Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2011 Apr 25;108(19):7757–7762. doi: 10.1073/pnas.1018857108

Structural basis of differential ligand recognition by two classes of bis-(3′-5′)-cyclic dimeric guanosine monophosphate-binding riboswitches

Kathryn D Smith a, Carly A Shanahan a, Emily L Moore b, Aline C Simon b, Scott A Strobel a,b,1
PMCID: PMC3093532  PMID: 21518891

Abstract

The bis-(3′-5′)-cyclic dimeric guanosine monophosphate (c-di-GMP) signaling pathway regulates biofilm formation, virulence, and other processes in many bacterial species and is critical for their survival. Two classes of c-di-GMP-binding riboswitches have been discovered that bind this second messenger with high affinity and regulate diverse downstream genes, underscoring the importance of RNA receptors in this pathway. We have solved the structure of a c-di-GMP-II riboswitch, which reveals that the ligand is bound as part of a triplex formed with a pseudoknot. The structure also shows that the guanine bases of c-di-GMP are recognized through noncanonical pairings and that the phosphodiester backbone is not contacted by the RNA. Recognition is quite different from that observed in the c-di-GMP-I riboswitch, demonstrating that at least two independent solutions for RNA second messenger binding have evolved. We exploited these differences to design a c-di-GMP analog that selectively binds the c-di-GMP-II aptamer over the c-di-GMP-I RNA. There are several bacterial species that contain both types of riboswitches, and this approach holds promise as an important tool for targeting one riboswitch, and thus one gene, over another in a selective fashion.

The ability of bacteria to adapt to their environment is essential for their survival. The bis-(3′-5′)-cyclic dimeric guanosine monophosphate (c-di-GMP) signaling system is one key pathway that allows bacteria to alter their behavior in response to changing environmental conditions (reviewed in refs. 1–4). Cyclic-di-GMP is a second messenger that is synthesized by diguanylate cyclases (DGCs) and degraded by specific phosphodiesterases in response to a variety of stimuli, including light, oxidative conditions, cell density, and antibiotics. Cyclic-di-GMP controls critical transitions such as the switch from a planktonic state to the formation of a biofilm (1, 2). It also controls virulence in many pathogenic bacterial species (5, 6).

To alter gene expression or protein activity in response to external or internal signals, c-di-GMP interacts with several proteins (1, 2, 7–17). In addition, two classes of RNAs have been reported that bind this second messenger, the c-di-GMP-I (class I) and c-di-GMP-II (class II) riboswitches (18, 19). Riboswitches are noncoding RNA elements usually located in the 5′ UTR of a mRNA that regulate gene expression by selectively binding small molecules. These RNAs have a two-domain architecture, with the 5′ domain (aptamer) responsible for ligand binding. Most often, genetic regulation is achieved when ligand binding in the first domain is coupled to an RNA structural rearrangement in the second domain leading to control at the transcriptional or translation level (reviewed in refs. 20–22).

Both class I and class II riboswitches have been shown to bind c-di-GMP and regulate gene expression (18, 19). The affinity of these RNAs for their ligand is very high; class I riboswitches have a Kd as tight as 10 pM (23) and the class II riboswitches have Kds between 200 pM and 2 nM (ref. 19 and Table 1). These Kds are substantially tighter than those of many of the known protein receptors (1, 2, 9–16). Whereas a few bacteria contain examples of both classes, several species lacking a class I riboswitch use c-di-GMP for signaling and have exclusively a class II riboswitch (19). The majority of c-di-GMP-II riboswitches are found in organisms from the bacterial class Clostridia, including the prominent pathogen Clostridium difficile, which is responsible for severe gastrointestinal infection. One of the four class II riboswitches found in C. difficile strain 630 allosterically regulates a downstream group I intron, demonstrating that these RNAs can be involved in sophisticated forms of genetic control (19). Molecular understanding of conserved bacterial pathways, such as c-di-GMP signaling, is critical to manipulate the growth of these organisms, a particularly desirable goal in the case of pathogens like C. difficile.

Table 1.

Cyclic-di-GMP-II riboswitch binding pocket mutants

RNA Kd (nM) Fold loss
Wild-type 2.2 ± 0.2 —
A69C > 50,000 > 25,000
A69G 190 ± 40 90
A69U 360 ± 10* 160
A70C 4.5 ± 0.4 2
A70G > 50,000 > 25,000
A70U > 50,000 > 25,000
G73A > 50,000 > 25,000
G73C 51 ± 9* 20
G73U > 50,000 > 25,000
Open in a new tab

*These values are the average of two trials ± s.d. All other values are the average of three trials ± s.d.

