Engineering a Sinorhizobium meliloti chassis with monopartite, single replicon genome configuration

Multipartite bacterial genomes pose challenges for genome engineering and establishment of additional replicons. We simplified the tripartite genome structure (3.65 Mbp chromosome, 1.35 Mbp megaplasmid pSymA, 1.68 Mbp chromid pSymB) of Sinorhizobium meliloti. Strains with bi- and monopartite genome configurations were generated by targeted replicon fusions. Our design preserved key genomic features, such as replichore ratios, GC skew, and KOPS and coding sequence distribution. Under standard culture conditions, growth rates of these strains and the wild type were nearly comparable. Spatiotemporal replicon organization and segregation were maintained in the triple replicon fusion strain. Deletion of the replication initiator-encoding genes including the oriVs of pSymA and pSymB from this strain resulted in a monopartite genome with oriC as the sole origin of replication, a strongly unbalanced replichore ratio, slow growth and an aberrant cellular localization of oriC. Suppressor mutation R436H in the cell cycle histidine kinase CckA and a 3.2 Mbp inversion, both individually, largely restored growth. These strains will facilitate integration of secondary replicons in S. meliloti, and thus be useful for genome engineering applications, such as generating hybrid genomes. Graphical Abstract


◼ INTRODUCTION
It is estimated that about 10 % of the bacterial species represented in genome databases carry their genomic information divided on several replicons, i.e. have multipartite genomes. (1)Their genome architectures are diverse in terms of number and size of DNA molecules.In addition to the main chromosome, these bacteria harbor one or more large secondary replicons classified as secondary chromosomes, chromids or megaplasmids according to their origin, presence of core essential genes, and size. (1,2)The presumed advantages of multipartite genomes include promotion of horizontal gene transfer, genome plasticity, and exploitation of new ecological niches. (2,3)(6)(7)(8)(9) All these fields of research and application have in common that new or modified genetic material has to be established in the bacterial cell, often using extrachromosomal replicons (ECRs).However, this genome engineering is complicated by a multipartite genome configuration.Multiple replicons can, for instance, restrict the establishment of further ECRs through incompatibility, or lead to genomic reorganization and instability through recombination events.The advantages of a streamlined genomic architecture with fewer replicons became evident from comparing two natural M. exorquens isolates, leading to strain AM1 (chromosome, 4 ECRs) increasingly being replaced by strain PA1 (chromosome, no ECRs) in research and application. (10)ndamentally different approaches to the rational streamlining of multipartite genomes are the reduction of the number of replicons through replicon curing or the fusion of replicons, with both approaches also being combinable.The former approach is associated with the loss of genetic information, while the latter approach can retain the entire genetic information.
S. meliloti, a soil-dwelling diazotrophic representative of α-proteobacteria with multipartite genomes, exemplifies both approaches, either through targeted genetic interventions or genetic events in natural isolates.The S. meliloti type strain Rm1021 carries a main chromosome of 3.65 Mbp, megaplasmid pSymA of 1.35 Mbp, and chromid pSymB of 1.68 Mbp. (11)Vertical transmission of the secondary replicons is mediated by replicon-specific repABC loci (12) that provide a complete replication and segregation system. (13)The repABC operon codes for the DNA replication initiator protein RepC and the ParAB-type partitioning proteins RepA and RepB.Together with centromer-like sequences (parS) these partitioning proteins are crucial for segregation of the replicated DNA. (14)The replication origin of repABC-type secondary replicons (oriV) is located within an AT rich region of repC. (15)In contrast, the chromosomal origin of replication (oriC) is characterized by DnaA-boxes and DNA unwinding elements (DUE). (16)trains cured for pSymA, pSymB or both secondary replicons were generated, with the latter resulting in a genome reduction of 45 %. (17) However, the lack of pSymB is associated with considerable fitness loss. (17)In addition, the loss of secondary replicon-harbored genes can strongly limit the application range of replicon-cured strains, as these genes are e.g.important for symbiotic nitrogen fixation as well as for fitness and a broad substrate spectrum in the natural habitat. (18,19)Yet, such strains have already been used as a platform for the construction of hybrid genomes by complementation with secondary replicons from closely related strains (in-cis hybrids). (20)3)(24)(25)(26)(27) In bacterial cells, the genomic DNA is highly organized, and replication and segregation are coordinated in space and time. (28)Tethering of the chromosomal origin and terminus regions to cell poles plays an important role for spatial chromosome organization and segregation. (29)Previously, a longitudinal oriC (old pole)-terC (new pole) configuration of the main chromosome and subpolar localization (old pole) of the oris of the secondary replicons was reported for S. meliloti G1-phase cells with tripartite genome configuration. (13,30)Additionally, it was shown that segregation of the replicons follows a predominant order, with segregation of the ori of the main chromosome preceding those of the secondary replicons. (13)is raises the question whether replicon fusions impair the spatial organization and spatiotemporal segregation of genomic DNA.
In this study, we constructed S. meliloti strains with a bi-and monopartite genome configuration, either with the genomic DNA molecules including one or more origins of replication.We show that a cointegrate of the three genomic replicons that per design preserved typical features of bacterial genome organization, maintained growth properties and key features of spatial organization and segregation dynamics of each replicon under standard laboratory culture conditions.Furthermore, we show that an engineered strain with monopartite genome configuration, harboring oriC as sole origin of replication and characterized by a strongly unbalanced replichore ratio is viable, but impaired in growth.Suppressor mutants of this strain with a missense mutation in the cell cycle histidine kinase CckA or a partial genomic inversion showed significantly restored growth, making them chassis candidates for facilitated integration of ECRs and for future construction of hybrid genomes.

Replicon fusions preserving strand asymmetry patterns
To study the properties of a triple-replication origin chromosome in S. meliloti in terms of spatial organization and spatiotemporal dynamics as well as propagation stability during the cell cycle, we merged the tripartite genome of this α-proteobacterium into a single DNA molecule.Targeted fusions of the three replicons were achieved by Cre/lox-mediated recombination in S. meliloti SmCreΔhsdR. (31)This strain, further referred to as wild type in this study, lacks hsdR encoding a restriction endonuclease and carries chromosomally integrated cre encoding the Cre recombinase.This recombinase catalyzes site-specific recombination of DNA between loxP sites.Replicon cointegrations sporadically occurring in S. meliloti by recombination between nodPQ1 and nodPQ2 as well as between the algI paralogues SMb20843 and SMc01551 (21) acted as blueprint for selection of the fusion sites, as these sites were compatible with a nearly symmetric structure of the fusion construct with respect to the distribution of three origins of replication.
Following integration of lox sites, in close proximity to nodPQ1 and nodPQ2, megaplasmid pSymA and chromid pSymB were initially merged to generate pSymAB by induced Cre/lox-mediated site-specific recombination.Thus, the resulting strain SmAB harbored a bipartite genome composed of the main chromosome and pSymAB (Fig 1A, S1 Fig).SmAB and SmABC genomic DNA configurations were designed to display wild type-like replichore ratios and distribution of GC content (Fig 1A, S2 Fig).(34) Each of the S. meliloti replichores contains a higher number of protein-coding genes on the leading than on the lagging strand (S3 Fig).This bias was preserved in SmAB and SmABC, even though two and three of the six replichores, respectively, were composed of segments originating from two or three different replicons of the wild type (Fig 1C).Also, the distribution of sequence motifs matching the E. coli KOPS consensus reflected the replicon structure and showed a strand asymmetry pattern with increasing motif density from ori towards ter in the replicons of wild type, SmAB and SmABC (Fig 1D, S4 Fig).

