Field release of genetically modified Pseudomonas putida WCS358r : molecular analysis of effects on microbial communities in the rhizosphere of wheat

Abstract

Genetically modified microorganisms (GMMs) with enhanced biocontrol activity are attractive to apply in agriculture. To investigate potential ecological risks of field introduction of GMMs, effects of P. putida strain WCS358r and two genetically modified derivatives of this strain on the indigenous microflora of wheat were investigated. The two modified strains produce phenazine-1-carboxylic acid (P. putida WCS358r::phz) or 2,4-diacetylphloroglucinol (P. putida WCS358r::phl), compounds known to be active against plant pathogenic microorganisms. The strains were applied as seed coating on wheat seeds and field trials were performed in 1999, 2000, 2001, and 2002. Treatments were introduced in the same plots in all years. The effects of the GMMs on the microflora were compared with effects of the parental strain, an untreated control and effects resulting from a common agricultural practice, crop rotation of wheat and potato. Denaturing gradient gel electrophoresis (DGGE) was used to detect changes in community profiles. In the first year, all bacterial treatments transiently affected the composition of the bacterial communities, without differences between the wild-type and the GMMs. After reintroduction, the DAPG-producing GMM influenced the bacterial microflora significantly different from the wild type in the following years. The PCA-producing derivative of WCS358r showed an effect different from the parental strain only after four consecutive introductions. Cropping potato in alternation with wheat had a lasting effect on the bacterial communities compared to continuous wheat cropping. Throughout this long-term field trial, effects of the GMMs never exceeded those of the crop rotation. Specific bacterial species that were affected by the introduction of GMMs were also identified by using additional molecular techniques. PCR-amplified DNA of rhizosphere samples was hybridized to Affymetrix GeneChips® containing 16S rDNA sequences of approximately 9000 operational taxonomic units (OTUs), representative of specific bacterial species. Cluster analysis of DGGE banding patterns and GeneChip data gave similar results, in agreement with the data described in Chapter 3. Neither P. putida WCS358r, nor the GMMs led to loss of any bacterial species identified. Some bacterial species were even stimulated by all bacterial treatments; a few were stimulated only by the DAPG-GMM. Bacteria-specific 16S rDNA amplicons of rhizosphere samples from 2002 treated with the wild type or the DAPG-producing strain were also analyzed using suppression subtractive hybridization (SSH). With this technique it is possible to identify unique sequences in one of two DNA samples. In a test system it was demonstrated that SSH can be applied to rhizosphere samples, since unique DNA sequences derived from digested phage ? X174 were detected after subtraction of a rhizosphere sample spiked with phage DNA against a non-spiked rhizosphere sample. However, when SSH was applied to rhizosphere samples treated with the wild type or the DAPG-producing strain, no obvious differences were detected. Apparently, this technique is not sufficiently optimized to identify shifts in microbial communities in these selected samples. Repeated introduction of genetically modified, antibiotic-producing Pseudomonas bacteria led to only transient effects on the soil microflora that are minor compared to natural variation in soil properties and crop species grown.