Differential Regulation of the Overlapping Kaposi's Sarcoma-Associated Herpesvirus vGCR (orf74) and LANA (orf73) Promoters

Cell lines.

HeLa, CV-1, SLK, NIH 3T3, and 293 cells (all from the American Type Culture Collection) were maintained in Dulbecco's modified Eagle medium (DMEM) (Cellgrow, Inc.) supplemented with 10% calf serum (Gibco-BRL), 2 mM l-glutamine, penicillin (0.05 μg/ml), and streptomycin (5 U/ml; Gibco-BRL) at 37°C under 5% CO₂. SLK cells are KS tumor-derived cells that grow indefinitely in culture and exhibit cobblestone, i.e., endothelial cell, morphology. They do not contain KSHV. SAOS-2, (10)1 (courtesy of G. Zambetti, St. Jude Children's Hospital), and BHK (American Type Culture Collection) cells were maintained in DMEM supplemented with 15% fetal bovine serum (Gibco-BRL), penicillin (0.05 μg/ml), and streptomycin (5 U/ml; Gibco-BRL) at 37°C under 5% CO₂. Cells were passaged at subconfluency (approximately every 3 days) in order to maintain a 3T3-like phenotype (16, 57). A fresh aliquot was thawed every 30 passages. LnCAP, BJAB, and BCBL-1 cells were cultured in a solution containing RPMI-1860, 25 mM HEPES (pH 7.55), 10% fetal bovine serum, 0.05 mM 2-mercaptoethanol, 1 mM sodium pyruvate, 2 mM l-glutamine, penicillin (0.05 μg/ml), and streptomycin (50 U/ml) (all Gibco-BRL) at 37°C under 5% CO₂. Cells were split every 5 days to 2 · 10⁵ cells/ml.

Real-time RT-PCR.

Quantitative DNA analysis and reverse transcription (RT)-PCR were carried out in duplicate using Taqman RT and Taqman PCR with Amplitaq Gold reagents (PE Biosystems Inc.). RT was carried out using a 2.5 μM concentration of random hexamers at 25°C for 10 min, 48°C for 30 min, and 95°C for 5 min. Real-time PCR was carried out using universal cycle conditions (2 min at 50°C, 10 min at 95°C, and then 40 cycles of 15 s at 95°C and 1 min at 60°C) on an ABI PRIZM 7000 sequence detector (18). Primers are described in Table 1. To prevent contamination, all PCRs were assembled in a segregated space in which neither KSHV virions nor cloned KSHV DNA was handled. Carryover of the amplification product was avoided using positive displacement pipettes and UNGglycosylase in the amplification reaction mixture (24).

Plasmids.

