Peptidoglycan remodeling improves salt resilience of Zymomonas mobilis

ABSTRACT

The alpha-proteobacterium Zymomonas mobilis is one of the most efficient microbial producers of ethanol and has the potential to be established as biofuel producer at industrial scale. However, a bottleneck hindering the full use of Z. mobilis in biorefinery is its sensitivity to environmental stresses, including sodium chloride (NaCl), a common component present in biomass from various sources. To address this limitation, we need to deepen our understanding of the cell envelope, the crucial barrier between bacteria and external stressors. To date, the cell envelope of Z. mobilis has remained largely uncharacterized. Here, we show that the deletion of ctpA, which encodes a periplasmic protease, increases the salt resilience. Salt resilience is mediated by an increased level of the peptidoglycan endopeptidase MepM, which we identified as a substrate of CtpA through comparative proteomics. Supporting this, the overexpression of MepM in the wild type enhanced salt resilience. We also discovered that the peptidoglycan of Z. mobilis is O-acetylated at MurNAc residues, a modification usually associated with virulence in pathogenic bacteria. Interestingly, O-acetylation was crucial for salt resilience, supporting a role in PG growth regulation under the stress condition. Overall, this study highlights the importance of investigating cell envelope biology in Z. mobilis as a foundation for engineering strains with improved resilience to environmental stress and, more generally, studying the cell envelopes of non-model bacteria to expand our fundamental understanding of cell function.

IMPORTANCE

Fossil fuels have negative impacts on the environment and will become limited in the next decades. Hence, alternative, sustainable energy sources need to be urgently established. Microbial fermentation of biomass for biofuel production presents a promising avenue. The gram-negative alpha-proteobacterium Zymomonas mobilis exhibits a superior capacity to convert sugars into ethanol, a clean, renewable, and widely used fuel. However, Z. mobilis has not been used as a first choice as a biofuel producer. The ethanol producer, baker's yeast Saccharomyces cerevisiae, serves as a model species in cell biology, but we lack fundamental understanding of the cell envelope biology of Z. mobilis, which would be critical to engineer a strain with increased resilience. Here, we demonstrate that knowledge about cell envelope biogenesis factors in Z. mobilis can help engineering optimized strains that grow under conditions of biofuel production.

INTRODUCTION

The gram-negative alpha-proteobacterium Zymomonas mobilis has been used in the traditional production of alcoholic beverages in Central America for thousands of years. It is well known for its exceptional ethanologenic physiology, which enables it to convert simple sugars such as glucose into ethanol at near-maximal theoretical yield and with a rapid production rate (1). Thus, Z. mobilis has been considered one of the most promising microbial platforms for industrial-scale bioethanol production (2). Over the last two decades, metabolic engineering has led to strains capable of utilizing various plant feedstock for the production of ethanol and other high-value chemicals (3–5).

The central metabolism of Z. mobilis is well understood, but little is known about the structure and biogenesis of its cell envelope, the physical barrier that separates the bacterial cytoplasm from the external environment. The gram-negative cell envelope comprises an outer membrane (OM), a periplasm with a thin layer of peptidoglycan (PG), and a cytoplasmic membrane. PG consists of glycan chains carrying peptides that can be cross-linked with each other, forming a mesh-like layer (6). PG surrounds the cytoplasmic membrane and, together with the firmly attached OM, provides mechanical strength to prevent bursting of the cell caused by the turgor (7). The PG also maintains the shape of the cell, and the OM protects the cell from being killed by many toxic compounds (e.g., detergents and antibiotics) and antibacterial lysins. As Z. mobilis remains sensitive to various environmental stresses, which is a bottleneck for its industrial use, we need to gain a better understanding of the key cell envelope components.

One of the known stressors affecting the growth of Z. mobilis is common salt, sodium chloride (NaCl). Even at a relatively low concentration of 225 mM, NaCl severely inhibits the growth and fermentation capacity of Z. mobilis for unknown reasons (8, 9). The effect of NaCl is not the result of an osmotic stress, as Z. mobilis is able to grow without morphological defects in environments with higher osmolality, for example, in the presence of 0.56 M glucose (8, 10). Interestingly, using adaptive evolution under elevated NaCl conditions, our previous work identified a truncation of the ctpA gene associated with enhanced salt resilience (9). The truncation likely led to the expression of a short, inactive CtpA version lacking the whole C-terminal domain (amino acid residues 289–459). CtpA is a member of the Carboxyl-terminal processing protease (Ctp) family of ubiquitous bacterial serine proteases. Originally, Ctp was shown to remove short C-terminal, flexible regions from its protein substrates, but further work established that it also cleaves within N-terminal domains or degrades whole substrates, regulating their cellular amount and/or activity (11). In gram-negative bacteria, the Ctp called Prc degrades periplasmic peptidoglycan hydrolases with different cleavage sites. For example, in the model species Escherichia coli, Prc degrades peptidoglycan endopeptidases and lytic transglycosylases, facilitated by the Prc-activator and endopeptidase adaptor protein NlpI (12–15). If a similar regulatory mechanism exists in Z. mobilis, the absence of CtpA might lead to changes in the cell envelope, which could explain the observed phenotype of the salt-adapted strain with truncated ctpA gene.

