Metatranscriptomics reveals the active bacterial and eukaryotic fibrolytic communities in the rumen of dairy cow fed a mixed diet

micb-08-00067 Janua y 28, 2017 Time: 11:52 # 1 ORIGINAL RESEARCH published: 31 Janua y 2017 doi: 10.3389/ micb.2017.00067 Edi ed by: Da id Be y, Uni e si y o Vienna, Aus ia Re iewed by: Seungha Kang, Commonweal h Scien i ic and Indus ial Resea ch O ganisa ion (CSIRO), Aus alia Le Luo Guan, Uni e si y o Albe a, Canada *Co espondence: E elyne Fo ano [email p o ec ed] †P esen add ess: Nicolas Pa iso , UMR203 BF2I, Uni Lyon, INSA-Lyon, INRA, Villeu banne, F ance; Pascale Lepe cq, LISBP-INSA Toulouse, Toulouse, F ance ‡Deceased Special y sec ion: This a icle was submi ed o Mic obial Symbioses, a sec ion o he jou nal F on ie s in Mic obiology Recei ed: 01 July 2016 Accep ed: 10 Janua y 2017 Published: 31 Janua y 2017 Ci a ion: Com e -Ma e S, Pa iso N, Lepe cq P, Chauchey as-Du and F, Mosoni P, Pey e aillade E, Baya AR, Shing ield KJ, Pey e P and Fo ano E (2017) Me a ansc ip omics Re eals he Ac i e Bac e ial and Euka yo ic Fib oly ic Communi ies in he Rumen o Dai y Cow Fed a Mixed Die . F on . Mic obiol. 8:67. doi: 10.3389/ micb.2017.00067 Me a ansc ip omics Re eals he Ac i e Bac e ial and Euka yo ic Fib oly ic Communi ies in he Rumen o Dai y Cow Fed a Mixed Die Sophie Com e -Ma e1, Nicolas Pa iso 2†, Pascale Lepe cq1†, F édé ique Chauchey as-Du and1,3, Pascale Mosoni1, E ic Pey e aillade2, Ali R. Baya 4, Ke in J. Shing ield4,5‡, Pie e Pey e 2and E elyne Fo ano1* 1UR454 Uni é de Mic obiologie, INRA, Sain -Genès-Champanelle, F ance, 2EA4678 CIDAM, Cle mon Uni e si é, Uni e si é d’Au e gne, Cle mon -Fe and, F ance, 3Lallemand Animal Nu i ion, Blagnac, F ance, 4Nu i ional Physiology, G een Technology, Na u al Resou ces Ins i u e Finland (Luke), Jokioinen, Finland, 5Ins i u e o Biological, En i onmen al and Ru al Sciences, Abe ys wy h Uni e si y, Abe ys wy h, UK Ruminan s ha e a unique abili y o de i e ene gy om he deg ada ion o plan polysaccha ides h ough he ac i i y o he umen mic obio a. Al hough his p ocess is well s udied in i o, knowledge gaps emain ega ding he ela i e con ibu ion o he mic obio a membe s and enzymes in i o. The p esen s udy used RNA-sequencing o e eal bo h he exp ession o genes encoding ca bohyd a e-ac i e enzymes (CAZymes) by he umen mic obio a o a lac a ing dai y cow and he mic oo ganisms o ming he ibe -deg ading communi y. Func ional analysis iden i ied 12,237 CAZymes, accoun ing o 1% o he ansc ip s. The CAZyme p o ile was domina ed by amilies GH94 (cellobiose-phospho ylase), GH13 (amylase), GH43 and GH10 (hemicellulases), GH9 and GH48 (cellulases), PL11 (pec inase) as well as GH2 and GH3 (oligosaccha idases). Ou da a suppo he pi o al ole o he mos cha ac e ized ib oly ic bac e ia (P e o ella,Ruminocccus and Fib obac e ), and highligh a subs an ial, al hough mos p obably unde es ima ed, con ibu ion o ungi and cilia e p o ozoa o polysaccha ide deg ada ion. Pa icula ly hese esul s may mo i a e u he explo a ion o he ole and he unc ions o p o ozoa in he umen. Mo eo e , an impo an pa o he ib oly ic bac e ial communi y emains o be cha ac e ized since one hi d o he CAZyme ansc ip s o igina ed om dis an ly ela ed s ains. These indings a e used o highligh limi a ions o cu en me a ansc ip omics app oaches o unde s and he unc ional umen mic obial communi y and oppo uni ies o ci cum en hem. Keywo ds: umen, ibe deg ada ion, glycoside hyd olases, ca bohyd a e es e ases, polysaccha ide lyases, me a ansc ip omics INTRODUCTION The umen ha bo s an amazing di e si y o mic oo ganisms, comp ising p oka yo es (bac e ia, a chaea) and euka yo es (p o ozoa, ungi), which co e essen ial unc ions o hei hos . Pa o hese mic oo ganisms a e specialized in he deg ada ion o plan polysaccha ides and he eby cons i u e a pi o al communi y p o iding a supply o ene gy o he hos animal (Hobson and S ewa , 1997;Whi e e al., 2014). O e all, he genomes o ib oly ic mic oo ganisms ha bo 100s o genes encoding ca bohyd a e ac i e enzymes (CAZymes) (Be g Mille e al., 2009;Pu ushe e al., 2010;Suen e al., 2011;Yousse e al., 2013), F on ie s in Mic obiology | www. on ie sin.o g 1Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 2 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion mainly glycoside hyd olases (GH), ca bohyd a e es e ases (CE), and polysaccha ide lyases (PL) which ac syne gis ically o decons uc die a y cellulose, hemicellulose, s a ch, and pec in (Can a el e al., 2009;Flin e al., 2012). Fib obac e succinogenes, Ruminococcus la e aciens, and Ruminococcus albus a e among he i s celluloly ic bac e ia isola ed om he umen and ha e long been conside ed o play a majo ole in ibe deg ada ion due o hei p e alence in uminan s and abili y o solubilize e icien ly plan cell wall ma e ial in i o (Kobayashi e al., 2008; Flin e al., 2012). Ne e heless, ecen e o s o isola ing no el umen celluloly ic bac e ia (Cai e al., 2010;Dodd e al., 2011; Nyonyo e al., 2014;Zieme , 2014) and analysis o he di e si y o bac e ial CAZymes h ough me agenomic s udies (B ulc e al., 2009;Hess e al., 2011;Wang e al., 2013), ha e p o ided e idence ha bac e ia, o he han he h ee mos ex ensi ely s udied, a e also in ol ed. Bac e ia a e usually conside ed as he main ib oly ic mic oo ganisms in he umen because o hei p edominance in his ecosys em. In con as , umen ungi a e hough o be only mino con ibu o s o plan deg ada ion because o a low biomass in he umen, al hough hey p oduce enzymes wi h a e y high speci ic ac i i y (Wilson and Wood, 1992;O pin and Joblin, 