The discovery that RNA molecules play key roles in converting the genetic information contained in DNA into the proteins that perform structural and catalytic functions in cells motivated the quest to determine the molecular structure of RNAs. In the early 1970s, Alexander Rich, Aaron Klug, and Sung-Hou Kim independently solved the structures of transfer RNAs, revealing how tRNAs carry the amino acids that are used in protein synthesis on the ribosomes. The field of RNA structural biology then languished for almost 20 years, due in part to the technical difficulty of preparing RNA samples for study in the laboratory and the lack of awareness of the wide variety of biological functions of RNA.

The situation changed in 1990, driven by the discoveries of many new kinds of RNA molecules in cells and viruses. In fact, many viral genomes, including those of HIV, HCV, and the influenza viruses, are made entirely of RNA, which can be either single-stranded or double-stranded. Questions about the stability of RNA genomes and how their structures might differ from those of genomic DNA increased the urgency of discovering how RNA molecules are structured.

The wide-ranging functions of RNA reflect a structural diversity much richer than that observed in DNA molecules. The propensity of RNA to form compact folded shapes was first revealed in the 1970s with the determination of the molecular structures of several tRNA molecules (Figure 6-22). These analyses showed that a single strand of RNA folds back on itself to form short base-paired or partially base-paired segments connected by unpaired regions (Figure 6-23). This property, called RNA secondary structure, enables RNA molecules to fold into many different shapes that lend themselves to many different biological functions.

As in DNA, the paired strands in RNA are antiparallel and tend to assume a right-handed helical conformation dominated by base-stacking interactions. Unlike DNA, however, the base-paired segments of RNA are interspersed with a variety of other, non-Watson-Crick, base pairings such as A–A and G–U (Figure 6-24; also see the How We Know section at the end of this chapter). In addition, RNA secondary structures include regions of unpaired nucleotides, which can interact with noncontiguous sequences to stabilize the three-dimensional folding. Such interactions produce compact shapes containing surfaces or crevices that bind other molecules or form sites capable of catalyzing chemical reactions, much like protein enzymes.

193

The greater structural variety in RNA relative to DNA reflects the three main chemical differences between the two polynucleotides: the pentose (2′-deoxyribose in DNA vs. ribose in RNA), the base composition (thymine in DNA vs. uracil in RNA), and the sugar pucker of the pentose (C-2′ endo in DNA vs. C-3′ endo in RNA). The presence of the 2′-hydroxyl group on the sugar of RNA nucleotides provides an extra site for hydrogen bonding, potentially stabilizing three-dimensional folding. This hydroxyl group also influences the sugar pucker, leading to more closely spaced phosphates on the 5′ and 3′ sides of each sugar and hence to a more compact, A-form helical structure (see Figure 6-17).

As in the DNA double helix, RNA base stacking is made energetically favorable by the resulting burial of hydrophobic surfaces away from the hydrophilic surroundings. In tRNA, as well as in the more recently discovered structures of catalytic RNA molecules and ribosomes, virtually all of the bases are stacked, even when they are not part of Watson-Crick base pairings. In each case, structural motifs, such as base-triple interactions (see Figure 6-21a, b) and helix-helix packing, allow stable three-dimensional folding.

Double-stranded RNAs, such as those that form the genomes of some viruses, do exist in nature. These RNAs exist as long helical structures analogous to the DNA double helix. In addition, some RNAs do not seem to form stable three-dimensional structures through local base-pairing interactions. For example, mRNAs seem to perform their function as transient carriers of genetic information without adopting any specific three-dimensional structure. These RNAs may fold into three-dimensional structures only in the presence of bound proteins, forming complexes called ribonucleoproteins (RNPs).

