Proc. Nati. Acad. Sci. USA Vol. 88, pp. 637-641, January 1991 Genetics
The Bradyrhizobium japonicum nolA gene and its involvement in the genotype-specific nodulation of soybeans (modulation gene/gene-for-gene interactions/symbiosis/plant-bacteria interaction)
MICHAEL J. SADOWSKY*t*, PERRY B. CREGANt, MICHAEL GOTTFERT§, ARUN SHARMA$, DAVID GERHOLD$, FRANCISCO RODRIGUEZ-QUINONESt, HAROLD H. KEYSERII, HAUKE HENNECKE§, AND GARY STACEY¶ *Soil Science and Microbiology Departments, University of Minnesota, St. Paul, MN 54108; tNitrogen Fixation and Soybean Genetics Laboratory, U.S. Department of Agriculture-Agriculture Research Service, BuildingS 011, BARC-West, Beltsville, MD 20705; §Mikrobiologisches Institut, Eidgendssische Technische Hochschule Zentrum, CH-8092 Zurich, Switzerland; ICenter for Legume Research, Department of Microbiology and Graduate Program in Ecology, University of Tennessee, Knoxville, TN 37996; and I1The NifTAL Project, University of Hawaii, P.O. Box 0, Paia, HI 96779
Communicated by Harold J. Evans, October 8, 1990
ABSTRACT Several soybean genotypes have been identified which specifically exclude modulation by members of Bradyrhizobiumjaponicum serocluster 123. We have identified and sequenced a DNA region from B. japonicum strain USDA 110 which is involved in genotype-specifc modulation of soybeans. This 2.3-kilobase region, cloned in pMJS12, allows B. japonicum serocluster 123 isolates to form nodules on plants of serogroup 123-restricting genotypes. The nodules, however, were ineffective for symbiotic nitrogen fixation. The nodulation-complementing region is located approximately 590 base pairs transcriptionally downstream from nodD2. The 5' end of pMJS12 contains a putative open reading frame (ORF) of 710 base pairs, termed nowl. Transposon Tn3-HoHo mutations only within the ORF abolished modulation complementation. The N terminus of the predicted noL4 gene product has strong similarity with the N terminus of MerR, the regulator of mercury resistance genes. Translational IacZ fusion experiments indicated that nolA was moderately induced by soybean seed extract and the isoflavone genistein. Restriction fragments that hybridize to pMJS12 were detected in genomic DNAs from both nodulation-restricted and -unrestricted strains. In soybean production areas of the midwestern United States, members of Bradyrhizobium japonicum serocluster 123 are the dominant indigenous competitors for soybean nodulation (1-3). Cregan and Keyser (4) identified several soybean genotypes [including plant introductions (PIs) and cultivars] which restricted nodulation and reduced the competitiveness of B. japonicum strain USDA 123. These genotypes nodulate normally with strain USDA 110. Three of the genotypes, PIs 377578, 371607, and 417566, were subsequently shown to restrict nodulation by different serocluster 123 isolates (5, 6). In B. japonicum, the essential common nodulation (nod) genes (nodA, nodB, nodC, and nodDi), as well as nodY, nodD2, and nodSUIJ, are closely linked on the chromosome (7-9). Host-specific nodulation (hsn) determinants have also been reported in Rhizobium melioti (10), Rhizobium leguminosarum biovars trifolii (11) and viceae (12), in the broad host-range Rhizobium sp. strain MPIK3030 (13), and in B. japonicum (7, 14, 15). In Rhizobium and Bradyrhizobium, nod, hsn, and other symbiosis genes are induced by flavonoid exudates from the host plant (16-18). While hsn genes have been localized in Rhizobium and Bradyrhizobium, there have been no reports of the identification of genotype-specific nodulation (GSN) genes in bradyrhizobia. The GSN genes would refer to those bacterial sequences which allow nodulation of specific plant genotypes
within a given legume species. One example of a GSN-like gene is nodX. This gene, in R. leguminosarum biovar viceae strain TOM, is not needed for nodulation of commercial pea cultivars, but it is essential for nodulation of the more primitive Afghanistan pea (19). Another GSN-like locus has also been reported in Rhizobiumfredii strain USDA 257 (20). In this report, we describe the identification and sequence analysis** of a gene, nolA, from B. japonicum strain USDA 110, that allows serocluster 123 isolates to nodulate soybean plants having USDA 123-restricting PI genotypes. We show that sequences in this region are induced by the isoflavone genistein and that the predicted amino acid sequence of the deduced open reading frame (ORF) contains a motif that may be similar to the helix-turn-helix motif associated with DNAbinding regulatory proteins.
MATERIALS AND METHODS Bacterial Strains, Plasmids, and Growth Conditions. The B. japonicum strains used in this study were obtained from the U.S. Department of Agriculture-Agricultural Research Service (Beltsville, MD). The nodulation phenotypes of B. japonicum strains on soybean hosts are as previously described (5, 6, 21). Cosmids pRJUT10, pR32, and pR29 were as previously described (22). The B. japonicum strains were maintained and mated on AG medium (23). The Escherichia coli strains were grown on LB medium (ref. 24, p. 440). As required, media were supplemented with tetracycline (Tc; 25-100 Ag/ml), kanamycin (50 pug/ml), spectinomycin (50 ptg/ml), carbenicillin (25 pug/ml), or ampicillin (30 Ag/ml). DNA Biochemistry. Total genomic B. japonicum DNA was isolated and Southern hybridizations were done as described (23). Plasmid DNA was isolated by using the alkaline rapid lysis procedure (ref. 24, pp. 90-91). Cloning and Identification of Nodulation and GSN Genes. The HindIll fragments of cosmid pR32 were subcloned in pVK102 as described (25, 26). Cosmids containing fragments corresponding to the nodDl YABC, nodSUIJ, nodZ, nodD2, or other gene regions (7-9, 25, 27) were used to transform E. coli strain S17-1 (ref. 24, pp. 254-255) (see Fig. 1). Conjugal transfer of pR29, pR32, and pMJS HindIII subclones of pR32 to B. japonicum strains USDA 162 and SD6-lc (serogroup 123) was done by using helper plasmid pRK2073 (28). Transconjugants were selected on AG medium containing Tc (60 Ag/ml). A Sal I partial-digestion genomic DNA library from Abbreviations: PI, plant introduction; GSN, genotype-specific nodulation; ORF, open reading frame; Tc, tetracycline. tTo whom reprint requests should be addressed at: Soil Science Department, University of Minnesota, 246 Borlaug Hall, 1991 Upper Buford Circle, St. Paul, MN 55108. **The sequence reported in this paper has been deposited in the GenBank data base (accession no. M58360).
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
637
Sadowsky et al.
Genetics:
638
Proc. Natl. Acad. Sci. USA 88 (1991)
strain DE3-la was constructed in cosmid pVK102 as described (26). Cosmid clones corresponding to the GSN and nod gene regions were identified by hybridization to the HindIII fragments of pMJS12 and pMJS18 and mated into B. japonicum strain SD6-lc as described above. DNA Sequence Analysis. The 1.9-kilobase (kb) Sma I-Bgl II fragment (see Fig. 1) was digested with the appropriate restriction enzymes and fragments were inserted into the corresponding restriction sites of phage vectors M13mp18 and M13mpl9 (29). The DNA sequence was determined by using the chain-termination method of Sanger et al. (30) and a DNA sequencer (Applied Biosystems model 370A). The UWGCG (University of Wisconsin Genetics Computer Group) software package (release 6.1) was used for computer-aided DNA and protein sequence analysis and the amino acid alignment was done by using the MULTALIN program (31). Data base searches were done by using the TfastA program in conjunction with GenBank and EMBL databases. Construction of lacZ Fusions. Transcriptional/translational IacZ fusions were made in the 1.7- and 2.3-kb HindIII fragments of pR32 (see Fig. 1) using Tn3-HoHol as described by Stachel et al. (32). The location and orientation of inserts were determined by restriction mapping. Fusions were transferred to B. japonicum strains SD6-lc and USDA 110 as described above. fi-Galactosidase Assays. Assays to monitor the induction of lacZ fusions by soybean seed extract and genistein were done as described (8). Strains containing IacZ fusions were induced for 14-16 hr with 2.0 gM genistein (ICN) or 10% (vol/vol) Glycine max cv. Essex seed extract (8). Duplicate cultures were assayed for each treatment and values are in P-galactosidase units (8, 16). Nodulation Studies. Plant assays were done in a Monmouth fine sandy loam soil, containing less than one B. japonicum cell per g of soil (5, 6, 25). Pots were planted with four seeds each of Glycine max PI 377578 [serogroup 123 restricting (4)] and a nonrestrictive host, cv. Williams. Transconjugants (about 10 per mating) and nodulation-positive (MN1-lc, DE3-la, and USDA 110) and -negative (USDA 162, SD6-lc, and USDA 123) controls were inoculated onto at least four replicate plants. Plants were watered with nitrogen-free nutrient solution (5) and harvested 35 days after inoculation, and nodule numbers were determined (4). Nodulationcomplementing transconjugants were isolated from surfacesterilized nodules (ref. 33, p. 7-8) on AG medium containing Tc (100 Ag/ml) or kanamycin (100 ,ug/ml). The identity of B. japonicum transconjugants and nodule isolates was -verified by using strain-specific fluorescent antibodies (34) or spotR
R H H
H H
R H H H
pRJUTIO
R H
R
H
R
pR29 pR32
H.__ D2 4
DG62 DG73 131_ DG66
D1
C
S
U
I
J
DG1
1 11
-I HIRBBg BHBSS H >-~~~^
Sequenced pMJSRegion
IM
pMJS 12
2.3
YA B
-
4
1.7
R
R
H pMJS 18
3.9
RR
1 kb
FIG. 1. Physical-genetic map of B. japonicum strain 1110 nodulation loci. The H locations of the GSN gene nolA, nodDI YABC, nodD2, nodS UIJ, and nodZ on pR32 are shown. The relationship of H subclones pMSJ9, pMJS12, pMJS18, 1 kb :1 pMJS22, and pMJS24 used in complementation and hybridization studies to the physz _ ical-genetic map of nodulation loci on a 17-kb portion ofpR32 are indicated. Arrows denote transcriptional orientation of genes. Numbers below subclone lines are HindIII fragment sizes in kb. The locations and H H orientations of Tn3-4acZ fusions in the 1.7pMJ'S 2 2 and 2.3-kb regions of pR32 are shown by stalked arrows. The region sequenced is denoted by the broken line. Restriction enzyme sites: H, HindIII; B, BamHI; Bg, Bgl II; R, EcoRl; and S, Sma 1. )
