Probing the Influence of the Protein Scaffold on H-Cluster Reactivity via Gain-of-Function Studies—Improved H₂ Evolution and O₂ Tolerance through Rational Design of [FeFe] Hydrogenase

Abstract

[FeFe] hydrogenases make up a structurally diverse family of metalloenzymes that catalyze proton/dihydrogen interconversion. They can be classified into phylogenetically distinct groups denoted A–G, which differ in structure and reactivity. Prototypical Group A hydrogenases have high turnover rates and remarkable energy efficiency. As compared to Group A enzymes, the putatively sensory Group D hydrogenase from Thermoanaerobacter mathranii (TamHydS) has a thousand-fold lower H₂ evolution rate and a high overpotential requirement to drive catalysis (irreversible) but shows increased inhibitor tolerance. This divergence in structure and activity between hydrogenases makes them ideal models for studying second (active-site environment) and outer (e.g., substrate transport) coordination sphere effects on metal cofactors. Herein, we generated three TamHydS-based variants, each mimicking proposed key structural features of Group A hydrogenase: the “active site” (AS), “proton-transfer pathway” (PTP), and “combined” (CM = AS + PTP) variant. A fourth single-point variant, A137C, which introduces a proposed critical cysteine in the active site, was characterized as a reference. No change in isolation resulted in Group A-like behavior; i.e., no positive impact on catalytic performance was observed. The CM variant, however, showed increased H₂ evolution activity but retained the overpotential requirement. Additionally, the CM variant improved the already relatively high stability of TamHydS against O₂ and CO inhibition. These findings show that activity rates, (ir)reversibility, and susceptibility to gaseous inhibitors are decoupled. Moreover, the results highlight the importance of exploring hydrogenase diversity as a path toward understanding the structural factors that enable the outstanding catalytic properties of [FeFe] hydrogenases.

Introduction

Hydrogenases are metalloenzymes serving central functions in energy metabolism, where they catalyze the interconversion of protons (H⁺), electrons, and hydrogen gas (H₂).^1,2 The so-called [FeFe] hydrogenases are generally considered as the most active, with turnover frequencies (TOFs) of 10⁴–10⁵ s^–1 reported for H₂ evolution and oxidation.^3,4 Remarkably, the enzyme can sustain these rates very close to the thermodynamic potential of the reaction. Consequently, [FeFe] hydrogenases have attracted considerable attention in the context of renewable energy research and serve as an inspiration for the design of synthetic H₂-evolving catalysts.^5,6

A unifying feature of all [FeFe] hydrogenases is their organometallic active-site cofactor, the “H-cluster” (Figure 1A). This biologically unique cofactor consists of a canonical iron–sulfur cluster ([4Fe-4S]_H) linked via a bridging cysteinyl thiolate to a diiron subsite ([2Fe]_H) with CO and CN^– ligands that stabilize the iron ions in a low-spin state.⁷⁻¹¹ The two iron ions of the [2Fe]_H subsite are further ligated by a bridging azadithiolate ligand (⁻SCH₂NHCH₂S^–, ADT).^12,13 [FeFe] hydrogenases often feature additional ferredoxin-type iron–sulfur clusters, commonly referred to as “F-clusters”. These clusters serve as molecular wires enabling electron transfer between the H-cluster and external redox partners, and tune the catalytic properties of the enzyme.¹⁴⁻¹⁶

Figure 1.

Structural comparison between TamHydS and the prototypical [FeFe] hydrogenase CpI. (A) Alignment of the YASARA-generated homology model (Alignment RMSD: 1.541 Å) of TamHydS¹⁷ (green cartoon) with the crystal structure of CpI (gray cartoon, PDB ID: 4XDC(18)), including an overlay of F-clusters (orange: Fe, yellow: S for TamHydS, gray for CpI). The H-cluster is encircled in solid black. Note: one [4Fe-4S]-cluster is not modeled here. Figure S1 presents the AlphaFold model of TamHydS, including the third accessory [4Fe-4S]-cluster located in the C-terminal domain. (B) Top: Amino acids (gray: C, red: O, blue: N) constituting the proton-transfer pathway (C299, E279, S319, E282, and R286) or interacting with the H-cluster through H-bonding and other electrostatic interactions (C299, M353, and M497) in CpI. Hydrogen turnover is catalyzed at Fe_d (*), i.e., the iron ion most distal to the [4Fe-4S]_H cluster. The azadithiolate (ADT) bridgehead is labeled as blue text. Bottom: The corresponding residues from TamHydS (green: C, red: O, blue: N). The following variants were investigated: (1) single-point variant A137C, (2) “active site” or AS (A137C, L291M, and S191M in gray labels); (3) “proton-transfer pathway” or PTP (A137C, F119E, L157S, and K126R in cyan labels); and (4) “combination” of AS and PTP variations or CM. The glutamate residue (E122) in TamHydS was not substituted, given that the analogous residue in CpI, E282, is conserved as glutamate across both enzymes.

In addition to the inorganic cofactors, the influence of the protein matrix on [FeFe] hydrogenase catalysis and inhibition has been extensively studied.¹⁹ Structural and functional studies, in combination with site-directed mutagenesis, have revealed how the second coordination sphere controls the spatial arrangement of the ligands (i.e., CO, CN^–, ADT) and tune the reactivity of the H-cluster.²⁰⁻²³ Moving beyond the active site, the protein also ensures controlled delivery of protons and gaseous reactants, which further impacts the enzyme’s overall catalytic performance.^24,25 Still, the structural factors that enable the remarkable reactivity of [FeFe] hydrogenase remain opaque. The role of second and outer coordination sphere effects is also receiving attention in the design of synthetic systems for H₂ catalysis.²⁶⁻²⁸ Thus, improved understanding of the biological blueprints is expected to also benefit molecular design efforts.

To date, studies of the second coordination sphere have focused primarily on the structurally similar “prototypical” [FeFe] hydrogenases of Group A, including Chlamydomonas reinhardtii HydA1 (CrHydA1), Clostridium pasteurianum HydA1 (CpI), and the closely related Clostridium acetobutylicum HydA1 (CaHydA1), as well as the dimeric enzyme from Desulfovibrio desulfuricans (DdHydAB).²⁹ However, [FeFe] hydrogenases form a structurally diverse enzyme family, which can be divided into distinct phylogenetic groups denoted as Group A–G.^2,29−32 The immediate protein environment around the H-cluster is conserved in each group, but can vary significantly between groups.^29,33 Similarly, the proton-transfer pathway is altered between groups.^17,34−36 Comparative studies between different groups of [FeFe] hydrogenase are expected to provide a wealth of new information concerning the role of the protein scaffold in tuning the hydrogenase activity.

