[Univ. of Illinois]  [School of Chemical Sciences]  [Dept. of Chemistry]  [Organic Chemistry]  [Chemical Biology]

[Journals List]  [Group Meetings]  [Group Pictures]  [CV & Contact Info]

[SKS Lab Home]  [SKS Lab People]  [SKS Lab Publications]

Fundamental studies of RNA folding and catalysis

RNA conformational changes are intimately related to the roles of RNA in biochemistry. For example, the ribosomal RNA-protein machinery that synthesizes proteins in vivo relies on catalytic RNA during protein synthesis. Of course, such catalysis requires changes in RNA conformation. We would like to understand how RNA molecules change conformation and how these conformational changes affect catalysis. Several projects in our laboratory are focused on understanding and controlling fundamental aspects of RNA folding. Some of these projects are described below, and others are included on our webpage for DNA as a macromolecular conformational constraint.

The hierarchical nature of RNA folding

The foldedness of an RNA molecule may be described in several structural levels, as is the case for proteins. The "primary structure" of an RNA molecule is its sequence of nucleotides (A, G, U, and C), just as the primary structure of a protein molecule is its sequence of amino acids. The secondary structure of an RNA molecule is dictated by Watson-Crick base pairing (A with U and G with C), which leads to double-helical regions interspersed with single-stranded loops, bulges, and other non-Watson-Crick elements. In proteins, secondary structure is primarily a-helix or b-sheet, although other secondary structure motifs can also be formed. Finally, the tertiary structure of an RNA or protein molecule—its three-dimensional shape—is formed by interactions among the various secondary structure elements. Two major differences between RNA and proteins are the roles of metal ions and the degree of hierarchy in structure formation. First, metal ions are sometimes found in protein structures and can be important for catalysis, but proteins do not generally require metal ions for their structure or function. In contrast, RNA is a polyanion and absolutely requires metal ions for structure and function. Although monovalent metal ions like Na⁺ and K⁺ contribute to RNA structure, divalent metal ions (M²⁺) are often crucial. In nature, Mg²⁺ is the divalent metal ion most commonly found associated with RNA. Second, proteins and RNA differ in the hierarchical relationship among primary, secondary, and tertiary structure. For proteins, secondary structure elements (a-helix and b-sheet) are rarely stable in isolation; indeed, it is quite an achievement to design an isolated protein secondary structure element. Formation of the three-dimensional protein structure stabilizes the secondary structure elements, and in a kinetic sense both levels of structure form at approximately the same time. However, for RNA, secondary structure is generally stable by itself and independent of three-dimensional tertiary structure. Because of this relatively strict hierarchy of RNA structure, many RNA folding experiments begin with "unfolded" RNA that has extensive secondary structure but no tertiary structure and proceed to "folded" RNA that has both secondary and tertiary structure, often by addition of metal ions such as Mg²⁺.

Fluorescence techniques to monitor RNA folding

For studies of protein folding, the naturally occuring amino acid tryptophan is often used as a fluorescent reporter group. In the simplest version of such an experiment, the fluorescence emission intensity of tryptophan increases or decreases as its environment changes due to folding of the protein to which it is attached. Therefore, the tryptophan fluorescence emission intensity correlates with the folded state of the protein. This type of experiment is very commonly performed, using either tryptophan residues that are naturally found in a protein or using tryptophan residues that are inserted into a protein by site-directed mutagenesis (this technique allows another amino acid to be replaced with tryptophan simply by changing the DNA that encodes the amino acid). For analogous studies of RNA folding, none of the four RNA nucleobases are appreciably fluorescent, so an unnatural chromophore must be provided. This requires a substantial amount of synthetic organic chemistry as well as synthetic biochemistry, because site-directed mutagenesis cannot insert an unnatural chromophore-containing nucleotide into RNA. In our laboratory, many efforts have focused on pyrene as the chromophore for monitoring RNA folding.

The synthetic organic chemistry necessary for attaching pyrene to RNA requires the preparation of a modified RNA nucleotide that has an unnatural chromophore such as pyrene attached directly. Alternatively, the synthesis of a modified nucleotide (e.g., with an amine or thiol nucleophile tethered to the ribose ring) enables derivatization of the RNA post-synthetically using a suitable chromophore reagent. Using the latter approach, we have synthesized two complete series of pyrene-derivatized RNAs: the "A" series (which has the pyrene directly connected to the ribose ring via a 2'-amido linkage) and the "T" series (which has the pyrene connected via a 2'-tethered amide).

Using fluorescence approaches, we have shown that pyrene is a generally useful probe of RNA folding. These studies used the P4-P6 domain of the Tetrahymena group I intron RNA, which is an independently folding domain that adopts a characteristic folded structure in the presence of sufficient Mg²⁺ (on the order of 1 mM under typical incubation conditions). The P4-P6 domain, which comprises 160 nucleotides and has a molecular weight of ~51000, is an outstanding model system for exploring RNA structure and folding. The fluorescently labeled P4-P6 derivatives were prepared using one of two assembly pathways, depending on the location of the chromophore (5'-end, path 1 or 3'-end, path 2).

In one study, we attached pyrene at ten different positions of P4-P6 on fourteen different tether combinations, and the results support the general utility of pyrene to monitor RNA structure (see Smalley & Silverman, Nucleic Acids Res. 2006, 34, 152-166 [#46] on the Full publications listing). The pyrene fluorescence increases as Mg²⁺ is added to the RNA; and detailed biophysical experiments demonstrate that that the fluorescence increase is due to RNA folding. Moreover, any structural perturbations as detected by pyrene fluorescence correlate well with the same structural perturbations as quantified by nondenaturing gel electophoresis (native PAGE). These observations indicate that pyrene fluorescence will be useful for study many RNAs, including those for which high-resolution structure information is lacking. In ongoing work, we are applying the pyrene fluorescence method to large RNAs other than P4-P6. This includes some recently discovered riboswitches, which are genetic control elements that respond to fluctuating metabolite concentrations without the intervention of proteins.

