Volume 19, Issue 1 p. 29-41
Minireview
Free Access

Formation of propionate and butyrate by the human colonic microbiota

Petra Louis

Corresponding Author

Petra Louis

Rowett Institute of Nutrition and Health, University of Aberdeen, Aberdeen, Foresterhill UK.

For correspondence. E-mail [email protected]; Tel. +44 1224 438735; Fax +44 1224 438699.Search for more papers by this author
Harry J. Flint

Harry J. Flint

Rowett Institute of Nutrition and Health, University of Aberdeen, Aberdeen, Foresterhill UK.

Search for more papers by this author
First published: 05 November 2016
Citations: 1,031

Summary

The human gut microbiota ferments dietary non-digestible carbohydrates into short-chain fatty acids (SCFA). These microbial products are utilized by the host and propionate and butyrate in particular exert a range of health-promoting functions. Here an overview of the metabolic pathways utilized by gut microbes to produce these two SCFA from dietary carbohydrates and from amino acids resulting from protein breakdown is provided. This overview emphasizes the important role played by cross-feeding of intermediary metabolites (in particular lactate, succinate and 1,2-propanediol) between different gut bacteria. The ecophysiology, including growth requirements and responses to environmental factors, of major propionate and butyrate producing bacteria are discussed in relation to dietary modulation of these metabolites. A detailed understanding of SCFA metabolism by the gut microbiota is necessary to underpin effective strategies to optimize SCFA supply to the host.

Introduction

Short chain fatty acids (SCFA) are the major metabolic products of anaerobic fermentation by microbial communities that colonize the mammalian gut, typically reaching total concentrations of 50–200 mM in the human large intestine. They are taken up efficiently by the gut mucosa and have important impacts upon host physiology as sources of energy, as regulators of gene expression and as signalling molecules that are recognized by specific receptors (Morrison and Preston, 2016; Koh et al., 2016). New mechanisms by which SCFA regulate immune cell development and suppress inflammation have been uncovered recently (Louis et al., 2014; Richards et al., 2016). It is apparent however that the three major SCFA, acetate, propionate and butyrate, differ considerably in their potential effects upon host physiology. First, they differ in their fate and tissue distribution, with butyrate being used preferentially as an energy source by the gut mucosa, propionate contributing to gluconeogenesis in the liver and acetate achieving the highest systemic concentrations in blood (Morrison and Preston, 2016). Second, there are differences in interactions with host proteins (e.g., inhibition of histone deacetylases by butyrate and propionate) and receptors (Bolognini et al., 2016). This makes it particularly relevant to consider the microbial origin of these major fermentation products and the potential for changes in diet and gut physiology to affect their relative production rates and concentrations.

This brief review will focus on butyrate and propionate as these two acids are most often considered to benefit health, including protection against colorectal cancer in the case of butyrate and promotion of satiety and reduction in cholesterol in the case of propionate (Morrison and Preston, 2016). Acetate is a net fermentation product for most gut anaerobes that is also produced by reductive acetogenesis, and almost invariably achieves the highest concentrations among the SCFA in the gut lumen. In contrast, propionate and butyrate are produced by distinct subsets of gut bacteria. We consider here what is currently known about the phylogenetic distribution of pathways leading to the formation of these two SCFA within the human colonic microbiota and the potential for diverse dietary and environmental factors to differentially modulate their production. Some fermentation products, including lactate, succinate and 1,2-propanediol, do not usually accumulate to high levels in the human colon of healthy adults, as they can also serve as substrates for other bacteria, including propionate and butyrate producers. As the microbial metabolism of these compounds is intricately linked to the degradation of the main dietary substrates, it will be discussed together with propionate and butyrate formation from carbohydrates and proteins respectively.

Pathways and bacterial groups contributing to butyrate formation from carbohydrates

Butyrate is produced from carbohydrates via glycolysis from the combination of two molecules of acetyl-CoA to form acetoacetyl-CoA, followed by stepwise reduction to butyryl-CoA. Two different pathways are known for the final step in butyrate formation from butyryl-CoA, which proceeds either via butyryl-CoA:acetate CoA-transferase or via phosphotransbutyrylase and butyrate kinase (Louis and Flint, 2009) (Fig. 1). Butyrate-producing species are found interspersed with butyrate non-producing species in the two predominant families of human colonic Firmicutes, Ruminococcaceae and Lachnospiraceae, as well as in other Firmicutes families, including Erysipelotrichaceae and Clostridiaceae (Barcenilla et al., 2000; Louis et al., 2004). We will briefly consider the characteristics of butyrate-producers that belong to the two most abundant families of Firmicutes. We should note that, as summarized in Table 1, many dominant human colonic Firmicutes (e.g., Blautia spp., Eubacterium eligens, Ruminococcus spp.) lack the ability to form butyrate from carbohydrates.

