Great Times for Small Molecules: c-di-AMP, a Second Messenger Candidate in Bacteria and Archaea

See allHide authors and affiliations

Science Signaling  19 Aug 2008:
Vol. 1, Issue 33, pp. pe39
DOI: 10.1126/scisignal.133pe39


Successful cell division in pro- and eukaryotes is ensured by checkpoints that regulate cell cycle progression. Structural and biochemical analyses of the DNA integrity scanning protein (DisA) have recently shown that its domain of unknown function, DUF147 [renamed DAC (for diadenylate cyclase)], has diadenylate cyclase activity. This diadenylate cyclase activity is abolished when DisA binds to branched DNA substrates, which arise during DNA double-strand breaks that can spontaneously occur during DNA replication. This finding identifies cyclic di(3′→5′)-adenylic acid (c-di-AMP) as a second messenger candidate that signals DNA integrity in Bacillus subtilis during sporulation, a specialized cell division process that leads to formation of a dormant cell called a spore. The DAC domain is widespread in Bacteria and Archaea; moreover, it is found in proteins containing diverse domains, suggesting that c-di-AMP acts as a second messenger molecule in response to various signals besides branched DNA. To elucidate the biological importance and molecular mechanisms of action for c-di-AMP and the recently recognized second messenger c-di-GMP will require a multidisciplinary approach.

Just a few years ago, the intracellular signaling landscape in bacteria looked quite deserted, resembling a pared-down version of the eukaryotic signaling landscape. In particular, small secondary signaling molecules seemed to be quite uncommon; these were restricted to cyclic adenosine 3′,5′-monophosphate (cAMP), used to signal changes in glucose concentration (1), and guanosine-tetraphosphate (ppGpp), an alarmone produced in response to nutrient stress (2). This view of bacteria as nutrient-responsive bags changed recently when c-di-GMP [cyclic di(3′→5′)-guanylic acid], the synthesis of which is stimulated by a broad range of signals, was recognized as a second messenger used by many bacteria that mainly affects biofilm-related phenotypes (35). A recent report from the group of K.-P. Hopfner (6) strongly suggests the presence of another previously unrecognized cyclic dinucleotide signaling molecule in bacteria, namely, c-di-AMP [cyclic di(3′→5′)-adenylic acid]. In addition, their work assigns diadenylate cyclase activity to the domain of unknown function DUF147.

The DUF147 Domain of DisA Is a Diadenylate Cyclase

Progression through the cell cycle in eukaryotes and prokaryotes is regulated by DNA damage checkpoints (7). The Hopfner group solved the crystal structure of DisA (DNA integrity scanning protein), a bacterial protein that scans the chromosome for DNA double-strand breaks (8). In the soil bacterium Bacillus subtilis, DisA provides a checkpoint for DNA damage during sporulation, a specialized process involving DNA replication and unequal cell division that leads to the development of a dormant cell called a spore (9).

DisA is a dumbbell-shaped protein that has the domain structure DUF147–domain 2–HhH (helix-hairpin-helix). Although domain 2 has no known structural homologs, the HhH domain is a nonspecific DNA binding domain that is associated with DNA phosphate backbone binding in numerous DNA-interacting proteins (10). In the crystal structure and in solution, DisA assembles as an octamer (Fig. 1). Four DisA molecules, which are arranged in parallel, interact head to head with another DisA tetramer through their N-terminal DUF147 domains. Surprisingly, electron density data of a substantial region between two opposing DUF147 domains, in combination with the chemical characteristics of the corresponding DUF147 pockets, suggested that c-di-AMP was located between the opposing DUF147 domains of DisA monomers, a prediction that was indeed confirmed by mass spectrometric analysis of the purified DisA protein.

Fig. 1.

Proposed c-di-AMP signaling pathway regulating the initiation and inhibition of Spo0A-mediated sporulation events in B. subtilis. A multicomponent phosphorelay with the sensor kinases KinA through KinE and the phosphotransfer proteins Spo0F and Spo0B leads to the phosphorylation of Spo0A (27, 28), a key transcriptional response regulator that initiates the transcription of genes involved in sporulation. On the intact chromosome, DisA loosely slides on the DNA, thereby producing c-di-AMP. c-di-AMP positively signals to the Spo0A phosphorelay either directly or through binding to a target, thereby initiating sporulation. After DNA damage, DisA arrests at the site of DNA double-strand breaks and binds tightly to branched DNA, thereby inhibiting c-di-AMP production (6) and subsequently sporulation (8). Arrows without frame are assigned to hypothetical pathways.

