Inhibiting a dynamic viral protease by targeting a non-catalytic cysteine

Summary

Viruses are responsible for some of the most deadly human diseases, yet available vaccines and antivirals address only a fraction of the potential viral human pathogens. Here, we provide methodology for managing human herpesvirus (HHV) infection by covalently inactivating the HHV maturational protease via a conserved, non-catalytic cysteine (C161). Using human cytomegalovirus protease (HCMV Pr) as a model, we screened a library of disulfides to identify molecules that tether to C161 and inhibit proteolysis, then elaborated hits into irreversible HCMV Pr inhibitors that exhibit broad-spectrum inhibition of other HHV Pr homologues. We further developed an optimized tool compound targeted toward HCMV Pr and used an integrative structural biology and biochemical approach to demonstrate inhibitor stabilization of HCMV Pr homodimerization, exploiting a conformational equilibrium to block proteolysis. Irreversible HCMV Pr inhibition disrupts HCMV infectivity in cells, providing proof of principle for targeting proteolysis via a non-catalytic cysteine to manage viral infection.

Graphical Abstract

eTOC

Hulce et al. develop an irreversible inhibitor targeting a non-catalytic cysteine in human herpesvirus proteases. Inhibition leads to disrupted infectivity, validating irreversible protease inactivation as a method to manage herpes viral infection.

Introduction

Viral infections are responsible for several of the most deadly outbreaks in human history, yet there are vaccines for only about 10% of the total human viral infections, and antiviral treatments available for even fewer (Razonable, 2011; Small and Ertl, 2011). Consequently, many viral infections remain a severe threat to human health with no effective routes to management. The human herpesviruses (HHV) belong to one of the most prevalent viral families and are associated with oral/genital herpes, chicken pox and shingles, infectious mononucleosis, roseola infantum, cancer and neurodegenerative disease (Gable, Acker and Craik, 2014; Fierz, 2017; Eimer et al., 2018; Hogestyn, Mock and Mayer-Proschel, 2018; Readhead et al., 2018; Tzeng et al., 2018). In particular, human cytomegalovirus (HCMV) causes serious symptoms or even death in immunodeficient patients, especially organ transplant patients and newborn infants (McIntosh, Hauschild and Miller, 2016; Francis et al., 2017; Rawlinson et al., 2017; El Helou and Razonable, 2019). The only currently approved HHV vaccine is against Varicella zoster virus to prevent chickenpox and shingles, and HHV antiviral drugs are toxic and suffer from emergent resistance (Gable, Acker and Craik, 2014; Lischka, Michel and Zimmermann, 2016).

An attractive new HHV therapeutic target is the viral maturational protease. The human herpesvirus proteases (HHV Prs) are a family of structurally conserved serine proteases essential for capsid maturation and viral replication (Gao et al., 1994; Dunn et al., 2001). The HHV Prs are expressed N-terminally fused to the assembly protein precursor (pAP, pUL 80) which is autoproteolyzed to release the free protease (Pr) during replication (Gable, Acker and Craik, 2014). Purification of free recombinant protease has allowed for extensive structural and functional studies (Burck et al., 1994). We and others have demonstrated that small molecule inhibition of proteolysis disrupts HHV infectivity, but to date no clinical compounds targeting HHV Pr have been developed (Waxman and Darke, 2000; Borthwick et al., 2003; Gable, Acker and Craik, 2014; Acker et al., 2017). HHV Pr drug discovery efforts focused on electrophilic small molecules targeting the catalytic serine (Gable, Acker and Craik, 2014). This approach requires high catalytic activity, which is challenging for HHV Pr due to the weak nucleophilicity of catalytic serine caused by a non-canonical His-His-Ser catalytic triad (Chen et al., 1996; Qiu et al., 1996; Shieh et al., 1996; Tong et al., 1996; Gable, Acker and Craik, 2014).

All HHV Prs are dynamic obligate homodimers with two independent active sites (Darke et al., 1996; Hoog et al., 1997; Qiu et al., 1997; Reiling et al., 2000; Buisson et al., 2002). Dimerization causes a disorder to order transition within the dimer interfacial helix and adjacent C-terminus (helices α5, α6), allosterically ordering the active site and allowing for catalysis (Batra, Khayat and Tong, 2001; Nomura et al., 2005; Zühlsdorf et al., 2015). The HHV Pr dimerization affinity is weak, with a K_d in the micromolar range, ensuring HHV Pr remains monomeric and inactive until packaged into the capsid, where a local concentration increase allows for homodimerization and protease activation (Nomura et al., 2005). This monomer-dimer equilbrium presents an opportunity to target protein dynamics to inhibit proteolysis via conformational capture.

Recent years have witnessed a renewed interest in the development of covalent protein inhibitors, which have been instrumental in managing challenging human diseases (Singh et al., 2011; De Cesco et al., 2017; Roeten, Cloos and Jansen, 2018). While best known for applications in cancer, covalent protein inhibition also presents a powerful method for managing infectious diseases (Bauer, 2015). Covalent HHV Pr inhibition presents an efficient route to capture and inactivate these dynamic enzymes, but history highlights the challenges associated with targeting the catalytic serine (Gable, Acker and Craik, 2014). We chose to take an alternative approach to covalent HHV Pr inhibition by targeting a non-catalytc cysteine (Liu et al., 2013; Miller and Taunton, 2014; Hallenbeck et al., 2017; Janes et al., 2018; Canon et al., 2019). Protease inhibition via a non-catalytic cysteine has shown promise for the treatment of both viral and bacterial infections (Hagel et al., 2011; Turner et al., 2021). Here, we expand this approach to target the HHV Pr family.

We developed covalent inhibitors that target a conserved, non-catalytic cysteine (C161) adjacent to the protease active site. Focusing on HCMV Pr, we performed a disulfide tethering screen to identify molecules that tether to C161 and inhibit proteolytic activity (Burlingame, Tom and Renslo, 2011; Turner, Tom and Renslo, 2014; Hallenbeck et al., 2018). We assigned the HCMV Pr ¹H-¹⁵N amide backbone and ¹H-¹³C ILVM methyl sidechain resonance shifts to develop an HSQC platform suitable for parallel inhibitor binding studies. We converted disulfide-tethered hits into irreversible inhibitors that also inhibit herpes simplex virus 1 protease (HSV1 Pr) and Epstein-Barr virus protease (EBV Pr), demonstrating this approach as a generalizable method for inactivating HHV Pr family members. We elaborated irreversible inhibitors into an optimized tool compound that disrupts HCMV Pr activity by exploiting a conformational equilbrium to stabilize the inhibited homodimer. Irreversible protease inhibition blocks HCMV infectivity in cells, validating targeting a non-catalytic cysteine as a method to manage viral infection. This work lays the foundation for an inhibitor development pipeline exploiting conserved, non-covalent cysteines to capture dynamic enzymes and inactivate a family of essential viral proteases.

Results

Disulfide tethering to C161 disrupts HCMV Pr activity

HCMV Pr contains five endogenous cysteine residues, all of which are reduced (Figure 1a) (Burck et al., 1994; Waxman and Darke, 2000). Titration of native HCMV Pr with 5,5’-dinitrobenzoic acid indicates three to four of the cysteines are reactive, and treatment with iodoacetamide reduces proteolytic activity (Stevens et al., 1994; Pinko et al., 1995; Flynn, Abood and Holwerda, 1997). HCMV Pr is inhibited by molecules that mediate disulfide linkage between endogenous cysteine residues or by small molecule cysteine alkylation (Baum, Ding, et al., 1996; Baum, Siegel, et al., 1996; Pinto et al., 1999; Waxman and Darke, 2000; Khayat et al., 2003). Of the HCMV Pr Cys residues, C161 is the only one conserved across the HHV Pr family, making it an ideal candidate for covalent inhibitor development (Figure 1b) (Waxman and Darke, 2000). Residue 161 is strategically positioned near the enzyme active site, providing the potential to take advantage of interactions with the substrate binding pocket (Khayat et al., 2003).

Figure 1. Disulfide tethering to C161 disrupts HCMV Pr activity.

a) The five endogenous HCMV Pr cysteine residues are mapped onto the X-ray crystal structure (PDB code: 1NJU). One monomeric chain is blue, the second is greencyan. The catalytic triad is red and the cysteine residues are yellow. b) Multiple sequence alignment of HHV Pr family members, showing that C161 is absolutely conserved. c) Disulfide-tethering screening and hit validation workflow. See also Figure 2 and Table S1.

Though it is conserved, mutagenesis studies indicate C161 is not required for proteolytic activity (Welch et al., 1993; Cox et al., 1995; Baum, Siegel, et al., 1996; Waxman and Darke, 2000). To probe HCMV Pr enzymatic activity, we synthesized a donor/quencher fluorescent substrate then determined k_cat/K_m of WT HCMV Pr (658 s⁻¹M⁻¹, Figure S1a). We mutated C161 to alanine (C161A) and determined k_cat/K_m to be 313 s⁻¹M⁻¹, which is only two-fold decreased compared to WT HCMV Pr, confirming that C161 makes some contribution to activity, but is not required for catalysis. To specifically target C161 for inhibition, we mutated the four remaining cysteines (C84, C87, C138 and C202) to alanine. For simplicity, the HCMV C84A, C87A, C138A, C202A variant is referred to as HCMV C-A in the remainder of this work. We determined k_cat/K_m of HCMV C-A to be 600 s⁻¹M⁻¹ (Figure S1a). Since HCMV C-A retains over 90% activity compared to WT HCMV Pr, the variant remains structurally and functionally comparable to the WT enzyme.

We screened a library of 1579 disulfide-capped fragments for molecules that disulfide tether to C161 and identified twenty-nine unique hits that showed greater than 52.4% labeling (3-sigma from average labeling, 1.9% hit rate, Figure 1c, Table S1) (Burlingame, Tom and Renslo, 2011; Turner, Tom and Renslo, 2014; Hallenbeck et al., 2018). Fifteen of the 29 hits were comprised of a five-membered heterocyclic core with a terminal aryl moiety. We selected ten aryl triazole hits with diverse aryl substitutions and evaluated their inhibition of in vitro substrate turnover by HCMV C-A. These compounds inhibit proteolytic activity with IC₅₀ values ranging from 41–334 μM, validating that crosslinking to C161 is an effective method to inhibit HCMV Pr activity (Table S1). For comparison, we selected eleven additional hits with chemically distinct scaffolds and tested for inhibition (Table S1). While inhibition is observed, IC₅₀ values were overall less potent than the aryl triazole inhibitors. We thus chose to focus on the aryl triazole scaffold for further study.

Aryl triazole inhibitors engage the HCMV Pr active site

We developed an HSQC NMR platform for medium-throughput screening of the aryl triazole inhibitor binding modes. We utilized two previously engineered obligate monomer constructs, L222D and a C-terminal truncation (Δ221), as starting points to assign our drug-bound HCMV spectra (Lee et al., 2011). In this present study, we provide more complete coverage of the HCMV Pr structure, by assigning 81% of the HCMV Δ221 ¹H-¹⁵N amide backbone resonances as well as the methyl/amide/indole/guanidine side chains of several Met, Val, Leu, Thr, Asn, Gln, Trp and Arg residues (BRMB accession number 51211).

Because the C-terminal truncation would remove residues nearby C161 and possibly disrupt inhibitor binding (Figure 1a), we chose to conduct HSQC experiments with a full-length HCMV C-A obligate monomer construct (HCMV C-A L222D) (Figure S1b-d). We collected the ¹H-¹⁵N HSQC of HCMV C-A L222D in the presence of four aryl triazole analogs (15, 18, 23, 24, Figure 2a). We inspected the DMSO control spectrum and selected peaks that do not show significant spectral crowding and peaks that have a signal to noise ratio greater than three (Table S2). We calculated chemical shift perturbations (CSP) in parts per million (ppm) (Lee et al., 2011) and determined normalized peak volumes (Sanulli et al., 2019), then took the ratio V_I/V_DMSO to compare peak volumes in the presence and absence of inhibitor. Any residue that demonstrated a CSP of 0.05 ppm or greater or V_I/V_DMSO less than 0.15 was considered significantly perturbed.

Figure 2. Aryl triazole inhibitors engage the HCMV Pr active site.

a) ¹H-¹⁵N HSQC NMR spectra of HCMV-L222D bound to disulfide tethered inhibitors (15, 18, 23 and 24). The seven annotated residues (H63, G130, L133, A159, G168, T169, R175, V180) were significantly perturbed (CSP of 0.05 ppm or above, or V_I/V_DMSO of less than 0.15, see Methods). See also Table S2. b) Mapping of the perturbed residues onto the HCMV Pr crystal structure (PDB code: 1NJU). Navy blue residues are perturbed by all four of the tested inhibitors, marine blue are perturbed by three, pale cyan by two.

Similar residues were perturbed by most of the aryl triazole inhibitors tested (Figure 2b). Except for two residues located on the backside of the protease (R175, V180), the majority of the perturbed residues cluster to the HCMV Pr active site and include members of the catalytic triad (H63), substrate binding pocket (L133, A159) and oxyanion hole loop (G168, T169). G130 is located behind C161 and CSP of this residue is a readout for covalent labeling at C161. Based on HSQC data, we conclude that aryl triazole inhibitors react with C161 and engage the HCMV Pr active site, sharing a conserved competitive mode of inhibition.

Chloromethyl ketone electrophiles inhibit HCMV, HSV1 and EBV proteases

With an understanding of the tethered inhibitor binding mode, we next sought to develop disulfide tethered hits into irreversible HCMV Pr inhibitors. Starting with a 4-phenyl-1,2,3-triazole scaffold, we synthesized analogs bearing a cysteine-reactive moiety (epoxide, chloromethyl ketone or tetrafluorophenoxymethyl ketone) in place of the disulfide and used intact protein LC/MS to test for labeling of HCMV C-A (Figure S2a). Only the chloromethyl ketone (CMK) inhibitor 31 showed reactivity with C161. We determined 31 inhibition of HCMV C-A and WT HCMV Pr (IC₅₀ values of 25 μM and 33 μM respectively, Table 1), confirming that inhibitors identified against HCMV C-A are translatable to WT HCMV Pr. We determined the K_I and k_inact of 31 against HCMV C-A to be 38 μM and 0.026 min⁻¹, respectively (Table 1 and Table S3).

Table 1. Inhibition of HHV Pr with chloromethyl ketone inhibitors.

Several molecules were synthesized containing an aryl triazole core and chloromethyl ketone warhead, with C-4 substitution (R1, compounds 34-40) or a fixed C-4 phenyl with varying C-5 substituents (compounds 41-47). IC₅₀ values are reported at one hour incubation with WT HCMV Pr, HCMV C-A, HSV1 Pr or EBV Pr. K_I and k_inact values are reported for HCMV C-A (Table S3). The symbol ---indicates the value was not tested for that particular protease/inhibitor.

			WT HCMV Pr	HSV1 Pr	EBV Pr	HCMV C-A
Compound	R1	R2	IC50 (μM)	IC50 (μM)	IC50 (μM)	IC50 (μM)	KI (μM)	kinact (min−1)	t1/2 (min)	kinact/KI (min−1 μM−1)
31		H	33	---	---	25	38	0.026	26	6.8 × 10−4
34		H	6	51	---	22	---	---	---	---
35		H	11	207	21	33	---	---	---	---
36		H	5	35	17	25	---	---	---	---
37		H	5	98	11	32	---	---	---	---
38		H	6	145	8	30	---	---	---	---
39		H	18	188	52	46	---	---	---	---
40		H	21	160	82	67	---	---	---	---
41			68	>1000	7	60	---	---	---	---
42			33	>1000	---	101	---	---	---	---
43			4	>1000	2	33	9	0.024	29	2.7 × 10−3
44			13	840	12	54	29	0.022	31	7.6 × 10−4
45			53	670	4	66	26	0.020	35	7.7 × 10−4
46			---	---	---	182	---	---	---	---
47			---	---	---	135	21	0.016	44	7.6 × 10−4

We next synthesized several CMK analogs with structures that mimic our original disulfide-tethered hits and tested for inhibition of HCMV C-A and WT HCMV Pr (Table 1, compounds 34-40). Inhibition is maintained for all CMK inhibitors tested, and potencies against WT HCMV Pr were two to six-fold better than against HCMV C-A. We tested 31 for dose-dependent labeling of WT HCMV Pr using intact protein LC/MS and observed up to three cysteine labeling events (Figure S2b). Mutation of C161 to alanine abolishes one of these labeling events, confirming that this residue is targeted in WT HCMV Pr. All three HCMV Pr contructs (WT, HCMV C-A and C161A) are at least partially inhibited by 31, and additional cysteine labeling likely contributes to the improved potency against WT HCMV Pr (Figure S2c). In addition to C161, three other HCMV Pr cysteine residues (C87, C138 and C202) are known to be reactive with covalent small molecule inhibitors (Baum, Siegel, et al., 1996). Of these, C138 is located adjacent to the S4 substrate binding pocket and is known to be involved in HCMV Pr inhibition with covalent benzimidazolylmethyl sulfoxides (Flynn, Abood and Holwerda, 1997). Based on this, we hypothesize the residual inhibition observed with HCMV C161A is due to C138 reactivity, allowing for obstruction of the active site and contributing to the improved potency against WT HCMV Pr.

