Activation of transcription by guanosine 5'-diphosphate,3'-diphosphate, transfer ribonucleic acid, and novel protein from Escherichia coli

Regulation through the secondary channel--structural framework for ppGpp-DksA synergism during transcription

Bacterial transcription is regulated by the alarmone ppGpp, which binds near the catalytic site of RNA polymerase (RNAP) and modulates its activity. We show that the DksA protein is a crucial component of ppGpp-dependent regulation. The 2.0 A resolution structure of Escherichia coli DksA reveals a globular domain and a coiled coil with two highly conserved Asp residues at its tip that is reminiscent of the transcript cleavage factor GreA. This structural similarity suggests that DksA coiled coil protrudes into the RNAP secondary channel to coordinate a ppGpp bound Mg2+ ion with the Asp residues, thereby stabilizing the ppGpp-RNAP complex. Biochemical analysis demonstrates that DksA affects transcript elongation, albeit differently from GreA; augments ppGpp effects on initiation; and binds directly to RNAP, positioning the Asp residues near the active site. Substitution of these residues eliminates the synergy between DksA and ppGpp. Thus, the secondary channel emerges as a common regulatory entrance for transcription factors.

Mechanism of regulation of transcription initiation by ppGpp. I. Effects of ppGpp on transcription initiation in vivo and in vitro

To determine the role of ppGpp in both negative and positive regulation of transcription initiation during exponential growth in Escherichia coli, we examined transcription in vivo and in vitro from the growth-rate-dependent rRNA promoter rrnB P1 and from the inversely growth-rate-dependent amino acid biosynthesis/transport promoters PargI, PhisG, PlysC, PpheA, PthrABC, and PlivJ. rrnB P1 promoter activity was slightly higher at all growth-rates in strains unable to synthesize ppGpp (deltarelAdeltaspoT) than in wild-type strains. Consistent with this observation and with the large decrease in rRNA transcription during the stringent response (when ppGpp levels are much higher), ppGpp inhibited transcription from rrnB P1 in vitro. In contrast, amino acid promoter activity was considerably lower in deltarelAdeltaspoT strains than in wild-type strains, but ppGpp had no effect on amino acid promoter activity in vitro. Detailed kinetic analysis in vitro indicated that open complexes at amino acid promoters formed much more slowly and were much longer-lived than rrnB P1 open complexes. ppGpp did not increase the rates of association with, or escape from, amino acid promoters in vitro, consistent with its failure to stimulate transcription directly. In contrast, ppGpp decreased the half-lives of open complexes at all promoters, whether the half-life was seconds (rrnB P1) or hours (amino acid promoters). The results described here and in the accompanying paper indicate that ppGpp directly inhibits transcription, but only from promoters like rrnB P1 that make short-lived open complexes. The results indicate that stimulation of amino acid promoters occurs indirectly. The accompanying paper evaluates potential models for positive control of amino acid promoters by ppGpp that might explain the requirement of ppGpp for amino acid prototrophy.

Mechanism of regulation of transcription initiation by ppGpp. II. Models for positive control based on properties of RNAP mutants and competition for RNAP

Strains containing ppGpp, a nucleotide whose synthesis is dependent on the RelA and SpoT proteins of Escherichia coli, display slightly lower rRNA promoter activity and much higher amino acid biosynthesis/transport promoter activity than deltarelAdeltaspoT strains. In the accompanying paper, we show that ppGpp directly inhibits rRNA promoter activity in vitro by decreasing the lifetime of the rrn P1 open complex. However, ppGpp does not stimulate amino acid promoter activity in vitro. We show here that RNA polymerase (RNAP) mutants, selected to confer prototrophy to deltarelAdeltaspoT strains, mimic the effects of ppGpp on wild-type RNAP. Based on the positions of the mutant residues that confer prototrophy in the structure of core RNAP, we suggest molecular models for how the mutants, and by analogy ppGpp, generally decrease the lifetime of open complexes. We show that amino acid promoters require higher concentrations of RNAP for function in vitro and in vivo than control promoters, and are more sensitive to competition for RNAP in vivo than control promoters. Furthermore, we show that the requirement of an amino acid promoter for ppGpp in vivo can be alleviated by increasing its rate-limiting RNAP-binding step. Our data are consistent with a previously proposed passive model in which ppGpp inhibits stable RNA synthesis directly by reducing the lifetime of the rrn P1 open complex, liberating enough RNAP to stimulate transcription from amino acid promoters. Our data also place considerable constraints on models invoking hypothetical factors that might increase amino acid promoter activity in a ppGpp-dependent fashion.

