Intestinal microbiota variation in Senegalese sole (Solea senegalensis) under different feeding regimes

Intestinal microbiota variation in Senegalese sole

(Solea senegalensis) under different feeding regimes

Beatriz Martin-Antonio1, Manuel Manchado1, Carlos Infante1, Ricardo Zerolo1, Alejandro Labella2,Carmen Alonso2 & Juan J Borrego2

1IFAPACentro El Torun� o (Junta deAndaluc|¤ a), CaminoTiro de Picho¤ n s/n, CaŁ diz, Spain2Department of Microbiology, Faculty of Sciences, University of Malaga, Malaga, Spain

Correspondence: J J Borrego, Department of Microbiology, Faculty of Sciences, University of Malaga, Campus Universitario Teatinos,

29071Malaga, Spain. E-mail: [email protected]

Abstract

The aimof this study was to determine the in£uence ofthe feeding regimes in Senegalese sole (Solea senegalen-sis) culturedunderextensive, semi-extensive and inten-sive production systems. Atotal of 254 bacterial isolatesfrom guts of ¢sh cultured under di¡erent productionsystems and feeding regimes were tested. Biochemicaltests and genetic analyses based on the 16S rDNA se-quence analysis were conduced to identify bacterialstrains.Vibrio species were the most represented taxo-nomic group in the culturable microbiota of S. senega-lensis guts tested. Particularly,Vibrio ichthyoenteri wasthe most frequently isolatedVibrio species. Comparisonamong diets showed a signi¢cant reduction (Po0.05)in vibrio percentages and a higher occurrence of She-wanella species in Senegalese soles fed polychaeta. Inaddition, a major in£uence of environmental tempera-ture on microbiota composition was detected. Coldtemperatures brought about a change in the percen-tages ofVibrio species and a higher representation of a-Proteobacteria in both outdoor systems (extensive andsemi-extensive).The signi¢cant di¡erences between in-testinal bacterial composition in Senegalese soles fedcommercial diets and natural preys (polychaeta) revealthe necessity to develop speci¢c optimized diets for theintensive rearing of this ¢sh species.

Keywords: Solea senegalensis, intestinal microbio-ta, 16S rDNA, feeding regimes, culture productionsystems

Introduction

Flat¢sh species share in common an asymmetricalbody development and a bottom-dwelling mode of

life. Depending on their feeding behaviour, they canbe classi¢ed into three groups: ¢sh-feeders, crusta-cean-feeders and polychaete/mollusc-feeders (deGroot 1971). Senegalese sole (Solea senegalensis) pos-sesses a high commercially valuable with an expand-ing aquaculture industry in Southern Europe(Imsland, Foss, Conceic� ao, Dinis, Delbare, Schram,Kamstra, Rema & White 2003). During intensiverearing, this species mainly feeds commercial dietsbased on ¢shmeal, which are quite distinct from thefeeding regimes in its natural environment, com-posed of di¡erent types of prey, the most relevantbeing crustaceans, polychaeta and molluscs (Garcia-Franquesa, Molinero, Valero & Flos 1996; Cabral2000; Sa, Bexiga,Vieira,Veiga & Erzini 2003). In addi-tion, Senegalese sole feeding habits seem to be highlyin£uenced by sex, age and season, witha feeding spe-cialization depending on food availability (Sa et al.2003).The intensive rearing of ¢sh species inaquaculture

has revealed intimate relationships between ¢sh andbacteria that eventually may a¡ect establishment ofan intestinal microbiota or result in disease out-breaks (Hansen & Olafsen1999). This process is com-plex and seems to be in£uenced by the microbialcontent of ¢sh eggs, the live feed and the bacteria pre-sent in thewater (Ringo & Birbeck1999). Gram-nega-tive facultative anaerobes prevail in the digestivetract of ¢sh (Clements 1997),Vibrio and Pseudomonasbeing the most common genera in marine ¢sh (Saka-ta 1990; Toranzo, Novoa, Romalde, Nun� ez, Devesa,Marin� o, Silva, Martinez, Figueras & Barja 1993). Fishintestinal tract is a dynamic ecosystemwhere severalfactors, such as food regimes, water quality orjust handling, determine changes in the microbiota

Aquaculture Research, 2007, 38, 1213^1222 doi:10.1111/j.1365-2109.2007.01790.x

r 2007 TheAuthorsJournal Compilationr 2007 Blackwell Publishing Ltd 1213

composition that may induce opportunistic bacterialpathogens to proliferate and provoke infectionand ulterior disease (Verschuere, Rombaut, Sorgeloos& Verstraete 2000;Winton 2001). However, coloniza-tion of the gut by a pathogen can be hinderedby the interactions between the pathogen and the in-testinal microbiota. In this respect, bacterial patholo-gies, caused mainly by Photobacterium damselaesubsp. piscicida, Vibrio harveyi and V. alginolyticus,are considered to be the most important factors af-fecting the culture of S. senegalensis in SouthernSpain (Zorrilla, Balebona, Morin� igo, Sarasquete &Borrego 1999; Zorrilla, Arijo, Chabrillon, Diaz,Martinez-Manzanares, Balebona & Morin� igo 2003).Moreover, in a recent study on microbiota composi-tion in Senegalese sole, it has been demonstratedthat diets based on polychaeta considerably promotethe number of bacterial strains with antimicrobialactivity in the intestine (Makridis, Martins,Tsalavou-ta, Dionisio, Kotoulas, Magoulas & Dinis 2005).However, neither the in£uence of feeding regimesnor the production system has been evaluated untilnow.The aim of the present study was to investigate the

microbiota composition in S. senegalensis in di¡erentaquaculture production systems (extensive, semi-ex-tensive and intensive) as well as the changes inducedby di¡erent diets.

