Crystal Structure of African Swine Fever Virus dUTPase Reveals a Potential Drug Target

African swine fever virus (ASFV), an Asfivirus affecting pigs and wild boars with up to 100% case fatality rate, is currently rampaging throughout China and some other countries in Asia. There is an urgent need to develop therapeutic and preventive reagents against the virus. Our crystallographic and biochemical studies reveal that ASFV E165R is a member of trimeric dUTP nucleotidohydrolase (dUTPase) family that catalyzes the hydrolysis of dUTP into dUMP. Our apo-E165R and E165R-dUMP structures reveal the constitutive residues and the configuration of the active center of this enzyme in rich detail and give evidence that the active center of E165R is very similar to that of dUTPases from Plasmodium falciparum and Mycobacterium tuberculosis, which have already been used as targets for designing drugs. Therefore, our high-resolution structures of E165R provide useful structural information for chemotherapeutic drug design.

effective vaccine or drug for this disease, and it is urgent to develop preventive and therapeutic reagents against it.
ASFV encodes a dUTP nucleotidohydrolase (dUTPase) (dUTP pyrophosphatase [Dut]; EC 3.6.1.23), called E165R. The enzyme degrades dUTP in the cytoplasm and thus minimizes misincorporation of uracil into viral DNA, which plays an essential role in maintaining the fidelity of genome replication. It is indispensable for reproductive infection of ASFV in swine macrophages, the natural host cell for the virus (7). Therefore, inhibition of the dUTPase activity can be detrimental to ASFV replication and an effective measure of treating ASFV infection.
Here, we present the crystal structures of apo E165R and its complex with dUMP. We show that E165R is a homotrimer and can specifically hydrolyze dUTP. The overall structure and active center of the enzyme are very similar to those of other homotrimeric dUTPases, especially from M. tuberculosis. Our results provide important clues for structure-based drug design by targeting E165R.

RESULTS
Characterization of recombinant ASFV E165R protein. Recombinant E165R protein was expressed in E. coli cells as tagged protein with a C-terminal His tag. The Ni-affinity-purified E165R protein was eluted from a Superdex 200 10/300 GL column with an elution volume of around 16 ml, indicating that it has a molecular weight of approximately 50 kDa. SDS-PAGE showed that the molecular weight of the E165R protein monomer is 17 kDa (Fig. 1A). Sedimentation velocity analytical ultracentrifugation analyses further confirmed that the E165R protein exists as a homotrimer (ϳ50.5 kDa) in solution (Fig. 1B). In the enzyme activity assay, we found that E165R specifically hydrolyzed dUTP and produced pyrophosphate (PPi), but did not hydrolyze dATP, dTTP, dCTP, or dGTP (Fig. 1C). We further determined the value of Michaelis constant (K m ), which was 3.788 Ϯ 0.191 m⌴ (Fig. 1D). These results suggested that the purified ASFV E165R protein is a homotrimeric dUTPase with strict substrate specificity.
Crystal structure of ASFV E165R. The structure of ASFV E165R was solved by molecular replacement, using the structure of P. falciparum dUTPase as a search model, to a resolution of 2.30 Å, with R work and R free values of 22.2% and 25.5%, respectively ( Table 1). The structure is a closely packed 3-fold symmetric homotrimer, with one E165R trimer in the asymmetric unit ( Fig. 2A). There are 447 visible residues (residues 1 to 149 of the total 165 residues in each subunit) in this structure, which account for 90.3% of the intact protein, and 16 residues at the C terminus of each subunit are missing due to the poor electron density. The E165R protomer in the trimer is mainly composed of ␤-strands. The 14 ␤-strands (␤1 to ␤14) form four sheets: ␤1, ␤2, ␤6, ␤7, and ␤14 form a parallel/antiparallel mixed sheet, ␤3, ␤8, ␤10, and ␤13 form an antiparallel sheet, ␤4 and ␤12 form another antiparallel sheet, and ␤5, ␤9 and ␤11 form the third antiparallel sheet. One ␣-helix (␣1) and two 3 10 helices (1 and 2) appear after ␤1, ␤13, and ␤14, respectively ( Phylogenetic analysis shows that primary sequences of dUTPases from different species are divided into three clusters, with each cluster corresponding to an oligomeric form (monomer, dimer, or trimer). ASFV E165R falls into the cluster of trimeric dUTPases (Fig. 2C). Therefore, combining the ultracentrifugation assay, the enzymatic assay, the crystal structure, and the phylogenetic analysis, we can conclude that ASFV E165R is a member of the trimeric dUTPase family.
