J. Mol. Microbiol. Biotechnol. (2002) 4(3): 331–340. JMMB Symposium

Regulation of Ribosomal RNA Synthesis in E. coli: Effects of the Global Regulator Guanosine Tetraphosphate (ppGpp)

Rolf Wagner* structure of the RNA polymerase complex obtained at high resolution considering most of the Institut fu¨ rPhysikalische Biologie, Heinrich-Heine- available data on ppGpp-dependent regulation. Universita¨ tDu¨ sseldorf, Universita¨ tsstr. 1, D-40225 Du¨sseldorf Ribosomal RNA Synthesis is Tightly Coupled to Cell Growth Abstract In cell growth and the capacity for aredirectly linked. Changes in the translational The global regulatory (p)ppGpp are capacity are generally reflected by changes in the major effectors for the control of ribosomal RNA number of . This is a consequence of the in bacteria. The effector molecules accumulate to fact that the rate of translation is relatively constant. different cellular levels at deprivation Since ribosomes are extremely large ribonucleopro- or during different growth rates. They change the tein complexes they represent a very costly machin- activity of RNA polymerase to transcribe from ery for the cell. Their number is thus carefully sensitive promoters (e.g. ribosomal RNA promo- adapted to the growth demands of the cell. The ters). Sensitive promoters are characterized by a biosynthesis of ribosomes is determined, however, by GC-rich discriminator element in addition to the rateofrRNAsynthesis while the synthesis of the further structural requirements not completely protein components is a subordinate process. Hence, understood. ppGpp must also be regarded as a rRNA synthesis is intricately coupled to cell growth mediator for growth rate control although it and largely determines the translational capacity of appears that ppGpp-independent regulatory me- bacterial cells. It is not surprising that a process so chanisms exist. Inhibition occurs at various steps central for the cell is tightly and efficiently regulated. during initiation but also during elongation where In fact, the synthesis of rRNAs is controlled by a RNA polymerase pausing is observed. From the complex set of interacting regulatory networks existing data a mechanistic model for the action (Wagner, 1994; Condon et al., 1995; Gourse et al., of ppGpp is suggested considering structural 1996; Wagner, 2000). Here I will focus on the global details of RNA polymerase obtained at high regulation governed by hormone-like effector nucleo- resolution. tides guanosine 30,50-bis(diphosphate) (ppGpp) and guanosine 30-diphosphate 50-triphosphate (pppGpp). Introduction These two regulatory compounds are major players in the adaptation of rRNA synthesis upon changes in The goal of this review is to serve three aspects. First the environmental conditions and during nutritional it should provide the reader with some basic informa- deprivation. tion on the global network of ppGpp-dependent Despite many years of intensive research the regulation and explain how it affects ribosomal RNA molecular mechanisms of (p)ppGpp-dependent tran- (rRNA) transcription in bacteria. This part should scriptional regulation are still unsolved and many highlight results that are undisputed but also present controversial ideas have been put forward to explain findings that are controversial and need further theregulatory effects. In the following section I like to clarification. I will then summarize and review results focus on those (regrettably few) steps of the regulatory collected over a period of several years in my own cascade which are more or less undisputed and for group undertaken to better understand some of the which consensus has been reached by several crucial steps of ppGpp-dependent regulation of rRNA independent scientific approaches. transcription. These findings will be discussed with The response provoked by the effector nucleotides results or models obtained by other groups working on (p)ppGpp is rapid and can either be dramatic, for related subjects. Finally, I like to introduce some new example, under conditions of amino acid starvation ideas on a possible mechanism of ppGpp-dependent when rRNA synthesis is shut-off immediately. On the control. The ideas are based on the three-dimensional other hand, the response can lead to a gradual and balanced adaptation of rRNA synthesis rates to accom- modate different steady state growth rates of the cell. *For correspondence. Email [email protected]; Both reactions are linked to changes in the levels Tel. 49 211 811 4928; Fax. 49 211 811 5167. of (p)ppGpp but can be distinguished physiologically.

# 2002 Horizon Scientific Press Further Reading

Caister Academic Press is a leading academic publisher of advanced texts in microbiology, molecular biology and medical research. Full details of all our publications at caister.com

• MALDI-TOF Mass Spectrometry in Microbiology Edited by: M Kostrzewa, S Schubert (2016) www.caister.com/malditof • Aspergillus and Penicillium in the Post-genomic Era Edited by: RP Vries, IB Gelber, MR Andersen (2016) www.caister.com/aspergillus2

• The Bacteriocins: Current Knowledge and Future Prospects Edited by: RL Dorit, SM Roy, MA Riley (2016) www.caister.com/bacteriocins • Omics in Plant Disease Resistance Edited by: V Bhadauria (2016) www.caister.com/opdr

• Acidophiles: Life in Extremely Acidic Environments Edited by: R Quatrini, DB Johnson (2016) www.caister.com/acidophiles

• Climate Change and Microbial Ecology: Current Research and Future Trends Edited by: J Marxsen (2016) www.caister.com/climate

• Biofilms in Bioremediation: Current Research and Emerging Technologies • Flow Cytometry in Microbiology: Technology and Applications Edited by: G Lear (2016) www.caister.com/biorem Edited by: MG Wilkinson (2015) www.caister.com/flow • Microalgae: Current Research and Applications • Probiotics and Prebiotics: Current Research and Future Trends Edited by: MN Tsaloglou (2016) www.caister.com/microalgae Edited by: K Venema, AP Carmo (2015) www.caister.com/probiotics • Gas Plasma Sterilization in Microbiology: Theory, Applications, Pitfalls and New Perspectives • Epigenetics: Current Research and Emerging Trends

Edited by: H Shintani, A Sakudo (2016) Edited by: BP Chadwick (2015) www.caister.com/gasplasma www.caister.com/epigenetics2015

