The role of heat-shock
proteins W illiam
University
of California,
as molecular
chaperones
J. Welch San Francisco,
California,
USA
Recent
studies have revealed that protein folding and assembly events require the participation of accessory components, now being referred to as ‘molecular chaperones’. A number of chaperones have been identified as members of the heat-shock (or stress) protein family. This review discusses the roles of two classes of chaperones, the heat-shock protein 70 and groEL/ES families, in facilitating protein maturation, and describes how such events are perturbed in the cell subjected to metabolic stress.
in vivo
Current
Opinion
in Cell Biology
Introduction Classic studies by Anfmsen [ 11 and others demonstrated process that protein folding in vitro is a spontaneous dictated primarily by the linear sequence of amino acids present within the polypeptide. Recent studies, however, have revealed that such processes in vivo require the participation of accessory components that are now being referred to as molecular detergents or chaperones (reviewed in [2,3**]). A number of chaperones have now been identified and some shown to be equivalent to members of the so-called heat-shock or stress protein family. Most of the stress proteins are constitutively expressed in all cells, but also show increased expression in cells exposed to elevated growth temperatures or to amino acid analogs and various heavy metals (reviewed in [4**]). In general, the stress response appears to represent a cellular defense mechanism. For example, cells first subjected to a mild, sublethal stress event (e.g. mild heat shock), sufficient to upregulate the levels of stress proteins, are able to survive a subsequent, and what would otherwise be lethal, stress challenge. Because so many of the agents/treatments that induce the stress response fall into the category of ‘protein chaotropes’, it has been suggested that the intracellular accumulation of abnormal proteins represents the trigger by which the response is initiated. Support for this idea comes from the observation that simply injecting a collection of denatured proteins into living cells is sufficient to activate a stress response [5]. In what follows, I will discuss how some of the stress proteins serve as molecular chaperones in facilitating protein maturation events. In addition, I consider why these particular chaperones exhibit increased expression in cells experiencing metabolic stress.
BiPdinding
The heat-shock protein 70 family: related ATP-binding proteins present throughout the cell Owing to its extremely high induction in cells subjected to metabolic stress, considerable attention has been directed towards the characterization of the so-called heatshock protein (hsp)70. Via both biochemical and genetic approaches we now know that there are related forms of hsp70 distributed throughout various cellular compartments. In addition to hsp70, which is expressed at high levels in cells experiencing stress, related forms of the protein are constitutively expressed in all cells grown under normal conditions. These include hsp73, which is highly related to hsp70 and is present within the nucleus and cytoplasm, the glucose-regulated protein (grp)78 or binding protein (BiP), present within the lumen of the endoplasmic reticulum, and grp75 present within the matrix of either mitochondria or chloroplasts. All of these proteins exhibit considerable sequence homology and bind avidly to ATP. All of the available evidence indicates that the related proteins of this family interact transiently with other proteins undergoing various aspects of maturation (reviewed in [4**] >. Using metabolic pulse-chase radiolabeling techniques, members of the hsp70 family have been observed to interact with nascent chains of proteins undergoing synthesis on the ribosome [6-a]. In a similar scenario, grp78/BiP and grp75, present within either the endoplasmic reticulum or mitochondria, respectively, appear to bind to proteins that are being translocated from the cytoplasm into their respective organeUe (Fig. 1) [ 70-9’1. It is thought that the related hsp70 proteins bind to ma-
Abbreviations protein; g-glucose-regulated
@ Current
Biology
1991, 3:1033-1038
protein; hsp-heat-shock
Ltd ISSN 0!355*74
protein.
10:
W
Post-transcriptional
processes
Cytoplasm lnterrnembrane
space Y
Mitochondrial matrix
ADP+Pi
0
Cc) + ADP+Pi
Fig. 1. A model describing the possible role of grp75 and hsp58 in mitochondrial protein import, folding and assembly. Newly synthesized proteins destined for the mitochondria are maintained in an unfolded or translocation-competent state within the cytoplasm by virtue of their interaction with the cytosolic hsp72/73. Translocation of the polypeptide into the mitochondria is accompanied by the ATPdependent release of hsp72/73. As the translocating and unfolded polypeptide enters into the mitochondria, it becomes complexed with grp75, and the mitochondrial signal sequence is removed by signal peptidase. Once entirely inside the mitochondria, folding of the polypeptide commences, accompanied by the ATP-dependent release of grp75 fa,i). For some monomeric mitochondrial proteins, it remains possible that folding is also dependent on an interaction with hsp58 fa,ii). For the assembly of oligomeric proteins, grp75 is released from the monomer (a,iii) and the monomer then moves to hsp58 fb,ii) and is assembled into its oligomeric form (CL Alternatively, the monomer, still bound to grp75, moves to hsp58 fb,ii) and is assembled into its oligomeric form (CL
turing polypeptides and, by doing so, prevent the premature folding of the target polypeptide until its synthesis or translocation event has been completed. Once synthesis or translocation is complete, the target polypeptide is releasedfrom its hsp70 chaperone and now commences along its final folding pathway. Releaseof hsp70 from its target appearsto require ATP and probably ATP hydrolysis.At least in some cases,the hydrolysis of ATP is accompanied by the autophosphotyiation of the hsp70 chaperone and its release from the target substrate [9*]. How the hsp70 proteins recognize and bind to unfolded proteins, and subsequently know when to release their target, remains unclear. In vifro studies have demonstrated that hsp70 will indeed bind stably to an unfolded protein, but not to its properly folded counterpart [lo*].
