Origin of Expression of the Herpes Simplex Virus Type 1 Protein U_S1.5

Abstract

ICP22, an immediate-early protein of herpes simplex virus type 1 (HSV-1), is required for viral replication in nonpermissive cell types and for expression of a class of late viral proteins which includes glycoprotein C. An understanding of the mechanism of ICP22 function has been complicated by the coexpression of the full-length protein with an in-frame, C-terminus-specific protein, U_S1.5. In this report, we confirm that the U_S1.5 protein is a bona fide translation product since it is detected during infections with three laboratory strains and two low-passage clinical isolates of HSV-1. To clarify the expression patterns of the ICP22 and U_S1.5 proteins, we examined their synthesis from plasmids in transient expression assays. Because previous studies had identified two different U_S1.5 translational start sites, we attempted to determine which is correct by studying the effects of a series of deletion, nonsense, and methionine substitutions on U_S1.5 expression. First, amino acids 90 to 420 encoded by the ICP22 open reading frame (ORF) migrated at the mobility of U_S1.5 in sodium dodecyl sulfate-polyacrylamide gels. Second, introduction of a stop codon downstream of M90 ablated expression of both ICP22 and U_S1.5. Finally, mutation of M90 to alanine (M90A) allowed expression of full-length ICP22 while dramatically reducing expression of U_S1.5. Levels of U_S1.5 but not ICP22 protein expression were also reduced in cells infected with an M90A mutant virus. Thus, we conclude that expression of IC22 and that of U_S1.5 can occur independently of each other and that U_S1.5 translation initiates at M90 of the ICP22 ORF.

The first genes of herpes simplex virus type 1 (HSV-1) to be expressed, the immediate-early (IE) genes, encode infected cell protein 0 (ICP0), ICP4, ICP22, ICP27, and ICP47, which prime the cell to promote efficient viral transcription and DNA replication. While four IE proteins (ICP0, -4, -27, and -47) have been studied extensively, the functional roles of ICP22, a 420-amino-acid (aa) phosphoprotein, in both productive infection and latency have not been determined. Elucidation of the functions of ICP22 has been hindered by several factors. First, ICP22 was originally described as nonessential for viral replication in cell culture and as such was largely ignored (22). Second, transient expression of ICP22 in the absence of other viral factors has proven to be highly inefficient and has been reported for only a few isolated studies (9, 23). Third, genetic analysis of ICP22 has been complicated by coexpression of the in-frame, C-terminal variant of ICP22, U_S1.5. The purpose of this study was to obtain information sufficient to develop techniques adequate for the genetic and functional analysis of ICP22 and U_S1.5 as separate entities.

Although information regarding the functions of ICP22 and/or U_S1.5 is limited, several physical properties have been described for the two overlapping proteins during productive infection. ICP22 is expressed from the U_s1 transcript, which initiates within the short internal repeat sequence of the HSV-1 genome (Fig. 1). The ICP22 open reading frame (ORF) is located entirely within the unique short (U_S) region of the genome. The U_S1.5 protein has been reported to be expressed from its own transcript, which initiates within the ICP22 ORF at codon 146 (6). Both ICP22 and U_S1.5 localize to the nuclei of infected cells, and ICP22 localizes to dense foci which costain for ICP4 and RNA polymerase II (9, 14). ICP22 migrates in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as multiple bands ranging in size from 65 to 90 kDa, much larger than the size predicted for the 420-aa protein (24). Similarly, U_S1.5 is also detected in multiple forms ranging in size from 40 to 50 kDa, once again larger than the 30 kDa predicted for the 249-aa protein (16). The multiple forms of both proteins and their reduced mobilities in SDS-PAGE are due in large part to extensive posttranslational modification by both viral and cellular enzymes. Infection of cells with viral mutants lacking the two viral Ser/Thr protein kinases (U_s3 and U_l13) results in a loss of most of the modified forms of ICP22 and U_S1.5, suggesting that both proteins are extensively phosphorylated by the two viral kinases (16, 24). Consistent with this hypothesis, direct phosphorylation of ICP22 by U_s3 (11) and U_L13 (1) has been demonstrated in vitro. In addition to being phosphorylated by viral kinases, ICP22 is phosphorylated on Tyr116 by an unknown cellular enzyme (5). It has also been reported that ICP22 is nucleotydilated by casein kinase II (3).

FIG. 1.

(A) Map of the HSV-1 genome showing the location of ICP22 and Us1.5. The positions of unique and repeat sequences are indicated by lines and boxes, respectively. The U_s1 transcript initiates within the internal repeat short (IR_s) segment of the genome and extends into unique short (U_s) sequences. (B) The 3.2-kb EcoRI-to-KpnI fragment that contains the U_s1 gene is shown. The published transcriptional start sites for the U_s1 and U_s1.5 transcripts are indicated by the arrows. The ICP22 and U_s1.5 ORFs (gray boxes) and their published start sites, M1 and M171, respectively, are indicated. Both ORFS are located entirely within U_s sequences.

Although little is known about the functions of ICP22 and U_S1.5, functional analysis of mutant viruses that fail to express both proteins has provided some insight into their roles in productive infection. The absence of both ICP22 and U_S1.5 in mutant viruses results in the failure of these viruses to replicate efficiently in a number of commonly used cell lines, including Rab-9, human embryonic lung cells, human foreskin fibroblasts, and rabbit skin cells but not Vero, Hep-2, or HeLa cells (17, 20, 26). From a molecular perspective, this cell type-specific defect in replication is coincident with a decrease in the rate of synthesis of viral transcripts, the failure to express wild-type (WT) levels of a subset of late viral proteins, and changes in the composition and density of mature extracellular virions (18, 24, 25). In addition, ICP22/U_S1.5 null viruses replicate poorly in the mouse ocular model during the acute phase of infection, establish latency with reduced efficiency, and exhibit reduced virulence (18, 19, 26).

While these multiple and various phenotypes are intriguing, the mechanism by which ICP22 facilitates WT replication is unclear. In addition, the individual contributions of ICP22 and/or U_S1.5 to each of the phenotypes are unknown. To begin to address these questions genetically and at the molecular level, expression and characterization of each protein independently of the other is necessary. Although previous studies have succeeded in expressing U_S1.5 in the absence of ICP22, all the ICP22-negative, U_S1.5-positive mutant viruses generated to date have been characterized only for the efficiency of expression of a subset of late viral genes and for the mobility of U_S1.5 in SDS-PAGE gels and not for any of the other properties ascribed to ICP22 mentioned above (16). In addition, the ICP22-negative, U_S1.5-positive mutant viruses were generated using cosmid libraries, and the integrity of the complete mutant viral genomes was not analyzed. Finally, mutant viruses that express ICP22 but not U_S1.5 were not generated or compared to mutants expressing only U_S1.5. These studies have been further complicated by the fact that the start site originally reported by Ogle and Roizman for U_S1.5, M147 (16) was later reported by the same group to be incorrect, and they reported that U_S1.5 actually initiates at M171 (21).

