Divergence
of
a
strain
of
Pseudomonas
aeruginosa
during
an
outbreak
of
ovine
mastitis
Elli
A.
Wright
a,1,
Valeria
Di
Lorenzo
b,1,
Claudia
Trappetti
b,
Manuele
Liciardi
c,
Germano
Orru
d,
Carlo
Viti
e,
Christina
Bronowski
a,
Amanda
J.
Hall
a,
Alistair
C.
Darby
f,
Marco
R.
Oggioni
b,g,h,2,**
,
Craig
Winstanley
a,2,*
a
InstituteofInfectionandGlobalHealth,UniversityofLiverpool,LiverpoolL697BE,UK
b
LaboratoriodiMicrobiologiaMolecolareeBiotecnologia,DipartimentodiBiologiaMolecolare,Universita` diSiena,Siena,Italy
c
IstitutoZooprofilatticoSperimentaledellaSardegna,Italy
d
DipartimentodiScienzeChirurgiche,ServizioSequenziamentoDNA,Policlinico,Universita` diCagliari,Cagliari,Italy
eDip.diBiotecnologieAgrarie,Universita` diFirenze,Firenze,Italy
fInstituteofIntegrativeBiology,UniversityofLiverpool,LiverpoolL697BE,UK g
UOCBatteriologia,AziendaOspedalieraUniversitariaSenese,Siena,Italy
h
DepartmentofGenetics,UniversityofLeicester,LeicesterLES17RH,UK
ARTICLE INFO
Articlehistory:
Received3September2014
Receivedinrevisedform4November2014 Accepted12November2014 Keywords: Mastitis Pseudomonasaeruginosa Diversity Geneexpression Biofilm Straindivergence ABSTRACT
Bacterialinfectionscausingmastitisinsheepcanresultinsevereeconomiclossesfor farmers.A large survey of milk samples from ewes with mastitis in Sardinia, Italy, indicatedanincreasing prevalenceof Pseudomonasaeruginosa infections.It hasbeen shown previously that during chronic, biofilm-associated infections P. aeruginosa populationsdiversify.We reportthephenotypicandgenomiccharacterisation oftwo clonalP.aeruginosaisolates (PSE305andPSE306)fromamastitis infectionoutbreak, representingdistinctcolony morphologyvariants.In additionto pigmentproduction, PSE305andPSE306differedinphenotypiccharacteristicsincludingbiofilmformation, utilisationofvariouscarbonandnitrogensources,twitchingmotility.Wefoundhigher levelsofexpressionof genesassociated withbiofilmformation (pelB) and twitching motility (flgD) in PSE305, compared to the biofilm and twitching-defective PSE306. Comparativegenomicsanalysis revealedsinglenucleotidepolymorphisms(SNPs)and minor insertion/deletion variations between PSE305 and PSE306, including a SNP mutationinthepilPgeneofPSE306.Byintroducingawild-typepilPgenewewereableto partiallycomplementthedefectivetwitchingmotilityofPSE306.Therewerealsothree largerregionsofdifferencebetweenthetwogenomes,indicatinggenomicinstability. Hence,wehavedemonstratedthatP.aeruginosapopulationdivergencecanoccurduring anoutbreakofmastitis,leadingtosignificantvariationsinphenotypeandgenotype,and resemblingthebehaviourofP.aeruginosaduringchronicbiofilm-associatedinfections.
ß2014ElsevierB.V.Allrightsreserved.
* Correspondingauthorat:InstituteofInfectionandGlobalHealth,UniversityofLiverpoolRonaldRossBuilding,WestDerbyStreet,LiverpoolL697BE, UK.Tel.:+4401517959642.
E-mailaddress:C.Winstanley@liv.ac.uk(C.Winstanley).
** Correspondingauthorat:DepartmentofGenetics,UniversityofLeicester,AdrianBuilding,Room115,UniversityRoad,LeicesterLE17RH,UK. E-mailaddress:mro5@leicester.ac.uk(M.R.Oggioni).
1
Theseauthorscontributedequallytothework.
2
Theseauthorscontributedequallytothework.
ContentslistsavailableatScienceDirect
Veterinary
Microbiology
j our na l h ome p a ge : w ww . e l se v i e r . com / l oc a te / v e t m i c
http://dx.doi.org/10.1016/j.vetmic.2014.11.011
1. Introduction
Mastitis is an inflammation of the mammary gland and is a common condition in many sheep-rearing countries (Conington et al., 2008). The economic losses due to mastitis are of major consequence for sheep farmers. In particular, mastitis can result in poor milk yields,changesinmilkcomposition,prematurecullingof ewesand poor growthratesof lambs(Conington etal., 2008). Mastitis isfrequentlycaused bytheintroduction andmultiplicationofpathogenicbacteriainthemammary glands.CausativebacteriaincludetheCoagulaseNegative Staphylococci,Staphylococcusaureus,Streptococci, Enter-obacteriaceaeandPseudomonasaeruginosa.Alargesurvey of 27,218 milk samples from ewes with mastitis (1214flocksand 1634mastitisevents)inSardinia,Italy indicated that the prevalence of P. aeruginosa cases increasedover the7 yearsamplingperiod(unpublished data).ThemanagementofP.aeruginosainfections repre-sents an increased challenge compared to infections involving other aetiological agents because of the tendencyforlongpersistenceandspread from subclini-cally infected animals throughout a herd. Furthermore with no effective antimicrobials available, intervention strategiesarelimited.
Asaconsequenceofitslargegenome,P.aeruginosaisa versatileandadaptable opportunisticpathogenthatcan causeinfections ofboth medical and veterinary impor-tance(WinstanleyandFothergill,2009;Hertletal.,2011). TheabilityofP.aeruginosatocoloniseandinfectavariety of sites has been attributed to the many virulence factors that it can produce. These include various exotoxins, andseveral secreted products thatare under thecontrolof complex interconnecting quorum-sensing (QS)regulatorynetworks,includingpyocyanin,elastases, LasA, alkaline proteases and rhamnolipid (Winstanley and Fothergill, 2009). Biofilm formation and alginate overproduction,arealsoimportantaspectsofP.aeruginosa pathogenicity,especially in thecontext of chronic lung infections of cystic fibrosis (CF) patients, where they contribute to the intrinsic resistance of P. aeruginosa infections to antimicrobials (Winstanley and Fothergill, 2009).