Because C161 is conserved across the HHV Pr family (Figure 1b), we hypothesized that CMK inhibitors could disrupt activity of HHV Pr homologues. To test this, we determined dose-dependent inhibition of herpes simplex virus 1 protease (HSV1 Pr) and Epstein-Barr virus protease (EBV Pr) (Table 1). Both proteases are inhibited with potencies of 35–207 μM (HSV1 Pr) and 8–82 μM (EBV Pr). This indicates that HCMV Pr inhibitors are translatable to other HHV Pr homologues, presenting the possibility for developing broad-spectrum covalent HHV Pr inhibitors.

Development of an optimized tool compound to study HCMV Pr inhibition

With proof of concept for inhibiting HHV Prs with electrophilic CMK inhibitors, we next sought to optimize inhibitors into tools suitable for mechanistic studies against HCMV Pr. HSQC data indicates aryl triazole disulfide tethered hits engage the HCMV Pr active site (Figure 2b) and we hypothesized that elaboration at C-5 of the core 1,2,3-triazole would allow for further engagement of the surrounding protein pocket and lead to improved inhibitory potency. We synthesized several 4-phenyl-1,2,3-triazole CMKs with an additional carbamate side chain at C-5 using regioselective, ruthenium-catalyzed azide-alkyne cycloaddition that directs internal alkynes with hydrogen bond donating groups towards the C-5 position in the resulting 1,4,5-trisubstituted products (Boren et al., 2008). These C-5 substituted molecules exhibited a five-fold range of IC₅₀ values against HCMV C-A and a wider 17-fold range against WT HCMV Pr (Table 1, compounds 41-47). We also tested several C-5 substituted molecules for inhibition of HSV1 and EBV Prs. While inhibition is maintained for EBV Pr (IC₅₀ values 2–12 μM), these molecules show weak to no inhibition of HSV1 Pr. Thus, inhibitor selectivity can be modulated by substitutions within the core 1,2,3-triazole. We determined K_I and k_inact against HCMV C-A of 43, 44, 45 and 47 (Table 1 and Table S3). While analogs 44, 45 and 47 have K_I values similar to 31, which lacks a C-5 side chain, analog 43 was 4-fold more potent, with a K_I of 9 μM. Based on this improved K_I value, we selected 43 to carry forward for further structural and mechanistic studies.

Inhibitor 43 binding mode is influenced by HCMV Pr dimerization

To determine binding mode of inhibitor 43, we used a combination of HSQC, X-ray crystallography and molecular dynamics (MD) simulations. We collected the ¹H-¹⁵N HSQC of 43 bound to HCMV C-A L222D, then quantified CSP and integrated peak volume (Figure S3a and Table S4). Four residues (G130, S132, G164 and T169) are significantly perturbed (Figure 3a). These residues cluster to the S₁ substrate pocket, suggesting that 43 engages the HCMV Pr active site.

Figure 3. Inhibitor 43 binding mode is influenced by HCMV Pr dimerization.

a) The 43/HCMV C-A co-crystal structure solved to 2.67 Å. Chain A of the homodimer is white, chain B is marine blue. Inhibitor 43 is aquamarine. The four residues perturbed in HSQC (G130, S132, G164, T169) are magenta. The S₁-S₁’ substrate pockets and the dimer interfacial helix α5 are labeled. b) The 2Fo-Fc electron density map (blue, 1.0 RMSD) and the Fo-Fc difference map (3.0 RMSD) is shown. c) The polder OMIT map (green, 2.5 RMSD) is shown. d) MD simulations modeling 43 binding to HCMV C-A. Left: Static views of the most representative binding poses of 43 in a single HCMV C-A chain, accounting for 86–100% of the MD trajectories from four independent MD simulations (Video S1, 2). Right: Static views of the most representative binding poses of 43 in the HCMV C-A homodimer, accounting for 99–100% of the MD trajectories from three independent MD simulations (Video S3). e) The interaction between R109 (green) on chain B and the C-5 phenyl of 43 (aquamarine) bound to chain A. The 2Fo-Fc electron density map (blue, 1.0 RMSD) and the Fo-Fc difference map (3.0 RMSD) is shown. The distance from the R109 guanidine sidechain to the C-5 phenyl is 3.7 Å.

To directly address the mode of binding, we co-crystallized HCMV C-A with 43 and solved the structure to 2.67 Å, using a CC_1/2 standard of 0.3 to determine optimal resolution cut-off (PDB code 7TCZ, Table S5)(Karplus and Diederichs, 2012). Inspection of the 2Fo-Fc electron density map at C161 indicates partial density corresponding to 43 binding (Figure 3b). To verify the presence of inhibitor 43, we calculated a polder OMIT map (Liebschner et al., 2017), which also reveals non-continuous density at C161 (Figure 3c). We attempted co-crystallization with other CMKs (31, 35, 39-47) and an analogue of 43 containing a 4-phenyl moiety para-substituted with bromine as methods to improve electron density within the inhibitor binding site. Unfortunately, these efforts yielded structures with no density for the inhibitor. Efforts to process the data to alternative resolution cutoffs of 2.43 Å or 2.93 Å resulted in poor data reduction statistics or loss of electron density at C161, respectively. We thus selected the 2.67 Å 43/HCMV C-A co-crystal structure as a balance between acceptable model quality and maintaining enough information to analyze inhibitor binding mode, carrying this structure forward for further analysis.

Taking the 2Fo-Fc and polder OMIT maps together, we conclude that 43 is present in the crystal structure. The incomplete density may be either due to incomplete labeling at C161 or sampling of multiple binding modes. The 43/HCMV C-A co-crystal structure was obtained by diluting HCMV C-A to 0.5 μΜ in Tris buffer then labeling with excess 43 overnight, followed by concentration and crystallization (see Methods). To determine whether the incomplete density at C161 is due to partial labeling at C161, we incubated diluted HCMV C-A (5 μM in Tris) with 50 μM 43 or DMSO overnight, then used intact protein LC/MS to determine labeling (Figure S3b). When HCMV C-A is incubated with 43, the peak corresponding to apo protein (28,924 Da) disappears and is replaced by an M + 363 Da peak corresponding to 43-bound protein (29,287 Da). Thus, HCMV C-A is fully labeled with 43 under the conditions used for crystallography. To verify that labeling is indeed occurring at C161, we generated a protein construct lacking all cysteine residues (HCMV C84A, C87A, C138A, C161A, C202A, referred to as HCMV C-A C161A for the remainder of this text) then repeated overnight labeling experiments with 43. The identified mass (28,892 Da) is the same for the DMSO or 43 treated samples, indicating no labeling occurs when C161 is removed. Thus, in samples generated for crystallography, we conclude 43 labels HCMV C-A to 100% at C161.

Based on LC/MS labeling experiments, we conclude that the partial 43 occupancy observed in our structure is due to inhibitor flexibility and sampling of multiple binding modes. We performed occupancy refinement to determine a final inhibitor occupancy of 86%. The remaining 14% could be explained by inhibitor flexibility, leading to loss of well-defined electron density. In both the 2Fo-Fc and polder OMIT maps there is density present at C161 as well as at the HCMV Pr dimer interface. Based on this, we modeled in 43 with a final binding mode that engages the S₁’ substrate pocket as well as the dimer interface (Figure 3a-c).

The HSQC and X-ray crystallography data suggest two alternative 43 binding modes, where 43 engages either the S₁ substrate pocket or the dimer interface. Realizing that the HSQC experiments are performed with obligate monomeric HCMV C-A L222D, whereas 43 co-crystallizes with the HCMV C-A homodimer, we explored if the oligomerization state could explain the observed differences in binding mode. A series of MD simulations were performed to model 43 binding to a single monomeric HCMV C-A chain or to the HCMV C-A homodimer (sample simulations provided, Videos S1-3). Clustering analysis was used to identify the most representative binding poses in each simulation. As shown in Figure 3d (left), the C-5 sidechain of 43 is oriented toward either the dimer interface or the S₁ substrate pocket in the HCMV C-A monomer, explaining the CSP observed in HSQC experiments. Because the HCMV Δ221 construct used to assign the ¹H-¹⁵N amide backbones lacks the dimer interfacial helix α5, the residues within the dimer interface are unassigned, which is why CSP are not observed in this region. In these MD simulations, the aryl triazole core makes contacts with R165, the main chain of which is known to form the HCMV Pr oxyanion hole (Videos S1, 2) (Tong et al., 1998).

For the HCMV C-A homodimer (Figure 3d, right), the MD results indicate that 43 makes contacts with both monomers simultaneously, keeping the inhibitor oriented toward the dimer interface. In addition to contacts with R165, as observed in the monomer, 43 makes additional contact with R109 in the second monomeric chain through its C-5 phenyl group, anchoring the inhibitor in an orientation facing the dimer interface (Video S3). We inspected the X-ray crystal structure to determine whether these interactions are observed. While R165 is flexible and undefined in the X-ray crystal structure, R109 shows well-resolved density and is located 3.7 Å from the C-5 phenyl, in proximity to make a cation-π interaction (Figure 3e). This stabilizing interaction could explain why density for the C-5 phenyl is observed in the X-ray crystal structure, as R109 would anchor this moiety, while the aryl triazole core is able to rotate around the carbamate linker. Taking the structural and MD data together, we conclude that alternative 43 binding modes are possible, and that binding mode is influenced by HCMV Pr dimerization.

Inhibitor 43 mechanism exploits HCMV Pr conformational equilibrium

Structural and MD simulation data suggest that HCMV Pr dimerization influences the binding mode of 43, leading us to consider whether inhibitor binding alters the HCMV Pr monomer/dimer equilibrium. To test this, we analyzed dimerization of HCMV C-A using size-exclusion chromatography (SEC) (Darke et al., 1996). In addition to the L222D obligate monomer, we identified a point mutation within helix α5 that promotes dimer formation (S225M). Both the HCMV C-A L222D monomer and the HCMV C-A S225M dimer elute as a single peak with retention times of 18.3 min (L222D monomer) and 16.8 min (S225M dimer) (Figure 4a). By contrast, HCMV C-A elutes as a broad peak that is a mixture of monomeric and dimeric species. HCMV C-A fully labeled by 43 elutes as a mixture favoring the dimeric species. To further study this equilibrium, we incubated HCMV C-A with either DMSO or 43 then analyzed dimerization at several protein concentrations using SEC (Figure 4b). Reaction with compound 43 at C161 biases the equilibrium toward the dimeric species at all protein concentrations, confirming that 43 indeed stabilizes the HCMV Pr dimer, but in a nonproductive state, as evidenced by the loss of activity (Table 1).

Figure 4. Inhibitor 43 mechanism exploits HCMV Pr conformational equilibrium.

a) SEC traces of the HCMV C-A L222D, HCMV C-A S225M, HCMV C-A and HCMV C-A fully labeled with 43, all at 0.9μΜ protein. b) SEC traces of HCMV C-A at varying concentrations incubated with either DMSO or 43. c) Percent labeling of 43 over time, as determined by intact protein LC/MS for HCMV C-A (blue), HCMV C-A L222D (black) and HCMV C-A S225M (red). Data are reported in triplicate from technical replicates as the mean ± SD.

The observation that 43 binding stabilizes the HCMV Pr dimer led us to consider whether the reaction with C161 might be more favorable for the monomeric vs. dimeric protein state. We used intact protein LC/MS to analyze time-dependent 43 labeling of HCMV C-A, the HCMV C-A S225M dimer and the HCMV C-A L222D monomer (Figure 4c). The rate of labeling (k_obs) was 0.093 min⁻¹ for HCMV C-A, and approximately 1.5-fold slower for both the HCMV C-A S225M dimer (k_obs = 0.059 min⁻¹) and the HCMV C-A L222D monomer (k_obs = 0.050 min⁻¹). Both L222 and S225 are located within helix α5 and are far away from 43, thus we conclude that changes to k_obs are due to changes to the monomer/dimer equilibrium rather than direct interactions between L222 or S225 and bound inhibitor (Figure S3c). This indicates that 43 reacts faster with HCMV C-A in equilibrium than either the monomer or dimer in isolation.

Disfavored reactivity with the HCMV Pr dimer may explain the premature saturation of labeling/inhibition of HCMV C-A S225M. Dithiothreitol (DTT) was used to quench the above reactions and could be potentially reactive with the compound 43/HCMV C-A complex. To test whether DTT affects 43 reactivity with C161, we performed time dependent labeling experiments in the absence of DTT (Figure S3d). Under these conditions, HCMV C-A S225M reacts with 43 at a two-fold slower rate than HCMV C-A, but in both cases protein labeling proceeds to 100%. This suggests the formation of an unstable covalent adduct that is susceptible to DTT attack in one of the HCMV C-A S225M chains. To further probe whether compound reactivity with the HCMV C-A S225M dimer is disfavorable, we determined K_I and k_inact of 43 against HCMV C-A S225M (Table S3). The K_I is 13μΜ, similar to the K_I of 9μM against HCMV C-A, indicating that binding affinity is not altered between the two protein constructs. However, k_inact is reduced from 0.024min⁻¹ against HCMV C-A to 0.011min⁻¹ against HCMV C-A S225M, verifying that reactivity with the pre-formed dimer is disfavorable.

Chloromethyl ketone inhibitors disrupt ΗCMV infectivity

We next used compound 43 to test the feasibility of managing HCMV infectivity by inhibiting protease activity via C161. We began by monitoring dose-dependent cell killing of a panel of human cell lines, including human foreskin fibroblasts (HFF-1), human fetal lung fibroblasts (MRC-5), human hepatocellular carcinoma cells (Huh7) and human cervical adenocarcinoma cells (HeLa) (Figure S4a). The 50% cytotoxic concentration (CC₅₀) values range from 4–20 μM across the cell lines tested. The CC₅₀ of approved HCMV antivirals ganciclovir (GCV) and letermovir (LET) are typically >30 μΜ, whereas that of the general protein synthesis inhibitor cycloheximide (CHX) ranges from mid to high nanomolar (Takashi E. Komatsu, 2016; Takashi E. Komatsu, Eric F. Donaldson, Anamaris M. Colberg Poley, 2017). Thus, while the cytotoxicity of 43 is greater than that of GCV and LET, it is significantly less than that of a non-selective cytotoxin like CHX.

Of the two cell lines tested that are amenable to HCMV infection (MRC-5, HFF-1), 43 was less toxic to HFF-1 cells, with a CC₅₀ value of 20 μM. We therefore elected to test 43 antiviral activity in HFF-1 cells (Hippenmeyer and Dilworth, 1996). Compound 43 causes a dose-dependent decrease in HCMV replication, with an EC₅₀ of 5 μM (Figure S4b). The EC₅₀ of GCV under the same assay conditions was 4 μM, similar to that of 43. We determined the CC₅₀ value in parallel to antiviral EC₅₀ to control for general cell killing over the course of the assay (Figure S4b). The CC₅₀ value of compound 43 is 11 μΜ, two-fold higher than the antiviral EC₅₀.

