Proc. Nad. Acad. Sci. USA Vol. 82, pp. 7379-7383, November 1985 Genetics

Isolation and characterization of modulation genes from Bradyrhizobium sp. (Vigna) strain IRc 78 (cosmid cloning/TnS mutagenesis/functional complementation) JOHN D. NOTI, BRIGITTA DUDAS* AND ALADAR A. SZALAYt Boyce Thompson Institute for Plant Research at Cornell University, Tower Road, Ithaca, NY 14853 Communicated by H. E. Umbarger, July 8, 1985

ABSTRACT An 11.76-kilobase-pair (kb) segment of DNA slow growers. However, very little has been reported about from Bradyrhizobium sp. (Vigna) strain IRc 78 that hybridizes the structural and functional organization of nod genes in to nodulation genes ofRhizobium meliloti strain 41 was isolated. slow-growing strains. Hybridization of the 11.76-kb DNA fragment to DNA from Whereas the fast-growing Rhizobium species are generally other Bradyrhizobium species revealed a high degree of se- host-specific, the slow-growing Bradyrhizobium species are quence conservation in this region. Transfer of the 11.76-kb promiscuous and can form effective symbioses with a wide segment to nodulation-defective (Nod-) mutants of R. meliloi range of legume hosts (1). Strains of Bradyrhizobium that restored their ability to induce nodules on Medicago saiva infect legumes either by invasion of root hairs or by (alfalfa). Mutants of strain IRc 78 generated by TnS mutagen- apoplastic movement into the root cortex are of particular esis of the 11.76-kb segment fell into three classes according to interest for determining whether the same nodulation genes their symbiotic reaction with Vigna unguiculata (cowpea). are involved in both infection processes. Class I mutants ofstrain IRc 78 were unable to induce root-hair In this paper we describe the identification and cloning of curling or to nodulate; class II induced small, ineffective a region ofDNA from Bradyrhizobium sp. (Vigna) strain IRc nodules; and class III showed delayed and decreased nodula- 78 that can complement the Nod- phenotype of mutant R. tion with reduction in amount of nitrogen fixed. Furthermore, meliloti strains on Medicago sativa (alfalfa). We report that in contrast to the wild-type strain, class I mutants could not TnS mutagenesis ofthis cloned region resulted in the isolation induce nodules on Glycine mar (soybean), Cajanus cajan of mutants affected in either the ability to nodulate or the (pigeon pea), or Arachis hypogaea (peanut). This finding ability to fix nitrogen. We also present evidence that nodula- suggests a common function of the 11.76-kb region in the tion genes essential for the typical root-hair infection mode infection of host plants by Bradyrhizobium either through root are also required for crack entry. This finding suggests that hairs or by "crack entry." nodulation functions are also conserved between different infection mechanisms. The formation of nitrogen-fixing root nodules on legumes by Rhizobium, in general terms, requires the following steps: (i) MATERIALS AND METHODS recognition and invasion ofthe root, (ii) nodule development and differentiation, (iii) multiplication of Rhizobium and Strains and Plasmids. R. meliloti strains 1021 (Fix', str) bacteroid formation, and (iv) expression of nitrogenase (1). (10) and 1027 (Nod-, IsRml::nod str neo) (10) were obtained Infection of most legume species proceeds after the attach- from F. Ausubel. R. meliloti strains 41 (Fix') (6) and ZB157 ment ofRhizobiumn to root hairs, which subsequently curl and (Nod-, Anod) (6) were obtained from A. Kondorosi. B. invaginate to form infection threads. The infection thread japonicum strains 1110, USDA 123, and RCR 3407 (Fix') and grows into the root cortex, branches, and releases the Bradyrhizobium sp. (Vigna) strain IRc 78 (Fix') were pro- bacteria into the cortical cells (2). In Arachis hypogaea vided by A. Eaglesham (Boyce Thompson Institute). The (peanut), however, the bacteria enter the