INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Herpes simplex virus type 1 (HSV-1) is a neurotropic virus that causes a variety of clinically significant diseases ranging from superficial cutaneous lesions to life-threatening encephalitis (49). The designation of HSV-1 as a neurotropic virus is based on its ability to produce both lytic and latent infections of neurons of the peripheral and central nervous systems. The extent to which a given strain of HSV-1 can gain access to and travel within neurons (neuroinvasiveness) and the relative efficiency with which a strain produces neuronal damage and death (neurovirulence) are major determinants of the disease-producing capabilities of that strain. Identification of the viral genes that control neuroinvasiveness and neurovirulence and elucidation of the functions of the protein products of these genes are necessary to understand the mechanism by which HSV-1 causes disease.

By definition, any viral gene which, when mutated, results in the failure of the virus to enter or spread in neurons or to replicate in epithelial or neuronal cells can be said to affect neuroinvasiveness or neurovirulence, respectively. Some viral genes, however, have unique properties which appear to function exclusively or predominantly in neurons. Notably, a number of viral genes which affect neurovirulence and which, collectively, determine whether infection will be productive or latent are located within the abb'a'c'ca repeat regions flanking the unique long and short regions of the HSV-1 genome (Fig. 1) (reviewed in reference 50).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Physical map of the internal repeat region of HSV-1 DNA. (A) Diagram of the HSV-1 genome. TRL, UL, and IRL, terminal repeat, unique, and internal repeat sequences, respectively, of the long segment; IRS, US, and TRS, internal repeat, unique, and terminal repeat sequences, respectively, of the short segment; b', internal inverted repeat sequence bracketing UL; c', internal inverted repeat sequence bracketing US; a', inverted sequence located between the b' and c' sequences and, in reverse orientation, at the genomic termini. (B) Expanded map of the internal repeat sequence b'-a'-c' lying between kbp 117 and 134 on the HSV-1 genome. Beneath the scale of kilobase pairs are shown the locations of the b'-a'-c' sequence and oriS. (C) Beneath the b-a-c repeats are shown the locations of genes contained within kbp 117 to 134. Specifically, the map shows the locations of sequences specifying the transcripts encoding ICP0, ICP34.5, ICP4, and ICP22 and sequences specifying the LATs. ORFs, including ORF-O and ORF-P, are shown as hatched boxes beneath ORF34.5. The locations of sequences specifying the L/STs are shown beneath ORF34.5 and ORF-P. Beneath the L/STs are shown sequences specifying TR/UL RNA, the X and X transcripts, and oriS RNAs 1 and 2. Arrows indicate the direction of transcription, and the polyadenylation status is designated A+, A or ? (D) DNA sequences specifying the riboprobes (arrows) used in this study. The arrows represent the orientation of these sequences in pGEM vectors as driven by the SP6 promoter. (E) Sequences specifying the DNA probes used for Southern blot analysis. (F to L) Physical map locations of the HSV-1 DNA sequences located in plasmids pBamK (F); pNco, pL/ST-n38, and pL/ST-4BSm (G); pL/ST-Nxi and pL/ST-NxiLM3 (H); pWR-CAT+, pL/ST-U-CAT, and pL/ST-4BSm-CAT (heterologous CAT sequences are shown as boxes) (I); p0V, p0V-n38, p0V-4BS, and p0V-4/38 (J); pG-L/ST and pG-L/ST-C (heterologous GST sequences are shown as boxes) (K); and p38Nco and p4BSNco (L). (M) Locations of the L/ST mutations constructed in this study. The position of the 4BS mutation is represented by the triangle. The position of the n38 mutation is represented by the rectangle. Below the physical map is the wild-type (KOS) sequence, beneath which are the mutated (L/ST-4BS and L/ST-n38) sequences. Each point mutation is designated by a vertical line. The Eco47III restriction site generated by the mutation in L/ST-4BS and the PflMI restriction site destroyed by the L/ST-n38 mutation are underlined.

The genes specifying the L/S junction-spanning transcripts (L/STs) and the ₁34.5 gene, are located within these repeat regions and constitute the subject of this paper. The L/STs are a series of 5'-coterminal polyadenylated RNAs transcribed from the same DNA strand as the latency-associated transcripts (LATs) (Fig. 1) (50). The L/STs range in size from 2.3 to >9.0 kb and are expressed at detectable levels only by mutant viruses that express no functional ICP4 or that lack an intact ICP4 binding site at the transcription initiation site of the L/ST gene (50). These properties suggest that ICP4 bound to its cognate binding site in the L/ST promoter blocks the expression of the L/STs. The L/STs overlap the 3' terminal 2.3 kb of the 8-kb minor LAT on the same strand, and, critical to the subject of this paper, the L/STs overlap and are antisense to the ₁34.5 gene (Fig. 1).

The first and largest open reading frame (ORF) in the L/STs is ORF-P, which specifies the ORF-P protein (OPP) (Fig. 1C and Fig. 10). ORF-P is conserved in all HSV-1 strains whose genomes have been sequenced in this region. OPP is expressed at detectable levels only under conditions which allow expression of the L/STs (26, 50). Overlapping the 5' half of ORF-P, and encoded by the same DNA strand but in another reading frame, is ORF-O, which encodes the ORF-O protein (OOP) (Fig. 1C). In strain 17syn+, ORF-O is approximately 300 codons in length. In strains KOS, F, MGH10, and CVG2, a single-base-pair substitution terminates ORF-O after approximately 160 codons. Randall et al. have recently provided evidence suggesting that, like ORF-P, ORF-O is downregulated by ICP4 (34). The translational start codon of ORF-O lies immediately downstream of the TATA box in the L/ST promoter; thus, ORF-O is probably encoded by the LATs rather than the L/STs. Recently, however, Randall et al. have suggested that in strain F, ORF-O and ORF-P initiate from the same methionine codon and that the initiation of ORF-O at this codon results from frameshift or editing of its RNA (34).