Details are in the caption following the image

Microbial pathways for butyrate formation from carbohydrates, organic acids, glutamate and lysine in gut communities. Carbohydrate fermentation to pyruvate via glycolysis is shown in green, butyrate formation from acetyl-CoA in black, amino acid fermentation pathways in blue (intermediates after which the different glutamate pathways are named are highlighted), and lactate and succinate fermentation in purple and pink, respectively. See main text for key enzymes and bacteria harbouring the different pathways. Redox reactions which involve electron carriers are indicated by [H]. CoA-transferase-mediated reactions are indicated by . As indicated, co-substrates other than acetate may operate in CoA-transferase reactions in some bacteria (for further detail see main text). CoA, coenzyme A; P, bound phosphate; Pi, inorganic phosphate; PEP, phosphoenolpyruvate; (B12), enzyme dependent on vitamin B12. Dotted line indicates that several intermediate steps are involved.

Table 1. Capabilities for butyrate and propionate production among dominant bacterial species detected in faecal samples of human subjects (Qin et al., 2010; Zhernakova et al., 2016).
Phylum (family) Species Butyratea Propionateb
Bacteroidetes (Bacteroidaceae) Bacteroides uniformis + (Suc)
Bacteroides vulgatus + (Suc)
Bacteroidetes (Prevotellaceae) Prevotella copri + (Suc)
Bacteroidetes (Rikenellaceae) Alistipes putredinis + (Suc)
Firmicutes (Lachnospiraceae) Eubacterium rectale + (CoAT)
Roseburia inulinivorans + (CoAT) + (Pdu)
Roseburia intestinalis + (CoAT)
Dorea longicatena
Eubacterium hallii + (CoAT) + (Pdu)
Anaerostipes hadrus + (CoAT)
Ruminococcus torques
Coprococcus eutactus + (ButK)
Blautia obeum + (Pdu)
Dorea formicigenerans
Coprococcus catus + (CoAT) + (Acr)
Firmicutes (Ruminococcaceae) Faecalibacterium prausnitzii + (CoAT)
Subdoligranulum variabile + (ButK)
Ruminococcus bromii
Eubacterium siraeum
Firmicutes (Veillonellaceae) Dialister invisus + (Suc)
Firmicutes (Acidaminococcaceae) Phascolarctobacterium succinatutens + (Suc)
Firmicutes (Erysipelotrichaceae) Eubacterium biformec + (CoAT)
Actinobacteria (Bifidobacteriaceae) Bifidobacterium adolescentis
Bifidobacterium longum
Actinobacteria (Coriobacteriaceae) Collinsella aerofaciens
Verrucomicrobia (Verrucomicrobiaceae) Akkermansia muciniphila + (Suc)
  • a. −, absent; +, present; ButK, butyrate kinase route; CoAT, butyryl-CoA:acetate CoA-transferase route.
  • b. −, absent; +, present; Acr, acrylate pathway; Pdu, 1,2-propanediol pathway; Suc, succinate pathway (succinate may be the major product formed instead of propionate in some species and/or under some growth conditions).
  • c. Reclassified as Holdemanella biformis (De Maesschalck et al., 2014). CoA-transferase route is proposed based on closely related butyrate producers within the Erysipelotrichaceae (see also main text, section on CoA-transferases).