1997). P o ozoa a e no essen ial o su i al o he hos animal bu hei emo al om he umen o sheep and ca le has been shown o lowe eed deg ada ion in he umen and dec ease eed con e sion e iciency in se e al s udies (Newbold e al., 2015). None heless, he mode o ac ion emains unclea as p o ozoa may con ibu e di ec ly h ough he sec e ion o ib oly ic enzymes o indi ec ly by c ea ing a o able condi ions o ib oly ic bac e ia in he umen (Jouany, 2006). Cu en ly, me a ansc ip omics is conside ed a eliable app oach o in es iga ing me abolically ac i e mic obial communi ies which a e no necessa ily he dominan ones (To s ik and Ø eås, 2002). The i s published umen me a ansc ip omic s udy was es ic ed o he euka yo ic communi y om he muskoxen umen (Qi e al., 2011). Recen ly, Dai e al. (2015) and Shinkai e al. (2016) conduc ed a me a ansc ip omic su ey o he ibe -a ached mic oo ganisms in he umen o ca le con i ming he ac i e ib oly ic s a us o well-known dominan bac e ial deg ade s. Ne e heless, he ela i e con ibu ion o euka yo es o uminal ibe deg ada ion was no in es iga ed ex ensi ely in hese s udies. In he p esen s udy, RNA sequencing was used o p o ide a unique insigh in o he con ibu ion o me abolically ac i e umen mic oo ganisms o he deg ada ion o ou plan polysaccha ides (cellulose, hemicellulose, pec in and s a ch) in he umen o a lac a ing cow ed a mixed die ep esen a i e o comme cial a ms. The main objec i e o his s udy was o deciphe he exp ession o genes encoding CAZymes o unde s and be e how he ib oly ic communi y ac s in he umen in i o and o iden i y po en ial mic obial con ibu o s o his pi o al unc ion. RNA was isola ed om a sample o o al umen con en s, a he han solids, o in es iga e bo h ibe -a ached and ee loa ing plank onic popula ion. Compa ed wi h p e ious me a ansc ip omic s udies (Shinkai e al., 2016), he p esen wo k p o ides a de ailed analysis o bo h euka yo es and p oka yo es o he CAZyme ansc ip s. In addi ion o CAZyme ansc ip s, we also e ie ed sequences known o be in ol ed in ibe -deg ading sys ems such as cellulosomes o Polysaccha ide U iliza ion Loci (PUL; Flin e al., 2012). Da a gene a ed we e used o highligh some o he limi a ions o high- h oughpu app oaches o in es iga e he ac i i y o he umen mic obial popula ion and oppo uni ies o ci cum en hem based on sequencing o o al RNA and a mRNA-en iched RNA ac ion ob ained using an in- house p ocedu e. Taxonomic analysis o he umen mic obio a was based on small subuni (ssu) RNA om he o al RNA, ollowed by unc ional and axonomic anno a ion o pu a i e mRNA om he mRNA-en iched RNA ac ion. Ou esul s show ha he well-cha ac e ized CAZyme amilies and ib oly ic bac e ia a e he majo con ibu o s o polysaccha ide deg ada ion in he umen o a cow ed a mixed die , bu also unde sco ed ha he con ibu ion o euka yo es is unde es ima ed. RESULTS AND DISCUSSION Insigh in o he Ac i e Rumen Communi y Based on RNA and Pu a i e mRNA Taxonomic mining o mic obial communi ies by small subuni RNA analysis has been widely applied o deciphe ac i e membe s. He e we used 16S and 18S RNA sequences om he o al RNA sample o highligh he ac i e p oka yo es (bac e ia and a chaea) and euka yo es (p o ozoa and ungi) o he umen. Bac e ia ep esen ed he majo i y o he umen ssu RNA eads (77.5%; Figu e 1). They exhibi ed a high di e si y ha included FIGURE 1 | Taxonomic epa i ion o he ac i e popula ion o he umen mic obial communi y de i ed om RNA sequences om he me a ansc ip omic da a (ou e ing) and he ela i e abundances o main mic obial g oups as measu ed by qPCR a he DNA and RNA le el (inne ings). The same colo code is used o bo h da a. F on ie s in Mic obiology | www. on ie sin.o g 2Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 3 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion 23 iden i ied phyla (Supplemen a y Table S1) o which only 6 we e de ec ed a a ela i e abundance abo e 1% (i.e., Fi micu es, Bac e oide es, Fib obac e es, P o eobac e ia, Spi ochae ae, and Len isphae ae; Table 1). While Li e al. (2016) ound he phylum P o eobac e ia ep esen ing be ween 5 and 90% o bac e ia ssu in o al RNA, i ep esen ed only 5.5% in he animal s udied he e. A he amily le el, P e o ellaceae (20%), Ruminococcaceae (13.2%), Lachnospi aceae (9.1%), and Fib obac e aceae (6.8%) we e he mos abundan in he bac e ial popula ion (Table 1). While a chaea con ibu ed o only 0.7% o o al ssu eads, euka yo es accoun ed o 21.8% (Table 1). Sequences om euka yo es we e domina ed by he In amac onuclea a subphylum (90.4% o he euka yo a) associa ed o he umen cilia e p o ozoa and he Neocallimas igomyco a phylum (3.3%) co esponding o he umen anae obic ungi (Table 1). O he euka yo ic amilies pe ained o he phylum Exca a a eg ouping some lagella e p o ozoa (Hampl e al., 2009) o which he commensal s a us is no con i med because none ha e been ex ensi ely cha ac e ized (Fon y and Chauchey as-Du and, 2007). The high ela i e con ibu ion o he In amac onuclea a subphylum o he ssu RNA eads, con i med by RT-qPCR, con as s wi h i s low ela i e abundance on a DNA basis (Table 2), leading o a ssu RNA:DNA a io o 542. This sugges s ha p o ozoa we e e y ac i e in he umen o cow unde he speci ied condi ions o his expe imen (die con aining 50:50 g ass silage:concen a es on a d y ma e basis and sampling o umen con en s be o e mo ning eeding). Ne e heless, his inding canno be con i med using axonomic binning o non- RNA eads, whe e pu a i e mRNA TABLE 1 | Taxonomic ep esen a ion o he ac i e umen mic obio a based on small subuni (ssu) RNA sequence analysis om RNA-sequencing da a