The L/STs overlap and are antisense to the ₁34.5 gene. The ₁34.5 gene is located between the genes specifying two major immediate-early (IE) transcriptional regulatory proteins, ICP0 and ICP4 (Fig. 1). The ~8-kb LAT, specified by the opposite strand, overlaps and is antisense to the entire ICP0 and ₁34.5 genes. The ₁34.5 gene encodes ICP34.5, a protein with an apparent molecular mass of 44 kDa (2, 11, 15). The ₁34.5 ORF (ORF-34.5) is conserved in all HSV-1 strains that have been sequenced to date (14, 18). Viruses with deletions in ORF-34.5 are replication competent in vitro; however, they are highly attenuated for neurovirulence in vivo following intracranial inoculation of mice (6, 14, 35, 46-48). Other properties of ₁34.5 mutant viruses include (i) more efficient replication at some primary sites of infection (e.g., the footpad and vagina) than at others (e.g., the eye) (49), (ii) defective replication in the sensory ganglia of mice, guinea pigs, and rabbits and in the central nervous system of mice (6, 14, 30-32, 35, 45, 48); (iii) wild-type growth in most cell types but restricted growth in SK-N-SH (human neuroblastoma) (13), CV-1 (African green monkey kidney) (32), murine 10T1/2 (stationary-phase primary mouse embryo), and 3T6 cells (6, 7); (iv) premature shutoff of all protein synthesis (cellular and viral) following infection of SK-N-SH cells (12, 13); and (v) reduced establishment of latency compared to wild-type virus (31, 45).

During the course of the studies of ICP34.5 just described, a number of mutant viruses were constructed and characterized. By virtue of the complete overlap of the ₁34.5 gene with the L/ST and the 8-kb LATs, mutations introduced into the ₁34.5 gene were also introduced into sequences specifying the L/STs and the LATs. Consequently, "₁34.5 mutant viruses" are either ORF-34.5 ORF-P double mutants or ORF-34.5 ORF-P ORF-O triple mutants (Fig. 1).

Because ORF-O is probably contained in the LATs rather than the L/STs, as noted above, we focused the present investigation on the contribution of ORF-P to the phenotype of ORF-34.5 ORF-P double-mutant viruses. We approached this problem by using two genetic strategies. In the first, we mutated ORF-P without altering ORF-34.5, and in the second, we altered the transcriptional regulation of the L/STs by greatly increasing the levels of L/ST and ORF-P expression. We report here the construction and characterization of two mutant viruses: one, L/ST-n38, containing a nonsense mutation that terminates ORF-P translation at amino acid 38 but does not alter the sequence of ICP34.5 encoded by the complementary DNA strand (Fig. 1G; also see Fig. 10), and another, L/ST-4BS, containing a mutation in the L/ST promoter that derepresses L/ST expression by abrogating ICP4 binding to its cognate binding site near the transcription initiation site of the L/STs (Fig. 1G; also see Fig. 10). The properties of L/ST-n38 demonstrate that expression of a truncated OPP produces no observable phenotype relative to wild-type virus in vitro or in vivo. By contrast, L/ST-4BS, which is mutated in the ICP4 binding site in the L/ST promoter and overexpresses the L/STs and OPP, is highly attenuated following intracranial inoculation of juvenile mice, yet replicates efficiently in the mouse eye and trigeminal ganglia (TGG) without eliciting an adverse host response. Of special interest was the observation that L/ST-4BS induced cell-type-specific expression of both ₁34.5 transcripts and LATs since both types of transcript were abundant in Vero cells but barely detectable in mouse neuroblastoma cells (NB41A3). The significance of these findings to the assignment of specific functions to OPP and ICP34.5 can be summarized as follows: (i) ICP34.5 and not ORF-P is the neurovirulence factor responsible for attenuation of ORF-P ORF-34.5 double mutants and (ii) a factor or element affected by the 4BS mutation mediates cell-type-specific expression of the ₁34.5 gene and the LATs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells and viruses. African green monkey kidney cells (Vero; ATCC CCL 81) and human embryonic lung cells (HEL cells) were grown and maintained in Dulbecco's modified Eagle's medium as previously described (38). Mouse neuroblastoma cells (NB41A3; ATCC CCL147) were propagated in Ham's F10 medium supplemented with 2.5% fetal calf serum and 15% horse serum. Human neuroblastoma cells (SK-N-SH) were kindly provided by Patrick Vaughn (University of California, Irvine, Calif.) and were propagated as described above for Vero and HEL cells. Rat pheochromocytoma cells (PC12) were the gift of John Wagner (Cornell University Medical College, New York, N.Y.) and were propagated and maintained as described previously (21).

Plasmids. For subcloning purposes, pBamK (Fig. 1F), containing the KOS BamHI K fragment from pSG28 (40), and pNco (Fig. 1G), containing the 1.75-kbp NcoI subclone of pBamK, were constructed as described previously (50). Plasmid pL/ST-n38 (Fig. 1G) is identical to pNco, except that it contains a single-base-pair transition in ORF-P. Mutagenesis was accomplished by in vitro phagemid mutagenesis with pNco as the parental plasmid and the mutant oligonucleotide 5'-ACCAGGTGACGCACCCGG-3'. The transition mutation from G to A (underlined) eliminates the cleavage site for the restriction enzyme PflMI. Plasmid pL/ST-4BS^m (Fig. 1G) is identical to pNco except for four single-base-pair changes in the ICP4 binding site near the transcriptional start site of the L/STs. The mutations were introduced by in vitro phagemid mutagenesis with pNco as the parental plasmid and the mutant oligonucleotide 5'-GCGGCTCCTGCCAGCGCTCCTCCGGAGAGC-3'. The mutation of ATCGTCTCT to AGCGCTCCT generates a new restriction site for the endonuclease Eco47III.

Plasmid mutagenesis. Competent Escherichia coli CJ236 (dut ung) was transformed with plasmid pNco, and single-stranded phagemid DNA was prepared as instructed by the manufacturer (Promega). Oligonucleotide-directed mutagenesis was then performed by the Kunkel method (23). The mutations were confirmed by the absence (pL/ST-n38) or appearance (pL/ST-4BS^m) of specific restriction enzyme sites and by DNA sequencing.

