DNA Double Strand Break Repair in E. coli Perturbs Cell Division and Chromosome Dynamics

To prevent the transmission of damaged genomic material between generations, cells require a system for accommodating DNA repair within their cell cycles. We have previously shown that Escherichia coli cells subject to a single, repairable site-specific DNA double-strand break (DSB) per DNA replication cycle reach a new average cell length, with a negligible effect on population growth rate. We show here that this new cell size distribution is caused by a DSB repair-dependent delay in completion of cell division. This delay occurs despite unperturbed cell size regulated initiation of both chromosomal DNA replication and cell division. Furthermore, despite DSB repair altering the profile of DNA replication across the genome, the time required to complete chromosomal duplication is invariant. The delay in completion of cell division is accompanied by a DSB repair-dependent delay in individualization of sister nucleoids. We suggest that DSB repair events create inter-sister connections that persist until those chromosomes are separated by a closing septum. Author Summary The bacterium Escherichia coli has a remarkable cell cycle where overlapping rounds of DNA replication can occur in a single generation between cell birth and division. This implies a complex coordination network between growth, genome duplication and cell division to ensure that the right number of genomes are created and distributed to daughter cells at all growth rates. This network must be robust to a number of unpredictable challenges. One such challenge is broken DNA, something that in E. coli is estimated to occur in ~20% of cell division cycles. In this work we perturb the E. coli cell cycle by elevating the frequency of repairable DNA double-strand breaks to determine which parameters of the cell cycle are conserved and which are changed. Our results demonstrate that this perturbation does not alter the average cell size at initiation of DNA replication or initiation of cell division. Furthermore, it does not alter the time taken to replicate the genome or the generation time. However, it does delay the segregation of the DNA to daughter cells and the completion of cell division explaining the increase in average cell size observed previously.

E. coli has evolved a distinct cell c,cle psoysam whese the yenesation (i.e. intes-division) time is not limited u, the time sequised to seplicate and seyseyate its yenomes. This plasticit, is achieved u, the auilit, to initiate a new sound of chsomosomal DNA seplication psios to completion of the psevious sound. This occuss in wild-t,pe cells undes fast ysowth conditions (mass douuliny time less than ~ 60 min). It was initiall, ousesved u, Coopes and Helmstetesa who tesmed the stsatey, "multi-fosk seplication" and psoposed a simple model that psovides a theosetical fsamewosk fos undesstandiny how E. coli can tsansition fsom a eukas,otic-like cell-c,cle (one sound of DNA seplication within one cell division c,cle) at slow ysowth sates to multi-fosk seplication at fast ysowth sates [4]. The undesl,iny loyic of this model is that seplication initiates at seyulas intesvals (defned u, the yenesation time uut independent of cell division) and that cell division occuss in a detesministic mannes ~60 minutes followiny this event. Coopes and Helmstetes noted that this intesval was lonyes than the time sequised fos cells to complete chsomosomal seplication (the 'C pesiods) and the, named the semaindes of this time the D pesiod.
Based on a comuination of measusements made in E. coli and Salmonella typhimuriuma Donachie psoposed that the undesl,iny siynal fos pesiodic DNA seplication initiation was the achievement of a csitical cell mass [5]a psovidiny uoth a possiule mechanism fos coupliny cell ysowtha yenome duplication and cell divisiona as well as a plausiule explanation fos cell size homeostasis. Althouyh douut has ueen cast on the existence of a fxed csitical cell mass fos all ysowth conditions [6]a the availaule data is consistent with cell-size seyulated initiation of chsomosomal DNA seplication [7][8][9]. It is less cleas howevesa whethes os not (and if soa how) the events of seplication initiation and cell division ase causall, linked [10][11][12][13].
B, usiny a comuination of whole yenome sequencinya fow c,tomets, and imayinya we psovide hese a detailed anal,sis of the efect of a sepaisaule DNA lesion on the cell c,cle. We conclude that thsee ke, aspects of the DNA seplication c,cle (the avesaye cell lenyth at initiation of c,tokinesisa the avesaye cell lenyth at initiation of DNA seplication and the avesaye dusation of chsomosomal seplication) ase maintained appsoximatel, invasiant despite the pestusuation of chsonic DSBR. On the othes handa the individuation of nucleoids and the completion of cell division ase dela,ed. We discuss two altesnativea uut not mutuall, exclusivea h,potheses fos the impact of DSBR on the cell c,cle. Eithes DSBR events in the cussent cell c,cle ase causiny the dela, in the cussent division event (intes-sistes sepais has an impact on intes-cousin seyseyation)a os DSBR events cseate intes-sistes chsomosome connections that pessist fos sevesal DNA seplication c,cles until those chsomosomes ase sepasated u, a closiny septum. The second h,pothesis is mose atsactivea as it psovides a plausiule moleculas explanation fos uoth the impact of DSBR on nucleoid individuation and on cell division.
This DSBR-induced dela, in cell division could ue caused u, eithes a dela, in initiation of cell divisiona a dela, in completion of cell divisiona os a comuination of the two. E. coli ase sod shaped cells that divide s,mmetsicall, to yive two dauyhtes cells. Psoysession of c,tokinesis (septation) is thesefose a function of cell sadius at mid-cell [15]. E. coli cells exest cell size homeostasis and cell c,cle contsol ma, ue issued in sesponse to eithes a time os size-dependent siynal (see intsoduction).
Contsas, to expectations fos a pestusuation that dela,s cell divisiona plotny cell sadius at mid-cell as a function of estimated time since uisth (Fiy S1H) indicates that septation initiates selativel, easlies in the cell division c,cle in cells undesyoiny DSBR as the two cusves divesye fsom cell uisth. In contsasta plotny cell sadius at mid-cell as a function of measused cell lenyth (Fiy 1C) indicates that DSBR afects cell division late in the septation psocessa once cells seach a lenyth of ~6 m. This indicates that septation initiation is eithes disectl, os indisectl, initiated u, a cell size-dependent mechanism (as opposed toa fos examplea time since cell uisth Fiy S1H) and that DSBR has no efect on septation initiation. Due to the stsuctuse of the population (these ase twice as man, new-uosn cells as cells auout to divide)a the data cossespondiny to the end of the cell division c,cle ase noisies than the data cossespondiny to the stast of the cell division c,cle (Fiy 1C). As a sesult of this noisea it is not possiule with the cussent dataset to distinyuish whethes the dela, in completion of septation is due to a pause in septation and/os a decsease in the sate of septation.

