Selective degradation of mRNAs by the HSV host PNAS PLUS shutoff RNase is regulated by the UL47 tegument

Minfeng Shu, Brunella Taddeo, Weiran Zhang, and Bernard Roizman1

Marjorie B. Kovler Viral Oncology Laboratories, The University of Chicago, Chicago, IL 60637

Contributed by Bernard Roizman, March 21, 2013 (sent for review February 25, 2013) 1 (HSV-1) encodes an endoribonuclease that is mRNAs are rapidly degraded 5′ to 3′ (2, 17). One host response responsible for the shutoff of host protein synthesis [virion host to infection is the synthesis of numerous stress response mRNAs shutoff (VHS)-RNase]. The VHS-RNase released into cells during carrying adenylate-uridylate (AU)-richelements(AREs)intheir3′ infection targets differentially four classes of mRNAs. Thus, (a) VHS- UTR. In uninfected cells, AREs’ mRNAshaveashortlifeof∼1h. RNase degrades stable cellular mRNAs and α (immediate early) viral In cells infected with wild-type virus, those mRNAs are dead- mRNAs; (b) it stabilizes host stress response mRNAs after deadeny- enylated and cleaved close to the AREs. Though the 3′ portion is lation and subsequent cleavage near the adenylate-uridylate (AU)- rapidly degraded, the 5′ domain remains stable in the cytoplasm. rich elements; (c) it does not effectively degrade viral β or γ mRNAs; Finally, recent studies from our laboratory (18) showed that though and (d) it selectively spares from degradation a small number of the VHS-RNase degrades α (immediate early) mRNAs, it essen- cellular mRNAs. Current evidence suggests that several viral and at tially spares from degradation β (early) and γ (late) viral mRNAs least one host protein (tristetraprolin) regulate its activity. Thus, vi- accumulating during productive infection with wild-type virus. rion protein (VP) 16 and VP22 neutralize the RNase activity at late The degradation of stable host mRNAs has been attributed to the times after infection. By binding to AU-rich elements via its interac- binding of VHS-RNase to the cap structure via its affinity for eIF4H tion with tristetraprolin, the RNase deadenylates and cleaves the (19, 20). Studies done to date suggest that VHS-RNase decaps the mRNAs in proximity to the AU-rich elements. In this report we show ′ ′

mRNA, and that the uncapped mRNA is degraded 5 to 3 (21, 22). MICROBIOLOGY that another virion protein, UL47, brought into the cell during infec- With respect to the AREs’ mRNAs, the available data focused on tion, attenuates the VHS-RNase activity with respect to stable host tristetraprolin, a protein that is induced in infected cells, and whose and viral α mRNAs and effectively blocks the degradation of β and γ mRNA is not degraded by VHS. In uninfected cells, tristetraprolin mRNAs, but it has no effect on the processing of AU-rich mRNAs. The binds AREs and redirects the mRNAs to P bodies or exosomes. α properties of UL47 suggest that it, along with the protein infected Tristetraprolin binds VHS-RNase (23). One interpretation of the cell protein 27, attenuates degradation of mRNAs by the VHS-RNase data is that VHS is deflected from the cap structure, binds triste- through interaction with the enzyme in polyribosomes. Mutants traprolin at the AREs, and cleaves the mRNA 5′ to the bound en- α lacking both VHS-RNase and UL47 overexpress genes and delay zyme. The cleaved 3′ domain is not protected by a cap and is rapidly the expression of β and γ genes, suggesting that overexpression of degraded. The 5′ domain remains capped and lingers in the cyto- α genes inhibits the downstream expression of early and late genes. plasm for many hours. There are no data to explain the resistance to degradation of viral mRNA in cells infected with wild-type virus. innate immunity | host response The hypothesis we chose to pursue is that the VHS-RNase has still another partner responsible for the suppression of degradation of ttachment and entry of 1 (HSV-1) pro- viral mRNAs. In this report we show that VHS-RNase interacts with Avokes powerful innate immune responses to the virus. HSV-1 UL47, and that UL47 attenuates the degradation of viral and stable blocks the responses in two fundamentally different ways, by host mRNAs but has no effect on the stability of AREs mRNAs. bringing into cells at the time of entry both tegument and Relevant to this report are the following: UL47 is a 693-residue prepackaged mRNAs and by encoding proteins made immediately tegument protein known also as VP13/14 (24). UL47 is not after infection with numerous functions designed to block host essential for virus replication in cell culture (25). The protein is innate immune responses (1–5). Among the best-known and highly extensively posttranslationally modified by glycosylation and effective tegument proteins actively blocking host responses is the nucleotidylylations (26, 27). It has been reported that UL47 binds virion host shutoff RNase encoded by the UL41 ORF (6–8). This RNase is an endoribonuclease with the specificity of RNase A (9), Significance and it is active immediately after entry of the virus into cells and during the initial stages of infection. The primary function of VHS- Herpes simplex virus encodes an RNase linked to the shutoff of RNase appears to be the degradation of mRNAs made in response host gene expression early in infection (virion host shutoff-RNase), to infection, but it also prevents, by a mechanism not fully un- – and recent studies showed that its activity is highly selective: it derstood, the activation of protein kinase R (10 12). At late stages targets stable host mRNAs and α (immediate early) mRNAs but in the viral replicative cycle, virion host shutoff (VHS)-RNase spares other viral mRNAs. Here we report that RNase’sfunc- activity is blocked or neutralized by two late proteins, virion pro- tion early in infections is regulated by a desig- tein (VP) 16 and VP22, encoded by UL48 and UL49 ORFs, re- nated UL47. UL47 is packed in virions; it binds the RNase and spectively (13, 14). The RNase cleaves mRNA in polyribosomes attenuates the degradation of all viral mRNAs but has no ef- (15). Another partner of the VHS-RNase present in polyribosomes α fect on the processing of the stress response mRNAs targeted by is infected cell protein (ICP) 27, an (immediate early) multi- the viral RNase. functional regulatory protein (16) fi The VHS-RNase was initially thought to be nonspeci c with Author contributions: M.S., B.T., and B.R. designed research; M.S., B.T., and W.Z. per- respect to its target substrate, thus cleaving indiscriminately both formed research; M.S. and B.T. analyzed data; and B.R. wrote the paper. viral and cellular mRNAs. More recent studies have shown that The authors declare no conflict of interest. this is not the case. Thus, some mRNAs, notably those encoding 1To whom correspondence should be addressed. E-mail: [email protected]. tristetraprolin and GADD45β are spared, whereas stable host edu.