roots in a distinctly Escherichia coli strains used were HB101 (Smr, hsdM- different manner. No infection threads have been found in the hsdR- recA pro leuB6 thi) (11), JA221 (hsdM' hsdR- recA root hairs or the nodules of this species. The bacteria initiate leuB6) (12), NS433(X Eam4 b2 red3 cIts857 Sam7) (13), NS428 infection at the bases of root hairs in the axils of emerging (A Aamll b2 red3 cIts857 Sam7) (13), S605 (thi thr leu supF lateral roots through spaces between epidermal cells and lac met: :TnS, Km9 (14), and SM10 (recA thi thr leu supf lac, subsequently proliferate in intercellular spaces before invad- RP-4-2-Tc::Mu, TRA+) (14). Plasmids pSUP106 (Tcr Cmr ing the cortical cells ("crack entry") (3). Mob+) (14) and pSUP202 (Apr Tcr Cmr) (14) were provided Although the nodulation process is characterized by a high by A. Puhler. Plasmid pEK5121 (6) was provided by A. degree of Rhizobium/legume specificity, nodulation (nod) Kondorosi. genes in several fast-growing Rhizobium species are con- DNA Isolations. Plasmid DNA from E. coli was prepared served in function (4-6) and DNA homology (7, 8). These either on a large scale as described by Hadley and Szalay (15) genes, designated nodABC ("common" nod genes), are or on a small scale as described by Birnboim and Doly (16). required for root-hair curling and are clustered on symbiotic Total DNA from Bradyrhizobium cells was prepared accord- (sym) plasmids linked to genes required for nitrogen fixation ing to Jagadish and Szalay (17). Total DNA for construction (4-6). When DNA fragments from a slow grower, of the gene library was isolated as described by Rosenberg et Bradyrhizobium sp. (parasponia), were introduced into a al. (18). Nod& R. meliloti strain, nodulation ability was restored (9). Construction of Strain IRc 78 Gene Library. Total DNA, This finding indicates that the functional conservation of partially digested with EcoRI, was fractionated by centrifu- some nodgenes in fast-growing strains may also extend to the Abbreviations: kb, kilobase pair(s); Fix, nitrogen fixation; Nod, nodulation. The publication costs of this article were defrayed in part by page charge *Present address: Attila Jozsef University, Department of Genetics, payment. This article must therefore be hereby marked "advertisement" Szeged, Hungary. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 7379 Downloaded by guest on September 25, 2021 7380 Genetics: Nod et al. Proc. Natl. Acad. Sci. USA 82 (1985) gation in 10-40% sucrose gradients to enrich for fractions in as described (17). Activities from M. sativa were determined the size range 20-40 kilobase pairs (kb), as described by from the combined root systems in each DiSPo bottle (5-6 Morris et al. (19). Plasmid pSUP106 was ligated to fraction- roots per jar). ated strain IRc 78 DNA and packaged in vitro (19). E. coli Root-Hair Curling Assay. Root-hair curling assays were transductants containing plasmids were screened by colony done in two ways. In the first method, root-hair curling was hybridization as described by Hanahan and Meselson (20). assayed on seedlings from surface-sterilized seeds germinat- Hybridization Conditions. DNA fragments were twice pu- ed between moist filter papers. The seedlings were inoculated rified from agarose gels (21), labeled with [a-32P]dTTP by after 2 days imbibition and examined by light microscopy 2 nick-translation (22), and hybridized under either high- days later. In the second method, root hairs of seedlings stringency conditions [0.3 M NaCl/0.03 M sodium citrate, pH grown in sand were examined 7, 9, 11, 13, and 16 days after 7/50% (vol/vol) formamide/0.2% Ficoll/0.2% polyvinylpyr- inoculation. rolidone/0.2% bovine serum albumin/0.1% NaDodSO4/ salmon sperm DNA (100 g/ml)] or low-stringency condi- tions (3-fold higher NaCl and sodium citrate concentrations, RESULTS formamide reduced to 33%) at 370C for 18-24 hr. Isolation of a