metabonomicsbased analysis of brachyspira pilosicoli's...
TRANSCRIPT
Metabonomicsbased analysis of Brachyspira pilosicoli's response to tiamulin reveals metabolic activity despite significant growth inhibition Article
Accepted Version
Creative Commons: AttributionNoncommercialNo Derivative Works 4.0
Le Roy, C. I., Passey, J. L., Woodward, M. J., La Ragione, R. M. and Claus, S. P. (2017) Metabonomicsbased analysis of Brachyspira pilosicoli's response to tiamulin reveals metabolic activity despite significant growth inhibition. Anaerobe, 45. pp. 7177. ISSN 10759964 doi: https://doi.org/10.1016/j.anaerobe.2017.03.018 Available at http://centaur.reading.ac.uk/70003/
It is advisable to refer to the publisher’s version if you intend to cite from the work. See Guidance on citing .Published version at: http://www.sciencedirect.com/science/article/pii/S1075996417300677
To link to this article DOI: http://dx.doi.org/10.1016/j.anaerobe.2017.03.018
Publisher: Elsevier
All outputs in CentAUR are protected by Intellectual Property Rights law, including copyright law. Copyright and IPR is retained by the creators or other copyright holders. Terms and conditions for use of this material are defined in the End User Agreement .
www.reading.ac.uk/centaur
CentAUR
Central Archive at the University of Reading
Reading’s research outputs online
Metabonomics-basedanalysisofBrachyspirapilosicoli’sresponseto1
tiamulinrevealsmetabolicactivitydespitesignificantgrowth2
inhibition.3
Caroline Ivanne Le Roya,b, Jade Louise Passeyc, Martin John Woodwarda, Roberto4
MarcelloLaRagionec,SandrinePauleClausa#.5
6
a Department of Food and Nutritional Sciences, University of Reading, Whiteknights,7
Reading,UK66AP,UK8
PresentaddressbDepartmentofTwinResearch&GeneticEpidemiology,King’sCollege9
London,LondonSE17EH,UK10
c FacultyofHealth andMedical Sciences, SchoolofVeterinaryMedicine,Universityof11
Surrey,Guilford,SurreyGU27XH,UK12
#Correspondingauthor:[email protected],Tel+44118378871713
14
Keywords:Metabonomics,Brachyspirapilosicoli,Tiamulin,Antibiotics.15
16
Abstract17
Pathogenic anaerobes Brachyspira spp. are responsible for an increasing number of18
IntestinalSpirochaetosis(IS)casesinlivestockagainstwhichfewapprovedtreatments19
areavailable.TiamulinisusedtotreatswinedysenterycausedbyBrachyspiraspp.and20
recentlyhasbeenusedtohandleavianintestinalspirochaetosis(AIS).Thetherapeutic21
doseusedinchickensrequiresfurtherevaluationsincecasesofbacterialresistanceto22
tiamulin have been reported. In this study, we evaluated the impact of tiamulin at23
varying concentrations on the metabolism of B. pilosicoli using a 1H-NMR-based24
metabonomics approach allowing the capture of the overall bacterial metabolic25
response to antibiotic treatment. Based on growth curve studies, tiamulin impacted26
bacterialgrowthevenatvery lowconcentration(0.008µg/ml)although itsmetabolic27
activitywasbarelyaffected72hpostexposuretoantibiotictreatment.Onlythehighest28
doseof tiamulin tested(0.250µg/ml)causedamajormetabolicshift.Resultsshowed29
that below this concentration, bacteria couldmaintain a normalmetabolic trajectory30
despitesignificantgrowthinhibitionbytheantibiotic,whichmaycontributetodisease31
reemergencepostantibiotictreatment.Indeed,weconfirmedthatB.pilosicoliremained32
viableevenafterexpositiontothehighestantibioticdose.Thispaperstressestheneed33
to ensure new evaluation of bacterial viability post bacteriostatic exposure such as34
tiamulin to guarantee treatment efficacy and decrease antibiotic resistance35
development.36
37
Highlight38
• B.pilosicolimetabolismwascharacterizedusing1HNMR-basedmetabonomics39
• Tiamulin inhibited B.pilosicoli growth at very low dose (respectively < 0.01640
µg/mLand>0.032µg/mL)41
• B.pilosicoli metabolism is not inhibited for tiamulin concentration superior to42
0.032µg/mL43
• B. pilosicolimetabolism is completely repressed at 0.250 µg/mL, but remain44
viable45
1.Introduction46
Brachyspira pilosicoli is a gram-negative bacterium of the Spirochaetes family. It47
colonizesthelowerpartofthegastrointestinaltrackofalargerangeofhostsincluding48
pigs,birds,humans,monkeys,dogsandhorses[1-4].Onceintheintestinal lumen,the49
bacteriumisattractedviachemotaxistothemucinbarrier[5,6]throughwhichitswims50
mediated by its unique “corkscrew” shape and rotation of its periplasmic flagella [7]51
aidedby thesecretionofmucinedegradingenzymes [5,8].B.pilosicoli attaches to the52
enterocytesinanendonfashionandmayalsoinfectthesecells[9-12].Colonizationby53