The source of genomic DNA for all clones was a KSHV genomic lambda library derived from a KS lesion (63). All nucleotide sequence positions are according to the numbering of Russo et al. (45). pDD130B contains a 1,879-bp PCR fragment (nt 127300 to 129179) amplified with primer 7326 (5′-TCGGGAAAGCTTGTCTGACA; nt 127300 to 127319 with an engineered HindIII site [underlined]) and with primer 7327 (5′-ctcgagCGGCCGCTAGCTTGTCACTCCCCTGA; nt 129161 to 129179 with XhoI [lowercase letters], NotI [underlined], and NheI [boldface type] sites) and inserted into pCR 2.1 (Invitrogen, Inc.). PCR was performed using Ready-To-Go PCR beads (Amersham) under the following conditions: 30 cycles of 30 s at 94°C, 1 min at 56°C, and 2 min at 74°C followed by 10 min at 74°C. pDD121B (nt 127539 to 128164) is derived from pDD130B by internal deletion of a 1,013-bp NheI fragment. pDD124 was constructed by subcloning the 1,879-bp HindIII/NotI fragment of pDD130B into pβgeo (courtesy of Limin Li, University of California—San Francisco). pDD125 was derived from pDD124 by an internal NheI deletion. pDD41 was constructed by subcloning the 1,763-bp NcoI fragment (nt 127607 to 129370) into the NcoI site of pGLbasic (Promega); pDD43 contains the same fragment in the opposite orientation. pDD53 and pDD83 were derived from pDD41 by an internal SmaI and an internal NheI deletion, respectively. pDD213, pDD268, and pDD270 were constructed by PCR amplification of a 373-bp fragment (nt 127546 to 127919), a 389-bp fragment (nt 127546 to 127935), and a 422-bp fragment (nt 127546 to 127968) from pDD39 using primer 7304 (5′-AGTCCTGGTGGCTCACCTGCC) and primer 7306 (5′-GCGGCGCCCGGGAC AATC), primer 7304 and primer 7331 (5′-CTCCGCCCTCCACTAC), and primer 7304 and primer 7332 (5′-AGCTGCCTCCAAATGATACACA), respectively. PCR products were gel purified and cloned into pCR2.1 (Invitrogen, Inc.). pDD271 contains a 349-bp NcoI/XhoI fragment of pDD213 in pGLbasic; pDD272 contains a 365-bp NcoI/XhoI fragment of pDD268 in pGLbasic. pDD274 contains a 398-bp NcoI/XhoI fragment of pDD270 in pGLbasic. pDD154 contains a 586-bp PCR fragment (nt 127297 to 127883) in pCRII-topo (Invitrogen, Inc.). pDD159 contains a 583-bp PCR fragment (nt 127300 to 127883) in pCRII-topo in the opposite direction. pDD163 and pDD168 were derived from pDD154 and pDD159 by subcloning the respective HindIII/XhoI fragments into pGL3basic (Promega, Inc.). pDD383 is a SacI deletion mutant of pDD163, with nucleotides 127300 to 127394 removed. pDD395 is an NheI/AvrII deletion mutant of pDD163, with nt 127300 to 27616 removed. All plasmids were sequenced in both orientations at the Oklahoma University Health Sciences Center sequencing core facility.

Transfection.

At day 1, cells were seeded to 2 · 10⁵ cells/10 ml/100-mm-diameter dish, 10⁵ cells/3 ml/35-mm-diameter dish (6-well plate), or 2 · 10⁴ cells/0.5 ml/well (12-well plate) to reach 50% confluency after 24 h. At day 2, 2,000 ng of total DNA together with 400 ng of pDD173 (pCDNA3.1-hislacZ; Invitrogen) was suspended in 200 μl of DMEM (no serum, no antibiotics) and 7.5 μl of Superfect (Qiagen, Inc.) was added. To minimize variability all plasmids were kept as 100-ng/μl stock solutions in Tris-EDTA, and normalizing plasmid was added to bulk DMEM before aliquoting for single transfection. The total DNA concentration was held constant to 2,400 ng by adding appropriate amounts of filler plasmid (pBluescript KS). We verified (data not shown) that our transfection data fell within the linear range of the assay using pDD83 reporter plasmid. Increasing amounts of reporter resulted in a commensurate increase in luciferase activity (normalized for transfection efficiency using 400 ng of cotransfected lacZ reporter plasmid). As little as 500 ng of pDD83 was sufficient to detect significant luciferase activity, while overall transfection efficiency dropped when more than 2,500 ng of DNA was added. The transfection mixture was incubated for 30 min at room temperature. For transfection in 35-mm-diameter dishes, the volume was adjusted to 1 ml with complete medium (DMEM, 10% fetal calf serum [FCS]). Cells were washed once, and 1 ml of transfection mixture per 35-mm-diameter dish was added. For transfection in 12-well plates, the volume was adjusted to 1.5 ml with complete medium (DMEM, 10% FCS). Cells were washed once with DMEM and 0.5 ml of transfection mixture was added per triplicate well. For transfection in 10-cm dishes, the volume was adjusted to 3 ml with complete medium (DMEM, 10% FCS). Cells were washed once with DMEM, and transfection mixture was added. Cells were incubated overnight at 37°C under 5% CO₂, and complete medium was exchanged. Luciferase activity was measured at 72 h after transfection using the Promega luciferase kit in a Turner TD20/20 luminometer according to the manufacturers' instructions.

Real-time quantitative RT-PCR distinguishes between KSHV latent and lytic mRNAs.

