Kinetic, Mutational, and Structural Studies of the Venezuelan Equine Encephalitis Virus Nonstructural Protein 2 Cysteine Protease

Abstract

The Venezuelan equine encephalitis virus (VEEV) nonstructural protein 2 (nsP2) cysteine protease (EC 3.4.22.-) is essential for viral replication and is involved in the cytopathic effects (CPE) of the virus. The VEEV nsP2 protease is a member of MEROPS Clan CN and characteristically contains a papain-like protease linked to an S-adenosyl-l-methionine-dependent RNA methyltransferase (SAM MTase) domain. The protease contains an alternative active site motif, (475)NVCWAK(480), which differs from papain's (CGS(25)CWAFS), and the enzyme lacks a transition state-stabilizing residue homologous to Gln-19 in papain. To understand the roles of conserved residues in catalysis, we determined the structure of the free enzyme and the first structure of an inhibitor-bound alphaviral protease. The peptide-like E64d inhibitor was found to bind beneath a β-hairpin at the interface of the SAM MTase and protease domains. His-546 adopted a conformation that differed from that found in the free enzyme; one or both of the conformers may assist in leaving group departure of either the amine or Cys thiolate during the catalytic cycle. Interestingly, E64c (200 μM), the carboxylic acid form of the E64d ester, did not inhibit the nsP2 protease. To identify key residues involved in substrate binding, a number of mutants were analyzed. Mutation of the motif residue, N475A, led to a 24-fold reduction in kcat/Km, and the conformation of this residue did not change after inhibition. N475 forms a hydrogen bond with R662 in the SAM MTase domain, and the R662A and R662K mutations both led to 16-fold decreases in kcat/Km. N475 forms the base of the P1 binding site and likely orients the substrate for nucleophilic attack or plays a role in product release. An Asn homologous to N475 is similarly found in coronaviral papain-like proteases (PLpro) of the Severe Acute Respiratory Syndrome (SARS) virus and Middle East Respiratory Syndrome (MERS) virus. Mutation of another motif residue, K480A, led to a 9-fold decrease in kcat and kcat/Km. K480 likely enhances the nucleophilicity of the Cys. Consistent with our substrate-bound models, the SAM MTase domain K706A mutation increased Km 4.5-fold to 500 μM. Within the β-hairpin, the N545A mutation slightly but not significantly increased kcat and Km. The structures and identified active site residues may facilitate the discovery of protease inhibitors with antiviral activity.

Figure 1

Organization of the VEEV nonstructural protein 2 (nsP2). (A) The nsP2 contains: an N-terminal region of unknown function, a helicase domain, a cysteine protease domain, and a SAM methyltransferase (MTase) domain. The nsP2 cysteine protease cleaves the nonstructural polypeptide into 4 segments (nsP1, nsP2, nsP3, nsP4). The cleavage is essential for viral replication. (B) The cysteine protease and SAM MTase domains are shown in white and green, respectively. The substrate is thought to bind beneath a β-hairpin (colored red) which forms a beak-like protrusion. The peptide-like E64d inhibitor was found beneath the β-hairpin and is shown in blue stick (PDB 5EZS).

Figure 2

(A) Time-dependent inhibition of the VEEV nsP2 protease by E64d. (B) Structure of the E64d prodrug. In cells esterases convert the E64d ester to the E64c acid. E64c is a prototypical cysteine protease inhibitor, but did not significantly inhibit the protease. E64d inhibited the enzyme in vitro, however, no inhibition of viral replication was observed with E64d or E64c in cell-based assays. (C) Proposed mechanism for covalent inhibition based upon the observed density in the E64d-inhibited enzyme structure. (D) Structure of the E64d VEEV nsP2 cysteine protease adduct (PDB 5EZS). The carbonyl oxygen of the ester was within hydrogen bonding distance of the Cys-477-NH (2.8 Å), this interaction likely stabilizes the TS in the enzyme-catalyzed proteolytic reaction. Other nearby NH groups were not directed towards the carbonyl oxygen. Asn-475 and Arg-662 were found hydrogen bonded in the adduct and no conformational change occurred in these residues upon inhibitor binding.

Figure 3

Proposed mechanism of the VEEV nsP2 Clan CN cysteine protease. (A) The mechanism is based upon the structures of the bound (PDB 5EZS) and free (PDB 5EZQ) enzyme described here. His-546 was found in two different conformations (lower panels) which may be relevant to different steps in the catalytic cycle. In one conformation His-546 could form an ion pair with the Cys-Sγ thiolate. In the free enzyme the side chain of Lys-480 was also directed towards the Cys-Sγ. During the catalytic cycle His-546 may rotate away from the Cys-Sγ in order to donate its proton to the amide nitrogen and facilitate the collapse of the tetrahedral TS and release of the free amine. (B) In overlays, the second conformation of the His-546-Nδ1 was near the E64d ester oxygen and was well positioned to donate its hydrogen. Notably, the Asn-475 side chain did not rotate towards the carbonyl oxygen. Asn-475 forms the base of the P1 binding site and remained hydrogen bonded to Arg-662 and to the backbone carbonyl oxygen of Asp-507.

Figure 4

Binding models of the VEEV nsP2 P34 substrates. (A) The C-terminal portion of the substrate (colored in light pink) was found directed downwards towards the SAM MTase domain, or (B) upwards towards the protease domain. The N-terminal portion of the substrate is colored in magenta. In both models the P4 residue (color in blue) interacted with K706. In the P34 model on the right the carbonyl of the scissile bond was closer to the Cys-Sγ. The hydrogen bond between R662 and N475 was maintained in the model on the right, and broken in the other. Other substrate bound models can be found in the Supplemental Information. (C) The 25-residue cleavage site sequences which were embedded between the YFP and CFP fluorescent proteins to make the substrates are shown for VEEV. The 17- and 25-residue SFV substrate sequences are also shown. The cleavage sequences for CHIKV are shown for comparison. The nomenclature of Berger and Schechter is used to identify residues on the amino (P1, P2, etc.) or carboxy (P1′, P2′, etc.) termini of the scissile bond. The P4 residue is colored, and the arrow indicates the location of the cleavage site. Cleavage sites in between the nonstructural proteins contain a common motif, AG(A/C)↓(G/Y/A). The P4 residue (colored) differs significantly between the Old and New World alphaviruses and is thought to be specifically recognized by residues in the SAM MTase domain (K705 and K706 in the VEEV nsP2).

Figure 5

Comparison of the active sites of (A) papain, the corona viral PLpro proteases from (B) SARS (C112S variant, PDB 4M0W) (74) and (C) MERS (PDB 4PT5) (75), and the (D) alphaviral nsP2 cysteine protease of VEEV (PDB 5EZS). Cysteine proteases utilize the backbone NH of the Cys and in some cases the side chains of various residues to stabilize the transition state. The architecture of the active site is distinctly different from that of a serine hydrolase which utilizes two to three backbone NH groups to form an oxyanion hole. Notably, the Asn-110 in the SARS protease, the Asn-111 in the MERS protease, and the Asn-475 in the VEEV protease are present at the n-2 position relative to the Cys and are conserved. The Asn does not hydrogen bond to the carbonyl oxygen of the substrate or inhibitor in the SARS PLpro or in the VEEV nsP2 cysteine protease, respectively.