ob ained wi h o al RNA. Supe kingdom Phylum Family/Subphylum Bac e ia (77.5%) Fi micu es (33.2%) Ruminococcaceae (13.2%) Lachnospi aceae (9.1%) Ch is ensenellaceae (4.8%) Bac e oide es (30.1%) P e o ellaceae (20.0%) BS11 gu g oup (1.7%) Rikenellaceae (1.6%) Unclassi ied Bac e oide es (1.7%) RF16 g oup (1.2%) Fib obac e es (6.8%) Fib obac e aceae (6.8%) P o eobac e ia (5.5%) Succini ib ionaceae (4.3%) Spi ochae ae (2.3%) Spi ochae aceae (2.2%) Len isphae ae (1.1%) RFP12 gu g oup (0.4%) Vic i allaceae (0.2%) BS5 g oup (0.2%) WCHB1-25 g oup (0.1%) A chaea (0.7%) Eu ya chaeo a (100%) Me hanobac e iaceae (49.7%) Me hanomassiliicoccaceae (50.3%) Euka yo a (21.8%) SAR g oup (96.7%) In amac onuclea a (90.4%) Opis hokon a (3.3%) Neocallimas igomyco a (2.7%) Only phyla iden i ied wi h a ela i e abundance highe han 1% and co esponding amilies a e epo ed. A he supe kingdom le el, ela i e abundances (in pa en heses) a e exp essed as a pe cen age o o al ssu RNA eads, whe eas o phyla and amilies, hese alues a e epo ed as a pe cen age o ssu RNA eads in he co esponding supe kingdom. TABLE 2 | Quan i ica ion and ela i e abundances o main popula ions o umen mic oo ganisms e alua ed by qPCR and RNA-sequencing. qPCR RNA-Seq DNA cDNA ssu RNA/DNA a io Rela i e abundance Copies/µg DNA ±SD Rela i e abundance Copies/µg cDNA ±SD Rela i e abundance Bac e ia (16S) Bac e ia (16S) 1.20 ×1010 ±3.53 ×10999.56% 5.35 ×1010 ±2.05 ×10976.07% 4.46 77.73% Me hanogenic a chaea (16S) Eu ya cheo a (16S) 2.12 ×107±2.88 ×1060.18% 1.20 ×108±1.41 ×1030.17% 5.66 0.68% P o ozoa (18S) In amac onuclea a (18S) 3.08 ×107±1.17 ×1070.25% 1.67 ×1010 ±1.06 ×10623.74% 542.2 20.45% Neocallimas igomyco a (ITS1) Neocallimas igomyco a (18S) 1.47 ×106±4.63 ×1050.01% 1.20 ×106±1.34 ×1010.02% 0.82 0.61% Rela i e abundances a e gi en as a pe cen age o o al ssu copies (qPCR) o o al ssu eads (RNA-Seq). Fo qPCR esul s, each alue is he mean (n =6) ±s anda d de ia ion. The p ecise denomina ion o mic obial g oups and he gene a ge ed a e indica ed in he i s column, in i alics o RNA-Seq da a. F on ie s in Mic obiology | www. on ie sin.o g 3Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 4 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion o umen p o ozoa accoun ed o only 0.70%. Low da abase comple ion may bias ou o e all iew o he ac i e mic obiome as ew e e ence genomes and genes om umen euka yo es a e cu en ly a ailable. Indeed no umen p o ozoal genome and only one ungal genome (O pinomyces sp. C1A, Yousse e al., 2013) a e a ailable. Recen ly, Jiang e al. (2016) highligh ed ha he umen mic obiome lacks su icien ep esen a i e e e ence genomes o pe o m obus sea ches o sequence simila i ies. We hypo hesized ha sequences o igina ing om he gu o o he mammals could ex end he anno a ion o umen me a ansc ip omics da a compa ed wi h gene alized da abases, such as he NCBI NR da abase. We compa ed BLAST esul s ob ained wi h NR and wo e sions o he human gu mic obio a gene ca alog om he Me aHIT p ojec as e e ence da abases (Qin e al., 2010;Li e al., 2014). BLAST based simila i y sea ch using he NR da abase esul ed in 38.8% signi ican BLAST hi s, which we e only inc eased by a u he 4.5% when he Me aHIT da abases we e used, wi h he bes esul s ob ained wi h Me aHIT V3 (Supplemen a y Figu e S1A). E alua ion o he bi sco e o he bes hi s which e lec s he quali y o he alignmen indica ed ha he NCBI NR da abase p o ided only a ound 50% o he bes bi sco es (Supplemen a y Figu e S1B). Thus, a ailable gene ca alogs om o he mammals p o ide only a limi ed enhancemen o anno a ion o sequenced da a. Fu he mo e, axonomic anno a ion o sequences om mammalian ca alogs is ex emely limi ed and did no allow comple e axonomic binning o pu a i e mRNA in he p esen wo k. Consequen ly, gene ca alog speci ic o umen is eally needed especially o umen euka yo es ha do no ha e ep esen a i es in cu en gu gene ca alogs. Taxonomic binning o pu a i e mRNA was only picked om BLAST esul s ob ained wi h he NCBI NR da abase. Based on a lowes common ances o (LCA) analysis, pu a i e mRNA we e p ima ily ela ed o bac e ia (93.7% o eads wi h a BLAST hi (e- alue <1e- 05), euka yo a (3.0%), a chaea (Eu ya cheo a; 1.5%), and i uses (0.05%). Rega ding bac e ia, he phyla o Bac e oide es (37.6%), Fi micu es (38.9%), Spi ochae es (3.6%), P o eobac e ia (3.3%), and Fib obac e es (2.9%) we e he mos ac i e, wi h an abundance o mo e han 1% o pu a i e mRNA. A he amily le el, P e o ellaceae (19.3%), Lachnospi aceae (5.3%), Ruminococcaceae (5.0%), Spi ochae aceae (3.0%), Bac e oidaceae (2.7%), Fib obac e aceae (2.8%), and Clos idiaceae (1.3%) we e among he mos ac i e mic oo ganisms. O e all, bo h app oaches based on RNA and non- RNA analysis esul ed in he same ep esen a ion o dominan ac i e bac e ia in he uminal communi y. Howe e , abou hal o he eads we e assigned o axonomic anks highe han amily using he LCA me hod. The LCA assignmen s o high-le el axa a e usually associa ed o conse ed sequences among se e al axa, and hus could bias he esul s. In con as o he analysis o RNA in which euka yo es ep esen ed a qua e o he ac i e mic obio a (Figu e 1), only a low numbe o eads a ibu ed o euka yo es (3.02%) we e e ie ed. Sequences we e mainly associa ed wi h umen p o ozoa (Oph yoscolecidae; 0.70%), Amoebozoa (0.15%), umen ungi (Neocallimas igaceae; 0.11%), and Ascomyco a (0.02%), whils he emaining euka yo ic axa we e disca ded due o e oneous iden i ica ion o sequence sou ces in he da abase. The unc ional binning o pu a i e mRNA (based on KEGG o hology) indica ed ha he main unc