The Potential of Pectins to Modulate the Human Gut Microbiota Evaluated by In Vitro Fermentation: A Systematic Review

Pectin is a dietary fiber, and its health effects have been described extensively. Although there are limited clinical studies, there is a growing body of evidence from in vitro studies investigating the effect of pectin on human gut microbiota. This comprehensive review summarizes the findings of gut microbiota modulation in vitro as assessed by 16S rRNA gene-based technologies and elucidates the potential structure-activity relationships. Generally, pectic substrates are slowly but completely fermented, with a greater production of acetate compared with other fibers. Their fermentation, either directly or by cross-feeding interactions, results in the increased abundances of gut bacterial communities such as the family of Ruminococcaceae, the Bacteroides and Lachnospira genera, and species such as Lachnospira eligens and Faecalibacterium prausnitzii, where the specific stimulation of Lachnospira and L. eligens is unique to pectic substrates. Furthermore, the degree of methyl esterification, the homogalacturonan-to-rhamnogalacturonan ratio, and the molecular weight are the most influential structural factors on the gut microbiota. The latter particularly influences the growth of Bifidobacterium spp. The prebiotic potential of pectin targeting specific gut bacteria beneficial for human health and well-being still needs to be confirmed in humans, including the relationship between its structural features and activity.


Introduction
Pectin constitutes a family of complex heteropolysaccharides present in the primary cell walls and middle lamella of higher plants ( Figure 1). Pectin is widely used as a food additive, and its chemical structure depends on its botanical origin, part of the plant used for extraction, and extraction method. Structurally, pectin polysaccharides share some similar features. Homogalacturonan (HG) or 'smooth region' is the most abundant domain (approx. 60%) which is primarily composed of a homopolymer of partially methyl-esterified α-1,4-D-galacturonic acid (GalA) units. Rhamnogalacturonan I (RGI) or 'hairy' region consists of a backbone of repeating disaccharides based on GalA and rhamnose units. RGI regions account for about 15-20% of pectin molecules and are highly branched structures with neutral sugars (mainly arabinose and galactose) and side chains (arabinans, galactans and arabinogalactans) attached to the rhamnose units. RGI is highly present in carrots, okra, tomatoes, and potatoes, where its side chains principally contain arabinan in apples and sugar beet, and galactan in potatoes. Other pectin domains include rhamnogalacturonan II (RGII) and xylogalacturonan, the former being an overly complex branched structure composed of a HG backbone branched with L-rhamnose, D-galactose and other minor sugars. RGII typically accounts for 10% of pectin, and it has been described as the most complex polysaccharide known [1,2]. As a food additive, pectin on the current market is commonly obtained from citrus, apple, and sugar beet, containing a minimum of 65% of GalA. The technical classification of pectin is based on its degree of methyl esterification (DM), (i.e., the percentage of GalA units esterified with methanol). High-methoxy (HM) pectin has over 50% of its carboxyl groups esterified with methanol whereas low-methoxy (LM) pectin has less than 50% [2]. Aside from DM, other structural features govern the suitability of pectin for specific applications, including the molecular weight, GalA content, neutral sugars content, and proportion of HG:RG regions [4].
Pectin qualifies as a dietary fiber, since it is neither digested in the stomach nor the small intestine, but largely fermented in the large intestine [5]. Fiber-associated health benefits have been shown with pectin in vitro such as enhanced antihypertensive effect in fermented food products [6], as well as in vivo, such as a reduction in postprandial glycemic response and the maintenance of blood cholesterol in a normal range [7,8]. Dietary fibers can modulate the gut microbiome and specific bacterial groups, and their variety might be key to supporting it through cross-feeding interactions [9]. It could then be questioned if pectic substrates can be classified as prebiotics, a term defined as a "substrate that is selectively utilized by host microorganisms conferring a health benefit" [10]. Several human clinical studies have shown the potential benefits of pectin on gut health [11,12] (e.g., by reducing digestive symptoms such as regurgitation in infants, and the alleviation of diarrhea or intolerance symptoms in adults fed enteral nutrition) [13,14]. Overall, it is unclear how these beneficial effects are related to the fermentation of pectin in the large intestine. Until now, human clinical studies that have studied the effect of the dietary supplementation of pectin on gut microbiota composition are still scarce [15,16].
In vitro gut models have been widely used to study the impact of diet on the gut microbiota since they allow gaining insight into the fermentation processes mediated by the gut microbiota [17]. In vitro gut models vary in design, from simple batch incubations to more complex semi-continuous or multi-compartmental continuous models representing distinct parts of the human colon [18]. These fermentation models coupled with recent advances in high-throughput sequencing techniques and culture-independent methodologies, such as sequencing of the 16S rRNA gene have allowed the extensive investigation of the diversity, function, and dynamics of the gut microbial communities. Earlier (and widely employed) cultivation-independent 16S rRNA-based methods such as quantitative (realtime) PCR and fluorescent in situ hybridization (FISH) target specific bacterial groups and exhibit an overall good taxonomic resolution and sensitivity. High-throughput methods including next-generation sequencing techniques provide sequencing of PCR amplicons of the 16S rRNA gene or fragmented total (meta)genomic DNA from the whole community (e.g., Illumina, PacBio, ion semiconductor sequencing (Ion Torrent), and Nanopore sequencing, among others). Although 16S rRNA gene-based technologies pose different advantages and limitations, their adaptation allows superior monitoring of changes in the overall microbial community diversity due to fiber consumption [19].
Most of the research investigating the effects of pectin and pectin-derived substrates on the human gut microbiota has been performed in vitro, and to this date, this evidence has not been systematically reviewed [20,21]. Therefore, this review aims to be the first to systematically evaluate in vitro fermentation studies using human fecal samples to (1) determine the state of evidence of the potential effects of pectic substrates on the gut microbiota composition and their fermentative activities, and (2) clarify the potential structure-function relationships based on the complexity of the pectin molecular structure. This knowledge could aid in the better design of human clinical studies, and the development of pectin-derived ingredients with the greatest prebiotic potential.