B.pilosicoli can lead to the development of intestinal spirochaetosis (IS), the signs of54
which are diarrhea, poor overall condition, dehydration and decreased growth rate.55
Mortalityisoftensignificantwhenthediseaseisleftuntreated[13-15],aconsequence56
thatmakesISaseriouseconomicandwelfareprobleminfarming.57
Tiamuliniseffective in treating IS causedbyBrachyspirahyodystenteriae,B.hampsonii58
andB.pilosicoli in swine [16-18] and in poultry [19-21]. Tiamulin is a bacteriostatic59
derived from a natural pleuromutilin that binds the 50S region of the ribosome to60
inhibit protein synthesis [22]. The antibiotic blocks peptide bond formation by61
interfering with substrate binding [22-25]. Tiamulin treatment in farms generally62
resultsinclearanceofinfectionandassociatedsymptoms.However,reoccurrenceofthe63
disease canbe observedpost treatment indicating incomplete clearance andpossibly64
decreased susceptibility [26,27] in response to treatment. The reason may be an65
inappropriate dosing as there is currently there is a lack of an internationally66
recognized standardized method to determine tiamulin minimum inhibitory67
concentration (MIC) for this bacterium, which has impacts upon selection of an68
appropriate treatment dose. Furthermore, recent studies have indicated that69
Brachyspiramay acquire resistance against tiamulin and, other than blocking protein70
synthesis,nothingisyetknownofthemetabolicresponseofB.pilosicolitotiamulin.We71
argue that evaluating this using a metabonomics approach would allow a better72
understanding of the bacterial response to tiamulin and give insights into improving73
selectionofeffectivedosingregimes.74
Metabonomicsallowsnon-targetedevaluationofthemetabolicmodificationsoccurring75
in abiological system in response to a stress [28],which in this study is exposure to76
tiamulin. By providing a general overview of the metabolic response, this technique77
allows the generation of new hypotheses and to evaluate metabolic in response to78
environmental stress or genetic modification. In this study, we used an NMR-based79
metabonomicsapproachcoupledwithmultivariatestatisticstoevaluatethemetabolic80
dose-response ofB.pilosicoli to tiamulin. Bacteriawere exposed to gradual antibiotic81
doses and media were sampled over 120h in order to evaluate the evolution of its82
metaboliccompositionduringgrowth.Thisallowedtosnapshotthemetabolicresponse83
ofB.pilosicolitotiamulin.84
85
2.MaterialandMethods86
2.1.Bacterialgrowthandantibioticassay87
B.pilosicoliB2904isolatedfromchickenpresentingclinicalsignsofAISintheUK[29]88
were grown from frozen stock on agar solidified BEB plate for four days under89
anaerobic conditions (94% N2and 6% CO2) at 37°C. Colonies were transferred into90
Brachyspiraenrichmentbrothmedia(BEBsupplementedwithheartinfusion)forthree91
daysundersimilarconditions.ThebacterialconcentrationwasthenadjustedinBEBto92
1 x106CFU/ml and transferred into24well plates (2mlperwell) and incubated as93
above for 120h. Every 24 h (with a first time point at 0 h growth), the entire well94
contentwastakenandcentrifugedfor2minat2400gtoseparategrowthmediumfrom95
bacteria. The supernatant was kept at - 80°C for further analysis. This process was96
repeatedateachtimepointinsextuplettodelivertheappropriatepowerforstatistical97
analysis.98
Thesamemethodwasusedforthetiamulinassay.Bacterialcellsweregrownasabove99
and bacterial pellets were resuspended in BEB with antibiotic at six concentrations100
(0.008, 0.016, 0.031, 0.062, 0.125 and 0.250 µg/ml plus control). Bacteriawere then101
inoculated into 24 well plates as previously described and incubated for 120h. For102
metabolicanalysis,eachcondition(tiamulinconcentration)andtimepoint(every24h103
for 120 h) were also repeated in sextuplet. The medium was not changed for the104
durationoftheexperimentsothatantibioticexposurewascontinuous.105
ToevaluatetheviabilityofB.pilosicolipost-antibioticexposure,theaboveexperiment106
wasrepeatedintriplicate.Following120hincubation,B.pilosicoliwasinoculatedonto107
fastidiousanaerobicbloodagarandincubatedat370C,anaerobicallyfor48h.Following108
incubationallplateswerevisuallyinspectedforbacterialgrowth.109
B. pilosicoli growth was evaluated using the same experimental design as the one110
previously described. Bacteria were grown in a 96 well plate (0.2 ml per well) and111
bacterialgrowthwasevaluatedevery2hfor120hatanabsorbanceof600nmusinga112
