242, 128±137 (1998) ARTICLE NO. VY979006

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector The Immediate-Early ICP27 Shuttles between Nucleus and Cytoplasm

Wendy E. Mears and Stephen A. Rice1

Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2H7 Received July 15, 1997; returned to author for revision September 3, 1997; accepted December 19, 1997

ICP27 is an essential type 1 (HSV-1) nuclear protein which regulates viral early and late genes during infection. The exact mechanism by which ICP27 modulates viral is unknown, but considerable evidence suggests that it functions posttranscriptionally. In this study, we have asked whether ICP27, like some other viral and cellular posttranscriptional regulatory , shuttles between the nuclear and cytoplasmic compartments of the cell. Using an interspecies heterokaryon assay, we demonstrate that ICP27, but not the HSV-1 nuclear proteins ICP4 or ICP8, is an efficient shuttling protein. ICP27's shuttling ability does not depend on viral infection or other HSV-1 proteins, as it shuttles even when transiently expressed in uninfected cells. To understand the importance of shuttling for ICP27's regulatory functions, we examined several mutant forms of ICP27 to see whether they exhibited altered shuttling. We identified three ICP27 mutations which partially disrupt shuttling, as well as one mutation, M15, which completely abrogates this activity. The M15 mutation alters residues 465 and 466 near the carboxyl terminus of ICP27 and was previously shown to inactivate ICP27's ability to induce certain viral late mRNAs. These results suggest that ICP27's nuclear shuttling activity is involved in its viral late gene activation function. © 1998 Academic Press

INTRODUCTION We have been investigating the role of the IE protein ICP27 in HSV-1 gene regulation. ICP27, also known as Herpes simplex virus type 1 (HSV-1), an extensively IE63, is a nuclear protein consisting of 512 amino acid studied human alphaherpesvirus, serves as a very useful residues. It plays an essential role in viral gene expres- model for studying how herpesviruses use the RNA poly- sion, being required for the progression of the viral in- merase II machinery and other cellular fection into the L phase (Sacks et al., 1985; McCarthy et components for their gene expression. During lytic infec- al., 1989; Rice and Knipe, 1990). Based on the phenotypic tion, the Ͼ75 HSV-1 genes are expressed in a temporally regulated cascade consisting of three major waves of analysis of viral mutants, it can be concluded that ICP27 viral gene expression (Honess and Roizman, 1974; re- carries out two distinct functions that contribute to L viewed in Roizman and Sears, 1996). These are referred phase activation (Rice and Knipe, 1990; Rice et al., 1993). First, ICP27 increases the level of viral DNA replication to as the immediate-early (IE or ␣), delayed-early (DE or ␤), and late (L or ␥) phases. The IE genes are expressed up to 10-fold (Rice and Knipe, 1990), most likely by trans- immediately upon infection and encode four gene regu- activating a subset of DE genes involved in DNA repli- latory factors (ICP0, ICP4, ICP22, and ICP27) and an cation (Uprichard and Knipe, 1996; McGregor et al.,1996). immunological modulator (ICP47). DE genes are ex- Second, ICP27 stimulates L gene mRNA expression by a pressed next; these mostly encode proteins involved in mechanism that does not involve the stimulation of viral viral DNA replication. Finally, L genes are expressed. DNA replication (Rice and Knipe, 1990; Rice and Lam, These predominantly encode virion structural proteins, 1994; Hibbard and Sandri-Goldin, 1995). In addition to its and their expression is stimulated by the process of viral stimulatory effects on viral genes, ICP27 also down- DNA synthesis. At the same time that the virus uses the regulates the expression of some viral IE and DE genes cellular machinery to express its own genes, it efficiently during infection (Sacks et al., 1985; McCarthy et al., 1989; represses host gene expression, at both transcriptional Rice and Knipe, 1990). Furthermore, ICP27 contributes to and posttranscriptional levels (reviewed in Smiley et al., the shutoff of host gene expression, as both cellular 1991; Spencer et al., 1997). protein synthesis (Sacks et al., 1985) and mRNA levels (Hardwicke and Sandri-Goldin, 1994) are elevated in ICP27 mutant infections compared to wild-type (WT) HSV-1 infection. 1 To whom reprint requests should be addressed at Department of Biochemistry, University of Alberta, 474 Medical Sciences Building, The exact mechanism(s) by which ICP27 regulates Edmonton, Alberta, Canada, T6G 2H7. Fax: (403) 492-0886. E-mail: viral and cellular genes during HSV-1 infection is un- [email protected]. known. However, considerable evidence indicates that it

0042-6822/98 $25.00 Copyright © 1998 by Academic Press 128 All rights of reproduction in any form reserved. NUCLEAR SHUTTLING OF HSV ICP27 129 functions, at least in part, at the posttranscriptional level. Under various experimental circumstances, ICP27 can modulate the efficiency of pre-mRNA polyadenylation (McLauchlan et al., 1992; Sandri-Goldin and Mendoza, 1992; McGregor et al., 1996), alter the intranuclear dis- tribution of pre-mRNA splicing factors (Phelan et al., 1993; Sandri-Goldin et al., 1995), inhibit pre-mRNA splic- ing (Hardy and Sandri-Goldin, 1994), increase the stabil- ity of specific cellular mRNAs (Brown et al., 1995), and inhibit the nuclear export of intron-containing viral tran- scripts (Phelan et al., 1996). Consistent with these post- transcriptional effects, recent studies have shown that ICP27 can bind directly to RNA (Brown et al., 1995; Ingram et al., 1996; Mears and Rice, 1996), although the nature of ICP27's putative in vivo RNA target(s) is not yet known. In the past few