to be determined.

Disulfide Bond Formationin Eukaryotes

In eukaryotic cells, disulfide bonds are formedin the lumen of the ER, which is a specializedcompartment for protein folding and assembly.Two proteins are primarily responsible for con-trolling the process of oxidative folding: proteindisulfide isomerase (PDI) and Ero1p.

PDI, a 57-kDa soluble protein, was one ofthe first identified thiol-disulfide oxidoreduc-tases (Goldberger et al. 1963) and has been wellcharacterized. Depending on the redox envi-ronment and the nature of the substrates, PDIcan catalyze the formation, reduction, and iso-merization of disulfide bonds (Gilbert 1994).Tian et al. (2006) recently determined the crys-tal structure of yeast PDI. The protein hastwo thioredoxin domains (a and a′) as its ac-tive site, with the sequence Cys-Gly-His-Cys.The redox potentials of the a and a′ domains are−188 mV and −152 mV, respectively. Ero1p


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


Membrane

Lumen

CCyyttososolol

SHSH

Ero1p

FAD

S S

Ero1p

S S

Erv2p

SH SH

Erv2p

PDIPDI PDI

FAD FAD FAD FAD

Misfoldedsubstrate

FoldedsubstrateSubstrateMisfolded

substrate

SH

SH SHS S S– SH

O2

O2

SHSSS SSHS SSH

H+

a b

Figure 8Disulfide bond formation in the lumen of the endoplasmic reticulum (ER) (Kersteen & Raines 2003, Sevier& Kaiser 2002). There are two pathways for disulfide bond formation in the ER. In the first pathway (a),oxidizing equivalents are transferred to protein disulfide isomerase (PDI) from the membrane-associatedEro1p-FAD (flavin adenine dinucleotide) complex. PDI transfers the oxidizing equivalents to the reducedsubstrate. In the second pathway (b), oxidizing equivalents can be transferred by another membrane-associated protein, Erv2p. Misfolded substrates are isomerized by the thiolate form of the reduced PDI. Forsimplicity, only one of the two active sites of Ero1p, Erv2p, and PDI is shown.

is a 65-kDa, membrane-bound, flavin adeninedinucleotide (FAD)-containing oxidase thatacts on PDI (Frand & Kaiser 1998, Pollard et al.1998, Sevier et al. 2001, Tu et al. 2000).

There are two pathways for disulfide bondformation in Saccharomyces cerevisiae (Figure 8).In the first pathway, oxidizing equivalents aretransferred from Ero1p to PDI, which in turnoxidizes substrate proteins. The flow of oxidiz-ing equivalents in this pathway occurs througha series of thiol-disulfide exchange reactions(Frand & Kaiser 1998, Tu et al. 2000). The sec-ond pathway involves the protein Erv2p. Erv2ptransfers oxidizing equivalents from molecu-lar oxygen to PDI via the FAD cofactor toform disulfide bonds (Sevier & Kaiser 2002).Therefore, in eukaryotic disulfide bond forma-

tion electrons appear to flow from the pro-tein to PDI to Ero1p (or Erv2p) and ulti-mately to an electron acceptor. The secondpathway has been observed only in the absenceof Ero1p and in the presence of a plasmid over-producing Erv2p. The normal role of Erv2p isunknown.

CATALYSIS OF OXIDATIVEPROTEIN FOLDING BY PROTEINFOLDING FACTORS IN VITRO

To determine how disulfide bond formation oc-curs in the cell, extensive investigation has beenconducted on the in vitro refolding of modelproteins by protein folding factors. Knowledgefrom these studies can be used to determine


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


how oxidative folding occurs in vivo. Here, wediscuss the influence of PDI and DsbA/DsbCon the in vitro oxidative folding pathways, us-ing BPTI, RNase A, and hirudin as modelproteins.

Catalysis of Oxidative Folding by PDI

BPTI. Early studies on BPTI unfolding andrefolding (Figure 4b) indicated that PDI in-creased the rates of oxidation and the reduc-tion of BPTI in the presence of DTTox. Incontrast, PDI had only a small effect on thekinetics of folding when used alone or in thepresence of GSSG, cystamine, or hydroxyethyldisulfide (Creighton et al. 1980). Subsequently,working with reversibly trapped intermedi-ates and using a redox buffer (GSSG/GSH),Weissman & Kim (1993) demonstrated thatPDI dramatically increased both the yield andrate of formation of native BPTI. PDI showeda modest increase in the rate of formationof the two kinetically trapped intermediatesand N∗. However, PDI dramatically acceler-ated the formation of native protein (N) fromthe kinetic traps N′ and N∗ (accelerations of3500- and 6000-fold, respectively). The ef-fect of PDI on the reduced protein and theone-disulfide intermediates [30-51] and [5-55]was negligible, indicating that the effects ofPDI are specific to the kinetically trappedintermediates.

