Structural and functional studies of the papain-like protease 2 from mouse hepatitis virus
Our goal is to establish a system to investigate how the deubiquitinating (DUB) and deISGylating activities of coronavirus (CoV) papain-like protease domains (PLPs) are involved in virus immune evasion. To this end, we chose PLP2 from mouse hepatitis virus (MHV) as our target of study because MHV has historically served as a model system for the study of CoVs, and it has undeniable advantage of ease in culturing in comparison to human coronaviruses. It is reported here the expression and purification of a region of MHV nsp3 that contains the catalytic core of the PLP2 domain and its neighboring domains. More than 20 mg of pure and active protein could be obtained from 1 L of E. coli cell culture. Further kinetic characterization of MHV PLP2 using fluorescence-based assays revealed that PLP2 possesses protease (RLRGG-AMC as the substrate), DUB (Ub-AMC as the substrate) and deISGylating activities (ISG15-AMC as the substrate). With the three substrates tested, MHV PLP2 can only be saturated by Ub-AMC with Km of 1.3 mM, but not by RLRGG-AMC or ISG15-AMC where Km is not determinable. This suggests that PLP2 has a higher binding affinity to Ub than to ISG15 or the peptide substrate (RLRGG). In addition, MHV PLP2 is a better DUB than SARSCoV PLpro and MERS-CoV PLpro, while its deISGylating activity is lower than the two PLpro’s from human CoVs. Moreover, survey with di-Ubs cleavage revealed that MHV PLP2 shows promiscuous recognition of the ubiquitin chain linkage. Although structures of several CoV PLPs have been reported, no structure of PLP from betacoronavirus genogroup 2a is available. The 2.6 Å structure of MHV PLP2 determined in Chapter 3 through single-wavelength dispersion method (SAD) with SeMet-substituted PLP2 presents the very first structure of PLP from a betacoronavirus 2a group. It reveals a thumb-palm-finger architecture of the PLP2 catalytic core with a well-aligned Cys-His-Asp catalytic triad, and an ubiquitin-like domain (Ubl2) at the N-terminus of the PLP2 catalytic core. What is significant about this structure is that it uncovers a new MHV nsp3 domain, designated as “Domain Preceding Ubl and PLP2” or “DPUP”. The DPUP domain is structurally similar to the SARS unique domain (SUD), which raises doubts towards the unique presence of the SUD domain in SARS-CoV. The kinetic (Chapter 2) and structural (Chapter 3) characterization of the DPUP-Ubl2-PLP2 construct reported in this dissertation is the largest portion of coronaviral nsp3 that has ever been characterized in vitro. In order to investigate the interactions between PLP2 and Ub, a computational model of PLP2–Ub complex was generated through molecular dynamics. X-ray crystal structure of PLP2 in complex with Ub was also determined (1.85 Å), where a non-covalent complex was generated using PLP2 with a catalytic cysteine to serine mutation and mono Ub (Chapter 4). Combining the information from the computational model and the crystal structure of the complex, residues of PLP2 that are involved in the binding of Ub but are at a distance from the active site were identified. A series of mutants were then generated targeting these residues in order to selectively disrupt the DUB and/or deISGylating activity of PLP2 while maintain its protease activity. Further in vitro activity evaluation classified the mutants into four groups (Chapter 4): mutants with decrease in both DUB and deISGylating activity (Group I), mutants with decrease only in deISGylating activity (Group II) or DUB activity (Group III), and unstable mutants (Group IV). Although not all mutants maintain intact protease activity, there are mutants with good protease activity in each class, such as F290R and R281A (numbering based on the DPUPUbl- PLP2 construct) in Class 1, R253E/R257A and R253A/R257A in Class 2, and I249R and D250A in Class 3. Corresponding mutant viruses were then generated by our collaborators through reverse genetics method to further investigate the effect of the mutation on virus pathogenesis. In addition, since CoV nsp3 is a large, membrane-associated, multi-domain protein, it is very likely that there is crosstalk between the adjacent domains of nsp3. In order to explore this possibility, mutations were introduced into the Ubl2 domain, which results in the discovery of a temperature-sensitive mutant (V787S). This mutant contains a valine to serine mutation at position 787 (V787S, numbering based on MHV nsp3) in the Ubl2 domain that compromises the stability and activity of the adjacent PLP2 domain (Chapter 5). The V787S mutant has a melting temperature (Tm) approximately 7°C lower than the Tm of WT protein, and the mutant is gradually inactivated when incubated at elevated temperature. The corresponding mutant virus (AM2 virus) has also been shown to be attenuated and capable of eliciting protective immune response in mice by our collaborators. (Abstract shortened by ProQuest.)
Mesecar, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our