Structure of human 71 in complex with a capsid-binding inhibitor

Pavel Plevkaa,1, Rushika Pereraa,1, Moh Lan Yapa, Jane Cardosab, Richard J. Kuhna, and Michael G. Rossmanna,2

aDepartment of Biological Sciences, Purdue University, West Lafayette, IN 47907-2032; and bSentinext Therapeutics, 10050 Penang, Malaysia

Edited by Robert A. Lamb, Northwestern University, Evanston, IL, and approved February 22, 2013 (received for review December 21, 2012) Human enterovirus 71 is a causing hand, foot, and sialylated glycans (23), and P-selectin glycoprotein ligand-1 (24– mouth disease that may progress to fatal encephalitis in infants and 26) were shown to function as possible receptors for EV71. small children. As of now, no cure is available for enterovirus 71 Capsids of numerous such as most HRVs, infections. Small molecule inhibitors binding into a hydrophobic , and EV71 contain a pocket factor that plays an es- pocket within capsid viral protein 1 were previously shown to effec- sential role in regulating particle stability (14, 27). The release of tively limit infectivity of many picornaviruses. Here we report a 3.2- the pocket factor, induced by receptor binding, leads to particle Å-resolution X-ray structure of the enterovirus 71 virion complexed destabilization, externalization of VP4 and the N-termini of the with the capsid-binding inhibitor WIN 51711. The inhibitor replaced VP1 subunits, and subsequently to genome release. Therefore, the the natural pocket factor within the viral protein 1 pocket without presence of the pocket factor functions as a switch in initiating inducing any detectable rearrangements in the structure of the cap- picornavirus infection. Some such as HRV14 and sid. Furthermore, we show that the compound stabilizes enterovirus HRV3 lack a pocket factor, when crystallized (3, 28). Although it 71 virions and limits its infectivity, probably through restricting dy- has been suggested that the pocket factor might be an artifact of namics of the capsid necessary for genome release. Thus, our results the purification procedure (29) it is improbable that the VP1 provide a structural basis for development of antienterovirus 71 pocket is a purely accidental feature of picornavirus capsids, as capsid-binding drugs. compounds that bind into the pocket are capable of regulating the stability and infectivity (30). Small molecule inhibitors that stability | virus bind into the hydrophobic pocket within subunit VP1 can inhibit irrespective of whether they contain a pocket factor in the uman enterovirus 71 (EV71) is a picornavirus associated native state or not (14). In the case of viruses containing the pocket Hwith hand, foot, and mouth disease (1). Nevertheless, EV71 factor, the inhibitor competes with the putative lipid molecule for infections may progress to severe encephalitis that causes various binding to the virus. For viruses that lack the pocket factor, the types of neurological complications including polio-like paralysis. binding of the inhibitor induces a conformational change in the Furthermore, the infection may be fatal for infants and young capsid protein VP1 resulting in an opening of the hydrophobic children (2). EV71 outbreaks are reported throughout the world, cavity between the two β-sheets of the jelly roll fold (31, 32). but have been especially severe in the Asia-Pacific region. Small molecules that bind into the hydrophobic pocket within Picornaviruses are small, icosahedral, nonenveloped animal VP1 were shown to be potent inhibitors of many picornaviruses viruses with positive sense RNA genomes. Picornavirus capsids including rhinoviruses (31), Coxsackie viruses A and B (17, 33), have pseudo T = 3 symmetry with 60 copies of each of four viral and polioviruses (12, 16). The inhibitors bind into the hydro- proteins—VP1, VP2, VP3, and VP4—that form an ∼300 Å di- phobic pocket within VP1 with high affinity and therefore cannot ameter icosahedral shell. The surface of a picornavirus capsid is be dislodged upon receptor binding. Several putative capsid- formed by subunits VP1, VP2, and VP3 that each has a β-sandwich binding inhibitors of EV71 were synthesized and shown to be jelly roll fold, whereas VP4 is a small protein attached to the inner effective (34, 35), but molecular details of their interactions with surface of the capsid. There is a circular depression or “canyon” the capsid proteins were not known. BIOCHEMISTRY around each icosahedral fivefold axis of symmetry on the surface of Although the disease caused by EV71 is a public health threat, many picornaviruses (3, 4). The EV71 canyon is shallower than no vaccines or antiviral compounds are currently available. Here that in polioviruses and rhinoviruses (5, 6). we present the crystal structure of EV71 virions in complex with Receptors with an Ig-like fold have been shown to bind into the the capsid-binding inhibitor WIN 51711 at 3.2 Å resolution. Fur- canyon (7–13). When receptor molecules bind into the canyon, thermore, we show that the compound can limit EV71 infection. they dislodge a “pocket factor” from a pocket within VP1 im- mediately below the floor of the canyon. The shape of the pocket Results and Discussion factor and the hydrophobic environment of the pocket suggest WIN 51711 Limits EV71 Infectivity and Stabilizes Virions. WIN 51711 that the pocket factor is probably a lipid (14–17). Analysis of lipid had previously been shown to be an inhibitor of rhinoviruses and molecules extracted from virions of bovine enterovirus had shown polioviruses (31, 36). Using FACS assay, we have now shown that μ A that the pocket factor is not a single type of molecule but rather the EV71 infectivity is inhibited by 600 M WIN 51711 (Fig. 1 ). a mixture of lipids with the most represented molecules being The decrease in infectivity of EV71 when treated with WIN palmitic acid and myristic acid (15). When a receptor binds into the canyon, it depresses the floor of the canyon corresponding to the roof of the pocket. Similarly, when a lipid or antiviral com- Author contributions: P.P., R.P., R.J.K., and M.G.R. designed research; P.P., R.P., and M.L.Y. performed research; P.P. and R.P. analyzed data; J.C. contributed new reagents/analytic pound binds into the pocket, it expands the roof of the pocket, tools; and P.P., R.P., and M.G.R. wrote the paper. fl corresponding to the oor of the canyon (14, 17). Thus, canyon- The authors declare no conflict of interest. binding receptors and the pocket factor compete with each other This article is a PNAS Direct Submission. for binding to the virion. Nevertheless, not all receptors of Data deposition: The EV71 native and inhibitor complex coordinates, observed structure picornaviruses bind into the canyon. A minor group of human amplitudes, and phases derived by the phase extension have been deposited in the Pro- rhinoviruses (HRV) bind to low-density lipoprotein receptors (18, tein Data Bank, www.pdb.org (PDB ID codes 3ZFE, 3ZFF, and 3ZFG). 19), whereas Coxsackie and echoviruses use decay-accelerating 1R.P. and P.P. contributed equally to this work. factor as a cellular receptor (20, 21). Scavenger receptor B2 (22), 2To whom correspondence should be addressed. E-mail: [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1222379110 PNAS | April 2, 2013 | vol. 110 | no. 14 | 5463–5467 Downloaded by guest on September 24, 2021 Fig. 1. Impact of WIN 51711 on EV71 infectivity and virion stability. (A) FACS analysis of EV71 infectivity. Cells were infected with EV71 in the presence or absence of varying concentrations of WIN 51711 for 30 h and were then fixed and stained with an anti-EV71 antibody and fluorescein-conjugated secondary antibody. The graph shows distribution of fluorescence intensities in the respective cell populations. (B) Sybr green fluorescent assay to measure stability of EV71 particles. EV71 virions were mixed with Sybr green dyes I and II and heated to indicated temperatures. The fluorescent signal increases as the dye binds to RNA that is release from thermally destabilized particles. Green line, EV71 virions; red line, EV71 with 300 μM WIN 51711; purple line, control without virus. See Materials and Methods for details.