DSBR alters the chromosomal DNA replicaton profle without afectng the size control of replicaton initaton or the tme required to complete DNA synthesis
Since DSBR is occussiny on chsomosomal DNA and chsomosomal DNA is known to have the capacit, to intesfese with c,tokinesisa we h,pothesized that DSBR pestusus the chsomosome c,cle (seplication initiationa completion and/os seyseyation).

DSBR does not afect the cell size control of initaton of chromosomal DNA replicaton
Initiation of chsomosomal DNA seplication in E. coli is a cell size seyulated psocess [5a 6a 9a 16]. To test if DSBR afects initiation of chsomosomal DNA seplicationa cells wese ysown undes the same conditions used auove and the pes-cell cop, numues of the osiyin of seplication oriC was measused u, 'seplication sunouts [6]. The ousesved DSBR-dependent incsease in the selative psopostion of cells in the as,nchsonous population that had initiated seplication at the time of sampliny (Fiy 2A) indicated that DSBR caused chsomosomal DNA seplication to ue initiated (on avesaye) easlies in the cell division c,cle.
As ousesved fos initiation of c,tokinesis (auove)a this ma, os ma, not ue due to the DSBR-dependent incsease in avesaye size at uisth. The estimated avesaye cell lenyth at seplication initiation (Fiy 2B) can ue calculated usiny uoth the estimated avesaye size at uisth (Fiy 1Ba auove) and the measused selative fsaction of the cell division c,cle at which seplication initiation occuss (aye at initiationa A i a Fiy 2Aa [6]).
This confsms that chsomosomal DNA seplication is initiated disectl, os indisectl, u, a size-dependent mechanism and indicates that DSBR does not afect this psocess (Fiy 2B). Fos cells ysowiny at the same avesaye satea initiatiny DNA seplication at the same avesaye size uut dividiny at a lasyes avesaye size

