Identi cation and Deletion of The Genes Responsible for Hydrogen Production in Thermoanaerobacter Ethanolicus JW200


 Background

 Thermoanaerobacter ethanolicus produces a considerable amount of ethanol from a range of carbohydrates and is an attractive candidate for applications in bioconversion processes. Due to the coupling of hydrogenase activity with fermentation product distribution, understanding hydrogen production of T. ethanolicus, particularly the genes responsible, is valuable for metabolic engineering of the species.
Results

Utilizing the hydrogenases reported in Thermoanaerobacterium saccharolyticum and Pyrococcus furiosus as templates, BLAST search identified five hydrogenase gene clusters, including two membrane-bound [NiFe] hydrogenases ech and mbh, two cytoplasmic [FeFe] hydrogenases hyd and hydII, and one cytoplasmic [NiFe] hydrogenase shi. The combined deletion of ech, mbh, shi and hydG resulted in a strain that did not produce hydrogen and showed no methyl viologen hydrogenase activity in cell extracts. Strains with deletions of all the hydrogenases except one showed normal hydrogen production. Methyl viologen hydrogenase activity was greatly reduced in all combined deletion strains except the strain with an intact hydG gene.
Conclusion

High hydrogen production and hydrogenase activities have been observed for T. ethanolicus. Five hydrogenases have been identified. Hydrogen production was eliminated by deleting genes required for all five hydrogenases. Each individual hydrogenase was verified to be capable of producing hydrogen during fermentation, indicating a high degree of redundancy and flexibility in the hydrogenase systems of T. ethanolicus. A large portion of hydrogenase activity is encoded by the [Fe-Fe] hydrogenases.


Background
Biofuel production from lignocellulosic biomass is a signi cant constituent of the global fuel supply in scenarios for a sustainable energy future [1], but improved biocatalysts are needed to reduce the costs of biomass conversion. Thermoanaerobacter ethanolicus is a gram-positive, anaerobic thermophilic bacterium that produces ethanol as the primary fermentation product from a wide range of polymeric and soluble carbohydrates [2] and is of interest for bioconversion processes [3]. It is also of interest as a co-culture companion for use with cellulolytic thermophilic microorganisms such as Clostridium thermocellum in one-step consolidated bioprocessing [4][5][6]. Due to the coupling of hydrogenase activity with fermentation product distribution, manipulation of hydrogenase activity has been identi ed as a method to direct metabolic ux to ethanol in Thermoanaerobacterium saccharolyticum [7]. Thus, understanding hydrogen production of T. ethanolicus, particularly the genes responsible, is valuable for further metabolic engineering of the species.
Hyrdrogenases are broadly classi ed according to the metal cofactors of their active sites as [Fe], [FeFe], and [NiFe] hydrogenases, but have deep evolutionary origins and ful ll a wide array of metabolic and energetic roles for bacteria and archea [8,9]. Depending on enzyme and cofactor properties as well as reactant concentrations, the hydrogenase reaction can potentially proceed in either direction. In the case of sugar fermenting anaerobes like T. ethanolicus, the reaction is often H 2 generating and proton and electron consuming. Such bacteria use hydrogenases to eliminate excess reduced cofactors such as NAD(P)H or ferredoxin formed during sugar oxidation.
Energy conservation is imperative, so hydrogenase reactions are tightly regulated, and diverse means have evolved for coupling the reaction energy to other cellular processes. For example, many membrane-associated [NiFe] hydrogenases conserve energy by coupling H 2 production with generation of a proton or sodium gradient across the cell membrane. Speci c examples are the Energy Conserving Hydrogenase (Ech) of T. tengcongensis and the Membrane Bound Hydrogenase (Mbh) of Pyrococcus furiosus [10,11]. The ech hydrogenase has a cluster of hyp genes in the same operon [7,10]. The mbh and mbx hydrogenases have a cluster of antiporters [12][13][14]. Hydrogenases also occur in the cytoplasm, where energy is conserved by the coupling of H 2 production and transhydrogenation in the bifurcating [FeFe] hydrogenases [15,16].
Genetic manipulation via gene deletion has been applied to study hydrogenases in thermophilic bacteria. In T. saccharolyticum, [NiFe] hydrogenase ech-hyp and [FeFe] hydrogenase hyd were deleted individually or in combination, and the mutants were characterized with respect to hydrogen production and hydrogenase activity [7]. A [FeFe] hydrogenase with putative sensory function called hfs was responsible for most of the hydrogen production under the conditions tested. Further studies of the four hfs genes showed that deletion of hfsA or hfsB resulted in high ethanol yield [17]. In Clostridium thermocellum, [FeFe] hydrogenase activities were eliminated by deleting the hydrogenase maturase gene hydG and [NiFe] hydrogenase was eliminated by deletion of ech-hyp [18]. In P. furiosus, H 2 production was eliminated by deletion of genes for the membrane bound hydrogenase Mbh and cytoplasmic hydrogenases SHI and SHII [19].
A markerless gene deletion and integration system has been developed for T. ethanolicus JW200 [20]. Three alcohol dehydrogenases have been characterized for their roles in ethanol production via gene deletions [21]. In this study, we identi ed hydrogenases in T. ethanolicus through BLAST searching, and deleted genes individually and in combination to gain insight into the activities and functions of the identi ed enzymes.

