Short‐term direct contact with soil and plant materials leads to an immediate increase in diversity of skin microbiota

Mic obiologyOpen. 2019;8:e645.   | 1 o 13 h ps://doi.o g/10.1002/mbo3.645 www.Mic obiologyOpen.com 1 | INTRODUCTION Immune- media ed diseases, such as as hma, alle gy, ype I diabe es, and in lamma o y bowel diseases (IBD), ha e inc eased du ing he las decades in u ban en i onmen s (Le ne , Je emias, & Ma hias, 2015; Okada, Kuhn, Feille , & Bach, 2010). In de eloped coun ies, as high as 21% p e alence o as hma (To e al., 2012), 8%–35% o ood alle gies (Osbo ne e al., 2012) and 5% o o he au oimmune diseases (Hay e & Cook, 2012) ha e been epo ed. In Uni ed S a es alone, i has been es ima ed ha he annual cos s o one indi idual immune- media ed disease may ange om 1 o 20 billion US dol- la s (e.g., AARDA, 2011). The hygiene hypo hesis and i s ex ensions (No e & Hu nagle, 2005; Rook e al., 2004; S achan, 1989) pos- ula e ha he eason o he inc eased p e alence o hese diseases is he inc eased hygiene le el in ou e e yday li e. High hygiene le el and mode n u ban li e- s yle ha e led, o example, o dec eased Recei ed:9Janua y2018 | Re ised:23Ma ch2018 | Accep ed:27Ma ch2018 DOI: 10.1002/mbo3.645 ORIGINAL ARTICLE Sho - e m di ec con ac wi h soil and plan ma e ials leads o an immedia e inc ease in di e si y o skin mic obio a Mi a G ön oos1 | Ani ud a Pa ajuli1 | Olli H. Lai inen2 | Ma ja I. Roslund1 | Heli K. Va i1 | Heikki Hyö y2,3 | Riikka Puhakka1 | Aki Sinkkonen1 This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed. © 2018 The Au ho s. Mic obiologyOpen published by John Wiley & Sons L d. 1Ecosys ems and En i onmen Resea ch P og amme, Facul y o Biological and En i onmen al Sciences, Uni e si y o Helsinki, Lah i, Finland 2Depa men o Vi ology, School o Medicine, Uni e si y o Tampe e, Tampe e, Finland 3Fimlab Labo a o ies, Pi kanmaa Hospi al Dis ic , Tampe e, Finland Co espondence Riikka Puhakka, Ecosys ems and En i onmen Resea ch P og amme, Facul y o Biological and En i onmen al Sciences, Uni e si y o Helsinki, Lah i, Finland. Email: iikka.puhakka@helsinki. i Aki Sinkkonen, Ecosys ems and En i onmen Resea ch P og amme, Facul y o Biological and En i onmen al Sciences, Uni e si y o Helsinki, Lah i, Finland. Email: aki.sinkkonen@helsinki. i Funding in o ma ion Tekes, G an /Awa d Numbe : 40333/14 Abs ac Immune- media ed diseases ha e inc eased du ing he las decades in u ban en i on- men s. The hygiene hypo hesis sugges s ha inc eased hygiene le el and educed con ac s wi h na u al biodi e si y a e ela ed o he inc ease in immune- media ed diseases. We es ed whe he sho - ime con ac wi h mic obiologically di e se na u e- based ma e ials immedia ely change bac e ial di e si y on human skin. We es ed di ec skin con ac , as wo olun ee s ubbed hei hands wi h six een soil and plan based ma e ials, and an exposu e ia ab ic packe s illed wi h moss ma e ial. Skin swabs we e aken be o e and a e bo h exposu es. Nex - gene a ion sequencing showed ha exposu es inc eased, a leas empo a ily, he o al di e si y o skin mi- c obio a and he di e si y o Acidobac e ia, Ac inobac e ia, Bac e oide es, P o eobac e ia and Alpha- , Be a- and Gammap o eobac e ia sugges ing ha con ac wi h na u e- based ma e ials modi y skin mic obiome and inc ease skin mic obial di- e si y. Un il now, app oaches o cu e o p e en immune sys em diso de s using mic obe- based ea men s ha e been limi ed o use o a ew mic obial species. We p opose ha na u e- based ma e ials wi h high na u al di e si y, such as he ma e ials es ed he e, migh be mo e e ec i e in modi ying human skin mic obiome, and e en- ually, in educing immune sys em diso de s. Fu u e s udies should in es iga e how long- e m changes in skin mic obio a a e achie ed and i he exposu e induces ben- e icial changes in he immune sys em ma ke s. KEYWORDS biodi e si y hypo hesis, human heal h, hygiene hypo hesis, na u e-based ma e ials 2 o 13 | GRÖNROOS e al. exposu e o mic oo ganisms and pa asi es and inc eased pe u ba- ions o mic obio a, o example, due o ex ensi e use o an ibio ics. These changes may a ec he no mal de elopmen o immune sys- em in ea ly childhood (S iemsma, Reynolds, Tu ey, & Finlay, 2015). Mos ecen ly, he hygiene hypo hesis has been ex ended o a bio- di e si y hypo hesis ( on He zen, Hanski, & Haah ela, 2011), which sugges s ha he apid global biodi e si y decline is ela ed o he inc ease in immune- media ed diseases: he dec eased mic obial di- e si y in he u ban li ing en i onmen con ibu es o he educ ion in na u al exposu e o mic oo ganisms (Pa ajuli e al., 2018) and p e- en he na u al de elopmen o immune sys em. Hygiene and biodi e si y hypo heses ha e gained suppo in nume ous s udies. These s udies p o ide a leas h ee ypes o e - idence: (1) educed di e si y o pe sonal mic obio a (e.g., s ool o skin mic obio a) is associa ed wi h many immune- media ed diseases (Hanski e al., 2012; Manichanh, 