Pairs of Vp1 Cysteine Residues Essential for Simia
http://www.100md.com
病菌学杂志 2005年第6期
Department of Molecular, Cell and Developmental Biology and Molecular Biology Institute, University of California at Los Angeles, Los Angeles, California
ABSTRACT
Transient disulfide bonding occurs during the intracellular folding and pentamerization of simian virus 40 (SV40) major capsid protein Vp1 (P. P. Li, A. Nakanishi, S. W. Clark, and H. Kasamatsu, Proc. Natl. Acad. Sci. USA 99:1353-1358, 2002). We investigated the requirement for Vp1 cysteine pairs during SV40 infection. Our analysis identified three Vp1 double-cysteine mutant combinations that abolished viability as assayed by plaque formation. Mutating the Cys49-Cys87 pair or the Cys87-Cys254 pair led to ineffective nuclear localization and diminished accumulation of the mutant Vp1s, and the defect extended in a dominant-negative manner to the wild-type minor capsid proteins Vp2/3 and an affinity-tagged recombinant Vp1 expressed in the same cells. Mutating the Cys87-Cys207 pair preserved the nuclear localization and normal accumulation of the capsid proteins but diminished the production of virus-like particles. Our results are consistent with a role for Cys49, Cys87, and Cys254 in the folding and cytoplasmic-nuclear trafficking of Vp1 and with a role for Cys87 and Cys207 in the assembly of infectious particles. These findings suggest that transient disulfide bond formation between certain Vp1 cysteine residues functions at two stages of SV40 infection: during Vp1 folding and oligomerization in the cytoplasm and during virion assembly in the nucleus.
TEXT
During simian virus 40 (SV40) infection, the newly synthesized major capsid protein Vp1 forms transient intrachain and interchain disulfide linkages in the cytoplasm and is transported to the nucleus (10). We have hypothesized that the disulfide bonds serve to facilitate the orderly interdigitation of Vp1 secondary structural elements, thereby facilitating the folding of the Vp1 monomer and the formation of the Vp1 pentamer (10). At least one type of cysteine pair, consisting of two different Vp1 cysteine residues, is required for forming the disulfide-linked monomer and oligomers observed (10). Neither the identity of the cysteine pair(s) nor their necessity during infection is yet known. Individual mutations of the seven Vp1 cysteines largely preserve viral viability (8), raising the possibility that the cysteines may redundantly function in a redox-coupled Vp1 folding and oligomerization pathway. In SV40 virion, disulfides are not present within individual Vp1 pentamers; the distance between any two cysteine sulfur atoms on a pentamer is at least 10 ?, much longer than the typical bond lengths of disulfides (11). Cys104-Cys104 disulfide bonds are observed between some Vp1 pentamers on SV40 (11, 15), though we have shown previously that they are not important for viral infectivity (8). To identify candidate Vp1 cysteine pairs for the transient disulfide bonding and determine their importance in infection, the cysteine residues were mutagenized in pairs and in larger combinations. All mutant combinations were first created in the pBS-Vp1 plasmid (8) via fragment exchanges among the single-cysteine mutant pBS-Vp1s (8) before the entire Vp1 coding region of the resulting plasmids was inserted into the nonoverlapping SV40-harboring plasmid, NO-pSV40 (6), via XbaI and SacI sites. Then, mutant viral genomes in the form of NO-SV40 were prepared from respective NO-pSV40 DNAs via BamHI digestion and recircularization, as described previously (6).
Three Vp1 double-cysteine mutants, C49A-C87A, C87A-C207S, and C87A-C254A, are nonviable. All possible Vp1 double-cysteine mutants and select multiple-cysteine mutants of NO-SV40 were examined for overall viability by the plaque formation assay, using lysates of the mutant genome-transfected cells. Among the total 21 double mutants (Fig. 1, DM series), three categories of plaque-forming ability can be seen. First, a majority of 15 double mutants had infectious titers that were reduced no more than 10-fold from the wild-type value of 1.5 x 108 PFU per U of transfected lysate. Nine of these mutants, DM1, DM6, DM8, DM10, DM12, DM17, DM19, DM20, and DM21, had average plaque diameters of at least 3.2 mm (classified as large in Fig. 1), and the remaining six, DM3, DM4, DM5, DM7, DM15, and DM16, had average plaque diameters of at least 1.5 mm (classified as medium). Note that all pairwise mutant combinations involving Cys9 or Cys104 are in this rather viable category. The triple mutant C9A-C104A-C207S is similarly viable (MM4 in Fig. 1; see below). Hence, Cys9 and Cys104, are the Vp1 cysteine residues least important for viability. This result contrasts with the finding of Gharakhanian et al. (4) that the simultaneous mutation of cysteines 9 and 104, 9 and 207, or 9, 104, and 207 in SV40, in which a part of the Vp1 coding sequence overlaps with that of Vp2/3, abolishes plaque formation. We believe this difference arises from the concomitant Vp2/3 codon change when the Vp1 Cys9 codon is mutated in the context of SV40, a complication avoided by the use of the nonoverlapping NO-SV40 genome in our study.
In the second category, three double mutants, DM2 (C49A-C87A), DM9 (C87A-C207S), and DM14 (C87A-C254A), were nonviable, having no measurable plaque-forming ability (Fig. 1). These three cysteine pairs, involving Cys87 in combination with Cys49, Cys207, or Cys254, are thus essential for the formation of infectious virions in vivo. These four Vp1 cysteines might engage in transient disulfide bonding at some stages of infectious virion formation (see further analysis of the double mutants below) and be partially redundant in this function.