In vitro transcription and translation. Linearized template DNA (2 µg) was transcribed in vitro with SP6 polymerase to generate capped RNA as specified by the manufacturer (Promega). Half of the RNA was used for in vitro translation; the other half was used for Northern blot analysis. In vitro translation was performed, as specified by the manufacturer (Promega), with rabbit reticulocyte lysate and either [³⁵S]cysteine or [³⁵S]methionine (New England Nuclear, Boston, Mass.). Aliquots (2.5 µl) of in vitro-translated proteins were analyzed by denaturing polyacrylamide gel electrophoresis.

Gel mobility shift assays. To generate probes for use in-gel mobility shift assays and competition experiments, single-stranded 30-mer oligonucleotides corresponding to the plus and minus strands of the wild-type and mutated ICP4 binding sites were synthesized by the Molecular Biology Core Facility at the Dana-Farber Cancer Institute. The nucleotide sequence of the wild-type binding site is 5'-GCGGCTCCTGCCATCGTCTCTCCGGAGAGC-3', and that of the mutated binding site is 5'-GCGGCTCCTGCCAGCGCTCCTCCGGAGAGC-3'. (The core ICP4 binding site is underlined.) Complementary single-stranded probes were annealed by heating to 98°C for 5 s followed by slow cooling to 68°C, incubation at 68°C for 1 h, and slow cooling to room temperature. Double-stranded probes were separated from unannealed single-stranded oligonucleotides by polyacrylamide gel electrophoresis (PAGE), cut out of the gel, and electroeluted overnight at 35 V. Gel-purified double-stranded probes were radiolabeled with T4 polynucleotide kinase (New England Biolabs, Beverley, Mass.) and [-³²P]ATP (New England Nuclear).

Transient-expression assays. Vero cells (10⁶ cells) were plated in 60-mm-diameter petri dishes 24 h prior to transfection. The medium was replaced 3 h prior to transfection. Reporter and effector plasmids were transfected by the calcium phosphate-bis-ethane sulfonic acid method described by Ausubel et al. (4). Cell lysates for determination of CAT activity were prepared approximately 40 h posttransfection and assayed by the phase extraction method (42) as modified by Frazier et al. (20). For transfection/superinfection assays, monolayers were infected with virus at an MOI of 10 PFU/cell 24 h after transfection and lysates were prepared at 3, 6, 12, and 24 h p.i.

Isolation of mutant and rescuant viruses. To isolate mutant viruses, L/ST-n38 and L/ST-4BS, Vero cells (5 × 10⁵ cells) were seeded in 60-mm-diameter dishes and cotransfected 24 h later with 1 µg of infectious 7134 DNA and 10 µg of a DNA fragment containing the mutation (i.e., the 7,002-bp PstI-EcoRV fragment from p0V-n38 or the 7,129-bp PstI-EcoRI fragment from p0V-4BS [Fig. 1J]). Six days later, when cytopathic effects were generalized, the cultures were harvested, frozen and thawed, sonicated for 1 min, clarified by low-speed centrifugation, plated on Vero cell monolayers, and overlaid with methylcellulose. After incubation at 37°C for 4 days, 2 ml of 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal) (300 µg/ml) in neutral red was added to each 35-mm petri dish. Parental 7134 virus would yield blue plaques because the ICP0 sequence has been replaced by lacZ. Recombinant plaques would be white if the wild-type ICP0 sequences had replaced lacZ sequences. In this way, the n38 and 4BS mutations were "piggybacked" simultaneously into the genome in a subset of recombinants. White plaques were picked and assayed by Southern blot analysis for the presence of the desired mutation in both copies of the b repeat sequences. Mutants were subjected to two additional rounds of plaque purification. One viral isolate obtained from cotransfection of infectious 7134 and linearized p0V-n38 DNAs was designated L/ST-n38, whereas an isolate from the cotransfection of 7134 and p0V-4BS DNAs was designated L/ST-4BS.

Western blot analysis. Cells were infected at a multiplicity of 10 PFU/cell and harvested at the indicated times p.i. Lysates were prepared as follows. Infected monolayers were washed twice with cold phosphate-buffered saline (PBS) (39), and scraped into PBS in the presence of protease inhibitors (10 µg of TPCK [N-tosyl-L-phenylalanine chloromethyl ketone] per ml and 2 mM phenylmethylsulfonyl fluoride [PMSF]). The cells were pelleted and resuspended in a small volume of PBS-TPCK-PMSF, aliquoted, and quick-frozen in liquid nitrogen. Aliquots were thawed only once for use. The protein concentration was determined with the Bio-Rad protein assay reagent, as specified by the manufacturer. Cell lysis was completed by the addition of SDS-PAGE loading buffer (50 mM Tris, 2% SDS, 0.7 M -mercaptoethanol, 2.75% sucrose). Proteins were separated by denaturing polyacrylamide gel electrophoresis (10) and transferred to Immobilon-P membranes (Millipore Corp., Bedford, Mass.).

G-L/ST-C fusion protein purification. To prepare the G-L/ST-C fusion protein, an overnight culture of bacteria transformed with pG-L/ST-C was diluted 1:50 and allowed to grow until its optical density at 595 nm reached 1.0. Fusion protein expression was induced by the addition of isopropyl--D-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. At 2 h after induction, bacteria were pelleted and resuspended in Tris-buffered saline (TBS) (39) containing 10 mM EDTA, 75 µg of TPCK per ml, 2 mM PMSF, and 2 mM leupeptin. The bacteria were lysed by sonication, incubation with Triton X-100 (final concentration, 1%), and rocking for 30 min at 4°C. The lysate was cleared by centrifugation for 15 min at 26,890 × g in an SS34 rotor. Glycerol was added to the supernatant fluid to a final concentration of 10%. The G-L/ST-C fusion protein was purified on glutathione-Sepharose 4B beads (Pharmacia) as specified by the manufacturer instructions and eluted from beads with 20 mM glutathione in 100 mM Tris [pH 8]-120 mM NaCl.

