Effect of Calcium on the Phosphorus Nutrition of Rhizobium meliloti

Effects of calcium at 300 and 1500 μM on P nutrition were assessed in eight strains of Rhizobium meliloti in defined liquid medium. Evaluations included: P storage from "luxury" external concentration (1000 μM P); utilization of stored P after transfer to unreplenished low-P medium (0.06 μM); and growth at low concentrations of P buffered at 5, 0.5, and 0.06 μM with an iron oxide dialysis system. The strains stored P in luxury medium, but unlike other rhizobia, they needed high Ca to utilize the stored P. They either grew or died following transfer to low-P medium, depending on the Ca concentration and the Ca concentration at which they had grown previously. Ability to grow in media buffered at low P concentrations also contrasted with that of other rhizobia, in two respects: no strain of R. meliloti grew at 0.06 μM P, regardless of Ca concentration; and some strains needed high Ca to grow at 0.5 and 5 μM P. Two isolates from Medicago rugosa and Melilotus indica were less Ca-demanding than six isolates from Medicago sativa. Previous reports that R. meliloti has low calcium requirements may be correct only for the luxury P levels that are conventional in defined media. Our evidence for high Ca requirement at realistic P concentrations agrees with data from soil experiments. Additional Index Words. symbiotic N fixation, Fe oxide, goethite, P-buffering, uptake, utilization. Beck, D.P. and D.N. Munns. 1985. Effect of calcium on the phosphorus nutrition of Rhizobium meliloti. Soil Sci. Soc. Am. J. 49:334337. __________________________________________ MINERAL NUTRITION STUDIES of plants and soil microbes in simplified artificial media become especially valuable when their results appear to contradict observations made in soils. Study of such discrepancies can uncover a previously unsuspected role for soil factors incorrectly reproduced in the artificial system. This paper reports such a case involving Ca and P nutrition of microbes important in nitrogen cycling, members of the genus Rhizobium. There has long been evidence suggesting that in acid soils Ca deficiency can limit growth of Rhizobium, especially R. meliloti (9, 10,11). Yet all controlled studies of Ca nutrition of rhizobia in artificial media, even with R. meliloti, indicated requirements so low that Ca would seem unlikely ever to limit the organisms' growth in soil (2,14). The artificial media, as is conventional, contained millimolar concentrations of phosphate, hundreds of times higher than the concentrations actually encountered in soil solutions. Since ___________ ' Contribution from Dep. of Land, Air and Water Resources, Univ. of California, Davis CA 95616. Supported in part by a grant from NSF RANN.Received 8 July 1983. Approved 31 Oct. 1984. 2 Respectively, Research Assistant (now Microbiologist, Univ. of Hawaii NifTAL Project, Maui, HI 96779), and Professor of Soil Science, UC Davis. Ca and P can interact in their transport into cells and organelles (5), we postulated a similar interaction in R. meliloti, such that high Ca becomes necessary when orthophosphate is lowered to concentrations relevant to the soil environment. This paper reports experiments testing the postulate, using a recently published procedure (1,6) for controlling phosphate concentrations in the micromolar range. MATERIALS AND METHODS Seven strains of Rhizobium meliloti were used: five isolates from Medicago sativa (Nitragin Co. strains 102F70 and 102F28 and United States Dep. of Agric. strains 1021a, 1029, and 1031), one isolate from Medicago rugosa (Nitragin 102H1), and one isolate from Melilotus indica (Nitragin 104B4). All the strains effectively nodulated Medicago saliva var. Moapa in aseptic agar tube culture. Cultures were maintained under refrigeration on yeast-agar slants similar in composition to our liquid luxury-P medium (below). Arabinose-galactose (0.3% each) as an energy source was chosen because it did not interfere with the phosphomolybdateblue determination of P. The liquid media also contained 1.1 g/L sodium glutamate, 0.1 mg/L biotin, and inorganic nutrients at the following micromolar concentrations: MgS04 300; FeEDTA 50; MnS04 