NADP⁺ or CO₂ reduction by frhAGB-encoded hydrogenase through interaction with formate dehydrogenase 3 in the hyperthermophilic archaeon Thermococcus onnurineus NA1

ABSTRACT

It has been reported that the frhAGB-encoded hydrogenase from Thermococcus onnurineus NA1 is homologous to the F₄₂₀-reducing hydrogenase from methanogens and can reduce thioredoxin reductase (TrxR) via direct electron transfer from H₂ oxidation. In this study, to find other interaction targets of frhAGB-encoded hydrogenase, we searched for structural homologs of TrxR using the Position-Specific Iterative Basic Local Alignment Search Tool (PSI-BLAST) protein database search program and AlphaFold Protein Structure Database. Fdh3B (TON_0542), a subunit of the formate dehydrogenase 3 (Fdh3), showed the most similar structure to TrxR in its domains. The interaction potential of Fdh3B with frhAGB-encoded hydrogenase was demonstrated by measuring H₂-dependent NADP⁺ reduction by a mixture of purified proteins. Similarly, the H₂-dependent NADP⁺-reducing activity of Fdh3 whole complex and two other TrxR homologs encoded by TON_0702 and TON_1376 was determined. It was also demonstrated that the Fdh3 complex could reduce CO₂ to formate in the presence of the frhAGB-encoded hydrogenase and H₂, as shown for the H₂-dependent CO₂ reductase (HDCR) of acetogen. The in vivo contribution of the frhAGB-encoded hydrogenase and Fdh3 to H₂-dependent CO₂ reduction was investigated using a resting cell assay, and formate production was significantly decreased in fdh3A or frhA deletion mutants. In conclusion, the frhAGB-encoded hydrogenase was identified to directly transfer electrons to Fdh3 without an electron carrier, similar to the mechanism observed for TrxR. These results expand our understanding of the functional role of the frhAGB-encoded hydrogenase in non-methanogenic Thermococcus species.

IMPORTANCE

The strategy using structural homology with the help of structure prediction by AlphaFold was very successful in finding potential targets for the frhAGB-encoded hydrogenase of Thermococcus onnurineus NA1. The finding that the hydrogenase can interact with FdhB to reduce the cofactor NAD(P)⁺ is significant in that the enzyme can function to supply reducing equivalents, just as F₄₂₀-reducing hydrogenases in methanogens use coenzyme F₄₂₀ as an electron carrier. Additionally, it was identified that T. onnurineus NA1 could produce formate from H₂ and CO₂ by the concerted action of frhAGB-encoded hydrogenase and formate dehydrogenase Fdh3.

INTRODUCTION

Strains of the order Thermococcales, known as hyperthermophilic heterotrophs, can survive on a variety of peptides or carbohydrates, producing H₂S in the presence of sulfur or H₂ in the absence of sulfur (1). [NiFe]-hydrogenases containing orthologs of cytoplasmic SulfI hydrogenase and membrane-bound hydrogenase (Mbh) are commonly found in Thermococcales strains, including Pyrococcus furiosus (2), Pyrococcus horikoshii (3), Thermococcus kodakarensis KOD1 (4), Thermococcus onnurineus NA1 (5, 6), and Thermococcus gammatolerans (7). Some strains, including T. onnurineus NA1 (6), Thermococcus barophilus Ch5 (8), and Thermococcus paralvinellae (9), have additional [NiFe]-hydrogenases such as membrane-bound CO-dependent hydrogenase (Mch), membrane-bound formate-dependent hydrogenase (Mfh) (10), and cytoplasmic frhAGB-encoded hydrogenase (FrhAGB) (10). Mch and Mfh were identified to play key roles in energy conservation from CO and formate utilization in T. onnurineus NA1, respectively (11, 12).

