Sttr

RaIra Cb R

1.

opminbire20dupr

Veterinary Microbiology 169 (2014) 80–88

A

Art

Re

Re

Ac

Ke

Ra

Sk

Sta

Vib

Pa

*

03

htt

aphylococcus warneri, a resident skin commensal of rainbowout (Oncorhynchus mykiss) with pathobiont characteristics

mi Musharrafieh a, Luca Tacchi a, Joshua Trujeque a, Scott LaPatra b,ene Salinas a,*

enter for Evolutionary and Theoretical Immunology, University of New Mexico, Albuquerque, NM, USA

esearch Division, Clear Springs Food Inc., Buhl, ID, USA

Introduction

Commensal microorganisms provide metabolic, devel-mental and immunological functions to the host (Som-er and Backhed, 2013). but they are, however, not alwaysnocuous. When homeostasis is breached, invasive sym-onts, or pathobionts, prompt abnormal inflammatorysponses potentially causing disease (Belkaid and Naik,13; Kamada et al., 2013). Pathobionts are overrepresentedring dysbiosis situations, which arise from geneticedispositions, exposure to environmental or metabolic

stressors, or alteration of the normal microbial communities(Round and Mazmanian, 2009; Packey and Sartor, 2009).

The delineation between pathogens and commensals isnot always easy to make and it is mostly defined by theimmune responses they trigger. One of the current dogmasestablishes that successful colonization by a commensalmicroorganism relies on the induction of anti-inflammatoryresponses, often mediated by the cytokine TGF-b (Detour-nay et al., 2012) that leads to tolerance of the commensal bythe host. Other theories point to the opposite, postulatingthat commensals must generate a stereotypical inflamma-tory cascade when establishing a symbiosis with their host(Nussbaum and Locksley, 2012).

In teleost fish, the gills, gut and skin are the main mucosalsurfaces harboring diverse microbial communities. The skinis thought to be the largest immunologically active organ

R T I C L E I N F O

icle history:

ceived 4 September 2013

ceived in revised form 10 December 2013

cepted 13 December 2013

ywords:

inbow trout

in

phylococcus warneri

rio anguillarum

thobiont

A B S T R A C T

Commensal microorganisms live in association with the mucosal surfaces of all

vertebrates. The skin of teleost fish is known to harbor commensals. In this study we

report for the first time the presence of an intracellular Gram positive bacteria,

Staphylococcus warneri that resides in the skin epidermis of rainbow trout (Oncorhynchus

mykiss). S. warneri was isolated from healthy hatchery trout skin epithelial cells. In situ

hybridization confirmed the intracellular nature of the bacterium. Skin explants exposed

in vitro to S. warneri or the extracellular pathogen Vibrio anguillarum show that S. warneri is

able to induce an anti-inflammatory cytokine status via TGF-b1b compared to the pro-

inflammatory responses (IL-1b, IL-6 and TNF-/) elicited by V. anguillarum. In vivo

experiments showed that S. warneri is not pathogenic to rainbow trout when injected

intraperitoneally at high concentrations. However, S. warneri is able to stimulate V.

anguillarum growth and biofilm formation on rainbow trout scales. Our results

demonstrate that rainbow trout skin commensals such as S. warneri have the potential

to become indirect pathobionts by enhancing growth and biofilm formation of pathogens

such as V. anguillarum. These results show that fish farming practices (i.e. handling and

other manipulations) can alter the skin microbiota and compromise the skin health of

rainbow trout.

� 2013 Elsevier B.V. All rights reserved.

Corresponding author. Tel.: +1 505 2770039; fax: +1 505 2773411.

E-mail address: isalinas@unm.edu (I. Salinas).

Contents lists available at ScienceDirect

Veterinary Microbiology

jou r nal h o mep ag e: w ww .e ls evier . co m/lo c ate /vetm i c

78-1135/$ – see front matter � 2013 Elsevier B.V. All rights reserved.

p://dx.doi.org/10.1016/j.vetmic.2013.12.012

abdckSeasm

toliTfiCrMd(Oa(Gw

apinn

2

2

afrfiuNb

T

P

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–88 81

nd commensals may often assist in the homeostasis of thisarrier and contribute to the hosts’ repertoire of immuneefense mechanisms against pathogens. In fish, the skin isomposed of a layer of living epithelial cells, with noeratinization and abundant mucus-secreting cells. TeleostALT (skin-associated lymphoid tissue) is continuouslyxposed to diverse microbial stimuli (including commensalsnd pathogens) as well as environmental and mechanicaltressors (Salinas et al., 2011; Esteban, 2012) and it is able toount gut-like immune responses (Xu et al., 2013).

