Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Mar 20.
Published in final edited form as: Chem Biol. 2014 Feb 20;21(3):345–356. doi: 10.1016/j.chembiol.2014.01.004

Structure-guided design of fluorescent S-adenosylmethionine analogs for a high-throughput screen to target SAM-I riboswitch RNAs

Scott F Hickey 1,3, Ming C Hammond 1,2,3
PMCID: PMC4074398  NIHMSID: NIHMS569726  PMID: 24560607

Summary

Many classes of S-adenosylmethionine (SAM)-binding RNAs and proteins are of interest as potential drug targets in diverse therapeutic areas, from infectious diseases to cancer. In the former case, the SAM-I riboswitch is an attractive target because this structured RNA element is found only in bacterial mRNAs and regulates multiple genes in several human pathogens. Here we describe the synthesis of stable and fluorescent analogs of SAM in which the fluorophore is introduced through a functionalizable linker to the ribose. A Cy5-labeled SAM analog was shown to bind several SAM-I riboswitches via in-line probing and fluorescence polarization (FP) assays, including one from Staphylococcus aureus that controls the expression of SAM synthetase in this organism. A fluorescent ligand displacement assay was developed and validated for high-throughput screening of compounds to target the SAM-I riboswitch class.

Introduction

Riboswitches are structured RNA elements typically found in bacterial mRNA untranslated regions that are capable of sensing cellular metabolites (Breaker 2011). Binding of the cognate ligand to the riboswitch aptamer domain affects a conformational change in the expression platform, resulting in regulation of downstream gene expression. To date, more than 20 classes of natural riboswitches have been discovered (Serganov and Nudler, 2013) that bind a wide variety of metabolites. These metabolites range in structure from simple amino acids such as glycine, to large and elaborate cofactors such as adenosylcobalamin, to the fluoride ion. The diversity of riboswitch ligands reveals the intrinsic potential of structured RNA elements to form binding pockets for specific molecular recognition.

Due to their roles in bacterial gene regulation and prospects of targeting the ligand binding pocket of the aptamer domain, riboswitches have generated interest for the development of novel antibiotics (Blount and Breaker, 2006; Deigan and Ferré-D'Amaré, 2011). Subversion of the normal regulatory response of the riboswitch by an exogenous compound may disable the function of essential genes and slow bacterial growth. The availability of ligand-bound crystal structures for many riboswitch classes has made it possible to identify ligand analogs by structure-based rational design (reviewed Deigan and Ferré-D'Amaré, 2011). For example, bactericidal compounds have been identified that target the lysine riboswitch (Blount et al., 2007; Budhathoki et al., 2012) and guanine riboswitch (Kim et al., 2009; Mulhbacher et al., 2010). Computational docking also has been employed to predict new ligands for the guanine riboswitch, however so far there is not a strong correlation between docking score and ligand affinity (Daldrop et al., 2011).

Alternatively, the development of high-throughput screening methods offers a separate approach for riboswitch ligand discovery. High-throughput methods allow rapid screening of more diverse sets of compounds, enabling discovery of candidate leads that would not be predicted by rational design. In addition, this approach takes advantage of the large and diverse chemical compound libraries that are available. Both fluorescence polarization and FRET-based assays have been developed for screening activators of the glmS riboswitch using fluorescently tagged oligonucleotides (Mayer and Famulok, 2006; Blount and Breaker, 2006). These screening techniques take advantage of the natural self-cleaving ability of the glmS ribozyme, however they are limited by slow ribozyme kinetics and false positives from ligand-independent cleavage (De Silva and Walter, 2009). Although an allosteric ribozyme was engineered for the Vc2 c-di-GMP riboswitch in order to develop a similar screening assay (Furukawa et al., 2012), the method requires extensive optimization for each riboswitch sequence of interest.

Other screening techniques have been developed for riboswitches that do not have ribozyme activity. An equilibrium dialysis assay was employed to screen a library of 1300 fragments for activity against the E. coli thiM TPP riboswitch (Cressina et al., 2011; Chen et al., 2010). However, the method requires tritium-labeled thiamine and the fragments were screened in sets of five to increase throughput, necessitating an additional deconvolution step after the initial screen. We recently have described the development of a microfluidics-based mobility shift assay (μMSA) that has promising quantitative analytical capabilities in addition to being time and resource sparing (Karns et al., 2013). Adaptation of this technology to a multiplex format would be highly desirable for high-throughput screening.

Recently, a dual molecular beacon assay was described for the B. subtilis pbuE adenine riboswitch in which ligand analogs were screened for their ability to regulate transcription (Chinnappan et al., 2013). In contrast to other methods, this system enables co-transcriptional screening of ligand analogs in a functional context, which is important as kinetic parameters are known to be critical for riboswitch function (Wickiser et. al, 2005). Thus, this method identifies compounds that not only bind the riboswitch aptamer, but also affect the transcriptional regulatory activity of the riboswitch. Although this assay has been demonstrated only in a low-throughput format, the technique is potentially amenable to high-throughput screening. However, optimization of the molecular beacons for each individual riboswitch sequence is still required and compounds with intrinsic fluorescence, transcriptional inhibitory activity, or that affect the molecular beacons themselves must be avoided (Chinnappan et al., 2013).

So far, SAM-binding riboswitches have been largely unexplored as targets for novel antibiotics, even though seven different classes have been reported (SAM-I through SAM-V, SAM-I/IV, and SAM/SAH) that each have distinct structural features (Wang and Breaker, 2008; Batey 2011). These riboswitches have been extensively characterized both structurally (Baird and Ferré-D'Amaré, 2010; Montange and Batey, 2006; Gilbert et al., 2008; Lu et al., 2008; Stoddard et al., 2010; Wilson et al., 2011) and mechanistically (reviewed in Garst et al., 2011; Breaker 2012; Haller et al., 2011). Rationally designed SAM analogs to target the SAM-II riboswitch aptamer have been described (Ham et al., 2011), and S-2,6-diaminopurinylmethionine (DAPSAM) was shown to bind the SAM-III riboswitch aptamer (Ottink et al., 2010). However, while the SAM-I riboswitch class has been specifically proposed as an antibacterial drug target (Blount and Breaker, 2006), it has posed a challenge to existing screening methodology.

In this study, we set out to design fluorescently-tagged SAM analogs for use in a ligand displacement assay for targeting SAM-I riboswitches (Figure 1A). A fluorescence-based assay was recently developed for the purine riboswitch using 2-aminopurine as the fluorescent ligand, and has considerable advantages over other methods such as in-line probing and isothermal titration calorimetry (Daldrop et al., 2013). Fluorescent SAM analogs based on 2-aminopurine have been prepared (Ottink et al., 2010), however it is known that the 6-amino group is important for SAM-I riboswitch binding affinity (Lim et al., 2006). Instead, we considered developing fluorescently-tagged SAM analogs via installation of a functionalizable linker at different positions.

Figure 1. Structure-based rational design of a fluorescently-tagged ligand analog that targets the SAM-I riboswitch.

Figure 1

Open in a new tab

(A) Schematic representation of a fluorescent ligand displacement assay for the SAM-I riboswitch. Fluorescent SAM analog is displaced by unlabeled competitor, leading to decrease in fluorescence polarization in a dose-dependent manner. Chemical structures of SAM 1, SAH sulfone 4, SAH 5, and C8-Cy5 18 are shown. (B) The x-ray crystal structure of a B. subtilis SAM-I riboswitch bound to SAM (PDB code 3NPB) shows that the SAM ribose hydroxyls are solvent accessible. The 2′-hydroxyl makes hydrogen bonding contacts to C48 and U112 of the riboswitch. See also Figure S3.

Analysis of the three-dimensional crystal structures of the SAM-I riboswitch aptamer bound to SAM (Montange and Batey, 2006, Lu et al., 2010) suggested the 2′ and 3′-ribose hydroxyl groups of SAM 1 as candidate positions for fluorophore attachment. Although numerous SAM analogs have been prepared for use as protein substrates or inhibitors, there is currently only one example of chemical modification at the 2′- or 3′-hydroxyl of the ribose moiety (Li et al., 2011). In this work, we report the synthesis of SAM analogs containing functionalized linkers installed onto the 2′ and 3′ ribose hydroxyls, and show that these analogs retain binding to the SAM-I riboswitch. We further describe the development of a fluorescence polarization assay using a fluorescently-tagged SAM analog for targeting SAM-I riboswitches. We demonstrate a fluorescent ligand displacement assay for targeting SAM-I riboswitches, including the S. aureus metK SAM-I sequence, that may be useful as a high-throughput screening method for antibacterial drug discovery.

