
==== Front
J Biol Chem
J Biol Chem
The Journal of Biological Chemistry
0021-9258
1083-351X
American Society for Biochemistry and Molecular Biology

S0021-9258(24)02176-8
10.1016/j.jbc.2024.107675
107675
Research Article
Characterization of alternate encounter assemblies of SARS-CoV-2 main protease
Aniana Annie 1
Nashed Nashaat T. 1
Ghirlando Rodolfo 2
Drago Victoria N. 3
Kovalevsky Andrey 3
Louis John M. johnl@niddk.nih.gov
1∗
1 Laboratory of Chemical Physics, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, DHHS, Bethesda, Maryland, USA
2 Laboratory of Molecular Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, DHHS, Bethesda, Maryland, USA
3 Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA
∗ For correspondence: John M. Louis johnl@niddk.nih.gov
14 8 2024
9 2024
14 8 2024
300 9 1076753 7 2024
6 8 2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
The assembly of two monomeric constructs spanning segments 1-199 (MPro1-199) and 10-306 (MPro10-306) of SARS-CoV-2 main protease (MPro) was examined to assess the existence of a transient heterodimer intermediate in the N-terminal autoprocessing pathway of MPro model precursor. Together, they form a heterodimer population accompanied by a 13-fold increase in catalytic activity. Addition of inhibitor GC373 to the proteins increases the activity further by ∼7-fold with a 1:1 complex and higher order assemblies approaching 1:2 and 2:2 molecules of MPro1-199 and MPro10-306 detectable by analytical ultracentrifugation and native mass estimation by light scattering. Assemblies larger than a heterodimer (1:1) are discussed in terms of alternate pathways of domain III association, either through switching the location of helix 201 to 214 onto a second helical domain of MPro10-306 and vice versa or direct interdomain III contacts like that of the native dimer, based on known structures and AlphaFold 3 prediction, respectively. At a constant concentration of MPro1-199 with molar excess of GC373, the rate of substrate hydrolysis displays first order dependency on the MPro10-306 concentration and vice versa. An equimolar composition of the two proteins with excess GC373 exhibits half-maximal activity at ∼6 μM MPro1-199. Catalytic activity arises primarily from MPro1-199 and is dependent on the interface interactions involving the N-finger residues 1 to 9 of MPro1-199 and E290 of MPro10-306. Importantly, our results confirm that a single N-finger region with its associated intersubunit contacts is sufficient to form a heterodimeric MPro intermediate with enhanced catalytic activity.

Keywords

SARS-CoV-2 main protease
protein folding
monomer-dimer equilibrium
inhibitor binding
dimer interface
swapped dimer
structure comparison
alternate folding pathways
precursor processing
protein fragment assembly
fold-switching
Abbreviations

ESV ensitrelvir

FS fold-switching

ipTM interface predicted TM score

Kd inhibitor dissociation constant

Kdimer dimer dissociation constant

MPro main protease

NAC nickel-affinity chromatography

NMV nirmatrelvir

nsp nonstructural protein

pLDDT predicted local distance difference test

pTM predicted TM score

SARS-CoV-2 severe acute respiratory syndrome coronavirus 2

SEC size-exclusion chromatography

SEC-MALS size-exclusion chromatography with in-line multi-angle light scattering

SV-AUC sedimentation velocity analytical ultracentrifugation

Reviewed by members of the JBC Editorial Board. Edited by Joseph Jez
==== Body
pmcIn SARS-CoV-2, a single copy of the main protease (MPro) is encoded as a part of large polyproteins (pp) 1a and 1ab surrounded by other nonstructural proteins (nsp) 1 to 10 and 1 to 16, respectively (1, 2, 3, 4). Like most viruses, polyprotein processing mediated by MPro (nsp5) is a critical step for the assembly and production of replication competent progeny virion (5). In vitro studies to understand the early mechanisms of viral protease regulation and ordered processing of the polyproteins are limited by the difficulty in attaining intact active precursors due to the efficient autocatalytic maturation upon expression in Escherichia coli (4). However, understanding such mechanisms and identifying the intermediate species in the pathway to MPro maturation may provide additional insights into defining novel targets and strategies for future drug design.

The mature MPro consists of three domains, I-III (Fig. 1A). The catalytic region includes domains I (residues 8–101) and II (residues 102–184), which exhibits an α-chymotrypsin–like fold. They are connected to a helical segment (domain III, residues 201–306) through a long loop region (residues 185–200), which constitutes part of the substrate-binding subsite S5 (4, 6, 7). The N-terminal residues 1 to 14 (termed N-finger) form the dimer interface through conformational changes and interactions with domain II and III (8, 9, 10, 11, 12, 13). Self-cleavage (autoprocessing) at the N-terminal nsp4/nsp5 site is pivotal for stable dimer formation and mature-like catalytic activity (10, 14). This is evident from the dimeric structure of the mature MPro showing that the free N-terminal S1 residue enables forming the substrate subsite S1 pocket (15) accompanied by the decrease in the dimer dissociation constant [(Kdimer), (14)]. Dimer formation is associated with (1) stabilizing the active site oxyanion loop (residues 138–145) into a wound active conformation [E∗-state, (11)] and (2) ∼40 ° rotation of the helical domain in relation to its orientation in monomeric MPro (9, 16) to facilitate dimer interface contacts. Likewise, substrates or inhibitors binding to MPro favor the E∗ over the inactive E conformation independent of dimer formation (10, 17, 18) as well as stabilize the dimer form. Among the interface residues R4, S10, G11, E14, N28, S139, F140, S147, E290, and R298 (8), which contribute to dimerization, a pivotal interface in the mature MPro dimer is created through two intersubunit salt bridges between residues E290 and R4 (10, 19). Accordingly, an E290A mutation in full-length MPro increases the Kdimer by ∼270-fold (4).Figure 1 Domain organization of mature MPro dimer and rationale for this investigation.A, one monomer [PDB 7JUN, (15)] is shown in white. In the other, the catalytic region composed of domains I and II and the helical region (domain III) are colored goldenrod and yellow green, respectively. The connecting loop between the catalytic and helical regions and the N-terminal residues 1 to 9 (N-finger) are shown in black. The catalytic dyad H41/C145 residues (red) are shown as ball and stick representations and the active site oxyanion loop in red. B, steps in the mechanism of MPro maturation from its precursor analog (4). C, the two deletion constructs used in this study to examine the transient pathway for MPro subunit assembly concomitant with the appearance of enhanced catalytic activity (4). Residues 1 to 199 make up the catalytic domain represented by the construct MPro1-199. Ovals, MPro catalytic and helical domains; black line, MPro N-terminal residues; dashed black and continuous blue lines, flanking nsp4 and nsp5 regions, respectively; red circles, catalytic dyad residues; E∗, active (wound) conformation of the active site oxyanion loop.

