Polyprotein Context Regulates the Activity of Poliovirus 2C Bound to Bilayer Nanodiscs

1 Positive strand RNA viruses generally replicate in large membrane associated complexes. For 2 poliovirus, these replication complexes are anchored to the membrane via the viral 2B, 2C, and 3 3A proteins. 2C is an AAA+ family ATPase that plays a key role in host cell membrane 4 rearrangement, is a putative helicase, and is implicated in virion assembly and packaging. 5 However, the membrane binding characteristics of all these viral proteins have made it difficult 6 to elucidate their exact roles in virus replication. In this work we show that small lipid bilayers 7 known as nanodiscs can be used to chaperone the in vitro expression of soluble poliovirus 2C, 8 2BC, and 2BC3AB polyproteins in a membrane bound form. ATPase assays on these proteins 9 show that the activity of the core 2C domain is stimulated ≈60-fold as compared to the larger 10 2BC3AB polyprotein, with most of this stimulation occurring upon removal of 2B. The proteins 11 are active over a wide range of salt concentrations, exhibit slight lipid headgroup dependence, 12 and show significant stimulation by acetate. Our data lead to a model wherein the replication 13 complex can be assembled with a minimally active form of 2C that then becomes fully activated 14 by proteolytic cleavage from the adjacent 2B viroporin domain. 15 16 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C INTRODUCTION 17 The picornaviruses are a family of human and mammalian pathogens that include hepatitis A 18 virus, the rhinoviruses, foot-and-mouth disease virus, coxsackie virus, and poliovirus (PV). PV 19 remains a health threat in many parts of the world, where it is being kept in check by extensive 20 immunization efforts aimed at its eventual eradication (1). Poliovirus has a 7.5 kb single21 stranded, positive-sense RNA genome that encodes a single ≈250 kD polypeptide. Upon 22 translation, this viral polyprotein undergoes a series of proteolytic cleavages to generate a total of 23 eleven different proteins plus several important functional intermediates (Figure 1A). The four 24 proteins in the P1 region (VP1-VP4) are the structural proteins that make up the virion capsid 25 while the remaining seven non-structural proteins in the P2 and P3 regions (2A-2C and 3A-3D, 26 respectively) are involved in viral replication biochemistry (2). Processing of the polyprotein 27 plays an important role in mediating protein function as several viral precursor proteins have 28 different functional roles and biochemical activities than the fully processed proteins. For 29 example, the cleavage of poliovirus 3CD generates and activates 3D, the RNA-dependent 30 RNA polymerase that carries out replication of the viral RNA (3-5). Similarly, the 2C and 31 precursor 2BC proteins act in concert to trigger a massive rearrangement of intracellular 32 membranes that results in the formation of extensive “membranous web” structures (6-8), on the 33 surfaces of which these proteins form the viral replication centers that are the sites of RNA 34 synthesis and virion assembly. 35 The viral 2B, 2C and 3A proteins all contain membrane-binding regions. Both 2BC and 2C 36 are able to induce membrane rearrangement in living cells (9-11); however, co-expression of a 37 combination of these membrane binding proteins is needed to recapitulate the double walled 38 membranes observed during virus infection (12). 2B contains a pair of predicted integral 39 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C membrane helices and is thought to form a viroporin structure (13-15), 2C is anchored to the 40 membrane via a N-terminal amphipathic helix (10, 16-18), and 3A contains an amphipathic 41 sequence within its C-terminal 22 residues (19, 20). There is evidence that these three 42 membrane-anchored proteins interact with one another both in vivo and in vitro; substitution of 43 the 2C amphipathic helix from other picornaviruses into a poliovirus background results in 44 compensatory mutations in the membrane anchoring sequences of both 2B and 3A (21, 22). In 45 addition to its roles in membrane rearrangements, 2C has also been implicated in initiation of 46 negative-strand RNA synthesis (23), interacting with the VP3 capsid protein (24), and virion 47 packaging (25). 48 2C is comprised of 329 amino acids and contains an N-terminal amphipathic helix, two 49 predicted RNA binding regions, three motifs implicated in NTP binding, and a cysteine rich 50 putative zinc finger motif (Figure 1B) (18, 26, 27). Two of the NTP binding motifs are Walker A 51 and B, common to AAA+ ATPases, while the third is motif C/Sensor-1 that shares homology 52 with the superfamily III helicases (28). 2C has been shown to have ATPase activity and may also 53 have low level GTPase activity (26, 28, 29). Electron microscopy images of 2C proteins from 54 poliovirus and foot-and-mouth disease virus (FMDV) show hexameric ring structures and 55 activity assays indicate that oligomerization of FMDV 2C is important for ATP hydrolysis (30, 56 31). There are no atomic level structures for the picornaviral 2C proteins and sequence homology 57 searches do not reveal convincing matches for proteins other than those from closely related 58 viruses. Structure prediction using the Phyre engine (32) that utilizes profile-profile sequence 59 matching and structure based threading to do homology modeling shows a high probability that 60 residues ≈100-260 fold into a AAA+ family ATPase. The AAA+ (ATPases associated with 61 diverse cellular activities) superfamily of enzymes are often considered protein workhorses 62 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C because they are involved in a multitude of basic cellular functions, including protein 63 degradation, membrane rearrangement, DNA replication, and disassembly of protein complexes 64 (33, 34). These enzymes tend to oligomerize (33), forming structures that are consistent with the 65 aforementioned electron microscopy data showing 2C ring assemblies. 66 The propensity of 2C to associate with membranes has made it difficult to study and much 67 remains unknown about both its structure and its function during viral replication. Thus far, 68 poliovirus 2C has been generated using GST (28) or maltose binding protein (30) fusions to 69 improve expression and solubility, but those proteins remained difficult to work with. FMDV 2C 70 has been purified in a soluble form by deleting the N-terminal membrane binding region to 71 generate a soluble and biochemically active core ATPase fragment (31). Attempts by our 72 laboratory to express poliovirus 2C using similar approaches in both E. coli and baculovirus 73 systems resulted in protein that was largely insoluble and had little or no ATPase activity. 74 Additionally, our GST-2C fusion construct could not be efficiently cleaved using PreScission 75 protease, a derivative of the picornaviral rhinovirus 3C protease, suggesting the protein may not 76 have been natively folded. 77 To circumvent these difficulties with expression and purification, we explored a new 78 approach for isolating 2C by expressing it in vitro on small lipid bilayers known as nanodiscs 79 (Figure 1C). Developed by the Sligar laboratory, nanodiscs are soluble nanolipoprotein particles 80 comprised of a lipid bilayer held together by an amphipathic helix membrane scaffold protein 81 (MSP) that forms a belt around the lipid acyl chains (35, 36). Nanodiscs provide a native bilayer 82 surface to which membrane proteins can adhere, thus reducing the likelihood of aggregation, 83 while being small enough to remain fully soluble and tumble rapidly enough to enable a variety 84 of spectroscopy experiments (37-39). Most importantly, this system has allowed us to 85 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C biochemically characterize 2C activity when bound to true membrane bilayers instead of 86 detergent micelles. Additionally, this allowed us to explore 2C activity in the context of its 87 membrane-associated polyprotein neighbors to examine how processing may affect 2C function. 