Materials and Methods
This systematic review was conducted to elucidate the available evidence on the effects of pectin and pectin-derived substrates on the human gut microbiota using an in vitro fermentation setup. This review was conducted in line with the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [22].

Search Strategy and Eligibility Criteria
A literature search was performed in the electronic databases PubMed and Scopus on 29 January 2021. The detailed search strategies used for each database are described in Supplementary Table S1. Peer-reviewed publications were also identified by searching the reference lists of other papers, and they were identified after the search date and until submission. To be included in the systematic review, the following eligibility criteria were used: (1) an in vitro fermentation system was used with human gut microbiota as inoculum (fecal and from ileostomy), (2) pectic substrates (pectin and/or pectin-derived substrates extracted from different raw materials, and in some cases further modified by different treatments) were tested individually (not in a blend), (3) the gut microbiota composition was assessed using comprehensive molecular biological methodologies, (4) the gut microbiota composition and fermentation activity (if also studied) were included as results of the study, (5) the articles were published in the English language, and (6) the articles were published after 1 January 2010, to evaluate the most recent data available.

Study Screening
Two authors (F.G. and N.P.) performed the primary screening (i.e., title and abstract). The articles were assessed for their eligibility, and disagreements were resolved through consensus with a third author (F.R.). Supplementary literature searches involved examining the reference lists of all relevant studies and review articles to identify articles that were not captured in the initial search. Additional articles were selected for inclusion when deemed necessary. The full text of all potential eligible articles was retrieved, and consensus between the three investigators determined the final eligibility of each reference. Data from each eligible article were extracted by two authors (N.P. and F.R.) and included the author, subject (number, health status, and age), fecal inoculum (pooled and from a single donor), test product description and comparators included in the study, concentration of the ingredient used, pH (controlled and non-controlled), sampling time (hours and days), method of analysis of the gut microbiota, and main outcomes in terms of microbiota composition and fermentative activities.

