Synthesis and Target Identification of a Novel Electrophilic Warhead, 2

Communication Cite This: Biochemistry 2019, 58, 2715−2719

pubs.acs.org/biochemistry

Synthesis and Target Identification of a Novel Electrophilic Warhead, 2‑Chloromethylquinoline Feng Ni,†,‡,∥,⊥ Arunika Ekanayake,†,‡,∥ Bianca Espinosa,†,‡ Caiqun Yu,†,‡ Jacob N. Sanders,§ John Perino,†,‡ K. N. Houk,§ and Chao Zhang*,†,‡ †

Department of Chemistry, University of Southern California, Los Angeles, California 90089, United States Loker Hydrocarbon Research Institute, University of Southern California, Los Angeles, California 90089, United States § Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095, United States Downloaded via GUILFORD COLG on July 18, 2019 at 01:00:16 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: Despite its power in identifying highly potent ligands for select protein targets, conventional medicinal chemistry is limited by its low throughput and lack of proteomic selectivity information. We seek to develop a chemoproteomic approach for discovering covalent ligands for protein targets in an unbiased, highthroughput manner. Tripartite probe compounds composed of a heterocyclic core, an electrophilic “warhead”, and an alkyne tag have been designed and synthesized for covalently labeling and identifying targets in cells. We have developed a novel condensation reaction to prepare 2-chloromethylquinoline (2-CMQ), an electrophilic heterocycle. These chloromethylquinolines potently and covalently bind to a number of cellular protein targets, including prostaglandin E synthase 2 (PTGES2), a critical regulator of cell proliferation, apoptosis, angiogenesis, inflammation, and immune surveillance. The 2-CMQs that we have developed here are novel PTGES2 binders that have the potential to serve as therapies for the treatment of human diseases such as inflammation.

Figure 1. Structures and experimental and calculated activation free energies of a series of electrophiles, including 2-chloromethylquinoline. (A) Structures of the acrylamide-containing cancer drug afatinib, a chloromethyl ketone-containing caspase inhibitor Ac-YVAD-CMK, and a highly specific EphB3 kinase inhibitor containing chloroacetamide. (B) Computed (red) and experimental (blue) free energies of activation for five electrophiles reacting with glutathione, computationally modeled as methyl sulfide.

S

mall molecules that are able to form covalent bonds with their target proteins are becoming increasingly important as both clinical therapeutic agents and research tools for biological studies.1 In particular, the covalent modification of proteins at cysteine (Cys) residues by small molecules has widespread applications in drug discovery.2−4 Because of the strength and often irreversible nature of covalent interactions, such drugs or tool compounds tend to possess strong binding affinity for their target proteins and elicit long-lasting effects on cells and organisms. For example, afatinib,5 a secondgeneration EGFR inhibitor, affords a more potent and durable blockade of EGFR activity than the first-generation EGFR inhibitors such as gefitinib because it contains an acrylamide group, an electrophilic warhead, for targeting a nonconserved Cys residue in EGFR (Figure 1A). Similarly, Ac-YVAD-CMK,6 a tool compound commonly used to inhibit the activity of caspase-1 and -4, harbors chloromethylketone, a different electrophilic warhead, for covalently targeting a catalytic Cys residue in caspases. EphB3 inhibitor 10, which contains a chloroacetamide warhead, is an ultraspecific that was © 2019 American Chemical Society

previously developed in our lab to target a unique cysteine in the EphB3 kinase (Figure 1A).7 A major advantage of covalent drugs and probes lies in the fact that the covalent linkages formed between probes and their target proteins can be exploited to facilitate the enrichment and identification of the targets.2,8 For example, our lab previously discovered that a chloroacetamidecontaining quinazoline potently and specifically bound to subunit A of the vacuolar ATPase.9 These applications rely on the availability of an arsenal of electrophiles with tunable reactivity and selectivity. The common electrophilic warheads are limited to a few electrophiles such as acrylamide, vinylsulfonamide, chloroacetamide, and halomethyl ketone (Figure 1A). Among these, the latter ones contain a halogen Received: April 21, 2019 Revised: May 29, 2019 Published: June 4, 2019 2715