2006; Sche e al., 2015), (2) cha ac- e is ics o he li ing en i onmen , such as he le el o u baniza ion, a e associa ed wi h he p e alence o hese diseases (Kond asho a, Seiska i, Ilonen, Knip, & Hyö y, 2012; Kond asho a e al., 2005; Ruokolainen e al., 2015) and (3) mic obial di e si y in he li ing en i- onmen is associa ed wi h hese diseases (Ege e al., 2012; Valkonen, Wou e s, Täubel, Rin ala, & Len e s, 2015). An example o he i s ype is p o ided by Sche e al. (2015), who ound ha gu mic obio a in pa ien s wi h pso ia ic a h i is and skin pso iasis was less di e se compa ed o ha in heal hy indi iduals. Kond asho a e al.’s (2005) s udy ep esen s he second ype o e idence as hey ound ha he incidence o Type 1 diabe es was six old highe in he mo e u banized Eas e n Finland wi h a high hygiene le el compa ed o he adjacen Russian Ka elia al hough he equency o he p edisposing geno- ypes did no di e be ween he wo popula ions. Ege e al.’s (2012) s udy is an example o he hi d ype: hey ound ha , compa ed o a e e ence g oup, child en li ing on a ms had a lowe p e alence o as hma and a opy coinciding wi h a highe mic obial di e si y in he dus collec ed om hei homes. E en hough many s udies ha e ound ha he o al mic obial di e si y ma e s (O e al., 2004; de Pai a e al., 2016; Sche e al., 2015), o he s ha e iden i ied ce ain mic obial g oups o be mo e impo an . Fo example, pa ien s wi h C ohn’s disease had lowe di e si y o phylum Fi micu es in hei s ool (Manichanh, 2006), in- an s wi h eczema had lowe di e si ies o phyla Bac e oide es and P o eobac e ia in s ool (Ab ahamsson e al., 2012) and adolescen s wi h a opy had lowe di e si y o Gammap o eobac e ia on skin (Hanski e al., 2012) compa ed o hose o heal hy indi iduals. Based on he hygiene and biodi e si y hypo heses i can be ex- pec ed ha inc easing he exposu e o di e se and na u al mic obial communi ies would di e si y and change he composi ion o human mic obio a and ha his change would educe he isk o immune- media ed diseases. Su p isingly, al hough se e al exposu e s udies using sepa a e bac e ial s ains ha e been conduc ed (S iemsma e al., 2015), di e se na u al ma e ial has ha dly been es ed expe - imen ally and, o ou knowledge, ne e o humans. Soils including compos ed ga dening ma e ials hos especially di e se mic obial communi ies (Yu e al., 2015) and hus inc easing con ac wi h such soils o u ban ci izens could p o ide a means o inc easing he di- e si y o hei mic obio a and u he dec ease he p e alence o as hma and a opies ( on He zen & Haah ela, 2006). This is espe- cially signi ican in he ligh o ou ea lie indings, which sugges ha u baniza ion and pollu ion could lead o changes in soil mic o- bio a and u baniza ion educes mic obial ans e indoo s (Pa ajuli e al.,2017,2018). He e, we used wo simple expe imen al se ups o s udy whe he a sho - ime con ac wi h na u al ma e ials can change bac e ial di- e si y on human skin. We concen a ed on skin mic obio a, as i has complex in e ac ions wi h immune sys em; o example, commensal mic obes can p omo e immune homeos asis and pa hogen esis ance (Chen, Fischbach, & Belkaid, 2018). Indeed, se e al skin diso de s ha e been linked wi h imbalance o skin mic obio a (e.g., Rod iques Ho man,2017).Wecollec edse e alcompos edga deningma e i- als, wo moss ma e ials and one pea ma e ial om comme cial soil p oduce s and ook bac e ial skin swab samples om wo olun ee s be o e and a e sho - ime skin exposu e o hese ma e ials. Hands we e i s exposed di ec ly o all he aw ma e ials o s udy i bac- e ial di e si y on skin inc eased a e exposu e. Then, we selec ed one ma e ial, moss, modi ied i and pu i in ab ic packe s o es i simila e ec can s ill be seen wi h he ma e ial illed in ab ica es, which can be conside ed mo e con enien o use in e e yday li e. In addi ion o he o al bac e ial di e si y, we inspec ed he di e si y wi hin bac e ial phyla domina ing ei he in soils (i.e., P o eobac e ia, Acidobac e ia, and Ac inobac e ia) (Janssen, 2006) o on human skin (i.e., Ac inobac e ia, Fi micu es, Bac e oide es, and P o eobac e ia) (Rod igues Ho mann, 2017). As P o eobac e ia is o en he mos  abundan phylum in soils and he di e si y o some P o eobac e ial classes has been associa ed wi h immune de ense (Hanski e al., 2012), we also inspec ed he di e si y wi hin majo P o eobac e ial classes (i.e., Alpha- , Be a- , Gamma- and Del ap o eobac e ia). We hypo hesized ha di ec soil con ac wi h ga dening ma e ials, moss and pea inc ease skin mic obial di e si y immedia ely a e he ex- posu e. In acco dance wi h his hypo hesis, we de ec ed an inc ease in o al di e si y as well as in he di e si y o all es ed axonomic g oups excluding he phylum Fi micu es. The esul s suppo he co e assump ions o hygiene and biodi e si y hypo heses ha con- ac wi h di and mic obially ich na u e a ec s he human mic o- bio a. He e, we also p opose ha daily mic obial exposu e could be ailo ed