Arginine Methylation of the RGG Box Does Not Appear To Regulate ICP27 Import during Herpes Simplex Virus Infection

Abstract

Arginine methylation can regulate protein import and export and can modulate protein interactions. Herpes simplex virus 1 (HSV-1) ICP27 is a shuttling protein involved in viral mRNA export. We previously reported that ICP27 is methylated on three arginines within its RGG box and that arginine methylation regulates ICP27 export and its interaction with SRPK1 and Aly/REF. Here, we report that ICP27 was efficiently imported into the nucleus when hypomethylated as determined by Fluorescence Recovery After Photobleaching (FRAP). Furthermore, coimmunoprecipitation of ICP27 with β-importin was not significantly affected by ICP27 hypomethylation. Thus, ICP27 import does not appear to be regulated by arginine methylation.

TEXT

Protein-arginine methylation is a posttranslational modification commonly found in RNA binding proteins that shuttle between the nucleus and cytoplasm (2, 3). Protein-arginine methylation is catalyzed by a family of enzymes known as protein arginine methyltransferases (PRMT), of which at least nine members have previously been identified (1). PRMT1 is the main methyltransferase in human cells (24), and this enzyme usually recognizes arginines within a glycine-arginine-rich region, which is a motif that is present in many RNA and DNA binding proteins. Arginine methylation has been shown to be important for modulating protein-protein interactions (3, 14, 21, 32, 33), for regulating export of shuttling proteins (11, 20, 28, 35), and for regulating nuclear importation of several regulatory proteins in mammalian cells (9, 15, 25, 29).

ICP27 is a multifunctional regulatory protein that shuttles between the nucleus and cytoplasm in herpes simplex virus 1 (HSV-1)-infected cells. ICP27 interacts with a number of cellular and viral proteins, and it binds to RNA to facilitate viral mRNA export through the TAP/NXF1 export receptor (27). ICP27 encodes an RGG box RNA binding domain, and ICP27 has been shown to be methylated on arginine residues in vivo (22). Using mass spectrometric analysis, we identified three arginine residues within the RGG box that were methylated in vivo: specifically, the arginines at positions 138, 148, and 150 (30). We constructed viral mutants in which these arginines were replaced with lysines either singly, in pairs, or in sets of all three, and we found that growth of these viral mutants was impaired compared to growth of wild-type (WT) HSV-1 (30). Furthermore, the interaction of ICP27 with two cellular proteins, SRPK1, an SR-splicing protein-specific kinase, and Aly/REF, an RNA export adaptor protein, was diminished under conditions of hypomethylation that resulted from infection with the substitution mutants or from adding the methylation inhibitor adenosine dialdehyde (AdOx) to HSV-1 KOS-infected cells (31). Although arginine methylation has been shown to affect the RNA and DNA binding activities of several proteins (4, 14), we recently reported that hypomethylation of ICP27 does not affect its ability to bind GC-rich sequences in vitro; neither does it appear to affect RNA export during HSV-1 infection (8). However, we found that arginine methylation does regulate ICP27 export to the cytoplasm. In vitro export assays showed that hypomethylated ICP27 was exported to the cytoplasm earlier and at a higher rate during infection (30).

The role of arginine methylation in regulating the export of shuttling proteins has been well documented (2, 3, 11, 20, 21, 23, 28, 30). It has also been reported that arginine methylation regulates the import of several regulatory proteins with diverse functions, including Sam68, an RNA binding protein involved in export of unspliced HIV RNAs (9); RNA helicase A, which is involved in unwinding doubled-stranded RNA and DNA (29); high-molecular-weight forms of fibroblast growth factor 2 (25); and adenovirus 100K protein, which inhibits translation of cellular mRNA in the cytoplasm and regulates virion assembly in the nucleus (15). We previously reported that the RGG box of ICP27, which is the primary site of arginine methylation, contributes to ICP27's efficient nuclear localization (13). Here, we sought to determine whether import of ICP27 might also be regulated by arginine methylation.