The putatively sensory [FeFe] hydrogenase from Thermoanaerobacter mathranii (TamHydS) is an example of a Group D enzyme.¹⁷ In comparison to the previously studied Group A [FeFe] hydrogenases, TamHydS displays significantly lower H₂ evolution rates. Moreover, although TamHydS functions as a bidirectional catalyst under electrochemical conditions, both proton reduction and H₂ oxidation display significant overpotentials.^17,37 Such irreversible bidirectional behavior is rare among reported hydrogenases and related molecular catalysts.^29,37−40TamHydS also displays unusual reactivity toward known hydrogenase inhibitors. Previously studied Group A [FeFe] hydrogenases show a complete loss of the H-cluster upon contact with O₂,⁴¹⁻⁴³ with few group members as exceptions.^42,44,45 Meanwhile, exposing TamHydS to O₂ transforms the [2Fe]_H subsite into a mononuclear species, which is stable under prolonged exposure to O₂ and retains limited H₂ evolution capacity.¹⁷ In addition to its unusual reactivity toward O₂, TamHydS was found to have a significantly higher tolerance toward CO. An elevated CO tolerance has also been reported for TmHydS from Thermotoga maritima, a representative example of a Group C [FeFe] hydrogenase (Table S1).³⁴ As the H-cluster is identical in all groups, the reactivity differences must be attributed to variations in second and/or outer coordination sphere, e.g., the active-site environment surrounding the H-cluster and/or the proton-transfer pathway.^17,35

Considering second coordination sphere effects, methionine residues M497_CpI and M353_CpI (CpI numbering; see Figure 1B), are proposed to be critical in modulating the reactivity of the H-cluster in Group A [FeFe] hydrogenases.^21,38 In the case of TamHydS, these residues are exchanged to leucine (L291_TamHydS) and serine (S191_TamHydS), respectively. Additionally, cysteine residue C299_CpI, which serves as the entry point into the well-conserved proton-transfer pathway of Group A,²⁵ is not conserved in Group D [FeFe] hydrogenases and is exchanged to alanine in TamHydS (A137_TamHydS). The corresponding cysteine-to-alanine variants in Group A enzymes (C169A_CrHydA1,^46,47 C299A_CpI,²⁴ and C298A_CaHydA1⁴⁸) exhibit reduction in H₂ turnover, highlighting the cysteine residue’s importance in catalysis.⁴⁹ Except for a glutamic acid (E122_TamHydS and E282_CpI), all residues considered crucial for proton transfer in CpI (C299, E279, S319, and R286)^25,50,51 are exchanged in TamHydS (to A137, F119, L157, and K126, respectively). An alternative proton-transfer pathway has been proposed that is conserved within Group D, but variable in other groups.³⁵ This striking difference in the proton-transfer pathway, or outer coordination sphere, was previously proposed as a possible factor rationalizing the irreversible catalytic behavior of TamHydS.³⁷ However, kinetic modeling has revealed that the irreversibility can be assigned to a one-electron reduced H-cluster intermediate, H_red, being stable over an unusually large potential range.⁵² This latter proposal is further supported by spectro-electrochemical titrations performed on the structurally related Group C enzyme, TmHydS.³⁴

The diverging structural features and catalytic properties of Group A and Group D [FeFe] hydrogenases provide a new entry point for investigating the complex interplay between the H-cluster and protein environment. Herein, we utilize the design principles of Group A [FeFe] hydrogenase to improve the catalytic performance of the Group D [FeFe] hydrogenase, TamHydS, via rational design. Three new enzyme constructs based on TamHydS were generated to mimic key features strictly conserved in Group A [FeFe] hydrogenases. As summarized in Figure 1B, these constitute an “active-site” (AS) variant and a “proton-transfer pathway” (PTP) variant. The two variants recreate important features of the active-site pocket and proton-transfer pathway of a prototypical [FeFe] hydrogenase, respectively. The third variant combines these modifications and is denoted the “combined” (CM) variant. For reference, a variant introducing only the active-site cysteine (A137C_TamHydS, equivalent to C299_CpI) was characterized (A137C). Critically, this approach allowed us to investigate the design principles of hydrogenase via gain-of-function variations rather than loss-of-function studies. Through a combination of spectroscopic, electrochemical, and biochemical analyses, we show that mimicking neither the Group A active site nor the proton-transfer pathway in isolation has a positive impact on the catalytic performance of TamHydS. Only the combined modification generates a hybrid enzyme displaying aspects of both Group A and Group D hydrogenases. Compared to TamHydS wild-type (WT), the CM variant improved H₂ evolution rates up to 170-fold, accompanied by improved stability toward inhibition by CO and O₂.

Results

Isolation and Spectroscopic Characterization of Variants

A137C, AS, PTP, and CM variants were generated based on a comparison between TamHydS and the crystal structure of CpI (Figure 1).¹⁸ The AS triple variant introduces two methionine residues (S191M and L291M) and the proton relay cysteine (A137C). The PTP quadruple variant includes the proton relay cysteine (A137C) in combination with additional proton-transfer pathway variations (F119E, K126R, and L157S). The CM sextuple variant incorporates all of the aforementioned variations. Lastly, the reference variant, A137C, featured only the proton relay cysteine mutation. The individual genetic constructs were transformed into BL21(DE3) Escherichia coli cells to produce the respective apo-protein, i.e., a form of the hydrogenase lacking the [2Fe]_H subsite (Figures S2 and S3). As no changes were made in the F-cluster binding domains, the variants are expected to incorporate three [4Fe-4S] clusters in addition to the [4Fe-4S]_H cluster. Albeit the fact that cells were lysed and the enzyme was isolated under anaerobic conditions, the iron content of the as-isolated variants differed slightly. Still, all variants shared the same final Fe:protein ratio of 15.9 ± 0.2 after reconstitution (Table S2).¹⁷ The successful assembly of TamHydS’ four [4Fe-4S] clusters was further confirmed by UV/visible spectroscopy and, in the case of AS, PTP, and CM, also by electron paramagnetic resonance (EPR) spectroscopy (Figures S4 and S5). The EPR spectra recorded on dithionite-reduced apo-forms of all constructs revealed two rhombic EPR signals attributable to the F-clusters ([4Fe-4S]⁺, S = 1/2) highly similar to that previously reported for WT,¹⁷ showing that the variations did not significantly alter the electron transfer relays (Figure S5). The reconstituted apo-proteins were subsequently treated with the synthetic [2Fe]_H site precursor [(ADT)Fe₂(CN)₂(CO)₄]^2– [([2Fe]^ADT)] and purified using desalting columns to generate the corresponding holo-enzymes. The Fe:protein ratio of the WT and variants was 17.9 ± 0.4, in good agreement with the addition of two iron ions in the active-site cofactor (Table S2).

