OPEN No difference in small bowel microbiota

SUBJECT AREAS:
between patients with irritable bowel
MICROBIOME
INFECTION
syndrome and healthy controls
Aldona Dlugosz1, Björn Winckler2, Elin Lundin3, Katherina Zakikhany4, Gunnar Sandström5, Weimin Ye2,
Received Lars Engstrand6 & Greger Lindberg1
11 November 2014
Accepted 1
Karolinska Institutet, Department of Medicine and Center for Digestive Diseases, Karolinska University Hospital Huddinge,
22 January 2015 Stockholm, Sweden, 2Karolinska Institutet, Department of Medical Epidemiology and Biostatistics, Stockholm, Sweden, 3Stockholm
University, Department of Biochemistry and Biophysics, Stockholm, Sweden, 4Public Health Agency of Sweden, Unit for Laboratory
Published Surveillance of Vaccine Preventable Diseases, Solna, Sweden, 5Karolinska Institutet, Department of Laboratory Medicine,
17 February 2015 Stockholm, Sweden, 6Karolinska Institutet, Department of Microbiology, Tumor & Cell Biology and Science for Life Laboratory,
Stockholm, Sweden.

Correspondence and
Several studies have indicated that colonic microbiota may exhibit important differences between patients
requests for materials with irritable bowel syndrome (IBS) and healthy controls. Less is known about the microbiota of the small
should be addressed to bowel. We used massive parallel sequencing to explore the composition of small bowel mucosa-associated
A.D. (aldona. microbiota in patients with IBS and healthy controls. We analysed capsule biopsies from the jejunum of 35
dlugosz@karolinska. patients (26 females) with IBS aged 18-(36)-57 years and 16 healthy volunteers (11 females) aged 20-(32)-48
years. Sequences were analysed based on taxonomic classification. The phyla with the highest total
se)
abundance across all samples were: Firmicutes (43%), Proteobacteria (23%), Bacteroidetes (15%),
Actinobacteria (9.3%) and Fusobacteria (7.0%). The most abundant genera were: Streptococcus (19%),
Veillonella (13%), Prevotella (12%), Rothia (6.4%), Haemophilus (5.7%), Actinobacillus (5.5%), Escherichia
(4.6%) and Fusobacterium (4.3%). We found no difference among major phyla or genera between patients
with IBS and controls. We identified a cluster of samples in the small bowel microbiota dominated by
Prevotella, which may represent a common enterotype of the upper small intestine. The remaining samples
formed a gradient, dominated by Streptococcus at one end and Escherichia at the other.

T
he human gastrointestinal microbiota consists of about 100 trillion microbial cells that outnumber our own
cells by a factor of 101. In a healthy host, bacteria colonize the alimentary tract soon after birth, and the
composition of the intestinal microflora is believed to remain relatively constant throughout life2. The adult
human intestine is home to more than 1000 species of microbes, which normally remain confined to the distal gut
(colon) where the concentration of organisms is approximately 1011 organisms per gram of content3,4. Because of
peristalsis and the antimicrobial effects of gastric acidity, the stomach and proximal small intestine contain small
numbers of bacteria in healthy individuals. The bacterial counts of coliforms rarely exceed 103 colony-forming
units (CFU)/mL in jejunal juice2.
In a study based on human faecal samples that spanned several nations and continents (totally 39 individuals)
Arumugam et al.5 demonstrated the existence of enterotypes in the human gut microbiome and identified three of
them that varied in species and functional composition. Each of these three enterotypes are identifiable by the
dominance of one of three genera: Bacteroides (enterotype 1), Prevotella (enterotype 2) and Ruminococcus
(enterotype 3). The relationships between microbiota and different diseases and conditions have been studied
especially in the colonic microbiota and some significant associations have been observed for inflammatory bowel
diseases (IBD)4,6, metabolic syndrome7, and obesity8.
The small bowel microbiota has not been fully described, partly because small bowel samples are relatively
difficult to obtain9. The microbiota of effluents from the distal small bowel was found to vary with the intake of
carbohydrates in a single patient with ileostomy10. Another study found a substantial difference between morning
and afternoon samples of the ileal effluent in one subject with ileostomy and the variation in the microbiota profile
during the day was larger than that seen in repeat samples obtained at the same time of the day over 9–28 days in 4
subjects11. The suspicion that host-microbe interaction may underlie observed immune activation in IBS makes
the mucosa-associated microbiota more interesting as target for research than luminal microbiota12.

