INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

After ocular herpes simplex virus type 1 (HSV-1) ocular infection, the virus travels up nerves to the trigeminal ganglia (TG), where it establishes a latent infection. Latency lasts for the life of the infected individual. HSV-1 can reactivate from latency and travel back to the eye, where it can be detected in tears and cause recurrent corneal disease. Recurrent ocular HSV-1 can lead to scarring of the cornea and loss of sight. In developed nations, HSV-1 is the most common cause of corneal blindness due to an infectious agent (12). How HSV-1 establishes, maintains, and reactivates from latency remains unknown.

The latency-associated transcript (LAT) is the only viral gene that is abundantly transcribed during latency (20). LAT is located in the long repeat region of the viral genome and thus is present in two copies per genome. LAT is initially transcribed as an 8.3-kb RNA (4, 26). This primary LAT transcript gives rise to a family of LAT RNAs, including the very stable 2-kb LAT (20, 24, 25), which appears to be an intron produced by splicing (6). LAT transcription-negative mutants have been shown to reactivate poorly by explant or induced reactivation in the mouse (9, 10, 21), by induced reactivation in the rabbit (1, 23), and by spontaneous reactivation in the rabbit (14, 17). Thus, LAT is essential for efficient, wild-type reactivation from sensory neurons.

The mechanism by which LAT functions remains unknown. No LAT-encoded protein has been detected during latency and there does not appear to be a LAT open reading frame that is well conserved among LAT genes capable of sustaining spontaneous reactivation (5). Thus, in the absence of undetected, atypical splicing, it is unlikely that LAT's function is due to a LAT protein. LAT's function also does not appear to be due to antisense downregulation of the important immediate-early gene ICP0, which LAT overlaps in an antisense direction. We recently showed that the first 1.5 kb of LAT alone is sufficient for wild-type levels of spontaneous reactivation (17). This region does not overlap any portion of any known HSV-1 gene.

Mapping a LAT spontaneous reactivation function to the first 1.5 kb of LAT was done by inserting the LAT promoter and the first 1.5 kb of LAT into an ectopic location in the genome of a LAT null mutant between HSV-1 genes UL37 and UL38 (17). This completely restored wild-type levels of spontaneous reactivation. Another LAT mutant, dLAT371, containing a StyI-StyI deletion that removed LAT nucleotides 76 to 447 was also wild type for spontaneous reactivation (18). Thus, this 371-nucleotide region within the first 1.5 kb of the primary LAT transcript did not appear to be essential for efficient spontaneous reactivation.

We report here on a mutant that is a combination of LAT3.3A and dLAT371. This virus, designated LAT2.9A, contains a LAT insert at the same ectopic location as LAT3.3A. This insert is identical to that in LAT3.3A (i.e., the LAT promoter and the first 1.5 kb of LAT), except that LAT nucleotides 76 to 447 are deleted. As with LAT3.3A, the LAT promoter and the first 1.67 kb of LAT are deleted from both copies of LAT in the long repeat and therefore the only LAT produced originates from the ectopic insert. We predicted that LAT2.9A, like LAT3.3A and dLAT371, would have wild-type spontaneous reactivation. Instead, LAT2.9A had a significantly reduced spontaneous reactivation rate that was comparable to that of the LAT null mutant dLAT2903. LAT2.9A also was significantly more virulent than wild-type virus. This was also unexpected, since dLAT2903, LAT3.3A, and dLAT371 all appeared to have wild-type virulence.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Cells and virus. Rabbit skin (RS) cells were grown in Eagle minimal essential media (MEM) supplemental with 5% fetal calf serum (FCS). CV-1 cells were grown in MEM supplemented with 10% FCS. CV-1 cells were used for growth kinetic studies. RS cells were used for all other tissue culture procedures, including the preparation of virus stocks. All mutants were derived from HSV-1 strain McKrae. The parental McKrae virus and all mutants were triple plaque purified and passaged only one or two times prior to use. The construction and properties of dLAT2903, LAT3.3A (previously designated LAT1.5a), and dLAT371 have been previously described (14, 17, 18).

