HOST MICROBE INTERACTIONS Molecular Analysis of Microbiota Along the Digestive Tract of Juvenile Atlantic Salmon (Salmo salar L.) P. Navarrete & R. T. Espejo & J. Romero Received: 8 February 2008 / Accepted: 28 August 2008 / Published online: 17 September 2008 # Springer Science + Business Media, LLC 2008 Abstract Dominant bacterial microbiota of the gut of juvenile farmed Atlantic salmon was investigated using a combination of molecular approaches. Bacterial community composition from the stomach, the pyloric caeca, and the intestine was assessed by extracting DNA directly from each gut compartment. Temporal temperature gradient gel electrophoresis (TTGE) analysis of 16S ribosomal DNA (rDNA) amplicons showed very similar bacterial composi- tions throughout the digestive tract. Band sequencing revealed a narrow diversity of species with a dominance of Pseudomonas in the three compartments. However, cloning revealed more diversity among the Pseudomonas sequences. To confirm these results, we analyzed the bacterial community by amplifying the variable 16S–23S rDNA intergenic spacer region (ITS). Similar ITS profiles were observed among gastrointestinal compartments of salmon, confirming the TTGE results. Moreover, the dominant ITS band at 650 bp, identified as Pseudomonas, was observed in the ITS profile from fish collected in two seasons (July 2003 and 2004). In contrast, aerobic culture analysis revealed Shewanella spp. as the most prevalent isolate. This discrepancy was resolved by evaluating 16S rDNA and ITS polymerase chain reaction amplification efficiency from both Shewanella and Pseudomonas iso- lates. Very similar efficiencies were observed in the two bacteria. Hence, this discrepancy may be explained by preferential cultivation of Shewanella spp. under the experimental conditions. Also, we included analyses of pelleted feed and the water influent to explore environmen- tal influences on the bacterial composition of the gut microbiota. Overall, these results indicate a homogeneous composition of the bacterial community composition along the gastrointestinal tract of reared juvenile salmon. This community is mainly composed of Pseudomonas spp., which could be derived from water influent and may be selectively associated with salmon in this hatchery. Introduction The gastrointestinal tract is a composite ecosystem con- taining a complex and dynamic consortium of micro- organisms, usually called microbiota, which appear to play a key role in the nutrition and health of the host [4, 16]. Evidence for the role of microbiota in fish was recently revealed [5, 34, 35]. Using germ-free zebrafish, these reports showed that gut microbiota might be involved in important processes like stimulation of epithelial prolifera- tion, promotion of nutrient metabolism, and innate immune responses. A key aspect of these results was the specificity of the host response at the gene expression level, which depended on the bacterial composition of the digestive tract [34]. Therefore, it may be relevant to know the composition of microbiota of reared fish, especially salmonids, which constitute an important economic industry in Chile. Current knowledge of the diversity in the bacterial composition of salmon microbiota is largely based on the use of classical culture-dependent techniques and the contribution of this approach has been reviewed [3, 10, 19, 36]. However, it has been shown that a large proportion of bacteria are not isolated on traditional agar substrates [1], and it is currently accepted that these culture-based methods detect only a small fraction of bacteria present in the gut [46]. As a possible alternative, molecular methods allow Microb Ecol (2009) 57:550–561 DOI 10.1007/s00248-008-9448-x P. Navarrete : R. T. Espejo : J. Romero (*) Laboratorio de Biotecnología, Instituto de Nutrición y Tecnología de los Alimentos, Universidad de Chile, El Líbano 5524, Macul, P.O. Box 138-11, Santiago, Chile e-mail: [email protected]
Transcript
HOST MICROBE INTERACTIONS
Molecular Analysis of Microbiota Along the Digestive Tractof Juvenile Atlantic Salmon (Salmo salar L.)
P. Navarrete & R. T. Espejo & J. Romero
Received: 8 February 2008 /Accepted: 28 August 2008 /Published online: 17 September 2008# Springer Science + Business Media, LLC 2008
Abstract Dominant bacterial microbiota of the gut ofjuvenile farmed Atlantic salmon was investigated using acombination of molecular approaches. Bacterial communitycomposition from the stomach, the pyloric caeca, and theintestine was assessed by extracting DNA directly fromeach gut compartment. Temporal temperature gradient gelelectrophoresis (TTGE) analysis of 16S ribosomal DNA(rDNA) amplicons showed very similar bacterial composi-tions throughout the digestive tract. Band sequencingrevealed a narrow diversity of species with a dominanceof Pseudomonas in the three compartments. However,cloning revealed more diversity among the Pseudomonassequences. To confirm these results, we analyzed thebacterial community by amplifying the variable 16S–23SrDNA intergenic spacer region (ITS). Similar ITS profileswere observed among gastrointestinal compartments ofsalmon, confirming the TTGE results. Moreover, thedominant ITS band at 650 bp, identified as Pseudomonas,was observed in the ITS profile from fish collected in twoseasons (July 2003 and 2004). In contrast, aerobic cultureanalysis revealed Shewanella spp. as the most prevalentisolate. This discrepancy was resolved by evaluating 16SrDNA and ITS polymerase chain reaction amplificationefficiency from both Shewanella and Pseudomonas iso-lates. Very similar efficiencies were observed in the twobacteria. Hence, this discrepancy may be explained bypreferential cultivation of Shewanella spp. under theexperimental conditions. Also, we included analyses ofpelleted feed and the water influent to explore environmen-
tal influences on the bacterial composition of the gutmicrobiota. Overall, these results indicate a homogeneouscomposition of the bacterial community composition alongthe gastrointestinal tract of reared juvenile salmon. Thiscommunity is mainly composed of Pseudomonas spp.,which could be derived from water influent and may beselectively associated with salmon in this hatchery.
Introduction
The gastrointestinal tract is a composite ecosystem con-taining a complex and dynamic consortium of micro-organisms, usually called microbiota, which appear toplay a key role in the nutrition and health of the host [4,16]. Evidence for the role of microbiota in fish was recentlyrevealed [5, 34, 35]. Using germ-free zebrafish, thesereports showed that gut microbiota might be involved inimportant processes like stimulation of epithelial prolifera-tion, promotion of nutrient metabolism, and innate immuneresponses. A key aspect of these results was the specificityof the host response at the gene expression level, whichdepended on the bacterial composition of the digestive tract[34]. Therefore, it may be relevant to know the compositionof microbiota of reared fish, especially salmonids, whichconstitute an important economic industry in Chile.
