ParA and ParB coordinate chromosome segregation with cell elongation and division during Streptomyces sporulation

In unicellular bacteria, the ParA and ParB proteins segregate chromosomes and coordinate this process with cell division and chromosome replication. During sporulation of mycelial Streptomyces, ParA and ParB uniformly distribute multiple chromosomes along the filamentous sporogenic hyphal compartment, which then differentiates into a chain of unigenomic spores. However, chromosome segregation must be coordinated with cell elongation and multiple divisions. Here, we addressed the question of whether ParA and ParB are involved in the synchronization of cell-cycle processes during sporulation in Streptomyces. To answer this question, we used time-lapse microscopy, which allows the monitoring of growth and division of single sporogenic hyphae. We showed that sporogenic hyphae stop extending at the time of ParA accumulation and Z-ring formation. We demonstrated that both ParA and ParB affect the rate of hyphal extension. Additionally, we showed that ParA promotes the formation of massive nucleoprotein complexes by ParB. We also showed that FtsZ ring assembly is affected by the ParB protein and/or unsegregated DNA. Our results indicate the existence of a checkpoint between the extension and septation of sporogenic hyphae that involves the ParA and ParB proteins.


Introduction
Robust cell proliferation requires precise coordination of key cell-cycle processes. In most rod-shaped bacteria, chromosome replication and segregation occur during longitudinal cell elongation and are shortly followed by cell division. Active chromosome segregation begins soon after the initiation of replication in the oriC region (origin of replication). The ParAB-parS system mediates the migration of newly replicated oriCs either unidirectionally (e.g. in Caulobacter crescentus) or bidirectionally (e.g. in vegetatively growing Bacillus subtilis) towards the cell pole(s) [1]. Nucleoprotein complexes named segrosomes are formed by ParB binding to parS sites, which are usually scattered in the proximity of the oriC region [2][3][4][5]. The movement of the segrosomes requires an interaction between ParB and ParA as well as the ATPase activity of ParA, its dimerization and interaction with DNA and/or polymerization [6][7][8]. Interestingly, parAB deletion is associated with diverse phenotypes in different bacterial species; in C. crescentus the genes are essential, but in B. subtilis their deletion has only a minor effect on the growth. The explanation for these differences is that ParAB proteins are involved in the coordination of chromosome segregation with different cell-cycle processes, such as chromosome replication (in B. subtilis and Vibrio cholerae), cell growth and/ or cell division (C. crescentus, Mycobacterium smegmatis and Streptomyces coelicolor) [9][10][11][12][13]. Thus, the cell-cycle coordinating mechanisms involving ParAB appear to be specific to particular bacterial species.
Chromosome segregation must be synchronized with the formation of the division septum, which separates the two daughter cells. The septation starts with the polymerization of FtsZ into a ring structure (Z-ring) that determines the division plane [7,14]. In model rod-shaped bacteria (Escherichia coli, B. subtilis), two systems, Min and nucleoid occlusion (NO), prevent guillotining of the unsegregated chromosomes by the division septum [15,16]. Both systems use inhibitors of FtsZ polymerization (MinC and Slm in E. coli or Noc in B. subtilis) that control the timing and positioning of the Z-ring formation [17]. However, these systems are not universal in all bacteria. Many bacteria, such as those belonging to the Actinobacteria, do not appear to possess homologues of Min or NO proteins [18]. The spatio-temporal control of cell division in these bacteria remains to be elucidated.
In mycelial Streptomyces, a member of the phylum Actinobacteria, chromosome segregation and cell division, followed by cell separation, occur exclusively during sporulation. Unlike model rod-shaped bacteria, Streptomyces form hyphae, which extend apically and form branches. During vegetative growth, chromosomes remain uncondensed and their replication is not followed by cell division [19]. This mode of growth leads to the formation of elongated multigenomic compartments that are not physically separated, although they are delimited by rare septa. Interestingly, cell division is not essential for S. coelicolor vegetative growth [20 -22], but it is crucial for sporulation. A sporulating colony produces multigenomic sporogenic hyphae, named aerial hyphae, which subsequently transform into chains of unigenomic exospores [19]. The conversion of hyphae to spore chains involves synchronized, multiple septation accompanied by the condensation and segregation of tens of chromosomes into prespore compartments. In aerial hyphae, FtsZ forms a ladder-like array of regularly spaced Z-rings along the hyphal compartment [23,24]. Synchronized septation imposes special requirements on the coordination of Z-ring positioning with multiple chromosome segregation in Streptomyces. As mentioned above, Streptomyces do not have homologues of Min proteins and nucleoid occlusion proteins (Noc and SlmA). Recent experiments indicated that in S. coelicolor, FstZ is recruited to the septum site by SsgB, which, unlike the system based on Min and NO, is an example of positive control of septal position [25]. However, it is still not known if a negative control mechanism(s) of septum positioning is also present in Streptomyces.
In Streptomyces, as in other rod-shaped bacteria, the ParAB proteins mediate segregation of multiple chromosomes in the sporogenic hyphal compartment [26,27]. The parAB operon (similarly to ftsZ [28]) is under the control of developmentally regulated promoters, which are strongly induced in sporulating hyphae [29]. In vegetative hyphae of S. coelicolor, both proteins are mostly associated with the tips, while during sporulation ParA extends along the hyphae, and ParB forms an array of regularly spaced complexes [26,30]. In the septated hyphae, each prespore contains a single ParB complex, positioning the chromosome between nascent septa. The elimination of ParA and/or ParB from S. coelicolor disturbs chromosome segregation and leads to the formation of anucleate spores [30]. ParAB deletions also affect sporulation septation, suggesting that the distribution of chromosomes determines septa positioning [30].
