Introduction

Recently, indigestible dietary oligo- and polysaccharides have attracted attention as functional food ingredients that provide health benefits beyond basic nutrition (de Jesus Raposo et al. 2016). These polysaccharides often stimulate the growth and activity of beneficial gut microbes, potentially qualifying as prebiotics. They may act as substrates for fermentation in the large bowel, leading to the production of short chain fatty acids (SCFA) with multiple functions that help maintain health (Conlon and Bird, 2015). Commonly used prebiotics include inulin, fructooligosaccharides, galactooligosaccharides, and lactulose (Al-Sheraji et al. 2013). A growing consumer awareness of the benefits of prebiotics is leading to commercial interest in the isolation and development of polysaccharides and other compounds from novel sources such as marine seaweeds for use as prebiotics.

There is a high diversity of brown seaweeds (Phaeophyceae) in Southern Australia; of the 231 species reported, 57% are considered endemic (Womersley 1990). They are a rich source of functional food ingredients and bioactive compounds, with polysaccharides being a major component, accounting for up to 70% of the dry weight (Holdt and Kraan 2011). As a result, brown seaweeds have a high total dietary fibre content of 33–75%, with soluble dietary fibre accounting for 26–60% of the dry weight (Lahaye 1991). Brown seaweed-derived carbohydrates have been shown to be fermented by gut microbiota and so could act as prebiotics and provide a health benefit to humans (O’Sullivan et al. 2010; Zaporozhets et al. 2014). Recent studies indicate that polysaccharides and oligosaccharides derived from seaweeds can modulate intestinal metabolism including fermentation, inhibit pathogen adhesion and evasion, and potentially treat inflammatory bowel disease (Devillé et al. 2007; Ramnani et al. 2012; Kuda et al. 2015; Lean et al. 2015). Ingested polyphenols with complex structures can also reach the large intestine where they can be converted into beneficial bioactive metabolites by microbes (Cardona et al. 2013), and this has been shown to occur for brown seaweed phlorotannins (Corona et al. 2016). Hence, brown seaweed phlorotannins may also be valuable as a food ingredient capable of providing a health benefit.

Several studies have examined the prebiotic activity of brown seaweeds using in vitro fermentation systems which mimic the human gut (Michel et al. 1996; Devillé et al. 2007; Ramnani et al. 2012; Li et al. 2016; Rodrigues et al. 2016) but none examined the digestibility and prebiotic potential of different fractions extracted from brown seaweed. Our previous study demonstrated for the first time that crude extracts of the brown seaweed Ecklonia radiata have the potential to be used as a source of dietary supplements with gut health benefits in humans (Charoensiddhi et al. 2016a). The aim of the present study was to further understand the prebiotic potential of some key fractions of E. radiata extract by refining the extraction process. Of particular novelty is the use of microwave-assisted enzymatic extraction to disassemble the high degree of structural complexity and rigidity of seaweed cell walls reported by Deniaud-Bouët et al. (2014) and liberate target polysaccharides and polyphenols without significant degradation. We will evaluate the digestibility and prebiotic potential of four major extract fractions, namely a crude extract fraction (CF), a phlorotannin-enriched fraction (PF), a low molecular weight (MW) polysaccharide-enriched fraction (LPF), and a high MW polysaccharide-enriched fraction (HPF). The four seaweed fractions were tested for their prebiotic potential by including them in an in vitro anaerobic fermentation containing human faecal inocula. The production of SCFA and the growth of selected gut microorganisms are used as indicators of prebiotic potential.

Materials and methods

Seaweed

Brown seaweeds (Ecklonia radiata—identification confirmed by the State Herbarium of South Australia) were collected from freshly deposited beach-cast seaweed in Rivoli Bay, Beachport, South Australia, in March 2013. They were rinsed in fresh water to remove any visible surface contaminants and placed on mesh racks to dry. All the seaweed materials were collected at the one time to provide consistent samples for all these studies. They were blended (Blendtec, USA) and dried in an oven at 45 °C. The ground powder was stored at −20 °C prior to extraction and fermentation.