Cell growth and morphology were not affected by the genome rearrangements
Like other α-proteobacteria, S. meliloti undergoes asymmetric cell division which leads to siblings of uneven cell size. (13,35,36)Cell shape and size of the S. meliloti replicon fusion strains and the wild type were indistinguishable by microscopic analysis (Fig 2A).A few cells with strongly asymmetrically placed constriction sites, probably resulting in minicells after cell division, were observed in cultures of the S. meliloti wild type, SmAB and SmABC strains in TY (complex) and high salt TY media (S5A Fig) .A quantitative analysis identified a maximum of 0.3 % minicells in wild type and replicon fusion strain  When cultivated in TY or low phosphate minimal MOPS (defined) media, cell growth of both replicon fusion strains was similar to that of the wild type (Fig 2B) indicating that the new genome configuration does not significantly interfere with cell cycle processes.To analyze the effect of high-salinity and hyperosmotic stress on these strains, TY medium was supplemented with 0.4 M NaCl and 0.5 M sucrose, respectively.Compared to the wild type, SmAB and SmABC appeared slightly impaired in growth in these media (Fig 2B).Supplementing TY medium with 0.6 M NaCl or 0.7 M sucrose enhanced the difference in growth between wild type and replicon fusion strains (S6B Fig) .High salt conditions are known to influence the supercoiling state of genomic DNA, (37) thus, the triple replicon fusion strain may be less robust to conditions affecting DNA condensation.
The secondary replicon oriVs were dispensable in the strain with monopartite genome configuration We asked if a single replication origin would be sufficient for proper vertical transmission of the AB and ABC DNA molecules in SmAB or SmABC.To this end, the repB-repC intergenic region and the repC coding region (repC2 of pSymA and repC1 of pSymB) were targeted for deletion (S7A  ).Apart from these modifications, we identified only up to three single nucleotide variations (SNVs) per engineered genome compared to the wild type genome (S6 Table ).Considering the nature of these changes, we suppose they are unlikely to be relevant to bacterial strain survival or fitness.
The sequential inactivation of the secondary replicon oriVs was associated with a gradually increasing growth defect, which appeared to correlate with the number of oriVs per DNA molecule and are possibly due to replication obstacles arising from the new replichore structures.Deletion of a single repC copy in SmAB resulted in a more severe growth defect than in SmABC, whose monopartite genomic DNA still carried two oris (repC and oriC) after deletion of one of the repCs.However, deletion of both repC copies in SmABC caused the most severe growth defect (Fig 3A), even though cell morphology was inconspicuous in a microscopic snap shot image analysis (Fig 3B).
To test if replication is initiated at all three oris and to determine the replication termination regions of the triple replicon fusion molecule in SmABC a Marker Frequency Analysis (MFA) was performed.The MFA data indicate that all three oris mediated bidirectional replication (Fig 3C, S8B Fig).Furthermore, this data suggests that replication terminated in regions close to or overlapping the MFA minima for chromosome, pSymA and pSymB identified in the parental strain with tripartite genome configuration (S8A Fig), which roughly match with predicted ter regions. (38)MFA analysis of SmABCΔoriV shows clear sequence read enrichment with a maximum at oriC but no additional enrichment maxima at the mutated repABΔC (ΔoriA/ΔoriB) regions.The latter suggests loss of function of oriA and oriB in SmABCΔoriV (Fig 3C).This analysis also suggests replication termination in a region close to or overlapping terC, which separates two replichores largely differing in size with a ratio of approximately 3:1 (S8C Fig) .The conservation of terC in the single replicon strain SmABCΔoriV, with even extremely skewed replichores, provides evidence for mechanisms defining this region for termination in the replication process.
Differences in replication fork progression are maintained, e.g., by replication fork trapping sequences that slow down or stall the replisome. (39)In contrast, blocking of replications fork progression was not found for the terminus regions of either of the two chromosomes of V. cholerae. (40) (38)Note that SmABCΔoriV lacks oriA and oriB due to repC deletions (ΔoriA and ΔoriB).(D) True to scale scheme of the secondary replicon derived replication and partitioning loci in S. meliloti.Short lines perpendicular to the sequence line depict the position of palindromic sequences (parS) involved in the process of DNA partitioning.Black arrowheads point to the area of gene disruption.The lines with a capital delta in between represent the regions that were targeted for deletion.
Plateaus or decrease in locus frequencies, including ΔoriB, terB and terA regions, might indicate replication obstacles in the single-replicon monopartite genomic DNA.Nevertheless, replisomes that had initiated replication at oriC do not appear to have been ultimately stalled in the original pSymA and pSymB terminus regions.
In addition, fluorescence microscopy-based analysis of mCherry-fused DNA polymerase III beta subunit DnaN revealed a maximum of two fluorescent foci in SmABCΔoriV (only oriC present) in constrast to four to five foci observed in parallel over the cell cycle in SmABC (oriC and both oriVs present) (S9 Fig).This is indicative of a reduced number of replisome formations in SmABCΔoriV and further confirms inactivation of both oriVs in this strain.Collectively, MFA and this microscopy data indicate that SmABCΔoriV contains a single 6.7 Mbp replicon and that the initiator protein-encoding repC genes including the oriVs of the secondary replicons are dispensable in this strain.