Putative nod Region from Strain IRc 78. An Conjugation Experiments. Plasmid DNA was transferred 8.5-kb EcoRI fragment isolated from pEK5121 (6), a cosmid from E. coli SM10 cells to Rhizobium orBradyrhizobium cells clone containing the nod genes from R. meliloti 41, was as described by Jagadish and Szalay (17). Transconjugants hybridized with EcoRI-digested total DNA from Brady- were selected on plates supplemented with the appropriate rhizobium sp. (Vigna) strain IRc 78 and B. japonicum strains antibiotics as described (17, 23). I110, USDA 123, and RCR 3407 (Fig. lA). Prolonged expo- Site-Directed TnS Mutagenesis. Isolated fragments ofstrain sure (96 hr) of the autoradiogram revealed several major IRc 78 DNA were ligated into pSUP202 made linear with bands and a number of less intense bands of hybridization EcoRI. TnS mutagenesis was performed in E. coli S605 present in the total DNA digests of each strain. When the according to Jagadish and Szalay (17), and the mutated same digests were hybridized to a 3.5-kb BamHI-EcoRI fragments were transferred from E. coli SM10 to strain IRc fragment (6, 7) isolated from the 8.5-kb EcoRI fragment, only 78 and exchanged for the corresponding wild-type DNA by the major bands (arrowheads, Fig. 1A) remained. The 3.5-kb double-reciprocal crossover as described (17). BamHI-EcoRl subfragment contains the nodA, -B, and -C Plant Growth and Isolation of Bacteria from Nodules. The genes (7). Two clones, pJN78-19 and pJN78-24, were isolated procedures described by Jagadish and Szalay (17) were from a cosmid library of strain IRc 78 total DNA probed by followed for all plant-growth experiments with modification colony hybridization with the 3.5-kb BamHI-EcoRI frag- for the growth ofM. sativa. M. sativa seeds were germinated ment. Hybridization ofthe 3.5-kb BamHI-EcoRI fragment to for 5-7 days at 22TC in 6-ounce DiSPo bottles (Scientific EcoRI-digested pJN78-19 and pJN78-24 showed the presence Products) (5-6 seeds per bottle) containing vermiculite and of two bands, at 7.08 kb and 4.68 kb (Fig. 1B), which were 100 ml of nitrogen-free nutrient solution, inoculated, and also present in the hybridization with total DNA from strain covered with paraffin-treated sand. The level of inoculum IRc 78 (Fig. LA, lanes 4 and 8). (1.5 x 104 cells per plant), determined empirically, did not The 4.68-kb putative nod fragment from strain IRc 78 result in mixed infections in nodules. Nutrient solution was hybridized with both the 5.0-kb and 3.5-kb BamHI-EcoRI added after 21 days and, thereafter, as needed. Harvesting fragments (the products of the 8.5-kb nod fragment) in occurred 35 days after inoculation. Bacteria were isolated pEK5121. The 7.08-kb putative nodfragment from strain IRc from nodules as described (17). Three or four nodules per 78 hybridized only to the 3.5-kb BamHI-EcoRP fragment. plant and six isolates per nodule were analyzed unless stated Furthermore, the 7.08-kb putative nod fragment also is otherwise. homologous to an additional uncharacterized fragment (1.4 Acetylene Reduction Determination. Acetylene-reduction kb) in pEK5121 (Fig. 2). The plasmid pEK5121 contains %42 activities from Vigna unguiculata (cowpea) were determined kb of DNA from the R. meliloti 41 sym plasmid and includes by assaying the root systems in 500-mljars (one root perjar) the common nod genes, host-specificity genes, and some fix