88 In this work we demonstrate the in vitro expression and subsequent purification of poliovirus 89 2C, 2BC, 2C3A, 2BC3AB and 3AB proteins on nanodiscs composed of dimyristoyl 90 phospholipid bilayers with choline, glycine, and serine headgroups. Using a NADH coupled 91 ATPase assay, we show that the 2C is active over a broad range of pH values and its 92 activity is stimulated by the presence of acetate while being slightly inhibited by monovalent 93 salts. Notably, we observe a ≈60-fold activation of 2C ATPase activity by the fully processed 2C 94 as compared to the 2BC3AB polyprotein precursor, suggesting that proteolytic processing of the 95 viral polyprotein plays a major role in activating 2C. 96 97 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C MATERIALS AND METHODS 98 Cloning. Constructs were cloned from pUC57-2BC3AB, a bacterial codon optimized 99 synthetic gene made by Genscript Inc. (www.genscript.com) encoding for poliovirus 2BC3AB 100 followed by a C-terminal GSSS-His6 tag sequence. The construct includes a SacII restriction site 101 immediately preceding the start codon and a NotI site following the His6 sequence. Initially, the 102 2C-containing constructs were cloned into a pET26 expression vector developed by the Cameron 103 laboratory for expression of poliovirus 3Dpol (40). With this vector the protein is initially 104 translated as a fusion with an N-terminal ubiquitin domain (Ub), but this domain is cleaved off 105 by the ubiquitin-specific carboxyl-terminal protease Ubp1 that is co-expressed in the E. coli 106 pCG1 strain, resulting in a processed polymerase with a N-terminal glycine residue that mimics 107 the natural cleavage of the poliovirus polyprotein at Gln-Gly sites. The 3Dpol gene was removed 108 from this vector and replaced with the gene for 2BC3AB-GSSS-His6 via the SacII and NotI sites, 109 resulting in pET26-Ub-2BC3AB-GSSS-His6. 5’and 3’deletions within the open reading frame 110 were then used to generate expression vectors for Ub2BC, Ub2C, Ub2C3A and Ub3AB. A 111 second set of these expression vectors where the N-terminal ubiquitin fusion was replaced by a 112 single methionine start codon were also generated, and the integrity of all constructs was verified 113 by sequencing. 114 For the in vitro translation experiments we found that the pET26-based constructs resulted in 115 very low protein yield as compared to the Elongation Factor Ts produced from the pEFT control 116 plasmid provided in the EasyXpress kit (Qiagen Inc.). Further analysis showed this to be the 117 result of inefficient RNA transcription, which can have a major effect on E. coli based in vitro 118 translation systems due to high endogenous RNase activity. To increase RNA levels in the 119 reactions, we cloned all our constructs into the pEFT control vector by simply replacing the 120 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C elongation factor Ts open reading frame. The subsequent mutations to the predicted AAA+ 121 ATPase NTP binding motifs in 2C were made by the Quickchange mutagenesis protocol to 122 produce K135A (Walker A), D177A (Walker B), and T222A and N223A (Motif C/Sensor-1) in 123 both the pEFT-Ub2C and pEFT-Ub2BC vectors. 124 Nanodisc Assembly. Nanodiscs were expressed and assembled with the MSP1E3D1 protein 125 (41) and DMPC lipids (unless noted) using 1:150 molar ratio of MSP1E3D1:DMPC as per the 126 protocols provided by the Sligar laboratory (sligarlab.life.uiuc.edu). Briefly, MSP1E3D1 was 127 expressed from the vector distributed by Addgene as plasmid number 20066 128 (www.addgene.org/20066) and purified by nickel affinity chromatography. The N-terminal His6 129 tag on the purified MSPs was removed using TEV protease that also contained a His6 tag. After 130 cleavage, the TEV/MSP mixture was run over a second nickel affinity column, from which the 131 only the MSPs lacking the His6 tag were collected in the unbound fraction. MSP and DMPC 132 were then co-solubilized using cholate detergent, cholate was removed using hydrophobic gel133 permeation BioBeads (Biorad, Inc.) to initiate nanodisc assembly, and the resulting discs were 134 purified over a Superdex S-200 gel filtration column where elution also confirmed that the 135 nanodiscs were of the expected size with a Stokes diameter of ≈12 nm (42). Nanodiscs 136 containing DMPG or DMPS phospholipids were assembled using the same MSP protein and 137 assembly protocols, with the difference that lipid solubilization and assembly were carried out 138 just above the appropriate lipid transition temperatures of 23°C for DMPC and DMPG and 35°C 139 for DMPS. We attempted to assemble nanodiscs using DMPE, but they did not remain soluble 140 when brought below the 50°C transition temperature. All lipids were purchased from Avanti 141 Polar Lipids (www.avantilipids.com). 142 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C In vitro Expression and Purification. 2C and the other 2BC3AB-based proteins were 143 produced using the Invitrogen Expressway in vitro bacterial expression system. 0.1-2 mL protein 144 expression reactions were incubated in an Eppendorf Thermomixer at 30°C and 1200 rpm for 145 three hours using 1 μg of DNA per 100 μL final reaction volume (RV) and nanodiscs were added 146 to a final concentration of 15 μM. Half of the required small molecule “feed mix” was added 30 147 minutes after starting the reaction and the other half was added after two hours. The completed 148 reactions were then centrifuged for 10 minutes at room temperature. The supernatant was batch 149 purified with one-quarter RV of settled Ni-NTA resin (Qiagen) pre-equilibrated in low imidazole 150 buffer (50 mM Tris, pH 8.0, 300 mM NaCl, 10 mM Imidazole, 0.02% NaN3 and 10-20% 151 glycerol) by incubation at 4°C for at least 30 min. After mixing, the supernatant was removed 152 and the resin washed with three RVs of low imidazole buffer. Protein was batch eluted in 153 multiple sequential quarter-RV fractions of high-imidazole (500 mM) buffer. TCEP reducing 154 agent (Pierce) was added to each elution to a final concentration of 10 mM, samples were 155 analyzed by SDS-PAGE, and elution fractions containing significant amounts of protein were 156 combined, aliquoted and stored at -80°C for future use. Note that since the His6 tag had 157 previously been cleaved from the MSPs, only nanodiscs bound to the 2C constructs are co158 purified in the elution fractions and this represents a small portion of the 15 μM nanodiscs that 159 were added to the translation reaction. 160 Several constructs were later subjected to more rigorous rounds of purification in order to 161 reduce the amount of contaminant proteins that co-eluted with 2C during the batch purification 162 process. In this purification process, 1-2 mL (final RV) of the Expressway reaction were purified 163 by liquid chromatography using an AKTA Purifier System first over a 1 mL HisTrap HP Nickel 164 resin column (GE Life Sciences) in the same low and high imidazole buffers used for batch 165 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C purification. Peak fractions were pooled, diluted 5-fold and further purified over a 1 mL HiTrap 166 Q HP ion exchange column (GE Life Sciences) with a gradient of low to high salt Q buffer (25 167 mM Tris, pH 8.5, 50 mM/1 M NaCl or KCl, 20% glycerol). 