Materials and methods

Feeding experiments and sampling

Extensive and semi-extensive production systems

Di¡erent batches of juvenile Senegalese soles weremass reared at IFAPA‘El Torun� o’ (El Puerto de SantaMaria, Cadiz, Spain) facilities from eggs obtainedfrom natural spawnings of captive breeders. Larvaewere reared with rotifers, artemia and dry food(Minipro; MariproAS, Bergen, Norway) according toCan� avate and Fernandez-Diaz (1999) up to 60 days.At this stage, ¢sh specimens (average weight 70mg)were transferred to outdoor experimental ponds (July2005). In the extensive trials, the initial density was10 individuals m�2, whereas 15 individuals m�2

were used as the initial stocking density in semi-ex-tensive trials (Fig.1a). Rectangular ponds with a sur-face of 1600m2 and 0.5m water depth were used forextensive nursery trials with juveniles. Before solestocking, these ponds were emptied and allowed todry to achieve predator elimination and bottom re-generation. For semi-extensive growing trials, smal-ler rectangular tanks (100m2) were used. Pondbottomwas also of the samemuddy nature andwaterdepth as in larger ponds. A constant water £ow of200m3 day�1 was maintained. After ¢lling up, allponds were allowed to stagnate for 3 weeks in orderto promote planktonand benthos development before

(a)

(b)

Figure 1 Experimental design used for di¡erent production systems. (a) Extensive, and semi-extensive productionsystems. (b) Intensive production system.

In£uence of feeding on intestinal microbiota BMartin-Antonio et al. Aquaculture Research, 2007, 38, 1213^1222

r 2007 TheAuthors1214 Journal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222

sole juvenile stocking. Polychaeta species in thebenthos were evaluated weekly. Three months afterstarting the experiment, when the concentration ofpolychaeta was estimated to have been reduced by90%, a commercial diet (Microbaq; Dibaq-Diproteg,Segovia, Spain) was supplied daily (2% total biomass)in the semi-extensive growing trials. Four to eightsoles from each experimental tank were caughtusing traps in December (2005) and March (2006),6 and 9months after starting the experiment respec-tively. They were killed by immersion in tricainemethanesulphonate (MS-222, Sigma Chemical, St.Louis, MO, USA), measured and the guts were rapidlydissected for bacterial analyses as described below.

Intensive production system

Twelve sole specimens, approximately 15 months old(between 51.1 and 57.3 g), reared at IFAPA‘El Torun� o’,were transferred to each of three 100 L £at-bottomtanks (commercial diet, CD; no diet, ND and polychae-ta diet, PD).Theywere fed anarti¢cial commercial dietuntil starting the experiment. The temperature andsalinity remained stable during the experiment (18 1Cand37 g L�1). During the experiment, ¢sh in tanks CDand ND were fed a commercial diet (Microbaq) for 3weeks. Animals in the PD tank were fed commercialfrozen polychaeta Nereis virens (Seabait

s

, Northum-berland, UK) (Fig. 1b). During this period, food quan-tity was approximately 2% of the body weight, and itwas completely ingested every day. In the fourthweek,no foodwas provided to tank ND. Fish in tanks CD andPD were fed the commercial diet and polychaeta, re-spectively, for an additional week. After this, no morefood was provided until ¢nishing the experiment.Samplings for microbiota composition were carriedout at days 0,1,8 and11after starting the experiment.At each sampling, two soles from each tank wereweighed and killed by immersion in MS-222.

Bacterial strains

Gut from each juvenile sole was dissected under asep-tic conditions following the technique described byWesterdahl, Olsson, Kjelleberg and Conway (1991).After weighing, it was homogenized and diluted to1/10 (w/v) in a sterile saline solution. The total viablecounts (expressed as colony-forming units, CFUg�1)of bacteriawere estimated by making10-fold serial di-lutions (10�1^10�4) of intestinal homogenates. Ali-quots of 100 mL were plated onto Marine agar (MA)(Cultimed Panreac Quimica SA, Barcelona, Spain)

and Thiosulphate^Citrate-Bile salt^Sucrose (TCBS)agar (Biolife, Milan, Italy) and incubated at 25 1C for2^3 days. A statistically signi¢cant number of coloniesfromeachgroup classi¢ed on the basis of their culturalcharacteristics (Bianchi & Bianchi1982) were selectedand subcultured to pure cultures for further DNA se-quencing analyses. Biochemical characterization wasperformed using the API 20NE system (BioMerieux,Madrid, Spain) following the manufacturer’s instruc-tions. The concentration of colonies for each taxonwas expressed as CFUg�1of intestinal content.In order to investigate bacterial composition in

water, samples of approximately 100mL were col-lected from the inlet water in each production system.They were homogenized and 100 mL of non-dilutedwater samples or appropriate dilutions of water(10�1^10�3) were plated onto MAandTCBS in dupli-cate. Each type of colony was counted and the con-centration was expressed as CFUmL�1. For analysisof commercial feed and polychaeta, they were homo-genized, diluted to 1/10 (w/v) in a sterile saline solu-tion and plated onto MA andTCBS as reported above.The Mann^Whitney U-test was used for data analy-sis, with a level of signi¢cance at a P-valueo0.05.