c-di-GMP was identified more than 20 years ago as an allosteric activator of cellulose synthase in the fruit-degrading bacterium Gluconacetobacter xylinus (11). However, because cellulose biosynthesis in bacteria was considered atypical, this discovery evoked interest in plant biologists but was ignored by microbiologists. c-di-AMP, on the other hand, had only been synthesized chemically (12) and had never previously been identified in living cells. Faced with the alternatives of a copurification artifact or of the DUF147 domain having c-di-AMP–synthesizing activity, the authors unambiguously confirmed the latter by biochemical analysis. After depleting DisA from copurified c-di-AMP, Witte et al. (6) demonstrated that the resulting apoDisA protein converted adenosine triphosphate (ATP) to c-di-AMP and showed a strong preference for ATP over other nucleotides as a substrate. They therefore renamed the DUF147 domain DAC (for diadenylate cyclase).

Witte et al. (6) observed allosteric regulation of DisA diadenylate cyclase activity. DisA binding to branched nucleic acids through its C-terminal HhH domain inhibited its diadenylate cyclase activity, whereas binding to double-stranded DNA was substantially weaker and did not affect diadenylate cyclase activity. Branched DNA is formed as an intermediate structure during the repair of DNA double-strand breaks by homologous recombination (13). DNA double-strand breaks occur spontaneously during the cell cycle, for example, during segregation of chromosomes or when the replication fork is stalled (14, 15). DNA double-strand breaks can also by induced by exogenous agents. Irrespective of how DNA double-strand breaks are created, damaged DNA interferes with successful sporulation.

Implications of DisA Diadenylate Cyclase Activity

In vivo observations show that DisA diffuses through the bacterial cell as a single focus (probably consisting of one or several octamers) that pauses at sites of DNA double-strand breakage, thereby delaying sporulation (8). On the basis of the biochemical findings, one can envisage that octameric DisA produces c-di-AMP while scanning the DNA with its outwardly oriented HhH domains and that DisA binding to branched nucleic acids subsequently inhibits c-di-AMP production. The question then arises as to whether c-di-AMP is the messenger molecule that couples DNA damage to sporulation arrest (Fig. 1). This question could be answered by complementing a disA deletion mutant with a form of DisA with constitutive diadenylate cyclase activity. However, even if c-di-AMP is not the messenger mediating this particular event, it seems likely that c-di-AMP targets must exist that connect the c-di-AMP signal to a physiological response. In addition, c-di-AMP must be degraded, exported out of the cell, or both to confer the highly dynamic adjustment in concentration required to provide the sufficient temporal and spatial specificity for a second messenger signal.

The DAC Domain Is Widespread in Bacteria and Archaea

c-di-AMP signaling does not seem to be restricted to the soil bacterium B. subtilis. The 120–amino acid–long DAC domain is found in 332 protein sequences from 275 species in the Pfam database of protein families and domains (16), suggesting that c-di-AMP is a widespread signaling molecule. DAC domain proteins are encoded by both bacterial and archaeal genomes, suggesting that c-di-AMP signaling occurs in at least two kingdoms of life. In bacteria with highly reduced genomes, intriguingly, the DAC domain is found in some human-associated bacteria such as Chlamydia spp. and Mycoplasma spp.

A substantial number of bacterial proteins, although not the majority, display the same domain architecture as DisA: DAC–domain 2–HhH, suggesting that DNA double-strand breakage is a signal that is commonly sensed by c-di-AMP–producing enzymes. DisA homologs occur mainly in Gram-positive bacteria. However, not all species that harbor a DisA homolog undergo sporulation like B. subtilis, suggesting that the coupling of c-di-AMP synthesis to sensing of DNA double-strand breaks is a cell cycle control mechanism that is not restricted to sporulating organisms.

Domains of Unknown Function Associated with DAC

Most domains associated with DAC are of unknown function, demonstrating that the signaling pathways that regulate and respond to c-di-AMP are mainly unexplored. Consequently, the connection of c-di-AMP signaling with other cellular signals and pathways is not obvious from bioinformatic analysis. Equally likely, microbial physiology and behavior that has integrated c-di-AMP signaling is not well explored. The YbbR domain, a domain of unknown function found in B. subtilis and other Gram-positive bacteria, and a domain structurally homologous to the noncatalytic C-terminal domain of pyruvate kinase (SCOP) are uncharacterized domains frequently associated with DAC (Fig. 2). The unexplored diversity of signal input and output domains suggests that various extra- and intracellular stimuli regulate c-di-AMP signaling.

Fig. 2.

