INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Herpes simplex virus type 1 (HSV-1) encodes seven proteins required for replication of its 152-kb double-stranded DNA genome (27). During productive infection, initiation of HSV DNA replication occurs at viral origins which include one copy of oriL, located in the unique long region of the genome, and two copies of oriS, located in the repeat sequences flanking the unique short region of the genome (22, 25). The core element of oriS consists of a 90-bp sequence that includes a 45-bp imperfect palindrome, whereas the core element of oriL consists of a 144-bp perfect palindrome (Fig. 1). Despite these differences, oriL and oriS share extensive nucleotide sequence homology, and both origins contain binding sites for the origin binding protein (OBP), encoded by the UL9 gene (8, 9, 19). OBP binds specifically to sequences within the origins, termed sites I, II, and III, which differ slightly in nucleotide sequence and thus in binding affinity for OBP (site I > site II > site III) (11). The imperfect oriS palindrome contains one copy each of sites I, II, and III, whereas the perfect oriL palindrome contains two copies each of sites I and III (Fig. 1).

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Comparison of sizes and sequences in HSV-1 oriL and oriS. The DNA sequences of the core origin of oriL (a 144-bp perfect palindrome) and oriS (a 45-bp imperfect palindrome) of HSV-1 KOS are shown. Black dots indicate the nucleotide differences between oriL and oriS. OBP binding sites are shaded in light gray and labeled I, II, and III. The GRE in oriL is boxed, and the origin factor 1 (OF-1) bipartite binding site in oriS is indicated by the thin black line. The sequence of the degenerate GRE in oriS is adjacent to the perfect GRE in oriL. The early genes that flank oriL (UL29 and UL30) and the immediate-early genes that flank oriS (ICP4 and ICP22/47) are shown in dark gray. The locations of the transcriptional start sites of these genes relative to the base of the origin palindrome are also shown. The 24-bp oriS site I probe used in gel shift assays throughout this study is represented by the thick vertical line.

The functional significance of the presence of three origins of DNA replication within the HSV-1 genome is not clear. Mutant viruses lacking either oriL or both copies of oriS have no obvious growth defects during productive infection of cells in culture, suggesting that the two types of origin are able to substitute functionally for each other (13, 21). However, there is evidence that oriL and oriS differ with respect to the efficiency of origin function in neural (PC12) and nonneural (Vero) cells. Specifically, in undifferentiated PC12 cells and in Vero cells, oriL and oriS function with similar efficiency in in vitro assays. In nerve growth factor (NGF)-differentiated PC12 (Nd-PC12) cells, the efficiency of oriS function is significantly reduced whereas oriL function is the same as in undifferentiated PC12 or Vero cells. Furthermore, addition of the synthetic glucocorticoid dexamethasone (DEX) enhances oriL function and further represses oriS function. Although the mechanism responsible for the repression of oriS function in PC12 cells in response to NGF-induced differentiation is unclear, the enhancement of oriL function in Nd-PC12 cells in response to DEX was shown to be mediated through a perfect glucocorticoid response element (GRE) present in oriL (10). oriS contains a degenerate GRE that may be responsible for the DEX-induced repression of oriS function in Nd-PC12 cells. Based on these findings, the differences in nucleotide sequence between oriL and oriS clearly have significant functional consequences. Moreover, cell-type-specific functional differences likely have significant implications for the biology of HSV (e.g., the ability to establish and reactivate from latent infection in neurons). One possible explanation for these functional differences between oriL and oriS is that formation of OBP-containing protein-DNA complexes differs at oriL versus oriS and/or in cells of neural versus nonneural lineage.

