The Effect of Pea Hull Fiber on Microbiota Composition Ambika Anand College of Liberal Arts and Sciences, Microbiology & Cell Science University of Florida Faculty Mentor: Dr. Volker Mai Department of Epidemiology University of Florida
Anand 1 Abstract Gut microbiota has been associated with proper digestive function, immunity, and other metabolic functions. Microbiota differs among individuals and over time, and it is influenced by environmental factors, including diet. Distortions in microbiota composition have been linked to inflammatory disorders, metabolic diseases, and conditions such as obesity. Dietary fiber is one of the main fuels utilized by gut microbes that results in the generation of various fermentation pr oducts such as short chain fatty acids. In our study, gut microbi ota was analyzed to determine how added pea hull fiber influenced the composition of bacteria. Five healthy male participants completed a 45 day, fully controlled fe eding, weight maintenance study separated into three, 15 day intervention periods. During period 1, participants received a high protein diet. During period 2, pa rticipants received 20 g of pea hull fiber in addition to the high protein diet. During period 3, subjects returned to r eceiving only the high protein diet. A follow up sample was collected after 30 days of normal diet after the last intervention period The addition of pea hull did not affect overall fecal microbiota diversity measures in a consistent pattern but appeared to affect the frequency at which specific bacterial signature sequences were observed and resulted in changes in prevalence. The high protein diet suppressed numbers of bifidobacteria as the proportion of bifidobacteria present after the washout period was significantly higher than during the study Supplementation with bifidobacteria may help to maintain a healthy microbiota composition when consuming high protein diets. Further research is required to confirm the findings of this pilot study.
Anand 2 Introduction The human gut contains a larg e population of microbiota that promotes metabol ic functions and immune health. Commensal microbes reside primarily in the large intestine due to an abundance of nutrients, slow flow rate and generally neutral to slightly acidic pH which contrasts with the faster flow rate and higher bile concentrations in the small intestine 1 The colonization of bacteria in the gastrointestinal tract immediately after birth is influenced by environmental factors such as diet, leading to distinct individual microbiota composition that after developing into a mature composition at an age of 1 2 years varies little over time Normal g ut microflora has a symbiotic relationship with the host and can help to ex clude potentially pathogenic competitors by preventing entr y and synthesizing bacteriocins 2 Fecal samples are thought to reflect the composition present in the large intestine. They can be affected by the addition of probiotic, live bacteria, or prebiotic substrate material for bacteria, measures 1 From metagenomic sequencing, the three typically dominant phyla are Firmicutes, Bacteroidetes, and Actinobacteria. Due to proposed correlations between microbiota composition and health, modifying microbiota co mposition might be a method to reduce the risk of infections 1 D iet has been associated with microbiota composition and can be used as a tool to alter composition towards a healthier distribution 3 4 Because of these diet associations, it is suspected that both diet and potential changes in microbiota contribute to increased incidence of diseases such as obesity and inflammatory bowel disease 4 Gut microbiota manipulation, particularly through the use of probiotics and prebiotics, has received renewed inte rest for influencing metabolic and immune functions While probiotics involve consumption of a bacterium for health benefits, prebiotics are non digestible food ingredients that seek to promote increased activity of specific bacteria of the colon 1 Dietary fiber has frequently been recommended to improve general health and well being 5 Evidence suggests
