(p)ppGpp: Magic Modulators of Bacterial Physiology and Metabolism

1Bacteria must adapt to fluctuations in their ever-changing surroundings to survive. In order to accomplish optimal resource allocation upon facing environmental shifts, the pathways through which microorganisms rewire their metabolism need to be finely and promptly tuned. An efficient control of metabolism senses stress, adjusts growth accordingly, mandates which genes – and to which extent – are expressed, and ultimately provides a fitness advantage over poorly-adapted microbial competitors. The second messenger molecules (p)ppGpp are compounds that can accomplish all this.

2More than 50 years ago, the “magic spots”, which were identified as guanosine 5'-diphosphate 3'-diphosphate (ppGpp) and guanosine 5'-triphosphate 3'-diphosphate (pppGpp) and collectively referred to as (p)ppGpp or “alarmones”, were discovered by. Since then, the synthesis and degradation of (p)ppGpp as well as the (p)ppGpp-mediated response to nutrient limitations, a phenomenon known as the “stringent response”, have been subject of extensive studies. Central in the metabolism of (p)ppGpp are members of the RelA/SpoT homology (RSH)-type protein family. Briefly, when bacteria encounter nutrient-limiting conditions, RSH proteins utilize ATP as a donor substrate and, through transfer of its β‐ and γ-phosphates onto the 3'-hydroxy group of the acceptor substrate guanosine 5'-diphosphate (GDP) or guanosine 5'-triphosphate (GTP), generate ppGpp or pppGpp, respectively. RSH proteins also degrade (p)ppGpp through removal of the 3'-pyrophosphate moiety of (p)ppGpp, thereby regenerating GDP/GTP. “Long” RSH proteins harbor (p)ppGpp hydrolase and synthetase domains whose reciprocal activities are controlled by further regulatory domains. “Short” RSH proteins consisting only of a (p)ppGpp synthetase or hydrolase domain constitute the class of small alarmone synthetases (SAS) and hydrolases (SAH), respectively. Enzymes of the GppA/PPX family are able to convert pppGpp to ppGpp opening the avenue for more intricate differential regulation by the two alarmone species.

3The activity of RSH proteins is subject to regulation by various mechanisms, and (p)ppGpp-inducing conditions include, e.g., the presence of stalled ribosomes (Rel/RelA;; ), fatty acid starvation, and carbon downshifts (SpoT;, ), enhanced transcription of (p)ppGpp synthetases through cell wall stress stimuli (SAS2/RelP;; ) or allosteric stimulation through the alarmone pppGpp itself (SAS1/RelQ;; ). The intracellular concentrations of (p)ppGpp during growth of Escherichia coli amount approximately 10–40 μM during logarithmic growth – unfortunately those basal levels are still not robustly determined because they typically fall beneath or close to the limit of quantification as in two recent studies – and peak at 800 μM at the onset of stationary phase. Full induction of the stringent response during acute amino acid starvation [e.g., induced by a transfer RNA (tRNA) synthetase inhibitor] gives rise to an increase of the intracellular (p)ppGpp concentrations to approximately 1 mM, ultimately causing growth arrest.

4Substantial progress has been made in the identification and characterization of (p)ppGpp binding targets (reviewed in:;; recent original works:;; ), which fall into different cellular processes such as DNA replication, transcription, translation, ribosome biogenesis, or nucleotide metabolism. Firstly, these studies provide insights as to how (p)ppGpp affect virulence, pathogenicity, persister cell and biofilm formation, heat shock response, and cell growth. Secondly, they supply a wealth of biochemical data that quantitatively describe (p)ppGpp/protein interactions.

