Specific role of two NlpC/P60 endopeptidases in cell division and membrane vesicle formation in Deinococcus radiodurans

TL;DR

This study explores how specific enzymes in Deinococcus radiodurans help with cell division and stress response, revealing their roles in cell survival and structure.

Contribution

The paper identifies and characterizes the roles of three NlpC/P60 endopeptidases in D. radiodurans, linking cell wall remodeling to genotoxic stress response.

Findings

CwlA is essential for cell viability and cell wall integrity.

CwlB is involved in cell septation, while CwlC is not required for cell shape or division.

CwlA expression is repressed under DdrO depletion, linking genotoxic stress to cell wall remodeling.

Abstract

•Three NlpC/P60 endopeptidases identified in Deinococcus radiodurans.•CwlA is essential for cell viability and cell wall integrity.•CwlB is involved in cell septation whereas CwlC is dispensable.•Regulatory link between CwlA expression and the genotoxic stress response pathway.

Three NlpC/P60 endopeptidases identified in Deinococcus radiodurans.

CwlA is essential for cell viability and cell wall integrity.

CwlB is involved in cell septation whereas CwlC is dispensable.

Regulatory link between CwlA expression and the genotoxic stress response pathway.

The bacterial cell wall is composed of peptidoglycan (PG), a sugar polymer cross-linked by short peptide stems. PG determines cell morphology and protects it from environmental stresses. Cell growth and division require a balance between synthesis and hydrolysis of the PG. One class of PG hydrolase is the NlpC/P60 superfamily which is broadly distributed in bacteria, archaea, eukaryotes and viruses. Deinococcus radiodurans, characterized by its extreme radioresistance, is a diderm bacterium with a thick layer of PG located between the inner and outer membrane. D. radiodurans exhibits three NlpC/P60 endopeptidases, their role in morphogenesis and cell cycle remains unexplored. In this work, we investigated the role of each endopeptidase in cell division to assess their specific role. Here, we showed that the CwlB protein is involved in cell division and that cwlA gene is essential for cell viability. The CwlC protein is not required for cell shape maintenance or cell division. We showed that CwlA protein is homogeneously localized around the cell except on septal region. CwlA-depleted cells lost viability, displayed morphological changes, and produced numerous membrane vesicles, similarly to cells exposed to sublethal mitomycin C (a DNA-damaging agent) or to DdrO depletion, the transcriptional repressor of the main genotoxic stress response in Deinococcus. We showed that cwlA expression is highly repressed under DdrO depletion. These data suggested a link between response pathway to genotoxic conditions and cell wall remodeling.

Image, graphical abstract

Figures

CwlA is essential for *D. radiodurans* viability. Strains were grown in liquid medium with spectinomycin at 28 °C. Sequential dilutions of cells were spotted on TGY plates in the presence or absence of spectinomycin at 28 °C (A) or 37 °C (B). Lane 1: strain GY13739 containing non-thermosensitive plasmid p11554 (p*repU*); lane 2: strain GY13747 containing thermosensitive plasmid p13841 (p*repU*Ts); lanes 3 and 4: GY18249 and GY18250 respectively, two independent clones [Δ*cwlA* (pr*epU*Ts*-cwlA*^+^)]; lane 5: strain GY19040 [Δ*cwlA* (p*repU*-*cwlA*^+^)]; lane 6: strain GY14164 [Δ*ddrO*(p*repU*T

CwlB is required for the degradation of septa. A. Microscopy analyses of stationary phase cells from wild type, Δ*cwl*B and Δ*cwl*B/*cwl*B^+^mutant strains. The different structures observed are indicated by coloured arrows: dyads in red, tetrads in green, double tetrades in blue and quadruple tetrades in yellow. B The percentage of cells found in dyads (2 cells, red), tetrads (4 cells, green), double tetrads (8 cells, blue) and quadruple tetrads (16 cells, yellow) are illustrated in the pie charts.Fig 3: dummy alt text

Formation of membrane vesicles following depletion of CwlA. Cells (∆*cwlA*/p*repU*TS::*cwlA*^+)^ in exponential growth phase cultivated at 30 °C in medium supplemented with spectinomycin were harvested by centrifugation, diluted in antibiotic-free medium and incubated at 30 °C (upper part) or at 37 °C (lower part). In each part, the first line contains the pictures of the Nomarski interference contrast (DIC), the second line, the pictures of membrane DNA staining (DAPI) and the third line the pictures of membrane staining (FM4–64). All pictures are the same scale (bar = 5 μm). Membrane vesicle

Cellular localization of the endopeptidase CwlA. Microscopy images of CwlA-neongreen cells under standard growth conditions. DIC: phase contrast; DAPI: DNA staining; FM4–64: membrane staining.Fig 6: dummy alt text

Full text

Introduction

1

The bacterial cell wall is dynamic and undergoes reorganization during vegetative growth, development, and cell division. During these processes, the wall is dismantled by a diverse set of enzymes that hydrolyze different bonds within the peptidoglycan (PG). These enzymes include glycosidases, such as lysozymes, which target the polysaccharide backbone, and peptidases which break down the cross-linking peptides. One class of PG hydrolase is the NlpC/P60 superfamily (Anantharaman and Aravind, 2003; Ishikawa et al., 1998). NlpC/P60-type endopeptidases are widely distributed among bacteria and participate in various processes such as cell growth, enabling the development of peptidoglycan through the insertion of new units, the maturation and recycling of peptidoglycan and its cleavage at the septum during cell division. Since over- or underactivity of these proteins lead to significant cellular damage, ranging from septation defects to cell lysis, the synthesis and activity of these proteins are tightly regulated throughout the cell cycle and cellular development (Vollmer et al., 2008).