F₄₂₀-reducing hydrogenase (FrhABG) from methanogens reduces coenzyme F₄₂₀, which supplies four electrons for CO₂ reduction and plays an essential role during methanogenesis (13, 14). The frhAGB-encoded hydrogenase in Thermococcales strains was assumed to have a unique role distinct from FrhABG from methanogens. The frhAGB-encoded hydrogenase of T. onnurineus NA1 has a large subunit FrhA with a binuclear [NiFe]-center, a small subunit FrhG with three [4Fe4S] clusters, and a flavoprotein FrhB with a [4Fe4S] cluster and FAD-binding site, as found in F₄₂₀-reducing hydrogenase from methanogens (14–16). However, the F₄₂₀-binding site of FrhB subunit of methanogens is not well conserved in the corresponding subunit of frhAGB-encoded hydrogenase, and F₄₂₀-reducing activity was not detected in the purified frhAGB-encoded hydrogenase (17). The co-immunoprecipitation assay and affinity pull-down assay suggested that the frhAGB-encoded hydrogenase could interact with thioredoxin reductase (TrxR). The redox cascade from the hydrogenase down to a protein disulfide oxidoreductase (Pdo) followed by a sulfur response regulator (SurR) was identified in T. onnurineus NA1 (18, 19). Deletion or overexpression of the frhAGB-encoded hydrogenase gene showed distinct phenotypes, indicating that the enzyme may be involved in the transcriptional regulation of various genes (20, 21). However, the mechanism of action of the frhAGB-encoded hydrogenase is not yet clearly understood.

In this study, we searched for interaction partners other than TrxR to understand better the function of frhAGB-encoded hydrogenase in T. onnurineus NA1. Sequence- and structure-based homologs of TrxR were targeted as interaction partners of the frhAGB-encoded hydrogenase, and three candidates (Fdh3B, TON_0702, and TON_1376) were selected through homology search and artificial intelligence (AI)-assisted structural modeling. The interaction and direct electron transfer between frhAGB-encoded hydrogenase and TrxR homologs were experimentally evaluated. The significance of the activity of the frhAGB-encoded hydrogenase through interaction with Fdh3 was investigated in an in vivo gene deletion experiment.

RESULTS

In silico prediction of interaction partners for frhAGB-encoded hydrogenase

To search for other interaction partners of the frhAGB-encoded hydrogenase, homologs of thioredoxin reductase (TrxR, TON_1603) were analyzed using the Position-Specific Iterative Basic Local Alignment Search Tool (PSI-BLAST) (Table 1). Eleven proteins were selected based on the E-values. Homologous proteins had very low sequence similarity with 23%–33% identity ranges. Multiple sequence alignment of these homologs showed that the active site (CXXC) of TrxR was not present in the other homologs but that the flavin adenine dinucleotide (FAD)-binding or nicotinamide adenine dinucleotide phosphate (NADPH)-binding residues are well conserved (Fig. S1) (22, 23). To determine the structural similarities between TrxR and its sequence homologs, structural models were retrieved from the AlphaFold Protein Structure Database (24). First of all, the structural model of TrxR of T. onnurineus NA1 was found to be highly similar to the TrxR structures of Bacillus cereus and Sulfolobus solfataricus (25, 26), with root-mean-square deviation (RMSD) values (27) of less than 1 Å for FAD and NADPH-binding domains. Then, the structural similarity between TrxR and its homologs was compared based on the RMSD values for the two domains (Fig. 1). TON_0542 was predicted to be most similar to TrxR among the homologs, an NADPH domain with an RMSD value of 0.601 Å and a FAD domain with an RMSD value of 1.545 Å. Several other gene products encoded by TON_0702, TON_1336, TON_0057, TON_0129, TON_0305, and TON_1265 were also similar to TrxR with RMSD values less than 2 Å. However, TON_1376, TON_0204, TON_0865, and TON_1271 were similar to TrxR in one of the FAD or NADPH domains, but less similar in the other. Therefore, TON_0542, TON_0702, and TON_1376 were selected as candidates to test their interaction with the frhAGB-encoded hydrogenase.

TABLE 1.

Thioredoxin reductase (TrxR, TON_1603) homologs from T. onnurineus NA1 based on PSI-BLAST analysis

