Master’s Thesis 2016 30 ECTSDepartment of Chemistry, Biotechnology and Food Science

A Comparative Analysis of Lactic Acid Bacteria Isolated from Honeybee Gut and Flowers, with Focus on Phylogeny and Plasmid Profile

Marte Sverdrup LinjordetChemistry and Biotechnology, Molecular Biology

AcknowledgementsThework presented in thisMSc thesiswas performed at the Laboratory ofMicrobial

Gene Technology (LMG), Department of Chemistry, Biotechnology and Food Science,

NorwegianUniversityofLifeSciences.

Iwould first like to thankmysupervisors,DzungBaoDiepandDanielMünch, for the

opportunityandguidanceprovidedforthisMScthesis.AthanktoKariOlsenforcarry

outtheGC-analysis,andtoClausKreibichforprovidingtheinformationonwhichplant

speciestocollect.IwouldalsoliketothankLindaGodagerandMay-BrittSelvågHovet

forallthehelpprovidedinthelab.Aspecialthanksgoestomyfellowstudent, Ingvild

Gallefoss,whohaskeptmecompanythelastfewmonths(andyears)andalsoprovided

mewithdatafromherownthesis.

Last, but not least, I want to thank my family and boyfriend for all the comfort and

encouragement through the past five years, and especially to my dad who is always

thereformewhenhelpisneeded.

MarteSverdupLinjordet

Oslo,May2016

AbstractApismellifera (honeybee) areofhugevalue as theyare themost importantpollinator

worldwide.Declinesinhoneybeepopulationshavemadethehoneybeesubjecttomuch

scientificresearch.Lacticacidbacteria(LAB)havebeendiscoveredinthehoneybeegut

and are believed to be of great importance to the honeybee health, protecting them

againstbeepathogens.ComparingLABcommunitiesinthehoneybeeguttothosefound

onflowersmayhelphighlighttherouteandimportanceoffloraltransmission.

Using cultivation techniques selective for Lactobacillus, plasmid and fermentation

profiling,and16SrDNAsequencing,LABwereisolatedfromdandelion,apple,rapeseed,

raspberryandwillowherbandcomparedtotheLABisolatedfromhoneybeeguts.

The results showed that Lactobacillus kunkeei and Fructobacillus fructosus were the

mostabundantofall the identifiedspecies inbothhoneybeegutand flowerssamples.

TheLABfloraofthehoneybeegutseemstoshiftfromL.kunkeeitoF.fructosusthrough

MaytolateJune,andF.fructosuswasalsofoundasthemostabundantLABinoneofthe

samples collected from honeybee guts in August. This is in accordancewith the LAB

florafoundondandelionandappleinMayandraspberryinlateJune,givingindications

towardapositivecorrelationbetweentheLABmicrobialflorainhoneybeegutandthe

flowerswithin their foraging area. The results also showed theplasmidprofileswere

more diverse in the honeybee gut samples, and although there seemed to be four

profiles that also occurred in the flower samples, indicating possible relatedness on

strainlevel,thedatasetsarerelativelysmallandfurtherinvestigationsareneeded.

Sammendrag

Apismellifera (honningbie) er av stor verdi og er den viktigste pollinatoren over hele

verden.Reduksjoneravhonningbienspopulasjonhargjortathonningbienerunderlagt

mye vitenskapelig forskning. Melkesyrebakterier har blitt oppdaget i tarmen til

honningbien og antas å være av stor betydning for honningbiens helse, hvor de

beskytter dem mot bie-patogener. Sammenligning av melkesyrebakteriene i

honningbienstarmogdesomfinnespåblomsterkankanskjehjelpetilmedåmarkere

veienogviktighetenavbakteriellblomsteroverføring.Vedhjelpavdyrketeknikkersom

er selektive for Lactobacillus, plasmid- og fermenteringsprofilering, og 16S rDNA

sekvensering, ble melkesyrebakteriene isolert fra løvetann, eple, raps, bringebær og

geitramsogsammenlignetmedmelkesyrebakterieneisolertfratarmentilhonningbien.

ResultatenevisteatLactobacilluskunkeeiogFructobacillusfructosusvardemesttallrike

av alle de identifiserte artene i både honningbietarmen og blomsterprøvene.

Melkesyrebakteriefloraen i tarmen ser ut til å skifte fra L. kunkeei til F. fructosus

gjennommai til slutten av juni, ogF. fructosus ble også funnet som denmest tallrike

melkesyrebakterien i en av prøvene samlet inn fra bie-tarmer i august. Dette er i

samsvarmedfloraensomblefunnetpåløvetannogepleimaiogpåbringebærislutten

av juni, noe som gir indikasjoner mot en positiv korrelasjon mellom

melkesyrebakteriefloraen i tarmen og blomstene i bienes pollineringsområde.

Resultatetvisteogsåatplasmidprofilenevarmervariert ihonningbieprøvene,ogselv

omdetsåuttilåværefireprofilersomogsåopptråddeiblomsterprøvene,noesomkan

indikereslektskappåstammenivå,vardatasettenerelativtsmåogvidereundersøkelser

erderfornødvendig.

Abbreviations

CHX Cycloheximide

FLAB FructophileLacticAcidBacteria

GC GasChromatography

GI Gastrointestinal

LAB LacticAcidBacteria

MRS DeMan-Rogosa-Sharp

NMBU NorwegianUniversityofLifeSciences

ON Overnight

PCR PolymeraseChainReaction

RT RoomTemperature

1

Contents

1. INTRODUCTION......................................................................................................................................2

2. MATERIALSANDMETHODS................................................................................................................62.1 SAMPLINGOFBACTERIAFROMFLOWERS..........................................................................................................62.2 GROWINGOFBACTERIALSTRAINS......................................................................................................................72.3 PLASMIDPROFILING...............................................................................................................................................72.3.1 IsolationofPlasmids......................................................................................................................................82.3.2 AgaroseGelElectrophoresis.......................................................................................................................82.3.3 RestrictionEndonucleaseCutting............................................................................................................9

2.4 PHYLOGENETICIDENTIFICATION......................................................................................................................102.4.1 PolymeraseChainReaction(PCR)........................................................................................................102.4.2 SequencingandProcessingofData.....................................................................................................11

2.5 SUGARFERMENTATIONPROFILING..................................................................................................................112.6 CYCLOHEXIMIDE...................................................................................................................................................12

3. RESULTS.................................................................................................................................................123.1 GROWINGOFBACTERIALSTRAINSANDPLASMIDPROFILING....................................................................123.2 PLASMIDPROFILING............................................................................................................................................133.3 PHYLOGENETICIDENTIFICATION......................................................................................................................153.4 SUGARFERMENTATION......................................................................................................................................163.5 CYCLOHEXIMIDE...................................................................................................................................................16

4. DISCUSSION...........................................................................................................................................184.1 LACTICACIDBACTERIAISOLATEDFROMFLOWERS.....................................................................................184.2 COMPARINGTHERESULTSTOLACTICACIDBACTERIAISOLATEDFROMHONEYBEEGUT...................194.2.1 ComparisonoftheLABSpecies..............................................................................................................194.2.2 ComparisonofthePlasmidProfiles.....................................................................................................21

4.3 BACTERIALTRANSMISSIONBETWEENFLOWERANDHONEYBEES............................................................224.4 CONCLUSION.........................................................................................................................................................23

5. REFERENCES.........................................................................................................................................24APPENDIX1–GROWTHMEDIA,BUFFERSANDTABLES......................................................................................27APPENDIX2-RESULTS...............................................................................................................................................28APPENDIX3-PROTOCOLS..........................................................................................................................................29

2

1. IntroductionApismellifera(honeybee)areofhugevalue,notonlyfortheirproductionofhoney,but

more importantly, because they are themost important pollinatorworldwide. Recent

declines in both wild and domestic pollinators have been a major concern and have

madethehoneybeesubjecttomuchscientificresearch.

Thegastrointestinaltractinbothhumansandanimalsisknowntoharboracollectionof

microorganisms called the gut microbiota. By extensive research on the human gut

microbiota, there is evidence that the gut bacteria influence human and animal

physiology,metabolism,nutrition,andimmunefunction(1).Thecompositionofthegut

microbiotaissubjecttodynamicchangesduringthecourseofthehost´sdevelopment,

physiologicalstatusorhealth.Forexample,differentgastrointestinal(GI)disordersare

oftenlinkedwithdegeneratingchangestothegutmicrobiota.Correspondingly,several

recentstudieshave investigated thegutmicrobiotaofhoneybeesasa response to the

alarmingdeclineinpollinators(2).