The bacterium Deinococcus radiodurans is renowned for its extraordinary ability to withstand a broad spectrum of genotoxic treatments, including ionizing and ultraviolet radiation, mitomycin C, desiccation, and oxidative stress. When exposed to extreme doses of γ-radiation causing hundreds of DNA breaks, D. radiodurans reassemble its genome within 2 to 3 h. This remarkable ability is supported by highly efficient DNA repair mechanisms, as well as other survival strategies, including significant nucleoid compaction, mechanisms protecting against protein oxidation and the induction of specific proteins following irradiation (Confalonieri and Sommer, 2011; Slade and Radman, 2011; Daly, 2012; Blanchard et al., 2017; Lim et al., 2019). In Deinococcus, the main response pathway to genotoxic conditions is regulated by a constitutively expressed metalloprotease IrrE and a transcriptional repressor DdrO (Ludanyi et al., 2014; Devigne et al., 2015; Wang et al., 2015; Lu et al., 2024). Upon exposure to genotoxic stress conditions, IrrE is activated and cleaves DdrO, leading to expression of about 40 repressed genes constituting the Radiation Desiccation Response (RDR) regulon (Makarova and Daly, 2010; Eugénie et al., 2021). The ddrO gene is essential for cell viability of D. radiodurans and its prolonged depletion by a conditional deletion system induced an apoptotic-like response (DNA degradation, defects in chromosome segregation, membrane vesicle formation) (Devigne et al., 2015).

Although D. radiodurans is Gram-positive, the structural organisation of its membrane is closer to that of Gram-negative bacteria, with the presence of a thick peptidoglycan layer topped by a structurally complex layer rich in lipids, proteins and carotenoids, itself covered by a dense carbohydrate layer (Farci et al., 2022 ; Rothfuss et al., 2006; Sexton et al., 2021). The peptidoglycan has a unique composition due to the presence of l-ornithine, a very rare component only found in a few bacterial membranes (Work and Griffiths, 1968). However, D. radiodurans lacks a lipopolysaccharide (LPS) layer, which is characteristic of Gram-negative bacteria (Beaud Benyahia et al., 2025). This bacterium, and more generally the Deinococcus-Thermus phylum, represents an intermediate stage in the evolutionary transition between bacteria with a single membrane and those with two (Gupta, 2011).

Beyond its well-documented resistance to high doses of ionizing radiation and long-term desiccation, D. radiodurans has garnered significant attention in recent years in the field of astrobiology. This bacterium is one of the few microorganisms capable of surviving harsh conditions of outer space, including exposure to cosmic and solar radiation, extreme vacuum, temperature fluctuations, desiccation, and microgravity (Ott et al., 2020). Notably, D. radiodurans cells have been shown to survive for three years on the exterior of the International Space Station. The formation of membrane vesicles (MVs) is observed when the bacteria were subjected to such extreme stress (Ott et al., 2020). Bacterial extracellular vesicles are nanoscale, membrane-bound structures ranging in size from 10 to 500 nm released by both Gram-negative and Gram-positive bacteria into their surrounding environment. These MVs play a critical role in a wide range of biological processes. They are key mediators of intercellular communication by acting as molecular carriers, enabling the transfer of genetic material such as DNA and RNA. Additionally, they are involved in the transport of virulence factors, toxins, and other bioactive molecules (Brown et al., 2015; Schwechheimer and Kuehn, 2015; Toyofuku et al., 2019).

In this work, we aimed to dissect the role of three endopeptidases containing NlpC/P60 domains (CwlA, CwlB, and CwlC) by analyzing the phenotype of single or double mutants during both exponential and stationary phases. The cwlC mutant displayed no detectable phenotype. In contrast, CwlB was found to play a key role in septal degradation, highlighting its involvement in cell division. Finally, a reduction in the amount of CwlA endopeptidase was associated with the formation of extracellular MVs, indicating a potential link between CwlA activity and membrane dynamics.

Materials and methods

2

Bacterial strains, plasmids, oligonucleotides, media

2.1

Bacterial strains and plasmids are listed in Table 1, Table 2, respectively. The Escherichia coli strain DH5α was used as the general cloning host and strain SCS110 was used to propagate plasmids prior to introduction into D. radiodurans via transformation (Meima et al., 2001). All D. radiodurans strains were derivatives of the wild-type strain R1 ATCC 13939. The deletion mutants were constructed by the tripartite ligation method (Mennecier et al., 2004). An antibiotic cassette (hygromycin or chloramphenicol resistance gene) and two regions of around 500 bp genomic fragments from upstream and downstream of the coding region of target gene were amplified by PCR using primer pairs that introduced BamHI and XbaI restriction sites. The antibiotic cassette flanked by the two genomic fragments were ligated together in molecular ratio 1/1/1 (100 ng of 500 bp fragments). The constructs were then introduced into D. radiodurans by transformation selecting for antibiotic resistance. This led to the replacement of the wild-type allele by the mutant counterpart via homologous recombination. D. radiodurans contains from 4 to 10 genome equivalents. Homogenotes of the deletion allele were obtained after two cycles of purification on selective medium. The same strategy was used to construct strains expressing neon green-tagged CwlA or CwlB proteins. The genetic structure and purity of the mutants were verified by PCR (Figure S1, S2 and S9). The sequence of oligonucleotides used for strain construction and diagnostic PCR are listed in Table S1. Chromosomal DNA of D. radiodurans was extracted using the NucleoSpin DNA Microbial Mini kit (Macherey-Nagel). PCR amplification of DNA fragments, using plasmid or genomic DNA as a template, was performed using Phusion DNA polymerase (Thermo Scientific) or GoTaq Flexi G2 (Promega).Table 1. Bacterial strains used in this study. ^a^: strains were constructed by the tripartite ligation method.Table 1: dummy alt textStrainsDescriptionsource or referenceE. coliDH5αsupE44ΔlacU(Φ80lacZΔM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1laboratory stockSCS110**endA dam dcm supE44 Δ(lac-proAB) (F’traD36 proAB lacI^q^ZΔM15)laboratory stockD. radioduransR1 / GY9613ATCC 13939laboratory stockGY13739R1/p11554laboratory stockGY13747R1/p13841laboratory stockGY14164∆ddrO Ωcat / p11891 (prepUTs::ddrO^+^)Devigne et al.(2015)GY18246∆cwlAΩhph non-homogenotizedThis work ^a^GY18249∆cwlAΩhph/p17263 (prepUTs::cwlA^+^) clone 1This work ^a^GY18250∆cwlAΩhph/p17263 (prepUTs::cwlA^+^) clone 2This workGY18875cwlA::DrNeongreenΩhphThis work ^a^GY19037∆cwlB ΩcatThis work ^a^GY19042∆cwlBΩcat/ p18736 (cwlB^+^)This workGY19040∆cwlAΩhph/ p18731 (prepU::cwlA^+^)This workGY19044∆cwlC ΩhphThis work ^a^GY19047∆cwlB Ωcat ∆cwlC ΩhphThis workGY19075∆cwlB Ωcat ∆cwlC Ωhph/ p18736 (cwlB^+^)This workGY19094cwlB::DrNeongreenΩhphThis work^a^Table 2. Plasmids used in this study.Table 2: dummy alt textplasmidsdescriptionreferencepPS6Source of cat cassettelaboratory stockp11554Shuttle vector E. coli/ D. radiodurans, Spc^R^laboratory stockp11830Vector thermosensitive for replication in D. radiodurans, Spc^R^, prepUTsNguyen et al. (2009)p12625Source of hph cassettelaboratory stockp13841p11830 P_Spac_-term 116Nguyen et al. (2009)p11891p13841: prepUTs::ddrO^+^Devigne et al.(2015)p17235p11554 BamHI/BglII + multiple cloning sitethis workp17263p13841 NdeI /XhoI+ drO_1728this workp17291pEX-128 + DrNeongreenEurofinsp17292p17291 AgeI/XhoI + hph cassette, source of DrNeongreen-hphthis workp18731p17235 NdeI/XhoI + PcwlA-cwlAthis workp18736p17235 BamHI/XbaI + PcwlB-cwlBthis work