www.pnas.org/cgi/doi/10.1073/pnas.1305475110 PNAS Early Edition | 1of7 Downloaded by guest on September 28, 2021 RNA and shuttles between the cytoplasm and the nucleus (28). In injunction with ICP27, UL47 binds and displaces PABC1 from the poly(A) binding protein complex at the cap structure (29). Results

Ectopically Expressed UL47 Is Modified in the Presence of ICP27 and Is Pulled Down by VHS-RNase in the Absence of Other Viral Proteins. In this series of experiments, HEK 293T cells were transfected with plasmids encoding full-length UL47taggedwithMycatthe N-terminal (UL47-N-Myc) and/or full-length ICP27. The cells were harvested 48 h after transfection, and total cell lysates were in- cubated with GST alone or GST fused to full-length VHS (GST- VHS) as previously reported (30). The electrophoretically sepa- rated precipitates were reacted with antibody to the Myc tag. The results shown in Fig. 1 were as follows. The ectopically expressed pUL47 formed a broad band in lane 1 that resolved in two closely migrating bands in lane 9. The pUL47accumulatedinhigher amounts in cells cotransfected with the plasmid encoding ICP27 (lane 3). In addition, the doubly transfected cells accumulated a protein that reacted with the Myc antibody and migrated faster than the doublet seen in lane 1. GST-VHS pulled down the Myc- tagged pU 47 (lane 5). On the basis of electrophoretic mobility, L Δ the GST-VHS pulled down primarily if not exclusively the slower- Fig. 2. Immunoblot analysis of wild-type HSV-1(WT), UL47 deletion ( UL47), Δ migrating component of the doublet. However, GST-VHS pulled and repair (UL47-R) recombinant virus. Construction of UL47 and UL47-R recombinant virus was carried out as described in Materials and Methods. Vero down the doublet plus the faster-migrating band from the cells that Δ N cells were mock infected or infected with HSV-1(WT), UL47 mutant (A, lanes 1 are transfected with plasmids encoding both UL47- -Myc and and 3, respectively), or UL47-R (B, lane 3) at a multiplicity of infection of ICP27. We conclude that (a) in the presence of ectopically ex- 10 pfu/cell for 24 h. Cells were collected and equal amounts of total proteins pressed ICP27, several forms of pUL47 differ in electrophoretic were electrophoretically separated on a 10% denaturing polyacrylamide gel, mobility and that (b)VHSinteractswithUL47 in the absence of transferred to a nitrocellulose sheet, and probed with rabbit polyclonal anti- other viral proteins. The focus of the remainder of this report is the body to UL47 (A) and mouse polyclonal antibody to myc (B). role of UL47 on the degradation of viral and host mRNA. Δ Validation of the Deletion and Repair of U 47 in HSV-1 Strain F DNA the UL47 mutant, or the Myc-tagged repaired mutant (UL47-R) L were electrophoretically separated in denaturing gels and probed Background. The objective of these studies was to construct a mu- A B Δ with antibody to UL47 (Fig. 2 )ortotheMyctag(Fig.2 ). The tant lacking the UL47 coding sequence ( UL47) and to repair the a Δ results show that ( )UL47 was missing from lysates of cells infected UL47 with a Myc-tagged UL47 coding sequence. The procedures Δ A b for the construction of the mutants in a BAC system encoding with the UL47 mutant (Fig. 2 , lane 3), and ( ) a band reacting Materials and with the anti-Myc antibody was present in electrophoretically sep- HSV-1 strain F [HSV-1(F)] DNA are described in – B Methods. Lysates of Vero cells infected with wild-type parental virus, arated lysates of UL47-R infected cells (Fig. 2 ,lane3).

UL47 Attenuates the Degradation of Viral mRNA by Wild-Type Virus. The experimental design of the studies reported here follows that reported earlier (18). As illustrated schematically in Fig. 3A, replicate cultures of human epidermoid (HEp-2) cells were ex- posed to 10 pfu of either wild-type parental virus or ΔUL47 or UL47-R viruses per cell for 1 h before the inoculum was replaced with fresh medium. After an additional 2 h at 37 °C, the cells were exposed to medium containing Actinomycin D (ActD). The cells were harvested and total RNA extracted at the time of ActD ad- dition (time 0) and at 0.5, 1, 3, or 6 h thereafter. The levels of different viral transcripts belonging to the three kinetic classes were analyzed by real-time quantitative PCR (RT-qPCR). The results were normalized with respect to the 18S rRNA and then expressed as fold change compared with the RNA levels present in the cells at the time of ActD addition. The results shown in Fig. 3B were as follows. As expected on the basis of earlier studies (18), the rates of degradation of α mRNAs (e.g., ICP0, ICP22, and ICP27) were faster than those of β (e.g., ICP8) or γ (VP16 and UL41) mRNAs. Interestingly, in cells infected by ΔUL47 mutant virus, the viral mRNAs belonging to the three kinetic classes were Fig. 1. VHS pulls down UL47 from transfected cell lysates. HEK 293T cells were degraded faster than those present in wild-type virus-infected cells. transiently transfected with plasmids encoding full-length UL47 tagged with Moreover, the rates of degradation of viral mRNA from cells Myc at N terminus (UL47-N-Myc) and/or full-length ICP27. Soluble cell extracts infected with the repaired virus, U 47-R, could not be differentiated were prepared 48 h after transfection, and reacted with glutathione–agarose L beads bound to GST (lanes 4, 6, and 8) or GST–VHS (lanes 5, 7, and 9). The from that observed in cells infected with the wild-type virus. To validate the results obtained with RT-qPCR, the level of viral proteins bound to the beads were subjected to electrophoresis on a denaturing α β polyacrylamide gel, transferred to a nitrocellulose membrane, and reacted with RNAs representative of the three kinetic classes ( : ICP0, :ICP8, γ antibody against Myc. Immunoreactivity of UL47-N-Myc (lanes 1 and 3) from a and : VP16) were also analyzed by Northern blot hybridization 1/10th volume of whole-cell lysates is also shown. pcDNA, pcDNA 3.1(+). (Fig. 3C) and quantified by densitometric scanning (Fig. 3D). The