We previously reported the structure and mutational analysis of the class I aptamer, showing that c-di-GMP is bound as part of a duplex (23, 24). The second messenger binds at the junction of three helices and is recognized asymmetrically through base pairing interactions. One guanine of the ligand, Gα, is contacted on the Hoogsteen face by a G of the riboswitch whereas the second guanine, Gβ, forms a Watson–Crick base pair with a conserved C. This base pair formed with c-di-GMP is the first of a helical stem (23).

The predicted secondary structure of the class II riboswitch is distinct from that of class I. The class II aptamer was predicted to form three helices, a kink turn and a pseudoknot (19). The presence of these latter two elements, as well as the positions of the helices with respect to one another, suggested that the tertiary structure of the class II riboswitch is different from the class I aptamer. Phylogenetic sequence alignments showed that the majority of the conserved nucleotides were in the pseudoknot and in the regions immediately flanking this structural element. Additionally, in-line probing analysis revealed that these are the residues that become more structured upon c-di-GMP binding (19). Taken together these observations suggested that the pseudoknot region may have a role in ligand binding. If the class II aptamer recognizes c-di-GMP similarly to class I, we would expect that a conserved, unpaired C would be present within the secondary structure. However, there are no unpaired Cs in this RNA, suggesting that class II riboswitches recognize c-di-GMP through a different mechanism than used by class I (19).

Given these anticipated differences in second messenger recognition, we set out to determine the crystal structure of the class II riboswitch aptamer from Clostridium acetobutylicum bound to c-di-GMP. We used this structure to design a second messenger analog that is able to discriminate between class I and class II c-di-GMP-binding RNAs.

Results and Discussion

Structure Determination and Global Architecture of the c-di-GMP-II Aptamer.

We solved the structure of the aptamer domain of a class II c-di-GMP riboswitch from C. acetobutylicum at 2.5 Å resolution (Fig. 1, Fig. S1, and Table S1). This riboswitch controls a gene that contains a carbohydrate binding domain along with cellulase and glycosyl hydrolase domains (19), implying a role in carbohydrate processing. To promote crystallization and phasing, we increased the GC content of the variable region of the P1 stem and added an optimal iridium hexammine binding site into a nonconserved region of the P3 helix (26). The RNA sequence that was amenable to structure determination contains a six base-pair P1 helix and one additional nucleotide on the 3′ end (Fig. 1; see Fig. S2 for wild-type sequence). Iridium hexammine bound tightly to the engineered site in P3, as well as to several other G•U wobble pairs in the RNA. The structure was solved by single isomorphous replacement with anomalous scattering (SIRAS) phasing. In addition to binding within the RNA, c-di-GMP mediates intermolecular contacts in the crystal lattice (SI Text and Fig. S3). Two molecules are present in the asymmetric unit with an extensive RNA interface formed between them (Fig. S4).

Fig. 1.

Fig. 1.

Open in a new tab

Structure of the c-di-GMP-II aptamer domain from C. acetobutylicum. (A) Secondary structure of the c-di-GMP-II aptamer. P1 is shown in blue, the kink turn is shown in yellow, P3 is shown in purple, the pseudoknot helix (P4) is shown in green, and c-di-GMP is shown in red. Base pairs are shown using standard symbols (25). (B) Crystal structure of the c-di-GMP-II aptamer. Coloring is the same as in panel A. (C) Binding pocket. Coloring is the same as in panel A. 2Fo-Fc density is shown in gray, contoured at 1σ.

The class II aptamer folds into a compact structure with c-di-GMP bound at the junction of P1, P2, and P4 (Fig. 1). The predicted helices, kink turn, and pseudoknot are all present in the structure. The pseudoknot is formed from nucleotides in the loops of the P3 and P1/P2 helices and is therefore an example of an H-H-type pseudoknot, also called an intramolecular kissing hairpin (27–30). Helices P1 and P4 form a stack, with P2 at a right angle. The kink turn bends the backbone, allowing P3 to span across the RNA and connect with P4. One of the nucleotides in J3/4, A34, forms a noncanonical base pair with U45, the nucleotide at the junction of P3 and P4. This base pair stacks directly on the P4 helix and forms a base quadruple with the U31/G46 pair in P3 (Fig. S4). This quadruple may help facilitate the sharp bend that occurs at the P3/P4 junction (Fig. 1B).

A triple helix is formed in the major groove of P4 by nucleotides from J2/4 as well as c-di-GMP and a residue from P4 itself (Fig. 2 A and B). These triples are noncanonical interactions. Many additional noncanonical interactions are observed in this RNA. Twelve out of 37 pairings are not Watson–Crick or G•U wobble pairs, with six occurring in and around the binding pocket. Several of these are made through a single hydrogen bond.

Fig. 2.

Fig. 2.