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 1

 www.nature.com/scientificreports

Figure 1 | Relative abundance of the five most common phyla among IBS patients and controls.

Figure 2 | Relative abundance of the eight most common genera among IBS patients and controls.

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 2

 www.nature.com/scientificreports

Table 1 | OTUs showing a trend towards differential expression between patients with IBS and controls. SEM 5 standard error of the mean
Abundance (%) mean 6 SEM

OTU Taxonomy Patients Controls p (unadjusted)

4391262 Proteobacteria, Escherichia (genus) 2.78 6 1.73 10.29 6 2.56 0.080
4465561 Bacteroidetes, Prevotella (genus) 3.16 6 0.51 1.56 6 0.76 0.025
4425214 Firmicutes, Streptococcus (genus) 0.94 6 0.29 2.27 6 0.42 0.054
271159 Firmicutes, Carnobacteriaceae (family) 1.54 6 0.41 0.73 6 0.60 0.081
31235 Fusobacteria, Leptotrichia (genus) 1.14 6 0.78 0.11 6 1.15 0.077

IBS is a common gastrointestinal disorder characterized by abdom- culture of jejunal fluid, with results expressed as CFU/mL of jejunal
inal pain or discomfort and altered bowel function13. The potential fluid has been regarded by many investigators as the gold standard
differences of the intestinal microbiota between IBS patients and for the diagnosis of SIBO, but molecular techniques suggest that as
healthy controls have mostly been studied using stool samples, as this much as 80% of the normal flora is not identified by culture-based
is the most accessible source of the GI microbiota14,15. In a detailed methods2. Profiling the microbiome using methods based on the 16S
faecal microbiota analysis of a well-characterized cohort of IBS ribosomal RNA gene is less biased than cultivation based approaches.
patients Jeffery et al.16 identified several clear associations with clinical In particular, pyrosequencing using parallel bar-coded sequence tags
data and a distinct subset of IBS patients with alterations in their enables deep sequencing of multiple samples and provides high taxo-
microbiota that did not correspond to IBS subtypes, as defined by nomic resolution21. The power of pyrotag sequencing for exploration
the Rome-II criteria13. of the human microbiome has been shown for different body sites21.
Kerckhoffs et al.17 found decreased Bifidobacteria levels in both As faecal samples are not representative of the entire intestine, the
faecal and duodenal brush samples of IBS patients compared to aim of the present study was to deeply explore the composition of
healthy subjects. Small intestinal bacterial overgrowth (SIBO) was small bowel mucosa-associated microbiota using 454-barcoded pyr-
proposed to be common in IBS18. Bacterial overgrowth is a condition osequencing in patients with IBS compared to healthy controls.
caused by an abnormal number of bacteria in the small intestine,
exceeding 105 organisms/ml (5 log colony-forming units (CFU)/ml) Results
owing to different predisposing conditions, such as impaired motility The amplicon reads were preprocessed as described in the ‘‘Bioin-
or failure of the gastric-acid barrier19,20. The direct aspiration and formatics pipeline’’ section, resulting in a data set consisting of 51

Figure 3 | Correlation between the relative abundance of Prevotella and Veillonella (Spearman rho 5 0.55 r < 0.55, p < 4 3 1025).

