Interactions among Four Proteins Encoded by the Human Cytomegalovirus UL112-113 Region Regulate Their Intranuclear Targeting and the Recruitment of UL44 to Prereplication Foci

Cell culture and virus infection.

HF, Vero, and 293T cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. The HCMV(Towne) virus stocks used in this study were prepared as previously described (12). In order to initiate infection, subconfluent HF cells were infected with HCMV at a multiplicity of infection (MOI) of 2.0 PFU per cell. The input supernatant virus was then adsorbed for 1.5 h at 37°C, and the inoculum was replaced with fresh warm medium at time zero.

Plasmid construction.

The pΔ60SV, pΔ6SV, and pΔ45SV plasmids, which contain AD169 UL112-113 cDNAs encoding 43-, 50-, and 84-kDa proteins, respectively (23), were kindly provided by Deborah H. Spector (University of California, San Diego, Calif.). The cDNA encoding the 34-kDa protein was amplified by PCR (top primer, 5′ GAAGATCTATGGATCTCCCTACTACCGTC 3′; bottom primer, 5′ GAAGATCTTCACGAGTCCACTCGGAAAGC 3′) from pRTS26, which harbors the AD169 UL112-113 genomic gene (19), and cloned into the pENTR vector (Invitrogen). The cDNAs for the 43-, 50-, and 84-kDa proteins were amplified by PCR (top primer, 5′ GAAGATCTATGGATCTCCCTACTACCGTC 3′; bottom primer, 5′ GAAGATCTTTAATCGTCGAAAAACGCCGCGAT 3) from the pΔ60SV, pΔ6SV, and pΔ45SV plasmids, respectively, and cloned into the pENTR vector. Expression plasmids encoding 5′ hemagglutinin (HA)- or Flag-tagged 34-, 43-, 50-, and 84-kDa proteins were then generated in a background of pSG5 (7), using Gateway technology (Invitrogen).

Two large truncation mutants were produced in the background of the 84-kDa protein. The 84(ΔExon-1) mutant, harboring a deletion of the N-terminal 252 amino acids (exon 1), was amplified by PCR (top primer, 5′ GAAGATCTATGGATCTCCCTACTACCGTC 3′; bottom primer, 5′ GAAGATCTTTAATCGTCGAAAAACGCCGCGAT 3′) and placed into pENTR. The 84(ΔC) mutant has a stop codon (UGA) at amino acid position 355, and the mutation was generated by PCR (top primer, 5′ CTACCCATTGAGCGCTGAGCGGTGGTTTCGTCG 3′; bottom primer, 5′ CGACGAAACCACCGCTCAGCGCTCAATGGGTAG 3′) according to the Stratagene QuickChange site-directed mutagenesis protocol. In addition, a series of small deletion mutants, each harboring a 25-amino-acid deletion within exon 1, were also generated by PCR in the same 84-kDa protein background. Expression plasmids encoding 5′ HA- or Flag-tagged mutant variants were generated in pSG5 as described above. In order to construct template plasmids for in vitro transcription/translation reactions, we cloned the wild-type or mutant cDNAs into the pSPUTK (without a tag) (Stratagene) or pCS3-MT (with a six-Myc tag) plasmid (18), using Gateway technology. pCS3-MT-driven plasmids were also employed to express proteins in the cultured cells via transfection.

The pJHA309 plasmid, which expresses 5′ Flag-tagged UL112-113 proteins together from the genomic gene, has been previously described (2). To disrupt the reported nuclear localization signal (NLS) (13) in the pJHA309 background, we replaced the Lys-Arg-Gln-Lys residues, which correspond to positions 261 to 264 of the 43-, 50-, and 84-kDa proteins, with His-Met residues using PCR mutagenesis, resulting in pJHA331. The Δ1-25 and ΔC mutations in the 84-kDa protein background were also introduced into a plasmid which harbors the HA-tagged genomic gene (pRYK21), and this resulted in the formation of pMY69 encoding HA-UL112-113(Δ1-25) and pRYK60 encoding HA-UL112-113(ΔC).

The pSP38 plasmid harboring the HCMV replication origin DNA and plasmids which express six core replication proteins (UL54, UL44, UL57, UL105, UL70, and UL102) and auxiliary proteins (UL36-38, IE1/IE2, and UL84) have been previously described (19) and were provided by Gary S. Hayward (Johns Hopkins University School of Medicine, Baltimore, MD).

Transient DNA transfection.

