Human papillomavirus type 16 E7 perturbs DREAM to promote cellular proliferation and mitotic gene expression

Abstract

Study of the small DNA tumor viruses continues to provide valuable new insights into oncogenesis and fundamental biological processes. While much has already been revealed about how the human papillomaviruses (HPVs) can transform cells and contribute to cervical and oropharyngeal cancer, there clearly is much more to learn. In this issue of Oncogene, Pang et al. demonstrate that the high-risk HPV16 E7 oncogene can promote cellular proliferation by interacting with the DREAM (DP, RB-like, E2F and MuvB) complex at two distinct phases of the cell cycle (1). Consistent with earlier work, HPV16 E7 can bind to the retinoblastoma tumor suppressor (RB) family member p130 (RBL2) protein and promote its proteasome-mediated destruction thereby disrupting the DREAM complex and prevent exit from the cell cycle into quiescence. In addition, they demonstrate that HPV16 E7 can bind to MuvB core complex in association with BMYB and FOXM1 and activate gene expression during the G2 and M phase of the cell cycle. Thus, HPV16 E7 acts to prevent exit from the cell cycle entry and promotes mitotic proliferation and may account for the high levels of FOXM1 often observed in poor risk cervical cancers.

HPV E7 is one of the two viral oncogenes (the other, E6, is not discussed here) that are expressed in nearly all cervical cancers, a variety of other anogenital tract carcinomas, and in a large fraction of squamous cell carcinomas of the oropharynx. While the human papillomavirus family contains more than 200 genotypes, only a small number have been identified as high-risk in cervical cancer. HPV16 is the most prevalent high-risk HPV and is found in approximately 50% of cervical carcinoma. HPV16 E7 is a relatively small protein of 98 residues that does not contain any intrinsic enzymatic properties. Instead, nearly all of E7’s transforming activity is mediated through direct interactions with cellular proteins (recently reviewed in (2)). Among the first identified cellular host targets of the high risk HPV16 E7 were RB (RB1) and the RB-related proteins p130 (RBL2) and p107 (RBL1), known collectively as the pocket proteins. HPV16 E7 contains a short stretch of residues known as the LXCXE motif. HPV16 E7 shares the canonical LXCXE motif with other DNA tumor viruses including polyomavirus large T antigen and adenovirus E1A. These viral oncogenes use the LXCXE motif to bind directly to all three RB family members. In addition to the LXCXE motif, E7 employs residues in its N-terminus to promote the proteasome-mediated destruction of the RB proteins. Mutation of residues 6–10 retains HPV16 E7 binding to pocket proteins but reduces its ability to promote their destruction. Proteasome mediated destruction of the pocket proteins is likely mediated by recruitment of the CUL2 E3 ubiquitin ligase and RB pocket proteins to the same HPV16 E7 molecule (3). The Thierry group uses these mutant forms to tease apart the specific effects of HPV16 E7 on the DREAM complex (1).

The DREAM complex contains the MuvB group of 5 proteins, LIN9, LIN37, LIN52, LIN54 and RBBP4, that remain together throughout the cell cycle (4). Components of the DREAM complex were originally identified in genetic experiments in C. elegans with a multi-vulvar phenotype and have also been identified in Drosophila and vertebrates. Within the MuvB core, the LIN54 protein contains a CXC domain that likely serves to bind to a specific DNA sequence called CHR or cell cycle genes homology region, present in the promoters of many genes with expression peaking during the G1/S phase or G2/M phase of the cell cycle (5, 6). Phosphorylation of LIN52, the smallest subunit, is required for the MuvB core to bind specifically to p130 (7). No specific functions for LIN9, LIN37 or RBBP4 have been determined. When the MuvB core binds to p130, DP1 and E2F4, it forms the DREAM complex that binds to the promoter of hundreds of E2F-regulated genes during the G0 or quiescent phase of the cell cycle (Figure 1A). When a quiescent cell is stimulated to enter the cell cycle, p130 and perhaps other components of the DREAM complex are phosphorylated by Cyclin D/CDK4 as well as Cyclin E/CDK2 resulting in release of p130, E2F4 and DP1 from the MuvB core. Release of p130, E2F4 and DP1 from the MuvB core permits expression of genes required for DNA synthesis including DNA polymerase (POLA2), thymidine kinase (TK1) and MCM6 as well as the activating E2Fs. The activating E2Fs, E2F1, E2F2 and E2F3, also drive expression of the transcription factors BMYB (MYBL2) and FOXM1. During S phase, the MuvB core complex recruits BMYB to the promoters of genes that are expressed during the G2/M phase. The BMYB-MuvB complex sequentially recruits FOXM1 that is required for transactivation of genes required for mitosis such as aurora kinase B (AURKB), polo-like kinase 4 (PLK4) and cyclin B1 (CCNB1) (Figure 1B). Therefore, the MuvB core serves two independent roles in the cell cycle: binding to p130 during quiescence to repress most if not all cell cycle regulated genes and binding to BMYB and FOXM1 to promote expression of genes during the G2/M phase of the cell cycle. Notably, RB1 itself does not bind to the MuvB complex or form the DREAM complex. Instead, RB1 binds to the activating E2Fs, represses gene expression during the G1/S transition, and establishes a checkpoint preventing cell cycle entry and progression independent of the DREAM complex. Importantly, HPV16 E7 targets all 3 RB family members for proteasome mediated destruction thereby disrupting the p130-containing DREAM complex in G0/G1 as well as disrupting RB1’s ability to repress the activating E2F during G1/S phase (8, 9). Targeted expression of HPV16 E7 can contribute to cervical cancer development in a mouse model (10). At least part of this activity is due to functional inactivation of all 3 RB family members. However, it was recently reported that combined loss of RB1, p107 and p130 resulted in development of high-grade pre-malignant lesions but was not sufficient to cause formation of cervical carcinomas (10). This result indicates that the ability of HPV16 E7 to cause cancer likely involves more than disrupting the RB family proteins. Part of E7’s transforming activity could be specific activation of the BMYB-MuvB-FOXM1 transcription factor complex.