H
,lToIA 4
RESULTS Identification of Genes Involved in the GSN of Soybeans. To determine if DNA regions within and clustered around nodDI YABC were involved in the GSN of soybeans, we conjugated pR29 and pR32 (both derived from strain USDA 110) (Fig. 1) into the PI 377578 nodulation-restricted strain SD6-lc (serogroup 123). When the SD6-lc(pR29) and -(pR32) transconjugants were inoculated onto PI 377578, only the pR32 transconjugants nodulated. The SD6-lc(pR32) transconjugants nodulated PI 377578 to an extent equal that of the nodulation-unrestricted strains, MN1-lc and USDA 110 (Table 1) and produced about 30-45 times the number of nodules (range of 90-138 nodules per plant) on the PI genotypes that the control strains SD6-lc or SD6-lc(pVK102) produced. The transconjugants, however, were unstable, since only about 4 out of 10 plate transconjugants were able to nodulate the PI genotypes. While all of the nodule isolates grew on AG medium with Tc and reacted with antiserum specific for serogroup 123 strains, only 7 out of 10 could form nodules when retested on the PI genotypes. Moreover, the SD6lc(pR32) transconjugants formed ineffective symbioses with PI 377578. Nodules were white, suggesting a lack of leghemoglobin production. While some nodules were large and had surface patterning and color typical of soybean nodules, others were small and lacked surface ribs. Since the SD6lc(pR29) transconjugants did not nodulate PI 377578 (data not shown), our results suggested that the DNA region in the area of nonoverlap (about 8.5 kb) was involved in the genotype-specific nodulation of soybeans. To test this, HindIII fragments from pR32 were subcloned in pVK102 (Fig. 1) and individually mated to SD6-lc, and tratsconjugants were tested for nodulation ability. Only the genomic DNA region contained in plasmid pMJS12 (2.3 kb) allowed SD6-lc to nodulate the restrictive soybean genotype (Table 1). The SD6-lc(pMJS12) transconjugants produced about 35 times more nodules (range of 97-120 nodules per plant) than the uncomplemented parent strain, SD6-lc(pVK102), and were also ineffective for nitrogen fixation on the PI genotypes. The GSN region is located about 590 bp transcriptionally downstream from the end of nodD2 (Fig. 1). Sequences present on pMJS12 did not hybridize to B. japonicum nodDI YABC' (pMJS18), R. meliloti hsnABCD, or Nod box consensus sequence gene probes (7, 35). The other SD6-lc
H
H
,
blot immunoassays (21). The USDA 162 and SD6-lc transconjugants isolated from nodules were retested for nodulation ability as described above.
pMJS24 5.6
Genetics:
Sadowsky et al.
Table 1. Nodulation of Glycine max PI 377578 by the serogroup 123 strain SD6-1c and its pR32 and pMJS12 transconjugants Nodules on Nodules on cv. Williams PI 377578 Dry wt, Dry wt, No. mg Isolate No. mg 1.0 7.0 3.5 14.0 Uninoculated 114.7 453.0 4.0 26.0 SD6-1c 2.8 9.0 133.2 361.0 SD6-lc(pVK102) 139.0* 111.8 111.0 113.5* SD6-lc(pR32) 153.8 210.0 109.8* 194.0* SD6-lc(pMJS12) 96.7 284.0 116.7 290.0 MN1-1c 215.0 142.8 351.0 107.75 USDA 110 Nodule number and dry weight are per plant. Values for transconjugants are means of 8-24 replicates. *Significantly different from the uninoculated or negative controls (P < 0.05) by Duncan's new multiple range test.
transconjugants (containing pMJS9, pMJS18, pMJS22, or pMJS24) failed to nodulate the restricting PI genotype (data not shown). Strains SD6-lc(pVK102), SD6-lc(pR32), and SD6-lc(pMJS12) nodulated and fixed nitrogen normally with the nonrestrictive commercial soybean genotype Williams (Table 1). A majority (6 out of 10) of SD6-lc(pMJS12) nodule-isolated transconjugants retained the ability to nodulate PI 377578, suggesting that pMJS12 most likely contributed to the formation of these nodules. Plasmid pMJS12 also allowed another serogroup 123 isolate, USDA 162, to nodulate PI 377578 (range of 126-145 nodules per plant), suggesting that its involvement in GSN is not limited to a particular serogroup strain. To determine whether a functional nodulation complementing GSN DNA region was unique to strain USDA 110, we made a genomic DNA library from strain DE3-la [nodulation unrestricted, serogroup 127 (5, 25)] and mated cosmid clones which hybridized to the pMJS12 gene probe to strain SD6-lc. One of the transconjugants, SD6-lc(pFR27), was found to nodulate PI 377578 (mean of 22 nodules per plant) to the same extent as DE3-la. Cosmid pRF27 contains sequences analogous to those extending from the USDA 110 nodZ gene to the end of the GSN gene region (data not shown). A pFR27 subclone, pFR2501, containing only the DE3-la GSN region, was later found to allow SD6-lc to nodulate the PI genotypes (mean of 21 nodules per plant), suggesting that other unrestricted strains contain functional GSN gene regions. The DE3-la GSN region also failed to complement nitrogen fixation effectiveness in the SD6-lc transconjugants. In several cases, SD6-lc(pFR2501) transconjugants were also ineffective on both permissive (cv. Williams) and restrictive soybean hosts. The DE3-la-derived GSN region was more unstable than the homologous 110 region, and cosmids containing the GSN region mated to recipients at frequencies 1/10th to 1/100th of those with DNA from other regions (10-5 to 1O-7 per recipient), and 60-100%o of transconjugants contained cosmids with deletions in the GSN gene region. The instability phenomenon also apparently prevented us from making site-directed mutations in the GSN gene region. While strain SD6-lc is restricted for nodulation of PI 377578, it still forms a small number (Table 1) of crown nodules on this PI genotype (5). To determine if these nodules were produced by a nodulation-competent subpopulation of serogroup 123 strains, nodule isolates (16 and 19 from cv. Williams and PI 377578, respectively) of strain USDA 123 were inoculated onto PI 377578. Nodulation by these isolates was not significantly different (P < 0.05) than by the USDA 123 parent. Thus, the observed "leakiness" of the restrictive nodulation phenotype is not due to a phenotypically stable
Proc. Natl. Acad. Sci. USA 88 (1991)
639
nodulation-competent subpopulation of serogroup 123 isolates. Expression of the GSN Region. To determine if sequences in and around pMJS12 are induced by host plant-derived signal compounds, we constructed Tn3-lacZ fusions in the 1.7- and 2.3-kb fragments of pR32 (Fig. 1) and studied their induction (in strain USDA 110) with soybean seed extract and genistein. Only fusion DG62 was induced by soybean seed extract (95 ± 5 B-galactosidase units) or genistein (89 ± 6 units). This fusion is in the 2.3-kb HindIII fragment contained in pMJS12 and maps within the GSN ORF described below. Fusion DG62 also had a small amount of 3-galactosidase activity (35 ± 7 units) in the absence of inducer, suggesting that this region is constitutively transcribed at a low level. None of the other lacZ fusions examined (DG1, DG6, or DG73) were induced by seed extract or genistein (mean of 5.6 ± 1 units). These results suggest that transcription of pMJS12 is in the same direction as nodD2 and is independent of transcripts originating from adjacent upstream regions. DNA Sequence Analysis. To determine the coding capacity of the GSN gene region and to analyze the DNA region downstream of nodD2, we sequenced the 1.9-kb Sma I-Bgl II fragment of pR32 (see Fig. 1). This fragment contains the 5' end of the 2.3-kb HindIII fragment cloned in pMJS12 (beginning at nucleotide 630) and areas 3' of the end of the nodD2 transcript (27). We identified one ORF (nucleotides 959-1669) within the sequenced region on the basis of potential start codons (Fig. 2). To determine if the 710-bp ORF was involved in genotype-specific nodulation, we tested Tn3-HoHo insertion mutations within or downstream of the ORF (fusions DG62 and 131, respectively; see Fig. 1) for the production of nodules on PI 377578. Only SD6-lc transconjugants with an insertion mutation within the ORF of pR32 (DG62) failed to nodulate, suggesting that the ORF gene product is most likely involved in the extended host-range phenotype. We have named this sequence nolA. Amino Acid Sequence of the Deduced nolA Gene Product. The putative translational start site at the first available ATG codon of the ORF, at position 956, results in a protein with a predicted molecular weight of 26,871. The predicted NolA protein product would have a fairly high arginine content (14.3%), which would result in a net positive charge. Since this positive charge might facilitate an interaction of the protein with the negatively charged DNA backbone, we looked for a helix-turn-helix motif, which is characteristic of many DNA-binding proteins. Using the method of Dodd and Egan (36), we identified such a motif in amino acids 13-32 (Fig. 2), and we suggest that nolA may code for a DNAbinding regulatory protein. A data base search revealed that the N terminus of NoIA is similar to the N terminus of MerR. MerR has been shown to be a metalloregulatory DNAbinding protein that activates or represses transcription of mercury resistance genes (37). The alignment of NolA to the MerR proteins of Bacillus sp. strain RC607 and Staphylococcus aureus (38, 39) is shown in Fig. 3. While the homology between the two MerR proteins extends to the very end of both proteins, the homology between NolA and MerR is restricted to the first 70 amino acid residues. However, since this is the region in which the helix-turn-helix motif is located, NolA and MerR could share a mode of regulating gene
expression. Conservation of GSN Gene Region in B. japonicum Strains. Since results from previous studies indicated that pMJS12 hybridized to genomic DNA from nodulation-restricted and -unrestricted serocluster 123 strains (25), it was of interest to determine if the GSN gene region was present in other B. japonicum isolates. The pMJS12 gene probe hybridized to genomic DNA from several serologically distinct B. japonicum strains. While pMJS12 hybridized predominantly to 6.1-kb EcoRI fragments from B. japonicum strains in DNA
Genetics:
640
Sadowsky et al.
