The papillomaviruses are epitheliotropic small DNA tumor viruses that cause both benign and malignant lesions in humans and animals. These viruses productively infect only squamous epithelia and progression of the viral life cycle is dependent on differentiation of the host cell, the keratinocyte. An oncogenic subset of the human papillomaviruses (including HPV-16, 18, 31, and 33) is associated with the vast majority of cervical cancers as well as squamous cell carcinomas in other locations. The E6 and E7 viral oncogenes interfere with the functions of p53 and pRb, respectively, and thus disrupt regulation of the cell cycle. The role of a third viral oncogene, E5, in the viral life cycle is less well understood. The L1 and L2 ORFs encode the major and minor capsid proteins, respectively. Expression of papillomavirus genes is regulated at both transcriptional and posttranscriptional levels. In the process of studying the post-transcriptional regulation of papillomavirus gene expression, we have developed considerable expertise in RNA processing and in the design and execution of isoform-specific real time QRT-PCR assays. We are continuing to apply this expertise to the analysis of viral and cellular mRNA levels and alternative splicing in collaborations with other investigators. Some of these collaborations are described below. In addition, we are also exploiting HPV alternative splicing to develop novel RNA-based antiviral and anticancer therapies. Post-transcriptional regulation of the papillomavirus L1 and L2 capsid proteins: The late transcription unit of the bovine papillomavirus (BPV-1), including both early and late poly(A) sites, was inserted into a replication defective adenoviral vector to aid in the study of BPV post-transcriptional regulation in keratinocytes. This recombinant vector produces high levels of L1 and L2 mRNAs in 293 cells which complement the replication defect of the adenoviral vector. Interestingly, there is an early to late switch in BPV-1 splice site and poly(A) site usage in concert with the early to late switch of the adenovirus life cycle. Surprisingly, no L1 or L2 protein could be detected in infected cells at 48 hrs post infection. The BPV-1 L1 and L2 mRNAs have a noncoding first exon with four out of frame AUG translation initiation codons. Short upstream open reading frames (uORFs) have been shown to regulate translation initiation in some systems. We therefore mutated the four late leader AUGs in the context of the adenoviral vector and demonstrated that these mutations are sufficient to allow L1 and L2 protein expression. These studies indicate that PV L1 and L2 capsid protein expression is regulated at many post-transcriptional levels including translation initiation. Targeting HPV infected cells using specific trans-splicing: The ideal cancer therapy would kill cancer cells and have no effect on normal cells. The expression of spliced human papillomavirus E6/E7 pre-mRNAs by the vast majority of cervical cancers provides an absolute difference between cancer cells and normal cells that can be exploited for the specific targeting of these cells. Through a CRADA between NCI and Intronn Inc., we are investigating the use of Intronn's Spliceosome Mediated RNA Trans-splicing (SMaRT) Technology to develop a novel RNA-based suicide gene therapy for cervical cancer. SMaRT is a platform technology that has the ability to reprogram genetic information at the mRNA level. SMaRT involves the trans-splicing of a pre-trans-splicing molecule (PTM) to a target pre-mRNA. PTMs are RNAs that are designed to contain an antisense binding domain (AsBD) complementary to a target pre-mRNA and 3' splice elements necessary for efficient and specific trans-splicing. Target specificity is obtained by tethering the PTM to the target pre-mRNA through the AsBD in the PTM. SMaRT can be used to replace or repair 5' exon(s), internal exon(s), and 3' terminal exon(s) and to make chimeric mRNAs. We have previously designed PTMs that target HPV-16 E6 and E7 pre-mRNAs and have used these to study the factors that are important for trans-splicing efficacy. Trans-splicing to both exogenous and endogenous target pre-mRNAs is efficient as long as delivery is efficient and the PTM is highly expressed. We have shown that the amount of trans-spliced product is directly proportional to the amount of target over at least a 300-fold range. This is an important property of SMaRT that is directly relevant to our goal of using this technology for a cancer therapy. We have also confirmed that the AsBD confers both target specificity as well as splice site specificity for a target with multiple exons. Over the past two years we have put considerable effort into the optimization of 5'RACE protocols and used these to assess trans-splicing specificity. Although SMaRT shows considerable target specificity, the vast excess of all other cellular pre-mRNAs over the specific target pre-mRNA leads to significant amounts of non-specific trans-spliced product. We have made a number of changes to the PTM expression vectors, including the splice site region, and assessed the affect of these changes on trans-splicing efficiency and specificity. These studies have improved our ability to rationally design optimal PTM vectors. Finally we are currently designing new combinatorial strategies to improve the overall specificity of a SMaRT-based cancer therapy. Establishment of papillomavirus infection is enhanced by promyelocytic leukemia protein (PML) expression:In a collaboration with the Schiller and Lowy labs at the NCI, we have used our expertise in splice-specific real time QRT-PCR to demonstrate that the PML protein is required for efficient expression of spliced papillomavirus mRNAs at early times after infection. Previous studies have suggested that most papillomaviruses enter the host cell via clathrin-dependent receptor-mediated endocytosis but have not addressed later steps in viral entry. To examine these events, the localization of L2 and packaged DNA was followed after entry of infectious virions or L1/L2 pseudovirions. Confocal microscopic analyses of HeLa cells showed a time-dependent uncoating of capsids in cytoplasmic vesicles and the accumulation of both L2 and viral DNA at distinct nuclear domains identified as nuclear domain 10 (ND10). Both L2 and the pseudogenome had a punctate distribution and localized to ND10 in promyelocytic leukemia protein (PML)-expressing cells, whereas L2 had a diffuse nuclear distribution in PML-/- cells. The number of pseudovirus-infected cells was an order of magnitude higher in the PML+ cells compared with the PML-/- cells, and viral genome transcription after infection with authentic bovine papillomavirus virions was similarly elevated in PML+ cells. The results identify a role for PML in the enhancement of viral infectivity in the early part of the life cycle. These studies suggest a model in which L2 chaperones the viral genome to ND10 to efficiently initiate viral transcription. cis-Acting and trans-acting modulation of equine infectious anemia virus alternative RNA splicing:In collaboration with the Derse lab at the NCI, we helped design splice-specific real time QRT-PCR assays that were used to study the regulation of equine infectious anemia virus (EIAV) alternative splicing. EIAV, a lentivirus distantly related to HIV-1, encodes regulatory proteins, EIAV Tat (ETat) and Rev (ERev), from a four-exon mRNA. Exon 3 of the tat/rev mRNA contains a 30-nucleotide purine-rich element (PRE) which binds both ERev and SF2/ASF, a member of the SR family of RNA splicing factors. To better understand the role of this element in the regulation of EIAV pre-mRNA splicing, we quantified the effects of mutation or deletion of the PRE on exon 3 splicing in vitro and on alternative splicing in vivo. We also determined the branch point elements upstream of exons 3 and 4
