Gut Microbiome and Metabolism

35.2.1

Carbohydrate Degradation

Mammalian hosts lack many enzymes needed to digest complex carbohydrates present in many of the plants they ingest. Hence, ruminants are largely, if not absolutely, dependent upon gut bacteria to breakdown and ferment ingested complex polysaccharides. While other herbivores and omnivores are less reliant on microbial fermentation, it has nonetheless been estimated that 10% of the dietary energy intake of humans is dependent upon gut bacteria. Breakdown of carbohydrates, including glycoproteins, oligosaccharides, and polysaccharides, is mediated by a group of enzymes collectively referred to as carbohydrate-active enzymes (CAZymes). The chemistry of CAZymes is a highly complex area whose understanding continues to evolve rapidly and thus can best be monitored by the related web sites cazy.org and cazypedia.org , which are maintained by a collective of carbohydrate chemists. Classes of CAZymes include glycoside hydrolases, which hydrolyze or rearrange glyosidic bonds, polysaccharide lysases, which catalyze the nonhydrolytic cleavage of glyosidic bonds, and carbohydrates esterases, which catalyze the deacylation of saccharides ( Fig. 35.1 ). As of 2017, there are 145 families of glycoside hydrolases, 27 families of polysaccharide lysases, and 16 families of carbohydrates esterases. For each of these families, hundreds of distinct members have been identified in bacteria and comparatively few in eukaryotes. They are also important in the digestion of carbohydrate binding molecules (CBM), which play a diverse role in the sequestration of these molecules and are presumed to aid in their breakdown. The CAZY database lists over 100,000 identified CBM classified into 81 families. Analogous to CAzymes, most known CBM are of bacterial origin.

Identification of CAZ and CBM has largely relied on sequence-based metagenomics approaches, which make it very difficult to assign roles of specific molecules in carbohydrate breakdown; but it is presumed to be a highly collaborative process involving many enzymes in many bacterial species whereas individual bacteria lack the wherewithal to efficiently metabolize many individual complex saccharides. Such interactions are best understood in the rumen wherein, for example, degradation of the plant cell wall component xylan requires five distinct enzymatic activities suggested to be provided by a range of Ruminococcus and Prevotella species. Microbiota also play a central role in the degradation of complex carbohydrates in humans. While simple carbohydates are readily broken down in the upper GI tract, approximately 20–60 g of ingested carbohydrates are thought to have reached the colon. Such carbohydrate includes resistant starch, which consists largely of plant-derived oligosaccharides. Enzymes and bacteria involved in the metabolism of resistant starch in the human intestine are not well defined but, nonetheless, such resistant starches are fermented in the colon and give rise to beneficial molecules discussed below. The breakdown of plant cell wall components such as cellulose in the human gut is even less well understood. Briefly, humans lack enzymes needed to degrade cellulose but yet are able to breakdown this complex carbohydrate. For example, a study wherein participants were adapted to a cellulose-enriched diet found that about 70% of cellulose was degraded as it passed through the GI tract although this result is variable in terms of both individuals studied and the specific type of fiber utilized. In any case, bacteria with the capacity to degrade cellulose have been isolated from the human intestine. Such bacteria were classified as Ruminococcus , Clostridia , and unknown species.

In contrast to cellulose, a range of dietary fibers are readily fermented by gut bacteria and are, accordingly referred to as fermentable or soluble fibers. Such readily fermentable fibers include fructans, galactooligosaccharides (GOS), and fructooligosaccharides (FOS). These fibers are efficiently metabolized by a range of gut bacterial species whose abundance is, consequently, increased upon the consumption of such fibers. Such changes in microbiota composition, and how it impacts metabolism, will be discussed below. Here, in terms of the role of the microbiota in breaking down such products, it is clear from mouse studies that, in the absence of a microbiota, fibers such as inulin that are efficiently fermented in conventional mice are not degraded in germfree mice. The breakdown of such fibers requires a range of distinct glycosidase activities. GOS are especially abundant in human breast milk and accordingly, a range of CAZ needed to degrade GOS are expressed by a range of Bifidobacterium species. Specifically, Bifidobacteria produce multiple glycosidases, transporters, and metabolic enzymes that enable it to use an unusually wide variety of carbohydrates. Collectively, these proteins comprise a central fermentative pathway referred to as the “Bifid shunt,” which is presumed to give Bifidobaterium a great growth advantage in a GOS-enriched diet and may produce a range of molecules that benefit the host.