• Virus Evolution: Current Research and Future Directions • Corynebacterium glutamicum: From Systems Biology to Biotechnological Applications Edited by: SC Weaver, M Denison, M Roossinck, et al. (2016) www.caister.com/virusevol Edited by: A Burkovski (2015) www.caister.com/cory2 • Arboviruses: Molecular Biology, Evolution and Control Edited by: N Vasilakis, DJ Gubler (2016) • Advanced Vaccine Research Methods for the Decade of www.caister.com/arbo Vaccines Edited by: F Bagnoli, R Rappuoli (2015) • Shigella: Molecular and Cellular Biology www.caister.com/vaccines Edited by: WD Picking, WL Picking (2016) www.caister.com/shigella • Antifungals: From Genomics to Resistance and the Development of Novel Agents

• Aquatic Biofilms: Ecology, Water Quality and Wastewater Edited by: AT Coste, P Vandeputte (2015) Treatment www.caister.com/antifungals Edited by: AM Romaní, H Guasch, MD Balaguer (2016) www.caister.com/aquaticbiofilms • Bacteria-Plant Interactions: Advanced Research and Future Trends Edited by: J Murillo, BA Vinatzer, RW Jackson, et al. (2015) • Alphaviruses: Current Biology www.caister.com/bacteria-plant Edited by: S Mahalingam, L Herrero, B Herring (2016) www.caister.com/alpha • Aeromonas Edited by: J Graf (2015) • Thermophilic Microorganisms www.caister.com/aeromonas Edited by: F Li (2015) www.caister.com/thermophile • Antibiotics: Current Innovations and Future Trends Edited by: S Sánchez, AL Demain (2015) www.caister.com/antibiotics

• Leishmania: Current Biology and Control Edited by: S Adak, R Datta (2015) www.caister.com/leish2

• Acanthamoeba: Biology and Pathogenesis (2nd edition) Author: NA Khan (2015) www.caister.com/acanthamoeba2

• Microarrays: Current Technology, Innovations and Applications Edited by: Z He (2014) www.caister.com/microarrays2

• Metagenomics of the Microbial Nitrogen Cycle: Theory, Methods and Applications Edited by: D Marco (2014) www.caister.com/n2

Order from caister.com/order 332 Wagner

One distinction can be made by the concentration of RNA transcription is clearly the most dramatic effect. the mediator that is reached in the cell. The fast and However, a number of additional cellular changes can dramatic response that is triggered by high (millimolar) be observed as direct or indirect consequence of the concentrations of the effector is generally termed increased (p)ppGpp level. For instance, concerted stringent regulation. In contrast, the steady state adap- inhibition is observed for the synthesis of the compo- tation of rRNA synthesis to an altered growth rate, for nents, which constitute the translation apparatus which the concentrations of the effector nucleotides (including ribosomal RNAs, ribosomal proteins trans- vary in the micromolar range, is termed growth rate lation initiation and elongation factors as well as control (Cashel et al., 1996). It is one characteristic of tRNAs). In addition, the synthesis of RNA polymerase growth rate regulation that changes in the rate of rRNA subunits is repressed while the expression of stress- synthesis are roughly proportional to the square of specific sigma factors, like sS and sH,isincreased. the changes in growth rates. It should be noted at this The transcription of amino acid biosynthetic is point, however, that growth rate control of rRNA pro- activated. Moreover, proteolysis is activated and trans- moters has also been described in strains devoid of port, with the exception of the uptake for branched (p)ppGpp (see below). Stringent control, on the other amino acids, is inhibited. DNA synthesis, cell wall syn- hand, summarises a complex set of metabolic reactions thesis and phospholipid metabolism are all inhibited. triggered by amino acid deprivation. Again, the most Peptidoglycan synthesis and penicillin tolerance are outstanding and direct effect during the stringent con- also changed. Furthermore, the fidelity of translation trol is the immediate repression of stable RNA (rRNAs is reduced and an enhanced mutation frequency has and tRNAs) transcription. been observed. Many of the reactions must be regar- In summary, the hormone-like effector molecules ded as indirect consequences. There appears to be a (p)ppGpp are able to trigger two types of transcrip- common logic behind this set of pleiotropic changes. tional regulation, stringent and growth rate control. They all seem to serve one crucial purpose, namely to Both types of regulation are linked to changes in the overcome the amino acid limitation and to adjust the cellular concentration of the effector nucleotides, protein synthesizing capacity to an altered cellular which are either in the micro- or the millimolar range. substrate pool (for further reviews see Galant, 1979; We will see below that two different enzymatic Cashel et al., 1996; Wagner, 2000). pathways are responsible for the accumulation or maintenance of the different effector What is the Target for (p)ppGpp Action? levels. Within the long list of cellular responses to increasing The Chain of Events Leading to (p)ppGpp levels transcription is clearly the most pro- Stringent Control minent step affected directly. Hence, the most likely candidate where (p)ppGpp exerts its effect is RNA The represents the consequence of polymerase. This assumption is supported already amino acid deprivation in the cell. A lack in the from the early analysis of mutants which show reduced availability of one or several amino acids causes that sensitivity to enhanced ppGpp levels (Little et al., the corresponding tRNAs cannot be sufficiently 1983; Tedin and Bremer, 1992). Several such muta- charged at their 30 acceptor ends. Hence, the ratio of tions have been mapped in the RNA polymerase b amino acylated versus non-amino acylated tRNAs subunit (some have also been characterized within the changes. The appearance of deacylated tRNAs, not s70 subunit (Hernandez and Cashel, 1995)). This had free amino acids, represents the signal for the cell to initiated many attempts to show ppGpp-RNA polymer- change a large number of metabolic activities. The ase interactions directly. Cross-linking experiments, actual sensors for the stringent control are translating CD-measurements and fluorescence labelling with ribosomes (Gallant, 1979; Krohn and Wagner, 1995). ppGpp analogues indicate that the b subunit is the When a non-amino acylated tRNA is bound in a codon- target for interaction. The experiments also show that dependent way to the ribosomal A-site the - alikelysiteofinteraction is in the vicinity of the active associated RelA protein (ppGpp synthetase I or PSI) center for nucleotide addition (Reddy et al., 1995). catalyses the synthesis of pppGpp by transferring Together the results are fairly convincing. They do not pyrophosphate from ATP to the 30 OH group of GTP. provide absolute proof, however. For instance, the The is then rapidly con- direct binding of genuine ppGpp has never been verted to ppGpp by a pppGpp 50 phosphohydrolase achieved and a certain degree of unspecific binding (GPPase) (see Figure 1). It is not completely clear of the analogues cannot be completely eliminated. whether the penta- or the tetraphosphate exhibit Attempts to cross-link radioactive ppGpp directly by different regulatory features. Based on in vitro studies, UV irradiation were not conclusive, for example, but however, both effector molecules appear to be indis- rather indicated that contaminating GTP was cross- tinguishable with respect to their regulatory activity linked totheNTPbinding site (Heinemann and (Heinemann and Wagner, 1997). Wagner, unpublished). Moreover, in a recent cross- With the rapid increase of the cellular (p)ppGpp linking study the b0 not the b subunit was identified as level from micromolar to roughly millimolar concentra- themajor target for ppGpp binding. It was proposed tions pleiotropic changes in many metabolic activities therefore that part of the b together with part of can be observed. Immediate repression of stable the b0 subunit constitute a modular ppGpp binding site ppGpp-Dependent Transcription Regulation 333

