Chromatin accessibility is associated with CRISPR-Cas9 efficiency in the zebrafish (Danio rerio)

RESEARCH ARTICLE Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) Me i I. E. Uusi-Ma ¨kela ¨ 1 *, Ha lan R. Ba ke 2 , Ca ina A. Ba ¨ue lein 1 , Tomi Ha ¨kkinen 3 , Ma i Nyk e 3 , Mika Ra ¨me 1,4,5 * 1Labo a o y o Expe imen al Immunology, BioMediTech Ins i u e and Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland, 2Labo a o y o Ana omy, Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland, 3Labo a o y o Compu a ional Biology, BioMediTech Ins i u e and Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland, 4PEDEGO Resea ch Uni , Medical Resea ch Cen e Oulu, and Depa men o Child en and Adolescen s, Oulu Uni e si y Hospi al, Oulu, Finland, 5Depa men o Pedia ics, Tampe e Uni e si y Hospi al, Tampe e, Finland *[email p o ec ed] (MR); me i.uusi-m[email p o ec ed] (MIEUM) Abs ac CRISPR-Cas9 echnology is ou inely applied o a ge ed mu agenesis in model o ganisms and cell lines. Recen s udies indica e ha he p oka yo ic CRISPR-Cas9 sys em is a ec ed by euka yo ic ch oma in s uc u es. He e, we show ha he likelihood o success ul mu a- genesis co ela es wi h ansc ip le els du ing ea ly de elopmen in zeb a ish (Danio e io) emb yos. In an expe imen al se ing, we ound ha guide RNAs di e in hei onse o mu a- genesis ac i i y in i o. Fu he mo e, some guide RNAs wi h high in i o ac i i y possessed poo mu agenesis ac i i y in i o, sugges ing he p esence o ac o s ha limi he mu agen- esis in i o. Using open access da ase s gene a ed om ea ly de elopmen al s ages o he zeb a ish, and guide RNAs selec ed om he CRISPRz da abase, we p o ide u he e i- dence o an associa ion be ween gene exp ession du ing ea ly de elopmen and he suc- cess o CRISPR-Cas9 mu agenesis in zeb a ish emb yos. In o de o u he inspec he e ec o ch oma in on CRISPR-Cas9 mu agenesis, we analysed he ela ionship o selec ed ch oma in ea u es on CRISPR-Cas9 mu agenesis e iciency using publicly a ailable da a om zeb a ish emb yos. We ound a co ela ion be ween ch oma in openness and he e i- ciency o CRISPR-Cas9 mu agenesis. These esul s indica e ha CRISPR-Cas9 mu agen- esis is in luenced by ch oma in accessibili y in zeb a ish emb yos. In oduc ion Since i s disco e y in S ep ococcus pyogenes, he CRISPR-Cas9 (Clus e ed egula ly in e - spaced sho palind omic epea s–CRISPR associa ed 9) sys em has been ex ensi ely applied o modi y he euka yo ic genome in a a ge ed manne [1,2]. CRISPR-Cas9 echnology akes ad an age o he bac e ial Cas9 endonuclease, which gene a es a double s anded b eak in i s DNA a ge [1]. The epai o he b eak by he e o p one epai machine y o non- homologous end joining o en leads o he inco po a ion o mu a ions and pe manen modi i- ca ions o he genome [2]. PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 1 / 15 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Ci a ion: Uusi-Ma¨kela¨MIE, Ba ke HR, Ba¨ue lein CA, Ha¨kkinen T, Nyk e M, Ra¨me M (2018) Ch oma in accessibili y is associa ed wi h CRISPR- Cas9 e iciency in he zeb a ish (Danio e io). PLoS ONE 13(4): e0196238. h ps://doi.o g/10.1371/ jou nal.pone.0196238 Edi o : B uce B. Riley, Texas A&M Uni e si y, UNITED STATES Recei ed: Feb ua y 2, 2018 Accep ed: Ma ch 9, 2018 Published: Ap il 23, 2018 Copy igh : ©2018 Uusi-Ma¨kela¨e al. 