A genetic and metabolic approach to redirection of biochemical pathways of Clostridium butyricum for enhancing hydrogen production

Abstract

Clostridium butyricum, a well known H₂ producing bacterium, produces lactate, butyrate, acetate, ethanol, and CO₂ as its main by-products from glucose. The conversion of pyruvate to lactate, butyrate and ethanol involves oxidation of NADH. It was hypothesized that the NADH could be increased if the formation of these by-products could be eliminated, resulting in enhancing H₂ yield. Herein, this study aimed to establish a genetic and metabolic approach for enhancing H₂ yield via redirection of metabolic pathways of a C. butyricum strain. The ethanol formation pathway was blocked by disruption of aad (encoding aldehyde-alcohol dehydrogenase) using a ClosTron plasmid. Although elimination of ethanol formation alone did not increase hydrogen production, the resulting aad-deficient mutant showed approximately 20% enhanced performance in hydrogen production with the addition of sodium acetate. This work demonstrated the possibility of improving hydrogen yield by eliminating the unfavorable by-products ethanol and lactate. Biotechnol. Bioeng. 2013; 110: 338–342. © 2012 Wiley Periodicals, Inc.

Clostridium butyricum is a Gram-positive, spore-forming, anaerobic bacterium used for fermentative hydrogen production. Clostridium species possess an additional pathway that allows them to produce hydrogen from NADH under specific conditions (Hallenbeck, 2009; Cai et al., 2011). The evolution of H₂ from NADH by a potential bifurcating hydrogenase has been hypothesized (Schut and Adams, 2009). This enzyme is capable of oxidizing NADH and ferredoxin simultaneously to produce H₂ (Schut and Adams, 2009). NADH is usually generated during glycolysis of glucose. Various extracellular metabolites including acetate, butyrate, lactate, and ethanol can be produced in association with production of H₂ and biomass. The conversion of pyruvate to butyrate, lactate and ethanol involves oxidation of NADH. The NADH concentration should increase if the formation of the reduced by-products can be blocked. H₂ evolution through the NADH pathway is driven by the necessity for deoxidizing the residual NADH of metabolic reactions such as NADH + 2Fd^red + 3H⁺ → NAD⁻ + 2Fd^ox + 2H₂ (Fd^red: reduced ferredoxin; Fd^ox: oxidized ferredoxin). Thus, if metabolic reactions can be controlled to increase the amount of residual NADH, the H₂ yield may be enhanced further.

C. butyricum W5, a recently isolated hydrogen producer, is capable of using glucose, starch and molasses to produce hydrogen along with a diversity of acids (lactate, acetate, and butyrate) and other by-products (Wang and Jin, 2009; Wang et al., 2007). The butyrate formation pathway is recognized as the main competing pathway which consumes more NADH, reducing the H₂ yield (Hallenbeck, 2009; Cai et al., 2011). Hallenbeck (2009) suggested that elimination of the butyrate pathway may improve H₂ yield by Clostridium sp. In our recent study, we inactivated hbd (encoding β-hydroxybutyryl-CoA dehydrogenase) in C. butyricum W5 to eliminate the butyrate formation pathway, resulting in a significant increase in ethanol production and a slight decrease in H₂ yield compared with the wild type strain (Cai et al., 2011). Thus, we hypothesized that blocking the ethanol formation pathway may be more beneficial for enhancing hydrogen production yield.

The ethanol formation pathway consists of two reductive steps from acetyl-CoA to acetaldehyde and then to ethanol, which are catalyzed by acetaldehyde CoA dehydrogenase (ACDH) and alcohol dehydrogenase (ADH), respectively. The aad gene of C. butyricum encodes a bifunctional aldehyde-alcohol dehydrogenase, which has ACDH and ADH activities. This study aimed to examine whether H₂ yield can be enhanced via reducing ethanol formation. Herein, this article describes the successful interruption of the ethanol formation pathway by inactivation of aad using the ClosTron system. We for the first time investigated and used a genetic and metabolic approach of blocking the ethanol formation pathway in conjunction with addition of sodium acetate to improve H₂ production yield. We experimentally investigated how these genetic and metabolic alterations could affect metabolism during fermentative hydrogen production.

Disruption of aad in C. butyricum W5 was accomplished following the protocols outlined in the Methods section. An antisense insertion site was selected that would yield an unconditional disruption. This site is 683 bp downstream of the predicted start codon and resides in the ACDH domain on aad. The pMTL007C-E2 was used because it contains FRT-flanked ermB RAM and can be used for multiple gene disruption (Heap et al., 2010). Erythromycin resistance (Em^R) colonies were confirmed to be aad-deficient mutants by PCR using primers flanking the 683/684 site of aad.

