Functional Attributes and Health Benefits of Novel Prebiotic Oligosaccharides Derived from Xylan, Arabinan, and Mannan

2.1 Enzymes and nutrient absorption in the digestive tract

The small intestine of the human gastrointestinal system contains amylases (from the pancreas) to break down glycogen and starch, a 6-carbon sugar with α-1,4 bonds, and brush border enzymes (lactase, maltase, dextrinase, sucrase) to break down short chain glucooligosaccharides and disaccharides such as sucrose, lactose, and maltose into glucose, fructose, and galactose [2]. Other enzymes aid in the digestion of fats and proteins. The epithelium in the jejunum and ileum of the small intestine is also specifically designed to absorb 6-carbon sugars such as glucose, fructose and galactose via passive, facilitated and active transport systems. Dimers must be broken down into monomers before absorption, and oligomers can persist further into the colon, where they feed microbes that contain enzymes and transport systems to break down complex polymers and oligomers such as xylan and inulin. Short chain fatty acids (SCFAs) can be absorbed from the small intestine, or from the colon if produced as metabolites of bacterial fermentation. It is estimated that >90% of the SCFAs produced in the colon are absorbed, where they can influence, e.g., hepatic regulation of glucose and lipids, and hormones that regulate satiety [3].

2.2 Microbial communities in the digestive tract

The digestive tract is proposed to contain about 10¹³ bacteria, with 100–300 different taxa and thousands of phenotypes [4]. The dominant gut bacteria identified by 16S RNA are Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria, Verrucomicrobia. The small intestine, where simple carbohydrates are abundant, contains about 1–10000 bacteria per gram of intestinal content, primarily Clostridium, Bacteroides, and Streptococcus species. Zoetendal et al. [5] sampled the small intestine of four healthy subjects, and observed that Bacteroidetes, Clostridium (clusters XIVa, IV, XI, and IX) and Proteobacteria dominated, with some Actinobacteria and Bacilli also present. Zoetendal et al. [5] provides evidence that the microbiota, particularly Streptococcus species, in the small intestine adapts rapidly to changes in dietary intake, particularly carbohydrates, based upon the presence of transport systems and enzymes that quickly and efficiently utilize simple carbohydrates.

The colon, which survives on complex carbohydrates that are not digested or absorbed in the small intestine, contains about 10¹¹ bacteria per gram of intestinal content, mainly Prevotella, Ruminococcus, and Bacteroides. Mucin degraders such as Akkermansia mucinophila are commonly found in the mucous layer. The GI tract, for various reasons, may also contain pathogenic bacteria at levels that may or may not be clinically significant. Furthermore, there is increasing recognition of the impact of the gut microbiota on the efficacy of drugs, since some microbes contain enzymes similar to those in the liver [6, 7], or may initiate breakdown of prodrugs. Such effects have been observed with metformin and L-DOPA, among other drugs, and may account for at least part of the interindividual variation in drug efficacy and side effects.

Most microbes evolved to process conventional 6 carbon sugars, particularly monomeric glucose, fructose and galactose. A smaller fraction is able to use mannose, arabinose, and xylose [8].

3.1 Definition, brief description of different types of prebiotics, their structure, and function

Historically, the definition of prebiotics was specific to oligosaccharides affecting the gut microbiome, and health impacts arising therein. Recently, the International Scientific Association for Probiotics and Prebiotics modified the consensus definition to include other types of compounds that may act as prebiotics, and also included prebiotics that could work outside of the digestive tract [9]. Even so, it is essential that that a prebiotic must selectively stimulate the growth of beneficial bacteria, and that there must be a health benefit arising from the consumption/application of the prebiotic. Molecules such as antibiotics that modify the microbiome by acting as antimicrobial agents against undesirable bacteria would not be considered prebiotics, since they do not act as substrates for beneficial bacteria.