D. radiodurans strains were grown at 30 °C (or 28 °C, 37 °C) in TGY2X (1 % tryptone, 0.2 % dextrose, 0.6 % yeast extract), or plated on TGY1X containing 1,5 % agar, and E. coli strains were grown at 37 °C in LB (Lysogeny Broth). When necessary, media were supplemented with the appropriate antibiotics used at the following final concentrations: chloramphenicol 3 µg/mL; spectinomycin 75 µg/mL, hygromycin 50µg/mL for D. radiodurans and 40 µg/mL of spectinomycin for E. coli.

Assay of genes essentiality

2.2

The essentiality of genes was evaluated in a growth experiment, in which the strains grown at 28 °C in liquid medium with spectinomycin were serially diluted, plated on TGY agar and incubated at 28 °C or 37 °C in the presence or the absence of spectinomycin.

CwlA and Ddro depletion

2.3

ΔcwlA and ΔddrO strains complemented by cwlA or ddrO, respectively, expressed from a plasmid with thermosensitive replication (prepUTs) were grown at permissive temperature (30 °C) with selective antibiotics (spectinomycin and hygromycin or choramphenicol), until they reached an A_650nm_ 0.25∼0.3. Cells were harvested by centrifugation and resuspended at the previous A_650nm_ levels in fresh medium without antibiotics and grown at permissive (30 °C) or non-permissive (37 °C) temperature. At 24 h, aliquots were removed for further analyses.

Fluorescence microscopy

2.4

Cell membranes were stained with N-(3-triethylammoniumumpropyl)-4-(6-(4-(diethylamino)phenyl)hexatrienyl) pyridinium dibromide (FM 4–64) at 0.01 mg/mL, and the nucleoid with 4,6-diamidino-2-phenylindole dihydrochloride (DAPI) at 2 mg/mL. FM 4–64 stains the lipid membranes with red fluorescence (excitation/emission ∼515/640 nm) and DAPI stains the nucleoid with blue fluorescence (excitation/ emission ∼350/470 nm). The stained cells were observed using a Leica DM RXA microscope. Images were captured with a CDD camera 5 MHz Micromax 1300Y (Roper Instruments).

Transmission electron microscopy (TEM)

2.5

Cells were fixed in 0.1 M cacodylate pH 7.4 buffer containing 2 % paraformaldehyde and 2 % glutaraldehyde at room temperature for 1 hour. Samples were washed 5 times in 0.1 M cacodylate buffer for 5 min and post-fixed in buffer A (1 % osmium tetroxid, 1.5 % potassium ferrocyanide) at room temperature for 1 hour. Then, cells were embedded in 2 % low melting agarose and dehydrated in a graded series of ethanol-water solution (30, 50, 70, 90 %) for 10 min each. Subsequently, cells are dehydrated with HMPA 90 % (hexamethylphosphoramide) – ethanol 10 %, HMPA 95 %-ethanol 5 %, HMPA 95 % - ethanol 3 % solution for 10 min each. Thin sections of the samples were stained with uranyless followed by lead citrate. Samples were viewed using a JEOL 1400 transmission electron microscope operated at 120 kV.

RNA seq analysis

2.6

RNA-seq data published in Eugénie et al. (Eugénie et al., 2021) were processed in a similar way. Only reads corresponding to RNA strands (R2 sequences) were used for these studies. Sequence alignments on the genomic sequence were performed by Bowtie2 software and differential analysis of gene expression between ∆ddrO/prepUTSddrO^+^strain compared to the ∆ddrO/prepUddrO^+^ strain for each time point was performed using the DESeq2 package using (FC ≥2, p-value ≤0.01) as cutoff/parameters to highlight upregulated or downregulated genes.

