Review

The Devil Lies in the Details: How Variations

in Polysaccharide Fine-Structure Impact the
Physiology and Evolution of Gut Microbes

Eric C. Martens 1 , Amelia G. Kelly 1 , Alexandra S. Tauzin 2 and Harry Brumer 2

1 - Department of Microbiology and Immunology, University of Michigan Medical School, Ann Arbor, MI 48109, USA
2 - Michael Smith Laboratories and Department of Chemistry, University of British Columbia,
2185 East Mall, Vancouver, BC, V6T 1Z4, Canada

Correspondence to Eric C. Martens and Harry Brumer: [email protected]; [email protected]

http://dx.doi.org/10.1016/j.jmb.2014.06.022
Edited by J. L. Sonnenburg

Abstract
The critical importance of gastrointestinal microbes to digestion of dietary fiber in humans and other mammals
has been appreciated for decades. Symbiotic microorganisms expand mammalian digestive physiology by
providing an armament of diverse polysaccharide-degrading enzymes, which are largely absent in mammalian
genomes. By out-sourcing this aspect of digestive physiology to our gut microbes, we maximize our ability to
adapt to different carbohydrate nutrients on timescales as short as several hours due to the ability of the
gut microbial community to rapidly alter its physiology from meal to meal. Because of their ability to pick up
new traits by lateral gene transfer, our gut microbes also enable adaption over time periods as long as centuries
and millennia by adjusting their gene content to reflect cultural dietary trends. Despite a vast amount of
sequence-based insight into the metabolic potential of gut microbes, the specific mechanisms by which symbiotic
gut microorganisms recognize and attack complex carbohydrates remain largely undefined. Here, we review the
recent literature on this topic and posit that numerous, subtle variations in polysaccharides diversify the spectrum
of available nutrient niches, each of which may be best filled by a subset of microorganisms that possess the
corresponding proteins to recognize and degrade different carbohydrates. Understanding these relationships
at precise mechanistic levels will be essential to obtain a complete understanding of the forces shaping
gut microbial ecology and genomic evolution, as well as devising strategies to intentionally manipulate the
composition and physiology of the gut microbial community to improve health.
© 2014 Elsevier Ltd. All rights reserved.

Introduction symbionts' largely fermentative metabolism, which

produces short-chain fatty acids and other products
Humans consume a broad range of polysaccharide- that are absorbed in the colon as nutrients [2].
rich foods, not only in the form of plant material (cell Of the dozens of different phyla of bacteria and
walls and storage polymers) but also as animal archaea that exist on Earth, less than 10 are abundant
connective tissue, food additives and even microbial in the guts of humans [3–5]. The Gram-positive
and fungal products. Our intrinsic ability to digest Firmicutes are typically most numerous, followed by
the available repertoire of complex carbohydrate Gram-negative Bacteroidetes. Other common but
molecules remains limited to just starch, lactose and less numerically abundant phyla include Actinobac-
sucrose [1]. This metabolic decrement is due to a teria, Verrucomicrobia and Proteobacteria, among
paucity of fiber-degrading enzymes encoded in the others. The selection for a few taxonomic groups was
genomes of humans and other animals (for a recent probably ancient, since these same phyla are shared
overview, see Ref. [1]). Fortunately, we have co- among other mammals and many invertebrates
evolved with a dense consortium of symbiotic distal [4]. Moreover, at finer taxonomic levels (genus and
gut microorganisms (microbiota), many of which have species), the microorganisms found in human and
adapted to target these polysaccharides for their own animal guts are typically not present in environmental
nutrition. In return, we reap the benefits of these gut reservoirs, leading to the hypothesis that we have

0022-2836/© 2014 Elsevier Ltd. All rights reserved. J. Mol. Biol. (2014) 426, 3851–3865
3852 Gut microbial adaptations to carbohydrates

co-evolved with many of these organisms and provide substantial viable recoveries (up to 46–93%) of the
their only habitats [6]. bacterial cells observable by direct microscopic
The genomes of sequenced human gut bacteria counts [13,14]. As a testament to the experimental
and the metagenomes of the communities they skill of these scientists, the lists of most abundant
compose reveal that our microbial symbionts have human gut bacterial species that they found in early
much more extensive armaments of polysaccharide- cultivation studies share substantial overlap with
degrading enzymes than we do [1,5,7–9]. This is taxon lists generated using more modern techniques,
evident in both the numbers of enzymes present and such as 16S rDNA amplicon sequencing and
the diversity of catalytic activities [1]. As a particularly metagenomics [14–16]. A much more recent study,
striking example, the recently published 7.1-Mbp which compared large-scale anaerobic culturing
genome of Bacteroides cellulosilyticus WH2 contains to direct molecular ecology-based community enu-
a total of 424 glycoside hydrolases, polysaccharide merations, also supports the idea that most human
lyases and carbohydrate esterases, which is ~ 25 gut bacteria can be readily grown outside of the host
times the number of human genome-encoded en- [17]. Taken together, these observations reinforce
zymes that are thought to be secreted into the the imperative for continued culturing of microorgan-
gastrointestinal tract [10]. Of the 76 different CAZyme isms from the human gut so that their physiology,
(carbohydrate-active enzyme) families (as defined both alone and in communities, can be studied and
in the Carbohydrate-Active Enzymes Database [11]) understood in great depth [18].
present in B. cellulosilyticus WH2, 56 are not Bacteriological studies into the abilities of human gut
represented in the human genome, highlighting the bacteria to degrade dietary fiber and mucin polysac-
amount of metabolic expansion that even a single gut charides did not initiate in earnest until the 1970s,
bacterium adds. Without this help from symbiotic catalyzed by the seminal efforts of Salyers, Wilkins
bacteria, humans and other animals, ranging from and colleagues, which involved fermentation studies
termites to ruminants, would simply be incapable of on a large collection of cultured human gut bacteria
assimilating nutrients from a substantial portion of [15,16]. Such analyses continue to be extended to a
dietary polysaccharides. range of polysaccharides, as well as oligosaccharide
Despite a vast—and expanding—amount of components accessible through specific enzymatic
sequence-based insight, precise mechanistic rela- treatment and fractionation [19–23]. These in vitro
tionships between the enormous diversity of poly- surveys have highlighted that the Bacteroidetes
saccharides that enter our digestive system and the possess notably broad abilities to digest a diverse
microbes that degrade them have been slower to array of mostly soluble polysaccharides. In contrast,
develop. In this review, we consider several emerging more recent work by Flint and co-workers has
facets of how symbiotic gut microorganisms assist suggested that members of the Firmicutes, which
humans and other animals with polysaccharide demonstrated more limited catabolic breadth in early
digestion. We focus first on the evolutionary benefit studies, possess the ability to attack more insoluble
of this digestive symbiosis, subsequently outline the substrates that are characteristic of the natural plant
sensory and enzymatic mechanisms employed by fibers in our diets and may serve as “keystone”
various gut bacteria to distinguish these nutrients and polysaccharide degraders [24,25].
conclude by discussing some recent data that imply Moving beyond descriptive growth studies, several
the presence of finely adapted and niche-specific researchers have provided molecular insight into
microbe–polysaccharide interactions in the gut, some the enzyme-based strategies through which human
of which are being driven by the lateral transfer of gut bacteria process complex carbohydrates, such
genes involved in polysaccharide degradation. as starch, inulin and many other polysaccharides
[26–31]. These paradigms extend to various Gram-
negative and Gram-positive bacteria and have
Old Questions Still in Need of provided a framework for understanding the molec-
Detailed Answers ular processes involved. Still, however, numerous
questions remain, including (i) which species com-
The critical role of intestinal microorganisms in pete most efficiently for each available polysaccha-
polysaccharide degradation became appreciated ride, (ii) to what extent and how do the species
around the early 1940s when Robert Hungate, a present cooperate during polysaccharide degradation
pioneer in the field of anaerobic microbiology, and (iii) are dominant rumen bacterial strategies, such
explored the phenomenon of microbial cellulose as deployment of cellulosomes by Gram-positive
degradation in the bovine rumen and termite gut [12]. species, at work in the human colon? An additional
With the advent of more facile anaerobic culturing question that serves as a central focus for this review
techniques—including the development of the an- is how finely tuned are the relationships between
aerobic chamber by Freter and colleagues in the members of our gut microbiota and the myriad
1960s [13]—pioneers in human gut microbiology, chemical differences present in the polysaccharides
such as Freter, Holdeman and Moore, reported that are so important to their biology?
Gut microbial adaptations to carbohydrates 3853