Current knowledge of the diversity in the bacterialcomposition of salmon microbiota is largely based on theuse of classical culture-dependent techniques and thecontribution of this approach has been reviewed [3, 10,19, 36]. However, it has been shown that a large proportionof bacteria are not isolated on traditional agar substrates [1],and it is currently accepted that these culture-based methodsdetect only a small fraction of bacteria present in the gut[46]. As a possible alternative, molecular methods allow
P. Navarrete :R. T. Espejo : J. Romero (*)Laboratorio de Biotecnología,Instituto de Nutrición y Tecnología de los Alimentos,Universidad de Chile,El Líbano 5524, Macul, P.O. Box 138-11, Santiago, Chilee-mail: [email protected]
bacterial identification independent of their ability to growin synthetic media. These methods have been used to studythe bacterial composition of haddock larvae [18], halibutlarvae [23], and the intestinal contents of coho salmon [42]and rainbow trout [24]. They also allowed for theidentification of Acinetobacter junii and a novel Mycoplas-ma phylotype as the predominant bacterial populations infarmed and wild adult salmon [20].
The salmon gut is composed of separate anatomicalcompartments: the esophagus, a U-shaped stomach, thepyloric caeca, and the intestine (Fig. 1a) [47]. As in manyfish, the pyloric caeca are blind appendages attached to theintestine in the region near the stomach [6]. The pH isreported to be between 3.0 and 4.5 in the stomach, between7.0 and 7.5 in the pyloric caeca, and a pH of 9.0 has beenreported in the distal part of the intestine [32, 37]. Ourhypothesis was that anatomical and physiological differ-ences along the gut would favor a specific microbiota ineach compartment. There are no reports about the molec-ular analysis of microbiota associated with the differentcompartments of the gastrointestinal tract of salmon, itscomposition, its complexity, or its stability, especiallyfollowing dietary changes or treatment with antibiotics,which are routine practices in aquaculture. Since differentmicroorganisms induce different responses in the host andsome confer potential benefits [34], knowledge of thebacteria commonly associated with different gastrointestinalcompartments may be useful for manipulating microbiotaas a strategy to improve nutrition or prevent pathogenicinfection. In this study, we used a ribosomal marker-basedapproach to determine the bacterial composition of thedifferent compartments of the gastrointestinal tract ofjuvenile Atlantic salmon (Salmo salar). Bacterial composi-tion was determined by analysis of 16S ribosomal DNA(rDNA) amplicons using temporal temperature gradient gelelectrophoresis (TTGE) and by cloning and sequencing themain amplicons. To assess diversity at the intraspecieslevel, the main ribosomal intergenic space region (ITS)amplicons were identified by sequencing. These resultswere then compared to those obtained by culture to get amore comprehensive overview of the bacterial populationspresent. Since bacterial communities were extracted fromeach gut compartment that contained epithelium and feeddigesta, the microbiota analyzed was a combination of bothautochthonous (able to colonize the epithelial surface ormucus of the host gut) and allochthonous (transient orassociated with digesta) bacteria. The main objective of thepresent work was to examine bacterial composition alongthe digestive tract of salmon with special emphasis on thedominant bacterial population present in each gut compart-ment. Also, the analyses of water influent and feed wereincluded to explore the environmental influences. To ourknowledge, the present work is the first examination of
bacterial microbiota of the entire gut by molecular analysisof 16S rDNA and ITS.
Methods
Samples Collection and Processing
Salmo salar juvenile specimens of 1–30 g were collectedfrom a hatchery (latitude 32° S, Chile) in two seasons:between July and October, 2003 and during July, 2004. Inboth seasons, fish were reared in freshwater with atemperature of 16°C year-round and normal densities wereabout 50 kg/m3. The fish collected were healthy juvenilesalmon that had never been treated with antibiotics. Theywere fed with conventional pelleted feed without probiotic,prebiotic, immunomodulatory, or inhibitory agents. Theaverage daily intake of pelleted feed from EWOS was 2.8%of body weight for all fish. This was accomplished bymanual feeding seven to eight times throughout the day andobservation of fish to check diet acceptance and satiety. Thefish were anesthetized using benzocaine and killed by ablow to the head. Collected salmon were size-graded. Thebiggest salmon (30 g; collected in July 2004) were analyzedindividually, meaning that every fish was dissected sepa-rately from the others and every gut compartment (stomach,pyloric caeca, and intestine) was analyzed separately.Smaller fish were divided into eight groups composed often specimens each, with an average weight of 1.5, 3, 4,and 14±0.5 g (July–October 2003) and 4, 14, 17, and 23±0.5 g (July 2004). The entire gut of ten specimens per groupwas homogenized and referred to as pooled samples.Pooling samples was performed because it is a commonpractice to study the gut microbiota in fish [2, 21] andprevious studies showed that individual microbiota are wellrepresented by pooled samples of individual gut microbiota[42]. Dead fish were placed in sterile plastic containers andtransported to the laboratory on ice. Samples wereprocessed immediately upon arrival in the laboratory,usually within 2 h. Gastrointestinal samples were obtainedby aseptically dissecting the fish and carefully extractingthe entire gastrointestinal tract under a stereomicroscope.For individual analysis of the biggest fish (30 g), the threeanatomical compartments (U-shaped stomach, a developedpyloric region, and intestine) were carefully separatedduring dissection. Each gut section (stomachs, pyloriccaeca, and intestines) was sliced and weighed separately,and an equal amount of cold sterile phosphate-bufferedsaline was added. For the smallest fish, the entire guts fromten specimens of the same weight were sliced andhomogenized together. These sliced guts were subsequentlyhomogenized in an ice bath with vigorous vortexing for3 min. Since all samples included gut epithelia, mucus, and
Bacterial Microbiota in the Salmon Gut 551
digesta, the microbiota analyzed was a combination of bothautochthonous (able to colonize the epithelial surface ormucus of the host gut) and allochthonous bacteria (transient).
