Restriction of nodulation by the broad host range Rhizohium tropici strain CIAT899 in wild accessions of Phaseolus vulgaris L.

Transcription

1 New Phytol. (1992), 120, 489^94 Restriction of nodulation by the broad host range Rhizohium tropici strain CIAT899 in wild accessions of Phaseolus vulgaris L. BY J. A. KIPE-NOLT, C. M. MONTEALEGRE M. AND J. TOHME International Centre for Tropical Agriculture (CIAT) A.A. 6713, Cali, (Received 24 June 1991; accepted 8 January 1992) SUMMARY Common bean (Phaseolus vulgaris L.) is nodulated by a heterogeneous group of Rhizobium strains. In a search for host plant restriction of nodulation by some of these strains, 50 wild and cultivated bean accessions were evaluated for the length of time taken to form effective nodules with three diverse bean strains. As bean genotypes vary substantially for this character, preference or restriction was defined as a relative parameter between the three Rhizobium strains for a given bean genotype. Many of the bean genotypes evaluated showed no preference for particular strains of Rhizobium though Rhizobium strain CIAT8002 was often slowest to nodulate. Amongst the wild and landrace bean accessions, there was a tendency for materials of Mesoamerican origin to form active nodules more rapidly with CIAT632 (a Rhizobium leguminosarum bv. phaseoli strain isolated from Guatemala) than with CIAT899 (a broad host-range, R. tropiei from ). In contrast, in those cases in which some preference was observed, wild and landrace bean genotypes from Peru and Argentina (Andean origin) nodulated more rapidly with Rhizobium strain CIAT899. There were exceptions to this pattern among bred lines. Three wild bean accessions: G10002 from Mexico, G23418 from Costa Rica and G21117 from showed strong resistance to nodulation with CIAT899. Formation of effective nodules took more than 10 d longer with CIAT899 than with CIAT632. This strain preference was not altered by changing the time of inoculation between 0 and 8 d after sowing. Within an accession, plant to plant variation was observed in the time taken to nodulate with a given strain oi Rhizobium. However this variation persisted in the progenies of contrasting individual plants of G Key words: Founder effect, nodulation restriction, Phaseolus vulgaris (common bean), Rhizobium leguminosarum bv. phaseoli, Rhizobium tropici. INTRODUCTION Martinez & Selander, 1988). Martinez et al. (1985) ^^^ g^.^^ ^^ ^^ ^^ggg^ defined two broad groups Cotnmon bean (Phaseolus vulgaris L.) forms an Nj- based on host range and copies of the nifh gene: fixing symbiosis with bacteria of the genus Rhi- Type I strains have a narrow host range and multiple zobium. Considerable diversity has been demon- copies of nifh, while Type II strains nodulate a strated in the effectiveness of this symbiosis due in number of different legumes, including Phaseolus part to the great variability within P. vulgaris and and Leucaena, and have a single copy of nifh. The among Rhizobium strains that nodulate bean. Two Type II strains were recently given a new species centres of origin/domestication, Mesoamerica and name, R. tropici (Martinez-Romero et al, 1991). the Andean zone, have been defined for common Strains of Rhizobium that nodulate beans are bean (see Gepts, 1988; Debouck & Tohme, 1989, for present in almost all soils in which this crop is grown reviews). Materials from these centres vary in and nodulation by highly effective inoculant strains morphological characters such as seed size, time to is frequetitly low (Dowling & Broughton, 1986; flowering, growth habits and in molecular markers Thies, Singleton & Bohlool, 1991). Much effort is such as phaseolin, isozymes and mitochondrial now being directed towards identifying highly DNA. competitive strains and towards understanding what Studies in a number of laboratories have demon- makes a Rhizobium strain competitive, strated the extreme heterogeneity of the bacteria that An alternative is to approach the problem from the nodulate common bean (for example Pinero, plant side. Variation in the time required to form