Interestingly, chromosome segregation and septation are also affected by modified DNA supercoiling, which can be induced by topoisomerase I (TopA) depletion [31]. Moreover, TopA assists in the regular distribution of segrosomes. The aberration in segrosome separation observed in a TopA depletion strain is associated with complete blockage of cell division [31]. Interestingly, moderate TopA depletion increased DNA compaction and did not inhibit septation but affected the distance between the septa. The aberrant septation in strains with chromosome missegregation suggests a role for the nucleoid in the regulation of septa placement in Streptomyces.
The ParAB system in S. coelicolor was also reported to be involved in the control of tip extension based on the interaction between ParA and Scy [11,32]. Scy is a tip-associated protein involved in polar growth. It also affects ParA polymerization and localization, while ParA presumably influences the Scy complex assembly. This idea is supported by our observation that the sporogenic hyphae of the S. coelicolor parA deletion strain are elongated. Although the analysis did not clarify whether parA deletion extends the time of hyphal growth and delays sporulation or affects the rate of hyphal elongation, the dynamic interplay between ParA and Scy at the tips of hyphae was suggested to provide the molecular switch from aerial hyphal extension to sporulation [11].
Thus, previous studies of Streptomyces sporulation have indicated that ParAB proteins play an important role in synchronizing the events during aerial hyphae sporulation. These studies suggested that ParA is involved in the control of sporogenic hyphal extension. They also indicated that by determining the position of nucleoids, segregation proteins affect the placement of the septa. This suggests a negative regulatory mechanism controlling cell division. However, those studies were performed using S. coelicolor as the model species. S. coelicolor aerial hyphae develop exclusively at the interface between medium and air, which impedes (due to impaired sporulation at the interface between agar and coverslip) live cell time-lapse imaging of the whole differentiation process, from the emergence of sporogenic hyphae to the separation of spores [33]. Here, to verify the hypotheses that ParA and ParB have a role in coordination of sporulation events, we used real-time single cell analysis of differentiation, including analyses of hyphal growth, division and chromosome segregation, in a novel model species, Streptomyces venezuelae. An advantage of S. venezuelae as a model species for studies of differentiation is its ability to sporulate in liquid culture (and under the coverslip) [34]. Fluorescence time-lapse microscopy enabled us to precisely measure the time of hypha elongation and sporulation as well as to determine the timing of ParA and FtsZ appearance during differentiation. We showed that segregation proteins control hyphae elongation and division in Streptomyces.

Construction of Streptomyces venezuelae mutant strains
Streptomyces venezuelae mutants were constructed using PCR targeting (electronic supplementary material, table S2), similar to the method described for S. coelicolor [37]. For strain construction details, see the electronic supplementary material. Conjugation from E. coli to S. venezuelae was performed as previously described [38]. All modified strains were verified by PCR and Southern and western blotting analyses to confirm gene deletions. The intensity of EGFP and YPet fluorescence was measured using the 'profile' function of ZEN 2012 software. A straight line was placed along the hyphae length. The corresponding background fluorescence was subtracted from the fluorescence measured for each hypha. The increase of the FtsZ-YPet and ParA-EGFP fluorescence was detected when the average value of the fluorescence intensity along the hyphae increased more than 10% over the fluorescence intensity in the same hyphae at earlier time points. The Z-ring fluorescence intensity was calculated as an average of the seven highest peaks present in the fluorescence profile (referring to the seven most representative Z-rings). Statistical analyses were performed using STATISTICA (StatSoft Inc. 2011, v. 10, www.statsoft.com). Because the assumption of normal distribution was violated, the non-parametric counterpart of ANOVA, the Kruskal-Wallis test, was applied for the comparisons of FtsZ-YPet fluorescence and parameters describing hyphae growth and length in the Streptomyces strains. For comparison of the TopA-depleted strain with the wild-type strain, unpaired Student's t-tests and Mann-Whitney tests were applied accordingly. Differences were considered significant when p-values were lower than 0.05. The average growth rate of each hypha was calculated as the slope of a line fitted to the relationship between the hypha length and growth time.

Microscopy and image analysis
Super-resolution three-dimensional (3D)-SIM imaging was performed with a V3 DeltaVision OMX 3D-SIM Blaze system (Applied Precision/GE Healthcare) equipped with a Â60/1.42 oil UPlanSApo objective (Olympus), 405 nm and 488 nm diode lasers and three sCMOS cameras (PCO). Each 3D-SIM stack was composed of 225 images (512 Â 512 pixels) consisting of 15 z-sections (125 nm z-distance), with 15 images per z-section, the striped illumination pattern rotated to three angles (2608, 08, þ608) and shifted in fivephase steps. Acquisition settings were 100 ms exposure with a 488 nm laser (attenuated to 100% transmission). The reconstruction of 3D-SIM raw data were performed with SOFTWORX 6.0 (Applied Precision) using a Wiener filter setting of 0.002. Reconstructed image-stacks were 3D rendered and visualized using SOFTWORKS and IMAGEJ software. The fluorescence intensity along the Z-rings was measured using the profile function in IMAGEJ software.