EPR and in situ attenuated total reflection Fourier transformed infrared (ATR-FTIR) spectroscopy supports the formation of the H-cluster in all variants (Figures 2 and 3). In agreement with earlier studies,¹⁷ the reference EPR spectrum of a freshly activated sample (denoted “as-isolated”) of WT displayed a mixture of H-cluster states in the g ∼ 2 region (Table S3). Similar preparations of PTP and CM showed two highly similar rhombic “resting state” H_ox components (Figure 2B,C, black spectra). Simulation of the spectra revealed the following g-values for PTP H_ox-R1: g_1,2,3 = 2.100, 2.050, 2.012; H_ox-R2: g_1,2,3 = 2.109, 2.034, 2.008; for CM H_ox-R1: g_1,2,3 = 2.101, 2.051, 2.005; for CM H_ox-R2: g_1,2,3 = 2.109, 2.035, 2.009 (simulated spectra of H_ox-R1 and H_ox-R2 are shown as magenta and dark gray traces in Figure 2, respectively). There was also a significant contribution by an axial signal attributed to the CO-inhibited state H_ox-CO (g_II = 2.038, g_⊥ = 2.020, simulated H_ox-CO spectra are shown as blue traces in Figure 2). As expected, the latter signal dominated the spectra recorded of samples incubated under a CO atmosphere (Figure 2B,C, cyan spectra). The H_ox-CO signals from PTP and CM displayed a highly similar set of g-values but differed in their relaxation properties (Figure S6A,B). As-isolated AS presented the simplest spectrum, with the dominant signal similar to the H_ox-R2 state of WT, with g_1,2,3 = 2.109, 2.048, and 2.011 (Figure 2A, black spectrum). An axial signal attributed to H_ox-CO was observable for AS after prolonged exposure to CO (g_II = 2.030, g_⊥ = 2.019; Figure 2A, cyan spectrum). The narrower, more isotropic H_ox-CO signal in AS, compared to those in PTP and CM, partially overlaps with the residual [3Fe-4S]⁺ signal. However, its distinct shape and relaxation characteristics allow for differentiation (Figure S6C). Overall, the observed g-values for the H_ox and H_ox-CO states exhibit a high degree of similarity across the variants, with the most notable alterations observed in AS. The g-values consistently remain higher than those reported for prototypical [FeFe] hydrogenases (Table S3).

Figure 2.

EPR characterization of TamHydS variants. EPR spectra of the as-isolated (black) and CO-flushed (cyan) TamHydS variants (A) AS, (B) PTP, and (C) CM. The simulations for the as-isolated spectra are depicted as light gray traces overlaid with the respective experimental spectrum. All spectra reveal features attributable to the H-cluster residing in the H_ox state (component H_ox-R1 g-values are shown in magenta; component H_ox-R2 g-values are shown in dark gray). The axial signal is attributed to H_ox-CO (g-values in blue), which increases in the signal amplitude after an hour of flushing with CO. Spectra recorded with the following settings: T = 21 K; modulation frequency = 100 kHz; amplitude = 10 G; microwave frequency = 9.4 GHz. Microwave power: 16 μW for all spectra, with the exception of the spectrum of the CO-flushed PTP recorded at 253 μW. The individual components contributing to the overall simulated spectra are shown in magenta (H_ox-R1, simulated from black spectra), dark gray (H_ox-R2, simulated from black spectra), and blue (H_ox-CO), simulated from cyan spectra with the exception of PTP for which the H_ox-CO simulation was derived from the spectrum of CO-flushed PTP recorded at 64 μW (Figure S6D). The asterisk (*) in panel A corresponds to a trace signal from a [3Fe-4S]⁺-cluster (g ≈ 2.02; see Figure S6).

Figure 3.

ATR-FTIR characterization of TamHydS variants. Infrared spectra of (A) A137C, (B) AS, (C) PTP, and (D) CM for the reaction with H₂ (red traces) and CO (blue traces). The upper row shows absolute spectra, and the lower row shows the resulting “H₂–N₂” and “CO–N₂” difference spectra. The H-cluster states observed under N₂ (black labels, H_ox), H₂ (red labels, H_red), and CO (blue labels, H_ox-CO) are annotated. A137C, AS, and CM adopt H_red during H₂ exposure. AS and CM adopt H_ox-CO when exposed to CO. Arrows pointing upward (↑) or downward (↓) highlight increasing or decreasing bands. The magenta stars in the absolute spectrum of A137C indicate the band positions for State 2. The transition from State 1 to State 2 observed in the PTP “H₂–N₂” difference spectrum is annotated in green and magenta, respectively. Note that PTP in State 1 does not bind to CO. A summary of the bands associated with the H_ox, H_ox-CO, and H_red states together with a comparison to Group A enzymes is provided in Table S4. All data were recorded at room temperature.

In all three variants, incubating the enzyme under 1 atm of H₂ resulted in the expected decrease in H_ox and H_ox-CO signal intensities. Instead, new signals appeared, consisting of broad rhombic features attributable to reduced [4Fe-4S]-cluster species in combination with a narrow axial signal (Figure S7). The latter signal is assigned to a state denoted as “State 2”. State 2 refers to a still only partially characterized H-cluster state that forms in WTTamHydS, as well as previously studied variants, under H₂. Earlier studies have proposed the State 2 signal to reflect a species with a redox-state configuration similar to H_ox (i.e., an oxidized [4Fe-4S]_H-cluster coupled to a mixed valent Fe(I)Fe(II) [2Fe]_H site; see Supporting Note 1 for further discussion on this state).³⁵ Albeit present in all samples, differences are observed between the variants of TamHydS with regard to the intensity of the State 2 signal, implying the presence of contrasting amounts of reduced EPR silent species such as H_red (Figure S7). Similar to the H_ox signals, comparing the simulated EPR spectra of H₂-reduced CM and WT revealed that the State 2 signals are highly similar, indicating only minor shifts in their electronic structure.