An example of an interesting fundamental physical-organic study that we have reported (on the P4-P6 RNA, which has molecular weight ~51,000!) was to dissect the thermodynamic and kinetic contributions of an RNA structure interaction named the "tetraloop/receptor" (see Young & Silverman, Biochemistry 2002, 41, 12271-12276 [#18]). This study combined native gel electrophoresis with stopped-flow fluorescence measurements. We examined both the thermodynamic and kinetic effects upon RNA folding when the tetraloop portion of this interaction is altered by changing its RNA sequence. We found that alterations which strongly affect the RNA folding thermodynamics change the folding kinetics very little. In physical-organic terms, the change in free energy of activation (DDG^‡) is small even when the change in free energy of folding (DDG°) is large. The ratio DDG^‡/DDG° is only ~0.07, and thus P4-P6 folding has an early transition state. We infer that the tetraloop/receptor interaction serves to "clamp" the RNA structure after folding has occurred, but this interaction does not contribute much to the folding pathway itself. 

RNA folding landscapes: Phototriggered and photoswitchable RNA folding

The experiments described above explore RNA structure using fluorescence as a signal that correlates with folding. An important overall challenge in understanding RNA folding is to map the entire "RNA folding landscape", using fluorescence as well as other techniques. Consideration of an RNA folding lanscape leads to the realization that it is very easy to prepare RNA in an "unfolded" state. Because RNA folding requires a divalent metal ion like Mg²⁺ as described above, removing all of the Mg²⁺ — or simply not adding it in the first place — forces the RNA into a conformation in which secondary structure interactions (i.e., Watson-Crick base pairs A-U and G-C) are present, but most of the higher-order tertiary contacts are not properly formed. Similarly, we know how to prepare RNA in a "correctly folded" state: add sufficient Mg²⁺ and ensure that the proper pH and temperature are maintained. However, RNA is notorious for forming "misfolded" states, which are conformations that are not correctly folded, yet not entirely unfolded. These misfolded states are important in nature partially because they can act as "kinetic traps", in which the RNA spends significant time in one or more misfolded conformations before ultimately folding correctly.

From the chemistry viewpoint, we would like to study RNA folding that begins from misfolded states (red arrows in diagram above). However, we face what at first sounds like a simple chemical challenge: how can we prepare a stable sample of misfolded RNA from which to initiate our folding experiments? The kinetic traps that occur naturally are transient (meta-stable) conformations; more generally, misfolded conformations represent local but not global minima on the RNA folding landscape. This means that preparing misfolded RNA is not as simple as it might appear. Our strategy to prepare misfolded RNA uses covalent modifications to enforce misfolded states. This strategy should succeed because all of the interactions that control whether an RNA molecule is unfolded, misfolded, or correctly folded are noncovalent; indeed, the stability of the correctly folded form of even a large RNA molecule is only about 5-15 kcal/mol, similar to the stability of a folded protein. Thus, introduction of a covalent modification (~100 kcal/mol within a covalent bond) should easily be able to overwhelm the correct RNA folding and enforce the formation of a misfolded state.

Perhaps the simplest strategy to create a misfolded RNA state is to introduce a single covalent modification that is structurally incompatible with the folded RNA conformation. If the covalent modification is removable with light (i.e., by photocleavage), then the misfolded RNA is a photochemically "caged RNA" (the term "caged" refers not to physical enclosure but rather the entrapment of a specific functional group in an unnatural chemical linkage). We have synthesized caged versions of each of the four RNA nucleotides U, C, A, and G, which are incorporated into folded RNAs to make misfolded versions of RNA. These stable, misfolded RNAs can then be studied in detail. Photocleavage of RNAs that include a caged nucleotide — i.e., phototriggered RNA folding — will allow us to study the misfolded-to-folded RNA folding pathways in ways that cannot be accomplished by starting from entirely unfolded RNA states. For our first effort in this area, see Höbartner & Silverman, Angew. Chem. Int. Ed. 2005, 44, 7305-7309 [#42].

A synthetically more complicated strategy that, in return for the synthetic effort, offers greater control over the RNA structure is to attach a covalent constraint via two positions on a large RNA molecule. One approach to this goal is shown schematically below. Because the correctly folded RNA conformation is geometrically incompatible with the covalent tether, the RNA misfolds, but not entirely; i.e., the constraint enforces a misfolded state and not an unfolded state. The illustrated covalent constraint incorporates the well-known nitroveratryl photocleavable moiety, which permits phototriggered RNA folding by release of the covalent constraint upon irradiation. These studies share some conceptual similarities with our efforts to use double-stranded DNA as a conformational constraint (see DNA as a macromolecular conformational constraint).

Alternatively, the covalent tether may be designed with a photoisomerizable (rather than photocleavable) group. A photoisomerizable group that is widely used in many contexts is azobenzene, which switches from its more stable trans conformation to the less stable cis form upon photolysis with light of the appropriate wavelength. If this isomerization reaction changes the geometrical compatibility of the tether with the folded RNA conformation, then the folded state of the RNA has been rendered photoswitchable.

Click here to return to the Silverman lab overview of research

Click here for a full publications list

Research description last updated January 26, 2008