To quantify ICP27 nuclear and cytoplasmic levels during infection under conditions of hypomethylation, we performed nuclear and cytoplasmic fractionation analysis. HeLa cells were infected with HSV-1 KOS at a multiplicity of infection (MOI) of 10. AdOx (20 μM) was added at 1 h after infection. For samples to be labeled with [³⁵S]methione or l-[methyl-³H]methionine, we added methionine-free media to infected cells with or without AdOx (20 μM) 1 h after infection. Three sets of infection experiments were performed. One infection set was not radiolabeled and was destined for immunoblot analysis. The second set was labeled with 150 μCi [³⁵S]methionine added 1 h before harvesting. For the third set, cycloheximide (100 μg/ml) was added for 30 min to halt new protein synthesis and 45 μCi l-[methyl-³H]methionine was then added for 1 h before harvesting to monitor protein arginine methylation as described previously (30). In this way, [³H]methionine is not incorporated into newly synthesized protein but is transferred to proteins through methylation by PRMTs. Infected cells were harvested at 3, 5, and 7 h after infection. Nuclear and cytoplasmic fractionation was performed as described previously (26), and immunoprecipitation (IP) was performed using anti-ICP27 P1119 antibody (Virusys). Samples were fractionated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Samples that were not radiolabeled were transferred to nitrocellulose, and the resulting blot was probed with anti-ICP27 antibody (Fig. 1A, top panels). For samples labeled with [³⁵S]methionine, the gel was dried after fractionation and was exposed to film (Fig. 1A, middle panels). For samples labeled with [³H]methionine (Fig. 1A, bottom panels), the gel was analyzed by fluorography as described previously (30). Methylation of ICP27 was reduced by about 50 to 70% in the AdOx-treated samples (Fig. 1A, bottom panels), which is consistent with results reported previously (24).

Fig. 1.

Hypomethylation does not prevent import of ICP27 into the nucleus. (A) HeLa cells were infected with HSV-1 KOS at an MOI of 10 in the absence or presence of AdOx (added at 1 h after infection). In three parallel sets of infection experiments, cells were left unlabeled or were labeled with [³⁵S]methionine or [³H]methionine as indicated. Infected cells were harvested at 3, 5, and 7 h after infection. Nuclear (Nuc) and cytoplasmic (Cyto) fractionation was performed as described previously (26), and immunoprecipitation (IP) was performed using anti-ICP27 antibody. The top panels show a Western blot probed with anti-ICP27 antibody. The asterisks mark the position of heavy-chain IgG. The middle panels show autoradiographs of [³⁵S]methionine-labeled samples, and quantification of the nucleus-to-cytoplasm ratios obtained by densitometer scanning of the autoradiograph is shown in the right panel. The bottom panels show the fluorograph of [³H]methionine-labeled samples. β-Tubulin is shown as a control for the nuclear-cytoplasmic fractionation. (B) HeLa cells were infected with HSV-1 KOS or viral mutant ΔRGG,R138,150K or R138,148,150K for 5 or 8 h. Nuclear (N) and cytoplasmic (C) fractions were immunoprecipitated with anti-ICP27 antibody, and Western blots were probed with the same antibody. β-Tubulin data are shown as a control for the nuclear-cytoplasmic fractionation. The blots were scanned by densitometry, and the nucleus-to-cytoplasm ratios are shown.

ICP27 was seen in the nuclear fraction at 3 h after infection in the AdOx-treated samples and in the untreated samples (Fig. 1A). The autoradiograph of the ³⁵S-labeled samples was analyzed by densitometry, and nucleus-to-cytoplasm ratios are shown at the right side of the panels. The nucleus-to-cytoplasm ratios for ICP27 were similar at 3 h for untreated (1.8) and AdOx-treated (1.9) samples, indicating that ICP27 was efficiently imported at early times after infection under both sets of conditions. At 5 and 7 h after infection, the nucleus-to-cytoplasm ratio for the untreated samples remained at 1.9 to 1 whereas the ratios were about 1 to 1 and 1.3 to 1 for the AdOx-treated samples at 5 and 7 h, respectively (Fig. 1A). This could mean that, whereas AdOx-treated ICP27 was imported as efficiently as untreated ICP27 at early times after infection, methylated ICP27 was imported more efficiently as infection proceeded. Alternatively, this result could reflect the faster export of ICP27 under conditions of hypomethylation that we reported previously (30).

Next, we examined the nuclear and cytoplasmic distribution of ICP27 in viral mutants in the RGG box. HeLa cells were infected with wild-type KOS, viral mutant ΔRGG, in which the RGG box is deleted (19), mutant R138,150K, with lysine substitutions in the RGG box at positions R138 and R150, and mutant R138,148,150K, with three substitutions in the RGG box (Fig. 1B). These mutants were described previously (30, 31). Infected cells were harvested at 5 and 8 h after infection, and nuclear and cytoplasmic fractions were immunoprecipitated with anti-ICP27 antibody. Western blot analysis showed that ICP27 was detected in the nuclear fractions by 5 h after infection with RGG box mutants (Fig. 1B). In similarity to the results seen with AdOx presented in Fig. 1A, the nucleus-to-cytoplasm ratios for the RGG box mutants at 5 and 8 h after infection ranged from a ratio of 1.6 to 1 to a ratio of 1.1 to 1, whereas, with KOS infection, the nucleus-to-cytoplasm ratio at 5 h after infection was 3.3 to 1 and at 8 h after infection was 2.4 to 1 (Fig. 1B, right panel). Again, this could mean that hypomethylated ICP27 is imported less efficiently at 5 and 8 h after infection or it could reflect the earlier export we previously reported when ICP27 was hypomethylated (6, 30). We explored this further to differentiate between these possibilities, as is shown in subsequent figures.

