Copyright 0 1989 by the Genetics Society of America
A Family of Activator Genes Regulates Expressionof Rhixobium meliloti Nodulation Genes
John T. Mulligan and Sharon R. Long Department of Biological Sciences, Stanford University, Stanford, Calfornia94305-5020 Manuscript received August 29, 1988 Accepted for publication January 27, 1989
ABSTRACT Nodulation (nod) gene expression in Rhizobium meliloti requires plant inducers and the activating protein product of the nodD gene. We have examined three genes in R. meliloti which have nodD activity and sequence homology. These three nodD genes are designated nodD1, nodD2 and nodD3, and have distinctive properties. The nodDl gene product activates expression of the nodABC operon, as measured by a nodC-lacZ fusion or by transcript analysis, in the presence of crude seed or plant wash or the inducer, luteolin. The nodD3 gene product can cause a high basal (uninduced) level of nodC-lacZ expression and nodABC transcripts which is relatively unaffected by inducers. The effect of nodD3 is dependent on the presence of another gene, syrM hmbiotic yegulator). By primer extension analysis we determined that the transcription start site is the same for nodDl plus luteolin or nodD3- syrM mediated expression of nodA and nodH mRNAs. syrM also enhances the expression of another symbiotically important trait, production of extracellular polysaccharide. This regulatory effect of syrM requires locus syrA, which is linked to nodD3 and syrM. The syrM-syrA mediated increase in polysaccharide production requires at least some of the previously identified exo genes and may be a parallel regulatory event to the syrM-nodD3 control of nod promoters.
ACTERIA in the genus Rhizobium invade specific tion is under the control of an additional set of sym- B host plants and stimulate the development of biotic genes, the exo genes, are also requiredfor nitrogen-fixing root nodules. This complex and highly normal nodule development(FINAN et al. 1985; CHAK- specific process occurs between a given species of RAVORTY et al. 1982; CHENet al. 1985; BORTHAKUR Rhizobium and a set of plants which defines its host et al. 1986; LEIGH, SIGNERand WALKER1985; DYLAN range. Rhizobium meliloti establishes a productive sym- et al. 1986. Most exo genemutations of R. meliloti biosis with alfalfa (Medicago sativa)and related plants. (deficient in acidic exopolysaccharides) block the nor- The early events in the development of nodules in- mal bacterial invasion of the plant root. When alfalfa clude bacterial stimulation of celldivisions in the plant is inoculated with exo- strains, nodules develop but root cortical layer, giving rise to a new meristem, and few or no infection threads are formed, fewif any bacterial invasion of epidermal root hairs and under- bacteria enter the nodules, and no nitrogen fixation lying cells (BAUER198 1; NEWCOMB198 1). occurs (FINANet al. 1985). At least three sets of R. meliloti genes are involved A family of regulatory genes, which are referred to in mediating the early stages of alfalfa nodule forma- as nodD, has been identified in all Rhizobium species tion. One set, nodABC, is present inall Rhizobium so far examined (MULLIGANand LONG 1985; EGEL- species so far examined, and mutationsin these genes HOFF and LONG1985; ROSSENet al. 1985; GOTTFERT lead to a Nod- phenotype: the bacteria neither stim- et al. 1986; SPAINKet al. 1987a, b; SHEARMANet al. ulate cortical cell divisions, nor deform orinvade root 1986). Although some Rhizobium species carry only a hairs (ROSTASet al. 1986; DEBELLEet al. 1986; DUD- single copy of nodD (DOWNIEet al. 1984; SCHOFIELD LEY, JACOBSand LONG 1987; JACOBS,EGELHOFF and and WATSON1985), many species carry two or three LONG 1985). The members of a second set of nodu- similar but not identical copies of nodD (RODRIGUEZ- lation genes (in R. meliloti, nodFE, nodH, nodG, nodPQ QUINONESet al. 1987; APPELBAUMet al. 1988). and others) are notnecessarily present in or function- Expression of nodABC, nodFE and nodH is induced in ally conserved among all Rhizobium species; these af- the presence of an exudate of legume seeds or roots; fect the rate and frequency of nodule formation and this activation requires a nodD gene product. R. meli- influence host range (HORVATHet al. 1986; DJORD- loti has three copies of nodD, designated nodDl, nodD2 JEVIC, SCHOFIELDand ROLFE 1985; DEBELLEet al. and nodD3 (PUTNOKYand KONDOROSI1986; GOTT- 1986; RODRIGUEZ-QUINONESet al. 1987; SWANSONet FERT et al. 1986; HONMAand AUSUBEL1987). al. 1987; SCHWEDOCKand LONG 1989). The flavone, luteolin, isolated from alfalfa seed The Rhizobium exopolysaccharides, whose produc- wash, is the most active inducer of R. meliloti nod
Genetics 122: 7-18 (May, 1989) 8 J. T. Mulligan and S. R. Long genes when nodDl is overexpressed (PETERS,FROST of pRmJT5 carrying an insertion of Tn5 which was pro- and LONG 1986).A wide variety of flavones and duced as described in SWANSONet al. (1987) and screened as described in RESULTS. The construction of pRmM 113 flavanones are present in legume exudates, some of involved two steps; first, pRmS507 was digested with ClaI which are active in inducing nod gene expression,and and the fragment bearingthe Tn5 insertion was cloned into others in blocking induction (DJORDEVICet al. 1987; the ClaI site of pBR322. The resulting plasmid was digested REDMONDet al. 1986; FIRMINet al. 1986; ZAAT et al. with Sal1 and XhoI, and the fragment carrying syrM and a 1987; PETERSand LONG 1988). Several lines of evi- segment of Tn5 which encodes neomycin resistance was ligated into pLAFR3 digested with BamHI, by the method dence indirectly suggest that the nodD gene products of partial fill-in of each cohesive end (ZABROVSKYand AL- interact with the inducing molecules. First, the substi- LIKMETS 1986), toform pRmM 1 13.pRmS5 1 1was digested tution of the naturally occurring nodD gene of one with XhoI and religated at low concentrationto form Rhizobium species for the nodD gene of another species pRmM 136, which lacksthe sequences between the Tn5 and in otherwise isogenic strains can determine the range the left-most XhoI site of pRmJT5. pRmM137 (FISHERet al. 1988) was digested with XbaI, partially filled-in and ligated of compounds which will induce expression of the nod to a partially filled-in Hind111 fragment which carries the genes (SPAINKet al. 1987a).Second, a mutation in Tn5 neomycin resistance gene to formpRmM 14 1. pRmJT5 nodD can broaden the spectrum of plant or synthetic was digested with XbaI, partially filled-in and ligated to a compounds which will induce nod gene expression partially filled in Hind111 fragment carrying the uidA gene (BURN,ROSSEN and JOHNSTON 1987). Third, the spe- and spectinomycin resistance to form pRmM142. A 2-kb BglII fragment from pRmS5 1 1was purified and cloned into cies source of nodD in a transgenic construct in some the BamHI site of pUCll9to create pRmM151 and cases influences or determines the host range of the pRmM152. A 2.1-kb ClaI fragmentfrom pRmS511 was Rhizobium strain (SPAINKet al. 1987b; HORVATHet al. cloned into theAccI site of pUCll9 toestablish pRmM 147. 