The distinction between lytic and latent mRNAs is fundamental to herpesvirus gene regulation. For KSHV Sarid et al. (47) described three classes of differentially transcribed messages in KSHV-infected BC-1 cells. Class I mRNAs can be detected in untreated (latent) cells and are not increased after 12-O-tetradecanoylphorbol-13-acetate (TPA) treatment, which reactivates KSHV. Class II mRNAs can be detected in untreated cells and are greatly increased after TPA treatment. Finally, class III mRNAs can only be detected in TPA-induced cells. Translated into the customary herpesvirus classification (44), class III mRNAs include lytic (alpha, beta, and gamma) and class I latent transcripts. Class II mRNAs present a conundrum, since it is not clear whether the mRNA detected in uninduced cells stems from the approximately 3% of cells that at any given time undergo spontaneous lytic replication or whether they represent latent mRNAs that are also induced during lytic replication. Here we employ quantitative real-time RT-PCR to assign KSHV mRNAs to their respective class: lytic or latent. Real-time quantitative PCR measures the amount of PCR product that is generated at each PCR cycle using a fluorescently labeled oligonucleotide which only fluoresces upon annealing to the amplified product. So-called C_t values indicate the cycle at which the fluorescence crosses a particular threshold. Hence, C_t values indicate the abundance of any given mRNA on a log scale. A low C_t value represents a highly abundant target mRNA. BCBL-1 cells are latently infected with KSHV and can be induced to release KSHV by TPA (42). Cells were either treated with TPA or mock treated, and total RNA was isolated at 72 h after induction. We then performed real-time quantitative RT-PCR on a dilution series of cells using primers listed in Table 1. It is important to note that all KSHV primers used in this experiment cross intron-exon borders and therefore do not amplify viral DNA. Figure 1 shows the result of our analysis. For all primers the real-time quantitative RT-PCR signal was linear over 4 orders of magnitude. The amount of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA remained unchanged in the presence or absence of TPA (Fig. 1A). This is indicated by the corresponding C_t values. The GAPDH C_t values for TPA-treated cells overlap the C_t values for mock-treated cells at every cell concentration tested. Hence, the input RNA concentration is the same. An identical pattern was observed using primers specific for the KSHV k-cyclin mRNA (Fig. 1B). Again, the C_t values and regression lines for TPA- and mock-treated samples superimpose. This establishes, quantitatively, that k-cyclin mRNA is transcribed independently of TPA or KSHV lytic genes, corroborating its classification as a latent transcript. Next, we tested the transcription pattern of orf50 and K14/vGCR, which are the subject of our subsequent analysis (Fig. 1C and D). Their real-time PCR analysis revealed a marked difference to the k-cyclin data. C_t values obtained from TPA-treated cells were much lower than those for mock-treated cells, indicating that orf50 and vGCR were induced upon TPA treatment. At the extreme end, quantitative real-time RT-PCR for orf29 (Fig. 1E) failed to detect any mRNA unless at least 500,000 latent BCBL-1 cells were used as input, while orf29 mRNA was readily detected in 1,000 TPA-treated cells. This corroborates earlier data (5, 10), which showed orf29 to be a true gamma-2 transcript. We calculated the induction to be 500-fold (calculated as exp[C_t(TPA) − C_t(mock)], where C_t(TPA) and C_t(mock) are C_t values for TPA- and mock-treated cells, respectively). Previous reports showed that about 2% of BCBL-1 cells undergo spontaneous lytic reactivation (63). Consistent with this observation, 10⁶ mock-treated BCBL-1 cells are needed to give a C_t value equivalent to 10⁴ TPA-induced cells in the orf29, orf50, and vGCR assay (Fig. 1C). Note that a quantitative comparison is only valid when the same primer pair is used but not between different probes. This is because the amplification efficiency may be different for each primer pair. In sum, these experiments represent the first quantitative analysis of LANA mRNA expression and introduce a new assay to classify KSHV mRNAs. Important to our studies, it quantitatively establishes K14/vGCR as a lytic (TPA-inducible) and LANA/v-cyclin as a latent (resistant to TPA induction) message.

Fine mapping of LANAp.