In the current model of HSV DNA replication, OBP functions as a DNA replication initiator protein by binding to HSV origins, initiating unwinding of origin DNA, and recruiting additional viral DNA replication proteins to the initiation site (reviewed in reference 4). Thus, the ability of OBP to bind to viral origins was shown to be essential for HSV DNA replication, as mutations within the origins themselves or within OBP that abrogate origin binding inhibit origin-dependent DNA replication (12, 23, 24). DNA footprinting and electron microscopy have demonstrated that binding of OBP to sites I and II in oriS loops and distorts the A+T-rich apex of the origin, which is thought to allow for subsequent DNA unwinding via the ATP-dependent helicase activity of OBP (14, 15). Binding of OBP to viral origins is also thought to facilitate recruitment of other essential replication proteins (i.e., the product of the UL8, UL42, and UL29 genes) to the sites of initiation via direct protein-protein interactions (3, 16, 17).

A truncated form of OBP, termed OBPC, has also been shown to bind to oriL and oriS (2). OBPC, encoded by the UL8.5 gene, is the product of a unique delayed-early transcript that originates within the open reading frame (ORF) of the gene encoding OBP (UL9) (1). Because the genes encoding OBP and OBPC are translated in the same reading frame, the amino acid sequence of OBPC is identical to the C-terminal 487 amino acids of OBP; however, OBPC lacks the N-terminal domain of OBP, which includes five of the six conserved helicase motifs, the ATP-binding and leucine zipper motifs, as well as domains that mediate interactions with a component of the helicase-primase complex (UL8) and the DNA polymerase processivity factor (UL42). Based on the observations that OBPC localizes to the nuclei of HSV-infected cells and binds to origin DNA, it has been postulated that OBPC may play a role in HSV DNA replication (2). Notably, overexpression of OBPC or C-terminal peptides of OBP has been shown to inhibit origin-dependent viral DNA replication and plaque formation by infectious HSV-1 DNA, presumably by occupying OBP binding sites and interfering with the initiation process (2, 20). Ultimately, determination of the precise role of OBPC in HSV DNA replication will require isolation of an OBPC mutant. Construction of this virus has not yet been achieved, however, because the genes encoding OBP and OBPC overlap and their ORFs are translated in the same reading frame. Consequently, mutagenesis of the nucleotide sequence of the OBPC promoter and transcriptional and translational start sites (which would be needed to abrogate expression of OBPC) also alters the amino acid sequence of OBP. Since maintaining a functional form of OBP is critical to differentiating the roles of OBP and OBPC, construction of an OBP⁺ OBPC virus is a formidable challenge under these circumstances. Numerous efforts to introduce mutations into the wobble position of OBP codons (i.e., into the promoter and start sites of the OBPC gene) that do not alter the amino acid sequence of OBP but eliminate OBPC expression have met with only limited success. Given the difficulty in isolating a suitable OBPC mutant virus, we have been forced to evaluate OBPC function by using alternative approaches, including characterization of mutations in site I that eliminate binding of OBP but not OBPC to HSV origins.

To characterize the formation of protein complexes at oriS and to determine whether these complexes differ when proteins are derived from neural versus nonneural cells, we used nuclear extracts of HSV-infected Nd-PC12 and Vero cells as the source of protein and a DNA sequence containing oriS site I (Fig. 1) as the probe in gel shift assays. Using nuclear extracts of both cell types tested, three HSV-specific protein complexes, A, B, and C, were shown to bind specifically to the site I probe. The three complexes exhibited similar migration patterns when proteins were derived from Nd-PC12 or Vero cells. Mapping of the precise nucleotide binding sites of the three complexes revealed that complex A (containing only OBPC) binds to a site lying entirely within the shared binding site of two OBP-containing complexes, B and C. We generated single-nucleotide substitution mutations which eliminate formation of all three complexes or of complexes B and C only and determined their effects on oriS-dependent DNA replication. Both mutations reduced oriS-dependent DNA replication significantly in in vitro assays.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and viruses PC12 cells (a gift from John Wagner, Cornell University Medical College, New York, N.Y.) were grown as described previously (10). PC12 cells were induced to differentiate by incubation in medium containing NGF (2.5S; Collaborative Biomedical Products, Bedford, Mass.) at a concentration of 100 ng/ml for 6 days, with one medium change on day 3 postplating. In experiments in which PC12 cells were treated with DEX, 0.5 µM DEX was added to the medium at the time of infection as previously described (10). Vero cells (ATCC CCL-81) were propagated and maintained as described previously (7). The wild-type strain of HSV-1, KOS, was grown and assayed as previously described (7).