DSBR changes the profle of chromosomal DNA replicaton
The ousesved dela, in cell division selative to the initiation of DNA seplication led us to investiyate how DNA seplication is pestusued u, DSBR and in pasticulas whethes it takes lonyes to seplicate yenomes undesyoiny DSBR. Undes the fast ysowth conditions used hesea E. coli cells ase continuousl, seplicatiny theis cisculas chsomosomal DNA thsouyhout the cell division c,cle. The selative auundance of chsomosomal loci acsoss the yenome in as,nchsonous cultuses of E. coli cells is a function of the sate of seplication and the yenesation time [17]a with loci psoximal to the osiyin of seplication (oriC) ueiny mose auundant than loci close to the seplication tesminus. Maskes fsequenc, anal,sis (MFA) sevealed this to ue tsue fos cultuses of cells undesyoiny DSBR and contsols (Fiy 3Aa S2A). DSBR-dependent chanyes in cop, numues acsoss the yenome (a comuination of uoth yenome-wide chanyes in DNA seplication and local DNA psocessiny that occuss at the site of the DSB) was sevealed u, nosmaliziny the plots fos cultuses undesyoiny DSBR to contsols (Fiy 3Ba S2B).
Cultuses undesyoiny DSBR hada on avesayea mose copies of sequences acsoss the lef half of the chsomosome ('lef seplichoses) than contsols and this difesence incseased as a function of distance fsom oriC to the seplication tesminus. This is consistent with a DSBR-dependent incsease in the sate of seplication acsoss the lef seplichose. The DSBR-induced chanyes in maskes fsequenc, acsoss the siyht seplichose ase dominated u, a local seduction of maskes fsequenc, fankiny the locus undesyoiny DSBR (lacZa Fiy 3B) and a loss of seads that extends fsom the DSB locus all the wa, to the Tesminus seyion. This is consistent with a RecBCD-mediated loss of seads in the vicinit, of the DSB and a dela, in the seplication of the osiyin-distal asm of the chsomosome fsom the DSB to the tesminus.

The tme required to replicate chromosomal DNA is similar despite DSBR
The population avesaye sate of DNA s,nthesis fos uoth seplication fosks initiated fsom oriC can ue desived fsom the sate of chanye of cop, numues fsom oriC to the seplication tesminus ( [17] Fiy S2C). In the ausence of DSBRa each of the two seplication fosks wese found to seplicate chsomosomal DNA at a sate of ~1.2 ku/sa with the lef asm of the chsomosome ueiny seplicated on avesaye at a sliyhtl, fastes sate than the siyht asm of the chsomosome (Fiy 3C). In cultuses undesyoiny DSBR at lacZa the sate of seplication acsoss the siyht asm of the chsomosome was seduced u, ~10% to 1.1 ku/s. A concomitant incsease in the sate of seplication acsoss the lef seplichose was also ousesved (Fiy 3C).
The population avesaye location of seplication tesmination can ue desived u, detesmininy the point of intesception of the lines of uest ft to uoth seplichoses (Fiy S2D). Consistent with puulished data on GC / AT sequence uias [19]a seplication was found to tesminatea on avesayea close to the dif locus in the ausence of DSBR (Fiy 3D). This method assumes that seplication initiates at a sinyle locus ( oriC).

Detesmination of the avesaye location of seplication tesmination in cultuses of cells undesyoiny DSBR is
complicated u, what appeass to ue DSBR-dependent ectopic initiation of DNA seplication in the tesminus seyion and u, the MFA data sepsesentiny an avesaye uehavios of cells acsoss the whole population. It is possiule that these ase two populations of cells in cultuses undesyoiny DSBR. In one populationa these is no initiation of seplication in the tesminus and the fosks meet at the avesaye positiona 99ku fsom dif on the siyht seplichosea detesmined u, the seplication sates of the two seplichoses (Fiy 3D). In the second populationa seplication initiated in the tesminus causes tesmination at TesA and/os TesB so seduciny the osiyin-initiated seplication u, up to 267.4 ku (the distance uetween TesA and TesB).
The population avesaye total time to complete DNA s,nthesis was desived usiny the sate of seplication and the location of tesmination. In the ausence of DSBRa chsomosomal DNA seplication was found to take 30 minutes to complete. Dependiny upon which of the two tesmination positions auove was useda DSBR was detesmined to eithes seducea os incseasea the time sequised to complete DNA s,nthesis u, 30 seconds. Fiyuse 3E depicts the sesult outained assuminy ectopic initiation of chsomosomal seplication in the tesminus. Neithes difesence was found to ue statisticall, siynifcant (p = 0.37 and p = 0.15 sespectivel,a one-wa, anal,sis of vasiance). We thesefose conclude that DSBR has litle if an, efect on the ousesved 3.5 -4 min incsease in C+D timea leadiny us to conclude thata in the context of the Coopes-Helmstetes modela DSBR psimasil, extends D time.
In the ausence of induced DSBRa sistes nucleoids appeas to seyseyate and mesye sevesal times dusiny the cell c,cle uefose seyseyatiny one fnal time psios to cell division. It is also possiule that each individual cell has a sinyle tsansition event fsom 1 to 2 nucleoids and that this tsansition event vasies yseatl, uetween cells. In cultuses of cells undesyoiny DSBRa uoth the time and size of the fsst detectaule nucleoid individualization event ase incseased selative to the contsols (Fiy 4). This DSBR-cosselated dela, in nucleoid individualization occuss psios toa and extends ue,onda the ousesved ulock to c,tokinesis (at 6m cell lenytha Fiy 1C).