years, several nuclear-localized RNA- binding proteins involved in posttranscriptional gene reg- ulation, including HIV-1 Rev (Kalland et al., 1994; Meyer and Malim, 1994) and several cellular hnRNP proteins (PinÄol-Roma and Dreyfuss, 1992; Flach et al., 1994), have been shown to be nuclear shuttling proteins. That is, although the steady-state localization of these proteins is predominantly nuclear, they can be demonstrated to move continuously between the nuclear and the cyto- plasmic compartments of the cell. Given ICP27's nuclear localization, RNA-binding activity, and posttranscriptional FIG. 1. Schematic diagram showing the interspecies heterokaryon assay used to test the shuttling of HSV-1 nuclear proteins. The appear- effects on gene expression, we asked in this study ance of a in the mouse nucleus of an interspecies hete- whether ICP27 is a nuclear shuttling protein. Our results reokaryon suggests that the protein has shuttled from the infected demonstrate that ICP27 is indeed such a protein, and we monkey nucleus. provide evidence that this activity is important for ICP27's gene regulation function. mouse cells by residual virus. One hour later, the local- ization of viral nuclear proteins in the interspecies het- RESULTS erokaryons is assessed by indirect immunofluores- cence. Since protein synthesis is inhibited, the appear- Nuclear shuttling of ICP27 in interspecies ance of an HSV-1 nuclear protein in a mouse nucleus of heterokaryons an interspecies heterokaryon suggests that the viral pro- ICP27's nuclear localization, RNA-binding activity, and tein has shuttled out of the infected monkey nucleus and posttranscriptional effects on gene expression suggest been re-imported into the uninfected mouse nucleus. the possibility that it shuttles between the nucleus and In the first set of experiments, the interspecies hetero- the cytoplasm of cells, similar to what has been ob- karyon assay was used to examine the ability of ICP27 to served for some other posttranscriptional regulatory pro- shuttle. After cell fusion, fixation, and staining, cellular teins including several hnRNP proteins and HIV-1 Rev heterokaryons were identified by phase microscopy (Fig. (PinÄol-Roma and Dreyfuss, 1992; Flach et al., 1994; Kal- 2A) and microscopic visualization of stress fibers using land et al., 1994; Meyer and Malim, 1994). To see if ICP27 FITC±phallocidin, which stains polymerized actin (not is a shuttling protein, we utilized an interspecies hetero- shown). To differentiate monkey nuclei from mouse nu- karyon assay that has been developed to identify and clei, cells were also stained with Hoechst 33258 dye. study shuttling proteins (Borer et al., 1989; PinÄol-Roma Monkey nuclei stain relatively diffusely with this reagent, and Dreyfuss, 1993). We modified the assay slightly, as whereas mouse nuclei exhibit a characteristic speckled indicated in Fig. 1, to apply it to HSV-1 nuclear proteins. pattern (Fig. 2B; the mouse nucleus is indicated by an In our modified assay, Vero monkey cells are infected arrow). Examination of numerous interspecies hetero- with HSV-1 for 4 h and then treated with cycloheximide to karyons indicated that nearly all which contained ICP27 prevent further protein synthesis. The infected Vero cells in their monkey nuclei also displayed ICP27 in their are then fused, using polyethylene glycol, to uninfected mouse nuclei (Fig. 2C). As expected, ICP27 was never 3T6 mouse cells in the presence of cycloheximide and observed in unfused mouse cells, indicating that the human immune globulin to prevent infection of the experimental protocol had successfully prevented pro- FIG. 2. Trafficking of ICP27, ICP4, and ICP8 in interspecies heterokaryons. A±C, D±F, and G±I show fields of cells stained for ICP27, ICP4, and ICP8, respectively. Each series shows a single field of cells containing one or more interspecies heterokaryons. The arrows denote the mouse nuclei in the interspecies heterokaryons. (A), (D), and (G): Phase-contrast images. (B), (E), and (H): Hoechst staining. Note that 3T6 mouse nuclei show an intense speckled pattern, while Vero monkey nuclei exhibit more diffuse staining. (C) ICP27 staining, using mouse monoclonal antibody H1119. ICP27 staining in the nucleus indicated was scored as ``strong.'' (F) ICP4 staining, using mouse monoclonal antibody H1114. The nuclei indicated were negative for ICP4 staining. (I) ICP8 staining, using mouse monoclonal antibody H1115. ICP8 staining in the nucleus indicated was scored as ``weak.'' Bar, 20 ␮m. 130 NUCLEAR SHUTTLING OF HSV ICP27 131

TABLE 1 excluded from the analysis presented in Table 1, as it Trafficking of HSV-1 Nuclear Proteins seemed possible that the ICP27 in the mouse nuclei within Interspecies Heterokaryons could have originated from the Vero cell cytoplasm, rather than from Vero cell nuclei. The results for ICP4 and Staining of mouse nuclei in ICP8, which were predominantly nuclear in their local- interspecies heterokaryons ization, were quite different than those for ICP27. These (%) proteins were detected in mouse nuclei in only a minority No. of Positive of interspecies heteryokaryons (ϳ15 and ϳ32% for ICP4 heterokaryons and ICP8, respectively), and in these cases, staining was a b Viral protein examined ϩϩ ϩ Negative nearly invariably weak. Two representative interspecies heterokaryons stained for ICP4 are shown in Figs. 2D± ICP27 114 94.7 5.3 0 ICP4 114 0 14.9 85.1 2F. Note that no ICP4 is detectable in these mouse ICP8 115 0.9 31.3 67.8 nuclei. A representative heterokaryon stained for ICP8 is shown in Figs. 2G±2I. In this case, weak ICP8 staining a (ϩϩ) Strong staining within mouse nuclei (visually comparable in can be detected in the mouse nucleus of the hetero- intensity to that observed in Vero nuclei of the same heterokaryon). b karyon. These results demonstrate