https://orcid.org/0000-0001-8784-1491
Figures
Abstract
The necrotrophic plant pathogenic bacterium Dickeya solani emerged in the potato agrosystem in Europe. All isolated strains of D. solani contain several large polyketide synthase/non-ribosomal peptide synthetase (PKS/NRPS) gene clusters. Analogy with genes described in other bacteria suggests that the clusters ooc and zms are involved in the production of secondary metabolites of the oocydin and zeamine families, respectively. A third cluster named sol was recently shown to produce an antifungal molecule. In this study, we constructed mutants impaired in each of the three secondary metabolite clusters sol, ooc, and zms to compare first the phenotype of the D. solani wild-type strain D s0432-1 with its associated mutants. We demonstrated the antimicrobial functions of these three PKS/NRPS clusters against bacteria, yeasts or fungi. The cluster sol, conserved in several other Dickeya species, produces a secondary metabolite inhibiting yeasts. Phenotyping and comparative genomics of different D. solani wild-type isolates revealed that the small regulatory RNA ArcZ plays a major role in the control of the clusters sol and zms. A single-point mutation, conserved in some Dickeya wild-type strains, including the D. solani type strain IPO 2222, impairs the ArcZ function by affecting its processing into an active form.
Author summary
The development of new antibacterial molecules is critical in tackling the emergence of new pathogens or bacterial strains resistant to already available antibiotics. Bacterial phytopathogens can potentially synthesize novel compounds capable of targeting a specific type of microorganism. Dickeya solani is the only of the twelve described Dickeya species that has the three secondary metabolic pathways sol, ooc and zms. An investigation of the functions of these three clusters allowed us to identify the anti-yeast activity of the sol cluster, a potential new molecule of clinical importance. By comparing the antimicrobial activity of several Dickeya solani strains, we identified the small RNA regulator ArcZ as a critical regulator in the activation of the sol and zms clusters. Our study showed that single-nucleotide polymorphisms of sRNA encoding genes can have huge impacts on bacterial phenotypes. It is thus critical to pay attention to the allele diversity of sRNA genes.
Citation: Brual T, Effantin G, Baltenneck J, Attaiech L, Grosbois C, Royer M, et al. (2023) A natural single nucleotide mutation in the small regulatory RNA ArcZ of Dickeya solani switches off the antimicrobial activities against yeast and bacteria. PLoS Genet 19(4): e1010725. https://doi.org/10.1371/journal.pgen.1010725
Editor: Pascale Romby, UPR 9002 CNRS-ARN, FRANCE
Received: November 25, 2022; Accepted: March 30, 2023; Published: April 27, 2023
Copyright: © 2023 Brual et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by grants from Agence Nationale de la Recherche to L.A. (Project RNAchap, ANR-17-CE11-0009-01) and to E.G. (Project Tn-Phyto, ANR-19-CE35-0016). E.G was also supported by the FR BioEEnVis, and by annual credits from the University Lyon I and the CNRS at regular basis. T.B. was supported by a PhD grant from the Ministère de l’Enseignement Supérieur, de la Recherche et de l’Innovation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Bacterial phytopathogens of the genus Dickeya and Pectobacterium are pectinolytic necrotrophic bacteria with a broad host plant spectrum [1,2]. These members of the family Pectobacteriaceae [3] cause substantial agricultural losses worldwide by affecting many vegetables, ornamentals and crops, of which the potato is the most important economically. These bacteria are able to invade and degrade the plant tissues through the coordinated expression of genes encoding virulence factors, with a major role of pectate lyases that dissociate the plant cell wall constituents [4].
The species D. solani was officially established in 2014 [5] but D. solani isolates have attracted attention since their emergence on the potato agrosystem in Europe in the early 2000s. It causes symptoms in both subtropical and temperate climates. Many scientific efforts have been made to provide information on this phytopathogen, resulting in 76 D. solani genomes available in May 2021 [6]. Comparative genomics analysis was performed to identify the genetic basis for the different levels of virulence between D. solani strains [6–11]. Most D. solani strains isolated from different regions show a low level of genetic variation, suggesting a clonal origin [11]. The D. solani genomes share a high similarity and synteny with those of the model strain D. dadantii 3937, prompting a comparison between the two species. Only a few hundred genes were specific to each species, including a few dozen distinctive genomic regions [7,8]. Three of these regions encode polyketide synthases (PKS), non-ribosomal peptide synthetases (NRPS) and amino acid adenylation domain proteins, which are typically involved in the production of secondary metabolites [7].
PKSs and NRPSs are able to synthesize molecules by sequential condensation of acyl-CoAs and amino acids, respectively. PKS and NRPS modules can combine together to form hybrid PKS/NRPS systems capable of producing compounds of great structural diversity [12]. The molecules synthesized may have siderophore, antibiotic or phytotoxic properties that promote the virulence of a plant pathogen. Three PKS/NRPS clusters are present in all sequenced D. solani strains [13] and found in a few other Dickeya species and related genera: ooc, zms, and a cluster of unknown function at the beginning of our study, which was recently named sol [14]. In Serratia plymuthica, the cluster ooc is involved in the synthesis of oocydin A, a halogenated macrolide with antioomycete and antiascomycete activities [15,16]. The cluster zms, previously found in the genomes of S. plymuthica and Dickeya oryzae, leads to the biosynthesis of a polyamino-amide antibiotic, zeamine [17,18]. The cluster sol (for solanimycin) is found in a few Dickeya species. To elucidate the contribution of the sol, ooc and zms clusters of D. solani in competition with other living organisms, we constructed mutants of each of the three loci involved in the biosynthesis of secondary metabolites in the highly virulent D. solani strain D s0432-1. We showed that the D. solani D s0432-1 clusters zms and ooc, encoding zeamine and oocydin biosynthesis, are involved in growth inhibition of bacteria and fungi, respectively. Recently, we, and others identified the third cluster sol as implicated in Dickeya solani’s ability to inhibit the growth of different yeast species. We had previously named this cluster ssm (for solani secondary metabolite) [19], and it was later renamed sol [14]. Finally, the comparative analysis of the genomes and phenotypes of several wild-type D. solani strains revealed the fundamental role of the small RNA ArcZ. This sRNA plays a major role in the regulation of the sol and zms clusters, which can be disrupted by a mutation in the arcZ sequence, conserved in some natural isolates.
Results
Description and distribution of the three selected PKS/NRPS secondary metabolite clusters of D. solani
The gene clusters encoding complex NRPS and PKS involved in the production of secondary metabolites were named sol, ooc and zms, respectively (Fig 1A). Their distribution in Dickeya species is summarized in Fig 1B (see also S1 Table for a complete list of the Dickeya strains with the indication of the presence or absence of each cluster).
(A) Organization of the clusters sol, ooc, and zms. Genes are indicated using the NCBI nomenclature of the NCBI reference genome sequence NZ_CP017453.1 of D. solani D s0432-1. XXXXX is the digital number in BJD21_RSXXXXX, which corresponds to the locus tag. Arrowheads show gene orientations. Color code indicates gene function. The red-framed arrows indicate genes targeted for in-frame deletion performed in this study. (B) MLSA tree positioning strains within the Dickeya genus. The evolutionary history was inferred by using the Maximum Likelihood method. The presence (green frame) or absence (red frame) of the cluster sol, ooc, and zms are indicated in front of each strain. The distribution of these clusters in 155 Dickeya strains whose genome has been sequenced is given in S1 Table.
The ∼42-kbp cluster sol contains the 12 genes solABCDEFGHIJKL (Fig 1A). It is widely conserved in the genus Dickeya, i.e., in all sequenced D. solani, D. aquatica, D. fangzhongdai, D. poaceiphila and D. zeae genomes, in some D. dadantii strains (such as NCPPB 898, NCPPB 3537, DSM18020, but not the model strain 3937), in some D. undicola strains (FVG1, FVG10), and in some D. oryzae strains (such as S20 and FVG03). The structure of the metabolite produced from this cluster has not been elucidated yet but the function of this cluster has been described by us in our 2021 preprint article [19] and others in 2022 [14]. The sol cluster produces solanimycin that targets fungi and yeast of the Ascomycota phylum.
The ∼80-kbp cluster ooc (Fig 1A) is highly similar to the oocBCDEFGJKLMNOPQRSTUVW cluster of S. plymuthica A153. It is present in all the sequenced genomes of D. solani and D. dianthicola, in some D. oryzae strains (such as ZYY5, EC1, DZ2Q and ZJU1202) and in a few other species. In S. plymuthica A153, disruption of this gene cluster abolished bioactivity against the ascomycete Verticillium dahliae and the oomycetes Pythium ultimum. This cluster produces oocydin A [15], a chlorinated macrolide, powerfully active against plant pathogenic oomycetes [20]. Since various D. solani strains inhibit V. dahliae and P. ultimum growth [16], it was suggested, on the basis of gene sequence homologies and similar cluster organization, that D. solani also produces oocydin A.