Why Do We Rely on Gut Microorganisms Our symbiotic microbes also enable adaptation
to dietary changes over much longer time periods.
for Polysaccharide Digestion? Different human cultures adopt unique dietary habits
that may exist in geographically restricted popula-
Before considering more detailed mechanistic tions and persist on timescales of centuries or
aspects of gut microbiota function, it is worth reflecting millennia. These protracted periods may still be too
on this fundamental question. As discussed above, short for the human genome to evolve appropriate
the human genome, as well as the genomes of most enzymes, but given the genomic plasticity of bacte-
animals, does not inherently encode a plethora of ria, adaptation by members of the gut microbiota is
carbohydrate digestive enzymes; animal genes that more easily achieved. A striking example of such
encode plant cell wall glycosidases are particularly adaptation is thought to have occurred in bacterial
exceptional [32,33]. Why is this so? By out-sourcing members of the Japanese microbiota. In Japan and
complex carbohydrate metabolism to our symbiotic surrounding Asian cultures, the consumption of red
gut microbes, we greatly enhance our ability to algae (e.g., nori in sushi) has been commonplace for
respond to nutritional changes on timescales both over a thousand years. Likely, as a consequence of
short (minutes to hours) and long (centuries to this availability in the gut, one cultured member of the
millennia). Since the amounts and types of polysac- healthy Japanese microbiota, Bacteroides plebeius,
charides in our omnivorous diet vary from meal to has obtained the capacity to degrade porphyran, a
meal, our gut symbionts must be capable of shifting sulfated polymer of galactose that is abundant in nori
their metabolism to accommodate the diversity that is [39,40]. Further molecular analysis suggests that
offered within this limited period. The presence of such plasticity is a common feature of gut microbial
“generalists”, such as B. cellulosilyticus (vide supra), genomes, as will be elaborated below.
Bacteroides thetaiotaomicron and Bacteroides ova-
tus, highlights short-term adaptability by individual
species. These organisms each possess the ability to The Benefits (and Limitations) of “Omics”
degrade over a dozen different polysaccharides
[10,29] and therefore have the option to switch to The advent of the sequencing era has significantly
metabolism of different nutrients as they become bolstered studies of the content and dynamics of the
available. Metabolic re-organization by these species microbiota. Culture-independent approaches, such
in complex nutrient conditions has been validated as 16S rDNA and metagenomics, have provided
both in vitro [34] and in vivo in gnotobiotic mice [10,35]. a less biased and more comprehensive view of
However, it is likely that much more complex, the microbial taxa that are present. In turn, these
community-driven metabolic changes also occur on data can guide the isolation of organisms not yet
short timescales. Although it remains to be explored at represented in pure culture [18]. Sequencing of
functional levels—or on the very short timescales individual microbial genomes has provided hun-
(hours) that reflect meal-to-meal variation—short-term dreds of reference blueprints for human gut bacteria
diet change experiments in humans and mice clearly [41], which can be analyzed using bioinformatics
show that the composition of the gut community approaches to predict functionality. The degradation
changes frequently and within a day after diet switch of most complex polysaccharides necessarily requires
[10,36–38]. These rapid changes likely reflect tran- the concerted expression of multiple CAZymes [11,42]
sient proliferation of the species that are best and other proteins. However, precisely predicting the
equipped to metabolize components of new dietary phenotypic abilities of gut bacteria to degrade and
items and, upon introduction of a different meal, are utilize carbohydrates based on sequence data alone
likely to shift again. remains difficult.
Despite the presence of some generalist species A key issue is that we currently have much less
noted above, no one microorganism that can “do it all” biochemical data vis a vis sequence data—and this
has yet been identified, suggesting that a fitness gap is only widening [11,42]. Our ability to predict the
trade-off exists between the cost of encoding a diverse function of gene products is therefore perilously
array of metabolic responses and the advantage of stretched in light of the broad sequence diversity
possessing additional metabolic options. Indeed, as exhibited by families of CAZymes and carbohydrate-
the metabolic abilities of more cultivated gut microor- binding proteins. Additionally, these families are often
ganisms are systematically interrogated, it will be “polyspecific” and include members with variations in
interesting to determine which polysaccharides are glycosidic linkage specificity. Given the subtle evolu-
targeted by metabolic specialists that have more tionary changes involved in altering specificity, many
limited abilities. Such nutrients, for example, host glycosidase families are composed of members with
mucin glycoproteins, likely represent the most con- non-absolute preferences for related monosaccha-
sistent resources that fuel gut microbial growth and rides, for example, epimers (glucose versus galac-
may therefore be the most important with respect to tose) or substituent variants (galactose versus
supporting community stability. N-acetylgalactosamine). Subfamily classification has
3854 Gut microbial adaptations to carbohydrates

GH97 (exo)
(a) B. thetaiotaomicron - Starch Utilization System
GH13
BT3698 BT3705
α1,4 α1,4 α1,6

su
su (GH
su
su

su

su

su

su GH
sG
sF 13
sE
sD

sC

sB

sA 97)

sR 13
(reducing end)

(G

(
H
α1,4 α1,4 α1,4 α1,4 α1,4

)
)
GH13 (endo)
(b) B. thetaiotaomicron - Fructan Utilization Locus GH32 (endo) GH32 (exo)
BT1754 BT1765

Fr

HX

GH

GH xo)

GH
α1,2 β2,6 β2,6

uK

32

32

32
(e

(e

(e
xo
nd

)
o )
(c) B. ovatus ATCC 8384 - Xyloglucan Utilization Locus
GH2 (exo) β1,2 α1,2 GH43 (exo)
BACOVA_02644 BACOVA_02659

GH31 (exo) α1,6 α1,6 GH31 (exo)

GH

GH

GH

GH

GH
GH
GH

GH
3

31

5
43
43

3
(reducing end)

β1,4 β1,4 β1,4 β1,4

GH3 (exo) GH5/GH9 (endo)

α1,2 GH95 (exo)?

(d) B. uniformis ATCC 8492 - Xyloglucan Utilization Locus GH2 (exo) β1,2 β1,2 GH2 (exo)
BACUNI_00326 BACUNI_00315
?
GH31 (exo) α1,6 α1,6 α1,6 GH31 (exo)
GH

GH

GH
GH

GH

GH
95

31

43
3

(reducing end)

β1,4 β1,4 β1,4 β1,4

GH3 (exo) GH5 (endo)

(e) B. plebeius - Porphyran Utilization Locus

6-O-sulfate

(reducing end)
BACPLE_01667
L
β1,4 α1,3
GH
GH 6A

GH

GH

GH

GH

GH

GH16B/GH86A (endo)
1
11

50

88

16
7

3,6-anhydro
BACPLE_01706

(reducing end)
GH
GH A

Su
GH

GH

GH

L
β1,4 α1,3
l
86
86

29

16

2
B

GH16A (endo)

(f) B. intestinalis DSM17393 - Xylan Utilization Locus (6-O-ferulate)

BACINT_04208 BACINT_04220 α1,2 α1,3 α1,2

IN
GH

CE

CE 95
GH

GH

β1,4 β1,4 β1,4 β1,4 β1,4 β1,4 β1,4

T

6/G

1
8

10
H

GH8

(g) C. canimorsus - N-glycan Utilization Locus α2,6 β1,4 β1,6

Ccan_08700 Ccan_08740
α1,6

β1,4 β1,6 Asn

GH

CB

CB

α2,6
P

P
18

β1,4 β1,4 β1

α2,6 β1,4 β1,2 α1,3

GH18 (endo)

α1,2

Ser/Thr
(h) B. thetaiotaomicron - Mucin Utilization Locus α1,3 β1,4 β1,3 β1,3 α1

BT4240 BT4252 ? GH109 GH2 A blood group

GH
GH

HX

PR

GH

GH

α1,2
10

O
T
2

11

10
9

Ser/Thr
α1,3 β1,4 β1,3 β1,3 α1

GH110 GH2 B blood group

Key

Glucose N-acetyl Galactose N-acetyl Mannose

glucosamine galactosamine
N- acetyl
Xylose Arabinose Fucose Fructose
neuraminic acid

