Structure and Biochemical Characteristics of the Methyltransferase Domain of RNA Capping Enzyme from African Swine Fever Virus

African swine fever (ASF) is a highly contagious hemorrhagic viral disease in pigs that is caused by African swine fever virus (ASFV). There have been no effective drugs or vaccines for protection against ASFV infection until now.

ABSTRACT

African swine fever virus (ASFV) is a complex nucleocytoplasmic large DNA virus (NCLDV) that causes a devastating swine disease, and the development of effective anti-ASFV vaccines and drugs is urgently needed. The process of mRNA 5′-end capping is a common characteristic in eukaryotes and many viruses, and the cap structure is required for mRNA stability and efficient translation. The ASFV protein pNP868R was found to have guanylyltransferase (GTase) activity involved in mRNA capping. Here, we report the crystal structure of the pNP868R methyltransferase (MTase) domain (termed pNP868R^MT) in complex with S-adenosyl-l-methionine (AdoMet). The structure shows the characteristic core fold of the class I MTase family, and the AdoMet is bound in a negative, deep groove. Remarkably, the N-terminal extension of pNP868R^MT is ordered and far away from the AdoMet-binding site, distinct from the close conformation over the active site of the poxvirus RNA capping D1 subunit or the largely disordered conformation in most cellular RNA capping MTases. Structure-based mutagenesis studies based on the pNP868R^MT-cap analog complex model revealed essential residues involved in substrate recognition and binding. Functional studies suggest that the N-terminal extension may play an essential role in substrate recognition instead of AdoMet binding. A positively charged path stretching from the N-terminal extension to the region around the active site was suggested to provide a favorable electrostatic environment for the binding and approaching of substrate RNA to the active site. Our structure and biochemical studies provide novel insights into the methyl transfer process of the mRNA cap catalyzed by pNP868R.

IMPORTANCE African swine fever (ASF) is a highly contagious hemorrhagic viral disease in pigs that is caused by African swine fever virus (ASFV). There have been no effective drugs or vaccines for protection against ASFV infection until now. The protein pNP868R was predicted to be responsible for the process of mRNA 5′-end capping in ASFV, which is essential for mRNA stability and efficient translation. Here, we solved the high-resolution crystal structure of the methyltransferase (MTase) domain of pNP868R. The MTase domain structure shows a canonical class I MTase family fold, and the AdoMet binds into a negative pocket. Structure-based mutagenesis studies revealed critical and conserved residues involved in AdoMet binding and substrate RNA binding. Notably, both the conformation and the role in MTase activities of the N-terminal extension are distinct from those of the previously characterized poxvirus MTase domain. Our structure-function studies provide the basis for potential anti-ASFV inhibitor design targeting the critical enzyme.

INTRODUCTION

African swine fever virus (ASFV) belongs to the complex nucleocytoplasmic large DNA viruses (NCLDVs) and is the sole member of the Asfarviridae family that was recently classified into the Nucleocytoviricota phylum (1, 2). In swine, ASFV preferentially replicates in cells of the monocyte/macrophage lineage (3). African swine fever (ASF) is caused by ASFV infection and is a highly contagious and often lethal swine disease of domestic and wild pigs (4). ASF was first reported in Africa in the 1920s and has spread to many countries of Europe and Asia in recent decades, especially China, where there was a recent outbreak (5–8). No vaccine for protection against ASFV infection or effective treatments have been available until now. ASFV has a large double-stranded DNA genome of 170 to 190 kbp comprising 151 to 167 open reading frames (ORFs) depending on the viral strain. The ASFV particle contains about 70 different polypeptides, including multiple structural components, as well as a full set of enzymes and factors involved in viral transcription (9–10). The virion is ∼200 nm in diameter and possesses a multilayered structure consisting of the nucleoid, core shell, inner envelope, capsid, and a host-derived outer envelope (11). The recent cryo-electron microscopy (cryo-EM) structures of the ASFV capsid confirmed the five-layer ASFV structure and showed that the virion has a radially averaged diameter of ∼2,080 Å (12–14). The capsid is composed mainly of one major (p72) and four minor (M1249L, p17, p49, and H240R) capsid proteins, organized into pentasymmetrons and trisymmetrons.