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, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed. Da a A ailabili y S a emen : All ele an expe imen al da a a e wi hin he pape and i s Suppo ing In o ma ion iles. All iles a e a ailable om he CRISPRz and he A ayExp ess da abases (accession codes E-GEOD-45706; E-GEOD-4863; E-GEOD-52110; E-GEOD-74231). Funding: This s udy was suppo ed wi h he ollowing g an s: Tampe e Tube culosis Founda ion (h p://www. ube kuloosisaa io. i/) (MU;MR); Finnish Conco dia Fund (h p://www.konko dia- lii o.com/) (MU); Sig id Juselius Founda ion Cas9 is di ec ed o bind i s a ge sequence by a single chime ic guide RNA molecule (sgRNA), which ecognizes an app oxima ely 20 nucleo ide a ge si e, ollowed by he h ee nucleo ide p o- ospace adjacen mo i (PAM)-sequence (5’-NGG-3’) [1–3]. The sgRNA sequence is conside ed he limi ing s ep in mu agenesis design, as he genomic a ge si e needs o be unique. An op imal GC-con en and speci ic nucleo ides a key posi ions in he a ge sequence can also al e he e i- ciency and he speci ici y o mu agenesis [4–8]. The e iciency and unspeci ic, o - a ge binding o he nuclease a e no easy o p edic . As a esul , mul iple algo i hms and online ools ha e been c ea ed o he iden i ica ion o guide RNA a ge s wi h op imal Cas9 loading sco es and he leas amoun o o - a ge s [9–17]. Howe e , he in silico p edic ions do no always co ela e wi h he obse ed mu agenesis e iciency and speci ici y [11,18–20]. Euka yo ic gene exp ession is egula ed a he epigene ic le el by packing o DNA in o nucleosomes, which a e o med by w apping 146bp o DNA a ound a his one oc ame [21]. These euka yo ic ch oma in s uc u es undamen ally di e om bac e ial DNA packing, and being a p oka yo ic enzyme, i is plausible ha Cas9 canno ully ope a e a ound all ch oma in s uc u es. Indeed, ecen e idence indica es ha ch oma in in luences Cas9 binding by limi - ing he accessibili y o he a ge si e [10,18,22–25]. Cas9 akes longe o scan o he a ge si es bu ied in he e och oma in, whe eas a ge s loca ed in euch oma in a e mo e accessible, and hus easie o loca e [24]. Howe e , he e och oma in does no en i ely p e en Cas9 om binding o po en ial a ge si es and despi e binding, clea age does no necessa ily occu [22,24]. Ta ge si e accessibili y is e lec ed in he endency o Cas9 o ac on seconda y a ge s, so i plays an impo an ole when designing e ec i e sgRNAs wi h maximum e iciency and a minimal numbe o o - a ge s [10,17,18]. I he in ended a ge is bu ied in he e och oma in, i is mo e p obable ha Cas9 binds o seconda y a ge s and is mo e likely o ind hose in he exon egions in euch oma in [18]. E idence suppo ing he in ol emen o ch oma in accessi- bili y in Cas9 binding has eme ged in in i o models, cell lines and in he zeb a ish (Danio e io) [10,17,23–26]. Howe e , de ailed unde s anding on which ch oma in ea u es con ibu e o ch oma in accessibili y his is s ill lacking. Compa ed o cell lines, zeb a ish can p esen addi ional challenges o genome edi ing. Compa ed o o he e eb a es, he eleos speci ic genome duplica ion has esul ed in mul iple simila genes o pseudogenes and his can, in some ins ances, complica e he iden i ica ion o unique a ge s o sgRNA. Secondly, o gene a e mu an zeb a ish, he sgRNA and Cas9 a e mic oinjec ed in o he e ilized emb yo, and mu agenesis occu s du ing he i s hou s o de elopmen [27]. Compa ed o cell lines, he e ilized, CRISPR-injec ed zygo e p esen s a challenge o all mu agenesis echniques as i unde goes de elopmen al and di e en ia ion p ocesses ha equi e global changes in ch oma in. Las ly, he i s cell di ision in zeb a ish akes place e y apidly (40 minu es a e e iliza ion), when compa ed o he cell di isions o example in mice ( eaching E1.5 a 24 hou s pos e iliza ion, hp ). Mu agenesis occu ing a e his i s cell di ision may mo e likely lead o mosaicism. Du ing de elopmen , he ch oma in landscape is unde cons an change in o de o enable coo dina ed g ow h and di e en ia ion [28–30]. The zygo e is suppo ed by he a ailable ma e - nal ansc ip s and he zygo ic genome emains ansc ip ionally inac i e un il he ma e nal o zygo ic genome ac i a ion (MZT) a he mid