Locus tag	Annotation	Coverage (%)	E-value	Identity (%)	Accession numbera
TON_0129	FAD-dependent oxidoreductase	84	6e−25	31.74	WP_012571087.1
TON_1336	NADPH-dependent glutamate synthase	96	5e−22	29.34	WP_012572298.1
TON_1265	FAD-dependent oxidoreductase	92	4e−16	27.38	WP_012572225.1
TON_1376	NADPH-dependent glutamate synthase	90	e −15	27.63	WP_012572338.1
TON_0057	NADPH-dependent glutamate synthase	93	e −12	27.25	WP_012571014.1
TON_0204	FAD-dependent oxidoreductase	93	3e−12	29.25	WP_048055134.1
TON_0542	FAD-dependent oxidoreductase (Fdh3B)	48	2e−11	32.79	WP_012571501.1
TON_0305	CoA-disulfide reductase	72	3e−10	28.35	WP_012571262.1
TON_1271	FAD-dependent oxidoreductase	86	e −6	23.03	WP_012572231.1
TON_0702	FAD-dependent oxidoreductase	94	2e−5	28.57	WP_012571662.1
TON_0865	FAD-dependent oxidoreductase	58	2e−5	27.59	WP_012571825.1

^a GenBank (www.ncbi.nlm.nih.gov/genbank/) accession numbers were used.

Fig 1.

Structure models of three TrxR homologs of T. onnurineus NA1, obtained from the AlphaFold Protein Structure Database. RMSD values for two domains, FAD domain and NADPH domain, were compared with the TrxR (TON_1603) model, and the numbers of calculated atoms are presented.

In vitro experiments on the interaction between frhAGB-encoded hydrogenase and Fdh3

In our previous report, TON_0542 (Fdh3B) was identified as the NAD(P)-binding subunit of the whole Fdh3 complex, which includes a tungsten-containing catalytic subunit (Fdh3A) and two Fe-S proteins (Fdh3G1 and Fdh3G2) (28). Since the Fdh3 complex can reduce NAD(P)⁺ by coupling with formate oxidation, it was assumed that electrons derived from H₂ oxidation by the frhAGB-encoded hydrogenase might be transferred to Fdh3B for NAD(P)⁺ reduction. To evaluate this possibility, in vitro assays were performed. The frhAGB-encoded hydrogenase in the form of a three-subunit complex, FrhAGB, and Fdh3 complex was purified from T. onnurineus NA1, and Fdh3B subunit alone was heterologously expressed and purified in Escherichia coli (Fig. S2). Fdh3B mixed with FrhAGB clearly displayed NADP⁺-reducing activity in the presence of H₂ (Fig. 2A), and the activity was measured to be 25.5±14.4 U/µmol. The H₂-dependent NADP⁺-reducing activity of the reaction mixture of FrhAGB and Fdh3 (211.1±36.2 U/µmol) (Fig. 2D) was eightfold higher than that of FrhAGB and Fdh3B (25.5±14.4 U/µmol), which is consistent with the pull-down assay results for FrhAGB and Fdh3 (Fig. S3). In addition to Fdh3B, other subunits of Fdh3 appear to contribute to the interaction with FrhAGB. To investigate whether other TrxR homologs interact with FrhAGB, the proteins encoded by TON_0702 and TON_1376 were additionally purified (Fig. S2), and their activities were determined (Fig. 2B and C). Despite their structural similarities to Fdh3B (Fig. 1), the H₂-dependent NADP⁺-reducing activities of TON_0702 (2.4±0.02 U/µmol) and TON_1376 (3.1±0.17 U/µmol) with FrhAGB were estimated to be about 10-fold lower than that of Fdh3B. Nevertheless, the frhAGB-encoded hydrogenase can clearly transfer electrons from H₂ to the three TrxR homologs.

Fig 2.

H₂-dependent NADP⁺ reduction by interactions between FrhAGB and TrxR homologs, Fdh3B (TON_0542) (A), TON_0702 (B), TON_1376 (C), and Fdh3 complex (D). Each line represents FrhAGB alone purged with H₂ (yellow), TrxR homolog alone purged with H₂ (blue), TrxR homolog alone purged with N₂ (green), FrhAGB and TrxR homolog purged with H₂ (red), and FrhAGB and TrxR homolog purged with N₂ (black).