The Gram positive Staphylococcus spp. can be pathogenic their fish hosts, causing exophthalmia and septicaemia-

ke symptoms in fish that have been infected (Shah andyagy, 1986), although they have also been reported fromsh in the absence of disease (Spanggaard et al., 2000;antas et al., 2012). S. warneri includes several strainseported as pathogenic to humans (Campoccia et al., 2010).

oreover, S. warneri has been isolated and grown fromiscolored kidneys and livers of diseased rainbow troutncorhynchus mykiss) that displayed ulcerations on the fins

nd exophthalmia, along with ascetic fluid in the abdomensil et al., 2000). In this study, we identify for the first time S.

arneri as a resident commensal of rainbow trout skin. Theim of the study was to examine the role S. warneri as aossible pathobiont for rainbow trout as well as theteractions between S. warneri and the common Gram

egative pathogen, Vibrio anguillarum.

. Materials and methods

.1. Isolation and identification of S. warneri

The bacterium was isolated from the skin of healthydult triploid rainbow trout (mean weight 250 g) obtainedom Lisboa Springs Hatchery, Pecos, New Mexico wheresh were maintained in concrete raceways with contin-ous water flow. Fish were sampled during the months ofovember and December when the water temperature isetween 8 and 13 8C. Health status was only evaluated

based on external signs: no ulcers, no bleeding, lack ofexternal parasites, active swimming and feeding behavior.No further tests were conducted to assess presence ofinternal infections. Both skin mucus samples and skinsamples without mucus were used for bacterial isolation.Bacteria present in the mucus were isolated as explainedelsewhere (Xu et al., 2013). 10 ml aliquots were plated inLuria broth (LB) agar plates or Tryptic Soy Agar (TSA)plates. Additionally, three fish were used to isolate possiblebacteria living in the skin. To that end, after the mucus hadbeen scraped, the skin was sprayed with 70% ethanol and a2 cm2 section of skin was immediately dissected andplaced in HBSS containing penicillin and streptomycin(100 U/ml and 100 mg/ml, respectively). After shaking atroom temperature for 2 h, the skin sample was transferredto a Petri dish containing HBSS without antibiotics. At thispoint the skin was finely minced (over 100 times) and cellswere further lysed using a 1 ml syringe with a needle. Thesuspension was vortexed and then centrifuged at 3000 � g,10 min. The pellet was resuspended in sterile PBS andplated in either LB or TSA plates. After 72 h, three differenttypes of colonies could be observed. One of these was sub-cultured onto TSA plates. Colonies were small, pale yellowwith round, smooth edges. The cultures matched thedefinition of S. warneri according to Bergey’s Manual ofSystematic Bacteriology (2011). The same type of colonywas found in the mucus-derived cultures. The pure isolateculture was identified at the Tricore laboratories (Albu-querque, NM) by means of Gram stain and MALDI-TOF.These tests revealed that the isolate was a Gram positivecocci, S. warneri. The identity was further confirmed by PCRusing S. warneri specific 16s rDNA primers (Table 1) (Iwaseet al., 2007), cloning and sequencing. PCR used thefollowing cycles: 94 8C for 5 min, then 30 cycles of 94 8Cfor 30 s, 50 8C for 30 s, 72 8C for 1.5 min and a finalextension of 72 8C for 10 min. All PCR products were clonedand sequenced to confirm their identity as explainedpreviously (Tacchi et al., 2013). The obtained sequence hada homology of 100% with S. warneri.

able 1

rimers used in the present study.

