VprBP/ DCAF1 is overexpressed and catalyzes H2AT120p in melanoma cells
Because VprBP expression is often dysregulated in cancer cells [7, 8], we reasoned that aberrant expression of VprBP could also be observed in melanoma cells. To explore this possibility, cell lysates were prepared from four melanoma (G361, MeWo, SK-MEL-5, and A375) and one melanocyte (NHEM2) cell lines and analyzed by Western blotting with VprBP antibody. Our analysis detected VprBP expression at much higher levels in the melanoma cells compared to the melanocyte cells (Fig. 1a). In checking whether the observed VprBP overexpression is associated with altered H2AT120p, we observed much higher levels of H2AT120p in chromatin fractions extracted from the melanoma cells. Further, stable depletion of VprBP with shRNA knockdown almost completely eliminated H2AT120p, and these changes could be rescued by ectopic expression of VprBP wild type but not VprBP K194R kinase-dead mutant in G361 and MeWo melanoma cells, as assessed by immunostaining and Western blot analyses (Fig. 1b, c and Additional file 1: Fig. S1a, b). As another approach to check a possible link between VprBP and H2AT120p in melanoma cells, we carried out immunostaining with melanoma and adjacent normal tissue samples. As shown in Fig. 1d, our immunostaining demonstrated that VprBP protein level was significantly elevated in melanoma patient tissue samples, and its expression was highly correlated with H2AT120p. Adding support for this observation, our analysis of VprBP expression levels in five stages of melanoma detected VprBP overexpression in all five stages of melanoma with only minor variations (Additional file 2: Fig. S2).
We previously screened a library of 5000 compounds and identified B32B3 as a selective inhibitor targeting VprBP catalytic domain and thus blocking VprBP kinase activity [7]. Given the demonstrated reliance of H2AT120p on VprBP in melanoma cells, it was reasonable to expect that B32B3 treatment will recapitulate the effects of VprBP knockdown. Toward this end, we treated G361 and MeWo cells with increasing concentrations of B32B3, and evaluated changes in H2AT120p by Western blot. The range of B32B3 concentrations used in these initial assays is based on our recent studies to determine its IC50 values using colon and prostate cancer cell lines [7, 8]. When the melanoma cells were exposed to six different concentrations (0, 0.03, 0.1, 0.3, 1, and 3 µM) of B32B3, B32B3 was able to potently block H2AT120p with a half-maximal inhibitory concentration (IC50) of 0.1 µM (Fig. 1e and Additional file 1: Fig. S1c). Consistent with these results, our immunostaining of G361 and MeWo cells after B32B3 treatment at the IC50 concentration detected a significant reduction in H2AT120p levels (Fig. 1f and Additional file 1: Fig. S1d). Together, these initial observations link VprBP overexpression to H2AT120p in melanoma cells and rationalize further studies on their possible contributions to melanomagenesis.
VprBP/DCAF1 knockdown and inhibition suppress melanoma cell growth
The data presented above confirmed that VprBP is overexpressed and is responsible for H2AT120p event in melanoma cells; however a potential significance of VprBP-mediated H2AT120p with respect to melanoma cell growth remains unclear. In an effort to address this question, we monitored changes in cell growth rates in response to VprBP depletion daily over a period of five days by MTT assays. As summarized in Fig. 2a and Additional file 3: Fig. S3a, our assays revealed an apparent decrease in cell growth rate after stable knockdown of VprBP in G361 and MeWo melanoma cells. To examine whether VprBP capacity to mediate H2AT120p is necessary for the observed effects, we also conducted rescue experiments. The expression of VprBP wild-type restored the original growth rate of VprBP-depleted melanoma cells. On the other hand, VprBP kinase dead mutant failed to recover the growth capacity of VprBP-depleted G361 and MeWo cells (Fig. 2a and Additional file 3: Fig. S3a), underscoring the notion that VprBP-mediated H2AT120p is critical for VprBP function in promoting melanoma cell growth. For the purpose of gaining support for the MTT assay results, we also evaluated the impact of VprBP knockdown on melanoma cell growth by colony formation assays. The results from these experiments indicate that VprBP depletion adversely affects the potential of melanoma cells to grow into a colony when evaluated after 14 days of culture (Fig. 2b and Additional file 3: Fig. S3b). Since VprBP-mediated H2AT120p displays a sharp reduction in G361 and MeWo melanoma cells treated with VprBP inhibitor B32B3 (Fig. 2c and Additional file 3: Fig. S3c), we wondered whether the growth of melanoma cells is also suppressed by such B32B3 treatments. In initially exploring this possibility, we discovered that exposing G361 and MeWo melanoma cells to B32B3 caused a marked impediment to the growth ability of the cells (Fig. 2c and Additional file 3: Fig. S3c). In performing colonogenic assays with G361 and MeWo cells, we also observed the colony forming capacity of the cells to reduce by up to 40% after B32B3 treatment (Fig. 2d and Additional file 3: Fig. S3d). These results underscore the direct action of VprBP on melanomagenesis and support the concept that VprBP exerts its melanomagenic function via H2AT120p.
