Formation of Novel Polysaccharides by Bradyrhizobium japonicum ...

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Feb. 1992, p. 607-613

Vol. 58, No. 2

0099-2240/92/020607-07$02.00/0 Copyright © 1992, American Society for Microbiology

Formation of Novel Polysaccharides by Bradyrhizobium japonicum Bacteroids in Soybean Nodules STREETER,'*

SALMINEN,1 ROBERT E. WHITMOYER,2 AND RUSSELL W. CARLSON3 Department of Agronomy' and Department of Plant Pathology,2 The Ohio State UniversitylOhio Agricultural Research & Development Center, Wooster, Ohio 446914096, and Complex Carbohydrate Research Center, The University of Georgia, Athens, Georgia 306023

JOHN G.

SEPPO

0.

Received 31 October 1991/Accepted 10 December 1991

Certain strains of Bradyrhizobium japonicum form a previously unknown polysaccharide in the root nodules of soybean plants (Glycine max (L.) Merr.). The polysaccharide accumulates inside of the symbiosome membrane-the plant-derived membrane enclosing the bacteroids. In older nodules (60 days after planting), the polysaccharide occupies most of the symbiosome volume and symbiosomes become enlarged so that there is little host cytoplasm in infected cells. The two different groups of B. japonicum which produce different types of polysaccharide in culture produce polysaccharides of similar composition in nodules. Polysaccharide formed by group I strains (e.g., USDA 5 and USDA 123) is composed of rhamnose, galactose, and 2-O-methylglucuronic acid, while polysaccharide formed by group II strains (e.g., USDA 31 and USDA 39) is composed of rhamnose and 4-O-methylglucuronic acid. That the polysaccharide is a bacterial product is indicated by its composition plus the fact that polysaccharide formation is independent of host genotype but is dependent on the bacterial genotype. Polysaccharide formation in nodules is common among strains in serogroups 123, 127, 129, and 31, with 27 of 39 strains (69%) testing positive. Polysaccharide formation in nodules is uncommon among other B. japonicum serogroups, with only 1 strain in 18 (6%) testing positive.

Bradyrhizobium japonicum is a gram-negative bacterium which fixes dinitrogen in symbiotic association with soybean plants. B. japonicum and other rhizobia are prolific producers of extracellular polysaccharide (EPS) in culture. The composition and structure of many of these EPSs have been established (1, 8, 9, 19, 23, 33). B. japonicum has been divided into two different groups based on several characteristics, including DNA homology, formation of a toxic amino acid analog rhizobitoxine, ability to synthesize nitrogenase ex planta, synthesis of the plant hormone indoleacetic acid, and type of EPS formed in culture (11, 12, 16, 17). Specifically, group I genotypes produce EPS containing glucose, mannose, galactose, galacturonic acid, and 4-0methylgalactose, whereas group II genotypes produce EPS containing rhamnose and 4-O-methylglucuronic acid (12, 16, 19). Recent studies have elucidated a second class of rhizobial polysaccharides, namely, the cyclic P-glucans, which are deposited in the periplasmic space (5, 15, 29). We report here a third class of polysaccharides, namely, those produced by B. japonicum in the soybean nodule. To our knowledge, there has been only one previous study of polysaccharide formation in nodules, and the results showed that the polysaccharide present in nodules was clearly of plant origin (30). The bacterial strain used in the prior study (30) was USDA 110, a strain which we now know does not form nodule polysaccharide (NPS) (results presented here).

Midwest cultivars are, on average, 25% inbred (25), the four cultivars used have only 6% interrelatedness. Cowpea (Vigna unguiculata (L.) Walp.) seeds were supplied by Kent Peters of the Biotechnology Center, Ohio State University. Cultures of B. japonicum were kindly provided by Peter van Berkum at the Nitrogen Fixation & Soybean Genetics Laboratory, Beltsville Agricultural Research Center, Beltsville, Md. Bacteria were grown in liquid medium containing, per liter of water, 0.35 g of K2HPO4, 0.35 g of KH2PO4, 0.20 g of MgSO4 7H20, 0.10 g of NaCl, 0.50 g of yeast extract, 2 g of Na gluconate, and 10 g of mannitol (pH 6.8 to 7.0). Plants were grown in autoclaved silica sand in a greenhouse, irrigated with a complete nutrient solution lacking combined nitrogen, and provided with supplementary light during the winter months (26). Liquid cultures of B. japonicum were diluted 1:1 with water and applied to seeds on the planting date, providing about 108 cells per seed. Prior work with nitrate reductase mutants of B. japonicum has shown that this system results in all nodules in a pot being formed by the inoculant strain (27). For electron microscopy, whole nodules were fixed in a solution of 3% glutaraldehyde, 2% paraformaldehyde, and 1.5% acrolein in a 0.2 M phosphate buffer at pH 7.3. The nodules were washed with buffer and postfixed in 2% osmium tetroxide in the same buffer for 3 h, washed, and stained overnight in 1% uranyl acetate. Specimens were dehydrated in an ethanol-acetone series and embedded in Spurr plastic (24). The plastic was polymerized at 70°C. Blocks were sectioned on a Sorvall MT-2B ultramicrotome and were picked up on Formvar-coated grids. Sections were poststained with 0.5% uranyl acetate and 0.1% lead citrate and viewed with a Philips 201 electron microscope. Routine analysis of NPS. Fresh nodules, generally 4 to 10 g (fresh weight), were ground with a mortar and pestle in 50%o ethanol by using a liquid volume of two times the fresh weight. (This ethanol concentration was chosen because it will not precipitate any polysaccharide; lower ethanol con-

