A membrane‐bound [NiFe]‐hydrogenase large subunit precursor whose C‐terminal extension is not essential for cofactor incorporation but guarantees optimal maturation

Abstract [NiFe]‐hydrogenases catalyze the reversible conversion of molecular hydrogen into protons end electrons. This reaction takes place at a NiFe(CN)2(CO) cofactor located in the large subunit of the bipartite hydrogenase module. The corresponding apo‐protein carries usually a C‐terminal extension that is cleaved off by a specific endopeptidase as soon as the cofactor insertion has been accomplished by the maturation machinery. This process triggers complex formation with the small, electron‐transferring subunit of the hydrogenase module, revealing catalytically active enzyme. The role of the C‐terminal extension in cofactor insertion, however, remains elusive. We have addressed this problem by using genetic engineering to remove the entire C‐terminal extension from the apo‐form of the large subunit of the membrane‐bound [NiFe]‐hydrogenase (MBH) from Ralstonia eutropha. Unexpectedly, the MBH holoenzyme derived from this precleaved large subunit was targeted to the cytoplasmic membrane, conferred H2‐dependent growth of the host strain, and the purified protein showed exactly the same catalytic activity as native MBH. The only difference was a reduced hydrogenase content in the cytoplasmic membrane. These results suggest that in the case of the R. eutropha MBH, the C‐terminal extension is dispensable for cofactor insertion and seems to function only as a maturation facilitator.