Chromatin immunoprecipitation and bioinformatics analysis
Chromatin immunoprecipitation assays followed by sequencing (ChIP-seq) were performed as previously described by Al-Bassam et al. [39] with minor changes. Each strain was cultured in MYM medium supplemented with trace elements for an appropriate time (14 and 20 h for the wild-type, DparA, DparB::apra and DparAB strains; 30 h for the TopA depletion mutant). After cross-linking (formaldehyde added to a final concentration of 1% (v/v)) and lysis, the samples were sonicated for six to seven cycles (20 s each) to shear the chromosome into fragments ranging from 200 to 600 bp. The samples were centrifuged and pre-cleaned with Protein A Sepharose (Sigma). Then, 2 mg of purified polyclonal anti-ParB antibodies was added to each cell lysate, and the mixtures were incubated on a rotating wheel at 48C overnight. Protein A Sepharose was added to cell lysates and incubated for the next 4 h. The samples were centrifuged, and the pellets were washed with IP buffer (50 mM Tris -HCl pH 8.0, 250 mM NaCl, 0.80% Triton, protease inhibitors) and eluted by overnight incubation at 658C with IP elution buffer (50 mM Tris -HCl pH 7.6, 10 mM EDTA, 1% SDS). After centrifugation, the pellets were re-extracted with TE buffer (50 mM Tris -HCl pH 8.0, 10 mM EDTA), incubated with proteinase K (Roche) for 1.5 h at 558C and extracted twice with phenol and once with chloroform. Finally, samples were purified with QiaQuick columns (Qiagen). For each immunoprecipitation sample, we used its corresponding input control of total DNA. DNA concentration rsob.royalsocietypublishing.org Open Biol. 6: 150263 was quantified using a NanoDrop spectrophotometer (ThermoScientific). Libraries were constructed and sequenced by the Karlsruher Institut fü r Technologie (KIT). Experiments were performed in duplicate. ChiP-seq sequence data were aligned (mapped) to the S. venezuelae chromosome using two aligners, Bowtie 1.1.1 and Novoalign 3.02.08 (Novocraft Technologies Sdn Bhd, www.novocraft.com) [40,41]. With Bowtie, we selected options reporting only alignments in the best alignment 'stratum', that is, having the least number of mismatches, which was assumed to be one up to three. Using Novoalign, we applied a method reporting all alignment locations. SAMtools 1.1 and BEDTools 2.17 were used to manipulate the obtained files and convert formats [42]. To find peaks, that is, regions with significant numbers of mapped reads, we used Peak Finder MetaServer 1.3 (PFMS), which collects results from several peak finders and produces consensus peaks [43]. We applied three recommended peak finders: MACS 1.4.2 [44], CisGenome 2.0 [45] and SISSRs 1.4 [46]. Finally, Integrative Genomics Viewer (IGV) 2.3.40 [47] was used to visualize data.
To search Streptomyces genomes for potential parS sequences, we constructed a position weight matrix (PWM) based on sequences (GTTTCACCTGAAAC) identified experimentally in S. coelicolor by Jakimowicz et al. [26]. Then, the PWM was used by the Motif Occurrence Detection Suite (MOODS), 1.0.2 [48] to find corresponding sequences in the Streptomyces chromosomes, which were downloaded from the GenBank database under accession numbers NC_003888 (S. coelicolor) and NC_018750 (S. venezuelae). The DNA walk method was used to visualize the A þ T content along the chromosome [49].

Sporogenic hyphal growth stops at the time of Z-rings formation and ParA accumulation along the hyphae
Previous reports have indicated that the extension of the sporogenic hyphae stops before the commencement of further stages of development, that is, chromosome condensation, segregation and hyphae septation [50]. The switch from the extension of aerial hyphae to the formation of spores was suggested to be mediated by ParA [11]. To verify this hypothesis, we used the novel model organism, S. venezuelae, in which sporulation is not limited to the interface between solid medium and air [34]. Therefore, we could use single cell time-lapse microscopy to monitor the whole sporulation process in detail. The S. venezuelae merodiploid reporter strain expressing ftsZ-ypet (from the native promoter) as a marker of Z-rings (MD100) was used to monitor the timing of sporulation septation in relation to the extension of the sporogenic aerial hyphae. Growth of this strain was similar to the growth of the wild-type parent strain (electronic supplementary material, figure S1), although its sporulation was slightly delayed. Images of the S. venezuelae sporulating hyphae were captured every 15 min for 16-20 h. Selected hyphae, which underwent the whole sporulation process and could be observed from branch emergence to the separation of individual spores, were analysed in subsequent images. These analyses enabled us to measure the parameters describing the growth and differentiation of sporogenic hyphae, such as the duration of hyphal extension (calculated as time from branch emergence to the cessation of its growth), the final length of sporulating hyphae and the hyphal tip extension rate. The appearance of Z-ring ladders was regarded as the initiation of septation, while the time period during which regular Z-rings were clearly detectable was considered the septation time. Additionally, we calculated the spore maturation time and the overall sporulation and differentiation time (figure 1a, tables 1 and 2).