1987; GYORGYPAL, IYERand KONDOROSI 1988). How- Strain construction: Marker exchange was carried out as ever, there are as yet no direct biochemical demon- described by JACOBS,EGELHOFF and LONG (1985) except that pR751 was used to exclude the IncP plasmids. Trans- strations of NodD-inducer interaction.The nodD gene ductions were carried out using phage N3 as described in product, NodD, appears to beDNA-binding a protein MARTINand LONG(1 984).JM 139 was produced by marker by the criterion thatNodD-containing extracts (HONG exchange of the spectinomycin resistance insertion in et al. 1987) and substantially purified NodD (FISHER pRmM139 into R. meliloti 1021. JM80 and JM86 were et al. 1988) affect the electrophoretic mobility of nod produced by marker exchange of the Tn5 insertions from pRmS303 and pRmS701, respectively, into JM6l. JM81 gene promoters. Competition studies show that this and JM83 were produced by marker exchange of the Tn5 binding occursat least partially at thenod box (FISHER insertions from pRmS9B7 and pRmS7O 1, respectively, into et al. 1988), a conserved sequence found upstream of JM57.JM85 was produced by markerexchange of the each inducible nod operon (ROSTASet al. 1986; SCHO- neomycin resistance insertion of pRmM141 intoJM57. FIELD and WATSON 1986; SPAINK et al. 1987b). Tran- JM88, JM204 and JM207 were produced by transduction of the Tn5 insertions from JT303, JT801 and JT709, scription of the inducible nod genes initiates 26-28 bp respectively, into JM57. JM216 was produced by transduc- downstream from the nod box (FISHERet al. 1988). tion of the Tn5 insertion from JT80 1 into JM6 1. JM90 was In this study, we examined the regulatory proper- produced by cotransduction of the nodD3 Tn5 insertion and ties of the threeR. meliloti nodD genes. We found that the nodD1-lac2 fusion of JM80 into JM139. JM96 was pro- while each of the nodD genes is capable of affecting duced by cotransduction of the syrM Tn5 insertion and the nodD1-lacZ fusion of JM86 into JM139. JM201 was pro- the expression of a node-lac2 fusion, the three genes duced by cotransduction of the nodD3 Tn5 insertion and are not equivalent. the nodC-lacZ fusion of JM88 into JM139. The described nodD? requires a second gene, syrM (symbiotic reg- insertions in all of the strains were confirmed by SOUTHERN ulator), for its activating function, but is not depend- blot analysis. ent on flavones to activate expression of nodC-lacZ. Assays: @-galactosidaseassays were carried out as de- scribed by MILLER(1 972) with the modifications previously The transcription start sites for both nodA and nodH described (MULLIGANand LONG1985). Four replicate sam- are unchanged, whether their expression is mediated ples were assayed for each condition and each strain. The by nodDl and luteolin, or by nodD? and syrM. syrM @-galactosidaseunits reported in the text and tables are the can affect both the expression of the nod genes and average of the results from the four assays. The results of the expression or activity of the exo genes required the individual assays are all within 15% of the reported averages. The endogenous @-galactosidaseactivity was 2 for invasion of the plant roots, and thus may play a units under all the conditions tested. Seed washwas pre- key role in aregulatory systemwhich coordinates pared as described by MULLIGANand LONG(1985). Nodu- several symbiotic functions. lation tests were carried out as described by JACOBS, EGEL- HOFF and LONG(1985). Western blots were performed as MATERIALS AND METHODS previously described by EGELHOFFand LONG(1985). For SOUTHERNblots, DNA was transferred to Biotrans nylon Strains and plasmids: Strains and plasmids used in this membrane (ICN) by the method of REEDand MANN(1985). study are listed in Table 1. The described insertions in all DNA fragments used as probes were labeled by the hexamer of the listed strains were confirmed by restriction digestion labeling method(FEINBERG and VOGELSTEIN 1983), with and SOUTHERN(1975) analysis. the modification that reactions were carried out in Klenow Plasmid constructions: Plasmid pRmM1 1 1 is a derivative buffer (50 mM Tris(pH 8.0) and50 mM MgC12). High R. meliloti nod Gene Regulation 9
TABLE 1 Strains and plasmids
Strains Relevant characteristics Source characteristics Relevant Strains R. meliloti Rm1021 Str'SU47 MEADEet al. (1982) JM57 nodC-lacZ derivative of RmlO2 1 MULLIGANand LONG(1 985) JM6l nodDl-lacZ derivative of Rm 1021 MULLIGANand LONG(1 985) JM80 nodDl-lacZ, nodD?::Tn5 #303 This study JM8l nodDl::Tn5, nodC-lacZ This study JM83 syrM::Tn5, nodC-lacZ This study JM85 nodD2-Nm', nodC-lacZ This study JM86 nodD1-lacZ, syrM::Tn5 #701 This study JM88 nodD?::Tn5 #303, nodC-lacZ This study JM90 nodD1-lacZ, nodD2:uidA-Sp', nodD?::Tn5, #303 This study JM96 nodD1-lacZ, nodD2:uidA-Sp', syrM::Tn5 #701 This study JM139 nodD2:uidA-Sp' This study JM201 nodD2:uidA-Sp', nodD?::Tn5 #303, nodC-lac2 This study JM204 nodD?::Tn5 #80 1, nodC-lacZ This study JM207 nodD3::Tn5, #709, nodC-lacZ This study JM216 nodD1-lacZ, nodD3::Tn5 #801 This study JT303 nodD?::Tn5 #303 SWANSONet al. (1 987) JT709 nodD?::Tn5 #709 SWANSONet al. (1987) JT80l nodD?::Tn5 #801 SWANSONet al. (1 987) Plasmids pRmJ30 pLAFRl, Tc', nodDl+ JACOBS,EGELHOFF and LONG (1 985) pR75 1 IncP, Tp' F. M. AUSUBEL pRmJT5 pLAFRl , Tc', nodD?+, syrM+, syrA+ SWANSONet al. (1987) pLAFR3 IncP, Tc', lacZ B. STASKAWICZ pRmS303 pRmJT5, Tc', Nm', nodD?:Tn5, yM+,syA+ SWANSONet al. (1 987) pRmS3 1 1 pRmJT5::Tn5, Tc', Nm', nodD?+, syrM::Tn5, syrA' SWANSONet al. (1987) pRmS507 pRmJT5::Tn5, Tc', Nm', nodD?+, syrM+, syA+ SWANSONet al. (1 987) pRmS5 11 pRmJT5::Tn5, Tc', Nm' nodD3+, yM+,syA+ SWANSONet al. (1 987) pRmS7Ol pRmJT5, Tc', Nm', nodD?+, syrMxTn5, syrA+ SWANSONet al. (1 987) pRmS709 pRmJT5, Tc', Nm', nodD3::Tn5, syrM+, syA+ SWANSONet al. (1 987) pRmS7 12 pRmJT5, Tc', Nm', nodD3+, syrM+::Tn5, syA+ SWANSONet al. (1 987) pRmS7 15 pRmJT5, Tc', Nm', nodD3+, syrM+::Tn5, syA+ SWANSONet al. (1987) pRmS8Ol pRmJT5, Tc', Nm', nodD3::Tn5, syrM+, qrA+ SWANSONet al. (1987) pRmM 11 1 pRmJT5, Tc', Nm', nodD?+, syrM+, syrA::Tn5 This study pRmMll3 pLAFR3, Tc', Nm', syrM+, nodH+ This study pRmM 136 pLAFRl, Tc', nodD3+, syrM-, syrA+ This study pRmM 137 pLAFR3, Tc', nodD2+ FISHERet al. (1 988) pRmM141 pLAFRl, Tc', Nm', nodD2- This study pRmM 142 pLAFR1, Tc', nodD?+, syrA+, syrM-, uidA, Sp' This study pRmE43 pLAFRI, Tc', p(trp) nodD1' FISHERet al. (1988) pRmM 139 pLAFR1, nodD2:uidA, SpR FISHERet al. (1988) pRmS9B7 pLAFR1, nodDI::Tn5 JACOBS,ECELHOFF and LONG(1 985)
