Packaging of Dinoroseobacter shibae DNA into Gene Transfer Agent Particles Is Not Random.

Packaging o Dino oseobac e shibae DNA in o Gene T ans e Agen Pa icles Is No Random Ju¨ genTomasch 1,3, *, Hui Wang 1 , Ap il T.K. Hall 2 ,DianaPa zel 1,6 , Ma hias P eusse 3 ,Jo¨ nPe e sen 4 , Henne B inkmann 4 , Boyke Bunk 4 , Sabin Bhuju 5 ,MichaelJa ek 5 , Robe Ge e s 5 , And ew S. Lang 2 ,and I ene Wagne -Do¨ble 1 1 G oup Mic obial Communica ion, Helmhol z-Cen e o In ec ion Resea ch, B aunschweig, Ge many 2 Depa men o Biology, Memo ial Uni e si y o New oundland, S John’s, New oundland and Lab ado , Canada 3 Depa men o Molecula Bac e iology, Helmhol z-Cen e o In ec ion Resea ch, B aunschweig, Ge many 4 Depa men o Mic obial Ecology and Di e si y Resea ch, Leibniz Ins i u e DSMZ—Ge man Collec ion o Mic oo ganisms and Cell Cul u es, B aunschweig, Ge many 5 G oup Genome Analy ics, Helmhol z-Cen e o In ec ion Resea ch, B aunschweig, Ge many 6 P esen add ess: Sa o ius S edim Bio ech GmbH, Go¨ ingen,Ge many *Co esponding au ho : E-mail: jue gen. omasch@helmhol z-hzi.de. Accep ed: Janua y 8, 2018 Abs ac Gene ans e agen s (GTAs) a e phage-like pa icles which con ain a agmen o genomic DNA o he bac e ial o a chaeal p oduce and deli e his o a ecipien cell. GTA gene clus e s a e p esen in he genomes o almos all ma ine Rhodobac e aceae (Roseobac e s) and migh be impo an con ibu o s o ho izon al gene ans e in he wo ld’s oceans. Fo all o ganisms s udied so a , no ob ious e idence o sequence speci ici y o o he non andom p ocess esponsible o packaging genomic DNA in o GTAs has been ound. He e, we show ha knock-ou o an au oinduce syn hase gene o Dino oseobac e shibae esul ed in o e p oduc ion and elease o unc ional GTA pa icles (DsGTA). Nex -gene a ion sequencing o he 4.2-kb DNA agmen s isola ed om DsGTAs e ealed ha packaging was no andom. DNA om low-GC conjuga i e plasmids bu no om high-GC ch omids was excluded om packaging. Se en ch omosomal egions we e s ongly o e ep esen ed in DNA isola ed om DsGTA. These packaging peaks lacked iden i iable conse ed sequence mo i s ha migh ep esen ecogni ion si es o he GTA e minase complex. Low-GC egions o he ch omosome, including he o igin and e minus o eplica ion, we e unde ep esen ed in DNA isola ed om DsGTAs. DNA me hyla ion educed packaging equency while he le el o gene exp ession had no in luence. Ch omosomal egions ound o be o e - and unde ep esen ed in DsGTA-DNA we e egula ly spaced. We p opose ha a “head ul” ype o packaging is ini ia ed a he si es o co e age peaks and, a e linea iza ion o he ch omosomal DNA, p oceeds in bo h di ec ions om he ini ia ion si e. GC-con en , DNA-modi ica ions, and ch oma in s uc u e migh in luence a which sides GTA packaging can be ini ia ed. Key wo ds: gene ans e agen , GTA, Roseobac e , ho izon al gene ans e . In oduc ion P oka yo es, e en di e en s ains o one species, o en di e ema kably in hei genomic in en o ies. This a iabili y in ge- nome con en canno be explained by gene loss alone (Dagan and Ma in 2007) bu is he esul o a conside able amoun o ho izon al gene ans e (HGT), which has been shown o be a main d i e o p oka yo e e olu ion (McIne ney e al. 2017). Mechanisms o HGT in p oka yo es include ans o ma ion, ha is, di ec up ake o naked DNA om he en i onmen , conjuga ional ans e o plasmids ha equi es di ec con ac o dono and ecipien cells (Thomas and Nielsen 2005), and ansduc ion by phages (Canchaya e al. 2003). In 1974 an addi ional, unique way o HGT o genomic DNA, independen o cell–cell con ac o he p esence o naked DNA in he en i- onmen , was disco e ed in Rhodobac e capsula us (Ma s 1974). La e , he esponsible R. capsula us gene ans e agen (RcGTA) was ound o esemble a small ailed phage ha con- ains agmen s o cellula genomic DNA ins ead o i al DNA ßThe Au ho (s) 2018. Published by Ox o d Uni e si y P ess on behal o he Socie y o Molecula Biology and E olu ion. ThisisanOpenAccessa icledis ibu edunde he e mso heC ea i eCommonsA ibu ionNon-Comme cialLicense(h p://c ea i ecommons.o g/licenses/by-nc/4.0/),whichpe mi snon- 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 p o ec ed] Genome Biol. E ol. 10(1):359–369. doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 359 GBE Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 (Yen e al. 1979). Since hen, Gene ans e agen (GTA) pa - icles o di e en e olu iona y o igins ha e been iden i ied in a ious o he bac e ia (Humph ey e al. 1997;Bie s e al. 2008) and he a chaeon Me hanococcus ol ae (Be ani 1999). In Alphap o eobac e ia, wo phylogene ically un ela ed ypes o GTAs ha e been iden i ied, namely he Rhodobac e ales GTA, wi h he bes s udied RcGTA, and he Ba onella GTA (BaGTA). Those ela ed o RcGTA ha e been iden i ied in many Alphap o eobac e ia and a e pa ic- ula ly widesp ead in membe s o he o de Rhodobac e ales whe e hey a e o ganized in a conse ed gene clus e o ap- p oxima ely 15 kb (Lang and Bea y 2007;New on e al. 2010;Huang e al. 2011). As RcGTA packages only app ox- ima ely 4 kb o DNA i has o be conside ed non-sel - ans- missible (Lang e al. 2017). GTAs isola ed om Rhodobac e aceae o he Roseobac e g oup ans e ed e- sis ance ma ke s wi h high equencies o bo h cul u es and na u al communi ies o ma ine bac e ia, sugges ing an impo - an ole o GTAs o he gene