A positively charged surface patch on the pestivirus NS3 protease module plays an important role in modulating NS3 helicase activity and virus production

Pestivirus nonstructural protein 3 (NS3) is a multifunctional protein with protease and helicase activities that are essential for virus replication. In this study, we used a combination of biochemical and genetic approaches to investigate the relationship between a positively charged patch on the protease module and NS3 function. The surface patch is composed of four basic residues, R50, K74 and K94 in the NS3 protease domain and H24 in the structurally integrated cofactor NS4APCS. Single-residue or simultaneous four-residue substitutions in the patch to alanine or aspartic acid had little effect on ATPase activity. However, single substitutions of R50, K94 or H24 or a simultaneous four-residue substitution resulted in apparent changes in the helicase activity and RNA-binding ability of NS3. When these mutations were introduced into a classical swine fever virus (CSFV) cDNA clone, a single substitution at K94 or a simultaneous four-residue substitution (Qua_A or Qua_D) impaired the production of infectious virus. Furthermore, the replication efficiency of the CSFV variants was partially correlated with the helicase activity of NS3 invitro. Our results suggest that the conserved positively charged patch on NS3 plays an important role in modulating the NS3 helicase activity invitro and CSFV production.

NS3 has serine protease and RNA helicase/nucleotide triphosphatase (NTPase) activity, and both are essential for virus replication [13,27,36,42,43]. Circumstantial evidence suggests that the protease and helicase/NTPase domains of NS3 are functionally interdependent [26,29]. NS3 protease activity requires NS4A PCS as a structurally integrated cofactor [36,49]. In the case of hepatitis C virus (HCV), the NS3 protease domain enhances its helicase activity [3,4]. The HCV NS3 helicase activity is modulated by NS5B, an RNA-dependent RNA polymerase (RdRP), and the protease domain is required for the interaction between NS3 and NS5B [1,48]. In the case of CSFV, a truncated NS3 protein (NS3Hel) containing only the helicase domain has been shown to exhibit similar NTPase activity and significantly decreased helicase activity when compared to the full-length NS3 (NS3fl) [33,39,42,43], thereby further demonstrating that the protease and helicase domains of NS3 are functionally coupled. However, the exact mechanism of this intramolecular regulation needs to be addressed further.
Our previous study revealed an intramolecular proteasehelicase interface with a positively charged groove in the pestivirus NS3 structure. Four basic residues (R50, K74 and K94 in NS3 and H24 in NS4A PCS ) in the protease part of the groove form a positively charged surface patch, which maybe coordinate with the helicase part of the groove to modulate the RNA-binding ability and helicase activity of NS3 [49]. In this study, we further investigated the relationship between this positively charged surface patch and NS3 function. Our data suggest that, in this natural proteasehelicase fusion protein, the positively charged surface patch of NS3 plays an important role in modulating NS3 helicase activity and infectious virus production.

Plasmid construction
To prepare the wild-type (WT) NS3 and its mutated proteins, two sets of NS3 constructs containing a single substitution or simultaneous substitution of the four basic residues (R50, K74 and K94 in the protease domain and H24 in NS4A PCS ) in the positively charged surface patch to alanine or aspartic acid (Fig. 1A) were generated by using an NS3 expression plasmid (pET28a-NS3 S163A /NS4A PCS ) as a template, using a QuikChange Site-Directed Mutagenesis Kit [49,50]. The variants generated by mutating the basic residue to alanine (A) or aspartic acid (D) are referred to as set 1 or set 2, respectively. When the four basic residues were simultaneously replaced by alanine or aspartic acid, the variant was correspondingly named Quad_A (Quadruple_A) or Quad_D (Quadruple_D). A helicase-only construct (NS3Hel, residues 204 to 683) and a variant harboring a K232A mutation in NS3 to abolish its ATPase and helicase activities were used as negative controls. We also introduced the mutations of both sets into a full-length CSFV infectious clone [20] as described previously [45] to investigate the effect of the basic residue substitutions on infectious virus production. All variants were confirmed by sequencing.