The 55-kbp cluster zms encodes mixed fatty acid synthase FAS/PKS and hybrid NRPS/PKS enzymes (Fig 1A). Its genomic organization is identical to the D. oryzae EC1 zeamine cluster and related to the S. plymuthica AS12 zeamine cluster [17,21]. While the zmsABCDEFGIJKLMNPQRS cluster directing zeamine biosynthesis is present only in a few D. oryzae strains, it is conserved in all the sequenced D. solani and D. fangzhongdai genomes, which suggest secondary acquisition by horizontal gene transfer in D. oryzae [13,21]. After reclassification of several D. zeae strains in the novel species D. oryzae [22], the zms cluster appeared to be absent in the genomes of true D. zeae strains. It is present in some, but not all, D. oryzae rice strains (ZYY5, EC1, DZ2Q and ZJU1202) and in only one D. dadantii strain (A622-S1-A17). The zeamine biosynthetic clusters from D. oryzae EC1 and D. solani Ds0432-1 share from 59 to 94% identity at individual protein level [21]. Zeamine-related antibiotics are polyamino-amide molecules toxic to a wide range of pro- and eukaryotic organisms such as bacteria, fungi, oomycetes, plants, and nematodes [23]. Mutation of the zeamine synthase gene zmsA in D. oryzae EC1 attenuates the inhibition of rice seed germination [17] and suppresses antibacterial activity against E. coli [17]. Zeamine produced by S. plymuthica kills nematodes and ascomycete yeast [24]. D. solani IPO 2222 can also kill the nematode Caenorhabditis elegans but not as quickly as S. plymuthica [24].
To interrupt the synthesis of the secondary molecules produced by these three clusters, we constructed from a NalR GmR derivative of WT D. solani D s0432-1 in-frame deletion mutants inactivating a PKS or NRPS gene of each cluster, which is predicted to interrupt metabolite synthesis. The inhibitory effects of the mutants ΔsolG, ΔoocL and ΔzmsA against fungi, bacteria and yeasts were compared with that of the parental strain. To further analyze possible synergistic effects, double mutants ΔsolGΔoocL, ΔsolGΔzmsA and ΔoocLΔzmsA as well as the triple mutant (hereafter called Δ3) were also constructed.
The D. solani oocydin cluster ooc inhibits growth of phytopathogenic ascomycete fungi, and the sol cluster is also involved in the inhibition of some fungi
We compared the ability of the WT D. solani strain D s0432-1 and its derived mutants to inhibit the growth of Botrytis cinerea, Magnaporthe oryzae and Sclerotinia sclerotiorum, three fungi eukaryotes of the phylum Ascomycota. The center of a potato dextrose agar (PDA) plate was inoculated with fungal mycelium, and 5 μl of overnight bacterial culture of each D. solani strain grown in M63 minimal medium with sucrose was deposited at the periphery of the plate (Fig 2). The different D. solani strains showed similar overall growth in liquid culture (S1 Fig). After incubation at 25°C for several days, we observed a growth inhibition of the three fungi by the D. solani WT and the ΔsolG and ΔzmsA mutants. In contrast, the ΔoocL mutant seemed less effective: B. cinerea and M. oryzae mycelium growth was completely unaffected, while a small zone of inhibition was observed for S. sclerotiorum (Fig 2). This suggests that the ΔoocL mutant may also produce another molecule that inhibits this fungus. The phenotypes were fully restored in the reverted strain ΔoocL/oocL+. Double and triple mutants of the secondary metabolite clusters containing the oocL deletion showed no differences from the ΔoocL mutant with B. cinerea and M. oryzae. However, the ΔoocLΔsolG double mutant and the Δ3 triple mutant were completely unable to inhibit S. sclerotiorum growth (Fig 2). These data suggest that, in our assay on PDA plate, the ooc cluster is mainly responsible for the majority of ascomycete inhibition and that the sol cluster is secondary involved in the inhibition of some fungi. In contrast, the zms cluster does not seem to be involved in fungal inhibition.
5 μL of bacterial culture at OD600nm = 2 from D. solani Ds0432-1 (WT) or derivatives were spotted onto PDA plates inoculated with plugs of S. sclerotiorum, B. cinerea or M. oryzae mycelium. Plates were incubated at 25°C until the mycelium covers the plate. Lengths of fungi inhibition zone were measured in 4 independent experiments. A statistical difference was significant only between the WT and ΔoocL mutants (Mann-Whitney test; * p-value < 0.05).
The D. solani zeamine cluster inhibits bacterial growth
The zeamine cluster of D. oryzae EC1 is responsible for a bactericidal activity against Escherichia coli DH5α [17]. D. oryzae EC1 and D. solani zeamine biosynthetic genes share a high degree of similarity [21]. We thus evaluated the capacity of the D. solani strain D s0432-1 and its derived mutants to inhibit the growth of Gram-positive and Gram-negative bacteria (Fig 3). Only the ΔzmsA mutant was unable to inhibit Bacillus subtilis growth, indicating that WT D. solani produces an active zeamine antibiotic (Fig 3). The WT phenotype was restored in the reverted ΔzmsA/zmsA+ strain. We also tested the ability of D. solani to inhibit growth of the other Gram-positive bacteria Streptomyces scabiei, a plant pathogen causing the potato disease common scab [25]. An inhibition of S. scabiei growth was observed, except with the ΔzmsA mutant. No phenotype difference was observed between the single ΔzmsA mutant and the double and triple mutants containing the zmsA deletion, suggesting that, out of the three clusters, only zms is required for Gram-positive bacterial growth inhibition. We also tested the ability of D. solani to inhibit growth of the Gram-negative bacteria E. coli, D. dadantii and Pectobacterium atrosepticum (Fig 3). E. coli DH5α was, like B. subtilis and S. scabiei, inhibited by the WT strain through zms activity, albeit much less potently (Figs 3 and S2). No inhibition of the two other pectinolytic bacteria D. dadantii and P. atrosepticum growth was observed. In conclusion, out of the three clusters analyzed, only the zeamine-like molecule potentially produced by the zms cluster has an antibacterial activity against some Gram-positive and Gram-negative bacteria.
Bioassay plates were prepared by mixing bacterial culture (B. subtilis, S. scabiei, E. coli, D. dadantii and P. atrosepticum) with warm LB agar. 5 μL of bacterial cultures at OD600nm = 2 from D. solani Ds0432-1 (WT) or derivatives were spotted onto the plate and incubation was performed at 30°C during 48 h. A slight inhibition zone was observed with B. subtilis and E. coli, except with the Δzms mutants. All experiments were performed in 4 replicates.
The D. solani clusters sol and zms inhibits ascomycete yeast growth
Since zeamine produced by S. plymuthica A153 has previously been shown to be toxic to the ascomycete yeast Saccharomyces cerevisiae [24], we tested the capacity of D. solani D s0432-1 and its derivative mutant to inhibit the growth of the yeasts S. cerevisiae, Kluyveromyces lactis, a predominant eukaryote during cheese productions [26], and Candida albicans, an opportunistic human pathogen. In a growth inhibition assay performed in Yeast Peptone Dextrose (YPD) solid agar medium, we observed a strong inhibition of K. lactis by WT D. solani, a lower inhibition of S. cerevisiae, and a slight inhibition of C. albicans (Fig 4). The ΔsolG mutant was almost completely deficient in its ability to inhibit yeast growth, while the ΔsolG/solG+ revertant strain showed a restored WT phenotype. The cluster sol is thus primarily responsible for the antieukaryotic action against these yeasts on YPD (Fig 4). Nevertheless, a weak halo of inhibition of the yeasts K. lactis and S. cerevisiae was observable around the mutant ΔsolG. This halo disappeared around the double mutant ΔsolGΔzmsA and the triple mutant Δ3, suggesting that the zms cluster is also implicated in yeast inhibition, but at a lower level than that observed for the sol cluster. In conclusion, although the zeamine cluster of D. solani D s0432-1 is involved in the inhibition of yeast growth, as shown for S. plymuthica A153 [24], D. solani D s0432-1 also possesses the cluster sol that produces another antiyeast molecule.