The partitioning system of pSymB was dispensable in the replicon fusion strains
It is thought that the RepAB proteins are specific to partitioning of the cognate repABC-type replicons. (41) study if the gene products of the repA and repB genes play a role for partitioning of the AB and ABC fusion replicons, we aimed to delete or inactivate the repAB gene regions individually and together in SmAB and SmABC (Fig. 3D).pSymA contains the single rep locus repA2B2C2, whereas pSymB harbors the complete repA1B1C1 and the truncated repA3B3 locus.
We found striking differences regarding the necessity of these elements in wild type, SmAB and SmABC strains.Deletion of repA1B1C1 with adjacent partitioning sites and of repA3B3 was possible in SmAB and SmABC (Fig. 3D).In the wild type, deletion or inactivation was achieved for repA3B3 only (Fig. 3D).Also, we obtained double mutants by deletion of repA1B1C1 and inactivation of repA3B3 in SmAB and SmABC but not in the wild type.Thus, the pSymB partitioning system appears to be dispensable for S. meliloti harbouring AB or ABC replicon fusions, in contrast to the results reported for Agrobacterium tumefaciens. (23)Deletion of pSymA-associated repA2B2C2 with adjacent partitioning sites and of the partitioning sites alone was not successful in wild type and both replicon fusion strains (Fig. 3D).We also did not achieve inactivation of this rep locus by plasmid integrations in these strains.This suggests that the pSymA partitioning genes are required in S. meliloti, even if pSymA and pSymB or all three replicons are fused.An important role for the pSymA-derived partitioning system-encoding genes in the triple replicon fusion strain was unexpected since S. meliloti can be cured from pSymA. (17,42)triple color fluorescent labeling system for simultaneous microscopic monitoring of oriC

and two further freely selectable genomic loci
To enable live cell imaging analyses of spatial organization and spatiotemporal dynamics of the rearranged genomic DNA in the replicon fusion strains SmAB, SmABC and SmABCΔoriV, we established a triple color fluorescent labeling system.As this required using several antibiotic resistance markers, we eliminated by homologous recombination the inactive lox sites together with the downstream resistance markers that remained from the replicon fusion procedure in the AB and ABC DNA molecules (S1 Fig, S1A Data).This gave rise to derivatives of SmAB, SmABC and SmABCΔoriV termed SmABΔR, SmABCΔR and SmABCΔoriVΔR, respectively.The genome architecture of these strains was validated by PFGE (S1B Data) and Illumina genome sequencing.
ParB is known to specifically recognize and bind cognate parS sequences typically located around the chromosomal origin of replication in various bacterial species. (43)This protein was previously reported to be essential in S. meliloti (44) and used as marker for oriC in this bacterium. (13,45) In addition, a fluorescence repressor operator system (FROS) for labeling of two freely selectable genomic loci was established.This system combined tetO and lacO operator arrays (46) with tetR-mVenus and lacI-mCherry placed under control of the PtauA promoter on the replicative and mobilizable low copy vector pFROS (S10A Fig) .In this setup, the basal activity of PtauA (47) was sufficient to generate a well detectable fluorescent signal over analysis periods of at least up to 6 hours in our study ( S10D Fig), which makes the system particularly suitable for time lapse applications.

Polar localization of oriC and subpolar localization of both oriVs at the old cell pole, as well as positioning of terC at the new cell pole was conserved in SmABC
We asked for the effect of replicon fusions on the spatial organization of the genomic DNA in S. meliloti by comparing monopartite and tripartite genome architectures.To this end, SmABCΔR and wild type derivatives, both carrying the parB-cerulean marker, were equipped with tetO and lacO arrays for simultaneous labeling of oriC/oriA/oriB, oriC/terC, oriC/oriA/terA and oriC/oriB/terB (Fig 4A, S11A Fig).
In addition, we constructed derivatives of both strains for simultaneous labeling of oriC via ParB-cerulean and one of 12 further genomic loci either by integration of a tetO or lacO array.The labeled positions included genomic loci adjacent to the sites used for fusion of the replicons (Fig 4A, S11A Fig).Prior to fluorescence microscopy analysis, these strains were validated by PFGE for their correct genome configuration (S1C Data, S1D Data), and pFROS was introduced to mediate fluorescent labeling of the genomically integrated tetO and lacO arrays.For subcellular 2D mapping of the labeled genomic loci by snap-shot imaging of live cells, we filtered cells by size with a maximum cell length of 2.0 µm and displaying a single ParB-Cerulean-mediated fluorescent focus in one of the two cell pole regions only (Fig 4B).This configuration is indicative of a G1-phase cell with oriC localized at the old cell pole. (13,30) average, the cell lengths were 1.8 ± 0.1 µm for both wild type and SmABCΔR.Plotting the subcellular location of individual markers as function of the genomic position showed a tripartite triangle pattern, indicating arrangement of DNA segments between ori and ter along the longitudinal cell axis (Fig 4C).This is in agreement with a recent study also suggesting a longitudinal oriter distribution of a natural cointegrate of the circular and linear chromosomes in A. tumefaciens. (23)In wild type and SmABCΔR, oriC and oriA/oriB occupied polar and subpolar areas, respectively (S11B Fig) .As expected from our filtering approach and in agreement with previous reports, (13,30) in both strains, the oriC signal showed a very low positional variance at one cell pole (S1 Table ).Subpolar localization of oriA and oriB signals with similar variances in both strains suggests a spatial confinement of these elements to this region of the cell (S1 Table ).The terC fluorescence signal was enriched with low positional variance at the cell pole opposite to the pole exhibiting the ParB-cerulean signal in wild type and SmABCΔR (S11B Fig, ).In contrast, the average spatial positions of terA and terB signals differed between these strains (S11B Fig, S1 Table).Especially, the spatial shift of terA in the triple replicon fusion strain argues against active positioning.

S1 Table
Consistent with tethering of oriC and terC to opposite cell poles, the average subcellular signal positions of chromosomal markers 2, 3 and 18, which map in close vicinity to oriC or terC, did not much differ between both strains (Fig 4D, S1 Table).Changes in the average spatial positions were found for the fluorescent signals of markers 5/6, 9/10 and 15/16, which flank the replicon fusions sites in SmABCΔR compared to the wild type (Fig 4C, S1 Table).This was expected since the markers of each pair are situated on different DNA molecules in the wild type, whereas they were brought into close proximity by the replicon fusions.
Tethering of ori and ter regions to factors localized at cell poles was previously reported for several bacteria, including the α-proteobacteria C. crescentus and A. tumefaciens. (48)(55) It can be assumed that orthologous proteins and similar mechanisms are also involved in the spatial confinement of oriC and terC at the old and new cell poles, respectively, in S. meliloti.However, factors responsible for the spatial confinement of oriA and oriB to the subpolar region of the old pole are still enigmatic.Factors confining the three oris and terC to cell poles are probably decisive for the spatial organization of the merged three replicons.
To test the hypothesis that the spatial organization of DNA in SmABCΔR is only determined by the positioning of locally confined ori and ter regions and the inherent features of the DNA as semi-flexible polymer of compacted units, we simulated the DNA arrangement for genome configurations differing in number and position of confined loci.We expanded a previously described model (56) by implementing not only one origin and one terminus as possible fixpoints, but three of each.By ergodic sampling over 200 configurations (Fig 4E) using the MOS-algorithm (57) the average genomic organization was obtained.

