This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sullivan, C. S. Articles by Pipas, J. M. Search for Related Content PubMed PubMed Citation Articles by Sullivan, C. S. Articles by Pipas, J. M.

 Previous Article  |  Next Article 

Microbiology and Molecular Biology Reviews, June 2002, p. 179-202, Vol. 66, No. 2
1092-2172/02/$04.00+0     DOI: 10.1128/MMBR.66.2.179-202.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.

T Antigens of Simian Virus 40: Molecular Chaperones for Viral Replication and Tumorigenesis

Department of Biological Sciences, University of Pittsburgh, Pittsburgh, Pennsylvania 15260

SUMMARY

INTRODUCTION

SV40 as a Model System

TUMOR SUPPRESSORS

pRB

p53

p300

T-Antigen Interactions with pRB and p53

WHAT IS A CHAPERONE?

DnaK/Hsc70 Families of Proteins

T Antigen Is a J-Protein Chaperone

FUNCTIONS OF T-ANTIGEN CHAPERONE ACTIVITY

Replication

Role of J Domain in pRB Complexes

Non-J-domain-mediated T-antigen effects on the pRB family.

Role of J domain in transactivating E2F promoters.

Transformation

Role of J domain in carboxyl-terminal transforming functions.

J Domain and Virion Assembly

J Domain and Transactivation

J Domain and Tumorigenesis

Role of the J Domain in Other T Antigens

SPECIFICITY OF T-ANTIGEN J-DOMAIN INTERACTIONS

CANCER AND CHAPERONES

THE FUTURE

ACKNOWLEDGMENTS

REFERENCES

Top Next References

INTRODUCTION

Top Previous Next References

 
Simian virus 40 (SV40), a member of the Polyomaviridae family, has a rich history of discovery. Its use as a model system has led to fundamental insights into the molecular processes of genome replication, gene expression, posttranscriptional processing, and cell cycle regulation. Its simple genome structure coupled with the ease that it can be grown and manipulated in cell culture has made SV40 an ideal system for coupling genetic and biochemical approaches to complex cellular control problems. Progress in each of these areas has been aided by studies of other viral systems, particularly the adenoviruses and papillomaviruses, since they face common obstacles to successful infection. Members of each of these families can induce tumors in experimental animals or their natural host, and thus they are classified as tumor viruses. That is, they are viruses with DNA genomes that contribute to tumorigenesis.

Each of these viruses encodes a set of proteins that function to (i) take over key cellular regulatory circuits and (ii) directly contribute to viral genome replication, expression, and/or virion assembly and release. SV40 encodes a 708-amino-acid, multifunctional large tumor antigen (T antigen) that plays several of these roles in viral infection and tumorigenesis. Recently, T antigen was shown to be a DnaJ molecular chaperone, and the chaperone activity is required for T antigen to function in viral replication, transcriptional control, and virion assembly, as well as transformation. In this article we review the chaperone activity of T antigen and discuss its role in each of these viral and cellular processes.

More than 30 years after the discovery that SV40 induces tumors in rodents, the debate continues as to whether SV40 is an etiologic agent of human cancer. In its native host species, the rhesus macaque, SV40 forms a persistent infection in the kidneys with no apparent harmful side effects (except in circumstances when the monkeys become immunocompromised due to SIV infection [164, 214]). However, recent PCR data suggest a potential role for SV40 in human mesotheliomas, osteosarcomas, choroid plexus tumors, and other brain tumors (reviewed in references 22 and 26). Furthermore, live SV40 has been cultured from patients with neurological disease (22). Currently, however, there is no consensus as to whether SV40 is a causative agent of human cancer or whether it merely propagates well in malignant tissue. Given the potent transforming activities of SV40 proteins in multiple cellular environments, it is not difficult to envision at least a collaborative role for SV40 in tumorigenesis when it is present in human neoplasias.

SV40 is a member of the Polyomaviridae family of viruses, named after the founding member's ability to induce multiple tumorigenic lesions in newborn rodents. BK virus (BKV) and JC virus (JCV) are human polyomaviruses which form a persistent infection in 90% of the human population early in life (reviewed in reference 105). BKV has been found in human tumors (105), and PCR data suggest that JCV may be linked to neural cancers (23). However, the contribution of these viruses to cancer remains controversial. Nevertheless, it is clear that JCV is the causative agent of progressive multifocal leukoencephalopathy, a neurodegenerative disease which infects immunocompromised individuals, including approximately 5% of all AIDS patients (8). Given the high degree of sequence similarity (approximately 70%) between SV40, JCV, and BKV (65), it is not surprising that these viruses have similar life cycles. Polyomavirus family viruses replicate in differentiated cells and generally form persistent infections in their native hosts (42). Differences among the polyomavirus family members in the non-protein-coding regions are thought to account for at least some of the variations in tissue tropisms (31).

SV40 has a relatively simple genetic architecture which consists of seven gene products organized into a circular DNA genome (Fig. 1A). The late genes VP1, VP2, and VP3 are the structural proteins that form the viral capsid. Agno protein is expressed late in infection and may have a role in capsid assembly and/or release (42). The early genes—which code for large T antigen, small t antigen, and 17kT antigen—share amino acid identity in the amino-terminal 82 amino acids and are encoded by three alternatively spliced products (reviewed in reference 16). These proteins are responsible for numerous functions, including regulation of early and late gene transcription, induction of host cell transcription (to up-regulate necessary enzymes for DNA synthesis), viral DNA replication, and virion assembly (42) (see map of the T antigens [Fig. 1B]).

View larger version (14K): [in this window] [in a new window]   FIG. 1. SV40 genomic organization and early gene products. (A) SV40 genomic organization of the alternatively spliced early (large T antigen [LT], small t antigen [St], and 17k T antigen) and late (VP1 to -3) proteins. The early (PE) and late (PL) promoters exist in opposite orientations that flank the SV40 origin (Ori) of replication. (B) T antigen is a large 708-amino-acid multidomain protein. It consists of an amino-terminal J domain (that directly contacts Hsc70), followed by an pRB family binding motif (LXCXE) which binds to all three pRB family members, pRB, p107, and p130; a nuclear localization signal (NLS); a specific DNA binding domain (ori binding); a zinc finger motif (Zn); ATPase domain; a bipartite p53 binding domain that is also essential for mediating interaction with the transcriptional adapter protein p300; and an HR specificity region. 17kT antigen is comprised of the first 131 amino acids of large T antigen (including the J domain [J]) and an pRB-binding (LXCXE) motif, plus an additional four unique amino acids. The amino terminus of small t antigen is comprised of the J domain (amino acids 1 to 82), plus an additional carboxyl-terminal domain that binds to the multimeric protein phosphatase 2A (pp2A) comprised of a catalytic (c) and a regulatory (a) peptide.