In order to resolve these issues and begin to elucidate the roles of ICP22 and U_S1.5 individually during infection, we utilized a recently described transient expression system (4) for ICP22 to study mutants lacking one protein or the other. Using this system, we have attempted to identify the bona fide start site of U_S1.5. Specifically, we introduced a panel of deletion, nonsense, and site-directed point mutations into the ICP22 ORF. We found that while U_S1.5 expression was not dependent on expression of full-length ICP22, mutation of M171 to alanine failed to alter U_S1.5 expression. More-extensive mutational analysis of the ICP22 ORF suggests that translation of U_S1.5 initiates at M90, much further upstream than indicated by all previous reports. Initiation at M90 would also preclude expression of U_S1.5 from the previously described U_S1.5 transcript, raising questions concerning the mechanism that is responsible for U_S1.5 translation.

MATERIALS AND METHODS

Cells and viruses.

Vero cells (ATCC CCL-81) were propagated in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (HyClone, Logan, UT), 100 mM penicillin-streptomycin (Invitrogen), and 2 mM l-glutamine (Invitrogen) at 37°C in 5% CO₂. The WT virus used in these studies was HSV-1 strain KOS at passage 8 from original isolation (27). All viruses used in the study were propagated in Vero cells by standard procedures (8).

Plasmids.

To map the translational start site of the U_S1.5 gene, a series of point and deletion mutant plasmids were constructed.

(i) Site-directed mutagenesis.

All point mutations were introduced into the pAlter22 plasmid by the Altered Sites mutagenesis system (Promega, Madison, WI) using the primers indicated in Table 1. Each mutation was designed to introduce or remove a unique restriction site so that mutant plasmids could be readily identified. The presence of each mutation was confirmed by restriction enzyme analysis and DNA sequencing.

TABLE 1.

Site-directed mutagenesis of the ICP22 ORF

Plasmid	Primer sequencea	Restriction site alterations
pAlter22:3xStop	GGCGCTTTTGCGCCTTGTGTATGATAATAGAAACCGCGGCGTCCCGCTCTC	SacII added
pAlter22:Stop75	GCCCCCCGCATCGGTTGACGTAGGGCCCCCCGG	HincII site added
pAlter22:Stop124	GACATTCCCCCACGACCCTAGCTAGCCCGGGTAAACCTGCG	NheI site added
pAlter22:Stop152	CCTAAGATGGGGCGGGTCCGGTAAACCCGGGAAACGCAGCCC	AccI site removed
pAlter22:M90A	GGCGGTTTTTTCTGGACGCGTCGGCGGAATCC	PciI site removed
pAlter22:M147A	GGATGGGGTTATTTTTCCTAAGGCGGGGCGGGTCCGGTCTACCCGG	Bsu36I site added
pAlter22:M171A	CGGCCCCAAGCCCAAATGCAGCCCTACGGCGCTCGGTGCGCCAGG	BsrdI site removed
pAlter22:M194A	CCCGACCTGGGCTACGCGCGCCAGTGTATCAATCAGC	NspI site removed

^aThe loss or gain of novel restriction sites is noted by underlining.

(ii) N-terminal deletion mutants.

Full-length and truncated ICP22 ORFs were PCR amplified from infectious KOS DNA (with the exception of pCDNA3:3xStop, in which pAlter22:3xStop was used as a template) and cloned into the pTOPO:blunt vector (Invitrogen) as per the manufacturer's instructions using forward primers (1, ATGGCCGACATTTCCCC; 90, ATGTCGGCGGAATCCACC; 147, ATGGGGCGGGTCCGGTC; 171, ATGCTCCGGCGCTCGGTG; 194, ATGCGCCAGTGTATCAATCAGC; and 228, ATGGGATACTGCCGAGCC) and the reverse primer (TCACGGCCGGAGAAACG). The PCR products were excised from pTOPO:blunt with EcoRI and subcloned into pCDNA3.1. The orientations of the inserts were determined by sizes following digestion with XhoI and separation by agarose gel electrophoresis. Constructs containing the ICP22 ORF or deletion products in the correct orientation were confirmed by DNA sequencing.

pTOPO:22ORF.

In order to make a 3′-specific ICP22 RNA probe, the ICP22 ORF was PCR amplified from infectious KOS DNA using primers to the 5′ (ATGGCCGACATTTCCCC) and 3′ (TCACGGCCGGAGAAACG) boundaries of the ICP22 ORF. The PCR product was gel purified using the Wizard SV gel and PCR cleanup system (Promega) and cloned into the pTOPO:blunt vector as per the manufacturer's instructions to make the pTOPO:22ORF vector. The orientation of the insert was determined by digestion with XhoI and separation by agarose gel electrophoresis. A clone with ICP22 in the correct orientation was amplified and cut with XhoI. The linearized plasmid was gel purified for use as a template for RNA probe synthesis using the Riboprobe combination system (Promega).

TOPO:ORF-PA.

In order to generate a plasmid that contains the mapped U_S1.5 promoter and transcriptional start site but not ICP22 promoter regulatory elements, the ICP22 ORF and 3′ regulatory sequences were PCR amplified using the primers ATGGCCGACATTTCCCC and CGGTTCCTGGCCTTTTGC. The PCR product was gel purified using the Wizard SV gel and PCR cleanup system (Promega) and ligated into the pTOPO:blunt vector as per the manufacturer's instructions (Invitrogen) to generate pTOPO:ORF-PA. The location and orientation of the PCR product were determined by restriction digestion and confirmed by DNA sequencing.

Construction of mutant KOS-ICP22:M90A.

In order to introduce the M90A mutation into the viral genome, Vero cells (3 × 10⁵) were transfected with 250 ng of infectious d22:GFP DNA, 100 ng of the ICP0 expression vector, pSH (ICP0 aids in viral gene expression and recombination), and 2.65 μg linearized pAlter22:M90A. At 48 h posttransfection (hpt), total virus was harvested and plated on Vero cell monolayers. Introduction of the M90A mutation into the viral genome by homologous recombination was detected by the loss of green fluorescent protein (GFP) expression in viral plaques. Individual white plaques were isolated and amplified on Vero cell monolayers. Total cellular DNA was isolated and recombination of ICP22-specific sequences confirmed by PCR. A single white plaque containing ICP22-specific sequences was plaque purified three times. Each of the plaques was isolated with virus from the previous passage that had been passaged through a 22-μm filter to ensure that each resulting plaque was initiated by a single virus particle. The final isolate was amplified, and the presence of ICP22-specific sequences and the desired mutation were determined by Southern blotting. Specifically, 3 × 10⁶ Vero cells were infected with 10 PFU/cell of KOS, d22:GFP, or KOS:ICP22:M90A. Twenty-four h later, total cellular DNA was isolated using the Qiaquick DNA purification system (Qiagen, Valencia, CA). One μg of each sample was digested with BamHI or BamHI and PciI overnight at 37C. The DNA was separated on a 1% agarose gel and transferred to nitrocellulose. ICP22-specific probes were synthesized using the Ready-to-go DNA labeling system (GE Healthcare, Piscataway, NJ) using linearized pCDNA3:ICP22 as a template. In addition, the ICP22 ORF was PCR amplified from mutant viral genomes and sequenced to ensure the location and orientation of the expected mutation.

