Coordination of chromosome segregation and cell division in the archaeon Sulfolobus acidocaldarius Authors: Rachel Y Samson, Naomichi Takemata, Stephen D Bell Published: 2025 ABSTRACT AbstractDespite its fundamental importance for the propagation of all cellular life, very little is known about the molecular basis of chromosome segregation in archaea. This is a particularly intriguing issue in members of the genus Sulfolobus which encode homologs of a “bacterial-like” candidate segregation machinery, termed SegAB, related to the bacterial ParA-based systems, yet rely on a “eukaryotic-like” ESCRT machinery for cell division. While investigating the temporal dynamics of Sulfolobus chromosome architecture, we uncovered conformational changes at the candidate segregation locus. Loss of the Seg system results in altered nucleoid morphology and induction of a DNA damage response. Significantly, loss of the Seg system also leads to precocious assembly of the key cell division protein CdvA. Finally, we observe interactions between SegA and CdvA, indicating dual roles for the Seg system in effecting chromosome segregation and defining the positioning of the cell division apparatus.Subject terms: Chromosome segregation, Archaea, Cellular microbiology Introduction The accurate replication of chromosomes and their subsequent segregation to daughter cells are essential for the propagation of all cellular life on the planet. In bacteria, these processes are temporally linked1,2, with the combined actions of partitioning (par) systems and Structural Maintenance of Chromosomes (SMC)—superfamily members playing pivotal roles in many, but not all, species. In contrast, eukaryotic cells generally have defined gap phases with G2 providing a temporal spacer between the completion of DNA replication and the initiation of chromosome segregation and cell division in mitosis. Interestingly, members of the crenarchaeal kingdom of Archaea have long been known to have a more “eukaryotic-like” cell cycle, with defined gap phases separating DNA synthesis and cell division3,4. Furthermore, cell division in the genus Sulfolobus is dependent on homologs of the eukaryotic ESCRT machinery5,6. The available data indicate that the initiating event in cell division is the formation of a ring of the Sulfolobus-specific factor, CdvA7. CdvA then recruits a specific ESCRT-III protein. The subsequent recruitment of further ESCRT-III paralogs leads to membrane ingression in a process dependent on the Vps4 ATPase and the activity of the proteasome6,8. We have shown previously that the CdvA ring is detectable in cells in which no discernable segregation of the chromosomes has occurred. Notably, embellishment of the CdvA ring with ESCRT-III was only observed in cells with segregated nucleoids, suggesting a functional interplay between segregation and cell division machineries7. However, the process of chromosome segregation into the two eventual daughter cells remains poorly understood. Sulfolobus homologs of the bacterial chromosome partitioning “par” system have been identified and termed SegA and SegB9. SegA is a homolog of the ParA ATPase, a member of the Walker-class ATPases, and SegB is a ribbon-helix-helix—containing DNA-binding protein10. SegB is related to bacterial Par-system-associated proteins, such as the plasmid-encoded ParG. Accordingly, this system has been proposed to be the Sulfolobus chromosome segregation machinery. However, while extensive biochemical and structural characterizations of these proteins have been performed, little is known about their in vivo role(s)9,10. The ParA-like proteins form a broad family, representatives of which generally appear to play roles in determining positioning signals in bacterial cells1,11,12. For example, the well-characterized Min system in E. coli regulates cell division by inhibiting the ability of the key cell-division protein, FtsZ, to polymerize at regions other than mid-cell12. In Bacillus subtilis and other bacteria, ParB proteins mediate loading of the bacterial SMC-containing condensin complex onto DNA at parS sites13,14. These sites are usually origin-proximal, and the ultimate action of SMC is to bring about the processive zipping-up of chromosome arms during the linked processes of replication and segregation15–17. In eukaryotes, the actions of SMC-superfamily proteins also play key roles in shaping chromosomes, with condensins playing pivotal roles in effecting major structural changes as chromosomes process through mitosis18. Intriguingly, in Sulfolobus, a very distinctive chromosome architecture is observed in asynchronous cell populations, with a compartmentalized organization, reminiscent of that seen in metazoan organisms19. The compartments, termed “A” and “B”, are most defined during exponential growth and are marked by higher and lower levels of transcription, respectively. The Sulfolobus-specific SMC-superfamily protein, coalescin (ClsN), plays a major role in establishing and maintaining the integrity of the B compartment19. However, the extent to which chromosome conformation is modulated during the Sulfolobus cell cycle is unresolved. In the current work, we reveal that the compartmentalization of the Sulfolobus chromosome is a prevalent feature throughout the cell cycle. However, the structure at the seg locus is modulated, with a transition from an isolated structural element to absorption of the locus into a broader domain of the chromosome as cells progress from division into S-phase. This insulated feature correlates with an extensive region of high occupancy of the SegB protein and is absent in mutant cells lacking the Seg system. Loss of SegAB results in the altered distribution of nucleoid morphologies within the population and a loss of culture viability. Intriguingly, chromosome compartmentalization and ClsN positioning are essentially unaffected by loss of the SegAB system. However, loss of the Seg system triggers a DNA damage response. This correlates with the presence of a sub-population of cells with less than a single copy of the chromosome, suggesting inaccurate cell division and/or DNA degradation. Intriguingly, we observe dramatically elevated levels of cells with CdvA structures in cells lacking the seg system, and we reveal that SegA and CdvA functionally interact, with SegA solubilizing CdvA in a heterologous system. Taken together, our data implicate the Seg system in playing dual roles in chromosome segregation and cell division. First, it functions like canonical bacterial ParA systems by facilitating chromosome segregation, and