Figure 1. ppGpp metabolism and control of rRNA synthesis. Details see text. 334 Wagner

(Toulokhonov et al., 2000). There is also some mutations within the region support the evidence that the stability of s-core interaction may conclusion that the –10 recognition element and the be affected by (p)ppGpp (Hernandez and Cashel, downstream AT-rich discriminator sequence are both 1995; Michels and Wagner, unpublished). No indica- important for ppGpp-dependent regulation (Riggs tion for binding of ppGpp to the s subunit has been et al., 1986; Shand et al., 1989). On the other hand, presented so far. positive stringent control has often been explained as In conclusion, there is little doubt that RNA aconsequence of a passive increase in RNA poly- polymerase is the target for ppGpp interaction and merase concentration as consequence of the repres- binding very likely occurs close to the RNA polymerase sion of stable RNA promoters. In the his case, active center constituted by the b and partly by the b0 however, studies performed with a coupled transcrip- subunit. Unequivocal proof where exactly the effector tion translation system did not support passive regula- molecules bind is still pending, however. tion. Under mixed-template conditions activation was shown to be independent from simultaneous repres- The Promoter Structure Discriminates Stringently sion (Choy, 2000). Passive regulation as a result of or Non-Stringently Transcribed Genes changing RNA polymerase activity may still be impor- tant. It requires, however, that RNA polymerase must Whether a gene is under stringent control or not is be limiting, a condition that has not been rigorously determined by the promoter structure. A conserved proven yet. In a very recent study no stimulation of GC-rich consensus sequence, the GCGC discrimin- amino acidpromoterscouldbeshownin vitro.Instead, ator motif, localized immediately downstream to the amino acid promoters appear to require higher con- 10 promoter element has been recognized as centrations of RNA polymerase for function in vitro importantÀ for negative stringent control (Travers, and in vivo than control promoters. The authors pro- 1984). The involvement of the discriminator element pose that liberated RNA polymerase from stable RNA in the stringent control as well as during growth rate promoters stimulate transcription from the amino acid regulation has been demonstrated in a number of promoters. Hence their data supports the passive cases (Mizushima-Sugano and Kaziro, 1985; Zachar- model (Barker et al., 2001a, b). In conclusion, there ias et al., 1989; Davies and Drabble, 1996). We had is good evidence that also for positive stringent control shown, for example, that an unregulated promoter thepromoter structure is crucial for the regulation could be converted to stringency by changing the mechanism. Passive effects may contribute signifi- non-consensus discriminator into a GCGC consensus cantly to positive stringent control, however. sequence. Interestingly, the same base change also resulted in growth rate sensitivity of the altered pro- Which Steps During the Initiation Cycle moter. However, introduction of the GCGC sequence are Affected by ppGpp? did not convert the synthetic and unregulated tac promoter to stringency (nor to growth rate depen- Transcription initiation is a multistep process that can dence). In conclusion, the GCGC discriminator must formally be divided in four major steps (see Figure 1). be viewed as an element, necessary but not sufficient In the first step RNA polymerase binds to the promoter for negative stringent control (Zacharias et al., 1989). DNA forming a binary closed complex (RPc). This Fusion of different sequence elements upstream and complex isomerizes to an open complex (RPo) with downstream of the discriminator site obtained from about 12 nucleotides of the core promoter sequence in stringently and non-stringently regulated promoters single stranded conformation. When the initial sub- revealed that not a defined sequence element but the strate NTPs are bound to this complex a ternary complete promoter structure in unknown sterical con- initiating complex is formed(RPinit). This ternary sequence defines whether the resulting promoter is complex can either be converted into an elongating stringently regulated or not (Zacharias et al., 1990). complex (EC) by successive addition of 9 to 12 NTPs. Whether the GCGC promoter determinants are At this point the s subunit leaves the complex. The generally identical for stringent control and growth ternary initiation complex can alternatively undergo rate-dependent regulation cannot be answered unam- repeated rounds of abortive cycling releasing abortive biguously today. Although the involvement of the transcription products between 3 and 12 nucleotides. GCGC discriminator sequence in growth rate-depen- Depending on the respective promoter, each step dence has been demonstrated in quite a number of during this cycle can be rate limiting and in principle cases (see above) a common function of this promoter each step may be affected during stringent control. element in both types of control is controversial. From Promoter melting (open complex formation) is studies with different promoter mutants it was con- considered to depend on the stability and thus on the cluded that stringent regulation and growth rate control GC-content of the promoter and flanking downstream have non-identical promoter sequence requirements discriminator. The presence of GC- or AT-rich dis- (Josaitis et al., 1995). criminators in positive and negative regulated promo- Interestingly, for genes under positive stringent ters, respectively have led to the conclusion that open control, like the his operon, the sequence correspond- complex formation is affected by ppGpp. In fact, for ing to the stringent discriminator is AT-rich. In vitro and several promoters under stringent control open com- in vivo analyses revealed that ppGpp is responsible for plex formation has been shown to be the step which is the activating effect at the his promoter. Moreover, primarily affected by ppGpp (Ohlson and Gralla, 1992; ppGpp-Dependent Transcription Regulation 335