Crystal structure of the E165R-dUMP complex. To further define the active center and explore the catalyzing mechanism of E165R as a dUTPase, we made an attempt to determine the crystal structure of E165R in complex with the dUTP substrate. However, we obtained the crystal structure of E165R-dUMP complex rather than E165R-dUTP complex at a resolution of 1.70 Å, with R work and R free values of 18.8% and 21.5%, respectively (Table 1). This result further confirmed that E165R has a pyrophosphatase activity that degrades dUTP into dUMP and pyrophosphoric acid. In the complex, the E165R has an overall similar trimeric three-fold symmetric structure to that of apo-E165R. Each of the three dUMP binds to the interface of two protomers of the E165R trimer (Fig. 3A).
All three E165R subunits in the crystal structure of E165R-dUMP complex contain 1 to 144 residues of the total 165 residues of E165R, and the 21 residues at the C terminus are invisible due to poor electronic density. The secondary structure of the E165R protein in the E165R-dUMP complex is very similar to that of apo-E165R, except that the C-terminal 3 10 helix in apo-E165R is missing in the E165R-dUMP complex (Fig. S1B).
By aligning one single subunit from apo-E165 and from the E165R-dUMP complex, we showed that these two subunit structures overlap well. Only their C termini show obvious conformational differences (Fig. S1C), suggesting high flexibility of the C terminus of E165R. By aligning the apo-E165R and the E165R-dUMP complex trimers, Crystal Structure of ASFV dUTPase we found that their subunits do not precisely overlap. There are certain anticlockwise rotations of each E165R-dUMP subunit relative to their corresponding subunits in apo-E165R. The angle degrees between the principal axes of the corresponding subunits in apo-E165R and in the E165R-dUMP complex are 1.96°, 4.87°, and 4.97° (  Fig. S1D).
dUMP interacts with N85, G88, L89, I90, D91, Y94, and M99 of one E165R protomer and R71, S72, and Q120 of another protomer. Among these interactions, N85, D91, and M99 from one protomer form hydrogen bonds with the deoxyuridine part of dUMP, while S72 and Q120 from the other protomer form hydrogen bonds with the phosphate group of dUMP (Fig. 3B).
dUTPases contain five conserved sequence motifs, which line the substrate binding pocket (active center) of the enzymes (25). By aligning E165R and other trimeric dUTPases, the five motifs in E165R were defined (see Fig. S2). As for E165R, motifs I, II, and IV of the active center are contributed by one protomer while motif III is contributed by another protomer. Like in many other dUTPase structures, the C termini (which contain motif V) of the three protomers were omitted due to poor electron density (Fig. 3C). By comparing the primary sequences of the five motifs from E165R and other trimeric dUTPases, we found that S72 from motif II, D91 from motif III, and Q120 from motif IV, which are responsible for forming hydrogen bonds with dUMP, are highly conserved among dUTPases from different species (Fig. S2).
where I i is the observed intensity and is the average intensity from multiple measurements.
where F o and F c are the structure-factor amplitudes from the data and the model, respectively. R free is the R factor for a subset (5%) of reflections that was selected prior to refinement calculations and was not included in the refinement.
By comparing the substrate/product binding pockets of apo-E165R and the E165R complex, we found that the binding pocket of apo-E165R is much larger than that of the E165R complex. As calculated by CASTp (26), the solvent accessible area and volume of the binding pockets in apo-E165R are 249.3 Å 2 and 163.2 Å 3 , respectively (Fig. S1E). In contrast, these values become 127.0 Å 2 and 62.2 Å 3 , respectively, for the binding pockets in the E165R-dUMP complex (Fig. S1F). Obviously, the binding pocket of E165R remarkably shrinks after binding the substrate/product for better coordinating them. There are several water molecules in the ligand binding pockets of both apo-E165R and the E165R complex structures. Some water molecules in the ligand binding pocket of the E165R-dUMP complex form hydrogen bonds with the phosphoric acid group of dUMP. They may be the active water molecules that are involved in the hydrolysis mechanism (Fig. S1F). Magnesium ions are important for the catalysis mechanism by coordinating the hydrolysis intermediate conformation and maximizing the catalyzing efficiency (25). However, we did not found any magnesium ion, free pyrophosphoric, or phosphoric acid groups in the binding pocket of the E165R-dUMP complex, just like in other dUTPase-dUMP complexes (11). The magnesium ion and the PP i (one of the hydrolysis products) may have been released from the active center after the hydrolysis reaction finished.
Comparison of the structure of E165R and other trimeric dUTPases. In the E165R-dUMP complex, the dUMP is mainly coordinated by the active center motifs I, II, Crystal Structure of ASFV dUTPase III, and IV (Fig. 4A). Based on the root mean square deviation (RMSD) values between the active center motifs I, II, III, and IV of E165R and those of other trimeric dUTPases, we found that the configuration of the active center of E165R is most similar to those of M. tuberculosis and P. falciparum dUTPases (Fig. 4B).