AlthoughseveralP.aeruginosagenomesequenceshave been reported, these are mostly from human infection isolates(Silbyetal.,2011).Genomesequencingstudieson P. aeruginosa isolates from human CF lung infections haveshownthatmutationandpopulationdivergenceisa commonfeatureofbacteriaduringthesechronic, biofilm-associated,infections,oftenleadingtochangesin pigmen-tationorcolonymorphology(Mowatetal.,2011).Similar diversificationhasbeenobservedinexperimentalbiofilm systems(McElroyetal.,2014).
Having observed colonial morphology variations amongstP.aeruginosaisolatedfromanoutbreakofovine mastitis,theaimofthisstudywastodeterminewhether duringmastitisP.aeruginosapopulationsshowbehaviour characteristicofchronicbiofilm-associatedinfectionsby characterisingthephenotypicandgenotypicpropertiesof two P. aeruginosa isolates representing distinct colony morphologyvariants.
2. Methods 2.1. Bacterialstrains
Theisolateschosenforthisstudywerefromaninfected flockofSardinianewesofabout350animalsnotifiedfor casesofmastitisduringthelactationperiod2004–2005.In a6monthperiod,morethan100animalsbecame culture-positiveforP.aeruginosa.Towardstheendofthelactation period,agroupofsevensheepshowingsignsofmastitis werechosenformoreextensivesampling.Duplicatemilk samplesandswabswerecollectedforeachhalfudder.In fourcases,cultures fromewes showingclinical signsof mastitis yielded a P. aeruginosa with green fluorescent pigmentation(exampleisolatePSE305).Fromtwosheep,a red-pigmentedpyorubin-positive,non-fluorescentP. aer-uginosawasisolated(exampleisolatePSE306).Fromeach sampleallcolonieswereeitherredorgreen-pigmented, and PSE305-like isolates were also obtained from the milkingmachine.
TheIstitutoZooprofilatticoSperimentaleistheofficial Italian public body operating in the frame of National Health Servicewithdutiesrelated toanimal health and welfareandfoodsafety.Animalsamplesweretakenaspart of thenormal surveillance of a regional programme in Sardegnaforreductionofovinemastitis.
2.2. ArrayTubegenotyping
IsolatesPSE305andPSE306werecharacterisedusing the ArrayTube (CLONDIAG, Alere Technologies, Ko¨ln, Germany) genotypingmethod(Wiehlmann etal.,2007). Data from the 13 SNPs, flagellin type (a/b), and the presence of the mutually exclusive type III secretion exotoxins(S orU), wereconvertedinto a ‘‘hexadecimal code’’ representedbyfourdigits,allowingthepublished databasetobesearchedasdescribedpreviously( Wiehl-mannetal.,2007).
2.3. Bacterialculture,RNAextractionandcDNAsynthesis GrowthcurveswereperformedonPAO1,PSE305and PSE306toconfirmthestageatwhichthebacterialcellsare at the mid-exponential and stationary phase. PAO1, PSE305 and PSE306 were inoculated into LB to give a starting A600 of 0.05. The cultures weregrown to
mid-exponential(A600of0.5)andstationaryphase(A600of1.5)
shakingat180rpmandat378C.PSE305andPSE306were alsogrownstaticallyin6-wellplatesinLBat378Cfor24h. RNAextractionswerecollectedfromtheseisolatesat mid-exponential,stationaryphaseandduringconditionsthat promote biofilmgrowth inPAO1 and PSE305 usingthe NucleospinRNAIIkit(Macherey-Nagel).Thepelletswere washedin1MTrispH8(0.1MEDTA)andresuspendedin 100
m
l lysis buffer (Nucleospin RNA II kit; Macherey-Nagel).FollowingRNAextraction,1m
gofRNAwasusedas atemplateforcDNAsynthesis,usingtheTranscriptorFirst Strand cDNA Synthesis kit (Roche Diagnostics). RNA samples thatwere notexposed toreverse transcriptase wereusedascontrols,andthesewereconfirmedashaving negligiblelevelsofgenomicDNApresent.2.4. GeneexpressioninPSE305andPSE306usingSYBR Green1
TherealtimePCRwasperformedinavolumeof20
m
l containing 1 LightCycler DNA Master SYBR Green I mastermix (RocheDiagnostics), 500nM ofeach primer and 2m
l of the appropriate cDNA sample. Each cDNA sample(includingreversetranscriptasenegativecontrols) wasdiluted1:2withsterile,RNaseandDNase-freewater andanalysedinduplicate.Thereactionswereperformedin 96-wellplatesinaLightCycler480realtimePCRmachine (Roche diagnostics). Amplification was carried out for 45cyclesof958Cfor10s,608Cfor5sand728Cfor30s. Fluorescence was recorded at a wavelength between 510nm and 550nm at theend of each PCR cycle. PCR products weresubsequently analysed by melting curve analysisrangingfrom658Cto958C.Foldchangesingene expressionwascalculatedusing the2DDCt(Schmittgenand Livak, 2008). The referencegene was gyrBand the referencecondition wasexponentiallygrown PSE305or PSE306. The fold changes in gene expression during stationary phase and biofilmgrowthwere comparedto PSE305 and PSE306grown toexponentialphase. Genes selected for analysis were pilT, algD, pslD, pelB, lasR, pslA and flgD (for primer details, see Supplementary Table1).Toconfirmsuitabilityofusingthe2DDCt
method toanalysethequantitativerealtimePCRdata,astandard curvewasconstructedforeachtargetandreferencegene (SchmittgenandLivak,2008).SerialdilutedP.aeruginosa genomic DNA was amplified with the corresponding primer sets and the obtained threshold crossing point cycles (CT) were plotted against logarithms of initial genomicDNAamountsutilisedineachreaction.AllPCRs exhibited a strong linear correlation (R2<0.9) between
template amounts and CT values, with PCR efficiencies rangingfrom95%to110%confirmingthesuitabilityofthe 2DDCTmethodfortheanalysisofrelativegeneexpression
(SchmittgenandLivak,2008). Furthermore,all quantita-tive PCRs had single melting points following melting curveanalysis.