Raghavan et al., 1998). We have shown for the rrnB P1 is consistent with the finding that certain RNA promoter that not the amount or the stability of open polymerase mutations in rpoB which destabilize initia- complexes formed but the rate of formation is affected tion complexes at stringently controlled promoters in by high ppGpp concentrations (Jo¨res, unpublished the absence of ppGpp behave like normal RNA results). For other promoters steps of the initiation polymerases in the presence of ppGpp (Zhou and cycle later than open complex formation can be rate Jin, 1998). Open complexes at inversely growth rate- limiting, however. Hence, open complex formation dependent regulated amino acid promoters are formed seems to be at least one but probably not the only more slowly and have longer life times than rrnB P1 step which is influenced by ppGpp (see below). open complexes. While ppGpp decreased the half- Anumberofadditional observations at rRNA lives of open complexes for all promoters a direct promoters are consistent with the conclusion that open effect of the mediator was only measurable for the complex formation is one of the crucial steps affected negatively regulated rrnB P1 promoter which has a by ppGpp during transcription initiation. For instance, very short half-life. No direct effects of ppGpp on the ribosomal RNA promoters have a characteristically rates of association or escape from amino acid suboptimal 16 bp spacing between the –35 and –10 promoters was detected in vitro. recognition elements (underwound). The promoters While the above study offers a mechanistic are therefore twist-sensitive and open complex explanation for many observations with respect to formation or isomerization from the closed to the stringent control there are many questions still unan- strand-opened structure is facilitated by negative swered. Comparison of different results obtained supercoiling. Interestingly, the his promoter under in vitro have shown that the system depends critically positive stringent control has a suboptimal larger on the particular experimental conditions, indicating spacer of 18 bp (overwound). It has been proposed that other steps during the initiation cycle may also that for many promoters the supercoil sensitivity contribute if the efficiency of the in vitro transcription parallels the responsiveness to stringent control system is different. Therefore, to analyse ppGpp- (Figueroa-Bossi et al., 1998). Interestingly, super- dependent inhibition in a reliable manner it is essential coiled templates have been found to resist ppGpp- to adjust the in vitro transcription conditions in a dependent inhibition in vitro consistent with the view defined way. This has been exemplified by varying the that open complex formation is now more efficient at concentrations of the NTPs necessary to form the high superhelical density and no longer rate limiting, initial open complex. If the concentration of these regardless if ppGpp is present or not. substrates is too high the regulatory effect of ppGpp There are examples indicating that steps different can easily be masked by the efficiency of the system. from open complex formation may be essential for The concentration of the initial NTP substrates is thus ppGpp-dependent regulation. This was concluded essential for the efficient in vitro determination of the from a detailed kinetic analysis with the rrnB P2 ppGpp-dependent repression. This observation differs promoter and a P2 promoter variant with high ppGpp from the proposal that growth rate-dependent regula- sensitivity. The study supported the view that a tion in the absence of ppGpp is mediated by limiting combination of effects at initiation, promoter clearance substrate NTP concentrations (see below). It only and elongation are mediated by ppGpp. During initia- means that ppGpp-dependent inhibition cannot be tion an alternative pathway is triggered by ppGpp measured at conditions where in vitro transcription is involving stabilized initial closed complexes and tooefficient. The finding is in accordance with the impeded open complexes. According to this study effect described above that supercoiled templates do ppGpp-modified RNA polymerases are trapped in the not support significant inhibition when analyzed in vitro closed complex which is formed at higher rates in (see above). We propose that inhibition by ppGpp is presence of ppGpp. The open complex, on the other only apparent in vitro when the step(s) affected are hand, is destabilised and the reaction back to the limiting and not too effective. Suboptimal transcription closed complex is facilitated. Hence, discrimination may be the consequence of a structural alteration of between stringently and non-stringently controlled the RNA polymerase in the transcription complex or promoters occurs at the early steps of initiation. This result from other destabilising or limiting factors which finding has led to the proposal of a trapping mechan- may facilitate a backward reaction instead of yielding ism (Heinemann and Wagner, 1997). Efficient inhibi- productive transcription complexes, for instance. tion is caused at later steps, very likely during promoter clearance where the apparent KM values for The Role of ppGpp as Growth Rate Regulator substrate NTP binding are increased by ppGpp (see below). Already from early studies of bacterial physiology it Recent data presented from the Gourse laboratory was concluded that (p)ppGpp represents the link (Barker et al., 2001a, b), where both positive and between the physiological state of the cell and the negative transcription regulation of the rRNA P1 activity of promoters that respond to growth rate promoter and a series of inversely growth rate- changes. The evidence is indirect, however, and dependent amino acid biosynthesis/transport promo- largely based on the observation that a linear inverse ters had been compared, revealed that the major correlation has been noted between the growth rate difference between the two systems is related to and the cellular ppGpp concentration. This observation the stability of the open complexes formed. This holds true under almost all conditions (Baracchini and 336 Wagner

