Pseudomonas fluorescens MFE01 delivers a putative type VI secretion amidase that confers biocontrol against the soft-rot pathogen Pectobacterium atrosepticum

The type VI secretion system (T6SS) is a contractile nanomachine wide-spread in Gram-negative bacteria. The T6SS injects effectors into target cells including eukaryotic hosts and competitor microbial cells and thus participates in pathogenesis and intermicrobial competition. Pseudomonas fluorescens MFE01 possesses a single T6SS gene cluster that confers bio-control properties by protecting potato tubers against the phytopathogen Pectobacterium atrosepticum (Pca). Here, we demonstrate that a functional T6SS is essential to protect potato tuber by reducing the pectobacteria population. Fluorescence microscopy experiments showed that MFE01 displays an aggressive behaviour with an offensive T6SS characterized by continuous and intense T6SS firing activity. Interestingly, we observed that T6SS firing is correlated with rounding of Pectobacterium cells, suggesting delivery of a potent cell wall targeting effector. Mutagenesis coupled with functional assays then revealed that a putative T6SS secreted amidase, Tae3 Pf , is mainly responsible for MFE01 toxicity towards Pca. Further studies finally demonstrated that Tae3 Pf is toxic when produced in the peri-plasm, and that its toxicity is counteracted by the Tai3 Pf inner membrane immunity protein.


INTRODUCTION
Pectobacterium atrosepticum (Pca) ranks in the top 10 soft-rot phytopathogenic bacteria (Mansfield et al., 2012).Pca infects potato plant in cold and temperate countries, causing tuber soft-rot and blackleg disease on the plant stem leading to huge crop production and economic losses (Crépin et al., 2012;Dupuis et al., 2021;Mansfield et al., 2012;Pérombelon, 2002;Smadja et al., 2004;Toth et al., 2021).After invading plants via wounded sites or natural openings, Pca cells undergo a transition to a pathogenic state, associated with increased growth rate.Following this phase dictated by quorum sensing, Pca secretes a massive amount of plant cell wall degrading enzymes (PCWDEs) to degrade plant tissues (Barnard & Salmond, 2007;Van Gijsegem et al., 2021).Although preventive treatments exist, such as potato seed sterilization or utilization of certified potato seeds, there is no option to cure these deleterious diseases if they occur in the field.One promising answer to combat these phytopathogens is the development of biocontrol strategies using antagonistic microorganisms (Diallo et al., 2011).Among them, bacteria grouped into the term 'Plant Growth-Promoting Rhizobacteria' (PGPR) develop beneficial relationships with plants (Turner et al., 2013).PGPR can directly stimulate plant growth by synthesizing hormones (phytostimulators) or by supplying the plants with nutrients (biofertilizing).Growth stimulation can also be indirectly achieved by suppressing or preventing the deleterious effects of pathogens (Diallo et al., 2011).Among PGPR, fluorescent pseudomonads are well recognized for their beneficial properties to protect plants including biocontrol activities (Haas & Défago, 2005).Most Pseudomonas species produce a type VI secretion system (T6SS), a multiprotein transport machine involved in bacterial antagonism (Bernal et al., 2018).By eliminating microbial rivals, the T6SS confers an advantage to adapt to different environmental niches (Gallegos-Monterrosa & Coulthurst, 2021;Gallique, Bouteiller, et al., 2017;Lin et al., 2017).
The T6SS is a widespread apparatus in Gramnegative bacteria (Bingle et al., 2008;Boyer et al., 2009;Cascales, 2008).The T6SS uses a contractile mechanism to inject a needle loaded with effectors directly into target eukaryotic or bacterial cells or to expel effectors into the environment to collect metals (Basler, 2015;Coulthurst, 2019;Hernandez et al., 2020;Ho et al., 2014;Jurenas & Journet, 2021;Yang et al., 2021).The T6SS is comprised of 13 corecomponent proteins and, in some instances, of additional, accessory subunits.T6SS assembly starts with the formation of the TssJLM membrane complex, which recruits the baseplate that serves as assembly platform to polymerize a tail structure (Cherrak et al., 2019;Cianfanelli et al., 2016;J. Wang et al., 2019).The tubular tail is made of an inner tube wrapped by a sheath structure assembled in an extended conformation.Contraction of the TssBC sheath propels the inner Hcp tube, tipped by a spike and loaded with effectors outside of the cell (Basler, 2015;Brackmann et al., 2017;Cherrak et al., 2019;Cianfanelli et al., 2016;Taylor et al., 2018;J. Wang et al., 2019).A large arsenal of T6SS effectors has been described with a broad repertoire of activities targeting essential macromolecules such as phospholipases, nucleases, proteases, peptidoglycan hydrolases, or ADP-ribosyltransferases (Hernandez et al., 2020;Jurenas & Journet, 2021;Russell et al., 2014).
In this study, we carried out tuber assays to provide further insights into the role of MFE01 T6SS for tuber protection and bacterial antagonism against Pca.We then analysed MFE01 T6SS dynamics through timelapse fluorescence microscopy.Finally, we identified and characterized a major antibacterial effector with putative amidase activity, implicated in Pca antagonism.