Study Selection and Characteristics
The study identification and selection are detailed in the PRISMA flow chart ( Figure 2). The initial electronic search generated 1634 unduplicated records across 2 databases. Based on the inclusion and exclusion criteria applied to the titles and abstracts of all studies, 50 publications along with 4 manually added articles were selected for a full-text review, and of those, 42 articles were included in this systematic review. All 42 studies investigated the effect of pectic substrates on the human gut microbiota via in vitro fermentations, and of those, 5 studies also investigated their fermentation in upper gastrointestinal tract conditions. Overall, 31 studies were performed using batch fermentations and 11 studies used continuous fermenters, such as the Simulator of Human Intestinal Microbial Ecosystem (SHIME ® ) and the TNO in vitro colon model (TIM-2). Differences were found regarding the fecal inoculum of single donor vs. pooled, where most batch fermentations were performed using fecal samples from single donors (n = 1-17 donors). In contrast, approximately half of the continuous fermenter studies were performed with pooled fecal inocula (n = 3-8 donors) (Figure 3). Another difference found between the fermentation systems was the duration of the fermentation process. Batch fermentations were commonly performed for between 10 and 72 h, while continuous processes were mostly carried out in from 3 days to 21 days. Furthermore, pH differences were found among the batch fermentation and continuous fermenter studies, and interestingly, most batch fermentations (23 studies) were not pH-controlled. From the studies that were pH-controlled, 6 studies used a pH range resembling distal colon (DC) condi-tions (i.e., pH = 6.7-7.0), 1 study used a pH range of 5.8-6.3, and 1 study used 2 different conditions (i.e., pH = 5.5 and 6.5). In contrast, continuous fermenters (all pH-controlled) were either performed at a specific pH (6 studies, pH = 5.8 or 6.2) or at a pH that gradually changed during fermentation, resembling the different parts of the luminal colon (5 studies, pH = 5.6-6.9). Regarding the approach used in the technologies for analysis of the gut microbiota composition, targeted and non-targeted high-throughput technologies were used in the batch fermentation and continuous fermenter studies. Non-targeted approaches were more commonly used (18 batch fermentation studies and 9 continuous fermenter studies) than targeted approaches, and among the latter, qPCR was the most-used technology. Thirty-three studies were conducted with fecal samples from healthy adults with a normal range for their body mass index (BMI, kg/m 2 ) who did not receive antibiotics for the last 2-6 months (3 months in general). One study was performed with samples from ileostomy and not fecal samples. One study was conducted with fecal samples from the elderly. Three studies evaluated the impact of pectin on the gut microbiota from overweight or obese subjects, and two of them compared lean and obese subjects in their experiments. Two studies were conducted using fecal samples from patients with ulcerative colitis (UC), one study was conducted with fecal samples from patients (condition not defined) with no gastrointestinal disorders, and one study was conducted using fecal samples from patients with cirrhosis.

Tested Products
The type of pectic substrates tested varied among the studies based on raw material and structural differences (Table 1). In terms of structure, pectin was the most tested substrate, followed by oligosaccharides, RGI, and other fibers rich in pectin ( Figure 4). In terms of raw material, most studies included pectic substrates from citrus and sugar beet, followed by apple and potato.

Effects on Gut Microbiota Composition and SCFA Production
The fermentation of pectic substrates was relatively slow (18-30 h) but complete, regardless of their structure, compared with other fibers and prebiotics. The pectic substrates were able to modulate specific populations of fecal microbiota, promoting the production of SCFA (Tables 2 and 3). In most studies, the composition of the gut microbiota after fermentation of the pectic substrates was remarkably different compared with other prebiotics, and this specific effect was evidenced by an increased abundance at the genus, species, or even strain level.
In the phylum Bacteroidetes, the relative abundances of the genus Bacteroides and the Bacteroides-Prevotella group were increased in fermentations with pectin extracted from different raw materials or with polygalacturonic acid (18 studies), as well as in fermentations with RGI (4 studies). Furthermore, two studies evidenced the specific effect of pectin on the relative abundance at the species level, where increases in Bacteroides vulgatus, Bacteroides stercoris and Bacteroides dorei were found.
In the phylum Firmicutes, the Ruminococcaceae family was promoted in fermentations with native pectins (11 studies). At the genus level, Lachnospira was commonly increased in fermentations of pectic substrates with a variety of structures (10 studies). Lachnospira eligens, previously known as Eubacterium eligens [23], is particularly increased in pectin fermentations (four studies). In six studies, the genus Roseburia was promoted in fermentations with different pectic substrates, and in some of them, increased abundances of R. intestinalis and R. hominis were observed. The genus Faecalibacterium was commonly increased in fermentations with pectic substrates (14 studies), and of those, pectin particularly promoted an increased abundance of F. prausnitzii (5 studies).
A structure function relationship can be observed, and some bacteria will preferentially use a sub-part of the pectin structure (linear or branched) or a pectin with either a high or low DM. Fermentation of the pectic substrates resulted in moderate SCFA levels, with acetate being the highest and most abundant, as evidenced in 24 studies, followed by lower levels of propionate and butyrate. Furthermore, fermentation of the pectic substrates resulted in higher acetate levels (15 studies) compared with fructans. Propionate production was either low or moderate, depending on the substrate tested, and butyrate production was generally lower compared with other fibers or prebiotics and fructans in particular.