_____~______~~~ ~~~ ~ stringency washes were 0.1 X SSC at 65O,and low stringency the RNA preparation toa final concentration of 0.1 mg/ml washes were 1X SSC at 65O. DNase. Primer extensions were carried out by the method Primer extensions: To determine the transcription start Of WILLIAMS and MASON(1985), except that the primerwas sites, an oligonucleotide primer complementary to a section hybridized to 10 pg of RNA overnight at 42 O. Sequencing of the coding sequence of each gene was synthesized. The ladders were obtained by using the same primers in dideoxy sequence of theprimer for nodA was 5'-TAG sequencing reactions on single stranded clones of the up CTTCCACTGCAC-3',and for nodH, 5"GCAGCG- stream region for each gene. The sequencing ladder and TGGAATGGG-3'. End-labeling was carried out in a vol- RNA-complementary primer extension products were run ume of 10 pl with 2 picomoles ofDNA primerand 8 in parallel on sequencing gels to establish the position of the picomoles (405 Ci/mmol) of [y-'*P]ATP for 4 hr at 37" transcription initiation site as previously described for nodF using T4 polynucleotide kinase in the buffer suggested by and nodH (FISHERet al. 1987b). the supplier. RNA was prepared as described by FISHERet al. (1987a). RNase-free DNase was prepared as described RESULTS by TULLISand RUBIN(1 980). DNase (Sigma) and proteinase K (Boehringer Mannheim),each at a concentration of 1 Identification and isolationof nodD2 and nodD3: mg/ml, were incubated together in 20 mM Tris (pH 7.5) In R. meliloti 102 1, nodD1 is adjacent to nodABC on and 10 mM CaC12 for 2 hr at 37 O , and then added fresh to an 8.7-kb EcoRI fragment (EGELHOFFet al. 1985). 10 J. T. Mulligan and S. R. Long
GOTTFERTet al. (1986)and HONMA and AUSUBEL were tested fortheir ability to induceconstitutive (1987) demonstrated thattwo homologous genes, des- expression of the chromosomal nodC-lac2 fusion and ignated nodD2 and nodD3, are present, respectively, to cause a mucoid colony morphology. Insertions in ona 6.8-kb and a 15.5-kb EcoRI fragment in R. three regions of the plasmid affected one or both of meliloti. these phenotypes (Figure 2) and defined three loci; SWANSONet al. (1987) describe the isolation of a syrM, syrA and nodD3 (syr for symbiotic regulation). cosmid clone, pRmJT5,which carries nodH, nodFand Eachlocus is delimited by flanking Tn? insertions nodE on a 15.5-kb EcoRI fragment. This same frag- with no regulatory phenotype.Plasmids bearing inser- ment hybridizes to an internal nodDl probe (data not tions in syrM (pRmS701, pRmS712, pRmS715 and shown), indicating that nodD3 resides on this same pRmS311) showed reduced nodC expression and 15.5-kb fragment. The remaining 6.8-kb nodDl ho- failed to cause the mucoid colony morphology (Table mologous fragment was not present on any of our 2, line 5, shows results representative of the three previously identified clones (data not shown). As de- insertions in syrM). The syrA insertion(pRmM111) scribed by FISHERet al. (1988), we identified a clone had no effect on the expression of nodC, but failed to containing this 6.8-kb fragment from a cosmid clone cause the mucoid colonial morphology (Table 2, line bank of R. meliloti 1021. The 6.8-kb fragment was 6). Three Tn5insertions in the 2.2-kb nodD3-contain- subcloned into pLAFR3 to createpRmM 137. ing BgllI fragment of pRmJT5 (pRmS303, pRmS709 Plasmid-borne copies of nodD2 and nodD3 stim- and pRmS801) reduced expression of nodC but still ulate expression of nodC Our previous work (MUL- rendered the strain mucoid (Table 2, lines 7 and 8, LIGAN and LONG 1985)showed that a plasmid-borne show results of two of the three insertions in nodD3). copy of nodDl enhanced the induction of the nodC- None of the remaining 83 Tn5 insertions in pRmJT5 lacZ fusion by plant exudate (see also Table 2, lines 1 affected nodC expression or colony morphology (data and 9). PETERS,FROST and LONG (1986)showed that not shown). Since several Tn5 insertions lying be- the flavone luteolin is the major inducing molecule in tween syrM and nodD3 display no regulatory pheno- alfalfa plant exudate when nodDl is plasmid-borne. type, the two genes are probably in separate transcrip- We therefore tested the effect of plasmid-borne copies tional units. Genomic Tn5 insertions at each of these of the other two nodD genes on the induction of the genes individually did not block or substantially delay nodC-lacZ fusion by luteolin and by alfalfa seed wash nodulation (SWANSONet al. 1987; HONMA and AUSU- (exudate). Seed wash is a complex mixture including BEL 1987). The lack of requirement for syrA and syrM luteolin and a wide variety of related compounds (S. may be due to redundant function, as in the case of LONGand N. K. PETERS,unpublished observations), the nodD group of genes; alternatively, syrM and syrA some of which may potentially interact with the three may affect regulation of functions in a way which is nodD gene products. not critical for nodulation, at least in standard labo- The plasmid carrying nodD2, pRmM 137, enhanced ratory assays. the expression of the nodC-lacZ fusion in the absence The same exo genes, such as exoA,exoB and exoE, of any inducer(Table 2, lines 1 and 2). Luteolin which are required for bacterial invasion of the plant stimulated expression of the nodC-lacZ fusion when roots (FINANet al. 1985; LEIGH, SIGNERand WALKER no plasmid was present (therefore reflecting theaction 1985) are also required for the mucoid colonial phe- of the normal genomic copies of nodD) but not when notype of strains