2.1 CsmR is a novel potential regulator of archaellation in H. volcanii, distinct from EarA

The gene encoding CsmR (hvo_1209), a small 13 kDa protein, is located upstream of the archaellin genes arlA1 and arlA2 in the H. volcanii genome. The presence of a predicted winged-helix DNA-binding domain suggested that similar to the previously identified positive regulators EarA in M. maripaludis and P. furiosus [22,25], this protein would be involved in the regulation of the archaellum cluster expression in H. volcanii. However, homology searches revealed no sequence or structural similarity between CsmR and EarA, suggesting a distinct regulatory mechanism in H. volcanii. Phylogenetic distribution analysis showed that csmR is predominantly found in Halobacteria and some members of Thaumarchaeota (S1A Fig). Moreover, gene cooccurrence investigation showed that csmR always cooccurs with the archaellin genes arlA1 and arlA2 in Halobacteria but not in Thaumarchaeota hinting to another regulatory function in the latter phylum than motility control (S1B Fig). Structural modeling classified CsmR within the Lrp/AsnC family of transcriptional regulators, known for their roles in metabolic and stress-response pathways in archaea and bacteria [38,39]. The highest prediction of confidence of the AlphaFold 3 model supported dimer formation (pTM 0.79), consistent with the dimeric nature of Lrp/AsnC regulators. The predicted dimerization interface of CsmR is mediated by a β-sandwich fold, further aligning it with Lrp/AsnC regulators rather than the previously characterized archaellum regulators (S2 Fig).

2.2 Deletion of csmR results in the formation of elongated cells and a complete loss of motility whilst overexpression renders cells to be hypermotile and constantly rod-shaped

To investigate the function of the potential archaellum regulator in Haloferax volcanii we generated a marker less deletion strain of csmR, that grows like the H26 control (S3A Fig) and assessed a potential effect on cell locomotion via motility plates. For all experiments strains, if not mentioned differently, were transformed with plasmid pTA1392 to complement the uracil auxotrophy promoting more reliable cell growth and a more homogeneous cell shape. In comparison to the control strain H26 the csmR deletion strain fully lost the ability to swim on semi solid agar plates (Fig 1A) indicating that csmR controls archaellum expression. Indeed, transmission electron microscopy of a ΔpilB3ΔcsmR strain showed cells completely devoid of archaellum filaments (Fig 1B), whilst the ΔpilB3 control strain showed archaellation. Since the discrimination between pili and archaellum filaments due to their homology is difficult, a deletion strain of the pilin assembly ATPase PilB3 [40] of H. volcanii was used to obtain cells that only have archaella as cell appendages. Uncropped TEM images are provided in the supplementary information (S4 Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Deletion of csmR leads to a complete loss of motility whilst overproduction results in hypermotility.

(A) Motility assay comparing wildtype H26 with the csmR deletion strain (transformed with pTA1392) and both strains overexpressing csmR (transformed with pSVA13922). Exemplary motility halos of all strains are shown. The area of the motility halos was measured and normalized to the average area of wildtype halos showing a significant reduction (p ≤ 0.0001) in cell locomotion of the csmR deletion strain compared to H26 and a significant (p ≤ 0.0001) increase of motility by csmR overproduction. Samples were measured in biological and technical triplicates and all nine single data points per strain were plotted. The middle line indicates the mean and the upper and lower line the standard deviation. Scale bar 2 cm. (B) Transmission electron microscopy images of the pilB3 + pTA1392 and the pilB3 csmR + pTA1392 double deletion strain (left) and of both strains overexpressing csmR from pSVA13922 (right). The filament of the archaellum is not assembled in the double deletion strain whilst overexpression of csmR resulted in assembly of an archaellum filament in the double deletion strain. Scale bar 1 µm. (C) Cell shape analysis of the wildtype H26 + pTA1392 and the csmR + pTA1392 deletion strain at low, mid and stationary OD₆₀₀ and under csmR overexpression condition. Both strains show a transition from rod- to plate-shaped cells. In contrast both overexpression strains stay rod-shaped over the whole growth cycle and do not transition from rod- to plate shaped cells. Scale bar 4 µm. (D) Cell shape was analyzed using MicrobeJ and the summarized results of three independent biological replicates per strain and OD₆₀₀ value plotted as violin-plots. The median is indicated by the middle black line, dotted lines indicate the first and third quartile. For each condition more than 1000 cells were analyzed. The dotted line indicates the length below which cells are considered plate shaped.

https://doi.org/10.1371/journal.pgen.1012198.g001

Several studies showed that motility of H. volcanii strictly depends on its cell shape [28]. Haloferax volcanii is pleomorphic being rod shaped and motile in lag- growth phase whilst during the exponential growth phase cells undergo a morphologic transition to plate shaped cells that are not motile anymore even though the archaellum motor is still present [28]. To exclude the option that the csmR deletion strain is immotile due to cell shape effects, we imaged cells of the deletion strain at different growth stages and compared them to strain H26 (early (OD₆₀₀ of 0.02), mid (OD₆₀₀ of 0.2), and stationary growth (OD₆₀₀ of 2.0)). Both strains showed the typical cell shape of H. volcanii during growth, with rod-shaped cells at early growth stages and a transition to plate-shaped cells when they reach the stationary phase. However, ΔcsmR cells were more elongated during growth than H26 cells hinting to a regulatory function of CsmR not only in cell locomotion but also cell shape determination (Fig 1C and 1D). To exclude a potential effect on cell shape and motility of the investigated strains due to the transformation with pTA1392 experiments were also conducted in absence of the plasmid with the exact same result as with plasmid (S3B and S3C Fig).

Overexpression of the euryarchaeal archaellum regulator (EarA) in other Euryarchaeota led to drastically increased archaellum formation [25]. To test if overexpression of CsmR in H. volcanii has a similar effect, H26 and the csmR deletion strain were transformed with a plasmid that harbored csmR under the control of a strong xylose inducible promoter [41]. Assessment of the optimal inducer concentration showed that the csmR deletion phenotype, in regard to motility, is already complemented without addition of xylose. Derived from motility assays under induced and non-induced conditions, we assume that under non-induced conditions, CsmR synthesis is very low (S5A Fig) suggesting that relatively low CsmR concentration already have an impact on the cell. The more the inducer concentration was increased the more strains transformed with the overexpression plasmid showed motility on the plates (S5B and S3C Figs). To obtain maximum expression of csmR 12 mM xylose was used for induction. In the ΔcsmR strain overexpression complemented the non-motile phenotype of the deletion strain not only back to wildtype levels but caused hypermotility (Fig 1A). The same was observed when csmR was overexpressed in H26, causing motility halos more than twice the size of the motility halo of the control strain. Increased cellular protein concentration of EarA in P. furiosus cells lead to hyperarchaellation [25]. However, in contrast to P. furiosus, ΔpilB3 H. volcanii cells transformed with the csmR overexpression plasmid showed no hyperarchaellation phenotype. The ΔpilB3ΔcsmR double deletion strain assembled archaellum filaments upon csmR overexpression, as observed with the transmission electron microscope (Fig 1B); however, the number of archaella was comparable to the control strains with or without csmR overexpression plasmid (Fig 1B).

As mentioned before, cell motility of H. volcanii strongly depends on its cell shape. Whilst the control strain H26 and the csmR deletion strain transition from rod to a plate-like shape during growth, both strains are locked in a rod-shaped appearance upon csmR overexpression (Fig 1C and 1D). Therefore, CsmR not only seems to affect archaellum expression but possibly also regulates genes that are involved in rod formation and/or maintenance.