Fluostar(Info).Waterwasusedasblankandbrothmediawithoutbacteriaasanegative113
control.Eachcondition(tiamulinconcentration)wasrepeatedintriplicateandresults114
arepresentedasanaverageofthelogofthebacterialconcentrationcalculatedfromthe115
absorbance observed at each tiamulin concentration per time point after correction116
withstandardcurve.117
118
2.2.NMRspectroscopy119
ForNMRspectroscopy,0.4mlofmediawasaddedto0.2mlofNMRphosphatebuffer120
(madeinD2Ocontaining10%waterand0.05%sodium3-(tri-methylsilyl)propionate-121
2,2,3,3-d4(TSP)asa1HNMRreference)and0.5mlofthesolutionwastransferredinto122
5mmofouterdiameterNMRtubes.1H-NMRspectrawereacquiredonaBrukerAvance123
DRX700MHzNMRSpectrometer(BrukerBiopsin,Rheinstetten,Germany)operatingat124
700.19 MHz and equipped with a cryogenic probe from the same manufacturer. A125
standard 1-dimensional (1D) pulse sequence [recycle delay (RD)-90°-t1-90°-tm-90°-126
acquire free induction decay (FID)]withwater suppression applied during RD of 2s127
andamixing time (tm)of100msanda90°pulse setat10μswasapplied.Foreach128
spectrum 128 scans were recorded on a total of 32K data points. A broadening line129
function of 0.3 Hz was used to multiply all FIDs. After acquisition, all spectra were130
manually phased and baseline corrected using the software MestReNova® (version131
2.1.8-11880,MestreLab,Spain).Finally,spectrawerecalibratedtothechemicalshiftof132
TSP(δ0.00).Inordertofacilitatemetaboliteidentificationbasedonliterature,aseries133
of2Dspectraonselectedsampleswereacquiredusingcorrelationspectroscopy(COSY)134
NMRspectroscopy.135
136
3.3.Statisticalanalysis137
All spectra were scaled on unit variance and mean centered prior to analysis. To138
evaluatemetabolicvariationbetweensamples,principalcomponentanalysis(PCA)was139
used.Orthogonal projection to latent structure discriminant analysis (O-PLS-DA)was140
alsoperformed,where1H-NMRspectrawereusedasamatrixofindependentvariables141
(X)andtimeorantibioticconcentrationwereusedaspredictionvectors(Y)tocapture142
themetabolicvariations linear to timeandantibioticconcentration.O-PLSDAmodels143
were generated between each tiamulin concentration at every time point144
independently.Aheatmapwasgeneratedusingeachofthismodelstrengthinorderto145
visualizewhentiamulinimpactedbacterialmetabolismincomparisontocontrolandif146
clustersrelatedtodosecouldbeobserved.147
148
3.Results149
3.1.ModificationsofB.pilosicolimetabolismduringgrowth150
The aim of this study was to characterize B. pilosicoli metabolism under optimum151
growthconditions.Figure1presentsB.pilosicoligrowthandmetabolicactivityinbroth152
media.PC1,whichcaptured49%ofthemetabolicvariation,indicatedthatasignificant153
metabolic shiftwas recordedafter96hof incubation.Distinctionsbetween0h, 24h154
and 48 h were observed on the 3rd component, representing only 9% of the total155
variation, which suggests amodest effect on the composition of the culturemedium156
overthefirst48h.Thismetabolictrajectoryindicatesthatbacterialmetabolismmight157
changedependingonthegrowthphase(Figure1AandB).Scoresfromthesametime158
pointwereclusteredtogetherindicatinggoodreproducibilityoftheexperiment.159
In theantibiotic freeculturemedium, thegrowthofB.pilosicoliwasassociatedwitha160
decrease in glucose and an increase in amino acids (phenylalanine, alanine, tyrosine,161
lysine, valine and methionine), fermentation products (lactate, acetate, butyrate and162
isovalerate),aswellasothercompoundsinvolvedintheregulationofcellosmosissuch163
asmyo-inositolandtrimethylamine(TMA)asobservedinthePCAresultspresentedin164
Figures1AandC.165
Glucosewastheonlyreadily identifiablesubstrate thatshowedareductionover time166
(Figure1C).Decreasedconcentrationofothersubstratescouldnotbedetectedanditis167
possible that some may be below the detection limit of the NMR instrument.168
Nevertheless,itisnotunreasonabletoassumethatglucosewastheonlycarbonsource169
used for bacterial anabolism and growth. B.pilosicoli and more especially the strain170
used for this experiment (B2904) is able to use a wide range of carbohydrates and171
hexosesasprimarycarbonsources[10]butitwouldseeminthisstudythatglucosewas172
usedpreferentially.173
174
3.2.TiamulinimpactsB.pilosicoligrowthevenatverylowdoses175
Having characterized the metabolic footprint of B. pilosicoli when grown in optimal176
conditions,wethenchallengeditwithincreasingtiamulindoses.TiamulinimpactedB.177
pilosicoli’s growth at the lowest concentrations tested (0.008 and 0.016 µg/mL) as178