Chemicals and substrates

All chemicals used are of analytical or chromatography grade from Merck and Sigma. The positive control substrate for comparison purposes in batch fermentation is inulin (Sigma). Blank and α-cellulose (Sigma) were used as negative controls. A commercial carbohydrate hydrolytic enzyme (Viscozyme®L) used for the preparation of seaweed extracts was kindly provided by Novozymes (Bagsvaerd, Denmark).

Preparation of seaweed fractions

From our preliminary study (data not shown), the crude extract prepared by a microwave-intensified Viscozyme treatment process (Fig. 1a) showed the potential for improving gut health as this extract induced high levels of SCFA production and growth of beneficial bacteria. In this study, a process was developed to sequentially obtain three extract fractions (Fig. 1b) in order to understand which specific fractions in the seaweed are responsible for the prebiotic activity.

Fig. 1
figure 1

Overview of the process flow chart to prepare a the crude extract fraction (CF) and b the phlorotannin-enriched fraction (PF), the low MW polysaccharide-enriched fraction (LPF), and the high MW polysaccharide-enriched fraction (HPF) by sequential extraction and fractionation of brown seaweed E. radiata

Full size image

Crude extract fraction

The CF was selected for inclusion in this experiment in order to demonstrate the impact of a mixture of components in the seaweed extract on prebiotic potential. This extract was prepared according to the method of Charoensiddhi et al. (2015) and Charoensiddhi et al. (2016b) with some modifications. Briefly, the dried and ground seaweed was dispersed in the pH-adjusted water in the ratio 1:10 (w/v). The pH was adjusted using 1 M HCl to achieve the optimum pH of Viscozyme at 4.5. The enzyme solution was added at 10% (v/w), and the enzymatic hydrolysis was performed under optimal conditions at 50 °C for 30 min in a StartSYNTH Microwave Synthesis Labstation (Milestone Inc., USA). The temperature was controlled using an infrared sensor and automatic power adjustment. The enzyme was inactivated by boiling the sample at 100 °C for 10 min and cooling immediately in an ice bath. The extract was centrifuged at 8000×g for 20 min at 4 °C. The supernatant was collected and adjusted to pH 7.0, freeze dried, and stored at −20 °C to be denoted as CF for further analysis and fermentation.

Phlorotannin-enriched fraction

The complex phlorotannins in brown seaweed have the potential to be fermented by gut microbes. They can be extracted efficiently with 90% ethanol with a small amount of polysaccharide contamination (data not shown). The other advantage is that ethanol is the preferred organic solvent for the extraction of food-grade components. In this study, dried and ground seaweed was first extracted with 90% (v/v) ethanol at an alga solid to solvent ratio of 1:10 (w/v). The suspension was incubated at room temperature for 24 h under continuous shaking in an orbital mixer incubator (Ratek Instruments, Australia), and then centrifuged at 8000×g for 20 min at 4 °C in order to separate (1) the supernatant and (2) the residue 1. The ethanol in the supernatant was evaporated in a rotary evaporator (Rotavapor-R, Switzerland), and the residual aqueous extract was freeze dried and stored at −20 °C denoted as PF for further use. The residue 1 was dried in an oven at 50 °C overnight for further microwave-assisted enzymatic extraction of seaweed polysaccharides.

Low and high MW polysaccharide-enriched fractions

The indigestible dietary polysaccharides are the most important sources for prebiotics, and their characteristics such as low and high MW may also affect the fermentability by gut microorganisms. LPF and HPF from seaweeds were prepared by the same method of preparing crude extract, but using the dried residue 1 after PF extraction instead of dried seaweed. After the enzymatic hydrolysis and centrifugation at 8000×g for 20 min at 4 °C to separate the supernatant and the residue 2, the supernatant was adjusted to pH 7.0. Ethanol was then added to the supernatant to a concentration of 67% (v/v) to precipitate HPF. HPF was precipitated at 4 °C overnight (modified method from Lorbeer et al. (2015)), and then collected from LPF (supernatant) by centrifugation at 8000×g for 20 min at 4 °C. The ethanol in the supernatant (LPF) was evaporated in a rotary evaporator. Both LPF and HPF were freeze dried and stored at −20 °C until further use.