Anand 3 that supplementing a diet with dietary fiber can protect against cardiovascular disease 6 obesity 7 and perhaps even type 2 diabetes 8 Dietary fiber is composed of non starch polysaccharides that cannot be digest or absorbed. It generally passes through the body in a soluble or insoluble state 9 Fiber supplementation has been associated with an increase in b ifidobacteria suggesting that it can promote a healthy composition of the intestinal microbiome 10,11 The addition of fiber has also has been associated with improved intestinal barrier function, which is important for immune health 12 The subject of gut microbiota requires further research to dete rmine the scope of the implications of its behavior. The variation between individual microbiota prevents successful comparisons between groups and limits generalizations about findings. However, further research into establishing correlations between bact eria and disease may prove promising 13 For this study, pea hull fiber was provided as a supplement in addition to a high protein diet. Pea hulls also have a low amount of carbohydrates 14 and high fiber and resistant starch 15 content, making it ideal for s upplementation. The majority of pea hull fiber is made up of insoluble fiber. Pea hull fiber in particular has been highly associated with increased bowel frequency and might affect gut microbiota, potentially due to the passage of high microbial mass 1 6 The objective was to observe microbiota composition for any changes due to the intervention of pea hull fiber or potentially the maintenance of the high protein diet. Materials and Methods Study Design The study design had four time points. First, the healthy participants were given a high protein diet for 15 days. During the last three days of this period, the first stool sample was taken. For the second period, the participants received 20g of pea hull fiber to take in addition to
Anand 4 the high protein di et for 15 days. The second stool sample was taken in the last three days of this period. For the next 15 days, the additional fiber was removed, leaving the participants with just the high protein diet. The third stool sample was taken in the last three da ys of this period. The fourth sample was taken one month after the study had finished and the participants had returned to their normal diet. In total t wenty stool samples were received from five participants four samples each. DNA Isolation Total DNA from the stool samples was extracted using the QIAGEN DNA Stool Mini Kit and a bead beating step. P olymerase Chain Reaction DNA Amplification The extracted bacterial DNA was amplified with primers for the V6 V8 region of the 16s rDNA portion of the genome. To amplify, a master mix containing MgCl 2 10x buffer, dNTPs, reaction. The samples underwent denaturation at ~94 C to separate the strands annealing at ~5 5C for primer attachment to ion at ~68 C for synthesis of new, complementary strands using the free nucleotides These three steps were repeated for 35 cycles, lasting between 2 and 2.5 hours. was used to check for product by running it on an agarose gel. Multiple rounds of PCR were performed in order to amplify the DNA for sequencing and to have enough DNA to run DGGE gels.
Anand 5 Denaturing Gradient Gel Electrophoresis ( DGGE ) and Microbiota Analys is DGGE was performed on an acrylamide gel with a denaturing gradient of 40% at the top of the gel and 50% at the bott om consisting of the chemical agents formamide and urea. Samples were run through the DGGE gel at 65 volts, 60C, for 16 hours. After el ectrophoresis was complete, the gels were removed and stained with SYBR Gold for ~45 minutes and analyzed using the Diversity Database software. Shannon and Simpson Diversity Indexes were calculated. Quantitative PCR (qPCR) qPCR was used to amplify three di fferent targets: the V3 region, lactic acid bacteria, and B ifidobacteria The master mix contained a SYBR Green Master Mix and water provided by Q uantiTech SYBR Green PCR Kit and forward and reverse primer for an overall 12.5 reaction. The qPCR reactions were carried out with an initial melting step at 95C for 10 minutes, followed by 40 cycles of 95C for 30 seconds for denaturation 58C for 60 seconds (for B ifidobacteria (Bif) and LAB, 56C for the V3 universal primer set ) for annealing and 72C for 1 minute for elongation GATTCTGGCTCAGGATGAACG CGGGTGCTICCCCACTTTCATG ACGAGTAGGGAATCTTCCA ATTYCACCGCTACACATG CCTACGGGAGGCAG CAG ATTACCGCGGCTGCTGG Based on the amplification of each of these targets, it was possible to calculate the number of genomes present of each type and the proportions of bacteria for comparison between per iods.