that, unlike ICP27, (ϩ) Weak staining within mouse nuclei (visually less than half the staining intensity observed in Vero nuclei of same heterokaryon). ICP4 and ICP8 are not efficient nuclear shuttling pro- teins. ductive infection of the mouse cells. These results indi- Analysis of viral ICP27 mutants cate that ICP27 is a nuclear shuttling protein. If ICP27's ability to shuttle between nucleus and cyto- plasm is biologically important, then this activity may be Analysis of other HSV-1 nuclear proteins disrupted by mutations which affect ICP27's regulatory To see if other HSV-1 nuclear proteins behave similarly functions. To investigate this possibility, we examined to ICP27 in the interspecies heterokaryon assay, we the shuttling ability of several mutant polypeptides en- carried out additional experiments in which we exam- coded by replication-defective HSV-1 ICP27 mutants. The ined the viral nuclear proteins ICP4 and ICP8, in addition mutants used were d1-2, which encodes an ICP27 to ICP27. The IE protein ICP4 is an essential transcrip- polypeptide having a deletion of an acidic region near tional transactivator with sequence-specific DNA-bind- the amino (N)-terminus (Rice et al., 1993), and n504R, ing activity, whereas the DE protein ICP8 is an essential M11, M15, and M16, which encode molecules having DNA replication protein with nonspecific single-stranded alterations in the carboxyl (C)-terminal half of ICP27. The DNA-binding activity (Roizman and Sears, 1996). Intranu- n504R mutant has a nonsense mutation near the C- clear trafficking was assessed semiquantitatively by ex- terminus of the ICP27 coding region and expresses only amining numerous (Ͼ100) interspecies heterokaryons 504 of ICP27's 512 residues (Rice and Knipe, 1990). The for each of the proteins under investigation. The degree M11, M15, and M16 mutations result in amino acid sub- of viral protein staining within the mouse nucleus of each stitutions at residues 340/341, 465/466, and 488, respec- interspecies heterokaryon was determined visually and tively (Rice and Lam, 1994). In preliminary experiments, scored as either strong or weak, based on its staining we found that several of the mutant molecules were relative to the infected monkey nucleus (or nuclei) in the expressed poorly under the conditions of prolonged cy- same heterokaryon. Although the division of viral protein cloheximide treatment (data not shown), perhaps as a staining into these two categories was somewhat sub- result of reduced protein stability relative to the WT jective, we found this distinction helpful, because in protein. To increase the levels of ICP27 for analysis, we some cases staining of mouse nuclei was quite weak modified the conditions of infection by using a ``cyclohex- (see below). imide release'' protocol to enhance the expression of The results of this analysis are shown in Table 1. For viral IE proteins (Honess and Roizman, 1974). To do this, ICP27, the results were consistent with those of the Vero cells were infected in the presence of cyclohexi- previous experiment. In this case, 100% (114/114) of the mide for4htoallow IE transcripts to accumulate to high interspecies heterokaryons which were scored had levels. The cells were then washed, and protein synthe- mouse nuclei which stained positively for ICP27, in most sis was allowed to proceed for 2 h. Uninfected mouse cases (ϳ95%) strongly. It should be noted that a number cells and cycloheximide were then added and hetero- of HSV-1-infected monkey cells, both fused and unfused, karyons were formed as previously described. This showed a low level of cytoplasmic ICP27 staining, in change to the protocol resulted in comparable levels of addition to strong nuclear staining. Interspecies hetero- WT and mutant molecules immediately prior to the cell karyons that contained significant cytoplasmic ICP27 fusion step. This is shown for the WT, M15, and n504R invariably showed positive mouse nuclei. These were polypeptides in Figs. 3A, 3E, and 3I. 132 MEARS AND RICE

FIG. 3. Nuclear trafficking of mutant ICP27 polypeptides. Vero cells were infected with either WT or mutant HSV-1 using the cycloheximide release protocol described in the text and fused to uninfected 3T6 cells. (A±D) WT-infected cells; (E±H) M15-infected cells; (I±L) n504R-infected cells. The photomicrographs show either cells 30 min prior to fusion (A, E, and I) or single fields of cells after fusion (B±D, F±H, and J±L). Cells were stained with anti-ICP27 monoclonal antibody H1119 (A, E, I and D, H, L) or Hoechst stain (C, G, and K). B, F, and J show phase-contrast images. The arrows indicate 3T6 mouse nuclei within interspecies heterokarons. ICP27 staining in the nuclei indicated in D and L was scored as ``strong,'' while ICP27 staining in the nuclei indicated in H was negative. Bar, 20 ␮m.

Using the cycloheximide reversal protocol, the shut- staining was strong. An example of WT ICP27's ability to tling abilities of the five mutant forms of ICP27 were shuttle between nuclei in interspecies heterokaryons is compared to that of the WT protein. For each virus, shown in Figs. 3B±3D. The mutant ICP27 polypeptides numerous (Ͼ200) interspecies heterokaryons were ex- fall into three categories with respect to their shuttling amined microscopically, and the degree of ICP27 stain- abilities. First, the n504R-encoded protein is as capable ing within mouse nuclei of interspecies heterokaryons as the WT polypeptide of shuttling, with greater than 93% was scored as either strong or weak using the criteria of interspecies heterokaryons showing strong staining of described earlier. The results of this analysis are shown mouse nuclei (Table 2). An example of an interspecies in Table 2. As expected from the previous experiments, heterokaryon which shows shuttling of the n504R WT ICP27 efficiently shuttled between nuclei, showing polypeptide is shown in Figs. 3J±3L. Second, the staining in the great majority (ϳ94±98%) of mouse nuclei polypeptides encoded by the d1-2, M11, and M16 mu- of interspecies heterokaryons. In most cases (Ͼ83%), tants all show modest deficiencies in shuttling. Although NUCLEAR SHUTTLING OF HSV ICP27 133