displayed in Figure 2. For these two doses, the bacterial count observed at the179
stationary phasewas one log lower than for the control demonstrating the ability of180
tiamulin to reduce the growth ofB.pilosicoli at low concentrations. Up to 54 hours,181
growth curves of the two lowest concentrations (0.008 and 0.016 µg/mL) were182
identicaltothecontrol(T1andT2onthegraph)buttheystoppedgrowingshortlyafter183
and entered into the stationery phase. No bacterial growth was detected for higher184
tiamulinconcentrations (over0.032µg/mL)confirming itsefficiency tostopbacterial185
proliferation.Interestingly,nogradualtiamulindoseresponseofbacterialgrowthwas186
observed. Indeed,growthratesweresimilar for thetwo lowestconcentrations(0.008187
and 0.016 µg/mL) while higher doses induced a complete inhibition ofB.pilosicoli’s188
growth.189
190
3.3.MetabolicresponseofB.pilosicolitotiamulin191
A clear metabolic response of B. pilosicoli to tiamulin could be observed when the192
antibiotic dose exceeded to 0.032 µg/mL. At lower doses (0.008 and 0.016 µg/mL),193
although B. pilosicoli growth was decreased by 1 log (Figure 2) compared with194
untreated control, the metabolic trajectories remained unaffected (Figure 3A,195
SupplementalFig.1and2).196
When bacteria were exposed to 0.032 µg/mL of tiamulin (Supplemental Fig. 3), a197
disruptionofthemetabolictrajectorywasobservedwhichwasduetomodificationsof198
amino acid concentration. A noticeable increase of tyrosine, methionine, valine,199
phenylalanineandlysineintothemediumfrom0to96hwasobserved.Afterthattime,200
theirconcentrationreduced,indicatingconsumptionoftheseaminoacidsuntiltheend201
of theexperiment.After120hofgrowth, themetaboliccompositionof themediawas202
comparable to control, indicating the full recovery of the metabolism following203
antibioticexposure.204
Higherdosesoftiamulin(0.062and0.125µg/mL)inducedsimilarresponsestothose205
observed at 0.032 µg/mL (Figure 3A and Supplemental Fig. 4. A and B). Amino acid206
metabolismwasaffectedinagreaterextent.207
At themaximumdosetested(0.250µg/mL)themetabolic trajectoryobservedfor the208
mediawasdrasticallymodifiedincomparisontothosedescribedpreviously(Figure3209
andSupplementalFig.5).Themetabolictrajectoryfollowedacircularshapewherethe210
scoresofthesamplescollectedafter120hofbacterialgrowthwereclusteredwiththe211
ones observed at T0, indicating metabolic similarities with the baseline time-point.212
Onceagain,aminoacidswerereleased into themediumaswellasbutyrateandmyo-213
inositol.214
215
3.4.B.pilosicolisurvivespost-antibioticexposure216
ToevaluatethesurvivingpotentialofB.pilosicoliaftertiamulinexposure,samplesfrom217
replicationofthegrowthcurveexperimentwereplatedonagarplateattheendofthe218
antibioticchallenge(120h).ForalltiamulindosesapplieditwaspossibletoobserveB.219
pilosicolicolonyformationonagarplates(SupplementTable1).Theseresultsindicate220
thatB.pilosicoli isabletorecoverfromtiamulinexposure(evenatthehighestdoseof221
0.250µg/mL)oncebackinoptimalgrowthconditions.222
223
4.Discussion224
The results obtained fromB.pilosicoli growth in a control mediumwithout tiamulin225
providenewinsightsaboutitsgeneralmetabolism.Thebacteriawereabletoproduce226
lactateandacetate fromglucose fermentationwithout secretingmethanol, suggesting227
theuseofthebifidumpathwayaccordingtothefollowingequation:glucoseà3acetate228
+ 2 lactate [30]. However, lactate was generally found in very small quantity in229
comparisontoacetate,indicatingitspotentialuseinothermetabolicreactions.Bacteria230
werealsoabletosecretebutyricacidbutnotpropionicacid.Bothoftheseshortchain231
fattyacidswerefoundtobepotentialcarbonsourcesforB.pilosicoli[10].Bacteriaalso232
released a largenumber of amino acids that couldbe caused either by synthesis and233
active secretion of these amino acids, or more likely due to exogenous protein234
degradation.ThisresultisinaccordancewiththegeneticresultspublishedbyMappley235
et al. [10] that indicated a strong proteolytic capacity of the bacterium. This specific236
strainofB.pilosicoliwasalsoshowntobeable touseaminoacidsasprimarycarbon237
source [10].However, as the bacterium favors glucose if available as primary carbon238
source, amino acidsmay here only be used for protein synthesis andmay therefore239
become in excess in the culturemediumwhere they accumulate. Finally, thebacteria240
secretedTMA.GutbacteriagenerallyproduceTMAfromdietaryL-carnitine,betaineor241