containing pRmJT5. We found that the nodD2 plasmid was present. In contrast, seed wash Tn5 insertion mutations in these genes suppress the stimulated nodC expression four fold (line 2). The mucoid colonial morphology caused by pRmJT5 (data increase in nodC expression in the presence of the not shown), suggesting that pRmJT5 causes the mu- plasmid was dependent on nodD2: an insertion in the coid phenotype by affecting the activity or expression plasmid borne copy of the nodD2 coding region, re- of these exo genes. This suggests that syrM can affect duced the expression of nodC (Table 2, line 3). This the regulation or activity of two classes of symbiotic mutated plasmid, pRmM 14 1, had a consistent small genes, the nod genes and the exo genes. (2-fold) effect on the uninduced level of nodC expres- Transcription initiates at the same site whether sion, but caused no enhancement of seed wash induc- induced by nodDl plus luteolin or by syrM plus tion of nodC (Table 2, lines 1 and 3). nodD3: We examined the transcription start sites for The cosmid clone pRmJT5, which carries nodD3, nodABC and fornodH under two regulated conditions. in addition to nodE, nodF and nodH, caused very high Figure 3 shows that for both nodA and nodH, tran- expression of the nodC-lacZ fusion with or without scription initiates at the same nucleotide whether in- inducer (Table 2, line 4). In addition, strains which duction is mediated by nodD1 in the presence of contain pRmJT5 displayed a mucoid colonial mor- luteolin, or by plasmid-borne copies of nodD3 and phology (Figure1). Ninety derivatives of pRmJT5 syrM. The reactions in panel A of Figure 3 use an with different Tn5 insertions (SWANSONet al. 1987) oligonucleotide primer that is complementary to the R. melilotiRegulation nod Gene 11
TABLE 2
Effect of cloned genes on inducible nod gene expression and colony morphology
@-(hlactosidase activityb
Line Plasmid Mininnl + Minimal + Colonial No. Srrain Genotype"luteolin Mininul seed wash morphology
1 JM.57 None 3 15 1.5 Normal 2 JM57 pRmM 137 D2+ 16 9 67 Normal 3 JM.57 pKmM141 D2- 6 10 IS Normal 4 JM.57 pRnlJT.5 DS'. M+.A+ 260 265 240 Mucoid 5 JM57 pKmS701 8 DS+,M-, A+ 8 14 Normal 6 JM57 pRmM1 11 IN+, M+,A- 220 NT 240 Normal 7 JM57 pRmS303 DS-. M+,A+ 45 NT 75 Mucoid 8 JV.57 pRnlS80I D3-, M+, A+ 70 NT 75 Mucoid 9 JM.57 pRmJS0 Dl+ 3 87 75 Normal ,, Gene abbrevi;ltions: Dl = nodDl; D2 = nodD2; DS = nodD3; M = s~nbf;A = 5yrA. ' In this and the following tables the activities listed are theaverages of the results from four ;ISS;I~S. In each case the individual assays give V:IIIIP< \vitIlin 1.5% nf tlw :IVC~:ICTC.The .-ndnm=nnm R-v:llarroidase :Irrivitv in the ahsmre of introduced fusions is 2 units. NT = not tested.
I~XRF2.-l'lasmitls and l'n5 insertions. Mapped positions of 1'115 insertions are shown ;IS determined bv SWANSONel al. (1987). EroRl restriction sites are shown forthe entire DNA segment. M'ithin the 15.5-kh BroRl fi.agIncnt, ClaI (C) Xbal (X) ancl RglIl (81) sitesarc idso shown. Map ofnodD2 is based on results of klONMA and ALWREI.(1987) and our results (J. MUI.LIGAN, unpublished ohser\.;ltions): in rhis fragnlent, RamHl (B) and Xho (X) sites are shown. The gap shown in the EroRI map is approrin1;ltely 47 kb. Positions of gene fusions to 1arZ ant1 uidA are shown. pKmM I4 1 corwtins an insertion which inactiv;rres nodD2, but which is not a I~IGL~RFI .-Aluroid colonial nlorphology of an R. me1i1ofi strain fusion. carrying phsmidpKmJT.5. I?. mdilofi 102 I carrying pRmS1.26 (left) or pRnl.lT5 (right) was strcaketl for single colonies on an IBplatc. sequence. Again, the primer extension products syn- thesized using the two induced RNA samples are initial portion of the nodA codingsequence. This similar (lanes 11 and 12), and indicate that transcrip- primer was usedin dideoxy sequencingreactions tion initiates 28 bp from the nodHnod box. Thus (SANCER,NICKLEN and COULSON1977) on a clone of inducible nod promoters areaffected indistinguishably the putative promoter regionto provide the size stand- bv the two regulatory systems: nodDl in the presence ards in lares 1-4. The same primer was end-labeled and used in primer extension reactions on RNA from of luteolin, and nodD3 with syrM. two strains. In lane .5 the RNA was purified from a Effect of genomic mutations onnodC expression: straincarrving a nodD1 plasmid and induced with The experiments described above used the overex- luteolin. In lane 6 the RNA was isolated from a strain pression of plasmid-borne copies of regulatory genes with plasmid copies of nodD3 and syrM. The primer to demonstrate gene location and function. We also extension products from the two reactions are similar studiedregulatory consequences of genomic inser- and indicate that transcription initiates 26 bp from tional mutations in these genes. A chromosomal copy the nodli nod box. Panel B shows analogous reactions of the nodC-lac2 fusion in an otherwise wild-type carriedout with an oligonucleotideprimer that is background was induced 5-fold by purified luteolin compIementar~ to thebeginning of the nodH coding and 5- 12-fold by alfalfa seed wash (Table 3, line 1). 12 J. T. Mulligan and S. R. Long 123456 TABLE 3 Regulatory behavior of mutations in nodD and syt" genes
Cllrotllosotlle @-Gal;lcaosir1;lsc;miviry gmolypc
Line Minin1;tI + Minim;d + No. Strain Dl D2 Df M Minimal Iuteolino seed u.;&' 1 JM57 ++++ 3 15 1 2 JM81 -+++ 3 5 4 3 JM83 +++- R 22 18 4 JM85 +"+ 3 75 15 5 JM88' ++-+ 3 30 18 6 JM207' ++-+ 2 33 NT 7 JM204' ++-+ 5 84 17 8 JM201 +--+ 4 I10 20" 9 JM57 ++ + + - 3 130 145 pRlnJ30
a In a11 the experinlents reported herethe bacterial cultures were exposed to inducer for 4 hr. When JM201 (line 8) w'as induced for 8 hours with seed wash the final activity was 80 units. Batches of seed wash are variable. Repetition of inductions with a single batch gave consistenr results, but occasional batches gave enhanced induction. The remaining inductions were done with FIGURE3."Primer extension to deternlinc transcription start seed wash which induced JM57 to 15 units. sites for nodA (lanes 1-6) and nodH (lanes 7-12) when induced by ' These strains each carry a different nodD3::Tn5 insertion. nodDl plus luteolin or nodD3 plus 5yrM. Oligonucleotide primers JM88 carries Tn5 insertion #30S. JM207 carries Tn5 insertion were used to direct DNA synthesis complementary to transcripts #709. and.lM204 carries Tn5 insertion #801 (SWANSONet al. 1987). isolated fronl luteolill-inducecl cellscarrying pRnlSL26, which con- tains nodDl (lanes 5 and 1 I), and luteolin uninduced cells carrying pRnlJT5, which contains nodD3 and syrM (lanes 6 and 12). The nodA complementary oligomer was used in lanes 1-6 and the nodH 200 - complementary oligomer in lanes 7- 12. For both nodA and nodH, transcription is initiated at the same site in cells induced by nodDI plus luteolin, and in cells induced by nodD3 plus 5yrM. Correspond- 92- ing sequencing ladders (lanes 1-4 ;~nd7-10) permit identification of the transcription start sites. 69-
