Temporal and spatial changes in phyllosphere microbiome of acacia trees growing in arid environments

Abstract

Background: The evolutionary relationships between plants and their microbiome are of high importance to the survival of plants in extreme conditions. Changes in microbiome of plants can affect plant development, growth and health. Along the arid Arava, southern Israel, acacia trees ( Acacia raddiana and Acacia tortilis ) are considered keystone species. In this study, we investigated the ecological effects of plant species, microclimate (different areas within the tree canopies) and seasonality on the endophytic and epiphytic microbiome associated with these two tree species. 186 leaf samples were collected along different seasons throughout the year and their microbial communities were studied using the diversity of the 16S rDNA gene sequenced on the 150-PE Illumina sequencing platform.
Results: our results showed amplifying V4 region of the 16S rDNA better presented the bacterial communities of both end and epiphytes of Acacia trees than V2, V3 and V5 regions of the 16S r DNA. When comparing the bacterial diversity of endo and epiphytes of the two acacia trees (shannon, choa1, PD and observed number of OTU’s), the epiphytes diversity indices showed about twice higher diversity compared to endophytes. The bacterial community compositions comparing both end and epiphytes were also significantly different. Interestingly, Acacia tortilis (umbral canopy shape) had a higher epiphytes bacterial diversity compared to Acacia raddiana, but were not statistically different. However the endophyte bacterial communities were significantly different compared to the two Acacia species (Firmicutes dominated Acacia raddiana and Proteobacteria dominated the Acacia tortilis ) . Alongside the biotic factor, Abiotic factors such as air temperature and precipitation also showed to significantly effect endo and epiphytes bacterial communities, while air humidity only affected the epiphytes bacterial communities.
Conclusions: These results shed light on the unique desert phyllosphere microbiome in mitigating stress conditions highlighting the importance of epiphytic and endophytic microbial communities which are driven by different genotypic and abiotic factors. This paper also shows only a few bacteria species (OTUS’s) to dominate both epi and endophytes highlighting the importance of climate change (precipitation, Air temperature) in affecting arid land ecosystems where acacia trees are considered keystone species in many arid regions.

Full text

1
Preprint:Pleasenotethatthisarticlehasnotcompletedpeerreview.
Temporalandspatialchangesinphyllosphere
microbiomeofacaciatreesgrowinginarid
environments
CURRENTSTATUS:UNDERREVIEW
AshrafAl-Ashhab
ADSSC
ashraf@adssc.orgCorrespondingAuthor
ORCiD:https://orcid.org/0000-0002-3716-2279
ShiriMeshner
adssc
RivkaAlexander-Shani
adssc
HanaDimerets
adssc
MichaelBrandwein
adssc
YaelBar-Lavan
adssc
GidonWinters
adssc
DOI:
10.21203/rs.3.rs-15477/v1
SUBJECTAREAS
Agroecology EcologicalModeling
KEYWORDS
Acaciaraddiana,Acaciatortilis,Phyllosphere,Desertplants,Microbiome,Endophytes
Epiphytes

2
Abstract
Background:Theevolutionaryrelationshipsbetweenplantsandtheirmicrobiomeareofhigh
importancetothesurvivalofplantsinextremeconditions.Changesinmicrobiomeofplantscan
affectplantdevelopment,growthandhealth.AlongthearidArava,southernIsrael,acaciatrees(
AcaciaraddianaandAcaciatortilis)areconsideredkeystonespecies.Inthisstudy,weinvestigated
theecologicaleffectsofplantspecies,microclimate(differentareaswithinthetreecanopies)and
seasonalityontheendophyticandepiphyticmicrobiomeassociatedwiththesetwotreespecies.186
leafsampleswerecollectedalongdifferentseasonsthroughouttheyearandtheirmicrobial
communitieswerestudiedusingthediversityofthe16SrDNAgenesequencedonthe150-PEIllumina
sequencingplatform.
Results:ourresultsshowedamplifyingV4regionofthe16SrDNAbetterpresentedthebacterial
communitiesofbothendandepiphytesofAcaciatreesthanV2,V3andV5regionsofthe16Sr
DNA.Whencomparingthebacterialdiversityofendoandepiphytesofthetwoacaciatrees(shannon,
choa1,PDandobservednumberofOTU’s),theepiphytesdiversityindicesshowedabouttwicehigher
diversitycomparedtoendophytes.Thebacterialcommunitycompositionscomparingbothendand
epiphyteswerealsosignificantlydifferent.Interestingly,Acaciatortilis(umbralcanopyshape)hada
higherepiphytesbacterialdiversitycomparedtoAcaciaraddiana,butwerenotstatisticallydifferent.
Howevertheendophytebacterialcommunitiesweresignificantlydifferentcomparedtothetwo
Acaciaspecies(FirmicutesdominatedAcaciaraddianaandProteobacteriadominatedtheAcacia
tortilis).Alongsidethebioticfactor,Abioticfactorssuchasairtemperatureandprecipitationalso
showedtosignificantlyeffectendoandepiphytesbacterialcommunities,whileairhumidityonly
affectedtheepiphytesbacterialcommunities.
Conclusions:Theseresultsshedlightontheuniquedesertphyllospheremicrobiomeinmitigating
stressconditionshighlightingtheimportanceofepiphyticandendophyticmicrobialcommunities
whicharedrivenbydifferentgenotypicandabioticfactors.Thispaperalsoshowsonlyafewbacteria
species(OTUS’s)todominatebothepiandendophyteshighlightingtheimportanceofclimatechange
(precipitation,Airtemperature)inaffectingaridlandecosystemswhereacaciatreesareconsidered

3
keystonespeciesinmanyaridregions.
1.Background
Theabove-groundsurfacesofplants(thephyllosphere)harboradiversevarietyofmicroorganisms,
includingbacteria.Theplantphyllospheremicrobiomehasbeenshowntoplayanimportantrolein
theadaptationoftheplanthosttodifferentenvironmentssuchastolerancetoheat,cold,drought
andsalinity[1 − 5].Manystudieshavedemonstrateddesertplantseco-physiologicaladaptationin
microbialfunctionaldiversity[2,6].Whiletheexactcorrelationofphyllospheremicrobial
communitiesandtheseuniqueadaptationsisyettobeclarified,growingfindingsindicatethateach
planttypeprovidesasuitableanduniquemicroenvironment.Plantphyllospheremicrobeswerefound
todifferamongdifferenthabitatandclimateconditionswhencomparedbetweenarid,semiaridand
temperatehabitat(8).Forinstance,recentstudyinvestigatedtheadaptationofthreeNegevdesert
plantspeciesfindingBacteroidetestodominatetheleavesofH.scopariaandwerenotabundantin
theotherspeciestheyinvestigated[7].Thesemicrobeswerealsofoundtocorrelatewithhigh
temperature,droughtsandUNradiation[8,9],regardlessoftheirgeographicaldistance[10].Inthis
context,Desertphyllospheremicrobiomeshowedtointerveneinplantgrowthandalternativewaysin
themetabolismofsomenutrientssuchas:fixingnitrogen(N)fromatmosphericsources[1,11],orby
utilizingphosphorus(P)throughsolubilizingenzymes[12,13]andbyproductionofSiderophoresto
bindiron[14,15],evenincreasingplanthealthagainstotherbacteriapathogenssuchasblight
disease[16],botrytisfungalinfection[17].
Alongsidewiththeeffectofseasonality[6,18,19]andcanopystructure[20]onplantphyllosphere
microbiome,otherstudiesshowedabiotic(climate-related)andbiotic(plantgenotype)factorsplaya
pivotalroleinstructuringthephyllospheremicrobialcommunities[21].Infact,endophytic(“insidethe
leavesofplants”)andepiphytic(“outsidetheleavesofplants”)microbialcommunitiesshowedtobe
differentinthemicrobialcommunitycomposition,wereepiphyticbacterialcommunitieswerericher
andmoreabundantcomparedtotheendophyticbacterialcommunities,moreover,abioticfactors
wereshowntohavedifferenteffectsonendophyticandepiphyticbacterialcommunities.Seasonwas
themajordriverofcommunitycompositionofepiphyteswhilewindspeed,rainfall,andtemperature