2C protein typically eluted in ≈300 168 mM salt and was then stored as stated above. 169 2C Yield and Concentration determination. Determination of the specific activity of each 170 ATPase construct requires knowing the precise concentration of ATPase in the purified samples. 171 Since the purified samples contained the 2C construct, the nanodisc MSP, and some contaminant 172 proteins from the batch-based purification scheme, we could not accurately determine the 2C 173 concentration by simple 280 nm absorbance or other overall protein content assays. We therefore 174 determined the concentrations of both the viral proteins and MSP by separating them on 12-15% 175 SDS-PAGE gels alongside known concentrations of the ProSieve Unstained Protein Marker II 176 standards (Lonza Group, Ltd.). The gel was stained with IRDye Blue Protein Stain (LiCor 177 Bioscience) and imaged on an Odyssey Infrared Imaging System at 700 nm. The intensities of 178 the protein standard bands were quantified and plotted against total amount of protein in each 179 band to yield a linear relationship between staining intensity and protein content (Figure 2). The 180 protein concentration of the in vitro expressed His-tagged viral proteins and the nanodisc MSPs 181 that copurified with them were then obtained by interpolating their band intensities on this 182 standard curve. 183 ATPase Activity assay. The rate of ATP hydrolysis was determined using a ATP/NADH 184 coupled assay (Figure 5A) combined with an ATP regeneration system that allows us to assay 185 steady state 2C activity over a wide range of ATP concentrations (43). In this assay the 186 conversion of ATP → ADP + Pi by an ATPase is coupled to ATP regeneration via the pyruvate 187 kinase (PK) catalyzed conversion of phosphoenolpyruvate (PEP) + ADP → pyruvate + ATP and 188 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C the lactate dehydrogenase (LDH) catalyzed conversion of pyruvate + NADH → lactate + NAD . 189 The rate of ATP hydrolysis is then monitored by the loss of NADH absorbance at 340 nm 190 (Figure 5B) with a molar extinction coefficient of 6.22 mM cm. Coupled reactions consisted 191 of 5 mM PEP, 0.024 U/μL PK, 0.036 U/μL LDH, and 0.4 mM NADH, resulting in an initial 192 absorbance of ≈2.0 at 340 nm that is linearly reduced to ≈0.25 upon complete oxidation of the 193 NADH as a result of ATPase turnover. 194 The ATP hydrolysis assays were carried out at 37°C in a Varian Bio50 spectrophotometer 195 with a thermostatted multi-cell holder using 0.5-40 nM 2C-domain in a base buffer 196 containing 75 mM KCl, 50 mM MgOAc, 20 mM HEPES (pH 7.0), 400 μM NADH, and 5 mM 197 DTT. Steady state data were collected for 30-60 minutes with points measured every ≈30-45 198 seconds, depending on how many assays were being conducted in the 18-cell multicell holder at 199 the time. Experiments to characterize the monovalent salt, pH, and magnesium dependence of 200 2C activity were done by varying individual components of this buffer as needed. To verify 201 that the buffer conditions tested did not adversely affect the coupled reporter system assay, 0.2 202 mM ADP was directly added to control reactions and the subsequent loss of NADH absorbance 203 was instantaneous and always significantly faster than the rate of ADP production by 2C. 204 205 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C RESULTS 206 Lipid Dependence of Nanodisc Based in vitro Expression. Nanodiscs were critical for viral 207 protein synthesis and we could not detect poliovirus proteins from expression reactions done in 208 the absence of nanodiscs (Figure 3). We expressed both Ub2BC and Ub2C in the presence of 209 nanodiscs made with dimyristoyl phospholipids attached to choline (DMPC), serine (DMPS) or 210 glycerol (DMPG) head groups and found there were marked differences in the expression levels 211 of the different proteins in the presence of the different lipids. Ub2BC expressed the best with 212 DMPC and DMPS nanodiscs, and expression levels were reduced about 2-fold in the presence of 213 DMPG nanodiscs (Figure 3A). Alternatively, the best Ub2C expression was seen with DMPG 214 discs and yield was reduced ≈2-3-fold when in the presence of DMPS or DMPC (Figure 3B). 215 Activity assays showed that the headgroup type had up to ≈3-fold effects on the ATPase kcat 216 values and the lipid giving maximal activity varied depending on the polyprotein being assayed; 217 Ub2C shows the highest activity with DMPC, Ub2BC prefers DMPG, and both proteins exhibit 218 half of their respective maximal activities with DMPS nanodiscs (Figure 3C). Ultimately, we 219 decided to use DMPC nanodiscs for further experiments as Ub2C demonstrated the highest 220 ATPase activity on these discs and Ub2BC activity was only reduced by ≈25% on DMPC discs 221 as compared to its maximal activity observed on DMPG discs. 222 Expression of 2C Constructs. In the presence of DMPC nanodiscs, the average final batch 223 purification yield for the various protein constructs was ≈11 μg per mL of translation reaction, 224 with the different constructs varying from 1–40 μg/mL in their yields. The lowest yield was 225 observed for 2C alone and this protein had a strong tendency to aggregate into a pelleting 226 fraction, but yield was improved by including polyprotein neighbor proteins, i.e. 2B and 3A 227 (Figure 4A). N-terminal Ub fusion proteins generally resulted in higher expression levels and 228 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C better yield of soluble protein (Figure 4B). Indeed, the soluble protein yield was typically lowest 229 for 2C at ≈1 μg per mL of in vitro translation reaction, but this was increased ≈5-fold for Ub2C, 230 and even further by the addition of viral polyprotein neighbors, resulting in up to ≈40 μg per mL 231 of reaction mixture for Ub2BC3AB. Thus, we primarily used the Ub fusion constructs for our 232 ATPase activity assays. We also made alanine point mutations to the AAA ATPase motifs 233 predicted to exists within the 2C domain: Walker A (K135A), Walker B (D177A), and motif C 234 (T222A, N223A). These mutants demonstrated a 2-30 fold increase in protein yield when 235 compared to the wild type, with the most significant increase seen for the Walker A mutant. 236 Furthermore, Ub3AB expressed at higher levels than any of the 2C-containing constructs, 237 resulting in up to 60 μg of protein per mL of in vitro expression reaction. 238 To reduce the probability of having multiple viral proteins associated with a single disc, the 239 in vitro translation reactions were carried out in the presence of a large excess nanodiscs (15 240 μM), only 2-5% of the which were recovered by co-purification with the expressed hexahistidine 241 tagged 2C constructs. Quantitation of the purified protein bands from multiple expressions 242 showed overall mole ratios of one to two 2C constructs per pair of MSPs needed for each 243 nanodisc. The protein-nanodisc complexes were soluble and little sample loss was observed 244 upon centrifugation at 4°C, except for the 2C protein that was prone to precipitation. This was 245 exacerbated in the FPLC purified samples that had a clear tendency to aggregate upon storage, 246 but was improved somewhat by the addition of the N-terminal ubiquitin fusion. Note that we 247 were able to stoichiometrically cleave the Ub domain from Ub2C using exogenously added 248 Ubp2 protease (generously provided by Dr. Robert Cohen), consistent with the fusion protein 249 being properly folded (data not shown). This is in contrast to our earlier attempts to express 2C 250 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C as a GST fusion in insect cells where the resulting protein was not very soluble and could not be 251 cleaved from GST, suggesting those proteins may have been misfolded or aggregated. 