Anand 6 DNA Sequencing As our main interest was in the effects of pea hull fiber we compared microbiota in the samples collected during the three intervention periods. The samples were amplified in 50 PCR reactions The16s rRNA V1 V3 region was amplified using the forward primer 27F 1 with CTATGCGCCTTGCCAGCCC GCTCAGTCAGAGTTTGATCCTGGCTCA CGTATCGCCTCCCTCGCGCCATCAGNNNNNNNNCATTACCGCGGCTGCTG GCAC The samples were then purified for sequencing using the Q iaQuick PCR Purification Kit with the centrifuge method From the raw sequence set obtained by 454 Titanium sequencing, low quality sequences or sequences with a length less than 150 nucle otides were removed. Sequen ces were binned using ESPRIT at 98% and 95% similarity levels. QIIME was used to calculate Chao rarefaction diversity and UniFrac distances. Statistics Statistical significance for OTU values between periods was done using two tailed z tests. Heatmaps were generated using the matrix2png interface DGGE analysis was done using Diversity Database software. The Shannon Weiner and Simpson (1/D) diversity indexes generated by the Chang Bioscience calculator were used to find microbiota diversity A two tailed t test was used to analyze for statistical significance. All stat istical analysis w as completed using Microsoft Excel.
Anand 7 Results Denaturing Gradient Gel Electrophoresis (DGGE) and Microbiota Analysi s Figure 1 shows the DGGE gels for all subjects. It appears that each individual has their own specific microbiota. N o distinct changes due to diet were observed Shannon diversity indices and Simpson diversity indices were calculated based on the number of bands revealing no significant difference in the diversity of microbiota at each time point. Results are shown in Figures 2 and 3. Quantitative PCR Figures 4 and 5 show the proportion s of B ifidobacteria and lactic acid bacteria present in the samples for each subject at each time point. The proportions were calculated from the qPCR data and genomes per sample and divided by the total genomes as found through use of a primer set targeting the V3 region. All subjects showed a decrease in lactic acid bacteria between periods 1 and 2, indicating a correlation with the addition of the pea hull fiber. For a majority of subjects, Bifidobacteria increased between th e same two periods. For time point 4, after the subjects returned to their free living diet, Bifidobacteria increased by a large amount while lactic acid bacteria mostly decreased. DNA Sequencing From 454 Titanium Sequencing 73,685 sequences were retain ed after removing low quality and short reads (sequences with less than 150 nucleotides). For each sample we generated
Anand 8 an average of 4909 sequences with an average length of 486 nucleotides. The results of sequencing show numerous operational taxonomic uni ts (OTUs) based on a similarity rate (95 or 98%) used to compare the sequences to one another and to the database. The data was separated by sample to see how the bacterial composition changed over the three periods and can be found in Figure 6. The figure shows that Firmicutes dominated the microbiota for the most part. There is no clear pattern of changes in distribution shared by all subjects. The graphs by phylum (Figure 7) show the different microbiota for each individual. There was no clear pattern of change that could be pinpointed to the time periods aside from Firmicutes dominating the microbiota at all time points Also from the sequencing data, rarefaction curves and UniFrac plots were generated. Rarefaction curves shown in Figure 8, did not diff er between diet and fiber periods. UniFrac plots represent the distances between organisms based on their phylog e nic relationship. The plots (Figures 9 and 10) show that the samples group by subject and not by period or by diet This is supported by the in dividualized diversities shown by the DGGE gels acid bacteria and B ifidobacteria counts from qPCR. Heatmaps were created based on comparisons by z test, based on whether an OTU was present at one time period a nd not another. First, periods 1 and 2 were compared. All OTUs with z scores greater than 1.96 or less than 1.96 corresponding to p< 0.05, were considered significant. Those OTUs were plotted in a heatmap for all three periods. Both comparisons are marke d in the same heatmap for the similarity level (Figures 11 and 12) The heatmaps show that there are some OTUs that only appear in the second period, perhaps due to the addition of fiber. There are also OTUs that are there initially but disappear in the second period. Some rea ppear in the third, but not all, indicating again an effect of the fiber addition that is limited to the duration of the supplement Of the significant OTUs from both z