Simultaneously, freshwater samples were obtained di-rectly from the water influent corresponding to thehatchery’s water source and the feed used at the hatcherywere collected and transported on ice. Samples wereprocessed immediately upon arrival in the laboratory.
DNA Extraction and Purification
DNA from the gastrointestinal compartments was obtainedfrom homogenates by lysis using sodium dodecyl sulfateand incubation at 70°C. The lysates were subsequentlyextracted with phenol/chloroform and precipitated withethanol as previously described [41]. A final purificationwas carried out using Wizard DNA Clean Up (Promega,
Figure 1 TTGE profiles of 16SrDNA amplification productsfrom Atlantic salmon gut. a Aschematic view of the salmondigestive tract. b TTGE profilesfrom stomach, pyloric caeca,and intestine of five individuals(with a body weight of 30 g).Each lane corresponds to asingle compartment, each de-rived from individual fish num-bered from 1 to 5. Commonbands showing the same migra-tion patterns are indicated (B1–B6). c TTGE profiles of theentire digestive tracts from fishof 4, 14, 17, and 23 g (lanes 4,14, 17, and 23) compared tothose from separated compart-ments obtained from a singleindividual (lanes S: stomach,PC: pyloric caeca, and I:intestine)
552 P. Navarrete et al.
Madison, WI, USA). DNA from water influent wasobtained by filtering 5 l of water through a 0.2-mm filter.Bacteria were then resuspended in TE buffer (Tris 0.01 M,EDTA 0.001 M, NaCl 0.15 M, pH 7.8), and lyses wereperformed as described above. DNA from pelleted feed wasobtained by homogenizing 15 g in TE buffer followed byDNA extraction as described above.
PCR Amplification and Analysis of the Products
Amplification of the V3–V5 region of 16S rDNA wascarried out to obtain profiles of the bacterial communitiespresent in different compartments, water influent, andpelleted feed. The extracted DNA was polymerase chainreaction (PCR)-amplified using conserved 16S rDNAbacterial domain-specific primers 341F (5′-GCCTACGGGAGGCAGCAG-3′ with GC clamps at the 5′ end) and 907R(5′-CCGTCAATTCMTTTGAGTTT-3′) as previously de-scribed [27]. PCR reactions were performed as described[42] with a reaction mixture (30 μl) containing 0.2 mM ofeach deoxynucleoside triphosphate, 0.05 U/ml PlatinumTaq DNA polymerase (Invitrogen, San Diego, CA, USA), 1×polymerase reaction buffer, 2 mM MgCl2, and 0.25 pmol/mlof each primer.
Amplification of the 16S–23S rDNA intergenic regionswas performed as described by Espejo and Romero [15]using primers G1 and L1 [22]. Semi-quantitative amplifi-cation was performed as previously described [14]. Ampli-fication of 16S rRNA together with its neighboring 16S–23S rDNA spacer was performed as described [41], usingprimers 357F and L1 [28]. For purification of this lastamplicon, the product was separated on a 1% agarose gel,and the band migrating at 1.8 kb was excised. The DNAwas subsequently extracted from the agarose gel bycentrifugation through glass wool at 12,000 g for 20 s ina microcentrifuge. The extract was diluted 1:200 in distilledwater, and 15 μl was used for amplification. Products wereanalyzed using polyacrylamide gel electrophoresis andsilver nitrate staining as previously described [13].
TGGE Analysis
PCR products obtained from 341F and 907R primers wereseparated by TTGE on a 6% (w/v) polyacrylamide gel in 1×TAE running buffer (Tris 0.04 M, acetate 0.002 M, EDTA0.001 M, pH 8.5) and a temperature gradient from 66°C to70°C [26]. Electrophoresis was run for about 20 h at 65 Vin a D-Code System (Bio-Rad, Hercules, CA, USA). Afterelectrophoresis, gels were stained for 1 h by incubationwith Sybr Green at room temperature. The dominant bandswere recognized as intense bands in each TTGE pattern andwere excised from the gel and eluted overnight in 50 μl ofMilliQ water; 1 μl was used for reamplification. To test for
the presence of similar amplicons, bands showing the samemigration in different lanes were digested with AluI andHaeIII as described below.
Restriction Fragment Length Polymorphism Analysis
Products of the 16S rDNA or ITS PCR amplification weredigested for 2 h at 37°C in 1.5 U of AluI or HaeIIIrestriction endonuclease (Invitrogen). The resulting frag-ments were subsequently analyzed by polyacrylamide gelelectrophoresis and silver nitrate staining as described.
Cloning and Sequence Analysis
PCR products were cloned into a TOPO TA vectoraccording to the procedure indicated by the manufacturer(Invitrogen). The resulting plasmids were purified using theWizard Plus SV Miniprep System (Promega). Plasmidscontaining an insert were selected, and their 16S rDNA orITS were PCR-amplified as described above and treatedwith restriction enzymes for characterization.
16S rDNA from the cloning library, the re-amplifiedbands, amplified ribosomal spacers, and bacterial isolateswere purified using Wizard PCR Preps (Promega) and thensequenced with an Applied Biosystems 310 automaticsequencer (Foster City, CA, USA). ABI Prism dyeterminator sequencing kits were used with primers 907Rand L1 for 16S rDNA genes and ITS, respectively. Se-quences were deposited in GenBank (EF587700–EF587702,DQ889968–DQ889982, DQ067264–DQ067274 ,EU794379–EU794397) and aligned with reference sequen-ces using Sequence Match software from the RibosomalDatabase Project II (RDP II) website [12]. Distance matriceswere constructed from the aligned sequences and correctedfor multiple base changes at single nucleotide positions usingthe Jukes and Cantor method included in the TREECONprogram [48]. Using the same program, a phylogenetic treewas constructed by a neighbor-joining method. Bootstrap-ping was performed using the bootstrap modus of theprogram, and values above 40% are reported.