The 5′-end cap structure is a common and conserved characteristic of eukaryotic cellular and viral mRNA that is essential for efficient splicing, export, translation, and stability of the mRNA (15–16). The capping process (Cap 0) usually requires three enzymatic reactions: the 5′-triphosphate end of the pre-mRNA is first hydrolyzed to a diphosphate by RNA 5′-triphosphatase (TPase), the diphosphate RNA is then capped with GMP by RNA guanylyltransferase (GTase) by forming GpppRNA, and finally, the GpppRNA cap is converted to a 7-methylguanosine RNA cap (m⁷GpppRNA) by RNA (guanine-N7)-methyltransferase (MTase) (17–19). Although these three activities are encoded by two or three different genes in eukaryotes (20), there are multifunctional polypeptides for mRNA capping in several large DNA viruses (21–23). A bifunctional TPase-GTase capping enzyme lacking a cap guanine-N7 MTase domain exists in chlorella virus and baculovirus (22). The vaccinia virus (VACV) RNA capping machinery is composed of the large D1 subunit and the stimulatory D12 subunit (24–25). The D1 subunit has been shown to be composed of an N-terminal TPase domain, a central GTase domain, and a C-terminal MTase domain, and a similar organization has also been observed in the capping enzyme (MimiCE) from mimivirus (21, 23). The domains containing three active sites of VACV capping enzyme have been well characterized by in vitro and in vivo mutagenesis studies for its TPase, GTase, and MTase activities (26–32). The crystal structure of the full-length D1-D12 heterodimer further revealed the TPase and GTase domains as members of the triphosphate tunnel metalloenzyme (TTM) and covalent nucleotidyltransferase (NTase) superfamilies, respectively, while the MTase domain exhibits the characteristic core fold of the class I family of S-adenosyl-l-methionine (AdoMet)-dependent MTases (23).

In ASFV, the protein pNP868R was previously found to covalently bind to GTP (α-³²P labeled), and the complex can be reversed by pyrophosphate, indicating that the pNP868R gene encodes a GTase involved in RNA capping (33). Bioinformatics analysis showed that the domain organization of pNP868R resembles the D1 subunit of the VACV capping enzyme (Fig. 1A), and its MTase activity may not require the stimulatory role of certain proteins like the D12 subunit. However, the MTase characteristic of pNP868R in the mRNA capping process has remained uncharacterized until now. In this work, we biochemically characterize the MTase domain of pNP868R by solving its crystal structure in complex with AdoMet and by performing structure-based mutagenesis studies. We found that both the conformation and the role in MTase activities of the N-terminal extension are distinct from those of the poxvirus MTase domain. The key residues involved in AdoMet binding and MTase activities are identified by mutagenesis studies. The structure-function studies provide novel insights into the catalytic mechanism of the NP868R MTase domain.

FIG 1.

Overview of the pNP868R MTase domain (pNP868R^MT) in complex with AdoMet at a 2.70-Å resolution. (A) Domain architecture of full-length pNP868R protein. It is predicted to be composed of a TPase domain, an GTase-OB fold domain, and an MTase domain (colored in blue, orange/gray, and green, respectively). (B) Oligomeric state of recombinant pNP868R^MT (residues 586 to 868) in solution determined by AUC. (C) Overall crystal structure of the pNP868R^MT-AdoMet complex. The two protein molecules (in green and magenta, respectively) formed a homodimer by crystal packing in the asymmetric unit. The electron density map (2F_o − F_c) of the N-terminal extension (residues 586 to 597) is shown at a 1.5σ level. The two binding AdoMet molecules are shown as yellow sticks.

RESULTS

Overall structure of the pNP868R^MT-AdoMet complex.

The structure of pNP868R^MT (residues ranging from M586 to N868) in complex with AdoMet was solved by molecular replacement using the mRNA cap (guanine-N7)-MTase Ecm1 (Protein Data Bank [PDB] ID 1RI1) from Encephalitozoon cuniculi as the search model and was refined to a final R/R_free factor of 0.19/0.23 at 2.70-Å resolution (Table 1). Two protein molecules form a homodimer by crystal packing in the asymmetric unit (Fig. 1C). However, the oligomeric state study by analytical ultracentrifugation (AUC) showed that the recombinant pNP868R^MT has a sedimentation boundary at 29.5 kDa (close to the theoretical molecular weight of 32.6 kDa), suggesting that pNP868R^MT is a monomer in solution (Fig. 1B).

TABLE 1.