Third, three double mutants, DM11 (C207S-C267L), DM13 (C49A-C207S), and DM18 (C87A-C267L), had infectious titers that are reduced by 4 to 5 logs and average plaque sizes in the medium range (Fig. 1). The greatly diminished viabilities of these mutants suggest that Cys49 and Cys207 can partially substitute for each other and that Cys267 can partially substitute for Cys87 or Cys207. That is, Cys267 may partake in an auxiliary function in infectious virion formation besides the four cysteines identified above. If so, one would expect that mutating Cys267 together with Cys49 and Cys207 would diminish viability much more than the individual double mutations of DM11 and DM13. This possibility is confirmed by the multiple-cysteine mutant analysis described below.
Among the 11 multiple-cysteine mutants (MM series), there are also three categories of plaque-forming efficiencies. The first group consists of four triple mutants, MM2 through MM5, which had PFU values 1 to 2 logs reduced from that of the wild type and average plaque diameters in the medium range (Fig. 1). That all four mutants were rather viable despite being mutated at both Cys9 and Cys104 confirms that these two cysteines are unimportant. The second group are six multiple-cysteine mutants that were unable to form plaques, including MM1, MM6, MM7, MM8, MM10, and MM11 (Fig. 1). The nonviability of MM1 (mutated at residues 9, 49, and 87), MM6 (mutated at 9, 49, 87, and 104), MM8 (mutated at 9, 87, 104, 207, and 267), MM10 (mutated at 9, 87, 104, 254, and 267), and MM11 (all cysteines mutated) can be attributed to the fact that they all harbor at least one nonviable double-cysteine mutant combination of DM2, DM9, or DM14. Though MM7 (mutated at residues 9, 49, 104, 207, and 267) lacks a lethal double mutation (Fig. 1), it too was nonviable, consistent with the prediction above that a triple mutation of Cys49, Cys207, and Cys267 would be highly detrimental.
In the third category of multiple mutants is the quintuple-mutant MM9, mutated at residues 9, 104, 207, 254, and 267. This mutant had a 5-log-reduced PFU value compared to the wild type, similar to DM11 (C207S-C267L), but had a much smaller average plaque diameter of 0.8 mm (Fig. 1). The greatly diminished viability of MM9 suggests that the remaining Cys49 and Cys87 residues can support SV40 infection to a limited extent, perhaps via forming transient disulfide bonds with each during Vp1 folding and assembly. Cys207, Cys254, and sometimes Cys267 appear to additionally contribute to these processes.
The nonviable Vp1 double-cysteine mutants fall into two phenotypic classes. To further characterize the three nonviable cysteine-pair mutants, DM2, DM9, and DM14, we examined their ability to replicate viral DNA and produce capsid proteins. The extent of DNA replication was normal for all mutants. The amounts of DpnI-resistant viral DNA extracted from all mutant DNA-transfected cells (Fig. 2B, lanes 2, 3, and 4) were comparable to that extracted from the wild type-transfected cells (Fig. 2B, lane 1). On the other hand, the intracellular levels of Vp1 and Vp2/3 were substantially lower for DM2 (Fig. 2A, lanes 3 and 4) and DM14 (lanes 5 and 6) than for the wild type (lanes 1 and 2) or for DM9 (lanes 7 and 8).
The double mutants also differed in the subcellular localization of the capsid proteins expressed from the mutant genomes. Mutant DM9-expressed Vp1 and Vp2/3 largely localized in the nucleus (Fig. 3A, panels g and h), as did the capsid proteins expressed from the wild-type genome (panels a and b). In contrast, mutants DM2 and DM14 had a striking defect in the nuclear localization of the capsid proteins. In cells into which these mutant genomes were introduced, the mutant Vp1s were largely distributed in the cytoplasm as irregular specks, though some were distributed in the nucleus either diffusely or in clumps or granules (Fig. 3A, panels c and e). Although the mutants encode wild-type Vp2/3, Vp2/3 exhibited a localization pattern indistinguishable from that of the coexpressed mutant Vp1s (panels d and f). This colocalization suggests that the mutant Vp1s exert a dominant-negative influence on Vp2/3. Therefore, the differences in capsid protein localization and accumulation divide the nonviable double-cysteine mutants into two classes: C87A-C207S (DM9), whose defect occurs during virion assembly in the nucleus; and C49A-C87A (DM2) and C87A-C254A (DM14), whose defect occurs during the trafficking and possibly folding of the capsid proteins in the cytoplasm.
Mutant C87A-C207S is defective in the nuclear phase of virion assembly. Given that mutant DM9 could replicate the viral genome and accumulate a substantial level of the capsid proteins that localize to the nucleus, we asked whether the DM9 capsid proteins could package viral genome into virion-like particles (VLPs). Lysates were prepared from wild-type and mutant genome-transfected cells and treated with DNase I. (We noted that it was difficult to detect the VLP peak in sucrose gradient without the nuclease treatment.) The DNase I-resistant fraction, expected to contain VLPs but not assembly intermediates, was then sedimented through a sucrose gradient. Whereas the wild-type sample exhibited abundant cosedimented viral DNA and Vp1 (Fig. 2C, fractions 1 through 10) around the expected position for virions (fraction 5), the DM9 sample gave less viral DNA in the same region of the gradient (Fig. 2D), even though twice as much transfected material was analyzed. Much of the mutant Vp1 did not cosediment with the DNA but was distributed heterogeneously in fractions 8 through 16. This sedimentation profile suggests the formation of heterogeneous assembly intermediates (75S to 160S), in which the packaging of minichromosomes has been initiated but virion particles are not formed. In SV40 capsid structure, Cys87 is located on the BC2 loop, and Cys207 resides in the E''' region of the EF loop (15). These structural elements are part of Vp1 structural region I, to which temperature-sensitive (ts) B mutations have recently been mapped (7). The phenotype of these ts mutants is the accumulation of 100S to 160S assembly intermediates (1, 2, 13). The similarity of DM9 to these ts mutants, in terms of the structural location of the mutations and the type of assembly intermediates accumulated (Fig. 2D), suggests that the Cys87-Cys207 pair may similarly function in the addition of Vp1 pentamers to the initiated viral minichromosome. Thus, the defect of mutant C87A-C207S consists of a deficiency in the assembly of virus-like particles coupled to a lack of infectivity of any particles that are formed. Further studies are needed to elucidate the role of this cysteine pair in virion assembly.