Ruminococcaceae

Faecalibacterium prausnitzii, one of the most abundant species present in the healthy human microbiota, produces butyrate via butyryl-CoA:acetate CoA-transferase with net consumption of acetate, and acetate stimulates its growth on carbohydrate energy sources (Duncan et al., 2002). While F. prausnitzii strains are obligate anaerobes, they also show growth stimulation by low concentrations of oxygen in the presence of riboflavin and reduced compounds such as cysteine or glutathione (Khan et al., 2012). It is hypothesized that this ability may provide a niche for the bacterium to thrive in the proximity of the colonic wall, where oxygen is diffusing in from the bloodstream. Oxygen consumption is accompanied by a decrease in butyrate formation (Khan et al., 2012). F. prausnitzii isolates show limited ability to utilize dietary polysaccharides such as starch and hemicellulose for growth, but some strains utilize inulin and pectin derivatives and the ability to utilize uronic acids is widespread (Lopez-Siles et al., 2012). F. prausnitzii is depleted in inflammatory bowel disease patients, especially Crohn's disease, and evidence that it has anti-inflammatory action has attracted interest in this species as a potential therapeutic (Quévrain et al., 2016). Similarly Butyricicoccus pullicaecorum is also reported to be less abundant in inflammatory bowel disease patients, and might also have therapeutic potential (Eeckhaut et al., 2013). Butyrate production has been reported for other Ruminococcaceae (Table 1), but rather little is known about most of these organisms.

Lachnospiraceae

Eubacterium rectale and the closely related Roseburia species constitute a major group of butyrate-producing Firmicutes that share the butyryl-CoA:acetate CoA-transferase route for butyrate production and the same genomic organization of their butyrate synthetic genes from acetyl-CoA to butyryl-CoA (Louis and Flint, 2009). In some Roseburia strains, particularly at mildly acidic pH, butyrate is almost the sole fermentation acid produced, with net consumption of acetate typically accompanying the formation of butyrate (Kettle et al., 2015). Other strains and species produce formate and lactate in addition to butyrate (Louis and Flint, 2009). Genome analysis reveals that there is considerable capacity within this group to utilize diet-derived polysaccharides including starch, arabinoxylan and inulin, that varies substantially between strains and species (Sheridan et al., 2016).

Butyrate-producing Lachnopiraceae show considerable divergence in their phylogeny, gene organization and physiology (Louis and Flint, 2009) (Table 1). Other Lachnospiraceae that possess the butyryl-CoA:acetate CoA-transferase gene include Eubacterium hallii, Anaerostipes hadrus, Coprococcus catus, un-characterised species related to isolates SS3/4 and M62/1, and some uncultured organisms (Louis et al., 2010; Reichardt et al., 2014). Two species of Coprococcus, in common with many Clostridium species that belong to other families of Firmicutes, use the butyrate kinase rather than CoA-transferase enzyme for the final step in butyrate formation (Louis et al., 2004; Louis and Flint, 2009). E. rectale and E. hallii are among the 10 most abundant species reported in the human faecal microbiota (Qin et al., 2010; Walker et al., 2011) (Table 1) and together accounted for 44% of butyryl-CoA:acetate CoA-transferase sequences amplified from faecal samples of 10 healthy volunteers (Louis et al., 2010).

Lactate can be produced from carbohydrates by many different gut bacteria (Duncan et al., 2004). In vitro incubations of 13C lactate with human intestinal microbiota show that the label is recovered in acetate, propionate and butyrate. The proportions of these products can vary widely, however, with acidic pH favouring butyrate (Belenguer et al., 2007). In addition, there is considerable inter-individual variation in the fate of 13C lactate, which is assumed to reflect differences in the relative abundance of lactate-utilising species within the microbiota (Bourriaud et al., 2005; Morrison et al., 2006). Certain Lachnospiraceae have the ability to grow in the presence of lactate and acetate to produce butyrate, showing an overall net stoichiometry of 4 mol of lactate and 2 mol of acetate producing 3 mol of butyrate (Duncan et al., 2004). These include the abundant species A. hadrus, which uses only D-lactate (Allen-Vercoe et al., 2012) and E. hallii, which is able to utilize both lactate isomers (Duncan et al., 2004; Muñoz-Tamayo et al., 2011). Lactate oxidation to pyruvate by direct reduction of NAD+ is energetically unfavourable. Anaerobic lactate utilisers carry a lactate dehydrogenase that operates in complex with an electron-transferring flavoprotein that couples the endergonic NAD+ reduction to ferredoxin oxidation, in a process called electron confurcation (Weghoff et al., 2014).