Results

Synthesis and Evaluation of Two Functionalizable SAM Analogs

For targeting SAM-I riboswitches, we sought to develop a synthetic route that was amenable to the installation of functionalizable linkers on the SAM ribose hydroxyl groups. The riboswitch aptamer makes extensive contacts to both the nucleobase and amino acid moieties of SAM 1 (Montange and Batey, 2006, Lu et al., 2010), and functional group transformations at these positions are not well tolerated (Lim et al., 2006). However, the 2′ and 3′ ribose hydroxyls of SAM 1 are located within a solvent-accessible cavity, and only the 2′ hydroxyl is directly contacted through hydrogen bonding interactions (Figure 1B).

Therefore, we designed two prototype analogs 2 and 3 containing either a 2′-3′-O-acetal or 3′-O-propionate linker, respectively (Scheme 1A). We based our design on a stable sulfone analog of SAM 1 (SAH sulfone, 4) that has been shown to bind the SAM-I riboswitch from the yitJ transcript of Bacillus subtilis with high affinity (Kd value of ∼30 nM) (Lim et al., 2006). The electropositive character imparted on the sulfur atom by oxidation to the sulfone leads to a higher affinity relative to S-adenosylhomocysteine (SAH, 5), which has a Kd value of ∼400 nM (Lim et al., 2006). Thus, we reasoned that fluorescently tagged versions of 4 would be useful in competition binding assays to identify novel riboswitch ligands.

Scheme 1.

Scheme 1

Open in a new tab

(A) Synthetic route to prototype compounds containing 2′-3′-O-acetal (2, top) and 3′-O-propionate (3, bottom) linkers to SAH sulfone. See also Figure S2. (B) Synthetic route to C3 and C8 linker-containing intermediates 16 and 17. See also Figure S2.

It was envisioned that linkers first could be introduced to the adenosine nucleoside followed by coupling to the homocysteine portion. The 2-(methoxycarbonyl)ethylidine group has been described as an unusually stable acetal protecting group for 1,2-diols (Ariza et al., 2000). As previously shown for 5′-O-trityluridine (Ariza et al., 2000), acetal formation was achieved cleanly by reacting 5′-chloroadenosine 6 with methyl propiolate and DMAP. Since formation of the acetal is non-stereoselective, we isolated and carried forward both diastereomers of 7 (60:40 R/S, diastereomers assigned by NOESY and ratio determined by 1H NMR).

Thioether formation was effected by coupling 7 with Boc-L-homocysteine following a procedure developed in our lab for efficient preparation of SAH 5 and other analogs (manuscript in preparation). These conditions also lead to hydrolysis of the methyl ester. Oxidation of the thioether to the sulfone using urea-hydrogen peroxide and trifluoroacetic anhydride (Balicki, 1999) followed by trifluoroacetic acid deprotection yielded the 2′-3′-O-acetal linked analog 2. Thus, in addition to stability to strong acid and methanolysis (Ariza et al., 2000), we have shown that this acetal is stable to aqueous base even at slightly elevated temperatures and mild oxidation conditions. Some depurination was observed during acid deprotection, contributing to the relatively low yield of the last step. Unfortunately, we were unable to adequately separate the two diastereomers by HPLC, so the binding analysis was performed on the mixture.

In the x-ray crystal structure of the B. subtilis 118 yitJ SAM-I riboswitch aptamer, the 2′ hydroxyl of SAM 1 is observed to make two hydrogen bonds, one with the ribose sugar moiety of residue C48 and the other with the O2-carbonyl of residue U112 (Lu et al., 2010). The 3′-O-propionate linked SAH sulfone analog 3 is proposed to retain these interactions. Since the synthesis of this analog requires mono-alkylation of the 3′ hydroxyl, we first considered methods for selective protection of the 2′ hydroxyl. It was found that the TIPDS (tetraisopropyldisilyl) group, which masks both 5′ and 3′ hydroxyls (Puffer et al., 2008), was not stable to subsequent reaction conditions. Instead, it was more straightforward to react adenosine with PMB chloride directly (Benito et al., 2008), as the 2′-O-PMB protected product selectively recrystallizes from solution. If desired, the 3′-O-PMB protected product remaining in the filtrate can be recycled back to the starting material, adenosine.

While the 5′ chlorination proceeded smoothly to give 9, the remaining 3′ hydroxyl group was quite unreactive. To install the 3′-O-propionate linker, the best yields were obtained by running the reaction in neat methyl acrylate with tert-butoxide. The strong base was added in portions during the course of the reaction to avoid a second alkylation, presumably occurring on the N6 amino group of adenosine, which otherwise leads to a less polar di-alkylated byproduct observed by TLC and LCMS. Under these conditions, we observed preferential alkylation of the 3′ hydroxyl, as the sole mono-alkylated product exhibits a broad singlet integrating to two protons in the 1H NMR spectrum, which is indicative of the N6 amino group. Finally, the 3′-O-propionate linked SAH sulfone analog 3 was obtained from the standard coupling reaction, followed by oxidation to the sulfone and deprotection with acid to remove both Boc and PMB protecting groups.

Compounds 2 and 3 were designed as stable analogs of SAM with two different linkers in order to probe whether it was possible to functionalize the ligand without blocking binding to the SAM-I riboswitch. Thus, we compared the binding affinities of SAH sulfone 4 and these prototype linker-containing compounds to the B. subtilis 124 yitJ SAM-I riboswitch aptamer (Bs SAM-I) (Figure 2A). Dissociation constants for each ligand were determined by performing in-line probing analysis (Regulski and Breaker, 2008) on Bs SAM-I in the presence of different concentrations of compound (Figure 2B). The in-line cleavage pattern was identical for all compounds tested and matched that of SAM 1, with the exception that residue C11 shows increased cleavage in the presence of SAH sulfone 4 and its analogs relative to SAM 1, consistent with previous observations (Lim et al., 2006). The Kd value for SAH sulfone 4 synthesized using our method (29 nM) matched the previously reported value of 30 nM (Lim et al., 2006).

Figure 2. Prototype linker-containing analogs of SAH sulfone maintain affinity for Bs SAM-I.

Figure 2

Open in a new tab

(A) Comparison of binding affinities to the riboswitch aptamer for analogs 2 and 3 versus SAH sulfone 4 determined by in-line probing analysis. Data shown are averages for several sites of modulation normalized to the cleavage observed at maximum ligand concentration, or the inverse value (one minus normalized cleavage observed) for sites of increasing cleavage. Also shown are best-fit curves used to calculate Kd values. Error bars represent SD of values for sites of modulation analyzed from a single gel. (B) Representative in-line probing gel showing change in spontaneous cleavage pattern of Bs SAM-I in the presence of 2. Sites of modulation analyzed are marked by black triangles and several G nucleotides mapped by T1 digest are labeled. NR = no reaction, T1 = partial digest with RNase T1, -OH = partial digest with alkali. See also Figure S4.

Furthermore, we found that both 2 and 3 maintain strong binding affinity to the SAM-I riboswitch aptamer. The acetal linker reduces affinity by 4.7-fold (Kd value of ∼140 nM). This loss of affinity could be due at least partially to the absence of the 2′ hydroxyl, which is involved in hydrogen-bonding interactions (Figure 1B). In support of this hypothesis, analog 3, which maintains the 2′ hydroxyl and bears the 3′-O-propionate linker, binds the SAM-I riboswitch aptamer with higher affinity than 2 (Kd value of ∼59 nM). While both model compounds 2 and 3 retain binding affinity to the RNA, we considered the 3′-O-propionate linker to be more promising to pursue, as it has little negative effect on binding and does not generate a new stereocenter.