51711 results in an increase of the virion to plaque-forming unit tallographic averaging showed clear features of amino acid side ratio. Furthermore, the RNA genome in EV71 virions incubated chains and of carbonyl oxygens. Models of the capsid proteins with 300 μM WIN 51711 was less accessible to RNA binding VP1, VP2, VP3, and VP4 were built except for residues 1 and 298 fluorescent dyes Sybr green I and II than genome of native virions of VP1, 1–9 of VP2, and 1–12 of VP4. Had it been calculated, B (Fig. 1 ). The virions stabilized by WIN 51711 had to be heated crystallographic Rfree would have been similar to crystallographic 3 °C higher than native virions to obtain equal staining of the Rwork, because of the high 20-fold noncrystallographic symmetry genome. It is therefore likely that the mechanism of inhibition is (37, 38). Therefore, all measured reflections were used in the due to the stabilizing effect of WIN 51711 on the EV71 capsid, structure refinement (Table 1). The structure of the icosahedral thereby preventing uncoating of the virion. asymmetric unit of EV71 consists of 840-aa residues. Nine addi- – “ ” X-Ray Structures of Native EV71 Virion and Its Complex with WIN tional residues (135 143) of the VP2 puff loop exposed on the 51711. The crystal structure of EV71 strain MY104-9-SAR-97 particle surface were visible in the electron density map in com- (GenBank DQ341368.1) was determined to 2.7 Å resolution. parison with the previously determined EV71 I212121 structure Structures of the EV71-WIN 51711 complex were determined [Protein Data Bank (PDB) ID code 4AED]. The rmsd between the independently from two datasets that included data to 3.2 Å and positions of Cα atoms in the current and previously determined 3.4 Å resolution. The maps resulting from 20-fold noncrys- EV71 structures were between 0.2 and 0.5 Å.