Preparation of antibodies to OPP. Rabbit polyclonal antibodies were raised against the purified G-L/ST-C fusion protein by HRP, Inc (Denver, Pa.). Two New Zealand White rabbits were inoculated intradermally with 250 µg of G-L/ST-C each and given booster doses three times at 2-week intervals by subcutaneous dorsal injections of 125 µg of G-L/ST-C prior to obtaining antisera. Antisera from both rabbits reacted with G-L/ST-C in Western blot analysis. Anti-G-L/ST-C antibodies were purified from one of the antisera, DF111, by the blot purification method of Harlow and Lane (22).

Analysis of infected-cell polypeptides. Vero cells (5 × 10⁵ cells) or SK-N-SH cells (1 × 10⁷ cells) were seeded in 35-mm-diameter petri dishes 24 h prior to infection. The cells were infected with the indicated viruses at a multiplicity of 10 PFU/cell. At the designated times p.i., the culture medium was replaced with 1 ml of methionine-minus medium (ICN Biomedicals, Inc., Costa Mesa, Calif.) supplemented with 50 µCi of [³⁵S]methionine (New England Nuclear). Monolayers were labeled for 30 min at 37°C and then processed as described above to obtain protein lysates. Individual proteins in lysates (5 µg of Vero and 10 µg of SK-N-SH cell lysate) were separated by SDS-PAGE and analyzed by autoradiography as described previously (24).

PAGE. SDS-PAGE of infected cell lysates was performed by the method of Laemmli (24). Protein samples were boiled for 5 min in sample buffer (50 mM Tris, 2% SDS, 0.7 M -mercaptoethanol, 2.75% sucrose, Coomassie blue). In vitro-translated proteins were also analyzed on Tricine gels. The running gel contained 20% acrylamide (acrylamide/bisacrylamide ratio, 30:1), 1 M Tris (pH 8.45), 0.1% SDS, and 16.7% glycerol. The stacking gel contained 4% acrylamide, 0.75 M Tris (pH 8.45), and 0.075% SDS. Samples were boiled in Tricine gel sample buffer (0.125 M Tris [pH 6.8], 4% SDS, 0.7 M -mercaptoethanol, 10% glycerol, Coomassie blue). The anode buffer contained 200 mM Tris (pH 8.9). The cathode buffer contained 100 mM Tris, 100 mM Tricine, and 0.1% SDS. The pH of the cathode buffer was approximately 8.25 unadjusted. All the polyacrylamide gels were cross-linked with ammonium persulfate and N,N,N',N'-tetramethylethylenediamine (TEMED).

Northern blot analysis. Total-cell RNA was isolated and Northern blot analysis was performed as previously described (50) with two exceptions. First, for Northern blot analysis, 7.5 µg of RNA was used per lane, and second, after overnight hybridization of the ³²P-labeled probe, the final wash in 0.1× SSC-0.1% SDS was performed twice for approximately 50 min per wash. (0.1× SSC is 15 mM NaCl plus 1.5 mM sodium citrate [pH 7].) Riboprobes were transcribed from the SP6 promoter in the vector as instructed by the manufacturer (Promega).

Animal procedures. For neurovirulence assays, 3-week-old female BALB/C mice (Charles River Breeding Laboratories, Inc., Kingston, N.Y.) were anesthetized with sodium pentobarbital and inoculated intracranially with 10-fold dilutions of virus stock (n = 6 to 10 per dilution) in a volume of 30 µl. The mice were monitored for 21 days, and the 50% lethal dose was calculated by the method of Reed and Muench (36).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Sequence analysis of the 2.3-kb L/ST region. The DNA sequences used throughout these studies were derived from HSV-1 strain KOS. Since only the sequence of the L/ST promoter region of strain KOS had been published prior to these studies (5), we subcloned and sequenced the region corresponding to the 2.3-kb L/ST. Comparison of the KOS DNA sequence with that of HSV-1 strains 17syn+, F, MGH-10, and CVG-2 revealed a nearly identical sequence, confirming that this region of the genome is highly conserved among various strains of HSV-1 (data not shown). The L/ST promoter region, as well as sequences containing the 5' end of the L/STs, are also highly homologous to the corresponding region in HSV-2 strain HG52 (26, 29). As shown by the alignment of amino acid sequences, OPP of strain KOS is nearly identical to OPP of other HSV-1 strains (Fig. 2). The only major strain-specific difference lies in the variable number of AGV repeats near the amino terminus of the protein. These repeats correspond to the previously reported ATP repeats in the carboxy terminus of ICP 34.5 (14, 18, 26, 29).

View larger version (80K): [in this window] [in a new window]   FIG. 2.   Alignment of the ORF-P amino acid sequence in five strains of HSV-1. The amino acid sequences of HSV-1 strain KOS ORF-P are aligned with ORF-P amino acid sequences from strains 17syn+ (18), F, MGH10, and CVG2 (26). Consensus amino acids are denoted by asterisks beneath the sequences. Periods represent gaps. The codon number of the last amino acid on each line is shown on the right. The numbers at the lower right represent the length of ORF-P in each viral strain.

The nonsense mutation (n38) terminates the translation of ORF-P at amino acid 38. To elucidate the functions of the L/STs and OPP, we introduced (i) nonsense mutations into both copies of ORF-P to yield a truncated protein consisting of only the amino terminus of OPP (the n38 mutation does not affect the amino acid sequence of ICP34.5 or the regulation of L/ST expression), and (ii) mutations into both copies of the ICP4 binding sites near the transcriptional start site of the L/STs that significantly enhance L/ST expression (and hence OPP expression) by eliminating ICP4-mediated repression.

