Introduction

Antimicrobial resistance (AMR) is a critical global health crisis, with bacterial AMR implicated in approximately 4.95 million deaths in 2019, and predictions estimating up to 10 million deaths per year by 2050 (1, 2). Developing novel antimicrobials is thus of paramount importance but also a highly challenging task. Bacterial cell walls, mainly composed of peptidoglycan (PG) and crucial for survival, serve as a primary target for antimicrobials. However, the discovery of conventional PG-inhibiting small molecules is impeded by fewer choices and increasing resistance (3, 4). The naturally occurring PG-lysing enzymes, secreted by bacteria for competition in polymicrobial environments (5–8), remain underexplored for AMR treatment.

As a widespread molecular weapon in Gram-negative bacteria, the type VI secretion system (T6SS) can directly inject PG-lysing enzymes and other antibacterial toxins into neighboring microbes and anti-eukaryotic toxins to yeast and mammalian cells(7, 9–15). The conserved structure of T6SS consists of a double tubular structure, a baseplate TssEFGK complex, and a transmembrane TssJLM complex (16–19). A spear-like inner tube, made of stacks of Hcp hexamers, is ejected outward upon contraction of an outer sheath made of stacks of TssB/C hexamers, and a VgrG-PAAR spike complex sitting atop the Hcp tube is delivered concomitantly (12, 20, 21). Although the ejected Hcp spear can penetrate neighboring bacterial cells, T6SS-elicited damages are largely dependent on the secreted effectors rather than the penetration (22–24). Numerous T6SS effectors have been predicted in Gram-negative bacteria, and known effectors display diverse functions including damaging the cell wall and membrane, DNase, and metal scavenging (6, 25–30). Despite the functional diversity of T6SS effectors, PG-targeting effectors seem to be most commonly associated with the T6SSs.

The conserved structure of PG comprises repeating disaccharide units of N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM), with a short peptide (most commonly _L-Ala-_D-Glu-m-DAP-_D-Ala-_D-Ala) attached to each NAM (31). Adjacent peptide chains of PG are cross-linked, forming a continuous network that envelops bacterial cells (31). T6SS effectors known to target PG fall into two categories, T6SS amidase/endopeptidase (Tae) families cleaving the peptides and T6SS glycoside hydrolase (Tge) families cleaving the glycan backbone. Tae effectors have been found to cleave bonds between _D-Glu and mDAP, between mDAP and _D-Ala, and between NAM and _L-Ala (7, 27, 32, 33). Tge effectors possess β-(1–4)-N-acetyl-muramidase activities and N-acetylglucosaminidase activities (5, 6, 27, 34). The recent discovery of Tse4 in Acinetobacter baumannii, cleaving both the glycan strand and the crosslinked peptide cross-bridge, suggests dual enzymatic activities could be found in a single effector (10). However, such bifunctionality among effectors is rare.

As a ubiquitous waterborne pathogen, Aeromonas dhakensis can cause soft tissue infection and bacteremia in fish and humans (35, 36). The T6SS of A. dhakensis type strain SSU is active under laboratory conditions and can secrete three known antimicrobial effectors, a membrane-targeting TseC, a lysozyme-like TseP, and a nuclease TseI (25, 37–40). Our previous findings have established that TseP serves a structural role because its expression can restore the T6SS secretion in the Δ3eff mutant lacking all three effector genes, and the C-terminal domain, TseP^C, possesses lysozyme activities (38). However, the function of the N-terminal domain, TseP^N, which lacks any recognizable conserved motif, has remained elusive. This study unveils TseP^N is capable of cleaving the peptide link between NAM and _L-Ala as an amidase. Both TseP^N and TseP^C can be secreted separately by the T6SS, but only TseP^N can compensate for T6SS defects, indicating its critical role as a structural component. Additionally, TseP^N exhibits approximately 10% higher GC content than TseP^C, and homologs of TseP^N and TseP^C exist independently in divergent species. This disparity in GC content and the presence of domain-specific homologs suggest an evolutionary fusion event. We further show that TseP^C can be engineered to hydrolyze resistant Gram-positive Bacillus subtilis cells, conferring the T6SS with the capability to kill Gram-positive bacteria in a contact-independent manner. This research identifies a novel class of amidase-lysozyme bifunctional effectors, provides insights into effector evolution, and highlights the potential of engineering these effectors as antimicrobials.