Anand 9 tests at both levels, there are two common ones: Mogibacterium_t imidum (85.17) and Clostridium_sp._strain_Z6 (88.1). Both were more prevalent in period 2, or with the addition of pea hull fiber. Of greater interest are the OTUs that appear or increase in period 2 and remain elevated in period 3 or that disappear or decrease in period 2 and remain reduced in period 3 because they demonstrate potential long term effects of the addition of pea hull fiber Ruminococcus species and Streptococcus thermophilus decrease in period 2 and stay lowered in period 3. Faecaliba cterium prausnitzii Lactococcus lactis and Oscillibacter valericigenes increase in period 2 and stayed elevated in period 3. Discussion Addition of dietary fiber to a high protein diet did not significantly affect o verall microbiota diversity Through diversity based analyses, subjects showed individualized microbiota, and any minor changes observed were by subject, not change in diet. However, pea hull fiber addition did correlate with a decrease of lactic acid bac teria and a slight increase of b ifidobacteria R eturn to normal diet from a high protein study diet r esulted in a large increase in b ifidobacteria This change suggests that the high protein diet had an effect on bifidobacteria that should be studied further as the data proposes that the diet decreased the amount of beneficial bacteria. The increase in bifidobacteria after return to normal diet also suggests that a bifidobacteria probiotic or a prebiotic that supports bifidobacteria maintenance such as inulin or oligofructose 17 should be provided alongside a high protein diet to maintain healthy levels of beneficial bacteria. Bifidobacteria assists with protection from enteric infection, suppression of pathogenic bacteria, stimulation of the immune system, and other activities w ithin the gut, making its decrease potentially harmful to the human host 17 Specific strains of bifidobacteria are
Anand 10 regularly used as probiotic supplements due to their beneficial role within the gut and non pathogenic profile 18 Unfortunately we don't hav e data on how overall microbiota composition was affected after return from high to normal protein content. While overall diversity seemed unaffected by pea hull fiber supplementation, t he OTU distributions shown in the heatmaps indicates that some bacter ia were indeed different between periods. Several Ruminococcus species decrease in period 2 and remained at lower numbers in period 3 It appears that once dietary changes results in a disappearance of these bacteria, returning to normal diet will not imme diately result in recurrence of these bacteria. Ruminococcus species are known for degrading cellulose of animals such as cattle and sheep, and more recently, humans 1 9 The species should increase in number when dietary fiber is added due to the cellulose present in dietary fiber. However, the opposite occurs, indicating a potential effect of an extended period with the high protein diet or another unknown reason that requires more research. Also notable was the increase of Faecalibacterium prausnitzii and Oscillibacter valericigenes in period 2 that was maintained in period 3. Once these bacteria increased, potentially through positive selection by pea hull fiber, they stay elevated even after the selection pressure (pea hull fiber) is removed. F. prausnit zii has been strongly associated with a healthy disease. The bacterium may assist in maintaining immunity and has the potential to be a probiotic treatment for maintain ing a healthy gut 20 The increase of this species suggests that the addition of dietary fiber promotes a beneficial bacterium. O. valericigenes has al so been dis ease proposing another marker of the benefits of the dietary fiber addition 21
Anand 11 Another change is the increase of Lactococcus lactis in period 2 that is maintained into period 3. In the same periods, Streptococcus thermophilus decreases and stays lowered. Both are lactic acid bacteria of the same family and commonly associated with dairy production and the conversion of lactose into lactic acid. It is possible that the addition of pea hull fiber promotes L. lactis over S. thermophilus as they compete for t he same source. However, it is also possible that the provided diet had more dairy products with L. lactis primarily cheeses, than with S. thermophilus which is primarily associated with yogurt production 2 2 2 3
Anand 12 Figure 1 : DGGE gels for all five subjects after each period. S stands for
Anand 13 Figure 2 : Shannon indices for a ll subjects were averaged by tim e point. There is no significant difference between the subjects. Figure 3 : Changes in Shannon indices between consecutive periods was calculated. As shown, there was no clear pattern in changes in diversity. The most variation was found between periods 3 and 4, indi cating the effect of the free living diet.