2.5. Cellattachmentassay
P.aeruginosaisolatesPSE305andPSE306were inocu-latedintoPM1F-0solution(Biolog)untilthesolutionwas at42%transmittance(Biologturbidimeter).Thissolution inavolumeof15mlwasfurtherdilutedin75mlof PM1F-0solution.Thedilutedbacterialsuspension,inavolumeof 100
m
l,wasaddedtoeachwellofPM1andPM2A96-well plates(Biolog)andincubatedfor24hat378C. Thiswas performed intriplicate.Following incubation,themedia and planktonic cells were removed and the adhered biofilms were washed three times with 150m
l of phosphate-bufferedsaline. Thebiofilmswereallowedto dryand 125m
lof0.5%(w/v)crystalvioletinwaterwas added to each well and stained for 15min at room temperature.Thebiofilmswerewashedthoroughlywith water to ensure excess stain was removed. Following which,125m
lofethanolwasaddedtoeachwelltodissolve thecrystalviolet.Thesolutionwastransferredtoanew microtiterplateandreadatA550nm.2.6. Twitchingmotilityassay
P.aeruginosastrainswerestabbedintoLBagarusinga steriletoothpick,throughtothebottomofthepetridish. Theplateswereincubatedovernight at378C.Following incubation,theagarwasremovedfromthepetridishand the dish was stained with0.5% (w/v) crystal violet for 20minatroomtemperature.Thestainwaswashedoffand the diameter of the stained zones was measured. The twitching motility assay was performed six times. For complementation, the plasmid pDFR1C (Martin et al., 1995), containing pilM, pilN, pilO and pilP, was kindly suppliedbyProfessorJohnMattick,Universityof Queens-land. The plasmid was introduced into P. aeruginosa PSE306 by electroporation, selecting for carbenicillin resistance(400
m
g/ml).2.7. Genomesequencing
DNA was extracted from PSE305 and PSE306 grown planktonically in LB broth overnight at 378C using the Wizard1
Genomic DNA Purification Kit (Promega). The purityoftheextractedbacterialDNAwasassessedusingthe NanodropND-1000andthequantityofbacterialDNAwas measuredusingtheQubit1
2.0Fluorometer.StrainPSE305 wassequencedusingtheRoche454GenomeSequencerFLX (GS-FLX)bypreparinga6kbpaired-endlibraryusingthe standardtitaniumchemistryfor454.Contigswere assem-bled from the reads, and scaffolded using the Roche 454Newblerassembler(version2.5)withdefaultsettings. ThescaffoldswereorderedandorientatedwithrespecttoP. aeruginosaPAO1(AE004091)usingAbacas( http://abacas.-sourceforge.net/)andthealignmentprogrammeMUMmer (Kurtzetal.,2004).Glimmerv3.02wasusedtocallputative ORFs(Delcheretal.,2007).Aputativefunctionwasthen assigned to each gene by BLAST (Altschul et al., 1990) comparisonwithadatabaseofsequencesgeneratedfrom previously annotated P.aeruginosa genomes.In orderto identifyregions ofdifference, the pseudochromosome of strain PSE305was aligned to the genome of PAO1 and analysedusingtheARTEMISComparisonTool(Carveretal., 2005). In addition, the genome of strain PSE306 was sequencedusingtheSOLiDv4.0systeminordertoidentify single nucleotide polymorphism (SNP) mutations and insertions/deletions(INDELs)withinthisgenomein com-parisontoPSE305.TheSOLiD readsweremapped using BWA(LiandDurbin,2009)incolorspaceusingsequence quality.TheresultingBAMwasthenprocessedusingGATK (McKennaetal.,2010)basequalityscorerecalibration,indel realignment, and duplicate removal. SNP and INDEL discovery was performed using standard hard filtering parameters (DePristo et al., 2011). The VCF output was filteredusingVCFTOOLS(Daneceketal.,2011)andviewed inARTEMIS(Carveretal.,2005).Genomesequencingwas performedbytheCentreforGenomicsResearch,University ofLiverpool.
2.8. Multi-locussequencetyping(MLST)
ForMLSTtyping,regionsoftheacsA,aroE,guaA,mutL, nuoD, pspAand trpEgenes were extracted directly from
genomesequencedata.Sequencetypeswereassignedusing theavailabledatabase(http://pubmlst.org/paeruginosa/). 2.9. BIOLOGphenotypicassays
Isolates PSE305 and PSE306, and the control P. aeruginosa strainsPAO1 and PA14 were subjected to a phenotypicmicroarray(PM)usingBIOLOGPM1andPM2 microtiterplates,whichtogethertesttheabilitytoutilise 190differentcarbonsources,andPM3,whichtestsforthe ability to utilise 95 different nitrogen sources. The P. aeruginosa were prepared as described previously (Viti etal.,2009).DatawerereadbytheOmniLogPMSoftware andat-testwasappliedtothreeindependentreplicatesto determinestatisticalsignificance.
2.10. Genomesequenceaccession
ThegenomesequencedataforstrainPSE305hasbeen deposited in the ENA-EMBL database under accession numberHG974234.
3. Results
3.1. Genotypingofmastitis-associatedisolatesrepresenting thetwocolonymorphologyvariants
TheArrayTubegenotypingmethodandMLST(derived fromgenomesequencedata) wereusedtoconfirmthat PSE305(green-pigmented isolate)and PSE306 (the red-pigmented isolate) were of the same clone type. The ArrayTubegenotypingmethod(Wiehlmannetal.,2007), whichgivesagenotypicprofilethatcanbeusedtosearcha largedatabaseofP.aeruginosastrains,showedthatPSE305 andPSE306haveanidenticalprofile,whichtheyalsoshare withP.aeruginosacloneV(hexadecimalcode:0812).Clone Visnotoneofthemajorclonesinthedatabase,butithas beenfoundamongstthecollectiontested,mostly associ-atedwithisolatesfromhumanbacteraemiacases( Wiehl-mannet al.,2007). Previousanalysesof clonal complex structuresbasedonthegenotypingdataidentifycloneVas an outlier rather than a member of a major cluster (Wiehlmannetal.,2007).Inadditiontheanalysisofthe PSE305 genomesequence identified thestrain as MLST typeST-244.
