Genetic traits required by bacterial phytopathogens for plant infection and fitness in non agricultural habitats using high-throughput transposon sequencing (Tn-seq) – TNPHYTO

The 1st task was to determine the bacterial strains we were going to use in our Tn-seq screenings in plants and river water. To do this, they had to meet several criteria: 1- Be highly mutagenisable; 2- Be capable of multiplying in the same host plants; 3- Be capable of multiplying in vitro in the same culture medium. The 1ᵉʳˢ months consisted of selecting, from a collection of Dickeya, Pectobacterium and P. syringae strains, the strains that met these criteria. Where necessary, the DNAs of these strains were prepared by partner 3 and sent for both Illumina and Nanopore sequencing. The coding sequences of complete genomes were predicted using the RAST server and the table of orthologous genes obtained with the Oma v.2.4.2 software was sent to partner 4. At the same time, the mutant libraries of the 9 selected strains were prepared and the Tn-seq libraries sequenced in duplicate by the MGX platform in Montpellier. The mutant libraries were all of good quality and could therefore be used for genetic screening in vivo, in plants and in river water.

In parallel with the production of the mutant libraries, we evaluated the use of broad beans and melons in a controlled environment, as common hosts for the multiplication of the 9 bacterial strains. On the bean, they were all shown to be capable of multiplying after injection into a stem wound. On melon seedlings, partner 2 found that the Dickeya and Pectobacterium strains did not produce sufficiently large multiplications in this plant, unlike the P. syringae strains. The lettuce tested as a replacement plant did not prove suitable for our project. Broad beans were chosen as the only plant tested.

For the river water tests, partners 2 and 3 collected 100 litres of Durance water, which was filtered and autoclaved for the growth experiments. The carbon concentration was determined and other measurements (nitrate, nitrite, phosphate, water chemistry) were taken. Analysis of bacterial growth in the recovered Durance water was studied by Partners 2 and 3.

With regard to data analysis, partner 4 has laid the foundations for a bioinformatics pipeline that identifies essential and conditionally essential genes from the raw sequencing data, and visualises and compares the results between strains. This development is being carried out in Python, R, shiny and docker. The first data sets have already been processed. Several genes involved in plant virulence or growth in water have been identified. The construction of mutants in these genes will enable the Tnseq results to be validated.

Development of TnSEEK

Partner 4 has designed TNSEEK, a bioinformatics pipeline for identifying essential and conditionally essential genes from raw sequencing data. Developed using Python, R, Shiny and Docker, it can also be used to visualise and compare results between different strains. TNSEEK has been used to analyse the first data sets and will be improved on the basis of the feedback received.

Selection of strains and sequencing

We selected key strains for Tn-seq screening in plants and river water: P. aquaticum A212, P. versatile A73, D. fangzhongdai B16 and D. parazeae A586. These new strains complete an already established panel. Genome sequencing using Illumina and Nanopore technologies was used to predict the coding sequences.

Mutant libraries for nine strains were created, with Tn-seq libraries sequenced in duplicate using the MGX platform. The insertion density of transposons in TA sites ranged from 36.8% to 69.8%, guaranteeing good coverage. These analyses identified essential genes, validating TNSEEK for exploring essential genomes within closely related genera or families. These genomes show similarities with that of Escherichia coli.

Tests on plants

The mutant libraries were used to infect broad beans, the only plant able to host the nine strains studied. After a few days, bacterial populations in the plant tissue reached 10⁹ to 10¹¹ cells, allowing robust analysis of the Tn-seq libraries. The analysis revealed genes involved in virulence, some common to all three bacterial species, others specific to a given strain. These results are currently being validated by constructing the corresponding mutants. A key point concerns the small arcZ RNA, identified as essential for colonisation. In D. solani, a ∆arcZ mutant showed reduced virulence and a loss of ability to inhibit competing microorganisms. ArcZ also regulates secondary metabolites (Brual et al., 2023), highlighting its central role in virulence and inter-bacterial competition.

Tests in water

A genomic analysis led to the description of P. quasiaquaticum, a new species distinct from P. aquaticum on genomic, physiological and ecological grounds (Ben Moussa et al., 2021). These aquatic species have abandoned the plant niche (Ben Moussa et al., 2023). A comparative study of 80 strains determined their variable abundance in rivers (Ben Moussa et al., 2022). The Tn-seq results show the importance of phosphate assimilation for the survival of Dickeya and Pectobacterium in an aquatic environment (in preparation).