Polymerase chain reaction (PCR)ampli¢cation and sequencing

A fragment of16S rDNAwas ampli¢ed using the uni-versal primers 63f and1387r (Marchesi, Sato,Weight-man, Martin, Fry, Hiom & Wade 1998). Polymerasechain reactions were carried out in a 25 mL reactionmixture that included 5 pmol of each primer, 200 mMdNTPs, 1 � PCR bu¡er, 2mM MgCl2, 1.5 U EcoTaqpolymerase (EcoTaq; Ecogen, Barcelona, Spain) and1 mL of a boiled colony suspension. The PCR pro¢lewas as follows: 2min at 94 1C, 30 cycles of 30 s at94 1C, 30 s at 55 1C and 1min at 72 1C. Polymerasechain reaction products were electrophoresed on a2% agarose gel and visualized via ultraviolet transil-lumination. Following the PCR reaction, uncon-sumed dNTPs and primers were removed using thePCR product puri¢cation kit (Marlingen Bioscience,Ijamsville, MD, USA). Cycle sequencing was per-formed with Bigdye

s

Terminator v3.1 kit (AppliedBiosystems, Foster City, CA, USA). All sequencing re-actions were performed according to the manufac-turer’s instructions using the ABI3130 GeneticAnalyzer (Applied Biosystems). DNA sequences wereanalysed using Seqman v5.53 (DNASTAR). The 16SrDNA gene sequences were used in a BLAST search inorder to retrieve the most closely related sequences.

Aquaculture Research, 2007, 38, 1213^1222 In£uence of feeding on intestinal microbiota BMartin-Antonio et al.

r 2007 TheAuthorsJournal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222 1215

Sequences for the ¢nal analysis were selected accord-ing to the results of the BLAST query.When several se-quences were available for a type species, allsequences were included in a preliminary analysis,although a single one was chosen for the tree that isshown. To select the model of DNA substitution thatbest ¢tted the data, a hierarchical likelihood ratio testapproach implemented in the program MODELTEST 3.5(Posada & Crandall, 1998) was used. The selectedmodel was theTrN1I1G, with a g-distribution shapeparameter (G) of 0.2816 and a proportion of invariablesites (I) of 0.2354. The TrN model assigns di¡erenttransition rates to purines (2.4278) and pyrimidines(3.6458) and gives a third rate (1.0) to all transver-sions. A neighbour-joining analysis using model-testparameters was carried out using PAUP� version4.0b10. Bootstrap values were obtained through1000 replicates using the same model as above. Se-quences have been deposited in GenBank/EMBL/DDBJwith accession numbers AB274733^AB274782.

Results

Microbiota composition in Senegalese solecultured in earth ponds under extensive andsemi-extensive production systems

In order to evaluate the microbiota composition inSenegalese sole cultured in earth ponds, two ¢shgroups with di¡erent feeding regimes were com-pared. In the extensive group (EXT), soles were exclu-sively fed wild prey species similar to the naturalfeeding. Polychaeta (worms) and crustaceans,mainly amphipoda, could be found in the gut.Capitella capitata was identi¢ed as the most fre-quently found polychaeta species in these earthponds. In the semi-extensive group (SEM), animalswere initially fed wild polychaeta (mainly Capitella).Three months later, when worm populations were

considerably reduced, they were daily supplied witha commercial diet. During samplings, pellets and am-phipoda could be detected in the gut.The ¢rst sampling was carried out 6 months after

sole stocking (December). The ¢sh mean weight was4.2 � 1.2 and 6.2 � 2.1g for EXTand SEM groups re-spectively (Table1).The gut meanweight was 0.2 and0.4 g for EXT and SEM groups respectively. A secondsampling was carried out 3 months later (March).The ¢sh mean weight was 5.5 � 0.9 and10.1 � 3.2 gfor EXT and SEM groups respectively. The gut meanweight was 0.2 and 0.8 g for EXTand SEM groups re-spectively.Water temperature £uctuated considerablyduring the experiment. At initial stocking in July, themean water temperature was 27.6 � 2.3 1C; it de-creased to 10.6 � 1.5 1C in December and increasedup to15.4 � 3.3 1C in March.No signi¢cant di¡erences (P40.05) with respect to

the bacterial numbers in ¢sh gut were found betweenproduction systems and temporal samplings (Table1).The bacterial number varied between 2.3 � 105 and6.7 � 106 CFUg�1 on MA. Lower values (5.0 � 103^3.9 � 105 CFUg�1) were obtained on TCBS. In addi-tion, themeanbacterial concentration inwater variedin bothgroups from1.3 � 103 to 6.0 � 103 CFUmL�1,and from 6.2 � 102 to 1.2 � 103 CFUmL�1 on MAand TCBS respectively. Bacterial titres in commercialfeed and polychaeta obtained on MAwere 1.8 � 103

and1.0 � 102 CFUg�1respectively.

Species composition of bacterial isolates

A total of 78 representative isolates were selected forfurther identi¢cation based on colony morphologyon MA and TCBS. Polymerase chain reaction frag-ments of 16S rDNA ranged from 1167 to 1243 bp inlength. Bacterial isolates were assigned to di¡erenttaxa from BLAST searches performed with the 16SrDNA sequences. Species belonging to genus Vibrio

Table 1 Estimated bacterial numbers (CFUg�1) on Marine Agar (MA) and TCBS plates from gut samples of Senegalese solecultured in extensive (EXT) and semi-extensive (SEM) production systems

Production system

December March

Mean (SD) fishweight (g) MA TCBS Mean (SD) fishweight (g) MA TCBS

EXT 4.2 (1.2) 5.0 � 106 5.0 � 103 5.5 (0.9) 6.7 � 106 3.9 � 105

0.2a 0.2

SEM 6.2 (2.1) 2.3 � 105 4.8 � 104 10.1 (3.2) 1.8 � 106 2.7 � 105

0.4 0.8

aGut weight (g).TCBS,Thiosulphate^Citrate-Bile salt^Sucrose.