choline. Yet, it was not possible to detect a decrease in concentration of these242
compoundsindicatingthatB.pilosicolimightnotusethesemoleculesasprecursorsor243
thatthetechniqueusedwasnotsensitiveenoughtodetectsuchvariations.244
More importantly, thisworkconfirmed tiamulinability to significantly reduceatvery245
low doses (0.008 and 0.016 µg/mL) and inhibit at higher concentrationsB.pilosicoli246
growth. Decreased growth rate at such low antibiotic doses were unexpected, as247
previousevaluationofminimuminhibitoryconcentration(MIC)valuesforthisspecific248
strainwereof0.250µg/mL[31],furthermore,10-15%ofB.pilosicoliisolatespresented249
MICs>4µg/mL[31].Differences in theMICvaluescanbeexplainedbyexperimental250
conditions. Growth curveswere acquiredwhenB.pilosicoliwas grown in BEBmedia251
ratherthanonagarplatesforMICtests.TheB.pilosicolistrainB2904usedinthisstudy252
isknowntohaveanMICof0.250µg/mL[31,32]butshowedclearinhibitionofgrowth253
with concentrations below this value. Thus, our findings confirm the previously254
reported observation that lower tiamulin MIC values are generally found in broth255
comparedtoagarforB.hyodysenteriae[33].256
Despite these encouraging results regarding tiamulin efficiency to inhibit pathogen257
growth, the evaluation of B. pilosicoli metabolic viability in response to antibiotic258
treatment revealed that classic MIC calculations might not be sufficient to assess259
antibiotic efficiency. Indeed, B. pilosicoli metabolic rate reduction did not mirror260
previouslycommentedreducedgrowthrate in response to tiamulin treatment.At the261
two lowest dosesused (0.008 and0.016µg/mL),B.pilosicoli growthwas reducedby262
onelogincomparisontocontrol.However,themetabolictrajectoriesobservedbythe263
mediawere identical. Indicating that tiamulinwas able to impact growth but thatB.264
pilosicolibasicmetabolismremainedunaffected.Higherdoseswereabletoreducebut265
notsilentB.pilosicolimetabolismdespitecompletegrowthinhibition.Furthermore,the266
metabolismofB.pilosicoliappears toslightlyrecover fromall tiamulinconcentrations267
exceptfrom0.250µg/mlafter120hofgrowth.Thismightbeduetotheapparitionof268
resistance,whichisknownasbeingaslowbacterialdevelopmentprocess[34,35].269
In addition, it was demonstrated that B. pilosicoli was able to survive post-tiamulin270
exposure even at the highest antibiotic dosewhen plated on agar (after 120 h drug271
exposure). Indeed B. pilosicoli colonies were identified 48h after the end of the272
antibiotic treatment for all doses tested (0.008-0.250µg/mL).These results illustrate273
thepotentialofB.pilosicolitoenteradormancystatewhenexposedtotiamulinthatis274
reversibleoncethetreatmentperiodisover.275
Besides,resultsdemonstratedaslowresponseofthebacteriatoantibiotictreatmentas276
modification of the metabolic footprint was only observed after more than 48 h of277
growth in presence of tiamulin. From these results, it seems that metabolism was278
stressed during the exponential phase, when bacterial division is compromised.279
Metabolism modification was mainly associated with increased amino acid280
consumption(thatwereproducedinthecontrol).However,astheprovenanceofthese281
aminoacidsremainsunclear,twohypothesescanbeformulated.Firstly,inresponseto282
antibiotic stress bacteria could use amino acids as alternative energy substrates.283
Secondly, B. pilosicoli might not be able to hydrolyse proteins present in the media284
becausenewproteinsynthesis,suchassecretedproteases,isblockedattheribosome.285
Thespecificityof theaminoacidsused indicates the firstoption is themostprobable286
and that catabolism repression could be overridden to secure energy from multiple287
sources. This is an interesting hypothesis that needs confirmation by alternative288
techniquessuchastranscriptomics.289
ThefactthatB.pilosicoliremainviableandmetabolicallyactivewithoutdividingdespite290
the antibiotic treatment and is able to recover after antibiotic exposure could partly291
explaintheISrelapseobservedinfarmsaftertiamulinintervention.Indeed,itseemsto292
arisefromtheseresultsthatbacteriaremainviablebutarenotabletodivideentering293
thereforeadormancystage. It ishighlypossible that suchphenomenonoccurs in the294
intestinal lumen, where bacteria could suffer from inactivation of cell division but295
remain viable. This bacterial state might be associated with a decrease in their296
pathogenicity explaining the disappearance of associated symptoms. Nevertheless,297