4
werethemajordriversforendophyticcomposition[22].
Thesecomplexinteractionsbetweenplantmicrobiome(bothendophyticandepiphytic)anddifferent
bioticandabioticconditionswithinaridecosystems,isofparticularinterestconsideringthecurrent
scenariosofclimatechangeanddesertification[23].Additionally,studiesonmicrobiomesinarid
plantscouldshednewlightonimportantkeymicrobialgroupsthatmightbeofpotentialuseinarid
agriculturalpractices,biotechnologyandplantadaptationstrategiestoclimatechange[24].Inthis
study,wefocusedontheNegevdesert(Fig.1A)byinvestigatingtheendophyticandepiphytic
microbiomeassociatedwiththephyllosphere(leaves)ofAcaciaraddiana(Savi)andAcaciatortilis
(Forssk)(Fig.1B).
ThesetwotreespeciesarefoundgrowinginsomeofthehottestanddriestplacesonEarth.Within
thearidAravavalleyalongtheSyrian-Africantransform(GreatRiftvalley)insouthernIsraeland
Jordan,AcaciaraddianaandAcaciatortilisarethetwomostabundantand,inmanyplaces,theonly
treespeciespresent[25].Inthesearidhabitats,acaciasarefoundmostlygrowinginthechannelsof
ephemeralriverbeds[26].BothAcaciaraddianaandA.tortilisareconsideredkeystonespeciesthat
supportthemajorityofthebiodiversitysurroundingthemandlocallyimprovesoilconditionsforother
plantspecies[26 − 30].Wehypothesizedthatvariationsinthebacterialcommunitiesofphyllosphere
wouldbeassociatednotonlywiththehostspecies(A.raddianaandA.tortilis),butalsowithsampling
season(temporalchanges),treemicroclimate(leavesgrowingonthenorthorsouthsideoftree
canopythatareexposedtodirectsunradiationvs.shadedleaves).
2.Results
Atotalof186acacialeavessamplewerecollectedforbothepiphyticandendophyticmicrobial
communities.Afterdataprocessing,eachofthefiveprimersetswereexaminedseparatelyfortheir
coverageacrossthesamplesandtheirretainedsequencenumbers(TableS4).
Resultsshowthatwhenusingthethirdprimerset,wewereabletoretainthelargestnumberof
sequencesforallthesamples13,944 ± 13,710(n = 186,min = 122)comparedtotherestofthe
primers(Table1).Thus,webasedallourfurtheranalysisofbacterialcommunitiesonthisprimerset.

5
Table1
Sequencenumberofthesplitfilesbasedonthedifferentprimersetwithsomebasicstatisticsincluding
sequencesaverage±SD,numberoffinalsamplesaftercuration(n),andtheminimumsequence
numberineachsample(min).
Firstprimerset 15,812 ± 14,109(n = 186) 4,569 ± 6,594(n = 175,min =
102)
Secondprimerset 1,540 ± 3,631(n = 186) 589 ± 961(n = 75,min = 104)
Thirdprimerset 36,451 ± 18,831(n = 186) 13,944 ± 13,710(n = 186,min =
122)
Forthprimerset 7,774 ± 8573(n = 186) 1,565 ± 3,513(n = 145,min =
100)
Fifthprimerset 998 ± 2,389(n = 183) 1,441 ± 2,434(n = 40,min = 104)
2.1.Acaciabacterialcommunitycompositionofendophyticcomparedto
epiphytic
Thediversityestimatesofepiphyticandendophyticbacterialcommunities,forbothA.raddianaand
A.tortilisatSouth“S”canopysidesareshowninTable2.Foralldiversityestimates,thediversityof
endophyticbacteriadiversitywashalfofthatfoundontheepiphytic(Table2)indicatingadifferent
bacterialcommunityanddiversitypatternthatexistsinthesetwomicrobialcommunities.
Table2
Averagediversityestimates(±SD)measuredacrosstheentiresamplingmonthsforsouthside(S)of
thetreecanopyfortheepiphyteandendophytemicrobialcommunitiesofA.raddianaandA.tortilis.
ShownareobservednumberofOTUs,Chao1species’diversityestimate,microbialcommunities
phylogeneticdiversityandtheShannonbacterialcommunities’diversity.Diversitymetricswere
calculatedinQIIME-1software.
Species Canopy Observed
numberofOTUs
Chao1species’
diversity
estimate
microbial
communities
phylogenetic
diversity
Shannon
bacterial
communities’
diversity
A.raddiana S-epiphyte 375 ± 158 644 ± 284 18.5 ± 5.7 5.2 ± 1.1
S-endophye 171 ± 41 350.3 ± 70 10.3 ± 2.1 2.7 ± 0.5
A.tortilis S-epiphyte 410 ± 164 702 ± 277 20.4 ± 6.5 5.0 ± 1.1
S-endophyte 135 ± 35 275.0 ± 73 8.7 ± 2.1 2.7 ± 0.7
Tocomparethediversitiesofepiphyticandendophyticbacterialcommunitiesextractedfromleaf
samples,acaciasamplesfromsouthcanopysides(Table2)wereanalyzedandplottedusingNMDS
basedonBray-Curtisdistancematrix(Fig.2).Twoseparateclustersforendophyticandepiphytic
bacterialcommunities(Fig.2A)werefoundtobesignificantlydifferent(p = 0.005).However,while
bothacaciaspecies(A.raddianaandA.tortilis)demonstratedseparateclusterswithintheendophytic
bacterialcommunities(p-value = 0.006,Fig.2AandB),theydidnotseparateintodifferentclustersin
theepiphyticsamples(p-value = 0.585,Fig.2A).Toillustratethesedifferences,majorbacterialphyla
wereplottedforbothspeciesinepi-andendophyticsamples(Fig.3).Epiphyticsamplesshowed

6
significantlyhigherabundanceofActinobacteria,Cyanobacteriaandsignificantlylowerabundanceof
FirmicutesandProteobacteriacomparedwithendophyticsamplesfromthesameleaves.While
epiphyticbacterialcommunitiesshowednosignificantchangesinphylumcompositionbetweenthe
hostspecies(A.raddianaorA.tortilis),theendophyticbacterialcommunitiesdifferedbetweenacacia
species(Fig.3).Inendophyticbacterialcommunities,abundanceofFirmicuteswassignificantly
higheronA.raddianacomparedwithleavessampledfromA.tortilistrees(61.2 ± 32.0%and32.0 ±
27.9%,respectively),whileA.tortilishadasignificantlyhigherabundanceofProteobacteriathanA.
raddiana(60.9 ± 26.4%and27.7 ± 21.3%,respectively).Interestingly,Comamonadaceaecomposed
morethan88%andBacillaceaecomposedmorethan90%ofproteobacteriaandactinobacteria
respectively.
2.2.Acaciatemporalandcanopyvariationofphyllospherebacterial
communities
Tocheckthetemporaleffectonepi-andendophyticbacterialcommunities,acacialeafsamplesfrom
differentsamplingtimes(Months;TableS1)wereanalyzedandplottedusingNMDSbasedonBray-
Curtisdistancematrix.Results(Fig.4A)showedseparateclustersforepiphytesbacterialcommunities
atdifferentsamplingmonths(p-value = 0.001).Forendophyticbacterialcommunities(Fig.4B),
differentsamplingmonthsshowednoclearseparation(p-value = 0.574),neverthelessaseparate
clusterwasnoticedforthesamplescollectedinJuly-SeptemberandNovembercomparedwiththe
endophyticbacterialcommunitiescollectedinJanuary,February,andinApril-June(Fig.4B).
Toinvestigatetheeffectofmicroclimate(differentsidesofthetreecanopy)ontheepiphytebacterial
communities,thediversityestimatesofepiphyticbacterialcommunitiesforbothA.raddianaandA.
tortilisin(i)northcanopyside,(ii)centercanopyshadedareawere,and(iii)southcanopysidewere
compared(Table3).BothDiversityestimates(Table3)andbacterialcommunitycomposition
ordination(Fig.5)showednocleardifferencesbetweenthedifferentcanopysidesforbothA.tortilis
andA.raddiana(p-value = 0.728),norbetweenthetwospecies(A.raddianaandA.tortilis)(p-value =
0.123),indicatingasimilarepiphyticbacterialdiversityacrossthedifferentsidesofthetrees.