Bremer, 1988; Zacharias et al., 1989). The low ppGpp responsible for this type of control. This has led to a concentration range between fast and slow growth model according to which the concentration of NTPs is rates (micromolar) is maintained by a cellular activity sensed by the instability of the initiation complexes different from RelA. The spoT gene product was found before productive transcription occurs (Gaal et al., to be responsible for synthesis and maintenance of the 1997). The NTP sensing model rests on the premise RelA-independent (p)ppGpp synthesis (Hernandez that the NTP concentrations are limiting during slow and Bremer, 1991). Hence, SpoT represents a second growth and vary notably when the growth rate ppGpp synthetase (PSII) (Figure 1). The involvement changes. This could not be verified in different studies, of (p)ppGpp in growth rate regulation is further however (Petersen and Møller, 2000). supported by experiments with isogenic strains which have different mutations in the spoT gene. These Transcription Initiation and Elongation mutations cause different basal cellular ppGpp con- are Affected by ppGpp centrations. Interestingly, the different SpoT mutations cause different growth rates of otherwise isogenic It is clear today that the effect of ppGpp on strains when cultured in the same medium. The growth transcription is not restricted to the initiation cycle. rates of the different mutants correlate perfectly with Measurements in vitro and in vivo have revealed that the cellular ppGpp concentration present in each transcription elongation rates are reduced at en- mutant (Sarubbi et al., 1988). Moreover, the results hanced ppGpp concentrations. This is consistent with outlined above that promoter mutations changing the theknown coupling between transcription and trans- sensitivity for stringent regulation affect growth rate lation, for instance when the substrates for one of the control in a very similar manner has strengthened the processes become scarce or if premature termination view that (p)ppGpp is the common mediator for both occurs (e.g. during attenuation or transcription polar- types of control. It is quite obvious, therefore, that ity). The reduction in transcription elongation rates at ppGpp must play a role in growth rate regulation. On elevated ppGpp concentrations is not simply the the other hand, several reports have shown that consequence of a general increase in the step time growth rate regulation of rRNA transcription occurs in forthe nucleotide addition reaction (Kingston et al., strains deficient in both the relA and spoT genes 0 1981). Reduced elongation rates are rather brought (ppGpp phenotype) (Gaal and Gourse, 1990). This about by discrete RNA polymerase pausing events at finding means that there must be at least a second defined template positions (Figure 1). Although the (p)ppGpp-independent mechanism for growth rate exact mechanism(s) leading to enhanced pausing still regulation. A feedback control mechanism had been need some clarifications several conclusions have proposed to explain rRNA shut down. According to the been reached with the aid of special techniques to original suggestion repression should be independent quantify RNA polymerase pausing during in vitro of ppGpp and mediated by ribosomes. However, the transcription. Specifically stalled ternary complexes activity of ribosomes or any other components of the which were elongated under defined conditions have translation machinery as repressors could not be helped to answer a number of questions (Theißen verified. In controversial studies evidence for an et al., 1990; Krohn et al., 1992). According to such increase in ppGpp under the feedback conditions studies RNA polymerase pausing depends on the was presented reinforcing the view that ppGpp is also nature of the promoter and on the transcription of the responsible for feedback control. The discussion is not closely adjacent sequence downstream of the tran- yet settled and attempts to identify specific feedback scription start site. For instance, strong enhance- repressors still continue. In principle, growth rate ments of transcriptional pauses predominate for RNA regulation, in addition to the ppGpp pathway, could polymerases initiated at rRNA promoters, which have be maintained by cellular transcription factors like FIS transcribed through the early sequence region. This and H-NS which are known as activator and repressor is true even if the downstream region has been for rRNA transcription, respectively (Afflerbach et al., replaced by sequences unrelated to the rRNA 1998; Schro¨ der and Wagner, 2000). It is interesting in transcription unit. On the other hand, pauses are this regard that transcription of one of the factors (FIS) only weakly or moderately affected if transcription is shows a strictly growth phase-dependent expression initiated from a promoter which is not ppGpp and is itself under ppGpp control (Ninnemann et al., dependent. Moreover, not all RNA polymerase paus- 1992). ing sites are affected in the same way. Pausing can Based on the finding that growth rate-dependent be enhanced, unchanged or in some rare cases, regulation can be seen for rRNA promoters in the pausing may even be reduced in presence of ppGpp absence of ppGpp and the fact that rRNA promoters (Krohn and Wagner, 1996). The same study also form notoriously unstable transcription initiation com- showed that in order to enhance transcriptional plexes an alternative model for growth rate-dependent pauses the presence of ppGpp was only required rRNA synthesis was proposed. Since open complex 70 during elongation when s has already left the formation at rRNA promoters in vitro requires the transcription complex. presence of the starting substrate NTPs it was The conclusion that transcription elongation is assumed that the cellular concentration of the initiating affected by ppGpp is supported by a different set of nucleotides for rRNA transcription (ATP and GTP) are studies performed in vivo (Vogel and Jensen, 1994a, b). ppGpp-Dependent Transcription Regulation 337

