A new metabolic trait in an acetogen: Mixed acid fermentation of fructose in a methylene‐tetrahydrofolate reductase mutant of Acetobacterium woodii

Abstract To inactivate the Wood–Ljungdahl pathway in the acetogenic model bacterium Acetobacterium woodii, the genes metVF encoding two of the subunits of the methylene‐tetrahydrofolate reductase were deleted. As expected, the mutant did not grow on C1 compounds and also not on lactate, ethanol or butanediol. In contrast to a mutant in which the first enzyme of the pathway (hydrogen‐dependent CO2 reductase) had been genetically deleted, cells were able to grow on fructose, albeit with lower rates and yields than the wild‐type. Growth was restored by addition of an external electron sink, glycine betaine + CO2 or caffeate. Resting cells pre‐grown on fructose plus an external electron acceptor fermented fructose to two acetate and four hydrogen, that is, performed hydrogenogenesis. Cells pre‐grown under fermentative conditions on fructose alone redirected carbon and electrons to form lactate, formate, ethanol as well as hydrogen. Apparently, growth on fructose alone induced enzymes for mixed acid fermentation (MAF). Transcriptome analyses revealed enzymes potentially involved in MAF and a quantitative model for MAF from fructose in A. woodii is presented.


INTRODUCTION
The Wood-Ljungdahl pathway (WLP) is one of the seven known pathways for the fixation of carbon dioxide found in nature (Garritano et al., 2022).It is present in strictly anaerobic microorganisms such as methanogenic archaea, sulphate reducing and acetogenic bacteria and catalyses the anabolic formation of acetyl-CoA from two molecules of CO 2 (Schauder et al., 1988;Wolfe, 1993;Wood et al., 1986).In acetogens, the WLP is also involved in catabolism and acetyl-CoA is further converted via acetyl phosphate to acetate, the characteristic and name-giving end product of CO 2 reduction in this group of bacteria (Drake, 1994;Müller, 2003).The pathway is linear with two branches.In the carbonyl branch, one molecule of CO 2 is first reduced to carbon monoxide that is bound to the central enzyme of the pathway, the CO dehydrogenase/acetyl-CoA synthase (CODH/ACS; Pezacka & Wood, 1984;Raybuck et al., 1988).In the methyl branch, one molecule of CO 2 is first reduced to formate by a formate dehydrogenase; in the model species Acetobacterium woodii and Thermoanaerobacter kivui, hydrogen is the electron donor for CO 2 reduction and hence, the enzyme was named hydrogen-dependent CO 2 reductase (HDCR; Dietrich et al., 2022;Schuchmann & Müller, 2013;Schwarz et al., 2018).Formate is then bound to the C1 carrier tetrahydrofolate (THF) giving formyl-THF (Himes & Harmony, 1973;Lovell et al., 1988), water is split off, and the resulting methenyl-THF is reduced via methylene-THF to methyl-THF (Bertsch et al., 2015;Ragsdale & Ljungdahl, 1984).The methyl group is then transferred to the CODH/ACS and combined with CO and CoA to acetyl-CoA (Ragsdale, 2008).
Acetogens are metabolically very versatile and can also grow heterotrophically on sugars, carboxylic acids and aldehydes (Bache & Pfennig, 1981;Drake et al., 2008;Gößner et al., 1994;Moon et al., 2019;Seifritz et al., 1999;Trifunovic et al., 2020;Weghoff et al., 2015).Actually, the WLP was discovered during glucose fermentation in Moorella thermoacetica (formerly Clostridium thermoaceticum; Fontaine et al., 1942).This bacterium oxidises glucose via the Embden-Meyerhoff-Parnas pathway to pyruvate and then further to 2 moles of acetate, 2 moles of CO 2 , and 8 electrons are generated (Drake & Daniel, 2004).In fermentation balances, no other product but a third mole of acetate had been