A strain containing a Tn5 insertion in nodDI showed 46- almost no induction ofnodC by luteolin or seed wash (Table 3, line 2). Tn5 insertions in syrM have little effect on the regulation ofnodC (line 3). An insertion in nodD2 enhances the induction of nodC by luteolin 30- (line 4). Three different Tn5 insertionsin the nodD3 region enhanced the induction ofnodC by luteolin (Table 3, 21 - lines 5, 6, 7 compared with line 1). The enhancement of nodC induction by nodD3 insertions is unlikely to be due to polar effects on another gene because our FIGURE4.-Use of a Western blot to estimate nodDl protein mapping indicates that insertion #lo05 (SWANSONet levels. Protein extracts from R. meliloti strains were subjected to al. 1987), which has no regulatory phenotype (data SDS-polyacrylamidegel electrophoresis, transferred to nitrocellu- not shown), is fewer than 500 bp downstream of the lose and probed with a NodD-specific polyclonal antibody. Strains nodD3 coding region. from which extracts were prepared are as follows: Lane I, MJ57; lane 2, JM57 (pRrnJR0);lane 3, JM201 (nodD2- nodD3-). The large Luteolin-mediated induction of the nodC-lac2 fu- arrow denotes the nodDl protein: the small arrows indicate the sion by nodDl was dramatically enhanced in a back- expected positions of the nodD2 and nodD3 proteins. ground in which nodD2 and nodD3 both were absent. nodC induction in JM201, a derivative of the nodC nodDs, could be due either to enhancedexpression of fusion strain with only nodDI intact (nodD2:uidA in- nodDl or to an increase in the activity of a constant sertion, nodD3:Tn5 #303), was similar to thatallowed amountof NodD1. To testthese possibilities, we by a plasmid copy of nodDl (Table 3, lines 8 and 9). analyzed the level of NodDlin the two strains. Figure The increased induction in JM201 (nodD2- nodD3-) 4 shows a Western blot loaded with an equal amount relative to JM57, which has intact copies of all three of protein from thenodC-lac2 fusion strain JM57 (lane R. meliloti nod Gene Regulation 13 TABLE 4 BglII fragments were subcloned into pUCll9 in Esch- nodDl expression in mutant backgrounds erichia coli (Figure 5) and protein extracts were pre- pared from thesestrains. The extracts were separated Chromosome by SDS polyacrylamide gel electrophoresis, electro- genotype blottedonto nitrocellulose, and hybridized with Line Plasmid @-Galactosidase No. Strain Dl D2 D3 M genotypeactivity polyclonal anti-NodD antiserum (FISHERet al. 1988). The Western blot shown in Figure 5A indicates that + + + None 54 1 JM61 lacZ the nodD3 codingsequence is located between the 2 JM80 lacZ + -a + None 52 - -a 3 JMSO lacZ + None 58 leftmost ClaI and rightmost BglII sites (Figure 5B). A 4 JM96 lacZ - + - None 50 product the size of authentic NodD3 (FISHERet al. 5 JM216 lacZ + -b + None 61 1988) was seen in extracts from strains in which the + + + nodDl 48 6 JM61 pRmE43 lacZ vector promoter directs transcription from the left of 7 JM61 pRmM137 lacZ + + + nodD2 45 8 JM6l pRmJT5 lacZ + + + nodD? 105 the BglII and ClaI fragments (Figure 5A, lanes 1 and 3). No such product was detected in the extract from a nodD?::Tn5 insertion #303. ’ nodD3::Tn5 insertion#80 1. the strain with a non-coding insert vector alone (lane 4) or in the extract from the strain with the vector l),the same strainwith a plasmid-bornecopy of nodDl promoter reading from right to left across the BglII (pRmJ3O) (lane 2), and JM201 (lane 3). The blot was fragment (lane2). This indicates that the open reading probed with a polyclonal antibody preparation which frame for nodD3 reads from left to right. binds all three NodDs (FISHERet al. 1988) and the The nodD3 coding region is bounded by the ClaI arrows indicate the expected positions of the three site on one end and Tn5 insertion #1005, 1.6 kb to proteins. The amount of NodDl present in JM201 the right. Based on the apparent molecularweight of (nodD2- nodD3-, lane 3) was similar to that in JM57 the nodD3 geneproduct (FISHERet al. 1988) the (nodD2+ nodD3+,lane 1) and is substantially less than coding sequence is probably slightly longer than the that in the strain carrying nodDl on a plasmid (lane l-kb nodDl coding sequence (EGELHOFFand LONG 2). The nodD3 protein is faintly detectable in strains 1985). Insertion #80 1, which is internal to this 1.6 kb where a normal nodD3 gene is present (lanes 1 and region, is thus likely to interrupt the gene. The two 2), but nodD2 protein is not detectable, as previously insertions with the less pronounced regulatory effects, observed (FISHERet al. 1988). In addition, expression #303 and#709, are probably not in the nodD3 coding of a nodD1-lac2 fusion was unaffected by mutations sequence, although as they are upstream of nodD3) in either or both of the other two nodD genes (Table 4, lines 2-5), and relatively little by elevated expres- they may interrupt or interfere with its transcription. sion of any of the nodD genes (lines 6-8). Taken This is consistent with nodulation assays whichsuggest together,the Western blot analysis, nodD1-lac2 that insertion #303 allows residual expression of expression values, and nodC induction data, indicate nodD3: JM90, a strain with insertion #303 and also that the products of nodD2 and nodD3 interfere with insertions in nodDl and nodD2, nodulates R. meliloti the activity of NodD 1 rather than reduceits synthesis. at low frequency(Figure 6). HONMA andAUSUBEL In strains carrying any one of the nodD genes on a (1987) have shown that strains with insertions in all plasmid, the expression of the nodC fusion was rela- three of the nodD genes are Nod-. tively unaffected by mutations in the chromosomal The interactionof syrM and nodD3: Two plasmids copies of the other nodD genes (Table 5; compare were constructed to test theinterdependence of lines 1-4 