second, it acts in a manner reminiscent of the E. coli Min system by defining the mid-cell region and modulating polymerization of key cell division proteins. Results Architectural changes at the putative Sulfolobus partitioning locus To measure changes in chromosome conformation during the Sulfolobus cell cycle, we initially performed Hi-C studies on cultures synchronized using a baby machine20. This enriches newborn G1 cells. These cells are accumulated at 4 °C, then released into growth by rapid elevation of temperature. Samples were taken at 0 min (G1-enriched), 40 min (S-phase enriched), 100 min (G2-enriched) and 180 min (dividing cell-enriched). The results shown in Fig. 1a reveal the plaid-pattern indicative of compartmentalization at all time points assayed. However, principle component analysis reveals a region of B-compartment in G1-enriched cells, with boundaries at approximately ~160 kb and 500 kb (Fig. 1b). This region of B-compartment has diminished by the S-phase enriched 40-min timepoint and vanished by 100 min, when the majority of cells are in G2. However, by 180 min, coincident with a subset of cells undergoing cell division in the population (Fig. 1f), this region of the B-compartment has reappeared. Note that while the baby machine gives strong enrichment of a G1 cell population, which mostly progresses synchronously through S-phase, the exit from G2 is quite heterogeneous, with the dividing cell population spread out over a broad temporal window. Our previous work has shown that this window for G2 exit is about 1 h in length20. Conformational changes in the Sulfolobus acidocaldarius chromosome during the cell cycle. a Hi-C contact matrices representing interaction frequencies from S. acidocaldarius cells that were synchronized by elution from a baby machine. Samples were collected prior to resumption of growth (G1-enriched) at 78 °C or following growth for 40 min (S-phase enriched), 100 min (G2-enriched) or 180 min (enriched in dividing cells). b Principal Component (PC) analysis of the data shown in (a). The first principal component is shown as a compartment index. c Difference heat map of the Hi-C matrices comparing interaction scores at the indicated time points. d Difference heat map of the Hi-C matrices comparing interaction scores at the 40 and zero minute time points (left) and 180:100 min (right), zoomed in and recentered to show the alterations in the contact pattern emanating from the segAB locus. e Schematic showing the alteration of contacts at the segAB locus as cells transition from G1 to S phase. Illustration created by Steve Bell, The Ohio State University, using Illustrator software. f Quantitative RT-PCR was performed on RNA isolated from baby machine synchronized cultures. Primer pairs for the segA gene transcript and that for the transcription elongation factor nusA were utilized. The experiment was performed on three biological replicates, and data are represented as mean values ± error bars derived from the standard deviation. Gray and black dots represent the individual data points for the segA and nusA reactions, respectively. Flow cytometry profiles of the samples are shown below. g Analysis of loops detected in previously published high-resolution 3C-Seq analyses of the S. acidocaldarius chromosome21. Loops anchored on one side at the segAB locus are shown in purple, and all the other loops are in red. A difference heat map comparing contact matrices in cells collected at 0 and 40 min (Fig. 1c, left panel) reveals a number of changes in local domain organization and elevation of contacts between the domains containing oriC1 and oriC3. By contrast, more modest changes in chromosome conformation are observed between the S-phase enriched cells harvested at 40 min and the G2 population at 100 min (Fig. 1c, middle panel). In addition, a red plume-like feature is visible, emanating from the vicinity of genome coordinate 160 kilobases as cells transition from G1 into S phase (Fig. 1c, left panel and zoomed in and recentered in Fig. 1d, left panel). Significantly, this region encompasses the segAB locus (green triangle in Fig. 1c, d). The plume-like signal indicates that this region undergoes a transition from an isolated state in the G1 population to making more extensive contacts with neighboring sequences (diagrammed in Fig. 1e). This plume-like feature is also apparent in a comparison of the 100 and 0- time points (Supplementary Fig. 1) By 180 min a sub-population of cells have begun to undergo cell division (see Fig. 1f), examination of the difference heat map between 100 and 180 min reveals a blue plume indicative of a re-isolation of this locus and consequent reduction of contacts between the locus and more distal sites. (Fig. 1c, right panel, and zoomed in on the seg locus in Fig. 1d, right panel) The intensity of the blue plume is not as marked as the corresponding red plume seen between 0 and 40 min presumably reflecting the fact that cells within the population are heterogeneously exiting G2, as discussed above. In light of the conformational modulation of the seg locus as a function of cell cycle position, we profiled expression of segAB using qPCR with RNA purified from baby machine-synchronized cells and observed a robust elevation of RNA levels in populations enriched for late G2/early M-phase cells (Fig. 1f). The lowest levels were seen in cells in early S-phase. Thus, expression of the segAB locus drops concomitant with the reabsorption of this locus into a broader architectural domain. In light of these twin observations of cell-cycle modulated conformation of the chromosome at the segAB locus and the cell-cycle dependent expression of the locus, we speculated that the segAB gene products might play a role in chromosome conformational dynamics. Given that the G1-enriched cells had been collected at 4 °C, we performed a control in which asynchronous cells were subjected to a parallel extended incubation at 4 °C prior to crosslinking and Hi-C analysis (Supplementary Fig. 2a). As can be seen in Supplementary Fig. 2b, with magnified view in Supplementary Fig. 2c, the difference heat map comparing G1 cells with chilled asynchronous cells also reveals the same plume from the vicinity of the segAB locus. Our data, therefore, reveal that the segAB-proximal region undergoes a conformational change in which it transitions from a locally insulated feature to making broader contacts with its encompassing domain. In doing so, it removes a boundary element, thus consolidating the extended A-compartment domain seen in G2 and asynchronous