Fructans such as inulin and fructooligosaccharides (FOS) are most common among prebiotics in the marketplace, although galactooligosaccharides (GOS) produced from lactose are also available, and xylooligosaccharides (XOS) were available in Japan since the 1980s [10]. Recently, forms of resistant starch (RS), isomaltooligosaccharides (IMOS), arabinoxylooligosaccharides (AXOS) and mannooligosaccharides (MOS) have become available commercially.

Fundamentally, these prebiotics are all materially different, even though the end goal – promoting growth of beneficial bacteria, is the same. The types of bacteria that can be fed by each of these prebiotics depend upon enzymes and transport systems present in the bacteria, which can vary considerably. The selectivity of a prebiotic is also tied to these enzyme and transport systems; if a high percentage of the bacteria have the necessary enzymes and transport systems, then the prebiotic will feed a diverse array of bacteria, including beneficial bacteria along with undesirable bacteria. Conversely, a highly selective prebiotic may not feed many types of bacteria, because fewer types of microbes have the right microbial “machinery” to utilize these prebiotics. Such prebiotics are less likely to directly feed undesirable bacteria if these bacteria do not have the right transport systems and enzymes. Below, we describe the different chemical structures of prebiotics, along with the enzymes and transport systems responsible for their utilization.

3.2 Chemical structures of prebiotics

Prebiotics are typically comprised of oligomers of 6-carbon and/or 5-carbon sugars, with different bonding structures, and different chain lengths (see Figure 1 for examples) [11].

Fructans from inulin are typically linear oligomers primarily made up of fructose monomers connected by β-2,1 bonds. Fructans from agave tend to have a more complex structure with multiple side branches and β-2,6 linkages in addition to β-2,1 bonds [12]. FFn-type fructans such as inulin are comprised entirely of fructose subunits, whereas GFn-type fructans (typically shorter chain oligosaccharides) may have a glucose subunit connected by an α-2,1 bond onto the main fructan chain. These short-chain GFn-type FOS molecules are typically produced by enzymatically adding fructose subunits onto sucrose (glucose-fructose) as the starting substrate. In contrast, short chain FFn FOS would be produced by hydrolysis of inulin (long chain FFn) using endoinulinases [11, 12].

Galactooligosaccharides (GOS) may also be comprised exclusively of galactose subunits, or it may have a glucose terminal subunit arising from use of lactose (glucose-galactose) as the initial substrate to produce GOS using β-galactosidases. The galactose subunits may be connected by β-1,3, β-1,4, or β-1,6 bonds, and may include branched structures, depending upon the type of β-galactosidase [11, 13].

Xylooligosaccharides (XOS) are primarily comprised of xylose subunits connected by β-1,4 bonds, although longer chain XOS may contain branches of arabinose subunits, acetyl groups, or uronic acids (originally present in the xylan source material) that can influence their functionality. XOS that includes arabinose sub-groups are frequently referred to as arabinoxylanoligosaccharides, or AXOS [11, 14].

Mannooligosaccharides are derived from mannan in biomass, and thus are typically made up of mannose subunits connected by β-1,4 bonds [15]. Isomaltooligosaccharides (IMOS) contain glucose subunits, but vary in their bonding structure (affected by manufacturing method). α-1,6 bonds are typical, with either linear or branched structures [11].

Complex starch structures that resist breakdown by pancreatic enzymes into glucose can persist into the colon – thus leading to the concept of “resistant starch” as a prebiotic. Resistant starch (RS) is available in four forms, depending upon method of manufacture and bond structure [16, 17]. Like starch, most of the bonds are of the α-1,4 variety; the presence of other types of bonds may confer “resistance”. RS1 is conventional starch that may be trapped within whole grains. RS2 is typically starch with more complex branching or bond structures, rendering it less accessible to amylase. RS3 is produced when starch undergoes retrogradation, i.e., cooked starch is cooled below its gelatinization temperature. RS4 starches have undergone chemical modification, usually by acidification or cross-linking [16].