Honeybeesharboranumberofcommensalorbeneficialbacteriadistributedthroughout

the different compartments of their GI tracts. The GI tract of an adult honeybee is

dividedintofourmajorcompartments:crop(honeystomach),midgut,ileumandrectum.

Each compartment has a distinct environment favoring specific microorganisms (3).

Severalfindingshaveindicatedthatthehoneybeegutiscolonizedbyadistinctivesetof

bacterial species designated as the core gut microbiota (4). Because the community

compositionhasshowntochangethroughthebee´slifecycle,thecolonizationofthegut

is believed to be influenced by the honeybee´s age (3). During the course of their

lifespan,honeybeeworkersperformmanydifferent tasks that can contribute to these

variations. Newly emerged worker bees nurse larvae within the hive whereas older

worker bees build and maintain the wax combs, defend the colony, and receive and

process food that is collected by foragers. Foragers are specialized worker bees and

theirjobistobringbacknectarandpollenfromdifferentflowerstheyvisitduringthe

season.

In addition to the coremicrobiota in the gut, a novel lactic acid bacterial (LAB) flora

composed of 13 taxonomically well-defined Lactobacillus andBifidobacterium species

3

werediscoveredinthehoneystomachofhoneybees(5,6).Thehoneystomachfunctions

asaninflatablebagthatcantransportthenectarbacktothehiveforstorageandhoney

production.ItishypothesizedthatLABplayakeyroleintheconversionofbothnectar

tohoneyandpollentobeebread(storedfoodrichinprotein)duetotheirfermentation

properties(5,7)TheLABmicrobiotais,besidestheimportanceinbeefoodprocessing,

believedtobeofgreatimportancetothehoneybeehealth,protectingthemagainstbee

pathogens(8,9)andcontributingtotheantimicrobialpropertiesofhoney(10).

LAB are Gram-positive, usually non-motile rods or cocci that do not form spores and

produce lactic acid as their major or sole fermentation product. They require

environments rich in sugars, amino acids, nucleic acid derivatives, minerals and

vitamins (11). LAB reside in a diversity of different habitats that includes the

gastrointestinal tract of humans and animals, as well as in a number of other

environments suchasplantsanddifferentprocessed foodproducts.LABare foundas

commensals within humans, animals and insects. They are considered as beneficial

organisms commonly found in healthy individuals by protecting their hosts through

antimicrobialmetabolites(12).ManyLABspeciesarealsoindispensabletothefoodand

dairy industry because of their key role in in the formation of taste, texture and

preservation effects due to the production of antimicrobial peptides such as

bacteriocins,propionicacidandotherorganicacidsthatlowersthepH(11).

LABarefoundintwodistinctphyla:FirmicutesandActinobacteria.Themostimportant

genera of LAB within the Firmicutes are Enterococcus, Lactobacillus, Lactococcus,

Leuconostoc,Pedicoccus,StreptococcusandWeissella,whichallhavealowG+Ccontent.

LAB in theActinobacteria phylum only includes species of theBifidobacterium genus

thatincontrasttotheFirmicutesmembershaveahighG+Ccontent.(13,14)

Fructophilic LAB (FLAB) are a special group of LAB that prefers fructose instead of

glucose as growth substrate (15). Fructobacillus spp. and Lactobacillus kunkeei are

representativesofthesemicroorganisms.Inpreviousstudies,L.kunkeeihasbeenfound

to be one of the most predominant LAB in honeybees (16, 17). FLAB are found in

fructoserichnichesandhave–inadditiontohoneybees–beenisolatedfrombeehives,

fruits,andflowers(15,16,18).

4

Flowers areoneof themanydiversehabitatsprovidedbyplants formicroorganisms.

The availability of nutrients differs betweenplant parts such as roots, leaves, flowers

and fruits, andwithinplant organs (19), and therefor contributes to varyingbacterial

communities.Thebacterialcommunitiescolonizingrootsandleaveshavebeenthemost

studiedformanyplantspecies.Onlyrecently, therehavebeenstudiesonthebacterial

compositionof floralnectar,whichinitiallywasconsideredasnotsuitableasbacterial

habitats due to their antimicrobial properties (20). Studies found that bacteria are

common inhabitants of floral nectar, and that the bacterial communities differed

between plant species andwere characterized by low species richness andmoderate

phylogenetic diversity (21, 22). For bees, the floral nectar is the main source of

carbohydrates and energy (‘fuel’),while pollenprovides proteins, lipids, vitamins and

mineralsforbroodrearinganddevelopment(23).

Specializedhoneybeeworkertypes,nectarforagers,canlearnflowerattributessuchas

shape,colorandodor,calledpollinatorsyndromesandusethisinformationtoselectfor

aparticularflower.Theindividualhoneybeetendstosticktoonetypeofflowerovera

certain period of time(24), and can also discriminate between small differences in

nectarconcentration.Theeffectthebacterialcommunitiesmayhaveonthefloralnectar

theycolonizeismainlyunexplored.Goodetal.(25)hypothesizesthatbacteriacanalter

thechemicalprofileofthenectar,thusaffectingthenectarattractivenesstopollinators

suchashoneybees. Inaddition,bacteria found in floralnectarare frequently found in

thehoneybeehiveenvironmentandalimentarytract(26),raisingquestionsconcerning

the transmission effect between flowers andpollinators visiting them for their nectar

andpollen.

Foodsshapethegutmicrobiotainanimalsandinsects.Thegutmicrobiotainhoneybees

has previously been shown to change in a season-dependent manner (27). Our

hypothesis is thatthiscanberelatedtothetypeof foodthehoneybeeshaveaccessto

along theaxisof time.During thewinter seasonhoneybees inNorway still have their

pollenstores,butreceiveartificialnectarfromthebeekeeperstoproducehoney,asno

flowers are available. During foraging season (spring and summer) honeybees fly out

and collect nectar and fresh pollen directly from flowers. The biochemistry and

5

complexity of food, and the microbial flora associated in foods, are therefor quite

different between winter season and foraging season, and may likely affect the gut

microbiota in honeybees. LAB are frequent in vegetables and fruits, and are also

commoninhabitantsingutmicrobiotaofmostvertebratesandinvertebrates.Theaimof

this MSc thesis is to characterize LAB in selected flowers on or near campus of the

NorwegianUniversityofLifeSciences.Thismaycontributetoabetterunderstandingof

howplantLABarerelatedtothosepresentinhoneybees.

Consequently,theaimsoftheworkofthisMSc-projectisto:

1. Isolate and characterize lactic acid bacteria from selected spring, summer and

autumnflowersusingPCR,plasmid-andfermentationprofiling

2. Compare the resultswith respect to lactic acid bacteria found in honeybee gut

sampledinparallelwiththeflowercollection

6

2. MaterialsandMethods

2.1 SamplingofBacteriafromFlowers

Flowerswerecollectedfromdifferent locationsonandnearcampusat theNorwegian

University of Life Sciences, Ås, Norway. To represent different nectar secretions and

relevantspeciesfortheproductionofhoneyinthedistrict,flowerswerecollectedfrom

five plant species: dandelion (Taraxacum officinale), rapeseed (Brassica napus),

willowherb (Chamerion angustifolium), apple (Malus) and raspberry (Rubus idaeus)

(Table2.1).

Table2.1Nectarproductionand importance for thedevelopmentofhoney fromsomeplant species in thedistrictsurroundingtheUniversityofLifeSciences,Ås,Norway.

Plantspecies Nectarproduction

Importanceforthehoneydevelopmentinthedistrict

Blooming

Dandelion Medium Importantforthedevelopment May–JuneFruit Medium–large Importantforthedevelopment May–JuneRaspberry Verylarge Cruciallyimportant70-90%ofthehoney June-JulyAutumnRapeseed

Medium–large Small May

Willowherb Large Small July-August

Toreducecontamination, flowerswerepickedusingrubberglovesandtransported in

plastic bags back to the laboratory the sameday. There, theywere cutwith a scalpel

understerileconditions.andtransferredtotwo50mlFalcontubesbeforetheaddition

of30ml0.9%NaClineachtube.Thenumberofflowersineachtubewas5,20,30,50

and110fordandelion,apple,rapeseed,willowherbandraspberryrespectively,toadjust

fordifferentflowermass/volumesastheyrangedgreatlyinsize.

Thetubesweremixedvigorouslyonavortexbeforethe liquidcontainingthebacteria

fromeachtubewastransferredtotwonew50mlFalcontubes.Aftercentrifugationat

5000gfor5minutes(Eppendorf5804R)thepelletinonetubewassuspendedin4ml

20%glycerolandthenmixedthoroughlywiththepelletintheothertube.Thebacterial

suspensionwasdistributedonfour2mlmicrotubes(Sarstedt),1mlineachtube,and

storedat-80°C.