Gene Primer Sequence (50–30) Application

IL1b IL1bF ACATTGCCAACCTCATCATCG qPCR

IL1bR TTGAGCAGGTCCTTGTCCTTG qPCR

TNFa TNFaF GGGGACAAACTGTGGACTGA qPCR

TNFaR GAAGTTCTTGCCCTGCTCTG qPCR

IL6 IL6F ACTCCCCTCTGTCACACACC qPCR

IL6R GGCAGACAGGTCCTCCACTA qPCR

TGFb1a TGFb1aF CTCACATTTTACTGATGTCACTTCCTGT qPCR

TGFb1aR GGACAACTGCTCCACCTTGTG qPCR

TGFb1b TGFb1bF CATGTCCATCCCCCAGAACT qPCR

TGFb1bR GGACAACTGTTCCACCTTGTGTT qPCR

EF-1a EF-1aF CAACGATATCCGTCGTGGCA qPCR

EF-1aR ACAGCGAAACGACCAAGAGG qPCR

Swar 16s SwarF TGTAGCTAACTTAGATAGTGTTCCTTCT RT-PCR

SwarR CCGCCACCGTTATTTCTT RT-PCR

2.2

thSptrafro

2.3

hy

thpr50

ge(5bithanscRe

2.4

str

wsecuunsage(2orpe

2.5

wExorpoextoprsidDNcatrastrin

2.6

in

1bcyus

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–8882

. Fish and stress experiments

Rainbow trout (n = 6) were either sampled directly frome hatchery rafts or after a 5 h transport from the Lisboaring hatchery to Tingley beach, Albuquerque. Thensport was conducted in a truck, using raceway waterm the hatchery without any sedation.

. Localization of S. warneri by fluorescent in situ

bridization (FISH)

In order to investigate the localization of S. warneri ine skin of rainbow trout, a FISH probe was designed. Theobe (Eurofins MWG Operon) was labeled with Cy5 in theend and targeted the superoxide dismutase A (sodA)ne of S. warneri. Control trout skin cryosections (n = 6)

mm-thick) were fixed in formalin for 10 min, permea-lized overnight in 70% ethanol, hybridized at 37 8C withe S. warneri probe, stained with the nuclear stain DAPId then observed under a Nikon Ti fluorescent micro-ope. Images were analyzed using Nis Elements Advancedsearch software.

. Quantification of S. warneri in the skin of control and

essed rainbow trout by PCR

The presence of S. warneri in the skin of rainbow troutas quantified by qPCR using a standard curve made ofrial dilutions of a pure S. warneri culture. The standardrve ranged from 109 to 10 S. warneri colony formingits (cfu). Total DNA was obtained from skin tissuemples of control or stressed rainbow trout. The sodAne of S. warneri was amplified according to Iwase et al.,007) using specific primers (Table 1). 100 ng of S. warneri

trout skin DNA were amplified by qPCR. The qPCR wasrformed as previously described (Tacchi et al., 2013).

. In vitro exposure of skin explants to S. warneri

Control rainbow trout skin explants (n = 5) (0.5 cm2)ere surface sterilized and placed in 24-well plates.plants were exposed to 0, 102, 104 or 106 S. warneri cells

106 V. anguillarum cells for 6, 24 or 48 h. At each timeint, explants were collected and placed in Trizol for RNAtraction. cDNA synthesis was performed using 1 mg oftal RNA, which was denatured (65 8C, 5 min) in theesence of 1 ml of oligo-dT17, 1 ml dNTP (deoxynucleo-e triphosphate mix 10 mM each (Promega) and RNA/A free water (Sigma) in a volume of 13 ml. Synthesis was

rried out using 1 ml Superscript III enzyme reversenscriptase (Invitrogen) in the presence of 5 ml of 5� firstand buffer, 1 ml 0.1 M DTT (final volume of 25 ml) and

cubated at 55 8C for 1 h.

. qPCR studies: expression of pro-inflammatory and anti-

flammatory cytokines

The expression of the pro-inflammatory cytokines IL-, Il-6 and TNF-a as well as the anti-inflammatory

tokines TGF-b1a and TGF-b1b was studied by RT-qPCRing specific primers (Table 1). The qPCR was performed

as described above using 3 ml of a diluted cDNA template.Rainbow trout elongation factor EF-1a was used as controlgene for normalization of expression. The relative expres-sion level of the genes was determined using the Pfafflmethod (Pfaffl, 2001) as previously described (Tacchi et al.,2013). The sequences of all the PCR products amplified byqPCR were further confirmed by cloning.

