Glucomannan-mediated attachment of Rhizobium leguminosarum to pea root hairs is required for competitive nodule infection

Abstract

The Rhizobium leguminosarum biovar viciae genome contains several genes predicted to determine surface polysaccharides. Mutants predicted to affect the initial steps of polysaccharide synthesis were identified and characterized. In addition to the known cellulose (cel) and acidic exopolysaccharide (EPS) (pss) genes, we mutated three other loci; one of these loci (gmsA) determines glucomannan synthesis and one (gelA) determines a gel-forming polysaccharide, but the role of the other locus (an exoY-like gene) was not identified. Mutants were tested for attachment and biofilm formation in vitro and on root hairs; the mutant lacking the EPS was defective for both of these characteristics, but mutation of gelA or the exoY-like gene had no effect on either type of attachment. The cellulose (celA) mutant attached and formed normal biofilms in vitro, but it did not form a biofilm on root hairs, although attachment did occur. The cellulose-dependent biofilm on root hairs appears not to be critical for nodulation, because the celA mutant competed with the wild-type for nodule infection. The glucomannan (gmsA) mutant attached and formed normal biofilms in vitro, but it was defective for attachment and biofilm formation on root hairs. Although this mutant formed nodules on peas, it was very strongly outcompeted by the wild type in mixed inoculations, showing that glucomannan is critical for competitive nodulation. The polysaccharide synthesis genes around gmsA are highly conserved among other rhizobia and agrobacteria but are absent from closely related bacteria (such as Brucella spp.) that are not normally plant associated, suggesting that these genes may play a wide role in bacterium-plant interactions.

FIG. 1.

Bacterial colonies grown on TY agar containing Congo red to stain for cellulose. (A to D) Strain 3841 (wild type) (A) and its derivatives A1060 (celA) (B), A1073 (pssA1) (C), and A1104 (pssA1 celA) (D). (E and F) Colonies of A168 (pssA1) (E) and A1077 (pssA1) (F), which are derivatives of A34.

FIG. 2.

Effects of mutations affecting surface polysaccharides on formation of biofilms in static culture. (A) Strain 3841 (wild type) formed a distinctive biofilm after 4 days of growth in static culture in Y medium, with microcolonies separated by voids. (B) Like panel A but with ×8 digital zoom, showing that cells within the microcolonies are in a close-packed hexagonal array. (C to F) The pssA mutant (C) formed a flat loosely attached lawn, whereas the celA (A1060) (D), gelA (A1090) (E), and (F) gmsA (A1045) (F) mutants formed biofilms indistinguishable from those formed by the wild-type strain.

FIG. 3.

Role of surface polysaccharides in root hair attachment. Strain 3841 (wild type) forms a root hair cap when it is incubated with vetch root hairs at pH 6.5 (A) or pH 7.5 (B). The glucomannan mutant A1045 (gmsA) was unable to attach or form a root hair cap at pH 6.5 (C) but formed root hair caps at pH 7.5 (D). Individual cells of the cellulose mutant A1060 (celA) attached to root hairs, but no cap was formed at pH 6.5 (E) or pH 7.5 (not shown). Normal attachment and cap formation was seen with A1090 (gelA) at pH 6.5 (F) and pH 7.5 (not shown). The acidic EPS mutant A1073 (pssA) did not attach to root hairs, but it did attach at root epidermal cell boundaries (G), as did wild-type strain 3841 (H).

FIG. 4.

Nodulation competitiveness of surface polysaccharide mutants. Each of the mutants shown was coinoculated onto peas with equal numbers of wild-type strain 3841 cells. The nodule occupancy of each mutant is expressed as a percentage of the bacteria recovered from individual nodules based on scoring using antibiotic resistance (streptomycin for the wild type, gentamicin for A1208 [gmsA] and A1209 [exoY], spectinomycin for A1247 [gelA], and kanamycin for A1248 [celA]). The gmsA mutant is significantly less competitive than the wild type, but the other mutants are not significantly different from the wild type, based on a chi-square test.