Mutant C49A-C87A and C87A-C254A Vp1s have a dominant-negative effect on the nuclear localization of Vp2/3. The subcellular distribution patterns of mutant DM2 and DM14 capsid proteins (Fig. 3A) show that these capsid proteins are ineffective in nuclear localization despite the intact nuclear localization signals (NLSs) harbored by both the mutant Vp1s and the wild-type Vp2/3. That the coexpressed wild-type Vp2/3 cannot rescue the nuclear localization of mutant DM2 and DM14 Vp1s is in contrast to the functional complementation observed for some mutants of the Vp1 NLS and DNA-binding domain, in which the wild-type Vp2/3 can effectively piggyback the mutant Vp1s to the nucleus (6, 9). The Vp1 double mutations C49A-C87A and C87A-C254A apparently override the intact NLSs not only of mutant Vp1s but also of the wild-type Vp2/3 in a dominant-negative manner. Our preliminary evidence shows that DM2 Vp1 coimmunoprecipitated with Vp2/3 from NO-SV40-DM2-transfected cell lysate (P. P. Li and H. Kasamatsu, unpublished data). Thus, DM2 mutant Vp1 appears to form a complex with Vp2/3 in the cytoplasm. We speculate that DM2 and DM14 mutant Vp1s fold into somewhat different structures than does wild-type Vp1. This altered folding might either mask the NLSs of Vp1 and associated Vp2/3 or cause the trafficking of these viral proteins to be altered, leading to their trapping and accelerated degradation in the cytoplasm. As might be expected from the blocked nuclear entry of the capsid proteins, mutant DM2 failed to form VLPs: no post-DNase I viral DNA was detected in the sucrose fractions, even though a large amount of the mutant transfected lysate was sedimented (Fig. 2E).
Mutant C49A-C87A and C87A-C254A Vp1s are dominant-negative to wild-type Vp1. To further explore the dominant interfering nature of DM2 and DM14 Vp1 mutations in the cytoplasm, a derivative of NO-SV40, NO-SV40-Vp1C58-H6, was used. It encodes a truncated Vp1 in which the carboxy-terminal 58 amino acids are replaced by a (Gly-Gly-Gly-Gly-Ser)3 flexible linker, followed by a His6 affinity tag. We believe that this truncated Vp1 forms pentamers, since a similar recombinant Vp1 minus the flexible linker was previously expressed in Escherichia coli and purified as a pentamer (9). The construction of NO-pSV40-Vp1C58-H6 proceeded through a series of intermediate plasmids as follows. A PCR fragment generated by using pBS-Vp1C58 (8) as a template, 5'-GTTTACCAACACTAGTGGAAC as a sense primer, and 5'-AAATGATGGGATCCACCAAAGCTAGCTGGGCCGGGGTTTTTCACAGACCGCTTTC as an antisense primer was inserted between the SpeI and BamHI sites (underlined above) of pBS-Vp1C58, producing pBS-Vp1C58-NheI (NheI site in italics above). A fragment encoding the flexible linker (sense strand, 5'-GGTGGAGGCGGTTCAGGCGGAGGTGGCTCTGGCGGTGGCGGATCT), the polyhistidine tag (sense strand, 5'-CATCACCATCACCATCAC), and a stop codon (TAA) was then inserted between the NheI and BamHI sites of pBS-Vp1C58-NheI, producing pBS-Vp1C58-linker-H6. Replacing the SpeI-to-SacI fragment of NO-SV40-BSM (8) with that of pBS-Vp1C58-linker-H6 resulted in NO-pSV40-Vp1C58-H6, which was digested with BamHI and recircularized to yield NO-SV40-Vp1C58-H6. This DNA expressed Vp1C58-H6, which effectively localized in the nucleus, similar to full-length Vp1, as detected by both anti-Vp1 and antitetrahistidine antibodies (Fig. 3B, panels a and b).