A kb , n 5 6 7 8 B.. 3 4 kb 23.13- 23.13- 9.42- 9.42- Allikohmim1'.- 6.56- 6.56- qw FIG. 1. (A) Hybridization of R. meliloti nod region to total DNA from Bradyrhizobium strains. 4-35- 4.35- 4w EcoRJ restriction pattern (visualized with ethidium bromide, lanes 1-4) and corresponding hybridization pattern with 32P-labeled (4 x 108 dpm/pg) 8.5-kb EcoRI fragment from pEK5121 as probe (low-strin- 2.32- 2.32- gency conditions, 96-hr autoradiographic exposure; lanes 5-8) are shown. Arrowheads indicate hybrid- 2.03- 2.03- ization bands observed when the probe was the 3.5-kb BamHI-EcoRI fragment. Lanes: 1 and 5, I110; 2 and 6, USDA 123; 3 and 7, RCR 3407; 4 and 8, IRc 78. (B) Hybridization of the 3.5-kb BamHI- EcoRI subfragment (3 x 10i dpm/ttg; low-stringen- cy conditions, 16-hr autoradiographic exposure) to EcoRI-digested pJN78-19 (lane 3) and pJN78-24 (lane 4). The corresponding restriction patterns vi- sualized with ethidium bromide are shown in lanes 1 and 2. Markers at left in A and B represent positions 0-57- offragments generated by HindIII digestion ofphage X DNA. Downloaded by guest on September 25, 2021 Genetics: Noti et al. Proc. Natl. Acad. Sci. USA 82 (1985) 7381

1.4 4-3.5-b 2 3 4-- 5.0 -3 EcoRi BamHI EcoI BamHI 5.(

R. meliloti | b,,,,,, ; IA 1a__ nod region A'

FIG. 2. Physical map of strain IRc m 78 DNA in pJN78-19 and pJN78-24. Schematic indicates regions ofhybrid- ization to the 8.5-kb EcoRI fragment of pEK5121. (Photo inset) BamHI/ EcoRI-digested pEK5121, stained with ethidium bromide (lane 1) or I I I I I I I 1 I hybridized with 32P-labeled 4.68-kb kb 0 10 20 30 40 46 (lane 2) and 7.08-kb (lane 3) EcoRI pJN78-19 fragments from strain IRc 78. Arrow- head, vector sequences present in pJN78-24 probe preparation (lane 3).

genes (on the 5.0-kb BamHI-EcoRI fragment, adjacent to the From 32 nodules (six isolates per nodule) picked at random nod genes) (6). However, only the 4.68-kb and 7.08-kb from three plants inoculated with R. meliloti 1027-24, only 3 putative nod fragments showed homology to pEK5121 (data nodules contained bacteria that had lost ISRmJ. In two not shown). similar experiments, we examined two nodules from each of Complementation of Nod- R. meliloti Mutants with IRc 78 20 and 25 plants inoculated with R. meliloti 1027-24, and only DNA. To determine whether the putative nod genes carried one nodule was found that contained revertant bacteria. on pJN78-19 and pJN78-24 were functionally homologous to Therefore, we conclude that pJN78-19 and pJN78-24 contain R. nodABC, isolates of the Nod- meliloti strains 1027 and genes that can complement the Nod functions in R. meliloti. ZB157 containing these plasmids were inoculated onto M. Plasmids pJN78-19 and pJN78-24 in R. meliloti strains 1027 sativa. Plants inoculated with R. meliloti ZB157 (Nod-) did not form nodules, whereas those inoculated with R. meliloti and ZB157 were not stably maintained through plant passage (data not shown). Although these are stable in ZB157-19 or R. meliloti ZB157-24 (strain ZB157 containing plasmids pJN78-19 or pJN78-24, respectively) formed normal-sized Rhizobium cultures grown under selective conditions, once nodules unable to fix nitrogen (Table 1). This finding was not infection occurs the plasmids may no longer be essential and unexpected, as R. meliloti ZB157 contains a deletion that are, thus, subsequently lost. Long et al. (10) have reported includes the common nod genes and extends into afix region similar findings with plasmid RP4. (6). Plants inoculated with R. meliloti 1027 (Nod-) were not Site Specific-Insertion of Tn5 Within the nod Region. Tn5 completely free of nodules. Inoculation with R. meliloti 1027-19 or R. meliloti 1027-24 (strain 1027 containing pJN78- 19 or pJN78-24, respectively) restored the nodule number to wild type and resulted in effective nodules comparable to

those with wild-type R. meliloti 1021 (Table 1). Nodules from I plants inoculated with R. meliloti strains 1027, 1027-19, and I 1027-24 contained bacteria that were neomycin-resistant (Nm') (Table 1). Thus, these nodules were not induced by a contamination with wild-type R. meliloti 1021. The restriction pattern of DNA isolated from nodule I ...... bacteria of all 54 plants (Table 1) inoculated with R. meliloti I III 1027 revealed that the insertion sequence (ISRml) within M I I r | nodC had been excised. Thus, the few nodules present on these plants were most likely the result of Nod' revertants.

