Two inverted repeats in the nodD promoter region are involved in nodD regulation in Rhizobium leguminosarum

Two inverted repeats in the nodD promoter region are involved in nodD regulation in Rhizobium leguminosarum

Gene, 144 (1994) 87-90 0 1994 Elsevier Science B.V. All rights reserved. 87 0378-l 119/94/$07.00 GENE 07993 Two inverted repeats in the nodD promo...

524KB Sizes 0 Downloads 7 Views

Gene, 144 (1994) 87-90 0 1994 Elsevier Science B.V. All rights reserved.


0378-l 119/94/$07.00

GENE 07993

Two inverted repeats in the nodD promoter region are involved in nodD regulation in Rhizobium leguminosarum (NodD protein; regulation; repression; RNA polymerase; transcription)

Chengjian Mao’, J. Allan Downieb and Guofan Hong” “Shanghai Institute of Biochemistry, Shanghai 200031, China: and bJohn Innes Institute, Colney Lane, Norwich NR4 7UH, UK. Tel. (44-603) 52-571 Received by K.F. Chater:

21 July 1993; Revised/Accepted:

16 December

1993/4 January

1994; Received at publishers:

24 March



In Rhizobium leguminosarum (R. 1.) biovar viciae, the nodulation gene nodD encodes a transcriptional activator (NodD) which binds to highly conserved DNA sequences (nod-boxes) in the promoters of other nod operons. In addition, NodD represses nodD transcription and this occurs at the divergent and overlapping nodd-nodD promoters. We mutagenised this region with hydroxylamine, and by cloning the mutagenised DNA into a vector carrying the 1acZ reporter gene downstream from the cloning site identified mutations affecting nodD expression and repression. The resulting plasmids were transferred to R. 1. uiciae strains containing or lacking nodD. Two classes of promoter mutants were identified: those in which nodD transcription was altered and those in which NodD-dependent repression was altered. The nucleotide (nt) changes in the promoter region were found to be located within two inverted repeat sequences (A2 and A3) which are about 70 bp apart. A2 is important for nodD transcription and A3 (which is upstream from A2) is involved in NodD-dependent repression. The nt sequence at A3 shows some homology to the nod-box region of the nodA promoter. It is proposed that the NodD-dependent repression occurs as a result of NodD binding to both A3 and the nodA nod-box, forming a loop which prevents transcription of nodD from its promoter, AZ, which lies between A3 and the nod-box. This model is supported by the observation that there are at least three sites for NodD binding in the nodA-nodD promoter region.


The nodulation of leguminous plants by rhizobia is dependent on the expression of a group of genes collectively known as the nod or no1 genes. Several of these Correspondence to: Dr. G.F. Hong, Shanghai Institute of Biochemistry, Shanghai 200031, China. Tel. (86-21) 437-4430; Fax (86-21) 433-8357. Abbreviations: A, absorbance (1 cm); BGal, B-galactosidase; bp, base pair(s); Del, deletion; Km, kanamycin; nod, nodulation genes; nod-box, highly conserved DNA sequences in nod promoters; NodD, protein encoded by nodD; nodLIp. nodD promoter; nt, nucleotide(s); R. l., Rhizobium leguminosarum; Sm, streptomycin; TBE buffer, 0.09 M Tris/0.09 M boric acid/l mM EDTA pH 8.3; Tc, tetracycline; TE buffer, 10 mM Tris/l mM EDTA pH 8.0; TY medium, complete medium for Rhizobium; wt. wild type; XGal, 5-bromo-4-chore-3-indolyl-B-ogalactopyranoside; [I, denotes plasmid-carrier state. SSDI 0378-1119(94)00204-6

genes are involved in the synthesis of highly specific lipooligosaccharide nodulation factors which induce nodule morphogenesis and determine the range of legumes nodulated (Dtnarit et al., 1992). Most of the operons containing the nod and no1 genes are under the control of a group of positively acting transcriptional regulators, the products of nodD genes. These NodD proteins belong to the LysR family of transcriptional regulators and bind to the promoter regions of nod gene operons (Schlaman et al., 1992). These promoters regions contain a highly conserved DNA sequence (known as a nod-box) to which NodD binds. The NodD proteins probably bind flavonoid or isoflavonoid molecules secreted from legume roots and then induce the expression of the other nod and no1 genes. The R. 1. viciae nodD gene is expressed in the absence