5We were wondering whether any prioritization in the order of regulation between those targets/processes would exist. We collected and analyzed binding affinities (exemplified by the dissociation constants, Kd, of the (p)ppGpp/protein complexes), inhibitory constants (Ki) and EC50/IC50 values for the targets of (p)ppGpp and depicted a scheme of prioritization in (p)ppGpp modulation (Figure 1 and Supplementary Table S1). Proteins involved in the metabolism of amines and amino acids exhibit very high-affine (p)ppGpp binding with Kd’s as low as 0.01 μM (E. coli LdcI,,). They are followed by a multitude of enzymes involved in the metabolism of nucleotides featuring Kd values below 10 μM, the tRNA modification GTPase TrmE and riboswitches of the YkkC type 2a. Notably, their Kd values below the assumed basal levels of (p)ppGpp imply a certain degree of regulation by (p)ppGpp even under non-stringent conditions (see below). (p)ppGpp binding to targets involved in transcription [i.e., E. coli RNA polymerase (RNAP); ] and Francisella tularensis MglA-SspA proceeds between 2 and 25 μM. It shall be noted that (p)ppGpp elicit control over E. coli RNAP through modulation of relative transcription rates from different promoters instead of a general reduction or induction of RNAP activity. These targets are succeeded by further enzymes of nucleotide metabolism with Kd values ranging between 30 and 80 μM. Except for GpmA exhibiting a Kd of 52 μM, proteins associated with carbon metabolism feature dissociation constants/IC50 values of 132–800 μM. In a similar range falls the inhibition of DNA replication via (p)ppGpp binding to DnaG. Enzymes of fatty acid metabolism exhibit weak binding affinities, suggesting regulation by (p)ppGpp only to take place under full stringent control (i.e., 1,000 μM (p)ppGpp). Some enzymes partaking in (p)ppGpp metabolism, i.e., the (p)ppGpp synthetases RelQ and Rel/RelA, are stimulated in their activity by pppGpp and to a lesser extent by ppGpp. Furthermore, the enzymes GdpP, Pde2, and PgpH involved in the degradation pathway of the second messenger cyclic-diadenosine-monophosphate (c-di-AMP) display Ki values for inhibition by ppGpp between 35 and 400 μM.

6Importantly, in proteins where (p)ppGpp bind competitively to another compound (Figure 1 and Supplementary Table S1), the intracellular concentration of this compound also affects the fraction of the (p)ppGpp-bound protein. The influence of this other compound rises with its intracellular concentration. For example, the concentration of GTP under non-stringent conditions is assumed to fall between approximately 1 and 5 mM but may drop heavily due to (p)ppGpp-dependent inhibition of GTP anabolism. Thereby, inhibition of GTPases is governed by the (p)ppGpp to GTP ratio that is indicative for the metabolic state of the cell. Assuming, for example, similar Kd values for (p)ppGpp and GTP binding, only approximately 1% of a protein would be inhibited at 10 μM (p)ppGpp/1 mM GTP while this fraction rises to 50% at equal concentrations of both nucleotides. In fact, the Ki for the inhibition of the GTPase RbgA of 300 and 800 μM (ppGpp and pppGpp in presence of 50S ribosomal subunits determined at 1 mM GTP;; Supplementary Table S1) is a better estimate than the Kd value. Similar considerations apply to some enzymes of nucleotide metabolism, e.g., GMK with its substrate GMP or Gpt with the substrate guanine. In contrast to GTP, however, the intracellular concentrations of those nucleotides are lower, i.e., 24 μM GMP and 19 μM guanine, suggesting 90% inhibition at (p)ppGpp concentrations of approximately 200 μM. In contrast, the enzymes YgdH (PpnN; ), PurF, and LdcI are allosterically inhibited by (p)ppGpp, rendering inhibition by (p)ppGpp independent of the substrate concentration. Summarized, the kinetic parameters of (p)ppGpp interaction with target proteins indicate the following hierarchy of adaptation: amine and amino acid metabolism, nucleotide metabolism, translational and ribosome biogenesis GTPases, DNA replication, carbon metabolism, and fatty acid synthesis (Figure 1).

