INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In this report, we describe the identification of a new open reading frame (ORF) in the genomes of herpes simplex virus types 1 and 2 (HSV-1 and HSV-2). What makes this ORF particularly interesting is its location antisense to U_L27, the gene encoding glycoprotein B (gB) (5, 22). The initial objectives of this study were quite different from its outcome.

The sequence of the HSV-1 genome (22) corroborated the existence of several glycoproteins and led to the discovery of others, bringing the total to 11 (gB, gC, gD, gE, gG, gH, gI, gJ, gK, gL, and gM) (reviewed in references 28, 30, and 32). Of these, gJ was reported to be dispensable for viral replication in cell culture and was the least well understood (4, 21, 33). To initiate these studies, we immunized rabbits with a fusion protein consisting of maltose binding protein fused in frame to the entire sequence of gJ. The resulting antibody reacted with a protein (presumed to be gJ) whose gene maps in the S component and with higher-M_r proteins in lysates of HSV-1- and HSV-2-infected cells. Extensive mapping studies led ultimately to the conclusion that these protein bands are encoded by a gene mapping antisense to gB. A possible explanation for the reactivity of the polyclonal antibody with both proteins rests on the observation that HSV-1 and HSV-2 U_L27.5 proteins have a short amino acid sequence common to the predicted amino acid sequence of gJ.

Transcripts antisense to known ORFs were first reported by Stevens et al. (31). The latency-associated transcript, however, does not appear to affect the expression of ICP0, the gene to which it is partially antisense. In the past several years, this laboratory reported two sets of antisense genes. Thus, ORFs O and P map antisense to the ₁34.5 gene, and the U_L43.5 gene maps antisense to the U_L43 gene (6, 20, 26). The striking feature of the antisense genes is that their expression is mutually exclusive. Thus, derepression of ORF P leads to its expression early in infection and grossly reduces the normal expression of the ₁34.5 gene (27). U_L43 and U_L43.5 are also expressed at different times after infection (6). The gB/U_L27.5 genes are the third set of genes located antisense to each other. gB is expressed relatively early in infection, whereas the U_L27.5 gene belongs to the ₂ group in that its expression is totally dependent on viral DNA synthesis (14, 15).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells and viruses. Rabbit skin cells and Vero cells were obtained from John McLaren and American Type Culture Collection, respectively, and were maintained in Dulbecco's modified Eagle medium supplemented with 5% newborn calf serum. BHK(TK+) cells (American Type Culture Collection) and 143TK cells (obtained from Carlo Croce) were maintained in the same medium supplemented with 5% fetal bovine serum. Infected cells were maintained in mixture 199 supplemented with 1% calf serum (199V) unless indicated otherwise. HSV-1(F) and HSV-2(G) are prototypes of HSV-1 and HSV-2 strains, respectively, used in this laboratory (10). Intertypic recombinants HSV-1(F) × HSV-2(G) were described previously (1, 8). In R7015 the HSV-1 S component was replaced by the homologous HSV-2(G) sequences. HSV-1(F)305 was derived from HSV-1(F) and has a 501-bp deletion in U_L23 (thymidine kinase) and U_L24 genes (25). The HSV-1(KOS) mutant, lacking the U_L26 gene (11), was the generous gift of Steven Weinheimer (Bristol-Myers Squibb).

Antibodies. Mouse monoclonal antibody CH28 to the human cytomegalovirus (HCMV) gB epitope or G1102 to ICP35 was purchased from Goodwin Institute (Plantation, Fla.). ICP0 polyclonal antibody was described previously (18). Goat anti-mouse or anti-rabbit immunoglobulin G alkaline phosphatase-conjugated secondary antibodies were purchased from Bio-Rad (Hercules, Calif.).

Generation of U_S5 polyclonal antibody. The U_S5 maltose binding protein expression plasmid pRB5154, containing the entire coding region of U_S5, was transfected into Escherichia coli. The induction and purification procedure of maltose binding fusion protein was done as recommended by the manufacturer (New England BioLabs, Beverly, Mass.). A New Zealand White female rabbit was injected subcutaneously five times, each time with 250 to 400 µg of fusion protein emulsified with Freund adjuvant.