2.3 Cell shape transition does not alter membrane lipid composition

To ascertain if morphological changes impact the composition of membrane lipid types, the abundance of several lipid species present in H. volcanii was assessed during different growth phases (S6 Fig). Overall, only small changes were observed, which predominantly involve the lower abundant lipid species PA (phosphatidic acid), PE (phosphatidylethanolamine), MGD (monogalactosyldiacylglycerol), as well as the cardiolipins. The only major abundant lipid species that seems to be mildly influenced by the growth stage is PG (phosphatidylglycerol). This might be explained by the fact that PG serves as a substrate for all cardiolipins, which in turn seem to be slightly upregulated at later growth stages, which is a common phenomenon observed in prokaryotes. Next, analysis of the lipid tail distribution of the four most abundant lipid species in H. volcanii was performed (S7 Fig). C40:0 (i.e., two times a phytanyl chain) is the predominantly observed form, which is in accordance with the literature [42]. Other lipid tail configurations do exist, in which both the level of desaturation as well as the carbon chain length can differ. In general, there seems to be a trend that with increasing optical densities, there is a decrease in saturated lipid tails, except for PG. Again, this might be explained by the fact that unsaturated PG serves a substrate for cardiolipins, which is more prominently produced during later growth stages. Altogether though, the overall changes in lipid composition (headgroup as well as tail configuration) are minor. Hence, it is unlikely that changes in cell morphology during growth are caused by an altered lipid composition.

2.4 CsmR deletion induces widespread upregulation of archaellum and chemotaxis genes

To further investigate how CsmR regulates motility and cell shape transitions, we analyzed its impact at the transcriptional level. We performed RNA-sequencing on both wild type and ∆csmR strains under three distinct growth conditions to: (1) assess the overall transcriptional effect of csmR deletion and uncover potentially associated regulatory networks and (2) determine whether these effects vary under conditions where cells typically lack appendages and adopt a plate-like morphology (Fig 2A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Deletion of csmR leads to upregulation of archaellum and chemotaxis genes.

(A) Schematic of the experimental design, comparing the Haloferax volcanii wild type strain H26 and a deletion strain of the potential archaellum regulator CsmR, across different growth phases. Note that during normal growth conditions, Haloferax exhibits pleiomorphism, transitioning from rod-shaped, motile cells to sessile plate-like cells. (B) Principal component analysis from wild type (WT, grey) and ∆csmR (archaellum regulator, purple) strains, based on total read counts after outlier removal. Different growth phases are indicated by different cell shapes, illustrating the primary source of transcriptomic variance. (C) Volcano plot showing differential gene expression in low OD samples. Genes are represented as circles, with log₂-fold changes color-coded from dark brown (downregulated in deletion strain) to dark green (upregulated in deletion strain), expression levels indicated by circle size and archaellum-related genes highlighted with dashed outlines. (D) Mean coverages of strand-specific, TPM (transcripts per million)-normalized reads are shown for wild type (grey) and csmR deletion strain (purple). The middle tracks display archaellum and chemotaxis-related genes, according to their relative gene size (scale bar depicted). Genes are color-coded based on log₂-fold changes from dark brown (downregulated in deletion strain) to dark green (upregulated in deletion strain) under low-OD conditions and strand orientation indicated by rectangle direction.

https://doi.org/10.1371/journal.pgen.1012198.g002

Principal component analysis revealed that the growth phases of the harvested cultures were the primary driver of transcriptomic variation, while the deletion of csmR had only a smaller impact. This supports the hypothesis that CsmR functions as a dedicated regulator of archaellation rather than on a global scale (Fig 2B). To further investigate gene-specific effects, we performed differential gene expression analysis across the growth phases (early (OD₆₀₀ of 0.02), mid (OD₆₀₀ of 0.2), and stationary growth (OD₆₀₀ of 2.0)). In early exponential growth, ∆csmR cells showed significant upregulation of shape-relevant genes, including sph3 (hvo_2175) and rdfA (hvo_2174), as well as chemotaxis-associated genes (Fig 2C). Despite the complete loss of motility, RNA-seq analysis revealed a significant upregulation of archaellum genes in ∆csmR (Fig 2C). H. volcanii features a unique gene arrangement, where archaellation and chemotaxis genes are interspersed and transcribed in different orientations. Despite this organization, ∆csmR exhibited consistent upregulation across entire transcriptional units, including ArlA1/A2, ArlC/E/F/G/G/I/J/CheF2, CheB/A/R/F1, CheY/C/D, and ArlD/HVO_1202 (Fig 2D). This contrasts with the expected results, given the strain’s impaired swimming ability, and deviates from regulatory patterns observed in other euryarchaeal archaellum regulators [24,25]. Rather than implying that CsmR acts as a direct transcriptional repressor of these genes, this paradoxical upregulation suggests a more complex regulatory mechanism in which CsmR may function as a regulator whose absence indirectly derepresses archaellum gene expression, possibly through dysregulation of counterbalancing regulators.

2.5 CsmR deletion leads to persistent activation of motility and signalling pathways across growth phases

To assess the impact of the csmR deletion across different growth conditions, we next examined differential gene expression in mid and stationary samples. Pearson’s correlation analysis (R = 0.27) revealed a weak positive correlation between log₂-fold changes in early and exponential (mid) growth phases, indicating that only a subset of genes was consistently up- or downregulated across both conditions (Fig 3A). Notably, several of the most upregulated genes in early growth, particularly those linked to archaellation and cell shape regulation, remained upregulated in the exponential phase. Overall, during early phase, 8% of genes were upregulated and 7% downregulated, with similar proportions in the exponential phase (6% upregulated, 8% downregulated), though there was minimal overlap between these gene sets (Fig 3B). In the stationary phase, where cells adopt a different morphology, gene regulation was significantly more pronounced, with 26% of the genes upregulated and 25% downregulated. Notably, archaellation and chemotaxis-associated genes remained consistently upregulated across all growth phases (Fig 3C). However, no clear functional pattern among downregulated genes could be detected. To further investigate the biological significance of these transcriptional changes, we performed functional enrichment analysis using archaeal clusters of orthologous groups (arCOGs). This analysis revealed overrepresentation trends in groups N (cell motility) and T (signal transduction mechanisms) among upregulated genes across all growth phases. Additionally, categories Q (secondary metabolites biosynthesis, transport, & catabolism), F (nucleotide transport & metabolism), and C (energy production & conversion) were enriched in stationary phase, while category E (amino acid transport & metabolism) was consistently underrepresented among downregulated genes across all conditions (Fig 3D).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. CsmR deletion leads to a shared functional response across growth conditions.

(A) Scatter plot showing the log₂-fold changes in gene expression between early and mid-exponential samples for the ∆csmR strain relative to the wild type. Archaellum genes (dashed outline, light pink) and other genes that are consistently regulated under both conditions are highlighted. Genes with significant changes (adjusted P-value < 0.05) under both conditions are colored in pink. (B) Flow diagram illustrating the relative number and interconnectedness of genes between all growth conditions (x-axis) for each regulatory group (upregulated: green, downregulated: brown, rest: off-white). (C) Heatmap displaying the top 10 genes with the highest mean log₂-fold changes across all conditions. Significance (adjusted P-value < 0.05) is indicated by triangles, with log₂-fold changes color-coded from dark brown (downregulated in deletion strain) to dark green (upregulated in deletion strain). (D) Gene set enrichment analysis of archaeal clusters of orthologous groups (arCOGs) across all three growth conditions. Color intensity (white to dark purple) indicates -log10(raw overrepresentation P-value), while circles denote categories with raw overrepresentation P-value < 0.05. Circle size reflects the total number of differentially regulated genes assigned to each category. These enrichment results are intended as exploratory.

https://doi.org/10.1371/journal.pgen.1012198.g003

2.6 CsmR directly binds promoter regions of motility-associated genes

To distinguish between direct and indirect regulatory effects and to identify genomic binding sites of the potential global regulator CsmR, we performed chromatin immunoprecipitation followed by sequencing (ChIP-seq). Distance analysis of peak summits relative to annotated gene start positions revealed a strong enrichment of CsmR binding sites in close proximity to gene starts, consistent with promoter-proximal binding (S8A Fig). In total, 65 genes were associated with high-confidence CsmR peaks, with the majority of sites located in intergenic regions (41 intergenic vs. 24 intragenic). Inspection of individual loci revealed pronounced CsmR enrichment at promoter regions of genes associated with archaellation and chemotaxis, including the arl operon (Figs 4A and S8B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. CsmR binds promoter regions of motility-associated genes and is linked to transcriptional changes.