Results

TseP^N and TseP^C can be independently secreted by the T6SS

To delineate the regions of TseP crucial for T6SS-mediated secretion, we constructed two TseP truncations, TseP^N and TseP^C, and then expressed them in the Δ3eff mutant using the pBAD arabinose-inducible vectors. Consistent with our previous results, protein secretion assays show that the Δ3eff mutant, akin to the T6SS null ΔvasK, failed to secrete Hcp (Figure 1A). Complementation with TseP and TseP^N, but not TseP^C, restored Hcp secretion, despite at a reduced level relative to the wild type (Figure 1A). We next tested whether the T6SS assembly is restored using a chromosomal sheath-labeled construct TssB-sfGFP. Fluorescence microscopy analysis showed that complementation with both TseP and TseP^N, but not the TseP^C, restored sheath assembly (Figure 1, B and C). Because both Hcp secretion and sheath assembly are hallmarks of T6SS assembly, these findings highlight the important structural role of the TseP N-terminus.

Figure 1.

Interestingly, we noticed that TseP^N, but not TseP^C, was secreted (Figure 1A). The lack of TseP^C secretion might be due to a defective T6SS. To test this, we expressed different plasmid-borne TseP constructs in a panel of SSU mutant strains. Western blot analysis not only confirmed the secretion of TseP and TseP^N but also detected the secretion of TseP^C, albeit to a lesser extent, in both wild-type and the ΔtseP cells (Figure 1D). Considering that VgrG2 is the carrier protein for TseP delivery (38), we next tested whether TseP^N and TseP^C could interact with VgrG2. Pull-down analysis showed direct interaction of VgrG2 with TseP^N and TseP^C, in contrast to the sfGFP and MBP negative controls (Figure 1E). These results indicate TseP^N and TseP^C can be secreted independently by the T6SS.

TseP^N is a Zn²⁺-dependent amidase

When analyzing TseP-mediated PG hydrolysis products using ultraperformance liquid chromatography coupled with mass spectrometry (UPLC-MS), we detected two distinctive peaks and determined their corresponding products to be NAM-NAG disaccharides and Tri/Tetra peptides (Figure 2, A and B), the latter suggesting an amidase activity. The released peptides were also detected in samples treated with TseP^N and the catalytically inactive C-terminal lysozyme mutant TseP^E663A, but not with TseP^C or its inactive TseP^C-E663A mutant, indicating that the amidase activity is attributed to TseP^N.

Figure 2.

Sequence comparison with known amidases suggests that the TseP^N contains a predicted Zn²⁺ binding domain with several conserved residues H19, H23, H109, H339, H359, and E539 (Figure 2C). To test the effect of these residues on the TseP^N amidase, we replaced these sites with alanines using site-directed mutagenesis and then purified a set of TseP^N mutant proteins. Subsequent PG-digestion assays, coupled with UPLC-MS analysis, showed significantly reduced amidase activity in these mutants compared to TseP^N (Figure 2D).

To test the role of metal ions for TseP’s amidase activity, we added the chelating agent EDTA to the PG-digestion mixture, which abolished the amidase activity (Figure 2E). To further identify the specific metal ion involved, we also purified TseP in the presence of EDTA and then supplemented it with Zn²⁺, Ca²⁺ and a combination of both. PG-digestion assays revealed that adding Zn²⁺, but not Ca²⁺, restored the amidase function (Figure 2E).