X-ray data collection and refinement statistics

Parameter	Value(s) for pNP868R MTase domain
Beamline	SSRF 17U1
Data collection statistics
Wavelength (Å)	0.9788
Space group	P6522
Unit cell parameters	a = 127.3 Å, b = 127.3 Å, c = 208.2 Å, α = β = 90°, γ = 120°
Resolution (Å)	2.70 (2.75–2.70)a
No. of unique reflections	28,114 (1,364)
Completeness (%)	100 (99.7)
Redundancy	35.9 (25.8)
Mean I/σ(I)	27.0 (3.0)
Molecules in asymmetric unit	2
Rmerge (%)	13.1 (75.2)
Rmeas (%)	13.3 (76.6)
CC1/2	100 (87.0)
Structure refinement statistics
Reflections used in refinement	27,986
Resolution range (Å)	47.07-2.70
Rwork/Rfree (%)	18.7/22.5
Protein atoms	4,649
Protein residues	566
Waters	51
Avg B factor (Å2)
Protein	48.05
Ligand (AdoMet)	51.22
Ramachandran plot (%)
Most favored	96.1
Allowed	3.0
Disallowed	0.9
RMSDs
Bond length (Å)	0.011
Bond angle (°)	1.009

^aThe values in parentheses are those for the highest-resolution shell.

The structure of pNP868R^MT displays a canonical class I MT family fold comprising a seven-strand β-sheet (described as β4, β3, α2, β5, β6, β12, and β11) with three helices on each side (α1, α2, and α3 on one side and α4, α5, and α7 on the other side) (Fig. 1C and 2). The two extensions include an N-terminal extended β-sheet (β1), followed by the 9-residue β1-α1 loop, and a flap domain between β6 and α7 that contains a four-strand antiparallel β-sheet (β7, β8, β9, and β10), flanked by four extra helices (α6, α8, α9, and α10).

FIG 2.

Structure-based sequence alignment of pNP868R^MT with its representative homologs from different species performed using Clustal X (version 1.81) and ESPript 3. The homologs include the mRNA cap guanine N7 MTases from VACV (D1 from vaccinia virus, residues 545 to 844), E. cuniculi (Ecm1), Homo sapiens (Hcm1; residues 127 to 476), and yeast (Abd1 from Saccharomyces cerevisiae; residues 100 to 436). The conserved residues are boxed in blue, and identical conserved and low conserved residues are highlighted by a red background and red letters, respectively. The residues involved in AdoMet binding and possibly involved in substrate RNA binding are highlighted using black arrows and triangles, respectively, and used in the subsequent mutagenesis studies.

Structural comparison of pNP868R^MT with its homologs.

A DALI search (http://ekhidna.biocenter.helsinki.fi/dali_server) for globally similar proteins was performed within the Protein Data Bank. Significant structural similarity was found between pNP868R^MT and the MTases from different species in spite of ∼13% to 20% amino acid sequence identities. The top three closest structural homologs are the RNA guanine-7 cap MTases Ecm1 from E. cuniculi (PDB ID 1RI1; with a Z-score of 28.2 and a root mean square deviation [RMSD] of 1.9 Å for 240 Cα atoms), Hcm1 from human (PDB ID 5E9W; with a Z-score of 27.7 and an RMSD of 2.3 Å for 248 Cα atoms), and the MTase domain of the D1 subunit from VACV (PDB ID 2VDW; with a Z-score of 26.6 and an RMSD of 2.1 Å for 241 Cα atoms), although they share only ∼20% sequence identities.

Structural comparisons of pNP868R^MT with Ecm1 and the D1 MTase domain revealed that the most significant difference is the various conformations of their N-terminal extensions (Fig. 3A and B). In pNP868R^MT, the extended peptide is well structured (β1 and the following loop) (Fig. 1C) and far away from the AdoMet-binding site, whereas the corresponding extension closes over the active site of D1 MTase domain or is largely disordered in Ecm1 (the N-terminal 40 residues cannot be observed in the structure). Another striking difference is that in NP868R MTase, the adenosine nucleoside of the methyl donor adopts an anti conformation, as observed in Ecm1 (Fig. 3B) and many other AdoMet-dependent MTases (34), whereas a syn conformation exists in the poxvirus structure (Fig. 3A). In addition, the major differences also include the conformations of the helices α4/α8/α9 and several loops, especially those involved in D12 binding in the poxvirus structure.

FIG 3.