Detection of viral proteins in lysates electrophoretically separated in a denaturing polyacrylamide gel. Mock-infected or HSV-1-infected Vero or BHK (TK+) cells were harvested in disruption buffer (50 mM Tris-HCl [pH 7.0], 2% sodium dodecyl sulfate (SDS), 0.7 M -mercaptoethanol, 2.75% sucrose), electrophoretically separated in a denaturing polyacrylamide gel cross-linked with N,N'-diallyltartardiamide, electrophoretically transferred to a nitrocellulose membrane (Schleicher & Schuell), and probed with appropriate antibodies. The immunoblotting procedure was as described elsewhere (7). The molecular weight marker was obtained with the LMW electrophoresis calibration kit (Pharmacia Biotech, Uppsala, Sweden).

Generation of plasmids and recombinant viruses. All molecular cloning was done by standard techniques as described elsewhere (29). The maltose binding protein-U_S5 chimeric construct was cloned as pRB5154 and contained the entire U_S5 coding sequence. The U_S5 coding sequence with an insertion at the EcoO109 site of an oligomer encoding the HCMV gB epitope (5' AAAAGGG ACAGAAGCCCAACCTGCTAGACCGACTGCGACACCGCAAAAACGG GTACCGACACC 3') was ligated to the U_L26.5 promoter sequence (599 to +44, relative to the transcription start site of the U_L26.5 gene) to create pRB4152. The U_L26.5 promoter-U_S5 construct was cloned into the HSV-1(F) thymidine kinase (tk) gene in place of the BglII/EcoNI sequence of the coding domain of the tk gene to create pRB5175. In turn, the HSV sequence in pRB5175 was recombined into HSV-1(F) as previously described to generate the recombinant virus R5175.

Analyses of viral DNA by hybridization. Cytoplasmic DNAs were purified from infected cells as described elsewhere (16), digested with either BamHI or EcoRV, separated by agarose gel electrophoresis, and transferred to a nylon membrane (Bio-Rad). The hybridization procedure was done according to the manufacturer's recommendations. Either the HSV-1(F) BamHI Q fragment cloned as pRB165 or the gel-purified HSV-1(F) EcoRV D fragment was used as a probe. All probes were internally labeled with [³²P]dCTP (800 Ci/mmol; DuPont, NEN, Boston, Mass.) with a nick translation kit (DuPont, NEN).

Generation of GST-U_L27.5 fusion protein. The sequence encoding U_L27.5 codons 299 to 412 was amplified by PCR and cloned into glutathione S-transferase (GST) expression vector pGEX-KG as pRB5132 (12). The induction and purification of a GST fusion protein were done as recommended by the manufacturer (Promega, Madison, Wis.). Affinity-purified fusion protein was loaded onto an SDS-15% polyacrylamide gel and stained with Coomassie blue or electrophoretically transferred to a nitrocellulose membrane and subjected to immunoblot analysis.

In vitro transcription and translation. The U_L27.5 ORF amplified by PCR was cloned into pGEM3Zf(+) as pRB5169 and used as a template for in vitro transcription and translation. The reaction was carried out in a TNT coupled transcription/translation system (Promega) with T7 RNA polymerase in the presence of [¹⁴C]leucine (0.25 µCi, 342 mCi/mmol; DuPont, NEN). The final product was denatured in disruption buffer and subjected to electrophoresis on an SDS-15% polyacrylamide gel and electrophoretically transferred to a nitrocellulose sheet. The membrane was treated with En³Hance spray (Dupont, NEN) and subjected to autoradiography.