Proc. Natl. Acad. Sci. USA 88 (1991)
1
CCCGGGACATTTCGTGCGATCCGTCTCACCGTTGCCGTAAGCACATTCGCGAAGTCTACGAACCTCACATTCCAACACTTGTCTCAACAC
91
ACAATTCCTGCTCGCCTGGCTTGCTCTTATTGCCGGGGGAAAGGATCCTTGCGGCAGGCAAGAGGCGCTTTCGCTCTTCGCACCTCCGCC
181
TCGGGGCGAGAAACGGCAGACTGCCGACTTGATGCTTTTAGGGACGGGAACGGAACGGGGCCGGCGAGCGACGGAGCAATGCCTTCGCA
271
CTGAAGCTAGGCGTTTACGTTGCAACGGCCGCGGCAGGGCGCAGCTCTCTCAAGuAGGTAGGCCAGGCACGGAGCTCTTGCCGTCAACTC
361
CCTGGGTCGGTAGTTGCTCCCGATGAACGGAGATGATCAGCTTATTCGGATTCCTACCACCGTTCATTGTGAGCATAGAGATCGTCGGAC
451
TTGCCGTGCCGTCGAAGCTAGCATCGCGAACGCGTGTGGCCTGAACGCATCAC
CACATCACATCCTCATCGCTTTGCGCCGCA
541
GGTATGACTTCTTTTGGCCCTACCGCAGCCAAACGTCTTGACGACGTGCGCTGCGCTTGACTCAATTTCAAACAAGGAGCTGAACGTTGA
631
AGCTTTGCAGCGGGGTGTTCGATCCAGAAGAGCTATCTGCCCTCGGACGGCTTTATGACGGCGCCGTCGACGCGCTACCTCCGTCGATGC
721
GGAGTCCAGAAAACTGCCAGCAATCGCCAAACTTATCCTGGAGCGTACGGCGGCTGGCGAGGCTGAACTGGCACGCCTGGCGAACTTGT
811
TGATCACCCTTTCTCCTGAGGGTTGATCAGCCGCCATTCTAAGGATAGTCACCACTCGCGCGCATGAATTAAAGGCGATTCAACTGGAGC
901
TTAGAAATCCGCTTGAATCATACGTCACGTCAGGTTGTAGGCTTTGAAATTGAACAACATGAACAGAGCTACACCAAGACGGCGTCGATG
991
GCGCATTGGCGAACTTGCGAGGCGACCGG>AACGGTGCGCACGCTGCACCATTATGAGCACACTGGACTGCTAGCAGCGACAGAGCG
N R
M
R.| I
1081
a
T
G
V
T
V
R
T
L
H
H
Y
Z -H- |T
G
L
L
A
A
R T
R W E
R
E G
G
H
R
M
Y
R E
D
S
G
Q
R
V
Q
H
I
A
R
L
R E
L
G
F
S
L
V
I
R K A
M E
G
T
T
S
L
T
D
L
L
R K H
L E R
I
B
V
Q V A R
T
L
L
R
D
R
L
R
N
M
T
I
D
S
E
A
Q
V
S
D
V
E
L
P
A
T
L
N
A
M
R
A
E
T
R
S
Q
T
S
R
C
T
C
N
L
A
A
E
R
E
D
R W R
R
I
R
D
D
TCTTCGCGACTGCATGGATGGGGGCGAGCACCCTTGCGGAGAGCGCGCAAGGCTGTCGCGGTTGCAGCACGCTTGCTGATCAGCGAAAT L
1531
E
R
GTCGCGCGCCGAGACGCGGAGTCAGACGTCCCGGTGCACCTGCAATTTAGCTGCGG&GCGAGAGGACCGCTGGCGGCGAATCCGTGATGA S
1441
A
R
AACTCTGTTGCGCGACCGTTTGCCGACATGACCATCGACAGCGAGGCGCAGGTAAGTGTGGATGAGCTGCCTGCCQCCCTGAACGCCAT T
1351
L
P
GGAGATCCGTAIAGCTATGGAGGGGACGACCTCACTCACGGACTTGCTGCGCAAACATCTAGAGCGCATAGAAGTCCAAGTCGCGCGCAC E
1261
E
T
CACTGAGGGCGGTCACCGGATGTATGACCGCGAAAGCGGGCAACGGGTTCATCAGATTCGCGCGCTGCGTGAGCTTGGCTTCTCGCTCGT T
1171
G
A
R
D
C N D
G
G
E
H
P
C
G
E
R
A
K
A
V
A
V
A
A
L
R
L
I
S
E
I
CGCCGGCGACGATTCACGTGTTTCGATGATCCTGAAGGTACTTGCACGATTAAGCGCGCCCCGCAGTCTGGCTGGTTGGGATCCCTGTCT A
G
D
D
S
R
V
S
M
I
L K V
L
A
R
L
S
A
P
R
S
L
G W D
A
P
C
L
1621
AATGCAATATCTCGATCTTGCCCTTGGCGGATTGGAGGTCAGCCTTACTGACGCATTCATGGCGGGTTCCACACGCCGGTTAGGGCAAG
1711
CTAAGAAACGCTAGGTGAACCGGCCGCTCTCCGACGCGGCTTGGAGTTTTACGGAATCAAACGGAAGGATTGTCACCGCCGGGCGATCCG
1801
CGTCCCTGACGCCTCCATCGCTCTTCCCA^AAGGGCTTGATCCGCACGTAACGTCAAGTTGTAGATCT
M
Q
Y
L
D
L
A
L
G
G
L
E
D
Q
P
Y
homology (40) [or fatty acid (41)] groups I and IA (USDA 4, 6, 38, 62, 110, 122, and 124), little or no hybridization was observed with DNAs from strains USDA 31, 46, 61, 76, 94, 130, or 135 (data not shown). With these strains, however, there was no apparent relationship between hybridization profiles and nodulation restriction. A GSN gene probe did not hybridize to genomic DNA from R. leguminosarum biovars trifolii and viceae strains or from Agrobacterium tumefaciens (not shown).
conceptually different than Bradyrhizobium and Rhizobium hsn genes (9-13, 18), which are involved in host-range specificity at the legume genus level. In contrast to other Bradyrhizobium nod (7, 27) and Rhizobium nod and hsn (35) gene loci, the nolA gene locus did not hybridize or have sequence similarity to the 25-mer nod box consensus sequence (7, 27), which has been implicated to have a role in the coordinated regulation of nodulation genes (35). This suggests that nolA, and several other B. japonicum nodulation genes, may be regulated by nod-box-independent promoters (7, 27). While nolA allows serogroup 123 strains to nodulate the PI genotypes, complementation is not complete and transconjugants fail to fix nitrogen. Also, while most SD6-lc transconjugants fix nitrogen with cv. Williams, the nodules produced are smaller than the wild-type parent. In some extreme cases, nolA totally interfered with the nitrogen fixation ability of SD6-lc(pFR27) transconjugants on the permissive soybean host. We are not sure why the nolA locus interacts in this manner, but there may be physical or functional incom-
DISCUSSION Using a nodulation-complementation approach, we have identified a B. japonicum USDA 110 gene, nolA, which is involved in the GSN of soybeans. The nolA gene, cloned in pMJS12, allows serogroup 123 isolates to nodulate soybean plants with restrictive genotypes. Since sequences present on this fragment allow for the nodulation of plants with a specific genotype within a legume species, we have designated this a GSN (genotype-specific nodulation) locus. This locus is SaMerR
1 10 20 30 40 50 MGMKISELAKACDVNKETVRYYERKGLIAGPPRNESGYRIYSEETADRVRFI
FIG. 3. Amino acid sequence of the NolA protein of B. alignment ****** ** * ** ** **** * *** * * * japonicum (BjNolA), MerR of BaMNRATPRRRRWRIGELAEATGVTVRTLHHYEHTGLLAATERTEGGHRMYDRESGQRVHQI cillus sp. strain RC607 [BsMerR 60 70 80 90 (38)], and MerR of Staphylococcus 100 110 KRMKELDFSLKEIHLLFGVVDQDGERCKDMYAFTVQKTKEIERKVQGLLRIQRLLEELKE aureus [SaMerR (39)]. Further de*** ** * * ** * **** * * *** ** * * ** * ** * * *** tails are given in the text. Identical KRMQELGFTLNEIDKLLGVVDRDEAKCRDMYDFTILKIEDIQRKIEDLKRIERMLMDLKE **** * ** * * * amino acid residues are marked by RALRELGFSLVEIRKAMEGTTSLTDLLRKHLERIEVQVARTTLLRDRLRNMT IDSEAQVS asterisks. Dots at the end of the NolA sequence indicate that amino 120 135 acids extend farther but the seKCPDEKAMYTCPIIETLMGGPDK quence is not homologous. The puRCPENKDIYECPIIETLMKK tative helix-turn-helix motif is underlined. VDELPATLNAMSRAETRSQTSRC... *
BsMerR
BjNolA SaMerR BsMerR
BjNolA SaMerR BsMerR
BjNolA
1868
FIG. 2. Nucleotide sequence of the nolA gene region. The sequence starts with the Sma I site at the end of nodD2 (see Fig. 1). The HindIlI site at the start of pMJS12 (position 630) is underlined. The deduced amino acid sequence of the nolA gene product is given in one-letter code. The putative helix-turn-helix motif at the N terminus of the NolA protein is boxed.
