Site-specific activation of the proton pump inhibitor rabeprazole by tetrathiolate zinc centres

Abstract

Proton pump inhibitors have become top-selling drugs worldwide. Serendipitously discovered as prodrugs that are activated by protonation in acidic environments, proton pump inhibitors inhibit stomach acid secretion by covalently modifying the gastric proton pump. Despite their widespread use, alternative activation mechanisms and potential target proteins in non-acidic environments remain poorly understood. Employing a chemoproteomic approach, we found that the proton pump inhibitor rabeprazole selectively forms covalent conjugates with zinc-binding proteins. Focusing on DENR, a protein with a C4 zinc cluster (that is, zinc coordinated by four cysteines), we show that rabeprazole is activated by the zinc ion and subsequently conjugated to zinc-coordinating cysteines. Our results suggest that drug binding, activation and conjugation take place rapidly within the zinc coordination sphere. Finally, we provide evidence that other proton pump inhibitors can be activated in the same way. We conclude that zinc acts as a Lewis acid, obviating the need for low pH, to promote the activation and conjugation of proton pump inhibitors in non-acidic environments.

Fig. 1. Identification of rabeprazole target proteins.

a, Backbone structure of PPIs (1) and their conventional mechanism of protonation-mediated activation (1–5) and target protein conjugation (6). The 2-substituted pyridyl moiety is shown in green, the benzimidazolyl moiety is black and the sulfinyl group is purple. b, The chemoproteomic approach used to isolate (left branch) or fluorescently label (right branch) rabeprazole-reactive proteins. SPAAC, strain-promoted azide–alkyne cycloaddition; SDS-PAGE, SDS polyacrylamide gel electrophoresis. c, Identification of Rabazi-conjugated proteins isolated from HEK293 MSR cells treated with three different concentrations of Rabazi (left, 10 μM; middle, 1 μM; right, 0.1 μM). An empirical Bayes moderated t-test from the limma package was used to test for differential abundance. Significant hits are coloured purple. Proteins with at least one zinc-binding site are labelled by name. d, Enrichment of STRING database categories for target proteins identified from HEK293 MSR cells, relative to the total proteome. Based on a hypergeometric test using the Benjamini–Hochberg method to correct P values for multiple testing. e, Identification of Rabazi-conjugated proteins isolated from PACO17 cells treated with 10 μM Rabazi. An empirical Bayes moderated t-test from the limma package was used to test for differential abundance. Significant hits are coloured dark blue. Proteins with at least one zinc-binding site are labelled by name. f, Enrichment of STRING database categories for target proteins identified from PACO17 cells, relative to the total proteome. Based on a hypergeometric test using the Benjamini–Hochberg method to correct P values for multiple testing. g, Structure of DENR–MCTS1 complex (Protein Data Bank (PDB) no. 6MS4) with the DENR C4 zinc-binding site expanded. Source data

Fig. 2. Intracellular rabeprazole conjugation depends on Zn²⁺ binding.

a, Workflow for analysing the intracellular conjugation of Rabazi to SBP-tagged DENR expressed in HEK293 MSR cells. b, Rabazi conjugates to SBP-DENR in intact cells. HEK293 MSR cells transfected with SBP-DENR wild type (WT) or empty vector (EV) were treated for 15 min with 10 µM Rabazi or Rabazi thioether (Rabazi-ΔO). The lower band (25 kDa) on the Coomassie gel represents (crosslinked) streptavidin dimers released from beads. Representative blot for n > 7 independent experiments. c, Replacement of zinc-coordinating cysteines by alanines in SBP-DENR* diminishes Rabazi conjugation (left panel). Quantitation of the Cy5 signal (right panel). HEK293 MSR cells ectopically expressing SBP-DENR* or mutants were treated for 15 min with 10 µM Rabazi or Rabazi thioether (Rabazi-ΔO). Data are presented as mean ± s.d., based on n = 3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001; P = 0.0022, 0.0001, 0.0002 and 0.012, based on a two-tailed paired t-test. d, Replacement of C44, but not of C53, by histidine in SBP-DENR* leads to the loss of Rabazi conjugation (left panel). Quantitation of the Cy5 signal (right panel). HEK293 MSR cells ectopically expressing SBP-DENR* or mutants were treated for 15 min with 10 µM Rabazi or Rabazi thioether (Rabazi-ΔO). Data are represented as mean ± s.d., based on n = 4 independent experiments. NS, not significant; **P < 0.01, ***P < 0.001; P = 0.0001, 0.0001, 0.0017 and 0.1753, based on a two-tailed paired t-test. Source data

Fig. 3. Conjugation of Rabazi to recombinant DENR depends on Zn²⁺ binding.