To verify that antiviral activity is due to true disruption of viral replication, we tested for 43 inhibition of HCMV reinfectivity. We infected HFF-1 cells with HCMV in the presence of varying concentrations of 43 as described above to determine antiviral EC₅₀. On day three, instead of measuring viral replication, we collected the media and used it to reinfect HFF-1 cells in the absence of inhibitor. Since media contains infectious viral progeny, any decrease in reinfectivity should be a result of decreased infectious virions, reflecting disruption of lytic replication caused by 43. On day three post-reinfection, we observed a dose-dependent decrease in viral replication (EC₅₀ 0.8 μM, Figure 5a). The EC₅₀ of GCV under the same conditions was slightly better, with EC₅₀ of 0.1 μM. We also determined cell death three days post-reinfection to be minimal, indicating no carryover of the cytotoxic effect of 43. From this, we conclude that 43 disrupts HCMV replication and the production of infectious viral progeny.

Figure 5. Chloromethyl ketone inhibitors disrupt HCMV infectivity.

a) The effects of 43, 48 and GCV on HCMV replication or HFF-1 cell viability three days post-reinfection. Data are reported in triplicate (reinfection) or duplicate (cytotoxicity) as the mean ± SD. b) In-gel fluorescence competition assay monitoring 43 labeling of exogenous HCMV C-A introduced into HFF-1 cell lysates (Figure S4c). Normalized fluorescence intensity was fit to determine binding affinity (blue curve). c) Alkyne-bearing CMK analog inhibitor 48. In vitro IC₅₀ against WT HCMV Pr is reported for 43 and 48. d) In-gel fluorescence assay monitoring reactivity of 48 with HCMV C-A in HFF-1 cell lysates, quantified as described for panel b. e) 48 labeling of FLAG-tagged HCMV Pr overexpressed in HEK293T cells. Left: Cy3/Cy5 fluorescence; middle: Coomassie stained gel; right: Anti-FLAG western blot (Figure S4d). HCMV Pr is labeled, as are two control proteins used for quantification. f) Normalized fluorescence intensity of HCMV Pr, Control 1 and Control 2.

To validate that engagement of HCMV Pr by compound 43 occurs in the HFF-1 cellular environment, we used a fluorescence competition assay to monitor protease labeling in cell lysates. We introduced purified HCMV C-A into HFF-1 cell lysate then treated with varying concentrations of 43 for 60 min (Figure S4c). After incubation, we treated the lysate with a pan-Cys reactive maleimide alkyne probe and performed click-chemistry to TAMRA-azide. Separation of protein targets by SDS-PAGE allowed for direct in-gel fluorescence readout of the ability of 43 to block maleimide labeling. We quantified labeling of HCMV C-A and found 43 to have an EC₅₀ of 3 μM, which falls within similar ranges to previously measured antiviral and in vitro inhibitor potency (Figure 5b). This validates 43 engages HCMV C-A in a complex biological system.

To further validate that HCMV Pr is engaged by CMK inhibitors in cells, we developed a new CMK probe containing an alkyne handle in place of the C-5 phenyl sidechain (48), allowing for direct detection of HCMV Pr labeling (Figure 5c). Inhibitor 48 disrupts in vitro proteolysis of WT HCMV Pr with an IC₅₀ of 8 μM and blocks HCMV reinfectivity with an EC₅₀ of 1.8 μM, indicating that incorporation of the alkyne decreases potency by ~two-fold but that overall inhibition is conserved (Figure 5a, c). Inhibitor 48 is also reactive with HCMV C-A in HFF-1 cell lysates and exhibits an EC₅₀ of 4 μM (Figure 5d and Figure S4c).

To determine whether CMK inhibitors are able to permeate the cell membrane and react with HCMV Pr in cells, we treated HEK293T cells overexpressing FLAG-tagged WT HCMV Pr with varying concentrations of 48 then used our in-gel fluorescence assay to determine protease labeling (Figure S4d). Several proteins exhibit a fluorescent signal increase, indicating 48 is able to permeate the cell and react with cellular targets, likely explaining the cytotoxicity observed in antiviral assays (Figure 5e). We noticed that the FLAG-tagged HCMV Pr construct travels at a higher than expected molecular weight compared to His-tagged purified enzyme, presumably due to the introduction of the highly negative FLAG sequence (Figure S4d). To verify the location of HCMV Pr on the gel, we performed anti-FLAG western blot and identified HCMV Pr as a strong fluorescence band located at the 35 kDa molecular weight marker (Figure 5e and Figure S4d). Comparing the fluorescently labeled bands to their counterparts in the Coomassie-stained gel indicates that, while HCMV Pr is not the most abundant protein expressed in HEK293T cells, it produces one of the strongest fluorescence signals that is enriched over more abundantly expressed proteins (Figure 5f). This indicates preferential reactivity of 48 with HCMV Pr. Taken together, in-gel fluorescence data validates inhibitor cell permeability and on-target reactivity, indicating the antiviral activity observed for both 43 and 48 is likely due to inhibition of HCMV Pr activity.

Discussion

We demonstrated irreversible inhibition of HCMV Pr targeting a conserved, non-catalytic cysteine (C161), expanding the toolkit for managing viral infection. Irreversible inhibitors are cross-reactive with homologous herpes simplex virus 1 (HSV1) and Epstein-Barr virus (EBV) proteases (Table 1), and this approach has the potential to be broadly applicable across the HHV Pr family. This provides an opportunity to further develop our current molecules into tools to study the link between HHV and other forms of disease progression, such as cancer and neurodegeneration, as well as therapeutics against additional HHV infections (Gable, Acker and Craik, 2014; Readhead et al., 2018). Other viral proteases, such as those encoded by SARS CoV-2 virus, Zika virus, West Nile virus and HIV, also contain non-catalytic cysteine residues targetable for covalent inhibition, and this approach may be expanded beyond the HHV family (Logsdon et al., 2004; Hammamy et al., 2013; Quek et al., 2020; Gao et al., 2021).

While C161 is conserved across the HHV Pr family, it is unclear what functional role this residue plays. Though not essential for activity, we found that mutation of C161 to alanine causes a two-fold decrease in k_cat/K_m, and that this effect is K_m-driven (Figure S1a). As C161 is located within the S₁’ substrate pocket, it may be important for substrate recognition. The dispensability of C161 for activity raises the concern that resistance may occur via mutation of C161. However, C161 is conserved not only within the HHVs, but across all herpesvirus proteases (Waxman and Darke, 2000). This indicates some essential function for C161, and we therefore expect mutation would be less likely than another non-conserved residue.

From initial CMK irreversible aryl triazole inhibitors, we developed an optimized tool compound (43, Table 1) and integrated X-ray crystallography, HSQC, MD simulations and biochemistry to study inhibitor binding mode and mechanism of action. Structural data indicate alternative 43 binding modes whereby the inhibitor engages the S₁’ substrate pocket and either positions itself in the S₁ substrate pocket or the dimer interface (Figure 3). MD simulations indicate 43 engages R165, which is part of the oxyanion hole, thus disruption of the HCMV Pr active site architecture may contribute to inhibition in addition to direct substrate blockage (Tong et al., 1998). X-ray crystallography and MD simulations also indicate that the orientation of 43 in the binding pocket is influenced by dimerization. When 43 is bound to chain A of the HCMV C-A homodimer, the C-5 phenyl makes a cation-π interaction with R109 in chain B (Figure 3). This prevents 43 from rotating into the S₁ substrate pocket, keeping it oriented toward the dimer interface when the homodimer is formed. Inhibitor 43 labeling stabilizes the HCMV C-A homodimer (Figure 4), leading us to conclude that, of the 43 binding poses observed in MD, the most relevant to inhibition is 43 bound to the HCMV Pr homodimer, oriented toward the dimer interface.

Labeling assays indicate that 43 reacts more efficiently with HCMV C-A in equilibrium than either the monomer or dimer in isolation (Figure 4). The orientation of R109 in the homodimer creates a 43-binding pocket that would be sterically occluded if dimer were to form prior to inhibitor binding (Figure 3). Conversely, helices α5 and α6 are expected to be partially disordered in the monomeric state, possibly explaining the slower reactivity with HCMV C-A L222D (Nomura et al., 2005). The mechanism of 43 inhibition is unique from conformational selection, highlighting how equilibrium can influence inhibition by requiring all three components of the system (both monomeric subunits and 43) to come together at the same time to yield optimal inhibition. This underscores the importance of factoring in conformational equilibrium when designing inhibitors against dynamic proteins.

Using 43 and alkyne-derivative 48, we demonstrated inhibition of HCMV viral replication using CMK inhibitors (Figure 5), presenting a mechanistically and chemically distinct starting point to develop more selective HCMV antiviral agents (Gable, Acker and Craik, 2014). While the CMK warhead served as a proof of principle for managing viral infection by irreversibly targeting C161, an immediate path forward for inhibitor optimization is to modulate the electrophilic warhead, including the use of reversible covalent warheads to mitigate off-target effects (Serafimova et al., 2012). The X-ray crystal structure indicates that rigidifying the 43 scaffold is a parallel approach to improve inhibitor potency and selectivity. Furthermore, other disulfide tethering hits containing chemical structures distinct from aryl triazoles exhibited inhibition of HCMV C-A, and could represent additional starting points for future inhibitor development (Table 1). This work serves as a launch pad to leverage the integrative pipeline we developed and make rapid progress on inhibitor optimization, improving target selectivity and cytotoxicity and further improving antiviral activity.

Limitations of the study

While in-gel fluorescence presents compelling evidence for on-target engagement of CMK inhibitors with HCMV Pr (Figure 5b-f), the off-target reactivity and observed cytotoxicity (Figure S4) make it difficult to fully decouple antiviral activity driven by HCMV Pr inactivation from antiviral activity due to general cytotoxic effects. This limitation will be improved as inhibitors are further optimized and the CMK warhead is replaced with less reactive electrophiles.

STAR Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Charles Craik (Charles.Craik@ucsf.edu).

Materials availability

• This study did not generate unique reagents

Data and code availability

• NMR assignments have been deposited to the Biological Magnetic Resonance Bank and are publicly available as of the date of publication. Accession numbers are listed in the key resources table.

• X-ray crystallography data have been deposited to the Protein Data Bank and are publicly available as of the date of publication. PDB codes are listed in the key resources table.

• Any additional data reported in this paper will be shared by the lead contact upon request.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit monoclonal anti-DYKDDDDK tag	Cell Signaling Technology	D6W5B
Bacterial and virus strains
XL-10 Ultracompetent cells	Agilent Technologies	200315
Rosettat™ 2(DE3) Singles™ Competent cells	MilliporeSigma	71400
BL21(DE3)pLysS Singles™ Competent Cells	MilliporeSigma	70236
Human herpesvirus 5	American Type Culture Collection	VR-2356
Biological samples
Chemicals, peptides, and recombinant proteins
NH2-Lys(MCA)-Tbg-Tbg-Asn-Ala-Ser-Ser-Arg-Leu-Lys(Dnp)-Arg-OH	GenScript USA Inc.	N/A
NH2-Lys(MCA)-His-Thr-Tyr-Lys-Gln-Ala-Ser-Glu-Lys-Phe-Lys-Lys(Dnp)-OH	This paper	N/A
Human Cytomegalovirus Protease	This paper	N/A
Human Cytomegalovirus Protease Δ221	This paper	N/A
Human Cytomegalovirus Protease L222D	This paper	N/A
Human Cytomegalovirus Protease C161A	This paper	N/A
Human Cytomegalovirus Protease C84A, C87A, C138A, C202A	This paper	N/A
Human Cytomegalovirus Protease C84A, C87A, C138A, C161A, C202A	This paper	N/A
Human Cytomegalovirus Protease C84A, C87A, C138A, C202A, L222D	This paper	N/A
Human Cytomegalovirus Protease C84A, C87A, C138A, C202A, S225M	This paper	N/A
Epstein-Barr virus protease	This paper	N/A
Herpes simplex virus 1 protease	This paper	N/A
1-Chloro-3-(4-phenyl-1H-1,2,3-triazol-1-yl)propan-2-one	This paper	N/A
1-(Oxiran-2-ylmethyl)-4-phenyl-1H-1,2,3-triazole	This paper	N/A
1-(4-Phenyl-1H-1,2,3-triazol-1-yl)-3-(2,3,5,6-tetrafluorophenoxy)propan-2-one	This paper	N/A
1-Chloro-3-(4-(4-nitrophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one	This paper	N/A
4-(1-(3-Chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)benzonitrile	This paper	N/A
1-(4-(4-Bromophenyl)-1H-1,2,3-triazol-1-yl)-3-chloropropan-2-one	This paper	N/A
1-Chloro-3-(4-(4-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one	This paper	N/A
1-Chloro-3-(4-(3-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one	This paper	N/A
N-(3-(1-(3-Chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)phenyl)acetamide	This paper	N/A
3-((1-(3-Chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)methyl)-1H-indene-1,2(3H)-dione	This paper	N/A
(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl methylcarbamate	This paper	N/A
(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl dimethylcarbamate	This paper	N/A
(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl benzylcarbamate	This paper	N/A
(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-methoxyethyl)carbamate	This paper	N/A
(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl morpholine-4-carboxylate	This paper	N/A
(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-morpholinoethyl)carbamate	This paper	N/A
(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (3-(2-oxopyrrolidin-1-yl)propyl)carbamate	This paper	N/A
1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl prop-2-yn-1ylcarbamate	This paper	N/A
Critical commercial assays
QuikChange Lightning Site Directed Mutagenesis Kit	Agilent Technologies	210518
The PEGs suite, condition #7	Qiagen	134307
CellTiter-Glo® Luminescent Cell Viability Assay	Promega	G7570
β-Galactosidase Enzyme Assay System with Reporter Lysis Buffer kit	Promega	E2000
Beta-Glo® Assay System kit	Promega	E4720
TransIT®−293 Transfection Reagent	Mirus Bio	MIR 2700
Deposited data
Human Cytomegalovirus protease Δ221 1H-13C and 1H-15N assignments	Biological Magnetic Resonance Bank	Accession number 51211
Human Cytomegalovirus Protease C84A, C87A, C138A, C202A co-crystal structure with inhibitor 43	Protein Data Bank	PDB ID: 7TCZ
Complex structure of Human Cytomegalovirus Protease and a peptidomimetic inhibitor	Khayat et al., 2003	PDB ID: 1NJU
Experimental models: Cell lines
Human: MRC-5 cells	American Type Culture Collection	CCL-171
Human: HFF-1 cells	American Type Culture Collection	SCRC-1041
Human: Huh7 cells	Chan Zuckerberg Biohub, Andreas Puschnik	N/A
Human: HeLa cells	Chan Zuckerberg Biohub, Andreas Puschnik	N/A
Human: HEK293T cells	American Type Culture Collection	CRL-3216
Experimental models: Organisms/strains
Oligonucleotides
Recombinant DNA
Plasmid: 6xHis Human Cytomegalovirus Protease	Lee et al., 2011	N/A
Plasmid: 6xHis Human Cytomegalovirus Protease Δ221	Lee et al., 2011	N/A
Plasmid: 6xHis Human Cytomegalovirus Protease L222D	Lee et al., 2011	N/A
Plasmid: 6xHis Human Cytomegalovirus Protease C161A	This paper	N/A
Plasmid: 6xHis Human Cytomegalovirus Protease C84A, C87A, C138A, C202A	This paper	N/A
Plasmid: 6xHis Human Cytomegalovirus Protease C84A, C87A, C138A, C161A, C202A	This paper	N/A
Plasmid: 6xHis Human Cytomegalovirus Protease C84A, C87A, C138A, C202A, L222D	This paper	N/A
Plasmid: 6xHis Human Cytomegalovirus Protease C84A, C87A, C138A, C202A, S225M	This paper	N/A
Plasmid: FLAG Human Cytomegalovirus Protease	GenScript USA Inc.	N/A
Plasmid: 6xHis Epstein-Barr virus protease	Gable et al., 2014	N/A
Plasmid: 6xHis Herpes simplex virus protease	GenScript USA Inc.	N/A
Software and algorithms
MassLynx 4.1	Waters	https://www.waters.com/waters/en_US/MassLynx-Mass-Spectrometry-Software-/nav.htm?cid=513164&locale=en_US
HiTS	UCSF Small Molecule Discovery Center	hits.ucsf.edu
BioTek Gen5 2.03	Agilent	https://www.biotek.com/products/software-robotics-software/gen5-microplate-reader-and-imager-software/
Prism 8	GraphPad	https://www.graphpad.com
REDiii	Bohn and Schiffer, 2015	DOI:10.1107/S139900471500303X
xia2	Diamond Light Source	https://xia2.github.io
Phenix.Refine 1.11.1–2575	Phenix	https://phenix-online.org
Coot 0.9.5 EL	MRC Laboratory of Molecular Biology	https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/
TopSpin 2 and 4	Bruker	https://www.bruker.com/en/products-and-solutions/mr/nmr-software/topspin.html
NMRpipe and NMRdraw 8.2	NIST IBBR	https://www.ibbr.umd.edu/nmrpipe
Sparky 1.414	NMRFAM	https://nmrfam.wisc.edu/nmrfam-sparky-distribution
Gaussian 16	Frisch et al. 2016	https://gaussian.com
Amber 16	D.A. Case et al., 2016	https://ambermd.org
ModLoop	ModBase	https://modbase.compbio.ucsf.edu/modloop/
Unicorn 6.3	Cytvia	https://www.cytivalifesciences.com/en/us/shop/unicorn-6-3-p-01118
Image Lab 5.0	BioRad	https://www.biorad.com/en-us/product/image-lab-software?ID=KRE6P5E8Z
ImageJ 1.53a	Schneider et al., 2012	https://imagej.nih.gov/ij/
ChemDraw 18.2	PerkinElmer	https://perkinelmerinformatics.com/products/research/chemdraw/
Other
Acquity LC/Xevo G2-XS QTof MS	Waters	https://www.waters.com/waters/en_US/Xevo-G2-XS-QTof-Quadrupole-Time-of-Flight-Mass-Spectrometer/nav.htm?cid=134798222&locale=en_US
Bruker Avance DRX 500 MHz	UCSF Nuclear Magnetic Resonance Laboratory	https://pharm.ucsf.edu/nmr/instruments/bruker-500
Bruker Avance AV 800 MHz	UCSF Nuclear Magnetic Resonance Laboratory	https://pharm.ucsf.edu/nmr/instruments/bruker-800
ÄKTA™ pure	Cytvia	https://www.cytivalifesciences.com/en/us/shop/chromatography/chromatography-systems/akta-pure-p-05844
ChemiDoc Imaging System	BioRad	https://www.biorad.com/en-us/product/chemidoc-imaging-system?ID=OI91XQ15

