Expression of the fixR-nifA operon in Bradyrhizobium japonicum depends on a new response regulator, RegR

Abstract

Many nitrogen fixation-associated genes in the soybean symbiont Bradyrhizobium japonicum are regulated by the transcriptional activator NifA, whose activity is inhibited by aerobiosis. NifA is encoded in the fixR-nifA operon, which is expressed at a low level under aerobic conditions and induced approximately fivefold under low-oxygen tension. This induction depends on a -24/-12-type promoter (fixRp1) that is recognized by the sigma54 RNA polymerase and activated by NifA. Low-level aerobic expression and part of the anaerobic expression originates from a second promoter (fixRp2) that overlaps with fixRp1 and depends on an upstream DNA region (UAS) located around position -68 (H. Barrios, H. M. Fischer, H. Hennecke, and E. Morett, J. Bacteriol. 177:1760-1765, 1995). A protein binding to the UAS was previously postulated to act as an activator. This protein has now been purified, and the corresponding gene (regR) has been cloned. On the basis of the predicted amino acid sequence, RegR belongs to the family of response regulators of two-component regulatory systems. We identified upstream of the regR gene an additional gene (regS) encoding a putative sensor kinase. A regR mutant was constructed in which neither a specific UAS-binding activity nor fixRp2-dependent transcript formation and fixR'-'lacZ expression was detected in aerobically grown cells. Anaerobic fixR'-'lacZ expression was also decreased in regR mutants to about 10% of the level observed in the wild type. Similarly, regR mutants showed only about 2% residual nitrogen fixation activity, but unlike nodules induced by nifA mutants, the morphology of those nodules was normal, displaying no signs of necrosis. While regR mutants grew only slightly slower in free-living, aerobic conditions, they displayed a strong growth defect under anaerobic conditions. The phenotypic properties of regS mutants differed only marginally, if at all, from those of the wild type, suggesting the existence of a compensating sensor activity in these strains. The newly identified RegR protein may be regarded as a master regulator in the NifA-dependent network controlling nif and fix gene expression in B. japonicum.

FIG. 1

Regulatory scheme and dual promoter of the B. japonicum fixR-nifA operon. P1 and P2 are the transcriptional start sites of transcripts T1 and T2 (see text) that originate from the overlapping promoters fixRp₁ and fixRp₂, respectively.

FIG. 2

Purified UBP. (A) Silver-stained SDS–13% PAGE gel of protein fractions taken after elution from the DNA-Sepharose affinity column, the last purification step (see Materials and Methods). One microgram of protein from the fractions that were eluted with 250 mM (lane 1), 300 mM (lane 2), 350 mM (lane 3), and 400 mM KCl (lane 4) and 0.5 μg of protein from the fraction that was eluted with 450 mM KCl (lane 5) were loaded. The prominent 21-kDa bands in fractions 2 to 4 were excised and subjected to N-terminal amino acid sequencing as described in Materials and Methods. (B) Gel retardation assay of the samples shown in panel A. Protein (0.1 μg) was mixed with 0.1 ng of ³²P-labeled double-stranded 32-bp oligonucleotide spanning the fixR-nifA −68 region (UAS) and incubated in the presence of poly(dI-dC) as a nonspecific competitor. The binding products were separated on a 6% nondenaturing polyacrylamide gel and visualized by phosphorimager analysis.

FIG. 3

Physical map of the B. japonicum regSR region and genetic structure of regS and regR mutations. (A) The physical map shows relevant restriction sites (B, BamHI; E, EcoRI; H, HindIII; M, MscI; N, NotI; Nd, NdeI; P, PstI; Pv, PvuII, S, SmaI), the location and orientation of regS and regR, and one additional open reading frame, orf1. Numbers at the top denote the nucleotide positions starting from the PvuII site upstream of regS. The hatched bar indicates the extent of the sequence submitted to the EMBL Nucleotide Sequence Database. (B) The inserts of plasmids pRJ2400 and pRJ2403 are depicted. (C) The structure of regS and regR mutations is shown along with the corresponding strain numbers. Horizontal arrows indicate the orientation of the inserted resistance cassettes, Ω (Sm^r Sp^r) or aphII (Km^r). Restriction sites in parentheses were lost during cloning.