3.2. PhenotypicvariationbetweenisolatesPSE305and PSE306
Inadditiontopigmentproduction,therewereanumber ofphenotypicdifferencesbetweentheP.aeruginosaisolates PSE305andPSE306.PSE306didnotformstablebiofilmsin thepresenceofvarioussubstrates,unlikePSE305(Fig.1). Therewerestatisticallysignificantdifferencesinthelevels of crystal violet staining between PSE305 and PSE306, following growth in conditions that promote biofilm formationandinthepresenceofvarioussubstrates (two-tailedStudentst-test,noadditional substrate:t[4]=3.13, p<0.05; with
a
-D-glucose: t[4]=10.68, p<0.001; with butyric acid: t[4]=39.62, p<0.001; with caproic acid: t[4]=7.051, p<0.01; with capric acid: t[4]=51.28,p<0.001).PSE305growninthepresenceof
a
-D-glucose,butyric acid, caproic acid and capric acid significantly increasedbiofilmformationinthisstrain,whencompared togrowth withoutadditional substrates(two-tailed Stu-dentst-test,with
a
-D-glucose:t[4]=8.24,p<0.001;with butyric acid:t[4]=18.87, p<0.001;with caproicacid: t[4]=6.53, p<0.01; with capric acid: t[4]=30.00, p<0.001).PSE305andPSE306differedintheirutilisationofcertain nitrogenandcarbonsources(SupplementaryTable2).In particular,PSE305exhibitedmoremetabolicactivityinthe presenceof
a
-D-glucose,butyricacid,caproicacid,capricacidandL-histamine,thanPSE306grownunderthesame conditions. This was true of other nitrogen and carbon sources that were assayed (data not shown). However, interestinglybothmastitisstrainsdidnotutiliselactosetoa significantextent(SupplementaryTable2).
OnewayANOVA(Bonferroniadjusted)indicatedthat PSE306wasdeficientin twitchingmotilitycomparedto PSE305(Fig.2)(p<0.001).
3.3. ComparativegenomicsofisolatesPSE305andPSE306 We obtainedthegenomesequenceofisolatePSE305 using454pyrosequencingandthepseudochromosomeof strainPSE305wasalignedtothegenomeofstrainPAO1. InsertionsanddeletionsinthegenomeofstrainPSE305in comparisontothegenomeofstrainPAO1aresummarised in Supplementary Tables 3 and 4. Many of these were locatedinpreviouslyreportedregionsofgenomeplasticity (Mathee et al., 2008). Most of the larger insertions corresponded to known genomic islands or prophages. The largest additional region in the genome of strain PSE305wasa115kbgenomicisland.
InadditionwesequencedthegenomeofstrainPSE306 using the SOLiD v4.0 system in order to identify SNP mutationsandINDELswithinthisgenomeincomparison toPSE305.Threelargegeneclustersidentifiedaspresentin thegenomeofstrainPSE305butabsentfromthegenome
Fig.1.PSE305formsstablebiofilms,unlikePSE306.PSE305andPSE306 werestaticallygrowninthepresenceofvarioussubstrates,andthe resultantbiofilmswerestainedwithcrystalviolet(CV)toassessthe extentofbiofilmformationinthesecultures.1.Noextrasubstrate,2.a-D -glucose,3.D9a-D-lactose,4.butyricacid,5.caproicacid,6.capricacid(all groups:n=3).TheblackandwhitebarsrepresentPSE305andPSE306, respectively.Errorbars representthestandard errorofthemean.(* p<0.05,**p<0.01and***p<0.001).
ofstrainPSE306wereidentified.Twoofthese(insertionsE andNinSupplementaryTable3)containedseveralORFs sharingsimilaritywithphageproteins,andareputative prophages. A third large deletion (approximately 91kb; position2070401to2161555)wasofaclusterofgenes correspondingtothePAO1genesPA1981-2059.Thislatter regionhasnotpreviouslybeenreportedasdivergent.PCR assayshavebeendesignedtothesethreedeletedregions. In allcases,strain PSE305wasPCR-positive andPSE306 wasPCR-negative,confirmingthegenomesequencedata (datanotshown).
SNPandminorINDELvariationsinthegenomeofstrain PSE305whencomparedtothegenomeofstrainPSE306are summarisedinTable1.OnlysixSNPspassingthequality threshold were found in the genome of PSE306. These includedSNPsinanumberofORFswithputativefunctions thatmightimpactonimportantphenotypessuchasasensor histidine kinase, siderophore receptor, type IV pilus biogenesis protein,alginateregulatory protein, permease andphenazinebiosynthesisprotein.Onlyasmallnumberof minorINDELsweredetected,oneofwhichwasinanORF withaputativeroleinhemeutilisationoradhesion(Table1). 3.4. GeneexpressionvariationsbetweenPSE305andPSE306 BecausePSE305andPSE306exhibiteddifferent pheno-typic propertieswith regardsto biofilm-forming ability andtwitchingmotility,weusedgeneexpressionanalysis toascertainwhetherdifferencesingeneexpressioncould have contributed to these observed phenotypic differ-ences. Therefore, the expression of genes involved in biofilm formationandtwitching motilityinPSE305 and PSE306werequantitativelyassessed(Fig.3).Specifically, theexpressionprofileindifferentinvitromodelsofgrowth (planktonic and biofilm mode of growth) were charac-terisedtoidentifywhethertherearedifferencesbetween
thePSE305 (which can undergo twitching motilityand biofilm formation) and PSE306 (which cannot). The expression profiles of the genes pilT and flgD both of which are involved in type IV fimbrial biogenesis and motility (Bertrand et al., 2010) was characterised. In addition, pslD, pslA, pelB and algD expression profiles, whichareinvolvedinbiofilmarchitectureandformation (Campisano et al., 2006; Ghafoor et al., 2011)and lasR which is involved in the regulation of QS network (WinstanleyandFothergill,2009)weredetermined.
