Evaluation of the potential prebiotic effect of Himanthalia elongata, an Atlantic brown seaweed, in an in vitro model of the human distal colon

Food Research International 156 (2022) 111156 Available online 17 March 2022 0963-9969/© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Evaluation of the potential prebiotic effect of Himanthalia elongata, an Atlantic brown seaweed, in an in vitro model of the human distal colon Aroa Lopez-Santamarina, Alejandra Cardelle-Cobas * , Alicia del Carmen Mondragon , Laura Sinisterra-Loaiza , Jose Manuel Miranda , Alberto Cepeda Laboratorio de Higiene Inspecci´ on y Control de Alimentos, Departamento de Química Analítica, Nutrici´ on y Bromatología, Universidade de Santiago de Compostela, 27002 Lugo, Spain ARTICLE INFO Keywords: Gut microbiota In vitro Fermentation Fibre H. elongata Algae Digestion 16S rRNA ABSTRACT Until now, although different studies have shown the potential prebiotic effect of seaweed carbohydrates, no studies with the whole seaweeds have been carried out. In addition, the prebiotic effect throughput sequencing remains poorly investigated since most of the published works used qPCR or FISH to estimate bacterial changes. In this work, an in vitro model of the human distal colon was used to determine, for the first time, the potential prebiotic effect of a brown whole seaweed Himanthalia elongata. The whole seaweed was characterized in basis of its nutritional and mineral composition and submitted to the entire gastrointestinal digestion. The prebiotic effect was evaluated by the microbial modulation through 16S rRNA amplicon sequencing, qPCR and short-chain fatty acid analysis. The obtained results indicated that the colonic fraction of H. elongata was used selectively by the Bacteroides genus, more specifically by the specie Bacteoides ovatus, whereas inulin was used mainly by the Parabacteroides genus, being Parabacteroides distasonis the most abundant identified specie. Selective use of inulin by P. distasonis is, therefore, reported by the first time. qPCR analysis shown no significative differences in Bifidobacterium population and a decrease in Lactobacillus along the fermentation assays with both substrates. Regarding to the short-fatty acid production, maximal concentration, 56.11 ±20.48 mM, was achieved for H. elongata, at 24 h of fermentation whereas for inulin total acid production was 93.66 ±21.82 mM at 48 h of assay. The metabolic pathways associated with bacterial genera were not significantly different between the two tested substrates. Although more studies are necessary to elucidate the prebiotic character of H. elongata, the results presented in this work are promissory and could open new opportunities of research and application in the area of Nutrition and Food Chemistry. 1. Introduction Dietary fiber (DF) is substantially composed of non-digestible polysaccharides, whose consumption in adequate amounts has numerous benefits for human health (Cronin, Joyce, O’Toole, & OConnor, 2021). In this sense, an increase in DF consumption has been recommended persistently in Western societies since the 1970 s, with the aim to achieve an average daily intake of 25–35 g/day in adults (Stephen, Champ, Cloran, Fleith, van Lieshou, Mejborn et al., 2017). Nevertheless, the average consumption by people in Western countries has continuously remained below these recommendations (Stephen et al., 2017). Currently, most of the non-digestible polysaccharides included in the Western diet come from the cell walls of terrestrial plants (LopezSantamarina, Miranda, Mondragon, Lamas, Cardelle-Cobas et al., 2020). However, due to intense population growth worldwide in recent decades, freshwater, an essential commodity for agriculture, is becoming increasingly scarce (Cosgrove & Loucks, 2015). According to the Food and Agriculture Organization (FAO, 2017a,b), global agricultural productivity is declining at around 1% per year. Because of this decline, it is reasonable to expect that the demand for food from other resources will increase in the coming years. Among nutritional components that can modulate the gut microbiota (GM) composition and function, DF can play a central role since it may selectively promote the growth of certain bacteria, such as Lactobacillus or Bifidobacterium, and exert beneficial effects for the host’s health (Lopez-Santamarina et al., 2020). This phenomenon, when occurs, is * Corresponding author at: Laboratorio de Higiene Inspecci´ on y Control de Alimentos, Facultad de Veterinaria, pabell´ on 4 p.b., Campus Universitario, 27002 Lugo, Spain. E-mail address: [email protected] (A. Cardelle-Cobas). Contents lists available at ScienceDirect Food Research International journal homepage: www.elsevier.com/locate/foodres https://doi.org/10.1016/j.foodres.2022.111156 Received 17 January 2022; Received in revised form 13 March 2022; Accepted 15 March 2022 Food Research International 156 (2022) 111156 2 known as the prebiotic effect (Salminen, Collado, Endo, Hill, Lebeer, Quigley et al., 2021). However, it is important to highlight that the current definition of prebiotic does not refer to any bacterial genus, since different genus can exert beneficial effects to the host and even, different genus can exert the same functionality. Thus, prebiotics are “selectively fermented ingredients that allow specific changes, both in the composition and/or activity in the GM that confer benefits upon host wellbeing and health” (Gibson, Scott, Rastall, Tuohy, Hotchkiss, Dubert-Ferrandon et al., 2010). Currently different bacterial genera such Enterococcus, Streptococcus, Pediococcus, Leuconostoc, Bacillus or specific strains of Escherichia coli are considered beneficial and even are used as probiotics in different commercial preparations. An adequate DF intake provides the host with substrates for fermentation by gut microbes that possess the necessary enzymatic machinery to degrade complex carbohydrates (Cronin et al., 2021) and therefore, depending on the DF structure different bacteria can be stimulated. Regarding seaweed composition, DF represents up to 76% expressed on a dry matter basis, in some seaweed species (Charoensiddhi, Conlon, Vuaran, Franco, & Zhang, 2017). In addition, seaweeds contain phenolic compounds and proteins that has been proposed as potential beneficial ingredients for human health (Charoensiddhi et al., 2017). Among all the potential benefits of these ingredients, the prebiotic activity has been one of the most studied and brown seaweeds the group seaweeds more employed (Lopez-Santamarina et al., 2020). This potential prebiotic effect has been evaluated mainly in in vitro assays simulating the human large intestine using faeces as inoculum and different carbohydrate fractions, without a previous step of digestion in the upper gastrointestinal tract (Charoensiddhi et al., 2017; Chen et al., 2018; Fu et al., 2018; Strain et al., 2020). In all these assays an increase in SCFAs was observed, however, regarding the bacterial population, the results are contradictory, in some cases Bifidobacterium and Lactobacillus increased whereas in others, decreased, depending on the carbohydrate extract tested. However, all these studies indicate the potentiality of the carbohydrate seaweeds as prebiotics. Several in vivo studies with