_ _ _ NOD ++ I+/-I + + Table 1. Complementation of Nod- mutants of R. meliloti FIX -- 1+4-A + No. of Inoculant No. of nodules C2H2-reduction Lin' 7 c? T IcT ,- p I I I strain Phenotype plants per plant activity* IV t It I* t 0<

Table 2. Plant tests of strain IRc 78::TnS strains III mutants, because ofthe delayed nodulation, were difficult Nodules Shoot C2H2- to assess, and observations were inconclusive. nodGenes from Strain IRc 78 Are Homologous Within Other No. per Position dry reduction Bradyrhizobium. Although the nod genes from R. meliloti Strain plant Color on roots* weight, g activityt hybridized to total DNA of various Bradyrhizobium species Class I (Fig. 1), the apparent homology was low. Hybridization All strains 0 0.51 <0.1 conditions that allowed for at least a 25% base-pairing Class II mismatch (low stringency, ref. 24) and long exposure (96 hr) 4-5 63 White 1° + 20 0.66 <0.1 of the x-ray film were required to visualize the hybridization 4-4 74 White 10 + 20 0.80 <0.1 pattern. For determination of whether the homology of the Class III nod genes among the slow-growing strains is greater than the 6-7 41 Pink 20 1.12 10.1 homology between fast and slow growers, the 4.68-kb and 6-8 18 Pink 20 0.95 7.2 7.08-kb EcoR-I nod fragments from strain IRc 78 were used as 6-6 91 Pink 20 1.52 12.1 probes in hybridizations with DNA from three B. japonicum IRc 78t 192 Pink 10 + 20 7.55 45.1 strains. Hybridization of these fragments to total DNA from None 0 0.75 <0.1 strains I110, USDA 123, and RCR 3407 showed the presence Above results are an average from five plant replicates per strain. ofthree (10.21 kb, 5.52 kb, and 1.51 kb), two (6.55 kb and 1.51 Plants were harvested 43 days after inoculation. kb), and two (6.50 kb and 1.51 kb) bands, respectively, that *10, Primary root; 20, secondary roots. correspond in size to bands seen when the probe was the WValues represent PM0ol of C2H4 produced per hr per plant. *Wild type. 3.5-kb BamHI-EcoRI (nod fragment) of R. meliloti (Fig. 4). However, the intensities of the bands were greater in hybrid- izations with the IRc 78 nod fragments under conditions that mutagenesis was performed on the two nod-bearing frag- did not allow for a 25% base-pairing mismatch than with the ments (4.68 kb and 7.08 kb) inserted into pSUP202. Thirteen R. meliloti nod fragment. Therefore, on the basis of the independent TnS insertions were mapped in these fragments intensities of the bands, there is greater sequence homology (Fig. 3). The mutated fragments, carried on pSUP202, were among slow growers than between the fast- and slow-growing transferred into strain IRc 78 and exchanged for the wild-type strains. regions by homologous recombination on either side of Tn5 Infection of Alternative Hosts with Nod- Mutants of Strain (17). On the basis of their reaction with V. unguiculata, the IRc 78. Strain IRc 78 can form nodules with more than one mutants could be grouped into three classes (Fig. 3, Table 2). host species (17). If infection of plants within the host range Class I mutants were Nod-. Microscopic analysis ofthe roots of a strain proceeded through the expression of a common showed no unusual swellings and they appeared to be gene(s), then interruption of this gene(s) would be expected indistinguishable from uninoculated roots. Class II mutants to affect the ability of the strain to nodulate all of its hosts. were Nod' but Fix-. The nodules were white owing to the The effect ofTnS insertions within the nod