In recent studies, we showed that N-terminal autoprocessing at the nsp4/nsp5 site of a model MPro precursor precedes the C-terminal nsp5/nsp6 cleavage (4). Although the isolated catalytic domain (MPro1-199) is monomeric and exhibits very low catalytic activity compared to that of the MPro dimer, a precursor of MPro1-199 undergoes N-terminal autoprocessing in E. coli and in vitro (4, 11). A mechanism was proposed based on the analysis of various precursor constructs of MPro with cleavage sites at both termini (4). In that mechanism, the MPro precursor undergoes N-terminal intramolecular cleavage from either the monomer or the dimer. The product derived from N-terminal cleavage with a free N terminus enables forming a heterodimeric intermediate and thereby increasing the catalytic activity and faster processing of the second N-terminal nsp4/nsp5 site of the heterodimer [Fig. 1B, see Figure 8 in reference (4) for more details].

Like MPro1-199, MPro when lacking the N-terminal residues 1 to 9 (MPro10-306, Fig. 1C) is monomeric and exhibits catalytic activity comparable to that of MPro1-199 (20). In this study, we show that a mixture of MPro10-306 and MPro1-199 assemble to form a heterodimer with increased catalytic activity. The heterodimer is formed through cooperative assembly mediated by the interaction of the N-finger residues of MPro1-199 with domain III of MPro10-306 (Fig. 1C). Such an assembly mimics the heterodimer suggested in the proposed mechanism of N-terminal autoprocessing of the MPro precursor. The addition of inhibitor to the mixture leads to trimer and tetramer formation concomitant with further increase in catalytic activity. Intriguingly, such encounter assemblies appear to mimic minor populations that accumulate upon expression of an MPro precursor construct appended to nsp4 residues as well as a deletion construct of MPro lacking residues 1 to 9, involving the dimerization of the helical domain. Structural models are proposed describing the assembly of dimers and higher order assemblies based on existing atomic structures and AlphaFold 3.

Results

Assembly of a heterodimer comprising MPro1-199 and MPro10-306 in the absence and presence of inhibitors

Previous studies allude to the presence of a transient heterodimeric intermediate in the N-terminal autoprocessing pathway of MPro from model precursors (4). In such a heterodimer, one subunit with a free N-terminal region (N-finger), like that of mature MPro, interacts with the helical region of a precursor subunit and thereby enhances the dimer population leading to faster cleavage of the unprocessed nsp4/nsp5 site. To provide experimental support for the existence of such heterodimer intermediates through interdomain complementation, two monomeric constructs were expressed and purified as described in the Experimental procedures. The scheme for their purification and Kdimer values are listed in Table S1. MPro10-306 bears a deletion of the N-finger region (residues 1–9), which makes crucial interface contacts with domain II and III essential for dimer stability. MPro1-199 carries a complete deletion of the helical region (residues 200–306) and thus excludes the interaction of the free N-terminal residues with domain III (Fig. 1C).

Plots of the absorbance c(s) distributions by sedimentation velocity analytical ultracentrifugation (SV-AUC) before and after mixing MPro1-199 and MPro10-306 at equimolar concentration (40 μM) are shown in Figure 2A. The S values and corresponding molecular masses are summarized in Table 1. Individually, both proteins sediment as a single species corresponding to the expected monomeric mass. The mixed sample points to an association between the two proteins as evidenced by a broadening of the MPro10-306 peak with a corresponding decrease in the integrated MPro1-199 peak. However, no distinct complex sedimenting at a higher S value is observed. Thus, all experiments reported in this work were carried out at or below ∼50 μM.Figure 2 Assembly of MPro1-199and MPro10-306monitored by SV-AUC. Absorbance c(s) distribution in the absence (A) and presence (B–D) of the inhibitor in buffer B at 25 °C. Proteins and inhibitor were prepared at three concentrations (12.5–50 μM) as shown in B through D. Actual measured concentrations by AUC are listed beside each plot. Estimated S and corresponding mass values are listed in Table 1.

Table 1 SV-AUC analysis of MPro1-199 and MPro10-306 in the absence and presence of inhibitors and the corresponding S values and molecular mass

Constructs and mixtures (μM)	s20,w (S)	Mass (kDa)	
MPro1-199			
40	2.26	21	
MPro10-306			
40	2.75	33	
MPro1-199 + MPro10-306 (1:1)			
40	2.29, 2.83	23, 31	
MPro1-199 + MPro10-306 + GC373 (1:1:4)			
41	5.62	95	
21	5.52	75	
10	5.02	47	
MPro1-199 + MPro10-306 + NMV (1:1:4)			
43	5.65	96	
22	5.57	95	
11	5.45	86	
MPro1-199 + MPro10-306 + ESV (1:1:4)			
54	5.67	102	
S and molecular mass values extracted from sedimentation absorbance distributions shown in Figure 2. Theoretical molecular mass of MPro1-199 and MPro10-306 is 21,885 and 32,775 Da, respectively.

kDa, kilo Daltons.

Monomeric structures of MPro1-199 reveal an inactive E-state and those bound to inhibitors [GC373 or nirmatrelvir (NMV) or ensitrelvir (ESV)] reveal an active E∗-state typical of the inhibitor-free dimer and inhibitor–dimer complexes as described (11, 20). Therefore, an inhibitor was used to establish the active (E∗) conformation of the oxyanion loop and thereby, enhance the assembly of MPro1-199 and MPro10-306 to heterodimers. The covalent inhibitor GC373 binds to both monomeric proteins with Kd’s of 32 ± 5 and 44 ± 8 μM, respectively, determined by ITC (11). Our previous SV-AUC results of both, MPro1-199 and MPro10-306, when analyzed individually at ∼50 μM concentration in the presence of 2-fold molar excess of the covalent inhibitor GC373 showed a distribution corresponding to a monomer under these conditions (11). The Kdimer of MPro1-199 and MPro10-306 with GC373 was estimated to be 1.2 ± 0.2 and >>1.2 mM, respectively [Table S1, (11)].

Interestingly, like as shown for the predominantly monomeric full-length MPro [MProM, Table S1, (10)], which forms a dimer upon GC373 addition [Kdimer = 6.2 μM, (10)], MPro1-199 and MPro10-306 show association into higher order forms. Assemblies of MPro1-199 and MPro10-306 in the presence of inhibitors GC373, NMV, and ESV are shown in Figure 2, B–D (Table 1). The AUC absorbance distributions are shown for three protein concentrations ranging from 10 to 50 μM with inhibitor in molar excess. The profiles appear to be about the same for covalent inhibitor NMV and the noncovalent ESV with complex formation reaching a near maximum S value at 40 to 54 μM concentration. The associated form corresponds to ∼47 kDa complex in the presence of GC373 at 10 μM proteins as compared to ∼86 kDa complex with NMV. The largest species at ∼40 μM of each protein in the presence of GC373 and NMV is 95 to 96 kDa and that for ESV at 54 μM is ∼102 kDa.