mice has also been carried out (Kim, Yu, Kim, Choi, Lee, Hong et al., 2016; Nguyen, Kim, Guevara, Lee, Kim et al., 2016). From these studies, using mainly extracts of the brown seaweed Ascophyllum nodosum, it is not possible to extract clear conclusions because the results are contradictory. Some authors indicate no microbial changes after the extract consumption, whereas other reported an increase in the microbiota diversity and an increase in Bacteroides and Parabacteroides species. It is important to indicate that the conditions of the studies, notably vary from one to another in terms of extracts composition, dosage and time of exposure. These interesting previous results, positive or not, together with the fact that seaweed consumption has increased in Spain and the European Union makes the seaweeds an interesting product to research about its functionality. Thus, in the present work, the evaluation of the potential prebiotic effect of a whole brown oceanic seaweed, Himanthalia elongata, was carried out, for the first time, in an in vitro model of the human colon. This specie, also known as sea spaghetti, was selected because it is predominant in the Rias Baixas coast of Galicia (northwest Spain), which is the Spanish region with the highest seaweed production (Des, Martinez, de Castro, Viejo, Sousa, Gomez-Gesteira et al., 2020), only in the year 2021a total of 38,149 kg of brown seaweed were collected. 2. Materials and methods 2.1. Seaweed H. elongata was obtained, dehydrated, from Portomui˜ nos (Cerceda, A Coru˜ na, Spain). Two hundred fifty grams of seaweed were sampled, crushed, and freeze-dried prior to testing. The sample was stored at room temperature for further analysis. 2.2. Proximate and mineral composition of H. elongata The nutritional composition of H. elongata was determined prior to and after in vitro digestion. The nutritional analysis of the seaweed was carried out following official methodologies established by the Association of Official Analytical Chemists (AOAC, 2002). The moisture content (g/100 g of product) was determined by drying in a porcelain capsule at 100–105 ◦C in a muffle furnace (SNOL 8.2/1100-1, Umega Group AB, Lithuania). The protein content was determined by measuring the nitrogen content (g/100 g of dry matter) by the Dumas method. The fat content (g/100 g of dry matter) was determined by the Soxhlet extraction method using petroleum ether. The DF content (g/ 100 g of dry matter) was determined by using an enzymatic–gravimetric method via the Megazyme® total dietary fiber assay kit (Megazyme, Wicklow, Ireland). The ash content (g/100 g of dry matter) was determined by muffle furnace (SNOL 8.2/1100-1, Umega Group AB, Lithuania) incineration at 500 ◦C The sodium content (mg/100 g of dry matter) was determined by atomic absorption spectrometry (AES, Agilent 5900, Agilent Technologies, Santa Clara, California, USA). The carbohydrate (g/100 g of dry matter) and caloric contents (kcal/100 g of product) were determined through calculations. Regarding minerals, seaweed was analyzed for 75 As, 43 Ca, 111 Cd, 63 Cu, 56 Fe, 127 I, 208 Pb, and 66 Zn (mg/kg) by inductively coupled plasmamass spectrometry (ICP-MS, Agilent 7700x, Agilent Technologies, Santa Clara, CA, USA). Sample blanks were prepared in the laboratory in a similar manner to the seaweed samples. Proximate and mineral composition analyses were performed in triplicate both before and after digestion. 2.3. In vitro simulation of oral, gastric, and small intestinal digestion In vitro oral, gastric, and small intestinal digestion was carried out in accordance with the INFOGEST protocol (Brodkorb, Egger, Alminger, Alvito, & Assunçao, 2019). All chemicals were purchased from SigmaAldrich (Sant Louis, MO, USA). Calculations for the different fluids and enzymes using the protocol were made for 10 g of sample. At the end of the in vitro digestion process, the beakers with the digested seaweed solution were cooled in an ice-bath to stop the enzymatic reactions. After that, small intestinal absorption was simulated by dialysis (molecular weight cutoff of 1000 Da, Spectra/Por®, Waltham, MA, USA) against distilled water at 4 ◦C for 2 days with agitation.. This process allows separate molecules in solution by the difference in their rates of diffusion through the semipermeable membrane. Then, the membrane content was frozen at −18 ◦C for subsequent freeze-drying. The freeze-drying process was performed by using a Vacuum Freeze Drier (Labconco TM 77560-LYPM-LOCK6), under a vacuum pressure ≤ 140 ×10 -3 Mbar and a condenser temperature of −46 ◦C. The entire in vitro digestion process, including dialysis, was performed in triplicate. 2.4. Volunteers and preparation of stool samples Stool samples were obtained from three healthy human volunteers (one male and two females, 32–50 years old), participating in a trial authorized by the Galician Bioethics Committee (trial 270/2018). These volunteers did not ingest antibiotics or pharmaceutical preparations of pre/pro/postbiotics in the 6 months prior to sample collection, and they had no gastrointestinal disorder. All of them signed an informed consent document in which they were informed how their samples would be used, about compliance of the study with the Declaration of Helsinki, and about the Spanish law about personal data protection. Stool samples (between 10 and 30 g each) were collected by volunteers in sterile containers and given to the laboratory within 2 h of their collection. Once received, the stool samples were diluted 1:10 with phosphate-buffered saline (PBS; 0.1 M, pH 7.0) and then homogenized in a sterile bag using a paddle homogenizer (MIX2, AES, Combourg, France) for 5 min. The diluted feces were stored in sterile jars and frozen A. Lopez-Santamarina et al. Food Research International 156 (2022) 111156 3 at −20 ◦C until use. 2.5. In vitro human colonic simulation The in vitro human colonic simulation was carried out according to Cardelle-Cobas, Olano, Corzo, Villamiel, Collins, Kolida et al. (2012). Briefly, a sterilized fermentation vessel with a capacity of 300 mL containing a basal culture medium without any source of carbon was used to simulate human distal colonic fermentation. Three experiments were carried out simultaneously: one for H. elongata, one for inulin from chicory (Sigma-Aldrich, St Louis, USA, product number I2255), which was selected as positive control, and one without a carbon source, without substrate (negative control). Different parameters were adjusted to simulate the conditions of the human distal colon. Thus, anaerobiosis was achieved by using a continuous supply of O 2 -free N 2 (Nippon Gases, Madrid, Spain) through a 0.2 μ m polytetrafluorethylene filter (Sartorius Stedim Biotech GmbH, Gottingen, Germany). A thermostatic bath (Pharmacia Biotech, Amsterdam, Netherlands) was used to recirculate water at 37 ◦C throughout the vessel’s water jackets to simulate the human internal body temperature. The pH of the colon was set at 6.8 because that is the pH of the colon in a situation of eubiosis. This pH was controlled by a pH regulator (Hanna Instruments, Eibar, Spain), adding 1 M NaOH or HCl as appropriate. Each fermentation vessel was filled, under aseptic conditions, with 200 mL of autoclaved nutrient basal medium, prepared according to Cardelle-Cobas et al. (2012). All included chemicals were purchased from Sigma-Aldrich, Merck (Darmstadt, Germany), or Panreac (Barcelona, Spain). Then, the medium was adjusted to pH 6.8 and left overnight with a stream of O 2 -free N 2 with stirring. Next, substrates were sterilized by boiling a recipient containing them during 10–15 min (H elongata or inulin) and dissolved in 52 mL of the same autoclaved medium and added to the vessels at a final concentration of 1% (w/v). This concentration was used based on previous studies about the evaluation of the prebiotic effect in vitro, where the substrate to be tested were used at 1-2% (Bajury, Rawi, Sazali, Abdullah, & Sardini, 2017; CardelleCobas et al., 2012). Finally, the vessels were inoculated with 10% (v/ v) (28 mL) of the previously prepared diluted feces. Aliquot samples (3 mL) were removed from each vessel after 0, 10, 24, and 48 h of incubation for real-time polymerase chain reaction (qPCR) analysis, 16S ribosomal RNA (rRNA) amplicon sequencing, and short-chain fatty acid (SCFA) analysis. 