blas ula ansi ion (MBT) [31]. Ou cu en unde - s anding o zygo ic ch oma in is limi ed, bu i has been shown ha a speci ic his one modi ica- ion p e-pa e ning ma ks de elopmen ally ac i e and inac i e genes du ing de elopmen [32]. The nuclease accessibili y o he de eloping, ch oma in-packed genome o emb yos emains poo ly unde s ood. P e iously, i was obse ed ha ch oma in does no in luence CRISPR-Cas9 a ge ing in zeb a ish emb yos in an MNase assay (Mic ococcal nuclease assay), bu la e ATAC-seq (Assay o T ansposase-Accessible Ch oma in using sequencing) esul s sugges ed ha CRISPR-Cas9 is mo e likely o be success ul when a ge ing open ch oma in [9,17]. Mo e Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 2 / 15 (h p://sig idjuselius. i/apu aha /) (MR); Uni e si y o Tampe e Doc o al School (h p://www.u a. i/ english/doc o alschool/index.h ml) (MU); Finnish Cul u al Founda ion – Maili Au io Fund (h ps://sk . i/) (HB). The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . Compe ing in e es s: The au ho s ha e decla ed ha no compe ing in e es s exis . in o ma ion on he in luence o ch oma in on CRISPR-Cas9 mu agenesis in model o ganisms is needed in o de o imp o e he e iciency o genome enginee ing me hodologies. In his s udy, we obse ed disc epancies be ween he in i o and in i o ac i i ies o sgRNAs, and ha selec ed sgRNAs di e o hei onse o mu agenesis. We saw an associa ion be ween success ul mu agenesis and he ansc ip le els du ing ea ly de elopmen . We looked u he in o he in ol emen o gene ac i a ion and ch oma in in explaining he CRISPR-Cas9 mu agenesis e iciency in zeb a ish emb yos. Ou esul s indica e ha gene exp ession and ch oma in openness a e associa ed wi h he e iciency o CRISPR-Cas9 mu agenesis. Howe e , we saw no associa ion o mu agenesis e iciency wi h ei he exon me hyla ion o his one H3 Lysine 4 ime hyla ion (H3K4me3) a p omo e s. Resul s Good in i o ac i i y o sgRNA does no assu e in i o e icacy Analyzing he e icacy o di e en sgRNAs in i o is labo ious. To imp o e he sc eening o e icien sgRNAs, in i o diges ion o he a ge sequence can be used. We analyzed he mu agenesis ac i i y o six sgRNAs i s in i o and hen selec ed h ee o analysis in i o. As shown in Fig 1, some sgRNAs wi h good in i o e iciency p esen ed low o no in i o ac i i y. This sugges s ha ac o s p esen in i o p e en Cas9 om ac ing on i s a ge si e. Fig 1. In i o and in i o CRISPR-Cas9 mu agenesis e iciencies do no co ela e o all genes. a) An in i o diges ion assay shows ha sgRNAs di e in hei e iciencies. Below he gene name, + and - indica e he p esence o absence o Cas9 p o ein in he eac ion. On he igh he wild ype (w ) and he mu an p oduc s a e indica ed. b) The in i o CRISPR-Cas9 mu agenesis isualized o ca6, cxc 2 and pyca d wi h a he e oduplex mobili y assay, wi h he wild ype (w ) and he mu an p oduc s indica ed. 5 emb yos we e collec ed pe sample a 8hp . h ps://doi.o g/10.1371/jou nal.pone.0196238.g001 Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 3 / 15 Impo an ly, cxc 2 had nei he de ec able gene exp ession no mu agenesis e iciency, whe eas he genes pe missi e o mu agenesis (pyca d,ca6) showed ea ly exp ession (Fig 1,S1 Fig). This led us o hypo hesize ha he onse and he le el o gene exp ession could in luence he CRISPR-Cas9 mu agenesis. The co esponding esul s using he T7 Endonuclease I assay a e displayed in S2 Fig. In ou hands he T7 Endonuclease I assay has a lowe esolu ion compa ed o he he e oduplex mobili y assay, especially wi h sgRNAs o lowe e iciency. On he o he hand, he T7 Endonuclease I assay can be eadily used o quan i a ion o mu agenesis e i- ciency, especially wi h sgRNAs o highe e iciency. The onse o mu agenesis di e s be ween sgRNAs in i o As we saw a disc epancy