In this study, we investigated the role of CtpA and determined the composition of Z. mobilis peptidoglycan. We show that modulating peptidoglycan hydrolase activity can significantly improve cell fitness under salt stress. In addition, we discovered that the peptidoglycan is O-acetylated, and this modification is important for salt resilience.

RESULTS

Loss of CtpA improves salt resilience in Z. mobilis

The previously evolved salt-resilient strain had a truncated ctpA gene that likely rendered the protein inactive (9). To verify that the absence of CtpA causes a salt-resilient phenotype in Z. mobilis, we first aimed to generate a mutant strain that lacked the complete ctpA gene using a homologous recombination-based method (16). We readily obtained mutant colonies and confirmed the absence of the ctpA gene by PCR analysis. As Z. mobilis is a polyploid species with up to 20 chromosome copies per cell (17–20), we also checked for the deletion of ctpA by whole-genome sequencing, which verified the deletion in all copies of chromosome (variant ratio 100%).

We next compared the growth characteristics and cell morphology of the ctpA deletion strain (ΔctpA) and the parental strain under both salt and no-salt conditions. As previously reported, we confirmed that low concentrations of 225 mM NaCl in the growth medium resulted in severely distorted morphology and impaired growth in the wild-type strain Z6 (8, 9); the cells displayed thick, filamentous morphology with pronounced bulging at one of the poles. In contrast, ΔctpA cells grew significantly faster and to a higher optical density than the wild-type under the same salt condition (Fig. 1A). The mutant cells tended to be more elongated but exhibited less severe polar bulging than wild-type cells (Fig. 1B). To exclude polar effects caused by the deletion of ctpA, we complemented the ΔctpA strain by inserting ctpA with its upstream region into a different locus of chromosome. The complementation (strain ΔctpA + ctpA) restored the salt sensitivity and morphology of the wild type (Fig. 1A and B), verifying that the observed phenotype in the ctpA mutant was due to the deletion of ctpA. We attempted to measure the cell area of stressed cells; however, this was not technically feasible due to aberrant cell shape and the clumping/aggregation of large cells on the agar pad.

Fig 1.

Deletion of ctpA improves salt resilience in Zymomonas mobilis. (A) Growth of Z. mobilis strains Z6 (wild-type), ΔctpA, and ΔctpA + ctpA (complementation strain) under salt stress conditions (left) and regular, no-salt conditions (right). Biological replicates N = 3. The error bars represent the standard deviation. (B) Phase contrast images of growing Z6, ΔctpA, and ΔctpA + ctpA (complementation strain) under salt conditions (left) and regular, no-salt growth conditions (right). Scale bars, 10 μm. (C) Box-and-whisker plot showing the area of Z. mobilis cells of the three different strains grown to an OD₆₀₀ of 0.7–1.5 under no-salt conditions. The cell area was measured using MicrobeJ (21). Biological replicates N = 3. Sample size, n = 1,356 cells (Z6), n = 1,239 cells (ΔctpA), n = 1,140 cells (ΔctpA + ctpA). NS, not significant; ****, P ≤ 0.0001 (unpaired t-test).

Under no-salt, standard growth conditions, ΔctpA and wild-type cells grew with similar rates (Fig. 1A, right side). However, a microscopic analysis revealed that the ΔctpA cells were by 14.7% smaller (measured by cell area) than the wild-type cells (Fig. 1C). This phenotype was reversed by the complementation, showing that the size reduction was also due to the loss of CtpA. These findings suggest that CtpA plays an unidentified role in cell morphology and size, presumably by affecting the synthesis or maintenance of the cell envelope or peptidoglycan, even under no-stress conditions.