Pathways and bacterial groups contributing to propionate formation from carbohydrates

Two pathways are known for the formation of propionate from sugar fermentation by gut bacteria. Most hexose and pentose sugars are processed through the succinate pathway (Fig. 2) whereas the deoxy sugars fucose and rhamnose are metabolized by the propanediol pathway (Fig. 3).

Details are in the caption following the image

Microbial pathways for propionate formation from carbohydrates, organic acids and amino acids. As indicated, amino acids capable of conversion to pyruvate can also give rise to butyrate (Fig. 1). Carbohydrate fermentation to pyruvate via glycolysis is shown in green, propionate formation via the succinate pathway in black, amino acid fermentation pathways in blue, and acrylate pathway for lactate utilisation in purple. See main text for key enzymes and bacteria harbouring the different pathways. Redox reactions which involve electron carriers are indicated by [H]. CoA-transferase-mediated reactions are indicated by • (*may be performed by a CoA-transferase or CoA-ligase reaction). Propionate formation from propionyl-CoA in the succinate pathway may involve either a CoA-transferase or phosphate propanoyltransferase/propionate kinase reaction. CoA, coenzyme A; PEP, phosphoenolpyruvate; (B12), dependent on vitamin B12. Dotted lines indicate that several intermediate steps are involved.

Details are in the caption following the image

Microbial pathways for propionate formation via 1,2-propanediol. Carriage of the different pathways in gut microbes is indicated by colour. Redox reactions which involve electron carriers are indicated by [H]. Propionate formation from propionyl-CoA may involve either a CoA-transferase or phosphate propanoyltransferase/propionate kinase reaction (indicated by a dashed line). Grey hexagon indicates that the reaction is carried out in polyhedral bodies to sequester toxic intermediate propionaldehyde. CoA, coenzyme A; P, bound phosphate; (B12), dependent on vitamin B12. Dotted line indicates that several intermediate steps are involved.

The succinate pathway is found mainly in Bacteroidetes and in the Negativicutes class of Firmicutes (Reichardt et al., 2014). It is the major route for propionate formation from dietary carbohydrates driven by the abundant Bacteroidetes, and relative Bacteroidetes abundance was found to correlate with relative faecal propionate levels in human volunteers (Salonen et al., 2014). Succinate is a precursor of propionate, but can accumulate in cultures of Bacteroides spp. under growth conditions where phosphoenolpyruvate carboxykinase is repressed, for example, at high pCO2 and high dilution rates (Caspari and Macy, 1983). Conversion of succinate to propionate also requires vitamin B12 and succinate has been shown to accumulate in B12-depleted cultures of Prevotella ruminicola (Strobel, 1992). Some species of Bacteroidetes, notably Prevotella copri, apparently produce succinate rather than propionate as their main fermentation product and succinate accumulation has been reported particularly in the rat gut (De Vadder et al., 2016). The succinate pathway is known to be present in some Ruminococcaceae, such as Ruminococcus flavefaciens, which also produces succinate rather than propionate as the end product (Macfarlane and Gibson, 1997). On the other hand, some human colonic bacteria belonging to the Negativicutes class of Firmicutes (e.g., Phascolarctobacterium succinatutens; Watanabe et al., 2012), have the ability to convert succinate to propionate (Flint et al., 2014; Reichardt et al., 2014). This activity may explain why succinate accumulation is infrequently reported for human faecal samples, although 3 of the 14 overweight human volunteers in one recent dietary study showed elevated faecal succinate concentrations (>30 mM) in samples from a non-starch polysaccharide-supplemented diet (reported in Salonen et al., 2014, Supporting Information). Other Negativicutes bacteria convert lactate to propionate either via the succinate pathway (e.g., Veillonella spp.) or via the acrylate pathway (Megasphaera elsdenii) (Reichardt et al., 2014) (Fig. 2). The acrylate pathway has also been shown to operate recently in a species of Lachnospiraceae, Coprococcus catus (Reichardt et al., 2014).