7Increasing evidence recommends that bacteria require basal (p)ppGpp levels to maintain homeostasis and strains that are completely devoid of any (p)ppGpp (known as (p)ppGpp0) reveal multiple disabilities in cell division, transcription, translation, GTP homeostasis, and antibiotic tolerance. In Bacillus subtilis, the lack of (p)ppGpp causes elevated GTP levels, which through dysregulation of the GTP-dependent transcriptional repressor CodY result in auxotrophies for branched-chain amino acids. Similar CodY-mediated detrimental effects of ppGpp0-strains were observed for Enterococcus faecalis and Staphylococcus aureus. In a rel-deletion strain of Synechococcus elongatus, the transcript levels of 52–67% of all genes were upregulated at least 3-fold and the levels of rRNA elevated indicating a “transcriptionally relaxed” state. Furthermore, cell size and volume increased but could be restored by synthetic compensation of (p)ppGpp. In Vibrio cholera basal (p)ppGpp has been linked to the expression of virulence factors and cell motility. Basal (p)ppGpp was also required for the tolerance of E. faecalis against vancomycin. These reports emphasize the importance of basal (p)ppGpp levels, which through high-affine binding of the alarmones (Kd < 10 μM; Figure 1) exert a regulatory function in the absence of a (p)ppGpp-stimulating trigger.

8Intuitively, a prolonged mismatch of (p)ppGpp synthesis and degradation resulting in over-accumulation of (p)ppGpp, is detrimental for the cell. In E. coli, RelA and SpoT are responsible for the synthesis of (p)ppGpp, with only SpoT being able to hydrolyze the alarmones. As such, the role of SpoT is pivotal for preventing toxic (p)ppGpp accumulation as exemplified by SpoT-deletion strains being characterized by reduced growth rates and distorted amino acid requirements. Bacteria of the Firmicutes phylum, instead of possessing RelA and SpoT, harbor the bifunctional enzyme Rel. The hydrolase activity of Rel is required to prevent toxic concentrations of (p)ppGpp, which results in severe growth defects. This effect appears more notorious in the presence of the two small alarmone synthetases, RelP and RelQ, the (p)ppGpp synthetase activity of which is not properly contained in absence of the (p)ppGpp-degrading Rel. In Mycobacterium tuberculosis, (p)ppGpp degradation is equally essential since deletion of (p)ppGpp hydrolase activity, while retaining (p)ppGpp synthesis, lead to lethal toxicity evidenced by irregularities in colony formation irrespective of the nutritional environment. Notably, these M. tuberculosis strains were also impaired in their virulence in a mouse model during both acute and chronic infection, highlighting the interference with (p)ppGpp metabolism as a potential antimicrobial therapeutic strategy. Taken together, a tight regulation of the antagonistic (p)ppGpp producing and degrading activities is indispensable for many, if not all bacteria for (i) correctly adjusting (p)ppGpp levels in response to environmental cues, (ii) coordinating various cellular processes in an orchestrated manner, and (iii) avoiding lethal consequences due to (p)ppGpp over-accumulation.

9Bacteria possess a thorough toolbox of nucleotide-based second messengers to efficiently respond to diverse external cues. Among those, c-di-AMP is a signaling molecule mainly synthetized by Gram-positive bacteria, which fulfills a pivotal role in the osmotic homeostasis. The dynamic range of intracellular c-di-AMP is much lower than for (p)ppGpp and lies in the one-digit μM range (in B. subtilis 1.7 μM during vegetative growth to 5.1 μM 2 h after sporulation; ). The importance of c-di-AMP for bacterial physiology is also highlighted by the observation that both the complete absence and over-accumulation of c-di-AMP impede growth of B. subtilis. Comparison of the transcriptional profiles of S. aureus cells at high c-di-AMP or (p)ppGpp concentrations revealed an overlap of 27.9% between the two regulons. Intriguingly, (p)ppGpp inhibit the activity of the S. aureus c-di-AMP phosphodiesterase GdpP with a Ki of 129.7 ± 42.8 μM in vitro correlating with higher levels of c-di-AMP observed in vivo. Inhibition by ppGpp with an IC50 of 139 ± 5.6 μM was also reported for the second S. aureus c-di-AMP phosphodiesterase Pde2. Similar observations made for the c-di-AMP hydrolases YybT (GdpP) from B. subtilis and PgpH from Listeria monocytogenes. These Ki/IC50 values suggest that (p)ppGpp concentrations must rise above their basal levels to inhibit c-di-AMP degradation. Deletion of S. aureus GdpP with concomitant increase of c-di-AMP evokes elevated (p)ppGpp, thus implying that induction of one second messenger, (p)ppGpp or c-di-AMP, augments the other. Conversely, the depletion of the only c-di-AMP synthetase, dacA, in L. monocytogenes also induced a toxic accumulation of (p)ppGpp. This apparent twist might be caused in the narrow dynamic range of c-di-AMP whereby deflection in either direction triggers (p)ppGpp synthesis or indicate species-specific differences in the regulatory circuits. Nevertheless, the functional connection of c-di-AMP and (p)ppGpp is further substantiated by the observation that the essentiality of dacA in L. monocytogenes during growth in rich medium was abrogated in a (p)ppGpp0-strain or in the wild-type strain grown in minimal medium. Hereby, the activity of the GTP-dependent transcriptional regulator CodY appears critical to dacA essentiality in rich media in that mutation of codY in the (p)ppGpp0-strain renders dacA again essential. This implies that elements of the CodY regulon may be toxic in the absence of c-di-AMP-producing DacA. The GTP-loading state and activity of CodY are in turn linked to (p)ppGpp through the inhibition of GTP anabolic enzymes involved in nucleotide metabolism by the alarmones (see above). Thus, as a whole, the interdependencies between c-di-AMP and (p)ppGpp conceivably require a high degree of coordination and involve a number of different cellular processes.