The related hsp70 proteins are composed of two functional domains. The highly conserved amino terminus contains an ATP-binding site with a three-dimensional structure similar to that of both the ATP-binding site of actin and the nucleotide-binding cleft of various sugar kinases (e.g. hexokinase) [ 11.1. Via computer modeling, the carboxy termini of the related hsp70 proteins have been proposed to be similar in structure to the peptidebinding cleft of the class I histocompatibility proteins [ 12.1. This proposal is especiallyinteresting considering that some of the genesencoding hsp70 are located within the major histocompatibility complex [ 13). Additionally, both hsp70 and the class I histocompatibility proteins bind somewhat promiscuously to a variety of short peptides in vitro [ 141.
The role of heat-shock
The groEUES proteins: a workbench folding and assembly
for protein
proteins as molecular
chaperones
Welch
ranged as two rings, often referred to as a ‘double donut’. Similarly, the smaller groES subunit, at least in bacteria, also exists as a ring of sevenidentical subunits. In E. coli, groEL/ESwere shown to facilitate the orderly assembly of the bacteriophagehead complex in infected bacteria. Subsequentstudies also implicated a role for groEI,&S in a number of other normal bacterial processes,including DNA replication, cell division and protein secretion (reviewed by Georgopoulos in [ 3**] ).
In contrast to membersof the hsp70 family, which appear to bind to and stabilize unfolded proteins, the groEI/ES family seem to actually promote or catalyzeprotein folding and/or assemblyevents.This family of proteins, first described in Escberi~ziz coli, consists of a larger, 60kD subunit (groEL) and a smaller but related 10.kD polypeptide (groES) (reviewed in [3**] ). In eukaryotic cells, homologs of both groEL and groES have been described and at least the larger-molecular-weightsubunit is upregulatedin cells subjected to stress.Thus, this protein is often referred to as hsp58 or hsp60. So far, hsp60 has only been describedwithin either chloroplasts or m itochondria. Most investigatorssuspect that related forms of the groEI/ES proteins will be discovered within other intracellular compartments. The hsp60- or groEL-related proteins exist as large oligomeric complexes, consisting of 14 subunits ar-
A groEL-related protein has been isolated from plant chloroplasts and shown to be equivalent to the so-called Rubisco-binding protein [ 151. Rubisco is the plant enzyme involved in the first step of CO2 iixation. It is a large oligomer, consisting of eight large and eight smaller subunits. The Rubisco-binding protein binds to the newly synthesizedRubisco large and small subunits and facilitates their orderly transition into the final, higher ordered Rubisco complex. Further support for this proposed pathway comes from recent studies showing that groEI/ES, purikd from E. cob, can substitute for the re-
rable 1. Stress proteins acting as molecular chaperones. Class
Location
Properties
Homo-oligomer of seven subunits
Bacteria, plastids and mitochondria
Arranged as seven identical subunits in a ring; approximately 50 % homologous to groEL; binds to groEL in the presence of MgATP. Postulated function: regulate the interaction of groEL with unfolded protein targets and/or regulate the ATPase activity of groEL. Likely to be related forms in other intracellular compartments.
- 60-70
Homo- or heterooligomer of 14 subunits
Bacteria, plastids and mitochondria
Distinctive ‘double-donut’ structure of 14 subunits arranged as two rings of seven subunits each; bind and hydrolyze ATP. Postulated function: in conjunction with groES and ATP, provides a surface for monomeric protein folding and/or higher ordered protein assembly. Likely to be related forms within other intracellular compartments.
hsp73
- 70
hsp72
- 70
w75
- 70
Monomer and/or dimer Monomer and/or dimer Monomer and/or dimer
grp7B (BiP)
- 70
Cytosol and nucleus Cytosol and nucleus Matrix of mitochondria and plastids Lumen of endoplasmic reticulum
All members are highly related and bind to ATP; most are constitutively synthesized, although hsp72 is expressed primarily only in the cell after stress. Postulated function: recognize and bind to unfolded proteins such as nascent polypeptides undergoing synthesis and/or translocation into organelles. After metabolic stress, likely to be involved in denatured proteins.
Size Apparent
(kDj
Native
1~~60 (groEL/ESj groES
hsp60 or groEL
-
10
1sp70
3iP, binding protein; grp, glucose-regulated
Monomer and/or dimer
protein;
hsp, heat-shock
protein.
1035
36
Post-transcriptional
processes
.
lated plant Rub&co-binding protein in facilitating the assembly of the Rub&o complex in vim 1161. In yeast, groEL-relatedproteins (hsp60) have been described and again appear to be integral in facilitating higher ordered protein assembly events. For example, via both genetic and biochemical studies, yeast hsp60 has been shown to interact with a variety of newly synthesized m itochondrial proteins, catalyzing their assembly into their linal oligomeric and enzymatically active forms [17). Such processesappear to be dependent on ATP and probably require the assistanceof additional components, one possibility being a eukaryotic homolog of groES [18*]. On the basis of all these studies, it has been suggestedthat groEL provides a surface, or workbench which binds to unfolded proteins and, through a series of ATP-hydrolysisevents, perhaps regulated in part by groES, results in the proper folding of the target polypeptide [3**,15,19**]. What remains unclear is how unfolded targets are recognized by groEL, how the energy of ATP hydrolysis is coupled to protein folding and, iinally, the exact role of groEI/ES in facilitating the assembly of monomeric proteins into their final higher ordered, oligomeric structure.