Identi cation of membrane-bound [NiFe] and cytoplasmic [FeFe] hydrogenases
To identify potential hydrogenases, a BLAST search of the T. ethanolicus genome was conducted using the hydrogenases reported for T. saccharolyticum [7]. The search identi ed four gene clusters with similarity to ech and hyd genes, as shown in Table 3. No matches were found in T. ethanolicus for the hfs genes of T. saccharolyticum. The genomic organization of the identi ed hydrogenase gene clusters in T. ethanolicus is shown in Figure 1. Based on the genetic grouping and similarity to known hydrogenases, the rst two are likely to be membrane-bound [NiFe] hydrogenases while the last two are likely to be cytoplasmic [FeFe] hydrogenases. The rst gene cluster similar to ech is composed of 12 genes with six genes for ech and six more matching the hyp [NiFe] maturation genes. The second gene cluster similar to T. saccharolyticum ech is composed of seven genes, similar to the mbh genes of P. furiosus, as described below. The ve hyd genes of T. saccharolyticum match ve similar genes in T. ethanolicus, and appear to encode hydABCD, the four-subunit cytoplasmic bifurcating hydrogenase. The gene named hydII in T. saccharolyticum has a match in T. ethanolicus but is the only hydrogenase in its genomic neighborhood. In T. saccharolyticum, hydII is 2 Kbp upstream of hyd.
In addition to hydrogenase genes from T. saccharolyticum, hydrogenase genes from two other species were used as BLAST queries as well. The C. thermocellum hydG maturase gene was used to identify a similar gene (TheetDRAFT_1696) in the T. ethanolicus genome. A BLAST search using the 14-gene mbh operon from P. furiosus identi ed a 13-gene cluster in T. ethanolicus, of which seven genes are the second cluster of ech-like genes previously identi ed using T. saccharolyticum sequences. Immediately upstream of those seven ech-like genes are a cluster of six genes annotated as cation/H+ antiporters. A similar genetic organization occurs for the mbh membrane-bound [NiFe] hydrogenase genes in P. furiosus [13]. Besides mbh, an mbx hydrogenase and two foursubunit [NiFe] Soluble Hydrognases SHI and SHII have been reported in P. furiosus [19]. A BLAST search using SHI and SHII as queries identi ed a single cluster of four genes which are annotated as sul te reductase. The gene con guration of this cluster is shown in Figure 1. The coding sequences for the rst two subunits (A and B) overlap, and in some other species occur as a single gene. Subunit C is an oxidoreductase FAD/NAD(P)-binding domain protein with electron transfer subunit and iron-sulfur cluster binding domain. Subunit D is a 4Fe-4S ferredoxin ironsulfur binding domain-containing protein. The gene cluster was named shi for Soluble Hydrogenase I. It likely encodes an NADPH-linked cystoplasmic hydrogenase based on its similarity to the genes in P. furiosus, The SHI and SHII hydrogenases have been well characterized in P. furiosus, and due to in vitro sulfur reductase activity were previously thought to play a role in sulfur metabolism [22]. However, that is no longer the case due to observed down regulation of the corresponding genes when elemental sulfur is present [19]. Our search for hydrogenases in T. ethanolicus led us to P. furiosus when we noticed the smell of hydrogen sul de from spent cultures, but we believe that the shi genes are in fact unrelated to sulfur metabolism.

