The research goal of this section is to understand how nutrient availability coordinates global patterns of gene expression as well as pathway-specific regulation in bacteria. Such complex networks integrate synthesis of macromolecules and regulate expression of the genomic repertoire. We focus on the roles of two regulatory nucleotides that are widespread in bacteria and that have recently been found in plants. Collectively called (p)ppGpp, they are analogs of GTP and GDP with pyrophosphate residues on the ribose 3'-hydroxyl group. Nutrient limitation, whether starving for amino acids, phosphate, nitrogen, or energy sources, causes fluctuations of (p)ppGpp levels. A regulatory role is assigned to (p)ppGpp because artificial induction of (p)ppGpp without nutrient limitation gives many of the same regulatory effects as starvation itself. Fluctuation of (p)ppGpp can also be an important element in adaptive responses to starvation, such as insuring survival by induction of the stationary phase-specific sigma factor (RpoS) of RNA polymerase. Regulatory responses to (p)ppGpp are thought to occur by effects on transcription, translation, and metabolism. We wish to understand these mechanisms as well as how nutrient limitation leads to changes in (p)ppGpp levels..

(p)ppGpp Effects on rpoS mRNA Translational Efficiency
Brown, Szalewska-Palasz, Murphy, Cashel in collaboration with T. Elliotta
We find that induction of RpoS by (p)ppGpp occurs by changing the efficiency of translation of rpoS mRNA instead of through the transcriptional effects on mRNA abundance. We have come to believe the dksA gene is involved; in the absence of (p)ppGpp, a dksA deletion blocks (p)ppGpp induction of RpoS, and overexpression of DksA induces RpoS. We have found that the effects of (p)ppGpp and DksA are exerted on a region of RpoS mRNA far upstream of the rpoS mRNA leader region, where the SD sequence is known to be sequestered by RNA folding and regulated by other factors. At the same time, DksA effects are puzzling because increased (p)ppGpp does not induce DksA nor increased DksA induce (p)ppGpp. Though at the transcriptional level, another suggestion of a relation between DksA and (p)ppGpp is based on the finding that the five amino acids required in a dksA deletion are a subset of those nine amino acids that are required in a complete (p)ppGpp deficiency [(p)ppGpp0]. Finally, RNA polymerase mutants that suppress the multiple amino acid auxotrophy of (p)ppGpp0 strains also suppress the five amino acid requirements of dksA deleted strains.

Suppressors of a (p)ppGpp-Deficiency Occur Exclusively in RNA Polymerase rpoB, rpoC and rpoD Subunit Genes
Szalewska-Palasz, Murphy, Cashel
While most (p)ppGpp regulation is thought to act directly on RNA polymerase initiation or elongation, verification of these notions with pure transcription components has proven elusive. Therefore, we have taken an alternative genetic approach of isolating mutants that suppress cellular phenotypes in response to a complete deficiency of (p)ppGpp. Accordingly, we have identified over 50 distinct lesions in the two largest RNA polymerase subunits. Interestingly, by using the crystal structure of the RNA polymerase core enzyme recently resolved by the Darst laboratory, we find that the sites of these amino acid changes are almost exclusively on enzyme surfaces deduced to involve DNA contacts. We have now isolated strains with multiple lesions and can ask whether there are spatially grouped subclasses of lesions that are functionally synergistic or antagonistic.

A (p)ppGpp-Synthesizing and -Degrading Enzyme from Streptococcus Displays Reciprocal Regulation of Bifunctionality as an Intrinsic Enzymatic Property
Mechold, Brown, Murphy, Cashel in collaboration with T. Hoggb
Most Gram-positive bacteria possess a single enzyme capable of both (p)ppGpp synthesis and degradation. A futile metabolic cycle would occur if both were fully active. This means that net activity requires a switch within the enzyme that activates one activity while simultaneously curtailing the opposing activity. We have found that the function of the C-terminal half of the 84 kDa protein is to regulate the opposing functions present in the N-terminal half protein in a reciprocal manner. To explore this switch further, we have isolated within the N-terminal half of the protein single missense suppressor mutations that restore reciprocal activation and repression of separate domains encoding (p)ppGpp degradation and synthesis. The N-terminal half protein, in which reside opposing catalytic domains, has been crystallized and its structure determined. Our intragenic allele-specific suppressors appear to map on the boundaries separating the two domains, but the mechanism of reciprocal activation and inhibition built into the structure remains to be elucidated.

Mapping the Ribosomal Binding Sites of RelA, the Major (p)ppGpp Synthetase from E. coli
Glaser, Murphy, Cashel in collaboration with laboratory of Al Dahlbergc
Amino acid starvation by codon-specified uncharged tRNA binding has long been known to induce synthesis of (p)ppGpp on the ribosome by the RelA protein, but few details are available on the RelA-ribosome interaction. We have now mapped RelA differences in dimethylsulfate modification of rRNA by RelA binding. With recently solved ribosome structures from the Noller laboratory as a guide, we have localized perturbations of rRNA structure on the interior surfaces of both the large and small ribosomal subunits. The DMS modifications accompanying RelA binding are found to occur in the region of the CCA end of bound tRNA as well as in the decoding region of A site bound tRNA, where the codons pair with anticodons. The conditions for binding examined are not physiological because neither mRNA nor codon-specified uncharged tRNA are present. Our next goal is to ask how rRNA conformational perturbations are altered by the presence of more physiological conditions and how specifically the RelA protein senses the difference between the binding of codon-specified uncharged and charged tRNA.