Structure comparisons of pNP868R^MT with its representative homologs. (A) Structural superimposition of pNP868R^MT (cyan) with the MTase domain of VACV D1 subunit (light gray) in complex with its activating subunit D12 (dark gray as surface) (PDB ID 2VDW). The N-terminal extension of the D1 subunit MTase domain that closes over the active site is highlighted in magenta. The AdoMet molecules in pNP868R^MT and D1 are shown as yellow and green sticks, respectively. The adenosine nucleoside of AdoMet in pNP868R^MT and D1 adopts an anti and a syn conformation, respectively. (B) Structural superimposition of pNP868R^MT (cyan) with E. cuniculi RNA cap MTase Ecm1 (PDB ID 1RI1) (light gray). The N terminus (residues 1 to 40) is largely disordered and is not observed in the current structure. The AdoMet molecules in pNP868R^MT and D1 are shown as yellow and green sticks, respectively. The adenosine nucleoside of AdoMet in both pNP868R^MTand Ecm1 adopts an anti conformation. The bound cap analog GpppG in Ecm1 is shown as orange sticks.

AdoMet binding in the active site of the MTase domain.

In the NP868R^MT structure, the AdoMet molecule is bound tightly in the pocket and consists mainly of several loops, including β2-α2, β3-α3, β4-α4, and β5-η1, forming a negative, deep groove (Fig. 4A and B). The main part of the methionine side chain of AdoMet overhangs into a large cavity with dominantly positive charges, which is most likely a binding site for the substrate guanine (see Discussion). The cavity is large enough to accommodate a guanine ring and position its amino group (N7) next to the active site residues (Fig. 4B). The directly interacting amino acids surrounding the AdoMet-binding region (within 3.8 Å) include K607, G624, D646, D680, and N709.

FIG 4.

AdoMet binding in the active site. (A) Contact analysis between AdoMet (yellow sticks) and pNP868R^MT (cyan cartoon). An electron density map (2F_o − F_c) of AdoMet is shown at a 1.5σ level. (B) AdoMet binds into a deep pocket with a dominantly negative charge seen from the surface electrostatic potential of pNP868R^MT (blue, +6.3 KT; red, −6.3 KT). KT, K_bT/e_c.

The hydroxyl group and carboxyl group of AdoMet methionine are coordinated by the two salt bridges to the side chain of K607, while the amino group is stabilized by forming two hydrogen bonds (H bonds) with the main chains of G624 and N709. K607 and N709 are not conserved in cellular or poxvirus MTases (Fig. 2), indicating that pNP868R^MT may have different AdoMet binding characteristics. The AdoMet-binding motif 620-VIDLGIG-626 in ASFV corresponds to the conserved motif 68-VLDLGCG-74 in cellular MTases and nonconserved 594-VLAIDFG-600 in poxvirus MTases (Fig. 2). In poxvirus MTases, this nonconserved motif has been suggested to be a key correlate for the rarer syn conformation of AdoMet (25). D646 stabilizes the ribose ring by forming several H bonds via its two side chains. The aspartic acid is highly conserved in both cellular MTases and poxvirus MTases (Fig. 2). F711 and Y714 may generate a hydrophobic aromatic stacking to further stabilize the AdoMet adenosine moiety and for the approach of guanine in the substrate cap. Polar contacts also exist between the side chain of D680 and the adenosine N6 and N1 atoms by the formation of two H bonds.

Thermodynamic analyses of pNP868R^MT with AdoMet.

The interaction of wild-type pNP868R^MT with AdoMet was characterized by isothermal titration calorimetry (ITC), and the integrated heat data could be fitted well by using the one-site binding site model, with a binding affinity (K_a) of 1.75 × 10⁵ M⁻¹ (Fig. 5A and B). Structure-based mutagenesis of pNP868R^MT was performed to confirm the key residues involved in AdoMet binding. The K607E mutant, which would abolish the two H-bonds mediated by its side chain with AdoMet methionine and significantly alter the surface charge of the AdoMet-binding pocket (from negative to positive), has a much more severe loss of binding affinity (K_a = 1.93 × 10⁴ M⁻¹). Surprisingly, the G624A mutant shows no detectable binding to AdoMet, suggesting that the motif 620-VIDLGIG-626 plays a vital role in AdoMet binding. The binding affinity of the D646A mutant was significantly affected (K_a = 1.23 × 10⁴ M⁻¹), confirming its essential role in stabilizing the AdoMet sugar. Unexpectedly, the F711A and N709A mutants, which are not conserved in cellular MTases and poxvirus MTases, have a more severe loss of binding affinity (with K_a values of 3.25 × 10⁴ M⁻¹ and 2.92 × 10⁴ M⁻¹, respectively). The F711A mutant disrupts the hydrophobic interaction with the AdoMet adenosine moiety, indicating that such contact may be important for AdoMet stabilization. N709 can stabilize the AdoMet methionine by directly binding to the amino group and providing the charge complementation, and the significantly low binding affinity of the mutant suggests its essential role in AdoMet binding. The binding affinity of the Y714A mutant was almost not affected (K_a = 1.40 × 10⁵ M⁻¹), indicating that its hydrophobic interactions with the AdoMet adenosine moiety may not be important for AdoMet binding. It is noted that we were unable to obtain the soluble forms of the D646A and D680A mutants in the denaturation-renaturation experiments of their inclusions, and their binding affinities were not determined.