by designing plan and soil based ma e ials ha comp ise ich mic obial lo a. 2 | MATERIAL AND METHODS 2.1 | Expe imen al design: Di ec hand- exposu e (expe imen 1) In he i s expe imen , wo u ban olun ee s es ed eigh com- pos ed, soil and plan based ma e ials. Al oge he 16 ma e ials we e es ed. Ma e ials es ed we e collec ed om i e comme cial en- e p ises p oducing ga dening ma e ials: Biolan Oy/No a b o Oy, Humuspeh oo i Oy, Kekkilä Oy, Suomen Kun apiha/SnowWay Oy | 3 o 13 GRÖNROOS e al. and Vapo Oy. Ma e ials included six comme cial soil p oduc s ( ade names: Mus a Mul a, Nii ymul a, Nu mikkomul a, Pe ennamul a, Puis omul a, Viljelymul a), i e compos ed soil en ichmen p oduc s which a e ei he sold di ec ly o end use s o used as aw ma e ial in o he p oduc s, wo di e en kind o o es u s (i.e., ans e able o es loo ) and wo moss ma e ials and one pea ma e ial ga he ed om p is ine Finnish pea bogs. The compos ed soil en ichmen p oduc s we e p oduced om animal dung (e.g., ho se and chicken dung), deciduous lea li e , plan deb is and sludge om was ewa e ea men plan . The ing edien s o he comme cial soil p oduc s in- clude compos ed dung, sludge and plan ma e ials, pea , wood mulch and mine al soils. P oduc s we e compos ed and con olled acco d- ing o E.U. egula ions. We used non- iden i iable codes o he ma- e ials o ollow he guidelines o ag eemen s wi h collabo a o s. Volun ee s ubbed hei hands in a es ma e ial o 20 s, washed hei hands wi hou soap in ap wa e o 5 s, and d ied he hands wi h pape owels. The p ocedu e was epea ed o all es ma e i- als sepa a ely. Ma e ials we e es ed in andom o de . No mo e han wo ma e ials we e es ed in he same day and he e was a leas i e hou s in e al be ween he es s. A skin swab (back- side o he igh hand, 3 × 3 cm a ea, 9 wipes) was aken wice, jus be o e exposu e and immedia ely a e d ying hands wi h a pape owel. A co on wool s ick was i s soaked in Tween® 20, used in sampling and cu o a s e ile polye hene sam- ple ube. Using he abo e- men ioned p o ocol, hands seemed clean h ough naked eye bu he co on used o aking he swab was seemingly da ke a e exposu e han be o e exposu e. We wan o emphasize ha his s udy did no handle he sa e y o no using soap in e e yday li e. He e, ou pu pose was only o es how mic o- bial di e si y on skin changes when a ious biodi e se ma e ials a e handled. 2.2 | Expe imen al design: Exposu e ia ab ic packe s (expe imen 2) In he second expe imen , wo u ban olun ee s es ed ab ic pack- e s illed wi h d ied, c ushed, and sie ed Sphagnum moss (pa icle size less han 1 mm). A laye o he ma e ial was placed inside a ab- ic packe o size 10 × 10 cm. Th ee di e en ypes o ab ics we e es ed: ai laid ma e ial ST047DIA ( hickness 0.44mm; Sha pCell, Kausala, Finland), ai laid ma e ial DS100 ( hickness 0.85 mm; Sha pCell, Kausala, Finland), and co on ab ic (Ma imekko, Helsinki, Finland). Bo h olun ee s es edonepacke madeo ST047DIAandone packe made ei he o DS100 o co on ab ic. The packe s we e placed on he inne o ea ms o he olun ee s. Packe s we e ied on wi h a clean disposable sel - adhesi e bandage (Pha maca e Spo Bandage). The olun ee s we e exposed o he packe s o 3 h and 45 min. Skin swabs (5 × 5 cm a ea, 10 s) we e aken jus be o e plac- ing he packe s and immedia ely a e hey we e emo ed. A s ick wi h a co on wool ip was i s we ed in Tween® 20, used in sam- pling and placed in o a s e ile polye hylene sample ube. We wan o poin ou , ha ou expe imen al se up does no ake in o accoun he possible dis u bance caused by he packe i sel . Howe e , as he exposu e ime is sho , we belie e ha he dis u bance e ec is low compa ed o he ans e o mic obes. Bac e ial composi ion on skin be o e and a e use o he packe s was compa ed o he bac e ial composi ion in he Sphagnum moss ha was used as a aw ma e ial o he packe s. Expe imen al p o ocol was app o ed by he e hics commi ee o Tampe e Uni e si y Hospi al (case numbe : ETL R15081). The s udy was ca ied ou in acco dance wi h he ele an guidelines and egu- la ions in Finland and he olun ee ’s signed in o med consen s. 2.3 | Sample p epa a ion o MiSeq sequencing Skin swab samples we e s o ed in deep eeze  (<−70°C) in ubes con aining Tween® 20 (MP Biomedicals) (0.1%) + NaCl (0.1 mol/L, J.T.Bake ) be o e DNA ex ac ion. To al DNA was ex ac ed om samples using Powe Soil® DNA Isola ion Ki (MoBio Labo a o ies, Inc., Ca lsbad, CA, USA) acco ding o he manu ac u e ’s s anda d p o ocol. The swab was ans e ed o he Powe Bead ube o ho- mogeniza ion and lysis. Fo he moss sample used in expe imen 2, app oxima ely 0.25 g o moss was used o DNA ex ac ion. DNA was checked wi h aga ose gel (1.5%) elec opho esis. To al DNA concen a ion was measu ed wi h Quan - iT™ PicoG een® dsDNA eagen ki (The mo scien i ic, MA, USA). In expe imen 1, DNA was analyzed o bac e ial (16S) com- muni iesusinga wo-s ep PCR app oach o a oid a 3′-end ampli- ica ion bias esul ing om he sample- speci ic DNA