The concept
of molecular
intimately with the carboxy-terminal domain? Moreover, what prevents the amino-terminal domain from inadvertently folding with components of the translational machinery? Similarly, during protein translocation, as proteins enter into the lumen of an organelle in an unfolded state (as has been shown), much of the polypeptide chain is still present within the cytosol. Hence, again you have a situation in which all of the information required for folding is not yet present within the organelle lumen. What prevents the partially translocatedpolypeptide from folding prematurely? It has been suggested that members of the hsp70 family function to bind to and prevent the premature folding of proteins that exist in an intermediate, unfolded state. Members of the groEI/ES family may then take over this role and catalyze the proper folding and assemblyof the target protein. Thus, for any given polypeptide, acquisition of its final or mature structure may be dependent upon one or more chaperones, perhaps working in tandem. In addition, one m ight suspect that the chaperones play an integral role in facilitating dynamic changes in mature proteins whose as sembly/disassemblyis regulated in response to different biological cues. Whether all proteins require the assistance of molecular chaperones for their maturation still remains to be determined.
chaperones
Owing to their role in facilitating protein folding and assembly events, members of the hsp70 and hsp60 families are now often referred to as ‘molecular chaperones’(reviewed in [3-l). In general terms, a chaperone is defined as ‘one delegated to insure proper behavior’. In molecular terms, molecular chaperones are broadly dehned as a class of unrelated proteins that mediate the correct assemblyof other polypeptides. In no known instancedo molecular chaperones themselvesactively provide the information for folding or assembly; nor do they become part of the linal folded structure itself. Rather, they seem to function by preventing nonproductive folding and/or higher ordered assemblypathways, and thereby ensure that productive folding pathways occur quickly and with high fidelity. Finally,the concept in no way contradicts the principle of self-assembly. Specifically,proteins do assume their fmal conformation as a consequenceof their primary amino acid sequence and relevantpost-translationalmodifications (phosphory lation, glycosylation,etc.), but may require the participation of molecular chaperones to ensure that folding and assemblyoccur correctly and quickly inside the cell. A particularly illustrative example of the requirement of molecular chaperonesis provided by examining the steps of protein synthesisand/or protein transkxation. For example, during the course of protein synthesis, as the nascent chain emerges from the ribosome, a situation arises in which all of the information necessaryfor folding of the polypeptide chain is not yet present. Whereas individual protein domains can fold independently of one another, how does the cell cope with an aminoterminal domain whose fate is to fold with or interact
Why are some molecular after stress?
chaperones
induced
Having discussed the concept of molecular chaperones, the question of why the expression of some chaperones is increased in cells subjected to metabolic stress, should be addressed A clue to the answer is the fact that many of the agents/treatmentsthat induce the stress response interfere with protein folding/assembly. For example, whereas molecular chaperones appear to interact transiently with their protein targets in cells maintained under normal growth conditions, we have found that such interactions no longer appear transient, but are instead stable in cells experiencing stress (6**]. Specifically, in cells exposed to amino acid analogs, a potent inducer of the stress response, newly synthesized proteins are again found to interact with hsp70. Unlike the situation in normal cells, however, the analog-containing proteins are not immediately releasedfrom their chaperone once their synthesisis complete. Presumably,when the newly synthesizedprotein incorporates an amino acid analog, it is unable to commence along its proper folded pathway and, as a consequence,continues to appear unfolded. With time, and as more proteins are made but are unable to properly fold, they remain bound to hsp70, and as a consequence the available levels of hsp70 are reduced. The cell then responds by upregulating the expression of hsp70, in effect, by the induction of a stress response. In a similar but neverthelessdistinct scenario, when cells are subjected to heat-shocktreatment, a number of mature cellular proteins begin to denature and consequently now appear as unfolded targets for hsp70. Again, the net result is a reduction in the availablelevels
The role of heat-shock
of hsp70 and a corresponding induction of the stress response. A similar scenario for groEI./hspbOmay also be operative 191. What remains to be determined is the fate of such proteins unable to commence folding, or mature proteins which become unfolded. Whether these denatured proteins can be rescued,that is refolded [ 200,21*], or whether they are simply maintained in a soluble form via binding to molecular chaperones such as hsp70, and subsequentlydegraded [22], remains to be established in vivo. Depending upon the extent of the protein denaturation, both of these scenarios will probably prove operative in the cell experiencing stress.
New insights regarding the details of protein maturation have emerged over the past few years and have raised new and exciting questions concerning the molecular mechanismsgoverning protein folding and assembly events.It is now well establishedthat such processesare facilitated by a class of proteins referred to as molecular chaperones,many of which represent members of the socalled heat-shockor stress protein family. In view of the themes outlined above, and with the appropriate chaperones now availablein purified form, the exact molecular details by which proteins assumetheir final structures will surely be forthcoming.
and recommended
mitochondria. 8.
.
special interest, published highlighted as: interest outstanding interest
within
the annual period
of review,
Demonstrates a general requirement for functional BiP in the import of proteins from the cytosol into the endoplasmic reticulum. 9. .
2.
ROMMAN JE: Polypeptide Chain Binding Roteins: Catalysts of Protein Folding and Related Processes in Cells. Cell 1989,
10.
.
the Folding of Protein
EUS RJ, VANDER VLES SM: Molecular Chaperones. Annu RetI .. Biocbem 1991, 60:321-347. Excellent up-todate review of the concept and description of molecular chaperones. 4. ..
MORIMOTO RI, TIZ+IERES A, GEORCOPOLIUIS C (EDS): Stress Proteins in Biology and Medicine [book]. New York: Cold Spring Harbor Laboratory Press, 1990. Excellent collection of review articles covering different aspects of the stress response. m J, G~~DEERG AL, VOEUMY R: Abnormal Roteins Serve as Eukaryotic Stress Signals and Trigger the Activation of Heat Shock Genes. Science 1986, 232:252-254. BECKMANN RP, MIZZEN L4. WELCH WJ: Interactions of hsp 70 with Newly Synthesized Proteins: Implications for Rotein Folding and Assembly. Science 1990, 248:850-854.
Biochemical evidence implicating a role for hsp70 in the co-translational binding of proteins under synthesis on the ribosome. Discusses the possible mechanism by which the stress response is regulated.
.
Mi-
PAILEROS DR, WELCH WJ, FV‘IK AL: Interaction of hsp 70 with Unfolded Roteins: Effects of Temperature and Nucleotides on the Kinetics of Binding. Pnx: Natl Acad Sci I/ S A 1991, 88:571+5723.
the role of nucleotides 11.