DOI: 10.1021/acs.biochem.9b00359 Biochemistry 2019, 58, 2715−2719

Communication

Biochemistry

lines to yield 2-CMQs also suffers from limited access to the required starting materials.15 Unfortunately, the methods described above either have limited functional group tolerance or require multistep transformation to yield one 2-CMQ (Figure 1A). A more efficient method is needed to generate probes of diverse functionality and substitution pattern to enable their biological evaluation. Herein, we report an efficient condensation reaction that yields 2-CMQs from readily accessible chloroacetamide precursors in a single step. We further identified three cellular targets of 2-CMQs and mapped the binding site for one target, prostaglandin E synthase 2 (PTGES2). We developed a one-step condensation method for the preparation of 2-chloromethylquinolines by adapting a previously described reaction for constructing quinolines16 (Scheme 1). In this method, a mixture of PCl5 and POCl3 is

atom as a leaving group and an adjacent carbonyl group as an electron-withdrawing group to activate the methylene carbon for nucleophilic attack from the Cys sulfhydryl. We reasoned that an electron-deficient aryl ring such as quinoline might replace the carbonyl’s role in enhancing the electrophilicity of the halomethyl group. Importantly, such 2-halomethylquinolines are expected to have unique reactivity profiles in the proteome because their structures are distinct from those of the commonly used electrophiles such as chloromethylketone and chloroacetamide. To investigate the feasibility of this idea, we performed kinetics experiments with glutathione (GSH) serving as a thiolbased nucleophile (Figure 1B and Table S1). We first measured the rate constants for the reactivity of compounds containing chloroacetamide, acrylamide, 2-chloromethylpyridine (2-CMP), or 2-chloromethylquinoline (2-CMQ) with GSH. It was found that the reactivities of 2-CMP and 2-CMQ were similar to that of chloroacetamide, verifying the feasibility of these novel arene-based electrophiles being used as warheads in chemical biology research. Interestingly, model reactions indicate that 2-CMQs can react with not only thiols but also amines in biological systems (Table S2), which broadens the application scope of 2-CMQ as an electrophilic warhead. We conducted further quantum mechanical computations on this series of five electrophiles under the conditions used in the experimental measurement (Figure 1B). First, the computations reveal that 2-chloromethylpyridines and 2chloromethylquinolines (3−5) preferentially react via the more electrophilic N-protonated species (3H−5H), even though the free energy barriers are adjusted upward on the basis of the pKa to account for the population of the protonated species at pH 7.4. Second, although the computations systematically overestimate the free energy barriers, both the experiments and the computations reveal that 2-chloromethylquinolines 4H and 5H have free energies of activation that are lower than those of acrylamide (1), chloroacetamide (2), or 2-chloromethylpyridine (3H) electrophiles. The discrepancy between experiment and computation is likely related to the use of a continuum solvation model for water in the computations, which makes it challenging to account for differences in the entropy of solvation between the starting materials and the transition state. Taken together, both the computations and the experiments establish that 2chloromethylquinolines constitute a privileged scaffold for reaction with sulfur nucleophiles and that they likely react through their N-protonated forms. These results also suggest that 2-CMQ’s reactivity is comparable to that of chloroacetamide, a commonly used warhead for the thiol group.7,10,11 Nearly four dozen FDA-approved drugs contain a quinoline ring, indicating that the quinoline as a scaffold binds to proteins from diverse families and affords druglike properties (www.fda.gov). However, a limited number of methods have been reported for the preparation of 2-CMQs. Traditionally, 2CMQs are synthesized from quinaldines by their conversion to N-oxides, followed by the reaction with sulfonyl chlorides.12 Another typical method is the radical reaction of 2methylquinolines with N-chlorosuccinimide;13 however, this reaction suffers from the formation of dihalogenated, trihalogenated, or other undesired halogenated products.14 Although 2-quinolinemethanols can be readily converted into 2-CMQs via halogenation, the required starting materials, 2quinolinemethanols, are usually not commercially available. Similarly, the recently reported chlorination of 2-methylquino-