Expression of T antigen is sufficient to induce transformation in multiple cell types (for examples, see references 17, 36, 82, and 103). Small t antigen assists in transforming some cell types, and in some situations expression of small t antigen by itself is sufficient to induce a transformed phenotype (162, 180, 181, 192). 17kT antigen has not been studied in depth, but it is known that it is expressed during infection in small amounts relative to the other T antigens (270). It is proposed that 17kT antigen may function in fine-tuning SV40-mediated cell cycle control (270).

Study of T antigen has led to the discovery of important insights in the fields of alternative RNA splicing, nuclear localization of proteins, DNA repair, gene regulation, and tumor suppressor functions. Current research uses T antigen as a model system to understand the process of DNA replication. The advantage to using this system is that only one viral protein (T antigen) is required to initiate DNA replication, which in combination with a wealth of characterized mutants and in vitro assays has led to a better understanding of how eukaryotic DNA replication is initiated (21).

Thus, T antigen is a potent model for understanding cellular functions. A recurring theme in SV40 research is that T antigen serves as a molecular divining rod to point to the essential factors required for cellular processes, serving to navigate the researcher through the immensely complicated interactions of the cellular milieu. T antigen has played an important role in identifying and elucidating the function of tumor suppressor proteins and the following sections deal solely with this issue.

TUMOR SUPPRESSORS

Top Previous Next References

 
Cancer is a complex, multifactorial disease that occurs when the balance between growth-promoting and growth-inhibitory signals is disrupted (reviewed in reference 84). Tumor suppressors are growth-inhibitory proteins that are often found mutated in human tumors. Two of the better-studied tumor suppressors are retinoblastoma binding protein (pRB) and p53. Both are the founding members of different families of genes related by structure and function. Why are pRB and p53 important in cancer? They are key regulatory components of cellular pathways involved in growth control. E2F can activate proliferation, and p53 inhibits abnormal cellular proliferation (Fig. 2). T antigen and other tumor virus proteins, such as the early proteins of the adenoviruses and papillomaviruses, have served as tools in elucidating the function of both p53 and pRB.

View larger version (24K): [in this window] [in a new window]   FIG. 2. T antigen inhibits both the pRB and p53 tumor suppressor pathways. Mitogenic stimulation triggers phosphorylation of pRB by cyclin D/CDK4,6 complexes. This releases pRB-mediated repression of E2F transactivation, thus allowing the synthesis of the enzymes necessary for cell cycle progression and DNA replication. T antigen induces free E2F. p53, "the guardian of the genome," inhibits cell cycle progression; one way this is accomplished is via p21, which inhibits phosphorylation of pRB. Additionally, p53 induces apoptosis when activated by genotoxic stresses. T antigen inhibits multiple activities of p53. Hsc70 inhibits apoptosis and may play an additional role in regulating the activities of both pRB and p53. See the text for more details.

In normal cells, pRB is negatively regulated by phosphorylation, which occurs in a cell cycle-dependent manner (254). This may be an oversimplification, because some level of phosphorylation is required to activate pRB function, with additional phosphorylation inhibiting pRB function (57). Thus, cyclin/cyclin dependent kinase (CDK) complexes negatively regulate pRB function and are, in turn, themselves negatively regulated by cell cycle inhibitors such as p16 (Fig. 2) (211). Therefore, cancers possess genetic mutations in many parts of the pRB pathway, including pRB itself and CDK inhibitors or overexpression of cyclins and E2Fs (212).

pRB is part of a family (pRB, p130, and p107) that act as transcriptional repressors but do not interact with specific DNA target sequences. The individual pRB family members share a high degree of sequence similarity in conserved regions, for example the A and B domains (Fig. 3.). The A and B domains are essential for repression of certain transcription factors (see below) (189, 206). pRB consists of multiple domains, some of which bind directly to specific transcription factors while others recruit histone deacetylases (HDACs) (reviewed in references 52, 86, 146, and 163). pRB represses transcription directly through repression domains and indirectly by recruiting HDACs and chromatin-rearranging complexes (86). An important way pRB recruits HDAC activity is through binding to RBP1 (126). In turn, RBP1 also possesses a transcriptional repression activity independent of recruiting HDAC activity (Fig. 4).

View larger version (29K): [in this window] [in a new window]   FIG. 3. pRB binding to T antigen and chaperones. (A) The crystal structure of pRB bound to an E7 peptide (from papillomavirus) containing an LXCXE motif (PDB code 1Gux [131]). Note that pRB binds to LXCXE entirely through its B domain, even though the A domain is required for efficient complex formation with LXCXE motif-containing proteins. For comparison, the conserved amino acids of the T antigen and E7 RB binding (LXCXE) motifs are shown in black. (B) The crystal structure of pRB bound to the first 117 amino acids of T antigen (PDB code 1GH6 [122]). Notice the LXCXE motif of T antigen binds to pRB in a manner similar to the LXCXE E7 peptide. Additionally, the J domain of T antigen is depicted as four multicolored helices: helix 1 (yellow), helix 2 (dark blue), the highly conserved HPD loop in red connecting helices 2 and 3, helix 3 (green), and helix 4 (light green). (C) Domain map demonstrating that the A and B domains of pRB are highly conserved among the other pRB family members, p130 and p107. The essential regions of pRB required for various activities, such as binding to Hsp70 or T antigen, are diagrammed with black lines corresponding to particular regions of pRB (35, 106, 131).

View larger version (17K): [in this window] [in a new window]   FIG. 4. pRB repression of the E2F transcription factors. In growth-arrested cells the pRB family of proteins (pRB, p107, and p130) can mediate transcriptional repression in at least two ways, via direct repression domains and by recruiting the activity of the protein RBP1 which directly represses transcription and indirectly represses transcription by recruiting HDAC (126). In dividing cells, pRB family members are no longer bound to E2F, thus allowing for transcriptional activation of promoters containing E2F binding sites.

pRB family members interact with multiple transcription factors (for a review, see reference 146) such as MyoD (79), Pax-3 (262), c-Abl (256), and CDP/cut (248, 253). However, the best understood, and perhaps the transcription factor most important for RB's growth-suppressive functions, is E2F (Fig. 1). E2F refers to any member of a family of transcription factors (presently there are six members; E2F-1 to -6) that are related by conserved regions of sequence similarity, as well as by the ability to bind to the same consensus DNA binding site (52).