Bacterial Counts and Cultivation
Total bacterial counts present in salmon and water sampleswere performed by epifluorescence microscopy usingacridine orange, as previously described [40]. For bacterialcultivation, serial dilutions of gut homogenates from eachentire compartment that included mucus and epithelialattached bacteria (autochthonous bacteria) and bacteriafrom the digesta content (allochthonous) were plated ontrypticase soy agar (TSA, Difco, Sparks, MD, USA), andthe plates were incubated for 10 days at 17°C in aerobicconditions. Serial dilutions of water and feed samples were
Bacterial Microbiota in the Salmon Gut 553
also plated and incubated as described above. Colonieswere counted after 10 days, and the colony-forming unit(CFU) per gram of gastrointestinal tract was calculated.Plates containing 50–200 colonies were selected for analysis.A random selection of 80 colonies from each compartment(stomach, pyloric caeca, and intestine) [38] was subculturedand characterized according to the colony morphology, 16SrDNA restriction fragment length polymorphism (RFLP),and ITS profiles, revealing coincidence among the threeapproaches. Representative isolates from each RFLP groupwere identified by 16S rDNA partial sequencing.
Statistical Analysis
Assuming 100% efficiency when the DNA template wasdoubled in each cycle, the PCR efficiency was calculated asE=10(−slope)−1, where E was the PCR efficiency. PCRefficiencies were expressed as means±SE of triplicatestandard curves. PCR efficiencies were analyzed with apaired t test, and differences were concluded to besignificant when p<0.05.
Results
Composition of Bacterial Communities Assessedby PCR-TTGE
The bacterial composition of stomach, pyloric caeca, andintestine from ten juvenile (30 g) salmon was analyzed by16S rDNA PCR-TTGE. This approach revealed a simpleand similar bacterial composition in each gastrointestinalcompartment for all individuals. Five bands were observedin the stomach samples, and two to three bands wereobserved in the pyloric caeca and intestinal samples(Fig. 1b). TTGE bands with the same migration patternshowed the same restriction profile (AluI and HaeIII)indicating that they corresponded to the same sequence(data not shown). At least one band with the samemigration pattern was sequenced for identification (Table 1).The main bands (B1 and B2) obtained from stomach,pyloric caeca, and intestine samples corresponded toPseudomonas spp. Three bands observed in the stomachsamples corresponded to 16S rDNA of chloroplasts fromGlycine max (bands B3 and B4) and chloroplasts ormitochondrion from Zea mays (band B5), derived fromthe vegetal material included in the pelleted feed, whichcontained approximately 10% soybean meal and 5% corngluten. Also, intestinal samples showed the presence of aweak band (band B6) with a 16S rDNA sequence related toComamonas sp. This band was detected in 70% (7/10) ofthe fish, showing that this species, though not alwayspresent, was a common inhabitant of the intestine.
The presence of Pseudomonas was examined in fishsmaller than 30 g. The TTGE profiles obtained from these fish(4 to 23 g) did not differ from the TTGE of 30 g (Fig. 1c).Restriction enzyme analysis confirmed that band B1 in thesesmaller fish also corresponded to Pseudomonas. Thisassessment also revealed the presence of an additional band(band B7) that was closely related to Stenotrophomonas sp.
Diversity Assessed by 16S rDNA Cloning
To study the molecular diversity of bacteria present in thesalmon gut, the 16S rDNA amplicons of the stomachsample showing the larger complexity in TTGE werecloned. Eighty clones were analyzed by RFLP using AluIand HaeIII, resulting in six RFLP groups (Table 1). Themost abundant group corresponded to soybean chloroplasts(35% of the clones). Another three RFLP groups (I–III),representing 30%, 20%, and 9% of the clones correspondedto the genus Pseudomonas. Two other RFLP groups(IV–V), together representing 6% of the clones analyzed,corresponded to Acinetobacter. Phylogenetic analysisshown in Fig. 2 illustrates the molecular diversity amongPseudomonas sequences and revealed that Pseudomonassequences from RFLP groups III and I were similar toTTGE band B1 and band B2, respectively.
Composition of Bacterial Communities Assessedby ITS Analysis
Coincident with the TTGE analysis, ITS profiles from gutcompartments were similar among individuals, showing adominance of a widespread band (650 bp) in all gastroin-testinal compartments (Fig. 3). However, the presence ofseveral weaker bands in the ITS profiles may indicate morediversity in microbiota composition. Sequence analysis ofthe dominant band using BlastN indicated that the closestmatch—with 98% identity—corresponded to Pseudomonasfluorescens (accession number EF5877001) and showed thesame organization of tRNAs (tRNAIle and tRNAAla) and non-coding regions as whole genome-sequenced Pseudomonas.
To test if the presence of Pseudomonas corresponded toa particular or a general observation, ITS analysis wasperformed in gut samples from fish collected from the samehatchery but during different seasons (Fig. 4). The presenceof the dominant 650 bp band and the similar ITS-RFLPprofiles among the fish suggests that a dominant bacterialpopulation related to Pseudomonas was present in salmonregardless of the season (Fig. 4).
Bacterial Counts
The average total bacterial density yielded 1×107, 8×106,and 5×107 bacteria/g of stomach, pyloric caeca, and
554 P. Navarrete et al.
Table 1 Nearest-match identification of 16S rDNA sequences obtained from TTGE bands, cloning approaches, and bacterial isolates fromsalmon gut compartments, water influent, and pelleted feed, with known sequences in the RDP II database
Name RFLPgroupa
Accessionnumber
Percentidentity
Affiliation phylum/class Closest sequence Gram
TTGE bands from individualsBand 1 (B1) DQ889976 99 Proteobacteria/γ-
intestine, respectively. The average counts of cultivablebacteria were 2.2×104, 3.1×103, and 1.9×105 CFU perstomach, pyloric caeca, and intestine, respectively. The totalbacterial count from the water supply was 1×106 bacteria/ml, whereas the cultivable count was 6×102 CFU/ml. Inpelleted feed samples, 4.3×102 CFU/g was obtained.Considering all samples assessed, the cultivability was <1%.