Construction of LAT2.9A. The parental virus for this construct was dLAT2093, a mutant of HSV-1 strain McKrae containing a 1.8-kb (EcoRV-HpaI) deletion in both copies of LAT that removed 0.2 kb of the LAT promoter and 1.6 kb of the 5' end of the primary 8.3-kb LAT transcript (14). The previously cloned EcoRI A fragment from HSV-1 strain McKrae (15) was digested with BamHI, and the products were separated by agarose gel electrophoresis. A resulting 7.5-kb band containing the McKrae genomic region including UL37 and UL38 was isolated by electroelution and cloned into the BamHI site of plasmid pEV-vrf3 (3, 15) to produce the plasmid pV375. pV375 was digested with AflII, the overhang was filled in using the Klenow fragment, and the blunt ends were self-ligated to create a unique PacI site in the plasmid between the sequences for UL37 and UL38. The resulting plasmid, designated pV375Pac, was amplified by transformation into Escherichia coli RR1CI857 according to standard protocol. An HpaI-HpaI restriction fragment consisting of 1.8 kb of the LAT promoter and the first 1.5 kb of the LAT RNA (17) was cloned into the PacI site of pNEB193 and further digested with StyI to remove a 371-nucleotide StyI-StyI region corresponding to LAT nucleotides 76 to 447. The plasmid was then self-ligated and digested with PacI, and the resulting 2.9-kb band was cloned into the PacI site of pV375Pac to produce pV375LAT2.9.

Replication of virus in tissue culture. CV-1 cell monolayers at approximately 70 to 80% confluency were infected with virus at 0.01 PFU/cell, and all monolayers were refed with exactly the same amount of MEM containing 10% FCS. Virus was harvested for titration at various times by two cycles of freeze-thawing the monolayers plus media (80°C to room temperature). The PFU/milliliter values were determined by standard plaque assays on RS cells.

Rabbits. Eight- to ten-week old New Zealand White female rabbits (Irish Farms) were used for all experiments. Rabbits were treated in accordance with ARVO (Association for Research in Vision and Ophthalmology), AALAC (American Association for Laboratory Animal Care), and NIH (National Institutes of Health) guidelines.

Rabbit model of ocular HSV-infection, latency, and spontaneous reactivation. Rabbits were bilaterally infected without scarification or anesthesia by placing 2 × 10⁵ PFU of HSV-1 per eye into the conjunctival cul-de-sac, closing the eye, and rubbing the lid gently against the eye for 30 s (20). At this dose of HSV-1 McKrae virtually all of the surviving rabbits harbor a bilateral latent HSV infection in both trigeminal ganglia, resulting in a high group rate of spontaneous reactivation with the McKrae strain of HSV-1. Latency is assumed to have been established by 28 days postinfection. Acute ocular infection of all eyes was confirmed by HSV-1 positive tear film cultures collected on days 3 and 4 postinfection.

Detection of spontaneous reactivation by ocular shedding. Beginning on day 31 postinfection, tear film specimens were collected daily from each eye for 26 days as previously described (22), using a nylon-tipped swab. The swab was then placed in 0.5 ml of tissue culture medium and squeezed, and the inoculated medium was used to infect primary rabbit kidney cell monolayers. These cell monolayers were observed in a masked fashion by phase light microscopy for up to 30 days to monitor HSV-1 cytopathic effects (CPE). All positive monolayers were blind passaged onto fresh cells to confirm the presence of virus. DNA was purified from randomly selected positive cultures derived from latently infected rabbits and analyzed by restriction enzyme digestion and Southern blots to confirm that the CPE was due to reactivated HSV-1 and that the reactivated virus was identical to the input virus.

Virus replication in rabbit eyes. Tear films were collected as described above on various days postinfection. The amount of virus in each tear film was determined by standard plaque assays on RS cells.

RT-PCR. Reverse transcriptase (RT)-PCR was done as we previously described (17) with minor modifications. Briefly, RNA was isolated with Trizol (Gibco-BRL, Grand Island, N.Y.) from individual TG from latently infected rabbits or from infected CV-1 cell monolayers and treated with DNase I (12 U, 37°C, 30 min; Stratagene, La Jolla, Calif.). RNA was isolated with an RNeasy Mini-Kit (Quiagen, Santa Clarita, Calif.). The purified RNAs were subjected to first-strand cDNA synthesis by Superscript II (Gibco-BRL) according to the manufacturer's protocol. The primer for first-strand cDNA synthesis from the LAT RNA was 5'-CTTTGTTGAACGACACCGGGGCGCCCTCGA-3'. The cDNA product was then amplified by PCR with the primer 5'-CCACAACGGCCCGGCGCATGCGCTGTGGTT-3' and the first-strand primer. These primers generate a 160-bp product specific for LAT nucleotides 471 to 631. The amplified products were fractionated by gel electrophoresis, transferred to a nylon membrane, and hybridized to the ³²P-labeled internal probe 5'-TCTCCCCCCCCCCTTCTTCACCCCCAGTAC-3' corresponding to LAT nucleotides 550 to 580.