Given the dual functionality of TseP, we sought to determine if TsiP, known as the immunity protein neutralizing the lysozyme activity, could inhibit the amidase, by mixing purified TsiP with TseP, TseP^N, and TseP^C in varying molar ratios. Enzymatic assays using purified PG showed that TsiP could effectively inhibit both amidase activity and peptidoglycan hydrolysis activity of TseP, TseP^N and TseP^C (Figure 2F, Supplemental Figure S1). Additionally, pull-down assays showed that TsiP interacts with not only TseP^C but also TseP^N (Figure 2G). Collectively, these results indicate that TseP^N is a Zn²⁺-dependent amidase.

Amidase activity is not required for T6SS assembly or lysozyme function

To determine whether the amidase activity is important for the structural role of TseP^N, we expressed these amidase-inactive mutants in the Δ3eff TssB-sfGFP mutant and tested T6SS-related functions. Secretion analysis revealed that expression of TseP^H339A, but not the other mutants, restored Hcp secretion similar to wild-type TseP (Figure 3A). As control, we examined the expression of these TseP variants and found comparable signals in the cell lysates and secretion samples. Observations of TssB-sfGFP sheath assembly in the Δ3eff TssB-sfGFP mutant further supported these findings, with the TseP^H339A strain showing dynamic assembly of T6SS similar to the strain expressing the wild-type protein (Figure 3B, Supplemental Figure S2A). Further analysis through bacterial competition assays, with E. coli MG1655 strain as prey, mirrored these results (Figure 3C). The expression of TseP^H339A and TseP showed comparable killing abilities while the expression of other mutants demonstrated severely impaired killing activities as the T6SS defective Δ3eff (Figure 3C). Although it is unclear why the other mutations failed to complement, the H339A amidase-inactive mutation exhibiting a similar complementary role to wild type TseP suggests that the amidase activity is not required for T6SS assembly.

Figure 3.

Additionally, we tested whether there is any synergistic effect between amidase and the TseP^C-mediated lysozyme activities by comparing PG degradation treated with TseP, TseP^C, and TseP^H339A. Results showed that TseP exhibited higher activities than the TseP^C alone, but there was no difference between TseP and the TseP^H339A mutant (Figure 3D, Supplemental Figure S2B). This indicates that amidase and lysozyme activities can operate as independent modules within the protein.

TseP homologs exhibit similar dual functions

Homolog searching analysis reveals a broad distribution of TseP homologs across Gram-negative bacteria (Figure 4A, Supplemental Figure S3). To examine the dual functionality of these homologs, we selected two representative proteins that have not been previously characterized: AHA_1849 from Aeromonas hydrophila which is a closely related homolog from the same genus, and PSPTO_5204 from Pseudomonas syringae, a more distantly related homolog. To determine whether AHA_1849 and PSPTO_5204 share amidase and lysozyme activities as TseP, we expressed and purified their respective N- and C-terminal truncated mutants, AHA_1849^N (1-640aa), AHA_1849^C (641-851aa), PSPTO_5204^N (1-343aa), and PSPTO_5204^C (344-588aa) (Supplemental Figure S3). We also constructed inactivating mutations in their predicted catalytic residues for each protein to serve as controls (Supplemental Figure S3). Subsequent PG-hydrolysis and UPLC-MS analysis confirmed that both AHA_1849 and PSPTO_5204 exhibit dual amidase and lysozyme activities, capable of cleaving both peptide and glycan chains (Figure 4, B and C).

Figure 4.

Additionally, by analyzing the GC-contents of the two Aeromonas homologs, we found that the GC-content of TseP^N is about 10% higher than that of TseP^C and 10% lower than the upstream VgrG (Figure 4D). This suggests that these two domains might have been horizontally transferred and subsequently fused during evolution. To explore this hypothesis further, we searched for homolog proteins with similar structures to either TseP^N or TseP^C in the Foldseek Search AFDB50 database (41), yielding 1000 proteins with comparable predicted structures. We identified TseP^N-only homologs in various Aeromonas species, including a specific homolog in A. schubertii possessing only the TseP^N domain, whereas TseP^C-homologs were discovered in several other species, such as Fulvimonas soli and Vibrio campbellii (Figure 4E). The comparison of their predicted structures reveals high similarity (Figure 4F, Supplemental Table S2 and S3), further supporting the notion that TseP^N and TseP^C domains may have been fused through a recombination event, resulting in a chimeric, dual-functional TseP.