Putative substrates of CtpA

We next asked what the underlying mechanisms of the CtpA-dependent salt sensitivity and cell size maintenance are. As CtpA is a predicted protease, we aimed to identify its substrate(s) that could influence growth and cell morphology. To this end, we performed comparative proteomics to search for putative substrates of CtpA among the total cellular proteins (comparing wild type vs ΔctpA, under regular or salt conditions). As expected, we did not detect CtpA in the ΔctpA strain under both conditions. Substrates of CtpA are expected to be present in the periplasm and enhanced in the ctpA mutant (or decreased in the wild type). Our proteomics analysis identified 42 possible substrates that were enhanced at least 1.5-fold in ΔctpA cells under salt conditions (Fig. 2A; Table S4). Among these, only two proteins were also present at significantly higher amounts under regular growth conditions. We identified the product of the ZZ6_1677 gene as the top candidate for a CtpA substrate under both conditions (Fig. 2). ZZ6_1677 encodes an M23 endopeptidase with homology to MepM in E. coli (20.7% amino acid identity and 30.6% similarity) (Fig. S1). Interestingly, MepM has been previously identified as a substrate of CtpA in Pseudomonas aeruginosa and Prc in E. coli, respectively (22, 23). Another protein significantly enhanced under both conditions was FtsL (encoded by ZZ6_0468), an essential cell division protein (Fig. 2). Other proteins enhanced only under salt conditions included three TonB-dependent proteins (encoded by ZZ6_1483, ZZ6_1535, and ZZ6_1024), an ammonium transporter (ZZ6_0900), and a potassium transporter (ZZ6_0125) (Table S4). These proteins might contribute to the enhanced salt resistance by transporting osmolytes from the environment.

Fig 2.

MepM and FtsL are enhanced in ΔctpA under salt and no-salt conditions. Volcano plot presenting log2 fold change (FC) of abundance of individual protein (ΔctpA vs Z6 under salt (A) and no-salt (B) conditions) in X-axis and its statistical significance (P value from t-test) in Y-axis. Four biological replicates were measured for each strain and conditions. Red indicates significantly enhanced proteins in ΔctpA, and blue indicates significantly reduced proteins in ΔctpA. Table S4 lists enhanced proteins in ΔctpA under salt conditions.

The altered cell morphology and proteomics suggested that the enhanced MepM endopeptidase level could contribute to salt resilience in ΔctpA. We therefore focused on investigating the role of MepM in salt-related phenotypes.

MepM overexpression increases salt resilience

To explore the potential role of MepM in salt resilience, we first examined the effect of mepM overexpression in Z. mobilis. We constructed a strain carrying the pBBR plasmid with the mepM gene under the constitutively active pdc promoter (ZZ6_1712) (24). We were unable to quantify the exact level of MepM due to the lack of specific antibodies and our inability to express a stable FLAG-tagged chromosomal version (data not shown). However, the pdc promoter has been used for the overexpression of other genes in Z. mobilis (24); hence, we predict that the strain with pBBR-mepM has a higher MepM level than the corresponding strain with empty pBBR.

Interestingly, compared to the strain with empty pBBR (Z6 pBBR), the strain carrying pBBR-mepM (Z6 pdc-mepM) exhibited enhanced growth under salt conditions (Fig. 3A and B), albeit the level of salt resilience did not fully reach that of ΔctpA (Fig. 3A). This result strongly indicates that MepM is indeed overproduced in Z6 pdc-mepM and demonstrates the importance of MepM in salt resilience.

Fig 3.

MepM overexpression increases salt resilience in Zymomonas mobilis cells. (A) Growth profiles of Z. mobilis strains Z6 (wild type) carrying pBBR, ΔctpA carrying pBBR, Z6 carrying pBBR pdc-mepM, and Z6 carrying pBBR pdc-mepM H287A under salt conditions (left) and regular, no-salt growth conditions (right). Biological replicates N = 3. The error bars represent the standard deviation. (B) Phase contrast images of growing, Z6 (wild type) carrying pBBR, ΔctpA carrying pBBR, Z6 carrying pBBR pdc-mepM, and Z6 carrying pBBR pdc-mepM H287A under salt conditions (left) and under regular, growth conditions (right). Scale bars, 10 μm. (C) Box-and-whisker plot showing the area of Z. mobilis cells of the three different strains grown to an OD₆₀₀ of 0.7–1.5 under no-salt conditions. The cell area was measured using MicrobeJ (21). Biological replicates N = 3. Sample size, n = 585 cells (Z6 + pBBR), n = 597 cells (ΔctpA + pBBR), n = 701 cells (Z6 + pBBR pdc-mepM), and n = 593 cells (Z6 + pBBR pdc-mepM H287A). NS, not significant; ****, P ≤ 0.0001 (unpaired t-test).

We next investigated whether the catalytic activity of MepM is essential for the observed phenotypes in Z6 pdc-mepM. Previous reports identified six critical active site residues (R286, H297, D301, H345, H376, and H378) for zinc-binding in the M23 endopeptidase ShyA, a homolog of MepM in Vibrio cholerae (25). Sequence alignment confirmed that five of these six critical catalytic residues (except R286 in V. cholerae) are conserved in Z. mobilis MepM. We mutated the mepM gene in pBBR pdc-mepM to express MepM (H287A) in which one of the putative critical residues, Histidine 287, is replaced by Alanine. We then transformed the plasmid into wild-type Z6 to generate strain Z6 pdc-mepM H287A. Under salt conditions, cells of this strain exhibited a similar poor growth phenotype as the wild-type cells, which markedly differed from the improved growth when expressing active wild-type MepM (Fig. 3A). Under no-salt conditions, Z6 pdc-mepM H287A cells were similar in size to wild-type cells (Fig. 3C). This indicates that the H287A mutation abolished MepM-dependent effects and that the endopeptidase activity of MepM is essential for the salt resilience and cell size phenotypes.