.
in mediating
such events.
HOLMESKC: Siiry of the Three-Dimensional Structures of Actin and the ATPase Fragment of a 70-kDa Heat Shock Cognate Rotein. Prcrc Nat1 Acad Sci US A 1991, 88:5041-5045.
FLAHER-IY KM, McKr;v DB, KAEXH W,
Discusses the three-dimensional structure of the hsp70 ATPase domain and its similarities to other ATP-binding proteins such as actin and hexokinase. 12.
A Hypo thetical Model for the Peptide Biding Domain of hsp 70 Based on the Peptide Binding Domain of HLA EML30 / 1991, 10:10531059.
R~PPMAN F, TAYLOR WR, ROTHBARD JB, GREEN NM:
paper proposing that the peptide-binding similar to that of class I histocompatibility
site of hsp70 is proteins.
DUNHAM I, TROYV~DAII J, CAMPBEU RD: Human Major Histocompacibility Complex Contains Genes for the Major Heat Shock Rotein hsp 70. Prcc Natf Ad Sci
13.
SARGENT CA,
14.
G, CHAPPEU T, ROTHMAN JE: Peptide Binding and Release by Roteins Implicated as Catalysts of Rotein Assembly. Science 1989, 245:38>350.
15.
HEMMINGSEN SM, WOOLFORO C, VAN DER VLES SM, ‘I?uY K, DENNIS DT, GEORCXPOUU~ CP, HENDRD( RW, Ews RJ: Homo-
KANc PJ, OSIXRMANN J, SHILLING J, NEUPERTW, Crwc F& PFANNER N: Requirement for hsp 70 in the Mitochondrial
FLY
logous Plant and Bacterial Proteins Chaperone Oligomeric Protein Assembly. Nature 1988, 333:330-334.
3.
7.
Mammalian
tochondrial Stress Roteins, grp 75 and hsp 58, Transiently Interact with Newly Synthesized Mitochondrial Roteins. Cell Regulation 1991, 2165-179.
Demonstrates the affinity of hsp70 for unfolded proteins and discusses
59:591-l.
..
MIZZEN L4, KABILING AN, WELCH WJ: The Two
U S A 1982, 86:1968-1972.
ANFINSEN CB: Principles that Govern Chains. Science 1973, 8:223-230.
6.
MI% LM, ROSE MD: Loss of Bii/Grp 78 Function Blocks Translocation of Secretory Roteins in Yeast. / Cell Biol 1990, 110:188%1895.
VOGEL JP,
Theoretical remarkably
reading
1.
5.
Welch
Matrix for Translocation and Folding of Recursor Roteins. Nature 1990, 348:137-143. Genetic evidence that the mitochondrial form of hsp70 is required for the import of most, if not all, proteins that are translocated into the
.
Papers of have been . of .. of
chaperones
Biochemical study examining the interactions of the mitochondrial stress proteins with newly synthesized mitochondrial proteins. Demon. strates an interaction between the two mitochondrial stress proteins and proposes that autophosphotyiation of grp75 is coupled with the release of its target.
Conclusions
References
proteins as molecular
16.
GOLOUBINOFF P, CHIU~TEUER JT, GATENBY AA, LLXI~IER GH: Reconstitution of Active Dimeric Ribulose Biiphosphate Carboxylase &om an Unfolded State Depends on Two Chap eronin Proteins and Mg-ATP. Nahtre 1989, 342:884-889.
17.
CHENG MY, m FU, M,wm J, POUAXK RA, KAIOUSEK F, NEUPERT W, HALLBERG EM, BERG RL, HORWICH AL Mito-
chondrial Heat-Shock Rotein hsp 60 is Essential for Assembly of Roteins Imported into Yeast Mitochondria. Nature 1989, 337:62&25. 18. .
LABEN TH, GATENBY A4 Do NNDSON GK, tiRlhIER GH, VIJTANEN PV: Identification of a groES-Like Chaperonin in Mitochon-
dria that Facilitates Rotein Folding. Pmc Nat1 Ad U S A 1990. 87:76837687. One of the first identifications of a eukaryotic cusses the biochemical function of groES.
19. ..
homolog
of groES. Dis-
M A R T W J, IANGER T. BOTEIVA R, !XHRAMEL 4 HORWICH AL, m FU: Chaperonin-Mediated Protein Folding at the Surface
groEL Through a ‘Molten Globule’-Lie ture 1991, 352:3&42.
Sci
of Intermediate. Na-
10X
lb38
Post-transcriptional
processes
One of the first in vitm biochemical studies demonstrating a role for groEIjES in facilitating monomeric protein folding. Discussesintermediates during the folding process, and the amount of ATP required for the folding of each monomer. 20. .
G~rr~~~rusGA, PAPAVASSILIOLI AC, RUBOCKP, SILVERSTEIN SJ. G~TTESMAN ME: Renaturation of Denatured Lambda Repressor Requites Heat Shock Proteins. Cell 1990, 61:10131020. A report on the ability of various bacterial heat-shock proteins to facilitate renaturation of a denatured protein. 21. .
SKOWRAD, GEORGOPOUUX C, Zvucz M: The E. coIi dnaK Gene Product, the hsp 70 Homolog. can Reactive Heat-
Inactivated RNA Polymerase in an ATP Hydrolysis-Dependent Manner. cel( 1990, 62:939-944. Reports that dnaK (hsp70) protects proteins from heat denaturation as well as facilitating renarurarion of a denatured protein. 22. CH~ANGHL TERLECKY SK, PLANTCP. DICEJP: A Role for 70 KiIodaIton Heat Shock Protein in Lysosomal Degradation of IntraceIIuIar Proteins. Science 1989, 246:381-385.