In£uence of feeding on intestinal microbiota BMartin-Antonio et al. Aquaculture Research, 2007, 38, 1213^1222

r 2007 TheAuthors1216 Journal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222

were represented in both experimental groupsduring all samplings. OnMA, their abundance variedbetween 15.1% and 60.0% in the EXT group, andbetween 27.9% and 94.1% in the SEM group inDecember and March respectively (Table 2). The a-Proteobacteria group was highly represented inDecember (84.7% and 46.5% in EXTand SEM groupsrespectively), when the water temperature was verylow (9^13 1C). It is important to note the signi¢cantincrease in Shewanella genus abundance in the EXTgroup, which was not found in December, whereas itconstituted 39.9% of the bacterial isolates in March.In contrast, this genus was only 5.9% of the bacterialisolates characterized in March in the SEM group,where genus Pseudoalteromonaswas detected mainlyin December (18.0%). Analysis of water compositiondemonstrated that Vibrio was the most frequentlyisolated taxonomic group (42.3%), followed byPseudoalteromonas (8.1%).The phylogenetic analysis showed that 16S rDNA

sequences from 25 representative isolates could beclassi¢ed into three groups belonging to Bacteroi-detes, Firmicutes and Proteobacteria phyla (Fig. 2).The latter was the most represented with species ofa-Proteobacteria and g-Proteobacteria classes. For a-Proteobacteria, strains EXT2 and SEM4 were themost abundant, with high homology to the mucusbacteria 107 [BLAST identity (BI)51167/1171] and 36(BI51234/1234). For g-Proteobacteria, strain EXT13,phylogenetically related to the mucus bacterium 6(AY654802; BI51233/1234) and Vibrio splendidusbiovar II (AB038030; BI51233/1234), were over-represented in December. Strains SEM8 and EXT6,with high similarity to Vibrio ichthyoenteri (1234/1234) and Vibrio chagasii (1233/1234), respectively,were the most represented in March.With respect toShewanella, three strains were identi¢ed. The mostabundant was EXT8 in the extensive system, 100%

identity with Shewanella sairae (BI51234/1234).Strains EXT13 and SEM3, phylogenetically related toV. splendidus biovar II, were detected in the water ofboth production systems. Strains EXT18 and SEM8,closely related toV. ichthyoenteri, were also detectedin both groups in March. In addition, strains EXT5and SEM14 and EXT22 and SEM2, homologous toVi-brio sp. V776 (DQ146991; BI51234/1234) andVibriosp. V201 (DQ146980; BI51234/1234), respectively,were also detected in both aquaculture systems.All bacterial isolates on TCBS plates corresponded

toVibrio, similar to the results obtained onMAplates.In water, the bacterial composition on MAwas quiteheterogeneous including mainly species ofVibrio andPseudoalteromonas genera, although other speciesbelonging to Shewanella, Jannaschia or Flexibactergenera could also be identi¢ed.

Microbiota composition in Senegalese soleunder intensive conditions

For comparative purposes, the microbiota composi-tion in animals cultured under an intensive aquacul-ture system was investigated. In addition, changesbrought about by replacing the commercial dieteither by polychaeta or no food were evaluated.Microbiota composition in the three groups was ana-lysed at days 0, 1, 8 and 11 after starting the experi-ment (Fig. 1). Sole mean weight was 53.7 � 20.1,57.3 � 22.3 and 51.1 � 19.2 g for ND, CD and PDgroups respectively. The gut mean weight was 0.8and1.2 g for PD and CD groups respectively.Figure 3 shows the counts (CFUg�1) of bacteria on

MA and TCBS media. The type of feeding was shownto a¡ect the total viable bacterial counts. Soles thatreceived a commercial feed showed higher bacterialtitres (CFU) than those that fed polychaeta duringdays 0 and 1 (Mann^Whitney U-test, Po0.05).

Table 2 Percentages of bacterial groups isolated in the extensive (EXT) and semi-extensive (SEM) production systems, andwater (W)

EXT SEM

WD M D M

Vibrio spp. 15.1 60.0 27.9 94.1 42.3

Shewanella spp. – 39.9 – 5.9 0.8

Pseudoalteromonas spp. – 0.1 18.0 – 8.1

Bacillus spp. – – 0.6 – –

Alphaproteobacteria 84.7 – 46.5 –

Others 0.1 – 7.0 – 48.8

Bacterial groups not detected are represented by ‘^’.Values for December (D, sixth month) and March (M, ninth month) samplings are indicated.

Aquaculture Research, 2007, 38, 1213^1222 In£uence of feeding on intestinal microbiota BMartin-Antonio et al.

r 2007 TheAuthorsJournal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222 1217

Although ¢sh from the ND group showed lower bac-terial titres than those from the CD group during thisperiod, these di¡erences were not signi¢cant. Bacter-ial numbers decreased in the CD group at days 8 and11 (Po0.05). In contrast, CFU remained stable in thePD group during the experiment.

Regarding bacteria composition, 176 isolates wereselected and assigned to di¡erent taxa from BLAST

searches performed using the 16S rDNA sequences.Vibrio species were the most frequently identi¢ed inthe three groups on MA plates (Table 3). In the CDgroup, percentages were between 85% and 99% in

Figure 2 Neighbour-joining tree of microbiota isolates from soles cultured in earth ponds under extensive andsemi-extensive conditions. Reference strains have been included in the alignment. Bootstrap values470 are indicated.Rhodothermus marinuswas used as an outgroup to root the tree.