found, and later it was demonstrated that 2 moles of CO 2 were reduced with the eight electrons to a third mole of acetate by a novel pathway, the WLP (Fontaine et al., 1942).This type of fermentation was called homoacetogenesis and proceeds in two steps: The reduction of CO 2 with electrons derived from molecular hydrogen conserves energy in the form of ATP by a respiratory mechanism, but in A. woodii only 0.3 ATP per mol of acetate are formed (Hess, Schuchmann, & Müller, 2013;Matthies et al., 2014); this is only 7.5% of the ATP gain of glycolysis plus pyruvate oxidation.The energetic benefit of using the WLP clearly is that all the acetyl-CoA produced from glucose by glycolysis and pyruvate oxidation can be used for energy-conserving acetate formation.This allows the bacteria to produce the maximum amount of ATP that can be obtained by glycolysis and pyruvate oxidation, 4 ATP/mol glucose.Glucose oxidation to acetate, CO 2 and hydrogen (Equation 5) gives the highest possible ATP yield for a fermentation but is performed only by a few bacteria such as, for example Thermotoga maritima, in pure culture (Schröder et al., 1994): Oxidation of NADH coupled to reduction of protons to hydrogen is highly endergonic and only bacteria that have an electron-bifurcating hydrogenase can overcome this steep energetic barrier (Schut & Adams, 2009).Such an enzyme is also present in A. woodii (Schuchmann & Müller, 2012;Wiechmann et al., 2020), but fermentation of glucose according to Equation 5 has never been observed in pure cultures.In contrast, if the hydrogen concentration is kept low by a syntrophic, methanogenic partner, A. woodii converted fructose according to Equation 5 and the hydrogen and CO 2 produced were used by the syntrophic methanogenic partner to produce methane, demonstrating the presence of the enzymatic machinery to catalyse this fermentation (Winter & Wolfe, 1980).
Electrons could also be disposed by reducing an intermediate of the pathway, that is, by classical fermentation.If intermediates of the pathway such as acetyl-CoA or pyruvate would have to act as electron acceptor, less acetate could be formed and for any acetate not produced, one ATP is lost.Nevertheless, a mixed acid fermentation (MAF) is well known, for example, in Escherichia coli (Xu et al., 1999), but this has never been observed in A. woodii although this model acetogen has all the enzymes required for a MAF, that is, CoA-dependent acetaldehyde and alcohol dehydrogenases, lactate dehydrogenases, pyruvate:formate lyase and formate dehydrogenases (Poehlein et al., 2012).
The strict requirement of the WLP for glucose/ fructose oxidation and/or removal of hydrogen by T. kivui and A. woodii were demonstrated recently by a genetic approach.HDCR deletion mutants could no longer convert H 2 + CO 2 to formate, the first step in the WLP, as expected.But they also did not grow on glucose/fructose anymore (Jain et al., 2020;Moon et al., 2023).We speculated that growth on hexoses may be restored by allowing the cells to recapture hydrogen and thus have attempted to genetically delete an enzyme downstream of the HDCR, the methylene-THF reductase.This mutant could be generated, and we discovered that the mutant performs so far unknown fermentative metabolism during growth on fructose.

Cultivation of A. woodii
A. woodii strains DSM1030 (wild-type), ΔpyrE or ΔmetVF were cultivated under anoxic conditions at 30 C in bicarbonate-buffered complex medium described previously (Heise et al., 1989).As substrates for growth, 20 mM of fructose or 20 mM of fructose + 80 mM of glycine betaine were used.Growth was monitored by measuring the optical density at 600 nm (OD 600 ).