2 490 J. A. Kipe-Nolt, C. M. Montealegre M. andj. Tohme Table 1. Some characteristics of the Rhizobium strains used in this study Strain number CIAT899 CIAT8002 CIAT632 Origin Classification* Effective nodulation of Leucaena sp. Minimum ph tolerance Colony size (mm) on YMA mediumf at 72 h, 28 C Growth at 36 C (Antioquia) II (Narino) I Guatemala I * Brom et al. (1988). t Vincent (1970). nodules by particular strains of Rhizobium may play a key role in determining competitiveness and has been demonstrated among genotypes of Trifolium (Jones & Hardarson, 1979), Glycine max (L.) Merr. (Cregan & Keyser, 1986) and Medicago sativa L. (Hardarson et al, 1982). Bromfield (1984) also demonstrated differences within a single genotype in Rhizobium meliloti strain preference. If simply inherited, Rhizobium strain specificity can be transferred to commercial plant varieties and would help to overcome problems of the lack of 'competitiveness ' of highly effective inoculant strains (Triplett, 1990); indigenous strains or groups of strains could be excluded by the host plant and nodulation by the inoculant Rhizobium favoured. The objective of this study was to identify genotypes of P. vulgaris that exclude or show delayed nodulation by strains representative of different bean Rhizobium groups. Nodulation of a number of wild, landrace and bred bean lines by one Type II and two Type I (one from Mesoamerica and the other from the Andean zone) Rhizobium strains was evaluated. Antioquia,. CIAT632 is a R. leguminosarum bv. phaseoli strain from Guatemala (Mesoamerica). CIAT8002 is also a R. leguminosarum bv. phaseoli strain but quite distinct, and from the Andean Zone (Narino, ). Some morphological and physiological characteristics of these strains are shown in Table 1. Experiments Between 7 and been genotypes were tested against the three Rhizobium strains in each of nine trials. Genotypes showing some strain preference were evaluated several times. In two experiments plants were inoculated at 0, 4 and 8 d after planting in order to determine whether the time of inoculation affected Rhizobium strain preference. Plants were grown in Erlenmeyer flasks in eight of the trials and in plastic growth pouches in one confirmation trial. Although nodulation can be monitored more easily in growth pouches, management of the indeterminant, climbing grow'th habit is more difficult. MATERIALS AND METHODS Germplasm Fifty P. vulgaris accessions from the CIAT (Centro Internacional de Agricultura Tropical) germplasm bank were evaluated for nodulation by three Rhizobium strains. The bean genotypes included wild and landrace accessions from the two centres of origin/domestication and cultivated materials representative of a range of different market classes (Voysest & Dessert, 1988). Information on the wild accessions, including origin, grain size and colour, growth habit and phaseolin type can be found in Toro, Tohme & Debouck (1990). For cultivated accessions see CIAT (1984). Three strains of bean Rhizobium were selected as representatives of different diversity groups. CIAT899 is the type strain for R. tropici and is from Erlenmeyer fiasks Plants were grown in 250 ml Erlenmeyer flasks containing sterile Perlite and 0 ml of ph 6-7 Sandman's nutrient solution (Sandman, 1970). The flasks were covered with dark paper and incubated in a light room with a 12 h photoperiod, C day temperature and C night temperature, and at a mean relative humidity of 70%. Plants were watered with sterile distilled water when necessary. Seeds were surface sterilized by washing for 3 min in 0" o (w/v) sodium hypochlorite, for 3 min in 95 % (v/v) ethanol and then rinsing six times with sterile distilled water. Benlate (1 g 1~^) was added to the final rinse to prevent fungal contamination. Seeds of wild accessions of P. vulgaris were scarified to facilitate germination. One seed was sown per flask and inoculated with 5x10'* Rhizobium from a 3- d-old culture in yeast mannitol broth (Vincent,