7
Table3
Averagediversityestimates±SDacrosstheentiresamplingmonthsforepiphyticbacterial
communitiesofA.raddianaandA.tortilis.Shownarethedifferentcanopiesfromwhichsampleswere
takenfrom(N=northcanopyside,C=centercanopyshadedarea,S=southcanopyside),observed
numberofOTUs,Chao1species’diversityestimate,microbialcommunitiesphylogeneticdiversityand
theShannonbacterialcommunities’diversity.DiversitymetricswerecalculatedinQIIME-1software.
Species Canopy Observed
numberofOTUs
Chao1species’
diversity
estimate
microbial
communities
phylogenetic
diversity
Shannon
bacterial
communities’
diversity
A.raddiana N-epiphyte 382 ± 204 676 ± 279 19.0 ± 7.1 5.1 ± 1.6
C-epiphyte 342 ± 181 604 ± 292 17.7 ± 6.4 5.0 ± 1.4
S-epiphyte 375 ± 158 644 ± 284 18.5 ± 5.7 5.2 ± 1.1
A.tortilis N-epiphyte 480 ± 189 801 ± 291 22.3 ± 6.5 5.5 ± 1.1
C-epiphyte 437 ± 180 760 ± 271 20.8 ± 6.4 4.8 ± 1.4
S-epiphyte 410 ± 164 702 ± 277 20.4 ± 6.5 5.0 ± 1.1
Tobetterillustratedifferentepiphytebacterialphylaandtheirseasonalitychanges,weplottedthe
differentbacterialphylacompositionalongwiththedifferentsamplingmonthsandcanopysides
(Fig.6).TemporalfluctuationswerefoundbetweenActinobacteriaandFirmicutescompositions,with
thephylaFirmicutesfoundtobemostlydominantinJanuaryandJulyforbothCenterandSouth
canopysides(72.4 ± 14.6%),butnotintheNorthcanopysidewhichwasdominatedbyActinobacteria
phyla(51.2 ± 17.7%)(Fig.6A).DifferentpatternsofbacterialdiversitywerealsofoundinJuly,where
bothNorthandSouthweredominatedbyActinobacteria(52.7 ± 5.8%)comparedtothecenter
canopysidethatwasdominatedbyProteobacteria(70.2 ± 21.7%).InSeptember,differentclusters
formedintheSouth,NorthandCentercanopysides,andallthesecanopysidesweredominated
mostlybyAcidobacteria(57.0 ± 8.7%)andFirmicutes(23.5 ± 5.5%).Itshouldalsobenotedthatin
September,thedifferentcanopysideshostedmoresimilarproportionsofthesedominatingbacterial
groups.
Totestwhetherotherabioticfactorsaffectthemicrobialcommunitiesdifferentlyforcanopysides,
canonicalcorrespondenceanalysis(CCA)[31]wasperformedfortheepiphytic(Fig.7A)and
endophytic(Fig.7B)bacterialcommunitiesofA.raddianaandA.tortilis.Onlythoseabioticfactors
withsignificantvalues(p-value ≦ 0.005)wereplotted.Resultsshowthattemperature,VPD,humidity
andprecipitationhadasignificanteffectontheepiphytesbacterialcommunitiesregardlessofthe
canopyside(Fig.7A),whiletemperatureandprecipitationhadasignificanteffectontheendophytic

8
bacterialcommunities(Fig.7B).
Totestforthemajorchangesinbacterialspecies,theabundantOTU’sandtheirrelativebacterial
familieswereplottedasaheatmapforendophyticandepiphyticsouthcanopyside(Fig.8).Results
showthatonlyfewbacterialOTU’sweredifferentiallyabundantcomparingepi-andendophytic,or
whencomparingwithintheendophyticbetweenA.raddianaandA.tortilis.Themajordifferences
occurredin5majorunclassifiedOTU’sbelongingtoBacillus,Comamonadaceae,Geodematophilaceae
andMicrococcaceaebacterialfamilies.
3.Discussion
Aimingtoimproveourunderstandingofthefunctionsthatphyllospheremicrobialcommunitiesmight
playinplantsgrowinginextremearidenvironments,weappliedahighresolutionsamplingscheme
tostudyingthephyllospheremicrobialcommunitiesoftwodesertkeystonetrees(Acaciaraddiana
andAcaciatortilis).Weinvestigatedboththeendophyticandepiphyticbacterialcommunitiesto
understandthe:(i)intra-andinter-individualspatialvariationofthemicrobialcommunitywithina
tree(thevariationwithinthesametreecausedbydifferentsidesofthecanopy,andthevariation
betweenneighboringtreesofthesamespeciessampledatthesametimeandsite)(ii)hostspecies
variation(variationofthemicrobialcommunitycausedbythehost(tree)species(i.e.,Acacia
raddianacomparedwithneighboringAcaciatortilis),(iii)temporalvariationofthemicrobial
communitywithinthesametreespeciesandcanopyside(samplescollectedfromthesametreesbut
atdifferentseasons).
Ourresultsdemonstratethattheepiphyticbacterialcommunitiesweremoresensitivetochangesin
theexternalenvironmentalconditions,comparedwiththeendophyticbacterialcommunitiesthat
weremorestablebetweendifferentenvironmentalconditions(e.g.,seasons)butvariedamonghost
treespecies.Surprisingly,upto60%ofthetotalbacterialcommunities(thecombinedendophyticand
epiphyticmicrobiomepopulations)wereunclassifiedbelowfamilylevel,highlightingtheuniqueness
ofthemicrobiomeassociatedwithacaciatreesinthearidenvironmentintheArava.
Theepiphyticbacterialdiversitywasfoundtobesignificantlyhigherthantheendophyticbacterial
community(Table2).IntermsoftheoverallobservednumberofOTU’s,theepiphyticbacterial

9
communitywasshowntohavedoublethediversitycomparedtoitsendophyticbacterialcommunity
counterpart.SimilarfindingsatearlyandlateleavesdevelopmentinOriganumvulgarealsofoundthe
totalnumberofcolony-formingunits(CFU)ofendophyticcommunities(1.8 ± 0.1)waslessthanhalf
oftheCFUofepiphyticbacterialcommunities(5.0 ± 0.2)[32].However,ourresultscontradict
previousworkonmicrobiomesassociatedwithArabidopsisthaliana)thatshowedepiphyticbacterial
diversityindexeswerelowerthanthosemeasuredfortheassociatedendophyticbacterial
communities[33].Arecentstudyontheepiphyticandendophyticfungaldiversityinleavesofolive
treesgrowinginMediterraneanenvironments,showedthattheepiphyticfungalcommunitieshad
higherdiversityindicescomparedtotheendophyticdiversityestimates[22].Thefactthatour
epiphyticOTUdiversitywashigherthantheendophyticdiversityisparticularlysurprising,considering
previouspublicationindicatedthatastheconditionsinsidetheplantmightbemorefavorable
comparedtothehostileconditionsoutside[34].Thismightexplainthedifferentfindingscomparing
epiphyticandendophyticbacterialabundanceanddiversityinotherstudies,butinourcase,bothA.
raddianaandA.tortilishadalowerepiphyticbacterialdiversitycomparedtoepiphyticbacterial
diversitythroughoutthesamplingmonth,includingthehotandharshconditionsofthedesert
summer(TableS5).Thisdiscrepancyfindinginourresultsandpreviouslydocumentfindingscouldbe
uniquetodesertplants.Plantsgrownindesertenvironmentsaresubjectedtocontinuousstress
conditionsincludingincreasedsaltconcentrationinendophyticcompartments[35],decreasing
stomatalconductanceandincreasedconcentrationofabscisicacid[36]andmanyothermetabolites
andenzymes[37].Theseplantresponseswereshowntoaffectplants-microbiomecolonization[32,
38,39].Moreover,ourresultsshowedthattheendophyticandepiphyticbacterialcommunitieswere
significantlydifferentfromeachother(Fig.2A).Infact,endophyticbutnotepiphyticbacteria
communities,differedbetweenthetwoacaciaspecies(Fig.3A,3B,4)–specifictothehost(acacia
tree).Thispotentiallyindicatesthatendophyticbacteriawerehorizontallytransmittedandthatthey
mightbemoreaffectedbygenotypicfactorsratherthanabioticfactors[4,5,21].
Similartootherfindingsindicatingthechangesinbacterialcommunitiesinphyllospherefollowing
differentenvironmentalandbioticfactors[38,39],ourresultsshowseasonalitytobethemajordriver