By superimposing the active center motifs of ASFV E165R with those of M. tuberculosis, we can see that the conformations of the backbones of the four motifs constituting the active centers of E165R and the M. tuberculosis dUTPase are very similar. In addition, the side chains of key residues that interact with the substrate in these motifs have the same orientations with those of M. tuberculosis dUTPase. Specifically, the side chains of S72, N85, D91, M99, and Q120 in E165R, which form hydrogen bonds with dUMP, have the same orientations as those of S65, N77, D83, K91, and Q114 in the M. tuberculosis dUTPase. These results further demonstrate the similarity of the active centers of E165R and M. tuberculosis dUTPase (Fig. 4C).
Similarly, alignment of the four active center motifs (motifs I, II, III, and IV) of ASFV E165R with those of the P. falciparum dUTPase shows that conformations of the backbones of these active center motifs as well as the side chain orientations of the substrate-interacting residues in these motifs of E165R are also very similar to those of P. falciparum dUTPase (Fig. 4D).
We also calculated the solvent accessible surface areas and volumes of the substrate binding pockets of E165R in complex with dUMP, M. tuberculosis dUTPase in complex with dUTP, and P. falciparum dUTPase in complex with ␣,␤-imido-dUTP (dUPNPP) by CASTp (26), and the results show that areas and volumes of E165R and M. tuberculosis dUTPase substrate binding pockets are very similar and that the shapes of the substrate binding pockets of these two proteins are also very similar ( Fig. 5A and B). However, the shape, area, and volume of the binding pocket of dUPNPP in P. falciparum dUTPase show obvious differences from those of E165R and M. tuberculosis dUTPase (Fig. 5C). In addition, we found the electrostatic potentials are distinct among these three pockets (Fig. 5D to F). These observations suggest that the unique property of the substrate binding pocket should be considered in the development of inhibitors against ASFV E165R.

DISCUSSION
Emerging and reemerging viruses keep challenging public health (27). ASFV is highly contagious and causes lethal diseases in both domestic pig and wild boar victims. Identification of drug targets from ASFV and structure-based drug design are important strategies for ASFV control and prevention. As a dUTPase, E165R plays an essential role in maintaining viral genome fidelity and integrity. Depletion of E165R aborts productive replication of ASFV in swine macrophages (28). Due to the high level of dUTP or a high ratio of dUTP/dTTP in macrophages, sufficient dUTPase activity is critical for viruses to replicate successfully in macrophages (29,30). In addition to strictly regulating the dUTP level and keeping uracil out of DNA, dUTPases from some virus species have another function as a signaling protein. Epstein-Barr virus (EBV) dUTPase binds to cellular Toll-like receptor 2 and activates NF-B expression, which leads to Crystal Structure of ASFV dUTPase immune dysregulation in the host (31). Murine herpesvirus 68 (MHV-68) dUTPase is essentially involved in the viral infective process by blocking the type I interferon signaling pathway via a mechanism independent of its intrinsic dUTPase activity. Whether ASFV E165R has similar functions needs to be explored in the future.
Our biochemical and crystallographic results demonstrate that ASFV E165R is a trimeric dUTPase with pyrophosphatase activity. The product of the hydrolysis catalyzed by E165R, dUMP, binds in the active center composed of motifs from the two adjacent protomers. The high-resolution structure of E165R-dUMP gives in-depth architecture information of the E165R active center and therefore provides a unique opportunity for designing inhibitors of this enzyme. By comparing the active center of E165R with those of other trimeric dUTPases, we found that both the architecture and substrate binding geometry of the E165R active center are very similar to those of M. tuberculosis and P. falciparum dUTPases, despite the binding pockets having distinct electrostatic potentials. Some chemicals have been designed to inhibit the M. tuberculosis dUTPase, such as ␣,␤-imido-dUTP (dUPNPP) (32). P. falciparum dUTPase has also been chosen as a target for antimalarial drug design, and a series of triphenylmethane derivatives of deoxyuridine were found to have antimalarial activities (21). These chemicals may also be effective for inhibiting E165R dUTPase activity or could serve as lead compounds for designing novel anti-ASFV drugs. All these observations need further investigation in the future.