Plasmids and mutagenesis. A plasmid containing wild-type oriS (pOS822 [26]) was used in this study. Two single-nucleotide substitution mutations described below, termed mutOBP and mutR(C-G), were introduced into pOS822 by the Quick Change mutagenesis method (Stratagene, La Jolla, Calif.), with the addition of 8 µl of 25% glycerol and 3 µl of dimethyl sulfoxide to each 50-µl reaction. Reactions underwent 12 amplification cycles in a thermal cycler block (MHJ Research, Watertown, Mass.), using the following parameters: 95°C for 30 s, 55°C for 1 min, and 68°C for 12 min.

Gel mobility shift assays. Nuclear extracts were prepared from 3 × 10⁶ HSV-infected (multiplicity of infection of 10 PFU/cell) Nd-PC12 and Vero cells at 12 h postinfection (hpi), and protein concentrations were measured as described previously (10).

Bacterial production of HisOBPC. The UL8.5 gene was cloned into the bacterial expression vector pTrcHisA (Invitrogen, Carlsbad, Calif.) such that a six-histidine tag was translated in frame with OBPC. Induction of HisOBPC expression was performed by addition of isopropyl--D-thiogalactopyranoside (1 mM) to XL1-Blue cells (Stratagene) as outlined in the Xpress system protein expression manual (Invitrogen). After 4 h, cells were lysed by addition of 1 mg of lysozyme per ml. Sarkosyl (1.5%) was added to cell lysates to improve protein solubility. HisOBPC was bound to a Ni²⁺ column containing a 50% slurry of ProBond resin (Invitrogen), and fractions were eluted by addition of 20 mM phosphate buffer (pH = 4) as outlined in the ProBond resin purification manual (Invitrogen). The purity of OBPC in each fraction was verified by sodium dodecyl sulfate-PAGE followed by Coomassie blue staining. The fraction in which HisOBPC was the only protein detectable by Coomassie blue staining was collected, and its protein concentration was measured by the method of Bradford (Bio-Rad, Hercules, Calif.), using a standard curve generated with known amounts of bovine serum albumin as the standard.