Open in a new tab

Cyclic-di-GMP is bound as part of a triple helix. (A) Part of the pseudoknot of the c-di-GMP-II aptamer. Purple nucleotides are from P3, green nucleotides are from P4, yellow nucleotides are from the kink turn, and blue nucleotides form a triple helix with P4. (B) Triple helix formed with P4. Coloring is the same as in panel A. Hydrogen bonds are shown as black dashed lines. (C) The U37/A69 base pair is antiparallel to the rest of the P4 helix. Coloring is the same as in panel A. The major groove is highlighted in orange. Hydrogen bonds are shown as black dashed lines.

The binding pocket is composed of nucleotides from P1, P2, J2/4, and P4, although the majority of the nucleotides that directly contact c-di-GMP are from P4 (Fig. 1C). Residues from P1 and J2/4 provide stacking contacts. P2 forms the edge of the pocket behind the phosphodiester backbone of c-di-GMP but does not directly contact the ligand.

Recognition of c-di-GMP in the Class II Aptamer.

A unique feature of the pseudoknot allows the class II riboswitch to bind c-di-GMP as part of a triplex. Five noncanonical base triples are formed with P4 by nucleotides 61–63, Gα, and residue 70 (Fig. 2 A and B and Fig. S5). These bases form a continuous stack that extends into the P1 helix. Even though Gα stacks with the other residues that form the triple helix, it is antiparallel with respect to these nucleotides. To accommodate this one antiparallel base, the c-di-GMP-II aptamer uses two RNA motifs, a U-turn and an S-turn, to insert a single base pair into P4 upside down and backward with respect to the rest of the helix (Fig. 2C, Figs. S6 and S7, and SI Text). The result is that whereas four out of the five base triple interactions are in the major groove of P4, the base triple formed between Gα, U37 and A69 is a minor groove contact (Fig. S5). This groove reversal results because U37 is inserted into the P4 helix from P3 in an antiparallel fashion. The turn in the RNA strand that follows U37 positions G38, the next nucleotide, to form the first base pair of P4 (Fig. 2C). The S-turn flips A69 so that it can form a Watson-Crick pair with U37 (Fig. 2C and Fig. S7). This A69/U37 pair therefore presents its minor groove edge to Gα (Fig. S5). All RNA structures in the Protein Data Bank (PDB) were examined for this feature using the program AMIGOS II (31) but no structures were found that contained a similar U-turn/S-turn architecture. Incorporation of c-di-GMP into this triplex is a key recognition feature of the class II aptamer and is facilitated by a unique groove-reversed pseudoknot.

Specific recognition of c-di-GMP is achieved through stacking interactions as well as noncanonical pairings to the guanine bases. Three conserved adenosines stack between and on either side of c-di-GMP. A70 intercalates between the two guanine bases of the ligand whereas A13 stacks below Gβ and A61 stacks above Gα (Fig. 3A). Stacking is continued into the P1 helix as well as the third strand of the P4 triplex. In the few sequences where these nucleotides are not adenosines they are guanines, consistent with their role in stacking (Fig. S8). Gα forms a minor groove base triple by interacting with the sugar edge of A69 in the P4 helix. No contacts are made to the Hoogsteen face of Gα (Fig. 3B). Gβ is contacted on its Watson–Crick and Hoogsteen faces by RNA atoms as well as a hydrated magnesium ion (Fig. 3C). G73 forms a single hydrogen bond to the N1 of Gβ and the 2′-OH of A70 contacts the N7. A fully hydrated magnesium ion makes outer sphere contacts with the O6 of Gβ. Interestingly, no contacts are made to the phosphodiester backbone of c-di-GMP with the single exception of a hydrogen bond between the exocyclic amine of A70 and one of the nonbridging phosphate oxygens of the dinucleotide (Fig. 3A).

Fig. 3.

Fig. 3.

Open in a new tab

Recognition of c-di-GMP by the class I and class II riboswitch. Coloring is the same as in Fig. 1 except c-di-GMP is colored by atom with carbon as white, oxygen as red, nitrogen as blue, and phosphorus as orange. Hydrogen bonds are shown as black dashed lines. (A) Stacking interactions in the class II aptamer. (B) Recognition of Gα by the class II aptamer. The designation of Gα was chosen based on the same relative orientation to P1 as that in the class I aptamer. (C) Recognition of Gβ by the class II aptamer. The waters of the hydrated magnesium are shown as red spheres and the magnesium as a yellow sphere. (D) Stacking interactions in the class I aptamer (PDB ID 3MXH). (E) Recognition of Gα by the class I aptamer. (F) Recognition of Gβ by the class I aptamer.

Previous data from in-line probing (19), as well as the structural data presented here, show that c-di-GMP binding to the class II aptamer stabilizes the formation of the pseudoknot and triplex structures, and facilitates stacking between P1 and P4. Riboswitch sequence alignment reveals that in many cases a rho-independent terminator stem can form using the 3′ end of P1 and, in several cases (19), the nucleotides that directly contact c-di-GMP. This provides a clear mechanism whereby ligand binding stabilizes a conformation of the RNA in which both P1 and P4 are intact, leading to genetic control.