ags (Be y, Mah ouldh, Wa ne , & Loy, 2011). The a iable egions 1–3 wi hin he 16S ibosomal RNA ( RNA) gene was ampli ied by p ima y PCR ( h ee eplica es om each sample) using pA and PD Illumina p ime s wi h o e hangs. P ima y PCR was ca ied ou in a eac- ion mix u e ( eac ion olume 50 μl) consis ing o 1 μl each o 10 mmol/L deoxynucleo ide iphospha es (dNTPs; The mo sci- en i ic, MA, USA), 5 μl o wa d p ime pA_Illum_FP (10 μmol/L; ATCTACACTCTTTCCCTACACGACGCTCTTCCGATCTAG AGTTTGATCMTGGCTCAG) and 5 μl e e se p ime  pD′_Illum_RP (10 μmol/L; GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCTGTA TTACCGCGGCTGCTG), 0.5 μl 2 U/μl Phusion G een Ho S a II High- Fideli y DNA polyme ase (The mo scien i ic, MA, USA), 10 μl 5× G een HF PCR bu e  (F-537), 5μl empla e DNA, and 23.5 μl s e ile wa e . The PCR eac ion was pe o med in a he mocycle (MJResea ch,MA,USA)as ollows:ini ialdena u a iona 98°C o  5min, ollowedby30cycleswi hdena u a iona 94°C o 1min, annealing o 10sa 50°Candex ension o 1mina 72°C,and hen a inalex ensiona 72°C o 10min.Aposi i econ ol(Cup ia idus neca o JMP134, DSM 4058) was included in PCR uns and a nega- i e con ol (s e ile wa e ) was un o de ec any possible con amina- ion. DNA was de ec ed wi h aga ose gel (1.5%) elec opho esis. The PCR p oduc s we e pu i ied using Agencou AMPu e XP solu ion (Beckman Coul e Ins.) o educe ca yo e o p ima y PCR p im- e s. Illumina adap e o e hang nucleo ide sequences we e added o he 16S RNA gene-speci ic sequences in he seconda y PCR. The seconda y PCR and sequencing we e pe o med a The Ins i u e o 4 o 13 | GRÖNROOS e al. Bio echnology (Uni e si y o Helsinki) using Illumina MiSeq pla o m. In he seconda y PCR, ull leng h adap e s and Indexes we e in o- duced. The PCR p o ocol was as desc ibed by Koskinen, Hul man, Paulin, Au inen, and Kankaanpää (2011). The sequencing was done as pai ed- end (300 bp+300 bp) on a MiSeq Illumina ins umen using a 3 eagen ki . In expe imen 2, he V4 egion wi hin he 16S ibosomal RNA ( RNA) gene was ampli ied by p ima y PCR as iplica es using 505F and 806R p ime s (Capo aso e al., 2012). P ima y PCR was ca ied ou in a eac ion mix u e ( eac ion olume 50 μl) consis ing o 1 μl each o 10 mmol/L dNTPs (The mo sci- en i ic, MA, USA) 5 μl o wa d p ime 505F (10 μmol/L; 5′– GTGCCAGCMGCCGCGGTAA-3′) and 5μl e e se p ime 806R (10 μmol/L; 5′–GGACTACHVGGGTWTCTAAT-3′), 0.5μl 2 U/μl Phusion G een Ho S a II High- Fideli y DNA polyme ase (The mo scien i ic, MA, USA), 10 μl 5× G een HF PCR bu e  (F-537), 5μl empla e DNA and 23.5 μl s e ile wa e . The PCR eac ion was pe - o med in a he mocycle (MJ Resea ch, MA, USA) as ollows: ini ial dena u a iona 98°C o 5min, ollowedby30cycles(only25cycles wasused o  hemosssample)wi hdena u a iona 94°C o 1min, annealing o 10sa 50°Candex ension o 1mina 72°C,and hen a inalex ensiona 72°C o 10min.Aposi i e(Cup ia idus neca o JMP134, DSM 4058) and a nega i e con ols (s e ile wa e ) we e in- cluded in PCR and DNA was de ec ed wi h aga ose gel elec opho- esis. The PCR p oduc s we e pu i ied using Agencou AMPu e XP solu ion (Beckman Coul e Ins.). T iplica es o he cleaned amplicons we e pooled and dilu ed 1:5. Cleaned and dilu ed p ima y PCR p od- uc s we e a ge ed in he seconda y PCR (TagPCR). Reac ion mix u e o he TagPCR was equal as abo e excep e e se p ime included a 12 bp unique Mul iplexing Iden i ie ag (MID- 806R). Ampli ica ion p og am was he same as abo e excep only en cycles we e used (7 cycles o  he moss sample). TagPCR p oduc s we e de ec ed on aga ose gel elec opho esis, pu i ied wi h Agencou AMPu e, pooled and DNA concen a ion was measu ed wi h PicoG een. The sequencing was conduc ed a he Kansas S a e Uni e si y using Illumina MiSeq pla o m wi h a 2 × 300 bp e sion 3 sequencing ki acco ding o manu ac u e ’s p o ocol and he GeneRead DNA Lib a y I Co e Ki (Qiagen, ca alog # 180432) was used o liga e Illumina’s T uSeq adap e s o amplicons. 2.4 | Sequence p ocessing Raw sequencing da a we e p ocessed using Mo hu - p og am ( e - sions 1.36.1 and .1.38.1). The da ase s o expe imen s 1 and 2 we e p ocessed sepa a ely al hough using almos he same se- quence p ocessing p o ocol. The p o ocol pa ly ollowed he pipeline sugges ed by Schloss, Ge e s, and Wes co (2011) and Kozich, Wes co , Bax e , Highlande , and Schloss (2013). The pai ed sequences in e e se and o wa d as q iles we e aligned in o con igs. Sequences we e immed and sc eened o emo e any mis- ma ches wi h p ime o DNA- ag sequences, ambiguous bases and homopolyme s la ge han 8 bp long. Sequences we e aligned using Mo hu e sion o SILVA bac e ial e e ence sequences ( e sion 132) (P uessee al.,2007)and hesequenceswhichwe eno aligned o a e e ence alignmen o he co ec sequencing egion we e e- mo ed. Unique sequences and hei equency in each sample we e iden i ied, and hen, almos iden ical sequences (>99% simila ) we e p eclus e ed o minimize sequencing e o s (Huse, Welch, Mo ison, & Sogin, 