be ween in i o and in i o mu agenesis e iciencies o some sgRNAs, we nex analyzed whe he he onse o mu agenesis co ela es wi h he onse o gene exp ession. To a oid he delay o mRNA ansc ip ion o Cas9 ac i i y, we used a eady Cas9 p o ein in ou expe imen s wi h app op ia e p eincuba ion s ep o allow he sgRNA o complex wi h Cas9. Th ee o ou unc ional sgRNAs we e chosen o he analysis. The sgRNAs a ge ing ca10a, sema4gb, o ca6 we e co-injec ed wi h he Cas9 p o ein in o he 1-cell s age emb yo and he onse o mu agenesis was analyzed using bo h a he e oduplex mobili y assay and a T7 Endonu- clease I mu a ion de ec ion assays. As shown in Fig 2 using he he e oduplex mobili y assay, he i s mu a ions become de ec able as soon as 1hp o ca10a and sema4gb, whe eas he i s mu a- ions o ca6 appea ed a 3hp (Fig 2). These esul s indica e ha he onse o mu agenesis di e s depending on he sgRNAs in zeb a ish emb yos. Based on hese esul s, we analyzed he ela ion- ship o ea ly gene exp ession and mu agenesis e iciency in mo e de ail wi h all ou sgRNAs. We we e able o de ec mu agenesis ac i i y a 1hp ( oughly co esponding o 4-cell s age). Likelihood o success ul mu agenesis in ela ion o he exp ession le el o he a ge gene in zeb a ish emb yos Al oge he , we ha e designed 86 sgRNAs using he c isp .mi .edu, ChopChop (V1 and V2) and CRISPRscan so wa es [9,14,15]. O hese sgRNAs, 30% showed de ec able in i o ac i i y (S1 Fig 2. Onse o mu agenesis di e s be ween sgRNAs. He e oduplex mobili y assay o demons a e he onse o mu agenesis using high e iciency guide RNAs a ge ing h ee di e en genes wi h di e en gene exp ession pa e ns in ea ly de elopmen . Emb yos we e collec ed a imepoin s 1, 2, 3, 4, 6hp (15– 20 emb yos pe g oup). The gene name abo e he gel image indica es CRISPR-Cas9 injec ed emb yos and con ol indica es uninjec ed con ols. The legend on he side indica es he posi ions o w (wild ype) and mu an bands in he gel. Red a ows indica e he poin a which i s mu a ions can be de ec ed. h ps://doi.o g/10.1371/jou nal.pone.0196238.g002 Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 4 / 15 Table). As GC-con en (%) has been sugges ed o in luence he e ec i eness o CRISPR-Cas9 mu agenesis, we analyzed he GC-con en o ou sgRNAs (S1 Table) [18]. The GC-con en o ou unc ional sgRNAs was ound o be simila (Mann-Whi ney U- es ; p- alue 0.452) o ha o he non- unc ional sgRNAs. When we compa ed he exp ession o he genes ha we we e able o mu a e o hose we we e no , he genes esis an o mu agenesis mo e o en had a e y low exp ession le el (Fig 3). Howe e , he di e ence did no each s a is ical signi icance (Fische ’s exac es ; no signi - ican ). Mo eo e , a majo i y o genes (79%) pe missi e o mu agenesis unde wen an inc ease in he numbe o ansc ip s a ound he MZT (iden i ied he e as a posi i e change in he num- be o ansc ip s be ween he oblong sphe e s age and 50% epiboly). This occu ed mo e o en han in he genes esis an o mu agenesis (50%). Howe e , his obse a ion was no s a- is ically signi ican (Fische ’s exac es ) (Fig 3). To examine whe he he lack o s a is ical sig- ni icance was due o a ype wo e o , we decided o de e mine whe he he e is a co ela ion be ween a ge gene exp ession and mu agenesis e iciency using la ge da ase s. Fig 3. Rela ionship o ansc ip le els in ea ly de elopmen and low mu agenesis e iciency. Pie cha s o he RNA-seq da a co esponding o g aphs in S1 Fig. a) Numbe o ansc ip s o he genes esis an o (le ) o pe missi e ( igh ) o mu agenesis be ween he oblong sphe e and he 15-somi e s age (Fische ’s exac es ; no signi ican ). 