With HF cells, DNA transfection was conducted using Lipofectamine 2000 reagents (Invitrogen). 293T cells were transfected via the N,N-bis-(2-hydroxyethyl)-2-aminoethanesulfonic acid-buffered saline (BBS) version of the calcium phosphate method, as described previously (12). With Vero cells, the DNA mixtures were introduced into the cells using FuGene 6 reagents (Roche).

Antibodies and indirect immunofluorescence assay (IFA).

A rabbit antipeptide polyclonal antibody (PAb) specific to the 84-kDa form of the UL112-113 protein was described previously (2). Mouse monoclonal antibody (MAb) M23, which recognizes the shared exon 1 of all four UL112-113 proteins (10), was kindly provided by Kanji Hirai (Tokyo Medical and Dental University, Tokyo, Japan). The anti-HA rat MAb 3F10, either conjugated with peroxidase or labeled with fluorescein, and the anti-Myc mouse MAb 9E10 were purchased from Roche. The anti-Flag mouse MAb M2 was obtained from Sigma. The anti-UL44 (p52) mouse MAb was obtained from Advanced Biotechnologies, Inc. The rabbit anti-PML PAb, referred to as PML(C), has been previously described (1).

For IFA, the cells were fixed with 1% paraformaldehyde and permeabilized with 0.2% Triton X-100. All of the subsequent procedures performed have been previously described (12). All slides were examined and photographed with a Zeiss Axiophot microscope.

In vitro transcription/translation and cross-linking assays.

The UL112-113 proteins were synthesized in vitro by using the TNT Quick Coupled transcription/translation system (Promega). For cross-linking assays, 5-μl in vitro-translated protein samples were diluted with 95 μl of 10 mM potassium phosphate buffer (pH 8.0). The mixtures were then incubated in 0.0001% glutaraldehyde (Sigma) for 30 min at room temperature. Fifteen microliters of each sample was then boiled for 5 min and separated by sodium dodecyl sulfate-8% polyacrylamide gel electrophoresis (SDS-8% PAGE). The gels were dried, and the protein bands were detected by autoradiography.

Coimmunoprecipitation assays.

DNA-transfected cells were harvested and sonicated in coimmunoprecipitation buffer (50 mM Tris-Cl, pH 7.4, 50 mM NaF, 5 mM sodium phosphate, and 0.1% Triton X-100) supplemented with protease inhibitor cocktail (Sigma). Cell lysates or proteins synthesized by coupled in vitro transcription/translation reactions were then incubated with the appropriate Ab. After 2 h of incubation at 4°C, 30 μl of a 50% slurry of protein A- and G-Sepharose (Amersham) was added and adsorbed for 16 h at 4°C. The mixture was pelleted and washed six times with coimmunoprecipitation buffer. The beads were then resuspended and boiled for 5 min in loading buffer. The samples were analyzed by SDS-PAGE and immunoblotting when necessary.

Metabolic labeling and immunoprecipitation.

HF cells (1 × 10⁷ cells) in 150-mm tissue culture dishes were infected with HCMV(Towne) at an MOI of 2.0. For [³⁵S]methionine-cysteine labeling, the cells were washed with phosphate-buffered saline at 48 h and incubated for 30 min with methionine- and cysteine-free Dulbecco's modified Eagle's medium supplemented with 5% dialyzed fetal bovine serum. The medium was then replaced with fresh medium containing 300 μCi of [³⁵S]methionine-cysteine (14.3 mCi/ml) (Promix; Amersham). After 4 h of incubation at 37°C, the cells were harvested and lysed with RIPA buffer (1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 10 mM Tris-HCl [pH 7.4]) for 30 min on ice. The cell lysates were then utilized in immunoprecipitation experiments.

Immunoblot analysis.

DNA-transfected cells were washed with phosphate-buffered saline, and the total extracts were prepared by incubation of the cells with RIPA buffer. The clarified cell extracts were separated in an SDS-8% polyacrylamide gel, followed by performance of the standard procedure using an enhanced chemiluminescence system (Amersham).

The UL112-113 proteins both self-interact and interact with each other via the common N-terminal region in vitro.