of ~avonoid inducers (Rossen et al., 1985), and the ~~~~ gene product represses the expression of [email protected] (i,e,, it is autoregulatory), a characteristic that is not seen with R, meliloti. The R. 1. viciae nodD gene is expressed divergently from the nodABCIJ genes, which are under NodD control (Rossen et al., 1985) and it is likely that the two promoters overlap. To analyse the regions of the n&D promoter (nodDp) important for nodD transcription and autoregulation, we carried out a random mutagenesis of nodDp region from R. 1. viciae and identified nt changes which affected nodD regulation.


(a) Isolation of nodD promoter mutants Following hydroxylamine mutagenesis (Humphreys et al., 1976) of pND487 DNA (Ni et al., 1991), the EcoRIPstI fragment carrying the R. I. uiciae nodD-nodA promoter region was excised and recloned into EcoRI + PstIcut pMP220 (Spaink et al., 19X7), such that the fucZ gene in pMP220 was under the control of nodDp. The recombinant plasmids were used to transform E. coEi and then transferred by conjugation to R. 1. strain 84Ol{pIJ1518] (Lamb et al., 1982), which carries the cloned nodD gene from R. 1. viciae (Rossen et al., 1985) cloned in the vector pKT230 (Bagdasarian et al. 1981). Approx. 1000 individual transconjugants were randomly picked and screened for quantitative alterations in BGal activity (Rossen et al., 1985) to identify mutations in nodDp. More than 40 potential mutants were obtained. Plasmid DNA from each was isolated and re-transferred into R. 1. strain 84Ol[pIJl518~ (+nodD) and into 84Oi[pKTZ30] (-nodf)). Subsequently, 21 potential mutants were identified as having either higher or lower leveis of @Gal than the control 8401[pCMl], (pMP220 carrying the unmutagenised EcoRI-PstI fragment). The plasmids were analysed by gel electrophoresis and by sequencing the DNA insert. Those that had no inserts, had inserts containing deletions or had multiple mutations, were discarded. Ten plasmids with single mutations were identified. The relevant BGal activities are shown in Table I. In the absence of nodD, two of the mutant plasmids, pCM27 and pCM25, gave activities lower than the wt control pCM1. pCM19, pCM26, pCM31, pCM32 and pCM33 exhibited levels of activity that were higher than the control. In the presence of nodD, pCM11, pCM20 and pCM21 gave lower levels of activity than the control, indicating an increased level of NodD-mediated repression. In the presence of nodD, pCM19, pCM25, pCM26, pCM31, pCM32 and pCM33 exhibited levels of activity higher than the control, implying that nodD-mediated repression was partially relieved.

TABLE I Measurements of the effects of mutations on the activities of nodD-lucZ transcriptional fusions Plasmid”



Units of j3Gal activityd

promoterb change”

pMP220 pCMl pCMl1 pCM20 pCM21 pCM26 pCMl9 pCM25 pCM27 pCM31 pCM32 pCM33

none wt nodDp1 tl0dDp2

nod&2 nodDp3

nodDp4 nodDp5

nodDp6 nodDp7 nodDp7 nodDp1

179;G+A 179;G+A 179:G-+A 179;GjDel 26&C+T 257:G+A 263;G+A 255;G+A ?SS;G+A 255:G-tA

pKT230 (- nodD)

(+ nodQ)

22 358 353 314 347 452 750 236 69 539 606

25 78 45 49 57 202 117 111 51 298 277




“The plasmids carrying the mutated nodD promoters upstream of 1ucZ in pMP220 (Spaink et al., 1987) were transferred to R. 1. strain 8401 (Lamb et al., 1982) containing no& on pIJl518 (Rossen et al., 1985) or the control pKT230 vector (Bagdasarian et al.. 1981) which lacks FZOdD.

bThe different mutant promoter alleles are defined in the legend to Fig. 1. ‘The nucleotide changes are numbered based on the sequence presented by Shearman et al. (1986). dMeasurements of @al were made as described by Rossen et al. (1985) using an automated microplate reader (BioTEK) to measure the cell A 540 nm and the hydrolysis of o-nitrophenyl galactose at AJo5 ,,m. The cells were grown on TY medium (Beringer, 1974) containing 20 ug Km/ml, 200 ng Sm/ml and 5 pg Tc/ml to an A,,, nm of 0.5. The data are the averages of five individual measurements. The standard errors were less than 10% of the average values shown.

