- Email: [email protected]
Second messenger – Sensing riboswitches in bacteria
Accepted Manuscript Title: Second messenger-sensing riboswitches in bacteria Author: Arati Ramesh PII: DOI: Reference:
S1084-9521(15)00216-5 http://dx.doi.org/doi:10.1016/j.semcdb.2015.10.019 YSCDB 1849
To appear in:
Seminars in Cell & Developmental Biology
Please cite this article as: Ramesh A, Second messenger-sensing riboswitches in bacteria, Seminars in Cell and Developmental Biology (2015), http://dx.doi.org/10.1016/j.semcdb.2015.10.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Second messenger-sensing riboswitches in bacteria Arati Ramesh* *National Center for Biological Sciences, GKVK Campus, Bellary Road, Bangalore, India- 560065 *Corresponding author. Email address [email protected] Tel: 91-80-2366 6930, Fax: 91 80 23636662
an
us
cr
ip t
Abstract Signal sensing in bacteria has traditionally been attributed to protein-based factors. It is however becoming increasingly clear that bacteria also exploit RNAs to serve this role. This review discusses how key developmental processes in bacteria, such as community formation, choice of a sessile versus motile lifestyle, or vegetative growth versus dormant spore formation may be governed by signal sensing RNAs. The signaling molecules that affect these processes, the RNAs that sense these molecules and the underlying molecular basis for specific signal-response are discussed here. keywords: riboswitch, second messenger, sporulation, biofilms, differentiation
M
Abbreviations: c-di-GMP, 3’,5’-cyclic-di-guanosine monophosphate; c-di-AMP, 3’,5’cyclic-di -Adenosine monophosphate; 3’,3’ c-GAMP, 3’,3’ cyclic guanosine monophosphate adenosine monophosphate; 2′,3′-c-GAMP, 2’,3’ cyclic guanosine monophosphate adenosine monophosphate; 5’ UTR, 5’ Untranslated Regions.
Ac ce pt e
d
1. Introduction: Riboswitches as signal sensors in bacteria Bacteria sense and respond to their environment. This response occurs not only through protein-based sensors but also via RNAs that directly sense signaling molecules and in turn control gene expression. Such signal sensing RNAs, called riboswitches, typically reside in the 5’ UTRs of mRNA, in a cis configuration with respect to the gene that they control. Riboswitches are characterized by an aptamer domain that directly binds a metabolite, and an expression platform that transmits metabolite-induced effects to the downstream gene[1-3] (Figure 1A). Metabolite binding may turn on or off the downstream gene in one of many ways; by modulating a transcriptional terminator to affect genes at the post transcriptioninitiation level, by modulating the ribosome binding site to control genes at the level of translation initiation, or in rare cases control mRNA stability. More recently it has been shown that a class of riboswitches can coordinate with proteins such as the Rho transcription termination factor to elicit gene regulation[4]. Riboswitches have also been shown to control the expression of other non-coding RNAs[5], suggesting a complex network of RNA-based regulation in bacteria. In eukaryotes, the extent of riboswitch-based regulation is as yet unclear, though a class of thymine pyrophosphate sensing riboswitches has been shown to function via the control of mRNA splicing[6]. Based on the types of metabolites recognized, over 20 families of riboswitches have been identified thus far. Ligands sensed by riboswitches range from complex 1
Page 1 of 15
an
us
cr
ip t
metabolites such as vitamins (cobalamine[7], thymine pyrophosphate[8], flavin mononucleotide[9]) and cofactors (S-adenosyl methionine, S-adenosyl homocysteine), to small ions such as cations[10-12] (Mg2+, Ni2+/Co2+, Mn2+) and anions[13] (F-). RNAs have also been shown to sense nucleotides[14] (adenine[15], guanine, 2'-deoxyguanosine and the guanine analog PreQ1), amino acids (glutamine[16], lysine[17], glycine[18]), sugars[19] (glucosamine 6-phosphate) and charged tRNAs[20]. These classes of riboswitches were identified first through computational methods that predict conserved, structured classes of RNAs in sequenced genomes. The genomic contexts of the predicted RNAs have since then been used to rationally predict and test metabolites that are likely to be sensed by the RNA[21]. The majority of riboswitches appear to control genes encoding transport proteins or involved in the synthesis of metabolites. However, some classes of RNAs are associated with such a diverse collection of genes that it has been particularly challenging to identify their cognate ligand simply based on the genomic context of the RNA. This has been especially so for RNAs that sense molecules influencing developmental processes in bacteria, such as biofilm formation or sporulation since there is no single class of genes that control these processes. These riboswitch classes and the metabolites affecting these developmental processes are the primary focus of this review.
Ac ce pt e
d
M
2. Riboswitches that sense second messengers Bacteria respond to their environment by sensing signals, whose transduction eventually leads to developmental, physiological and behavioral changes in the bacterial cell. Signaling molecules thus play a central role in the manifestation of a variety of developmental processes in bacteria. While signal receptors form the very first layer of sensing, small molecule second messengers then carry signals from receptors to downstream effectors, which could be protein or DNA targets. In eukaryotes, nucleotide-based second messengers such as cyclic AMP and cyclic GMP are commonly used as signaling molecules. In bacteria, cyclic AMP and another intracellular molecule called guanosine tetra or penta phosphate (ppGpp) are more commonly used, though these are restricted to the response to nutrient limitation and carbon metabolism. More recently, the roles of novel second messengers, which are often dinucleotide molecules such as cyclic di-AMP, cyclic di-GMP and the recently discovered cyclic-AMP-GMP, have gained prominence. These molecules are circularly fused nucleotides where two molecules of AMP or GMP or both, form 3’,5’bonds (Figure 1B). Distributions of these purine derived cyclic nucleotides vary across organisms. Some like Cyclic-di-GMP are present in bacteria but have not been detected in eukaryotes thus far. Others like cGAMP have been detected across organisms, but are present as different stereoisomers. While the 3’,3’ linkage isomer of cGAMP has been detected in bacteria, eukaryotes show the presence of the 2’,3’ linkage isomer of cGAMP (Figure 1B). Although the biology around these molecules is not yet fully understood, it is clear that they affect myriad aspects of developmental processes. Recently, three classes of signal-sensing RNAs have been discovered that recognize bacterial second messengers with incredibly high affinity and specificity while discriminating against other similar molecules in the cell.
2
Page 2 of 15
These discoveries substantially expand the role and scope of riboswitches in bacterial biology.
us
cr
ip t
2. 1. The c-di-AMP riboswitch The ydaO/yuaA RNA was first predicted computationally over a decade ago[22]. Since then, over 3000 representatives of this RNA have been found in bacterial genomes[23], upstream of genes involved in cell wall metabolism, osmotic stress and sporulation. However, the sheer diversity of genes associated with the RNA long prevented its cognate metabolite from being identified. Recently, using a biochemical pull-down approach, a metabolite from yeast extract was shown to interact with the RNA[24]. Mass spec analysis identified this metabolite as c-di-AMP, a cyclic dinucleotide made of two adenosines joined with a 3’,5’ phosphodiester linkage.
Ac ce pt e
d
M
an
The biosynthesis of c-di-AMP itself involves the joining of 2 ATP molecules by proteins containing- diadenylyl cyclase domains (DAC)[25]. These proteins bear little homology in sequence or structure to the c-di-GMP or c-GAMP synthesizing enzymes. C-di-AMP is involved in regulating a range of diverse processes in bacteria, and is implicated in cell wall homeostasis, fatty acid synthesis in mycobacterial species[26], growth of Staphylococcus aureus in low K+ conditions[27] and DNA integrity in Bacillus subtilis[28]. In S. aureus high c-di-AMP levels are also linked to decreased cell size. In B. subtilis and Listeria monocytogenes decrease in c-di-AMP corresponds to higher sensitivity to cell wall targeting antibiotics. Listeria in the cytosol of the host secretes c-di-AMP, which is sensed by the host and used as a cue to mount a type I IFN response, activating defense mechanisms in the host[29]. Although DAC domain proteins are widely present in bacterial genomes, c-di-AMP itself has been experimentally detected only in some bacteria, with intracellular concentrations estimated to be in the 1.7 to 2 μM range in B. subtilis and S. aureus[30]. In addition to bacteria, DAC domain-containing proteins have been found in archaea, suggesting a possible role for c-di-AMP in archaea[31]. With the identification of c-di-AMP RNAs and the genes they control, we can now perhaps learn more about the roles of c-di-AMP itself in bacterial developmental processes. The 165 nucleotide ydaO riboswitch from Bacillus subtilis showed a strong binding affinity of 10 pM and an estimated stoichiometry of 1:1 for c-di-AMP[24] (Figure 1C). This stoichiometry and affinity however is not conserved within the ydaO riboswitch class since some ydaO homologs from other bacteria have been found to bind 2 molecules of c-di-AMP per RNA and display weaker binding affinities[32]. This could indicate diverse structural and functional modes for c-di-AMP recognition within the same riboswitch class, in tune with the needs of the particular species of bacteria. Strikingly, molecules related to c-di-AMP such as isoforms of AMP, ADP, cAMP, c-di-IMP, c-di-GMP and a deoxy-derivative c-di-dAMP are strongly discriminated against by this class of RNAs, suggesting that these RNAs are molecular machines of exquisite specificity. Paradoxically, the ydaO RNA binds ATP in vitro with an affinity of 0.6 mM[33]. This falls within the range of intracellular ATP concentrations and it been proposed that this RNA might sense 3
Page 3 of 15
ip t
ATP in the cell. However the significantly higher affinity for c-di-AMP and genes associated with the RNA class indicate that the cellular signal being sensed is likely to be c-di-AMP. The intracellular concentrations of c-di-AMP have been estimated at ~2 uM in the cytoplasm of B. subtilis and S. aureus. This is orders of magnitude higher than the picomolar binding affinity of the riboswitch, strongly suggesting that the riboswitch function in this case is driven by the kinetics of the system rather than thermodynamics wherein the RNA would function at equilibrium with cellular c-di-AMP.