2.7. Effect of S. warneri on the growth of the pathogen V.

anguillarum

GFP-V. anguillarum was grown in the presence orabsence of S. warneri. V. anguillarum was grown for 24 h inLB at 24 8C and cultures were adjusted to an absorbance of1.4 at 600 nm for a concentration of 8.4 � 108 cfu/ml. Atotal of 104 cfu were added to each well of 96-well plates intriplicate. S. warneri cultures grown for 24 h (also in LB)were added to the wells at 10, 102, 104 or 106 cfu/well.Positive control consisted of wells containing V. anguil-

larum only, while negative controls consisted of wellscontaining S. warneri only and LB only. GFP fluorescencewas measured at 3, 6, 21 and 29 h in a Synergy H1 platereader at an excitation wavelength of 485 nm and emissionwavelength of 538 nm.

2.8. Effect of S. warneri on V. anguillarum biofilm formation

Biofilm formation on rainbow trout scales was assayedas described previously by Croxatto et al. (2007). Briefly,scales were collected from the lateral line and kept at 12 8Cfor 48 h before infection. GFP-expressing V. anguillarum

was added (104 cells/well) as well as S. warneri (in dilutionsof 102, 104, 106 cells/well). After 48 h the media wasremoved and the scales were fixed with 100% methanol,air-dried and mounted. The images acquired wereanalyzed using NIS Elements Advanced Research software.The mean green fluorescence intensity of each capturedimage was divided by the surface area value. This ratio wasused to quantify the mean relative biofilm formation valuefor each treatment.

2.9. In vivo injection of S. warneri

S. warneri was grown overnight in LB, washed in PBSand injected intraperitoneally (i.p.) into healthy rainbowtrout (mean weight = 3 g). Rainbow trout (n = 40) wereobtained from Clear Springs Foods Inc. (Buhl, ID) andreceived a 50 ml intraperitoneal (i.p.) injection containing1 � 107, 3 � 108 or 1 � 109 S. warneri cfu/fish. Control groupreceived an i.p. injection of PBS. Fish were monitored forany signs of disease or mortality for four weeks after theinjection.

2.10. Statistical analysis

Data is presented as mean � standard error (SE). qPCRmeasurements were analyzed by t-test by comparing valueswith the control group. One-way ANOVA analysis followed byTukey’s post hoc test was used to identify differencesbetween treatments in the biofilm formation experiments.p-Values < 0.05 were considered significant.

3

3

r(Fc

F

5

fi

c

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–88 83

. Results

.1. Localization of S. warneri in the skin of rainbow trout

FISH staining on rainbow trout skin cryosectionsevealed the presence of S. warneri both in the mucusig. 1A) and in the skin tissue, mostly in the dermis

ompartment (Fig. 1B).

3.2. The numbers of S. warneri cfu in the skin increase during

a transport event

Using a standard curve and a qPCR assay, we wereable to quantify the number of cfu present in the skin ofrainbow trout before and after a transport event. AsFig. 2 shows, the numbers of S. warneri cfu present in theskin of rainbow trout were significantly higher in the

ig. 1. Localization of S. warneri in hatchery rainbow trout skin by FISH. Skin cryosections were hybridized with a S. warneri specific probe labeled with cynanine

(magenta). DAPI-stained cell nuclei are shown in blue. The fluorescent images were overlaid with a differential interference contrast (DIC) image of the same

eld. (A) S. warneri is present in the dermis layer of the skin and (B) S. warneri is present in the mucus layer of rainbow trout. Arrows point S. warneri cells. CHR:

hromatophores, EC: epithelial cells. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig

the

a t

Fig

co

su

dif

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–8884

‘‘handled’’ group (post transport) compared to the controlgroup.

3.3. Expression of pro-inflammatory and anti-inflammatory

cytokines in rainbow trout skin explants exposed to S. warneri

or V. anguillarum

The expression of three pro-inflammatory (IL1b, TNFaand IL6) and two anti-inflammatory cytokines (TGFb1aand TGFb1b) was measured in the skin of rainbow trout atthree different time points following in vitro exposure tothree different concentrations of S. warneri. As shown inFig. 3A–D, the commensal S. warneri and the pathogen V.

anguillarum induced different cytokine expression profilesin the skin of rainbow trout. At 6 h, S. warneri stimulatedcells showed a suppression of the inflammatory responsewith all the pro-inflammatory molecules (and TGFb1a)down-regulated and the down-regulation decreased withincreasing concentrations of the commensal. However, the

. 2. Quantification of the number of S. warneri colony forming units in

skin (both mucus and tissue) of rainbow trout before (control) or after

ransport event by qPCR (n = 6).