•
PGalA fermented rapidly overall (within 4 h), and more slowly when present in a mixture of fibers rather than alone (contrary to arabinoxylan), and 40% remained intact after 24 h. • ↑ Bacteroidetes (all fibers), ↑ Bacteroides (all fibers), and Lachnospira (≈ 3 to 7-15% of relative abundance). The latter was unique to PGalA but not increased by the mixture of fibers.

Discussion
To our knowledge, this is the first systematic review aiming to elucidate the prebiotic effect of pectic substrates investigated in vitro using human gut microbiota. Pectin and other pectin-derived substrates such as hydrolyzed pectins, RG1, and oligosaccharides have been widely studied, and recent studies even indicate that many of the structural characteristics of pectin have been shown to influence the utilization of these substrates by the gut microbiota. This systematic review demonstrates that fermentation of pectic substrates with human gut microbiota appears to have beneficial effects in terms of specific gut microbiota modulation and SCFA production.

Influence of the Methodology on the Fermentation of Pectic Substrates
Simple fermentation models such as batch fermentations and more complex continuous fermenter models can capture the complex microbial interactions driven by dietary substrates [18]. As of today, no standardized in vitro methodology exists for studying the gut microbiome. Differences in the methodology of these fermentation systems can influence the microbial abundances and metabolite production. The main differences observed between the fermentation systems were pH control, DNA-based techniques for gut microbiota analysis, and fecal inoculum.
The pH conditions play a key role in the competition between bacteria from different phyla or families that share the ability to utilize the same substrate. In pectin fermentations, a slightly acidic pH (pH < 6) suppresses the growth of Bacteroides spp. [55]. Even though most batch fermentations and continuous fermenters can detect an increase in this genus with different pectic substrates regardless of the pH, the effect is more obvious in pH-controlled studies [4,[38][39][40]42,43,49,[54][55][56]. Furthermore, growth inhibition of F. prausnitzii in fermentations with apple pectin at a mildly acidic pH has been reported to be strain-dependent [63]. However, the effect on this species in fermentations of pectic substrates has been detected regardless of pH control in both batch [34,39] and continuous fermenters [4,55].
In terms of differences in the DNA-based techniques for gut microbiota analysis used in the studies, most studies using a non-targeted approach detected an increased abundance of the genus Lachnospira [9,24,29,[51][52][53]58,59,62]. The fact that the genus Lachnospira was not investigated in studies targeting specific bacterial groups could be a possible explanation for its non-detection in these studies. On the contrary, studies using a targetedapproach detected the increased abundances of the genus Lactobacillus and the Lactobacillus-Enterococcus group [32][33][34][35]40,43,46,57], as well as of C. histolyticum [32,34,35,40,42]. Changes in Bifidobacterium abundances were observed in most studies regardless of the DNA-based technique used for gut microbiota analysis. However, this finding is more obvious in studies using a targeted approach, suggesting that this method might be more accurate for detecting the change in this genus [4,[31][32][33][34][35][39][40][41][42][43]46,57].
Using a pooled fecal inoculum rather than the microbiota from a single individual remains controversial, particularly with the concern of how representative such an inoculum is in regard to the colonic ecosystem, considering the abundance and the variety of bacterial species [17,64]. Overall, no major differences were observed between the studies performed with a fecal inocula from single donors compared with pooling. One study recently compared both ways of preparing a fecal inoculum (pooled and single) in an in vitro system monitoring the composition and activity of the gut microbiota under a standard TIM-2 fermentation [17]. Despite some differences observed in certain groups of bacteria in the single-donor fermentations, no major differences were obtained in terms of diversity in the gut microbiota. The majority of the operational taxonomic units (OTUs) were shared among both types of inocula (e.g., Oscillibacter, Alistipes, Bacteroides and Dorea genera were detected to have similar levels under both conditions). Furthermore, a similar metabolic activity in terms of SCFA production in the experiments using fecal inocula prepared from individuals and a pool of these was observed, indicating that the metabolic response of the gut microbiota present in both types of inocula was similar under both conditions.