Fig. 1 PULs and their substrates. (a) The Starch Utilization System of B. thetaiotaomicron is the archetypal PUL [53,78].
(b and c) The B. thetaiotaomicron Fructan Utilization Locus [31] (b) and the B. ovatus XyGUL [30] (c) represent two other
PULs for which all of the glycoside hydrolases (GH) and several substrate-binding proteins have been biochemically
characterized. (d) A partially homologous and syntenic B. uniformis XyGUL is predicted to degrade both (arabinogalacto)
xyloglucan and (fucogalacto)xyloglucan due to an “upgrade” with a predicted fucosidase from GH95 [30]. (e and f) The
B. plebeius Porphyran Utilization Locus [40] (e) and the B. intestinalis Xylan Utilization Locus [83] (f) represent complex loci
with only partially characterized enzyme complements; for some gene products, function may be broadly predicted by
family membership, but exact specificities are currently unknown. Notably, the B. plebeius Porphyran Utilization Locus
encodes individual enzymes with specificity for either porphyran or agarose but is only induced by the former. (g and h) The
Capnocytophaga canimorsus N-glycan Degradation Locus [84] (g) and one of several B. thetaiotaomicron mucin O-linked
glycan utilization loci [46,85] (h) represent PULs in which the GH complement does not reflect the full complexity of the
substrate; complete saccharification necessarily requires the input of other loci. Gene color-coding by putative or
confirmed function is as follows: blue, GHs; purple, SusC-like; orange, SusD-like; yellow, SusE-positioned (but lacking
sequence similarity); salmon, sensor/transcriptional regulator; red, two-component systems associated with mobile
elements; cyan, non-catalytic carbohydrate binding proteins; green, other proteins with characterized homologs; gray,
unknown function. Abbreviations: glycoside hydrolase (GH), carbohydrate-binding protein (CBP), carbohydrate esterase
(CE), fructokinase (FruK), hexose transporter (HXT), integrase (INT), mucin protease (PRO) and sulfatase (Sul). Sugars
are represented by Consortium for Functional Glycomics standard symbols.
Gut microbial adaptations to carbohydrates 3855

been employed to further refine specificity prediction much of the apparent redundancy in genes encoding
for a handful of families, but even this is not absolute some enzyme families that are heavily represented
[43–45]. Thus, an individual gut enzyme may be in the genomes of symbiotic gut bacteria. When over-
hypothetically associated with degradation of multiple expressed and analyzed in vitro (away from their
polysaccharides, leading to ambiguous functional normal regulatory control and associated enzymatic
predictions. partners), such enzymes may display overlapping
Since many bacteria tightly control the expression activities on a range of substrates that harbor a
of enzyme systems that target various nutrients, particular target linkage [51]. However, in the context
whole genome transcriptional analyses of bacteria of a living cell, deployment of these overlapping
grown in pure carbohydrates can be very useful to enzymes in response to specific substrates becomes
guide identification of gene products that work much more specialized.
together [10,28,29,46,47]. Such functions are often Generalized polysaccharide degradation schemes
grouped together—especially in Bacteroidetes—into for Gram-negative and Gram-positive gut bacteria
genomically linked PULs (polysaccharide utilization are illustrated in Fig. 2, based on knowledge of
loci; Fig. 1) [48]. Empirical delineation of these gene several different catabolic systems in Bacteroidetes
clusters, together with reverse genetic and biochem- and Firmicutes/Actinobacteria, respectively (see also
ical studies, provides a powerful route to the discovery recent reviews [52,53]). Aside from the need for
of new metabolic pathways, especially those for sensory specificity, an additional prerequisite is
complex carbohydrate utilization. Given the ever- access to the target substrate in order to initiate
expanding breadth of sequence-based data, such enzymatic attack and release cues that activate a
comprehensive findings can be particularly useful corresponding response. For some bacteria, substrate
because they can be rapidly fit into a much larger solubility may create a substantial barrier to access.
biological/ecological picture using comparative ge- Both B. thetaiotaomicron and Eubacterium rectale can
nomics and metagenomics. grow on soluble forms of starch but not on insoluble
“resistant starch” [15,16,25]. In contrast, one Firmicute,
Ruminococcus bromii, has been shown to degrade
Gut Symbionts Have Specialized resistant starch in vitro, revealing that it possesses an
Sensory and Enzymatic Toolkits alternative mechanism of substrate access relative to
for Polysaccharide Digestion other bacteria [25]. At least two other Ruminococci,
Ruminococcus champanellensis and Ruminococcus
Substrate specificity is a key facet of microbial torques, possess the ability to utilize insoluble cellu-
responses to complex carbohydrates. It might be lose and carbohydrate components of human colonic
expected that an organism such as B. cellulosilyticus, mucin, respectively [54,55]. These species stand in
which harbors 424 different carbohydrate-degrading contrast to Bacteroides that have the ability to use
enzymes, would carefully regulate expression of these soluble portions of these same substrates, such as
enzymes and their associated non-enzymatic factors cellobiose and chemically released mucin O-linked
or risk needlessly expending energy for superfluous glycans [29,55]. Future elucidation of the molecular
protein synthesis in the competitive gut environment. mechanisms that mediate these phenotypes will
Indeed, this is the case in B. cellulosilyticus and other undoubtedly provide novel insight into the first crucial
gut and rumen Bacteroidetes that have been studied step in insoluble fiber digestion.
[29,31,49]. In these organisms, expression of gene When accessible polysaccharides are encountered,
clusters ranging in size from 6 to several dozen open initial attack presumably proceeds via constitutively
reading frames is regulated via inducible promoters produced enzymes that are present at low “surveil-
that are usually activated by transcription factors lance” levels [46]. Indeed, without such enzymes, the
present within or adjacent to the responsive gene bacterium may be very well incapable of generating
cluster. A common theme that has emerged in the the required cues to amplify a catabolic response in
Gram-negative Bacteroidetes is that the sensor the first place. This phenomenon of oligosaccharide
proteins extend outside of the cytoplasm to interact substrate-induced up-regulation is readily observed
with enzymatically released saccharides either in in B. thetaiotaomicron. In the presence of glucose, a
the periplasm or during transport across the outer monosaccharide substrate that does not trigger acti-
membrane [29,31,48,50]. By sensing carbohydrate vation of the Starch Utilization System (Fig. 1a), a
cues before degradation to individual monosaccha- punctate pattern of the key SusD protein can be ob-
rides, these organisms are able to integrate addi- served at the cell surface (inset micrographs in Fig. 2a).
tional information about the polysaccharide that has In cells exposed to starch for just 30 min, expression
been encountered, including sugar sequence, linkage of these proteins is dramatically up-regulated due to
stereochemistry and linkage regiochemistry. This the production of malto-oligosaccharides that bind the
strategy allows bacteria to couple a particular sensor sensor/regulator SusR [34]. Thus, in the model shown
to the expression of a more specific set of polysac- in Fig. 2., each polysaccharide substrate can initiate
charide-degrading enzymes and may account for a positive feedback loop in which small amounts of
3856 Gut microbial adaptations to carbohydrates

(a)
Carbohydrate surveillance Active degradation

Soluble/accessible
polysaccharides

binding
Binding & binding GH/PL GH/PL binding
GH/PL GH/PL
hydrolysis
binding GH/PL

binding

Transport

outer
GH/PL GH/PL
Hydrolysis to GH/PL
GH/PL membrane
GH/PL
simple sugars
GH/PL
and transport

Recognition

inner
Transcript activation & membrane
increased expression

α-B. theta SusD (glucose) α-B. theta SusD (starch)

(b)

GH/PL

binding
GH/PL (SBP) binding
(SBP)

ATP ATP cell

ADP ADP membrane/
cell wall
GH/PL

Fig. 2 (legend on next page )