Statistical analysis. Statistical analyses were performed by using Instat, a personal computer software program. For analyses with either the Student t test, the Mann-Whitney rank sum test, the chi-square test, or the Fisher exact test the results were considered statistically significant when the P value was <0.05.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Structure of LAT2.9A. The genomic structures of wild-type HSV-1 McKrae, LAT2.9A, and the other viruses used in this study are shown in Fig. 1A. All of the viruses were derived from HSV-1 strain McKrae. The construction and properties of dLAT2903 and its marker-rescued virus dLAT2903R, dLAT371 and its marker-rescued virus dLAT371R, and LAT3.3A (previously designated LAT1.5a) have been described previously (14, 17, 18). The LAT transcript(s) made by each virus are detailed in Fig. 1B. Wild-type McKrae and the marker-rescued viruses dLAT2903R and dLAT371R contain two copies of LAT, one in each viral long repeat (Fig. 1A, top). The viral long repeats are expanded (in dashed lines) to show the relative location and status of the LAT gene. In the topmost panel, the location of the ICP0 and ICP34.5 genes are shown for reference. The primary LAT transcript in the wild-type and the marker-rescued viruses is approximately 8.3 kb (Fig. 1A, top panel, and Fig. 1B) (24, 25). A very stable and easily detected 2-kb LAT (solid rectangle) appears to be an intron derived by splicing of the primary LAT (6). dLAT2903 contains a deletion in both copies of LAT from 161 to +1,667 relative to the start of the primary LAT transcript (Fig. 1A, indicated by "XXXXX"). This virus is missing key promoter elements, makes no LAT RNA (Fig. 1B), and is a true LAT null mutant. dLAT371 contains a 371-nucleotide deletion of LAT nucleotides 76 to 447, corresponding to a StyI-StyI region prior to the 2-kb LAT (Fig. 1A, third panel, indicated by "X"). This virus makes a normal primary LAT transcript except that it is missing LAT nucleotides 76 to 447 (Fig. 1B). LAT3.3A is derived from dLAT2903 by insertion of 1.8 kb of the LAT promoter and the first 1.5 kb of LAT into a unique PacI site that was constructed between UL37 and UL38 (Fig. 1A). LAT3.3A and LAT2.9AR make no LAT RNA from either copy of LAT in the long repeats, but they do make a 1.5-kb LAT RNA from the ectopic insert that corresponds to the first 1.5 kb of the primary LAT (Fig. 1B). LAT2.9A is identical to LAT3.3A, except that LAT nucleotides 76 to 447 have been deleted from the inserted 1.5-kb LAT region (Fig. 1A). LAT2.9A therefore transcribes an RNA of 1,128 nucleotides corresponding to LAT nucleotides 1 to 76 and 447 to 1,499. The construction of all the above mutants except LAT2.9A and LAT2.9AR have been previously described (14, 17, 18). Additional details of the construction of LAT2.9A and LAT2.9AR are given in Materials and Methods.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   Construction and structure of LAT2.9A. (A) The top panel (wt McKrae, dLAT2903R, and dLAT371R) shows a schematic representation of the genome of wild-type McKrae and the wild-type marker-rescued viruses dLAT2903R and dLAT371R, which are identical to wild-type HSV-1. The prototypic orientation of HSV-1 shown here contains a unique long region and a unique short region, each bounded by inverted repeats. The unique regions are shown as solid lines. The repeats are shown as open rectangles. UL, unique long; US, unique short; TRL, terminal repeat long; IRL, internal repeat long; IRS, internal repeat short; TRS, terminal repeat short. The lines with arrows under the genome indicate the locations and directions of the LAT, ICP34.5, and ICP0 transcripts. The solid rectangle within the primary 8.3-kb LAT transcript indicates the location of the stable 2-kb LAT. TATA indicates the location (in the genomic DNA) of the LAT promoter TATA box. The next panel shows that dLAT2903, the previously described LAT null mutant (14), has a deletion extending from LAT nucleotides 161 to +1667. This deletion encompasses the LAT promoter, and dLAT2903 therefore cannot transcribe any LAT RNA. The dashed rectangle and the preceding dashed line covered by "XXXXX" represent the portion of LAT that is deleted from both long repeats. Panel dLAT371 shows a previously described mutant containing a deletion of LAT nucleotides 767 to 447 in both copies of LAT (18). The blow-up in panel LAT3.3A and LAT2.9R shows the location of the 3.3-kb LAT fragment inserted between genes UL37 and UL38 in the unique long region of the LAT deletion mutant dLAT2903 to generate the virus LAT3.3A (17). The insert contains 1.8 kb of the LAT promoter and the first 1.5 kb of the primary LAT. The insertion site is outside the domains of the UL37 and UL38 promoters, and these genes are not affected (17). The structure of LAT2.9AR, a marker-rescued virus derived from LAT2.9A (see below) is identical to LAT3.3A. Panel LAT2.9A shows that LAT2.9A contains the same insert between UL37 and Ul38 as LAT3.3A, except that the StyI-StyI region (LAT nucleotides 76 to 447) has been deleted from the insert. This deletion corresponds to the deletion in dLAT371 described above. (B) The LAT RNAs made by each of the viruses shown in panel A are indicated. Wild-type McKrae, dLAT2903R, and dLAT371R all make the complete primary LAT transcript. The solid rectangle starting at LAT nucleotide 662 indicates the location of the stable 2-kb LAT. dLAT2903 makes no LAT RNA. dLAT371 transcribes all of the LAT except nucleotides 76 to 447. LAT3.3A and LAT2.9AR make no LAT from the original LAT region (one in each long repeat), but they do transcribe the first 1.5 kb of LAT from the LAT insert between UL37 and UL38. LAT2.9A also makes not LAT from the original LAT region. It transcribes LAT nucleotides 1 to 76 and 447 to 1499 (i.e., the first 1.5 kb of LAT minus the StyI-StyI region).