The study of guanosine 5'-diphosphate 3'-diphosphate-mediated transcription regulation in vitro using a coupled transcription-translation system

The effects of the "alarmone" guanosine 5'-diphosphate 3'-diphosphate (ppGpp) on regulation of the Salmonella typhimurium histindine operon and the Escherichia coli tRNA(leu) operon were analyzed in vitro using a DNA-dependent transcription-translation system, S-30. The expression of the hisG promoter is positively regulated by ppGpp, whereas that of the leuV promoter (of tRNA(1eu)) is negatively regulated by ppGpp. In an attempt to understand the global regulatory mechanism of ppGpp control, interrelationship between ppGpp-dependent activation and repression of gene expression was examined using these promoters as models. It has been traditionally supposed that the ppGpp-dependent regulation, at least for the activation, is by a passive mode of control: the activation of gene expression by ppGpp is a consequence of the repression of stable RNA gene expression in the condition of RNA polymerase limiting. To test this model, the ppGpp-dependent regulations of both an activable promoter (hisGp) and a repressible promoter (leuVp) were determined in vitro simultaneously using a mixed template setup. The rationale for this exercise was to see whether the ppGpp-dependent activation and repression are inversely correlated in the in vitro condition in which RNA polymerase is limiting. No correlation was observed. It was concluded that the ppGpp-dependent activation is independent of the repression. Moreover, it was proposed that ppGpp-dependent activation and repression are mediated by titratable factors, each of which operate independently.

Genetic regulation of glycogen biosynthesis in Escherichia coli: in vitro effects of cyclic AMP and guanosine 5'-diphosphate 3'-diphosphate and analysis of in vivo transcripts

Glycogen accumulation in Escherichia coli is inversely related to the growth rate and occurs most actively when cells enter the stationary phase. The levels of the three biosynthetic enzymes undergo corresponding changes under these conditions, suggesting that genetic control of enzyme biosynthesis may account for at least part of the regulation (J. Preiss, Annu. Rev. Microbiol. 38:419-458, 1984). We have begun to explore the molecular basis of this control by identifying factors which affect the expression of the glycogen genes and by determining the 5'-flanking regions required to mediate the regulatory effects. The in vitro coupled transcription-translation of two of the biosynthetic genes, glgC (ADPglucose pyrophosphorylase) and glgA (glycogen synthase), was enhanced up to 26- and 10-fold, respectively, by cyclic AMP (cAMP) and cAMP receptor protein (CRP). Guanosine 5'-diphosphate 3'-diphosphate stimulated the expression of these genes 3.6- and 1.8-fold, respectively. The expression of glgB (glycogen branching enzyme) was affected weakly or negligibly by the above-mentioned compounds. Assays which measured the in vitro formation of the first dipeptide of glgC showed that a restriction fragment which contained 0.5 kilobases of DNA upstream from the initiation codon supported cAMP-CRP-activated expression. Sequence-specific binding of cAMP-CRP to a 243-base-pair restriction fragment from the region upstream from glgC was observed by virtue of the altered electrophoretic mobility of the bound DNA. S1 nuclease protection analysis identified 5' termini of four in vivo transcripts within 0.5 kilobases of the glgC coding region. The relative concentrations of transcripts were higher in the early stationary phase than in the exponential phase. Two mutants which overproduced the biosynthesis enzymes accumulated elevated levels of specific transcripts. The 5' termini of three of the transcripts were mapped to a high resolution. Their upstream sequences showed weak similarity to the E. coli consensus promoter. These results suggest complex transcriptional regulation of the glycogen biosynthesis genes involving multiple promoter sites and direct control of gene expression by at least two global regulatory systems.