Assessment of the approximate stoichiometry of the complex

The assembly of MPro1-199 and MPro10-306 in the absence and presence of NMV was monitored also by size-exclusion chromatography with in-line multi-angle light scattering (SEC-MALS). NMV was chosen primarily because it forms a chemically more stable adduct (11, 21, 22) and exhibits good solubility in buffer (up to 4 mM). Unlike SV-AUC, significant dilution of the sample that occurs during the size-exclusion chromatography (SEC) step is unavoidable and the estimated concentrations of a slice at the top of the peak are indicated. Figure 3A shows the fractionation of a premixed amount of 50 μM of each protein in an injection volume of 125 μl on Superose-12 (1 × 30 cm) column (blue trace). No major association of the proteins based on the molecular mass is evident under these conditions (Fig. 2A). However, in the presence of molar excess of NMV, the proteins assemble to various reversible interconverting species reaching an estimated mass of ∼80 kDa (Fig. 3A, black). Figure 3B shows a similar fractionation with a premixed amount of 100 μM of each protein with NMV in a 140 μl injection volume. Complex formation is observed for the peak with the retention volume between 10.5 to 11.5 ml. The estimated mass is ∼80 kDa with a protein concentration of ∼2.5 μM at the top of the peak. SDS-polyacrylamide gel analysis of these fractions followed by quantifying the band intensities indicates the presence of both proteins at a ratio of 2:1.4 of MPro10-306/MPro1-199 complex. This ratio was determined by comparing the band intensities of predetermined amount of these proteins in three different ratios on gels under identical conditions marked 2:2, 2:1, and 1:2. The difference in the molecular mass of the ∼80 kDa complex at 2.5 μM observed by SV-AUC and ∼47 kDa at 10 μM by SV-AUC could be attributed to the differences in sample preparation and the method itself. The maximum size of ∼102 kDa observed by SV-AUC approaches a composition of two molecules of MPro1-199 and two molecules of MPro10-306 with a calculated mass of 109.3 kDa (Table 1).Figure 3 Molecular mass estimation of MPro1-199and MPro10-306association by SEC-MALS.A, Fifty micromolars of each protein was mixed either in the absence (blue trace) or presence of 10-fold molar excess of NMV (black) in a total injection volume of 125 μl. Black (without NMV) and red (with NMV) circles indicate estimated mass range. B, Hundred micromolars of each protein mixed with 10-fold molar excess of NMV (black) in a total injection volume of 140 μl. Samples were fractionated on Superose-12 column (1 × 30 cm) in buffer B at a flow rate of 0.5 ml/min at 25 °C. Red circles indicate estimated mass range. C, collected fractions from (B, 10.5–11.5 ml) showing complex formation were analyzed by SDS-PAGE. The estimated ratio (2:1.4) by quantifying the band intensities indicates a complex approaching a stoichiometry of two molecules of MPro10-306 and one molecule of MPro1-199 at ∼2.5 μM concentration estimated at the top of the peak. This ratio was confirmed by the analyses of predetermined mixed amounts (2:2, 2:1 and 1:2) of the two proteins (where 2 denotes 1 μg of protein) by SDS-PAGE. M and kDa denote molecular weight standards and kilo Daltons, respectively. Molecular mass was calculated with the Astra software provided with the instrument.

Appearance of catalytic activity is concomitant with MPro1-199 MPro10-306 heterodimer formation

Unlike the dimeric MProWT, the catalytic activity of MPro1-199 is linearly dependent on the protein concentration indicating that the catalytic activity is that of a monomer (11). No dimer was detectable for inhibitor-free MPro1-199 up to 200 μM as evaluated by SV-AUC [Fig. S1A, (11)]. Dimerization was even more restricted for MPro10-306 with no dimer detectable up to ∼200 μM even in the presence of inhibitor GC373 and NMV [Fig. S1B, (11)]. The estimated Kdimer values are listed in Table S1.

Catalytic activities of MPro1-199 and MPro10-306 assayed individually or when mixed in equimolar ratios in the absence and presence of GC373 are shown in Figure 4. The raw progress curves are shown in Fig. S2. Comparison of the initial rates observed using a synthetic substrate is shown in Table 2. Consistent with earlier observations, the catalytic activities of MPro1-199 and MPro10-306 individually are very low and similar, compared to that of MProWT, but interestingly an equimolar mixture of both proteins exhibits a 13-fold increase in catalytic activity. Addition of GC373 to an equimolar amount of the protein mixture leads to an additional 7-fold increase in catalytic activity (Fig. 4A, Table 2).Figure 4 Catalytic activity of various assembled compositions of MPro1-199and MPro10-306and their mutants. The concentrations of each protein mixed at a ratio as indicated above the plots are (A, E) 25 μM and (B) 10 μM. A, bar plot of the initial rates calculated from progress curves of catalytic activity in the absence and presence of GC373, respectively, shown in Fig. S2. B, a plot of the rate versus GC373 concentration showing the rise and fall in catalytic activity of the protein mixture. C, plots of the catalytic activity of 40 μM MPro1-199 mixed with 80 μM GC373 versus increasing concentration of MPro10-306 (red) and vice versa (black). D, a plot of the rate/MPro1-199 concentration versus MPro1-199 concentration showing the mid-point of the dissociation of the complex (∼6 μM, dashed intercept) accompanied by the loss in catalytic activity. E, SV-AUC analysis of similar mixtures with GC373 as shown in (A). kDa denotes kilo Daltons.

Table 2 Catalytic activity of MPro constructs and their assemblies

Constructs and mixtures (μM)	Activity (μM/min)	Activity/μM of total protein (min-1)	Reference	
MProWT (at 0.5 μM)	12.7	25.4	(10)	
MPro1-199 (50 μM)	0.01	0.00020	This work	
MPro10-306 (50 μM)	0.0093	0.00019	This work	
MPro1-199 (25 μM) + MPro10-306 (25 μM)	0.131	0.00262	This work	
MPro1-199 (25 μM) + MPro10-306 (25 μM) + GC373 (25 μM)	0.972	0.01944	This work	
MProM (10 μM)	0.0086	0.00086	(10)	
MProM (10 μM) + GC373 (10 μM)	0.243	0.02430	(10)	
MProTM (50 μM)	0.016	0.00032	(20)	
MPro10-306 atypical dimer (25 μM)	0.0124	0.00049	This work	
MPro10-306 atypical dimer (25 μM) + GC373 (25 μM)	0.0056	0.00023	This work	
MPro1-199 (25 μM) + MPro10-306 atypical dimer (25 μM)	0.215	0.0043	This work	
MPro1-199 (25 μM) + MPro10-306 atypical dimer (25 μM) + GC373 (25 μM)	0.420	0.0084	This work	
Initial rate measurement at the indicated concentration of enzyme and 200 μM substrate in buffer B at 25 °C. See enzyme assays section under Experimental procedures for details. MPro10-306 atypical dimer was purified as shown in Figure 5A. But for the MPro10-306 atypical dimer, the 6H-tag was removed for the rest of the assayed constructs (Table S1).