Recently, Satoh et al. (2005) reexamined thecatalysis of BPTI folding by PDI. They usedacid quenching to trap the folding interme-diates and determined the influence of PDIon each of the intermediates. PDI efficientlycatalyzed the folding reaction from the fullyreduced form to the native form (N). PDI read-ily converted N′ to N and partially to N∗, in-dicating that PDI catalyzed the reduction ofthe [14-38] disulfide bond and the formationof the [5-55] disulfide bond and then recreatedthe [14-38] disulfide bond. PDI also convertedN∗ to N, demonstrating that PDI catalyzed thereduction of the [14-38] disulfide bond and theformation of the [30-51] disulfide bond, fol-lowed by the recreation of the [14-38] disul-

fide bond. In both cases, the reduction of the[14-38] disulfide bond preceded the disulfidebond formation. The structure of intermedi-ate N′ is very similar to that of the native pro-tein, and the free thiols are presumably buriedin the conformational structure. Therefore, theintermediate N′ must be unfolded to form N(Weissman & Kim 1991). PDI accelerates thefolding by rearrangement of N′ to N∗, NSH

SH.Because it is important for the rearrangementof N′ to N to lose structure, PDI acceleratesthe folding of BPTI by promoting both un-folding and disulfide bond rearrangements instructured intermediates.

RNase A. The rate of regeneration of RNaseA catalyzed by PDI depends on the compo-sition of the GSSG/GSH redox buffer (Lyles& Gilbert 1991a,b). The rate of formation ofnative RNase A with DTTox/DTTred as a re-dox agent markedly increases in the presenceof PDI (9-fold at 15◦C, 6-fold at 25◦C, and62-fold at 37◦C). Although major changes wereobserved in the distribution of some disulfideintermediates with the rapid accumulation ofthe des species [65-72] and [40-95], PDI didnot alter the two major pathways of RNaseA regeneration (Figure 3a) (Shin & Scheraga1999). A subsequent study on the regenerationof RNase A at 25◦C confirmed that PDI accel-erates the formation of RNase A by catalyzingeach of the intermediate steps without chang-ing the folding mechanism (Shin & Scheraga2000).

During the regeneration of RNase A, fourpossible three-disulfide intermediates can form:des[65-72], des[40-95], des[26-84], and des[58-110] (Figure 3a). As discussed above, at 25◦C,the majority of the three-disulfide intermedi-ates are made up of des[65-72] and des[40-95],which regenerate native RNase A (Rothwarfet al. 1998a). However, at 15◦C, des[26-84] anddes[58-110] accumulate (Welker et al. 1999)as long-lived kinetic traps that slowly reshuffleback to the 3S ensemble (Welker et al. 2001).Shin et al. (2002) regenerated RNase A in thepresence of catalytically active and inactive PDIto determine the effects of PDI as an oxidase


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


and as a chaperone. Both forms of PDI in-creased the rate constant for the formation ofRNase A when compared with the rate con-stant for the case involving the redox reagentalone (∼17-fold and 2-fold for catalyticallyactive and inactive PDI, respectively). In thepresence of catalytically active PDI, the pop-ulations of des[40-95], des[26-84], and des[58-110] decreased as the reaction proceeded, mak-ing the des[65-72] species the prominent desspecies responsible for the regeneration of na-tive protein. Shin et al. (2002) also determinedthat the catalytically inactive PDI did not affectthe populations of des[65-72], des[26-84], anddes[58-110], indicating that noncatalytic bind-ing of the enzyme has no effect on the con-formation of these des species. However, theconcentration of only des[40-95] was signifi-cantly reduced, with a corresponding forma-tion of the native protein, indicating that thenoncatalytic binding of PDI induced a confor-mational change that led to the formation ofthe fourth disulfide in des[40-95]. Both disul-fide bonds [26-84] and [58-110] are buried inthe native protein; therefore, the 26 and 84 thiolgroups in des[58-110] and the 58 and 110 thiolgroups in des[26-84] must be exposed before di-rect oxidation takes place. The concentration ofthe species was similar in the presence and ab-sence of catalytically inactive PDI and the redoxagent DTTox, making direct oxidation of thesedes species unlikely. This finding suggests thatnoncatalytic binding of PDI to the substratesinduces the exposure of thiol groups in thedes[26-84] and des[58-110] species. Therefore,PDI converts the kinetically trapped des[26-84] and des[58-110] species into des[40-95] bydisulfide rearrangement through the 3S ensem-ble (Figure 3a). Thus, PDI plays a dual role asan oxidase/isomerase and a chaperone in the re-generation of RNase A.

Hirudin. The role of PDI in hirudin fold-ing is to promote the process of disulfiderearrangement in the scrambled species (con-solidation) (Chang 1994). PDI alone has no ef-fect on the consolidation process. However, inthe presence of free thiols (cysteines), PDI dis-

plays an additive effect in reshuffling the scram-bled species. By optimizing the mixture of PDIand the redox reagent (Cys/Cys-Cys), PDI cancomplete the folding of hirudin within 30 s.Therefore, the in vitro efficiency of hirudin re-folding is comparable to the scale of in vivo pro-tein folding. PDI thus can accelerate the speedof hirudin folding in vitro. In general, PDI canaccelerate the oxidation and isomerization ofdisulfide bonds but does not greatly alter fold-ing pathways.