Bioassay plates were prepared by mixing yeast culture (S. cerevisiae, K. lactis, C. albicans) with warm YPD agar. (A) 5 μL of bacterial culture at OD600nm = 2 from D. solani Ds0432-1 (WT) or derivatives were spotted onto the plates and incubation was performed at 30°C during 48 h. (B) The radius of each inhibition zone was measured. A statistical difference was significant only between the WT and the Δsol mutants (Mann-Whitney test; * p-value < 0.05). For C. albicans, the inhibition zone was too small to be quantified. All experiments were performed in 4 replicates.
The WT D. solani strains IPO 2222 and IFB0223 are unable to inhibit yeast and bacterial growth
We showed that D. solani D s0432-1 has the capacity to inhibit the growth of a wide range of microorganisms, including filamentous fungi, bacteria, and yeasts. We therefore tested whether other WT D. solani strains isolated from the environment in Europe, either from lesions of potato tubers or plants, or from potato rhizosphere, possess this ability. We focused our work on D. solani strains available in national strain collections and whose genome sequence is available. These strains are D s0432-1, IFB0099, IFB0158, IFB0223, IFB0484, IPO 3337, IPO 3494, IPO 3793 and the D. solani type strain IPO 2222. We tested their capacity to inhibit the bacterium B. subtilis, the yeast K. lactis and the fungus S. sclerotiorum (Fig 5A). All strains were able to inhibit the growth of S. sclerotiorum. The inhibition of B. subtilis and K. lactis was clearly observed for all the strains except IPO 2222 and IFB0223. These strong phenotypic differences between IPO 2222 and IFB0223 versus the other WT D. solani strains could not be explained by any nucleotide variabilities in the sol, zms, and ooc clusters. All the selected strains carry the sol, zms, and ooc clusters, with 100% conservation at the nucleotide level (including in the promoter regions). Nonetheless, there are certain changes at the genome level, most notably some nucleotide variability with the presence of SNPs (single-nucleotide polymorphism) and InDel (insertion or deletion), outside the clusters sol, zms, and ooc. Hence, we searched for variations in the whole genomes by comparing two groups: strains IFB0223 and IPO 2222 that are defective for the growth inhibition of yeast and bacteria, versus strains D s0432-1, IPO 3337 and IFB0099 that are efficient for these inhibitions (S2 Table).
(A) Inhibition of B. subtilis, K. lactis and S. sclerotiorum by diverse WT D. solani strains. The tests were performed as previously described using the nine strains IFB 0099 (NZ_CP024711), IFB 0158 (NZ_PENA00000000), IFB 0223 (NZ_CP024710), D s0432-1 (NZ_CP017453), IPO 3494 (NZ_CM001842), IFB 0484 (NZ_CM001860), IPO 3337 (NZ_CP016928), IPO 3793 (NZ_CP017454) or the type strain IPO 2222 (NZ_CP015137). Only IFB 0223 and IPO 2222 were unable to inhibit B. subtilis and K. lactis. (B) Organization and conservation of the arcZ locus in the nine D. solani strains. The single mutation G90A in IPO 2222 arcZ is symbolized by a dark spot in the red arcZ arrow. (C) Zoom on the nucleotide sequence around arcZ of D. solani D s0432-1. This allele is named arcZ1. Positions and orientations of the genes elbB, arcZ and arcB are shown. The location of the G90A mutation present in IPO 2222 is indicated. This allele is named arcZ2.
A SNP was observed with an identical nucleotide in the strains D s0432-1, IPO3337 and IFB0099 (a G at position 90 of arcZ) but different nucleotide in the strain IPO2222 (a A at position 90 of arcZ) (position 2530087 according to IPO 2222 genome). This variation is positioned in the 3’ region of the small RNA arcZ, known to be a key area interacting with targeted mRNA for post-translational regulation in other bacterial species (Fig 5C). The ArcZ variants of D s0432-1 and IPO 2222 will be referred to as ArcZ1 and ArcZ2 respectively in the remainder of this article. In the strain IFB0223, we observed a 3-kbp chromosomal deletion covering the gene arcZ, and extending between the genes arcB and osmY (Fig 5B). In other D. solani strains, this region includes arcZ, elbB, mtgA, and a gene encoding an alginate lyase domain (pfam05426). These discoveries led us to hypothesize that the inability of strains IFB0223 and IPO 2222 to inhibit yeast and bacterial growth could be due to an ArcZ deficiency.
Before going further, we ordered the D. solani IPO 2222 type strain stored in the BCCM collection (LMG 25993) to analyze its genotype and phenotype. Using PCR amplification and Sanger sequencing of arcZ, we confirmed that the arcZ allele of LMG 25993 is arcZ2, similarly to the IPO 2222 strain used in our experiments. Our laboratory strain IPO 2222 and LMG 25993 have the same inhibition phenotype, i.e., an inability to inhibit yeast and bacterial growth (S3 Fig).
The arcZ1 allele is required for zms and sol antimicrobial activities
sRNAs are posttranscriptional regulators that most commonly influence gene translation by base-pairing to target mRNAs. In E. coli, this regulation often requires the RNA chaperones Hfq or ProQ [27–29]. ArcZ belongs to what is called the “core sRNAs” because it is highly conserved among enterobacterial species [30]. In order to determine if ArcZ is required for secondary metabolite production, the ΔarcZ mutants of D. solani strains D s0432-1 and IPO 2222 were constructed, and their phenotypes compared to those of the WT strains (Fig 6A). In M63 sucrose, the mutants grew as well as the parental WT strains (S4 Fig). In our inhibition assay against B. subtilis and K. lactis, no inhibition zones were observed around the ΔarcZ1 mutant of D s0432-1 (Fig 6A), demonstrating that ArcZ1 is required for the synthesis of the metabolites produced from the sol and zms clusters in D s0432-1. However, the ΔarcZ1 mutant of D. solani D s0432-1 was still able to inhibit the growth of the fungus S. sclerotiorum, although a slight decrease in the antifungal effect was observed. Thus, ArcZ does not appear to be fully required for ooc cluster expression, which is consistent with the results obtained with D. solani IFB0223, the naturally arcZ deletion mutant capable of inhibiting S. sclerotiorum growth (Figs 5A and 6A). D. solani IPO 2222 WT inhibited neither bacteria nor yeast but inhibited the growth of fungi, and the phenotype of the IPO 2222 ΔarcZ2 mutant was similar to that of the IPO 2222 WT or the D s0432-1 ΔarcZ1 mutant. This suggests that ArcZ2 might be inactive.
(A) Yeast bioassay plates were prepared by mixing culture of K. lactis with warm YPD agar. Bacteria bioassay plates were prepared by mixing culture of B. subtilis with warm LB agar. PDA plates were inoculated with plugs of S. sclerotiorum. For each experiment, 5 μL of bacterial culture at OD600nm = 2 from the different D. solani strains were spotted onto the plate. Inhibition zones observed with D s0432-1, IPO 2222, their respective ΔarcZ mutants and IFB0223. ArcZ1 is required for the synthesis of metabolites produced by the clusters sol and zms clusters (inhibition of K. lactis and B. subtilis, respectively) but not for the metabolite produced by the cluster ooc (inhibition of S. sclerotiorum). (B) Expression of sol::luxCDABE, ooc::luxCDABE and zms::luxCDABE fusions in the different D. solani strains. Cultures of the different strains with the plasmid pSEVA421 carrying a transcriptional fusion of each promoter of the 3 clusters into a 96-well plate. The luminescence values shown here were measured when all cultures reached an OD600nm of 0.3. A, B, C, D represent groups with significantly different means (MannWhitney test, p-value < 0.05). The expression of the 3 clusters is higher in D s0432-1, especially for the sol and zms clusters. (C) Complementation assays and heterologous expression of the alleles arcZ1 or arcZ2 in the three D. solani strains. Yeast bioassay plates were prepared as in (A). ArcZ1 activates the anti-microbial activities of sol and zms clusters in IPO 2222 and IFB0223. All experiments were performed in 4 replicates.
The sol and zms clusters’ expression is promoted by the arcZ1 allele, but not arcZ2
To further demonstrate that ArcZ1 controls the sol and zms clusters, we constructed transcriptional fusions between the promoter region of the sol, zms, and ooc clusters, and the luxCDABE operon. We introduced sol::luxCDABE, zms::luxCDABE and ooc::luxCDABE into WT D s0431-1, WT IPO 2222, WT IFB0223, and the mutants ΔarcZ1 D s0432-1 and ΔarcZ2 IPO 2222. Luminescence of the strains were measured (Fig 6B). We detected a strong transcriptional activity of the sol cluster in WT D s0432-1, while its activity was 229-fold lower in the ΔarcZ1 background. WT IPO 2222, its ΔarcZ2 mutant, and IFB0223 all exhibited low sol transcriptional activity similar to the ΔarcZ1 D s0432-1 mutant. Regarding the transcriptional activity of the cluster zms, WT D s0432-1 had a significant ∼3-fold stronger transcriptional activity than the ΔarcZ1 D s0432-1 mutant, WT IPO 2222, and its ΔarcZ2 mutant. Surprisingly, the ΔarcZ1 D s0432-1 mutant, WT IPO 2222, and its ΔarcZ2 mutant had a stronger zms transcriptional activity compared to IFB0223, suggesting the existence of other genomic differences between IPO 2222 and IFB0223 that could modulate zms transcriptional activity. Finally, a slight 1.8-fold decrease of ooc transcriptional activity was measured between WT D s0432-1 and its ΔarcZ1 D s0432-1 mutant. These findings indicate that ArcZ has a modest effect on ooc transcriptional activity while promoting the expression of zms and sol clusters.