M
an
us
cr
High-resolution crystal structures have been solved for three ydaO aptamers from different bacteria [32, 34, 35]. The overall architecture of the RNA comprises a fivehelical arrangement that contains a zippered up bubble and a pseudoknot that stabilizes long-range contacts in the RNA (Figure 2A). Two c-di-AMP molecules are bound at sites lined with conserved nucleotides, suggesting that these are distinct and functionally important ligand binding sites. One ligand molecule binds at the collinearly arranged P4-P5 helices and pseudoknot end of the molecule while the other binds at the collinearly arranged P3-P2 helices and the zippered bubble region. Distances between coordinating RNA functional groups are organized in a way that would exclude the amino group of guanine nucleotide derivatives, explaining why the c-di-AMP riboswitch does not interact with c-di-GMP or c-GAMP molecules.
Ac ce pt e
d
C-di-AMP assumes an extended conformation with Aα and Aβ wedged between nucleotide bases of the RNA, likely stabilizing junction regions of the RNA[32, 34, 35]. Both within and between the two ligand-binding sites, there is pseudosymmetry, a rather unique feature of the RNA wherein two near identical sites are created for ligand binding. Whether or not the two sites interact cooperatively is yet to be determined and would be relevant in the context of intracellular c-di-AMP concentrations. The principles of c-di-AMP recognition appear largely conserved between c-di-AMP binding proteins and RNAs. In crystal structures of both the DisA protein[36] and the ydaO riboswitch[32, 35], the ligand contacts the macromolecule with two adenine bases and the sugar rings of the ligand are puckered as a 3’-endo configuration. The sugar, phosphate as well as nucleobases of the ligand make extensive hydrogen bonding contacts with the protein and the riboswitch. This therefore appears to be an illustrative example of proteins and RNAs evolving similar binding sites for an important biomolecule. The premise of riboswitch function relies on an expression platform being modulated by the metabolite-bound aptamer. The c-di-AMP bound structure of the RNA shows a P1 helix platform that abrogates an antiterminator, such that in the presence of c-di-AMP, an intrinsic transcription terminator stem would form distal to the riboswitch. The absence of c-di-AMP would allow disruption of the P1 helix, thereby forming an antiterminator structure that results in expression of the downstream gene. This agrees with in vitro and in vivo observations for the B. subtilis ydaO RNA, wherein the presence of c-di-AMP results in a shorter transcript and decreased expression of a reporter gene fusion[24]. In S. coelicolor, instead of a 4
Page 4 of 15
transcriptional terminator platform, a ribosome-binding site overlaps with the pseudoknot, suggesting that c-di-AMP binding may regulate translation and not transcription in some cases[37].
an
us
cr
ip t
The majority of ydaO RNAs associate with genes that encode proteins involved in cell wall metabolism. This is followed by proteins for the transport and synthesis of osmoprotectants[37] and genes are involved in peptidoglycan breakdown and remodeling. Another commonly associated set of genes encodes transporters of potassium or other solutes implicated in cellular osmosis. In B. subtilis, transcription of the RNA associated genes ydaO and ktrAB shows upregulation ~10minutes after spore germination[38] where as in Streptomyces coelicolor the cell wall hydrolase enzymes associated with this riboswitch appear to be expressed throughout development[39]. The distribution of this riboswitch class strongly correlates with genes encoding diadenylate cyclase proteins that are involved in the synthesis of cdi-AMP. In Staphylococcus aureus, genes downstream of ydaO often encode proteins that bind c-di-AMP. Many of these processes are integral to bacterial sporulation and germination in a range of diverse bacteria.
Ac ce pt e
d
M
2.2. The c-di-GMP riboswitch Discovered in the 1980s, c-di-GMP is now recognized to be a key determinant of developmental decisions in bacteria such as switching between motile and sessile biofilm states. [40]. C-di-GMP binding proteins such as the GGDEF domain (Pfam:PF00990)-containing diguanylate cyclases and the EAL domain (Pfam:PF00563)-containing phosphodiesterases together maintain the intracellular concentration of c-di-GMP. These proteins were initially linked to biofilm matrix formation, rugose colony morphology, virulence, twitching motility etc., suggesting a link between c-di-GMP and these various biological processes. High intracellular cdi-GMP has since been shown to promote sessile biofilm lifestyle whereas low c-diGMP promotes motile planktonic lifestyle. Other processes such as exopolysaccharide production, flagellum biosynthesis, pilus assembly, cell-cell communication, quorum sensing, expression of virulence factors etc. have been linked to c-di-GMP levels suggesting a very broad role for this second messenger[40, 41]. RNAs that sense c-di-GMP were identified upstream of genes encoding diguanylate cyclases and phosphodiesterases and other c-di-GMP related proteins. These RNAs contain the GEMM (Genes for the Environment, Membranes and Motility) motif[42], now known to occur in ~545 RNA representatives that are further classified as type I (500 RNAs) and type 2 (45 RNAs) based on differences in RNA structural elements. Unlike the unbiased approach leading to the discovery of c-di-AMP as the ligand for the ydaO riboswitches, in the case of GEMM motifs, strong gene associations with diguanylate cyclases and phosphodiesterases led to the finding that this RNA recognizes the bacterial second messenger c-di-GMP[43].
5
Page 5 of 15
us
cr
ip t
In vitro binding analysis of the c-di-GMP riboswitches showed a dissociation constant for c-di-GMP between 10 pM to 1 nM (Figure 1C, estimated using different methods) for type 1 and 200 pM to 2 nM for type 2 RNAs with a stoichiometry of 1:1[43]. Analogs of c-di-GMP and breakdown products of c-di-GMP such as pGpG are discriminated against by a difference in affinity greater than 3-fold for the type 1 RNA, suggesting high specificity for ligand recognition similar to that seen for c-diAMP riboswitches. Interestingly, the GEMM motif RNAs bind c-di-GMP with much higher affinity than other c-di-GMP binding proteins (EAL domains, GGDEF domains etc.) that typically display affinities only in the range of 50 nM to low μM. Detailed kinetic analyses of this RNA reveal that the half life of the ligand bound to RNA is in the order of a month, suggesting that within biological contexts this complex is essentially irreversible[44], committing the bacteria to a specific regulatory response.
Ac ce pt e
d
M
an
Given that cyclic dinucleotides are structurally and chemically similar, how then do RNAs build in specificity in their binding sites? Structural studies have dissected the molecular mechanisms of c-di-GMP binding to this class of riboswitches in detail, revealing this distinction. Crystal structures of the type 1 riboswitch show a Yshaped architecture where three helices P1, P2 and P3 come together to form a junction that harbors the ligand-binding site[44] (Figure 2B). Tetraloop-receptor interactions between P2 and P3 hold them in a parallel orientation to each other. This region is highly stabilized through multiple interactions and does not undergo local ligand-induced changes in secondary structure. Unlike the c-di-AMP riboswitch, here the ligand is recognized asymmetrically wherein Gβ of the ligand makes Watson Crick contacts with a conserved cytidine, while Gα is stabilized by Hoogsteen interactions. The conserved cytidine prevents an adenosine from binding the RNA, thus excluding c-di-AMP from being recognized by this RNA. Interactions occur with the base as well as the sugar phosphate backbone of Gβ while the Watson-Crick surface of Gα is not recognized by the RNA and lies in a solvent accessible cavity at the helical junction. In addition to hydrogen bonding, intercalated base stacking and metal ion-mediated interactions with the phosphate group of the ligand further stabilizes the ligand in the binding pocket. A similar theme is observed in the thymine pyrophosphate riboswitch wherein the negatively charged functional group of the ligand is coordinated by a metal ion in order to facilitate RNA binding. Mutations made to the binding pocket nucleotides (C to U at the Gβ interaction and G to A mutation at the Gα interaction) resulted in RNAs that can select for c-di-AMP over c-di-GMP by over 4-fold. However, naturally occurring c-di-AMP riboswitches appear to have architectures and binding site configurations that are unique from c-di-GMP sensing RNAs[32, 34, 35, 44]. C-di-GMP sensing riboswitches show two unique architectures capable of binding this molecule. Crystallographic studies of the Clostridium acetobutylicum type 2 aptamer bound to c-di-GMP, reveal a tertiary architecture distinct from type I RNAs[45, 46] (Figure 2C). In addition to the three helices seen in type I c-di-GMP riboswitches, type 2 RNAs harbor a kink-turn motif and a pseudoknot whose
6
Page 6 of 15
us
cr
ip t
nucleotides are tightly conserved and show ligand-induced changes in secondary structure. The ligand is sequestered into a triplex at the edge of a pseudoknot. Although both type 1 and type 2 RNAs sequester the ligand in asymmetric binding sites, type 2 RNAs do not make Watson-Crick or Hoogsteen contacts with the Gβ and lack the critical, conserved cytidine present in type I. Also, unlike the type I riboswitches, only few contacts are made with the ribosyl phosphate backbone of the ligand, providing a possible explanation for the weaker binding affinities seen for type 2 RNAs. Distribution analysis of this RNA class across bacteria shows that although some bacteria have both type 1 and 2 switches, the majority seems to prefer one type over the other with type 2 being more prevalent in Clostridia[46]. Since these two types of riboswitches do not commonly co-occur in bacteria, this suggests variations in intracellular c-di-GMP concentrations, which may in turn govern processes differently.
d
M
an
Proteins such as PilZ also bind c-di-GMP asymmetrically and make sidechain stacking interactions with the Gα and Gβ of c-di-GMP[47]. However, these interactions are unlikely to be as stabilizing as those with RNA where the ligand is incorporated into the core of the RNA molecule. Thus the conformation of the ligand itself is different when bound to protein partners versus RNAs and would explain the significantly higher binding affinity seen for RNAs compared to proteins. Binding of the ligand likely nucleates formation of the P1 helix expression platform formed by the 5’ and 3’ ends of the aptamer, the sequence of which is not strictly conserved among representatives of this RNA class. The GEMM motif is associated downstream with both transcriptional terminators as well as translation controlling sites, suggesting that this RNA likely controls genes by common riboswitch mechanisms.