The exact functions of the individual E2F family members are currently unclear, and this topic is the focus of ongoing research. Some E2F family members have overlapping functions in the induction of S phase, promoting expression of genes such as cyclins E and A, cdc2, CDK2, and enzymes essential for DNA synthesis such as DNA polymerase and thymidine kinase (52). However, in certain contexts, such as when bound by pRB family members, the role of E2F is not to activate transcription, but rather the pRB-E2F complex serves to repress transcription (Fig. 4). Thus, in theory, the same E2F peptide can be a growth-promoting or growth-inhibitory agent, depending on its binding partners. Recently, it has been proposed that pRB-E2F complexes may bind near origins of DNA replication and thus play a direct inhibitory role in DNA replication (120) as well as an indirect role by preventing synthesis of the enzymes necessary to replicate DNA and drive cells to cycle.

All known E2Fs function by associating with a heterodimeric DNA binding partner called DP. There are two known DPs, and both display sequence similarity to each other as well as to the E2F family members (52). It is thought that both DP1 and DP2 can bind to all six E2Fs, enhancing their DNA binding function (271, 272).

The E2F family is commonly grouped into three subclasses based on functional and structural similarities. One class includes E2F-1 to -3, which encode their own nuclear localization signal and most frequently bind pRB (52). E2F-1, -2, and -3 are commonly thought to activate cellular division. In support of this notion, E2F-1 knockout mouse embryo fibroblasts (MEFs) are defective for exit from G₀ (252), and E2F-3 is required for cell proliferation induced by loss of pRB (276). A second class includes E2F-4 and E2F-5, which do not contain a nuclear localization signal and are most commonly found associated with p130 and p107. E2F-4 and E2F-5 are not required for exit from G₀ in MEFs, and unlike E2F-1 to -3, at least in some cell types, their role is not to activate cellular division but rather to mediate cell cycle inhibitory signals transduced by p16 (71). The third class includes E2F-6, which does not bind to pRB family members and lacks a transcriptional activation domain. Therefore, it has been proposed to play a predominantly inhibitory role in transcription (27, 72, 245).

The fact that the pRB and E2F families are the focus of intensive study is a testament to their essential function in cellular growth control. Their central role in the molecular mechanisms underlying human cancer have led them to be a focal point for genetic and drug therapeutic approaches (37, 93). The depth of our present understanding of these proteins is just a fraction of what is to come; however, much of what we do know has been enhanced either directly or indirectly by the study of DNA tumor viruses.

As is the case with the pRB family, the field of p53 research finds its roots in the study of DNA tumor viruses. p53 was initially identified as a coprecipitating protein in immunoprecipitation assays of T antigen (30, 125, 127, 145).

p53 is a specific transcription factor, referred to as the "guardian of the genome," whose function can prevent DNA synthesis, cause G₂ and G₁ growth arrest, and induce apoptosis (reviewed in reference 135). The transcriptional adapter protein referred to as p300 (also known as CREB-binding protein [CBP]) is sometimes found associated with p53, and it is thought to assist in numerous cell cycle regulatory functions, including those of p53 (see below and references 2, 80, 142, and 217). p53 is ubiquitously expressed but is not stable; however, upon genotoxic stresses such as irradiation, chemotoxin exposure, or virus-induced unscheduled DNA synthesis, p53 steady-state levels increase. An increase in p53 levels leads to a cascade of events including the transcriptional activation of the CDK/cyclin kinase inhibitor p21 and the ubiquitin E3 ligase shuttling protein MDM2 (reviewed in reference 157). p21 inhibits the cell cycle-promoting functions of the CDKs (156), and MDM2 down-regulates p53 function, inducing the degradation of p53 (Fig. 2) (157). Thus, down-regulation of p53 function by MDM2 provides a feedback loop mechanism that allows the restoration of normal cell function when the genotoxic stress is diminished.

Recently several p53-related proteins have been identified, such as p63 and p73 (reviewed in reference 150). These proteins share regions of sequence similarity with p53 and can bind to the same consensus DNA binding site. Several different proteins are encoded by alternatively spliced gene products of p63 and p73. The exact biological functions of these polypeptides are still to be defined. Some p53 family members have functions similar to, and overlapping with, those of p53, including transactivation, while other members are antagonistic to p53 function (150).

Given the central roles of pRB and p53 in cellular growth control, it is not surprising that both pathways cross talk with numerous cellular signaling pathways such as Ras, MYC, and Egf (13, 136, 215). Shown in Fig. 2 is a bare bones map of the pRB and p53 signaling pathways. Note that each pathway "communicates" with the other; for example, E2F induces S phase but also stimulates p53 activity, which can inhibit cellular division. The overlapping nature of these pathways is an important issue to keep in mind when addressing the transforming activites of T antigen (see sections below).

Until recently, the mechanism of how T antigen disrupts the function of pRB and p53 was thought to be analogous to that of an absorbent sponge; T antigen binds to the tumor suppressor protein and "soaks up" the available pools of pRB and p53. This hypothesis has been referred to in the literature as a "sequestration model." It is now understood that the mechanisms by which T antigen inhibits pRB and p53 function are more complicated. For example, T antigen down-regulates p53 function, but at least part of this activity does not require the p53 binding domain of T antigen. In fact, an amino-terminal fragment of T antigen that lacks the p53 binding domain can still inhibit p53 function in some assays (74, 185, 194). This argues that sequestration alone cannot account for the effects of T antigen on p53. In addition to the pRB binding motif, T antigen requires a functional chaperone J domain to down-regulate some activities of the pRB family members (16, 47). Additionally, a transforming activity(ies) in the carboxyl terminus of T antigen also requires the function of the J domain (223). Several lines of new evidence demonstrate that T antigen is not a static sponge, but rather a dynamic machine, with many of its activities depending on its chaperone function. The rest of this review focuses on the role of the chaperone functions of T antigen in SV40 biology.

WHAT IS A CHAPERONE?

Top Previous Next References

 
Chaperone proteins promote the proper folding of proteins and prevent protein aggregation during periods of cellular stress (90). Historically, some of the first chaperones identified were the heat shock class of proteins identified by the work of Polissi et al. as proteins that are required for the ability of phage to infect Escherichia coli (179). This work identified two structurally unrelated classes of chaperones: the chaperonins (including the Hsp60 family of proteins) and the DnaK/DnaJ families of cochaperones. Both families are involved in promoting the proper folding of protein substrates and are especially important under conditions of cellular stress, such as exposure to extreme temperatures or chemotoxic agents that may lead to denaturation of protein tertiary structure (20). Homologues of both the chaperonins and the DnaK/DnaJ families of chaperones are found in a broad range of species, including all known eukaryotes (186).