In vitro origin-dependent DNA replication assays. In vitro oriS-dependent DNA replication assays were performed as described previously (10). Briefly, 3.5 × 10⁶ PC12 cells were seeded in collagen-coated 100-mm-diameter plates. Twenty-four hours after plating, cells were transfected with 10 µg of wild-type or mutant pOS822 by the Lipofectin method (Gibco, Grand Island, N.Y.). After 5 h, cells were washed, and fresh medium containing NGF (100 ng/ml) was added. NGF differentiation proceeded for 6 days, with one medium change on day 3. After 6 days, cells were infected with KOS at a multiplicity of 10 PFU/cell. Cells were harvested at 18 hpi, and total cellular DNA was isolated. For Vero cells, 3.5 × 10⁸ cells were plated, transfected 24 h later with the indicated plasmid, and infected 24 h after transfection. Five micrograms of total Nd-PC12 or Vero cell DNA was digested with HindIII (to linearize the vector) and either MboI (to cleave unmethylated DNA) or DpnI (to cleave methylated DNA). Digested DNA was resolved by electrophoresis on a 0.8% agarose gel and transferred to a nylon membrane. After UV cross-linking, the membrane was prehybridized for 1 h at 55°C in ExpressHyb solution (Clontech, San Francisco, Calif.) and hybridized for 3 h at 55°C with a ³²P-labeled probe (3 × 10⁶ cpm/ml) generated by nick translation of pGEM7Zf⁺ (vector backbone of pOS822). The membrane was then washed according to the ExpressHyb protocol (Clontech) and exposed in a Phosphorlmager cassette.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Three HSV-specific complexes bind to oriS site I when nuclear extracts of Vero or Nd-PC12 cells are used as the source of protein. Previously published gel shift assays from this laboratory using total Vero cell extracts as the source of protein described binding of two HSV-specific complexes (complexes A and B) to oriS site I (5, 6, 9). In the present study, we used nuclear extracts of Vero, Nd-PC12 cells, or Nd-PC12 cells treated with DEX and observed binding of three HSV-specific complexes (complexes A, B, and C) to a double-stranded DNA probe containing oriS site I (Fig. 1). Specifically, to compare whole-cell versus nuclear extracts as the source of protein, we prepared extracts from KOS-infected Vero cells. Using the site I probe, the pattern produced using total Vero cell extract (Fig. 2A, lane 1) was nearly identical to that observed in previous studies using total Vero cell extract (5, 6, 9). Several differences were observed, however, when total Vero cell extract was compared with nuclear extract as the source of protein (Fig. 2A). Complex C was much less prominent in total cell extract (lane 1) than in nuclear extract (lane 2), suggesting that nuclear extracts are greatly enriched for the proteins that comprise complex C. The intensity of complex A was also reduced in total cell relative to nuclear extracts, whereas the intensity of complex B was considerably greater in total cell than in nuclear extracts. Moreover, the mobility of complex B in the two types of extract differed. The enhanced intensity of complex B formed from total cell extracts suggests that these extracts are enriched for components of complex B relative to nuclear extracts. A fourth complex that migrated slightly below complex B was prominent in gel shifts using total cell extracts but barely visible in shifts using nuclear extracts. Taken together, these observations indicate that the efficiency of formation of specific complexes at oriS site I is dependent on the presence of individual proteins within the cytoplasm or nucleus and the concentrations of the proteins within these two compartments.

View larger version (53K): [in this window] [in a new window]   FIG. 2.   Identification and specificity of oriS binding complexes. (A) Vero cells were infected with KOS at a multiplicity of 10 PFU/cell. At 12 hpi, total cell or nuclear extracts were prepared and used in gel shift assays. (B) NGF-differentiated PC12 cells were infected with KOS at a multiplicity of 10 PFU/cell and harvested at the times indicated. Nuclear extracts were prepared, and protein-DNA complexes were analyzed as described in Materials and Methods. The letters to the right correspond to specific complexes (A, B, and C). (C) Quantitation and comparison of the intensities of bands corresponding to complexes A, B, and C observed during the time course shown in panel B, expressed as Phosphorlmager units. (D) HSV-infected Nd-PC12 cell nuclear extracts were incubated with the site I probe either alone (KOS) or in the presence of a 100-fold molar excess of unlabeled specific competitor DNA (+100X site I) or unlabeled nonspecific competitor DNA (+100X NF-1).

Do complexes A, B, and C contain OBP and/or OBPC? To determine whether complexes A, B, and C contain OBP and/or OBPC, we performed antibody supershift experiments in which antibodies specific for OBP and OBPC were added to gel shift binding reactions (Fig. 3A). Addition of antibody specific for OBP and OBPC shifted the mobility of all three complexes (lane 3, two supershifted bands are marked with asterisks; a third is frequently visible below the second supershifted band), indicating that complexes A, B, and C contain OBP and/or OBPC. Although addition of the OBP/OBPC-specific antibody also appears to shift the faint band which migrates slightly more slowly than complex C, this observation was not reproducible. Addition of antibody specific for ICP8 (provided by M. Zweig) had no effect on complex formation (data not shown).