To gain further functional insights into MepM in Z. mobilis, we aimed to delete the mepM gene in Z. mobilis. However, several attempts to delete the gene were unsuccessful. In total, we have examined 81 colonies that underwent the excision of integrated plasmid and found that they were all reverted to wild type, consistent with a previous systematic study using the Z. mobilis type strain Z4, which identified mepM as an essential gene (26). To the best of our knowledge, essentiality of a single peptidoglycan endopeptidase has not been experimentally demonstrated in other bacteria. This indicates that Z. mobilis might possess a unique peptidoglycan hydrolase system with lower genetic redundancy, differing from that of well-characterized model bacteria (27).

Muropeptide profile of Z. mobilis

In E. coli, MepM hydrolyzes DD(4-3)-cross-links in PG peptides, contributing to making space for the incorporation of new PG during cell elongation (28). We asked if the PG composition is altered in Z. mobilis strains with enhanced MepM, that is, ΔctpA or Z6 pdc-mepM. To date, the PG profile of Z. mobilis has not been reported, and we therefore wanted to establish it for wild-type Z6, using the standard method for the gram-negative model bacterium E. coli (29). Z6 cells were grown anaerobically to mid-exponential phase, harvested, and boiled in the presence of 4% sodium dodecyl sulfate, followed by the recovery and purification of PG sacculi. Muropeptides (disaccharide peptide subunits) were released by the muramidase cellosyl and reduced with sodium borohydride, followed by their separation by reversed-phase high-performance liquid chromatography (HPLC) using the standard conditions for E. coli muropeptides (Fig. S2). This analysis showed that the major muropeptides present in the E. coli PG profile, disaccharide-tetrapeptide (Tetra) and bis-disaccharide-tetratetrapeptide (TetraTetra), were also present in Z. mobilis PG (Fig. S2). Interestingly, Z. mobilis PG exhibited additional major muropeptides that were absent in muramidase-digested PG from E. coli (Fig. S2). Mass spectrometry (MS) analysis identified these additional peaks as O-acetylated muropeptides (Fig. S2). O-acetyl groups are known to be present in the PG of several pathogenic bacteria (6). They are labile and lost at acidic and alkaline conditions and are most stable around pH 6.0 (30). The standard HPLC buffers for E. coli muropeptide analysis have acidic pH (4.31 and 4.95), which likely causes the loss of some of the O-acetyl groups. We therefore repeated the PG preparation and analysis using buffers with a pH of 6.0 in all purification and analysis steps to preserve O-acetyl groups. As expected, these conditions gave rise to enhanced peaks of the O-acetylated muropeptides, but the use of HPLC buffers with a pH of 6.0 resulted in broader, less resolved peaks (Fig. S3).

The muropeptide analysis combined with MS confirmed that Z. mobilis possesses mature stem peptides typical for gram-negative bacteria, L-Ala-D-iGlu-mDAP-D-Ala (Fig. 4A; Fig. S2). Cross-linked dimeric muropeptides comprised 45.9% of the total muropeptides, which is similar to those in other gram-negative bacteria (29, 31, 32). O-acetylation was present in both monomeric and dimeric muropeptides (peak 4, 7, and 9 in Fig. 4B), affecting 55.9% of all muropeptides in the wild-type under regular growth conditions. We subsequently analyzed the PG from salt-treated wild-type cells, revealing that the saline stress altered the muropeptide profile. Changes included an increase in Tri (from 1.7% to 4.1%) and Tri(OAc) (from 1.0% to 6.3%), and a decrease in Tetra (from 19.5% to 11.9%), TetraTetra (from 5.3% to 1.8%), TetraTri(OAc) (from 3.4% to 1.8%) and TetraTetra(OAc) (from 13.4% to 7%) (Fig. 4C). Overall, the dimers decreased from 45.9% to 43.1% and O-acetylation increased from 55.9% to 63.0% by the salt stress. Thus, salt stress induces changes in the PG in Z. mobilis (Fig. 4D).

Fig 4.

Muropeptides profile of Zymomonas mobilis strain Z6 (wild type) and ΔctpA. (A) Cells of Z6 (wild type) and ΔctpA were grown under no-salt and salt (+NaCl) conditions. The PG was isolated, and the muropeptides were released by cellosyl, reduced by sodium borohydride, and separated by HPLC under O-acetyl group preserving conditions. Muropeptides are numbered, and their structures are shown in Fig. S4. (B) Proposed structures of the major muropeptides released from Z. mobilis PG, illustrating the presence of O-acetyl groups at MurNAc in some of the muropeptides. (C) Relative quantification of muropeptides from Z6 and ΔctpA grown under no-salt and salt conditions. Biological replicates N = 2. The error bars represent the standard deviation. (D) Relative proportion of monomeric, dimeric, and O-acetylated muropeptides. Colors indicate the strains/growth conditions as in (C). Biological replicates N = 2. The error bars represent the standard deviation.