2010) and sc eened o chime as wi h UCHIME (Edga , Haas, Clemen e, Quince, & Knigh , 2011) which uses he abundan sequences as a e e ence. The chime ic sequences we e emo ed. We calcula ed a pai wise dis ance ma ix o unique sequences and clus e edOTUsa 97%sequencesimila i yusing henea es neigh- bo algo i hms. Sequences we e classi ied using he Mo hu e sion o Bayesianclassi ie (Wang,Ga i y,Tiedje,&Cole,2007)wi h he RDP aining se e sion 16 (Cole e al., 2009). Sequences classi ied o Chlo oplas , Mi ochond ia, unknown, A chaea, and Euka yo a we e emo ed om he analyses. Ra e OTUs ha we e ep esen ed by 10 o ewe sequences in he whole da a we e emo ed as sug- ges ed by Oli e , B own, Callaham, and Jumpponen (2015). Fo each OTU he numbe o sequences ound in nega i e con ols we e sub- ac ed om each sample. This kind o p ocedu e was a comp omise be ween wo ex emes: emo ing all OTUs ound in con ols and no emo ing any. I has been sugges ed no o emo e OTUs iden i ied in nega i e con ols i hey a e biologically expec ed (Sal e e al., 2014). In his s udy, i was e y di icul o iden i y biologically ex- pec ed bac e ia because bac e ia could o igina e om human skin, soil o plan ma e ial. On he o he hand, o ally igno ing he OTUs ound in nega i e con ols migh ha e esul ed in ela i ely la ge pe - cen age o alse OTUs in be o e samples as hese samples had much lowe DNA concen a ion compa ed o he a e samples (Figu e 3.). Finally,all hesampleswe e a e ied o7,286sequencesinexpe i- men 1 and o 11,511 sequences in expe imen 2, which we e he lowes numbe o sequences in each expe imen . 2.5 | Quan i a i e PCR In expe imen 1, we used he quan i a i e PCRs (q- PCRs) o bac e ial 16S RNA gene based on SYBR g een de ec ion. PCRs we e ca ied ou wi h he Ligh Cycle 96 Quan i a i e eal- ime PCR machine (MJ Resea ch,MA,USA).The o wa dp ime usedwaspE5′-AAACTC AAAGGAATTGACGG-3`and he e e sep ime pF5′-ACGAGC TGA CGA CAG CCA TG- 3` (Kan o Öq is e al., 2008). All samples we e ampli ied in iplica es in 20 μl eac ions con aining 10 μl 2× Powe Up SYBR G een Mas e Mix (The mo scien i ic, MA, USA), 0.2 μl 20 mg/ml BSA, 0.5 μl o each p ime (10 μmol/L), and he sam- ple empla e. A s anda d cu e was included in e e y un o allow quan i ica ion o he numbe o bac e ial 16S copies p esen in he o iginal sample. The q- PCR cycling was as ollows: ini ial dena u a- iona 95°C o 2min, ollowedby40cycleso dena u a iona 95°C o 10s,annealing o 20sa 53°Candex ension o 30sa 72°C. Mel ingcu eanalysison heampliconwasas ollows:95°C o 10s, 65°C o 60s,97°C o 1s,37°C o 30swi hcon inuousmeasu e- men o he luo escence signal. DNA o Cup ia idus neca o JMP134 (DSM 4058) was used as a s anda d ha wo ked also as a posi i e con ol while s e ile wa e was used as a nega i e con ol. | 5 o 13 GRÖNROOS e al. 2.6 | S a is ical me hods, expe imen 1 Fo expe imen 1, he analyses we e conduc ed o he whole communi y a di e en axonomic le els (i.e., OTU, genus, amily, o de , class, and phylum). A OTU le el, analyses we e also con- duc ed wi hin phyla Acidobac e ia, Ac inobac e ia, Bac e oide es, Fi micu es and P o eobac e ia and classes Alpha- , Be a- , Gamma- and Del ap o eobac e ia. This kind o app oach is some imes called as decons uc ing species di e si y and i means ha he whole communi y da a a e pa i ioned in o smalle g oups o example, by axon, guild o o he g ouping (Ma que , Fe nández, Na a a e, & Valdo inos, 2004; Tolonen e al., 2016). We used his app oach because ine axonomic g oupings may show pa e ns di e ing om he pa e ns ound o he whole communi y (Ma que e al., 2004). This commonly used app oach has p e iously been used also wi hPERMANOVAandPERMDISP(An hony,F ey,&S inson,2017; G ön oos e al., 2013). Pai ed Wilcoxon signed- ank es o S uden ’s T- es was used o compa ing he numbe o bac e ial 16S copies, numbe o bac e- ial OTUs (i.e., ichness) and Shannon di e si y index o he bac e ial communi y in hands be o e and immedia ely a e exposu e. T- es was conduc ed when he da a we e no mally dis ibu ed based on Shapi o- Wilk es and Wilcoxon signed- ank es was conduc ed when he da a we e no no mally dis ibu ed. The di e ence be ween bac e ial composi ion in hands be o e and immedia ely a e exposu e was s udied using Pe mu a ional Mul i a ia e Analysis o Va iance (PERMANOVA) (Ande son, 2001), Mul i a ia e Homogenei y o G oup Dispe sions (PERMDISP) (Ande son, 2006; Ande son, Ellingsen, & McA dle, 2006) and isu- ally illus a ed using Non- Me ic Mul idimensional Scaling (NMDS). PERMANOVA was un using wo ac o s: be o e/a e ea men and pe son. PERMDISP was un using bias co ec ion (S ie , Geange, Hanson, & Bolke , 2013) and spa ial median as he g oup cen oid. In all he h ee me hods, B ay- Cu is dis ance was used. PERMANOVA and PERMDISP we e un using 999 pe mu a ions. PERMANOVA, PERMDISP, and NMDS we e no conduc ed o Acidobac e ia and Del ap o eobac e ia because se e al samples did no con ain any bac e ia belonging o hese wo axa. Fo Bac e oide es, one sample included only one sequence which masked he di e ences among all he o he samples. Thus, o Bac e oide es, he gi en sample pai was emo ed when unning NMDS, PERMANOVA, and PERMDISP. Analyses we e done using R ( e sion 3.3.3) and package egan ( e sion2.4-3)(Oksanene al.,2017). 