Independence of cyclic AMP and relA gene stimulation of glycogen synthesis in intact Escherichia coli cells

Previous studies from our laboratory established that in Escherichia coli, glycogen synthesis is regulated by both the relA gene, which mediates the stringent response, and by cyclic AMP. However, those studies raised the question of whether this dual regulatory system functions in an independent or a dependent manner. We show here that this regulation is independent, i.e., each regulatory process can express its action in the absence of the other. Triggering the stringent response by amino acid starvation increased glycogen synthesis even in mutants lacking the ability to synthesize cyclic AMP or lacking cyclic AMP receptor protein; and cyclic AMP addition stimulated glycogen synthesis in relA mutant strains. We also show that physiological concentrations of GTP inhibit ADP-glucose synthetase (glucose-1-phosphate adenylyltransferase, EC 2.7.7.27), the rate-limiting enzyme of bacterial glycogen synthesis, in vitro. Because the stringent response is known to cause an abrupt decrease in the cellular level of GTP, modulation of ADP-glucose synthetase activity by this nucleotide could account for a substantial portion of the step-up in the cellular rate of glycogen synthesis observed when the stringent response is triggered.

Characterization of RNA synthesis in an Escherichia coli mutant with a temperature-sensitive lesion in stable RNA synthesis

Previous experiments with Escherichia coli strain 2S142 have shown that the synthesis of stable RNA is preferentially blocked at the restrictive temperature. In this paper, we have examined the capacity of this mutant strain to synthesize RNA in vitro. Growth of the strain for as short a period as 10 min at 42 degrees C resulted in a 40 to 60% loss of RNA synthetic capacity and a fourfold decrease in percent rRNA synthesized in toluenized cell preparations. The time course for the loss and recovery of this RNA synthetic capacity correlated very well with the changes in RNA synthesis observed in vivo. We found no difference in temperature sensitivity of the purified RNA polymerase from the mutant and the parental strains. Moreover, there was no detectable alteration in the amount of enzyme, specific activity of the enzyme, or electrophoretic mobility of the subunits when the mutant strain was grown at 42 degrees C. The capacity for rRNA synthesis was also measured with the Zubay in vitro system (Reiness et al., Proc. Natl. Acad. Sci. 72:2881-2885, 1975). Supernatant fractions (S-30) prepared from cells grown at 30 degrees C were capable of up to 31.2% rRNA synthesis, using phi 80d3 DNA as template. S-30 fractions from cells grown at 42 degrees C synthesized 8.6% rRNA. The bottom one-third of the S-100 fraction and the ribosomal salt wash from 30 degrees C cells contained one or more factors which partially restored preferential rRNA synthesis in S-30 fractions from cells grown at 42 degrees C. Preliminary evidence suggests that the factor(s) is protein in nature.

The energy-dependent degradation of guanosine 5'-diphosphate 3'-diphosphate in Escherichia coli. Lack of correlation with ATP levels in vivo and role of the transmembrane proton gradient

Different inhibitors of the energy metabolism have been assayed in Escherichia coli K12 for their ability to increase the level of guanosine 5'-diphosphate 3'-diphosphate (ppGpp) as a consequence of a restriction of its degradation. Inhibitors of the respiration and uncouplers of oxidative phosphorylations had effects similar to carbon-source-induced downshifts while the ATPase inhibitor dicyclohexylcarbodiimide was less efficient. The effects of dicyclohexylcarbodiimide and of the uncoupler carbonylcyanide p-fluoro methoxyphenylhydrazone (FCCP) on ppGpp degradation were compared in a drug-permeable envelope mutant. At concentrations of inhibitors sufficient to deplete the pool of ATP by 50%, only FCCP was able to block ppGpp degradation. Moreover, FCCP also inhibited ppGpp degradation in a ATPase-deficient strain growing on glucose as carbon source while, as expected, it did not change the level of ATP. It is concluded, according to Mitchell's chemiosmotic hypothesis, that, in vivo, the integrity of the transmembrane proton gradient rather than the ATP pool size is a prerequisite for the normal processing of the energy-dependent degradation of ppGpp.