Scheme 1. Efficient and Divergent Synthesis of 2Chloromethylquinolines in a One-Step Condensation Reaction

used to condense two arylacetamides, readily available materials with tremendous choices of structural diversity, to construct the quinoline ring in a single step. By condensing one arylacetamide containing a propargyloxy group and another arylacetamide of divergent substitution position, we were able to quickly generate a panel of 16 2-CMQs that all contained a clickable tag and thus could be readily used to investigate their proteomic reactivity profiles. The same condensation reaction can be also used to produce 2-CMQs that do not contain an alkyne tag. We first examined the proteome reactivity of the 16 2-CMQ probes (Figure 2A and Figure S1). HEK293H (293) cells were treated with each probe at 0.1 or 1.0 μM for 1 h before being lysed and clicked with TAMRA-azide using CuAAC. The samples were resolved by sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE) and then scanned for in-gel fluorescence (Figure 1B, lanes selected). To our delight, 2-CMQs labeled a few distinct bands in the proteome. For example, 1.0 μM probe 15 labeled three prominent bands with sizes between 20 and 50 kDa (Figure 1B). The different probes have subtly different labeling patterns indicating the different spectrum of protein targets. These data demonstrate that the proteome reactivity of 2-CMQs can be finely adjusted by modulating the electronics and sterics of the quinoline ring system, thereby providing a highly tunable electrophile for chemoproteomic studies. We then chose probe 15 for further investigation of its proteomic targets because it labeled three prominent bands (A−C) with sizes between 20 and 50 kDa with high intensity (Figure 2B and Figure S1). To determine the binding saturability, we prepared compound 18, which is a close analogue of probe 15 and contains a methoxy group instead of the propargyloxy group, using the same condensation reaction. The 293 cells were pretreated with various concentrations of 18 for 1 h before being washed out and further incubated with 3 μM 15 for 1 h (Figure 3A). This competition assay showed that all three prominent bands were almost abolished by high 2716

DOI: 10.1021/acs.biochem.9b00359 Biochemistry 2019, 58, 2715−2719

Communication

Biochemistry

results revealed IC50 values of ∼1 μM for binding of the 2CMQs to targets A−C, a good starting point considering that no chemical optimization was done yet. We next set out to identify 2-CMQ’s prominent covalent targets in live cells. The 293 cells were treated with probe 15, lysed, clicked to biotin-azide, and enriched with streptavidin resins before elution and resolution with SDS−PAGE. The sections corresponding to the region containing bands A−C were cut off from the gel and subjected to MS analysis (Figure S2 and Tables S3 and S4). This led to the identification of glutathione S-transferase ω-1 (GSTO1), heme oxygenase 2 (HMOX2), and prostaglandin E synthase 2 (PTGES2). Previous studies have identified GSTO1, HMOX2, and several PTGES enzymes as proteins with hyperactive cysteines.17,18 A study focusing on covalent kinase inhibitors was able to identify HMOX2 and PTGES2 as non-kinase targets of the probes described in this study.19 The potency of our CMQ in engaging GSTO1 is lower than that of the optimized probe reported by Cravatt. We chose to focus on one target, PTGES2, an isomerase responsible for the conversion of prostaglandin H2 to prostaglandin E2, a critical regulator of cell proliferation, apoptosis, angiogenesis, inflammation, and immune surveillance.20,21 PTGES2 is expressed as a Golgiassociated protein with a full length of 42 kDa, and spontaneous cleavage of N-terminal residues 1−87 produces the catalytically active cytosolic fragment of approximately 34 kDa.22 As band C had an approximate size of 37 kDa, we suspected that what 2-CMQ labeled was the soluble fragment of PTGES2 (sPTGES2). To test this, we transfected either full-length PTGES2 (flPTGES2) or sPTGES2 into cells before they were labeled with probe 15 and subjected to in-gel fluorescence analysis. Transfection of flPTGES2 produced a new band that was labeled by probe 15, while transfection of sPTGES2 enhanced the existing band C (Figure S3). This result strongly supports the soluble fragment of PTGES2 as the protein responsible for forming band C. We purified recombinant sPTGES2 and confirmed that it could be labeled by probe 15 during in-gel fluorescence experiments (Figure S4). Furthermore, the labeling of sPTGES2 by probe 15 could be competed off by compound 18 (Figure S4). We went on to identify the residue in PTGES2 that was covalently modified by probe 15. Two cysteine residues are present in the cytoplasmic fragment of PTGES2. We mutated either of them to alanine, transfected these mutants, and analyzed labeling patterns of the resulting cells by probe 15. The C110A mutation dramatically decreased the level of labeling of the 37 kDa band, while the C113A mutation had little effect, implicating C110 as the site of probe modification, which is also the essential residue for PTGES2 activity (Figure 3B).23 Consistent with the mutagenesis results, a docking model generated using Schrodinger covalent docking software predicts that probe 15 fits in a large pocket and forms a covalent bond with the side chain of C110 of PTGES2 (Figure S5). In conclusion, we have developed a novel condensation reaction to prepare 2-chloromethylquinoline, a novel electrophile, and explored its reactivity in the proteome. These chloromethylquinolines covalently bind to diverse cellular protein targets, including GSTO1, HMOX2, and PTGES2. Our results thus suggest that 2-CMQs have target profiles rather distinct from those of common electrophiles such as chloroacetamide and acrylamide. The fact that 2-CMQs appear to have a half-life shorter than that of acrylamide (Table S1)