and iron ions in the active site are coordinated to the protein backbone via four cysteine-stemming thiolates. Two cysteines ligate both metals and therefore serve as bridging ligands. Furthermore, two cyanides (CN) and one carbon monoxide (CO) belong to the ligand sphere of the iron and keep the metal in a low-spin Fe 2+ state. Maturation and insertion of the NiFe(CN) 2 (CO) cofactor into the apo-form of the large subunit require a sophisticated maturation machinery that consists of at least six auxiliary Hyp proteins (Böck, King, Blokesch, & Posewitz, 2006;Lacasse & Zamble, 2016).
First, the Fe(CN) 2 (CO) moiety is assembled with the aid of the HypE and HypF proteins, which synthesize the cyanide ligands out of carbamoyl phosphate (Blokesch et al., 2004;Reissmann et al., 2003). The metabolic origin of CO under anaerobic conditions remains, however, unclear (Bürstel et al., 2011;Nutschan, Golbik, & Sawers, 2019), while under aerobic conditions, this diatomic ligand is derived from formyltetrahydrofolate (Bürstel et al., 2016;Schulz et al., 2020). Assembly takes place on a scaffold complex, consisting of the HypC and HypD proteins, from which the Fe(CN) 2 (CO) moiety is transferred to the apo-large subunit (Bürstel et al., 2012;Stripp et al., 2013). Nickel is subsequently delivered to the active site with the help of the metallochaperones HypA and HypB.
HypB was shown to feed HypA with nickel, which is compatible with a model in which HypA donates nickel to the large subunit (Lacasse, Summers, Khorasani-Motlagh, George, & Zamble, 2019;Watanabe et al., 2015). The assumption that HypA delivers the active site nickel was recently supported by observations made for the apo-large subunit of the cytoplasmic [NiFe]-hydrogenase from Thermococcus kodakarensis that has been crystallized in a complex with the HypA protein (Kwon et al., 2018). Interestingly, the interaction of HypA with the flexible N-terminus of the large subunit brought the chaperone in close vicinity of the still vacant active site. This was in fact a surprising result, because so far only the C-terminal extension of the large subunit stood in the focus of [NiFe]-hydrogenase maturation. The apo-large subunit is usually synthesized with a C-terminal extension comprising 3-68 amino acids (Greening et al., 2015), which is cleaved off by a specific endopeptidase once the complete NiFe site has been incorporated (Böck et al., 2006;Fritsch, Lenz, & Friedrich, 2013;Theodoratou, Huber, & Böck, 2005). Modifications of this C-terminal extension, including amino acid exchanges, truncation (Theodoratou, Paschos, Mintz-Weber, & Böck, 2000), or even complete removal (Massanz, Fernandez, & Friedrich, 1997;Senger, Stripp, & Soboh, 2017;Thomas, Muhr, & Sawers, 2015), by genetic engineering generally lead to the formation of inactive hydrogenase. Interestingly, while exchanges and truncations revealed a premature large subunit that was unable to form a complex with the small subunit, genetic removal of the entire extension allowed the formation of hydrogenase with canonical subunit composition. This phenomenon has been observed for the soluble, NAD + -reducing [NiFe]hydrogenase of R. eutropha (Massanz et al., 1997) and, more recently, for membrane-bound [NiFe]-hydrogenase 2 (Hyd-2) from Escherichia coli (Thomas et al., 2015). In both cases, however, the large subunit was at least devoid of nickel (Massanz et al., 1997).
Nickel-free Hyd-2 was also shown to lack the CN and CO ligands of the Fe(CN) 2 (CO) moiety, which are easily detectable by infrared spectroscopy (Senger et al., 2017). These observations suggest an essential role of the C-terminus in active site maturation.
However, it has to be stressed at this point that numerous [NiFe]-hydrogenases are naturally devoid of a C-terminal extension, yet they seem to employ the canonical maturation machinery to acquire a NiFe active site. Prominent members are the H 2sensing hydrogenases (belonging to groups 2b and 2d according to (Greening et al., 2015)), CO dehydrogenase-associated hydrogenases (group 4c), Ech-type hydrogenases (group 4e), and certain so far uncharacterized hydrogenases (group 4g) (Greening et al., 2015). The parallel occurrence of [NiFe]-hydrogenase large subunits with and without C-terminal extension leaves the importance of the C-terminal extension in the maturation of [NiFe]hydrogenases ambiguous.
To obtain further information, we investigated the role in maturation of the C-terminal extension of the large subunit of the  (Lenz, Lauterbach, Frielingsdorf, & Friedrich, 2015). Three of them, the soluble cytoplasmic NAD + -reducing [NiFe]-hydrogenase, the membrane-bound [NiFe]-hydrogenase (MBH), and the actinobacterial-like [NiFe]-hydrogenase, harbor large subunits whose apo-forms carry C-terminal extensions. The large subunit of the H 2 -sensing regulatory hydrogenase (RH), by contrast, is devoid of a C-terminal extension (Kleihues, Lenz, Bernhard, Buhrke, & Friedrich, 2000), although being equipped with a canonical NiFe(CN) 2 (CO) center (Bernhard et al., 2001), which is incorporated by the Hyp machinery of R. eutropha ).
The basic hydrogenase module of the MBH of R. eutropha consists of the large subunit HoxG carrying the NiFe(CN) 2 (CO) cofactor and the small subunit HoxK comprising three iron-sulfur (Fe-S) clusters . Upon incorporation of the NiFe(CN) 2 (CO) cofactor into HoxG by the Hyp machinery, the specific endopeptidase HoxM cleaves off the C-terminal extension.

Deletion of the gene encoding the MBH-specific endopeptidase
HoxM results in the accumulation of a HoxG preform still carrying the C-terminal extension (Bernhard, Schwartz, Rietdorf, & Friedrich, 1996;Hartmann et al., 2018).
In this study, we deleted the C-terminal extension of HoxG by genetic engineering and show that the resulting truncated version of HoxG still receives a NiFe(CN) 2 (CO) cofactor and forms, together with the corresponding HoxK subunit, catalytically active MBH.

| Genetic constructions
All bacterial strains and plasmids used in theis study are listed in Table 1. The sequence encoding the C-terminal extension of hoxG (amino acids 604-618 of HoxG, ) was eliminated by PCR amplification using the primers SFP43 5′-AAGAATGTATACGTGCCAGACGTG-3′ and SFP44 5′-ACTAAGCTTTTAGTGAGTCGAACACGCCAGAC-3′ using pJH5415 as template. The PCR product was digested with AccI/HindIII and ligated into AccI/HindIII-cut pJH5415 yielding pSF8.14. pSF8.14 was digested with SpeI/XbaI, and the resulting 3595-bp fragment was ligated into XbaI-cut pEDY309 yielding pSF10.8. This plasmid was transformed into E. coli S17-1 (AK2429) for subsequent conjugative transfer to R. eutropha strain HF1063 yielding strain HP9. The wild-type control strain was generated by digesting pJH5415 with SpeI/XbaI and transfer of the resulting fragment into pEDY309, yielding pJH5437. Plasmid pJH5437 was transferred by conjugation into strain HF1063, yielding strain HP3.