bacteriamightremainviablebutata“dormancy”stateintheintestinallumenoranimal298
faecesuntil the environmentbecomes lesshostile (endof antibiotic treatment)when299
they can recover theirpathogenicproperty.However, it is impotent topoint out that300
here B. pilosicoli was recovered on agar plates that represent optimal growth301
conditions. Thismight not be the case in the intestinal lumen that is amore hostile302
environmentwhenthebacteriumalsoneedstocompetewithothercommensalbacteria303
andisconfrontedtothehost’simmunesystem.304
As increasing antibiotic resistancemechanisms developed by pathogenic bacteria are305
arising,manystudiesandreviewshavestressed theconcerns linked to inappropriate306
antibioticusage.Indeed,thisisstronglylinkedtodevelopmentofantibioticresistance,a307
burden forhealth and industry. There is therefore a surge for redefining appropriate308
antibiotic use that would help minimizing the current concern linked to decreased309
antibiotic efficiency. Thus, importance should be given to newmethods development310
aiming at better assessing antibiotic efficiency and to detect potential antibiotic311
resistance factor development. This study indicates thatmetabonomics could be and312
easy and practical way to evaluate bacterial metabolic activity and therefore assess313
antibiotic efficiency to totally inactivate pathogens and therefore avoid infection314
relapse.315
316
5.Conclusion317
ThisworkgaveaclearerunderstandingofB.pilosicolimetabolisminoptimumgrowth318
conditions, including indication regarding favored fermentation pathways and amino319
acids metabolism. It supports the fact that tiamulin can inhibit efficiently bacterial320
growthatlowconcentrations.However,itwassurprisingtoobservethattiamulincould321
impactB.pilosicoligrowthwithoutinfluencingitsbasicmetabolism.Italsorevealsthat322
thebacteriumtrytomaintainmetabolichomeostasisdespiteanobviousstressvisible323
on the growth curve, demonstrating that in response to xenobiotic stress, bacterial324
division is the first mechanism to be suspended. This indicates that tiamulin might325
presentagoodsolutionagainstAISoutbreaks,asitisabletosignificantlyreduceorstop326
bacterial growth, provided that the efficient dose is achieved in the gut of every327
individual.Evenso,thetreatmentmaynotbesufficienttoavoidrelapseofthedisease.328
Suchfindingssuggestthatmeasurementofbacterialactivitymightbeneededinorder329
to assess antibiotic efficiency against potential reoccurrence of the disease. In that330
prospect, metabonomics appeared as a potential solution to evaluate if antibiotic331
treatmentcaninactivatemicrobialmetabolicactivity.332
333
References334
[1] Stanton,T.B.&Hampson,D.J.PhysiologyofRuminalandintestinalspirochaetes.335CABInt.7–45(1997).336
[2] Duhamel,G.E.,Stryker,C.J.,Lu,G.,Wong,V.J.&Tarara,R.P.Colonicspirochetosis337ofcolony-raisedrhesusmacaquesassociatedwithBrachyspiraandHelicobacter.338Anaerobe.9,45–55(2003).339
[3] Hidalgo,A.,Rubio,P.,Osorio,J.&Carvajal,A.PrevalenceofBrachyspirapilosicoli340and ‘Brachyspira canis’ in dogs and their association with diarrhoea. Vet.341Microbiol.146,356–60(2010).342
[4] Trott,D.J.,Stanton,T.B.,Jensen,N.S.,Duhamel,G.E.,Johnson,J.L.,&Hampson,D.343J.Serpulinapilosicoli sp., theagentofporcine intestinal spirochetosis. Int.J.Syst.344Bacteriol.46,206–15(1996).345
[5] Naresh, R. & Hampson, D. J. Attraction of Brachyspira pilosicoli to mucin.346Microbiology.156,191–7(2010).347
[6] Hopwood,D.,Pethick,D.&Hampson,D.Increasingtheviscosityoftheintestinal348
contents stimulates proliferation of enterotoxigenic Escherichia coli and349Brachyspirapilosicoliinweanerpigs.Br.J.Nurition.88,523–532(2007).350
[7] Prapasarakul, N., Lugsomya, K., Disatian, S., Lekdumrongsak, T., Banlunara,W.,351Chetanachan,P.,&Hampson,D. J. Faecal excretionof intestinal spirochaetesby352urbandogs,andtheirpathogenicityinachickmodelofintestinalspirochaetosis.353Res.Vet.Sci.91,e38–43(2011).354
[8] Li,C.,Motaleb,A.,Sal,M.,Goldstein,S.F.&Charon,N.W.Spirocheteperiplasmic355flagellaandmotility.J.Mol.Microbiol.Biotechnol.2,345–54(2000).356
[9] Dworkin,M.Theprokaryotes:Vol.7:Proteobacteria:DeltaandEpsilonSubclasses.357S.Falkow,E.Rosenberg,K.H.Schleifer,&E.Stackebrandt(Eds.).SpringerScience358&BusinessMedia.2006.359
[10] Mappley,L.J.,Black,M.L.,AbuOun,M.,Darby,A.C.,Woodward,M.J.,Parkhill, J.,360Turner,A.K.,Bellgard,M.I.,La,T.,Philips,N.D.,LaRagione,R.M.&Hampson,D.361J. Comparative genomics of Brachyspira pilosicoli strains: genome362rearrangements, reductions and correlation of genetic compliment with363phenotypicdiversity.BMCGenomics.13,454(2012).364