We and others previously identified the transcription start site for the LANA (orf73) and k-cyclin (orf72) mRNAs (9, 47, 56). The major transcription start site as determined by nuclease protection analysis and 5′ rapid amplification of cDNA ends is at nt 127880, with less-prominent transcripts initiating at nt 127900 and 127948 (47). Since no consensus TATA box was evident 30 bp upstream of nt 127880 but at nt 127934, the hypotheses were proposed either that the LANAp is TATA independent (like EBV EBNA Qp [37, 49]) or that the start site represents premature termination of the reverse transcriptase reaction. To identify the minimal cis sequence necessary for basal promoter activity we undertook an extended deletion analysis. Figure 2 schematizes the structure of our deletion clones. Plasmids pDD124 and pDD125 contain the entire 5′ untranslated region (5′-UTR) from the LANA translation initiation site at nt 127296 fused to lacZ and extend out to position −1299 (nt 129179) and position −279, respectively (nucleotide positions refer to GenBank accession number U75698 [45]). Both plasmids exhibited promoter activity (data not shown). We therefore used our previously published plasmids (9) together with additional deletion mutants to define the minimal LANAp cis sequence. Plasmids pDD41 to pDD53 contain progressive deletions and use the Ncol site at position +271 (nt 127609) to drive a luciferase reporter gene. Luciferase assays were performed in triplicate, and transfection efficiency was controlled by cotransfection with a constitutive lacZ expression plasmid. Figure 3 depicts the result of our promoter analysis in 293, SLK, (10)1, and BJAB cells. In all cell lines tested, deletion clones up to plasmid pDD274 extending just to position −88 exhibited significant promoter activity. In contrast, pDD273 extending to position −55 had lost all ability to drive reporter gene transcription. Reporter pDD83, extending to position −273, showed maximal expression in the three fibroblast cell lines, while in BJAB cells we observed a gradual decrease in promoter activity proportional to the length of the upstream region. These experiments establish that cis elements between positions −55 (nt 127935) and −88 (nt 127968) are essential for minimal LANAp activity, whereas sequences up to position −273 are needed for robust expression in some instances.

To determine whether additional, perhaps KSHV-specific, sequence elements were located further upstream, we compared the promoter activity of pDD83 (position −273) to that of pDD44 (position −1490) in several additional cell lines: HeLa, LnCAP, or KSHV-positive BCBL-1 and BC-3 cells (Table 2). Independent of the cell line, pDD41 and pDD83 showed 5- to 20-fold higher activity than vector alone or pDD53. Plasmid pDD43, which contains the reporter in the opposite direction of pDD41, showed no activity. Overall, LANAp activity was significantly higher than that of a simian virus 40 minimal promoter (pGL3 promoter; Invitrogen). Except in BC-3 cells no significant differences could be discerned between LANAp extending to positions −279 (pDD83) and −1490 (pDD41), suggesting that pDD83 contains all the cis elements that are necessary for LANAp function. Therefore, we decided to use pDD83 in all subsequent experiments.

LANAp activity is independent of other KSHV functions.

Unlike other KSHV mRNAs, LANA message is not upregulated after induction of the KSHV lytic cycle by phorbol ester, butyrate, or gamma interferon (5, 9, 47, 48, 56). To determine whether the isolated LANAp maintained this property, we transfected 293 cells in the presence or absence of TPA. Figure 4A shows that the activity of the pDD83 promoter fragment remained the same regardless of the presence or absence of TPA. This is in contrast to the activity of an NF-κB reporter (Fig. 4D) which was dramatically induced by TPA. Figure 4B and C show negative controls and indicate the lack of discernible luciferase activity when the LANA 5′-UTR fragment (pDD168) or K14/vGCR promoter (pDD163) fused to luciferase was transfected. This establishes that LANAp is not affected by cellular signaling pathways, which induce KSHV immediate early proteins.

To verify that these results also apply to B cells, we performed a similar experiment in BJAB cells, which are negative for EBV and KSHV. Figure 5C shows fold induction of LANAp with increasing concentrations of TPA. LANAp activity did not change. In this experiment we also transfected the KSHV gB promoter (26) into BJAB cells as a control (Fig. 5D). Since the KSHV gB promoter is a true late promoter it was similarly not induced by TPA in this KSHV-negative cell line. Note that the activity of LANAp in the absence of TPA is 10-to 100-fold higher than that of gB in either cell line (data not shown). Hence, we calculated fold induction in either case.