Fluorescently Tagged SAM Analogs Target the SAM-I Riboswitch

To prepare fluorescent analogs, we considered that installing a terminal amine on the 3′ linker would allow us to take advantage of the multitude of commercial amine-reactive fluorescent probes (Scheme 1B). Saponification of 10 followed by HATU coupling to either mono Boc-protected 1,3-diaminopropane or 1,8-diaminooctane (Chadwick et al., 2010) gave 12 and 13, respectively, showing that overall linker length can be modified by simply exchanging the diaminoalkane used in the coupling reaction. The corresponding SAH sulfone analogs were obtained after employing similar thioether formation, oxidation, and deprotection conditions as established for 3.

Mono protection of the diaminoalkane linker with Cbz also can be effected, but we were concerned with having to perform late-stage deprotection of this group in the presence of a thioether. As the Fmoc protecting group likely is not stable to the conditions for thioether formation, the Boc-protected diaminoalkanes were used. Thus, the final deprotection of 14 and 15 removed both Boc groups and revealed two potentially reactive amines. Efforts to couple to the fluorophore NHS ester in the presence of copper (II) sulfate to block the amino acid (Wiejak et al., 1999) were unsuccessful. However, the ε-amine of lysine has been shown to react with an NHS ester selectively over the α-amine (Chauhan, 2009), so a similar effect was expected with 16 and 17. Coupling of 5(6)-carboxytetramethylrhodamine (TAMRA) NHS ester with 16 gave a 70:30 mixture of C3-TAMRA regioisomers as monitored by HPLC, but with 17, which has a longer alkyl chain, only the desired regioisomer (C8-TAMRA) was observed by HPLC. Assignment of the regiochemistry of C8-TAMRA was confirmed by correlation to the 1H NMR chemical shifts for the more extensively characterized Cy5-tagged compound 18.

Other riboswitch-small molecule interactions have been interrogated using a radiolabeled ligand in native polyacrylamide gel electrophoresis (PAGE) (Shanahan et al., 2011). Similarly, binding of the TAMRA-tagged SAM analogs to Bs SAM-I was analyzed using native PAGE. Both fluorescently tagged analogs non-covalently labeled Bs SAM-I, however the fluorescent ligand-riboswitch complex was detectable at lower RNA concentrations for C8-TAMRA (250 nM) than for C3-TAMRA (5 μM) (Figure S1). Inline probing analysis corroborates the native PAGE results (Figure S1), which together suggest that the longer C8 linker is better tolerated for binding to the riboswitch aptamer.

Coupling of sulfo-Cy5 NHS ester with 17, which was analyzed by 2D COSY experiments in order to assign peaks in the 1H NMR (Figure S2), yielded the desired single regioisomer of C8-Cy5 18. Consistent with fluorophore coupling to the C8 amine linker instead of the amino acid, we observed a decrease in the number of protons attached to the C8 amine from three to one (Figure S3). In addition, the chemical shift of the adjacent methylene changes from 2.7 to 3.0 ppm (Figure S3), which was also observed for acetylated 16 prepared from mono-acetyl diaminopropane.

Interestingly, the binding affinities of fluorescent ligands C8-TAMRA (Kd value of ∼3 μM) and the related C8-Cy5 18 (Kd value of ∼2.8 μM) to Bs SAM-I (Figure 3A, Figure S1) are similar, in spite of the latter having an additional alkyl spacer. Both exhibit weaker than expected binding to the riboswitch aptamer. This effect is not due to the presence of the linker, as 17 and acetylated 16 both retain relatively strong binding affinities (Kd values of ∼142 nM and 102 nM, respectively) (Figure S4).

Figure 3. P3 stem length analysis reveals major effect on relative ligand binding affinity.

Figure 3

Open in a new tab

On left, sequences and secondary structure models mapped with in-line probing patterns are shown for (A) Bs SAM-I, (B) Pi SAM-I, and (C) Tt SAM-I. P3 stem lengths are labeled next to each sequence. On right, corresponding in-line probing data and best-fit curves used to calculate Kd values are shown. Error bars represent SD of values for sites of modulation analyzed from a single gel. Relative fold-change in affinity between compounds 4 and 18 is labeled in red. (D) Plot of the difference in free energy for the dissociation of 18 and 4 versus length of the riboswitch aptamer P3 stem. See also Figure S6 and Table S1.

Riboswitch P3 Stem Length Affects Binding of a Fluorescently Tagged SAM Analog

The x-ray crystal structure of the B. subtilis 118 yitJ SAM-I riboswitch aptamer (Lu et al., 2010) shows that the P3 stem drapes over the solvent-accessible cavity leading to the ribose portion of the ligand (Figure S5A, B). Thus, we propose that steric hindrance with the P3 stem due to adding the TAMRA or Cy5 fluorophores to the flexible alkyl linker leads to the observed detrimental effect on riboswitch binding. Alternatively, the fluorophores may affect the conformation of the SAM analogue in solution, such that binding to the riboswitch aptamer is less favorable. It also is possible that both factors contribute to the loss in affinity for the fluorescently tagged ligand.

To determine whether the P3 stem contributes to steric clash, we examined three previously characterized SAM-I riboswitch aptamers with varying P3 stem lengths (Figure 3). The published consensus sequence for the SAM-I riboswitch aptamer, which is based on 4757 representative sequences from different bacteria, shows that the length of the P3 stem is highly variable (Barrick and Breaker, 2007). The Bs SAM-I riboswitch aptamer has a total P3 stem length of 40 nucleotides, while the 115 nucleotide riboswitch aptamer sequence from the O-acetylhomoserine sulfhydrylase transcript of P. irgensii (Pi SAM-I) has a total P3 stem length of 29 nucleotides. Both of these RNAs exhibit similar binding affinities for SAM (Kd values of 4 nM and 3 nM, respectively) (Winkler et al., 2003, Karns et al., 2013). In addition, we analyzed the crystallography construct used to obtain the T. tengcongensis SAM-I riboswitch aptamer structure (Tt SAM-I) (Montange and Batey, 2006). This 94 nucleotide sequence derived from the metF transcript has an artificially shortened P3 stem of 19 nucleotides (Figure S5) and poorer binding affinity for SAM (Kd of 130 nM as determined by ITC) (Montange and Batey, 2006).

To deconvolute the effect of sulfone substitution from the effect of the fluorophore tag, the affinities for each riboswitch aptamer to SAH sulfone 4 and fluorescently tagged 18 were measured by the in-line probing assay (Figure 3 and S6). Interestingly, the sulfone substitution had variable effects on binding affinity to the different riboswitch aptamers (Table S1). Comparing the binding affinities for SAM and SAH sulfone, replacement of the sulfonium group with the sulfone leads to similar fold increases in dissociation constants for the B. subtilis and T. tengcongensis riboswitch aptamers (7 and 6-fold, respectively). This result matches the observation that the ligand binding pockets for these two aptamers are virtually superimposable (Lu et al., 2010). In contrast, there is a much larger effect of sulfone substitution for the P. irgensii riboswitch aptamer (25-fold), but the cause of this is not entirely clear without further structural information. As expected, the effect of the sulfone substitution does not correspond to the P3 stem length, which does not play a direct role in recognition of the sulfonium group (Montange and Batey, 2006; Lu et al., 2010).

In contrast, comparison of dissociation constants for 4 and 18 shows that addition of the fluorophore tag has a detrimental effect on binding affinity that increases with P3 stem length (Figure 3 and Table S1). Strikingly, this correlation is linear (Figure 3D), and together with analysis of the two crystal structures (Figure S5) support the hypothesis that a longer P3 stem may cause steric interference with binding of the fluorescently tagged SAM analog. However, this effect is masked by differences in ligand recognition, as the P3 stem also is involved in recognition of the adenosine base of SAM, so the binding affinity of 18 for Tt SAM-I is still poorer than for Bs SAM-I. Similarly, we have observed that SAM-I riboswitch aptamers with naturally short P3 stems also exhibit poor binding affinity for 18 (Figure S5C, D).