cells. This prompted us to re-examine our previous high-resolution asynchronous cell 3C-Seq datasets in which we had mapped anchored loop structures in the S. acidocaldarius chromosome21. Notably, we observed the presence of loop structures emanating from the immediate vicinity of the segAB locus (highlighted in purple in Fig. 1g). Examination of the gene environment of the distal anchor points of these loops does not reveal any conserved features with respect to identity, compartment location or predicted function of the genes within ±10 kb of the anchor point. Previous work has revealed that segAB gene expression is modulated following resumption of growth from a G2-arrest mediated by acetic acid administration to the cultures of Sulfolobales species22,23. In S. acidocaldarius, segAB expression peaks between 80 and 100 min following resumption of growth22. A very recent study revealed that in Saccharolobus islandicus REY15A, segAB expression was highest at 150 min after release from the acetic acid block23. In both species, these time points show the highest levels of dividing cells (M/G1) in the population, in good agreement with our baby machine expression data. The different absolute time points presumably reflect the different doubling times of the two species. While the mechanism of the acetic acid-mediated arrest is unknown, it has been speculated to arise via uncoupling of the proton gradient, leading to a premature stationary phase-like arrest. (While the growth medium of Sulfolobus is below the pKa of acetic acid (4.76), the intracellular pH is between 6 and 6.5). We performed 3C-Seq in both S. acidocaldarius and S. islandicus, comparing arrested cells with populations enriched for dividing cells (Fig. 2). In both species, we observed a reduction in the definition of the compartmentalized architecture of the chromosomes following the acetic acid-mediated arrest (Fig. 2a, b, left-hand panels). Notably, a similar loss of compartment definition is seen as cells enter the stationary phase19 and is presumably reflective of the starvation-like response induced by respiration uncoupling. As cells resume growth and division, we observe a sharpening of the compartmentalized nature of the chromosomes (Fig. 2a, b, right-hand panels) and, importantly, the appearance of a locally insulated boundary element at the segAB locus in both species, as seen in the difference heat maps shown in Fig. 2c (S. acidocaldarius) and Fig. 2d (S. islandicus). In S. acidocaldarius, we observe a plume-like feature, similar in scale to that observed with baby-machine arrested cells. In the case of S. islandicus, the segAB locus is coincident with a major boundary element that is seen in the population enriched in dividing cells. Conformational changes in the Sulfolobus acidocaldarius and Saccharolobus islandicus chromosomes in acetic acid-arrested and dividing cells. a Pearson correlation matrices based on 3C-Seq data binned at 10-kb resolution (upper panels). Red indicates a correlation between bins, and blue indicates an anti-correlation. Principal component analyses based on the Pearson correlations are shown in the bottom panels for arrested and dividing S. acidocaldarius cells. The first principle component is shown as the compartment index. Raw data from two biological replicates were merged for analysis. b Pearson correlation matrices (upper panels) and principal component analyses (bottom panels) for arrested and dividing S. islandicus cells. Analyses were performed as in (a). c Difference heat map of the contact matrices comparing dividing and arrested S. acidocaldarius cells. The whole chromosome is shown on the left, and the right panel is zoomed in and recentered on the segAB locus. Green triangles point to the segAB locus. d Difference heat map of the contact matrices comparing dividing and arrested S. islandicus cells. Green triangles point to the segAB locus. Loss of the segAB genes impacts cell viability and induces DNA damage responses In agreement with a genome-wide mutational study performed in the related species Saccharolobus islandicus24, we were able to delete the Sulfolobus acidocaldarius segAB operon (Supplementary Fig. 3). The mutant cells grew with a reduced rate (generation time 5.6 ± 0.06 h) compared to wild-type (generation time 4.8 ± 0.06 h) (Fig. 3a). Live-dead staining revealed a doubling in the number of dead cells in exponentially growing cultures compared to that seen with wild-type. This ratio was markedly reduced upon entry into stationary phase and concomitant cessation of cell division (Fig. 3b). A 2-fold loss of viability was also seen in measurements of plating efficiency (Supplementary Fig. 4a). In agreement with this, flow cytometry analyses demonstrated that ΔsegAB cells showed aberrant DNA content profiles with elevated levels of cells with sub-1C chromosome content and a small population of >2 C content cells, not seen in the wild-type cells (Fig. 3c and Supplementary Fig. 4b). Importantly, however, these aberrant DNA content cells were in the minority. Furthermore, the overall proportions of G1, S and G2 phase cells were similar to those seen with wild-type cells, indicating that there is no arrest of the cell cycle at any particular stage in the absence of the segAB gene products. Phenotype of ΔsegAB mutant strain. a Growth of triplicate cultures of wild-type (black and gray squares) and ΔsegAB mutant cells (red triangles) was monitored by following absorption at 600 nm. b Live/Dead staining was performed according to the Invitrogen protocol on cultures of wild-type (black) or ΔsegAB mutant cells (red), sampling at early, mid or late logarithmic growth or in the stationary phase. The experiment was performed on three biological replicates for each strain, and data are represented as mean values of dead/live ratios within each culture. Error bars are derived from standard deviations, and gray and black dots represent the individual data points for the SK-1 wild-type strain and the ΔsegAB strain, respectively. c Flow cytometry profiles of cultures of wild-type (upper panel) and ΔsegAB (lower panel) mutant cells during culture growth. The positions of cells with less than one chromosome content (<1C), one (1C), and two (2C) chromosome content are indicated. Supplementary Fig. 4b shows flow cytometry profiles for additional biological replicates. d Marker frequency analyses comparing the replication profiles of wild-type (black) and ΔsegAB (red) cells. The positions of the three origins of DNA replication are shown. We additionally profiled replication of the