Composition of Bacterial Communities Assessedby Analysis of Cultivable Bacteria
Eighty isolates from each gut compartment were character-ized by 16S rDNA RFLP. Five groups were obtained
according to the similarity of their patterns. Representativeisolates from each group were identified by 16S rDNApartial sequencing (Table 2). Among them, the isolatesclustering with Shewanella were most frequently obtainedin the gut compartments, whereas Pseudomonas was onlyrecovered from the intestine at very low levels. Othergenera such as Microbacterium, Cellulomonas, and Serra-tia were observed with different relative abundances(Table 2).
The ITS amplicons from these isolates were comparedwith the previous ITS profiles obtained after amplificationof DNA extracted directly from the gut compartments(Fig. 3b). ITS bands with similar electrophoretic migration
Figure 2 Phylogenetic tree of16S rDNA sequences obtainedfrom juvenile Atlantic salmon.Neighbor-joining phylogenetictree showing the relationshipbetween sequences retrievedfrom the TTGE profiles, clones,and their closest relativesequences deposited in the RDPII database. The tree was con-structed based on the 341–907region of the 16S rDNA genes,using TREECON version 1.3b.A bootstrap analysis was per-formed with 100 repetitions, andvalues greater than 40% areshown
a Abundance (%) of RFLP group: I: 30%, II: 20%, III: 9%, IV: 3%, V: 3%, and VI: 35%
556 P. Navarrete et al.
to ITS from Shewanella (500 bp) were observed in salmongut ITS profiles. Pseudomonas isolates showed ITS bandswith similar size (650 bp) and sequence (EF587700–EF587702) to that previously observed in all compartmentsof the salmon gut. Altogether, these observations confirmthat Pseudomonas is abundant and common in thesesalmon guts.
PCR Amplification Efficiency for Prevailing Bacteria
The discrepancy between the results from molecular andcultural approaches may be attributed to either a low
culturing efficiency of Pseudomonas or a low PCRamplification efficiency of Shewanella. In order to resolvethis issue, we evaluated the matching of primers in silicoand PCR amplification efficiency for both bacteria. Primers341 and 907 used for 16S amplification were revised byusing the Probe Match tool in the RDP II database. Noamplification preference was inferred since both primersmatched correctly with 70% of the sequences for Pseudo-monas and Shewanella. Primers G1 and L1 used for ITSamplification were evaluated in the GenBank Genomedatabase using 16S and 23S rDNA present in the sequencedgenomes of Pseudomonas and Shewanella (13 strains
Figure 3 Analysis of ITS(16S–23S rDNA intergenicspacer) amplification productsand their RFLP profiles fromAtlantic salmon (S. salar). a ITSprofiles from stomach, pyloriccaeca, and intestine of threeindividuals from a total of tenanalyzed, with a body weight of30 g each. The number in eachlane corresponds to each indi-vidual. b ITS profiles of stom-ach, pyloric caeca, and intestinefrom a single fish (30 g; lanes S:stomach, PC: pyloric caeca, andI: intestine) compared to ITSprofiles of bacterial isolates(lanes Ps: Pseudomonas, Sw:Shewanella, Mb: Microbacte-rium, Ce: Cellulomonas). xindicates the presence of acommon band between isolatesand the direct profiles
Bacterial Microbiota in the Salmon Gut 557
each). Both primers had identical sequence targets in bothbacteria. Simultaneously, a semi-quantitative PCR approachshowed similar amplification efficiency for both bacteria.ITS amplification efficiency was 53.4±2.05% for Pseudo-monas and 55.3±3.3% for Shewanella, (t test; p=0.43997),whereas 16S rDNA amplification efficiency was 89.3±2.4% for Pseudomonas and 89.7±3.0% for Shewanella(t test; p=0.85508).
Origin of Dominant Pseudomonas Retrieved from SalmonDigestive Tracts
To determine the possible sources of Pseudomonas foundin the digestive tract of juvenile salmon, bacterial DNAfrom water influent and pelleted feed was analyzed byPCR-TTGE. All bands from water influent and pelletedfeed TTGE profiles were excised and identified bysequencing (Table 1). The results showed that the dominantbacterial populations from water and feed are different fromthose in the digestive tract of salmon and Pseudomonas wasnot detected in these samples. The water sample wasdominated by bacteria belonged to Flavobacteria (Class)whereas pelleted feed was dominated by bacteria belong toFusobacteria, Bacilli, and Actinobacteria (Class; Table 1).Bacterial isolates recovered from pelleted feed showed onlyGram-positive bacteria, whereas only Gram-negative bac-teria were retrieved from water influent (Table 1). Indeed, aPseudomonas bacterium was isolated from water influentshowing 97% identity with band 1 retrieved from thesalmon digestive tract (Table 1).
Discussion
In this study, we analyzed the bacterial composition alongthe gastrointestinal tract of juvenile salmon using differentmolecular methods. The emphasis was placed on thedominant bacterial populations because they may have aspecific function associated with each anatomical compart-ment of the digestive tract. A culture approach wasincluded to obtain a comprehensive visualization of themicrobiota.
The combined results of the molecular fingerprints fromPCR-TTGE and ITS and the results of the cloning analysesclearly demonstrated that bacterial composition along thegut of juvenile salmon was similar and dominated byPseudomonas spp. In order to rule out the possibility thatthis narrow diversity may be due to primer bias, 16S rDNAprimers were analyzed using the informatics tool ProbeMatch in RDP II. This database contains over 450,000sequences, which are distributed in 33 phyla [11]. Eachprimer matched with over 70% of the total sequencesdeposited in RDP II, recognizing all phyla. More than 70%of the sequences assigned in each phylum were recognizedby these primers. Therefore, no specific preference duringamplification should be expected, suggesting narrowdiversity is not a consequence of primer bias.