The microbial ability to utilize specific substrates is dependent upon the enzymes and transporters encoded within the cells. Some microbes can utilize a broad set of substrates, while others are more selective. This also points to the selectivity of the prebiotic; preferably, the prebiotic preferentially feeds beneficial bacteria, with limited growth of undesirable bacteria. Table 1 summarizes the key hydrolytic enzymes and the corresponding substrates/reactions [19]. These enzymes may be extracellular, or intracellular, which can influence substrate utilization; intracellular enzymes require a corresponding transporter system.

There are various types of transport systems that move prebiotics into the cell, although microbes are likely to only possess a subset, targeted towards a narrower set of substrates. Some transport systems are specific to (certain) monomers, while others target specific oligosaccharides. Chemical structures must be matched to the structure of the transport system in order to be transported into the cell. Table 2 summarizes key membrane transport systems involved in utilization of prebiotics and other carbohydrates.

From a selectivity perspective, it is advantageous if the substrate is used intracellularly – thus, the requisite transport system plus intracellular fructofuranosidases would have an advantage (i.e., FOS that can be transported intracellularly would have an advantage over FOS/inulin that can only be processed via extracellular enzymes). The restricted capacity of most transporter systems precludes use of long chain fructans by many Lactobacillus species [11].

3.3 Types of microbes with enzyme/transport systems of various types

Microbes that have fructofuranosidase encoded extracellularly are able to use long-chain FFN-type FOS, and inulin. Some species of Lactobacillus (casei, paracasei) and Streptococcus are examples [11]. The resulting short chain fructans may be transported intracellularly for utilization. Mao et al. [23] identified 19 strains from human feces that were capable of metabolizing FOS, including multiple strains of E. coli, Enterobacter cloacae, Bifidobacterium spp., and Lactobacillus spp. Additional bacteria also proved capable of growth on FOS. This includes additional strains of E. coli and Bifidobacteria, along with several strains of Streptococcus, Clostridia, Roseburia, Klebsiella, and Enterococcus [23].

According to Rossi et al. [24], virtually all Bifidobacteria are able to grow on short-chain FOS. However, most Bifidobacteria grew poorly on inulin (only 8 out of 55), because most of the fructofuranosidases are intracellular, and inulin cannot be transported intracellularly. Scott et al. [25] made similar observations when inulin with a DP >25 were fed to Bifidobacteria. However, in a mixed culture system, inulin may be broken down into shorter chain FFn-FOS, and then used by Bifidobacteria. Thus, growth of Bifidobacteria in the presence of inulin is primarily due to cross-feeding in the presence of other microbes that act as primary degraders, rather than direct feeding by inulin.

Different species of Bifidobacteria contain ABC transporters, sucrose permeases, fructose PTS transporters, and MFS transporters. The type(s) of available transporters dictate substrate utilization, whether GFn-type short chain FOS, FFn-type FOS, or analogous substrates. Although several strains of Bifidobacteria can utilize GOS, there are significant differences between strains [26]. Certain strains of B. breve and B. longum contain an extracellular galactanase that breaks down long-chain GOS and galactan in plant fiber, producing di- and tri-saccharides that can be transported into the cell and converted into galactose via intracellular β-galactosidase [27]. B. lactis BI-04 contains lactose permease and ABC transport systems, along with β-galactosidase [28], that enable utilization of GOS. L. acidophilus has the ability to utilize many different prebiotics, with various monomeric subunits and bond structures [28]. Such broad utilization is due to a multiplicity of molecular transport systems and hydrolytic enzymes, including up to nine different enzymes from the GH13 family that act on α-glucan [11].

Starch, owing to its high DP and complex branched structure, would be degraded in the presence of extracellular enzymes that can act on α-1,4 and α-1,6 linkages between glucose subunits. Certain Bifidobacteria, including B. pseudolongum and B. breve have the necessary enzymes for extracellular starch utilization [29].