Gut microbial adaptations to carbohydrates 3857

surveillance enzymes produce substrate-specific cues of β(1-3) and β(1-4) linkages in mixed-linkage glucans
to amplify expression of more enzymes until either the has a significant effect on solubility [62]. The effect
bacterial surface is saturated or the substrate wanes in of substituents on polysaccharide physical and
abundance. chemical properties is likewise considerable: amylo-
An additional facet of microbial competition for pectin has significantly greater water solubility than
available nutrients, which has remained largely amylose due to numerous α(1-6) branch points
unexplored for most symbiotic gut bacteria beyond extended with α(1-4)-glucan chains [59,63]. Like-
Escherichia coli, is the propensity of many bacteria to wise, the xyloglucan family of plant cell wall matrix
prioritize some nutrients with respect to others. This glycans can be considered as soluble cellulose
idea is predicated on the classic example of repres- derivatives, in which the common β(1-4)-glucan
sion of E. coli lactose utilization, discovered by Jacob backbone is highly substituted with diverse monosac-
and Monod [56]. A recent study on B. thetaiotaomicron charides extended from α(1-6)-xylosyl branches
revealed that this generalist engages in metabolic (Fig. 1) [64,65]. Numerous other such examples of
prioritization of the more than 12 different polysac- branched polysaccharides (with an equally diverse
charides that it is equipped to consume [34]. Since range of solubilities) are well known [57,66,67]. An
this species is just one of many hundred that are additional level of structural complexity, which impacts
common to the human gut, an important frontier in physical–chemical properties of many polysaccha-
understanding the behavior of these organisms will rides, is the presence of ether (e.g., methyl) and ester
be to understand the programming of different nutrient (e.g., acetyl, ferulate, sulfate) substituents [39,67,68].
prioritization schemes. Indeed, many organisms Indeed, the large repertoire of monosaccharide
may exhibit similar or identical metabolic potential building blocks and the manifold possible stereo-
in vitro on pure polysaccharide substrates but deploy chemical and regiochemical linkage variations com-
these traits in very different ways in the complex and bine to generate a tremendous diversity of saccharide
dynamic nutrient mixtures that are likely to be the norm structures in terrestrial and marine biomass relevant
in the gut. to the microbiota; this includes polysaccharides,
proteoglycans/ glycoproteins and glycolipids [57,69–
71]. A key corollary is that a similarly large cohort of
Closely Related Polysaccharides Harbor saccharide-specific CAZymes, binding proteins,
Variations That Impact Their Availability transporters and sensor regulators is required by gut
to Gut Microbes microbial communities to address this diversity.
Functional genomics and enzymological studies of
From the discussion mentioned above, it follows the microbiota have necessarily been guided—and
that detailed knowledge of polysaccharide structure likely limited—by our current knowledge of the
is central to understanding substrate prioritization structural diversity of complex carbohydrates in
and the potential catabolic niches that are available Nature (as well as the general availability of these
to the microbiota. In particular, variations in polysac- substrates for experimentation) [12,19–22,29,46,72].
charide linkage and side-chain branching affect Given the intrinsic technical difficulties in determining
physical properties such as solubility and solution oligosaccharide and polysaccharide sequences, vis a
rheology [57], which are important in distinguishing vis polypeptides or polynucleotides, accurate struc-
“soluble” and “insoluble” forms of dietary “fiber” [58]. tural data are challenging to obtain and generally
For example, it is readily appreciated that cellulose limited to abundant, readily extracted representatives
from plant cell walls [β(1-4)-glucan], amylose from of a given polysaccharide family. Polysaccharide
starch [α(1-4)-glucan] and mixed-linkage β(1-3)/ biosynthesis is often not tightly regulated, with the
β(1-4)-glucans from cereal endosperm have widely result that variations in composition and degree of
differing solubilities and susceptibilities to biological polymerization arise across individual tissues and
degradation due to differences in secondary structure; growth conditions [62,66]. This is as true for plant cell
however, all are “simple” homopolymers of glucose wall polysaccharides as it is for animal glycoproteins,
[59–61]. Even the relative abundance and distribution such that all “xylan” and all “mucin” are not created

Fig. 2. Generic models for polysaccharide acquisition by gut bacteria. (a) Gram-negative model based on Bacteroidetes
Sus-like systems showing the transition from a surveillance state to active degradation. Monosaccharides are illustrated,
alone or in chains, as red circles. Key sequential steps, beginning with the first encounter with soluble or accessible
polysaccharides and ending with increased expression of degradative and assimilatory machinery, are noted in blue.
Insets at the bottom of (a) shows B. thetaiotaomicron cells stained with an antibody against SusD, involved in starch
degradation (see the text for details). (b) Gram-positive model based on ABC transporter-based systems present
in Firmicutes and Actinobacteria. Opening and closing of the cytoplasmic-membrane spanning permease allows
monosaccharide and oligosaccharide substrates that are bound by solute-binding proteins (SBPs) to be transported
as ATP is hydrolyzed to ADP and inorganic phosphate. Abbreviations: glycoside hydrolase (GH) and polysaccharide
lyase (PL).
3858 Gut microbial adaptations to carbohydrates

equal (to name just two examples) [62,73]. This starch-binding proteins. In addition to the genetics
heterogeneity poses a significant challenge to the and biochemistry of this system, the structural biology
fractionation of pure compounds for structural deter- of all these starch-binding proteins and enzymes,
mination, with the result that monosaccharide and except for the GH13 α(1-6)-glucosidase, has been
linkage compositions most often represent average elucidated through a detailed series of studies
values. As such, particular substructures may escape [48,78–82]. The selected examples shown in Fig. 1
detection, either due to technical limitations or due to further highlight the correlation between polysaccha-
source-sampling bias. Fortunately for microbiota ride and PUL complexity for a number of systems in
research, insight into complex carbohydrate structure which functional predictions have been supported by
continues to increase, buoyed by a rising global rigorous reverse genetic and biochemical analyses
biotechnological interest in functional foods/prebiotics [30,31,40,83,84].
and biomass saccharification for fuel and chemical An overarching problem in generating robust
production. functional predictions for novel PULs is our general
Work on the xyloglucan family provides one example inability to precisely map putative specificities of
in which both strictly analytical and combined genetic/ enzyme cohorts (e.g., deduced from CAZy family
analytical approaches continue to uncover novel membership) one-to-one with known polysaccharide
polysaccharide motifs. Indeed, the structural variation substructural motifs. Both the B. plebeius Porphyran
in this family across plant species has been probably Utilization Locus [40] and the Bacteroides intestinalis
the most widely and precisely explored of all the matrix Xylan Utilization Locus [83] are complex loci, which
polysaccharides [64,65,74]. Nonetheless, recent de- encode gene products whose correlation to polysac-
tailed analysis of xyloglucans from the Ericaceae, charide structure is not immediately obvious despite
notably including the edible bilberry fruit (Vaccinium partial characterization of their enzyme comple-
myrtillus L.), revealed an unusual di-xylosylated motif ments. Observations of this type highlight gaps in
[β- D -Xylp-(1-2)-α- D -Xylp-(1-6)-β- D -glucan] as a our current knowledge of both enzyme function and
dominant structural component [75,76]. Likewise polysaccharide structural diversity. These systems
surprising, previously unknown galacturonic acid provide opportunities to leverage information from
(GalpA)-containing branches [β-D-GalpA-(1-2)-α-D- bacterial genomic sequences (i.e., groups of co-
Xylp-(1-6)-β-D-glucan and α-L-Fucp-(1-2)-β-D-Gal- regulated enzymes with partially unknown targets)
pA-(1-2)-α-D-Xylp-(1-6)-β-D-glucan] have been re- to inform detailed studies of new polysaccharide
vealed recently in root tissues of the model dicot substructures. In contrast to these complex loci,
Arabidopsis thaliana. Moreover, the glycosyltrans- other PULs are less extensive and clearly lack the
ferase family 47 enzyme responsible for adding complete enzyme cohort necessary to full degrade
the GalpA residue has been identified [77], which their cognate substrates (Fig. 1g and h) [84,85],
may in turn facilitate the possible discovery of this suggesting that they may work together with other
type of xyloglucan in other plant species, including systems during catabolism.
food crops. Although such structural differences Recent studies have demonstrated how a “change
may seem subtle, on one hand, they may impart in scale” of PUL characterization to encompass
significant differences in solubility and accessibility functional characterization of most, if not all, of the
by the microbiota to a given polysaccharide, as gene products can bring a more holistic understanding
outlined above. On the other hand, each novel of PUL function in the context of organism and
glycosidic linkage in a polysaccharide or family of ecosystem [86]. One of the first examples of this
polysaccharides will also further complicate attack approach elegantly revealed the molecular basis of
by a given microbiota species or consortium due to levan [β(2-6)-fructan] specificity of B. thetaiotaomicron,
the requirement for an additional glycosidase (and which otherwise grows poorly on the isomeric inulin
possibly substitution-tolerant sensor/regulators, [β(2-1)-fructan] polysaccharide. Here, functional char-
transporters and substrate-binding proteins) in the acterization of the GH32 endo-levanase and the
microbiome. SusD-like carbohydrate-binding protein encoded by
The concept that enzyme cohorts must increase in a Fructan Utilization Locus (Fig. 1b), both of which are
complexity in step with polysaccharide fine-structure located on the cell surface, revealed a strict specificity
is exemplified by a number of recent biochemical for levan polysaccharide. In striking contrast, neither
studies of specific PULs. As a point of reference, the hybrid two-component system sensor regulator
the archetypal PUL, the Starch Utilization System of nor the three GH32 exo-fructosidases (including one
B. thetaiotaomicron, encodes three glycoside hydro- encoded outside of this PUL) discriminated against
lases sufficient to cleave all the linkages in starch: either linkage [31].
a GH13 endo-amylase [endo-α(1-4)-glucanase], a A similarly extensive study from our own collabo-
GH13 α(1-6)-glucosidase and a GH97 exo-α(1-4)- rative work recently revealed the functions of all eight
glucosidase (Fig. 1a). This PUL also encodes one glycoside hydrolases and two carbohydrate-binding
transcriptional regulator, one TonB-dependent proteins encoded by a unique Xyloglucan Utilization
receptor/transporter and three outer-membrane Locus (XyGUL) in B. ovatus (Fig. 1c) [30]. The
Gut microbial adaptations to carbohydrates 3859