0.5 RPKM (Reads pe Kilobase o ansc ip pe Million mapped eads) was used as a limi o low exp ession. b) The numbe o genes esis an (le ) o pe missi e ( igh ) o CRISPR-Cas9 mu agenesis in which he numbe o ansc ip s is inc eased o dec eased be ween he oblong sphe e- s age and 50% epiboly (a ound he MZT) (Fische ’s exac es ; no signi ican ). h ps://doi.o g/10.1371/jou nal.pone.0196238.g003 Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 5 / 15 Mu agenesis e iciency co ela es wi h gene exp ession and ch oma in accessibili y in zeb a ish emb yos In sea ching o ac o s ha would explain he poo in i o ac i i y o some sgRNAs, we in es- iga ed a ailable open access da ase s. As genes wi h low exp ession alues ended o be mo e di icul o mu a e in ou se ing (Fig 3), we analyzed he associa ion be ween exp ession le els and mu agenesis e iciency in g ea e dep h, using la ge da ase s in o de o a oid ype 2 e o . We ob ained CRISPR-Cas9 sgRNA e iciency da a om CRISPRz da abase o all analyses [33]. We used open access RNA-seq da a (E-GEOD-45706) o ou p ima y analysis o he co - ela ion be ween CRISPR-Cas9 mu agenesis and gene exp ession [33,34]. We ound signi i- can co ela ions in ea ly de elopmen (be ween 64-cell s age and 36hp ), a he oblong sphe e s age (3.66hp , Spea man co ela ion 0.227; p- alue 0.001) and a 36hp (Spea man co ela ion 0.230; p- alue 0.001). A s ong co ela ion was obse ed a he oblong sphe e s age which occu s sho ly a e MBT, a ound he ime o zygo ic genome ac i a ion. These esul s sugges ha ansc ip ional ac i i y in luences CRISPR-Cas9 mu agenesis a ea ly de elopmen (Table 1). As me hyla ion is known o co ela e wi h ansc ip ional ep ession, we used zeb a ish exon me hyla ion da a o analyze whe he he e is any co ela ion be ween exon me hyla ion and he success o CRISPR-Cas9 mu agenesis [33,35]. As is shown in Table 2, he e was no sig- ni ican co ela ion be ween exon me hyla ion and CRISPR-Cas9 mu agenesis e iciency a he 1-cell s age o a MBT (Table 2). Simila ly, using open access da a on emb yonic his one me h- yla ion, we analyzed whe he he e is a co ela ion o H3K4me3 a p omo e s wi h CRISPR-- Cas9 mu agenesis e iciency. As shown in Table 2, he e seemed o be a co ela ion bu his did no each s a is ical signi icance (Spea man co ela ion 0.263; p- alue = 0.074) [33,36]. ATAC-sequencing is a ecen nex gene a ion sequencing me hod, which can be used o di ec ly analyze ch oma in accessibili y. Open access ATAC-seq da a o he zeb a ish emb yo is a ailable a he 4hp imepoin [37]. We compa ed mu agenesis e iciency da a wi h ATAC- seq da a a ansc ip ion s a si es o a o al o 263 genes. We disco e ed a signi ican , albei a he weak co ela ion, indica ing ha ch oma in accessibili y appea s o be one o he ac o s ha explain he e iciency o CRISPR-Cas9 mu agenesis in zeb a ish emb yos (Table 2). Table 2. Co ela ion be ween mu agenesis e iciency and ch oma in ea u es a di e en de elopmen al s ages and imepoin s. Ch oma in ea u e n De elopmen al s age/Timepoin Spea man co ela ion p- alue Exon me hyla ion 263 1-cell 0.115 0.063 Mid blas ula ansi ion 0.107 0.084 H3K4me3 47 75–80% epiboly 0.263 0.074 Ch oma in accessibili y 263 4hp 0.182 0.003 h ps://doi.o g/10.1371/jou nal.pone.0196238. 002 Table 1. Co ela ion be ween mu agenesis e iciency and gene exp ession a di e en de elopmen al s ages du - ing ea ly de elopmen . n = 209. De elopmen al s age Spea man co ela ion p- alue 64-cell 0.190 0.006 oblong-sphe e 0.227 0.001 50%-epiboly 0.187 0.007 15-somi e 0.210 0.002 36hp 0.230 0.001 48hp 0.182 0.008 60hp 0.188 0.006 72hp 0.131 0.058 h ps://doi.o g/10.1371/jou nal.pone.0196238. 