Microbes such as the L. acidophilus cluster (including L. johnsonii, L. helveticus, L. reuteri and L. plantarum) contain LacS permease and β-galactosidase which allow these microbes to transport GOS into the cell, then break it down into glucose and galactose for metabolism [28].

Several Bifidobacteria contain the hydrolytic enzymes needed to break down the β-1,4 linkages present in xylooligosaccharides (XOS) and XOS with arabinose side groups (AXOS). Key enzymes include β-xylosidase and β-xylanase, the latter which breaks down longer chain XOS into shorter chains, ultimately xylobiose, that may be converted into xylose using β-xylosidase. AXOS requires arabinofuranosidase enzymes to process the arabinose side group. Some carbohydrate esterases may also be present to deal with acetyl or feruloyl side groups. The enzymes may be intracellular or extracellular; intracellular enzymes also require transporters such as an ABC transport system to act on the longer chain oligomers. Ejby et al. [30] noted that ABC transporters specific to XOS are exclusive to Bifidobacteria. B. lactis, B. breve, and B. bifidum are among the many species of Bifidobacteria that have the requisite enzymes and transport systems for utilization of short and longer chain XOS and AXOS. Crossfeeding of Bifidobacteria is aided by Bacteroides and Prevotella, which act as primary degraders that break down insoluble xylan in plant fiber into soluble oligosaccharides.

The wide variation in structures of prebiotics, along with the different transport and enzyme systems, ultimately dictate the selectivity of the prebiotic in a mixed culture. Monoculture systems provide some useful insights into the utilization of prebiotics by various substrates. Makelainen et al. [31] conducted a thorough study of the growth of >15 microbes, some beneficial, some pathogenic, in the presence of 11 different carbohydrate sources. Aggregate growth over 24 hours was reported as the area under the curve of DP600 measurements. Figures 2–6 show growth of various probiotics and pathogenic microbes on glucose, FOS, GOS, scXOS, and XOS, respectively. As expected, all bacteria grew well on glucose (Figure 2), consistent with the widespread ability of microbes to utilize simple 6-carbon sugars.

Low DP FF_n-FOS proved to be a good substrate for several strains of Bifidobacteria, along with L. paracasei and L. acidophilus, but was also used by E. coli EHEC, S. epidermis, and C. perfringens, among pathogenic bacteria tested. GOS also grew well on Bifidobacteria and Lactobacilli, but also proved to be an excellent substrate for several pathogenic bacteria, including E. coli and C. perfringens.

Consistent with the unique chemical structure and enzyme/transporter requirements, there was less microbial growth on scXOS and XOS. Fewer strains of Bifidobacteria utilized XOS, along with select strains of Lactobacillus. Li et al., in a study using 29 Lactobacillus strains and 35 strains of Bifidobacterium, observed that all Bifidobacterium strains tested grew on a high dose of XOS, and 30 of 35 strains grew on low dose XOS [32]. They also noted that Lactobacillus strains were able to utilize XOS, albeit with fewer strains and at lower efficiency compared to Bifidobacteria.

However, importantly, Makelainen et al. [31] noted minimal growth of pathogenic bacteria in the presence of XOS (Figures 5 and 6), consistent with a much higher selectivity of XOS for beneficial bacteria. This is a key advantage in a mixed microbial environment such as the GI tract. Bifidobacteria and Lactobacillus species have to compete with many more bacteria for FOS and GOS, which thus increases the dose required for efficacy. XOS, conversely, is better targeted to Bifidobacteria, and in a mixed culture, could be efficacious at a lower dose.