. 3. Expression of pro- and anti-inflammatory cytokines in rainbow trout skin explants following incubation with S. warneri or V. anguillarum at different

ncentrations by qPCR. (A) Cytokine expression levels after 6 h, (B) cytokine expression levels after 24 h, (C) cytokine expression levels after 48 h, (D) visual

mmary of all results from (A)–(C). Expression levels were put into eight different categories according to the figure legend. Asterisks indicate significant

ferences compared with the control treatment.

arwoTstimnrtrvva

w

3

a

S

2logeetwati1a

F

m

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–88 85

nti-inflammatory cytokine TGF1bb was found to be up-egulated significantly at 6 h (P values < 0.05) when cellsere stimulated with S. warneri. V. anguillarum, on the

ther hand, induced the expression of IL1b, TNFa, IL6 andGFb1a and a suppression of TGF1bb. 24 h followingtimulation, both bacteria induced a dramatic up-regula-on of IL1b, TNFa, IL6 and TGFb1a (in a dose dependentanner in S. warneri stimulated cells), whilst TGF1bb was

ot modified in expression. At 48 h, a significant down-egulation of the genes studied was observed in theeatment with the lowest concentration of S. warneri (P

alue < 0.05) and in the V. anguillarum treatment (P

alue < 0.01). The suppression was more evident in V.

nguillarum-stimulated skin explants compared to S.

arneri stimulated skin explants.

.4. Direct effects of S. warneri on the growth of GFP-V.

nguillarum

The growth of V. anguillarum in the presence of different. warneri concentrations was measured over a period of9 h. Four concentrations of S. warneri were tested. Thewest dose (10 cfu/ml) had no effect on V. anguillarum

rowth but the rest of the concentrations tested, allnhanced V. anguillarum growth (Fig. 4A–D). The greatestnhancement occurred after 29 h of culture although the

o highest concentrations tested already had an effectfter 21 h. At 29 h, V. anguillarum growth was �4 times, 7mes and 6 times higher in the presence of 102, 104 and06 S. warneri cfu/ml, respectively, compared to V.

nguillarum alone.

3.5. Effect of S. warneri on V. anguillarum biofilm formation

on trout scales

We measured the ability of GFP-V. anguillarum to formbiofilms on the scales of control rainbow trout in thepresence or absence of S. warneri. Biofilm formation wasenhanced after 48 h by the presence of S. warneri at alldoses tested (Fig. 5).

3.6. S. warneri is not pathogenic to rainbow trout

Challenge experiments using three different doses of S.

warneri by i.p. injection caused no mortality or morbidityin rainbow trout over a period of four weeks. Only one fishdied at the highest assayed dose (109 cfu/fish) 6 days post-injection. Bacterial colonies with similar morphologicalcharacteristics to S. warneri were recovered on TSA platesfrom the kidney of the only fish that died. The ID of thebacterial colonies was confirmed by 16s rDNA sequencing(sequence homology 100%).

4. Discussion

Commensal microorganisms bring many benefits to thevertebrate host colonizing different compartments withinthe mucosal epithelia but they are not always innocuous tothe host if homeostasis is breached at the mucosal barriers.

In the present study, we report for the first time thepresence of a Gram positive bacterium, S. warneri thatresides in the skin of hatchery-reared rainbow trout in NewMexico. S. warneri is a common resident of the human skin

ig. 4. V. anguillarum growth curves in the presence of absence of S. warneri at (A) 10 cfu/ml, (B) 102 cfu/ml, (C) 104 cfu/ml and (D) 106 cfu/ml. Growth was

easured by quantifying the GFP fluorescence units at each time point in a plate reader. Results are representative of three independent experiments.

thprInfoskwa

ceau

th(Fwcuofskprtro

cotracaba

Fig

flu

DIC

48

ind

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–8886

at can cause disease or even be fatal in immunocom-omised patients and infants (Campoccia et al., 2010).terestingly, in the present study, S. warneri was not onlyund in association with the mucus but also inside thein. This was shown both by FISH staining with an S.

rneri specific probe and also by culturing lysates of skinlls on agar plates. Other Staphylococcus spp., such as S.