Crystallographic analysis of TseP^C reveals a wide active groove

While testing TseP-mediated hydrolysis of purified PG from Gram-negative bacteria, we noticed that both TseP and TseP^C displayed more efficient digestion than the commonly used hen-egg lysozyme (Figure 5A). This enhanced activity suggests that TseP may utilize a unique mechanism for cell wall degradation. To further investigate this, we determined the X-ray crystal structure of TseP^C to a resolution of 2.27Å (R_work =0.189 and R_free=0.227), by molecular replacement with one molecule in the asymmetric unit (Table 1).

Figure 5.

Table 1.

The crystal structure revealed that TseP^C comprises a smaller lobe of seven helices and a larger lobe of ten helices, separated by a concave groove (Figures 5B, Supplemental Figure S4A). The surface electrostatic potential of TseP^C, calculated using the Adaptive Poisson-Boltzmann Solver (APBS), consists of both electronegative and electropositive regions, unlike the predominantly electronegative groove of lysozyme (Figure 5C, Supplemental Figure S4B). The electropositive surface of TseP^C consists of residues R723 and K796, with an adjacent residue Y720 corresponding to W62 in lysozyme, altering which is known to affect substrate binding (42). Additionally, the previously predicted catalytic dyad (E655 and E663) was located near the groove. Although this structural arrangement resembles that of other lysozyme family members, TseP^C exhibits a less compact conformation with a notably wider groove compared to the typical lysozyme (Figure 5D). Previous studies have shown that the lysozyme possesses catalytic sites at E53 and D70 in the substrate-binding groove (42), while the active residues in the TseP^C structure are E654 and E663, respectively (Figure 5D). To test the effect of E663 for TseP^C activity, we purified a mutant E663D (TseP^C-E663D) protein and perform PG digestion assays. The results show that TseP^C-E663D lost the activity, suggesting the greater distance within the groove requires this glutamate residue (Figure 5E). These results collectively highlight the key structural features that may account for the increased activity of TseP^C relative to the hen-egg lysozyme.

Engineered TseP^C can lyse Gram-positive cells

To test whether TseP can degrade various cell walls, we purified PG from B. subtilis, a model Gram-positive bacterium, for use as substrate. However, we failed to detect any PG hydrolysis treated with TseP or TseP^C, in contrast to the lysozyme (Figure 6A). Further purification of PG with NaOH treatment, removing PG-associated peptides, led to effective hydrolysis of PG by TseP and TseP^C, suggesting a difference in substrate accessibility between the lysozyme and TseP. By comparing protein surface electrostatic potentials between the lysozyme and TseP^C, we noticed a more electronegative surface patch in TseP^C, suggesting a potential electrostatic repulsion from negatively charged PG (Figure 6B, Supplemental Figure S4B). To increase electropositive regions while minimizing impact on activity, we mutated four negative charge residues distant from the active site and obtained a quadruple mutant TseP^C4+ (Figure 6B). Activity assays confirmed that TseP^C4+ retained PG-hydrolyzing activities (Figure 6C). To test whether the engineered TseP^C4+ was able to lyse B. subtilis, we treated B. subtilis cells with purified TseP^C and TseP^C4+ and found that cells were lysed effectively by TseP^C4+ but not TseP^C (Figure 6D, Supplemental Figure S4C). These results indicate that modulating the surface charge is an effective strategy to expand the target range of PG-lysing effectors.

Figure 6.