chromosome using marker frequency analysis (MFA), which revealed that all three replication origins were being used with similar efficiency to wild-type cells. The absolute amplitude of peaks in MFA is dictated by the relative proportion of replicating to non-replicating cells. Thus, the MFA studies support the flow cytometry data indicating that there is no arrest in cell cycle progression at a specific stage in the mutant cell line (Fig. 3d). Epifluorescence microscopy revealed elevated levels of cells with altered morphology or nucleoid appearance in the mutant line when compared with the wild-type counterpart (Fig. 4a). More specifically, comparing mutant to wild-type cells, we observed a three-fold increase in cells with diffuse nucleoids. Additionally, there was a four-fold increase in nucleoid-free cells that failed to stain with DAPI. We also observed a 9-fold increase in cells where the nucleoid was noticeably asymmetric in appearance. Examples of these morphologies are shown in Fig. 4a. Consequences of loss of SegAB on nucleoid morphology, cell size distribution and gene expression. a Distribution of nucleoid morphologies as adjudged by epifluorescence microscopy of DAPI-stained wild-type (left panel) or ΔsegAB cells (right panel). Cells were binned into the six classes described at the bottom of the panel. Examples of the appearance of these cell types are shown for the ΔsegAB cells. Analysis was performed on n = 1078 wild-type cells and n = 1055 ΔsegAB cells. b A Coulter Counter Multisizer 4e was used to measure cell volume distributions of wild-type (blue) or ΔsegAB cultures (red). c DESeq2 analysis was performed on the results of RNA-Seq performed with RNA extracted from exponentially growing wild-type or ΔsegAB cells. The volcano plot compares the transcriptome profile of the mutant and wild-type cells, with genes undergoing greater than 2-fold changes in expression in ΔsegAB cells highlighted in red. d Comparison of the overlap between genes whose mRNA levels were greater than 2-fold elevated by loss of segAB and those induced by exposure to UV-induced DNA damage. The p-value (<2.1e-14) was calculated using the exact hypergeometric probability formula (representation factor 8.7). Next, we utilized a Coulter Counter to measure the size distribution in exponentially growing wild-type and ΔsegAB cells. As seen in Fig. 4b, the wild-type cells demonstrated a distribution of volumes centered around a median of 0.546 µm3. The mutant line showed an altered distribution, with a median of 0.485 µm3 and an extended size distribution ranging from 0.2–5 µm3. During the course of these analyses, we noticed that the mutant culture, but not the wild-type, had a tendency to clog the 20 µM pore on the Coulter Counter. This was overcome by vortexing the cells prior to application and constant stirring during cell counting. We investigated this property further by performing live cell imaging at 75 °C. As can be seen in Supplementary Movies 1, 2, the mutant cells formed extensive clumped aggregates of cells. Previous work has revealed that Sulfolobus cells undergo clumping in response to DNA-damaging insults25,26. Expression profiling and genetic studies have demonstrated that this is due to the induction of pili of the Ups (UV-inducible pili of Sulfolobus) system that then facilitates DNA uptake via the Ced (Crenarchaeal system for exchange of DNA) system25,26. Together, these machineries have been proposed to effect a community-based DNA repair system, presumably via homologous recombination27–31. We speculated that a consequence of the loss of the segAB system could be that double-strand DNA breaks might be generated by cell division, cutting through inaccurately segregated chromosomes. Accordingly, we profiled gene expression in wild-type and ΔsegAB cells using RNA Seq. DESeq2 analysis revealed only 50 genes were differentially-regulated between the two strains, with only a single gene down-regulated in ΔsegAB (Fig. 4c and Supplementary Dataset 1). Significantly (p < 2.1 × 10−14), of the 49 upregulated genes, 19 overlapped with the 96 genes known to be induced by UV-damage (Fig. 4d)25. Notably, this subset of UV-induced genes includes those encoding the Ups and Ced DNA exchange systems together with the ferritin-like protein DPS. All three of these are likely involved in the repair of, or response to, double-strand breaks arising as a result of DNA damage. Thus, in combination with our flow cytometry, Coulter Counter analyses and live cell imaging, these data suggest that in the absence of the segAB system, there is activation of DNA repair systems, presumably in response to the generation of DNA breaks. SegB binds to multiple sites on the chromosome While it seems plausible that the induction of the repair system is a consequence of DNA damage detection, it is also formally possible that the DNA repair system could be under the direct control of the SegAB machinery. Accordingly, to investigate the in vivo localization of the sequence-specific DNA-binding SegB protein on the chromosome, we generated antisera against SegB and performed ChIP-Seq studies. Previous in vitro studies indicated SegB bound specifically to sites in the vicinity of its gene. In agreement with the biochemical data, we found that SegB ChIP generated 17 peaks of occupancy enrichment between genome coordinates 150 and 170 kb (Fig. 5a). We also detected peaks of lower amplitude at other regions of the chromosome (marked with asterisks in Fig. 5b). Binding sites were predominantly intragenic (Supplementary Table 1). Motif identification using the MEME suite revealed an extended consensus site, termed segSX, compared to the dyad symmetric site suggested by the in vitro work9,10. While 23 of the top 25 enrichment peaks of SegB contained the segSX site, two did not. Instead, they possessed a second, purine-rich site that we termed segSR (Fig. 5c). Interestingly, further examination of the ChIP-Seq data revealed 13 enrichment peaks that possessed at least two copies of segSX and/or segSR. These ChIP signals of SegB association with its binding sites were lost when experiments were performed with the ΔsegAB mutant cells (Fig. 5b and Supplementary Fig. 5). Importantly, no association of SegB with the promoters of genes associated with the Ups and Ced systems was detected (Supplementary Fig. 6). We also note that no SegB binding sites were found in the vicinity of the distal anchor points for the loops emanating from the segAB locus. Additionally, there was no significant alteration in transcription in the vicinity of the