1 Introduction

Endogenous rumen microbiota can convert fiber to volatile fatty acid (VFA), and also result in the production of hydrogen and carbon dioxide (Kamra et al., 2012). When hydrogen builds up, it inhibits the oxidation of reduced nicotinamide adenine dinucleotide (NADH), leading to the accumulation of lactic acid and reduction of the fermentation (Wolin et al., 1997). To maintain a balanced environment for fermentation in the rumen, methanogens can perform methanogenesis, which converts hydrogen and carbon dioxide to methane. However, this process results in 2%–15% loss of feed energy (Johnson and Johnson, 1995) and exacerbation of greenhouse effects.

Recent studies have shown that hydrogen is less used for methanogenesis in the gastrointestinal tract of termites (Breznak, 1994) and Australian marsupials (Ouwerkerk et al., 2009) due to the predominant acetogenesis in the gut. Reductive acetogens are a group of bacteria that can produce acetate from hydrogen and carbon dioxides (4H2+2CO2→CH3COOH+2H2O) using the acetyl-CoA pathway (Drake et al., 2006). This suggests that acetogens may serve as hydrogen sinks in experiments, which seek to lower methane emissions (van Nevel and Demeyer, 1995; Morvan et al., 1996; Joblin, 1999) in cattle. Joblin et al. (1999) has reported that methane production decreased by 97% and 64%, respectively, when a Methanobrevibacter smithii sp. isolated from a grazing sheep was grown on H2/CO2 in vitro in the presence of a rumen acetogen isolate. Although an acetogen originating from sewage-sludge, Peptostreptococcus productus, competed successfully against methanogens in a simulated gastro-intestinal fermenter (Nollet et al., 1997a), such competition failed when in vitro ruminal digesta was used (Nollet et al., 1997b; 1998). Chaucheyras-Durand et al. (1995b) have shown that the presence of yeast cells could stimulate utilization of hydrogen by acetogens and enhance acetogenesis in an experiment utilizing a co-culture of acetogen and methanogen. Therefore, we hypothesized that the methanogenesis in the rumen is reduced by the enhanced acetogenesis with a supplement of Saccharomyces cerevisiae fermentation product (XP) and exogenous acetogens.

TWA4, a novel reductive homoacetogen isolated from forestomach contents of female tammar wallabies, could outcompete M. smithii in high hydrogen or heterotrophic growth on glycerol (with low hydrogen generated during fermentation) (Gagen et al., 2014). Therefore, in this study we assessed the effects of TWA4 and XP on methane production, rumen fermentation, methanogen, and acetogen population, as well as acetogenic diversity using the in vitro rumen fermentation system.

2 Materials and methods

2.1 Acetogen and S. cerevisiae fermentation product

TWA4 strain was kindly provided by Dr. Chris MCSWEENEY (Commonwealth Scientific and Industrial Research Organization (CSIRO), Australia) and was revived from anaerobic glycerol medium using the modified AC11.1 medium (Gagen et al., 2010) with H2 and CO2 in a proportion of 1:3 at 120 kPa. After growing three generations, the cell density was counted using a hemocytometer (Strober, 2001). Original XP was supplied by Diamond V (Cedar Rapids, IA, USA).

2.2 Experimental design

The effects of TWA4 and XP supplementation on ruminal fermentation, methane production, methanogen population, abundance, and diversity of rumen acetogens were determined using a 2×2 factorial design using in vitro rumen fermentation system. The four treatments were the Control (without TWA4 or XP), TWA4 (2×107 cells/ml TWA4 without XP), XP (2 g/L XP without TWA4), and TWA4XP (2×107 cells/ml TWA4 with 2 g/L XP).