Next, we tested whether the engineered TseP^C4+ could be delivered by the T6SS for inhibiting B. subtilis. Protein secretion analysis showed a significant TseP^C4+ signal in secretion samples, despite at a lower level than the wild type TseP^C (Figure 6E). When B. subtilis was co-incubated with the ΔtseP mutant expressing the engineered TseP^C4+ in liquid culture, survival of B. subtilis was significantly reduced in a T6SS-dependent manner (Figure 6F, Supplemental Figure S4D). These data collectively suggest that the T6SS can be armed with engineered TseP^C4+ to gain Gram-positive killing capabilities in a contact-independent manner.

Discussion

The T6SS plays a pivotal role in bacterial competition, using a variety of antibacterial effectors to kill competing cells. The exploration of effector functions holds the promise of unlocking a largely untapped source of novel antimicrobials, presenting innovative approaches to address the escalating antimicrobial resistance crisis. Here we report that TseP represents a novel family of dual-functional chimeric effectors, exhibiting both amidase and lysozyme activities. Our study reveals that the N-terminus of TseP not only serves a critical structural role in T6SS assembly but also acts as a Zn²⁺-dependent amidase. The N- and C-domains are both secreted and functionally independent, and may be subject to an evolutionary fusion event to form the chimeric full-length TseP. Such multifunctionality underscores a complex evolutionary strategy aimed at promoting T6SS-mediated competitiveness.

Among the known T6SS effectors, those targeting the bacterial PG layer represent a prevalent category. Certain T6SSs are capable of translocating multiple PG-targeting effectors. For instance, the H1-T6SS in Pseudomonas aeruginosa deploys Tse1 and Tse3, while Vibrio cholerae secretes VgrG3 and TseH (5–7, 9, 43, 44). Tse1 and TseH exhibit endopeptidase activities, whereas Tse3 and VgrG3 have potent lysozyme activities (5–7, 9, 43, 44). Yet, it remains relatively uncommon to identify effectors that possess dual functionality, capable of cleaving both the glycan chains and the cross-linked peptides within the bacterial cell wall. To date, the sole established example is Tse4 from A. baumannii, which acts both as a transglycosylase to cleave the glycan strand and as an endopeptidase to cut the peptide cross-bridge (10, 45, 46). Our enzymatic assays reveal that TseP distinguishes itself from Tse4 with dual roles as a lysozyme and an amidase, as well as a structural role for T6SS assembly. Further homology searches and biochemical validations of homolog activities substantiate that TseP constitutes a new family of dual-functional effectors.

Our data demonstrate that the N-terminus of TseP can modulate T6SS assembly, which provides further evidence that effectors play a crucial structural role in T6SS assembly (23, 37, 47, 48). Although this notion has now been shown in several species, the molecular details remain elusive. The assembly of a non-contractile mutant T6SS in V. cholerae in the absence of effectors suggests that effectors might be needed to stabilize a polymerizing T6SS (49). In comparison to TseP^C, the N-terminus not only contributes to T6SS assembly but also exhibits 10% higher GC content. Additionally, both N and C-terminus can be independently secreted by the T6SS. Using structure-assisted homolog search, we have identified TseP^N and TseP^C-only proteins in different species. These data suggest that these two domains may result from independent effector proteins and have undergone an evolutionary fusion event.

Previous studies have shown that expressing TseP in the periplasm is highly toxic, but this toxicity is mitigated when catalytic mutations are introduced into the C-terminal lysozyme domain (38). This indicates that the N-terminal amidase is either less toxic or is tightly regulated by the concentration of Zn²⁺ in the growth medium. Although the total intracellular Zn²⁺ concentration during growth in LB has been reported to be around 0.1mM—approximately 100 times higher than the Zn²⁺ level in the LB medium—the bioavailable (free) Zn²⁺ concentration within E. coli cells remains exceedingly low (50). A similar dependency on metal ions was observed with another cell-wall-targeting endopeptidase effector, TseH, whose activity is dependent on the environmental concentrations of Mg²⁺ and Ca²⁺ (51). Such metal ion dependencies might lead to misbelief about these effectors being cryptic or inactive, complicating the discovery and functional analysis of these important proteins.