The above results do not eliminate the possibility that addition of the linked fluorophore changes the conformation of the SAM analog in solution. It was previously observed that the fluorescence of flavin adenine dinucleotide (FAD) is decreased relative to flavin mononucleotide (FMN) due to the adenine moiety stacking with the isoalloxazine ring (Li and Glusac, 2008). Similarly, we considered that if Cy5 has the propensity to stack with the adenine portion of the SAM analog, one would observe an increase in fluorescence going from this conformation in solution to unstacked upon binding to the riboswitch aptamer. However, there was little difference between the fluorescence intensity of 18 upon addition of water or Bs SAM-I (Figure S7), which suggests that binding does not significantly alter ligand conformation. For the same samples, formation of the riboswitch-ligand complex was confirmed by an observed increase in fluorescence polarization (FP).

We further considered that if the fluorescently tagged ligand was prone to aggregation, one would observe an increase in fluorescence polarization upon titration of the ligand. However, while fluorescence intensity increases linearly with respect to compound concentration, the fluorescence polarization was insensitive to changes in ligand concentration up to the value used in binding experiments (Figure S7). The latter observation reveals that the fluorescently tagged ligand is not aggregated under our assay conditions. Thus, these data together suggest that the length of P3 stem of the SAM-I riboswitch affects accommodation of the fluorophore tag attached via the ribose 3′ hydroxyl, however we cannot rule out that other differences in the aptamer domains contribute to disrupt binding to 18.

Development of Direct and Competitive Fluorescence Polarization (FP) Binding Assays for SAM-I Riboswitches

With the C8-Cy5 fluorescent ligand 18 in hand, we sought to develop an FP binding assay for SAM-I riboswitches. Previous analysis by the in-line probing assay showed that the fluorescently tagged ligand binds to different SAM-I riboswitch aptamers (Figure 3). Correspondingly, we observed formation of the riboswitch-ligand complex by increase in FP signal upon titration of the WT Bs SAM-I RNA, while no change in FP signal was observed upon titration of a mutant Bs SAM-I RNA harboring a disruptive mutation in the P2 stem (Heppell and Lafontaine, 2008) (Figure 4A, B). Furthermore, the dissociation constants determined from the FP binding assay (4.5 ± 1.2 μM and 2.6 ± 0.1 μM for Bs and Pi SAM-I, respectively) are in good agreement with those determined by in-line probing assays.

Figure 4. Application of C8-Cy5 18 to fluorescence polarization binding assays with multiple SAM-I riboswitch aptamers.

Figure 4

Open in a new tab

Fluorescence polarization of 18 is shown upon titration with (A) WT Bs SAM-I, (B) P2 mutant Bs SAM-I, (C) Pi SAM-I, and (D) Sa SAM-I. Different symbols indicate independent replicates. Change in polarization (ΔP) was normalized to the maximal difference observed at the highest RNA concentration. The reported Kd value is the average of those calculated from best-fit curves with margin of error reported as SD. See also Figure S1.

S. aureus is the causative pathogenic agent of a significant portion of hospital-acquired infections (Otto, 2012). The S. aureus metK gene encodes S-adenosylmethionine synthetase, which has been shown to be essential in E. coli (Wei and Newman, 2002) and B. subtilis (Kobayashi et al., 2003) and is predicted to be essential in S. aureus according to shotgun antisense RNA analysis (Forsyth et al., 2002). The expression of this gene is regulated by a computationally predicted SAM-I riboswitch, which represents a potential target for the development of new antibiotics (Blount and Breaker, 2006). Using the FP binding assay, we found that the S. aureus 114 metK SAM-I riboswitch aptamer (Sa SAM-I) binds to the fluorescently tagged SAM analog 18 with a Kd of 5.2 ± 0.6 μM, which is similar to the dissociation constants determined for the other two SAM-I riboswitches.

We next used 18 to develop a competitive ligand displacement assay for targeting SAM-I riboswitches. The ability of SAH, SAH sulfone, and adenosine to displace 18 from Bs SAM-I and Sa SAM-I was assayed (Figure 5). The first two compounds are expected to be competitive ligands, and indeed titration of these compounds leads to a decrease in FP signal. In contrast, adenosine does not bind Bs SAM-I at concentrations up to 25 μM (Winkler et. al. 2003), and our FP results are consistent with it being a non-binder (Figure 5C, F). The data were fit to equations for complete competitive binding (Roehrl et al., 2004) and IC50 values were calculated for SAH and SAH sulfone using nonlinear regression analysis. The IC50 values correlate to but do not match the Kd values for these two ligands (30 and 400 nM, respectively, for Bs SAM-I) (Lim et al., 2006). The difference is due to the recommended design of competition experiments, in which the starting polarization represents 50-80% of the maximal value (Kishor et al., 2013). Under these conditions, total RNA concentration is high and the IC50 values can be overestimations of the Kd values (Cheng and Prusoff, 1973, Swillens, 1995). Nevertheless, the IC50 values for SAH and SAH sulfone with both riboswitch aptamers show the expected trend, which is that SAH is a poorer affinity ligand than SAH sulfone. These results suggest that the fluorescent ligand analog 18 can be used to screen compound libraries for lead compounds that bind to Sa SAM-I, a potential antibiotic target.

Figure 5. Competitive ligands displace C8-Cy5 18 in a dose-dependent manner as observed by fluorescence polarization.

Figure 5

Open in a new tab

Calculated fraction bound are graphed for 18 upon displacement by SAH sulfone 4 (white boxes), SAH 5 (black boxes), or adenosine (grey boxes) from (A-C) Bs SAM-I and (D-F) Sa SAM-I. Different symbols indicate independent replicates. The reported IC50 value is the average of those calculated from nonlinear regression analysis according to a single-site competition model with margin of error reported as SD. See also Figure S5.

To further validate that the FP competitive ligand displacement assay can be adapted for high-throughput screening, we determined the Z′ factor for performing the assay in a 384-well plate in place of the single cuvette format. The Z′ factor provides a measure of assay quality for HTS and accounts for the dynamic range and well-to-well variation of an assay (Zhang et al., 1999). An excellent HTS assay by industry standards has a Z′ factor between 0.5-1.0, however HTS assays with Z′ factors between 0.2-0.6 can still be informative (Zhang et al., 1999).

To determine Z′ factors for the competition assay, 16 well replicates on a 384-well plate containing either no competitor or 100 μM of SAM were evaluated in the presence of Bs, Sa, or Pi SAM-I and 18 by FP signal. Unfortunately, calculated Z′ values for Bs and Sa SAM-I were 0.31 and 0.22, respectively (Figure 6A, B). We hypothesized that the low Z′ values for these RNAs were attributed to the modest changes in FP signal (ΔmP) for Bs and Sa SAM-I upon competitive displacement of 18 (61 and 46 mP, respectively, as measured by platereader, similar to the overall signal change observed for the cuvette-based assay, see Table S2). In contrast, the Pi SAM-I gave the largest ΔmP upon competitive displacement of 18 (110 ± 22 mP), and consistent with this result, the Z′ value for Pi SAM-I was calculated as 0.57 ± 0.01, representing an excellent HTS assay (Figure 6C).

Figure 6. Z′ values determined through 384-well format fluorescence polarization competitive ligand displacement assay.

Figure 6

Open in a new tab

Fluorescence polarization of 18 graphed for 16 well replicates containing (A) Bs, (B) Sa, and (C) Pi SAM-I with no competitor or 100 μM of SAM, SAH sulfone, or adenosine. The dashed black line indicates the mean (μ) and the solid black bracket indicates the standard deviation (SD) for each replicate set. Also shown is the signal window for identifying hits defined by μ - 3SD for no competitor (red dashed line) and μ + 3SD for the SAM control (blue dashed line). The Z value shown is the average of 3 plate replicates for Pi, or single plate replicates for Bs and Sa. See also Figure S7 and Table S2.

In addition, we evaluated 16 well replicates of two other compounds, SAH sulfone or adenosine, at 100 μM concentrations (Figure 6). Interestingly, at this high concentration, we observe some minimal binding for adenosine to the SAM-I riboswitch aptamers, which is consistent with the cuvette-based assay results (Figure 5C, F). While SAH sulfone lies beyond the “no hit” threshold as defined by 3 standard deviations (SD) away from the mean (μ) of the no competitor sample, adenosine does not meet this threshold and represents a non-hit compound. In contrast, SAH sulfone also lies beyond the “hit” threshold as defined by 3 SD away from μ of the positive control, SAM, so represents a true competitive ligand. Similar ligand binding results are observed for each of the RNAs, but only the assay employing Pi SAM-I has a wide signal window for distinguishing between hits and non-hits, as reflected by a demonstrated Z′ factor in the range for a statistically reliable high-throughput screen.