7

2.2 GrowingofBacterialStrains

ThemediumandtheanaerobicconditionssetwerechosentofavorthegrowthofLAB.

Thebacterial flowersampleswerethawedatroomtemperature(RT)and100µlwere

seriallydilutedin0.9%NaClwithintherange10-1–10-6.

Volumes of 100 µl were plated out on De Man-Rogosa-Sharp (MRS) agar plates and

incubatedat30°Cinananaerobicchamberuntilvisiblecolonieswereformed,generally

within72hours.ThechambercontainedanAnaeroGensachet (ThermoScientific) for

thegenerationofanaerobicconditions.

Forsamplesthatshowednocoloniesontheagarplateswithin72hours,anewroundof

cultivation was performed with either non-diluted or concentrated bacterial flower

samples. For concentrated bacterial flower samples, 500 µl of the sample was

centrifugedonmaximumspeedfor1minuteinatablecentrifugebeforethesupernatant

wasdiscarded.Thecellpelletwasre-suspendedin200µl0.9%NaClbeforeplatedout

onnewMRSagarplatesandincubatedanaerobically.

Forpurecultureisolation,40colonieswerepickedwithinarandomandlimitedareaon

the agar plate, and streaked on newMRS agar plates and incubated anaerobically as

described above. Samples that got fewer colonies than 40were left at RT for aerobic

growthfor48hoursinanattempttoyieldmoreisolates.

Culture tubes containing 6 ml MRS broth supplemented with 40 % fructose were

inoculated with single colonies from the pure cultures using sterile toothpicks. The

culture tubes were then incubated at 30 °C for 24 to 72 hours for the isolation of

plasmidsdescribedinsection2.3.

For storing samples, glycerol stocks were made from pure cultures grown in MRS +

fructosebyadding400µl45%Glycerolto800µlbacterialcultureina2mlmicrotube.

The stockswere stored at -80 °C until further use. See Appendix 1 for description of

preparationofMRS+fructose.

2.3 PlasmidProfiling

Plasmidprofileanalysishasbeenappliedinclinicalstudiestoinvestigatethespreadof

antibiotic resistance, where patterns of plasmids are compared between disease

8

outbreak strains and non-outbreak control strains (28). Plasmids are first separated

according to size by agarose gel electrophoresis, and then cleaved by restriction

endonucleases generating multiple fragments that make interpretation of strain

relatednessmorefeasible(29).

2.3.1 IsolationofPlasmids

PlasmidswereisolatedusingtheE.Z.N.A.PlasmidDNAMiniKitI(Omega)accordingto

theprotocolprovidedby themanufacturerwithminormodificationsdescribed in the

following.

PureculturesgrowninMRS+fructoseovernight(ON)werecentrifugedat4000rpmfor

5minutesbeforethepelletwasre-suspendedin500µlTE(Tris-ClEDTA)-buffer.The

suspension was transferred into a 1.5 ml micro centrifuge tube followed by another

centrifugationstepatmaximumspeedfor1minuteinatablecentrifuge.SolutionIwas

addedtogetherwith5µl lysozyme(40mg/ml)and5µlmutanolysine(1000U/ml) to

weaken the cell walls, and incubated on water bath at 37 °C for 10 minutes before

resumingtotheprotocol.BoundDNAwaselutedtwicewith30µlElutionBufferandthe

DNAsampleswerekeptat-20°Cuntilphylogeneticidentification(section2.4).Astep-

by-stepprotocolisprovidedinAppendix3.

2.3.2 AgaroseGelElectrophoresis

AgarosegelelectrophoresiswasusedfordetectionandseparationofDNAaccordingto

molecularsizeatvariousstagesthroughoutthestudy.

To prepare two 1% agarose gels, amixture of 1 g agarose and 100ml 1XTAE (Tris-

acetate-EDTA)bufferwasheatedinamicrowaveovenuntiltheagarosehaddissolved.

The agarose solution was cooled down to below 60 °C before 4 µl Peq-Green (VWR

Peqlab)wasaddedasafluorescentdyetomaketheDNAvisiblewhenexaminedunder

UV-light.Thesolutionwasmixedbygentlyswirlingtheflask.Thesolutionwaspoured

evenlyintotwomoldingtraysthatheldcombsfortheformationofwellsandallowedto

cool until the gels solidified. The solid gels were placed in electrophoresis chambers

(BioRad)andthechamberswerefilledwith1XTAEBuffer.

Small volumes, generally 5 or 15 µl, of the samplesweremixedwith the appropriate

amountof6XLoadingBuffercontainingavisibledyeand loaded intothewellsonthe

9

gel.Fiveor10µl1kbLadderfromNewEnglandBiolabswereprovidedasamolecular

sizemarkeronbothsidesofthegel.Thegelswerethenrunat75Vuntiltheloadingdye

hadtraveledabout2/3ofthelengthofthegel.DNAbandsonthegelswerevisualizedby

UV-lightusingaGelDocimagerfromBioRad.

2.3.3 RestrictionEndonucleaseCutting

Restriction endonucleases are enzymes that recognize specific nucleotide sequences

known as restriction sites and cleave the target DNA at defined positions at or near

theserestrictionsites.Tomakeplasmidprofiles,restrictionenzymeswereusedtocut

plasmidsintosmallerfragmentsthatwouldmakeauniquebandpatternwhenrunonan

agarosegel.

Duringthepilotstudy,twoenzymes,XhoIandSpeI,withdifferentrecognitionsiteswere

testedfortheirabilitytocuttheisolatedplasmids.Theyweretestedbothtogetherand

separately.BecauseXhoIperformedpoorer thanSpeI, adecisionwasmade to replace

XhoI with XbaI. This decision was based on the difference in G+C content of the

recognitionsite.LacticacidbacteriahavealowG+CcontentandXhoIhasahigherG+C

contentthanbothSpeIandXbaI,seeAppendix1;TableA1.1.

Reactionmixtureswereprepared,whileonice,accordingtoTable2.2.Thecomponents

weremixedgentlybeforeashortcentrifugationtogetallthecomponentsatthebottom

ofthetube,andthenincubatedat37°Cfor2h.Afterincubation,thesampleswererun

ona1%agarosegelaccordingtotheproceduredescribedinsection2.3.2.

Table 2.2 Reaction setup for endonuclease cutting showing the components with their correspondingvolumesandfinalconcentrationsfora20µlreaction

Component Volume[µl] Finalconcentration10XCutSmartBuffer 2.0 1X20000units/mlXbaI 0.5 10units/20µlreaction10000units/mlSpeI 0.5 5units/20µlreactiondH2O 7.0 PlasmidDNA 10

10

2.4 PhylogeneticIdentification

2.4.1 PolymeraseChainReaction(PCR)Forphylogeneticidentification,16SrRNAgeneamplificationbyPCRwasperformedon

a selection of the collected bacterial strains using the universal primers 11F: (5´-TAA

CACATGCAAGTCGAACG-3´)and4R:(5´-ACGGGCGGTGTGTRC-3´).TheDNAsamples

fromthe isolationofplasmids(section2.3)wereusedasDNAtemplates.Thereaction

setup shown in Table 2.3 and the thermocycling conditions shown in Table 2.4were

applied.

Table 2.3 Reaction setup for PCR showing the components with their corresponding volumes and finalconcentrationsfora50µlreaction.

Component Volume[µl] FinalConcentration20µMForwardPrimer 1 0.4µM20µMReversePrimer 1 0.4µM5XPCRStandardReactionBuffer 10 1X10mMdNTPmix 1 0.2mM5000units/mlTaqDNAPolymerase 0.3 1.5units/50µlreactiondH2O 35.7 TemplateDNA 1 Table2.4Thermo-cyclingconditionsofPCRfor16S

Steps Temperature[°C] Time[s] Cycles DescriptionStep1 95 30 1 InitialdenaturingStep2

95 15 30

Denaturing54 60 Primerannealing72 90 Primerextension

Step3 72 300 1 Completion

Inorder todetect thepresenceofamplicons,5µlof thePCRproductofeverysample

were runon a1%agarose gel according to theproceduredescribed in section2.3.2.

SamplescontainingampliconswerepurifiedusingtheNucleoSpin®GelandPCRClean

up kit (Macherey-Nagel GmbH, Düren, Germany) according to the manufacturers

instructions, and the DNA concentration was determined using a NanoDrop 2000

SpectrophotometerfromThermoScientific.