Formation of propionate and propanol from the deoxy sugars rhamnose and fucose via the propanediol pathway has been demonstrated in dominant gut commensal bacteria belonging to the Lachnospiraceae, including Roseburia inulinivorans and Blautia species (Scott et al., 2006; Reichardt et al., 2014) (Table 1, Fig. 3). Metabolism of rhamnose and fucose via this pathway has also been reported for Salmonella and Listeria species (Xue et al., 2008). Other bacteria, including Bacteroides species, Escherichia coli and Anaerostipes rhamnosivorans, are able to degrade deoxy sugars via the propanediol pathway, but produce the pathway intermediate 1,2-propanediol as the final product (Saxena et al., 2010; Rodionova et al., 2013; Bui et al., 2014). 1,2-Propanediol can also be produced from other sugars via the glycolysis intermediate dihydroxyacetone-phosphate and methylglyoxal by microbes including Escherichia coli, Clostridium sphenoides and the yeast Saccharomyces cerevisiae (Bennett and San, 2001; Saxena et al., 2010). Methylglyoxal is further metabolised to 1,2-propanediol either via lactaldehyde or via hydroxyacetone (Fig. 3). In C. sphenoides it has been shown that 1,2-propanediol formation via dihydroxyacetone-phosphate operates under phosphate limitation and it remains to be established whether it plays a major role in the gut environment. A third pathway for 1,2-propanediol production via lactaldehyde operates from lactate in Lactobacillus buchneri. The pathway has been elucidated in a strain isolated from maize silage (Gänzle, 2015), but this species has also been detected in the human gut (Mikelsaar et al., 2016).

E. hallii and Lactobacillus reuteri, although unable to grow on fucose or rhamnose, are nevertheless able to utilise 1,2-propanediol to produce propionate and propanol (Gänzle, 2015; Engels et al., 2016) (Fig. 3). Furthermore, metagenomic mining for dehydratases has indicated that further gut anaerobes, including Flavonifractor plautii, Intestinimonas butyriproducens and Veillonella spp. may also be able to produce propionate from this substrate (Engels et al., 2016). Thus, cross-feeding of the intermediate 1,2-propanediol between different bacteria may play an important role in the production of propionate from deoxy sugars. The conversion of 1,2-propanediol to propionate, which is dependent on vitamin B12, takes place in polyhedral bodies, microcompartments that sequester the toxic pathway intermediate propionaldehyde (Chowdhury et al., 2014). Interestingly, glycerol is converted to 1,3-propanediol and 3-hydroxypropionate in L. reuteri and E. hallii by the same dehydratase that acts on 1,2-propanediol (Gänzle, 2015; Engels et al., 2016) indicating that glycerol utilization may be the primary function of this enzyme in these species. It is also worth noting that the pathway intermediate 3-hydroxypropionaldehyde, also known as reuterin, is a potent antimicrobial compound (Gänzle, 2015).

Butyrate and propionate formation from proteins and amino acids

Propionate and butyrate are also formed as products from peptide and amino acid fermentation (Figs. 1 and 2), although the numbers of amino acid-fermenting bacteria have been estimated to constitute less than 1% of the large intestinal microbiota (Smith and Macfarlane, 1998; Dai et al., 2011). It is estimated that the colon receives approximately 13 g of protein and peptides per day, and large amounts of soluble protein and peptides were found in intestinal contents of sudden death victims (Smith and Macfarlane, 1998). Peptides seem to be preferred over free amino acids by gut bacteria. Low gut pH and the presence of carbohydrates reduces peptide and amino acid fermentation in vitro, which helps to explain why microbial amino acid fermentation is higher in the distal than the proximal colon contents (Smith and Macfarlane, 1998). Amino acid fermentation leads to the production of potentially harmful metabolites (for example phenolic and indolic compounds, amines, ammonia) in addition to branched-chain fatty acids (BCFA) and SCFA (Smith and Macfarlane, 1997; Dai et al., 2011).