In£uence of feeding on intestinal microbiota BMartin-Antonio et al. Aquaculture Research, 2007, 38, 1213^1222

r 2007 TheAuthors1218 Journal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222

the four samplings. Similar values were obtained forthe ND group (93.2^100%). In contrast, this repre-sentation declined to 51.7^85.9% in the PD group.Vibrio ichthyoenteri strains were themost representedin the three groups, with 100% identity with theEXT18 and SEM8 strains described above. OtherVibrio species showed high similarity toV. splendidus,V. probioticus,V. chagasii,V. campbelli,V. harveyi andV.alginolyticus. The latter two species were detectedonly in the CD and ND groups. The same pro¢le ofVibrio species was obtained on MA and TCBS plates,withV. ichthyoenteri being the most represented spe-cies in all cases. Of particular interest is the identi¢-cation of di¡erent Shewanella strains exclusively inthe PD group. They appeared at days 8 and 11 andthe percentages £uctuated between15% and 48% re-spectively. Strains closely phylogenetically related toShewanella schlegeliana (BI51221/1233) and Shewa-nella pneumatophori (BI51232/1233) were the mostrepresented. Pseudoalteromonas spp. were also identi-¢ed in the three groups. Their abundance was be-tween 0.1% and 15%, depending on the group andsampling.

The water analysis showed species belonging toPseudoalteromonas genus and phylogenetically relatedto the mucus bacterium 6 (AY654802) as being themost abundant during the experiment. However, otherspecies belonging to Alcanivorax, Idiomarina, Marino-bacter, as well as di¡erentVibrio species could be identi-¢ed. In commercial feed and polychaeta, strains with99% (BI51219/1222) and100% (1246/1246) similarityto Kocuria rhizophila (AF542072) and Bacillus hwajin-poensis (AF541966), respectively, were isolated.

Discussion

Intestinal microbiota of ¢sh may play a key role infood digestion and disease control, competing withopportunistic bacteria and hindering pathogen colo-nization (Moriarty 1990; Hansen & Olafsen1999). Toassess the changes induced by diet in microbiota, weinvestigated the bacterial composition in Senegalesesole fed withwild prey from the natural environment(EXT trial) and subsequently fed with commercialdiet (SEM trial); we identi¢ed 78 intestinal bacteriabased on 16S rDNA sequences. The total viable

Figure 3 Counts of di¡erent bacterial groups isolated in the CD, PD and ND experimental groups and water. Mean bac-terial titres, expressed as CFUg�1 � SD, on MA andTCBS plates for samples collected at days 0,1,8 and11are indicated.CD, commercial diet; ND, no diet; PD, polychaeta diet; MA, Marine agar;TCBS,Thiosulphate^Citrate-Bile salt^Sucrose.

Table 3 Percentages of bacterial groups isolated in Senegalese soles cultured under intensive conditions

CDa PD ND

0 1 8 11 0 1 8 11 0 1 8 11 W

Vibrio spp. 99.0 98.7 95.9 85.0 73.7 85.9 84.9 51.7 99.0 93.2 100 100 11.8

Shewanella spp. – – – – – – 15.1 48.3 – – – – 0.1

Pseudoalteromonas spp. 1.0 0.1 4.1 15.0 3.0 14.1 – – 1.0 3.8 – – 52.8

Bacillus spp. – 1.0 – – 23.3 – – – – – – – –

Others – 0.2 – – – – – – 3 – – 35.3

aCD and PD groups were fed a commercial diet and polychaeta for 4 weeks, respectively. The ND group was fed a commercial diet for 3weeks.Values from samplings at days 0, 1, 8 and 11, as well as mean water values (W), are reported.Bacterial groups not detected are represented by ‘^’.

Aquaculture Research, 2007, 38, 1213^1222 In£uence of feeding on intestinal microbiota BMartin-Antonio et al.

r 2007 TheAuthorsJournal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222 1219

counts onMAplates were similar in the EXTand SEMproduction systems, ranging from 2.3 � 105 to6.7 � 106 CFUg�1. These values fall within therange described for Dover sole (5.2 � 105^9.8 � 106 CFUg�1) (MacDonald, Stark & Austin1986), Japanese £ounder (3.6 � 105 to 6.0 �107 CFUg�1) (Sugita, Okano, Suzuki, Iwai, Mizukami,Akiyama & Matsuura 2002) and summer £ounder(average 1.3 � 105 CFUg�1) (Eddy & Jones 2002).However, a lower bacterial number was detected in aprevious study using Senegalese sole juveniles(2.4 � 103^1.2 � 105 CFUg�1) (Makridis et al. 2005).Endogenous and exogenous factors, such as the devel-opmental stage of thehost ¢sh (Sugita et al.2002), typeof food (Ringo & Vadstein1998; Eddy & Jones 2002) orbacterial host speci¢ty (Cerda-Cuellar & Blanch 2002),could explain these di¡erent results.The Vibrio genus is the predominant microbial

group detected in the intestinal content of marine¢sh (Olafsen 2001; Eddy & Jones 2002; Sugita et al.2002; Sugita & Ito 2006). Our results con¢rmed theprevalence ofVibrio species in the culturable micro-biota of S. senegalensis under the production systemsand feeding regimes tested. An increase in the isola-tion of Vibrio sp. and Pseudoalteromonas sp. was ob-tained in Senegalese sole fed with a commercial diet(SEM system) comparedwith animals fedwith natur-al preys (EXT system). In contrast, in the EXT systema higher detection of Shewanella sp. at the ninthmonth of feeding was achieved (Table 2).The mainVibrio species identi¢ed in Senegalese sole