Transfection.

Vero cells were seeded in six-well plates at a density of 3 × 10⁵ cells/well. Twenty-four h postseeding, DMEM was replaced with 2 ml Optimem (Invitrogen). Cells were transfected with Lipofectamine 2000 (Invitrogen) as per the manufacturer's instructions using 6 μl of Lipofectamine 2000 and 3 μg of DNA per well. Six h after addition of the transfection reagent, Optimem was replaced with 2 ml of DMEM and cultures were incubated until protein lysates were prepared as described for Western blots below.

Antibodies.

Antibodies 372 and 413, used throughout these studies, were manufactured by Bethyl Laboratories (Montgomery, TX). Antibody 372 was raised against the peptide corresponding to aa 372 to 384 of the ICP22 ORF. Antibody 413 was raised against the peptide corresponding to the last eight amino acids of the ICP22 ORF. Both antibodies were conjugated to keyhole limpet hemocyanin and generated in rabbits. Rabbit serum was harvested, and both antibodies were affinity purified by Bethyl Laboratories. The specificity of the purified antibody was tested by enzyme-linked immunosorbent assay by Bethyl Laboratories.

Western blots.

Vero cells were seeded in six-well plates at a density of 3 × 10⁵ cells/well. Twenty-four h postseeding, cells were transfected in Optimem with Lipofectamine 2000 as per the manufacturer's instructions (Invitrogen). Six h after addition of the transfection reagent, Optimem was replaced with 2 ml growth medium. Twenty-four hours hpt, cells were infected with 10 PFU/cell of the indicated viruses. Six h postinfection (hpi), cells were washed with phosphate-buffered saline and lysed in 250 μl RIPA buffer (25 mM Tris·HCl [pH 7.6], 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS) with rocking for 30 min at 4°C. Lysates were transferred to 1.5-ml Eppendorf tubes and clarified by centrifugation at 14,000 × g for 10 min. The supernatant fluid was transferred to a new tube containing 4× sample buffer (0.25 M Tris-HCl [pH 6.8], 8% SDS, 40% glycerol, 0.02% bromophenol blue, 10% beta-mercaptoethanol) and boiled for 3 min. Proteins in lysates equivalent to 2 × 10⁴ cells were separated by SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes. Membranes were blocked for 1 h at room temperature in TBST (25 mM Tris, pH 7.4, 3.0 mM KCl, 140 mM NaCl, and 0.1% Tween 20) with 5% nonfat milk. Blots were incubated with the ICP22-specific antibodies overnight in TBST with 5% nonfat milk at 4°C at dilutions of 1:250, 1:2,000, and 1:10,000 for the antibodies N-term, 372, and 413, respectively. The next morning, membranes were washed three times by rocking them for 10 min with TBST at room temperature. Goat anti-rabbit conjugated to horseradish peroxidase secondary antibodies (Jackson Laboratories, Bar Harbor, Me) were diluted 1:25,000 in TBST plus 5% nonfat milk and incubated with the blots for 45 min at room temperature. Blots were washed six times with TBST for 30 min per wash, treated with Millipore (Millipore, Bedford, MA) ECL reagent, and exposed to X-ray film (Pierce, Rockford, IL).

Northern blots.

In order to detect ICP22 and U_S1.5 transcripts in infected cells, Northern blotting was performed essentially as described by Lee and Schaffer (13). Briefly, 3 × 10⁶ Vero cells were plated in 100-mM dishes. Twenty-four h later, cells were mock infected or infected with d22:GFP, KOS, or strain F as indicated. Total RNA was harvested at the time indicated using Trizol (Invitrogen) as per the manufacturer's instructions. RNA was separated in 1% agarose gels and transferred to nitrocellulose. Blots were UV cross-linked, dried, and blocked for 1 h at 60°C. 3′-specific ICP22 RNA probes were generated using the Promega Riboprobe combination system as per the manufacturer's instructions. The template, pTOPO:22ORF, was linearized using XhoI. The probes were hybridized to blots overnight at 68°C and washed as reported previously (13). Blots were exposed to a PhosphorImager screen (Molecular Dynamics) and data analyzed using Imagequant software (Molecular Dynamics).

RESULTS

U_S1.5 expression is conserved among multiple HSV-1 strains and low-passage clinical isolates.

At the time we initiated these studies, U_S1.5 had been reported to exist only in strain F-infected cells. In order to determine if U_S1.5 expression is conserved among a series of HSV-1 strains, we generated two new antibodies specific for peptides within the C terminus of ICP22 corresponding to aa 372 to 385 and 413 to 420. We then tested and optimized the binding and washing conditions of the antibodies for reactivity with ICP22 by Western blot analysis using protein lysates generated from strain F-infected Vero cells. Both antibody 372 and antibody 413 detected full-length ICP22, as well as faster-migrating bands similar to those described as U_S1.5 (data not shown).

Having confirmed the synthesis of U_S1.5 in strain F-infected Vero cells with the new C-terminus-specific antibodies, we asked whether U_S1.5 expression is conserved among many strains of HSV and is not an artifact of genome rearrangement specific to strain F. For this purpose, replicate cultures of Vero cells were infected with 10 PFU/cell of KOS, strain F, strain 17, low-passage isolates Seibert and Gayle, or as a negative control, mutant virus d22:GFP, in which the ICP22 gene is replaced by a GFP cassette. At 6 hpi, whole-cell lysates were prepared and expression of ICP22 and U_S1.5 was examined with rabbit polyclonal antibodies. As shown in Fig. 2A, using the N-terminus-specific antibody, multiple forms of ICP22 were detected in lysates of cells infected with all three laboratory strains and both clinical isolates of HSV-1 but not in lysates of the ICP22 null virus, d22:GFP. A unique, faster-migrating, N-terminus-specific band was also detected in strain 17-infected cells. With the C-terminus-specific antibody, 413, in addition to the bands corresponding to ICP22, three to four faster-migrating bands ranging in size from 40 to 60 kDa were also detected in all laboratory strains and clinical isolates but not in d22:GFP-infected cells (Fig. 2B). These faster-migrating bands, corresponding to U_S1.5, were also detected with the other C-terminal antibody, 372 (data not shown). As expected, the faster-migrating bands were not detected when the blots were probed with an antibody specific for the N terminus of ICP22, but the slower-migrating bands corresponding to ICP22 were (Fig. 2A). These findings demonstrate that U_S1.5 expression is conserved among different HSV-1 strains and primary isolates in that the bands synthesized in infected cells are of similar number and size.