Results

3

Sequence analysis of endopeptidase NlpC/P60 in D. radiodurans

3.1

NlpC/P60 endopeptidases belong to a cysteine peptidases family that hydrolyze peptide bonds within the peptidoglycan layer. NlpC/P60 enzymes often exhibit a modular architecture, combining the catalytic domain with additional domains such as LysM, SH3 or choline binding domain (Anantharaman and Aravind, 2003), which facilitate substrate recognition and protein localization. D. radiodurans, encodes three putative NlpC/P60 endopeptidases (Anantharaman and Aravind, 2003): DRO_1728 (DR1749), DRO_1316 (DR1325), and DRO_1350 (DR1316). We named these proteins CwlA (cell wall hydrolase A), CwlB, and CwlC, respectively. Search for known domains in the InterPro Database showed that these endopeptidases may be classified into two groups based on their N-terminal domain organization (Fig. 1A). The first group, comprising CwlA and CwlB, possesses a putative signal peptide, indicative of extracellular localization or membrane targeting, and one or two LysM domains, respectively. The LysM (Lysin Motif) domain is a small globular domain of approximately 40 amino acids, originally identified in enzymes degrading bacterial cell walls. This widespread protein module binds peptidoglycan by interacting with N-acetylglucosamine residues. The third protein, CwlC, contains two bacterial SH3-like domains but lacks the signal peptide. The bacterial SH3-like domain also interacts with peptidoglycan, typically recognizing pentapeptide cross-bridges. This non-covalent binding facilitates cell wall targeting (Kurochkina and Guha, 2012). All these three proteins harbor a C-terminal NlpC/P60 catalytic domain that hydrolyzes specific peptide cross-bridges within the peptidoglycan layer using a cysteine residue as a nucleophile. A highly conserved catalytic triad composed of cysteine, histidine, and a polar residue as asparagine, glutamine or histidine carries their enzymatic activity (Anantharaman and Aravind, 2003). CwlA, CwlB, and CwlC, as well as YdhO from E. coli and CwlT from B. subtilis, all sequences contain this catalytic triad, along with three conserved residues that contribute to the formation of catalytic core (tyrosine, aspartic acid, and serine) (Fig. 1B). Due to the conservation of catalytic sites, their modular organization and the functional roles of their respective domains, CwlA, CwlB and CwlC are likely involved in peptidoglycan remodelling in D. radiodurans, thus playing a potential role in processes such as cell wall maintenance and cell division.Fig. 1In silico analysis of NlpC/P60 endopeptidases of D. radiodurans.A. Domain architecture of the three putative endopeptidases CwlA, CwlB, and CwlC from D. radiodurans. The signal peptide (S), suggesting potential secretion or membrane targeting. LysM and SH3 domains are known to be involved in peptidoglycan binding. The NlpC/P60 domain represents the catalytic core responsible for peptidoglycan hydrolase activity. Domain positions and lengths were identified using the InterPro database (Blum et al., 2025).B. The NlpC/P60 catalytic domains of YdhO from E. coli, CwlT from B. subtilis, and CwlA, CwlB, and CwlC from D. radiodurans. These domains were aligned using MAFFT (Katoh et al., 2019). Symbols below the alignment indicate the degree of conservation at each position: an asterisk (*) denotes positions with fully conserved residues across all sequences; a colon (:) indicates conservation between residues with strongly similar biochemical properties; and a period (.) marks positions with weakly similar residues. The strictly conserved catalytic triad—cysteine, histidine, and asparagine (or histidine/glutamine) —is highlighted in dark pink. Additionally, three conserved residues contributing to the formation of the catalytic core are highlighted in cyan blue.Fig 1 dummy alt text

cwlA is an essential gene for cell viability

3.2

To investigate the role of Cwl proteins in D. radiodurans, deletion mutants were constructed for each gene. Since D. radiodurans contains between 4 and 10 genome copies, the transformants were purified on selective medium to obtain homogenotes with the deleted allele present in all genome copies. Homogenotes of cwlB and cwlC deletion mutants and the ΔcwlBΔcwlC double mutant were easily obtained after two rounds of purification on selective medium, and the purity of the strains was confirmed by PCR (Fig. S1). For cwlA, hygromycin-resistant colonies were obtained, but PCR analysis of two candidates revealed that, even after three rounds of purification on hygromycin plates, both still retained the wild-type allele in addition to the ΔcwlA::Ωhph allele (Fig. S2). Further purification attempts failed to eliminate the wild-type allele, strongly suggesting that the CwlA protein is essential for cell viability. To further confirm that the cwlA gene is essential in D. radiodurans, we used a conditional gene inactivation system (Nguyen et al., 2009). In this system, ΔcwlA cells expressed the CwlA protein at 30 °C from a temperature-sensitive (repUTs) replication vector. We used, as control, similar construct encoding the essential DdrO protein (Fig. 2, lane 6). Shifting the culture to 37 °C, a non-permissive temperature, resulted in an inability of the plasmid to replicate during successive cell divisions, leading to the depletion of CwlA or DdrO. The cells failed to grow at the non-permissive temperature (Fig. 2), confirming that cwlA is essential for cell viability.Fig. 2. CwlA is essential for D. radiodurans viability. Strains were grown in liquid medium with spectinomycin at 28 °C. Sequential dilutions of cells were spotted on TGY plates in the presence or absence of spectinomycin at 28 °C (A) or 37 °C (B). Lane 1: strain GY13739 containing non-thermosensitive plasmid p11554 (prepU); lane 2: strain GY13747 containing thermosensitive plasmid p13841 (prepUTs); lanes 3 and 4: GY18249 and GY18250 respectively, two independent clones [ΔcwlA (prepUTs*-cwlA*^+^)]; lane 5: strain GY19040 [ΔcwlA (prepU-cwlA^+^)]; lane 6: strain GY14164 [ΔddrO(prepUTs-ddrO^+^)].Fig 2: dummy alt text

CwlB is involved in the degradation of septa

3.3

To further investigate the function of CwlB and CwlC proteins, we examined by epifluorescence microscopy the cell morphology of wild type, ΔcwlB, ΔcwlC and ΔcwlBΔcwlC mutant bacteria. In the absence of CwlC, the cell morphology was comparable to that of the wild-type strain in the exponential or stationary phase of growth (Fig. S3) whereas CwlB depleted cells exhibited markedly abnormal shapes (Fig. 3). The cell morphology of this mutant was strongly affected in the exponential or stationary phase of growth. In stationary growing phase, the ΔcwlB mutant displayed a severe cytokinesis defect by the presence of enlarged multicellular structures, including double tetrads (eight cells) and quadruple tetrads (sixteen cells). In the wild-type strain during this phase, 23.4 % of the cells were present as dyads and 73.8 % as tetrads. In contrast, no dyads were observed in the ΔcwlB mutant. Only 1.6 % of the cells formed tetrads, while the majority were organized as double tetrads (48 %) or quadruple tetrads (50 %); the latter morphology being absent in the wild-type strain (Fig. 3). During the exponential phase, ΔcwlB mutant cells also displayed noticeable defects in cell division (Fig. S4); however, these abnormalities were difficult to be classified into distinct morphological forms. Normal cell segregation was restored in the cwlB mutant upon complementation with a plasmid expressing the cwlB gene under the control of its native promoter (Fig. 3). In the double ΔcwlBΔcwlC mutant, the proportion of quadruple tetrads (80 %) increased compared to that observed in the single cwlB mutant (40 %) (Fig. S5). As well as for the single mutant ΔcwlB, the wild-type phenotype was restored when the ΔcwlBΔcwlC double mutant was complemented with a plasmid carrying the cwlB gene (Fig. S6). When CwlB is present, its activity appears sufficient to degrade the peptidoglycan at the septum; in its absence, CwlC may partially compensate for this function, albeit to a limited extent.Fig. 3. CwlB is required for the degradation of septa. A. Microscopy analyses of stationary phase cells from wild type, ΔcwlB and ΔcwlB/cwlB^+^mutant strains. The different structures observed are indicated by coloured arrows: dyads in red, tetrads in green, double tetrades in blue and quadruple tetrades in yellow. B The percentage of cells found in dyads (2 cells, red), tetrads (4 cells, green), double tetrads (8 cells, blue) and quadruple tetrads (16 cells, yellow) are illustrated in the pie charts.Fig 3: dummy alt text