TABLE 2 nor why ICP27 should show more cytoplasmic localiza- Trafficking of Mutant ICP27 Molecules tion than ICP4 or ICP8. It is possible that these obser- within Interspecies Heterokaryons vations are related to the fact that ICP27 is a nuclear shuttling protein. ICP27 staining within mouse nuclei of interspecies Nuclear shuttling of ICP27 in the absence of viral heterokaryons (%) infection No. of Positive To determine if viral infection or other viral proteins are heterokaryons required for ICP27's shuttling, we examined the ability of Virus examined ϩϩa ϩb Negative ICP27 to shuttle when expressed by transient expression Expt. 1 in uninfected cells. Vero cells were transfected with WT 220 85.0 8.6 6.4 pC27, a plasmid which contains the ICP27 open reading M15 254 0.4 0.8 98.8 frame under the control of a human cytomegalovirus- d1-2 231 19.6 60.9 19.6 derived promoter. Transfected Vero cells were fused to Expt. 2 WT 254 83.5 14.6 1.9 3T6 mouse cells in the presence of cycloheximide, and n504R 247 93.5 5.3 1.2 the location of ICP27 was assessed in numerous inter- M11 281 33.1 58.7 8.2 species heterokaryons which exhibited ICP27 within M15 287 0 11.5 88.5 their Vero nuclei. ICP27 was found to be present within M16 249 32.1 62.7 5.2 the mouse nuclei of the great majority of these hetero- a (ϩϩ) Strong ICP27 staining within 3T6 nuclei (visually comparable karyons. A representative interspecies heterokaryon is in intensity to that observed in Vero nuclei of same heterokaryon). shown in Fig. 4. We conclude that the ability of ICP27 to b (ϩ) Weak ICP27 staining within 3T6 nuclei (visually less than half shuttle between nucleus and cytoplasm does not de- the staining intensity observed in Vero nuclei of same heterokaryon). pend on viral infection or the expression of other HSV-1 factors. Rather, nuclear shuttling appears to be an inher- ent property of the ICP27 polypeptide. these proteins were present in Ͼ80% of mouse nuclei within positive interspecies heterokaryons, the intensity DISCUSSION of ICP27 staining was significantly less than that for the WT protein, with the majority of interspecies heterokary- Recently, it has become appreciated that some nu- ons displaying weak staining (Table 2). Third, the mutant clear proteins are not confined to the nucleus once they protein encoded by M15 appears to be completely un- are transported there. Rather, these proteins continu- able to shuttle. In two experiments, ϳ99 and ϳ89% of ously shuttle between the nuclear and the cytoplasmic mouse nuclei in interspecies heterokaryons were nega- compartments. Such nuclear shuttling proteins include tive for M15 ICP27 staining. Moreover, those few mouse some of the heterogenous nuclear RNP (hnRNP) proteins nuclei which were positive stained only weakly. A repre- (PinÄol-Roma and Dreyfuss, 1992; Flach et al., 1994), sev- sentative interspecies heterokaryon, exhibiting no de- eral nucleolar proteins (Borer et al., 1989; Meier and tectable shuttling of the M15 polypeptide, is shown in Blobel, 1992), the human immundeficiency virus type 1 Figs. 3F±3H. The results of this analysis show that some (HIV-1) Rev protein (Meyer and Malim, 1994; Kalland et lethal mutations in the ICP27 polypeptide can either al., 1994), U1 small nuclear RNP-specific protein U1A partially (d1-2, M11, and M16) or completely (M15) com- (Kambach and Mattaj, 1992), steroid hormone receptors promise its ability to shuttle between the nucleus and (Giuochon-Mantel et al., 1991), heat shock-related pro- cytoplasm. teins (Mandell and Feldherr, 1990), and the protein ki- In the experiment shown in Fig. 3, very little if any nase A inhibitor peptide (PKI) (Fantozzi et al., 1994). ICP27 was seen in the infected monkey cytoplasm prior Although the physiological significance of nuclear shut- to the fusion step, for either the WT or mutant proteins tling is not known in all cases, some shuttling proteins (Figs. 3A, 3E, and 3I). However, after cell fusion, both the which bind RNA appear to have roles in mRNA export. WT- and the n504R-encoded proteins efficiently localized Thus, the shuttling protein hnRNP A1 is associated with to mouse nuclei. This clearly indicates that the bulk of RNA polymerase II-derived transcripts in both the nu- the protein imported into mouse nuclei must have origi- cleus and cytoplasm and may directly mediate the nu- nated from the infected monkey nuclei, rather than from cleocytoplasmic transport of these molecules (PinÄol- the cytoplasm, demonstrating bona fide shuttling. As Roma and Dreyfuss, 1992). HIV-1 Rev is involved in the mentioned earlier, in some experiments, significant cy- nuclear export of viral intron-containing mRNAs, medi- toplasmic staining was seen for WT ICP27 and some ated via binding of Rev to the Rev-response element mutant ICP27 molecules, but not for ICP4 and ICP8. It is within these mRNAs (Cullen, 1994; Fischer et al., 1994). not known why the degree of ICP27 cytoplasmic staining HSV-1 ICP27, an essential regulatory factor, has re- should vary depending on the experimental conditions, cently been shown to possess RNA-binding activity 134 MEARS AND RICE