In vitro incubations of faecal slurries with individual amino acids showed that propionate was produced mainly from aspartate, alanine, threonine and methionine, whereas butyrate was a major fermentation product from glutamate, lysine, histidine, cysteine, serine and methionine (Smith and Macfarlane, 1997). While several Bacteroidetes have major roles in proteolysis and in propionate formation from peptides (Macfarlane and Macfarlane, 1995), certain Firmicutes species also show high activity on amino acids, notably Intestinimonas AF211, which ferments glucose and lysine to butyrate via distinct pathways (Bui et al., 2015) (Fig. 1). Several different pathways exist for glutamate degradation in butyrate-producing bacteria, which have mainly been studied in Clostridium species not originating from gut environments (Barker, 1981; Buckel, 2001). However, there is genomic and metagenomic evidence that they are also present in some gut bacteria (Potrykus et al., 2008; Vital et al., 2014). The glutamate degradation pathways enter the main butyrate pathway either via pyruvate (3-methylasparate pathway; Clostridium limosum, Fusobacterium spp.) or crotonyl-CoA (4-aminobutyrate pathway, discussed in more detail below, and 2-hydroxyglutarate pathway, found in different Firmicutes including Acidaminococcus fermentans, Clostridium sporosphaeroides, Clostridium symbiosum, Fusobacterium spp. and Peptostreptococcus asaccharolyticus) (Fig. 1). Some bacteria belonging to the Acidaminococcaceae also degrade glutamate via the 3-methylasparate pathway, but produce propionate rather than butyrate from the intermediate pyruvate (Buckel, 2001) (Fig. 2).

Glutamate degradation to 4-aminobutyrate (gamma-aminobutyrate, GABA) is carried out under acid stress to maintain intracellular pH homeostasis in a number of gut bacteria (Feehily and Karatzas, 2013), and a bacterial isolate exclusively growing on GABA has recently been found (http://www.abstractsonline.com/pp8/#!/4060/presentation/18619). As GABA also acts as a neurotransmitter, the abundance of microbes involved in the production or consumption of GABA may influence mood and behaviour. The pathway for GABA degradation is shared with succinate degradation via succinate semialdehyde and 4-hydroxybutyrate (Fig. 1), and butyrate production from succinate via this pathway has been demonstrated in Porphyromonas gingivalis and Clostridioides difficile (Ferreyra et al., 2014; Yoshida et al., 2016).

The fermentation routes of other amino acids are less well understood. Histidine is converted to glutamate (Potrykus et al., 2008; Kanehisa et al., 2016), which is in agreement with high levels of butyrate being formed from histidine by faecal microbiota (Smith and Macfarlane, 1997). Alanine, serine and cysteine are broken down to pyruvate (Potrykus et al., 2008; Carbonero et al., 2012), thus product formation depends on the bacterium utilizing those amino acids and their corresponding fermentative pathways. For example, in Clostridium propionicum, alanine fermentation leads to the production of propionate via pyruvate, lactate and the acrylate pathway (Buckel, 2001) (Fig. 2). Threonine and methionine are converted to 2-oxobutyrate, which leads to propionate formation (Fig. 2) (Barker, 1981; Smith and Macfarlane, 1997; Kanehisa et al., 2016). Several routes for the breakdown of asparate exist, via alanine, threonine, oxaloacetate or fumarate (Smith and Macfarlane, 1997; Kanehisa et al., 2016) (Fig. 2), which accounts for the fact that it is mainly converted to propionate in in vitro incubations.

Role of CoA-transferases in SCFA metabolism

Propionate and butyrate can be generated from their respective CoA thioesters either by transfer of the CoA-moiety onto another metabolite, or by conversion via propionyl-phosphate or butyryl-phosphate. The second (kinase) route leads to the generation of ATP, but the CoA-transferase route also conserves the energy of the CoA bond in the newly formed CoA-derivative of the co-substrate. Acetate is a common co-substrate in CoA-transferase reactions, and the high acetate concentrations in the large intestine provide a possible explanation for the prevalence of the butyryl-CoA:acetate CoA-transferase route in gut microbes (Louis et al., 2004) (see also section on pH below). Bacteria often carry multipble different CoA-transferases in their genomes, with Intestinimonas AF211 encoding at least 14 such enzymes (Bui et al., 2015). It can be difficult to pin-point which gene is responsible for SCFA formation, especially as CoA-transferases tend to have broad substrate specificity. For example, the purified butyryl-CoA:acetate CoA-transferase (butCoAT gene product) from Roseburia hominis has a similar affinity for butyryl-CoA and propionyl-CoA although the enzyme is clearly responsible for butyrate formation in this species (Charrier et al., 2006) (Table 2). Gene expression evidence in Intestinimonas AF211 suggested that the enzyme AtoD-A, responsible for butyryl-CoA:acetoacetate CoA-transferase activity, plays a key role in conversion of lysine to butyrate, while the butCoAT gene product mediated the final step in butyrate formation from glucose (Bui et al., 2015). In Clostridium aminobutyricum, a CoA-transferase that acts on 4-hydroxybutyrate and butyryl-CoA links the final step of butyrate production to the formation of 4-hydroxybutyryl-CoA further up in the glutamate fermentation pathway (Buckel, 2001). Similarly, C. propionicum links the formation of lactoyl-CoA in the acrylate pathway to propionate formation via a CoA-transferase (Buckel, 2001). There are also instances where different CoA-transferases appear to have evolved for the same enzymatic reaction. Thus, bacteria belonging to the Erysipelotrichaceae do not carry a gene closely related to the butyryl-CoA:acetate CoA-transferase identified in other Firmicutes. Instead, a gene more closely related to propionate CoA-transferases is thought to be responsible for butyrate formation in these organisms (Eeckhaut et al., 2011).