The IR signatures of typical H-cluster states (H_ox, H_red, and H_ox-CO) were verified in the absolute and difference spectra of AS and CM and again differ only slightly from WTTamHydS (Figures 3, S8 and Table S4), supporting the EPR data. As previously observed for WT,¹⁷ both AS and CM adopted H_red as a semistable resting state when exposed to the H₂-containing atmosphere of the glovebox housing the FTIR instrument (1% H₂ in 99% N₂). Purging under a neat N₂ atmosphere resulted in the conversion of H_red into the oxidized species, H_ox. The rate of H-cluster oxidation differed between the variants but consistently occurred on a time-scale of hours (Figure S9). Exposure to H₂ reformed the H_red state on a seconds time-scale, although AS and CM reacted with H₂ 2–8 times slower than WT (Figure S10). Similarly, formation of H_ox-CO was found to be 3–5 times slower than WT when the enzyme films were exposed to a neat CO atmosphere (99% CO, 1% N₂, Figures S9 and S10). No H_ox-CO formation was observed when 1% H₂ was included in the headspace gas (data not shown).

Conversely, for PTP, only a negligible population of the H-cluster resided in the H_ox and H_red states. Instead, the spectrum was dominated by a signature that is strikingly similar to a species denoted State 1, recently observed in other variants of TamHydS targeting the enzyme’s native proton-transfer pathway (Figure 3C).³⁵ The exact nature of State 1 remains to be elucidated, but it is likely to represent an “over-oxidized” form of the H-cluster similar, but not identical, to H_inact (i.e., an oxidized [4Fe-4S]_H-cluster coupled to a diferrous [2Fe]_H site). Formation of State 1 in the PTP variant can be promoted through incubation under both a CO containing as well as a neat N₂ atmosphere, and consequently does not appear to involve binding of extraneous CO. As State 1 is also readily formed in the absence of an external sulfide, and in TamHydS variants lacking any cysteines in the active site, binding of a thiol ligand to the [2Fe]_H site can be ruled out. Instead, we have previously proposed that the [2Fe]_H subsite is coordinated by an aqua or hydroxido ligand (see also Supporting Note 1).³⁵ Reduction of PTP by H₂ yielded a spectrum typical of State 2.

While there is overall a good correlation between the FTIR and EPR data in the case of AS and CM, noticeable differences are observed for the PTP variant. State 1 dominates the FTIR spectrum, while the EPR spectrum of as-isolated PTP (Figure 2B) shows the distinct signatures of H_ox and H_ox-CO. This is at least partially attributable to the EPR silent (S = 0) nature of State 1, preventing its detection by EPR. By extension, this in turn facilitates detection of the mixed valent H_ox and H_ox-CO states by EPR even if present in relatively low concentrations. The apparent absence of these states in FTIR is instead attributed to their rapid loss upon exposure of the enzyme to the H₂-containing atmosphere of the glovebox housing the FTIR instrument. However, we note that changes in state populations arising from differences in sample conditions cannot be ruled out, e.g., the large temperature differences in EPR and FTIR (cryogenic vs room temperature, respectively).

Lastly, the A137C variant displayed a mixture of three H-cluster states (Figure 3A). Analogous to the CM and AS variants, the FTIR spectrum was dominated by signals that could be attributed to reduced species. More specifically, A137C featured a significant population of H_red, but in contrast to CM and AS, a large population of State 2 was also evident in the spectrum. A minor H_ox population was also discernible. The band signatures of the H_red and H_ox signatures were similar to those observed in WT, AS, and CM, while the State 2 bands align with those observed in PTP (Figure 3) and our earlier study.³⁵ Exposure of the A137C variant to higher H₂ concentrations yielded an increase in the H_red population with concurrent depopulation of H_ox, while the State 2 signal remained practically unchanged.

A summary of the observed FTIR bands associated with the H_ox, H_ox-CO, and H_red states together with selected reference values reported for Group A [FeFe] hydrogenases is shown in Table S4. Overall, recreating either the active-site second coordination sphere (AS) or the proton-transfer pathway (PTP) resulted in variants with clearly distinct properties. AS gave rise to a single form of H_ox, in contrast to the two forms observed for WT. On the other hand, PTP yielded an enzyme that appeared to be dominated by H-cluster species State 1 and State 2, while A137C featured a mixture of State 2 and H_red. Clearly, altering the second and outer coordination sphere impacts the relative stability of different H-cluster states in TamHydS, in line with earlier work on Group A [FeFe] hydrogenases.^24,53,54 In TamHydS, a tendency to preferentially populate State 1 and State 2 appears to be primarily associated with changes in outer coordination sphere residues connected to proton transfer, as observed here with the PTP variant and to a lesser extent in A137C, as well as in earlier work.³⁵ However, CM restored the spectroscopic properties of WT, although the kinetics governing the formation of various states were altered.

Catalytic Properties of Variants

The catalytic properties of the variants were evaluated using a combination of solution assays and protein film electrochemistry (PFE) to probe the effect of the variations on specific activity, apparent catalytic bias, and overpotential (η) for H₂ oxidation and proton reduction. WT was included in all of the assays as a reference.

Cyclic voltammetry (CV) traces were recorded following enzyme adsorption on a pyrolytic graphite electrode (PGE) in the presence of polycationic polymyxin B sulfate, as previously reported.¹⁷ To ensure that the observed currents closely approximate steady-state conditions, the potential was swept at a slow scan rate (2 mV/s). Substrate mass transport limitations for H₂ oxidation at high overpotentials were alleviated by saturating the cell with constant H₂ flow (1 atm) and by rotating the electrode at high speed (3000 rpm).

The CV traces, recorded under conditions of direct electron transfer (DET), clearly show that both CM and PTP are bidirectional catalysts, displaying catalytic currents related to both proton reduction and H₂ oxidation (Figure 4A). Under the conditions of the PFE experiment, both variants displayed a significant increase in the apparent catalytic bias relative to WT, i.e., the preference to catalyze the 2H⁺/H₂ interconversion in a specific direction. Indeed, the ratio of the proton reduction rates relative to H₂ oxidation increases 2–3-fold in CM and PTP when compared to WT, as determined from the currents observed at a high driving force with η ≈ 200 mV. More specifically, under the employed conditions, 40 °C and a pH of 6.0, the E°′ is approximately −0.37 V vs the normal hydrogen electrode (NHE), and the currents were determined at E = −0.57 and −0.15 V vs NHE for proton reduction and H₂ oxidation, respectively. This increase in the apparent bias toward proton reduction appears at least partially attributable to the inhibition of CM and PTP at strongly oxidizing potentials. This is observable as a decrease in current despite an increase in the driving force, apparent from approximately E > −0.12 V vs NHE (see also Figure S13). It is also in line with the more sluggish reactivity of CM toward H₂ oxidation as observed in the time-resolved FTIR experiments (Figure S9). Although AS was reactive toward H₂ oxidation as determined by FTIR and EPR spectroscopy (Figures 3B, S7, and S9), we found it to be catalytically inactive under the employed PFE conditions, as reflected in negligible current densities at low and high potentials (Figure S14A). The A137C variant also demonstrated a large reduction in both oxidative and reductive currents relative to those of WT. However, the currents observed for A137C did increase above the blank measurements (Figure S14B).