ensemble of enzymes encoded by this locus is degrade the red and brown algal polysaccharides
consistent with a particular specificity for solana- agarose and alginate have revealed, using compara-
ceous (arabinogalacto)xyloglucan (as found in to- tive genomics, that enzymes for degrading these
matoes, peppers, eggplants, olives, etc.) [64,87–89] polysaccharides are present in the genomes of some
due to the presence of two α-L-arabinofuranosidases gut Bacteroides. While putative mechanisms for these
from GH43. Notably, the B. ovatus XyGUL lacks a transfer events remain unclear, these observations
fucosidase, as would be required to fully degrade lend further support for the idea that genetic material
the (fucogalacto)xyloglucan widespread among can be transferred, perhaps along with foods
other dicot plants (e.g., leafy vegetables). In contrast, containing the target polysaccharides, into gut bacteria
partially homologous XyGULs in other prevalent to enable catabolism of novel nutrients. Interestingly,
human and termite gut Bacteroidetes encode both a cTn that is related to the one implicated in transfer
a GH43 member, as well as predicted fucosidase of porphyran utilization into the B. plebeius genome
from GH95 or GH29, which indicates the evolution of is also present in the sequenced type strain of B.
these XyGULs to address both types of xyloglucans thetaiotaomicron but carries genes that are expressed
from dietary plants (Fig. 1d). Strikingly, analysis of during degradation of Saccharomyces cerevisiae cell
metagenomic data revealed that at least 92% of the wall α-mannan instead of porphyran [40,46].
humans sampled (n = 250) across North America, Taken together, the observations of porphyran and
Europe and Japan carry at least one or more α-mannan-degrading abilities associated with mo-
homologous XyGULs, underscoring the particular bile elements suggest that dietary polysaccharides
importance of xyloglucan metabolism in our diet that are novel to regional human diets or increase
[30]. historically as a result of new dietary customs can
Together, the levan and xyloglucan PUL studies elicit adaptive genetic changes in members of the gut
highlight the comprehensive insight that a combi- microbiota. To explore this point further in the context
nation of biochemical, enzymological, genetic and of this review, we searched fully sequenced or draft
structural analyses can bring. We anticipate that genomes for conjugative elements related to the two
similar multidisciplinary approaches will bring sig- that are present in the genomes of B. plebeius and
nificant new insight into PUL functional diversity in B. thetaiotaomicron, which revealed that numerous
future studies. Further, as the polysaccharides additional site-specific cTns exist and that some of
studied become more complex or variable (arabi- these newly identified elements harbor genes likely
noxylans, pectins, mucin glycoproteins, etc.), it is to be involved in bacterial polysaccharide degrada-
likely that similar or greater levels of fine-structural tion (Fig. 3). Inspection of a subset of cTns related
adaptations by the cognate bacterial enzyme to that conferring porphyran degradation further
systems will be revealed. revealed that all of the related elements are found
in gut isolates and are separate from sequences
found in environmental isolates (compare blue
How Are Gut Bacterial branch highlighted in light green with dark green
Genomes Upgraded to Target branches in Fig. 3a). This latter observation sug-
New Polysaccharides? gests that transfer of these genes, which are
ordinarily found in phylogenetically more distant
Moving beyond the effective deployment of existing marine bacteria [39], may have been more complex
enzyme and protein complements, the evolution than simple transfer of a single cTn from a marine
or acquisition of genes enabling the degradation organism to one residing in the gut. Interestingly,
of alternate polysaccharides can allow access to another common “genetic cargo” that we found
new or unused nutrient niches. In some cases, the located in cTns are gene clusters predicted to be
mechanism of gene acquisition is apparent, while involved in capsular polysaccharide synthesis,
in others, this evolution is less clear. For example, suggesting that both catabolic and anabolic func-
delineation of the genes required for porphyran tions related to polysaccharides are being mobilized
degradation by B. plebeius and analysis of the by this mechanism.
surrounding genomic region revealed that a mobile Notably, our recent investigation of xyloglucan
element belonging to the conjugative transposon (cTn) utilization in gut Bacteroidetes [30] revealed marked
family, a family of mobile elements long associated heterogeneity in carriage of this PUL among closely
with transfer of antibiotic resistance among gut related strains of B. ovatus and Bacteroides xylani-
bacteria [90], likely mediated transfer of this ability solvens. In this case, appearance of the gene cluster
into B. plebeius (Fig. 1e) [40]. Moreover, similar genes required for xyloglucan utilization was not associated
involved in porphyran degradation are present in the with any discernable mobile element that might have
fecal metagenomes of Japanese subjects but are less been responsible for its transfer. Rather, it appeared
abundant or absent in the microbiota of westerners to be precisely inserted, with only a few hundred
[39]. Additional studies that have initially focused on base pairs or less of flanking sequence, into a single
the enzymatic abilities of marine Bacteroidetes to homologous region of the B. ovatus genome that
3860 Gut microbial adaptations to carbohydrates

also contained additional PULs in some isolates. different in several cases, revealing that enzyme
The order and specificity of enzymes contained in encoding genes are likely being duplicated, re-
other B. ovatus xyloglucan utilization PULs were arranged and probably diversified via mutations

(a)
Legend (panel B):
Bacteroidetes conjugative
transposons (cTns): Cargo genes:
gut bacteria polysaccharide utilization loci
capsular polysaccharide synthesis
oral bacteria
antibiotic / heavy metal resistance
environmental
Core cTn functions:
bacteria
mobilization (tra) genes
genitourinary tract integrase
bacteria rteA/B two component regulators
other bacteria unknown cargo
topoisomerase
excisionase
other genes

(b) mobilization cargo

B. plebeius
90

65

*
*
* tRNAlys insertion sequence:
* ccaggcggatcac
*
*
*
*
*
*
76 82 84 93 00

B. thetaiotaomicron
50

*
*

*
*
*
*
*
*
* tRNAphe insertion sequence:
gttcgattcctggtggcaccac
*
*
*

Fig. 3 (legend on next page )

Gut microbial adaptations to carbohydrates 3861

within closely related gene clusters (Fig. 1c and d). evolved by the human gut microbiota in parallel with
Such phenomena remain to be studied in detail. dietary habits.
However, the presence of new and re-arranged
genes within closely related PULs suggests that,
once a gene cluster is acquired, it can be subse- Prospectus: How Fine Is the Level of
quently subjected to additional, fine-level tailoring to
modify or refine its function. Polysaccharide Niche Adaptation by
In the future, it will be interesting to determine if the Gut Microbes?
polysaccharide utilization abilities associated with
other Bacteroidetes cTns and other divergent PULs The interactions between our symbiotic gut
enable metabolism of polysaccharides that have microorganisms and dietary polysaccharides play
been introduced into some human diets on relatively important roles in several aspects of health
recent timescales or those that are fine-level variants [52,53,94]. Polysaccharides represent the major
of more commonly consumed plant and animal class of nutrients that escape digestion in the
polysaccharides. Some of these items exist uniquely proximal intestine, illuminating dietary fiber or
in other regional foods, such as seeds from Salvia purified polysaccharide prebiotics as promising
hispanica (chia), which have been cultivated and non-invasive avenues to intentionally manipulate
consumed in Mexico and Guatemala since pre- the gut community. In order to achieve a highly
Columbian times and contain a unique polymer sophisticated ability to employ such manipulations,
composed of xylose, glucose and 4-O-methyl glucu- we must first understand the precise connections
ronic acid [91]. Other novel polysaccharides such as between the many different dietary and endogenous
xanthan gum (approved for human consumption mucosal carbohydrates and the microorganisms
in 1968) are much more recent introductions to the that directly degrade them.
human food supply and are often used as food Beginning in the 1950s, Freter and colleagues
additives or thickeners [92]. In the context of dietary began formulating the nutrient-niche theory of gas-
red seaweed galactan utilization, it is interesting to trointestinal colonization by microorganisms [95–97],
note that a screen of nearly 300 common gut which postulates that a diverse, but finite, number
Bacteroides strains revealed only two, Bacteroides of niches are present in the gut, each defined by
uniformis NP1 and B. thetaiotaomicron VPI-3731, availability of a subset of particular nutrients. When
that were able to grow on agarose and kappa-carra- these niches are filled by endogenous symbionts,
geenan, respectively [40]. These capacities stand in invasion of the intestine by outside species, including
contrast to the strict porphyran specificity of B. plebeius pathogens, is more difficult and leads to the phenom-
[40], which implies that each strain harbors a unique enon of colonization resistance. When nutrient niches
ensemble of sensory proteins and enzymes to address are opened, either due to large-scale loss of microbial
these distinct, complex polysaccharides (see Ref. [57] colonization during antibiotic use or perhaps due
for a structural summary). Indeed, several families of to smaller perturbations such as diet shift, the gut
enzymes specific for agars (comprising agarose and environment may become transiently susceptible
porphyrans) and the diverse types of carrageenans to invasion by pathogens, such as Clostridium difficile,
have already been characterized from marine bacteria E. coli, Salmonella enterica or Vibrio cholera. While
[93]. It will be interesting to ascertain to what extent most of these pathogens may not be metabolically
these enzymes may have been acquired and further equipped to directly attack polysaccharides (a notable