Next, we analyzed the PG composition of the ΔctpA strain for comparison. Interestingly, despite the morphological differences, the majority of muropeptides remained at similar levels between the two strains under both conditions (Fig. 4A and C). However, the dimers were reduced from 45.9% to 43.4% (no-salt) and 43.1% to 41.4% (salt) in ΔctpA (Fig. 4D). The muropeptide profile of Z6 pdc-mepM was similar to that of ΔctpA under regular growth conditions (Fig. S5A and B), with the exception of an overall 8.9% higher O-acetylation level (Fig. S5C). Under salt conditions, the dimers decreased from 42.4% to 39.0% (Fig. S5C).

PG O-acetylation is important for salt stress response in Z. mobilis

Both the wild-type and ΔctpA strains contained increased levels of PG O-acetylation under salt conditions, compared to no-salt conditions (Fig. 4D). In addition, ΔctpA cells also contained higher PG O-Ac levels than the wild type under salt conditions (Fig. 4D). These observations indicate that the O-Ac contents might be important for Z. mobilis to resist salt stress. To test the idea, we generated a mutant lacking the homolog of patA (ΔpatA), which encodes a PG O-acetyl transferase present in several gram-negative bacteria (33). The muropeptide analysis confirmed that O-acetylated muropeptides were absent in ΔpatA cells, while the unmodified Tetra and TetraTetra were dominant in the profile (Fig. S6). As expected, the complemented ΔpatA + patA strain (with patA on another locus on the chromosome) showed a wild-type-like muropeptide profile with O-acetylated peaks (Fig. S6). This confirmed that PatA functions as a PG O-acetyltransferase in Z. mobilis.

We then assessed the salt resilience of the patA mutant. Interestingly, the strain exhibited compromised growth under salt conditions while it had no growth defects under no-stress conditions (Fig. 5A). This phenotype was reversed to that of the wild type upon complementation with the patA gene, excluding polar effects on adjacent genes (Fig. 5A). Interestingly, compared to wild type, the ΔpatA mutant cells showed a lower degree of polar bulging (Fig. 5B and D). Under regular, no-salt growth conditions, ΔpatA cells were slightly smaller than wild-type cells (Fig. 5C). These results indicate that O-acetylation likely affects PG synthesis and remodeling.

Fig 5.

O-acetylation of PG is important for salt resilience. (A) Growth profiles of Zymomonas mobilis strains Z6 (wild type), ΔpatA, and ΔpatA + patA (complementation strain) under salt (left) and regular, no-salt growth conditions (right). Biological replicates N = 3. The error bars represent the standard deviation. (B) Phase contrast images of growing, Z6 (wild-type), ΔpatA, and ΔpatA + patA under salt conditions (left) and regular, no-salt growth conditions (right). Scale bars, 10 μm. (C) Box-and-whisker plot showing the area of Z. mobilis cells of the three different strains grown to an OD₆₀₀ of 0.7–1.5 under no-salt conditions. The cell area was measured using MicrobeJ (21). Biological replicates N = 3. Sample size, n = 1,356 cells (Z6), n = 963 cells (ΔpatA), and n = 577 cells (ΔpatA + patA). *, P ≤ 0.05; ***, P ≤ 0.001 (unpaired t-test). (D) A box-and-whisker plot presenting the maximum diameter of bulged pole of indicated Z. mobilis cells under salt conditions. NS, not significant, ****P ≤ 0.0001 (unpaired t-test). Biological replicates N = 3. Sample size, n = 63 cells (Z6), n = 58 cells (ΔpatA), and n = 45 cells (ΔpatA + patA).

Elevated MepM increases salt resilience independent of PG O-acetylation

Our work has so far identified two key factors contributing to salt resilience in Z. mobilis, elevated active MepM and O-acetylation of PG. We were curious to know if these two phenomena are linked with each other. Specifically, we wanted to examine whether MepM-dependent salt resilience depends on the presence of O-acetyl groups in the PG. To address this question, we introduced the mepM overexpression plasmid (pBBR pdc-mepM) into non-PG-O-acetylated strain ∆patA and assessed its growth under the salt stress conditions. The analysis revealed that ΔpatA grew better with elevated MepM level (Fig. 6A and B). The ΔpatA pBBR pdc-mepM cells were smaller than wild type and ∆patA under no-salt growth conditions (Fig. 6C). Furthermore, under salt conditions, the “thin pole” phenotype of ΔpatA was reverted towards that of the wild type by overexpressing mepM (Fig. 6D). These results indicate that MepM-mediated salt resilience and PG remodeling are independent of the presence of O-Ac in PG.