FIG 5.

Binding characteristics analysis of pNP868R^MT wild type and mutants to AdoMet by ITC. (A) ITC data for titration of pNP868R^MT variants with AdoMet. K_a, binding affinity; ΔH, change in enthalpy; ΔS, change in entropy; N, number of binding sites; WT, wild type. (B) ITC spectra for pNP868R^MT variants. Baseline-subtracted raw ITC data for injections of AdoMet are indicated in the upper panels of each of the ITC profiles shown. The peaks normalized to the ligand/protein molar ratio were integrated as shown in the bottom panels. The solid squares represent the experimental data, and the best fit to the experimental data was obtained from a nonlinear least-squares method of fitting by use of a one-site binding model depicted by a solid line.

Meanwhile, a variant (the Δ586–597 mutant) lacking 12 residues (residues 586 to 597) was created to investigate the role of the N-terminal extension in AdoMet binding. The results showed that the binding of the deletion mutant is similar to that of the wild type, consistent with the observation that the N-terminal extension is far away from the AdoMet-binding site in the structure.

In vitro activity of NP868R^MT mutants.

Considering the high structural similarity between NP868R^MT and Ecm1 (Fig. 3B), structural superposition was performed to determine the implications on the substrate-binding characteristic features of NP868R^MT (Fig. 3B and 6A). The cap analog (m⁷GpppG) from Ecm1 can be docked into the putative substrate binding pocket with predominantly positive charges without steric clash by their superposition. Moreover, the distance between the AdoMet methyl group and the substrate N7 atom is ∼4.0 Å in this model, suggesting that a small shift may be sufficient to trigger catalysis (Fig. 6A). The docked m⁷GpppG showed generally favorable direct interactions with several conserved residues in NP868R^MT. For example, the interacting residues K54, K81, R106, and Y284 in Ecm1 correspond to K607 (also involved in AdoMet binding), R633, R658, and F860 in NP868R^MT, respectively, indicating that they may have similar substrate-binding characteristics. The residues I603, S604, R633, and R658 may form several hydrogen bonds or provide charge complementation with the phosphate groups of the cap analog, while the hydrophobic interaction between F860 and the guanosine ring may further stabilize the m⁷GpppG molecule. In addition, there are several conserved residues distributed around the active site, including the positive R592, R597, and K627 residues, as well as Y714, which may bridge the cap guanine.

FIG 6.

Identification of the key residues involved in the MTase activities of pNP868R^MT. (A) Close-up view of the potential MTase active sites of pNP868R^MT (cyan) and Ecm1 (gray). The cap analog m⁷GpppG is docked into the active site of pNP868R^MT by its superimposition with Ecm1 as shown in Fig. 3B. The substrate-interacting residues in Ecm1 and the corresponding residues in pNP868R^MT are shown as gray and cyan sticks, respectively. The distance between the AdoMet methyl group and the N7 atom of m⁷GpppG is ∼4.0 Å, which may require a slight shift to trigger catalysis. (B) In vitro MTase activities of the pNP868R^MT variants. The recombinant proteins were assayed for MTase activity by label transfer from [³H-CH₃]AdoMet to GpppA cap dinucleotide as described in Materials and Methods. Data are presented as the average (±standard error of the mean) from duplicate experiments. (C) Close-up view of the surface charge surrounding the potential MTase active sites of pNP868R^MT (blue, +6.3 KT; red, −6.3 KT), Ecm1 (blue, +7.4 KT; red, −7.4 KT), and VACV D1 (blue, +7.4 KT; red, −7.4 KT), colored by their local electrostatic potential. The regions around the active site that directly contact the potential substrate RNA are predominantly electropositive in the three RNA capping enzymes.