Previously, it was also reported that Fdh3 of T. onnurineus NA1 could reduce CO₂ with NADPH or reduced ferredoxin (28). Therefore, we wondered if Fdh3 interacting with FrhAGB could reduce CO₂ through the direct transfer of electrons derived from H₂ oxidation by FrhAGB. To address this issue, FrhAGB and Fdh3 were mixed in the presence of H₂/CO₂ (80:20, vol/vol) without an electron mediator, and formate production was measured. As a result, formate was detected in the reaction mixture of FrhAGB and Fdh3 (Fig. 3). Since the FrhB subunit of FrhAGB was not absolutely required for TrxR interaction (19), formate production was also tested with a combination of FrhAG and Fdh3. Clearly, the reaction mixture of FrhAG and Fdh3 could generate formate. Comparing the formate-producing activity, the initial activity measured after 1-hour incubation was very similar. FrhAGB-Fdh3 and FrhAG-Fdh3 produced 0.357±0.174 mmol/L/h and 0.338±0.031 mmol/L/h of formate, respectively. However, they showed some differences over time. The FrhAG–Fdh3 showed 43.5% less formate-producing activity than FrhAGB-Fdh3 at 2 hours.

Fig 3.

H₂-dependent CO₂ reduction by interaction between FrhAG(B) and Fdh3. FrhAGB and Fdh3 purged with H₂/CO₂ (▲); FrhAGB and Fdh3 purged with N₂ (△); FrhAG and Fdh3 purged with H₂/CO₂ (◆); FrhAG and Fdh3 purged with N₂ (◇); Fdh3 purged with H₂/CO₂ (○); FrhAGB purged with H₂/CO₂ (■); FrhAG purged with H₂/CO₂ (□); blank purged with H₂/CO₂ (●).

Table 2 shows the electron transfer activity of the combination of FrhAGB and Fdh3 (or Fdh3B) compared to other enzymes. It is worth noting that the activity appeared at a similar rate to the Pdo reduction by FrhAGB and TrxR (19). The H₂-dependent NADP⁺-reducing activity of the combination of FrhAGB and Fdh3 (or Fdh3B) was significantly lower than those of SulfI hydrogenases from T. kodakarensis or P. furiosus, but comparable to that of SulfII hydrogenase from P. furiosus (29, 30). The H₂-dependent CO₂-reducing activity of FrhAGB–Fdh3 was shown to be similar to that of formate hydrogenlyase (FHL) from E. coli (31), but significantly lower than those of other CO₂-reducing enzymes such as HDCR from Thermoanaerobacter kivui or Acetobacterium woodii and electron-bifurcating NADP- and ferredoxin-dependent [FeFe] hydrogenase in complex with formate dehydrogenase (Hyt-Fdh) from Clostridium autoethanogenum (32, 33).

TABLE 2.

Comparison of H₂-dependent NADP⁺- or CO₂-reducing activity

Electron acceptor	Hydrogenase	Organisma	Specific activity (Ub/mg)	Reference
NADP+	FrhAGB and Fdh3	T. onnurineus NA1	0.78 ± 0.1	This study
	FrhAGB and Fdh3B	T. onnurineus NA1	0.19 ± 0.1	This study
	FrhAGB and TON_0702	T. onnurineus NA1	0.018 ± 0.001	This study
	FrhAGB and TON_1376	T. onnurineus NA1	0.021 ± 0.001	This study
	SulfI	T. kodakarensis KOD1	48.7	(29)
	SulfI	P. furiosus	75	(30)
	SulfII	P. furiosus	0.3	(30)
CO2	FrhAGB and Fdh3	T. onnurineus NA1	0.011 ± 0.005	This study
	HDCR	Th. kivui	900	(32)
	HDCR	A. woodii	10	(32)
	FHL	E. coli	0.03	(31)
	Hyt-Fdh	C. autoethanogenum	41	(33)
Pdo	FrhAGB and TrxR	T. onnurineus NA1	0.011	(19)

^a T, Thermococcus; P, Pyrococcus; Th, Thermoanaerobacter; A, Acetobacterium; E, Escherichia; C, Clostridium.

^b One unit (U) is equivalent to the transfer of 2 μmol of electrons per minute.

In vivo assay for H₂-dependent CO₂-reducing activity

To determine whether FrhAGB and Fdh3 are involved in the H₂-dependent CO₂-reducing activity of T. onnurineus NA1 in vivo, deletion mutants of the frhA or fdh3A genes were constructed. The H₂-dependent CO₂-reducing activities of wild-type strain and two mutants were measured using resting cells. Whereas the wild-type strain produced 1.42±0.12 mmol/L/h of formate, the ΔfrhA and Δfdh3A mutants produced 0.13±0.03 mmol/L/h and 0.8±0.03 mmol/L/h, respectively, indicating that FrhAGB and Fdh3 might contribute to formate production in T. onnurineus NA1 (Fig. 4). More experiments are needed to determine whether this result is due to the actual interaction between frhAGB-encoded hydrogenase and Fdh3 in vivo.

