Structural Analysis of JC Polyomavirus Inhibitors and Receptors

Polyomaviruses are non-enveloped, double-stranded DNA viruses that have been found to infect mammals, birds and fish. The first two discovered human polyomaviruses, JC polyomavirus (JCPyV) and BK polyomavirus (BKPyV), are closely related, and the majority of the population is seropositive for these viruses. While they generally cause an asymptomatic, persistent infection in healthy individuals, they can trigger severe diseases in immunocompromised patients. In the case of JCPyV, viral reactivation results in infection of astrocytes and oligodendrocytes in the central nervous system. This in turn leads to the fatal disease Progressive Multifocal Leukoencephalopathy (PML), which is characterized by the demyelination of these cells. For BKPyV, infection can lead to polyomavirus-associated nephropathy (PVAN) and hemorrhagic cystitis in renal transplant patients. About 10% of kidney transplant patients suffer a graft loss thereafter. So far, there is no efficient cure other than the reconstitution of the immune system for either disease. The icosahedral polyomavirus capsid is comprised of 72 pentamers of the major capsid protein VP1. Each VP1 pentamer is typically associated with one copy of a minor capsid protein, VP2 or VP3. The structure of a VP1 monomer is characterized by a jelly-roll fold consisting of a conserved β-barrel core and highly divergent loops connecting the β-strands. Thus, the loops confer receptor specificity and polyomaviruses have been shown to rely on sialylated carbohydrate receptors for host cell attachment. For JCPyV, it has been shown that the engagement of lactoseries tetrasaccharide c (LSTc) is required for the attachment of JCPyV to host cells. More recently, the adipocyte plasma membrane-associated protein APMAP was identified as putative receptor for JCPyV. Little is known about this type II transmembrane protein, and its extracellular region, which is responsible for binding to JCPyV, was investigated in the course of this work. Secondary structure prediction of the extracellular region indicated the presence of a six-bladed β-propeller, with each blade containing a twisted four-stranded antiparallel β-sheet. This prediction is supported by CD spectroscopic analysis of purified protein. The work performed in this project provides a platform for future studies about the structure of APMAP, its function and its molecular interactions with JCPyV VP1. Since there is still no efficient cure for diseases caused by JCPyV nor BKPyV, an important goal is to develop strategies to interfere with virus receptor binding or to block viral assembly. One possibility for preventing virus receptor interactions or viral assembly is the design of small molecule compounds that can act as viral attachment or assembly inhibitors. This was one project investigated in this work. Fragment-based drug discovery was conducted in order to find small molecule fragments that were analyzed in crystal soaking experiments with JCPyV VP1. Two compounds were shown to bind to JCPyV VP1, both of which display the same binding site. The binding site is located on the inside of the VP1 pentamer and does partially overlap with the VP2 binding site. Based on the identified compounds, molecular modeling studies were performed to find new compounds displaying a higher affinity towards JCPyV VP1. Since none of these compounds was shown to bind, another approach was pursued. Based on the sequence of JCPyV VP2, peptides that were hypothesized to bind to JCPyV VP1 with high affinity were designed and crystallized with JCPyV VP1. Although difference omit maps were observed in multiple data sets, assignment to any peptide structure was not possible and binding of these peptides to JCPyV VP1 could thus not be confirmed. In a third project, JCPyV VP1 protein in complex with single-chain variable fragments of the cross-neutralizing, monoclonal antibody 29B1 was structurally analyzed. The crystal structure revealed the antibody epitope and shed light onto the neutralization escape of characteristic PML mutants. Furthermore, comparison with the complex structure of BKPyV VP1 allowed for the detection of subtle differences in binding of 29B1 between JCPyV VP1 and BKPyV VP1. Together with the analysis of previously solved crystal structures of JCPyV VP1 – antibody fragment complexes, the different binding profiles of these antibodies to JCPyV VP1 variants could be explained. This knowledge could provide a platform for a rational design of vaccines and therapeutic antibodies