Table 2. Activity of the butyryl-CoA:acetate CoA-transferase of Roseburia hominis A2-183 (purified recombinant butCoAT gene product expressed in Escherichia coli (Charrier et al., 2006)).
Km (mM) Vmax (µmol/min/mg protein) Inhibition by competition with acetate (%)a
Acetate 6.4
Butyryl-CoA 0.098 112
Propionyl-CoA 0.099 51
Butyrate 75
Propionate 70
Isobutyrate 56
Valerate 28
  • a. No significant inhibition was found for caproate, 3-hydroxybutyrate, 4-hydroxybutyrate, 4-aminobutyrate, lactate, acetoacetate and succinate.

Impact of the gut environment

pH

Gut pH has a major impact on competition between different groups of bacteria within the microbial community. In pH-controlled in vitro continuous culture experiments with soluble polysaccharide provided as the main energy source, mildly acidic pH has been shown to curtail the growth of Bacteroides spp. relative to Firmicutes and Actinobacteria (Walker et al., 2005; Chung et al., 2016). This is because human colonic Bacteroides spp. are generally less able than many dominant Firmicutes to tolerate the presence of short chain fatty acids at pH 5.5 (Duncan et al., 2009). This selective inhibition and the resulting shift in community composition has the consequence of limiting propionate formation and enhancing butyrate production by the community at pH values around 5.5 compared with 6.5–6.8 (Walker et al., 2005; Chung et al., 2016). The impact of pH shifts upon experimentally observed butyrate and propionate concentrations has been successfully modelled mathematically, based on the differing tolerance to low pH of the major bacterial functional groups that comprise the human colonic microbiota (Kettle et al., 2015).

For bacteria that use the butyryl-CoA:acetate CoA-transferase route, acetate consumption and butyrate production are reported to increase at mildly acidic pH compared with near neutral pH (Kettle et al., 2015). Although conversion of glucose to butyrate, 2 CO2 and 2 H2 can occur with no net uptake of acetate (Gottschalk, 1979), net acetate uptake is typically observed for species of Roseburia and F. prausnitzii. Theoretical stoichiometries involving net acetate uptake are shown in Fig. 4A, which also assumes that some of the reducing power that is generated drives proton export, increasing the ATP yield per glucose fermented (Buckel and Thauer, 2013). Incorporation of exogenous acetate via the CoA-transferase reaction results in some loss of ATP production via acetyl-phosphate, but this is more than compensated by the additional ATP formed from proton export, giving a potential maximum of 4 ATP formed per glucose metabolized when 2 mol of acetate are taken in for each mol of glucose fermented. Interestingly, Fig. 4B shows that the predictions from these stoichiometries (based on the generalised equation shown in Fig. 4A) fit experimental data for the impact of pH on metabolites produced by F. prausnitzii and two Roseburia spp. in anaerobic batch culture (Kettle et al., 2015). Thus low pH (5.5) tends to increase acetate uptake and butyrate production while near neutral pH (6.7) has the opposite effect. It seems possible that the increased ATP gain associated with net acetate uptake helps to compensate for the effects of low pH and might account for the reliance on the CoA-transferase route for butyrate formation in these bacteria.

Details are in the caption following the image

Butyrate production in bacteria that use the butyryl-CoA:acetate CoA-transferase route. A: General equation for the relationship between acetate consumption and butyrate production, assuming no lactate or formate are produced (modified from Louis and Flint, 2009 and Kettle et al., 2015). Etf, electron-transferring flavoprotein; Fd, ferredoxin; P, bound phosphate; Pi, inorganic phosphate. B: Alternative stoichiometries for butyrate production based on A. Experimental data (coloured symbols) refer to R. intestinalis L1-82, R. hominis A2-183 and F. prausnitzii A2-165 grown at three different initial pH values (5.5, 6.2, 6.7) (Kettle et al., 2015 and Sylvia Duncan, personal communication).