The higher-order assemblies of cells observed in the ΔcwlB mutant suggest that, while septation may initiate, it fails to proceed to completion, resulting in incomplete separation of daughter cells. We hypothesized that CwlB may be involved in degrading peptidoglycan at the septum during cell division. To determine the subcellular localization of CwlB, we generated strains expressing CwlB fused to fluorescent tag (NeonGreen). These strains displayed a morphology indistinguishable from the wild-type strain (Fig. S7); however, no clear fluorescent signal was detected.

Membrane vesicle formation under CwlA depletion conditions

3.4

To investigate potential effects of CwlA in cell division, we used the conditional gene inactivation system. Fig. 4 illustrates the impact of CwlA depletion on the cellular morphology, visualized through fluorescent microscopy after 24 h of growth at either 30 °C or 37 °C. At 30 °C, no cells exhibited significant structural abnormalities and extracellular membrane vesicles were not observed under these conditions. In contrast, cells incubated at 37 °C displayed marked morphological alterations. DIC images revealed a subset of cells with irregular shapes and evidence of membrane protrusions (highlighted by blue arrows). The FM4–64 staining clearly showed the presence of extracellular MVs surrounding many of the cells. To better visualise the membrane vesicles, electron microscopy experiments were carried out. Compared to wild-type morphology, CwlA-depleted cells exhibited pronounced production of outer membrane vesicles, seen as numerous spherical structures budding from or surrounding the cellular surface (Fig. 5 and Fig. S8).Fig. 4. Formation of membrane vesicles following depletion of CwlA. Cells (∆cwlA/prepUTS::cwlA^+)^ in exponential growth phase cultivated at 30 °C in medium supplemented with spectinomycin were harvested by centrifugation, diluted in antibiotic-free medium and incubated at 30 °C (upper part) or at 37 °C (lower part). In each part, the first line contains the pictures of the Nomarski interference contrast (DIC), the second line, the pictures of membrane DNA staining (DAPI) and the third line the pictures of membrane staining (FM4–64). All pictures are the same scale (bar = 5 μm). Membrane vesicles are indicated by blue arrows.Fig 4: dummy alt textFig. 5Transmission electron microscopy photograph of membrane vesicles in D. radiodurans. Impact of CwlA depletion on D. radiodurans cells. A and C: Cells in exponential growth phase cultivated at 30 °C in medium supplemented with spectinomycin were harvested by centrifugation, diluted in antibiotic-free medium and incubated at 37 °C for 24 h. A1 and A2 -GY 18249 (ΔcwlA/ prepUTS::cwlA^+^); C – GY14164 (ΔddrO/prepUTS::ddrO^+^). B: Membrane vesicle formation in wild type cells in the presence of mitomycin C.Fig 5: dummy alt text

The widespread production of MVs observed upon CwlA depletion reveals a strong association between loss of CwlA and cell envelope alterations. This phenotype is consistent with defects in envelope homeostasis, potentially linked to impaired peptidoglycan remodeling, although a more general envelope stress response cannot be excluded.

The absence of phenotypes such as cellular aggregation or mislocalization of septa during the depletion of CwlA suggests that this endopeptidase is not specifically involved in cell division but rather acts continuously throughout cell development. To support this hypothesis, we constructed a strain expressing a recombinant CwlA fused to a neongreen tag. Easy production of homozygous strains cwlA::neogreen (Fig. S9) and the absence of particular phenotype in terms of growth or cellular morphology (Fig. 6) showed that the addition of neongreen tag did not impact the functionality of the protein. The protein CwlA is homogeneously localized around the cell (Fig. 6). When compared to cells with membranes labeled with FM4–64, septal regions were devoid of fluorescent signal as long as cell separation was not fully completed, suggesting that CwlA is involved throughout cell development and not specifically during septation.Fig. 6. Cellular localization of the endopeptidase CwlA. Microscopy images of CwlA-neongreen cells under standard growth conditions. DIC: phase contrast; DAPI: DNA staining; FM4–64: membrane staining.Fig 6: dummy alt text

Membrane vesicle formation in D. radiodurans following DNA damage by mitomycin C

3.5

CwlA depletion leads to the formation of MVs. Previous work has also shown their formation when D. radiodurans cells were subjected to stress (Devigne et al., 2015; Li et al., 2017; Pospíšil et al., 2020; Ott et al., 2020). For instance, a treatment with a sub-lethal dose of mitomycin C (15 µg/ml for 40 min, resulting in a survival of 90 %), a DNA damage compound, led to morphological changes including membrane blebbing (Li et al., 2017). In order to observe the structures of the MVs, transmission electron microscopy (TEM) experiments were carried out. TEM images revealed the formation of numerous membrane vesicles at the cell surface of D. radiodurans following treatment with the DNA-damaging agent mitomycin C (Fig. 5B and Fig. S10). Notably, small and spherical vesicular structures were observed budding from the outer membrane, uniformly distributed around the cells. The emergence of these MVs in response to mitomycin C-induced genotoxic stress suggests a link between DNA damage and cell wall remodeling. In Deinococcus, the main response pathway to genotoxic conditions is regulated by a constitutively expressed metalloprotease IrrE and a transcriptional repressor DdrO. We previously showed that the depletion of DdrO expressed from a heat-sensitive replication plasmid leads to a loss of viability of the strain when cells grow at non permissive temperature and is accompanied by the development of numerous cellular phenotypes: condensation and disrupted nucleoids, DNA fragmentation and budding of MVs (Devigne et al., 2015). Under the same DdrO depletion conditions, the electron microscopy analyses performed here showed that the MVs formed are similar to those observed in the presence of mitomycin or under CwlA depletion conditions (Fig. 5C).