Protein expression and purification
The expression and purification of NS3 S163A /NS4A PCS and its variants were performed as described previously [22,49]. Briefly, E. coli strain BL21-CodonPlus (DE3)-RIL was transformed with the expression plasmid, and the bacteria were then cultured at 37°C in terrific broth (TB) medium containing 50 µg of kanamycin and 25 µg of chloramphenicol per ml. When the optical density at 600 nm (OD 600 ) of the culture reached 0.8, isopropyl-β-D-1thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM. After an additional incubation at 25

ATPase assay
The ATPase activity was measured using a malachite-greenbased method as described previously [39,43,49]. Briefly, the 90-μl reaction mixture except for the ATP substrate was incubated at 37 °C for 5 min. The reaction was initiated by addition of a 10-μl ATP solution to yield a final reaction mixture containing 10 nM NS3, 50 mM Tris (pH 7.5), 2.5 mM MgCl 2 , 50 mM NaCl, and 5 to 500 μM ATP. After an additional incubation at 37 °C for 15 min, the malachite green mixture (water, 0.081% [wt/vol] malachite green, and 5.7% [wt/vol] ammonium molybdate in 6 M HCl at a ratio of 3:2:1 [vol:vol:vol]) was added, and the absorbance was immediately measured at 630 nm on a Multiskan MK3 microplate reader (Thermo Fisher Scientific). Initial ATPase catalytic rates were determined based on the slope of the initial absorbance change and the reference standard curve of absorbance versus phosphate concentration determined independently. The observed ATP hydrolysis rates at various Schematic diagram of the CSFV polyprotein and the NS3/NS4A PCS construct (top) and the conserved amino acids in the basic surface patch of NS3 with its cofactor NS4A PCS (bottom). (B) A continuous surface groove (indicated by a green dashed circle) with high positive potential (blue) on the three-dimensional structure of the NS3-pssRNA complex (left). Scale bar: − 2 kT/e to 2 kT/e, where k is the Boltzmann constant, T is the temperature in Kelvin, and e is the charge of an electron. The black discontinuous line indicates the boundary between the protease and helicase/ATPase domains of NS3/ NS4A PCS . The colors in the ribbon structure are as follows: red, the basic residues R50, K74, K94 and H24; green, NS4A PCS ; wheat, the protease module; yellow, lime green, and cyan, the domains D1, D2, and D3, respectively; purple, pssRNA [49]. (C) Analysis of purified WT NS3 and its mutants by SDS-PAGE ◂ ATP concentrations were fitted to Michaelis-Menten kinetics to yield the ATPase parameters (K M app and k cat ).

Fluorescence polarization (FP)-based RNA binding assay
To measure the RNA-binding ability of NS3 and its variants, we designed pssRNA-2 by annealing T40 (mentioned above) and a 33-mer release strand (R33, 5′-UCC ACC AAU CAA GUA-UCU CGU AUG CAU GAU UGG -3′) labeled with a 5′-FAM for use in an FP-based RNA binding assay to estimate the affinity of each protein for RNA as described previously [28,31]. Reactions were performed in a 96-well plate (Corning; flat bottom, non-binding surface, black polystyrene). For each reaction, a 72-μl reaction buffer mixture was incubated at 37 °C for 5 min, and the reaction was initiated by adding NS3 to make the final reaction mixture containing 10 nM 5′-FAM-labeled pssRNA-2, 50 mM MOPS-NaOH (pH 7.0), 2.5 mM MgCl 2 , 50 mM NaCl, 1 to 1000 nM NS3 (with a molar ratio of NS3 to pssRNA-2 of 0.1:1 to 100:1). After incubation at 37 °C for 30 min, polarization was monitored using a Cytation 3 Cell Imaging Multi-Mode Reader (Bio Tek) by exciting at 485 nm (20 nm bandwidth) and measuring total fluorescence intensity and parallel and perpendicular polarized light at 528 nm (20 nm bandwidth). The G-factor (the instrument calibration factor) was calculated using readings from wells with 10 nM pssRNA-2 alone. The data at different NS3 concentrations ([S]) were fitted to the quadratic equation where the "A" represents the amplitude of the FP value change, 10 is the concentration of pssRNA-2 (10 nM), and K d is the dissociation constant of the NS3-RNA binding complex.