TheexpressionofflgDwassignificantlylowerinPSE306 at stationary phase and under conditions that promote biofilmformation,comparedtoexpressioninPSE305 (two-tailed Students t test; stationary phase: t(6)=2.54, p=0.04.Biofilmgrowth:t(6)=2.87, p=0.03).In addition, theexpressionofpelBwasalsosignificantlylowerinPSE306 at stationary phase and under conditions that promote biofilmformation (two-tailed Studentst test; stationary phase: t(6)=2.72, p=0.03. Biofilm growth: t(6)=2.99, p=0.03).Thisdifferencebetweenthetwoisolateswasalso trueofalgD(two-tailedStudentsttest;stationaryphase: t(6)=6.86,p=0.0005.Biofilmgrowth:t(6)=3.96,p=0.007). Therewasnostatisticalsignificantdifferencebetweenthe geneexpressionprofilesofpilT,pslD,pslAandlasRbetween PSE306and PSE305grownin eitherstationaryphase or underconditions that promote biofilmformation. There was,however,asignificantincreaseinpslDexpressionin PSE306between stationary phase andduring growthin conditions that promote biofilm formation (two-tailed Studentsttest;t(6)=3.57,p=0.01).ThisdifferenceinpslD expressionwasnotobservedinPSE305.
3.5. Partialcomplementationoftwitchingmotility
HavingidentifiedamutationinthepilPgene(Table1) in the twitching motility-defective strain PSE306, we
Fig.2.Comparisonoftwitchingmotilities.TwitchingmotilitywascomparedbetweenstrainsPAO1(controlstrain),PSE305,PSE306andPSE306containing theplasmidpDFRIC.Errorbarsrepresentthestandarderrorofthemean.pvaluesone-wayanalysisofvariance(Bonferroniadjustment)comparisonsare shown(n=6);ns:notsignificant.
introducedplasmidpDFR1C,containinganintactpilP,into the strain. The presence of the plasmid led to partial complementationofthelossoftwitchingmotility(Fig.2). 4. Discussion
Littleis known about thebehaviour ofP. aeruginosa during mastitis infections in sheep. Having observed variations in colony morphology amongst isolates, we hypothesised that during such infections P. aeruginosa showbehaviourcharacteristicofchronic, biofilm-associ-atedinfections.In thisstudy,wefocusedon representa-tivesofthetwomajorcolonymorphologytypesidentified during the mastitis outbreak and have demonstrated considerablevariationsbetween them. Thetwo isolates (PSE305and PSE306)fromthe same outbreak of ovine mastitisina flock ofSardinian eweswereconfirmedas beingofthesameclone-type,bothgenotype0812(clone V),whichhasbeenfoundinanumberofdifferentclinical sourcesandamongstisolates fromwater (Crameretal., 2012). Previous analyses of clonal complex structures
basedonthegenotypingdataidentifycloneVasanoutlier ratherthanamemberofamajorcluster(Wiehlmannetal., 2007).MLSTanalysisshowedthatPSE305belongedtothe subtype, ST-244. This subtype has been identified in a collectionofhumanwoundisolatesfromthe Mediterra-nean,andincasesofbacteraemiainAsia(Maatallahetal., 2011; Wooetal.,2011). Hence,it islikely thatthetwo isolatescomparedinthisstudyarederivedfromthesame outbreak. Despitebeingof thesameclone-type,PSE305 and PSE306 showed evidence of both phenotype and genotype variation. PSE305 and PSE306 differed in pigmentproduction,twitchingmotilityandintheirability toformstablebiofilms,suggestingthatstrain diversifica-tionoccurredwithinthisoutbreakofovinemastitis.
Some of these phenotypic differences could be explained bythe SNPmutationspresentin PSE306 and theothergenomicdifferencesbetweenthesetwoisolates. Inparticular,theSNPinpilPofPSE306couldexplainthe deficiencyintwitchingmotilityseeninthisisolate.PilPis a lipoprotein that localises to the inner membrane of P.aeruginosaandpromotestheassemblyandfunctionof
Table1
Non-synonymousSNPs/INDELsinPSE306relativetothechromosomeofPSE305ThepositionreferstothelocationoftheSNPandINDELinthePSE305 pseudochromosome.Comp.indicatesthatthegeneisonthecomplementarystrandandonlySNPswithaminimumqualityscoreof100areshown.
Position Description Quality PSE305CDS PAO1ORF Comments
1051625 SNPT!C 246.81 PSE305_10100 PA1098 FlagellarsensorhistidinekinaseFleS(402AAin length); changeofVal57!Ala57;databaseP.
aeruginosasequencesareallAlainthisposition 3880538 SNPA!G(comp.T!C) 128.06 PSE305_36890c PA3408 Hemophore HasAouter membrane receptor HasR; iron siderophore receptor protein (883AAinlength);changeofSer860!Pro860;
databaseP.aeruginosasequencesareallProin thisposition
5237941 SNPT!G(comp.A!C) 157.63 PSE305_49740c PA5041 TypeIVpilusbiogenesisproteinPilP(174AAin length);changeofThr107!Pro107;databaseP.