View larger version (52K): [in this window] [in a new window]   FIG. 3.   Characterization of HSV-specific complexes that form at oriS. (A) HSV-infected PC12 cell nuclear extracts were incubated with the site I probe. After 5 min of incubation, no antibody (lane 1), antibody specific for GR (lane 2), or antibody specific for OBP (lane 3) was added to the binding reaction. Incubation was continued for an additional 25 min, and protein-DNA complexes were resolved by nondenaturing PAGE. Complexes are labeled A, B, and C, and supershifted complexes are indicated with asterisks. (B) Nd-PC12 cells were infected with KOS either in the absence (lane 1) or presence (lane 2) of PAA (400 µg/ml), cells were harvested at 12 hpi, and nuclear extracts were prepared; 5 µg of nuclear extract was incubated with the site I probe, and complexes were resolved and visualized as described for panel A. (C) HSV-infected PC12 cell nuclear extract (lane 1) or bacterially expressed His OBPC (lane 2 and 3) was incubated with the oriS site I probe. Complexes were resolved and visualized as described for panel A. Lane 3 is a darker exposure of lane 2.

Nucleotide mapping of the binding sites of complexes A, B, and C. To identify the precise nucleotides required for formation of complexes A, B, and C with the site I probe, we analyzed the ability of these complexes to form with probes that contained sequential two-base-pair substitution mutations (Fig. 4A, mutations are indicated in bold). In contrast to complexes produced using the wild-type probe, complexes B and C were barely detectable upon incubation of nuclear extracts with probes Mut3 through Mut7, indicating that the mutations contained within these probes abrogate formation of these complexes. The intensity of complexes B and C was reduced to a lesser extent upon incubation of extracts with probes Mut8, -11, and -12 relative to the wild-type probe, implicating the wild-type nucleotides mutated in these probes as essential for efficient formation of complexes B and C with site I DNA. Formation of complex A was reduced significantly upon incubation of extracts with probes Mut4, -5, and -6 and slightly upon incubation with Mut12. The enhanced intensity of two bands that migrate below complex B in tests using probes Mut3 through -8 and Mut12 was notable. The composition of these bands (i.e., whether they contain viral or cellular proteins) is unknown. These results indicate that 10 nucleotides (represented by probes Mut3 through Mut7) are required for formation of complexes B and C, whereas only 6 nucleotides (represented by probes Mut4 through Mut6) are required for efficient formation of complex A. Interestingly, the dinucleotide represented by Mut12 appeared to be required for efficient formation of all three complexes. Whether this result is real or a consequence of the location of the mutated nucleotides at the 3' end of the probe remains to be determined. Additional tests in which unlabeled mutant probes were used to compete for formation of complexes A, B, and C to the wild-type site I probe confirmed these results (data not shown). Based on these findings and as shown schematically in Fig. 5B, the binding site for complex A is contained entirely within the shared binding site for complexes B and C. 

View larger version (86K): [in this window] [in a new window]   FIG. 4.   Mapping of the nucleotide binding sites of complexes A, B, and C. (A) Nucleotide sequence of the wild-type (wt) and mutant site I probes. Two-base-pair substitution mutations (indicated in bold) were introduced throughout the probe, yielding a total of 12 mutant probes (labeled Mut1 through Mut12). (B) HSV-infected PC12 cell nuclear extract was incubated with equal counts of labeled wild-type or mutant site I probe; protein-DNA complexes were resolved by nondenaturing PAGE, gels were dried, and complexes were visualized with a Phosphorlmager. The profile of complexes that form on the wild-type probe is shown in the far left and right lanes. Complexes A, B, and C are indicated in these lanes.