Mutagenesis of Binding Pocket Nucleotides.

To further investigate the interactions made between the class II aptamer and c-di-GMP we mutated the nucleotides that stack with and directly contact the bases of the ligand and assessed the effect on binding affinity using a native gel-shift assay (Fig. S9). The nucleotides directly involved in c-di-GMP recognition are conserved in at least 90% of sequences (Fig. S8). There are only three examples in which A69 is not an A. In these cases it is a guanine and there is a compensatory U to C transition at position 37, which maintains base pairing between these residues. Similarly, in only two cases is A70 variable, and in both examples it is a guanine. G73 is about 90% conserved and is usually substituted with a pyrimidine (19).

In accordance with the high level of sequence conservation of nucleotides 69, 70, and 73, mutations at these positions caused large decreases (> 25,000-fold) in c-di-GMP affinity. The interpretation is complicated however, because A69 and A70 are involved in base triples, which makes it difficult to ascribe the decreases in affinity solely to the loss of ligand contacts. Mutation of A69, the residue that interacts with Gα, to a guanine caused a 90-fold loss in binding affinity. Other mutations to this base result in larger losses in affinity, with A69C not binding at all (Table 1). All mutations to the adenosine that stacks between Gα and Gβ eliminate c-di-GMP binding with the surprising exception of A70C that binds within twofold of the wild-type affinity (Table 1). A cytosine at this position may be able to maintain the contact to the backbone of c-di-GMP as well as present a hydrogen bond acceptor to the C71/G39 base pair, perhaps making up for the loss in favorable stacking energy. When G73, the nucleotide that contacts Gβ, was mutated, G73C was the only variant that retained any binding.

Recognition Is Distinct from the Class I Riboswitch.

Two classes of riboswitches have been discovered that regulate gene expression in response to the second messenger c-di-GMP. Both are important macromolecular targets in this signaling pathway. We have characterized the atomic level interactions made by each class of RNA, and in both cases, the symmetrical ligand is recognized asymmetrically by the riboswitch, but in very different ways.

Both riboswitches incorporate the bases of c-di-GMP into structural elements, a duplex in class I and a triplex in class II. In class I, Gβ is incorporated as the first base pair of a helix and in class II Gα is part of a triple helix formed with a pseudoknot. The ability of RNA to integrate c-di-GMP into stable structures may explain the high affinity of the c-di-GMP-binding riboswitches compared with protein receptors. Members of both classes of c-di-GMP-binding RNAs have picomolar affinity for their ligand (19, 23), whereas protein receptors bind with nanomolar to micromolar affinity (1, 2, 9–16). Because c-di-GMP can participate in RNA structures, riboswitches seem uniquely suited to serve as high-affinity targets for this second messenger.

When comparing the binding pockets of both types of c-di-GMP-binding RNAs, we observe four distinct ways in which the guanine bases are recognized (Fig. 3 B, C, E, and F). In class I riboswitches the interactions are Watson–Crick and Hoogsteen pairs (23), whereas in class II no canonical pairings are present. Despite the diversity in base recognition, we observe similar stacking contacts in both classes, suggesting that these interactions are critical for binding (Fig. 3 A and D). Additionally, the backbone recognition is quite different in both classes. In class I, contacts from both the RNA and a hydrated magnesium are observed whereas in class II, no contacts are made to the backbone with the exception of the hydrogen bond made with A70. Both Gs as well as the phosphodiester backbone are more heavily contacted in class I riboswitches than in class II. This may be reflected in the higher affinity of class I for c-di-GMP (23). It is clear that at least two independent solutions have evolved for RNA recognition of this second messenger.

A c-di-GMP Analog Selectively Binds the Class II Riboswitch.

Given the differential approaches to ligand binding we hypothesized that analogs of c-di-GMP may be able to distinguish between the two RNAs. Based on the lack of recognition of the c-di-GMP phosphodiester backbone by the class II riboswitch, we sought to design an analog that would specifically bind this RNA. Both 2′-OHs of c-di-GMP are contacted by either an RNA atom or a well coordinated solvent molecule in the c-di-GMP-I riboswitch. It appears there is insufficient space for any additional steric bulk on either 2′-OH without rearrangement of the RNA (23, 24). However, in the c-di-GMP-II aptamer both 2′-OHs face directly into solvent. We hypothesized therefore that a 2′-O-methyl analog of c-di-GMP would target class II over class I. We synthesized this derivative and tested for binding to both class I and class II riboswitches using a competition assay.