low in he wo ld’s oceans (McDaniel e al. 2010). Fu he mo e, phylogenomic analysis iden i ied GTAs as he main sou ce o genes acqui ed h ough HGT in membe s o he genus Phaeobac e , consis ing mainly o ma ine su ace colonize s (F eese e al. 2017). The Rhodobac e ales GTAs p obably e ol ed om a p o- phage o he Sipho i idae amily ha los genes o in eg a ion in o and excision om he hos DNA as well as o eplica ion, whe eas genes o DNA packaging and s uc u al componen s we e e ained (Huang e al. 2011), and hese genes we e sub- sequen ly b ough unde he con ol o he hos gene egula- o y sys ems (Me ce e al. 2012).InR. capsula us,genes coding o he compe ence sys em, needed o up ake o DNA om GTAs by he ecei ing cell, a e unde con ol o he same eg- ula o s as he GTA genes (B imacombe e al. 2015). Recen ly i has been ound ha he RcGTA is induced by nu ien deple- ion, possibly as a consequence o inc eased le els o guanosine- e aphospha e (ppGpp) (Wes bye e al. 2017). No ably, only a small subse o a R. capsula us cul u e (<3%) exp esses RcGTA genes, leading o lysis o he p oducing cells (Fogg e al. 2012;Hynes e al. 2012). A second phylogene ically dis inc alphap o eobac e ial GTA has been ound in he in acellula bac e ia om he genus Ba onella (Be glund e al. 2009;Tama i e al.2017). As in R. capsula us, BaGTA p oduc ion is limi ed o a small ac ion o he popula ion and hese p oduce s lyse o elease GTA pa icles (Queba e e al. 2017). The au ho s sugges ed ha he i es subpopula ion o B. henselae cells, indica ed by low le els o ppGpp, sac i ices i sel o sp ead bene icial mu a ions ia GTAs h oughou he popula ion o ac i ely di iding cells. The mechanism o packaging DNA in o GTA pa icles is cu en ly no known. Based upon homology o key compo- nen s like e minase and po al p o eins i is assumed ha i esembles he “head ul” ype o packaging ha is used by many double-s anded DNA phages (Lang e al. 2017). The pa s o he genomes ha a e mobilized by packaging in o GTAs ha e been de e mined o R. capsula us as well as B. g ahamii and B. henselae. These s udies sugges andom packaging o hos genomic DNA in o GTAs. Ma ke ans e and DNA diges ion analysis in he p ege- nomic e a sugges ed andom packaging o genomic DNA in o RcGTAs (Ma s 1974;Humph ey e al. 1997;Be ani 1999). These ini ial indings la e gained suppo by a s udy employ- ing whole-genome mic oa ay hyb idiza ion o RcGTA-DNA, which showed ha only he GTA gene clus e i sel was un- de ep esen ed, by 25%, in packaged DNA, whe eas he emaining genomic DNA was andomly packaged (Hynes e al. 2012). Since he GTA clus e is highly ansc ibed in he subse o he popula ion ha is esponsible o p oducing he GTA pa icles, i was hypo hesized ha access o he GTA encoding DNA was blocked by he RNA polyme ase complex (Hynes e al. 2012). Fo BaGTA, a i ulence egion impo an o hos - adap abili y o he mouse pa hogen B. g ahamii was ound o be o e ep esen ed in GTA-DNA (Be glund e al. 2009). Genomic DNA o his egion was p esen in highe copy num- be in he cells due o un-o eplica ion om a phage-de i ed o igin o eplica ion, hus packaging in his case can be conside ed andom bu dependen on he copy numbe o he DNA (Be glund e al. 2009). A ecen ansposon se- quencing s udy in B. henselae con i med he ela ionship be ween DNA copy numbe and equency o packaging o ch omosomal DNA bu addi ionally showed ha DNA om plasmids is no ans e ed by BaGTAs (Queba e e al. 2017). The genome o he dino lagella e-associa ed bac e ium Dino oseobac e shibae consis s o a 3.79-Mb ch omosome and i e ex ach omosomal eplicons (Wagne -Do¨ ble e al. 2010). Two o hem (153 and 72 kb) ha e a simila high GC- con en (65–68%) and codon-usage as he ch omosome and a e he e o e classi ied as “ch omids”’ acco ding o Ha ison e al. (2010). The o he h ee (191, 126, and 86 kb) ha e a lowe GC-con en (60–61%) and de ia ing codon-usage and ep esen ypical plasmids (Pe e sen e al. 2013). The 191- and 126-kb plasmids ha bo ype IV sec e ion sys ems ha a e employed o hei conjuga ional ans e (Pa zel e al. 2016). A comple e GTA gene clus e is p esen on he D. shibae ch omosome (Wang e al. 2014). Simila o R. capsu- la us (Me ce e al. 2012), he D. shibae GTA (DsGTA) gene clus e is unde he con ol o he C A phospho elay (Wang e al. 2014). This signal- ansduc ion sys em is in eg a ed in o he D. shibae quo um sensing (QS) sys em as i is induced by he p oduc o he au oinduce syn hase LuxI 1 (Pa zel e al. 2013) and hen induces exp ession o wo addi ional QS e- sponse egula o /au oinduce syn hase ope ons, luxR/I 2 ,and luxB/I 3 (Wang e al. 2014). He e, we show ha knockou o he au oinduce syn hase gene luxI 2 esul s in o e p oduc ion o DsGTAs. Using nex -gene a ion sequencing, we analyze pack- aging o D. shibae DNA in o GTA pa icles. Tomasch e al. GBE 360 Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 Ma e ials and Me hods Bac e ial S ains and Plasmids Dino oseobac e shibae s ains used in his s udy, E. coli s ains, plasmids and p ime s used o cons uc ion o he D. shibae DluxI 2 (Dshi_2852) dele ion s ain a e lis ed in supple- men a y able S1,Supplemen a y Ma e ial online. D. shibae DluxI 2 was gene a ed by eplacing he gene wi h a gen amy- cin esis ance casse e ia double homologous ecombina ion as p e iously desc ibed (Pa zel e al. 2013). The knock-ou ec o