3 Restriction of nodulation by Rhizobium in Phaseolus vulgaris ). Three or four replicate flasks were prepared for each of three harvest dates for each genotypestrain combination. These were arranged in a completely randomized design. Non-inoculated controls of each bean genotype were included. Plants were harvested at three different times between 10 and 29 d after inoculation. Owing to the heterogeneity of bean germplasm evaluated, germination and seedling growth rates varied substantially, and harvest dates for different sets of genotypes had to be adjusted accordingly. Roots were evaluated for tbe presence, quantity and colour of nodules. Shoot appearance was also noted. For each genotype and harvest date the numbers of effective nodules formed by the three strains were compared by analysis of variance. Growth pouches Seven bean genotypes were evaluated in one trial using polyethylene growth pouches (16-6 x 20-3 cm) containing a sheet of filter paper and 100 ml of onethird strength Sandman's solution. Six replicate pouches per genotype strain combination were prepared. Three non-inoculated controls per genotype were also included. Seed sterilization, inoculation and incubation conditions were as described for Erlenmeyer flasks. Roots were examined daily for number and colour of nodules. For eacb genotype data for tbe time (days) to appearance of effective nodules by tbe different strains were compared by analysis of variance. RESULTS No Rhizobium strain preferences were detected for many of the bean genotypes evaluated in these trials. Responses of wild and landrace accessions from Mexico, Peru and, that were conflrmed in at least two experiments, are shown in Table 2. The Mexican genotypes G10002, G11032 and G2464 formed effective (red or pink) nodules more quickly with strain CIAT632 while the Peruvian genotypes G21245, G16119 and G73O2 showed some preference for CIAT899. All these Mexican and Peruvian genotypes were slowest to nodulate with CIAT8002 except for the wild Mexican accession G10002, which was much slower with CIAT899. Of the n materials evaluated, G21117 was of most interest as it presented a pattern similar to that of G10002, resistance to nodulation with CIAT899. G8181 showed a slight preference for CIAT632 but no strain preferences were detected for G8185. The response of wild and landrace genotypes from a given country was not uniform but, as can be seen from Table 2, tbere was a tendency for Mexican and n accessions to form effective nodules more rapidly with CIAT632 while those from Peru preferred CIAT899. One wild accession from Argentina (G19889) also showed a slight preference for CIAT899, but no strain differences were detected for the other two Argentinean accessions evaluated in a single experiment (data not shown). Strain preferences of G21117 and G10002 were not affected by the time of inoculation; however plants inoculated after germination, 4 d after planting, nodulated more rapidly with all strains than when seed was inoculated when planted (data not shown). Diversity for both morphological and molecular traits is frequently observed within a P. vulgaris accession, especially in the case of wild materials. G21117, a genotype that had demonstrated resistance to nodulation with CIAT899, was used to test if plant to plant variability in time taken to establish effective nodulation with a given strain was true genetic variability or was simply a consequence of environmental or seedling vigour differences. Seed was harvested from three individual plants, two that had no effective CIAT899 nodules and one that had no effective CIAT632 nodules at 30 d. These three seed lots, as well as a bulk from the germplasm bank, were inoculated witb the three test strains of Rhizobium. There were no differences between seed lots in strain response: all showed delayed nodulation with CIAT899, and in all lots there was plant to plant variation in time to the appearance of nodules and the time at which nodules turned red. Twelve days after inoculation, 5 of 56 plants inoculated with CIAT632 had effective nodules and 38 had white nodules in formation. None of the plants inoculated with CIAT899 showed any reaction. A week later all Table 2. Relative speed of effective nodule formation of wild and landrace Phaseolus vulgaris accessions from Mexico, Peru and Plant genotype Origin Status Relative speed of effective nodule formation* G10002 Mexico 632 > 8002 > 899 G ^899^8002 G2464 Landrace 632 ^ 899 > 8002 G6119 Landrace 632 = 899^8002 G21245 Peru 899 ^ 632 > 8002 G = 632 ^ 8002 G16119 Landrace 899 > 632 > 8002 G7302 Landrace 899 ^ 632 > 8002 G12128 Landrace 899 = 632^8002 G > 8002 > 899 G8181 Lancrace 632 > 8002 = 899 G8185 Landrace 632 = 8002 = 899 * > Significantly {P < 0-05) more effective nodules at one harvest date at least. > Nodulated more rapidly in the majority of repetitions but numbers of effective nodules not significantly different {P < 0 05). = Time to the formation of effective nodules was not different.