These potential interacting residues were mutated to identify key residues in substrate recognition and catalysis by determining their MTase activities in vitro. The results showed that all mutants exhibit reduced activities compared to the wild-type NP868R^MT (Fig. 6B). Mutation of S604 to alanine moderately reduced the MTase activity (∼45% of that of the wild-type protein), suggesting its role in stabilizing the cap triphosphate bridge, whereas mutation of I603 had little effect on the MTase activity. Mutations of the positive residues R607, R633, and R658, which may provide charge complementation or form direct contacts with the phosphate groups of the cap analog, resulted in severe losses in MTase activity, to ∼16%, ∼21%, and ∼18% of that of the wild type, respectively. Mutation of Y714 to alanine also resulted in severe loss of activity, to ∼14% of that of the wild type, whereas the specific activity of the F860A mutant was ∼60% of that of the wild type, suggesting that Y714 may play a more important role in bridging interactions between the cap guanine and AdoMet (Fig. 6A). Meanwhile, mutation of R592 and R597 also reduced the MTase activity to ∼73% and ∼21% of that of the wild type, respectively, suggesting that the N-terminal extension may play an important role in substrate binding.

DISCUSSION

In this study, the structure of pNP868R^MT showed it has a class I MTase fold with remarkable similarities to many cellular and viral cap MTases. Our sequence alignment and mutation studies showed that many residues essential for pNP868R^MT activity are highly conserved in cellular and viral cap MTases. For example, the positive residues K607, R633, and R658 are critical for the cap MTase activity of pNP868R^MT. These residues correspond to K54, K81, and R106 in Ecm1, respectively, which provide the positive charges around the cap-binding pocket. Mutations of these residues were lethal in vivo (K81A and K54A) or caused a severe growth defect (R106A) (34). Similarly, the three corresponding residues in VACV D1 MTase, R573, K607, and R632, were also found essential for cap methylation in vivo, being involved in interactions with AdoMet carboxyl, cap guanosine, and cap triphosphate, respectively (30, 31). Another example is that mutation of Y714 significantly affected the MTase activity of pNP868R^MT. The corresponding residue Y145 makes bridging interactions between the cap guanine and AdoMet ligands in the Ecm1 structure (25). Moreover, substitutions of alanine for the equivalent tyrosine (Y683) of VACV MTase reduced the activity to 0.05% and indicated that the aromatic group was essential for cap methylation (29). Mutations at the equivalent tyrosine residues of Saccharomyces cerevisiae MTase Abd1 (Y254) and human cap MTase Hcm1 (Y289) were lethal in vivo and elicited severe defects in cap MTase activity in vitro (29, 30, 35, 36). The high conservation of the key residues suggests that the substrate cap recognition characteristics of pNP868R^MT are similar to those of cellular and viral cap MTases.

A significant difference between pNP868R^MT and cellular and viral cap MTases is the conformation of their N-terminal extensions. It is ordered and far away from the MTase active site in pNP868R^MT, distinct from the close conformation over the active site in the MTase domain of poxvirus D1 subunit (25) or the largely disordered conformation (such as that of Ecm1) in many cellular RNA capping MTases (34). In the structure of VACV D1 MTase (residues 548 to 844), the N-terminal peptide (residues 545 to 563) is ordered and forms a surface loop that overlays S-adenosyl-homocysteine (AdoHcy) and virtually buries the methyl donor and acceptor (25). The mutational analysis of the N-terminal lid peptide showed that it is essential for both AdoMet binding and cap binding in vivo (32, 36, 37), in which R560 and R562 (equivalent to R597 in pNP868R^MT) are responsible for bridging the cap triphosphate. Moreover, deletion mutant analysis showed that the N-terminal residues 30 to 40 in Ecm1 are essential for the MTase activity in vivo, although they are not visualized in the crystal structure (34). In our study, deletion of the N-terminal extension seems not to be important for AdoMet binding (Fig. 5), which may be explained by its relatively far distance from the AdoMet-binding site in the pNP868R^MT structure. Importantly, our functional studies suggest that the N-terminal extension (R592 and R597) plays an essential role in substrate recognition by in vitro MTase assays (Fig. 6B). Therefore, it is also possible that the N-terminal extension may be induced to get close to the active site upon substrate binding.