Generation of A. woodii ΔmetVF mutant
For the generation of a ΔmetVF mutant, the plasmid pMTL84151_AW_dmetVF was constructed in E. coli HB101 (Promega, Madison, WI) and transformed into A. woodii ΔpyrE strain, as described previously (Westphal et al., 2018).The plasmid was modified from pMTL84151 (Heap et al., 2009) as suicide plasmid which lacks a Gram-positive replicon.In pMTL84151_AW_dmetVF, 450 bp of upstream flanking regions (UFR) of metV (Awo_c09300) and 700 bp of downstream flanking regions (DFR) of metF (Awo_c09310) were cloned into the multiple cloning sites for deletion of the metVF genes via homologous recombination.Moreover, this plasmid contains a catP marker for chloramphenicol/thiamphenicol resistance from Clostridium perfringens (Werner et al., 1977) and a heterologous pyrE from Eubacterium limosum (Wiechmann et al., 2020) as a counter selectable marker.The first selection was achieved on an agar plate with complex medium supplemented with 20 mM of fructose and 30 ng/μL of thiamphenicol after transformation of pMTL84151_AW_dmetVF into A. woodii ΔpyrE strain by electroporation (625 V, 25 μF, 600 Ω, in 1 mm cuvettes).The second selection was carried out on an agar plate with minimal medium (Westphal et al., 2018) supplemented with 20 mM of fructose, 1 μg/mL of uracil and 1 mg/ mL of 5-FOA.The deleted region was analysed by PCR with oligonucleotides which bind in front of UFR and behind DFR of the respective gene: aus_metVF_for (5 0 -ATGATTGCTGATGAAAGAGGATTTTT-3 0 ) and aus_-metVF_rev (5 0 -AAGTCCCGCCAAGTTCATC-3 0 ).To check the purity of the mutant, oligonucleotides which bind in metV and metF genes were used for PCR: in_metVF_for (5 0 -GAGTGTTACCTGAATGGGACC-3 0 ) and in_metV-F_rev (5 0 -CGTCTGACTGTAAGGCAATACG-3 0 ).The deleted region of the mutant was further verified by Sanger sequencing (Sanger et al., 1977).

Transcriptome analyses
For transcriptome analyses, the wild-type and the ΔmetVF mutant were grown on 20 mM of fructose and harvested in the exponential growth phase at an OD 600 of 0.2.RNA preparation and RNA sequencing were carried out as described previously (Moon et al., 2021).Subsequently, gene expression levels of the ΔmetVF mutant were compared to those of the wild-type.Genes with a log 2 -fold change of transcript levels of +2/À2 and a p-adjust value <0.05 were considered differentially expressed.

Preparation of resting cells
Cells of the ΔmetVF mutant were grown either on 20 mM of fructose or 20 mM of fructose + 80 mM of glycine betaine in 0.5-3 L of bicarbonate-buffered complex medium to late exponential growth phase (on 20 mM of fructose, OD 600 of 0.25; on 20 mM of fructose +80 mM of glycine betaine, OD 600 of 1.5).After harvest by centrifugation (Avanti™J-25 and JA-10 Fixed-Angle Rotor; Beckman Coulter, Brea, CA) at 8000 rpm and 4 C for 10 min, cells were washed with 30 mL of buffer containing 50 mM of imidazole (pH 7.0), 20 mM of KCl, 20 mM of MgSO 4 , 4 mM of DTE and 4 μM of resazurin and pelleted by centrifugation at 8500 rpm and 4 C for 10 min (Avanti™J-25 and JA-25.50Fixed-Angle Rotor; Beckman Coulter, Brea, CA).Subsequently, the cells were resuspended in 5 mL of imidazole buffer and transferred to 16-mL Hungate tubes.All steps were performed under strictly anoxic conditions in an anoxic chamber (Coy Laboratory Products, Grass Lake, MI) filled with N 2 /H 2 (96%-98%/2%-4%; v/v).The gas phase of the cell suspensions was changed to 100% N 2 to remove residual H 2 from the anoxic chamber.The total protein concentration of the resting cells was determined according to Schmidt et al. (1963).