| Media composition and cell cultivation
Ralstonia eutropha strains HF649, HF1063, HP3, and HP9 were cultivated in FGN mod medium as described elsewhere (Hartmann et al., 2018;Lenz et al., 2018). Cells were harvested by centrifugation (11,500 g, 4°C, 12 min), and the resulting cell pellet was frozen in liquid nitrogen and stored at − 80°C. Lithoautotrophic cultivation in liquid and on agar-solidified minimal medium devoid of organic carbon sources was carried out as described previously (Lenz et al., 2018) with the exception that a gas atmosphere of 10% (v/v) CO 2 , were inoculated with a preculture to an initial OD 436 nm of 0.1 and shaken at 120 rpm and 30°C for 16 days. Agar plates were incubated for 6 days at 30°C.

| Protein extract preparation, polyacrylamide gel electrophoresis, and immunological analysis
To analyze the MBH subunit content in different cellular protein fractions, cell pellets were resuspended in 3 ml (per gram wet weight) of resuspension buffer (50 mM K 2 HPO 4 /KH 2 PO 4 , pH 7.3, 150 mM NaCl) containing cOmplete EDTA-free protease inhibitor mixture (Roche Diagnostics) and DNase I (Roche Diagnostics).
The resuspended cells were disrupted in a chilled French pressure cell at 124.11 MPa. This procedure resulted in whole-cell lysate ("lysate" sample). Unbroken cells and cell debris were subsequently sedimented by centrifugation (4,000 g, 4°C, 20 min), yielding an emulsion composed of membranes and soluble proteins. The emulsion was ultracentrifuged (100,000 g, 4°C, 1 hr), yielding a dark-brown membrane pellet and a brownish liquid supernatant ("soluble extract" sample). The membrane pellet was washed by homogenization with a Potter-Elvehjem homogenizer in 10 ml of resuspension buffer (per gram, wet weight, of the membrane) containing protease inhibitor cocktail. The suspension was then ultracentrifuged (100,000 g, 4°C, 35 min), yielding clean membranes as a pellet. Its resuspension in resuspension buffer yielded the sample "membrane". Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli, 1970) was used for protein separation, which was followed by Western blot analysis (Towbin, Staehelin, & Gordon, 1979). Proteins in gels were either stained with Coomassie brilliant blue G-250 (Diezel, Kopperschläger, & Hofmann, 1972) or transferred to a nitrocellulose membrane (BioTrace, Pall Corp.) using a fast semidry transfer buffer. (Garić et al., 2013) Polyclonal antibodies raised against the MBH large subunit (anti-HoxG, (Bernhard et al., 1996)) in combination with a goat anti-rabbit secondary antibody (coupled with alkaline phosphatase, Dianova) were used for detection of HoxG. Protein concentrations were determined with the Pierce BCA Protein Assay Kit (Thermo Scientific), using bovine serum albumin as standard.

| MBH purification
Purification of the MBH Strep variants derived from R. eutropha HP3, HP9, and HF649 was carried out as described previously (Goris et al., 2011;Lenz et al., 2018). The protein samples were frozen in liquid nitrogen and stored at − 80°C.

| H 2 oxidation activity assay
H 2 -mediated reduction of methylene blue was determined spectrophotometrically as previously described (Lenz et al., 2018) using a Cary50 UV-Vis spectrophotometer (Varian, Agilent). The specific activity was given in Units (U) per mg of protein, where 1 U corresponds to the turnover of 1 µmol H 2 per minute.