2.3 In vitro rumen fermentation

The rumen fluid was collected from three ruminally fistulated lactating Chinese Holstein cattle (raised in Hangzhou Zhengxing Animal Husbandry Co. Ltd., China) fed twice daily with a mixed diet (roughage:concentrate=55:45) before morning feeding and strained through four layers of gauze into a pre-warmed and insulated bottle. The care and use of fistulated cattle was approved by the Animal Care Committee of Zhejiang University (Hangzhou, China). Rumen fluid was processed under continuous flushing with CO2. The 120 ml serum bottle containing 0.5 g dry substrates (50:50 (w/w) mixture of Chinese wild rye meal and corn silage), 40 ml buffered medium, and 10 ml rumen fluid was anaerobically incubated at 39 °C using the semi-automated reading pressure technique (Mauricio et al., 1999). To get the final concentration of 2 g/L XP for the XP and TWA4XP treatments, 0.1 g XP was added into the serum bottle following the supplement of 0.5 g dry substrates. The 2.5 ml growing TWA4 medium was injected into the bottle for TWA4 and TWA4XP treatments, while 2.5 ml medium without TWA4 inoculation was injected into the control and XP treatments. A pressure transducer, connected with a computer, was used to measure the accumulated head-space gas pressure through the in vitro incubation. Gas pressure was recorded after 6, 12, 24, and 48 h of incubation and was subsequently converted to the volumes of gas production (GP). At the same time intervals, 2 ml of head-space gas was collected with an airtight needle (SGE Analytical Science, Australia) to measure the methane production using gas chromatography (GC-2010, Shimadzu, Kyoto, Japan) equipped with a Flame Ionization detector (Hu et al., 2005). For each treatment, triplicate bottles were included and three blanks were included simultaneously to correct the GP values for gas release from endogenous substrates.

At the end of incubation (48 h), 3 ml mixed rumen samples were collected from each bottle and three bottles were sampled under oxygen-free CO2, and immediately stored at −80 °C to await further determination of the quantity of acetogens and methanogens, the diversity of acetogens, and the measurement of VFAs. VFAs were determined using the procedure described by Hu et al. (2005). Dry matter degradation (DMD) was measured using the modified nylon bag method (Goering and van Soest, 1970).

2.4 Total DNA extraction and real-time quantitative polymerase chain reaction (PCR)

Total DNA was extracted from rumen fluid collected from a 48-h incubation period using the bead-beating method as previously reported (Gagen et al., 2010). The primers specific to formyltetrahydrofolate synthetase gene (ƒhs) and methyl coenzyme-M reductase A (mcrA) genes (Table 1) were used to enumerate microorganisms with formyltetrahydrofolate synthetase (FTHFS) (Xu et al., 2009) and methanogens (Denman et al., 2007), respectively. The 16S rRNA gene of total bacteria was amplified with the primers as reported by Denman and McSweeney (2006) as shown in Table 1, and the copy number of total bacterial 16S rRNA gene was used as the reference to calculate the relative quantification of target. Real-time PCR was performed with SYBR green in ABI 7500 using the program: one cycle of initial denaturation at 95 °C for 10 s, 40 cycles of denaturation at 95 °C for 15 s, and annealing at 60 °C for 1 min. The relative abundance of each marker gene was estimated as: relative quantification of target=2−(CT target−CT total bacteria).