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 3

 www.nature.com/scientificreports

samples and 350 OTUs. The median depth of samples was 3771 and more samples would be needed to determine whether differences are
the minimum depth was 628. More than 99% of all reads could be real or due to chance.
taxonomically classified to phylum rank and 90% up to genus rank. We found that the abundance of certain genera exhibited patterns
The phyla with the highest total abundance across all samples of co-dependence. In particular: (1) Prevotella and Veillonella were
were: Firmicutes (43%), Proteobacteria (23%), Bacteroidetes (15%), correlated (Spearman r < 0.55, p < 4 3 1025) (Figure 3), (2)
Actinobacteria (9.3%) and Fusobacteria (7.0%) (Figure 1). These Escherichia and Rothia seemed to be mutually exclusive in the sense
phyla were present in all samples. The most abundant genera were: that more than 5% of one typically implied less than 4% of the other
Streptococcus (19%), Veillonella (13%), Prevotella (12%), Rothia (Figure 4), and (3) samples in which the total abundance of Prevotella
(6.4%), Haemophilus (5.7%), Actinobacillus (5.5%), Escherichia and Streptococcus was high exhibited an inverse relationship between
(4.6%) and Fusobacterium (4.3%) (Figure 2). These genera were these genera (Figure 5). The distribution of Prevotella abundance was
present in more than 98% of all samples, except Actinobacillus, which bimodal (Figure 6), indicating that samples may be naturally sub-
was present in 82%. divided into two distinct subgroups according to whether they have
In order to test if there were any differences in terms of bacterial low or high Prevotella abundance.
diversity within samples (a-diversity) we looked for associations A cluster analysis was performed in order to check for evidence of
between the disease phenotype (control, IBS, C-IBS, D-IBS) and a Prevotella ‘‘enterotype’’ as described by Arumugam et al.5. We
the number of OTUs, as well as the Chao1 and Shannon diversity. considered Bray-Curtis distance (BC) and Jensen-Shannon diver-
It is well known that these measures are highly dependent on sequen- gence (JSD) on the relative abundance aggregated to genus level, as
cing depth22, so we first randomly sub-sampled all samples to have well as weighted and unweighted UniFrac (after subsampling all
even depth before evaluating the above measures of diversity. We samples to have even depth). Coordinates for each dissimilarity
found no statistically significant difference between disease pheno- matrix were estimated using classical multidimensional scaling
types in terms of a-diversity. (MDS/PCoA) and these coordinates were clustered using partition-
We used Metastats23 in order to detect bacteria that were differ- ing around medoids (PAM). We found one robust cluster consisting
entially expressed across controls and patients, as well as across the of roughly 20%–27% of all samples, including all the samples that
different disease phenotypes. Although we found no statistically sig- were enriched for Prevotella (in line with the bimodal distribution of
nificant difference between IBS and controls, we did find some OTUs Prevotella abundance mentioned above). By ‘‘robust’’ we mean that
that exhibited a trend towards differential expression (Table 1). None this cluster appeared for all dissimilarities apart from unweighted
of these differences was significant after adjusting for multiple test- UniFrac (Supplementary Figure 1) and for a range of choices of the
ing. Since these OTUs are abundant we can rule out the possibility number of clusters (which is a parameter to the PAM algorithm). We
that observed differences were due to insufficient sampling depth but also tested the clustering strength using prediction strength (PS),

Figure 4 | The relative abundance of Escherichia and Rothia exhibit a mutually exclusive relationship.

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 4

 www.nature.com/scientificreports

Figure 5 | The relative abundance of Prevotella and Streptococcus exhibit an inverse relationship in samples with high total abundance of these two
genera.

silhouette index (SI) and Calinski-Harabasz coefficient (CH)24–26. In order to look for differences between individual IBS phenotypes
While PS and SI were never strong (maxima around 0.6 and 0.3, we performed a multinomial logistic regression (adjusted for age and
respectively) we always found the robust cluster for the value at gender) with the first five MDS coordinates as independent variables
which CH (or PS, or SI) were maximal. and the disease phenotype as dependent variable. We did not find any
The fact that the robust cluster appears for several dissimilarities indication that different IBS phenotypes could be separated in this way.
and different choices of the number of clusters we take as evidence
in favour of a Prevotella enterotype. However, the low clustering Discussion
strength indicates that the separation from other samples is not Our study is the first attempt at characterizing small bowel micro-
definite. Samples outside the robust cluster seemed to be distributed biota with massive parallel sequencing. We detected representatives
more along a gradient whose extremes typically were enriched for of several genera with a predominance of Firmicutes. The prevalence
Streptococcus on the one end and Escherichia on the other. A principal of phyla in the jejunum was different from that in the stomach
component analysis supported the above observations (Figure 7). The analysed by Andersson et al.21 using the same methodology. At the
first two principal components were roughly organized along three taxon level Streptococcus spp., Veillonella spp., Prevotella spp., Rothia
directions determined by enrichment for Prevotella, Streptococcus or spp., Haemophilus spp., Actinobacillus spp., Escherichia spp. and
Escherichia. Fusobacterium spp. were the dominating species. Our results differ
We investigated the possible association between the classification from those obtained by Zilberstain et al.27 in a study based on the
into control or patient and MDS coordinates by plotting two coor- direct aspiration and culture of jejunal fluid, describing domination
dinates at a time. Looking at such plots it seemed that controls had a of Veillonella spp., Lactobacillus spp., Proteus spp., and Bacteroides spp.
tendency to be spread out away from patients. To confirm this visual in proximal jejunum. This difference is expected, since Zilberstain
indication, we performed logistic regressions with the first five MDS used culturing as the method of choice to characterize the microbiota
coordinates as independent variables and the classification into con- and it is known that only a part of the bacterial colonizers of the gut
trol or patient as the dependent variable. The statistical models for can be cultured28,29. One previous study that used PCR-amplified 16S
BC (Supplementary Figure 2A), JSD (Supplementary Figure 2B) and rDNA clone libraries to characterize the microbial diversity of a
UniFrac (Supplementary Figure 2C) all indicated that MDS coord- jejunum biopsy from a single healthy subject found 78% Firmicutes,
inate 2 exhibited a non-significant trend towards separation of con- 13% Proteobacteria, 3% Bacteroidetes, 1% Actinobacteria, and 3%
trols from patients, p < 0.08 (BC, UniFrac) and p < 0.07 (JSD) after Fusobacteria30. Although these abundances were somewhat different
adjusting for age and gender. No evidence of association was found from our mean values, they all fell within the ranges observed in our
for weighted UniFrac. material.