Ac ce pt e
Besides guanylate cyclase genes that are involved in c-di-GMP biosynthesis, and phosphodiesterases that are involved in c-di-GMP breakdown, the GEMM motif RNAs also associate with flagellum biosynthesis genes in Vibrio cholera. Here c-diGMP is linked to virulence wherein the production of cholera toxin is upregulated under conditions of low c-di-GMP concentrations[48]. This suggests that in some cases, riboswitch based sensing is coupled to organelle biosynthesis and lifestyle choices such as a switch between motile and sessile forms. The Vc2 riboswitch of Vibrio cholera is associated with a transcription regulator protein (VC1722) that is itself known to control pathways involved in competence[49]. The c-di-GMP riboswitches also control genes for cell differentiation, morphology, cell cell communication, exopolysaccharide production, and are linked to quorum sensing[40]. 2.3. The c-GAMP riboswitch The most recent member to join the family of cyclic dinucleotide messengers is cyclic GMP-AMP (c-GAMP). The discovery of the c-GAMP riboswitches was itself serendipitous. Analysis of GEMM domain containing type I c-di-GMP riboswitches revealed the presence of a single point mutation in the Gα-coordinating guanine nucleotide in the ligand-binding pocket (G20A mutation in the Vc2 riboswitch). This
7
Page 7 of 15
us
cr
ip t
rendered the RNA capable of binding both c-di-GMP and c-GAMP with equal affinities[50]. Remarkably ~ 23% of the GEMM motif containing RNA representatives naturally contain an adenine in this position, making it likely that some of these might show discrimination for c-GAMP over c-di-GMP. For example, one candidate from Geobacter metallireducens, Gm0970 showed high selectivity in vitro for c-GAMP with an affinity of 16 nM over 26 μM for c-di-GMP (Figure 1C). Although highly specific, it is clear that this affinity is not as high as the c-di-AMP or c-di-GMP riboswitches, and may hint at higher cellular concentrations for c-GAMP. This class of RNAs (named GEMM 1b) do not bind c-di-AMP or even the 2’,3’ linkage isomer of cGAMP, and are highly selective for 3’,3’cGAMP. Interestingly, in Geobacter cells low amounts of 3’,3’cGAMP have been detected using mass spectrometry even though these cells do not appear to possess a homolog of DncV, the predominant cGAMP synthase. This hints at the possibility of yet to be discovered c-GAMP synthases.
Ac ce pt e
d
M
an
As seen with the other second messenger RNAs, structural studies have revealed binding mechanisms for this RNA class. Crystallographic analysis of the Geobacter RNA Gm1761 reveals a high degree of similarity with the type 1 c-di-GMP riboswitches, with a three helical junction and multiple bulged nucleotides in each of the three helices[51] (Figure 2D). The tuning fork structure has the P1 helix as base and P2 and P3 helices as prongs. As seen in the c-di-GMP RNA, the ligand binds at the junction and is stabilized by extensive base stacking and hydrogen bonding interactions. Adenine nucleotides at the junction show altered C2’endo sugar puckers with one of these adenines being sandwiched between the two bases of the ligand, making intermolecular contacts both with the ligand bases and sugar phosphate groups. The terminal base pairs of the P2 and P1 helices bracket the ligand. While Aα of c-GAMP makes a non-canonical hydrogen bond with A14 and A42 (weak), Gα makes multiple hydrogen bonds through its bases. Although the number of nucleotides in the junction is the same for c-GAMP and c-di-GMP-specific riboswitches, the sequence is not identical. An important difference is the presence of an adenine in place of the Hoogsteen-pairing guanine from c-di-GMP RNAs. This adenine allows recognition of the Aα of c-CAMP and excludes c-di-GMP from binding at this site. C-di-AMP is also excluded from this site since that an adenine replacement in place of the c-GAMP Gβ would disrupt base pairing with the conserved cytosine in the junction. Representatives of this RNA class possess transcriptional terminators distal to the aptamer, the formation of which is modulated by 3’,3’c-GAMP. Disruption of the terminator happens with a half maximum ligand concentration of 10nM 3’,3’cGAMP, and is comparable to the binding affinities seen in vitro. Genes associated with the c-GAMP riboswitch in Geobacter include pgcA, which encodes a periplasmic cytochrome c protein and is highly expressed during growth on insoluble Fe3 oxide versus Fe3 citrate. Two GEMM-1b riboswitches lie in operons with genes encoding pilins and pilus assembly proteins through which direct electron shuttling has been shown to occur. Based on these observations, the GEMM-1b RNA may be implicated
8
Page 8 of 15
in extracellular electron transfer[52]. Some bacteria use 3’,3’ c-GAMP for intestinal colonization in the host. Just like c-di-AMP or c-di-GMP riboswitches, the c-GAMP riboswitches are also associated with diverse gene classes suggesting broad roles for these RNAs and the second messenger, in biology.
us
cr
ip t
While c-GAMP is implicated in host colonization by bacteria, eukaryotes use the same molecule to sense the invasion of pathogens, to mount an innate immunity response. In response to cytosolic DNA, c-GAMP is synthesized in the cytoplasm[53, 54]. The DNA triggers a eukaryotic c-GAMP synthase protein called c-GAS to produce 2’,3’ cGAMP from ATP and GTP. C-GAMP binds to an adapter protein called STING (stimulator of interferon genes), which then facilitates innate immune signaling by activation of IRF3 and interferon-β. Why do bacteria and metazoans use two different stereoisomers of the same c-GAMP molecule? And how does this tie in to host-pathogen relationships? These questions are exciting areas of future work.
M
an
3. Conclusions Second messenger molecules are key determinants in the ability of an organism to adopt distinct lifestyles or forms. In bacteria, messengers derived from cyclicdinucleotides and not mononucleotides appear to play a more prominent role. The discovery of three riboswitch classes with extremely high specificity for each of the three known cyclic dinucleotide second messengers is only the first step in revealing how widespread their roles might be. This also shows the remarkable versatility of RNAs as sensory molecular machines.
Ac ce pt e
d
Acknowledgements A.R. is funded by institutional support from the National Center for Biological Sciences and the Wellcome Trust-Department of Biotechnology India Alliance Intermediate Fellowship. References
[1] Breaker RR. Prospects for riboswitch discovery and analysis. Molecular cell 2011;43:867-79. [2] Nudler E, Mironov AS. The riboswitch control of bacterial metabolism. Trends Biochem Sci 2004;29:11-7. [3] Mellin JR, Cossart P. Unexpected versatility in bacterial riboswitches. Trends in genetics : TIG 2015;31:150-6. [4] Proshkin S, Mironov A, Nudler E. Riboswitches in regulation of Rho-dependent transcription termination. Biochimica et biophysica acta 2014;1839:974-7. [5] DebRoy S, Gebbie M, Ramesh A, Goodson JR, Cruz MR, van Hoof A, et al. Riboswitches. A riboswitch-containing sRNA controls gene expression by sequestration of a response regulator. Science 2014;345:937-40. [6] Cheah MT, Wachter A, Sudarsan N, Breaker RR. Control of alternative RNA splicing and gene expression by eukaryotic riboswitches. Nature 2007;447:497-500.
9
Page 9 of 15
Ac ce pt e
d
M
an
us
cr
ip t
[7] Nahvi A, Barrick JE, Breaker RR. Coenzyme B12 riboswitches are widespread genetic control elements in prokaryotes. Nucleic Acids Res 2004;32:143-50. [8] Winkler W, Nahvi A, Breaker RR. Thiamine derivatives bind messenger RNAs directly to regulate bacterial gene expression. Nature 2002;419:952-6. [9] Winkler WC, Cohen-Chalamish S, Breaker RR. An mRNA structure that controls gene expression by binding FMN. Proc Natl Acad Sci U S A 2002;99:15908-13. [10] Cromie MJ, Shi Y, Latifi T, Groisman EA. An RNA sensor for intracellular Mg(2+). Cell 2006;125:71-84. [11] Dann CE, 3rd, Wakeman CA, Sieling CL, Baker SC, Irnov I, Winkler WC. Structure and mechanism of a metal-sensing regulatory RNA. Cell 2007;130:878-92. [12] Furukawa K, Ramesh A, Zhou Z, Weinberg Z, Vallery T, Winkler WC, et al. Bacterial riboswitches cooperatively bind ni(2+) or co(2+) ions and control expression of heavy metal transporters. Molecular cell 2015;57:1088-98. [13] Baker JL, Sudarsan N, Weinberg Z, Roth A, Stockbridge RB, Breaker RR. Widespread genetic switches and toxicity resistance proteins for fluoride. Science 2012;335:233-5. [14] Kim JN, Breaker RR. Purine sensing by riboswitches. Biology of the cell / under the auspices of the European Cell Biology Organization 2008;100:1-11. [15] Mandal M, Breaker RR. Adenine riboswitches and gene activation by disruption of a transcription terminator. Nat Struct Mol Biol 2004;11:29-35. [16] Ames TD, Breaker RR. Bacterial aptamers that selectively bind glutamine. RNA Biol 2011;8:82-9. [17] Winkler WC, Nahvi A, Sudarsan N, Barrick JE, Breaker RR. An mRNA structure that controls gene expression by binding S-adenosylmethionine. Nat Struct Biol 2003;10:701-7. [18] Mandal M, Lee M, Barrick JE, Weinberg Z, Emilsson GM, Ruzzo WL, et al. A glycine-dependent riboswitch that uses cooperative binding to control gene expression. Science 2004;306:275-9. [19] Winkler WC, Nahvi A, Roth A, Collins JA, Breaker RR. Control of gene expression by a natural metabolite-responsive ribozyme. Nature 2004;428:281-6. [20] Green NJ, Grundy FJ, Henkin TM. The T box mechanism: tRNA as a regulatory molecule. FEBS letters 2010;584:318-24. [21] Yao Z, Weinberg Z, Ruzzo WL. CMfinder--a covariance model based RNA motif finding algorithm. Bioinformatics 2006;22:445-52. [22] Barrick JE, Corbino KA, Winkler WC, Nahvi A, Mandal M, Collins J, et al. New RNA motifs suggest an expanded scope for riboswitches in bacterial genetic control. Proc Natl Acad Sci U S A 2004;101:6421-6. [23] Gardner PP, Daub J, Tate JG, Nawrocki EP, Kolbe DL, Lindgreen S, et al. Rfam: updates to the RNA families database. Nucleic Acids Res 2009;37:D136-40. [24] Nelson JW, Sudarsan N, Furukawa K, Weinberg Z, Wang JX, Breaker RR. Riboswitches in eubacteria sense the second messenger c-di-AMP. Nat Chem Biol 2013;9:834-9. [25] Corrigan RM, Grundling A. Cyclic di-AMP: another second messenger enters the fray. Nature reviews Microbiology 2013;11:513-24. [26] Zhang L, Li W, He ZG. DarR, a TetR-like transcriptional factor, is a cyclic di-AMPresponsive repressor in Mycobacterium smegmatis. J Biol Chem 2013;288:3085-96.