ions exp essed by he umen mic obio a we e p ima ily ela ed o me abolism (47.6% o assigned eads) and gene ic in o ma ion p ocessing (14.4%). O he unc ional assignmen s we e linked o bac e ial and i al i ulence ac o s associa ed o human in ec ious diseases (7.5%), en i onmen al in o ma ion p ocessing (5.4%), o ganismal sys ems (4.9%, mos ly plan -pa hogen in e ac ion), and cellula p ocesses (4.7%). Wi hin me abolic unc ions, ca bohyd a e me abolism was he mos impo an (16.1% o o al assigned eads), ollowed by ene gy me abolism (10.1%), whe eas o he me abolic p ocesses accoun ed o less han 7% o o al assigned eads. Me abolic unc ions, me abolism o ca bohyd a es in pa icula , a e usually he mos exp essed by gu mic obio a (Gosalbes e al., 2011;Dai e al., 2015;Lee e al., 2015) unde lining he impo ance o his ac i i y in gu ecosys ems. O e iew o CAZymes Exp essed by he Rumen Mic obio a Fo mally, he “CAZyme” e m emb aces all CAZymes and hei associa ed non-ca aly ic ca bohyd a e-binding modules (CBM), in ol ed in he syn hesis o deg ada ion o complex ca bohyd a es (Can a el e al., 2009). While glycosyl ans e ase (GT) ansc ip s we e ound in he me a ansc ip ome (3816 eads) o he o al umen con en , hese we e no aken in o accoun since hey we e ela ed o ca bohyd a e syn hesis and no deg ada ion. Simila ly, amilies o auxilia y ac i i ies (AA), mainly ep esen ed by lignin deg ading enzymes, de ec ed wi h only 161 eads in o al (9 amilies) we e no u he analyzed (Supplemen a y Table S2). In he p esen analysis, emphasis was placed on he sea ching o ca aly ic domains o GH, CE, and PL ( u he designa ed as CAZymes in he manusc ip ) which a e key enzymes in ol ed in he decons uc ion o polysaccha ides. A o al o 12,237 sequences o CAZymes, including GH (10,209), CE (1,404), and PL (624) we e e ie ed, ha collec i ely ep esen ed abou 1.0% o non- RNA eads. Based on he LCA analysis, a ound 41% o non- RNA eads we e ela ed o 89 di e en gene a, and he emaining eads we e linked o highe axonomic le els (Supplemen a y Table S3). CAZyme eads pe ained p edominan ly o bac e ia om he phyla Bac e oide es (43.8%; P e o ella sp., 18.8%; Bac e oides sp., 3.5%), Fi micu es (23.4%; Ruminococcus sp., 7%; Clos idium sp., 0.8%), Fib obac e es (4.6%; Fib obac e sp., 4.6%), and Spi oche es (0.9%; T eponema sp., 0.8%), wi h lowe numbe s assigned o euka yo es (5.4%; Neocallimas ix sp., 0.3%; Pi omyces sp., 0.15%; Epidinium sp., 0.3%, and Polyplas on sp., 0.3%; Supplemen a y Table S3). Di e si y o hese CAZymes was ela i ely high gi en ha 99 GH (ou o he 135 lis ed in he CAZy da abase), 14 CE (ou o 16), and 13 PL (ou o 23) amilies we e iden i ied, wi h a b oad ange o subs a es anging om complex polysaccha ides o oligosaccha ides (Supplemen a y Table S2). O e all, de ec ed CAZymes we e domina ed by amilies known o exhibi enzyma ic ac i i ies o hyd olysis o plan polysaccha ides (hemicellulose, cellulose, pec in, and F on ie s in Mic obiology | www. on ie sin.o g 4Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 5 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion s a ch; Supplemen a y Table S2), consis en wi h isola ion o umen con en om a cow ed a die con aining bo h g ass silage and ce eals as sou ces o plan cell walls and s a ch (Baya e al., 2015). Majo CAZyme ansc ip s encoded polysaccha idases (GH13, 9, and 10 o example) o oligosaccha idases (GH2 and GH3), he la e o en being he mos ep esen ed in umen me agenomes (B ulc e al., 2009;Hess e al., 2011). The CAZyme p o ile epo ed he e di e s subs an ially om wo ecen epo s on he umen me a ansc ip ome (Dai e al., 2015;Shinkai e al., 2016). Fo example, he GH13 amily (amylases) was consis en ly ound in he me a ansc ip omes o wo cows o ep esen abou 20% o he GH (Dai e al., 2015), bu only 6% in he p esen analysis. Simila ly, he p o ile o hemicellulases also di e ed be ween he p esen and p e ious s udies (Shinkai e al., 2016). Se e al ac o s may con ibu e o hese appa en disc epancies including animal die , me hodology (RNA ex ac ion), and sou ce o umen mic obiome (sampling si e and ime), as well as he ela i e abundance o plank onic and a ached communi ies in ol ed in ibe deg ada ion. Al hough ew a ia ions we e obse ed be ween he wo samples om he same die in he p e ious me a ansc ip ome o Dai e al. (2015), animal a iabili y could also be in ol ed. Dis inc domains can be associa ed o he ca aly ic domains men ioned abo e. Typically, CBM enable o po en ia e and op imize enzyma ic hyd olysis by main aining he ca aly ic domain nea he subs a e (Abbo and an Bue en, 2014). A o al o 1,780 pu a i e CBM we e e ie ed and assigned o 52 amilies, wi h CBM37, CBM50, CBM4, CBM48, and CBM6 amilies being he mos abundan (Supplemen a y Table S2). While CBM50 is common wi hin bac e ia1, CBM37 (260 eads in his me a ansc ip ome) has only been ound in Ruminococcus albus and hough o acili a e adhesion o he bac e ium o he subs a e (Eze e al., 2008;Rako oa i onina e al., 2009). The de ec ion o pu a i e mRNA con aining docke in and cohesin domains sugges s ha cellulosomes a e employed by umen mic oo ganisms in i o (Supplemen a y Table S2). Cellulosomes ep esen mul i-enzyma ic machine ies ha enable he ac i i y o se e al ib oly ic enzymes o ac syne gis ically (Flin e al., 2012). The de ec ion o susC and susD homologs also unde lines he use o PUL by he umen mic obio a (Supplemen a y Table S2). These gene ic loci ha e been epo ed in se e al umen me agenomes (Pope e al., 2012) and p oposed as a mechanism o polysaccha ide decons uc ion in he umen (Naas e al., 2014). In he p