*
***
*
*****
*****
***
*
*
*
***
**
*
**
**
MKFRIGELADKCGVNKETIRYYERLGLIPEPERTEKGYRMYSQQTVDRLHFI
Genetics:
Sadowsky et al.
patibility between nolA and the indigenous nodulation and nitrogen fixation genes (42). This instability may also be reflected in the low transfer frequencies we obtained with nolA-containing cosmids, the presence of deletions in the nolA region of transconjugants, and our inability to construct stable mutations in the strain USDA 110 nolA locus. Though the function of the nblA gene is unknown, the N terminus of its product shows similarity to the N terminus of the MerR proteins from Bacillus (38, 43), Staphylococcus (31), and several Gram-negative bacteria. NolA and these proteins all possess helix-turn-helix motifs which are characteristic of many DNA-binding proteins. Thus, NolA and MerR may all be members of a unique family of gene regulators. Our results indicate that the genetic components of both partners interact to influence nodulation and competition in the soybean/Bradyrhizobium symbiosis. Serogroup-specific nodulation restriction has been reported for several B.japonicum strain/host combinations and could possibly be mechanistically related to race specificity systems found in several plant pathogens. In these systems, there is a gene-for-gene relationship between bacterial and plant genotypes. Djordjevic et al. (44) proposed that such a mechanism may be operating in the R. leguminosarum strain TOM/Afghanistan pea symbiosis, where nodulation restriction is determined by a single recessive host gene (45) and a single dominant bacterial gene (19). A conceptually similar gene-for-gene system is most likely functioning to control nodulation of serogroup 123 isolates with these specific soybean genotypes. In these soybean genotypes host-controlled nodulation restriction appears to be, at least partially, conditioned by a single dominant plant genett and a dominant bacterial locus.
ttCregan,
P. B., Keyser, H. H. & Sadowsky, M. J., American Society of Agronomy Annual Meeting, Nov. 27-Dec. 2, 1988, Anaheim, CA, p. 78.
We thank Pamela Prigg and Fuyuhiko Nishijima for technical assistance, Ray Tully for illustrations, and Renee Kosslak, Carol Vance, and Adam Kondorosi for their helpful comments and suggestions. This study was supported in part by U.S. Department of Agriculture Competitive Research Grants 86-CRCR-1-2029 and 8837231-4003 (to M.J.S., P.B.C., and H.H.K.) and 88-37262-3894 and 62-600-1636 (to G.S.), by U.S. Public Health Service Grants GM33494-03 and 40183-01 (to G.S.), by the Swiss Federal Institute of Technology (to H.H.), by the Deutscher Akademischer Austauschdienst (to M.G.), and by the Fundacion Juan March-Madrid (to F.R.-Q.). This is contribution no. 18,524 of the Minnesota Agriculture Experiment Station scientific publication series. 1. Moawad, H. A., Ellis, W. R. & Schmidt, E. L. (1984) Appl. Environ. Microbiol. 47, 607-612. 2. Schmidt, E. L., Zidwick, M. J. & Abebe, H. H. (1986) Appl. Environ. Microbiol. 51, 1212-1215. 3. Gibson, A. H., Dudman, W. F., Weaver, R. W., Horton, J. C. & Anderson, I. C. (1971) Plant Soil Spec. Vol. 1971, 33-37. 4. Cregan, P. B. & Keyser, H. H. (1986) Crop Sci. 26, 911-916. 5. Keyser, H. H. & Cregan, P. B. (1987) Appl. Environ. Microbiol. 53, 2631-2635. 6. Cregan, P. B., Keyser, H. H. & Sadowsky, M. J. (1989) Crop Sci. 29, 307-312. 7. Nieuwkoop, A. J., Banfalvi, Z., Deshmane, N., Gerhold, D., Schell, M. G., Sirotkin, K. M. & Stacey, G. (1987) J. Bacteriol. 169, 2631-2638. 8. Banfalvi, Z., Nieuwkoop, A., Schell, M., Besl, L. & Stacey, G. (1988) Mol. Gen. Genet. 214, 420-424. 9. Gottfert, M., Hitz, S. & Stacey, G. (1990) Mol. Plant-Microbe Interact. 3, 308-316.
Proc. Natl. Acad. Sci. USA 88 (1991)
641
10. Horvath, B., Bachem, C. W. B., Schell, J. & Kondorosi, A. (1987) EMBO J. 6, 841-848. 11. Schofield, P. R. & Watson, J. M. (1986) Nucleic Acids Res. 14, 2891-2903. 12. Spaink, H. P., Okker, R. J. H., Wijffelman, C. A., Pees, E. & Lugtenberg, B. J. J. (1987) Plant Mol. Biol. 9, 27-39. 13. Bachem, C. W. B., Banfalvi, Z., Kondorosi, E., Schell, J. & Kondorosi, A. (1986) Mol. Gen. Genet. 203, 42-48. 14. Deshmane, N. & Stacey, G. (1989) J. Bacteriol. 171, 33243330. 15. Gottfert, M., Grob, P. & Hennecke, H. (1990) Proc. Nail. Acad. Sci. USA 87, 2680-2684. 16. Kosslak, R. M., Bookland, R., Barkei, J., Paaren, H. E. & Applebaum, E. R. (1987) Proc. Natl. Acad. Sci. USA g4, 7428-7432. 17. Sadowsky, M. J., Olson, E. R., Foster, V. E., Kosslak, R. M. & Verma, D. P. S. (1988) J. Bacteriol. 170, 171-178. 18. Long, S. (1989) Cell 56, 203-214. 19. Davis, E. O., Evans, I. J. & Johnston, A. W. B. (1988) Mol. Gen. Genet. 212, 531-535. 20. Heron, D. S., Ersek, T., Krishan, H. B. & Pueppke, S. G. (1989) Mol. Plant-Microbe Interact. 2, 4-10. 21. Cregan, P. B., Keyser, H. H. & Sadowsky, M. J. (1989) Appl. Environ. Microbiol. 55, 2532-2536. 22. Russel, P., Schell, M. G., Nelson, K. K., Halverson, L. J., Sirotkin, K. M. & Stacey, G. (1985) J. Bacteriol. 164, 13011308. 23. Sadowsky, M. J., Tully, R. E., Cregan, P. B. & Keyser, H. H. (1987) Appl. Environ. Microbiol. 53, 2624-2630. 24. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY). 25. Sadowsky, M. J., Cregan, P. B. & Keyser, H. H. (1990) Appl. Environ. Microbiol. 56, 1768-1774. 26. Knauf, V. C. & Nester, E. W. (1982) Plasmid 8, 45-54. 27. Gottfert, M., Lamb, J. W., Gasser, R., Semenza J. & Hennecke, H. (1989) Mol. Gen. Genet. 215, 407-415. 28. Leong, S. A., Ditta, G. S. & Helinski, D. R. (1982) J. Biol. Chem. 257, 8724-8730. 29. Norrander, J., Kempo, T. & Messing, J. (1983) Gene 26, 101-106. 30. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463-5467. 31. Corpet, F. (1988) Nucleic Acids Res. 16, 10881-10890. 32. Stachel, S. E., An, G., Flores, C. & Nester, E. W. (1985) EMBO J. 4, 891-898. 33. Vincent, J. M. (1970) A Manual for the Practical Study qf Root-Nodule Bacteria, IBP Handbook No. 15 (Blackwell Scientific, Oxford). 34. Schmidt, E. L., Bankole, R. 0. & Bohlool, B. B. (1968) J; Bacteriol. 95, 1987-1992. 35. Rostas, K., Kondorosi, E., Horvath, B., Simoncsits, A. & Kondorosi, A. (1986) Proc. Natl. Acad. Sci. USA 83, 17571761. 36. Dodd, I. B. & Egan, J. B. (1987) J. Mol. Biol. 194, 557-564. 37. O'Halloran, T. V., Frantz, B., Shin, M. K., Ralston, D. M. & Wright, J. G. (1989) Cell 56, 119-129. 38. Wang, Y., Moore, M., Levinson, H. S., Silver, S., Walsh, C. & Mahler, I. (1989) J. Bacteriol. 171, 83-92. 39. Laddaga, R. A., Chu, L., Misra, T. K. & Silver, S. (1987) Proc. NatI. Acad. Sci. USA 84, 5106-5110. 40. Hollis, A. B., Kloos, W. E. & Elkan, G. H. (1981) J. Gen. Microbiol. 123, 215-222. 41. Kuykendall, L. D., Roy, M. A., O'Neill, J. J. & Devine, T. E. (1988) Int. J. Syst. Bacteriol. 38,, 358-361. 42. Sadowsky, M. J. & Bohlool, B,-B. (1985) Symbiosis 1, 125138. 43. Hellman, J. D., Wang, Y., Mahler, I. & Walsh, C. (1989) J. Bacteriol. 171, 222-229. 44. Djordjevic, M. A., Gabriel, D. W. & Rolfe, B. G. (1987) Annu. Rev. Phytopathol. 25, 145-168. 45. Holl, F. B. (1975) Euphytica 24, 767-770.