Another class of chaperones is that of the Hsp90-related proteins. Hsp90-like proteins are found in prokaryotes and eukaryotes and are involved in activating specific protein signaling molecules, including kinases and steroid hormone binding receptors (19). Hsp90 proteins are the most abundant cytosolic chaperones and are involved in the general stress response functioning to assist in proper protein folding and prevention of aggregation. Perhaps not surprisingly, Hsp90s interact at multiple levels with the Hsp70 chaperone machine (19, 66). For the purposes of this review, discussion is limited to the DnaK/DnaJ families of proteins and their homologues. (For in-depth reviews of the different classes of chaperones see references 19, 20, and 90).

View larger version (40K): [in this window] [in a new window]   FIG. 5. Domain structure of Hsc70. (A) The peptide binding domain of the E. coli Hsc70 protein DnaK (PDB code 1DKZ [275]). The substrate binding domain is shown in purple, the lid (or "variable") domain is shown in blue, and the bound peptide substrate is shown in yellow. (B) NMR structure of the J domain of E. coli DnaJ (PDB code 1BQZ [100]). The D35 residue is modeled with space filling in red to highlight its important role in directly contacting the ATPase domain of DnaK (see the text for more details). (C) The crystal structure of the ATPase domain of DnaK is shown as four alpha-helical regions in purple (PDB code 1DKG [89]). Residue R167 is modeled with red space filling to underscore the importance of the surrounding region in directly binding to residue D35 of the DnaJ J domain. (D) Domain map of Hsc70, the ATPase (amino acids 1 to 386), and the substrate binding (P [for peptide]) domains are shown in purple. The lid (L) domain is shown in blue.

Crystallographic studies show that the DnaK ATPase domain consists of four alpha-helical domains (Fig. 5C) (89). The crystal structure for the carboxyl-terminal substrate binding and lid domain has also been solved for DnaK (275). This structure, shown in Fig. 5A, reveals that the first half (depicted in purple in Fig. 5A) has a ß-sandwich structure that forms the channel which interacts with peptide substrates (depicted in yellow), followed by an alpha-helical region (depicted in blue) that closes over the channel. This alpha-helical region, therefore, has been proposed to be a lid that, when in the appropriate conformation, traps bound substrates in the ß-channel (275). The crystal structures for mammalian Hsc70 ATPase domain (60, 225) and the nuclear magnetic resonance (NMR) structure for the substrate binding domain (159) reveal extensive similarities with the prokaryotic DnaK, suggesting a conservation of the mechanism of function between diverse species.

The ATP-bound form of Hsc70 is in a state of rapid flux between binding and release of substrate (61, 62, 152, 172, 176, 201, 243). Structural evidence suggests that this is accomplished through the action of the extreme carboxyl-terminal lid domain of Hsc70, which clamps shut or open, trapping and untrapping substrates bound by the substrate binding domain in response to completion of the ATPase cycle (Fig. 5A and 6) (275). Upon J-protein stimulation of Hsc70-mediated ATP hydrolysis, a global conformational change takes place in Hsc70 (5, 6, 18, 112, 140), causing the lid domain to clamp shut, thus trapping the substrate in a bound conformation (Fig. 5A and 6).

View larger version (10K): [in this window] [in a new window]   FIG. 6. The ATPase cycle of Hsc70. When bound to ADP, Hsc70 has a high substrate affinity; conversely when bound to ATP, Hsc70 displays a weak affinity for peptide substrates. BAG-1 and GrpE are nucleotide exchange factors that promote the exchange of ATP for ADP, increasing the steady-state ATPase activity of Hsc70. J domain-containing proteins (J proteins) stimulate the ATPase activity of Hsc70, which is inhibited by CHIP. Hip promotes the stabilization of the ADP-bound form of Hsc70.

View larger version (48K): [in this window] [in a new window]   FIG. 7. J domain structure. (A) NMR structures of the J domains of E. coli DnaJ (PDB code 1BQZ [100]), human HDJ1 (PDB code 1HDJ [184]), polyomavirus T antigen (PYV) (PDB code 1Faf [9]) and crystal structure of SV40 T antigen (SV40) (PDB code 1GH6 [122]). Alpha-helix I is shown in yellow, alpha-helix II is shown in blue, alpha-helix III is shown in dark green, and alpha-helix IV (DnaJ and HDJ1 only) is shown in light green. The absolutely conserved HPD tripeptide comprising the loop between helices II and II is shown in red. (B) Amino acid alignment of DnaJ, HDJ1, PYV, and SV40 is shown. The amino acids that make up the particular helices are indicated by colored boxes. The absolutely conserved HPD tripeptide is shown in red. (C) Amino acids of key SV40 T antigen mutants. Alpha-helix 4 of the SV40 J domain is omitted to better show the location of three distinct mutants of SV40 T antigen, representing three different phenotypes (Table 1). The locations of the D44N point mutant and the L19F,P28S double point mutant are indicated by highlighting these residues and their side chains in cyan. The region that corresponds to the small deletion mutant, 17-27, is shown in black.

J proteins may regulate Hsc70 function by binding to substrates of Hsc70 and recruiting and/or stabilizing the substrates into a complex with Hsc70 (116). Tertiary complex formation between Hsc70, a peptide substrate, and a J protein induces maximal stimulation of the ATPase activity of Hsc70 (113, 129, 155). Hsc70-mediated hydrolysis of ATP "powers" the desired "work" being performed, by changing the conformation of the bound substrate and Hsc70 itself. Mutational analysis has revealed that the J domain of J proteins is essential for the induction of increased ATPase stimulation by Hsc70 (223, 247, 249). Structure and function as well as NMR perturbation analysis demonstrated that the J domain directly contacts the ATPase domain of Hsc70 through alpha-helix 2 and the HPD motif (70, 77, 231). J proteins may also contact the carboxyl-terminal region of Hsc70 since mutations of Hsc70 in the substrate binding or the EEVD motif of the lid domain impair productive DnaJ/Hsc70 interactions in various species (63, 113, 232).