CwlA is repressed under Ddro depletion conditions

3.6

In our previous work (Eugénie et al., 2021), we mapped the DdrO regulon by ChIP-seq and RNA-seq analysis. The transcriptome of D. radiodurans was analysed at 37 °C, during DdrO depletion over a 24 h-kinetic (1 h, 4 h, 6 h, 8 h, 16 h, 24 h) using 1 h as the reference point. Here, differential analysis was designed to highlight genes whose expression level varies at the same time point between the ΔddrO/prepUTS ddrO^+^ strain developing an apoptotic-like cell death phenotype after a long lasting DdrO depletion (16 h and 24 h) and not in the ΔddrO/ddrO^+^control strain.

Tables 3 and S2 list the most highly expressed genes at 16 h and 24 h, respectively, after DdrO depletion. The prevalence of RDR (radiation/desiccation response) regulon genes among the most overexpressed genes was found at the both time points, with some of them showing high expression levels, such as the ddrA gene. Among the most repressed genes (Tables 4 and S3), cwlA was approximately 18 times less expressed under DdrO depletion conditions at 16 h, being the most down-regulated gene. This repression is likely indirect, since no DdrO-binding site was identified in the cwlA regulatory region and previously performed DdrO ChIP-seq analyses did not detect any interaction peak upstream of cwlA (Eugénie et al., 2021). In contrast, the expression of cwlB and cwlC remained unchanged. Deregulation of cwlA under DdrO depletion may disrupt the structural organization of the membrane (disrupted peptidoglycan formation and recycling, imbalance of membrane components, activation of autolysins), ultimately leading to MV formation. Since DdrO depletion led to a strong reduction in cwlA expression, the MVs observed under these conditions are most likely a consequence of the decreased amount of CwlA.Table 3list of the 50 most upregulated genes at 16 h of DdrO depletion.Table 3: dummy alt textID gene Eugénie et al.ID gene WhiteFunctionFold changeP-valueDRO_C0016DRC0016Hypothetical protein483,028,43E-229DRO_0421DR0423DdrA454,330DRO_0003DdrC311,390DRO_0070DR0070DdrB138,410DRO_A0342DRA0346PprA132,100DRO_C0011DRC0012helix-turn-helix transcriptional regulator, GerE family117,011,87E-162DRO_C0010Hypothetical protein91,641,43E-156DRO_0323DR0326DdrD61,110DRO_1140Hypothetical protein41,741,37E-83DRO_2308DR2338competence/damage-inducible protein cinA40,252,73E-273DRO_0899DR0906GyrB35,723,88E-223DRO_2310DR2340RecA27,009,52E-150DRO_1139Hypothetical protein26,348,08E-79DRO_0596DR0596RuvB25,271,78E-150DRO_C0021Hypothetical protein22,282,26E-114DRO_2249DR2275UvrB21,232,39E-194DRO_A0073DRA0075Transposase, putative16,714,04E-79DRO_A0077DRA0079Hypothetical protein16,032,90E-33DRO_1880DR1902Exodeoxyribonuclease V, subunit RecD, putative15,663,12E-93DRO_2145DR2174Leucyl-tRNA ligase14,832,39E-168DRO_2144DR2173Hypothetical protein13,611,82E-62DRO_1891DR1913GyrA13,501,20E-94DRO_0420DR0422Trans-aconitate methyltransferase13,326,50E-127DRO_2415DR2441Acetyltransferase12,892,93E-59DRO_2251DR2277Nickel transporter, permease protein12,783,57E-155DRO_1899DR1921sbcD12,751,57E-81DRO_A0167DRA0165Hypothetical protein12,681,29E-84DRO_1459DR1477RecN12,561,24E-132DRO_C0022DRC0024Hypothetical protein12,353,39E-104DRO_A0084DRA0085Hypothetical protein12,359,78E-27DRO_C0009DRB0005Transposase12,095,84E-83DRO_0595DR0595Hypothetical protein10,971,58E-73DRO_1279Hypothetical protein10,919,93E-58DRO_2346Hypothetical protein10,893,95E-32DRO_1900DR1922SbcC10,822,04E-52DRO_1673DR1696MutL10,781,58E-72DRO_A0081DRA0083Hypothetical protein10,652,15E-36DRO_A0080DRA0082Hypothetical protein10,621,29E-18DRO_1879DR1901Hypothetical protein10,556,26E-65DRO_A0082DRA0084Hypothetical protein10,467,55E-28DRO_A0341DRA0345Erythromycin esterase10,356,03E-123DRO_0614DR0614Hypothetical protein10,191,23E-91DRO_1796DR1817Phospho-2-dehydro-3-deoxyheptonate aldolase9,78NADRO_A0085DRA0086Hypothetical protein10,191,23E-91DRO_0438DR0440RuvC-like9,602,18E-79DRO_2142DR2171Hypothetical protein9,583,31E-58DRO_0437DR0439Heme biosynthesis protein HemY9,492,05E-69DRO_2416DR2442Acetylglutamate kinase9,292,42E-49DRO_0820DR082550S ribosomal protein L319,192,66E-39DRO_C0008DRC0010Hypothetical protein8,623,42E-72Table 4list of the 50 most downregulated genes at 16 h of DdrO depletion.Table 4: dummy alt textID gene Eugénie et al.ID gene WhiteFunctionFold changeP-valueDRO_1728DR1749Endopeptidase-related protein−18,021,15E-196DRO_1645DR1668Potassium transporter KtrB−7,786,72E-63DRO_0564DR0566Hypothetical protein−7,002,05E-51DRO_2236DR2262Hypothetical protein−6,501,33E-33DRO_1964DR1990Hypothetical protein−6,504,43E-04DRO_1965DR1991Pseudouridine synthase−5,542,01E-36DRO_0015DR001416S rRNA methyltransferase−5,412,11E-27DRO_0363DR0364Peptide ABC transporter permease−5,374,43E-67DRO_2032DR2058Hypothetical protein−5,215,27E-46DRO_B0027DRB0025Sigma-B regulator RsbS−5,051,05E-45DRO_1839DR1862Translation factor Sua5−4,818,12E-06DRO_1017DR1022Hypothetical protein−4,768,58E-44DRO_2518DR2546Hypothetical protein−4,722,11E-21DRO_2324DR2353L-asparaginase−4,722,77E-31DRO_0127DR0128Nucleotide exchange factor GrpE−4,462,17E-30DRO_2579DR2607Molybdenum cofactor biosynthesis protein Moa−4,444,63E-20DRO_1963DR1989Hypothetical protein−4,434,41E-17DRO_0465DR0466Diguanylate cyclase−4,424,68E-50DRO_1810DR1831Hypothetical protein−4,367,03E-30DRO_0137DR0137Restriction endonuclease−4,344,54E-28DRO_1439DR1455MerR family transcriptional regulator−4,311,11E-21DRO_0364DR0365Peptide ABC transporter permease−4,317,85E-43DRO_2318DR2347Hypothetical protein−4,281,11E-25DRO_2209DR2235Hypothetical protein−4,252,80E-37DRO_1683DR1705Hydrolase−4,197,34E-37DRO_1958DR1984Thymidine kinase−4,121,49E-24DRO_0391DR0391Hypothetical protein−4,102,92E-05DRO_B0026DRB0024Sigma-B regulator RsbR−4,091,11E-30DRO_B0028DRB0026Sigma-B regulator RsbT−4,096,68E-18DRO_B0055DRB0053Dihydroxyacetone kinase subunit DhaK−4,077,21E-26DRO_0949DR0956Phosphoesterase−4,021,60E-16DRO_1941DR1967Enoyl-ACP reductase−3,964,88E-24DRO_0665DR0670Hypothetical protein−3,918,11E-14DRO_0125DR0127Hypothetical protein−3,865,37E-29DRO_2045DR2072Hypothetical protein−3,821,47E-23DRO_B0054Hypothetical protein−3,821,02E-16DRO_1809DR1830Hypothetical protein−3,805,61E-28DRO_1227DR1231Hypothetical protein−3,783,53E-26DRO_1836DR1858Permease−3,731,60E-30DRO_1837Hypothetical protein−3,721,93E-21DRO_2238DR2264MBL fold metallo-hydrolase−3,711,04E-27DRO_0799DR0804RNA polymerase subunit sigma−3,683,13E-27DRO_0123DR0125Acetyltransferase−3,661,58E-34DRO_1850DR1873Hypothetical protein−3,654,52E-22DRO_1824DR1845Hypothetical protein−3,623,64E-31DRO_0483DR0484Hypothetical protein−3,612,35E-25DRO_0118DR0120DNA processing protein DprA−3,591,63E-09DRO_1813DR1834rRNA (guanine-N2)-methyltransferase−3,594,44E-15DRO_2117ABC transporter−3,573,99E-33DRO_0870DR0875Zinc metalloprotease−3,562,09E-29DRO_1887DR1909Lactate utilization protein−3,551,71E-27