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 5

 www.nature.com/scientificreports

Figure 6 | The distribution of Prevotella abundance is bimodal, indicating that samples may be naturally subdivided into two distinct subgroups
according to whether they have low or high Prevotella abundance.

454-Barcoded pyrosequencing is a powerful method to explore the on the basis of various techniques for carbohydrate breath testing36.
diversity within the human gut ecosystems but it also has some However, in a recent study Yu et al.37 demonstrated that lactulose
limitations. One of them is the inability to distinguish between live breath testing detects oro-caecal transit, not small intestinal bacterial
and dead bacteria. Using jejunum biopsies instead of jejunal fluid we overgrowth in patients with IBS.
analysed mucosa-associated bacteria, thus diminishing the possibil- The achieved results do not reveal pronounced and reproducible
ity of sequencing bacteria just passing the gastrointestinal tract. The IBS-related deviations of entire phylogenetic or functional microbial
mucosa-associated bacteria are of particular interest since they, gen- groups. The lack of apparent similarities in the taxonomy of micro-
erally speaking, are more likely to have a direct effect on the host biota in IBS patients may partially arise from the fact that the applied
through the mucosal layer than the bacteria only passing through the molecular methods, the nature and location of IBS subjects, and the
intestinal tract. We cannot exclude the contamination from oral or statistical power of previous studies have varied considerably15. It is
oesophageal flora although the probability is very low. We washed unclear whether IBS is a disorder of the small intestine or the large
the Watson capsules before opening them and our method for DNA intestine, or both. Our findings do not support a role for microbiota
extraction was designed to extract only mucosa-associated flora. of the upper small intestine in the pathogenesis of IBS. However, we
However, we found similarities at both phylum and taxon level with cannot rule out that changes in microbiota of the distal small bowel
previously described microbiota from the distal oesophagus31. might influence development of IBS or IBS symptoms. Several stud-
Franzosa et al.32 identified a subset of abundant oral microbes that ies have indicated that both epithelial barrier function and entero-
routinely survive transit to the gut, but with a minimal transcrip- endocrine function of the small intestine are important players in the
tional activity there. Although DNA from oral species was detectable crosstalk between intestinal microbes and the host that is believed to
in the gut it did not form a dominant component of that community. have an important role in the development of IBS38.
We found no significant difference in small bowel microbiota Regarding the degree of inter-individual variation of the gut
between IBS patients and healthy controls. Our results do not cor- microbiota in IBS, previous studies reported a highly significant loss
relate with some previous studies that reported differences between of variation in IBS patients39 or suggested that the microbiota of IBS
IBS patients and healthy controls in the composition of faecal micro- subjects was more heterogeneous than that of healthy controls40. We
biota14,33,34. Moreover, our findings do not support a role for SIBO in could not confirm such findings in our study.
IBS. We did not find any qualitative differences in jejunum micro- Our study lends some support to the existence of a Prevotella
biota between patients and controls although we cannot exclude enterotype in the small bowel but not the other two enterotypes
quantitative differences because the 16S amplicon cannot measure described by Arumugam et al.5 This observation is in line with
the absolute number of bacteria. The relationship between SIBO and new data suggesting that the boundaries between the enterotypes
IBS is highly inconsistent among studies35. SIBO is often diagnosed may be fuzzier than previously suggested and the communities of

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 6

 www.nature.com/scientificreports

Figure 7 | A principal component analysis of all samples. The first two principal components were roughly organized along three directions determined
by enrichment for Prevotella, Streptococcus or Escherichia.