2.6. Bacterial DNA extraction from fermentation samples Bacterial DNA was extracted from the fermentation samples by using the DNA Realpure Spin Food-Stool Kit® (Real, Durviz S.L, Valencia, Spain) following the instructions provided by the manufacturer for fecal samples. The pellet obtained after centrifugation (Centrifuge 5415D, Eppendorf, Hamburg; Germany; 6100 g) of 1.2 mL of sample (fermentation vessels) was recovered and used for DNA extraction. The obtained DNA was quantified by using a Qubit™4 fluorometer (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA), and the DNA HS Assay Kit (Invitrogen, Thermo Fisher Scientific, Eugene, OR, USA). DNA samples were stored at −20 ◦C until further analysis. 2.7. Bacterial quantification by qPCR qPCR assays were carried to characterize fecal bacteria using phylumand species-specific primers for total bacteria, Firmicutes, Bacteroidetes, Actinobacteria, Proteobacteria, Lactobacillus, and Bifidobacterium, previously described (Murri, Leiva, Gomez-Zumaquero, Tinahones, Cardona, Soriguer et al., 2013). Briefly, qPCR experiments were performed in a QuantStudio 12 K Flex (Applied Biosystems, Life Technologies Holding, Singapore, Singapore) equipment using the fast SYBR TM green master mix (Applied Biosystems, Vilnius, Lithuania). All PCR tests were carried out in triplicate, with a final volume of 10 μ L containing 1 μ L of each sample DNA, primers (0.4 μ L) added at a concentration of 200 nM (for each primer), 5 μ L of fast SYBR TM green master mix (Applied Biosystems), and 3.2 µL of molecular biology grade water. The thermal cycling conditions used were as follows: an initial DNA denaturation step at 95 ◦C for 10 min, followed by 45 cycles of denaturation at 95 ◦C for 10 s, primer annealing at an optimal temperature for 20 s, and extension at 72 ◦C for 15 s. Finally, melt curve analysis was performed by slowly cooling the reactions from 95 ◦C to 60 ◦C (0.05 ◦C per cycle) with simultaneous measurement of the SYBR green signal intensity. Melting-pointdetermination analysis allowed confirmation of the specificity of the amplification products. The bacterial concentration (copies/mL) from each sample was calculated by comparing the threshold cycle (Ct) values obtained from the standard curves with the Quant Studio 12 K Flex Software (Applied Biosystems). Standard curves were constructed for each experiment by using 10-fold serial dilutions of bacterial genomic DNA (of known concentration) from pure cultures, corresponding to 10 1 to 10 10 copies/ mL of fermentation media. The pure cultures used to construct the standard curves were obtained from different collections of type cultures—from the Spanish Collection (CECT), the Belgian Co-ordinated Collections of Microorganims (LMG), and the German Collection of Microorganisms and cell Cultures GmbH (DSM)—as follows: Enterobacter cloacae CECT 194, Clostridium perfringens CECT 376, Bifidobacterium longum CECT 4503, Bacteroides vulgatus LMG 17767, and Lactobacillus reuteri DSM 20016. Each bacterial strain was grown in its required culture media and growth conditions and then, DNA was extracted and diluted to construct the standard curves. The final data are expressed as an average of the duplicate values obtained in the analyses. The efficiency of the reaction for all pairs of probes was determined by using the slope of the calibration curve obtained for each of the bacterial groups analyzed, namely E =10^(-1/ slope). For the primer pairs used in this study, the efficiency ranged from 95% (E =1.90) to 104% (E =2.07), with slopes in the range of −3.59 to −3.16. 2.8. 16S rRNA amplicon sequencing For 16S rRNA amplicon sequencing, 12 μ L of DNA extracted from each sample was used to construct the libraries and the Ion GeneStudio TM S5 System (Life Technologies, Carlsbad, CA, USA) was used. For this purpose, the 16S hypervariable regions were amplified with two sets of primers, v2-4–8 and v3-6,7–9, and the libraries were prepared by using the Ion 16S TM Metagenomics Kit (Life Technologies) and the Ion Xpress TM Plus Fragment Library Kit (Life Technologies). Libraries containing equal amounts of PCR products pooled with a barcode were prepared by using the Ion Xpress TM Barcode Adapters Kit (Life Technologies). Then, these libraries were quantified by using the Ion Universal Library Quantitation Kit (Life Technologies). Next, 10 pM of each library was pooled and loaded on an Ion OneTouch™ 2 System (Life Technologies), which automatically performs template preparation and enrichment. Template-positive ion sphere particles were enriched with Dynabeads™ MyOne™ Streptavidin C1 magnetic beads (Invitrogen, Carlsbad, CA, USA) by using an Ion One Touch ES instrument. Finally, an Ion 520 TM chip (Life Technologies) was loaded with the samples on an Ion GeneStudio TM S5 System sequencer using the Ion 520™ & Ion 530™ Loading Reagents supplied in the OT2-Kit (Life Technologies). 2.9. Short-chain fatty acids analysis SCFA analysis was carried out following the protocol of Gull´ on, Gull´ on, Sanz, Alonso, and Paraj´ o (2011). One mL of fermentation samples obtained after 0, 10, 24, and 48 h were centrifuged for 7 min at 6100 g. The supernatants were removed and filtered through 0.2 µm cellulose A. Lopez-Santamarina et al. Food Research International 156 (2022) 111156 4 acetate membranes. Then, 20 μ L of sample was injected in an Agilent 1200 series HPLC instrument equipped instrument with a refractive index detector (Agilent, Waldbronn, Germany). Separations were carried out on an Aminex HPX-87H column (Bio-Rad, Hercules, California, USA). The elution system consisted of sulfuric acid (0.003 M) operating isocratically with a flow rate of 0.6 mL/min. Samples were run at 50 ◦C. Standards of organic acids (lactic, formic, acetic, butyric, propionic, isobutyric, valeric, isovaleric and succinic acids) were obtained from Sigma (Poole, Dorset, UK). Peaks were identified by comparison with the retention times of the standards and quantified by regression formula obtained with the standards. 