Analyses of major compositions of dried seaweed fractions

The composition of dried seaweed fractions (expressed as g 100 g-1 dry fraction) were investigated according to the methods described in Charoensiddhi et al. (2016a, b). Total protein, starch, dietary fibre, and non-digestible non-starch polysaccharide (NNSP) of all four seaweed fractions were determined using established AOAC methods. Total phlorotannin was analysed by Folin Ciocalteu’s phenol reagent, and the results were expressed as gram phloroglucinol equivalent. Total sugar content was determined by HPLC after sulfuric acid hydrolysis and 1-phenyl-3-methyl-5-pyrazolone derivatisation according to the method of Comino et al. (2013). Briefly, the derivatives were separated and analysed by HPLC (system: Prominence UFLC XR, Shimadzu; column: Kinetex 2.6u C18 100A, 100 × 3 mm, Phenomenex; detection: Prominence SPD-20A UV-VIS Detector, Shimadzu) column, operated at a flow rate of 0.8 mL min−1 and 30 °C. The gradient eluents used were (A) 10% acetonitrile, 40 mM ammonium acetate, and (B) 70% acetonitrile with absorbance detection at 250 nm. Compounds were identified and quantified based on their retention time and peak areas, respectively, and comparison with monosaccharide standards (d-mannose, d-ribose, l-rhamnose, d-glucuronic acid, d-galacturonic acid, d-galactose, d-xylose, l-arabinose, l-fucose, d-glucose, and the internal standard 2-deoxyglucose). Sodium alginate from Sigma was subjected to the same procedure to allow for identification and quantification of the alginate-derived uronic acids.

In vitro batch fermentations

Anaerobic batch fermentations were used to assess the effect of seaweed fractions on fermentation characteristics and composition of gut microbiota. The preparation of faecal sample and fermentation medium, SCFA analysis, and bacterial enumeration were conducted according to the method described in Charoensiddhi et al. (2016a) with some modifications.

In brief, seaweed fractions were fermented in vitro in a batch system under anaerobic conditions for 6, 12, and 24 h. Fresh faecal samples from three healthy individuals 10% (w/v) in sterile anaerobic solution of glycerol 20% (v/v) and cysteine hydrochloride 0.05% (w/v) were used as inoculum. Anaerobic conditions were maintained throughout the fermentations using an anaerobic chamber. Seaweed fraction substrates at a concentration of 1.5% (w/v) in fermentation media were used in each test. For comparative purposes, no substrate was added for the blank, but positive and negative controls supplemented with inulin and cellulose, respectively, were included at the same concentration. Test substrates and controls were fermented in triplicate. Each fermentation test was inoculated with 10% (w/v) of faecal inoculum, and all fermentation tests were incubated in an orbital shaker incubator at 37 °C, 80 rpm.

Fermentations were sampled at 6, 12, and 24 h for SCFA analysis using gas chromatography. Fatty acid concentrations were calculated in micromole per milliliter by comparing their peak areas with standards (acetic, propionic, butyric, isobutyric, valeric, isovaleric, and caproic acids). The bacterial numbers were determined by quantitative real-time PCR (Q-PCR) after 24 h fermentation with a series of microbe-specific primer pairs according to the modified methods from Guo et al. (2008) for Firmicutes, Bartosch et al. (2004) for Bacteroides-Prevotella, and Charoensiddhi et al. (2016a) for the rest of the target bacteria (Bacteroidetes, Bifidobacterium, Clostridium coccoides, Entercoccus, Escherichia coli, Faecalibacterium prausnitzii, and Lactobacillus). These target bacterial genera were selected in the study due to their relevance to gut health.

In vitro digestibility of seaweed components

The digestibility of seaweed fractions were determined using an in vitro human digestion model developed by CSIRO Health and Biosecurity. Briefly, approximately 4.5 g of each seaweed fraction was treated successively with protease, pancreatic, and amyloglucosidase enzymes to mimic upper gastrointestinal digestion. After 16 h at 37 °C, the undigested material was precipitated via the addition of acetonitrile to 66% (v/v) and collected by centrifugation at 2000×g for 10 min. Precipitates were washed with acetone and dried to give the final products. Under these conditions, soluble and insoluble dietary fibre and larger peptides should be precipitated. Sugars, small oligosaccharides, and fructans would be removed with the supernatant. Digestibility was calculated by subtracting the dried weight of sample residue after precipitation from 4.5 g of dried seaweed fraction and was expressed as a percentage.