Spatiotemporal dynamics of origin and terminus regions in replicon fusions strains
To study the effect of bi-and monopartite genome configurations and reduced number of replication initiation sites on DNA segregation, we analyzed the spatiotemporal dynamics of the ori and ter regions in SmABΔR, SmABCΔR and SmABCΔoriVΔR in comparison to the wild type.For this purpose, we followed the same labeling strategy as described above, and validated the strain's genome configurations by PFGE (S1 Data).At the level of individual cells, microscopic snap-shot and time lapse data of combinations of labeled oriC/oriA/oriB, oriC/terC, oriC/oriA/terA and oriC/oriB/terB were generated.
The spatiotemporal choreography of origins and termini was similar in the replicon fusion strains and wild type.A previous study showed a clear spatiotemporal pattern of duplication and segregation of the three oris in the S. meliloti wild type. (13)This was characterized by initial occurrence of two oriC copies at the old pole, indicative of the start of chromosome segregation (approximately in the first quarter of the cell cycle), and followed by translocation of one oriC copy to the new cell pole.The oriVs translocate from the subpolar region of the old cell pole to the midcell area, where, after oriC segregation is complete,  ).
The previous study indicated a preferential temporal order of oriA and oriB segregation. (13)We asked if this order is conserved in cells with fused replicons.Upon detailed examination, we found that, duplication of oriA and oriB foci was very close in time in both the mother and daughter cells of wild type, In the remaining proportion of cells that showed sequential segregation of these origins, oriA segregation preceded that of oriB in three-fourth of these cells (Fig 5E), indicating that the previously observed temporal order (13) of duplication and segregation first of oriA and then of oriB was less stringent both in wild type and the replicon fusion strains.
In conclusion, the time-lapse analyses of the replicon fusion strains SmABΔR and SmABCΔR indicate that their spatiotemporal choreography of the origin and terminus regions is similar to that of the wild type with a tripartite genome configuration.This suggests that regulation of replication initiation and the processes relevant to segregation have been conserved.

The chronology of segregation and spatial position of oriC and ΔoriA/B regions are altered in
SmABCΔoriVΔR.Strain SmABCΔoriVΔR carries monopartite genomic DNA containing oriC but lacking both secondary replicon origins (repA1B1C1, oriB and repA2B2C2, oriA).We studied the segregation chronology of active oriC, inactive oriA/B, and terminus regions in this strain to learn about the segregation properties of the 6.7 Mbp single replicon.
In an initial fluorescence microcopy snapshot series, we analyzed putative G1-phase SmABCΔoriVΔR cells filtered with a cut-off of 2.0 µm cell length and deduced the relative distance of oriC and both ∆oriA/B foci to the cell equator.Striking differences were found between oriC localization in SmABCΔoriVΔR   Furthermore, the frequent colocalization of these regions (Fig 6B) even after loss of their original cellular position would be consistent with ori-ori clustering as described for the replicons in A. tumefaciens. (52) addition to the spatial alterations in segregation patterns in SmABCΔoriVΔR, we observed also alterations in temporal patterns associated to ori and ter regions.Translocation of the terC focus occurred already in the beginning of the second cell cycle quarter, and not shortly after the midpoint of the cell cycle (Fig 5Cii).This shift in timing was also reflected in earlier doubling of terC and also terA foci (Fig 5A).
Moreover, duplicated foci representing the oriA regions appeared later (halfway point of the cell cycle vs.second quarter of the cell cycle (Fig 5A).Thus, in SmABCΔoriVΔR cells, visible segregation of terA occurred before that of the oriA region.Also the order of duplication of terB and oriB foci was changed, with duplication of terB occurring on average before that of oriB (Fig. 5A).These alterations in chronology of foci duplication were observed in mother and in daughter cells (S13A Fig, S3A Table), and correlate with the position on the replicon sequence and the distance to oriC.

A missense mutation in cell cycle kinase-encoding cckA increased fitness of SmABCΔoriVΔR, but is unlikely to be responsible for aberrant ori localizations
In addition to genome sequencing of SmAB, SmABC, and SmABCoriV (see above), we also determined the genome sequence of SmABΔR, SmABCΔR, and SmABCΔoriVΔR (S6 Table ).In particular, a SNV in the coding sequence SMc00471 of the cell cycle histidine kinase CckA in SmABCΔoriVΔR attracted our attention.This SNV that causes amino acid substitution R436H (S18A Fig, S6 Table) was not found in the precursor strains.
To test whether this missense mutation was responsible for the aberrant oriC localization pattern in strain SmABCΔoriVΔR we attempted to revert the SNV to the wild-type cckA sequence in this strain, and failed.We then introduced this missense mutation into the wild type and SmABCΔR and analyzed the  In addition, we used ParB-Cerlulean to analyze the oriC localization pattern in SmABCΔoriV, the direct precursor of SmABCΔoriVΔR.SmABCΔoriV carries the cckA wild type sequence and is already deleted for both secondary replicon replication origins (oriA/B).We found that the oriC localization pattern in SmABCΔoriV already deviates from the wild type-like pattern in SmABC, which has three intact replication origins (Fig 7B).This implies that CckA R436H did not cause the observed mislocalization of oriC in SmABCΔoriVΔR.However, we found a strong difference in growth between SmABCΔoriV (CckA) and Cell cycle regulation by the CckA-ChpT-CtrA signaling pathway is wide-spread in α-proteobacteria. (58)e CckA phosphorelay controls the phosphorylation status of ChpT which in turn regulates activity and stability of the cell cycle master regulator CtrA through phosphorylation. (59)CckA phosphatase activity enables initiation of DNA replication through dephosphorylation and degradation of CtrA.(59,60) The S. meliloti (S18A Fig) and C. crescentus CckA domain composition (61) is very similar.The R436H substitution in S. meliloti CckA locates in one of the PAS domains, which in C. crescentus were shown to regulate switching between the CckA kinase and phosphatase activities. (61,62)Narayanan and coworkers (63) identified a point mutation in the PAS-B domain of C. crescentus CckA that suppresses a topoisomerase IV inhibitor-induced chromosome segregation defect, possibly by slowing down the chromosome replication cycle.CckA R436H is therefore proposed to mitigate chromosome replication and/or segregation defects in a similar manner.We speculate that the R436H substitution promotes kinase or reduces phosphatase activity of S. meliloti CckA, which results in higher levels of CtrA-P repressing replication initiation and slowing down the cell cycle.