Cell suspension experiments
The cells were resuspended in 10 mL of imidazole buffer (50 mM of imidazole, 20 mM of KCl, 20 mM of NaCl, 20 mM of MgSO 4 , 60 mM of KHCO 3 , 4 mM of DTE, 4 μM of resazurin, pH 7.0) in 120 mL serum flasks to a final protein concentration of 2 mg/mL under an atmosphere of N 2 /CO 2 (80:20, v/v).20 mM of fructose was given as carbon and energy source, and if necessary, 80 mM of glycine betaine, 4 mM of caffeate or H 2 (100%, 1 bar) was added to the resting cells.The cells were preincubated at 30 C in water bath with shaking (150 rpm) and the experiments were started by the addition of the substrate(s).One millilitre of samples were taken at each time point for determination of metabolites.

Deletion of the metVF genes of A. woodii
The MTHFR of A. woodii consists of the three subunits RnfC2, MetV and MetF (Bertsch et al., 2015), encoded by rnfC2 (Awo_c09290), metV (Awo_c09300) and metF (Awo_c09310) that are clustered with other genes encoding the enzymes of the methyl branch of the WLP, such as fhs1 (Awo_c09260; formyl-THF synthetase), fchA (Awo_c09270; methenyl-THF cyclohydrolase) and folD (Awo_c09280; methylene-THF dehydrogenase; Poehlein et al., 2012).We deleted the metVF genes through allelic exchange mutagenesis leaving only 42 bp including the start codon of metV and 36 bp including the stop codon of metF (Figure 1A).PCR experiments with primers binding outside the deleted region showed that the metVF genes were successfully deleted (Figure 1B).Consequently, the metV and metF genes could not be amplified with primers binding inside of metV and metF (Figure 1C).Subsequently, the absence of metVF in the chromosome was confirmed by DNA sequencing (Sanger et al., 1977).

Growth phenotype of the ΔmetVF mutant
The mutant did not grow on H 2 + CO 2 , formate, methanol or glycine betaine, demonstrating that the WLP is essential for C1 metabolism.The mutant also did not grow on ethanol, 2,3-butanediol, acetoin or ethylene glycol, demonstrating the essentiality of the WLP as electron sink for the oxidation of alcohols.However, the situation was different with substrates that are metabolised via pyruvate.Cells grew on pyruvate and dihydroxyacetone but did not grow on lactate or alanine.Of special interest was the metabolism of hexoses such as fructose since these are common substrates for acetogens and the ΔmetVF mutant was obtained on fructose-containing agar plates.
The ΔmetVF mutant grew on fructose but growth was poor with a growth rate of 0.05 h À1 and a final OD 600 of only 0.3, which is only 26% and 10% of the wild-type values (Figure 2A).During 90 h of incubation, only 7.1 ± 0.5 mM of fructose was consumed, and only 8.9 ± 0.0 mM of acetate was produced with a fructose:acetate ratio of 1:1.25.This implies that acetate was not formed via the WLP, instead, 8.1 ± 1.9 mM of H 2 , 6.6 ± 0.2 mM of formate, 1.0 ± 0.2 mM of ethanol and 2.1 ± 0.3 mM of lactate were produced.
Alternative electron acceptors restore growth of the ΔmetVF mutant on fructose to wild-type levels Apparently, the ΔmetVF mutant did grow on fructose, but growth was rather poor.If this phenotype was due to the loss of an efficient pathway that enables reoxidation of reduced electron carriers, that is, the WLP, growth should be restored by addition of other electron acceptors.The substrates that feed carbon into the pathway after methylene-THF are methyl groupcontaining substrates.Indeed, we did show very recently, that methanol + CO 2 can be used as an electron sink; excess electrons are used to reduce CO 2 to CO which is then condensed with the methyl group and CoA to acetyl-CoA (Kremp et al., 2018;Litty et al., 2022).Here, we used glycine betaine as methyl group donor.The mutant did not grow on glycine betaine alone but in the presence of glycine betaine (plus CO 2 present in the medium), growth on fructose was restored to wild-type levels (Figure 2B).Ethanol, formate, lactate and H 2 were no longer formed.Caffeate is another alternative electron acceptor that is reduced by A. woodii to hydrocaffeate (Dilling et al., 2007;Hess, Gonzalez, et al., 2013).Addition of caffeate to the culture grown on fructose fully restored growth to the wildtype level and, again, formate, ethanol, lactate and H 2 were no longer formed (Figure 2C).Apparently, production of short chain fatty acids, ethanol and H 2 was turned off in the presence of an energetically more advantageous external electron acceptor.