WJ Welch, Department of Medicine and Physiology, University of Califomia, San Francisco, California 94143.Ot354,USA
proteins W illiam
University
of California,
as molecular
chaperones
J. Welch San Francisco,
California,
USA
Recent
studies have revealed that protein folding and assembly events require the participation of accessory components, now being referred to as ‘molecular chaperones’. A number of chaperones have been identified as members of the heat-shock (or stress) protein family. This review discusses the roles of two classes of chaperones, the heat-shock protein 70 and groEL/ES families, in facilitating protein maturation, and describes how such events are perturbed in the cell subjected to metabolic stress.
in vivo
Current
Opinion
in Cell Biology
Introduction Classic studies by Anfmsen [ 11 and others demonstrated process that protein folding in vitro is a spontaneous dictated primarily by the linear sequence of amino acids present within the polypeptide. Recent studies, however, have revealed that such processes in vivo require the participation of accessory components that are now being referred to as molecular detergents or chaperones (reviewed in [2,3**]). A number of chaperones have now been identified and some shown to be equivalent to members of the so-called heat-shock or stress protein family. Most of the stress proteins are constitutively expressed in all cells, but also show increased expression in cells exposed to elevated growth temperatures or to amino acid analogs and various heavy metals (reviewed in [4**]). In general, the stress response appears to represent a cellular defense mechanism. For example, cells first subjected to a mild, sublethal stress event (e.g. mild heat shock), sufficient to upregulate the levels of stress proteins, are able to survive a subsequent, and what would otherwise be lethal, stress challenge. Because so many of the agents/treatments that induce the stress response fall into the category of ‘protein chaotropes’, it has been suggested that the intracellular accumulation of abnormal proteins represents the trigger by which the response is initiated. Support for this idea comes from the observation that simply injecting a collection of denatured proteins into living cells is sufficient to activate a stress response [5]. In what follows, I will discuss how some of the stress proteins serve as molecular chaperones in facilitating protein maturation events. In addition, I consider why these particular chaperones exhibit increased expression in cells experiencing metabolic stress.
BiPdinding
The heat-shock protein 70 family: related ATP-binding proteins present throughout the cell Owing to its extremely high induction in cells subjected to metabolic stress, considerable attention has been directed towards the characterization of the so-called heatshock protein (hsp)70. Via both biochemical and genetic approaches we now know that there are related forms of hsp70 distributed throughout various cellular compartments. In addition to hsp70, which is expressed at high levels in cells experiencing stress, related forms of the protein are constitutively expressed in all cells grown under normal conditions. These include hsp73, which is highly related to hsp70 and is present within the nucleus and cytoplasm, the glucose-regulated protein (grp)78 or binding protein (BiP), present within the lumen of the endoplasmic reticulum, and grp75 present within the matrix of either mitochondria or chloroplasts. All of these proteins exhibit considerable sequence homology and bind avidly to ATP. All of the available evidence indicates that the related proteins of this family interact transiently with other proteins undergoing various aspects of maturation (reviewed in [4**] >. Using metabolic pulse-chase radiolabeling techniques, members of the hsp70 family have been observed to interact with nascent chains of proteins undergoing synthesis on the ribosome [6-a]. In a similar scenario, grp78/BiP and grp75, present within either the endoplasmic reticulum or mitochondria, respectively, appear to bind to proteins that are being translocated from the cytoplasm into their respective organeUe (Fig. 1) [ 70-9’1. It is thought that the related hsp70 proteins bind to ma-
Abbreviations protein; g-glucose-regulated
@ Current
Biology
1991, 3:1033-1038
protein; hsp-heat-shock
Ltd ISSN 0!355*74
protein.
10:
W
Post-transcriptional
processes
Cytoplasm lnterrnembrane
space Y
Mitochondrial matrix
ADP+Pi
0
Cc) + ADP+Pi
Fig. 1. A model describing the possible role of grp75 and hsp58 in mitochondrial protein import, folding and assembly. Newly synthesized proteins destined for the mitochondria are maintained in an unfolded or translocation-competent state within the cytoplasm by virtue of their interaction with the cytosolic hsp72/73. Translocation of the polypeptide into the mitochondria is accompanied by the ATPdependent release of hsp72/73. As the translocating and unfolded polypeptide enters into the mitochondria, it becomes complexed with grp75, and the mitochondrial signal sequence is removed by signal peptidase. Once entirely inside the mitochondria, folding of the polypeptide commences, accompanied by the ATP-dependent release of grp75 fa,i). For some monomeric mitochondrial proteins, it remains possible that folding is also dependent on an interaction with hsp58 fa,ii). For the assembly of oligomeric proteins, grp75 is released from the monomer (a,iii) and the monomer then moves to hsp58 fb,ii) and is assembled into its oligomeric form (CL Alternatively, the monomer, still bound to grp75, moves to hsp58 fb,ii) and is assembled into its oligomeric form (CL
turing polypeptides and, by doing so, prevent the premature folding of the target polypeptide until its synthesis or translocation event has been completed. Once synthesis or translocation is complete, the target polypeptide is releasedfrom its hsp70 chaperone and now commences along its final folding pathway. Releaseof hsp70 from its target appearsto require ATP and probably ATP hydrolysis.At least in some cases,the hydrolysis of ATP is accompanied by the autophosphotyiation of the hsp70 chaperone and its release from the target substrate [9*]. How the hsp70 proteins recognize and bind to unfolded proteins, and subsequently know when to release their target, remains unclear. In vifro studies have demonstrated that hsp70 will indeed bind stably to an unfolded protein, but not to its properly folded counterpart [lo*].