Domain architecture of DAC domain proteins. More than 50 proteins harbor the domain architecture of DisA from B. subtilis (DisA_BACSU) and Thermotoga maritima, DAC-domain 2-HhH (10). DisA homologs are encoded by chromosomes of actinobacteria (high G+C, Gram positive), firmicutes (low G+C, Gram positive), and Thermotogae, a group of thermophilic Gram-negative bacteria. TM-DAC domain proteins are predominantly found in firmicutes. DAC-YbbR proteins are found in Geobacter spp. (δ-proteobacteria) and Streptococcus spp. Proteins with the domain architectures SCOP-DAC (Methanococcus spp.) and REC-PAS-PAS-DAC (Methanoregula boonei) are solely found in archaea. The right-hand column gives the names of these proteins in the InterPro database (29). PAS, Per periodic clock protein; Arnt, Ah receptor nuclear translocator protein; Sim, single-minded protein; REC, CheY-homologous receiver domain; TM, transmembrane. For YbbR and SCOP domain definitions, see text. The scale bar represents number of amino acid residues.

c-di-AMP Signaling Network in B. subtilis

Most of the bacterial species that encode DAC domain–containing proteins express only a single representative. Intriguingly, however, B. subtilis harbors three DAC domain–containing proteins (Fig. 3): DisA, YbbP (a protein predicted to be membrane-localized), and YojJ (a protein predicted to be cytosolic). Aside from the DAC domain, YbbP and YojJ have domain compositions distinct from that of DisA, suggesting that their putative diadenylate cyclase activity is regulated by signals other than branched nucleic acids. The apparent redundancy in enzymatic activity, through the presence of the DAC domain combined with the diversity of domain structure, suggests the presence of spatially and temporally distinct c-di-AMP pools. Indeed, DisA is regulated by the salt stress–responsive sigma factor SigM (17) and YbbP is a member of the alkaline-inducible SigW regulon (18), whereas YojJ does not belong to any of these regulatory networks, indicating that expression of the three proteins occurs in response to different environmental signals.

Fig. 3.

c-di-AMP signaling in B. subtilis. B. subtilis has three DAC domain proteins that have different predicted localizations in the cell: the DNA binding DisA protein, the membrane-bound YbbP protein, and the cytosolic YojJ protein.

c-di-AMP and c-di-GMP as Second Messengers

Therefore, in a short time span, two previously unidentified second messengers, cyclic di-GMP and c-di-AMP, have been identified in bacteria. The report by Witte et al. has uncovered many conceptual parallels in the synthesis of these two signaling molecules. For example, both molecules are synthesized by domains that are usually part of multidomain proteins, so more than one input signal may affect their enzymatic activity. Indeed, tight allosteric control of the diguanylate cyclase activity of the Gly-Gly-Asp-Glu-Phe (GGDEF) domain is the rule rather than the exception (19, 20). As noted above for B. subtilis, some bacteria and archaea harbor more than one DAC domain protein, suggesting spatially and temporally distinct c-di-AMP pools. Spatial and temporal regulation of individual signaling cascades leading to a specific phenotype has also been suggested (3) [and partially proven (21, 22)] for the c-di-GMP signaling pathway. In the future, multidisciplinary approaches will be required in order to unravel the biological importance and regulatory mechanisms of these newly studied signaling molecules.

A Fresh View on Bacterial Signaling

Our view on bacterial signaling pathways, especially with respect to small second messenger molecules, has been restricted. The model second messenger pathway in bacteria, the cAMP signaling pathway in the laboratory work horse Escherichia coli, is remarkably simple, with one cAMP synthetic enzyme, one cAMP receptor protein, and one cAMP-degrading enzyme (1). However, there exists a complexity in the nature and network structure of signaling system(s) even in an average bacterial cell, which has a volume more than 1000 times smaller than that of a eukaryotic cell. For example, the chromosomes of Vibrio vulnificus and Shewanella spp. encode more than 60 potential c-di-GMP–synthesizing proteins, suggesting that the c-di-GMP signaling network in these marine bacteria is extremely complex (23). But c-di-GMP signaling is already sufficiently complex in E. coli with 31 proteins in total involved in c-di-GMP metabolism.

Broader Perspectives

The report by Witte et al. has opened up a novel research field on c-di-AMP–related signaling processes. However, the impact of their findings probably goes beyond future investigations of c-di-AMP signaling in organisms known to harbor DAC. In c-di-GMP signaling, degradation of the messenger can be performed by two structurally unrelated c-di-GMP–specific phosphodiesterases (24, 25). Although the alarmone ppGpp was discovered nearly 40 years ago (2), novel ppGpp synthetic enzymes have been identified in B. subtilis only recently (26). Once c-di-AMP has been demonstrated in vivo, screens can be set up to investigate the presence of c-di-AMP even in organisms that do not harbor the DAC domain, possibly identifying alternative pathways for the synthesis of c-di-AMP.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
View Abstract

Stay Connected to Science Signaling

Navigate This Article