a, The thiol alkylating agent NEM improves the detection of Rabazi-conjugated recombinant WT DENR in the Cy5-based gel assay. NEM does not interfere with Rabazi conjugation even when applied at very high concentration (100 mM). Representative blot of n = 2 independent experiments. b, Conjugation of Rabazi to recombinant DENR is insensitive to EDTA (2.5 mM), diminished by the stronger metal chelator TPEN (2.5 mM) and completely abolished by the combination of EDTA (2.5 mM) and SDS (1%). The experiment was conducted in the presence of NEM. Representative blot of n = 2 independent experiments. c, Workflow for assessing the interaction between Rabazi and the recombinant His-DENR–MCTS1 complex. d, Mutation of the zinc-coordinating cysteines in DENR diminishes Rabazi conjugation, with the exception of C53A (left panel). Quantitation of the Cy5 signal (right panel). Data are represented as mean ± s.d., based on n = 4 (C34A, n = 3) independent experiments. **P < 0.01, ***P < 0.001; NS, not significant; P = 0.0018, 0.0003, 0.0001, 0.2494, 0.0012, 0.0002 and 0.0002, based on a two-tailed paired t-test. e, Recombinant WT DENR conjugates to Rabazi more rapidly than DENR(C34A). Representative blot of n = 2 independent experiments. Source data

Fig. 4. Rabeprazole conjugates to zinc-coordinating cysteines 44 and 34.

a, Representative MS/MS spectrum of DENR peptide 31–52, demonstrating the presence of hydrolysed rabeprazole on C44. The corresponding list of fragment matches is provided in Extended Data Fig. 4d. a.u., arbitrary units; CAM, carbamidomethyl; RT, retention time; m/z, mass-to-charge ratio. b–d, Differential alkylation suggests C44 as the preferred conjugation site of rabeprazole (RAB). b, Workflow used to identify the preferred target cysteine of rabeprazole on DENR*. c, Analysing tryptic peptide 31–52, containing C34, C37 and C44, the fully alkylated form is predominantly found when reduction (Red) precedes alkylation (Alk). d, By contrast, the doubly alkylated form (lacking alkylation on C44) is predominantly found when alkylation (Alk) precedes reduction (Red). Data in bar graphs are represented as mean ± s.d., based on n = 2 independent experiments with 2 and 3 technical replicates. Source data

Fig. 5. Rabeprazole causes major conformational changes in DENR.

a, Time (t) series of ¹H spectra of rabeprazole in buffer, in the absence (left) and presence (right) of the ¹⁵N-labelled DENR^26–98–MCTS1 complex. b, Time series of ¹H spectra of rabeprazole thioether (rabeprazole-ΔO) in buffer, in the absence (left) and presence (right) of the ¹⁵N-labelled DENR^26–98–MCTS1 complex. c, A 2D ¹H–¹⁵N HSQC spectrum of the ¹⁵N-labelled DENR^26–98–MCTS1 complex in the presence of rabeprazole (RAB) collected over time. d, A 2D ¹H–¹⁵N HSQC spectrum of the ¹⁵N-labelled DENR^26–98–MCTS1 complex in the presence of rabeprazole thioether (RAB-ΔO) collected over time. e, Hydrophobic interactions promote the initial coordination of rabeprazole to the zinc cluster. The left shows the structure of the DENR–MCTS1–rabeprazole complex with four possible rabeprazole configurations (highlighted in transparent colours) aligned on the zinc cluster. The inset shows one rabeprazole configuration obtained from MD simulations together with interacting DENR residues. The right shows DENR residues interacting with rabeprazole based on MD simulations of four initial rabeprazole configurations. For every trajectory snapshot, residues with an atom–atom distance below 3.5 Å were classified as interacting. Only snapshots with reactive configurations (pyridine N to benzimidazole C distance within 20% of the sum of van der Waals radii) were considered. Data are presented as mean ± s.e.m., based on n = 3 simulations of four different configurations. Source data

Fig. 6. All PPI variants react with the DENR C4 zinc finger.

a, Release of Zn²⁺ ions from His-DENR* (7.5 µM), in response to a 50-fold molar excess of rabeprazole or rabeprazole thioether (rabeprazole-ΔO), as monitored by chelation with 4-(2-pyridylazo)resorcinol (PAR; 100 µM). Based on the mean of n = 3 technical replicates. Shading indicates s.d. b, Same experiment as in a, testing the release of Zn²⁺ ions from DENR in response to all six FDA-approved PPIs. Based on the mean of n = 3 technical replicates. Shading indicates s.d. Source data

Fig. 7. Proposed mechanism for the reaction of a PPI with a C4 zinc finger.

The possible pathways promoting PPI conjugation to a C4 zinc cluster cysteine are shown. Following binding and activation (11–15), the PPI may either conjugate to the cysteine that first dissociated from the zinc (16; path a); conjugate to a zinc-coordinated cysteine (17; path b); or produce a sulfenamide intermediate, which can react with any nearby cysteine (18, path c). The results obtained in this work support path b.

Extended Data Fig. 1. Identification of rabeprazole target proteins.