The related hsp70 proteins are composed of two functional domains. The highly conserved amino terminus contains an ATP-binding site with a three-dimensional structure similar to that of both the ATP-binding site of actin and the nucleotide-binding cleft of various sugar kinases (e.g. hexokinase) [ 11.1. Via computer modeling, the carboxy termini of the related hsp70 proteins have been proposed to be similar in structure to the peptidebinding cleft of the class I histocompatibility proteins [ 12.1. This proposal is especiallyinteresting considering that some of the genesencoding hsp70 are located within the major histocompatibility complex [ 13). Additionally, both hsp70 and the class I histocompatibility proteins bind somewhat promiscuously to a variety of short peptides in vitro [ 141.
The role of heat-shock
The groEUES proteins: a workbench folding and assembly
for protein
proteins as molecular
chaperones
Welch
ranged as two rings, often referred to as a ‘double donut’. Similarly, the smaller groES subunit, at least in bacteria, also exists as a ring of sevenidentical subunits. In E. coli, groEL/ESwere shown to facilitate the orderly assembly of the bacteriophagehead complex in infected bacteria. Subsequentstudies also implicated a role for groEI,&S in a number of other normal bacterial processes,including DNA replication, cell division and protein secretion (reviewed by Georgopoulos in [ 3**] ).
In contrast to membersof the hsp70 family, which appear to bind to and stabilize unfolded proteins, the groEI/ES family seem to actually promote or catalyzeprotein folding and/or assemblyevents.This family of proteins, first described in Escberi~ziz coli, consists of a larger, 60kD subunit (groEL) and a smaller but related 10.kD polypeptide (groES) (reviewed in [3**] ). In eukaryotic cells, homologs of both groEL and groES have been described and at least the larger-molecular-weightsubunit is upregulatedin cells subjected to stress.Thus, this protein is often referred to as hsp58 or hsp60. So far, hsp60 has only been describedwithin either chloroplasts or m itochondria. Most investigatorssuspect that related forms of the groEI/ES proteins will be discovered within other intracellular compartments. The hsp60- or groEL-related proteins exist as large oligomeric complexes, consisting of 14 subunits ar-
A groEL-related protein has been isolated from plant chloroplasts and shown to be equivalent to the so-called Rubisco-binding protein [ 151. Rubisco is the plant enzyme involved in the first step of CO2 iixation. It is a large oligomer, consisting of eight large and eight smaller subunits. The Rubisco-binding protein binds to the newly synthesizedRubisco large and small subunits and facilitates their orderly transition into the final, higher ordered Rubisco complex. Further support for this proposed pathway comes from recent studies showing that groEI/ES, purikd from E. cob, can substitute for the re-
rable 1. Stress proteins acting as molecular chaperones. Class
Location
Properties
Homo-oligomer of seven subunits
Bacteria, plastids and mitochondria
Arranged as seven identical subunits in a ring; approximately 50 % homologous to groEL; binds to groEL in the presence of MgATP. Postulated function: regulate the interaction of groEL with unfolded protein targets and/or regulate the ATPase activity of groEL. Likely to be related forms in other intracellular compartments.
- 60-70
Homo- or heterooligomer of 14 subunits
Bacteria, plastids and mitochondria
Distinctive ‘double-donut’ structure of 14 subunits arranged as two rings of seven subunits each; bind and hydrolyze ATP. Postulated function: in conjunction with groES and ATP, provides a surface for monomeric protein folding and/or higher ordered protein assembly. Likely to be related forms within other intracellular compartments.
hsp73
- 70
hsp72
- 70
w75
- 70
Monomer and/or dimer Monomer and/or dimer Monomer and/or dimer
grp7B (BiP)
- 70
Cytosol and nucleus Cytosol and nucleus Matrix of mitochondria and plastids Lumen of endoplasmic reticulum
All members are highly related and bind to ATP; most are constitutively synthesized, although hsp72 is expressed primarily only in the cell after stress. Postulated function: recognize and bind to unfolded proteins such as nascent polypeptides undergoing synthesis and/or translocation into organelles. After metabolic stress, likely to be involved in denatured proteins.
Size Apparent
(kDj
Native
1~~60 (groEL/ESj groES
hsp60 or groEL
-
10
1sp70
3iP, binding protein; grp, glucose-regulated
Monomer and/or dimer
protein;
hsp, heat-shock
protein.
1035
36
Post-transcriptional
processes
.
lated plant Rub&co-binding protein in facilitating the assembly of the Rub&o complex in vim 1161. In yeast, groEL-relatedproteins (hsp60) have been described and again appear to be integral in facilitating higher ordered protein assembly events. For example, via both genetic and biochemical studies, yeast hsp60 has been shown to interact with a variety of newly synthesized m itochondrial proteins, catalyzing their assembly into their linal oligomeric and enzymatically active forms [17). Such processesappear to be dependent on ATP and probably require the assistanceof additional components, one possibility being a eukaryotic homolog of groES [18*]. On the basis of all these studies, it has been suggestedthat groEL provides a surface, or workbench which binds to unfolded proteins and, through a series of ATP-hydrolysisevents, perhaps regulated in part by groES, results in the proper folding of the target polypeptide [3**,15,19**]. What remains unclear is how unfolded targets are recognized by groEL, how the energy of ATP hydrolysis is coupled to protein folding and, iinally, the exact role of groEI/ES in facilitating the assembly of monomeric proteins into their final higher ordered, oligomeric structure.