The electrostatic surface of the MTase structures could most likely reflect the region where the RNA or DNA substrates will bind. Notably, on the electrostatic surface of pNP868R^MT (Fig. 6C), there is a long positively charged surface patch surrounding the MTase catalytic site (starting from K592 in the N-terminal extension). The positive path may mediate the binding and approaching of the negative substrate RNA to the active site of pNP868R^MT. In Ecm1, there is a positively charged surface extending outward from the triphosphate bridge, which was suggested to provide a path for the incoming RNA strand to interact with Ecm1 (Fig. 6C). Mutation of the residues constituting the positive surface, such as R47, K75, K81, and R106, were found to cause a severe growth defect or to be lethal in yeast (34). Similarly, there is also a positively charged region close to the active site of VACV D1 (Fig. 6C). The residues contributing to the charged surface patch are composed mainly of N-terminal R560 and R562 (R597 in pNP868R^MT), K573 (K607 in pNP868R^MT), and K607 (R633 in pNP868R^MT), which are conserved in the D1 orthologs encoded by numerous diverse genera of vertebrate and invertebrate poxviruses (25). The mutational analysis showed that all of them are essential for MTase activity in vivo and are suggested to interact with the cap triphosphate bridge (31, 37). Considered together, the comparisons suggest a common essential role of the positive surface patch around the active site in the three capping enzymes in substrate RNA binding.

Conclusions.

In the present study, we performed structure-function studies of the MTase domain of ASFV mRNA cap enzyme pNP868R. The structure revealed a class I MTase fold with remarkable structural similarities to poxvirus and cellular mRNA capping MTases in spite of low sequence identities. Notably, the N-terminal extension is far away from the MTase active site, unlike the close conformation over the active site in the MTase domain of the poxvirus D1 subunit or the largely disordered conformation in most cellular RNA capping MTases. The in vitro MTase assays and ITC analysis revealed that the N-terminal extension plays an essential role associated with MTase activity instead of AdoMet binding. Structure-guided mutagenesis also revealed potential key residues involved in substrate binding, and many of them are conserved in poxvirus and cellular mRNA capping MTases. A positively charged path stretching from the N-terminal extension to the active site nearby was suggested to provide a favorable electrostatic environment for the incoming RNA strand interacting with pNP868R^MT. The systematic studies give us a deeper molecular understanding of the substrate recognition and catalytic mechanism of the pNP868R MTase domain.

MATERIALS AND METHODS

Cloning.

The optimized full-length DNA sequence encoding ASFV pNP868R was synthesized by GenScript Corp. (Nanjing, China). The sequence encoding the MTase domain (Met586 to Asn868) of pNP868R (referred to as pNP868R^MT) was PCR amplified from the synthesized NP868R gene and cloned into the modified pET28a-SUMO (Novagen, USA) in which the cleavable ubiquitin-like-specific protease 1 (ULP1) fused with an N-terminal His₆ tag was introduced. The recombinant plasmid was transformed into an Escherichia coli DH5α cloning strain and plated onto LB kanamycin plates. The plasmid was isolated and transformed into an E. coli BL21(DE3) star expression strain (Invitrogen). Site-directed mutagenesis of pNP868R^MT was performed by a PCR-based technique according to the QuikChange site-directed mutagenesis strategy (Stratagene) in accordance with the manufacturer’s instructions. The mutant genes were sequenced and found to contain only the desired mutations.

Protein expression and purification.

Bacterial cells were grown to mid-log phase in LB medium at 37°C in the presence of 50 mg/ml kanamycin. Induction of protein expression was initiated by adding isopropyl-1-thio-β-d-galactopyranoside (IPTG) to the culture to a final concentration of 0.4 mM, and the cells were grown at 16°C. Cells were pelleted after 20 h by centrifugation at 6,000 × g for 10 min at 4°C. The cell pellet was resuspended in a buffer containing 20 mM Tris (pH 8.0), 100 mM NaCl, 2 mM β-mercaptoethanol, and 1 mM phenylmethylsulfonyl fluoride (PMSF) and lysed by ultrasonication on ice. The cell debris and membranes were pelleted by centrifugation at 20,000 × g for 60 min at 4°C. The soluble N-terminally His-SUMO-tagged pNP868R^MT was purified by affinity chromatography with nickel-nitrilotriacetic acid resin (Bio-Rad, USA), and the tag was removed by protease ULPI hydrolysis overnight and reloading with 20 mM imidazole. pNP868R^MT was further purified by gel filtration (Superdex 75; GE Healthcare, USA) equilibrated in a buffer containing 20 mM Tris (pH 8.0), 300 mM NaCl, and 2 mM dithiothreitol (DTT) using an ÄKTA purifier system (GE Healthcare, USA). Highly purified protein fractions were pooled and concentrated. Protein concentrations were determined using the Bio-Rad protein assay kit, and crystallization trials were performed by ultrafiltration in an Amicon cell (Millipore, USA).