figure Tab1

2.5 Acetyl-CoA synthase gene sequencing

To investigate the effect of TWA4 and XP on acetogen diversity, acetyl-CoA synthase (ACS) genes were amplified and sequenced from the triplicate of DNA extracted from a 48-h incubation period as previously described (Gagen et al., 2010). The amplicon triplicates from each treatment were then pooled on an equal concentration basis (Checked by Qubit 2.0, Invitrogen, USA) for clone library construction using the pGEM-T Easy® vector (Promega Co., Madison, Wisconsin, USA) and Escherichia coli competent cell DH5α (TaKaRa, Dalian, China) according to manufacturer’s instructions. In total, 96, 96, 94, and 92 clones were randomly sequenced from the control, TWA4, XP, and TWA4XP libraries, respectively, and were then sequenced with T7 primer (Beijing Genomic Institute, China). The Primer Premier 5.0 (PREMIER Biosoft International, CA, USA) was used to translate DNA sequences to amino acid sequences before alignment. Sequence similarity to ACS was determined by BLASTP analysis (Gish and States, 1993) using all existing bacterial ACS amino acid sequences in NCBI database. The ACS amino acid sequences were grouped into phylotypes at a distance of 0.01 using Jones-Taylor-Thornton Matrix of MEGA6 (Tamura et al., 2013), because Blautia schinkii and Ruminococcus obeum shared 96.3% ACS amino acid identity, while 0.03 and 0.02 distance levels could not be clustered in our ACS amino acid sequences. The richness of ACS amino acid sequences was evaluated by the number of phylotypes and Chao1 index. The evenness was analyzed by Simpon’s diversity and Pielou’s evenness indices, calculated from Shannon’s diversity index (Felsenstein, 1993). Relative abundance of the phylotypes was determined as the sequence numbers in the phylotypes/total number of sequences. The Chao1, Simpon’s and Shannon’s diversity indices, Good’s coverage (1−nscp/ntotal, nscp is the number of single clone phylotypes and ntotal is the total number of sequences) were estimated using the summary single command in MOTHUR (Schloss et al., 2009). Unifrac.weighted in MOTHUR was used to compare the structure of acetogen community based on ACS amino acid sequences, and then to run the principal coordinates analysis (PCoA) with the pcoa command in MOTHUR. Bootstrapped neighbour joining tree of deduced ACS amino acid sequences was constructed with MEGA 6.06 with 100 resamplings. Similar amino acid sequences identified by BLASTP analysis and ACS of TWA4 (AEL12814 and AEL12815) and Methanococcoides methylutens (KGK98586) were selected as references for tree construction. Putative ACS amino acid sequences determined in the present study have been submitted to the GenBank database under accession numbers KR152340 to KR152636.

2.6 Statistical analyses

Data for DMD, methane production, VFA, and abundance of acetogens and methanogens were analyzed by two-way analysis of variance in SAS 9.1 (SAS Institute Inc., Cary, NC, USA) with individual bottles as the experimental unit, TWA4 and XP supplementation as main effects, where the TWA4×XP interaction was significant, and a secondary test was conducted to separate the efficacy of TWA4 within XP (Robinson et al., 2006). Multiple comparisons means among treatments were completed by Duncan’s multiple range tests. Significance was declared if P<0.05.

3 Results

3.1 Rumen fermentation characteristics

3.1.1 Methane production

The methane production (ml/g substrate) was not affected (P>0.05) by TWA4 or XP alone except for TWA4 at 6 h incubation (Table 2). When supplemented with TWA4XP, methane production was increased (P<0.05) at all time points by 20% to 107% compared with the control.

figure Tab2

3.1.2 Fermentation parameters

DMD was not affected (P>0.05) by treatments (Table 2). XP increased (P<0.05) total VFA and acetate and butyrate concentrations and TWA4XP increased (P<0.05) acetate, propionate, butyrate, and total VFA concentrations compared with the control. TWA4 showed no effect (P>0.05) on VFA production.

3.2 FTHFS and methanogen abundance

After a 48-h incubation period, the relative abundances of FTHFS and methanogens were increased (P<0.05) with the addition of TWA4, XP, and TWA4XP, though relative abundance of FTHFS was much lower than methanogen abundance (Table 3). The increase in the abundance of FTHFS was greater than that of methanogens (P<0.05) for XP and TWA4XP compared with TWA4 (Table 3). The increases of FTHFS and methanogens were 382.1% and 209.4% for TWA4, 656.0% and 478.1% for XP, and 679.1% and 421.9% for TWA4XP, respectively, compared with the control.

figure Tab3

3.3 Acetogen diversity

In total, 73, 70, 74, and 80 ACS amino acid sequences were detected in the control, TWA4, XP, and TWA4XP groups, respectively. Further amino acid similarity (99%) analysis identified 15 phylotypes for the control, 10 phylotypes for TWA4, 12 phylotypes for XP, and 7 phylotypes for TWA4XP with 3 common phylotypes among the four treatments (Table 4). Coverage analysis and a rarefaction curve indicated that the sequences obtained covered the majority of acetogen communities in all treatments with 89.0% for the control, 94.3% for TWA4, 93.2% for XP, and 95.0% for TWA4XP groups, respectively (Table 4, Fig. 1). From the results of numbers of phylotypes and Chao1, the richness of deduced ACS amino acid sequences was highest in the control group, followed by XP, TWA4, and TWA4XP. The control and XP had the highest Pielou’s evenness followed by TWA4, while TWA4XP had the highest Simpson indices value.