Figure 4.

Catalytic properties of TamHydS WT enzyme and variants. (A) Cyclic voltammetry (CV) traces (first scan) recorded for WT (black), A137C (gold), AS (light gray), PTP (cyan), and CM (red) subtracted with the blank electrode (no immobilized enzyme). Inset: Zoom in near the formal reduction potential of the H⁺/H₂ couple, emphasizing the apparent oxidative inactivation of PTP and CM at higher applied potentials. Second scans and current densities as a function of the scan number are shown in Figures S11 and S12. Experimental conditions: pH 6.0, 40 °C, 1 atm H₂, scan rate 2 mV/s, electrode rotation 3000 rpm. (B) Comparison of the specific activities for H₂ oxidation and evolution at pH 6.8 using methyl viologen (MV: 10 mM MV + 100 mM sodium dithionite, red bars), benzyl viologen (BV: 1 mM, gray bars), and methylene blue (MB: 30 μM, cyan bars). H₂ oxidation was measured in colorimetric assays at 25 °C and H₂ evolution by gas chromatography at 28 °C. Specific activity (U/mg) is one unit (U) of activity that catalyzes 1 μmol of H₂ oxidized or evolved per minute per mg of enzyme. One-way ANOVA (Tukey’s Honestly Significant Difference Test) results are indicated above the bars: significantly different at the p = 0.05 level (*) and not significantly different (ns) based on mean exam scores between WT and variants (Table S5). Data are presented as a mean of two biological replicates, each with 2–3 technical replicates with error bars showing the standard error of the mean (SEM). Not detected: H₂ evolution rate in AS. Not determined was H₂ oxidation of A137C using MB.

As previously observed for WT,¹⁷ the proton reduction currents of CM and PTP do not plateau but rather increase with the driving force, a behavior that is commonly attributed to random binding of the enzymes to the electrode surface and therefore dispersion of interfacial electron transfer rates.⁵⁵ This rate dispersion makes an exact determination of the catalytic onset potential challenging, but it is evident that both CM and PTP maintain a substantial potential window around E°′(2H⁺/H₂) devoid of significant current (Figure 4A, inset). The irreversible catalytic behavior of all TamHydS variants is in stark contrast to prototypical [FeFe] hydrogenases, which with few exceptions operate with negligible overpotential requirements in both directions (η ≈ 0 mV).^17,37 Earlier PFE studies of WT have confirmed that this is an intrinsic property of the enzyme and not due to slow interfacial electron transfer.³⁷ This is further underscored under the mediated electron transfer (MET) conditions. The addition of methyl viologen (MV) to the electrolyte enhanced the proton reduction currents significantly (Figure S15). However, it did not remove the irreversible nature of the catalysis. Instead, the relative rate enhancement increased with an increasing driving force. When comparing currents recorded during MET and DET at pH 6.8, a 3-fold increase in relative rate enhancement for WT was observed at −0.60 V that increased to 4-fold at −0.65 V vs NHE. In the case of CM, this potential-dependent rate enhancement was even more pronounced: the addition of MV resulted in a 7-fold increase of the current at −0.60 V, while a 12-fold increase was observed at −0.65 V vs NHE.

The redox mediator effect reveals that the maximum currents are limited by electron transfer. The MET data further shows that facilitating interfacial electron transfer through the addition of MV does not fully remove the dependence of the catalytic currents on the overpotential.⁵⁶ The structural rationale behind this irreversible catalytic behavior remains to be fully elucidated, and it is evidently not alleviated in either CM or PTP. Moreover, the difference in relative rate enhancement as a function of potential when comparing CM and WT under conditions of MET suggests a higher possible catalytic rate for CM given conditions of a sufficiently strong driving force.

The specific activities for H₂ oxidation and proton reduction were determined in solution assays (Figure 4B). H₂ oxidation rates were based on the initial rate of change (V₀) determined through colorimetric assays. The assays were performed at 25 °C under anaerobic conditions following injection of H₂-saturated buffer to a pH 6.8 solution of the respective enzyme. Considering the overpotential requirement of the variants, the effect of driving force was probed using different electron acceptors, i.e., benzyl viologen (BV, E^0′ = −0.359 V vs NHE,⁵⁷Figure S16) and methylene blue (MB, E^0′ = 0.008 V vs NHE,⁵⁷Figure S17). In parallel, H₂ evolution activities were determined via gas-chromatography assays.^58,59 The amount of H₂ in the reaction mixture headspace was measured after 15 min of incubation at pH 6.8 in the presence of methyl viologen (MV, E^0′ = −0.446 V vs NHE⁵⁷) as electron mediator and sodium dithionite as a sacrificial electron donor. The measured specific activities (units per milligram) are summarized in Table S5, and one unit of activity indicates 1 μmol of H₂ oxidized or evolved per minute under the employed assay conditions.

At low driving force, when BV was used as the electron acceptor, the specific activities for H₂ oxidation displayed by PTP (2.43 ± 0.52 U/mg) and CM (1.67 ± 0.03 U/mg) are comparable to that of WT (1.95 ± 0.27 U/mg). However, the H₂ oxidation capacity of WT substantially increased (22.1 ± 0.8 U/mg) when MB was used as an electron acceptor in the solution assay, outcompeting PTP and CM (3.70 ± 0.29 and 2.88 ± 0.89 U/mg, respectively). The deviating activities between the variants with an increased driving force for oxidation are in good agreement with the effect of overpotential on H₂ oxidation current densities measured in PFE (Figure 4A).

A distinctly different trend was observed for the H₂ evolution rates. One-way ANOVA (Analysis of Variance) with Tukey’s Honestly Significant Difference Test⁶⁰ was used to compare the means of the variants including WT. At the p = 0.05 level, there is no significant difference in mean H₂ evolution activity rates between WT and PTP, displaying specific activities of 0.25 ± 0.03 and 0.60 ± 0.01 U/mg, respectively. As expected from the PFE experiments, AS displayed negligible activities in both oxidation and reduction assays, while A137C showed only ≈10 and 25% residual activity for H₂ evolution and oxidation, respectively (Figure 4B). Lastly, CM increased the activity up to 170-fold versus WT, with CM having a specific activity for H₂ evolution of 36.2 ± 1.8 U/mg. This improved capacity for H₂ production in CM is in line with the PFE experiments, where CM showed a higher relative current enhancement compared to WT under MET conditions (Figure S15).