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human cell lines

HEK293T cells (CRL-3216), MRC-5 cells (CCL-171) and Human Foreskin Fibroblasts (HFF-1, SCRC-1041), were obtained from the American Type Culture Collection (Bethesda, MD). These cells were maintained in high-glucose Dulbecco’s Modified Eagle Medium (DMEM, HyClone) with 20% Fetal Bovine Serum (Gibco) and 1% penicillin/streptomycin. Human hepatocellular carcinoma line (Huh7) and human cervical adenocarcinoma cells (HeLa) were generous gifts from Andreas Puschnik’s group (Chan Zuckerberg Biohub). These cells were maintained in high-glucose Dulbecco’s Modified Eagle Medium (DMEM, HyClone) with 10% Fetal Bovine Serum (Gibco) and 1% penicillin/streptomycin. All cells were grown at 37°C and 5% CO₂. HEK293T, HFF-1 and MRC-5 cells were authenticated prior to purchase from ATCC. Huh7 and HeLa cells were not authenticated by the authors of this study prior to use. HEK293T and HeLa cells are female, Huh7 and HFF-1 cells are male (Shah, McCormack and Bradbury, 2014).

Viruses

A recombinant, Towne-strain, human cytomegalovirus bearing the lacZ gene (ATCC VR-2356) was obtained from the American Type Culture Collection (Bethesda, MD). Viral stocks were expanded and titered according to previously described methods (Britt, 2010).

E. coli strains

PCR products generated during mutagenesis were propagated in XL10 Gold Ultracompetent cells (Agilent). Protein expression was performed in Rosetta 2(DE3) competent cells (MilliporeSigma) or BL21 DE3 pLysS competent cells (MilliporeSigma). Details on specific growth conditions are included in the Methods Details within Protein Mutagenesis and Protein Expression sections.

METHODS DETAILS

Protein Mutagenesis

Point mutagenesis was performed using a QuikChange Lightning Site Directed Mutagenesis Kit (Agilent). PCR products were transformed into XL-10 Gold competent E. coli cells (Agilent) and selected using Ampicillin (AMP) and Chloramphenicol (CAM). Cultures were grown in 5 mL Luria Broth (LB) supplemented with AMP (100 μg/mL final concentration) and CAM (34 μg/mL final concentration) while shaking at 37°C overnight. DNA was isolated using a Qiagen Miniprep Kit and sequenced.

Protein Expression

HCMV Pr:

All HCMV Pr constructs bear a N-terminal His6 tag. WT HCMV Pr and all subsequent mutant constructs contain A141V, A143V, P144A and A209V mutations to stabilize against autoproteolysis. Constructs were expressed in Rosetta 2(DE3) competent E. coli cells (MilliporeSigma).

Unlabeled:

Cells were grown in 50 mL LB supplemented with AMP (100 μg/mL) and CAM (34 μg/mL) while shaking at 37°C overnight. The next day, 10–50 mL of culture was used to inoculate 1 L LB supplanted with antibiotics and shaken at 37^oC to an OD₆₀₀ of 0.6. Isopropyl β-d-1-thiogalactopyranoside (IPTG) was added (1 mM final concentration) and cultures were shaken at 16°C overnight.

Isoptopically labeled:

M9 minimal media (MM) was prepared from the following recipe (amounts per 1 L media): 3.0 g KH₂PO₄, 6.8 g Na₂HPO₄, 1.5 g NaCl, 1 mL vitamin solution, 2 mL trace metal solution, 10 mL 20% glucose, 1 mL 1 M MgSO₄ and 30 μL 1 M CaCl₂ and 1 g NH₄Cl. For U-¹⁵N labeling, ammonium chloride was substituted with 1 g ¹⁵NH₄Cl (Cambridge Isotopes). For selective ¹³C-ILVM labeling, 100 mg [¹³C₂-methyl] α-ketoisovalerate, 50 mg [¹³C1-methyl] α-ketobutyrate and 250 mg ¹³C-methyl Methionine (Cambridge Isotopes) was added ~1 h prior to IPTG induction. Cells were grown in 5 mL LB + antibiotics while shaking at 37°C overnight. The 5 mL culture was used to inoculate 50 mL M9 MM + antibiotics and shaken at 37°C overnight. The 50 mL culture was added to 1 L M9 MM and shaken at 37^oC to an OD₆₀₀ of 0.6 then induced with IPTG (1 mM final concentration) and shaken at 16°C overnight.

EBV Pr:

An N-terminal His6-tagged construct (Uniprot A0A0C7TBJ5, AA 1–235) was transformed into Rosetta 2(DE3) competent E. coli cells (MilliporeSigma). Cells were grown and induced using the same procedure described for unlabeled HCMV Pr.

HSV1 Pr:

A pET15b vector (Genescript, Inc) containing the protease (Uniprot P10210, AA 1–247) with an N-terminal His6-tag was transformed into BL21 DE3 pLysS competent E. coli cells (MilliporeSigma). Bacteria were grown in 50 mL LB media containing AMP (100 ug/mL) for 16 h shaking at 30 °C, and subsequently inoculated 1:100 into 1 L LB media containing AMP (100 ug/mL). When the culture reached A₆₀₀= 0.6 −0.7, protein expression was induced overnight at 16 °C with 1 mM IPTG. Cells were harvested at 4000 xg for 20 minutes and either used immediately or stored frozen at −80 °C.

Protein Purification

HCMV Pr:

All purification steps were performed at 4 °C or on ice. Cells were harvested, pelleted and suspended in a buffer containing 50 mM potassium phosphate pH 8.0, 300 mM KCl and 25 mM imidazole. Buffer also included 5 mM β-mercaptoethanol (BME) for WT HCMV Pr, HCMV C161A and HCMV Δ221. For HCMV C-A, HCMV C-A L222D and HCMV C-A S225M, protein was prepared without reducing agent. Cells were lysed by microfluidization and pelleted, and the supernatant was purified on a Cytvia Äkta Explorer FPLC. Protein was eluted over two stacked 5 mL HisTrap Nickel columns with gradient elution into a buffer containing 25 mM potassium phosphate pH 8.0, 150 mM KCl and 300 mM imidazole (± 5 mM BME). Eluate was collected and dialyzed overnight against a buffer containing 25 mM potassium phosphate pH 8.0, 150 mM KCl and 0.1 mM EDTA (± 1 mM BME). Dialyzed protein was concentrated to ~2 mL and purified over a HiLoad 26/60 Superdex 75 (Cytvia) into the same buffer. Protein bands were analyzed by SDS-PAGE buffer and pure protein was collected, flash frozen and stored at −80 °C.

EBV Pr:

Protein was purified as previously described with the following alterations (Buisson et al., 2001). All purification steps were performed at 4 °C or on ice. Cells were harvested, pelleted and suspended in buffer containing 20 mM Tris-HCl pH 7.9, 10 mM imidazole and 0.5 M NaCl then lysed by microfluidization and pelleted. The lysate supernatant was incubated with ~5 mL Ni-NTA agarose resin (Qiagen) for three hours. Beads were washed with five column volumes of lysis buffer containing 50 mM imidazole then eluted with three column volumes of lysis buffer containing 500 mM imidazole. Eluate was dialyzed overnight into buffer containing 20 mM Tris-HCl pH 7.5, 100 mM Na Cl, 1 mM EDTA and 10 mM BME. The next day, ~2 mL of protein was purified over a HiLoad 26/60 Superdex 75 (Cytvia) into the same buffer. SDS-PAGE was used to select pure protein bands and purified EBV Pr was collected, concentrated, flash frozen and stored at −80 °C.

HSV1 Pr:

All purification steps were performed at 4 °C or on ice. Cells were resuspended in Buffer A (50 mM Tris pH 8.0, 500 mM NaCl), supplemented with Pierce™ Protease Inhibitor Tablets, EDTA-free (ThermoScientific), and subsequently lysed by two rounds of sonication (2 s on/4 s off at 50 W output). Lysate was centrifuged at 14,000 xg for 40 min, and then soluble cell extract was bound to Ni-NTA agarose resin (Qiagen) and washed with Buffer A plus 20 mM imidazole for 20 column volumes. The bound protein was eluted with Buffer A plus 200 mM imidazole over 8 column volumes and subsequently dialysed against 4 L Buffer B (50 mM HEPES pH 8.0, 50 mM KCl, 1 mM EDTA, 5 mM TCEP) overnight. The dialyzed solution was then concentrated to 4 mL (using an Amicon™ Ultra-15 Centrifugal Filter unit, 10 kDa MWCO spin concentrator) and subjected to size-exclusion chromatography using a HiLoad 26/600 Superdex 75 prep grade gel filtration column (Cytvia) in Buffer B. Fractions were analyzed by SDS-PAGE and those containing pure protein were collected, concentrated to >2 mg/mL in Buffer B plus 10% glycerol, and flash-frozen for storage at −80 °C.

Intact protein LC/MS

Intact protein liquid chromatography/mass spectrometry (LC/MS) was performed on a Waters Acquity LC/Xevo G2-XS QTof MS and data were processed using MassLynx as previously described (Hallenbeck et al., 2018). The predicted masses for His6-tagged HCMV Pr constructs were calculated using ExPASy ProtParam:

• WT HCMV Pr: 29,051 Da

• HCMV C161A: 29,019 Da

• HCMV C-A: 28,922 Da

• HCMV C-A L222D: 28,924 Da

• HCMV C-A S225M: 28,966 Da

• HCMV C-A C161A: 28,890 Da

The normalized intensity of the apo protein peak and the protein + compound adduct peak (I_adduct) were used to calculate percent labeling:

$$\%Labeling=\left(\frac{I_{adduct}}{I_{adduct}+I_{apo}}\right)\times 100$$

Disulfide Tethering Screen

Screening was performed in 384 well plates with each well containing 25 μL HCMV C-A at 200 nM in 25 mM ammonium acetate buffer at pH 8.0 containing 50 μM BME. A library of disulfide capped fragment molecules housed in the UCSF Small Molecule Discovery center (https://pharm.ucsf.edu/smdc), stored at 50 mM in DMSO was used for endpoint assay screening. A Biomek EL406 plate dispenser was used to deliver 50 nL compounds to each protein well, for a final concentration of 100 μM compound and 0.1% DMSO. Plates were incubated for 1 h at room temperature before injection onto LC/MS. Adduct formation was analyzed and compiled onto the HiTS webserver as previously described (hits.ucsf.edu) (Hallenbeck et al., 2018). Hits were characterized by percent labeling greater than 52.36% (3-sigma from the average percent labeling).

Dose-dependent inhibitor labeling

Assays were performed in 384 well plates with each well containing 20–25 μL WT HCMV Pr, HCMV C-A or HCMV C161A at 350 nM in a 100 mM ammonium acetate buffer at pH 8.0. Compound stocks were diluted from 50 mM in DMSO to generate a 12-point dilution series with the last point in the series being a DMSO-only control. 1 μL compound was delivered to each well (highest final concentration 500 μM). Protein was incubated with compound at room temperature for 1 h then analyzed by LC/MS.

Time-course inhibitor labeling

All assays were run using 100 mM ammonium acetate buffer at pH 8.0. Final labeling curves were fit to one-phase decay curve in GraphPad Prism.

Quenched:

Protein was diluted to 4 μM in ammonium acetate buffer. Compound 43 was combined with protein to a final concentration of 20 μM and incubated at room temperature. At varying time points (1–180 min) aliquots were removed and diluted 1:2 into ammonium acetate buffer containing 2 mM dithiothreitol (DTT). When the assay was complete, quenched sample was diluted 1:4 into ammonium acetate (500 nM protein and 250 μM DTT final) and analyzed by LC/MS. Data were collected in triplicate from technical replicates and are reported in figures as the mean, including error bars depicting standard deviation.

Unquenched:

Protein was diluted to 500 nM in ammonium acetate buffer. Compound 43 was combined with protein to a final concentration of 10 μM and incubated at room temperature. At varying time points 25 μL was removed and immediately analyzed by LC/MS. Data were collected as single replicates.

Michaelis-Menten Kinetics

WT HCMV Pr, HCMV C-A or HCMV C161A was diluted to 500 nM in buffer containing 25 mM potassium phosphate at pH 8.0, 150 mM KCl and 0.1 mM EDTA, 10% glycerol, 0.01% TWEEN-20 and 1 mM BME. An internally quenched fluorescent substrate was used: NH2-Lys(MCA)-Tbg-Tbg-Asn-Ala-Ser-Ser-Arg-Leu-Lys(Dnp)-Arg-OH, where Tbg is L-tert-leucine, Lys(MCA) is a lysine residue with side chain linked to a 7-methoxycoumarin-4-acetic acid and Lys(Dnp) is a lysine linked to 2,4-dinitrophenyl. Substrate was diluted in DMSO in serial dilution (150–1 μM final in assay). Protein and substrate were combined in a 96-well black untreated polystyrene plate and enzyme velocity was monitored by fluorescence increase (excitation: 328, emission: 393) on a BioTek Synergy H4 Bioreader at 30°C. Mean velocity (RFU/s) during steady state was fit using BioTek Gen5 data analysis software. The mean velocity of control wells containing only buffer and substrate was averaged and subtracted from calculated enzyme velocity, then subtracted values were used for further processing. To convert velocity values from RFU/s to μM/s, a serial dilution of free MCA was prepared, ex/em at 30°C was determined and fluorescence (RFU) was plot in triplicate against MCA concentration (μΜ) then fit to a linear regression to get a slope of 11053 RFU/μM. This value was used to convert initial velocities to μM/s. Substrate concentration vs. velocity curves were plotted in triplicate from technical replicates and fit in GraphPad Prism using the standard k_cat equation and resulting values were used to calculate k_cat/K_m with standard propagation of error.