View larger version (79K): [in this window] [in a new window]   FIG. 3.   In vitro transcription/translation of wild-type (WT) and mutant ORF-P. (A) Restriction map of the region of the 2.3-kb L/ST. Beneath the restriction map is shown the location of ORF-P. The plasmid template was linearized at the restriction sites indicated before in vitro transcription (B) and translation (C). The positions of the methionine (Met) and cysteine (Cys) residues used to radiolabel in vitro-translated proteins are shown. (B) Capped RNA from wild-type and n38 mutant plasmids was transcribed in vitro, separated electrophoretically, and transferred to Magnagraph paper. The resulting transcripts were detected by Northern blot analysis with riboprobe EBN9-LAT (Fig. 1D). The positions of RNA size markers (in kilobases) are shown on the left; the expected nucleotide (nt) sizes of the in vitro-transcribed RNA are indicated beneath each lane. (C) In vitro-transcribed wild-type and n38 mutant RNA from panel B were translated in the presence of [35S]Cys, separated electrophoretically, and transferred to a polyvinylidene difluoride membrane. Proteins were visualized by autoradiography. The positions of molecular mass markers (in kilodaltons) are shown on the left. A.A., amino acids.

The ICP4 binding-site mutation (4BS) abrogates ICP4 binding at the L/ST transcriptional start site and significantly enhances expression from the L/ST promoter. To disrupt ICP4-mediated repression of L/ST expression, we introduced four single-base-pair substitution mutations into the ICP4 binding site in the L/ST promoter (Fig. 1M). These mutations created a new Eco47III restriction enzyme site, facilitating the identification of this mutation in plasmid and recombinant viral constructs. The effects of these mutations were first evaluated in vitro by gel mobility shift and transient-expression assays before they were introduced into the viral genome.

View larger version (40K): [in this window] [in a new window]   FIG. 4.   Effects of the mutant ICP4 binding site on expression from the L/ST promoter. (A) Gel shift analysis. Radiolabeled 30-bp probe DNAs containing either the wild-type (WT) or mutated (4BS Mutant) ICP4 binding site were incubated with mock-, n12-, or KOS-infected Vero cell lysates prepared at 8 h p.i. and enriched for nuclear proteins. Anti-ICP4 monoclonal antibody (58S MAb) was added to the incubation mixture where indicated (lanes 2, 4, 6, 9, 11, and 13). The four major complexes detected are labeled a through d. (B to D) Transient-expression assays. Vero cells were transfected with pL/ST-U-CAT (WT), containing the wild-type ICP4 binding site, or pL/ST-4BSm-CAT (4BS MUT), containing the mutated ICP4 binding site. Reporter plasmids were cotransfected with increasing amounts of the ICP4-expressing plasmid, pn11 (16) (B) or the ICP0-expressing plasmid, pSH (9) (C). Cells were harvested at 40 h posttransfection or at the times indicated, and CAT enzyme activity was assayed. Data points represent the average activities in two separately transfected plates. Transfection of pWR-CAT+, the promoter-less reporter plasmid, in the presence or absence of the effector plasmids, yielded background levels of CAT activity (data not shown). At 24 h after transfection of reporter plasmids alone, Vero cells were mock superinfected or superinfected with n12 or KOS at an MOI of 10 PFU/cell (D). The cells were harvested at 3, 6, 12, and 24 h p.i. A 9 h p.i., an additional data point was determined for 4BS reporter-transfected and KOS-superinfected Vero cells.

L/ST mutant viruses replicate efficiently in cells of neural and nonneural origin. ORF-34.5/ORF-P double mutants replicate to wild-type levels in permissive cells (e.g., Vero cells) but are growth restricted in certain neural cell lines (e.g., SK-N-SH cells) (12, 13). To assess the ability of mutant viruses L/ST-n38 and L/ST-4BS to replicate in Vero and SK-N-SH cells, single-cycle growth experiments were performed with the wild-type virus, the two mutants, and the rescuant viruses. In Vero cells, L/ST-n38, L/ST-4BS, and the rescuants replicated as efficiently as wild-type virus, indicating that no spurious mutations in genes that affect replication efficiency were present in any of the viruses (Fig. 5A). In SK-N-SH cells, however, L/ST-n38 replicated to wild-type levels whereas L/ST-4BS replicated to somewhat lower levels (Fig. 5B). This experiment was repeated twice with identical results. The modest reduction in replication efficiency of L/ST-4BS in SK-N-SH cells was not observed with another ICP4 binding-site mutant virus, R7530 (25).

View larger version (22K): [in this window] [in a new window]   FIG. 5.   Growth curves of L/ST mutant viruses in cell culture. Monolayers of Vero cells (A) or SK-N-SH cells (B) were infected at an MOI of 2.5 PFU/cell with KOS, L/ST-n38, L/ST-n38R, L/ST-4BS, or L/ST-4BSR as described in Materials and Methods. Total virus was harvested at the indicated times p.i. and quantified by a standard plaque assay on Vero cell monolayers. Titers are the average obtained in two tests.

Transcription of genes in the abb'a'c'ca repeats in cells of nonneural and neural lineage by L/ST mutant viruses. L/ST expression by the mutant viruses, KOS, and the rescuant viruses was examined by Northern blot hybridization of total RNA harvested from infected Vero and NB41A3 cells (Fig. 6A). The ICP4 nonsense mutant n12 was included as an L/ST-expressing control (50).

View larger version (54K): [in this window] [in a new window]   FIG. 6.   Transcription of viral genes in the b'-a'-c' sequence in L/ST mutant virus-infected Vero and NB41A3 cells. Vero and NB41A3 cells were mock infected or infected with 10 PFU of KOS, n12 (an ICP4 nonsense mutant virus), L/ST mutant viruses L/ST-n38 or L/ST-4BS, or L/ST rescued viruses L/ST-n38R or L/ST-4BSR per cell. Total-cell RNA (Vero [lanes 1 to 7] or NB41A3 [lanes 8 to 14]) was isolated at 18 h p.i. (A, B, and D) or 6 h p.i. (C), separated electrophoretically, and transferred to Magnagraph paper. The positions of RNA size markers (in kilobases) are shown on the left. (A) L/STs were detected with riboprobe EBN9-LAT (Fig. 1D). (B) 134.5 RNAs were detected with riboprobe EBSN34.5 (Fig. 1D). (C) ICP0 and ICP4 RNAs were detected concurrently with riboprobes 0Hc-Xh and 4Sma (Fig. 1D), respectively. The specificity of each riboprobe was determined independently (data not shown). ICP0 and ICP4 RNAs from NB41A3 cells (lanes 8 to 14) were detected after a twofold-longer autoradiographic exposure than RNAs from Vero cells (lanes 1 to 7). The positions of ICP4 RNA and ICP0 are marked by black and white arrowheads, respectively. (D) LATs were detected with riboprobe BbSLAT (Fig. 1D).