more, ICP27 can shuttle even when transiently ex- pressed in uninfected cells, indicating that its shuttling does not require viral infection or any other viral compo- nents. Our finding that low levels of ICP4 and ICP8 are im- ported into mouse nuclei in some interspecies hetero- karyons raises the possibility that HSV-1 nuclear pro- teins ICP4 and ICP8 also are shuttling proteins, albeit weak ones. However, an alternate and more trivial inter- pretation of the data is that there is a low level of ICP8 and ICP4 in the cytoplasm of infected monkey cells prior to heterokaryon formation and that this cytoplasmic pro- tein is imported into mouse nuclei after cell fusion, re- sulting in some weakly positive nuclei. At present, we cannot distinguish between these two possibilities. The capacity of a nuclear protein to shuttle demon- strates that its nuclear import is counterbalanced to some extent by its nuclear export. Although much has been elucidated with respect to nuclear protein import, the process of nuclear protein export is less well under- stood (for recent reviews, see GoÈrlich and Mattaj, 1996; Moore, 1996; Nigg, 1997). Two pathways for nuclear protein export have been proposed (Michael et al., 1995). First, it has been suggested that nuclear export can be a passive process which occurs for nuclear proteins which do not have a means of nuclear retention (Laskey and Dingwall, 1993; Schmidt-Zachmann et al., 1993). Other nuclear proteins possess specific nuclear export se- quences (NESs) which mediate their active export, pre- sumably via interactions with cellular transport machin- ery. Such sequences have been identified in Rev (Fischer et al., 1995; Wen et al., 1995; Meyer et al., 1996), PKI (Wen et al., 1995), hnRNP A1 (Michael et al., 1995), RANBP1 (Richards et al., 1996), and HTLV-1 Rex (Kim et al., 1996; Palmieri and Malim, 1996). The Rev, Rex, RANBP1, and PKI NESs consist of short leucine-rich segments, whereas the hnRNP A1 NES, termed M9, is a 38-residue sequence which contains no leucine residues and is FIG. 4. Shuttling of ICP27 in transfected cells. Vero cells were transiently transfected with a plasmid expressing ICP27, and the trans- glycine-rich. In general, NES-containing proteins exhibit fected cells were subjected to the interspecies heterokaryon assay as rapid nuclear shuttling (within 1 h), whereas the shuttling described in the text. A field of cells containing a representative of proteins lacking NESs is a slower process. interspecies heterokaryon is shown. (A) Phase-contrast image; (B) Although we have previously characterized ICP27's Hoechst staining; (C) ICP27 staining using monoclonal antibody H1119. nuclear localization signals (Mears et al., 1995), the sig- Arrow indicates 3T6 nucleus in heterokaryon. Bar, 20 m. ␮ nals and mechanisms involved in ICP27's nuclear export are unknown. ICP27 exhibits rapid shuttling, in that a (Brown et al., 1995; Ingram et al., 1996; Mears and Rice, substantial fraction of ICP27 can traffic between nuclei in 1996). In addition, it is known to have a number of the 1-h time frame of the interspecies heterokaryon as- posttranscriptional effects on viral and cellular gene ex- say. Based on this, we speculate that ICP27's nuclear pression, including alterations to the nucleocytoplasmic export is mediated by an NES. ICP27 does not contain transport of viral mRNAs (Phelan et al., 1996). Given the any sequence which is obviously related to the M9 NES, similarities between ICP27 and some of the known nu- but it does possess several short sequences which bear clear shuttling proteins, we asked in this work whether some resemblance to leucine-rich NESs. It is also pos- ICP27 shuttles between the nucleus and cytoplasm. The sible that ICP27 does not possess an NES that directly results of our experiments demonstrate that ICP27 is a interacts with the cellular export machinery, but rather very active shuttling protein, as it can efficiently traffic shuttles because it associates with a protein or macro- between nuclei in interspecies heterokaryons. Further- molecular complex which does possess such a signal. NUCLEAR SHUTTLING OF HSV ICP27 135

Given ICP27's ability to bind to RNA, it is interesting to shuttling. These results suggest that the C-terminal por- speculate that ICP27 might bind to pre-mRNA in the tion of ICP27 is involved in both the shuttling activity and nucleus and be transported to the cytoplasm in direct an additional function required for late gene expression. association with these molecules. How might the shuttling of ICP27 aid in the transactiva- Our analysis of mutant ICP27 polypeptides provides tion of viral L genes? One attractive idea is that ICP27 information about the determinants in ICP27 which affect stimulates the nuclear export of L gene transcripts by its nuclear export. We find that the ICP27 molecules binding to them in the nucleus and escorting them to the encoded by d1-2, M11, and M16 are significantly reduced cytoplasm. Experiments are currently in progress in our in their shuttling ability. It is interesting that the d1-2 laboratory to test this model. mutation results in the deletion of residues 12±63, dis- rupting a potential leucine-rich NES near the N-terminus MATERIALS AND METHODS of ICP27 (residues 7±20: MLIDLGLDLSDSDL). The most deleterious mutation, however, is the C-terminal muta- Cells, , and infections tion M15, which completely abrogates shuttling. The M15 Vero (African green monkey kidney) cells and 3T6 mutation alters residues 465 and 466 from PG to LE and (mouse fibroblast) cells were obtained from the Ameri- is in a region of the protein which is very closely con- can Type Culture Collection (Rockville, MD) and were served among herpesviral ICP27 homologues (Brown et propagated in Dulbecco's modified Eagle's medium con- al., 1995). It is possible that the M15 mutation disrupts a taining 5% heat-inactivated fetal bovine serum (GIBCO). critical NES or affects ICP27's ability to interact with an HSV-1 infections were carried out at a m.o.i. of 10 PFU NES-containing protein or complex. It is perhaps note- per cell, and infected cells were maintained in 199V worthy that there is a NES-resembling leucine-rich