IC₅₀ value determination

HCMV Pr:

WT HCMV Pr or HCMV C-A was diluted to 500 nM in buffer containing 25 mM potassium phosphate pH 8.0, 150 mM KCl and 0.1 mM EDTA, 10% glycerol, 0.01% TWEEN-20 and 50 μΜ BME (note, assays with aryl triazole disulfide tethered molecules did not contain TWEEN-20). Disulfide tethered or CMK inhibitors were diluted in DMSO to produce 2-fold serial dilutions (1 mM or 250 μM top concentration, final in assay) with the last point a DMSO-only control. Diluted inhibitors were added to a final DMSO concentration of 2% and protein + inhibitor was incubated at room temperature for 1 h. The same fluorescent substrate was used as described in Michaelis-Menten kinetic assays. The final concentration of substrate in the assay was 15 μM (disulfide tethered molecules) or 20 μM (CMK electrophiles). Data was collected and processed as described for Michaelis-Menten kinetic assays. The velocity (V) of DMSO controls was averaged and used to normalize data as percent activity:

$$\%\text{Activity}=\left(\frac{{\text{V}}_{\text{compound}}}{{\text{V}}_{\text{DMSO}}}\right)\times 100$$

The log₁₀[Inhibitor] was plotted vs. percent activity and curves were fit in GraphPad Prism with a standard four parameter log(inhibitor) vs. response with variable hill slope equation. All data were collected in duplicate or triplicate from technical replicates and are reported in figures as the mean, including error bars depicting standard deviation.

EBV Pr:

The setup for EBV Pr inhibition assays using CMK inhibitors was identical to HCMV Pr except the final enzyme concentration in the assay was 125 nM. The same FRET substrate was used as for HCMV Pr, and the final concentration in the assay was 20 μM (K_m is 13 μM for EBV Pr).

HSV1 Pr:

The setup for HSV1 Pr inhibition assays using CMK inhibitors was similar to HCMV Pr, except for the following. The assay buffer consisted of 50 mM potassium phosphate pH 8.0, 50 mM KCl, 1 mM EDTA, 1 mΜ BME, 10% glycerol, 0.01% Tween-20 and 500 mM citrate. Final assay enzyme concentration was 5 μM. Substrate used was NH₂-Lys(MCA)-His-Thr-Tyr-Lys-Gln-Ala-Ser-Glu-Lys-Phe-Lys-Lys(Dnp)-OH. The final concentration of substrate in the assay was 50 μΜ (K_m of this substrate is 58 μM for HSV1 Pr).

K_I and k_inact determination

HCMV C-A was diluted to 500 nM in buffer containing 25 mM potassium phosphate pH 8.0, 150 mM KCl and 0.1 mM EDTA, 10% glycerol and 0.01% TWEEN-20. CMK electrophiles were diluted as described for IC₅₀ determination, then protein and inhibitors were incubated at room temperature for varying times (1–240 min). Assay setup and data collection were as in the sections above that describe the Michaelis-Menten kinetics and IC₅₀ determination assays. The final concentration of substrate in the assay was 20 μM.

All data were collected in duplicate or triplicate from technical replicates. Time vs. % Activity for each compound concentration was fit in GraphPad Prism to determine k_obs. Irreversible inhibitors exhibit time-dependent inhibition kinetics. Under such conditions, steady-state enzyme analysis over varying pre-incubation times of HCMV Pr with a set concentration of inhibitor can be performed, and the determined enzyme reaction progress curve is fit to the following equation, where V_t is enzyme velocity at inhibitor pre-incubation time t, V₀ is enzyme velocity at pre-incubation time zero, and k_obs is the pseudo-first-order rate constant:

$${\text{V}}_{\text{t}}={\text{V}}_0\left({\text{e}}^{-k_{obs}\times \text{t}}\right)$$

The above method is applied for varying inhibitor concentrations ([I]), to determine k_obs for each concentration of inhibitor tested. Inhibition of HCMV Pr with irreversible CMK inhibitors exhibits a two-step mechanism of inactivation, such that a plot of k_obs vs. [I] is saturating at high [I]. Under these conditions, the plot of k_obs vs. [I] is fit to the following equation to determine K_I and k_inact :

$$k_{obs}=\frac{k_{inact}\times \left[\text{I}\right]}{{\text{K}}_{\text{I}}+\left[\text{I}\right]}$$

This analysis was performed to determine K_I and k_inact of each inhibitor tested. The ratio of these values was taken to determine k_inact/K_I and t_1/2 was determined by dividing 0.693 by k_inact.

X-ray crystallography

HCMV C-A (purified in the absence of reducing agent) was diluted to 500 nM in buffer containing 25 mM Tris pH 8.0 and 100 mM KCl then combined with 51 μΜ 43 (0.1% DMSO) and stirred gently at 4 °C overnight. The next day, the protein sample was concentrated to a final concentration of 6.2 mg/mL. All crystallization plates were set up as 96-well hanging drop screens with wells containing 80 μL crystallization buffer. Using a TTP LabTech Mosquito Nanoliter Dropsetter, 0.1–0.2 μL protein and 0.1–0.2 μL crystallization buffer was combined (1:1 ratio) and plates were sealed. Plates were incubated at room temperature and crystal formation was monitored using a Formulatrix Rock Imager. Crystals formed within 7 days as rectangular prisms with two short axes and one long axis extending to ~200–500 μm. The final 43/HCMV C-A co-crystal structure was collected using a crystal formed from condition #7 of Qiagen PEGS suite with the following components: 0.1 M MES pH 6.5 with 40% v/v PEG 200. Crystals were flash frozen in liquid nitrogen and dipped in Al’s Oil cryoprotectant prior to data collection. Crystals were diffracted at Beamline 8.3.1 at the Advanced Light Source. Initial data processing and refinement was done using REdiii with HCMV Pr as search model (PDB code:1NJT) (Bohn and Schiffer, 2015). Reflections were processed using xia2 in space group P2₁2₁2₁ (Winter, Lobley and Prince, 2013). After initial inspection, final reprocessing was done in P4₁22 with a single protease chain per asymmetric unit. After iterative rounds of model building with Coot and refinement using Phenix.Refine inhibitors were fitted to density surrounding C161 using models and restraints files generated using Phenix.Elbow (Afonine et al., 2012). All final figures were produced using PyMol 2.3.2 (Schödinger LLC).

NMR Spectroscopy

The standard buffer for NMR studies contained 25 mM potassium phosphate pH 7.0, 150 mM KCl and 0.1 mM EDTA. All NMR spectra were acquired at 300 K on either a Bruker Avance DRX 500 MHz spectrometer equipped with a QCI CryoProbe and a 60-slot B-ACS sample changer or a Bruker Avance AV 800 MHz spectrometer equipped with a TXI CryoProbe and a 16-slot SE Lite sample changer. Spectral processing analysis was performed with NMRpipe and NMRFAM-Sparky (Lee, Tonelli and Markley, 2015).

¹H-¹⁵N amide backbone assignments:

NMR samples used for the triple-resonance experiments consisted of ~ 230 μM uniformly ¹³C/¹⁵N/²H-labeled HCMV Δ221 in standard NMR buffer supplemented with 5 mM DTT. Backbone amide and side chain carbon resonances were acquired using a suite of three-dimensional triple-resonance experiments as previously described (Lee et al., 2012). Selectively ¹⁵N-labeled (¹⁵N-Ala, ¹⁵N-Gly, ¹⁵N-Leu, ¹⁵N-Val, or ¹⁵N-Tyr) HCMV Δ221, as well as ¹⁵N/¹³C-methyl ILVM labeled HCMV Δ221 Leu/Val-to-Ile constructs were also used for resonance assignments. A three-dimensional ¹⁵N-¹H NOESY-HSQC spectrum acquired on a uniformly ¹⁵N/¹H-labeled HCMV Δ221 sample was used to determine the tryptophan indole HN and aliphatic methyl group assignments.

Inhibitor binding assays:

For disulfide tethered hits, U-¹⁵N ¹³C-ILVM labeled HCMV C-A L222D was diluted to 50 μM in standard NMR buffer supplemented with 50 μM BME, combined with d₆-DMSO or 500 μM inhibitor (15, 18, 23 or 24) and incubated at 4 °C overnight. The next day, deuterium oxide was added to 10% v/v. Final samples were prepared and mounted on the auto-sampler for overnight spectra collection. For 43-bound protein samples, U-¹⁵N ¹³C-ILVM labeled HCMV C-A L222D was diluted to 1 μM and stirred overnight at 4 °C in the presence of d6-DMSO or 43 (25 μM final, 2% d₆-DMSO). The next day, samples were concentrated and prepared to 42 μM in NMR assay buffer and deuterium oxide was added to 10% v/v. Peak lists were manually analyzed and centered onto peaks in Sparky. Signal to noise ratio (S/N) was determined in Sparky and peaks were inspected in DMSO control spectra to verify S/N was greater than three. Peaks with sufficient S/N and little spectral crowding were included in the final analysis. Chemical shift perturbations (CSPs) were determined as previously described, with a significance cutoff of 0.05 ppm (1-sigma from the average CSP) (Lee et al., 2011). Peak volume was integrated in Sparky and used in analysis as previously described (Sanulli et al., 2019). To account for global changes in peak intensity and to allow for comparison between spectra, all peaks within one spectrum were normalized to a single peak value (W179). Then, the ratio of protein + inhibitor (V_Inhibitor) and protein + DMSO (V_DMSO) was taken. Peaks that disappear upon inhibitor binding were taken to have a V_I of zero.

Molecular Dynamics Simulations

Parameterization of inhibitor 43-tethered cysteine:

Since inhibitor 43 forms cross-linkage to C161, the entire molecule is constructed as a non-standard residue. The N-terminal and C-terminal of Cys was capped with Acetyl (ACE) and N-methyl (NME) groups, and inhibitor 43 bonded to the Cys through a disulfide bond. The geometry optimization and electrostatic potential were performed using the Hartree-Fock theory and 6–31G(d) basis set on Gaussian 16 (Frisch et al., 2016), and antechamber was used to derive RESP charges for the molecule in the amber forcefield (Bayly et al., 1993; Wang et al., 2006).

Molecular Dynamics Simulation:

MD simulations and analyses were carried out using Amber 16 (D.A. Case et al., 2016). The initial structures (monomer and dimer) were obtained from the 43/HCMV C-A co-crystal structure, and the missing residues were constructed from the 1NJU crystal structure and modeled using the ModLoop server (Fiser, Do and Sali, 2000; Fiser and Sali, 2003; Khayat et al., 2003). The general Amber force field (GAFF) (Wang et al., 2004) and parameters generated from Gaussian 16 and Amber 16 were used for inhibitor-linked C161, and the Amber ff14SB force field was used for the protein in MD simulations (Maier et al., 2015). Each structure was solvated in a truncated octahedron box of TIP3P water molecules with a minimum distance of 10 Å between protein and box boundary, and potassium and chloride ions were added to keep neutral and match experimental conditions. The simulations started with energy minimization for 30,000 steps to remove bad contacts followed by heating of the system smoothly from 0 to 300 K. The equilibration of the systems was carried out with gradual release of harmonic restraints on heavy atoms for 8 ns in the NPT ensemble (1 atm, 300 K). The final production run was performed for 150 ns, and the MD trajectories were analyzed using the CPPTRAJ module of Amber 16 (Roe and Cheatham, 2013), and cluster analysis was performed using DBscan clustering method to characterize the representative binding modes of inhibitor 43 (Ester, M. et al., 1996).

Size-Exclusion Chromatography

A Superdex 75 10/300 GL column was equilibrated at 4 °C into a standard SEC buffer containing 25 mM potassium phosphate pH 8.0, 150 mM KCl, 0.1 mM EDTA and 10% glycerol.

Protein dimerization controls for NMR:

HCMV C-A or HCMV C-A L222D was diluted into SEC buffer to varying concentrations (0.4–141 μM) and incubated at room temperature for 1–7 h.

Construct comparison:

HCMV C-A, HCMV C-A previously labeled with 43 for X-ray crystallography, HCMV C-A L222D or HCMV C-A S225M were diluted to 0.9 μΜ and incubated at room temperature for 1–2 h.

Comparison of HCMV CA C161 +DMSO or +43:

HCMV C-A was diluted to 1 μM in SEC buffer and combined with DMSO or 50 μM 43 (2% DMSO) and stirred overnight at 4°C. Samples were concentrated then diluted to varying concentrations (0.4–35 μM) and incubated at room temperature for 1–7 h. For all protein samples, 500 μL was injected onto the column and protein eluted over 1 CV at 0.6 mL/min flow rate while monitoring absorbance at 280 nm (A280). For each individual sample run, A280 was plotted against time (min). The minimum absorbance value across the full spectrum was calculated in Microsoft Excel and subtracted from each absorbance value to correct the baseline to zero. The maximum absorbance between 14–23 min was calculated, then all data were normalized to this value using the following equation, where Ax is the A280 at time x and Amax is the maximum A280 calculated between 14–23 min:

$$NormalizedIntensity=\left(\frac{A_x}{A_{max}}\right)\times 100$$

Cell Toxicity

50 μL of compound 43 of varying concentrations (25–0.038 μM) in DMEM+5% FBS+PenStrep+1% DMSO was plated into 96 well plates. 50 μL of cells in DMEM+5% FBS+PenStrep at varying cell densities of the previously described cell lines was plated onto the compound plates (0.5% DMSO final). Cells were incubated at 37 °C and 5% CO2 for two days. Cell viability was measured using the standard CellTiter-Glo® Luminescent Cell Viability Assay (Promega). Data were plotted in triplicate and are reported in figures as the mean, including error bars depicting standard deviation. Data were analyzed using Graphpad Prism to determine CC₅₀ values.

Antiviral Activity

100 μL of compounds (43 or ganciclovir) of varying concentrations (25–0.038 μM) in DMEM+5% FBS+PenStrep+1% DMSO were plated into 96 well plates. HFF-1 cells were prepared and diluted to 350,000 cells/mL and the lacZ-bearing Towne virus was added at a MOI of 0.05 to the suspended cells. 100 μL of infected cells were plated into the compound plate. Cells were incubated at 37 °C and 5% CO₂ for three days. Viral replication was measured indirectly using a β-Galactosidase Enzyme Assay System with Reporter Lysis Buffer (Promega) kit. Data were plotted using five replicates, which were a combination of biological and technical replicates, and are reported in figures as the mean ± SD. Data were analyzed using Graphpad Prism to determine EC₅₀ values.

Reinfection Assays

100 μL of compounds (43, 48 or ganciclovir) of varying concentrations (25–0.038 μM final concentration) in DMEM+5% FBS+PenStrep+.5% DMSO were plated into 96 well plates. HFF-1 cells were prepared and diluted to 350,000 cells/mL and the lacZ-bearing Towne virus was added at a MOI of 0.05 to the suspended cells. 100 μL of infected cells were plated into the compound plate. Cells were incubated at 37°C and 5% CO₂ for three days. On day 3 post-infection, 80 μL of supernatant was removed from infected cells, and diluted with 150 μL of fresh DMEM+5% FBS+PenStrep. 100 μL of uninfected HFF-1 cells at 350,000 cells/mL were mixed with 100 μL of diluted virus in duplicate. No additional compound treatment occurred during reinfection. Remaining supernatant and cells from the initial infection were used to measure viral replication using a Beta-Glo® Assay System (Promega) kit and cytotoxicity measured using CellTiter-Glo® Luminescent Cell Viability Assay kit (Promega). Three days post second infection (six days after the initial infection), viral replication was assessed using a Beta-Glo® Assay System (Promega) kit and cytotoxicity measured using CellTiter-Glo® Luminescent Cell Viability Assay kit (Promega) to ensure antiviral activity was not due to residual compound toxicity. Data were plotted using three biological replicates, and are reported in figures as the mean, including error bars depicting standard deviation. Data were analyzed using Graphpad Prism to determine EC₅₀ values.