We next looked for predictions of ArcZ1 interactions with mRNA targets in the -200/+100 region around the start codon (+1) of each D. solani D s0432-1 gene using the CopraRNA algorithm [31] (S3 Table). No putative interactions were predicted around the start codons of the first genes of each cluster (solA, zmsA and oocJ). This result, together with transcriptional fusion analyses, suggests that ArcZ may not interact directly with the 5’ ends of the sol and zms mRNAs but rather indirectly by regulating the expression of a transcriptional factor that regulates these clusters at the transcriptional level.
The G90A mutation in arcZ2 of D. solani IPO 2222 causes a loss of function of zms and sol clusters
To verify the effect of the G90A mutation in arcZ, we performed a heterologous complementation experiment by transferring the allele arcZ1 from D s0432-1 or arcZ2 from IPO 2222 in D. solani WT strains D s0432-1 or IPO 2222 or their ΔarcZ derivatives. The alleles arcZ1 and arcZ2 were also transferred into D. solani IFB0223 naturally deleted for arcZ.
The two allelic forms arcZ1 and arcZ2 were cloned with their own promoter in the plasmid pWSK29-oriT, a mobilizable low copy plasmid with pSC101 origin and ampicillin resistance gene. DNA sequencing confirmed that the only difference between the two plasmids is the G90A mutation in pWSK29-oriT-arcZ2. The recombinant plasmids and the empty vector were transferred by conjugation into WT or ΔarcZ D. solani D s0432-1 or IPO 2222 strains, as well as WT D. solani IFB0223 (naturally deleted for arcZ). The phenotypes of the conjugants were determined by testing their antimicrobial activity against B. subtilis, K. lactis and S. sclerotiorum (Fig 6C).
The transfer of arcZ1 in ΔarcZ1 Ds0432-1 restored the WT phenotypes, proving that the allele arcZ1 activates the antimicrobial activities of the sol and zms clusters, while the antifungal activity was unaffected. In contrast, the transfer of arcZ2 in the ΔarcZ1 DS0432-1 strain did not restore the WT phenotypes. Thus, the difference in complementation between arcZ1 and arcZ2 is due to the presence of the G90A mutation that impairs ArcZ2 function.
An identical result was obtained with the ΔarcZ2 mutant of IPO 2222 or the IFB0223 strain carrying the pWSK29-oriT, pWSK29-oriT-arcZ1 and pWSK29-oriT-arcZ2. Indeed, the transfer of arcZ1 in strains WT IPO 2222, ΔarcZ2 IPO 2222 or IFB0223 activates the antimicrobial activities of the sol and zms clusters. In contrast, the transfer of arcZ2 in these strains did not. These data demonstrate that the presence of arcZ and the nature of the arcZ allele are the main factors governing the difference in phenotypes for antimicrobial activities between the D. solani strains D s0432-1 and IPO 2222 or IFB0223.
Finally, we questioned if an arcZ mutation could be found in additional D. solani natural isolates. We extended this genomic analysis to 57 D. solani genomes, including IPO 2222. It revealed that only four strains have an arcZ single-nucleotide mutation, all at different positions in the 3’ region of arcZ. It is not known whether these other mutations affect the antimicrobial activity of these strains. The arcZ DNA sequence of D. solani IPO 2222 was also used as a query in a BlastN search against the NCBI nucleotide database. Three D. fangzhongdai strains DSM 101947 (CP025003.1), QZH3 (CP031507.1), and LN1 (CP031505.1) had the G90A mutation in arcZ. A G90T mutation was also discovered in D. parazeae Ech586 (CP001836.1) (S5 Fig). Therefore, the G90 mutation to A or T of arcZ can be found in other Dickeya species.
The G90A mutation in arcZ2 of D. solani IPO 2222 alters ArcZ2 processing
To gain further insight into the consequence of the G90A mutation in arcZ, Northern-blot experiments were performed to detect the sRNAs ArcZ1 in WT D s0432-1 and ArcZ2 in WT IPO 2222 (Fig 7). A probe specific to the 3’ part of ArcZ1 and ArcZ2 was used (annealing from nucleotide 91 to 132). An abundant smaller transcript (~60nt) representing the processed 3′ fragment of ArcZ1 was detected in WT D s0432-1 but not in the mutant D s0432-1 ΔarcZ1 (Fig 7, compared lanes 1 and 2). The full-length form corresponding to the ArcZ1 precursor was barely detected. This confirms that the processed form of ArcZ1 is predominant, like observed in E. coli, Salmonella, and Photorhabdus [30,32,33]. In IPO 2222, only the full-length ArcZ2 was detected in the WT strain whereas no ArcZ2 transcript was detected in the IPO 2222 ΔarcZ2 mutant (Fig 7, compared lanes 5 and 6). This result suggests that the G90A mutation present in the arcZ2 allele of IPO 2222 prevents ArcZ2 processing in a shorter functional transcript.
(A) Detection of the sRNA ArcZ1 and ArcZ2 by Northern Blot in D. solani strains. A small transcript (~60 nt) corresponding to the processed 3′ fragment of ArcZ1 was detected in D s0432-1 and in all strains expressing arcZ1 on plasmid. In IPO 2222, only the full-length form (~132nt) ArcZ2 was detected, as well as in strains expressing the allele arcZ2 on plasmid. The 5S RNA was used as a loading control. This experiment was performed twice with biological replicates and gave identical results. (B) Prediction of the secondary structure of the precursor forms of ArcZ1 and ArcZ2 by RNA fold. Bases are colored according to base-pairing probabilities (see color scale). The G90A mutation present in ArcZ2 seems to modify the structure of the ArcZ sRNA precursor.
Next, we realized Northern-blot with the D. solani ΔarcZ strains expressing either the arcZ1 or arcZ2 allele from the plasmid pWSK29-oriT-arcZ. On one hand, Northern blot showed that D s0432-1ΔarcZ1, IPO 2222 ΔarcZ2 and IFB0223 (natural ΔarcZ) expressing arcZ1 from D s0432-1 accumulate the processed short transcript of ArcZ1 (Fig 7, lanes 3, 7 and 10) and have the ability to inhibit yeast and bacteria (Fig 6). On the other hand, D s0432-1ΔarcZ1, IPO 2222 ΔarcZ2 and IFB0223 expressing arcZ2 from IPO 2222 showed the accumulation of the non-processed transcript of ArcZ2 (Fig 7, lanes 3, 7 and 10) and an inability to inhibit yeast and bacteria (Fig 6). These results support the hypothesis that the G90A mutation in ArcZ alters the processing of the ArcZ precursor in a shorter form, leading to its inability to play a correct regulatory function on the sol and zms clusters.
Discussion
PKS/NRPS secondary metabolite pathways are a great source of molecules with antieukaryotic or antibacterial activity, giving the bacteria that synthesize them a competitive advantage over other organisms. Targeted mutagenesis of the D. solani D s0432-1 chromosome has allowed us to specifically study the involvement of three secondary metabolite biosynthesis clusters encoded by all the D. solani strains for which genome sequence is available. We focused our work on strain D s0432-1, one of the most virulent D. solani strains [9]. We have demonstrated that D. solani D s0432-1 is able to inhibit the growth of a variety of living microorganisms. We tested bio-activities against Gram-negative bacteria (E. coli, P. atrosepticum, D. dadantii), Gram-positive bacteria (B. subtilis, S. scabiei), ascomycete yeasts (S. cerevisiae, K. lactis, C. albicans), and ascomycete fungi (B. cinerea, M. oryzae and S. sclerotiorum). The cluster ooc encodes a biosynthesis pathway that produces an oocydin-like molecule, a chlorinated macrocyclic lactone molecule having antifungal, antioomycete and antitumor activities in S. plymuthica [15,16]. It was previously shown that the D. solani strains MK10, MK16 and IPO 2222, all encoding the oocydin cluster, inhibit the growth of the ascomycete fungus V. dahliae and the oomycete P. ultimum [16]. Our study showed that D. solani D s0432-1 prevents the growth of the ascomycete fungi B. cinerea, M. oryzae and S. sclerotiorum, and that inactivation of the oocL gene suppressed the inhibition of the three fungi without any visible effect on the bacteria or yeasts tested. We can thus conclude that the ooc cluster of D. solani produces an antiascomycete molecule.