pJB5603luxI 2 :: Gm was subcloned in E. coli ST18 and in oduced in o D. shibae DFL12 ia conjuga ion. Resequencing o he knock-ou s ain con i med dele ion o he wild- ype gene, inse ion o he esis ance casse e in o he co ec locus and iden i ied one seconda y mu a ion. De ails on knock-ou cons uc ion and con i ma ion can be ound in supplemen a y me hods, no e S1 and igu e S1, Supplemen a y Ma e ial online. Cul i a ion Dino oseobac e shibae s ains we e g own in 1=2Di co Ma ine B o h 2216 (BD, Spa ks, MD) o a i icial seawa e medium (ASM, supplemen a y me hods)a 30Cand 160 pm. Ruege ia pome oyi was g own in 1=2yeas ex ac - yp one-sea sal (YTSS) medium (4 g l 1 yp one, 2.5 g l 1 yeas ex ac , and 15 g l 1 sea sal s) a 20 C and 160 pm. De ec ion o he GTA Majo Capsid P o ein by Wes e n Blo Dino oseobac e shibae s ains we e g own o 19 h in ASM. Samples om Ruege ia pome oyi,g ownin1=2YTSS o 19 h, we e included as posi i e con ols as his o ganism is known o p oduce GTA pa icles (Bie s e al. 2008) and has de ec able GTA capsid p o ein (Fu e al. 2010). Cells om 0.4 ml o each cul u e we e pelle ed and esuspended in an equal olume o TE bu e . Cul u e supe na an s we e p epa ed om cul u es by cen i uga ion o 2 min a 17,000 g, wi h emo al o a ac ion o he esul ing supe na an o a new ube. The sam- ples (5 and 10 ll o cells and supe na an s, espec i ely) we e mixed wi h SDS-PAGE sample bu e (NEB) and incuba ed 5mina 98C be o e elec opho esis. SDS-PAGE and immu- noblo ing we e pe o med as desc ibed p e iously (Fu e al. 2010) using he p ima y an ibody a ge ing a egion o he GTA majo capsid p o ein (Ag ise a AB, V€ ann€ as, Sweden) ha is conse ed among many Rhodobac e ales.Theimageswe e cap u ed using he ImageQuan LAS 4000 ins umen (GE Li e Sciences, Uppsala, Sweden) and we e each uni o mly ad- jus ed o b igh ness and con as . Gene T ans e Bioassays Gene ans e agen dono s ains we e g own o 24 h in ASM a 30 C and 160 pm. The ecipien s ain, D. shibae DFL12, was g own unde he same condi ions in 1/2 Mb. The dono cul u es we e passed h ough 0.22-mm il e sand 0.5 ml o his il a e was added o 0.1 ml o he ecipien cells. The mix u e was hen incuba ed a 30 C o 1 h wi hou shaking, o 4 h and 160 pm, ollowed by addi ion o 0.9 ml ASM and a u he incuba ion a 30 C and 160 pm o 18 h. The cells we e hen pla ed on 1/2 Mb wi h gen a- mycin sul a e (150 mgml 1 ) and incuba ed a 30 C o 2–4 days un il colonies we e isible o coun ing. The nega i e con ols o hese assays we e il a es o he gen amycin- sensi i e s ain, DFL12, and he gen amycin- esis an s ain, DluxI 1 , added o he ecipien cells. Fil a e-only con ols ha con ained no ecipien cells we e also included o con i m no cells passed h ough he 0.22-mm il e s. Pu i ica ion o DsGTA Pa icles o Elec on Mic oscopy Dino oseobac e shibae DluxI 2 was g own 1/2 Mb o 29 h a 30 C wi h shaking a 160 RPM. The cells we e pelle ed by cen i uga ion and NaCl was added o he supe na an o a inal concen a ion o 1 M. Polye hylene glycol (PEG) 6000 was added o make a inal concen a ion o 10% w/ and dissol ed by s i ing a oom empe a u e. The solu ion was hen incu- ba ed a 4 C o 18 h and he p ecipi a ed ma e ial pelle ed by cen i uga ion a 10,000 g o 20mina 20C. The supe - na an was pou ed o and pelle ed ma e ial was esuspended in 0.1 M ammonium ace a e (pH 7) wi h shaking a oom em- pe a u e o 2 h ollowed by gen le pipe ing. This ma e ial was cen i uged a 2,400 g o 2 min a oom empe a u e and he supe na an collec ed in o a new ube. T i on X-100 was added o a inal concen a ion o 10% ( / ) wi h gen le mixing by in e sion. This was cen i uged a 1,10,000 g o 2 h a 20 C, and mos o he supe na an hen emo ed. The pelle was esuspendedin he emainingsolu ion(0.5 ml) by gen le pipe ing and he ube e illed wi h 0.1 M ammonium ace a e. The ube was cen i uged again o 1 h and he p e ious s eps epea ed. A e his hi d cen i uga ion, almos all o he su- pe na an was emo ed and he pelle le o esuspend in he small amoun o emaining liquid wi h shaking a 4 C o 18 h. The ma e ial was homogenized by gen le pipe ing, ans- e ed o a mic o uge ube, and cen i uged a 17,000 g o 1 min a oom empe a u e. The supe na an was collec ed o elec on mic oscopy imaging a he Uni e si y o Guelph (Canada) Molecula and Cellula Imaging Facili y. A 5-mlsample was placed on o a 200 mesh coppe g id wi h o m a /ca bon coa ing, adso bed o 1 min, and blo ed o wi h il e pape . The g id was hen loa ed on a d op o 2% u anyl ace a e. The sample was iewed in a FEI Tecnai F20 elec on mic oscope a 200kV.Theimageswe ecollec edwi haGa an4Kbo om- moun came a using he Ga an Digi alMic og aph so wa e. Isola ion o DsGTA-DNA DsGTA-DNA was isola ed om h ee independen cul u es. The cul u es we e ea ed as o he Elec on mic oscopy up o Gene T ans e Agen s o Dino oseobac e shibae GBE Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 361 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 he poin o esuspension o he PEG p ecipi a es. In his case, he pelle s we e esuspended in TE bu e wi h gen le shaking a oom empe a u e o 2 h. F ee nucleic acids we e emo ed by ea men wi h DNase I (0.01 U ll 1 ) and RNase A (0.08 U ll 1 )a 37Cin1DNase bu e (New England Biolabs) o 1 h. The sample was hea ed a 75 C o 15min a e addi ion o EDTA o a inal concen a ion o 5 mM. Nucleic acids we