Comparison of the Rta/Orf50 Transactivator Proteins of Gamma-2-Herpesviruses

Alignment of the rhadinovirus Rta/Orf50 proteins.

The three Rta/Orf50 proteins of KSHV, RRV, and MHV-68 are well conserved, raising the possibility that the transactivation functions of the Rta/Orf50 proteins are likewise conserved among this group of proteins. Figure 1A depicts the alignment of the amino acid sequences encoded by the Rta/Orf50 proteins of KSHV, RRV, and MHV-68, which was performed using the ClustalW program. The amino acid similarity between KSHV and RRV Rta/Orf50 is 55%, the amino acid similarity between KSHV and MHV-68 Rta/Orf50 is 43%, and the amino acid similarity between RRV Rta/Orf50 and MHV-68 Rta/Orf50 is 47%. Amino acids 1 to 380 are conserved in all three proteins, followed by a more deletion-prone region and two more conserved blocks at amino acids 500 to 570 and 615 to 660. Expression of truncated versions of the KSHV Rta/Orf50 protein revealed that DNA binding was mediated by the amino-terminal 272 amino acids (33), which is consistent with our alignment in Fig. 1A, where this region is conserved among all three Rta/Orf50 rhadinovirus proteins. Figure 1B depicts a Western blot of the three Rta/Orf50 proteins expressed in 293 cells. All three proteins exhibit a higher apparent molecular weight than that predicted by their amino acid sequence. This is consistent with prior reports that showed that the KSHV and MHV-68 proteins are posttranslationally modified by phosphorylation (8, 31, 34, 47, 49, 57).

Comparison of KSHV, RRV, and MHV-68 Rta/Orf50 activities on a panel of KSHV promoters.

In order to evaluate the impact of the different Rta/Orf50 proteins with regard to the KSHV viral life cycle, we compared the ability of the KSHV, RRV, and MHV-68 transactivators to activate KSHV promoters in three relevant cell lines: KSHV-negative SLK endothelial cells, KSHV-negative BJAB B lymphocytes, and 293 epithelial cells. KSHV has been shown to be tropic for both B cells and endothelial cells and has also previously been shown to infect and replicate in 293 epithelial cells (17, 41, 42, 60). The KSHV promoters we used have been previously described (28, 31, 33-35). These include the immediate-early/early promoters KSHV Rta/Orf50, K8, and Mta/Orf57. The set of delayed early promoters included promoter regions for KSHV Nut-1 (also called Pan), viral G-protein-coupled receptor (vGPCR), DNA binding protein (DBP), vIL-6, TK, kaposin (or K12), and DNA polymerase (POL). The set of late gene promoters included those for glycoprotein B (gB) and assembly protein (AP). To control for experimental variation, transfections were performed in duplicate and repeated five times in each cell line. To control for transfection efficiency, the results were normalized using a β-galactosidase expression vector, which was cotransfected with the Rta expression plasmids and all the luciferase reporter plasmids in each experiment. The same cell extracts were used to measure the β-galactosidase activity and luciferase activity of each sample. To control for similar expression levels of all three Rta/Orf50 proteins, the three gamma-2-herpesvirus Rta/Orf50 cDNAs were all expressed from a pcDNA3 expression vector (Invitrogen). The three Rta/Orf50 proteins showed similar levels of expression in each cell line when expressed from this vector (Fig. 1C). To determine the effect of Rta/Orf50 on the different promoters, transactivation by Rta/Orf50 was calculated as fold activation compared to pcDNA3 vector alone. This normalization procedure allowed us to compare the relative effect of the three Rta/Orf50 proteins on all promoters in all cell lines.