distal anchor loops when comparing wild-type and ΔsegAB mutant cells (Supplementary Fig. 7). ChIP-Seq profile of SegB binding sites in vivo. a View of binding profile of SegB in the vicinity of the segAB operon. The identity, number and orientation of the two binding site consensus sequences (c) are indicated by arrows above each peak. b Expanded view of the whole chromosome occupancy profile of SegB. The identity, number and orientation of the two binding site consensus sequences (c) are indicated by arrows above each peak for sites distal to the segAB locus. The ChIP seq profile for ChIP with anti-SegB antisera with the ΔsegAB strain is shown in the lower panel. c The two binding site consensus sequences for SegB identified by MEME suite analyses. Arrows indicate short direct or inverted repeat elements within segSX. Chromosome conformation in cells lacking the Seg system Next, we performed 3C-Seq on wild-type and ΔsegAB cells. At the gross level, the plaid pattern that we have previously described in wild-type cells19 was observed for the mutant cells, and thus compartmentalization was maintained in the mutant lines (Fig. 6a and Supplementary Fig. 8). Notably, while our manuscript was in revision, a study of SegAB was published32 that suggested, based on an enrichment of SegB binding sites within the A compartment, that SegB may play an important role in organizing the A compartment. However, our data indicate that the hypothesis is not correct. Principle component analysis revealed a modest increase in noise in the compartment index in the mutant line. It seems likely that this could be attributed to the loss of viability of the mutant strain and our evidence, described above, for enhanced DNA damage in the mutant strain. Nevertheless, these results demonstrate that the SegAB system is not required for the generation or maintenance of compartmentalization per se. Loss of SegAB does not alter chromosome compartmentalization or ClsN occupancy but does results in loss of insulation at the segAB locus. a Pearson correlation matrices (upper panels) and corresponding principal component analyses (lower panels) based on merged 3C-Seq datasets performed with duplicate asynchronous, exponentially growing wild-type (left) or ΔsegAB (right) cultures. The first principle component is shown as the compartment index. b Difference heat map comparing the contact scores between wild-type and ΔsegAB cultures. Green triangles point to the location of the segAB locus and anchor symbols indicate the anchor points of the five loops emanating from the segAB locus. c Difference heat map of the contact matrices of wild-type and ΔsegAB cultures, zoomed in and recentered on the segAB locus. Green triangles point to the location of the segAB locus. d Anchored loop structures detected by Chromosight in 3C analyses of wild-type and ΔsegAB cultures. The positions of the distal loop-anchors emanating from the segAB locus are indicated with anchor symbols in the wild-type cells. These loops are lost in the mutant strain as indicated by the crossed-out anchor symbols. The positions of the three origins of replication are also indicated. e Analysis of the distribution of ClsN in wild-type and ΔsegAB mutant cells. T = total in whole cell extract, S = soluble fraction of protein and C = chromatin-associated fraction. Proteins were separated on SDS-PAGE and transferred to a PVDF membrane. The membrane was probed with antisera against ClsN and Vps4, the latter serving as a marker for the soluble fraction. This experiment was performed in triplicate, and uncropped blots and molecular weight standards are located in Source Data. f ChIP-Seq profiles of ClsN occupancy in exponentially growing cultures of wild-type (upper) or ΔsegAB (lower) cultures. A difference heat map (Fig. 6b and zoomed in and re-centered in Fig. 6c), comparing the contact maps for the two strains, reveals a number of differences between the two strains, most notably in the vicinity of the segAB locus. More specifically, we observe a plume-like feature of local contacts emanating from the segAB locus that is elevated in the mutant strain. Importantly, this is reminiscent of what we observed when cells transition from G1 into S phase (Fig. 1c). Additionally, we observe a loss of contacts between the seg locus and more distal points in the mutant line, as indicated by the blue dots in the heat map with anchor symbols above them in Fig. 6b. These correlate with the distal loop anchor points for loops emanating from the segAB locus seen in wild-type cells. Further analysis of loop structures by Chromosight confirms that the segAB-anchored loops seen in the wild-type cells are lost in the mutant line (Fig. 6d). Taken together, our data support the hypothesis that the extensive binding of SegB to multiple sites surrounding segAB locus generates a locally individualized or insulated structure encompassing this locus, which also makes more specific long-range contacts in a SegAB-dependent manner. Following the completion of cell division and concomitant with the reduction of segAB expression (or in the absence of the SegAB system in the deletion strain) (Fig. 1f), this locus establishes a broader range of local contacts resulting from its (re)absorption into a larger architectural domain. We profiled the binding occupancy of the SMC-superfamily protein, ClsN, in the two strains. First, we performed fractionation experiments revealing that the total amount of ClsN and the amount associated with chromatin were comparable in the two strains (Fig. 6e). Next, we performed ChIP-Seq experiments to map sites of ClsN enrichment along the chromosome. In agreement with the preservation of compartmentalization, we observed a high similarity (R = 0.982) in ClsN binding profiles in wild-type and ΔsegAB cells (Fig. 6f and Supplementary Fig. 9). Thus, unlike the case in many bacteria where ParB serves as a loading factor for the SMC-containing bacterial condensin, SegB is not required for recruitment of the Sulfolobus SMC-superfamily protein, ClsN, to the chromosome. Interplay between the Seg system and the cell division machinery Our previous data demonstrated that as Sulfolobus cells prepare for cell division, the first event is the assembly of a ring of CdvA7. We found that CdvA structures assembled prior to the visible segregation of nucleoids. We could only observe decoration of CdvA rings with ESCRT-III proteins and/or ring ingression in cells that had already segregated nucleoids. While CdvA transcription is tightly cell-cycle regulated, the protein is detectable throughout the cell cycle. Notably, however, we previously had to perform synchronization experiments to generate enriched populations of late G2 cells in order to detect cells containing CdvA structures7. In agreement with this, immunolocalization studies of asynchronous cultures of wild-type cells revealed very low levels of cells with CdvA structures (<1%). Very strikingly, and in marked contrast to the case in wild-type cells, in asynchronous cultures of the ΔsegAB strain, 66% of cells (n = 448) contained either partial or complete ring-structures of CdvA, as detected by epifluorescence microscopy (Fig. 7a and Supplementary Fig. 10, Supplementary Movies 3–5). Precocious CdvA structure formation in the absence of SegAB. a Epifluorescence microscopy of ΔsegAB cells. The top panel shows phase contrast; second from top, DAPI staining; third from top, immunolocalization with anti-CdvA antisera. The bottom panel shows a merge of the three panels above. Five independent cultures were analyzed by immunofluorescence microscopy. b Yeast 2-hybrid analysis with yeast containing plasmids expressing the indicated activation domain or DNA-binding domain fusions. The left panel shows growth on plates selecting for the plasmids (conferring leucine and tryptophan prototrophy). The right panel shows growth under additional selection for histidine prototrophy, conferred by the interaction of the protein fusions. c The top panel shows a Coomassie-stained Bis-Tris gradient gel of soluble (S) and insoluble (I) fractions of whole cell extracts prepared from E. coli expressing neither SegA nor CdvA (control), either protein individually or both together. The lower panel shows a western blot generated by sequential probing with SegA and CdvA antisera. This experiment was performed in triplicate, and uncropped western blots and molecular weight markers are deposited in Source Data. d A model for the action of the Seg system and its interplay with CdvA. See text for details. Illustration created by Anthony S. Baker, The Ohio State Library Medical Visuals. Reproduced with the permission of The Ohio State University. We speculated that there might be a direct interplay between the SegAB system and CdvA. To test this, we performed yeast 2-bybrid assays to determine whether components of the Seg system might physically interact with CdvA. We observed the previously described ability of CdvA to self-interact and also to interact with the ESCRT-III paralog encoded by saci_1373 (also called CdvB). We did not detect any interaction between CdvA and SegB. Importantly, however, we did observe interaction between SegA and CdvA (Fig. 7b). We further tested the functional significance of the SegA-CdvA interaction by heterologously expressing these proteins individually or in combination in E. coli. Notably, expression of CdvA resulted in approximately 43% of the protein pelleting with the insoluble fraction. However, co-expression of SegA with CdvA led to 88% of the CdvA being soluble (Fig. 7c). Thus, it appears that SegA prevents the formation of insoluble, higher-order forms of CdvA. Discussion Taken together, our data implicate the SegAB system in playing a role in the fidelity of chromosome segregation. Loss of this system results in impaired cell growth and viability, yet 50% of cells divide normally. Our observations of changes in the chromosome contact pattern at the seg locus as synchronized wild-type cells progress through the cell cycle indicate that this locus is conformationally malleable. The alteration of contact intensity seen around the segAB locus upon entry into S-phase indicates that this locus is elevating contacts with the chromosome body, suggesting it might be transitioning from a more isolated state to being more homogeneously absorbed into the main chromosome body following the completion of cell division. These conformational changes are compatible with this locus playing a centromere-like role. In light of our transcription, ChIP-Seq and 3C data, we suggest that, as SegB is expressed in the late G2 phase (Fig. 1f), it binds to the 25 kb region encompassing the segAB genes (Fig. 5a) and that this extensive nucleoprotein structure acts to sequestrate this region. Following cell division, as expression of the segAB genes decreases, protein turnover will gradually remove SegB, and consequently, the 25 kb region will be able to reform contacts with its surrounding region. The Seg system is clearly not essential for accurate segregation in all cells. We do, however, observe an elevated population of cells with less than a single chromosome content in flow cytometric studies of the mutant. This observation, coupled with the constitutive expression of DNA-damage responsive genes, leads us to speculate that the SegAB system plays a role in coupling chromosome segregation to cell division. In support of this, our microscopy reveals a dramatic elevation of cells that contain CdvA structures in asynchronous cultures of the strain lacking the Seg system. These observations could be accounted for by proposing that in the absence of the Seg system, cell cycle progression is delayed, leading to an accumulation of cells that have built CdvA rings but in which cell division has not initiated. This may reflect a simple steric inhibition if the cell division machinery is unable to ingress through the mass of an unsegregated nucleoid. Alternatively, there could be a checkpoint functioning to prevent cell division structure ingression until nucleoids are segregated. We do, however, observe cells with altered, non-integral chromosome contents. In the absence of SegAB, we also observe the induction of genes previously implicated in DNA damage repair25,26. Thus, it appears that nucleoid scission can occur, at least in a subset of cells, presumably resulting in the generation of DNA double-strand breaks. Furthermore, our flow cytometry and marker frequency do not provide support for any substantial cell cycle arrest in the cells lacking the SegAB system. How then can we reconcile our various lines of data? Our expression profiling of the transcription of the segAB operon shows a nearly identical profile to that seen for CdvA7. Indeed, independent studies by the Shen, Peng, and She labs have shown that both CdvA and SegAB are under the direct control of the candidate cell-cycle regulator aCcr133,34. Thus, the promoters for these genes are likely coordinately regulated. We propose that the spatially dynamic nature of the SegA protein may provide an answer to the elevated levels of CdvA structures. By analogy to the well-studied