Cytoplasmic and nuclear fractionation. BHK(TK+) cells were mock infected or exposed to 10 PFU of HSV-1(F) or HSV-2(G) per cell and harvested at 18 h after infection. Infected cells were collected by low-speed centrifugation (1,000 × g for 5 min), resuspended in buffer A (10 mM HEPES buffer [pH 7.4], 10 mM NaCl, 1.5 mM MgCl₂), and stored on ice for 10 min. The cells were subjected to five strokes of Dounce homogenization and incubated on ice for 10 min. The cytoplasmic fraction was collected after centrifugation (12,000 × g at 4°C for 20 min). The pellet was washed once with buffer A and resuspended in buffer B (10 mM HEPES buffer [pH 7.4], 420 mM NaCl, 1.5 mM MgCl₂). The pellet was briefly sonicated to facilitate resuspension, and the nuclear fraction was collected after the removal of any insoluble material by centrifugation (12,000 × g at 4°C for 5 min). The proteins in the cytoplasmic and nuclear fractions were denatured by the addition of disruption buffer, boiled for 5 min, and subjected to SDS-polyacrylamide gel electrophoresis.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Genomic structure of recombinant viruses R4351, R4492, and R5175. Recombinant viruses were constructed by double homologous recombination as described in Materials and Methods, and their structures were verified by Southern blot analyses. Cytoplasmic DNAs from infected cells were digested with either the BamHI or the EcoRV enzyme, separated on an agarose gel, transferred to a nylon membrane, and hybridized with appropriate probes. As illustrated in Fig. 1C, lines 2 and 3, the replacement of the BglII/EcoNI fragment of the tk gene with a U_L26.5 promoter-U_S5-HCMV tag construct increased the size of HSV-1(F) BamHI-Q from 3.5 to 4.2 kb, as detected when ³²P-labeled pRB165 was used as the probe (Fig. 2B). To verify the predicted genomic structure of R4351 and R4492, electrophoretically separated EcoRV digests of the recombinant virus DNAs were probed with a ³²P-labeled HSV-1(F) EcoRV D fragment. There are two closely positioned EcoRV sites in the 27-tk construct. The EcoRV-D probe showed, as predicted, that insertion of the 27-tk construct and the 4 promoter into R4492 (Fig. 1C, lines 4 and 5) replaced the original EcoRV D fragment (14.8 kb) with two fragments of 2.4 and 14.6 kb in length (Fig. 2A, lanes 1 and 3). In the same fashion, the insertion of the 27-tk cassette into the R4351 viral genome (Fig. 1C, line 6) resulted in the replacement of the EcoRV D fragment with two fragments of 4.6 and 11.8 kb in length (Fig. 2A, lane 2). We conclude that the three recombinant viruses made in these studies show the expected genotype.

View larger version (19K): [in this window] [in a new window]   FIG. 1.   (A) Schematic diagram of genome structures of HSV-1(F) and HSV-1(F) × HSV-2(G) intertypic recombinants. Line 1, HSV-1(F) genome arrangement. Open rectangles represent inverted repeats, and the single lines in between represent unique long (UL) and unique short (US) regions of the genome. Lines 2 to 7, schematic diagram of genome arrangement of intertypic recombinants used in this study. Lines labeled 1 represent HSV-1(F) sequences, and lines labeled 2 represent HSV-2(G) sequences. Crossovers are indicated by bold-faced lines. R7015 (line 2) has the HSV-1(F) unique short region replaced with the HSV-2(G) unique short sequence. Lines 3 to 7 are a set of intertypic recombinants that are basically HSV-1(F), with various portions between map units 0.3 and 0.45 replaced with the HSV-2(G) sequence. (B) Schematic diagram of HSV-1(F) and various recombinants used in this study. Line 1, genome arrangement of HSV-1(F) and location of ORFs UL23 to UL28. Arrow indicates the polarity and position of each ORF. Line 2, position of the HSV-2(G) KpnI H fragment relative to the location of ORFs. Lines 3 and 4, positions of proposed new ORF UL27.5 in the HSV-1(F) and HSV-2(G) genomes, respectively. Lines 5 to 9, genome organization of various recombinants. Short vertical lines represent deletions (lines 5, 7, and 9). Triangles represent insertions (lines 6, 8, and 9). Filled rectangle represents HCMV epitope tag (line 9). (C) Line 1, sequence arrangement of HSV-1(F). Line 2, ORFs in HSV-1(F) BamHI Q fragment. Line 3, in R5175, an 0.8-kb EcoNI/BglII fragment within the UL23-to-UL24 region is replaced by a 1.5-kb UL26.5 promoter-US5 construct. Line 4, ORF arrangement in HSV-1(F) EcoRV D fragment. Line 5, in R4492, a 2.2-kb 27tk-4 promoter construct is inserted between the UL25 and UL26 genes. Line 6, in R4351, a 1.6-kb 27tk construct is inserted between the UL26 and UL27 genes. B, BamHI; Bg, BglII; E, EcoRV; En, EcoNI; N, NcoI.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Autoradiography of electrophoretically separated DNA hybridized with 32P-labeled probe. DNA of HSV-1(F) or recombinant viruses was digested with EcoRV (A) or BamHI (B), electrophoretically separated in agarose gel, transferred to a nylon membrane, and hybridized with labeled EcoRV D fragment (A) or BamHI Q fragment (B).

