The cellular growth rate controls overall mRNA turnover, and modulates either transcription or degradation rates of particular gene regulons

Published online 29 Decembe 2015 Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 3643–3658 doi: 10.1093/na /gk 1512 The cellula g ow h a e con ols o e all mRNA u no e , and modula es ei he ansc ip ion o deg ada ion a es o pa icula gene egulons Jos´ eGa c ´ ıa-Ma ´ ınez1,2,†, Lidia Delgado-Ramos3,4,†, Guille mo Ayala5, Vicen Pelechano6, Daniel A. Medina2,7, Fany Ca asco2,7,Ram´ on Gonz´ alez8, Edua do And ´ es-Le´ on3, La s S einme z6,9,10, Jonas Wa inge 11, Sebas i´ an Ch´ a ez3,4,* and Jos´ eE.P ´ e ez-O ´ ın2,7,* 1Depa amen o de Gen´ e ica, Facul ad de Ciencias Biol´ ogicas, Uni e si a de Val` encia. C/D . Moline 50. E46100, Bu jasso , Spain, 2ERI Bio ecmed, Facul ad de Ciencias Biol´ ogicas, Uni e si a de Valencia. C/D . Moline 50. E46100, Bu jasso , Spain, 3Ins i u o de Biomedicina de Se illa (IBiS), Hospi al Vi gen del Roc´ ıo-CSIC-Uni e sidad de Se illa, C/An onio Mau a Mon ane , E41013 Se illa, 4Depa amen o de Gen´ e ica, Uni e sidad de Se illa, A enida de la Reina Me cedes s/n, E41012, Spain, 5Depa amen o de Es ad´ ıs ica e In es igaci´ on Ope a i a, Facul ad de Ma em´ a icas, Uni e si a de Val` encia. C/D . Moline 50. E46100, Bu jasso , Spain, 6Eu opean Molecula Biology Labo a o y (EMBL), Genome Biology Uni , Meye ho s asse 1, 69117 Heidelbe g, Ge many, 7Depa amen o de Bioqu´ ımica y Biolog´ ıa Molecula , Facul ad de Ciencias Biol´ ogicas, Uni e si a de Valencia. C/D . Moline 50. E46100, Bu jasso , Spain, 8Ins i u o de Ciencias de la Vid y del Vino (CSIC, Uni e sidad de La Rioja, Gobie no de La Rioja), Finca La G aje a LO-20 Salida 13, Au o ´ ıa del Camino de San iago, E26007 Log o˜ no, Spain, 9S an o d Uni e si y School o Medicine, Depa men o Gene ics, S an o d, CA 94305, USA, 10S an o d Genome Technology Cen e , 3165 Po e D . Palo Al o, CA 94305, USA and 11Depa men o Chemis y and Molecula Biology, Uni e si y o Go henbu g, Medicina ega an 9 c, 40530 G¨ o ebo g, Sweden Recei ed No embe 04, 2015; Re ised No embe 30, 2015; Accep ed Decembe 16, 2015 ABSTRACT We analyzed 80 di e en genomic expe imen s, and ound a posi i e co ela ion be ween bo h RNA poly- me ase II ansc ip ion and mRNA deg ada ion wi h g ow h a es in yeas . Thus, in spi e o he ma ked a ia ion in mRNA u no e , he o al mRNA con- cen a ion emained app oxima ely cons an . Some genes, howe e , egula ed hei mRNA concen a ion by uncoupling mRNA s abili y om he ansc ip ion a e. Ribosome- ela ed genes modula ed hei an- sc ip ion a es o inc ease mRNA le els unde as g ow h. In con as , mi ochond ia- ela ed and s ess- induced genes lowe ed mRNA le els by educing mRNA s abili y o he ansc ip ion a e, espec i ely. We also de ec ed hese egula ions wi hin he he - e ogenei y o a wild- ype cell popula ion g owing in op imal condi ions. The ansc ip omic analysis o so ed mic ocolonies con i med ha he g ow h a e dic a es al e na i e exp ession p og ams by modu- la ing ansc ip ion and mRNA decay. The egula ion o o e all mRNA u no e keeps a cons an a io be ween mRNA decay and he dilu- ion o [mRNA] caused by cellula g ow h. This eg- ula ion minimizes he indisc imina e ansmission o mRNAs om mo he o daugh e cells, and a o s he esponse capaci y o he la e o physiological signals and en i onmen al changes. We also con- clude ha , by uncoupling mRNA syn hesis om de- cay, cells con ol he mRNA abundance o hose gene egulons ha cha ac e ize as and slow g ow h. INTRODUCTION Cells mus adap o changing en i onmen al condi ions o main ain i ness and o compe e wi h o he geno ypes du - ing he na u al selec ion p ocess. As i ness is he ne g ow h o a geno ype o e ime, one o he mos impo an a i- ables o single cell o ganisms du ing he cou se o adap a- ion is he g ow h a e (GR) o a popula ion composed only o one geno ype in ela ion o he g ow h a e o o he iso- genic popula ions. I is o en assumed ha highe popula- ion g ow h a es in mic oo ganisms equi e highe p o ein syn hesis a es (1–5). This is because p o eins cons i u e a *To whom co espondence should be add essed. Tel: +34 963543467; Fax +34 963544635; Email: jose.e.pe ez@u .es Co espondence may also be add essed o Sebas i´ an Ch´ a ez. Tel: +34 955923127; Fax +34 955 923 101; Email: scha [email protected] †These au ho s con ibu ed equally o his wo k as i s au ho s. C The Au ho (s) 2015. Published by Ox o d Uni e si y P ess on behal o Nucleic Acids Resea ch. This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by-nc/4.0/), which pe mi s non-comme cial e-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. Fo comme cial e-use, please con ac jou nals.pe [email protected] 3644 Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 la ge ac ion o d y mass in bo h p oka yo es (6) and eu- ka yo es (e.g. yeas (7)) and, he e o e, hei syn hesis is he mos ene ge ically demanding p ocess (8). This p ocess is no exclusi e o unicellula o ganisms and p obably as - p oli e a ing umo cells a e also go e ned by hese ules (9,10). The ansla ion machine y includes he mos abundan noncoding RNAs: RNA and RNAs. Thus he euka y- o ic RNA polyme ases (RNA pol) de o ed o he syn he- sis o RNA and RNA (RNA pol I and III) mus inc ease hei ansc ip ion a es (TRs, see he explana ion abou