4 492 y. A. Kipe-Nolt, C. M. Montealegre M. andy. Tohme Table 3. Time {days) to the appearance of effective nodules on seveti Phaseolus vulgaris accessions iyioculated with three Rhizobium strains Rhizobium strain Genotype Origin Status CIAT899 C1AT632 CIAT8002 G21117 G10002 G23418 G10018 G8185 G21245 G16119 Mexico Costa Rica Mexico Peru Peru Cultivated Cultivated * i * Evaluated to day 29, at which time no plants had effective nodules, and only a few white nodules were present. t For a given genotype, differences between strains of more than 1 d were significant (P < ()-05). ;-- W ;"/ "-'.;-*''.'. " ',7''. ' - -, : - ' '.,.',;','-.'- ' : ', ' *..- '"'_ 4 -V -f - ', - I Figure L Nodulation of the wild bean accession, G21117, with CIAT899 and CIAT632, 25 d after inoculation. CIAT632-inoculated plants, but no CIAT899- inoculated plants, had effective nodules. Twentyeight days after inoculation 13 of 56 plants inoculated with CL^T899 had effective nodules. Results with CLAT8002 were similar to those for CL\T632. In one experiment, P. vulgaris cultivars representative of different grain classes were evaluated for nodulation with the three test strains. As with the wild and landrace genotypes, nodulation with CIAT8002 was delaved. Moreover the tendencv for Mesoamerican materials to form effective nodules more rapidly with CIAT632 than with CIAT899 was maintained. Several Andean accessions formed effective nodules more rapidly with CIAT899 but others showed preference for CIAT632. A final experitnent was conducted with seven accessions that had shown significant strain preferences: G21117 and G8185 from, G10002 and G10018 from Mexico, G21245 and G16119 from Peru, and G23418 from Costa Rica. Plants were

5 Restriction of nodulation by Rhizobium in Phaseolus vulgaris 493 grown in plastic growth pouches so nodulation could be monitored daily. G21117, G10002 and G23418 all showed a delay in effective nodulation with CIAT899 of longer than 10 d when compared to CIAT632 (Table 3). This delay was most striking for G21117 (see Fig. 1). G10018 had effective nodules 6 d later with CIAT899 as compared to the other two strains. In this experiment G16119 did not form effective nodules significantly faster with CIAT899 than with CIAT632, however in previous trials (e.g. Table 2), preference for the R. tropici was significant. No differences between CIAT899 and CIAT632 were detected for G8185 and G21245, although preference for CIAT899 had been observed in one of the two previous experiments with these accessions. DISCUSSION Resistance of the bean plant to indigenous soil Rhizobium strains or groups of strains may offer a means of overcoming the problem of scarce nodulation by highly effective inoculant strains. Fifty wild and cultivated accessions of common bean were evaluated in this study for reaction to three diverse Rhizobium strains. Plants grown in Erlenmeyer flasks were harvested at 3-5 d intervals so only fairly major differences in strain preferences could be detected. Bean genotypes vary substantially in the time taken to form active nodules (CIAT 1985, 1989), so harvest dates were varied and strain preference of a given genotype was chosen as a relative parameter among the three strains in the trials. The study identified three wild bean genotypes that consistently restricted nodulation by CIAT899, a promiscuous R. tropici strain. These accessions were G21117 from, G10002 from Mexico and G23418 from Costa Rica. G21117 nodulated more slowly than the other two genotypes but the delay of about 10 d in nodulation with CIAT899 was similar in all three genotypes (Table 3). Strain preference was maintained when plants were inoculated between 0 and 8 d after planting, suggesting that the trait is not controlled by a transient plant signal. Plant to plant variation in time taken to form