esen me a ansc ip omics da a, hei occu ence was highe han cellulosome associa ed domains (docke ins and cohesins; Supplemen a y Table S2). Deg ada ion o Plan Polysaccha ides in he Rumen o a Cow Fed a Mixed Die : CAZymes and Mic obial Communi ies In ol ed The expe imen al cow ecei ed a mixed die , simila o comme cial condi ions, con aining 23% o acid de e gen ibe 1www.cazy.o g (comp ising cellulose and lignin) and 17% o hemicellulose, eaching o 40% o ibe , among which 32% was po en ially diges ible. The die con ained also 12% o s a ch and 3.3% suga . In acco dance wi h he die composi ion, cellulase and hemicellulase CAZyme amilies p edomina ed, wi h 26.5 and 43.2% o CAZyme eads, espec i ely. Cellulases we e ep esen ed by 13 amilies (Figu e 2 and Supplemen a y Table S2), mainly encoding endoglucanases and cellobiohyd olases, excep GH1 and GH3 (in majo i y β-glucosidases) and GH94 (cellobiose/cellodex in phospho ylase). The po en ial o igin o endo-ac ing cellulase ansc ip s was mos ly om he bac e ial gene a Ruminococcus and Fib obac e , ungi (Neocallimas ix sp., Pi omyces sp., O pinomyces sp.), and p o ozoa (Epidinium sp., Polyplas on sp.) while oligosaccha idases (GH1, GH3, and GH94) o igina ed mainly om P e o ella sp. (Figu e 3 and Supplemen a y Table S3). GH94 is no usually epo ed as abundan in he umen me agenome, bu was he mos exp essed amily in he p esen me a ansc ip ome. Cellobiose phospho ylase ac i i y has been demons a ed o umen s ains o Ruminococcus,P e o ella, and Fib obac e (Wells e al., 1995;Lou e al., 1996, 1997). P esen esul s highligh he possibili y ha phospho yla ion could be a common mechanism o he deg ada ion o oligosaccha ides eleased du ing cellulose hyd olysis. Pu a i e mRNA encoding endo-ac ing cellulases om GH9, GH48, and GH5 amilies, ha a e he mos widely s udied in umen mic obiology, we e among he mos abundan in his da ase . Al hough GH9 and GH5 sequences we e o e - ep esen ed in umen me agenomic da a (B ulc e al., 2009;Hess e al., 2011;Wang e al., 2013), GH48 sequences ha e been always iden i ied a low equency. Indeed, genes encoding hese amilies a e usually ound in high copy numbe s in celluloly ic bac e ial and ungal genomes, excep ha GH48 genes a e p esen in monocopies in bac e ia1. In he p esen s udy, exp ession o GH48-encoding genes was a he high (3.0% o CAZyme ansc ip s in ou sample), consis en wi h ea lie epo s (Dai e al., 2015), p o iding u he suppo o GH48 enzymes playing an essen ial ole in plan cell wall deg ada ion. Su p isingly, GH48 ansc ip s we e no iden i ied in Shinkai e al. (2016) me a ansc ip ome. Cellulases om GH48 a e pi o al enzymes in bac e ial (Clos idium he mocellum,Ruminococcus la e aciens) and ungal cellulosomes (S eenbakke s e al., 2002;Baye e al., 2004). In ou da a, he majo i y o GH48 ansc ip s we e om bac e ia (89.4%; Supplemen a y Table S3) and mainly a ilia ed o Ruminococcus (Supplemen a y Table S3), and o a lesse ex en , umen ungi (Neocallimas igomyco a). In e es ingly he majo cellulase genes exp essed by Fib obac e in he p esen bo ine umen sample encode GH45 enzymes (Supplemen a y Figu e S2). F. succinogenes S85 genome con ains 4 GH45 genes while o he cellulase genes a e ound in highe numbe s1(12 GH5, 9 GH9, and 6 GH8). This sugges s ha F. succinogenes GH45 genes we e exp essed a high le els in he umen sample analyzed p esen ly. Only a small numbe o GH45 genes ha e been iden i ied in bac e ia1and ew o hem ha e been cha ac e ized (Gilbe e al., 1990;Seon Pa k e al., 2007). F. succinogenes GH45 enzymes me i u he in es iga ion gi en he likelihood o F on ie s in Mic obiology | www. on ie sin.o g 5Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 6 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion FIGURE 2 | Rela i e abundances o ansc ip s encoding ca bohyd a e-ac i e enzymes (CAZymes) amilies wi h known ac i i ies in ol ed in he b eakdown o cellulose, hemicelluloses, pec ins, and s a ch. Families o endo-ac ing enzymes which may play a pi o al ole in hese p ocesses a e highligh ed by ha ches. GH, glycoside hyd olase; CE, ca bohyd a e es e ase; PL, polysaccha ide lyases. FIGURE 3 | Pu a i e mic oo ganisms in ol ed in he b eakdown o cellulose, hemicelluloses, pec ins, and s a ch. Taxonomic binning o ansc ip s encoding CAZymes amilies wi h known ac i i ies in ol ed in he b eakdown o each polysaccha ide ype was pe o med using he Lowes Common Ances o algo i hm, leading o a di e en axonomic ank (gene a a e highligh ed by do s). Rela i e abundances we e calcula ed om he o al o CAZymes a ge ing each plan polysaccha ide. ha ing a majo ole in ibe deg ada ion by his bac e ium in i o. Wi h espec o hemicellulose hyd olysis, pu a i e mRNA encoding enzymes implica ed in he deg ada ion o xylans, xyloglucans, and mannans we e e ie ed o 32 GH amilies and 8 CE amilies (Figu e 2 and Supplemen a y Table S2). None heless, only 13 amilies we e obse ed wi h a ela i e abundance g ea e han 1%, accoun ing o 32.4% o CAZymes. These amilies include endoxylanases (GH43, GH10, GH11, GH51, GH5, GH26, and GH8), mannanases (GH5 and GH26), oligosaccha idases as well as deb anching enzymes (GH2, GH43, GH3, GH51, GH95, GH36, and GH8), and e uloyl and ace yl xylan es e ases (CE1). The high ela i e abundance o ansc ip s encoding he amilies GH10 and GH11 (exclusi ely comp ised o endoxylanases) wi hin he CAZyme me a ansc ip ome unde lines hei pi o al ole in xylan deg ada ion in i o. T ansc ip s encoding hese wo amilies p incipally o igina ed om p o ozoa (24.1% o he GH10-GH11 ansc ip s), Ruminococcus sp. (13.8%), Fib