The GC373-mediated concentration-dependent increase in catalytic activity of an equimolar mixture of 10 μM proteins reaches a maximum at 10 μM GC373 followed by a decrease in catalytic activity upon further increase in GC373 concentration (Fig. 4B). Clearly, the increase in catalytic activity is dependent on the catalytic C145 in MPro1-199 and residues E290 and R298 in MPro10-306 confirmed from mixtures of MPro1-199/C145A/MPro10-306 and MPro1-199/MPro10-306/E290A/R298A. Of the latter mutations, E290A and R298A, R298A is unlikely to be of significance because M6, with which R298 makes an intrasubunit hydrogen bond does not exist because of the deletion of residues 1 to 9 (4). Mixtures of MPro1-199/MPro10-306/C145A without and with GC373 exhibit the same trend in catalytic activities as that of MPro1-199/MPro10-306 mixtures. In control assays, MPro10-306 and MPro1-199 show no increase in catalytic activity in the presence of GC373 under identical assay conditions.

At a constant concentration of GC373 mixed with one of the proteins, the rate of hydrolysis of the synthetic substrate is linearly dependent on the concentration of the other protein (Fig. 4C). These results indicate that the catalytically active complex contains equimolar amount of MPro1-199, MPro10-306, and GC373. Since the catalytic activity is observed only from the active site of MPro1-199 in a complex comprising equimolar amount of MPro1-199 and MPro10-306, Figure 4D shows the dependency of the rate of the catalyzed-hydrolysis of the synthetic substrate per molar concentration of MPro1-199 on the concentration of MPro1-199 at a mixture comprising a constant ratio of 1:1:4 of MPro1-199, MPro10-306, and GC373. The increase in catalytic activity displays a titration curve for the formation of the three-component complex with a mid-point of ∼6 μM.

Comparison of equimolar mixtures of proteins with GC373 by SV-AUC, similar in concentrations to Figure 4A, indicate the formation of heterodimer populations evident from the broadening of the peaks (Fig. 4E) with MPro1-199/MPro10-306 and MPro1-199/MPro10-306/C145A mixtures exhibiting propensities to form higher order assemblies, consistent with the SV-AUC profiles shown in Figure 2.

Identification of atypical dimers of monomeric MPro10-306, MPro197-306, and MPro model precursor

It is plausible that encounter assemblies larger than a heterodimer of MPro1-199 and MPro10-306 involve a fold-switching (or fold-switched, termed FS) pathway of MPro10-306 through dimerization of its domain III, like that shown for the isolated domain III (23, 24) as well as a deletion mutant of MPro lacking its N-terminal seven residues [(MPro-Δ7, (25)], both derived from SARS-CoV. An interesting consensus emerges from our analyses of various MPro expression constructs during our studies producing monomeric MPro precursor mimetics and deletion mutants to assess dimerization and autoprocessing in E. coli. The initial construct, and the product when released upon autoprocessing or protease cleavage, are indicated on top of each panel (Figure 5, Figure 6 and 6 and Table S1). All MPro constructs shown display excellent solubility. Proteins following nickel-affinity chromatography (NAC, step 1, see Experimental procedures) were subjected to SEC (step 2). Comparable to the expression results of just domain III of SARS-CoV, expression of the monomeric MPro10-306 and MPro197-306 results in observing a minor population of an atypical dimer (Figure 5, A and C). Unlike the MProWT dimer or its interface mutants (10, 11, 14, 20), which exhibit a faster reversible interconversion to the monomer, the atypical dimers are stable to permit their isolation and refractionation as shown in Figure 5, B and D. The purified atypical dimers of MPro10-306 and MPro197-306 exhibit slower conversion to the monomer over days of incubation at room temperature or at 37 °C (Fig. 5B) like that of the isolated SARS-CoV domain III (23, 24) and MPro-Δ7 (25) FS-dimers. Mass estimations of MPro197-306 atypical dimer and monomer peaks (Fig. 5C) upon refractionation are shown in Fig. 5D.Figure 5 SEC profiles of atypical dimer forms of MPro10-306and MPro197-306, mass estimations, and conversion to the monomer.A and C, expressed proteins after initial purification by NAC (10–12 mg) were subjected to SEC on Superose-12 column in buffer B at ambient temperature. B and D, SEC-MALS analyses of atypical dimer and monomer forms. Estimated concentrations at the top of the peak are indicated. Injection concentrations for SEC-MALS range from 25 to 336 μg in 125 μl. Conversion of the MPro10-306 atypical dimer to the monomer [traces other than red in (B)] was monitored by injecting 1/10th the injection concentration of 250 μg in 125 μl (red) following incubation at 25 and 37 °C for the specified duration. Tbin and kDa denote thrombin and kilo Daltons, respectively.

Figure 6 SEC profiles of MProWTand MProH41Aexpressed as precursors appended to N-terminal nsp4 residues.A and B, expressed proteins after initial purification by NAC [(A): 6 mg, (B): 2 mg)] were subjected to SEC on Superose-12 column in buffer B at ambient temperature. (-102) and (-6) denote the length of nsp4 residues and asterisk, the Q to E mutation (see Table S1 for details). C and D, pooled peak fractions of dimer and monomer from (A) and (B) were subjected to SEC-MALS as described in Experimental procedures. Estimated concentrations at the top of the peak are indicated. Injection concentrations for SEC-MALS were 250 μg in 125 μl. Conversion of the (-6∗)MProH41A atypical dimer to the monomer [traces other than red in (C)] was monitored by injecting 1/10th the injection concentration of 250 μg in 125 μl (red) following incubation at 25 °C for the specified duration. kDa denotes kilo Daltons.

In the above context, it is relevant to note that non-native residues or solubility/affinity tags appended to the N terminus of full-length MPro and which negatively influence MPro dimerization (14) prior to their removal have led to observing oligomeric populations of MPro up to an octamer upon expression (23, 26, 27). Such an oligomeric population of MProWT precursor appended to 102 amino acids of nsp4 is not observed because of its efficient autoprocessing within 15 to 20 min of expression resulting in a single dimer form of mature MProWT (Fig. 6A). Reproducibly, when N-terminal autoprocessing is restricted as in (-6∗)MProH41A resulting in a mainly monomer form of MPro [Kdimer = >340 μM, (14)], a minor atypical dimer peak is observed as confirmed by SEC-MALS (Fig. 6, compare B-D). Recent studies show that mutations Q to E (denoted by an asterisk) in the P1 position of the nsp4/nsp5 cleavage site and H41A restrict autoprocessing while the six nsp4 residues (-6) increase the Kdimer (14). Like the atypical dimers above, (-6∗)MProH41A atypical dimer also exhibits a slow equilibrium of conversion to the monomer (Fig. 6C). Mass estimation of (-6∗)MProH41A monomer peak (Fig. 6B) upon refractionation is shown in Fig. 6D.