2of7 | www.pnas.org/cgi/doi/10.1073/pnas.1305475110 Shu et al. Downloaded by guest on September 28, 2021 PNAS PLUS MICROBIOLOGY

Fig. 3. Effects of UL47 on viral mRNAs degradation. (A) Experimental design diagram. HEp-2 cells were exposed to 10 pfu per cell of HSV-1(WT), ΔUL47 mutant, or UL47-R virus for 1 h. The inoculum was then replaced with fresh complete medium, and the cell cultures were incubated at 37 °C for two additional hours. was then stopped by addition of 10 μg of ActD per milliliter (time 0). Cells were harvested at the indicated time points. (B) Total RNAs were extracted, reverse-transcribed, and used for quantification by RT-qPCR of different viral transcripts representative of the three kinetic classes: α genes

mRNA: ICP0, ICP22, ICP27; β gene mRNA: ICP8; γ genes mRNA: VP16, UL41. The results were normalized with respect to the 18S rRNA and then expressed as fold change compared with the RNA levels present in the cells at the time of ActD addition. (C) Northern blot analysis was done as described in Materials and Methods. The ethidium bromide staining of 18S rRNA was used as loading control. The intensities of the bands at various time points were quantified by phosphorimager analysis, and the mRNAs remaining were normalized with respect to 18S rRNA and plotted relative to the time of addition of ActD (D). The mRNAs assayed were representative of the three kinetic classes: ICP0 (α), ICP8 (β), and VP16 (γ).

results clearly indicate that the data obtained by Northern blot UL47 Has No Apparent Effect on the Stability of AU-Rich mRNAs analyses mirrored those obtained by RT-qPCR (compare Fig. 3 B Accumulating After Infection. The design of this experiment is with D). Thus, (a) ICP0 mRNA was more rapidly degraded than shown in Fig. 5A.Briefly, replicate cultures of HEp-2 cells were VP16 or ICP8 mRNAs, (b) the rates of degradation of mRNAs in exposed to ActD 3 h after infection. The cells were harvested at c cells infected with wild-type virus or UL47-R were similar, and ( ) time of addition of ActD (time 0) and at time intervals shown. Δ RNA accumulating in cells infected with the UL47 mutant was Total RNA was subjected to Northern blot analyses. The auto- degraded more rapidly than those accumulating in cells infected by radiographic images of the intact (upper band) RNA shown in Fig. wild-type virus or repaired virus. The conclusion from this set of 5B were scanned with the aid of a General Dynamics phosphor- data is that UL47 attenuates the degradation of all classes of viral imager and normalized with respect to 18S ribosomal RNA and mRNAs accumulating in cells at 3 h after infection. RNA detected at time 0 (Fig. 5C). The results indicate that UL47 does not contribute to the stability of the AU-rich RNAs accu- UL47 Attenuates the Degradation of Stable Host mRNA After Infection. The experimental design of these studies is shown in Fig. 3A. mulating in infected cells. In brief, HEp-2 cells were exposed to 10 pfu of wild-type HSV-1, Validation of the Mutant Virus Deleted in both U 41 and U 47 (ΔU 41/ ΔU 47, or U 47-R viruses per cell for 1 h. The inoculum was L L L L L ΔU 47). To assess the contribution of U 47 to the rates of accu- replaced with fresh medium and the cells incubated for additional L L mulation of viral mRNAs, it was of interest to construct a mutant 2 h before ActD was added to the cultures. The cells were har- lacking both the VHS-RNase and the UL47 gene. The construction vested at the time of infection (mock), at the time of addition of Materials and Methods ActD (time 0), and at 0.5, 1, or 3 h thereafter. Total RNA was of the double-deletion virus is detailed in .In extracted and either analyzed by RT-qPCR for β-actin mRNA the experiments described here, Vero cells were mock infected Δ Δ (Fig. 4A) or subjected to Northern blot hybridization for GAPDH or exposed to 10 pfu of wild-type HSV-1 or the UL41/ UL47 mRNA accumulation (Fig. 4B), and the data obtained by densi- mutant virus per cell. The cells were harvested 24 h after infection, tometric scanning were plotted as fraction of mRNA remaining and equal amounts of total proteins were electrophoretically sep- compared with mock-infected cells (Fig. 4C). As expected, host arated on a 10% denaturing polyacrylamide gel, transferred to mRNA is rapidly degraded during the 3 h after infection. The key a nitrocellulose sheet, and probed with mouse monoclonal anti- finding of Fig. 4 is that UL47 attenuated the degradation of both body to US11 or rabbit polyclonal antibodies to UL47 and UL4. As stable host mRNAs assayed in this study. shown in Fig. 6, lane 3, the double-mutant virus did not express

Shu et al. PNAS Early Edition | 3of7 Downloaded by guest on September 28, 2021 Fig. 4. Effects of UL47 on stable host mRNAs degradation. (A) Total RNAs were extracted, and β-actin mRNA was analyzed by RT-qPCR. The results were normalized with respect to the 18S rRNA and then expressed as fraction of total mRNA relative to mock-infected cells. (B) GAPDH mRNA was analyzed by Northern blot. The ethidium bromide staining of 18S rRNA was used as loading control, and the value obtained by densitometric scans was normalized with respect to 18S rRNA and plotted relative to the RNA levels present in the mock infected cells (C).