Figure 2. Series of 2-chloromethylquinolines and the proteome labeling patterns for eight of them. (A) Probes tested in protein labeling. (B) In-gel fluorescence images of HEK293H proteomes labeled by a portion of the 2-CMQ probes. Bands A−C represent the major targets labeled by the CMQ probes.

Figure 3. Binding of 2-CMQ to the prominent cellular targets is saturable, and the site of binding was mapped for one target. (A) Competition of probe 15 by a competitor to reveal binding affinity in cells: (top) in-gel fluorescence image of probe-labeled HEK293H proteomes and (bottom) colloidal blue staining image. (B) Identification of the covalent interacting residue in PTGES2 via mutagenesis: (top) in-gel fluorescence image of probe-labeled HEK293H proteomes and (bottom) immunoblots for FLAG-tagged PTGES2.

concentrations of the competitor compound, suggesting that the labeling is saturable. IC50 values derived from the dose− response curve signify the concentration at which 50% of the target protein is occupied by the probe in situ (Figure S1). Our 2717

DOI: 10.1021/acs.biochem.9b00359 Biochemistry 2019, 58, 2715−2719

Communication

Biochemistry

(2) Evans, M. J., Saghatelian, A., Sorensen, E. J., and Cravatt, B. F. (2005) Target discovery in small-molecule cell-based screens by in situ proteome reactivity profiling. Nat. Biotechnol. 23, 1303−7. (3) Visscher, M., Arkin, M. R., and Dansen, T. B. (2016) Covalent targeting of acquired cysteines in cancer. Curr. Opin. Chem. Biol. 30, 61−67. (4) Zhao, Z., Liu, Q., Bliven, S., Xie, L., and Bourne, P. E. (2017) Determining Cysteines Available for Covalent Inhibition Across the Human Kinome. J. Med. Chem. 60, 2879−2889. (5) Li, D., Ambrogio, L., Shimamura, T., Kubo, S., Takahashi, M., Chirieac, L. R., Padera, R. F., Shapiro, G. I., Baum, A., Himmelsbach, F., Rettig, W. J., Meyerson, M., Solca, F., Greulich, H., and Wong, K. K. (2008) BIBW2992, an irreversible EGFR/HER2 inhibitor highly effective in preclinical lung cancer models. Oncogene 27, 4702−11. (6) Mathiak, G., Grass, G., Herzmann, T., Luebke, T., Zetina, C. C., Boehm, S. A., Bohlen, H., Neville, L. F., and Hoelscher, A. H. (2000) Caspase-1-inhibitor ac-YVAD-cmk reduces LPS-lethality in rats without affecting haematology or cytokine responses. Br. J. Pharmacol. 