The Bradyrhizobium japonicum nolA gene and its involvement in the genotype-specific nodulation of soybeans (modulation gene/gene-for-gene interactions/symbiosis/plant-bacteria interaction)
MICHAEL J. SADOWSKY*t*, PERRY B. CREGANt, MICHAEL GOTTFERT§, ARUN SHARMA$, DAVID GERHOLD$, FRANCISCO RODRIGUEZ-QUINONESt, HAROLD H. KEYSERII, HAUKE HENNECKE§, AND GARY STACEY¶ *Soil Science and Microbiology Departments, University of Minnesota, St. Paul, MN 54108; tNitrogen Fixation and Soybean Genetics Laboratory, U.S. Department of Agriculture-Agriculture Research Service, BuildingS 011, BARC-West, Beltsville, MD 20705; §Mikrobiologisches Institut, Eidgendssische Technische Hochschule Zentrum, CH-8092 Zurich, Switzerland; ICenter for Legume Research, Department of Microbiology and Graduate Program in Ecology, University of Tennessee, Knoxville, TN 37996; and I1The NifTAL Project, University of Hawaii, P.O. Box 0, Paia, HI 96779
Communicated by Harold J. Evans, October 8, 1990
ABSTRACT Several soybean genotypes have been identified which specifically exclude modulation by members of Bradyrhizobiumjaponicum serocluster 123. We have identified and sequenced a DNA region from B. japonicum strain USDA 110 which is involved in genotype-specifc modulation of soybeans. This 2.3-kilobase region, cloned in pMJS12, allows B. japonicum serocluster 123 isolates to form nodules on plants of serogroup 123-restricting genotypes. The nodules, however, were ineffective for symbiotic nitrogen fixation. The nodulation-complementing region is located approximately 590 base pairs transcriptionally downstream from nodD2. The 5' end of pMJS12 contains a putative open reading frame (ORF) of 710 base pairs, termed nowl. Transposon Tn3-HoHo mutations only within the ORF abolished modulation complementation. The N terminus of the predicted noL4 gene product has strong similarity with the N terminus of MerR, the regulator of mercury resistance genes. Translational IacZ fusion experiments indicated that nolA was moderately induced by soybean seed extract and the isoflavone genistein. Restriction fragments that hybridize to pMJS12 were detected in genomic DNAs from both nodulation-restricted and -unrestricted strains. In soybean production areas of the midwestern United States, members of Bradyrhizobium japonicum serocluster 123 are the dominant indigenous competitors for soybean nodulation (1-3). Cregan and Keyser (4) identified several soybean genotypes [including plant introductions (PIs) and cultivars] which restricted nodulation and reduced the competitiveness of B. japonicum strain USDA 123. These genotypes nodulate normally with strain USDA 110. Three of the genotypes, PIs 377578, 371607, and 417566, were subsequently shown to restrict nodulation by different serocluster 123 isolates (5, 6). In B. japonicum, the essential common nodulation (nod) genes (nodA, nodB, nodC, and nodDi), as well as nodY, nodD2, and nodSUIJ, are closely linked on the chromosome (7-9). Host-specific nodulation (hsn) determinants have also been reported in Rhizobium melioti (10), Rhizobium leguminosarum biovars trifolii (11) and viceae (12), in the broad host-range Rhizobium sp. strain MPIK3030 (13), and in B. japonicum (7, 14, 15). In Rhizobium and Bradyrhizobium, nod, hsn, and other symbiosis genes are induced by flavonoid exudates from the host plant (16-18). While hsn genes have been localized in Rhizobium and Bradyrhizobium, there have been no reports of the identification of genotype-specific nodulation (GSN) genes in bradyrhizobia. The GSN genes would refer to those bacterial sequences which allow nodulation of specific plant genotypes
within a given legume species. One example of a GSN-like gene is nodX. This gene, in R. leguminosarum biovar viceae strain TOM, is not needed for nodulation of commercial pea cultivars, but it is essential for nodulation of the more primitive Afghanistan pea (19). Another GSN-like locus has also been reported in Rhizobiumfredii strain USDA 257 (20). In this report, we describe the identification and sequence analysis** of a gene, nolA, from B. japonicum strain USDA 110, that allows serocluster 123 isolates to nodulate soybean plants having USDA 123-restricting PI genotypes. We show that sequences in this region are induced by the isoflavone genistein and that the predicted amino acid sequence of the deduced open reading frame (ORF) contains a motif that may be similar to the helix-turn-helix motif associated with DNAbinding regulatory proteins.
MATERIALS AND METHODS Bacterial Strains, Plasmids, and Growth Conditions. The B. japonicum strains used in this study were obtained from the U.S. Department of Agriculture-Agricultural Research Service (Beltsville, MD). The nodulation phenotypes of B. japonicum strains on soybean hosts are as previously described (5, 6, 21). Cosmids pRJUT10, pR32, and pR29 were as previously described (22). The B. japonicum strains were maintained and mated on AG medium (23). The Escherichia coli strains were grown on LB medium (ref. 24, p. 440). As required, media were supplemented with tetracycline (Tc; 25-100 Ag/ml), kanamycin (50 pug/ml), spectinomycin (50 ptg/ml), carbenicillin (25 pug/ml), or ampicillin (30 Ag/ml). DNA Biochemistry. Total genomic B. japonicum DNA was isolated and Southern hybridizations were done as described (23). Plasmid DNA was isolated by using the alkaline rapid lysis procedure (ref. 24, pp. 90-91). Cloning and Identification of Nodulation and GSN Genes. The HindIll fragments of cosmid pR32 were subcloned in pVK102 as described (25, 26). Cosmids containing fragments corresponding to the nodDl YABC, nodSUIJ, nodZ, nodD2, or other gene regions (7-9, 25, 27) were used to transform E. coli strain S17-1 (ref. 24, pp. 254-255) (see Fig. 1). Conjugal transfer of pR29, pR32, and pMJS HindIII subclones of pR32 to B. japonicum strains USDA 162 and SD6-lc (serogroup 123) was done by using helper plasmid pRK2073 (28). Transconjugants were selected on AG medium containing Tc (60 Ag/ml). A Sal I partial-digestion genomic DNA library from Abbreviations: PI, plant introduction; GSN, genotype-specific nodulation; ORF, open reading frame; Tc, tetracycline. tTo whom reprint requests should be addressed at: Soil Science Department, University of Minnesota, 246 Borlaug Hall, 1991 Upper Buford Circle, St. Paul, MN 55108. **The sequence reported in this paper has been deposited in the GenBank data base (accession no. M58360).
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
637
Sadowsky et al.
Genetics:
638
Proc. Natl. Acad. Sci. USA 88 (1991)
strain DE3-la was constructed in cosmid pVK102 as described (26). Cosmid clones corresponding to the GSN and nod gene regions were identified by hybridization to the HindIII fragments of pMJS12 and pMJS18 and mated into B. japonicum strain SD6-lc as described above. DNA Sequence Analysis. The 1.9-kilobase (kb) Sma I-Bgl II fragment (see Fig. 1) was digested with the appropriate restriction enzymes and fragments were inserted into the corresponding restriction sites of phage vectors M13mp18 and M13mpl9 (29). The DNA sequence was determined by using the chain-termination method of Sanger et al. (30) and a DNA sequencer (Applied Biosystems model 370A). The UWGCG (University of Wisconsin Genetics Computer Group) software package (release 6.1) was used for computer-aided DNA and protein sequence analysis and the amino acid alignment was done by using the MULTALIN program (31). Data base searches were done by using the TfastA program in conjunction with GenBank and EMBL databases. Construction of lacZ Fusions. Transcriptional/translational IacZ fusions were made in the 1.7- and 2.3-kb HindIII fragments of pR32 (see Fig. 1) using Tn3-HoHol as described by Stachel et al. (32). The location and orientation of inserts were determined by restriction mapping. Fusions were transferred to B. japonicum strains SD6-lc and USDA 110 as described above. fi-Galactosidase Assays. Assays to monitor the induction of lacZ fusions by soybean seed extract and genistein were done as described (8). Strains containing IacZ fusions were induced for 14-16 hr with 2.0 gM genistein (ICN) or 10% (vol/vol) Glycine max cv. Essex seed extract (8). Duplicate cultures were assayed for each treatment and values are in P-galactosidase units (8, 16). Nodulation Studies. Plant assays were done in a Monmouth fine sandy loam soil, containing less than one B. japonicum cell per g of soil (5, 6, 25). Pots were planted with four seeds each of Glycine max PI 377578 [serogroup 123 restricting (4)] and a nonrestrictive host, cv. Williams. Transconjugants (about 10 per mating) and nodulation-positive (MN1-lc, DE3-la, and USDA 110) and -negative (USDA 162, SD6-lc, and USDA 123) controls were inoculated onto at least four replicate plants. Plants were watered with nitrogen-free nutrient solution (5) and harvested 35 days after inoculation, and nodule numbers were determined (4). Nodulationcomplementing transconjugants were isolated from surfacesterilized nodules (ref. 33, p. 7-8) on AG medium containing Tc (100 Ag/ml) or kanamycin (100 ,ug/ml). The identity of B. japonicum transconjugants and nodule isolates was -verified by using strain-specific fluorescent antibodies (34) or spotR
R H H
H H
R H H H
pRJUTIO
R H
R
H
R
pR29 pR32
H.__ D2 4
DG62 DG73 131_ DG66
D1
C
S
U
I
J
DG1
1 11
-I HIRBBg BHBSS H >-~~~^
Sequenced pMJSRegion
IM
pMJS 12
2.3
YA B
-
4
1.7
R
R
H pMJS 18
3.9
RR
1 kb
FIG. 1. Physical-genetic map of B. japonicum strain 1110 nodulation loci. The H locations of the GSN gene nolA, nodDI YABC, nodD2, nodS UIJ, and nodZ on pR32 are shown. The relationship of H subclones pMSJ9, pMJS12, pMJS18, 1 kb :1 pMJS22, and pMJS24 used in complementation and hybridization studies to the physz _ ical-genetic map of nodulation loci on a 17-kb portion ofpR32 are indicated. Arrows denote transcriptional orientation of genes. Numbers below subclone lines are HindIII fragment sizes in kb. The locations and H H orientations of Tn3-4acZ fusions in the 1.7pMJ'S 2 2 and 2.3-kb regions of pR32 are shown by stalked arrows. The region sequenced is denoted by the broken line. Restriction enzyme sites: H, HindIII; B, BamHI; Bg, Bgl II; R, EcoRl; and S, Sma 1. )