Southern analysis of the structure of LAT2.9A. Southern analyses of viral DNAs were performed to confirm the structure of LAT2.9A (Fig. 2). DNAs were individually digested with PacI and probed with a ³²P-labeled restriction fragment (StyI-HpaI; LAT nucleotides 447 to 1,499) that is completely within the LAT region deleted in dLAT2903 and that therefore can hybridize only with LAT sequences in the LAT2.9A insert (Fig. 2A). The wild-type virus (lane 1) produced a single large band of genome size as expected, since PacI does not cut wild-type HSV-1. The dLAT2903 DNA produced no band (lane 2), since the sequences corresponding to the probe are deleted from this virus. PacI should cut the PacI sites flanking the 2.9-kb LAT fragment inserted between UL37 and UL38 in LAT2.9A to produce a single band of 2.9 kb that hybridizes to the probe as is seen in lane 3. Similarly, PacI cuts the correspondingPacI sites in LAT2.9AR and LAT3.3A, producing bands of 3.3 kb (lanes 4 and 5). These results indicate that LAT2.9A contains the expected 2.9-kb LAT insert between UL37 and UL38 and that the corresponding region in LAT2.9AR has been rescued back to the wild-type size (lane 4) seen in LAT3.3A (lane 5).

View larger version (92K): [in this window] [in a new window]   FIG. 2.   Southern analysis of LAT2.9A. (A) Viral DNAs were isolated, individually digested with PacI, and hybridized to a 32P-labeled probed corresponding to part of the LAT region deleted in dLAT2903 but present in wild-type McKrae and the LAT insert in LAT2.9A (see text). (B) The viral DNAs were digested with BamHI and probed with the BamHI-H restriction fragment that hybridizes to the UL37-UL38 region, but not to LAT (see text). (C) The viral DNAs were digested with BamHI and probed with a cloned HpaI-MluI restriction fragment corresponding to LAT nucleotides 1,667 to 2850 (see the text) that should hybridize to LAT DNA from the long repeats but not to the LAT inserts in the UL37-UL38 region (see text). Lanes: 1, wild-type McKrae; 2, dLAT2903; 3, LAT2.9A; 4, LAT2.9AR; 5, LAT3.3A.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Replication of LAT2.9A (A) Semiconfluent monolayers of CV-1 cells were infected with 0.01 PFU of the indicated virus per cell. At various times, the infected cell monolayers were harvested by freeze-thawing and the amount of virus was determined by plaque assay on RS cells. Each time point is the average of two determinations. (B) Tear swabs were performed on the eyes of rabbits infected with the indicated virus at various times postinfection. The amount of virus present in individual tear swabs was determined by plaque assay on RS cells. Each point represents the mean titer from five eyes, each from a different rabbit. Error bars indicate the standard deviation.