To investigate interactions among the four proteins encoded by the UL112-113 region of the HCMV genome, we generated expression plasmids encoding the 34-, 43-, 50-, and 84-kDa proteins with N-terminal Flag tags (Fig. 1A). The expression of these proteins was verified by immunoblot analysis of the cell extracts prepared from transfected Vero cells (Fig. 1B). Notably, the cDNA for the 84-kDa protein also expressed a small amount of the 43-kDa protein. We also generated ³⁵S-labeled 34-, 43-, 50-, and 84-kDa proteins via coupled in vitro transcription and translation. When the individually translated proteins were incubated with glutaraldehyde, cross-linked dimeric variants were detected for all of the 34-, 43-, 50-, and 84-kDa proteins (Fig. 2A). Similar cross-linking assays were conducted with two mutant variants of the 84-kDa protein. The ΔExon-1 mutant, which is devoid of the 252 amino acids at the N terminus, and the ΔC mutant, which lacks the C-terminal region encompassing amino acids 348 to 684, were both synthesized with N-terminal six-Myc tags (Fig. 2B). Interestingly, the ΔExon-1 mutant failed to form dimers, whereas the ΔC mutant continued to generate dimers (Fig. 2A). This clearly suggests that the UL112-113 proteins may self-interact via their common 252 N-terminal amino acids.

We then attempted to determine whether the different UL112-113 proteins were capable of interacting with each other. The 34-, 43-, 50-, and 84-kDa proteins were cotranslated with the 84-kDa protein. When the ³⁵S-labeled proteins were immunoprecipitated with a rabbit PAb which recognizes only the 84-kDa form of the UL112-113 proteins, we observed coprecipitation of the 34-, 43-, and 50-kDa proteins (Fig. 2C). This also suggests that the 84-kDa UL112-113 protein may interact with other 34-, 43-, and 50-kDa UL112-113 proteins.

Interactions among four UL112-113 proteins in vivo.

The interactions occurring among the different UL112-113 proteins were investigated further by using coimmunoprecipitation assays with DNA-cotransfected cells (Fig. 3). When the Myc-tagged 34-, 43-, 50-, and 84-kDa proteins were coexpressed simultaneously with the HA-tagged 50-kDa protein in 293T cells, immunoprecipitation of the cell extracts with anti-Myc Ab resulted in coprecipitation of the HA-tagged 50-kDa protein. When the myc-84(ΔExon-1) and myc-84(ΔC) mutants were employed, the ΔExon-1 mutant was not coprecipitated with the HA-tagged 50-kDa protein, although the ΔC mutant continued to be coprecipitated with it. These results showed that four different UL112-113 proteins could interact with one another via their common N-terminal 252-amino-acid regions (exon 1). In a control experiment, the IE1 protein was determined not to interact with the HA-tagged 50-kDa protein. To exclude the possibility that the lack of interaction of the ΔExon-1 mutant with the 50-kDa protein was attributable to its cytoplasmic retention, we assessed the cellular localization of the ΔExon-1 mutant by IFA. We found that although some ΔExon-1 mutants were distributed in the cytoplasm, most of the ΔExon-1 mutants continued to exhibit nuclear localization (data not shown).

To confirm whether or not these interactions between the different UL112-113 proteins occur in virus-infected cells, the proteins of HCMV-infected HF cells were labeled in vivo with [³⁵S]Met-Cys, followed by immunoprecipitation with an 84-kDa form-specific Ab. In order to identify the UL112-113 proteins, we also conducted immunoprecipitation with the mouse MAb M23, which was reported to detect all four of the UL112-113 proteins together (10). We determined that immunoprecipitation with an anti-84-kDa-protein Ab coprecipitated the 43-kDa protein and also probably the 34-kDa protein (Fig. 4). The 50-kDa UL112-113 protein, which might be coprecipitated with the 84-kDa protein, could not be identified. This was attributed to interference of the immunoglobulin G (IgG) heavy chains. This result generally appears to support the idea that the four UL112-113 proteins are able to interact with one another and may manifest as a complex in virus-infected cells. Note that a protein with a molecular mass of about 40 kDa could be coprecipitated efficiently with the anti-84-kDa-protein Ab but not with the M23 Ab (Fig. 4, arrowhead).

The highly conserved N-terminal 125-amino-acid region of the 84-kDa protein is required for both self-interaction and nuclear localization as foci.

The sequence of the N-terminal 252 amino acids (exon 1) encoded by HCMV UL112 appears to be well conserved among betaherpesviruses, including chimpanzee CMV, rat CMV, mouse CMV, human herpesvirus 6 (HHV-6), and HHV-7 (Fig. 5). Interestingly, the N-terminal region of exon 1 exhibits a higher degree of homology than does the C-terminal portion.