The 2′-O-methyl analog is able to discriminate between class I and class II riboswitches (Fig. 4). The Kd of the class II complex with the 2′-O-methyl analog is 4.3 nM, which is very close to the affinity for the parental dinucleotide (2.2 nM) (Table 2). The class I RNA we have used in previous studies has a very slow approach to equilibrium due to slow binding and dissociation rates (23). Therefore, to reach equilibrium in class I we used a mutant that has an affinity equivalent to that of the class II riboswitch (24) (Table 2). In an equilibrium competition assay, we found that very high concentrations of 2′-O-methyl analog (approximately 250 μM) are necessary to even partially displace c-di-GMP. We estimate the Kd to be approximately 60 μM, which corresponds to at least a 30,000-fold loss of affinity relative to c-di-GMP (Table 2). This is a nearly 6 kcal/mol loss in binding energy compared to the native ligand. Whereas further experiments will be necessary to determine if this differential binding can be exploited in vivo, these experiments show that the class II RNA reaches equilibrium in 1 hour for the c-di-GMP binding reaction, indicating that it may have a faster association rate constant than the class I riboswitch. If a significant difference in kon does exist, an even more pronounced difference in the concentration of analog necessary to see an effect in vivo may be expected.

Fig. 4.

Fig. 4.

Open in a new tab

A 2′-O-methyl c-di-GMP analog selectively binds the class II riboswitch. (A) Structure of c-di-GMP and the analog. (B) Competition assay. Radiolabeled c-di-GMP is in trace and RNA is at 25 nM in all cases.

Table 2.

Binding of a 2′-O-methyl analog to class I and class II riboswitches

RNA c-di-GMP Kd (nM) 2′-O-methyl Kd (nM) Fold loss
c-di-GMP-I 2.2 ± 0.8* approximately 60,000 approximately 30,000
c-di-GMP-II 2.2 ± 0.2 4.3 ± 0.7 2
Open in a new tab

*This RNA corresponds to the GUAA mutant reported in ref. 24. All values are the average of three trials ± s.d.

These results demonstrate the difference in recognition between the two c-di-GMP-binding aptamers and provide evidence that it may be possible to exploit this difference to selectively bind one aptamer over the other. The ability to selectively target one riboswitch could provide a useful tool for studying c-di-GMP signaling in organisms that use both classes of second messenger binding RNAs. Additionally, some organisms use only one type of riboswitch so an analog able to discriminate between the two classes could be used to selectively target one organism while not affecting RNA based signaling in others. Furthermore, structures of several protein targets of c-di-GMP have been determined and analogs may be designed that can differentiate not only between class I and class II riboswitches but also between RNA and protein receptors.

Materials and Methods

Materials.

RNAs.

The wild-type RNA sequence (Fig. S2) was cloned into a pUC19 vector using overlapping oligonucleotides. The T7 RNA polymerase promoter was placed immediately 5′ of the riboswitch sequence and the antigenomic hepatitis delta virus (HDV) ribozyme was cloned directly 3′ of the riboswitch sequence to generate homogenous 3′ ends. Mutations necessary for crystallization as well as those used for biochemical analysis were generated using site directed mutagenesis. All RNAs were in vitro transcribed using T7 RNA polymerase and after posttranscriptional HDV ribozyme cleavage were purified by 6% denaturing polyacrylamide gel electrophoresis (PAGE). RNAs were excised from the gel and eluted overnight in 300 mM NaOAc, pH 5.2 after crushing.

Cyclic-di-GMP.

A modified diguanylate cyclase (tDGC) was purified and used to generate c-di-GMP enzymatically from GTP as described (32). Briefly, 1 mM GTP was added every hour to 5 μM protein in reaction buffer (50 mM Tris, pH 7.5, 250 mM NaCl, 20 mM MgCl2) in a total reaction volume of 10 mL for approximately 8 hours. A heavy white precipitate corresponding to pyrophosphate was observed. The protein was heat-denatured and removed and the supernatant was purified by reverse phase HPLC on a C18 column using a gradient from 100% 50 mM triethylammonium acetate, pH 6.0 to 10% acetonitrile over 30 minutes.

Methods.

Crystallization.

RNA (150 μM) was refolded in buffer (10 mM NaCac, pH 6.8, 10 mM MgCl2, 10 mM KCl) in the presence of 375 μM c-di-GMP. This complex was allowed to come to equilibrium and then was mixed in a 1∶1 ratio with well solution consisting of 13% PEG 8000, 200 mM Mg(OAc)2, and 100 mM NaCac, pH 6.0. Crystals were grown at 25 °C using hanging drop vapor diffusion and grew to maximum size in approximately 1–2 weeks. Crystals were frozen by equilibrating in mother liquor with the amount of PEG 8000 increased to 33% plus 150 μM c-di-GMP and then flash freezing in liquid nitrogen. For iridium hexammine soaks, crystals were transferred to freezing solution supplemented with 1 mM iridum hexammine and allowed to equilibrate for 4–5 hours prior to flash freezing.