TseP is predicted to contain an EF-hand calcium-binding domain with a “helix-loop-helix” motif, a feature previously unreported in T6SS effectors (Supplemental Figure S5A). However, mutations introduced to this domain, aimed at exploring its role in T6SS assembly and TseP toxicity, did not reveal any critical effect on TseP functions under the conditions tested (Supplemental Figure S5, B and C). It is possible that this EF-hand motif is active under more Ca²⁺ stringent conditions or during infections. Further study is required to decipher its physiological function. Additionally, while constructing TseP variants with different residue changes, we noticed that even single residue changes could reduce the expression, secretion or the structural role of TseP, highlighting the technical challenge of heterologous effector engineering (Figure 3A, Supplemental Figure S6, A and B).

Because the outer membrane of Gram-negative bacteria generally prohibits the entry of cell-wall lysing proteins, the relatively exposed Gram-positive bacteria are likely more amenable to such treatments. However, the T6SS, exclusively found in Gram-negative bacteria, primarily targets Gram-negative species. To date, only two T6SSs, identified in Acidovorax citrulli and A. baumannii (10, 13), have shown the capability to kill Gram-positive bacteria. Additionally, no Gram-positive cell wall lysing effectors have been reported; for instance, Tse4 from A. baumannii cannot directly lyse B. subtilis cultures but requires T6SS delivery. Our biochemical characterization and engineering of TseP^C not only establishes its exceptional enzymatic activity but also demonstrates an effective strategy to acquire Gram-positive-cell lysing activities, as well as to equip the T6SS of A. dhakensis with a contact-independent killing ability. These results highlight the potential of antimicrobial effector engineering in addressing antimicrobial resistance.

In conclusion, our study unveils a new family of dual functional chimeric T6SS effectors and demonstrates the potential for bioengineering these proteins to broaden their antimicrobial spectrum. The engineered TseP^C represents the first effector variant capable of directly lysing B. subtilis cultures, paying the way for engineering the diverse set of T6SS cell-wall targeting effectors. The urgent need for new antimicrobials necessitates the exploration of innovative strategies. The insights from our research into T6SS and its diverse effectors signal promising directions for future work in microbial biology and antimicrobial therapy. When enhanced through protein engineering, these enzymes have the potential to address bacterial infections not only in medical settings but also in food safety and animal health.

Materials and Methods

Bacterial strains and growth conditions

All strains and plasmids used in this study are displayed in Supplemental Table S1. Strains were routinely cultivated in Lysogeny Broth (LB) ([w/v] 1% tryptone, 0.5% yeast extract, 0.5% NaCl) under aerobic conditions at 37 ℃. Antibiotics were supplemented at the following concentrations: streptomycin (100 µg/ml), kanamycin (50 µg/ml), gentamicin (20 µg/ml), chloramphenicol (25 µg/ml for E. coli, 2.5 µg/ml for SSU). All plasmids were constructed using standard molecular techniques, and their sequences were verified through Sanger sequencing. All strains and plasmids are available upon request.

Protein secretion assay

Secretion assays were performed as previously described (38). Briefly, cells were grown in LB to OD₆₀₀=1 and collected by centrifugation at 2,500 × g. The pellets were resuspended in fresh LB and incubated at 28 ℃ for 2 h with 0.01% (w/v) L-arabinose. After incubation, cells were pelleted again by centrifugation at 2,500 × g for 8 min. Cell pellets were then resuspended in SDS-loading dye and used as cell lysate samples. The supernatants were centrifuged again at 12,000 × g for 30 s and then precipitated in 10% (v/v) TCA (trichloroacetic acid) at −20 ℃ for 30 min. Precipitated proteins were collected by centrifugation at 15,000 × g for 30 min at 4 °C, and then washed twice with acetone. The air-dried pellets were resuspended in SDS-loading dye and used as the secreted protein samples. All the samples were boiled for 10 min before SDS-PAGE analysis.