When Vp1C58-H6 was coexpressed with DM2 Vp1 (Fig. 3B, panels c and d) or with DM14 Vp1 (panels e and f), derived from respective mutant NO-SV40s, the localization of Vp1C58-H6, detected by antitetrahistidine, changed from mostly nuclear (panel b) to mostly cytoplasmic (panels d and f), as did that of DM2 and DM14 Vp1s detected by anti-Vp1 (panels c and e). This result shows that the aberrant localization of DM2 and DM14 Vp1s extends to the coexpressed Vp1C58-H6, rather than Vp1C58-H6 rescuing the nuclear localization of DM2 and DM14 Vp1. Hence, the C49A-C87A and C87A-C254A double-cysteine mutations exert a dominant interference on the Vp1 NLS function both in cis and in trans, demonstrating that they are dominant-negative Vp1 mutations with respect to the protein's nuclear localization. We interpret this phenomenon as a result of hetero-oligomer formation among the truncated wild-type Vp1C58-H6 and the DM2 or DM14 Vp1 synthesized in the same cells. The blocked nuclear entry of the hetero-oligomers could result because their NLSs are not displayed properly or because the aberrant structural features of the mutant Vp1 members, similarly to those of DM2 or DM14 Vp1 alone, can trigger an altered trafficking pattern of the protein complexes. A connection between aberrant folding and altered fate of a protein might be mediated by proteins such as Hsc70, a molecular chaperone that has been found to associate with Vp1 of murine polyomavirus during infection (3). Besides its role in cytosolic protein folding, Hsc70 also participates in the nuclear transport process (5, 14) and can interact with the E3 ubiquitin ligase CHIP, a component of the ubiquitin-proteasome system (see reference 12 and references therein)
In summary, our present Vp1 cysteine mutant study has revealed an essential role for three pairs of Vp1 cysteine residues in two stages of SV40 morphogenesis. First, Cys49-Cys87 and Cys87-Cys254 may be involved at the cytoplasmic stage of Vp1 folding and trafficking. Second, Cys87-Cys207 appears to function at the nuclear stage of virion formation. Our results, though not ruling out other means by which the cysteine residues may contribute to these viral processes, are consistent with the notion we have previously raised that transitory disulfide bonding between certain pairs of Vp1 cysteines is necessary for guiding proper Vp1 folding (10), as well as for guiding virion assembly. Whether redox reactions involving Cys49-Cys87 and Cys87-Cys254 pairs are responsible for the sequential formation of the intrachain and interchain disulfide-linked cytoplasmic Vp1 intermediates (10) remains to be studied, along with the possible involvement of enzymatic or chaperone machineries in such processes.
ACKNOWLEDGMENTS
We thank Mary A. Tran for assistance in plasmid construction and plaque assays.
This work was supported by Public Health Service grant CA50574 from the National Institutes of Health (NIH). P.P.L. was supported in part by an award from the University of California, Los Angeles (UCLA) Jonsson Comprehensive Cancer Center.
Present address: Department of Bioscience and Biotechnology, Tokyo Institute of Technology, Midori-ku, Yokohama, Kanagawa 226-8503, Japan.
REFERENCES
Bina, M., S. C. Ng, and V. Blasquez. 1983. Simian virus 40 chromatin interaction with the capsid proteins. J. Biomol. Struct. Dyn. 1:689-704.
Blasquez, V., S. Beecher, and M. Bina. 1983. Simian virus 40 morphogenetic pathway. An analysis of assembly-defective tsB201 DNA protein complexes. J. Biol. Chem. 258:8477-8484.
Cripe, T. P., S. E. Delos, P. A. Estes, and R. L. Garcea. 1995. In vivo and in vitro association of hsc70 with polyomavirus capsid proteins. J. Virol. 69:7807-7813.
Gharakhanian, E., C. L. Fasching, S. J. Orlando, and A. R. Perez. 2001. Cys(9), Cys(104) and Cys(207) of simian virus 40 Vp1 are essential for infectious virion formation in CV-1 cells. J. Gen. Virol. 82:1935-1939.
Imamoto, N., Y. Matsuoka, T. Kurihara, K. Kohno, M. Miyagi, F. Sakiyama, Y. Okada, S. Tsunasawa, and Y. Yoneda. 1992. Antibodies against 70-kD heat shock cognate protein inhibit mediated nuclear import of karyophilic proteins. J. Cell Biol. 119:1047-1061.
Ishii, N., A. Nakanishi, M. Yamada, M. H. Macalalad, and H. Kasamatsu. 1994. Functional complementation of nuclear targeting-defective mutants of simian virus 40 structural proteins. J. Virol. 68:8209-8216.
Kasamatsu, H., J. Woo, A. Nakamura, P. Müller, M. J. Tevethia, and R. C. Liddington, submitted for publication.
Li, P. P., A. Nakanishi, M. A. Tran, A. M. Salazar, R. C. Liddington, and H. Kasamatsu. 2000. Role of simian virus 40 Vp1 cysteines in virion infectivity. J. Virol. 74:11388-11393.
Li, P. P., A. Nakanishi, D. Shum, P. C. Sun, A. M. Salazar, C. F. Fernandez, S. W. Chan, and H. Kasamatsu. 2001. Simian virus 40 Vp1 DNA-binding domain is functionally separable from the overlapping nuclear localization signal and is required for effective virion formation and full viability. J. Virol. 75:7321-7329.
Li, P. P., A. Nakanishi, S. W. Clark, and H. Kasamatsu. 2002. Formation of transitory intrachain and interchain disulfide bonds accompanies the folding and oligomerization of simian virus 40 Vp1 in the cytoplasm. Proc. Natl. Acad. Sci. USA 99:1353-1358.
Liddington, R. C., Y. Yan, J. Moulai, R. Sahli, T. L. Benjamin, and S. C. Harrison. 1991. Structure of simian virus 40 at 3.8-? resolution. Nature 354:278-284.
Murata, S., T. Chiba, and K. Tanaka. 2003. CHIP: a quality-control E3 ligase collaborating with molecular chaperones. Int. J. Biochem. Cell Biol. 35:572-578.
Ng, S. C., and M. Bina. 1984. Temperature-sensitive BC mutants of simian virus 40: block in virion assembly and accumulation of capsid-chromatin complexes. J. Virol. 50:471-477.
Shi, Y., and J. O. Thomas. 1992. The transport of proteins into the nucleus requires the 70-kilodalton heat shock protein or its cytosolic cognate. Mol. Cell. Biol. 12:2186-2192.