VprBP/DCAF1-mediated H2AT120p inactivates growth regulatory genes in melanoma cells
In order to examine whether VprBP-mediated H2AT120p described above plays any role in regulating melanomagenic transcription program, genome-wide RNA sequencing (RNA-seq) analysis was performed with total RNA isolated from control and VprBP-depleted G361 cells. In the principal component analysis (PCA) of RNA-seq data, samples for each group were found to be markedly separated from each other, but close clustering of replicates from groups indicated minimal variability in the quality of analyzed replicates (Fig. 3a). Using a 2-fold cutoff, our transcriptome profiling revealed a total of 1941 genes differentially expressed upon stable knockdown of VprBP in melanoma cells (Fig. 3b, c). Among those genes, 674 genes were downregulated and 1267 genes upregulated in response to VprBP knockdown (Fig. 3c and Additional file 4: Fig. S4). Consistent with our previous publications implicating VprBP-mediated H2AT120p in oncogenic gene silencing, gene ontology analysis of 1267 upregulated targets also identified cell growth and proliferation as the most activated biological pathways in VprBP-depleted melanoma cells (Fig. 3d). Our Gene Ontology (GO) and Reactome pathway enrichment analyses of the genes that were upregulated upon VprBP knockdown revealed a significant enrichment of metabolic process genes (Additional file 5: Fig. S5, S6 and Additional information 2). The functional enrichment analysis using a list of significantly enriched GO terms also links VprBP to metabolic processes (Additional file 6: Fig. S6). Given the well-established importance of metabolic processes in melanoma development and progression [27], these findings strongly implicate VprBP in the pathogenesis of melanoma. The role for VprBP in melanomagenesis was further supported by the fact that our analysis of the leading-edge subset in the gene set detected 20 genes encoding negative regulators of melanoma growth and proliferation (Fig. 3e). To validate our RNA-seq data, we conducted reverse transcription quantitative PCR (RT-qPCR) analysis of 8 target genes that encode factors acting as tumor suppressors in several types of cancers including melanoma (Fig. 4a and Additional file 7: Fig. S7a). Our analysis with total RNA from G361 and MeWo cells demonstrated the up- and down-regulation of the selected targets after VprBP knockdown and rescue expression (Fig. 4a and Additional file 7: Fig. S7a), respectively. Importantly, if VprBP K194R kinase-dead mutant is expressed in VprBP-depleted cells, target genes were still expressed at high levels, underscoring the importance of VprBP-mediated H2AT120p for target gene inactivation in melanoma cells. Similar RT-qPCR assays after treatment with VprBP inhibitor B32B3 also detected the active state of target genes in G361 and MeWo cells (Fig. 4c and Additional file 7: Fig. S7c), results of the incapability of VprBP to induce H2AT120p and thus target gene depotentiation.
To check whether the observed function of VprBP in suppressing target gene transcription reflects its stable recruitment, we next investigated its occupancy at INPP5J, ZNF750, and TUSC1 genes by chromatin immunoprecipitation (ChIP) assays. Crosslinked chromatin was isolated from control and VprBP-depleted G361 and MeWo cells, and the precipitated DNA was amplified by quantitative real-time PCR (qPCR) using primer sets specific for promoters, transcription start sites, and coding regions of INPP5J, ZNF750, and TUSC1 genes. Consistent with our previous observation [7, 8], VprBP ChIP signals were much more enriched at the promoter region than at transcription start site and coding region in mock-depleted control melanoma cells (Fig. 4b and Additional file 7: Fig. S7b). This result is supportive of the idea that VprBP targets the process of initiating transcription for its repressive action. VprBP distribution patterns across the target genes were perfectly correlated with H2AT120p enrichment patterns, suggesting a major role for VprBP in mediating H2AT120p at target genes. Indeed, VprBP occupancy of the target genes were significantly reduced after VprBP knockdown, and such changes also diminished the levels of H2AT120p. It was also apparent in our parallel ChIP-qPCR assays that ectopic expression of VprBP wild type, but not VprBP K194R kinase-dead mutant, in VprBP-depleted cells restored H2AT120p to levels quantitatively similar to those observed with mock-depleted control cells (Fig. 4b and Additional file 6: Fig. S6b). These observations were further corroborated by additional ChIP- and RT-qPCR experiments in which exposure of G361 and MeWo melanoma cells to VprBP inhibitor B32B3 almost completely crippled H2AT120p and triggered target gene reactivation (Fig. 4d and Additional file 7: Fig. S 7d). Taken together, these data argue persuasively that VprBP-mediated H2AT120p is a necessary step to downregulate growth-regulatory genes and to drive melanomagenesis.