Statistical analysis

Results were expressed in triplicate as mean ± SEM for SCFA analysis and bacterial enumeration. One-way analysis of variance (ANOVA) was used to compare the means. Differences in SCFA and bacterial enumeration were considered significant at P < 0.05 by Duncan’s test in the IBM SPSS Statistics 22 (IBM Corporation Software Group, USA).

Results

Composition of four brown seaweed fractions

The levels of the main fermentable components in each seaweed extract fraction are shown in Table 1 Fibre and sugar were the major components (of components tested) in CF. Relative to CF, HPF contained approximately fourfold higher total dietary fibre and NNSP content and twofold higher sugar content, while the LPF contained lower levels of fibre. PF contained the highest total phlorotannin content, but the lowest fibre and sugar contents. Protein was not detected in any seaweed fraction, and a small amount of starch was found in LPF and HPF. Each of the four seaweed extract fractions had different sugar composition profiles. Glucose was the major component of all seaweed fractions. Apart from glucose, CF also consisted of fucose and small amounts of mannuronic acid, mannose, galactose, and xylose, while relatively high proportions of fucose and mannuronic acid were found in HPF, followed by guluronic acid, mannose, glucuronic acid, galactose, and xylose.

Table 1 Main nutrients and potential fermentable components (g (100 g)−1 dry fraction) of four seaweed fractions, the crude extract fraction (CF), the phlorotannin-enriched fraction (PF), the low MW polysaccharide-enriched fraction (LPF), and the high MW polysaccharide-enriched fraction (HPF). Values are means of analytical duplicate analyses (n = 2)
Full size table

SCFA production

Total SCFA including acetic, propionic, and butyric acids were monitored during 6, 12, and 24 h fermentation (Fig. 2). Acetic and propionic acids were the predominant SCFA produced in all fermentations. All seaweed fractions and controls produced low levels of isobutyric, valeric, isovaleric, and caproic acids (0.02–3.43 μmol mL−1), so these results were not shown. Concentrations of acetic acid, propionic acid, and butyric acid fermentations increased over the 24-h fermentations of all four seaweed fractions, with the highest concentrations observed at 24 h in each case. No significant increase in butyric acid concentration was observed in fermentations with PF, the positive control (inulin), and negative controls (cellulose and blank). After 24 h fermentation, total SCFA production of CF supplemented fermentation (97.3 μmol mL−1) was significantly higher than that of other seaweed fractions (19.6–89.0 μmol mL−1). Fermentations with all seaweed fractions, except PF, significantly increased total SCFA production (68.9–97.3 μmol mL−1) by almost two to threefold compared to negative controls (cellulose 39.7 μmol mL−1 and blank 36.3 μmol mL−1). Relative to positive control (inulin), total SCFA produced from CF and LPF supplemented fermentations was approximately 10–20% higher, while fermentation with HPF was 20% lower than that of inulin. Of the seaweed fractions tested, the highest levels of acetic (40.8 μmol mL−1), propionic (54.6 μmol mL−1), and butyric (17.3 μmol mL−1) acids resulted from fermentations with the HPF, CF, and LPF, respectively. Importantly, higher levels of propionic and butyric acids were observed in CF, LPF, and HPF supplemented fermentations compared to both positive and negative controls. The acetic acid to propionic acid ratio was different among tested fractions and controls, valued at 0.47 ± 0.04, 1.32 ± 0.09, 0.62 ± 0.06, 2.15 ± 0.06, 4.08 ± 0.18, 5.73 ± 0.13, and 5.64 ± 0.05 for CF, PF, LPF, HPF, inulin, cellulose, and blank, respectively. The initial pH of this batch fermentation was 7.2. The pH of all seaweed fractions decreased, in agreement with the increase of SCFA production during fermentation (data not shown).