252 Initially, all the 2C constructs were only batch purified using nickel affinity resin; however, 253 we consistently noticed a ≈70 KDa contaminant protein co-eluting with the purified protein. This 254 band was excised from a SDS-PAGE gel, analyzed by liquid chromatography tandem mass 255 spectrometry (LC-MS/MS) at the Proteomics and Metabolomics Facility at Colorado State 256 University, and identified as the bacterial chaperone protein DnaK (Hsp70), a known protein 257 folding chaperone and ATPase (44). Concerned that DnaK could be contributing to the observed 258 ATPase activity in our assays, we subjected several constructs to a more rigorous FPLC-based 259 purification scheme that included both a nickel affinity column and an ion exchange column 260 (Figure 4C,D). Ub2C, Ub2BC, Ub2BC3AB, Ub3AB and Ub2C-K135A were all purified in this 261 manner and used for the final Michaelis-Menten kinetic analyses (Table 2). Batch purified 262 proteins were used to determine ATPase assay conditions and for initial kinetic comparisons 263 (Table 1). 264 ATPase Assay. Enzymatic activity was determined using a coupled assay in which the rate 265 of ATP hydrolysis is linked to a stoichiometric conversion of NADH to NAD (Figure 5A). 266 Figure 5B shows the time dependence of NADH depletion measured by absorbance at 340 nm in 267 an assay containing Ub2BC with ATP concentrations ranging from 0 to 10 mM. The lowest ATP 268 concentration tested in this experiment, 0.25 mM, results in a NADH oxidation rate that is about 269 twice that of background, and the rates then increase with increasing ATP concentration. At 4 270 mM ATP and above the total amount of ATP hydrolyzed in 30-60 minutes often exceeded the 271 0.4 mM NADH added to the assay and the absorbance data therefore reached a minimal signal 272 plateau at ≈0.25 absorbance units. The raw data curves do not have the same Y-intercept because 273 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C there is contaminating ADP in the ATP solution that is immediately converted back to ATP at 274 the expense of NADH upon assay assembly. The concentration of this contaminating ADP 275 increases along with that of the ATP, resulting in lower starting absorbance values at higher ATP 276 concentrations, as shown in Figure 5B. 277 Salt, pH, and Temperature Dependence. To determine the optimal conditions for 2C 278 ATPase activity assays we examined the rate of ATP hydrolysis by batch-purified Ub2C under 279 different monovalent salt (KCl and NaCl), magnesium, acetate, pH, and temperature conditions 280 (Figure 5C-F). To verify that the temperature and pH conditions being tested were not affecting 281 the enzymatic activity of the PK and LDH components of the coupled assay, ADP was added to 282 a final concentration of 0.2 mM, corresponding to half of the starting NADH concentration. In 283 every case, the loss of NADH absorbance at 340 nm in the control reactions was immediate, 284 indicating that turnover within the reporter system was not rate limiting. 285 The activity of Ub2C bound to nanodiscs showed a gradual but weak monovalent salt 286 dependence with a fairly flat plateau between 25 and 100 mM KCl and a drop in activity at salt 287 concentrations higher than that (Figure 5C). The effect was small in the presence of KCl, with 288 activity only reduced by ≈10% in the presence of 200 mM salt, while NaCl exhibited a greater 289 inhibitory effect with activity decreasing linearly with salt concentration to reach ≈60% activity 290 in 200 mM NaCl. We opted to use 75 mM KCl as a standard condition for further ATPase assays 291 as it demonstrated only a ≈5% decrease in activity and is fairly close to physiological salt 292 concentrations. We examined the pH dependence of batch purified Ub2C activity at eleven 293 different values and found a fairly flat activity profile from pH 6 to 7 followed by a decrease in 294 activity as the pH was increased to 9 (Figure 5D). 295 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C The activity of membrane proteins is often affected by the fluidity of the bilayer and can 296 therefore be influenced by lipid phase transitions, and conversely the presence of proteins can 297 affect phase transition temperatures. The normal phase transition temperature of DMPC is 298 ≈23°C, but it is increased by 3–4°C in the context of a nanodisc (45). To examine if a lipid phase 299 transition near 27°C affected 2C function we measured ATPase activity at a range of 300 temperatures between 15 and 40 °C. This was done using both Ub2C that is anchored to the 301 membrane via an amphipathic surface helix and Ub2BC that is a more integral membrane protein 302 due to the addition of a pair of predicted integral membrane segments. The data show an 303 essentially linear temperature dependent increase in activity for both proteins with a slight 304 inflection point in the vicinity of 30°C that may correspond to a weak effect from the DMPC 305 phase transition of a nanodisc bilayer with a bound 2BC or 2C protein (Figure 5E). The ≈3-fold 306 increase in ATPase activity between 15-40 °C is in contrast with prior studies of GST-2C that 307 have shown a drastic loss of ≈25% activity between 37 oC and ≈40 oC (28). To ensure that the 308 rates were not reflecting assay reagent instability (e.g. PEP, NADH, ATP), the temperature 309 dependence experiments were repeated without the addition of Ub2C and these background rates 310 were subtracted from the observed Ub2C rates to yield the data shown in Figure 5E (correction 311 was 5-10% of observed signal). 312 Magnesium and Acetate Dependence. The presence of magnesium acetate had a substantial 313 stimulatory effect on the batch-purified Ub2C ATPase activity and this effect was further 314 attributed to the addition of acetate. We initially carried out experiments using MgOAc in a fairly 315 standard manner of maintaining the Mg concentration 1 mM higher than the ATP 316 concentration, but discovered that we could not obtain a reliable Vmax for ATP hydrolysis 317 without going to unusually high ATP concentrations in the millimolar range. We then 318 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C independently examined the magnesium and the acetate dependence of ATP turnover by titrating 319 magnesium acetate, magnesium chloride, or potassium acetate in the reaction. The results show a 320 clear stimulation by both magnesium and potassium acetate with activation midpoints of 10-15 321 mM and saturation at ≈50 mM concentrations; in contrast, magnesium delivered as MgCl2 322 strongly inhibited activity (Figure 5F). We also verified that the magnesium acetate activation 323 was not due to low level contamination by another divalent metal, such as Zn, that could 324 perhaps be binding to the Cys-rich zinc finger-like motif in 2C. To test for this possibility we 325 added a low concentration of EDTA (1 mM) that would preferentially chelate Zn or other 326 divalent ions even in the presence of 50 mM Mg (46, 47) and observed that this did not 327 significantly change the ATPase rates (data not shown). High concentrations of magnesium 328 chloride, magnesium acetate, or potassium acetate did not affect the response from the coupled 329 assay reporter system. 