Virus rescue and titration
The virus was rescued as described previously [20,45,46]. Briefly, PK-15 cells were transfected with 2 μg of the WT CSFV cDNA clone or its variants containing two sets of mutations, using Lipofectamine 3000 (Invitrogen). After incubation at 37°C for 72 h, virus production was monitored by indirect immunofluorescence assay (IFA) using an anti-NS3 rabbit polyclonal antibody as the primary antibody [20] and an Alexa Fluor 488-conjugated secondary antibody (goat anti-rabbit IgG, Invitrogen). The culture supernatant was harvested and clarified by centrifugation, and virus titration was performed by IF staining [20] with anti-NS3 antibody in 96-well plates using the Reed-Muench method [32]. Virus titers were expressed as tissue culture infectious doses (50% endpoint, TCID 50 ) per milliliter.

RT-qPCR
Viral RNA copy numbers were determined using a reverse transcription quantitative PCR (RT-qPCR) [19,30]. PK-15 cell monolayers in 24-well plates were infected with the virus at an MOI of 0.001. Total RNA was extracted from the infected cells at 6, 12, and 24 hpi using a TaKaRa Min-iBEST Universal RNA Extraction Kit (TaKaRa) and 500 ng of total RNA was reverse transcribed using a ReverTra Ace qPCR RT Kit (TaKaRa), with a specific primer (5′-TAG CCT AAT AGT GGG CCT CTG-3′). Then, the cDNA transcribed from 50 ng of total RNA was analyzed by qPCR using a THUNDERBIRD Probe qPCR Mix kit (TaKaRa) with a 5′-FAM-labeled probe (5'-TCA GGT CGT ACT CCC ATC ACG TGG TGTGA-3') and the primers CSF-F174 (5′-ACA GGA CAG TCG TCA GTA GTTC-3′) and CSF-R345 (5′-TAG CCT AAT AGT GGG CCT CTG-3′). A known quantity of cDNA containing the 5'UTR of the CSFV genome was used as a standard for qPCR. The cycling parameters consisted of denaturation at 95 °C for 60 s, followed by 40 amplification cycles (95 °C for 15 s and 60 °C for 60 s), and fluorescent signals were detected using a Bio-Rad CFX Connect Real-Time PCR Detection System. The RNA copy numbers (log 10 copies/μg) were calculated from three independent experiments.

Statistical analysis
Statistical analysis was performed using Student's t-test. A p-value less than 0.05 was considered significant.

Expression and purification of NS3 and its mutants
To investigate the effect of amino acid substitutions in the positively charged surface patch of NS3 on enzyme activities, expression plasmids harboring a single substitution or simultaneous substitution of four basic residues (H24, R50, K74 and K94) ( Fig. 1) were constructed using the plasmid pET28a-NS3/NS4A PCS as a template [49]. These basic amino acids are highly conserved in pestivirus NS3 proteins (Fig. 1A, bottom). The four basic amino acids were simultaneously changed to alanine (A) or aspartic acid (D) in the NS3/NS4A PCS surface patch to generate the corresponding mutant, Quad_A or Quad_D. Truncated NS3 (NS3Hel) and the mutant NS3/K232A were used as controls in enzyme assays [18,42]. The recombinant proteins were expressed in E. coli, purified, and analyzed by SDS-PAGE (Fig. 1C).

The positively charged surface patch modulates the helicase activity of NS3 independently of its ATPase activity in vitro
To assess the effect of basic amino acid substitutions in the NS3/NS4A PCS surface patch on the helicase and ATPase activities of NS3, we first measured the ATPase activity of WT NS3 and its mutants in vitro. ATP hydrolysis by NS3 (WT) and its mutants (for simplicity, hereinafter collectively referred to by the mutated amino acid symbol) were determined under steady-state conditions, and the data were fit  Fig. 2A and Table 1). These results indicate that perturbation of the positively charged surface patch of NS3 had no influence on its ATPase activity.
We also measured the RNA helicase activity of NS3 and its mutants at a saturating ATP concentration. A partially single-stranded RNA (pssRNA) T40:R20 was used as a helicase substrate. The percentage of the unwound release strand relative to the total amount of release strand was calculated based on the fluorescent signals from native polyacrylamide gel electrophoresis (PAGE) separating free R20 from the T40:R20 complex. WT NS3 unwound 80% of the R20 under the tested conditions, while NS3Hel unwound only 15% of the substrate (Fig. 2B and Table 1). For the mutant NS3/K232A (NC), an extremely low unwinding activity was observed (Fig. 2B, 6% unwound). In set 1, the R50A, K74A, and K94A mutants unwound 71%, 86%, and 64% of the substrate, respectively. The mutants H24A and Quad_A exhibited a significantly decreased unwinding activity (Fig. 2B, 50% and 54% unwound) compared to WT NS3. In set 2, the mutant K74D unwound 71% of the substrate, and the unwinding activities of the remaining mutants significantly decreased compared to WT NS3 ( Fig. 2B; 47-64% unwound). The mutant Quad_D had the lowest helicase activity (Fig. 2B). These observations collectively suggest that the basic residues in the NS3 surface patch modulate the helicase activity in vitro.