The aggregate area under the curve data reported by Makelainen et al. [31] do not, however, capture changes in growth rates, which can vary over time. Similarly, any issues with viability of microbes could be masked by rapid early growth, which may not be sustained. Figure 7 illustrates growth of a strain of B. breve on FOS, XOS, and inulin, showing temporal effects. A noteworthy observation is that viability of B. breve decreased significantly after ~12–16 hours if grown on FOS or inulin, whereas growth on XOS sustained B. breve for a longer period, even up to 48 hours. This may have important implications in terms of sustaining key microbes in the digestive tract.

In the next section, we describe health impacts of prebiotics, with a particular emphasis on studies with XOS, AXOS, and MOS, due to their distinct chemical structures and selectivity for beneficial bacteria.

4.1 XOS/AXOS

As noted previously, XOS and AXOS are oligomers of 5-carbon sugars, connected by β-1,4 bonds. Fewer types of microbes contain the enzymes and transport systems necessary to utilize XOS and AXOS, thus conferring greater selectivity to beneficial bacteria (see Figures 5 and 6). Mannooligosaccharides (MOS), although based upon a 6-carbon sugar backbone, are connected via β-1,4 bonds that also render them less susceptible to microbial utilization. In this section, we summarize clinical trial results from these novel prebiotic oligosaccharides.

Tables 3 and 4 summarize clinical trial results with XOS and AXOS, respectively, outlining impacts on the microbiome and various clinical biomarkers. Clinical trials with XOS were conducted with doses ranging from 0.4 to 8 g per day; most were in the range from 1 to 3 g per day. A statistically significant increase in Bifidobacteria was observed in most trials with XOS [34, 35, 36, 37, 38, 39]. Improvements in laxation were noted by Childs et al. [34], Chung et al. [35], Iino et al. [40], and Tateyama et al. [41]; the latter was a study in constipated pregnant women. Studies by Childs et al. [34], Na and Kim [39], and Sheu et al. [42] observed improvements in triglyceride and cholesterol levels, at doses as low as 2.8 g/d, whereas Yang et al. [43] did not observe changes in triglycerides when dosing XOS at 2 g/d (2.8 g/d of a 70% purity product). Yang et al. [43] observed a tendency towards a reduction in OGTT insulin in prediabetes patients dosed with 2.8 g/d of 70% XOS, but no effect on blood glucose, unlike Na and Kim [39], who noted a reduction in blood glucose at a XOS dose of 2.8 g/day. Sheu et al. [42], in a trial with diabetic patients receiving 4 g/d of XOS, observed a statistically significant reduction in blood glucose and HbA1C.

Table 3.

Summary of clinical trials with XOS.

Table 4.

Summary of clinical trials with AXOS.

Studies with AXOS complement studies with XOS, since most AXOS products are comprised of at least ~50% XOS. All studies noted an increase in Bifidobacteria [44, 45, 46, 47, 48]. Francois et al. [45] and Walton et al. [48] observed increased fecal levels of SCFAs (acetate, propionate, and butyrate). Reductions in cresol, indicative of reduced protein fermentation, were observed by Cloetens et al. [44] and Francois et al. [45] with AXOS, and Lecerf et al. [38] with XOS.

The bifidogenic effect and health benefits observed in the various clinical trials with XOS and AXOS where generally observed at a dose less than 4 g/d. By comparison, approximately 10–20 g/d of FOS is needed to trigger a bifidogenic effect, and the required inulin dose is stated to be at least 15 g/d [49, 50]. Alfa et al. [51] required a dose of 30 g/d of resistant starch to enhance Bifidobacterium levels.

Miremadi et al. [52] summarized clinical trials in which various prebiotics were evaluated to assess cardiovascular impacts, particularly reductions in cholesterol, lipids and blood pressure. No improvements to lipid profile were observed with 5.5 g of GOS or 20 g/d of FOS (type not specified), whereas triglyceride levels improved following consumption of 10 and 20 g/d of inulin. Similarly, Alles et al. were unable to detect changes in blood glucose and lipid profiles following administration of 15 g/d FOS [53].