BJAB is an Epstein-Barr virus- and KSHV-negative B-cell line. BJAB cells were electroporated with the promoter constructs described above and one of the three Rta/Orf50 expression plasmids. Table 1 depicts the transcriptional activation seen with these constructs in BJAB B lymphocytes. (i) Promoter activity, which was measured in luciferase units for each of the plasmids and normalized to β-galactosidase activity, is represented as fold activation over the pcDNA3 vector control (see also Fig. 2A). (ii) To compare the efficiency of the three Rta/Orf50 proteins, the activation of each KSHV promoter by the KSHV Rta/Orf50 transactivator was set at 100%, and the ability of RRV and MHV-68 transactivators to activate the same promoter was calculated as a percentage of the activity seen with the KSHV Rta/Orf50 protein. Although the RRV and MHV-68 proteins could transactivate the KSHV promoters of all three groups of immediate-early, early, and late genes, overall transactivation efficiencies were lower than that seen with the autologous KSHV Rta/Orf50 transactivator. In BJAB cells, KSHV Rta/Orf50 was able to transactivate the immediate-early promoter K8 significantly, whereas the other two Rta/Orf50 homologs failed to do so to any significant degree above background. This was also true for the delayed early promoters, vGPCR and kaposin, and the gB promoter. This phenotype is consistent with a report by Chang et al. (4) and KSHV array analysis (27, 40) that found gB mRNA and the gB promoter fragment used here exhibit viral early kinetics. The late AP promoter was not significantly activated by any of the Rta/Orf50 proteins. Both primate Rta/Orf50 proteins could activate the KSHV Nut-1 promoter, whereas the MHV-68 Rta/Orf50 protein was not able to significantly transactivate this promoter.

An identical set of experiments were performed in 293 epithelial cells. 293 epithelial cells have been previously shown to support KSHV infection in vitro (17, 41, 42, 60). We transfected the aforementioned promoter constructs into these cells along with the KSHV, RRV, or MHV-68 Rta/Orf50 expression plasmids as described above. Promoter activity was measured by luciferase units for each of the plasmids and depicted as fold activation over the pcDNA3 vector control (Table 1 and Fig. 2B). As can be seen in Table 1, although the RRV and MHV-68 proteins could transactivate all three classes of promoters, their efficiencies were lower compared to the autologous KSHV protein, as was the case in BJAB cells. In 293 cells, the KSHV Rta/Orf50 protein was able to transactivate the immediate-early/early K8 promoter significantly, whereas the other Rta/Orf50 transactivators could not. This was also true for the delayed early promoter, vGPCR, and the late gB promoter. However, in 293 epithelial cells, the RRV Rta/Orf50 transactivator was able to moderately activate the Nut-1, DBP, kaposin, TK, and AP promoters, whereas MHV68 Rta/Orf50 did not significantly activate these promoters.

SLK is a KS tumor-derived endothelial cell line that is KSHV negative (25, 36, 41). It was of particular interest to measure promoter activity in these cells since KSHV can infect endothelial cells in vitro and KS lesions are comprised of a large number of infected endothelial cells. Furthermore, KSHV Rta/Orf50 can induce KSHV lytic transcription experimentally in in vitro KSHV- infected SLK cell populations (R. Renne and D. Dittmer, unpublished observation). We transfected the aforementioned promoter constructs into these cells along with the KSHV, RRV, or MHV-68 Rta/Orf50 expression plasmids as described above. Table 1 and Fig. 2C depicts the transcriptional activation seen with these constructs in SLK endothelial cells. Once again, the promoter activity for each construct is represented as fold activation over the pcDNA3 vector control. As can be seen in Table 1, although the RRV and MHV-68 proteins could transactivate all three classes of promoters, their transactivation efficiencies were lower than that seen with the autologous KSHV protein, a pattern similar to the results we obtained in BJAB and 293 cells. Furthermore, the KSHV Rta/Orf50 protein was able to transactivate the immediate-early/early promoter, K8, significantly, whereas the two other Rta/Orf50 transactivators could not. This was also true for the delayed early promoter, vGPCR, as well as the kaposin promoter and the gB promoter. In SLK cells, the RRV Rta/Orf50 transactivator was able to moderately activate the Nut-1 promoter whereas MHV-68 Rta/Orf50 did not significantly activate this promoter.