Fig 4.

H₂-dependent CO₂ reduction in the wild-type and mutant strains. The genotypes of mutant strains are described in Table 3. Formate production was quantified after incubating resting cells at 80°C for 1 hour (upper panel). Expression of Fdh3A and FrhA proteins was detected by Western blot (lower panels).

DISCUSSION

A number of biochemical or genetic methods have been used for screening protein interactions, such as co-immunoprecipitation, tandem affinity purification, chemical cross-linking, two-hybrid screen, and phage display (34, 35). In our previous study, the interaction between frhAGB-encoded hydrogenase and TrxR was identified via co-immunoprecipitation, and direct electron transfer between the proteins could be demonstrated in T. onnurineus NA1 (19). In this study, we tried to find a new target protein for direct electron transfer to broaden our understanding of the function of frhAGB-encoded hydrogenase. If the protein structures of frhAGB-encoded hydrogenase and TrxR had been determined through X-ray crystallography, nuclear magnetic resonance (NMR), or cryo-electron microscopy (cryo-EM), the interaction partner of frhAGB-encoded hydrogenase could have been easily predicted using computational methods such as protein–protein docking, based entirely on three-dimensional protein structures (36, 37). However, their protein structures have not yet been determined. Therefore, we envisioned that proteins with TrxR-mimetic structures could interact with frhAGB-encoded hydrogenase. Two sequential analyses were attempted to select amino acid sequence homologs via PSI-BLAST and then structure homologs via AlphaFold (24, 38). Our simple premise has allowed us to find new combinations of hydrogenase and NADP⁺-reducing protein either alone (Fdh3B) or as a subunit of formate dehydrogenase (Fdh3).

Our previous report showed that frhAGB-encoded hydrogenase can interact with TrxR and directly transfer electrons from H₂ to TrxR and that the TrxR/Pdo redox cascade can reduce SurR, a sulfur-responsive transcription factor (Fig. 5), suggesting a new transcriptional regulatory role for the frhAGB-encoded hydrogenase in T. onnurineus NA1 (18, 19). Transcriptome analysis of hydrogenase gene deletion or overexpression mutants supported the possibility (20, 21). Nevertheless, many questions remained about the role of frhAGB-encoded hydrogenase in vivo, such as whether the redox cascade of the hydrogenase is restricted to the TrxR/Pdo/SurR cascade. In this regard, novel combinations of the hydrogenase and NADP⁺-reducing proteins may provide clues to the role of the hydrogenase. In particular, it is intriguing to note that direct electron transfer from frhAGB-encoded hydrogenase to Fdh3B or Fdh3 resulted in NAD(P)⁺ reduction or CO₂ reduction, respectively, whereas frhAGB-encoded hydrogenase itself did not show NAD(P)⁺ or CO₂ reduction (17 and Fig. 2). Based on transcriptomic and proteomic analysis data, frhAGB-encoded hydrogenase is constitutively expressed in a variety of culture media including formate medium in T. onnurineus NA1 (39–41). In addition, it seems that there might be a correlation between the presence of frhAGB-encoded hydrogenase and formate dehydrogenases in many Thermococcus strains (42), suggesting that frhAGB-encoded hydrogenase probably plays an important role along with formate hydrogenase. For example, the interaction and electron relay between frhAGB-encoded hydrogenase and Fdh3 whole complex suggest that a role for the hydrogenase in reducing-equivalent disposal seems possible. During the heterotrophic growth of Thermococcales strains, as with most anaerobes, reducing equivalent disposal is a key issue for sustained growth. Thermococcales strains mostly use elemental sulfur (S⁰) or proton as electron acceptors, producing H₂S or H₂, respectively (1). Mbh plays a key role in recycling reduced ferredoxins generated from carbohydrate or peptide breakdown (43–45). CO-dependent membrane-bound hydrogenase (CODH-Mch-Mnh3) and formate-dependent membrane-bound hydrogenase (Fdh2-Mfh2-Mnh2) also play important roles in carboxydotrophic and formatotrophic growth, respectively (11, 12, 46). Since energy conservation by these enzymes correlates with H₂ production, excessive accumulation of H₂ gradually inhibits the use of protons as an electron acceptor thermodynamically (47, 48). In situations where the H₂ partial pressure is high and electron acceptors are limited, H₂-dependent CO₂-reducing activity to ameliorate H₂ partial pressure will be undoubtedly beneficial for cell growth. The decrease in H₂-dependent CO₂-reducing activity in ΔfrhA or Δfdh3A mutants indicates that frhAGB-encoded hydrogenase and Fdh3 are actually involved in formate generation in vivo. However, formate generation was still detected in Δfdh3A or ΔfrhA mutants. So, considering redundant gene copies encoding formate dehydrogenase and hydrogenase in T. onnurineus NA1, it is worth noting that the combination of other FDHs (Fdh1 or Fdh2) and hydrogenases might be also involved in H₂-dependent CO₂-reducing activity (5).