region on six host absence of leghemoglobin and they were generally smaller species is shown in Table 3. A. hypogaea was included in this (1.5-2 mm) than the wild-type nodules (2.5-3 mm). The study because nodulation of this reduction in nodulation with class II mutants was apparently plant occurs without the not a result of a delay in modulation. In general, nodules will invasion of root hairs. Wild-type strain IRc 78 induced first appear on the primary root (the first root to emerge) and, nodules, pseudonodules (disorganized proliferation of thereafter, will form preferentially on the developing second- uninfected cortical cells), or root swellings on the various ary roots. The number of nodules induced on the primary hosts. In contrast, no nodules (or pseudonodules) were roots with class II mutants was comparable to that obtained formed on plants inoculated with mutant strains 4-2 and 6-4. with wild-type IRc 78 (data not shown). Further, examination Root-hair curling appeared normal on plants inoculated of the roots of plants at 7, 9, 11, 13, and 16 days after with wild-type strain IRc 78. No root-hair curling was seen on inoculation with class II mutants showed the appearance of 1 2 3 4 5 6 7 8 9 10 11 12 nodules at about the same time (day 11) as they appeared on kb plants inoculated with wild-type strain IRc 78. Class III 23.13- mutants showed Plants inoculated with .- variable phenotypes. 9.42- _I~a _ mutant strains 6-7 and 6-8 had reduced nodule numbers (21% 6.56- | 4 I~ 4 and 9%, respectively, of wild-type nodule levels, Table 2), 4.35- whereas the number of nodules induced with strain 6-6 was considerably higher (47% of wild-type levels). Nodule size 2.32- also was variable but was generally smaller with strain 6-6 2.03- (1.5-2 mm) than with strains 6-7 and 6-8 (2.5-3 mm). The plants exhibited Fix' phenotypes as evidenced by the leaf color (Fig. 3), shoot dry weights, and acetylene-reduction activities (Table 2), but overall plant growth even 51 days after planting was indicative of poor nitrogen fixation (data not shown). Class III mutants formed nodules mainly on the 0.57- secondary roots and only a few or none at all on the primary root (Table 2), a finding indicating a delay in nodule initiation. Examination of the roots of plants 11 days after inoculation FIG. 4. Hybridization of strain IRc 78 nod region to total DNA showed no evidence of nodulation. from other Bradyrhizobium. The EcoRI restriction pattern and Root-hair curling on V. unguiculata was not observed corresponding hybridization pattern with 32P-labeled 4.68-kb (108 dpm/ptg) or 7.08-kb (8 x 107 dpm/jLg) EcoRI fragments of strain IRc when class I strains 4-2, 6-4, 6-3, and 6-1 were used. Class I 78 as probes are shown. Lanes: 1, I110; 2, USDA 123; 3, RCR 3407; strains 4-1 and 4-3 showed root-hair deformation but not the 4, IRc 78; 5-8, hybridization pattern with 4.68-kb probe; 9-12, typical "shepherd's crook" formation. Class II mutants were hybridization pattern with 7.08-kb probe. (High-stringency condi- normal in their root-hair-curling response to infection. Class tions; autoradiogram exposed 17 hr). Downloaded by guest on September 25, 2021 Genetics: Noti et al. Proc. Natl. Acad. Sci. USA 82 (1985) 7383