A point mutation eliminates formation of complexes B and C without affecting formation of complex A at oriS site I. To further test the observation that the binding site for complex A is contained entirely within the shared binding site of complexes B and C, we generated a probe containing a single-nucleotide substitution mutation outside the binding site for complex A but within the binding site for complexes B and C, this mutant probe was designated mutOBP (Fig. 5B). Similarly, we generated a probe that contained a single-nucleotide substitution mutation within the binding site for all three complexes; this probe was designated mutR(C-G). mutR(C-G) differs from mutR(C-A) described previously by Dabrowski et al. (5) in that in this study, C was changed to G rather than A (Fig. 5B). Nuclear extracts from KOS-infected Nd-PC12 cells were incubated with the wild-type oriS site I, mutOBP, or mutR(C-G) probe, and protein-DNA complex formation was evaluated by gel shift analysis (Fig. 5A). As expected, complexes A, B, and C formed efficiently with the wild-type probe; however, complex A, but not complex B or C, formed with mutOBP, and none of the three complexes formed with mutR(C-G). These findings confirm those presented in Fig. 4 and demonstrate that a single-nucleotide substitution mutation is sufficient to eliminate binding of all three complexes. The nucleotide sequences of the wild-type and mutant probes and of the complexes that bind to each probe are summarized schematically in Fig. 5B.

View larger version (43K): [in this window] [in a new window]   FIG. 5.   Point mutations in the site I probe eliminate formation of specific protein-DNA complexes. (A) HSV-infected Nd-PC12 cell nuclear extract was incubated with wild-type (wt) or mutant [mutOBP or mutR(C-G); sequences shown in panel B] site I probes. Binding reactions were subjected to nondenaturing PAGE, gels were dried, and complexes were visualized with a Phosphorlmager. Complexes are marked A, B, and C. (B) DNA sequences of the wild-type, mutOBP, and mutR(C-G) probes. Single-base-pair substitution mutations are indicated in bold. The binding sites for complex A (smaller box) and complexes B and C (larger box) are indicated (based on Fig. 4B). Above each probe sequence are indicated the complexes that form with each probe (an "X" through the complex designates the failure of the complex to form with that probe).

Effect of mutOBP and mutR (C-G) on oriS-dependent HSV DNA replication. To examine the effect of formation of complexes A, B, and C at site I on origin-dependent HSV DNA replication, the mutOBP and mutR(C-G) mutations were introduced into the wild-type oriS-containing plasmid, pOS822, and the effects of these mutations were evaluated in in vitro origin-dependent DNA replication assays (Fig. 6A). Undifferentiated PC12 cells were transfected with pOS822 (lanes 1 to 4) or pOS822 containing the mutR(C-G) (lanes 5 to 8) or mutOBP (lanes 6 to 10) mutation. PC12 cells transfected with the pOS822 backbone (pGEM) which does not contain oriS were used as a negative control for plasmid amplification (data not shown). Following transfection, PC12 cells were differentiated with NGF for 6 days and then infected with KOS. At 18 hpi, total cell DNA was extracted, subjected to restriction enzyme digestion with MboI or DpnI to distinguish between input or newly replicated plasmid DNA, respectively, and analyzed by Southern blot. The probe used in these tests hybridizes to a DNA sequence contained in the vector component of all pOS822-derived plasmids. The blot in Fig. 6A shows the results of each plasmid tested in duplicate, and the data are presented quantitatively in Fig. 6B. The intensity of the band corresponding to newly replicated pOS822 (DpnI resistant) (Fig. 6A, lanes 2 and 4, arrow) was ~11-fold greater than that of input pOS822 (MboI resistant (lanes 1 and 3), indicating that the wild-type oriS-containing plasmid was amplified efficiently (Fig. 6B). In contrast, amplification of plasmids containing the mutR(C-G) (lanes 6 and 8) or mutOBP (lanes 10 and 12) mutation was considerably less than that of the wild-type plasmid (lanes 1 and 3) (Fig. 6). Similar results to mutR(C-G) were obtained for mutR(C-A) by Dabrowski et al. (5). The levels of newly replicated mutant plasmids were only 1.4- and 2.0-fold, respectively, above the input plasmid levels. Therefore, the single-nucleotide substitution mutations in mutR(C-G), mutR(C-A), and mutOBP reduced the efficiency of plasmid replication to 10% [mutR(C-G); Fig. 6B], 4% [mutR(C-A)] (5), and 25% (mutOBP; Fig. 6B) of the wild-type level, indicating that mutations in the shared region of the binding sites of complexes A, B, and C [mutR(C-A), mutR(C-A)] or B and C (mutOBP) greatly reduce the ability of oriS to support efficient origin-dependent DNA replication.