Fig 5.

Proposed model for direct electron transfer between FrhAGB and its interaction partners.

The H₂-dependent CO₂-reducing activity of T. onnurineus NA1 determined in this study was much lower than those of the H₂-dependent CO₂-reducing enzymes. Among the known enzymes, HDCR is composed of four subunits: formate dehydrogenase, hydrogenase, and two iron–sulfur (Fe-S) cluster proteins, and catalyzes reversible H₂-dependent CO₂ reduction to formate by electron transfer between the two catalytic subunits (32, 49). The enzyme of Th. kivui is a highly active enzyme that reduces CO₂ to formate with a specific activity of 900 U/mg and produces H₂ and CO₂ from formate with a specific activity of 930 U/mg (32). The Hyt-Fdh from C. autoethanogenum catalyzes H₂-dependent CO₂ reduction and also NADPH- and ferredoxin-dependent confurcating CO₂ reduction to formate (33). Recently, FHL-1 from E. coli, consisting of soluble formate dehydrogenase (FDH-H), two Fe-S cluster proteins, and a nickel-dependent hydrogenase module (Hyd-3), has been shown to work in reverse and exhibit H₂-dependent CO₂ reductase activity (31, 50). Structural features of these enzymes lead us to wonder how FrhAGB transfers electrons to Fdh3 for CO₂ reduction, as the interaction between FrhAGB and Fdh3B does not necessarily mean that Fdh3 can only interact with FrhAGB through Fdh3B subunit. There may be other routes for electron transfer. The question of how FrhAGB and Fdh3 can interact together or how electrons can be transferred from FrhAGB to Fdh3 awaits further study.

It is worth noting that SulfI hydrogenase is known to reduce NADP⁺ in T. kodakarensis KOD1 and P. furiosus (29, 30). Their homologs, SulfI (TON_0534 to 0537) and SulfII (TON_0052 to 0055), are found in T. onnurineus NA1. The NADP⁺-reducing activity of FrhAGB and Fdh3 is much less than that of SulfI. Therefore, it is unlikely that FrhAGB and Fdh3 play a major role in NADP⁺ reduction in the presence of SulfI or SulfII. However, SulfI, SulfII, and FrhAGB showed different expression patterns depending on the growth substrate (39, 40). Therefore, it is assumed that FrhAGB and Fdh3 may provide redundancy in reducing NADP⁺, especially in the absence of SulfI and SulfII.

It is not easy to pinpoint the reason for the increase in absorbance above baseline at 340 nm using FrhAGB and Fdh3 purged with N₂ (black line) or Fdh3 alone (blue line) in Fig. 2. H₂ present in the anaerobic chamber may be dissolved in the buffer during the experiment, or the purified Fdh3 may be contaminated with trace amounts of hydrogenase. Nevertheless, it was very clear that NADP⁺ reduction activity was maintained for longer in the presence of H₂ and was higher when FrhAGB was added.

In this study, efforts were made to find more interaction targets for electron transfer by frhAGB-encoded hydrogenase. Fdh3B, TON_0702, and TON_1376 were selected among TrxR homologs, and their H₂-dependent NADP⁺-reducing activity was assayed. As a result of the in vitro and in vivo assay, new functions of FrhAGB were identified: NADP⁺ reduction and CO₂ reduction through protein–protein interaction with Fdh3B and its whole complex, Fdh3. Other TrxR homologs may be direct targets of FrhAGB or intermediates in the redox cascade. Further studies will expand our understanding of the functional role of the frhAGB-encoded hydrogenase in T. onnurineus NA1.