(i) L/STs. The L/STs were not detected in RNA from L/ST-n38-infected Vero or NB41A3 cells (Fig. 6A, lanes 4 and 11); this was expected since the L/ST promoter in L/ST-n38 contains a wild-type ICP4 binding site and is therefore subject to repression by ICP4. Similarly, the L/STs were not detected in KOS-infected (lanes 2 and 9) or L/ST-n38R-infected (lanes 5 and 12) cells for the same reason. As expected from the results of transient-expression assays, the L/STs were expressed in L/ST-4BS-infected cells (lanes 6 and 13). In Vero cells, L/ST-4BS expressed lower levels of the 2.3-kb L/ST than did n12, whereas the opposite was true in NB41A3 cells. L/ST-4BS also expressed higher levels of L/STs than did n12 in differentiated and undifferentiated PC12 cells and in SK-N-SH cells (data not shown). In addition to the abundant 2.3-kb transcript, two smaller transcripts (<1.35 kb) were detected in L/ST-4BS-infected Vero and NB41A3 cells (lanes 6 and 13, respectively). The presence of these smaller transcripts in both infected Vero and NB41A3 cells argues that they represent independent transcripts and not degradation products of the larger L/STs. Finally, L/ST expression was not detected in L/ST-4BSR infected cells, confirming that the expression of the L/STs in L/ST-4BS was due to the 4-bp substitution in the ICP4 binding site (lanes 7 and 14).

(ii) ₁34.5 transcripts. Using a DNA probe specific for ₁34.5 transcripts, Chou and Roizman (15) detected two infected-cell-specific RNAs of 1.3 and 5.2 kb and one nonspecific 3.5-kb RNA in wild-type virus (strain F)-infected Vero cells. As shown in Fig. 6B, the abundant 1.3-kb ₁34.5 RNA and two less prominent 2.5- and 4-kb transcripts were detected in infected Vero cells with the EBSN34.5 riboprobe (lanes 2 and 4 through 7). Further mapping experiments are needed to determine whether the 2.5- and 4-kb transcripts are ₁34.5 specific or nonspecific. Levels of the 1.3-kb ₁34.5 RNA detected in cells infected with wild-type virus and L/ST-n38 were similar (lanes 2 and 4) but somewhat higher in L/ST-4BS-infected cells, indicating that the n38 mutation had no major effect and that the 4BS mutation had, if anything, an enhancing effect on the levels of ₁34.5 RNA in Vero cells. As anticipated, transcripts were not detected in n12-infected cells with the ₁34.5 probe (lane 3), since ICP4-null viruses do not express late genes.

(iii) ICP0 and ICP4 transcripts. Northern blot analysis of total RNA harvested from infected Vero cells at 6 h p.i. demonstrated that ICP0 and ICP4 RNAs were expressed by both L/ST-n38 and L/ST-4BS at levels comparable to those of wild-type virus (Fig. 6C, lanes 2, 4, and 6). Because these transcripts were present at very low levels in NB41A3 cells at 6 h p.i., a considerably longer autoradiographic exposure was necessary to visualize the bands in lanes 8 to 14. As in Vero cells, levels of ICP0 and ICP4 transcripts were similar to those seen in KOS-infected cells. Although abundant ICP4 transcripts were present in n12-infected Vero cells, no functional ICP4 protein was made because of the nonsense mutation in the ICP4 gene; hence, ICP0 and ICP4 transcription is not repressed and these transcripts are made at higher-than-wild-type levels. Notably, overexpression of ICP0 and ICP4 transcripts by n12 relative to KOS was evident in Vero cells but not in NB41A3 cells (or PC12 cells [data not shown]). Thus, overexpression of these two transcripts is cell type specific.

(iv) LATs. Because the L/STs may function during HSV latency (50), it was of interest to determine if the n38 or 4BS mutations have an effect on LAT expression in vitro. We therefore performed Northern blot analysis with a LAT-specific riboprobe, BbSLAT, with total RNA from Vero and NB41A3 cells harvested at 18 h p.i. with the L/ST mutant and rescuant viruses (Fig. 6D). RNAs from KOS-infected cells (lanes 2 and 9) and n12-infected cells (lanes 3 and 10) were included as LAT-positive and LAT-negative controls, respectively.

OPP expression by L/ST mutant viruses. Having demonstrated high-level expression of the L/STs by L/ST-4BS but not by L/ST-n38 in both Vero and NB41A3 cells, we attempted to assess the kinetics and levels of expression of OPP in Vero cells. To facilitate these studies, rabbit antibody (DF111) was raised against the C-terminal 52 amino acids of ORF-P fused to GST. Using this antibody, we first examined the ability of the L/ST mutant viruses and their respective rescuants to express OPP. Extracts of KOS- and n12-infected Vero cells were included as OPP-negative and -positive controls, respectively. As shown in Fig. 7, OPP was not detected in Vero cells infected with KOS, L/ST-n38, and L/ST-4BSR, viruses that do not express the L/STs as demonstrated by Northern blot analysis (Fig. 6A). In cells infected with L/ST-4BS, three proteins of approximately 24, 29, and 33 kDa were detected (Fig. 7, lanes 10 to 12). The size of the 24-kDa band is consistent with the molecular mass of OPP (23 kDa) and with the size of the OPP band observed by in vitro transcription-translation (26 kDa) (Fig. 3). The size and relative abundance of each protein species at 6, 12, and 18 h p.i. suggest that the 24-kDa protein is modified posttranslationally during the course of lytic infection. As anticipated, the 24-kDa protein was also detected in n12-infected Vero cells, although the appearance of the larger (29-kDa) species was delayed and these proteins were present in lower abundance relative to that in L/ST-4BS-infected cells. OPP expression by n12 and L/ST-4BS was also detected in NB41A3 cells, nerve growth factor-differentiated and undifferentiated PC12 cells, and SK-N-SH cells (data not shown), demonstrating that like expression of the L/STs, OPP expression is not cell type specific.

