br Materials and methods br Results
Materials and methods
Discussion We recently detected phosphorylation of tyrosine Y102 in BPV-1 E2 and reported that FGFR3 binds to E2 and limits E1 dependent viral DNA replication (Culleton et al., 2017, Xie et al., 2017). In the current study, we addressed the potential role of other FGFRs in regulating the replication functions of bovine and human PV E2. We observed the interaction between FGFR2 and E2 by co-immunoprecipitation and PLA and demonstrated that FGFR2 phosphorylates tyrosine residues in BPV E2 that results in decreased PV DNA replication. To identify which FGFR may interact with HPV E2, these were Myc-epitope tagged and co-transfected along with HPV-16 and HPV-31 E2. While HPV-16 and HPV-31 E2 co-ip’d all the Myc-FGFRs, FGFR2 and FGFR4 were strongest respectively, yet only endogenous FGFR2 could be co-ip’d with HPV-16 E2. Endogenous FGFR-3 also complexes with HPV-16 E2 (Xie et al., 2017). Further, over-expression of FGFR2 suppressed BPV-1 and HPV-31 replication, indicating that FGFR2 and FGFR3 (Xie et al., 2017) are the most likely tyrosine kinases that phosphorylated E2 protein in vivo. Several FGFR inhibitors are commercially available, however these are not FGF receptor type specific and have off target effects (Katoh, 2016). We tested several FGFR inhibitors but the data were not robust as these also inhibited cell proliferation. Bioinformatic analysis to identify a potential kinase and target tyrosine residues of E2 has been challenging. Most of the predictive tools have given inconsistent results; hence we approached the problem by directly studying the interaction between E2 protein and the FGF family of receptors. Our results are consistent with the expression profiles of these kinases. Human protein atlas (HPA) data SB408124 analysis suggests high expression of FGFR2 (97.5 transcripts per million – TPM) when compared to FGFR1 (17.7 TPM) or FGFR4 (1.1 TPM) in skin tissue samples (Uhlen et al., 2005, Uhlen et al., 2015, Uhlen et al., 2010). Only FGFR3 has higher expression (334.6 TPM) than FGFR2. It is likely that PV E2 interacts with and is regulated by these relatively abundant skin tissue tyrosine kinases. Interestingly, the median expression of FGFR2 and FGFR3 across several skin tissues is reduced in primary tumor samples compared to normal samples as analyzed using Metabolic gEne Rapid Visualizer (Shaul et al., 2016), suggesting growth inhibitory properties for these two kinases. In an effort to identify the tyrosine residues phosphorylated by FGFR2, we performed mass spectrometry with immunoprecipitated BPV-1 E2 protein from 293TT cells. The majority of the tyrosine phosphorylations detected were located throughout the transactivation domain (at amino acids 32, 44, 131, 158, 159, 169, and 170), suggesting these may influence protein-protein interactions. For example, HPV-18\'s Y36 (homologous to BPV Y32) in the E2/E1 co-crystal structure (PDB 1TUE) (Abbate et al., 2004) directly faces E1, and therefore could disrupt binding. Similarly HPV16 Y44 in the E2/Brd4 co-crystal structure (PDB 2NNU) (Abbate et al., 2006) is oriented externally towards the Brd4 fragment. Residue Y131 mediates binding to the cellular protein Chlr1. Phosphorylation at this residue could disrupt this protein-protein interaction, which is required for efficient viral genome segregation and maintenance in cells (Harris et al., 2017). We speculate that at least some if not all phosphorylations among these are inhibitory to viral replication and are specifically induced by FGFR2 and FGFR3. Due to the qualitative nature of mass spectrometry, the lack of detected phospho-tyrosines at other positions does not exclude the possibility of a phosphorylation. Future studies are necessary to characterize and validate each of these detected tyrosine residues as potential phosphorylation sites for FGFR2.
Acknowledgements We appreciate the generosity of the following for plasmids: Alison McBride (NIAID/NIH) for the codon optimized HPV31 E2, Jacques Archambault (McGill Univ.) for BPV1 and HPV31 luciferase replicons and Leslie Thompson (UC Irvine) for the FGFR constructs. Research reported in this publication was supported by the National Cancer Institute, National Institute of Arthritis and Musculoskeletal and Skin Diseases, and National Institute of Allergy and Infectious Diseases of the National Institutes of Health: R01CA058376 (EJA), T32AI060519 (MD), T32AR062495 (TG), T32AI007637 (TG), and Beijing Natural Science Foundation grant # 7174347 (FX). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.