H
,lToIA 4
RESULTS Identification of Genes Involved in the GSN of Soybeans. To determine if DNA regions within and clustered around nodDI YABC were involved in the GSN of soybeans, we conjugated pR29 and pR32 (both derived from strain USDA 110) (Fig. 1) into the PI 377578 nodulation-restricted strain SD6-lc (serogroup 123). When the SD6-lc(pR29) and -(pR32) transconjugants were inoculated onto PI 377578, only the pR32 transconjugants nodulated. The SD6-lc(pR32) transconjugants nodulated PI 377578 to an extent equal that of the nodulation-unrestricted strains, MN1-lc and USDA 110 (Table 1) and produced about 30-45 times the number of nodules (range of 90-138 nodules per plant) on the PI genotypes that the control strains SD6-lc or SD6-lc(pVK102) produced. The transconjugants, however, were unstable, since only about 4 out of 10 plate transconjugants were able to nodulate the PI genotypes. While all of the nodule isolates grew on AG medium with Tc and reacted with antiserum specific for serogroup 123 strains, only 7 out of 10 could form nodules when retested on the PI genotypes. Moreover, the SD6lc(pR32) transconjugants formed ineffective symbioses with PI 377578. Nodules were white, suggesting a lack of leghemoglobin production. While some nodules were large and had surface patterning and color typical of soybean nodules, others were small and lacked surface ribs. Since the SD6lc(pR29) transconjugants did not nodulate PI 377578 (data not shown), our results suggested that the DNA region in the area of nonoverlap (about 8.5 kb) was involved in the genotype-specific nodulation of soybeans. To test this, HindIII fragments from pR32 were subcloned in pVK102 (Fig. 1) and individually mated to SD6-lc, and tratsconjugants were tested for nodulation ability. Only the genomic DNA region contained in plasmid pMJS12 (2.3 kb) allowed SD6-lc to nodulate the restrictive soybean genotype (Table 1). The SD6-lc(pMJS12) transconjugants produced about 35 times more nodules (range of 97-120 nodules per plant) than the uncomplemented parent strain, SD6-lc(pVK102), and were also ineffective for nitrogen fixation on the PI genotypes. The GSN region is located about 590 bp transcriptionally downstream from the end of nodD2 (Fig. 1). Sequences present on pMJS12 did not hybridize to B. japonicum nodDI YABC' (pMJS18), R. meliloti hsnABCD, or Nod box consensus sequence gene probes (7, 35). The other SD6-lc
H
H
,
blot immunoassays (21). The USDA 162 and SD6-lc transconjugants isolated from nodules were retested for nodulation ability as described above.
pMJS24 5.6
Genetics:
Sadowsky et al.
Table 1. Nodulation of Glycine max PI 377578 by the serogroup 123 strain SD6-1c and its pR32 and pMJS12 transconjugants Nodules on Nodules on cv. Williams PI 377578 Dry wt, Dry wt, No. mg Isolate No. mg 1.0 7.0 3.5 14.0 Uninoculated 114.7 453.0 4.0 26.0 SD6-1c 2.8 9.0 133.2 361.0 SD6-lc(pVK102) 139.0* 111.8 111.0 113.5* SD6-lc(pR32) 153.8 210.0 109.8* 194.0* SD6-lc(pMJS12) 96.7 284.0 116.7 290.0 MN1-1c 215.0 142.8 351.0 107.75 USDA 110 Nodule number and dry weight are per plant. Values for transconjugants are means of 8-24 replicates. *Significantly different from the uninoculated or negative controls (P < 0.05) by Duncan's new multiple range test.
transconjugants (containing pMJS9, pMJS18, pMJS22, or pMJS24) failed to nodulate the restricting PI genotype (data not shown). Strains SD6-lc(pVK102), SD6-lc(pR32), and SD6-lc(pMJS12) nodulated and fixed nitrogen normally with the nonrestrictive commercial soybean genotype Williams (Table 1). A majority (6 out of 10) of SD6-lc(pMJS12) nodule-isolated transconjugants retained the ability to nodulate PI 377578, suggesting that pMJS12 most likely contributed to the formation of these nodules. Plasmid pMJS12 also allowed another serogroup 123 isolate, USDA 162, to nodulate PI 377578 (range of 126-145 nodules per plant), suggesting that its involvement in GSN is not limited to a particular serogroup strain. To determine whether a functional nodulation complementing GSN DNA region was unique to strain USDA 110, we made a genomic DNA library from strain DE3-la [nodulation unrestricted, serogroup 127 (5, 25)] and mated cosmid clones which hybridized to the pMJS12 gene probe to strain SD6-lc. One of the transconjugants, SD6-lc(pFR27), was found to nodulate PI 377578 (mean of 22 nodules per plant) to the same extent as DE3-la. Cosmid pRF27 contains sequences analogous to those extending from the USDA 110 nodZ gene to the end of the GSN gene region (data not shown). A pFR27 subclone, pFR2501, containing only the DE3-la GSN region, was later found to allow SD6-lc to nodulate the PI genotypes (mean of 21 nodules per plant), suggesting that other unrestricted strains contain functional GSN gene regions. The DE3-la GSN region also failed to complement nitrogen fixation effectiveness in the SD6-lc transconjugants. In several cases, SD6-lc(pFR2501) transconjugants were also ineffective on both permissive (cv. Williams) and restrictive soybean hosts. The DE3-la-derived GSN region was more unstable than the homologous 110 region, and cosmids containing the GSN region mated to recipients at frequencies 1/10th to 1/100th of those with DNA from other regions (10-5 to 1O-7 per recipient), and 60-100%o of transconjugants contained cosmids with deletions in the GSN gene region. The instability phenomenon also apparently prevented us from making site-directed mutations in the GSN gene region. While strain SD6-lc is restricted for nodulation of PI 377578, it still forms a small number (Table 1) of crown nodules on this PI genotype (5). To determine if these nodules were produced by a nodulation-competent subpopulation of serogroup 123 strains, nodule isolates (16 and 19 from cv. Williams and PI 377578, respectively) of strain USDA 123 were inoculated onto PI 377578. Nodulation by these isolates was not significantly different (P < 0.05) than by the USDA 123 parent. Thus, the observed "leakiness" of the restrictive nodulation phenotype is not due to a phenotypically stable
Proc. Natl. Acad. Sci. USA 88 (1991)
639
nodulation-competent subpopulation of serogroup 123 isolates. Expression of the GSN Region. To determine if sequences in and around pMJS12 are induced by host plant-derived signal compounds, we constructed Tn3-lacZ fusions in the 1.7- and 2.3-kb fragments of pR32 (Fig. 1) and studied their induction (in strain USDA 110) with soybean seed extract and genistein. Only fusion DG62 was induced by soybean seed extract (95 ± 5 B-galactosidase units) or genistein (89 ± 6 units). This fusion is in the 2.3-kb HindIII fragment contained in pMJS12 and maps within the GSN ORF described below. Fusion DG62 also had a small amount of 3-galactosidase activity (35 ± 7 units) in the absence of inducer, suggesting that this region is constitutively transcribed at a low level. None of the other lacZ fusions examined (DG1, DG6, or DG73) were induced by seed extract or genistein (mean of 5.6 ± 1 units). These results suggest that transcription of pMJS12 is in the same direction as nodD2 and is independent of transcripts originating from adjacent upstream regions. DNA Sequence Analysis. To determine the coding capacity of the GSN gene region and to analyze the DNA region downstream of nodD2, we sequenced the 1.9-kb Sma I-Bgl II fragment of pR32 (see Fig. 1). This fragment contains the 5' end of the 2.3-kb HindIII fragment cloned in pMJS12 (beginning at nucleotide 630) and areas 3' of the end of the nodD2 transcript (27). We identified one ORF (nucleotides 959-1669) within the sequenced region on the basis of potential start codons (Fig. 2). To determine if the 710-bp ORF was involved in genotype-specific nodulation, we tested Tn3-HoHo insertion mutations within or downstream of the ORF (fusions DG62 and 131, respectively; see Fig. 1) for the production of nodules on PI 377578. Only SD6-lc transconjugants with an insertion mutation within the ORF of pR32 (DG62) failed to nodulate, suggesting that the ORF gene product is most likely involved in the extended host-range phenotype. We have named this sequence nolA. Amino Acid Sequence of the Deduced nolA Gene Product. The putative translational start site at the first available ATG codon of the ORF, at position 956, results in a protein with a predicted molecular weight of 26,871. The predicted NolA protein product would have a fairly high arginine content (14.3%), which would result in a net positive charge. Since this positive charge might facilitate an interaction of the protein with the negatively charged DNA backbone, we looked for a helix-turn-helix motif, which is characteristic of many DNA-binding proteins. Using the method of Dodd and Egan (36), we identified such a motif in amino acids 13-32 (Fig. 2), and we suggest that nolA may code for a DNAbinding regulatory protein. A data base search revealed that the N terminus of NoIA is similar to the N terminus of MerR. MerR has been shown to be a metalloregulatory DNAbinding protein that activates or represses transcription of mercury resistance genes (37). The alignment of NolA to the MerR proteins of Bacillus sp. strain RC607 and Staphylococcus aureus (38, 39) is shown in Fig. 3. While the homology between the two MerR proteins extends to the very end of both proteins, the homology between NolA and MerR is restricted to the first 70 amino acid residues. However, since this is the region in which the helix-turn-helix motif is located, NolA and MerR could share a mode of regulating gene
expression. Conservation of GSN Gene Region in B. japonicum Strains. Since results from previous studies indicated that pMJS12 hybridized to genomic DNA from nodulation-restricted and -unrestricted serocluster 123 strains (25), it was of interest to determine if the GSN gene region was present in other B. japonicum isolates. The pMJS12 gene probe hybridized to genomic DNA from several serologically distinct B. japonicum strains. While pMJS12 hybridized predominantly to 6.1-kb EcoRI fragments from B. japonicum strains in DNA
Genetics:
640
Sadowsky et al.