obac e sp. (8.0%), and P e o ella sp. (7.1%) (Supplemen a y Table S3). Fu he mo e, GH26, a amily o mannanases, was ela i ely en iched in he p esen da ase and mos ly ela ed o he same bac e ial gene a. The high exp ession o genes encoding GH43 in he p esen analysis as well as ansc ip omic da a om pu e s ains g own on xylan (Dodd e al., 2010;Sawhney e al., 2015) sugges ed an F on ie s in Mic obiology | www. on ie sin.o g 6Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 7 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion impo an ole in hemicellulose deg ada ion. Indeed, GH43 amily con ains enzymes wi h a ious speci ici ies including oligosaccha idases and deb anching enzymes2(β-xylosidase, α-L- a abino u anosidase, a abinanase). Deb anching enzymes a e key componen s o he enzyma ic machine y o hemicelluloly ic mic oo ganisms ha acili a e he access o endo-ac ing enzymes o hei subs a e. Due o hei in ima e associa ion wi h cellulose and hemicellulose, deg ada ion o pec ins may accele a e he p ocess o comple e plan cell wall decons uc ion. Cu en ly, e y sca ce me a ansc ip omic da a o pec in deg ada ion a e a ailable. P e ious umen me a ansc ip omes iden i ied a ew PL eads, bu he ull a ay o enzymes implica ed in pec in deg ada ion was no analyzed in de ails (Shinkai e al., 2016; Dai e al., 2015). Con e sely, 14 CAZymes (6 GH amilies, 6 PL amilies and 2 CE) in ol ed speci ically in pec in deg ada ion we e iden i ied in his da ase , ep esen ing 12.6% o o al CAZyme eads. Pu a i e mRNA encoding endo- and exo-ac ing enzymes om he amilies PL11 (2.2% o CAZymes), GH28 (1.6%), PL1, and PL9 (∼1.0%), and deb anching enzymes om he amily CE8 (1.2%) we e he mos abundan (Figu e 2 and Supplemen a y Table S2). They ha e been exclusi ely assigned o bac e ia, mainly o Bac e oide es (46.3%; P e o ella sp., 21.5%; Bac e oides sp., 6.2%), and he Ruminococcus (7.4%) and Fib obac e (6.6%) gene a (Figu e 3 and Supplemen a y Table S3). S ains o hese gene a can deg ade and u ilize pec in suga s, excep o Fib obac e succinogenes ha canno u ilize p oduc s o pec in hyd olysis (G adel and Deho i y, 1972;Osbo ne and Deho i y, 1989;Ma ounek and Dusko a, 1999;Dongowski e al., 2000). The subs an ial p opo ion o pec inase ansc ip s in ou da ase may be due o he p esence o pec ins in he die o he animal (less han 3%), bu also o co ansc ip ion o pec inase and cellulase/hemicellulase genes by umen mic oo ganisms, as shown p e iously o Ruminococcus la e aciens by ansc ip ome analysis (Be g Mille e al., 2009). Polygalac u onase and pec a e lyase genes ha e been iden i ied by me agenomics analysis o he umen mic obial communi y in a sheep (Yuan e al., 2012), in which pec inase genes we e a ilia ed o Bu y i ib io, P e o ella, Bac e oides, and Fib obac e consis en wi h he p esen s udy, o he han e y ew eads close o Bu y i ib io sequences we e de ec ed. I appea s ha pec inoly ic ac i i y is mainly associa ed o he bac e ial gene a known o be ac i e on he main plan cell wall polysaccha ides, i.e., cellulose and hemicelluloses. None heless, ce ainly due o limi a ions in da abases, a hi d o he ansc ip s associa ed o pec in deg ada ion we e no axonomically a ilia ed in he p esen analysis (Figu e 3 and Supplemen a y Table S3), aising he p ospec ha o he axa comp ising umen ungi (Kopeˇ cný and Hod o á, 1995) may also be in ol ed in uminal hyd olysis o pec in polyme s. Simila o o he plan polysaccha ides, o al hyd olysis o s a ch equi es he conce ed ac ion o se e al enzymes comp ising deb anching enzymes, endo- and exo-amylases. In he p esen s udy, wi h a die con aining 12% o s a ch, ansc ip s o amyloly ic enzymes we e ound in 11 GH amilies 2www.cazy.o g (Figu e 2 and Supplemen a y Table S2) and accoun ed o 16.9% o CAZyme eads. They we e ela ed o Bac e oide es (43.0%; P e o ella, 24.1%; Bac e oides, 3.8%), Fi micu es (17.6%; Clos idium, 2.4%; Ruminococcus, 1.6%), and Gammap o eobac e ia (1.4%) (Figu e 3 and Supplemen a y Table S3). The main cul i a ed amyloly ic umen s ains belong o P e o ella and Bu y i ib io species, S ep ococcus bo is, and Selenomonas uminan ium (Fon y and Chauchey as-Du and, 2007). These esul s demons a e ha P e o ella is he majo playe in he s udied cow umen, bu Bac e oides and Clos idium may also be signi ican amyloly ic gene a, al hough e y ew, i any, umen isola es om hese gene a ha e been cha ac e ized. P o ozoa unc ion as amyloly ic mic oo ganisms in he umen (Williams and Coleman, 1997), and wi h a die con aining 12% s a ch we expec ed o ind mo e ansc ip s a ilia ed o p o ozoal amylases. None heless, hey we e de ec ed a only a low ela i e abundance (GH13; 0.1%). Only 2 p o ozoal GH13 a e a ailable in he NR da abase. Gi en ha iden i ica ion o eads depends la gely on sequences a ailable in da abases, a high p opo ion o p o ozoal amylase sequences a e mos p obably no e ie ed and emain unassigned. Con ibu ion o Euka yo es o Fib oly ic P ocesses in he Rumen The con ibu ion o euka yo es o uminal ibe deg ada ion in i o has no been ully elucida ed. Me a ansc ip omics is he e o e a ele an app oach o add ess his knowledge gap. CAZyme ansc ip s po en ially o igina ing om euka yo es ha e been in es iga ed (Qi e al., 2011), bu he da a gene a ed did no pe mi o e alua e he abundance o ansc ip s om euka yo es ela i e o bac e ia. Shinkai e al. (2016) did no de ec euka yo ic ansc ip s du ing he analysis o a cow umen sample. In con as , Dai e al. (2015) epo ed ha euka yo es can con ibu e o 18% o cellulase (GH5, GH6, GH9, GH44, GH45, and GH48) and 3.5% o hemicellulase (GH8, GH10, GH11, GH26, GH28, GH51, GH53, GH67, and GH78) ansc ip s. Ne e heless, abou 20% and 40% o cellulase and hemicellulase-encoding ansc ip s wi h as low as 30 o 60% o iden i y wi h known sequences we e included, aising he likelihood o alse posi i es. In he p esen s udy, CAZyme ansc ip s om euka yo ic o igin and encoding he same GH amilies ep esen ed 20.5 and 15.2%, espec i ely, wi h a leas 60% iden i y wi h a known sequence. In e oga ion o CAZyme amilies o which euka yo ic sequences we e a ailable (Figu e 4), indica ed ha ungal sequences con ibu ed up o 10% o cellulase (GH9 and GH48) and hemicellulase (GH11) amilies. The GH6 amily appea ed o be speci ic o ungi. The mos no able euka yo ic con ibu ion was de ec ed o GH11, wi h p o ozoal eads ep esen ing up o 48% o o al GH11 ansc ip s. P o ozoa also con ibu ed signi ican ly o cellulase and hemicellulase sequences om he GH5, GH9 and GH10 amilies. Taking in o accoun he low numbe o euka yo ic CAZyme sequences in da abases3and he p opo ion o p o ozoa and ungi in he umen mic obial popula ion (a ound 0.00001%), he abundance o pu a i e ansc ip s encoding CAZyme amilies om euka yo ic o igin suppo s he hypo hesis ha , unde ou 3h p://www. mgne wo k.o g/hunga e1000.h ml F on ie s in Mic obiology | www. on ie sin.o g 7Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 8 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion FIGURE 4 | Rela i e con ibu ion o bac e ia, p o ozoa, and ungi o pu a i e mRNA encoding CAZyme amilies o which sequences o euka yo es a e a ailable. Abundances we e calcula ed ela i e o he o al CAZyme (GH+CE+PL) eads. These da a indica e ha euka yo es, p o ozoa in pa icula , may con ibu e subs an ially o hemicellulose and cellulose deg ada ion in he umen. expe imen al condi ions, euka yo es con ibu ed signi ican ly o uminal ibe deg ada ion. Such a con ibu ion is p obably unde es ima ed as many o he p o ozoal and ungal CAZyme sequences a e cu en ly una ailable. The ole o p o ozoa in ib olysis has long been con o e sial, as hey ha e been p ima ily conside ed as e y ac i e p o ein and s a ch deg ade s (Williams and Coleman, 1997). The e y high RNA/ DNA a io obse ed o p o ozoa (Table 2), sugges ing hei high ac i i y, ecommends ha p o ozoa unc ion and ecology should dese e u he in es iga ions. E idence o he Con ibu ion o Uncha ac e ized and Unknown Bac e ial Communi ies o Fib oly ic P ocesses in he Rumen Using a bi sco e h eshold o 90 o he LCA analysis, he p esen analysis concen a ed on sequences ha sha ed he highes iden i y wi h known e e ences o cha ac e ize wi h a high deg ee o ce ain y he axonomic con ibu ion o he umen ib oly ic communi y. Using his app oach, i was possible o con i m he impo an ole o well-s udied umen bac e ia o deg ada ion o he ibe s included in ou expe imen al die (i.e., P e o ella sp., 18.7% o ibe deg ading CAZymes; Ruminococcus sp., 5.6%, and Fib obac e sp., 4.0%). In gene al he CAZyme exp ession p o ile co ela ed well wi h he abili y o deg ade cellulose, hemicellulose, pec in o s a ch in i o (Supplemen a y Figu e S2) (Bé a-Maille e al., 2000;Be g Mille e al., 2009;Dodd e al., 2010;Yousse e al., 2013;Bu ne e al., 2015;Couge e al., 2015). O he bac e ia may well pa icipa e, gi en ha sequences o Bac e oides sp. (3.4%), Clos idium sp. (0.8%), T eponema sp. (0.5%), and Bu y i ib io sp. (0.4%) we e also iden i ied (Supplemen a y Table S3). While hese gene a a e equen ly e ie ed om umen me agenomes (Pa el e al., 2014), ew uminal isola es a e a ailable o biochemical cha ac e iza ion. A hi d o he sequences e ie ed in he p esen umen me a ansc ip ome may o igina e om mo e dis an ly ela ed s ains o unknown mic oo ganisms, since 26% o eads wi h as low as 30% o iden i y wi h a known sequence we e no assigned using MEGAN (bi sco e <90) and 3% did no esul in a signi ican BLAST hi (e- alue >1e-05) (Supplemen a y Table S3). Gene a ion and a ailabili y o mo e genomes is equi ed o ad ance knowledge on he umen ib oly ic communi y. Fo his eason, he use o d a genomes econs uc ed om a me agenome o swi chg ass-a ached bac e ia (Hess e al., 2011) ha may well include ib oly ic bac e ia was also in es iga ed. Using only 15 genomes, 1.7% o he CAZyme eads we e e ie ed wi h an iden i y highe han 90% (Supplemen a y Table S4). Genomes ha esul ed in imp o emen o simila i y sea ch we e ela ed mainly o Bac e oidales (72.1%) and Clos idiales (20.6%) (Supplemen a y Table S4). CONCLUSION E en hough sho -gun me agenomics app oaches ha e been success ully applied in he pas o deciphe he ib oly ic po en ial o he umen mic obio a, me a ansc ip omics app oaches o e he oppo uni y o in es iga e he ela i e con ibu ion o me abolically ac i e membe s o he umen mic obial communi y. Cu en ly, only wo me a ansc ip omes ocusing on he umen ib oly ic communi ies ha e been published (Dai F on ie s in Mic obiology | www. on ie sin.o g 8Janua y 2017 | Volume 8 | A icle 67 micb-08-00067 Janua y 28, 2017 Time: 11:52 # 9 Com e -Ma e e al. Rumen Me a ansc ip omics and Fibe Deg ada ion e al., 2015;Shinkai e al., 2016). Using di e en expe imen al condi ions (mixed die , sampling be o e mo ning eeding), he p esen s udy comple es he a ailable snapsho s o he key ea u es o uminal ibe deg ada ion by con i ming he main con ibu o s o his unc ion, and gene a ing new in o ma ion highligh ing he ole o euka yo ic mic oo ganisms. Despi e a di e si ied CAZyme epe oi e ha bo ed by he genome o umen ib oly ic mic oo ganisms, ew enzyme amilies we e highly exp essed in i o. Thei high ela i e abundance sugges s an ac i e con ibu ion o he