10
Page 10 of 15
Ac ce pt e
d
M
an
us
cr
ip t
[27] Corrigan RM, Campeotto I, Jeganathan T, Roelofs KG, Lee VT, Grundling A. Systematic identification of conserved bacterial c-di-AMP receptor proteins. Proc Natl Acad Sci U S A 2013;110:9084-9. [28] Mehne FM, Gunka K, Eilers H, Herzberg C, Kaever V, Stulke J. Cyclic di-AMP homeostasis in bacillus subtilis: both lack and high level accumulation of the nucleotide are detrimental for cell growth. J Biol Chem 2013;288:2004-17. [29] Woodward JJ, Iavarone AT, Portnoy DA. c-di-AMP secreted by intracellular Listeria monocytogenes activates a host type I interferon response. Science 2010;328:1703-5. [30] Corrigan RM, Abbott JC, Burhenne H, Kaever V, Grundling A. c-di-AMP is a new second messenger in Staphylococcus aureus with a role in controlling cell size and envelope stress. PLoS pathogens 2011;7:e1002217. [31] Romling U. Great times for small molecules: c-di-AMP, a second messenger candidate in Bacteria and Archaea. Science signaling 2008;1:pe39. [32] Ren A, Patel DJ. c-di-AMP binds the ydaO riboswitch in two pseudo-symmetryrelated pockets. Nat Chem Biol 2014;10:780-6. [33] Watson PY, Fedor MJ. The ydaO motif is an ATP-sensing riboswitch in Bacillus subtilis. Nat Chem Biol 2012;8:963-5. [34] Jones CP, Ferre-D'Amare AR. Crystal structure of a c-di-AMP riboswitch reveals an internally pseudo-dimeric RNA. EMBO J 2014;33:2692-703. [35] Gao A, Serganov A. Structural insights into recognition of c-di-AMP by the ydaO riboswitch. Nat Chem Biol 2014;10:787-92. [36] Witte G, Hartung S, Buttner K, Hopfner KP. Structural biochemistry of a bacterial checkpoint protein reveals diadenylate cyclase activity regulated by DNA recombination intermediates. Molecular cell 2008;30:167-78. [37] Block KF, Hammond MC, Breaker RR. Evidence for widespread gene control function by the ydaO riboswitch candidate. J Bacteriol 2010;192:3983-9. [38] Keijser BJ, Ter Beek A, Rauwerda H, Schuren F, Montijn R, van der Spek H, et al. Analysis of temporal gene expression during Bacillus subtilis spore germination and outgrowth. J Bacteriol 2007;189:3624-34. [39] Haiser HJ, Yousef MR, Elliot MA. Cell wall hydrolases affect germination, vegetative growth, and sporulation in Streptomyces coelicolor. J Bacteriol 2009;191:6501-12. [40] Shanahan CA, Strobel SA. The bacterial second messenger c-di-GMP: probing interactions with protein and RNA binding partners using cyclic dinucleotide analogs. Organic & biomolecular chemistry 2012;10:9113-29. [41] Romling U, Gomelsky M, Galperin MY. C-di-GMP: the dawning of a novel bacterial signalling system. Mol Microbiol 2005;57:629-39. [42] Weinberg Z, Barrick JE, Yao Z, Roth A, Kim JN, Gore J, et al. Identification of 22 candidate structured RNAs in bacteria using the CMfinder comparative genomics pipeline. Nucleic Acids Res 2007;35:4809-19. [43] Sudarsan N, Lee ER, Weinberg Z, Moy RH, Kim JN, Link KH, et al. Riboswitches in eubacteria sense the second messenger cyclic di-GMP. Science 2008;321:411-3. [44] Smith KD, Lipchock SV, Ames TD, Wang J, Breaker RR, Strobel SA. Structural basis of ligand binding by a c-di-GMP riboswitch. Nat Struct Mol Biol 2009;16:121823.
11
Page 11 of 15
Ac ce pt e
d
M
an
us
cr
ip t
[45] Smith KD, Shanahan CA, Moore EL, Simon AC, Strobel SA. Structural basis of differential ligand recognition by two classes of bis-(3'-5')-cyclic dimeric guanosine monophosphate-binding riboswitches. Proc Natl Acad Sci U S A 2011;108:7757-62. [46] Smith KD, Strobel SA. Interactions of the c-di-GMP riboswitch with its second messenger ligand. Biochemical Society transactions 2011;39:647-51. [47] Benach J, Swaminathan SS, Tamayo R, Handelman SK, Folta-Stogniew E, Ramos JE, et al. The structural basis of cyclic diguanylate signal transduction by PilZ domains. EMBO J 2007;26:5153-66. [48] Tischler AD, Camilli A. Cyclic diguanylate regulates Vibrio cholerae virulence gene expression. Infect Immun 2005;73:5873-82. [49] Meibom KL, Blokesch M, Dolganov NA, Wu CY, Schoolnik GK. Chitin induces natural competence in Vibrio cholerae. Science 2005;310:1824-7. [50] Kellenberger CA, Wilson SC, Hickey SF, Gonzalez TL, Su Y, Hallberg ZF, et al. GEMM-I riboswitches from Geobacter sense the bacterial second messenger cyclic AMP-GMP. Proc Natl Acad Sci U S A 2015;112:5383-8. [51] Ren A, Wang XC, Kellenberger CA, Rajashankar KR, Jones RA, Hammond MC, et al. Structural Basis for Molecular Discrimination by a 3',3'-cGAMP Sensing Riboswitch. Cell reports 2015;11:1-12. [52] Nelson JW, Sudarsan N, Phillips GE, Stav S, Lunse CE, McCown PJ, et al. Control of bacterial exoelectrogenesis by c-AMP-GMP. Proc Natl Acad Sci U S A 2015;112:5389-94. [53] Wu J, Sun L, Chen X, Du F, Shi H, Chen C, et al. Cyclic GMP-AMP is an endogenous second messenger in innate immune signaling by cytosolic DNA. Science 2013;339:826-30. [54] Sun L, Wu J, Du F, Chen X, Chen ZJ. Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science 2013;339:786-91. FIGURE LEGENDS: Figure 1. A. General mechanism of a metabolite-sensing riboswitch. In low concentrations of metabolite, the aptamer (blue) adopts a structure that allows formation of the downstream transcription terminator thus turning off the downstream gene. Under high metabolite conditions an altered conformation of the aptamer sequesters a portion of the downstream terminator to form an antiterminator that results in gene expression. B. Structural representations of second messengers c-di-AMP, c-di-GMP, 3’,3’ c-GAMP and 2’,3’ c-GAMP. C. Riboswitch classes with their ligands and binding affinities. Figure 2. Architectures of second messenger-binding riboswitches. A. The c-di-AMP binding riboswitch binds 2 molecules of ligand (magenta) at the junction of collinearly stacked helices. The inherent pseudo-symmetry is seen both within and between the two ligand binding sites. The P1 helix (cyan) is stabilized by multiple interactions in the ligand bound state of the RNA. B. The Type I c-di-GMP riboswitch forms a Y-shaped structure with P1 (cyan) at the base and P2 and P3 as parallel helices. The ligand (magenta) is bound at the junction of the helices and is stabilized by hydrogen bonding and stacking interactions. C. The Type II c-di-GMP riboswitch is distinct from Type I in that it harbors a kink-turn motif and a pseudoknot, both of
12
Page 12 of 15
Ac ce pt e
d
M
an
us
cr
ip t
which help stabilize ligand (magenta) in the binding pocket. D. The 3’,3’ c-GAMP riboswitch resembles the global architecture of the Type I c-di-GMP RNA (panel B) but lacks the conserved nucleotides required for recognition of two guanine bases from the ligand and instead preferentially binds 3’,3’ c-GAMP (magenta) to stabilize formation of the P1 helix (cyan).