Reactivity toward O₂ of Variants

The sluggish reactivity of CM toward CO suggests that enhanced activity can be achieved without negatively impacting inhibitor tolerance. To further probe its inhibition properties, the susceptibility of CM toward O₂ was analyzed. The effect of O₂ on catalytic capacity was tested by incubating dilute aliquots (50 μL, 2 μM enzyme) of WT and CM under air, in the absence of any reductant in the buffer. Samples were treated with air for 0.5–10 min before being flushed with argon and returned to the glovebox. Subsequently, residual H₂ oxidation activity was determined by using the BV assay described above. As shown in Figure 5A, O₂ caused irreversible inactivation of the enzymes, and the process followed apparent first-order kinetics. A half-life time (t_1/2) of 0.84 ± 0.40 min was determined for WT. CM substantially improved the O₂ tolerance of WT, showing an approximate 3-fold extension of the half-life (t_1/2 = 2.45 ± 0.07 min). PTP displayed intermediate O₂ tolerance, with t_1/2 = 1.45 ± 0.20 min (Figure S18).

Figure 5.

Probing O₂ sensitivity of the TamHydS variants via enzymatic assays and spectroscopy. (A) Stability toward air was monitored via H₂ oxidation assays (Figure S16). The relative specific activities of WT (black) and CM (red) were measured following incubation under air (approximately 21% O₂) for 0.5–10 min. The data agrees with exponential decay fits (solid lines). The half-life values (t_1/2) extracted from the single exponential decay curves are in parentheses. Data are presented as a mean of 2–3 technical replicates, with error bars showing SEM. (B) Prolonged O₂ tolerance in CM (red) compared to WT (black) and CrHydA1 (gray) is reproduced by in situ ATR-FTIR spectroscopy following the decay of the H_red state over time. The presence of 1% H₂ in the gas phase stabilizes H_red and prolongs the deactivation of TamHydS, which is most clearly observed with CM. A similar protection was observed in CO inhibition (Figures S8–S10).

The increased O₂ tolerance of CM was verified by ATR-FTIR spectroscopy. Samples of WT and CM were pre-equilibrated in the H_red state under a 1% H₂ atmosphere. The subsequent introduction of 20% O₂ into the atmosphere resulted in a loss of the H_red population on a minute time-scale for both the WT and CM. The rate of H_red loss increased as the H₂ component of the atmosphere was removed. As seen in Figure 5B, already the WT was more stable than the prototypical CrHydA1 [FeFe] hydrogenase, and this stability was further improved in CM. Monitoring the products of the addition of O₂ revealed further differences between WT and CM (Figure S19). In all experiments, H_ox is the initial product of O₂-induced oxidation. However, while WT displayed parallel formation of the H_air state, reflecting a partially degraded H-cluster,¹⁷CM instead forms State 1. The PTP variant was also found to adopt State 1 in the presence of 20% O₂ albeit from State 2 rather than H_red or H_ox (Figure S19).

Discussion

The results presented in this work describe the spectroscopic, electrochemical, and biochemical features of rationally designed TamHydS variants, providing new insights into how interactions between the protein environment and the H-cluster modulate the overall catalytic performance and inhibitor tolerance of [FeFe] hydrogenase.

Optimizing the active-site environment or the proton-transfer pathway in isolation was not sufficient to improve the catalytic properties of TamHydS. Only when the two structural features of AS and PTP are combined, a clear synergistic effect is observed resulting in significantly enhanced H₂ evolution rates in CM. The improved capacity for proton reduction was clearly evident in solution assays (Figure 4B) and further supported by PFE through the observed current enhancements under the conditions of MET (Figure S15). The latter observation is important considering the recent report that [FeFe] hydrogenases can be inhibited by oxidation products of NaDT formed in standard hydrogenase solution assays,⁶¹ as it argues against a rate enhancement originating from improved inhibitor tolerance in CM.

The fact that CM maintains an irreversible catalytic response with midpoint potentials of the reductive and oxidative catalytic waves similar to those of the WT variant (Figure 4A) strongly suggests that the H-cluster reduction potentials remain largely unaltered relative to those of WT. This conclusion is supported by a recent study demonstrating that the midpoint potentials are directly related to the redox potentials and pK_a values of the associated protonation steps at the active site of TamHydS.⁵² Since it is highly unlikely that the redox potentials and pK_a values both change to compensate for each other’s contribution, other underlying mechanisms should be at play in influencing the catalytic properties of the variants.

A priori, it could be speculated that recreating the sulfur-rich environment of prototypical [FeFe] hydrogenase will alter the electronic structure and geometry of the H-cluster, improving its catalytic properties. However, the AS variant shows that this is clearly not sufficient to improve catalysis in the Group D enzyme. A related observation was recently reported for the low-activity Group C enzyme TmHydS, where modifications of amino acids in the active-site pocket did not increase the catalytic rates.⁶² The absence of current in cyclic voltammetry and the lack of detectable activity in solution assays instead provide evidence of the loss of catalytic function with AS. The EPR and FTIR spectroscopy data confirm that the H-cluster of AS has retained the capacity for H₂ activation. Instead, we attribute the loss of catalytic function to modifications in the active-site hydrogen-bonding network, due to steric displacements and changes in the electrostatic interactions. Collectively, we expect these changes to cause a disruption of the TamHydS’ native proton-transfer pathway and effectively isolate the H-cluster from the bulk solution.³⁵

The characterization of the single-point variant A137C allows us to propose a model for how the triple variant AS completely abolished catalytic activity. A137_TamHydS corresponds to the active-site cysteine of Group A (i.e., C299_CpI, C169_CrHydA1, C298_CaHydA1), which is known to be essential for coupling the H-cluster to the proton-transfer pathway, and likely fine-tunes the spatial positioning of the amine bridgehead of the ADT ligand through H-bonding.^21,48 Assuming that A137C behaves like the active-site cysteine in Group A enzymes, we can consider two different scenarios: when it is acting alone or in conjunction with the methionines in AS. A structural visualization of A137C and its potential effect on the positioning of the H-cluster, independently and when combined with other variations, is depicted in Figure 6.

Figure 6.

Proposed effect of single-point variant, A137C, in isolation and combined with other variants. (A) Homology model of TamHydS highlighting residues L291 and A137 (green sticks), targeted for variation, and the proposed native proton relay residue of TamHydS E252 (green stick) located ≈3.4 Å from the amine bridgehead of ADT (green dashed lines).³⁵ The H-cluster (diiron subsite is only shown for clarity) is depicted as a wire with C: green, O: red, N: blue, S: yellow, and Fe: orange. (B) A137C, as a lone variant, where the position of the A137C residue (pink stick) corresponds to C299_CpI (the position of all residues labeling in pink are based on the crystal structure of CpI, PDB ID 4XDC(18)). The introduced cysteine potentially interacts with the amine via electrostatic interactions (yellow dashed lines) or H-bonding (3.5 Å).^18,21 The proposed modified position of the H-cluster in this variant is depicted as a wire with C: pink, O: red, N: blue, S group: yellow, and Fe: orange. (C) In the triple variant AS, A137C is combined with L291M (and S191M, not shown). The L291M (pink stick) can also interact with the amine bridgehead potentially via H-bonding (3.6–3.7 Å) or increase flexibility due to decrease of steric hindrance.^18,21 This dual H-bonding interaction scheme is proposed to be detrimental due to complete disconnect of the ADT-amine from the TamHydS’ native proton-transfer pathway. (D) In the quadruple variant PTP, A137C is combined with other residues (pink sticks), steering A137C away from the amine bridgehead (note that here the A137C residue deviates from the crystal structure of CpI as it is manually shifted to illustrate the proposed effect). This shift reduces the interaction of A137C with the ADT-amine, allowing for continuous operation of the native proton-transfer pathway of WT. (E) In the sextuple variant CM, reintroducing L291M enables A137C to reconnect with the amine bridgehead and now also connects the ADT-amine to the improved PTP.

In TamHydS and other Group D [FeFe] hydrogenases, a glutamate positioned at E252_TamHydS has been proposed to replace the active-site cysteine of Group A enzymes (C299_CpI) as the proton relay residue (Figure 6A).³⁵ As a standalone variant, A137C may interact with ADT-amine, likely via H-bonding (Figure 6B). Previous reports indicate longer than expected H-bonding distances between the cysteine and the amine bridgehead: 3.5 and 3.2 Å for CpI^7,18 and DdH,⁸ respectively, although computational studies suggest these distances could be shorter (2.1 Å).⁶³ We propose that introduction of this cysteine into the active site of TamHydS causes an elongation of the distance between the ADT-amine and the native proton relay (E252_TamHydS) but not enough to completely sever their connection and entirely abolish catalytic activity (Figure 6B).

Introducing the methionine L291M (corresponding to M497_CpI and M415_CrHydA1) in triple-variant AS opens up the possibility of another H-bonding interaction to the ADT-amine (Figure 6C). In combination, A137C and L291M may modulate the spatial positioning and proton conducting ability of the ADT-amine (via NH–S or N–HS bonds) and contribute to the increased flexibility of the active site (Figure 6C). This dual interaction negatively impacts the catalysis of TamHydS, as evidenced by the observed loss of catalytic activity in AS, most likely due to the disconnect of the amine bridgehead from the native proton-transfer pathway. The potential influence of S191M (corresponding to M353_CpI and M223_CrHydA1) has not been probed in detail here, leaving its role in tuning the H-cluster somewhat ambiguous. However, the methionine in this position (Figure 1B) can interact electrostatically with the bridging CO, and it has been previously suggested that this interaction is crucial for achieving high turnover rates, adjusting affinity for H₂, and modulating H-cluster reactivity.^21,38

More broadly, considering the highly conserved nature of the active-site methionines in Group A [FeFe] hydrogenase, H-bonding is unlikely to be their sole role. Moreover, the distances observed by crystallography in Group A [FeFe] hydrogenases between the ADT-interacting methionine (M497_CpI) sulfur and the ADT-amine are again somewhat longer than would be expected for a H-bonding interaction, 3.6–3.7 Å for CpI^7,18 and 3.9 Å for DdH.⁸ Instead, the relatively soft nature of the methionine sulfur is likely to further allow the ADT ligand to optimize its position relative to Fe_d and active-site cysteine. The side chain of methionine is known for its significant conformational adaptability, due to the length and low rotational energy barrier of the C_methyl–S bond,⁶⁴ while leucine, albeit having a fairly large side chain, is still less flexible than methionine due to higher rotational energy barriers. For Group A [FeFe] hydrogenases, it has been proposed that M497_CpI may act as a “cushion” between the H-cluster and the peptide structure by absorbing the impact of their relative movements.²¹ The specific substitution of methionine with leucine in the putatively sensory TamHydS WT raises intriguing questions. We hypothesize that the increased rigidity conferred by leucine might enhance detection of external stimuli (e.g., H₂ pressure) and ensure consistent signaling. Conversely, catalytic enzymes might require a more flexible active site to accommodate substrates more effectively and allow for rapid transitions between conformational states.⁶⁵ In another sensory hydrogenase, Group C TmHydS, the methionine in this position is replaced by serine.⁶² While serine can form H-bonds, it is generally considered to have less conformational flexibility than methionine. Similar to L291M, the methionine side chain of S191M can also introduce additional flexibility within the active-site environment. Prior studies on Group A have shown that replacing this methionine (M353_CpI) with a more rigid leucine led to reduced activities compared to wild-type enzymes, though not as severely as variations in the two residues interacting with the amine bridgehead (i.e., C299_CpI and M497_CpI).³⁸ Exploring the balance between rigidity and flexibility in sensory and catalytic metalloenzymes and their impact on the physiological function presents an interesting point of further investigation.

Moving beyond the immediate active-site environment, introducing the Group A-type proton-transfer pathway could accelerate catalysis through a faster long-range proton transfer. However, no enhanced catalytic activity was observed in the PTP variant. We speculate that the H-cluster is not positioned to benefit from the engineered proton trajectory as long as the active-site methionines interacting with the H-cluster are missing (Figure 6D). When A137C is not paired with L291M but instead linked with other residues more distant from the H-cluster, the interaction of A137C with the amine bridgehead is likely reduced due to redirection of the cysteine thiol by electrostatic or H-bonding interactions with other PTP residues or water molecule(s) (Figure 6D). As a consequence, the incorporation of additional PTP residues does not result in faster catalysis but alleviates the negative effects of A137C, as the variant enables the native proton-transfer pathway via E252_TamHydS to function continuously, thereby, preserving the catalytic activity of WT.

With regard to PTP, it is also noteworthy that the FTIR spectra of this variant is dominated by State 2 under a H₂ atmosphere, in contrast to WT and CM that instead display primarily H_red with a small population of State 2 only discernible by EPR spectroscopy under equivalent conditions. The single-point variant A137C also exhibits a clearly discernible population of State 2, albeit H_red also forms upon H₂ treatment (Figure 3A). The origin of this difference in the H-cluster state accumulation has not been identified. However, considering that PTP displays similar H₂ formation rates as WT, it suggests that State 2 is catalytically competent.

Considering the reactivity of the PTP and AS variants, the increased H₂ formation rates observed for CM highlight the influence of the protein matrix on H-cluster reactivity, as well as the complexity associated with macromolecular catalyst design. The rate enhancement in CM is in good agreement with the notion that the catalytic performance of WTTamHydS is limited by long-range substrate (proton) transport during H₂ formation. However, the faster proton-transfer kinetics provided by the Group A-type proton-transfer pathway must evidently be coupled with fine-tuning of the active-site structure for the rate enhancement to become apparent (Figure 6E). Following the reasoning above, the two methionine residues would then serve to reposition the H-cluster so as to connect it to the alternative “faster” proton-transfer pathway. Nevertheless, CM continues to display overpotential requirements that are significantly larger than those of Group A enzymes. This means that although a 170-fold increase in the H₂ evolution rate was observed for CM, the specific activities as determined by the MV-based assay possibly underestimate the maximum rate. Indeed, the CV measurements with MV as a mediator further reveal that the proton reduction rate enhancement in CM is dependent on the applied overpotential. While the oxidation assay could be readily adapted to increase the driving force (i.e., employing benzyl viologen or methylene blue as the electron acceptor), the reduction assay is limited by the lack of suitable viologen redox mediators with reduction potentials below that of MV. The driving force dependence on H₂ evolution rates in solution presents an important line of inquiry moving forward.

A key feature of TamHydS is its relatively high tolerance toward inhibitors, as compared to previously studied prototypical [FeFe] hydrogenases.¹⁷ Efforts to improve catalytic features and O₂ tolerance of hydrogenases to O₂, based on proposed molecular mechanisms of proton, gas, and electron transfer, are still limited. Most protein engineering studies to shift the catalytic bias toward proton reduction and increase O₂ tolerance have been performed on [NiFe] hydrogenases, and some success has been achieved.⁶⁶⁻⁷³ Parallel attempts of optimizing Group A [FeFe] hydrogenases have so far proven challenging,^23,74 although it has been shown that there is no strict inverse correlation between O₂ tolerance and high catalytic rates.⁷⁵ In line with the latter observation, PTP increased the level of tolerance of O₂ with retained H₂ production activities relative to WT, while CM provides a striking example that rates can be improved concomitantly with inhibitor tolerance. Three main mechanisms for O₂ protection have been observed in hydrogenases, including (i) complete reduction of O₂ to benign H₂O, avoiding the formation of reactive oxygen species (ROS);^42,76,77 (ii) restricting access of inhibitors via proteinaceous gas filters;^78,79 and (iii) external ligand binding, such as the thiol-inhibited state H_inact^44,80,81 or hydroxide-inhibited Ni–A and Ni–B states of [NiFe] hydrogenase.⁸² The slow reaction with CO observed for CM could be due to more limited gas access, arguing for inhibition protection via a gas filter effect in the variant. However, our findings suggest a further protective mechanism that involves the binding of an external ligand, as CM and PTP exhibited FTIR spectra akin to those of H_inact, i.e., State 1 forming upon exposure to O₂ (Figure S19, Supporting Note 1).³⁵ We speculate that the earlier proposed aqua (or hydroxido) ligand binding in State 1 can protect CM and PTP against O₂ attack,³⁵ similar to [NiFe] hydrogenase.⁸² It follows that State 1 would be an inhibited state, and the formation of State 1 upon oxidation could also rationalize the oxidative inactivation observed for CM and PTP in PFE experiments (Figure 4A).

Conclusions

This study demonstrates how design aspects of two phylogenetic groups can be employed for improving the performance of a macromolecular catalyst through rational design. It shows the importance of exploring a broader range of [FeFe] hydrogenases to probe the impact of different protein structural features on H-cluster reactivity. By extension, these efforts will also provide new paths toward metalloenzyme engineering and alternative model systems for biomimetic synthetic catalysts. More specifically, the current study provides gain-of-function support for both the proposed proton-transfer pathway of prototypical [FeFe] hydrogenase and the importance of the sulfur-rich active-site canopy for fast catalysis. We hypothesize that the observed rate enhancement in CM arises from optimized coupling between the engineered faster proton-transfer pathway and the H-cluster, which relies on the precise positioning of the central proton relay cysteine as well as the [2Fe]_H subsite for optimal engagement with ADT-amine. This interaction is critically contingent on the presence of another sulfur-containing residue, methionine, that also interacts with the amine bridgehead. While the role of the second methionine, positioned on the opposite side of [2Fe]_H close to the μCO ligand, remains to be fully elucidated, it is conceivable that the introduced sulfur-rich residues collectively result in a relatively more flexible active-site environment for the H-cluster, allowing faster conformational state transitions. We note that the identified coupled effects between structural features, i.e., the proton-transfer pathway and active-site environment, would be difficult to identify via classical loss-of-function studies. The factors that govern the irreversible catalytic response of TamHydS remain elusive, and solving this issue represents an intriguing molecular design challenge. Nonetheless, the study highlights that activity rates, overpotential, and susceptibility to gaseous inhibitors are decoupled and thus can be tuned separately. Indeed, [FeFe] hydrogenases can evidently be tailor-made to become faster and more robust catalysts through rational design. Lastly, it highlights a critical design challenge with regard to the current trend of immobilizing molecular catalysts in larger (electrocatalytic) polymers, as the catalyst positioning relative to the proton (substrate) delivery pathway needs to be optimized in tandem.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.4c17364.

• Experimental procedures, Alphafold model of TamHydS, AGE/PAGE gels, Fe/protein content, additional EPR and FTIR data, PFE data, absorbance traces for reconstitution and solution assays, and an extended discussion on the properties of State 1 and State 2 (PDF)

The Swedish Energy Agency (STEM, Grant No. 48574-1 to G.B.), the Olle Engkvist stiftelse (Grant No. 220-0226 to G.B. and M.S.), and the European Union Horizon Europe—the Framework Programme for Research and Innovation (2021–2027) under the Grant Agreement Number 101070948 (Project PhotoSynH2 to G.B.) are gratefully acknowledged for funding.

The authors declare no competing financial interest.