The higher doses required for a bifidogenic effect or clinical efficacy compared to XOS and AXOS is likely due to the greater selectivity of XOS and AXOS for beneficial bacteria, versus the widespread utilization of FOS and GOS [11, 23, 31]. For example, Mao et al. [23] found that 237 out of 453 strains (114 genera) of gut bacteria contained the necessary transporters and enzymes for some type/degree of FOS utilization, suggesting fairly widespread utilization of FOS, which will impact the effective dose required to grow the targeted beneficial bacteria. Similarly, Goh et al. [11] note the widespread ability of microbes to utilize lactose, suggesting the presence of LacS (or similar) transporters and enzymes in many microbes that would allow use of the GGa_n type of GOS. This is consistent with the observations of Makelainen et al. [31], who also observed excellent growth of many species, beneficial and pathogenic, on GOS (see also Figure 4). Ultimately, increased sharing of prebiotic substrates among bacteria means that a higher dose is needed to trigger sufficient growth of the targeted beneficial bacteria that produce the desired health benefits.

4.2 MOS

Manno-oligosaccharides (MOS), extracted from yeast cell walls, coffee or biomass rich in mannan, are typically comprised of up to 6 mannose subunits connected by β-1,4 bonds. MOS have not been as extensively evaluated in humans, but have been recognized as a nutritional supplement for livestock.

Salinardi et al. [54] observed a reduction in body volume and adipose tissue in men consuming 4 g/d MOS, attributed to an increase in fat excretion in the feces or inhibition of hepatic lipogenesis. Kumao et al. [55] observed a reduction in fat utilization and an increase in fecal fat excretion in people consuming 3 g/d of MOS for 7 days. St.-Onge et al. [56] also observed a reduction in body weight and adipose tissue following consumption of 4 g/d MOS for 12 weeks. Umemura et al. evaluated the impact of 1 g/d of MOS consumption on fecal microbiota and laxation [57]. They noted an increased ratio of Bifidobacterium spp., and enhanced defecation frequency/volume, reducing constipation. A study in mice by Zheng et al. [58] suggested that MOS acted synergistically with metformin, altering the gut microbiota in a manner that would decrease clinical diabetic parameters, including a reduction in blood glucose. This may have promise for future clinical trials and application of MOS plus metformin for management of type II diabetes.

Jahromi et al. [59] examined MOS supplementation (1 g/kg) to broiler chicks, and observed increased levels of Lactobacillus spp. and Bifidobacteria spp., while reducing levels of E. coli and Enterobacter by >50%. Navidshad et al. [60] compared MOS derived from yeast cell walls with MOS from palm kernel expeller, assessing their efficacy as a supplement (2 g/kg) to the diet of broiler chicks. The yeast-derived MOS improved weight gain, while reducing the intestinal percentage of E. coli in the birds. MOS has been reported to have receptors for fimbriae on E. coli, which can help to control or limit colonization within the digestive tract [61]. Jahromi et al. [62] also evaluated In Ovo injection of MOS, and feeding of MOS to chicks. The single In Ovo injection had some short term but limited long term effects, other than an increase in Bifidobacterium spp. at 14 days. Feeding MOS in the diet markedly improved levels of Lactobacillus and Bifidobacteria, while reducing levels of pathogenic strains of Salmonella, E. coli, and Campylobacter. Adding MOS to the diet also improved levels of serum immunoglobulins IgA, IgM, and IgG. The immunomodulatory effect of MOS in chicks was stated to be a significant benefit that could help improve productivity, and reduce disease (thereby reducing use of antibiotics). Zhao et al. [63] studied the impact of supplementing 0.1 wt% MOS or 0.1 wt% FOS in weanling pigs over 28 days. They observed greater average daily weight gain and average daily feed intake in pigs consuming MOS compared to controls. Nutrient digestibility also improved, along with diarrhea score (potentially by inhibiting E. coli). Collectively, the authors concluded that MOS could enhance piglet growth and health.