Discussion

Fluorescence polarization has been widely adopted as a high-throughput assay, as screening experiments can be fully automated using commercially available instrumentation (Jameson and Ross, 2010). FP is commonly employed for analysis of protein-ligand and protein-nucleic acid interactions (Moerke, 2009) and has been applied to detect metabolite-dependent activation of the GlmS riboswitch-ribozyme (Mayer and Famulok, 2006). However, prior to our study, it had not yet been explored for standard riboswitch-ligand interactions.

For select riboswitch classes, the cognate ligand has intrinsic fluorescence (Serganov et al., 2009) or can be directly substituted with a fluorescent analog (Budhathoki et al., 2012; Ottink et al., 2010; Daldrop et al., 2011). In each case, it was found that fluorescence is quenched upon binding, and this change in fluorescence has been used to measure ligand binding and competitive displacement. We now have shown that a fluorescently tagged riboswitch ligand can be used for measuring ligand binding via fluorescence polarization. Similar to other fluorescence-based methods, the FP assay enables rapid and direct sample analysis with high sensitivity and without needing a separation step. Furthermore, the reproducibility of the FP signal is increased by its relative insensitivity to changes in fluorophore concentration (Lea and Simeonov, 2011), which improves statistical significance parameters for high-throughput formats. Also, in contrast to assays that rely on changes in fluorescence intensity, ligands can be covalently tagged with a convenient and bright fluorophore at any position that does not disrupt binding. These advantages have made FP-based competition binding assays a workhorse high-throughput screening method in the clinical and biomedical fields (Jameson and Ross, 2010), and we expect that it will prove useful for riboswitch-targeted drug discovery as well.

For FP assays, minimal effects upon binding on the quantum yield of the fluorophore are preferable, so the fluorescent tag needs to be placed outside of the ligand binding pocket. It was challenging to rationally design a fluorescently tagged ligand for the SAM-I riboswitch class, because almost every functional group of SAM is recognized by the RNA and the electropositive sulfonium center is hydrolytically unstable. Using two prototype analogs based on SAH sulfone, we explored the possibility of installing a linker to the ribose hydroxyls and found that the 3′ position tolerated functionalization. As planned, fluorophore conjugation to the alkyl amine linker was the last synthetic step, which enables modification of fluorescent properties through choice of the NHS ester reagent (e.g. TAMRA, Cy5) and minimizes the required amount of the expensive fluorophore reagent.

Although RNA-protein interactions typically display larger changes in fluorescence polarization, the FP signal for 18 bound to SAM-I riboswitch aptamers is similar to what has been observed for other RNA-small molecule interactions (Tok et al., 1999). Interestingly, under the conditions of the cuvette-based FP assay, the signal for 18 bound to Bs SAM-I is larger than for Sa SAM-I (68 versus 43 mP, respectively, see Table S2), in spite of the two RNAs having similar molecular weights and structural folds. Importantly, the dynamic range of the FP signal of 18 was sufficient for determination of Kd and IC50 values for ligands and competitors, respectively, for both SAM-I riboswitch aptamers.

To translate the FP competition assay from cuvette-based to 384-well format, we found it necessary to improve the dynamic range of the FP signal for increased statistical significance as measured by the Z′ factor. This was accomplished using Pi SAM-I, which had a much larger ΔmP (110 ± 22 mP) compared to Sa SAM-I (∼46 mP). Thus, fluorescently tagged 18 enabled facile screening of multiple SAM-I riboswitch aptamer sequences for assay optimization. While Pi SAM-I itself is not an interesting target for drug discovery, we expect that compounds that bind this RNA found by HTS would be a promising set of candidate ligands for screening against Sa SAM-I.

Preliminary hits from a high-throughput binding assay like the one we have developed would need to be further validated, as binding alone does not necessarily confer modulation of the riboswitch regulatory function. However, our assay has the potential to identify compounds that bind within the ligand binding pocket of the riboswitch, as the assay signal is correlated to displacement of the fluorescently tagged SAM analog. Thus there is the strong likelihood of identifying competitor compounds that stabilize the P1 in a manner analogous to the cognate ligand SAM. In addition, two SAM-I expression platforms recently were shown to trigger transcriptional termination in the presence of either theophylline or tetracycline, when the expression platforms were fused to the corresponding in vitro evolved aptamer (Ceres et al., 2013). The results from that study suggest that a simple binding event is sufficient to induce regulation through stabilization of the P1 helix. Nevertheless, it is expected that compound leads could be further validated using more low-throughput functional screens, such as the molecular beacon assay (Chinnappan et al., 2013) or the two-piece switching assay recently developed for the Thermoanaerobacter tengcongensis metF SAM-I (Hennelly and Sanbonmatsu, 2011).

In this study, we have demonstrated that there is a signal threshold for adapting FP assays for RNA-small molecule interactions to the HTS format. The FP assay requires optimization of the fluorescently tagged ligand and we further showed that it may be beneficial to screen related RNA targets in the HTS format. We recommend a practical limit for ΔmP in the 384-well format of above 80 mP when using a manual multichannel pipette, as the SD between replicates is typically 10-15%. However, employing an automated pipetting system should decrease SD to 3-4% and allow the ΔmP to be considerably lower (∼40 mP) (Mayer and Famulok, 2006). We expect that these results may inform FP-based HTS screening efforts for small molecule drug candidates against other therapeutic RNA targets in general (Thomas and Hergenrother, 2008).

Although knowledge of the structure of the riboswitch-ligand complex aided in the design of the Cy5-tagged analog 18, the binding affinity of this compound was poorer than expected given the results for the linker-containing prototype 3. However, individual members of the SAM-I riboswitch class with higher affinity for 18 may exist. The FP assay readily allows screening for such variants in a riboswitch aptamer library, as it does not require chemical modification or labeling of each RNA. Furthermore, additional riboswitch classes that bind SAM as the cognate ligand have been described (Gilbert et al., 2008; Lu et al., 2008; Weinberg et al., 2008) that may bind 18 as well. Experiments for targeting other SAM riboswitch classes are currently in progress.

Additionally, SAM-binding proteins such as histone methyltransferases (HMTs) represent interesting therapeutic targets and have been the focus of drug discovery efforts in recent years. For example, the lysine methyltransferase SET7/9 has generated interest as a drug target for type II diabetes, AIDS, and hormone-dependent breast cancer (Wagner and Jung, 2012). Also, the G9a methyltransferase is upregulated in lung, prostate, and hepatocellular carcinoma (Wagner and Jung, 2012). X-ray crystal structures of each of these enzymes with inhibitor bound suggest that the SAM ribose hydroxyls are solvent accessible when bound to the protein (PDB codes 4E47 for SET7/9 and 3K5K for G9a) (Walker J.R. unpublished, Liu et al., 2009). Furthermore, hydrogen bonding interactions to the ribose hydroxyls are not conserved in SET domain HMTs (Campagna-Slater et al., 2011). While SAH sulfone 4 does not inhibit SET7/9 and G9a (Minkui Luo, personal communication), it has been previously shown that these enzymes are inhibited by their byproduct SAH 5 (Ibanez et al., 2010). Thus, from intermediate 15 of our synthetic route, one could access a fluorescently tagged SAH analog akin to 18 that potentially could be used to identify inhibitor compounds in an FP-based ligand displacement assay that bind in the cofactor binding site, as opposed to the substrate binding site. Such inhibitors would likely be broad spectrum rather than selective inhibitors, which may be desirable for therapeutic development (Wagner and Jung, 2012).

Significance

In conclusion, we have used structure-guided rational design to develop a fluorescently tagged SAM analog that can be used in FP-based screening of compound libraries for binding to SAM-I riboswitches or other relevant SAM-binding RNAs and proteins. Our synthetic route enables facile modification of the ligand and installation of other functional probes such as biotin, terminal alkyne, or a crosslinking agent for synthetic biology and chemical biology applications. Furthermore, we demonstrate the utility of analyzing multiple riboswitch sequences for the development of a statistically reliable FP-based HTS assay for identifying competitive binders. In addition to the ligand tagging approach being generally applicable to target many riboswitches, the adaptation of HTS methodologies to interrogate RNA-small molecule interactions will be critical to efforts to target other therapeutically relevant RNA structures.