Transcriptome of fructose metabolism in the ΔmetVF mutant
Comparative genome-wide expression profiling was performed with the wild-type and the ΔmetVF mutant during growth on fructose to address regulation of key enzymes and to shed light on the enzymes involved in formation of ethanol, lactate, formate and H 2 .There is only one hydrogenase present in A. woodii that can produce H 2 from NADH and reduced ferredoxin, the electronbifurcating hydrogenase HydABC (Schuchmann & Müller, 2012).However, expression of the encoding T A B L E 1 Transcript abundance of genes encoding key enzymes for fructose metabolism in the ΔmetVF mutant.

Gene
Annotation Function genes (Awo_c26970-Awo_c27010) in the mutant did not change in a physiologically significant range compared to the wild-type, but transcript counts were rather high in the mutant and the wild-type (Table 1).Expression of the HDCR (Awo_c08190-Awo_c08260) also did not change to a physiologically significant extent, but the two iso-genes encoding the formyl-tetrahydrofolate synthetase (fhs1, Awo_c09260; fhs2, Awo_c08040) and the following genes of the WLP were upregulated 2-to 3.5-fold.A gene encoding the potential formate transporter (Awo_c08050) was also upregulated 3.1-fold.Energy conservation via Rnf complex and ATP synthase seemed not to be affected at the level of gene expression.
A. woodii encodes an electron-bifurcating lactate dehydrogenase-Etf-complex (Weghoff et al., 2015) but apparently no other lactate dehydrogenase.The transcript levels of these genes together with the genes encoding a lactate racemase and a lactate permease (Awo_c08710-Awo_c08750) were highly upregulated (7.1-to 8.4-fold) in the mutant compared to the wildtype.Ethanol can be formed from acetyl-CoA by CoAdependent acetaldehyde dehydrogenases and alcohol dehydrogenases.A. woodii contains several genes encoding CoA-dependent acetaldehyde dehydrogenases and alcohol dehydrogenases, but the transcriptome data did not reveal a striking differential regulation of any of these genes.However, transcription of a gene cluster encoding a potential pyruvate dehydrogenase/ acetoin dehydrogenase (Awo_c01690-Awo_c01720) showed the highest level of upregulation in our studies.