2.7 | S a is ical me hods, expe imen 2 Numbe o OTUs and Shannon di e si y index we e calcula ed using unc ions specnumbe and di e si y in R package egan, espec- i ely. P incipal Coo dina e Analysis was pe o med using cmdscale unc ion in R package s a s and i was based on B ay- Cu is dis ance calcula ed o bo h abundance and p esence- absence da a and using unc ion egdis in package egan. FIGURE1 Rela i e abundances o he i e mos abundan phyla in skin swab samples be o e and immedia ely a e exposu e o plan and soil ma e ials (expe imen 1). Each o he wo olun ee s es ed eigh ma e ials. In o al, 16 di e en ma e ials we e es ed. Each ba shows esul s o one indi idual sample 1Be o e A e 2Be o e A e 3Be o e A e 4Be o e A e 5 Be o e A e 6Be o e A e 7 Be o e A e 8Be o e A e 9 Be o e A e 1 0Be o e A e 1 1Be o e A e 1 2Be o e A e 1 3Be o e A e 1 4Be o e A e 1 5Be o e A e 1 6Be o e A e 0 20 40 60 80 100 P opo ion (%) Ac inobac e ia P o eobac e ia Acidobac e ia Fi micu esBac e oide es O he s 6 o 13 | GRÖNROOS e al. 0.41.2 Phylum = 5.73, p < .001 0.51.5 Class = 6.85, p < .001 0.52.0 O de = 7.91, p < .001 1.02.5 4.0 Family = 7.21, p < .001 1.03.0 Genus = 6.62, p < .001 135 OTU = 6.72, p < .001 0.01.5 3.0 Phyl. Acidobac e ia = 4.51, p < .001 0.52.0 3.5 Phyl. Ac inobac e ia = 8.09, p < .001 02 Phyl. Bac e oide es = 3.39, p = .004 1.02.5 Phyl. Fi micu es = 1.41, p = .178 13 Phyl. P o eobac e ia = 5.56, p < .001 1.53.0 Alphap o eobac e ia = 5.6, p < .001 0.52.0 Be ap o eobac e ia = 2.18, p = .045 0.52.0 Gammap o eobac e ia = 4.7, p < .001 Be o e A e Be o eA e Be o e A e Be o e A e Be o eA e Be o e A e Be o e A e Be o eA e Be o e A e Be o e A e Be o eA e Be o e A e Be o e A e Be o eA e Be o e A e 0.01.5 Del ap o eobac e ia = 3.16, p = .006 Shannondi e si y index | 7 o 13 GRÖNROOS e al. 3 | RESULTS 3.1 | Di ec hand- exposu e In he i s expe imen , wo u ban olun ee s es ed eigh soil and plan based ma e ials by ubbing he ma e ials in hei hands. Al oge he six een di e en ma e ials we e es ed. Each ma e- ial was es ed sepa a ely and skin swab sample was aken be o e and a e he exposu e. Illumina MiSeq sequencing o bac e ial 16S RNA gene showed ha he mos common bac e ial phyla in hands be o e he exposu e we e Ac inobac e ia, P o eobac e ia, and Fi micu es (Figu e 1). A e he exposu e, he ela i e abundance o Acidobac e ia and Bac e oide es inc eased. The ela i e abundance o se e al less common phyla inc eased (g oup “o he s” in Figu e 1), which also led o an inc ease in Shannon index and ichness o he phyla (Figu es 2 and S1). Shannon di e si y index (Figu e 2) and axon ichness (i.e., num- be o axa, Figu e S1) we e signi ican ly highe (p≤.045)a e expo- su e a all es ed axonomic le els as well as o all es ed axonomic g oups excep o he phylum Fi micu es. Quan i a i e PCR and Wilcoxon signed- ank es also showed ha he o al bac e ial abun- dance (i.e., numbe o bac e ial 16S copies) was signi ican ly highe a e exposu e compa ed o ha be o e exposu e (p = .001, V = 9) (Figu e 3). Bac e ial communi y composi ion in hands was signi ican ly di e en a e exposu e compa ed o ha be o e exposu e a all es ed axonomic le els and in all es ed axonomic g oups (Table 1, Figu e 4). Pe mu a ional Mul i a ia e Analysis o Va iance (PERMANOVA) showed ha he e ec o exposu e ea men (be- o e/a e ) was highly signi ican . The iden i y o pe son was also signi ican in many cases in PERMANOVA analysis, bu he in e - ac ion be ween ea men and pe son was signi ican only a he phylum le el (p=.017)and o  hephylumAc inobac e ia(p = .011). The lack o in e ac ion sugges s ha he change in communi y com- posi ion did no depend on he pe son conduc ing he expe imen . We also plo ed NMDS igu e wi h he sampling o de (Figu e S2). This showed ha al hough he ma e ials we e es ed consecu i ely (minimum o 5 h be ween wo es occasions), in mos cases, he skin bac e ial composi ion had eco e ed be ween he sampling occasions. Finally, Mul i a ia e Homogenei y o G oup Dispe sions (PERMDISP) showed ha he a ia ion in bac e ial composi ion was signi ican ly highe a e exposu e compa ed o he a ia ion be o e exposu e o he whole da a a o de , amily, genus, and OTU le els as well as o OTU le el da a wi hin Ac inobac e ia, Fi micu es, P o eobac e ia, and Gammap o eoba e ia (Table 1, Figu e 4). 