001 Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 6 / 15 Discussion In his s udy, we ound disc epancies be ween he in i o and in i o e iciencies o some sgRNAs. These disc epancies sugges ed he p esence o cellula ac o s which limi mu agene- sis, and encou aged us o analyze ch oma in in ol emen in mo e de ail a he ansc ip omic and epigenomic le els. Because he ansc ip coun s o he ea ly emb yo can be masked by he p esence o ma e nal ansc ip s, i is di icul o es ablish he exac ela ionship be ween gene exp ession and mu agenesis e iciency [38]. Howe e , we ound weak bu signi ican co ela- ions o gene exp ession wi h mu agenesis e iciency du ing ea ly de elopmen , wi h he s on- ges co ela ion a he oblong sphe e s age (3.66hp , Spea man co ela ion 0.227; p- alue 0.001) and la e a 36hp (Spea man co ela ion 0.230; p- alue 0.001). The co ela ion a he oblong sphe e s age sugges s ha genes which become ac i e a he MZT a e mo e accessible o Cas9 and hence unde go mo e e icien mu agenesis. As ch oma in s uc u e is complex, i s e ec on a ge si e accessibili y has o be de e mined o each s uc u al le el, s a ing wi h di ec modi ica ions o DNA bases, con inuing wi h analysis o his one modi ica ions signaling o open ch oma in, and ending wi h analysis o ch oma in accessibili y. A de ailed analysis is equi ed in o de o unde s and how CRISPR-- Cas9 mu agenesis ac i i y could be manipula ed a molecula le el using o example chemical inhibi o s o his one deace ylase ac i i y. DNA me hyla ion is known o ma k ansc ip ional inac i i y and ec ui modi ied his ones a he exons [39]. In ou s udy, exon me hyla ion was no ound o signi ican ly in luence he ac i i y o mu agenesis in zeb a ish emb yos. In con- i ma ion, i has p e iously been sugges ed ha Cas9 can ac independen ly om DNA me h- yla ion in cell lines, and ha , in gene al, mos p o ein-DNA in e ac ions a e independen o DNA me hyla ion [7,40]. I DNA me hyla ion is no a limi ing ac o , we hypo hesized ha mu agenesis migh co ela e wi h highe o de s uc u es, speci ically his one modi ica ions. Va ious his one modi ica ions media e ansc ip ional ac i a ion and ep ession, and o m nucleosome s uc u es, which bind ch oma in in o an inac i e he e och oma in s a e. H3K4me3 is a well known modi ica ion occu ing in ea ly de elopmen [41]. The mos s ongly sugges i e, albei no signi ican , co ela ion be ween expe imen al da a and CRISPR-Cas9 mu agenesis e iciency was ound wi h H3K4me3 da a (Spea man co ela ion 0.263, p- alue 0.07) [32]. This was expec ed, gi en he associa ion wi h ansc ip ional ac i i y a ea ly de elopmen al s ages. As mu agenesis can al eady be de ec ed a 1hp i is possible ha we ail o see a s onge co ela ion because he inspec ed imepoin is la e and he his one landscape a 75–80% epiboly is dissimila o ha which is p esen be o e he MBT. In addi ion, i da a om mul iple imepoin s would be a ailable i would p o ide a mo e comp ehensi e iew o opening o local ch oma in s uc u es. Also, obse ing only H3K4me3 signals migh no accu a ely e lec he ch oma in s a e in ea ly emb yos, as he e a e also o he his one ma ks o open and closed ch oma in, including H3K9me3 and H3K27me3 as well as H3K27ac a p omo e s [28,32,42]. A wide scale analysis o his one modi ica ions could p o- ide mo e insigh in o he associa ion o CRISPR-Cas9 e iciency wi h his one landscape. A highe o de s uc u e abo e he his one landscape is shaped by modi ied his ones o ga- nizing in o nucleosomes. Nucleosome occupancy, b ea hing and emodeling ha e p e iously been ound o a ec he clea age ac i i y o Cas9 and consequen ly, CRISPR-Cas9 mu agenesis is mo e success ul when a ge ing he sequences deple ed in nucleosomes [25,26,43]. The posi- ion o he PAM-sequence ela i e o nucleosomes has been ound o be a key de e minan o he Cas9 endonuclease ac i i y in i o bu no in zeb a ish [17,23]. Nucleosomes a ec ch o- ma in accessibili y, which can be measu ed using ATAC-seq [44]. This s a e-o - he-a me hod has been used o iden i ica ion o accessible ch oma in egions du ing ea ly de elopmen [37]. Using he publicly a ailable da a, we ound a weak bu signi ican co ela ion be ween Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 7 / 15 ch oma in accessibili y and mu agenesis e iciency a he MBT, indica ing ha ch oma in in luences he e iciency o CRISPR-Cas9 mu agenesis in zeb a ish emb yos, e en hough i is no he sole de ining ac o (Table 2) [37]. Ou esul s a e in line wi h hose by o he s [17] wi h di e en analysis me hod and da ase . Mo eo e , ou esul s sugges CRISPR-Cas9 mu agene- sis e iciency o be independen o exon me hyla ion and H3K4me3 a p omo e s. Deciphe ing he e ec o de elopmen al ch oma in on he ac i i y o CRISPR-Cas9 mu a- genesis model o ganisms ul ima ely leads us o an unanswe ed ques ion abou he egula ion o zygo ic genome ac i a ion and he signals ha egula e his e en a ea ly s ages be o e he MTZ [45]. The genome emains in a ansc ip ionally inac i e s a e be o e he MZT, and i is likely ha his inac i e ch oma in also limi s he access o mu agenesis eagen s such as Cas9. I is also likely ha Cas9 can gain access du ing eplica ion, and a si es ha con ain mo e pe - missi e his one modi ica ions o a e deple ed in nucleosomes, bu only wi h limi ed e icacy. Wi h u he cell di isions, ch oma in ep essi e signals hen become dilu ed, leading o ch o- ma in opening a he MZT and ini ia ion o ansc ip ion [45]. Despi e he biological signi i- cance o he MBT and MZT, we we e al eady able o see mu agenesis aking place a 1hp o some genes, so we p opose ha (when designing CRISPR-Cas9 mu agenesis s a egies) ch o- ma in s uc u e should be aken in o accoun a a e y ea ly imepoin (Fig 2). Se e al s udies ha e looked in o he co ela ion o in silico p edic ions and in i o ac i i y o sgRNAs and ound ha CRISPR-sgRNA design ools o en ail o accu a ely p edic sgRNA ac i i y [11,20,25]. Mo eo e , i has been obse ed, ha he in silico p edic ions which a e e i- cien o model o ganisms a e no e icien o cell line based assays and ice e sa [11]. As Haeussle e al. (2016) obse ed, CRISPR-Cas9 e iciency in mice is be e p edic ed by he algo i hms ha ha e been ained on zeb a ish expe imen al da a, han by cell line based algo- i hms. I is logical o assume his is a leas in pa due o he ac ha mice and zeb a ish unde go simila , conse ed de elopmen al dynamics a he ansc ip omic and epigenomic le el (a he ime when CRISPR-mu agenesis is aking place), and a ge si e accessibili y is la gely de ined by ea ly ch oma in. Thank ully, design ools, which also ake in o accoun a - ge si e accessibili y, ha e ecen ly become a ailable [11,14,16,17]. De ailed analysis is equi ed o pinpoin which a e he mos impo an ch oma in s uc u es impac ing CRISPR-Cas9 ac i - i y. Wi h a be e unde s anding o hese, we will hope ully achie e imp o emen s in p edic- ions o expe imen al design especially in he in i o models. E en ually, i migh be possible o modi y local ch oma in o inc ease a ge si e accessibili y and simul aneously dec ease he likelihood o o - a ge binding. Ou esul s con i m he in ol emen o ch oma in in de ining CRISPR-Cas9 mu agenesis e iciency in a e eb a e model in i o. Ma e ials and me hods Zeb a ish main enance Wild ype AB ish we e main ained in a low- h ough sys em wi h a ligh /da k cycle o 14h/ 10h acco ding o he s anda d p ocedu e. Emb yos and la ae we e g own in an incuba o (28.5˚C) in emb yonic medium/E3 wa e (5mM NaCl, 0.17mM KCl, 0.33mM CaCl 2 , 0.33mM MgSO 4 , and 10–15% Me hylene Blue). E hics s a emen and da a a ailabili y All expe imen s we e ca ied ou in acco dance wi h he EU-di ec i e 2010/ 63/EU on he p o- ec ion o animals used o scien i ic pu poses, and wi h he Finnish Ac on he P o ec ion o Animals Used o Scien i ic o Educa ional Pu poses (497/2013) and he Go e nmen Dec ee on he P o ec ion