The Tn-seq screening carried out as part of this project identified a large number of genes essential to the virulence of the three bacterial species studied. Several of these genes encode transcriptional regulators, which play a central role in all the species. In-depth study of these regulators is a major prospect. We will focus on their target genes in order to decipher the molecular mechanisms underlying their contribution to bacterial virulence.

With regard to ArcZ, an sRNA that has already been well studied, we are making progress in understanding its role as a global regulator of virulence and the production of secondary metabolites in Dickeya solani. This function appears to be specific to D. solani and absent in other species of the Dickeya genus. One of its identified targets is conserved in the Pectobacteriaceae, but its role varies between species. To explore these functional differences, a comparative transcriptomic study using RNA-seq will be essential. This will enable us to decipher the networks regulated by ArcZ and explain the variability of its impact within this bacterial family.

Genes of unknown function involved in virulence represent a significant proportion of the targets identified. Their precise role remains to be elucidated. An approach combining protein structure prediction via AlphaFold and interactomic analyses will lay the foundations for exploring their function. Structural modelling will identify similarities with known proteins or predict active sites, while interactomics will reveal their molecular partners, offering clues to their integration in biological networks. These efforts will contribute to a better understanding of virulence mechanisms and could reveal new targets for control strategies.

Certain transcriptional regulators appear to be essential for virulence in some species but not in others, suggesting species-specific regulatory mechanisms. A comparative study using RNA-seq and ChIP-seq will identify the specific targets of these regulators and their interaction networks. Although not included in the initial deliverables of the ANR, this analysis is crucial for deciphering the virulence mechanisms specific to each species and identifying targets for innovative strategies.

Finally, certain genes whose absence increases virulence play a key role in survival in oligotrophic environments, such as river water. These results underline the importance of a holistic approach, considering the pathogen in its interactions with host plants and its natural habitats. In-depth exploration of these genes will provide a better understanding of bacterial adaptation to various environments and pave the way for ecologically targeted control strategies.

Plant pathogenic bacteria are a main threat for food production worldwide and no efficient means of control exist. Crop protection methods require knowledge of the genetic and phenotypic diversity of plant pathogens, of the extent of their host range and of habitats in which reservoirs of inoculum are found. Studies on plant pathogenic bacteria have often been conducted on a limited number of model laboratory strains and model plants. While some bacteria are found only in infected plants, others can be isolated in various environments such as soil, water or healthy plants. These strains could represent a potential source of disease outbreak and nothing is known on how environmental reservoirs weight on their survival, multiplication and pathogeny. From the perspective of developing more comprehensive approaches to plant disease studies, including environment, we propose to investigate genetic determinants of bacterial lifestyle both in planta and in fresh water, a major vector of pathogen dissemination through the river and irrigation networks. Availability of large collections of environmental strains and new high throughput sequencing techniques now allow to identify simultaneously, in a large panel of strains, genes required for growth in a given condition, like in planta, in water or in soil. In this project, we will use Tn-Seq, an innovative mutagenesis and high throughput sequencing technique to find such genetic determinants, in three wide spectrum pathogenic bacteria present in France: Pseudomonas syringae, Pectobacterium spp and Dickeya spp. Strains of these bacteria will be selected in our collections to represent the largest phylogenetic diversity and origin. The Tn-seq analysis will be performed on several plants and in river water. To analyse this data, we will develop a dedicated open-source bioinformatics pipeline, that will address the challenge of multiple conditions and multiple strains comparison. Validation of the results obtained will necessitate construction of a large number of mutants. In addition to the techniques commonly used to make mutants, we will use CrispRi gene silencing method in the three species. All these approaches will allow the identification of new genetic determinants involved in virulence on plants and those required for fitness in water, a major environmental habitats. Comparison of the two habitats will identify genetic determinants in common and lead to hypothesis on how environmental reservoirs could drive growth, survival and pathogeny of pathogens. Studying various pathogenic bacteria will allow to identify the virulence genetic determinants conserved at the genus, species or strain level. The results obtained in this work could help predict strain virulence, develop antibacterial strategies.