By testing other mutants of the other secondary metabolite clusters, we showed that the mutant ΔsolG was nearly completely unable to prevent the growth of the ascomycete yeast S. cerevisiae and K. lactis. Therefore, the sol cluster appears to be responsible for yeast inhibition. D. solani D s0432-1 also inhibits the human pathogen yeast C. albicans, but its inhibitory activity appears to be weaker than that observed with S. cerevisiae and K. lactis in our inhibition assay. This observation on the sol cluster paves the way for the identification of a new molecule with highly specific inhibitory activity against yeasts. The sol cluster is conserved in several Dickeya genomes but rarely in other bacterial genera, except in some Rouxiella species. The structure of the molecule and its target remains to be elucidated.
Our study also showed that D. solani D s0432-1 has an antibacterial activity linked to the cluster zms, which encodes a zeamine biosynthetic pathway. Zeamine produced by S. plymuthica kills B. subtilis, the yeasts S. cerevisiae and Schizosaccharomyces pombe, and the nematode Caenorhabditis elegans [24]. We showed that D. solani D s0432-1, but not the ΔzmsA mutant, can clearly inhibit the growth of two Gram-positive bacteria, B. subtilis and S. scabiei. Since D. solani and S. scabiei are two potato pathogens, they might be in competition for the same ecological niche. We also observed a slight inhibition exerted by D. solani D s0432-1 against E. coli, but no inhibition towards D. dadantii and P. atrosepticum. Zeamine resistance of D. dadantii could be explained by the presence in its genome of the desAB genes encoding an RND pump identified in D. oryzae EC1 and involved in zeamine efflux [34]. In-frame deletion of desA or desB in D. oryzae EC1 leads to a zeamine sensitive phenotype [34]. In in planta Tn-seq experiments with D. dadantii 3937 [35], mutants of the genes desA and desB (Dda3937_00787 and Dda3937_00786 respectively) did not display a significant negative or positive variation (Log2 fold-changes -0.09 and +0.44, respectively), suggesting that this RND efflux pump does not play a significant role during D. dadantii chicory infection, contrary to the D. dadantii RND efflux pump AcrAB that appeared to be essential for virulence [35]. While P. atrosepticum is not inhibited by D. solani D s0432-1, it does not have the desAB genes. Zeamine resistance could be provided by another efflux pump or by a different mechanism in P. atrosepticum. It is worth remembering that the D. solani D s0432-1 ΔzmsA mutant grows slightly faster than the WT strain after reaching OD600nm 0.4 (S1 Fig). This small difference in growth might be explained by the lack of need of the mutant to adapt to the presence of zeamine in the medium. We also observed that D. solani D s0432-1 ΔsolG produced a very slight halo of inhibition against S. cerevisiae and K. lactis, barely visible in our tests. This weak halo disappears in the double mutant ΔsolGΔzmsA and the triple mutant ΔsolGΔoocLΔzmsA, showing that the zeamine produced by the D. solani zms cluster has also an anti-yeast action. It is important to note that zeamine produced by S. plymuthica A153 is bioactive against S. cerevisiae and S. pombe [24]. Thus, the presence of three clusters encoding secondary metabolites allows D. solani D s0432-1 to produce an arsenal of bioactive secondary metabolites against a variety of living microorganisms. The cluster sol produces an unknown molecule active against ascomycete yeasts, the cluster ooc produces an oocydin-like molecule active against ascomycete fungi and the cluster zms produces a zeamine-like molecule active against bacteria and ascomycete yeast. It is now clear that D. solani synthesizes several molecules to inhibit the growth of ascomycetes. Through a variety of regulatory mechanisms, the bacterium may always be able to express at least one of the sol, zms, and ooc clusters. This can make it particularly effective in the struggle against rival microorganisms. In future studies, it would be instructive to investigate D. solani behavior toward other organisms found in the environment, such as amoebas, nematodes, or even insects.
We then expanded our studies to a larger set of D. solani strains isolated in different European countries and years. Through the combination of phenotypic analysis and comparative genomics, confirmed by mutant construction, we demonstrated that the ArcZ sRNA plays a major role in activating the antimicrobial activities of the clusters sol and zms. We first observed that some D. solani WT strains are defective for yeast inhibition and possess different arcZ alleles, a deletion in the case of strain IFB0223, or the presence of the SNP G90A in the 3’ part of arcZ in the type strain IPO 2222. We demonstrated that these arcZ alterations significantly impair the D. solani ability to inhibit other bacteria and yeasts. These arcZ-related genomic changes represent a minority of D. solani isolates, only two of the nine strains phenotypically tested. Hence, our results show that despite a probable clonal origin, environmental strains of D. solani can present very different characteristics due to a single mutation into a regulatory sRNA. The arcZ polymorphism may be one of the explanations for the variability of phenotypes and virulence observed between genetically close D. solani strains. A global investigation of the polymorphism of all the small RNAs within the genomes of the same species could give surprising results. This study shows that it is critical to pay attention to the allelic diversity due to SNPs present not only in protein coding genes but also in sRNA genes.
ArcZ is a well-known regulatory sRNA that is associated with the chaperone Hfq. The ArcZ primary transcript (121 to 134nt-long) is converted by the essential endoribonuclease RNAse E [36] to a shorter stable of about 60 nt RNA retaining the seed region. The precise length of this short processed RNAs depends on the bacterial species [30,33,37,38]. In E. coli, the processed short form of ArcZ promotes rpoS translation while suppressing the expression of several other genes. In anaerobic conditions, arcZ is repressed and translation of RpoS is low [37]. RpoS is the RNA polymerase alternative sigma factor that governs the general stress response, which is activated by a wide range of conditions. It controls 10% of the E. coli genome [39]. RpoS can control positively or negatively the production of metabolites such as antibiotics [16,40]. The rpoS mRNA is not the only ArcZ direct target. In Photorhabdus and Xenorhabdus, ArcZ pairs with the mRNA encoding HexA, a transcriptional repressor of the expression of specific metabolite gene clusters [33]. In Dickeya species, the regulator gene pecT is the orthologous gene to hexA. In D. dadantii 3937, the pecT mRNA is also a direct target of ArcZ. Short processed ArcZ binds to Hfq and basepairs with the 5’UTR region of pecT mRNA, repressing its translation [38]. PecT is, directly and indirectly, involved in the expression of genes encoding plant cell wall degrading enzymes and virulence factors [38]. In the phytopathogen Erwinia amylovora, ArcZ also participates in the positive control of T3SS, exopolysaccharide production, biofilm formation, and motility [41]. Finally, another known direct target of ArcZ is the flhD mRNA [42,43]. FlhD forms with FlhC the master regulator FlhDC that controls motility in bacteria [44]. Although we demonstrated the critical positive role of ArcZ in the regulation of sol and zms expression, and secondary metabolite production in D. solani, it remains to understand the precise mechanism of ArcZ action on the sol and zms clusters. Our CopraRNA analysis indicates no putative binding between the processed form of ArcZ1 and the 5’UTR region of solA and zmsA. However, the pecT 5’UTR, but not flhD and rpoS 5’UTRs, is detected as a putative target of ArcZ1 (S3 Table). It should be kept in mind that bioinformatics analyses remain predictions. For example, CopraRNA does not detect the interaction between E. coli ArcZ and the 5’UTR of the rpoS mRNA. Much work remains to be done to elucidate the mode of action of ArcZ in Dickeya solani.
Another point to consider is a more global role of D. solani ArcZ exerted on other cellular functions such as motility and virulence. ArcZ targets in D. solani need to be investigated in the future. It is worth pointing out that the G90 nucleotide is in the 3’ region of ArcZ that interacts with the target mRNAs rpoS, flhD, or pecT in other bacteria (S5 Fig). Northern-blot analyses revealed that ArcZ2 of D. solani IPO 2222 is not processed in the stable shorter sRNA as ArcZ1 of D. solani D s0432-1. The G90A mutation in IPO 2222 ArcZ2 is predicted to modify the secondary structure of the ArcZ precursor form (Fig 7B). Since maturation by RNase E is essential for target regulation by the ArcZ sRNA in Salmonella [36], we conclude that the absence of a short form of ArcZ in IPO 2222 is probably the main cause of the loss of regulatory function of the IPO 2222 ArcZ2.