(A) Mean log₂(ChIP/Input) signal across the arl operon region showing CsmR enrichment at promoter-proximal sites. Individual replicate ratios are shown in grey. Peak summits are indicated by vertical lines, and gene annotations are shown below according to genomic orientation. (B) Sequence logo of the CsmR binding motif identified by de novo motif analysis of peak-associated regions. The semi-palindromic consensus sequence TATCA(N₄)TGATA is enriched upstream of target genes and positioned relative to the archaeal TATA box. (C) Strand-normalized positions of motif matches relative to ChIP-seq peak summits, showing that the CsmR motif occurs in a defined position near the centre of enriched regions. (D) Integration of ChIP-seq and RNA-seq data. Differential gene expression in the ∆csmR strain is shown as log₂ fold change versus expression level. Genes associated with CsmR peaks are highlighted and grouped based on binding characteristics (motif presence, strand orientation, and promoter distance). (E) Comparison of log₂ fold changes across gene classes. Genes associated with motif-containing, promoter-proximal, strand-consistent CsmR peaks show the strongest upregulation in the ∆csmR strain, followed by genes within the same transcriptional units. Statistical significance was assessed using Wilcoxon rank-sum tests.

https://doi.org/10.1371/journal.pgen.1012198.g004

De novo motif analysis identified a ~ 34 bp promoter-proximal sequence containing a semi-palindromic consensus motif, TATCA(N₄)TGATA, present in 50 of 65 peaks. Interestingly, adjacent Canonical archaeal promoter elements like the TATA box are located adjacent to the CsmR binding site (Fig 4B). Comparison of the core CsmR motif to known transcription factor binding sites did not reveal significant similarity to previously described Lrp/AsnC motifs. After accounting for motif orientation, these sequences showed a consistent positional relationship relative to peak summits (Fig 4C).

Integration of ChIP-seq and transcriptome data showed that genes strongly upregulated in the ∆csmR mutant were predominantly associated with promoter-proximal CsmR peaks or belonged to transcriptional units containing such peaks (Fig 4D). Genes with motif-containing, strand-consistent peaks at promoter-compatible distances exhibited the strongest upregulation, followed by operon-linked genes, whereas genes without detectable binding showed little or no change (Fig 4E). These differences were statistically significant across all tested conditions. Direct CsmR targets showed consistent growth-phase-dependent expression patterns, with similar trends in wild type and ∆csmR strains, while CsmR transcript levels remained stable across all conditions, suggesting that its activity is not primarily controlled at the transcriptional level (S9 Fig).

Following the identification of the CsmR binding sites by ChIP-seq, we further investigated its structural features to assess whether they are consistent with DNA-binding and with membership in the Lrp/AsnC family of transcription factors. Structure-guided comparisons with characterized Lrp/AsnC family members, including E. coli Lrp/AsnC and B. subtilis LrpC, showed that the strongest structural conservation is confined to the predicted N-terminal DNA-binding region of CsmR, whereas the remaining part of the protein is more divergent (S10 Fig). Thus, rather than displaying close full-length similarity to canonical Lrp/AsnC regulators, CsmR appears to retain a conserved DNA-binding core within an otherwise divergent architecture. In addition, Foldseek identified the P. furiosus Phr regulator (PDB ID: 2P4W) as a structurally related hit, further supporting similarity in the overall fold and arrangement of the predicted DNA-binding region (S11 Fig). Consistently electrostatic surface mapping identified a positively charged surface patch in this region, compatible with a DNA binding interface (S12 Fig). A multiple sequence alignment (MSA) across haloarchaeal homologs further showed that this N-terminal region is the most conserved part of the protein and contains several conserved basic residues, supporting its role in direct transcriptional regulation (S13 Fig). Analysis with the Protein Structure Transformer (PeSTo) tool [43] likewise predicted a DNA binding interface in this conserved and positively charged region (S14 Fig). In contrast, the remainder of CsmR showed substantially lower sequence conservation, consistent with a more divergent family associated region that may contribute to oligomerization or regulatory protein interactions rather than sequence specific DNA recognition (S14 Fig).

2.7 CsmR and CirA co-regulate archaellation and chemotaxis through overlapping but distinct pathways

Given that csmR deletion resulted in widespread transcriptional upregulation of archaellation and chemotaxis genes, we next sought to identify additional regulatory factors that may interact with or modulate CsmR function. In particular, cir genes (cirA/ cirD) emerged as key candidates, given their established role in archaella regulation and the observed transcriptional shifts in ΔcsmR. To further explore these connections, we examined cir gene expression patterns in ΔcsmR. In the wild type, transcript levels for cirA and cirD were comparable across growth phases. However, in the ∆csmR strain, cirD exhibited consistent upregulation across all phases, while cirA showed reduced expression in early and stationary phases but remained unchanged during mid-exponential conditions (S15 Fig). To further analyze their roles, we generated ∆cirD and ∆cirA deletion strains. Phenotypically, ∆cirD resembled ∆csmR, with non-motile cells and shape dependency, while ∆cirA displayed hypermotility as previously described [33] and a persistent rod-like morphology across all growth phases (Fig 5A–5C). Additionally, transmission electron microscopy revealed archaellation of the cirA deletion strain even at stationary growth phase in contrast to the cirD deletion strain that was not archaellated at all (Fig 5D and 5E). Notably, a double deletion strain of ∆csmR ∆cirA resulted in a phenotype identical to ∆csmR alone, suggesting that CirA acts upstream or in parallel to CsmR in motility regulation (Fig 5F–5H). With respect to cell shape the double deletion of csmR cirD showed no difference to the wild type. However, the ∆csmR ∆cirD strain showed slight motility (S16A–S16C Fig). This was surprising as both single deletions were fully non-motile. Despite these morphological differences in ∆cirA, growth phase remained the primary driver of transcriptomic variance across all strains (S17A and S17B Fig). However, ∆cirA exhibited significantly more deregulated genes during early and mid-exponential growth, whereas ∆csmR and ∆cirD displayed stronger transcriptomic shifts during stationary phase (S18 Fig). Gene set enrichment analysis revealed that ∆cirA and ∆csmR share similar functional effects, particularly among upregulated genes during early and mid-exponential phases (S19 Fig). However, in stationary phase, ∆csmR more closely resembled ∆cirD, except for motility-related genes, which were not enriched in ∆csmR. Focusing on CsmR targets identified through ChIP-seq, we found that ∆cirA deletion led to strong upregulation on the respective genes, whereas ∆cirD had minimal impact (S20 Fig). While overall expression correlations between the strains were moderate, CsmR target genes were among the most differentially expressed genes in both ∆cirA and ∆csmR strains (S17C Fig). Notably, despite high correlation between ∆csmR and ∆cirD during stationary phase, chemotaxis and archaellin genes showed distinct regulation (S21A Fig). Using publicly available ΔrosR transcriptomic data, we did not observe transcriptional changes associated with archaellation that correlated with the patterns seen for ΔcsmR, ΔcirA, or ΔcirD (S21B and S12C Fig). To validate these findings, we analyzed the data independently of the previously identified promoter CsmR targets and detected 49 genes significantly upregulated in ∆csmR and ∆cirA across all growth phases (padj < 0.1) (S17D Fig). STRING network analysis revealed a highly interconnected cluster of archaella assembly, chemotaxis, and signaling genes (S17E Fig). This core included well-characterized genes such as ArlA, ArlB, ArlC, and CheF, alongside transcriptional regulators like TFB and CirA. Interestingly, pilin genes were upregulated in ∆csmR and ∆cirA, suggesting a compensatory shift towards type IV pili-related functions, such as surface adhesion, in response to disrupted archaella regulation. Glycoproteins and signaling proteins were also linked to this network, reinforcing the influence of both CsmR and CirA on chemotaxis and cellular signaling. Notably, CirD-regulated genes were absent, highlighting its distinct and phase-specific role.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. CirA and CirD impact cell locomotion of H. volcanii but the csmR deletion phenotype is dominant over the ΔcirA phenotype.