Western blotting analysis

Proteins were separated using SDS-PAGE and subsequently transferred onto a PVDF membrane (Bio-Rad). Prior to antibody incubation, the membrane was blocked with a solution of 5% (w/v) non-fat milk dissolved in TBST buffer (consisting of 50 mM Tris, 150 mM NaCl, 0.1% [v/v] Tween-20, pH 7.6) for 1 h at room temperature. The membrane was then incubated sequentially with primary and secondary antibodies. The Clarity ECL solution (Bio-Rad) was utilized for signal detection. Monoclonal antibodies specific to RpoB (RNA polymerase beta subunit) were sourced from Biolegend (Product # 663905). Additionally, polyclonal antibodies to Hcp were customized by Shanghai Youlong Biotech (39). HRP-linked secondary antibodies were purchased from Beyotime Biotechnology (Product # A0208 and # A0216, respectively).

Protein expression and purification

Proteins were expressed using pET28a vector with the His tag in E. coli BL21 (DE3). Strains containing the recombinant plasmid were grown in LB at 37 °C. When the optical density of the culture reached OD₆₀₀ of 0.6, isopropyl-β-_D-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM, and the culture was incubated at 20 °C for 16 h. The cells were collected by centrifugation at 2,500 × g for 10 minutes at 4°C, resuspended in lysis buffer (20 mM Tris-HCl, 300 mM NaCl, 10 mM imidazole, pH 8.0), and lysed by sonication. After sonication, the lysate was centrifuged at 12,000 × g for 60 min at 4 °C, and the supernatant was collected. Ni-NTA affinity chromatography was used to purify the target protein, followed by gel filtration using Superdex 200 pg 10/300 column (GE Healthcare). The protein was finally concentrated to 15 mg/ml. All samples were visualized by SDS-PAGE and stained with Coomassie brilliant blue dye. Gel images for all purified proteins used in this study are shown in Supplemental Figure S7.

Crystallization, X-ray data collection and structure determination

TseP^C was crystallized at 18 °C using the hanging drop method. Equal volumes (1:1) of the protein solution and reservoir solution, containing 0.1 M sodium formate (pH 7.0) and 12% (w/v) PEG 3350, were mixed. The crystals were cryoprotected in a mother liquor containing 20% v/v glycol and were flash-cooled at 100 K.

An X-ray diffraction data set of TseP^C was collected at beamline BL19U1 (wavelength, 0.97852 Å) at the Shanghai Synchrotron Radiation Facility. Diffraction data was auto-processed by an aquarium pipeline. The crystals were diffracted to 2.27 Å and belonged to the space group C 2 2 21 with unit cell dimensions of a = 44.48 Å, b = 138.8 Å and c = 95.83 Å. This structure was solved by molecular replacement with PHASER using the previously reported SPN1S endolysin structure (PDB ID: 4OK7) as the searching model and further polished by COOT and PHENIX. refine (52, 53). All structure figures were generated using PyMOL (https://pymol.org/2/) software (54).

Peptidoglycan purification

Peptidoglycan was extracted as previously described (55). Bacteria were cultured at 37 °C for 8 h. Cells were then centrifuged at 4 °C for 30 min at 3,000 × g and the cell pellets were washed with ultrapure water. After washing, the bacteria were resuspended in ultrapure water to an OD₆₀₀ of 70-100 and then added dropwise to an equal volume of boiling 8% SDS solution. The mixture was slowly stirred while boiling for 3 h and then cooled naturally to room temperature. The precipitate was collected at 20 °C for 1 h at 100,000 × g and washed with ultrapure water five times. To remove the residual proteins and DNA, the precipitate was resuspended in 10 ml of 10 mM Tris-HCl (pH 7.0) and treated with 5 mg DNase I and 16 mg trypsin at 37 °C for 16 h. After treatment, the mixture was centrifuged at room temperature for 1 h at 100,000 × g, and the pellets were dissolved and treated with an equal volume of boiling 8% SDS solution for another 3 h. The purified peptidoglycan was obtained after washing with ultrapure water, freeze-dried, and stored at −20℃.