Fig 6.

The mepM overexpression provides salt resilience in the patA mutant. (A) Growth profiles of Zymomonas mobilis strains Z6 (wild type) carrying pBBR plasmid, ΔpatA carrying pBBR plasmid, and ΔpatA carrying pBBR pdc-mepM, under salt conditions (left) and regular, no-salt growth conditions (right). (B) Phase contrast images of actively growing, Z6 (wild type) carrying pBBR plasmid, ΔpatA carrying pBBR plasmid, and ΔpatA carrying pBBR pdc-mepM under salt conditions (left) and regular, no-salt growth conditions (right) Scale bars, 10 μm. (C) A box-and-whisker plot showing the area of Z. mobilis cells of the three different strains grown to an OD₆₀₀ of 0.7–1.5 under no-salt conditions. The cell area was measured using MicrobeJ (21). Biological replicates N = 3. Sample size, n = 585 cells (Z6 + pBBR), n = 722 cells (ΔpatA + pBBR), and n = 677 cells (ΔpatA + pBBR pdc-mepM). NS, not significant; ****, P ≤ 0.0001 (unpaired t-test). (D) A box-and-whisker plot presenting the maximum diameter of bulged pole of indicated Z. mobilis cells under salt conditions. Biological replicates N = 3. Sample size, n = 70 cells (Z6 + pBBR), n = 58 cells (ΔpatA + pBBR), and n = 63 cells (ΔpatA + pBBR pdc-mepM).

DISCUSSION

Modulation of protein levels by the serine proteinase Ctp has been shown to be important in various biological processes across bacterial species, including virulence in pathogenic bacteria (11, 22, 34, 35), spore formation in Bacillus subtilis (36), and photosynthesis in cyanobacteria Synechocystis sp. PCC 6803 (37). Notably, Ctp was shown to play a significant role in opposing salt stresses in two gram-negative gamma-proteobacteria. A P. aeruginosa ctpA mutant grew poorer in LB lacking NaCl compared to the parental strain, while growth was comparable in LB with NaCl, suggesting a role of CtpA under low-salt conditions (22). In E. coli, the prc mutant exhibits slow growth, altered cell morphology, and frequent formation of spheroplasts under high-salt conditions (38), implying weakened peptidoglycan integrity in response to osmotic stress. The prc mutant is also sensitive to hypo-osmotic shock (39). Intriguingly, in sharp contrast to these two species, the Z. mobilis ctpA mutant grows significantly better than the parental strain under high-salt conditions (Fig. 1). Despite these differences, a conserved theme appears to exist in all three species: Ctp degrades the endopeptidase MepM. Our study provides the first indication that the regulation of a PG endopeptidase is conserved beyond the gamma-proteobacteria and perhaps throughout proteobacteria. However, more experiments are needed to test the idea of a generalized endopeptidase degradation pathway across bacteria.

We showed that increased levels of active MepM alone significantly enhance salt resilience in Z. mobilis (Fig. 3), albeit the level of salt resilience in Z6 + pBBR pdc-mepM did not fully reach that of ΔctpA (Fig. 3A). This difference in salt resilience could be caused by one of the following reasons or both: (i) overproduction of MepM from the plasmid pdc-mepM was not at the same level as in the ctpA mutant, and (ii) other possible CtpA substrates contribute to the salt resilience phenotype. It remains unclear which of these explanations accounts for the observed difference. Interestingly, the mepM overexpressing strains also exhibited another phenotype of ΔctpA, a reduction of cell size under no-salt condition (Fig. 3C). Taken together with the proteomics data (Fig. 2), these results indicate that elevated MepM activity is a major factor underlying both the enhanced salt resilience and the reduced cell size observed in the ΔctpA mutant. As MepM appears to be essential for viability, it poses the intriguing question of why the essential endopeptidase is constantly degraded by CtpA. One possibility is that CtpA might have a spatial control of MepM, directing its activity to specific sites within the periplasm and degrading it elsewhere.

While the exact mechanism of MepM-mediated salt resilience remains unclear, it is possible that salt stress compromises the total endopeptidase activity, leading to inefficient peptidoglycan remodeling. Insufficient endopeptidase activity, possibly together with disturbed cell division, likely contributes to the elongated morphology with bulged cell shape under the stress (Fig. 1B). The elevated MepM abundance may compensate for an impaired activity of PG hydrolases needed for PG growth, explaining the enhanced growth observed in ΔctpA and Z6 pdc-mepM cells (Fig. 7).

Fig 7.

