Mutations in VP1 and 5'-UTR affect enterovirus 71 virulence

Abstract

Enterovirus 71 (EV71) is a major cause of hand, foot and mouth disease (HFMD). The current EV71 propagating in Vero (EV-V) or sub-passaged in RD (EV-R) cells was used as a pathogen. Interestingly, EV-R exhibited differential virulence; challenging human scavenger receptor class B2-expressing (hSCARB2-Tg) mice with EV71 revealed that EV-V was more virulent than EV-R: 100% of mice that received lethal amounts of EV-V died, while all the mice that received EV-R survived. Severe pathogenesis correlated with viral burdens and proinflammatory cytokine levels were observed in EV-V-challenged mice, but controversy in EV-R-challenged mice. Consensus sequence analysis revealed EV-R rapidly acquired complete mutations at E145G and S241L and partial mutations at V146I of VP1, and acquired a T to C substitution at nucleotide 494 of the 5'-UTR. EV-R exhibited higher binding affinity for another EV71 receptor, human P-selectin glycoprotein ligand-1 (hPSGL-1), than EV-V. Both EV71s exhibited no significant difference in binding to hSCARB2. The molecular modelling indicate that these mutations might influence EV71 engagement with PSGL-1 and in vivo virulence.

Figure 1

Challenging hSCARB2-Tg mice with EV-V, but not EV-R, causes lethal pathogenesis. Fourteen-day-old hSCARB2-Tg mice were challenged s.c. with 5 × 10⁶ pfu of EV-V (■) or EV-R (●), and animals were monitored daily for (A) scoring of CNS-related paralysis and (B) survival for 15 days. The number (N) of transgenic mice per group is shown. Two-way ANOVA and log-rank test were used for statistical analysis in (A and B), respectively.

Figure 2

Tissue pathogenesis is induced by EV-V, but not EV-R, in hSCARB2-Tg mice.Serial sections of spinal cord and the surrounding muscle from naïve mice and hSCARB2-Tg mice challenged with 5 × 10⁶ pfu of EV-V or EV-R were obtained on day 4 pi and then stained with dyes or antibodies. Representative images of H/E-stained (A) spinal and (B) muscle sections are shown. (C) Representative images of VP2 of EV71 in respective muscle sections that were stained with Mab979, a VP2-specific antibody. All images were obtained at 10× magnification, and the scale bars represent 200 µm. Similar results were obtained from two independent experiments with four mice per group; one is shown.

Figure 3

Expression of VP1 and proinflammatory cytokines in the tissues of EV71-infected Tg-mice. Four days afterchallenginghSCARB2-Tg mice with 5 × 10⁶ pfu of EV-V or EV-R s.c., RNA was extracted from the brainstem, spinal cord, and muscle, and quantitative RT-PCR analysis was conducted to quantify (A) VP1, (B) TNF-α, (C)CCL3, and (D) CXCL10 expression. hSCARB2-Tg mice that received medium were used as the negative control. The number of PCR cycles required for fluorescent detection of target genes was calculated and is presented as the relative expression after normalization to the internal control of β-actin expression from the respective tissue as described in the Materials and Methods. A schematic representation of target gene expression and the statistical average from 5 mice per group are shown. One tail unpaired student’s t-test was used for statistical analysis.

Figure 4

Binding activity of EV-R vs. EV-V in the interaction of human PSGL-1 and SCARB2. Ninety-six-well plates coated with live EV-V or EV-R (10³ pfu) were incubated with various amounts of recombinant (A) hSCARB2-Fc or (B) hPSGL-1-Fc per well for 1 hour at room temperature. After incubation, the plates were washed three times with PBS, and then the EV71 particles in the wells were quantified by incubating with MAB979 antibody in an ELISA assay as described in the Materials and Methods. Three independent experiments were performed, and one of results is shown. One tail unpaired student’s t-test was used for statistical analysis.

Figure 5

Susceptibility of EV-R vs. EV-V in human receptor-expressing mouse cells. L929-PSGL-1 and NIH3T3-SCARB2 cells (4 × 10⁵ per well) were cultured in 6-well plates overnight and subsequently infected separately with EV-R or EV-Vat an MOI = 0.1. Cells were washed after 1 hour of infection and cultured for another 1 or 3 hrs for viral RNA extraction or 12, 24, or 48 hrs for a plaque-forming assay. (A,C) RNA was extracted from EV71-infected (A) L929-PSGL-1 or (C) 3T3-SCARB2 cells and subjected to quantification of EV71 transcripts with real-time RT-PCR using specific primers against the VP1 region and β-actin. The relative expression of the target gene was calculated as described in the Materials and Methods. Each VP1 2^Ct value was normalized by calculating the ratio to the respective β-actin 2^Ct value, and then the means of normalized 2^Ct values were calculated. A schematic representation of the expression ofEV71 transcripts is shown. (B,D) The supernatants from EV71-infected (B) L929-PSGL-1 or (D) 3T3-SCARB2 cells were harvested and used to assay for viral titre with a plaque-forming assay as described in the Materials and Methods. Data represent one of three independent experiments. One tail unpaired student’s t-test was used for statistical analysis.

Figure 6

Modelling of the steric structure of EV71 particles and EV71 mutants. EV-V and EV-R particles were prepared for TEM imaging as described in the Methods. (A) The atomic structure of VP1 extracted from VP1-4 complex (PDB: 3VBS) was fitted into cryoEM structure of EV71 (EMDB: 5558). The spheres showed the residues 145 (colored in red), 146 (colored in green), and 241 (colored in orange) surrounding the 5-fold axis of the structure. The N, C terminus are indicated. The scale bar = 10 nm. (B) The 3-D configurations of the side chains of the residues at VP1-145, 146, and 241 from EV-V and EV-R are shown. The VP1 of EV-R was modelled from SWISSMODE website, a protein structure homology-modelling server. The residues were colored as previous figure. (C) The cryoEM images (upper panel) and 2-D classification (lower panel) of EV71 particles derived from Vero cell line (EV-V, left panel) and from RD cell lone (EV-R, right panel). The image analyses showed that the empty (E) and full (F) particles indicated by white and black arrows, respectively existed in the samples and the diameter of EV71 particles were approximately 30 nm. The scale bar in the micrography is 200 nm. The particle picking and 2-D classification analyses were done by EMAN2 and Xmipp where the contrast was inverted.