Structure determination.

Native and single-wavelength anomalous dispersion (SAD) datasets were collected at Brookhaven National Laboratory beamline X29A. Data were processed using HKL2000 (33). The Matthew’s coefficient suggested that there were two molecules present in the asymmetric unit. Native Patterson maps revealed a translational twofold noncrystallographic symmetry (NCS) axis between the two molecules. SIRAS phasing was performed using a native dataset and a SAD dataset with a Riso of 24%. The SHELX suite (through the hkl2map interface) was used to find four initial sites at a resolution cutoff of 3.2 Å (34). These sites were refined in MLPHARE (35) and initial phases were calculated. Difference Fourier methods were then employed to find additional sites. A total of 8 sites were included to produce the maps used for building. Density modification with twofold NCS averaging was performed with RESOLVE (36). The resolution was extended to 2.5 Å during density modification. Model building was done in COOT (37) and refinement was performed in Refmac (38). Figures were made using Pymol (39).

Synthesis of the 2′-O-methyl c-di-GMP analog.

Cyanoethyl phosphate protected 2′-O-methoxy guanosine phosphoramidite and the 3-(4,4′-dimethoxytrityloxy)-2,2-(dicarboxymethylamido)propyl-1-O-succinoyl-long chain alkylamino-CPG (3′-CPR II CPG) solid support were purchased from Glen Research. Synthesis was performed on solid-phase (2 μmol scale) as previously described with slight adaptations (40–42) (see SI Text).

Measurement of dissociation constants.

Radiolabeled c-di-GMP was enzymatically synthesized as previously described (23, 43) and incubated with RNA in folding buffer containing 10 mM NaCac, pH 6.8, 10 mM MgCl2, and 10 mM KCl. Binding reactions were allowed to come to equilibrium at room temperature (approximately 21–23 °C) and bound and free c-di-GMP were separated by native PAGE and data were quantified as reported previously (23, 24). Competition experiments with the 2′-O-methyl c-di-GMP analog were performed under similar conditions, only the radiolabeled c-di-GMP and unlabeled competitor ligand were premixed before adding wild-type class II RNA or GUAA mutant class I RNA (24) to a final concentration of 25 nM. Data were fit to the following equation as previously reported (44) to determine the dissociation constant of the 2′-O-methyl analog:

graphic file with name pnas.1018857108eq1.jpg

where FB = fraction bound; [ ∗ cdiG] = 5 pM; [Omethyl] = concentration of 2′-O-methyl analog; [RNA] = 25 nM.

In all cases, binding reactions were allowed to equilibrate for increasing amounts of time until no change in the final fraction bound was observed. This was approximately 1 hour for the wild-type class II sequence and mutants. For competition experiments, the class II RNA reached equilibrium after incubation overnight, whereas the class I RNA required a 48 hour incubation.

Supplementary Material

Supporting Information
supp_108_19_7757__index.html (991B, html)

Acknowledgments.

We thank S. Myers, A. Soares, A. Héroux, and the beamline staff at X25 and X29A at the National Synchrotron Light Source at Brookhaven National Laboratory; M. Strickler and D. Keller (Yale); J. Wang for data processing and phasing help and advice; E. Butler for help with data collection; V. Singh for synthesis of iridium hexammine; Z. Weinberg, E. Lee, N. Sudarsan, R. Breaker, and other members of the Breaker Lab for helpful advice and discussions; K. Keating and A. Pyle for the AMIGOS II software; E. Butler, S. Lipchock, D. Hiller and other members of the Strobel Lab for helpful discussions; Z. Liang for the gift of the tDGC expression plasmid; and U. Jenal (University of Basel) for the gift of the PleD* expression plasmid. This work was supported by National Institutes of Health Grant GM022778 and US Department of Defense National Security Science and Engineering Faculty Fellowship Grant 1N00244-09-1-0070.

Footnotes

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The atomic coordinates and structure factors have been deposited in the Protein Data Bank, www.pdb.org (PDB ID code 3Q3Z).

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1018857108/-/DCSupplemental.