Schematic view of improved salt resilience in ΔctpA and Z6 pdc-mepM strains. (Top) Under salt condition, wild-type cells exhibit poor growth possibly because the cells are unable to efficiently cleave cross-linked peptides in peptidoglycan. This deficiency hinders the insertion of nascent glycan chains for PG growth. The level of the essential (26) endopeptidase MepM is low and in part regulated by CtpA-mediated proteolysis. (Bottom) Both ΔctpA and Z6 pdc-mepM grow better than the wild-type under salt conditions due to higher MepM abundance and total endopeptidase activity. This allows more efficient space-making for nascent PG chains to be inserted, supporting cell growth. Another crucial factor contributing to salt resilience is peptidoglycan O-acetylation, which appears to act independently of endopeptidase-mediated stress resilience.

Our study also revealed that the non-pathogenic gram-negative Z. mobilis O-acetylates its PG and this modification is important for the salt resilience, a function never observed in previously studied bacteria with O-acetylated PG. PG O-acetylation at MurNAc has long been known to confer resistance to lysozyme in pathogenic Gram-positive bacteria (40). Gram-negative bacteria possess an outer membrane that limits the access of lysozyme to PG, and O-acetylation has been proposed to modulate the activity of lytic transglycosylases (40). Under salt conditions, O-acetylation might become crucial in regulating activity of lytic transglycosylases. Furthermore, we showed that salt resilience conferred by enhanced MepM does not rely on the presence of O-Ac (Fig. 6), implying that multiple factors contribute to salt resilience.

Overall, this study highlights the importance of dissecting cell envelope biology as a promising strategy for engineering Z. mobilis strains with improved stress tolerance to enhance their potential for industrial applications. Furthermore, from a fundamental perspective, Z. mobilis appears to exhibit a distinct mode of cell envelope organization, which will likely offer unique insights that cannot be obtained by studying model organisms.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions

Bacterial strains and plasmids used in this study are listed in Tables S1 and S2. Oligonucleotides used in this study are listed in Table S3.

Z. mobilis strain ATCC 29191 (Z6) was grown at 30°C in RM medium (Bacto yeast extract [5 g/L], glucose [20 g/L], NH₄SO₄ [1 g/L], KH₂PO₄ [1 g/L], MgSO₄ [0.5 g/L]) in a cap-enclosed standing tube (15 mL). For muropeptide profiling and growth assay, the RM growth medium was flushed with nitrogen gas for 45 min to remove oxygen. To follow growth, we measured the optical density of the cultures over time, which has previously been shown to be a proxy for biomass increase in elongated Z. mobilis cells, validated by comparison with cell dry weight measurements (9).

Construction of Z. mobilis strains

For genetic manipulations, we employed the method as described in reference 16, with a modification of the selection of cells that had undergone second homologous recombination (plasmid excision from the chromosome). To construct plasmids for gene manipulation, the upstream and downstream regions (500–1,000 bp) of the gene of interest were amplified by Q5 DNA polymerase (New England Biolabs). The obtained cassettes were inserted into the plasmid pPK15534 by Gibson Assembly (New England Biolabs). Sequence of constructed plasmid was verified by Sanger sequencing. The constructed suicide vectors were introduced into Z. mobilis strain Z6 via mating with the E. coli strain WM6026. Z. mobilis cells carrying the integrated plasmid in their chromosomal DNA were selected by chloramphenicol 90 μg/mL in RM medium agar. The integration of the plasmid into the chromosomal DNA was confirmed by colony PCR. For excision of the plasmid, the Z. mobilis cells were grown in liquid RM medium without selection for overnight, and the grown cultures were subsequently selected for the loss of plasmid. In addition to fluorescence-activated cell sorting for selection, we used B. subtilis SacB-based counter selection, a widely used method in other bacterial systems. This method was proven to be effective in Z. mobilis (manuscript in preparation). The genomic DNA of ΔctpA and ΔpatA mutants was sequenced by whole-genome sequencing (MicrobesNG) and analyzed by CLC Genomics Workbench (Qiagen).

Microscopy and image analysis

A culture with growing Z. mobilis (0.5 µL) was spotted onto a 1% agarose pad made of RM growth medium and covered by a cover glass. Phase contrast image of Z. mobilis cells was captured using a Nikon Eclipse Ti microscope (Nikon) equipped with Prime sCMOS camera (Teledyne Photometrics). Image was acquired using Nikon NIS elements AR software. Image analysis was performed using MicrobeJ (21). Unpaired t-test was applied to assess statistical significance.

Proteomics

Samples for proteomics were prepared as follows. Z. mobilis strains were grown anaerobically in RM medium cap-enclosed standing tubes (15 mL) to an OD₆₀₀ of 1.0. For salt stress conditions, RM medium was supplemented with 225 mM NaCl. Cells were harvested by centrifugation at 15,000 × g for 1 min at 4°C, washed once with ice-cold phosphate-buffered saline (PBS), and the cell pellets were stored at −80°C until proteomics analysis. Frozen cell pellets were lysed in cell lysis buffer containing 5% sodium dodecyl sulfate (SDS, Sigma-Aldrich), 50 mM triethylammonium bicarbonate (TEAB, Sigma-Aldrich), pH 8.5. Samples were sonicated on ice using a probe sonicator (3 × 30 s pulses with 30-s intervals) to ensure complete protein extraction and kept on ice between processing steps. Protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) with bovine serum albumin as a standard, performed in triplicate and averaged. Protein amounts were normalized to 50 μg per sample, and sample volumes were adjusted to 23 μL with lysis buffer when necessary for consistent downstream processing.