Recently, Chang et al. (2006) studied ox-idative folding of hirudin in human serum.Consistent with earlier studies on hirudin(Chatrenet & Chang 1993), Chang et al. (2006)reported that the major rate-limiting step forthe regeneration of native protein is the disul-fide shuffling of the scrambled three-disulfideintermediates, which requires a reductant as theinitiator. Native hirudin was completely regen-erated in undiluted human serum (in a period of48 h) without any redox supplement, indicatingthat human serum may contain unidentified ox-idases like PDI that can catalyze disulfide bondformation.

Catalysis of Oxidative Foldingby DsbA and DsbC

BPTI. Zapun & Creighton (1994) examinedthe influence of DsbA on the folding of BPTIin the presence of GSSG/GSH as the redoxagent. DsbA differs significantly from PDI inits effect on the refolding of BPTI. Unlike PDI,DsbA induced a marginal increase in the rateof formation of native BPTI. The marginal in-crease was attributed to the direct oxidation ofBPTI by DsbA, as indicated by the disappear-ance of the reduced protein. However, DsbAwas unable to catalyze intramolecular disulfidebond rearrangements in the two-disulfide in-termediates [30-51, 14-38] and [5-55, 14-38],whereas PDI had a dramatic effect on thesetwo reactions. Thus, DsbA can rapidly oxidizeBPTI but is unable to eliminate the kinetic trapsin the folding pathway (Zapun & Creighton1994). Disulfide bond rearrangements usingDsbA were albeit observed when folding was


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


carried out at low pH, at which the rate of thiol-disulfide exchange using redox buffers with nor-mal pKa values is very slow. This activity ofDsbA is likely due to the very low pKa of itsactive-site cysteine and the consequential highreactivity at low pH. However, stoichiometricamounts of DsbA are required, and folding oc-curs on the hour timescale.

DsbC has no influence on the rate of ox-idation of reduced BPTI and therefore doesnot appear to catalyze disulfide bond formation.However, DsbC exhibits a marked effect on therate of appearance of native BPTI. DsbC cancatalyze rearrangements in [30-51, 14-38] and[5-55, 14-38] intermediates forming the nativeprotein. The efficiency of DsbC in rearrangingthe disulfide bonds of BPTI is greater than thatof DsbA but less than that of PDI (Zapun et al.1995).

RNase A. DsbA stimulates oxidation of re-duced RNase A in the presence of GSSG/GSHas the redox buffer but is very inefficient in pro-moting isomerization (Akiyama et al. 1992). Inlater studies, by using the reconstituted DsbA-DsbB system in vitro with DsbB as the redoxagent, Bader et al. (2000) showed that DsbA wasable to oxidize RNase A to the point at whichno free thiols were detectable; however, RNaseA gained negligible activity, indicating that thebulk of the protein had not been folded prop-erly, presumably owing to misoxidation of thi-ols by DsbA. But with the addition of reducedDsbC, the reactivation of RNase A increaseddramatically, suggesting that DsbC acts as anisomerase on a misoxidized 4S ensemble.

Hirudin. Catalytic amounts of DsbA accel-erate the overall folding of native hirudinand decrease its half time of formation bytwo- to threefold in a GSSG/GSH redoxbuffer (pH 8.7) without changing the rela-tive distribution of intermediates. At acidicpH, substoichiometric quantities of DsbAwere able to catalyze hirudin folding, indi-cating that DsbA is required for the forma-tion of disulfide bonds when bacteria are ex-

posed to acidic pH (Wunderlich et al. 1993,1995).

CONCLUSION AND FUTUREPERSPECTIVES

Disulfide bonds are critical posttranslationalmodifications of proteins. They not only sta-bilize protein structures but also are requiredfor the proper folding and biological activityof several proteins. Because the formation ofdisulfide bonds is tightly linked to the confor-mational folding of the protein, the problemof protein folding can be addressed by inves-tigating disulfide-linked protein folding. Sev-eral in vitro studies on small disulfide-rich pro-teins have helped elucidate the mechanism ofdisulfide-linked protein folding. However, pro-tein folding in the cell is not spontaneous andrequires the presence of protein folding cata-lysts. The discovery and characterization of theoxidative folding machinery in both prokary-otes and eukaryotes have opened up avenuesto exploit the system for folding disulfide-richproteins in the cell. Future research now canprobably be aimed at utilizing this knowledgeto engineer efficient systems for the expres-sion of pharmacologically important proteins.How proteins fold in vivo is one of the keyunsolved problems in basic biology. The vastmajority of detailed folding studies have notbeen done in vivo but rather in vitro, simplybecause most of the techniques that are usedto follow folding, such as circular dichroism,fluorescence, and hydrogen/deuterium (H/D)exchange, work well only in isolated systems.There are, however, big differences between thein vivo and in vitro environments. The presenceof folding factors is the most obvious difference,but factors such as molecular crowding may beequally important. Fortunately, thiol trappingis one technique by which folding can be mon-itored and that can be used almost as well invivo as in vitro. Future studies are expected toinvestigate in vivo folding pathways, perhaps byusing some of the same tools that have beenused so successfully to analyze disulfide-linkedfolding pathways in vitro.