11

2.4.2 SequencingandProcessingofData

Purified PCR products were, if necessary, diluted with sterile dH2O to the final DNA

concentrationof20to80ng/µl.Volumesof5µlwerethenaddedtoa96-wellskirted

traycontaining5µl5µMprimer11Fineachwell.SequencingwasperformedbyGATC

Biotech(Constance,Germany).

ThegeneratedsequenceswereeditedusingBioEditv7.2.5SequenceAlignmentEditor

(Ibis Biosciences, Carlsbad, CA, USA). Blast analysis was performed on the edited

sequencesandalignedusingthefreelyavailablemultiplesequencealignmentprogram

MAFFT. The phylogenetic tree constructed by MAFFT was used to pick isolates for

fermentationprofilingbasedontheirclusteringinordertogetarandomsampling.

2.5 SugarFermentationProfiling

Twenty isolateswerechosenforcarbohydrate fermentationprofilingusingtheAPI50

CHL identification system from BioMèrieux, Inc. The API 50 CH strip consists of 50

microtubes that containonenegative control and49different substratesbelonging to

thecarbohydratefamilyanditsderivatives.TheAPI50CHstripisusedinconjunction

withAPI50CHLmediumthatrehydratesthesubstratesandthatcontainsapHindicator

thatdetectstheproductionofacidswhenthesubstrateisfermented.

Steriletoothpickswereusedtoscrapeasmallamountoffofthetopofthefreezingstock

ofeachisolateanddroppedintoculturetubeswith5mlMRSmedium.Theisolateswere

culturedONand1mloftheON-culturewascentrifugedfor5minutesat13000rpm.The

supernatantwasdiscardedandthepelletsurfacewaswashedwithasmallamountof50

CHLmedium.Thecellpelletwasre-suspendedin2mlof50CHLmediumbefore300µl

of thesuspensionwastransferredto6ml50CHLmediumandmixedthoroughly.The

APIcarbohydratefermentationstripswerepreparedbyplacingthemintheincubation

trayandthebacterialsuspensionwasdistributedintothe50tubes.Mineraloilwasused

tofillthecupuletocreateanaerobicconditionsandtheincubationtraywasplacedina

plastic box with moistened paper and incubated at 30 °C. Color changes due to the

productionofacidsinthecupuleswererecordedduringthecourseofoneweek.

12

2.6 Cycloheximide

Cycloheximide (CHX) is a eukaryotic translation inhibitor and is the most common

laboratory reagent used to inhibit protein synthesis. CHX was therefore used to

determinewhethersomeofthecoloniesgrownaerobicallywereyeasts.

Twoisolatesof inall23thatweresuspectedtobeyeastwere inoculated in5mlMRS

brothON.OneisolatethatwasproventobeLactobacilluskunkeeiwasalsoinoculatedin

5mlMRSbrothtofunctionasacontrol.

To make a CHX stock solution, 0.157 g CHX was dissolved in 6 mL sterile dH2O by

pipettingthesuspensionwhilstimmersedinhotwater.Thesolutionwasthenfiltrated

throughFiltropurS0.2syringe(Sarstedt)filterbytheuseofa10mlsyringe.

To achieve a working concentration of approximately 5 mg/ml, 200 µl of the stock

solutionwasdilutedin800µlsteriledH2O.

Fiveculturetubeseachcontaining5mlMRSbrothwerepreparedforeachisolate,and

the working solution of CHX was added in various volumes to achieve various final

concentrations,seeAppendix1;TableA1.2.

The culture tubes were then thoroughly mixed before 50 µl of ON-cultures of each

isolatewereaddedtotheirrespectivetubes.Theculturetubeswereagainmixedbefore

incubatedat30°CandcelldensitywasmeasuredonaUltraspec10CellDensityMeter

fromAmershamBiosciencesafter24hours.

3. Results

3.1 GrowingofBacterialStrainsandPlasmidProfiling

TypicalcoloniesfromalldifferentflowersamplesgrownonMRSagarplateswerewhite,

smoothandvaried little in size.Appleanddandelionsamplesgrownaerobicallywere

bigger insizeandhadroughcolonyedges.Thedurationof incubationvariedbetween

thedifferentbacterialflowersamples,andMRSagarplateswereleftatRTinanattempt

toyieldmoreisolatesfrombothappleanddandelionsamples.Inaddition,isolatesfrom

appleproducedaredpigmentaroundtheedgesofcoloniesontheperiphery.Lateron,

these isolates were confirmed to be yeast and are therefore excluded from further

estimationsdoneonthetotalsetofisolates.

13

3.2 PlasmidProfiling

Onehundredandseventy-sevenisolateswereisolatedfromflowers,40fromeachplant

species, except for apple samples with a yield of 17 isolates only. All isolates were

assessedforplasmidcontent.Ofthe177isolates,37hadplasmids,whichaccounts for

approximately21%.

Themajorityoftheisolatesthatcontainedplasmidswerecollectedfromraspberry.The

resultsfromraspberrysamplesarerepresentedinFigure3.1containingpicturesofall

theagarosegelsthatwererun.Somelaneshaddistinctivebandsandotherswhereonly

smearsofDNA.Differentbandintensitiesmadeitdifficulttoobtainasingleimageofall

thebandsonthegel.Lane11containsabandatapproximately11kbwhichinFigure

3.1 (B) appears as a smear. For the remaining plant species, samples that contained

plasmidswere collectedand representedona single agarosegel shown inFigure3.2.

Thisisbecausemanyofthegelslackedplasmidbandsandsomeonlyhad-atmost-two

plasmids.

Noplasmidswereobservedforisolatesisolatedfromrapeseed,whilethetotalnumber

of plasmid containing isolates fromdandelion, apple andwillowherbwere 4, 2 and 5

respectively(Figure3.2).Closeexaminationofthebandpatternsrevealed14different

profiles, where one profile, consisting of a single band of approximately 3.7 kb, was

observedfor10oftheisolates.

14

BA1kb123456789101kb

1kb111213141516171819201kb

1kb212223242526272829301kb D

1kb313233343536373839401kb

1kb4546515368721631641691721881kb

C

Figure3.1Plasmidprofilesofbacteriaisolatedfromraspberry.Theplasmidprofilesof40 isolates from raspberry are shown from (A) numbers 1-10, (B) numbers 11-20, (C)numbers21-30,and(D)numbers31-40.

Figure3.2Plasmidprofilesfromisolatesgatheredfromdandelion(numbers45,46,51and53),apple(68and72),andwillowherb(163,164,169,172and188).

15

3.3 PhylogeneticIdentification

Approximately 20 isolates from each flowerwere chosen for further identification by

16S rRNA gene sequencing. Some strainswere deliberately chosen tomake sure that

everyplasmidprofilewasrepresented. Intotal,eighty-nine isolateswere identifiedby

BLASTanalysisofsequencesofapproximately1100bp.

Fromdandelion,10oftheisolateswereidentifiedasL.kunkeeiwhiletheother10ofthe

isolates, that were grown aerobically, had the closest percent identity to the

Enterobacteriaceae family. All isolates from rapeseed were identified as Enterococcus

spp.withhigh sequence similarities (99%) tobothEnterococcushaemoperoxidusand

Enterococcusplantarum.Fromapple,2of the isolateswere identifiesasL.kunkeei (18

%;2of11isolates)while9isolateswereF.fructosus(82%;9/11).Thisspecieswasalso

found in raspberry,which represented78%(14/18)while the remaining22%were

Lactococcus lactis (4/18). Samples of willowherb revealed the most diverse bacterial

composition,with6isolatesidentifiedasL.lactis,2asLactobacillussakei(10%;2/20),

6 as Lactococcus garvieae (30 %; 6/20), 5 asWeissella ceti (25 %; 5/20) and 1 as

Weissellaviridescens (5%;1/20).Thedistributionof the identified isolatesamongthe

differentflowersisillustratedinFigure3.3.

Figure3.3Distributionofallidentifiedisolatesamongvariousflowers.IsolatesfromdandelionwereidentifiedasL.kunkeeiandEnterobacteriaceae,rapeseedisolateswereEnterococcusspp.,appleisolateswereF.fructosusandL.kunkeei,raspberryisolateswereF.fructosusandL.lactis,andisolatesfromwillowherbwereidentidiedasL.lactis,L.sakei,L.garvieae,W.cetiandW.viridenscens.

0

5

10

15

20

25

Dandelion Rapeseed Apple Raspberry Willowherb

Num

berofisolates

Enterobacteriaceae

W.viridenscens

W.ceti

L.garvieae

L.sakei

L.kunkeei

L.lactis

F.fructosus

Enterococcusspp.