Artificial tethering of VprBP/DCAF1 to target genes drives H2AT120p-induced transrepression
Through RNA-seq and ChIP/RT-qPCR studies described above, we identified a group of growth regulatory genes inactivated by VprBP and enriched for H2AT120p in melanoma cells. These results support the conjecture that H2AT120p mainly contributes to VprBP-induced transcriptional suppression leading to melanomagenesis. However, these studies do not exclude the possibility that VprBP generates the inactive state of target genes through some other mechanisms. If VprBP exerts its suppressive effects mainly by catalyzing H2AT120p at target genes, we can predict that artificially tethering VprBP to target genes is sufficient to re-establish an inactive state of transcription in VprBP-depleted melanoma cells. In exploring this possibility, we chose to use CRISPR/dCas9-based system for directing VprBP-mediated H2AT120p to target genes [28–31]. For this cellular manipulation of H2AT120p, we constructed a series of pPlatTET-gRNA2 all-in one vectors expressing dCas9-VprBP wild type (wt) or kinase dead mutant (K194R) and single guide RNAs (sgRNAs) recognizing the promoters or coding regions of INPP5J, ZNF750, and TUSC1 genes (Fig. 5a). We then tested the impact of these dCas9-VprBP fusions on the transcription of INPP5J, ZNF750, and TUSC1 genes in G361 melanoma cells. As summarized in Additional file 8: Fig. S8, individual expression of promoter-binding sgRNA 1 and 2 together with dCas9-VprBP wild type in VprBP-depleted G361 cells led to a detectable inactivation of target gene transcription, whereas coding region-binding sgRNA 3 or 4 generated no obvious changes in transcription. Also, by directing dCas9-VprBP to the promoter region with sgRNA 1 and 2 pair, we were able to generate a more pronounced repression of INPP5J, ZNF750, and TUSC1 genes (Additional file 8: Fig. S8a, b). Under identical assay conditions, sgRNA 3 and 4 pair minimally altered the levels of target gene transcription, thus implying that transcription initiation is the step mainly regulated by VprBP-mediated H2AT120p (Additional file 8: Fig. S8b). Consistent with expectations from these results, targeting dCas9-VprBP to both promoter and proximal coding regions by sgRNA 1, 2, 3 and 4 together established target gene silencing at levels comparable with those observed with sgRNA 1 and 2 pair (Additional file 8: Fig. S8b). In additional experiments, B32B3 treatment significantly impaired the suppressive activity of dCas9-VprBP wild type at INPP5J, ZNF750, and TUSC1 genes and dCas9-VprBP K194R kinase-dead mutant failed to exert any effects on transcription (Fig. 5b), clearly indicating the requirement of VprBP kinase activity for target gene silencing in melanoma cells.
To gain support for the transcription results, we next investigated dCas9-VprBP localization and its impact on H2AT120p in the promoter and coding regions of target genes by ChIP-qPCR analysis. The expression of dCas9-VprBP wild type or K194R kinase-dead mutant with sgRNA 1 and 2 pair generated a specific accumulation of the dCas9 fusion proteins in the promoter regions of the INPP5J, ZNF750, and TUSC1 genes (Fig. 5c). When sgRNA 1 and 2 were replaced by coding region-binding sgRNA 3 and 4, the dCas9-VprBP fusions were mainly localized in the proximal coding regions. In monitoring the extent of H2AT120p, we found that H2AT120p levels were increased in the promoter regions in VprBP-depleted G361 cells expressing dCas9-VprBP wild type together with sgRNA 1 and 2 pair (Fig. 5c). Parallel analysis on the proximal coding regions repeatedly demonstrated efficient accumulation of H2AT120p in the cells transfected with dCas9-VprBP wild type and sgRNA 3 and 4 (Additional file 9: Fig. S9a). As another evidence supporting the precise targeting of dCas9-VprBP fusions, H2AT120p was established at both promoter and coding regions after simultaneously transfecting dCas9-VprBP wild type and all four sgRNAs into the cells (Additional file 9: Fig. S9b). However, in agreement with transcription data, all the sgRNAs failed to increase H2AT120p levels at the target genes in dCas9-VprBP K194R-expressing cells as well as B32B3-treated cells.