(A) Cell shape analysis of the wildtype H26 + pTA1392, the cirA + pTA1392 and cirD + pTA1392 deletion strain at low, mid and stationary OD₆₀₀. The cirA deletion strain stayed rod-shaped over the whole growth cycle and did not transition from rod- to plate shaped cells as the control strain or the cirD deletion strain. Scale bar 4 µm. (B) Cell shape was analyzed using MicrobeJ and the summarized results of three independent biological replicates per strain and OD₆₀₀ value plotted as violin-plots. The median is indicated by the middle black line, dotted lines indicate the first and third quartile. For each condition more than 1000 cells were analyzed. The dotted line indicates the length below which cells are considered plate shaped. (C) Motility assay comparing wildtype H26 + pTA1392 with the cirA + pTA1392 and cirD + pTA1392 deletion strain. Exemplary motility halos of the three strains are shown. The area of the motility halos was measured and normalized to the average area of wildtype halos showing a significant reduction (p ≤ 0.0001) in cell locomotion of the cirD deletion strain compared to H26. In contrast the deletion strain of cirA showed increased (p ≤ 0.0001) motility compared to the wildtype. Samples were measured in biological and technical triplicates and all nine single data points per strain were plotted. The middle line indicates the mean and the upper and lower line the standard deviation. Scale bar 2 cm. (D) Transmission electron microscopy images of the pilB3 cirA and the pilB3 cirD double deletion strains showing that the cirA deletion strain is archaellated while the cirD deletion strain is not. Scale bar1 µm. (E) Transmission electron microscopy images of the pilB3 cirA double deletion strain at the stationary growth phase being still archaellated. Scale bar 1 µm. (F) Cell shape analysis of the wild type H26 + pTA1392 and the csmR cirA + pTA1392 double deletion strain at low, mid and stationary OD₆₀₀. Scale bar 4 µm. (G) Cell shape was analyzed using MicrobeJ and the summarized results of three independent biological replicates per strain and OD₆₀₀ value plotted as violin-plots. The median is indicated by the middle black line, dotted lines indicate the first and third quartile. For each condition more than 1000 cells were analyzed. The dotted line indicates the length below which cells are considered plate shaped. (H) Motility assay comparing wild type H26 + pTA1392 with the csmR cirA + pTA1392 double deletion strain. Exemplary motility halos of both strains are shown. The area of the motility halos was measured and normalized to the average area of wild type halos showing significantly (p ≤ 0.0001) decreased motility in the double deletion strain compared to the H26 control. Samples were measured in biological and technical triplicates and all single data points per strain were plotted. The middle line indicates the mean and the upper and lower line the standard deviation. Scale bar 2 cm.

https://doi.org/10.1371/journal.pgen.1012198.g005

2.8 Posttranscriptional regulation of cirA by the RNA hvo_1211s

Since CirA appears to be prominently involved in archaellum and cell shape regulation we had a closer look at its genetic neighborhood. Very recently, the transcriptional regulator RosR was shown by ChIP-seq to bind to the 3’ end of the arlA2 gene that is located downstream of cirA [44]. Interestingly, the cells in the ΔrosR mutant were also non-motile ([44], S22A and S22B Fig) most likely due to the disability to form rod-shaped cells (S22C–S22E Fig) that is essential for H. volcanii to swim. Northern blot analysis of RNA isolated from the rosR deletion strain showed enriched cirA RNA levels compared to the wildtype (S23A and S23B Fig). This might explain the disability of ΔrosR cells to form rod shaped cells that are motile since CirA is involved in the formation of plate-shaped cells. Analysis of transcription start site data at OD₆₀₀ of 0.8 under standard growth conditions [45] revealed a transcription start site downstream of the 3’ end of arlA2 and the RosR binding site. Northern blot analysis confirmed the expression of an RNA with a size of about 420 nt (S23C Fig). We term this hitherto unannotated gene hvo_1211s. With its 420 nt the gene reaches from the intergenic region between arlA2 and cirA into the 3’ end of cirA (S24 Fig). Since the hvo_1211s RNA is complementary to almost half of the cirA RNA it has the potential to act as an anti-sense RNA to regulate cirA translation or mRNA stability. To test this hypothesis, we deleted the part of the hvo_1211s RNA in the intergenic region leaving cirA intact.

Similar to ∆cirA, partial deletion of hvo_1211s showed a hypermotile phenotype compared to the wild type (Fig 6A and 6B). However, unlike in the cirA deletion strain cell shape transition during growth was not affected (Fig 6C and 6D). Overexpression of the full length hvo_1211s blocked rod formation during early growth and hampered motility in the wildtype and partially in the hvo_1211s deletion strain (Fig 6E and 6F), similar to the rosR deletion strain (S22 Fig). Hence, the strain carrying a deletion of rosR and the strain with hvo_1211s overexpression show a comparable phenotype suggesting an inhibitory function of RosR on hvo_1211s transcription that in turn might have a promoting effect on cirA transcript abundancy as detected by Northern blotting (S23A and S23B Fig). Additionally, to exclude that the hypermotility phenotype of ∆cirA is a result of the partial deletion of hvo_1211s, we generated a partial deletion of cirA leaving hvo_1211s full sequence intact. The partial cirA deletion showed the same phenotype as the full cirA knock-out. Cells were rod-shaped throughout the complete growth cycle and hypermotile (S25A–S25D Fig). This implies that the cause for the hypermotility of ΔcirA is its constant rod-shaped cell shape whilst the cause of the hypermotility of Δhvo_1211s might be a direct impact on the archaellum operon since the deletion strain kept its cell shape like the wild type. Taken together, it is possible that hvo_1211s might act as a regulatory RNA that impacts archaellum genes thereby adding an additional layer of post-transcriptional regulation on cell shape and motility.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Motility of H. volcanii is suppressed by a regulatory RNA.