Figure 1. HPV16 E7 targets the DREAM complex twice.

A. The DREAM complex is composed of the RB-related protein p130 bound to the repressor E2F4 and DP1 in complex with a group of 5 proteins referred to as the MuvB core. MuvB contains LIN9, LIN37, LIN52, LIN54 and RBBP4. The DREAM complex binds and represses the promoters of genes with peak expression during either the G1/S phase or G2/M phase during cellular quiescence. HPV16 E7 binds directly to p130 and promotes its destruction by the proteasome thereby disrupting the DREAM complex.

B. During S and G2 phase of the cell cycle, the MuvB core sequentially recruits BMYB and then FOXM1 to the promoters of genes with peak expression during G2/M phase. HPV16 E7 can bind to this complex and promote expression of the G2/M phase genes.

Pang et al. tested the ability of HPV16 E7 to independently disrupt the DREAM complex and RB1 by using a combination of HPV16 E7 mutants and short hairpin RNA interference (shRNA) constructs that target all 3 pocket proteins (1). They report that while both the LXCXE and N-terminal motifs of HPV16 E7 were required to bind and inactivate the pocket proteins, only the LXCXE motif was required to promote mitotic gene expression in the absence of all 3 pocket proteins (Figure 1A). Remarkably, Pang et al. provide evidence that HPV16 E7 could bind to the BMYB-MuvB-FOXM1 complex in the absence of pocket proteins (Figure 1B) (1). Among the many earlier reports of HPV16 E7 binding proteins, it had been reported that HPV16 E7 could bind to FOXM1 (then known as M-phase phosphoprotein 2 (MPP2) (11). The Thierry group significantly extends this earlier report and shows that HPV16 E7 can also bind to BMYB and MuvB. Notably, they show that HPV16 E7 can bind to the BMYB-MuvB-FOXM1 promoters using chromatin immunoprecipitation and increase expression of genes regulated by this complex. Therefore, HPV16 E7 perturbs the function of the DREAM complexes during at least 2 distinct phases of the cell cycle.

The multiple activities of HPV16 E7 on cell cycle regulated gene expression serve to illustrate the often observed and apparently redundant activities of the small DNA tumor virus oncogenes. While degradation of RB as well as the RB-related proteins should remove all repression of E2F dependent gene expression, HPV16 E7 binding to the BMYB-FOXM1-MuvB complex seems to add an extra kick to the late cell cycle gene expression. Overexpression of FOXM1 has been reported in cervical cancer and associated with poor prognosis (12, 13). This raises the interesting question of whether HPV16 E7 directly contributes to the poor prognostic features of FOXM1 expression and whether specific downstream targets of FOXM1 are being activated by E7.

It is not clear where and how HPV16 E7 binds to the BMYB-FOXM1-MuvB complex. While the LXCXE motif is required, it is not clear what component(s) HPV16 E7 binds. An earlier report found that HPV16 E7 bound to FOXM1 in a yeast 2-hybrid assay suggesting the possibility of a direct interaction (11). If so, then HPV16 E7 could bind indirectly to BMYB and MuvB though FOXM1. It should be noted that the BMYB-MuvB complex does not contain p130 or any other pocket protein; BMYB binds to the MuvB core in a mutually exclusive pattern with p130 (14). Therefore, the LXCXE motif requirement for binding to BMYB-FOXM1-MuvB is not due to p130 or other RB pocket family member remaining as part of this complex. It is also not clear how HPV16 E7 can seemingly increase the activity of the BMYB-MuvB-FOXM1 complex resulting in higher levels of gene expression during G2/M phase. Perhaps HPV16 E7 recruits additional cellular proteins to FOXM1-MuvB-BMYB to increase the overall transcriptional activity. Alternatively, HPV16 E7 could perturb an auto-inhibitory activity within FOXM1 (15).

It should be noted that several features of the DREAM complex remain poorly understood. It is not known how the MuvB core actively contributes to repression of E2F dependent genes during quiescence or to the activation of genes bound by BMYB and FOXM1. Does the MuvB core bind to additional gene regulators or chromatin modifying factors to perturb the function of the p130-E2F4-DP, BMYB and FOXM1 containing complexes? Perhaps continued study of the interaction HPV16 E7 and the DREAM complex will provide additional insight into how this complex is regulated throughout the cell cycle and its role in cellular transformation and oncogenesis.