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


SUMMARY POINTS

1. Disulfide bonds are formed in proteins by two thiol-disulfide exchange reactions with aredox reagent or another protein.

2. Disulfide bond formation is intimately linked to the conformational folding of a protein.The folded conformation stabilizes a disulfide bond to the same extent to which thedisulfide bond is stabilized by that particular conformation.

3. The technique of direct oxidative folding can be used to study the in vitro disulfide-linkedfolding pathway of proteins. Reduced proteins are folded in the presence of a suitableredox reagent, and the intermediates are trapped by a suitable quenching method. Theintermediates are then isolated and characterized by the use of biochemical and structuraltechniques.

4. Protein folding catalysts called oxidoreductases are required to form disulfide bonds invivo. Oxidoreductases accelerate oxidative folding by eliminating kinetic traps in proteinfolding.

DISCLOSURE STATEMENT

The authors are not aware of any biases that might be perceived as affecting the objectivity of thisreview.

LITERATURE CITED

Abkevich VI, Shakhnovich EI. 2000. What can disulfide bonds tell us about protein energetics, function andfolding: simulations and bioninformatics analysis. J. Mol. Biol. 300:975–85

Akiyama Y, Kamitani S, Kusukawa N, Ito K. 1992. In vitro catalysis of oxidative folding of disulfide-bondedproteins by the Escherichia coli dsbA (ppfA) gene product. J. Biol. Chem. 267:22440–45

Arolas JL, Aviles FX, Chang J-Y, Ventura S. 2006. Folding of small disulfide-rich proteins: clarifying thepuzzle. Trends Biochem. Sci. 31:292–301

Bader M, Muse W, Ballou DP, Gassner C, Bardwell JCA. 1999. Oxidative protein folding is driven by theelectron transport system. Cell 98:217–27

Bader MW, Xie T, Yu C-A, Bardwell JCA. 2000. Disulfide bonds are generated by quinone reduction. J. Biol.Chem. 275:26082–88

Bardwell JCA, McGovern K, Beckwith J. 1991. Identification of a protein required for disulfide bond formationin vivo. Cell 67:581–89

Beld J, Woycechowsky KJ, Hilvert D. 2007. Selenoglutathione: efficient oxidative protein folding by a dise-lenide. Biochemistry 46:5382–90

Cabrele C, Fiori S, Pegoraro S, Moroder L. 2002. Redox-active cyclic bis(cysteinyl)peptides as catalysts for invitro oxidative protein folding. Chem. Biol. 9:731–40

Chang J-Y. 1983. The functional domain of hirudin, a thrombin-specific inhibitor. FEBS Lett. 164:307–13Chang J-Y, Knecht R. 1991. Direct analysis of the disulfide content of proteins: methods for monitoring the

stability and refolding process of cystine-containing proteins. Anal. Biochem. 197:52–58Chang J-Y, Lu B-Y, Lai P-H. 2006. Oxidative folding of hirudin in human serum. Biochem. J. 394:249–57Chang JY. 1994. Controlling the speed of hirudin folding. Biochem. J. 300(Pt. 3):643–50Chang JY. 2004. Evidence for the underlying cause of diversity of the disulfide folding pathway. Biochemistry

43:4522–29Chatrenet B, Chang JY. 1992. The folding of hirudin adopts a mechanism of trial and error. J. Biol. Chem.

267:3038–43


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


Chatrenet B, Chang JY. 1993. The disulfide folding pathway of hirudin elucidated by stop/go folding experi-ments. J. Biol. Chem. 268:20988–96

Collet J-F, Riemer J, Bader MW, Bardwell JCA. 2002. Reconstitution of a disulfide isomerization system.J. Biol. Chem. 277:26886–92

Creighton TE. 1986. Disulfide bonds as probes of protein folding pathways. In Methods in Enzymology, ed.CHW Hirs, SN Timasheff, 131:83–106. San Diego: Academic

Creighton TE. 1988. On the relevance of nonrandom polypeptide conformations for protein folding. Biophys.Chem. 31:155–62

Creighton TE. 1990. Protein folding. Biochem. J. 270:1–16Creighton TE, Darby NJ, Kemmink J. 1996. The roles of partly folded intermediates in protein folding.

FASEB J. 10:110–18Creighton TE, Hillson DA, Freedman RB. 1980. Catalysis by protein-disulfide isomerase of the unfolding

and refolding of proteins with disulphide bonds. J. Mol. Biol. 142:43–62Darby NJ, Creighton TE. 1993. Dissecting the disulphide-coupled folding pathway of bovine pancreatic

trypsin inhibitor: forming the first disulphide bonds in analogues of the reduced protein. J. Mol. Biol.232:873–96

Elisha H. 2005. The study of protein folding and dynamics by determination of intramolecular distancedistributions and their fluctuations using ensemble and single-molecule FRET measurements. Chem.Phys. Chem. 6:858–70

Fink M, Nieves P, Chang S, Narayan M. 2008. Non-redox-active small-molecules can accelerate oxidativeprotein folding by novel mechanisms. Biophys. Chem. 132:104–9

Flory P. 1953. Principles of Polymer Chemistry. Ithaca: Cornell Univ. PressFrand AR, Kaiser CA. 1998. The ERO1 gene of yeast is required for oxidation of protein dithiols in the

endoplasmic reticulum. Mol. Cell 1:161–70Gilbert HF. 1994. The formation of disulfide bonds. In Mechanisms of Protein Folding, ed. RH Pain, pp. 104–35.