Proc. Natl. Acad. Sci. USA 88 (1991)
1
CCCGGGACATTTCGTGCGATCCGTCTCACCGTTGCCGTAAGCACATTCGCGAAGTCTACGAACCTCACATTCCAACACTTGTCTCAACAC
91
ACAATTCCTGCTCGCCTGGCTTGCTCTTATTGCCGGGGGAAAGGATCCTTGCGGCAGGCAAGAGGCGCTTTCGCTCTTCGCACCTCCGCC
181
TCGGGGCGAGAAACGGCAGACTGCCGACTTGATGCTTTTAGGGACGGGAACGGAACGGGGCCGGCGAGCGACGGAGCAATGCCTTCGCA
271
CTGAAGCTAGGCGTTTACGTTGCAACGGCCGCGGCAGGGCGCAGCTCTCTCAAGuAGGTAGGCCAGGCACGGAGCTCTTGCCGTCAACTC
361
CCTGGGTCGGTAGTTGCTCCCGATGAACGGAGATGATCAGCTTATTCGGATTCCTACCACCGTTCATTGTGAGCATAGAGATCGTCGGAC
451
TTGCCGTGCCGTCGAAGCTAGCATCGCGAACGCGTGTGGCCTGAACGCATCAC
CACATCACATCCTCATCGCTTTGCGCCGCA
541
GGTATGACTTCTTTTGGCCCTACCGCAGCCAAACGTCTTGACGACGTGCGCTGCGCTTGACTCAATTTCAAACAAGGAGCTGAACGTTGA
631
AGCTTTGCAGCGGGGTGTTCGATCCAGAAGAGCTATCTGCCCTCGGACGGCTTTATGACGGCGCCGTCGACGCGCTACCTCCGTCGATGC
721
GGAGTCCAGAAAACTGCCAGCAATCGCCAAACTTATCCTGGAGCGTACGGCGGCTGGCGAGGCTGAACTGGCACGCCTGGCGAACTTGT
811
TGATCACCCTTTCTCCTGAGGGTTGATCAGCCGCCATTCTAAGGATAGTCACCACTCGCGCGCATGAATTAAAGGCGATTCAACTGGAGC
901
TTAGAAATCCGCTTGAATCATACGTCACGTCAGGTTGTAGGCTTTGAAATTGAACAACATGAACAGAGCTACACCAAGACGGCGTCGATG
991
GCGCATTGGCGAACTTGCGAGGCGACCGG>AACGGTGCGCACGCTGCACCATTATGAGCACACTGGACTGCTAGCAGCGACAGAGCG
N R
M
R.| I
1081
a
T
G
V
T
V
R
T
L
H
H
Y
Z -H- |T
G
L
L
A
A
R T
R W E
R
E G
G
H
R
M
Y
R E
D
S
G
Q
R
V
Q
H
I
A
R
L
R E
L
G
F
S
L
V
I
R K A
M E
G
T
T
S
L
T
D
L
L
R K H
L E R
I
B
V
Q V A R
T
L
L
R
D
R
L
R
N
M
T
I
D
S
E
A
Q
V
S
D
V
E
L
P
A
T
L
N
A
M
R
A
E
T
R
S
Q
T
S
R
C
T
C
N
L
A
A
E
R
E
D
R W R
R
I
R
D
D
TCTTCGCGACTGCATGGATGGGGGCGAGCACCCTTGCGGAGAGCGCGCAAGGCTGTCGCGGTTGCAGCACGCTTGCTGATCAGCGAAAT L
1531
E
R
GTCGCGCGCCGAGACGCGGAGTCAGACGTCCCGGTGCACCTGCAATTTAGCTGCGG&GCGAGAGGACCGCTGGCGGCGAATCCGTGATGA S
1441
A
R
AACTCTGTTGCGCGACCGTTTGCCGACATGACCATCGACAGCGAGGCGCAGGTAAGTGTGGATGAGCTGCCTGCCQCCCTGAACGCCAT T
1351
L
P
GGAGATCCGTAIAGCTATGGAGGGGACGACCTCACTCACGGACTTGCTGCGCAAACATCTAGAGCGCATAGAAGTCCAAGTCGCGCGCAC E
1261
E
T
CACTGAGGGCGGTCACCGGATGTATGACCGCGAAAGCGGGCAACGGGTTCATCAGATTCGCGCGCTGCGTGAGCTTGGCTTCTCGCTCGT T
1171
G
A
R
D
C N D
G
G
E
H
P
C
G
E
R
A
K
A
V
A
V
A
A
L
R
L
I
S
E
I
CGCCGGCGACGATTCACGTGTTTCGATGATCCTGAAGGTACTTGCACGATTAAGCGCGCCCCGCAGTCTGGCTGGTTGGGATCCCTGTCT A
G
D
D
S
R
V
S
M
I
L K V
L
A
R
L
S
A
P
R
S
L
G W D
A
P
C
L
1621
AATGCAATATCTCGATCTTGCCCTTGGCGGATTGGAGGTCAGCCTTACTGACGCATTCATGGCGGGTTCCACACGCCGGTTAGGGCAAG
1711
CTAAGAAACGCTAGGTGAACCGGCCGCTCTCCGACGCGGCTTGGAGTTTTACGGAATCAAACGGAAGGATTGTCACCGCCGGGCGATCCG
1801
CGTCCCTGACGCCTCCATCGCTCTTCCCA^AAGGGCTTGATCCGCACGTAACGTCAAGTTGTAGATCT
M
Q
Y
L
D
L
A
L
G
G
L
E
D
Q
P
Y
homology (40) [or fatty acid (41)] groups I and IA (USDA 4, 6, 38, 62, 110, 122, and 124), little or no hybridization was observed with DNAs from strains USDA 31, 46, 61, 76, 94, 130, or 135 (data not shown). With these strains, however, there was no apparent relationship between hybridization profiles and nodulation restriction. A GSN gene probe did not hybridize to genomic DNA from R. leguminosarum biovars trifolii and viceae strains or from Agrobacterium tumefaciens (not shown).
conceptually different than Bradyrhizobium and Rhizobium hsn genes (9-13, 18), which are involved in host-range specificity at the legume genus level. In contrast to other Bradyrhizobium nod (7, 27) and Rhizobium nod and hsn (35) gene loci, the nolA gene locus did not hybridize or have sequence similarity to the 25-mer nod box consensus sequence (7, 27), which has been implicated to have a role in the coordinated regulation of nodulation genes (35). This suggests that nolA, and several other B. japonicum nodulation genes, may be regulated by nod-box-independent promoters (7, 27). While nolA allows serogroup 123 strains to nodulate the PI genotypes, complementation is not complete and transconjugants fail to fix nitrogen. Also, while most SD6-lc transconjugants fix nitrogen with cv. Williams, the nodules produced are smaller than the wild-type parent. In some extreme cases, nolA totally interfered with the nitrogen fixation ability of SD6-lc(pFR27) transconjugants on the permissive soybean host. We are not sure why the nolA locus interacts in this manner, but there may be physical or functional incom-
DISCUSSION Using a nodulation-complementation approach, we have identified a B. japonicum USDA 110 gene, nolA, which is involved in the GSN of soybeans. The nolA gene, cloned in pMJS12, allows serogroup 123 isolates to nodulate soybean plants with restrictive genotypes. Since sequences present on this fragment allow for the nodulation of plants with a specific genotype within a legume species, we have designated this a GSN (genotype-specific nodulation) locus. This locus is SaMerR
1 10 20 30 40 50 MGMKISELAKACDVNKETVRYYERKGLIAGPPRNESGYRIYSEETADRVRFI
FIG. 3. Amino acid sequence of the NolA protein of B. alignment ****** ** * ** ** **** * *** * * * japonicum (BjNolA), MerR of BaMNRATPRRRRWRIGELAEATGVTVRTLHHYEHTGLLAATERTEGGHRMYDRESGQRVHQI cillus sp. strain RC607 [BsMerR 60 70 80 90 (38)], and MerR of Staphylococcus 100 110 KRMKELDFSLKEIHLLFGVVDQDGERCKDMYAFTVQKTKEIERKVQGLLRIQRLLEELKE aureus [SaMerR (39)]. Further de*** ** * * ** * **** * * *** ** * * ** * ** * * *** tails are given in the text. Identical KRMQELGFTLNEIDKLLGVVDRDEAKCRDMYDFTILKIEDIQRKIEDLKRIERMLMDLKE **** * ** * * * amino acid residues are marked by RALRELGFSLVEIRKAMEGTTSLTDLLRKHLERIEVQVARTTLLRDRLRNMT IDSEAQVS asterisks. Dots at the end of the NolA sequence indicate that amino 120 135 acids extend farther but the seKCPDEKAMYTCPIIETLMGGPDK quence is not homologous. The puRCPENKDIYECPIIETLMKK tative helix-turn-helix motif is underlined. VDELPATLNAMSRAETRSQTSRC... *
BsMerR
BjNolA SaMerR BsMerR
BjNolA SaMerR BsMerR
BjNolA
1868
FIG. 2. Nucleotide sequence of the nolA gene region. The sequence starts with the Sma I site at the end of nodD2 (see Fig. 1). The HindIlI site at the start of pMJS12 (position 630) is underlined. The deduced amino acid sequence of the nolA gene product is given in one-letter code. The putative helix-turn-helix motif at the N terminus of the NolA protein is boxed.