U_S5 rabbit polyclonal antibody reacted with two viral proteins by immunoblot analyses. Although the rabbit polyclonal antibody was made against the U_S5 protein, in immunoblot assays the anti-U_S5 polyclonal antibody reacted with two bands containing different proteins. Mock-infected or HSV-1(F)-infected BHK(TK+) cells were harvested in disruption buffer at 18 h after infection, electrophoretically separated in an SDS-15% polyacrylamide gel, electrophoretically transferred to a nitrocellulose sheet, and probed with the U_S5 polyclonal antiserum. The results (Fig. 3A, lane 5) reproducibly showed that the serum reacted with two virus-specific HSV-1 proteins with apparent M_rs of 43,000 and 23,000, respectively (designated band 2 and band 3, respectively). The antibody did not react with lysates of mock-infected cells (Fig. 3A, lane 1). In HSV-2(G)-infected cell lysate, only one virus-specific signal, with a mobility on a denaturing gel equivalent to that of HSV-1(F) band 2, was detected as a heterogeneous group of bands with apparent M_rs ranging from 43,000 to 50,000. The HSV-2 equivalent of HSV-1(F) band 3 was not detected (Fig. 3A, lane 3). Overdeveloped immunoblots occasionally showed another faint band with an apparent M_r of greater than 100,000 (designated band 1).

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Photograph of immunoblot of electrophoretically separated proteins reacted with proper antibodies. Vero cells mock infected or infected with HSV were harvested at 18 h postinfection, electrophoretically separated onto an SDS-15% polyacrylamide gel, electrically transferred to a nitrocellulose membrane, and reacted with US5 polyclonal antibody (panel A, lanes 2 to 8, and panels B and C) or monoclonal antibody recognizing the HCMV epitope (CH28) (panel A, lane 9). (B) Lane 3, HSV-1(KOS)UL26 (UL26 deletion virus m100); lane 4, HSV-1(KOS)UL26R (UL26 repaired virus). Molecular weight markers are shown in lane 1, and bands 1 to 4 are indicated at the right of panel A.

The gene encoding protein in band 2 maps in U_L. In this section, we describe three series of experiments designed to map the gene encoding band 2 protein. We took advantage of the difference in the electrophoretic mobilities of the HSV-1 and HSV-2 homologs of band 2 protein.

View larger version (43K): [in this window] [in a new window]   FIG. 4.   Photograph of immunoblot of electrophoretically separated proteins reacted with US5 polyclonal antibody. Vero cells mock infected or infected with wild-type HSV or intertypic recombinants, harvested at 18 h postinfection, electrophoretically separated in a denaturing polyacrylamide gel, transferred to a nitrocellular membrane, and reacted with US5 polyclonal antibody.

View larger version (39K): [in this window] [in a new window]   FIG. 5.   Photograph of immunoblot of electrophoretically separated proteins reacted with US5 polyclonal antibody. Vero cells were infected with various intertypic recombinants and harvested at 18 h postinfection. Proteins were electrophoretically separated on an SDS-15% polyacrylamide gel, transferred to a nitrocellulose membrane, and reacted with US5 polyclonal antibody. The bands formed by the UL27.5 and US5 proteins are identified at the right.