Growth requirements

It has been show in a rodent model that limitation of dietary iron intake can dramatically decrease the production of both butyrate and propionate as lactobacilli and Proteobacteria are favoured (Dostal et al., 2012). Populations of Roseburia-related butyrate producers appear particularly sensitive to iron availability, while in pure cultures of R. intestinalis butyrate production was favoured at high iron concentrations with a switch to lactate production under iron-deficient conditions (Dostal et al., 2015). It remains to be established whether other growth factors also have a major impact on SCFA formation.

Intestinal gases

SCFA formation is also likely to be affected by differences in oxygen concentration in different regions and micro-compartments of the gut due to differences in oxygen sensitivity and metabolic capacity between microbes, as exemplified by the peculiar relationship of F. prausnitzii with oxygen (discussed above). Furthermore, the abundance of microbes consuming hydrogen and thereby influencing the hydrogen partial pressure in the gut also influences SCFA formation, as this affects the overall balance of fermentation products formed (Macfarlane and Macfarlane, 2003; Wolf et al., 2016).

Concluding remarks

Huge advances have been made in recent years in our understanding of SCFA metabolism in the human gut, and many of the dominant propionate- and butyrate-producing bacteria are available in culture, enabling detailed investigations into their metabolism. Recent work has emphasized that butyrate and propionate can arise from fermentation both of amino acids and of carbohydrates, but the relative contributions of protein and carbohydrate fermentation in vivo over the wide range of ‘normal’ human dietary intakes is not yet clear. We know that high protein, low carbohydrate weight loss diets lead to a disproportionate decrease in butyrate among total faecal SCFA, together with an increased proportion of branched chain fatty acids that are wholly derived from branched chain amino acids and therefore provide an indicator of protein fermentation (Duncan et al., 2007; Russell et al., 2011). This suggests strongly that butyrate production is mainly determined by the supply of non-digestible carbohydrates, rather than by protein fermentation. This may however reflect the particular ecology of butyrate-producing bacteria, as discussed above. In the case of propionate, on the other hand, the major producers of propionate from dietary carbohydrates, the Bacteroidetes, are also important peptide fermenters and the propionate proportion among faecal SCFA was not decreased by such low carbohydrate diets (Duncan et al., 2007). It is also clear that compounds normally regarded as intermediates (e.g., succinate, lactate) may accumulate in certain individuals or in particular conditions. This makes it important also to consider the impacts of these metabolites on the host, as for example in the case of succinate which it is suggested may provide health benefits (De Vadder et al., 2016). Lactate is detected as a major fermentation product in breast-fed infants whose microbiota is dominated by Bifidobacterium spp. In adults, however, lactate accumulation is associated with dysbiosis, for example, in severe colitis (Hove et al., 1994), that may result in part from a lack of lactate-utilizing bacteria (Belenguer et al., 2007).

The ever-increasing availability of genomic and metagenomics sequences is a highly useful resource to foster our understanding of microbial metabolism in the gut, but care has to be taken with assigning function to genes by sequence analysis, which should ideally be complemented by evidence from genetic or enzymatic studies. A renewed interest in isolation and study of gut bacteria (Walker et al., 2014, Browne et al., 2016) together with novel systems for gene transfer and knockout on the horizon will enable a thorough understanding of the different members of the microbial community. This will benefit in vitro and in vivo microbial community-based studies to foster our understanding of the different ecological niches of the community members, how they interact with each other and how we can modulate the system by dietary means to optimize SCFA production. The fact that, in general, different phylogenetic groups of bacteria are responsible for butyrate and propionate production suggests that there may be scope for differentially manipulating their production by the gut microbiota.

Acknowledgements

The authors receive financial support from the Scottish Government Rural and Environmental Sciences and Analytical Services (RESAS). We would like to thank Sylvia Duncan for SCFA data of F. prausnitzii grown under different pH regimes and for critically reading the manuscript.

The authors declare that they have no conflicts of interest.