Because arginine methylation can also affect protein interactions (2, 3, 31) and because we showed previously that interaction with cellular proteins SRPK1 and Aly/REF in the presence of hypomethylated ICP27 was reduced (31), we sought to determine whether hypomethylation might affect the interaction of ICP27 with β-importin, with which it interacts for import into the nucleus (18). We performed coimmunoprecipitation experiments. HeLa cells were mock infected or were infected with HSV-1 KOS in the absence or presence of AdOx added 1 h after infection. At 5 and 8 h after infection, cell lysates were immunoprecipitated with anti-ICP27 antibody, immune complexes were fractionated by SDS-PAGE, and proteins were transferred to nitrocellulose. Blots were probed with anti-β-importin 3E9 antibody (Pierce) as shown in Fig. 2A. The experiment was performed three times, and densitometer analysis data were averaged for the three experiments as shown in Fig. 2B. Approximately 20% more β-importin coprecipitated with ICP27 in the AdOx-treated samples at 5 h than in the untreated samples, whereas approximately 30% less β-importin was found to coimmunoprecipitate with ICP27 in AdOx-treated samples than in untreated samples at 8 h after infection. We consider it unlikely that either a 20% increase or a 30% decrease in the amount of β-importin that coimmunoprecipitated with ICP27 under conditions of hypomethylation would have a significant effect on ICP27 import.

Fig. 2.

Coimmunoprecipitation of ICP27 with β-importin. (A) HeLa cells were infected with HSV-1 KOS in the absence or presence of AdOx for 5 or 8 h. Cell lysates were immunoprecipitated with antibody to ICP27, and the Western blot was probed with anti-β-importin antibody. After extensive washing, the same blot was then probed with anti-ICP27 antibody. Western blots of input cell extracts are also shown. β-Actin was used a loading control. (B) Immunoprecipitated β-importin normalized to β-importin in the input samples from three experiments was analyzed by densitometry. Standard deviations are shown.

To further attempt to distinguish import and export of ICP27 under conditions of hypomethylation, we performed immunofluorescent staining experiments. HeLa cells were infected with HSV-1 KOS at an MOI of 5 in the absence or presence of AdOx added at 1 h after infection and in the absence or presence of cycloheximide (100 μg/ml) added 2 h after infection. Cycloheximide was added to inhibit new protein synthesis so that the localization of ICP27 could be observed in the absence of newly synthesized protein. Samples were fixed and stained for ICP27 at 3, 4, 5, 6, 7, and 8 h after infection (Fig. 3A). In KOS-infected cells in the absence of AdOx, ICP27 was observed in the nucleus at 3, 4, and 5 h after infection and cytoplasmic fluorescence was not observed until 6 h after infection, in accord with our previous findings (6, 7, 16). Similar results were seen with KOS-infected cells treated with cycloheximide (Fig. 3A). In contrast, in the presence of AdOx, whereas ICP27 was seen in the nucleus at 3 h after infection, cytoplasmic florescence was observed by 4 h after infection, and this was also the case for KOS-infected cells treated with AdOx and cycloheximide to halt new protein synthesis. This indicates that ICP27 was efficiently imported into the nucleus under conditions of hypomethylation and that nuclear ICP27 was exported earlier under conditions of hypomethylation. This was corroborated in experiments performed with ICP27 viral mutants dLeu and n406, which are unable to be exported to the cytoplasm because these mutants cannot interact with TAP/NXF1 (6, 17) (Fig. 3B). At 8 h after infection with dLeu and n406, ICP27 was seen in the nucleus, and this was also what was observed in the presence of AdOx (Fig. 3B). This indicates that ICP27 was imported under conditions of hypomethylation but that ICP27 could not be exported because these mutant proteins cannot interact with TAP/NXF1, which is required for ICP27 export to the cytoplasm (17).

Fig. 3.

ICP27 is imported into the nucleus under conditions of hypomethylation. (A) HeLa cells were infected with HSV-1 KOS at an MOI of 5. Infected cells were treated with AdOx, cycloheximide (CH), or AdOx plus CH as indicated or left untreated. At 3, 4, 5, 6, 7, and 8 h after infection, cells were fixed and stained with anti-ICP27 P1119 antibody. White arrows mark cytoplasmic ICP27. (B) HeLa cells were infected with HSV-1 dLeu or n406 in the absence or presence of AdOx as indicated for 8 h, at which time the cells were fixed and stained. Images were viewed on a Zeiss Axiovert microscope at ×100 magnification. At least 5 fields of 10 or more cells were viewed for each set of conditions and time points, and the images shown are representative of the majority of cells in each field.