Sequence analysis predicts new ORF antisense to U_L27. We next searched for potential ORFs in the target regions stated above. We focused on new ORFs conserved in HSV-1 and HSV-2 with the capacity to encode at least 200 amino acids. As shown in Fig. 1B, lines 3 and 4, only one ORF, designated U_L27.5 and predicted to encode an HSV-1 protein of 575 amino acids, and an HSV-2 protein of 985 amino acids met these criteria. Sequence comparison showed that, except for the amino-terminal region of the predicted HSV-2 U_L27.5 ORF, the HSV-1 and HSV-2 U_L27.5 ORFs were homologous (Fig. 6). Since the U_S5 polyclonal antibody cross-reacted with the denatured band 2 protein, the cross-reacting epitope is predicted to be linear and could potentially be deduced by comparing primary amino acid sequences. Amino acid sequence comparison of HSV-1(F) and HSV-2(G) U_L27.5 and U_S5 revealed very limited sequence homology (Fig. 7). The sequence of the U_S5 gene predicts a hydrophobic protein. Therefore, the observation that the limited homology resided in a rather hydrophobic stretch of amino acids was not unexpected.

View larger version (54K): [in this window] [in a new window]   FIG. 6.   Amino acid sequence alignment of the proposed new ORF UL27.5 from HSV-1(F) (top line) and HSV-2(G) (middle line). The consensus sequence is shown on the bottom line. Numbers indicate the amino acid number based on the sequence of the HSV-2(G) ORF. Note the slight sequence variation between HSV-1(F) and HSV-2(G) from 855 to 934, which gives the HSV-2(G) UL27.5 ORF a total amino acid residue count of 985.

View larger version (30K): [in this window] [in a new window]   FIG. 7.   Amino acid sequence alignment between HSV-2(G) UL27.5 amino acids 701 to 847 (the second line), the equivalent sequence in HSV-1(F) UL27.5 (top line), and the entire coding sequence of HSV-1(F) US5 (the third line). The consensus sequence is shown at the bottom.

A GST-U_L27.5 fusion protein expressing the U_S5 homologous region of U_L27.5 reacted with the U_S5 polyclonal antibody by immunoblot analysis. A GST-U_L27.5 fusion protein expressing the homologous region of the HSV-1 U_L27.5 coding sequence (Fig. 6, HSV-1 amino acids 299 to 412) was constructed, purified on affinity columns, and subjected to electrophoresis in denaturing polyacrylamide gels. The eluted GST-U_L27.5 chimeric protein, the chimeric protein remaining bound to the affinity resin and eluted by solubilization in disruption buffer, and the eluted GST protein were each separated on a 15% denaturing gel and stained with Coomassie blue (Fig. 8A, lanes 1, 2, and 3, respectively). Portions of the same preparation were electrophoretically separated in the same gel, electrically transferred to a nitrocellulose membrane, and probed with U_S5 polyclonal antibody (Fig. 8B). The results were that the GST-U_L27.5 fusion protein migrated at an expected M_r of 42,000 on denaturing gel (panel A, lanes 1 and 2) (2). Both eluted GST-U_L27.5 and resin-associated GST-U_L27.5 reacted with the U_S5 polyclonal antibody on an immunoblot, whereas GST alone did not (Fig. 8B, lanes 1, 2, and 3). These results are consistent with the results of the mapping studies and indicate that U_L27.5 encodes a protein containing an epitope common to the U_S5 protein.

View larger version (42K): [in this window] [in a new window]   FIG. 8.   Photograph of gel stained with Coomassie blue (A) or immunoblot of proteins electrophoretically separated in an SDS-polyacrylamide gel and reacted with US5 polyclonal antibody (B). GST-UL27.5 fusion protein was induced as described in Materials and Methods. The eluent (lane 1), the remaining fusion protein bound to the affinity resin (lane 2), and the eluent of GST protein (lane 3) were electrophoretically separated in an SDS-15% polyacrylamide gel and stained with Coomassie blue (A). Equal amounts of each sample were separated in the same fashion, electrically transferred to a nitrocellulose membrane, and reacted with US5 polyclonal antibody (B). Molecular weight marker is shown at the left.