13
Page 13 of 15
Page 14 of 15
ed
pt
ce
Ac M
an
cr
us
ip
Page 15 of 15
ed
ce pt
Ac M an
cr
us
ip t
S1084-9521(15)00216-5 http://dx.doi.org/doi:10.1016/j.semcdb.2015.10.019 YSCDB 1849
To appear in:
Seminars in Cell & Developmental Biology
Please cite this article as: Ramesh A, Second messenger-sensing riboswitches in bacteria, Seminars in Cell and Developmental Biology (2015), http://dx.doi.org/10.1016/j.semcdb.2015.10.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Second messenger-sensing riboswitches in bacteria Arati Ramesh* *National Center for Biological Sciences, GKVK Campus, Bellary Road, Bangalore, India- 560065 *Corresponding author. Email address [email protected] Tel: 91-80-2366 6930, Fax: 91 80 23636662
an
us
cr
ip t
Abstract Signal sensing in bacteria has traditionally been attributed to protein-based factors. It is however becoming increasingly clear that bacteria also exploit RNAs to serve this role. This review discusses how key developmental processes in bacteria, such as community formation, choice of a sessile versus motile lifestyle, or vegetative growth versus dormant spore formation may be governed by signal sensing RNAs. The signaling molecules that affect these processes, the RNAs that sense these molecules and the underlying molecular basis for specific signal-response are discussed here. keywords: riboswitch, second messenger, sporulation, biofilms, differentiation
M
Abbreviations: c-di-GMP, 3’,5’-cyclic-di-guanosine monophosphate; c-di-AMP, 3’,5’cyclic-di -Adenosine monophosphate; 3’,3’ c-GAMP, 3’,3’ cyclic guanosine monophosphate adenosine monophosphate; 2′,3′-c-GAMP, 2’,3’ cyclic guanosine monophosphate adenosine monophosphate; 5’ UTR, 5’ Untranslated Regions.
Ac ce pt e
d
1. Introduction: Riboswitches as signal sensors in bacteria Bacteria sense and respond to their environment. This response occurs not only through protein-based sensors but also via RNAs that directly sense signaling molecules and in turn control gene expression. Such signal sensing RNAs, called riboswitches, typically reside in the 5’ UTRs of mRNA, in a cis configuration with respect to the gene that they control. Riboswitches are characterized by an aptamer domain that directly binds a metabolite, and an expression platform that transmits metabolite-induced effects to the downstream gene[1-3] (Figure 1A). Metabolite binding may turn on or off the downstream gene in one of many ways; by modulating a transcriptional terminator to affect genes at the post transcriptioninitiation level, by modulating the ribosome binding site to control genes at the level of translation initiation, or in rare cases control mRNA stability. More recently it has been shown that a class of riboswitches can coordinate with proteins such as the Rho transcription termination factor to elicit gene regulation[4]. Riboswitches have also been shown to control the expression of other non-coding RNAs[5], suggesting a complex network of RNA-based regulation in bacteria. In eukaryotes, the extent of riboswitch-based regulation is as yet unclear, though a class of thymine pyrophosphate sensing riboswitches has been shown to function via the control of mRNA splicing[6]. Based on the types of metabolites recognized, over 20 families of riboswitches have been identified thus far. Ligands sensed by riboswitches range from complex 1
Page 1 of 15
an
us
cr
ip t
metabolites such as vitamins (cobalamine[7], thymine pyrophosphate[8], flavin mononucleotide[9]) and cofactors (S-adenosyl methionine, S-adenosyl homocysteine), to small ions such as cations[10-12] (Mg2+, Ni2+/Co2+, Mn2+) and anions[13] (F-). RNAs have also been shown to sense nucleotides[14] (adenine[15], guanine, 2'-deoxyguanosine and the guanine analog PreQ1), amino acids (glutamine[16], lysine[17], glycine[18]), sugars[19] (glucosamine 6-phosphate) and charged tRNAs[20]. These classes of riboswitches were identified first through computational methods that predict conserved, structured classes of RNAs in sequenced genomes. The genomic contexts of the predicted RNAs have since then been used to rationally predict and test metabolites that are likely to be sensed by the RNA[21]. The majority of riboswitches appear to control genes encoding transport proteins or involved in the synthesis of metabolites. However, some classes of RNAs are associated with such a diverse collection of genes that it has been particularly challenging to identify their cognate ligand simply based on the genomic context of the RNA. This has been especially so for RNAs that sense molecules influencing developmental processes in bacteria, such as biofilm formation or sporulation since there is no single class of genes that control these processes. These riboswitch classes and the metabolites affecting these developmental processes are the primary focus of this review.
Ac ce pt e
d
M
2. Riboswitches that sense second messengers Bacteria respond to their environment by sensing signals, whose transduction eventually leads to developmental, physiological and behavioral changes in the bacterial cell. Signaling molecules thus play a central role in the manifestation of a variety of developmental processes in bacteria. While signal receptors form the very first layer of sensing, small molecule second messengers then carry signals from receptors to downstream effectors, which could be protein or DNA targets. In eukaryotes, nucleotide-based second messengers such as cyclic AMP and cyclic GMP are commonly used as signaling molecules. In bacteria, cyclic AMP and another intracellular molecule called guanosine tetra or penta phosphate (ppGpp) are more commonly used, though these are restricted to the response to nutrient limitation and carbon metabolism. More recently, the roles of novel second messengers, which are often dinucleotide molecules such as cyclic di-AMP, cyclic di-GMP and the recently discovered cyclic-AMP-GMP, have gained prominence. These molecules are circularly fused nucleotides where two molecules of AMP or GMP or both, form 3’,5’bonds (Figure 1B). Distributions of these purine derived cyclic nucleotides vary across organisms. Some like Cyclic-di-GMP are present in bacteria but have not been detected in eukaryotes thus far. Others like cGAMP have been detected across organisms, but are present as different stereoisomers. While the 3’,3’ linkage isomer of cGAMP has been detected in bacteria, eukaryotes show the presence of the 2’,3’ linkage isomer of cGAMP (Figure 1B). Although the biology around these molecules is not yet fully understood, it is clear that they affect myriad aspects of developmental processes. Recently, three classes of signal-sensing RNAs have been discovered that recognize bacterial second messengers with incredibly high affinity and specificity while discriminating against other similar molecules in the cell.
2
Page 2 of 15
These discoveries substantially expand the role and scope of riboswitches in bacterial biology.
us
cr
ip t
2. 1. The c-di-AMP riboswitch The ydaO/yuaA RNA was first predicted computationally over a decade ago[22]. Since then, over 3000 representatives of this RNA have been found in bacterial genomes[23], upstream of genes involved in cell wall metabolism, osmotic stress and sporulation. However, the sheer diversity of genes associated with the RNA long prevented its cognate metabolite from being identified. Recently, using a biochemical pull-down approach, a metabolite from yeast extract was shown to interact with the RNA[24]. Mass spec analysis identified this metabolite as c-di-AMP, a cyclic dinucleotide made of two adenosines joined with a 3’,5’ phosphodiester linkage.
Ac ce pt e
d
M
an
The biosynthesis of c-di-AMP itself involves the joining of 2 ATP molecules by proteins containing- diadenylyl cyclase domains (DAC)[25]. These proteins bear little homology in sequence or structure to the c-di-GMP or c-GAMP synthesizing enzymes. C-di-AMP is involved in regulating a range of diverse processes in bacteria, and is implicated in cell wall homeostasis, fatty acid synthesis in mycobacterial species[26], growth of Staphylococcus aureus in low K+ conditions[27] and DNA integrity in Bacillus subtilis[28]. In S. aureus high c-di-AMP levels are also linked to decreased cell size. In B. subtilis and Listeria monocytogenes decrease in c-di-AMP corresponds to higher sensitivity to cell wall targeting antibiotics. Listeria in the cytosol of the host secretes c-di-AMP, which is sensed by the host and used as a cue to mount a type I IFN response, activating defense mechanisms in the host[29]. Although DAC domain proteins are widely present in bacterial genomes, c-di-AMP itself has been experimentally detected only in some bacteria, with intracellular concentrations estimated to be in the 1.7 to 2 μM range in B. subtilis and S. aureus[30]. In addition to bacteria, DAC domain-containing proteins have been found in archaea, suggesting a possible role for c-di-AMP in archaea[31]. With the identification of c-di-AMP RNAs and the genes they control, we can now perhaps learn more about the roles of c-di-AMP itself in bacterial developmental processes. The 165 nucleotide ydaO riboswitch from Bacillus subtilis showed a strong binding affinity of 10 pM and an estimated stoichiometry of 1:1 for c-di-AMP[24] (Figure 1C). This stoichiometry and affinity however is not conserved within the ydaO riboswitch class since some ydaO homologs from other bacteria have been found to bind 2 molecules of c-di-AMP per RNA and display weaker binding affinities[32]. This could indicate diverse structural and functional modes for c-di-AMP recognition within the same riboswitch class, in tune with the needs of the particular species of bacteria. Strikingly, molecules related to c-di-AMP such as isoforms of AMP, ADP, cAMP, c-di-IMP, c-di-GMP and a deoxy-derivative c-di-dAMP are strongly discriminated against by this class of RNAs, suggesting that these RNAs are molecular machines of exquisite specificity. Paradoxically, the ydaO RNA binds ATP in vitro with an affinity of 0.6 mM[33]. This falls within the range of intracellular ATP concentrations and it been proposed that this RNA might sense 3
Page 3 of 15
ip t
ATP in the cell. However the significantly higher affinity for c-di-AMP and genes associated with the RNA class indicate that the cellular signal being sensed is likely to be c-di-AMP. The intracellular concentrations of c-di-AMP have been estimated at ~2 uM in the cytoplasm of B. subtilis and S. aureus. This is orders of magnitude higher than the picomolar binding affinity of the riboswitch, strongly suggesting that the riboswitch function in this case is driven by the kinetics of the system rather than thermodynamics wherein the RNA would function at equilibrium with cellular c-di-AMP.