he ac onyms used in M&M) in pa allel o he GR (11). RNA pol II, howe e , ansc ibes a much la ge se o mRNAs subjec ed o mul iple egula o y in luences, many o which lowe in concen a ion wi h he GR (12). So al hough a la ge se o mRNAs, which a e highly ansc ibed, is also de- o ed o ibosome biosyn hesis and ansla ion ac o s (8), he ela ionship be ween he GR and RNA pol II TR is no ob iously p edic able. Unlike RNAs and RNAs, which a e s able molecules, mRNAs ha e a sho e hal -li e, and bo h hei syn hesis and decay signi ican ly con ibu e o egula e hei abundance (13). The RNA pol II ansc ip- ion a e, especially o single cell o ganisms, is highly a i- able and a p oxy o hei physiology and me abolism (3). The in luence o he cell cycle (14) and cell size (15)on he ansc ip ion a e has been in es iga ed in he model yeas Saccha omyces ce e isiae. Cell cycle leng h and cell size a e ela ed o he GR (15). T ansc ip ome dependence on he yeas GR has been ho oughly in es iga ed in D. Bo - s ein’s labo a o y (16–19). These au ho s ha e ound ha he mRNA amoun /concen a ion (RA) o abou one hi d o he genes a ies in a GR-dependen manne . In ac based on an RA signa u e o 72 genes, i is possible o p edic he ins an aneous GR wi h easonable accu acy (17). The [mRNA] o a gi en gene is con olled by he ac ion o wo opposi e o ces: syn hesis by RNA pol II and deg ada- ion by se e al exonuclease pa hways (20). In he exponen- ial phase, he GR is a dis inc i e ea u e o any yeas s ain and en i onmen al condi ion. In his phase, mos mRNAs a e in a s eady s a e (21), in which syn hesis and deg ada ion a es (DR) a e iden ical. Syn hesis is a ze o-o de eac ion, whe eas deg ada ion is a i s -o de eac ion ha depends on a kine ic deg ada ion cons an (kd) and he mRNA con- cen a ion. The e o e, TR =DR =kd·RA whe e kdis a pa icula ea u e o each mRNA species and is in e sely ela ed wi h he mRNA hal -li e (HL): kd=ln2/HL. These simple equa ions allow any o he h ee kine ic pa- ame e s (RA, TR, o kd) o be de i ed om he o he wo (22). Gi en he appea ance o se e al genome-wide echniques o measu e he syn hesis and deg ada ion a es and he mRNA concen a ions, in se e al model o ganisms, i is possible o quan i y he espec i e con ibu ions o an- sc ip ion and deg ada ion o he ac ual concen a ion o each indi idual mRNA (13,23), and also o he sum o hem all. To do his, we used mul iple da a se s o di e - en mu an s and al e ed he key pa hways in gene egu- la ion and unde di e en g ow h condi ions ob ained by ou sel es and o he s o es he ela ionship o he h ee ki- ne ic pa ame e s wi h he g ow h a e a global and single gene le els. We ound ha TR and DR gene ally ended o inc ease when g owing as e , and in such a way ha he o al mRNA abundance o mos genes emained indepen- den o he GR in a yeas cell. Howe e , some genes, which co esponded o pa icula p o ein unc ions, showed in- c eases and/o dec eases exclusi ely on one side o he equi- lib ium, which p o oked changes in he ne abundance o speci ic mRNAs. Wi h an inc easing GR, he abundance o he ansla ion- ela ed mRNAs and he s ess-induced mRNAs inc eased and dec eased, espec i ely, due o an- sc ip ion in bo h cases. The abundance o he mi ochond ia- ela ed mRNAs also dec eased wi h an inc easing g ow h a e, which led o a highe p opo ion o e men a ion o e espi a ion a a highe GR, e en in he same high glucose medium. Howe e , his was caused by inc eased mRNA des abiliza ion, and no by educed ansc ip ion. We con i med he la e by a o ally di e en expe imen- al app oach. We ook ad an age o he in insic he e o- genei y o he GR wi hin wild- ype yeas popula ions un- de op imal en i onmen al condi ions (24) ose upanew in-house de eloped p ocedu e o allow he sepa a ion o yeas mic ocolonies acco ding o hei GR om a popula- ion o gene ically iden ical cells. When compa ing mRNA exp ession le els, we ound ha slow- and as -p oli e a ing subpopula ions exhibi ed di e en ial ansc ip omic signa- u es. This sugges s he exis ence o dis inc exp ession p o- g ams ha a e closely linked o he GR and a e independen o he en i onmen . The mi ochond ia- and espi a ion- ela ed genes we e o e exp essed in he cells om slowe g owing mic ocolonies. The PUF3 gene, which encodes an RNA-binding p o ein (RBP) ha egula es he s abili y and ansla ion o many mRNAs ela ed wi h mi ochon- d ial unc ions, was also o e exp essed a a low g ow h a e in bo h he mu an collec ion and slow-g owing mic o- colonies. Slowe g owing mic ocolonies we e speci ically en- iched in long mRNA iso o ms wi h Pu 3-binding si es a hei 3UTRs. We p opose ha he g ow h a e in luences gene exp es- sion by ac ing on bo h sides o he [mRNA] equilib ium: syn hesis and deg ada ion. The impo ance o hese wo op- posi e p ocesses is dis inc o di e en classes o gene unc- ions. In pa icula o espi a ion- ela ed mRNAs, Pu 3 ex- e s a s abilizing in luence and p omo es espi a ion a a low GR in a manne ha is independen o ex e nal glucose con- cen a ions. MATERIALS AND METHODS Me a-analysis o expe imen al da a se s In he i s me a-analysis we used se e al S. ce e isiae an- sc ip omic da a se s (published and unpublished) ob ained by us. These da a included 42 da a se s o nascen an- sc ip ion a es and 21 da a se s o mRNA amoun s, bo h ob ained by he GRO p o ocol (25). All hese da a se s we e ob