2; ZnS04 1; CUS04 0.5; Na2MoO4 0.1; CoC12 0.02. Calcium was supplied as chloride at two concentrations: 300 μM (low-Ca) and either 1500 μM or 3000 μM(high-Ca). The pH was adjusted to 5.5 with HCl before autoclaving. To maintain buffered concentrations of phosphate at low levels in solution, an "Fe oxide dialysis" system was used (6). Concentration of P in solution depended on the amount added with the iron oxide (powdered limonite from Ward's Natural Science Establishment, Rochester, NY). The oxide was phosphated in 150 g lots by shaking it for 6 d in 1.5 L of 10 mM CaC12 with the appropriate amount of KH2PO4, followed by filtration and air-drying. Each culture received a slurry of 3 g oxide and 4 mL water contained in a section of dialysis tubing knotted at both ends, added to 37 mL of medium in a 125 mL Erlenmeyer, culture flask, autoclaved 30 min, and left 3 d at 25 to 27°C to equilibrate. The desorption isotherm was then determined by analysis of the liquid media. The luxury-P control contained KH2P04-K2HP04 at a P concentration of 1000 μM. The low-P treatment, with no P added, and the oxide cultures, received 250 μM K2SO4 to ensure an adequate supply of K. All liquid cultures were incubated at 25°C on an orbital shaker. Three sets of experiments were done: l. Phosphorus storage in cells was measured at both a luxury level of P (1000 μM) and a high level representing that found in solution in a fertile soil (5 μM). The 5 μM level was maintained using the oxide dialysis system. At each level of P, two replicates each of two levels of Ca (300 and 1500 μM) were inoculated and grown to 10 to 10 cells per mL, then centrifuged at relative centrifugal force 10 000Xg for 20 min, resustions in high-Ca low-P medium, and growth following storage correlated fairly well with amount of P stored R = 0.85 (Table 1). In the low-Ca low-P cultures, however, the number of viable cells decreased with time. The level of calcium at which the inocula were grown had a significant effect on subsequent growth to low P (Fig. 2 and 3). High calcium inoculum reached higher populations in high-Ca low-P medium, and remained viable longer in low-Ca low-P. Multiplication or survival varied with strain as well as calcium regime and amount of P stored. Strain 102H1, isolated from Medicago rugosa, behaved similarly to the alfalfa strains in runout culture, while 104B4, isolated from Melilotus indica, would not grow at all in the low P media, regardless of cal BECK & MUNNS: EFFECT OF CALCIUM ON PHOSPHORUS NUTRITION OF RHIZOBIUM MELILOTI 335 pended in 10 mM CaC12 and recentrifuged. The pellet was dried overnight at 60°C, then analyzed after Kjeldahl digestion by a phosphomolybdate-blue procedure (13). 2. Growth was measured following transfer from luxury P (1000 μM) to unbuffered low-P (0.06 μM) medium. High-P cells grown at two levels of Ca were used to inoculate low-P medium containing the two levels of calcium in a factorial experimental design. The inocula were added to 37 mL media to give an initial density of about 10° cells per ml. Cell growth was followed by conventional drop counts on yeast arabinose-galactose plates. At each time (see Fig. 2 and 3), six replicate 41μL drops were counted from each triplicate culture. Analysis of variance was done on logo transforms of the counts. 3. Growth was measured at buffered concentrations of phosphate likely to be encountered by bacteria in the soil, with the high level representing a P-fertile soil (5 μM), medium a P-deficient soil (0.5 μM), and low a P-depleted rhizosphere (0.06 μM). The inocula were from cultures grown in high-Ca, low-P medium to eliminate P storage effects, and diluted to provide 10 to 10cells per mL at the beginning of the experiment. Growth of strains 102F28, 102111, 104B4 and 1021a was monitored by drop counts at intervals for 7 d after inoculation into media at two levels of Ca and the three buffered levels of P, plus a 2000 μM P luxury control. The least significant difference (LSD, p=0.05) was calculated from the analysis of variance on log10 transforms of the counts. The culture medium was analyzed at 3 and 5 d for P to check the P-buffering performance of the oxide, and pH was checked near the end of each