View larger version (39K): [in this window] [in a new window]   FIG. 6.   Effects of mutations mutOBP and mutR(C-G) on oriS-dependent HSV DNA replication. (A) PC12 cells were transfected with pOS822-derived plasmids containing wild-type (wt; lanes 1 to 4), mutR(C-G) (lanes 5 to 9), and mutOBP (lanes 9 to 12) or mutant [mutR(C-G) and mutOBP] oriS sequence. The pOS822 vector (pGEM) plasmid was used as a negative control (data not shown). Following transfection, PC12 cells were differentiated with NGF for 6 days and infected with KOS at a multiplicity of 10 PFU/cell. At 18 hpi, total cellular DNA was extracted, subjected to digestion with MboI (lanes 1, 3, 5, 7, 9, and 11) or DpnI (lanes 2, 4, 6, 8, 10, and 12) to differentiate newly replicated DNA, and analyzed by Southern blot hybridization using a 32P-labeled, nick-translated probe which recognizes pGEM vector sequence present in the wild-type and mutant oriS-containing plasmids. Each sample was tested in duplicate. oriS-containing plasmid DNA, which is resistant to enzyme digestion, is indicated by an arrow. (B) The bands shown in panel A were quantitated by Phosphorlmager analysis, and fold replication was calculated by dividing the sum of the MboI-resistant and DpnI-resistant bands by the MboI-resistant band for each sample. Mutant plasmid amplification is expressed as a percentage of wild-type plasmid amplification (set to 100%). The graph shown represents the average of the duplicate samples shown in panel A.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

OBP and OBPC-containing complexes that form at oriS site I. The results presented here confirm and extend previous studies in which total Vero cell extracts were used to demonstrate formation of two protein complexes, complexes A and B, at oriS site I by gel shift analysis (5, 6, 9). Comparison of the site I-binding complexes from total cell and nuclear extracts revealed differences in the intensities of complexes A and B as well as the presence of an additional complex, complex C, whose formation appears to be dependent on nuclear factors (Fig. 2A). The differences in complex formation noted when total cell versus nuclear extracts were used may reflect differences in the presence or concentration of nucleus- or cytoplasm-specific proteins present in these complexes. Alternatively, the conformation, oligomerization, or posttranslational modification of proteins contained within these complexes may differ in the cytoplasm versus the nucleus. That the interaction of OBP-containing complexes with oriS is dependent on factors present in the nucleus and/or cytoplasm likely has important implications for the functional role of OBP during HSV infection. Notably, however, no differences in complex formation were noted when proteins used in gel shift assays were derived from nuclei of Vero or Nd-PC12 cells or Nd-PC12 cells treated with DEX. Thus, the basis for the differential effects of NGF and DEX on oriS function noted previously (10) remains unclear.

Roles of OBP and OBPC in HSV DNA replication. Extensive mutagenesis of the oriS site I probe revealed that complexes B and C share a 10-bp binding site which was previously described as OBP binding site I (11). Despite the fact that OBP and OBPC share the same C-terminal DNA binding domain, the DNA binding site of complex A, which appears to consist solely of OBPC, was found to comprise only 6 bp which lie totally within the 10-bp binding site of complexes B and C. That the OBP-containing complexes require a larger binding site than OBPC may reflect the larger size of OBP relative to OBPC, dimer formation of OBP but not OBPC, or the presence of additional cellular proteins in complexes B and C but not A which are necessary for efficient contact with origin DNA. Notably, the sizes of complex A and B binding sites determined in this study are contrary to previous findings in which the nucleotide binding site of complex A was several nucleotides larger than that of complex B (6). These differences are likely due to differences in the composition of the cell extract used (total cell extract [6] versus nuclear extract [this study]) as well as differences in the mutant probes used for mapping studies (nucleotide deletions [6] versus dinucleotide substitutions [this study]).