330 Michaelis-Menten Kinetics of 2C Constructs. To examine the extent to which proteolytic 331 processing modulates the ATPase activity of the core AAA+ domain within 2C, we determined 332 the Michaelis-Menten enzyme kinetic parameters of ubiquitin fusion proteins for 2BC, 2C, 333 2C3A, 2BC3AB and 3AB and the non-ubiquitin versions of 2BC and 2C (Figure 6A,B). We also 334 made a series of mutations to the putative AAA+ ATPase and analyzed their effects on ATPase 335 activity in the context of both Ub2C and Ub2BC (Figure 6C,D). Based on the characterization 336 experiments described above, these assays were carried out using the coupled ATP/NADH assay 337 containing 0.5–6 nM 2C protein, 0.1–20 mM ATP, 50 mM MgOAc, and 75 mM KCl at pH 7.0 338 and 37°C. The observed ATP hydrolysis rates were converted to specific activities (μmol 339 ATP/μmol protein/minute, hereafter referred to as min) by normalizing to the protein 340 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C concentrations determined by quantitative measurements of band intensities on SDS-PAGE gels 341 (Figure 2). 342 Initial ATPase assays were carried out using batch purified protein; however, these samples 343 contained the contaminant protein ATPase DnaK, a bacterial chaperone. DnaK has a reported 344 Km of ≈1 μM and a kcat ranging from 0.03-3 min depending on the type of peptide substrate and 345 whether cofactor proteins are present (44). Since our ATPase assays were carried out at 0.1-15 346 mM ATP concentrations that are well above the Km for DnaK, we chose to semi-quantitatively 347 background correct the data by using a floating offset for the Michaelis-Menten curve fits, thus 348 making the assumption that this offset represents a Vmax rate for any contaminant DnaK in the 349 reaction. This background correction represented only ≈3% of the maximal 2C activity for the 350 batch purified proteins and the resulting data and kinetic parameters are shown in Figure 351 6A,C&D; and Table 1. When the same curve fit was used to determine the Michaelis-Menten 352 parameters for the FPLC purified proteins the fitted offset values were within error of zero, 353 indicating that these protein preparations did not contain contaminant ATPase activity. The data 354 obtained using the FPLC purified proteins are shown in Figure 6B and the resulting kcat and Km 355 values for Ub2C, Ub2BC, Ub2BC3AB, Ub3AB and Ub2C-K135A are reported in Table 2. 356 Proteins obtained by both purification methods demonstrate very similar activity patterns 357 among the various 2C constructs and mutants and these data show that proteolytic processing on 358 both sides of the core 2C domain modulate its ATPase activity. For the more reliable FPLC 359 purified proteins (Figure 6B, Table 2), we found that Ub2BC3AB has the lowest turnover rate at 360 ≈160 min, this increased modestly to 440 min upon removal of 3AB to yield Ub2BC, and 361 then increased greatly upon removal of the 2B domain to give the highest kcat of ≈9,000 min for 362 Ub2C. This 2C rate corresponds to a ≈20-fold activation compared to Ub2BC and ≈60-fold 363 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C increase compared to Ub2BC3AB. Only background activity of 5 min is observed for Ub3AB, 364 which lacks the 2C domain. The ATPase activity modulation is primarily due to changes in kcat 365 as the Km values of the purified proteins are consistent at ≈1.1±0.2 mM ATP for all the 366 constructs. Note that low yield of soluble Ub2C from the FPLC based protocol made protein 367 quantitation less reliable and this in turn increases the error of the specific activity measurements 368 for the Ub2C protein to ≈10%. 369 Among the mutations of the predicted canonical AAA+ ATPase motifs, the Walker A 370 mutation (K135A) has the most substantial effect on 2C ATPase activity with kcat being reduced 371 by 90-95%. Interestingly, the K135A only reduces activity by ≈50% in the context of 2BC, 372 suggesting that the active site geometries of 2C and 2BC may be slightly different. The 373 remaining mutations in the Walker B (D177A) and motif C (T222A and N223A) regions were 374 not as deleterious and their kcat values are reduced to 40-80% of those observed for the wild type 375 proteins in the batch purified samples. 376 377 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C DISCUSSION 378 Among the membrane anchored poliovirus proteins, both 2C and its precursor 2BC are 379 involved in forming the host membrane structures that serve as the site of viral replication (9380 12); however, the membrane binding, oligomerization, and aggregation propensities of the 381 picornaviral 2C proteins generally make them very difficult to express and purify for 382 biochemical and structural studies. In this work we have shown that nanodisc bilayers can be 383 used to chaperone the expression of soluble poliovirus 2C by providing membrane surfaces that 384 allow for proper folding and help reduce aggregation by sequestering the proteins on separate 385 bilayers, allowing us to carry out basic biochemical characterization of the enzyme. Processing 386 of the viral polyprotein can be important for regulating protein function, as has been shown for 387 3D where processing from the 3CD protease precursor protein activates the polymerase and 388 even minor changes to its N-terminal residues can have significant deleterious effects on activity 389 (4, 5, 48). To determine whether proteolytic processing plays a similar role in mediating 2C 390 behavior, we also examined the ATPase activity of 2C when expressed as a known functional 391 precursor protein (2BC), with a polyprotein neighbor (2C3A), or in the context of the entire 392 surrounding polyprotein (2BC3AB). 393 Nanodisc-Bound 2C is a Highly Active ATPase. The in vitro expression results clearly 394 show that nanodiscs effectively chaperone viral protein expression, resulting in highly active 2C 395 that could be purified as a 2C-nanodisc complex whose maximal turnover rate ≈200-fold higher 396 than that previously reported while retaining an ATP Km of ≈0.6 mM that is quite similar to the 397 0.5 0.7 mM values reported for both poliovirus GST-2C and FMDV 2C proteins (28, 31). 398 Previous studies of poliovirus 2C expressed as a GST fusion (28) and FMDV 2C expressed with 399 an N-terminal truncation (31) have reported 50% inhibition of ATPase activity in the presence of 400 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C 20 and 50 mM NaCl, respectively. In contrast, we found a weaker salt dependence that resulted 401 in only a ≈10% reduction of ATPase activity in the presence of 200 mM KCl. Sodium had a 402 stronger inhibitory effect, but even at 200 mM NaCl the nanodisc anchored proteins retained 403 ≈60% activity (Figure 5C). When testing the magnesium dependence, we found that Mg in the 404 form of magnesium acetate stimulated ATPase activity while Mgin the form of MgCl2 405 inhibited 2C activity (Figure 5F), the latter of which is consistent with prior findings of GST-2C 406 fusion construct activity (28). Further analysis showed that the activation was due to acetate 407 stimulation of 2C, where saturatable maximal 2C activity was observed at ≈50 mM 408 concentrations of either potassium acetate or magnesium acetate (Figure 5F). The origin of this 409 stimulatory effect is not yet known, but ligand-induced activation is common among AAA+ 410 ATPases where the presence of substrates is often necessary to promote oligomerization and 411 activity (33), suggesting that high concentrations acetate may mimic a natural ligand. 412 We also found that phospholipid composition can modulate expression levels and ATPase 413 activities of both 2C and 2BC, although the 2to 3-fold effects on each are fairly minor (Figure 414 3). Expression of the peripherally membrane anchored 2C was highest with DMPG lipids while 415 2BC preferred DMPS or DMPC. While subtle, the combined ≈10-fold effects of phospholipid 416 composition on expression levels and ATPase activity suggest that lipid headgroups can play a 417 role in modulating host membrane interactions with 2C and 2BC. This could perhaps influence 418 the membrane rearrangements that take place during poliovirus infection, and the shift toward 419 DMPS for higher 2BC expression is consistent with prior studies showing that 2B preferably 420 associates with anionic lipids such as phosphatidylserine and phosphatidylinositol that are 421 enriched in the Golgi and ER membranes targeted by the replication proteins (14). 