The concept
of molecular
intimately with the carboxy-terminal domain? Moreover, what prevents the amino-terminal domain from inadvertently folding with components of the translational machinery? Similarly, during protein translocation, as proteins enter into the lumen of an organelle in an unfolded state (as has been shown), much of the polypeptide chain is still present within the cytosol. Hence, again you have a situation in which all of the information required for folding is not yet present within the organelle lumen. What prevents the partially translocatedpolypeptide from folding prematurely? It has been suggested that members of the hsp70 family function to bind to and prevent the premature folding of proteins that exist in an intermediate, unfolded state. Members of the groEI/ES family may then take over this role and catalyze the proper folding and assemblyof the target protein. Thus, for any given polypeptide, acquisition of its final or mature structure may be dependent upon one or more chaperones, perhaps working in tandem. In addition, one m ight suspect that the chaperones play an integral role in facilitating dynamic changes in mature proteins whose as sembly/disassemblyis regulated in response to different biological cues. Whether all proteins require the assistance of molecular chaperones for their maturation still remains to be determined.
chaperones
Owing to their role in facilitating protein folding and assembly events, members of the hsp70 and hsp60 families are now often referred to as ‘molecular chaperones’(reviewed in [3-l). In general terms, a chaperone is defined as ‘one delegated to insure proper behavior’. In molecular terms, molecular chaperones are broadly dehned as a class of unrelated proteins that mediate the correct assemblyof other polypeptides. In no known instancedo molecular chaperones themselvesactively provide the information for folding or assembly; nor do they become part of the linal folded structure itself. Rather, they seem to function by preventing nonproductive folding and/or higher ordered assemblypathways, and thereby ensure that productive folding pathways occur quickly and with high fidelity. Finally,the concept in no way contradicts the principle of self-assembly. Specifically,proteins do assume their fmal conformation as a consequenceof their primary amino acid sequence and relevantpost-translationalmodifications (phosphory lation, glycosylation,etc.), but may require the participation of molecular chaperones to ensure that folding and assemblyoccur correctly and quickly inside the cell. A particularly illustrative example of the requirement of molecular chaperonesis provided by examining the steps of protein synthesisand/or protein transkxation. For example, during the course of protein synthesis, as the nascent chain emerges from the ribosome, a situation arises in which all of the information necessaryfor folding of the polypeptide chain is not yet present. Whereas individual protein domains can fold independently of one another, how does the cell cope with an aminoterminal domain whose fate is to fold with or interact
Why are some molecular after stress?
chaperones
induced
Having discussed the concept of molecular chaperones, the question of why the expression of some chaperones is increased in cells subjected to metabolic stress, should be addressed A clue to the answer is the fact that many of the agents/treatmentsthat induce the stress response interfere with protein folding/assembly. For example, whereas molecular chaperones appear to interact transiently with their protein targets in cells maintained under normal growth conditions, we have found that such interactions no longer appear transient, but are instead stable in cells experiencing stress (6**]. Specifically, in cells exposed to amino acid analogs, a potent inducer of the stress response, newly synthesized proteins are again found to interact with hsp70. Unlike the situation in normal cells, however, the analog-containing proteins are not immediately releasedfrom their chaperone once their synthesisis complete. Presumably,when the newly synthesizedprotein incorporates an amino acid analog, it is unable to commence along its proper folded pathway and, as a consequence,continues to appear unfolded. With time, and as more proteins are made but are unable to properly fold, they remain bound to hsp70, and as a consequence the available levels of hsp70 are reduced. The cell then responds by upregulating the expression of hsp70, in effect, by the induction of a stress response. In a similar but neverthelessdistinct scenario, when cells are subjected to heat-shocktreatment, a number of mature cellular proteins begin to denature and consequently now appear as unfolded targets for hsp70. Again, the net result is a reduction in the availablelevels
The role of heat-shock
of hsp70 and a corresponding induction of the stress response. A similar scenario for groEI./hspbOmay also be operative 191. What remains to be determined is the fate of such proteins unable to commence folding, or mature proteins which become unfolded. Whether these denatured proteins can be rescued,that is refolded [ 200,21*], or whether they are simply maintained in a soluble form via binding to molecular chaperones such as hsp70, and subsequentlydegraded [22], remains to be established in vivo. Depending upon the extent of the protein denaturation, both of these scenarios will probably prove operative in the cell experiencing stress.
New insights regarding the details of protein maturation have emerged over the past few years and have raised new and exciting questions concerning the molecular mechanismsgoverning protein folding and assembly events.It is now well establishedthat such processesare facilitated by a class of proteins referred to as molecular chaperones,many of which represent members of the socalled heat-shockor stress protein family. In view of the themes outlined above, and with the appropriate chaperones now availablein purified form, the exact molecular details by which proteins assumetheir final structures will surely be forthcoming.
and recommended
mitochondria. 8.
.
special interest, published highlighted as: interest outstanding interest
within
the annual period
of review,
Demonstrates a general requirement for functional BiP in the import of proteins from the cytosol into the endoplasmic reticulum. 9. .
2.
ROMMAN JE: Polypeptide Chain Binding Roteins: Catalysts of Protein Folding and Related Processes in Cells. Cell 1989,
10.
.
the Folding of Protein
EUS RJ, VANDER VLES SM: Molecular Chaperones. Annu RetI .. Biocbem 1991, 60:321-347. Excellent up-todate review of the concept and description of molecular chaperones. 4. ..
MORIMOTO RI, TIZ+IERES A, GEORCOPOLIUIS C (EDS): Stress Proteins in Biology and Medicine [book]. New York: Cold Spring Harbor Laboratory Press, 1990. Excellent collection of review articles covering different aspects of the stress response. m J, G~~DEERG AL, VOEUMY R: Abnormal Roteins Serve as Eukaryotic Stress Signals and Trigger the Activation of Heat Shock Genes. Science 1986, 232:252-254. BECKMANN RP, MIZZEN L4. WELCH WJ: Interactions of hsp 70 with Newly Synthesized Proteins: Implications for Rotein Folding and Assembly. Science 1990, 248:850-854.
Biochemical evidence implicating a role for hsp70 in the co-translational binding of proteins under synthesis on the ribosome. Discusses the possible mechanism by which the stress response is regulated.
.
Mi-
PAILEROS DR, WELCH WJ, FV‘IK AL: Interaction of hsp 70 with Unfolded Roteins: Effects of Temperature and Nucleotides on the Kinetics of Binding. Pnx: Natl Acad Sci I/ S A 1991, 88:571+5723.
the role of nucleotides 11.