| RE SULTS
Previous studies revealed that the genetic removal of the C-terminal extension of the large subunit results in hydrogenase devoid of a functional NiFe active site (Massanz et al., 1997;Thomas et al., 2015). To test whether the absence of the C-terminal extension of the membrane-bound hydrogenase (MBH) of R. eutropha also leads to inactive protein, we deleted the sequence encoding the amino acid residues Val604-Arg618 of the HoxG subunit. The deletion resulted in a preprocessed HoxG subunit, henceforth designated as HoxG proc , terminating with residue His603 that also represents the very last residue upon cleavage of the native subunit with the protease HoxM . The corresponding hoxGproc gene was cloned together with hoxK Strep , which encodes the  According to the current model, the C-terminus of the large subunit is cleaved off only if nickel had been inserted properly into the active site (Böck et al., 2006). In 2015, Sawers and coworkers have challenged this model. They observed that the genetically processed large subunit of E. coli Hyd2 lacks the native NiFe cofactor but forms a complex with the small subunit. The resulting inactive Hyd2 was even accepted by the Tat translocation apparatus and appropriately inserted into the cytoplasmic membrane (Thomas et al., 2015). Therefore, we analyzed the catalytic activity and the cofactor content of the MBH proc purified from the membrane fraction of R. eutropha HP9 and compared the results with those of R. eutropha HP3, synthesizing native MBH.
The MBH yield from R. eutropha HP9 and HP3 was (115 ± 28) µg and (208 ± 11) µg, respectively, of protein per gram of cells (wet weight). Thus, membranes of strain HP9 had a ~45% lower MBH content than strain HP3, which is in line with the Western blot results (Figure 2b and c). Both MBH versions showed, however, identical specific activities for H 2 -mediated methylene blue reduction, with (87.8 ± 3.4) U/mg for native MBH and (88.5 ± 4.7) U/mg for MBH proc . Thus, despite the genetic removal of the C-terminal extension, the NiFe active site of MBH proc seemed to be correctly assembled.
To investigate the integrity of the active site further, we performed infrared (IR) spectroscopy, which probes the C≡O and C≡N stretching vibrations associated with the CO and CNligands of the NiFe site. These vibrations are very sensitive to structural and redox modifications of the active site (Bagley, Duin, Roseboom, Albracht, & Woodruff, 1995). The resulting IR spectra of as-isolated, oxidized native MBH and MBH proc are shown in Figure 3. To obtain quantitative information on the loading of the proteins with the NiFe cofactor, the spectra were normalized based on the intensity of the amide II band, which is proportional to the protein concentration. Both spectra were almost identical to that of aerobically purified MBH (Goris et al., 2011)