*
***
*
*****
*****
***
*
*
*
***
**
*
**
**
MKFRIGELADKCGVNKETIRYYERLGLIPEPERTEKGYRMYSQQTVDRLHFI
Genetics:
Sadowsky et al.
patibility between nolA and the indigenous nodulation and nitrogen fixation genes (42). This instability may also be reflected in the low transfer frequencies we obtained with nolA-containing cosmids, the presence of deletions in the nolA region of transconjugants, and our inability to construct stable mutations in the strain USDA 110 nolA locus. Though the function of the nblA gene is unknown, the N terminus of its product shows similarity to the N terminus of the MerR proteins from Bacillus (38, 43), Staphylococcus (31), and several Gram-negative bacteria. NolA and these proteins all possess helix-turn-helix motifs which are characteristic of many DNA-binding proteins. Thus, NolA and MerR may all be members of a unique family of gene regulators. Our results indicate that the genetic components of both partners interact to influence nodulation and competition in the soybean/Bradyrhizobium symbiosis. Serogroup-specific nodulation restriction has been reported for several B.japonicum strain/host combinations and could possibly be mechanistically related to race specificity systems found in several plant pathogens. In these systems, there is a gene-for-gene relationship between bacterial and plant genotypes. Djordjevic et al. (44) proposed that such a mechanism may be operating in the R. leguminosarum strain TOM/Afghanistan pea symbiosis, where nodulation restriction is determined by a single recessive host gene (45) and a single dominant bacterial gene (19). A conceptually similar gene-for-gene system is most likely functioning to control nodulation of serogroup 123 isolates with these specific soybean genotypes. In these soybean genotypes host-controlled nodulation restriction appears to be, at least partially, conditioned by a single dominant plant genett and a dominant bacterial locus.
ttCregan,
P. B., Keyser, H. H. & Sadowsky, M. J., American Society of Agronomy Annual Meeting, Nov. 27-Dec. 2, 1988, Anaheim, CA, p. 78.
We thank Pamela Prigg and Fuyuhiko Nishijima for technical assistance, Ray Tully for illustrations, and Renee Kosslak, Carol Vance, and Adam Kondorosi for their helpful comments and suggestions. This study was supported in part by U.S. Department of Agriculture Competitive Research Grants 86-CRCR-1-2029 and 8837231-4003 (to M.J.S., P.B.C., and H.H.K.) and 88-37262-3894 and 62-600-1636 (to G.S.), by U.S. Public Health Service Grants GM33494-03 and 40183-01 (to G.S.), by the Swiss Federal Institute of Technology (to H.H.), by the Deutscher Akademischer Austauschdienst (to M.G.), and by the Fundacion Juan March-Madrid (to F.R.-Q.). This is contribution no. 18,524 of the Minnesota Agriculture Experiment Station scientific publication series. 1. Moawad, H. A., Ellis, W. R. & Schmidt, E. L. (1984) Appl. Environ. Microbiol. 47, 607-612. 2. Schmidt, E. L., Zidwick, M. J. & Abebe, H. H. (1986) Appl. Environ. Microbiol. 51, 1212-1215. 3. Gibson, A. H., Dudman, W. F., Weaver, R. W., Horton, J. C. & Anderson, I. C. (1971) Plant Soil Spec. Vol. 1971, 33-37. 4. Cregan, P. B. & Keyser, H. H. (1986) Crop Sci. 26, 911-916. 5. Keyser, H. H. & Cregan, P. B. (1987) Appl. Environ. Microbiol. 53, 2631-2635. 6. Cregan, P. B., Keyser, H. H. & Sadowsky, M. J. (1989) Crop Sci. 29, 307-312. 7. Nieuwkoop, A. J., Banfalvi, Z., Deshmane, N., Gerhold, D., Schell, M. G., Sirotkin, K. M. & Stacey, G. (1987) J. Bacteriol. 169, 2631-2638. 8. Banfalvi, Z., Nieuwkoop, A., Schell, M., Besl, L. & Stacey, G. (1988) Mol. Gen. Genet. 214, 420-424. 9. Gottfert, M., Hitz, S. & Stacey, G. (1990) Mol. Plant-Microbe Interact. 3, 308-316.
Proc. Natl. Acad. Sci. USA 88 (1991)
641
10. Horvath, B., Bachem, C. W. B., Schell, J. & Kondorosi, A. (1987) EMBO J. 6, 841-848. 11. Schofield, P. R. & Watson, J. M. (1986) Nucleic Acids Res. 14, 2891-2903. 12. Spaink, H. P., Okker, R. J. H., Wijffelman, C. A., Pees, E. & Lugtenberg, B. J. J. (1987) Plant Mol. Biol. 9, 27-39. 13. Bachem, C. W. B., Banfalvi, Z., Kondorosi, E., Schell, J. & Kondorosi, A. (1986) Mol. Gen. Genet. 203, 42-48. 14. Deshmane, N. & Stacey, G. (1989) J. Bacteriol. 171, 33243330. 15. Gottfert, M., Grob, P. & Hennecke, H. (1990) Proc. Nail. Acad. Sci. USA 87, 2680-2684. 16. Kosslak, R. M., Bookland, R., Barkei, J., Paaren, H. E. & Applebaum, E. R. (1987) Proc. Natl. Acad. Sci. USA g4, 7428-7432. 17. Sadowsky, M. J., Olson, E. R., Foster, V. E., Kosslak, R. M. & Verma, D. P. S. (1988) J. Bacteriol. 170, 171-178. 18. Long, S. (1989) Cell 56, 203-214. 19. Davis, E. O., Evans, I. J. & Johnston, A. W. B. (1988) Mol. Gen. Genet. 212, 531-535. 20. Heron, D. S., Ersek, T., Krishan, H. B. & Pueppke, S. G. (1989) Mol. Plant-Microbe Interact. 2, 4-10. 21. Cregan, P. B., Keyser, H. H. & Sadowsky, M. J. (1989) Appl. Environ. Microbiol. 55, 2532-2536. 22. Russel, P., Schell, M. G., Nelson, K. K., Halverson, L. J., Sirotkin, K. M. & Stacey, G. (1985) J. Bacteriol. 164, 13011308. 23. Sadowsky, M. J., Tully, R. E., Cregan, P. B. & Keyser, H. H. (1987) Appl. Environ. Microbiol. 53, 2624-2630. 24. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY). 25. Sadowsky, M. J., Cregan, P. B. & Keyser, H. H. (1990) Appl. Environ. Microbiol. 56, 1768-1774. 26. Knauf, V. C. & Nester, E. W. (1982) Plasmid 8, 45-54. 27. Gottfert, M., Lamb, J. W., Gasser, R., Semenza J. & Hennecke, H. (1989) Mol. Gen. Genet. 215, 407-415. 28. Leong, S. A., Ditta, G. S. & Helinski, D. R. (1982) J. Biol. Chem. 257, 8724-8730. 29. Norrander, J., Kempo, T. & Messing, J. (1983) Gene 26, 101-106. 30. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463-5467. 31. Corpet, F. (1988) Nucleic Acids Res. 16, 10881-10890. 32. Stachel, S. E., An, G., Flores, C. & Nester, E. W. (1985) EMBO J. 4, 891-898. 33. Vincent, J. M. (1970) A Manual for the Practical Study qf Root-Nodule Bacteria, IBP Handbook No. 15 (Blackwell Scientific, Oxford). 34. Schmidt, E. L., Bankole, R. 0. & Bohlool, B. B. (1968) J; Bacteriol. 95, 1987-1992. 35. Rostas, K., Kondorosi, E., Horvath, B., Simoncsits, A. & Kondorosi, A. (1986) Proc. Natl. Acad. Sci. USA 83, 17571761. 36. Dodd, I. B. & Egan, J. B. (1987) J. Mol. Biol. 194, 557-564. 37. O'Halloran, T. V., Frantz, B., Shin, M. K., Ralston, D. M. & Wright, J. G. (1989) Cell 56, 119-129. 38. Wang, Y., Moore, M., Levinson, H. S., Silver, S., Walsh, C. & Mahler, I. (1989) J. Bacteriol. 171, 83-92. 39. Laddaga, R. A., Chu, L., Misra, T. K. & Silver, S. (1987) Proc. NatI. Acad. Sci. USA 84, 5106-5110. 40. Hollis, A. B., Kloos, W. E. & Elkan, G. H. (1981) J. Gen. Microbiol. 123, 215-222. 41. Kuykendall, L. D., Roy, M. A., O'Neill, J. J. & Devine, T. E. (1988) Int. J. Syst. Bacteriol. 38,, 358-361. 42. Sadowsky, M. J. & Bohlool, B,-B. (1985) Symbiosis 1, 125138. 43. Hellman, J. D., Wang, Y., Mahler, I. & Walsh, C. (1989) J. Bacteriol. 171, 222-229. 44. Djordjevic, M. A., Gabriel, D. W. & Rolfe, B. G. (1987) Annu. Rev. Phytopathol. 25, 145-168. 45. Holl, F. B. (1975) Euphytica 24, 767-770.