3.2 | Exposu e ia ab ic packe s In he second expe imen , wo u ban olun ee s used moss- illed ab ic packe s on hei skin. Bac e ial ichness and Shannon di e - si y index we e highe in samples aken a e he use o packe s con aining he Sphagnum moss compa ed o hose aken be o e he use o  he packe s excep  o  he ichness o  ST047DIA_2 (Table S1). P incipal Coo dina e Analysis showed ha a e he expe imen , he bac e ial communi y composi ion on skin was di - e en om he samples aken be o e he expe imen (Figu e S3). In addi ion, a e he exposu e, he skin bac e ial composi ion was mo e simila o he sample o moss ma e ial han be o e he expo- su e sugges ing ha some bac e ial OTUs we e ans e ed om he packe s o he skin. 4 | DISCUSSION While p e ious e o s ha e shown ha mic obes can be ans- plan ed om one body si e o he o he (e.g., Cos ello e al., 2009), we show he e ha also mic obes in na u al ma e ials can be a ached o skin o inc ease i s mic obial di e si y. The easoning was ha he icini y o di e se mic obial communi ies has been sugges ed o educe he isk o immune- media ed diseases ( on He zen e al., 2011). The immedia e esponse ha we ound is ele an as e en a ansien change is plausibly o u mos impo ance a imes o dining and ouching lips o nos ils by hand. One po en ial way o a ec mic obial exposu e o o he o gans is o modi y skin mic obio a o hands. He e, we exposed hands wi h soil and plan ma e ials and ob- se ed a d as ic inc ease in skin mic obial di e si y. Un il now, app oaches o cu e o p e en immune sys em diso - de s using mic obe- based ea men s ha e been limi ed o he use o li ing p obio ic bac e ia (Ab ahamsson, Jakobsson, Bjö ks én, Oldaeus, & Jenmalm, 2013; Ma schan e al., 2008), inac i a ed bac- e ia (Be h- Jones e al., 2006; B o he s, Ashe , Jaksic, & S ewa , 2009),bac e ialpa s(Kline,2007),bac e ialo helmin hexc e ions (McSo ley e al., 2012) and ecal ansplan a ions. O hese me h- ods, only ecal ansplan a ion has been p o en o be e ec i e bu i is es ic ed o only a ew diseases, such as se e e Clos idium di icile dia hea and in lamma o y bowel diseases, and i is ha dly sui able o ea men s aimed a modula ing immune sys em o p e- en immune-media eddiseases(Cohen&Maha shak,2017).O he  FIGURE2 Shannon di e si y o bac e ia in hands inc eased a e exposu e o plan and soil ma e ials (expe imen 1). Resul s a e shown o he whole da a a six di e en axonomic le els (phylum, class, o de , amily, genus and OTU) as well as o OTUs wi hin i e majo phyla (Acidobac e ia, Ac inobac e ia, Bac e oide es, Fi micu es, and P o eobac e ia) and ou classes wi hin phylum P o eobac e ia. Boxplo s show medians ( hick line), uppe and lowe hinges (box), minimum and maximum alues (whiske s) and ou lie s (poin s; alues mo e han 1.5 imes he in e qua ile ange om he hinges). Da a we e no mally dis ibu ed based on Shapi o- Wilk es and hus pai ed T- es was used. Deg ees o eedom is 15 o all es s (n = 16, each o he wo olun ee s es ed eigh ma e ials) 8 o 13 | GRÖNROOS e al. app oaches a e based on exposu e o ce ain mic obes, such as ce - ain bac e ial species, hei componen s o exc e ions, o modula e immune sys em. Some s udies ha e ound hese me hods p omising (Kline, 2007; Ma schan e al., 2008; McSo ley e al., 2012) while o he s ha e ound no e ec s (Ab ahamsson e al., 2013; Be h- Jones e al., 2006; B o he s e al., 2009). O hese me hods, use o p obio ics has ecei ed subs an ial amoun o esea ch in e es and se e al s udies ha e shown ha , when applied o in an s o du ing p egnancy, hey educe he isk o a opy and alle gy, al hough con- adic o y esul s ha e also been epo ed (Ne mes, Salminen, & Isolau i, 2013). In he cu en s udy, we subs i u ed he s ain- based app oach wi h na u e- based ma e ials comp ising di e se mic obial com- muni ies. We ound a p onounced inc ease in di e si y o skin mic obio a immedia ely a e sho - e m di ec exposu e o hands o na u al soil and plan based ga dening ma e ials. Also, he o al abundance o bac e ia was highe a e exposu e han be o e expo- su e. These esul s indica e ha bac e ia in soil and plan ma e ials a e a ached o skin and emain he e e en a e washing hands wi h wa e . In he second expe imen , we es ed a mo e con e- nien way o applying he exposu e ma e ial and ound ha skin mic obial di e si y also inc eased when using ab ic packe s illed wi h d ied and c ushed moss. Toge he , hese esul s sugges ha in oducing an inc eased exposu e o na u e- based ma e ials has a po en ial o ansien ly change he skin mic obio a. These indings open new possibili ies o s udy exposu es ha could be used o modula e unc ions o he immune sys em, pa icula ly i he iming o he exposu e is op imized. In p e ious s udies, con adic ing esul s when using app oaches based on single o a ew mic obial species, migh be due o a mis- ma ch be ween indi idual disease pheno ype and he mic obial spe- cies o s ain used (Ab ahamsson e al., 2013; Ne mes e al., 2013; S iemsma e al., 2015). Thus, using ma e ials wi h na u al di e si y, such as he soil and plan based ma e ials es ed he e, migh be a mo e e ec i e app oach o enhance mic obial di e si y and he e- o e p e en and cu e immune sys em diso de s. The app oach p e- sen ed he e simula es wo king in ga den o in he ield—p ac ices p e iously common in human e