10
ofcommunitycompositioninepiphyticbacteria(Fig.4AandFig.6),includingaspecificabiotic
parameterssuchas;humidity,temperature,precipitationandVPD(Fig.7AandB).Whiletheseresults
highlightthesignificanteffectoftemperatureonbothepiphyticandendophyticbacterial
communities,theeffectofmicroclimate(differentcanopysides)ontheepiphyticbacterialdiversity
(Table3)andcommunitycomposition(Figs.5and6)showednosignificantvariationforthedifferent
canopysidesforbothspecies.Thiscouldbeexplainedbythedifferencebetweenmonthly
temperature,humidityandprecipitationhindersbacktheseeffectsofcanopysidevariation.
Wealsoshowedthatthebacterialcommunitycompositionsfoundinthisstudy,differfromother
epiphyticorendophyticmicrobiomefoundintropical,subtropicalandtemperateregions,whichare
mostlydominatedbyhighabundanceofAlphaproteobacteria,BacteroidetesandAcidobacteria[1,40,
41].Inourstudy,themajordifferencesbetweenepiphyticandendophyticbacterialcommunities
wereduetothedifferentialabundanceoffourmajorunclassifiedOTU’sbelongingthebacterial
familiesofBacillaceae(Firmicutesphylum)andComamonadaceae(Betaproteobacteriaphylum)for
theendophyteofA.raddianaandA.tortilis,respectively(Fig.8).OtherunclassifiedOTU’sbelonging
tothebacterialfamiliesofGeodematophilaceaeandMicrococcaceae(bothbelongingto
Actinobacteriaphylum)werefoundintheepiphytebacterialcommunities(Fig.8).Thesebacterial
familieswerealsofoundinotherstudiesinvestigatingextremeconditionsthatinvestigatedthe
metagenomicsignaturesofTamarixphyllosphere[10,42,43]andotherdesertshrubs[7],
highlightingtheimportanceandtherelationshipofthesefoundbacterialcommunitiesindesertplants
adaptationtoaridenvironment[7].However,theexactlinkbetweenthesedifferentbacterialgroups
andtheirfunctionaldiversityisstilltobeinvestigated,suchstudiescouldshedthelightofspecific
metabolitesandenzymesthattheseadaptivebacterialgroupsexhibitinsuchanenvironmentandat
differentstressconditions.Learningfromthelongcoevolvedplants-microbiomeformnaturally
occurringplantsinharshconditionsisinvitalunderthecurrentrateofclimatechangeandtheurgent
needfornewinnovativesolutionsthatcanbelearnedfromtheseinteractionsformoreadaptivearid
landagriculture.
4.Conclusion

11
Theevolutionaryrelationshipandinteractionbetweenplantsandtheirmicrobiomeisofhigh
importancetotheiradaptationtoextremeconditions.Changesinplantsmicrobiomecanaffectplant
development,growthandhealth.Inthisstudyweexploredtherelationshipbetweennaturally
occurringdesertplantsandtheirmicrobiomealongseasonalandabioticconditions.Theseresults
shedlightontheuniquedesertphyllospheremicrobiomeinmitigatingstressconditionshighlighting
theimportanceofepiphyticandendophyticmicrobialcommunitieswhicharedrivenbydifferent
genotypicandabioticfactors.Nevertheless,morestudiesutilizingthefunctionaldiversityofthese
differentplants-microbiomeinteractionsinaridclimateisinvitalwithdesertificationandglobal
warmingprocessesinmind.Thepotentialagritechoftheseuniquemicrobialcommunitiescallsfor
moreresearchonthistopicinthefutureexploringthefunctionaldiversityofeachendoandepiphytic
microbialcommunitiesalongsidewithplantsmetabolitesatdifferentstressconditions.
5.MaterialsAndMethods
5.1.Studyareaandsamplingscheme
ThisstudywasconductedintheAravavalley,ahyper-aridregionalongtheSyrian-Africanriftin
southernIsraelandJordan.Theelevationofthearearangesfrom230mabovesealevelto419m
belowsealevel(Fig.1A).TheclimateintheAravavalleyishotanddry:30-yearaverageminimum,
mean,andmaximumairtemperatureofthehottestmonthwas26.2°C,33.2°C,and40.2°C,
respectively;averageminimum,mean,andmobotemperaturesensorsmaximumairtemperatureof
thecoolestmonthwererecordedas9.1°C,14.4°C,and19.6°C,respectively,andannual
precipitationofonly20–70mmisrestrictedtotheperiodbetweenOctoberandMay[30]withlarge
year-to-yearvariations[44].Thecombinationoftheveryhighairtemperaturesandtheverylow
relativehumidityvaluesof6%cancausesummermiddayvaporpressuredeficit(VPD)toreachupto
9kPa[30].Vegetationintheregionisusuallyconfinedtowithinwadis(ephemeralriverbeds[45]),
wherethemainwatersupplycomesfromundergroundaquifers[46,47]andwinterflashfloods[48].
MultipleindividualtreesofA.raddianaandA.tortilisarescatteredthroughouttheSheizafwadi
(Fig.1A),butneverformingacontinuouscanopy.Toinvestigatetheeffectofdifferentcanopysides
onphyllospheremicrobiome,leafsampleswerealsocollectedfromthreedifferentcanopysides

12
(north,centerandsouth;Fig.1D).
WadiSheizafisadrysandystreambedatthenorthernedgeoftheAravaValley,Israel(Fig.1A;
30°44'N,35°14'E;elevation− 137m).Meteorologicaldata(airtemperatureandhumiditylogged
every3hours)forthissitewereobtainedfromtheIsraeliMeteorologicalService(IMS)forstation
340528atHatzeva,located7kmnorthofWadiSheizaf(Fig.1C).
Forsamplingbacteriafromacaciatrees,twoneighboringtrees(> 20mawayfromeachother)ofA.
tortilis(T023andT300)andtwoneighboringtreesofA.raddiana(R284andR286)inWadiSheizaf,
weresampledmonthlybetweenJanuaryandDecember2015fortheirNorth,SouthandCentral
canopysides(Fig.1DandTableS1).Thissamplingschemewaschosentoenableustoinvestigate
theeffectofhavingtwodifferenthost(tree)species,inadditiontothevariationcausedbythe
samplingseasonandthemicroclimateeffect(differentcanopysites(central,north,andsouth-facing
sidesoftree)onthephyllospheremicrobiome.
Duringallsamplingmonths,sampleswerecollectedfromtreesusingsterilegloves(changedbetween
eachsample).Leaves(20–25gfreshweight)werecollectedmonthly(seeTableS1forexactdates)
andinsertedinto15mlsteriletubesplacedonice.Uponreachingthelaboratory(within< 2hrs)
samplesweremovedtofreezers(-20°C)wheretheywerekeptuntilsubjectedtoDNAextraction.
5.2.DNAextraction
AllDNAextractionswereperformedusingtheMoBio96wellplatePowerSoilDNAIsolationKits(MO
BIOLaboratories,California,USA).Forepiphyte“outsideofplantsleaves”,0.15g(FW)ofleaveswere
weighedandplacedin1.5mlEppendorftubesfilledwith500µlMoBioPowerbeadSolutionand
sonicated(DG-1300Ultrasoniccleaner,MRCLAB,Israel)for5minandthenthesolutionwas
transferredtothePowerbeadTubesandtherestofthestepsforDNAextractionwerecarriedout
followingthemanufacturerprotocol.Fortheextractionofendophytic(“insideplantleaves”)microbial
communities,leaveswerewashedusing1mlofDNA/RNAfreewaterthreetimestogetridofasmuch
oftheepiphytemicrobiomefraction.Thewashedleaveswerethencutintosmallpiecesusinga
sterilescalpelandplacedintotheMoBio96wellPowerbeadplateforDNAextractionfollowingthe
manufacturer’sprotocol.AllstepsofDNAextractionwerecarriedoutinasterileUV-hood(DNA/RNA