Oxford, UK: Oxford Univ. PressGoldberger RF, Epstein CJ, Anfinsen CB. 1963. Acceleration of reactivation of reduced bovine pancreatic

ribonuclease by a microsomal system from rat liver. J. Biol. Chem. 238:628–35Gomez G, Mansouraty G, Gardea J, Narayan M. 2007. Acceleration of oxidative protein folding by curcumin

through novel nonredox chemistry. Biochem. Biophys. Res. Commun. 364:561–66Gough JD, Barrett EJ, Silva Y, Lees WJ. 2006. ortho- and meta-substituted aromatic thiols are efficient redox

buffers that increase the folding rate of a disulfide-containing protein. J. Biotechnol. 125:39–47Grauschopf U, Winther JR, Korber P, Zander T, Dallinger P, Bardwell JCA. 1995. Why is DsbA such an

oxidizing disulfide catalyst? Cell 83:947–55Grutter MG, Priestle JP, Rahuel J, Grossenbacher H, Bode W, et al. 1990. Crystal structure of the thrombin-

hirudin complex: a novel mode of serine protease inhibition. EMBO J. 9:2361–65Guddat LW, Bardwell JCA, Glockshuber R, Huber-Wunderlich M, Zander T, Martin JL. 1997. Structural

analysis of three His32 mutants of DsbA: support for an electrostatic role of His32 in DsbA stability.Protein Sci. 6:1893–900

Guddat LW, Bardwell JCA, Martin JL. 1998. Crystal structures of reduced and oxidized DsbA: investigationof domain motion and thiolate stabilization. Structure 6:757–67

Hogg PJ. 2003. Disulfide bonds as switches for protein function. Trends Biochem. Sci. 28:210–14Inaba K, Murakami S, Suzuki M, Nakagawa A, Yamashita E, et al. 2006. Crystal structure of the DsbB-DsbA

complex reveals a mechanism of disulfide bond generation. Cell 127:789–801Iwaoka M, Juminaga D, Scheraga HA. 1998. Regeneration of three-disulfide mutants of bovine pancreatic

ribonuclease A missing the 65-72 disulfide bond: characterization of a minor folding pathway of ribonu-clease A and kinetic roles of Cys65 and Cys72. Biochemistry 37:4490–501

Kadokura H. 2006. Oxidative protein folding: many different ways to introduce disulfide bonds. Antioxid.Redox Signal. 8:731–33

Kersteen EA, Raines RT. 2003. Catalysis of protein folding by protein disulfide isomerase and small-moleculemimics. Antioxid. Redox Signal. 5:413–24


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


Kibria FM, Lees WJ. 2008. Balancing conformational and oxidative kinetic traps during the folding of bovinepancreatic trypsin inhibitor (BPTI) with glutathione and glutathione disulfide. J. Am. Chem. Soc. 130:796–97

Kobayashi T, Kishigami S, Sone M, Inokuchi H, Mogi T, Ito K. 1997. Respiratory chain is required to maintainoxidized states of the DsbA-DsbB disulfide bond formation system in aerobically growing Escherichia colicells. Proc. Natl. Acad. Sci. USA 94:11857–62

Konishi Y, Ooi T, Scheraga HA. 1982. Regeneration of ribonuclease A from the reduced protein. Rate-limitingsteps. Biochemistry 21:4734–40

Laity JH, Lester CC, Shimotakahara S, Zimmerman DE, Montelione GT, Scheraga HA. 1997. Structuralcharacterization of an analog of the major rate-determining disulfide folding intermediate of bovinepancreatic ribonuclease A. Biochemistry 36:12683–99

Lawrence KL, Hang-Cheol S, Mahesh N, William JW, Harold AS. 2000. Acceleration of oxidative folding ofbovine pancreatic ribonuclease A by anion-induced stabilization and formation of structured native-likeintermediates. FEBS Lett. 472:67–72

Low LK, Shin H-C, Scheraga HA. 2002. Oxidative folding of bovine pancreatic ribonuclease A: insight intothe overall catalysis of the refolding pathway by phosphate. J. Protein Chem. 21:19–27

Lyles MM, Gilbert HF. 1991a. Catalysis of the oxidative folding of ribonuclease A by protein disulfide iso-merase: dependence of the rate on the composition of the redox buffer. Biochemistry 30:613–19

Lyles MM, Gilbert HF. 1991b. Catalysis of the oxidative folding of ribonuclease A by protein disulfideisomerase: presteady-state kinetics and the utilization of the oxidizing equivalents of the isomerase.Biochemistry 30:619–25

Martin JL, Bardwell JCA, Kuriyan J. 1993. Crystal structure of the DsbA protein required for disulphide bondformation in vivo. Nature 365:464–68