In-gel fluorescence competition assays

HFF-1 cell lysate (0.5 mg/mL total protein) ± purified HCMV C-A (0.025 mg/mL) was incubated with 43 (50 – 1.6 μM, 2-fold dilutions) at 30 °C for 60 min. Alkyne maleimide probe (Lumiprobe, Cat # 51780) was added to 10 μM final in solution and incubated at 30 °C for 30 min. Then, TAMRA-azide (5-isomer, Lumiprobe Cat # C7130; 60 μM final in solution), L-ascorbic acid (5 mM final), Tris(3-hydroxypropyltriazolylmethyl)amine (THPTA, 2 mM final) and Copper(II) sulfate (4 mM final) were added and incubated in the dark at room temperature for 1 h. Reactions were quenched into 2X SDS-PAGE running buffer supplemented with 1 M BME, boiled then run on SDS-PAGE. Fluorescence was visualized on a BioRad Chemidoc gel-imaging system using Cy3 and Cy5 detection channels. After fluorescence visualization, gels were stained with InstantBlue coomassie protein stain and imaged via Chemidoc. Fluorescence intensity was quantified in ImageJ then normalized such that maximum and minimum values were between 100 and 0 (Schneider, Rasband and Eliceiri, 2012). Data were plot against 43 concentration in Graphpad Prism and fit to a [Agonist] vs. response four parameter variable slope equation to determine EC₅₀.

In-gel fluorescence direct detection assays

HFF-1 cell lysates:

Experiments were set up using inhibitor 48 as described above for competition assays, except the incubation step with maleimide probe was excluded.

HEK293T whole cells:

HEK293T cells were plated in 10 cm plates, 4 million cells/plate in high-glucose Dulbecco’s Modified Eagle Medium (DMEM, HyClone) with 2% Fetal Bovine Serum (Gibco) and 1% penicillin/streptomycin then incubated at 37°C and 5% CO2 overnight. Cells were transfected with a pcDNA3.1 (+) plasmid containing N-terminally FLAG-tagged HCMV Pr (15000 ng/plate) using TransIT®−293 Transfection Reagent (Mirus Bio) and cells were incubated overnight. The next day, cells were treated for two hours with 1, 5, or 10 μM 48 or DMSO, then trypsinized washed with PBS and pelleted. Cell pellets were lysed, treated with TAMRA-azide and analyzed as described above for in-gel fluorescence competition assays. Samples were used to run two identical gels simultaneously. Gel #1 was imaged for Cy3/Cy5 fluorescence then carried forward for anti-FLAG western blot (Cell Signaling Technology D6W5B). Gel #2 was stained with InstantBlue Coomassie protein stain. Fluorescence intensity of HCMV Pr or control proteins were quantified in ImageJ then normalized to the band corresponding to HCMV Pr + 10 μM 48.

Inhibitor Synthesis

All inhibitors were diluted to 50 mM in d₆-DMSO, aliquoted and stored at −20^oC. Synthesis and product analysis of all inhibitors are as follows:

General Methods:

Unless otherwise noted, reactions were conducted using commercially available reagents and solvents, which were used as received. ¹H and ¹³C NMR spectra were recorded on Bruker Avance III HD 400 spectrometers. Chemical shifts are reported in δ units (ppm). NMR spectra were referenced relative to residual NMR solvent peaks. Coupling constants (J) are reported in hertz (Hz). Solvent removal was accomplished with a rotary evaporator at ca. 10–50 Torr. Column chromatography was carried out using a Biotage Isolera flash chromatography system and silica gel cartridges from Silicycle. Mass analyses and compound purity were determined using Waters Micromass ZQ, equipped with Waters 2795 Separation Module, Waters 2996 Photodiode Array Detector, and Waters 2424 ELS detector Separations were carried out with an XTerra® MS C18, 5µm, 4.6 × 50 mm column, at ambient temperature (unregulated) using a mobile phase of water-methanol containing a constant 0.1 % formic acid.

Method A-Triazole Synthesis:

Epichlorhydrin (1.2 equiv) was added to a solution of sodium azide (1.2 equiv) in water. After stirring at room temperature for 5 min, copper(II) sulfate pentahydrate (0.1 equiv) and sodium ascorbate (0.2 equiv) were added to the reaction mixture, followed by addition of the alkyne (1 equiv) in THF. After stirring at room temperature for 18 h, the reaction mixture was diluted with ethyl acetate, filtered and separated the layers of the filtrate. The organic phase was washed with brine, dried over magnesium sulfate and purified by flash column chromatography.

Method B-Triazole Synthesis:

Epichlorhydrin (1.2 equiv) was added to a solution of sodium azide (1.2 equiv) in water/THF. After stirring at room temperature for 5 min, 0.3 M aqueous solution of copper(II) sulfate pentahydrate (0.1 equiv) and 1M aqueous solution of sodium ascorbate (1 equiv) were added to the reaction mixture, followed by addition of the alkyne (1 equiv). After stirring at room temperature for 18 h, the reaction mixture was diluted with ethyl acetate and the organic phase was washed with brine, dried over magnesium sulfate and purified by flash column chromatography.

Method C-Ketone Synthesis:

The alcohol (1 equiv.) and Dess Martin periodinone (2 equiv.) were taken in dichloromethane. After stirring at room temperature for 18 h, the reaction mixture was washed with saturated aqueous sodium thiosulfate, saturated aqueous sodium bicarbonate and brine. The organic phase was dried over magnesium sulfate, concentrated and purified by flash column chromatography.

Method D-Carbamate Synthesis:

The p-nitrophenylcarbonate (1 equiv) and the amine (2 equiv) were stirred in THF at room temperature. The reaction mixture was diluted with ethyl acetate and the organic layer was washed with saturated aqueous sodium bicarbonate solution, water and brine, dried over magnesium sulfate and purified by flash column chromatography.

1-(Oxiran-2-ylmethyl)-4-phenyl-1H-1,2,3-triazole (32)

Phenylacetylene (0.3 g, 3.0 mmol), epichlorhydrin (0.46 mL, 6.0 mmol), sodium azide (0.76 g, 12.0 mmol), copper(II) sulfate pentahydrate (0.073 g, 0.3 mmol) and sodium ascorbate (0.115 g, 0.6 mmol) were taken in water (5.0 mL). A precipitate was formed immediately, and the mixture was stirred at room temperature for 2 h. The sticky precipitate was filtered, washed with water, dried and purified by flash column chromatography (35% ethyl acetate/hexanes) to isolate 1-chloro-3-(4-phenyl-1H-1,2,3-triazol-1-yl)propan-2-ol (15 mg) and 1-(oxiran-2-ylmethyl)-4-phenyl-1H-1,2,3-triazole (32) (36 mg).

1-chloro-3-(4-phenyl-1H-1,2,3-triazol-1-yl)propan-2-ol :¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.89 (s, 1H), 7.73–7.86 (m, 2H), 7.34–7.48 (m, 3H), 4.63–4.75 (m, 1H), 4.47–4.61 (m, 1H), 4.41 (dt, J = 6.2, 3.0 Hz, 1H), 3.45–3.64 (m, 2H)

LC-MS: m/z = 237.99 [M+H]⁺.

1-(oxiran-2-ylmethyl)-4-phenyl-1H-1,2,3-triazole (32) ¹H NMR (400 MHz, DMSO-d6, 25ºC): δ = 8.56 (s, 1H), 7.87 (d, J = 7.3 Hz, 2H), 7.46 (t, J = 7.7 Hz, 3H), 4.52–4.71 (m, 2H), 4.33–4.52 (m, 3H)

LC-MS: m/z = 201.95 [M+H]⁺.

1-Chloro-3-(4-phenyl-1H-1,2,3-triazol-1-yl)propan-2-one (31)

1-Chloro-3-(4-phenyl-1H-1,2,3-triazol-1-yl)propan-2-ol (15 mg, 0.063 mmol) and Dess Martin periodinone (0.064 g, 0.15 mmol) were stirred in dichloromethane (1.5 mL) for 24 h. The reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–10% acetone in dichloromethane) to obtain 6.5 mg (44%) of 1-chloro-3-(4-phenyl-1H-1,2,3-triazol-1-yl)propan-2-one (31) as a white solid. ¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.86–7.89 (m, 3H), 7.38–7.48 (m, 3H), 5.57 (s, 2H), 4.26 (s, 2H)

LC-MS: m/z = 235.95 [M+H]⁺, 268.04 [M+MeOH]⁺

1-(4-Phenyl-1H-1,2,3-triazol-1-yl)-3-(2,3,5,6-tetrafluorophenoxy)propan-2-one (33)

2,3,5,6-Tetraflouro phenol (0.52 g, 3.1 mmol), allyl bromide (0.27 mL, 3.1 mmol) and potassium carbonate (0.86 g, 6.2 mmol) were taken in DMF (5 mL). After stirring at room temperature for 2 h, the reaction mixture was diluted with ethyl acetate, washed with water and brine. The organic layer was dried over magnesium sulfate and concentrated to obtain about 0.6 g of crude 3-(allyloxy)-1,2,4,5-tetrafluorobenzene which was used without further purification.

3-Chloroperoxybenzoic acid (0.55 g 3.2 mmol) was added to a cooled (0 °C) solution of 3-(allyloxy)-1,2,4,5-tetrafluorobenzene (0.6 g, 3.0 mmol) in DCM (15 mL). After stirring at room temperature for 18 h, the reaction mixture was washed with saturated aqueous sodium thiosulfate, saturated aqueous sodium bicarbonate and brine. The organic layer was dried over magnesium sulfate, concentrated and isolated by flash column chromatography (0–25% ethyl acetate/hexanes) to obtain 95 mg (14mg) of 2-((2,3,5,6-tetrafluorophenoxy)methyl)oxirane as a white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 6.76–6.88 (m, 1H), 4.50 (dd, J = 11.6, 2.8 Hz, 1H), 4.15 (br dd, J = 11.3, 5.8 Hz, 1H), 3.34–3.41 (m, 1H), 2.84–2.94 (m, 1H), 2.68–2.78 (m, 1H)

2-((2,3,5,6-Tetrafluorophenoxy)methyl)oxirane (0.095 g, 0.43 mmol), sodium azide (0.28 g, 4.3 mmol) and ammonium chloride (0.042 g, 0.86 mmol) were taken in 8:1 mixture of methanol/water (4.5 mL). After stirring at room temperature for 72 h, the reaction mixture was diluted with ethyl acetate, washed with brine. The organic layer was dried over magnesium sulfate and concentrated to obtain about 108 mg of the crude 1-azido-3-(2,3,5,6-tetrafluorophenoxy)propan-2-ol which was used without further purification.

1-Azido-3-(2,3,5,6-tetrafluorophenoxy)propan-2-ol (0.05g, 0.19 mmol), phenylacetylene (0.021 mL, 0.19 mmol), copper(II) sulfate pentahydrate (5 mg, 0.019 mmol) and sodium ascorbate (0.037 g, 0.19 mmol) were taken in 1:1 mixture of water/methanol (2 mL). After stirring at room temperature for 18 h, the reaction mixture was diluted with ethyl acetate, washed with water and brine. The organic layer was dried over magnesium sulfate, concentrated and purified by flash column chromatography (35% ethyl acetate/hexanes) to obtain 52 mg (74%) of 1-(4-phenyl-1H-1,2,3-triazol-1-yl)-3-(2,3,5,6-tetrafluorophenoxy)propan-2-ol as a white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.87 (s, 1H), 7.70 (d, J = 7.1 Hz, 2H), 7.27–7.45 (m, 3H), 6.79–6.91 (m, 1H), 4.69–4.80 (m, 1H), 4.52–4.63 (m, 2H), 4.13–4.39 (m, 3H)

LC-MS: m/z = 368.10 [M+H]⁺

1-(4-Phenyl-1H-1,2,3-triazol-1-yl)-3-(2,3,5,6-tetrafluorophenoxy)propan-2-ol (0.035 g, 0.095 mmol) and Dess Martin periodinone (0.08 g, 0.19 mmol) was stirred in dichloromethane (10 mL) at room temperature for 18 h. The reaction mixture was washed with washed with saturated aqueous sodium thiosulfate, saturated aqueous sodium bicarbonate and brine. The organic layer was dried over magnesium sulfate, concentrated and isolated by reverse phase HPLC to obtain 19 mg (58%) of 1-(4-phenyl-1H-1,2,3-triazol-1-yl)-3-(2,3,5,6-tetrafluorophenoxy)propan-2-one (33) as a white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.84–7.93 (m, 3H), 7.33–7.55 (m, 3H), 6.83–7.02 (m, 1H), 5.68 (s, 2H), 4.99 (s, 2H)

LC-MS: m/z = 366.02 [M+H]⁺, 384.11 [M+H2O]⁺, 398.10 [M+MeOH]⁺

1-Chloro-3-(4-(4-nitrophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one (34)

Epichlorhydrin (0.13 mL, 1.63 mmol) as added to a solution of sodium azide (0.11g, 1.63 mmol) in water (6 mL). After stirring at room temperature for 5 min, copper(II) sulfate pentahydrate (34 mg, 0.143mmol) and sodium ascorbate (54 mg, 0.27 mmol) were added to the reaction mixture, followed by addition of 1-ethynyl-4-nitrobenzene (0.2 g, 1.36 mmol) in THF (2 mL). After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method A and purified by flash column chromatography (35 % ethyl acetate/hexanes) to obtain 122 mg (32%) of 1-chloro-3-(4-(4-nitrophenyl)-1H-1,2,3-triazol-1-yl)propan-2-ol as a cream colored solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 8.28 (d, J = 8.9 Hz, 2H), 8.08 (s, 1H), 7.95 (d, J = 8.8 Hz, 2H), 4.74 (dd, J = 14.1, 3.3 Hz, 1H), 4.57 (dd, J = 14.1, 7.0 Hz, 1H), 4.35–4.50 (m, 1H), 3.56–3.75 (m, 2H), 3.50 (br s, 1H)

LC-MS: m/z = 283.01 [M+H]⁺

1-Chloro-3-(4-(4-nitrophenyl)-1H-1,2,3-triazol-1-yl)propan-2-ol (65 mg, 0.23 mmol) and Dess Martin periodinone (0.2 g, 0.46 mmol.) were taken in dichloromethane (10 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–25% ethyl acetate/hexanes) to obtain 13 mg (20%) 1-chloro-3-(4-(4-nitrophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one (34).

¹H NMR (400 MHz, acetone, 25ºC): δ = 8.59 (s, 1H), 8.31–8.38 (m, 2H), 8.17–8.25 (m, 2H), 5.83 (s, 2H), 4.72 (s, 2H)

LC-MS: m/z = 299.01 [M+H2O]⁺, 313.05 [M+MeOH]⁺

4-(1-(3-Chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)benzonitrile (35)

A solution of 4-cyanophenylacetylene (0.2 g, 1.57 mmol) in THF (2 mL) was added to epichlorhydrin (0.15 mL, 1.86 mmol), sodium azide (0.12 g, 1.86 mmol), copper(II) sulfate pentahydrate (0.04g, 0.16 mmol) and sodium ascorbate (0.06 g, 0.31 mmol) in water (6 mL). The reaction mixture was stirred at room temperature for 18 h and worked up as described in Method A. Purification by column chromatography (35% ethyl acetate/hexanes) yielded 63 mg (15%) of 4-(1-(3-chloro-2-hydroxypropyl)-1H-1,2,3-triazol-4-yl)benzonitrile as a white solid.

¹H NMR (400 MHz, DMSO-d6, 25ºC): δ = 8.76 (s, 1H), 8.00–8.14 (m, 2H), 7.86–8.00 (m, 2H), 5.71–5.87 (m, 1H), 4.62 (dd, J = 13.9, 3.8 Hz, 1H), 4.43 (dd, J = 13.9, 7.9 Hz, 1H), 4.08–4.29 (m, 1H), 3.59–3.74 (m, 2H)

LC-MS: m/z = 262.97 [M+H]⁺

4-(1-(3-Chloro-2-hydroxypropyl)-1H-1,2,3-triazol-4-yl)benzonitrile (0.04g, 0.15 mmol) and Dess Martin periodinone (0.125 g, 0.3 mmol) were stirred in dichloromethane (1 mL) at room temperature for 18 h. The reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–35% acetone/dichloromethane) to obtain 24 mg (62%) of 4-(1-(3-chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)benzonitrile (35).