The type strain D. solani IPO 2222 is largely used in academic laboratories as a model to study the D. solani species. In light of our findings, it could be important to compare data obtained with the strain IPO 2222 with those obtained with other D. solani strains. The arcZ mutations may be responsible, at least partially, for the lower virulence of IPO 2222 and IFB0223 observed previously in comparison to D s0432-1 [9]. ArcZ is expressed in aerobic condition in stationary phase. Thus, in IFB0223 and the type strain IPO 2222, one can assume that each gene controlled by ArcZ in this condition will not be properly regulated at the post-transcriptional level. IPO 2222 exhibited a low number of genes with oxygen-dependent transcription [45]. It might be possible that the missing functional ArcZ in this strain prevents the oxygen-dependent response of some genes. Besides, IPO 2222 was isolated from infected seed potatoes in the Netherlands in 2007 [46]. It is less virulent than D s0432-1 [9] but can still successfully colonize potato and bittersweet nightshade (Solanum dulcamara) plants [47,48]. These observations imply that ArcZ might not be necessary for a successful plant infection. However, when D. solani is present on the phylosphere, where oxygen is available, ArcZ might be crucial for outcompeting other microorganisms. It will be necessary to conduct more research to validate these hypotheses.
Even strains with the same arcZ, sol, ooc, and zms genomic sequences as D s0432-1, showed heterogeneity in their ability to inhibit other microorganisms. For example, IPO 3494 and IFB0484 inhibited B. subtilis better than other D. solani strains (Fig 5A). Other specific mutations in the genomes of these strains, which could also modify the activity of secondary metabolite clusters, could explain these phenotypic variations.
We first described the inhibitory function of the sol cluster of the D. solani strain D s0432-1 against the yeasts S. cerevisiae and K. lactis in a 2021 bioRxiv preprint [19]. Very recently, in November 2022, Matilla et al. published results on the sol cluster of the D. solani strain MK10 [14]. The authors described the antifungal activity of the uncharacterized molecule that they called solanimycin [14]. Matilla et al. also reported the inhibition observed with the D. solani strain IPO 2222 against Verticillium dahliae and S. pombe ([14], S1 Fig of the article). In our conditions, strain IPO 2222 was able to inhibit the fungus S. sclerotiorum but not the yeast K. lactis (Fig 5). In comparison to our study, Matilla et al used different fungal species but also different growth conditions (medium, temperature, exposure time, etc). By random mutagenesis of strain MK10, they identified RsmA and the two Dickeya quorum-sensing systems ExpI/R and Vfm as regulators of the sol cluster [14]. Interestingly, the Rsm regulatory pathway of D. dadantii 3937 is regulated by PecT, which is under the ArcZ control [38]. Our CopraRNA analysis identified pecT, and vfmX, one of the Vfm quorum sensing genes, as putative targets of ArcZ, but not rsmA, expI and expR, (S3 Table).
Further studies are now necessary to specify the biological targets of solanimycin and to clarify the role in sol expression of the sRNA ArcZ, quorum-sensing Vfm and Exp, and transcriptional regulators PecT or RsmA, in order to better understand the regulatory network controlling the production of secondary metabolites in D. solani. We are currently investigating the regulatory pathway of secondary metabolite clusters involving ArcZ.
In conclusion, our results provide the basis for further investigation of a novel metabolite capable of limiting yeast growth and, at the same time, open doors to understand how the ArcZ sRNA acts as a global regulator to regulate two different clusters of secondary metabolites in D. solani. It also shows how a single mutation in a small RNA can greatly modify the phenotype of bacterial wild-type strains that are very close phylogenetically.
Material and methods
Bacterial and fungal strains, plasmids and growth conditions
The E. coli and Dickeya bacterial strains, plasmids and oligonucleotides used in this study are described in S4 and S5 Tables. The genome accession number of D. solani D s0432-1 is NZ_CP017453. The following strains have been used in the study: Sclerotinia sclerotiorum S5, Botrytis cinerea B05.10, Magnaporthe oryzae Guy11, Saccharomyces cerevisiae BY4743 (MATa/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 LYS2/lys2Δ0 met15Δ0/MET15 ura3Δ0/ura3Δ0), Kluyveromyces lactis MWL9S1 [49], Candida albicans SC5314, Dickeya dadantii 3937, Pectobacterium atrosepticum SCRI1043, Streptomyces scabiei CFBP4517, Bacillus subtilis PY79. E. coli was grown routinely at 37°C in LB. Fungus strains were grown at 25°C onto Potato Dextrose Agar (PDA). S. scabiei was grown in tryptic soy broth (TSB) medium at 28°C. B. subtilis, P. atrosepticum and the Dickeya strains were cultivated in LB unless specified. Yeast cells were grown at 30°C in rich medium consisting of complete yeast extract-peptone (YP) medium containing 1% Bacto yeast extract, 1% Bacto peptone (Difco) supplemented with 2% glucose (yeast extract-peptone-dextrose [YPD] medium). For the bacteria, yeast and fungi inhibition assay, M63 medium supplemented with sucrose (2 g (NH4)2SO4, 13.6 g KH2PO4, 2.5 mg FeSO47H2O, 0.2 g MgSO47H2O, 10 g sucrose, per liter) were employed to grow overnight the D. solani strains before performing the assay.
When required, antibiotics were added at the following concentrations: ampicillin (Amp), 100 μg/L; nalidixic acid (Nal), 10 μg/mL. streptomycin (Sm), 50 μg/mL; Diaminopimelic acid (DAP) (57 μg/mL) was added for the growth of the E. coli MFDpir strain. Media were solidified with 12 g/L agar.
Growth inhibition assay of bacteria and yeast
D. solani strains were grown for 24 h at 30°C with shaking in M63 medium supplemented with 1% sucrose. B. subtilis, E. coli, D. dadantii and P. atrosepticum were grown in LB at 30°C with shaking overnight. The next day, the OD600 of the cultures of D. solani were adjusted to 2. The temperature of melted LB agar was lowered to around 40°C (just before the agar resolidified). 100 mL of the LB agar in surfusion were mixed with 100 μl of the OD600 1 culture of B. subtilis, E. coli, D. dadantii or P. atrosepticum. 30 mL of inoculated LB agar were poured in 12- by 12-cm square plates. Then 5 μl of the OD600 2 cultures of D. solani were spotted onto the inoculated square plates which were incubated at 30°C for 24–48 h before visualization of the inhibition zone.
The same experiment was conducted with S. scabiei, except that it was grown for 3 days in TSB at 28°C and TSB agar was poured in the square plates. With the yeasts S. cerevisiae, K. lactis, and C. albicans, a similar protocol was used with YPD medium and with OD600 2 culture of yeast per 30mL of medium.
Growth inhibition of fungal strains
S. sclerotiorum, B. cinerea and M. oryzae were grown onto PDA plates at 25°C for 5, 7 and 10 days respectively. D. solani strains were grown for 24 h at 30°C with shaking in M63 medium supplemented with 1% sucrose. The OD600 of these overnight cultures were adjusted to 2. Then, 5 μl of the bacterial suspensions were spotted onto PDA plates with 5 mm agar plugs of fungus at the center of the Petri dish. The radius of fungus inhibition zone was measured.
Construction of the MLSA tree positioning strains within the Dickeya genus
Details on the methods used for the construction of MLSA tree positioning strains are given in S1 Supplementary Methods.
Constructions of the mutant strains and plasmids
All the material and methods for the construction of the strains and plasmids used in this study are given in S1 Supplementary Methods.
RNA isolation and Northern detection
All the material and methods for RNA isolation and Northern detection performed in this study are given in S1 Supplementary Methods.
Supporting information
S1 Fig. Growth curves of D. solani strains and mutant derivatives.
A 96-well plate containing M63 medium with 1% sucrose was inoculated with the strains used in this study at an OD of 0.06. The growth of each strain was determined by measuring OD600 every 20 minutes during 2 days in a TECAN device. The different strains showed similar overall growth, except for the light grey points where a low but significant fitness gain was observed for the mutants Δzms, Δsol Δzms, Δooc Δzms and Δsol Δooc Δzms (Mann-Whitney test, p-value<0.05).
https://doi.org/10.1371/journal.pgen.1010725.s001
(TIF)
S2 Fig. Inhibition of E. coli growth by D. solani D s0432-1 and mutant derivatives (zoom on drop borders).
Bioassay plates were prepared by mixing E. coli culture with melted LB agar as described in the experimental procedures. 5 μL of bacterial culture at OD600nm = 2 of D. solani D s0432-1 (WT) or derivatives were spotted onto the plate and incubated at 30°C during 48 h. A slight inhibition zone was observed except with the Δzms mutant. All experiments were carried in 4 replicates.
https://doi.org/10.1371/journal.pgen.1010725.s002
(TIF)
S3 Fig. Phenotypes and arcZ sequence of WT D. solani IPO 2222 from the BCCM collection (LMG 25993).