10In the 1970s, the accumulation of two further magic spots designated as (p)ppApp was observed in B. subtilis during the early stages of sporulation and the (p)ppApp synthetic activity retrieved from the ribosomal fraction, which indicates the contribution of an RSH. The first RSH to catalyze (p)ppApp formation was recently identified in Methylobacterium extorquens, however, (p)ppApp-producing activity was not evidenced for the E. coli RelA enzyme. Another source of (p)ppApp is the nucleoside 5'-diphosphate kinase from Streptomyces morookaensis although the enzyme being secreted raises the question of its physiological relevance for this organism. The high similarity of (p)ppGpp and (p)ppApp – they only differ in their nucleobase – advises a putative overlap of their target spectra. Recent studies evidence that E. coli PurF and RNAP, both of which are validated targets of (p)ppGpp, also bind (p)ppApp. However, while in PurF both second messengers bind in similar fashion at the same site and inhibit the enzyme with equal potency, the binding sites on RNAP are discrete and (p)ppApp enhance transcription from, e.g., the rrnB P1 promoter as opposed to (p)ppGpp. In one study, ppApp also inhibited the (p)ppGpp synthetic activity of E. coli RelA with an IC50 of 24.5 ± 3.5 μM raising the possibility that, in fact, (p)ppApp and (p)ppGpp are antagonists. Besides potentially fulfilling regulatory functions in their host cell, (p)ppApp also serve as toxins during interbacterial warfare. Hereby, the type VI secretion system (T6SS) effector protein Tas1 of Pseudomonas aeruginosa PA14, a (p)ppApp synthetase, is injected into the target cell and depletes the cellular ATP pools, resulting in dysregulation of the metabolome and, ultimately, cell death. Endogenous ppApp, the metabolism of which is embedded in toxin-antitoxin systems, might represent an intricate pathway to control growth. Whether the (p)ppApp molecules exert growth control independent from (p)ppGpp via a separated protein target spectrum or in a competitive manner to (p)ppGpp remains to be investigated.

11The functions of (p)ppGpp are often simplistically portrayed as a biphasic switch between relaxed and stringent environmental conditions, the latter triggering a major metabolic rearrangement through induction of (p)ppGpp synthesis. The presence of gradation in (p)ppGpp-dependent regulation of transcription is an approved mechanism. However, the wide range of binding affinities among the (p)ppGpp-affected protein targets covering four orders of magnitude furthermore suggests a post-translational adaptational program activated hierarchically by increasing (p)ppGpp. The depicted scheme illustrates the importance of high-affine (p)ppGpp targets that explicitly require regulation, thus suggesting a protagonism of the alarmones on cell homeostasis. We believe that the post-translational sequential “dimming” of protein activities by (p)ppGpp, a conception potentially also true for related second messengers like (p)ppApp and c-di-AMP, plays a major role in successful adaptation of microorganisms.

12The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

13WS and GB conceived the study. WS and VZ performed the analysis. WS, VZ, and GB wrote the manuscript. All authors contributed to the article and approved the submitted version.

14The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.