ained om exponen ially g owing yeas popula- ions. Di e ences in GRs we e due o dis inc cul u e con- di ions (ca bon sou ce, empe a u e), and also o he single dele ions o indi idual genes. In 34 cases, he TR o a e - e ence sample was simul aneously de e mined. The nascen TR and RA alues o indi idual genes we e measu ed as a ios wi h ega d o a ele an e e ence sample: a wild- ype (mos ly BY4741, see Supplemen a y Table S1) s ain Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 3645 a 28◦C in YPD (yeas ex ac 1%, pep one 1%, glucose 2%) medium. In his way, da a can be compa ed ac oss expe i- men s. Indi idual alues we e summed o ob ain he global es ima es o all he genes. Fo he o al TR (RNA pol I + II + III), he o al adioac i i y inco po a ed du ing a un-on expe imen was used. A ecen ly published se o 44 yeas mu an s was also used, in which he au ho s employed uni o m cul u e con- di ions: exponen ial phase cells g own a 30◦CinYPD(26) o which he ma u e syn hesis a e (TR), RA and kdwe e de e mined acco ding o he cDTA p o ocol. Since di e en mu an s ha e dis inc cell olumes, he TR and RA alues we e co ec ed by di iding each TR and RA gene alue by each mu an /w old o cell olume compa ed o i s wild- ype s ain. The da a o he di e en yeas mu an s and hei wild ype we e ob ained om a ious sou ces (27,28). Nascen TR and ma u e mRNA syn hesis a e ep esen di e en aspec s o he same phenomenon (22). In his pa- pe , o he sake o simplici y, we used he ac onym TR o e- e o he molecula p ocess o RNA syn hesis by any RNA polyme ase. Mos o he ime, we e e only o he RNA pol II TR, o he wise we indica e wha we e e o. In GRO ex- pe imen s TR is calcula ed as nascen ansc ip ion a e bu in he cDTA me hod (26) a ma u e mRNA syn hesis a e is de e mined by in i o RNA labeling. The e m syn hesis a e is used when alking abou chemical equilib ium (as in he equa ion desc ibed in he In oduc ion sec ion); whe e i e e s o he a e o change o he ma u e mRNA concen a- ion in he cy oplasm (22). The syn hesis a e can be in e ed om nascen TR da a by assuming ha a ixed pe cen age o nascen mRNA molecules each he cy oplasm. G ow h a e es ima ions Fo all ou expe imen al condi ions, we ob ained he GR by g owing 50 ml o yeas cul u es in 250-ml lasks wi h shaking (190 pm) a he desi ed empe a u e. Aliquo s we e aken e e y 30 min in he exponen ial phase and hei OD600 ( om 0.05 o 0.7) we e measu ed. The GR (in h−1)in he exponen ial phase was calcula ed om g ow h cu es (Sup- plemen a y Table S1A). Simila ly o he 44 yeas mu an s used om Sun e al. (26), he GRs o 38 o hese s ains we e calcula ed om he g ow h cu es ob ained by mic o- cul i a ion in a Biosc een C eade , as desc ibed elsewhe e (29) (Supplemen a y Table S1B). The exponen ial phase GRs we e ex ac ed om g ow h cu es and a e aged o e eplica es (n =2) (Supplemen a y Table S1A), as p e iously desc ibed (30). Expe imen se and da a no maliza ion We employed da a se s om many di e en expe imen s ob- ained om a ious pla o ms. The exp ession alues ac- qui ed om hese expe imen s we e no malized by means o quan ile no maliza ion (31), and we e implemen ed in o he unc ion No malizeBe weenA ays o he Limma R package (32). Co ela ion analysis be ween pa ame e s and he g ow h a e: global alues The global endencies be ween di e en pa ame e s and he GR o ou da a se s and o hose o Sun e al. (26)we ecal- cula ed by plo ing he o al RNA pol II ansc ip ion, and also o he mRNA concen a ion and he mRNA deg ada- ion cons an (kd). They we e all ep esen ed as being ela- i e o hei wild- ype o e e ence condi ion. The obse ed Pea son’s co ela ion coe icien , , and he P- alue o he s a is ically signi ican de ia ion om he null hypo hesis o no co ela ion ( =0) we e calcula ed. A mul iple eg ession model was also applied (see s a is ical appendix in Supple- men a y In o ma ion). En ichmen analyses o gene ca ego ies Fo each sample in hese expe imen s, we had he co- a iable GR and he co esponding exp ession p o ile pe gene. The GR-gene exp ession co a ia ion was quan i ied using a modi ied e sion o Pea son’s co ela ion ( o de- ails, see R. Tibshi ani, G. Chu, B. Na asimhan and J. Li, 2011, SAM, Signi icance Analysis o Mic oa ays. R package e sion 2.0. h p://CRAN.R-p ojec .o g/package= sam ) be ween bo h alues. A gene- o-gene analysis o he di e en ial exp ession was pe o med by he Signi icance Analysis o Mic oa ays (SAM) me hod (33,34) wi h a alse disco e y a e o q <0.05. A gene se en ichmen analysis was applied o p e iously de ec ed se s o di e en ially ex- p essed genes. A unila e al Fishe ’s exac es was applied whe e he gene se s o be compa ed we e he Gene On ology g oups. Analyses we e un wi h R packages (R. Gen leman, wi h con ibu ions om S. Falcon and D. Sa ka , 2014, Ca - ego y Analysis. R package e sion 2.32.0 and M. Ca lson, 2014), o g.Sc.sgd.db: Genome wide anno a ion o Yeas . R package e sion 3.0.0). The whole R code used in his pape is ound as Supplemen a y Ma e ial. A second en- ichmen analysis was un by he Gene Se Analysis (GSA) me hod (B. E on and R. Tibshi ani, 2010 GSA:, Gene se analysis. R package e sion 1.03. h p://CRAN.R-p ojec . o g/package=GSA)(35). A de ailed s a is ical p o ocol is gi en as a s a is ical appendix in Supplemen a y In o ma- ion. Gene on ology ep esen a ions The genes om he global expe