2.10. Statistical and bioinformatic analysis A two-way ANOVA was used to determine significative differences for time and substrates as the two covariates in the general linear model using Tukey’s analysis. For significant differences (p <0.05) a one-way ANOVA was conducted for each substrate comparing 0, 5, 10 and 24 h. similarly, each substrate was compared if two-way ANOVA results were significant (p <0.05). SPSS® v.27 for Windows (SPSS Inc., Chicago, IL, USA) was used for these analyses. For the analysis of 16S rRNA amplicon sequencing, the raw sequencing reads were obtained from the Torrent Suite software (v.5.12.2.) as BAM files, which were converted to fastq files with BEDTools wrapped into the Public Galaxy Server (https://usegalaxy.eu/, v. 21.05) (Afgan, Baker, Batut, van den Beek, Bouvier, Cech et al., 2018) by Gruening (2014) Galaxy wrapper (https://github. com/bgruening/galaxytools). The fastq files were processed with QIIME 2 software v. 2021.8 (Bolyen, Rideout, Dillon, Bokulich, Abnet, Al-Ghalith et al., 2019). To produce amplicon sequence variants (ASVs), the DADA2 method was used for quality filtration (Q score >30), trimming, denoising, and dereplication. Samples with features (taxa) with a total abundance (summed across all samples) of <10 were removed. Then, ASVs were aligned with mafft and used to construct a phylogenetic tree with fasttree. α - and βdiversity metrics were estimated by using q2-diversity core-metrics-phylogenetic after samples were rarefied to a sequencing depth of 33,000 reads. Taxonomy was assigned to ASVs by using the q2-feature-classifier classify-sklearn naïve Bayes taxonomy classifier against the Greengenes 13_8 99% operational taxonomic unit (OTU) reference sequences. The PICRUSt online Galaxy version on the Huttenhower Lab (v1.0.0) server was used to predict the metagenome functional content from marker gene surveys and full genomes. Functional metagenomes were categorized based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database at hierarchy level 3. STAMP software (v 2.1.3) for the “Statistical Analysis of Taxonomic and Functional Profiles” (Parks, Tyson, Hugenholtz, & Beiko, 2014) was used for determinate differences in metabolic functions. Welch’s t-tests with Bonferroni correction were used to determine significant differences in the relative abundance of 20 selected KEGG pathways (level 3). In addition, differences in the relative abundance of the most common species were determined by using a G-test (with Yates’ correction) + Fisher’s exact test with Bonferroni correction. 3. Results and discussion 3.1. Proximate and mineral composition of H. elongata The proximate composition of H. elongata can be found in Table 1. Although some factors such as seasonality can alter the nutritional composition of seaweeds, in global terms, the results obtained were close to those previously reported by Fern´ andez-Segovia, Lerma-García, Fuentes, & Barat, 2018 for the same seaweed species collected at the same location (protein: 7.50% vs. 6.8%; DF: 36% vs. 39%; fat: <0.5% vs. 0.1%; ash: 33% vs. 30%). After oral, gastric, and small intestinal digestion, including dialysis, the weight of the seaweed was reduced by 51.82 ±1.58% (w/w), which would be equivalent in an in vivo assay to the loss due to food hydrolysis and absorption (Asensio-Grau, Calvo-Lerma, Heredia, & Andr´ es, 2021). Namely, only around a 48% of the seaweed would reach the colon to exert its effect on GM. After digestion, the seaweed exhibited a different nutrient composition compared with the raw seaweed. It is well known that during digestion, several key factors contribute to the progressive alteration of the food matrix, including the enzymes involved throughout the gastrointestinal tract, as well as the acidic conditions at each stage (Asensio-Grau et al., 2021). Because of these actions, the fat, carbohydrate, and ash contents significantly decreased during this digestion process, whereas the protein, DF, and caloric contents significantly increased. One of the reasons because the protein content increased in relation to the other macronutrients is due to the formation of organic complexes in the seaweed, for example with iodine, as it has been indicated in previous works (Romarí-Hortas, Bianga, MoredaPi˜ neiro, Bermejo-Barrera, & Szpunar, 2014; Domínguez-Gonz´ alez, Chiocchetti, Herbello-Hermelo, V´ elez, Devesa, Bermejo-Barrera et al., 2017). From these studies, it is possible to indicate that in seaweeds as wakame an important proportion of iodine is under the form of organic iodine complexes (Shah, Wuilloud, Kannamkumarath, & Caruso, 2005), which can reach molecular weights greater than 10 kDa what can lead, in consequence, to a low dialyzability of iodine and protein (DomínguezGonz´ alez et al., 2017). Thus, from an initial H. elongata composition in which the two main macronutrients are fiber and carbohydrates, after digestion, protein and fiber are the main components and, therefore, act at the colonic level on GM. Table 1 shows the mineral content in H. elongata before and after the digestion process. With respect to the maximum residue limits for minerals, in the European Union (EU) only a specific limit for iodine in seaweed was established (20 mg I/kg dry seaweed, Commission Recommendation [EU] 2018/464), especially for populations with endemic iodine deficiency. Table 1 shows that the iodine content of dried H. elongata is below this limit (4.49 mg/100 g). Considering that in 2006 the EU established an upper limit for iodine intake of 600 µg/day for adults, consumption of 15 g of dried H. elongata would reach this amount. Iodine is essential for human health; it may be therapeutic in the case of nutritional deficiencies, but high values due to excessive consumption could cause adverse health effects. Hence, these kind of seaweed products must be consumed with caution. In this work, the Table 1 Comparison of the nutritional composition (g/100 g) and mineral content (mg/ 100 g) of sea spaghetti (Himanthalia elongata) raw and after in vitro upper intestinal digestion. The caloric content is expressed in kcal/100 g. The results are expressed as the mean ±standard deviation. Different letters indicate significant differences (p <0.05). Nutritional composition Raw H. elongata H. elongata after upper in vitro digestion Fat 1.00 a ±0.20 0.64 b ±0.12 Protein 7.50 b ±1.43 12.26 a ±2.44 Carbohydrates 15.00 a ±2.56 9.87 b ±2.17 Sugars <0.5 <0.5 Dietary fiber 36.00 b ±3.71 63.6 a ±4.31 Ash 33.20 a ±3.22 16.63 b ±2.89 Caloric content 170.00 b ±10.22 221.48 a ±15.67 Minerals Ca 1,512.10 a ±332.80 1,2980 a ±185.72 Fe 4.30 a ±3.76 11.73 b ±4.58 Cu 0.21 a ±0.19 0.22 a ±0.03 Zn 4.37 a ±3.34 5.94 a ±0.77 As 0.60 a ±0.25 0.45 a ±0.15 Cd 0.03 a ±0.01 0.02 a ±0.02 Pb 0.02 a ±0.02 0.05 a ±0.06 I 4.49 a ±6.77 194.67 b ±9.41 A. Lopez-Santamarina et al. Food Research International 156 (2022) 111156 5 objective was to show that small amounts of these products (specifically H. elongata) could exert a beneficial effect on human health by modulating GM and acting as a potential prebiotic. Most prebiotics for the gut require an oral dose of around 3 g per day or more to confer a benefit. Typically, around 5 g is the target for fructooligosaccharides (FOS) and galactooligosaccharides (GOS) in the daily diet, and this includes plant sources of prebiotics (Salminen, et al., 2021). Consumption of 5 g of H. elongata would be equivalent to consuming 200 μ g of I, not exceeding the established