e pu i ied by phenol: chlo o o m: isoamyl alcohol (25:24:1) ex ac ion and e hanol p ecipi a ion. The pu i ied DNA was subjec ed o es ic ion enzyme diges ion using he enzymes EcoRI and B aI (New England Biolabs), acco ding o he manu ac u e ’s p o ocol. The un ea ed DNA sample was incuba ed unde he same condi ions (i.e., in 1 eac ion bu e a 37 C) wi h he addi ion o dH 2 Oin place o enzyme. The app oxima ely 4.2-kb band o DsGTA- DNA was excised om he gel and pu i ied wi h he Wiza d SV Gel and PCR Clean-Up Sys em (P omega, Fi chbu g, WI), ollowed by e hanol p ecipi a ion. Isola ion o Genomic DNA Genomic DNA was isola ed om 5 ml samples o wo o he D. shibae cul u es used o isola ion o DsGTA-DNA using he NucleoSpin issue ki (Mache y-Nagel, Du¨ en, Ge many) acco ding o manu ac u e ’s p o ocols. Isola ion o RNA RNA o ansc ip ome analysis o he h ee di e en D. shibae s ains was isola ed om wo independen cul u es. RNA o analysis o he ela ionship o DsGTA packaging and ansc ip- ion was isola ed om wo cul u es om which DsGTA-DNA was isola ed. Cells we e collec ed by cen i uga ion a 12,000 g o 1mina 4C, co e ed wi h 1 ml T izol (Ambion, Ge many), immedia ely ozen in liquid N 2 and s o ed a 70 C un il p ocessing. Fo RNA ex ac ion, cells we e ho- mogenized wi h 0.3 g o glass beads in he Fas P ep-24 in- s umen (MP Biomedicals, CA) a 6.0 m/s o 3 min and hen incuba ed o 5 min a oom empe a u e. Samples we e cen- i uged a 12,000 g o 10 min a 4 C and he supe na- an s we e ans e ed o esh ubes, ollowed by addi ion o 100 ml o 1-b omo-3-chlo op opane (Sigma-Ald ich, S . Louis, MO) and incuba ion o 10 min a oom empe a u e. Samples we e cen i uged a 12,000 g o 10 min a 4 C, a e which he aqueous phase was ans e ed o new ubes and mixed wi h 500 ml o absolu e e hanol. Ex ac s we e pu- i ied using RNeasy Mini ki (Qiagen, Hilden, Ge many) acco ding o he manu ac u e ’s ins uc ions. In addi ion, samples we e ea ed wi h DNAse I (Qiagen, Hilden, Ge many). Sequencing o RNA To al RNA was ea ed wi h RiboZe o (Bac e ia) ki (Illumina, San Diego, CA) ollowing he manu ac u e ’s p o ocol in o de o emo e ibosomal RNA om he samples. Single end, s and speci ic cDNA lib a ies we e p epa ed om RNA deple ed o al RNA using Sc ip seq 2 RNA-Seq Lib a y P epa a ion Ki (Illumina) ollowing he manu ac u e s p o ocol. Fo sequencing equal olume o lib a ies (12 PM) was mul iplexed on a single lane. Sequencing was done on he HiSeq 2500 (Illumina) using T uSeq SBS Ki 3—HS (Illumina) o 50 cycles esul ing in 50-bp eads. Image analysis and base calling we e pe o med using he Illumina pipeline 1.8 (Illumina). Sequencing o Genomic and DsGTA-DNA Pai ed end cDNA lib a ies we e p epa ed using he NEBNex Ul a DNA Lib a y P epa a ion Ki o Illumina (NEB, Ipswich, MA). Sequencing was done on he MiSeq 2000 (Illumina) using MiSeq Reagen Ki s 2 (Illumina) o 250 cycles esul ing in 2250 bp pai ed end eads. Image analysis and base calling we e pe o med using he Illumina pipeline 1.8 (Illumina). PacBio-Sequencing and DNA Modi ica ion Calling PacBio-sequencing was pe o med as p e iously desc ibed (Ba ling e al. 2017). Me hylome analysis was pe o med using he “RS_Modi ica ion_and_Mo i _Analysis.1” p o ocol in- cluded in SMRT Po al e sion 2.3.0. 99.5 and 96.6% o he iden i ied m6A and m4C mo i s we e me hyla ed, espec i ely. P ocessing o Sequence Da a The demul iplexed aw as q- iles we e quali y-con olled us- ing he FASTQ-mc sui e (h ps://gi hub.com/Exp ession Analysis/ea-u ils; las accessed Janua y 15, 2018). Low quali y bases (Ph ed-sco e <30) and iden i ied Illumina adap o s we e clipped om he sequences. Reads we e mapped o e e ence genomes using bow ie2 (Langmead and Salzbe g 2012) wi h de aul pa ame e s o single o pai ed end eads. Ambiguously mapping eads we e andomly dis ibu ed be- ween all egions o which hey could be assigned. The esul - ing sam- iles we e con e ed o indexed bina y o ma and pile-up o ma using sam ools (Li e al. 2009). Accession num- be s o e e ence D. shibae DNA sequences: ch omosome, 3.79 Mb [NC_009952.1]; pDSHI01, 191 kb [NC_009955.1]; pDSHI02, 153 kb [NC_009956.1]; pDSHI03, 126 kb [NC_009957.1]; pDSHI04, 86 kb [NC_009958.1]; pDSHI05, 72 kb [NC_009959.1]. S a is ical Analysis Pile-up iles we e loaded in o he R s a is ical en i onmen (R e sion 3.4.0). Rep oducibili y o biological eplica es was assessed by isual inspec ion o sca e plo s and boxplo s and by calcula ing Spea man ank co ela ion (supplemen a y igs. S7–S9,Supplemen a y Ma e ial online). Dis ibu ion o DsGTA and genomic co e age was compa ed wi h heo e ical da a ollowing a no mal dis ibu ion using he qqno m() Tomasch e al. GBE 362 Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 unc ion in R. The mean co e age by DsGTA-DNA, genomic DNA, and RNA as well as he mean GC con en we e calcu- la ed o sliding windows o 4 kb using he ollapply unc ion o he zoo package (Zeileis and G o hendieck 2005). GC- con en o sliding windows was calcula ed using he le e F equency unc ion o he Bios ings package (Page`s e al. 2017). Only me hyla ed si es wi h an iden i ica ion Ph ed sco e >30 (99.9% p obabili y o co ec iden i ica ion) we e used in his s udy and summed up o he sliding window ange. Spea man ank co ela ion was calcula ed o DsGTA co e age wi h GC-con en and log 2 - ans o med