Fig. 2
figure 2

Concentration (μmol mL−1 ± SEM) of a acetic acid, b propionic acid, c butyric acid, and d total SCFA in in vitro faecal fermentations supplemented with four seaweed fractions, the crude extract fraction (CF), the phlorotannin-enriched fraction (PF), the low MW polysaccharide-enriched fraction (LPF), and the high MW polysaccharide-enriched fraction (HPF), and controls at 6 h, 12 h, 24 h fermentation; values are means of experimental triplicate analyses (n = 3), and bars with different letters are significantly different from each other (P < 0.05)

Full size image

Bacterial enumeration

The bacterial populations resulting from fermentations with four seaweed fractions and controls are shown in Table 2 Relative to negative controls at 24 h fermentation, the numbers of beneficial bacteria such as Bifidobacterium, Lactobacillus, and C. coccoides in the LPF supplemented fermentation significantly increased approximately 10-fold, while increased numbers of C. coccoides in CF supplemented fermentation were also observed. Although the positive control inulin significantly increased the numbers of most target bacteria compared to negative controls and other seaweed fractions, higher numbers of Lactobacillus, F. prausnitzii, C. coccoides, Firmicutes, and E. coli were observed for the LPF supplemented fermentation compared to inulin fermentation. In contrast, the numbers of Enterococcus in fermentations with PF and CF significantly decreased, approximately 10-fold, but the numbers of E. coli in most seaweed fractions and positive control increased in comparison to negative controls. A significant increase in the bacterial populations of F. prausnitzii was observed for PF and LPF supplemented fermentations compared to both negative controls and only cellulose, respectively. Most target bacteria tested could not be detected in HPF supplemented fermentation determined by Q-PCR except for Bacteroidetes, Firmicutes, and E. coli. The Firmicutes to Bacteroidetes ratio was calculated for each fermentation substrate. These ratios were 1.14 ± 0.001, 1.20 ± 0.008, 1.18 ± 0.006, 1.08 ± 0.008, 1.13 ± 0.004, 1.22 ± 0.004, and 1.25 ± 0.005 for CF, PF, LPF, HPF, inulin, cellulose, and blank, respectively.

Table 2 Bacterial populations (log10 cells mL−1 ± SEM) in batch fermentations at 24 h of four seaweed fractions, the crude extract fraction (CF), the phlorotannin-enriched fraction (PF), the low MW polysaccharide-enriched fraction (LPF), and the high MW polysaccharide-enriched fraction (HPF), and controls; values are means of experimental triplicate analyses (n = 3) from which DNA was extracted in duplicate. PCR amplification was carried out in triplicate from each of these 6 DNA extracts
Full size table

The digestibility test of seaweed fractions

The digestibility of four seaweed fractions in the simulated human digestive system was assessed and the results are shown in Fig. 3. When HPF was subjected to hydrolytic enzymes imitating the upper gastrointestinal, no change was observed in dried weight of the residue. Therefore, HPF was not digestible by enzymes present in the small intestine whereas the CF, PF, and LPF were digested at 71.5, 87.3, and 86.1%, respectively.

Fig. 3
figure 3

The digestibility of four seaweed fractions, the crude extract fraction (CF), the phlorotannin-enriched fraction (PF), the low MW polysaccharide-enriched fraction (LPF), and the high MW polysaccharide-enriched fraction (HPF), tested by in vitro simulation of the enzymes involved in upper gastrointestinal digestion; values are means of analytical duplicate analyses (n = 2)

Full size image

Discussion

Our previous study (Charoensiddhi et al. 2016a) using an in vitro anaerobic fermentation system containing human faecal inocula demonstrated that crude extracts of the brown seaweed E. radiata could be fermented to produce beneficial SCFA and also promoted the growth of some targeted beneficial bacteria and could be considered to be prebiotics. However, the fractions of the complex extract mixture responsible for prebiotic activity were not clearly defined. In this study, key potential fermentable components, particularly low and high MW polysaccharides and polyphenols, were fractionated from E. radiata for further investigation of their prebiotic potential in vitro. The digestibility of these seaweed fractions was also tested to understand the likelihood of components reaching the large bowel and the resident microbiota.