View larger version (105K): [in this window] [in a new window]   FIG. 7.   Expression of ORF-P protein by the L/ST mutant viruses. Vero cells were mock infected or infected with KOS, n12, L/ST-n38, L/ST-4BS, L/ST-4BSR, or L/ST-4/38 at a multiplicity of 10 PFU/cell. Infected-cell proteins were prepared at the indicated times p.i., separated by SDS-PAGE, and transferred to a polyvinylidene difluoride membrane. ORF-P protein was detected by Western blot analysis with rabbit polyclonal antibody, DF111, diluted 1:1,000.

L/ST-4BS but not L/ST-n38 induces premature shutoff of protein synthesis in SK-N-SH cells. Gradual shutoff of host cell protein synthesis is observed after HSV infection of cultured cells (19). In contrast to wild-type virus infection, ORF-34.5/ORF-P double-mutant viruses induce the shutoff of all protein synthesis (viral and host cell) at early times (3 to 6 h p.i.) in SK-N-SH cells relative to other cell types (13). To examine protein shutoff in L/ST-n38- and L/ST-4BS-infected Vero and SK-N-SH cells, we performed metabolic labeling experiments. An ICP34.5-null mutant, 1771, derived from strain 17syn+ was included as a positive control, and wild-type strains KOS and 17syn+ were included as negative controls. In Vero cells, infection with the L/ST mutant viruses, both rescuants and the positive and negative control viruses, produced wild-type (KOS or 17syn+) protein profiles, characterized by a gradual shift from host to viral protein synthesis over the 12-h period (Fig. 8A). In SK-N-SH cells, L/ST-n38 and both rescuant viruses produced KOS wild-type-like protein profiles (Fig. 8B). In contrast, L/ST-4BS and 1771 infection resulted in rapid and extensive shutoff of both viral and cellular protein synthesis relative to their respective wild-type viruses. Rescue of the 4BS mutation to produce the wild-type genotype (L/ST-4BSR) returned the protein synthesis profile to that seen with wild-type virus. The results obtained with L/ST-n38 suggest that the premature protein shutoff phenotype of the ORF-P ORF-34.5 double-mutant viruses in SK-N-SH cells is not likely to be a consequence of mutations in ORF-P but, rather, is a consequence of mutations in ORF-34.5. The molecular basis for the premature protein shutoff phenotype of L/ST-4BS is not clear and will be discussed below.

View larger version (92K): [in this window] [in a new window]   FIG. 8.   Kinetics of the shutoff of viral and cellular protein synthesis in Vero and SK-N-SH cells infected with L/ST mutant viruses. Cells were mock infected (M) or infected with 10 PFU of the indicated viruses per cell, labeled with [35S]methionine for 30 min at 3, 7, or 12 h p.i., and processed for PAGE as described in Materials and Methods. (A) Vero cells, 5 µg of protein per lane. (B) SK-N-SH cells, 10 µg of protein per lane.

L/ST-4BS but not L/ST-n38 is attenuated for neurovirulence in mice. Having shown that the protein shutoff phenotype is characteristic of L/ST-4BS but not L/ST-n38, we tested these mutants in vivo for their neurovirulence phenotypes and their ability to establish and reactivate from latency in mice.

L/ST-4BS replicates in the mouse eye and in TGG during acute infection following corneal inoculation. Because the attenuation of L/ST-4BS in mice could, in theory, be due to a general defect in the ability of this virus to replicate in mice, we examined the ability of KOS, L/ST-4BS, and L/ST-4BSR to replicate in the mouse eye and in TGG following corneal inoculation. Tear film from four mice and TGG from two mice were collected daily during acute infection (the first 7 days p.i.) and assayed for infectious virus by a standard plaque assay on Vero cell monolayers. The arithmetic mean titers of multiple samples at each time point are shown in Fig. 9. In tear film, KOS and L/ST-4BSR reached peak titers of approximately 10⁵ PFU/ml on day 1 p.i., and these titers were maintained at high levels through day 6 p.i., dropping slightly on day 7 p.i. (Fig. 9A). Replication of these viruses in TGG was evident by day 2 and peaked between days 3 and 4 p.i. (Fig. 9B). In both tear film and TGG, L/ST-4BS replicated to near-wild-type levels during the first few days p.i. but exhibited an earlier and more rapid decline in titer thereafter. Rescue of the mutation in L/ST-4BS restored wild-type levels of replication in TGG as in eyes. Notably, these results demonstrate that L/ST-4BS is capable of significant replication (albeit less efficiently than the wild type) in the peripheral nervous system.

View larger version (34K): [in this window] [in a new window]   FIG. 9.   Production of infectious virus in mouse eyes and TGG during acute infection and in explanted ganglia during reactivation. Mice were inoculated with 2 × 106 PFU of KOS, L/ST-n38, L/ST-4BS, or L/ST-4BSR per eye. (A) Eye swabs were taken 1 to 3 h p.i. and then daily from 1 to 7 days p.i. Each point represents the arithmetic mean titer of eight eye swabs (four mice). (B) TGG were removed and assayed directly for the presence of infectious virus on days 1 through 7 p.i. Each point represents the arithmetic mean titer for TGG (two mice). The titers were determined by a standard plaque assay on Vero cells. Data points below the asterisk represent undetectable levels of virus. (C) At 30 days p.i., TGG were removed and cocultivated with Vero cells. Aliquots of the cocultivation medium were then assayed for infectious virus on the indicated days postexplant. Each point represents the cumulative percentage of cocultivated ganglia scoring positive for infectious virus. In experiment 1, the total numbers of TGG tested were as follows: KOS, 18; L/ST-n38, 18; L/ST-4BS, 24. In experiment 2, the total numbers of TGG used were as follows: KOS, 18; L/ST-n38, 19; L/ST-4BS, 18; L/ST-4BSR, 14.