imen se s o all he pa- ame e s (TR, RA, and kd) om bo h da a sou ces we e anked acco ding o hei co ela ion wi h he GR. The o e - ep esen ed unc ional ca ego ies (Gene On ology) ob- ained in he SAM and GSA analyses we e educed and i- sualized wi h he Re iGO web se e ((36)h p:// e igo.i b. h /). Calcula ing he espi a o y quo ien (RQ) index The p o ocol desc ibed in Qui ´ os e al. (37) was used wi h he ollowing modi ica ions. RQ de e mina ions we e made in MiniBio bio eac o s (250 ml nominal olume) (App- likon, Schiedam, The Ne he lands) equipped wi h Pel ie - e ige a ed gas condense s, illed wi h YPD b o h (200 ml). Cul u es we e spa ged wi h ai a a gas low o 600 ml/h. 3646 Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 The gas low was con olled wi h MFC17 mass low con- olle s (Aalbo g, O angebu g, NY, USA), whose calib a- ion was e i ied egula ly by a soap bubble lowme e . The ins an CO2and O2concen a ions in he exhaus gas we e eco ded e e y minu e in a BlueInOne gas analyze (Blue- sens, He en, Ge many). The measu ed pe cen ages o bo h gases we e con e ed in o ins an p oduc ion o consump- ion a es (exp essed in moles o gas pe li e o cul u e hou ly) by aking in o accoun he con ibu ion o CO2p o- duc ion and O2consump ion o he o al gas low. The accu- mula ed p oduc ion o consump ion o each gas was de e - mined as he in eg al o e ime o he ins an alues. The ime poin ha co esponded o 0.15 moles o CO2p o- duced pe li e o cul u e was chosen as a way o s anda d- ize he RQ alues o compa isons. A his ime poin , mos o he ini ial suga was al eady consumed, hus he co e- sponding RQ alues (RQ 0.15) would easonably summa- ize all he cul u e s ages. RQ 0.15 was calcula ed as he a io be ween he accumula ed amoun s (exp essed in moles) o he p oduced CO2and he consumed O2 o his ime poin . Cell mic oencapsula ion and mic ocolony assay The mic oencapsula ion o indi idual yeas cells in algi- na e mic osphe es (no mo e han one cell pe mic opa icle) was pe o med using a Cellena mic oencapsula o (Ingeni- a ics), as desc ibed elsewhe e (38). Encapsula ed cells we e incuba ed oge he in YPD medium unde s anda d cul u - ing condi ions. A e e hanol ixa ion, mic ocolonies we e analyzed and so ed using a BioSo e la ge-pa icle low cy ome e (Union Biome ica). A de ailed p o ocol o he so ing p ocedu es is desc ibed in (39). T ansc ip omic analysis o mic ocolonies The o al RNA ex ac ed om he so ed mic ocolonies and om he non-so ed con ols was sequenced ollowing he 3T- ill p o ocol as p e iously desc ibed (40). The s a ing ma e ial used was 500–3000 ng o o al RNA. The aw nex gene a ion sequencing eads o be i s as- sessed o hei quali y wi h he FASTQC ool ki Bad qual- i y eads (ph ed sco e <20) we e immed. The genome se- quence o S. ce e isiae S288c ( e sion R64) and i s anno- a ions we e e ie ed om SGD (Saccha omyces Genome Da abase, h p://yeas genome.o g/) and used o all he analyses. Raw eads we e aligned agains he e e ence genome using BWA/Bow ie2. We quan i ied he gene ex- p ession om he mapped eads using he HTSeq-coun package (41) in o de o ob ain coun s o mapped eads pe gene (in he ‘in e sec ion-nonemp y’ mode). As he coun p ocess was dependen on sequencing dep h, hose samples wi h a small numbe o eads we e emo ed, whe eas hose samples wi h a huge numbe o eads we e down-sampled using he Fas qSample u ili y om he Sho ead package (42). The s a is ical en i onmen R ( e sion 3.1; (43)), was used o pe o m he s a is ical analysis. Hie a chical clus e - ing, boxplo s, and Mul idimensional Scaling (MDS) analy- ses we e pe o med be o e and a e no maliza ion o mea- su e he di e ences among samples. The edgeR package om R/BioConduc o (44) was used o no malize and i coun da a o he di e en ial gene exp ession analysis. In o de o a oid a bias in poo ly exp essed genes, hose genes wi h less han i e mapped eads pe million in a leas wo samples we e emo ed. An iso o m-speci ic analysis was pe o med and an assignmen was made o ansc ip s us- ing GSNAP ( e sion 2012–01–11) as p e iously desc ibed (40). RESULTS T ansc ip ion a e scales wi h he yeas g ow h a e To analyze he ela ionship be ween mRNA u no e and g ow h a e, we analyzed a se o 42 genome-wide ansc ip- ion a e expe imen s (25,45–48)(P ´ e ez-O ´ ın, Medina and An ´unez, unpublished), pe o med by he genomic un-on (GRO) me hod ha p o ides nascen TR da a (25). These expe imen s co esponded o he yeas s ains (wild- ypes and mu an s) ha g ow in di e en ca bon sou ces (YPD, YPGal, YPRa ) and a dis inc empe a u es ( om 23◦C o 37◦C; see Supplemen a y Table S1A). In each expe imen , we ob ained indi idual mRNA amoun s, nascen TR and kd o mos p o ein-coding genes, as well as he o al TR ( o all h ee nuclea RNA polyme ases). All he expe imen s we e compa able because he da a we e no malized by cell num- be s and cell olumes (48). We i s analyzed he global RNA pol II nascen TR and RA da a by summing all he p o ein-coding genes wi h a leas 32 alid da a poin s (5411 genes). Figu e 1A shows a clea posi i e co ela ion be ween he o al RNA pol II TR and he g ow h a e (in ed). Wi hin he analyzed GR ange, an inc ease in he TR was app oxima ely p opo ional o an inc ease in he GR. This dependence was no caused by he genes ha encoded he ansla ion machine y because he o e all endency emained almos