Each of the seaweed fractions tested in this study increased SCFA levels when fermented over 24 h in vitro. However, the rates and extent of production varied and are likely to reflect the different compositions and complex structures of components added (Rodrigues et al. 2016). The increase in SCFA levels was considerably more pronounced in the fermentations supplemented with CF, LPF, and HPF. As expected, the most abundant SCFA produced during each of the in vitro fermentations were acetic acid, propionic acid, and butyric acid, with low concentrations of the branched chain fatty acids. These SCFA can have different physiological impacts within the gut (Topping and Clifton 2001). Butyric acid is the primary energy source of cells lining the colon, and helps maintain colonic tissue integrity through stimulation of apoptosis in cells with high levels of DNA damage (Canani et al. 2011). Acetic acid can inhibit the growth of enteropathogenic bacteria (Fukuda et al. 2011), and propionic acid produced in the gut may influence hepatic cholesterol synthesis (Raman et al. 2016). Interestingly, the acetic acid to propionic acid ratio of fermentations with CF, LPF, and HPF were significantly lower than those for the positive control inulin and negative controls after 24 h fermentation. The decrease in this SCFA ratio has been proposed as a possible indicator of the inhibition of cholesterol and fatty acid biosynthesis in the liver, leading to a decrease in lipid levels in blood (Delzenne and Kok 2001; Salazar et al. 2008). In contrast to the fractions enriched in polysaccharides, fermentation of the PF fraction which was enriched with phlorotannins resulted in very low levels of SCFA in comparison to other seaweed fractions, levels also below those of the negative and positive controls. Phlorotannins in brown seaweed appear to have some antibacterial activities (Dierick et al. 2010) which could explain the low SCFA production, although our microbiology analyses suggest that the inhibition of growth occurs for selected populations. Some phlorotannins were present in other fractions tested, including those enriched in polysaccharides, but this did not prevent the polysaccharides from inducing significant production of SCFA. The influence of dietary fibre and NNSP content might have a greater contribution to SCFA production than from the phlorotannin content. However, elimination of phlorotannins from the polysaccharide mixes in the future could potentially enhance their prebiotic effects.

Several studies have demonstrated the degradation of polysaccharides from brown seaweeds in the human gastrointestinal tract. Salyers et al. (1978) reported that species of Bacteroides can induce production of enzymes which degrade laminarin and alginate. Bacteroides distasonis, Bacteroides thetaiotaomicron, and Bacteroides group “0061” are able to break down laminarin to glucose (G1) and higher oligomers (G2-G6), while Bacteroides ovatus is involved in alginate degradation. Studies have also reported the use of enzymes produced by bacteria from marine environments to hydrolyse fucoidan into oligomers (Kusaykin et al. 2016). However, the results from some studies also indicate the potential of fucoidan degradation by gut bacteria. A study in pigs (Lynch et al. 2010) demonstrated that a fucoidan-supplemented diet increased Lactobacillus populations and SCFA production, and Shang et al. (2016) demonstrated that fucoidan increased the abundance of Lactobacillus and Ruminococcaceae, and decreased the number of Peptococcus, in mice. We have used Q-PCR to investigate the effects of the seaweed fractions on selected bacteria present in human stool. Relative to the negative controls, all seaweed fractions increased the number of Bacteroidetes and/or Firmicutes, which together comprise about 90% of the large intestinal microbiota and represent the majority of the gut bacterial phyla in humans (Qin et al. 2010). Of the seaweed fractions tested, LPF induced the greatest increase in Bifidobacterium and Lactobacillus, which are the most commonly recognised bacterial markers of prebiosis (Kleerebezem and Vaughan 2009; Bird et al. 2010), and significantly enhanced the growth of the butyric acid producing C. coccoides group (Louis and Flint 2009) compared to controls, as well as the key butyrate producing bacteria F. prausnitzii relative to the cellulose control. In addition, the Bacteroides-Prevotella population, which plays important roles in the hydrolysis and fermentation of dietary fibre (Balamurugan et al. 2010), increased over 24 h of fermentation with CF, LPF, and the positive control, but not HPF. This group of microbes and numerous others that we targeted were not detectable in the HPF fermentation, suggestive of components in the fraction interfering with many of the Q-PCR assays. Nevertheless, Bacteroidetes and Firmicutes were detectable in samples taken from the HPF fermentations. Increases in Bacteroidetes but not Firmicutes relative to the controls indicate that bacteria of the former phylum may be responsible for the increased SCFA induced by HPF. The relative proportions of Firmicutes and Bacteroidetes appears to have implications for human health (Gerritsen et al. 2011) with a low Firmicutes to Bacteroidetes ratio associated with a reduced risk of obesity or excessive body weight (Ley et al. 2005, 2006). In our study, the lowest Firmicutes to Bacteroidetes ratio was observed in response to HPF fermentation. We also examined changes in Enterococcus and E. coli, which are often linked to poor gut health outcomes. A decrease in the numbers of Enterococcus was observed with PF and CF supplemented fermentations while an increase in the numbers of E. coli was observed with most seaweed fractions compared to negative controls. The PF and CF fractions contained a higher phlorotannin content in comparison to other seaweed fractions, and these components may be able to slow down or partially inhibit pathogenic bacterial growth (Eom et al. 2012). This potential to influence the growth of gut pathogens could at least partly explain the observation that brown seaweed-derived phlorotannins could also prevent inflammatory diseases of the mammalian intestine (Bahar et al. 2016). Although E. coli are commonly known for their pathogenic potential, they can also have some benefits as a consequence of non-pathogenic strains outcompeting the pathogenic forms, as is thought to occur for the probiotic E. coli Nissle 1917 (Iannitti and Palmieri 2010; Gerritsen et al. 2011). Hence, the increase in E. coli in response to the seaweed fractions in our fermentations could conceivably be beneficial, although impacts on pathogenic strains would need to be tested specifically in future studies.