Discussion

Primarily, results presented here identify the minimal interface arrangement required to form a plausible heterodimeric intermediate in the proposed mechanism of MPro autoprocessing based on recent in vitro studies of model precursors (Fig. 1B). Two monomeric proteins, MPro1-199 and MPro10-306, were analyzed under conditions when assembly can occur only through interdomain complementation. A mixture of the two proteins displays peak broadening of MPro10-306 in AUC analysis concomitant with enhanced catalytic activity over their individual activities (see below), indicating a heterodimeric population in the mixture. The association is further confirmed by the addition of inhibitors, which results in detectable higher order assemblies. Notably, inhibitor binding to both proteins favors stabilizing the oxyanion loop in the E∗-state, a feature that is part of the requirement for native-like dimer interface formation and can occur independent of dimerization (4, 10, 11). The catalytic activity of this heterodimer population of MPro1-199 and MPro10-306 is compared with previously characterized full-length predominantly monomeric MProM and monomeric MProTM (Tables 2 and S1).

MPro1-199 and MPro10-306 exhibit catalytic activity of 0.0002 min-1 and 0.000186 min-1, respectively, for the hydrolysis of the synthetic peptide substrate (Table 2). In contrast, a mixture of an equimolar amount of both shows a catalytic activity of 0.00262 min-1, which is > 13-times larger than either of the proteins. This suggests the formation of a population of heterodimers harboring at least one competent active site. A similar increase in catalytic activity is observed from an equimolar mixture of MPro1-199/MPro10-306/C145A, but not from equimolar mixtures of MPro1-199/C145A/MPro10-306 and MPro1-199/MPro10-306/E290A/R298A (Fig. 4A). This points to MPro1-199 being the catalytically competent species. The addition of the covalent inhibitor GC373 to an equimolar amount of MPro1-199/MPro10-306 and MPro1-199/MPro10-306/C145A increases the catalytic activity further by a factor ranging from 7 to 10 (Fig. 4A). The observed catalytic activity in the presence of GC373 displays first-order dependency on the concentration of both, MPro1-199 and MPro10-306, indicating that the catalytically active species is that of a heterodimer with one inhibitor molecule bound to the active site of MPro10-306. The catalytic activity increases upon the addition of GC373 in a concentration-dependent manner up to 10 μM followed by a decrease in catalytic activity upon further increase in GC373 concentration (see Fig. 4B) attesting to GC373, further occupying the MPro1-199 active site. The rise and fall in catalytic activity as a function of increasing GC373 concentration can be explained by a similar mechanism to that reported previously for MProM. Unlike MPro1-199 and MPro10-306, the predominately monomeric MProM displays catalytic activity from a homodimer. The addition of GC373 to MProM favors dimer formation with concomitant increase in catalytic activity up to a molar equivalent (10). Accordingly, GC373 stabilizes the dimer form of MProM in which one active site is occupied by GC373, while the other is available for catalytic function. Clearly, and differently from MProM, a heterodimer is present in the absence of GC373 in an MPro1-199/MPro10-306 mixture. The heterodimer population increases upon GC373 binding, providing a population of stabilized heterodimers in which the active site of MPro10-306 is occupied by the inhibitor leaving a competent active site of MPro1-199 to enable the hydrolytic reaction. It is noteworthy that the catalytic activity MPro1-199/MPro10-306 heterodimer is like that of MProM homodimer in the presence of equimolar amount of GC373, whereas the heterodimer is somewhat more active than MProM in the absence of the inhibitor. Also, the results in Table 2 show that constructs MProTM, MPro1-199, and MPro10-306 have comparable activity observed from the monomer.

Unexpectedly, we observe a progressive increase in the size of the complex up to ∼102 kDa corresponding to a 1:1 to nearly 2:2 molecules of MPro1-199 and MPro10-306 as shown by SEC-MALS and AUC. Attempts to crystallize the MPro1-199 and MPro10-306 complex at ∼400 μM with 2.5-fold molar excess of NMV, ESV, or GC373 were unsuccessful likely due to the presence of reversible interconverting species in equilibrium. Also, inhibitor-free MPro10-306 atypical dimer without the 6H-tag did not yield crystals. Therefore, the above results are described based on higher order assemblies reported for SARS-CoV MPro (23, 28) and AlphaFold 3 predictions. The pathway to forming alternate encounter assemblies requires perturbing the contacts contributed by the free N terminus either by adding or deleting residues at the N terminus of MPro. In accordance, dimer forms of (-6∗)MProH41A, MPro10-306, and MPro197-306 are observed that are distinct in their monomer-dimer (M-D) equilibrium from that of the native MProWT dimer and resembling a characterized domain III FS-dimer of SARS-CoV. The inherent faster equilibrium observed for the assembly of MPro1-199 and MPro10-306 in the presence of GC373 could be an attribute of this heterodimeric system forming a less stable assembly as compared to (-6∗)MProH41A, MPro10-306, and MPro197-306 atypical dimers (Figure 5, Figure 6 and 6). The largest higher order structure reported for SARS CoV MPro is an octamer (i.e., four dimers) held together also through an FS pathway of domain III connecting the assembly (23). The structure of an isolated MPro187-306 FS-dimer has also been described (28). In this fold, the MPro187-306 FS-dimer is held together by switching the location of helix 201 to 214 (α1) with a second subunit and vice versa [Fig. 7, A and B, (28)].Figure 7 Preserved interface contacts upon association of MPro1-199and MPro10-306exhibiting possible alternate conformation of the helical domain.A and B, previously described alternate [via fold-switching (FS)] conformation of isolated SARS-CoV helical domain [B, PDB: 3EBN, (28)] compared with the helical domain of SARS-CoV-2 MProWT native dimer [A, PDB: 7JUN, (15)]. The five helices are labeled α1 to α5. C, overlay of MProC145A dimer [white, PDB: 7N89, (35)] and FS-dimer of the helical domain [blue and yellow, PDB: 3EBN, (28)]. The box highlights the region critical for dimer interface stability of mature MPro. D, despite the helical residues 201 to 214 switching (α1) their location to the opposite subunit, the positioning of the helix (293–301, α5) with critical inter-monomer E290/R4’ (prime denotes the second subunit) is clearly maintained as in the native conformation, which does not exhibit fold-switching.