iden ical when hey we e excluded (Figu e 1A in blue). The o al TR (including RNA pol I, II and III), which was also ob ained in he un- on expe imen , co ela ed posi i ely wi h he g ow h a e (Supplemen a y Figu e S1). This was expec ed because ap- p oxima ely a 75% o he o al TR came om RNA pol I + III ansc ip ion (8), which was done o ansc ibe he non- coding componen s o he ansla ion machine y ( RNAs and RNAs). Gi en he e y di e se se o s ains and g ow h condi- ions (see Supplemen a y Table S1A), hey all e ealed hei own pa icula in luence on he o al TR. Thus he obse ed g ow h a e dependence o he TR mus be ega ded as e- ma kably obus . None heless, all he da a we e ob ained by a pa icula echnique, GRO, which could be subjec o echnique-speci ic sys ema ic biases. To independen ly con- i m ou esul s, we used an independen published da a se based on me abolic mRNA labeling, which di ec ly mea- su es he ma u e mRNAs ha appea in he cy oplasm. We analyzed he da a ha co esponded o 44 yeas mu an popula ions ha g ew exponen ially a 30◦C in YPD om which he TR, RA and kd o mos o he p o ein-coding genes we e ex ac ed (26). To complemen his da a se , we measu ed he GRs o mos o hese mu an s (38 s ains, see Supplemen a y Table S1B) by mic o-cul i a ion (30). As he TR should be measu ed as a change in mRNA concen a- ion o e ime (22), and he o iginal da a we e no co ec ed, we used he olume es ima es o hese mu an s ob ained by Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 3647 Figu e 1. The mRNA u no e co ela ed posi i ely wi h he yeas g ow h a e. (A) We plo ed 34 o he 42 di e en g ow h a ia ions (mu an s ains, empe a u e o ca bon sou ce a ia ions) o which we had ela i e ansc ip ion a es o a e e ence sample. A clea dependence o he o al nascen ansc ip ion a e (TR) o RNA pol II is obse ed, as measu ed by GRO (23) wi h he GR o he cul u e in ae a ed lasks. Values a e ela i e o he co esponding e e ence s ain. Red symbols co espond o he a e age o all he da a, whe eas blue symbols co espond o he a e age o all he da a excep genes RP and RiBi. Pea son’s co ela ion coe icien ( ) and he associa ed P- alue a e shown. (B–D). We used he da a o he a e age o al mRNA syn hesis a e (TR), mRNA le els (RA) o mRNA deg ada ion cons an (kd) om 38 mu an s ains (26) and he GR de e mined by mic o-cul i a ion. The o iginal TR and RA da a we e co ec ed by cell olume using da a om (27,28). Once again, we obse e a posi i e dependence o TR and in kd,which kep he mRNA le els app oxima ely cons an . Pea son’s co ela ion coe icien s ( ) and associa ed P- alues a e shown. (E) Mul iple Reg ession Model shows ha he RA can be p edic ed om he TR and he kdwi h highly signi ican P- alues bu i canno be p edic ed om GR (see S a is ical Appendix in Supplemen a y In o ma ion o u he explana ion). 3648 Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 Jo gensen e al. (27) and T uong e al. (28) o co ec hem. The plo s o he h ee pa ame e s a e shown in Figu e 1B– D. He e we see ha he TR s eadily and p opo ionally in- c eased wi h an inc easing GR (Figu e 1B), whe eas mRNA s abili y dec eased (kdinc eased, Figu e 1D). The e e se dependencies o he TR and mRNA s abili y on g ow h a e ensu ed ha he s eady-s a e mRNA le els, which e lec ed he join consequence o mRNA syn hesis and deg ada ion a es, we e la gely independen o he GR (Figu e 1C). A Mul iple Reg ession Model con i ms ha RA can be p e- dic ed om he TR and om he kd, as expec ed, wi h highly signi ican P- alues bu i canno be p edic ed om he GR (Figu e 1E). Simila esul s o he kdand RA dependence on GR we e ob ained o ou GRO expe imen s da a se om a subse o expe imen s (see Supplemen a y Table S4). Toge he hese esul s compellingly suppo ha he RNA pol II ansc ip ion a es, measu ed as bo h he nascen o ma u e TR, we e s ongly associa ed wi h he g ow h a e. The deg ada ion a e cons an (kd,seeIn o- duc ion sec ion) also inc eased wi h a ising g ow h a e, and o al [mRNA] emained cons an . Thus he mRNA u no e a e o yeas cells co ela ed di ec ly wi h he pop- ula ion g ow h a es. Mo eo e , he inc ease in mRNA u no e was quan i a i ely simila o ha o he GR, which sugges s ha he p opo ionali y be ween bo h should be main ained. P o ein syn hesis genes a e mo e ansc ibed and espi a o y gene ansc ip s a e des abilized a as g ow h The gene al endency o he RNA pol II TR ep esen s he a e age o e all p o ein-coding genes. The pa icula en- dencies o indi idual genes may di e ge subs an ially om his a e age and can be s udied om ou o iginal da a se s o om ha published by P. C ame ’s g oup (26). To in- es iga e he dominan biological ends among indi idual genes o each TR, kdand RA, we o de ed all he no mal- ized da a acco ding o hei co ela ion wi h he g ow h a e. Two di e en s a is ical analyses, he SAM and GSA me h- ods (see M&M), we e hen applied o iden i y he gene unc- ions (GO ca ego ies) en iched in he op (highes posi i e ) o he bo om (lowes posi i e o highes nega i e , de- pending on he plo ) o each anking lis . Bo h me hods and bo h da a se s ga e simila esul s. As he gene al TR and kd endencies esul ed in s ong posi i e co ela ions wi h he GR (Figu e 1), e y ew genes showed nega i e co ela- ions. Thus o he TR, and kddesc ibed below, ‘ op’ means he genes wi h he s onges posi i e co ela ion, whe eas ‘bo om’ gene ally indica es he genes wi h he lowes posi- i e (a ound ze o) co ela ion. In con as , he RA da a se s a e cen e ed on ze o gi en he gene al absence o a change wi haGR(Figu e1C). The e o e he ‘bo om’ g oup con- ains he genes ha uly and nega i ely co ela e wi h he g ow h a e. As all he da a se s we e no malized by hei median, he ound endencies ep esen biases o he unc- ional ca ego ies wi h ega d o he ypical beha io o he popula ion. The SAM s a is ical analyses o TR-GR dependence ound RiBi (Ribosome Biogenesis), and o he p o ein- syn hesis- ela ed genes, o be en iched among he genes whose nascen TRs co ela ed posi i ely wi h he GRs (Fig- u e 2A, Supplemen a y Table S2). In con as , he TRs o he espi a ion- and mi ochond ia- ela ed genes ended o be he leas posi i ely co ela ed wi h GRs (Figu e 2B). The s a is ical analyses o RA-GR dependence also ound ha he s eady-s a e mRNA abundance o RiBi- and he o he p o ein syn hesis- ela ed genes ended o inc ease wi h ising g ow h a es (Figu e 2C, Supplemen a y Ta- ble S2). In con as , he s eady-s a e mRNA abundance o he espi a ion- and mi ochond ia- ela ed genes dimin- ished wi h as e g ow h (Figu e 2D). We did no ana- lyze he mRNA s abili y endencies in his da a se because e y ew da a se s we e a ailable. Once again, he analy- sis o he TR-GR dependence in he al e na i e da a se (26) ound ha he mRNA syn hesis a e o he RiBi- and amino acid biosyn hesis- ela ed genes co ela ed s ongly and posi i ely wi h g ow h a e, whe eas he TRs o he cell cycle and egula ion genes ended o decouple om i (see Supplemen a y Figu e S2 and Table S3). Gi en he la ge numbe o es ima es in his case, we also analyzed kd- GR dependence o ind ha he mRNA deg ada ion o he RiBi and Glycosyla ion genes did no co ela e wi h g ow h a e, whe eas he mRNA deg ada ion o he acuole- and mi ochond ia- ela ed genes ended o inc ease wi h ising GR (see Figu e 2E–F and Supplemen a y Table S2). O he al e na i e sea ches be ween ou TR and RA da a and he TRandRAda ao Sune al. (26) we e made by he GSA me hod. Simila esul s we e ob ained. D. Bo s ein’s lab made a sys ema ic sea ch o genes whose RA co ela ed ei he posi i ely o nega i ely wi h he g ow h a e (16,17,49). In hese s udies, a signa u e o 72 genes was ob ained, which s ongly depended on he GR. We u he analyzed he esponse o hese genes and ound ha some 50 genes showed essen ially iden ical (de- c easing o inc easing) endencies o hei RA and TR p o- iles (Supplemen a y Figu e S3). This indica es ha mos g ow h a e- egula ed ansc ip s a e con olled manly a he syn hesis le el. Howe e , we also ound se en genes o which mRNA abundance and he TR a e da a co ela ed in e sely, which sugges s a s ong egula ion a he s abili y le el. In conclusion, apa om he gene al endency o cells o inc ease he o al mRNA syn hesis a e wi h as e g ow h, he inc eased ela i e ansc ip ion o some unc- ional classes o genes was pa icula ly p onounced. As a esul o he highe TRs, he s eady-s a e mRNA abun- dance o hese pa icula ypes o genes, unlike ha ob- se ed o he o e all ansc ip ome (Figu e 1C), seemed o also inc ease wi h as e g ow h. In con as , he s eady- s a e mRNA concen a ion o o he unc ional classes o genes ac ually dec eased as a esul o as e g ow h, caused appa en ly by a pa icula ela i e dec ease in hei mRNA s abili y (inc ease in kd). Ribosome biogenesis, mi ochond ia- ela ed and s ess- induced genes use di e en mechanisms o adjus mRNA le els o he g ow h a e We p e iously ound ha some GO ca ego ies di e ed om he gene al endency o no change in mRNA abundance by displaying changes in he GR. To shed u he ligh on o hese opposing ends, we sepa a ely conside ed he i- Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 3649 Figu e 2. Sea ch o he Gene On ology ca ego ies ha show enhanced dependence on he g ow h a e. GO en ichmen sea ches o he highes posi i e and nega i e GR dependences on he nascen TR, mRNA le els and deg ada ion a es a e shown. Apa om he gene al endency o he nTR and RA showninFigu e1, ce ain unc ional ca ego ies a e especially en iched among he highes posi i e co ela ions ( op) wi h he GR, o among he genes wi h he lowes co ela ion (bo om). We show he e (A–D) he analyses done wi h he da a se s used in Figu e 1 o he nTR and RA, and om he da a se o Sun e al. (26) o kd(E–F) using an SAM analysis and summa ized by he ReViGO p og am. Simila esul s o he TR and RA we e ob ained wi h he da a se s used in Figu e 1B–C (26); see Supplemen a y Figu e S3. 3650 Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 bosome biogenesis (GO:0042254) and mi ochond ial ibo- some (GO:0005761) unc ional ca ego ies. Figu e 3A illus- a es how he TR o he RiBi genes inc eased o esul in highe mRNA abundance and he eby, p esumably, main- ained he ibosome numbe s le el wi h GR demands. This is he clea es example o a egulon con olled a he an- sc ip ional le el. RiBi mRNA s abili ies did no change in a GR-dependen manne (Figu e 3B). This beha io was sim- ila in ibosomal p o ein (RP) genes (Supplemen a y Figu e S4). In con as , he mi ochond ial ibosome genes showed no egula ion a he ansc ip ional le el (Figu e 3C). Ne - e heless, s eady-s a e mRNA abundance diminished wi h he GR, which was achie ed by a ma ked educ ion in he s abili y o hese mRNAs (Figu e 3D). This is he clea es example