In summary, we tested a panel of 12 KSHV promoters in three different cell lines that are relevant for KSHV tropism/infection, which include B cells, endothelial cells, and epithelial cells. In all three cell lines, the KSHV viral transactivator activated the autologous KSHV promoter significantly, as has been previously described (3, 6, 7, 10, 19, 30, 33, 34, 39, 43, 46, 49, 51, 57, 59, 61). These studies provide for the first simultaneous comparison of primate Rta/Orf50 proteins. In contrast to KSHV, the RRV and MHV-68 Rta/Orf50 transactivators were notably reduced in their ability to transactivate this panel of KSHV promoters. In this set of KSHV promoters, the overall response patterns to KSHV Rta/Orf50 were similar in BJAB B-cells and SLK endothelial cells (Fig. 2). Surprisingly, in 293 cells the Mta/Orf57 promoter was less responsive to KSHV Rta/Orf50 relative to the Nut-1 promoter, while the opposite was true in SLK and BJAB cells. This suggests that the compositions of host factors that mediate Rta/Orf50's effects differ between 293 and the other two cell lines and underscores the need to conduct transactivation studies in multiple cell lines. Neither the RRV nor MHV-68 Rta/Orf50 proteins were able to activate the KSHV K8, vGPCR GPCR, or the gB promoter. This suggests that some interactions between human cellular transcription factors and KSHV Rta/Orf50 are not evolutionary conserved but are required for at least this subset of KSHV promoters. We would predict that a subset of basic transcription factors that interacts with KSHV Rta/Orf50 (20) includes proteins that would not interact with either the RRV or MHV-68 Rta/Orf50 proteins. The regulation of vGPCR, K8, and gB promoters by KSHV Rta/Orf50 appears to be unique to the human virus and may explain some aspects of its biology.

Rta/Orf50's interaction with p53 is restricted to the primate viruses.

KSHV Rta/Orf50 has been demonstrated to inhibit p53's transactivation function (23), which is independent of Rta/Orf50's specific DNA-binding function. We compared the ability of MHV-68 and RRV Rta/Orf50 transactivators to inhibit p53's transactivation function by using a well-established p53-responsive element (58). As seen in Fig. 2D, only RRV and KSHV Rta/Orf50 were able to interfere with human p53's transactivation function, but MHV-68 Rta/Orf50 could not. This suggests that the interaction between the primate Rta/Orf50 proteins and human p53 has evolved later during speciation. Similar to the interaction between the human papillomavirus E6 proteins and p53 (45, 54), which can be interrupted by slight amino acid variation between “high-risk” and “low-risk” strains, the interaction between Rta/Orf50 and p53 also shows high specificity.

Comparison of the genome-wide effects of KSHV, RRV, and MHV-68 Rta/Orf50 proteins on KSHV reactivation.

In order to evaluate the impact of the different Rta/Orf50 proteins on the KSHV life cycle, we analyzed the effect of expression of the individual Rta/Orf50 proteins on KSHV reactivation from latently infected BCBL-1 cells. KSHV Rta/Orf50 has previously been shown to efficiently induce KSHV lytic replication in these cells (34, 50, 51), and KSHV Rta/Orf50 alone is necessary and sufficient to induce the latent to lytic switch. We electroporated 10⁷ BCBL-1 cells with 20 μg of empty vector pcDNA3 or the three different pcDNA3-Orf50 expression constructs along with 5 μg of pEGFP plasmid (Clontech, Inc.). Forty-eight hours postelectroporation, the cells were sorted based upon GFP fluorescence. No cells were excluded based on scatter, since KSHV reactivation in BCBL-1 cells induces a blast phenotype, which changes the light scatter profile of the cells. In independent experiments, between 5 × 10³ and 10 × 10³ GFP-positive cells were sorted and mRNA was isolated, reverse transcribed, and analyzed using our previously developed real-time quantitative reverse transcription-PCR (RT-PCR) array for KSHV (12, 16). We previously showed that glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels do not change upon KSHV reactivation in BCBL-1 cells (16). We also tested several other cellular mRNAs in this system, including those for actin, 18S rRNA, and hypoxanthine phosphoribosyltransferase. These did not change in response to KSHV reactivation (data not shown). Since these mRNAs correlated with each other, as expected based on prior studies (12, 16), we chose to normalize all data based upon the level of GAPDH mRNA and subjected the normalized data to statistical analysis (52). Figure 3A shows a Treeview representation after hierarchical clustering of the data from cells sorted with a fluorescence-activated cell sorter 48 h after treatment (15). Overall, the repeat experiments showed an identical pattern: relative to mock-treated cells, KSHV Rta/Orf50 significantly induced the transcription of all KSHV lytic mRNAs (indicated by the dark shade of gray) but not cellular mRNAs or latency type I mRNAs. Electroporation of an RRV or MHV-68 Rta/Orf50 expression plasmid also induced KSHV transcription, but to a lesser degree than KSHV Rta/Orf50. This could presumably have occurred due to the nonhuman Rta/Orf50 proteins activating the KSHV Rta/Orf50 promoter and thus initiating a complete, albeit delayed, lytic replication cascade. Figure 3B shows the data from the total culture at 72 h postelectroporation. Here data were also normalized relative to mock-transfected cells as described in references 12 and 16. Relative to mock-transduced cultures, KSHV Rta/Orf50-transfected BCBL-1 cultures induced virtually all mRNAs, RRV Rta/Orf50-transfected cultures likewise induced all mRNAs, but to a lesser degree. Consistently, for every time point tested, MHV-68- Rta/Orf50-transfected BCBL cells exhibited the least amount of induction at this time point. At the 24-h time point, no significant differences in KSHV transcription were observed in cells transfected with three Rta/Orf50 homologs and cells transfected with a GFP expression plasmid (data not shown).

