Characterization of the nifA regulatory gene of Rhizonium leguminosarum PRE

This thesis describes the characterization of the nif A regulatory gene of the pea endosymbiont Rhizobiumleguminosarum PRE.Chapter I gives a general overview on the regulation of nitrogen fixation in diazotrophs, with special focus on the regulatory NifA protein. The regulation of genes involved in nitrogen fixation in two bacteria is discussed in detail: the free living Klebsiellapneumoniae and the endosymbiont of alfalfa R . meliloti . Major differences exist between these organisms where the onset of nitrogen fixation is concerned. K . pneumoniae has a general nitrogen regulatory circuitry which senses an internal biochemical signal i.e. the level of available ammonia as defined by the glutamine to 2-ketoglutarate ratio, a high ratio indicating a surplus, a low ratio a deficit. Sensing of a N-deficit results is translated, through a chain reaction of protein modifications, into activation of the regulatory NtrC product by phosphorylation. The resulting NtrC-P activates transcription of the regulatory nif LA operon, which encodes the inhibitor NifL and the activator NifA. The Klebsiella NifA thereupon activates transcription of the genes involved in nitrogen fixation. In a recently published paper David et al. (1988) suggest that the onset of nitrogen fixation in R . meliloti starts with the sensing of the external oxygen level. The FixL protein is hypothesized to sense a decrease in oxygen level. This protein is thus activated and in turn activates the FixJ protein, which directly or indirectly activates transcription of the nif A gene. The Rhizobium NifA protein activates transcription of the nitrogen fixation genes. In this overview we hypothesize that the oxygen sensing protein FNR instead of FixL senses the internal oxygen level. FNR then activates transcription of the fix LJ operon. The FixL protein may be a moderator of the activity of FixJ, comparable to the role of NtrB in activating NtrC. To date all rhizobial NifA proteins, in contrast to Klebsiella NifA, were shown to be oxygen sensitive. The structural analysis of the NifA protein is described and possible functions ascribed to domains identified in this protein are discussed. A model for NifA activity emerging from data presented for K . pneumoniae isdiscussed. At present a complete model cannot be presented for Rhizobiaceae . The similarities and differences between the models for K . pneumoniae and Rhizobium species are discussed.In Chapter 2 the DNA sequence and deduced amino acid sequence of R . leguminosarum PRE are presented. The amino acid sequence differs in 30 amino acids from that published for R . leguminosarum 3855 (Grönger et al., 1987). A possible explanation for this difference is discussed. The NifA Open Reading Frame (ORF) reveals two potential translation start sites, which in a heterologous E.coli background appear to be used both. The second translation start, which leads to a 488 amino acids, 53 kD protein, is preferred over the first, which leads to a 519 amino acids, 56.1 kD protein. The R . meliloti (Weber et al., 1985, Buikema et al., 1985) and B . japonicumnif A genes (Thöny et al., 1987) also have two translation start sites. It was shown for R . meliloti NifA (Beynon et al., 1988) the full length protein is the active form in an E . coli background. It is discussed that a translational preference for the second translational start site, leading to the inactive protein, as was found in pulse labeling experiments in E . coli may also exist in Rhizobium . We therefore suggest that the experiments presented by Beynon et al. (1988) are not conclusive as to the size of the functional protein in a Rhizobium background. Primer extension experiments and S 1 -nuclease protection were used to identify the putative nif A promoter. A transcription terminator was identified by S1-nuclease protection.Chapter 3 deals with a phenomenon reported by Hawkins and Johnston (1988) and Roelvink et al. (1988). A nif A::Tn 5 mutant can not be complemented by a plasmid having only the nif A coding DNA fragment. A detailed analysis of the nif A- nif B intergenic region is presented. The nif A gene has a transcriptional terminator typical of bacterial genes (Brendel et al., 1986) consisting of a four GC basepairs stem and a nine base loop followed by a thymidine rich DNA stretch. This terminator was sapped by S1-nuclease protection. The nif B gene has a RpoN dependent promoter, having all nucleotides thought to be crucial to its activity. The nif A terminator was fused to the Tet - promoter and this fusion was cloned in a low copytranscriptional lac Z vector. The results show that the nif A terminator allows 85% readthrough. RNA::DNA hybridisation studies show that the nif A gene is transcribed at a level twice of that of nif B. By using a plasmid, which has a DNA region encompassing nif A, nif B and a ferredoxin like gene downstream of nifB (Grönger et al., 1988, Klipp et al., 1988) it was shown that nif A::Tn 5 mutants can be fully complemented. Taken together these findings suggest that the nif A and the nifB gene are in one operon. The failure of plasmids having the nifA encoding DNA fragment alone to complement a nif A::Tn 5 mutant results from a polar effect of the Tn 5 transposon on nif B transcription.Chapter 4 deals with the nif H promoter region of R . leguminosarum PRE as one of the target sites of the NifA protein. We determined the nucleotide sequence of this region and identified a pseudo upstream activator sequence (UAS), a pseudo promoter, a consensus UAS and a consensus promoter. The promoter, mapped by primer extension experiments, differs from the consensus in one of the nucleotides thought to be invariant (see Gussin et al., 1986). The function of the nif H promoter elements was tested in a heterolo gous E . coli and a homologous Rhizobium background. Fusions of the nif H promoter region to lac Z, and fusions of deleted nif H promoter regions to lac Z, were used in activation studies byE .pneumoniae NifA in E . coli . Both high and low copy (deletion) nif H:: lac Z fusions were conjugated to Rhizobium . The activation study in an E . coli background showed that the pseudo UAS and the pseudo promoter are not involved in the function of the promoter. A different result was obtained with low copy nif H:: lac Z constructs in a Rhizobium background. The construct having both pseudo and consensus UAS, when compared with a construct having the consensus UAS only, seems to delay the onset of nitrogen fixation by three days. We suggest that this indicates that the presence of one or more UAS's modulates the expression of nif and fix genes, as was suggested for UAS's of B . Japonicumnif and fix genes (Gubler and Hennecke, 1988). A nif promoter region holding a UAS, when cloned in a multi copy vector, can inhibit nitrogen fixation by capturing the NifA activator needed for expression of nif and fix genes. A multicopy Inhibition study with (deleted) nif H:: lac Zfusions led to a surprising finding: deletion of part of the consensus UAS on the multicopy plasmid did not result in inhibition of nitrogen fixation. The relevance of this finding is discussed. We conclude that R . leguminosarumnif H can function without an UAS as was found forR .meliloti nifH inplanta (Better et al., 1985). We suggest that NifA way form a complex withRpoN-RNAPthat can bind directly at the promoter to activate transcription.