One can visualize this basic Hsc70 ATPase mechanism (Fig. 6) accounting for multiple Hsc70 activities. For example, transport of newly synthesized proteins into the lumen of the ER requires a luminal Hsc70 homologue, BiP (153). Various models depict BiP as either anchoring peptides that pass through the membrane pore via Brownian motion, or alternatively, BiP may act as a motor that actively pulls peptides through the pore (14). Either model requires the binding and release of peptide substrates by BiP. Another example of the versatility of the Hsc70 ATPase cycle is the disassembly of O-P-DnaB multiprotein complex at the origin of phage replication (73, 139). Phage uses the E. coli host DnaK and DnaJ proteins to assemble and disassemble the components of the replication machinery necessary for phage DNA replication. Thus, the role of Hsc70 in multiple contexts is to use the force generated by ATP hydrolysis to bind, alter, and then release the substrate so that Hsc70 is recycled to perform additional functions. Clearly, disparate results can occur by the same basic Hsc70 chaperone mechanism—depending on the context in which the Hsc70 is found.

There are several modulators of Hsc70 function that regulate different parts of the Hsc70 ATPase cycle (Fig. 6). For example, in prokaryotes there is a nucleotide exchange factor referred to as GrpE that promotes release of ADP and binding of ATP by DnaK (171). The result is a GrpE-induced enhancement of the steady-state ATPase activity of Hsc70. Thus far, except in the mitochondria and chloroplasts, no structural homologues to GrpE have been identified in eukaryotes. However, functional homologues that foster the nucleotide exchange of Hsc70 from the ADP-bound to the ATP-bound form have been found. These include isoforms of the Bag-1 protein. Like GrpE, Bag-1 binds to Hsc70 and stimulates the exchange of ADP for ATP, thus enhancing the steady-state ATPase activity of Hsc70 (96, 228, 240); however, Bag-1 also serves to negatively regulate some Hsc70 functions in cultured cells (166). Additionally, the multiple isoforms of Bag-1 are involved in different elements of hormone receptor regulation (81).

Several regulators of Hsc70 function contain tetratricopeptide motifs that (in part) foster binding to Hsc70. These include Hip, CHIP, and HOP (4, 49, 66). Hip stabilizes Hsc70 into the ADP-bound high substrate affinity form and facilitates the activation of hormone receptors (97, 183). CHIP performs the opposite function by promoting the stabilization of the low substrate affinity ATP-bound form of Hsc70 (4). Additionally, CHIP targets some proteins for proteasome-mediated degradation (reviewed in reference 154). HOP is a bridging molecule that fosters association of Hsc70 with Hsp90 (205, 216). The roles of some of the regulators of Hsc70 function are depicted in an Hsc70 ATPase cycle mechanistic model in Fig. 6. Note that these added points of regulation can allow for a subtle fine-tuning of the Hsc70 chaperone motor machine.

There are many DnaJ and Hsc70 homologues in the cell. In yeast there are more than 14 different Hsc70-like or DnaJ-like proteins, some of which are localized to the same cellular compartment (186). How then does a particular Hsc70 mediate interaction with its proper binding J-protein partner? Experiments in yeast have demonstrated that there is specificity to the interaction between DnaJ-like proteins and Hsc70 family members. The endoplasmic reticulum (ER) luminal DnaK homologue BiP but not Ssa1p (a cytosolic Hsc70) can associate with the J domain of the ER luminal chaperone Sec63p (153). Ydj1p, a cytosolic J protein, stimulates the ATPase activity of Ssa1p by 10-fold, but does so only 2-fold for BiP (153). When the J domain of the ER luminal chaperone Sec63p was replaced with the J domain of either Sis1p or Mdj1p, these cytosolic and mitochondrial J domains could not substitute for the J domain of Sec63p. However, changing only three amino acids in Sis1p restored function, presumably by fostering interaction with an ER luminal DnaJ homologue (200). Other experiments indicate that the J domains from E. coli DnaJ and other nonmitochondrial J proteins can substitute (to various degrees) for the J domain of the mitochondrial luminal J protein Mdj1p when their expression is targeted to the mitochondria (147). DnaJ restored wild-type function; Xdj1p, Ydj1p, and Sis1p restored function to an intermediate level; and Scj1p lacked the ability to restore viability and respiration. These results suggest that while the basic mechanism of Hsc70-J-protein function is conserved in the various orthologues, elements within and surrounding the J domain contribute the specificity of interaction of particular chaperone partners. Similar approaches, when applied to other Hsc70/DnaJ-like proteins, may provide an understanding of the determinants of specificity in chaperone interactions.

First, there is sequence similarity between the amino-terminal region of the SV40 T antigens and the conserved residues of the type 3 DnaJ-like proteins including the absolutely conserved HPD loop of the J domain (Fig. 7) (33; W. L. Kelley and S. J. Landry, Letter, Trends Biochem. Sci. 19:277-278, 1994). Recently, the crystal structure of SV40 T antigen bound to the A and B domains of the pRB protein has been reported (122). SV40 T antigen shares 44% identity in the J-domain region with polyomavirus T antigen, and the NMR structural determination of the first 79 amino acids of polyomavirus T antigen demonstrates that it bears extreme similarity to the T-antigen J domain (Fig. 7A) (9). Consistent with the NMR structures of DnaJ, HDJ1, and polyomavirus T antigen, the SV40 T antigen is composed of a finger-like projection of two antiparallel helices (Fig. 3). Like the polyomavirus T antigen, the first alpha-helix of SV40 T antigen is longer than the nonpolyomavirus J domains, suggesting the possibility of additional functions in the extreme amino-terminal region. Interestingly, alpha-helix 4 curls back in the opposite direction of other known nonpolyomavirus J domains and makes contacts with alpha-helix 2 and the HPD loop. The extended long loop connecting alpha-helices 3 and 4 is also novel to T antigen. Not surprisingly, the conserved residues of the LXCXE motif of T antigen directly contact the B domain of pRB in a manner analogous to the papillomavirus E7 peptide (Fig. 3). Interestingly, helices 3 and 4 of T antigen make additional hydrogen bond contacts with pRB. Because regions of T antigen that contact Hsc70 (HPD motif) are proximal to the regions of T antigen that bind to pRB, this structure supports the model that T antigen serves as a bridge to direct the action of Hsc70 to RB-E2F complexes (122).

Second, functional studies involving domain-swapping experiments show that the J domain of T antigen can functionally substitute for the J domains of E. coli DnaJ in a phage growth assay (118) and for Ydj1p in a yeast viability assay (S. Fewell and J. L. Brodsky, unpublished observation). Mutation in the amino-terminal region of T antigen in residues conserved with other J proteins renders T antigen defective for SV40 replication, transformation, and assembly. These mutants fail to substitute for the DnaJ J domain in the E. coli complementation assay (J. V. Vartikar and W. L. Kelley, unpublished observations). Furthermore, the J domain of human J proteins Hsj1 and DnaJ2 can functionally substitute for the T-antigen J domain in DNA replication, albeit DnaJ2 is less efficient than Hsj1 in this activity (24).

