Maize seeds were inoculated with a commercial inoculant containing 1.3 × 107 Azospirillum lipoferum CRT1 cells. After 24 or 48 h, bacteria were washed from the seed surface. Washed and unwashed seeds were then planted in pots containing perlite and grown for 28 days under greenhouse conditions. Whatever the density of Azospirillum at planting, the number of these bacteria at the end of the experiment was similar (1.9-8.0 × 107 bacteria·plant-1). However, comparison of root surface areas of the plants were different depending on the period of contact between seeds and the density of the inoculum. Twenty-four hours of contact was not sufficient to increase root growth surface areas. Contact for 48 h permitted us to obtain root surface areas comparable with those measured after a continuous contact. These results showed that in order to promote maize root surface areas, an optimal density of Azospirillum is not required during the whole cultural cycle. This optimal density is indispensable only up to the emergence of the radicle.Key words: Azospirillum, maize, inoculation, PGPR.

Azospirillum lipoferum CRT1 has been isolated from a maize rhizosphere by Fages and Mulard in 1988. Pot experiments carried out on maize showed a strong beneficial effect on root development and shoot dry weight. These results led them to test the efficiency of the strain in field experiments. Fages (1994) reported results from different field trials. The main result was a better absorption of nitrogen fertilizer. Moreover, grain yield improvement was frequently observed (particularly for suboptimal nitrogen doses). It was concluded that these improvements were probably due to a better root development. Strain CRT1 is now integrated in a peat-based inoculant for maize, named Azogreen-m® and marketed by LIPHA group.

In order to develop a reliable and specific tool for the detection of Azospirillum lipoferum CRT1, randomly cloned DNA fragments from this strain were used as hybridization probes to differentiate A. lipoferum CRT1 from 29 closely related Azospirillum strains. Two cloned fragments hybridizing only with DNA from A. lipoferum CRT1 (CRT1-5 and CRT1-7) were considered as specific probes of this strain. CRT1-7 fragment (1.4 kb) was further tested for purity control of the inoculant Azogreen-m by colony hybridization. The sequence of the CRT1-7 fragment has been determined and compared with those present in databases: no significant similarity with other sequences was detected. This probe permitted us to count specifically A. lipoferum CRT1 cells on maize roots during a field trial. During the first two weeks, A. lipoferum CRT1 remained at 107 CFU plant−1. Afterwards, bacterial concentration sharply decreased. We could not detect any CRT1 cells on maize roots 28 days after sowing. Concurrently, three plant parameters were estimated (plant height, primary root length and root fresh weight). The results showed that A. lipoferum CRT1 growth promotion effect began early on (from day 14) in plant development and increased in spite of a rapid decrease of bacterial density.

Striga spp. are obligate parasitic weeds of tropical cereals and generally have the same host range as rhizospheric bacteria of the genus Azospirillum. Four strains of Azospirillum brasilense, isolated from soil where sorghum is grown, have been tested for their effect on germination of Striga hermonthica seeds and on cereal (Sorghum vulgare) growth. Two out of four strains assayed significantly inhibited germination of the parasite. Moreover one of the two strains showed a plant growth promoting (PGPR) effect.

Activity of bacterial Azospirillum lipoferum laccase on phenolic derivatives was studied by spectrophotometry, HPLC and GC/MS. Phenolic compounds of the syringic type (aldehyde, acid or acetophenone) were transformed into 2,6-dimethoxy-1,4-benzoquinone (2,6-DMBQ). Comparison has been made with the fungal Pyricularia oryzae laccase. Transformation of other phenolic acid derivatives, related to lignin metabolism, have been studied with both bacterial and fungal enzymes, using spectrophotometric analysis.

Azospirillum lipoferum 4B and non-motile A. lipoferum 4T have been simultaneously isolated from rice rhizosphere at the same frequency. A. lipoferum 4T showed stable morphological and metabolic traits which are atypical for A. lipoferum species such as lack of motility, carbohydrate metabolism and laccase activity. Inoculation experiments showed that A. lipoferum 4T, but not A. lipoferum 4B, needed rice roots to stabilize in sterile soil. Both strains were able to colonize efficiently rice roots (108 cfu g−1 fresh roots) but motile form 4B remained dominant. In spite of their phenotypical differences, A. lipoferum 4B and 4T co-existed without exclusion in sterile soil (planted or not) and rice rhizosphere. Inoculation of rice roots with A. lipoferum 4B showed that rice rhizosphere enhanced the frequency of appearance of stable non-motile forms (40%). This percentage was weaker in plantlet growth medium (4%). However, these non-motile bacteria kept the same biochemical traits than the motile parental strain 4B (carbohydrates metabolism, laccase activity).

Partial sequences of the 16S rRNA molecules of nine strains belonging to four Azospirillum species were used to design species-specific oligonucleotide probes. Azospirillum strains sequences were analyzed and three homologous fragments containing 16 nucleotides were determined. These three probes were found to be characteristic of A. lipoferum (Al), A. irakense (Ai), and A. brasilense/amazonense species (Aba) and of few nontarget organisms. The specificity of these three probes was tested both against sequences in the GenBank data base and in numerous colony hybridization experiments. As a few non-target organisms hyridized with the different Azospirillum probes, the use of these probes in bulk soil hybridization is not permitted. However, their use together with specific isolation techniques is validated.

A mixture of 5-(12-heptadecenyl)-resorcinol as the main compound with four other alkyl resorcinols was isolated from rice root exudate. The chemical characterization was done by HPLC/MS coupling, 1H and 13C NMR.

In non-motile Azospirillum lipoferum 4T and A. lipoferum 4Bp, a laccase activity was characterized by several substrates and inhibitors. This laccase activity was required for the production of a brown-dark pigment, as melanin. Following random Tn5 mutagenesis in A. lipoferum 4T, several mutants affected in melanization and laccase activity were obtained. These mutations were located on the chromosome or on the cryptic 300 MDa plasmid of A. lipoferum 4T.