Fig. 3. Sampling of Bacteroidetes conjugative transposon (cTn) diversity in sequenced genomes. (a) A cladogram
based on alignment of 501 Bacteroidetes amino acid sequences for TraJ, a conserved protein involved in bacterial
conjugation. Terminal branches are color-coded based on the habitat from which the sequenced bacterial isolate
was taken. Two branches in the lower left quadrant that contain tRNA lys and tRNA phe targeting cTns related to the
characterized elements, which confer porphyran and fungal mannan-degrading abilities to B. plebeius and
B. thetaiotaomicron, are highlighted in light green and pink, respectively. Survey was from the publicly available
sequence database up to May 2012 and the cladogram was constructed using Clustal W [103]. (b) A more detailed view of
the cTn sequences and their putative functions associated with tRNA lys and tRNA phe insertion sites. The precise insertion
sequences corresponding to the 3′ end of each tRNA are listed; examples where a direct copy of this repeat could be
located at the adjacent (right) flank of the element are indicated with an asterisk. Variable cargo gene regions are
connected between different elements with gray shading to highlight the conserved location of cargo but the different size
and functions of transferred genes. The previously characterized elements from B. plebeius and B. thetaiotaomicron are
highlighted with red boxes. Five other elements contain putative PULs and are indicated by blue cargo genes. Breaks in
incomplete elements (a result of gaps in draft genome assemblies) are indicated as black vertical line.
3862 Gut microbial adaptations to carbohydrates

exception is Yersinia enterocolitica, which has been Received 13 April 2014;

shown to target plant cell wall pectin [98,99]), it is likely Received in revised form 13 June 2014;
that they compete for other limiting nutrients that are Accepted 29 June 2014
used or released by polysaccharide-degrading organ- Available online 12 July 2014
isms. Thus, the presence of endogenous bacteria
filling most available niches reinforces the ability of the Keywords:
community to exclude pathogens. Indeed, this is a microbiome;
likely mechanism behind the remarkable success microbiota;
of fecal transplant therapy and “rePOOPulation” polysaccharide;
to combat chronic C. difficile infection [94,100,101]. Bacteroides;
In light of the points discussed above, connecting lateral gene transfer
specific gut symbionts with certain fiber polysac-
charides, understanding the specific details of
nutrient–microbe relationships in the gut and deter-
mining downstream interactions between microbes
may represent a pinnacle step in comprehending References
Freter's theory. Such knowledge may reveal how
many niches actually exist to be both filled by
[1] El Kaoutari A, Armougom F, Gordon JI, Raoult D, Henrissat
endogenous microbes and subsequently competed
B. The abundance and variety of carbohydrate-active
for when the gut microbial ecosystem is perturbed enzymes in the human gut microbiota. Nat Rev Microbiol
and potential pathogens have an opportunity to gain 2013;11:497–504.
foothold. [2] McNeil NI. The contribution of the large intestine to energy
Our central hypothesis is that many polysaccha- supplies in man. Am J Clin Nutr 1984;39:338–42.
ride–gut microbe relationships have evolved to be [3] Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlefsen
highly specific, involving sensory and enzymatic L, Sargent M, et al. Diversity of the human intestinal
adaptations by particular microbes that best equip microbial flora. Science 2005;308:1635–8.
certain species or strains to utilize fine-level varia- [4] Ley RE, Hamady M, Lozupone C, Turnbaugh PJ, Ramey
tions that exist in polysaccharides. Support for this RR, Bircher JS, et al. Evolution of mammals and their gut
microbes. Science 2008;320:1647–51.
idea has been born out in the literature with a few
[5] Qin J, Li R, Raes J, Arumugam M, Burgdorf KS,
rigorously investigated examples [30,31] but more Manichanh C, et al. A human gut microbial gene catalogue
are needed. As the catalog of common gut micro- established by metagenomic sequencing. Nature 2010;464:
organisms expands along with knowledge of these 59–65.
species' genetic potential, defining precise relation- [6] Backhed F, Ley RE, Sonnenburg JL, Peterson DA, Gordon
ships between individual organisms and exogenous JI. Host-bacterial mutualism in the human intestine. Science
forces such as dietary polysaccharides will be an 2005;307:1915–20.
essential part of understanding the behavior of this [7] Gill SR, Pop M, Deboy RT, Eckburg PB, Turnbaugh PJ,
complex ecosystem. Such knowledge will be re- Samuel BS, et al. Metagenomic analysis of the human distal
quired to design strategies to manipulate the impact gut microbiome. Science 2006;312:1355–9.
[8] Xu J, Bjursell MK, Himrod J, Deng S, Carmichael LK,
of gut microbes on aspects of our health ranging
Chiang HC, et al. A genomic view of the human-Bacteroides
from nutrition to mitigation of diseases such thetaiotaomicron symbiosis. Science 2003;299:2074–6.
as inflammatory bowel disease, colon cancer, [9] Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL,
obesity and diabetes [102]. Duncan A, Ley RE, et al. A core gut microbiome in obese
and lean twins. Nature 2009;457:480–4.
[10] McNulty NP, Wu M, Erickson AR, Pan C, Erickson BK,
Martens EC, et al. Effects of diet on resource utilization
by a model human gut microbiota containing Bacteroides
cellulosilyticus WH2, a symbiont with an extensive glyco-
Acknowledgements biome. PLoS Biol 2013;11:e1001637.
[11] Lombard V, Ramulu HG, Drula E, Coutinho PM, Henrissat
H.B. is supported by faculty funding from the B. The carbohydrate-active enzymes database (CAZy) in
Michael Smith Laboratories of the University of 2013. Nucleic Acids Res 2014;42:D490–5.
British Columbia, the Natural Sciences and Engi- [12] Chung KT, Bryant MP. Robert E. Hungate: pioneer of
neering Research Council of Canada (Discovery anaerobic microbial ecology. Anaerobe 1997;3:213–7.
[13] Arank A, Syed SA, Kenney EB, Freter R. Isolation of anaerobic
Grant), the Canada Foundation for Innovation and
bacteria from human gingiva and mouse cecum by means of a
the British Columbia Knowledge Development Fund. simplified glove box procedure. Appl Microbiol 1969;17:568–76.
E.C.M. was supported by National Institutes of [14] Moore WE, Holdeman LV. Human fecal flora: the normal flora
Health grants DK084214 and GM099513 and Global of 20 Japanese-Hawaiians. Appl Microbiol 1974;27:961–79.
Probiotics Council Young Investigator Grant for [15] Salyers AA, Vercellotti JR, West SE, Wilkins TD. Fermen-
Probiotics Research. tation of mucin and plant polysaccharides by strains of
Gut microbial adaptations to carbohydrates 3863