Experimental Procedures

Preparation of DNA Constructs

Sequences for all DNA templates and primers used in this study are provided in Table S2. DNA templates corresponding to riboswitch sequences from different organisms were either amplified by PCR from the genomic DNA or from a single-stranded oligonucleotide ordered from IDT. The resulting PCR products were cloned into the pCR2.1-TOPO vector (Invitrogen) following manufacturer instructions and sequence confirmed.

In Vitro Transcription of RNAs

DNA templates containing the extended T7 promoter sequence and two additional G nucleotides for increased transcription efficiency were transcribed in vitro using T7 RNA polymerase (New England Biolabs) at 37 °C for 2-3 h. RNA was purified in a denaturing (7.5 M urea) 6% polyacrylamide gel and was extracted from gel pieces using Crush Soak buffer (10 mM Tris-HCl, pH 7.5, 200 mM NaCl and 1 mM EDTA, pH 8.0). RNAs were precipitated with ethanol and resuspended in TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA). RNA concentrations were measured by the absorbance at 260 nm using a microvolume UV-Vis spectrophotometer (Nanodrop ND-8000). Concentration values were corrected for hypochromicity by measuring the absorbance at 260 nm after hydrolysis with 50 mM sodium carbonate at 95 °C for 1-2 h (Wilson et. al., submitted).

In-Line Probing Assays

In-line probing analysis was performed following standard procedures (Regulski and Breaker, 2008, see supplemental experimental procedures). The signal intensity for multiple sites of ligand-induced modulation was normalized to the observed value at the highest ligand concentration, and fraction cleaved was taken as the average of the data analyzed for each modulation site. Kd values were determined by fitting the experimental data to a best-fit curve for a 1:1 RNA-ligand complex.

Cuvette-Based Fluorescence Polarization Binding Assays

Fluorescence polarization readings were carried out using a QuantaMaster fluorescence spectrofluorometer (excitation 646 nm, emission 662 nm, Photon Technology International). Samples were prepared in 50 μL of TBM buffer (90 mM Tris base, 89 mM boric acid, 10 mM MgCl2, pH 7.0) containing 1 μM of C8-Cy5, and saturation binding experiments were performed with RNA concentrations ranging from 0 - 40 μM. RNA was added successively to the sample cuvette, and concentration values were corrected for added volume. Samples were equilibrated at 30 °C for 2 min prior to each FP measurement using tubing connecting the cuvette holder to a water bath.

For competitive ligand displacement assays, samples were prepared in 50 μL of TBM buffer containing 1 μM of C8-Cy5 and 10 μM of either Bs SAM-I or Sa SAM-I RNA, corresponding to ∼65% fraction bound. Competitor compounds were added successively to the sample cuvette, and concentration values were corrected for added volume. Competitor binding experiments to Bs SAM-I were performed with compound concentrations ranging from 0 - 92 μM for 4 and 0 - 272 μM for 5. For Sa SAM-I, the compound concentration ranges were 0 - 94 μM for 4 and 0- 362 μM for 5. See also supplemental experimental procedures.

384-Well Fluorescence Polarization Competition Assays

For 384-well FP competitive ligand displacement assays, samples were prepared from a master mix of TBM buffer containing 1 μM of C8-Cy5 and 10 μM of either Bs, Pi, or Sa SAM-I RNA, with no competitor or with 100 μM of SAM, SAH sulfone, or adenosine. A multichannel repeat pipettor was used to manually add 50 μL of samples into 16 replicate wells on a 384-well plate (Corning low volume flat bottom polystyrene nonbinding surface). FP measurements were taken on a Molecular Devices FlexStation 3 Plate Reader at the UCSF Small Molecule Discovery Center.

Supplementary Material

01

Highlights.

  • Synthesis of stable and fluorescent analogs of SAM are described

  • Sulfone substitution and Cy5 label affect binding to different SAM-I riboswitches

  • Fluorescent ligand used in direct and competitive fluorescence polarization assays

  • Quality of the competitive assay for high-throughput screening was validated

Acknowledgments

We would like to acknowledge Michelle Arkin and Chris Wilson at the UCSF Small Molecule Discovery Center for assistance with HTS assay development and Minkuo Luo at Memorial Sloan-Kettering for sharing data on the methyltransferase inhibitor study. This work is supported in part by an NIH New Innovator Award (1DP2 OD008677) to M.C.H., a Career Award at the Scientific Interface from the Burroughs Wellcome Fund to M.C.H., the Aldo DeBenedictis fellowship from the UC Berkeley Department of Chemistry to S.F.H., and an NIH Chemistry-Biology Interface Institutional Training Grant (T32 GM066698). The Synthetic Biology Institute is supported by Agilent Technologies as an industry partner.

Footnotes

Supplemental Information: Supplemental Information includes all synthetic procedures and compound characterization data, seven figures, and two tables.

Publisher's Disclaimer: 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 citable 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.