figure Tab4

Eight out of 22 of the phylotypes showed high (98%–100%) similarity to putative ACS amino acid of uncultured bacterium at GenBank uploaded by Gagen et al. (2010). All these phylotype sequences showed 62%–90% similarity to the nearest valid taxa including Acetitomaculum ruminis, R. sp. CAG:9, R. obeum, Eubacterium limosum, B. schinkii, B. hydrogenotrophica, and B. wexlerae (Table 5). Only 3 phylotypes were common phylotypes among the four treatments (Fig. 2), which occupied about 79.1% of the total ACS sequences. Phylotypes 1 and 2 had 86% similarity with A. ruminis, 96% and 98% similarities with uncultured bacterium. Phylotype 4 had 76% similarity with B. schinkii (Table 5).

figure Tab5
Fig. 1
figure 1

Rarefaction curve of observed phylotypes of acetyl-CoA synthase generated at 99% identity cutoff value

Full size image

Based on amino acid identified phylotypes, uncultured bacteria sequences were predominant in all acetogen communities. The dominant acetogens were unchanged in the treatments (phylotypes 1 and 2; Table 5 and Fig. 3) while the acetogen communities were significantly different among four treatments (weighted significance <0.001; Fig. 3b), whose community changes were reflected in the appearance new phylotypes. For XP treatment, about all the four unique phylotypes were identified as low-E. limosum-like acetogens (Table 5). There were no TWA4 ACS amino acid sequences actually recovered in the TWA4 and TWA4XP libraries (Fig. 2).

Fig. 2
figure 2

Phylogenetic analysis of deduced acetyl-CoA synthase amino acid sequences

P is the phylotypes generated at 99% identity cutoff value. GenBank accession numbers of reference sequences are shown after the species names. Bootstap values of ≥50% are shown at the nodes. The low-A. ruminis-like sequences and low-E. limosum-like sequences are marked with solid circles (•) and squares (▪), respectively

Full size image

4 Discussion

Ruminal methane production through enteric fermentation is of global concern due to its contribution to greenhouse gas emission as well as accounting for dietary energy loss in animals. Since the 1950s much effort has been expended trying to reduce methane emission from ruminants. Although the use of agents including chemicals and antibiotics has been successful, their use has been limited due to the probability of high quantities of residuals remaining in the animal products and feces (Kobayashi, 2010; Patra, 2012). Methanogenesis is the primary method of facilitating the consumption of hydrogen in the rumen, followed by fumarate reduction (Asanuma and Hino, 2000; Mitsumori and Sun, 2008), sulfate reduction (Morvan et al., 1996; Sahakian et al., 2010), nitrate and nitrite reduction (Sakthivel et al., 2012), and reductive acetogenesis (Henderson et al., 2010). Evidence from in vitro incubations of rumen contents (Nollet et al., 1997a; le Van et al., 1998) and methanogen-free lambs (Fonty et al., 2007) indicates that acetogens can function as hydrogenotrophs to suppress methanogenesis. Thus, enhancing reductive acetogenesis in the rumen may be an effective strategy to mitigate methane.

Lopez et al. (1999) investigated the ability of six reductive acetogens to prevent the accumulation of methane in vitro and found that only two of them, E. limosum strains ATCC 8486 and Ser 5, decreased methane production by about 5% after a 24-h incubation period, while total VFA was not affected (P> 0.05) by the addition of the six acetogens. As with Lopez’s study, the addition of TWA4 alone did not change the methane and VFA production in our experiment. The TWA4, XP, and TWA4XP failed to improve the DMD, revealing that no additional hydrogen and nutrients were produced from the substrates in these treatments compared with the control.

Yeast, as a feed additive, has been confirmed as being able to provide nutrients and vitamins to microorganisms (Chaucheyras-Durand et al., 1995a). Chaucheyras-Durand and Fonty (2001) reported that the supplementation of S. cerevisiae CNCM I-1077 tended to improve the in sacco degradation of wheat straw, and increased VFA production. From that study, Lascano and Heinrichs (2009) reported S. cerevisiae yeast culture increased rumen VFA concentration without influencing DMD. They concluded that the increased total VFA concentration with unchanged molar proportions of individual VFA was the result of the increased fermentation rate created by yeast culture supplement. However, reasons for inconsistency between the increased fermentation rate and unchanged DMD were not explained. As with the above study, our results showed that DMD remained unchanged, suggesting that the increased VFA by XP addition or XP and TWA4 co-addition was not a product of fermentable organic matter, such as cellulose or starch.