Strain construction
Construction of the MFE01 chromosomal tssB-sfGFP translational fusion-A DNA fragment encompassing the 3 0 end of the tssA gene (460 bp), the complete tssB gene fused to the sfGFP sequence and the 5 0 end of the tssC gene (959 bp) was synthetized by Genewiz, amplified by PCR (oligonucleotides listed in Table S2), and cloned into the SmaI-digested pAKE604 suicide vector (El-Sayed et al., 2001).The construct, verified by DNA sequencing, was introduced into E. coli S17-1 (Simon et al., 1983) and transferred into P. fluorescens MFE01 by biparental mating, as previously described (Decoin et al., 2015).The sf gfp insertion downstream of the tssB gene by a double crossover event was verified by PCR and sequencing.
Construction of the MFE01 Δtae3 mutant strain-The markerless tae3 deletion mutant (Δtae3) was constructed as previously described (Decoin et al., 2015).Briefly, the tae3 upstream and downstream regions were amplified by PCR (Phusion ® High Fidelity DNA polymerase, NEB) using the oligonucleotide pairs muta1amidase/muta2amidase and muta3amidase/ muta4amidase (Table S2), respectively.The two fragments were cloned in tandem into the pAKE604 suicide vector before being transferred into MFE01 by biparental mating and selection of the double recombination event.This in-frame tae3 deletion mutant, containing a truncated version of the tae3 gene, was checked by PCR analysis and DNA sequencing.

Plasmid construction
Plasmid pJN-Tae3 was constructed by ligation of an EcoRI/XbaI-digested PCR fragment comprising the tae3 gene into the pJN105 vector (Newman & Fuqua, 1999) digested by the same enzymes, downstream the arabinose-inducible ParaBAD promoter.Plasmids pBAD-Tae3 Pf -VSVG, pBADtat-Tae3 Pf -VSVG and pTrc-ST-Tai3 Pf , encoding respectively Tae3 Pf fused to a C-terminal VSV-G tag, Tae3 Pf -VSVG fused to an N-terminal TorA signal sequence, and Tai3 Pf fused to an N-terminal StrepTag, were engineered by restriction-free cloning (van der Ent & Löwe, 2006) as previously described (Aschtgen et al., 2008).Briefly, the fragments encoding Tae3 Pf and Tai3 Pf were amplified by PCR using oligonucleotides (Table S2) that introduced extensions annealing to the target vector.The double-stranded products of the first PCR were then used as primers for a second PCR using the target vectors (pBAD33, pBAD33tat [D.Jurenas, unpublished vector] and pTrc99A [Pharmacia], respectively) as templates.PCR products were treated with DpnI to eliminate template plasmids and transformed into DH5α-competent cells.For Tae3 Pf cloning, plasmids were selected on LB agar plates supplemented with 1% of glucose.Substitutions were introduced by Quickchange site-directed mutagenesis using complementary oligonucleotides bearing the desired mutations.All plasmids have been verified by colony PCR and DNA sequencing.

In vitro antibacterial competition assay
Antibacterial competition assays were carried out as previously published (Murdoch et al., 2011), with slight modification (Decoin et al., 2014).To ensure the selection of attacker and recipient cells for counting, P. fluorescens MFE01 strains were transformed with pSMC2.1 (Kan R ; Davey et al., 2003) and Pca CFBP6276 was transformed with pME6000 (Tet R ; Maurhofer et al., 1998).Bacterial cells were grown overnight on LB agar plates and resuspended in LB broth.The optical density at λ = 580 nm (OD 580 ) was adjusted to 0.5, and cells were mixed in an attacker: recipient ratio of 5:1.About 25-μL drops of the mixture were spotted in triplicate onto 0.22-μm filters deposited onto a prewarmed LB agar plate.After 4 h of incubation at 28 C, bacteria on the filter were resuspended in 1 mL of sterile physiological water and serial dilutions were plated on LB agar supplemented with antibiotic.Colonies were counted after overnight incubation at 28 C.

Biocontrol assays on potato tubers
Solanum tuberosum cv Allians tubers were prepared for inoculation as previously described (Barbey et al., 2013).Inoculation was performed by injection into the intramedulla (to a depth of 1 cm) of 10 μL of a cell suspension containing 10 7 CFU of Pca CFBP6276 and 10 8 CFU of P. fluorescens MFE01 WT, ΔtssC, ΔtssC complemented with a wild-type allele of tssC (ΔtssC-R; Bouteiller et al., 2020), tae3 or tae3 complemented with a wild type allele of tae3 (tae3-R).As controls, tubers were inoculated with each strain alone.Inoculated tubers were incubated in a Minitron incubator (Infors) with 80% humidity at 25 C. Thirty tubers were used per strain or combination.At 1-, 2-and 7-day postinoculation, 10 tubers for each condition were sectioned across the middle, photographed, analysed by measuring average maceration thickness and cut for bacterial population analysis as previously described (Barbey et al., 2013).Experiments were performed in triplicate.A representative result is shown.