either UL47 (Fig. 6A)orUL41 (Fig. 6B). Both viruses expressed accumulation was noted in the case of VP16 and VP22 (both γ B v vi US11—alate(γ2) protein (Fig. 6C). genes) mRNAs (Fig. 7 , and ). Although both VP16 and VP22 mRNAs increased in amounts, they lagged behind the mRNA Δ Δ Accumulation of Viral mRNAs in Cells Infected by UL41/ UL47 encoded in cells infected with the wild-type virus or the ΔUL41 Mutant Virus. The experimental design of these studies is shown mutant. We conclude the following: as predicted from the results A fl schematically in Fig. 7 . Brie y, replicate HEp-2 cell cultures shown in Fig. 3, deletion of both UL41 and UL47 increased the rate were exposed to 10 pfu of wild-type HSV-1, ΔUL41, or ΔUL41/ of accumulation of α mRNAs. Concurrently, as could be pre- α β γ ΔUL47 mutant viruses per cell. After 2 h, the inoculum was dicted, the transition from to and mRNA synthesis was replaced with fresh medium. The cultures were harvested at the delayed. Though within the time frame of these studies the ac- times shown and purified total RNA was analyzed by RT-qPCR. cumulation of ICP8 and TK mRNAs ultimately reached the levels The results were as follows. As shown in Fig. 7B, in the absence of of mRNAs accumulating in wild-type virus-infected cells, the γ both UL41 and UL47, both ICP0 and ICP22 mRNAs, products of α levels of the mRNAs encoded by VP16 and VP22 genes con- genes, accumulated at a faster rate (Fig. 7B, i and ii). In cells tinued to lag behind those of wild-type virus. infected with the double-deletion mutant, the accumulation of The differences in the accumulation of mRNAs encoded by Δ fi ICP8 (an early β) and TK (a late β) mRNAs appeared to be bi- wild-type and UL41 viruses did not differ signi cantly or differed Δ phasic: there was an initial delay followed by a faster rate of ac- less extensively than those between the wild-type and UL41/ Δ cumulation (Fig. 7B, iii and iv). A more extensive delay in mRNA UL47 mutant viruses. These results mirror the results reported previously (31), suggesting that the rate of transition from α to β to γ genes expression in wild-type virus-infected cells is largely en- abled by UL47. Discussion The salient features of the results presented here are as follows: (i)UL47, a tegument protein made late in infection, is modified in transfected cells in the presence of ICP27 and interacts with UL41 in the absence of other viral proteins. (ii)UL47 attenuates the degradation of all kinetic classes of viral mRNAs. (iii)UL47 attenuates the degradation of stable host mRNA in infected cells but has no apparent effect on the stability of the AREs-mRNAs induced in infected cells. (iv) The transition from α to β and γ genes’ mRNA synthesis is significantly delayed in the absence of UL47. The significance of these findings is as follows: VP16 was the first viral protein shown to interact with VHS and, in the context of infected cells, shown to abrogate the RNase ac- tivity (13). The demonstration that neutralization of VHS activity required VP22 emerged from the observation that deletion of the gene encoding VP22 resulted in the selection of mutants defective in RNase activity (32). The evidence that VHS-RNase interacts with ICP27 emerged from studies designed to determine the role of this α protein in the degradation of viral mRNAs (33). We se- lected UL47 to study its role as a potential partner of VHS-RNase on the basis of the evidence that VHS-RNase degrades mRNAs in ’ Fig. 5. Effects of UL47 on AU-rich mRNAs degradation. (A) Experimental design polyribosomes and the report that UL47 binds the poly(A) binding Δ diagram. HEp-2 cells were exposed to 10 pfu per cell of HSV-1(F) or UL47 protein (29). The striking feature of the results obtained so far is mutant virus for 1 h. The inoculum was then replaced with fresh complete that three of the four known viral partners of VHS-RNase are medium, and the cell cultures were incubated at 37 °C for two additional hours. tegument proteins, and that the function of all three is to attenuate Transcription was then stopped by addition of 10 μg of ActD per milliliter (time 0). Cells were harvested at the indicated time points. (B)IEX-1mRNAwasana- or neutralize the RNase activity of VHS. One hypothesis that lyzed by Northern blot. The ethidium bromide staining of 18s rRNA was used as could explain the wealth of proteins that regulate the VHS-RNase loading control. (C) The value obtained by densitometric scans was normalized activity is that they function at different times during the replicative with respect to 18S rRNA and plotted relative to the time of addition of ActD. cycle or in different environments. For example, the RNase-