Table 1. Scaling and refinement statistics Structure

EV71 native EV71 WIN 51711 3.4Å EV71 WIN 51711 3.2Å

Space group I23 I23 I23 Unit cell dimensions, Å 594.5 591.0 592.5 Resolution limits (high-resolution bin), Å 30.4–2.7 (2.82–2.70) 27.5–3.4 (3.55–3.40) 33.2–3.2 (3.35–3.20) Completeness, % 74.9 (35.9) 53.3 (27.5) 66.8 (44.0)

Rmerge* 0.208 (0.709) 0.251 (0.538) 0.332 (0.974) Average redundancy 1.9 (1.2) 1.8 (1.5) 2.5 (2.2) /<σI> 3.05 (0.56) 2.60 (1.06) 2.50 (0.81) Reciprocal space correlation coefficient of 0.904 0.830 0.839

Fobs and Fcalc after convergence of map R-factor 0.240 (0.410) 0.243 (0.344) 0.249 (0.355) Average B-factor 31.0 30.6 36.9 † Ramachandran plot outliers, % 0.24 1.92 1.56 † Ramachandran plot most favored regions, % 95.43 89.54 89.66 Rotamer outliers, %† 1.83 4.77 5.61 rmsd, bonds, Å 0.005 0.008 0.008 rmsd, angles, ° 1.29 1.49 1.49 N of unique reflections 692,970 (34,593) 247,559 (12,721) 353,681 (15,538)

Fcalc, structure factor amplitudes calculated by Fourier inversion of averaged electron density map; Fobs, observed structure factor amplitudes. Values in parenthesesP representP highPP resolution bin. *Rmerge = h jjIhj-hIhij= jIhjj. †According to the criterion of Molprobity.

5464 | www.pnas.org/cgi/doi/10.1073/pnas.1222379110 Plevka et al. Downloaded by guest on September 24, 2021 Fig. 2. Binding of native pocket factor and WIN 51711 into the VP1 pocket. (A) Overview of EV71 protomer with capsid protein subunits VP1 (blue), VP2 (red), VP3 (green), and VP4 (yellow) shown in a cartoon representation. WIN 51711 is shown as a space-filling model in orange. Positions of the icosahedral symmetry elements are indicated. (B) WIN 51711 electron density (green), with WIN 51711 model shown in orange. VP1 is shown in cartoon representation in blue with side chains of residues forming the hydrophobic pocket shown as sticks. Side chain of Leu-24 of VP3 that forms the bottom of the pocket is shown in red. (C) Electron density of the native pocket factor (red). Superimposed WIN 51711 model is shown for comparison.