FIG. 2.

The U_s1.5 protein is expressed by multiple HSV-1 strains. Vero cells were either mock infected or infected with 10 PFU/cell of the viruses indicated. At 6 hpi, whole-cell lysates were prepared and probed for ICP22 and U_s1.5 using the ICP22 N-terminus-specific antibody (A) or the C-terminus-specific antibody, 413 (B).

Transient expression of the ICP22 and U_S1.5 proteins.

One of the major hurdles in studies of ICP22 has been the difficulty in demonstrating transient expression of this protein. In the absence of a transient expression system, researchers have constructed mutant viruses without prior screening of the mutations for alterations in ICP22 structure and function. This added layer of complexity has limited the number of well-characterized ICP22 mutant viruses available for study. Several procedures devised to express ICP22 in transient assays have been reported recently (4), although reagents capable of detecting expression of U_S1.5 using these procedures were not available. In order to determine if U_S1.5 can be expressed in transient assays in the presence of other viral proteins, replicate cultures of Vero cells were transfected with pAlter22, a plasmid containing the entire U_s1 gene (Fig. 3A), or as a negative control a GFP expression vector (pCMV:GFP). Twenty-four h later, the cells were mock infected or infected with KOS or d22:GFP as appropriate (Fig. 3B). Six h after infection, whole-cell lysates were prepared, and proteins were separated by SDS-PAGE, transferred to PVDF membranes, and probed with the C-terminal antibody 413. While ICP22 and U_S1.5 were not detected in cells transfected with pAlter:22 alone (Fig. 3B), both proteins were detected when transfected cells were infected with an ICP22-minus virus, d22:GFP. The bands reactive with antibody 413 migrated in a manner similar to that of those detected in KOS-infected samples (Fig. 2 and 3B). These observations demonstrate that both ICP22 and U_S1.5 can be expressed transiently following infection with an ICP22 null virus, i.e., in the presence of other viral proteins. Consequently, this procedure should be suitable for physical and functional characterization of mutant ICP22 proteins.

FIG. 3.

Expression of ICP22 and U_s1.5 from pAlter22, M171A, M194A, and 3×Stop mutants. (A) Predicted effects on ICP22 and U_S1.5 expression as a result of mutations introduced into U_S1 by pAlter mutagenesis. (B) Vero cells were transfected with the plasmids indicated. At 18 hpt, cells were infected with 10 PFU/cell of d22:GFP or KOS as indicated. At 6 hpi, whole-cell lysates were prepared and ICP22 and U_s1.5 levels measured by Western blotting using antibody 413. Overexposure of the U_s1.5 bands is also shown.

Separate expression of ICP22 and U_S1.5.

In order to express ICP22 independently of U_S1.5 and vice versa, the series of mutant expression plasmids shown in Fig. 3A was generated. To eliminate expression of ICP22, three in-frame stop codons were introduced shortly after the recognized ICP22 translational start site in pAlter22:3xStop. In order to eliminate expression of U_S1.5, M171, the reported U_S1.5 translational start site (21), was mutated to alanine. As a negative control, M194 was also mutated to alanine. Vero cells were transfected with the plasmids shown in Fig. 3B, and 18 h later, cultures were infected with 2.5 PFU/cell of either KOS or d22:GFP. At 6 hpi, whole-cell lysates were prepared and analyzed for ICP22 and U_S1.5 expression by Western blotting. ICP22 was detected in pCMV:GFP-transfected, KOS-infected samples and in cells transfected with pAlter22, M171A, and M194A and infected with d22:GFP (Fig. 3B). ICP22 was not detected in cells transfected with the 3xStop mutant, demonstrating that introduction of the three stop codons eliminated ICP22 expression. U_S1.5 was detected in pCMV:GFP-transfected cells infected with KOS, albeit at low levels, and in all d22:GFP-infected cultures that had been transfected with an ICP22 expression vector. Expression of U_S1.5 in cells transfected with pAlter22:3xStop, which does not express ICP22, argues strongly that U_S1.5 is a discrete translational product whose expression is independent of ICP22 expression. Transfection of neither M171A nor M194A eliminated expression of any of the U_S1.5-specific bands, suggesting that neither of these methionines serves as the translational start site of U_s1.5 in strain KOS.

Transient expression of N-terminal deletion mutants of ICP22.

Although U_s1.5 expression was shown to be independent of expression of full-length ICP22, mutation of the published start site, M171, had no effect on U_s1.5 expression, ruling out M171 as the actual U_s1.5 start site. In order to identify the true translational initiation site of U_s1.5, three strategies were used. The first was to compare the electrophoretic mobilities of the putative U_s1.5 bands expressed by N-terminal deletion mutants of ICP22 with those of bands detected in pAlter22-transfected, d22:GFP-infected cells. To this end, a panel of N-terminal ICP22 deletion mutants was constructed by PCR and cloned into pCDNA3 (Fig. 4A). Plasmids expressing full-length ICP22 (pCDNA3:22ORF) and ICP22 containing the 3× stop mutation (pCDNA3:3xStop) were included as positive and negative controls, respectively. Vero cells were transfected with the plasmids indicated in Fig. 4B. Twenty-four hpt, cells were infected with 10 PFU/cell d22:GFP, and 6 h later, whole-cell lysates were prepared and proteins separated by SDS-PAGE, transferred to PVDF, and probed with antibody 413. As expected, Vero cells transfected with pCMV:GFP did not express ICP22. In contrast, cells transfected with pCDNA3:ICP22ORF expressed high levels of ICP22. Vero cells transfected with pCDNA3:3xStop failed to express ICP22, as expected, but did express smaller, C-terminus-specific proteins that migrated with the same mobility as protein detected in WT transfected/infected cells (Fig. 4C and data not shown). These bands correspond in number and size to the bands specified by U_s1.5. Vero cells transfected with the N-terminal deletion plasmids expressed progressively smaller, faster-migrating proteins (with the exception of the 171-420 mutant, which failed to express detectable levels of protein) as increasing amounts of the N terminus of ICP22 were deleted. Notably, deletion of the first 89 amino acids of ICP22 resulted in expression of proteins that comigrated with those detected in cells transfected with pCDNA3:3xStop. Since M90 is the only methionine in this region of the protein, this observation suggests that initiation of translation of U_s1.5 occurs at M90.

FIG. 4.