Cloning showed genetic diversity within Pseudomonassequences with calculated nucleotide differences amongthese clones ranging from 0.4% to 4.8%. Considering thatintergenomic differences in the nine genomes of Pseudo-monas (available at Microbial Genomes NCBI) range from
Figure 4 a RFLP obtained by digestion with AluI of the ITSamplified from DNA extracted from pooled gastrointestinal tractsfrom five fish of varying sizes. Lane Ps corresponds to Pseudomonasisolate, and lanes P14, P4, P3, P1.5, and P4.1 correspond to pools offish with sizes of 14, 4.1, 4, 3, and 1.5 g, respectively. These fish werecollected in different seasons from the same location. M correspondsto a 100-bp molecular weight marker (Invitrogen). b RFLP obtainedby digestion with AluI of the ITS amplified from DNA extracted frompooled gastrointestinal tracts from five fish of various sizes comparedto stomach, pyloric caeca, and intestine from a single fish of 30 g(lanes S: stomach, PC: pyloric caeca, and I: intestine). Lanes P14, P4,P3, and P1.5 correspond to pools described in a. M corresponds to a100-bp molecular weight marker (Invitrogen)
558 P. Navarrete et al.
0.4% to 5.9% and also that intragenomic differences areless than 0.2%, the differences found among the Pseudo-monas clones suggest the presence of several distinguish-able Pseudomonas strains in the gut samples.
The dominance of a particular bacterial group waspreviously observed in salmonid guts using culture-independent methods. Holben et al. [20] reported that somegenera were highly abundant in reared Atlantic salmonfrom two different locations: in the Scottish hatchery,Mycoplasma corresponded to 81% of clones retrieved,whereas in the Norwegian hatchery, Acinetobacteraccounted for 55%. Although other genera were alsopresent, their abundance was closer to 2%. Interestingly,in wild salmon (entirely carnivorous), the abundance ofMycoplasma was 96% of the clones analyzed [20].Similarly, Pond et al. [31] described the intestinal micro-biota of rainbow trout by using a cloning approach. Theyreported only two major groups among 200 clonesanalyzed, which corresponded to Clostridium and Aeromo-nas. Furthermore, Kim et al. [24] reported that Clostridiumdominated microbiota in rainbow trout analyzed by DGGE.The carnivorous diet of salmon may explain in part the lownumber of taxa observed, since a recent study indicated thatdiet influences the bacterial diversity of the digestive tract.Bacterial diversity increases from carnivory to omnivory toherbivory [25].
Our knowledge about the fish microbiota was obtainedusing several culture approaches and these methodsrevealed a wide range of microorganisms inhabiting thedigestive tract of fish [3, 10, 19, 36]. Our culture analysisshowed, however, that less than 1% of the microscopicallyobserved bacteria in salmon gut compartments werecultivable, indicating that microbiota detected with theculture approach may represent only a small fraction of thetotal microbiota, as previously described [23, 42]. Byculture analysis, Shewanella appeared as the most abundantbacterium in the salmon digestive tract, which contrastedwith the molecular results. Since PCR pitfalls may includeefficiency bias [50], the PCR amplification efficiency of16S rDNA and ITS from both Shewanella and Pseudomo-nas isolates was evaluated. The same efficiency was
measured for both bacteria. Hence, these discrepant resultsmay be better explained by preferential cultivation ofShewanella spp. under the experimental conditions. It isalso important to note that strictly anaerobic bacteria werenot investigated in our study. Ringø et al. [39] havepreviously suggested that the predominant bacteria isolatedfrom the salmonid gut are aerobes or facultative anaerobes.In our study, molecular approaches did not reveal thepresence of anaerobes. Although they may be present, theirabundances are not high enough to be detected by thesemethods.
The salmon gastrointestinal tract is made up of a U-shaped stomach, pyloric caeca, and intestine. Thus, it maybe expected that the composition of bacterial microbiotaalong the salmon gastrointestinal tract would also bedifferent, showing regional specialization. However, thesimilar community composition found along the digestivetract of juvenile salmon suggests that conditions foundwithin each compartment are not sufficient to selectdifferent bacterial populations. The absence of majordifferences may be due to several factors, namely, thecontinuous presence of feed with high protein content mayneutralize the pH along the gut. Feed particles are notretarded in the salmon gut: the pyloric caeca fill within thesame time course as the intestine, thus limiting feedfermentation [7–9, 20, 43]. This contrasts with the caecain birds and mammals, which have larger retention timesfor fermentation functions [44]. Holben et al. [20] reportedlow concentrations of acidic bacterial metabolites (aceticacid, lactic acid) indicating that a decreasing redox gradientis absent in salmon, leading to low levels of bacterialfermentation. Furthermore, the presence of Pseudomonas,which have an aerobic metabolism, suggests that thejuvenile salmon gut is predicted to have significant levelsof oxygen along the gut, hence limiting the growth offermentative bacteria [31, 33].
Previous culture-based studies have reported that bacte-ria present in the hatchery environment may influence thecomposition of gastrointestinal microbiota [10, 36]. Specif-ically, it was suggested that bacteria from water or feed maysurvive and multiply in the digestive tract [30, 49]. As
Table 2 Relative abundance and nearest-match identification of bacterial isolates obtained from salmon gut compartments
Bacterial isolate groups Phylum/class ITS (bp) Relative abundance (%)
Salmon gut compartment
Genus/accession number Stomach Pyloric caeca Hindgut
revealed by the TTGE analysis of the water influent andsalmon feed, Pseudomonas do not represent the dominantbacteria in these samples, probably because they wereabsent or present in very small proportions (less than 1%)[29]. However, a bacterial isolate identified as Pseudomo-nas was retrieved from the water sample showing 97%identity with dominant band 1 from the salmon. Converse-ly, dominant bacteria from water and pelleted feed were notdetected in the salmon digestive tract. These results suggestthat the salmon digestive tract is a favorable habitat forharboring some bacteria (Pseudomonas) that may bederived from minor bacteria present in the water influent.This host selection was recently demonstrated by thereciprocal transplantation of gut microbiota between differ-ent species [33]. Moreover, Pseudomonas was detected as acommon and dominant component in most of the samplesindependent of both the date of collection and fish size,suggesting that these bacteria may be selectively associatedwith salmon in this hatchery. These data support thetraditional idea that the composition and diversity of gutmicrobiota may depend not only on host selection but alsoon local conditions. This statement is reinforced by ourprevious finding, which described Pseudomonas as adominant bacteria recovered from coho salmon reared inthe same hatchery [42].