4of7 | www.pnas.org/cgi/doi/10.1073/pnas.1305475110 Shu et al. Downloaded by guest on September 28, 2021 PNAS PLUS Assuming that the UL47 interacts randomly with the VHS-RNase in infected cells, as suggested by the observation that it attenuates the degradation of all classes of mRNAs, it is not surprising that it also attenuates the degradation of host mRNAs after infection. Two observations, however, are worthy of note. First, the pattern of degradation of stable cellular mRNAs appears to be similar to that of viral α mRNAs rather than that of β or γ mRNAs. We have no evidence to support the hypothesis that host and viral α mRNAs share features that differentiate them from β or γ mRNAs. A more likely scenario is that an α protein, possibly ICP27, plays a role in enhancing the stability of viral mRNAs made later in infection. ICP27 has been reported to enhance the translation of γ mRNAs (38), but the effect may be due to stabilization of the mRNAs rather than to a specific effect on the translational machinery of the infected cell. It is noteworthy, as shown in this report, that ICP27 mediates a posttranslational modification as well as enhancement of accumulation of UL47 in cotransfected cells (Fig. 1). The second noteworthy observation is that UL47 has no effect on the pattern of accumulation of AREs-mRNAs. One hypoth- esis that could explain this observation is based on the evidence that (a) VHS-RNase binds tristetraprolin, which in turn binds to the AREs, and (b) VHS-RNase performs two functions: it deadenylates the mRNA and cleaves the mRNA at or near AREs. It is conceivable that UL47 does not affect or participate in the Fig. 6. Immunoblot analysis of HSV-1(WT) and UL41/UL47 double-deletion interactions of VHS-RNase with tristetraprolin and the AREs. Δ Δ Δ Δ ( UL41/ UL47) recombinant virus. Construction of UL41/ UL47 recombinant The sum total of the available data on the function of VHS- virus was processed as described in Materials and Methods. Vero cells were RNase may be viewed from the prospective of the function of mock infected or infected with HSV-1(WT) or ΔU 41/ΔU 47 mutant viruses at L L VHS-RNase in the viral replicative cycle. Specifically, the data MICROBIOLOGY a multiplicity of infection of 10 pfu per cell for 24 h. Cells were collected, and suggest that VHS-RNase is a non–target-specific endoribonuclease equal amounts of total proteins were electrophoretically separated on a 10% that is tightly regulated to degrade only specific targets. The reg- denaturing polyacrylamide gel, transferred to a nitrocellulose sheet, and fi probed with the mouse monoclonal antibodies to US11 (C)orβ-actin (D) and ulators identi ed to date are VP16, VP22, UL47, and ICP27. The rabbit polyclonal antibodies to UL47 (A)orUL41 (B). net effects of the carefully orchestrated regulation are to degrade two groups of mRNAs and spare a third. One group, degraded schematically as illustrated in Fig. 8A, consists of stable host and neutralizing activities of VP16 and VP22 could be confined to viral α mRNAs. The objective, as discussed above, is perhaps to a late stage in the replicative cycle rather than to the proteins packaged and brought into the newly infected cells. This conclu- sion is consistent with the evidence that VHS is active following infection, even though all three proteins are brought into the newly infected cell. In contrast, the VHS-RNase–attenuating activities of UL47 and those associated with ICP27 could be expressed in polyribosomes assembled after the entry of the virus into cells. A preceding report from this laboratory demonstrated that with the possible exception of α mRNAs, viral RNAs are relatively resistant to degradation attributed to VHS-RNase (18). The results presented here indicate that UL47 attenuates the degradation of all kinetic classes of viral mRNAs and hence it most likely acts on VHS-RNase rather than on specific features of viral mRNAs. In the absence of both VHS-RNase and UL47 gene, α mRNAs ac- cumulate at a faster rate, whereas the accumulation of β and γ mRNAs is delayed. Additional evidence supporting the hypothesis that α mRNA accumulations are tightly controlled is the obser- vation that ICP4 inhibits its own transcription by binding to its transcription initiation site (34, 35). The hypothesis that α mRNAs outcompete viral mRNAs with respect to efficiency of synthesis or translation is tempered by the observation that HSV also controls the accumulation and function of key α proteins. Thus, (a)atearly time after infection ICP0, a key regulatory protein turns over with a half-life of less than 1 h. Between 6 and 9 h after infection Fig. 7. Effects of ΔU 41/ΔU 47 on viral gene transcription. (A)Experimental b L L the nucleus is totally depleted of ICP0 (36). ( ) The product of design diagram. HEp-2 cells were exposed to 10 pfu per cell of HSV-1(WT), ΔUL41, Δ Δ ORF O, a protein made late in infection, blocks the function of or UL41/ UL47 mutant virus for 2 h. The inoculum was then replaced with fresh ICP4 (37). It is conceivable that the relatively rapid turnover of α complete medium. Cells were harvested at the indicated time points. (B)Total β γ RNAs were extracted, reverse-transcribed, and used for quantification by RT-qPCR mRNAs compared with or mRNAs is a component of the α α of different viral transcripts representative of the three kinetic classes: genes design to tightly control the synthesis and accumulation of pro- mRNA: ICP0, ICP22; β genes mRNA: ICP8, TK; γ genes mRNA: VP16, VP22. The teins and that, by default, overexpression of α proteins is detri- results were normalized with respect to the 18S rRNA and then expressed as fold mental to the orderly transition from α to β and γ protein synthesis. change relative to the RNA levels present in the cells at 2 h after infection.

Shu et al. PNAS Early Edition | 5of7 Downloaded by guest on September 28, 2021 Fig. 8. A model of the role of UL47 on VHS-dependent mRNA degradation. Current models of actively translated mRNAs displaythattheyaremaintainedin a closed-loop conformation by the interaction of the translation initiation complex eIF4F bound to the 5′ cap with the poly(A)-binding protein (PABP) at the 3′ end of RNA. (A) For stable host and α viral mRNAs, VHS-RNase bound to eIF4H, a component of the eIF4F complex, decaps the mRNA, thus initiating the degradation of

mRNA by the 5′ exonuclease Xrn1. However, this process could be attenuated by UL47 bound to VHS. (B) For stress-response mRNAs, VHS-RNase binds preferentially to tristetraprolin (TTP) and cleaves the ARE-mRNA 5′ to the AREs. The product of cleavage inside the 3′ UTR is equivalent to a deadenylated mRNA accessible to the

multimeric complex of 3′ to 5′ exonucleases known as the exosome. (C)Forβ and γ viral mRNAs, studies in this report suggest that UL47 bound to VHS, which also can interact with ICP27, probably interferes with the decapping of the mRNA and consequently the degradation of mRNA by the 5′ exonuclease Xrn1.

control the synthesis and accumulation of α gene products and to 95 °C for 5 min, resolved by PAGE, transferred to nitrocellulose mem- enable efficient transition from α to β and γ genes’ expression. The brane, and immunoblotted with the mouse anti-Myc antibody. second group targeted by the VHS-RNase in a manner schemati- cally depicted in Fig. 8B contains stress-response mRNAs that Immunoblot Analysis. Cells were collected 24 h after infection. The procedures fi reflect, in large part, host innate responses. The group that is for harvesting, solubilization, protein quanti cation, SDS/PAGE, and transfer to nitrocellulose membranes were performed as previously reported (41). The spared, as illustrated schematically in Fig. 8C,consistsofviral membrane was probed for UL47, UL41, Myc tag, and β-actin with the anti- mRNAs that encode proteins required for viral DNA synthesis and bodies listed above. structural proteins. It would make little sense to degrade the mRNAs directing the synthesis of these proteins. Construction of Recombinant Viruses. The construction procedures were de- scribed previously (42). Briefly, The U 41 ORF is contained in two different HSV- Materials and Methods L 1 BamHI fragments, BamHI(0), and BamHI(I), and the UL47 ORF is contained in Cells and Viruses. HEp-2, Vero, and HEK 293T cell lines were obtained from the HSV-1 BamHI(F) fragment (43). The pRB129 containing BamHI(O) was digested American Type Culture Collection. All cell lines were grown in Dulbecco’s with HindIII and BamHI. The pRB130 containing BamHI(I) fragment was modified Eagle medium supplemented with 5% (vol/vol) FCS (HEp-2 cells), digested with SacI and BamHI. The 1.5-kb HindIII–BamHI fragment from 5% (vol/vol) newborn calf serum (Vero cells), or 10% (vol/vol) FBS (HEK 293T BamHI(O), along with the 2.0-kb SacI–BamHI fragment, was cloned in pUC19 cells). The BAC encoding the HSV-1(F) DNA was reported elsewhere (39, 40). to generate pRB9001. The pRB9001 was subsequently digested with HindIII and EcoRI. The 3.6-kb HindIII–EcoRI fragment from pRB9001 was subcloned in