McCarthy AA, Haebel PW, Torronen A, Rybin V, Baker EN, Metcalf P. 2000. Crystal structure of the proteindisulfide bond isomerase, DsbC, from Escherichia coli. Nat. Struct. Biol. 7:196–99

Nakamoto H, Bardwell JCA. 2004. Catalysis of disulfide bond formation and isomerization in the Escherichiacoli periplasm. Biochim. Biophys. Acta Mol. Cell Res. 1694:111–19

Narayan M, Welker E, Wedemeyer WJ, Scheraga HA. 2000. Oxidative folding of proteins. Acc. Chem. Res.33:805–12

Pace CN, Grimsley GR, Thomson JA, Barnett BJ. 1988. Conformational stability and activity of ribonucleaseT1 with zero, one, and two intact disulfide bonds. J. Biol. Chem. 263:11820–25

Pan JL, Bardwell JCA. 2006. The origami of thioredoxin-like folds. Protein Sci. 15:2217–27Pearson MA, Karplus PA, Dodge RW, Laity JH, Scheraga HA. 1998. Crystal structures of two mutants that

have implications for the folding of bovine pancreatic ribonuclease A. Protein Sci. 7:1255–58Pollard MG, Travers KJ, Weissman JS. 1998. Ero1p: a novel and ubiquitous protein with an essential role in

oxidative protein folding in the endoplasmic reticulum. Mol. Cell 1:171–82Rietsch A, Belin D, Martin N, Beckwith J. 1996. An in vivo pathway for disulfide bond isomerization in

Escherichia coli. Proc. Natl. Acad. Sci. USA 93:13048–53Roques BP, Rao R, Marion D. 1980. Use of nuclear Overhauser effect in the study of peptides and proteins.

Biochimie 62:753–73Rothwarf DM, Li YJ, Scheraga HA. 1998a. Regeneration of bovine pancreatic ribonuclease A: detailed kinetic

analysis of two independent folding pathways. Biochemistry 37:3767–76Rothwarf DM, Li YJ, Scheraga HA. 1998b. Regeneration of bovine pancreatic ribonuclease A: identification

of two nativelike three-disulfide intermediates involved in separate pathways. Biochemistry 37:3760–66Rothwarf DM, Scheraga HA. 1991. Regeneration and reduction of native bovine pancreatic ribonuclease A

with oxidized and reduced dithiothreitol. J. Am. Chem. Soc. 113:6293–94Rothwarf DM, Scheraga HA. 1993a. Regeneration of bovine pancreatic ribonuclease A. 2. Kinetics of regen-

eration. Biochemistry 32:2680–89Rothwarf DM, Scheraga HA. 1993b. Regeneration of bovine pancreatic ribonuclease A. 4. Temperature

dependence of the regeneration rate. Biochemistry 32:2698–703Ruoppolo M, Freedman RB, Pucci P, Marino G. 1996a. Glutathione-dependent pathways of refolding of

RNase T1 by oxidation and disulfide isomerization: catalysis by protein disulfide isomerase. Biochemistry35:13636–46


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


Ruoppolo M, Pucci P, Marino G. 2005. Folding and disulfide formation. In Protein Folding Handbook, ed. TKiefhaber, J Buchner, pp. 946–64. Weinheim, Ger.: Wiley

Ruoppolo M, Torella C, Kanda F, Panico M, Pucci P, et al. 1996b. Identification of disulphide bonds in therefolding of bovine pancreatic RNase A. Fold. Des. 1:381–90

Rydel TJ, Ravichandran KG, Tulinsky A, Bode W, Huber R, et al. 1990. The structure of a complex ofrecombinant hirudin and human alpha-thrombin. Science 249:277–80

Santoro J, Gonzalez C, Bruix M, Neira JL, Nieto JL, et al. 1993. High-resolution three-dimensional structureof ribonuclease A in solution by nuclear magnetic resonance spectroscopy. J. Mol. Biol. 229:722–34

Satoh M, Shimada A, Kashiwai A, Saga S, Hosokawa M. 2005. Differential cooperative enzymatic activitiesof protein disulfide isomerase family in protein folding. Cell Stress Chaperones 10:211–20

Scheraga HA, Konishi Y, Rothwarf DM, Mui PW. 1987. Toward an understanding of the folding of ribonu-clease A. Proc. Natl. Acad. Sci. USA 84:5740–44

Scheraga HA, Wedemeyer WJ. 2001. Bovine pancreatic ribonuclease A: oxidative and conformational foldingstudies. In Methods in Enzymology, ed. AW Nicholson, 341:189–221. San Diego: Academic

Schmidt B, Ho L, Hogg PJ. 2006. Allosteric disulfide bonds. Biochemistry 45:7429–33Sevier CS, Cuozzo JW, Vala A, Aslund F, Kaiser CA. 2001. A flavoprotein oxidase defines a new endoplasmic

reticulum pathway for biosynthetic disulphide bond formation. Nat. Cell Biol. 3:874–82Sevier CS, Kaiser CA. 2002. Formation and transfer of disulphide bonds in living cells. Nat. Rev. Mol. Cell

Biol. 3:836–47Sevier CS, Kaiser CA. 2006. Conservation and diversity of cellular disulfide bond formation pathways. Antioxid.