Anand 14 Figure 4 : This graph shows the percentage of bifidobacteria present in each subject at each time point. The percentages were determined by dividing the number of genomes of bifidobacteria by the total number of genomes in the sample.
Anand 15 Figure 5 : This graph shows the percentage of lactic acid bacteria present in each subject at each time point. The percentages were determined by dividing the number of genomes of lactic acid bacteria by the total number of genome s in the sample.
Anand 16 Figure 6 : The graph shows the distribution of bacteria by phylum at 95% similarity for each subject at the indicated time point. Figure 6 : The graph shows the distribution of bacteria by phylum at 95% similarity for each subject at the indicated time point.
Anand 17 Figure 7 : The ab ove graphs show a closer lo ok at proportions of bacteria by phyla at the 95% similarity level, allowing for comparison across subjects within a phylum.
Anand 18 Figure 8 : This rarefaction plot is based on the groups that were only on diet (samples from periods 1 and 3) and those that were on the diet and fiber (period 2). The curves did not differ between diet and fiber periods.
Anand 19 Figure 9 : Each shape (blue square, red triangle, or green circle) represents a different time period of t h e three sequenced. The groups created are clearly not by period or diet. Figure 10 : Each color (blue square, green triangle, cyan circle, red triangle, and y ellow triangle) represents a different subject. The dimensions are the same here as for the previous set of plots. It is easy to see the groups created by subject, showing individualized microbiota.
Anand 20 Figure 11 : This heatmap has OTUs chosen by the z test results comparing periods 1 and 2 and periods 2 and 3 for the 95% similarity results. The heatmap is separated into four sections: (1) compares periods 1 and 2 and has OTUs more prevalent in period 1; (2) compares periods 2 and 3 and has OTUs more prevalent in period 3; (3) compares periods 1 and 2 and has OTUs more prevalent in period 2; (4) compares periods 2 and 3 and has OTUs more prevalent in period 2.
Anand 21 Figure 12 : This heatmap has OTUs chosen by the z test results comparing periods 1 and 2 and periods 2 and 3 for the 98% similarity results. The heatmap is separated into four sections: (1) compares periods 1 and 2 and has OTUs more prevalent in period 1; (2) compares periods 2 and 3 and has OTUs more prevalent in period 3; (3) compares periods 1 and 2 and has OTUs more prev alent in period 2; (4) compares periods 2 and 3 and has OTUs more prevalent in period 2.
Anand 22 Acknowledgements I would like to thank Dr. Volker Mai for the opportunity to learn as a part of his research lab and Dr. Maria Ukhanova for her instruction and guidance in the lab References 1. Flint HJ, Scott KP, Louis P, Duncan SH. The role of the gut microbiota in nutrition and health. Nat Rev Gastroenterol Hepatol 2012;9:577 589. 2. Guarner F, Malagelada JR. Gut flora in health and disease. The Lancet 2003;3 61(9356):512 519. 3. Mai V, Colbert LH, Perkins SN, Schatzkin A, Hurst ing SD. Intestinal microbiota: a potential diet responsive prevention target in ApcMin mice. Mol Carcinog 2007;46(1):42 48. 4. David LA, Maurice CF, Carmody RN, Gootenberg DB, Button JE, W olfe BE, Ling AV, Devlin AS, Varma Y, Fischbach MA, Biddinger SB, Dutton RJ, Turnbaugh PJ. Diet rapidly and reproducibly alters the human gut microbiome. Nature 2014;505(7484):559 563. 5. Brownawell AM, Caers W, Gibson GR, Kendall CWC, Lewis KD, Ringel Y, Slavin JL. Prebiotics and the health benefits of fiber: Current regulatory status, future research, and goals. J Nutr. 2012;142(5):962 974. 6. Eshak ES, Iso H, Date C, Kikuchi S, Watanabe Y, Wada Y, Wakai K, Tamakoshi A, JACC Study Group. Dietary fiber intak e is associated with reduced risk of mortality from cardiovascular disease among Japanese men and women. J Nutr 2010;140(8):1445 1453. 7. Davis JN, Hodges VA, Gillham MB. Normal weight adults consume more fiber and fruit than their age and height matched overweight/obese counterparts. J Am Diet Assoc 2006;106(6):833 840. 8. Hopping BN, Erber E, Grandinetti A, Verheus M, Kolonel LN, Maskarinec G. Dietary fiber, magnesium, and glycemic load alter risk of type 2 diabetes in a multiethnic cohort in Hawaii. J Nu tr. 2010;140(1):68 74. 9. Turner ND, Lupton JR. Dietary fiber. Adv Nutr 2011;2(2):151 152. 10. Kuo SM. The interplay between fiber and the intestinal microbiome in the inflammatory response. Adv Nutr. 2013;4:16 28.