Third, biochemical evidence confirms that T antigen functions as a J domain in multiple assays, including stimulating the ATPase activity of bovine Hsc70 and Ssa1p (cytosolic yeast Hsc70) (223), promoting the release of a bound substrate from Ssa1p (223), and binding to Hsc70 (24, 198, 209, 234, 236). Thus, structural, functional, and biochemical assays demonstrate with clarity that T antigen contains a functional J domain.

FUNCTIONS OF T-ANTIGEN CHAPERONE ACTIVITY

Top Previous Next References

 
Mutational analysis has demonstrated that the T-antigen J domain is essential for multiple viral activities, including viral DNA replication, transformation, transcriptional activation, and virion assembly (Table 1). Each of these topics is discussed in detail below.

Currently the role of the J domain of T antigen in DNA replication is unknown, and two different explanations are possible. First, the J-domain requirement may be direct. In such a model, the T-antigen J domain directly recruits the activity of an Hsc70 homologue to rearrange the necessary replication machinery to drive DNA replication in a manner analogous to phage DNA replication. Second, the J domain may be indirectly required. There are several ways to envision this hypothesis. The J domain may be required to recruit a cellular Hsc70 homologue activity to disrupt chromatin complexes which are inhibitory to DNA replication in the cellular context but absent in the noncellular in vitro assays. Recent evidence demonstrating a possible inhibitory role for pRB-E2F complexes at origins of replication (120) lends credence to this hypothesis. Alternatively, the J domain may be indirectly required for replication to drive the production of a necessary cellular enzyme(s) required for DNA synthesis such as DNA polymerase or thymidine kinase. In support of these indirect hypotheses, in vitro assays in which the J domain of T antigen is mutated (43; P. G. Cantalupo and J. M. Pipas, unpublished observations), or even completely deleted (255), still replicate SV40 DNA in noncellular replication assays (albeit at a partially reduced level) (255). Furthermore, Hsc70 (or homologues) is not a major required component in a reconstituted noncellular in vitro replication assay (263). Finally, addition of purified Hsc70 to an in vitro replication reaction mixture does not enhance the replication efficiency (Cantalupo and Pipas, unpublished observations). Surely there are other ways to envision an indirect role for the J domain in replication; however, the actual mechanism remains undetermined, and only future experimentation will elucidate whether the J-domain requirement is direct, indirect, or some combination of both.

Because the J domain is required in cis with the pRB-binding motif to elicit transformation (223), a model in which the chaperone activity of the J domain directly acts on pRB-transcription factor complexes has developed. In this model (Fig. 8A) the J domain acts to recruit Hsc70 to the pRB-transcription factor complex (in this case E2F, but could easily apply to others such as MyoD or c-Abl) (16). The action of Hsc70 ATP hydrolysis induced by the J domain of T antigen induces the disruption of pRB-E2F, either directly by altering the conformation of pRB or E2F (Fig. 8A) or by recruiting cellular factors to the complex (Fig. 8B). Thus, free E2F is liberated and transcription of the genes necessary for viral replication proceeds.

View larger version (12K): [in this window] [in a new window]   FIG. 8. Chaperone models for disruption of pRB-E2F family complexes. The multiple pRB and E2F family members are represented by "RB" and "E2F." (A) In this model, T antigen recruits Hsc70 to pRB to directly act as a molecular machine that pries apart the pRB-E2F multiprotein complexes. (B) In this model, T antigen recruits Hsc70 to a multiprotein complex that requires the action of an additional unknown cellular protein (C [for cellular factor]) to disrupt pRB-E2F complexes (see the text for details). Factor C may posttranslationally modify pRB or E2F (denoted with asterisks) after they are separated from each other by the action of Hsc70. An alternative explanation is that factor C may enhance the activity of Hsc70 directly.

However, there exists a caveat to interpreting the above data. Since T antigen is a powerful mitogen with multiple growth-inducing activities, it is possible that the J domain is required to drive the cells to cycle in a manner that indirectly disrupts pRB-E2F complexes. For example, several different mitogens will induce free E2F and transcriptional activity even though they are not known to directly bind to the pRB family (83, 92, 169, 258). Biochemical evidence, however, clearly demonstrates that T antigen has the capability to disrupt pRB-E2F family complexes in vitro (233). When a lysate from cells that overexpressed p130-E2F complexes was incubated with T antigen, p130 remained bound to E2F-4. However, inclusion of exogenous Hsc70 and an ATP regeneration system in the reaction released a portion of the E2F from p130. These results indicate that disruption of p130-E2F requires a functional J domain. The released E2F is capable of binding DNA containing an E2F consensus-binding site, consistent with its role as a transcription factor. Interestingly, the chaperone-mediated release of pRB family members from E2F is more efficient (approximately sixfold) in the presence of an unknown proteinaceous factor (designated factor C, for cellular protein) (233; C. S. Sullivan and J. M. Pipas, unpublished observation). Because the release reaction is enhanced by Hsc70 and ATP, it is possible that factor C is a cochaperone such as a Hip or a Bag. Alternatively, it is known that the phosphorylation state of pRB and E2F family members changes as cells proceed through the cell cycle and divide (52, 148). Phosphorylation of pRB prevents its association with E2F, and phosphorylation of E2F prevents its ability to bind to DNA in vitro (50, 51). Therefore, another plausible possibility is that factor C is a phosphatase or a kinase that requires Hsc70 to potentiate the disruption of the pRB-E2F family complex. In this model, T antigen acts as a switch that uses its chaperone activity to propagate some secondary effector such as a kinase to p130 or E2F-4 (Fig. 8B). In support of this notion, the J domain of T antigen is required to alter the phosphorylation state of p130 in cultured cells (229). An in vitro system of defined components will greatly enhance the testing of these models.