Redox Signal. 8:797–811Shin H-C, Scheraga HA. 1999. Effect of protein disulfide isomerase on the regeneration of bovine ribonuclease

A with dithiothreitol. FEBS Lett. 456:143–45Shin H-C, Scheraga HA. 2000. Catalysis of the oxidative folding of bovine pancreatic ribonuclease A by

protein disulfide isomerase. J. Mol. Biol. 300:995–1003Shin H-C, Song M-C, Scheraga HA. 2002. Effect of protein disulfide isomerase on the rate-determining steps

of the folding of bovine pancreatic ribonuclease A. FEBS Lett. 521:77–80Thannhauser TW, Rothwarf DM, Scheraga HA. 1997. Kinetic studies of the regeneration of recombinant

hirudin variant 1 with oxidized and reduced dithiothreitol. Biochemistry 36:2154–65Thornton JM. 1981. Disulphide bridges in globular proteins. J. Mol. Biol. 151:261–87Tian G, Xiang S, Noiva R, Lennarz WJ, Schindelin H. 2006. The crystal structure of yeast protein disulfide

isomerase suggests cooperativity between its active sites. Cell 124:61–73Tu BP, Ho-Schleyer SC, Travers KJ, Weissman JS. 2000. Biochemical basis of oxidative protein folding in the

endoplasmic reticulum. Science 290:1571–74van Mierlo CP, Kemmink J, Neuhaus D, Darby NJ, Creighton TE. 1994. 1H NMR analysis of the partly-

folded non-native two-disulphide intermediates (30-51,5-14) and (30-51,5-38) in the folding pathway ofbovine pancreatic trypsin inhibitor. J Mol Biol 235:1044–61

Volles MJ, Xu X, Scheraga HA. 1999. Distribution of disulfide bonds in the two-disulfide intermediates in theregeneration of bovine pancreatic ribonuclease A: further insights into the folding process. Biochemistry38:7284–93

Wedemeyer WJ, Welker E, Narayan M, Scheraga HA. 2000. Disulfide bonds and protein folding. Biochemistry39:4207–16

Wedemeyer WJ, Xu X, Welker E, Scheraga HA. 2002. Conformational propensities of protein folding inter-mediates: Distribution of species in the 1S, 2S, and 3S ensembles of the [C40A,C95A] mutant of bovinepancreatic ribonuclease A. Biochem. 41:1483–91

Weissman JS, Kim PS. 1991. Reexamination of the folding of BPTI: predominance of native intermediates.Science 253:1386–93

Weissman JS, Kim PS. 1992. The pro region of BPTI facilitates folding. Cell 71:841–51Weissman JS, Kim PS. 1993. Efficient catalysis of disulphide bond rearrangements by protein disulphide

isomerase. Nature 365:185–88Weissman JS, Kim PS. 1995. A kinetic explanation for the rearrangement pathway of BPTI folding. Nat.

Struct. Mol. Biol. 2:1123–30


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.


Welker E, Narayan M, Volles MJ, Scheraga HA. 1999. Two new structured intermediates in the oxidativefolding of RNase A. FEBS Lett. 460:477–79

Welker E, Narayan M, Wedemeyer WJ, Scheraga HA. 2001. Structural determinants of oxidative folding inproteins. Proc. Natl. Acad. Sci. USA 98:2312–16

Wlodaver LAS, Sjoin L, Gilliland GL. 1988. Structure of phosphate-free ribonuclease A refined at 1.26 A.Biochemistry 27:2705–17

Woycechowsky KJ, Wittrup KD, Raines RT. 1999. A small-molecule catalyst of protein folding in vitro andin vivo. Chem. Biol. 6:871–79

Wunderlich M, Otto A, Maskos K, Mu M, Seckler R, Glockshuber R. 1995. Efficient catalysis of disulfideformation during protein folding with a single active-site cysteine. J. Mol. Biol. 247:28–33

Wunderlich M, Otto A, Seckler R, Glockshuber R. 1993. Bacterial protein disulfide isomerase: efficientcatalysis of oxidative protein folding at acidic pH. Biochemistry 32:12251–56

Xie T, Yu L, Bader MW, Bardwell JCA, Yu C-A. 2002. Identification of the ubiquinone-binding domain inthe disulfide catalyst disulfide bond protein B. J. Biol. Chem. 277:1649–52

Xu X, Rothwarf DM, Scheraga HA. 1996. Nonrandom distribution of the one-disulfide intermediates in theregeneration of ribonuclease A. Biochemistry 35:6406–17

Xu X, Scheraga HA. 1998. Kinetic folding pathway of a three-disulfide mutant of bovine pancreatic ribonu-clease A missing the [40-95] disulfide bond. Biochemistry 37:7561–71

Zander T, Phadke ND, Bardwell JCA. 1998. Disulfide bond catalysts in Escherichia coli. In Methods in Enzy-mology, ed. GH Lorimer, TO Baldwin, 290:59–74. San Diego: Academic

Zapun A, Bardwell JCA, Creighton TE. 1993. The reactive and destabilizing disulfide bond of DsbA, a proteinrequired for protein disulfide bond formation in vivo. Biochemistry 32:5083–92

Zapun A, Creighton TE. 1994. Effects of DsbA on the disulfide folding of bovine pancreatic trypsin inhibitorand α-lactalbumin. Biochemistry 33:5202–11

Zapun A, Missiakas D, Raina S, Creighton TE. 1995. Structural and functional characterization of DsbC, aprotein involved in disulfide bond formation in Escherichia coli. Biochemistry 34:5075–89


Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.