Shih et al. have identified a single residue mutation, Val192- structure of WIN 51711 than with sphingosine that was used to Met, that confers resistance to the presumed capsid binding in- model the pocket factor in the native structure of EV71 (6). Sim- hibitor BPR0Z-194 (39). Val192 is located in the middle of the ilarly, in the native crystals, the electron density within the pocket wild type VP1 pocket (Fig. 2B). It is therefore likely that the agrees better with the sphingosine model. Thus, WIN 51711 substitution for methionine, a residue with larger side chain, replaced the pocket factor in crystals cocrystallized with the in- prevents binding of BPR0Z-194 to the capsid. This observation hibitor. Nevertheless, it is possible that WIN 51711 did not replace verifies the role of VP1 pocket for the infectivity of the virus. all pocket factor molecules and that the observed electron density represents an average between the pocket factor and WIN 51711. Comparison of WIN 51711 with the Native Pocket Factor. The major The molecule of WIN 51711 is elongated in shape with two ar- difference between the native EV71 and the EV71-WIN 51711 omatic rings at one end and one ring at the other (Fig. 2B). The complex is that the native pocket factor density extends ∼2Å RSCC comparison of simulated electron density maps of WIN further toward the opening of the pocket into the canyon than the 51711 molecules refined in the two alternative orientations within WIN 51711 density (Fig. 2). To evaluate differences in the shape of the VP1 pocket shows that both orientations fit well into the ex-

the pocket factor and WIN 51711 density, real-space correlation perimental density, with slight preference for the orientation of the BIOCHEMISTRY coefficients (RSCC) were calculated to compare the electron inhibitor with the two rings pointing toward the opening of the density distributions within the VP1 pocket of the native and in- pocket (Table 3, Fig. 2B). Comparable ambiguity in determining hibitor complexes. The experimental electron density maps were the orientation of WIN 51711 was encountered in its complex with calculated with phases obtained by phase extension starting from HRV 14 (31, 40). Similar to the situation in EV71, the RSCC 10 Å resolution and are therefore free of model bias. The RSCCs analysis of the two possible orientations of WIN 51711 within the comparing electron density distributions of pocket factor to WIN pocket of HRV 14 indicates a slight preference for the two rings 51711 are less than 0.77, whereas RSCCs comparing different pointing toward the opening of the pocket (Table 3). datasets of the same object are greater than 0.89 (Table 2). The RSCCs comparing experimental electron density maps with those Structure of Residues Forming the VP1 Pocket. The RMSD com- derived from models were calculated to verify the nature of the parison of positions of residues that form the pocket in the native moiety in the pocket (Table 3). For crystals soaked with WIN and the two WIN 51711–containing complexes (Table 4) shows 51711, the density within the pocket correlates better with that there are no major rearrangements induced by WIN 51711

Table 2. RSCC comparison of electron densities within VP1 pockets of native and WIN 51711– derivative EV71 structures Structure Native 3VBF EV71 native WIN 51711 3.2Å WIN 51711 3.4Å

WIN 51711 3.4Å 0.75 0.72 0.91 — WIN 51711 3.2Å 0.77 0.73 — EV71 native 0.89 — Native-3VBF —

Numbers in bold indicate that the compared electron density maps correspond to the same compound. A dash indicates a comparison is meaningless as the structures are identical.

Plevka et al. PNAS | April 2, 2013 | vol. 110 | no. 14 | 5465 Downloaded by guest on September 24, 2021 Table 3. RSCC comparison of experimental electron densities within VP1 pocket with electron densities derived from refined models Structure Pocket factor model RSCC

WIN 51711 3.2 Å WIN 51711–2OUT* 0.85 WIN 51711–2IN† 0.82 Sphingosine 0.61 WIN 51711 3.4 Å WIN 51711–2OUT* 0.80 † WIN 51711–2IN 0.78 Sphingosine 0.77 EV71 native WIN 51711–2OUT* 0.80 Sphingosine 0.87 EV71 native 3VBF WIN 51711–2OUT* 0.80 Sphingosine 0.92 HRV14 WIN 51711–2OUT* 0.82 † WIN 51711–2IN 0.80