The successful permanence of Pseudomonas within fishgut compartments may be due the eventual attachment ofthe bacterium to digested food particles, mucus-likematerial, or to the host epithelium that prevent its microbialwashout [45]. The possible role of Pseudomonas within thedigestive tract of fish has been recently addressed: thisbacteria modulates the expression of some genes of the hostrelated to nutrient metabolism and innate immune response[34, 35].
Pseudomonads have been reported as common, oftendominant, members of the microbial communities of otherfish and of water [19, 35, 42]. Recently, it was demonstrat-ed that strains assigned to different Pseudomonas species,which shared >95% in 16S rDNA identity, may have verydissimilar genomes indicated by low values of DNA–DNAhybridization (7–37%) [17]. The stability of Pseudomonasspp. over the time evaluated in this hatchery may be due tothis high genome versatility, which could provide someadvantage in the gut environment.
Altogether, our results suggest that the Atlantic salmongut favors Pseudomonas establishment and that this bacterialpopulation dominated the gut of these salmon. However, itshould be noted that the gastrointestinal microbiota describedin this study may not be representative of all Atlanticsalmon, especially those in an open environment.
Acknowledgments This project was supported by a grant(FONDECYT No. 1061121) from CONICYT-Chile. P. Navarrete
acknowledges a scholarship from CONICYT-Chile and Dr. Stekel afellowship from INTA-Nestlé. Partial support was derived from anINNOVA CORFO grant (05CT6PPT-09) and FONDECYT 1080480.
References
1. Amann RI, Ludwig W, Schleifer KH (1995) Phylogeneticidentification and in-situ detection of individual microbial-cellswithout cultivation. Microbiol Rev 59:143–169
2. Andlid T, Vazquez-Juarez R, Gustafsson L (1998) Yeasts isolatedfrom the intestine of rainbow trout adhere to and grow in intestinalmucus. Mol Mar Biol Biotech 7:115–126
3. Austin B (2006) The bacterial microflora of fish, revised.ScientificWorldJournal 6:931–945
4. Backhed F, Ding H, Wang T, Hooper LV, Koh GY, Nagy A,Semenkovich CF, Gordon JI (2004) The gut microbiota as anenvironmental factor that regulates fat storage. Proc Natl Acad SciU S A 101:15718–15723
5. Bates J, Mittg E, Kuhlman J, Baden K, Cheesman S, GuindulainT (2006) Distinct signals from the microbiota promote differentaspects of zebrafish gut differentiation. Dev Biol 297:374–386
6. Bergot P, Solari A, Luquet P (1975) Comparaison des surfacesabsorbantes des caeca pyloriques et de l’intestin chez la truite arc-en-ciel (Salmo gairdneri Rich). Ann Hydrobiol 6:27–43
7. Buddington RK, Chen JW, Diamond J (1987) Genetic andphenotypic adaptation of intestinal nutrient transport to diet infish. J Physiol 393:261–281
8. Buddington RK, Diamond JM (1987) Pyloric ceca of fish: a“new” absorptive organ. Am J Physiol 252:G65–G76
9. Buddington RK, Krogdahl A, BakkeMcKellep AM (1997) Theintestines of carnivorous fish: structure and functions and therelations with diet. Acta Physiol Scand 161:67–80
10. Cahill MM (1990) Bacterial-flora of fish—a review. Microb Ecol19:21–41
11. Cole JR, Chai B, Farris RJ, Wang Q, Kulam SA, McGarrell DM,Garrity GM, Tiedje JM (2005) The Ribosomal Database Project(RDP-II): sequences and tools for high-throughput rRNA analysis.Nucleic Acids Res 33:D294–D296
12. Cole JR, Chai B, Farris RJ, Wang Q, Kulam-Syed-Mohideen AS,McGarrell DM, Bandela AM, Cardenas E, Garrity GM, Tiedje JM(2007) The ribosomal database project (RDP-II): introducingmyRDP space and quality controlled public data. Nucleic AcidsRes 35:D169–D172
13. Espejo RT, Escanilla D (1993) Detection of HIV1 DNA by asimple procedure of polymerase chain reaction, using “primer-dimer” formation as an internal control of amplification. Res Virol144:243–246
14. Espejo RT, Feijoo CG, Romero J, Vasquez M (1998) PAGEanalysis of the heteroduplexes formed between PCR-amplified16S rRNA genes: estimation of sequence similarity and rDNAcomplexity. Microbiology 144(Pt 6):1611–1617
15. Espejo RT, Romero J (1997) Bacterial community in coppersulfide ores inoculated and leached with solution from acommercial-scale copper leaching plant. Appl Environ Microbiol63:1344–1348
16. Gordon JI (2005) A genomic view of our symbiosis with membersof the gut microbiota. J Pediatr Gastroenterol Nutr 40(Suppl 1):S28
17. Goris J, Konstantinidis KT, Klappenbach JA, Coenye T,Vandamme P, Tiedje JM (2007) DNA–DNA hybridization valuesand their relationship to whole-genome sequence similarities. Int JSyst Evol Microbiol 57:81–91
18. Griffiths S, Melville K, Cook M, Vincent S, Pierre M, LanteigneC (2001) Profiling of bacterial species associated with haddock