The composition of the four seaweed fractions tested demonstrated that they all contained variable proportions of potentially fermentable components, mainly fibres, sugars, and polyphenols. Although components of the seaweeds were shown to stimulate SCFA production and growth of potentially beneficial gut bacteria in an in vitro system, it is possible that such effects will not manifest themselves in vivo following consumption of these components. The digestibility of components will determine how much material reaches the large bowel and is available for fermentation. In this study, HPF that contained a high dietary fibre and NNSP content was not digestible by the enzymes responsible for breakdown of polysaccharides in the small intestine. This result was in agreement with polysaccharide compositions of this fraction inferred from sugar analysis. HPF contained high proportions of fucose, guluronic and mannuronic acid, and glucose which might imply the presence of fucoidan, alginate, and laminarin. The mixture and cross-linkages between these sulphated, branched polysaccharides and other cell wall components such as proteins and polyphenols contribute to the more complex structure of seaweed cell wall polysaccharides compared to terrestrial plants (Jeon et al. 2012). The resistance of these polysaccharides from brown seaweeds to digestion by enzymes in the upper gastrointestinal tract have been reported in many publications (Devillé et al. 2004; Zaporozhets et al. 2014; de Jesus Raposo et al. 2016). This contrasts with the other fractions tested which were highly digestible. Although CF and LPF were better than HPF at stimulating SCFA production and some microbial growth using our in vitro fermentation system, HPF may prove to be more effective in promoting SCFA production in vivo as its minimal digestion will enable far more to reach the colon. HPF is also likely to have other advantages. In particular, the high levels of fibre are expected to increase the stool bulk (and consequently dilute toxins) due to their capacity to hold water (Praznik et al. 2015).

In conclusion, the results presented in this study have shown that different seaweed fractions, and hence components derived from the brown seaweed E. radiata, also differ in their prebiotic potential as assessed using an in vitro fermentation system containing human stool. The HPF constituents show promise as prebiotics as they were resistant to digestion by enzymes responsible for polysaccharide digestion in the small intestine and were readily fermentable, stimulating the production of beneficial SCFA, including butyric acid. However, the impacts on gut microbe populations below the phylum level were not as clear as for other fractions and may be a result of the narrow range of targets examined. The LPF constituents also show potential, as they were the most potent at stimulating growth of the traditional markers of prebiosis Lactobacillus and Bifidobacterium, and stimulated the production of SCFA in our in vitro system. However, the high digestibility of LPF indicates that far less of these components would reach the large bowel in vivo, reducing its effectiveness as a prebiotic relative to HPF components. Further investigations, which are carried out in vivo, are required to substantiate the prebiotic potential of HPF and LPF fractions derived from E. radiata. The phlorotannin fraction of the seaweed was also shown to influence the gut microbiota populations as evidenced by inhibition of the growth of potentially pathogenic bacteria and inhibition of fermentation in vitro, and this may have beneficial uses in vivo.