RNA co e - age, espec i ely. The au oco ela ion unc ion, a ailable om he R s a s package, calcula es he Pea son co ela ion om all pai s o obse a ions ha ha e he same dis ance om each o he . This is done o inc easing dis ances and makes i pos- sible o ind dis ances a which he co ela ion o wo obse - a ions (in his case DsGTA co e age) peaks, indica ing pe iodici y o he da a. Cumula i e GC/AT-skew o iden i i- ca ion o he o igin and e minus o eplica ion has been pe o med using he o iloc unc ion o he seqin package (Cha i and Lob y 2007). Code A ailabili y Linux shell and R sc ip s o analysis a e a ailable a h ps:// gi hub.com/Jue gen 79/GTA; las accessed Janua y 15, 2018. Iden i ica ion o Phage Genes The ool PHAST (Zhou e al. 2011) was used o iden i y phage de i ed genes in he genome o D. shibae. DNA Mo i Disco e y DNA sequence o he se en DsGTA co e age peaks was used as p ima y inpu o he mo i disco e y algo i hm o he MEME sui e 4.12.0 (Bailey e al. 2009). Mo i disco e y was un in he disc imina i e mode, wi h he se en peak egions indica ed in igu e 3Aas que y inpu and ei he he whole ch omosomal DNA sequence o DNA sequences o ch omo- somal egions wi h co e age be ween 200 and 500 eads/n as con ols (seconda y) inpu . The second con ol a oided in- clusion o low GC egions ha showed an ex emely low co e age. Da a A ailabili y RNA, genomic, and DsGTA Illumina sequencing da a ha e been made publicly a ailable a he Eu opean Nucleo ide A chi e (ENA, h p://www.ebi.ac.uk/ena; las accessed Janua y 15, 2018) unde he p ojec accession numbe PRJEB20656. PacBio da a used o e alua ion o D. shibae DluxI 2 (supplemen a y ig. S1 and no e S1, Supplemen a y Ma e ial online) has no been deposi ed in a public da abase ye bu will be a ailable om he au ho s upon eques . Resul s Only weak exp ession o he DsGTA gene clus e could be obse edinwild ypeD. shibae and i ually no exp ession was ound in he luxI 1 knockou . Howe e , when we knocked ou he C A-con olled luxI 2 gene (supplemen a y ig. S1 and no e S1, Supplemen a y Ma e ial online), s ong exp ession o he DsGTA gene clus e was obse ed ( ig. 1A). The DsGTA majo capsid p o ein could acco dingly be de ec ed in cell ex ac s and supe na an s o D. shibae DluxI 2 , indica ing ha his s ain p oduces and eleases GTA pa icles. A weak signal was also obse ed in cell-ex ac s, bu no supe na an s o he wild ype, indica ing a low GTA p oduc ion a e in he na i e s ain unde he condi ions s udied ( ig. 1B). The gen- amicin esis ance casse e used as he ma ke o knock-ou cons uc ion could be success ully ans e ed by cell- ee supe na an s o D. shibae DluxI 2 , bu no DluxI 1 , o he wild ype demons a ing ha DsGTAs a e unc ional and ha D. shibae cells a e able o ecei e and inco po a e DNA om he GTAs ( ig. 1C). A gen amycin- esis an s ain ob ained ia GTA-media ed ma ke ans e also p oduced and eleased inc eased numbe s o GTAs ela i e o he wild ype s ain, alida ing he ole o he luxI 2 gene in ep essing DsGTA p o- duc ion ( ig. 1D). DsGTA esembles a small ailed phage wi h a head diame e o app oxima ely 33 nm and a ail leng h o app oxima ely 48 nm ( ig. 1E). I packages DNA o 4.2 kb in size, simila o RcGTA-DNA. T ea men wi h es ic ion enzymes p oduced smea ing below his band, indica ing ha he DNA was double-s anded and ha i was no a speci ic sequence o DNA ( ig. 1F). We used nex -gene a ion sequencing o analyze whe he packaging o DNA in o DsGTA pa icles is andom, o shows p e e ences o dis inc genomic loci. To his end, DsGTA- DNA, genomic DNA, and RNA om he same samples we e sequenced and mapped on he D. shibae genome. This allowed us o de e mine he ela i e copy numbe o epli- cons, o add ess he possibili y o un-o eplica ion ound in in B. g ahamii and B. henselae (Be glund e al. 2009;Queba e e al. 2017), and o s udy he ela ionship be ween ansc ip- ional ac i i y and packaging equency simila o R. capsula us (Hynes e al. 2012). Fu he mo e, we used PacBio sequencing da a o de e mine me hyla ion o he D. shibae DNA. We calcula ed he mean co e age o sliding windows o 4 kb, o he ch omosome and ex ach omosomal eplicons. Co e age o he mul ipa i e genome o D. shibae by DNA isola ed om DsGTA pa icles showed a clea de ia ion om no mal dis ibu ion and a long ail o lowe co e age ( ig. 2A). Co e age by genome sequencing o he acco ding s ains mos ly ollowed a no mal dis ibu ion, wi h only ew egions being o e o unde ep esen ed (supplemen a y ig. S2,Supplemen a y Ma e ial online). Co e age dis ibu ion di e ed o he six D. shibae epli- cons ( ig. 2Band supplemen a y able S1,Supplemen a y Ma e ial online). The ch omosome showed he highes Gene T ans e Agen s o Dino oseobac e shibae GBE Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 363 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 a iabili y (be ween 10 and 763 eads/n , median: 181 eads/ n ); he 72-kb ch omid had a co e age simila o he ch omo- some (median: 192 eads/n ) while he 153-kb ch omid showed he highes abundance wi hin DsGTA-DNA (median: 1082 eads/n ). Co e age o he plasmids was low o e all (median be ween 4 and 6 eads/n ), bu indica ed ha a leas pa s o he 191- and 86-kb plasmids we e p esen in DsGTA- DNA (maximum: 237 and 258 eads/n , espec i ely) whe eas he 126-kb plasmid was almos absen (maximum: 8 eads/ n ). In he case o he 153-kb