As noted above, the largest complex observed is ∼102 kDa approaching two molecules each of MPro1-199 and MPro10-306 at the highest concentration of 54 μM (Fig. 2, Table 1). The scheme shown in Figure 8 summarizes the assembly of MPro1-199 and MPro10-306. The broad overlapping peaks observed by SEC and SV-AUC indicate slow exchange in the time scale of the measurements between the various observed species. SEC-MALS analysis of a mixture of MPro1-199/MPro10-306/GC373 indicates a stoichiometry with a higher amount of MPro10-306 than MPro1-199 leading up to forming complexes larger than a heterodimer. Thus, one possible pathway for such association is through fold-switching of domain III (Figure 7, Figure 8 and 8). An overlay of MPro dimer on MPro187-306 FS-dimer suggests that critical interface contacts of E290/R4′ and the free N-terminal Ser1 which forms the substrate subsite S1 pocket are clearly maintained even in a fold-switched conformation of domain III (Fig. 7C and D, 8E). Thus, this assembly will allow the association of a second MPro1-199 onto the trimer (Fig. 8, E and F). In this context, it is worth noting that the MPro10-306 atypical dimer exhibits the same low catalytic activity as that of MPro10-306 monomer even in the presence of GC373 (Table 2). These results indirectly suggest that MPro10-306 atypical dimer is formed through fold-switching of its domain III and therefore, cannot form a native-like dimer. Consistent with 1:1 heterodimer of MPro1-199 and MPro10-306, addition of MPro1-199 to MPro10-306 atypical dimer increases the activity by a factor of 9, which doubles in the presence of GC373 (Table 2, Fig. S2). This observation is consistent with the assembly of MPro1-199 onto MPro10-306 atypical dimer like as proposed in Fig. 8F. It remains to be explored if the intersubunit E290/R4′ salt bridge formation is linked to the disorder-to-order states of the α-helix 293 to 301 (α5, Fig. 7) as well as the domain III dynamics and thus, control the fold-switch pathway for domain III dimerization of MPro10-306 like as proposed for the isolated domain III of SARS-CoV (24).Figure 8 Predicted structures formed upon mixing MPro1-199and MPro10-306.A, room temperature X-ray structure of MProWT [gold, PDB: 7JUN, (15)] overlayed on cryo-X-ray monomer structure of SARS-CoV MProR298A [white, PDB: 2QCY, (16)]. The inactive conformation of the active site oxyanion loop and the rotated helical domain (residues 200–306), relative to the WT dimer, are shown in orange red. B and C, expected folded conformation of purified MPro10-306 and MPro1-199 prior to mixing either in the absence or presence of inhibitor. The scheme for their association is shown. FS denotes fold-switch. D and E, progressive assembly upon increasing protein concentration showing a stoichiometry of 1:1 (D), 2:1 (E), and 2:2 (F) of MPro10-306:MPro1-199. Theoretical molecular weights of the complexes are indicated in kilo Daltons (kDa). Molecular structure models were created using SARS-CoV MProWT octamer [PDB: 3IWM, (23)], helical domain of SARS-CoV [PDB: 3EBN, (28)], and SARS-CoV-2 MProWT and MProC145A [PDB: 7JUN and 7N89, respectively, (15, 35)].

It is reported that AlphaFold 3 demonstrates higher accuracy in predicting protein–protein interactions than AlphaFold-Multimer v2.3 (29, 30). However, it is noted that the predicted models are generally static structures and do not reflect the structural dynamics of proteins. Consequently, in some instances, the predicted structures are conformationally biased (29). Fig. S3 shows the predicted assemblies of MPro10-306 and MPro1-199. The predicted heterodimer is consistent with the enzyme activity results and resembles the native homodimer (Fig. S3A) like that shown in Fig. 8D. Interestingly, the higher order assembly of 2:1 displays domain III dimerization between two MPro10-306 molecules with an interface virtually identical to that of the mature dimer and by placing the catalytic domain of the second MPro10-306 in a random position (Figs. 9 and S3). We cannot exclude this possibility of a native-like interface formed by residues 280 and 283 to 286 through hydrophobic interactions of their side chains, with distances between carbon atoms of ∼4 Å, in conjunction with the intersubunit E290-R4′ salt bridge contact (Fig. 9). Such a conformation could allow the second MPro10-306 to pair with another catalytic domain to form the 2:2 assembly like that shown for the MPro10-306 FS-dimer associating with two molecules of MPro1-199 (Fig. 8F). The latter 2:2 assembly is very different from that predicted by AlphaFold 3 in which two molecules of MPro1-199 dimerize and form a symmetric interface with domain III provided by two molecules of MPro10-306 like that of a native dimer (Fig. S3, E and F). We exclude the possibility of MPro1-199 dimerization because of no detectable dimer population at 200 μM by AUC and the observed linear relationship between activity and protein concentration (11). The estimated Kdimer of MPro1-199 with GC373 is 1.2 ± 0.2 mM (11).Figure 9 Predicted structures formed upon mixing MPro1-199and MPro10-306. AlphaFold 3 model for the 2:1 complex of MPro10-306 and MPro1-199 showing the association of domain III through intermonomer contacts formed by residues 280 and 283 to 286 and the E290/R4′ salt bridge like that of the WT mature homodimer. All AlphaFold 3 generated assemblies are shown in Fig. S3.

In conclusion, we verified that a single N-finger region making interprotomer contacts mainly through the R4/E290′ salt bridge is sufficient to form a heterodimeric immature MPro intermediate with enhanced catalytic activity in the MPro autoprocessing pathway, as proposed in our earlier studies [Fig. 1, (4)]. Thus, N-terminal autoprocessing of the MPro appears to be regulated by an ensemble of conformations leading to stable dimer formation and mature-like catalytic activity.

Experimental procedures

MPro expression constructs and their designations

Constructs used in this work and their designations are listed in Table S1. The two constructs used in most experiments reported in this work are MPro1-199 and MPro10-306. New constructs prepared and reported for the first time are (GST-Tbin)MPro197-306-GS-6H, MPro10-306/C145A-GP-6H, MPro10-306/E290A/R298A-GP-6H, and (6H-TEV)MPro1-199/C145A. But for (GST-Tbin)MPro197-306-GS-6H, which was cloned into pD454-GST vector, the rest were cloned into pJ414 vector (ATUM).

Expression and purification

Plasmids were transformed into BL21-DE3 cells (Agilent) and induced for expression at an optical density (600 nm) of 0.7 to 0.8 with 1 mM IPTG, typically for 3 h at 37 °C. Proteins were purified from the cell lysate by NAC (step 1). The bound fraction was subjected to isocratic fractionation on Superose-12 column (step 2, Cytiva) in buffer A (25 mM Tris–HCl, pH 7.2, 150 mM NaCl, and 1 mM TCEP) or the corresponding buffer to facilitate protease cleavage. The 6His-tag was removed either with HRV-3C, TEV, or thrombin protease followed by NAC and SEC (step 3) as described previously for MProWT and MProM (11) in buffer B (25 mM Tris–HCl, pH 7.2, 50 mM NaCl, and 1 mM TCEP). Peak fractions were pooled and concentrated to the desired concentration using Amicon Ultra-15 or 0.5 ml centrifugal filters (Merck Millipore Ltd) and stored in aliquots at −30 °C and for long term storage at −80 °C. Purity and mass of constructs were verified both by SDS-PAGE and reverse-phase liquid chromatography with in-line electrospray ionization mass spectrometry (11). The same stock solutions of proteins, stored in aliquots, were used across different experiments. Protein concentrations were measured before storage and prior to the experiment at least in duplicate based on the extinction coefficient at 280 nm. Extinction coefficients (ε) of all constructs used in this study are listed in Table S1.

Sedimentation velocity analytical ultracentrifugation

Various constructs at the indicated protein concentrations of 10 to 50 μM in the absence and presence of up to 4-fold molar excess inhibitor were subjected to SV-AUC in buffer B (25 mM Tris–HCl, pH 7.2, 50 mM NaCl and 1 mM TCEP). Samples containing the inhibitor were prepared using a 1 to 4 mM stock solution of GC376 or NMV in buffer to achieve the desired protein and inhibitor ratios and incubated for a period of 1 to 2 h prior to filling the cells. ESV stock solution (100 mM in DMSO) was diluted in buffer to 1 mM and then added to the sample. GC376 and NMV (or PF-07321332) were purchased from Selleckchem and MedChemExpress, and ESV from MilliporeSigma.