(b) Identification of mutations in the nodl) promoter The nt sequences of the inserts in pCM20 and pCM21 turned out to be identical, as were those of pCM31, pCM32 and pCM33. It is possible that these identical plasmids are siblings. In all, seven different mutant alleles were identified and called nodDp1 through to nodDp7 (Table I). The different nt changes identified in the ten mutant plasmids are summarised in Fig. 1. Six of the changes are nt substitutions and one (nodDp4 in pCM26) is a single nt deletion. Significantly, all of the changes occurred within two regions of dyad symmetry, A2 and A3, which had been noted previously by Shearman et al. (1986), in the sequence upstream from the nodD gene. The mutant alleles nodDp1 and nodDp2 contain substitutions in the region of dyad symmetry A3 (Fig. 1). Significantly, these mutations did not influence nodDp activity when NodD was absent, although they did increase the level of NodD-mediated repression of nodDfacZ expression (Table I). This indicates that these two mutations influenced the interaction between NodD and


sion of nodD expression although at a lower level than normal.


(c) NodD may bind to region A3

Fig. 1. Mutations in the no&p region. The nt sequence shown extends from the proposed translation start of nodA (Rossen et al., 1984, allowing for a sequencing to the proposed scribed

by Shearman

and a region

error as corrected


in EMBL J. accession

of nodD showing


et al. (1986). The nodA nod-box

of dyad symmetry



nt 92-356

it is arrowed.

as de-

region is boxed The domains


which NodD binds (Lu et al., 1994) are underlined. The regions of dyad symmetry A? and A3 overlap with those described by Shearman et al. (1986). The altered

nt are indicated

above the vertical


and the

numbers in parentheses indicate the mutant nodD promoter allele numbers as presented in Table I (e.g., A( 1) corresponds with a G+A substitution

in the nodDp1 allele on pCMl1).

by a deletion

and is marked

double-stranded ribonucleotide

The nodDp3 allele was caused

Del. The nt sequence

was determined


DNA sequencing of the plasmids using the oligodeoxyprimer 5’-TATATCAACGGTGGTATATCCAGTG,

which is homologous to the nt sequence downstream from the PstI site in pMP220. Since the vector is a relatively large low-copy-number plasmid, it was necessary DNA fragments

to separate

the plasmid

before sequencing.

5 ml TE buffer containing 37’C for 2-4 h. The DNA


0.1 mg DNase-free

DNA from contaminating of DNA were dissolved


RNase and incubated


was then precipitated

with ethanol,


solved in 300 pl TE buffer and loaded onto a gel made up with 1.2% low-melting-point agarose in TBE. Samples were electrophoresed (3 V/cm) for 1 h at 4’C and then stained

with ethidium



plasmid DNA was excised, the DNA extracted from the melted agarose and taken up in 40 ~1 of TE buffer. Aliquots of 4 pl were used in the sequencing reactions using a Bio-Rad kit following the manufacturers instructions.

this region dyad symmetry. It should be noted that since the level of NodD-mediated repression of nodD-1acZ was already relatively strong, due to the multicopy nature of pIJ1518, any further repression of the activity (as in the case of nodDp1 and nodDp2) must be due to a relatively strong effect. Conversely, the deletion allele nodDp3, which is also in region A3 (Fig. l), significantly reduced the level of NodD-mediated repression (Table I), also indicating that this region may be involved in an interaction with NodD. Four of the mutations (nodDp4, nodDp5, nodDp6 and nodDp7) were in the region A2 (Fig. 1). One (nodDp6), essentially abolished the nodDp activity (Table I), indicating that this region is probably involved in the formation of the a transcription initiation complex. The nodDp5 allele also decreased transcription, although only by about 30%. The nodDp4 and nodDp7 alleles significantly increased the transcriptional activity, also implying that this region is important for initiation of nodD transcription. The three alleles with active promoters (nodDp4, nodDp5 and nodDp7) retained NodD-dependent repres-