References

  • 1.Hengge R. Principles of c-di-GMP signalling in bacteria. Nat Rev Microbiol. 2009;7:263–273. doi: 10.1038/nrmicro2109. [DOI] [PubMed] [Google Scholar]
  • 2.Schirmer T, Jenal U. Structural and mechanistic determinants of c-di-GMP signalling. Nat Rev Microbiol. 2009;7:724–735. doi: 10.1038/nrmicro2203. [DOI] [PubMed] [Google Scholar]
  • 3.Römling U, Amikam D. Cyclic di-GMP as a second messenger. Curr Opin Microbiol. 2006;9:218–228. doi: 10.1016/j.mib.2006.02.010. [DOI] [PubMed] [Google Scholar]
  • 4.Jenal U, Malone J. Mechanisms of cyclic-di-GMP signaling in bacteria. Annu Rev Genet. 2006;40:385–407. doi: 10.1146/annurev.genet.40.110405.090423. [DOI] [PubMed] [Google Scholar]
  • 5.Tamayo R, Pratt JT, Camilli A. Roles of cyclic diguanylate in the regulation of bacterial pathogenesis. Annu Rev Microbiol. 2007;61:131–148. doi: 10.1146/annurev.micro.61.080706.093426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Cotter PA, Stibitz S. c-di-GMP-mediated regulation of virulence and biofilm formation. Curr Opin Microbiol. 2007;10:17–23. doi: 10.1016/j.mib.2006.12.006. [DOI] [PubMed] [Google Scholar]
  • 7.Paul K, Nieto V, Carlquist WC, Blair DF, Harshey RM. The c-di-GMP binding protein YcgR controls flagellar motor direction and speed to affect chemotaxis by a “backstop brake” mechanism. Mol Cell. 2010;38:128–139. doi: 10.1016/j.molcel.2010.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Boehm A, et al. Second messenger-mediated adjustment of bacterial swimming velocity. Cell. 2010;141:107–116. doi: 10.1016/j.cell.2010.01.018. [DOI] [PubMed] [Google Scholar]
  • 9.Krasteva PV, et al. Vibrio cholerae VpsT regulates matrix production and motility by directly sensing cyclic di-GMP. Science. 2010;327:866–868. doi: 10.1126/science.1181185. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Leduc JL, Roberts GP. Cyclic di-GMP allosterically inhibits the CRP-like protein (Clp) of Xanthomonas axonopodis pv. citri. J Bacteriol. 2009;191:7121–7122. doi: 10.1128/JB.00845-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Chin K-H, et al. The c-AMP receptor-like protein CLP is a novel c-di-GMP receptor linking cell-cell signaling to virulence gene expression in Xanthomonas campestris. J Mol Biol. 2009;396:646–662. doi: 10.1016/j.jmb.2009.11.076. [DOI] [PubMed] [Google Scholar]
  • 12.Tao F, He Y-W, Wu D-H, Swarup S, Zhang L-H. The cNMP domain of Xanthomonas campestris global regulator Clp defines a new class of c-di-GMP effectors. J Bacteriol. 2009;192:1020–1029. doi: 10.1128/JB.01253-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hickman JW, Harwood CS. Identification of FleQ from Pseudomonas aeruginosa as a c-di-GMP-responsive transcription factor. Mol Microbiol. 2008;69:376–389. doi: 10.1111/j.1365-2958.2008.06281.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Duerig A, et al. Second messenger-mediated spatiotemporal control of protein degradation regulates bacterial cell cycle progression. Genes Dev. 2009;23:93–104. doi: 10.1101/gad.502409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Newell PD, Monds RD, O’Toole GA. LapD is a bis-(3′,5′)-cyclic dimeric GMP-binding protein that regulates surface attachment by Pseudomonas fluorescens Pf0-1. Proc Natl Acad Sci USA. 2009;106:3461–3466. doi: 10.1073/pnas.0808933106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Lee VT, et al. A cyclic-di-GMP receptor required for bacterial exopolysaccharide production. Mol Microbiol. 2007;65:1474–1484. doi: 10.1111/j.1365-2958.2007.05879.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Amikam D, Galperin MY. PilZ domain is part of the bacterial c-di-GMP binding protein. Bioinformatics. 2006;22:3–6. doi: 10.1093/bioinformatics/bti739. [DOI] [PubMed] [Google Scholar]
  • 18.Sudarsan N, et al. Riboswitches in eubacteria sense the second messenger cyclic di-GMP. Science. 2008;321:411–413. doi: 10.1126/science.1159519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Lee ER, Baker JL, Weinberg Z, Sudarsan N, Breaker RR. An allosteric self-splicing ribozyme triggered by a bacterial second messenger. Science. 2010;329:845–848. doi: 10.1126/science.1190713. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Garst AD, Edwards AL, Batey RT. Riboswitches: Structures and mechanisms. Cold Spring Harb Perspect Biol. 2010 doi: 10.1101/cshperspect.a003533. 10.1101/cshperspect.a003533. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Serganov A. The long and the short of riboswitches. Curr Opin Struct Biol. 2009;19:251–259. doi: 10.1016/j.sbi.2009.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Roth A, Breaker RR. The structural and functional diversity of metabolite-binding riboswitches. Annu Rev Biochem. 