Toxicity assays
For toxicity and rescue assays, E. coli DH5α cells were transformed with pBAD33 or derivatives, or cotransformed with pBAD33tat and pTrc99a or derivatives encoding the tae3 and tai3 genes, respectively.Transformants were selected on LB agar plates supplemented with 1% of glucose, chloramphenicol, or chloramphenicol and ampicillin.Overnight cultures of transformants were grown in the presence of antibiotics and glucose, serially diluted and 10-μL drops were spotted on LB agar plates supplemented with antibiotics and 1% of glucose (repression conditions) or 0.2% of arabinose (toxin induction conditions) and 0.5 mM of IPTG (immunity protein induction conditions).Plates were incubated at 37 C for 16 h.

Cell fractionation and membrane differential solubilization
About 10 10 exponentially growing cells producing Tai3 Pf from plasmid pTrc-ST-Tai3 Pf were resuspended in 0.5 mL of Tris-HCl 10 mM (pH 8.0), sucrose 30% and incubated for 10 min on ice.After addition of 100 μg/mL of lysozyme and 1 mM of EDTA and further incubation for 45 min on ice, 0.5 mL of Tris-HCl 10 mM (pH 8.0) supplemented with DNase (200 μg/mL) and MgCl 2 (4 mM) were added and cells were lysed by five cycles of freeze and thaw.Unbroken cells were removed by centrifugation, and soluble and membrane fractions were separated by ultracentrifugation for 40 min at 100,000g.For membrane differential solubilization, the membrane pellet was resuspended in 1 mL of Tris-HCl 10 mM (pH 8.0), EDTA 1 mM supplemented with 1% of sodium N-lauroyl sarcosinate (SLS; Sigma-Aldrich) and incubated on a wheel for 1 h at room temperature.SLS is an anionic detergent that selectively disrupts the inner membrane and solubilises inner membrane proteins (Filip et al., 1973).The insoluble (outer membrane) and soluble (inner membrane) fractions were collected by ultracentrifugation at 100,000g for 40 min.All samples (Total, Soluble, Total membranes, SLS insoluble and SLS soluble) were resuspended in loading buffer and analysed by SDS-PAGE and immunoblotting using anti-StrepTag (Classic, Bio-Rad).Anti-IscS, anti-TolA and anti-OmpA antibodies (laboratory collection) were used as cytoplasmic, inner membrane and outer membrane markers, respectively.

Substituted cysteine accessibility method
Topology assays using cysteine accessibility experiments were carried out as described (Aschtgen et al., 2010;Bogdanov, 2017) with modifications.Positions of the cysteine substitutions were based on (i) DeepTMHMM prediction of potential transmembrane segments (https://www.biorxiv.org/content/10.1101/2022.04.08.487609v1) and (ii) serine residues (to limit impact of substitutions on Tai3 Pf function).2 Â 10 10 cells producing the cysteine variant were harvested, resuspended in buffer A (Hepes 100 mM pH 7.5, sucrose 250 mM, MgCl 2 25 mM and KCl 0.1 mM) to a final OD 600 of 12. 3-(N-maleimidyl-propionyl) biocytin (MPB; Molecular Probes) was added to a final concentration of 100 μM (from a 20 mM stock freshly dissolved in DMSO) and the cells were incubated for 30 min at 25 C. β-Mercaptoethanol (20 mM final concentration) was added to quench the biotinylation reaction, and the cells were washed and resuspended in buffer A supplemented with 5 mM N-ethyl maleimide (Sigma-Aldrich) to block all free sulfhydryl residues.After incubation for 20 min at 25 C, cells were disrupted by sonication.Membranes recovered by ultracentrifugation for 40 min at 100,000g were resuspended in 1 mL of buffer B (Tris-HCl 10 mM pH 8.0, NaCl 100 mM and Triton X-100 1% (v/v)) supplemented with protease inhibitor cocktail (Complete, Roche).After incubation on a wheel for 1 h, insoluble material was discarded by centrifugation 15 min at 20,000g, and solubilized proteins were subjected to precipitation with 100 μL of Strep-Tactin Superflow resin (IBA Technology) equilibrated in buffer B. After 1 h of incubation on a wheel, and five washes with 1 mL of buffer B, Strep-tagged Tai3 Pf was eluted with 100 μL of buffer B supplemented with 2.5 mM of desthiobiotin, boiled in Laemmli buffer before SDS-PAGE analysis and immuno-detection with anti-StrepTag antibodies (to detect the proteins) or streptavidin (to detect the biotinylated proteins) coupled to alkaline phosphatase.Control experiments were performed on purified membranes instead of whole cells.