aeruginosasequencesareallThrinthisposition 5521552 SNPG!A(comp.C!T) 253.13 PSE305_52460c PA5261 AlginatebiosynthesisregulatoryproteinAlgR (248AAinlength);changeofPro16!Leu16;
databaseP.aeruginosasequencesareallLeuin thisposition
6105549 SNPG!A 241.13 PSE305_58030 PA0220 Amino acid permease (477 AA in length); changeofGly33!Asp33;databaseP.aeruginosa
sequencesareallGlyinthisposition 6752225 SNPT!G(comp.A!C) 122.6 PSE305_64430c PA1905and
PA4216
Phenazinebiosynthesis proteinPhzG1/PhzG2 (95 AA in length); changeof Lys24!Thr24;
databaseP.aeruginosasequencesareallArgin this position. Poordatabase matchesin the N-terminalfirst24residues
2300849 INDELGC!G 16/16 PSE305_22350 N/A Possible RBS region mutation upstream of ORF21710c(putativeNADH-flavinreductase); matchesPAO1ORFPA2176
3166266 INDELCT!C 11/11 – N/A BetweenORFs28800and28810
3297412 INDELG!GC(comp.C!CG) 11/11 PSE305_31300c PA2885 Acyclic terpenes utilisation regulator, AtuR, TetRfamily
4564584 INDELCAGGT!C Poor PSE305_43480 PA3980 (Dimethylallyl)adenosine tRNA methylthio-transferase;frameshiftmutationfromposition 125ina446AAprotein
4981298 INDELA!AC 7/7 – N/A BetweenORFs46340and46350
5153865 INDELGC!G(comp.CG!C) Poor PSE305_49100c PA4981 Aminoacidtransporter,AATfamily;frameshift mutationfromposition327ina470AAprotein 5907354 INDELC!CG 8/8 PSE305_56120 PA0041 Largeexoproteinsinvolvedinhemeutilisation
oradhesion
6245439 INDELTC!T 10/10 PSE305_59200 PA0299 Omega-aminoacid-pyruvateaminotransferase 6733229 INDELCA!C Poor – N/A BetweenORF62780and62790
Fig.3.RelativelevelsofgeneexpressioninPSE305andPSE306.RelativeexpressionlevelsofbiofilmandmotilitygenesbetweenPSE305andPSE306grown tostationaryphaseandunderconditionsthatpromotebiofilmformation,comparedtoexponentialphase(n=3–4).ThereferencegenewasgyrB.Blackand whitebarsrepresentPSE305andPSE306,respectively.Errorbarsrepresentthestandarderrorofthemean.(*p<0.05,**p<0.01and***p<0.001).
thetypeIVpilus(Ayersetal.,2009).Studieshaveshown thatdisruptionofthepilPgene,resultsinalossofsurface typeIVpiliandadeficiencyintwitchingmotility(Ayers et al., 2009). We demonstrated partial restoration of twitchingmotility followingtheintroductionof a wild-typepilP on aplasmid,implicatingthis mutationin the defectivetwitchingmotilityofPSE306.
Furthermore,therewerealsodifferencesinthe utilisa-tionofhistamineasanitrogensourcebetweenPSE305and PSE306. PSE305 utilised histamine to a greater extent comparedtoPSE306.Thiscouldhavebeenattributedtothe presenceof a SNPmutation in PSE306,in a regionwith homology tothe open reading frame, PA0220in PAO1. PA0220codes foran aminoacidpermease,which when knockedoutinPAO1causedadefectinhistamineutilisation (Johnson et al., 2008). Biolog analysis also showed the absenceoflactoseutilisationbybothPSE305andPSE306. Althoughlactoseisamajorcomponentofovinemilk(Jandal, 1996),theabsenceoflactoseutilisationwasnotsurprising as Pseudomonas sp. does not typically ferment lactose. Glucose appears to be the dominant carbon source of PSE305,unlikeinPSE306.Interestingly,capricacid utilisa-tionbyPSE306appearedtobegreaterthanbyPSE305.This fatty acid is also a significant component of ovine milk (BreckenridgeandKuksis,1967)andcouldbeasourceof carbonforPSE306.Studieshaveshownthatthefattyacid profileofovinemilkcanbemanipulatedbyusingdifferent feedingregimens(Addisetal.,2005).Itispossiblethatby changingthedietoftheovineherdinordertoreducethe fatty acid components of the milk, the P. aeruginosa populationsthatutilisethesecomponentsforgrowthcould becontrolled.Incontrast,capricacidhasbeenshownto haveantimicrobialpropertiesagainstotherGram-negative bacteria,includingCampylobacterjejuni,Salmonellaspp.and Escherichiacoli(Thormaretal.,2006).
ThethreemajordeletionsfromthegenomeofPSE306 compared to PSE305, and the insertions/deletions in PSE305comparedtoPAO1,suggestconsiderablygenomic instabilityduringtheoutbreak.GenomicinstabilityinP. aeruginosaCFinfectionshasbeenwidelyreported(Cramer etal.,2011).Howeverthisisthefirstevidenceofgenomic instabilityduringacuteP.aeruginosamastitisinfections.
BesidesthegenomicdifferencesbetweenPSE305and PSE306,therewerealsosignificantdifferencesinthegene expressionprofilesbetweenthesetwoisolates.Thesegene expression differences could also explain the observed phenotypes of PSE305 and PSE306, in particular the deficiencyintwitchingmotilityandtheabsenceofstable biofilm formation in PSE306. There was a significant differencebetweentheexpressionofpelBbetweenPSE305 andPSE306underthetwogrowthconditionstested.The pelgenesareinvolvedintheproductionofaglucose-rich polysaccharide that contributes to the formation of a matrix that surrounds the cells in a biofilm; PelB in particular has been predicted to be a transmembrane proteinlocatedwithinthecytoplasmicmembrane( Vas-seuretal.,2005).P.aeruginosawithmutationsinpelBare deficient in their ability to form biofilms, in particular duringearlystages of biofilmformation (Vasseur etal., 2005).ThissuggeststhatthereducedexpressionofpelBin PSE306comparedtoPSE306,couldhavecontributedtothe
deficiencyofPSE306toformstablebiofilms.AswithpelB, algD gene expression was also reduced in PSE306 comparedtoPSE305,underthegrowthconditionstested. ThealgDgeneispartofanoperonthatisresponsiblefor the synthesis of alginate and contributes to biofilm architecture(Stapperetal.,2004). ExpressionoftheflgD gene,involvedinflagellaassembly,wasreducedinPSE306. It has been shown that flagella are necessary for P. aeruginosabiofilmdevelopmentandmotility(O’Tooleand Kolter,1998).Therefore,thereducedexpressionofalgD, flgDandpelBcouldcontributetotheinabilityofPSE306to formstablebiofilms.Furthermore,reducedflgDexpression in PSE306 could have resulted in reduced flagella-dependant motility in this isolate. Interestingly, pilP transcription was not significantly reduced in PSE306 compared to PSE305 despite PSE306 containing a SNP mutationinthisgene.ThereforeitispossiblethattheSNP mutationaffectsthefunctionofPilP,butnotthelevelof pilP transcription. Overall,somekey differencesin gene expressionbetween PSE305and PSE306wereidentified thatcouldhavecontributedtothephenotypicdifferences betweenthesetwoisolates.