¹H NMR (400 MHz, acetone, 25ºC): δ = 8.52 (s, 1H), 8.09–8.17 (m, 2H), 7.83–7.90 (m, 2H), 5.81 (s, 2H), 4.71 (s, 2H)

LC-MS: m/z = 278.97 [M+H₂O]⁺, 293.02 [M+MeOH]⁺

1-(4-(4-Bromophenyl)-1H-1,2,3-triazol-1-yl)-3-chloropropan-2-one (36)

Epichlorhydrin (0.078 mL, 0.99 mmol) as added to a solution of sodium azide (0.065 g, 0.99 mmol) in water (5 mL) and THF (2 mL). After stirring at room temperature for 5 min, 0.3 M aqueous solution of copper(II) sulfate pentahydrate (0.28 mL, 0.083mmol) and 1M aqueous solution of sodium ascorbate (0.83 mL, 0.83 mmol) were added to the reaction mixture, followed by addition of 1-bromo-4-ethynylbenzene (0.15 g, 0.83 mmol). After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method B and purified by flash column chromatography (0–35% ethyl acetate/hexanes) to obtain 48 mg (18%) of 1-(4-(4-bromophenyl)-1H-1,2,3-triazol-1-yl)-3-chloropropan-2-ol as a white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.85 (s, 1H), 7.48–7.59 (m, 4H), 4.63–4.74 (m, 1H), 4.37–4.57 (m, 2H), 4.11 (br s, 1H), 3.67 (dd, J = 5.3, 1.6 Hz, 2H)

LC-MS: m/z = 315.95 [M+H ⁷⁹Br]⁺, 317.97 [M+H ⁸¹Br]⁺

1-(4-(4-Bromophenyl)-1H-1,2,3-triazol-1-yl)-3-chloropropan-2-ol (47 mg, 0.15 mmol) and Dess Martin periodinone (0.125 g, 0.3 mmol.) were taken in dichloromethane (10 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–35% acetone/dichloromethane) to obtain 28 mg (20%) the 1-(4-(4-bromophenyl)-1H-1,2,3-triazol-1-yl)-3-chloropropan-2-one (36) as a white powder.

¹H NMR (400 MHz, acetone, 25ºC): δ = 8.38 (s, 1H), 7.84–7.91 (m, 2H), 7.61–7.68 (m, 2H), 5.77 (s, 2H), 4.70 (s, 2H)

LC-MS: m/z = 315.95 [M+H]⁺, 333.97 [M+H2O]⁺, 346.00 [M+MeOH ⁷⁹Br]⁺, 347.96 [M+MeOH ⁸¹Br]⁺

1-Chloro-3-(4-(4-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one (37)

Epichlorhydrin (0.14 mL, 1.76 mmol) as added to a solution of sodium azide (0.11g, 1.76 mmol) in water (6 mL). After stirring at room temperature for 5 min, copper(II) sulfate pentahydrate (37 mg, 0.146 mmol) and sodium ascorbate (58 mg, 0.29 mmol) were added to the reaction mixture, followed by addition of 4-chlorophenyl acetylene (0.2 g, 1.46 mmol) in THF (2 mL). After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method A and purified by flash column chromatography (35 % ethyl acetate/hexanes) to obtain 48 mg (38%) of 1-chloro-3-(4-(4-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-ol as a white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.91 (s, 1H), 7.64 (d, J = 7.9 Hz, 2H), 7.28–7.41 (m, 2H), 4.64 (dd, J = 14.0, 3.5 Hz, 1H), 4.44 (dd, J = 14.0, 7.4 Hz, 1H), 4.23–4.35 (m, 1H), 3.58 (d, J = 5.6 Hz, 2H), 2.54 (s, 1H)

LC-MS: m/z = 271.98 [M+H ³⁵Cl]⁺, 273.93 [M+H ³⁷Cl]⁺

1-Chloro-3-(4-(4-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-ol (48 mg, 0.17 mmol) and Dess Martin periodinone (0.15 g, 0.35 mmol.) were taken in dichloromethane (10 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–35% acetone/dichloromethane) to obtain 30 mg (63%) 1-chloro-3-(4-(4-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one (37)as a white solid.

¹H NMR (400 MHz, acetone, 25ºC): δ = 8.37 (s, 1H), 7.90–7.97 (m, 2H), 7.42–7.56 (m, 2H), 5.77 (s, 2H), 4.70 (s, 2H)

LC-MS: m/z = 270.15 [M+H]⁺, 287.92 [M+H2O ³⁵Cl]⁺, 289.93 [M+H2O ³⁷Cl]⁺, 301.97 [M+MeOH, ³⁵Cl]⁺, 303.98 [M+MeOH, ³⁷Cl]⁺

1-Chloro-3-(4-(3-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one (38)

Epichlorhydrin (0.14 mL, 1.75 mmol) as added to a solution of sodium azide (0.11g, 1.75 mmol) in water (6 mL). After stirring at room temperature for 5 min, copper(II) sulfate pentahydrate (36 mg, 0.14 mmol) and sodium ascorbate (58 mg, 0.29 mmol) were added to the reaction mixture, followed by addition of 4-chlorophenyl acetylene (0.2 g, 1.46 mmol) in THF (2 mL). After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method A and purified by flash column chromatography (35 % ethyl acetate/hexanes) to obtain 130 mg (33%) of 1-chloro-3-(4-(3-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-ol as a white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.86 (s, 1H), 7.50–7.61 (m, 2H), 7.24–7.33 (m, 2H), 4.62–4.74 (m, 2H), 4.40–4.51 (m, 2H), 3.67–3.72 (m, 2H)

LC-MS: m/z = 271.98 [M+H ³⁵Cl]⁺, 273.93 [M+H ³⁷Cl]⁺

1-Chloro-3-(4-(4-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-ol (70 mg, 0.26 mmol) and Dess Martin periodinone (0.22 g, 0.51 mmol.) were taken in dichloromethane (10 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–50% acetone/dichloromethane) to obtain 55 mg (79%) 1-chloro-3-(4-(3-chlorophenyl)-1H-1,2,3-triazol-1-yl)propan-2-one (38) as a white solid.

¹H NMR (400 MHz, acetone, 25ºC): δ = 8.42 (s, 1H), 7.96 (t, J = 1.8 Hz, 1H), 7.87 (dt, J = 7.7, 1.3 Hz, 1H), 7.48 (t, J = 7.9 Hz, 1H), 7.38 (dd, J = 8.0, 2.1 Hz, 1H), 5.78 (s, 2H), 4.70 (s, 2H)

LC-MS: m/z = 270.15 [M+H]⁺, 287.92 [M+H2O ³⁵Cl]⁺, 289.93 [M+H2O ³⁷Cl]⁺, 301.97 [M+MeOH, ³⁵Cl]⁺, 303.98 [M+MeOH, ³⁷Cl]⁺

N-(3-(1-(3-Chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)phenyl)acetamide (39)

Epichlorhydrin (0.089 mL, 1.13 mmol) as added to a solution of sodium azide (0.074 g, 1.13 mmol) in water (5 mL) and THF (2 mL). After stirring at room temperature for 5 min, 0.3 M aqueous solution of copper(II) sulfate pentahydrate (0.31 mL, 0.094mmol) and 1M aqueous solution of sodium ascorbate (0.94 mL, 0.94 mmol) were added to the reaction mixture, followed by addition of N-(3-ethynylphenyl)acetamide (0.15 g, 0.94 mmol). After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method B and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 82 mg (30%) of N-(3-(1-(3-chloro-2-hydroxypropyl)-1H-1,2,3-triazol-4-yl)phenyl)acetamide as a white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.94 (s, 1H), 7.83 (s, 1H), 7.47–7.64 (m, 2H), 7.22–7.37 (m, 2H), 4.66 (dd, J = 12.9, 1.9 Hz, 1H), 4.35–4.55 (m, 2H), 3.70 (d, J = 5.1 Hz, 2H), 2.21 (s, 3H)

LC-MS: m/z = 295.04 [M+H]⁺

N-(3-(1-(3-Chloro-2-hydroxypropyl)-1H-1,2,3-triazol-4-yl)phenyl)acetamide (53 mg, 0.18 mmol) and Dess Martin periodinone (0.14 g, 0.36 mmol.) were taken in dichloromethane (10 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 4.6 mg (9%) N-(3-(1-(3-chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)phenyl)acetamide (39) as a pale pink solid.

¹H NMR (400 MHz, acetone, 25ºC): δ = 9.23 (br s, 1H), 8.16–8.32 (m, 2H), 7.60–7.70 (m, 1H), 7.57 (d, J = 7.8 Hz, 1H), 7.26–7.47 (m, 1H), 5.75 (s, 2H), 4.69 (s, 2H), 2.12 (s, 3H)

LC-MS: m/z = 293.02 [M+H]⁺, 311.04 [M+H2O]⁺, 325.09 [M+MeOH]⁺

3-((1-(3-Chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)methyl)-1H-indene-1,2(3H)-dione (40)

Epichlorhydrin (0.051 mL, 0.65 mmol) as added to a solution of sodium azide (0.042 g, 0.65 mmol) in water (5 mL) and THF (2 mL). After stirring at room temperature for 5 min, 0.3 M aqueous solution of copper(II) sulfate pentahydrate (0.18 mL, 0.054mmol) and 1M aqueous solution of sodium ascorbate (0.54 mL, 0.54 mmol) were added to the reaction mixture, followed by addition of 1-(prop-2-yn-1-yl)indoline-2,3-dione (0.1 g, 0.54 mmol). After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method B and purified by flash column chromatography (0–50% acetone/dichloromethane) to obtain 73 mg (42%) of 1-((1-(3-chloro-2-hydroxypropyl)-1H-1,2,3-triazol-4-yl)methyl)indoline-2,3-dione as a beige solid.

¹H NMR (400 MHz, DMSO-d6, 25ºC): δ = 8.13 (s, 1H), 7.64 (td, J = 7.8, 1.3 Hz, 1H), 7.58 (dd, J = 7.4, 0.7 Hz, 1H), 7.18 (d, J = 7.9 Hz, 1H), 7.14 (td, J = 7.5, 0.7 Hz, 1H), 5.69 (d, J = 5.5 Hz, 1H), 4.97 (s, 2H), 4.49 (dd, J = 13.9, 3.8 Hz, 1H), 4.32 (dd, J = 13.9, 7.8 Hz, 1H), 4.07 (td, J = 3.8, 1.5 Hz, 1H), 3.49–3.66 (m, 2H)

LC-MS: m/z = 321.06 [M+H]⁺

1-((1-(3-Chloro-2-hydroxypropyl)-1H-1,2,3-triazol-4-yl)methyl)indoline-2,3-dione (40 mg, 0.12 mmol) and Dess Martin periodinone (0.1 g, 0.25 mmol.) were taken in dichloromethane (10 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–35% acetone/dichloromethane) to obtain 6.5 mg (16%) the 3-((1-(3-chloro-2-oxopropyl)-1H-1,2,3-triazol-4-yl)methyl)-1H-indene-1,2(3H)-dione (40) as an orange colored solid.

¹H NMR (400 MHz, acetone, 25ºC): δ = 8.01 (s, 1H), 7.69 (td, J = 7.8, 1.3 Hz, 1H), 7.58 (dd, J = 7.5, 0.8 Hz, 1H), 7.29 (d, J = 7.9 Hz, 1H), 7.19 (td, J = 7.5, 0.7 Hz, 1H), 5.67 (s, 2H), 5.08 (s, 2H), 4.63 (s, 2H)

LC-MS: m/z = 318.98 [M+H]⁺, 336.99 [M+H2O]⁺, 351.04 [M+MeOH]⁺

((1-Azido-3-chloropropan-2-yl)oxy)(tert-butyl)dimethylsilane

Step1:

Epichlorhydrin (2 mL, 25.5 mmol) was added to a mixture of sodium azide (1.82 g, 28 mmol) and acetic acid (1.6 mL, 28 mmol) in 4:1 mixture of water/ethanol (10 mL). After stirring at room temperature for 18 h, to the reaction mixture, was added ethyl acetate and the brine. The organic layer was separated, dried over magnesium sulfate and concentrated to obtain about 3.0 g of 1-azido-3-chloropropan-2-ol as a colorless oil which was used without further purification.

Step 2:

tert-Butyldimethylsilyl chloride (2.7 g, 17.7 mmol) was added to a mixture of 1-azido-3-chloropropan-2-ol (2.0 g, 14.75 mmol) and imidazole (2.5 g, 36.88 mmol) taken in DMF (40 mL). After stirring at room temperature for 18 h, the reaction mixture was diluted with ethyl acetate, washed with water and brine. The organic layer was dried over magnesium sulfate, concentrated and purified by flash column chromatography (0–30% ethyl acetate/hexanes) to obtain 2.87 g (78%) of ((1-azido-3-chloropropan-2-yl)oxy)(tert-butyl)dimethylsilane as a colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 3.95–4.01 (m, 1H), 3.43–3.57 (m, 3H), 3.30–3.36 (m, 1H), 0.89–0.96 (m, 9H), 0.15 (d, J = 3.7 Hz, 6H)

¹³C NMR (100 MHz, CHLOROFORM-d, 25ºC): δ = 71.9, 54.1, 45.1, 25.7, 25.6, 17.9, - 4.8, −4.9

(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl methylcarbamate (41)

3-Phenylprop-2-yn-1-ol (0.53 g, 4.0mmol) and ((1-azido-3-chloropropan-2-yl)oxy)(tert-butyl)dimethylsilane (1.0 g, 4.0mmol) in toluene (5 mL) was added to a degassed mixture of chloro(1,5-cyclooctadiene)(η5-pentamethylcyclopentadienyl)ruthenium (0.077 g, 0.2mmol) in toluene (5 mL). After stirring at 70 °C for 18 h, the reaction mixture was diluted with ethyl acetate, washed with water and brine. The organic layer was dried over magnesium sulfate, concentrated and purified by flash column chromatography (0–50% ethyl acetate/hexanes) to obtain 1.16 g (76%) of (1-(2-((tert-butyldimethylsilyl)oxy)-3-chloropropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methanol as a green solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.67 (d, J = 6.6 Hz, 2H), 7.36–7.46 (m, 3H), 4.78–4.89 (m, 2H), 4.59–4.67 (m, 1H), 4.47–4.55 (m, 2H), 3.78 (br s, 1H), 3.54–3.62 (m, 2H), 0.80 (s, 9H), −0.04–0.12 (m, 3H), −0.27 (s, 3H)

¹³C NMR (100 MHz, CHLOROFORM-d, 25ºC): δ = 145.7, 133.3, 130.6, 128.8, 128.3, 127.8, 71.4, 52.4, 51.5, 46.1, 25.6, 17.9, −5.3

LC-MS: m/z = 382.10 [M+H ³⁵Cl]⁺, 384.12 [M+H ³⁷Cl]⁺

(1-(2-((tert-Butyldimethylsilyl)oxy)-3-chloropropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methanol (0.5 g, 1.3 mmol) and DIEA (0.23 mL, 1.3 mmol) in DCM (5 mL) were slowly added to a cooled (0 °C) solution of p-nitrophenylchloro formate (0.28g, 1.37 mmol) in DCM (10 mL). After sirring at room temperature for 18 h, the reaction mixture was washed with water and brine, dried over magnesium sulfate, concentrated and purified by flash column chromatography (0–40% ethyl acetate/hexanes) to obtain 0.41 g (58%) of (1-(2-((tert-butyldimethylsilyl)oxy)-3-chloropropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate as a pale yellow oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 8.29 (d, J = 9.3 Hz, 2H), 7.76 (d, J = 8 Hz, 2H), 7.38–7.54 (m, 5H), 5.61 (s, 1H), 5.58 (s, 1H), 5.52–5.68 (m, 1H), 5.44–5.49 (m, 1H), 4.81 (dd, J = 14.1, 3.7 Hz, 1H), 4.66 (dd, J = 14.1, 8.5 Hz, 1H), 4.49–4.55 (m, 1H), 3.58–3.66 (m, 2H), 0.81 (s, 9H), 0.04 (s, 3H), −0.28−-0.25 (m, 3H)

LC-MS: m/z = 547.06 [M+H ³⁵Cl]⁺, 549.08 [M+H ³⁷Cl]⁺

(1-(2-((tert-Butyldimethylsilyl)oxy)-3-chloropropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (0.41 g, 0.75 mmol) and 4M hydrochloric acid in dioxane (7.5 mL) were stirred at room temperature for 18 h. The reaction mixture was concentrated down to dryness to obtain 0.33 g of the crude (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate as a yellow foaming solid which was used without further purification.