On the basis that mutations in region A2 affected nodD transcription, it is reasonable to conclude that A2 is part of the nodDp. However, this would mean that region A3, involved in the NodD-mediated repression, is upstream from the promoter. A3 has similarity to the nod-box region that makes up part of the promoter of nodA, which is transcribed divergently from nodD (Fig. 1). Because of the dyad symmetry within A3 and within the nod-box it is possible to align A3 in both orientations with regard to the nod-box. Both alignments are shown (Fig. 2) and it is evident that one section of the nodA promoter region has 14 identities within a stretch of 21 nt from the A3 region when aligned in either orientation relative to the nodA nod-box (Fig. 2). Cao and Hong (1992) showed that increasing amounts of NodD protein added to constant amounts of the radioactively-labelled nodA-nodD promoter region revealed at least seven retarded bands. NodD bound to different individual sites on the nodA-nodD promoter region might be seen as complexes retarded to different extents as a result of the different conformations that might be formed. The presence of at least seven retarded bands indicates that there must be a minimum of three separate NodD-binding domains in the nodA-nodD promoter region. Fisher and Long (1993) and Lu et al. (1994) have demonstrated using footprinting analysis that nod-box regions have two separate NodD-binding domains. The effects of mutations in the A3 region and its similarity with the nodbox region suggest that region A3 would also be a likely candidate for NodD binding. Although Lu et al. (1994) did not observe such binding, milder conditions for footprinting are being tried to detect NodD binding to region A3. The mutation in nodDp2 alters the sequence of region A3 such that it comes closer to the consensus sequence

A3 invert part


of nodd nod-box

A3 direct


A(2);De1.(3) A(1) A L 5'GATTGCCGGTTAGGCAATCGA3' ******* 5'GATTGCCATC&2':;::3'






Fig. 2. Alignment of region A3 with part of the nodA nod-box. The nodbox iswrittensuch that the nodA gene is downstream from its 3’ end. The region A3 is aligned with the nod-box either as an inverted fragment (showing the opposite strand to that of the nod-box) or as a direct fragment (showing the same strand as the nod-box). The mutations associated with the three mutant nodDp promoter alleles 1, 2 and 3 are indicated with arrows following the convention in Fig. 1. Identical nt are marked

by asterisks.


of the nod-box when A3 is aligned in the opposite orientation to the nodA nod-box (Fig. 2). It might, therefore, be possible to detect binding between NodD and the mutant nodDp2 sequence. An explanation for the NodD-mediated repression of nodD expression in R. 1. viciae could be that NodD binds to the A3 region and interacts with NodD bound to the nodA nod-box. This would create a loop containing region A2. The RNA polymerase might, thus, be prevented from initiating transcription of nodD from A2, thereby explaining the NodD-dependent repression of nodD transcription via a site upstream from the presumed transcription start point of nodD. One prediction of such a model would be that some mutations in region A3 would relieve repression, and such an effect was observed with nodDp3. In contrast, the level of repression was increased with nodDp2. This mutation makes A3 (in the inverted orientation) more similar to the consensus nod-box sequence (Fig. 2). If this were correlated with an increase in the stability of NodD binding at A3, then it might result in increased stability of the proposed loop containing A2, thus increasing NodD-mediated repression of nodD expression. In the nodA-nodD1 promoter region of R. meliloti (Tiirok et al., 1984) there is not an A3-like region containing a partial nod-box and there is no NodD-mediated repression of nodD transcription (Mulligan and Long, 1985). A second prediction of this model might be that mutations within the nod-box region would reduce the NodD-mediated repression and initially we were surprised not to find such mutants. However, since the nodA nod-box has two NodD-binding domains (Fisher and Long, 1993; Lu et al., 1994), it would be very unlikely that a double mutant would be isolated in which both NodD-binding domains in the nodA nodbox region are lost. The biological significance of the NodD-mediated repression has not yet been established. Since overexpression of the nod genes in R. 1. viciae inhibits nodulation of both peas and vetch (Knight et al., 1986) the autoregulation may serve to reduce overall nod gene expression by reducing the availability of the nodD gene product.

him a Biotechnology Career Fellowship. J.A.D. was supported by a grant in aid from the British Agricultural&Food Research Council.