16

3.4 SugarFermentation

Carbohydrate fermentation reactions were recorded for representatives of every

species isolated using API50 CHL galleries (Table 3.1). In total, 19 isolates produced

acidsfrom29ofthe49carbohydratestested.TheresultsshowthatallstrainsexceptW.

ceti fermentedD-fructoseafter1day,andD-glucosewithin7days.F.fructosus strains

fermented only 3 sugarswithin oneweek,while strains ofL. lactis,L.garvieae andL.

sakeifermentedupto20sugarseach.Threestrainsidentifiedasmembersofthefamily

Enterobacteriaceae were also tested for carbohydrate fermentation using the API 50

CHL system, but gave only erratic results that are neither shownnor evaluatedmore

closely.Table3.1Carbohydratefermentationprofileoflacticacidbacteriaisolatedfromflowers.

Carbohydratesa 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16bGlycerol - - - - - - - - i i i - - - - -L-arabinose - - - - - - - - - - - - - - - 1dD-ribose - - - 1d 1d - - - 3d 3d 3d 1d 1d i w 1dD-xylose - - - 1d 1d - - - - - - - - - - 1dD-galactose - - - 1d 1d - - - i i i 2d 2d - - 1dD-glucose + 3d + 1d 1d + + + 1d 1d 1d 1d 1d i 1d 1dD-fructose 1d 1d 1d 1d 1d 1d 1d 1d 1d 1d 1d 1d 1d - 1d 1dD-mannose - - - 1d 1d - - - 1d 1d 1d 1d 1d - 2d 1dL-rhaminose - - - - - - - - - - - - - - - iD-mannitol + + + 1d 1d i + + - - - 2d 2d - - -N-acetylglucosamine - - - 1d 1d - - - 1d 1d 1d 1d 1d i 2d 1dAmygdallin - - - 2d 3d - - - i i i 1d 1d - - 2dArbutin - - - 1d 1d - - - 1d 1d 1d 1d 1d - - 2dEsculinferriccitrate - - - 1d 1d - - - 1d 1d 1d 1d 1d - 1d 1dSalicin - - - 1d 1d - - - 1d 1d 1d 1d 1d - w 2dD-cellobiose - - - 1d 1d - - - 1d 1d 1d 1d 1d - - 1dD-maltose - - - 1d 1d - - - + + + 2d 2d i 1d -D-lactose - - - 2d 3d - - - i i i - - - - 2dD-melibiose - - - - - - - - - - - - - - - 1dD-saccharose(sucrose) - - - - 1d + + + - - - - - - w 1dD-trehalose - - - 1d 1d + + + w w w 1d 1d - 1d 1dAmidon(starch) - - - + + - - - - - - - - - - -Gentiobiose - - - 1d 1d - - - 3d 3d 3d 1d 1d - - 2dD-turanose - - - - - - - - - - - - - - w -D-lyxose - - - - - - - - - - - - - w - iD-tagatose - - - - - - - - - - - 1d 1d - - -Potassiumgluconate - - - i i i i i - - - w w w i w”d”daysuntilpositivereaction,”+”positivereactionwithin4to7days,”i”intermediate,”w”weakreactionand”-”negative(nocolorchange).aCarbohydratesnotlistedwerenotfermentedbyanyofthestrains.bNumbers:1-3refertoF.fructosusstrainsB23,B29andB69;4-5,L. lactis strainsB37andB169;6-8,L.kunkeei strainsB45,B53andB72;9-11,Enterococcus sp. strainsB125,B139andB149;12-13,L.garvieaestrainsB167andB187;14,W.cetistrainB177;15,W.viridescensstrainB183;16,L.sakeistrainB188.

3.5 Cycloheximide

Fromthecultivationofbacterialstrainsfromapplesamples,therewereonly17colony-

formingunits on theMRSagarplates. In an attempt to yieldmore isolates, theplates

were left at room temperature for aerobic growth for three days. The new colonies

17

formed were sub-cultured for pure culture isolation and underwent all the same

experiments as the other isolates. Plasmid isolation and PCR amplification of the 16S

rRNA gene revealed neither plasmids nor amplicons. To address the theory of these

isolatesnotbeingprokaryote,anexperimentwithaeukaryotetranslationinhibitorwas

conducted(section2.8).

The results showed no growth in culture tubes containing CHX of concentrations

ranging from 5.2 to 156 µg/ml. There was, however, growth in the control tubes

containing the bacterium L. kunkeei. Microscopy also revealed that the cells were

oval/egg-shaped.Thisstronglysuggeststhattheisolateswereyeast.

When grown on MRS agar plates, the yeast produced a sweet-like odor that could

resemble the smell of apple cider. This prompted us to investigate if the pigment

productionandsmellweremediumdependent.Two isolatesdenoted47and48were

grownon4differentgrowthmedia(MRS,LA,GM17andBHI).MRSandLAwerethetwo

media that gave off the strongest smell and were therefore analyzed with gas

chromatography,alongwithblanksamplesofbothmediatocorrectforfalsepositives.

TheanalysiswasperformedbyKariOlsenatNMBU.

Many different alcohols together with some esters were detected; see Appendix 2;

FigureA2.1andA2.2.Whataffectthesecompoundscouldhaveonthefloralnectar,or

eveniftheyareproducedbytheyeastwhileinhabitingthenectar,isunknownandneed

moreinvestigation.Itcouldbeinterestingtoseeifanyofthecompoundsproducedina

significant amount could have any effect on the behavior of the honeybee. A deep

literatureresearchandpossiblyalearningexperimentconductedonhoneybeeswould

betorecommend.

18

4. DiscussionLABwere isolated fromfresh flowersof fivedifferentplantspecies.The isolateswere

characterized using plasmid profiling, 16S rDNA sequencing and fermentation profile.

Two FLAB species, L. kunkeei and F. fructosus, were identified in samples from

dandelion, apple and raspberry. Rapeseed harbored only Enterococcus spp. and

willowherbhadthemostdiverseLABfloraconsistingofL.garvieae,L.sakei,W.cetiand

W.viridescens.TheLABwerecomparedwithLABisolatedfromhoneybeegutsampled

inparallelwiththeflowersbloomingperiod.ComparisonoftheLABstrainsandplasmid

profilessuggestapositivecorrelationbetweentheLABmicrobialflorainhoneybeegut

andtheflowerswithintheirforagingarea.

4.1 LacticAcidBacteriaIsolatedfromFlowers

The results showed that the LAB flora of flowers from different plant species varied

significantly between some plant species and less between others. Dandelion, apple

flowerandraspberryharboredthefructophilicLABspeciesL.kunkeeiandF.fructosus,

which have in previous studies also been isolated from flowers, and are known to

inhabit fructose rich niches. Rapeseed harbored only Enterococcus spp. while

willowherbhadthemostdiverseLABfloraconsistingofL.garvieae,L.sakei,W.cetiand

W.viridescens.

L. kunkeei and F. fructosus have previously been isolated from several flowers, while

Enterococcusspp.arereadilyfoundintheenvironmentandarecommensalmembersof

gutcommunitiesinmammalsandbirds(30).TheremainingLABspecies,isolatedfrom

willowherb,haveasfarasweknownotbeenisolatedfromfreshflowersbefore.These

LAB species have for the most part been detected in processed meat, fish or dairy

products(31).L.garvieaehasalsobeenidentifiedinvegetablesprouts(32),andbothL.

garvieaeandW.cetihavebeenshowntobefishpathogens(33,34).However,arelative

ofW. ceti, W. confusus, has recently been isolated from the gut of honeybees (35).

Obviously,thereisalotmoretolearnaboutthemicrofloraofbeesandflowers.

Nectarcompositionisgreatlydiverseamongdifferentplantspeciesasmeanstoattract

different pollinators. Although every plant species collected are of importance for the

19

honeybee pollinator, indicating a similar nectar composition in general, the nectar

compositionsmay be different enough to cause the different LAB flora observed. For

example, nectar fromwillowherb and rapeseedmay not contain enough fructose for

FLAB growth. One the other hand, the results may also indicate that the bacterial

extractionmethodusedonlyextracts themostdominantbacterialspecies.Wedidnot

discriminateagainstdifferentpartsoftheflower,exceptforthestemandleaves,which

maybeconsideredasa source for thevariationofLABspeciesobservedbetween the

plantspecies,wheresomeLABmayoriginatefromeitherpetals,pollenornectar.

Thefermentationprofilescorrespondedreasonablywellwiththeliterature,andstrains

ofthesamespeciealsoshowedthesameprofile,withL.lactisastheonlyexception.