FIG. 4

Sequence alignments of the predicted B. japonicum RegR and RegS proteins to their homologs of R. capsulatus, R. sphaeroides, and S. meliloti. The consensus sequence (cons) in the bottom line was determined from amino acids identical in all four sequences. Putative essential domains are indicated by numbers in parentheses. Numbers above the sequence refer to positions in RegR and RegS. The sequences of RegA and RegB from R. sphaeroides, published in references and , which are very similar to those of PrrA and PrrB, have been omitted for clarity. (A) Alignment of B. japonicum RegR (BjRegR) to S. meliloti ActR (RmActR [59]), R. capsulatus RegA (RcRegA [53]), and R. sphaeroides PrrA (RsPrrA [13]). The experimentally determined N terminus of Ubp (RegR) is shaded in gray. The presumptive phosphorylation site is indicated by “(1),” and the helix-turn-helix motif is indicated by “(2).” (B) Alignment of B. japonicum RegS (BjRegS) to S. meliloti ActS (RmActS [59]), R. capsulatus RegB (RcRegB [38]), and R. sphaeroides PrrB (RsPrrB [14]). The presumptive autophosphorylation site is indicated by “(3)”; the putative kinase domain, composed of three essential regions, is indicated by “(4).”

FIG. 5

Anaerobic growth of B. japonicum wild type and nifA, regR, and regS mutants in YEM medium supplemented with KNO₃. Symbols: •, wild type; ○, nifA mutant A9; ▴, regS mutant 2409; ▵, regR mutant 2426. Samples were taken from three parallel cultures of each strain, and growth was determined by measuring the OD₆₀₀.

FIG. 6

UAS binding activity in extracts of regR and regS mutants analyzed by gel retardation. Crude extracts of aerobically grown cells of the strains indicated were prepared as described in Materials and Methods. Approximately 4 μg of protein was mixed with 0.1 ng of ³²P-labeled double-stranded 32-bp oligonucleotide spanning the fixR-nifA −68 region (UAS) and incubated in the presence of poly(dI-dC) as a nonspecific competitor. The binding products were separated on a 6% nondenaturing polyacrylamide gel and visualized by phosphorimager analysis. The identity of the nonspecific, slow-migrating complexes is not known.

FIG. 7

Primer extension analysis of the promoter fixRp₁- and fixRp₂-dependent fixR′-′lacZ transcripts in wild type and ΔUAS and regR mutants containing a chromosomally integrated fixR′-′lacZ fusion. Total RNA was purified from the indicated B. japonicum strains grown aerobically (+O₂) in PSY medium or anaerobically (−O₂) in YEM medium plus KNO₃. Hybridization to RNA with the ³²P-labeled oligonucleotides lac4B and PBj16S and reverse transcription of the fixR′-′lacZ mRNA and the 16S rRNA primary transcript, respectively, were performed as described in Materials and Methods. The products were separated on a 6% polyacrylamide gel next to a sequence ladder of plasmid pRJ7211 made with oligonucleotide lac4B. Transcripts T1, T2, and Bj16S (control) are marked. The origin of the unmarked reverse transcription products present in all lanes is not known; they had not been observed in similar, previous studies in which a shorter lacZ-specific oligonucleotide (lac4 [5]) was used.

FIG. 8

Electron micrographs showing the structure of soybean nodule cells infected by the regR mutant 2426 (A) and the wild-type 110spc4 (B). Black and white arrowheads in panel A mark starch granules and empty membrane vesicles, respectively. Bar = 5 μm.

FIG. 9

Transcriptional mapping of the regR promoter region. Total RNA was purified from B. japonicum wild type grown aerobically in PSY medium and used for primer extension. Hybridization to RNA with the ³²P-labeled oligonucleotide PEregR2 and reverse transcription of the regR mRNA were performed as described in Materials and Methods. The products were separated on a 6% polyacrylamide gel next to a sequence ladder of plasmid pRJ2403 made with oligonucleotide PEregR2. The sequence of the indicated region (positions 1448 to 1520 of the database-submitted nucleotide sequence) is denoted in the box. It contains the transcription start point (+1), the putative −10 and −35 regions, and the probable translation start site (TTG).