*Orientation of the WIN51711 compound in the VP1 pocket in which the end of the molecule that contains two aromatic rings points toward the opening of the pocket. † Orientation of the WIN51711 compound in the VP1 pocket in which the end Fig. 3. Comparison of the positions of side chains of residues forming the of the molecule that contains single aromatic ring points toward the open- VP1 hydrophobic pocket in the WIN 51711 complex (dark blue) and native ing of the pocket. structure (light blue). Main-chain trace of protein subunits is shown as rib- bon and side chains of residues forming the pocket are shown as sticks. VP1, blue; VP3, red. WIN 51711 is shown as a stick model in orange. binding within the pocket (Fig. 3). The inhibitor interacts mostly with hydrophobic side chains of residues forming the interior of the pocket. Similarly, changes in the structure of HRV16, where detector at beamline 14 C BioCARS at the Advanced Photon Source syn- the inhibitors replaced the pocket factor, were quite minor (41). chrotron. An oscillation range of 0.2° was used during data collection. The In contrast, binding of WIN 51711 to the collapsed pocket of diffraction images were processed and scaled using the HKL2000 package (43) (Table 1). HRV14 induced conformational changes within VP1 (31). X-Ray Structure Determination. Virions of EV71 crystallized in the body- Design of EV71 Capsid-Binding Inhibitors. Our results show the mo- centered cubic space group I23. The cell parameters differed from crystal to lecular details of the WIN 51711 interactions with EV71. Thus, crystal by up to 5 Å (Table 1). The crystallographic asymmetric unit contained WIN 51711 can be used as a scaffold for further development of one-third of a virion. The orientation of the virion about the crystallographic more potent inhibitors. For example, extra groups added to the threefold axes was determined with a one-dimensional locked-rotation five-membered oxazoline ring might interact with polar residues function search using the program GLRF (44). Orientations of the particles in that form the opening of the pocket into the canyon. Considering different crystals differed by up to 2°. The differences in particle orientation that EV71 receptors probably do not bind into the shallow EV71 could be correlated with differences in the unit cell parameters among in- canyon (22, 24), the EV71 pocket factor may function merely to dividual crystals. Position of the particle center was determined with a one- stabilize the particle and not as a switch regulating genome release dimensional translation function search using the program Phaser (45). Initial phases up to 10 Å resolution were calculated from a properly oriented and as in rhinoviruses, polioviruses, and Coxsackie viruses. Therefore, positioned model of EV71 (PDB ID code 4AED) with the program CNS (46). the critical property of future EV71 capsid-binding inhibitors is The atoms corresponding to the pocket factor were removed from the model their ability to stabilize the virus. before the phase calculation. The phases were refined with 10 cycles of 20-fold real-space noncrystallographic symmetry averaging with the program Materials and Methods AVE (47). The mask defining the volume of electron density to be averaged EV71 Preparation, Crystallization, and Diffraction Data Collection. EV71 was was derived from the EV71 atomic model by including all grid points within produced and purified as described previously (42). Crystals of EV71 were 5 Å of each atom. Grid points outside the capsid were set to the average obtained using the hanging drop technique with conditions analogous to value of the density outside the mask. Phase information for reflections im- those described by Wang et al. (6). The crystallization drops were prepared mediately outside the current resolution limit was obtained by extending the − by mixing 0.83 μL EV71 at ∼2 mg/mL in PBS with 0.17 μL 0.2 M sodium citrate, resolution in steps of (1/a) Å 1 followed by three cycles of averaging. This 0.1 M Tris pH 8.5, and 30% (vol/vol) PEG 400. The well solution contained 1.8 M procedure was repeated until the high-resolution limit of the available data sodium acetate and 0.1 M Bis-Tris propane, pH 7.0. Crystals formed within was reached. For each of the data sets, the particle orientation and position 1 wk. The crystals of EV71 WIN 51711 complex were prepared in the same were refined by searching for their best value as judged by the correlation way except that the virus solution included 2% (vol/vol) dimethyl sulfoxide coefficient between observed and calculated structure amplitudes. and 0.1 mg/mL WIN 51711. For data collection, crystals were flash frozen in The structure was built manually using the program O (48) starting from liquid nitrogen. The crystallization solution was a sufficient cryoprotectant. the native EV71 structure. Subsequently, the structure was subjected to co- Each dataset was collected from a single crystal at 100 K on the ADSC 315 ordinate and B-factor refinement using the program CNS. The refinement used noncrystallographic symmetry constraints. Other calculations used the CCP4 suite of programs (49). Table 4. rmsd of residues forming the hydrophobic pocket Data Deposition. The EV71 native and inhibitor complex coordinates together Structure WIN 51711 3.2Å WIN 51711 3.4Å EV71 native with the observed structure amplitudes and phases derived by the phase EV71 native 0.159 0.189 — extension have been deposited in the Protein Data Bank (PDB ID codes 3ZFE, WIN 51711 3.4 Å 0.139 — 3ZFF, and 3ZFG). WIN 51711 3.2 Å — FACS Infectivity Assay. The impact of WIN 51711 on the infectivity of EV71 was rmsd of residues that have at least one atom closer than 4.0 Å to any atom determined using a FACS assay. Approximately 1 × 106 pfu of EV71 were of WIN 51711. The program O (48) was used for superposition of the residues. incubated with varying concentrations of WIN51711 at 37 °C for 30 min. The A dash indicates a comparison is meaningless as the structures are identical. virus and compound mixture was then adsorbed onto rhabdomyosarcoma