(a) The target proteins (n = 4, n = 21 and n = 24) identified with three different concentrations of Rabazi (0.1, 1 and 10 μM) completely overlap as hierarchical nested sets. (b) Zinc-binding motifs of the nine most significantly enriched target proteins following treatment of cells with 1 μM Rabazi (logFC > 1.5 and FDR < 0.001). Cysteines belonging to a C4 zinc cluster are both highlighted and boxed. Cysteines and histidines belonging to other types of zinc clusters are highlighted without boxing.

Extended Data Fig. 2. Conjugation of Rabazi to recombinant DENR depends on zinc binding.

(a) Intact protein MS of recombinant DENR* following incubation with a 5-fold molar excess of rabeprazole for 15 min at r.t. The observed mass shift (–4 Da) indicates the formation of two intramolecular disulfide bonds in the presence, but not in the absence of rabeprazole. (b) Quantification of Zn²⁺ bound to recombinant His-DENR:MCTS1 and respective cysteine mutants. Zn²⁺ was released from 10 µM recombinant protein by incubation with 4 M guanidinium chloride (GdnHCl) and 100 mM NEM. PAR was added to a final concentration of 100 µM and formation of the zinc-PAR complex was quantified spectrophotometrically. Data are represented as mean ± s.d., based on n = 4 (C37A, C44A, C53A, C44H, C53H) or n = 5 (DENR*, C34A, ΔZF) independent experiments with n = 3 technical replicates each. (c) Zinc occupancy of His-DENR:MCTS1 and respective mutants as determined by PAR and ICP-OES. Data are represented as mean ± s.d., based on n = 3 (C34A: n = 2) technical replicates each. Source data

Extended Data Fig. 3. The C4 zinc cluster tends to disengage one of its ligands.

(a) Intact protein MS of recombinant DENR* (5 μM) following incubation with different concentrations of NEM (0, 0.5, 2.5 and 5 mM) for 15 min at r.t. (b) Release of Zn²⁺ ions from His-DENR* (7.5 µM), induced by a 100-fold molar excess of NEM, as monitored by chelation with 4-(2-pyridylazo)resorcinol (PAR; 100 µM). The 15 min time point relevant for comparison with (a) is highlighted by a dashed vertical line. Based on the mean ± s.d. of n = 3 technical replicates. Source data

Extended Data Fig. 4. Rabeprazole conjugates to zinc-coordinating cysteines 44 and 34.

(a) Intact protein MS of recombinant DENR*-MCTS1 complex incubated with a 5-fold molar excess of rabeprazole for 15 min. DENR predominantly conjugates one molecule of rabeprazole. (b) Hydrolysis of protein rabeprazole conjugates: the electrophilic central carbon atom of the benzimidazolyl group is susceptible to nucleophilic attack by water leading to loss of benzimidazolone. (c) List of fragment matches from scan number 24412 (XCorr = 5.14) indicating the presence of hydrolyzed rabeprazole on C34. Identified y-ions are marked blue and b-ions are marked red. A consecutive series of y-ions and b-ions cover the peptide sequence from the C- and N-terminal end, respectively. (d) List of fragment matches from scan number 24206 (XCorr = 3.51) indicating the presence of hydrolyzed rabeprazole on C44. Identified y-ions are marked blue and b-ions are marked red. A consecutive series of y-ions and b-ions cover the peptide sequence from the C- and N-terminal end, respectively. (e) List of fragment matches from scan number 24585 (XCorr = 3.42) indicating presence of non-hydrolyzed rabeprazole on C34. Identified y-ions are marked blue and b-ions are marked red. A consecutive series of y-ions covers the peptide sequence from the C-terminal end. (f) Accessibility of the DENR zinc cluster after dissociation of C53. Left panel: Surface rendering. Only the sulfur atoms of C34 and C44 are directly accessible on the surface. Right panel: Sectional view through the volume model of DENR, with the clipping plane just above the zinc cluster. The sulfur atom of C37 can be seen to be buried inside the protein. Source data

Extended Data Fig. 5. Rabeprazole causes major conformational changes in DENR.

(a) Complete 2D ¹H-¹⁵N HSQC spectrum for the experiment shown in main Figs. 5b and 5d. (b) Distance between the pyridine nitrogen and the benzimidazole carbon. MD simulations of the DENR-MCTS1-rabeprazole complex with either (R)- (upper panels) or (S)-rabeprazole (lower panels), with the benzimidazole nitrogen coordinated between C34 and C37 (coordination geometry I, left panels) or between C37 and C44 (coordination geometry II, right panels). Individual trajectories are shown for every initial configuration (n = 3 technical replicates). Source data

Extended Data Fig. 6. Structural organization of the DENR zinc-binding site.

(a) Topology of the DENR zinc-binding site. Blue: C34-C37 loop; Red: C37-C44 loop; Green: C44-C53 loop. The loop lengths (n) are indicated. (b) Structure of the zinc- binding site of DENR (based on PDB entry 6MS4) not including the C-terminal part beyond P48, to illustrate the appearance of the cluster after dissociation of C53. The DENR binding partner MCTS1 is shown in the background (shaded in gray). Please also see the corresponding surface rendering (Extended Data Fig. 4f), showing differences in the surface exposure of the sulfur atoms.