35.2.2

Role of Microbial-Mediated Carbohydrate Breakdown in Human Metabolism

The general importance of microbial-mediated carbohydrate breakdown in nonruminant mammals can be appreciated by the aforementioned estimate that it accounts for 10% of total caloric intake in humans. The impact of such a difference can be seen in the pioneering work of Jeff Gordon and colleagues, who compared conventional and germfree mice, which were fed the same autoclaved chow. Germfree mice ate over 35% more chow to achieve similar total body weights but yet stored less energy in that they had a 40% reduction in fat mass. Germfree mice also exhibit a near-complete loss of colonic short-chain fatty acids (SCFAs), which are an efficiently used energy source for mammalian hosts, and an increase in unharvested calories that remained in the feces of germfree mice. Together, these results strongly suggest that complete absence of a microbiota reduces efficiency of energy harvest in a manner that contributes to altered state of energy balance characterized by leanness despite high levels of food consumption. The extension of this logic to the less extreme case of a host that has a microbiota has led Gordon and colleagues to postulate that not all microbiotas are equal in their capacity to harvest energy but rather that some microbiotas do a better job of breaking down complex carbohydrates that are available to them and hence microbiota composition is a determinant of energy intake in general and, consequently, a risk factor of adiposity in particular. Evidence that supports this notion includes that both obese mice, including mice that are obese due to impaired satiety signaling, namely leptin-deficient mice (Ob/Ob) and obese humans have alterations in gut microbiota that are associated with the obese state, which is itself associated with higher levels of cecal SCFA and lower levels of calories in feces indicating that the microbiotas of the obese hosts are indeed associated with increased energy harvest. Furthermore, transfers of such microbiotas from obese humans to germfree mice transferred increased microbial fermentation, as read out by SCFA levels relative to germfree mice conventionalized with microbiotas from lean hosts. Furthermore, germfree mice receiving microbiotas from obese mice displayed reduced levels of energy in their feces, as assayed by bomb calorimetry, indicating they had more completely harvested the potential energy available in food they had ingested. Lastly, transplant of such microbiotas with “increased capacity for energy harvest” resulted in increased adiposity in mice receiving these microbiotas. Metagenomic analysis of feces both donors and recipients of such microbiotas indicate that such increased energy harvest correlates with a microbiota enriched in bacterial genes-related breakdown of complex carbohydrates.

The notion that microbiota composition is a determinant of energy harvest is supported by agricultural practices which find that feed efficiency, in this case, defined as total saleable meat/feed cost is enhanced by administering low-dose antibiotics although it is not clear if this increases energy harvest per se—this question is more difficult to address in large farm animals wherein many potentially lost calories are emitted in gasses such as methane—or might reflect that the antibiotics make the animals grow faster, thus reducing calories used to sustain life.

Another observation in humans that supports the importance of microbiota composition as a determinant of energy harvest is a study from Walker and colleagues, who observed that, among a group of healthy volunteers there was heterogeneity in microbiota composition that correlated with the ability to ferment dietary resistant starches. Specifically, while most individuals fermented over 96% of administered resistant starch, some donors, particularly ones who harbored very few Ruminococci species had over 60% of the resistant starch pass through their GI tracts unfermented. Together, these observations highlight the potential of the microbiota to impact metabolism by energy harvest. Yet, this is a complex area wherein some mouse studies have found microbiota not to be necessary for diet-induced obesity, while others have found it is strain dependent. Moreover, some human studies, including ones from Gordon’s group, report that more efficient metabolism of carbohydrates to SCFAs associates with a lean state. In any case, as will be discussed below, there are a number of other ways the microbiota can also impact metabolism and the relative importance of these different mechanisms in humans, in most situations, is not yet clear.

35.2.3

Role of Microbiota in the Breakdown of Proteins

The role of the microbiota in the metabolism of proteins has been studied much less extensively than that of carbohydrates although both nutrient types are largely catabolized simultaneously by host and microbial enzymes. In contrast to the case of carbohydrates wherein mammals, without help from bacteria, lack the wherewithal to degrade many complex polysaccharides, both bacteria and mammals are capable of catabolizing most proteins. Hence, the relative contributions of bacterial versus mammalian proteases in hydrolyzing proteins to amino acids and small peptides are not clear. In any case, a significant portion, perhaps 10%–20% of the amino acids derived from ingested proteins is thought to be initially be taken up by bacteria in the small intestine although tracer-based studies used to reach these estimates are prone to a wide variation. Amino acids taken up by bacteria can be used for anabolic purposes (largely protein synthesis) or catabolism to make various structural molecules (amines, polyamines) or deaminated to make ketoacids and various lipids, which are primarily used by the bacteria themselves ( Fig. 35.2 ). Additionally, bacteria are capable of de novo synthesis of most (to some extent) all amino acids and labeling studies have shown, for example, that a significant portion of nitrogen from ammonium chloride administered to mice ended up in host proteins in conventional but not germfree mice. Thus, while the net consequence of much of bacterial metabolism of proteins would seem to be influenced by the availability of amino acids and small peptides to the host although the extent to which is actually the case is not well defined. In any case, a number of products of amino acid breakdown are released by bacteria that have a range of impacts on host metabolism.