re- media [199 media (GIBCO), 2% heat-inactivated newborn gion very near to the M15 alterations (residues 469±476: calf serum (GIBCO)]. The HSV-1 strain KOS1.1 (Hughes MAGLIEIL). Another possibility is that the M15 mutation and Munyon, 1975) was the WT strain used in these decreases ICP27's nuclear export indirectly by increas- experiments. The HSV-1 ICP27 mutants d1-2, M11, M15, ing its nuclear retention, perhaps by trapping ICP27 in a M16, and n504R have been described previously (Rice nonproductive interaction in the nucleus. A precedent for and Knipe, 1990; Rice et al., 1993; Rice and Lam, 1994). this notion comes from studies of the hnRNP C proteins, which possess nuclear retention signals that can over- Plasmid construction ride the effects of NESs (Nakielny and Dreyfuss, 1996). To examine the ability of ICP27 to shuttle in the ab- While this paper was being revised, a study by Soli- sence of other viral proteins, ICP27 was expressed in man et al. was published which independently con- Vero cells by transient transfection. The plasmid used for cluded that ICP27 can shuttle between the nucleus and transfection was pC27, which contains the open reading cytoplasm of HSV-1-infected cells (Soliman et al., 1997). frame of the ICP27 gene under the control of a minimal These investigators demonstrated that ICP27 accumu- CMV promoter. To construct pC27, the plasmid pBS27 lates in the cytoplasm when cells are treated with acti- (Rice and Lam, 1994) was linearized by cleavage with nomycin D. This effect is characteristic of several shut- DrdI, which cuts just upstream of ICP27's initiation tling proteins and appears to reflect a linkage between codon. The DNA ends were made blunt by treatment with their nuclear import and ongoing RNA polymerase II with T4 DNA polymerase, and EcoRI linkers were ligated transcription (PinÄol-Roma and Dreyfuss, 1991, 1992). In on. After treatment with EcoRI, the 2.0-kb ICP27 gene- contrast to our findings, however, Soliman et al. reported containing fragment was cloned into the vector that ICP27 was unable to shuttle when expressed alone pUHD10-3 (Gossen and Bujard, 1992), previously linear- in transfected cells, but could be activated to shuttle by ized with EcoRI and treated with calf alkaline phos- the coexpression of an HSV-1 L gene transcript. The phatase. After cloning in Escherichia coli, the orientation reason for this apparent discrepancy is unknown, but of the ICP27 gene insert in pC27 was verified by restric- may reflect the two different methods used to assay tion enzyme digestion. shuttling. The key question raised by the discovery of ICP27's Interspecies heterokaryon assay for nuclear shuttling shuttling activity concerns its biological function. The finding that the lethal M15 mutation inactivates ICP27's To study the ability of HSV-1 nuclear proteins to shuttle shuttling argues that this activity is essential. The M15 between nucleus and cytoplasm, a modified version of virus is WT for viral DNA replication, but is unable to an interspecies heterokaryon assay (Borer et al., 1989; induce certain L gene mRNAs (Rice and Lam, 1994). This PinÄol-Roma and Dreyfuss, 1992) was used. Vero cells suggests that ICP27's shuttling plays an important role in grown on circular coverslips in 12-well trays (1.5 ϫ 105 the activation of HSV-1 L genes. It is interesting that the cells/well) were infected with HSV-1. Four hours postin- n504R mutation also destroys ICP27's ability to stimulate fection (p.i.), 3T6 cells were plated (3.5 ϫ 105 cells/well) L genes (Rice and Knipe, 1990) but does not affect its onto the infected Vero cells in 199V medium containing 136 MEARS AND RICE

50 ␮g/ml cycloheximide and 0.15% human immune glob- costained with 0.5 ␮g/ml Hoechst dye No. 33258 (Sigma) ulin (Armour Pharmaceutical Company, Kankakee, IL). diluted in PBS and a 1:100 dilution in PBS of Bodipy-FL Four hours later, the cycloheximide concentration was phallocidin (Molecular Probes Inc., Eugene, OR). The increased to 100 ␮g/ml. After 30 min, the cells were cells were visualized with a Zeiss Axioskop 20 fluores- washed in phosphate-buffered saline (PBS). Cell fusion cence microscope equipped with a PlanNeofluar 63ϫ was performed by the addition of 50% polyethylene gly- objective lens. Photographic negatives were converted col (w/w) in Hanks' balanced media (GIBCO) to each to electronic files using an ArcusII desktop scanner coverslip for 2 min. After fusion, the coverslips were (AGFA) and a Power Macintosh 8500/150 computer, and washed in PBS four times. All reagents and solutions composite figures were made using the Photoshop pro- involved in the fusion process were prewarmed to 37°C. gram (version 3.0, Adobe). After washing, the coverslips were returned to medium containing 100 ␮g/ml cycloheximide for 60 min, after ACKNOWLEDGMENTS which time the cells were fixed for immunofluorescence. This research was supported by the National Cancer Institute of In some experiments, the level of ICP27 in infected Canada with funds from the Canadian Cancer Society. S.A.R. is a cells was increased by using a ``cycloheximide release'' Senior Scholar of the Alberta Heritage Foundation for Medical Re- protocol (Honess and Roizman, 1974). This was done by search (AHFMR). W.E.M. was supported by a studentship award from infecting Vero cells in the presence of 50 ␮g/ml cyclo- the AHFMR. We gratefully acknowledge the expert editorial and scien- tific advice of Leslie Schiff. heximide to allow for the enhanced accumulation of IE mRNAs. Four hours p.i., the cycloheximide-containing REFERENCES medium was removed, the cells were washed, and new media without cycloheximide were added. This allowed Borer, R. A., Lehner, C. F., Eppenberger, H. M., and Nigg, E. A. (1989). the accumulated IE mRNAs to be translated. After 2 h, Major nucleolar proteins shuttle between nucleus and cytoplasm. 