o a pos - ansc ip ional egulon (50). Gi en ha he mRNA le els o he espi a o y- and o he mi ochond ia- ela ed genes lowe ed wi h as e g ow h, and by assuming no addi ional compensa o y eg- ula ion o he ansla ion o pos - ansla ional le els, we expec ed espi a ion o dec ease acco dingly. To es his p edic ion, we selec ed a se o nine s ains (mu an s and wild ypes) om ou da a se (see Figu e 1A and Sup- plemen a y Table S1A) and de e mined hei espi a o y quo ien ( e men a ion/ espi a ion, RQ) in a mic o e men- o (37). As seen in Supplemen a y Figu e S5, he lowe he mRNA le els o he genes ela ed wi h espi a ion (GO:0009060) o mi ochond ial ibosome (GO:0005761), he highe he RQs and, hus, he ela i e in luence o es- pi a ion on me abolism. This shows ha he educ ion in mRNA le els o he espi a o y and o he mi ochond ial genes wi h highe g ow h a es is physiologically ele an and se es o educe espi a ion. This s ongly suppo s he dependence o as e -g owing yeas cells on he e men a- i e me abolism. Finally, we paid special a en ion o he equen ly s ud- ied s ess-induced genes o he so-called ‘en i onmen al s ess esponse’ (ESR) (51). These genes ha e been shown o be up- egula ed in he RA in slow-g owing mu an s (12,16), which implies a nega i e co ela ion be ween hei mRNA abundance and g ow h. I was ou in en ion o es ablish whe he his was indeed he case, and i so, whe he o no i was achie ed by a change in he TR o DR. Indeed he mRNA abundance o he s ess-induced sec ion o he ESR diminished a as e g ow h (Figu e 3E). This dec ease was achie ed exclusi ely by lowe ing TRs, and no gene al change ook place in mRNA deg ada ion (Figu e 3F). Thus he s ess-induced pa o he ESR exempli ied a hi d egu- la o y s a egy: a nega i e e ec o g ow h a e on ansc ip- ion a es. mRNA dependence on he g ow h a e can be de ec ed wi hin he he e ogenei y o a wild- ype cell popula ion So a we analyzed he dependence o mRNA concen a- ions, and hei syn hesis and deg ada ion a es on cellula g ow h a e ac oss di e en geno ypes and en i onmen al condi ions, including mu an s ains wi h impai ed p oli - e a ion capaci y o wild- ype cells ha g ow in di e en en- i onmen s. Simila ly, p e ious expe imen s ha ha e de- sc ibed he in luence o he GR on gene exp ession ha e also been based on he compa ison o cells popula ions in di e en ca bon sou ces o unde chemos a - egula ed con- di ions (18,52). These expe imen al designs canno ule ou he ac ha he seeming dependence o TR, RA and kdon he GR migh be a con ounding e ec o bo h depending on a sha ed hi d ac o in he o m o speci ic mu a ions and speci ic ad e se en i onmen al e ec s. The connec ion be ween he g ow h a e and RNA me abolism should hap- pen a he single cell le el. In o de o ule ou ha any o hose e ec s occu ed only a he popula ion le el, we mea- su ed he GR o single cells as hei abili y o p oduce mi- c ocolonies, and we analyzed hei GR ansc ip omes. To do his, we analyzed he mic ocolonies gene a ed a - e he encapsula ion o he single cells ha belonged o a clonal wild- ype popula ion in algina e mic osphe es (Sup- plemen a y Figu e S6). Algina e encapsula ion allows he ee di usion o nu ien s, is ully compa ible wi h yeas p o- li e a ion, and helps so cells acco ding o hei GR po- en ial (38). No dec eased iabili y was de ec ed a e he encapsula ion p ocedu e (Supplemen a y Figu e S7A). En- capsula ed cells we e incuba ed in ich liquid medium un- de s anda d cul u e condi ions and we e allowed o p oli - e a e. A he di e en ime poin s, samples we e aken and obse ed by op ical mic oscopy. The e y di e en size ob- se ed in mic ocolonies indica ed an e iden a ia ion in he p oli e a ion capaci y wi hin his clonal popula ion (Figu e 4A). Quan i ica ion o he mic ocolony size by la ge pa - icle low cy ome y con i med an in insic he e ogenei y ha inc eased wi h incuba ion ime (Figu e 4B). We es ed whe he he di e en mic ocolony size was due o a dis inc size o he mic opa icle i sel o i i was a biased localiza- ion wi hin he mic opa icle. We ound no di e ence o ei he (Supplemen a y Figu e S7B and S7C). Mic ocolony size he e ogenei y could also be due o an unequal lag phase. The inc easing s a is ical dispe sion shown by mic ocolony size du ing he ime-cou se expe imen s, by main aining symme ical dis ibu ion, was ha dly compa ible wi h his possibili y (Figu e 4B). Ou esul s a e in ag eemen wi h p e ious epo s on p oli e a ion he e ogenei y conduc ed unde op imal cul u e condi ions in S. ce e isiae, de ec ed wi h echnical app oaches ha did no in ol e encapsula- ion (24). The combina ion o mic oencapsula ion and la ge pa i- cle low cy ome y allows sis e mic ocolonies o be physi- cally sepa a ed by so ing (Supplemen a y Figu e S6). We isola ed he 10% op and bo om subpopula ions o mic o- colonies acco ding o hei size (Supplemen a y Figu e S6). These wo subpopula ions exhibi ed an a e age cell num- be o 32 and 2519, espec i ely, which in ol es appa en doubling imes o app oxima ely 162 and 76 min acco d- ing o he incuba ion ime unde g owing condi ions (Figu e 4C). This longe duplica ion ime o he small mic ocolony cells ag ees wi h hei longe G1 phase and e lec s hei di e en ial cell-cycle egula ion (Delgado–Ramos, Mu˜ noz– Cen eno and Ch´ a ez, o be published elsewhe e). The a e - age cell numbe and appa en doubling