A 3.2 Mbp genomic inversion relocalizes terC opposite to oriC, and alleviates the growth deficit of SmABCΔoriV
While establishing strain SmABCΔoriVΔR, we detected a further derivative, which showed ~ 75 % of the growth rate of SmABC and the wild type (Fig. 8A).This clone showed altered banding patterns in PFGE analysis (Fig. 8B), suggestive of genomic rearrangement events.Combining nanopore and Illumina sequencing data, we de novo assembled the genome sequence of this strain, and detected a 3.2 Mbp large inversion.The inversion in this strain, hereafter called SmABCΔoriVΔRInv, occurred between the groEL1/S1 and groEL2/S2 paralogs, which map to the chromosome and pSymA sequences, respectively.This inversion positioned the terC region opposite to oriC (Fig. 8C), and realigned KOPS across the single replicon (Fig. 8D).The CckA R436H missense mutation characteristic for SmABCΔoriVΔR was not detectable in SmABCΔoriVΔRInv (S6 Tab.).Remarkably, the point mutation in cckA alleviated the growth deficiency of the ancestral strain almost as much as this massive genetic rearrangement (Fig. 8A).Previous studies of engineered E. coli strains indicate that disturbing chromosome organization patterns by integration of one or more additional oriC copies or re-localization of oriC to ectopic locations can cause replication-transcription conflicts and issues with replication fork trap regions, which affect DNA replication and segregation, and promote selection of phenotype-moderating genome rearrangements and genetic suppressions. (64,65)In naturally occurring single-chromosome Vibrio (NSCV) strains of V. cholerae, a second origin was found to be either active or silenced depending on the position of the cointegration event, (66) whereas in laboratory-generated fusions the chromosome 2 replication machinery was not functional. (67,68)Recently, a derivative of the ultrafast growing Vibrio natriegens with fused chromosomes 1 and 2 was constructed, which is not impaired in growth under standard laboratory culture conditions. (69)e design of this strain preserved only the ori and ter region of chromosome 1 and maintained genomic symmetries.The results studies in different bacterial species and our study in S. meliloti suggest that retaining replichore symmetry and replichore-orienting sequence element distributions as well as avoiding replication termination traps are important design rules for the fitness of replicon fusion strains.But even strains that do not sufficiently fulfill these rules but are still viable can be evolvable into fitter strains.

◼ CONCLUSION
Members of the α-, β-, or -proteobacteria are specialists in occupying ecological niches, not least due to their genetic diversity, which is enhanced by segmented genome configurations.Because of their metabolic diversity, several of these bacteria are already used extensively in the fields of biotechnology, climate protection, plant genetics and sustainable agriculture and are thus of great value for the transformation of a future bioeconomy.Although there is a benefit from the segmented genomic makeup in terms of functional gain through facilitated horizontal gene transfer, the genome complexity often limits genetic engineering efforts that are essential for strain development.Replicon fusions appear to be an effective strategy for reducing genome complexity while preserving gene content.A caveat of this strategy is that changes in genomic organization can lead to impairment of fundamental cell cycle functions and thus to reduced fitness.The challenge is therefore to avoid such negative effects through appropriate design and/or evolution of the engineered strains, and to well characterize the new chassis in the desired culture conditions.
Our study provides S. meliloti strains with bi-and monopartite genome configurations through sitespecific recombination-mediated targeted replicon fusions.In the fused replicons, we aimed at retaining organizational properties, such as replichore ratios, GC skew, as well as distribution and orientation of KOPS and coding sequences.In contrast to a natural S. meliloti isolate with triple replicon cointegrate that showed a high frequency of revertants by homologous recombination, (21) our replicon fusion strategy reduced the likelihood of revertants.This enabled investigating spatial DNA organization and spatiotemporal segregation patterns of a monopartite triple-replicon bacterial genome, mostly independent of spontaneously occurring genome rearrangements in the studied strains.These analyses and growth tests showed that this strain can be applied under standard laboratory culture conditions.The strongly reduced fitness of the derivative strain, which is characterized by oriC as the sole origin of replication and several deviations from the patterns of replicon architecture, underlines the importance of these architectural rules.
However, suppressor mutations that largely restored the fitness of this strain show that suitable strains can also be obtained from an unfavorable design through evolution.These S. meliloti strains with monopartite and single origin of replication genome configuration and good growth characteristics are now available for genome engineering applications that were previously hampered or prevented by the tripartite genome structure.

DNA manipulation and plasmid extraction.
Plasmids used in this study are listed in S8 Table .Standard molecular techniques were employed for cloning and transfer of nucleic acids. (70)DNA fragments were PCR amplified using Q5® High-Fidelity DNA Polymerase (New England Biolabs) or Taq DNA Polymerase (New England Biolabs).DNA oligonucleotides were provided by Sigma-Aldrich (USA) and Integrated DNA Technologies (USA) (S11 Table ).For DNA purification and gel extractions the E.Z.N.A.® Cycle-Pure Kit (Omega Bio-Tek) and illustra ™ GFX ™ PCR DNA and Gel Band Purification Kit (GE Healthcare Life Sciences) were used, respectively.For phosphorylation of the 5′ hydroxyl terminus of PCR amplicons and oligonucleotides T4 Polynucleotide Kinase (Thermo Scientific) was applied.Dephosphorylation of DNA ends was performed by use of FastAP ™ Thermosensitive Alkaline Phosphatase (Thermo Scientific).Filling in of 5′-overhangs in double stranded DNA to form blunt ends was achieved using the large fragment of DNA Polymerase I (Klenow fragment) (Thermo Scientific).T4 Ligase (Thermo Scientific) was used for ligation.Plasmid DNA was isolated using the "E.Z.N.A. Plasmid Mini Kit" (Omega Bio-Tek).All enzymatic catalyzation and purification steps were performed according to the manufacturer's protocols and instructions.For sequence verification of plasmid and amplified DNA the sanger sequencing service of Eurofins Genomics (Germany) was used.For detailed information on the construction of individual plasmids refer to S10 Table .Strain construction.Strains generated in this study are listed in S9 Table .Transfer of plasmids to S. meliloti was achieved by conjugation using E. coli S17-1 (73) or by electroporation as previously described. (31)Cells for electroporation were prepared as described in Ferri et al.. (74) Markerless integrations through double homologous recombination were carried out using pK18mobsacB derivatives and sucrose selection. (75) the basis of a Cre/lox toolbox and S. meliloti Rm1021 derivative SmCreΔhsdR (31) the tripartite genome was merged in two consecutive steps.First, pSymA and pSymB were fused with each other, giving rise to the megaplasmid hybrid pSymAB harbored by S. meliloti strain SmAB.Therefore, pK18mobsacB derivatives pJD98 and pJD99 were used to integrate lox sites and antibiotic selection markers into SmCreΔhsdR for the site-specific recombination.After removal of active lox sites, SmAB was sequentially transformed with constructs pJD130 and pJD126 again providing lox sites and an additional antibiotic selection marker for integration of the chromosome and pSymB.Cre-mediated integration of pSymAB into the chromosome gave rise to SmABC with monopartite genome configuration.Cre/lox applications were performed as described before. (31)Illustration of SmAB and SmABC strain generation and detailed information about the construction process is given in S1 Fig and S9 Table, respectively.
In order to remove DNA replication origins of strain SmABC, deletion constructs pJD201 and pJD202 were used for sequential excision of the megaplasmid-encoded copies of repC and corresponding repBC intergenic regions.The deletion of both regions in SmABC resulted in strain SmABCΔoriV.For functional inactivation of the inherent partitioning system in pSymA and pSymB, the constructs pMW210 (repA1B1C1), pMW211 (repA2B2C2) and pJD225 (repA3B3) were used to delete the hole operon and the most proximal partitioning sites in SmCreΔhsdR, SmAB and SmABC.In an attempt to delete the partitioning sites upstream of repA2B2C2, pMW230 was used.Additionally, the constructs pJD206 (repA2), pJD207 (repA1) and pJD266 (repA3) were used to inactivate the partitioning system and the entire operon by single crossover integrations into the respective RepA coding regions.Gene deletions and plasmid integrations were verified by PCR.By use of deletion constructs pJD222 and pJD229, S. meliloti strains SmAB, SmABC and SmABCΔoriV were further cured from spectinomycin and gentamicin resistance cassettes, respectively, giving rise to strains SmABΔR, SmABCΔR and SmABCΔoriVΔR accessible for constructs of the replicon labeling system.For in vivo studies of DNA organization and spatiotemporal dynamics of origin and terminus by fluorescence microscopy, a triple label system based on the fluorescent reporter gene fusions tetR-mVenus and lacI-mCherry (derived from the FROS (46) ) and parB-cerulean was developed.Initially, S. meliloti strains SmCreΔhsdR, SmABΔR, SmABCΔR and SmABCΔoriVΔR were transformed with pMW198.