MATERIALS AND METHODS

Culture media

T. onnurineus NA1 wild-type and mutant strains were cultivated as previously described with slight modifications (11). The MM1 medium routinely used in this study contained 1 g yeast extract, 35 g NaCl, 0.7 g KCl, 3.9 g MgSO₄, 0.4 g CaCl₂∙H₂O, 0.3 g NH₄Cl, 0.15 g Na₂HPO₄, 0.03 g Na₂SiO₃, 0.5 g NaHCO₃, 0.5 g cysteine-HCl, 1 mL of trace element solution, 2 mL of 500X Fe-EDTA solution (51), 1 mL of vitamin solution (52), and 1 mL of 5% (wt/vol) Na₂S∙9H₂O solution per liter. For protein purification, 5 g/L of maltodextrin and 0.1 M Bis-Tris (pH 6.5) were added to the MM1 medium, and yeast extract was increased to 4 g/L supplemented. To prepare cell suspensions, 5 g/L of sodium pyruvate was added to the MM1 medium.

Construction of mutants

All mutants were created by applying the gene disruption and gene introduction system as previously described with slight modifications (53). The 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase gene from P. furiosus (hmg_pfu) was integrated into the genome with a strong promoter of TON_0157 (P₀₁₅₇) as a selection marker.

The TON_0542, TON_0702, and TON_1376 genes were amplified from the genomic DNA of T. onnurineus NA1 and then inserted into the pET-28a(+) vector (Merck KgaA, Darmstadt, Germany). Each recombinant plasmid was transformed into the E. coli Rosetta (DE3) strain (Merck KgaA, Darmstadt, Germany) to obtain strains RTN0542, RTN0702, and RTN1376, respectively. The strains used in this study are summarized in Table 3, and the primers used in this study are given in Table 4.

TABLE 3.

Strains and their genotypes

Strain	Parent strain	Genotype	Reference
Thermococcus onnurineus NA1
Δfdh3A	Wild type	Δfdh3A	This study
ΔfrhA	Wild type	ΔfrhA::bPgdh-chmgpfu	(20)
MF01	DF01	aP0157-hmgpfu-fdh3	(28)
FrhAGB-overexpressing strain	Wild type	P 1559 ::P 0157 -hmg pfu	(19)
FrhAG-overexpressing strain	ΔfrhB strain	P 1559 ::P 0157 -hmg pfu	(19)
Escherichia coli
RTN0542	Rosetta (DE3)	fdh3B gene cloned in pET-28a(+)	This study
RTN0702	Rosetta (DE3)	TON_0702 gene cloned in pET-28a(+)	This study
RTN1376	Rosetta (DE3)	TON_1376 gene cloned in pET-28a(+)	This study

^a P₀₁₅₇, promoter of TON_0157 gene of T. onnurineus NA1.

^b P_gdh, promoter of glutamate dehydrogenase gene of T. kodakarensis KOD1.

^c hmg_pfu, HMG-CoA reductase gene from P. furiosus.

TABLE 4.

Primers used in this study

Mutant	Primer	Sequence (5' → 3')
Δfdh3A	fdh3A_LA_F	AATTCGAGCTCGGTACCCGGCTGGTGCATACATAGCCTTTGGATT
	fdh3A_LA_R	CCAAGCTTGCATGCCTGCAGATACGGCTCTATTCCTATGGGCTGC
	fdh3A_RA_F	CCATAGGAATAGAGCCGTATGGGGCCAATATACTCACGAACGATG
	fdh3A_RA_R	CCAAGCTTGCATGCCTGCAGCAAACCTCATCCCGCTCATACTCAT
RTN0542	pET_TON0542_F	GCGCGGCAGCCATATGAGCGGGATGAGGTTTGCG
	pET_TON0542_R	GCTCGAATTCGGATCCTCACCTCCCCATCAGCCA
RTN0702	pET_TON0702_F	GTGCCGCGCGGCAGCCATATGGTGAGGTTCTACATCTGTGAG
	pET_TON0702_R	GGTGGTGGTGGTGGTGCTCGAGGCACTTACCGCTGAGAACC
RTN1376	pET_TON1376_F	GTGCCGCGCGGCAGCCATATGATGGCGAGAGCCAAACCTAAG
	pET_TON1376_R	GTGGTGGTGGTGGTGCTCGAGAGCGTTCGCCTTCTTGGC