As in our transfection experiments, we could discern genes that were responsive to the Rta/Orf50 viral transactivator of any species. We also identified genes that were only activated by KSHV Orf50/Rta. This is exemplified in Fig. 3C, which shows the results of two independent experiments. Shown here are the normalized (ΔCTgapdh) values for a latent mRNA, v-cyclin (vCYC), and two early gene mRNA transcripts, Orf57/Mta and vGPCR. In this representation, lower ΔCT (change in cycle threshold [CT]) values correspond to higher mRNA levels on a logarithmic scale. As expected, v-cyclin mRNA was unaffected by KSHV, RRV, or MHV-68 Rta/Orf50 as compared to the vector-only-transfected BCBL-1 cells. We had previously shown that the LANA promoter, which also drives v-cyclin expression, is unaffected by KSHV Rta/Orf50 (28). The mRNA for Mta/Orf57 is significantly induced by RRV or MHV-68 Rta/Orf50 and even more so by KSHV Rta/Orf50 as compared to the vector-transfected cells. In contrast, only the KSHV Rta/Orf50 protein, but not the RRV or MHV-68 Rta/Orf50 protein, induces the mRNA for vGPCR, as evidenced by a 5-U decrease in ΔCT (corresponding to a 1.8⁵- or ∼19-fold difference). The pattern of transactivation of KSHV mRNA transcripts in the context of the viral genome parallels the pattern seen with the individual KSHV promoters, as shown in Table 1.

Figure 3D shows a plot of the levels of induction (ΔCT) by RRV or MHV-68 Rta/Orf50 for each of the KSHV genes (y axis) relative to the induction that was achieved by KSHV Rta/Orf50 alone (x axis). The values are represented on a logarithmic scale, to allow for a more robust statistical analysis (18). As can be gleaned from this two-dimensional representation, neither the MHV-68 nor RRV Rta/Orf50 proteins induced KSHV lytic mRNAs to the levels seen after transfection of KSHV Rta/Orf50 expression plasmid. However, both RRV Rta/Orf50 and MHV-68 Rta/Orf50 induced the complete KSHV reactivation program. Although the RRV Rta/Orf50 protein could activate KSHV promoters better than the MHV-68 protein in both reporter assays (Table 1) and the viral microarray (Fig. 3A and B), the primate Rta/Orf50 protein was not dramatically better at inducing KSHV reactivation as a whole than the MHV-68 protein, as evidenced by their overlapping expression profiles (Fig. 3D). This suggests a mechanism whereby MHV-68 or RRV Rta/Orf50 cannot fully substitute for KSHV Rta/Orf50 but may instead initiate lytic replication by inducing the KSHV immediate-early promoters, such as the KSHV Rta/Orf50 promoter. In this scenario, the rate-limiting step for complete, ordered lytic replication may also be dependent on the time postelectroporation of the BCBL-1 cells.