Acetate and hydrogen formation from fructose in resting cells of the ΔmetVF mutant grown on fructose + glycine betaine
In cells of the ΔmetVF mutant growing on fructose, metabolism was shifted towards reduced end products due to the block in the WLP.Fructose metabolism of the ΔmetVF mutant was further investigated in resting cells to analyse catabolic electron flow only.For these experiments, cells were grown on 20 mM of fructose + 80 mM of glycine betaine until stationary growth phase, harvested, washed and resuspended in 10 mL of imidazole buffer as described in Experimental Procedures.When 20 mM of fructose was given to the resting cells of the ΔmetVF mutant, only 14.0 ± 0.2 mM of acetate was produced from 9.1 ± 0.7 mM with a fructose:acetate ratio of 1:1.5 (Figure 3A).A minor amount of formate (2.9 ± 0.2 mM) was produced but lactate and ethanol were not produced.Much to our surprise, H 2 was produced in huge amounts with 37.4 ± 2.2 mM, giving a fructose:H 2 ratio of 1:4.1.This is the maximum amount of H 2 that can be produced from fructose (see Equation 5).When H 2 was given in addition to fructose, fructose was no longer consumed, and acetate was not formed, indicating an inhibition of fructose oxidation by molecular hydrogen.However, formate was accumulated to high concentrations (24.3 ± 0.4 mM; Figure 3B) under these conditions, indicating that the HDCR converts bicarbonate/CO 2 (present in the buffer) with H 2 to formate with high rates, as observed previously (Kottenhahn et al., 2018;Schuchmann & Müller, 2013).When a methyl group (80 mM of glycine betaine) was added as electron sink (together with CO 2 /bicarbonate present in the medium), fructose consumption accelerated and was 6.6 times faster than in the absence of glycine betaine (Figure 3C).Resting cells performed homoacetogenesis and produced 109.4 ± 2.1 mM of acetate with a fructose:acetate ratio of 1:5.4.Hydrogen was no longer produced in the presence of glycine betaine.Addition of the alternative electron sink caffeate to cells metabolising fructose also increased fructose consumption rates and completely abolished H 2 production; the reducing equivalents were apparently used to reduce caffeate to hydrocaffeate (Figure 3D).2.2 ± 0.1 mM of acetate was formed from 1.1 ± 0.1 mM of fructose in the presence of 4.2 ± 0.1 mM of caffeate with a fructose:acetate ratio of 1:2.

MAF with fructose in resting cells of the ΔmetVF mutant grown on fructose
Whereas growing cells produced hydrogen as well as formate, lactate and ethanol in addition to acetate, resting cells described earlier did not.However, the cells used for these experiments were pre-grown on fructose + glycine betaine, conditions under which formate, lactate and ethanol were not formed, opening the possibility that the ability to produce formate, lactate and ethanol was turned off by the presence of glycine betaine as electron sink in the growth medium.To check this possibility, we grew cells on fructose only, prepared resting cells and analysed product formation from fructose.These cells converted 19.5 ± 1.0 mM of fructose to 27.6 ± 1.2 mM of acetate with a fructose:acetate ratio of 1:1.41 (Figure 4).However, as in growing cells, we found formate (4.2 ± 0.2 mM), ethanol (3.0 ± 0.2 mM) and lactate (5.8 ± 0.3 mM) along with hydrogen (48.7 ± 2.6 mM).Acetoin, glycerol and alanine were not produced.As observed in growing cells, intermediates (pyruvate and acetyl-CoA) as well as end products (CO 2 ) were reduced under these conditions, demonstrating the ability of A. woodii for a MAF of fructose.