o Animals Used o Scien i ic o Educa ional Pu poses (564/2013). We ha e only used zeb a ish p io o hei independen ly eeding la al s ages in his s udy, which hus Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 8 / 15 do no equi e animal pe mi s. Pe mi o he zeb a ish housing and main enance o he acil- i y a he Uni e si y o Tampe e is ESAVI/10079/04.10.06/2015. The compu a ional da a analysed in his s udy we e collec ed om open access sou ces, as de ailed in he app op ia e sec ions. Design and p oduc ion o sgRNAs o CRISPR/Cas9 media ed genome edi ing Ta ge sequences (S1 Table) o sgRNA design we e chosen using he online based CRISPR design ool (h p://c isp .mi .edu/), ChopChop.V1 o V2 [14,15] o CRISPRscan [9]. Ta ge si e uniqueness was e i ied wi h he NCBI BLAST analysis agains he zeb a ish genome (GRCz10). sgRNAs we e p oduced as desc ibed p e iously [46]. B ie ly, he sgRNA oligo (Sigma-Ald ich) and he T7 p omo e si e oligo (S1 and S2 Tables) (Sigma-Ald ich) we e annealed and in i o ansc ibed using he MEGAsho sc ip T7 T ansc ip ion Ki (Ambion Li e Technologies, CA, USA). The in eg i y and size o he p oduced sgRNAs we e analyzed wi h gel elec opho esis (1% aga ose in T is-ace a e-EDTA, TAE). The concen a ion o he sgRNAs was measu ed wi h he Qubi 1RNA BR Assay ki (The mo Fishe Scien i ic, MA USA 02451) and Nanod op 2000 (The mo Fische Scien i ic). sgRNA and Cas9 mic oinjec ion and genomic DNA ex ac ion The sgRNAs and he Cas9 p o ein (ToolGen Inc., Seoul, Sou h Ko ea) we e co-injec ed in o one-cell s age zeb a ish emb yos wi h a mic o injec o (PV830 Pneuma ic PicoPump, Wo ld P ecision Ins umen s) unde a Nikon mic oscope (SMZ645), using bo osilica e needles p e- pa ed wi h a Flaming/B own mic opipe e pulle . Needles we e calib a ed by injec ing solu ion in o a haloca bon oil d ople o achie e a diame e o 12μm (app oxima ely 1nl). The emb yos we e aligned on 1.2% aga ose E3 wa e pla es p io o he injec ion. An injec ion solu ion con- aining 130ng/μl sgRNA and 250ng/μl o he Cas9 p o ein in nuclease- ee wa e was incu- ba ed 37˚C 15min. Rhodamine dex an was added o he solu ion o he isualiza ion o he injec ions unde a Zeiss Luma V12 luo escence mic oscope. To analyze he onse o he mu agenesis 10–20 CRISPR-Cas9 injec ed emb yos we e collec ed and ozen in liquid ni o- gen o DNA ex ac ions a 1, 2, 3, 4, 6hp (hou s pos e iliza ion). To analyze he in i o mu agenesis e iciency, 5 emb yos we e collec eda 8hp and immedia ely ozen in liquid ni ogen. Fo DNA ex ac ion, he emb yos we e lysed 4h 55˚C in lysis bu e (10mM T is pH 8,2, 10mM EDTA, 200mM NaCl, 0.5% SDS, 200μg/ml P o einase K). DNA was p ecipi a ed 1h -20˚C using wo olumes o e hanol. DNA was hen pelle ed by cen i uging 16,000g 10min. The pelle was washed wi h 200μl o 70% e hanol be o e esuspending in 200μl o wa e . A pu i ica ion s ep wi h phenol-chlo o o m was pe o med a e ea men wi h 15u o RNase A (The mo Fische Scien i ic) pe 100μl o sample, 1h 37˚C. He e oduplex mobili y assay Ta ge ed loci we e ampli ied om he genomic DNA by PCR using he Maxima Ho S a DNA polyme ase (The mo Fische Scien i ic) acco ding o he manu ac u e ’s ins uc ions. The PCR p ime s (S3 Table) we e designed o anneal ups eam and downs eam o he expec ed cu ing si e. The PCR p oduc was pu i ied using Exo I and Fas AP (The mo Fische Scien i ic) ea men 15min 37˚C, hen 15min 85˚C. 10μl o he pu i ied PCR p oduc was annealed in a eac ion con aining 1x NEBu e 2 (New England Biolabs, MA, USA) and was un on a 10% polyac ylamide gel. The gel was s ained wi h GelRed (Bi ium Inc., F emon , CA). Ch oma in accessibili y is associa ed wi h CRISPR-Cas9 e iciency in he zeb a ish (Danio e io) PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0196238 Ap il 23, 2018 9 / 15