within panels 2, 3 or 4). However, plasmid- nodD3 and syrM: one carries only syrM (pRmM 1 13), borne copies of the nodD2 and nodD3 genes showed and one carries only nodD3 (pRmM 136). In a strain some unexpected behavior. While the genomic copies with an intact genomic copy of nodD3, the syrM plas- of nodD2 and nodD3 appeared to decrease nodD1- mid causes high uninduced expression (50 or more mediated induction by luteolin (Table 3, lines 4-8), units); the luteolin or seed wash induced expression is plasmid copies of nodD2 or nodD3 enhanced thebasal twofold higher, about 100-200 units (Table 5, panel level of nodC-lac2 expression (Table 5, panels3 5, lines 1-3 and line 6). However, in the strains lacking and 4). an intact nodD3 gene (Table 5, lines 4 and 5), the Definition of nodD3 gene: The phenotypes and syrM plasmid had almost no effect on nodC expression regulatory effects of the different nodD3 insertions (compare panel 5 with panel 1). Similarly, the nodD3 were correlated with their likely position within nodD3 plasmid caused high, constitutive expression of nodC by a variety of criteria. The insertions in JM88 and when syrM was intact but had almost no effect in a JM207, #303 and #709 respectively, are near the left syrM- background (Table 5, panel 4, compare line 1 border of a 1.6-kb BglII fragment (Figure 2). The and line 6). The behavior of plasmid-borne syrM or insertion in JM204, #801, is near the middle of the nodD3 was not affected by the presence or absence of same fragment. The nodDl homologous ClaI and other nodD genes (Table 5, panels 4 and 5). 14 J. T. Mulligan and S. R. Long
TABLE 5
nodC-lacZ" activities resulting from interaction of plasmid-borne and genomic nodD and syrM genes
Chromosome Plasmid, genotype, andnodC-lacZ fusion activity" genotypeb None pRmJ3O pRmM 137 pRmM136 pRmMll3 1 2 3 4 5 nodD2 nodD3nodDl nodD2 syrM LineStrain No. D2 Dl D3 M - Lut' SWd - Lut sw - I.ut sw - Lut sw - Lut sw
1 JM57 + + + 15+ 3 8715 3 75 16 9 67 107 113 105 51 150 120 2 JM8l - + +5 + 3 654 3 45 17 9 23 98 105 87 64 215 94 3 JM85 3 + - + 75 + 3 15 3 94 81 14 9 58 86 97 83 65 110 NT 4 JM88 + + - + 3 30 18 3 97 89 17 10 60 125 137 NT 7 14 11 5 JM204 + + - +84 5 17NT NT NT NT NT NT 155135 NT 8 57 16 6 JM83 + + + - 16 3 45 6222 3 18 1810 9 61 11140 79 121 7 JM201 + - - +110 4 20 5 111 17138 12 25 NT NT NT NT NT NT
a Numbers represent &galactosidase assay repeated at least twice. NT = not tested. ,L Dl, D2, D3,'M as in. ?able 2. ' Lut = luteolin. sw = alfalfa seed wash.
Expression of syrM was also required for nodD3- nodD3. Strains which lack all three nodD homologs dependent nodulation. R. meliloti strains which lack are Nod- (HONMA andAUSUBEL 1987). nodDI,nodD2 and nodD3 are Nod- (HONMAand We found that regulation of the nod genes may be AUSUBEL1987); we observed low nodulation for strain coordinated with the regulation of another set of JM90, which we ascribe to the leakiness of symbiotic genes, those required forexopolysaccharide nodD3::Tn5 insertion #303 (see above). However, synthesis. One gene,syrM, can increase the expression JM96, a strain which lacks nodD1, nodD2 and syrM of the nod genes in conjunction with nodD3, and also was completely Nod- (Figure 6). This indicates that can increase the expression or activity of the exo genes syrM is required for nodD3 to function as an activator in conjunction with another locus, syrA. Insertions in of the nod genes. syrM or nodD3 have little effect on nodulation in a wild-type strain (SWANSONet al. 1987) butcompletely block nodulation in an otherwise wild-type strain DISCUSSION which lacks nodDl and nodD2 (Figure 6). In the pres- ence of syrM, nodD3 induces sufficient nod gene Ofthe Rhizobium nodulation genes identified to expression to allow normalnodulation. Together, date, only the nodD genes have a well characterized syrM and syrA may induce the expression of exo genes. function: regulationof the expression of the inducible Plasmid pRmJT5, carrying syrM and syrA, can sup- nod genes. Most of the species of Rhizobium tested press the Exo- phenotype of Tn5 insertions in two have two or threecopies of sequences which are highly different exo genes which are probably regulatory homologous to nodD (RODRIGUEZ-QUINONESet al. (J. MULLIGANand S. LONG, unpublishedobservations; J. 1987). Mutations in nodD are Nod- in R. leguminosa- LEIGH and G. WALKER,personal communications). rum bv. trqolii and R. leguminosarum bv. viciae but are This suggests that, like the nod genes, expression of Nod+ in R. meliloti (DOWNIEet al. 1984; SCHOFIELD the exo gene functions can be controlled by more than and WATSON1985; JACOBS, EGELHOFFand LONG 1985; EGELHOFFet al. 1985). This difference is pos- one locus. tulated to occur because the two R. leguminosarum The datapresented here indicate that all three strains each have a single copy of nodD, but R. meliloti copies of nodD in R. meliloti are functional as regula- harbors three nodD homologs (PUTNOKY andKON- tors, and at least nodDl and nodD3 affect the same DOROSI 1986; HONMA and AUSUBEL1987). Eachof promoters (Figure3). Each copy is capable of affecting the threehomologous R. meliloti nodD genes will allow the expression of the nod genes when it is overpro- some nodulation in the absence of theother two duced (Table 2), and insertions in any copy will alter (HONMA andAUSUBEL 1987; GYORGYPAL, IYER and the expression of a nodC-lac2 fusion (Table 3). KONDOROSI1988), although strains which have only The regulatory activities of the three nodD homo- nodD2 cause delayed and poor nodulation on alfalfa logs are consistent with the nodulation phenotypes of and areNod- on Melilotus alba (GOTTFERTet al. 1986; theirmutants. Enhanced expression of nodDl or HONMA andAUSUBEL 1987). The copy of nodD adja- nodD3 dramatically affects nodC expression, and cent to nodABC in R. meliloti is referred to as nodDl strains which lack both of these genes nodulate very and the other two loci are referred to as nodD2 and poorly. Strains with insertions in nodD2 display a R. meliloti nod Gene Regulation 16
FIGURE6.-Plot of nodulation versus time for wild type R. mrliloti and nod11 mutants. The graph shows the number of plants nodul;ltetl from a group of 10 plants that were inoculated at day 0 for each of the six strains. w)I02 1 ; (6-6) JMfil (nodDI-); w JM139 (nodD2-); (0-0) JT303 (nodnj-): w JM90(nodnl- nodD2- nodD3#303-); and (W)JM96(nodDl- nodD2- syW).
