Identification of a Target Site for Covalent Inhibition of Protein Phosphohistidine Phosphatase 1

Abstract

Despite the recent discovery of numerous phosphohistidine (pHis) sites in mammalian proteomes, the functions of this labile post-translational modification (PTM) mostly remain unknown. Phosphohistidine phosphatase 1 (PHPT1), one of the few known protein pHis phosphatases, regulates important cellular processes, and its genetic knockdown attenuated cancer cell proliferation and a liver fibrosis model. Unfortunately, the lack of PHPT1 inhibitors has limited further understanding and the therapeutic potential of this unique enzyme. We report that PHPT1 can be covalently inhibited by targeting Cys73, a residue that is nonessential for the enzyme activity. We also determined the inhibition kinetics of various small molecule electrophiles as potential warheads against PHPT1. Our results lay a foundation for the development of more potent and specific PHPT1 inhibitors.

The phosphorylation of histidine (His) is a post-translational modification (PTM) much less explored than the O-phosphorylation of serine (Ser), threonine (Thr), and tyrosine (Tyr) owing to the chemical lability of the phosphohistidine (pHis) product.^1,2 In prokaryotes and plants, the histidine kinase is a critical component of the two-component systems (TCS), which sense environmental changes and control the adaptation responses such as antibiotic resistance and biofilm formation.³

In eukaryotes, this modification has been related to cell metabolism, signal transduction, and epigenetic changes; in addition, recent advances in phosphoproteomics have revealed numerous pHis sites in the human proteome.^4,5 Nucleoside diphosphate kinase (NDPK) isoforms, also known as NMEs, are well documented to have protein histidine kinase activities and to be involved in cancer metastasis and other essential processes such as G-protein signaling.⁶ For pHis-specific phosphatases, the following three enzymes are known: phosphohistidine phosphatase 1 (PHPT1), phospholysine phosphohistidine inorganic pyrophosphate phosphatase (LHPP),⁷ and phosphoglycerate mutase family member 5 (PGAM5).⁸

PHPT1, also known as PHP14, is the first cloned pHis-specific phosphatase, discovered in 2002.^9,10 It can dephosphorylate pHis in various proteins, including the KCa3.1and TRPV5 ion channels, ATP citrate lyase (ACLY), and the β-subunit of the heterotrimeric G protein, to regulate their functions.¹¹ Interestingly, these substrates are histidine-phosphorylated by NDPK, suggesting the opposite effect of PHPT1 on NDPK-mediated pathways (Figure 1a). Physiologically, PHPT1 has been linked to cytoskeletal reorganization, cell proliferation, hepatic stellate cell (HSC) migration, and adipocyte differentiation.¹²⁻¹⁶ Promisingly, the knockdown of PHPT1 led to significant attenuation of cell proliferation and liver fibrosis in mice, as well as promotion of brown adipocyte differentiation.¹³⁻¹⁶

Figure 1.

(a) Schematic of the function of phosphohistidine phosphatase 1 (PHPT1) and nucleoside diphosphate kinase (NDPK). (b) Catalytic mechanism of PHPT1. (c) Summary of PHPT1 inhibitor screening (including Z′ score). (d) Structure and IC₅₀ of ethacrynic acid (EA) under our standard assay conditions (n = 9).

Consequently, PHPT1 inhibition has been proposed as a potential therapeutic strategy against cancer, liver fibrosis, and obesity. As part of our long-standing interest in histidine phosphorylation,¹⁷ we previously reported a fluorescent probe for PHPT1 activity.¹⁸ This probe was successfully utilized for in vitro biochemical characterization and the specific detection of PHPT1 activity in cell lysates.