limits. Arsenic in biological matrices exists either in organic forms (e.g., arsenobetaine and arsenosugars) or as inorganic arsenic (iAs) (Edmonds & Francesconi, 2003). While organoarsenic forms are considered to be non-toxic or of low toxicity, iAs is regarded as the most toxic form of As for living organisms. Thus, the presence of As in seaweeds has safety implications for their use as food. However, regulations on As in food are currently limited in the EU, and no maximum allowed levels of As (either total As or iAs) in vegetables or food supplements exist (Petursdottir, Sloth, & Feldmann, 2015). Although As speciation in fish and seafood has attracted a lot of attention because of the high total As levels, these commodities are not generally important contributors to iAs dietary exposure, except for shellfish and seaweed in the regions where consumption levels of these items are appreciable (Petursdottir, Sloth, & Feldmann, 2015). Maximum limits for other heavy metals have been established by European Commission regulation (EC) 1881/2006, however, none are specific for seaweeds. Regarding the tolerable upper intake levels (UL) for adults, our data are far below: Ca (UL: 2500 mg/day), Fe (no UL), Cu (5 mg/day), Zn (25 mg/day) (European Food Safety Authority, EFSA, 2006), Pb (25 μ g/kg body weight [bw]/week) (EFSA, 2010), and Cd (2.5 mg/kg bw/weekly) (EFSA, 2011). Compared with previous work, the Ca composition in the current work (1.5 g/100 g) is like the values obtained by Fern´ andez-Segovia et al. (2018) for the same seaweed. Regarding Cd and Pb, the obtained concentrations are lower than those found by Filippini, Baldisserotto, Menotta, Fedrizzi, Rubini, Giglioti et al. (2021) (0.03 vs. 0.07 mg/kg Cd; 0.02 vs. 0.06 mg/kg Pb) for the same seaweed in Italy. One of the main risks that seaweed consumption can confer to humans is the potential excess of heavy metals and iodine. Elements such as Fe, Ca, Cu, and Zn are required for growth, enzymatic reactions, and metabolic activities of marine organisms (Velusamy, Satheesh Kumar, Ram, & Chinnadurai, 2014; Farias et al., 2021). However, they can also be toxic to humans at elevated concentrations, especially when seaweeds grow in polluted waters (Sudharsan, Seedevi, Ramasamy, Subhapradha, Vaira, & Shanmugam, 2012). High concentrations of iodine have been found in several countries for certain seaweed species (Charoensiddhi et al., 2017). Other elements such as As, Hg, Cd, and Pb can be toxic even at low concentrations, causing cellular damage, reduced reproduction and growth rates, or even death (Velusamy et al., 2014). Not all effects of heavy metals are harmful, because some minerals (such as Zn or Cu) are essential for many enzymes that promote maintenance, formation, and homeostasis of body tissues and perform metabolic activities (Farias et al., 2021). Moreover, GM requires certain heavy metals such as Fe, Cu, or Zn for their growth and metabolism (Sizentsov, Sizentsov, Kvan, Salnikova, & Salnikova, 2019). There were no significant changes in the mineral content of H. elongata before and after digestion in the upper intestinal tract, except for iodine and Fe, which are concentrated after digestion and dialysis. It is important to note that some studies have shown the limitations of dialysis membranes in the study of the bioavailability of elements (Domínguez-Gonz´ alez et al., 2017). On the one hand, the simulated gastrointestinal digestion may cause the breakdown of iodine complexes in macromolecules present in the seaweed matrix, increasing the soluble iodine fraction (Domínguez-Gonz´ alez et al., 2017). On the other hand, iodine complexes cannot cross the dialysis membranes when the pore is too small, a factor relevant to the present work given that the dialysis membrane has a 1 kDa cutoff. A previous study carried out investigating Table 2 Bacterial population (log 10 DNA copies/mL) in the in vitro colon model at 0, 10, 24, and 48 h of fermentation. The results are expressed as the mean ±standard error (n =6). nd, not detected. A two-way ANOVA using general lineal model, performed with time and substrates as two covariates for each bacterial group, indicated significant differences between both time and substrates, p <0.05. Then, a one-way ANOVA with Tukey’s test was performed to determine a significant increase/decrease of bacterial populations with time for the same substrate (different lowercase letters indicate significant differences), also a one-way ANOVA analysis was carried out to compare substrates at 10, 24 and 48 h (indicated with capital letters). H. elongata Inulin Negative control 0 h 10 h 24 h 48 h 10 h 24 h 48 h 10 h 24 h 48 h All bacteria 9.45 a ±0.18 10.08 a,A ±0.01 10.05 a,A ±0.01 11.78 b,A ±0.01 9.74 a,B ±0.02 8.39 b,B ±0.01 11.79 c,A ±0.01 10.66 b,C ±0.01 10.69 b,A ±0.03 10.62 b,B ±0.01 Firmicutes 6.29 a ±0.18 6.66 a,b,A ±0.01 7.80 b,c,A ±0.02 7.90 c,A ±0.01 6.46 a,B ±0.03 5.57 a,B ±0.02 8.02 b,A ±0.06 7.21 a, b ,C ±0.01 7.24 A ±0.01 7.12 a,b,B ±0.01 Lactobacillus 1.38 a ±0.03 nd nd nd 0.91 b,A ±0.01 0.26 c,A ±0.03 0.23 c,A ±0.02 0.46 b,B ±0.01 0.23 c,A ±0.01 0.01 c,B ±0.01 Actinobacteria 6.11 a ±0.34 5.87 a,A ±0.02 6.13 a,B,C ±0.08 5.77 a,A ±0.04 6.11 a,A ±0.14 5.65 a,B ±0.15 6.75 a,A ±0.01 7.37 a,B ±0.01 6.98 a,C ±0.02 6.65 a,B ±0.02 Bifidobacterium 5.11 a ±0.30 4.09 a,A ±0.14 5.94 aA ±0.21 4.94 a,A ±0.08 5.52 a,B ±0.01 5.22 a,A ±0.01 5.51 a,B ±0.04 4.66 a,A ±0.01 4.37ª ,B ±0.02 4.05 a,C ±0.01 Bacteroidetes 6.92 a ±0.12 4.83 b,A ±0.01 7.87 c,A ±0.02 7.41 a,c,A ±0.02 5.15 b,A ±0.05 5.58 b,B ±0.04 8.38 c,A ±0.01 6.03 b,B ±0.09 6.00 b,C ±0.01 5.87 b,B ±0.18 Proteobacteria 5.44 a ±0.28 8.62 b,A ±0.01 7.90 b,A ±0.02 7.60 b,A ±0.45 8.63 b,c,A ±0.02 7.41 b,B ±0.04 9.36 c,A ±0.02 9.77 b,B ±0.02 9.84 b,C ±0.06 9.57 b,A ±0.02 A. Lopez-Santamarina et al. Food Research International 156 (2022) 111156 6 the intestinal bioavailability and bioavailability of wakame seaweed confirmed this fact (Domínguez-Gonz´ alez et al., 2017). There is a similar situation for Fe: this metal is part of high-molecular-weight complexes such as ferritin and cannot cross the dialysis membranes, although humans absorb iron at the intestinal level (Fairweather-Tait et al., 2005). 