M
an
us
cr
High-resolution crystal structures have been solved for three ydaO aptamers from different bacteria [32, 34, 35]. The overall architecture of the RNA comprises a fivehelical arrangement that contains a zippered up bubble and a pseudoknot that stabilizes long-range contacts in the RNA (Figure 2A). Two c-di-AMP molecules are bound at sites lined with conserved nucleotides, suggesting that these are distinct and functionally important ligand binding sites. One ligand molecule binds at the collinearly arranged P4-P5 helices and pseudoknot end of the molecule while the other binds at the collinearly arranged P3-P2 helices and the zippered bubble region. Distances between coordinating RNA functional groups are organized in a way that would exclude the amino group of guanine nucleotide derivatives, explaining why the c-di-AMP riboswitch does not interact with c-di-GMP or c-GAMP molecules.
Ac ce pt e
d
C-di-AMP assumes an extended conformation with Aα and Aβ wedged between nucleotide bases of the RNA, likely stabilizing junction regions of the RNA[32, 34, 35]. Both within and between the two ligand-binding sites, there is pseudosymmetry, a rather unique feature of the RNA wherein two near identical sites are created for ligand binding. Whether or not the two sites interact cooperatively is yet to be determined and would be relevant in the context of intracellular c-di-AMP concentrations. The principles of c-di-AMP recognition appear largely conserved between c-di-AMP binding proteins and RNAs. In crystal structures of both the DisA protein[36] and the ydaO riboswitch[32, 35], the ligand contacts the macromolecule with two adenine bases and the sugar rings of the ligand are puckered as a 3’-endo configuration. The sugar, phosphate as well as nucleobases of the ligand make extensive hydrogen bonding contacts with the protein and the riboswitch. This therefore appears to be an illustrative example of proteins and RNAs evolving similar binding sites for an important biomolecule. The premise of riboswitch function relies on an expression platform being modulated by the metabolite-bound aptamer. The c-di-AMP bound structure of the RNA shows a P1 helix platform that abrogates an antiterminator, such that in the presence of c-di-AMP, an intrinsic transcription terminator stem would form distal to the riboswitch. The absence of c-di-AMP would allow disruption of the P1 helix, thereby forming an antiterminator structure that results in expression of the downstream gene. This agrees with in vitro and in vivo observations for the B. subtilis ydaO RNA, wherein the presence of c-di-AMP results in a shorter transcript and decreased expression of a reporter gene fusion[24]. In S. coelicolor, instead of a 4
Page 4 of 15
transcriptional terminator platform, a ribosome-binding site overlaps with the pseudoknot, suggesting that c-di-AMP binding may regulate translation and not transcription in some cases[37].
an
us
cr
ip t
The majority of ydaO RNAs associate with genes that encode proteins involved in cell wall metabolism. This is followed by proteins for the transport and synthesis of osmoprotectants[37] and genes are involved in peptidoglycan breakdown and remodeling. Another commonly associated set of genes encodes transporters of potassium or other solutes implicated in cellular osmosis. In B. subtilis, transcription of the RNA associated genes ydaO and ktrAB shows upregulation ~10minutes after spore germination[38] where as in Streptomyces coelicolor the cell wall hydrolase enzymes associated with this riboswitch appear to be expressed throughout development[39]. The distribution of this riboswitch class strongly correlates with genes encoding diadenylate cyclase proteins that are involved in the synthesis of cdi-AMP. In Staphylococcus aureus, genes downstream of ydaO often encode proteins that bind c-di-AMP. Many of these processes are integral to bacterial sporulation and germination in a range of diverse bacteria.
Ac ce pt e
d
M
2.2. The c-di-GMP riboswitch Discovered in the 1980s, c-di-GMP is now recognized to be a key determinant of developmental decisions in bacteria such as switching between motile and sessile biofilm states. [40]. C-di-GMP binding proteins such as the GGDEF domain (Pfam:PF00990)-containing diguanylate cyclases and the EAL domain (Pfam:PF00563)-containing phosphodiesterases together maintain the intracellular concentration of c-di-GMP. These proteins were initially linked to biofilm matrix formation, rugose colony morphology, virulence, twitching motility etc., suggesting a link between c-di-GMP and these various biological processes. High intracellular cdi-GMP has since been shown to promote sessile biofilm lifestyle whereas low c-diGMP promotes motile planktonic lifestyle. Other processes such as exopolysaccharide production, flagellum biosynthesis, pilus assembly, cell-cell communication, quorum sensing, expression of virulence factors etc. have been linked to c-di-GMP levels suggesting a very broad role for this second messenger[40, 41]. RNAs that sense c-di-GMP were identified upstream of genes encoding diguanylate cyclases and phosphodiesterases and other c-di-GMP related proteins. These RNAs contain the GEMM (Genes for the Environment, Membranes and Motility) motif[42], now known to occur in ~545 RNA representatives that are further classified as type I (500 RNAs) and type 2 (45 RNAs) based on differences in RNA structural elements. Unlike the unbiased approach leading to the discovery of c-di-AMP as the ligand for the ydaO riboswitches, in the case of GEMM motifs, strong gene associations with diguanylate cyclases and phosphodiesterases led to the finding that this RNA recognizes the bacterial second messenger c-di-GMP[43].
5
Page 5 of 15
us
cr
ip t
In vitro binding analysis of the c-di-GMP riboswitches showed a dissociation constant for c-di-GMP between 10 pM to 1 nM (Figure 1C, estimated using different methods) for type 1 and 200 pM to 2 nM for type 2 RNAs with a stoichiometry of 1:1[43]. Analogs of c-di-GMP and breakdown products of c-di-GMP such as pGpG are discriminated against by a difference in affinity greater than 3-fold for the type 1 RNA, suggesting high specificity for ligand recognition similar to that seen for c-diAMP riboswitches. Interestingly, the GEMM motif RNAs bind c-di-GMP with much higher affinity than other c-di-GMP binding proteins (EAL domains, GGDEF domains etc.) that typically display affinities only in the range of 50 nM to low μM. Detailed kinetic analyses of this RNA reveal that the half life of the ligand bound to RNA is in the order of a month, suggesting that within biological contexts this complex is essentially irreversible[44], committing the bacteria to a specific regulatory response.
Ac ce pt e
d
M
an
Given that cyclic dinucleotides are structurally and chemically similar, how then do RNAs build in specificity in their binding sites? Structural studies have dissected the molecular mechanisms of c-di-GMP binding to this class of riboswitches in detail, revealing this distinction. Crystal structures of the type 1 riboswitch show a Yshaped architecture where three helices P1, P2 and P3 come together to form a junction that harbors the ligand-binding site[44] (Figure 2B). Tetraloop-receptor interactions between P2 and P3 hold them in a parallel orientation to each other. This region is highly stabilized through multiple interactions and does not undergo local ligand-induced changes in secondary structure. Unlike the c-di-AMP riboswitch, here the ligand is recognized asymmetrically wherein Gβ of the ligand makes Watson Crick contacts with a conserved cytidine, while Gα is stabilized by Hoogsteen interactions. The conserved cytidine prevents an adenosine from binding the RNA, thus excluding c-di-AMP from being recognized by this RNA. Interactions occur with the base as well as the sugar phosphate backbone of Gβ while the Watson-Crick surface of Gα is not recognized by the RNA and lies in a solvent accessible cavity at the helical junction. In addition to hydrogen bonding, intercalated base stacking and metal ion-mediated interactions with the phosphate group of the ligand further stabilizes the ligand in the binding pocket. A similar theme is observed in the thymine pyrophosphate riboswitch wherein the negatively charged functional group of the ligand is coordinated by a metal ion in order to facilitate RNA binding. Mutations made to the binding pocket nucleotides (C to U at the Gβ interaction and G to A mutation at the Gα interaction) resulted in RNAs that can select for c-di-AMP over c-di-GMP by over 4-fold. However, naturally occurring c-di-AMP riboswitches appear to have architectures and binding site configurations that are unique from c-di-GMP sensing RNAs[32, 34, 35, 44]. C-di-GMP sensing riboswitches show two unique architectures capable of binding this molecule. Crystallographic studies of the Clostridium acetobutylicum type 2 aptamer bound to c-di-GMP, reveal a tertiary architecture distinct from type I RNAs[45, 46] (Figure 2C). In addition to the three helices seen in type I c-di-GMP riboswitches, type 2 RNAs harbor a kink-turn motif and a pseudoknot whose
6
Page 6 of 15
us
cr
ip t
nucleotides are tightly conserved and show ligand-induced changes in secondary structure. The ligand is sequestered into a triplex at the edge of a pseudoknot. Although both type 1 and type 2 RNAs sequester the ligand in asymmetric binding sites, type 2 RNAs do not make Watson-Crick or Hoogsteen contacts with the Gβ and lack the critical, conserved cytidine present in type I. Also, unlike the type I riboswitches, only few contacts are made with the ribosyl phosphate backbone of the ligand, providing a possible explanation for the weaker binding affinities seen for type 2 RNAs. Distribution analysis of this RNA class across bacteria shows that although some bacteria have both type 1 and 2 switches, the majority seems to prefer one type over the other with type 2 being more prevalent in Clostridia[46]. Since these two types of riboswitches do not commonly co-occur in bacteria, this suggests variations in intracellular c-di-GMP concentrations, which may in turn govern processes differently.
d
M
an
Proteins such as PilZ also bind c-di-GMP asymmetrically and make sidechain stacking interactions with the Gα and Gβ of c-di-GMP[47]. However, these interactions are unlikely to be as stabilizing as those with RNA where the ligand is incorporated into the core of the RNA molecule. Thus the conformation of the ligand itself is different when bound to protein partners versus RNAs and would explain the significantly higher binding affinity seen for RNAs compared to proteins. Binding of the ligand likely nucleates formation of the P1 helix expression platform formed by the 5’ and 3’ ends of the aptamer, the sequence of which is not strictly conserved among representatives of this RNA class. The GEMM motif is associated downstream with both transcriptional terminators as well as translation controlling sites, suggesting that this RNA likely controls genes by common riboswitch mechanisms.