References

  1. Ariza X, Costa AM, Faja M, Pineda O, Vilarrasa J. New protecting groups for 1,2-diols (Boc- and Moc-ethylidene). Cleavage of acetals with bases. Org Lett. 2000;2:2809–2811. doi: 10.1021/ol0062289. [DOI] [PubMed] [Google Scholar]
  2. Baird NJ, Ferré-D'Amaré AR. Idiosyncratically tuned switching behavior of riboswitch aptamer domains revealed by comparative small-angle X-ray scattering analysis. RNA. 2010;16:598–609. doi: 10.1261/rna.1852310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Balicki R. A mild and efficient procedure for the oxidation of organic sulfides to sulfones by use of urea-hydrogen peroxide/trifluoroacetic anhydride system. Synthetic Commun. 1999;29:2235–2239. [Google Scholar]
  4. Barrick JE, Breaker RR. The distributions, mechanisms, and structures of metabolite-binding riboswitches. Genome Biol. 2007;8:1–19. doi: 10.1186/gb-2007-8-11-r239. R239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bartholomew CH, Agrawal PK, Katzer JR. Sulfur poisoning of metals. Adv Catal. 1982;31:135–242. [Google Scholar]
  6. Batey RT. Recognition of S-adenosylmethionine by riboswitches. Wiley Interdiscip Rev RNA. 2011;2:299–311. doi: 10.1002/wrna.63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Benito D, Matheu MI, Morère A, Díaz Y, Castillón S. Towards the preparation of 2″-deoxy-2″-fluoro-adenophostin A. Study of the glycosylation reaction. Tetrahedron. 2008;64:10906–10911. [Google Scholar]
  8. Blount KF, Breaker RR. Riboswitches as antibacterial drug targets. Nat Biotechnol. 2006;24:1558–1564. doi: 10.1038/nbt1268. [DOI] [PubMed] [Google Scholar]
  9. Blount K, Puskarz I, Penchovsky R, Breaker RR. Development and application of a high-throughput assay for glmS riboswitch activators. RNA Biol. 2006;3:77–81. doi: 10.4161/rna.3.2.3102. [DOI] [PubMed] [Google Scholar]
  10. Blount KF, Wang JX, Lim J, Sudarsan N, Breaker RR. Antibacterial lysine analogs that target lysine riboswitches. Nat Chem Biol. 2007;3:44–49. doi: 10.1038/nchembio842. [DOI] [PubMed] [Google Scholar]
  11. Breaker RR. Riboswitches and the RNA world. Cold Spring Harb Perspect Biol. 2012;4:a003566. doi: 10.1101/cshperspect.a003566. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Breaker RR. Prospects for riboswitch discovery and analysis. Mol Cell. 2011;43:867–879. doi: 10.1016/j.molcel.2011.08.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Budhathoki P, Bernal-Perez LF, Annunziata O, Ryu Y. Rationally-designed fluorescent lysine riboswitch probes. Org Biomol Chem. 2012;10:7872–7874. doi: 10.1039/c2ob26160j. [DOI] [PubMed] [Google Scholar]
  14. Campagna-Slater V, Mok MW, Nguyen KT, Feher M, Najmanovich R, Schapira M. Structural chemistry of the histone methyltransferases cofactor binding site. J Chem Int Model. 2011;51:612–623. doi: 10.1021/ci100479z. [DOI] [PubMed] [Google Scholar]
  15. Ceres P, Garst AD, Marcano-Velazquez JG, Batey RT. Modularity of select riboswitch expression platforms enables facile engineering of novel genetic regulatory devices. ACS Synth Biol. 2013;2:463–472. doi: 10.1021/sb4000096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chadwick J, Jones M, Mercer AE, Stocks PA, Ward ST, Park BK, O'Neill PM. Design, synthesis and antimalarial/anticancer evaluation of spermidine linked artemisinin conjugates designed to exploit polyamine transporters in Plasmodium falciparum and HL-60 cancer cell lines. Bioorg Med Chem. 2010;18:2586–2597. doi: 10.1016/j.bmc.2010.02.035. [DOI] [PubMed] [Google Scholar]
  17. Chauhan SS. Enantioselective synthesis of (L)-Fmoc-α-Me-Lys(Boc)-OH via diastereoselective alkylation of oxazinone as a chiral auxiliary. Tetrahedron Lett. 2009;50:6913–6915. [Google Scholar]
  18. Chen L, Cressina E, Leeper FJ, Smith AG, Abell C. A fragment-based approach to identifying ligands for riboswitches. ACS Chem Biol. 2010;5:355–358. doi: 10.1021/cb9003139. [DOI] [PubMed] [Google Scholar]
  19. Cheng Y, Prusoff WH. Relationship between the inhibition constant (Ki) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochem Pharmacol. 1973;22:3099–3108. doi: 10.1016/0006-2952(73)90196-2. [DOI] [PubMed] [Google Scholar]
  20. Chinnappan R, Dubé A, Lemay JF, Lafontaine DA. Fluorescence monitoring of riboswitch transcription regulation using a dual molecular beacon assay. Nucleic Acids Res. 2013;41:e106. doi: 10.1093/nar/gkt190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Cressina E, Chen L, Abell C, Leeper FJ, Smith AG. Fragment screening against the thiamine pyrophosphate riboswitch thiM. Chem Sci. 2011;2:157–165. [Google Scholar]
  22. Daldrop P, Reyes FE, Robinson DA, Hammond CM, Lilley DM, Batey RT, Brenk R. Novel ligands for a purine riboswitch discovered by RNA-ligand docking. Chem Biol. 2011;18:324–335. doi: 10.1016/j.chembiol.2010.12.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. De Silva C, Walter NG. Leakage and slow allostery limit performance of single drug-sensing aptazyme molecules based on the hammerhead ribozyme. RNA. 2009;15:76–84. doi: 10.1261/rna.1346609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Deigan KE, Ferré-D'Amaré AR. Riboswitches: discovery of drugs that target bacterial gene-regulatory RNAs. Accounts Chem Res. 2011;44:1329–1338. doi: 10.1021/ar200039b. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Forsyth RA, Haselbeck RJ, Ohlsen KL, Yamamoto RT, Xu H, Trawick JD, Wall D, Wang L, Brown-Driver V, Froelich JM, Kedar GC, King P, McCarthy M, Malone C, Misiner B, Robbins D, Tan Z, Zhu Z, Carr G, Mosca DA, Zamudio C, Foulkes JG, Zyskind JW. A genome-wide strategy for the identification of essential genes in Staphylococcus aureus. Mol Microbiol. 2002;43:1387–1400. doi: 10.1046/j.1365-2958.2002.02832.x. [DOI] [PubMed] [Google Scholar]
  26. Furukawa K, Gu H, Sudarsan N, Hayakawa Y, Hyodo M, Breaker RR. Identification of ligand analogues that control c-di-GMP riboswitches. ACS Chem Biol. 2012;7:1436–1443. doi: 10.1021/cb300138n. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Garst AD, Edwards AL, Batey RT. Riboswitches: Structures and Mechanisms. Cold Spring Harb Perspect Biol. 2011;3:a003533. doi: 10.1101/cshperspect.a003533. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Gilbert SD, Pambo RP, Tyne DV, Batey RT. Structure of the SAM-II riboswitch bound to S-adenosylmethionine. Nat Struct Mol Biol. 2008;15:177–182. doi: 10.1038/nsmb.1371. [DOI] [PubMed] [Google Scholar]
  29. Hafner M, Vianini E, Albertoni B, Marchetti L, Grune I, Gloeckner C, Famulok M. Displacement of protein-bound aptamers with small molecules screened by fluorescence polarization. Nat Protoc. 2008;3:579–587. doi: 10.1038/nprot.2008.15. [DOI] [PubMed] [Google Scholar]
  30. Haller A, Altman RB, Soulière MR, Blanchard SC, Micura R. Folding and ligand recognition of the TPP riboswitch aptamer at single-molecule resolution. Proc Natl Acad Sci USA. 2013;110:4188–4193. doi: 10.1073/pnas.1218062110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Haller A, Soulière MF, Micura R. The dynamic nature of RNA as key to understanding riboswitch mechanisms. Accounts Chem Res. 2011;44:1339–1348. doi: 10.1021/ar200035g. [DOI] [PubMed] [Google Scholar]
  32. Ham YW, Humphreys DJ, Choi S, Dayton DL. Rational design of SAM analogues targeting SAM-II riboswitch aptamer. Bioorg Med Chem Lett. 2011;21:5071–5074. doi: 10.1016/j.bmcl.2011.04.033. [DOI] [PubMed] [Google Scholar]
  33. Hennelly SP, Sanbonmatsu KY. Tertiary contacts control switching of the SAM-I riboswitch. Nucleic Acids Res. 2011;39:2416–2431. doi: 10.1093/nar/gkq1096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Heppell B, Blouin S, Dussault AM, Mulhbacher J, Ennifar E, Penedo JC, Lafontaine DA. Molecular insights into the ligand-controlled organization of the SAM-I riboswitch. Nat Chem Biol. 2011;7:384–392. doi: 10.1038/nchembio.563. [DOI] [PubMed] [Google Scholar]
  35. Heppell B, Lafontaine DA. Folding of the SAM aptamer is determined by the formation of a K-turn-dependent pseudoknot. Biochemistry. 2008;47:1490–1499. doi: 10.1021/bi701164y. [DOI] [PubMed] [Google Scholar]
  36. Ibanez G, McBean JL, Astudillo YM, Luo M. An enzyme-coupled ultrasensitive luminescence assay for protein methyltransferases. Anal Biochem. 2010;401:203–210. doi: 10.1016/j.ab.2010.03.010. [DOI] [PubMed] [Google Scholar]