Unfortunately, the dearth of PHPT1 inhibitors has limited further exploration of the enzyme’s biological roles and therapeutic potential. The enzyme is inhibited by divalent zinc but is resistant to various broad-spectrum phosphatase inhibitors, including sodium vanadate and fluoride.¹⁹ As shown by a mechanistic model,²⁰ PHPT1 utilizes a histidine base to activate a water nucleophile to attack the substrate’s phosphoryl group (Figure 1b). This unique mechanism is consistent with the enzyme’s inertness toward many phosphatase inhibitors. Although the 3D structure of the enzyme is available, the shallow active site makes it hard to design high-affinity small-molecule ligands for PHPT1. Thus, we set out to find PHPT1 inhibitors through unbiased high-throughput screening. A pioneering study by Barrios and colleagues also reported the first small molecule inhibitors of PHPT1 through a library screening.²¹

Our high-throughput screening for PHPT1 inhibitors utilized a colorimetric assay with p-nitrophenyl phosphate (pNPP).²⁰ We screened 9265 compounds from the Korea Chemical Bank, and the assay behaved well (Figure 1c and Supplementary Table 3). Initially, 22 inhibitor hits (>30% inhibition at a single dose of 50 μM) were identified, but most of them did not significantly inhibit PHPT1 in a subsequent dose-dependent assay (Supplementary Table 4). To our delight, ethacrynic acid (EA) successfully inhibited PHPT1 with an IC₅₀ of 18.3 μM in our standard assay (Figure 1d). EA is clinically used as a diuretic and inhibits various enzymes, such as Na⁺K⁺-ATPase and glutathione-S-transferase (GST).²² As a Michael acceptor, it is reported to covalently modify many proteins, including MAP2K6 kinase, LEF-1,²³ and adenine nucleotide translocases.²⁴

Considering the existing literature, we tested if EA is a covalent inhibitor of PHPT1. In the pNPP assay, the reaction rate decreased over time, and longer preincubation of EA with PHPT1 resulted in more significant inhibition of PHPT1 (Figure 2a), implicating covalent inhibition. The same trend was observed for the dephosphorylation of histidine-phosphorylated BSA (Supplementary Figure 3). Electrospray ionization mass spectrometry (ESI-MS) analysis confirmed that the mass of PHPT1 increased by 303 Da when incubated with EA (Figure 2b). These data collectively indicate the covalent inhibition of PHPT1 by EA, potentially through Michael addition.

Figure 2.

(a) Hydrolysis of p-nitrophenyl phosphate (pNPP) by PHPT1. The preincubation of EA attenuated PHPT1 activity (n = 3). (b) ESI-MS spectra of native and EA-labeled PHPT1. (c) Effect of EA on the reaction rate of wild-type (WT) and mutant PHPT1 (n = 3, ns: not significant). (d) Tandem mass spectrometry (MS/MS) spectrum of the tryptic fragment (QGDCECLGGGR) from the EA-PHPT1 adducts. Only Cys73 was modified with EA.

Subsequently, we investigated the EA-modification site of PHPT1. PHPT1 has three cysteines, and we mutated each cysteine into an alanine (PHPT1 C69A, C71A, and C73A). Previously, Klumpp et al. showed that none of these cysteines was required for the phosphatase activity,²⁵ and we also observed this effect. Of these single Cys mutants, only the C73A mutant was inert to EA (Figure 2c and Supplementary Figure 4). ESI-MS data also showed that C69A and C71A mutants formed covalent EA adducts, unlike the C73A mutant (Supplementary Figures 5–12). Finally, we performed an LC-MS/MS analysis of the EA-PHPT1 adduct following Ye’s method (Supplementary Figure 13).²⁵ The tryptic fragment around Cys73 indeed confirmed the EA adduct formation at this cysteine (Figure 2d). These data show that Cys73 is the covalent modification site for EA.

Due to its diverse intracellular targets,²²⁻²⁴ however, EA itself is unlikely a specific inhibitor for PHPT1. Therefore, we also evaluated other types of Cys-reactive electrophiles as potential “warheads” for next-generation covalent PHPT1 inhibitors. N-ethylmaleimide (NEM), a Michael acceptor, also inhibited PHPT1 (Supplementary Figure 14). However, other commonly used cysteine-targeting electrophiles, such as iodoacetamide, 2-iodo-N-phenylacetamide, and methyl p-toluenesulfonate, did not significantly inhibit PHPT1 (Supplementary Figure 15), indicating the distinct reactivity pattern of Cys73 toward different classes of electrophiles. Glutathione disulfide (GSSG) was also inactive toward PHPT1. A recent comparative proteomic study on reactive cysteinomes also found that cysteines in PHPT1 only reacted with a Michael acceptor probe but not with an iodoacetamide-based probe.²⁶ This reactivity pattern could be exploited to develop more potent and specific PHPT1 inhibitors.