In vitro-translated product of the U_L27.5 coding sequence migrated as a protein with a molecular weight of 65,000. The primary sequence of the U_L27.5 ORF has the capacity to encode a protein of 575 amino acids. The apparent molecular weight of band 2 on denaturing polyacrylamide gel was 43,000. To test if the discrepancy between the predicted and observed size is due to the nature of the protein or to posttranslational modification, we translated the U_L27.5 open reading frame with a coupled in vitro transcription translation system and analyzed the protein in a denaturing gel. As shown in Fig. 9, lanes 3 and 4, the only [¹⁴C]leucine-labeled species, supposedly the full-length product, migrated with an apparent M_r of 65,000. A minor protein band with an apparent M_r of 65,000 was also observed in immunoblots reacted with the U_S5 antibody (Fig. 4).

View larger version (24K): [in this window] [in a new window]   FIG. 9.   Autoradiography of [14C]leucine-labeled in vitro-translated UL27.5 protein immobilized on a nitrocellulose membrane. Vector pGEM3Zf(+) (lanes 1 and 2) or pRB5169 containing the proposed HSV-1(F) UL27.5 coding sequence (lanes 3 and 4) was in vitro translated in a programmed rabbit reticulocyte lysate, labeled with [14C]leucine, electrophoretically separated in an SDS-15% polyacrylamide gel, electrically transferred onto a nitrocellulose membrane, and subjected to autoradiography. Lanes 1, 2, 3, and 4 represent four independent experiments.

U_L27.5 is a ₂ gene. In this series of experiments, replicate Vero cultures were either mock infected or infected with HSV-1(F) and either left untreated or incubated in medium containing phosphonoacetic acid (PAA) (300 µg/ml of medium; a gift of Abbott Laboratories) throughout the course of infection. The cells were harvested at 18 h after infection, solubilized in disruption buffer, subjected to electrophoresis, electrically transferred onto a membrane, and probed with appropriate antibodies to viral proteins. As shown in Fig. 10B, lanes 2 and 4, the treatment of PAA was effective, as evidenced by the reduced accumulation of ICP35 (a ₁ gene). In contrast to ICP0 and the ICP35 protein, U_L27.5 was not detected in infected cells treated with PAA (Fig. 10C). We conclude therefore that U_L27.5 is a ₂ gene totally dependent on viral DNA synthesis for its expression.

View larger version (32K): [in this window] [in a new window]   FIG. 10.   Photograph of immunoblot of electrophoretically separated infected-cell proteins reacted with indicated antibodies. Vero cells were mock infected or infected with HSV-1(F) in the presence or absence of PAA and harvested at 18 h postinfection. Lysates were electrophoretically separated in a denaturing polyacrylamide gel, transferred to a nitrocellulose membrane, and reacted with antibodies to ICP0 (A) and ICP35 (B) and the US5 polyclonal antibody (C).

U_L27.5 protein accumulates in the cytoplasm. BHK(TK+) cells were either mock infected or exposed to 10 PFU of HSV-1(F) or of HSV-2(G) per cell and incubated at 37°C. The cells were harvested at 18 h after infection. The cytoplasmic and nuclear fractions were collected as described in Materials and Methods, solubilized, electrophoretically separated onto a denaturing polyacrylamide gel, electrically transferred to a membrane, and reacted with the anti-U_S5 antibody. As shown in Fig. 11, lanes 1 and 2, the U_L27.5 protein accumulated exclusively in the cytoplasmic fraction. In HSV-2(G)-infected cell lysates, the majority of the signal was detected in the cytoplasmic fraction. A slightly more rapidly migrating protein band was detected in both the nuclear and cytoplasmic fractions whereas a less abundant and more rapidly migrating protein band was detected only in the nuclear fraction (Fig. 11, lanes 3 and 4).

View larger version (60K): [in this window] [in a new window]   FIG. 11.   Photograph of immunoblot of electrophoretically separated proteins reacted with US5 polyclonal antibody. BHK(TK+) cells were either mock infected or infected with HSV for 18 h. The cytoplasmic and nuclear fractions were collected as described in Materials and Methods, electrophoretically separated onto an SDS-15% polyacrylamide gel, electrically transferred to a nitrocellulose membrane, and reacted with US5 polyclonal antibody. Band 1, 2, and 3 proteins are indicated at the right.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

A polyclonal rabbit antibody made against gJ reacted with a protein expressed by both HSV-1 and HSV-2. The gene encoding this protein has been mapped to an ORF antisense to the gene encoding gB. In this article, we report on the mapping of the gene and preliminary characterization of the product of the ORF. The salient features of this report are as follows.