Oneof the twostrainsofL. lactis fermented sucrosewhile theotherdidnot.The two

strains of L. lactis were isolated from different flowers and had different plasmid

profiles.Thisisbecausethemetabolismofsucroseiscontrolledbyplasmidgenesandis

thereforestraindependent(36).

4.2 ComparingtheResultstoLacticAcidBacteriaIsolatedfromHoneybeeGut

TocomparehowtheLAB isolated from flowersare related toLAB found inhoneybee

gut, theresults fromthepresentstudywerecomparedto theresultsofaparallelMSc

thesisconductedbyIngvildGallefoss(37)(REF),characterizingthegutmicrobiotaofthe

honeybee,andwithsimilarmethodsandprocedures.Bacterialsamplesfromhoneybee

gut were provided by the LMG group at the Norwegian University of Life Sciences.

Honeybeeshadbeencollectedfromwithinthehiveatthesametimepoints for flower

sample collections, and each bacterial sample contained about 10 honeybee guts. All

resultsmentionedaboutLABisolatedfromhoneybeegutsampleswerekindlyprovided

by I. Gallefoss. The flowers collected represent amain pollination target at any given

timepoint.

4.2.1 ComparisonoftheLABSpecies

AnoverallcomparisonoftheLABspecies,regardlessofseasonalprogression,revealed

that45%of all isolates fromhoneybeegutwereL.kunkeeicompared to15%of the

isolatesfromflowersamples.Bothflowersamplesandhoneybeegutsamplescontained

20

about29%F.fructosus. Bifidobacteriumasteroidesaccountedfor25%ofallidentified

LABstrainsinhoneybeesamples,asdidEnterococcusspp.inflowersamples.

Withfocusontheprogressingseasonandthefollowingchangeofpollinationtargets,a

comparisonoftheLABisolatesfromflowersandhoneybeegutinrelationtotheirtime

of collection is shown in Figure 4.1. Some correspondence can be seen between the

samplescollectedinmiddleofMay,lateMayandlateJune.

In middle of May, the main flowering period of dandelion, L. kunkeei was the only

species found in both flower and honeybee samples. In lateMay, themain flowering

period for fruit flowers and autumn rapeseed, this species was still dominant in the

honeybeegutsample,whileF.fructosuswasdominantintheapplesample.

Figure4.1Thebacterial compositionof flower samples compared to samplesof thehoneybeeguton theirrespective time points of collection.Data of bacterial strains isolated fromhoneybee gut samples (C3T6, C3T7,C3T9, C3T12 and C3T13) were kindly provided by Gallefoss 2016 (REF). The relative abundance represents thenumberofisolatesnormalizedto100%ofLABspeciesanalyzed.*Numberofisolates.

In late June, almost 80% of the identified isolates from raspberry were F. fructosus,

while thiswas theonlyspecies identified in thehoneybeesample.Seemingly, theLAB

compositioninthehoneybeegutshiftsfromL.kunkeeitoF.fructosusbetweenmiddleof

May and late June when comparing C3T6, C3T7 and C3T9. One sample from August

(C3T12)alsocontainedalmostonlyF.fructosus,butthiscorrespondedpoorlywiththe

willowherb sample and the other honeybee gut sample (C3T13) collected in August.

0%10%20%30%40%50%60%70%80%90%100%

Dandelion

C3T6

Apple

Rapeseed

C3T7

Raspberry

C3T9

Willowherb

C3T12

C3T13

MiddleofMay LateMay LateJune August

Relativeabundance L.mesenteroides

B.asteroides

W.viridenscens

W.ceti

L.garvieae

L.sakei

L.kunkeei

L.lactis

F.fructosus

Enterococcus

1022112020187202023*

21

Actually,noneoftheidentifiedbacteriaonneitherrapeseednorwillowherbwerefound

in anyof the samples fromhoneybeegut. Likewise,noB.asteroideswereobserved in

anyoftheflowersamples.L.lactisandtheotherLABfoundinflowersamplesthatwere

absentinthehoneybeegut,hasasfarasweknownotbeenidentifiedinotherstudies

conductedonthehoneybeegut.

4.2.2 ComparisonofthePlasmidProfiles

In order to evaluate the relatedness between the LAB isolated from flowers and

honeybee guts, the plasmid profiles were compared. The plasmid profiles from

honeybeegutsamplesshowedgreaterdiversitythantheprofilesfromflowerssamples.

Atotalof36differentprofilesdistributedamong140isolates,whichaccountsforabout

74 % of the total number isolated bacteria from honeybee gut, were observed. In

contrast, only21%of the isolates from flowers containedplasmidswith14different

profiles.

L.kunkeeihavepreviouslybeenshowntocomprisemanyplasmidswithhighdiversity

among different strains (38). The same resultswere shown for theL.kunkeei strains

isolatedfromC3T6andC3T7.Eighteenofthetotal34differentprofileswereassignedto

L.kunkeei,andalmosteveryprofilewasobserved formore thanone isolate. In flower

samples, L. kunkeei was identified for 5 different plasmid profiles where no strains

contained the same profile. Although the diversity seemed greater for the plasmid

profiles of honeybee samples, the abundance of this species were also greater in

honeybeesamples,whichmaypartlyexplainthediversitydifference.

F.fructosusshowedsimilardiversityintheflowersamplestothehoneybeegutsamples.

SevenplasmidprofileswereidentifiedasF.fructosusinthehoneybeesamples,whileF.

fructosusstrainsfromflowersamplesshowed6differentprofiles.

Fouroftheplasmidprofilesfromflowersamplesgreatlyresembledthebandpatternof

fourprofilesobtainedfromthehoneybeesamples.Forthreeoftheseprofiles,assigned

L.kunkeei,therewashoweveradiscrepancybetweenthetimepointsofsamplingofthe

isolates that contained the same profile. One profile sampled from apple inMaywas

seen in an isolate from honeybee gut sampled two weeks earlier, and vise versa. A

possible explanation for thismay be that the blooming period of dandelion and fruit

trees,likeapple,arearoundthesametime.Individualhoneybees,orbeesfromthesame

22

hive,couldthereforehavevisitedbothflowersbeforetheygotsampled.Themajorityof

theF.fructosusstrainsfrombothflowersandhoneybeegutsamplescontainedaprofile

consisting of only one band of approximately 3.7 kb, indicating a prevalence of this

straininbothflowerandhoneybeegutsamples.

4.3 BacterialTransmissionbetweenFlowerandHoneybees

The rapeseed samplewas dominated by another flora than the other flowers, and no

enterococci were observed in the honeybee samples. Honeybees may prefer apple

flowerstorapeseedinthisperiod,orevendandelion,asthebloomperiodforallthree

flowersoverlap.Thefactthattherapeseedfieldwheretheflowerswerecollectedfrom

islocatedmorethan4kilometersfromNMBUcampusalmostcertainlymeansthatthe

honeybeeslocatedoncampusdidnotvisittheseflowers.Thereisapossibilitythatthere

are other rapeseed fields closer to the hive, but these flowers would have to be

examinedforLABfloraaswellbeforesomeconclusionscanbedrawnconcerningthese

samplingresults.

Due to the seemingly rapid loss of L. kunkeei in the honeybee gut, there are strong

indicationstowardsafloraltransmission.However,thereisnoclearevidenceforwhich

way the transmissionoccur.This is inagreementwith thehypothesisof severalother

studies.Ushioetal.(39)demonstratedthatthemicrobialcommunitycompositionona

flower surface changed after contact with an insect, and they suggested that the

microbesaretransferredfromtheinsecttothefloweralthoughtheydidnotexcludethe

possibility of the other way around. Aizenberg-Gershtein et al. (40) compared the

microbialcommunity inhoneybeegutwithbothcoveredanduncoveredflowers.They

found that the microbial flora in the honeybee gut differed more from the covered

flower than the uncovered flower, supporting the hypothesis that honeybees may

introducebacteriatotheflower.

The honeybees were collected from within the hive, meaning that they did not

discriminate against honeybee age and development stage. As the honeybee gut has

been shown to vary with age and also gut compartments, any discrepancies in the

honeybeesamplescanalsobe influencedby these factors.Onetheotherhand,even if

theindividualhoneybeeforagertendstosticktoonetypeoffloweratatime,bacterial

23

flora from different flowers can be transferred back to the hive by many different

foragers,makingthefloramorediversewithinthehivethanonaparticularflower.This

may contribute to the greater variety of plasmid profiles seen in the honeybee gut

samples compared to the flower samples. Possible contamination/transmission from

thesurroundingenvironmentisquiteobvious,butthecontaminationpressurefromthe

in-hive environment complicates interpretation of the results. The data sets are

relatively small to give any hard evidence, but there are many indications toward a

positive correlationbetween theLABmicrobial flora inhoneybeegut and the flowers

withintheirforagingarea.