Fluorescence microscopy and data analysis
Bacterial cells grown overnight in LB media were harvested and resuspended in LB to an OD 600 of 10. Cell resuspension or mixtures were then spotted on a thin 2% agarose pad containing LB, covered with a coverslip and incubated for 2-5 min at 28 C before microscopy acquisition.Fluorescence microscopy analysis was performed with a Nikon Eclipse Ti2 microscope equipped with an OrcaFusion digital camera (Hamamatsu) and a perfect focus system to automatically maintain focus to keep the point of interest within the specimen in sharp focus at all times despite mechanical or thermal perturbations.The microscope temperature was maintained at 30 C in a thermoregulated incubator.All fluorescence images were acquired with a minimal exposure time to minimize bleaching and phototoxicity effects.Exposure times were typically 200 ms for phase contrast and 20 ms for fluorescent fusions.The images shown in the figures are representative regions cropped from large fields and are from at least triplicate experiments.Images were analysed using ImageJ (C. A. Schneider et al., 2012).For analysis of Pca cell lysis, Pca and P. fluorescens MFE01 cells were mixed in a 1:1 ratio, disposed on an agar pad and followed using time-lapse fluorescence microscopy over a 4-h period.Pca cells that remained intact and the cells that lysed (including those lysing after rounding) were then numbered from five different fields (40-100 Pca cells per field), and the experiments were performed four times.

Statistical analyses
The Mann and Whitney test was used to assess differences in maceration symptoms between groups (pvalue < 0.05).The one-way ANOVA test was used to assess differences in prey recovery cells between groups ( p-value < 0.05).

A functional T6SS is required for MFE01 to protect potato tubers against Pca
Previous data have shown that the P. fluorescens MFE01 hcp2 gene, which encodes the major Hcp protein secreted in the medium, is important to confer protection of potato tubers against the phytopathogen Pca (Decoin et al., 2014).Due to the wealth of hcp genes in the P. fluorescens MFE01 genome (Figure S1) and their potential redundancy, we tested the role of the T6SS in potato tuber protection by carrying out in planta assays using a mutant of a gene essential for T6SS function, tssC.For this, bacterial suspensions containing 2 Â 10 7 Pca CFU per gram of potato tuber were inoculated into a wounded site, and the development of soft-rot symptoms was evaluated over a 7-day period (Figure 1A,B).When inoculated alone, Pca caused maceration of potato tubers, with damages appearing as soon as 2 days after inoculation.Maceration intensity increased with incubation time.When Pca was mixed with wild-type (WT) P. fluorescens MFE01, only maceration damages comparable to the wound in the absence of inoculum were observed, demonstrating that MFE01 protects the tuber against Pca.This protection is dependent on a functional T6SS, as co-inoculation of Pca with MFE01 ΔtssC cells did not prevent potato tuber maceration, with damages similar to inoculation with Pca alone.To measure the impact of MFE01 on Pca population, we counted surviving Pca CFU on selective plates (Figure 1C).When inoculated alone or mixed with MFE01 ΔtssC mutant cells, the Pca population reached $10 10 CFU/g of potato tuber after 2 days and then stabilized.In contrast, when mixed with WT MFE01, the Pca population reached 10 9 CFU/g of potato tuber.
These results suggest that the MFE01 T6SS partly inhibits the growth of Pca in planta.This conclusion is supported by in vitro experiments demonstrating that Pca survival is severely affected by MFE01 in a T6SSdependent manner (Figure 1D).Taken together, these data show that MFE01 deploys its T6SS to eliminate Pca cells, or to affect their growth, hence limiting Pca density in the potato tuber and the development of the disease.