5. Conclusion
PSE306andPSE305representisolatesofthesameclone thatvaryinbothgenotypeandphenotype.Thisprovides evidence that strain divergence occurred during this infection outbreak. P. aeruginosa population divergence has previouslynot beenreported in mastitis infections. Thefurther understandingofpopulation behaviourofP. aeruginosaduringmastitisinfections,inparticular regard-ingbiofilmformingability,mayaidthedevelopmentof noveltherapeuticstrategiestotreatmastitisinfections. Acknowledgments
TheworkwassupportedinpartbyagrantoftheMinistro della Saluteprogramma diricercacorrente IZS SA 01/04 and Progetti di ricercafondamentale o di base; Regione Sardegna.Anno2009CRP1_188.Thegenomesequencingand analysis work was supportedby a grantfrom the Royal CollegeofVeterinarySurgeonsGoldenJubileeTrustFund, UK.
AppendixA. Supplementarydata
Supplementarydataassociated withthisarticlecanbe found, in the online version, at http://dx.doi.org/10.1016/ j.vetmic.2014.11.011.
References
Addis,M.,Cabiddu,A.,Pinna,G.,Decandia,M.,Piredda,G.,Pirisi,A.,Molle, G.,2005.Milkandcheesefattyacidcompositioninsheepfed Medi-terraneanforages withreference toconjugated linoleicacid cis-9,trans-11.J.DairySci.88,3443–3454.
Altschul,S.F.,Gish,W.,Miller,W.,Myers,E.W.,Lipman,D.J.,1990.Basic localalignmentsearchtool.J.Mol.Biol.215,403–410.
Ayers,M.,Sampaleanu,L.M.,Tammam,S.,Koo,J.,Harvey,H.,Howell,P.L., Burrows,L.L.,2009.PilM/N/O/Pproteinsformaninnermembrane complexthataffectsthestabilityofthePseudomonasaeruginosatype IVpilussecretin.J.Mol.Biol.394,128–142.
Bertrand,J.J.,West,J.T.,Engel,J.N.,2010.Geneticanalysisoftheregulation oftypeIVpilusfunctionbytheChpchemosensorysystemof Pseudo-monasaeruginosa.J.Bacteriol.192,994–1010.
Breckenridge,W.C.,Kuksis,A.,1967.Molecularweightdistributionsof milkfattriglyceridesfromsevenspecies.J.LipidRes.8,473–478.
Campisano,A.,Schroeder,C.,Schemionek,M.,Overhage,J.,Rehm,B.H., 2006.PslDisasecretedproteinrequiredforbiofilmformationby Pseudomonasaeruginosa.Appl.Environ.Microbiol.72,3066–3068.
Carver,T.J.,Rutherford,K.M.,Berriman,M.,Rajandream,M.A.,Barrell,B.G., Parkhill,J.,2005.ACT:theArtemisComparisonTool.Bioinformatics 21,3422–3423.
Conington,J.,Cao,G.,Stott,A.,Bunger,L.,2008.Breedingforresistanceto mastitisinUnitedKingdomsheep,areviewandeconomicappraisal. Vet.Rec.162,369–376.
Cramer,N.,Klockgether,J.,Wrasman,K.,Schmidt,M.,Davenport,C.F., Tummler,B.,2011.Microevolutionofthemajorcommon Pseudomo-nasaeruginosaclonesCandPA14incysticfibrosislungs.Environ. Microbiol.13,1690–1704.
Cramer,N.,Wiehlmann,L.,Ciofu,O.,Tamm,S.,Hoiby,N.,Tummler,B., 2012. MolecularepidemiologyofchronicPseudomonasaeruginosa airwayinfectionsincysticfibrosis.PLoSONE7,e50731.
Danecek,P.,Auton,A.,Abecasis,G.,Albers,C.A.,Banks,E.,DePristo,M.A., Handsaker, R.E.,Lunter, G.,Marth,G.T.,Sherry,S.T., McVean,G., Durbin,R.,GenomesProjectAnalysisGroup,2011.Thevariantcall formatandVCFtools.Bioinformatics27,2156–2158.
Delcher,A.L.,Bratke,K.A.,Powers,E.C.,Salzberg,S.L.,2007.Identifying bacterialgenesandendosymbiontDNAwithGlimmer. Bioinformat-ics23,673–679.
DePristo,M.A.,Banks,E.,Poplin,R.,Garimella,K.V.,Maguire,J.R.,Hartl,C., Philippakis,A.A.,delAngel,G.,Rivas,M.A.,Hanna,M.,McKenna,A., Fennell,T.J.,Kernytsky,A.M.,Sivachenko,A.Y.,Cibulskis,K.,Gabriel, S.B.,Altshuler,D.,Daly,M.J.,2011.Aframeworkforvariation discov-eryandgenotypingusingnext-generationDNAsequencingdata.Nat. Genet.43,491–498.
Ghafoor,A.,Hay,I.D.,Rehm,B.H.,2011.Roleofexopolysaccharidesin Pseudomonasaeruginosabiofilmformationandarchitecture.Appl. Environ.Microbiol.77,5238–5246.
Hertl,J.A.,Schukken,Y.H.,Bar,D.,Bennett,G.J.,Gonzalez,R.N.,Rauch,B.J., Welcome,F.L.,Tauer,L.W.,Grohn,Y.T.,2011.Theeffectofrecurrent episodesof clinicalmastitiscaused by gram-positive and gram-negativebacteriaandotherorganismsonmortalityandcullingin Holsteindairycows.J.DairySci.94,4863–4877.
Jandal,J.M.,1996.Comparativeaspectsofgoatandsheepmilk.Small Rumin.Res.22,177–185.