Discussion

4

Although D. radiodurans is a Gram-positive bacterium, the organisation of its membrane structure is similar to that of Gram-negative bacteria. Cell wall hydrolases are ubiquitous enzymes that play critical roles in cell division and bacterial enveloppe remodeling. Here, we show that CwlB plays a key role in cell separation in D. radiodurans. In mycobacteria, the NlpC/P60 endopeptidase RipA family is required for septal degradation, and its deletion leads to cells with multiple septa (Chao et al., 2013). Similar phenotypes were observed in Corynebacterium, where deletion of the cg1735, another NlpC/P60 member caused elongated, multiseptated cells (Gaday et al., 2022). The predicted signal peptide at the N-terminal sequence of CwlB suggests an extracytoplasmic, likely periplasmic, localization. Although fluorescence was not detected for the CwlB–NeonGreen fusion protein, the wild-type phenotype of the strain indicates that this fusion protein is functional, suggesting technical limitations such as low protein abundance or inefficient folding or maturation of the fluorophore. The precise subcellular localization of CwlB remains to be determined through a protein isolation approach based on subcellular localization.

Peptidoglycan metabolism and hydrolase activity are tightly regulated throughout the cell cycle to prevent deleterious effects (Vermassen et al., 2019). In D. radiodurans, the DNA damage response gene ddrI encodes a transcriptional regulator belonging to the cAMP receptor protein (CRP) family. These proteins are known to act as positive or negative transcriptional regulators. In silico predictions identified a putative DdrI binding site upstream of the cwlB (dr1325) coding region (Meyer et al., 2018). DdrI is strongly upregulated in stationary-phase cells (Meyer et al., 2018), data confirmed by transcriptomic analyses showing a strong ddrI gene upregulation in stationary phase compared to exponential phase (Eugénie et al., 2021). Cells lacking DdrI exhibit a pleiotropic phenotype, including growth defects, increased sensitivity to DNA-damaging agents, oxidative stress, and heat shock (Yang et al., 2016; Meyer et al., 2018). In addition, DdrI deficiency leads to a significant increase in two-tetrad-forming subpopulations suggesting that DdrI is crucial for accurate completion of cell division, particularly during stationary phase (Meyer et al., 2018). The formation of double tetrads is even more pronounced in the absence of CwlB. Based on these observations, we propose that DdrI activates cwlB expression.

In this study, we demonstrated the essentiality of the cwlA gene. The loss of viability due to the absence of cwlA is quite surprising since the inability to obtain the deletion of a gene encoding an NlpC/P60 endopeptidase has rarely been reported in other bacteria. One of the few examples of essentiality for a hydrolase concerns PcsB in S. pneumoniae, were depletion of PcsB is bacteriostatic and leads to growth arrest, aberrant cell phenotypes and uncontrolled peptidoglycan synthesis (Ng et al., 2004). In Listeria monocytogenes, deletion of the cwhA gene results in abnormal septa and cell filamentation during the exponential phase (Pilgrim et al., 2003). Cell filamentation is also described after inactivation of the lytF gene in Bacillus subtilis, a phenotype exacerbated in the absence of the second peptidase NlpC/P60 LytE (Bisicchia et al., 2007). Only the absence of both LytE and the endopeptidase CwlO is lethal (Bisicchia et al., 2007). Finally, in E. coli, the three endopeptidases Spr, YdhO and YebA are functionally redundant (Singh et al., 2012). The inability to obtain a cwlA deletion mutant in D. radiodurans suggests that, unlike in other bacteria, CwlA activity cannot be replaced by other proteins, despite the presence of two other potential NlpC/P60 endopeptidases, CwlB and CwlC.