Non-J-domain-mediated T-antigen effects on the pRB family. pRB has multiple functions including repressing the activities of E2F and other transcription factors (see section "pRB" of this review). The role of the J domain on non-E2F transcription factors remains to be determined. Consistent with the multifaceted growth regulation mechanisms that pRB possesses, T antigen can alter pRB cellular growth regulation in a J-domain-dependent (see above) or J-domain-independent manner. M. J. Tevethia and coworkers demonstrated that fusing the pRB-binding motif to a heterologous site on a carboxyl-terminal T-antigen fragment (amino acids 128 to 708) enabled this fragment to induce cells to grow to a high density (242). Therefore, the pRB-binding motif can contribute some growth-enhancing function(s) without the presence of a J domain. In an established rat embryo fibroblast cell line, the T-antigen J-domain mutant 17-27 induces expression of B-myb mRNA as well as wild-type T antigen, which the authors interpret as a productive interaction with p130 (182). Sheng and coworkers (210) have shown that the polyomavirus T antigen induces apoptosis in C2C12 myoblast cells and that this depends on an intact pRB-binding motif (LXCXE), but not a J domain, since mutant H42Q induces apoptosis nearly as well as the wild type. In contrast, SV40 T antigen prevents apoptosis in a neural astrocyte precursor cell line upon growth factor withdrawal, and this activity requires both the pRB-binding motif as well as the J domain of T antigen (215). Therefore, pRB possesses multiple activities intimately associated with the balance of cell growth and death which T antigen can disrupt—some in a J-domain-independent manner.

In another example, wild-type T antigen and J-domain mutants of T antigen restore growth to a cell line that is growth inhibited by the conditional expression of p53 (74, 185). The pRB-binding motif is required for this activity, but the J-domain mutant H42Q functions as well as the wild type. Interestingly, a deletion mutant of the entire J domain (2-82) compromises the ability of T antigen to inhibit growth arrest by 80%, which the authors attribute to poor expression levels of 2-82 (74). The small-deletion mutant 17-27 is not functional in this assay, reflecting a more severe phenotype than point mutants in the J domain (see "Transformation" section below). Thus, while it is clear that the J domain is required to disrupt E2F from pRB, it has yet to be determined what activity the LXCXE motif possesses independent of the J domain. One interesting hypothesis is that the LXCXE motif by itself is able to disrupt HDACs from pRB (74, 210), thereby reducing the ability of pRB to inhibit transcription. There is support for such an idea since peptides corresponding to the LXCXE motif of T antigen inhibit or disrupt complex formation between HDAC-1 and pRB (149), but whether or not this occurs in the context of the cell remains to be determined. Another possibility is that the LXCXE motif or amino acid sequences surrounding it interact with some other cellular target that affects growth control independent of T antigen's actions on pRB.

Role of J domain in transactivating E2F promoters. Multiple studies of polyomavirus, BKV, and SV40 T antigen demonstrate that the J domain and pRB-binding motif of T antigen contribute to alleviating repression of E2F transactivation (32, 74, 88, 134, 209, 210, 269). These studies are performed by transiently transfecting reporter constructs containing an E2F site(s) upstream of a reporter gene in the context of a minimal (E2F site upstream of a TATA box) or a portion of a physiological promoter such as the E2F-1 promoter. In five of the six reports, a J-domain point mutant is diminished in inducing the transcriptional activity of E2F relative to wild type (32, 74, 88, 134, 209, 210, 269). Additionally, J-domain mutants are consistently less defective than pRB-binding mutants in these assays. This suggests that the LXCXE pRB-binding motif contains additional activities that alleviate pRB-mediated repression of E2F family members in a J-domain-independent fashion. An alternative interpretation is that the J-domain point mutants assayed are "leaky," thus retaining partial J-domain function. The J domain and LXCXE pRB-binding motif also act to alleviate the trans-repression activity of pRB that is independent of binding to E2F. This was demonstrated by experiments in which the repression activity of a fragment of pRB (containing the trans-repression domain) which cannot bind to E2F, is alleviated by T antigen (74). This activity was dependent on an intact LXCXE motif and J domain. These data suggest that in addition to alleviating pRB-mediated repression of E2F, both the J domain and LXCXE motif combine to alleviate pRB transcriptional repression that is independent of disrupting pRB-E2F complexes. Using a physiological cyclin A promoter with the E2F binding sites mutated, Sheng et al. demonstrated that LXCXE mutants transactivate cyclin A as well as wild-type T antigen (210). In this assay, a J-domain mutant is unable to transactivate cyclin A. These results suggests that the J domain has the ability to regulate some promoters independent of liberating E2F from pRB family repression.

The question arises as to why the J domain is required to transactivate promoters containing E2F to various degrees, ranging from no requirement with J-domain mutants transactivating reporter constructs as well as wild-type T antigen (32) up to J-domain mutants that are 10-fold defective relative to wild type (134). There are several experimental parameters that may account for at least some of these differences. Because many of these studies depend on transient transfections, proper normalization to account for discrepancies in transfection efficiencies must occur. Additionally, in the six studies (32, 74, 88, 134, 209, 210, 269), four different cell types were used, making it plausible that each may have differing pRB/E2F or chaperone levels. Finally, there is some suggestion that the stage of cell cycle division of the cells affects the degree to which the various domains of T antigen are required to transactivate E2F (210).

Different J-domain mutants displayed various degrees of penetrance for transformation ability (Table 1). The small-deletion mutant 17-27 (dl1135) (diagrammed in Fig. 7C) is 100% defective for transformation, while the D44N mutant (Fig. 7C), which contains a mutation in the highly conserved HPD motif, is capable of inducing transformation, albeit in a partially defective manner, inducing only 50% of the foci that wild-type T antigen does.

What can account for this discrepancy? There are at least three possible explanations. First, it is possible that the D44N mutation only partially disrupts J-domain function while the 17-27 mutation completely abolishes activity. The D44N mutation is analogous to the E. coli DnaJ mutant D35N, which is defective for J-domain function. A suppressor mutation in DnaK of the E. coli D35N mutant is R167H (Fig. 5B) (231). Suh et al. have shown that the cleft region surrounding R167 is a likely binding site for the HPD and helix 2 of the J domain (shown in Fig. 5C). It is reasonable to infer that the J domain of T antigen contacts an analogous cleft domain in Hsc70 and possibly other mammalian DnaK homologues. Extending this inference, the D44N mutation likely diminishes stimulation of Hsc70 activity in a manner analogous to that of the phenotypic null E. coli D35N mutant. Furthermore, studies using D44N confirm that it is defective for several chaperone-related activities, including binding to Hsc70 (234), stimulating Ssa1p ATP hydrolysis in a single turnover assay (although D44N still has approximately twofold more activity than a control mutant lacking the entire J domain) (233), and functioning for the E. coli DnaJ J domain in a phage replication assay (Vartikar and Kelley, unpublished observation). These data suggest that D44N retains little, if any, chaperone activity.