131, 383−6. (7) Kung, A., Chen, Y. C., Schimpl, M., Ni, F., Zhu, J., Turner, M., Molina, H., Overman, R., and Zhang, C. (2016) Development of Specific, Irreversible Inhibitors for a Receptor Tyrosine Kinase EphB3. J. Am. Chem. Soc. 138, 10554−60. (8) Speers, A. E., and Cravatt, B. F. (2004) Profiling enzyme activities in vivo using click chemistry methods. Chem. Biol. 11, 535− 46. (9) Chen, Y. C., Backus, K. M., Merkulova, M., Yang, C., Brown, D., Cravatt, B. F., and Zhang, C. (2017) Covalent Modulators of the Vacuolar ATPase. J. Am. Chem. Soc. 139, 639−642. (10) Liu, Q., Sabnis, Y., Zhao, Z., Zhang, T., Buhrlage, S. J., Jones, L. H., and Gray, N. S. (2013) Developing irreversible inhibitors of the protein kinase cysteinome. Chem. Biol. 20, 146−59. (11) Lonsdale, R., Burgess, J., Colclough, N., Davies, N. L., Lenz, E. M., Orton, A. L., and Ward, R. A. (2017) Expanding the Armory: Predicting and Tuning Covalent Warhead Reactivity. J. Chem. Inf. Model. 57, 3124−3137. (12) White, J. D., Yager, K. M., and Stappenbeck, F. (1993) Synthesis and Conformation of 2-[[3-(1-Hydroxyhexyl)Phenoxy]Methyl]Quino-Line, a 5-Lipoxygenase Inhibitor and Leukotriene Antagonist. J. Org. Chem. 58, 3466−3468. (13) Stock, N. S., Bain, G., Zunic, J., Li, Y., Ziff, J., Roppe, J., Santini, A., Darlington, J., Prodanovich, P., King, C. D., Baccei, C., Lee, C., Rong, H., Chapman, C., Broadhead, A., Lorrain, D., Correa, L., Hutchinson, J. H., Evans, J. F., and Prasit, P. (2011) 5-Lipoxygenaseactivating protein (FLAP) inhibitors. Part 4: development of 3-[3tert-butylsulfanyl-1-[4-(6-ethoxypyridin-3-yl)benzyl]-5-(5-methylpyridin-2-y lmethoxy)-1H-indol-2-yl]-2,2-dimethylpropionic acid (AM803), a potent, oral, once daily FLAP inhibitor. J. Med. Chem. 54, 8013−29. (14) Kuchar, M., Kmonicek, V., Panajotova, V., Jandera, A., Brunova, B., Junek, R., Bucharova, V., Cejka, J., and Satinsky, D. (2004) Derivatives of (phenylsulfanyl)benzoic acids with multiple antileukotrienic activity. Collect. Czech. Chem. Commun. 69, 2098−2120. (15) Xie, Y. Y., and Li, L. H. (2014) Microwave-assisted alphahalogenation of 2-methylquinolines with tetrabutylammonium iodide and 1,2-dichloroethane (1,2-dibromoethane). Tetrahedron Lett. 55, 3892−3895. (16) Braun, J. v., and Heymons, A. (1930) Imid- und Amidchloride nicht-aromatischer Säuren (V. Mitteil.). Ber. Dtsch. Chem. Ges. B 63, 502−507. (17) Weerapana, E., Wang, C., Simon, G. M., Richter, F., Khare, S., Dillon, M. B., Bachovchin, D. A., Mowen, K., Baker, D., and Cravatt, B. F. (2010) Quantitative reactivity profiling predicts functional cysteines in proteomes. Nature 468, 790−5. (18) Backus, K. M., Correia, B. E., Lum, K. M., Forli, S., Horning, B. D., González-Páez, G. E., Chatterjee, S., Lanning, B. R., Teijaro, J. R., Olson, A. J., Wolan, D. W., and Cravatt, B. F. (2016) Proteome-wide covalent ligand discovery in native biological systems. Nature 534, 570−4.