MFE01 T6SS activity causes Pca and E. coli cell rounding and lysis
To visualize T6SS sheath activity, we constructed a translational fusion between a component of the T6SS sheath, TssB and the superfolder green fluorescent protein (sfGFP).This fusion was engineered on the MFE01 chromosome, at the native locus.We then monitored T6SS sheath dynamics by time-lapse fluorescence microscopy (Figure 2A, Video S1).MFE01 cells assemble 5 sheaths in average and up to 12 sheaths per cell (Figure 2B) with dynamics similar to other species such as Vibrio cholerae, Serratia marcescens or enteroaggregative E. coli (EAEC) (Basler et al., 2012;Brunet et al., 2013;Gerc et al., 2015;Santin et al., 2018Santin et al., , 2019;;J. P. Schneider et al., 2019): assembly in $60 s (Figure 2C) and a residence time of $3 min (Figure 2D).To determine whether MFE01 T6SS presents a defensive (i.e., respond to an attack) or aggressive (i.e., fire against any cell in contact) behaviour, we then compared T6SS activity in single cells and cells established in microcolonies.The recordings and statistical analyses showed that there are no significant differences in the number of sheath assemblies (Figure 2E), contraction events (Figure 2F) or residence time (Figure 2G) when cells are isolated or in contact with other cells.These observations suggest that P. fluorescens MFE01 can be considered as an aggressive strain, with its T6SS categorized as an offensive weapon.
F I G U R E 1 MFE01 prevents potato tuber maceration by deploying its T6SS against Pca.Biocontrol activity of Pseudomonas fluorescens MFE01 wild-type (WT) (MFE01), P. fluorescens ΔtssC (ΔtssC) and ΔtssC producing TssC (ΔtssC-R) were monitored against Pca (Pca) for a 7-day period.Visual damages, Pca population and maceration thickness were assessed on Day 1 (D1), Day 2 (D2) and Day 7 (D7).(A) Potato tuber protection assay on a 7-day period was realized with different inoculation conditions: control (no inoculum, 0.9% NaCl solution), Pca alone (lane 1), Pca versus MFE01 (lane 2), Pca versus MFE01 ΔtssC (lane 3) and Pca versus MFE01 ΔtssC producing TssC (ΔtssC-R, lane 4).To gain information on the impact of MFE01 T6SS activity on target cells, MFE01 TssB-sfGFP cells were mixed with Pca cells in a 1:1 ratio, and analysed by time-lapse fluorescence microscopy.Figure 3A shows that Pca cells in contact with WT MFE01 change morphology, from a rod to a spherical shape (see white and orange arrows), before losing cytoplasmic density (Figure 3A, Video S2).Quantitative analyses further showed that about 35% of Pca cells lysed when in contact with MFE01 over a 4-h period, with $80% of these Pca cells experiencing cell rounding before lysis (Figure 3B).While cell rounding was the most frequent consequence, we also observed blebbing, plasmolysis and burst events (Figure S2).The morphological modification and lysis of Pca cells were not observed when co-cultured with MFE01 ΔtssC cells (Figure 3A,B, Video S3).Interestingly, target cell rounding appears after sheath contraction suggesting that Pca damages might be correlated with a T6SS firing event (Figure 3C, Video S4).Similar observations were made when MFE01 was co-cultured with E. coli K12 (Figure 3D).As we observed that Pca and E. coli rounding is the major phenotype associated with P. fluorescens MFE01 T6SS antibacterial activity and that cell rounding is usually associated with cell wall damages, these data suggest that the MFE01 T6SS delivers one or several effectors targeting the peptidoglycan.
To better visualize the cellular consequences of MFE01 on target cells, we performed fluorescence microscopy experiments using a strain of E. coli harbouring plasmid pTHV037dsbA-SP-msfTq2O_sol-Cytomcherry as recipient.This strain produces mCherry in the cytoplasm and sfTurquoise in the periplasmic space, allowing to differentiate the two compartments (Deghelt et al., 2023).Time-lapse fluorescence microscopy recordings showed that upon MFE01 T6SS firing, the periplasm of the attacked cell collapsed at one pole of the cell, eventually reached the other pole, before cell rounding (Figure 3E, Videos S5 and S6).Taken together, these images suggest that the peptidoglycan of the recipient cell is severely damaged after an MFE01 T6SS attack.