Ac ce pt e
Besides guanylate cyclase genes that are involved in c-di-GMP biosynthesis, and phosphodiesterases that are involved in c-di-GMP breakdown, the GEMM motif RNAs also associate with flagellum biosynthesis genes in Vibrio cholera. Here c-diGMP is linked to virulence wherein the production of cholera toxin is upregulated under conditions of low c-di-GMP concentrations[48]. This suggests that in some cases, riboswitch based sensing is coupled to organelle biosynthesis and lifestyle choices such as a switch between motile and sessile forms. The Vc2 riboswitch of Vibrio cholera is associated with a transcription regulator protein (VC1722) that is itself known to control pathways involved in competence[49]. The c-di-GMP riboswitches also control genes for cell differentiation, morphology, cell cell communication, exopolysaccharide production, and are linked to quorum sensing[40]. 2.3. The c-GAMP riboswitch The most recent member to join the family of cyclic dinucleotide messengers is cyclic GMP-AMP (c-GAMP). The discovery of the c-GAMP riboswitches was itself serendipitous. Analysis of GEMM domain containing type I c-di-GMP riboswitches revealed the presence of a single point mutation in the Gα-coordinating guanine nucleotide in the ligand-binding pocket (G20A mutation in the Vc2 riboswitch). This
7
Page 7 of 15
us
cr
ip t
rendered the RNA capable of binding both c-di-GMP and c-GAMP with equal affinities[50]. Remarkably ~ 23% of the GEMM motif containing RNA representatives naturally contain an adenine in this position, making it likely that some of these might show discrimination for c-GAMP over c-di-GMP. For example, one candidate from Geobacter metallireducens, Gm0970 showed high selectivity in vitro for c-GAMP with an affinity of 16 nM over 26 μM for c-di-GMP (Figure 1C). Although highly specific, it is clear that this affinity is not as high as the c-di-AMP or c-di-GMP riboswitches, and may hint at higher cellular concentrations for c-GAMP. This class of RNAs (named GEMM 1b) do not bind c-di-AMP or even the 2’,3’ linkage isomer of cGAMP, and are highly selective for 3’,3’cGAMP. Interestingly, in Geobacter cells low amounts of 3’,3’cGAMP have been detected using mass spectrometry even though these cells do not appear to possess a homolog of DncV, the predominant cGAMP synthase. This hints at the possibility of yet to be discovered c-GAMP synthases.
Ac ce pt e
d
M
an
As seen with the other second messenger RNAs, structural studies have revealed binding mechanisms for this RNA class. Crystallographic analysis of the Geobacter RNA Gm1761 reveals a high degree of similarity with the type 1 c-di-GMP riboswitches, with a three helical junction and multiple bulged nucleotides in each of the three helices[51] (Figure 2D). The tuning fork structure has the P1 helix as base and P2 and P3 helices as prongs. As seen in the c-di-GMP RNA, the ligand binds at the junction and is stabilized by extensive base stacking and hydrogen bonding interactions. Adenine nucleotides at the junction show altered C2’endo sugar puckers with one of these adenines being sandwiched between the two bases of the ligand, making intermolecular contacts both with the ligand bases and sugar phosphate groups. The terminal base pairs of the P2 and P1 helices bracket the ligand. While Aα of c-GAMP makes a non-canonical hydrogen bond with A14 and A42 (weak), Gα makes multiple hydrogen bonds through its bases. Although the number of nucleotides in the junction is the same for c-GAMP and c-di-GMP-specific riboswitches, the sequence is not identical. An important difference is the presence of an adenine in place of the Hoogsteen-pairing guanine from c-di-GMP RNAs. This adenine allows recognition of the Aα of c-CAMP and excludes c-di-GMP from binding at this site. C-di-AMP is also excluded from this site since that an adenine replacement in place of the c-GAMP Gβ would disrupt base pairing with the conserved cytosine in the junction. Representatives of this RNA class possess transcriptional terminators distal to the aptamer, the formation of which is modulated by 3’,3’c-GAMP. Disruption of the terminator happens with a half maximum ligand concentration of 10nM 3’,3’cGAMP, and is comparable to the binding affinities seen in vitro. Genes associated with the c-GAMP riboswitch in Geobacter include pgcA, which encodes a periplasmic cytochrome c protein and is highly expressed during growth on insoluble Fe3 oxide versus Fe3 citrate. Two GEMM-1b riboswitches lie in operons with genes encoding pilins and pilus assembly proteins through which direct electron shuttling has been shown to occur. Based on these observations, the GEMM-1b RNA may be implicated
8
Page 8 of 15
in extracellular electron transfer[52]. Some bacteria use 3’,3’ c-GAMP for intestinal colonization in the host. Just like c-di-AMP or c-di-GMP riboswitches, the c-GAMP riboswitches are also associated with diverse gene classes suggesting broad roles for these RNAs and the second messenger, in biology.
us
cr
ip t
While c-GAMP is implicated in host colonization by bacteria, eukaryotes use the same molecule to sense the invasion of pathogens, to mount an innate immunity response. In response to cytosolic DNA, c-GAMP is synthesized in the cytoplasm[53, 54]. The DNA triggers a eukaryotic c-GAMP synthase protein called c-GAS to produce 2’,3’ cGAMP from ATP and GTP. C-GAMP binds to an adapter protein called STING (stimulator of interferon genes), which then facilitates innate immune signaling by activation of IRF3 and interferon-β. Why do bacteria and metazoans use two different stereoisomers of the same c-GAMP molecule? And how does this tie in to host-pathogen relationships? These questions are exciting areas of future work.
M
an
3. Conclusions Second messenger molecules are key determinants in the ability of an organism to adopt distinct lifestyles or forms. In bacteria, messengers derived from cyclicdinucleotides and not mononucleotides appear to play a more prominent role. The discovery of three riboswitch classes with extremely high specificity for each of the three known cyclic dinucleotide second messengers is only the first step in revealing how widespread their roles might be. This also shows the remarkable versatility of RNAs as sensory molecular machines.
Ac ce pt e
d
Acknowledgements A.R. is funded by institutional support from the National Center for Biological Sciences and the Wellcome Trust-Department of Biotechnology India Alliance Intermediate Fellowship. References
[1] Breaker RR. Prospects for riboswitch discovery and analysis. Molecular cell 2011;43:867-79. [2] Nudler E, Mironov AS. The riboswitch control of bacterial metabolism. Trends Biochem Sci 2004;29:11-7. [3] Mellin JR, Cossart P. Unexpected versatility in bacterial riboswitches. Trends in genetics : TIG 2015;31:150-6. [4] Proshkin S, Mironov A, Nudler E. Riboswitches in regulation of Rho-dependent transcription termination. Biochimica et biophysica acta 2014;1839:974-7. [5] DebRoy S, Gebbie M, Ramesh A, Goodson JR, Cruz MR, van Hoof A, et al. Riboswitches. A riboswitch-containing sRNA controls gene expression by sequestration of a response regulator. Science 2014;345:937-40. [6] Cheah MT, Wachter A, Sudarsan N, Breaker RR. Control of alternative RNA splicing and gene expression by eukaryotic riboswitches. Nature 2007;447:497-500.
9
Page 9 of 15
Ac ce pt e
d
M
an
us
cr
ip t
[7] Nahvi A, Barrick JE, Breaker RR. Coenzyme B12 riboswitches are widespread genetic control elements in prokaryotes. Nucleic Acids Res 2004;32:143-50. [8] Winkler W, Nahvi A, Breaker RR. Thiamine derivatives bind messenger RNAs directly to regulate bacterial gene expression. Nature 2002;419:952-6. [9] Winkler WC, Cohen-Chalamish S, Breaker RR. An mRNA structure that controls gene expression by binding FMN. Proc Natl Acad Sci U S A 2002;99:15908-13. [10] Cromie MJ, Shi Y, Latifi T, Groisman EA. An RNA sensor for intracellular Mg(2+). Cell 2006;125:71-84. [11] Dann CE, 3rd, Wakeman CA, Sieling CL, Baker SC, Irnov I, Winkler WC. Structure and mechanism of a metal-sensing regulatory RNA. Cell 2007;130:878-92. [12] Furukawa K, Ramesh A, Zhou Z, Weinberg Z, Vallery T, Winkler WC, et al. Bacterial riboswitches cooperatively bind ni(2+) or co(2+) ions and control expression of heavy metal transporters. Molecular cell 2015;57:1088-98. [13] Baker JL, Sudarsan N, Weinberg Z, Roth A, Stockbridge RB, Breaker RR. Widespread genetic switches and toxicity resistance proteins for fluoride. Science 2012;335:233-5. [14] Kim JN, Breaker RR. Purine sensing by riboswitches. Biology of the cell / under the auspices of the European Cell Biology Organization 2008;100:1-11. [15] Mandal M, Breaker RR. Adenine riboswitches and gene activation by disruption of a transcription terminator. Nat Struct Mol Biol 2004;11:29-35. [16] Ames TD, Breaker RR. Bacterial aptamers that selectively bind glutamine. RNA Biol 2011;8:82-9. [17] Winkler WC, Nahvi A, Sudarsan N, Barrick JE, Breaker RR. An mRNA structure that controls gene expression by binding S-adenosylmethionine. Nat Struct Biol 2003;10:701-7. [18] Mandal M, Lee M, Barrick JE, Weinberg Z, Emilsson GM, Ruzzo WL, et al. A glycine-dependent riboswitch that uses cooperative binding to control gene expression. Science 2004;306:275-9. [19] Winkler WC, Nahvi A, Roth A, Collins JA, Breaker RR. Control of gene expression by a natural metabolite-responsive ribozyme. Nature 2004;428:281-6. [20] Green NJ, Grundy FJ, Henkin TM. The T box mechanism: tRNA as a regulatory molecule. FEBS letters 2010;584:318-24. [21] Yao Z, Weinberg Z, Ruzzo WL. CMfinder--a covariance model based RNA motif finding algorithm. Bioinformatics 2006;22:445-52. [22] Barrick JE, Corbino KA, Winkler WC, Nahvi A, Mandal M, Collins J, et al. New RNA motifs suggest an expanded scope for riboswitches in bacterial genetic control. Proc Natl Acad Sci U S A 2004;101:6421-6. [23] Gardner PP, Daub J, Tate JG, Nawrocki EP, Kolbe DL, Lindgreen S, et al. Rfam: updates to the RNA families database. Nucleic Acids Res 2009;37:D136-40. [24] Nelson JW, Sudarsan N, Furukawa K, Weinberg Z, Wang JX, Breaker RR. Riboswitches in eubacteria sense the second messenger c-di-AMP. Nat Chem Biol 2013;9:834-9. [25] Corrigan RM, Grundling A. Cyclic di-AMP: another second messenger enters the fray. Nature reviews Microbiology 2013;11:513-24. [26] Zhang L, Li W, He ZG. DarR, a TetR-like transcriptional factor, is a cyclic di-AMPresponsive repressor in Mycobacterium smegmatis. J Biol Chem 2013;288:3085-96.