gut bacteria may form a spectrum rather than falling into distinct Methods
groups41,42. The discrepancy can also be explained by different mate- Patients. All patients fulfilled Rome-II criteria for IBS13. A total of 35 patients (26
rials analysed in the two studies: faecal samples in the Arumugan females) with a median age of 36 (range 18–50) were investigated. Diarrhoea-
study and small bowel mucosa samples in ours. predominant IBS (D-IBS) was present in 13 patients (37%), while 9 patients (26%)
had constipation-predominant IBS (C-IBS) and 13 patients (37%) had IBS that did
It is unclear whether or not Bacteroides versus Prevotella entero- not fulfil the criteria for D-IBS or C-IBS.
types exist as distinct entities or rather represent a continuum where
the observed dietary associations occur at the extremes. Perhaps better Controls. The control group comprised 16 healthy volunteers (11 females) in whom
described as an enterogradient between abundance of Bacteroides- presence of IBS and all other functional bowel disorders had been excluded by
and Prevotella- dominant gut microbial communities, it currently medical interview and a validated questionnaire for the Rome-II symptom criteria.
The median age of the controls was 32 (range 20–48) years. Obesity was excluded in
appears that these two genera do not coexist well within the gut envir- both patients and controls. Neither patients nor controls had been treated with
onment. Organisms that are phylogenetically related and functionally antibiotics during one month prior to biopsy taking.
similar tend to coexist within the same environment consistent with
niche-driven community structures. Coexclusion of Bacteroides and Mucosa biopsy. Mucosa specimens from the proximal jejunum were taken in the
Prevotella, taxonomically and functionally similar genera, within the time period from January 2006 to December 2009 with a sterile Watson capsule in all
patients and controls. The Watson capsule was swallowed by the subject and brought
gut is an exception perhaps suggesting competition within the same by peristalsis to a position distal to the ligament of Treitz as determined by
niche43. We cannot exclude that Prevotella enterotype reflects small fluoroscopy. The Watson capsule was washed with sterile water before being opened.
bowel microbiota, as we found in the present study, and Bacteroides Biopsy samples obtained with capsules were divided into two pieces. One piece was
large bowel microbiota. The abundance of Bacteroides versus frozen in liquid nitrogen and stored at 280uC for future DNA extraction. The other
piece was fixed in formalin and mounted in paraffin blocks for histopathological
Prevotella may be an oversimplification of alternative states of the analysis. The presence of villus atrophy and other significant abnormalities were
gut microbiota in response to diet44. excluded in all biopsies.
Our study is the first to characterize small bowel microbiota with
massive parallel sequencing. We did not confirm significant differ- DNA extraction. Extraction of total genomic DNA (gDNA) from frozen biopsies was
ences in small bowel mucosa-associated microbiota between patients performed with the DNeasy Blood & Tissue Kit (Qiagen, Germany). Biopsy samples
were homogenized with a pestle in 1.5 ml tubes containing 200 ml freezing buffer.
with IBS and healthy individuals although we identified candidates 100 ml of the homogenate was added to 200 ml lyses buffer (180 ml ATL buffer, 20 ml
for potentially being differentially expressed between controls and Proteinase K) and incubated overnight at 56uC in a shaking incubator. A negative
patients. Further studies are required to verify our results. extraction control was included for each batch of DNA extraction. After extraction of