3.2. qPCR analysis Quantification of the main phyla as well as for Bifidobacterium and Lactobacillus species was carried out for all samples by using qPCR. The obtained results are shown in Table 2. There was an increase in total bacteria over time, with the maximum after 48 h of fermentation for all cases. Other works investigating the effects of total bacteria found a maximum after 24 h of fermentation (V´ azquez-Rodríguez, Santos-Zea, Heredia-Olea, Acevedo-Pacheco, Santacruz, Guti´ errez-Uribe et al., 2021). Proteobacteria was the phylum with the greatest increase. Comparison of the inulin and H. elongata assays after 24 and 48 h did not show significant changes. Firmicutes also increased with time. For H. elongata, it reached the maximum values after 24 and 48 h. For inulin, however, there was only a significant change after 48 h compared with 0 h. A slight increase was observed for the control. When comparing both substrates, significant changes were found at 24 h, being the higher value for H. elongata. For Bacteroidetes, there was a slight decrease during the first 10 h of assay may be due to an establishment or adaptation to the media of the bacteria present in the fecal inoculum. This decrease occurs in all cases, control, seaweed and inulin. After this 10 h, a significant increase in Bacteroidetes occurred at 24 and 48 h for the assay with H. elongata compared with 0 and 10 h. Similar results were found for the case of Bacteroides genus using different polysaccharides from seaweeds (Seong, Bae, Seo, Kim, Kim, & Han, 2019). For the assay with inulin, the maximum (and significant) value was observed at 48 h. On the contrary, there was a significant decrease in Bacteroidetes for the control assay for all fermentation times. Comparing the fermentation assays for inulin and H. elongata after 24 h, there was a significantly higher value for H. elongata after 24 h and a significantly higher value for inulin after 48 h. This could indicate a faster consumption of carbohydrates in the seaweed compared with inulin. There was a significant decrease in Lactobacillus, whereas there were no significant changes for Bifidobacterium. Other authors, contrariwise, also found significant increase in Bifidobacterium counts after seaweeds fermentation, such as Ecklonia radiata (Charoensiddhi et al., 2017) or different seaweed polysaccharides extracts (Seong et al., 2019). 3.3. Amplicon 16rRNA sequencing The GM composition was evaluated at the relative level (16S rRNA amplicon sequencing; Fig. 1) and the absolute level (quantification of the main phylum and Bifidobacterium and Lactobacillus species by qPCR, Table 3) before (time 0) and at various times during the fermentation assays with the substrates (H. elongata and inulin). The relative frequency (Fig. 1) at the phylum level showed that both inulin and H. elongata modified the GM composition compared with the control. Whereas at time 0 Firmicutes and Bacteroidetes were the predominant phyla (67% for Firmicutes and 29% for Bacteroidetes), after 10 h of fermentation, there was an increase in the phylum Proteobacteria for inulin (75%) and H. elongata (54%). This increase was also observed in the control (no source of carbon added) but at a lower percentage (47%). There was a decrease in Firmicutes compared with time 0 for both substrates; it was more pronounced for H. elongata, where only 10% of the total bacteria belonged to Firmicutes phylum. After 24 and 48 h of fermentation with H. elongata, the Bacteroidetes phylum increased notably, while both Firmicutes and Proteobacteria phyla decreased. For inulin, after 24 h of fermentation, the Bacteroidetes phylum increased with respect to 10 h of incubation (19% vs. 4%), but the major change occurred after 48 h, where this phylum represented 62% of the total bacteria. Of note, Bacteroidetes contains Bacteroides species with a wide range of glycoside hydrolases and carbohydrate metabolic pathways (Mahowald, Hamilton, Mackey, Moore, Baker, Scanza et al., 2019). The results obtained at time 0 are in accordance with the literature: the human GM is mostly composed of Firmicutes and Bacteroidetes, which represent more than 90% of the total community (Lopez-Santamarina et al., 2020). The changes caused by the substrates are also in Fig. 1. Relative abundance of different bacterial phyla and genera. Bacterial composition (relative abundance, %) determined using 16S rRNA amplicon sequencing at the phylum (a) and genus (b) levels. The x axis shows the different substrates evaluated at the different assay times (10, 24, and 48 h). 0 h indicates the bacterial composition before substrate addition. Due to the large number of reported families, only the top 15 most abundant genus are included in the legend. INU, inulin; CONT, negative control (no addition of substrate); SW, seaweed (H. elongata). A. Lopez-Santamarina et al. Food Research International 156 (2022) 111156 7 agreement with recent similar studies. V´ azquez-Rodríguez, et al., (2021) studied the effect of polysaccharide fractions of a brown seaweed (Silvetia compressa) on human GM by using a different in vitro colonic model. They found that the relative abundance of Proteobacteria increased in all treatments with the fermentation time for the control, inulin, and polysaccharide fractions, reaching similar percentages to those found in the present work. Those authors found an increase in Bacteroidetes phylum after 2 h of incubation with a later decrease with time. In this work, the maximum relative abundance of Bacteroidetes occurred at later times (24 and 48 h). This could be due to the differences in the substrates including inulin, which can possess different degrees of polymerization, or to the assay conditions, because V´ azquez-Rodríguez et al. (2021) used an in vitro colonic model without pH control, among other differences. Strain et al. (2020), in a similar study testing polysaccharide extracts of Laminaria digitata, also found an increase in the Bacteroidetes phylum, whereas contrariwise to those obtained in the current work, no differences in Proteobacteria were observed. Fu et al. (2018), who also evaluated a new polysaccharide extracted from another brown seaweed (Sargassum thunbergii), did not observe changes in the Proteobacteria phylum, whereas they noted a significant reduction in Firmicutes and an increase in Bacteroidetes. As commented before, these differences could be attributed to the differences in the in vitro colonic model and the structure of the substrates. These authors did not use any other prebiotic carbohydrate as positive control and did not control pH. Considering the genera of the Bacteroidetes phylum, Bacteroides and Parabacteroides were the most abundant. For H. elongata, the Bacteroides genus showed a higher relative abundance, whereas for inulin, Parabacteroides was the predominant genus. Previous studies have documented a few examples in which symbiotic bacteria belonging to Bacteroides, a dominant genus in the human GM, possess genes for degrading seaweed-derived porphyran agarose, alginate, and laminarin (Hemsworth, Thompson, Stepper, Sobala, Coyle, Larsbrink et al., 2016; Dejean, Tamura, Cabrera, Jain, Pereira et al., 2020). Thirty-seven different bacterial species were identified. The results obtained from the statistical analysis of the samples obtained after 24 and 48 h of fermentation showed significant differences between inulin and H. elongata for 14 species (Fig. 2a) after 24 h and 16 species after 48 h (Fig. 2b). After 24 h of fermentation, in the assay with H. elongata the predominant species belong to the genus Bacteroides, specifically Bacteroides ovatus, Bacteroides fragilis, Bacteroides plebeius, Bacteroides uniformis, and Bacteroides coprophilus, confirming the changes observed at the genus level. Moreover, these species were present in a significantly higher proportion than in the assay with inulin. For the assay with inulin, the predominant species were Parabacteroides distasonis and Clostridium perfringens, also present in the assay with H. elongata but at a much lower percentage. After 48 h of fermentation, the main differences observed indicated an increase in B. ovatus in the assay with inulin, whereas there was a decrease in C. perfringens compared with the same assay after 24 h. For the assay with H. elongata, there was an increase in P. distasonis together a decrease in the