DISCUSSION
The metabolism of acetogenic bacteria shows an extraordinary diversity.They can grow organotrophically as well as lithotrophically; central to both types of metabolism is the WLP.It allows growth on C1 compounds such as H 2 + CO 2 , formate, methyl groups and carbon monoxide (Diekert & Thauer, 1978;Genthner & Bryant, 1982;Kremp et al., 2018;Kremp & Müller, 2021;Lechtenfeld et al., 2018;Litty et al., 2022;Moon et al., 2021;Stupperich & Konle, 1993;van der Meijden et al., 1984;Weghoff & Müller, 2016;Wood et al., 1986).The WLP allows to conserve energy, either by the pathway itself, for example the formyl-THF synthetase in the oxidative reaction during methyl group oxidation (Kremp et al., 2018;Kremp & Müller, 2021;Lechtenfeld et al., 2018;Litty et al., 2022;Moon et al., 2023), or by a respiratory chain involved in redox balancing (Hess, Schuchmann, & Müller, 2013).However, the amount of ATP generated is small compared to the amount of ATP generated by glycolysis and the acetate kinase reaction.The use of the external electron acceptor CO 2 allows acetogenic bacteria that grow on hexoses to produce the maximum amount of ATP that can be obtained by hexose fermentation.Interestingly, many acetogens can use alternative electron acceptors such as nitrate, dimethylsulfoxide, fumarate or even phenylacrylates (Bache & Pfennig, 1981;Matthies et al., 1993;Rosenbaum et al., 2022;Seifritz et al., 1993).Here, we have identified another electron acceptor previously unknown for A. woodii and other mesophilic acetogens, protons.Apparently, A. woodii converted fructose and the reducing equivalents were released as H 2 .Hydrogenogenesis has been found before in several anaerobes (Schröder et al., 1994;O-Thong et al., 2008;Pradhan et al., 2015) and among them so far only one thermophilic acetogen, T. kivui (Moon et al., 2020).Hydrogenogenesis requires hydrogen formation from the reduced electron carriers involved in acetogenesis from fructose in A. woodii which are NADH and reduced ferredoxin.Hydrogen formation from NADH is F I G U R E 3 Conversion of fructose in resting cells of the ΔmetVF mutant.Cells of the ΔmetVF mutant were grown in bicarbonate-buffered complex media under a N 2 /CO 2 atmosphere (80:20, v/v) with 20 mM fructose + 80 mM glycine betaine and harvested in the early stationary growth phase.After washing, the cells were resuspended in 10 mL of cell suspension buffer (50 mM imidazole, 20 mM MgSO 4 , 20 mM KCl, 20 mM NaCl, 60 mM KHCO 3 , pH 7.0) in 120 mL serum flasks under a N 2 /CO 2 atmosphere at a total protein concentration of 2 mg/mL.(A) 20 mM fructose, (B) 20 mM fructose + H 2 , (C) 20 mM fructose + 80 mM glycine betaine was given to resting cells as carbon and energy source.(D) For conversion of 1 mM fructose + 4 mM caffeate, resting cells were prepared from the mutant grown on 20 mM fructose + 8 mM caffeate.Fructose (•), acetate (■), H 2 (▲), formate (▼) and caffeate (Â) were determined.Each data point indicates a mean with standard error of the mean; n = 2 independent experiments.thermodynamically restricted but made possible in A. woodii by the electron-bifurcating hydrogenase HydABC that catalyses the following reaction (Schuchmann & Müller, 2012, 2014) Hydrogenogenesis is also the preferred way of metabolism in co-culture of A. woodii with a methanogen (Heijthuijsen & Hansen, 1986;Winter & Wolfe, 1980).Under these conditions, the methanogen keeps the hydrogen concentration low, making hydrogen formation according to Equation 6 thermodynamically more favourable.Of course, high concentrations of the end product, H 2 , inhibit further hydrogen formation.This was observed in the presence of fructose + H 2 in A. woodii that inhibited further hydrogen formation and thus, fructose oxidation.