Further work is advised with a closer examination of the LAB flora from the plant

species collected in the study, where parallel samplings possibly could discard bias

relatedtomethodologicalorpersonalerrors.Samplingofhoneybeeswhiletheyvisitthe

flowercouldperhapsbypassagerelatedinfluenceincontrasttoin-hivesampling.

4.4 Conclusion

Flowers provide diverse habitats for microorganisms, and as shown in the present

study,theLABcompositioninfivedifferentplantspeciesvariedamongthem.L.kunkeei

andF.fructosuswerethemostabundantofalltheidentifiedspeciesandisincorrelation

withotherstudiesonLABfoundinflowers.TheLABfloraofthehoneybeegutseemsto

shift fromL.kunkeei toF.fructosus throughMayto late June,andF.fructosuswasalso

foundasthemostabundantLABinoneofthesamplescollectedfromhoneybeegutsin

August.This is inaccordancewiththeLABflora foundondandelionandapple inMay

andraspberryinlateJune.Theseresultsagreeswellwithearlierfindingofanumerical

variation of LAB in the honeybee crop across seasons (5) and a seasonal shift of L.

kunkeeiinthegutofsummerbeesandwinterbees(41)wheretheysuggestedthatfloral

transmissioncouldexplainthesporadicfindingsofthisspecie.

24

5. ReferencesReferanser

1. GuinaneCM,CotterPD.2013.Roleofthegutmicrobiotainhealthandchronic

gastrointestinaldisease:understandingahiddenmetabolicorgan.TherapAdvGastroenterol6:295-308.

2. BromenshenkJJ,HendersonCB,WickCH,StanfordMF,ZulichAW,JabbourRE,DeshpandeSV,McCubbinPE,SeccombRA,WelchPM,WilliamsT,FirthDR,SkowronskiE,LehmannMM,BilimoriaSL,GressJ,WannerKW,CramerRA,Jr.2010.Iridovirusandmicrosporidianlinkedtohoneybeecolonydecline.PLoSOne5:e13181.

3. MartinsonVG,MoyJ,MoranNA.2012.Establishmentofcharacteristicgutbacteriaduringdevelopmentofthehoneybeeworker.ApplEnvironMicrobiol78:2830-2840.

4. MartinsonVG,DanforthBN,MinckleyRL,RueppellO,TingekS,MoranNA.2011.Asimpleanddistinctivemicrobiotaassociatedwithhoneybeesandbumblebees.MolEcol20:619-628.

5. OlofssonTC,VasquezA.2008.DetectionandidentificationofanovellacticacidbacterialflorawithinthehoneystomachofthehoneybeeApismellifera.CurrMicrobiol57:356-363.

6. OlofssonTC,AlsterfjordM,NilsonB,ButlerE,VasquezA.2014.Lactobacillusapinorumsp.nov.,Lactobacillusmellifersp.nov.,Lactobacillusmellissp.nov.,Lactobacillusmelliventrissp.nov.,Lactobacilluskimbladiisp.nov.,Lactobacillushelsingborgensissp.nov.andLactobacilluskullabergensissp.nov.,isolatedfromthehoneystomachofthehoneybeeApismellifera.IntJSystEvolMicrobiol64:3109-3119.

7. VásquezA,OlofssonTC,SammataroD.2008.AscientificnoteonthelacticacidbacterialflorainhoneybeesintheUSA–AcomparisonwithbeesfromSweden.Apidologie40:26-28.

8. VasquezA,ForsgrenE,FriesI,PaxtonRJ,FlabergE,SzekelyL,OlofssonTC.2012.Symbiontsasmajormodulatorsofinsecthealth:lacticacidbacteriaandhoneybees.PLoSOne7:e33188.

9. ForsgrenE,OlofssonTC,VásquezA,FriesI.2009.NovellacticacidbacteriainhibitingPaenibacilluslarvaeinhoneybeelarvae.Apidologie41:99-108.

10. OlofssonTC,ButlerE,MarkowiczP,LindholmC,LarssonL,VasquezA.2014.Lacticacidbacterialsymbiontsinhoneybees-anunknownkeytohoney'santimicrobialandtherapeuticactivities.IntWoundJdoi:10.1111/iwj.12345.

11. HayekSA,IbrahimSA.2013.CurrentLimitationsandChallengeswithLacticAcidBacteria:AReview.FoodandNutritionSciences04:73-87.

12. ServinAL.2004.Antagonisticactivitiesoflactobacilliandbifidobacteriaagainstmicrobialpathogens.FEMSMicrobiolRev28:405-440.

13. SunZ,YuJ,DanT,ZhangW,ZhangH.2014.PhylogenesisandEvolutionofLacticAcidBacteria.doi:10.1007/978-94-017-8841-0_1:1-101.

14. LiuW,PangH,ZhangH,CaiY.2014.BiodiversityofLacticAcidBacteria.doi:10.1007/978-94-017-8841-0_2:103-203.

15. EndoA,Futagawa-EndoY,DicksLM.2009.Isolationandcharacterizationoffructophiliclacticacidbacteriafromfructose-richniches.SystApplMicrobiol32:593-600.

25

16. NevelingDP,EndoA,DicksLM.2012.FructophilicLactobacilluskunkeeiandLactobacillusbrevisisolatedfromfreshflowers,beesandbee-hives.CurrMicrobiol65:507-515.

17. McFrederickQS,WcisloWT,TaylorDR,IshakHD,DowdSE,MuellerUG.2012.Environmentorkin:whencedobeesobtainacidophilicbacteria?MolEcol21:1754-1768.

18. EndoA,SalminenS.2013.Honeybeesandbeehivesarerichsourcesforfructophiliclacticacidbacteria.SystApplMicrobiol36:444-448.

19. AndrewsJH,HarrisRF.2000.TheEcologyandBiogeographyofMicroorganismsonPlantSurfaces.AnnuRevPhytopathol38:145-180.

20. SasuMA,WallKL,StephensonAG.2010.AntimicrobialnectarinhibitsaflorallytransmittedpathogenofawildCucurbitapepo(Cucurbitaceae).AmJBot97:1025-1030.

21. FridmanS,IzhakiI,GerchmanY,HalpernM.2012.Bacterialcommunitiesinfloralnectar.EnvironMicrobiolRep4:97-104.

22. Alvarez-PerezS,HerreraCM,deVegaC.2012.Zooming-inonfloralnectar:afirstexplorationofnectar-associatedbacteriainwildplantcommunities.FEMSMicrobiolEcol80:591-602.

23. NicolsonSW.2011.Beefood:thechemistryandnutritionalvalueofnectar,pollenandmixturesofthetwo.doi:10.1080/15627020.2011.11407495.

24. GruterC,MooreH,FirminN,HelanteraH,RatnieksFL.2011.Flowerconstancyinhoneybeeworkers(Apismellifera)dependsonecologicallyrealisticrewards.JExpBiol214:1397-1402.

25. GoodAP,GauthierMP,VannetteRL,FukamiT.2014.Honeybeesavoidnectarcolonizedbythreebacterialspecies,butnotbyayeastspecies,isolatedfromthebeegut.PLoSOne9:e86494.

26. AndersonKE,SheehanTH,MottBM,MaesP,SnyderL,SchwanMR,WaltonA,JonesBM,Corby-HarrisV.2013.Microbialecologyofthehiveandpollinationlandscape:bacterialassociatesfromfloralnectar,thealimentarytractandstoredfoodofhoneybees(Apismellifera).PLoSOne8:e83125.

27. LudvigsenJ,RangbergA,AvershinaE,SekeljaM,KreibichC,AmdamG,RudiK.2015.ShiftsintheMidgut/PyloricMicrobiotaCompositionwithinaHoneyBeeApiarythroughoutaSeason.MicrobesEnviron30:235-244.

28. MayerLW.1988.Useofplasmidprofilesinepidemiologicsurveillanceofdiseaseoutbreaksandintracingthetransmissionofantibioticresistance.

29. SinghA,GoeringRV,SimjeeS,FoleySL,ZervosMJ.2006.Applicationofmoleculartechniquestothestudyofhospitalinfection.ClinMicrobiolRev19:512-530.

30. ByappanahalliMN,NeversMB,KorajkicA,StaleyZR,HarwoodVJ.2012.Enterococciintheenvironment.MicrobiolMolBiolRev76:685-706.