suggests that they are unlikely to be used as drugs in the clinic. However, they can still serve as powerful tool compounds for identifying protein targets susceptible to covalent modulation in basic research. In view of the role of PTGES2 in producing inflammatory molecules such as prostaglandin E2, the 2-CMQbased inhibitors that we discovered here have the potential to serve as novel anti-inflammatory agents. Beyond this case study, our methodology could be extended to create additional electrophilic probes for discovering covalent ligands across the proteome.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.9b00359. Tables S1−S3, Figures S1−S5, Methods, spectra, optimized quantum mechanical energies and geometries, full gel images for Figures 2B, 3A, 3B, and S3, and additional references (PDF) Tables of complete data and top hits (XLSX) Accession Codes

PTGES2, Q9H7Z7; HMOX, P30519; GSTO1, P78417.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

K. N. Houk: 0000-0002-8387-5261 Chao Zhang: 0000-0003-0251-8156 Present Address ⊥

F.N.: Institute of Drug Discovery Technology, Ningbo University, Ningbo, Zhejiang 315211, China.

Author Contributions ∥

F.N. and A.E. contributed equally to this work.

Funding

This work was supported by the University of Southern California, the National Science Foundation (CHE-1455306 and CHE-1764328), and the National Institutes of Health (F32GM122218). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Dr. N. A. Graham and Mr. A. Delfarah (University of Southern California) for their generous help with high-resolution mass spectrometry. Computational resources were provided by the UCLA Institute for Digital Research and Education (IDRE).

■

ABBREVIATIONS Cys, cysteine; FDA, U.S. Food and Drug Administration; CMQ, chloromethylqunoline; PTGES2, prostaglandin E synthase 2.

■

REFERENCES

(1) Potashman, M. H., and Duggan, M. E. (2009) Covalent modifiers: an orthogonal approach to drug design. J. Med. Chem. 52, 1231−46. 2718

DOI: 10.1021/acs.biochem.9b00359 Biochemistry 2019, 58, 2715−2719

Communication

Biochemistry (19) Lanning, B. R., Whitby, L. R., Dix, M. M., Douhan, J., Gilbert, A. M., Hett, E. C., Johnson, T. O., Joslyn, C., Kath, J. C., Niessen, S., Roberts, L. R., Schnute, M. E., Wang, C., Hulce, J. J., Wei, B., Whiteley, L. O., Hayward, M. M., and Cravatt, B. F. (2014) A road map to evaluate the proteome-wide selectivity of covalent kinase inhibitors. Nat. Chem. Biol. 10, 760−767. (20) Tanikawa, N., Ohmiya, Y., Ohkubo, H., Hashimoto, K., Kangawa, K., Kojima, M., Ito, S., and Watanabe, K. (2002) Identification and characterization of a novel type of membraneassociated prostaglandin E synthase. Biochem. Biophys. Res. Commun. 291, 884−9. (21) Watanabe, K., Ohmiya, Y., Ohkubo, H., Tanikawa, N., Kojima, M., and Ito, S. (2003) A novel type of membrane-associated prostaglandin E synthase. Adv. Exp. Med. Biol. 525, 107−11. (22) Murakami, M., Nakashima, K., Kamei, D., Masuda, S., Ishikawa, Y., Ishii, T., Ohmiya, Y., Watanabe, K., and Kudo, I. (2003) Cellular prostaglandin E2 production by membrane-bound prostaglandin E synthase-2 via both cyclooxygenases-1 and −2. J. Biol. Chem. 278, 37937−47. (23) Watanabe, K., Ohkubo, H., Niwa, H., Tanikawa, N., Koda, N., Ito, S., and Ohmiya, Y. (2003) Essential 110Cys in active site of membrane-associated prostaglandin E synthase-2. Biochem. Biophys. Res. Commun. 306, 577−81.

2719

DOI: 10.1021/acs.biochem.9b00359 Biochemistry 2019, 58, 2715−2719