.
in mediating
such events.
HOLMESKC: Siiry of the Three-Dimensional Structures of Actin and the ATPase Fragment of a 70-kDa Heat Shock Cognate Rotein. Prcrc Nat1 Acad Sci US A 1991, 88:5041-5045.
FLAHER-IY KM, McKr;v DB, KAEXH W,
Discusses the three-dimensional structure of the hsp70 ATPase domain and its similarities to other ATP-binding proteins such as actin and hexokinase. 12.
A Hypo thetical Model for the Peptide Biding Domain of hsp 70 Based on the Peptide Binding Domain of HLA EML30 / 1991, 10:10531059.
R~PPMAN F, TAYLOR WR, ROTHBARD JB, GREEN NM:
paper proposing that the peptide-binding similar to that of class I histocompatibility
site of hsp70 is proteins.
DUNHAM I, TROYV~DAII J, CAMPBEU RD: Human Major Histocompacibility Complex Contains Genes for the Major Heat Shock Rotein hsp 70. Prcc Natf Ad Sci
13.
SARGENT CA,
14.
G, CHAPPEU T, ROTHMAN JE: Peptide Binding and Release by Roteins Implicated as Catalysts of Rotein Assembly. Science 1989, 245:38>350.
15.
HEMMINGSEN SM, WOOLFORO C, VAN DER VLES SM, ‘I?uY K, DENNIS DT, GEORCXPOUU~ CP, HENDRD( RW, Ews RJ: Homo-
KANc PJ, OSIXRMANN J, SHILLING J, NEUPERTW, Crwc F& PFANNER N: Requirement for hsp 70 in the Mitochondrial
FLY
logous Plant and Bacterial Proteins Chaperone Oligomeric Protein Assembly. Nature 1988, 333:330-334.
3.
7.
Mammalian
tochondrial Stress Roteins, grp 75 and hsp 58, Transiently Interact with Newly Synthesized Mitochondrial Roteins. Cell Regulation 1991, 2165-179.
Demonstrates the affinity of hsp70 for unfolded proteins and discusses
59:591-l.
..
MIZZEN L4, KABILING AN, WELCH WJ: The Two
U S A 1982, 86:1968-1972.
ANFINSEN CB: Principles that Govern Chains. Science 1973, 8:223-230.
6.
MI% LM, ROSE MD: Loss of Bii/Grp 78 Function Blocks Translocation of Secretory Roteins in Yeast. / Cell Biol 1990, 110:188%1895.
VOGEL JP,
Theoretical remarkably
reading
1.
5.
Welch
Matrix for Translocation and Folding of Recursor Roteins. Nature 1990, 348:137-143. Genetic evidence that the mitochondrial form of hsp70 is required for the import of most, if not all, proteins that are translocated into the
.
Papers of have been . of .. of
chaperones
Biochemical study examining the interactions of the mitochondrial stress proteins with newly synthesized mitochondrial proteins. Demon. strates an interaction between the two mitochondrial stress proteins and proposes that autophosphotyiation of grp75 is coupled with the release of its target.
Conclusions
References
proteins as molecular
16.
GOLOUBINOFF P, CHIU~TEUER JT, GATENBY AA, LLXI~IER GH: Reconstitution of Active Dimeric Ribulose Biiphosphate Carboxylase &om an Unfolded State Depends on Two Chap eronin Proteins and Mg-ATP. Nahtre 1989, 342:884-889.
17.
CHENG MY, m FU, M,wm J, POUAXK RA, KAIOUSEK F, NEUPERT W, HALLBERG EM, BERG RL, HORWICH AL Mito-
chondrial Heat-Shock Rotein hsp 60 is Essential for Assembly of Roteins Imported into Yeast Mitochondria. Nature 1989, 337:62&25. 18. .
LABEN TH, GATENBY A4 Do NNDSON GK, tiRlhIER GH, VIJTANEN PV: Identification of a groES-Like Chaperonin in Mitochon-
dria that Facilitates Rotein Folding. Pmc Nat1 Ad U S A 1990. 87:76837687. One of the first identifications of a eukaryotic cusses the biochemical function of groES.
19. ..
homolog
of groES. Dis-
M A R T W J, IANGER T. BOTEIVA R, !XHRAMEL 4 HORWICH AL, m FU: Chaperonin-Mediated Protein Folding at the Surface
groEL Through a ‘Molten Globule’-Lie ture 1991, 352:3&42.
Sci
of Intermediate. Na-
10X
lb38
Post-transcriptional
processes
One of the first in vitm biochemical studies demonstrating a role for groEIjES in facilitating monomeric protein folding. Discussesintermediates during the folding process, and the amount of ATP required for the folding of each monomer. 20. .
G~rr~~~rusGA, PAPAVASSILIOLI AC, RUBOCKP, SILVERSTEIN SJ. G~TTESMAN ME: Renaturation of Denatured Lambda Repressor Requites Heat Shock Proteins. Cell 1990, 61:10131020. A report on the ability of various bacterial heat-shock proteins to facilitate renaturation of a denatured protein. 21. .
SKOWRAD, GEORGOPOUUX C, Zvucz M: The E. coIi dnaK Gene Product, the hsp 70 Homolog. can Reactive Heat-
Inactivated RNA Polymerase in an ATP Hydrolysis-Dependent Manner. cel( 1990, 62:939-944. Reports that dnaK (hsp70) protects proteins from heat denaturation as well as facilitating renarurarion of a denatured protein. 22. CH~ANGHL TERLECKY SK, PLANTCP. DICEJP: A Role for 70 KiIodaIton Heat Shock Protein in Lysosomal Degradation of IntraceIIuIar Proteins. Science 1989, 246:381-385.
WJ Welch, Department of Medicine and Physiology, University of Califomia, San Francisco, California 94143.Ot354,USA