Although the immunofluorescence studies suggested that methylation is not critical for ICP27 import, the results are qualitative and do not directly address the efficiency of import. To determine whether the rate of ICP27 import into the nucleus is affected by hypomethylation, we performed Fluorescence Recovery After Photobleaching (FRAP) (5, 10). HeLa cells were infected with vN-YFP-ICP27 (16) at an MOI of 5 in the absence or presence of AdOx added 1 h after infection. vN-YFP-ICP27 virus encodes ICP27 tagged at the N terminus with yellow fluorescent protein (YFP). N-terminally tagged ICP27 virus behaves like wild-type ICP27 in KOS-infected cells (12, 16). At 6 h after infection, photobleaching was performed on living cells, using a Zeiss LSM 510 confocal microscope and FRAP software. The laser was set at 100% power, and bleaching was performed using the 514-nm laser line. The acousto-optical tunable filter (AOTF) was set at 100%, and 50 iterations were performed. Images were captured every 125 s after bleaching. Figure 4A shows a representative field of cells treated with AdOx and a field not treated with AdOx. The left panels show the infected cells before bleaching. In the middle panels, the bleached regions, which encompass the entire nucleus of each cell, are circled. The right panel shows the cells after recovery. Over time, the amount of fluorescence in the photobleached area increases as unbleached molecules move into the bleached nuclei from the cytoplasm. The percentages of fluorescence recovery for the cell marked by the pink arrow in the +AdOx field and for the cell marked by the red arrow in the no-AdOx field are shown in Fig. 4B. Percent fluorescence recovery is the amount of fluorescence after photobleaching divided by the amount of fluorescence before photobleaching multiplied by 100 according to the formula (Y/X) × 100 = % recovery, where the percentage of fluorescence lost due to photobleaching is represented by X and the amount of fluorescence that returned to the bleached area is represented by Y. The lateral mobility is determined by the slope of the curve: the steeper the curve, the faster the recovery. There were no discernible differences in recovery in the no-AdOx cell compared to the +AdOx cell, indicating that YFP-ICP27 moved from the cytoplasm into the bleached nuclei with similar kinetics. This was also the case for the other cells in the fields shown that were bleached and for two other fields of AdOx-treated and untreated vN-YFP-ICP27-infected cells (data not shown). We conclude that hypomethylation does not affect nuclear import of ICP27.

Fig. 4.

FRAP analysis of import of ICP27 under conditions of hypomethylation. (A) HeLa cells were infected with vN-YFP-ICP27 at an MOI of 5. At 6 h after infection, photobleaching was performed using living infected cells, a Zeiss LSM 510 confocal microscope, and FRAP software. The entire circled nucleus in each cell was photobleached. The laser was set at 100% power, and bleaching was performed using the 514-nm laser line. The acousto-optical tunable filter (AOTF) was set at 100%, and 50 iterations of the procedure were performed. Images were captured every 125 s after bleaching. (B) The percentages of fluorescence recovery corresponding to the indicated time points for the cell marked by the pink arrow in the +AdOx field and for the cell marked by the red arrow in the no-AdOx field are shown. Percent fluorescence recovery is the amount of fluorescence after photobleaching divided by the amount of fluorescence before photobleaching multiplied by 100 according to the formula (Y/X) × 100 = % recovery, where the percentage of fluorescence lost due to photobleaching is represented by X and the amount of fluorescence that returned to the bleached area is represented by Y.

Arginine methylation has been shown to affect protein trafficking and protein interactions (2, 3). Using in vitro export assays, we showed previously that hypomethylated ICP27 is exported from the nucleus more rapidly and at earlier times after infection than wild-type ICP27 (30). We further showed that arginine methylation of ICP27 occurs primarily in the nucleus (30), which is in accord with a recent report that showed that ICP27 is methylated primarily by the cellular arginine methyltransferase PRMT1, which is predominantly nuclear (34). In this study, we found that hypomethylation of ICP27 did not affect import of ICP27 into the nucleus and that hypomethylation did not appear to affect the rate of import as measured by FRAP, which is in contrast to the results seen with ICP27 export. Therefore, defects in viral replication seen in RGG box substitution mutants can be attributed to effects of hypomethylation on export (30) and regulation of functional interactions with cellular proteins (31). Hypomethylation of ICP27 does not affect its ability to bind RNA (8), and, as we showed here, there is no demonstrable effect on ICP27 import.