Sporulation starts with the extension of sporogenic hyphae, which subsequently undergo synchronized division. The sporogenic branches of the strain MD100 ( ftsZ-ypet) extended for 299 + 61 min (mean + s.d.). The final length of the hyphae was 36 + 13 mm (table 1, see also figure 2a). These measurements allowed us to calculate average hyphal extension rate (table 1), which was 7.6 + 2.2 mm h 21 . The extension rate was constant during the growth of the aerial hyphae as shown by the uniform increase of hyphal length at subsequent time points. Only at the beginning of aerial hyphal growth and close to growth cessation was the rate slightly lower than the growth stage. The increase in FtsZ-YPet fluorescence was detected in the MD100 sporogenic hyphae 42 + 7 min prior to growth cessation (figure 1b,c; electronic supplementary material, movie S1). At first, FtsZ-YPet fluorescence was observed as dispersed fluorescence along the hyphae. The Z-rings became clearly visible at growth cessation (approx. 33 min after the increase in FtsZ-Ypet fluorescence) and remained noticeable for 84 + 26 min. The position of the rings did not change, but their fluorescence increased slightly over 29 + 17 min to reach the maximal level (more than three times higher than the initial fluorescence) and remained constant for the next 55 + 16 min (figure 1c). Approximately 63 + 37 min after the disappearance of the FtsZ rings, the spores rounded and separated (figure 1a,b, table 2; electronic supplementary material, movie S1). The total time of differentiation from the emergence of sporogenic hypha to spore separation was approximately 7.5 h (449 + 82 min), of which 5 h were required for hyphae to reach the full length. The total time of sporulation measured from the cessation of hyphal growth to spore separation was 2.5 h, of which septation lasted approximately 1.5 h, followed by 1 h spore maturation. The timing of FtsZ ring formation in relation to the arrest of tip extension showed very little variation (table 2), indicating the precise coordination between those processes.
To determine the timing of chromosome segregation, we attempted to construct fluorescent protein fusions of the segregation proteins, ParA and ParB. In aerial hyphae, ParA extends along the hyphae to accompany ParB complexes only during chromosome segregation. The attempts to construct functional ParB-EGFP fusions were not successful, and the constructed strain had a chromosome segregation defect similar to the DparB strain. The strain expressing parA-egfp from the native chromosomal locus (MD004) exhibited slightly impaired chromosome segregation (5% anucleate spores in MD004 compared with 10% in the DparA strain and 0.4% in the wild-type strain) but septation disturbances, which are associated with parA deletion, were not detected in MD004 (electronic supplementary material, figure S1b). The disturbances in segregation may be due to the fusion of EGFP to ParA resulting in reduced protein dynamics/stability and/or the slightly lower ParB level in the MD004 strain in comparison to the rsob.royalsocietypublishing.org Open Biol. 6: 150263 wild-type strain (electronic supplementary material, figure S1a). The ParA-EGFP fluorescence was visible in MD004 as a diffuse bright signal extending from the tip along the hyphae (figure 1b(ii); electronic supplementary material, movie S2). The same localization was observed when immunofluorescence was used for ParA localization (electronic supplementary material, figure S2a). Interestingly, the tipassociated ParA complexes that were previously described in S. coelicolor [11,30] were not detected in S. venezuelae. The ParA fluorescence along the hyphae was detected 54 + 14 min before the cessation of hyphal growth in all analysed hyphae, and its signal increased for the next 164 + 5 min and then remained constant (figure 1b(ii), figure 1d; electronic supplementary material, movie S2). The fluorescence exceeded the initial intensity by approximately three times at growth cessation and at spore separation reached a level approximately eight times higher than the initial intensity (figure 1d). These results suggest that ParA starts to accumulate almost 1 h before growth cessation. Its increasing fluorescence was visible for approximately 2.5 h until spore separation. rsob.royalsocietypublishing.org Open Biol. 6: 150263 In summary, our time-lapse analyses allowed determination of the growth parameters of developing S. venezuelae sporogenic hyphae. We showed that growth cessation is concurrent with Z-ring formation and the onset of chromosome segregation, which is marked by the accumulation of ParA proteins.

Segregation proteins affect the extension rate of sporogenic hyphae
A time-lapse analysis of the S. venezuelae sporulating hyphae confirmed that ParA appearance is associated with the cessation of hyphal growth. Our previous studies of S. coelicolor indicated that ParA affected hyphal extension [11,32]. To determine if the sporulation proteins influence the growth of S. venezuelae sporogenic hyphae, we analysed parA, parB and parAB deletion strains expressing ftsZ-ypet (MD111, MD021 and MD031, respectively). As observed previously in S. coelicolor [27,30], parA, parB and parAB deletions in S. venezuelae also resulted in chromosome missegregation and irregular septation during sporulation (restored in complemented strains, electronic supplementary material, figure S1b), although the disturbances were less pronounced than in S. coelicolor. The elimination of ParA in S. venezuelae led to an increased rate of tip extension in the sporogenic hyphae to approximately 160% of the wild-type rate (figure 2a). A time-lapse analysis of the parA deletion strain (MD111) growth revealed that its growth rate was 12.0 + 3.5 mm h 21 (7.6 + 2.2 mm h 21 in the wild-type, the difference was statistically significant with p , 0.0001; figure 2b, table 1). The time period of sporogenic hyphae extension was only marginally longer in the DparA strain than in the wild-type (335 + 85 min versus 299 + 61 min; table 1, figure 2c). As a result of the higher growth extension rate and the somewhat longer extension period, the length of the sporogenic hyphae of the DparA strain at growth cessation was approximately 170% of wild-type hyphae length (61 + 16 mm and 36 + 13 mm, respectively, significant with p , 0.0001, figure 2d, table 1). Complementation of parA deletion Table 1. Basic parameters describing sporogenic hyphal growth in the studied S. venezuelae strains. The data are based on 35 hyphae from at least four independent experiments. The mean + s.d. and the median (in parentheses) are shown; asterisks indicate statistically significant results in comparison with wild-type (*p , 0.1, **p , 0.01, ***p , 0.001). From growth cessation to spore separation. c From emergence of sporogenic hyphae to spore separation. rsob.royalsocietypublishing.org Open Biol. 6: 150263 restored the hyphal length (electronic supplementary material, figure S1b). Thus, our observations revealed that parA deletion significantly increases the length of the hyphae and the extension rate.