Protein crystallization.

A mixture of pNP868R^MT and AdoMet (Sigma, USA) at a molar ratio of 1:3 was prepared and incubated on ice for 6 h before performance of crystallization experiments. The initial crystallization condition of the complex was obtained under the no. 25 crystallization conditions in the Index kit (catalog no. HR2-134; Hampton Research, USA) by using the sitting drop vapor diffusion method at room temperature after 2 days. The crystal quality was optimized by adjusting the concentration of the precipitant and the pH value. The best crystal was obtained in a solution of 4.3 M sodium formate (pH 6.6) after 3 to 4 days.

Data collection, crystal structure determination, and refinement.

The diffraction data from a single crystal were collected on the BL17U1 beamline station of SSRF (Shanghai Synchrotron Radiation Facility) using an EIGER pixel detector at a wavelength of 0.9788 Å. The total oscillation was 360° with 1° per image, and the exposure time was 0.3 s per image. Before data collection, the crystals were soaked in the reservoir solution supplemented with 20% (vol/vol) ethylene glycol for a few seconds and then flash-frozen in liquid nitrogen. All the data were processed by the program HKL2000 (38). The initial phases were calculated using the program PHASER (39) with the mRNA cap MTase Ecm1 (PDB ID 1RI1) as the search model. The structure was refined with the program Phenix.refine (40) and manually corrected in Coot (41). The qualities of the final models were validated with the program MolProbity (42). Refinement statistics and model parameters are given in Table 1. The program PyMOL (http://www.pymol.sourceforge.net/) was used to prepare structural figures.

ITC.

Isothermal titration calorimetry (ITC) was applied to quantitatively determine the binding affinities of the pNP868R^MT wild type and mutants to AdoMet. For the titration experiments, the protein was purified by the same method as described above and dialyzed against the buffer containing 50 mM Na HEPES (pH 7.5), 0.15 M NaCl, 5% (vol/vol) glycerol, 2 mM MgCl₂, and 2 mM β-mercaptoethanol for 24 h. The ITC experiments were carried out using a high-sensitivity iTC-200 microcalorimeter from MicroCal (GE Healthcare) at 20°C using 20 to 100 μM AdoMet in the injector and 2.5 to 10 μM protein in the sample cell. All samples were thoroughly degassed and then centrifuged to eliminate precipitates. Injection volumes of 2 µl per injection were used for the different experiments, and for every experiment, the heat of dilution for each ligand was measured and subtracted from the calorimetric titration experimental runs for the protein. Consecutive injections were separated by 2 min to allow the peak to return to the baseline. Integrated heat data obtained for the ITCs were fitted in a one-site model using a nonlinear least-squares minimization algorithm to a theoretical titration curve, using the MicroCal-Origin 7.0 software package.

In vitro MTase assay.

Reaction mixtures (20 μl) containing 50 mM Tris-HCl (pH 8.0), 5 mM DTT, 500 μM GpppA (New England Biolabs), 40 μM [³H-CH₃]AdoMet (Perkin Elmer Life Sciences), and 0.5 μM recombinant pNP868R^MT were incubated for 60 min at 37°C. Aliquots (4 μl) of the reaction mixture were then spotted on polyethyleneimine-cellulose thin-layer chromatography (TLC) plates (Sigma-Aldrich), which were developed in 0.05 M (NH₄)₂SO₄. The AdoMet- and m⁷GpppA-containing portions of the lanes were cut out, and the radioactivity in each was quantified by liquid scintillation counting. A parallel control reaction to subtract background was done in which the MTase assay was performed in the absence of AdoMet. Each mutant sample was run in duplicate to ensure accuracy.

AUC.

The sedimentation velocity measurements were carried out using a Beckman Optima XL-I analytical ultracentrifuge (AUC) (Beckman-Coulter Instruments) with a Ti rotor at 20°C. The pNP868R^MT protein concentration was adjusted to an absorption of ∼0.8 at 280 nm. The SEDFIT program was used to analyze the sedimentation coefficient (43).

Data availability.

The atomic coordinates and structure factors of the pNP868R^MT-AdoMet complex were deposited in the RCSB PDB under the identifier 7D8U.