effective nodules was observed in all trials, so promising genotypes were evaluated in at least two experiments. This variation is apparently an effect of the micro-environment, rather than true genetic differences, as progeny of individual contrasting plants showed the same variation. Promiscuous Type \\ R. tropici strains were less abundant than the Type I strains in several soils evaluated in, Ecuador (P. Graham, personal communication) and Mexico (Martinez- Romero & Rosenblueth, 1990), suggesting that resistance of plants to them would be less effective in overcoming competition problems than would resistance to CIAT632 Type I strains. However, B. Anyango and K. E. Giller (personal communication) found that over 90 % of the bean Rhizobium isolated from an acid soil in Kenya were R. tropici. Genotypes of soybean that restrict nodulation by strains of the ubiquitous serogroup 123 have been identified and offer considerable potential for this legume (Cregan, Keyser & Sadowsky, 1989). CIAT8002, also a Type I strain, was slow to nodulate almost all of the plant genotypes evaluated and therefore was not very useful in the screening for restriction of 'competitive' Type I strains. G16119 from Peru showed a slight, but consistent, delay in nodulation with CIAT632 as compared to CIAT899. However, mixed strain competition studies would be needed to determine if this preference results in significantly more nodules being formed by a Type II (CIAT899) inoculant. Although no strain preference was detected for many of the bean accessions, there was a tendency for Mesoamerican materials to prefer CIAT632 and Andean accessions to form effective nodules more quickly with CIAT899. The three genotypes that showed restriction of CIAT899 are of Mesoamerican origin. So far no true wild Andean accession has excluded CIAT899. While more bean accessions and Rhizobium strains from the Andean region need to be studied, the present data are consistent with the concept of co-evolution of the Rhizobium and the wild germplasm of beans. Co-evolution of the Rhizobium leguminosarum Pisum sativum L. symbiosis has been demonstrated by Lie et al. (1987), with much more striking strain specificity as compared to bean. The results highlight the importance of wild germplasm as a source of useful traits for improvement of common bean. None of the cultivated bean accessions that were evaluated in these or previous trials in CIAT showed restriction of R. leguminosarum bv. phaseoli strains. There is evidence (Debouck & Tohme, 1989) that only a few populations were domesticated along the range of wild bean distribution. The three wild genotypes that showed restriction of CIAT899 have a phaseolin type not found in the cultivated common bean, suggesting a possible founder effect for that trait. accessions are also useful as geographical markers, allowing a more systematic search for a trait of interest in the cultivated germplasm. Exclusion of CIAT632 would, for example, seem more likely in Andean germplasm. The three bean accessions that showed restriction of CIAT899 are now being tested against other R. tropici strains. Mixed strain competition studies are also needed to confirm that the CIAT899 restriction is not overcome by the presence of other rhizobia. Crosses have been made and are being evaluated to determine if the same gene(s) control restricted nodulation in the three bean accessions. An understanding of the genetic control of CI AT899