Interestingly, in contrast with the DparA results, the DparB strain (MD021) exhibited a hyphal extension rate decreased to approximately 70% of the wild-type strain rate (5.5 + 2.1 mm h 21 , p , 0.01) and approximately 50% of the extension rate of the parA deletion strain ( p , 0.0001; figure 2a,b). Although the time period of hyphal extension in the DparB strain increased to 120% of the wild-type (367 + 88 min, p ¼ 0.011), the final length of MD021 hyphae did not differ significantly from the wild-type hyphae length (31 + 11 mm, in comparison with 36 + 13 mm, respectively; figure 2c,d, table 1). Thus, the effect of parB deletion on the hyphal extension rate is the opposite of the parA deletion.
If the effect of ParB on the extension rate results from the effect of ParB on ParA assembly, a double parAB deletion strain should exhibit similar characteristics to the DparA strain. However, the hyphal extension rate of the DparAB (MD031) strain was still 80% (6.3 + 2.1 mm h 21 , figure 2b, table 1) of the wild-type, which was similar to the DparB strain and almost two times lower than the DparA hyphal extension rate (only the latter difference was statistically significant; p , 0.0001). Additionally, the hyphal extension time was similar to the wild-type strain (319 + 115 min, figure 2c, table 1), while the hyphal length was, on average, the same as the DparB strain ( figure 2d, table 1). The analysis of the double deletion mutant suggests the increased extension rate caused by ParA elimination is dependent on ParB. Moreover, ParB elimination inhibited the growth rate in both the parB and parAB deletion strains, which suggests a ParA-independent mechanism.
In sum, our results revealed that ParA and ParB affect both the extension rate of the sporogenic hyphae and the period of hyphal extension. We also found that the two proteins have opposite effects on hyphal extension.

ParA induces the formation of massive ParB complexes during sporulation
Earlier studies of S. coelicolor demonstrated that regularly distributed ParB complexes are formed when ParA spreads along the sporulating hyphae [30], which is supported by our time-lapse microscopy studies of S. venezuelae. These results suggest that ParA accumulation along the hyphae promotes the initiation of chromosome segregation. Because we were not able to analyse the formation of ParB complexes during sporulation by time-lapse microscopy, to study their formation and dependence on ParA, we carried out ChIP-seq analysis of the wild-type and DparA strains (the DparB strain was used as a negative control). To determine the appropriate time points for the ChIP-seq analysis, we ana- Analysis of the binding profile confirmed that ParB bound the central region of the chromosome near the oriC ( figure 3). Almost all identified peaks corresponded to the parS sites (results based on reads mapped by two aligners, Bowtie and Novoalign, were very similar). Binding was also detected between the parS sites, which suggests the formation of large nucleoprotein ParB complexes encompassing approximately 40 kb proximal to oriC (electronic supplementary material, figure S4c). No binding was detected in the negative control (DparB mutant, MD002). Comparison of ParB binding profiles suggests more efficient binding in the 14th hour of culture than in the 20th hour of culture (the efficiency of immunoprecipitation, e.g. formaldehyde cross-linking should not be affected by the time of culture [51]). This was reflected by the higher number of peaks identified (78 and 73 for Bowtie and Novoalign-mapped reads in the 14th hour, respectively, and 37 and 33, respectively, in the 20th hour; figure 3). In the DparA strain, the binding profile suggests less efficient ParB binding than in the wildtype strain, which was confirmed by the smaller number of peaks identified ( figure 3). Additionally, the binding profile (and the number of identified peaks) in the 14th and 20th hour is very similar. It suggests that in the absence of ParA, ParB binding is diminished, is less extensive and is not enhanced at a particular time in development (14th hour) as in the wild-type strain.
In summary, ChIP-seq analysis confirmed the formation of large nucleoprotein ParB complexes encompassing parS sites and the regions between the binding sites at early sporulation time points. Additionally, the analysis showed that ParA is required for the formation of fully assembled ParB complexes at the early stage of sporulation, reinforcing the hypothesis that ParA accumulation facilitates formation of segrosomes.

Aberration of DNA segregation affects formation of Z-rings
The primary role of the ParB protein in Streptomyces is the formation of nucleoprotein complexes in sporulating hyphae, whereas ParA is required for their regular distribution along the hyphae and for their efficient assembly [30], which was supported by the ChIP-seq experiments.