13
UV-cleanerbox,UVT-S-ARbioSan,Ornat,Israel)toreduceexternalcontaminations.IneveryDNA
extraction96wellplate,DNAextractionnegativecontrolswereaddedbyplacing200µlofRNasefree
water(SigmaAldrich,Israel).AllsampleswereplacedrandomlyintheDNAextractionplateto
excludeanybias.
5.3.PCR,librarypreparationandIlluminasequencing
Inordertoobtainabetterphylogeneticresolutionanddiversityestimate,amultiplexPCRusingfive
differentsetsofthe16SrDNAgeneswasappliedtocoverabout1000bpofthe16SrRNAgene(Table
S2).
FirstPCR(PCR-I)reactionswereperformedintriplicates,whereeachPCR-Ireaction(total25µl)
contained;12.5µlofKAPAHiFiHotStartReadyMix(biosystems,Israel);0.4µlofequalv/vmixed
primersforwardandreverseprimers(TableS2);10µlofmoleculargradedDDW(Sigma,Israel)and;
2µlof(2-100ng/µl)DNAtemplate.PCR-IreactionswereperformedinBiometrathermalcycler
(Biometra,TGradient48)asthefollowing:initialdenaturation95°Cfor2min,followedby35cycles
of98°Cfor10sec,61°Cfor15sec,and72°Cfor7sec.EndingthePCR-Iroutinewasafinal
extensionfor72°Cfor1min.UponcompletionofPCR-I,anelectrophoresisgelwasruntoverifyall
thesamplesworkedsuccessfully.Followingsuccessfulandverifiedamplification,triplicatesamples
werepooledtogetherandwerecleanedusingAgencourt®AMPureXP(BeckmanCoulter,Inc,
Indianapolis,USA)beadsolutionbasedonmanufacturer’sprotocol.
5.3.1.LibraryPreparationandIlluminasequencing
LibrarypreparationwasperformedusingasecondPCR(PCR-II)toconnecttheilluminalinker,adapter
andunique8basepairbarcodeforeachsample[49].ThePCR-IIreactionswerepreparedbymixing
21µlofKAPAHiFiHotStartReadyMix(biosystems,Israel),2µlofmixedprimerswithilluminaadapter
(TableS3),12.6µlofRNasefreewater(Sigma,Israel),4µlofeachsamplefromthefirstPCRproduct
with2µlofbarcodedreverseprimer,andplacedinBiometrathermalcycler(Biometra,TGradient48)
asthefollowing:initialdenaturation98°Cfor2min,andthen8cyclesof98°Cfor10sec;64°Cfor
15sec;72°Cfor25sec;andfinalextensionof72°Cfor5min.ThenallPCR-IIproductwerepooled
togetherandsubjectedtocleaningusingAgencourt®AMPureXP(BeckmanCoulter,Inc,Indianapolis,

14
USA)beadsolutionbasedonmanufacturer’sprotocol,where50µlofpooledPCR-IIproductwere
cleanedusing1:1ratiowiththebeadsolutionformoreconservativesizeexclusionoffragmentsless
than200bp,andatthefinalstep,50µlofDDWwith10mMTris[pH = 8.5]wereaddedtoeach
sample.Thiswasfollowedbyaliquoting48µlofthesupernatantintosterilePCRtubesandstoredin
-80°C,whileanadditional15µlofthefinalproductweresenttotheHebrewUniversity(Jerusalem,
Israel)wheretheyweresequencedonfulllaneof150PEilluminaMiseqplatform.
5.4.Sequenceanalysisandqualitycontrol
Aseriesofsequencequalitycontrolstepswereappliedbeforedataanalysis.Theseincludedthe
followingsteps:allsampleswerefilteredforPhiXcontaminationusingBowtie2[50];incompleteand
low-qualitysequenceswereremovedbypairingthetworeadsusingPEARsoftware[51];lookingfor
ambiguousbasesandmissmergedsequencescarriedoutusingtheMOTHURSoftwareV.1.36.1[52].
Followingqualitycontrol,QIIME-1software[53]wasused.Sequenceswerealigned,checkedfor
chimericsequencesandclusteredtodifferentOTU’s(operationaltaxonomicunit)basedon97%
sequencesimilarity,thenthesequenceswereclassifiedbasedonGreengenesdatabaseV13.8[54],
andanOTUtablewasgenerated.Allsequencesclassifiedasf__mitochondria,c__Chloroplast,
k__ArchaeaandK__Unclassified,wereremovedfromtheOTUtable.
5.4.1.OTUrichnessanddiversityestimates
Foreachsample,fourdiversityestimatesweremeasured;(i)observednumberofOTU’s,(ii)Chao1
species’diversityestimate[55],(iii)Shannonbacterialcommunities’diversity[56],and(iv)microbial
communitiesphylogeneticdiversity[57].AllthesediversitymetricswerecalculatedinQIIME-1
software[58]usingtheparallel_alpha_diversity.pycommandontherarefactionsubsamplesto10,000
sequencesusingmultiple_rarefactions.pycommand.
5.4.2.Assessmentofcommunitycomposition
FromtheobtainedQIIMEclassifiedOTUtable,eachtaxonomicgroupwasallocateddowntothegenus
levelusingsummarize_taxa.pycommandinQiimeandrelativeabundancewassetasthenumberof
sequencesaffiliatedwiththattaxonomicleveldividedbythetotalnumberofsequences.Relative
abundanceswereplottedusingRstatisticalsoftware[59]whereeachphylumwasassigneda
distinguishedcolourandallgeneraunderthesamephylum,wereassignedtodifferentshadesofthe

15
samecolour.
5.4.3.StatisticalAnalysis
UsingRstatisticalsoftware[59]allsampleswereanalyzedbasedonthepreviouslygeneratedOTU
table.UsingVEGANpackage[60]inR,non-parametricmultidimensionalscaling(NMDS)wereusedto
produceordinationbasedonBray-Curtisdistancematrixusingatotalsumtransformedmatrixforthe
rowOTUtable[61,62].Statisticalcomparisonsweredonebasedonanalysisofthesimilaritybetween
thematricesofdifferentOTU’stableusingANOSIMcommandinR.
Declarations
Ethicsapprovalandconsenttoparticipate
Notapplicable

Consenttopublish
AllAuthorapprovehesubmission

Authors'information(optional)
Notapplicable

Availabilityofdataandmaterial
AllcuratedsequenceswerejoinedintoasinglefastafileandsubmittedtoMG-RASTunderprojectlink
(https://www.mg-rast.org/linkin.cgi?project=mgp92155).

Codeavailability
Allcodesforquarationsteps,qualitycontrolandsequenceanalysisuploadedtoGitHubrepository,
includingthemetadatafilesandmadepubliclyavailable(https://github.com/ashrafashhab/Desert-
plant-microbiome).
Competinginterests
Allauthorsdeclarenoconflictofinterest

16

Funding
ThisresearchwassupportedbyICAfundingagency,grantnumber:03-16-06A

Authors'contributions
Dr.AshrafAlAshhabwasinvolvedintheprojectconceptualization,datacuration,formalanalysis,
methodology,projectadministration,resources,visualizationandMSwriting.Dr.ShiriMeshner,
MichaelBrandweinandDr.GidonWinterswereinvolvedinFundingacquisition,project
Conceptualizationandprojectinvestigation.Inaddition,bothDr.GidonWintersandDr.YaelBar-lavn
alsotookanactivepartinMSandgraphicsrevisionandediting.WhileMs.RivkaAlexander-Shaniand
HanaDimeretshelpedinlaboratoryandfieldwork.