AR356-FM ARI 8 September 2008 23:12

Annual Reviewof Cell andDevelopmentalBiology

Volume 24, 2008

Contents

Microtubule Dynamics in Cell Division: Exploring Living Cells withPolarized Light MicroscopyShinya Inoue � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Replicative Aging in Yeast: The Means to the EndK.A. Steinkraus, M. Kaeberlein, and B.K. Kennedy � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �29

Auxin Receptors and Plant Development: A New Signaling ParadigmKeithanne Mockaitis and Mark Estelle � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �55

Systems Approaches to Identifying Gene RegulatoryNetworks in PlantsTerri A. Long, Siobhan M. Brady, and Philip N. Benfey � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �81

Sister Chromatid Cohesion: A Simple Concept with a Complex RealityItay Onn, Jill M. Heidinger-Pauli, Vincent Guacci, Elçin Ünal,

and Douglas E. Koshland � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 105

The Epigenetics of rRNA Genes: From Molecularto Chromosome BiologyBrian McStay and Ingrid Grummt � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 131

The Evolution, Regulation, and Function of Placenta-Specific GenesSaara M. Rawn and James C. Cross � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 159

Communication Between the Synapse and the Nucleus in NeuronalDevelopment, Plasticity, and DiseaseSonia Cohen and Michael E. Greenberg � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 183

Disulfide-Linked Protein Folding PathwaysBharath S. Mamathambika and James C. Bardwell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 211

Molecular Mechanisms of Presynaptic DifferentiationYishi Jin and Craig C. Garner � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 237

Regulation of Spermatogonial Stem Cell Self-Renewal in MammalsJon M. Oatley and Ralph L. Brinster � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 263

Unconventional Mechanisms of Protein Transport to the Cell Surfaceof Eukaryotic CellsWalter Nickel and Matthias Seedorf � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 287

viii

Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.

AR356-FM ARI 8 September 2008 23:12

The Immunoglobulin-Like Cell Adhesion Molecule Nectin and ItsAssociated Protein AfadinYoshimi Takai, Wataru Ikeda, Hisakazu Ogita, and Yoshiyuki Rikitake � � � � � � � � � � � � � � � � � 309

Regulation of MHC Class I Assembly and Peptide BindingDavid R. Peaper and Peter Cresswell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 343

Structural and Functional Aspects of Lipid Binding by CD1 MoleculesJonathan D. Silk, Mariolina Salio, James Brown, E. Yvonne Jones,

and Vincenzo Cerundolo � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 369

Prelude to a DivisionNeedhi Bhalla and Abby F. Dernburg � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 397

Evolution of Coloration PatternsMeredith E. Protas and Nipam H. Patel � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 425

Polar Targeting and Endocytic Recycling in Auxin-DependentPlant DevelopmentJurgen Kleine-Vehn and Jirı Friml � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 447

Regulation of APC/C Activators in Mitosis and MeiosisJillian A. Pesin and Terry L. Orr-Weaver � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 475

Protein Kinases: Starting a Molecular Systems View of EndocytosisPrisca Liberali, Pauli Rämö, and Lucas Pelkmans � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 501

Comparative Aspects of Animal RegenerationJeremy P. Brockes and Anoop Kumar � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 525

Cell Polarity Signaling in ArabidopsisZhenbiao Yang � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 551

Hunter to Gatherer and Back: Immunological Synapses and Kinapsesas Variations on the Theme of Amoeboid LocomotionMichael L. Dustin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 577

Dscam-Mediated Cell Recognition Regulates NeuralCircuit FormationDaisuke Hattori, S. Sean Millard, Woj M. Wojtowicz, and S. Lawrence Zipursky � � � � � 597

Indexes

Cumulative Index of Contributing Authors, Volumes 20–24 � � � � � � � � � � � � � � � � � � � � � � � � � � � 621

Cumulative Index of Chapter Titles, Volumes 20–24 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 624

Errata

An online log of corrections to Annual Review of Cell and Developmental Biology articlesmay be found at http://cellbio.annualreviews.org/errata.shtml

Contents ix

Ann

u. R

ev. C

ell D

ev. B

iol.

2008

.24:

211-

235.

Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

gby

Uni

vers

ity o

f M

ichi

gan

- A

nn A

rbor

on

02/1

1/11

. For

per

sona

l use

onl

y.

disulfide-linked protein folding pathwayssites.lsa.umich.edu/webbkeane/wp-content/uploads/... ·...

Documents