In vitro enzymatic assays and enzyme kinetics

To characterize the in vitro enzymatic activity of TseP and its homologs, 500 μg of purified E. coli PG was co-incubated with an equal amount of proteins in a 0.1 ml reaction system at 37 °C for 12 h. After centrifugation, the supernatant was collected and subjected to analysis by ultraperformance liquid chromatography-quadrupole time of flight mass spectrometry (UPLC-QTOF-MS). The Acquity UPLC HSS T3 column was chosen as the solid phase, with solvent A (water with 0.1% [v/v] formic acid) and solvent B (methanol) as the mobile phase. Analyte separation was achieved by a linear gradient of solvent B, which was gradually increased from 1% to 20% over 65 min. Thereafter, a rapid gradient adjustment was followed to increase the concentration of solvent B from 20% to 100% in 1 min, with a retention period of 5 min to ensure complete separation. Finally, the system was flushed with 99% solvent A for 4 min to complete the process.

To measure the enzyme kinetics of TseP^C, a dye-release method was used to detect the PG hydrolyzing. Prior to use, the purified PG was labelled by Remazol Brilliant Blue (RBB) following a previously described method (56). Briefly, purified PG was incubated with 20 mM RBB in 0.25 M NaOH for 12 h at 37°C. The RBB-labelled PG was collected via centrifugation for 10 min at 21,400 × g and washed with ultrapure water several times until no RBB dye could be detected in the supernatant. This RBB-labelled PG was then applied for the enzyme kinetics analysis. In each reaction, 10 nM enzyme was mixed with different concentrations of RBB-labelled PG at 37°C for 10 min. Undigested PG was pelleted by centrifugation at 21,400 × g for 10 min and the dye release in the supernatant was quantified by measuring its absorbance at OD₅₉₅.

Bacterial competition assays

Killer and prey strains were grown aerobically in liquid LB medium until they reached OD₆₀₀ values of 1 and 2, respectively. Later, the cells were harvested via centrifugation and gently resuspended in fresh LB broth. Killer and prey cells mixed at a ratio of 5:1 were deposited onto LB-agar plates supplemented with 0.01% (w/v) L-arabinose and co-incubated for 6 h at 28°C. Afterward, the mixed cells were recovered in fresh LB broth and vigorously vortexed to dislodge the agar. For the Gram-positive bacterial killing, the SSU and B. subtilis cells were mixed at a ratio of 20:1 with 0.01% L-arabinose in liquid LB medium for 24 h at 28°C. The mixtures were then serially diluted and plated onto specific antibiotic-containing plates.

Bioinformatics analysis

The gene sequences of A. dhakensis SSU were retrieved from the draft genome assembly (GenBank accession NZ_JH815591.1) and confirmed by Sanger sequencing. The resulting sequences were then aligned using Clustal Omega (57), and the alignment was visualized using ESPript with its default settings (https://espript.ibcp.fr)(58).

Data availability

The data that support the findings of this study are available within the article or from the corresponding author upon reasonable request.

Protein Data Bank and accession code

The atomic coordinate and structure factor of TseP^C have been deposited in the Protein Data Bank with accession code 8XCL.

Acknowledgements

This work was supported by funding from National Natural Science Foundation of China (32030001) and National Key R&D Program of China (2020YFA0907200). The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Author contributions

T.D. conceived the project. Z-H.W., T-T.P., D.Y.W., and X.L. performed research and data analysis. Z--H.W., Y.A., T.Z., and W.Q. performed the crystallography and structural analysis. T.D. and Z-H.W.. wrote the manuscript with assistance from Y.A.

Competing interests

The authors declare no competing interests.