Acknowledgments
WethankIsraelCharitableAssociation(ICA)thefundingagencyforitsgeneroussupportofthis
researchthatwasawardedundergrantnumber:03-16-06A.WealsothankDr.NoamShentalfromthe
DepartmentofComputerScienceattheOpenUniversityofIsrael,forhisgeneroussupportin
preliminarydataanalysis,selectionof16sPrimesandhisinsightintheexperimentaldesign.Also
specialthankstoMs.TalGalkerfromAravaStudiofordevelopingFig.1intheMS.
References
1.  K e mbelSW,O’ConnorTK ,  A r n o l d  H K,etal(2014)Relatio n s h i p s  b e tweenphyllosphere
bacterialcommunitiesa n d  p l a n t functionaltraitsinaneotr o p i c a l  f o r est.ProcNatl
AcadSci111:13715 – 1 3 7 2 0.https://doi.org/10.10 7 3 / p n a s.1216057111
2.  M artirosyanV,Steinberge r  Y  ( 2 0 1 4)Microbialfunctionald i v e r s i t y  i n  t h ephyllosphere
andlaimosphereofdif f e r e n t  d e sertplants.JAridEnviron  1 0 7 : 2 6 –33.
https://doi.org/10.101 6 / j . j a r i d e n v.2014.04.002
3.  S a leemM,MeckesN,P e r v a i z  Z H ,TrawMB(2017)M i c r o b i a l  i n teractionsinthe

17
phyllosphereincreasep l a n t  p e r f ormanceunderherbiv o r e  b i o t i c stress.Front
Microbiol8:.https://doi.o r g / 1 0 . 3 3 89/fmicb.2017.000 4 1
4.  W hipps,J.M.,Hand,P.,Pin k ,  D .  a n dBending,G.D.2008 .  P h y l l o s p h eremicrobiology
withspecialreferenceto  d i v e r s i t y andplantgenotype.Jou r n a l  o f  A pplied
Microbiology105(6),p p .  1 7 4 4 –1755.
5.  A g lerMT,RuheJ,KrollS,e t  a l  ( 2 0 1 6 )MicrobialHubTaxaL i n k  H o s t  andAbiotic
FactorstoPlantMicrobio m e  V a r ia tion.PLoSBiol14:.
https://doi.org/10.137 1 / j o u r n a l . p bio.1002352
6.  R e dfordAJ,BowersRM ,  K n i g h t  R , etal(2010)Theecolo g y  o f  t h e  phyllosphere:
Geographicandphylog e n e t i c  v a riabilityinthedistributionof  b a c t e r i a ontreeleaves.
EnvironMicrobiol12:28 8 5 – 2 8 93.https://doi.org/10.1 1 1 1 / j . 1 462-2920.2010.02 2 5 8 . x
7.  M artirosyanV,UncA,Mille r  G ,  e t  a l  (2016)DesertPerenn i a l  S h r u b s Shapethe
Microbial-CommunityM i s c e l l a n y  i n LaimosphereandPh y l l o s p h e r e Space.MicrobEcol.
https://doi.org/10.100 7 / s 0 0 2 4 8-016-0822-9
8.  Q v it-RazN,JurkevitchE,Be l k i n  S  ( 2 0 08)Drop-sizesodalak e s :  T r a n s ie ntmicrobial
habitatsonasalt-secretin g  d e s e r ttree.Genetics178:16 1 5 – 1 6 22.
https://doi.org/10.153 4 / g e n e t i c s.107.082164
9.  C a rvalho,S.D.andCastillo,  J . A .  2 0 1 8.InfluenceofLightonP l a n t - P h y l l o s phere
Interaction.Frontiersinp la n t  s c i e n c e9,p.1482.
10.  F i n k e lOM,BurchAY,EladT, e t  a l  ( 2 0 1 2)Distance-decayrela t i o n s h i p s  p artially
determinediversitypat t e r n s  o f  p hyllospherebacteriaon  T a m r i x  t reesacrossthe
sonorandesert.ApplE n v i r o n  M ic robiol78:6187–619 3 .
https://doi.org/10.112 8 / A E M . 0 0888-12
11.  L a m baisMR,BarreraSE,  S a n t o s  E C,etal(2017)Phyllosp h e r e  M e taproteomesof
TreesfromtheBrazilian  A t l a n t i c  F orestShowHighLevels  o f  F u n c t io nalRedundancy.

18
MicrobEcol.https://doi.o r g / 1 0 . 1 007/s00248-016-08 7 8 - 6
12.  M w a jitaMR,MurageH,Ta n i  A ,  K a h angiEM(2013)Evalu a t i o n  o f  r h izosphere,
rhizoplaneandphyllosp h e r e  b a c teriaandfungiisolatedf r o m  r i c e inKenyaforplant
growthpromoters.Sp r i n g e r P l u s.https://doi.org/10.11 8 6 / 2 1 9 3-1801-2-606
13.  B a t o olF,RehmanY,Hasna i n  S  ( 2 0 16)Phylloplaneassociat e d  p l a n t  b acteriaof
commerciallysuperior w h e a t  v arietiesexhibitsuperiorp l a n t  g r o w thpromoting
abilities.FrontLifeSci.http s : / / d o i . o r g /10.1080/215537 6 9 . 2 0 1 6.1256842
14.  S c a v ino,A.F.andPedraza, R . O .  2 0 1 3.TheRoleofSiderop h o r e s  i n  P lantGrowth-
PromotingBacteria.In: M a h e s h wari,D.K.,Saraf,M., a n d  A e r o n,A.eds.Bacteriain
agrobiology:cropprod u c t i v i t y .  B erlin,Heidelberg:Spring e r  B e r l i n  H eidelberg,pp.
265–285.
15.  F u  S F ,SunPF,LuHY,etal(2 0 1 6 )  P l a ntgrowth-promotingtr a i t s  o f  y e astsisolated
fromthephyllosphere a n d  r h i z o sphereofDroseraspa t u l a t a  L a b.FungalBiol.
https://doi.org/10.101 6 / j . f u n b i o . 2015.12.006
16.  L i  S B , FangM,ZhouRC,Hu a n g  J  ( 2 0 12)Characterizationa n d  e v a l u a tionofthe
endophyteBacillusB01 4  a s  a  p otentialbiocontrolagent  f o r  t h e  c o ntrolof
Xanthomonasaxonop o d i s  p v .  dieffenbachiae-Induce d  b l i g h t  o f Anthurium.Biol
Control.https://doi.org/1 0 . 1 0 1 6 /j.biocontrol.2012.06.0 0 2
17.  K e f i  A ,SlimeneIB,Karkouch I ,  e t  a l  ( 2 0 15)Characterizationof e n d o p h y ticBacillus
strainsfromtomatop la n t s  ( L y c o persiconesculentum)d i s p l a y i n g  antifungalactivity
againstBotrytiscinerea P e r s .  W o rldJMicrobiolBiotechnol.
https://doi.org/10.100 7 / s 1 1 2 7 4-015-1943-x
18.  C o p e landJK,YuanL,Layegh i f a r d  M ,  etal(2015)Seasonalc o m m u nitysuccessionof
thephyllospheremicro b i o m e .  M olPlantMicrobeInterac t  2 8 : 2 7 4 –285.
https://doi.org/10.109 4 / M P M I - 1 0-14-0331-FI