L/ST-4BS exhibits delayed reactivation from latency in the mouse ocular model. Because the L/STs and/or OPP may be involved in some way in the establishment or reactivation of latency, we examined the ability of the L/ST mutant viruses to reactivate from latently infected mouse TGG. In two independent experiments, mice were inoculated with 2 × 10⁶ PFU/eye. On day 30 p.i., TGG were removed and cocultivated with Vero cells. The cocultivation medium was then assayed daily for the presence of infectious, reactivated virus. The cumulative percentage of cocultivated ganglia which scored positive for infectious virus in both experiments is shown in Fig. 9. Consistent with its wild-type phenotype in vitro, L/ST-n38 reactivated with wild-type kinetics and efficiency in both experiments. L/ST-4BS also reactivated but with delayed kinetics and reduced efficiency relative to wild-type virus in both experiments. In experiment 1, reactivation of L/ST-4BS was delayed somewhat and was less efficient than that of the wild type (71% TGG [17 of 24]) by day 15 postexplant versus 100% of TGG [24 of 24] by day 6). In this experiment, it was not determined whether the TGG remaining on day 15 harbored latent L/ST-4BS virus; therefore, in the second experiment, explant cocultures were monitored for an extended period (21 days postexplant). In this experiment, L/ST-4BS ultimately reactivated from 89% of TGG (16 of 18), a level comparable to KOS (100% [18 of 18 TGG]) and L/ST-n38 (100% [19 of 19 TGG]) in the 21-day experiment. The delayed kinetics of L/ST-4BS reactivation noted in the first experiment were also evident in the second experiment. Rescue of the 4BS mutation restored the wild-type reactivation profile (L/ST-4BSR, experiment 2). Viral DNA isolated from half of the reactivated cultures in experiment 1 and all of the cultures in experiment 2 was analyzed by Southern blot analysis; the L/ST mutations were present in reactivated mutant viruses in all cases (data not shown).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The abb'c'a'ca repeats of the HSV-1 genome contain genes and cis-acting elements that play central roles in establishing the programs of gene expression characteristic of productive infection and latency. The large number of these genes and elements and the extent to which they overlap have complicated the genetic and functional analysis of this region of the genome in that mutations introduced into one gene or element invariably alter the nucleotide sequence of other overlapping genes and elements. Such has been the case in studies of the ₁34.5 gene, which is encoded on the DNA strand opposite that specifying the L/STs and OPP (Fig. 10). For these reasons, we endeavored to design mutations that have minimal effects on ORF-34.5 in an effort to elucidate the function(s) of the L/STs and OPP.

View larger version (20K): [in this window] [in a new window]   FIG. 10.   Physical map of viruses mutated in the ORF-34.5 ORF-P ORF-O region. Diagram of the HSV-1 genome at the junction between the b' and a' sequences in the same orientation as shown in Fig. 1. Shown above the bars representing the b' and a' sequences are the positions of relevant restriction enzyme sites. Beneath the b' and a' sequences are the locations of ORF-34.5, translated from a transcript transcribed in the leftward direction, and ORF-P and ORF-O, translated from transcripts transcribed in the rightward direction. Beneath the ORFs are shown the corresponding b'a' region in wild-type viruses KOS and F and mutant viruses L/ST-n38, L/ST-4BS, and R7530. On the line representing wild-type virus, the positions and sequences of the ICP4 binding site in the L/ST promoter and of codon 38 of ORF-P are shown. Beneath the line representing wild-type viruses are shown the specific mutations in L/ST-4BS, L/ST-n38, and R7530.

OPP does not play a major role in neurovirulence or latency in murine models. To examine the contributions of mutations in ORF-P to the in vitro premature protein shutoff and in vivo neurovirulence and latency phenotypes of ORF-34.5 ORF-P double mutants, we constructed mutant virus L/ST-n38, containing a single-base-pair single-frame nonsense mutation at codon 38 in ORF-P. Before the n38 mutation was introduced into the viral genome, the phenotype of a plasmid carrying this mutation was established in vitro. In transient-expression assays, the n38 mutation terminated the translation of ORF-P prematurely but had no effect on the transcription of the L/STs (Fig. 3). In the context of viral infection, the n38 mutation (i) had no detectable effect on virus replication in neural (SK-N-SH) or nonneural (Vero) cells (Fig. 5); (ii) had no effect on the levels of the L/STs, ₁34.5, ICP0, or ICP4 transcripts or the LATs in neural (NB41A3) or nonneural (Vero) cells (Fig. 6); and (iii) did not produce premature shutoff of infected SK-N-SH cell protein synthesis (Fig. 8). Consistent with the absence of a unique phenotype in in vitro tests, L/ST-n38 was wild type with regard to neurovirulence, latency, and reactivation in mice. Although the n38 mutation in ORF-P is a single-base-pair, single-frame nonsense mutation, the G-to-A nucleotide transition also induced an Arg-to-Thr change at amino acid 93 of ORF-O (Fig. 10). If OOP is involved in any of the functions measured by our in vitro or in vivo assays, the Arg-to-Thr substitution in L/ST-n38 had no detectable effect on these functions.

L/ST-4BS exhibits in vitro and in vivo properties characteristic of ORF-P ORF-34.5 double mutants. The second mutation used to examine the function of ORF-P ORF-34.5 double mutants, 4BS, alters the ICP4 binding site 4 bp upstream of the transcriptional start site of the L/STs. As anticipated, tests conducted before the transfer of this mutation into the viral genome demonstrated that the 4BS mutation abrogates the binding of ICP4 to its cognate repressive binding site (Fig. 4A). Notably, this mutation had no effect on the ability of ICP4 or ICP0 to activate the expression of a reporter gene from the same mutant promoter in cotransfection and transfection/superinfection assays, demonstrating that the sequences responsible for ICP0- ICP4-mediated activation of the L/ST promoter do not include the repressive ICP4 binding site (Fig. 4B to D).