(A) Inhibition assay of K. lactis, B. subtilis and S. sclerotiorum by WT D. solani IPO 2222 LMG 25993 from the BCCM collection, compared to the strains used in this study. (B) Sanger sequencing results of arcZ of D. solani IPO 2222 (LMG 25993). The mutation G90A is highlighted by a red frame.
https://doi.org/10.1371/journal.pgen.1010725.s003
(TIF)
S4 Fig. Growth curves of WT and ΔarcZ D. solani strains used in this study.
(A) Growth curves of D s0432-1, IPO 2222, their respective ΔarcZ mutants and IFB0223. The latter (in grey) has a slight growth delay (Mann-Whitney test, p-value<0.05). (B) Growth curves of strains used in complementation and heterologous expression tests. The plasmids do not cause any growth defect in D s0432-1, IPO 2222 and ΔarcZ mutants. IFB0223 derivatives containing a plasmid (in blue) have a growth delay already observed in the strain without plasmid (Mann-Whitney test, p-value<0.05). All experiments were performed in 4 biological replicates.
https://doi.org/10.1371/journal.pgen.1010725.s004
(TIF)
S5 Fig. Alignment of arcZ sequences.
The arcZ DNA sequences were retrieved by running a BlastN on the NCBI database using the arcZ sequence of D. solani D s0432-1 as query. Search was limited to the Dickeya genomes. Then, the Dickeya arcZ sequences were aligned with E. coli MG1655 arcZ by using Jalview [50] and Muscle [51]. Known regions of interactions with the pecT, flhD and rpoS 5’UTR mRNA in D. dadantii and E. coli are indicated.
https://doi.org/10.1371/journal.pgen.1010725.s005
(TIF)
S1 Supplementary Methods. DOCX file with supplementary material and methods.
https://doi.org/10.1371/journal.pgen.1010725.s006
(DOCX)
S1 Table. Distribution of the clusters sol, ooc, and zms in 155 Dickeya strains whose genome is sequenced.
For each Dickeya species, the strains were classified on the basis of the presence or absence of the cluster sol, ooc, and zms.
https://doi.org/10.1371/journal.pgen.1010725.s007
(XLSX)
S2 Table. Lists of SNPs in D. solani IFB0099, D S0432.1, IFB0223, IPO3337 and IPO 2222.
https://doi.org/10.1371/journal.pgen.1010725.s008
(XLSX)
S3 Table. List of the top 200 putative ArcZ/mRNA Dickeya solani D s0432-1 interactions.
https://doi.org/10.1371/journal.pgen.1010725.s009
(XLSX)
S4 Table. Strains and plasmids used in the study.
https://doi.org/10.1371/journal.pgen.1010725.s010
(DOCX)
S5 Table. Oligonucleotides used in the study.
https://doi.org/10.1371/journal.pgen.1010725.s011
(DOCX)
S1 Data. Raw data used to construct the figures, with their statistical analysis.
The file can be open with Prism (Graphpad). A free viewer mode is available and can be downloaded online.
https://doi.org/10.1371/journal.pgen.1010725.s012
(PZFX)
Acknowledgments
We thank Veronique Utzinger for technical assistance, the members of the Membrane Trafficking and Signaling in Bacteria (MTSB) team for discussions, Amélie De Vallée, Nathalie Poussereau and Christophe Bruel for advices on fungi growth and providing the strains, Alexandre Soulard, Marc Lemaire and Jade Ravent for advices on yeast growth and providing the strains and medium.
References
- 1.
Hugouvieux-Cotte-Pattat N, Condemine G, Gueguen E, Shevchik VE. Dickeya Plant Pathogens. eLS. Pettis G. (ed), Chichester, UK: John Wiley & Sons; 2020. pp. 1–10. https://doi.org/10.1002/9780470015902.a0028932
- 2. Charkowski A, Blanco C, Condemine G, Expert D, Franza T, Hayes C, et al. The role of secretion systems and small molecules in soft-rot Enterobacteriaceae pathogenicity. Annu Rev Phytopathol. 2012;50: 425–449. pmid:22702350
- 3.
Van Gijsegem F, Toth IK, van der Wolf JM. Soft Rot Pectobacteriaceae: A Brief Overview. In: Van Gijsegem F, van der Wolf JM, Toth IK, editors. Plant Diseases Caused by Dickeya and Pectobacterium Species. Cham: Springer International Publishing; 2021. pp. 1–11. https://doi.org/10.1007/978-3-030-61459-1_1
- 4. Hugouvieux-Cotte-Pattat N, Condemine G, Shevchik VE. Bacterial pectate lyases, structural and functional diversity: Bacterial pectate lyases. Environ Microbiol Rep. 2014;6: 427–440. pmid:25646533
- 5. van der Wolf JM, Nijhuis EH, Kowalewska MJ, Saddler GS, Parkinson N, Elphinstone JG, et al. Dickeya solani sp. nov., a pectinolytic plant-pathogenic bacterium isolated from potato (Solanum tuberosum). Int J Syst Evol Microbiol. 2014;64: 768–774.
- 6. Blin P, Robic K, Khayi S, Cigna J, Munier E, Dewaegeneire P, et al. Pattern and causes of the establishment of the invasive bacterial potato pathogen Dickeya solani and of the maintenance of the resident pathogen D. dianthicola. Mol Ecol. 2021;30: 608–624. pmid:33226678
- 7. Garlant L, Koskinen P, Rouhiainen L, Laine P, Paulin L, Auvinen P, et al. Genome Sequence of Dickeya solani, a New soft Rot Pathogen of Potato, Suggests its Emergence May Be Related to a Novel Combination of Non-Ribosomal Peptide/Polyketide Synthetase Clusters. Diversity. 2013;5: 824–842.
- 8. Pédron J, Mondy S, des Essarts YR, Van Gijsegem F, Faure D. Genomic and metabolic comparison with Dickeya dadantii 3937 reveals the emerging Dickeya solani potato pathogen to display distinctive metabolic activities and T5SS/T6SS-related toxin repertoire. BMC Genomics. 2014;15: 283. pmid:24735398
- 9. Golanowska M, Potrykus M, Motyka-Pomagruk A, Kabza M, Bacci G, Galardini M, et al. Comparison of Highly and Weakly Virulent Dickeya solani Strains, With a View on the Pangenome and Panregulon of This Species. Front Microbiol. 2018;9: 1940. pmid:30233505
- 10. Motyka-Pomagruk A, Zoledowska S, Misztak AE, Sledz W, Mengoni A, Lojkowska E. Comparative genomics and pangenome-oriented studies reveal high homogeneity of the agronomically relevant enterobacterial plant pathogen Dickeya solani. BMC Genomics. 2020;21: 449. pmid:32600255
- 11. Khayi S, Blin P, Pédron J, Chong T-M, Chan K-G, Moumni M, et al. Population genomics reveals additive and replacing horizontal gene transfers in the emerging pathogen Dickeya solani. BMC Genomics. 2015;16: 788. pmid:26467299
- 12. Cane DE, Walsh CT. The parallel and convergent universes of polyketide synthases and nonribosomal peptide synthetases. Chem Biol. 1999;6: R319–R325. pmid:10631508
- 13. Duprey A, Taib N, Leonard S, Garin T, Flandrois JP, Nasser W, et al. The phytopathogenic nature of Dickeya aquatica 174/2 and the dynamic early evolution of Dickeya pathogenicity. Environ Microbiol. 2019;21: 2809–2835. pmid:30969462
- 14. Matilla MA, Monson RE, Murphy A, Schicketanz M, Rawlinson A, Duncan C, et al. Solanimycin: Biosynthesis and Distribution of a New Antifungal Antibiotic Regulated by Two Quorum-Sensing Systems. Sperandio V, editor. mBio. 2022; e02472–22. pmid:36214559
- 15. Matilla MA, Stöckmann H, Leeper FJ, Salmond GPC. Bacterial Biosynthetic Gene Clusters Encoding the Anti-cancer Haterumalide Class of Molecules. J Biol Chem. 2012;287: 39125–39138. pmid:23012376
- 16. Matilla MA, Leeper FJ, Salmond GPC. Biosynthesis of the antifungal haterumalide, oocydin A, in Serratia, and its regulation by quorum sensing, RpoS and Hfq. Environ Microbiol. 2015;17: 2993–3008. pmid:25753587