In contrast to the above findings these studies indicate (Hillebrand and Wagner; Pohl and Wagner, unpub- that the presence of a nut-like boxA sequence lished results). element downstream from the transcription start site (as it is naturally found in the leader of all rRNA ANewMechanistic Proposal for ppGpp-Dependent operons) is crucial for a high elongation rate and Transcriptional Control reduced pausing (Vogel and Jensen, 1995). Only genes devoid of such a boxA signal (like normal According to the recently published three-dimensional mRNAs) show significantly reduced elongation rates structure bacterial core RNA polymerase can in presence of enhanced ppGpp levels. Since in the be described as a crab claw-shaped molecule with a studies described above transcription was always 27 A˚ wide internal channel formed by the arms of the started from the same inducible hybrid promoter the claw (Zhang et al., 1999). This channel harbours the results may not be directly comparable with our own upstream and downstream duplex DNA, oriented at a in vitro data. bend angle of about 90 which results from a kink In conclusion, altered transcription elongation formed between the downstream duplex and the rates as a consequence of ppGpp-dependent RNA RNA/DNA hybrid. One arm of the claw is primarily polymerase pausing must be considered in explaining the b subunit, the other primarily b0.Atthe inner end transcriptional control under conditions of stringent or of the claw the two subunits make extensive interac- growth rate regulation. tions with each other. Within this base of the channel lies the activecenter consisting of a highly conserved How do the Different rRNA Operons Respond to structure and a chelated Mg2+.Awall-like structure the Stringent Control and What do We Know from composed of conserved b0 domains F and G bifur- In Vivo Studies? cates the main channel into two separate cavities. The smaller secondary channel (10 to 12 A˚ in dia- The seven different E. coli rRNA operons have very meter, 45 A˚ in length) is too small to accommodate similar promoter sequences. The individual upstream duplex nucleic acid structures although it might be sequences (UAS regions) important for transcription broad enough to take up backtracked RNA (Korzheva regulation deviate significantly from each other, et al., 2000). It is believed that the secondary channel however. Because the different UAS regions com- functions in delivering nucleotide substrates to the prise binding sites for the growth rate-dependent active center (Zhang et al., 1999). It appears that the transcription factors FIS and H-NS a differential crowded nucleic acid elements in the major channel transcription of some of the rRNA operons is feasible. prevent direct access of substrate NTPs to the As a consequence, individual rRNA operons not only catalytic site. In contrast, the secondary channel is deviate in their basal transcription efficiencies but large enough and it provides a clear path for incoming also in their response to regulation (Hillebrand and NTPs. From its dimensions there is room enough for Wagner, unpublished results). In previous studies only one NTP at a time to diffuse through the NTP with the rrnB P1 promoter the sequences upstream of channel, however. Based on this structural considera- the core promoters were not considered to be tions I like to propose a model explaining the involved in stringent regulation. The differences in inhibitory effects of ppGpp during different stages of the individual UAS regions have led us to reinvesti- transcription by preventing free access of substrate gate the differential regulation of individual rRNA NTPs to the secondary channel. The model outlined operons in more detail. This question is currently below is illustrated in Figure 2. Due to its similarity being addressed by systematic in vitro and in vivo with the NTP substrates (p)ppGpp should be able to analyses employing isolated rRNA transcription units. enter the secondary channel. With the 30 diphos- The study is an extension of a previous in vivo phates it is unable, however, to gain access into the investigation of rRNA promoter activities. In this study catalytic center where the nucleotide addition reaction the effects of altered gene dosage (by plasmid copy occurs. The specific arrangement of phosphate number), different growth conditions, and strains with groups, the negative surface charge and the purine a defective fis gene were analyzed. In accordance base may provide the specificity for a transient with other investigators (Gafny et al., 1994; Josaitis interaction withexposed amino acid side chains of et al., 1995) we found that not only rRNA P1 pro- the b0 and b subunits. Bound ppGpp could thus moters but also P2 promoters were downregulated by temporarily block the channel and reduce the rate of ppGpp (although to a smaller extent). In the absence NTP access considerably. This will not stop but slow of FIS or at high ppGpp concentration almost all rRNA down the progress of NTP incorporation. The effect is transcribed from the P2 promoters. Furthermore, becomes dramatic at template positions characterized clear evidence for differential regulation between the by a high apparent KM value for NTP addition. This, rrnB and rrnD operons was apparent underlining that for instance, is the case for the first nucleotide position individual rRNA operons are controlled differentially at all stable RNA promoters. The conformation of the (Liebig and Wagner, 1995). With our present study we secondary channel, and thus its affinity for ppGpp, are able to exactly compare expression of all indivi- may be triggered by the promoter structure and dual rRNA operons. From this analysis a much more therefore be different for stringently or non-stringently detailed picture of the operon-specific regulatory controlled genes. The activity of phage polymerases, events under different growth conditions emerges which do not have a NTP entry channel are not 338 Wagner

Figure 2. ppGpp-dependent inhibition of RNA polymerase by blockage of the NTP substrate channel. RNA polymerase, the secondary channel as entrance for NTPs are presented as cartoon and indicated. The path of the upstream and downstream DNA duplex, the position of the non-template strand, the active center as well as the DNA-RNA hybrid are schematically illustrated. Substrate NTPs and ppGpp molecules compete for passage through the secondary channel.

regulated by ppGpp. The model also provides an polymerase is considerably slower as for phage poly- explanation for the effect of ppGpp during growth rate merases (5 to 10 times) or even DNA polymerases regulation where the concentration of ppGpp deter- (10 to 20 times). These all exceed NTP mines the rate of access of NTPs to the substrate incorporation rates at comparable NTP concentra- channel. Moreover, the model explains ppGpp-de- tions. The published proximity of the rifampicin resis- pendent inhibition during elongation where extensive tance site to the ppGpp binding site and the results RNA polymerase pausing is observed at defined from cross-linkingstudies are all consistent with the template positions. Note that pausing sites are almost model. The finding that stable RNA promoters require always characterized by a high apparent KM for the high concentrations of starting NTPs and the fact complementary nucleotide at which RNA polymerase that the presence of ppGpp increases the apparent is stalled. The model thus explains ppGpp-dependent KM for sensitive genes fit nicely into the model. The inhibition during initiation and elongation and equally chemical and structural similarity of ppGpp with sub- for genes under stringent control or during growth strate NTPs represents an important requirement to rate regulation. It even explains the weak general compete for access into the space limiting substrate reduction of the transcription efficiency observed for channel. The 30 pyrophosphate of ppGpp makes it genes not under stringent control. There are addi- distinct enough not to block the catalytic center tional arguments which support the above model. The irreversibly. Finally, the concentration range for property of the secondary channel as entry site for ppGpp during stringent and growth rate regulation substrate NTPs limiting free access to the catalytic and substrate NTPs are in line with the range of site is supported by the well known fact that the inhibition observed. Experiments are open to test the rate of transcription catalyzed by bacterial RNA validity of the model! ppGpp-Dependent Transcription Regulation 339