Transient expression of N-terminal deletion mutant plasmids of ICP22. (A) Maps of the full-length ICP22 gene and ICP22 mutants subcloned into pCDNA3. The coding sequences expressed by each mutant are indicated by the gray box. The sequences deleted in each construct are indicated by the white boxes. (B) Vero cells were transfected with the plasmids indicated. pCDNA3 clones drive expression from the cytomegalovirus IE promoter. At 24 hpt, cells were infected with 10 PFU/cell of d22:GFP. At 6 hpi, total cell lysate was prepared and probed for ICP22. (C) Vero cells were transfected with the N-terminal ICP22 deletion plasmid panel, pAlter22 and pAlter22:3xStop. pAlter22-derived clones drive expression of the ICP22 protein from the ICP22 promoter. Expression of ICP22 and U_s1.5 was determined as described in Fig. 4B. (D) Vero cells were transfected with the pCDNA3 constructs indicated and probed for ICP22 expression as in Fig. 4B.

To further test this possibility, the experiment was repeated to include expression of ICP22 from its native promoter in pAlter22 and pAlter22:3xStop. Vero cells were transfected with the plasmids indicated in Fig. 4C and infected with d22:GFP as before. ICP22 was not detected in the GFP control sample but was readily detected in cells transfected with pAlter22 or pCDNA3:22ORF. In addition, a faint signal corresponding to the location of U_s1.5 was detected in cells transfected with pAlter22. This band comigrated with the faster-migrating form of the C-terminal protein bands expressed in cells transfected with pAlter22:3xStop (Fig. 4C). Expression of the N-terminal deletion mutants yielded proteins that migrated progressively faster in the gel. Notably, expression of aa 90 to 420 yielded proteins that comigrated with the faster-migrating band detected in pAlter22- and pAlter22:3xStop-transfected cells, suggesting that proteins expressed in pCDNA3:90-420 are the same size as the WT U_S1.5 protein. These data also indicate that the preferred U_S1.5 translational start site is located at M90.

Finally, although expression of ICP22 from pCDNA3:171-420 was below the limit of detection in Fig. 4B and C, we were able to detect expression of proteins from this plasmid in some experiments. Since the start site for U_S1.5 has been reported to be M171, peptides expressed from M171 were compared with those initiated at M90. In order to minimize the amount of signal on the blot, the experiment was repeated with a smaller panel of deletion plasmids. Vero cells were transfected with the plasmids indicated in Fig. 4D. At 24 hpt, cells were infected with 10 PFU/cell of d22:GFP, and 6 h later, whole-cell lysates were prepared and probed for ICP22 and U_S1.5 expression. As observed previously, the full complement of ICP22-specific bands was detected in cells transfected with pCDNA3:22ORF. In addition, U_s1.5-specific bands were detected in both the 3xStop and 90-420 expression vectors. Expression of C-terminal proteins was also detected in cells transfected with pCDNA3:171-420, albeit at much lower levels than in cells transfected with the other ICP22 expression vectors. Moreover, these bands migrated considerably faster than those detected in lysates of cells transfected with pCDNA3:3xStop or pCDNA3:90-420, implying that aa 171 to 420 are too small to be WT U_s1.5. It is unclear at this time whether expression of ICP22 peptides from aa 171 to 420 is inefficient or if the peptides are unstable.

Nonsense mutation mapping of the ICP22 ORF.

As a second means of identifying the U_s1.5 translational start site, stop codons were introduced into the ICP22 ORF at positions 75, 124, and 152 using pAlter mutagenesis, as shown in Fig. 5A. pAlter22:3xStop and p199, a plasmid containing a linker insertion with multiple stop codons at position 199, were included as controls. Vero cells were transfected with the plasmids indicated in Fig. 5B and infected with d22:GFP as described previously. As before, the complete complement of ICP22 bands was detected in cells transfected with pAlter22 but not in cells transfected with GFP or the 3xStop mutant. In addition, introduction of stop codons at 75, 124, 152, and 199 abrogated expression of full-length ICP22. The multiple forms of U_s1.5 were detected at a low levels in cells transfected with pAlter22, consistent with the previous finding. U_s1.5-specific bands were detected at elevated levels in cells transfected with 3xStop and Stop75. In contrast, U_s1.5-specific bands were not observed when stop codons were introduced at position 124, 152, or 199. These data strongly suggest that translation of U_s1.5 initiates between positions corresponding to aa 76 and 123 of the ICP22 ORF. The only in-frame methionine within this region is M90, supporting results obtained with the N-terminal deletion mutant panel (Fig. 4).

FIG. 5.

Nonsense mapping of the ICP22 ORF. (A) Diagram of the ICP22 ORF. The positions of the in-frame methionine codons are indicated, as are the positions of the introduced stop codons (bars). (B) Vero cells were transfected with the plasmids indicated. At 24 hpt, cells were infected with 10 PFU/cell of d22:GFP. At 6 hpi, total cell lysates were prepared and probed for ICP22 with the 413 antibody by Western blotting.

M → A substitution mutations.

As a third approach to identifying the translational start site of U_S1.5, individual methionines potentially able to initiate a protein of similar size to that detected in infected cells were mutated. To this end, M90, M147, M171, and M194 were mutated to alanine using pAlter mutagenesis. Vero cells were transfected with the plasmids indicated in Fig. 6 and infected with d22:GFP as described previously. As before, full-length ICP22 was detected in cells transfected with pAlter22 but not in cells transfected with GFP or any of the mutants containing stop codons. ICP22 expression was not affected by mutation of the individual methionines, indicating that the mutations did not affect ICP22 expression or stability. U_s1.5 was detected in pAlter22-transfected cells and in cells transfected with 3xStop and Stop75, confirming that U_S1.5 initiates downstream of codon 75 of the ICP22 ORF. Mutation of M147, M171, and M194 had no effect on the expression of U_s1.5, whereas levels of U_S1.5 detected in cells transfected with the M90A mutant were greatly reduced. In addition, the bands detected in cells transfected with M90A migrated slightly faster than those expressed in cells transfected with WT pAlter22. The source of these faster-migrating bands is unclear; however, these findings suggest that M90 is the preferred site for initiation of WT U_s1.5 translation.

FIG. 6.

Identification of the U_S1.5 translational start site by M → A substitutions. Vero cells were transfected with the plasmids indicated. At 24 hpt, cells were infected with 10 PFU/cell of d22:GFP. Six h later, total cell lysates were prepared and probed for ICP22 by Western blotting. Overexposure of the gel to emphasize changes in U_s1.5 expression is shown below.