throughout these experimentswasV. ichthyoenteri.Thisbacterial species is closely related toV. scophthalmi, acommon inhabitant in the intestines of other £at¢shsuch as Scophthalmus maximus (Cerda-Cuellar &Blanch 2002) or Paralichthys olivaceus (Sugita & Ito2006). However, V. ichthyoenteri has been associatedwith intestinal necrosis in the early stages of Japanese£ounder when the stomach is not fully developed(Muroga,Yasunobu, Okada & Masumura1990). In fact,the growth of this bacterial species is highly in£uencedby pHand pepsin concentration (Kim, Han, Kim, Lee &Park 2004), as well as by its inability to digest chitin(Sugita & Ito 2006). This would explain the importantreduction in this species and the increase in Shewanellasp. in the EXT production system. In addition, the ab-sence of pepsin-like activity during the ¢rst 30 daysafter hatching (DAH) (Ribeiro, Zambonino-Infante,Cahy & Dinis 1999), and the lack of gastric glandsbefore 27 DAH (Ribeiro, Sarasquete & Dinis 1999) inS. senegalensis suggest thatV. ichthyoenteri may be anopportunistic pathogen in this ¢sh species.This is espe-

cially important in larvae fed with commercial dietsin an intensive aquaculture system. Further studiesare required to demonstrate the possible role ofV. ichthyoenteri as a pathogen in Senegalese sole larvae.To determine the development of intestinal micro-

biota in adult specimens under di¡erent diets, an in-tensive experimental approach was carried out (Fig.1). As has been described previously,Vibrio sp. is thepredominant microorganism found in the adult ¢shgut. However, it should be highlighted that higherpercentages were obtained in animals fed a commer-cial diet (CD group) and deprived of food (ND group)(85^100%) in comparison with those specimens fedwith polychaeta (PD group) (51.7^85.9%). Changesin the microbiota from wild to arti¢cial fed animalshave also been reported in juvenile cod. Approxi-mately 50% of the bacteria fromwild ¢sh guts belongtoVibrio genus. After1-year feeding with an arti¢cialdiet, the non-pathogenic vibrios could not be de-tected in the guts, assuming that the diet could in£u-ence the persistence of commensal, non-pathogenicstrains that may help to confer protection against re-lated pathogenic strains (Str�m & Olafsen1990).Microbial heterogeneity often helps to reduce the

susceptibility of farmed animals towards opportunis-tic colonization by bacteria (Olafsen 2001). This isbased on the reduction in opportunistic bacteria,such as certainVibrio species. Particularly,V. harveyiand V. alginolyticus are considered to be two impor-tant pathogenic bacterial species frequently isolatedfrom diseased cultured Senegalese sole (Zorrillaet al. 2003). Our results demonstrate that diets basedon polychaeta brought about signi¢cant changes inthe intestinal microbiota composition, reducingthe percentage of vibrios. In addition, diets based onpolychaeta (Hediste diversicolor) are able to increasethe percentages of bacteria showing antimicrobialactivity up to 2.7^2.8% in S. senegalensis (Makridiset al. 2005). Although most of the strains assayedwith antimicrobial activity were classi¢ed asVibrio,two Shewanella species showed inhibition activity. Inthis respect, a signi¢cant increase in Shewanella spe-cies was observed in soles fed mainly with C. capitataand N. virens. These bacteria are inhabitants of es-tuarine environments, including seawater and sedi-ments, with oxic^anoxic interphases. They possessdi¡erent enzymatic activities such as tyrosinase,alkylsulphatase, chitinase and elastase, and are ableto use avarietyof di¡erent electronacceptors (Brettar& H˛£e,1993; Khashe & Janda,1998; LeCleir, Buchan& Hollibaugh 2004). These factors may favour thesurvival and competition to colonize the gut of

In£uence of feeding on intestinal microbiota BMartin-Antonio et al. Aquaculture Research, 2007, 38, 1213^1222

r 2007 TheAuthors1220 Journal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222

estuarine bottom-dwelling species such as Senega-lese sole.Microbiota composition of ¢sh and shell¢sh may

also vary as a function of stress and environmentalconditions, such as water temperature (Gatesoupe1999; Olsen, Sundell, Hansen, Hemre, Myklebust,Mayhew & Ringo 2002). Facultative anaerobic halo-philic bacteria (i.e.Vibrionaceae) have been associatedwith Mediterranean oysters during the warmestmonths (fromMarch to October). However, in the coldseason (November to February), Gram-negative oxi-dative bacterial groups corresponding to a-Proteo-bacteria were dominant (Pujalte, Ortigosa, Macian &Garay 1999). In addition, the intra-generic composi-tion of genus Vibrio can also change conspicuouslydepending on the water temperature (Sugita, Iwata,Miyajima, Kubo, Noguchi, Hashimoto & Deguchi1989; Pujalte et al. 1999). In our trials with soles cul-tured outdoor in earth ponds, a-Proteobacteria spe-cies were also identi¢ed in December, when thewater temperature was colder. In addition, vibriosclosely related toV. splendidus, a typical psychrotrophable to grow at 4 1C, were identi¢ed preferentiallyduring this cold period. In contrast, otherVibrio spe-cies were detected in March from soles cultured un-der intensive conditions in warmer temperatures. Infact, mesophilic species unable to grow at 4 1C, suchasV. harveyi orV. alginolyticus, were only found in theexperiments with soles reared indoor at18 1C. Di¡er-ences in optimal temperature for growth couldexplain the di¡erences observed for temporal sam-plings within each experimental group.In summary, our data stress the in£uence of

feeding regimes and environmental conditions onmicrobiota composition in S. senegalensis. Signi¢cantdi¡erences in intestinal bacterial composition wereobserved when soles were fed commercial diets ornatural preys. More studies are required to evaluatethe possible impact of these microbiota changes onnutrient assimilation, disease resistance or theapplication of probiotics in aquaculture. A betterperception of these factors will provide relevantinformation for the optimization of feeding for the in-tensive rearing of this species.