Bacteroides from the human colon. Appl Environ Microbiol marine isopod in the absence of gut microbes. Proc Natl
1977;33:319–22. Acad Sci U S A 2010;107:5345–50.
[16] Salyers AA, West SE, Vercellotti JR, Wilkins TD. Fermen- [33] Larsson AM, Anderson L, Xu B, Munoz IG, Uson I, Janson
tation of mucins and plant polysaccharides by anaerobic JC, et al. Three-dimensional crystal structure and enzymic
bacteria from the human colon. Appl Environ Microbiol characterization of beta-mannanase Man5A from blue
1977;34:529–33. mussel Mytilus edulis. J Mol Biol 2006;357:1500–10.
[17] Goodman A, Kallstrom G, Faith JJ, Reyes A, Moore A, [34] Rogers TE, Pudlo NA, Koropatkin NM, Bell JS, Moya Balasch
Dantas G, et al. Extensive personal human gut microbiota M, Jasker K, et al. Dynamic responses of Bacteroides
culture collections characterized and manipulated in gnoto- thetaiotaomicron during growth on glycan mixtures. Mol
biotic mice. Proc Natl Acad Sci U S A 2011;108:6252–7. Microbiol 2013;88:876–90.
[18] Walker AW, Duncan SH, Louis P, Flint HJ. Phylogeny, [35] Sonnenburg JL, Xu J, Leip DD, Chen CH, Westover BP,
culturing, and metagenomics of the human gut microbiota. Weatherford J, et al. Glycan foraging in vivo by an
Trends Microbiol 2014;22:267–74. intestine-adapted bacterial symbiont. Science 2005;307:
[19] Hartemink R, VanLaere KMJ, Mertens AKC, Rombouts FM. 1955–9.
Fermentation of xyloglucan by intestinal bacteria. Anaerobe [36] David LA, Maurice CF, Carmody RN, Gootenberg DB,
1996;2:223–30. Button JE, Wolfe BE, et al. Diet rapidly and reproducibly
[20] Jonathan MC, van den Borne J, van Wiechen P, da Silva alters the human gut microbiome. Nature 2014;505:
CS, Schols HA, Gruppen H. In vitro fermentation of 12 559–63.
dietary fibres by faecal inoculum from pigs and humans. [37] Turnbaugh PJ, Ridaura VK, Faith JJ, Rey FE, Knight R,
Food Chem 2012;133:889–97. Gordon JI. The effect of diet on the human gut microbiome:
[21] Kabel MA, Kortenoeven L, Schols HA, Voragen AGJ. In vitro a metagenomic analysis in humanized gnotobiotic mice. Sci
fermentability of differently substituted xylo-oligosaccharides. Transl Med 2009;1:6ra14.
J Agric Food Chem 2002;50:6205–10. [38] Caporaso JG, Lauber CL, Costello EK, Berg-Lyons D,
[22] Van Laere KMJ, Hartemink R, Bosveld M, Schols HA, Gonzalez A, Stombaugh J, et al. Moving pictures of the
Voragen AGJ. Fermentation of plant cell wall derived human microbiome. Genome Biol 2011;12:R50.
polysaccharides and their corresponding oligosaccharides [39] Hehemann JH, Correc G, Barbeyron T, Helbert W, Czjzek
by intestinal bacteria. J Agric Food Chem 2000;48:1644–52. M, Michel G. Transfer of carbohydrate-active enzymes from
[23] Crittenden R, Karppinen S, Ojanen S, Tenkanen M, marine bacteria to Japanese gut microbiota. Nature
Fagerstrom R, Matto J, et al. In vitro fermentation of cereal 2010;464:908–12.
dietary fibre carbohydrates by probiotic and intestinal [40] Hehemann JH, Kelly AG, Pudlo NA, Martens EC, Boraston
bacteria. J Sci Food Agric 2002;82:781–9. AB. Bacteria of the human gut microbiome catabolize red
[24] Leitch EC, Walker AW, Duncan SH, Holtrop G, Flint HJ. seaweed glycans with carbohydrate-active enzyme up-
Selective colonization of insoluble substrates by human dates from extrinsic microbes. Proc Natl Acad Sci USA
faecal bacteria. Environ Microbiol 2007;9:667–79. 2012;109:19786–91.
[25] Ze X, Duncan SH, Louis P, Flint HJ. Ruminococcus bromii [41] Nelson KE, Weinstock GM, Highlander SK, Worley KC, Creasy
is a keystone species for the degradation of resistant starch HH, Wortman JR, et al. A catalog of reference genomes from
in the human colon. ISME J 2012;6:1535–43. the human microbiome. Science 2010;328:994–9.
[26] Shipman JA, Berleman JE, Salyers AA. Characterization of [42] Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard
four outer membrane proteins involved in binding starch to V, Henrissat B. The Carbohydrate-Active EnZymes data-
the cell surface of Bacteroides thetaiotaomicron. J Bacteriol base (CAZy): an expert resource for Glycogenomics.
2000;182:5365–72. Nucleic Acids Res 2009;37:D233–8.
[27] Garrido D, Kim JH, German JB, Raybould HE, Mills DA. [43] Aspeborg H, Coutinho PM, Wang Y, Brumer H, Henrissat B.
Oligosaccharide binding proteins from Bifidobacterium Evolution, substrate specificity and subfamily classification
longum subsp. infantis reveal a preference for host glycans. of glycoside hydrolase family 5 (GH5). BMC Evol Biol
PLoS One 2011;6:e17315. 2012;12:186.
[28] Scott KP, Martin JC, Chassard C, Clerget M, Potrykus J, [44] Lombard V, Bernard T, Rancurel C, Brumer H, Coutinho PM,
Campbell G, et al. Substrate-driven gene expression in Henrissat B. A hierarchical classification of polysaccharide
Roseburia inulinivorans: importance of inducible enzymes lyases for glycogenomics. Biochem J 2010;432:437–44.
in the utilization of inulin and starch. Proc Natl Acad Sci U S A [45] Stam MR, Danchin EGJ, Rancurel C, Coutinho PM,
2011;108:4672–9. Henrissat B. Dividing the large glycoside hydrolase family
[29] Martens EC, Lowe EC, Chiang H, Pudlo NA, Wu M, McNulty 13 into subfamilies: towards improved functional annota-
NP, et al. Recognition and degradation of plant cell wall tions of alpha-amylase-related proteins. Protein Eng Des
polysaccharides by two human gut symbionts. PLoS Biol Sel 2006;19:555–62.
2011;9. http://dx.doi.org/10.1371/journal.pbio.1001221. [46] Martens EC, Chiang HC, Gordon JI. Mucosal glycan
[30] Larsbrink J, Rogers TE, Hemsworth GR, McKee LS, Tauzin foraging enhances fitness and transmission of a sacchar-
AS, Spadiut O, et al. A discrete genetic locus confers olytic human gut bacterial symbiont. Cell Host Microbe
xyloglucan metabolism in select human gut Bacteroidetes. 2008;4:447–57.
Nature 2014;506:498–502. [47] Rey FE, Faith JJ, Bain J, Muehlbauer MJ, Stevens RD,
[31] Sonnenburg ED, Zheng H, Joglekar P, Higginbottom SK, Newgard CB, et al. Dissecting the in vivo metabolic potential
Firbank SJ, Bolam DN, et al. Specificity of polysaccharide of two human gut acetogens. J Biol Chem 2010;285:
use in intestinal bacteroides species determines diet-induced 22082–90.
microbiota alterations. Cell 2010;141:1241–52. [48] Martens EC, Koropatkin NM, Smith TJ, Gordon JI. Complex
[32] King AJ, Cragg SM, Li Y, Dymond J, Guille MJ, Bowles DJ, glycan catabolism by the human gut microbiota: the Bacter-
et al. Molecular insight into lignocellulose digestion by a oidetes Sus-like paradigm. J Biol Chem 2009;284:24673–7.
3864 Gut microbial adaptations to carbohydrates