To map domains within exon 1 required for self-interaction, we constructed a series of mutant 84-kDa variants, all containing a 25-amino-acid deletion within the exon 1 region. The untagged 84-kDa protein and the Myc-tagged wild-type or deletion mutant proteins were then cotranslated in vitro and were immunoprecipitated with anti-Myc Ab, with a subsequent immunoblot analysis with an anti-84-kDa-protein Ab (Fig. 6). The results indicated that like the ΔExon-1 mutant, the small deletion mutants Δ1-25, Δ26-50, Δ51-75, Δ76-100, and Δ101-125 did not bind to the intact 84-kDa protein, whereas the mutants Δ151-175, Δ201-225, and Δ226-250 continued to bind to the intact 84-kDa protein as efficiently as was observed for the wild-type protein. Our results show that the self-interaction of the 84-kDa protein requires the highly conserved 125 amino acids located in the N-terminal region.

The UL112-113 proteins were shown to be distributed in nuclear punctate forms (2). To determine whether self-interaction of the 84-kDa protein is required for its localization in nuclear foci, we expressed the above mutant proteins in Vero cells and characterized their localization patterns by IFA. Our results indicated that the mutant proteins containing 25-amino-acid deletions within the highly conserved N-terminal portion of exon 1, from positions 1 to 125 (Δ1-25, Δ26-50, Δ51-75, Δ76-100, and Δ101-125), were localized as uniform nuclear diffuse forms, whereas the mutant proteins harboring deletions at positions 126 to 250 (Δ151-175, Δ176-200, Δ201-225, and Δ226-250) yielded the typical localization pattern of the intact 84-kDa protein (Fig. 7A). The expression levels of the wild-type and mutant proteins in transfected Vero cells proved to be comparable to the results of immunoblotting analysis (Fig. 7B). This demonstrates that the highly conserved N-terminal region, located from positions 1 to 125 of exon 1 and essential for self-interaction, is also required for the formation of nuclear foci from the 84-kDa protein.

Differential localization of four UL112-113 proteins.

The localization pattern of each of the UL112-113 proteins was further examined in transfected HF cells by IFA. When the Flag-tagged UL112-113 proteins were individually expressed in HF cells, the 34-kDa protein was determined to be predominantly distributed in the cytoplasm, although some aggregates were also located in the nucleus. Double labeling with PML indicated that the nuclear forms of the 34-kDa protein were not associated with the PODs (Fig. 8A to D). The 43-kDa protein was distributed in nuclear punctate forms with some diffuse nuclear background staining, whereas the 50-kDa protein was found to be localized only as nuclear punctate forms (Fig. 8E and F). The nuclear punctate forms of these proteins partially overlapped with the PODs. The localization pattern of the 84-kDa protein was similar to that of the 43-kDa protein, but its nuclear foci appeared to overlap the PODs more efficiently than did those of the 43- and 50-kDa proteins in transfected cells (Fig. 8G). Considering that a small amount of the 43-kDa protein was also expressed from the cDNA for the 84-kDa protein (Fig. 1 and 3), this efficient POD targeting may also be explained by the coexpression of both the 84-kDa and 43-kDa proteins.

An expression plasmid containing the UL112-113 genomic gene expressing an N-terminal Flag tag was previously described and shown to encode four UL112-113 proteins simultaneously (2). When the cells received this plasmid, cell staining with anti-Flag Ab revealed that the UL112-113 proteins yielded a localization pattern reminiscent of that of the 84-kDa protein (Fig. 8H). Interestingly, no cytoplasmic form of the 34-kDa protein could be detected in these cells. This suggests that the simultaneous presence of four splice variants might have some influence on the localization patterns of individual UL112-113 proteins, thereby resulting in efficient targeting to the PODs.

A study involving the 43-kDa protein allowed for the mapping of an NLS to the region encompassing amino acid positions 245 to 272 and also revealed that the KRQK residues located from positions 261 to 264 are vitally important for the functioning of the NLS (13). Apparently, the failure of nuclear targeting by the 34-kDa protein in this study can be attributed to the absence of this reported NLS. When we deleted these KRQK residues within the background of the genomic gene expression plasmid, the mutant ΔNLS UL112-113 proteins were localized as cytoplasmic punctate forms (Fig. 8I), thereby verifying that these amino acids also perform an NLS function in both the 50- and 84-kDa proteins as well as the 43-kDa protein.

Colocalization interactions among the UL112-113 proteins.