Table 3. Nodulation of six hosts with wild type and mutants of genes have been found within a 30-kb region downstream of strain IRc 78 nifK nor within a region about 25 kb upstream of nifH Average nodule number (unpublished data). (phenotype*) Our results show that a common mechanism is involved in root-hair infection of V. unguiculata and crack-entry infec- Plant IRc 78 (wt) 4-2 6-4 tion of A. hypogaea and that this mechanism is dependent on Vigna unguiculata (cowpea) cv. the common nod genes. Infection of the non-legume California Black Eye 5 233 (Fix+) 0 0 Parasponia rigida involves elements ofboth the root-hair and Glycine max (American soybean) cv. crack-entry processes (29). Proliferation of the outer cortical Wilkin 25 (P) 0 0 cells disrupts the epidermis, allowing the bacteria to invade Glycine max (Asian soybean) cv. the intercellular spaces and then to penetrate the cortical cells TGm119 45 (Fix') 0 0 by infection-thread formation. Conceivably, the common Vigna radiata (mung bean) cv. nod genes may be involved in this third mode of infection. MB33 21 (S) 0 0 Cajanus cajan (pigeon pea) cv. We thank Dr. A. K. La Favre and Dr. A. R. J. Eaglesham for 15 (Fix-) 0 0 discussions and help in analyzing plant-test data; Dr. A. Kondorosi CITA-1 and Dr. E. Kondorosi for clones and initial incentive; Mrs. V. Schiff Arachis hypogaea (peanut) cv. NC7 32 (Fix+) 0 0 for technical assistance; and Dr. John Telford for photography. We Five plants of each cultivar were inoculated with each strain. especially thank Mrs. Suzette Noti for her excellent typing and Plants were harvested 42 days after inoculation. editing skills. This work was supported by a grant from Allied *P. pseudonodules; S, root swellings. Corporation (awarded to A.A.S.) and by the Boyce Thompson Endowment. plants inoculated with either strain 4-2 or strain 6-4 (data not 1. Verma, D. P. S. & Long, S. (1983) Int. Rev. Cytol. Suppl. 14, shown). 211-245. 2. Dart, P. J. (1974) in Biology of Nitrogen Fixation, ed. Quispel, A. DISCUSSION (North Holland, Amsterdam), pp. 382-429. Our data suggest that the structural and functional conser- 3. Chandler, M. R. (1978) J. Exp. Bot. 29, 749-755. vation between nod genes required for root-hair curling of R. 4. Banfalvi, Z., Sakanyan, V., Koncz, C., Kiss, A., Dusha, I. & meliloti and other fast-growing species, previously reported, Kondorosi, A. (1981) Mol. Gen. Genet. 184, 318-325. 5. Djordjevic, M. A., Schofield, P. R., Ridge, R. W., Morrison, can now be extended to include those of a slow grower, N. A., Bassam, B. J., Plazinski, J., Watson, J. M. & Rolfe, B. G. Bradyrhizobium sp. (Vigna) strain IRc 78 (Fig. LA). This (1985) Plant Mol. Biol. 4, 147-160. work substantiates the finding by Marvel et al. (9) that DNA 6. Kondorosi, E., Banfalvi, Z. & Kondorosi, A. (1984) Mol. Gen. from a Bradyrhizobium sp. (parasponia) strain can comple- Genet. 193, 445-452. ment the nodC gene of R. meliloti. Restoration of the Nod' 7. Torok, I., Kondorosi, E., Stepkowski, T., Posfai, J. & Kondorosi, to R. meliloti ZB157 (which has a total deletion of A. (1984) Nucleic Acids Res. 12, 9509-9524. phenotype 8. Rossen, L., Johnston, A. W. B. & Downie, J. A. (1984) Nucleic nodABC) by DNA from strain IRc 78 makes it seem likely Acids Res. 12, 9497-9508. that all three genes are conserved across genera. 9. Marvel, D. J., Kuldau, G., Hirsch, A. M., Park, J., Torrey, J. G. & Three distinct regions essential for nodulation or nitrogen Ausubel, F. M. (1984) in Advances in Nitrogen Fixation Research, fixation were revealed in the 11.76-kb region from strain IRc eds. Veeger, C. & Newton, W. E. (Nijhhoff-Junk, The Hague, The 78 (Fig. 3). The region specifying functions essential for Netherlands), p. 691. two types of nod genes 10. Long, S. R., Buikema, W. J. & Ausubel, F. M. (1982) Nature nodulation (class I mutants) contains (London) 298, 485-488. distinguishable by their effect on root-hair curling. 