ch omid, he high co e age could pa ly be explained by he 50% highe copy numbe o his eplicon, whe eas he copy numbe o all o he epli- cons was close o one ch omosome equi alen ( ig. 2C). Nex , we analyzed he co e age o each eplicon in de ail, s a ing wi h he ex ach omosomal eplicons. Di e ences in DNA con en o a sequencing bias could be excluded as a sou ce explaining co e age di e ences o ch omids and plasmids ( ig. 2D). Bo h ch omids we e comple ely, albei no e enly, co e ed by GTA-DNA. In pa icula , co e age o he 153-kb ch omid showed a conspicuous gap be ween 95 and 102 kb. The 126-kb plasmid showed i ually no mapping eads and hus can be conside ed no o be packaged in o DsGTA pa - icles. The 191 and 86 kb showed a simila low co e age ex- cep o one 7.5-kb egion wi h e y high co e age and sha p bounda ies. We ound ha he DNA sequences o hese egions we e iden ical o he low co e age egion on he 153-kb ch omid. I con ains a ansposable elemen wi h he gene hiC in he cen e ha has been sp ead o he di - e en eplicons ( hiC locus, supplemen a y ig. S3A, Supplemen a y Ma e ial online). Thus, andom mapping should lead o an app oxima ely equal dis ibu ion o he am- biguous eads be ween he h ee pa alogous a ge loci. Assigning all he eads o he locus on he 153-kb ch omid would ill he gap on his ch omid and esol e obse ed dis- c epancies o ead mapping (supplemen a y ig. S3 and no e S2, Supplemen a y Ma e ial online). The e o e, we conclude ha he la ge majo i y o eads assigned o he hiC locus o igina ed om he 153-kb ch omid, and ha plasmids a e i ually excluded om being packaged in o GTA pa icles. Plasmids showed a 60% and 40% highe equency o FIG.1.—Biosyn hesis o unc ional GTAs in D. shibae DluxI 2 .(A) Exp ession o he DsGTA gene clus e in D. shibae DFL12, DluxI 1 ,andDluxI 2 as de e mined by s and-speci ic RNA-seq. Co e age o he nega i e s and is shown. DsGTA genes conse ed in Rhodobac e ales a e displayed as g ey a ows. Genes a e numbe ed acco ding o he e e ence R. capsula us; p edic ed unc ions a e indica ed on he igh . Rep oducibili y o RNA sequencing is shown in supplemen a y ig. S7,Supplemen a y Ma e ial online. (B) De ec ion o he GTA majo capsid p o ein in cell ex ac s and cul u e supe na an s o Ruege ia pome oyi (Rp, posi i e con ol) and he a o emen ioned D. shibae s ains using wes e n blo ing. (C) T ans e o a gen amycin- esis ance casse e o D. shibae DFL12 by supe na an s om di e en D. shibae s ains. Indi idual da a om h ee (DFL12, DluxI 2 )and wo(DluxI 1 ) independen expe imen s is shown. (D) T ans e o he gen amycin- esis ance casse e dis up ion o luxI 2 ia GTA esul s in o e p oduc ion o GTAs in he esul ing ecipien s ain (AH1). (E) Elec on mic og aph showing wo DsGTA pa icles. The ilamen ous s uc u es a e mos likely lagella. (F) Visualiza ion o he DNA isola ed om GTAs: un ea ed DNA (U) and DNA diges ed wi h BamHI (B) and EcoRI (E) a e shown. Tomasch e al. GBE 364 Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 me hyla ed adenosines wi hin 4-kb sliding windows han he ch omids and he ch omosome, espec i ely. The equency o me hyla ed cy osines was simila o all eplicons (supple- men a y ig. S4 and ables S2 and S3, Supplemen a y Ma e ial online). The ch omosome showed a clea ly non andom co e age wi hin he DsGTA-DNA ( ig. 3A). In con as , genome se- quencing demons a ed a egula pa e n o co e age and hus un-o eplica ion, o some o he cause o di e ences in he amoun s o di e en egions o he ch omosome being p esen in cells, could be excluded as a po en ial explana ion o he obse ed pa e n in he DsGTA-DNA co e age ( ig. 3A). Fou almos egula ly spaced egions we e clea ly unde ep esen ed, including he o igin and e minus o epli- ca ion ( ig. 3A,supplemen a y ig. S5,Supplemen a y Ma e ial online). In addi ion, he GTA gene clus e i sel showed e y low packaging. In con as , se en egions we e p e e en ially packaged in o DsGTA pa icles. Be ween hose clea ly o e - and unde ep esen ed egions, se e al local max- ima and minima exis . Regions wi h low co e age ended o ha e a low GC-con en ( ig. 3B), consis en wi h he low co - e age o he low-GC plasmids. The cooccu ence o local min- ima in GC-con en and GTA packaging was mos ob ious o he o igin and e minus o eplica ion. A local minimum in GTA co e age wi hin peak 4 could be obse ed coinciding wi h a local minimum o GC-con en ( ig. 3A). On he o he hand, less equen packaging o he GTA clus e could no be explained by GC-con en . O e all, he s ong posi i e co ela- ion be ween packaging and GC-con en was highly signi i- can ( ig. 3D). We could no iden i y DNA mo i s such as possible ecogni ion si es o es ic ion enzymes, phage-like ecogni ion si es, o epea s ha we e en iched in egions wi h high packaging a e using he MEME sui e o sequence mo i disco e y. Besides he GTA gene clus e , one de ec i e phage egion, lacking a phage eplicase gene, was iden i ied in he D. shibae genome a posi ion 98.7–102.8 kb, jus ou - side peak 1 (supplemen a y able S4,Supplemen a y Ma e ial online). Genes encoding ansposable elemen s o RNAs we e also no ound wi hin he DsGTA co e age peaks (sup- plemen a y able S5,Supplemen a y Ma e ial online). The o - igin and e minus o eplica ion showed a highe equency o me hyla ed adenosines (m6A) ( ig. 3C), whe eas he e- quency o me hyla ed cy osines (m4C) lacked such a pa e n (supplemen a y ig. S6,Supplemen a y Ma e ial online). The nega i e co ela ion be ween packaging in o DsGTAs and he numbe o m6A-si es was o e all lowe han o he FIG.2.