Second, it is possible that 17-27 causes a gross structural abnormality in the T-antigen molecule, such that it would interfere with the other more carboxyl-terminal transforming functions of T antigen including binding to the pRB or p53 family. This seems unlikely, because purified 17-27 is functional in numerous biochemistry assays such as stimulation of DNA replication, double hexamer assembly, ATPase activity, helicase activity, and ori unwinding, suggesting that it has structural integrity (43). Furthermore, 17-27 can bind to both p53 and pRB (56, 151, 178) and induce lymphomas when expressed in mice (238). However, it has been suggested that the 17-27 may be less stable than wild-type T antigen (151). Thus, while maintaining global structural integrity, lower steady-state levels in some cells could account for the more severe phenotype of 17-27, independent of its J-domain activity.

Third, it is possible that 17-27 abolishes another transforming function in the amino terminus in addition to the J-domain activity. In support of this idea, Cavender et al. have shown that the amino-terminal first 82 amino acids convey an essential, independent function that does not require binding to pRB to transactivate a polI-dependent promoter (29). Additionally, mutants in the polyomavirus J domain in the extreme amino-terminal region are defective for transformation induced by middle T antigen (44), even though the NMR structure does not support a role for these residues in contacting Hsc70 (9). The amino terminus of T antigen has a stretch of amino acids with weak amino acid similarity to the conserved CR1 of adenovirus E1A (177). The CR1 of E1A is required to bind to p300 (250) and to inhibit pRB activities in vitro (58, 104). Because both E1A and T antigen bind to pRB and induce transformation, it has been proposed that the CR1-like region of T antigen may share functionality with the corresponding sequence in E1A (265). The 17-27 mutation deletes portions of both the CR1-like region as well as a part of alpha-helix 2 of the J domain. Thus, it is possible that this mutant targets a CR1-like function as well as a J-domain function of T antigen. However, the crystal structure of the SV40 large T antigen predicts that the CR1-like region of SV40 is buried in the hydrophobic core of alpha-helices 2 and 3 of the J domain (Fig. 7C), which makes it difficult to envision a separate transforming activity for this exact region (amino acids 17 to 27). The 17-27 mutant, however, could abolish additional transforming activities in the extreme amino terminus of SV40 T antigen, similar to that of polyomavirus middle T antigen. Notice in Fig. 7 that the J domains of polyomavirus and SV40 T antigens contain extra conserved sequences amino terminal to the start of alpha-helix 1 of the nonpolyomavirus J domains. Therefore, it is possible that the 17-27 mutant alters 2 independent functions: the J domain and a transforming function amino terminal to amino acid 17. Construction of mutants in the first 17 amino acids of SV40 T antigen should be informative regarding this possibility.

Further complicating matters, the degree of the J-domain transforming contribution is dependent on the assay used to measure cellular growth deregulation. The cis requirement for the J domain and pRB-binding motif to transform REF52 cells was determined using crystal violet staining for dense focus formation. In this assay, cells are allowed to grow for approximately 6 weeks. As mentioned above, 17-27 is totally defective for focus formation, but D44N is only partially defective, forming approximately 50% the number of foci. However, when D44N is examined in an assay with growth to high density, it is completely defective (229). This assay is performed on a much shorter time scale (approximately 2 weeks) than the dense-focus assay and therefore likely measures different aspects of cellular growth control. Importantly, another mutant in the conserved loop of the J domain, H42Q, is also defective at inducing cellular growth to high density (229). However, the D44N mutant induces growth in soft agar as well as wild-type T antigen (229). Furthermore, TN136, a fragment of T antigen consisting of the J domain and the pRB-binding (LXCXE) motif, is capable of inducing foci in C3H10T1/2 cells (at a much reduced level relative to the wild type) but is unable to induce foci in REF52 cells (223). Interestingly, unlike the full-length situation, a D44N mutant in the context of TN136 is completely defective at inducing foci in C3H10T1/2 cells. The above data point out the varied requirements for J-domain function depending on cell type and the transformation assay employed. Unlike the J domain, a functional pRB-binding motif of T antigen is required to induce transformation in most assays, including induction of foci in REF52 and growth in soft agar. Interestingly, the fact that the J domain is required for transformation of some cell types in the context of TN136, but not full-length T antigen, suggests that the carboxyl-terminal domain of T antigen may encode transforming activities redundant with the J domain. Work from Cavender et al. supports such a notion (28).

Role of J domain in carboxyl-terminal transforming functions. The J domain is required in cis with some as-yet-unknown carboxyl-terminal function for transformation. This was established using the TN136 amino-terminal fragment of T antigen and 17-27 which were unable to complement each other for full transformation of REF52 or C3H10T1/2 cells (223). As mentioned previously, the carboxyl-terminal function of T antigen that is required is unknown, but p53 is a likely candidate. However, J-domain mutants of T antigen function as well as the wild type in blocking p53 DNA binding in vitro (P. A. Carroll and J. M. Pipas, unpublished data), suggesting that the J domain has no role in this activity of T antigen. Another possible target of the J-domain function is the p300 protein, whose binding has been mapped to the carboxyl terminus of T antigen (143). Mutants in the amino terminus of T antigen are defective for altering the phosphorylation state of p300 and inhibiting its transactivation activity (54). This suggests that a carboxyl-terminally bound p300 substrate is a potential target for the J-domain activity. Alternatively, since p53 and p300 form a complex, it is possible that the J domain may be required to somehow alter this complex. Finally, it is possible that there are still undiscovered transforming functions in the carboxyl-terminal regions. Consistent with this idea, coexpression of TN136 and DD53, a dominant negative mutant of p53, is transformation defective compared to the p53 binding-competent full-length T antigen (196). Therefore, it is formally possible that the J domain acts independently of p53 and p300 on some other carboxyl-terminal function.

It should be noted that the J-domain function can be complemented in trans in several assays including induction of hepatic tumors and transactivation of a ribosomal promoter (7, 29). Why the J domain is required in cis for some activities (such as transformation of REF52 cells and transactivation of E2F responsive promoters) (209, 223) and not others is unknown. One possible determinant may be whether J-domain mutants of T antigen can functionally oligomerize with mutants of T antigen that contain a wild-type J domain. However, Cavender et al. argue that while oligomerization may be required for trans complementation, it is not sufficient for transactivation, since mutants that contain the regions of T antigen sufficient for oligomerization fail to complement in trans (29). Future experimentation is required to understand the differing cis and trans J-domain requirements in these multiple assays.