This study presents another example of the successful application of the ClosTron system in gene disruption in C. butyricum. We note that selection of a target site plays a key role in specific insertion, which is important to determine whether or not the gene could be successfully disrupted. Initially, a sense insertion site, which is 1,137 bp downstream of the predicted start codon, was used. However, the re-targeted intron failed to insert itself into the selected target site. Thus, we selected another antisense insertion site that could target to the correct site. Furthermore, we found that selection of an appropriate plasmid is another important step. The plasmid pMTL007C-E2 rather than pMTL007 contains a lazZα ORF, which can be replaced with a re-targeted intron. Clones containing successfully re-targeted plasmid can be easily distinguished from those which contained the parental plasmid by while instead of blue colony color on plates supplemented with X-gal. In term of this, it is more efficient to screen re-targeted pMTL007C-E2 compared to re-targeted pMTL007.

To examine the effect of aad inactivation on end-product metabolism, batch fermentations of the wild type and the aad-deficient mutant (M6) strains were performed under the same conditions. It was exciting to note that the strain M6 did not produce ethanol, which indicated that interruption of the ACDH domain of aad successfully blocked the ethanol formation pathway. The function of AAD (referred to as AdhE in some reports) varies among different microorganisms (Arnau et al., 1998; Green and Bennett, 1996; Kessler et al., 1992). Based on our results, it would be assumed that AAD in C. butyricum W5 may be responsible for ethanol production rather than butanol production. Another possibility is that the ADH domain of aad may play an important role in butanol production and other ADHs may be responsible for ethanol production, since there are several ADHs in Clostridium species. Since ACDH catalyses the first step in the production of ethanol from acetyl-CoA, inactivating the ACDH domain should completely block ethanol formation, which is the case for the ACDH deficient mutant M6 in this study.

Our results showed that the mutant strain M6 performed an enormously enhanced lactate production (over 484%) as well as 32% more acetate production, while generating 78% less hydrogen and 9% less butyrate compared with the wild type strain at pH 6.5 (Fig. 1). These unexpected results suggested that blockage of the ethanol pathway could redirect the metabolic flux toward the lactate formation. A similar result was observed in cultures of Lactococcus lactis MGKAS15 (adhE deficient), which showed an enhanced lactate production (Arnau et al., 1998). It was interesting to note that the enhanced lactate production resulted in a significant reduction in H₂ and butyrate yield. Our findings from this and previous studies (Cai et al., 2010; Cai et al., 2011) are suggestive that elimination of the butyrate formation may not be a promising approach to the enhancement of H₂ yield.

Figure 1. Product yields during fermentation using the wild type strain (W5) and the aad-deficient strain (M6). LA, lactic acid; AA, acetic acid; Eth, ethanol; BA, butyric acid; H2, hydrogen. (Value ± SD for n = 2 samples).

It seems that the reduction of lactate production is a key step to enhance H₂ yield by the mutant strain M6. Our recent metabolic flux analysis shown that operational pH is a crucial parameter which affected lactate production (Cai et al., 2010). Lactate yield decreased from 0.81 mol/mol glucose to 0.52 mol/mol glucose as pH decreased from 6.5 to 5.5. Lactate production increased slightly (0.54 mol/mol glucose at pH 5.0). This result could be linked to the decrease in the activity of NAD-independent lactate dehydrogenase (iLDH) responsible for lactate utilization at pH < 5.5 (Diez-Gonzalez et al., 1995). By contrast, the yields of acetate, butyrate, and hydrogen increased as pH varied from 6.5 to 5.0. However, a higher H₂ and lactate yield was observed in the mutant strain M6.

Here, we can assume that the H₂ yield should be improved if the lactate produced can be reutilized. The role of acetate in the utilization of lactate has been reported in several Clostridium species. Thus, we added sodium acetate (NaAc) into the fermentation medium in order to examine its effect on the utilization of lactate with respect to operational pH. Analytical data showed that NaAc inhibited the cell growth at pH 5.0, while NaAc improved cell growth at pH 5.5 or above (data not shown). The profiles of metabolite production from glucose with NaAc addition at pH 5.5 are shown in Figure 2. Lactate increased within 11 h and reached a maximum value of 1.41 g/L, and then followed a lactate reduction phase. This indicates that lactate produced during the exponential phase was consumed completely. Thus, we may conclude that addition of NaAc can substantially reduce lactate formation by the mutant strain M6, while leading to a pronounced increase in butyrate production, reaching a maximum value of 2.53 g/L at 24 h. Importantly, the mutant strain M6 grown in the medium with additional NaAc performed a significant enhancement in H₂ yield from 0.94 mol/mol glucose to 1.65 mol/mol glucose with the addition of NaAc at pH 5.5.

Figure 2. Time course profile of glucose consumption and product formation during fermentation using M6 under operational pH of 5.5 with the sodium acetate addition. LA, lactic acid; AA, acetic acid; Eth, ethanol; BA, butyric acid; Glu, glucose; H2, hydrogen. (Value ± SD for n = 2 samples).