3T6 cells were added in medium containing cyclohexi- Cell 56, 379±390. Brown, C. R., Nakamura, M. S., Mosca, J. D., Hayward, G. S., Strauss, mide, and the interspecies heterokaryon assay was car- S. E., and Perera, L. P. (1995). Herpes simplex virus trans-regulatory ried out as described above. protein ICP27 stabilizes and binds to 3Ј ends of labile mRNA. J. Virol. To examine the potential of transiently expressed 69, 7187±7195. ICP27 to shuttle, the interspecies heterokaryon assay Cullen, B. R. (1994). RNA-sequence-mediated gene regulation in HIV-1. was applied to transfected cells using a protocol similar Infect. Agents Dis. 3, 68±76. 2 Fantozzi, D. A., Harootunian, A. T., Wen, W., Taylor, S. S., Feramisco, J. R., to that of Michael et al. (1995). Vero cells grown in 25-cm Tsien, R. Y., and Meinkoth, J. L. (1994). Thermostable inhibitor of flasks were transfected with 8 ␮g of pC27 using the cAMP-dependent protein kinase enhances the rate of export of the calcium phosphate precipitation procedure as described kinase catalytic subunit from the nucleus. J. Biol. Chem. 269, 2676± previously (Rice and Knipe, 1990). Twenty-four hours 2686. Fischer, U., Huber, J., Boelens, W. C., Mattaj, I. W., and LuÈhrmann, R. later, the transfected Vero cells were transfered to cov- (1995). The HIV-1 Rev activation domain is a nuclear export signal 5 erslips in 12-well trays (2 ϫ 10 cells/well). After approx- that accesses an export pathway used by specific cellular RNAs. Cell imately 12 h, 3T6 mouse cells were plated (3 ϫ 105/well) 82, 475±483. in the presence of 50 ␮g/ml cycloheximide onto the Fischer, U., Meyer, S., Teufel, M., Heckel, C., LuÈhrmann, R., and Raut- transfected Vero cells. After 4 h, the cycloheximide con- mann, G. (1994). Evidence that HIV-1 Rev directly promotes the nuclear export of unspliced RNA. EMBO J. 13, 4105±4112. centration in the media was increased to 100 ␮g/ml, and Flach, J., Bossie, M., Vogel, J., Corbett, A., Jinks, T., Aker Willens, D., and the interspecies heterokaryon assay was performed as Silver, P. A. (1994). A yeast RNA-binding protein shuttles between the described above. nucleus and cytoplasm. Mol. Cell. Biol. 14, 8399±8407. GoÈrlich, D., and Mattaj, I. W. (1996). Nucleocytopasmic transport. Sci- ence 271, 1513±1518. Immunofluorescence Gossen, M., and Bujard, H. (1992). Tight control of gene expression in Cells were fixed with 3.7% formaldehyde, permeabil- mammalian cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89, 5547±5551. ized with acetone, and processed for immunofluores- Guiochon-Mantel, A., Lescop, P., Christin-maitre, S., Loosfelt, H., Perrot- cence as described by Quinlan et al. (1984). The primary Applanat, M., and Milgrom, E. (1991). Nucleocytoplasmic shuttling of antibodies used were the mouse monoclonals H1119, the progesterone receptor. EMBO J. 10, 3851±3859. H1115, and H1114, which are specific for ICP27, ICP8, Hardwicke, M. A., and Sandri-Goldin, R. M. (1994). The herpes simplex virus regulatory protein contributes to the decrease in cellular mRNA and ICP4, respectively. All three were obtained from the levels during infection. J. Virol. 63, 4797±4810. Goodwin Institute for Cancer Research (Plantation, FL). Hardy, R. W., and Sandri-Goldin, R. M. (1994). Herpes simplex virus The antibodies were diluted 1:600, 1:1000, and 1:800, inhibits host cell splicing and the regulatory protein ICP27 is required respectively, in PBS. The secondary antibody used for for this effect. J. Virol. 68, 7790±7799. indirect immunofluorescence was tetramethylrhodamine Hibbard, M., and Sandri-Goldin, R. M. (1995). Arginine-rich regions succeeding the nuclear localization region of herpes simplex virus isothiocyanate-conjugated anti-mouse immunoglobulin type 1 regulator protein ICP27 are required for efficient nuclear G (IgG) (Jackson ImmunoResearch Laboratories Inc., localization and late gene expression. J. Virol. 69, 4656±4667. Mississauga, Ontario), diluted 1:200 in PBS. Cells were Honess, R. W., and Roizman, B. (1974). Regulation of herpesvirus mac- NUCLEAR SHUTTLING OF HSV ICP27 137

romolecular synthesis. I. Cascade regulation of the synthesis of Phelan, A., Carmo-Fonseca, M., McLauchlan, J., Lamond, A. I., and three groups of viral proteins. J. Virol. 14, 8±19. Clements, J. B. (1993). A herpes simplex virus type 1 immediate-early Hughes, R. G., and Munyon, W. H. (1975). Temperature-sensitive mu- gene product, IE63, regulates small nuclear ribonucleoprotein distri- tants of herpes simplex virus type 1 defective in lysis but not in bution. Proc. Natl. Acad. Sci. USA 90, 9056±9060. transformation. J. Virol. 16, 275±283. Phelan, A., Dunlop, J., and Clements, J. B. (1996). Herpes simplex virus Ingram, A., Phelan, A., Dunlop, J., and Clements, J. B. (1996). Immediate type 1 protein IE63 affects the nuclear export of virus intron-contain- IE63 of herpes simplex virus type 1 binds RNA directly. ing transcripts. J. Virol. 70, 5255±5265. J. Gen. Virol. 77, 1847±1851. PinÄol-Roma, S., and Dreyfuss, G. (1991). Transcription-dependent and Kalland, K. H., Szilvay, A. M., Brokstad, K. A., Saetrevik, W., and Hauke- transcription-independent nuclear transport of hnRNP proteins. Sci- nes, G. (1994). The human immunodeficiency virus type 1 Rev protein ence 253, 312±314. shuttles between cytoplasm and nuclear compartments. Mol. Cell. PinÄol-Roma, S., and Dreyfuss, G. (1992). Shuttling of pre-mRNA binding Biol. 14, 7436±7444. proteins between nucleus and cytoplasm. Nature 355, 730±732. Kambach, C., and Mattaj, I. W. (1992). Intracellular distribution of the PinÄol-Roma, S., and Dreyfuss, G. (1993). hnRNP proteins: Localization U1A protein depends on active transport and nuclear binding to U1 and transport between the nucleus and cytoplasm. Trends Cell Biol. snRNA. J. Cell Biol. 118, 11±21. 