Although mutation of M90 altered U_s1.5 mobility and levels, it did not completely eliminate its expression. In order to determine if mutation of M90 allows for initiation at the next downstream methionine, an M90A/M147A double mutant was constructed and tested for expression of U_s1.5. As before, mutation of M90 to alanine resulted in alteration of the levels and mobility of U_s1.5. Mutation of both M90 and M147 did not result in a further loss of the U_s1.5 signal or alteration of U_s1.5 mobility, indicating that translation of the faster-migrating proteins does not initiate from the next downstream methionine (data not shown). Two cryptic translational start sites are also located downstream of M90, and efforts to mutate these sites to determine if they might be used when M90 is not available are under way.

Construction and analysis of an ICP22 M90A mutant virus.

In order to determine if the transient expression assay system described here reflects what occurs when the mutations are introduced into the virus, we introduced the M90A mutation into the KOS genome (Fig. 7). For this purpose, infectious d22:GFP DNA was cotransfected with the M90A mutant plasmid. Recombinant plaques were initially screened for the loss of GFP. The recombinant virus, KOS:ICP22:M90A, was plaque purified three times and the presence of the mutation in the correct location and orientation confirmed by Southern blotting. Specifically, digestion of WT or M90A viral DNA with BamHI resulted in a 5-kb fragment which contains the ICP22 ORF (Fig. 7A). Probes specific for the ICP22 ORF detected a 5-kb fragment in both KOS- and M90A virus-infected Vero cells but not in cells infected with d22:GFP (data not shown). To confirm the presence of the M90A mutation, DNA isolated from Vero cells infected with KOS, d22:GFP or KOS:ICP22:M90A was digested with BamHI and PciI and probed for ICP22 sequences by Southern blotting (Fig. 7B). As predicted, digestion of KOS resulted in the formation of two fragments; however, the mutation in M90A, which mutates the PciI site, yielded a 5-kb fragment, indicating that the mutation is indeed present. In addition, the ICP22 ORF was PCR amplified and sequenced, confirming that the M90A mutation is present. Unfortunately, the sequencing reaction identified a second point mutation in the ICP22 ORF that mutated alanine 137 to glycine. Although we were concerned about the effect of introduction of this second mutation, we decided to proceed with screening KOS:ICP22:M90A for expression of ICP22 and U_S1.5.

FIG. 7.

Construction and characterization of KOS:ICP22M90A. (A) Restriction maps of KOS and KOS:ICP22M90A. The positions of the ICP22 ORF and relevant methionines are noted. (B) Genotype of KOS:ICP22M90A. DNA isolated from KOS and KOS:ICP22:M90A was digested with the enzymes indicated and probed by Southern blotting. (C) Expression of ICP22 and U_S1.5 in KOS:ICP22M90A-infected Vero cells. Vero cells were infected with the viruses indicated at an MOI of 10. At 24 hpi, whole-cell lysates were prepared and levels of ICP22 and U_S1.5 measured by Western blotting. An overexposure of the U_S1.5 bands is presented below.

In order to characterize U_S1.5 expression in KOS:ICP22:M90A-infected cells, Vero cells were mock infected or infected with KOS, d22:GFP, or KOS:ICP22:M90A. At 24 hpi, whole-cell lysates were prepared and levels of ICP22 and U_S1.5 measured by Western blotting (Fig. 7C). As expected, neither ICP22 nor U_S1.5 was detected in mock-infected or d22:GFP-infected cells. In contrast, both ICP22 and U_S1.5 were detected in KOS-infected cells. Similarly, ICP22 was detected at similar levels in KOS:ICP22:M90A-infected cells, however, U_S1.5 levels were markedly reduced. In addition, the smaller bands detected in M90A virus-infected cells migrated faster than the bands detected in WT-infected cells. Similar results were observed when lysates were harvested at 6 hpi (data not shown). The reduced levels and the change in size of U_S1.5 are both consistent with the observations made in transient assays and strongly argue that M90 is the preferred translational start site for the U_S1.5 protein. Initial experiments addressing the effect of a loss of U_S1.5 expression suggest that U_S1.5 is not necessary for replication in restrictive cells (data not shown); however, full characterization will be done once we have isolated a clean M90A mutant virus.

Efforts to detect the U_S1.5 transcript.

Since M90 is upstream of the mapped U_S1.5 transcriptional start site, we were curious as to whether we could detect the U_S1.5 transcript. To begin this study, we looked for expression of ICP22 and U_S1.5 RNAs using a previously described ICP22 RNA probe, 22SS (28). This probe is specific for the 5′ half of the ICP22 transcript but contains significant (more than 200 bp) overlap with the U_S1.5 transcript mapped by Carter and Roizman (6). In order to detect the U_S1.5 transcript, we analyzed RNA isolated from Vero, Rab-9, HeLa, PC12 cells and primary cortical neurons infected with KOS by Northern blotting. We were unable, however, to detect any RNA species migrating faster than the ICP22 transcript (data not shown).

Because our initial efforts to detect the U_S1.5 transcript failed, we decided to reproduce the conditions originally reported by Carter and Roizman more closely. In order to increase the specificity and sensitivity of detection, we constructed a new RNA probe specific for the 3′ end of the ICP22 transcript. Vero cells were mock infected or infected with 20 PFU/cell of d22:GFP, KOS, or strain F. At 10 hpi, total cellular RNA was isolated and 30 μg separated on formaldehyde gels. The RNA was transferred to nitrocellulose and probed for RNA specific to the 3′ end of the ICP22 ORF (Fig. 8A). As expected, neither the ICP22 nor the U_s1.5 transcript was detected in either mock-infected or d22:GFP-infected samples. The primary ICP22 transcript and two previously described longer transcripts (2) were detected in cells infected with either KOS or strain F. In contrast, faster-migrating, ICP22-specific bands were not detected, even after overexposure of the Northern blot (lower panel). Similar results were obtained when RNA was isolated from infected Rab-9 cells or when replicate blots were probed with the 22SS probe (data not shown).

FIG. 8.

Analysis of ICP22 and U_S1.5 transcripts isolated from HSV-infected cells. (A) Vero cells were infected with the viruses indicated and total cellular RNA harvested at 10 hpi. Thirty μg of RNA was separated on formaldehyde gels and probed for ICP22 and U_S1.5 transcripts with an RNA probe specific for the region of the ICP22 ORF corresponding to the 3′ end. (B) Northern blot analysis of RNA isolated from Vero cells infected with the viruses indicated in the presence or absence of CHX. Molecular weight markers are noted for both blots. Overexposures of the blots are also presented. (C) Expression of U_S1.5 from pAlter22 and pTOPO:ORF-PA. Vero cells were transfected with the plasmids indicated and infected with d22GFP or KOS as noted. Levels of ICP22 and U_S1.5 were determined by Western blotting.