(A) Motility assay comparing wild type H26 with the hvo_1211s deletion strain either transformed with pTA1392 or a plasmid for overexpression of hvo_1211s. An exemplary motility plate is shown. Scale bar 2 cm (B) The area of the motility halos was measured and normalized to the average area of wild type halos showing significantly (p ≤ 0.05) increased motility in the hvo_1211s deletion strain compared to the H26 control. Overproduction of hvo_1211s suppresses motility in H26 significantly compared to the wild type control (p ≤ 0.05) and fully complements the hypermotility phenotype of the Δhvo_1211s strain back to wild type levels. Samples were measured in biological and technical triplicates and all single data points per strain were plotted. The middle line indicates the mean and the upper and lower line the standard deviation. (C) Cell shape analysis of the wild type H26 + pTA1392 and the hvo_1211s + pTA1392 deletion strain at low, mid and stationary OD₆₀₀. The hvo_1211s deletion strain showed the same phenotype as the control strain. Scale bar 4 µm. (D) Cell shape was analyzed using MicrobeJ and the summarized results of three independent biological replicates per strain and OD₆₀₀ value plotted as violin-plots. The median is indicated by the middle black line, dotted lines indicate the first and third quartile. For each condition more than 1000 cells were analyzed. The dotted line indicates the length below which cells are considered plate shaped. (E) Cell shape analysis of H26 with and without an hvo_1211s overexpression plasmid at low, mid and stationary OD₆₀₀. Overexpression of hvo_1211s inhibits rod formation of the cells. Scale bar 4 µm. (F) Cell circularity was analyzed using MicrobeJ and the summarized results of three independent biological replicates per strain and OD₆₀₀ value plotted as violin-plots. The median is indicated by the middle black line, dotted lines indicate the first and third quartile. For each condition more than 1000 cells were analyzed.

https://doi.org/10.1371/journal.pgen.1012198.g006

3.1 Summary

This study identifies CsmR as a novel transcriptional regulator integrating motility and cell shape control in H. volcanii. Our findings establish CsmR as essential for archaellation, connecting its function to both the transcriptional regulation of motility genes and the morphological transitions during growth. Unlike previously described archaellum regulators, which typically activate archaellation under favorable conditions [24,25], CsmR deletion leads to a paradoxical upregulation of archaellum and chemotaxis genes, despite the complete loss of motility. This paradox is best explained by the integration of ChIP-seq, transcriptomic and genetic data, which show that CsmR binds directly to promotor regions of archaellum-associated genes, while the increased transcript levels observed in ΔcsmR most likely reflect regulatory network effects involving the CirA axis rather than direct transcriptional repression by CsmR itself. Importantly, while motility and cell shape are functionally linked in H. volcanii under standard growth conditions, archaellation and cell morphology are not obligatorily coupled processes across haloarchaea. Reports from other species demonstrate that round or plate-shaped cells can retain archaellum filaments [46] and the relationship between cell shape and archaellation described in this study should therefore be understood as reflecting the specific biology of H. volcanii rather than a general principle of haloarchaeal motility.

3.2 Transcriptional control and evolutionary context

A defining feature of H. volcanii archaellum regulation is its dispersed operon organization, where motility and chemotaxis genes are interspersed and transcribed in opposing directions [10]. This contrasts with the clustered archaellum loci found in most Euryarchaeota, likely necessitating a complex regulatory framework. The extensive overlap of CsmR and CirA regulons supports a model in which CsmR acts as a central regulator, while CirA fine-tunes motility responses. The deletion of cirA results in hypermotility and a prolonged rod morphology ([33], our data) whereas csmR deletion eliminates motility despite transcriptional upregulation of archaellum genes. The csmR/cirA double deletion has the same phenotype as the ∆csmR mutant, reinforcing the hypothesis that CsmR occupies a dominant regulatory position, with CirA acting as a modulator rather than an independent transcriptional activator.

These findings parallel the multi-layered regulatory strategies observed in Thermoproteota, such as S. acidocaldarius, where motility control involves both transcriptional and post-translational mechanisms. In S. acidocaldarius, archaellation is regulated by ArnA and ArnB, which are modulated by phosphorylation, enabling rapid environmental responsiveness [14,16]. The potential for post translational regulation within the H. volcanii archaellum regulatory network is further supported by the observation that KaiC-like proteins are directly integrated into the archaellum machinery itself: ArlH, an essential archaellum motor component [7,8] is a KaiC-like protein whose ATPase and autophosphorylation activities are critical for motor assembly and function. This establishes a precedent for KaiC family proteins acting as regulatory switches within the archaellum system specifically. CirA, also a KaiC homolog, may therefore exert analogous post translational control over CsmR function, potentially through phosphorylation of the unstructured C-terminal region of CsmR, reminiscent of the phosphorylation-based regulation of archaellum regulators ArnA and ArnB in Sulfolobus [14–16]. Under this model CirA would function not merely as a transcriptional modulator but as post translational switch that fine tunes CsmR activity in response to growth phase signals adding an additional layer of regulation to the CsmR/ CirA axis beyond what is captured at the transcriptional level alone.

Additionally, the structural similarity of CsmR to Lrp/AsnC-type regulators - widely conserved in archaea and bacteria - suggests evolutionary parallels with bacterial motility regulators, which integrate transcriptional control with metabolic adaptations [47,48]. This is supported by a recent study already linking transcriptional regulation to metabolic control in H volcanii, revealing an interplay between cellular development and central metabolism. In this study TrmB and TbsP were identified as a key transcriptional regulators that influence motility and biofilm formation by controlling the glucose metabolism through co-regulation of gapII expression [49]. Furthermore, bacterial motility control involves both transcriptional and post-transcriptional regulation through flagellar-specific sigma factors, small RNAs, and kinases ensuring precise control over flagellum assembly, sometimes even morphology-dependent [50–55]. The identification of hvo_1211s includes a regulatory RNA in the archaellum and cell shape regulatory network of H. volcanii. Through its overlap with cirA - initially thought to act as an anti-sense RNA for cirA blocking its translation - our data suggest that it rather supports cirA expression resulting in increased levels of CirA that lead to the formation of plate shaped cells and the reduction of motility that is tightly coupled to cell shape. Thereby hvo_1211s might shield the cirA mRNA from endonucleolytic cleavage stabilizing the 3’ end of its target by interaction with their complementary parts supporting mRNA stability, a mechanism described for a regulatory sRNA in the archaeon Methanosarcina mazei as well [56].

Several environmental factors are known to influence cell shape and motility in H. volcanii and related haloarchaea, though their molecular intersection with the CsmR regulatory network remains to be established. Lower growth temperatures induce upregulation of archaellins in H. volcanii [27], suggesting that temperature-dependent signals feed into the archaellum regulatory network at some level. Salt concentration has been shown to influence archaellum filament composition in Haloarcula marismortui [26], pointing to osmotic conditions as an additional environmental input. Most directly relevant to the present study, the rod to plate transition in H. volcanii is triggered by increasing population density through a quorum sensing mechanism [30], which - as discussed above - likely intersects with the CsmR/CirA axis at the level of shared downstream targets. How temperature, nutrient availability and osmotic signals are sensed and transmitted to CsmR or its regulatory partners remains unknown but the Lrp/AsnC structural classification of CsmR is notable in this context, as members of this family in other organism are known to integrate metabolic and environmental signals into transcriptional responses [38,39].

The co-regulation between motility and cell shape in H. volcanii further emphasizes the importance of CsmR. The persistent rod morphology of ΔcirA cells and the normal shape transitions of ΔcsmR cells indicate that CirA suppresses CsmR-driven motility and shape-determining pathways. This is consistent with the broader picture of morphological control in H. volcanii, where dedicated shape determining factors such as RdfA, which drives rod formation, and Sph3, which promotes transition to plate morphology [32], act within regulatory landscape that CsmR appears to interface with as evidence by the upregulation of rdfA and sph3 in ΔcsmR cells, observed in our transcriptomic data. The coregulation of these shape determinants alongside archaellum genes further supports the view that CsmR coordinates morphology and motility as coupled outputs rather than independent processes in this organism. The upregulation of pilin genes in both ΔcsmR and ΔcirA mutants suggests regulatory crosstalk between archaella and type IV pili, potentially allowing cells to compensate for disrupted motility by shifting toward surface-associated behaviors [33,57]. Similar compensatory mechanisms have been observed in bacteria, where flagellar defects trigger alternative motility or adhesion pathways [58]. Additionally, ArlI and ArlJ have been implicated in pilus-dependent motility, hinting at deeper connections between archaellation and type IV pilin systems [33]. The ability of ArlI and ArlJ mutations to suppress pilin-related motility defects suggests a compensatory mechanism that fine-tunes appendage function, potentially through regulatory crosstalk with CsmR and CirA. Understanding how these systems interconnect will be critical in defining the full regulatory architecture of H. volcanii motility.