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 7

 www.nature.com/scientificreports

gDNA following the manufacturer’s instruction the gDNA was eluted in 100 ml 13. Thompson, W. G. et al. Functional bowel disorders and functional abdominal
Buffer AE and stored at 220uC until further usage. pain. Gut 45 Suppl 2, II43–47 (1999).
14. Kassinen, A. et al. The fecal microbiota of irritable bowel syndrome patients differs
PCR and template preparation for 454 sequencing. The protocol for barcoded significantly from that of healthy subjects. Gastroenterology 133, 24–33 (2007).
454-pyrosequencing has been adapted from Andersson et al.21 To create barcoded 15. Salonen, A., de Vos, W. M. & Palva, A. Gastrointestinal microbiota in irritable bowel
sequencing templates, PCR was performed using the forward primer 341f syndrome: present state and perspectives. Microbiology 156, 3205–3215 (2010).
(GCCTTGCCAGCCCGCTCAGCCTACGGGNGGCWGCAG) and a barcoded reverse 16. Jeffery, I. B. et al. An irritable bowel syndrome subtype defined by species-specific
primer 805r (GCCTCCCTCGCGCCATCAGACTACHVGGGTATCTAATCC) alterations in faecal microbiota. Gut 61, 997–1006, doi:10.1136/gutjnl-2011-
targeting the V4 region of the 16S rRNA gene. For each sample, a PCR mix was 301501 (2012).
prepared (triplicate) containing 1x PCR buffer, 1 mM dNTP’s (Finnzymes, Finland), 17. Kerckhoffs, A. P. et al. Lower Bifidobacteria counts in both duodenal mucosa-
1 U Phusion High-Fidelity DNA Polymerase (Finnzymes, Finland), 0.4 mM of each associated and fecal microbiota in irritable bowel syndrome patients. World J
primer and 1–5 ml template DNA. For the PCR reaction the following conditions were Gastroenterol 15, 2887–2892 (2009).
applied: after an initial denaturation step (95uC, 5 min), 30 cycles of 95uC for 40 sec, 18. Pimentel, M., Chow, E. J. & Lin, H. C. Normalization of lactulose breath testing
55uC for 40 sec and 72uC for 60 sec were followed by an final elongation step with 72uC correlates with symptom improvement in irritable bowel syndrome. a double-blind,
for 10 min. Following agarose gel electrophoresis (1% agarose in TBE buffer randomized, placebo-controlled study. Am J Gastroenterol 98, 412–419 (2003).
containing gel red (Sigma Aldrich, Sweden), PCR products with the right size (approx. 19. Stotzer, P. O., Bjornsson, E. S. & Abrahamsson, H. Interdigestive and postprandial
500 bp) were excised and purified using the QIAquick gel extraction kit (Qiagen, motility in small-intestinal bacterial overgrowth. Scand J Gastroenterol 31,
Germany). The DNA concentration was assessed using the Qubit Fluorometer 875–880 (1996).
(Invitrogen, USA). After an initial concentration determination, the PCR products 20. Bouhnik, Y. et al. Bacterial populations contaminating the upper gut in patients
were diluted to a concentration between 3–5 ng/ml and then pooled for the 454- with small intestinal bacterial overgrowth syndrome. Am J Gastroenterol 94,
sequencing run. The final concentration of the DNA pool was measured for further 1327–1331 (1999).
dilutions for the emulsion PCR (emPCR) in the 454 workflow. Pyrosequencing was 21. Andersson, A. F. et al. Comparative analysis of human gut microbiota by barcoded
performed according to the Genome Sequencer FLX System Methods Manual. pyrosequencing. PLoS One 3, e2836 (2008).
22. Lundin, D. et al. Which sequencing depth is sufficient to describe patterns in
bacterial alpha- and beta-diversity? Environ Microbiol Rep 4, 367–372,
Bioinformatics pipeline. Pyrosequenced amplicon reads were de-multiplexed by
doi:10.1111/j.1758-2229.2012.00345.x (2012).
stripping barcode and primer from each read. Reads which had one or more
23. White, J. R., Nagarajan, N. & Pop, M. Statistical methods for detecting
mismatches in barcode or primer were discarded. The remaining reads were quality
differentially abundant features in clinical metagenomic samples. PLoS Comput
filtered using the fastq_filter command of USEARCH v7.0.100145 with parameters set
Biol 5, e1000352, doi:10.1371/journal.pcbi.1000352 (2009).
to truncate reads to 200 basepairs whilst discarding shorter reads, and with maximum
expected errors set to 0.5. The script pick_open_reference_otus.py from QIIME 24. Caliński, T. & Harabasz, J. A dendrite method for cluster analysis. Commun Stat 3,
v1.7.0 was used to pick operational taxonomic units (OTUs)46. For this step the 1–27 (1974).
Greengenes May 2013 reference OTUs database at 97% identity was employed47. 25. Rousseeuw, P. J. Silhouettes: a graphical aid to the interpretation and validation of
Chimera filtering was performed on the representative OTU reads using the cluster analysis. J Comput Appl Math 20, 53–65 (1987).
uchime_ref command of USEARCH against the ChimeraSlayer ‘‘Gold’’ database 26. Tibshirani, R. & Walther, G. Cluster validation by prediction strength. J Comput
from the Broad Microbiome Utilities r2011051948. The resulting data was analysed Graph Stat 14, 511–528 (2005).
using R [http://www.R-project.org/] and the packages phyloseq 1.4.5, Metastats23, 27. Zilberstein, B. et al. Digestive tract microbiota in healthy volunteers. Clinics (Sao
cluster 1.14.4, fpc 2.1-7, ggplot2 0.9.3.1, plyr 1.8.1, grid 3.0.1 and VGAM 0.9-3. Paulo) 62, 47–54 (2007).
Analyses at phylum and genus level were performed by: (1) aggregating all OTUs 28. Amann, R. I., Ludwig, W. & Schleifer, K. H. Phylogenetic identification and in situ
which had identical classification at the given taxonomic level (suggested detection of individual microbial cells without cultivation. Microbiol Rev 59,