Stehle, T., S. J. Gamblin, Y. Yan, and S. C. Harrison. 1996. The structure of simian virus 40 refined at 3.1 ? resolution. Structure 4:165-182.(Peggy P. Li, Akira Nakani)
ABSTRACT
Transient disulfide bonding occurs during the intracellular folding and pentamerization of simian virus 40 (SV40) major capsid protein Vp1 (P. P. Li, A. Nakanishi, S. W. Clark, and H. Kasamatsu, Proc. Natl. Acad. Sci. USA 99:1353-1358, 2002). We investigated the requirement for Vp1 cysteine pairs during SV40 infection. Our analysis identified three Vp1 double-cysteine mutant combinations that abolished viability as assayed by plaque formation. Mutating the Cys49-Cys87 pair or the Cys87-Cys254 pair led to ineffective nuclear localization and diminished accumulation of the mutant Vp1s, and the defect extended in a dominant-negative manner to the wild-type minor capsid proteins Vp2/3 and an affinity-tagged recombinant Vp1 expressed in the same cells. Mutating the Cys87-Cys207 pair preserved the nuclear localization and normal accumulation of the capsid proteins but diminished the production of virus-like particles. Our results are consistent with a role for Cys49, Cys87, and Cys254 in the folding and cytoplasmic-nuclear trafficking of Vp1 and with a role for Cys87 and Cys207 in the assembly of infectious particles. These findings suggest that transient disulfide bond formation between certain Vp1 cysteine residues functions at two stages of SV40 infection: during Vp1 folding and oligomerization in the cytoplasm and during virion assembly in the nucleus.
TEXT
During simian virus 40 (SV40) infection, the newly synthesized major capsid protein Vp1 forms transient intrachain and interchain disulfide linkages in the cytoplasm and is transported to the nucleus (10). We have hypothesized that the disulfide bonds serve to facilitate the orderly interdigitation of Vp1 secondary structural elements, thereby facilitating the folding of the Vp1 monomer and the formation of the Vp1 pentamer (10). At least one type of cysteine pair, consisting of two different Vp1 cysteine residues, is required for forming the disulfide-linked monomer and oligomers observed (10). Neither the identity of the cysteine pair(s) nor their necessity during infection is yet known. Individual mutations of the seven Vp1 cysteines largely preserve viral viability (8), raising the possibility that the cysteines may redundantly function in a redox-coupled Vp1 folding and oligomerization pathway. In SV40 virion, disulfides are not present within individual Vp1 pentamers; the distance between any two cysteine sulfur atoms on a pentamer is at least 10 ?, much longer than the typical bond lengths of disulfides (11). Cys104-Cys104 disulfide bonds are observed between some Vp1 pentamers on SV40 (11, 15), though we have shown previously that they are not important for viral infectivity (8). To identify candidate Vp1 cysteine pairs for the transient disulfide bonding and determine their importance in infection, the cysteine residues were mutagenized in pairs and in larger combinations. All mutant combinations were first created in the pBS-Vp1 plasmid (8) via fragment exchanges among the single-cysteine mutant pBS-Vp1s (8) before the entire Vp1 coding region of the resulting plasmids was inserted into the nonoverlapping SV40-harboring plasmid, NO-pSV40 (6), via XbaI and SacI sites. Then, mutant viral genomes in the form of NO-SV40 were prepared from respective NO-pSV40 DNAs via BamHI digestion and recircularization, as described previously (6).
Three Vp1 double-cysteine mutants, C49A-C87A, C87A-C207S, and C87A-C254A, are nonviable. All possible Vp1 double-cysteine mutants and select multiple-cysteine mutants of NO-SV40 were examined for overall viability by the plaque formation assay, using lysates of the mutant genome-transfected cells. Among the total 21 double mutants (Fig. 1, DM series), three categories of plaque-forming ability can be seen. First, a majority of 15 double mutants had infectious titers that were reduced no more than 10-fold from the wild-type value of 1.5 x 108 PFU per U of transfected lysate. Nine of these mutants, DM1, DM6, DM8, DM10, DM12, DM17, DM19, DM20, and DM21, had average plaque diameters of at least 3.2 mm (classified as large in Fig. 1), and the remaining six, DM3, DM4, DM5, DM7, DM15, and DM16, had average plaque diameters of at least 1.5 mm (classified as medium). Note that all pairwise mutant combinations involving Cys9 or Cys104 are in this rather viable category. The triple mutant C9A-C104A-C207S is similarly viable (MM4 in Fig. 1; see below). Hence, Cys9 and Cys104, are the Vp1 cysteine residues least important for viability. This result contrasts with the finding of Gharakhanian et al. (4) that the simultaneous mutation of cysteines 9 and 104, 9 and 207, or 9, 104, and 207 in SV40, in which a part of the Vp1 coding sequence overlaps with that of Vp2/3, abolishes plaque formation. We believe this difference arises from the concomitant Vp2/3 codon change when the Vp1 Cys9 codon is mutated in the context of SV40, a complication avoided by the use of the nonoverlapping NO-SV40 genome in our study.
In the second category, three double mutants, DM2 (C49A-C87A), DM9 (C87A-C207S), and DM14 (C87A-C254A), were nonviable, having no measurable plaque-forming ability (Fig. 1). These three cysteine pairs, involving Cys87 in combination with Cys49, Cys207, or Cys254, are thus essential for the formation of infectious virions in vivo. These four Vp1 cysteines might engage in transient disulfide bonding at some stages of infectious virion formation (see further analysis of the double mutants below) and be partially redundant in this function.