Given that the INPP5J, ZNF750, and TUSC1 genes encode the components of growth control system, we subsequently tested whether delivering dCas9-VprBP with sgRNA 1 and 2 to their promoter regions can lead to changes in cell growth rate. In our MTT and colony formation assays, a marked increase in cell growth rates and colony numbers was evident, when dCas9- VprBP wild type was co-expressed with promoter-binding sgRNA 1 and 2 in VprBP-depleted G361 cells (Fig. 5d, e). Conversely, co-expression of dCas9-VprBP K194R kinase-dead mutant with sgRNA 1 and 2 failed to trigger a similar augmentation of growth and colony formation potential of VprBP-depleted G361 cells. Since B32B3 treatment also almost completely abolished the ability of dCas9-VprBP wild type to repotentiate the growth activity of sgRNA 1 and 2-transfected G361 cells (Fig. 5d, e), these results strongly suggest that VprBP can function to stimulate the growth of melanoma cells in a kinase activity-dependent manner. Together with ChIP-qPCR data above, these results also discount the possibility that VprBP drives melanomagenic gene silencing program in an H2AT120p-independent manner, and reinforce the conclusion that VprBP can accurately establish H2AT120p-induced gene silencing and growth stimulatory effects if stably recruited to the target genes.
VprBP/DCAF1 promotes melanoma tumorigenesis in its kinase activity-dependent manner
Based on our demonstration of VprBP being overexpressed and stimulating melanoma cell growth, an important question is whether VprBP knockdown or inhibition exhibits anti-melanomagenic efficacy through blocking VprBP kinase activity toward H2AT120 and thus decreasing VprBP transrepression potential. As a way to address this question, we decided to use xenograft mouse models derived from the G361 melanoma cell line. Accordingly, we subcutaneously injected 1x107 mock-depleted control or VprBP-depleted G361 cells into the right hind leg of nude mice, and monitored their growth at 3 days intervals for a period of 24 days. From this first set of experiments, we found that VprBP depletion significantly inhibits the growth of G361 melanoma xenografts when compared with mock-depleted control G361 xenografts (Fig. 6a-c). Moreover, the expression of VprBP wild-type, but not VprBP K194R mutant, in VprBP-depleted G361 xenografts resulted in a full recovery of the original xenograft growth rate (Fig. 6a-c), indicative a major role for VprBP-mediated H2AT120p in stimulating melanoma development and progression. To support the knockdown data, we also tested whether treatment with increasing concentrations (0.6, 1.2, 2.5, 5 and 10 mg/kg) of VprBP inhibitor B32B3 would influence the growth of G361 xenografts. As can be seen in Fig. 7a-c, the proliferative capacity of G361 melanoma xenografts was reduced by an average of 70% following the administration of 2.5 mg/kg B32B3 every three days for 24 days. At two higher doses of B32B3 (5 and 10 mg/kg), a greater impairment of melanoma growth compared with a dose of 2.5 mg/kg B32B3 was not observed (Fig. 7a-c). Also, that VprBP knockdown and B32B3 treatment were well tolerated without any significant changes in body weight (Additional file 10: Fig. S10) argues strongly that their inhibitory effects are generated by specifically targeting G361 melanoma cells in mice.
Another key question arising from the above-noted xenograft growth data is to what extent VprBP knockdown and inhibition affect H2AT120p in G361 melanoma xenograft models. In an attempt to address this question, we sacrificed mice, harvested melanoma xenograft tumors, and prepared xenograft lysates. Our Western blot analysis detected high levels of H2AT120p in lysates collected from control G361 melanoma xenografts (Fig. 6e, 7e). However, H2AT120p was almost completely disappeared upon shRNA-mediated knockdown and B32B3-induced inhibition of VprBP (Fig. 6e, 7e), again suggesting that VprBP stimulates the growth of G361 xenograft in a manner dependent on its kinase activity toward H2AT120. Considering the importance of H2AT120p in VprBP-driven transcriptional silencing of growth regulatory genes, we also examined whether VprBP depletion and inhibition-induced blocking of VprBP kinase activity also affect target gene expression in G361 xenografts. We found that VprBP knockdown generated, albeit to a somewhat varying extent, an active state of target gene expression at the level of transcription in G361 melanoma xenografts (Fig. 6d). Consistent with melanoma-promoting effects of VprBP-mediated H2AT120p, treating G361 melanoma xenograft with VprBP inhibitor B32B3 at doses of 2.5 mg/kg or higher over a period of 24 days also markedly increased target gene mRNA levels (Fig. 6d, 7d). B32B3 thus recapitulates anti-melanomagenic effects of VprBP knockdown and represent a potent molecular tool to negate VprBP-induced melanoma development and proliferation.