10
Page 10 of 15
Ac ce pt e
d
M
an
us
cr
ip t
[27] Corrigan RM, Campeotto I, Jeganathan T, Roelofs KG, Lee VT, Grundling A. Systematic identification of conserved bacterial c-di-AMP receptor proteins. Proc Natl Acad Sci U S A 2013;110:9084-9. [28] Mehne FM, Gunka K, Eilers H, Herzberg C, Kaever V, Stulke J. Cyclic di-AMP homeostasis in bacillus subtilis: both lack and high level accumulation of the nucleotide are detrimental for cell growth. J Biol Chem 2013;288:2004-17. [29] Woodward JJ, Iavarone AT, Portnoy DA. c-di-AMP secreted by intracellular Listeria monocytogenes activates a host type I interferon response. Science 2010;328:1703-5. [30] Corrigan RM, Abbott JC, Burhenne H, Kaever V, Grundling A. c-di-AMP is a new second messenger in Staphylococcus aureus with a role in controlling cell size and envelope stress. PLoS pathogens 2011;7:e1002217. [31] Romling U. Great times for small molecules: c-di-AMP, a second messenger candidate in Bacteria and Archaea. Science signaling 2008;1:pe39. [32] Ren A, Patel DJ. c-di-AMP binds the ydaO riboswitch in two pseudo-symmetryrelated pockets. Nat Chem Biol 2014;10:780-6. [33] Watson PY, Fedor MJ. The ydaO motif is an ATP-sensing riboswitch in Bacillus subtilis. Nat Chem Biol 2012;8:963-5. [34] Jones CP, Ferre-D'Amare AR. Crystal structure of a c-di-AMP riboswitch reveals an internally pseudo-dimeric RNA. EMBO J 2014;33:2692-703. [35] Gao A, Serganov A. Structural insights into recognition of c-di-AMP by the ydaO riboswitch. Nat Chem Biol 2014;10:787-92. [36] Witte G, Hartung S, Buttner K, Hopfner KP. Structural biochemistry of a bacterial checkpoint protein reveals diadenylate cyclase activity regulated by DNA recombination intermediates. Molecular cell 2008;30:167-78. [37] Block KF, Hammond MC, Breaker RR. Evidence for widespread gene control function by the ydaO riboswitch candidate. J Bacteriol 2010;192:3983-9. [38] Keijser BJ, Ter Beek A, Rauwerda H, Schuren F, Montijn R, van der Spek H, et al. Analysis of temporal gene expression during Bacillus subtilis spore germination and outgrowth. J Bacteriol 2007;189:3624-34. [39] Haiser HJ, Yousef MR, Elliot MA. Cell wall hydrolases affect germination, vegetative growth, and sporulation in Streptomyces coelicolor. J Bacteriol 2009;191:6501-12. [40] Shanahan CA, Strobel SA. The bacterial second messenger c-di-GMP: probing interactions with protein and RNA binding partners using cyclic dinucleotide analogs. Organic & biomolecular chemistry 2012;10:9113-29. [41] Romling U, Gomelsky M, Galperin MY. C-di-GMP: the dawning of a novel bacterial signalling system. Mol Microbiol 2005;57:629-39. [42] Weinberg Z, Barrick JE, Yao Z, Roth A, Kim JN, Gore J, et al. Identification of 22 candidate structured RNAs in bacteria using the CMfinder comparative genomics pipeline. Nucleic Acids Res 2007;35:4809-19. [43] Sudarsan N, Lee ER, Weinberg Z, Moy RH, Kim JN, Link KH, et al. Riboswitches in eubacteria sense the second messenger cyclic di-GMP. Science 2008;321:411-3. [44] Smith KD, Lipchock SV, Ames TD, Wang J, Breaker RR, Strobel SA. Structural basis of ligand binding by a c-di-GMP riboswitch. Nat Struct Mol Biol 2009;16:121823.
11
Page 11 of 15
Ac ce pt e
d
M
an
us
cr
ip t
[45] Smith KD, Shanahan CA, Moore EL, Simon AC, Strobel SA. Structural basis of differential ligand recognition by two classes of bis-(3'-5')-cyclic dimeric guanosine monophosphate-binding riboswitches. Proc Natl Acad Sci U S A 2011;108:7757-62. [46] Smith KD, Strobel SA. Interactions of the c-di-GMP riboswitch with its second messenger ligand. Biochemical Society transactions 2011;39:647-51. [47] Benach J, Swaminathan SS, Tamayo R, Handelman SK, Folta-Stogniew E, Ramos JE, et al. The structural basis of cyclic diguanylate signal transduction by PilZ domains. EMBO J 2007;26:5153-66. [48] Tischler AD, Camilli A. Cyclic diguanylate regulates Vibrio cholerae virulence gene expression. Infect Immun 2005;73:5873-82. [49] Meibom KL, Blokesch M, Dolganov NA, Wu CY, Schoolnik GK. Chitin induces natural competence in Vibrio cholerae. Science 2005;310:1824-7. [50] Kellenberger CA, Wilson SC, Hickey SF, Gonzalez TL, Su Y, Hallberg ZF, et al. GEMM-I riboswitches from Geobacter sense the bacterial second messenger cyclic AMP-GMP. Proc Natl Acad Sci U S A 2015;112:5383-8. [51] Ren A, Wang XC, Kellenberger CA, Rajashankar KR, Jones RA, Hammond MC, et al. Structural Basis for Molecular Discrimination by a 3',3'-cGAMP Sensing Riboswitch. Cell reports 2015;11:1-12. [52] Nelson JW, Sudarsan N, Phillips GE, Stav S, Lunse CE, McCown PJ, et al. Control of bacterial exoelectrogenesis by c-AMP-GMP. Proc Natl Acad Sci U S A 2015;112:5389-94. [53] Wu J, Sun L, Chen X, Du F, Shi H, Chen C, et al. Cyclic GMP-AMP is an endogenous second messenger in innate immune signaling by cytosolic DNA. Science 2013;339:826-30. [54] Sun L, Wu J, Du F, Chen X, Chen ZJ. Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science 2013;339:786-91. FIGURE LEGENDS: Figure 1. A. General mechanism of a metabolite-sensing riboswitch. In low concentrations of metabolite, the aptamer (blue) adopts a structure that allows formation of the downstream transcription terminator thus turning off the downstream gene. Under high metabolite conditions an altered conformation of the aptamer sequesters a portion of the downstream terminator to form an antiterminator that results in gene expression. B. Structural representations of second messengers c-di-AMP, c-di-GMP, 3’,3’ c-GAMP and 2’,3’ c-GAMP. C. Riboswitch classes with their ligands and binding affinities. Figure 2. Architectures of second messenger-binding riboswitches. A. The c-di-AMP binding riboswitch binds 2 molecules of ligand (magenta) at the junction of collinearly stacked helices. The inherent pseudo-symmetry is seen both within and between the two ligand binding sites. The P1 helix (cyan) is stabilized by multiple interactions in the ligand bound state of the RNA. B. The Type I c-di-GMP riboswitch forms a Y-shaped structure with P1 (cyan) at the base and P2 and P3 as parallel helices. The ligand (magenta) is bound at the junction of the helices and is stabilized by hydrogen bonding and stacking interactions. C. The Type II c-di-GMP riboswitch is distinct from Type I in that it harbors a kink-turn motif and a pseudoknot, both of
12
Page 12 of 15
Ac ce pt e
d
M
an
us
cr
ip t
which help stabilize ligand (magenta) in the binding pocket. D. The 3’,3’ c-GAMP riboswitch resembles the global architecture of the Type I c-di-GMP RNA (panel B) but lacks the conserved nucleotides required for recognition of two guanine bases from the ligand and instead preferentially binds 3’,3’ c-GAMP (magenta) to stabilize formation of the P1 helix (cyan).
13
Page 13 of 15
Page 14 of 15
ed
pt
ce
Ac M
an
cr
us
ip
Page 15 of 15
ed
ce pt
Ac M an
cr
us
ip t