3.3 Model: CsmR as a central regulator of archaellation

The central paradox of this study is that deletion of CsmR abolishes motility while simultaneously increasing transcript levels of archaellum-associated genes. Integration of ChIP-seq, transcriptomic and genetic data suggests that CsmR binds directly to promoter regions of key motility genes, whereas the elevated expression observed in ΔcsmR is more likely explained by indirect regulatory effects within the broader CsmR/CirA network.

Our findings suggest that CsmR serves as a central regulatory node within the archaellum regulatory hierarchy. The integration of ChIP-seq and RNA-seq data allows us to directly address whether these transcriptional changes reflect direct derepression of CsmR target genes or indirect compensatory responses arising from disrupted motility or signaling. Genes associated with motif-containing, strand-consistent, promotor proximal CsmR peaks, including the entire archaellum operon and chemotaxis associated genes, showed the strongest and most consistent upregulation in ΔcsmR across all growth phases, establishing them as high confidence direct transcriptional targets. In contrast genes lacking detectable CsmR binding showed little or no expression change arguing against a broad compensatory response as the primary driver of the observed transcriptional upregulation. We therefore conclude that the transcriptional changes observed in the csmR deletion predominantly reflect direct derepression of CsmR bound promotors rather than indirect effects of disrupted motility or downstream signaling. Based on the position of the CsmR binding motif upstream of BRE and TATA promoter elements in CsmR-targets, it can be inferred that CsmR acts as a positive regulator under standard early growth conditions, potentially enhancing TFB recruitment to ensure proper transcription of all necessary components for archaellum assembly. This positive regulation is counter-balanced by CirA, which functions as a strong local repressor to fine-tune the system for phase-specific regulation.

Deletion of CirA results in hypermotility and upregulation of overlapping regulons, suggesting that CirA prevents overactivation of motility pathways by CsmR. In the absence of CsmR, CirA expression is reduced, which might derepress its targets further. We additionally propose that CirA or downstream effectors of the CsmR/ CirA regulatory axis may act at a post-transcriptional or post-translational level to block functional archaellum filament assembly in the absence of CsmR, preventing motility despite elevated transcript levels. Under this model the transcriptional upregulation of archaellum genes in ΔcsmR reflects a decoupling of transcriptional output from assembly competence which is normally ensured by the coordinated activity of CsmR and CirA together. The ∆csmR ∆cirA double mutant phenotype, resembling ∆csmR alone, supports the idea that CsmR occupies a dominant position in the regulatory hierarchy, while CirA functions as a subordinate regulator. Overexpression of CsmR in this model overrides CirA’s repressive influence, resulting in archaellated constantly rod-shaped strains. CirA’s mRNA is influenced by the regulatory RNA hvo_1211s stabilizing it and hvo_1211s in turn is negatively regulated by RosR binding upstream of its promotor region fine tuning the interplay between cell shape transition and motility (Fig 7).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Archaellum and cell shape regulation in Haloferax volcanii.

CsmR (purple) acts as central regulatory node within the archaellum regulatory network working as positive regulator for cell locomotion during early-stage growth when cells are motile and rod shaped. CirA (blue) counterbalances the positive regulation of CsmR during early growth, fine tuning archaellum expression and rod shape formation. CirA biosynthesis is supported by the RNA transcript hvo_1211s (green) stabilizing the cirA mRNA by interaction through their homologous regions. Hvo_1211s is regulated by RosR (yellow) that binds to the 3’-end of arlA2 (yellow box) blocking transcription of hvo_1211s. The rod to plate transition in H. volcanii is triggered by a quorum sensing signal [30]. A functional intersection between the CsmR/CirA axis and the quorum sensing output pathway is proposed based on phenotypic overlap, though direct molecular evidence for this connection remains to be established. (Created in BioRender. Albers, S. (2026) https://BioRender.com/7gsc8mr).

https://doi.org/10.1371/journal.pgen.1012198.g007

While ΔcsmR cells retain the ability to undergo the rod to plate transition during growth, they display a consistent shape phenotype: cells are more elongated than wildtype cells at early and mid-growth phases indicating that even in the absence of CsmR the shape of rod phase cells is affected. This demonstrates that CsmR does contribute to cell shape determination in the knockout background, albeit in a subtler manner than the complete shape lock observed under overexpression conditions. We propose that this asymmetry reflects the architecture of the underlying regulatory network. In the absence of CsmR, the CirA/ hvo_1211s/ RosR axis remains intact and capable of driving the rod to plate transition in response to increasing population density, thus preserving the overall shape transition program despite the loss of CsmR. The elongated rod morphology of ΔcsmR cells during early and mid-growth likely reflects dysregulation of shape determination factors, as supported by the upregulation of rdfA and sph3 observed in our transcriptomic data at these growth stages. The upregulation of the rod promoting factor RdfA in particular may explain the more elongated rod morphology, as increased RdfA activity could drive exaggerated rod formation before the transition to plate shape is triggered. Under CsmR overexpression, elevated CsmR activity sustains expression of rod promoting factors and simultaneously overrides CirA mediated repression, locking cells in rod morphology regardless of growth phase and cell density. Taken together, these observations reveal a dosage dependent role for CsmR in cell shape determination: at normal expression levels CsmR fine tunes the dimensions of rod-shaped cells during early growth, while at elevated levels it overrides the shape transition machinery entirely. CsmR therefore contributes to cell shape regulation primarily through its influence on the balance of rod- and plate-promoting factors within the broader CsmR/ CirA regulatory network, rather than as a dedicated morphological regulator. Regarding the intersection between the CsmR/ CirA network and the quorum sensing system, we note that the rod to plate transition in H. volcanii is triggered by quorum sensing signals that accumulate with increasing population density [30]. The observation that CsmR overexpression and ΔcirA both result in persistent rod morphology across all growth phases, phenocopying conditions under which quorum sensing driven shape transitions are prevented, suggests functional intersection between the CsmR/ CirA regulatory axis and the quorum sensing output pathway. However, we emphasize that no direct molecular connection between CsmR, CirA or RosR and the quorum sensing machinery has been demonstrated in this study and the nature of this potential intersection remains speculative.

This hierarchical model is consistent with growth-phase-specific regulation, that is sensed by a recently described two-component system [59], enabling flexible fine-tuning of activation and repression. It also explains the stage-dependent repression of archaella-related genes, such as Sph3 and RdfA, which contribute to shape-dependent changes. The potential layered regulation by the TFB variant (HVO_0226), CsmR, CirA, hvo_1211s, RosR and two additional transcriptional regulators (HVO_A0465, HVO_2507) highlights the complexity of transcriptional control required for precise coordination of cell shape determination motility and archaellum assembly in H. volcanii.

3.4 Future perspectives and open questions

Our findings establish CsmR as a central transcriptional regulator coordinating motility and cell shape transitions in H. volcanii, yet critical questions remain. The identification of promotor-proximal ChIP-seq peaks together with a conserved consensus motif strongly supports direct DNA binding and transcriptional regulation, although the precise regulatory consequences of this binding remains to be fully resolved. Whether CsmR functions exclusively as a transcriptional activator, as a context-dependent dual-function regulator or through indirect mechanisms that vary across growth phases remains an important open question that will require further investigation. The interplay between CsmR and CirA suggests a layered regulatory system, raising the possibility that CirA modulates CsmR post-translationally, perhaps through phosphorylation. Given that phosphorylation-based archaellum regulation is well-documented in Thermoproteota, a similar mechanism in H. volcanii would suggest convergent strategies across archaeal lineages [60]. Beyond transcriptional control, the environmental factors governing CsmR activity remain unclear. Archaellation in H. volcanii is influenced by temperature and nutrient availability, suggesting that CsmR integrates external signals to balance motility and morphological transitions [27]. The connection between CsmR and other transcriptional regulators, including alternative TFBs and RosR, should be explored to map its broader regulatory network. Evolutionarily, the structural similarity between CsmR and bacterial Lrp/AsnC regulators highlights potential conserved functions linking motility and metabolism. Moreover, quorum sensing in H. volcanii, controlling cell morphology and motility by a small secreted molecule has recently been discovered [30]. However, the molecular mechanism by which this quorum sensing signal intersects with the transcriptional network described here, including CsmR, CirA and RosR, remains unknown and represents a priority for future investigation. The upregulation of pilin genes in ΔcsmR and ΔcirA mutants further suggests regulatory crosstalk between archaella and type IV pili systems, pointing to a more integrated control of cell appendages.