Sedimentation velocity experiments were conducted at 50,000 rpm and 25 °C on a Beckman Coulter ProteomeLab XL-I analytical ultracentrifuge following standard protocols (31). Samples were loaded in 2-channel centerpiece cells and scans were collected using both the absorbance (280 nm) and Rayleigh interference (655 nm) optical detection systems. Sedimentation data were time-corrected and analyzed in SEDFIT 16.1C (32) in terms of a continuous c(s) distribution of Lamm equation solutions. Solution densities ρ, solution viscosities η, and protein partial specific volumes were calculated in SEDNTERP (33).

Size exclusion chromatography with multi-angle light scattering

Proteins were fractionated by analytical SEC with in-line MALS (DAWN Heleos-II, Wyatt Technology Inc.), refractive index (Optilab T-rEX, Wyatt Technology Inc.), and UV (Waters 2487, Waters Corporation) detectors. Sample (125 μl) was applied onto a pre-equilibrated Superose-12 column (1.0 × 30 cm, Cytiva) and eluted at a flow rate of 0.5 ml/min in buffer B (25 mM Tris–HCl, pH 7.2, 50 mM NaCl, 1 mM TCEP) at 25 °C. Molecular mass was calculated using the Astra software provided with the instrument.

SDS-PAGE and quantifying band intensities

Proteins were analyzed by SDS-PAGE on 4 to 20% gradient mini-protean TGX precast gel (Bio-Rad), stained with InstantBlue Coomassie Protein stain (abcam), and imaged using Bio-Rad Gel Doc EZ Imager (Bio-Rad). Band intensities were quantified using the ImageJ software (https://imagej.net).

Enzyme assays

Activity assays using the FRET substrate Dabsyl-KTSAVLQ/SGFRKM-E(Edans)-NH2 (6, 10), where (/) denotes the scissile peptide bond, were carried out in a total volume of 100 μl in buffer B (25 mM Tris–HCl, pH 7.2, 20–50 mM NaCl, and 1 mM TCEP) at 28 °C. Assays were initiated by adding the reaction mixture (95 μl) to 5 μl of 4 mM substrate in 100% DMSO kept in the microplate well (Reference 655809, Greiner bio-one). The excitation and emission wavelengths were set to 336 nm and 490 nm, respectively, and the increase in emission fluorescence intensity was recorded 2 to 4 times per data point as a function of time in a Tecan Infinite M plex microplate reader (Tecan). After background correction of the average of no enzyme negative controls, concentration of substrate cleaved was determined from an EDANS standard plot, and instrument-specific inner filter correction values were applied prior to calculating the rates as described (6, 10, 11, 34). The reproducibility of enzyme activity was tested at least 2 to 3 times with freshly prepared enzyme and stock solutions of the substrate and inhibitor. Once this was determined to provide consistent reaction rates within an error limit of 5%, the final experiment for the data displayed in the manuscript was carried out in duplicate and four reads per well for each time point. The mean of the data points was used for fitting. The same stock solutions of proteins, substrate, and inhibitor were used across different experiments shown in Figure 4. The substrate was custom synthesized (Biomatik). Stock solution of GC373 (10 mM) was prepared in assay buffer and diluted accordingly.

AlphaFold3 predictions

Structural models for MPro10-306 and MPro1-199 assemblies of stoichiometries 1:1, 2:1, and 2:2 were predicted using AlphaFold 3. The amino acid sequences for MPro10-306 and MPro1-199 were used as the input to the AlphaFold 3 server, specifying the number of copies of each protomer. The models were ranked by their predicted TM-score (pTM), and average predicted local distance difference test (pLDDT) and interface predicted TM-score (ipTM) to gauge the reliability of the predictions. The pTM scores are used to assess the quality of the overall assembly predictions whereas the pLDDT is used to evaluate the structure’s local confidence. The interactions between residues of different chains are scored by the ipTM. The metrics for the best ranking 1:1 heterodimer model are as follows: pTM = 0.68, pLDDT = 87.1, and ipTM = 0.51. For the 2:1 MPro10-306/MPro1-199 trimer, the metrics are as follows: pTM = 0.58, pLDDT = 83.2, and ipTM = 0.5. The metrics for the best 2:2 heterotetramer model are as follows: pTM = 0.55, pLDDT = 83.1, and ipTM = 0.48.

Data availability

All data that support the findings of this study are contained within this article and its accompanying files.

Supporting information

This article contains supporting information (4, 6, 10, 11, 14, 15, 20).

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.

Supporting information

Supporting Table and Figures

Acknowledgments

This work was supported by the 10.13039/100030692 Intramural Research Program of 10.13039/100000062 National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), 10.13039/100000002 NIH and the 10.13039/100006206 Office of Biological and Environmental Research supported research at ORNL's Center for Structural Molecular Biology (CSMB), a 10.13039/100006132 DOE Office of Science User Facility. ORNL is managed by UT-Battelle LLC for DOE’s Office of Science.

Author contributions

A. A., N. T. N., R. G., V. N. D., A. K., and J. M. L. methodology; A. A., N. T. N., R. G., V. N. D., A. K., and J. M. L. investigation; A. A., N. T. N., R. G., V. N. D., A. K., and J. M. L. writing–review and editing; N. T. N., A. K., and J. M. L. writing–original draft; A. K. and J. M. L. conceptualization; J. M. L. funding acquisition.

Funding and additional information

Funding Source: NIDDK; Project number: DK075166-01 (J. M. L.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
==== Refs
References