Our results revealed that with the addition of NaAc (or acetate) C. butyricum aad-deficient strain M6 was able to utilize endogenetic end-product lactate to produce butyrate and hydrogen when glucose was nearly consumed. However, this phenomenon was not obvious when NaAc was not added. This suggested that it was the added NaAc instead of the endogenetic by-product acetate that may play a role in redirection of metabolic fluxes, leading to significant lactate consumption by the C. butyricum aad-deficient strain M6. The only difference between the added NaAc and endogenetic acetate is that the former was present from the beginning of the fermentation and the latter was only produced after 5 h fermentation. The requirement for acetate in lactate utilization by C. beijerinckii (Bhat and Barker, 1947) and C. tyrobutryicum (Woolford, 1984) has been reported in previous studies. The role of acetate in the utilization of lactate by C. acetobutylicum was defined as an alternative electron acceptor and acetate was completely consumed at the end of fermentation (Diez-Gonzalez et al., 1995). However, acetate consumption appeared to be minor when lactate was utilized in this study. Results from this study demonstrated that the added acetate can affect the cell growth (maximum OD_600nm increased from 3.15 to 5.02), and improve butyrate and gas production (Fig. 1). Colin et al. (2001) and Heyndrickx et al. (1991) suggested that acetate acted as an indirect proton acceptor during the early stage of glycerol fermentation by C. butyricum, but did not serve as an energy source. This can explain why the added NaAc played a role in stimulation of butyrate formation. The re-utilization of lactate may be explained by the essential reactions to consume excess oxidized nucleotides (NAD⁺) generated through stimulation of butyrate formation by acetate supplementation, leading to increased hydrogen production.

It appeared that H₂ yield was improved merely due to pH decrease and NaAc addition. To clarify if the aad disruption could contribute to hydrogen production improvement, we conducted comparison trials with NaAc addition using the wild type C. butyricum W5 strain and the hbd deficient mutant. The latter was genetically modified to eliminate butyrate production (Cai et al., 2011). It was found that the added NaAc did not significantly influence hydrogen production in both strains although increase butyrate production occurred in the wild type W5 strain (data not shown). Therefore, hydrogen production was enhanced via redirection of metabolic pathways as a result of aad disruption and NaAc addition.

From this study, we hypothesize that the presence of NaAc at the beginning of the fermentation may stimulate the conversion of acetate to butyrate via coenzyme A transferase. After glucose was nearly consumed, the re-utilization of lactate driven by the necessity to consume excess NAD⁺ generated through stimulation of butyrate production resulted in enhanced hydrogen production using the aad-deficient mutant.

Materials and Methods

The strains and plasmids used in this study are listed in Table I. Growth conditions and plasmid transfer procedures were as described previously (Cai et al., 2011). Mutants deficient in aad were generated according to the ClosTron system and procedures using pMTL007C-E2 (Heap et al., 2010). PCR using primer sets AADa-f and AADa-r (Table II) was performed to confirm intron insertion in the aad.

Table I. Strains and plasmids.

Organism/plasmids	Feature	Source/reference
Strains
Clostridium butyricum W5	Wild-type	Laboratory strain/ Wang et al. (2007)
Escherichia coli DH5α	E. coli cloning strain	Laboratory strain
E. coli CA434	E. coli HB101 carrying R702	The University of Nottingham
C. butyricum M6	aad::intron	This study
Plasmids
pMTL007C-E2	E. coli/Clostridium shuttle vector	The University of Nottingham/ Heap et al. (2010)
pMTL007C-E2::Cbu-aad-683a	pMTL007C-E2 with intron targeted to aad	This study

Table II. Primers used for DNA amplification in this study.

Oligonucleotide	Sequence (5′-3′)	Target
AADa-f	CTCCAGAAGGAATAATCGGATGG	aad
AADa-r	CTATTTCTTCGCCTTTAAGGATG
EBS universal	CGAAATTAGAAACTTGCGTTCAGTAAAC	Intron retargeting
IBS	AAAAAAGCTTATAATTATCCTTAACTGACTTTACAGTGCGCCCAGATAGGGTG
EBS1d	CAGATTGTACAAATGTGGTGATAACAGATAAGTCTTTACAGCTAACTTACCTTTCTTTGT
EBS2	TGAACGCAAGTTTCTAATTTCGATTTCAGTTCGATAGAGGAAAGTGTCT

Batch fermentations were performed in a laboratory-scale batch bioreactor BioFlo110 (New Brunswick Scientific, Enfield, CT) as described previously (Cai et al., 2011). The pH was controlled at four levels of 6.5, 6.0, 5.5, and 5.0, or 6.5 if not specified. For sodium acetate (NaAc) addition experiments, the fermentation medium was supplemented with 2 g/L NaAc. Sample preparation and analyses were performed as described by Cai et al. (2010).