e yday li e bu nowadays la gely lacking om li es o u ban ci izens. In addi ion, wi h he exposu e s a egy p esen ed he ein, he immune sys em encoun e s en i on- men al mic obial s imuli in a e y na u al ashion. We assume ha his kind o exposu e has a po en ial o s imula e he whole immune sys em including di e en ana omic loca ions such as skin, o al and espi a o y mucosa, Peye ’s pa ches, cilia in gas oin es inal ac , diges i e enzymes, and inally, he exposu e in e ac wi h di e en lymphocy ic cells. Indeed, se e al s udies ha e shown a ela ionship be ween di e si y o human mic obio a and educed isk o immune- media ed diseases (Ab ahamsson e al., 2012; Hanski e al., 2012; Manichanh, 2006; Sche e al., 2015) and u al li ing en i onmen has been shown o be posi i ely ela ed o he di e si y o human mic obio a (De Filippo e al., 2010; Schno e al., 2014; Ya sunenko e al., 2012). I is likely, ha he ue mechanism behind hese pa - e ns is no a single bac e ial species o a s ain, bu a he , a highe exposu e o na u ally di e se mic obio a. Ou esul s showed ha ouching na u al soil and plan based ma e ials inc eased, a leas empo a ily, he o al di e si y o skin mic obio a and he di e si y o phyla Acidobac e ia, Ac inobac e ia, Bac e oide es, P o eobac e ia as well as classes Alpha- , Be a- and Gammap o eobac e ia. We also ound ha o many g oups he a ia ion in communi y composi ion (i.e., be a di e si y) was highe immedia ely a e he exposu e. Because se e al di e en ma e ials we e used in he exposu e, i migh be possible o mix di e en na - u al soil and plan ma e ials and hus p oduce a composi e ma e ial ha would lead e en o a highe inc ease in skin mic obial di e si y. Ou esul s also showed ha he skin mic obio a eco e ed quickly close o he ini ial composi ion. This sugges s ha single exposu e may no be enough o p oduce long- las ing e ec s and hus e- pea ed exposu es migh be needed. The app oach p esen ed he e, is, a leas ini ially, concen a - ing on modi ying skin mic obio a. Di e se skin mic obio a can ha e di ec bene icial e ec s on human heal h (Rod igues Ho mann, 2017).Fo example,commensalmic obesinhibi g ow ho pa ho- genic mic obes by compe ing o nu ien s and space, hus educ- ing he g ow h o pa hogens (San o d & Gallo, 2013). In addi ion, a la ge numbe o commensal bac e ia di ec ly es ic he g ow h o FIGURE3 Quan i a i e PCR shows inc ease in bac e ial abundance in hands a e exposu e o soil and plan based ma e ials (expe imen 1). Six een ma e ials we e es ed and quan i a i e PCR was conduc ed o samples aken be o e (n = 16, each o he wo olun ee s es ed eigh ma e ials) and a e (n = 16) each exposu e. Fo isual easons, one ou lie in a e samples is le ou side he axis ma gins. Pai ed Wilcoxon signed- ank es o whole da a is p = .001 and V = 9. Fo da a wi hou he pai wi h he ou lie is p = .002 and V = 9. See boxplo desc ip ion in Figu e 2 Be o eA e 0e+001e+06 2e+063e+06 4e+0 6 Ou lie : 4.7e+07 Numbe o 16Scopiespe sample | 9 o 13 GRÖNROOS e al. TABLE1 Bac e ial communi y composi ion in hands changed a e exposu e o plan and soil based ma e ials (expe imen 1) PERMANOVA PERMDISP p- alue F 2 p- alue FMean be Mean a Be /A Pe son In e ac . Be /A Pe son In e ac . Be /A Pe son In e ac . Taxonomic le els Phylum 0.001 0.006 0.017 19.88 8.34 5.38 0.32 0.14 0.09 0.501 0.53 0.19 0.16 Class 0.001 0.004 0.055 18.32 6.29 2.65 0.33 0.11 0.05 0.197 1.74 0.20 0.25 O de 0.001 0.009 0.093 18.31 4.92 2.03 0.34 0.09 0.04 0.012 7.01 0.21 0.32 Family 0.001 0.001 0.093 14.80 5.81 2.08 0.29 0.11 0.04 0.035 4.60 0.30 0.41 Genus 0.001 0.003 0.086 13.22 5.65 1.89 0.27 0.12 0.04 0.029 5.42 0.32 0.44 OTU 0.001 0.002 0.085 8..68 5.55 1.78 0.20 0.13 0.04 0.001 11.74 0.34 0.51 OTU le el wi hin Phyla/Class Ac inobac e ia 0.001 0.001 0.011 9.63 7.39 2.97 0.20 0.15 0.06 0.02 6.77 0.38 0.49 Bac e oide es 0.001 0.007 0.524 2.21 1.69 0.95 0.07 0.05 0.03 0.145 2.18 0.63 0.67 Fi micu es 0.009 0.005 0.394 4.36 4.83 0.90 0.11 0.13 0.02 0.048 4.55 0.30 0.45 P o eobac e ia 0.001 0.023 0.536 8.45 3.18 0.77 0.21 0.08 0.02 0.006 8.32 0.31 0.49 Alphap o eobac e ia 0.001 0.693 0.861 9.10 0.71 0.59 0.24 0.02 0.02 0.944 0.01 0.49 0.49 Be ap o eobac e ia 0.001 0.003 0.473 4.48 2.53 0.95 0.12 0.07 0.03 0.745 0.11 0.59 0.57 Gammap o eobac e ia 0.002 0.009 0.492 4.34 3.71 0.86 0.12 0.10 0.02 0.006 9.50 0.29 0.51 Resul s o pe mu a ional mul i a ia e analysis o a iance (PERMANOVA) wi h wo ac o s (be o e/a e , s udy pe son and hei in e ac ion) and esul s o he analysis o mul i a ia e homogenei y o g oup dispe sions (PERMDISP). B ay- Cu is dissimila i y was used. Analyses we e no done o phylum Acidobac e ia and class Del ap o eobac e ia because se e al samples did no include any bac e ia belonging o hese wo axa. Fo PERMDISP, also he mean alues o he dis ances o g oup medians a e shown. Each o he six een ma e ials was es ed once by one o he wo s udy pa icipan s. P- alues less han 0.05 a e highligh ed in bold.