| D ISCUSS I ON
The MBH proc of R. eutropha is the first example of a [NiFe]hydrogenase that is equipped with a NiFe catalytic center, although the C-terminal extension of the large subunit was genetically removed. In fact, the purified MBH proc protein was indistinguishable from native MBH with respect to the active site architecture and catalytic activity. Thus, the C-terminal extension of the large subunit is not essential for Hyp protein-mediated insertion of the NiFe cofactor. Nevertheless, removal of the C-terminal extension led to significantly lowered MBH levels in the membrane. Thus, our results indicate that the C-terminal extension optimizes maturation efficiency. Notably, there was no indication for apo-MBH, that is, MBH without NiFe cofactor, in our membrane-derived protein preparation. This is in clear contrast to previous reports for E. coli Hyd-2, where catalytically inactive hydrogenase complexes were identified that contained genetically processed, but NiFe cofactor-free large subunits (Massanz et al., 1997;Senger et al., 2017;Thomas et al., 2015). It has been convincingly shown that the Fe-S cluster-containing hydrogenase small subunit, which is equipped with the Tat leader peptide, becomes transported through the cytoplasmic membrane only in complex with the large subunit (Rodrigue, Chanal, Beck, Müller, & Wu, 1999;Schubert et al., 2007). The results by Thomas et al. suggest that the mere attachment of the large subunit to the small subunit elicits the signal to initialize the Tat cofactor-free hydrogenase complexes may be subjected to proteolysis. The latter mechanism is rather unlikely, because the genetically processed large subunit of the soluble, NAD + -reducing [NiFe]-hydrogenase, SH, of R. eutropha forms a complex with the remaining hydrogenase subunit, although no nickel had been inserted into the active site (Massanz et al., 1997). Thus, the small subunits themselves might reject processed premature large subunits lacking a complete NiFe cofactor, and this capability seems not to be uniformly distributed among all hydrogenases. While the SH small subunit obviously cannot distinguish between processed mature and immature large subunits (Massanz et al., 1997), those of the MBH and the regulatory [NiFe]-hydrogenase (RH) obviously can. In fact, the RH belongs to the subclass of hydrogenases whose apo-large subunits natively lack the C-terminal extension, but are recognized by the canonical Hyp machinery that inserts the canonical NiFe cofactor . Purified RH protein consisted of the iron-sulfur cluster-containing small subunit and large subunit that was stoichiometrically loaded with nickel (Buhrke et al., 2005), indicating an intrinsic proofreading mechanism that prevents the complex formation of premature subunits.
The same seems to be true for MBH proc .
The fact that both HoxG proc and the RH large subunit properly receive the NiFe ( The protein backbones of the large subunits are depicted in blue with red N-termini and magenta C-termini (cartoon representation). The structural models with the PDB codes 4IUC (mature HoxG) (Frielingsdorf et al., 2014) and 5YXY (preform of apo-HyhL (Kwon et al., 2018)) were used. (a) The active site, including the coordinating cysteines and the C-terminal histidine of mature HoxG, are shown as stick models. The C-terminal extension is not visible, as it has been cleaved off. Note that the N-terminus (red) adapts to a β-sheet domain of the main protein, and one of the N-terminal β-strands is located at the position, which is occupied by a β-strand structure of the C-terminal domain in apo-HyhL (b). (b) The C-terminal extension (C ext) of apo-HyhL, which is cleaved off upon NiFe cofactor insertion, is shown in mint. The red N-terminus (the region between the two ends of the red line is structurally unresolved and depicted as a broken line) protrudes from the globular protein and forms a complex with HypA (gray). The four conserved cysteines that coordinate the NiFe(CN) 2 (CO) cofactor are represented as sticks. The terminal histidine residue (shown as sticks), which lies directly in front of the cleavage site of the C-terminal extension, was not resolved and therefore was modelled computationally into the structure using PyMol ( HyhL occupies parts of the position of the N-terminus in the mature structure, forcing the N-terminus in another direction (Kwon et al., 2018). As a consequence, the N-terminus acts like a "crane arm," which brings the HypA protein close to the active site cavity where it can deliver the nickel ion ( Figure 4). Nickel incorporation presumably leads to a dramatic conformational change in the C-terminal extension that upon cleavage moves close to the active site cavity, which unblocks the dedicated position of the N-terminus of the mature protein (Kwon et al., 2018). Therefore, a major role of the C-terminal extension might be providing indirectly the N-terminus with sufficient flexibility to interact with the Hyp machinery. A role of the C-terminal extension in enhancing structural flexibility to facilitate the interaction with Hyp proteins has also been proposed (Albareda, Pacios, & Palacios, 2019;Pinske et al., 2019). A function as a maturation facilitator would also explain why the removal of the C-terminal extension does not necessarily lead to immature hydrogenase. In case of R. eutropha MBH proc , the assembly process seemed to be just less efficient, resulting in a reduced amount of fully active MBH in the cytoplasmic membrane. It should be mentioned, however, that the genetic removal of the extension might result in reduced stability of the large subunit before it becomes equipped with NiFe(CN) 2 (CO) cofactor and oligomerizes with the small subunit.
To clarify the overall necessity of the C-terminal extension, more large subunits need to be tested for their capacity to tolerate the absence of the C-terminal extension in the course of NiFe cofactor insertion. Ingo Zebger for providing access to the IR spectrometer.

CO N FLI C T O F I NTE R E S T S
None declared.

E TH I C S S TATEM ENT
None required.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data are provided in full in the results section of this paper.