Since Carter and Roizman noted they detected the U_S1.5 transcript in the presence of cycloheximide (CHX) (reported as data not shown) and mapped the transcriptional start site in the presence of CHX, we attempted to determine if we could detect the U_S1.5 transcript under these conditions. To that end, Vero cells were mock infected or infected with d22:GFP, KOS, or strain F in the presence or absence of CHX. At 6 hpi (the same time RNA was harvested by Carter and Roizman), total cellular RNA was isolated and levels of ICP22 and U_S1.5 transcripts were measured by Northern blotting. As shown in Fig. 8B, no bands were observed with the ICP22-specific RNA probe in mock-infected or d22:GFP-infected cells. As expected, the primary ICP22 transcript was detected in KOS- and strain F-infected cells in the presence or absence of CHX. Interestingly, the slower-migrating bands observed with the ICP22-specific RNA probe were also detected in the presence of CHX, arguing that these transcripts are expressed without the need for viral protein synthesis. Similar to results observed at 10 hpi, faster-migrating, ICP22-specific bands were not detected in mock-treated cells. In contrast, overexposure of the blot revealed several faster-migrating bands in both strain F- and KOS-infected cells that had been treated with CHX. The levels and number of bands detected, however, make it difficult to conclude that they are the U_S1.5 transcript. Rather, it is just as likely that they could be stable breakdown intermediates of the full-length ICP22 transcript.

Having detected the U_S1.5 protein but not the U_S1.5 transcript in infected cells, we sought to detect the U_S1.5 transcript indirectly by looking for U_S1.5 protein expression in the absence of the ICP22 promoter. To this end, the full-length ICP22 ORF and 3′ regulatory sequences, including the polyadenylation signal, were PCR amplified from the viral genome and cloned in the pTOPO:Blunt vector to make pTOPO:ORF-PA. This insert contains the entire ICP22 ORF as well as the reported U_S1.5 promoter and 3′ regulatory sequences but lacks ICP22 promoter or regulatory sequences. It is predicted that transfection of cells with pTOPO:ORF-PA will not express full-length ICP22 since the ICP22 promoter is not present; however, since the U_S1.5 promoter and 3′ regulatory sequences are present, the U_S1.5 protein should be expressed.

To test this hypothesis, Vero cells were transfected with the plasmids indicated in Fig. 8C. At 24 hpt, cells were infected with KOS or d22:GFP as indicated, and 6 h later, whole-cell lysates were prepared and probed for ICP22 and U_S1.5 by Western blotting. As expected, infection of Vero cells with KOS resulted in expression of both ICP22 and U_S1.5. Similarly, infection of Vero cells transfected with pAlter22 with d22:GFP resulted in the expression of both full-length ICP22 and U_S1.5. In contrast, however, ICP22 and U_S1.5 were not detected in cells transfected with a GFP expression vector or with pTOPO:ORF-PA. This observation suggests that expression of U_S1.5 is dependent on regulatory elements upstream of the mapped U_S1.5 promoter and argues that U_S1.5 may not be expressed from a transcript that initiates within the ICP22 ORF. Collectively, these observations are concerning, since they raise the question as to whether the U_S1.5 transcript actually exists.

DISCUSSION

In this report, we have demonstrated that expression of the U_S1.5 protein is conserved among three commonly used laboratory strains and two low-passage clinical isolates of HSV-1. In addition, we have shown that U_S1.5 can be expressed by multiple means in transient transfection/infection assays. U_S1.5 can also be detected when ICP22 expression vectors are cotransfected with an ICP0 expression vector (data not shown). In order to separate the expression of ICP22 and that of U_S1.5, we introduced site-specific mutations into the ICP22 ORF. Introduction of nonsense mutations downstream of the ICP22 start codon resulted in the loss of ICP22 expression but enhanced levels of U_S1.5 expression, demonstrating that U_S1.5 expression is not dependent on expression of full-length ICP22. To our surprise, mutation of the reported U_S1.5 start site, M171 (21), to alanine did not alter expression of U_S1.5, indicating that its translation is initiated at an alternative site. We employed three different mutational strategies to determine the actual translational start site of U_S1.5. First, we examined the mobility of sequential deletions of the N terminus of ICP22 and found that peptides containing aa 90 to 420 encoded by the ICP22 ORF migrated with mobility similar to that of the WT U_S1.5 protein. Second, introduction of a stop codon at position 124 of the ICP22 ORF but not a stop at 75 resulted in a loss of U_S1.5 expression, indicating that U_S1.5 translation initiates between positions 76 and 124 of the ICP22 ORF. Third, mutation of the only methionine located between 75 and 124, M90, to alanine resulted in a large reduction and increased mobility of the faster-migrating, C-terminus-specific bands. In order to ensure that our transient expression system reflects events in infected cells, we introduced the M90A mutation into the viral genome. Relative to WT virus, infection of Vero cells with ICP22:M90A produced reduced levels of U_S1.5, which migrated more quickly in SDS-PAGE gels, confirming the observations made in transient assays. Finally, since M90 is upstream of the published U_S1.5 transcriptional start site, we attempted to determine if a bona fide U_S1.5 transcript is expressed in infected cells. To date, we have been unable to detect a faster-migrating RNA specific for the 3′ half for the ICP22 ORF, suggesting U_S1.5 translation may initiate by an alternative mechanism.

U_S1.5 translational start site.

While the published data regarding the start site of U_S1.5 are confusing, our findings argue that translation initiates further upstream than previously reported. One explanation for this difference is that U_S1.5 expression is different among different HSV-1 strains, since our work has utilized KOS almost exclusively whereas previous data were generated using strain F-infected cells. We believe this is unlikely since the amino acid sequence between the two strains is 100% conserved. In addition, U_S1.5 detected in KOS, strain F, strain 17, and two low-passage clinical isolates all migrated with similar mobilities. If U_S1.5 initiated at different sites in different strains, a shift in mobility would be expected, yet no shift was observed.

Alternatively, it is possible that the original data used to map the U_S1.5 protein may have been misinterpreted. Original mapping data indicated that U_S1.5 initiates at M147 of the genome based on the mobility of U_S1.5 expressed from three ICP22 mutant viruses, R7805, R7808, and R7815 (16). U_S1.5 expressed from R7805 migrated at the same mobility as U_S1.5 expressed in WT strain F-infected cells. It was concluded that M147 was the preferred start site, since R7805 should not contain M90 of the ICP22 ORF and therefore M147 would be the first available, in-frame methionine. In addition, the ICP22 promoter but not the mapped U_S1.5 promoter was reported to be deleted in R7805. Thus, it was reasoned that expression of U_S1.5 from R7805 would have to be from the U_S1.5 promoter. The method of construction of R7805 raises some concern regarding these conclusions, since it was constructed by transfecting an incomplete cosmid library lacking the sequences believed to be deleted in the final isolate. It was assumed that the fragments adjacent to the sequences not represented in the cosmid library were joined end to end to create the final isolate. Southern blots of the R7805 genome presented by Ogle and Roizman (16) indicate that the ICP22 ORF-containing BamHI fragment of R7805 is smaller than that of the WT, indicating that some sequence was deleted; however, no size markers were presented on the gel to indicate the extent of the deletion. In addition, the final isolate was not reported to have been sequenced. Without a more complete genotyping of R7805, it is difficult to believe that the mutant virus genotype is as predicted. Consequently, these observations are one source of confusion surrounding identity of the U_S1.5 start site.