Some deaminated amino acids are ultimately released as products used by enterocytes primarily for purposes of supplying energy such as short-chain and branched-chain fatty acids. Indeed, these are preferred fuels of enterocytes. Moreover, four other classes of compounds present in the colon are largely the consequence of bacterial metabolism, namely ammonia, hydrogen sulfide, polyamines, and indols/phenols. Ammonia is abundant in the colon in feces in most animals and there is strong evidence that the source is gut bacterial catabolism of peptides. Levels of ammonium, parallel concentration of bacteria increase with increased protein consumption and are greatly reduced by depletion of gut bacteria by antibiotics or germfree approaches. Indeed, the lack of fecal ammonium is the primary reason cages of germfree mice can be changed much less frequently than conventional mice. Ammonia is produced via urease acting on urea, which is resulting from the deamination of amino acids. While urease from Helicobacter pylori is the best-studied form of this enzyme, numerous other bacteria, including species common in the colon are able to produce ammonia suggesting these bacteria possess this enzyme. While ammonia can be taken up by enterocytes, and can be used to make glutamine, ammonia is generally thought to interfere with an array of basic intracellular metabolic processes and hence requires detoxification or excretion. Another intestinal gas produced by gut bacteria is hydrogen sulfide (H ₂ S). H ₂ S can be produced from sulfur-containing amino acids by fermentation by a well-defined set of chemical reactions. Production of H ₂ S from proteins can be mediated by a range of common gut bacteria including Clostridia and Enterobacter species. While there are a number of other routes for bacteria to produce H ₂ S, that elevation of protein consumption increases H ₂ S levels argues that breakdown of proteins play a significant role in vivo. Like ammonia, H ₂ S is considered a metabolic troublemaker. Hence, there are number of means by which the host sequesters and detoxifies H ₂ S but, nonetheless high levels of both H ₂ S and ammonia have been proposed to contribute to a range of disease states. Polyamines can be produced from decarboxylated amino acids by a range of bacteria, although synthetic pathways have primarily been studied in Escherichia coli. While human cells also produce a range of polyamines, some polyamines, for example, cadaverine, are microbial specific and accordingly, levels are eliminated by antibiotic treatment. While polyamines are increasingly appreciated to play an important role in a number of processes related to host-microbial interactions, including the formation of biofilms, the extent to which bacterial-produced polyamines play essential or detrimental roles in host metabolism is not well defined.

One of the starkest means by which microbiota metabolism of ingested food results in production of metabolites whose generation depends upon the presence of the microbiota is via acting upon aromatic amino acids. Specifically, studies by Smith and McFarlane demonstrated that a range of anaerobic bacteria present in the colon, including Bacteroides , Lactobacillus , Bifidobacterium , and Clostridia species, can ferment aromatic amino acids. As outlined in Fig. 35.3 , tyrosine is a range of compounds to end up as phenol, ethylphenol, propylphenol, and especially p-cresol, which is likely the major end product of tyrosine metabolism by bacteria in vitro and in vivo. Phenylalanine is metabolized to phenylacetate and phenylproprionate. Tryptophan is metabolized to a range of indols including indole, indolpyruvate, tryptamine, indolproprionate, and skatole. Many of these molecules are taken up by the host and can be found in blood and urine. Indeed, untargeted metabolomics studies of blood and urine generally find bacterial metabolites of aromatic amino acids to be the most enriched products in conventional mice relative to germfree animals. Specifically, a number of phenyl derivatives, including phenylsulfate, p-cresol sulfate, and phenylproprionylglycine, were readily detected in serum from conventional mice but were undetectable in germfree mice while levels of serum hippuric acid were 17-fold higher in conventional mice. Indolproprionic acids (IPA) were also found to be present only in conventional mice while serotonin, which can be derived from IPA, was 2.8-fold higher in conventional mice. Conversely, tryptophan itself was about twofold higher in germfree mice suggesting that lack of its metabolism in the absence of a microbiota increases the pool of this product for downstream metabolism by host pathways. Overall, while the relative roles of the most indol/phenolic compounds are not well understood that microbiota-derived phenols, especially hippuric acids and cresols, are excreted in urine argues that the host deems these bacterial-generated metabolites as somewhat toxic. Hence, much of the current research related to these molecules has focused on developing metabolomics based approaches to define metabolic signatures of microbial impacts upon health and disease. In accord with the general notion that altered microbiota compositions influence such outputs, transfer of distinct microbiotas to germfree mice correlated with clear changes in levels of these phenolic compounds in the host. Indols are increasing appreciated to play central roles in both bacterial-bacterial and bacterial-host communications in general and in maintaining a stable host-microbiota relationship in particular, at least in part by activating the aryl hydrocarbon receptor to drive a variety of immunomodulatory effects. A lot of further work is needed to understand how bacterial-derived indols influence host metabolism and how differences in bacterial populations in the gut influence the production of such metabolites. In any case, microbiota metabolism of aromatic amino acids seems likely to be one major means by which direct microbial action on ingested proteins broadly influences host metabolism.

35.2.4

Lipids

Numerous gut bacteria are capable of using a variety of lipids as a primary carbon source. Yet, the extent to which gut bacteria might directly take up and metabolize dietary lipids has been a subject of much research. Rather, gut bacteria are known to have an indirect but major impact on lipid metabolism via their modification of primary bile acids to secondary bile acids. Some of these effects involve lipid trafficking while others are mediated by localized and systemic receptor-mediated processes, which will be discussed below.