[49] Dodd D, Moon YH, Swaminathan K, Mackie RI, Cann IK. Spring Harbor, NY: Cold Spring Harbor Laboratory Press;
Transcriptomic analyses of xylan degradation by Prevotella 2009.
bryantii and insights into energy acquisition by xylanolytic [70] Carpita N, McCann M. The cell wall. Biochemistry and
bacteroidetes. J Biol Chem 2010;285:30261–73. Molecular Biology of Plants. Somerset, NJ: John Wiley &
[50] Lowe EC, Basle A, Czjzek M, Firbank SJ, Bolam DN. A scissor Sons; 2000. p. 55–108.
blade-like closing mechanism implicated in transmembrane [71] Albersheim P, Darvill A, Roberts K, Sederoff R, Staehelin A.
signaling in a Bacteroides hybrid two-component system. Plant cell walls. New York, NY: Garland Science; 2010.
Proc Natl Acad Sci USA 2012;109:7298–303. [72] Hill RRH. Digestion of mucin polysaccharides in vitro by
[51] Zhu Y, Suits MD, Thompson AJ, Chavan S, Dinev Z, Dumon bacteria isolated from the rabbit cecum. Curr Microbiol
C, et al. Mechanistic insights into a Ca2+-dependent family 1986;14:117–20.
of alpha-mannosidases in a human gut symbiont. Nat Chem [73] Marcobal A, Southwick AM, Earle KA, Sonnenburg JL. A
Biol 2010;6:125–32. refined palate: bacterial consumption of host glycans in the
[52] Flint HJ, Scott KP, Duncan SH, Louis P, Forano E. Microbial gut. Glycobiology 2013;23:1038–46.
degradation of complex carbohydrates in the gut. Gut [74] Tuomivaara ST, Yaoi K, O'Neill MA. Generation and
Microbes 2012;3:289–306. structural validation of a library of diverse xyloglucan-
[53] Koropatkin NM, Cameron EA, Martens EC. How glycan derived oligosaccharides, including an update on xyloglu-
metabolism shapes the human gut microbiota. Nat Rev can nomenclature. In: York WS, editor. Carbohydr Res
Microbiol 2012;10:323–35. 2014. http://dx.doi.org/10.1016/j.carres.2014.06.031 [in
[54] Chassard C, Delmas E, Robert C, Lawson PA, Bernalier- press].
Donadille A. Ruminococcus champanellensis sp. nov., a [75] Hilz H, de Jong LE, Kabel MA, Verhoef R, Schols HA,
cellulose-degrading bacterium from human gut microbiota. Voragen AGJ. Bilberry xyloglucan—novel building blocks
Int J Syst Evol Microbiol 2012;62:138–43. containing beta-xylose within a complex structure. Carbo-
[55] Png CW, Linden SK, Gilshenan KS, Zoetendal EG, hydr Res 2007;342:170–81.
McSweeney CS, Sly LI, et al. Mucolytic bacteria with increased [76] Ray B, Loutelier-Bourhis C, Lange C, Condamine E, Driouich
prevalence in IBD mucosa augment in vitro utilization of mucin A, Lerouge P. Structural investigation of hemicellulosic
by other bacteria. Am J Gastroenterol 2010;105:2420–8. polysaccharides from Argania spinosa: characterisation of a
[56] Jacob F, Monod J. Genetic regulatory mechanisms in the novel xyloglucan motif. Carbohydr Res 2004;339:201–8.
synthesis of proteins. J Mol Biol 1961;3:318–56. [77] Peña MJ, Kong YZ, York WS, O'Neill MA. A galacturonic
[57] Sinnott ML. Carbohydrate chemistry and biochemistry: acid-containing xyloglucan is involved in Arabidopsis root
structure and mechanism. 2nd ed. London, UK: Royal hair tip growth. Plant Cell 2012;24:4511–24.
Society of Chemistry; 2013. [78] Cameron EA, Maynard MA, Smith CJ, Smith TJ, Koropatkin
[58] McDougall GJ, Morrison IM, Stewart D, Hillman JR. Plant NM, Martens EC. Multidomain carbohydrate-binding pro-
cell walls as dietary fibre: range, structure, processing and teins involved in Bacteroides thetaiotaomicron starch
function. J Sci Food Agric 1996;70:133–50. metabolism. J Biol Chem 2012;287:34614–25.
[59] Buleon A, Colonna P, Planchot V, Ball S. Starch granules: [79] Koropatkin NM, Martens EC, Gordon JI, Smith TJ. Starch
structure and biosynthesis. Int J Biol Macromol 1998;23: catabolism by a prominent human gut symbiont is directed by
85–112. the recognition of amylose helices. Structure 2008;16:1105–15.
[60] Nishiyama Y. Structure and properties of the cellulose [80] Koropatkin NM, Smith TJ. SusG: a unique cell-membrane-
microfibril. J Wood Sci 2009;55:241–9. associated alpha-amylase from a prominent human gut
[61] Burton RA, Fincher GB. (1,3;1,4)-Beta-D-glucans in cell symbiont targets complex starch molecules. Structure
walls of the poaceae, lower plants, and fungi: a tale of two 2010;18:200–15.
linkages. Mol Plant 2009;2:873–82. [81] Gloster TM, Turkenburg JP, Potts JR, Henrissat B, Davies GJ.
[62] Burton RA, Gidley MJ, Fincher GB. Heterogeneity in the Divergence of catalytic mechanism within a glycosidase
chemistry, structure and function of plant cell walls. Nat family provides insight into evolution of carbohydrate metab-
Chem Biol 2010;6:724–32. olism by human gut flora. Chem Biol 2008;15:1058–67.
[63] Carciofi M, Blennow A, Jensen SL, Shaik SS, Henriksen A, [82] Kitamura M, Okuyama M, Tanzawa F, Mori H, Kitago Y,
Buleon A, et al. Concerted suppression of all starch Watanabe N, et al. Structural and functional analysis of
branching enzyme genes in barley produces amylose-only a glycoside hydrolase family 97 enzyme from Bacteroides
starch granules. BMC Plant Biol 2012;12:16. thetaiotaomicron. J Biol Chem 2008;283:36328–37.
[64] Hoffman M, Jia ZH, Peña MJ, Cash M, Harper A, Blackburn [83] Hong PY, Iakiviak M, Dodd D, Zhang M, Mackie RI, Cann I.
AR, et al. Structural analysis of xyloglucans in the primary Two new xylanases with different substrate specificities
cell walls of plants in the subclass Asteridae. Carbohydr from the human gut bacterium Bacteroides intestinalis DSM
Res 2005;340:1826–40. 17393. Appl Environ Microbiol 2014;80:2084–93.
[65] Zabotina OA. Xyloglucan and its biosynthesis. Front Plant [84] Renzi F, Manfredi P, Mally M, Moes S, Jeno P, Cornelis GR.
Sci 2012;3:134. The N-glycan glycoprotein deglycosylation complex (Gpd)
[66] Atmodjo MA, Hao ZY, Mohnen D. Evolving views of pectin from Capnocytophaga canimorsus deglycosylates human
biosynthesis. In: Merchant SS, editor. Annu Rev Plant Biol, IgG. PLoS Pathog 2011;7:e1002118.
64; 2013. [85] Kashyap PC, Marcobal A, Ursell LK, Smits SA, Sonnenburg
[67] Scheller HV, Ulvskov P. Hemicelluloses. Annu Rev Plant ED, Costello EK, et al. Genetically dictated change in host
Biol 2010;61:263–89. mucus carbohydrate landscape exerts a diet-dependent
[68] Pauly M, Gille S, Liu LF, Mansoori N, de Souza A, Schultink A, effect on the gut microbiota. Proc Natl Acad Sci USA
et al. Hemicellulose biosynthesis. Planta 2013;238:627–42. 2013;110:17059–64.
[69] Varki A, Cummings RD, Esko JD, Freeze HH, Stanley P, [86] Terrapon N, Henrissat B. How do gut microbes break down
Bertozzi GR, et al. Essentials of Glycobiology, 2nd ed. Cold dietary fiber. Trends Biochem Sci 2014;39:156–8.
Gut microbial adaptations to carbohydrates 3865

[87] Jia ZH, Qin Q, Darvill AG, York WS. Structure of the [96] Freter R. Human intestinal microflora in health and disease.
xyloglucan produced by suspension-cultured tomato cells. In: Hentges DJ, editor. New York, NY: Academic Press; 1983.
Carbohydr Res 2003;338:1197–208. [97] Freter R. Virulence mechanisms of bacterial pathogens.
[88] Vierhuis E, York WS, Kolli VSK, Vincken JP, Schols HA, Van Washington, DC: American Society for Microbiology;
Alebeek G, et al. Structural analyses of two arabinose 1988.
containing oligosaccharides derived from olive fruit xyloglucan: [98] Abbott DW, Hrynuik S, Boraston AB. Identification and
XXSG and XLSG. Carbohydr Res 2001;332:285–97. characterization of a novel periplasmic polygalacturonic
[89] York WS, Kolli VSK, Orlando R, Albersheim P, Darvill AG. acid binding protein from Yersinia enterolitica. J Mol Biol
The structures of arabinoxyloglucans produced by solana- 2007;367:1023–33.
ceous plants. Carbohydr Res 1996;285:99–128. [99] Bowen JH, Kominos SD. Evaluation of a pectin agar
[90] Salyers AA, Gupta A, Wang Y. Human intestinal bacteria as medium for isolation of Yersinia enterocolitica within 48
reservoirs for antibiotic resistance genes. Trends Microbiol hours. Am J Clin Pathol 1979;72:586–90.
2004;12:412–6. [100] Petrof E, Gloor G, Vanner S, Weese S, Carter D, Daigneault
[91] Lin KY, Daniel JR, Whistler RL. Structure of chia seed M, et al. Stool substitute transplant therapy for the eradication
polysaccharide exudate. Carbohydr Polym 1994;23:13–8. of Clostridium difficile infection: “RePOOPulating” the gut.
[92] Kool MM, Gruppen H, Sworn G, Schols HA. Comparison of Microbiome 2013;1:3.
xanthans by the relative abundance of its six constituent [101] Ettinger G, Burton JP, Reid G. If microbial ecosystem
repeating units. Carbohydr Polym 2013;98:914–21. therapy can change your life, what's the problem? BioEssays
[93] Martin M, Portetelle D, Michel G, Vandenbol M. Microor- 2013;35:508–12.
ganisms living on macroalgae: diversity, interactions, and [102] Kootte RS, Vrieze A, Holleman F, Dallinga-Thie GM,
biotechnological applications. Appl Microbiol Biotechnol Zoetendal EG, de Vos WM, et al. The therapeutic potential
2014;98:2917–35. of manipulating gut microbiota in obesity and type 2 diabetes
[94] Aroniadis OC, Brandt LJ. Fecal microbiota transplantation: mellitus. Diabetes Obesity Metab 2012;14:112–20.
past, present and future. Curr Opin Gastroenterol 2013;29: [103] Larkin MA, Blackshields G, Brown NP, Chenna R,
79–84. McGettigan PA, McWilliam H, et al. Clustal W and Clustal
[95] Freter R. Experimental enteric Shigella and Vibrio infections X version 2.0. Bioinformatics 2007;23:2947–8.
in mice and guinea pigs. J Exp Med 1956;104:411–8.