Protein purification

Fdh3 whole complex was purified from strain MF01 (28) by TALON immobilized metal affinity chromatography (IMAC) as previously described. Fdh3B subunit and the proteins encoded by TON_0702 and TON_1376 were purified from strains RTN0542, RTN0702, and RTN1376, respectively. FrhAGB and FrhAG proteins were purified from FrhAGB- and FrhAG-overexpressing strains, respectively, as previously described (19).

H₂-dependent NADP⁺- or CO₂-reducing assay

H₂-dependent NADP⁺-reducing activity was assessed using 1 µM of Fdh3B, TON_0702, or TON_1376 in the presence of 1 µM FrhAGB and 2 mM NADP⁺ dissolved in a buffer containing 50 mM Tris-HCl (pH 8.0), 0.5 M KCl, and 75 mM NaCl. In addition, combinations of 0.2 µM Fdh3 or FrhAGB were estimated with H₂-dependent NADP⁺-reducing activity. Three bars of H₂ gas was filled in a 2 mL sealed cuvette and incubated at 70°C. Absorbance change according to NADPH generation was monitored at 340 nm (ε_NADPH=6.22 mM⁻¹ cm⁻¹) (54) using a UV-VIS spectrophotometer (UV-2600, Shimadzu, Kyoto, Japan). To confirm that the NADP⁺ reduction activity depends on H₂ consumption, the buffer charged with N₂ was tested as a negative control.

The H₂-dependent CO₂-reducing activity of the frhAGB-encoded hydrogenase and Fdh3 was evaluated using 2 µM of each enzyme dissolved in a buffer containing 50 mM Tris-HCl (pH 8.0), 0.5 M KCl, and 75 mM NaCl. After charging 3 bar of H₂/CO₂ (80:20) gases, all serum vials were incubated at 70°C for 1–5 hours. Formate production was quantified using the Formate Assay Kit (ab111748, Abcam, Cambridge, UK).

Resting cell assay

To prepare cell suspensions, the cells were cultivated and harvested by centrifugation at 4,426 × g for 20 min at 10°C in an anaerobic chamber. Cells were washed twice and resuspended to an optical density at 600 nm (OD_{600 nm}) of 1.0 with modified buffer A containing 50 mM imidazole-HCl (pH 6.5), 600 mM NaCl, 30 mM MgCl₂, and 10 mM KCl (55). One milliliter of the cell suspension was placed in a 9 mL serum vial and charged with 2 bar of H₂/CO₂ (80:20) gas. All vials were then incubated at 80°C for 1 hour. Formate production was quantified using the Formate Assay Kit (ab111748, Abcam, Cambridge, UK).

Western blot analysis

Western blot analysis of protein expression was performed as previously described (19) using rabbit polyclonal anti-FrhA or anti-Fdh3A antibodies (Ab Frontier, Seoul, Republic of Korea). Blots were visualized with a Clarity Western ECL substrate (Bio-Rad, Hercules, USA). Chemiluminescent signals were visualized using a ChemiDoc MP imaging system (Bio-Rad, Hercules, USA).

Protein structure models

The sequence homologs of TrxR were determined by the PSI-BLAST at NCBI (https://www.ncbi.nlm.nih.gov/). Multiple sequence alignment of TrxR homologs was performed by Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). Protein structure models were compiled from the AlphaFold Protein Structure Database (https://alphafold.ebi.ac.uk/). Structural visualization and alignment were performed using the PyMOL program (https://pymol.org/2/).

Pull-down assay

Protein–protein interactions were validated using the pull-down approach as described by Jung et al. (19). Briefly, after attaching purified His-tagged Fdh3 or Fdh3B to Dynabeads His-Tag Isolation and Pull-down kit (Invitrogen, Waltham, USA), purified strep-tagged FrhAGB was applied and washed four times. Eluted proteins were analyzed by SDS-PAGE and Western blot by using anti-FrhAGB antibody (Ab Frontier, Seoul, Republic of Korea).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01474-23.

Fig. S1 to S3. aem.01474-23-s0001.docx.

Supplementary Figures.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.