reus, has been shown to invade host cells, replicating ine cytoplasm or in the phagolysosome of phagocytesraunholz and Sinha, 2012), and can also internalizeithin keratinocytes (Kintarak et al., 2004). From ourrrent study, it appears that S. warneri has similar abilities

internalizing within host; specifically in rainbow troutin epithelial cells. We found that S. warneri is bothesent in the skin mucus and the skin tissue of rainbowut.We quantified the number of S. warneri cfu in the skin of

ntrol and ‘‘handled’’ rainbow trout following a 5 hnsport event. Our results clearly indicate that transport

uses important changes in some skin commensalcteria. In the case of S. warneri, numbers were

considerably higher following transport than beforeindicating that this commensal has the ability to growbetter when the host is under stress. Our results are inagreement with a previous report on rainbow trouthindgut microbiota, which showed that acute stress altersthe concentration of Staphyloccocus spp. (Olsen et al.,2005). In contrast to our results, Staphylococcus spp.concentrations decreased compared to the increasedconcentrations reported here. Our observations in theskin have important implications for the management ofhatcheries and any other fish farming facilities, wheretransport events are common and therefore the composi-tion of fish skin-associated bacteria is likely to change.

Pathogens and commensals are known to inducedifferent signatures of cytokine production by the host(Maehr et al., 2013). Whilst pathogens induce pro-inflammatory responses, commensals, in turn, elicit anti-inflammatory cytokines. In rainbow trout, there are twoisoforms of the cytokine TGF-b1 currently known, TGF-b1a and TGF-b1b. Whether TGF-b plays a role in the skinof rainbow trout in response to commensals or pathogens

. 5. V. anguillarum biofilm formation on rainbow trout skin is enhanced by the presence of S. warneri. (A) Results are shown as the mean � SE of the GFP

orescence units per unit of area after analyzing 10 different scales per experimental group by fluorescence microscopy. (B) Fluorescent microscopy image and

overlay of a control scale incubated with S. warneri for 48 h. (C) Fluorescent microscopy image and DIC overlay of a scale incubated with GFP-V. anguillarum for

h. (D) Fluorescent microscopy image and DIC overlay of a control scale incubated with GFP-V. anguillarum and S. warneri (102 cfu/ml) for 48 h. Different letters

icate significantly different groups.

hhccew

spopkalolempbmthhpc

htoew

mw

eebbsbsoleew

sdcths

hdFrseptovscsiscc

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–88 87

ad not been investigated to date. In order to test ourypothesis that anti-inflammatory and pro-inflammatoryytokines may be differentially modulated in response toommensal or pathogen colonization, we studied thexpression of five cytokines in skin explants exposed to S.

arneri or V. anguillarum in vitro. Our results show that thekin of rainbow trout responds to commensals andathogens in a different way, by inducing the expressionf different cytokine subsets. Generally speaking, theathogen V. anguillarum induces a pro-inflammatory cyto-ine signature whereas the commensal S. warneri inducesnti-inflammatory cytokines (TGF-b1b) when present atw concentrations. High S. warneri concentrations, in turn,d to induction of IL1-b expression mirroring an inflam-atoryresponseagainstthecommensalsimilartothatofthe

athogen. Interestingly, the two TGF-b1 forms studiedehaved in different fashions, with TGF-b1a displaying aore pro-inflammatory like response, and TGF-b1b beinge stereotypical anti-inflammatory cytokine. These results

ighlight the delicate line that divides commensals andathogens and how, if present at high concentrations,ommensals evoke pathogen-like responses.

Extensive studies have revealed that disruption of theost commensal communities increases host susceptibility

pathogenic infections (Belkaid and Naik, 2013; Kamadat al., 2013). Our results indicate that the presence of S.

arneri has a direct effect on V. anguillarum growth. Thisay be due to production of certain metabolites by S.

arneri that are beneficial for V. anguillarum, although thexact mechanism for this interaction remains to belucidated. Next, we examined whether V. anguillarum

iofilm formation on rainbow trout scales may be affectedy S. warneri in vitro. Biofilms contain complex poly-accharide polymers within its matrix and can be producedy various bacteria. Biofilm formation can be tolerated andometimes beneficial to the host (Visick, 2009), butverproduction, especially in the case of pathogens, mayad to increase bacterial colonization and disease (Wahl

t al., 2012). Our results show that in the presence of S.

arneri the ability of V. anguillarum to form biofilms wasignificantly enhanced. Together, these experimentsemonstrate that the skin residing commensal, S. warneri,an act as a catalyst for the growth and biofilm formation of

e pathogen V. anguillarum, likely increasing infectionuccess.