5466 | www.pnas.org/cgi/doi/10.1073/pnas.1222379110 Plevka et al. Downloaded by guest on September 24, 2021 cells for 1 h at room temperature and overlaid with liquid media (MEM + necessary because interaction of Sybr green dyes with RNA is temperature- 10% FCS) containing the same concentrations of compound. Following in- dependent. cubation at 37 °C for 30 h, the cells were trypsinized and equal numbers of cells were fixed with 1% formaldehyde, permeabilized with PBS containing Calculation of RSCCs. RSCCs comparing two experimental maps or comparing 0.5% Triton X-100 and blocked with PBS + 10 mg/mL BSA (FACS buffer). The an experimental map with a map calculated from a refined model were cells were then stained with an anti-EV71 monoclonal antibody (EV20/5/A5- calculated using the program MAPMAN from the USF package (47). The 3/C10/C1, provided by MAb Explorations) and an FITC-conjugated anti- experimental maps were calculated using phases derived from phase ex- mouse secondary antibody for 30 min at 4 °C with extensive washing with tension as described previously. The model-derived maps were calculated FACS buffer between antibody incubations. The cells were then analyzed with the CNS program (46). The RSCC calculation included all grid points using a Beckman Coulter FC500 flow cytometer. within 1.5Å of any atom of a given residue (∼320 grid points were used for WIN 51711 and sphingosine molecules). Thermal Stability Assay. Virions of EV71 (final concentration 0.125 mg/mL) fi were mixed with WIN 51711 ( nal concentration 0.01 mg/mL) in NaCl-Tris- ACKNOWLEDGMENTS. We thank Vukica Srajer, Robert Henning, and the EDTA buffer (20 mM Tris, pH 8.0, 120 mM NaCl, and 1 mM EDTA) and in- other staff of the Advanced Photon Source (Argonne National Labora- cubated for 1 h at 37 °C. Sybr green I and Sybr green II dyes (Life Technol- tory) BioCARS beamline 14, and Robert Fischetti of GM/CA beamline 23 ogies) were added at 3× concentration (according to the manufacturer’s for help with data collection. Use of BioCARS Sector 14 was supported by information) together with ANTI-RNase (ambion) at 1 U/μL final concentra- the National Institutes of Health, National Center for Research Resources tion. The Sybr green dyes were previously shown not to penetrate into na- (NIH/NCRR) Grant RR007707. Use of the Advanced Photon Source was supported by the US Department of Energy, Office of Science, Office of tive virions (50). The Sybr green fluorescence of the mixture was Basic Energy Sciences, under Contract DE-AC02-06CH11357. The authors then analyzed in a real-time PCR machine (Applied Biosystems 7300). The acknowledge the use of the Flow Cytometry and Cell Separation Facility – mixture was heated to temperatures in the range of 37 90 °C in 1-°C steps. of the Bindley Bioscience Center at Purdue University (supported by After heating to a given temperature for 2 min, the mixture was cooled to NIH/NCRR Grant RR025761). This study was supported by NIH Grant 36 °C for 2 min, during which the fluorescence could be measured. This was AI11219 (to M.G.R.).