422 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C Electron microscopy data of both poliovirus and FMDV 2C proteins showing hexameric ring 423 structures characteristic of this class of enzymes (30, 31) and structural modeling using the Phyre 424 protein fold recognition server (32) indicates with high confidence that 2C has homology to the 425 superfamily III AAA+ ATPases. We mutated canonical residues within the characteristic Walker 426 A, Walker B and motif C sequences (34) and found that they all reduced activity (Figure 6, 427 Tables 1 and 2). In particular, K135 in the Walker A motif plays an important role in positioning 428 the ATP γ phosphate at an inter-subunit contact, linking oligomerization and ATP hydrolysis 429 (49). Mutating K135 to an alanine not only abolished activity but also significantly increased 430 Ub2C solubility and resulted in ≈30-fold increased protein yield from the FPLC purification 431 protocol, all of which are consistent with AAA+ domain oligomerization. Interestingly, the same 432 mutation was not as deleterious in the context of 2BC, where it resulted in only a ≈50% activity 433 loss, suggesting that the added 2B domain may not only prevent oligomerization but also cause 434 the ATPase domain to utilize a slightly different active site geometry that accounts for its lower 435 turnover rate. Furthermore, the Walker B and motif C/Sensor-1 mutations generally affect ATP 436 hydrolysis rates significantly more than ATP binding (34, 49), which is consistent with the 437 relatively small changes we observe in the ATP Km values for the various mutants. 438 Based on our findings, we believe that the nanodisc bound 2C protein is indeed forming 439 multimeric structures that are necessary for AAA+ ATPase activity, although we cannot say if 440 these are discrete hexamers or larger oligomers based on our data. Nanodiscs are quite effective 441 at preventing oligomerization of integral membrane proteins by sequestering the transmembrane 442 segments into individual discs, but it is important to note that 2C is a peripheral membrane 443 protein and that its oligomerization is based on protein-protein interactions outside the lipid 444 bilayer. As a result, the discs may chaperone proper protein folding, but do not necessarily 445 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C prevent oligomerization. Our expression and purification data indicate that the isolated 2C 446 protein is significantly more prone to aggregation or oligomerization than the larger 2BC and 447 2BC3AB constructs. We postulate that the propensity of 2C to oligomerize will bring multiple 448 nanodiscs together in close proximity, creating a high local concentration that facilitates fusion 449 of the disc membranes into larger bilayer structures and reducing solubility, in effect defeating 450 the advantage of using the nanodiscs. Consistent with this, the K135A mutation that is predicted 451 to disrupt oligomerization results in more soluble material and a significant increase in protein 452 yield. It is also possible that 2C itself may be acting to change the structure of the nanodiscs as it 453 is known to be involved in membrane rearrangement (9-11). 454 Proteolytic Control of 2C ATPase Activity. Comparative analysis of enzyme rates by the 455 different protein constructs shows that poliovirus 2C ATPase activity is strongly modulated by 456 the presence of its polyprotein neighbors. Starting with the 2BC-3AB protein, ATPase activity is 457 increased ≈3-fold by removal of 3AB to generate 2BC and then another ≈20-fold upon removal 458 of 2B to generate the fully processed 2C protein. This reflects an overall ≈60-fold increase in 459 activity for the fully processed 2C as compared to its fully embedded precursor form and 460 indicates that polyprotein neighbors exert allosteric effects on the core 2C domain that can 461 modulate the rate of ATP hydrolysis. It is also notable that both the 2C and Ub2C proteins have 462 higher activities than 2BC, showing that the allosteric effects are specific to having a 2B domain. 463 This is in contrast to the 3Dpol proteolytic activation event where any N-terminal modification, 464 even adding a single amino acid residue, will inhibit activity. 465 Overall, the results from this study lead to a model in which precursor forms of 2C, such as 466 2BC and 2BC3AB that contain all the membrane binding elements of the viral polyprotein, 467 assemble into the viral replication complex in a precursor form. During proteolytic processing, 468 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C the viral proteases carry out multiple cleavages of the polyprotein, the first of which is 2A 469 mediated cleavage of the P1 region from the polyprotein and this is followed by the 3C/3CD 470 mediated cleavages of the P2 and P3 regions (50). The exact order in which the P2 and P3 471 proteins are cleaved is not well established, but the intact P2-P3 protein is rarely observed during 472 infection and it is thought that the P2-P3 junction (i.e. 2C-3A linkage) is cleaved early during 473 processing. This would make 2BC the predominant precursor form, wherein 2C ATPase activity 474 is kept at a low basal level because the presence of the 2B domain sterically prevents 2C from 475 forming the oligomeric structures required for maximal ATP hydrolysis. This pre-assembly of 476 2BC proteins on the membrane surface is then followed by processing of the 2B-2C junction, 477 triggering 2C to assemble into larger oligomeric structures that increase ATPase activity ≈20478 fold. The true purpose of this ATPase activity during viral replication is not yet known and not 479 addressed by our data, but it may play a role in RNA helicase or translocation activity, capsid 480 assembly, RNA packaging, or perhaps a combination of these functions. 481 In conclusion, we have shown that the membrane associated poliovirus proteins can be 482 synthesized in an active and soluble form when anchored to small nanodisc membrane bilayers. 483 The core domain within these proteins is the 2C ATPase, which is shown to exhibit high ATPase 484 activity over a broad range of monovalent salt concentrations, pH values, and temperatures. The 485 activity of this core domain can be modulated by the addition of either Nor C-terminal 486 polyprotein neighbors and the different polyprotein constructs exhibit slight activity preferences 487 for different lipid headgroups. Although issues with oligomerization and solubility continue to 488 present a clear obstacle to structural studies of poliovirus 2C that would require large amounts of 489 purified protein, our nanodisc based expression data do set the stage for experiments to further 490 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C understand the interactions of the picornaviral 2C and 2BC proteins with RNA and to study their 491 effects on membrane dynamics by biophysical methods. 492 493 494 495 496 497 498 499 500 501 502 503 504 ACKNOWLEDGEMENTS 505 We thank Megan Filbin and Grace Campagnola for critical reading of the manuscript. This work 506 was supported by NIH grant R01 AI-059130 to OBP and an Azari Graduate Fellowship to CLS. 507 508 FOOTNOTES 509 To whom correspondence may be addressed: [email protected] 510 The abbreviations used are: PV, poliovirus; FMDV, foot-and-mouth disease virus; Ub, ubiquitin. 