We also reported that depletion of CwlA results in extensive formation of membrane vesicles (MVs), a phenotype previously observed under stress conditions, such as exposure to the DNA-damaging agent mitomycin C (Li et al., 2017) or to low earth orbit conditions outside the International Space Station (Ott et al., 2020). Transmission electron microscopy analysis in our study confirmed the presence of numerous spherical vesicles budding from the outer membrane following mitomycin C treatment. These MVs were evenly distributed around the cell surface, indicating a regulated process rather than random membrane disintegration.

Depletion of CwlA and DdrO, two proteins essential for cell viability, induces MV formation. In contrast, MVs were not detected when other essential proteins in D. radiodurans, such as DNA gyrase or HU protein, were depleted (Nguyen et al., 2009). Changes in cell morphology are observed, such as the appearance of anucleate cells in the absence of DNA gyrase or fragmentation of the nucleoid leading to cell lysis upon depletion of HU. The formation of MVs observed during CwlA or DdrO depletion therefore appears to depend on a specific pathway. In Pseudomonas aeruginosa, the formation of MVs has been observed when cells are exposed to ciprofloxacin or mitomycin C and explosive cell lysis is regulated by the RecA-dependent SOS response (Turnbull et al., 2016). DNA damage activates RecA, triggering the SOS response and leading to the production of endolysins that degrade the cell wall. While D. radiodurans exhibits a more controlled vesiculation process rather than explosive lysis, both systems highlight the critical role of DNA damage signaling in modulating cell envelope integrity and membrane vesicle production. In Deinococcus, the main response pathway to genotoxic conditions is regulated by the metalloprotease IrrE and the repressor DdrO (Wang et al., 2015; Blanchard et al., 2017). We show that DdrO depletion results in strong repression of cwlA. Given that loss of CwlA alone is sufficient to induce vesiculation, we propose that its downregulation during DdrO depletion contributes to the observed membrane abnormalities and outer MV formation. However, no RDR binding motif is present upstream of cwlA, suggesting that this effect is indirect. In many Gram-positive bacteria, regulation of peptidoglycan metabolism and the numerous associated hydrolases involves the essential two-component system WalK/WalR (Dubrac et al., 2008). In E. coli, the two-component system CpxR/CpxA regulates several amidases (Weatherspoon-Griffin et al., 2011). The regulation of the cwlA expression and other cell wall hydrolases in D. radiodurans may also involve a two-component regulatory system similar to WalK/WalR. About twenty genes predicted to encode sensor proteins with histidine kinase like domains along with numerous proteins containing a receptor domain are present in D. radiodurans (Makarova and Daly, 2010) but only a few have been studied (Im et al., 2013).

CwlA is distributed around the cells, probably within the periplasm. The presence of an NlpC/P60-type catalytic domain together with a LysM peptidoglycan-binding domain is consistent with a potential role for CwlA in peptidoglycan metabolism. Based on homology with other NlpC/P60 endopeptidases, CwlA may contribute to the hydrolysis of peptide bonds within PG and thus participate in PG remodeling processes associated with cell wall growth and/or turnover (Anantharaman and Aravind, 2003; Vermassen et al., 2019). Although no direct evidence for such a function is provided here, altered peptidoglycan homeostasis resulting from CwlA depletion could weaken the peptidoglycan layer. Given the thickness of the D. radiodurans cell wall, fragile areas could lead to localized breakage points across the entire wall resulting in the formation of bubbles and MVs. Alternatively, membrane vesicule formation could result from the accumulation of PG fragments or misfolded proteins in the periplasm, generating turgor pressure on the outer membrane and promoting vesiculation, as previously proposed (McBroom and Kuehn, 2007; Tashiro et al., 2009). Future experiments will be required to discriminate between these possibilities and to determine whether MV formation is a direct consequence of altered peptidoglycan remodeling or a secondary effect of envelope stress.

Finally, the MazEF toxin-antitoxin system may contribute to stress-induced vesiculation. MazEF mediates programmed cell death (PCD) under severe stress conditions and induces membrane blebbing and MV formation (Li et al., 2017), similar to phenotypes observed during CwlA or DdrO depletion . This suggests a potential interplay between MazEF, DdrO, and CwlA in coordinating stress responses. The convergence of phenotypes associated with CwlA or DdrO depletion, and MazEF activation suggests a regulatory network linking DNA damage response, cell wall homeostasis and PCD enabling D. radiodurans to balance survival and death under extreme conditions.

Beyond structural consequences, MV production may have functional implications. Vesicles are known to carry diverse biomolecules, including proteins, lipids, and nucleic acids, which can participate in stress response, signaling, and horizontal gene transfer. The MVs produced upon CwlA depletion or DNA damage may contain enzymes or signaling molecules involved in envelope remodeling or DNA repair. The overrepresentation of DNA repair proteins has recently been demonstrated in extracellular vesicles produced by Methanobrevibacter smithii (Baquero et al., 2025) suggesting a link between genotoxic stress response and vesiculation. The role of MV production remains to be fully elucidated, but it may serve dual purposes: expelling damaged components to promote survival to the bacterial population or facilitating cell death by disrupting membrane integrity.

CRediT authorship contribution statement

Tim Kamara: Investigation, Visualization, Writing – review & editing. Geoffrey Martinez: Investigation, Visualization, Writing – review & editing. Nicolas Eugénie: Investigation, Data curation. Murielle Dutertre: Resources. Fabrice Confalonieri: Funding acquisition, Data curation, Writing – review & editing. Esma Bentchikou: Conceptualization, Writing – review & editing. Pascale Servant: Conceptualization, Supervision, Writing – original draft, Writing – review & editing.

Declaration of competing interest

All authors have read and approved the final version of the manuscript.

We confirm that this work is not presently submitted to a journal different from Current Research in Microbial Sciences.