gene products may induce the nod genes insitu in #I09 @303 #8OI #1005 response to conditions which are not imitated by our Btl11 0.1 I I I induction assays. I I 1 I The physiological significance of this pattern of two LANE 1 + 1.C promot*. or three regulatory genes with similar functions in R. LANE 2 + meliloti is not yet clear, particularly in light of the fact lac promoter that many R. leguminosarum biovars have only one LANE 3 + 1.e promoter copy. GOTTFERT et al. (1986)proposed that nodD LANE 4 + l"e
LITERATURE CITED FINAN,T. M.,A. M. HIRSCH,J. A. LEIGH,E. JOHANSEN,G. A. KULDAU,S. DEEGAN,G. C. WALKERand E.R. SIGNER, 1985 APPELBAUM,E. R., D. V. THOMPSON,K. IDLERand N. CHARTRAIN, Symbiotic mutants of Rhizobium meliloti that uncouple plant 1988 Rhizobiumjaponicum USDS 191 hastwo nodD genes from bacterial differentiation. Cell 40 869-877. that differ in primary structure andfunction. FIRMIN,J. L., K. E. WILSON, L. ROSSEN and A. W. B. JOHNSTON, BAUER,W. D., 1981 Infection of legumes by Rhizobia. Ann. Rev. 1986 Flavonoid activation of nodulation genes in Rhizobium Plant Physiol. 32: 407-449. reversed by other compounds present in plants. Nature 324: BORTHAKUR,D., and A. W. B. JOHNSTON, 1987 Sequence ofpsi a 90-92. gene on the symbiotic plasmid of Rhizobium-phaseoli which FISHER,R. F., H. L. BRIERLEY,T. J. MULLIGANand S. R. LONG, inhibits exopolysaccharide synthesis and nodulation and dem- 1987a Transcription of Rhizobiummeliloti nodulation genes: onstration that its transcription is inhibited by psr another gene identification of a nodD transcription initiation site in vitro and on the symbiotic plasmid. Mol. Gen. Genet. 207: 149-154. in vivo.J. Biol. Chem. 262: 6849-6855. BORTHAKUR,D., C. E. BARBER,J. W. LAMB,M. J. DANIELS,J. A. FISHER,R. F., J. SWANSON,J. T. MULLIGANand S. R. LONG, DOWNIEand A. W. B. JOHNSTON. 1986A mutation that 1987b Extended nodulation gene region in Rhizobium meliloti blocks exopolysaccharide synthesis prevents nodulation of peas 1021. 11. Sequence, protein products, and transcription start by Rhizobium-leguminosarum but not of beans by Rhizobium- sites for nodFEGH. Genetics 117: 191-201. phaseoli and is corrected by cloned DNA from the phytopath- FISHER,R. F., T. T. EGELHOFF,J. T. MULLIGAN and S. R. LONG, ogen Xanthomonas. Mol. Gen. Genet. 203: 320-323. 1988 Specific binding of Rhizobium meliloti extracts contain- BURN,J., L. ROSSENand A.B. W. JOHNSTON,1987 Four classes ing NodD to DNA sequences upstream of inducible nodulation of mutations in the nodD gene of Rhizobiumleguminosarum genes. Genes Dev. 2: 282-293. biovar. viciae that affect its ability to autoregulate and/or acti- GOTTFERT, M.,B. HORVATH,E. KONWROSI, P. PUTNOKY,F. ROD- vate other nod genes in the presence of flavonoid inducers. RIGUEZ-QUINONESand A. KONWROSI, 1986 Atleast two Genes Dev. 1: 456-464. nodD genes are necessary for efficient nodulation of alfalfa by CHAKRAVORTY,A. K., W. ZURKOWSKI,J. SHINEand B. G. ROLFE, Rhizobium meliloti. J. Mol. Biol. 191: 41 1-420. 1982 Symbiotic nitrogen fixation molecular cloning of Rhi- GYORGYPAL,Z., N. IYER and A. KONDOROSI, 1988 Three regu- zobium-triyolii genes involved in exo polysaccharide synthesis latory nodD alleles of diverged flavonoid-specificity are in- and effective nodulation. J. Mol. Appl. Genet. 1: 585-596. volved inhost-dependent nodulation by Rhizobium meliloti. Mol. CHEN, H., M. BATLEY,J. REDMONDand B. G. ROLFE, 1985 Gen. Genet. 212: 85-92. Alteration of the effective nodulation properties of a fast- HONG,G.-F., J. E. BURNand A. W. B. JOHNSTON, 1987 Evidence growing broad host range Rhizobium due to changes in exo- that DNA involved in the expression of nodulation (nod)genes polysaccharide synthesis.J. Plant Physiol. 120: 331-350. in Rhizobium binds to the product of the regulatory gene nodD. DEBELLE,F., C. ROSENBERG,J. VASSE, F.MAILLET, E. MARTINEZ, Nucleic Acids Res. 159677-9690. J. DENARIEand G. TRUCHET,1986 Assignment of symbiotic HONMA,M. A., and F. M. AUSUBEL,1987 Rhizobiummeliloti has developmental phenotypes to common and specific nodulation three functional copies of the nodD symbiotic regulatory gene. (nod)genetic loci of Rhizobium meliloti. J. Bacteriol. 168: 1075- Proc. Natl. Acad. Sci. USA 84 8558-8562. 1086. HORVATH,B., E. KONWROSI,M. JOHN, J. SCHMIDT,I. TOROK,Z. DJORDJEVIC,M. A., P. R. SCHOFIELDand B. G. ROLFE,1985 Tn5 GYORGYPAL,I. BARABAS, U. WIENEKE,J. SCHELLand A. KON- mutagenesis of Rhizobium trifolii host-specific nodulation genes WROSI, 1986 Organization, structure and symbiotic function result in mutants with altered host-range ability. Mol. Gen. of Rhizobium meliloti nodulation genes determining host speci- Genet. 200: 463-471. ficity for alfalfa. Cell 46: 335-343. DJORDJEVIC,M. A,, J. W. REDMOND,M. BATLEYand B. G. ROLFE, HORVATH,B., C. W.B. BACHEM,J. SCHELLand A. KONWROSI, 1987 Clovers secrete specific phenolic compounds which 1987 Host-specific regulation of nodulation genes in Rhizo- either stimulate or repress nod gene expression in Rhizobium bium is mediated by a plant signal, interacting with the nodD trifolii. EMBO J. 6 1173-1 179. gene product. EMBO J. 6 841-848. DOWNIE,J. A., Q.4. MA, B. WELLS, C. D.KNIGHT, G. HOMBRECHER JACOBS,T. W., T. T. EGELHOFFand S. R. LONG,1985 Physical and A.W. B. JOHNSTON, 1984 Cloning of the symbiotic and genetic map of a Rhizobium meliloti nodulation gene region region of Rhizobium leguminosarum: the nodulation genes are and nucleotide sequence of nodC. J. Bacteriol. 162: 469-476. between the nitrogenase genes and a nfA-like gene. EMBO J. LEIGH, J. A., R. E. SIGNER and G. C. WALKER, 2: 947-952. 