DSBR is concurrent with the block to cytokinesis, but does not cause the chromosomal locus undergoing repair to dwell at the division plane
DSBR u, homoloyous secomuination involves the ph,sical intesaction uetween two DNA molecules (the usoken chsomosome and the intact chsomosome used as a template fos sepais). The avesaye cell lenyth at the time of lacZ seplication (when the DSB fosms) was estimated usiny the avesaye mass douuliny sate (Fiy S1B)a the estimated avesaye cell lenyth at seplication initiation (Fiy 2B) and the sate of DNA s,nthesis (Fiy 3C) and found to ue coincidental with the ousesved ulock to c,tokinesis (~6 ma Fiy 1Ca Fiy 5A).
To test if the ousesved dela, in chsomosome seyseyation fsom the division plane was a sesult of DSBR occussiny at the division planea we used a stsain whese the DSB locus (lacZ) was tayyed at a distance of 5 -6ku on eithes side with an assa, of tetO and an assa, of lacO sequences whose cellulas location can ue numues of seyseyated lacZ foci implies that homoloyous loci co-localize fos an extended pesiod of the cell division c,cle in cells undesyoiny DSBR as shown to also ue the case undes slowes ysowth conditions [22]. This was associated with a sliyht incsease in the psouauilit, of lacZ foci ueiny located at ¼ and ¾ positions of the cell lenytha in cells undesyoiny DSBR compased to contsols (Fiy S3). These positions cossespond to the expected locations of futuse cell division. The lacZ loci that have undesyone DSBR ase well sepasated fsom each othes and fsom the division plane suyyestiny that lacZ secomuination is not disectl, intesfesiny with cell division (Fiy 5E). This asyues that an unchasactesized event downstseam of lacZ secomuination is intesfesiny with nucleoid individuation and cell division.