—Packaging o DNA o igina ing om di e en eplicons. (A) Quan ile-quan ile plo compa ing log 10 - ans o med sequencing ead-co e age o DsGTA-DNA o a heo e ical no mal dis ibu ion (uppe panel) and his og am showing dis ibu ion o log 10 -co e age (lowe panel). Mean alues o 4-kb bins o e lapping by 2-kb a e shown. Red line in he uppe panel indica es no mal dis ibu ed da a. (B) Dis ibu ion o log 10 -co e age o DsGTA-DNA by eplicon. A e age GC-con en o each eplicon is indica ed. (C) Dis ibu ion o log 10 -co e age o D. shibae genomic DNA by eplicon. The boxplo s p esen ed in B and Cshow he 25 h and 75 h qua iles as whiske s, he in e qua ile ange as box and he median as line. These da a ha e been ep oduced (supplemen a y ig. S8,Supplemen a y Ma e ial online). (D) Co e age o ex ach omosomal eplicons by sequencing eads om DsGTA-DNA (blue and ed, acco ding o ype o eplicon) and genomic DNA (g ey). Posi ion o a pa alogous egion o which eads mapped ambiguously is indica ed by an as e isk (supplemen a y ig. S3 and no e S2, Supplemen a y Ma e ial online). Gene T ans e Agen s o Dino oseobac e shibae GBE Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 365 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 co ela ion wi h he GC-con en ( ig. 3E). Gene exp ession did no ha e a la ge in luence on packaging equency, nei he iewed globally o bo h s ands ( ig. 3D) no when exp ession o genes by s and was in es iga ed (supplemen a y ig. S6, Supplemen a y Ma e ial online). Only a weak, bu s ill signi i- can , nega i e co ela ion was ound ( ig. 3D). Wi hin peaks 3, 4, and 6, local minima co esponded o highly exp essed genes ( ig. 3Aand supplemen a y able S5,Supplemen a y Ma e ial online). The i s imp ession o a egula spacing o he di e en packaging equency egions gained suppo om an au oco - ela ion analysis. Fo his me hod, adap ed om ime-se ies analysis, Pea son co ela ion is calcula ed om all pai s o co - e age alues ha ha e he same dis ance om each o he and his is done o inc easing dis ances. The esul ing plo , showing he co ela ion o de ed by inc easing dis ance, indica ed a sig- ni ican pe iodici y o he DsGTA co e age da a wi h maximal co ela ion a dis ances sligh ly less han one qua e , one hi d, and hal he ch omosome size ( ig. 3F). Discussion He e, we demons a ed ha knockou o he au oinduce syn hase-encoding gene luxI 2 leads o o e p oduc ion o GTA pa icles by D. shibae. Thus, he p oduc o LuxI 2 likely ac s as a ep esso o GTA gene exp ession. T ansc ip ion o bo h he GTA gene clus e as well as he ope on coding o LuxI 2 and he AHL-binding ansc ip ion ac o LuxR 2 is in- duced by he same egula o y sys em, he C A phospho elay (Wang e al. 2014). Coac i a ion o he GTA gene clus e and i s ep esso migh o e an explana ion o bis abili y o GTA gene exp ession wi h only a subse o cells p oducing GTA pa icles (Fogg e al. 2012;Hynes e al. 2016;Queba e e al. FIG.3.—Packaging o ch omosomal DNA in o GTAs in compa ison o ch omosomal posi ion, GC-con en , and ansc ip ion. (A) Co e age o he ch omosome by DsGTA (black) and genomic DNA (g ey). O e and unde ep esen ed egions a e highligh ed in ed and blue, espec i ely. Minimum o cumula i e GC/AT-skew (yellow) indica es he e minus o eplica ion ( e ), which is iden ical wi h he posi ion o he gene cckA. The o igin o eplica ion (o iC) has been iden i ied be ween he genes dnaA and pa A (supplemen a y ig. S5,Supplemen a y Ma e ial online). These da a ha e been ep oduced (supple- men a y ig.S9,Supplemen a yMa e ialonline). (B) GC-con en o he ch omosome. (C) Numbe o me hyla ed Adenosines (m6A) wi hin he GANTC mo i . (D) Log 2 co e age o he ch omosome by RNA-seq eads o bo h s ands. Da a in A o Dha e been calcula ed o sliding windows o 4 kb. (E) Sca e plo s o DsGTA co e age e sus GC-con en , m6A me hyla ion and log 2 RNA-seq co e age. Spea man’s qis indica ed (P<0.001). (F) Au oco ela ion unc ion showing Pea son co ela ion calcula ed om DsGTA co e age o all pai s o sequences wi h he same dis ance om each o he so ed by inc easing dis ance. One qua e , one hi d, and hal he ch omosomal leng h away om o iC a e indic ed by ed lines. The dashed blue line indica es he 95% con idence in e al. Tomasch e al. GBE 366 Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018 2017) al hough we do no ha e di ec e idence om single cell expe imen s wi h D. shibae.Geneexp essionisas ochas- ic o noisy p ocess. Bis abili y can a ise om noise in he ex- p ession o ansc ip ional egula o s embedded in posi i e o nega i e eedback loops (Dubnau and Losick 2006). I is emp ing o specula e ha noisy exp ession o luxI/R 2 induces bi u ca ion o he popula ion in o cells exp essing GTA genes and hose ha do no . Howe e , u he expe imen s will be needed in o de o unde s and egula ion o GTA p oduc ion in D. shibae. Packaging o D. shibae genomic DNA in o GTA pa icles is s ikingly di e en om ha epo ed o he o he o ganisms o which whole genome s udies exis (Be glund e al. 2009; Hynes e al. 2012). In hese cases mic oa ays had been used o de e mine packaging equency o genomic DNA, a less sensi i e me