ib oly ic p ocess in he animal s udied. P esen da a highligh ed pi o al amilies in ol ed in he in i o decons uc ion o cellulose (GH9, GH48, GH5, and GH94), hemicellulose (GH10, GH11, and GH43), pec in (PL11 and GH28), and s a ch (GH13). Illus a ing he s eng h o me a ansc ip omic app oaches, GH amilies which a e ypically nea -absen in me agenomics da a such as GH48 and GH94 we e among he mos ep esen ed in CAZyme ansc ip s wi hin ou da ase . Howe e , i should be pinpoin ed ha ansc ip s quan i ica ion does no always e lec exac ly he enzyma ic ac i i ies p oduced as many pos - ansc ip ionnal egula ions can occu in he mic oo ganisms. O he wise i is highly p obable ha he CAZyme p o ile is dependen on he ela i e amoun s o o age and concen a e, and he ype o plan polysaccha ides in he hos uminan die . Also, he p o ile o he CAZyme ansc ip s may e ol e du ing diges ion o he meal and he coloniza ion o inges ed eed pa icles. Thus, he CAZyme exp ession pic u e ou lined he e mus be conside ed as a snapsho o he ib oly ic communi y o one cow a a single ime poin . Fu u e s udies should expand he p esen app oach o gi e a mo e comple e pic u e o ibe deg ada ion in he umen, o example by moni o ing CAZyme ansc ip s h oughou he diges ion p ocess, o by compa ing he deg ada ion ac i i y o di e en managemen sys ems. As ound by Dai e al. (2015), bac e ia om he mos s udied gene a o umen ib oly ic s ains (P e o ella,Ruminoccocus, Fib obac e ) we e ound o be among he mos ac i e in he s udied cow umen, unde lying ha p e ious cul u e-based analyses ha e ce ainly p o ided a aluable o e iew o he majo bac e ial ac o s in he umen. The main o iginali y o he p esen s udy is o use wi h g ea ca e me a ansc ip omics da a o assess he ela i e con ibu ion o euka yo es o polysaccha ide b eakdown, highligh ing he ac ha his app oach could o e come many o he cu en limi a ions o in i o and in i o s udies. Euka yo es, and especially p o ozoa, we e ound o con ibu e signi ican ly o he exp ession o he CAZyme amilies GH9, GH10, GH48, GH5, and GH11 despi e a low abundance in he umen and he limi ed numbe o e e ence sequences in a ailable da abases. These indings sugges ha he me abolic ac i i y o p o ozoa, and maybe also ungi, is mo e impo an in cellulose and hemicellulose deg ada ion han p e iously hough , a con ibu ion likely o be unde es ima ed due o he lack o gene ca alogs. These esul s may mo i a e u he explo a ion o he ole and he unc ions o p o ozoa in he umen. O e all he complexi y o ecology and unc ion o he umen mic obio a leads o huge challenges in acqui ing sequence da a. The lack o sequenced genomes s ill ep esen s a emendous ba ie o iden i y genes, especially CAZymes, and many o hem emained as pu a i e in ou s udy. We assessed oppo uni ies o ci cum en analy ical limi a ions (use o he human gu me agenome ca alog and d a genomes econs uc ed om me agenomic da a). Un o una ely, enhancemen o sequence anno a ion was e y limi ed. Comple ion o he umen mic obiome ca alog as well as be e anno a ion o genomes is eally undamen al o be e unde s and he umen mic obio a. MATERIALS AND METHODS Animal Expe imen , Die , and Sampling All expe imen al p ocedu es we e app o ed by he Na ional E hics Commi ee (Hämeenlinna, Finland) in acco dance wi h he guidelines es ablished by he Eu opean Communi y Council Di ec i e 86/609/EEC (Eu opean Union, 1986). The analyzed umen sample was ob ained om a lac a ing dai y cow ed wi h a o al mixed a ion ( o age:concen a e a io 50:50, on a d y ma e con en ) based on g ass silage. De ails o he die ed ha e been epo ed p e iously (Baya e al., 2015). In b ie , he o al mixed a ion con ained (g/kg d y ma e ) neu al de e gen ibe (401), acid de e gen ibe (228), and s a ch (120). Samples we e aken be o e he mo ning eeding om i e di e en si es in he umen, composi ed, and mixed ho oughly o ob ain a ep esen a i e sample o umen con en s. Fo molecula analysis, umen con en was subsampled (50 g) and mixed wi h 100 mL o RNAla e (The mo Fishe Scien i ic, Wal ham, MA, USA) o p e en RNA deg ada ion. The mix u e was main ained o e nigh a +4◦C and hen s o ed a –80◦C un il nucleic acid ex ac ion. DNA and RNA Ex ac ion om Rumen Sample DNA and RNA we e ex ac ed om he RNAla e -p ese ed sample ollowing ho ough mixing a e hawing. DNA isola ion was pe o med in iplica e om 250 mg o cen i uged umen con en acco ding o Baya e al. (2015). DNA samples we e assessed o pu i y and quan i y using a NanoD op 1000 spec opho ome e (The mo Fishe Scien i ic, Wal ham, MA, USA), and s o ed a –20◦C un il molecula analysis. To al RNA was ex ac ed in iplica e om 400 mg o cen i uged umen con en using T izol (The mo Fishe Scien i ic, Wal ham, MA, USA). In b ie , samples we e homogenized wi h ∼100 mg o 0.1 mm zi conia beads (BioSpec P oduc s, Ba les ille, OK, USA) and 1.2 mL o T izol using he Fas P ep-24 ins umen (MP Biomedicals, I ine, CA, USA). All subsequen s eps we e in acco dance wi h he ecommenda ions om he manu ac u e . RNA p ecipi a ion was pe o med using a mix o isop opanol and a saline solu ion (NaCl 1.2 M, disodium ci a e 0.8 M) as epo ed by Bé a-Maille e al. (2009). To al RNA was subjec ed o DNase ea men (Nucleospin DNase se , Mache ey-Nagel, Dü en, Ge many). The in eg i y o RNA was assessed wi h he Agilen 2100 Bioanalyze using he RNA Nano Chip (Agilen Technologies, San a Cla a, CA, USA). Quan i y o RNA was assessed wi h he Nanod op 1000 spec opho ome e (The mo Fishe Scien i ic, Wal ham, MA, USA). T iplica es o o al RNA F on ie s in Mic obiology | www. on ie sin.o g 9Janua y 2017 | Volume 8 | A icle 67