MFE01 delivers a putative amidase of the Tae3 family that is active against Pca and E. coli
Based on the previous observations, we searched for a potential effector targeting the peptidoglycan.T6SS effectors have been found to be additional domains associated with the Hcp, VgrG or PAAR proteins, or being independent proteins encoded at vicinity of hcp, vgrG or paar genes (Hernandez et al., 2020;Jurenas & Journet, 2021).We therefore conducted a bioinformatic survey to identify hcp, vgrG and paar genes and analysed their sequences and their genomic neighbourhoods (Figure S1).In addition to the main T6SS gene cluster containing a copy of the vgrG gene, we found one hcp, five vgrG and three hcp-vgrG islands (Figure S1).These islands encode putative effectors with various functions, including DNases, phospholipases, peptidases, PoNE-like phosphodiesterases and Rhs proteins (Figure S1).Interestingly, one of the hcp-vgrG islands encodes a putative amidase effector of the Tae3 family (called hereafter Tae3 Pf ) (Figure 4A).This gene is located downstream hcp3 and is followed by a gene of unknown function and a potential additional effector-immunity pair (Figure 4A).
To test whether Tae3 Pf is involved in MFE01 antibacterial activity, we engineered an in-frame deletion of the tae3 gene on the MFE01 chromosome and tested the ability of MFE01 Δtae3 cells to cause Pca and E. coli cell rounding and lysis.Time-lapse fluorescence microscopy recordings showed that MFE01 Δtae3 cells assemble dynamic sheath structures at levels comparable to the wild-type strain (Figure S3A,B), indicating that the T6SS remains active in the absence of Tae3 Pf and that Tae3 Pf is not required for functional assembly of the T6SS. Figure 4B shows that the tae3 mutation decreases MFE01 antibacterial activity against Pca in vitro but maintains a significant level of activity, suggesting that other potential effectors contribute to Pca killing.It is worth noting that the (over)production of Tae3 in the Δtae3 mutant strain significantly increases T6SS activity compared to the wild-type strain (Figure 4B).One may hypothesize that more Tae3 are loaded and delivered into Pca cells, increasing the potential of MFE01.These data were confirmed by time-lapse microscopy recordings of Pca or E. coli incubated with MFE01 Δtae3 cells.While some Pca and E. coli cell lysis events could be still observed in the absen (Figure 4C,D, Videos S7 and S8), quantitative analyses revealed a significant reduction in antibacterial activity with 8%-15% of Pca and E. coli lysed cells over a 4-h period (Figure 4E; to compare with $35% of lysis with WT MFE01; see Figure 3B).This decrease of T6SS activity was also accompanied by a reduction in the number of rounding events in both Pca and E. coli (Figure 4E).Potato tuber assays further showed that Pca co-inoculation with tae3 mutant cells do not confer tuber protection against Pca, although the maceration intensity was significantly decreased compared to Pca alone (Figure 4F,G).As shown in the in vitro antibacterial assay, (over)production of Tae3 significantly increases potato tuber protection compared to the WT strain (Figure 4F,G).Taken together, these results confirm that the Tae3 putative amidase is an important T6SS effector in MFE01 and that the remaining antibacterial activity could be attributed to other potential effectors secreted through the T6SS.
Tae3 Pf is toxic in the periplasm of E. coli and its toxicity is counteracted by the Tai3 Pf inner membrane immunity protein To test whether the putative Tae3 Pf amidase has toxic activity, the P. fluorescens MFE01 tae3 gene was cloned into the pBAD33 vector, fused, or not, to the TorA Tat signal sequence for its export into the periplasm.Figure 5A shows that the Tae3 Pf production in the cytoplasm has no impact on E. coli growth, while the production of Tae3 Pf in the periplasm of E. coli is toxic in the presence of arabinose.Substitution of the cysteine and histidine residues of the putative catalytic site of Tae3 Pf inactivated the toxin (Figure 5A), although these variants were produced at levels comparable to WT Tae3 Pf (Figure 5B).Genes encoding T6SS toxins are usually genetically linked to genes that confer immunity against the activity of the toxin (Durand et al., 2014;Jurenas & Journet, 2021;Russell et al., 2014).Indeed, co-expression of the gene located downstream of tae3, called hereafter tai3 Pf , protects E. coli from the action of Tae3 Pf (Figure 5C).Subcellular localization of the Tai3 Pf immunity protein showed that it cofractionates with membrane proteins (Figure 5D).Further differential membrane solubilization with sodium lauroyl-sarcosinate, a detergent that specifically solubilize inner membrane proteins (Filip et al., 1973), demonstrated that Tai3 Pf associates with the inner membrane (Figure 5D).Bioinformatic analyses suggest that Tai3 Pf comprises a single transmembrane segment between residues 29 and 51.To test this topological model, we performed the substituted cysteine accessibility method, which consists to probe cysteine thiol groups with N-(3-maleimidopropionyl) biocytin (MPB), a compound that readily crosses the outer membrane but unable to reach the cytoplasm (Bogdanov, 2017).We substituted serine residues located at positions 24 and 54 with cysteines.These Tai3 Pf variants were functional as they conferred protection against Tae3 Pf (Figure S4A), were produced at similar levels (Figure S4B), and are accessible (Figure S4C). Figure 5E shows that the S24C variant is not labelled while the S54C variant is labelled with MPB, confirming the presence of a transmembrane segment between these two positions, in an in-to-out topology (Figure 5F).
In this study, we showed that P. fluorescens MFE01 protects potato tubers against Pca and that this protection is promoted by its T6SS.This result confirms a previously published work defining one of the MFE01 hcp genes, hcp2, as an important determinant for potato protection (Decoin et al., 2014).A protective role of the T6SS was also shown in other PGPR belonging to the Pseudomonas genus.For example, P. protegens uses T6SS to exert antagonist activity against insect pathogens (Vacheron et al., 2019), P. fluorescens F113 and Pf29Arp T6SSs increase persistence in the rhizosphere and present antagonistic activity against fungi, respectively; while P. putida exerts a T6SS-dependant biocontrol activity against Xanthomonas campestris which is responsible for Nicotinia benthamiana leaf necrosis (Bernal et al., 2017;Dur an et al., 2021;Marchi et al., 2013).
It is interesting to note that our quantification analyses showed that Pca density in the potato tuber in the presence of MFE01 is only a log below the pathogen population when alone or in the presence of MFE01 ΔtssC.However, even such a limited difference has a significant impact on potato tuber maceration.As previously described with another biocontrol strain, Rhodococcus erythropolis, Pca has to reach a certain density, i.e. the quorum, to initiate a strong and synchronized PCWDE production leading to tissues maceration (Chane et al., 2019).For Pca, this density was estimated above 5.10 9 CFU/g of potato tuber (Chane et al., 2019).One may hypothesize that MFE01 deploys its T6SS to prevent Pca from reaching this density, and hence to prevent massive PCWDE production.
These in planta results were corroborated with data showing that MFE01 T6SS limits Pca and E. coli growth in co-cultures.However, it should be noted that the MFE01 T6SS-dependent impact on Pca differs in the potato tubers (elimination of half the population) compared to the in vitro activity (3-4 logs of antibacterial activity); this could be attributable to the presence of caches and crevices in the tuber parenchyma, which limits cell-to-cell contacts between the two strains.
Further time-lapse fluorescence microscopy recordings showed that MFE01 cells assemble five T6SS structures per cell on average and that, once assembled, these sheath structures remain extended for several minutes.As previously shown for V. cholerae, S. marcescens or EAEC (Basler et al., 2012;Brunet et al., 2013;Gerc et al., 2015;Ostrowski et al., 2018;Santin et al., 2018), MFE01 T6SS activity presents an aggressive behaviour.In co-culture, MFE01 T6SS activity leads to Pca or E. coli lysis, mainly through rounding of the target cell.Other cell damage consequences were also observed, such as blebbling, plasmolysis or burst.Similar lysis phenotypes were shown to be associated with V. cholerae T6SS activity (Basler et al., 2012).Analyses of the time-lapse images showed that target cell rounding correlated with T6SS contraction events, suggesting that T6SS translocates effector(s) targeting the cell wall.Indeed, peptidoglycan, which is one of the major structural components of the cell envelope that confers bacterial shape, is a common target for antibacterial strategies (Russell et al., 2014).T6SS effectors that target the peptidoglycan belong to enzymes of the amidase (Tae; hydrolysing peptide cross bridges) and glycoside hydrolase (Tge, hydrolysing the glycan backbone) families (Durand et al., 2014;Hernandez et al., 2020;Jurenas & Journet, 2021;Russell et al., 2014).In each family, Tae and Tge effectors segregate into subfamilies that diverge by the specific bridge they target.Based on the usual genomic environment of T6SS effector genes, we screened the MFE01 genome and identified a putative amidase of the Tae3 family, encoded downstream of an hcp gene in an hcp-vgrG island, and located upstream of a gene of unknown function.Peptidoglycan-targeting enzymes have been shown to be common T6SS effectors in pseudomonads, including the opportunistic pathogen P. aeruginosa (Chou et al., 2012;Radkov et al., 2022;Russell et al., 2011Russell et al., , 2012;;T. Wang et al., 2020;Whitney et al., 2013).While MFE01 T6SS antibacterial activity and T6SS-mediated potato protection are not abolished in a strain lacking Tae3, our results showed that Tae3 Pf contributes significantly to Pca and E. coli rounding and lysis.Tae3 Pf is indeed toxic when produced in the periplasm of E. coli and its toxicity is counteracted when the gene located downstream tae3 is co-expressed.This gene, which likely encodes a specific immunity to Tae3 Pf , was named tai3.Tai3 Pf is an inner membrane protein with a single transmembrane pass with in-to-out topology, and the majority of the protein locates in the periplasm.One may hypothesize that this periplasmic domain is responsible for neutralizing Tae3 Pf activity, likely by protein-protein interaction.
Due to the large number of highly dynamic T6SSs assembled per cell and their aggressive behaviour, MFE01 is likely to colonize very efficiently plants and soils by eliminating microbial rivals.Indeed, our results provide evidence for the important role of the T6SS in the biocontrol properties of P. fluorescens MFE01, which can be considered/envisioned as an alternative strategy to defeat phytopathogens and to protect plants.Interestingly, P. fluorescens MFE01 carries a single T6SS gene cluster and nine hcp, vgrG or hcp-vgrG islands, which encode >15 potential effector-immunity pairs.Characterization of the MFE01 effector repertoire will likely reveal the variety of activities and potential targets, and probably the broad potential of MFE01 for biocontrol.