Staphyloccocus spp. such as S. aureus and S. epidermidis

ave been associated with the aquatic environment andiseased aquatic organisms (Kusuda and Sugiyama, 1981).urthermore, S. warneri was once isolated from diseasedainbow trout in Spain (Gil et al., 2000). The isolate in thattudy, Y-13-L, was tested for pathogenicity in brown troutxposed to a high temperature stress (20 8C). However,athogenicity was not tested in rainbow trout. According

the present study, S. warneri is not pathogenic (or has aery low pathogenicity) to rainbow trout (at least to thetrain here used) even when injected at very highoncentrations (109/fish). The differences between bothtudies may be due to differences between the bacterialolates, the hosts (brown versus rainbow trout) and/or the

hallenge conditions (temperature stress versus thermal

In conclusion, S. warneri is part of the residentmicrobiota of the trout skin. At low concentrations, S.

warneri appears to be innocuous to the host inducing anti-inflammatory cytokines. However, dysbiosis or any otherbreak down of skin homeostasis may lead to overgrowth ofthis commensal. In this situation, an inflammatoryresponse in the skin takes place. Despite the fact that S.

warneri does not appear to be a direct pathobiont inrainbow trout (causing disease if reaching the systemiccirculation) it acts as an indirect pathobiont aidingpathogens (in this case V. anguillarum) to grow andcolonize the host. The present study underscores thedelicate balance between the fish mucosal immune systemand its associated commensal microbial communities andhighlights the potential role of commensals in the onset ofaquatic skin diseases.

Acknowledgements

We thank Dr D. Milton for kindly providing the GFP-V.

anguillarum strain and the Lisboa Spring Hatchery for therainbow trout. This work was funded by NIH COBRE grantP20GM103452.

References

Belkaid, Y., Naik, S., 2013. Compartmentalized and systemic control oftissue immunity by commensals. Nat. Immunol. 14, 646–653.

Campoccia, D., Montanaro, L., Visai, L., Corazzari, T., Poggio, C., Maso, A.,Pirini, V., Ravaioli, S., Cangini, I., Speziale, P., Arciola, C.R., 2010.Characterization of 26 Staphylococcus warneri isolates from orthope-dic infections. Int. J. Artif. Organs 33, 575–581.

Cantas, L., Sørby, J.R., Alestrom, P., Sørum, H., 2012. Culturable gutmicrobiota diversity in zebrafish. Zebrafish 9, e0712.

Croxatto, A., Lauritz, J., Chen, C., Milton, D., 2007. Vibrio anguillarumcolonization of rainbow trout integument requires a DNA locusinvolved in exopolysaccharide transport and biosynthesis. Environ.Microbiol. 9, 370–382.

Detournay, O., Schnitzler, C.E., Poole, A., Weis, V.M., 2012. Regulation ofcnidarians-dinoflagellate mutualisms, evidence that activation of ahost TGFb innate immune innate immune pathway promotes toler-ance of the symbiont. Dev. Comp. Immunol. 38, 525–537.

De Vos, P., Boone, D.R., Garrity, G.M., Castenholz, R.W., Brenner, D.J., Krieg,N.R., Staley, J.T., 2011. Bergey’s Manual of Systematic Bacteriology:Volume 3: The Firmicutes. Springer (Online).

Esteban, M., 2012. An overview of the immunological defenses in fishskin. ISRN Immunol. 2012 , 29 pp.

Fraunholz, M., Sinha, B., 2012. Intracellular Staphylococcus aureus: live-inand let die. Front. Cell Infect. Microbiol. 2, 43.

Gil, P., Vivas, J., Gallardo, C.S., Rodrıguez, L.A., 2000. First isolation ofStaphylococcus warneri, from diseased rainbow trout, Oncorhynchusmykiss (Walbaum), in Northwest Spain. J. Fish Dis. 23, 295–298.

Iwase, T., Seki, K., Shinji, H., Mizunoe, Y., Masuda, S., 2007. Development ofa real-time PCR assay for the detection and identification of Staphy-lococcus capitis, Staphylococcus haemolyticus and Staphylococcus war-neri. J. Med. Microbiol. 56, 1346–1349.