Common Features for Pectic Substrates in Terms of Fermentation Rate, Gut Microbiota Composition, and SCFA Production
In comparison with other fibers and prebiotics, (e.g., fructans and galacto-oligosaccharides), microbiota fermentation of pectic substrates is slower (within 18-30h) [38,44,51]. The gradual fermentation is also observed when comparing pectic oligosaccharides (POS) to other oligosaccharides, such as fructo-oligosaccharides (FOS) with a similar degree of polymerization (DP) [41]. This phenomenon may be accentuated when pectic substrates are fed to the human gut microbiota in mixtures with other fibers (which may resemble the common human diet) rather than individually [29]. The available literature in humans confirm the slow and complete fermentation of pectic substrates in the large intestine [5], being non-detectable in the feces of adults during intervention studies even at daily doses of up to 30-40 g [65]. Gas production during fermentation (including breath hydrogen, H 2 ) is commonly lower with pectin than with other fibers (e.g., wheat bran, fructans, and lactulose) in both in vitro [26] and human clinical studies [66,67]. Commonly used rapidly fermented prebiotics may lead to discomfort and flatulence in humans, especially at higher doses [68]. In contrast, non-digestible carbohydrates that are slowly fermented may reach the distal part of the colon and promote the growth of beneficial gut microflora, resulting in beneficial metabolites involved in the prevention of intestinal or metabolic diseases [41,69].
The phylum Firmicutes is variably influenced by pectin during in vitro fermentation. Within this phylum, the family Ruminococcaceae, mostly promoted by pectin in continuous fermenters, represents abundant members of the normal microbiome, reaching approximately 10-20% of abundance in healthy humans, where they break down indigestible carbohydrates and produce SCFA [73]. F. prausnitzii, a member of the Ruminococcaceae family, is one of the dominant butyrate producers in the human gut, with a prevalence of approx. 99% [74]. However, its relative abundance is reduced in case of ulcerative colitis or inflammatory bowel disease [75,76], and metabolic disorders [74,77]. Interestingly, F. prausnitzii is promoted in pectin fermentations with fecal samples from healthy subjects [34,39], but also with citrus pectin rich in RGI [27], RGII [39], as well as with oligosaccharides from citrus [34] or sugar beet [52]. In contrast, one study found that the abundance of F. prausnitzii varied in pectin fermentations based on its structural differences [59]. A plausible reason for these differences could be that substrate utilization by this species may be strain dependent, as previously reported [63]. The authors also confirm the ability of intestinal isolates of F. prausnitzii to utilize pectin from apple (but not citrus) in culture growth experiments and that it possesses small repertoires of carbohydrate-active enzymes (CAZyme) encoding genes involved in pectin degradation.
The Lachnospiraceae family presents enzymatic capabilities to degrade pectic substrates among its members [71]. Particularly, L. eligens possesses various enzymes able to degrade pectin, at least two glycoside hydrolases, two polysaccharide lyases, and two carbohydrate esterases. These enzymatic capabilities are also represented in Lachnospira spp. and Roseburia intestinalis, but are absent in the genera Coprococcus and Anaerostipes. This variation in enzymatic capabilities to degrade pectin could justify why the modulation of the Lachnospiraceae family varied among studies. A significant increase in the relative abundance of the genus Lachnospira in the gut microbiota of healthy subjects was observed with citrus pectic substrates and various DM [24,28,59], polygalacturonic acid from citrus pectin [29], RGI [51], as well as potato fiber (rich in pectin) [58]. Additionally, L. eligens can also be stimulated by apple pectin (with no clear effect of a pH level between 5.5 and 6.5-6.9) [39,55]. In these studies, the specific stimulation of L. eligens or Lachnospira was quite unique to pectic substrates and not observed with fructans (FOS or inulin) or resistant starch [24,58]. Lachnospira is not stimulated in microbiota from obese subjects after fermentation with citrus pectin, possibly due to its very low initial concentration [61]. Interestingly, one study demonstrated that the effect is less visible when polygalacturonic acid is fed to the gut microbiota within a blend of three or six fibers, than when presented alone [29].
The selective modulation of these bacterial groups by pectic substrates can be beneficial. For example, the ability of pectin to stimulate L. eligens could be particularly important from a health perspective, since this bacterium is known to exert strong anti-inflammatory effects in vitro to an even greater extent than F. prausnitzii [71]. L. eligens, which is around 90-92% prevalent with a relative abundance between 1-2% [74,78], was recently identified as one of the top 30 bacterial species strongly correlated with the healthy eating index [74], or the Mediterranean diet [79], similar to F. prausnitzii. In these studies, the L. eligens species was particularly negatively correlated with visceral fat, blood triglycerides, and very lowdensity lipoprotein (VLDL) at 6 h but also with fasting and post-prandial glycoprotein acetylation, a biomarker of inflammation. In another clinical study conducted in type 2 diabetes patients, L. eligens was selectively increased by a high dietary fiber diet and negatively associated with postprandial glucose and insulin, body weight, and waist circumference [80].