In vivo replication of LAT2.9A. Rabbits were infected with 2 × 10⁵ PFU of LAT2.9A, dLAT2903, dLAT2903R, or LAT3.3A per eye. Tears were collected at the indicated times (Fig. 3B), and the virus yield was determined by plaque assay. As we have previously shown (14, 17), replication of dLAT2903 and LAT3.3A in rabbit eyes was similar to that of wild type (dLAT2903R) virus. Replication of LAT2.9A (Fig. 3B, open circles) appeared to be slightly lower than that of LAT3.3A and dLAT2903R (solid circles and squares) on days 3 and 5 but not on day 7. However, as indicated by the overlapping error bars, these minor differences were not significant.

Reduced survival of rabbits ocularly infected with LAT2.9A. Eighteen rabbits per group were infected with 2 × 10⁵ PFU of LAT2.9A, dLAT2903, dLAT2903R, or LAT3.3A per eye. Only 17% of the LAT2.9A-infected rabbits survived for 21 days (Table 1, experiment 1, column 3) compared to 33-39% survival for the other groups. Since dLAT2903, dLAT2903R, and LAT3.3A all have wild type parental McKrae virulence in rabbits (14, 17), it appeared that survival of LAT2.9A may have been reduced compared to wild type. However, the differences were not significant (Table 1, experiment 1, column 4). To determine if the tendency in experiment 1 for LAT2.9A to be more virulent than wild type was meaningful, additional experiments were done. In a second experiment, 22 rabbits per group were infected with LAT2.9A or dLAT2903R (Table 1, experiment 2). Again, the percentage of rabbits surviving infection with LAT2.9A appeared to be reduced compared to wild type (32 versus 55%). However, as in experiment 1, the differences was not statistically significant (Table 1, column 3).

Reduced spontaneous reactivation of LAT2.9A. Beginning at 31 days postinfection (at which time latency had already been established), all eyes from the surviving rabbits in experiments 1, 2, and 3 were swabbed once a day for 26 days to collect tear films for analysis of reactivated virus as described in Materials and Methods. In experiment 1, LAT2.9A was compared to dLAT2903, which has reduced spontaneous reactivation, and to LAT3.3A and dLAT2903R, both of which have spontaneous reactivation rates identical to wild-type McKrae. The cumulative number of virus-positive tear film cultures is shown in Fig. 4A. Because of the different numbers of surviving rabbits in the different groups, the data were standardized to represent cumulative positive cultures per eye. The cumulative spontaneous reactivation rate in rabbits latently infected with LAT2.9A (Fig. 4A, open circles; approximately 1.8/eye on day 26) appeared to be less than dLAT2903R and LAT3.3A (Fig. 4A, solid squares and circles; approximately 3 to 3.5/eye) and slightly greater than dLAT2903 (Fig. 4A, open squares; approximately 1/eye).

View larger version (20K): [in this window] [in a new window]   FIG. 4.   Spontaneous reactivation of LAT2.9A. Each panel shows an independent experiment, each of which was performed at a different time. Rabbits were infected with 2 × 105 PFU of the indicated virus per eye. Rabbits surviving past day 21 are considered to have a latent HSV-1 infection in both TG. Beginning 31 days postinfection (day 1 of the collection period), tear swabs were collected daily for 26 days from all eyes and plated on RS to look for the presence of spontaneously reactivated virus. The cumulative number of virus-positive cultures divided by the number of eyes in the group is shown. Panels A, B, and C indicate the results of experiment 1, 2, and 3, respectively. dLAT2903R, LAT3.3A, dLAT371R, all of which have previously been shown to have wild-type McKrae spontaneous reactivation rates, were used as positive controls in various experiments.