  37. Jameson DM, Ross JA. Fluorescence polarization/anisotropy in diagnostics and imaging. Chem Rev. 2010;110:2685–2708. doi: 10.1021/cr900267p. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Karns K, Vogan JM, Qin Q, Hickey SF, Wilson SC, Hammond MC, Herr AE. Microfluidic screening of electrophoretic mobility shifts elucidates riboswitch binding function. J Am Chem Soc. 2013;135:3136–3143. doi: 10.1021/ja310742m. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Kim JN, Blount KF, Puskarz I, Lim J, Link KH, Breaker RR. Design and antimicrobial action of purine analogues that bind guanine riboswitches. ACS Chem Biol. 2009;4:915–927. doi: 10.1021/cb900146k. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Kishor A, Brewer G, Wilson GM. Analyses of RNA-ligand interactions by fluorescence anisotropy. Biophysical Approaches to Translational Control of Gene Expression. Biophysics for the Life Sciences. 2013;1:173–198. [Google Scholar]
  41. Kobayashi K, et al. Essential Bacillus subtilis genes. Proc Natl Acad Sci USA. 2003;100:4678–4683. doi: 10.1073/pnas.0730515100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Lea WA, Simeonov A. Fluorescence polarization assays in small molecule screening. Expert Opin Drug Discov. 2011;6:17–32. doi: 10.1517/17460441.2011.537322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Li G, Glusac KD. Light-triggered proton and electron transfer in flavin cofactors. J Phys Chem A. 2008;112:4573–4583. doi: 10.1021/jp7117218. [DOI] [PubMed] [Google Scholar]
  44. Li J, Wei H, Zhou MM. Structure-guided design of a methyl donor cofactor that controls a viral histone H3 lysine 27 methyltransferase activity. J Med Chem. 2011;54:7734–7738. doi: 10.1021/jm201000j. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Lim J, Winkler WC, Nakamura S, Scott V, Breaker RR. Molecular-recognition characteristics of SAM-binding riboswitches. Angew Chem Int Ed. 2006;45:964–968. doi: 10.1002/anie.200503198. [DOI] [PubMed] [Google Scholar]
  46. Liu F, Chen Z, Allali-Hassani A, Quinn AM, Wasney GA, Dong A, Barsyte D, Kozieradzki I, Senisterra G, Chau I, Siarheyeva A, Kireev DB, Jadhav A, Herold JM, Frye SV, Arrowsmitch CH, Brown PJ, Simeonov A, Vedadi M, Jin J. Discovery of a 2,4-diamino-7-aminoalkoxyquinazoline as a potent and selective inhibitor of histone lysine methyltransferase G9a. J Med Chem. 2009;52:7950–7953. doi: 10.1021/jm901543m. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Lu C, Ding F, Chowdhury A, Pradhan V, Tomsic J, Holmes WM, Henkin TM, Ke A. SAM recognition and conformational switching mechanism in the Bacillus subtilis yitJ S box/SAM-I riboswitch. J Mol Biol. 2010;404:803–818. doi: 10.1016/j.jmb.2010.09.059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Lu C, Smith AM, Fuchs RT, Ding F, Rajashankar K, Henkin TM, Ke A. Crystal structures of the SAM-III/SMK riboswitch reveal the SAM-dependent translation inhibition mechanism. Nat Struct Mol Biol. 2008;15:1076–1083. doi: 10.1038/nsmb.1494. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Mayer G, Famulok M. High-throughput-compatible assay for glmS riboswitch metabolite dependence. ChemBioChem. 2006;7:602–604. doi: 10.1002/cbic.200500490. [DOI] [PubMed] [Google Scholar]
  50. Moerke NJ. Fluorescence polarization (FP) assays for monitoring peptide-protein or nucleic acid-protein binding. Curr Protoc Chem Biol. 2009;1:1–15. doi: 10.1002/9780470559277.ch090102. [DOI] [PubMed] [Google Scholar]
  51. Montange RK, Batey RT. Structure of the S-adenosylmethionine riboswitch regulatory mRNA element. Nature. 2006;441:1172–1175. doi: 10.1038/nature04819. [DOI] [PubMed] [Google Scholar]
  52. Mulhbacher J, Brouillette E, Allard M, Fortier LC, Malouin F, Lafontaine DA. Novel riboswitch ligand analogs as selective inhibitors of guanine-related metabolic pathways. PLoS Pathog. 2010;6:1–11. doi: 10.1371/journal.ppat.1000865. e1000865. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Murray SG, Hartley FR. Coordination chemistry of thioethers, selenoethers, and telluroethers in transition-metal complexes. Chem Rev. 1981;81:365–414. [Google Scholar]
  54. Ottink OM, Nelissen FHT, Derks Y, Wijmenga SS, Heus HA. Enzymatic stereospecific preparation of fluorescent S-adenosyl-L-methionine analogs. Anal Biochem. 2010;396:280–283. doi: 10.1016/j.ab.2009.09.013. [DOI] [PubMed] [Google Scholar]
  55. Otto M. MRSA virulence and spread. Cell Microbiol. 2012;14:1513–1521. doi: 10.1111/j.1462-5822.2012.01832.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Puffer B, Moroder H, Aigner M, Micura R. 2′-methylseleno-modified oligoribonucleotides for X-ray crystallography synthesized by the ACE RNA solid-phase approach. Nucleic Acids Res. 2008;36:970–983. doi: 10.1093/nar/gkm880. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Regulski EE, Breaker RR. In-line probing analysis of riboswitches. Methods Mol Biol. 2008;419:53–67. doi: 10.1007/978-1-59745-033-1_4. [DOI] [PubMed] [Google Scholar]
  58. Roehrl MHA, Wang JY, Wagner G. A general framework for development and data analysis of competitive high-throughput screens for small-molecule inhibitors of protein-protein interactions by fluorescence polarization. Biochemistry. 2004;43:16056–16066. doi: 10.1021/bi048233g. [DOI] [PubMed] [Google Scholar]
  59. Serganov A, Huang L, Patel DJ. Coenzyme recognition and gene regulation by a flavin mononucleotide riboswitch. Nature. 2009;458:233–238. doi: 10.1038/nature07642. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Serganov A, Nudler E. A decade of riboswitches. Cell. 2013;152:17–24. doi: 10.1016/j.cell.2012.12.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Shanahan CA, Gaffney BL, Jones RA, Strobel SA. Differential analogue binding by two classes of c-di-GMP riboswitches. J Am Chem Soc. 2011;133:15578–15592. doi: 10.1021/ja204650q. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Stoddard CD, Montange RK, Hennelly SP, Rambo RP, Sanbonmatsu KY, Batey RT. Free state conformational sampling of the SAM-I riboswitch aptamer domain. Structure. 2010;18:757–758. doi: 10.1016/j.str.2010.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Swillens S. Interpretation of binding curves obtained with high receptor concentrations: practical aid for computer analysis. Mol Pharmacol. 1995;47:1197–1203. [PubMed] [Google Scholar]
  64. Thomas JR, Hergenrother PJ. Targeting RNA with small molecules. Chem Rev. 2008;108:1171–1224. doi: 10.1021/cr0681546. [DOI] [PubMed] [Google Scholar]
  65. Tok JBH, Cho J, Rando RR. Aminoglycoside antibiotics are able to specifically bind the 5′-untranslated region of thymidylate synthase messenger RNA. Biochemistry. 1999;38:199–206. doi: 10.1021/bi9819428. [DOI] [PubMed] [Google Scholar]
  66. Wagner T, Jung M. New lysine methyltransferase drug targets in cancer. Nat Biotechnol. 2012;30:622–623. doi: 10.1038/nbt.2300. [DOI] [PubMed] [Google Scholar]
  67. Wang JX, Breaker RR. Riboswitches that sense S-adenosylmethionine and S-adenosylhomocysteine. Biochem Cell Biol. 2008;86:157–168. doi: 10.1139/O08-008. [DOI] [PubMed] [Google Scholar]
  68. Wei Y, Newman EB. Studies on the role of the metK gene product of Escherichia coli K-12. Mol Microbiol. 2002;43:1651–1656. doi: 10.1046/j.1365-2958.2002.02856.x. [DOI] [PubMed] [Google Scholar]
  69. Weinberg Z, Regulski EE, Hammond MC, Barrick JE, Yao Z, Ruzzo WL, Breaker RR. The aptamer core of SAM-IV riboswitches mimics the ligand-binding site of SAM-I riboswitches. RNA. 2008;14:822–828. doi: 10.1261/rna.988608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Wickiser JK, Winkler WC, Breaker RR, Crothers DM. The speed of RNA transcription and metabolite binding kinetics operate an FMN riboswitch. Mol Cell. 2005;18:49–60. doi: 10.1016/j.molcel.2005.02.032. [DOI] [PubMed] [Google Scholar]
  71. Wiejak S, Masiukiewicz E, Rzeszotarska B. A large scale synthesis of mono- and di-urethane derivatives of lysine. Chem Pharm Bull. 1999;47:1489–1490. doi: 10.1248/cpb.49.1189. [DOI] [PubMed] [Google Scholar]
  72. Wilson RC, Smith AM, Fuchs RT, Kleckner IR, Henkin TM, Foster MP. Tuning riboswitch regulation through conformational selection. J Mol Biol. 2011;405:926–938. doi: 10.1016/j.jmb.2010.10.056. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Wilson SC, Cohen DT, Hammond MC. manuscript submitted. [Google Scholar]
  74. 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;10:701–707. doi: 10.1038/nsb967. [DOI] [PubMed] [Google Scholar]
  75. Zhang J, Chung TDY, Oldenburg KR. A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J Biomol Screen. 1999;4:67–73. doi: 10.1177/108705719900400206. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

01
NIHMS569726-supplement-01.pdf (1.7MB, pdf)

RESOURCES