What would happen in the presence of the virus and its complete set of immediate-early, early, and late proteins? To address this question, we transfected BCBL-1 cells. BCBL-1 cells harbor latent, episomal KSHV that can be induced to undergo lytic replication in response to TPA (42). This viral isolate is fully functional, since virions isolated from BCBL-1 cells are infectious in culture (11, 20, 29, 33, 41) as well as in vivo (10). TPA at 20 ng/ml represents the optimal dose for the induction of lytic KSHV replication (5, 42). The KSHV gB promoter exhibited a TPA dose-dependent increase in activity (Fig. 5B), demonstrating (i) that our TPA was biologically active and (ii) that KSHV transactivating functions are induced in BCBL-1 cells. In contrast, LANAp showed no significant increase in activity (Fig. 5A). Although suboptimal amounts of TPA upregulated the LANAp somewhat, no dose-response curve was obtained. The twofold induction by TPA might be the result of the well-documented increase in transfection efficiency in the presence of phorbol esters or butyrate (28). This establishes that in transient-transfection assays LANAp is indifferent to cellular or viral immediate-early, early, or late functions, mimicking the behavior of LANA mRNA.

Identification of K14/vGCR promoter.

The LANA mRNA +1 site is located within the K14 ORF, which is oriented in the opposite direction (Fig. 2). Recently, Kirshner et al. (19) showed that K14 and vGCR are contained on a bicistronic, spliced message that initiates within the LANA 5′-UTR at nt 127848. In other words, nt 127848 to 127880 are part of the LANA mRNA as well as—in antisense orientation—part of the K14/vGCR mRNA. The K14/vGCR mRNA, however, is an early transcript in KSHV and was never detected in latently infected BCBL-1 cells or KS tumors (19). To determine the minimal cis-acting region necessary to direct K14/vGCR transcription and regulation, we cloned the complete intervening region (nt 127296 to 127883) in both orientations upstream of a luciferase reporter. We then transfected 293 cells with the putative K14/vGCR 5′-proximal region (pDD163, vGCRp) containing the K14/vGCR start site at nt 127848 in the presence or absence of TPA. We did not detect any luciferase activity (Fig. 4B). In the same experiment transfection of a reporter under control of NF-κB response elements exhibited TPA-inducible activity, establishing the functionality of our assay (Fig. 4D). As expected, we failed to observe promoter activity when we transfected the construct containing the same region in the antisense orientation (pDD168, antisense vGCR, Fig. 4C). This evidences (i) that there is no cryptic latent promoter located in the LANA 5′-UTR downstream of +1 and (ii) that the 5′-proximal fragment of the K14/v-GCR promoter does not function in 293 cells, despite the presence of a predicted TATA box 30 bp upstream of +1.

To test the hypothesis that the K14/vGCR promoter might be tissue specific, we transfected NIH 3T3 cells (Fig. 6A) and BJAB cells (Fig. 6B) in the presence of TPA. The K14/vGCR promoter (pDD163) failed to exhibit any significant activity in either cell line compared to the antisense orientation control (pDD168) or vector (pGL3basic). This is in contrast to the human immunodeficiency virus (HIV) long terminal repeat (LTR), which served as a positive control. The result changed drastically when we assayed these constructs in KSHV-positive BCBL-1 cells (Fig. 6C). The putative K14/vGCR promoter fragment showed approximately 20-fold higher reporter activity compared to the antisense control or vector. In TPA-treated BCBL-1 cells this promoter was even stronger than the HIV LTR. This identifies a cis-regulatory region that is sufficient to function as a K14/vGCR promoter. Furthermore, it demonstrates that vGCR promoter activity is absolutely dependent on KSHV transactivating functions, which is consistent with the regulation of authentic K14/vGCR mRNA.

orf50 transactivates the K14/vGCR promoter.