19
19.  R e d f ordAJ,FiererN(2009 )  B a c t e r i a lsuccessionontheleafs u r f a c e :  A novelsystem
forstudyingsuccessiona l  d y n a m ics.MicrobEcol58:18 9 – 1 9 8 .
https://doi.org/10.100 7 / s 0 0 2 4 8-009-9495-y
20.  L e f f  J W ,DelTrediciP,Friedma n  W E ,  F iererN(2015)Spatial s t r u c t u r i n g ofbacterial
communitieswithinindiv i d u a l  G i n k gobilobatrees.Environ M i c r o b i o l  17:2352–2361.
https://doi.org/10.111 1 / 1 4 6 2 -2920.12695
21.  R a s t o giG,CoakerGL,Levea u  J H J  ( 2 0 13)Newinsightsintothe  s t r u c t u r e andfunction
ofphyllospheremicrob i o t a  t h r o u ghhigh-throughputmo l e c u l a r  a p proaches.FEMS
MicrobiolLett348:1–1 0 .  h t t p s : / / doi.org/10.1111/157 4 - 6 9 6 8 .12225
22.  G o m es,T.,Pereira,J.A.,Ben h a d i ,  J . ,  L i n o-Neto,T.andBaptista,  P .  2 0 1 8 .Endophytic
andepiphyticphyllosphe r e  f u n g a lcommunitiesaresha p e d  b y  d ifferent
environmentalfactorsin  a  m e d iterraneanecosystem.  M i c r o b i a lEcology76(3),pp.
668–679.
23.  S t r i n g erLC,DyerJC,ReedM S ,  e t  a l  ( 2 009)Adaptationstoclim a t e  c h a nge,drought
anddesertification:local i n s i g h t s  t o  enhancepolicyinsouthe r n  A f r i c a . EnvironSci
Policy.https://doi.org/10. 1 0 1 6 / j . e nvsci.2009.04.002
24.  V a c h erC,HampeA,Porté A J ,  e t  a l  ( 2 016)ThePhyllosphere:  M i c r o b i a lJungleatthe
Plant–ClimateInterface. A n n u  R e vEcolEvolSyst47:1–2 4 .
https://doi.org/10.114 6 / a n n u r e v-ecolsys-121415-03 2 2 3 8
25.  D a n i n ,A.1983.Desertveg e t a t i o n  o f IsraelandSinai.Jerusale m :  C a n a Publishing
House148p.-illus.,col.illu s . ,  m a p s..EnIcones,Maps.Ge o g  2 .
26.  M u n zbergovaZ,WardD(2 0 0 2 )  A caciatreesaskeystone  s p e c i e s  in negevdesert
ecosystems.JVegSci1 3 : 2 2 7 – 236.https://doi.org/10 . 1 1 1 1 / j. 1654-
1103.2002.tb02043.x
27.  M i l t o n SJ,DeanWRJ(1995)H o w  u s e fulisthekeystonespecie s  c o n c e pt,andcanitbe

20
appliedtoAcaciaeriolob a  i n  t h e  K alahariDesert?ZFürO k o l o g i e  N aturschutz4:147–
156
28.  W a r d,D.andRohner,C.1 9 9 7 .  A n thropogeniccausesofh i g h  m o r talityandlow
recruitmentinthreeAca c i a  t r e e  t axaintheNegevdeser t ,  I s r a e l .  B iodiversityand
Conservation6(6),pp. 8 7 7 – 8 9 3.
29.  R o d g erYS,GreenbaumG ,  S i l v e r  M ,etal(2018)Detecting h i e r a r c h i c allevelsof
connectivityinapopulatio n  o f  A c a ciatortilisatthenortherne d g e  o f  t h especies’
globaldistribution:Comb i n i n g  c l a s s icalpopulationgeneticsa n d  n e t w orkanalyses.
PLoSONE.https://doi.or g / 1 0 . 1 3 71/journal.pone.0194 9 0 1
30.  W i n t e rsG,OtienoD,CohenS ,  e t  a l  ( 2 0 18)Treegrowthand w a t e r - u s einhyper-arid
Acaciaoccursduringth e  h o t t e s t  a nddriestseason.Oec o l o g i a .
https://doi.org/10.100 7 / s 0 0 4 4 2-018-4250-z
31.  T e r  B raakCJF(1986)Cano n i c a l  c o r r espondenceanalysis:a  n e w  e i g envector
techniqueformultivaria t e  d i r e c t  g radientanalysis.Ecology .
https://doi.org/10.230 7 / 1 9 3 8 672
32.  P o n t o nioE,DiCagnoR,Tarra f  W ,  e t  a l(2018)Dynamicand  a s s e m blyofepiphyteand
endophytelacticacidba c t e r i a  d u ringthelifecycleofOriga n u m  v u lgareL.Front
Microbiol.https://doi.org/ 1 0 . 3 3 8 9/fmicb.2018.0137 2
33.  B o d e nhausenN,HortonM W ,  B e r g elsonJ(2013)Bacteria l  C o m m unitiesAssociated
withtheLeavesandthe  R o o t s  o fArabidopsisthaliana.P L o S  O N E .
https://doi.org/10.137 1 / j o u r n a l . p one.0056329
34.  T h a p a,S.andPrasanna,R .  2 0 1 8 . Prospectingthecharac t e r i s t i c s  a n dsignificanceof
thephyllospheremicro b i o m e .  A nnalsofmicrobiology6 8 ( 5 ) ,  p p . 229–245.
35.  A r n d tSK,ArampatsisC,Foe t z k i  A ,  e t  a l(2004)Contrastingpa t t e r n s  o f  leafsolute
accumulationandsalta d a p t a t i o ninfourphreatophyticd e s e r t  p l a ntsinahyperarid

21
desertwithsalinegroun d w a t e r . JAridEnviron.
https://doi.org/10.101 6 / j . j a r i d e n v.2004.01.017
36.  L i a n g J,ZhangJ,WongMH( 1 9 9 6 )  S tomatalconductancein  r e l a t i o n  t o xylemsap
abscisicacidconcentrat i o n s  i n  t w o tropicaltrees,Acaciaco n f u s a  a ndLitsea
glutinosa.PlantCellEnviro n .  h t t p s : / / d oi.org/10.1111/j.136 5 - 3 0 4 0 .1996.tb00230.x
37.  W a h id ,A.,Gelani,S.,Ashraf, M .  a n d  F oolad,M.2007.Hea t  t o l e r a n c einplants:An
overview.Environmen t a l  a n d  e xperimentalbotany61 ( 3 ) ,  p p .  199–223.
38.  R e m us-EmsermannMNP ,  S c h l e c h terRO(2018)Phyllosp h e r e  m i c robiology:atthe
interfacebetweenmic r o b i a l  i n d i v idualsandtheplanthos t .  N e w  P hytol218:.
https://doi.org/10.111 1 / n p h . 1 5 054
39.  S c h l e chter,R.O.,Miebach,M .  a n d  R emus-Emsermann,M . N . P .  2 0 19.Drivingfactorsof
epiphyticbacterialcomm u n i t i e s :  A review.Journalofadv a n c e d  r e search19,pp.57–
65.
40.  K i m  M ,SinghD,Lai-HoeA,e t  a l  ( 2 0 1 2)DistinctivePhyllosphere  B a c t e r i a l
CommunitiesinTropica l  T r e e s .  M icrobEcol63:674–68 1 .
https://doi.org/10.100 7 / s 0 0 2 4 8-011-9953-1
41.  L a f o r est-LapointeI,MessierC ,  K e m b e lSW(2016)Treephy l l o s p h e r e  bacterial
communities:exploring  t h e  m a g nitudeofintra-andinter - i n d i v i d u a l  variationamong
hostspecies.PeerJ4:e2 3 6 7 .  h t tps://doi.org/10.7717/ p e e r j . 2 3 67
42.  F i n k e lOM,DelmontTO,Post  A F ,  B e l k i n S(2016)Metageno m i c  s i g n a turesofbacterial
adaptationtolifeinthep h y l l o s p h e reofasalt-secretingdes e r t  t r e e .  ApplEnviron
Microbiol82:2854–2 8 6 1 .  h t t p s://doi.org/10.1128/A E M . 0 0 4 83-16
43.  F i n k e lOM,BurchAY,Lindow  S E ,  e t  a l (2011)Geographicallo c a t i o n  d e terminesthe
populationstructureinp h y l l o s p h e r emicrobialcommunitie s  o f  a  s a lt-excretingdesert
tree.ApplEnvironMicro b i o l  7 7 : 7 647–7655.https://do i . o r g / 1 0 . 1128/AEM.05565-1 1