By defining CsmR’s role within the archaeal transcriptional landscape, this study provides a foundation for future work on regulatory networks governing motility, environmental sensing, and cellular architecture, with broader implications for the evolution of prokaryotic transcriptional control.

4.1 Growth and media

Escherichia coli strains for cloning were grown in LB-medium [61] at 37°C. If necessary, LB-medium was supplemented with 100 µg/ml Ampicillin.

For growth of Haloferax volcanii either non-selective YPC-medium or selective Ca-medium [62] was used. For cell shape analysis, CA-medium was supplemented with a trace element solution [29] to improve cell shape homogeneity. Cultures smaller than 5 ml were grown in 15 ml tubes while rotating. Larger cultures for microscopy were grown in 20 ml medium in 100 ml flasks under constant agitation at 120 rpm. Growth temperature for H. volcanii was 45°C. Plates were prepared as described before [62]. Motility plates were exactly prepared like described by Patro et al, [63].

For growth quantification of the csmR deletion strain cells were grown in 20 ml Cab medium and growth was continuously measured using the cell growth quantifier (CGQ, Software (CGQuant 7.4)) from Aquila Biolabs GmbH. Cell were grown at 45 °C in biological triplicates and H26 was used as control.

Archaeal and bacterial strains used in this study are listed (S1 Table).

4.2 Cloning

For plasmid construction all inserts were amplified using the PhusionHigh-Fidelity DNA polymerase obtained from New England Biolabs (NEB) following the manufactures protocol. Plasmids were either assembled via restriction enzyme-based cloning or in vivo ligation [64]. Enzymes for cloning were also obtained from NEB and the manufacturers protocol was followed. For in vivo ligation respective plasmids were linearized via polymerase chain reaction and subsequently cleaned by passing the linearized plasmids through an agarose gel. Approximately 50 ng linearized plasmid was mixed with app. 100 ng insert, the mixture transferred to 50 µl chemically competent E. coli and cells were then stored on ice for 30 min. Next, cells were heat shocked at 42°C for 1 min, 450 µl LB-medium was added and the cells incubated for 1.5 h at 37°C under constant shaking. Afterwards the whole cell suspension was plated on LB-plates supplemented with 100 µg/ml Ampicillin and the plates incubated overnight at 37°C. Colonies were transferred to 5 ml LB-medium supplemented with Ampicillin, grown over night and screened via colony PCR for successful plasmid generation the next day. Possible correct assembled plasmids were isolated using the Nucleospin Plasmid Easy Pure Kit (Macherey-Nagel) and send for sequencing. Used plasmids and primers are listed (S2 and S3 Tables).

4.3 Transformation in Haloferax volcanii

For the genetic modification of Haloferax volcanii, cells were transformed using polyethylene glycol [62]. Respective H. volcanii strains were grown in 10 ml YPC medium over-night at 45°C under constant rotation. The next day cells were harvested and resuspended in 2 ml buffered spheroplasting solution (1 M NaCl, 27 mM KCl, 50 mM Tris HCl pH 8.5, 15% sucrose (sterile filtered)), transferred to a round bottomed 2 ml reaction tube and centrifuged for 8 min at 3380 g. Supernatant was discarded and the pellet resuspended in 600 µl buffered spheroplasting solution. For one transformation 200 µl of the cell suspension was used. To generate spheroplasts 50 mM ethylenediaminetetraacetic acid (EDTA, pH 8) was added, the tube inverted and incubated for 10 min at room temperature. During incubation 1 µg of demethylated plasmid was mixed with 5 µl EDTA (0.5 M, pH 8) and then filled up to a total volume of 30 µl with unbuffered spheroplasting solution (1 M NaCl, 27 mM KCl, 15% sucrose (sterile filtered)). After 10 min the DNA mixture was added to the tube, mixed by inversion and incubated for 5 min. Subsequently 250 µl 60% polyethylene glycol 600 (diluted in unbuffered spheroplasting solution) was added to the cells and the tube was shaken horizontally for mixing. Thirty minutes later 1.5 ml spheroplast dilution solution (23% salt water, 15% sucrose, 3.75 mM CaCl₂ (sterile filtered)) was added to the transformants, incubated for 2 min following centrifugation at 3380 g for 8 min. The supernatant was discarded and 1 ml regeneration solution (18% salt water, 1 x YPC-solution, 15% sucrose, 3.75 mM CaCl₂ (sterile filtered)) added whilst the cell pellet was scratched off the wall. Transformants were then incubated at 45°C for 1.5 h without any agitation. Subsequently, cells were resuspended by tapping the tube and incubated for another 3 h at 45°C under constant rotation. The regenerated transformants were then harvested at 3380 g for 8 min, the supernatant discarded and the cell pellet resuspended in 1 ml transformant dilution solution (18% salt water, 15% sucrose, 3.75 mM CaCl₂ (sterile filtered)). 100 µl of the resuspended cells were plated on selective Ca-plates and plates incubated in an airtight container at 45°C until colonies were visible.

4.4 Gene deletion in Haloferax volcanii

For the generation of deletion strains knock-out plasmids (S2 Table) of the genes to be knocked-out were transformed in the respective background strains as described above. Once colonies were visible after transformation, one colony per strain was transferred into 5 ml of non-selective YPC medium and grown over night at 45°C under constant rotation. The next day the growing culture was diluted back 1:500 into fresh 5 ml YPC-medium and grown over night. This process was repeated to a total of three times. Subsequently three dilutions of the growing culture, starting from 10^-2, were plated on selective Ca-plates supplemented with 50 µg/ml 5-FOA and 10 µg/ml Uracil. Plates were incubated in an airtight container at 45°C until colonies were visible. To screen possible deletion strains 100 colonies per knock-out attempt were streaked on a YPC-plate and grown for two days at 45°C. Colony PCR was then used to screen for successful deletions using primers indicated in S3 Table.

4.5 Light microscopy

In order to examine the effect of the absence or overexpression of csmR on the cell shape of Haloferax volcanii cells, phase-contrast microscopy was used. Control strain H26 and the csmR deletion strain HTQ289 were either transformed with an “empty” expression plasmid (pTA1392) to complement the auxotrophic marker used in the genetic system or with the csmR overexpression plasmid pSVA13922. For microscopy, precultures were inoculated in 5ml Cab-medium and grown overnight at 45°C under constant rotation. The next day cells were diluted back into 20 ml fresh Cab-medium and grown to specific optical densities (OD₆₀₀) under constant shaking at 45°C. When cells reached the desired ODs 5 µl cell culture was spotted on 1% agarose pads dissolved in 18% salt water. For cell shape analysis Fiji (Version 2.14.0) [65]plug-in MicrobeJ (Version 5.13l) [66]was used. Outliers in the cell shape datasets were identified by the ROUT method (Q = 1%) and removed. Cells of the cirA, cirD, rosR, regulatory RNA deletion strains as well as the double deletion strains of csmR cirA and csmR cirD were exactly prepared as the csmR single deletion strain.