(i) The protein is encoded by the U_L27.5 gene on the basis of three kinds of data, i.e., mapping of two independently derived intertypic recombinants, analyses of insertion and deletion mutants, and direct reaction of the antibody with a chimeric protein containing the relevant domain of the ORF expressed in E. coli.

(ii) The U_L27.5 ORFs predict an HSV-1 protein of 575 amino acids and an HSV-2 protein of 985 amino acids. The HSV-1 sequence predicts a single methionine codon, whereas the HSV- 2 sequence predicts two methionines at positions 1 and 669. Translation in vitro of the entire HSV-1 ORF yielded a protein with an apparent M_r of 65,000. However, this observation does not prove that translation initiation occurs at the initiator methionine of the ORF, and there is no substantive evidence that the accumulating HSV-2 proteins are derived from a high-molecular-weight precursor. Initiation at the second methionine codon of the HSV-2 U_L27.5 open reading frame would predict a protein much smaller than the HSV-2 protein accumulating in the infected cell. We cannot at this time exclude the possibility that translation initiation begins internally from an alternate translation initiation codon (19).

(iii) Most of the HSV-1 and HSV-2 U_L27.5 accumulated in the cytoplasm. A small amount of HSV-2 U_L27.5 protein migrating faster than the cytoplasmic protein partitioned in the nucleus. We should note that the amount of HSV-2 protein accumulating in infected cells is higher or reacts better with the antibody than the corresponding HSV-1 protein. It is conceivable that a small amount of HSV-1 U_L27.5 also accumulates in the nucleus.

Because U_L27.5 maps antisense to a gene essential for viral replication, it has not been possible at this time to determine whether the U_L27.5 gene product is also essential. Experiments now in progress should allow us to assess the role of this gene in cell culture and in an experimental animal system.

(iv) The discovery of a gene antisense to gB (U_L27) was totally unexpected. Most analytical tools used to analyze nucleotide sequences are based on the assumption that a coding domain is present in the form of a linear array of nucleotides on one strand only. The gB-U_L27.5 pair of antisense genes is the third set discovered within the HSV-1 genome (6, 20, 26). The ease with which they have been discovered in the past few years suggests that there may be more such pairs. Given the fact that the size of the capsids is conserved and virtually identical for all herpesviruses and that the capsids could package >240 kbp of DNA (e.g., the HCMV genome), the question arises as to why HSV encodes genes antisense to each other rather than stringing these ORFs in linear arrays. Among the many possible explanations, three are worthy of further discussion.

The first, less-interesting hypothesis is that even the large herpesviruses contain genes arranged antisense to each other and that the actual number of genes in HCMV is grossly underestimated. It is conceivable that the antisense arrangements antedate the divergence of the primordial herpesvirus into the various subfamilies now in existence.

The second, more-attractive hypothesis is that the antisense arrangement is a form of regulation of gene expression that determines both the timing of synthesis and the abundance of the gene product. In the two preceding cases, that is, ₁34.5/ORF P and ORF O and U_L43/U_L43.5, we have found that the expression of the genes situated antisense to each other was sequential or even mutually exclusive (6, 20, 26). In this instance, gB is expressed very early in infection in abundant amounts, whereas the U_L27.5 gene expression appears to be a late event, dependent on viral DNA synthesis. One test of the hypothesis would be to reverse the timing of the expression of the two genes to determine whether early expression of the U_L27.5 is deleterious.

Lastly, the possibility that the sequences of two ORFs fit such that both proteins contain only the amino acid sequences essential for their function is probably remote. It is more likely that key amino acid domains of one protein correspond to neutral or linker domains in the product of the antisense ORF. Once the precise sequence encoding the U_L27.5 protein is elucidated, it may be possible to probe more accurately the corresponding domains of the gB gene.