Although electron microscopy studies suggested that in S. coelicolor, septa may be formed over non-segregated DNA [24], our previous observations of minicompartments (delimited by closely spaced septa), which are particularly abundant in DparA and also present in the DparB strain [30], indicated that the chromosome organization affects the positioning of Z-rings. Here, using time-lapse microscopy, we determined whether chromosome missegregation  rsob.royalsocietypublishing.org Open Biol. 6: 150263 resulting from the deletion of parA and/or parB affects the timing and duration of Z-ring formation.
In the hyphae of DparB and DparAB merodiploid strains expressing ftsZ-ypet (MD021 and MD031), Z-rings formed significantly earlier than in the wild-type strain, on average 28 + 15 min and 39 + 12 min (mean + s.d.) prior to growth cessation, respectively, while in the wild-type, the Z-rings were visible exactly at the time of growth cessation ( p , 0.00001; figure 4a-c,e, table 2; electronic supplementary material, movie S3). Surprisingly, in the DparA strain (MD011), Z-rings were visible later, only 15 + 12 min after growth cessation (figure 4a-c,e, table 2; electronic supplementary material, movie S4). The level of FtsZ protein was similar in all analysed strains (electronic supplementary material, figure S3), showing that the differences in timing of the Z-ring formation are not clearly related to ftsZ overexpression but rather result from an accelerated assembly of the Z-rings. The earlier appearance of Z-rings was associated with their significantly longer 'lifetime' in the parB and parAB deletion strains compared with the wild-type (117 + 21 min in DparB, 135 + 28 min in DparAB and 84 + 26 min in wild-type strain, p , 0.00004; table 2, figure 4d). In DparA, the overall lifetime of Z-rings was similar to the wild-type (79 versus 84 min, respectively; figure 4d). No significant differences were observed between the wild-type and deletion strains in the spore maturation time, and the total time of sporulation measured from growth cessation to spore separation was similar in all analysed strains (table 2). This indicates that ParA and ParB do not influence the duration of the later stages of the Streptomyces cell cycle, but they visibly affect the timing of Z-ring formation.
In contrast with the timing of Z-ring formation, the timing of FtsZ fluorescence appearance was not significantly different in the hyphae of all analysed deletion strains and the wild-type (figure 4b,e, table 2). In all studied strains, the intensity of Z-ring fluorescence increased over time (figure 4b). A similar increase in overall FtsZ-YPet fluorescence was observed in the liquid culture time course (electronic supplementary material, figure S3). The deletion strains showed somewhat higher intensity of FtsZ-YPet fluorescence along the hyphae and higher intensity of Z-ring fluorescence (figure 4b). Notably, in the DparB strain, Z-rings along the hyphae exhibited a variation of intensity that was three times greater than in the wild strain ( p , 0.002) at the time of growth cessation (figure 5b).
In the Streptomyces hyphae, chromosomes condense and segregate during formation of Z-rings [24,30]. Because the Z-rings form earlier in the DparB strain, prior to growth cessation, we determined if the earlier formation of Z-rings precedes DNA segregation. Indeed, in contrast with the wild-type strain where Z-rings were detected in the hyphae with condensed nucleoids, in the DparB strain, approximately 10% of hyphae with Z-rings still contained unsegregated and uncondensed nucleoids (figure 5a).
We next investigated whether the accelerated formation of Z-rings in the hyphae with non-segregated chromosomes in the DparB mutant was associated with any aberrations in their Z-ring structure. To compare the Z-ring structure in the wild-type and DparB strains, high-resolution structured illumination microscopy was used. In the wild-type strain, Z-rings had a regular shape and uniform intensity of fluorescence, but in the DparB strain, they were often broken or included local accumulation of FtsZ, which affected the regular shape of the structures (figure 5c). Measurements of fluorescence intensity along the circumference of the rings in 10 randomly selected hyphae supported this observation (figure 5d ). In the parB deletion strain, the variation of the Z-ring fluorescence intensity along the Z-rings was two times higher than in the wild-type strain ( proved with the F-test for the intensities scaled up to similar values, with the ratio of variances 1.18 + 0.089; p ¼ 0.000037). Moreover, approximately 20% of the analysed rings were broken (noncontinuous) compared with 1%-2% in the wild-type strain (figure 5c). These observations suggest that the elimination of ParB promotes earlier polymerization of FtsZ, but the ring assembly is not uniform.
Overall, our analyses revealed that ParB elimination promotes earlier formation of Z-rings, but these rings are less uniform than those formed in the wild-type. In the absence of ParA, the Z-rings are delayed. This may be associated with reduced ParB complex formation in the absence of ParA. Our observations suggest that ParB complexes and/or DNA organization directly or indirectly affect the assembly of Z-rings.