511 512 on D ecem er 5, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Proteolytic Activation of Poliovirus 2C REFERENCES5131. Roberts, L. 2012. The Polio Emergency. Science 337:514-516.5142. De Jesus, N. 2007. Epidemics to eradication: the modern history of poliomyelitis. Virology515Journal 4:70-70.5163. Thompson, A. A., and O. B. Peersen. 2004. Structural basis for proteolysis-dependent517activation of the poliovirus RNA-dependent RNA polymerase. EMBO J 23:3462-3471.5184. Flanegan, J. B., and T. A. Van Dyke. 1979. Isolation of a soluble and template-dependent519poliovirus RNA polymerase that copies virion RNA in vitro. Journal of Virology 32:155-161.5205. Harris, K. S., S. R. Reddigari, M. J. Nicklin, T. Hammerle, and E. Wimmer. 1992.521Purification and characterization of poliovirus polypeptide 3CD, a proteinase and a precursor522for RNA polymerase. J Virol 66:7481-7489.5236. Belov, G. A., V. Nair, B. T. Hansen, F. H. Hoyt, E. R. Fischer, and E. Ehrenfeld. 2012.524Complex dynamic development of poliovirus membranous replication complexes. J Virol52586:302-312.5267. Limpens, R. W., H. M. van der Schaar, D. Kumar, A. J. Koster, E. J. Snijder, F. J. van527Kuppeveld, and M. Barcena. 2011. The transformation of enterovirus replication structures:528a three-dimensional study of singleand double-membrane compartments. mBio 2.5298. Schlegel, A., T. H. Giddings, Jr., M. S. Ladinsky, and K. Kirkegaard. 1996. Cellular530origin and ultrastructure of membranes induced during poliovirus infection. J Virol 70:6576-5316588.5329. Cho, M. W., N. Teterina, D. Egger, K. Bienz, and E. Ehrenfeld. 1994. Membrane533Rearrangement and Vesicle Induction by Recombinant Poliovirus 2C and 2BC in Human534Cells. Virology 202:129-145.53510. Teterina, N. L., A. E. Gorbalenya, D. Egger, K. Bienz, and E. Ehrenfeld. 1997.536Poliovirus 2C protein determinants of membrane binding and rearrangements in mammalian537cells. J Virol 71:8962-8972.53811. Aldabe, R., and L. Carrasco. 1995. Induction of membrane proliferation by poliovirus539proteins 2C and 2BC. Biochem Biophys Res Commun 206:64-76.54012. Suhy, D. A., T. H. Giddings, Jr., and K. Kirkegaard. 2000. Remodeling the endoplasmic541reticulum by poliovirus infection and by individual viral proteins: an autophagy-like origin for542virus-induced vesicles. J Virol 74:8953-8965.54313. Agirre, A., A. Barco, L. Carrasco, and J. L. Nieva. 2002. Viroporin-mediated membrane544permeabilization. Pore formation by nonstructural poliovirus 2B protein. J Biol Chem545277:40434-40441.54614. Agirre, A., M. Lorizate, S. Nir, and J. L. Nieva. 2008. Poliovirus 2b insertion into lipid547monolayers and pore formation in vesicles modulated by anionic phospholipids. Biochim548Biophys Acta 1778:2621-2626.54915. Nieva, J. L., A. Agirre, S. Nir, and L. Carrasco. 2003. Mechanisms of membrane550permeabilization by picornavirus 2B viroporin. FEBS Letters 552:68-73.551onDecemer5,2017bygesthttp/jvi.asm.rg/Downladedfom Proteolytic Activation of Poliovirus 2C 16. Paul, A. V., A. Molla, and E. Wimmer. 1994. Studies of a putative amphipathic helix in the552N-terminus of poliovirus protein 2C. Virology 199:188-199.55317. Echeverri, A. C., and A. Dasgupta. 1995. Amino Terminal Regions of Poliovirus 2C554Protein Mediate Membrane Binding. Virology 208:540-553.55518. Echeverri, A., R. Banerjee, and A. Dasgupta. 1998. Amino-terminal region of poliovirus5562C protein is sufficient for membrane binding. Virus Res 54:217-223.55719. Strauss, D. M., L. W. Glustrom, and D. S. Wuttke. 2003. Towards an understanding of the558poliovirus replication complex: the solution structure of the soluble domain of the poliovirus5593A protein. Journal of molecular biology 330:225-234.56020. Towner, J. S., T. V. Ho, and B. L. Semler. 1996. Determinants of membrane association561for poliovirus protein 3AB. J Biol Chem 271:26810-26818.56221. Yin, J., Y. Liu, E. Wimmer, and A. V. Paul. 2007. Complete protein linkage map between563the P2 and P3 non-structural proteins of poliovirus. J Gen Virol 88:2259-2267.56422. Teterina, N. L., E. Levenson, M. S. Rinaudo, D. Egger, K. Bienz, A. E. Gorbalenya, and565E. Ehrenfeld. 2006. Evidence for functional protein interactions required for poliovirus RNA566replication. J Virol 80:5327-5337.56723. Barton, D. J., and J. B. Flanegan. 1997. Synchronous replication of poliovirus RNA:568initiation of negative-strand RNA synthesis requires the guanidine-inhibited activity of protein5692C. J Virol 71:8482-8489.57024. Liu, Y., C. Wang, S. Mueller, A. V. Paul, E. Wimmer, and P. Jiang. 2010. Direct571interaction between two viral proteins, the nonstructural protein 2C and the capsid protein572VP3, is required for enterovirus morphogenesis. PLoS Pathog 6:e1001066.57325. Li, J. P., and D. Baltimore. 1990. An Intragenic Revertant of a Poliovirus-2c Mutant Has an574Uncoating Defect. Journal of Virology 64:1102-1107.57526. Mirzayan, C., and E. Wimmer. 1994. Biochemical Studies on Poliovirus Polypeptide 2C:576Evidence for ATPase Activity. Virology 199:176-187.57727. Rodriguez, P. L., and L. Carrasco. 1995. Poliovirus protein 2C contains two regions578involved in RNA binding activity. J Biol Chem 270:10105-10112.57928. Pfister, T., and E. Wimmer. 1999. Characterization of the nucleoside triphosphatase580activity of poliovirus protein 2C reveals a mechanism by which guanidine inhibits poliovirus581replication. Journal of Biological Chemistry 274:6992-7001.58229. Rodriguez, P. L., and L. Carrasco. 1993. Poliovirus Protein-2C Has Atpase and Gtpase583Activities. Journal of Biological Chemistry 268:8105-8110.58430. Adams, P., E. Kandiah, G. Effantin, A. C. Steven, and E. Ehrenfeld. 2009. Poliovirus 2C585protein forms homo-oligomeric structures required for ATPase activity. J Biol Chem586284:22012-22021.58731. Sweeney, T. R., V. Cisnetto, D. Bose, M. Bailey, J. R. Wilson, X. Zhang, G. J. Belsham,588and S. Curry. 2010. Foot-and-Mouth Disease Virus 2C Is a Hexameric AAA+ Protein with a589Coordinated ATP Hydrolysis Mechanism. Journal of Biological Chemistry 285:24347-24359.590onDecemer5,2017bygesthttp/jvi.asm.rg/Downladedfom Proteolytic Activation of Poliovirus 2C 32. Bennett-Lovsey, R. M., A. D. Herbert, M. J. Sternberg, and L. A. Kelley. 2008.591Exploring the extremes of sequence/structure space with ensemble fold recognition in the592program Phyre. Proteins 70:611-625.59333. Erzberger, J. P., and J. M. Berger. 2006. Evolutionary relationships and structural594mechanisms of AAA+ proteins. Annu Rev Biophys Biomol Struct 35:93-114.59534. Hanson, P. I., and S. W. Whiteheart. 2005. AAA+ proteins: have engine, will work. Nat596Rev Mol Cell Biol 6:519-529.59735. Bayburt, T. H., and S. G. Sligar. 2003. Self-assembly of single integral membrane proteins598into soluble nanoscale phospholipid bilayers. Protein Sci 12:2476-2481.59936. Civjan, N. R., T. H. Bayburt, M. A. Schuler, and S. G. Sligar. 2003. Direct solubilization600of heterologously expressed membrane proteins by incorporation into nanoscale lipid bilayers.601BioTechniques 35:556-563.60237. Katzen, F., J. E. Fletcher, J.-P. Yang, D. Kang, T. C. Peterson, J. A. Cappuccio, C. D.603Blanchette, T. Sulchek, B. A. Chromy, P. D. Hoeprich, M. A. Coleman, and W. Kudlicki.6042008. Insertion of Membrane Proteins into Discoidal Membranes Using a Cell-Free Protein605Expression Approach. Journal of Proteome Research 7:3535-3542.60638. Boldog, T., S. Grimme, M. Li, S. G. Sligar, and G. L. Hazelbauer. 2006. Nanodiscs607separate chemoreceptor oligomeric states and reveal their signaling properties. Proceedings of608the National Academy of Sciences 103:11509-11514.60939. Leitz, A. J., T. H. Bayburt, A. N. Barnakov, B. A. Springer, and S. G. Sligar. 