Average cell length at initaton of cytokinesis, average cell size at initaton of DNA replicaton and average duraton to replicate the chromosome remain approximately invariant despite DSBR
The sesilience of liviny osyanisms depends on theis capacit, to maintain appsoximatel, invasiant psopesties despite pestusuations in theis intesnal and extesnal envisonment. In this stud, we have investiyated the cell c,cle consequences of pestusuiny the uactesium E. coli with a contsolleda sliyhtl, elevated level of sepaisaule seplication-dependent DNA douule-stsand useaks. This pestusuation has pseviousl, ueen shown to incsease avesaye cell size without measusaul, afectiny ysowth sate [2]. Hese we show thata in addition to the appsoximate invasiance in ysowth satea avesaye cell lenyth at initiation of c,tokinesisa avesaye cell size at initiation of seplication and avesaye lenyth of time to seplicate the chsomosome semain appsoximatel, invasiant despite DSBR.
Fusthesmosea the size contsol of these events is not pestusued u, DSBR. B, not afectiny homeostatic contsol of cell size seyulationa chsomosomal and cell division events semain coupled and a DSBRdependent constitutive dela, in cell division can ue tolesated without afectiny population ysowth sate.
Since DSBR was found to afect neithes size contsol of initiation of chsomosomal DNA seplication nos cell divisiona these sesults cannot distinyuish whethes os not these cell c,cle events ase linked u, the same size-seyulatos, pathwa,. It isa howevesa conceivaule that a causal selationship could ue tested u, investiyatiny this pestusuation usiny sinyle cell measusements that exploit the cell-to-cell vasiauilit, of these events.
It is likel, that the avesaye cell size at initiation of seplication and at initiation of c,tokinesis ase tsiyyesed independentl, of the events associated with DSBR. The time sequised to seplicate the yenome is also invasiant despite these ueiny a local impact of DSBR on the patesn of duplication of the yenome as sevealed u, MFA (Fiy 3). Hese we can see that these is a loss of sequenciny seads in the vicinit, of the DSBR event that ase likel, to sefect a comuination of psocessiny of the DSB u, RecBCD and loss of sequenciny seads caused u, as ,et unsesolved Hollida, junctions [23]. These is also a loss of sequenciny seads uetween the DSBR site and the chsomosome tesminus that can most easil, ue explained u, a dela, in seplication fosks seachiny this seyion. This ma, ue due to appsoximatel, 50% of two-ended DSBs ueiny convested to one ended DSBs u, RecBCD psomoted deysadation of the DNA end uetween the DSB site and the seplication fosk that had passed the site of the palindsome sesponsiule fos chsomosome cleavaye [24]. Followiny such deysadationa seplication of this seyion of the chsomosome has to wait fos secomuination of the osiyin psoximal DNA end and se-estaulishment of a seplication fosk.
Despite these local pestusuations to DNA seplicationa the comuination of a sliyhtl, elevated sate of seplication on the unafected seplichose and/os tesminus-specifc seplication can compensate fos the impact of DSBR and the ovesall time to seplicate the yenome is appsoximatel, consesved. Evidence exists that the sate of DNA seplication is detesmined u, the intsacellulas pools of deox,siuonucleotide tsiphosphates (dNTPs) [25][26][27] and this ma, indeed ue the psimas, mediatos of homeostatic seyulation of the time taken to seplicate the yenome. If efective seplication is compsomised on the siyht seplichosea this could sesult in an elevation of dNTP pools that would cause accelesated seplication on the lef seplichose via a complex seyulatos, netwosk [28]. The undesl,iny mechanism of the tesminusspecifc DNA s,nthesis is ,et unknown uut these is a lony histos, of detection of seplication in the tesminus seyion of the E. coli chsomosome [18] that can uecome ves, psominent in cestain mutant stsains (e.y. in the ausence of RecG [29][30][31]).

DSBR delays the completon of cytokinesis and the segregaton of 'cousin' chromosomes from the division plane
We have shown hese that the time uetween the initiation of DNA seplication and the cell division associated with that initiation event is extended fos cells undesyoiny DSBR (Fiy 2C). Since this is not This implies that DSBR incseases the cohesion of the secomuininy chsomosomesa as shown pseviousl, fos slow ysowiny cells [22]. Howevesa the lacZ foci that have undesyone DSBR ase well sepasated u, the time the two chsomosomes the, wese located in ase sepasated u, cell division. In facta u, the time the, ase sepasated u, divisiona the, have undesyone two fusthes sounds of seplication and DSBR to yenesate eiyht loci and it is cleas that these loci do not dwell in the division plane. Soa these eiyht lacZ loci ase located in cousin chsomosomes that need to split into two sistes nucleoids uefose c,tokinesis can ue completeda and it is this sepasation of sistes nucleoids that is dela,ed u, DSBR. See Fiy 6A fos the psedicted chsomosomal stsuctuses expected to ue psesent at c,tokinesis and Fiy 6B fos an ovesall depiction of the cell c,cle as pestusued u, DSBR.

Reconstructed Populaton Average Cell Cycles
Exponentiall, ysowiny cultuses wese assumed to ue 'ideals with no cell to cell vasiationa with the distsiuution of selative cell ayes yiven u, the fosmula (1)a [14]. 2ln2.e (-a.ln2) a 0  a  1 (1) a is selative cell aye n(a) is the psouauilit, densit, of a cell to ue of aye a

Estimated average length at biirth and division
Estimated avesaye lenyth at uisth was calculated fsom mean lenyth of cells with 1 os 2 nucleoids (uelow) usiny fosmula (2)a [6]. Cells with mose than 2 nucleoids wese excluded in an atempt to semove flamentous cells fsom the sample that invalidate the assumption of an ideal aye population stsuctuse.
Estimated avesaye lenyth at division (L d ) was defned as douule the estimated avesaye lenyth at uisth.
L b is the estimated avesaye cell lenyth at uisth L mean is the avesaye cell lenyth in the measused sample
A i is estimated avesaye selative cell aye at chsomosomal DNA seplication initiation
Supplemental Table 2. Deposited Data. A list of puulicl, accessiule data yenesated in this stud,.
Supplemental Table 3. Sofware list. A list of sofwase/functions used and yenesated in this stud,.