1985 Exopolysaccharide-deficient mutants of Rhizobiummel- DUDLEY,M. E., T. W. JACOBS and S. R. LONG,1987 Microscopic iloti that form ineffective nodules. Proc. Natl. Acad. Sci. USA studies of cell divisions induced in alfalfa roots by Rhizobium 82: 6231-6235. meliloti. Planta 171: 289-301. MARTIN, K., L. Huo and R. F. SCHLEIF,1986 The DNA loop DYLAN, T.,L. IELPI,S. STANFIELD,L. KASHYAP,C. DOUGLAS,M. model for ara repression: AraC protein occupies the proposed YANOFSKY, E. NESTER,D. R. HELINSKIand G. DITTA, 1986 loop sites in vivo and repression-negative mutations lie in these Rhizobium-meliloti genes required for nodule development are same sites. Proc. Natl. Acad. Sci. USA 83: 3654-3658. related to chromosomal virulence genes in Agrobacterium-tu- MARTIN,M. O., and S. R. LONG,1984 Generalized transduction mefaciens. Proc. Natl. Acad. Sci. USA 83: 4403-4407. in Rhizobium meliloti. J. Bacteriol. 159 125-129. EGELHOFF,T. T., and S. R. LONG, 1985 Rhizobium meliloti MEADE,H. M., S. R. LONG,G. B. RUVKUN,S. E. BROWNand F. M. nodulation genes: Identification of nodDABC gene products, AUSUBEL,1982 Physical and genetic characterization of sym- purification of nodA protein, and expression of nodA in Rhizo- biotic and auxotrophic mutants of Rhizobium meliloti induced bium meliloti. J. Bacteriol. 164: 591-599. by transposon Tn5 mutagenesis. J. Bacteriol. 149: 114-122. EGELHOFF,T. T., R. F. FISHER,T. W. JACOBS and S. R. LONG, MILLER,J. H., 1972 Experimentsin Molecular Genetics, pp. 466. 1985 Nucleotide sequence of Rhizobium meliloti 1021 nodu- Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. lation genes: nodD is read divergently from nodABC. DNA 4: MULLIGAN,J. T., and S. R. LONG, 1985 Induction of Rhizobium 241-248. meliloti nodC expression by plant exudate requires nodD. Proc. FEINBERG,A. P., and B. VOGELSTEIN,1983 A technique for Natl. Acad. Sci. USA 82: 6609-6613. radiolabelling DNA restriction endonuclease fragments to high NEWCOMB,W., 1981 Nodule morphogenesis and differentiation, specific activity. Anal. Biochem. 132: 6-13. pp. 247-298 in Biology of the Rhizobiaceae, edited by K. L. GILFS 18 J. T. Mulligan and S. R. Long
and A. G. ATHERLY.Academic Press, New York. melilotia,nodP and nodQ. Mol. PlantMicrobe Interact. (in PETERS,N. K., and S. R. LONG,1988 Alfalfa root exudates and press). compounds which promote or inhibit induction of Rhizobium SHEARMAN,C. A., L. ROSSEN,A. W. B.JOHNSTONand J. A. DOWNIE, meliloti nodulation genes. Plant Physiol. 88: 396-400. 1986 The Rhizobiumleguminosarum nodulationgene nodF PErERS, N. K.,J. W. FRosTand S. R. LONG,1986 A plant flavone, encodes a polypeptide similar to acyl-carrier proteinand is luteolin, inducesexpression of Rhizobium meliloti nodulation regulated by nodD plus a factor in pea root exudate. EMBO J. genes. Science 233: 917-1008, 5: 647-652. PUTNOKY,P., and A. KONDOROSI,1986 Twogene clustersof SOUTHERN,E. M. 1975 Detectionof specific sequences among Rhizobium meliloti code for earlyessential nodulation functions DNA fragments separated by gel electrophoresis. J. Mol. Biol. and a third influences nodulation efficiency. J. Bacteriol. 167: 98: 503-5 17. 881-887. SPAINK,H. P., R. J. H. OKKER,C. A. WIJFFELMAN,E. PEESand B. REDMOND, J. W.,M. BATLEY,M. A. DJORDJEVIC,R. W. INNES, P. J. J. LUGTENBERG,1987a Promoters in the nodulation region L. KUEMPELand B. G. ROLFE,1986 Flavones induce expres- of the Rhizobiumleguminosarum Sym plasmid pRLlJI. Plant sion of nodulation genes in Rhizobium. Nature 323: 632-634. Mol. Biol. 9 29-37. REED,K. C., and D. A. MANN,1985 Rapid transfer of DNA from SPAINK, H.P., C. A. WIJFFELMAN,E. PEES, R. J. H. OKKER andB. agarose gels to nylon membranes. Nucleic Acids Res. 13: 7207- J. J. LUGTENBERG,1987b Rhizobium nodulation gene nodD as 722 1. a determinant of host specificity. Nature 328: 337-340. RODRIGUEZ-QUINONES,F., Z. BANFALVI, P.MURPHY and A. KON- SWANSON, J.,K. J. TU, J.M. OGAWA,R. SANGA,R. FISHERand S. DOROSI, 1987 Interspecieshomology of nodulation genes in R. LONG, 1987 Extendednodulation gene region in Rhizo- Rhizobium. Plant Mol. Biol. 8: 61-76. bium meliloti 1021. I. Physical and genetic map of loci in clone ROSSEN,L., C.A. SHEARMAN,A. W. B.JOHNSTONandJ. A. DOWNIE, pRmJT5. Genetics 117: 181-189. 1985 The nodD gene of Rhizobium leguminosarum is autoreg- TULLIS,R.H., and H. RUBIN, 1980 Calcium protects DNase I ulatory and in the presence of plant exudate induces the nod from proteinase K: a new method for the removal of contami- A, E, C genes. EMBO J. 4: 3369-3373. nating RNase from DNase I. Anal. Biochem. 107: 260-264. ROSTAS,K., E. KONDOROSI,B. HORVATH,A. SIMONCSITSand A. WILLIAMS,J. C., and P. J. MASON,1985 Hybridisation in the KONDOROSI,1986 Conservationof extendedpromoter re- analysis of RNA, pp. 139-160 in Nucleic Acid Hybridization: A gions of nodulation genes in Rhizobium. Proc. Natl. Acad. Sci. Practical Approach, edited by B. D. HAMS and S. J. HIGGINS. USA 83: 1757-1761. IRL Press, Washington, D.C. SANGER, F.,S. NICKLEN andA. R. COULSON,1977 DNA sequenc- ZAAT, S. A. J., C. A. WIJFFELMAN,H. P. SPAINK,A. A.N. VAN ing with chain-terminating inhibitors. Proc. Natl.Acad. Sci. BRUSSEL,R. J.H. OKKERand B. J.J. LUTENBERG, 1987 USA 74: 5463-5467. Induction of the nodA promoter of Rhizobium leguminosarum SCHOFIELD,P. R.,and J. M. WATSON,1986 DNA sequenceof Sytn pRL1Jl by plant flavanones and flavones. J. Bacteriol. Rhizobium trqolii nodulationgenes reveals a reiteratedand 169: 198-204. potentially regulatory sequence precedingnodABC and nodFE. ZABAROVSKY,E. R.,and R. L. ALLIKMETS,1986 An improved Nucleic Acids Res. 14: 2891-2903. technique for the efficient construction of gene libraries by SCHWEDOCK,J.,and S. LONG, 1989 Nucleotidesequence and partial filling-in of cohesive ends. Gene 42: 1 19-1 23. protein products of two new nodulation genes of Rhizobium Communicating editor: G. MOSIG