1. Chan LG, et al.; For the Outbreak Study Group (2000) Deaths of children during an 27. Rossmann MG, He Y, Kuhn RJ (2002) Picornavirus-receptor interactions. Trends Mi- outbreak of hand, foot, and mouth disease in Sarawak, Malaysia: clinical and path- crobiol 10(7):324–331. ological characteristics of the disease. Clin Infect Dis 31(3):678–683. 28. Zhao R, et al. (1996) Human 3 at 3.0 Å resolution. Structure 4(10): 2. Alexander JP, Jr., Baden L, Pallansch MA, Anderson LJ (1994) Enterovirus 71 infections 1205–1220. and neurologic disease—United States, 1977-1991. J Infect Dis 169(4):905–908. 29. Katpally U, Smith TJ (2007) Pocket factors are unlikely to play a major role in the life 3. Rossmann MG, et al. (1985) Structure of a human common cold virus and functional cycle of human rhinovirus. J Virol 81(12):6307–6315. relationship to other picornaviruses. Nature 317(6033):145–153. 30. Grant RA, et al. (1994) Structures of poliovirus complexes with anti-viral drugs: im- 4. Rossmann MG (1989) The canyon hypothesis. Hiding the host cell receptor attachment plications for viral stability and drug design. Curr Biol 4(9):784–797. – site on a viral surface from immune surveillance. J Biol Chem 264(25):14587 14590. 31. Smith TJ, et al. (1986) The site of attachment in human rhinovirus 14 for antiviral 5. Plevka P, Perera R, Cardosa J, Kuhn RJ, Rossmann MG (2012) Crystal structure of hu- agents that inhibit uncoating. Science 233(4770):1286–1293. man enterovirus 71. Science 336(6086):1274. 32. Chapman MS, Rossmann MG (1993) Comparison of surface properties of picornavi- 6. Wang X, et al. (2012) A sensor-adaptor mechanism for enterovirus uncoating from ruses: Strategies for hiding the receptor site from immune surveillance. Virology structures of EV71. Nat Struct Mol Biol 19(4):424–429. 195(2):745–756. 7. Belnap DM, et al. (2000) Three-dimensional structure of poliovirus receptor bound to 33. Xiao C, et al. (2005) The crystal structure of coxsackievirus A21 and its interaction with poliovirus. Proc Natl Acad Sci USA 97(1):73–78. ICAM-1. Structure 13(7):1019–1033. 8. Colonno RJ, et al. (1988) Evidence for the direct involvement of the rhinovirus canyon 34. Chang CS, et al. (2005) Design, synthesis, and antipicornavirus activity of 1-[5-(4- in receptor binding. Proc Natl Acad Sci USA 85(15):5449–5453. arylphenoxy)alkyl]-3-pyridin-4-ylimidazolidin-2-one derivatives. J Med Chem 48(10): 9. Olson NH, et al. (1993) Structure of a human rhinovirus complexed with its receptor – molecule. Proc Natl Acad Sci USA 90(2):507–511. 3522 3535. 10. He Y, et al. (2000) Interaction of the poliovirus receptor with poliovirus. Proc Natl 35. Shia KS, et al. (2002) Design, synthesis, and structure-activity relationship of pyridyl Acad Sci USA 97(1):79–84. imidazolidinones: A novel class of potent and selective human enterovirus 71 in- 11. He Y, et al. (2003) Complexes of poliovirus serotypes with their common cellular re- hibitors. J Med Chem 45(8):1644–1655. ceptor, CD155. J Virol 77(8):4827–4835. 36. Lentz KN, et al. (1997) Structure of poliovirus type 2 Lansing complexed with antiviral 12. Zhang P, et al. (2008) Crystal structure of CD155 and electron microscopic studies of its agent SCH48973: Comparison of the structural and biological properties of three complexes with polioviruses. Proc Natl Acad Sci USA 105(47):18284–18289. poliovirus serotypes. Structure 5(7):961–978. 13. Xing L, et al. (2000) Distinct cellular receptor interactions in poliovirus and rhinovi- 37. Arnold E, Rossmann MG (1990) Analysis of the structure of a common cold virus, BIOCHEMISTRY ruses. EMBO J 19(6):1207–1216. human rhinovirus 14, refined at a resolution of 3.0 Å. J Mol Biol 211(4):763–801. 14. Rossmann MG (1994) Viral cell recognition and entry. Protein Sci 3(10):1712–1725. 