ime o non-so ed mic ocolonies ga e in e media e alues. Howe e , when all he analyzed mic ocolonies we e conside ed as a single pop- ula ion, he esul ing popula ion doubling ime was compa- able o he s anda d liquid cul u es (Figu e 4C). This simple p ocedu e allowed us o isola e mRNA p epa a ions om he wo ypes o so ed colonies, and o Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 3651 Figu e 3. Ribosome biogenesis, mi ochond ia- ela ed and s ess-induced genes use di e en mechanisms o adjus mRNA le els o he g ow h a e. Using he da a om he Figu e 1A da a se , we ex ac ed da a o he RiBi egulon (GO: 0042254, panels Aand B), o he mi ochond ial ibosome (GO: 0005761, panels Cand D), and o he induced pa o he ESR (51)(panels Eand F). The genes in he Mi ochond ial ibosome GO ca ego y (and simila ly in he Pu 3 egulon, see Figu e 5B) seem o ha e lowe mRNA le els (RA) wi h he g ow h a e (GR) o he cul u e, mainly by mRNA des abiliza ion (inc eased deg ada ion cons an , kd). The mRNA le els o he RiBi genes (and he ibosomal p o ein genes, see Supplemen a y Figu e S4) end o ise wi h he GR due o inc easing ansc ip ion (nTR). The induced pa o he ESR con e sely lowe s, and RA educes wi h dec easing ansc ip ion. Pea son’s co ela ion coe icien s ( ) and associa ed P- alues a e shown. 3658 Nucleic Acids Resea ch, 2016, Vol. 44, No. 8 54. Lesage,P., Yang,X. and Ca lson,M. (1996) Yeas SNF1 p o ein kinase in e ac s wi h SIP4, a C6 zinc clus e ansc ip ional ac i a o : a new ole o SNF1 in he glucose esponse. Mol. Cell. Biol.,16, 1921–1928. 55. Vincen ,O. and Ca lson,M. (1998) Sip4, a Sn 1 kinase-dependen ansc ip ional ac i a o , binds o he ca bon sou ce- esponsi e elemen o gluconeogenic genes. EMBO J.,17, 7002–7008. 56. Gasmi,N., Jacques,P.E., Klimo a,N., Guo,X., Riccia di,A., Robe ,F. and Tu co e,B. (2014) The swi ch om e men a ion o espi a ion in Saccha omyces ce e isiae is egula ed by he E 1 ansc ip ional ac i a o / ep esso . Gene ics,198, 547–560. 57. Mille ,M.A., Russo,J., Fische ,A.D., Lopez Leban,F.A. and Oli as,W.M. (2014) Ca bon sou ce-dependen al e a ion o Pu 3p ac i i y media es apid changes in he s abili ies o mRNAs in ol ed in mi ochond ial unc ion. Nucleic Acids Res.,42, 3954–3970. 58. Lee,C.D. and Tu,B.P. (2015) Glucose- egula ed phospho yla ion o he PUF p o ein Pu 3 egula es he ansla ional a e o i s bound mRNAs and associa ion wi h RNA g anules. Cell Rep.,11, 1638–1650. 59. Gup a,I., Claude -Muns e ,S., Klaus,B., Ja elin,A.I., Aiya ,R.S., Benes,V., Wilkening,S., Hube ,W., Pelechano,V. and S einme z,L.M. (2014) Al e na i e polyadenyla ion di e si ies pos - ansc ip ional egula ion by selec i e RNA-p o ein in e ac ions. Mol. Sys . Biol.,10, 719. 60. Rio dan,D.P., He schlag,D. and B own,P.O. (2011) Iden i ica ion o RNA ecogni ion elemen s in he Saccha omyces ce e isiae ansc ip ome. Nucleic Acids Res.,39, 1501–1509. 61. Sain -Geo ges,Y., Ga cia,M., Dela eau,T., Jou d en,L., Le C om,S., Lemoine,S., Tan y,V., De aux,F. and Jacq,C. (2008) Yeas mi ochond ial biogenesis: a ole o he PUF RNA-binding p o ein Pu 3p in mRNA localiza ion. PloS One,3, e2293. 62. Po ua,O. and Lib i,D. (2015) T ansc ip ion e mina ion and he con ol o he ansc ip ome: why, whe e and how o s op. Na . Re . Mol. Cell Biol.,16, 190–202. 63. Pelechano,V., Cha ez,S. and Pe ez-O in,J.E. (2010) A comple e se o nascen ansc ip ion a es o yeas genes. PloS One,5, e15442. 64. Ke en,L., an Dijk,D., Weinga en-Gabbay,S., Da idi,D., Jona,G., Weinbe ge ,A., Milo,R. and Segal,E. (2015) Noise in gene exp ession is coupled o g ow h a e. Genome Res.,25, 1893-1902 65. Shalem,O., G oisman,B., Chode ,M., Dahan,O. and Pilpel,Y. (2011) T ansc ip ome kine ics is go e ned by a genome-wide coupling o mRNA p oduc ion and deg ada ion: a ole o RNA Pol II. PLoS Gene .,7, e1002273. 66. Molin,C., Jauhiainen,A., Wa inge ,J., Ne man,O. and Sunne hagen,P. (2009) mRNA s abili y changes p ecede changes in s eady-s a e mRNA amoun s du ing hype osmo ic s ess. RNA,15, 600–614. 67. Pe ez-O in,J.E., Alepuz,P.M. and Mo eno,J. (2007) Genomics and gene ansc ip ion kine ics in yeas . T ends Gene .,23, 250–257. 68. Ge be ,A.P., He schlag,D. and B own,P.O. (2004) Ex ensi e associa ion o unc ionally and cy o opically ela ed mRNAs wi h Pu amily RNA-binding p o eins in yeas . PLoS Biol.,2, E79. 69. Diaz-Ruiz,R., Rigoule ,M. and De in,A. (2011) The Wa bu g and C ab ee e ec s: on he o igin o cance cell ene gy me abolism and o yeas glucose ep ession. Biochim. Biophys. Ac a,1807, 568–576. 70. Lagunas,R. (1986) Misconcep ions abou he ene gy me abolism o Saccha omyces ce e isiae. Yeas ,2, 221–228. 71. Dashko,S., Zhou,N., Compagno,C. and Pisku ,J. (2014) Why, when, and how did yeas e ol e alcoholic e men a ion? FEMS Yeas Res., 14, 826–832. 72. Wang,J., A olia,E., Hua,B., Sa i ,Y., Escalan e-Chong,R. and Sp inge ,M. (2015) Na u al a ia ion in p epa a ion o nu ien deple ion e eals a cos -bene i adeo . PLoS Biol.,13, e1002041. 73. Ven u elli,O.S., Zule a,I., Mu ay,R.M. and El-Samad,H. (2015) Popula ion di e si ica ion in a yeas me abolic p og am p omo es an icipa ion o en i onmen al shi s. PLoS Biol.,13, e1002042. 74. Abe na hy,E., Gilbe son,S., Alla,R. and Glaunsinge ,B. (2015) Vi al Nucleases induce an mRNA deg ada ion- ansc ip ion eedback loop in mammalian cells. Cell Hos Mic obe,18, 243–253.