Increased chromosome compaction induces Z-ring formation
We previously observed that increased chromosome supercoiling induced by TopA depletion in S. coelicolor led to the inhibition of chromosome segregation and septation, while a moderate increase in chromosome compaction affected the positioning of the septa [31]. Here, our S. venezuelae studies showed that chromosome segregation affects the timing of sporulation septation. Thus, we tested whether the altered chromosome supercoiling resulting from decreased TopA disturbed the formation of Z-rings. The S. venezuelae TopAdepleted strain (MD013, in which the topA gene is under the control of the inducible p tet promoter) exhibited growth inhibition and sporulation blockage when cultured without an inducer, as previously described for the S. coelicolor strain [31]. Time-lapse microscopy of an MD013 derivative expressing the ftsZ-ypet gene (AKMD2 strain) confirmed that at the lowest level of TopA (no inducer added), the formation of Z-rings was completely abolished. Surprisingly, partial depletion of TopA ( protein level eight times lower than the wild-type level, electronic supplementary material, figure S4a) led to accelerated formation of the Z-rings, which were already detectable in the extending hyphae (figure 6a). The Z-rings were visible more than 3 h before growth cessation, and their duration was significantly increased compared with the wild-type strain (205 + 52 min and 84 + 26 min, respectively, table 2). Interestingly, the time between the arrest of hyphal extension and spore separation was shortened to only about one quarter of the wild-type sporulation time (table 2). Additionally, Z-rings at growth cessation were very irregularly spaced, and they formed between the irregularly spaced regions of condensed DNA (figure 6b,c). Our observations suggest that the increased chromosome condensation in TopA-depleted strains promotes Z-ring formation in chromosome-free regions.
To summarize, the analyses of the strain with the modified topoisomerase levels suggest that chromosome organization directly influences the FtsZ polymerization in rsob.royalsocietypublishing.org Open Biol. 6: 150263 the ring structure. Our data also suggest that the increased chromosome compaction promotes Z-ring formation.

Discussion
In various bacterial species, ParAB proteins synchronize chromosome segregation with other cell-cycle events, such as cell division, cell elongation or chromosome replication [9,12,52]. Thus, the disruption of the parAB genes results in a plethora of phenotypes ranging from disturbances in chromosome segregation to variation of the cell length to chromosome over-replication and inhibition of cell division.
Here, we determined how the ParAB proteins coordinate the cell-cycle events during Streptomyces sporulation, when tens of chromosomes are distributed evenly along every sporogenic cell. The cessation of hyphal growth coincides with the formation of Z-rings and the accumulation of ParA. The time of induction of parA and ftsZ expression suggests their involvement in a developmental checkpoint between hyphae extension and growth cessation. This type of checkpoint was suggested earlier by the analysis of S. coelicolor developmental mutants with long and unseptated hyphae, whiA and whiB [50]. Previously, both the parAB p2 and ftsZ p2 promoters were shown to be developmentally regulated by the WhiA and WhiB proteins [28,29]. Our time-lapse analysis confirmed that the initiation of septation and chromosome segregation is synchronized with hyphal growth cessation.
Elimination of ParA increased the rate of hyphal growth approximately 1.5 times and led to increased hyphae length at growth cessation and longer spore chains. Elongated spore chains were observed before in S. coelicolor, and this phenomenon was explained by the interaction of ParA with Scy, a tip-localized component of the polarisome complex (or TIPOC, tip organizing complex) responsible for hyphae elongation [11,20,32,53]. According to our model, cell extension is promoted when the balance between ParA and Scy is shifted towards Scy assembly. The induction of parA expression in sporulating hyphae is associated with the disassembly of Scy at the tip, which triggers the cessation of hyphal elongation and commencement of sporulation [11]. Interestingly, ParA signals in S. venezuelae could be detected only as diffuse fluorescence along the hyphae, but tip localization was not clearly detectable, which could possibly be due to a different experimental set-up or a weak transient ParA signal. The strict coordination of the ParA induction and growth cessation observed in S. venezuelae suggests that in this species, ParA is also engaged in the developmental checkpoint.
Surprisingly, time-lapse analyses showed that elimination of ParB decreases the hyphal growth rate. However, the hyphal extension time of the parB deletion strain was increased, allowing the hyphae to reach a length approximately similar to the wild-type hyphae. This suggests the presence of a checkpoint mechanism that is based on control of the hyphal length rather than the extension time. Such observation is reminiscent of recently reported constant size extension which was suggested to maintain size homeostasis in E. coli and C. crescentus [54]. The opposite effect of parB deletion and parA deletion on hyphal extension could be explained by the modulation of ParA ATPase activity, its nucleotide exchange or ability to interact with Scy by ParB. If that was the only mechanism accounting for the observed changes in growth rate, the hyphal extension of the double parAB deletion strain should be similar to the parA deletion strain. Interestingly, the double parAB mutant exhibited a decreased rate of hyphal extension, which suggests the involvement of ParB in an additional, ParA-independent mechanism(s) influencing rsob.royalsocietypublishing.org Open Biol. 6: 150263 growth. One possible mechanism could be the influence of ParB on chromosome replication or transcription of the genes in the region involved in complex formation, which could affect the rate of hyphae growth. The other possibility is the engagement of ParB with the polarisome (TIPOC) complex. A direct interaction between ParB and DivIVA has been suggested for Corynebacterium glutamicum, which also belongs to the Actinobacteria phylum and exhibits a polar mode of growth [55]. However, the relatively far distance from the ParB complex to the tip in Streptomyces hyphae (1.4 mm) challenges the hypothesis of direct interaction. Interestingly, in C. glutamicum, the reduced rate of polar cell elongation was caused by parA deletion [56]. It is difficult to account for the contrasting effects of parA and parB deletion in Streptomyces and Corynebacterium, but presumably, during the transition from the rod cell polar extension to hyphal growth, a new mechanism(s) controlling cell elongation evolved. Although at present we cannot provide an explanation for the observed ParB influence on the hyphal extension, our studies revealed that the hyphal elongation rate in Streptomyces is controlled by a not yet fully identified mechanism that involves both segregation proteins. In S. coelicolor, ParA spreads from the tip complex to an elongated structure and was suggested to trigger formation of regularly spaced ParB complexes along the hyphae. We showed that ParB-EGFP foci in the sporulating hyphae of the parA deletion strain are irregular compared with the wild-type and showed much weaker fluorescence [30]. The ChIP-seq analysis confirmed that ParB complexes encompassed large chromosomal fragments. It was suggested that ParB homologues are able to trap DNA loops by DNA bridging or binding DNA near parS over long distances [57,58]. Our findings are consistent with the observation made in B. subtilis that the ParB complex engages a chromosomal region of several kilobases in length. The presence of ParA stimulates the formation of large nucleoprotein complexes in S. venezuelae and initiates chromosome segregation.