22
44.  G i n a t  H,ShlomiY,BatersehS ,  V o g e l  J  (2011)ReductioninPre c i p i t a t i o n  L evelsinthe
AravaValley(Southern I s r a e l  a n d Jordan),1949-2009 .  J  D e a d - S eaAravaRes
45.  W a r d,D.2010.Theeffect s  o f  a p i c a lmeristemdamage o n  g r o w thanddefencesof
twoAcaciaspeciesinth e  N e g e v Desert.EvolutionaryEc o l o g y  R e search.
46.  S h e r AA,WiegandK,Ward  D  ( 2 0 1 0 )DoAcaciaandTam a r i x  t r e e s competeforwater
intheNegevdesert?JA r i d  E n v i r o n .https://doi.org/10.10 1 6 / j . j a r i d e nv.2009.09.007
47.  W i n t e rsG,RyvkinI,RudkovT ,  e t  a l  ( 2 0 15)Mappingunderg r o u n d  l a y ersinthesuper
aridGidronWadiusing e l e c t r i c a l  r esistivitytomography(ER T ) .  J  A r i d  Environ.
https://doi.org/10.101 6 / j . j a r i d e n v.2015.05.008
48.  S h r e sthaMK,StockWD,W a r d  D ,  G olan-GoldhirshA(2003)  W a t e r  s tatusofisolated
Negevdesertpopulatio n s  o f  A c a ciaraddianawithdiffe r e n t  m o r talitylevels.Plant
Ecol.https://doi.org/10.1 0 2 3 / A : 1 024431124954
49.  F u k s G,ElgartM,AmirA,et a l  ( 2 0 1 8 )Combining16SrRNA  g e n e  v a riableregions
enableshigh-resolutionm i c r o b i a l  c ommunityprofiling.Mic r o b i o m e .
https://doi.org/10.118 6 / s 4 0 1 6 8-017-0396-x
50.  L a n g meadB,SalzbergSL ( 2 0 1 2 )  Fastgapped-readalign m e n t  w ithBowtie2.Nat
Methods9:357–359
51.  Z h a n gJ,KobertK,FlouriT,Sta m a t a k i s A(2014)PEAR:Afas t  a n d  a c c urateIllumina
Paired-EndreAdmerg e R .  B i o i n formatics30:614–62 0 .
https://doi.org/10.109 3 / b i o i n f o r matics/btt593
52.  S c h l o ssPD,WestcottSL,Rya b i n  T ,  e t  a l(2009)Introducingm o t h u r :  O p en-source,
platform-independent, c o m m u nity-supportedsoftware  f o r  d e s cribingandcomparing
microbialcommunities.  A p p l  E n v ironMicrobiol75:7537 – 7 5 4 1 .
https://doi.org/10.112 8 / A E M . 0 1541-09
53.  C a p o rasoJG,KuczynskiJ,Stom b a u g h J,etal(2010)QIIMEallo w s  a n a l y sisofhigh-

23
throughputcommunity  s e q u e n cingdata.NatMethods  7 : 3 3 5 – 336
54.  D e S a ntisTZ,HugenholtzP,La r s e n  N ,  etal(2006)Greengen e s ,  a  c h i m era-checked
16SrRNAgenedata b a s e  a n d workbenchcompa t i b l e  w i t h  A RB.ApplEnvironMicrob i o l .
https://doi.org/10.112 8 / A E M . 0 3006-05
55.  H i l l ,  T . C .J.,Walsh,K.A.,Harris,J. A .  a n d  M offett,B.F.2003.Usin g  e c o l o g ic aldiversity
measureswithbacter i a l  c o m m unities.FEMSMicrobiolo g y  E c o l o gy43(1),pp.1–11.
56.  H a e gemanB,HamelinJ,M o r i a r t y  J ,  etal(2013)Robustes t i m a t i o n  o fmicrobial
diversityintheoryandin  p r a c t i c e .  ISMEJ.https://doi.org/10 . 1 0 3 8 / is mej.2013.10
57.  F a i t h , D.P.andBaker,A.M. 2 0 0 7 .  P hylogeneticdiversity(PD)  a n d  b i o d iversity
conservation:somebio i n f o r m a ticschallenges.Evolutiona r y  B i o i n f o rmaticsOnline2,
pp.121–128.
58.  K u c z y nski,J.,Stombaugh,J.,W a l t e r s ,  W .A.,González,A.,Capo r a s o ,  J . G .  andKnight,
R.2012.UsingQIIME  t o  a n a l y z e 16SrRNAgeneseq u e n c e s  f rommicrobial
communities.Currentp r o t o c o l s  i n microbiologyChapter 1 ,  p .  U n it  1E.5.
59.  R  C o reteam(2015)RCor e  T e a m
60.  O k s a nenJ,BlanchetFG,Frien d l y  M ,  e tal(2019)vegan:Com m u n i t y  EcologyPackage.
Rpackageversion2. 5 - 2 .  C r a n R
61.  M a r tín ezI,StegenJC,Maldon a d o - G ó m ezMX,etal(2015)Th e  G u t  M ic robiotaofRural
PapuaNewGuineans :  C o m p o sition,DiversityPatterns,a n d  E c o l o gicalProcesses.Cell
Rep.https://doi.org/10. 1 0 1 6 / j . c elrep.2015.03.049
62.  S i n c l a ir L,OsmanOA,Bertilsso n  S ,  E i l e r A(2015)Microbialcom m u n i t y  composition
anddiversityvia16SrR N A  g e n eamplicons:Evaluating  t h e  i l l u m i n aplatform.PLoS
ONE.https://doi.org/10 . 1 3 7 1 / j o urnal.pone.0116955
Figures

24
Figure1
SouthIsraeltopographymapshowingthestudiedsiteofWadiSheizaf(A),andacaciatrees
(A.raddianaandA.tortilis;B)sampledmonthlyduring2015.Airtemperatureandhumidity
werehourlyobtainedfromtheHazevameteorologicalstation(C).Ineachmonth,leaf
samples(exampleofleavescollectedshowninE)werecollectedfromthenorth,centerand
southsidesofthecanopies(D).withaclose-uppictureshowingthecollectedleavesfor
endophyticandepiphyticmicrobialcommunityanalysis.

25
Figure2
NMDSillustratingthephyllospherebacterialcommunityshowingseparateclustersof
bacterialcommunitiesbetween(A)theepiphytic(red)andtheendophytic(blue)bacterial
communitiesfromleavessampledfromsouthsidecanopyareasand(B)differentclusters
forA.raddiana(blue)andA.tortilis(green)forendophyticbacterialcommunities.

26
Figure3
Boxplotillustratingepi-andendophytemajorbacterialphyla.

27
Figure4
NMDSillustratesthephyllospherebacterialcommunitiesshowingseparateclustersof
bacterialcommunitiesbetweendifferentsamplingmonthsfor(A)epiphyticand(B)
endophyticbacterialcommunities.

28
Figure5
NMDSillustratestheepiphytesbacterialcommunityatnorth(red),south(green)andcenter
(blue)canopyside,forA.raddiana(circles)andA.tortilis(triangles).

29
Figure6
Epiphytesbacterialcommunitycompositionclustersarrangedprimarilybysamplingmonth
andcanopyside.A)Clusteringdendrogramofbacterialcommunities,and(B)bacterial
communitycompositionforthesamplesshowninpanelA.Differentbarcolorsrepresent
differentphylumwhileshadesofthesamecolorrepresentdifferentOTU’s.TheX-axistitles
werecolorcodedfordifferentmonths,Januaryinblue,Juneingreen,Julyinredand
Septemberinblack.

30
Figure7
CCAordinationillustrating(A)epiphyticbacterialcommunityatnorth(red),south(green)
andcenter(blue)canopysidesand(B)endophyticbacterialcommunities,forA.raddiana
(circles)andA.tortilis(triangles)withsignificantabioticfactorsaffectingthebacterial
communities.

31
Figure8
HeatmapshowingtheabundanceofOTUs>5%ofthetotalbacterialcommunities(x-axis)
foreachofthesampledepiphyticandendophyticbacterialcommunitiesatsouthcanopy
sideatdifferentsamplingmonths(1-12)during2015.
SupplementaryFiles
Thisisalistofsupplementaryfilesassociatedwiththispreprint.Clicktodownload.
TableS2(1).docx
TableS4(1).docx
TableS5(1).docx
TableS3(1).docx

32
TableS1(1).docx