Having established that, on the one hand, the LANAp fragment recapitulated the constitutive expression pattern previously observed for LANA mRNA and that, on the other hand, the K14/vGCR promoter was critically dependent on KSHV transactivation functions, we tried to identify transactivating factors specific for either promoter. First, we asked the question, is the K14/vGCR inducible by KSHV orf50? The immediate-early protein encoded by orf50 is the EBV R homolog of KSHV (25, 51, 55, 64). KSHV orf50 is necessary and sufficient to initiate complete lytic viral replication (15, 26, 58). We transfected BCBL-1 cells (Fig. 7A) with an orf50 expression plasmid and the K14/vGCR promoter directed toward the reporter gene (pDD163) or directed away from it (pDD168). orf50 induced expression from the K14/vGCR reporter more then 20-fold, whereas the same fragment located in the opposite direction had no activity (Fig. 7A). This implies that KSHV orf50 induces the vGCR promoter either directly or through other KSHV transactivating functions.

To determine whether orf50 alone was sufficient to activate the vGCR promoter we repeated the previous experiment in KSHV-negative cells. Shown are the results of one such experiment performed in (10)1 fibroblasts (Fig. 7); similar data were obtained in SLK cells (data not shown). Cotransfection of increasing amounts of orf50 expression plasmids increased K14/vGCR promoter activity in a dose-dependent manner (Fig. 7B). At maximum, 10-fold induction could be observed in the presence, compared to the absence, of orf50, while no effect was observed on the antisense orientation control plasmid (Fig. 7C). In the same experiment cotransfection of LANA had no effect (Fig. 7D and E). Since these experiments were performed in fibroblasts, which do not harbor KSHV, this establishes that orf50 alone is sufficient to activate the K14/vGCR promoter.

To determine whether orf50 activates the K14/vGCR promoter through interaction with the basal transcription machinery or whether the orf50 response is mediated through a separate cis site, we undertook a deletion analysis. The results are shown in Fig. 8. Deletion of the distal-most 97 bp (nt 127297 to 127394) completely abolished orf50 responsiveness, as did deletion of a larger (nt 127297 to 127619) part of the promoter. This maps the orf50 element to a 97-bp region at position −454 of the transcription start site for the K14/vGCR message.

Regulation of LANAp by LANA and p53.

Renne et al. present preliminary evidence that LANA positively autoregulates its own promoter, while at the same time inhibiting the HIV LTR (40a). We extend these experiments here, by narrowing the region for the presumed LANA responsive element. Figure 9 shows the results of transfecting LANA with the LANAp deletion constructs. Compared to vector alone, LANA increases transcription of plasmids pDD41 (position −1490) and pDD83 (position −279). In contrast, LANA did not significantly affect the minimal latent promoter pDD274 (position −88). As shown in Fig. 7D and E, LANA did not activate or repress the K14/vGCR promoter. This demonstrates promoter specificity for LANA transactivation and locates the LANA responsive element to a 200-nt region between position −88 and −279.

Recently, Friborg et al. showed that LANA could bind to the p53 tumor suppressor protein and could suppress the p53 transactivating function (12). Here we asked the converse question, namely, whether p53 could regulate the LANAp. Figure 10 shows that human p53 suppressed LANAp (position −1490 or −279) activity (Fig. 10A and B) compared to the vector control (Fig. 10D). This holds true to a lesser extent for the minimal promoter fragment pDD274 (position −89 [Fig. 10C]), suggesting that p53 exerts its effect on the basal transcriptional apparatus. Unfortunately, in (10)1 cells pDD274 basal activity is much lower than for the two larger promoter fragments. Consequently, fold repression in this case is marginal. These transfections were conducted in the established p53-negative (10)1 cell line (61). Following the protocol described in the initial reports on p53-mediated suppression (23, 27, 52), we chose to normalize for total protein concentration rather than rely on a cotransfected reporter. This was mainly for two reasons: (i) a wide range of promoters is suppressed or activated by p53, including the cytomegalovirus promoter-enhancer, which we used as transfection control in all other experiments; (ii) any additional promoter-enhancer could influence our results by binding cellular factors. We did not observe toxicity in transfected cells, and we verified the functionality of our p53 expression plasmid using a consensus p53 reporter construct (Fig. 10E). These experiments raise the possibility for p53's involvement in the regulation of KSHV latency.