Third, three double mutants, DM11 (C207S-C267L), DM13 (C49A-C207S), and DM18 (C87A-C267L), had infectious titers that are reduced by 4 to 5 logs and average plaque sizes in the medium range (Fig. 1). The greatly diminished viabilities of these mutants suggest that Cys49 and Cys207 can partially substitute for each other and that Cys267 can partially substitute for Cys87 or Cys207. That is, Cys267 may partake in an auxiliary function in infectious virion formation besides the four cysteines identified above. If so, one would expect that mutating Cys267 together with Cys49 and Cys207 would diminish viability much more than the individual double mutations of DM11 and DM13. This possibility is confirmed by the multiple-cysteine mutant analysis described below.
Among the 11 multiple-cysteine mutants (MM series), there are also three categories of plaque-forming efficiencies. The first group consists of four triple mutants, MM2 through MM5, which had PFU values 1 to 2 logs reduced from that of the wild type and average plaque diameters in the medium range (Fig. 1). That all four mutants were rather viable despite being mutated at both Cys9 and Cys104 confirms that these two cysteines are unimportant. The second group are six multiple-cysteine mutants that were unable to form plaques, including MM1, MM6, MM7, MM8, MM10, and MM11 (Fig. 1). The nonviability of MM1 (mutated at residues 9, 49, and 87), MM6 (mutated at 9, 49, 87, and 104), MM8 (mutated at 9, 87, 104, 207, and 267), MM10 (mutated at 9, 87, 104, 254, and 267), and MM11 (all cysteines mutated) can be attributed to the fact that they all harbor at least one nonviable double-cysteine mutant combination of DM2, DM9, or DM14. Though MM7 (mutated at residues 9, 49, 104, 207, and 267) lacks a lethal double mutation (Fig. 1), it too was nonviable, consistent with the prediction above that a triple mutation of Cys49, Cys207, and Cys267 would be highly detrimental.
In the third category of multiple mutants is the quintuple-mutant MM9, mutated at residues 9, 104, 207, 254, and 267. This mutant had a 5-log-reduced PFU value compared to the wild type, similar to DM11 (C207S-C267L), but had a much smaller average plaque diameter of 0.8 mm (Fig. 1). The greatly diminished viability of MM9 suggests that the remaining Cys49 and Cys87 residues can support SV40 infection to a limited extent, perhaps via forming transient disulfide bonds with each during Vp1 folding and assembly. Cys207, Cys254, and sometimes Cys267 appear to additionally contribute to these processes.
The nonviable Vp1 double-cysteine mutants fall into two phenotypic classes. To further characterize the three nonviable cysteine-pair mutants, DM2, DM9, and DM14, we examined their ability to replicate viral DNA and produce capsid proteins. The extent of DNA replication was normal for all mutants. The amounts of DpnI-resistant viral DNA extracted from all mutant DNA-transfected cells (Fig. 2B, lanes 2, 3, and 4) were comparable to that extracted from the wild type-transfected cells (Fig. 2B, lane 1). On the other hand, the intracellular levels of Vp1 and Vp2/3 were substantially lower for DM2 (Fig. 2A, lanes 3 and 4) and DM14 (lanes 5 and 6) than for the wild type (lanes 1 and 2) or for DM9 (lanes 7 and 8).
The double mutants also differed in the subcellular localization of the capsid proteins expressed from the mutant genomes. Mutant DM9-expressed Vp1 and Vp2/3 largely localized in the nucleus (Fig. 3A, panels g and h), as did the capsid proteins expressed from the wild-type genome (panels a and b). In contrast, mutants DM2 and DM14 had a striking defect in the nuclear localization of the capsid proteins. In cells into which these mutant genomes were introduced, the mutant Vp1s were largely distributed in the cytoplasm as irregular specks, though some were distributed in the nucleus either diffusely or in clumps or granules (Fig. 3A, panels c and e). Although the mutants encode wild-type Vp2/3, Vp2/3 exhibited a localization pattern indistinguishable from that of the coexpressed mutant Vp1s (panels d and f). This colocalization suggests that the mutant Vp1s exert a dominant-negative influence on Vp2/3. Therefore, the differences in capsid protein localization and accumulation divide the nonviable double-cysteine mutants into two classes: C87A-C207S (DM9), whose defect occurs during virion assembly in the nucleus; and C49A-C87A (DM2) and C87A-C254A (DM14), whose defect occurs during the trafficking and possibly folding of the capsid proteins in the cytoplasm.
Mutant C87A-C207S is defective in the nuclear phase of virion assembly. Given that mutant DM9 could replicate the viral genome and accumulate a substantial level of the capsid proteins that localize to the nucleus, we asked whether the DM9 capsid proteins could package viral genome into virion-like particles (VLPs). Lysates were prepared from wild-type and mutant genome-transfected cells and treated with DNase I. (We noted that it was difficult to detect the VLP peak in sucrose gradient without the nuclease treatment.) The DNase I-resistant fraction, expected to contain VLPs but not assembly intermediates, was then sedimented through a sucrose gradient. Whereas the wild-type sample exhibited abundant cosedimented viral DNA and Vp1 (Fig. 2C, fractions 1 through 10) around the expected position for virions (fraction 5), the DM9 sample gave less viral DNA in the same region of the gradient (Fig. 2D), even though twice as much transfected material was analyzed. Much of the mutant Vp1 did not cosediment with the DNA but was distributed heterogeneously in fractions 8 through 16. This sedimentation profile suggests the formation of heterogeneous assembly intermediates (75S to 160S), in which the packaging of minichromosomes has been initiated but virion particles are not formed. In SV40 capsid structure, Cys87 is located on the BC2 loop, and Cys207 resides in the E''' region of the EF loop (15). These structural elements are part of Vp1 structural region I, to which temperature-sensitive (ts) B mutations have recently been mapped (7). The phenotype of these ts mutants is the accumulation of 100S to 160S assembly intermediates (1, 2, 13). The similarity of DM9 to these ts mutants, in terms of the structural location of the mutations and the type of assembly intermediates accumulated (Fig. 2D), suggests that the Cys87-Cys207 pair may similarly function in the addition of Vp1 pentamers to the initiated viral minichromosome. Thus, the defect of mutant C87A-C207S consists of a deficiency in the assembly of virus-like particles coupled to a lack of infectivity of any particles that are formed. Further studies are needed to elucidate the role of this cysteine pair in virion assembly.