The Effects of Pectic Substrates on the Production of SCFA
The amount and type of fiber consumed can have dramatic effects on the composition of the intestinal microbiota, and consequently, on the type and amount of SCFA produced [81]. Overall, total SCFA production is generally moderate, but significant, during the fermentation of pectic substrates by the human gut microbiota. In the first hours of fermentation, the total SCFA levels were lower for pectic substrates compared with fructans. Furthermore, pectic substrates induce a larger amount of acetate produced by the gut microbiota from healthy adults and the elderly (60-83 years), regardless of their botanical source or structure when compared with other fibers [24,30,42,62]. One study showed that the SCFA levels that resulted from pectin fermentation (including acetate) were lower in the obese compared with the lean subjects [17]. Accordingly, two clinical studies showed higher acetate levels (in feces or blood) after dietary supplementation of 20-25 g of pectin, or citrus fiber containing pectin [82,83]. The large amounts of acetate commonly produced in the fermentations of pectic substrates may be in alignment with studies showing an increase in the relative abundance of Lachnospira, since some species of Lachnospira are known to produce acetate [84], and its level is negatively correlated with propionate in two in vitro studies [24,58].
Although more variable, the propionate and butyrate levels are lower compared with the fructans [26,29,38]. This finding is particularly observed in butyrate production in fermentations of different pectic substrates [26,34], hydrolyzed pectins and oligomers [40].
In terms of other metabolites, a limited amount of lactate was observed in pectin fermentations aligned with a limited increase of Bifidobacterium and Lactobacillus, which suggests that it could be further used by other bacteria via cross-feeding interactions [31,34,35]. Only five studies measured ammonia (NH 4 + ) production during pectin fermentations, and its levels were either reduced [4,60,61] or unchanged [17,31].

The Impact on Fermentative Activities Based on Donor Health Status
The effect of pectic substrates on the gut microbiota of overweight or obese subjects [17,49,61] is slightly different compared with healthy or normal weight subjects. For instance, B. vulgatus was stimulated by apple pectin and arabinogalactan only in the gut microbiota of normal weight children but not in overweight ones [49]. An increase in the relative abundances of Bacteroides spp., Bifidobacterium, Catenibacterium, Clostridium cluster XIVb, and Parasutterella were only observed with the gut microbiota from lean subjects with apple fiber and/or sugar beet pectin [17]. The authors argue whether the gut microbiota from overweight subjects might have lost their capacity to degrade the pectic structures due to depletion or a decrease in some bacterial species (e.g., B. vulgatus is reduced in overweight children compared with normal weight children) [49]. Interestingly, LM pectin promotes an increase in bacteria with potential anti-inflammatory effects (Succinivibrionaceae members) and SCFA levels, and a decrease in Lachnospiraceae in obese microbiota [61].
Two studies compared the effect of pectic substrates on the human gut microbiota from UC patients (in remission or relapse) with healthy adults [46,47]. In both studies, a treatment with arabino-oligosaccharides from sugar beet (DP 2-10) resulted in no significant differences in gut microbiota composition between the populations. However, slightly different effects were found between the studies. While both showed a reduction of Bacteroidetes, one study reported an increased relative abundance of Firmicutes (for oligosaccharides with a DP4 only) [47], and the other one reported a decrease in the same phylum [46]. Interestingly, a limited SCFA production was obtained in comparison with other studies, and the common production of acetate was only seen with microbiota from relapse subjects [46,47] The effect of DM of pectic substrates on the modulation of the gut microbiota was investigated in fives studies [27,30,32,40,59]. LM pectins were previously reported to be fermented faster than HM pectins [85]. Similar findings were observed in two studies comparing pectins with different botanical origins, production processes and molecular weights [27,30]. In contrast, HM pectins can induce a slightly higher production of SCFA, especially propionate [59]. Other authors also reported that DM does not influence the fermentability of pectic oligosaccharides or hydrolyzed pectin, as reflected by a similar SCFA yield and profile [40]. These contradictory findings may suggest that DM either has an overall moderate effect on gut microbiota modulation or that it might depend on the initial gut microbiota composition. The species Prevotella copri, and F. prausnitzii and the family Ruminococcaceae are positively correlated with DM, contrary to many genera such as Coprococcus, Oscillospira, Prevotella, Ruminococcus, and Lachnospira [59]. Regardless of whether these bacteria are considered fast or slow growers, they can impact fermentation kinetics and induce different results between LM and HM pectins in terms of the rate of SCFA production.