The basic residues in the surface patch regulate the RNA-binding ability of NS3 synergistically
To further investigate the mechanism by which the charged patch in the NS3 protease domain regulates RNA unwinding activity, we assessed the RNA-binding ability of NS3 and its mutants. We hypothesized that the 5′ region of the substrate release strand resides at the back of NS3 and allows itself to bind the charged patch on the protease domain. We prepared another pssRNA-2 by annealing the T40 template strand and a 5′-FAM-labeled 33-mer release strand (R33). An FP-based RNA binding assay [28,31] showed that the apparent equilibrium dissociation constant (K d ) of NS3 was 30.4 ± 4.3 nM and that the mutant NS3/K232A (NC) exhibited similar RNA-binding ability, with a K d of 39.4 ± 9.9 nM (Fig. 3 and Table 1). The truncated mutant NS3Hel bound very little of the RNA substrate, and the value of K d could not be detected under the tested conditions ( Fig. 3 and Table 1). In set 1, the mutant Quad_A exhibited a significantly higher K d value, 84 nM, but the K d values of remaining mutants ranged from 24 to 39 nM, similar to NS3 (Fig. 3 and Table 1). In set 2, the mutant Quad_D had a K d value of 160 nM, and the remaining mutants had K d values in the range of 27-58 nM (Fig. 3 and Table 1). Compared to Quad_A, Quad_D exhibited significantly decreased RNA-binding ability. These results suggest that the four basic residues in the protease domain regulate the helicase activity by synergistically affecting the RNA-binding ability of NS3.  . 2 The ATPase and helicase activities of CSFV NS3 and its mutants. (A) ATPase kinetics curves. ATPase activity was measured using a malachite-green-based method. The initial reaction rates at different ATP concentrations were fitted to a standard Michaelis-Menten curve (mean ± SD; n = 3). Ctr, control. (B) The helicase activity was measured using a T40:R20 pssRNA substrate. Annealed, annealed T40:R20 in the absence of NS3 (upper, the annealed form); denatured, boiled T40:R20 (lower, the released R20); Ctr, NC, NS3Hel, WT; set 1, the mutants harboring alanine substitutions; set 2, the mutants harboring aspartic acid substitutions. Measurement of helicase activity was carried out independently four times, and the percentage of mean unwound RNA is shown below the gels.