[11] Dassanayake,R.P.Biochemicalpropertiesofmembrane-associatedproteasesof365Brachyspira pilosicoli isolated from humans with intestinal disorders. J. Med.366Microbiol.53,319–323(2004).367
[12] Nakamura, S., Adachi, Y., Goto, T. & Magariyama, Y. Improvement in Motion368Efficiency of the Spirochete Brachyspira pilosicoli. Biophys. J. 90, 3019–3029369(2006).370
[13] Fellström, C. & Gunnarsson, A. Phenotypical characterisation of intestinal371spirochaetesisolatedfrompigs.Res.Vet.Sci.59,1–4(1995).372
[14] Duhamel, G. E., Hunsaker, B. D., Mathiesen, M. R., & Moxley, R. A. Intestinal373spirochaetosisandgiardasisinabeaglepupwithdiarrhoea.Vet.Pathol.33,360-374362(1996)375
[15] Taylor,D.J.,Simmons,J.R.&Laird,H.M.Productionofdirrhoeaanddysenteryin376pigs by feeding pure cultures of spirochaete differing from Treponemu377hyodysenteriae.Vet.106,326–332(1980).378
[16] Johnston,W.T.,Dewey,C.E.,Friendship,R.M.,Smart,N.,McEwen,B. J.,Stalker,379M.,&deLange,C.F.Aninvestigationoftheetiologyofamilddiarrheaobservedin380agroupofgrower/finisherpigs.Can.Vet.J.42,33–37(2001).381
[17] Wilberts, B. L., Arruda, P.H.,Warneke,H. L., Erlandson, K. R., Hammer, J.M.,&382Burrough, E. R. Cessation of clinical disease and spirochete shedding after383tiamulin treatment inpigsexperimentally infectedwith ‘Brachyspirahampsonii’.384
Res.Vet.Sci.97,341–347(2014).385
[18] Burch,D.TiamulinactivityagainstBrachyspirahyodysenteriae.Vet.Rec.163,698386(2008).387
[19] Burch, D. & Klein, U. Treatment ofBrachyspira species with highMICs against388Tiamulininlayers.6thInt.Conf.AvianIntest.SpyrochaetosisInfect.Anim.Humans.38951(2013).390
[20] Stephens, C. P. & Hampson, D. J. Evaluation of tiamulin and lincomycin for the391treatment of broiler breeders experimentally infected with the intestinal392spirochaeteBrachyspirapilosicoli.AvianPathol.31,299–304(2002).393
[21] Burch, D. G. S., Harding, C., Alvarez, R. & Valks,M. Treatment of a field case of394avian intestinal spirochaetosis caused by Brachyspira pilosicoli with tiamulin.395AvianPathol.35,211–6(2006).396
[22] Schlünzen,F.,Pyetan,E.,Fucini,P.,Yonath,A.&Harms,J.M.Inhibitionofpeptide397bond formation by pleuromutilins: the structure of the 50S ribosomal subunit398fromDeinococcusradioduransincomplexwithtiamulin.Mol.Microbiol.54,1287–3991294(2004).400
[23] Poulsen,S.M.,Karlsson,M.,Johansson,L.B.,&Vester,B.Thepleuromutilindrugs401tiamulinandvalnemulinbindtotheRNAatthepeptidyltransferasecentreonthe402ribosome.Mol.Microbiol.41,1091–1099(2001).403
[24] Long, K. S., Hansen, L. H., Jakobsen, L. & Vester, B. Interaction of pleuromutilin404derivatives with the ribosomal peptidyl transferase center. Antimicrob. Agents405Chemother.50,1458–1462(2006).406
[25] Forschungsinstitut, S. TheModeof actionof pleuromutilinderivatives effect on407cell-freepolypeptidesynthesis.Europ.J.Biochem.533,527–533(1974).408
[26] Stephens, C. P. & Hampson, D. J. Evaluation of tiamulin and lincomycin for the409treatment of broiler breeders experimentally infected with the intestinal410spirochaeteBrachyspirapilosicoli.AvianPathol.31,299–304(2002).411
[27] Burch, D. G. S., Harding, C., Alvarez, R. & Valks,M. Treatment of a field case of412avian intestinal spirochaetosis caused by Brachyspira pilosicoli with tiamulin.413AvianPathol.35,211–6(2006).414
[28] Sperling, D., Smola, J. & Cízek, A. Characterisation ofmultiresistantBrachyspira415hyodysenteriaeisolatesfromCzechpigfarms.Vet.Rec.168,215(2011).416
[29] Karlsson, M., Aspan, A., Landén, A., & Franklin, A. Further characterization of417porcine Brachyspira hyodysenteriae isolates with decreased susceptibility to418tiamulin.J.Med.Microbiol.53,281–285(2004).419
[30] Nicholson, J. K., Lindon, J. C. & Holmes, E. ‘Metabonomics’: understanding the420
metabolic responses of living systems to pathophysiological stimuli via421multivariatestatisticalanalysisofbiologicalNMRspectroscopicdata.Xenobiotica.42229,1181–9(1999).423
[31] Mappley,L.J.,Tchórzewska,M.a,Cooley,W.a,Woodward,M.J.&LaRagione,R.424M. Lactobacilli antagonize the growth, motility, and adherence of Brachyspira425pilosicoli: a potential intervention against avian intestinal spirochetosis. Appl.426Environ.Microbiol.77,5402–5411(2011).427
[32] Gottschalk,G.Bacterialmetabolism.SpringerScience&BusinessMedia.(2012).428