Acknowledgments

The authors thank Dr Jose Pedro Can� avate (IFAPACentro El Torun� o, El Puerto de Santa Mar|¤ a, Cadiz,Spain) for facilitating ¢sh specimens. This study wassupported bya JACUMAR grant and byan INTERRE-GIIIb project Atlantic Arc Aquaculture Group.

References

Bianchi M.A.G. & Bianchi A.J.M. (1982) Statistical samplingof bacterial strains and its use in bacterial diversity mea-surement.Microbial Ecology 8,61^69.

Brettar I. & H˛£e M.G. (1993) Nitrous oxide producing het-erotrophic bacteria from the water column of the centralBaltic: abundance and molecular identi¢cation. MarineEcology Progress Series 94, 253^265.

Cabral H.N. (2000) Comparative feeding ecology of sympa-tric Solea solea and S. senegalensis within the nurseryareas of the Tagus estuary, Portugal. Journal of FishBiology 57,1550^1562.

Can� avate J.P. & Fernandez-Diaz C. (1999) In£uence ofco-feeding larvae with live and inert diets on weaningthe sole, Solea senegalensis, onto commercial dry feeds.Aquaculture Research174, 255^263.

Cerda-Cuellar M. & Blanch A.R. (2002) Detectionand identi-¢cation ofVibrio scophthalmi in the intestinal microbiotaof ¢sh and evaluation of host speci¢city. Journal of AppliedMicrobiology 93, 261^268.

Clements K.D. (1997) Fermentation and gastrointestinalmicroorganisms in ¢shes. In: Gastrointestinal Microbiol-ogy,Vol. 1, Gastrointestinal Ecosystems and Fermentations(ed. by R.J. Mackie, B.A.Withe & R.E. Isaacson), pp. 156^198. Chapman and Hall Microbiology Series, Interna-tional Thomson Publishing, NewYork, USA.

de Groot S.J. (1971) On the interrelationships between mor-phology of the alimentary tract, food and feeding beha-viour in £at¢shes (Pisces: Pleuronectiformes). NetherlandJournal of Sea Research 5,121^196.

Eddy S.D. & Jones S.H. (2002) Microbiology of summer£ounder Paralichthys dentatus ¢ngerling production at amarine ¢sh hatchery. Aquaculture 211,9^28.

Garcia-Franquesa E., Molinero A.,Valero J. & Flos R. (1996)In£uence of sex, age and season on the feeding habits ofthe £at¢sh Solea senegalensis. Environmental Biology ofFish 47, 289^298.

Gatesoupe F.J. (1999) The use of probiotics in aquaculture.Aquaculture180,147^165.

Hansen G.H. & Olafsen J.A. (1999) Bacterial interactionsin early life stages of marine cold water ¢sh. MicrobialEcology 38,1^26.

Imsland A.K., Foss A., Conceic� ao L.E.C., Dinis M.T., DelbareD., Schram E., Kamstra A., Rema P. & White P. (2003) Areview of the culture potential of Solea solea and Soleasenegalensis. Review of Fish Biology and Fisheries 13,379^407.

Khashe S. & Janda M. (1998) Biochemical and pathogenicproperties of Shewanella alga and Shewanella putrefaciens.Journal of Clinical Microbiology 36,783^787.

Kim D.H., Han H.J., Kim S.M., Lee D.C. & Park S.I. (2004)Bacterial enteritis and the development of the larval di-gestive tract in olive £ounder, Paralichthys olivaceus(Temminck & Schlegel). Journal of Fish Diseases 27,497^505.

Aquaculture Research, 2007, 38, 1213^1222 In£uence of feeding on intestinal microbiota BMartin-Antonio et al.

r 2007 TheAuthorsJournal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222 1221

LeCleir G.R., Buchan A. & Hollibaugh J.T. (2004) Chitinasegene sequences retrieved from diverse aquatic habitatsreveal environment-speci¢c distribution. Applied andEnvironmental Microbiology 70,6977^6983.

MacDonald N.L., Stark J.R. & Austin B. (1986) Bacterial mi-cro£ora in the gastro-intestinal tract of Dover sole (Soleasolea L.), with emphasis on the possible role of bacteria inthe nutrition of the host. FEMS Microbiology Letters 35,107^111.

Makridis P., Martins S.,Tsalavouta M., Dionisio L.C., Kotou-las G., Magoulas A. & Dinis M.T. (2005) Antimicrobialactivity in bacteria isolated from Senegalese sole, Soleasenegalensis, fed with natural prey. Aquaculture Research5,1^9.

Marchesi J.R., Sato T.,Weightman A.J., Martin T.A., Fry J.C.,Hiom S.J. & Wade W.G. (1998) Design and evaluation ofuseful bacterium-speci¢c PCR primers that amplify genescoding for bacterial16S rRNA. Applied and EnvironmentalMicrobiology 64,795^799.