3, 151±155. Kim, F. J., Beeche, A. A., Hunter, J. J., Chin, D. J., and Hope, T. J. (1996). Quinlan, M. P., Chen, L. B., and Knipe, D. M. (1984). The intra-nuclear Characterization of the nuclear export signal of human T-cell lym- location of a herpes simplex virus DNA binding protein is determined photropic virus type 1 Rex reveals that nuclear export is mediated by by the status of viral DNA replication. Cell 36, 857±868. position-variable hydrophobic interactions. Mol. Cell. Biol. 16, 5147± Rice, S. A., and Knipe, D. M. (1990). Genetic evidence for two distinct 5155. transactivation functions of the herpes simplex virus a protein ICP27. Laskey, R. A., and Dingwall, C. (1993). Nuclear shuttling: The default J. Virol. 64, 1704±1715. pathway for nuclear proteins. Cell 74, 585±586. Rice, S. A., and Lam, V. (1994). Amino acid substitution mutations in the Mandell, R. B., and Feldherr, C. M. (1990). Identification of two HSP70- herpes simplex virus ICP27 protein define an essential gene regu- related Xenopus oocyte proteins that are capable of recycling across lation function. J. Virol. 68, 823±833. the nuclear envelope. J. Cell Biol. 111, 1775±1783. Rice, S. A., Lam, V., and Knipe, D. M. (1993). The acidic amino-terminal McCarthy, A. M., McMahan, L., and Schaffer, P. A. (1989). Herpes region of herpes simplex virus type 1 alpha protein ICP27 is required simplex virus type 1 ICP27 deletion mutants exhibit altered patterns for an essential lytic function. J. Virol. 67, 1778±1787. of transcription and are DNA deficient. J. Virol. 63, 18±27. Richards, S. A., Lounsbury, K. M., Carey, K. L., and Macara, I. G. (1996). McGregor, F., Phelan, A., Dunlop, J., and Clements, J. B. (1996). Regu- A nuclear export signal is essential for the cytosolic localization of lation of herpes simplex virus type 1 poly(A) site usage and the action the Ran binding protein, RANBP1. J. Cell Biol. 134, 1157±1168. of the immediate-early protein IE63 in the early-late switch. J. Virol. Roizman, B., and Sears, A. E. (1996). Herpes simplex viruses and their 70, 1931±1940. replication. In ``Fundamental Virology'' (B. N. Fields, P. M. Howley, and McLauchlan, J., Phelan, A., Loney, C., Sandri-Goldin, R. M., and Clem- D. M. Knipe, Eds.), 3rd ed., pp. 1043±1108. Lipincott±Raven, Philadelphia. ents, J. B. (1992). Herpes simplex virus IE63 acts at the posttranscrip- Sacks, W. R., Greene, C. C., Aschman, D. P., and Schaffer, P. A. (1985). tional level to stimulate viral mRNA 3Ј processing. J. Virol. 70, 6939± Herpes simplex virus type 1 ICP27 is an essential regulatory protein. 6945. J. Virol. 55, 796±805. Mears, W. E., Lam, V., and Rice, S. A. (1995). Identification of nuclear and Sandri-Goldin, R. M., and Mendoza, G. E. (1992). A herpesvirus regu- nucleolar localization signal sequences in the herpes simplex virus latory protein appears to act post-transcriptionally by affecting mRNA regulatory protein ICP27. J. Virol. 69, 935±947. processing. Genes Dev. 6, 848±863. Mears, W. E., and Rice, S. A. (1996). The RGG box motif of herpes Sandri-Goldin, R. M., Hibbard, M. K., and Hardwicke, M. A. (1995). The simplex virus ICP27 protein mediates an RNA-binding activity and C-terminal repressor region of herpes simplex virus type 1 ICP27 is determines in vivo methylation. J. Virol. 70, 7445±7453. required for redistribution of small nuclear ribonucleoprotein parti- Meier, U. T., and Blobel, G. (1992). Nopp140 shuttles on tracks between cles and splicing factor SC35; however, these alterations are not nucleolus and cytoplasm. Cell 70, 127±138. sufficient to inhibit host cell splicing. J. Virol. 69, 6063±6076. Meyer, B. E., and Malim, M. H. (1994). The HIV-1 trans-activator Schmidt-Zachmann, M. S., Dargemont, C., KuÈhn, L. C., and Nigg, E. A. shuttles between the nucleus and the cytoplasm. Genes Dev. 8, (1993). Nuclear export of proteins: The role of nuclear retention. Cell 1538±1547. 74, 493±504. Meyer, B. E., Meinkoth, J. L., and Malim, M. H. (1996). Nuclear transport Smiley, J. R., Panning, P., and Smibert, C. A. (1991). Regulation of cellular of human immunodeficiency virus type 1, visna virus and equine gene products by HSV products. In ``Herpesvirus Transcription and infectious anemia virus Rev proteins: Identification of a family of Its Regulation'' (E. K. Wagner, Ed.), pp. 151±179. CRC Press, Boca transferable nuclear export signals. J. Virol. 70, 2350±2359. Raton, FL. Michael, W. M., Choi, M., and Dreyfuss, G. (1995). A nuclear export Spencer, C. A., Dahmus, M. E., and Rice, S. A. (1997). Repression of host signal in hnRNP A1: A signal mediated, temperature dependent RNA polymerase II transcription by herpes simplex virus type 1. nuclear protein export pathway. Cell 83, 145±422. J. Virol. 71, 2031±2040. Moore, M. S. (1996). Protein translocation: Nuclear exportÐOut of the Soliman, T. M., Sandri-Goldin, R. M., and Silverstein, S. J. (1997). Shut- dark. Curr. Biol. 6, 137±140. tling of the herpes simplex virus type 1 regulatory ICP27 between the Nakielny, S., and Dreyfuss, G. (1996). The hnRNP C proteins contain a nucleus and cytoplasm mediates the expression of late proteins. nuclear retention sequence that can override nuclear export signals. J. Virol. 71, 9188±9197. J. Cell Biol. 134, 1365±1373. Uprichard, S. L., and Knipe, D. M. (1996). Herpes simplex ICP27 mutant Nigg, E. A. (1997). Nucleocytoplasmic transport: Signals, mechanisms, viruses exhibit reduced expression of specific DNA replication and regulation. Nature 386, 779±786. genes. J. Virol. 70, 1969±1980. Palmieri, D., and Malim, M. (1996). The human T-cell leukemia virus Wen, W., Meinkoth, J. L., Tsien, R. Y., and Taylor, S. S. (1995). Identifica- type 1 posttranscriptional trans-activator rex contains a nuclear ex- tion of a signal for rapid export of proteins from the nucleus. Cell 82, port signal. J. Virol. 70, 6442±6445. 463±473.