To validate our statistical analysis, we performed Gaussian clustering using a second different, genetic optimization algorithm with ArrayMiner 2.6 under Macintosh OsX (Optimal Design, Inc., Brussels, Belgium). We analyzed the complete data sets for these experiments followed by flow cytometry sorting for each rhadinovirus Rta/Orf50 as well as vector control (595 data points after exclusion of some PCR failures). The algorithm terminated with five essentially similar clusters (data not shown), as well as genes that did not change expression upon electroporation. The latter group included the cellular housekeeping mRNAs c-myc and hu-actin, as well as the KSHV latent mRNAs for LANA and v-cyclin (not shown). This is consistent with our prior observations, which found that neither LANA mRNA nor the LANA promoter is activated upon KSHV reactivation in BCBL-1 cells (11, 16, 28). All other KSHV mRNAs were induced by KSHV Rta/Orf50 relative to the vector. Four mRNAs that were most highly induced by KSHV Rta/Orf50 relative to vector were the KSHV K2, K4, Rta/Orf50, and K8 transcripts. RRV and MHV Rta/Orf50 also induced KSHV lytic mRNAs but much less efficiently.

Summary.

All herpesvirus lytic replication occurs in an ordered sequence of gene expression. Immediate-early mRNAs are transcribed in the absence of de novo protein synthesis, early mRNAs are activated by the immediate-early transactivating factors, and late mRNA transcription is blocked by inhibiting viral DNA replication. Once the immediate-early protein(s) is made, this transcriptional cascade culminates in capsid assembly, virion egress, and death of the host cell.

One question regarding the evolution of the gammaherpesvirus subfamily pertains to the degree of functional conservation that exists among the human, primate, and vole rhadinoviruses. We find that the immediate-early Rta/Orf50 proteins of KSHV, RRV, and MHV-68 are highly conserved in amino acid sequence (Fig. 1), as well as in their ability to activate relevant KSHV viral promoters (Fig. 2). KSHV Orf50 alone is necessary and sufficient to induce KSHV lytic reactivation in BCBL-1 cells (33-35, 50, 51). We demonstrate here that the RRV and MHV-68 Rta/Orf50 protein can function similarly. In addition, we found important qualitative as well as quantitative differences among the three homologous proteins: Some KSHV promoters (K8, vGPCR, gB, and kaposin) only responded to KSHV Rta/Orf50, but not to RRV or MHV-68 Rta/Orf50, while other promoters responded to all three (Rta/Orf50, Mta/Orf57, DBP, vIL-6, TK, Pol, and AP). This is not unexpected since Rta/Orf50 has multiple modes of action. It transactivates the Nut-1, Orf57, and K8 promoters by binding directly to DNA (33, 47, 49), while other promoters may be activated indirectly through protein-protein interactions. Interestingly, we had previously observed that KSHV Rta activated the RRV R8, Orf57, and gB promoters but only weakly activated the RRV v-IRF and Orf50 promoters (10).

In the context of latently infected BCBL-1 cells, addition of KSHV Rta/Orf50 was sufficient to induce the complete lytic transcription program, which agrees with a recent study by Nakamura et al. (38). Extending their work, we found that KSHV Rta/Orf50 was much better at inducing complete KSHV reactivation than either the RRV or MHV-68 Rta/Orf50 homologs (Fig. 3). This suggests that while either Rta/Orf50 homolog is sufficient to induce KSHV reactivation, other KSHV immediate-early or early proteins, such as KbZIP/K8 (56, 61), Mta/Orf57 (29), or Orf45 (61), cooperate with Rta/Orf50 to complete the lytic cascade. We surmise that this cooperation is species specific. The delayed replication kinetics in cells that are transfected by RRV or MHV-68 Rta/Orf50 suggests that enough KSHV Rta/Orf50 protein needs to accumulate in the BCBL-1 cells to circumvent this species barrier and activate complete lytic protein expression. Ultimately, mature KSHV virions are produced in this system, regardless of which rhadinovirus Rta/Orf50 protein initiates the process.