Chaucheyras-Durand et al. (1995b) showed that the presence of yeast cells stimulated the utilization of hydrogen by acetogens and enhanced acetogenesis, in a co-culture of acetogens and methanogens. In our XP and TWA4XP treatments, the increased VFA was mainly due to the acetate (Table 2). With the ƒhs primers of Xu et al. (2009), the rate of acetogen increase was higher than that of methanogens in XP and TWA4XP, suggesting that XP treatments may alter the acetogen communities and increase acetogenesis. However, according to Gagen et al. (2010), the ƒhs primers used to enumerate acetogens in our experiment could recover partial FTHFS sequence from a wider range of rumen acetogens, although multiple spurious amplicons could be generated from some acetogens and rumen microbial DNA. Until now, no appropriate real-time PCR primers were reported for acetogens, it could only deduce that XP treatments may increase the FTHFS biochemical pathway of rumen microorganisms, especially acetogens, which would cause an increase in VFA production.

The diversity of the acetogen community in our experiment showed that Lachnospiraceae was the dominant acetogen in the rumen fermentation system, but without close sequences from cultured isolates, which was similar to the results of Gagen et al. (2010). However, the species richness of our experiment was lower than that of Gagen et al. (2010). This difference may be due to the production accumulation in vitro system. It is possible that TWA4 could not survival after a 48-h fermentation period; however, the acetogen community was significantly changed by adding TWA4 alone or with XP, and therefore we posit that future studies are needed to track the fate of TWA4 with the change of fermentation time.

Acetogen phylotypes were increased by the addition of XP (Table 5 and Fig. 3a) with four unique phylotypes where amino acid was identified as low-E. limosum-like acetogen (Table 5). E. limosum and A. ruminis were all isolates from rumen by Sharak Genthner et al. (1981) and Greening and Leedle (1989), respectively. These bacteria can utilize H2/CO2, one-carbon compounds, to produce acetate. However, in contrast to A. ruminis, E. limosum can utilize amino acid to produce acetate and butyrate, and its growth can be stimulated by amino acids (Sharak Genthner et al., 1981; Pacaud et al., 1985). Amino acids were a substrate for methanogen growth and methanogensis (Mathrani and Boone, 1985; Chaucheyras-Durand et al., 1995b) as well. Therefore, enhancing acetogenesis by supplement with acetogen strain and/or yeast cells may be a way of mitigating methane, targeting acetogens such as E. limosum which utilize substrates such as amino acids to facilitate growth and acetogenesis. In order to further support our speculation, future experiments should next try isolating the low-E. limosum-like acetogens to test their nutritional characteristics and competition with methanogens.

Fig. 3
figure 3

Unique and common acetyl-CoA synthase phylotypes identified in the four treatments (a), and the principal coordinates analysis (PCoA) for the four treatments (b)

Full size image

5 Conclusions

The efficacy of adding TWA4 alone in the in vitro rumen fermentation system was limited, while with the substrate provided by XP, TWA4 could enhance acetogenesis by changing the acetogen community to low-A. ruminis-like acetogens. However, methanogenesis in rumen was enhanced by the co-addition. The XP addition could enhance acetogenesis with unchanged methanogenesis by changing the acetogen community to low-E. limosum-like acetogens, suggesting that enhancing acetogenesis by supplementation with acetogen strain and/or yeast cells may be a way to mitigate methane, with proper targeted acetogens such as the uncultured low-E. limosum-like acetogens.

Compliance with ethics guidelines

Chun-lei YANG, Le-luo GUAN, Jian-xin LIU, and Jia-kun WANG declare that they have no conflict of interest.

All management and experimental procedures were approved by the Animal Care Committee of Zhejiang University, China and were carried out in accordance with the University’s guidelines for animal research.