MATERIALS AND METHODS
Gene cloning, protein production, and purification. The ASFV E165R gene (GenBank accession number CBW46796.1), coding for the dUTPase, fused at its C terminus with a hexa-histidine tag was cloned into the pET-21a vector (Novagen) with NdeI and XhoI restriction sites (33). Transformed E. coli strain BL21(DE3) clones were grown in LB medium containing 100 g/ml ampicillin to an optical density at 600 nm (OD 600 ) of 0.6 to 0.8 at 37°C. Expression of the recombinant proteins was induced by the addition of 0.5 mM isopropyl-␤-D-1-thiogalactopyranoside (IPTG), and incubation was continued for a further 16 h at 16°C. Cells were harvested by centrifugation at 7,000 ϫ g for 15 min at 4°C, resuspended in lysis buffer (20 mM Tris-HCl [pH 8.5] and 150 mM NaCl), and further homogenized with a lowtemperature ultrahigh pressure cell disrupter (JNBIO, China). The lysate was clarified by centrifugation at 20,000 ϫ g for 60 min at 4°C. The supernatant was purified by metal affinity chromatography using a HisTrap HP 5 ml column (GE Healthcare). Proteins were eluted using the lysis buffer supplemented with 300 mM imidazole. The proteins were further purified by gel filtration chromatography using a HiLoad 16/60 Superdex 75 PG (GE Healthcare) with a running buffer of 20 mM Tris-HCl (pH 8.5) and 50 mM NaCl, and the collected protein fractions were concentrated to 10 mg/ml using a membrane concentrator with a molecular weight cutoff of 10 kDa (Millipore).
Crystallization, data collection, and structure determination. Crystallization trials were set up with commercial crystallization kits (Hampton Research) using the sitting drop vapor diffusion method. Typically, 1 l protein was mixed with 1 l reservoir solution. The resultant drop was then sealed, equilibrating against 100 l reservoir solution at 18°C. Diffractable crystals of ASFV E165R were obtained in 0.1 M ammonium citrate tribasic (pH 7.0), 12% (wt/vol) polyethylene glycol 3350 at 18°C. For the ASFV E165R-dUMP complex, 10 mg/ml E165R proteins were cocrystallized with 10 mM dUTP and 10 mM MgCl 2 in a buffer of 20 mM Tris-HCl (pH 8.5) and 50 mM NaCl at 18°C. Diffractable crystals of the ASFV E165R-dUMP complex were obtained in 0.1 M citric acid (pH 3.5), 14% (wt/vol) polyethylene glycol 1000. Crystals were flash-cooled in liquid nitrogen after a brief soaking in reservoir solution with the addition of 17% (vol/vol) glycerol. The X-ray diffraction data were collected under cryogenic conditions (100 K) at Shanghai Synchrotron Radiation Facility (SSRF), beamline BL19U1, and indexed, integrated, and scaled with HKL2000 (34).
The E165R and E165R-dUMP complex structures were solved by the molecular replacement method using Phaser (35) from the CCP4 program suite (36), with the structure of P. falciparum dUTPase in complex with 2,3-deoxy-3-fluoro-5-o-trityluridine (PDB 1VYQ) as the search model (21). Initial restrained refinement and manual model building were performed using REFMAC5 (37) and COOT (38), respectively. Further refinement was performed using Phenix (39). Final statistics for data collection and structure refinement are represented in Table 1.
Biochemical characterization of E165R protein. The purified protein was analyzed with an analytical gel-filtration assay with a calibrated Superdex 200 10/300 GL column (GE Healthcare). The sample was further analyzed with SDS-PAGE.
The analytical ultracentrifugation assay was performed according to a previously reported method (40). The proteins were prepared in 20 mM Tris (pH 8.5) and 150 mM NaCl at an A 280 of 0.8. The assay was performed on an optimal ProteomeLab XL-I analytical ultracentrifuge (Beckman Coulter) at a speed of 48,000 rpm. The molecular mass analysis was performed with the XL-I data analysis software.
Measurement of enzyme kinetics. The substrate specificity of dUTPase ASFV E165R was determined by monitoring the production of PP i according to a previously reported method (41). The dUTPase enzymatic activity assay was performed by monitoring the color change of cresol red using a SpectraMax M5 (Molecular Devices, USA) (42). Briefly, a total of 990 l of buffer solution (2 mM bicine, 100 mM KCl, 5 mM MgCl 2 , pH 8.0) including final concentrations of 25 M cresol red and 10 M dUTP was rapidly mixed with 10 l of 2.5 M dUTPase in a cuvette (optical path of 1 cm). The color change of cresol red was recorded by monitoring the absorbance at 573 nm at intervals of 1.0 s at 37°C. Finally, the K m values were calculated using the integrated Michaelis-Menten method.
Data availability. The accession numbers for the atomic coordinates and diffraction data reported in this paper are PDB 6KY8 (crystal structure of E165R) and 6KY9 (crystal structure of E165R/dUMP complex).