The time-lapse analysis revealed that ParAB proteins affect the formation of Z-rings. The earlier Z-ring formation in the parB and parAB mutant strains could not be directly related to the earlier increase in FtsZ levels. However, in the absence of ParB, the Z-rings had very irregular shapes and uneven fluorescence intensity. This suggests that either chromosome organization or ParB itself affects the assembly of the Z-rings. The effect of parAB deletion on the septation has been noted before, as has the uneven spacing between the septa in the parA and parB mutant strains [30]. Interestingly, the disturbances in chromosome segregation and septation observed in S. venezuelae parAB deletion strains were less severe than those described earlier for S. coelicolor [30]. This may result from either differences in growth characteristics between the species or the presence of plasmid-encoded segregation proteins (S. venezuelae genome contains plasmids with parAB homologues, but nothing is known about their expression) that complement the parAB deletion as observed earlier in the case of the S. coelicolor SCP1 plasmid [59]. In contrast with DparB, parA deletion delayed rsob.royalsocietypublishing.org Open Biol. 6: 150263 initiation of septation. In the DparA strain, ParB still formed complexes, although not efficiently, which remained irregularly spaced. The observed delay in Z-ring assembly in the DparA strain suggests the direct involvement of ParB complexes in control septation. Notably, the inhibitory effect of unsegregated chromosomes and ParB complexes on the septation was also suggested by our studies of a topoisomerase I (TopA)-depleted strain [31]. ChIP-seq analyses confirmed ParB binding to most of the parS sites in the TopA-depleted strain, although the binding profile differed from the one observed in the wild-type strain (electronic supplementary material, figure S4b). The other explanation for early Z-ring formation in the TopA depletion strain, which cannot be excluded, is that increased chromosome supercoiling induces transcription of some regulators that control septation. Our studies of S. venezuelae showed that moderately increased chromosome supercoiling promotes Z-ring assembly even in extending hyphae. We observed that the intensity and spacing of the Z-rings in the TopA-depleted strain at the time of growth cessation were very irregular. This also suggests that their position may be determined by irregular condensation of the chromosome.
Our observations suggest the involvement of the ParB protein in negative control of Z-ring formation. In fact, in some bacterial species, the ParAB proteins are involved in septum placement through interaction with regulators of FtsZ polymerization, such as MipZ in C. crescentus [12] or PomZ in Myxococcus xanthus [60]. MipZ (a ParA family protein), through interaction with ParB, forms a gradient extending from the pole and inhibits the formation of Z-rings close to the pole. Contrary to MipZ, PomZ (another ParA family protein) was suggested to positively regulate division site selection by recruitment of FtsZ. Interestingly, in B. subtilis, Noc is a ParB homologue that interacts with the chromosome and cell membrane [61]. In Streptomyces, neither potential FtsZ regulators interacting with ParB nor other ParB homologues that could act as NO proteins were identified [21]. To date, the only described mechanism controlling septa placement is the positive control by Ssg proteins [25]. SsgA and SsgB were detected prior to Z-ring formation at the aerial hyphae membrane, and SsgB is speculated to tether the FtsZ filaments [62]. Because an ssgB mutant strain also showed impaired growth cessation [63], it is plausible that the ParAB proteins are involved in the interaction with Ssg proteins. Interestingly, the increased variation of septa placement in parAB deletion mutants has been observed before in other bacteria, such as M. smegmatis or C. glutamicum [10,56]. For those species, it was also suggested that the nucleoid itself or the interaction with an unknown factor determined the positioning of the Z-rings [18]. Consistent with this idea, we suggest that Streptomyces, in the absence of nucleoid occlusion proteins (Noc or SlmA), developed another negative control mechanism to spatially and temporally regulate septum formation.
In summary, we demonstrated that the function of ParAB in Streptomyces extends from chromosome segregation to coordination of this process with cell elongation and cell division (figure 7). During Streptomyces sporulation, hyphae elongation ends before chromosome segregation and septation. The unique Streptomyces spatio-temporal separation of cell elongation and division imposes special requirements for synchronization checkpoints on both processes. Our data show that ParAB proteins are involved in these checkpoints and coordinate the transition from hyphal growth to segregation of tens of chromosomes associated with septation. The formation of ParB complexes determines the timing of septation. Moreover, we demonstrate that the absence of ParB affects the assembly and distribution of Z-rings. Thus, our analysis shows a close link between Z-rings and ParB complexes.