Transcription of LAT in tissue culture and the TG of rabbits latently infected with LAT2.9A. To confirm that transcription of LAT in LAT2.9A was as expected, RT-PCR analyses were done. CV-1 cells were infected at a multiplicity of infection of 2, and RT-PCR was performed on total RNA. The primers used generate a 160-bp product specific for LAT nucleotides 471 to 631. The RT-PCR products were subjected to Southern analysis by using an internal ³²P-labeled probe (LAT nucleotides 550 to 580) (Fig. 5A). As expected, a 160-bp RT-PCR product was detected in cells infected with dLAT371, LAT3.3A, or LAT2.9A (Fig. 5A). In addition, the apparent intensity of the RT-PCR bands in each lane was similar. Since we previously showed that the amount of LAT RNA in TG latently infected with dLAT371 or LAT3.3A is similar to that of wild-type McKrae (17, 18), these results suggest that in tissue culture the predicted LAT2.9A RNA was present in wild-type amounts. No RT-PCR product was produced from dLAT2903-infected cells, confirming that RNA corresponding to this region of LAT was not made by the dLAT2903 LAT null mutant.

View larger version (29K): [in this window] [in a new window]   FIG. 5.   Transcription of LAT in tissue culture cells and in rabbits latently infected with LAT2.9A. (A) CV-1 cell monolayers were infected with the indicated viruses at a multiplicity of infection of 2, and the total RNA was isolated, RT-PCR was performed with primers corresponding to LAT nucleotides within the predicted LAT2.9A RNA, and Southern analysis was performed with an internal 32P-labeled probe as described in Materials and Methods. (B) Total RNA was isolated from individual TG from rabbits latently infected with the indicated virus, and RT-PCR analyses were performed and analyzed by Southern analysis as above. Each lane shows the RT-PCR product from one TG. The lane labeled ddH2O is a negative control. The RT-PCR analyses shown in the last three lanes were done at a different time than the other analyses.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

In the experiments reported here, LAT2.9A had increased virulence (based on reduced rabbit survival) and decreased spontaneous reactivation. In contrast, its marker-rescued virus, LAT2.9AR, was wild type for both phenotypes. This very strongly suggests that the increased virulence and the decreased spontaneous reactivation of LAT2.9A were both due to the 371-nucleotide StyI-StyI deletion. Both of these phenotypes were unexpected.

In contrast to the StyI-StyI deletion in LAT2.9A, we previously showed that deletion of the StyI-StyI region in the context of the entire LAT gene (dLAT371) (18) did not effect spontaneous reactivation. In addition, we previously showed that insertion of the LAT promoter and the first 1.5 kb of LAT (LAT 3.3A) into an ectopic location in the LAT null mutant dLAT2903 restored wild-type spontaneous reactivation (17). LAT2.9A is identical to LAT3.3A except that it contains the StyI-StyI deletion. Thus, it was expected that LAT2.9A would have a wild-type rate of spontaneous reactivation.

Possible explanations for the reduced reactivation of LAT2.9A fall into at least three categories. (i) The first is the LAT copy number. It is possible that the StyI-StyI region is not essential for wild-type spontaneous reactivation when two copies of LAT are present in the genome. (ii) The second is the LAT location. It is possible that when LAT is located in its normal location in the long repeats, the StyI-StyI region is not required, but the StyI-StyI region is required for LAT function when LAT is not in its normal location. (iii) The third is the lack of LAT from 1.5 to 8.3 kb. The StyI-StyI region may not be required in the context of the otherwise-complete LAT gene, but it is needed when only the first 1.5 kb of LAT is present. This would suggest that there are one or more additional functional regions downstream of the first 1.5-kb region. Since both dLAT371 and LAT3.3A have wild-type rates of spontaneous reactivation, the region pre-1.5 kb and the region post-1.5 kb must each be able to produce wild-type levels of spontaneous reactivation, but the effects of this must not be additive. This would easily occur if each functional region resulted in the maximum spontaneous reactivation possible in the rabbit model. Although not easily detected in animal models, the presence of multiple independent functional regions within LAT may have beneficial effects in nature. It would also help explain why selective pressure has maintained the entire 8.3-kb LAT gene in all of the HSV-1 isolates examined, despite the fact that the first 1.5 kb of LAT alone is sufficient for maximum levels of spontaneous reactivation in the rabbit. Alternatively, functional regions after the first 1.5 kb of LAT may not be fully independently able to produce wild-type levels of spontaneous reactivation but may be able to compensate for the StyI-StyI deletion in the first 1.5 kb of LAT, perhaps via interactions with part of the remaining 1.5-kb region. It is also possible that a previously undetected splicing event occurs, either partially or completely within the StyI-StyI region, and that this splicing is required for LAT function in the context of the first 1.5 kb but that it can be compensated for by sequences downstream of the first 1.5 kb.