Our results suggest that the simultaneous coexpression of the different UL112-113 proteins might affect the localization patterns of individual proteins. We further evaluated this possibility by using a series of cotransfection assays. When the HA-tagged 50-kDa protein and the Flag-tagged 34-kDa protein were expressed simultaneously in HF cells, both of the proteins were colocalized in a nuclear punctate pattern typical of the 50-kDa protein, thereby suggesting that the majority of the cytoplasmic forms of the 34-kDa protein had been relocalized to the nucleus in the presence of the 50-kDa protein (Fig. 8J to L). In a similar fashion, when the 50-kDa and 43-kDa proteins were coexpressed simultaneously, they were colocalized in a pattern reminiscent of that of the 50-kDa protein (Fig. 8M to O). Interestingly, when the 50-kDa and 84-kDa proteins were coexpressed simultaneously, both proteins were colocalized in a distribution pattern reminiscent of that of the 84-kDa protein (Fig. 8P to R). These findings indicate that there may be some cooperative behavior inherent to the localization of different UL112-113 proteins and that either the expression of the 84-kDa protein or the coexpression of both the 84-kDa and 43-kDa proteins may regulate, to some degree, the localization patterns of these other UL112-113 proteins.

Effects of self-interaction and coexpression of the four UL112-113 proteins on the recruitment of UL44 protein to prereplication foci.

When the six core replication complex components (UL54, UL44, UL57, UL105, UL70, and UL102) were coexpressed along with the UL112-113 proteins, the UL44 protein (the polymerase processivity factor) was shown to be recruited to the POD-associated sites in DNA cotransfection assays (2). For this study, we investigated whether the UL112-113 proteins are able to do so when coexpressed along with the UL44 protein only. The Flag-tagged UL44 protein was localized in a uniform nuclear diffuse pattern in transfected Vero cells (Fig. 9A, panels a to c). When Flag-UL44 was cotransfected with the plasmid harboring the UL112-113 genomic gene, which encodes all four of the UL112-113 proteins simultaneously, the distribution pattern of Flag-UL44 changed to include typical punctate forms which were colocalized with the 84-kDa protein (Fig. 9A, panels d to f). This result demonstrates that the UL112-113 proteins are able to recruit UL44 into the POD-associated sites, even in the absence of other core replication complex components.

We next investigated the importance of UL112-113 self-interaction with regard to the recruitment of UL44 to the nuclear foci. Vero cells were cotransfected with plasmids containing the HCMV replication origin and plasmids encoding six replication core proteins. When we conducted IFA with anti-UL44 Ab, we observed localization of UL44 as a nuclear diffuse form (Fig. 9B, panel a). When four auxiliary proteins (such as UL36-38, IE1/IE2, UL84, and UL112-113) were added, we observed efficient generation of UL44-containing nuclear foci which resembled the viral prereplication foci in HCMV-infected cells (2) (Fig. 9B, panel b). However, those UL44 foci were not produced at all in cases in which the UL112-113 plasmid was omitted from the transfection mix (Fig. 9B, panel c) or the expression plasmid for the UL112-113 genomic gene was replaced with one containing a deletion of the N-terminal 25 amino acids, UL112-113(Δ1-25) (Fig. 9B, panel d). Considering that the Δ1-25 mutant produces self-interaction-deficient mutant variants of the UL112-113 proteins, our findings demonstrate that the self-interaction exhibited by UL112-113 is a prerequisite for the recruitment of the UL44 protein to the prereplication foci.

Finally, we investigated whether simultaneous coexpression of all four UL112-113 proteins exerts a cooperative effect on their capacities to form UL44 foci (Fig. 9C). When the UL44 protein was coexpressed with the 43-, 50-, or 84-kDa protein, the formation of UL44 foci was detected in 4, 9, and 21% of the UL44-positive cotransfected cells, respectively. When the four UL112-113 proteins were simultaneously coexpressed from the genomic gene, the efficiency with which the UL44 foci were formed increased drastically, to 88%, thereby suggesting that the expression of all four UL112-113 proteins facilitates the recruitment of UL44. We also introduced the ΔC mutation of the 84(ΔC) mutant form (Fig. 2B) into the background of the genomic gene. This mutant UL112-113(ΔC) gene was shown to encode the 34-, 43-, and 50-kDa proteins normally, but not the intact 84-kDa protein (Fig. 9D). When the UL112-113(ΔC) gene was employed instead of the intact UL112-113 gene, the formation of UL44 foci was reduced to 36%. The expression levels of the UL112-113 proteins in these experiments are shown in Fig. 9D. This immunoblot analysis verified both the expression of the 25-amino-acid-deleted UL112-113 proteins from the UL112-113(Δ1-25) gene and the lack of expression of the intact 84-kDa form of the UL112-113(ΔC) gene. Taken together, our results clearly demonstrate that the simultaneous coexpression of the four UL112-113 proteins results in a cooperative effect on their activity of recruiting UL44 into the prereplication foci.