Effect of guanosine 5'-diphosphate 3'-diphosphate and related nucleoside polyphosphates on induction of tryptophanase and beta-galactosidase in permeabilized cells of Escherichia coli

Exogenous addition of guanosine and adenosine 5'-(mono, di and tri) phosphate 3'-diphosphates (pppGpp, ppGpp, pGpp, pppApp, ppApp and pApp) stimulated the synthesis of tryptophanase and beta-galactosidase in permeabilized cells of Escherichia coli. From the results obtained with ppGpp and pppApp, this effect appeared to be at a transcriptional level and depended greatly on the growth condition; the largest effect was observed in cells under shiftdown or grown on poor enrgy source. ppGpp and pppApp, unlike cyclic AMP, did not act to overcome the inhibition of enzyme induction by glucose, but in combination with cyclic AMP caused a synergistic stimulation effect. In the shiftdown cells, ppGpp and pppApp gave 30% or more stimulation effect on tryptophanase induction while cyclic AMP did not stimulate induction. There was therefore a pronounced difference between cyclic AMP and ppGpp or pppApp in stimulatory function.

Effect of guanosine tetraphosphate on in vitro protein synthesis directed by E1 and E3 colicinogenic factors

The in vitro synthesis of colicin E3 was found to be stimulated by guanosine 5'-diphosphate 3'-diphosphate (ppGpp), while that of several other ColE3 plasmid-specific proteins was reduced in the presence of this nucleotide. The ColE1 plasmid-directed in vitro synthesis of colicin E1 was also found to be stimulated by ppGpp.

Physiological regulation of a decontrolled lac operon

The expression of the lac operon was studied under a variety of growth conditions in induced and in constitutive cells of Escherichia coli that carried different catabolite-insensitive lac promoters. Use of such "decontrolled" lac operons permitted a study of the expression of an operon that was presumably subject only to passive control. Since the use of toluenized cells was demonstrated not to be completely reliable, all enzyme assays were performed on sonic supernatant fluids. The cells contained different catabolite-insensitive promoters, which included the L1 and UV5 lac promoters, as well as others isolated in this study. There were three major observations. First, small but real carbon source effects were seen. Second, there was only a small change in beta-galactosidase specific activity with changes in the growth rate. This result implies a limited transcription and/or translation capacity within the cell. Third, at rapid growth rates, most promoters exhibited a decreased expression. The UV5 promoter, which was the "strongest" promoter, was an exception. A mechanism to explain this promoter-dependent control is discussed.

Biosynthesis of mammalian transfer RNA. Evidence for regulation by deacylated transfer RNA

The rate of tRNA synthesis in cultured Friend leukemia cells has been examined as a function of the variation in polyribosome structure produced by treatment with a variety of inhibitors of protein synthesis. The results indicate, in contrast to the conclusions of Bölcsföldi (Bölcsföldi, G. (1974) Exp. Cell Res., 88, 231--240), that no necessary relationship exists between the ribosome distribution and the rate of tRNA synthesis. Alternatively, it is observed that inhibitors of tRNA aminoacylation cause, in all cases, a decrease in the rate of tRNA synthesis whereas drugs which may stimulate the aminoacylation of tRNA cause, in all cases, an elevation of the rate of tRNA synthesis. It is concluded that tRNA synthesis in mammalian cells may be regulated by the relative levels of acylated and deacylated tRNA.

Specificity of the stimulation of in vitro ribonucleic acid synthesis by guanosine 5'-diphosphate 3'-diphosphate

The in vitro synthesis of ribonucleic acid (RNA) by S-30 extracts of Escherichia coli K-12 is stimulated from two-to fourfold by 0.16 mM to 0.32 mM guanosine 5'-diphosphate 3'-diphosphate (ppGpp) when either gammacI857St68h80 deoxyribonucleic acid (gammah80 DNA), gammah80dilv DNA or gammah80dlac DNA are employed as templates. Hybridization analysis of the 3H-RNA product transcribed from gammah80dilv DNA in the presence of ppGpp indicates that both bacteriophage- and bacterial-specific transcription is stimulated to an equivalent degree. In the absence of cyclic 3'-5'-adenosine monophosphate (cyclic AMP), correct lac-specifci RNA synthesis from gammah80dlac DNA is not stimulated by 0.32 mM ppGpp although total RNA synthesis is increased nearly twofold. In the presence of 0.5 mM cyclic AMP, correct lacspecific RNA synthesis is stimulated preferentially by ppGpp. These data suggest that ppGpp is capable of stimulating in vitro transcription in both a general and selective manner.