ParA protein of bacteria1, and supported by previous biochemical studies9,10, we suggest that SegA, as an ATP-bound dimer, interacts non-specifically with the nucleoid. SegB interacts with SegA and triggers ATP hydrolysis and release of SegA from DNA. SegB is then attracted to nearby nucleoid-associated SegA dimers. As SegB is bound primarily to sites in the vicinity of the SegAB locus, this locus will serve as a cargo, shuttled by SegB, chasing the gradient of SegA over the nucleoid surface. This dynamic model is well supported by a range of experimental data in bacterial systems. We suggest it also applies to the Sulfolobus system. To integrate CdvA into the model, we hypothesize that SegA dimers interact with CdvA and inhibit its polymerization, in agreement with the enhanced solubilization of CdvA seen in our heterologous expression experiments. However, as SegB triggers SegA dimer release from the nucleoid and chromosomes begin to segregate, a zone depleted of SegA dimers is created. This permits polymerization of CdvA and thus provides a permissive zone for mid-cell definition (Fig. 7d). Thus, in wild-type cells, CdvA would polymerize during the process of chromosome segregation, accounting for our previous observations that CdvA structures are discernable before full segregation of the nucleoids has occurred. Furthermore, in the absence of the SegAB system, CdvA polymerization would not be inhibited prior to the initiation of segregation, thus explaining our observation of precocious CdvA structures in the strains lacking the segAB genes. It is notable that there is considerable precedent from studies of bacterial cell division for interaction between systems based on ParA-type ATPases and the FtsZ-based cell division machinery35. For example, the Min system in E. coli acts to inhibit FtsZ ring formation, and in Caulobacter, the ATPase MipZ interacts directly with FtsZ and inhibits polymer formation. (See ref. 32 for a recent review.) However, it must be emphasized that the Sulfolobus cell division machinery is completely distinct from that of bacteria, being based on the ESCRT machinery rather than FtsZ36. Nevertheless, our data indicate a striking convergence in the mechanism of regulation of these entirely non-orthologous machineries. Furthermore, it underscores the central position of the archaeal-specific CdvA protein as a key mediator between the bacterial-like SegAB positioning system and the eukaryotic-like ESCRT proteins. Methods Strain construction The segAB genes (Saci_0204 and Saci_0203) were deleted in the Sulfolobus acidocaldarius SK-5 strain (ΔpyrE, ΔsuaI, ΔargD) by gene replacement with the Sulfolobus tokadaii argD gene, reverting the agmatine auxotrophy in this strain. Growth conditions All S. acidocaldarius strains were grown in Brock’s Medium containing 0.1% tryptone and 0.2% sucrose. The SK-1 and ΔsegAB strains were also supplemented with 10 µg/ml uracil, and the SK-5 strain was grown with 10 µg/ml uracil and 1 mg/ml agmatine. S. islandicus E233S was grown in TSV medium. Baby machine synchronization of S. acidocaldarius was performed as described20. S. acidocaldarius was synchronized with acetic acid by adding 3 mM [final] acetic acid to growing cells at OD600 = 0.2. After incubation at 78 °C for 3.5 h, the cells were washed with 20 mM sucrose and resuspended in pre-warmed Brock’s medium. A similar approach was taken to synchronize S. islandicus, except the cells were treated with 6 mM acetic acid [final] for 6 h. Viability measurements Two methods were used to compare cell viability in the ΔsegAB strain versus the SK-1 strain. First, 100 cells of each strain were plated on Brock’s (0.1% tryptone, 0.2% sucrose, uracil [10 µg/ml]) plates, and the resulting colonies were counted. Cells were simultaneously taken and stained with SYTO9 and propidium iodide dyes according to the LIVE/DEAD kit (Invitrogen). These samples were measured on a Varian Cary 100 Bio UV/Visible spectrophotometer (excitation 470 nm, emission 490–700 nm), and live/dead staining ratios were calculated for each sample. Flow cytometry and cell size measurements Cells were fixed with 72 % ethanol at 4 °C for a minimum of 1 h and then washed with 10 mM Tris, 10 mM MgCl2 (pH 7.5). The cells were treated with RNaseA [100 µg/ml] and stained with 10 µM SYTOX Green before being analyzed on an LSRII flow cytometer in the flow cytometry facility at Indiana University. Cell size parameters were measured with a Beckman Coulter Counter Multisize 4e. Wild-type cells were diluted 1/1200 in Beckman electrolyte solution and measured using a 20 µM aperture. Because the mutant cells are prone to clumping, ΔsegAB cells were diluted 1/2000 into a beaker of electrolyte solution and were stirred during measurement with the 20 µM aperture. Three experiments were performed for each strain, and a total of 90,625 wild-type cells and 90,885 ΔsegAB cells were measured. Chromosome conformation analysis Hi-C and 3C-Seq were performed and analyzed as detailed in refs. 37,38. RNA purification and analysis RNA samples from asynchronous cultures and cultures synchronized by the Baby Machine method were purified as described6. For RT-PCR, first-strand cDNA was synthesized from 250 ng of purified total RNA using 250 ng random hexamers (Invitrogen) and SuperScript III Reverse Transcriptase (Invitrogen) according to the manufacturer’s instructions. No RT controls were prepared using water in place of the RNA. SYBR detection assays were performed in triplicate using 5.5% of cDNA reaction as template, forward and reverse primers at 125 nM final concentration, and 2.5X RealMasterMix/SYBR Solution (5 PRIME) according to the manufacturer’s instructions. Standard curves were prepared in duplicate using purified PCR products as templates. qPCR reactions and melting curves were performed and analyzed using a Mastercycler ep realplex2 machine and included analysis software (Eppendorf). For RNA-Seq, strand-specific libraries were prepared following the protocol for mRNA with the NEBNext Ultra II Directional RNA Library Prep Kit (NEB). Paired-end reads were obtained on an Illumina NextSeq 75 at the Center for Genomics and Bioinformatics at Indiana University. Mapping was performed with Geneious Prime v2025.1.2 (www.geneious.com), and analysis was performed with SeqMonk v1.48.1(http://www.bioinformatics.babraham.ac.uk/projects/seqmonk/). Marker frequency analysis Genomic DNA was purified from Sulfolobus cells with phenol/chloroform and ethanol precipitation. Libraries were made using an NEBNext Ultra II kit (NEB), and next-generation sequencing