Antibodies. The rabbit polyclonal antibody against UL41 has been described pBlueScript KS(+) to generate pRB9002. The pRB9002 was then digested with elsewhere (30). Rabbit polyclonal antibody against UL47 was kindly provided BmgBI and self-ligated to eliminate 822 bp within the UL41 ORF; this gener- by D. M. Meredith (Leeds University, Leeds, United Kingdom). Mouse mono- ated pRB9003, which contains ΔUL41 ORF plus flanking sequences. The clonal antibody against actin and Myc tag were purchased from Sigma and pRB9003 was then digested with XhoI and XbaI, and the 2.8-kb fragment

Santa Cruz Biotechnology, respectively. All of the antibodies were used in containing ΔUL41 ORF plus flanking sequences was cloned into a pKO5 plas- a dilution of 1:1,000. mid (44) to generate pKO9003. The 1.2-kb upstream flanking sequence of

UL47 was amplified from pRB128 by PCR using the primers UL47-up-F (5′-CGG ′ Plasmids. The full length of UL47 tagged with Myc at the 5′ end and ICP27 GAT CCG GTG GCG ATA GAC GCG GGT TAT CGG ATG -3 )andUL47-up-R ORF was subcloned into pcDNA 3.1(+) transfer vector from previously (5′-TGC TCT AGA AGG TAC TAC CGA GAG ACC GCT CGT CTG -3′), which con-

generated constructs (30). The resulting plasmids were named pUL47 and tain a BamHI and a XbaI restriction size, respectively. The 1.2-kb fragment was pICP27, respectively. cloned in pBlueScript KS(+) to generate pRB9004. The 1.3-kb downstream

flanking sequence of UL47 was amplified from pRB128 by PCR using the pri- HEK 293T Cell Transient Transfection. The procedures for transfection HEK mers UL47-down-F (5′-CGG AAT TCA AGC TCC TCC CGA TAA AAA GCG CCC-3′) 293T cells were described previously (14). The cells were harvested 48 h after and UL47-down-R (5′-CCG CTC GAG CCC GAA CCA AGC CTT GAT GCT CAA transfection and lysed according to the protocol for the subsequent GST C-3′), which contain an EcoRI and a XboI restriction size, respectively. The 1.3-kb pull-down assay. fragment was subcloned into pRB9004 to generate pRB9005, which contains up- and downstream flanking sequences of UL47. The pRB9005 was then GST Pull-Down Assay. The procedures of pull-down were performed as pre- digested with XhoI and XbaI, and the 2.5-kb fragment containing UL47 viously described (14). Briefly, transfected HEK 293T cells were lysed in GST flanking sequences was cloned into pKO5 plasmid to generate pKO9005. The ′ fi lysis buffer [20 mM Tris (pH 8.0)/1 mM EDTA/1% Nonidet P-40/200 mM NaCl/ UL47 coding sequence tagged with Myc in the 3 end was ampli ed from 0.1 mM sodium orthovanadate/10 mM NaF/2 mM DTT/protease inhibitor pRB128 by PCR using the primers UL47-myc-F (5′-CCG GAA TTC TTA CAG ATC mixture (Complete Protease; Roche Diagnostics] on ice for 1 h. Insoluble TTC TTC AGA AAT AAG TTT TTG TTC TGG GCG TGG CGG GCC TCC CAG CCC-3′) material was pelleted by centrifugation at maximum speed in a 5415C and UL47-myc-R (5′-CGC GGA TCC ATG TCG GCT CGC GAA CCC GCG GGG-3′), centrifuge (Eppendorf) for 10 min at 4 °C. The supernatant fluids precleared which contain a EcoRI and Myc tag sequence and a BamHI restriction size, with 50 μL of a 50% slurry of glutathione beads for 3 h at 4 °C were in- respectively. The fragment of Myc-tagged UL47 coding sequence was cloned cubated overnight at 4 °C with equal amounts of a 50% slurry of glutathione into pRB9005 to generate pRB9006. The pRB9006 was then digested with XhoI

beads bound to GST alone or GST fused to the VHS protein. The beads were and XbaI, and the 4.5-kb fragment containing Myc-tagged UL47 ORF plus pelleted by centrifugation and rinsed five times with GST buffer. The pro- flanking sequences was cloned into a pKO5 plasmid to generate pKO9006. teins bound to the beads were solubilized in 50 μL of SDS gel-loading buffer Escherichia coli RR1 stain harboring HSV-1 BAC was electroporated with [2% SDS/5% 2-mercaptoethanol/50 mM Tris (pH 6.8)/2.75% sucrose), heated pKO9003, pKO9005, and pKO9006, respectively, and incubated at 43 °C on LB