F
I G U R E 2 MFE01 assembles highly dynamic T6SS with aggressive behaviour.(A) Representative fluorescence microscopy field of MFE01 TssB-sfGFP (scale bar, 1 μm) (left panel) and time-lapse fluorescence recordings of a single cell (upper panels) and microcolony (lower panels) (time in min).White arrowheads indicate T6SS sheath assembly events.Red and yellow arrowheads indicate contraction events.The corresponding video is provided in the Supporting Information (Video S1).(B-D) T6SS sheath dynamics.Box plot representation of the number of T6SS sheath per cell (B), of the assembly time (time necessary to assemble an extended sheath, C) and of the residence time (time for which the sheath remains extended before contraction, D).The internal horizontal line and cross represent the median and mean values, respectively; the boundaries of the internal box plot correspond to the 25th and 75th percentiles; the whiskers correspond to the 10th and 90th percentiles.Outliers are shown as closed circles.The number of analysed events (n) is indicated on the top of each graph.(E-G) Comparison of T6SS sheath dynamics in single cells and microcolonies.Violin plot representations representing the number of sheath per cell (E), the number of contraction events per cell over a 20-min period (F), and the residence time (G) of MFE01 TssB-sfGFP single cells (dark blue) or microcolonies (light blue).The distribution is represented by the outer shape.The broken line represents the median value.The dotted lines of the internal violin plot correspond to the 25th and 75th percentiles, respectively.The number of analysed events (n) is indicated on the top of each graph.F I G U R E 3 Microscopic analysis of MFE01 T6SS antibacterial activity.(A) Time-lapse fluorescence microscopy recordings of Pseudomonas fluorescens MFE01 wild-type (WT) or ΔtssC producing TssB-sfGFP against Pca (time in min).White and orange arrowheads indicate events of Pca cell damage (scale bar, 1 μm).Corresponding videos are provided in the Supporting Information (Videos S2 and S3).(B) Histograms reporting antibacterial activity of MFE01 WT and MFE01 ΔtssC against Pca after a 4-h incubation (green, living cells; orange, lysed cells; hatched orange, lysis following rounding).Only bacteria in contact with MFE01 or ΔtssC were counted.The number of analysed cells (n) is indicated on the top of each graph.(C) Time-lapse fluorescence microscopy recordings of P. fluorescens MFE01 TssB-sfGFP in contact with Pca (time in min) (scale bar, 1 μm).White arrowheads point an event of T6SS sheath assembly and contraction.The blue arrowhead highlights Pca cell rounding.The corresponding video is provided in the Supporting Information (Video S4).(D) Histograms reporting antibacterial activity of WT MFE01 against Escherichia coli after a 4-h incubation (green, living cells; orange, lysed cells; hatched orange, lysis following rounding).Only E. coli in contact with MFE01 were counted.The number of analysed cells (n) is indicated on the top of each graph.(E) Time-lapse fluorescence microscopy recordings of P. fluorescens MFE01 TssB-sfGFP (green) against E. coli producing mTurquoise in the periplasm and mCherry in the cytoplasm (time in min) (scale bar, 1 μm).The orange and blue arrows point to periplasm collapse at the cell pole and cell rounding, respectively.Corresponding videos are provided in the Supporting Information (Videos S5 and S6).