Anand 23 11. Zeng H, Lazarova DL, Bordonaro M. Mechanisms linking dietary fiber, gut microbiota and colon cancer prevention. World J Gastrointest Oncol 2014;6(2):41 51. 12. Chen H, Mao X, He J, Yu B, Huang Z, Yu J, Zheng P, Chen D. Dietary fibre affects intestinal mucosal barrier function and regulat es intestinal bacteria in weaning piglets. Br J Nutr. 2013;110(10):1837 1848. 13. Mai V, Draganov PV. Recent advances and remaining gaps in our knowledge of associations between gut microbiota and human health. World J Gastroenterol 2009;15(1):81 85. 14. Longst aff M, McNab JM. Digestion of fibre polysaccharides of pea ( Pisum sativum ) hulls, carrot and cabbage by adult cockerels Br J Nutr 1989;62(3):563 577. 15. Ratnayake WS, Hoover R, Warkentin T. Pea starch: c omposition, structure, and properties a review. Starch/Strke 2002;54(6):217 234. 16. Dahl WJ, Whiting SJ, Healey A, Zello GA, Hildebrandt SL. Increased stool frequency occurs when finely processed pea hull fiber is added to usual foods consumed by elderly residents in long term care. J Am Diet Assoc. 200 3;103(9):1199 1202. 17. Kolida S, Tuohy K, Gibson GR. Prebiotic effects of inulin and oligofructose. B J Nutr. 2002;87(Suppl 2):S193 S197. 18. Hardy H, Harris J, Lyon E, Beal J, Foey AD. Probiotics, prebiotics and immunomodulation of gut mucosal defences : homeostasis and immunopathology. Nutrients. 2013;5(6):1869 1912. 19. Chassard C, Delmas E, Robert C, Lawson PA, Bernalier Donadille A. Ruminococcus champanellensis sp. nov., a cellulose degrading bacterium from human gut microbiota Int J Syst Evol Microbol 2012;62(1):138 143. 20. Miquel S, Martn R, Rossi O, Bermdez Humarn LG, Chatel JM, Sokol H, Thomas M, Wells JM, Langella P. Faecalibacterium prausnitzii and human intestinal health Current Opinion in Microbiology 2013;16(3):255 261. 21. Mondot S, Kang S, Furet JP, Aguirre de Carcer D, McSweeney C, Morrison M, Marteau P, Dore J, Leclerc M. Highlighting new phylogenetic specificities of Crohn's disease microbiota Inflammatory Bowel Diseases 2011;17(1):185 192. 22. Dhaisne A, Guellerin M, Laroute V, Laguerre S Cocaign Bousquet M, Le Bour geois P, Loubiere P. Genotypic and phenotypic analysis of d airy Lactococcus lactis biodiversity in milk: volatile organic compounds as discriminating m arkers Appl Environ Microbiol. 2013;79(15):4643 4652. 23. Adolfsson O, Meydani SN, Russell RM. Yogurt and gut function. Am J Clin Nutr. 2004;80(2):245 256.