Johnson,D.A.,Tetu,S.G.,Phillippy,K.,Chen,J.,Ren,Q.,Paulsen,I.T.,2008.
High-throughputphenotypiccharacterizationofPseudomonas aeru-ginosamembranetransportgenes.PLoSGenet.4,e1000211.
Kurtz,S.,Phillippy,A.,Delcher,A.L.,Smoot,M.,Shumway,M.,Antonescu, C.,Salzberg,S.L.,2004.Versatileandopensoftwareforcomparing largegenomes.GenomeBiol.5,R12.
Li,H.,Durbin,R.,2009.Fastandaccurateshortreadalignmentwith Burrows-Wheelertransform.Bioinformatics25,1754–1760.
Maatallah,M.,Cheriaa,J.,Backhrouf,A.,Iversen,A.,Grundmann,H.,Do,T., Lanotte,P.,Mastouri,M.,Elghmati,M.S.,Rojo,F.,Mejdi,S.,Giske,C.G., 2011. Populationstructure of Pseudomonasaeruginosa from five Mediterraneancountries:evidenceforfrequentrecombinationand epidemicoccurrenceofCC235.PLoSONE6,e25617.
Martin,P.R.,Watson,A.A.,McCaul,T.F.,Mattick,J.S.,1995. Characteriza-tionofafive-geneclusterrequiredforthebiogenesisoftype4 fim-briaeinPseudomonasaeruginosa.Mol.Microbiol.16,497–508.
Mathee,K.,Narasimhan,G.,Valdes,C.,Qiu,X.,Matewish,J.M.,Koehrsen, M.,Rokas,A.,Yandava,C.N.,Engels,R.,Zeng,E.,Olavarietta,R.,Doud, M.,Smith,R.S.,Montgomery,P.,White,J.R.,Godfrey,P.A.,Kodira,C., Birren,B.,Galagan, J.E.,Lory,S.,2008. DynamicsofPseudomonas aeruginosagenomeevolution.Proc. Natl.Acad.Sci. U.S.A.105, 3100–3105.
McElroy,K.E.,Hui,J.G.,Woo,J.K.,Luk,A.W.,Webb,J.S.,Kjelleberg,S.,Rice, S.A.,Thomas,T.,2014.Strain-specificparallelevolutiondrives short-termdiversificationduringPseudomonasaeruginosabiofilm forma-tion.Proc.Natl.Acad.Sci.U.S.A.111,E1419–E1427.
McKenna,A.,Hanna,M.,Banks,E.,Sivachenko,A.,Cibulskis,K.,Kernytsky, A.,Garimella,K.,Altshuler,D.,Gabriel,S.,Daly,M.,DePristo,M.A., 2010.TheGenomeAnalysisToolkit:aMapReduceframeworkfor analyzingnext-generationDNAsequencingdata.GenomeRes.20, 1297–1303.
Mowat,E.,Paterson,S.,Fothergill,J.L.,Wright,E.A.,Ledson,M.J.,Walshaw, M.J.,Brockhurst,M.A.,Winstanley,C.,2011.Pseudomonasaeruginosa populationdiversityandturnoverincysticfibrosischronicinfections. Am.J.Respir.Crit.CareMed.183,1674–1679.
O’Toole,G.A.,Kolter,R.,1998.Flagellarandtwitchingmotilityare neces-saryforPseudomonasaeruginosabiofilmdevelopment.Mol. Micro-biol.30,295–304.
Schmittgen,T.D.,Livak,K.J.,2008.Analyzingreal-timePCRdatabythe comparativeC(T)method.Nat.Protoc.3,1101–1108.
Silby,M.W.,Winstanley,C.,Godfrey,S.A.,Levy,S.B.,Jackson,R.W.,2011.
Pseudomonasgenomes:diverseandadaptable.FEMSMicrobiol.Rev. 35,652–680.
Stapper,A.P.,Narasimhan,G.,Ohman,D.E.,Barakat,J.,Hentzer,M.,Molin, S.,Kharazmi, A.,Hoiby,N.,Mathee,K.,2004.Alginateproduction affectsPseudomonasaeruginosabiofilmdevelopmentand architec-ture,butisnotessentialforbiofilmformation.J.Med.Microbiol.53, 679–690.
Thormar,H., Hilmarsson,H., Bergsson, G.,2006.Stable concentrated emulsionsofthe1-monoglycerideofcapricacid(monocaprin)with microbicidalactivitiesagainstthefood-bornebacteriaCampylobacter jejuni,Salmonellaspp.,andEscherichiacoli.Appl.Environ.Microbiol. 72,522–526.
Vasseur,P.,Vallet-Gely,I.,Soscia,C.,Genin,S.,Filloux,A.,2005.Thepel genesofthePseudomonasaeruginosaPAKstrainareinvolvedatearly andlatestagesofbiofilmformation.Microbiology151,985–997.
Viti,C.,Decorosi,F.,Mini,A.,Tatti,E.,Giovannetti,L.,2009.Involvementof theoscA gene inthe sulphur starvation responseand in Cr(VI) resistanceinPseudomonascorrugata28.Microbiology155,95–105.
Wiehlmann,L.,Wagner,G.,Cramer,N.,Siebert,B.,Gudowius,P.,Morales, G.,Kohler,T.,vanDelden,C.,Weinel,C.,Slickers,P.,Tummler,B.,2007.
PopulationstructureofPseudomonasaeruginosa.Proc.Natl.Acad.Sci. U.S.A.104,8101–8106.
Winstanley,C.,Fothergill,J.L.,2009.Theroleofquorumsensinginchronic cysticfibrosisPseudomonasaeruginosainfections.FEMSMicrobiol. Lett.290,1–9.
Woo,P.C.,Tsang,A.K.,Wong,A.Y.,Chen,H.,Chu,J.,Lau,S.K.,Yuen,K.Y., 2011.Analysisofmultilocussequencetypingschemesfor35different bacteriarevealedthatgenelociof10bacteriacouldbereplacedto improvecost-effectiveness.Diagn.Microbiol.Infect.Dis.70,316–323.