- 17. Zhou J, Zhang H, Wu J, Liu Q, Xi P, Lee J, et al. A Novel Multidomain Polyketide Synthase Is Essential for Zeamine Production and the Virulence of Dickeya zeae. Mol Plant-Microbe Interactions®. 2011;24: 1156–1164. pmid:21899437
- 18. Masschelein J, Mattheus W, Gao L-J, Moons P, Houdt RV, Uytterhoeven B, et al. A PKS/NRPS/FAS Hybrid Gene Cluster from Serratia plymuthica RVH1 Encoding the Biosynthesis of Three Broad Spectrum, Zeamine-Related Antibiotics. PLOS ONE. 2013;8: e54143. pmid:23349809
- 19. Effantin G, Brual T, Rahbé Y, Hugouvieux-Cotte-Pattat N, Gueguen E. Dickeya solani D s0432-1 produces an arsenal of secondary metabolites with anti-prokaryotic and anti-eukaryotic activities against bacteria, yeasts, fungi, and aphids. bioRxiv; 2021. p. 2021.07.19.452942. https://www.biorxiv.org/content/10.1101/2021.07.19.452942v1
- 20. Strobel G, Li J-Y, Sugawara F, Koshino H, Harper J, Hess WM. Oocydin A, a chlorinated macrocyclic lactone with potent anti-oomycete activity from Serratia marcescens. Microbiology. 1999;145: 3557–3564. pmid:10627053
- 21. Zhou J, Cheng Y, Lv M, Liao L, Chen Y, Gu Y, et al. The complete genome sequence of Dickeya zeae EC1 reveals substantial divergence from other Dickeya strains and species. BMC Genomics. 2015;16: 571. pmid:26239726
- 22. Wang X, He S-W, Guo H-B, Han J-G, Thin KK, Gao J-S, et al. Dickeya oryzae sp. nov., isolated from the roots of rice. Int J Syst Evol Microbiol. 2020;70: 4171–4178. pmid:32552985
- 23. Masschelein J, Jenner M, Challis GL. Antibiotics from Gram-negative bacteria: a comprehensive overview and selected biosynthetic highlights. Nat Prod Rep. 2017;34: 712–783. pmid:28650032
- 24. Hellberg JEEU Matilla MA, Salmond GPC. The broad-spectrum antibiotic, zeamine, kills the nematode worm Caenorhabditis elegans. Front Microbiol. 2015;6. pmid:25767467
- 25. Loria R, Kers J, Joshi M. Evolution of Plant Pathogenicity in Streptomyces. Annu Rev Phytopathol. 2006;44: 469–487. pmid:16719719
- 26. Rodicio R, Heinisch JJ. Yeast on the milky way: genetics, physiology and biotechnology of Kluyveromyces lactis. Yeast. 2013;30: 165–177. pmid:23576126
- 27. Valentin-Hansen P, Eriksen M, Udesen C. MicroReview: The bacterial Sm-like protein Hfq: a key player in RNA transactions. Mol Microbiol. 2004;51: 1525–1533. pmid:15009882
- 28. Vogel J, Luisi BF. Hfq and its constellation of RNA. Nat Rev Microbiol. 2011;9: 578–589. pmid:21760622
- 29. Dutta T, Srivastava S. Small RNA-mediated regulation in bacteria: A growing palette of diverse mechanisms. Gene. 2018;656: 60–72. pmid:29501814
- 30. Papenfort K, Said N, Welsink T, Lucchini S, Hinton JCD, Vogel J. Specific and pleiotropic patterns of mRNA regulation by ArcZ, a conserved, Hfq-dependent small RNA. Mol Microbiol. 2009;74: 139–158. pmid:19732340
- 31. Wright PR, Georg J, Mann M, Sorescu DA, Richter AS, Lott S, et al. CopraRNA and IntaRNA: predicting small RNA targets, networks and interaction domains. Nucleic Acids Res. 2014;42: W119–123. pmid:24838564
- 32. Argaman L, Hershberg R, Vogel J, Bejerano G, Wagner EGH, Margalit H, et al. Novel small RNA-encoding genes in the intergenic regions of Escherichia coli. Curr Biol. 2001;11: 941–950. pmid:11448770
- 33. Neubacher N, Tobias NJ, Huber M, Cai X, Glatter T, Pidot SJ, et al. Symbiosis, virulence and natural-product biosynthesis in entomopathogenic bacteria are regulated by a small RNA. Nat Microbiol. 2020;5: 1481–1489. pmid:33139881
- 34. Liang Z, Huang L, He F, Zhou X, Shi Z, Zhou J, et al. A Substrate-Activated Efflux Pump, DesABC, Confers Zeamine Resistance to Dickeya zeae. mBio. 2019;10. pmid:31138747
- 35. Royet K, Parisot N, Rodrigue A, Gueguen E, Condemine G. Identification by Tn-seq of Dickeya dadantii genes required for survival in chicory plants. Mol Plant Pathol. 2019;20: 287–306. pmid:30267562
- 36. Chao Y, Li L, Girodat D, Förstner KU, Said N, Corcoran C, et al. In Vivo Cleavage Map Illuminates the Central Role of RNase E in Coding and Non-coding RNA Pathways. Mol Cell. 2017;65: 39–51. pmid:28061332
- 37. Mandin P, Gottesman S. Integrating anaerobic/aerobic sensing and the general stress response through the ArcZ small RNA. EMBO J. 2010;29: 3094–3107. pmid:20683441
- 38. Yuan X, Zeng Q, Khokhani D, Tian F, Severin GB, Waters CM, et al. A feed-forward signalling circuit controls bacterial virulence through linking cyclic di-GMP and two mechanistically distinct sRNAs, ArcZ and RsmB. Environ Microbiol. 2019;21: 2755–2771. pmid:30895662
- 39. Battesti A, Majdalani N, Gottesman S. The RpoS-mediated general stress response in Escherichia coli. Annu Rev Microbiol. 2011;65: 189–213. pmid:21639793
- 40. Wilf NM, Salmond GPC. The stationary phase sigma factor, RpoS, regulates the production of a carbapenem antibiotic, a bioactive prodigiosin and virulence in the enterobacterial pathogen Serratia sp. ATCC 39006. Microbiology. 2012;158: 648–658.
- 41. Zeng Q, McNally RR, Sundin GW. Global Small RNA Chaperone Hfq and Regulatory Small RNAs Are Important Virulence Regulators in Erwinia amylovora. J Bacteriol. 2013;195: 1706–1717. pmid:23378513
- 42. De Lay N, Gottesman S. A complex network of small non-coding RNAs regulate motility in Escherichia coli. Mol Microbiol. 2012;86: 524–538. pmid:22925049
- 43. Schachterle JK, Zeng Q, Sundin GW. Three Hfq-dependent small RNAs regulate flagellar motility in the fire blight pathogen Erwinia amylovora. Mol Microbiol. 2019;111: 1476–1492. pmid:30821016
- 44. Fraser GM, Hughes C. Swarming motility. Curr Opin Microbiol. 1999;2: 630–635. pmid:10607626
- 45. Lisicka W, Fikowicz-Krosko J, Jafra S, Narajczyk M, Czaplewska P, Czajkowski R. Oxygen Availability Influences Expression of Dickeya solani Genes Associated With Virulence in Potato (Solanum tuberosum L.) and Chicory (Cichorium intybus L.). Front Plant Sci. 2018;9: 374. pmid:29619040
- 46. Pritchard L, Humphris S, Baeyen S, Maes M, Van Vaerenbergh J, Elphinstone J, et al. Draft Genome Sequences of Four Dickeya dianthicola and Four Dickeya solani Strains. Genome Announc. 2013;1. pmid:23887905
- 47. Fikowicz-Krosko J, Czajkowski R. Systemic Colonization and Expression of Disease Symptoms on Bittersweet Nightshade (Solanum dulcamara) Infected with a GFP-Tagged Dickeya solani IPO2222 (IPO2254). Plant Dis. 2018;102: 619–627. pmid:30673477
- 48. Czajkowski R, de Boer WJ, Velvis H, van der Wolf JM. Systemic Colonization of Potato Plants by a Soilborne, Green Fluorescent Protein-Tagged Strain of Dickeya sp. Biovar 3. Phytopathology®. 2010;100: 134–142. pmid:20055647
- 49. Wésolowski-Louvel M. An efficient method to optimize Kluyveromyces lactis gene targeting. FEMS Yeast Res. 2011;11: 509–513. pmid:21627769
- 50. Clamp M, Cuff J, Searle SM, Barton GJ. The Jalview Java alignment editor. Bioinforma Oxf Engl. 2004;20: 426–427. pmid:14960472
- 51. Edgar RC. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics. 2004;5: 113. pmid:15318951