To quantitatively evaluate the inhibition efficiency, the inhibitors’ K_i and k_inact values were determined (Table 1 and Supplementary Figure 14).^27,28 Compared with norstictic acid, a recently covalent inhibitor of PHPT1,²¹ EA and NEM inhibited PHPT1 less efficiently (Table 1). However, norstictic acid’s target site on PHPT1 is unknown, although it likely binds to a lysine residue with its aldehyde functionality.

Table 1. PHPT1 Inhibition Kinetics of Ethacrynic Acid and Other Electrophiles (n = 3)^a.

inhibitor	kinact (s–1)	Ki (μM)	kinact/Ki (M–1 s–1)
ethacrynic acid	2.67 ± 0.02 × 10–3	183 ± 43	1.46 ± 0.37 × 101
N-ethylmaleimide	7.5 ± 1.6 × 10–3	209 ± 48	3.1 ± 1.1 × 101
iodoacetamide	N. I.	N. I.	N. I.
2-iodo-N-phenylacetamide	N. I.	N. I.	N. I.
methyl p-toluenesulfonate	N. I.	N. I.	N. I.
GSSG	N. I.	N. I.	N. I.
norstitic acid21	2.8 ± 0.2 × 10–2	90 ± 20	3.1 ± 1.0 × 102

^aN. I. = No appreciable inhibition after 1 h of incubation with the electrophile (25 μM or higher).

This study identified Cys73 of PHPT1 as a target site for electrophilic covalent inhibitors. Many phosphatases, such as protein tyrosine phosphatases, feature a nucleophilic Cys as the essential catalytic residue targeted by covalent inhibitors.²⁹ In contrast, all of human PHPT1’s cysteines (Cys69, Cys71, and Cys73) are dispensable for the phosphatase activity,²⁶ and covalent masking of a cysteine was not expected to inhibit the enzyme significantly. However, we note that Cys73 is only approximately 8 Å apart from the catalytic His53, so the binding of bulky electrophiles to this site might disrupt substrate binding or catalysis (Figure 3d).

Figure 3.

3D structure of PHPT1 (adapted from PDB: 2HW4). The active site His53 (orange) and Cys73 (green) are about 8 Å apart, without direct contact.

While EA can covalently inhibit PHPT1, EA itself is unlikely to be specific in vivo, considering EA’s diverse intracellular targets. Recently reported PHPT1 inhibitors demonstrated some specificity over other phosphatases, although their binding sites on PHPT1 are still unknown.²¹ If the binding sites are close to Cys73, it may be possible to reconfigure those inhibitors as more potent and selective covalent inhibitors of PHPT1 by linking an electrophilic warhead to target the cysteine.

Despite PHPT1’s unique biochemical activities and therapeutic potential, the lack of suitable inhibitors has hindered further investigations of its functions and therapeutic potential. Our study should be a starting point for developing next-generation PHPT1 inhibitors.

Glossary

Abbreviations

PHPT1

protein histidine phosphatase 1

NDPK

nucleoside diphosphate kinase

EA

ethacrynic acid

pNPP

para-nitrophenyl phosphate

BSA

bovine serum albumin

NEM

N-ethyl maleimide

GSSG

glutathione disulfide

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.2c00450.

• Figures S1–S15 and Tables S1–S4; detailed experimental procedures and characterization data; mass spectrometric characterization of the PHPT1-EA adduct; kinetic analysis of the covalent inhibition (PDF)

Author Contributions

^† H.J.K. and H.J. contributed equally to this work. J.-M.K. designed and conceptualized the project. H.J.K., H.J., and S.K. conducted experiments and analyzed data under the guidance of J.K.S. and J.-M.K. All authors contributed to the writing and review of the manuscript.

The authors declare no competing financial interest.