Although U_S1.5 expression from cells infected with R7815 was not compared to that from WT-infected cells, it was compared side by side with U_S1.5 expressed in R7805-infected cells (16). U_S1.5 expressed in R7815-infected cells comigrated with U_S1.5 expressed in R7805-infected cells and, by extension, with that in WT-infected cells. The first available, in-frame methionine in R7815 is M90. The authors concluded that M90 was skipped and M147 was the preferred start site. U_S1.5 expressed in R7808-infected cells initiated at M171 of the ICP22 ORF and migrated faster in SDS-PAGE than U_S1.5 expressed in WT-infected cells. In addition, U_S1.5 expressed in R7808-infected cells failed to be modified as extensively as the WT protein. Although these observations would indicate that M171 is not a candidate start site of U_S1.5, in a later report, Poon et al. stated (as data not shown) that M171 was the actual start site (21). It is unclear experimentally how this conclusion was reached; however, M171 is the first methionine available to initiate translation from the mapped U_S1.5 transcript (6). Our data strongly suggest that neither M147 nor M171 is the U_S1.5 translational start site, but rather, translation initiates at the position of the ICP22 ORF corresponding to M90. This is consistent with data presented when cells infected with the R7815 mutant were analyzed.

Although the data presented herein indicate that U_S1.5 most likely begins to be translated at M90, mutation of M90 to alanine did not completely eliminate detection of faster-migrating bands using a C-terminus-specific antibody. Notably, multiple mechanisms could lead to expression of these bands. We tested the hypothesis that they were the result of initiation at the next available methionine, M147, but mutation of both M90 and M147 did not alter expression of the faster-migrating bands. This experiment does not rule out the possibility that translation is initiated at cryptic translational start sites, however, and at least two potential cryptic start sites, codons 103 and 130, are located between the M90 and M147 positions. Efforts to mutate these sites alone and in combination with the M90 site are under way. An alternative explanation is that the bands detected in M90A mutant virus-transfected cells are cleavage products of the full-length ICP22 gene. In fact, Munger et al. reported that ICP22 is cleaved by caspase 3 in ICP4-minus virus-infected cells to yield a stable 38-kDa C-terminal protein (15). Although this cleavage was observed only in ICP4-null virus-infected cells, U_S1.5 was expressed to WT levels and thus may mask cleavage products observed in WT-infected cells. By eliminating U_S1.5 expression, we may be detecting a low level of cleavage that occurs in the presence of ICP4. Tests to determine if this is the case are ongoing. Finally, a third possibility is that translation initiates upstream of M90 and that mutation of M90 to alanine destabilizes the protein, resulting in the reduced levels and faster migration. Although we cannot rule out this possibility at this time, we believe it is unlikely since there are no other classical or cryptic translational start sites between codon 74 and the M90 position.

Mechanism of expression of Us1.5.

Initiation of translation at the M90 position of the ICP22 ORF also raises another question, “from what transcript is U_S1.5 translated?” The U_S1.5 transcript originally mapped by Carter and Roizman initiates at codon 146 of the ICP22 ORF and thus is too far downstream to allow translation to initiate at M90 (6). Since the U_S1.5 protein is detected in multiple strains of the virus and is not a cleavage product of the full-length ICP22 protein, it clearly must initiate from some transcript. We have not been able to reliably detect a faster-migrating transcript in infected cells, nor could we detect the U_S1.5 protein in the absence of the ICP22 promoter (Fig. 8). Analysis of the U_S1 gene sequence for splice donor and acceptor sites identifies only the previously identified splice site. This combined information suggests two possible mechanisms of expression. First, several larger ICP22-specific transcripts are detected in KOS-infected cells (2). It is possible that one or more of these transcripts may serve as a template for U_S1.5; however, data presented here would argue against this possibility. Specifically, cells transfected with pCDNA3:3xStop expressed proteins that migrated at the correct size for U_S1.5. Since this vector contains the ICP22 ORF under the control of the cytomegalovirus promoter and is lacking all ICP22 promoter and 5′ untranslated region sequences, expression of U_S1.5 from pCDNA3:3xStop must originate from the primary transcript.

Alternatively, U_S1.5 translation may initiate from the primary ICP22 transcript by one of several mechanisms, including ribosomal scanning, ribosomal shunting, or internal ribosomal entry. Although little experimental evidence is available to determine which mechanism is preferred, sequence analysis suggests that ribosomal scanning may be important. First, M90 is the first available methionine following the ICP22 translational start site. In addition, the Kozak consensus sequence for the ICP22 translational start site is only adequate since the +4 position is strong while the −3 position is weak (12). Given that the ICP22 Kozak sequence is not strong, we would predict that the ribosome skips the ICP22 translational start site at a low frequency and scans until it reaches M90. The Kozak sequence for M90 is also adequate, containing one strong site and one weak site, and should allow initiation of translation. Experiments aimed to test these possibilities are in progress.

Study of U_S1.5 protein function.

If translation indeed initiates at the M90 position of the ICP22 ORF, then research analyzing U_S1.5 function needs to be reevaluated. Published studies of U_S1.5 expression in transient assays have utilized either aa 147 to 420 or 171 to 420 as their defined boundaries of U_S1.5 (7, 9, 10). These experiments should be repeated to ensure that the same phenotypes are observed with the full-length protein. Although it seems likely that the phenotypes will be the same, since the 90-420 region contains all of the sequences defined by the other two boundaries, it is possible that the additional N-terminal sequences alter function. For example, one of the two mapped nuclear localization signal sequences of ICP22 lies between aa 90 and 147, suggesting that there may be a difference in U_S1.5 localization between the two proteins. In addition, Tyr116 is also now within the defined region of U_S1.5, suggesting that U_S1.5 may play a role in viral pathogenesis.

Role of U_S1.5 during infection.

While we have provided evidence for the location of the true translational start site of U_S1.5, it remains unclear why HSV encodes both ICP22 and U_S1.5. Having identified two mutations that separate expression of ICP22 and U_S1.5 (M90A and 3xStop), we are now in a position to begin to address this question. Construction of HSV-1 mutant viruses containing M90A, 3xStop, or both M90A and 3xStop is currently under way, and they will be tested for known properties of ICP22 during infection. We hope to use these reagents to begin to tease apart the functions of ICP22 and U_S1.5 during infection.