2009;78:305–334. doi: 10.1146/annurev.biochem.78.070507.135656. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Smith K, et al. Structural basis of ligand binding by a c-di-GMP riboswitch. Nat Struct Mol Biol. 2009;16:1218–1223. doi: 10.1038/nsmb.1702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Smith KD, Lipchock SV, Livingston AL, Shanahan CA, Strobel SA. Structural and biochemical determinants of ligand binding by the c-di-GMP riboswitch. Biochemistry. 2010;49:7351–7359. doi: 10.1021/bi100671e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Leontis NB, Westhof E. Geometric nomenclature and classification of RNA base pairs. RNA. 2001;7:499–512. doi: 10.1017/s1355838201002515. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Keel AY, Rambo RP, Batey RT, Kieft JS. A general strategy to solve the phase problem in RNA crystallography. Structure. 2007;15:761–772. doi: 10.1016/j.str.2007.06.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Sperschneider J, Datta A, Wise M. Heuristic RNA pseudoknot prediction including intramolecular kissing hairpins. RNA. 2011;17:27–38. doi: 10.1261/rna.2394511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Giedroc DP, Cornish PV. Frameshifting RNA pseudoknots: Structure and mechanism. Virus Res. 2009;139:193–208. doi: 10.1016/j.virusres.2008.06.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Brierley I, Pennell S, Gilbert RJC. Viral RNA pseudoknots: Versatile motifs in gene expression and replication. Nat Rev Microbiol. 2007;5:598–610. doi: 10.1038/nrmicro1704. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Staple DW, Butcher SE. Pseudoknots: RNA structures with diverse functions. PLoS Biol. 2005;3:956–959. doi: 10.1371/journal.pbio.0030213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Wadley L, Keating K, Duarte C, Pyle A. Evaluating and learning from RNA pseudotorsional space: Quantitative validation of a reduced representation for RNA structure. J Mol Biol. 2007;372:942–957. doi: 10.1016/j.jmb.2007.06.058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Rao F, et al. Enzymatic synthesis of c-di-GMP using a thermophilic diguanylate cyclase. Anal Biochem. 2009;389:138–142. doi: 10.1016/j.ab.2009.03.031. [DOI] [PubMed] [Google Scholar]
  • 33.Otwinowski Z, Minor W. Processing of X-ray diffraction data collected in oscillation mode. Method Enzymol. 1997;276:307–326. doi: 10.1016/S0076-6879(97)76066-X. [DOI] [PubMed] [Google Scholar]
  • 34.Pape T, Schneider T. HKL2MAP: A graphical user interface for phasing with SHELX programs. J Appl Crystallogr. 2004;37:843–844. [Google Scholar]
  • 35.Collaborative Computational Project 4. The CCP4 suite: Programs for protein crystallography. Acta Crystallogr D. 1994;50:760–763. doi: 10.1107/S0907444994003112. [DOI] [PubMed] [Google Scholar]
  • 36.Terwilliger T. Maximum likelihood density modification. Acta Crystallogr D. 2000;56:965–972. doi: 10.1107/S0907444900005072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Emsley P, Cowtan K. Coot: Model-building tools for molecular graphics. Acta Crystallogr D. 2004;60:2126–2132. doi: 10.1107/S0907444904019158. [DOI] [PubMed] [Google Scholar]
  • 38.Winn MD, Isupov MN, Murshudov GN. Use of TLS parameters to model anisotropic displacements in macromolecular refinement. Acta Crystallogr D. 2001;57:122–133. doi: 10.1107/s0907444900014736. [DOI] [PubMed] [Google Scholar]
  • 39.DeLano WL. Pymol. San Carlos, CA: DeLano Scientific; 2008. [Google Scholar]
  • 40.Kiburu I, Shurer A, Yan L, Sintim HO. A simple solid-phase synthesis of the ubiquitous bacterial signaling molecule, c-di-GMP and analogues. Mol Biosyst. 2008;4:518–520. doi: 10.1039/b719423d. [DOI] [PubMed] [Google Scholar]
  • 41.Alazzouzi E, Escaja N, Grandas A, Pedroso E. A straightforward solid-phase synthesis of cyclic oligodeoxyribonucleotides. Angew Chem Int Ed Engl. 1997;36:1506–1508. [Google Scholar]
  • 42.Micura R. Cyclic oligoribonucelotides (RNA) by solid-phase synthesis. Chemistry Eur J. 1999;5:2077–2082. [Google Scholar]
  • 43.Paul R, et al. Cell cycle-dependent dynamic localization of a bacterial response regulator with a novel di-guanylate cyclase output domain. Genes Dev. 2004;18:715–727. doi: 10.1101/gad.289504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Strobel SA, Cech TR. Tertiary interactions with the internal guide sequence mediate docking of the P1 helix into the catalytic core of the Tetrahymena ribozyme. Biochemistry. 1993;32:13593–13604. doi: 10.1021/bi00212a027. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information
supp_108_19_7757__index.html (991B, html)
1018857108_pnas.1018857108_SI.pdf (1.7MB, pdf)

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES