Redox tuning of the H-cluster by second coordination sphere amino acids in the sensory [FeFe] hydrogenase from Thermotoga maritima

[FeFe] hydrogenases are exceptionally active catalysts for the interconversion of molecular hydrogen with protons and electrons. Their active site, the H-cluster, is composed of a [4Fe–4S] cluster covalently linked to a unique [2Fe] subcluster. These enzymes have been extensively studied to understand how the protein environment tunes the properties of the Fe ions for efficient catalysis. The sensory [FeFe] hydrogenase (HydS) from Thermotoga maritima has low activity and displays a very positive redox potential for the [2Fe] subcluster compared to that of the highly active prototypical enzymes. Using site directed mutagenesis, we investigate how second coordination sphere interactions of the protein environment with the H-cluster in HydS influence the catalytic, spectroscopic and redox properties of the H-cluster. In particular, mutation of the non-conserved serine 267, situated between the [4Fe–4S] and [2Fe] subclusters, to methionine (conserved in prototypical catalytic enzymes) gave a dramatic decrease in activity. Infra-red (IR) spectroelectrochemistry revealed a 50 mV lower redox potential for the [4Fe–4S] subcluster in the S267M variant. We speculate that this serine forms a hydrogen bond to the [4Fe–4S] subcluster, increasing its redox potential. These results demonstrate the importance of the secondary coordination sphere in tuning the catalytic properties of the H-cluster in [FeFe] hydrogenases and reveal a particularly important role for amino acids interacting with the [4Fe–4S] subcluster.


Activity assays
Hydrogen production was measured by gas chromatography on a 6890 Series GC System (Agilent Technologies) using a molecular sieve 5 Å PLOT column using reduced methyl viologen as the electron donor. The 400 µl reactions were set up in 2.5 mL stoppered plastic tubes containing 10 µg -50 µg of artificially maturated TmHydS variants in 200 mM potassium phosphate buffer pH 8, 100 mM sodium dithionite, and 10 mM methyl viologen. The reactions were carried out at 70 °C and they were initiated by addition of the protein. The reaction vials were purged with argon for 5 min and incubated in a temperature-controlled water bath at 70 °C for 10 min before extraction of 0.3 mL of the headspace gas for analysis. Hydrogen content was quantified by comparison with a 100% H2 standard. All values are the average of three measurements after subtracting the value of a blank measurement.
Hydrogen oxidation was measured by following the reduction of 1 mM benzyl viologen in H2 -saturated 200 mM phosphate buffer pH 8 with 1.0 µg -10 µg artificially maturated TmHydS variants. The specific activity of the protein was measured by the initial rate of change of absorbance at 600 nm due to the reduction of benzyl viologen. The measurements were performed in 1.5 ml plastic cuvettes using an Ocean Optics DH-mini UV-Vis-NIR light source and a USB2000 + XR1-ES detector, operated by the SpectraSuite software. The reactions were carried out at 70 °C using a temperature controlled cuvette holder (CUV-QPOD-2E-ABSKIT, Ocean Optics). All values are the average of three measurements after subtracting the value of the blank measurement. The other details are described in the figure legends.

EPR spectroscopy
For X-band (9.63 GHz) EPR spectroscopy, 200 μl of samples were transferred to X-band quartz EPR tubes and frozen in liquid nitrogen. The spectra were recorded on a Bruker ELEXSYS E500 CW EPR spectrometer. Cryogenic temperatures were maintained with liquid He using an Oxford ESR900 helium flow cryostat. The measurement parameters were: modulation frequency, 100 kHz, modulation amplitude, 7.46 G, time constant, 81.92 ms, conversion time, 81.92 ms. All the spectra were processed using home written programs in the MATLAB TM environment. EPR simulations were also carried out in MATLAB TM using the 'esfit' fitting function from the Easyspin package. 6

FTIR spectroscopy and spectroelectrochemistry
FTIR spectra of the samples were recorded using a Bruker IFS 66v/S FTIR spectrometer equipped with a liquid nitrogen cooled Bruker mercury cadmium telluride (MCT) detector. For FTIR spectroscopy, 10 μl of sample were placed between two CaF2 windows (Korth Kristalle, Altenholz), separated by a 50 μm Teflon spacer. These windows were then accommodated in a FTIR cell and fixed with rubber rings. Spectra were collected at 15 °C in the double sided, forward-backward mode with 1000 scans, and a resolution of 2 cm -1 , an aperture setting of 2 mm and scan velocity of 20 kHz.
Spectroelectrochemical FTIR was carried out in the same spectrometer set-up, but using a home built electrochemical IR cell, constructed according to an original design by Moss et. al. 7 Protein samples (≈ 1 -1.5 mM, 30 µl) mixed with 0.5 mM of each redox mediator (potassium indigo trisulphonate, anthraquinone-1,5-disulfonic acid, anthraquinone-2-sulfonate, benzyl viologen, methyl viologen and 1,1´,2,2´-tetramethyl-[4,4´-bipyridine]-1,1´-diium iodide) and loaded between two CaF2 windows on an electrochemically reduced gold mesh working electrode (approximately 50 µm thick) in electrical contact with a platinum counter electrode. An Ag/AgCl (1 M KCl) electrode was used as a reference and was calibrated before and after measurement with (hydroxymethyl)ferrocine (E = + 436 mV vs SHE). The potential was set using an Autolab PGSTAT101 potentiostat controlled by Nova software. The sample was equilibrated at a particular potential until the current flowing through the cell reached an equilibrium plateau, which took ≈ 30 -75 min. This was followed by measurement of the FTIR spectrum at that potential. All the FTIR spectra were processed using home written routines in the MATLAB TM environment. Table S1. Sequences of the primers used for mutagenesis Primer name Primer sequence  Figure S1. WebLogo plot (prepared using 251 sequences from the HydDB server 8 ) depicting the conservation pattern of amino acids mutated in this study and their vicinity in sensory hydrogenases. The overall height of each stack of letters in the logo plot indicates sequence conservation at a particular position, and the height of each amino acid one-letter code in each stack shows the frequency of occurrence of those amino acids in that position.   Figure S5. Magnified view of the H-cluster and the proton channel in the CpI hydrogenase. 9 The water molecule involved in proton transfer is shown as a sphere (turquoise). All the proton channel residues, C299, S319, E279 and E282 are altered in TmHydS.  (1861 cm -1 ) states, the latter two were determined from fitting Gaussian peaks (see Figure S8), is plotted against the applied potential and fitted with a model based on the Nernst equation using n = 1 for the Hox/H red * (Em = -330 mV) and H red * /H sred * (Em = -625 mV) transitions. These data represent the reverse titration (re-oxidation) of the IR spectroelectrochemical titration presented in Fig. 6 of the main text. S12 Figure