Acknowledgements Korzheva, N., Mustaev, A., Kozlov, M., Malhotra, A., Nikiforov, V., Goldfarb, A., and Darst, S.A. 2000. A structural model of transcription I like to thank all the past and present members of my lab for their elongation. Science 289: 619–625. countless contributions which have enabled me to write this mini Krohn, M., Pardon, B., and Wagner, R. 1992. Effects of the template review. I apologise though, that many contributions were not men- topology on RNA polymerase pausing during in vitro transcription of tioned explicitly and not all the important literature could be cited – the the E. coli rrnB leader region. Mol. Microbiol. 6: 581–589. reference list is already much too long. I am grateful to the Deutsche Krohn, M., and Wagner, R. 1995. A procedure for the rapid preparation Forschungsgemeinschaft, which has funded the work over the past of guanosine tetraphosphate (ppGpp) from ribo- years, and I greatly acknowledge the Fonds der Chemischen Industrie somes. Anal. Biochem. 225: 188–190. forfinancial support. Krohn, M., and Wagner, R. 1996. Transcriptional pausing of RNA polymerase in the presence of guanosine tetraphosphate depends References on the promoter and gene sequence. J. Biol. Chem. 271: 23884–23894. Afflerbach, H., Schro¨ der, O., and Wagner, R. 1998. Effects of the Liebig, B., and Wagner, R. 1995. Effects of different growth conditions Escherichia coli DNA-binding protein H-NS on rRNA synthesis on the in vivo activity of the tandem Escherichia coli ribosomal RNA promoters P1 and P2. Mol. Gen. Genet. 249: 328–335. in vivo.Mol.Microbiol. 28: 641–654. Little, R., Ryals, J., and Bremer, H. 1983. rpoB mutation in Escherichia Baracchini, E., and Bremer, H. 1988. Stringent and growth control of coli alters control of ribosome synthesis by guanosine tetrapho- rRNA synthesis in Escherichia coli are both mediated by ppGpp. sphate. J. Bacteriol. 154: 787–792. J. Biol. Chem. 263: 2597–2602. Mizushima-Sugano, J., and Kaziro, Y. 1985. Regulation of the expres- Barker, M.M., Gaal, T., Josaitis, C.A., and Gourse, R.L. 2001a. sion of the tufB operon: DNA sequences directly involved in the Mechanism of regulation of transcription initiation by ppGpp. I. stringent control. EMBO J. 4: 1053–1058. Effects of ppGpp on transcription initiation in vivo and in vitro.J.Mol. Ninnemann, O., Koch, C., and Kahmann, R. 1992. The E. coli fis Biol. 305: 673–688. promoter is subject to stringent control and autoregulation. EMBO J. Barker, M.M., Gaal, T., and Gourse, R.L. 2001b. Mechanism of 11: 1075–1083. regulation of transcription initiation by ppGpp. II. Models for positive Ohlsen, K.L., and Gralla, J.D. 1992. Melting during steady-state control based on properties of RNAP mutants and competition for transcription of the rrnB P1 promoter in vivo and in vitro.J.Bacteriol. RNAP. J. Mol. Biol. 305: 689–702. 174: 6071–6075. Cashel, M., Gentry, D.R., Hernandez, V.J., and Vinella, D. 1996. The Petersen, C., and Møller, L.B. 2000. Invariance of the nucleoside stringent response. Escherichia coli and Salmonella Cellular and triphosphate pools of Escherichia coli with growth rate. J. Biol. Molecular Biology. Washington, D.C., ASM Press. Chem. 275: 3931–3935. Choy, H.E. 2000. The study of guanosine 50-diphosphate 30-diphos- Raghavan, A., Kameshwari, D.B., and Chatterji, D. 1998. The phate-mediated transcription regulation in vitro using a coupled differential effects of guanosine tetraphosphate on open complex transcription-translation system. J. Biol. Chem. 275: 6783–6789. formation at the Escherichia coli ribosomal protein promoters rplJ Condon, C., Squires, C., and Squires, C.L. 1995. Control of rRNA and rpsA P1. Biophys. Chem. 75: 7–19. transcription in Escherichia coli.Microbiol. Reviews 59: 623–645. Reddy, P.S., Raghavan, A., and Chatterji, D. 1995. Evidence for a Davies, I.J., and Drabble, W.T. 1996. Stringent and growth-rate- ppGpp-binding site on Escherichia coli RNA polymerase: proximity dependent control of the gua operon of Escherichia coli K-12. relationship with the rifampicin-binding domain. Mol. Microbiol. 15: Microbiology 142: 2429–2437. 255–265. Figueroa-Bossi, N., Gue´ rin, M., Rahmouni, R., Leng, M., and Bossi, L. Riggs, D.L., Mu¨ ller, R.D., Kwan, H.-S., and Artz, S.W. 1986. Promoter 1998. The supercoilling sensitivity of a bacterial tRNA promoter domain mediates guanosine tetraphosphate activation of the parallels its responsiveness to stringent control. EMBO J. 17: histidine operon. Proc. Natl. Acad. Sci. U.S.A. 83: 9333–9337. 