Most of the highly structured RNAs contain noncanonical base pairs and backbone conformations not observed in DNA. In many cases, the 2′-hydroxyl group on ribose, a chemical feature that distinguishes RNA from DNA, seems to be directly or indirectly responsible for these unique structural properties. Recall that the presence of this 2′ hydroxyl gives ribose its C-3′ endo geometry (its sugar pucker), as distinct from the C-2′ endo characteristic of deoxyribose. This seemingly small difference in chemical conformation leads to a distinct helical geometry for RNA relative to DNA. Furthermore, it enables direct or water-mediated hydrogen bonding between the 2′-hydroxyl of one RNA nucleotide and the adjacent ribose or phosphate (Figure 6-25). As a result, RNA helices are more thermodynamically stable than DNA helices of the same length and sequence.

194

Where complementary sequences are present in an RNA molecule, the predominant double-stranded structure in the RNA is an A-form, right-handed double helix (see Figure 6-17). The A-RNA helix has a wider, shallower minor groove and a narrower, deeper major groove than the B-form helix observed for most DNA. The A-form geometry in RNA has a shorter distance between adjacent phosphates in the sugar–phosphate backbone than in B-DNA, a consequence of the C-3′ endo sugar pucker of ribose. A B-form RNA helix has not been observed in nature. Z-RNA helices have been induced to form in the laboratory under high-salt or high-temperature conditions but are not known to occur in cells.

As mentioned previously, mismatched or unmatched bases are common in base-paired segments of RNA, locally disrupting the regular A-form helix and resulting in bulges or internal loops (see Figure 6-23). The potential for base pairing within a single strand of RNA frequently produces thermodynamically stable secondary structures that consist of hairpinlike conformations capped by connecting loops. Such hairpins are the most common type of RNA secondary structure, often containing specific short sequences at their ends (such as UUCG or GAAA) that form particularly energetically favorable loops. The nucleotides in the loops are arranged to maximize hydrogen bonding and base stacking, thereby enhancing thermodynamic stability. Important additional structural contributions are made by hydrogen bonds that are not part of canonical Watson-Crick base pairs. These properties, evident in the structures of tRNAs and catalytic RNAs, were also evident in the ribosome structures solved by Jamie Cate and others (see this chapter’s Moment of Discovery). The functions of these highly structured RNAs, like those of proteins, depend on their three-dimensional properties.

Weak interactions, especially base-stacking interactions, play a major role in stabilizing RNA structures, just as they do in DNA. The 2′ hydroxyl is often involved in hydrogen bonds and van der Waals interactions that stabilize alternative helical shapes and conformations, such as loops and kinks that require the close approach of the two phosphodiester backbones. Divalent and monovalent metal ions—such as Mg²⁺, Ca²⁺, K⁺, and Na⁺—bind to specific sites in RNA and help shield the negative charge of the backbone, allowing parts of the molecule to pack more tightly together (Figure 6-26).

The diversity of RNA three-dimensional structures found in nature is illustrated by riboswitches, non-protein-encoding RNA sequences with the unique ability to directly bind cellular metabolites and form alternative metabolite-free and metabolite-bound shapes. Found widely among bacteria and sporadically in other organisms, riboswitches can change their structure on binding to a particular metabolite and, as consequence, alter the efficiency of synthesis of proteins involved in pathways utilizing that metabolite. Such metabolite-sensing RNAs function through complex three-dimensional folding to create structures that can selectively recognize a small molecule within the mixture of molecules in a cell. Due to their ability to adopt a wide range of structures, riboswitches can bind to molecules ranging in size and type from tiny ions to large organic molecules. Although riboswitches employ common structural principles to form their metabolite-binding shapes, different riboswitch classes possess unique features for ligand recognition, even those that have evolved to recognize the same metabolites (Figure 6-27). We will encounter further examples of the wide-ranging structures of RNA when discussing ribosomes and catalytic introns in Chapters 16 and 18.

The analysis of RNA structure, along with the relationship between structure and function, is an emerging field of inquiry with many of the same complexities as the analysis of protein structure. The importance of understanding RNA structure has grown as we have become increasingly aware of the large number of functional roles for RNA molecules. For example, we now know that the extensive structural features found in the genomic RNA of HIV control viral gene expression (Highlight 6-2).

195