Progressive multifocal leukoencephalopathy-associated mutations in the JC polyomavirus capsid disrupt lactoseries tetrasaccharide c binding

Abstract

The human JC polyomavirus (JCPyV) is the causative agent of the fatal, demyelinating disease progressive multifocal leukoencephalopathy (PML). The Mad-1 prototype strain of JCPyV uses the glycan lactoseries tetrasaccharide c (LSTc) and serotonin receptor 5-HT2A to attach to and enter into host cells, respectively. Specific residues in the viral capsid protein VP1 are responsible for direct interactions with the α2,6-linked sialic acid of LSTc. Viral isolates from individuals with PML often contain mutations in the sialic acid-binding pocket of VP1 that are hypothesized to arise from positive selection. We reconstituted these mutations in the Mad-1 strain of JCPyV and found that they were not capable of growth. The mutations were then introduced into recombinant VP1 and reconstituted as pentamers in order to conduct binding studies and structural analyses. VP1 pentamers carrying PML-associated mutations were not capable of binding to permissive cells. High-resolution structure determination revealed that these pentamers are well folded but no longer bind to LSTc due to steric clashes in the sialic acid-binding site. Reconstitution of the mutations into JCPyV pseudoviruses allowed us to directly quantify the infectivity of the mutants in several cell lines. The JCPyV pseudoviruses with PML-associated mutations were not infectious, nor were they able to engage sialic acid as measured by hemagglutination of human red blood cells. These results demonstrate that viruses from PML patients with single point mutations in VP1 disrupt binding to sialic acid motifs and render these viruses noninfectious. IMPORTANCE Infection with human JC polyomavirus (JCPyV) is common and asymptomatic in healthy individuals, but during immunosuppression, JCPyV can spread from the kidney to the central nervous system (CNS) and cause a fatal, demyelinating disease, progressive multifocal leukoencephalopathy (PML). Individuals infected with HIV, those who have AIDS, or those receiving immunomodulatory therapies for autoimmune diseases are at serious risk for PML. Recent reports have demonstrated that viral isolates from PML patients often have distinct changes within the major capsid protein. Our structural-functional approach highlights that these mutations result in abolished engagement of the carbohydrate receptor motif LSTc that is necessary for infection. Viruses with PML-associated mutations are not infectious in glial cells, suggesting that they may play an alternative role in PML pathogenesis.

FIG 1

JCPyV with PML-associated mutations are not infectious. (A) PML-associated mutations are highlighted on the wild-type Mad-1 JCPyV VP1 pentamer (surface representation) in complex with LSTc (stick model) (41). Residues in pink indicate mutations used in this study, while those labeled in black represent other PML-associated mutations. (B) Growth of JCPyV VP1 wild-type (WT) and PML-associated mutant viruses. SVG-A cells were transfected with linearized DNA from wild-type and mutant JCPyV. Transfected cells were fixed and stained at day 4 posttransfection and then at 3-day intervals for 22 days by indirect immunofluorescence. Transfected or infected cells were quantified based on nuclear VP1 staining. Each data point represents the average number of infected cells per visual field for 6 fields of view for 3 independent experiments. Error bars indicate standard deviations. (C) Infectivity of supernatants from JCPyV VP1 wild-type and mutant viruses. SVG-A cells were inoculated with supernatants harvested from infected cells at day 22 from panel B. Cells were fixed and stained by indirect immunofluorescence at 72 h postinfection and quantified based on nuclear VP1 staining. The results are presented as the average number of infected cells per visual field for 6 visual fields from 3 individual samples performed in triplicate. Error bars indicate standard deviations.

FIG 2

VP1 pentamers of JCPyV Mad-1 with PML-associated mutations exhibit reduced binding to cells. (A) SVG-A cells were incubated with 100 µg/ml of His-tagged wild-type (WT) or mutant pentamers in PBS, washed, and then incubated with a penta-His Alexa Fluor 488 antibody. Pentamer binding was analyzed by flow cytometry. Histograms represent the fluorescence intensity of Alexa 488 for antibody alone (gray) and pentamer samples for 10,000 gated events. (B) Quantitation of binding of VP1 pentamers with PML-associated mutations. Bar graph represents the mean fluorescence intensity of VP1 pentamers binding to SVG-A cells for 3 independent experiments. Error bars indicate standard deviations.

FIG 3

JCPyV VP1 proteins with PML-associated mutations have altered oligosaccharide-binding sites. (A) JCPyV wild-type (WT) VP1 pentamer-LSTc complex. (B) S268F VP1 pentamer-LSTc complex. (C) Unliganded L54F VP1 pentamer. (D) Unliganded S268Y VP1 pentamer. The JCPyV VP1 proteins are shown in cartoon representation. Important residues contributing to ligand binding and specificity are shown in stick representation. Side chains of residues at positions 54, 266, and 268 are highlighted in color. Key hydrogen bonds are shown as blue dashes, and red arrows indicate important distances. Carbohydrate residues are shown in stick representation and colored in orange or light orange (light orange indicates binding with reduced affinity) when present in the complex structures. Carbohydrate moieties depicted in gray are shown only for reference purposes and are obtained by superposing with the JCPyV WT VP1 pentamer-LSTc complex structure.

FIG 4

JCPyV pseudoviruses with PML-associated mutations are not infectious. The cell types shown were plated in a 96-well plate O/N. Medium was aspirated, cells were infected with 1 × 10⁷ particles/ml of wild-type (Mad-1) or mutant pseudovirus in incomplete medium without phenol red at 37°C for 1 h, then complete phenol red-free medium was added, and cells were incubated at 37°C for 72 h. Supernatant from infected cells was collected and analyzed for secreted luciferase using a luminometer. The average relative luciferase units for quadruplicate samples are shown in log scale. Error bars represent standard deviations. These data are representative of 3 experiments performed in triplicate.

FIG 5

JCPyV pseudoviruses with PML-associated mutations do not bind to sialic acid. Wild-type (WT) and mutant pseudoviruses (1 × 10⁷ particles) were added to U-bottomed 96-well plates containing equal volumes of PBS, and 2-fold serial dilutions were made across the plate. Human type O RBCs were added to each well and incubated at 4°C for 4 h. Data are representative of 2 independent experiments.