In the absence of a suitable electron acceptor such as CO 2 or when the thermodynamics does not allow for hydrogen formation, A. woodii as well as T. kivui did not grow (Jain et al., 2020;Moon et al., 2023).This has always been puzzling since at least A. woodii has the genetic potential for MAF.Indeed, here we did observe MAF from fructose for the first time in the metVF mutant.A. woodii has three copies of a pyruvate:formate lyase gene and thus, formate may have been produced by the PFL (Transcript levels of pfl genes were upregulated 0.1-1.5 fold).The formate produced may then have been converted to H 2 and CO 2 by the HDCR enzyme.A quantitative model of MAF that reflects the concentrations of the products formed, is shown in Figure 5.If we take 0.3 mol of pyruvate for lactate formation, as determined experimentally, that leaves 1.7 mol of acetyl-CoA.0.15 mol are reduced to ethanol (experimentally determined: 0.15 mol of ethanol per mol of fructose), which leaves 1.55 mol of acetyl-CoA that can be oxidised to acetate (experimentally determined: 1.41 mol of acetate per mol of fructose).Pyruvate oxidation by PFL yields 1.7 mol of formate.Since only 0.2 mol of formate per mol of fructose were found, 1.5 could have been converted to 1.5 mol of H 2 and 1.5 mol of CO 2 .However, not 1.5 mol of H 2 but 2.5 were experimentally determined.In addition, 1.7 mol of NADH are left over from glycolysis in this model.Hydrogen production from NADH requires reduced ferredoxin but the only way to get reduced ferredoxin is via the PFOR reaction.If we assume 1.7 mol of pyruvate to be oxidised by PFOR instead of PFL, this would yield 1.7 reduced ferredoxin that together with 1.7 mol of NADH would give 3.4 mol of H 2 .Therefore, we assume the PFL way to be unlikely but further genetic analyses will shed light on the role of the PFL.The situation is even more complicated in light of the three potential pfl genes (Poehlein et al., 2012).
Here, production of lactate as an end product has been observed for the first time in A. woodii.Since A. woodii has only one lactate dehydrogenase, the electron-bifurcating Ldh/Etf complex, reduction of pyruvate is accompanied with the reduction of ferredoxin (Weghoff et al., 2015).Reoxidation of reduced ferredoxin with concomitant reduction of NAD by the Rnf complex conserves additional energy by a chemiosmotic mechanism (Hess, Schuchmann, & Müller, 2013).Ethanol formation could proceed via acetaldehyde to ethanol by aldehyde dehydrogenase ALDH/PduP plus alcohol dehydrogenase Adh4/6 (Moon & Müller, 2021).
In sum, we have demonstrated MAF from fructose in A. woodii when the WLP has been genetically inactivated.Genes encoding enzymes for the production of lactate, ethanol, formate and hydrogen are present in A. woodii, but their role in this new metabolic trait of F I G U R E 4 Fermentation in fructose-adapted resting cells of the ΔmetVF mutant.Cells of the ΔmetVF mutant were grown in bicarbonate-buffered complex media under a N 2 /CO 2 atmosphere (80:20, v/v) with 20 mM fructose and harvested in the stationary growth phase (OD 600 of 0.3).After washing, the cells were resuspended in 10 mL of cell suspension buffer (50 mM imidazole, 20 mM MgSO 4 , 20 mM KCl, 20 mM NaCl, 60 mM KHCO 3 , pH 7.0) in 120 mL serum flasks under a N 2 /CO 2 atmosphere at a total protein concentration of 2 mg/mL.20 mM fructose was given to resting cells as carbon and energy source.Fructose (•), acetate (■), H 2 (▲), formate (▼), ethanol (♦) and lactate (Â) were determined at each time point.Each data point indicates a mean ± SEM; n = 2 independent experiments.

F
I G U R E 1 Deletion of the metVF genes in the chromosome of Acetobacterium woodii.(A) Only 42 bp of the metV gene and 36 bp of the metF gene remained in the ΔmetVF mutant.(B) Genotypic analyses of the ΔmetVF mutant were performed using colony PCR with primers binding outside the deleted region (aus_metVF_for and aus_metVF_rev) or inside (C) (in_metVF_for and in_metVF_rev).DFR, downstream flanking regions; UFR, upstream flanking regions.

F
I G U R E 2 Growth of the ΔmetVF mutant on fructose with alternative electron acceptors.Growth experiments were carried out in 5 mL complex medium (containing bicarbonate under a N 2 /CO 2 atmosphere; 80:20, v/v) in 16 mL Hungate tubes at 30 C with (A) 20 mM fructose, (B) 20 mM fructose + 80 mM glycine betaine and (C) 20 mM fructose + 8 mM caffeate.Depicted are the optical densities of the wild-type (•) and the ΔmetVF mutant (▲).(C) 2 mM Caffeate was added at 0, 29, 53 and 78 h of incubation.Black, with caffeate; white, without caffeate.The growth experiments were performed in biological triplicates and one representative growth curve is presented.