[33] Pringle, M., Landén, A., Unnerstad, H. E., Molander, B. & Bengtsson, B.429Antimicrobial susceptibility of porcine Brachyspira hyodysenteriae and430BrachyspirapilosicoliisolatedinSwedenbetween1990and2010.ActaVet.Scand.43154,54(2012).432
[34] Woodward,M.J.,Mappley,L.,LeRoy,C.,Claus,S.P.,Davies,P.,Thompson,G.,&La433Ragione,R.M.DrinkingwaterapplicationofDenagard®Tiamulin forcontrolof434Brachyspirapilosicoliinfectionoflayingpoultry.Res.Vet.Sci.103,87–95(2015).435
[35] Rohde, J., Kessler, M., Baums, C. G. & Amtsberg, G. Comparison of methods for436antimicrobial susceptibility testing and MIC values for pleuromutilin drugs for437Brachyspira hyodysenteriae isolated in Germany. Vet. Microbiol. 102, 25–32438(2004).439
[36] Bock,A.,Turnowsky,F.,Hogenauer,G.,Universitat,M.D.&Germany,W.Tiamulin440resistancemutationsinEscherichiacoli.J.Bacteriol.151,1253–1260(1982).441
[37] Karlsson, M., Gunnarsson, A. & Franklin, A. Susceptibility to pleuromutilins in442Brachyspira(Serpulina)hyodysenteriae.Anim.Heal.Res.Rev.2,59–66(2001).443
444
Figuresandtables445
446
Figure 1: B. pilosicoli consumed glucose and released amino acids and fermentation447
products in its environment. (A) PCA scores plot. (B) B. pilosicoli growth curve in448
Brachyspiraenrichmentbrothmediaunderanaerobicconditionsmadeintriplicate.(C)449
Associatedloadingsofthefirstcomponent.Themetabolictrajectoriesdescribedbythe450
arrowsweredeterminedbythepositionofthecentroidscalculatedateachtimepoint451
usingthecoordinateoftheassociatedscoresontheprincipalcomponents.452
453
Figure 2: Impact of tiamulin onB.pilosicoli growth inBrachyspira enrichment broth454
mediaunderanaerobicconditions.Growthwasmeasuredintriplicateevery2hfor120455
h.456
457
Figure3:(A)MetabolictrajectoriesderivedfromthePCAanalysisperformedusingall458
thesamplepopulation(N=252)ofthestudy(i.e.controlplus6tiamulindilution)onPC1459
and PC3 displaying the centroids for each time points of the control and three460
concentrations of tiamulin. (B) Heatmap representing the O-PLS DAmodel strength461
existing between each tiamulin concentration at each time point, based on R2Y462
(goodness of fit of themodel) and Q2Y (goodness of prediction of themodel) values463
usingthefollowingformula:𝑀𝑜𝑑𝑒𝑙𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ = ./0×2/0./032/0
.464
465
Supplementalmaterial466
467
Supplement Fig. 1:Metabolic trajectories ofB.pilosicoli footprints in brothmedia for468
120hatatiamulinconcentrationof:0.008μg/mL(A)and0.016μg/mL(B).Thearrows469
indicate themetabolic trajectory. Themetabolic trajectories described by the arrows470
weredeterminedbythepositionofthecentroidscalculatedateachtimepointusingthe471
coordinateoftheassociatedscoresonthePCs.472
473
Supplement Fig. 2: Metabolic variation related to PCA scores plots presented in474
SupplementFig.1.(A)loadingsoftheprincipalcomponent1ofthemodelpresentedin475
Fig A.1.A. (B) loadings of the principal component 3 of the model presented in476
SupplementFig.1.A.(C)loadingsoftheprincipalcomponent1ofthemodelpresented477
in Supplement Fig. 1.B. (D) loadings of the principal component 4 of the model478
presentedinSupplementFig.1.B.479
480
SupplementFig.3:MetabolictrajectoriesofB.pilosicolifootprintinbrothmediafor120481
hatatiamulinconcentrationof0.032μg.mlonprincipalcomponent1and3.(A)PCA482
scoreplot.(B)AssociatedloadingplotforPC1.Themetabolictrajectoriesdescribedby483
the arrowswere determined by the position of the centroids calculated at each time484
pointusingthecoordinateoftheassociatedscoresonthePCs.485
486
SupplementFig.4:MetabolictrajectoriesofB.pilosicolifootprintinbrothmediafor120487
h at a tiamulin concentration of: 0.062 μg/ml (A) and 0.125 μg/ml (B). The arrows488
indicate themetabolic trajectory. Themetabolic trajectories described by the arrows489
weredeterminedbythepositionofthecentroidscalculatedateachtimepointusingthe490
coordinateoftheassociatedscoresonthePCs.491
492
SupplementFig.5:MetabolictrajectoriesofB.pilosicolifootprintinbrothmediafor120493
hatatiamulinconcentrationof0.250μg/ml.(A)PCAscoreplot.(B)Associatedloading494
plot of PC1. The arrows indicate themetabolic trajectory. Themetabolic trajectories495
describedbythearrowsweredeterminedbythepositionofthecentroidscalculatedat496
eachtimepointusingthecoordinateontheassociatedscoresonthePCs.497
498
SupplementTab.1:GrowthscoresofB.pilosicolionagarplatepost-antibioticexposure.499
DetectionofB.pilosicoli’sgrowth48hafterplatingareindicatedbyasign‘+’andby‘-‘if500
nogrowthwasobserved.501