6 494 J. A. Kipe-Nolt, C. M. Montealegre M. and J. Tohme restriction will also determine how readily the character can be transferred into cultivated commercial germplasm. ACKNOWLEDGEMENTS We thank G. I. Ocampo for her contribution in the preliminary experiments and G. Segura for technical assistance. REFERENCES BROM, S., MARTINEZ, E., DAVILA, G. & PALACIOS, R. (1988). Narrow and broad host-range symbiotic plasmids of Rhizobium spp. strains that nodulate Phaseolus vulgaris. Applied and Environmental Microbiology 54, BROMFIELD, E. S. P. (1984). Variation in preference for Rhizobium meliloti within and between Medicago sativa cultivars grown in soil. Applied and Environmental Microbiology 48, CIAT (CENTRO INTERNACIONAL DE AGRICULTURA TROPICAL) (1984). Catdlogo de Germoplasma de Frijol Comiin, Phaseolus vulgaris L. CIAT (CENTRO INTERNACIONAL DE AGRICULTURA TROPICAL) (1985). Annual Report of the Bean Program. CIAT (CENTRO INTERNACIONAL DE AGRICULTURA TROPICAL) (1989). Annual Report of the Bean Program. CREGAN, P. B. & KEYSER, H. H. (1986). Host restriction of nodulation by Bradyrhizobium japonieum strain USDA 123 in soybean. Crop Science 26, CREGAN, P. B., KEYSER, H. H. & SADOWSKY, N. J. (1989). Soybean genotype restricting nodulation of previously unrestricted serocluster 123 Bradyrhizobia. Crop Science 29, DEBOUCK, D. G. & TOHME, J. (1989). Implications for bean breeders of studies on the origins of common beans Phaseolus vulgaris. L. In: Current Topics in Breeding of Common Bean. pp Cali,, Bean Program. Centro Internacional de Agricultura Tropical. DowLiNG, D. N. & BROUGHTON, W. J. (1986). Competition for nodulation of legumes. Annual Review of Microbiology 40, GEPTS, P. (1988). A middle American and an Andean common bean gene pool. In: Genetic Resources of Phaseolus Beans, pp Kluwer Academic Publishers. HARDARSON, G., HEICHEL, G. H., BARNES, D. K. & VANCE, C. P. (1982). Rhizobial strain preference of alfalfa populations selected for characteristics associated with N2 fixation. Crop Science 22, JONES, D. G. & HARDARSON, G. (1979). Variation within and between white clover varieties in their preference for strains of Rhizobium trifolii. Annals of Applied Biology 92, LIE, T. A., GOKTAN, D. ENGIN, M., PIJNENBORG, J. & ANLARSAL, E. (1987). Co-evolution of the \egume-rhizobium association. Plant and Soil 100, MARTINEZ, E., PARDO, M., PALACIOS, R. & CEVALLOS, M. A. (1985). Reiteration of nitrogen fixation gene sequences and specificity of Rhizobium in nodulation and nitrogen fixation in Phaseolus vulgaris. Journal of General Microbiology 131, MARTINEZ-ROMERO, E. & ROSENBLUETH, M. (1990). Increased bean (Phaseolus vulgaris L.) nodulation competitiveness of genetically modified Rhizobium strains. Applied and Environmental Microbiology 56, MARTINEZ-ROMERO, E., SEGOVIA, L., MERCANTE, F. M., FRANCO, A. A., GRAHAM, P. & PARDO, M. A. (1991). Rhizobium tropici, a novel species nodulating Phaseolus vulgaris L. beans and Leucaena sp. trees. International Journal of Systematic Bacteriology 41, PiNERO, D., MARTINEZ, E. & SELANDER, R. (1988). Genetic diversity and relationships among isolates of Rhizobium leguminosarum biovar phaseoli. Applied and Environmental Microbiology 54, SANDMAN, W. P. L. (1970). Practical aspects of Rhizobium bacteriology. Rhodesian Ministry of Agriculture. THIES, J. E., SINGLETON, P. W. & BOHLOOL, B. B. (1991). Influence of the size of indigenous rhizobial populations on establishment and symbiotic performance of introduced rhizobia on field-grown legumes. Applied and Environmental Microbiology 57, ToRO, O., TOHME, J. & DEBOUCK, D. G. (1990). bean (Phaseolus vulgaris L.): Description and Distribution. International Board for Genetic Resources (IBPGR) and Centro Internacional de Agricultura Tropical (CIAT), Cali,. TRIPLETT, E. W. (1990). The molecular genetics of nodulation competitiveness in Rhizobium and Bradyrhizobium. Molecular Plant-Microbe Interactions 3, VINCENT, J. M. (1970). A Manual for the Practical Study of Root Nodule Bacteria. IBP Handbook. Blackwell Scientific Publications, Oxford. VOYSEST, O. V. & DESSERT, M. (1988). Market classes of dry edible beans grown commercially in the world. Bean Improvement Cooperative 31,

7