Deletion of all hydrogenase activities
Strains were constructed with deletions for the genes responsible for hydrogen production in T. ethanolicus (Table  1). All strains carry a deletion of the tdk gene encoding thymidine kinase, which allows counterselection against integrated gene cassettes and generation of markerless mutations [20]. Rather than deleting the hyd and hydII clusters individually, a single deletion was made for hydG, which has been shown to effectively eliminate [FeFe] hydrogenase activity in C. thermocellum [18].
The strain with all four identi ed hydrogenase systems inactivated was named H0. Fermentation products were measured in batch bottle fermentations for H0 and other intermediate strains and compared to the wildtype (WT) strain and the tdk deletion strain (WT_tdk). As shown in Table 4, hydrogen production is eliminated in strain H0, and acetate levels are 10-fold lower than the controls. Lactate was the major product of strain H0 with a mass yield about 84%, while ethanol was much lower than the control, at 7.6 mM vs. 18-34 mM. The shift from ethanol and acetate production to lactate production implies pyruvate is not being e ciently converted to acetyl-coA via pyruvate:ferredoxin oxidoreductase in strain H0. Hydrogen production by ferredoxin-linked hydrogenases results in re-oxidation of reduced ferredoxin [23], which is required by pyruvate:ferredoxin oxidoreductase. There is no pyruvate formate lyase or pyruvate dehydrogenase in T. ethanolicus. Thus, removing hydrogen production may shift the ux of pyruvate from acetyl-coA to lactate.

Analysis of strains with one intact hydrogenase
Fermentation pro les of intermediate strains with three out of four hydrogenase systems deleted are also listed in Table 4. These strains retain only one hydrogenase while the other three are deleted. As shown in Table 4, strains H1, H2, H3, and H4 retain only ech, mbh, hyd, and shi, respectively. All these strains produce signi cant amount of hydrogen, suggesting that all four hydrogenases are potentially active in WT and that any of them can compensate for deletion of the others. Production of ethanol, lactate, acetate, and hydrogen is similar for H1, H2, and H4, while H3 produces about 20% less ethanol and more organic acid and hydrogen. Relative to WT, the intermediate strains and WT_tdk showed lower lactate and higher ethanol levels. It is unknown why the fermentation products of WT_tdk differ somewhat from WT, but wide variation in the fermentation product pro le was observed in the original species description of T. ethanolicus [2].

Methyl viologen hydrogenase activity
During efforts to create an ethanologenic strain of T. ethanolicus, the authors noticed high in vitro hydrogenase activity, up to 50 times greater than that of Thermoanaerobacterium saccharolyticum reported previously [7]. Methyl viologen (MV) hydrogenase activity was assayed with cleared lysate cell extracts for strains H0, H1, H2, H3, H4, WT, and WT_tdk. MV acts as a universal electron acceptor-donor and can interact with hydrogenases that have either NAD(P)H or ferredoxin as a natural substrate [25]. Strain H0, which does not produce hydrogen, exhibited a small background hydrogenase activity. Strain H3, producing the highest amount of hydrogen, shows the highest hydrogenase activity among the mutants (H1, H2, H3, H4) with only one hydrogenase. However, with around 10% higher hydrogen production for H3 compared to the other three strains, its hydrogenase activity is 10 times higher.
Although it is expected that the majority of enzymatic activity present in whole cells is also present in the cleared lysates, it is possible that enzymatic activity present in the membrane fraction (H1 and H2) could be underrepresented [7]. For strains WT and WT_tdk, they both have all the hydrogenases and produce comparable amount of hydrogen. However, the WT strain shows a hydrogenase activity about ve times that of the WT_tdk strain. The WT strain gives twice the hydrogenase activity of H3, while the WT_tdk strain has 40% hydrogenase activity of H3. Nevertheless, WT, WT_tdk, and H3 produce almost the same amount of hydrogen. High hydrogenase activity does not correlate to high hydrogen production in T. ethanolicus, which has also been reported for T. saccharolyticum [7]. The hydrogenases in T. ethanolicus, all capable of producing hydrogen, seem to be at standby mode to balance redox reactions for pyruvate metabolism to acetyl-CoA. When fermentation goes in the direction of ethanol production, a small amount of hydrogen is produced. When fermentation goes in the direction of organic acids, more hydrogen is produced. This might explain why variations in product ratios occur in the fermentation of wild-type T. ethanolicus [2,20].