Composition of Neutral Sugars
The nature of neutral sugars and the types of linkages in the pectin structure may influence the composition and activity of the human gut microbiota. In this regard, gluco-oligosaccharides and galacto-oligosaccharides are fermented faster than arabinooligosaccharides, and even faster than oligogalacturonides, for which there is a lag time of 5-7 h in vitro [32,34,35]. Galactan molecules are degraded by glycoside hydrolases (GHs). For example, GH2 is present in the genome of many genera in the human gut, such as Bacteroides, Bifidobacterium, and Ruminococcus, and species such as F. prausnitzii, R. intestinalis, Akkermansia muciniphila and Escherichia coli. The genera Bifidobacterium and Lactobacillus possess enzymatic capabilities to degrade β-1,4-galactan chains as a carbon source in potato pulp [37]. The widespread presence of these enzymes in the gut commensals may explain why these molecules are quickly fermented. Furthermore, enzymes such as the GH families that can degrade arabinans (e.g., GH51, GH43, GH27, and GH127), are also largely present in Bacteroides spp. and several species of Bifidobacterium (e.g., B. longum subsp. longum or B. adolescentis). On the contrary, enzymes such as glycoside hydrolases and polysaccharide lyases (PLs) that can degrade the RGI backbone (e.g., PL11) are only present in some Bacteroides spp., and only the GH type appears to be present in F. prausnitzii, R. intestinalis, Klebsiella oxytoca, Enterobacter cloacae, and A. muciniphila [1].
Bifidobacterium spp. (at least B. adolescentis and B. breve) show a poor capacity to degrade pectin or the RGI backbone contrary to Bacteroides [71], but it seems to be more adapted to degrading side chains made of arabinan and galactan [1,37]. This is in agreement with a pure culture experiment showing that Bifidobacterium spp. can partially metabolize distinct types of apple pectin-derived oligosaccharides, except oligogalacturonides [35]. Similarly, Bifidobacterium exhibits a preference for oligosaccharides rich in galactose and arabinose, as shown by several in vitro studies [33,42,43]. Bifidobacterium further develops with alkali-extracted pectin containing the highest RGI ratio (≈60%) than with pectin obtained by acid extraction presenting a lower RGI ratio (≈26%) [27]. Previous genomic and in vitro studies have shown that Bifidobacterium preferentially utilizes arabinan (sugar beet) and arabinogalactan (potato) rather than polymers or oligomers of GalA [42]. Finally, this genus also increases in fermentations with sugar beet arabino-oligosaccharides [33,36,41], and apple-derived oligosaccharides [35]. This preference for RGI certainly because of its side chains has been confirmed in humans, where a dietary supplementation with 15 g of potato fiber did not induce a significant modification of fecal microbiota. However, RGI derived from the same fiber could increase the relative abundance of Bifidobacterium [86]. Additionally, in infants, the consumption of formulae containing citrus pectin-derived oligosaccharides could stimulate the growth of Bifidobacterium [87].
The presence of arabinose (and arabinan side chains) also seems to be positively correlated with higher counts of Prevotella, (and P. copri in particular) but negatively correlated with Lachnospira, Bacteroides ovatus, and members of Coprococcus [59]. In most studies, Prevotella is investigated jointly with other bacterial genera such as Bacteroides, but not on its own, thus not allowing the confirmation of a strong effect of pectic substrates on its growth. Furthermore, the low abundance of Prevotella could also be due to the lower degree of branching and/or arabinose content of the pectic substrates tested in these studies compared which RGI and sugar beet pectin. Since P. copri is generally associated with high plant fiber diets and favorable postprandial glucose metabolism [74], further studies are needed to provide evidence of its fermentative activities.
Interestingly, these structural differences in terms of the neutral sugar composition of pectic substrates, as described above, may result in different fermentation patterns and activities by the gut bacteria. In particular, galactan and arabinan and their respective oligosaccharides, may promote higher total SCFA levels compared with fermentations with polygalacturonic acid and pectin [42].

Distribution of HG and RG Fractions
The distributions of the linear and branched structural regions of pectin have been shown to influence its fermentation by the human gut microbiota. The Ruminococcaceae family and especially F. prausnitzii positively correlated with GalA and DM [4,17,59]. This suggests that different microorganisms may exhibit specific preferences for defined substrates, which may be the case for F. prausnitzii that preferentially utilized the HG backbone compared with RG, and HM compared with LM pectins. Previous research has demon-strated the enzymatic capabilities of F. prausnitzii to degrade pectin and RGI-backbone [1], as well as GalA [88]. The utilization of the latter in particular was also confirmed for several strains of F. prausnitzii [63]. These strains are able to compete with B. thetaiotaomicron and L. eligens for pectin degradation, especially in mildly acidic pH conditions. Interestingly, the ability of the specific strains of F. prausnitzii to function both as a consumer and a producer of acetate, depending on the conditions, has been previously reported, and in particular, a higher uptake of GalA as carbon source is facilitated by acetate production by the bacteria [88].