6of7 | www.pnas.org/cgi/doi/10.1073/pnas.1305475110 Shu et al. Downloaded by guest on September 28, 2021 plates containing 25 μg/mL Zeocin and 20 μg/mL chloramphenicol. The colo- Northern Blot Analyses. The Northern blot analysis was done as described pre- PNAS PLUS nies were diluted and plated on LB plates containing chloramphenicol (20 μg/mL) viously (14), with minor modifications. Briefly, 8 μg of RNA were loaded onto and 5% sucrose. Colonies grown on the sucrose plates were screened with denaturing formaldehyde gel and probed with random hexanucleotide-primed PCR or by colony hybridization. The recombinant BAC9003, BAC9005, and 32P-labeled fragment of indicated viral (ICP0, ICP8, and VP16) or cellular (GAPDH, BAC9006 obtained contain partial deletion of U 41 (ΔU 41), full deletion of L L IEX-1) genes upon transfer onto a nylon membrane. Prehybridization and hy- U 47 (ΔU 47), and Myc-tagged U 47 repair (U 47-R), respectively. Competent L L L L bridization were performed with the ULTRAhyb buffer (Ambion) supplemented cells of RR1 harboring BAC9003 (ΔUL41) were electroporated with pKO9005, with 200 μg of denatured salmon sperm DNA per milliliter (Stratagene). The the selection process was repeated, and BAC9007 (ΔUL41/ΔUL47) was membranes were prehybridized for 2 h at 42 °C and then overnight after the obtained, in which both the UL41 and UL47 were deleted. Plasmids isolated from E. coli were transfected into Vero cells. The resulting viruses addition of the 32P-labeled probe. The membranes were rinsed as suggested were plaque purified three times and confirmed by immunoblot analysis, as by the manufacturer of the ULTRAhyb buffer and exposed to film for signal detailed in Results. detection.

Total RNA Extraction and RT-qPCR Analysis. The total RNA extraction and RT- ACKNOWLEDGMENTS. Support for this work was provided by National qPCR analysis procedures were described previously (18). Cancer Institute Grant CA115662.

1. He B, Gross M, Roizman B (1997) The gamma(1)34.5 protein of herpes simplex virus 1 tristetraprolin, two cellular proteins that bind and destabilize AU-rich RNAs. J Virol 78 complexes with protein phosphatase 1alpha to dephosphorylate the alpha subunit of the (16):8582–8592. eukaryotic translation initiation factor 2 and preclude the shutoff of protein synthesis by 24. Spear PG, Roizman B (1972) Proteins specified by herpes simplex virus. V. Purification double-stranded RNA-activated protein kinase. Proc Natl Acad Sci USA 94(3):843–848. and structural proteins of the herpesvirion. J Virol 9(1):143–159. 2. Esclatine A, Taddeo B, Roizman B (2004) The UL41 protein of herpes simplex virus 25. Barker DE, Roizman B (1990) Identification of three genes nonessential for growth in mediates selective stabilization or degradation of cellular mRNAs. Proc Natl Acad Sci cell culture near the right terminus of the unique sequences of long component of USA 101(52):18165–18170. herpes simplex virus 1. 177(2):684–691. 3. van Lint AL, et al. (2010) Herpes simplex virus immediate-early ICP0 protein inhibits 26. Blaho JA, Mitchell C, Roizman B (1994) An amino acid sequence shared by the herpes Toll-like receptor 2-dependent inflammatory responses and NF-kappaB signaling. J simplex virus 1 alpha regulatory proteins 0, 4, 22, and 27 predicts the nucleotidylylation Virol 84(20):10802–10811. of the UL21, UL31, UL47, and UL49 gene products. JBiolChem269(26):17401–17410. 4. Sciortino MT, Suzuki M, Taddeo B, Roizman B (2001) RNAs extracted from herpes 27. Meredith DM, Lindsay JA, Halliburton IW, Whittaker GR (1991) Post-translational simplex virus 1 virions: Apparent selectivity of viral but not cellular RNAs packaged in modification of the tegument proteins (VP13 and VP14) of herpes simplex virus type 1 virions. J Virol 75(17):8105–8116. by glycosylation and phosphorylation. J Gen Virol 72(Pt 11):2771–2775. 5. Sciortino MT, Taddeo B, Poon AP, Mastino A, Roizman B (2002) Of the three 28. Donnelly M, Elliott G (2001) Nuclear localization and shuttling of herpes simplex virus tegument proteins that package mRNA in herpes simplex virions, one (VP22) tegument protein VP13/14. J Virol 75(6):2566–2574.