Conclusion
High hydrogen production and hydrogenase activities have been observed for T. ethanolicus. Five hydrogenases have been identi ed by sequence analysis including three [NiFe] hydrogenases and two [FeFe] hydrogenases. Hydrogen production was eliminated by deleting genes required for all ve hydrogenases. With the two [FeFe] hydrogenases grouped as one and their activities removed by deleting the maturase gene hydG, each individual hydrogenase was veri ed to be capable of producing hydrogen during fermentation, indicating a high degree of redundancy and exibility in the hydrogenase systems of T. ethanolicus. A large portion of hydrogenase activity is encoded by the [Fe-Fe] hydrogenases.

Strains and culturing conditions
Thermoanaerobacter ethanolicus JW200 (ATCC 31550) was obtained from ATCC. Mutant strains constructed in this study are listed in Table 1. The strains were cultured in CTFUD medium [26] with or without 0.8% (w/v) agar with an initial pH of 7 at 65 o C.

Construction of mutant strains
DNA fragments were ampli ed by PCR using the primers listed in Table 2, then puri ed by gel electrophoresis. Construction of vectors, transformation, and mutant selection were performed according to a markerless gene deletion and integration system reported previously (Shao et al., 2016). Gene deletion PCR products were ampli ed directly from the Gibson Assembly mixture using primers p29 and p30. The PCR products were then column puri ed and transformed into target strains.

Measurement of fermentation products
Wild-type and the mutant strains were cultured in CTFUD medium with cellobiose at an initial concentration of 5 g/L in serum bottles sealed with butyl rubber stoppers. The reaction volume was 10 mL with 27 mL headspace lled with ultra-pure nitrogen. Inoculum prepared in the same medium was added at 1% (v/v). The serum bottles were incubated at 65 o C for two days in a shaking incubator (Innova 4080, New Brunswick Scienti c, Edison, NJ) with a rotation speed of 200 rpm. Samples were taken for measurement of product concentrations. Liquid-phase fermentation products were measured using HPLC using an Aminex HPX-87H column (Bio-Rad, Hercules, CA) at 60°C, with RI (refractive index) detection and a 5-mM sulfuric acid solution eluent at a ow rate of 0.6 ml/min.
Hydrogen was measured using gas chromatography using an SRI 310C gas chromatograph with a HayeSep D packed column using a thermal conductivity detector and nitrogen carrier gas at a ow rate of 8.2 ml/min.

Hydrogenase activity assays
Wild-type and the mutant strains were cultured under the same conditions as for measurement of fermentation products, but cells were harvested in the exponential phase of growth. The procedures for preparation of cleared lysate extracts and methyl viologen-based hydrogenase activity assays were as reported previously (Shaw et al., 2009). One unit of enzymatic activity equals to one µmol of product formed per minute per mg of crude cell extract protein.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.
Competing interests CDH is supported by Enchi Corporation, and LRL is supported by and has a nancial interest in Enchi, which is a for-pro t company that seeks to commercialize C-CBP technology.      Figure 1 Genomic organization of hydrogenase gene clusters in T. ethanolicus