Although LAT may partially inhibit productive gene expression (2, 7, 11), there have been no previously reports of any LAT mutants with increased virulence. Over the course of numerous experiments in our laboratory, the virulence of the LAT null mutant, dLAT2903 (14), has always been similar to that of its wild-type parent. In addition, the virulence of LAT3.3A has been indistinguishable from that of both its direct parent, dLAT2903, and dLAT2903's wild-type McKrae parent (17). In contrast, LAT2.9A, which is effectively a deletion of the StyI-StyI region from LAT3.3A, had significantly increased virulence. Moreover, the increased virulence of LAT2.9A was restored to the less-virulent wild-type level by rescue of the StyI-StyI deletion in the ectopic insert (LAT2.9AR). Thus, the increased virulence of LAT2.9A appeared to be due to the 371-nucleotide StyI-StyI deletion in its ectopic LAT insert.

Why does the deletion of LAT nucleotides 76 to 447 in LAT2.9A increase virulence when the deletion of LAT nucleotides 161 to +1667 (dLAT2903) does not? Asked another way, why does the insertion of LAT nucleotides 1800 to +76 plus 447 to 1667 between UL37 and UL38 in dLAT2903 increase virulence, whereas the insertion of LAT nucleotides 1800 to +1667 does not? or, from yet another perspective, why does preventing transcription of LAT nucleotides 76 to 447 along with LAT nucleotides 1667 to 8323 increase virulence, whereas preventing all LAT transcription does not alter virulence? This phenomenon, in which a small deletion has a larger effect than a larger deletion that encompasses the smaller deletion, is reminiscent of results often seen when mapping promoter activity. This is because promoters often contain numerous functional elements, some of which work in concert and some of which act antagonistically. A series of deletions that removes successively larger regions of the 5' end of a promoter often produces a pattern of promoter activity that increases and decreases several times before all activity is gone. Since it is unlikely that LAT encodes a protein and since deletion of the StyI-StyI region does not appear to alter the expression of LAT itself, this parallel suggests that the LAT RNA may regulate the expression or function of one or more viral and/or cellular genes. This regulation could be direct or through the intervention of other gene products. In either situation, LAT could enhance spontaneous reactivation by being a key factor involved in intricate, controlled regulation of viral and/or cellular genes involved in the latency-reactivation cycle.

We previously reported that d34.5, a McKrae-based mutant deleted for both copies of 34.5, the gene for ICP34.5 (one in each long repeat), had dramatically decreased virulence and poor spontaneous reactivation after infection of rabbits with 2 × 10⁵ PFU/eye (19). However, after an extremely high-dose ocular infection with over 10⁸ PFU/eye, d34.5 was still avirulent but its spontaneous reactivation rate was wild type (16). A second mutant, derived from d34.5 and designated d34.5A, contains one copy of 34.5 inserted into the same ectopic location used for LAT2.9A. d34.5A, has wild-type spontaneous reactivation while remaining much less virulent than wild type (13). These previous studies showed that decreased virulence does not necessarily result in decreased spontaneous reactivation. The results reported here extend and complement the above findings by showing that increased virulence does not necessarily result in increased spontaneous reactivation. In 2.9A, increased virulence did not result in increased spontaneous reactivation and, in fact, was coincident with decreased spontaneous reactivation. These results confirm that the phenotypes for spontaneous reactivation and virulence are separable.

Nonetheless, since LAT2.9A shows that LAT affects virulence as well as reactivation, is of interest to consider the possibility that in this mutant the increased virulence and the decreased spontaneous reactivation were related. It is possible that the increased virulence resulted in the elimination of neurons that would otherwise have become latently infected. This might result in fewer surviving latently infected neurons, resulting in a smaller pool of latently infected neurons capable of reactivation. This, in turn, would be expected to result in reduced spontaneous reactivation. Thus, one possible hypothesis that could be derived from these studies is that the normal function of LAT is to protect acutely infected neurons from death, thereby producing a larger pool of latently infected neurons. This larger pool of latently infected neurons would subsequently allow for increased spontaneous reactivation. Consistent with this, it has been proposed that the bovine herpes virus LR-RNA (latency-related RNA, comparable to LAT) encodes products that promote neuronal survival (8, 22). Unfortunately, in the rabbit model it has not yet been possible to detect decreased amounts of latent viral DNA or decreased numbers of latent viral DNA-positive neurons in the TG (13), and so this hypothesis remains to be tested.