hod compa ed wi h DNA sequencing due o he limi ed dynamic ange o he ob ained luo escence signal. Howe e , un-o eplica ion and he esul ing highe packag- ing equency o he ampli ied egion could be eliably de ec ed by his me hod (Be glund e al. 2009). Fu he mo e, ansposon-sequencing suppo ed he mic oa - ay da a o BaGTA (Queba e e al. 2017). Thus, i is unlikely ha he di e ences in packaging equency be ween his and he p e ious s udies ep esen an a i ac o he di e en me hods used. The ques ion o exac ly how he packaging o bac e ial DNA in o GTAs is pe o med has no been add essed ye . Howe e , he a ailable da a would sugges a “head ul” ype o packaging, which is common in double-s anded DNA phages (Casjens 2011;Oli ei a e al. 2013;Black 2015). In his packaging sys em, he o en highly conca e- me ic and b anched phage DNA is channeled in o he emp y capsid by an enzyme complex loca ed a he capsid en y (po al) and i is condensed 10,000- old in he p ocess (Be ndsen e al. 2015). This biomo o , which is he as es and mos powe ul molecula mo o known o da e (Smi h 2011), consis s o h ee essen ial componen s, each o hem o ming oligome ic ing s uc u es: he po al p o ein, and he la ge and small subuni s o he e minase complex, Te L and Te S, espec i ely (Casjens 2011). Packaging is ini ia ed by Te S binding o ecogni ion si es wi hin he conca eme ic phage DNA. A ecen s udy sugges s ha he DNA is w apped a ound he oligome ic Te S ing (Gao e al. 2016)andTe S- bound DNA is hen ecognized and ini ially cu by Te L a a de ined si e. As he nuclease domain o Te L cu s DNA unspe- ci ically, he si e o DNA linea iza ion is de e mined by he s uc u e o he Te S-DNA nucleop o ein complex (Djacem e al. 2017). The ATP-dependen mo o domain o Te L hen anspo s he linea DNA h ough he po al ing in o he capsid. When he capsid is ull, ano he cu eleases linea DNA ha can be packaged in o he nex capsid (Black 2015). Te S p esen a ion o DNA owa ds Te L is indispensable o p oduc ion and packaging o linea DNA in i o. Howe e , PhageT4DNApackagingisup o100%e icien in i owi h h ee componen s only: small linea DNA o any sequence, he la ge e minase Te L, and pu i ied p ohead wi h po al p o eins (Black and Rao 2012). Thus, i linea DNA is p esen in he cell, packaging could be possible e en wi hou ecog- ni ion si es o Te S. How hen can andom packaging o ch omosomal DNA in o GTAs be accomplished? Genes coding o e minase and po al p o eins a e p esen in alphap o eobac e ial GTA gene clus e s (Lang and Bea y 2007;Tama i e al. 2017). On he o he hand, we we e no able o iden i y sequence mo i s ha could ac as ecogni ion si es o Te S wi hin he peaks o high packaging. Two hypo heses migh help o explain he ob- se ed packaging equency. Fi s , he D. shibae genome migh encode ecogni ion si es o Te S a he co e age max- ima ha a e so weakly conse ed ha we canno de ec hem, o he physical p ope ies o he DNA, o example, how easy i is o w ap he DNA a ound Te S, a e mo e im- po an han he sequence i sel . Second, pe haps he e a e mo e ecogni ion si es h oughou he genome, making i di icul o iden i y hose en iched in he packaging peaks by disc imina i e analysis. O ien a ion, h ee-dimensional s uc u e, DNA modi ica ion and he o ma ion o nucleop o- ein complexes o he D. shibae ch omosome and plasmids migh hen de e mine i a egion can be accessed o pack- aging. Many Alphap o eobac e ia, including D. shibae (Pa zel e al. 2013), u ilize a pola mode o cell di ision. In he bes - s udied model o ganism Caulobac e c escen us, he o igin o eplica ion is loca ed a one cell pole, and he a ms o he ch omosome ex end he leng h o he cell owa ds he e mi- nus o eplica ion which is bound o he opposi e cell pole (Bowman e al. 2008). Physical a achmen o he cell poles migh s e ically hinde DNA o be bound by Te S, explaining he i ually absen packaging o DNA om he ch omosomal o igin and e minus o eplica ion. Using ch omosome con- o ma ion cap u e and deep sequencing, he h ee- dimensional s uc u e o he C. c escen us ch omosome was mapped (Le e al. 2013). I showed mul iple spa ial domains which we e s able h oughou he cell cycle; he ch omosome con ained back-bone egions as well as so-called plec onemes (supe -coiled s uc u es). A simila ch oma in-like s uc u e was also ound in Bacillus sub ilis and Esche ichia coli (Dame and Ta k-Dame 2016). The na u e o ch omosomal o ganiza- ion in D. shibae is cu en ly unknown, bu i i is compa able o ha o hese well-s udied model o ganisms, such o gani- za ion in o domains ha a e dependen on localiza ion and in e ac ion equency migh esul in some egions in which he DNA is mo e exposed and he e o e mo e accessible o packaging in o GTAs. This migh explain he egula dis ibu- ion o packaging peaks h oughou he ch omosome. Rega dless o he de ails o DNA ecogni ion by Te S, once DNA is bound by he p o ein i would hen be cu by Te L, esul ing in linea DNA ha can be u he packaged in o a GTA head. As successi e packaging migh andomly abo a e his ini ial linea iza ion, he GTA co e age would Gene T ans e Agen s o Dino oseobac e shibae GBE Genome Biol. E ol. 10(1):359–369 doi:10.1093/gbe/e y005 Ad ance Access publica ion Janua y 9, 2018 367 Downloaded om h ps://academic.oup.com/gbe/a icle-abs ac /10/1/359/4794729 by Ges Bio echnologische use on 07 Feb ua y 2018