Degree of Branching
The degree and distribution of neutral sugar branches attached to the RGI region may influence the molecular conformation of pectins [89], and further influence their fermentability [42]. Pectins derived from sugar beet or apple are usually characterized by a higher degree of branching compared with those of a citrus origin [89]. Sugar beet-derived pectin and POS may be fermented more slowly, and not completely, than from apple or citrus [32]. Furthermore, the genus Lachnospira, particularly promoted in pectin fermentations, is positively correlated with the degree of branching [59]. The same study also showed a similar finding regarding the genus Coproccocus, contrary to the species P. copri.

Molecular Weight
The molecular structure, and even nature of monosaccharide units, may influence the kinetics and extent of SCFA production. POS are generally fermented faster and to a greater extent (resulting in the greater production of SCFA and pH reduction) than pectin polysaccharides from the same botanical origin [32,34,42]. A minor difference in the degree of polymerization (four vs. five) from the same type of substrate induces a different effect on the gut microbiota. While a decrease in the Bacteroidetes phylum was observed for both oligosaccharides, a significant increase in Firmicutes was only seen with the substrate with DP4 [47]. Furthermore, Bifidobacterium was more stimulated by oligosaccharides from potato-derived RG1 (without a difference between DP 2-12 and DP 13-70) than by polysaccharides [57]. A similar effect was observed in fermentations with modified pectins from citrus and artichoke compared with their parent pectins [43]. The genus Bifidobacterium exhibits a well-known preference for shorter molecules (e.g., the fermentation of fructanderived substrates), as was recently reviewed [90], even though this was not observed in a previous study with different pectic substrates [42]. In contrast to Bifidobacterium, the effect of the DP on F. prausnitzii may be strain-dependent. Certain F. prausnitzii strains possess some ability to utilize apple pectin, and in common with L. eligens, they can utilize the galacturonide oligosaccharides with DP4 and DP5 derived from sugar beet pectin [71].

Other Structural Characteristics
One study investigated the possible impact of the presence of amide groups in the pectin structure on the gut microbiota [59]. The authors reported that amidated citrus pectins had a similar effect on the gut microbiota in comparison to non-amidated pectins from citrus or sugar beet. However, both pectins shared a similar (low) DM, and this parameter might have had a stronger influence on the modulation of the gut microbiota than amidation. Another study investigated the effect of feruloyl substitution of short (DP 2-10) and long-chain (DP 7-14) arabino-oligosaccharides on the gut microbiota [36]. In this study, Bifidobacterium was similarly and selectively stimulated by both the feruloylated and non-feruloylated arabino-oligosaccharides. The potential specific effect of this structural difference on the gut microbiota may have been hidden by a greater effect of different DPs between the substrates.

Conclusions
Based on this comprehensive review, we found evidence to support the potential prebiotic effect of pectic substrates based on in vitro studies, as shown by a specific fermentation profile at the genus, species, and even strain levels. It is evident that most pectic substrates can stimulate the growth of Bacteroides and Lachnospira genera as well as species such as F. prausnitzii and L. eligens, and increase the production of SCFA (acetate in particular). The structural characteristics of pectic substrates have been shown to influence their utilization by the gut microbiota. Bifidobacterium in particular, shows a preference for the fermentation of shorter molecules, especially arabinose-rich side chains such as from RGI. F. prausnitzii prefers HG to RG, but its preferences are less clearly defined and might be determined at the strain rather than species level.
Even though these findings are promising, they should be carefully interpreted since direct extrapolation to humans cannot be performed. A recent clinical trial where a dietary intervention of 15 g/d of sugar beet pectin for 4 weeks resulted in no significant changes in gut microbiota composition or SCFA production [15]. Furthermore, the fermentation profile of pectic substrates by the human gut microbiota varies depending on whether the fiber is provided alone or in a blend, which is more representative of the typical human diet, as was previously reported [29]. Therefore, well-designed clinical trials are needed to prove the specific effect of these substrates on human gut microbiota and the potential health effects derived from it.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/nu14173629/s1, Table S1: The search strategy used for identification of the relevant studies included in the systematic review.