F
I G U R E 5 Tae3 Pf toxicity in Escherichia coli is counteracted by the Tai3 Pf inner membrane immunity protein.(A) Toxicity assay in the heterologous host E. coli.Cultures of E. coli cells bearing the pBAD33 plasmid-producing Tae3 Pf (cytoplasm), the empty pBAD33tat plasmid or the pBAD33tat plasmid-producing wild-type (WT) or catalytic mutants (C18A or H141A) of Tae3 Pf (exported to the periplasm) were serially diluted and 10 À1 to 10 À6 dilutions (from left to right) were spotted on LB-agar plates supplemented with 1% of glucose or 0.2% of L-arabinose to repress or induce expression from pBAD33tat vectors, respectively.(B) Western-blot analyses of WT and catalytic mutants of Tae3 Pf .Cell extracts of E. coli cells producing WT or catalytic mutants (C18A or H141A) of Tae3 Pf were analysed by SDS-PAGE, transferred onto nitrocellulose and immunodetected with anti-VSV-G monoclonal antibodies.Molecular weight markers are indicated on the left.(C) Toxicity and rescue assays in the heterologous host E. coli.Cultures of E. coli cells bearing the empty or tae3-encoded pBAD33tat plasmid and the empty or tai3-encoded pTrc99A plasmid were serially diluted and 10 À1 to 10 À6 dilutions (from left to right) were spotted on LB-agar plates supplemented with 0.2% of L-arabinose to induce expression of tae3, or supplemented with 0.2% of L-arabinose and 0.5 mM of IPTG to induce expression of tae3 and tai3, respectively.(D) Cell fractionation and differential solubilization.E. coli cells producing Strep-tagged Tai3 Pf (T) were subjected to cell fractionation to separate the soluble (S) from the membrane (M) fraction.The membrane fraction was then solubilized with SLS to separate inner (IM) and outer (OM) membranes.The different fractions were analysed by SDS-PAGE, transferred onto nitrocellulose and immunodetected, from top to bottom, with anti-TolA, anti-OmpA, anti-IscS and anti-Strep-Tag (Tai3 Pf ) antibodies.Molecular weight markers are indicated on the left.(E) Substituted cysteine accessibility method.E. coli cells producing Strep-tagged wild-type (WT) or S24C or S54C Tai3 Pf were treated with MPB to label accessible thiol groups.After Tai3 Pf precipitation, the samples were analysed by SDS-PAGE, transferred onto nitrocellulose and immunodetected with anti-StrepTag monoclonal antibodies (upper panel) and streptavidin coupled to alkaline phosphatase (lower panel).Molecular weight markers are indicated on the left.(F) Topological model of Tai3 Pf .The inner membrane (IM), which separates the cytoplasm (cyto) from the periplasm (peri), is indicated as well as the boundaries of the Tai3 Pf transmembrane segment (green).
a restriction site or sequence annealing on the target vector underlined b tag sequence in italics c mutagenized codon in bold