2359–2367. Sarubbi, E., Rudd, K.E., and Cashel, M. 1988. Basal ppGpp level Gaal, T., Bartlett, M.S., Ross, W., Turnbough, C., and Gourse, R.L. adjustment shown by new spoT mutants affect steady state growth 1997. Transcription regulation by initiating NTP concentration: rRNA rates and rrnA ribosomal promoter regulation in Escherichia coli. synthesis in bacteria. Science 278: 2092–2097. Mol. Gen. Genet. 213: 214–222. Gaal, T., and Gourse, R.L. 1990. Guanosine 30-diphosphate Schro¨ der, O., and Wagner, R. 2000. The bacterial DNA-binding protein H-NS represses ribosomal RNA transcription by trapping 50-diphosphate is not required for growth rate-dependent regulation of rRNA transcription in Escherichia coli.Proc.Natl.Acad.Sci.U.S. RNA polymerase in the initiation complex. J. Mol. Biol. 298: 737–748. A. 87: 5533–5537. Shand, R.F., Blum, P.H., Mu¨ ller, R.D., Riggs, D.L., and Artz, S.W. Gafny, R., Cohen, S., Nachaliel, N., and Glaser, G. 1994. Isolated P2 1989. Correlation between histidine operon expression and guano- rRNA promoters of Escherichia coli are strongpromoters that are sine 50-diphosphate-30-diphosphate levels during amino acid down- subject to stringent control. J. Mol. Biol. 243: 152–156. shift in stringent and relaxed strains of Salmonella typhimurium.J. Gallant, J.A. 1979. Stringent control in E. coli.Ann. Rev. Genet. 13: Bacteriol. 171: 737–743. Tedin, K., and Bremer, H. 1992. Toxic effects of high levels of ppGpp 393–415. Gourse, R.L., Gaal, T., Bartlett, M.S., Appleman, J.A., and Ross, W. in Escherichia coli are relieved by rpoB mutations. J. Biol. Chem. 267: 2337–2344. 1996. rRNA transcription and growth rate-dependent regulation of Theißen, G., Pardon, B., and Wagner, R. 1990. A quantitative ribosome synthesis in Escherichia coli.Ann. Rev. Microbiol. 50: assessment for transcriptional pausing of DNA-dependent RNA 645–677. polymerase in vitro.Anal. Biochem. 189: 254–261. Heinemann, M., and Wagner, R. 1997. Guanosine 3 ,5 -bis(diphos- 0 0 Toulokhonov, I.I., Shulgina, I., and Hernandez, V.J. 2000. Binding of phate) (ppGpp)-dependent inhibition of transcription from stringently the transcription effector ppGpp to E. coli RNA polymerase is controlled Escherichia coli promoters can be explained by an altered allosteric, modular, and occurs near the N-terminus of the b0-subunit. initiation pathway that traps RNA polymerase. Eur. J. Biochem. 247: J. Biol. Chem. 276: 1220–1225. 990–999. Travers, A.A. 1984. Conserved features of coordinately regulated Hernandez, V.J., and Bremer, H. 1991. Escherichia coli ppGpp E. coli promoters. Nucleic Acids Res. 12: 2605–2618. synthetase II activity requires spoT.J.Biol. Chem. 266: 5991–5999. Vogel, U., and Jensen, K.F. 1994a. Effects of guanosine 30,50-bisdi- Hernandez, V.J., and Cashel, M. 1995. Changes in conserved region 3 phosphate (ppGpp) on rate of transcription elongation in isoleucine- 70 of Escherichia coli s mediate ppGpp-dependent functions in vivo. starved Escherichia coli.J.Biol. Chem. 269: 16236–16241. J. Mol. Biol. 252: 536–549. Vogel, U., and Jensen, K.F. 1994b. The RNA chain elongation rate Josaitis, C.A., Gaal, T., and Gourse, R.L. 1995. Stringent control in Escherichia coli depends on the growth rate. J. Bacteriol. 176: and growth-rate-dependent control have nonidentical promoter 2807–2813. sequence requirements. Proc. Natl. Acad. Sci. U.S.A. 92: 1117–1121. Vogel, U., and Jensen, K.F. 1995. Effects of the antiterminator boxA on Kingston, R.E., Nierman, W.C., andChamberlin, M.J. 1981. A direct transcription elongation kinetics and ppGpp inhibition of transcription effect of guanosine tetraphosphate on pausing of Escherichia coli elongation in Escherichia coli.J.Biol. Chem. 270: 18335–18340. RNA polymerase during RNA chain elongation. J. Biol. Chem. 256: Wagner, R. 1994. The regulation of ribosomal RNA synthesis and 2787–2797. bacterial cell growth. Archives Microbiol. 160: 100–109. 340 Wagner

Wagner, R. 2000. Transcription Regulation in Prokaryotes. Oxford, to a particular sequence motif within the promoter region. Nucleic Oxford University Press. Acids Res. 18: 6271–6275. Zacharias, M., Go¨ ringer, H.U., and Wagner, R. 1989. Influence Zhang, G., Campbell, E.A., Minakhin, L., Richter, C., Severinov, K., of the GCGC discriminator motif introduced into the ribosomal and Darst, S.A. 1999. Crystal structure of Thermus aquaticus core RNA polymerase at 3.3 A˚ resolution. Cell 98: 811–824. RNA P2- and tac promoter on growth rate control and stringent Zhou, Y.N., and Jin, D.J. 1998. The rpoB mutants destabilizing sensitivity. EMBO J. 11: 3357–3363. initiation complexes at stringently controlled promoters behave like Zacharias, M., Go¨ ringer, H.U., and Wagner, R. 1990. The signal for ‘‘stringent’’ RNA polymerases in Escherichia coli.Proc.Natl.Acad. growth rate control and stringent sensitivity in E. coli is not restricted Sci. U.S.A. 95: 2908–2913.