Mutant C49A-C87A and C87A-C254A Vp1s have a dominant-negative effect on the nuclear localization of Vp2/3. The subcellular distribution patterns of mutant DM2 and DM14 capsid proteins (Fig. 3A) show that these capsid proteins are ineffective in nuclear localization despite the intact nuclear localization signals (NLSs) harbored by both the mutant Vp1s and the wild-type Vp2/3. That the coexpressed wild-type Vp2/3 cannot rescue the nuclear localization of mutant DM2 and DM14 Vp1s is in contrast to the functional complementation observed for some mutants of the Vp1 NLS and DNA-binding domain, in which the wild-type Vp2/3 can effectively piggyback the mutant Vp1s to the nucleus (6, 9). The Vp1 double mutations C49A-C87A and C87A-C254A apparently override the intact NLSs not only of mutant Vp1s but also of the wild-type Vp2/3 in a dominant-negative manner. Our preliminary evidence shows that DM2 Vp1 coimmunoprecipitated with Vp2/3 from NO-SV40-DM2-transfected cell lysate (P. P. Li and H. Kasamatsu, unpublished data). Thus, DM2 mutant Vp1 appears to form a complex with Vp2/3 in the cytoplasm. We speculate that DM2 and DM14 mutant Vp1s fold into somewhat different structures than does wild-type Vp1. This altered folding might either mask the NLSs of Vp1 and associated Vp2/3 or cause the trafficking of these viral proteins to be altered, leading to their trapping and accelerated degradation in the cytoplasm. As might be expected from the blocked nuclear entry of the capsid proteins, mutant DM2 failed to form VLPs: no post-DNase I viral DNA was detected in the sucrose fractions, even though a large amount of the mutant transfected lysate was sedimented (Fig. 2E).
Mutant C49A-C87A and C87A-C254A Vp1s are dominant-negative to wild-type Vp1. To further explore the dominant interfering nature of DM2 and DM14 Vp1 mutations in the cytoplasm, a derivative of NO-SV40, NO-SV40-Vp1C58-H6, was used. It encodes a truncated Vp1 in which the carboxy-terminal 58 amino acids are replaced by a (Gly-Gly-Gly-Gly-Ser)3 flexible linker, followed by a His6 affinity tag. We believe that this truncated Vp1 forms pentamers, since a similar recombinant Vp1 minus the flexible linker was previously expressed in Escherichia coli and purified as a pentamer (9). The construction of NO-pSV40-Vp1C58-H6 proceeded through a series of intermediate plasmids as follows. A PCR fragment generated by using pBS-Vp1C58 (8) as a template, 5'-GTTTACCAACACTAGTGGAAC as a sense primer, and 5'-AAATGATGGGATCCACCAAAGCTAGCTGGGCCGGGGTTTTTCACAGACCGCTTTC as an antisense primer was inserted between the SpeI and BamHI sites (underlined above) of pBS-Vp1C58, producing pBS-Vp1C58-NheI (NheI site in italics above). A fragment encoding the flexible linker (sense strand, 5'-GGTGGAGGCGGTTCAGGCGGAGGTGGCTCTGGCGGTGGCGGATCT), the polyhistidine tag (sense strand, 5'-CATCACCATCACCATCAC), and a stop codon (TAA) was then inserted between the NheI and BamHI sites of pBS-Vp1C58-NheI, producing pBS-Vp1C58-linker-H6. Replacing the SpeI-to-SacI fragment of NO-SV40-BSM (8) with that of pBS-Vp1C58-linker-H6 resulted in NO-pSV40-Vp1C58-H6, which was digested with BamHI and recircularized to yield NO-SV40-Vp1C58-H6. This DNA expressed Vp1C58-H6, which effectively localized in the nucleus, similar to full-length Vp1, as detected by both anti-Vp1 and antitetrahistidine antibodies (Fig. 3B, panels a and b).