transports the mRNA to uninfected cells for expression prior to viral infection. Proc 29. Dobrikova E, Shveygert M, Walters R, Gromeier M (2010) Herpes simplex virus MICROBIOLOGY Natl Acad Sci USA 99(12):8318–8323. proteins ICP27 and UL47 associate with polyadenylate-binding protein and control its 6. Kwong AD, Kruper JA, Frenkel N (1988) Herpes simplex virus virion host shutoff subcellular distribution. J Virol 84(1):270–279. function. J Virol 62(3):912–921. 30. Taddeo B, Zhang W, Roizman B (2006) The U(L)41 protein of herpes simplex virus 1 7. Kwong AD, Frenkel N (1989) The herpes simplex virus virion host shutoff function. J degrades RNA by endonucleolytic cleavage in absence of other cellular or viral Virol 63(11):4834–4839. proteins. Proc Natl Acad Sci USA 103(8):2827–2832. 8. Barzilai A, Zivony-Elbom I, Sarid R, Noah E, Frenkel N (2006) The herpes simplex virus 31. Zhang Y, McKnight JL (1993) Herpes simplex virus type 1 UL46 and UL47 deletion type 1 vhs-UL41 gene secures viral replication by temporarily evading apoptotic mutants lack VP11 and VP12 or VP13 and VP14, respectively, and exhibit altered viral cellular response to infection: VHS-UL41 activity might require interactions with thymidine kinase expression. J Virol 67(3):1482–1492. elements of cellular mRNA degradation machinery. J Virol 80(1):505–513. 32. Sciortino MT, et al. (2007) Replication-competent herpes simplex virus 1 isolates 9. Taddeo B, Roizman B (2006) The virion host shutoff protein (UL41) of herpes simplex selected from cells transfected with a bacterial artificial chromosome DNA lacking virus 1 is an endoribonuclease with a substrate specificity similar to that of RNase A. J only the UL49 gene vary with respect to the defect in the UL41 gene encoding host Virol 80(18):9341–9345. shutoff RNase. J Virol 81(20):10924–10932. 10. Leib DA, Machalek MA, Williams BR, Silverman RH, Virgin HW (2000) Specific 33. Taddeo B, Zhang W, Roizman B (2010) Role of herpes simplex virus ICP27 in the phenotypic restoration of an attenuated virus by knockout of a host resistance gene. degradation of mRNA by virion host shutoff RNase. J Virol 84(19):10182–10190. Proc Natl Acad Sci USA 97(11):6097–6101. 34. Leopardi R, Michael N, Roizman B (1995) Repression of the herpes simplex virus 1 11. Leib DA (2002) Counteraction of interferon-induced antiviral responses by herpes alpha 4 gene by its gene product (ICP4) within the context of the viral is simplex viruses. Curr Top Microbiol Immunol 269:171–185. conditioned by the distance and stereoaxial alignment of the ICP4 DNA binding site 12. Sciortino MT, et al. (2013) The virion host shut-off RNase plays a key role in blocking relative to the TATA box. J Virol 69(5):3042–3048. the activation of protein kinase R in cells infected with herpes simplex virus 1. J Virol 35. Michael N, Roizman B (1993) Repression of the herpes simplex virus 1 alpha 4 gene by 87(6):3271–3276. its gene product occurs within the context of the viral genome and is associated with 13. Lam Q, et al. (1996) Herpes simplex virus VP16 rescues viral mRNA from destruction by all three identified cognate sites. Proc Natl Acad Sci USA 90(6):2286–2290. the virion host shutoff function. EMBO J 15(10):2575–2581. 36. Gu H, Poon AP, Roizman B (2009) During its nuclear phase the multifunctional 14. Taddeo B, Sciortino MT, Zhang W, Roizman B (2007) Interaction of herpes simplex regulatory protein ICP0 undergoes proteolytic cleavage characteristic of polyproteins. virus RNase with VP16 and VP22 is required for the accumulation of the protein but Proc Natl Acad Sci USA 106(45):19132–19137. not for accumulation of mRNA. Proc Natl Acad Sci USA 104(29):12163–12168. 37. Randall G, Lagunoff M, Roizman B (1997) The product of ORF O located within the 15. Taddeo B, Zhang W, Roizman B (2009) The virion-packaged endoribonuclease of herpes domain of herpes simplex virus 1 genome transcribed during latent infection binds to simplex virus 1 cleaves mRNA in polyribosomes. Proc Natl Acad Sci USA 106(29):12139–12144. and inhibits in vitro binding of infected cell protein 4 to its cognate DNA site. Proc 16. Sandri-Goldin RM (2008) The many roles of the regulatory protein ICP27 during Natl Acad Sci USA 94(19):10379–10384. herpes simplex virus infection. Front Biosci 13:5241–5256. 38. Fontaine-Rodriguez EC, Knipe DM (2008) Herpes simplex virus ICP27 increases 17. Esclatine A, Taddeo B, Evans L, Roizman B (2004) The herpes simplex virus 1 UL41 translation of a subset of viral late mRNAs. J Virol 82(7):3538–3545. gene-dependent destabilization of cellular RNAs is selective and may be sequence- 39. Ye GJ, Roizman B (2000) The essential protein encoded by the UL31 gene of herpes specific. Proc Natl Acad Sci USA 101(10):3603–3608. simplex virus 1 depends for its stability on the presence of UL34 protein. Proc Natl 18. Taddeo B, Zhang W, Roizman B (2013) The herpes simplex virus host shutoff RNase Acad Sci USA 97(20):11002–11007. degrades cellular and viral mRNAs made before infection but not viral mRNA made 40. Horsburgh BC, Hubinette MM, Tufaro F (1999) Genetic manipulation of herpes after infection. J Virol 87(8):4516–4522. simplex virus using bacterial artificial chromosomes. Methods Enzymol 306:337–352. 19. Feng P, Everly DN, Jr., Read GS (2005) mRNA decay during herpes simplex virus (HSV) 41. Taddeo B, Esclatine A, Zhang W, Roizman B (2003) The stress-inducible immediate- infections: Protein-protein interactions involving the HSV virion host shutoff protein early responsive gene IEX-1 is activated in cells infected with herpes simplex virus 1, and translation factors eIF4H and eIF4A. J Virol 79(15):9651–9664. but several viral mechanisms, including 3′ degradation of its RNA, preclude expression 20. Sarma N, Agarwal D, Shiflett LA, Read GS (2008) Small interfering RNAs that deplete of the gene. J Virol 77(11):6178–6187. the cellular translation factor eIF4H impede mRNA degradation by the virion host 42. Gu H, Roizman B (2007) Herpes simplex virus-infected cell protein 0 blocks the shutoff protein of herpes simplex virus. J Virol 82(13):6600–6609. silencing of viral DNA by dissociating histone deacetylases from the CoREST-REST 21. Karr BM, Read GS (1999) The virion host shutoff function of herpes simplex virus complex. Proc Natl Acad Sci USA 104(43):17134–17139. degrades the 5′ end of a target mRNA before the 3′ end. Virology 264(1):195–204. 43. Post LE, Conley AJ, Mocarski ES, Roizman B (1980) Cloning of reiterated and 22. Perez-Parada J, Saffran HA, Smiley JR (2004) RNA degradation induced by the herpes nonreiterated herpes simplex virus 1 sequences as BamHI fragments. Proc Natl Acad simplex virus VHS protein proceeds 5′ to 3′ in vitro. J Virol 78(23):13391–13394. Sci USA 77(7):4201–4205. 23. Esclatine A, Taddeo B, Roizman B (2004) Herpes simplex virus 1 induces cytoplasmic 44. Lopez P, Van Sant C, Roizman B (2001) Requirements for the nuclear-cytoplasmic accumulation of TIA-1/TIAR and both synthesis and cytoplasmic accumulation of translocation of infected-cell protein 0 of herpes simplex virus 1. JVirol75(8):3832–3840.

Shu et al. PNAS Early Edition | 7of7 Downloaded by guest on September 28, 2021