MATERIALS AND METHODS General methods. Soybean (Glycine max (L.) Merr.) seeds were kindly supplied by Brian McBlain of the Agronomy Department, Ohio State University. Cultivar Beeson 80 was generally used. For the comparison of soybean genotypes, cultivars as unrelated as possible were chosen. Although *

Corresponding author. 607

608

STREETER ET AL.

centrations make it difficult to achieve a high ethanol concentration later in the procedure.) Assuming that nodules contain 80% water (unpublished data), this gives an approximate ethanol concentration of 36% (vol/vol) in the crude homogenate. Homogenate was transferred to 50-ml centrifuge tubes with 36% ethanol for rinses of the mortar. Samples were centrifuged at 48,000 x g for 15 min at 21°C. The supernatant was poured into a graduated cylinder, the pellet was resuspended in 2x volume of 36% ethanol, and this mixture was centrifuged a second time. The supernatant was added to the first supernatant, and the total volume was made up to five times the initial nodule fresh weight with 36% ethanol. This sample was thoroughly mixed and used for viscosity analysis at room temperature. Two viscometers were available to us, but neither would measure a sufficiently wide range of viscosities. We devised a viscometer consisting of a 1.1-m length of glass tubing with an inside diameter of 5.0 mm. A 1-m section was marked, one end was plugged, and the sample was loaded into the tube. BB shot 4 mm in diameter and weighing 346.6 ± 0.8 mg (standard error; n = 14) were passed through the sample, and the time required to traverse 1 m was recorded. Three or four passes were made, and the mean time in seconds was taken as the result; standard error values were generally 30-day-old nodules formed by B. japonicum USDA 123 and USDA 31. These two strains are in DNA homology groups I and II, respectively, and their NPSs will be referred to as type I NPS and type II NPS. Ion-exchange chromatography of extracts of NPS-containing nodules (e.g., USDA 123) showed a very large uronic acid-containing peak which was absent in nodules which lack NPS (Fig. 1). When this polysaccharide was rechromatographed with a different salt gradient, a single symmetrical peak of uronic acid was found. This result plus the absence of minor components in the analysis of composition indicated that the material analyzed as NPS was of high purity (data not shown). For the type II NPS, carboxyl reduction resulted in the appearance of a peak whose mass spectrum showed that it was derived from a 4-O-methyluronic acid. Its retention was identical to that of the alditol acetate of 4-O-methylglucose. The only other component of this sample was rhamnose, identified on the basis of its retention time and on GC-mass

BACTERIAL POLYSACCHARIDES IN LEGUME NODULES

VOL. 58, 1992

609

TABLE 2. Composition of polysaccharide formed by USDA 123 bacteroids in nodules of four diverse genotypes of soybean plants

250

0 1-

Carbohydrate

Compositiona (%) from soybean cultivar: Resnik

Ripley

Sloan

Beeson 80

200

51 54 61 55 Rhamnose 29 26 18 27 Galactose 12 11 12 13 2-0-Methylglucuronic acid a Composition was determined by using only partially purified polysaccha-

NaCI

1300

150

FPROTEIN

~200

00

-

0#j1OO:W 50

-

0

75 55 65 25 35 45 FRACTION NIJABER FIG. 1. Purification of NPS with a DEAE-agarose column. Protein concentration (A280 [squares]), uronic acid concentration (circles), and NaCl concentration (triangles) in 4-ml fractions are shown. Protein patterns differed among runs depending on the quality of the starting sample. The shape of the major uronic acid peak varied slightly among runs but was always centered on fraction 50; generally, fractions 44 through 56 were kept for analysis of composition. Some uronic acid-containing polysaccharide was present in nodules lacking NPS (e.g., formed by USDA 110), but these polysaccharides eluted almost entirely prior to fraction 45 and were assumed to be plant polysaccharides. Results shown are for the elution of USDA 123 polysaccharide, but the pattern of uronic acid elution was very similar for the polysaccharide formed by USDA 31.

15

spectrometry (Table 1). For the type I NPS, carboxyl reduction resulted in the appearance of a peak having a mass spectrum of the alditol acetate of a 2-0-methylhexose which contained two deuterium atoms at C-6 and one deuterium atom at C-1. The two C-6 deuteriums indicate that it was TABLE 1. Composition of polysaccharide produced by B. japonicum bacteroids in soybean nodules % of polysaccharide dry wt Carbohydrate

USDA 123

USDA 31

(group I) in:

(group II) in:

Culturea

Glucose Mannose Galacturonic acid Galactose

4-0-Methylgalactose Rhamnose 4-0-Methylglucuronic acid 2-0-Methylglucuronic acid

41 19 13 6.5 13

Nodule

Cultureb

Nodule

70 30

64

12

37

25

9.1c

12. b

See references

c

Estimated concentration is probably low because of low detector response

(see text).

9

and

16.

ride; small amounts (