Living without Fur: the subtlety and complexity of iron-responsive gene regulation in the symbiotic bacterium Rhizobium and other alpha-proteobacteria.

The alpha-proteobacteria include several important genera, including the symbiotic N(2)-fixing "rhizobia", the plant pathogen Agrobacterium, the mammalian pathogens Brucella, Bartonella as well as many others that are of environmental or other interest--including Rhodobacter, Caulobacter and the hugely abundant marine genus Pelagibacter. Only a few species--mainly different members of the rhizobia--have been analyzed directly for their ability to use and to respond to iron. These studies, however, have shown that at least some of the "alphas" differ fundamentally in the ways in which they regulate their genes in response to Fe availability. In this paper, we build on our own work on Rhizobium leguminosarum (the symbiont of peas, beans and clovers) and on Bradyrhizobium japonicum, which nodulates soybeans and which has been studied in Buffalo and Zürich. In the former species, the predominant Fe-responsive regulator is not Fur, but RirA, a member of the Rrf2 protein family and which likely has an FeS cluster cofactor. In addition, there are several R. leguminosarum genes that are expressed at higher levels in Fe-replete conditions and at least some of these are regulated by Irr, a member of the Fur superfamily and which has the unusual property of being degraded by the presence of heme. In silico analyses of the genome sequences of other bacteria indicate that Irr occurs in all members of the Rhizobiales and the Rhodobacterales and that RirA is found in all but one branch of these two lineages, the exception being the clade that includes B. japonicum. Nearly all the Rhizobiales and the Rhodobacterales contain a gene whose product resembles bona fide Fur. However, direct genetic studies show that in most of the Rhizobiales and in the Rhodobacterales it is a "Mur" (a manganese responsive repressor of a small number of genes involved in Mn uptake) or, in Bradyrhizobium, it recognizes the operator sequences of only a few genes that are involved in Fe metabolism. We propose that the Rhizobiales and the Rhodobacterales have relegated Fur to a far more minor role than in (say) E. coli and that they employ Irr and, in the Rhizobiales, RirA as their global Fe-responsive transcriptional regulators. In contrast to the direct interaction between Fe2+ and conventional Fur, we suggest that these bacteria sense Fe more indirectly as functions of the intracellular concentrations of FeS clusters and of heme. Thus, their "iron-omes" may be more accurately linked to the real-time needs for the metal and not just to its absolute concentration in the environment.