Strain validation.
Pulsed-Field Gel Electrophoresis (PFGE) was used as method to validate the genome architecture after major fusion and integration/deletion steps.The applied PFGE protocol for DNA preparation and digestion was carried out as described for strain validations in Checcucci et al.. (20) To gain fusion strain characteristic banding patterns the restriction digestion of genomic DNA was performed with PacI (New England Biolabs, USA).For PFGE analysis, ¼ agarose plug with treated genomic DNA was separated in an 0.7% agarose gel (Pulse Field Certified Agarose, Bio-Rad, USA) and 0.5x TBE buffer at 12°C (44.5mMTris-HCl, 44.5mM boric acid, 1mM EDTA) using the Rotaphor® System 6.0 (Analytik Jena, Germany) according to the manufacturer's instructions.Separation of DNA fragments was achieved with 130V-100V for 50-175sec at 130°-110° (run time 18h), 130V-80V for 175sec-500sec at 110° (run time 18h) and 80V-50V for 500sec-2000sec at 106° (run time 40h) with a logarithmic course of increase or decrease between varying parameters, respectively.
For sequence specific analysis, including verification of proper deletions and detection of singlenucleotide variants (SNVs), all basic strains (SmCreΔhsdR, SmAB, SmABC, SmABCΔoriV, SmABΔ, SmABCΔR and SmABCΔoriVΔR) were subjected to next generation DNA sequencing using the MiSeq™ System (Illumina, USA).For preparation of genomic DNA, S. meliloti cells were grown in TY supplemented with appropriate antibiotics and harvested at optical density600 (OD600) of 1.0 by centrifugation at 3000g (4°C).Sample preparation was performed as previously described. (76)The investigation for single nucleotide variations was carried out using the Basic Variant Detection tool (v.2.1) of CLC genomic workbench (v.20.0.4) with a minimum coverage of eight, minimum count of four and minimum frequency of 50% for mapped reads.Next, each SNV was manually analyzed by eye and, in case of doubt, additionally validated by Sanger sequencing.Assembly of the SmABC∆oriV∆RInv genome was performed using data from a whole genome nanopore sequencing approach (MinION; Oxford Nanopore Technologies plc) and the de novo assembly algorithm for long reads of the CLC Genomic Workbench (v.22.0.2).The single contig genomic scaffold was then refined with data (short read) from the MiSeq™ system using the polish with reads tool.
Growth experiments.Prior to inoculation, overnight cultures were washed with 0.9 % NaCl and adjusted to OD600 of 0.01-0.15 in TY or MOPS buffered medium supplemented with 600 mg/ml streptomycin.
Cultures were incubated in a 100 µl volume in a 96 well microtiter plate at 30°C and with shaking at 200 rpm.OD600 of cell cultures was measured every 30 min with a microplate reader (Tecan 200 PRO, Tecan, Switzerland).
Live cell microscopy and image analysis.S. meliloti cell cultures were grown in TY medium (glass tubes) supplemented with suitable antibiotics to OD600 of 0.25 for time lapse and OD600 of 0.5 for snapshot microscopy.To enrich cultures with G1-phase cells for snapshot analysis, strains were grown to OD600 of 1.6 -1.8, diluted to OD600 of 0.5 and subsequently used for microscopy. 1 µl of these cultures were then placed onto 1 % (w/v) molecular biology-grade agarose (Eurogentec, Belgium) pads containing ddH2O (snap shots) or MOPS minimal medium and suitable antibiotics (time lapse), covered with a cover glass and sealed with VALAP. (77)For visual examination of S. meliloti cells by phase contrast and epifluorescence microscopy an Eclipse Ti-E inverse research microscope (Nikon, Japan) equipped with a 100x CFI Plan Apo1 oil objective (numerical aperture of 1.45), a green DPSS solid state laser (561 nm, 50 mW; Sapphire) and a multiline Argonlaser (457/488/514 nm, 65 mW; Melles Griot) with AHF HC filter sets F36-513 DAPI (excitation band pass [ex bp] 387/11 nm, beam splitter [bs] 409 nm, emission [em] bp 447/60 nm), F36-504 mCherry (ex bp 562/40 nm, bs 593 nm, em 624/40 nm), F36-528 mVenus (ex bp 500/24 nm, bs 520 nm, and em bp 542/27 nm) was used.Exposure times ranged from 200 ms to 2 s.Image acquisition and adjustment was done with an Andor iXon3 885 electron-multiplyingcharge-coupled device (EMCCD) camera and the software NIS-Elements v.4.13 (Nikon, Japan), respectively.Time-lapse analysis was performed at 30°C in a microscope incubator and images were acquired every 2 or 5 minutes.Analysis of snap-shot and time-lapse microscopy images was performed using ImageJ plug-in MicrobeJ. (78)G1phase cells were filtered for a maximal length of 2.0 µm and presence a single ParB-cerulean focus indicative of non-segregated oriC.
Marker frequency analysis.For marker frequency analysis, S. meliloti strains SmABC and SmABCΔoriV were grown in 50 ml TY medium supplemented with 600 mg/ml streptomycin at an initial OD600 of 0.1.
After incubation at 30°C and 200 rpm, samples were taken at OD600 of 0.6 (exponential phase) or OD600 of ~ 2.6 (overnight culture, stationary phase).Cells were harvested by centrifugation (4000 g, 4°C) and immediately frozen in liquid nitrogen.Preparation and acquisition of Illumina Miseq data were performed as previously described. (76)Paired-end reads were then mapped by the QuasR R package (v1.6.2) onto the S. meliloti replicons.Only unique hits were considered.Subsequently, the coverage was determined from the obtained mapped genomic DNA reads using the genomecov from the bedtools toolbox (v2.25.0).The average coverage of all samples ranged between 18 and 23.The coverage was normalized by the total coverage (sum of coverage) of each sample.To identify minimal variations in the copy number along the replicons, we used sliding window averaging.The size of the window comprised 200 kb.After averaging, the value at a certain position reflects the average coverage of about 2 % of the replicon left and right of the indicated position.This process averages out random noise and local sequence specific variation.To determine the copy number without prior information about the terminus region, the lower 10 % quantile of all windows was used to determine the reads in the terminus region.All windows where then normalized by this value, resulting in a copy number relative to the terminus region.
Modeling.For the simulations we used a model for DNA described by Buenemann and Lenz. (56)The basic assumptions of the model are 1) DNA can be modeled as a sequence of compacted units (S1 Text); 2) compact units can be restricted in their spatial arrangement e.g. by the action of proteins; and 3) the measured organization of the chromosome in the cell results from averaging over many individual configurations that meet these constraints.For C. crescentus this model revealed that self-avoidance of DNA, specific positioning of the origin (and terminus) region and the compaction of DNA are sufficient to explain the strong linear correlation between specific positions on the chromosome and their longitudinal arrangement within the cell. (57)To predict the spatial organization of the merged replicons in the S. meliloti replicon fusion strain SmABC an expansion of the model by implementing not only one origin and one terminus as fixpoints, but three each was done.For realization, the A* algorithm was added to the model, which made it possible to generate random walks between any number of fixed points (S2 Text).Bioinformatic analysis.GC and GC c skew analysis of the tri-, bi-and monopartite S. meliloti genome was performed using GenSkew (http://genskew.csb.univie.ac.at,Feb. 2020).GC skew depictions of the individual replicons as shown in Fig 1 were generated using the CGView server (http://stothard.afns.ualberta.ca/cgview_server/Feb. 2020).Oligonucleotide skews for KOPS were calculated with fuzznuc (http://emboss.toulouse.inra.fr/cgi-bin/emboss/fuzznuc?_pref_hide_optional=1, Mar.2020).Functional domain analysis of S. meliloti CckA was done using the NCBI conserved domain database (CDSEARCH/cdd) (79) with low complexity filter, composition-based adjustment and an E-value threshold of 0.01 (Oct.2021).

