|Home | About | Journals | Submit | Contact Us | Français|
Trigeminal ganglia (TG) from rabbits latently infected with either wild-type herpes simplex virus type 1 (HSV-1) or the latency-associated transcript (LAT) promoter deletion mutant 17ΔPst were assessed for their viral chromatin profile and transcript abundance. The wild-type 17syn+ genomes were more enriched in the transcriptionally permissive mark dimethyl H3 K4 than were the 17ΔPst genomes at the 5′ exon and ICP0 and ICP27 promoters. Reverse transcription-PCR analysis revealed significantly more ICP4, tk, and glycoprotein C lytic transcripts in 17syn+ than in 17ΔPst. These results suggest that, for efficient reactivation from latency in rabbits, the LAT is important for increased transcription of lytic genes during latency.
During herpes simplex virus type 1 (HSV-1) latency in sensory neurons, there is an overall repression of transcription from the viral genome, with the exception of the latency-associated transcript (LAT) region. The latent genomes are maintained as nucleosome-associated episomes (4) that are not repressed through DNA methylation (6, 11). Rather, histone tail modifications appear to correspond with transcriptional permissiveness. Specifically, during latency in the mouse, theLAT promoter region is more enriched in acetylated histone H3 K9 or K14, a marker of transcriptional permissiveness, than are the lytic genes ICP27 and ICP0, and the LAT 5′ exon/enhancer region is more acetylated than the LAT promoter (11). Further, Wang et al. demonstrated that for mice infected with a wild-type virus, the HSV-1 genome's lytic gene promoters become associated with dimethyl histone H3 K9, a marker of transcriptional repression, during establishment of latency, while they become less associated with dimethyl H3 K4, a transcriptionally permissive mark (14). In contrast, a LAT promoter deletion virus has less-repressive histone modifications associated with lytic genes during latency (14). These patterns of histone modifications are consistent with a previous study demonstrating that a LAT promoter deletion mutant is more transcriptionally leaky during latency in mice, displaying increased ICP4 and tk transcript accumulation compared to those of wild-type HSV-1 (3). Taken together, these findings suggest that in the mouse LAT plays a role in the repression of lytic genes during latency.
While the mouse model is used extensively to study HSV-1 latency and has provided valuable insight into molecular events surrounding the infection, the rabbit eye model is generally considered a more relevant system to study reactivation and the events leading up to reactivation because it more closely parallels the biology of clinical HSV-1 reactivation in humans (for a review see reference 13). In the rabbit eye model, latency is established in the trigeminal ganglia (TG) following ocular infection with HSV-1. Reactivation can be efficiently induced by iontophoresis of adrenergic agents, such as epinephrine, and the reactivating virus can then be detected in the tears (7, 9). The rabbit is also one of the only HSV-1 reactivation models in which reactivation results in viral shedding and clinical lesions that recur at the primary site of infection (1). However, reactivation from latency in the rabbit eye model is more LAT dependent than that in mouse models; specifically, LAT promoter deletion mutants are severely reduced in reactivation relative to the wild type (8, 12). Since the role of LAT in facilitating efficient reactivation in the rabbit eye model is not known, we sought to determine whether a LAT promoter mutant exhibits alterations in its latent chromatin profile and RNA transcript accumulation, characteristics that could give new insight into the LAT's mechanism of action.
To determine the transcriptional permissiveness of the LAT region in latently infected rabbits, chromatin immunoprecipitation (ChIP) was performed using anti-dimethyl H3 K4 (Millipore) on TG from rabbits latently infected with 17syn+, following the procedure described previously (11). Bound and unbound fractions were analyzed in triplicate by TaqMan real-time PCR. The relative quantity of the bound fraction was normalized to that of the total of the bound plus unbound fractions. During latency in the rabbit, the chromatin profile of wild-type 17syn+ indicated that the LAT region was more transcriptionally permissive than lytic genes ICP0 and ICP27 (Fig. (Fig.1A;1A; Table Table1).1). This finding is consistent with the chromatin profiles observed in latently infected mice (11). We also assessed the viral chromatin profile of rabbits latently infected with 17ΔPst, a nonreactivating HSV-1 recombinant with a 202-bp deletion of the core LAT promoter (2, 5). A comparison between the wild type and this LAT promoter deletion virus showed that while the LAT region of 17ΔPst was more enriched in dimethyl H3 K4 than were the lytic genes examined (Fig. (Fig.1B;1B; Table Table1;1; P = 0.02 and P = 0.03 for ICP0 and ICP4, respectively), the level of enrichment was an average of 2.5-fold less (P = 0.004) than that observed for 17syn+. In addition, in 17ΔPst the level of H3 K4 dimethylation was approximately twofold lower for ICP0 (P = 0.1) and ICP27 (P = 0.1) than in 17syn+. These observations were in contrast to the chromatin profile of a LAT-negative mutant previously described for the mouse, in which the enhancer and ICP0 appear slightly more transcriptionally permissive than in the wild type (10).