Bacteroides species except for B. uniformis, which increased after 48 h compared with 24 h (Fig. S1, Supplementary material). Other significant differences were found for minor abundant species such as Roseburia faecis, Ruminococcus gnavus and Faecalibacterium prausnitzii being the two most abundant for the seaweed and the last one for inulin, at 24 h. The increase of Bacteroides species can be controversial because they have been reported in different clinical infections (Wexler, 2007). However, it is important to note that Bacteroides species are the most abundant in the human gut, generally maintaining a beneficial relationship with the host and their abundance has been related to the consumption of rich-fiber diets. When they escape this environment is when they can act as pathogens causing infections. As friendly commensal, Bacteroides species utilize simple and complex sugars and Table 3 Changes in lactate and short-chain fatty acids (SCFA) concentrations (mM) in the samples obtained for each substrate after fecal fermentation assays at 0, 10, 24, and 48 h. The results are expressed as the mean ±standard error (n =6). nd: not detected (LOD =0.016–0.03 mM). A two-way ANOVA, using general lineal model performed with time and substrates as two covariates for each bacterial group, indicated significant differences between both time and substrates, p <0.05. Then, a one-way ANOVA with Tukey’s test was performed to determine a significant increase/decrease of SCFA concentration with time for the same substrate (different lowercase letters indicate significant differences), also a one-way ANOVA analysis was carried out to compare substrates at 24 and 48 h (indicated with capital letters). H. elongata Inulin Negative control SCFAs 0h 10h 24h 48 h 10h 24h 48h 10h 24h 48h Succinic 0.31 a ±0.23 0.50 a ±0.14 nd 0.33 a,A ±0.24 0.11 a ±0.08 5.53 b,A ±2.25 0.80 a,b,A ±0.56 0.28 a, ±0.04 0.50 a,C ±0.20 nd Lactic 0.64 a ±0.48 0.40 a ±0.28 3.62 a,A ±2.56 nd 0.09 a,b ±0.07 6.74 b,A ±2.39 7.09 b,A ±2.00 0.09 a ±0.06 0.23 a,B ±0.17 0.18 a,B ±0.13 Formic 1.16 a ±0.90 1.57 a ±1.11 7.09 a,A ±5.02 nd nd 21.87 b,B ±5.22 15.39 b,A ±2.92 0.35 a, ±0.25 1.25 a,B ±0.89 1.06 a,B ±0 .75 Acetic 4.36 a ±0.66 1.59 a ±0.06 17.34 a,b,A ±2.54 25.93 b,A ±1.02 0.80 a ±0.47 21.42 b.A ±0.55 35.37 b,A ±5.84 0.55 a ±0.06 8.2 a.B ±0.46 9.67 a,B ±0.33 Propionic 2.42 a ±0.54 1.79 a ±0.05 8.19 b,A ±2.06 3.93 a,b,A ±0.82 1.06 a ±0.38 1.22 a,B ±0.44 16.61 b,B ±4.74 1.53 a, ±0.05 2.58 a,B ±0.38 2.34 a,A ±0.26 Isobutyric nd nd 4.28 a,A ±2.57 0.81 a,A ±0.29 nd nd nd nd nd 0.06 a,B ±0.05 Butyric 3.55 a ±1.34 1.80 a ±0.31 8.17 a,A ±1.30 4.30 a,A ±1.52 2.34 a ±0.92 9.65 a,b,A ±0.86 18.23 b,B ±5.62 0.82 a, ±0.32 1.25 a.B ±0.11 0.79 a,C ±0.18 Isovaleric nd nd 3.92 a,A ±2.21 0.84 a,A ±0.36 nd nd 0.18 a,A ±0.13 nd 0.18 a,B ±0.13 0.54 a,A ±0.25 Valeric nd nd 3.51 a,A ±2.22 0.69 a,A ±0.49 nd nd nd nd 0.16 a,B ±0.11 0.17 a,A ±0.12 Total 12.53 ±4.27 7.64 ±1.95 56.11 ±20.48 36.03 ±4.73 4.41 ±1.91 66.43 ±11.71 93.66 ±21.82 3.62 ±0.78 14.37 ±2.45 14.82 ±2.07 A. Lopez-Santamarina et al. Food Research International 156 (2022) 111156 8 polysaccharides for growth. In the present work, seawed carbohydrates are mainly used by Bacteroides ovatus, following by other species such as B. fragilis, B. plebeius and B. uniformis as previously indicate. B. ovatus has been previously reported as degrader of seaweed carbohydrates (Hemsworth et al., 2020). Although, in this case, the strain was not identified, is remarkable to indicate that are strains from B.ovatus considered beneficial as the case of B.ovatus strain ELH-B2 considered as a potential next-generation probiotic due to its preventive effects on lipopolysaccharides-associated inflammation and intestinal microbiota disorders in mice (Tan, Yu, Wang, Zhang, Zhao, Zhang et al., 2018). Analysis of α -diversity in terms of Shannon H diversity and the number of OTUs showed that for both substrates, the diversity decreased after 24 h and then slightly increased from 24 to 48 h. These changes in terms of diversity were significant in the assays at 24 and 48 h (with respect to time 0 h) for both substrates. With H. elongata, the diversity was lower after 24 h and increased after 48 h. By contrast, for inulin this diversity remained constant and similar to the α -diversity reached for H. elongata after 48 h. For β-diversity, the Bray–Curtis analysis showed four different groups: (1) 0 h; (2) control at 48 h; (3) H. elongata at 24 and 48 h and inulin at 24 and 48 h; and (4) control at 10 h, control at 24 h, H. elongata at 10 h, and inulin at 10 h. Finally, statistical analysis of the 20 most frequent metabolic pathways obtained with PICRUSt showed significant differences between 0 and 24 h for H. elongata and inulin (Fig. S2a and S2b respectively, Supplementary material). At 0 h, the most frequent metabolic pathways were related to sporulation, pore ion channels, lipopolysaccharide biosynthesis proteins, ribosome biogenesis, and the pentose phosphate pathway, among others. The statistical analysis obtained for the metabolic pathways obtained for H. elongata and inulin at 24 and 48 h showed no significant differences between them (Fig. S2c and S2d. Supplementary material). The same result was obtained from the analysis of the metabolic functions for H. elongata after 24 and 48 h (Fig. 3a) and inulin after 24 and 48 h (Fig. 3b). Among the 20 most frequent metabolic pathways for H. elongata were ABC transporters; twocomponent systems; transporters; DNA repair and recombinant proteins; secretion system; and different pathways related to carbohydrate metabolism such as fructose and mannose metabolism, pentose and glucuronate interconversions, and other glycan degradation. For inulin, the same pathways were present, with some exceptions related to carbohydrate metabolism, such as fructose and mannose metabolism, which did not appear among the 20 most frequent in the assays with inulin. Fig. 3. Main metabolic pathways. Statistical Analysis Functional Profile of the main 20 metabolic pathways for the samples obtained from the in vitro colonic model with a) H. elongata at 24 and 48 h and b) inulin at 24 and 48 h. The p-value obtained for all the metabolic functions (only the 20 more representative in graphics) was higher than 0.05, indicating no significant differences between the times of assay. Fig. 2. Statistical differences for the most abundant identified bacterial species. Significant differences (p <0.05) obtained from the analysis of samples (a) INU24H (fermentation assay with inulin for 24 h) and SW24H (fermentation assay with Himanthalia elongata for 24 h) and (b) INU48H (fermentation assay with inulin for 48 h) and SW48H (fermentation assay with H. elongata for 48 h) for the most common species using a G-test (with Yates’ correction) +Fisher’s exact test with Bonferroni correction in STAMP. A. Lopez-Santamarina et al. Food Research International 156 (2022) 111156 9 3.4. SCFAs analysis Table 3 shows the SCFA analysis. As expected, there was a greater SCFA production for the assays with either substrate compared with the negative control (without substrate). For inulin, the highest SCFA production occurred after 48 h, whereas for H. elongata, the maximum production occurred after 24 h (93.66 mM vs. 56.11 mM). However, when