31. Champomier-VergesMC,ChaillouS,CornetM,ZagorecM.2001.Lactobacillussakei:recentdevelopmentsandfutureprospects.ResMicrobiol152:839-848.

32. KawanishiM,YoshidaT,KijimaM,YagyuK,NakaiT,OkadaS,EndoA,MurakamiM,SuzukiS,MoritaH.2007.CharacterizationofLactococcusgarvieaeisolatedfromradishandbroccolisproutsthatexhibitedaKG+phenotype,lackofvirulenceandabsenceofacapsule.LettApplMicrobiol44:481-487.

26

33. VendrellD,BalcazarJL,Ruiz-ZarzuelaI,deBlasI,GironesO,MuzquizJL.2006.Lactococcusgarvieaeinfish:areview.CompImmunolMicrobiolInfectDis29:177-198.

34. FigueiredoHC,CostaFA,LealCA,Carvalho-CastroGA,LeiteRC.2012.Weissellasp.outbreaksincommercialrainbowtrout(Oncorhynchusmykiss)farmsinBrazil.VetMicrobiol156:359-366.

35. RabadjievY,ChristovaP,IlievI,IvanovaI.2015.IdentificationofaLacticAcidBacterialflorawithinthehoneyintestinaltractofApismelliferafromdifferentregionsofBulgaria.

36. DubravkaSamaržijaNA,JasminaLukaHavranek.2001.Taxonomy,physiologyandgrowthofLactococcuslactis:areview.

37. GallefossI.2016.AComparativeAnalysisonPhylogeny,GeneticsandSelectedPhenotypesofLacticAcidBacteriaIsolatedfromGutMicrobiotaofHoneyBeeVersusFlowers.

38. RangbergA,DiepDB,RudiK,AmdamGV.2012.Paratransgenesis:anapproachtoimprovecolonyhealthandmolecularinsightinhoneybees(Apismellifera)?IntegrCompBiol52:89-99.

39. UshioM,YamasakiE,TakasuH,NaganoAJ,FujinagaS,HonjoMN,IkemotoM,SakaiS,KudohH.2015.Microbialcommunitiesonflowersurfacesactassignaturesofpollinatorvisitation.SciRep5:8695.

40. Aizenberg-GershteinY,IzhakiI,HalpernM.2013.Dohoneybeesshapethebacterialcommunitycompositioninfloralnectar?PLoSOne8:e67556.

41. RangbergA,MathiesenG,AmdamGV,DiepDB.2015.TheparatransgenicpotentialofLactobacilluskunkeeiinthehoneybeeApismellifera.BenefMicrobes6:513-523.

27

APPENDIX1–GrowthMedia,BuffersandTables

PreparationofMRSMediaSupplementedwithFructose

Avolumeof7.5ml40%fructosewasaddedto26gMRSbroth.Waterwasthenadded

to the final volume of 500 ml giving the final concentration of 0.6 % fructose. The

solutionwasdistributedonculturetubesinvolumesof6mlbeforeautoclaving.

Buffers

TE-buffer(TrisEDTA):

1mMTris-HClpH8.0and100μMEDTApH8.0.

50XTAEstock:

Abottlewasfilledwithapproximately700mldH2Obefore242gTris-Base,100ml0.5

MEDTAand57.1mlGlacialAceticAcidwereadded.The flaskwas filledup to1 liter

withdH2Oandmixedonamagneticstirrer.

1XTAE-buffer:

100ml50XTAEwasmixedwith4.9litersdH2O.

TablesTableA1.1Restrictionendonucleasesandtheirrestrictionsites,incubation-andkilltemperature.

Enzyme Sequence Incubationtemperature HeatkillSpeI A/CTAGT 37°C 80°CXbaI T/CTAGA 37°C 65°CXhoI C/TCGAG 37°C 65°CTableA1.2Dilutionseriesshowingvolumesofcycloheximideandthefinalconcentrations.

Culturetube 1 2 3 4 55.2mg/mLCHXworkingSolution[µL] 0 5 20 50 150CHXFinalConcentration[µg/mL] 0 5.2 20.8 52 156

28

APPENDIX2-Results

DiagramsofGCanalysisresults

FigureA2.1.GCanalysisresultsofvolatilesproducedbytwoyeastisolatesfromappleflowergrownonMRSmediumcomparedwithablanksampleofMRSmedium.

FigureA2.2.GCanalysisresultsofvolatilesproducedbytwoyeastisolatesfromappleflowergrownonLAmediumcomparedwithablanksampleofLAmedium.

110100100010000100000100000010000000100000000100000000010000000000

Aceticacid

Pyrazine

Pyrazine,methyl-

Pyrazine,2,5-

2,3-Butanedione

Acetone

IsopropylAlcohol

Butanal,3-methyl-

EthylAcetate

1-Butanol,3-methyl-

ethanol

PhenylethylAlcohol

1-Propanol,2-methyl-

Butanenitrile,3-

Acetoin

1-propanol

1-butanol

Acetaldehyde

Aceticacid,2-

Methacrolein

2-Butanone

Butanal

Hexanal,3-methyl-

Propanal,2-methyl-

Area[log10]

MRSBlankMRS 47MRS 48MRS

110100100010000100000100000010000000100000000

Aceticacid

Pyrazine

Pyrazine,methyl-

Pyrazine,2,5-

2,3-Butanedione

Acetone

IsopropylAlcohol

Butanal,3-methyl-

EthylAcetate

1-Butanol,3-

ethanol

PhenylethylAlcohol

1-Propanol,2-

Butanenitrile,3-

Acetoin

1-propanol

1-butanol

Acetaldehyde

Aceticacid,2-

Methacrolein

2-Butanone

Butanal

Hexanal,3-methyl-

Propanal,2-methyl-

Area[log10]

LABlankLA 47LA 48LA

29

APPENDIX3-Protocols

PlasmidIsolationProtocol

1) Inoculateasinglecolonyfromapurecultureplatein6mlMRS+fructoseand

incubateat30°Covernight.

2) TransfertheON-culturetoa12mlfalcontubeandcentrifugeat400rpmfor5

minutes.

3) Decantanddiscardtheculturemedia.

4) Dissolvethecellpelletin500µlTE-bufferbyvortexorpipettingandtransferthe

suspensiontoanew1,5mlmicrocentrifugetube.

5) Centrifugeatmaxspeedfor1minuteinatablecentrifuge.

6) Decantanddiscardthesupernatant.

7) Makeamixofenzymesbyadding5µlLysozymeand5µlMutanolysinemultiplied

bythenumberofsamples.

8) Add250µlSolutionItothecellpellet.Vortextomixthoroughly.

9) Add10µloftheenzymemixandinvertthetubestomix.

10) Placethetubeonawaterbathwith37°Cfor10minutes.

11) Add250µlSolutionII.Invertandgentlyrotatethetubeseveraltimestoobtaina

clearlysate.A2-3minutesincubationtimemaybenecessary.NB!AvoidvigorousmixingtoavoiddissolvingchromosomalDNAandalowerplasmidpurity.Do

notexceed5minutesincubation.

12) Add350µlSolutionIII.Immediatelyinvertseveraltimesuntilaflocculentwhite

precipitateforms.NB!ItisimportantthatthesolutionismixedwellandimmediatelyaftertheadditionofSolution

IIItoavoidlocalprecipitation.

13) Centrifugeatmaximumspeedfor10minutes.Acompactwhitepelletwillform.

Promptlyproceedtothenextstep.

14) InsertaHiBindDNAMiniColumnintoa2mlCollectionTube.

15) TransfertheclearedsupernatantfromStep13bycarefullyaspiratingitintothe

HiBindDNAMiniColumn.

16) Centrifugethecolumnatmaxspeedfor1minute.

17) Discardthefiltrateandreusethecollectiontube.

18) Add500µlHBCBuffer.

19) Centrifugeatmaxspeedfor1minute.

30

20) Discardthefiltrateandreusethecollectiontube.

21) Add700µlDNAWashBuffer.

22) Centrifugeatmaxspeedfor30seconds.

23) Discardthefiltrateandreusethecollectiontube.

24) Repeatstep21through23.

25) CentrifugetheemptyHiBindDNAMiniColumnatmaxspeedfor2minutesto

drythecolumn.

26) TransfertheHiBindDNAMiniColumntoanew1,5mlmicrocentrifugetube.

27) Add30µlElutionBuffer.

28) LetthecolumnsitatRTfor1minute.

29) Centrifugeatmaxspeedfor1minute.

30) Repeatstep27through29.

31) StoretheelutedDNAat-20°C.

Postboks 5003 NO-1432 Ås, Norway+47 67 23 00 00www.nmbu.no