Figure 2 .
Figure 2. Basic characterization of S. meliloti replicon fusion strains SmAB and SmABC.(A) Phase contrast microscopy images of SmAB, SmABC and precursor strain SmCreΔhsdR (wt) representative cells at different stages of the cell cycle.Scale bar: 2 µm.(B) Growth of S. meliloti SmAB and SmABC compared to precursor strain SmCreΔhsdR (wt) in rich medium (TY), minimal medium (MOPS low Pi), high salt medium (TY + 0.4 M NaCl) and high sucrose medium (TY + 0.5 M sucrose).Data represent the mean ± standard deviation of three technical replicates.Growth curves of biological replicates are shown in S6A Fig.
Fig).In the wild type, deletion of these repC regions was not achieved.However, SmAB or SmABC lacking either of these regions (SmABΔrepC1, SmABΔrepC2 or SmABCΔrepC1, SmABCΔrepC2) or SmABC lacking both these regions (SmABCΔoriV) were obtained (S7B Fig).The configuration of the genomic DNA in all these strains was validated by PFGE (S7C Fig).Illumina sequencing verified the intended genetic changes in SmAB, SmABC and SmABCΔoriV, and revealed the modifications resulting from the construction process (S4Table, S5 Table

Figure 3 .
Figure 3. Deletion and mutational studies in S. meliloti SmCreΔhsdR (wt), SmAB and SmABC fusion strains.(A) Growth curves of S. meliloti SmAB and SmABC repC deletion strains.Prior to inoculation, overnight cultures were washed with 0.9 % NaCl and adjusted to an OD600 ~ 0.15 in TY medium.Mean and standard deviation was calculated from three technical replicates.(B) Morphology of the double repC deletion mutant strain SmABCΔoriV.(C) Marker frequency analysis of SmABC and SmABCΔoriV of logarithmic (OD600 of 0.6) vs. stationary (OD600 of 2.6) cultures.Trimmed and normalized marker frequencies are depicted in log2 as a function of the genome position in Mbp.Arrowheads indicate the position of origin and predicted terminus regions.(38)Note that SmABCΔoriV lacks oriA and oriB due to repC deletions (ΔoriA and ΔoriB).(D) True to scale scheme of the Indeed, we found good evidence for Caulobacter crescentus-like parS sites localized close to oriC in S. meliloti (S10C Fig).In our study, we employed a ParB-cerulean fusion to label oriC.To this end, parB at its native chromosomal locus was replaced by a parB-cerulean fusion in SmABΔR, SmABCΔR, SmABCΔoriVΔR, and the wild type (S10A Fig).Growth of the parental strains and corresponding derivative strains was indistinguishable in TY medium (S10B Fig), suggesting that the ParB-cerulean protein was functional.