LC-MS: m/z = 433.01 [M+H ³⁵Cl]⁺, 434.96 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (24 mg, 0.055 mmol) and 2M solution of methylamine in THF (0.055 mL, 0.11 mmol) were stirred in THF (1 mL) at room temperature for 5 h. The reaction mixture was worked up as described in Method D and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 18 mg(100%) of (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl methylcarbamate as colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.70 (br d, J = 7.2 Hz, 2H), 7.38–7.53 (m, 3H), 5.19–5.41 (m, 2H), 5.10–5.45 (m, 1H), 4.62–5.02 (m, 3H), 4.39–4.54 (m, 1H), 3.70 (d, J = 5.5 Hz, 2H), 2.73–2.91 (m, 3H)

¹³C NMR (100 MHz, CHLOROFORM-d, 25ºC): δ = 156.2, 146.9, 130.1, 129.6, 128.9, 128.6, 127.5, 70.4, 54.0, 51.5, 46.2, 27.7

LC-MS: m/z = 325.02 [M+H ³⁵Cl]⁺, 327.04 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl methylcarbamate (18 mg, 0.055 mmol) and Dess Martin periodinone (47 mg, 0.11 mmol.) were taken in dichloromethane (2mL) After stirring at room temperature for 5 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–35% acetone/dichloromethane) to obtain 15 mg (85%) of 1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl methylcarbamate (41) as white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.74–7.93 (m, 2H), 7.41–7.57 (m, 3H), 5.79 (s, 2H), 5.20 (s, 2H), 4.31 (s, 2H), 2.82 (d, J = 5.0 Hz, 3H)

LC-MS: m/z = 323.01 [M+H]⁺, 355.01 [M+MeOH]⁺

(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl dimethylcarbamate (42)

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (35 mg, 0.08 mmol) and 2M solution of dimethylamine in THF (0.08 mL, 0.16 mmol) were stirred in THF (1 mL) at room temperature for 24 h. The reaction mixture was worked up as described in Method D and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 23 mg (85%) of (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl dimethylcarbamate as colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.69–7.80 (m, 2H), 7.38–7.51 (m, 3H), 5.31–5.40 (m, 1H), 5.24–5.31 (m, 1H), 4.67–4.88 (m, 2H), 4.40–4.51 (m, 1H), 4.18 (br s, 1H) 3.66–3.75 (m, 2H), 2.92 (s, 3H), 2.87 ppm (s, 3H)

LC-MS: m/z = 339.01 [M+H ³⁵Cl]⁺, 341.03 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl dimethylcarbamate (20 mg, 0.06 mmol) and Dess Martin periodinone (50 mg, 0.12 mmol.) were taken in dichloromethane (2mL) After stirring at room temperature for 5 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–50% acetone/dichloromethane) to obtain 15 mg (75%) of 1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl dimethylcarbamate (42) as white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.84–7.90 (m, 2H), 7.41–7.57 (m, 3H), 5.82 (s, 2H), 5.21 (s, 2H), 4.32 (s, 2H), 2.93 (s, 6H)

LC-MS: m/z = 337.06 [M+H]⁺, 369.00 [M+MeOH]⁺

(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl benzylcarbamate (43)

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl benzylcarbamate (PJ169–121)

1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (20 mg, 0.046 mmol) and benzylamine (0.01mL, 0.092 mmol) were stirred in THF (1 mL) at room temperature for 3h. The reaction mixture was worked up as described in Method D and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 10 mg (54%) of (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl benzylcarbamate as colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.73 (br d, J = 7.1 Hz, 2H), 7.27–7.49 (m, 8H), 5.27–5.44 (m, 2H), 5.22 (br s, 1H), 4.55–4.79 (m, 2H), 4.27–4.52 (m, 2H), 3.73–3.94 (m, 1H), 3.61–3.73 (m, 2H)

LC-MS: m/z = 401.00 [M+H ³⁵Cl]⁺, 403.02 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl benzylcarbamate (10 mg, 0.025 mmol) and Dess Martin periodinone (21 mg, 0.05 mmol.) were taken in dichloromethane (1 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–25% acetone/dichloromethane) to obtain 7.5 mg (75%) (1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl benzylcarbamate (43) as white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.82 (br d, J = 7.3 Hz, 2H), 7.26–7.52 (m, 9H), 5.78 (s, 2H), 5.23 (s, 2H), 4.37 (d, J = 6.0 Hz, 2H), 4.26 (s, 2H)

LC-MS: m/z = 399.05 [M+H]⁺, 417.07 [M+H2O]⁺, 431.06 [M+MeOH]⁺

(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-methoxyethyl)carbamate (44)

1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (32 mg, 0.074 mmol) and 2-methoxyethylamine (0.013 mL, 0.15 mmol) were stirred in THF (1 mL) at room temperature for 3h. The reaction mixture was worked up as described in Method D and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 24 mg (54%) of (1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-methoxyethyl)carbamate as colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.72 (br d, J = 6.0 Hz, 2H), 7.39–7.52 (m, 3H), 6.94 (d, J = 9.1 Hz, 1H), 5.31–5.41 (m, 1H), 5.26 (br d, J = 14.1 Hz, 2H), 4.62–4.80 (m, 2H), 4.35–4.51 (m, 1H), 4.01 (br s, 1H), 3.64–3.78 (m, 2H), 3.43–3.51 (m, 2H), 3.31–3.43 (m, 4H)

LC-MS: m/z = 369.06 [M+H ³⁵Cl]⁺, 371.01 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-methoxyethyl)carbamate (24mg, 0.065 mmol) and Dess Martin periodinone (55 mg, 0.13 mmol.) were taken in dichloromethane (2 mL) After stirring at room temperature for 3 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–100% ethyl acetate/hexanes to obtain 22 mg (92%) of (1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-methoxyethyl)carbamate (44) as white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.83 (d, J = 7.2 Hz, 2H), 7.41–7.56 (m, 3H), 5.78 (s, 2H), 5.24 (br s, 1H), 5.19 (s, 2H), 4.31 (s, 2H), 3.42–3.51 (m, 2H), 3.31–3.42 (m, 5H)

LC-MS: m/z = 367.05 [M+H]⁺, 385.00 [M+H2O]⁺

(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl morpholine-4-carboxylate (45)

1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (32 mg, 0.074 mmol) and morpholine (0.013 mL, 0.15 mmol) were stirred in THF (1 mL) at room temperature for 3h. The reaction mixture was worked up as described in Method D and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 28 mg (100%) of (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl morpholine-4-carboxylate as colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.59–7.78 (m, 2H), 7.39–7.51 (m, 3H), 5.40 (d, J = 13.8 Hz, 1H), 5.30 (br d, J = 13.8 Hz, 1H), 4.62–4.79 (m, 2H), 4.45 (dt, J = 3.6, 1.7 Hz, 1H), 4.07–4.23 (m, 1H), 3.30–3.84 ppm (m, 10H)

LC-MS: m/z = 381.03 [M+H ³⁵Cl]⁺, 383.05 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl morpholine-4-carboxylate (28 mg, 0.074 mmol) and Dess Martin periodinone (62 mg, 0.15 mmol.) were taken in dichloromethane (2 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–100% ethyl acetate/hexanes followed by 0–25% acetone/dichloromethane) to obtain 16 mg (57%) of (1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl morpholine-4-carboxylate (45) as white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25ºC): δ = 7.76–7.90 (m, 2H), 7.41–7.58 (m, 3H), 5.78 (s, 2H), 5.25 (s, 2H), 4.31 (s, 2H), 3.66 (br d, J = 17.0 Hz, 4H), 3.46 (br s, 4H)

LC-MS: m/z = 379.08 [M+H]⁺, 397.10 [M+H2O]⁺

(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-morpholinoethyl)carbamate (46)

1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (32 mg, 0.074 mmol) and 4-(2-aminoethyl)morpholine (0.021 mL, 0.16 mmol) were stirred in THF (1 mL) at room temperature for 3h. The reaction mixture was worked up as described in Method D and purified by flash column chromatography (0–100% ethyl acetate/hexanes) to obtain 24 mg (70%) of (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-morpholinoethyl)carbamate as colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25°C): δ = 7.70 (br d, J = 7.2 Hz, 2H), 7.38–7.50 (m, 3H), 5.28–5.46 (m, 2H), 4.62–4.78 (m, 2H), 4.43 (br dd, J = 6.1, 4.1 Hz, 2H), 3.63–3.76 (m, 6H), 3.48 (s, 1H), 2.35–2.56 ppm (m, 6H)

LC-MS: m/z = 424.06 [M+H ³⁵Cl]⁺, 426.02 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-morpholinoethyl)carbamate (32mg, 0.075mmol) and Dess Martin periodinone (64 mg, 0.15 mmol) were taken in dichloromethane (2 mL) After stirring at room temperature for 18 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–100% ethyl acetate/hexanes followed by 0–10% methanol/dichloromethane) to obtain 2 mg (6%) of (1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (2-morpholinoethyl)carbamate (46) as a clear oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25°C): δ = 7.69–7.93 (m, 2H), 7.38–7.57 (m, 3H), 5.77 (s, 2H), 5.22 (s, 2H), 4.33 (s, 2H), 4.14 (d, J = 7.2 Hz, 2H), 3.75 (br s, 4H), 3.31 (br s, 2H), 2.33–2.61 (m, 4H)

LC-MS: m/z = 422.05 [M+H]⁺, 440.07 [M+H₂O]⁺

(1-(3-Chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (3-(2-oxopyrrolidin-1-yl)propyl)carbamate (47)

1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (30 mg, 0.07 mmol) and 1-(3-aminopropyl)-2-pyrrolidine (0.01 mL, 0.14 mmol) were stirred in THF (1 mL) at room temperature for 18h. The reaction mixture was worked up as described in Method D and purified by flash column chromatography (0–100% ethyl acetate/hexanes followed by 0–10% methanol/dichloromethane) to obtain 25 mg (82%) of (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (3-(2-oxopyrrolidin-1-yl)propyl)carbamate as colorless oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25°C): δ = 7.70–7.81 (m, 2H), 7.37–7.51 (m, 3H), 5.90–6.10 (m, 1H), 5.36 (br d, J = 13.5 Hz, 1H), 5.24 (br d, J = 13.6 Hz, 1H), 4.57–4.78 (m, 2H), 4.32–4.52 (m, 1H), 3.62–3.79 (m, 2H), 3.24–3.42 (m, 4H), 3.04–3.24 (m, 2H), 2.39 (br t, J = 8.0 Hz, 2H), 1.97–2.09 (m, 2H), 1.53–1.77 (m, 2H)

LC-MS: m/z = 436.10 [M+H ³⁵Cl]⁺, 438.12 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (3-(2-oxopyrrolidin-1-yl)propyl)carbamate (25mg, 0.057 mmol) and Dess Martin periodinone (49 mg, 0.11 mmol) were taken in dichloromethane (2 mL) After stirring at room temperature for 3 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–50% acetone/dichloromethane) to obtain 3.5 mg (14%) of (1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (3-(2-oxopyrrolidin-1-yl)propyl)carbamate (47) as a clear oil.

¹H NMR (400 MHz, CHLOROFORM-d, 25°C): δ = 7.73–7.95 (m, 2H), 7.38–7.59 (m, 3H), 6.05 (br s, 1H), 5.79 (s, 2H), 5.19 (s, 2H), 4.35 (s, 2H), 3.28–3.43 (m, 4H), 3.14 (q, J = 6.2 Hz, 2H), 2.29–2.51 (m, 2H), 1.87–2.12 (m, 2H), 1.68 (br t, J = 5.9 Hz, 2H)

LC-MS: m/z = 434.08 [M+H]⁺

1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl prop-2-yn-1ylcarbamate (48)

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl (4-nitrophenyl) carbonate (32 mg, 0.074 mmol) and propargylamine (0.01 mL, 0.15 mmol) were stirred in THF (1 mL) at room temperature for 3 h. The reaction mixture was worked up as described in Method D and purified by flash column chemoatography (0–100% ethyl acetate/hexanes) to obtain 24 mg (92%) of (1-(3-chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl prop-2-yn-1ylcarbamate as colorless oil. ¹H NMR (400 MHz, CHLOROFORM-d, 26°C): δ = 7.62–7.78 (m, 2H), 7.36–7.56 (m, 3H), 5.12–5.45 (m, 3H), 4.59–4.79 (m, 2H), 4.35–4.51 (m, 1H), 3.82–4.04 (m, 3H), 2.29 (s, 1H).

LC-MS: m/z = 349.18 [M+H ³⁵Cl]⁺, 351.08 [M+H ³⁷Cl]⁺

(1-(3-Chloro-2-hydroxypropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl prop-2-yn-1ylcarbamate (24 mg, 0.07 mmol) and Dess Martin periodinone (58 mg, 0.14 mmol.) were taken in dichloromethane (5mL) After stirring at room temperature for 5 h, the reaction mixture was worked up as described in Method C and purified by flash column chromatography (0–50% acetone/dichloromethane) to obtain 18 mg (75%) of 1-(3-chloro-2-oxopropyl)-4-phenyl-1H-1,2,3-triazol-5-yl)methyl prop-2-yn-1ylcarbamate as white solid.

¹H NMR (400 MHz, CHLOROFORM-d, 25°C): δ = 7.68–7.89 (m, 2H), 7.42–7.58 (m, 3H), 5.72–5.84 (m, 2H), 5.23 (s, 2H), 5.13 (br s, 1H), 4.31 (s, 2H), 3.98 (br dd, J = 5.5, 2.3 Hz, 2H), 2.29 (t, J = 2.2 Hz, 1H).

LC-MS: m/z = 347.08 [M+H]⁺

QUANTIFICATION AND STATISTICAL ANALYSIS

Average values were calculated from replicates (triplicate or duplicate, technical replicates for biochemical assays, biological replicates for antiviral assays) and are reported as the mean ± SD. Details on quantification are found in figure legends and in Methods Details sections. For determination of kinetic values (k_cat, K_m, k_inact, K_I), assay conditions were such that general kinetic assumptions are valid (e.g. steady state kinetics), allowing for mathematical modeling and determination of values. For IC₅₀ and EC₅₀ values, we do not compare potencies statistically and thus did not perform any statistical tests of validation.

Supplementary Material

3

Video S1. Sample MD simulation of 43 bound to monomer, binding mode 1, related to Figure 3. Video exemplifying MD simulations of 43 bound to one HCMV C-A chain, with inhibitor oriented toward the dimer interfacial helix α5. Video corresponds to Figure 3d, monomer panel 1 (left). R165 is highlighted in green, 43 is aquamarine.

4

Video S2. Sample MD simulation of 43 bound to monomer, binding mode 2, related to Figure 3. Video exemplifying MD simulations of 43 bound to one HCMV C-A chain, with inhibitor oriented toward the S₁ substrate pocket. Video corresponds to Figure 3d, monomer panel 3 (left). R165 is highlighted in green, 43 is aquamarine.

5

Video S3. Sample MD simulation of 43 bound to dimer, related to Figure 3. Video exemplifying MD simulations of 43 bound to the HCMV C-A homodimer, with inhibitor oriented toward the dimer interfacial helix α5. Video corresponds to Figure 3d, dimer panel 1 (right). Chain A is white, chain B is blue, 43 is aquamarine. R165 (chain A) and R109 (chain B) are both green.

Significance.

We have established an integrative pipeline to inhibit recalcitrant drug targets that 1) demonstrates irreversibly labeling a non-catalytic cysteine is a viable method for inactivating viral protease activity and disrupting viral infectivity, 2) demonstrates that targeting conserved residues is a method to develop broadly reactive inhibitors capable of inhibiting homologous enzyme families and 3) highlights the power of using covalent molecules to capture dynamic enzymes in inactivated conformational states. The chemical tools developed herein provide proof of principle that can be leveraged in the future to develop antiviral agents suitable for drug development.

Highlights.

• HHV proteases are inhibited by targeting a non-catalytic cysteine

• Inhibitors bind the protease active site and engage the dimer interface

• Protease inhibition stabilizes the inactivated homodimer

• Inhibition disrupts HHV infectivity in cells