Moriarty D.J.W. (1990) Interactions of microorganisms andaquatic animals, particularly the nutritional role ofthe gut £ora. In: Microbiology in Poecilotherms (ed. by R.Lesel), pp. 217^222. Elsevier Science Publishers B.V.,Amsterdam, the Netherlands.

Muroga K., Yasunobu H., Okada N. & Masumura K. (1990)Bacterial enteritis of cultured £ounder Paralichthys oliva-ceus larvae. Diseases of Aquatic Organisms 9,121^125.

Olafsen J.A. (2001) Interactions between ¢sh larvae and bac-teria in marine aquaculture. Aquaculture 200, 223^247.

Olsen R.E., Sundell K., HansenT., Hemre G.I., Myklebust R.,Mayhew T.M. & Ringo E. (2002) Acute stress altersthe intestinal lining of Atlantic salmon, Salmo salar L.:an electron microscopical study. Fish Physiology andBiochemistry 26, 211^221.

Posada D. & Crandall K. (1998) Modeltest: testing the modelof DNA substitution. Bioinformatics14,817^818.

Pujalte M.J., Ortigosa M., Macian M.C. & Garay E. (1999)Aerobic and facultative anaerobic heterotrophicbacteria associated to Mediterranean oysters andseawater. International Microbiology 2, 259^266.

Ribeiro L., Sarasquete C. & Dinis M.T. (1999) Histological andhistochemical development of the digestive system ofSolea senegalensis Kaup, 1858 larvae. Aquaculture 179,291^306.

Ribeiro L., Zambonino-Infante J.L., Cahy C. & Dinis M.T.(1999) Development of digestive enzymes in larvae of So-lea senegalensis, Kaup1858. Aquaculture179, 465^473.

Ringo E. & Birbeck T.H. (1999) Intestinal micro£ora of ¢shlarvae and fry. Aquaculture Research 30,73^93.

Ringo E. & Vadstein O. (1998) Colonization ofVibrio pelagiusand Aeromonas caviae in early developing turbot,Scophthalmus maximus (L.) larvae. Journal of AppliedMicrobiology 84, 227^233.

Sa R., Bexiga C., Vieira L., Veiga P. & Erzini K. (2003)Diets of the sole Solea vulgaris Quensel,1806 and Solea se-

negalensis Kaup, 1858 in the lower estuary of the Guadi-ana River (Algarve, southern Portugal): preliminaryresults. Boletin del Instituto Espan� ol de Oceanogra¢a 19,505^508.

Sakata T. (1990) Micro£ora in the digestive tract of ¢sh andshell¢sh. In: Microbiology in Poecilotherms (ed. byR. Lesel), pp. 171^176. Elsevier Science Publishers B.V.,Amsterdam, the Netherlands.

Str�m E. & Olafsen J.A. (1990) The indigenous micro£ora ofwild-captured juvenile cod in net-pen rearing. In: Micro-biology in Poecilotherms (ed. by R. Le¤ sel), pp. 181^185.Elsevier Science Publishers B.V., Amsterdam, the Nether-lands.

Sugita H. & ItoY. (2006) Identi¢cation of intestinal bacteriafrom Japanese £ounder (Paralichthys olivaceus) and theirability to digest chitin. Letters in Applied Microbiology 43,336^342.

Sugita H., Iwata J., Miyajima C., KuboT., Noguchi T., Hashi-moto K. & Deguchi Y. (1989) Changes in micro£ora of apu¡er ¢sh, Fugu niphobles, with di¡erent water tempera-tures.Marine Biology101, 299^304.

Sugita H., OkanoR., SuzukiY., Iwai D., MizukamiM., Akiya-ma N. & Matsuura S. (2002) Antibacterial abilities of in-testinal bacteria from larval and juvenile Japanese£ounder against ¢sh pathogens. Fisheries Science 68,1004^1011.

Toranzo A.E., Novoa B., Romalde J.L., Nun� ez S., Devesa S.,Marin� o E., Silva R., Martinez E., Figueras A. & Barja J.L.(1993) Micro£ora associated with healthy and diseasedturbot (Scophthalmus maximus) from three farms innorthwest Spain. Aquaculture114,189^202.

Verschuere L., Rombaut G., Sorgeloos P. & Verstraete W.(2000) Probiotic bacteria as biological control agents inaquaculture. Microbiology and Molecular Biology Reviews64,655^671.

Westerdahl A., Olsson J.C., Kjelleberg S. & Conway P.L. (1991)Isolation and characterization of turbot (Scophthalmusmaximus)-associated bacteria with inhibitory e¡ectsagainstVibrio anguillarum. Applied and Environmental Mi-crobiology 57, 2223^2228.

Winton J.R. (2001) Fish health management. In: FishHatchery Management, 2nd edn, ed. by G. Wedemeyer),pp. 559^639. American Fisheries Society, Bethesda,USA.

Zorrilla I., Balebona M.C., Morin� igo M.A., Sarasquete C. &Borrego J.J. (1999) Isolation and characterizationof the causative agent of pasteurellosis, Photobacteriumdamselae subsp. piscicida, from sole, Solea senegalensis(Kaup). Journal of Fish Diseases 22,167^172.

Zorrilla I., Arijo S., Chabrillon M., Diaz P., Martinez-Manza-nares E., Balebona M.C. & Morin� igo M.A. (2003) Vibriospecies isolated from diseased farmed sole, Solea senega-lensis (Kaup), and evaluation of the potential virulencerole of their extracellular products. Journal of FishDiseases26,103^108.

In£uence of feeding on intestinal microbiota BMartin-Antonio et al. Aquaculture Research, 2007, 38, 1213^1222

r 2007 TheAuthors1222 Journal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 1213^1222