REFERENCES Bagdasarian,

M., Lurz, R., Riickert,

B., Franklin,

F.C.H., Bagdasarian,

M.M., Frey, J. and Timmis, K.N.: Specific-purpose vectors.

II. Broad

vectors, and a host-vector Gene 16 (1981) 237-247. Beringer,

J.E.: R-factor

in Rhizobium



can form as many


RSFlOlO-derived in Pseudomonas.

system for gene cloning


Microbial. 84 (1974) 188-189. Cao, H.M. and Hong, G.F.: nodD gene product tergenic


host range, high copy number,

J. Gen.

and the nodA-D in-

as seven complexes.

Chin. Sci.

Bull. 37 (1992) 951-953. DCnariC, J., Debelle, F. and Rosenberg, C.: Signalling and host range variation in nodulation. Annu. Rev. Microbial. 46 (1992) 4977531. Fisher, R.F. and Long, S.R.: Interaction binds to two distinct

of NodD at the nod-box: NodD

sites on the same face of the helix and induces

a bend in the DNA. J. Mol. Biol. 233 (1993) 336-348. Humfreys, G.O., Willshaw, G.A., Smith, H.R. and Anderson, Mutagenesis

of plasmid

tants of multi-copy

DNA with hydroxylamine:


Mol. Gen. Genet.


E.S.: of mu-

145 (1976) 101-108.

Knight, C.D., Rossen, L., Robertson, J.G., Wells, B. and Downie, J.A.: Nodulation inhibition by Rhizobium leguminosarum multicopy nodABC




of early

J. Bacterial. 166 (1986) 552-558. Lamb, J.W., Hombrecher, G. and determined




of plant




abilities in Rhizobium

and nitrogen-fixation

phaseoli. Mol. Gen. Genet. 186 (1982) 449-452. Lu, Y.Y., Xu, Y.C. and Hong, G.F.: The DNA sequences NodD


have been determined.

in press. Mulligan, J.T. and Long, S.R.: Induction by plant



(1985) 6609-6613. Ni, F.D., Xu, Y.C. and Hong, DNA complexes



G.F.: The formation

36 (1991) 1126-1129. Rossen, L.. Johnston, A.W.B. and Downie, nodulation


of Rhizobium nodC expression

nodD. Proc.

in the solid electrophoresis

Rhizobium leguminosarum


Sci. Bull. 38 (1994)


Sci. USA 82

of specific protein-


Chin. Sci. Bull.

J.A.: DNA sequence

of the

genes nodA, B and C required

for root hair curling. Nucleic Acids Res. 12 (1984) 9497-9508. Rossen, L., Shearman, C.A., Johnston, A.W.B. and Downie, J.A.: The nodD gene of Rhizobium leguminosarum is autoregulatory and in the presence of plant exudate induces the nodABC genes. EMBO J. 4 (1985) 3369-3373. Schlaman, H.R.M., Okker, of nodulation

R.J.H. and Lugtenberg,

gene expression

by NodD

B.J.J.: Regulation

in rhizobia.

J. Bacterial.

174 (1992) 5177-5182. Shearman. C.A., Rossen, L., Johnston, A.W.B. and Downie, J.A.: The Rhizobium gene nodF encodes a protein similar to acyl carrier proACKNOWLEDGEMENTS

G.H. was supported by the Chinese ‘863 plan’ and ‘pandeng plan’. CM. was supported by a graduate student fellowship from the Chinese Academy of Sciences. G.H. would like to thank Rockefeller Foundation for providing

tein and is regulated by nodD plus a factor in pea root exudate. EMBO J. 5 (1986) 647-652. Spaink. H.P., Okker, R.J.H.. Wijffelman, C.A., Pees, E. and Lugtenberg, B.J.J.: Promoters in the nodulation region of the Rhizobium [eguminosarum Sym plasmid pRLlJ1. Plant Mol. Biol. 9 (1987) 27-39. Tiirok, I., Kondorosi, E., Stepkowski, T., Posfai, J. and Kondorosi, A.: Nucleotide sequence of Rhizobium meliloti nodulation genes. Nucleic Acids Res. 12 (1984) 9509-9524.