11. Boyer, H. W. & Roulland-Dussoix, D. (1969) J. Mol. Biol. 41, The region immediately to the left of the nod genes (class 459-472. II mutants; Fig. 3) affects nodule function (Table 2). Inter- 12. Clarke, L. & Carbon, J. (1979) Gene 5, 111-126. estingly, the 5.0-kb BamHI-EcoRI fragment to the left of the 13. Sternberg, N., Tiemeier, D. & Enquist, L. (1977) Gene 1, 255-280. 3.5-kb BamHI-EcoRI fragment containing the nod genes in 14. Simon, R., Priefer, U. & Puhler, A. (1983) in Molecular Genetics of the Bacteria-Plant Interactions, ed. Puhler, A. (Springer, Berlin), R. meliloti 41 (Fig. 2; see refs. 4 and 7) carries a gene(s) for pp. 98-106. nitrogen fixation. The 5.0-kb BamHI-EcoRI fragment of R. 15. Hadley, R. G. & Szalay, A. A. (1982) Mol. Gen. Genet. 188, meliloti 41 hybridized to the 4.68-kb EcoRI fragment 361-369. mutagenized with TnS in the Nod' Fix- strains 4-4 and 4-5 16. Birnboim, H. C. & Doly, J. (1979) Nucleic Acids Res. 7, (Figs. 2 and 3). However, it is not known whether the fix 1513-1522. gene(s) in these homologous regions are related. 17. Jagadish, M. N. & Szalay, A. A. (1984) Mol. Gen. Genet. 196, 290-300. The region to the right of the nod genes (class III mutants; 18. Rosenberg, C., Casse-Delbart, F., David, M., Dusha, I. & Fig. 3) influences both the time required by strain IRc 78 to Boucher, C. (1981) J. Bacteriol. 150, 402-406. infect the roots initially and the numbers of nodules induced. 19. Morris, D. W., Noti, J. D., Osborne, F. A. & Szalay, A. A. (1981) The variability of nodule number and size within the class III DNA 1, 27-36. mutants is similar to that reported by Stacey et al. (25). They 20. Hanahan, D. & Meselson, M. (1980) Gene 10, 63-67. two classes of B. japonicum mutants affected in 21. Yang, R. C. A., Lis, J. & Wu, R. (1979) Methods Enzymol. 68, identified 176-182. either the rate of nodule formation (class I) or the frequency 22. Rigby, P. W. J., Dieckmann, M., Rhodes, C. & Berg, P. (1977) J. of nodulation (class II). Mol. Biol. 113, 237-251. Symbiotic plasmids analogous to those found in Rhizobium 23. Ruvkun, G. B., Long, S. R., Meade, H. M., van den Bos, R. C. & species have not been found in strain IRc 78 or in other Ausubel, F. M. (1982) J. Mol. Appl. Genet. 1, 405-418. slow-growing strains (unpublished data and refs. 26 and 27). 24. Meinkoth, J. & Wahl, G. (1984) Anal. Biochem. 138, 267-284. A. S., Noel, D., Maier, R. J., Silver, L. E. & genes in strain IRc 78 are 25. Stacey, G., Paau, We assume, therefore, that the nod Brill, W. J. (1982) Arch. Microbiol. 132, 219-224. located on the chromosome. Further, whereas nod genes in 26. Hadley, R. G., Eaglesham, A. R. J. & Szalay, A. A. (1983) J. Mol. R. trifolii (28) and R. meliloti (6, 7) are separated from the Appl. Genet. 2, 225-236. nitrogenase (nif) genes by only about 16 kb and 25 kb, 27. Cantrell, M. A., Hickok, R. E. & Evans, H. J. (1982) Arch. Micro- respectively, linkage of the nod genes to the nif region in biol. 131, 102-106. 28. Schofield, P. R., Djordjevic, M. A., Rolfe, B. G., Shine, J. & strain IRc 78 is not apparent. In strain IRc 78 nifH is Watson, J. (1983) Mol. Gen. Genet. 192, 459-465. approximately 35 kb upstream of the niJDK operon. No nod 29. Lancelle, S. A. & Torrey, J. G. (1984) Protoplasma 123, 26-37. Downloaded by guest on September 25, 2021