comparing SCFA production after 24 and 48 h, there were no significant differences between the substrates. Formic, acetic, propionic, and butyric acids dominated when production was maximum. These results are similar to those obtained by Chen et al. (2018), where the concentration of SCFA increased in the group treated with the seaweed polysaccharide (predominantly acetic and propionic acids). Bajury, Rawi, Sazali, Abdullah, & Sardini, 2017 also reported an increase in acetic acid over the course of seaweed fermentation. The SCFA results are in concordance with the results obtained for the GM analysis. For the fermentation assays with inulin after 48 h and H. elongata after 24 h (highest SCFA production), there was also a higher representation of the phylum Bacteroidetes in GM. This phylum is known to be the main producer of SCFA (Fu et al., 2018). In the case of H. elongata, propionic acid increased considerably after 24 h, at which time there was greater representation of different Bacteroides species. This may be because Bacteroidetes can metabolize various carbohydrates to produce propionate through the succinate pathway (Chen et al., 2018). 4. Conclusions This is the first time that whole brown seaweed (H. elongata) has been tested in an in vitro model of the human colon to determine its potential prebiotic activity. The obtained results in terms of iodine and As composition indicate the suitability and safety of using a similar amount of the product to other known and marketed prebiotic compounds. H. elongata was selectively used by some members from GM, causing an increase in Bacteroides species; on the contrary, inulin led to a significant increase in the relative abundance of P. distasonis, which has been described for the first time. While taxonomic differences in GM were found for the two substrates employed, the metabolic pathways associated with the GM function was not significantly different between the substrates, indicating the potentiality of the use of the whole seaweed as prebiotic. Although these results have been obtained by means an in vitro assay with its limitations, the present work includes findings that encourage to continue with the research on the potential use of H. elongata (and maybe, other seaweeds) as potential prebiotic ingredient in supplements or food. Our future prospects include a complete chemical characterization of H.elongata, mainly polyphenols and carbohydrates, in vivo studies with animals and humans as well as identification of the bacterial strain B. ovatus, the main identified specie that use the seaweed togrow and elucidate its role. Funding The authors thank the Xunta de Galicia and European Regional Development Funds (FEDER), grant ED431C 2018/05, for covering the cost of the work. CRediT authorship contribution statement Aroa Lopez-Santamarina: Methodology, Formal analysis, Writing – original draft. Alejandra Cardelle-Cobas: Methodology, Formal analysis, Writing – original draft, Visualization. Alicia del Carmen Mondragon: Methodology. Laura Sinisterra-Loaiza: Methodology. Jose Manuel Miranda: Conceptualization, Investigation, Writing – review & editing, Visualization, Supervision, Project administration. Alberto Cepeda: Conceptualization, Funding acquisition, Investigation, Writing – review & editing, Visualization, Supervision, Project administration. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgement Authors would like to thank the use of RIAIDT-USC analytical facilities. Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi. org/10.1016/j.foodres.2022.111156. References Afgan, A., Baker, D., Batut, A., van den Beek, M., Bouvier, D., ˇ Cech, M., … Blankenberg, D. (2018). The galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2018 update. Nucleic Acids Research, 46(W1), W537–W544. https://doi.org/10.1093/nar/gky379 Asensio-Grau, A., Calvo-Lerma, J., Heredia, A., & Andr´ es, A. (2021). In vitro digestion of salmon: Influence of processing and intestinal conditions on macronutrients digestibility. Food Chemistry, 342, Article 128387. https://doi.org/10.1016/j. foodchem.2020.128387 Association of Official Analytical Chemists, & Horwitz, W. (AOAC). (2002). Official methods of analysis (17th ed.). Gaithersburg, DC: Association of Official Analytical Chenist. Bolyen, E., Rideout, J.R, Dillon, M.R., Bokulich, N.A., Abnet, C.C., Al-Ghalith, G.A., … Caporaso, J.G. (2019). Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. https://doi.org/10.1038/s41587-0190209-9857. Bajury, D. M., Rawi, M. H., Sazali, I. H., Abdullah, A., & Sardini, S. R. (2017). Prebiotic evaluation of red seaweed (Kappaphycus alvarezii) using in vitro colon model. International Journal of Food Sciences and Nutrition, 68(7), 821–828. https://doi.org/ 10.1080/09637486.2017.1309522 Brodkorb, A., Egger, L., Alminger, M., Alvito, P., & Assunçao, R. (2019). INFOGEST static in vitro simulation of gastrointestinal food digestion. Nature Protocols, 14(4), 991–1014. https://www.nature.com/articles/s41596-018-0119-1. Cardelle-Cobas, A., Olano, A., Corzo, N., Villamiel, M., Collins, M., Kolida, S., & Rastall, R. A. (2012). In vitro fermentation of lactulose-derived oligosaccharides by mixed fecal microbiota. Journal of Agricultural and Food Chemistry, 60(8), 2024–2032. https://doi.org/10.1021/jf203622d Charoensiddhi, S., Conlon, M. A., Vuaran, M. S., Franco, C. M. M., & Zhang, W. (2017). Polysaccharide and phlorotannin-enriched extracts of the brown seaweed Ecklonia radiata influence human GM and fermentation in vitro. Journal of Applied Phycology, 29, 2407–2416. https://doi.org/10.1007/s10811-017-1146-y Chen, L., Xu, W., Chen, D., Chen, G., Liu, J., Zeng, X., … Zhu, H. (2018). Digestibility of sulfated polysaccharide from the brown seaweed Ascophyllum nodosum and its effects on the human GM in vitro. International Journal of Biological Macromolecules, 112, 1055–1061. https://doi.org/10.1016/j.ijbiomac.2018.01.183 Commission Recommendation (EU) 2018/464 of 19 March 2018 on the monitoring of metals and I in seaweed, halophytes and products based on seaweed. Official Journal of the European Journal, L78, 16-18. Commission Regulation (EC) No 1881/2006 of 19 December 2006 setting maximum levels for certain contaminants in foodstuffs. Official Journal of the European Union, L364, 5-24. Cronin, P., Joyce, S. A., O Tolle, P. W., & OConnor, E. M. (2021). Dietary fibre modulates the gut microbiota. Nutrients, 13, Article 1655. https://doi.org/10.3390/ nu13051655. Dejean, G., Tamura, K., Cabrera, A., Jain, N., Pudlo, N. A., Pereira, G., .. Brumer, H. (2020). Synergy between cell surface glycosidases and glycan-binding proteins dictates the utilization of specific beta(1,3)-glucans by human gut Bacteroides. mBio 11, Article e00095-20. https://doi.org/10.1128/mBio.00095-20. Cosgrove, W. J., & Loucks, D. P. (2015). Water management: Current and future challenges and research directions. Water Resources Research, 51, 4823–4839. https://doi.org/10.1002/2014WR016869 Des, M., Martinez, B., de Castro, M., Viejo, R. M., Sousa, M. C., & Gomez-Gesteira, M. (2020). The impact of climate change on the geographical distribution of habitatforming macroalgae in the Rias Baixas. Marine Environmental Research, 161, Article 105074. https://doi.org/10.1016/j.marenvres.2020.105074 A. Lopez-Santamarina et al.