1 Wu F. Zhao S. Yu B. Chen Y.M. Wang W. Song Z.G. A new coronavirus associated with human respiratory disease in China Nature 579 2020 265 269 32015508
2 Yost S.A. Marcotrigiano J. Viral precursor polyproteins: keys of regulation from replication to maturation Curr. Opin. Virol. 3 2013 137 142 23602469
3 Yoshimoto F.K. The proteins of severe acute respiratory syndrome coronavirus-2 (SARS CoV-2 or n-COV19), the cause of COVID-19 Protein J. 39 2020 198 216 32447571
4 Aniana A. Nashed N.T. Ghirlando R. Coates L. Kneller D.W. Kovalevsky A. Insights into the mechanism of SARS-CoV-2 main protease autocatalytic maturation from model precursors Commun. Biol. 6 2023 1159 37957287
5 Xu J. Zhao S. Teng T. Abdalla A.E. Zhu W. Xie L. Systematic comparison of two animal-to-human transmitted human coronaviruses: SARS-CoV-2 and SARS-CoV Viruses 12 2020 244 32098422
6 Zhang L. Lin D. Sun X. Curth U. Drosten C. Sauerhering L. Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved alpha-ketoamide inhibitors Science 368 2020 409 412 32198291
7 Kneller D.W. Phillips G. O'Neill H.M. Jedrzejczak R. Stols L. Langan P. Structural plasticity of SARS-CoV-2 3CL M(pro) active site cavity revealed by room temperature X-ray crystallography Nat. Commun. 11 2020 3202 32581217
8 Goyal B. Goyal D. Targeting the dimerization of the main protease of coronaviruses: a potential broad-spectrum therapeutic strategy ACS Comb. Sci. 22 2020 297 305 32402186
9 Xia B. Kang X. Activation and maturation of SARS-CoV main protease Protein Cell 2 2011 282 290 21533772
10 Nashed N.T. Aniana A. Ghirlando R. Chiliveri S.C. Louis J.M. Modulation of the monomer-dimer equilibrium and catalytic activity of SARS-CoV-2 main protease by a transition-state analog inhibitor Commun. Biol. 5 2022 160 35233052
11 Nashed N.T. Kneller D. Coates L. Ghirlando R. Aniana A. Kovalevsky A. Autoprocessing and oxyanion loop reorganization upon GC373 and nirmatrelvir binding of monomeric SARS-CoV-2 main protease catalytic domain Commun. Biol. 5 2022 976 36114420
12 Chen S. Jonas F. Shen C. Hilgenfeld R. Liberation of SARS-CoV main protease from the viral polyprotein: N-terminal autocleavage does not depend on the mature dimerization mode Protein Cell 1 2010 59 74 21203998
13 Ferreira J.C. Fadl S. Rabeh W.M. Key dimer interface residues impact the catalytic activity of 3CLpro, the main protease of SARS-CoV-2 J. Biol. Chem. 298 2022 102023
14 Kovalevsky A. Coates L. Kneller D.W. Ghirlando R. Aniana A. Nashed N.T. Unmasking the conformational stability and inhibitor binding to SARS-CoV-2 main protease active site mutants and miniprecursor J. Mol. Biol. 434 2022 167876
15 Kneller D.W. Phillips G. Weiss K.L. Pant S. Zhang Q. O'Neill H.M. Unusual zwitterionic catalytic site of SARS-CoV-2 main protease revealed by neutron crystallography J. Biol. Chem. 295 2020 17365 17373 33060199
16 Shi J. Sivaraman J. Song J. Mechanism for controlling the dimer-monomer switch and coupling dimerization to catalysis of the severe acute respiratory syndrome coronavirus 3C-like protease J. Virol. 82 2008 4620 4629 18305031
17 Cheng S.C. Chang G.G. Chou C.Y. Mutation of Glu-166 blocks the substrate-induced dimerization of SARS coronavirus main protease Biophys. J. 98 2010 1327 1336 20371333
18 Silvestrini L. Belhaj N. Comez L. Gerelli Y. Lauria A. Libera V. The dimer-monomer equilibrium of SARS-CoV-2 main protease is affected by small molecule inhibitors Sci. Rep. 11 2021 9283 33927258
19 Chou C.Y. Chang H.C. Hsu W.C. Lin T.Z. Lin C.H. Chang G.G. Quaternary structure of the severe acute respiratory syndrome (SARS) coronavirus main protease Biochemistry 43 2004 14958 14970 15554703
20 Kovalevsky A. Aniana A. Coates L. Ghirlando R. Nashed N.T. Louis J.M. Visualizing the active site oxyanion loop transition upon ensitrelvir binding and transient dimerization of SARS-CoV-2 main protease J. Mol. Biol. 436 2024 168616
21 Owen D.R. Allerton C.M.N. Anderson A.S. Aschenbrenner L. Avery M. Berritt S. An oral SARS-CoV-2 M(pro) inhibitor clinical candidate for the treatment of COVID-19 Science 374 2021 1586 1593 34726479
22 Kneller D.W. Li H. Phillips G. Weiss K.L. Zhang Q. Arnould M.A. Covalent narlaprevir- and boceprevir-derived hybrid inhibitors of SARS-CoV-2 main protease Nat. Commun. 13 2022 2268 35477935
23 Zhang S. Zhong N. Xue F. Kang X. Ren X. Chen J. Three-dimensional domain swapping as a mechanism to lock the active conformation in a super-active octamer of SARS-CoV main protease Protein Cell 1 2010 371 383 21203949
24 Kang X. Zhong N. Zou P. Zhang S. Jin C. Xia B. Foldon unfolding mediates the interconversion between M(pro)-C monomer and 3D domain-swapped dimer Proc. Natl. Acad. Sci. U. S. A. 109 2012 14900 14905 22927388
25 Zhong N. Zhang S. Zou P. Chen J. Kang X. Li Z. Without its N-finger, the main protease of severe acute respiratory syndrome coronavirus can form a novel dimer through its C-terminal domain J. Virol. 82 2008 4227 4234 18305043
26 Noske G.D. Nakamura A.M. Gawriljuk V.O. Fernandes R.S. Lima G.M.A. Rosa H.V.D. A crystallographic snapshot of SARS-CoV-2 main protease maturation process J. Mol. Biol. 433 2021 167118
27 Noske G.D. Song Y. Fernandes R.S. Chalk R. Elmassoudi H. Koekemoer L. An in-solution snapshot of SARS-COV-2 main protease maturation process and inhibition Nat. Commun. 14 2023 1545 36941262
28 Zhong N. Zhang S. Xue F. Kang X. Zou P. Chen J. C-terminal domain of SARS-CoV main protease can form a 3D domain-swapped dimer Protein Sci. 18 2009 839 844 19319935
29 Abramson J. Adler J. Dunger J. Evans R. Green T. Pritzel A. Accurate structure prediction of biomolecular interactions with AlphaFold 3 Nature 630 2024 493 500 38718835
30 [preprint] Evans R. O’Neill M. Pritzel A. Antropova N. Senior A. Green T. Protein complex prediction with AlphaFold-Multimer bioRxiv 2022 10.1101/2021.10.04.463034
31 Zhao H. Brautigam C.A. Ghirlando R. Schuck P. Overview of current methods in sedimentation velocity and sedimentation equilibrium analytical ultracentrifugation Curr. Protoc. Protein Sci. 2013 Chapter 20, Unit20.12
32 Schuck P. Size-distribution analysis of macromolecules by sedimentation velocity ultracentrifugation and Lamm equation modeling Biophys. J. 78 2000 1606 1619 10692345
33 Cole J.L. Lary J.W. T P.M. Laue T.M. Analytical ultracentrifugation: sedimentation velocity and sedimentation equilibrium Methods Cell Biol. 84 2008 143 179 17964931
34 Liu Y.Y. Kati W. Chen C.M. Tripathi R. Molla A. Kohlbrenner W. Use of a fluorescence plate reader for measuring kinetic parameters with inner filter effect correction Anal Biochem. 267 1999 331 335 10036138
35 Kneller D.W. Zhang Q. Coates L. Louis J.M. Kovalevsky A. Michaelis-like complex of SARS-CoV-2 main protease visualized by room-temperature X-ray crystallography IUCrJ 8 2021 973 979