Figure 4 .
Figure 4. Investigation and modeling of the spatial DNA organization in SmABCΔR.(A) Schematic true to scale representation of the SmABCΔR monopartite genome with tetO and lacO integration sites 1-18 selected to reveal the spatial configuration of the genomic DNA.Color code: chromosome (grey), pSymA (green), pSymB (blue).Red circles: replication origins, black diamonds: terminus regions.(B) Example of snapshot images from labeled cells used for the 2D genome mapping study.Scale bar: 1 µm.(C) Normalized spatial localization of labeled loci within SmABCΔR (filled circles) compared to SmCreΔhsdR (non-filled circles) with wild type genome configuration as a function of the genome sequence coordinate.Old cell pole: 1, New cell pole: -1.(D) Scatter plots of selected strains illustrate examples of similar (Pos.no.: 2, 3, 18) and clearly different distribution (Pos.no.: 5, 10, 16) of marked loci within SmABCΔR (red dots) and SmCreΔhsdR (black dots).(E) Example configuration of simulations consolidating physical principles such as self-avoidance and compaction of DNA in a SmABCΔR cell represented as spherocylindrical shape.(F) Model of DNA self-organization in SmABCΔR compared to the experimental data.The model considers a spatial confinement (experimental standard deviation) for terA and terB in addition to oriABC and terC as fixpoints.Shown are the normalized locations in the cell as a function of the position on the genomic map.Experimental data of origins are indicated by red circles, terminus regions by black diamonds and remaining marker positions with non-filled circles.The red line depicts the model results averaged over 276 cells.Shaded areas represent the standard deviations.For the model a cell of 1800 nm and a loop-size of 1298 bp (DNA within a "blob") was used.
Initially, we generated a model including only oriC (S12A Fig) and oriC/terC (S12B Fig) as anchoring points of the 6.7 Mbp DNA molecule at the old and the new cell poles.We found that confining these two loci are not sufficient to describe the spatial arrangement of pSymA-and pSymB-derived DNA.By anchoring of all three origins to the experimentally determined average subcellular position, we gained a ternary model structure with highly variable DNA segments between the three points (S12C Fig).However, in this model structure, the DNA segment between oriC and oriB did not extend up to the opposite cell pole.We then confined terC to the new cell pole, and oriC to the polar, and oriA and oriB to the subpolar regions of the old cell pole (S12D Fig) since these loci showed the smallest positional variance in the experimental data (S12E Fig).With these four anchored points the generated model structure already reflected the experimentally determined positions of loci 2, 3 and 5 as well as locus 6 located on the chromosome and pSymB, respectively.However, it did not describe the experimentally determined average localization of terA and terB.Loci close to terA and terB showed a high positional variance in the experimental data (S12E Fig).By a spatial confinement of terA and terB within the experimental data variance we gained a tripartite triangle structure (Fig 4F), albeit this model did not reproduce the experimentally determined average localization of loci 13 to 18.

Figure 5 .
Figure 5. Spatiotemporal pattern of origin and terminus regions in native and reorganized multi-replicon backgrounds.(A) Temporal order of origin and terminus region segregation in SmCreΔhsdR (wt), SmABΔR (AB), SmABCΔR (ABC) and SmABCΔoriVΔR (ABCΔ).Colored bars with a black center indicate the standard deviation and the mean timepoint of sister foci separation within the cell cycle normalized to 100 %.Analyzed M-cells: 20 (oriC), 10 (oriA/B) and 5 (terA/B/C).(B) Localization of ori and ter foci segregation in SmCreΔhsdR (wt), SmABΔR, SmABCΔR and SmABCΔoriVΔR.The bar chart depicts the relative longitudinal position of a single focus before separation within the normalized cell (old pole: 0, new pole: 1, midcell: 0,5).Data represent the mean ± the standard deviation for oriC (n=20), terC (n=5), oriA (n=10), terA (n=5), oriB (n=10) and terB (n=5) in M-cells.(C) Spatiotemporal choreography of oriC and terC in SmCreΔhsdR (wt), SmABΔR, SmABCΔR and SmABCΔoriVΔR.(i) Time lapse series with oriC (ParB-cerulean) and terC (TetR-mVenus) in SmCreΔhsdR (wt) and SmABCΔR.Scale bar: 1 µm.(ii) Timepoint of oriC (cyan circles) arriving at the new cell pole in relation to terC (yellow circles) release from the same pole within a normalized cell cycle (0-100 %).Analyzed M-cells: 5 each.Black bar indicates the data mean.(D) Time lapse microscopy series of SmCreΔhsdR (wt) and SmABCΔR as examples of sequential segregation of oriA and oriB with varying order.Arrowheads depict the position of segregation start.(E) Percentage of cells with oriA foci (green) and oriB foci (blue) segregating first.Cells SmABΔR and SmABCΔR (S14A Fig).It was observed either simultaneously (S14B Fig) or sequentially (Fig 5D), with oriA starting segregation before oriB in the majority of cells.In up to a fifth of wild type, SmABΔR and SmABCΔR cells analyzed (n=200 each), we found the doubling of oriA and oriB foci within the same time frame of 5 minutes (S14C Fig).
compared to wild type, SmABΔR and SmABCΔR cells.In contrast to wild type, SmABΔR and SmABCΔR cells that showed oriC foci clustering predominantly in the cell pole area, a considerable proportion of SmABCΔoriVΔR cells showed oriC localization in the midcell area (Fig 6A, S15 Fig).We also observed that in SmABCΔoriVΔR, both ∆oriA/B foci localized more frequently in the midcell area than in the cell pole area (S15 Fig), and that the proportion of cells showing colocalization of both ΔoriA/B foci with oriC in the midcell area was considerably higher than in wild type, SmABΔR and SmABCΔR cells (Fig 6B, S16 Fig).Collectively, this data suggests aberrant localization of oriC in more than one third of SmABCΔoriVΔR G1 -phase cells.

Pattern 1
was characterized by visible oriC foci segregation in the midcell area (Fig 5B, Fig 6D), which ultimately resulted in localization of one oriC focus (defined as oriC2) at the new cell pole and the other (defined as oriC1) in the mid-area of the mother cell compartment (S17 Fig).Pattern 2 was characterized by duplication of the oriC focus (oriC1) at the old cell pole followed by translocation of one of the oriC foci (oriC2) to the new pole (Fig 6D).Right before cell division, oriC1 localized either at the old cell pole or in the mid-area of the mother cell compartment in 30% and 70% of the cells, respectively (S17 Fig).We identified the pattern 2 cells as the daughter cells since they most frequently adopted the oriC localization pattern of the ancestral cell (i.e.mother cell, pattern 1 cell) after one cell cycle.Moreover, we found that in mostly all mother cells analyzed displaying aberrant oriC localization also the ∆oriA and ∆oriB regions lost their subpolar localization (S2D Data).Notably, oriB duplication in SmABCΔoriVΔR cells was observed in the new cell pole compartment and not in the midcell region as in wild type, SmABΔR and SmABCΔR cells (Fig 5B).Loss of polar oriC and subpolar ∆oriA and ∆oriB localization suggests that the deletion of repC/oriV directly or indirectly affect mechanisms relevant for anchoring of these sequence elements.
localization of oriC right before cell division in G1-phase cells, using ParB-Cerlulean for fluorescent labeling (Fig 7A, S18B Fig, S18C Fig).Regarding polar (approx.90 %) and midcell (approx.10 %) localization of the ParB-Cerlulean mediated fluorescent focus, we found no difference between the strains carrying the SNV in cckA and the corresponding strains with wild type cckA (Fig 7A).However, in this comparison, growth of the strains with CckA R436H was slightly reduced (S18D Fig).
SmABCΔoriVΔR (CckA R436H ) (S18E Fig), indicating that the cckAR436H allele mitigates the strong growth deficiency caused by deletion of the oriVs.In a growth comparison, SmABCΔoriVΔR (CckA R436H ) reached ~ 65 % of the growth rate of SmABC and the wild type in the exponential phase (Fig 8A).