Kamada, N., Chen, G.Y., Inohara, N., Nunez, G., 2013. Control of pathogensand pathobionts by the gut microbiota. Nat. Immunol. 14, 685–690.

Kintarak, S., Nair, S.P., Speight, P.M., Whawell, S.A., 2004. A recombinantfragment of the fibronectin-binding protein of Staphylococcusaureus inhibits keratinocyte migration. Arch. Dermatol. Res. 296,250–257.

Kusuda, R., Sugiyama, A., 1981. Studies on the characters of Staphylococcusepidermidis isolated from diseased fish. 1. On the morphological,biological, and biochemical properties. Fish Pathol. 16, 15–23.

Maehr, T., Costa, M.M., Vecino, J.L., Wadsworth, S., Martin, S.A., Wang, T.,Secombes, C.J., 2013. Transforming growth factor-b1b: a second TGF-b1 paralogue in the rainbow trout (Oncorhynchus mykiss) that has alower constitutive expression but is more responsive to immune

stimulation. Fish Shellfish Immunol. 34, 420–432. omfort).

Nu

Ols

Pa

Pfa

Ro

Sa

Sh

R. Musharrafieh et al. / Veterinary Microbiology 169 (2014) 80–8888

ssbaum, J.C., Locksley, R.M., 2012. Infectious (Non)tolerance-frustratedcommensalism gone awry? Cold Spring Harb. Perspect. Biol., http://dx.doi.org/10.1101/cshperspect.a007328.

en, R.E., Sundell, K., Mayhew, T.E., Myklebust, R., Ringø, E., 2005. Acutestress alters intestinal function of rainbow trout, Oncorhynchus mykiss(Walbaum). Aquaculture 250, 480–495.

ckey, C., Sartor, R., 2009. Commensal bacteria, traditional and oppor-tunistic pathogens, dysbiosis and bacteria killing in inflammatorybowel diseases. Curr. Opin. Infect. Dis. 22, 292–301.

ffl, M.W., 2001. A new mathematical model for relative quantificationin real-time RT-PCR. Nucleic Acids Res. 29 , http://dx.doi.org/10.1093/nar/29.9.e45.

und, J., Mazmanian, S., 2009. The gut microbiota shapes intestinalimmune responses during health and disease. Nat. Rev. Immunol.9, 313–323.

linas, I., Zhang, Y.A., Sunyer, J.O., 2011. Mucosal immunoglobulins and Bcells of teleost fish. Dev. Comp. Immunol. 35, 1346–1365.

ah, K.L., Tyagy, B.C., 1986. An eye disease in silver carp, Hypophthal-michthys molitrix, held in tropical ponds, associated with the bacter-ium Staphylococcus aureus. Aquaculture 55, 1–4.

Sommer, S., Backhed, F., 2013. The gut microbiota – masters of hostdevelopment and physiology. Nat. Rev. Microbiol. 11, 227–238.

Spanggaard, B., Huber, I., Nielsen, J., Nielsen, T., Appel, K.F., Gram, L., 2000.The microflora of rainbow trout intestine: a comparison of traditionaland molecular identification. Aquaculture 182, 1–15.

Tacchi, L., Larragoite, E., Salinas, I., 2013. Discovery of J chain in AfricanLungfish (Protopterus dolloi, Sarcopterygii) using high throughputtranscriptome sequencing: implications in mucosal immunity. PLOSONE 8 (8) e70650, http://dx.doi.org/10.1371/journal.pone.0070650.

Visick, K.L., 2009. An intricate network of regulators controls biofilmformation and colonization by Vibrio fischeri. Mol. Microbiol. 74, 782–789.

Wahl, M., Goecke, F., Labes, A., Dobretsov, S., Weinberger, F., 2012. Thesecond skin: ecological role of epibiotic biofilms on marine organ-isms. Front. Microbiol., http://dx.doi.org/10.3389/fmicb.2012.00292.

Xu, Z., Parra, D., Gomez, D., Salinas, I., Zhang, Y-A., von Gersdorff Jørgen-sen, L., Heinecke, R.D., Buchmann, K., LaPatra, S., Sunyer, J.O., 2013.Tleost skin, an ancient mucosal surface that elicits gut-like immuneresponses. PNAS 110 (32) 13097–13102.