The positively charged surface patch of NS3 plays an important role in regulating CSFV production
To investigate the role of the positively charged surface patch on NS3 in infectious virus production, we introduced each of the mutations into an infectious full-length cDNA clone of the CSFV Shimen strain, pSPT I /SM [20]. PK-15 cells were transfected with the full-length cDNA construct to rescue infectious CSFV. The infectious rescued virus corresponding to each construct is signified by a lowercase 'v' preceding the construct name. The recovery of infectious CSFV was detected by IFA using an anti-NS3 antibody at 72 h posttransfection. The data showed that PK-15 cells transfected with a construct harboring the H24A, R50A,K74A, H24D, R50D, or K74D mutation were positive for viral antigen. No viral antigen was detected in PK-15 cells transfected with an infectious cDNA clone containing the K94A, Quad_A, K94D, or Quad_D mutation ( Fig. 4A and Table 1). Next, we determined the titers of the rescued CSFV mutants. The mutant vK74A had a virus titer similar to that of the wild-type virus vWT. However, the titers of the mutants vH24A, vR50A, vH24D, vR50D, and vK74D were significantly lower (Fig. 4A). Consistent with the viral antigen detection results, no infectious CSFV was rescued when using the K94A, Quad_A, K94D or Quad_D construct ( Fig. 4A and Table 1). Viral RNA copy numbers in infected cells were determined using RT-qPCR [19,30]. The viral RNA copy numbers were significantly lower for the mutants vR50A (6 to 24 hpi), vR50D and vK74D (12 and 24 hpi), and vH24A and vH24D (24 hpi) compared to  Fig. 3 The RNA-binding ability of NS3 and its mutants, measured using an FP-based assay. pssRNA-2 annealed to T40 and a synthetic 5′-FAM-labeled 33mer ssRNA (R33) was used in this assay. Data represent the mean value of three independent experiments, and curves were fitted to a quadratic equation using GraphPad Prism software.  infected with vWT or its variant at an MOI of 0.001. At 6, 12, and 24 hpi, total RNA was extracted from the infected cells, and the viral RNA copy number was determined by RT-qPCR. The data represent the mean ± SD of three independent experiments. *, P < 0.05; **, P < 0.01 vWT. As expected, the levels of genomic RNA were found to be similar for vK74A and vWT (Fig. 4B). The differences in of RNA copy numbers were consistent with those of the virus titers. These results suggest that the positively charged surface patch of CSFV NS3 plays an important role in regulating infectious virus production by modulating viral RNA replication.

Discussion
Although NS3 of the members of the family Flaviviridae is a natural fusion protein with two separable enzymatic modules, the N-terminal protease and C-terminal NTPase/ helicase of the protein are functionally coupled. The pestivirus NS3 protease domain catalyzes cleavage of host and viral proteins and is essential for the process of viral RNA replication [36,49]. The protease domain is required for RNA unwinding by NS3 helicase and greatly enhances the ability of NS3 to bind RNA. Intermolecular electrostatics in HCV NS3 plays an important role in this process [3]. For West Nile virus NS3, this crosstalk between the protease and helicase modules has an autoregulatory function [7]. Perturbation of the CSFV NS3 protease-helicase interface by point mutations impairs the helicase activity in vitro as well as virus production in vivo [49]. It has been widely reported that the NS3 protease domain of members of the family Flaviviridae stimulates the RNA unwinding activity of its helicase [1,23,39,43,44,48,49]. In HCV, NS3Hel has greatly decreased helicase activity and RNA-binding ability compared to NS3fl [3,11]. The electrostatics and allosteric contribution from the interaction interface between the CSFV NS3 helicase and protease modules play an important role in the enhanced RNA-binding ability of NS3 by the protease domain [49]. Here, we further addressed the effect of residue substitutions in the basic patch on infectious virus production. Interestingly, no direct correlation was observed between NS3 helicase activity and virus production. Unexpectedly, when the residue K94 was mutated to alanine or aspartic acid, no infectious CSFV was rescued although the mutants K94A and K94D still had moderate helicase activity in vitro. Previous studies have demonstrated that HCV NS3 interacts with NS2 and other viral proteins to form a replication complex, which is essential for virus replication and viral particle assembly [24,34,47]. Mutations at the interface between the BVDV NS3 protease domain and the NS4A-kink region impairs the NS3/4A-kink interaction, and the mutant is no longer capable of viral RNA replication [9]. We speculated that K94 on NS3 surface patch is essential for replication complex formation, independently of its helicase activity or RNA-binding ability. The precise mechanism of action of the residue K94 on NS3 surface patch needs to be addressed further. Among the remaining mutants with a single amino acid substitution, the virus titer was to a certain extent related to the helicase activity and RNA-binding ability of NS3. The simultaneous four-residue substitutions (Quad_A or Quad_D) resulted in significantly decreased RNA-binding ability and helicase activity of NS3 in vitro. As expected, no infectious virus could be rescued from the full-length cDNA clone harboring the Quad_A or Quad_D mutations. Collectively, our results suggested that the positively charged surface patch on the pestivirus protease module plays an important role in modulating NS3 helicase activity and virus production. These findings contribute to our understanding of the functional regulation of pestivirus NS3 and will be of potential use for the design of novel antiviral strategies.