Proteins were reduced with tris(2-carboxyethyl)phosphine (TCEP, Sigma-Aldrich) at a final concentration of 5 mM and incubated at 55°C for 15 min, then cooled to room temperature. Alkylation was performed with iodoacetamide (IAA, Sigma-Aldrich) at a final concentration of 20 mM and incubated at room temperature for 15 min. Following reduction and alkylation, samples were acidified to a final concentration of 2.5% phosphoric acid (85% wt in H₂O, Sigma-Aldrich). Acidified samples were loaded onto S-Trap micro spin columns (Protifi) and digested according to the manufacturer’s protocol using sequencing-grade modified trypsin (Promega Gold) at a 1:10 enzyme-to-protein ratio (wt/wt) and incubated for 2 h at 47°C. After digestion, samples were loaded onto EVOTIPS following the EVOSEP tip loading protocol. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis was performed using an EVOSEP ONE liquid chromatograph (EVOSEP) coupled to a timesTOF HT mass spectrometer (Bruker Daltonics) via the CaptiveSpray source. Chromatographic separation was achieved using an IonOpticks Aurora Elite 15 × 75 mm C18 UHPLC column operated in Whisper Zoom configuration. The EVOSEP 20SPD method was employed with mobile phase A consisting of aqueous (Sigma-Aldrich, UHPLC Grade) 0.1% formic acid and mobile phase B consisting of 100% acetonitrile (Sigma-Aldrich, UHPLC Grade) with 0.1% formic acid. The gradient profile ramped mobile phase B from 2% to 40% over 60 min at a flow rate of 200 nL/min during the gradient phase. Mass spectrometric analysis was performed in positive ion mode using data-independent acquisition (DIA) with a method optimized using the pydiAID algorithm (41). The ion source was operated with the following parameters: capillary voltage, 1.6 kV; dry gas flow, 3 L/min; and dry gas temperature, 180°C. DIA acquisition was performed with precursor ion isolation windows spanning m/z 400–1,000 with variable window widths optimized for even precursor distribution. The accumulation and ramp time were set to 100 ms per scan, resulting in a total cycle time of approximately 1.7 s. Trapped ion mobility spectrometry (TIMS) was enabled with 1/K₀ values ranging from 0.6 to 1.6 V·s/cm². The diaPASEF method employed for the LC-MS analysis of the samples is provided alongside the raw files in the PRIDE repository. Raw mass spectrometry data were processed using DIA-NN version 2.0.1 (42) with a predicted spectral library generated from the Z. mobilis Z6 database and contaminant database (cRAP). The following parameters were applied: precursor and protein false discovery rate (FDR) set to 0.01, maximum number of variable modifications set to 1, maximum missed cleavages set to 1, mass accuracy tolerance set to 15 ppm for both MS1 and MS2, N-terminal methionine excision enabled, and tryptic digest specificity (K*, R*). Variable modifications included oxidation of methionine (UniMod: 35, +15.994915 Da) and N-terminal acetylation (UniMod: 1, +42.010565 Da). Cysteine carbamidomethylation was applied as a fixed modification. Match-between-runs (MBR) was enabled for enhanced peptide identification across samples. Protein grouping was performed using implicit protein grouping based on protein names, and contaminant proteins tagged with “cRAP-” were excluded from normalization and quantification. Peptidoform scoring and RT profiling were enabled to improve identification confidence. Results of the proteomics, as well as the raw data and LC-MS methods (embedded in the raw data files) are made available on ProteomeXchange repository (43) via PRIDE database (44). Prediction of protein localization was assessed using Signal IP 6.0 (45)

Muropeptide analysis

A Z. mobilis culture (400 mL) was anaerobically grown to an optical density of 0.6–0.8. The cells were harvested by centrifugation, resuspended in ice-cold PBS, and subsequently boiled in 4% SDS. Purification of peptidoglycan and high-performance liquid chromatography (HPLC) was performed as described (29, 46), with a modification of use of buffers and mobile phase with a pH of 6.0 to preserve O-acetyl groups. HPLC buffer B contained 15% methanol instead of 30% methanol as in reference 46. Mass spectrometry (MS) analysis for the identification of muropeptide was performed as described in reference 46.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.02350-25.

Supplemental material. aem.02350-25-s0001.docx.

Fig. S1 to S6; Tables S1 to S4.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.