38. Kleywegt GJ, Brünger AT (1996) Checking your imagination: Applications of the free 15. Smyth M, Pettitt T, Symonds A, Martin J (2003) Identification of the pocket factors in R value. Structure 4(8):897–904. a picornavirus. Arch Virol 148(6):1225–1233. 39. Shih SR, et al. (2004) Mutation in enterovirus 71 capsid protein VP1 confers resistance 16. Filman DJ, et al. (1989) Structural factors that control conformational transitions and to the inhibitory effects of pyridyl imidazolidinone. Antimicrob Agents Chemother serotype specificity in type 3 poliovirus. EMBO J 8(5):1567–1579. 48(9):3523–3529. 17. Muckelbauer JK, et al. (1995) The structure of coxsackievirus B3 at 3.5 Å resolution. 40. Badger J, Minor I, Oliveira MA, Smith TJ, Rossmann MG (1989) Structural analysis of – Structure 3(7):653 667. antiviral agents that interact with the capsid of human rhinoviruses. Proteins 6(1): 18. Hofer F, et al. (1994) Members of the low density lipoprotein receptor family mediate 1–19. cell entry of a minor-group common cold virus. Proc Natl Acad Sci USA 91(5): 41. Hadfield AT, Diana GD, Rossmann MG (1999) Analysis of three structurally related 1839–1842. antiviral compounds in complex with human rhinovirus 16. Proc Natl Acad Sci USA 19. Querol-Audí J, et al. (2009) Minor group human rhinovirus-receptor interactions: 96(26):14730–14735. Geometry of multimodular attachment and basis of recognition. FEBS Lett 583(1): 42. Plevka P, Perera R, Cardosa J, Kuhn RJ, Rossmann MG (2012) Structure determination 235–240. of enterovirus 71. Acta Crystallogr D Biol Crystallogr 68(Pt 9):1217–1222. 20. Clarkson NA, et al. (1995) Characterization of the echovirus 7 receptor: Domains of 43. Otwinowski Z, Minor W (1997) Processing of X-ray diffraction data collected in os- CD55 critical for virus binding. J Virol 69(9):5497–5501. cillation mode. Methods Enzymol 276:307–326. 21. Lea SM, et al. (1998) Determination of the affinity and kinetic constants for the in- 44. Tong L, Rossmann MG (1997) Rotation function calculations with GLRF program. teraction between the human virus echovirus 11 and its cellular receptor, CD55. J Biol – Chem 273(46):30443–30447. Methods Enzymol 276:594 611. – 22. Yamayoshi S, et al. (2009) Scavenger receptor B2 is a cellular receptor for enterovirus 45. McCoy AJ, et al. (2007) Phaser crystallographic software. J Appl Cryst 40(Pt 4):658 674. 71. Nat Med 15(7):798–801. 46. Brünger AT, et al. (1998) Crystallography & NMR system: A new software suite for 23. Yang B, Chuang H, Yang KD (2009) Sialylated glycans as receptor and inhibitor of macromolecular structure determination. Acta Crystallogr D Biol Crystallogr 54(Pt 5): enterovirus 71 infection to DLD-1 intestinal cells. Virol J 6:141. 905–921. 24. Miyamura K, Nishimura Y, Abo M, Wakita T, Shimizu H (2011) Adaptive mutations in 47. Kleywegt GJ, Read RJ (1997) Not your average density. Structure 5(12):1557–1569. the genomes of enterovirus 71 strains following infection of mouse cells expressing 48. Jones TA, Zou JY, Cowan SW, Kjeldgaard M (1991) Improved methods for building human P-selectin glycoprotein ligand-1. J Gen Virol 92(Pt 2):287–291. protein models in electron density maps and the location of errors in these models. 25. Nishimura Y, et al. (2009) Human P-selectin glycoprotein ligand-1 is a functional re- Acta Crystallogr A 47(Pt 2):110–119. ceptor for enterovirus 71. Nat Med 15(7):794–797. 49. Collaborative Computational Project, Number 4 (1994) The CCP4 suite: Programs for 26. Nishimura Y, Wakita T, Shimizu H (2010) Tyrosine sulfation of the amino terminus of protein crystallography. Acta Crystallogr D Biol Crystallogr 50(Pt 5):760–763. PSGL-1 is critical for enterovirus 71 infection. PLoS Pathog 6(11):e1001174. 50. Brandenburg B, et al. (2007) Imaging poliovirus entry in live cells. PLoS Biol 5(7):e183.

Plevka et al. PNAS | April 2, 2013 | vol. 110 | no. 14 | 5467 Downloaded by guest on September 24, 2021