560 P. Navarrete et al.
larviculture by PCR amplification of 16S rDNA and denaturinggradient gel electrophoresis. J Aquat Anim Health 13:355–363
19. Hansen GH, Olafsen JA (1999) Bacterial interactions in early lifestages of marine cold water fish. Microb Ecol 38:1–26
20. Holben WE, Williams P, Gilbert MA, Saarinen M, Sarkilahti LK,Apajalahti JH (2002) Phylogenetic analysis of intestinal micro-flora indicates a novel Mycoplasma phylotype in farmed and wildsalmon. Microb Ecol 44:175–185
21. Hovda MB, Lunestad BT, Fontanillas R, Rosnes JT (2007)Molecular characterisation of the intestinal microbiota of farmedAtlantic salmon (Salmo salar L.). Aquaculture 272:581–588
22. Jensen MA, Webster JA, Straus N (1993) Rapid identification ofbacteria on the basis of polymerase chain reaction-amplifiedribosomal DNA spacer polymorphisms. Appl Environ Microbiol59:945–952
23. Jensen S, Øvreås L, Bergh O, Torsvik V (2004) Phylogeneticanalysis of bacterial communities associated with larvae of theAtlantic halibut propose succession from a uniform normal flora.Syst Appl Microbiol 27:728–736
24. Kim DH, Brunt J, Austin B (2007) Microbial diversity ofintestinal contents and mucus in rainbow trout (Oncorhynchusmykiss). J Appl Microbiol 102:1654–1664
25. Ley RE, Hamady M, Lozupone C, Turnbaugh PJ, Ramey RR,Bircher JS, Schlegel ML, Tucker TA, Schrenzel MD, Knight R,Gordon JI (2008) Evolution of mammals and their gut microbes.Science 320:1647–1651
26. Magne F, Abely M, Boyer F, Morville P, Pochart P, Suau A (2006)Low species diversity and high interindividual variability in faecesof preterm infants as revealed by sequences of 16S rRNA genesand PCR-temporal temperature gradient gel electrophoresisprofiles. FEMS Microbiol Ecol 57:128–138
27. McCracken VJ, Simpson JM, Mackle RI, Gaskins HR (2001)Molecular ecological analysis of dietary and antibiotic-inducedalterations of the mouse intestinal microbiota. J Nutr 131:1862–1870
28. Moreno C, Romero J, Espejo RT (2002) Polymorphism inrepeated 16S rRNA genes is a common property of type strainsand environmental isolates of the genus Vibrio. Microbiology148:1233–1239
29. Muyzer G, Dewaal EC, Uitterlinden AG (1993) Profiling of complexmicrobial-populations by denaturing gradient gel-electrophoresisanalysis of polymerase chain reaction-amplified genes-coding for16S ribosomal-RNA. Appl Environ Microbiol 59:695–700
30. Olafsen JA (2001) Interactions between fish larvae and bacteria inmarine aquaculture. Aquaculture 200:223–247
31. Pond MJ, Stone DM, Alderman DJ (2006) Comparison ofconventional and molecular techniques to investigate the intestinalmicroflora of rainbow trout (Oncorhynchus mykiss). Aquaculture261:194–203
32. Ransom D, Lannan C, Rohovec J, Fryer J (1984) Comparison ofhistopathology caused by Vibrio anguillarum and Vibrio ordalii inthree species of Pacific salmon. J Fish Dis 7:107–115
33. Rawls JF, Mahowald MA, Ley RE, Gordon JI (2006) Reciprocalgut microbiota transplants from zebrafish and mice to germ-freerecipients reveal host habitat selection. Cell 127:423–433
34. Rawls JF, Samuel BS, Gordon JI (2004) Gnotobiotic zebrafishreveal evolutionarily conserved responses to the gut microbiota.Proc Natl Acad Sci U S A 101:4596–4601
35. Rawls JF, Mahowald MA, Goodman AL, Trent CM, Gordon JI(2007) In vivo imaging and genetic analysis link bacterial motilityand symbiosis in the zebrafish gut. Proc Natl Acad Sci U S A104:7622–7627
36. Ringø E, Birkbeck T (1999) Intestinal microflora of fish larvaeand fry. Aquac Res 30:73–93
37. Ringø E, Olsen RE, Mayhew TM, Myklebust R (2003) Electronmicroscopy of the intestinal microflora of fish. Aquaculture227:395–415
38. Ringø E, Sperstad S, Myklebust R, Refstie S, Krogdahl A (2006)Characterisation of the microbiota associated with intestine ofAtlantic cod (Gadus morhua L.): the effect of fish meal, standardsoybean meal and a bioprocessed soybean meal. Aquaculture261:829–841
39. Ringø E, Strom E, Tabachek J (1995) Intestinal microflora ofsalmonids: a review. Aquac Res 26:773–789
40. Romero J, Espejo R (2001) The prevalence of noncultivablebacteria in oysters (Tiostrea chilensis, Philippi, 1845). J ShellfishRes 20:1235–1240
41. Romero J, Garcia-Varela M, Laclette JP, Espejo RT (2002)Bacterial 16S rRNA gene analysis revealed that bacteria relatedto Arcobacter spp. constitute an abundant and common compo-nent of the oyster microbiota (Tiostrea chilensis). Microb Ecol44:365–371
42. Romero J, Navarrete P (2006) 16S rDNA-based analysis ofdominant bacterial populations associated with early life stages ofcoho salmon (Oncorhynchus kisutch). Microb Ecol 51:422–430
43. Rønnestad I, Rojas-Garcia CR, Skadal J (2000) Retrogradeperistalsis; a possible mechanism for filling the pyloric caeca? JFish Biol 56:216–218
44. Rust M (2002) Nutritional physiology. In: Halver J, Hardy R (eds)Fish nutrition. Academic, San Diego, pp 368–446
45. Sonnenburg JL, Angenent LT, Gordon JI (2004) Getting a grip onthings: how do communities of bacterial symbionts becomeestablished in our intestine? Nat Immunol 5:569–573
46. Suau A, Bonnet R, Sutren M, Godon JJ, Gibson GR, Collins MD,Dore J (1999) Direct analysis of genes encoding 16S rRNA fromcomplex communities reveals many novel molecular specieswithin the human gut. Appl Environ Microbiol 65:4799–4807
47. Suyehiro Y (1942) A study on the digestive system and feedinghabits of fish. Jpn J Zool 10:1–303
48. vandePeer Y, DeWachter R (1997) Construction of evolutionarydistance trees with TREECON for Windows: accounting forvariation in nucleotide substitution rate among sites. Comput ApplBiosci 13:227–230
49. Verschuere L, Rombaut G, Sorgeloos P, Verstraete W (2000)Probiotic bacteria as biological control agents in aquaculture.Microbiol Mol Biol Rev 64:655–671
50. von Wintzingerode F, Göbel U, Stackebrandt E (1997) Determi-nation of microbial diversity in environmental samples: pitfalls ofPCR-based rRNA analysis. FEMS Microbiol Rev 21:213–229
