Résumé
Comparison of soils suppressive or not to take-all disease of wheat by 16S taxonomic microarray targeting plant-protecting rhizosphere bacteria. Take-all is an important wheat disease caused by the soil-borne fungus Gaeumannomyces graminis var. tritici. Disease severity can be high, but a decline of take-all disease may take place in the following years in case of wheat monocropping. Microbial populations known to be associated to take-all decline (disease suppressiveness) include culturable antagonistic fluorescent Pseudomonas spp. producing the antifungal compound 2,4-diacetylphloroglucinol. The objective of this study was to assess changes in the diversity of rhizosphere pseudomonads linked with take-all decline of wheat, following a culture-independent approach based on the use of a 16S rRNA-based taxonomic microarray. The microarray contains about 700 probes, which target bacteria at various taxonomic levels. Certain probes (about 70) target pseudomonads, including groups of biocontrol Pseudomonas strains (about 20 probes). The latter were defined in this work after phylogenetic analysis of 16S rRNA sequences available in databases for biocontrol strains and other pseudomonads, based on Neighbor Joining (with Kimura 2 parameter) and Maximum Parcimony methods. The microarray probes were defined using ARB software (www.arb-home.de) and they were checked in silico. Most have a melting temperature of 65 ± 5 °C, GC > 50 %, no secondary structure or a secondary structure with a -2 kcal mol-1 and melting temperature < 50 °C, and they do not form stable homoduplex. The probes were spotted by 5' C6-NH2 covalent link on slides. Each basic pattern for the probe set is replicated three times per slide. Rhizosphere samples were collected at the INRA station of La Gruche (Brittany, France) from plots grown with wheat for one year (treatment PI; low level of takeall disease), five years (treatment PV; high level of disease) or ten years (treatment PX; low level of disease, suppressiveness reached). This experimental set-up enabled comparison of treatments under same conditions of soil composition, microclimate, wheat cultivar and farming techniques. Rhizosphere DNA was extracted and subjected to PCR using priming conditions designed in this work to target biocontrol pseudomonads. Cluster analysis of microarray data obtained after PCR of Pseudomonas populations in the wheat rhizosphere discriminated mainly between treatments PI and PX, treatment PV being in an intermediate position. The results are in accordance with quantitative PCR data obtained for the total pseudomonads. Overall, treatment PX appears to be associated with a particular composition in biocontrol pseudomonads (which comprise biocontrol strains producing 2,4- diacetylphloroglucinol). These findings indicate that the 16S taxonomic microarray is a promising tool for analysis of bacterial diversity associated with diseasesuppressive soils.