When Vp1C58-H6 was coexpressed with DM2 Vp1 (Fig. 3B, panels c and d) or with DM14 Vp1 (panels e and f), derived from respective mutant NO-SV40s, the localization of Vp1C58-H6, detected by antitetrahistidine, changed from mostly nuclear (panel b) to mostly cytoplasmic (panels d and f), as did that of DM2 and DM14 Vp1s detected by anti-Vp1 (panels c and e). This result shows that the aberrant localization of DM2 and DM14 Vp1s extends to the coexpressed Vp1C58-H6, rather than Vp1C58-H6 rescuing the nuclear localization of DM2 and DM14 Vp1. Hence, the C49A-C87A and C87A-C254A double-cysteine mutations exert a dominant interference on the Vp1 NLS function both in cis and in trans, demonstrating that they are dominant-negative Vp1 mutations with respect to the protein's nuclear localization. We interpret this phenomenon as a result of hetero-oligomer formation among the truncated wild-type Vp1C58-H6 and the DM2 or DM14 Vp1 synthesized in the same cells. The blocked nuclear entry of the hetero-oligomers could result because their NLSs are not displayed properly or because the aberrant structural features of the mutant Vp1 members, similarly to those of DM2 or DM14 Vp1 alone, can trigger an altered trafficking pattern of the protein complexes. A connection between aberrant folding and altered fate of a protein might be mediated by proteins such as Hsc70, a molecular chaperone that has been found to associate with Vp1 of murine polyomavirus during infection (3). Besides its role in cytosolic protein folding, Hsc70 also participates in the nuclear transport process (5, 14) and can interact with the E3 ubiquitin ligase CHIP, a component of the ubiquitin-proteasome system (see reference 12 and references therein)
In summary, our present Vp1 cysteine mutant study has revealed an essential role for three pairs of Vp1 cysteine residues in two stages of SV40 morphogenesis. First, Cys49-Cys87 and Cys87-Cys254 may be involved at the cytoplasmic stage of Vp1 folding and trafficking. Second, Cys87-Cys207 appears to function at the nuclear stage of virion formation. Our results, though not ruling out other means by which the cysteine residues may contribute to these viral processes, are consistent with the notion we have previously raised that transitory disulfide bonding between certain pairs of Vp1 cysteines is necessary for guiding proper Vp1 folding (10), as well as for guiding virion assembly. Whether redox reactions involving Cys49-Cys87 and Cys87-Cys254 pairs are responsible for the sequential formation of the intrachain and interchain disulfide-linked cytoplasmic Vp1 intermediates (10) remains to be studied, along with the possible involvement of enzymatic or chaperone machineries in such processes.
ACKNOWLEDGMENTS
We thank Mary A. Tran for assistance in plasmid construction and plaque assays.
This work was supported by Public Health Service grant CA50574 from the National Institutes of Health (NIH). P.P.L. was supported in part by an award from the University of California, Los Angeles (UCLA) Jonsson Comprehensive Cancer Center.
Present address: Department of Bioscience and Biotechnology, Tokyo Institute of Technology, Midori-ku, Yokohama, Kanagawa 226-8503, Japan.
REFERENCES
Bina, M., S. C. Ng, and V. Blasquez. 1983. Simian virus 40 chromatin interaction with the capsid proteins. J. Biomol. Struct. Dyn. 1:689-704.
Blasquez, V., S. Beecher, and M. Bina. 1983. Simian virus 40 morphogenetic pathway. An analysis of assembly-defective tsB201 DNA protein complexes. J. Biol. Chem. 258:8477-8484.
Cripe, T. P., S. E. Delos, P. A. Estes, and R. L. Garcea. 1995. In vivo and in vitro association of hsc70 with polyomavirus capsid proteins. J. Virol. 69:7807-7813.
Gharakhanian, E., C. L. Fasching, S. J. Orlando, and A. R. Perez. 2001. Cys(9), Cys(104) and Cys(207) of simian virus 40 Vp1 are essential for infectious virion formation in CV-1 cells. J. Gen. Virol. 82:1935-1939.
Imamoto, N., Y. Matsuoka, T. Kurihara, K. Kohno, M. Miyagi, F. Sakiyama, Y. Okada, S. Tsunasawa, and Y. Yoneda. 1992. Antibodies against 70-kD heat shock cognate protein inhibit mediated nuclear import of karyophilic proteins. J. Cell Biol. 119:1047-1061.
Ishii, N., A. Nakanishi, M. Yamada, M. H. Macalalad, and H. Kasamatsu. 1994. Functional complementation of nuclear targeting-defective mutants of simian virus 40 structural proteins. J. Virol. 68:8209-8216.
Kasamatsu, H., J. Woo, A. Nakamura, P. Müller, M. J. Tevethia, and R. C. Liddington, submitted for publication.
Li, P. P., A. Nakanishi, M. A. Tran, A. M. Salazar, R. C. Liddington, and H. Kasamatsu. 2000. Role of simian virus 40 Vp1 cysteines in virion infectivity. J. Virol. 74:11388-11393.
Li, P. P., A. Nakanishi, D. Shum, P. C. Sun, A. M. Salazar, C. F. Fernandez, S. W. Chan, and H. Kasamatsu. 2001. Simian virus 40 Vp1 DNA-binding domain is functionally separable from the overlapping nuclear localization signal and is required for effective virion formation and full viability. J. Virol. 75:7321-7329.
Li, P. P., A. Nakanishi, S. W. Clark, and H. Kasamatsu. 2002. Formation of transitory intrachain and interchain disulfide bonds accompanies the folding and oligomerization of simian virus 40 Vp1 in the cytoplasm. Proc. Natl. Acad. Sci. USA 99:1353-1358.
Liddington, R. C., Y. Yan, J. Moulai, R. Sahli, T. L. Benjamin, and S. C. Harrison. 1991. Structure of simian virus 40 at 3.8-? resolution. Nature 354:278-284.
Murata, S., T. Chiba, and K. Tanaka. 2003. CHIP: a quality-control E3 ligase collaborating with molecular chaperones. Int. J. Biochem. Cell Biol. 35:572-578.
Ng, S. C., and M. Bina. 1984. Temperature-sensitive BC mutants of simian virus 40: block in virion assembly and accumulation of capsid-chromatin complexes. J. Virol. 50:471-477.
Shi, Y., and J. O. Thomas. 1992. The transport of proteins into the nucleus requires the 70-kilodalton heat shock protein or its cytosolic cognate. Mol. Cell. Biol. 12:2186-2192.
Stehle, T., S. J. Gamblin, Y. Yan, and S. C. Harrison. 1996. The structure of simian virus 40 refined at 3.1 ? resolution. Structure 4:165-182.(Peggy P. Li, Akira Nakani)