Statistical comparison of cell shape parameters between the control and mutant populations were performed using Welch’s two-sample t-test which accounts for unequal variance and unequal sample sizes. Effect sizes were quantified using Cohen’s d.

Due to the large sample sizes employed in this study, Welch’s t-test yields extremely high statistical power, rendering virtually any difference between populations statistically significant even when the magnitude of that difference is negligible. Under such conditions, p-values alone are insufficient to assess the biological relevance of observed differences [67,68]. Accordingly, significance annotations have been omitted from cell shape plots to avoid overstating the biological importance of statistically significant but practically trivial differences. Instead, effect sizes (Cohen’s d) are reported alongside p-values in S4 Table to provide a comprehensive and transparent account of both, the statistical significance and the practical magnitude of each comparison. Effect sizes were interpreted according to conventional thresholds: d < 0.2 as negligible, 0.2 ≤ d ≤ 0.5 as small, 0.5 ≤ d ≤ 0.8 as medium and d ≥ 0.8 as large [69].

Summarized statistical analysis are shown in S4 Table.

4.6 Motility plates

To assess the effect of the different deletion strains on the ability of cell locomotion semi solid Ca-agar (0.3% w/v) plates were prepared as described above. Control strain H26 and the deletion strains were either transformed with pTA1392 or the indicated plasmids for overexpression of csmR or the sRNA. One colony of each transformed strain was inoculated in 5 ml Cab-medium and grown over night at 45°C under constant rotation. The next day the OD₆₀₀ of the growing cultures was set to 0.2 by dilution with Ca-medium. Sterile toothpicks were used for inoculation of the motility plates. Plates that harbored strains containing the overexpression plasmid were supplemented with 12 mM Xylose, if not stated differently, to induce expression. Inoculated plates were airtight packed into plastic bags and incubated at 45°C for 3–4 days. Subsequently plates were scanned and the area of motility halos measured using Fiji. Statistical analysis was done using an unpaired Welch’s t-test.

4.7 Electron microscopy

Since the csmR deletion strain is non-motile electron microscopy was used to assess if archaella were present in the mutant strain. To ensure that the observed cell appendages on Haloferax volcanii were solely archaella a pilB3 deletion strain as control and a pilB3 csmR double deletion strain was used. Both strains contained plasmid pTA1392 and were grown in 5 ml Ca-medium at 45°C, rotating. The next day the cultures were diluted back into 20 ml Ca-medium and grown to an optical density of 0.01 at 45°C under constant shaking. Cells were harvested and resuspended to a theoretical OD₆₀₀ of 20 in Ca-medium. Next, 5 µl of cells were spotted on a glow discharged carbon coated copper grid (Plano GmbH, Wetzlar, Germany) and incubated for 10 sec. Excess liquid was removed using blotting paper and the cells were stained by addition of 2% uranyl acetate (w/v). Imaging was done with Hitachi HT7800 operated at 100 kV, equipped with an EMSIS Xarosa 20-megapixel CMOS camera. The effect of cirA and cirD deletion on archaellum formation was also assessed in a pilB3 deletion background as described for the csmR deletion. Since the cirA deletion strain showed hypermotility, the cirA pilB3 deletion strain was also imaged during stationary phase to see if archaella were still assembled. Each strain was imaged in biological triplicates.

4.8 Lipid extraction

Lipid were extracted from 3-14 mg freeze-dried Haloferax volcanii cells (variation caused by high salt levels) by applying an acidic Bligh and Dyer protocol, containing 5% trichloroacetic acid (TCA), as described elsewhere [70,71]. In short, freeze-dried cells were resuspended in 2:1:0.8 (v/v) MeOH:CHCl₃:5% TCA, spiked with 10 µg of the internal standard n-Dodecyl β-maltoside (DDM), vortexed and sonicated for 60 min. Phase separation was obtained by addition of CHCl₃ and 5% TCA. After centrifugation (2500 x g, 5 min at RT), the bottom phase was collected. To ensure total lipid extraction this procedure was repeated two times more. To remove TCA contaminations, the recovered bottom-phase was washed by adding 1:1:0.9 (v/v) MeOH:CHCl₃:MQ. After centrifugation (2500 x g, 5 min at RT), the bottom phase was collected, dried under a nitrogen (N₂) stream, and resuspended in 50 µl MeOH.

4.9 LC-MS analysis of lipids

Samples were analyzed using an Agilent Technologies 1290 Infinity high-performance liquid chromatography (HPLC) system, coupled to a heated electrospray ionization–mass spectrometer (Thermo Q Exactive Plus; Thermo Fisher Scientific). 5 µl was injected into an ACQUITY UPLC CSH C18 1.7 µm Column, 2.1 × 150 mm (Waters Chromatography Ireland Ltd) operating at 55 °C with a flow rate of 300 µl/min. Separation of the compounds was achieved by a changing gradient of eluent A (5 mM ammonium formate in water/acetonitrile 40:60, v/v) and eluent B (5 mM ammonium formate in acetonitrile/2-propanol, 10:90, v/v). The following linear gradient was applied: 1) 5% eluent B for 2.5 min, 2) a gradient from 5% to 90% eluent B over 36.5 min, 3) holding for 3 min, 4) returning to 5% eluent B in 0.5 min, 5) and holding for 8 min. The column effluent was injected directly into the Thermo Q Exactive Plus operating in positive and negative ion mode. Settings for positive mode: Spray Voltage: 3.20 |kV|, Capillary temperature: 230 °C, S-lens RF level: 50.0, sheath gas flow: 30, auxiliary gas flow: 5. Settings for negative mode: Spray Voltage: 3.20 |kV|, Capillary temperature: 230 °C, S-lens RF level: 50.0, sheath gas flow: 30, auxiliary gas flow: 20, sweep gas flow: 3, Aux gas heater temperature: 380 °C.

Spectral data constituting total ion counts were analyzed using the MacCoss Lab Software: Skyline. The following transition settings were used: Scan range: 133.4-2000 m/z, MS1 filtering: Orbitrap, resolving power 70,000 at m/z 400. MS/MS filtering: DDA, Orbitrap, resolving power: 17,500 at m/z 400. Total ion counts for extracted lipid species (S5 Table) were initially corrected for both the internal standard DDM (m/z 509.30 [M-H]− or m/z 528.34 [M + NH4]+) and the OD (due to high salt concentrations the freeze-dried mass could not be used). To further improve sample comparability, lipid species were normalized for total lipid ion count per sample and subsequently plotted on the y-axis.

4.10 RNA sequencing for differential gene expression analysis

4.11 Chromatin immunoprecipitation sequencing (ChIP-seq)

4.12 Northern blot analysis

Total RNA was prepared from H. volcanii strain as described for the transcriptome analysis. Cells were harvested at exponential phase (OD₆₅₀ = 0.5) and at stationary phase (OD₆₅₀ = 1.1-1.4). In addition, total RNA was isolated using the RNeasy Plus Kit (Quiagen) from the following strains: three biological replicates of H26 and ∆rosR, in each case at an OD₆₀₀ of 0.02 and of 0.2. From all different RNA preparations, 10 µg each were separated on 1.5% agarose-gel, which were subsequently transferred to a nylon membrane (Biodyne A, PALL). After transfer, the membrane was hybridized with specific oligonucleotides (primer sequences are listed in S3 Table), which were radioactively labelled with [γ-³²P]-ATP via polynucleotide kinase treatment to detect the hvo_1211s, cirA or 5S transcript. The signals were recorded by an autoradiography film. For quantification of the cirA signal, an additional imaging plate (FujiFilm) was used for recording and the intensities were analysed using the ImageLab Software (Bio-Rad). The cirA signal was normalized with the 5S signal, as a loading control.