Chen et al. previously reported that, during latency, a LAT promoter deletion virus displayed greater leakiness of lytic transcripts than did the wild type in mice infected via the ocular route (3). Because our ChIP analysis of rabbit TG revealed that the lytic genes of 17ΔPst are less transcriptionally permissive than those of 17syn+, we sought to determine the relative levels of lytic gene transcription of the two viruses during latency in the rabbit. RNA was isolated from latently infected rabbit TG; cDNA for each TG was synthesized simultaneously in four separate reactions using random decamers and then pooled to enable detection of low-abundance transcripts. The resulting cDNA was analyzed by real-time PCR. Relative quantities were first normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and the normalized values were further normalized to the relative viral genome quantity. For the 17ΔPst lytic genes, no transcripts were detectable by TaqMan real-time PCR at the limits of sensitivity. Therefore, values assigned in these cases represent the maximum possible quantity. As shown in Fig. Fig.2A2A and Table Table2,2, while there may have been some leaky expression of LAT in 17ΔPst during latency, it was almost 800-fold less (P = 0.017) than that observed for 17syn+. Assessment of lytic genes ICP4 (Fig. (Fig.2B),2B), thymidine kinase (tk) (Fig. (Fig.2C),2C), and glycoprotein C (gC) (Fig. (Fig.2D),2D), HSV-1 regions representative of the immediate-early, early, and late gene classes, respectively, confirmed that these RNAs are present at greater levels in rabbit TG latently infected with 17syn+ than in rabbit TG infected with 17ΔPst. Specifically, 17syn+ displayed averages of at least 3-, 35-, and 154-fold more RNA for ICP4, tk, and gC, respectively, than did 17ΔPst, strongly suggesting that the absence of LAT in 17ΔPst corresponds to a greater repression of lytic genes during latency in the rabbit. The observed differences were not due to variability in establishment, since the relative genome quantities were 0.0018 and 0.002 for 17syn+ and 17ΔPst, respectively (P = 0.84).
It had previously been shown that in mice latently infected with a LAT promoter deletion mutant, the HSV-1 genomes were less enriched in the repressive histone dimethyl H3 K9 than with the wild type (14). Therefore, in order to examine the association of a repressive histone modification with the latent HSV genome in the rabbit, a ChIP experiment was performed on latently infected rabbit TG using anti-trimethyl H3 K9 (Millipore). The 17ΔPst lytic genes tested (ICP4, tk, and gC) showed levels of enrichment in this histone modification similar to those of 17syn+ (data not shown). This indicates that, unlike during latency in the mouse, HSV-1 LAT-negative genomes in the rabbit do not become less enriched in repressive histone marks relative to the wild type, therefore indicating that the LAT does not seem to exert a repressive effect on the chromatin state of latent genomes in rabbits.
The findings presented here for the transcriptional status of a LAT-negative mutant during latency in the rabbit are the opposite of what has been previously observed for the mouse, where the LAT appears to play a role in repression of the latent HSV-1 genome (3, 14). To the contrary, in the rabbit LAT seems to exert a positive effect on facilitating transcriptional permissiveness of the lytic genes, both at the level of H3 K4 dimethylation and at the level of viral transcripts detected in latent ganglia. Further, when the average number of RNA molecules per viral genome (values calculated through extrapolation of a DNA standard curve derived from TaqMan real-time PCR) is compared with those determined by previous analyses in the mouse, the overall abundance of lytic transcripts detected is an order of magnitude less in the rabbit than in the mouse (Table (Table3),3), suggesting that control of viral transcription in the rabbit is more constrained than that in the mouse.
In summary, our results suggest that in contrast to what occurs in the mouse, in the rabbit LAT does not act as a repressor of lytic genes but instead acts to keep the lytic genome in a more transcriptionally activated state. We believe that in the rabbit eye model, the expression of LAT is important in establishing or maintaining the lytic genes in a state that is poised for reactivation.
This work was supported in part by NIH grants AI48633 (D.C.B.) and EY006311 (J.M.H.) as well as an Investigator in the Pathogenesis of Infectious Disease Award from the Burroughs Wellcome Fund (D.C.B.) and a Research to Prevent Blindness, Senior Scientific Investigator Award (J.M.H.), and an LSUHSC Translational Research Initiative Grant (P.S.B.). D.M.N. was supported by an individual NRSA, EY016316, and N.V.G. was supported by training grant AI07110 from the NIH.
We thank J. Feller, L. Watson, and Z. Zeier for helpful comments on the manuscript.
Published ahead of print on 9 April 2008.