2006.610Functional reconstitution of Beta2-adrenergic receptors utilizing self-assembling Nanodisc611technology. BioTechniques 40:601-602.61240. Gohara, D. W., C. S. Ha, S. Kumar, B. Ghosh, J. J. Arnold, T. J. Wisniewski, and C. E.613Cameron. 1999. Production of "Authentic" Poliovirus RNA-Dependent RNA Polymerase614(3Dpol) by Ubiquitin-Protease-Mediated Cleavage in Escherichia coli. Protein Expr Purif61517:128-138.61641. Denisov, I. G., B. J. Baas, Y. V. Grinkova, and S. G. Sligar. 2007. Cooperativity in617Cytochrome P450 3A4. Journal of Biological Chemistry 282:7066-7076.61842. Denisov, I. G., Y. V. Grinkova, A. A. Lazarides, and S. G. Sligar. 2004. Directed Self-619Assembly of Monodisperse Phospholipid Bilayer Nanodiscs with Controlled Size. Journal of620the American Chemical Society 126:3477-3487.62143. Barnett, R. E. 1970. Effect of monovalent cations on the ouabain inhibition of the sodium622and potassium ion activated adenosine triphosphatase. Biochemistry 9:4644-4648.62344. Russell, R., R. Jordan, and R. McMacken. 1998. Kinetic Characterization of the ATPase624Cycle of the DnaK Molecular Chaperone†. Biochemistry 37:596-607.62545. Denisov, I. G., M. A. McLean, A. W. Shaw, Y. V. Grinkova, and S. G. Sligar. 2005.626Thermotropic phase transition in soluble nanoscale lipid bilayers. J Phys Chem B 109:15580-62715588.62846. Nyborg, J. K., and O. B. Peersen. 2004. That zincing feeling: the effects of EDTA on the629behaviour of zinc-binding transcriptional regulators. Biochem J 381:e3-4.630onDecemer5,2017bygesthttp/jvi.asm.rg/Downladedfom Proteolytic Activation of Poliovirus 2C 47. Falke, J. J., S. K. Drake, A. L. Hazard, and O. B. Peersen. 1994. Molecular tuning of ion631binding to calcium signaling proteins. Quarterly reviews of biophysics 27:219-290.63248. Campagnola, G., M. Weygandt, K. Scoggin, and O. Peersen. 2008. Crystal structure of633coxsackievirus B3 3Dpol highlights the functional importance of residue 5 in picornavirus634polymerases. J Virol 82:9458-9464.63549. Wendler, P., S. Ciniawsky, M. Kock, and S. Kube. 2012. Structure and function of the636AAA+ nucleotide binding pocket. Biochim Biophys Acta 1823:2-14.63750. Palmenberg, A. C. 1990. PROTEOLYTIC PROCESSING OF PICORNAVIRAL638POLYPROTEIN. Annu. Rev. Microbiol. 44:603-623.639640641 onDecemer5,2017bygesthttp/jvi.asm.rg/Downladedfom Proteolytic Activation of Poliovirus 2C FIGURE LEGENDS642FIG. 1. Poliovirus Polyprotein Processing and Nanodisc Chaperoned Expression.643(A) Organization of the poliovirus single-stranded positive sense RNA genome encoding a single644open reading frame for the viral polyprotein. Dashed arrows indicate viral proteins known to645interact with the 2C and grey bars identify membrane binding (MB) regions.646 (B) Schematic representation of the five viral proteins that were generated for this study. All647 constructs contained a C-terminal hexahistidine tag and were made both with and without a648N-terminal ubiquitin domain (±Ub). Diagrammed below are known 2C functional domains,649including the NTP interaction regions that comprise the core AAA+ ATPase motifs.650(C) Depiction of the nanodisc based in vitro expression protocol showing the initial assembly of651discs from pure phospholipids and two copies of the amphipathic helix based membrane scaffold652protein (MSP) that are then added to bacterially based in vitro transcription/translation expression653reactions (derived with permission from (38, 39)).654 655656FIG. 2. Protein concentration determination.657(A) Protein concentrations were determined by running various dilutions of the ProSieve658Unstained Protein Marker II standard alongside dilutions of 2C protein constructs on reducing659acrylamide SDS gels. The gels were stained with LiCor IRDye Blue Protein Stain, imaged by660 scanning on a LiCor Odyssey infrared fluorescence imaging system, and multiple bands from the661standards, the 2C constructs, and nanodisc MSP were outlined for quantitation as shown.662 (B) Example standard curve generated from the correlation of the protein standards intensities663 (open circles) with the amount of protein loaded and then fit to a linear curve whose parameters664 are listed in the inset box. The concentrations of the various 2C construct dilutions were then665onDecemer5,2017bygesthttp/jvi.asm.rg/Downladedfom Proteolytic Activation of Poliovirus 2C interpolated on this line, as per the three proteins shown here (solid symbols), and averaged to666 obtain the concentration in the original sample. Independent gels with standards were analyzed667for every purified protein, allowing us to calculate specific activities in the ATPase assays based668on the amount of 2C protein added.669 670671FIG. 3. Phospholipid dependence of Ub2C and Ub2BC expression and activity.672(A,B) Batch nickel purification elution fractions E1 and E2 from 100 μL in vitro expression of673Ub2BC (A) and Ub2C (B) in the presence of nanodiscs assembled with DMPG, DMPS or674DMPC lipids or in the absence of nanodiscs entirely. Lane L contains ProSieve Unstained675Protein Markers II. (C) Relative expression levels and ATPase activities obtained with the676different nanodiscs for both Ub2C and Ub2BC proteins. ATPase assays were carried out at 37°C677in the presence of 1 mM ATP.678 679680FIG. 4. Purification of in vitro translated 2C proteins.681(A,B) Batch nickel resin elution fractions (E1-E4) from DMPC nanodisc chaperoned in vitro682expression of (A) wildtype 2C constructs and (B) N-terminal ubiquitin fusion constructs. The683 bands for the desired 2C protein are marked by single sided arrows and the nanodisc MSP684protein is denoted by double sided arrows. Constructs were expressed in 100 μL reactions,685purified in batch with Ni affinity resin, and 5 μL samples of each 25 μL elution fraction were686analyzed on 12% acrylamide SDS gels. The ladder molecular weight shown beside the 2BC3AB687 is the same for all gels except for Ub2C3A. (C,D) FPLC purification of Ub2C via HisTrap nickel688affinity (not shown) and HiTrap Q anion exchange columns. The primary 2C elution position is689 shown by the arrow on the HiTrap Q chromatogram (C) and SDS-PAGE analysis (D) of690onDecemer5,2017bygesthttp/jvi.asm.rg/Downladedfom Proteolytic Activation of Poliovirus 2C fractions 17-23 revealed two major proteins eluting immediately after 2C. Mass spectrometry691 was used to identify these proteins as bacterial DnaK and glucosamine fructose-6-phosphate692aminotransferase (GF6PA).693 694695FIG. 5. ATPase activity of 2C proteins bound to nanodisc bilayers.696 (A) Reaction scheme of ATPase assay in which ATP hydrolysis is coupled to oxidation of697 NADH via pyruvate kinase (PK) and lactate dehydrogenase (LDH). Loss of absorbance at 340698nm reflects the loss of NADH due to ATP hydrolysis. (B) Raw absorbance data for a set of699Ub2BC assays with various ATP concentrations that were done simultaneously using a Cary 50700Bio spectrophotometer with a thermostatted (37°C) multicell holder. (C) Monovalent salt701concentration dependence of Ub2C ATPase activity in the presence of NaCl or KCl. Reactions702were carried out with 3 mM ATP and 1 nM batch purified Ub2C protein and normalized to703highest activity in the absence of added salt. (D) pH dependence of Ub2C ATPase activity704examined with 2 mM ATP and 5 nM batch purified Ub2C, normalized to activity at pH 7.705(E) Temperature dependence of Ub2C and Ub2BC ATPase activity carried out with 2 mM ATP706and 5 nM batch purified protein. (F) Magnesium and acetate dependence of Ub2C ATPase707activity. The MgOAc data is a combination of three sets of assays that were carried out with7081 mM ATP and 2-4 nM batch purified Ub2C. MgCl2 and KOAc assays started with 75 μL709 reactions containing 2 mM ATP and 3 mM MgCl2 into which either MgCl2 or KOAc were710titrated with the total added volume added not exceeding 5 μL.711