was performed at the Center for Genomics and Bioinformatics at Indiana University. Approximately 10 million paired-end reads were obtained from an Illumina NextSeq 75. Mapping was performed with Geneious Prime v2025.1.2 (www.geneious.com), and analysis was performed with SeqMonk v1.48.1(http://www.bioinformatics.babraham.ac.uk/projects/seqmonk/). Read counts were grouped in one-kilobase bins. Microscopy Immunofluorescence imaging was performed as described6,39. Anti-CdvA antibodies were affinity-purified using immobilized recombinant CdvA7. For nucleoid morphology studies, cells were only stained with DAPI [1.5 ng/µ] after fixation and permeabilization. For live cell imaging, actively growing S. acidocaldarius cells were transferred to VAHEAT chambers (Interherence SmSL DR-16 smart substrates) set to 75 °C. Nile Red membrane stain was added [5.5 ng/µl] 30 min prior to visualization40. Phase contrast and fluorescent images were taken every 5 min for 3 h. All images were acquired with a Nikon Eclipse Ti2 inverted microscope using a 100X oil immersion objective, a Hamamatsu ORCA-Flash 4.0 LT digital camera, and NIS Elements software (Nikon). ChIP-Seq Chromatin immunoprecipitation (ChIP) was performed as described41. ChIP reactions were performed in triplicate and pooled. Libraries were constructed using NEBNext Ultra II kits (NEB), and sequencing was performed on an Illumina NextSeq 75 at the Indiana University Center for Genomics and Bioinformatics. Ten million paired-end reads were mapped to the S. acidocaldarius. Mapping was performed with Geneious Primer v2025.1.2 (www.geneious.com), and analysis was performed with SeqMonk v1.48.1(http://www.bioinformatics.babraham.ac.uk/projects/seqmonk/). Read counts were normalized to the input DNA Chromatin fractionation Cells in exponential growth were pelleted and resuspended in chromatin extraction buffer (25 mM HEPES, 15 mM MgCl2, 100 mM NaCl, 400 mM sorbitol, 0.5% Triton X-100, pH 7.5) using 100 µl of buffer for every milliliter of culture at OD600 = 0.1. The cells were incubated on ice for 10 min and then centrifuged at 14,000 × g, 4 °C for 20 min. The soluble fraction was transferred to a new tube, and the pelleted chromatin fraction was resuspended in a volume of chromatin extraction buffer equivalent to the volume of the soluble fraction. Extracts were run on SDS-PAGE gels, and proteins were detected by western blotting. Uncropped blots can be found in Source Data. Yeast 2-hybrid assay Yeast 2-hybrid analyses were performed using the Matchmaker system (Clonetech). The Saccharomyces cerevisiae Y2H Gold strain was co-transformed with plasmids encoding protein fusions to the GAL4 DNA-binding domain (pGBK) and the GAL4 activation domain (pGAD). Transformants were selected on SC-Leu-Trp agar. Liquid cultures of the transformants were diluted to OD600 = 0.1, and 5 µl were spotted on SC-Leu-Trp-His agar. Growth on this selective agar was indicative of potential protein–protein interactions. Solubility of heterologously expressed proteins E. coli Rosetta cells were transformed to create the following four expression strains: (1) pETDuet + pCDFDuet, (2) pETDuet-SegA + pCDFDuet, (3) pETDuet + pCDFDuet-CdvA, and (4) pETDuet-SegA + pCDFDuet-CdvA. 20 ml cultures were grown in LB (ampicillin, spectinomycin, chloramphenicol) and induced with 1 mM IPTG at OD600 ≈ 0.8. After 3 h of induction, 1 ml of each culture was pelleted and resuspended in ~100 µl of Pro-Lyse Cell Lysis Buffer (Protein Express) to normalize the OD/µl concentration among the samples. The samples were rotated at room temperature for 10 min and then centrifuged at 21 K × g for 5 min to separate the soluble from the insoluble fractions. The resulting fractions were run on 4–12% Bis-Tris gradient gels and either stained with Coomassie for total protein visualization or transferred to a PVDF membrane for western detection of the SegA and CdvA proteins. Uncropped blots can be found in Source Data. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Dataset 1 Supplementary Movie 1 Supplementary Movie 2 Supplementary Movie 3 Supplementary Movie 4 Supplementary Movie 5 Reporting Summary Transparent Peer Review file Source data Source Data Acknowledgements S.D.B.’s lab is funded by NIH R35GM152171. This work was additionally funded by NIH 1R01GM135178, the College of Arts and Sciences, Indiana University and the College of Arts and Sciences, The Ohio State University. We are indebted to Veronica A. Bell for invaluable assistance with microscopy. We thank Buzz Baum for discussions prior to publication. We are also grateful to Mart Krupovic and Miguel Gomez Raya Vilanova for sharing protocols and advice for acetic acid-mediated arrest and release of S. islandicus. Author contributions R.Y.S. and N.T. performed the baby machine synchronization, and N.T. performed the Hi-C experiments on this material. All other experiments were performed and analyzed by R.Y.S. S.D.B. assisted in experimental design, obtained funding and wrote the initial draft of the manuscript. All three authors contributed to the final text of the manuscript. Peer review Peer review information Nature Communications thanks Qihong Huang and the other anonymous reviewer for their contribution to the peer review of this work. A peer review file is available. Data availability The sequence reads for the 3C-Seq, RNA-Seq and ChIP-Seq data have been deposited in the Sequence Read Archive, BioProject PRJNA1272008; submissions SUB15364289 and SUB15720780. Source data will be provided with this paper. Strains are available on request from RYS. Source data are provided with this paper. Competing interests The authors declare no competing interests Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Rachel Y. Samson, Email: samson.11@osu.edu Stephen D. Bell, Email: bell.2007@osu.edu Supplementary information The online version contains supplementary material available at 10.1038/s41467-025-67934-8. References Associated Data Supplementary Materials Supplementary Information Description of Additional Supplementary Files Supplementary Dataset 1 Supplementary Movie 1 Supplementary Movie 2 Supplementary Movie 3 Supplementary Movie 4 Supplementary Movie 5 Reporting Summary Transparent Peer Review file Source Data Data Availability Statement The sequence reads for the 3C-Seq, RNA-Seq and ChIP-Seq data have been deposited in the Sequence Read Archive, BioProject PRJNA1272008; submissions SUB15364289 and SUB15720780. Source data will be provided with this paper. Strains are available on request from RYS. Source data are provided with this paper.