classifications were heeded, e.g. [Prevotella] was considered identical with Prevotella), 143–169 (1995).
(2) normalizing samples by their total abundance, (3) discarding unclassified taxa and 29. Vaughan, E. E. et al. A molecular view of the intestinal ecosystem. Curr Issues
taxa with mean relative abundance lower than 0.01%. Step (3) was also performed on Intest Microbiol 1, 1–12 (2000).
OTUs before analysis. Low abundance taxa were discarded to avoid having sampling 30. Wang, M., Ahrne, S., Jeppsson, B. & Molin, G. Comparison of bacterial diversity
depth effects influencing the analysis. Also, since the sample size was fairly small the along the human intestinal tract by direct cloning and sequencing of 16S rRNA
analysis focused on taxa that were expressed in most samples in order to get genes. FEMS Microbiol Ecol 54, 219–231, doi:10.1016/j.femsec.2005.03.012 (2005).
meaningful results. 31. Pei, Z. et al. Bacterial biota in the human distal esophagus. Proc Natl Acad Sci U S A
101, 4250–4255 (2004).
32. Franzosa, E. A. et al. Relating the metatranscriptome and metagenome of the
Ethical considerations. The Regional Board of Research Ethics in Stockholm human gut. Proc Natl Acad Sci U S A 111, E2329–2338, doi:10.1073/
approved all parts of the study. Informed consent was obtained from all patients and pnas.1319284111 (2014).
controls and all methods were carried out in accordance with approved guidelines. 33. Krogius-Kurikka, L. et al. Microbial community analysis reveals high level
phylogenetic alterations in the overall gastrointestinal microbiota of diarrhoea-
predominant irritable bowel syndrome sufferers. BMC Gastroenterol 9, 95 (2009).
1. Savage, D. C. Microbial ecology of the gastrointestinal tract. Annu Rev Microbiol 34. Lyra, A. et al. Diarrhoea-predominant irritable bowel syndrome distinguishable
31, 107–133 (1977). by 16S rRNA gene phylotype quantification. World J Gastroenterol 15, 5936–5945
2. Quigley, E. M. & Abu-Shanab, A. Small intestinal bacterial overgrowth. Infect Dis (2009).
Clin North Am 24, 943–959, viii-ix (2010). 35. Ford, A. C., Spiegel, B. M., Talley, N. J. & Moayyedi, P. Small intestinal bacterial
3. Gorbach, S. L. & Bengt E. Gustafsson memorial lecture. Function of the normal overgrowth in irritable bowel syndrome: systematic review and meta-analysis.
human microflora. Scand J Infect Dis Suppl 49, 17–30 (1986). Clin Gastroenterol Hepatol 7, 1279–1286 (2009).
4. Qin, J. et al. A human gut microbial gene catalogue established by metagenomic 36. Bures, J. et al. Small intestinal bacterial overgrowth syndrome. World J
sequencing. Nature 464, 59–65 (2010). Gastroenterol 16, 2978–2990 (2010).
5. Arumugam, M. et al. Enterotypes of the human gut microbiome. Nature 473, 37. Yu, D., Cheeseman, F. & Vanner, S. Combined oro-caecal scintigraphy and
174–180, doi:10.1038/nature09944 (2011). lactulose hydrogen breath testing demonstrate that breath testing detects oro-
6. Willing, B. et al. Twin studies reveal specific imbalances in the mucosa-associated caecal transit, not small intestinal bacterial overgrowth in patients with IBS. Gut
microbiota of patients with ileal Crohn’s disease. Inflamm Bowel Dis 15, 653–660 60, 334–340 (2011).
(2009). 38. Ohman, L., Tornblom, H. & Simren, M. Crosstalk at the mucosal border:
7. Vijay-Kumar, M. et al. Metabolic syndrome and altered gut microbiota in mice importance of the gut microenvironment in IBS. Nat Rev Gastroenterol Hepatol,
lacking Toll-like receptor 5. Science 328, 228–231 (2010). doi:10.1038/nrgastro.2014.200 (2014).
8. Bäckhed, F. et al. The gut microbiota as an environmental factor that regulates fat 39. Codling, C., O’Mahony, L., Shanahan, F., Quigley, E. M. & Marchesi, J. R. A
storage. Proc Natl Acad Sci U S A 101, 15718–15723 (2004). molecular analysis of fecal and mucosal bacterial communities in irritable bowel
9. El-Aidy, S., van den Bogert, B. & Kleerebezem, M. The small intestine microbiota, syndrome. Dig Dis Sci 55, 392–397 (2010).
nutritional modulation and relevance for health. Curr Opin Biotechnol 32, 14–20, 40. Rajilic-Stojanovic, M., Smidt, H. & de Vos, W. M. Diversity of the human
doi:10.1016/j.copbio.2014.09.005 (2015). gastrointestinal tract microbiota revisited. Environ Microbiol 9, 2125–2136 (2007).
10. Zoetendal, E. G. et al. The human small intestinal microbiota is driven by rapid 41. Yong, E. Gut microbial ‘enterotypes’ become less clear-cut. Nature, doi:10.1038/
uptake and conversion of simple carbohydrates. ISME J 6, 1415–1426, nature.2012.10276 (2012).
doi:10.1038/ismej.2011.212 (2012). 42. Koren, O. et al. A guide to enterotypes across the human body: meta-analysis of
11. Booijink, C. C. et al. High temporal and inter-individual variation detected in the microbial community structures in human microbiome datasets. PLoS Comput
human ileal microbiota. Environ Microbiol 12, 3213–3227, doi:10.1111/j.1462- Biol 9, e1002863, doi:10.1371/journal.pcbi.1002863 (2013).
2920.2010.02294.x (2010). 43. Faust, K. et al. Microbial co-occurrence relationships in the human microbiome.
12. Parkes, G. C. et al. Distinct microbial populations exist in the mucosa-associated PLoS Comput Biol 8, e1002606, doi:10.1371/journal.pcbi.1002606 (2012).
microbiota of sub-groups of irritable bowel syndrome. Neurogastroenterol Motil 44. Bushman, F. D., Lewis, J. D. & Wu, G. D. Diet, gut enterotypes and health: is there
24, 31–39, doi:10.1111/j.1365-2982.2011.01803.x (2012). a link? Nestle Nutr Inst Workshop Ser 77, 65–73, doi:10.1159/000351385 (2013).

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 8

 www.nature.com/scientificreports

45. Edgar, R. C. UPARSE: highly accurate OTU sequences from microbial amplicon massive parallel sequencing, analyzed the data and revised the paper, K.Z. extracted DNA
reads. Nat Methods 10, 996–998, doi:10.1038/nmeth.2604 (2013). and barcoded samples and revised the paper, G.S. designed the study, provided the
46. Caporaso, J. G. et al. QIIME allows analysis of high-throughput community laboratory resources and revised the paper, W.Y. analyzed the data and revised the paper,
sequencing data. Nat Methods 7, 335–336, doi:10.1038/nmeth.f.303 (2010). L.E. designed the study, provided the laboratory resources and revised the paper, G.L.
47. DeSantis, T. Z. et al. Greengenes, a chimera-checked 16S rRNA gene database and designed the study, obtained funding, collected the material, analyzed the data and wrote
workbench compatible with ARB. Appl Environ Microbiol 72, 5069–5072, the paper.
doi:10.1128/aem.03006-05 (2006).
48. Haas, B. J. et al. Chimeric 16S rRNA sequence formation and detection in Sanger
and 454-pyrosequenced PCR amplicons. Genome Res 21, 494–504, doi:10.1101/ Additional information
gr.112730.110 (2011). Supplementary information accompanies this paper at http://www.nature.com/
scientificreports
Competing financial interests: The authors declare no competing financial interests.
How to cite this article: Dlugosz, A. et al. No difference in small bowel microbiota between
Acknowledgments patients with irritable bowel syndrome and healthy controls. Sci. Rep. 5, 8508; DOI:10.1038/
The study was supported by grants from Ruth and Richard Julin’s Foundation and
srep08508 (2015).
Foundation Olle Engkvist Byggmästare. AD was the recipient of a Rome Foundation
Fellowship in Functional GI and Motility Disorders. None of the funding sources had any
involvement with the study. This work is licensed under a Creative Commons Attribution 4.0 International
License. The images or other third party material in this article are included in the
article’s Creative Commons license, unless indicated otherwise in the credit line; if
Author contributions the material is not included under the Creative Commons license, users will need
A.D. designed the study, collected the material, analyzed the data and wrote the paper, B.W. to obtain permission from the license holder in order to reproduce the material. To
performed statistical analyses, analyzed the data and wrote the paper, E.L. performed the view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

SCIENTIFIC REPORTS | 5 : 8508 | DOI: 10.1038/srep08508 9