The UA_handle: a versatile submotif in stable RNA architectures

Abstract

Stable RNAs are modular and hierarchical 3D architectures taking advantage of recurrent structural motifs to form extensive non-covalent tertiary interactions. Sequence and atomic structure analysis has revealed a novel submotif involving a minimal set of five nucleotides, termed the UA_handle motif (5'XU/AN(n)X3'). It consists of a U:A Watson-Crick: Hoogsteen trans base pair stacked over a classic Watson-Crick base pair, and a bulge of one or more nucleotides that can act as a handle for making different types of long-range interactions. This motif is one of the most versatile building blocks identified in stable RNAs. It enters into the composition of numerous recurrent motifs of greater structural complexity such as the T-loop, the 11-nt receptor, the UAA/GAN and the G-ribo motifs. Several structural principles pertaining to RNA motifs are derived from our analysis. A limited set of basic submotifs can account for the formation of most structural motifs uncovered in ribosomal and stable RNAs. Structural motifs can act as structural scaffoldings and be functionally and topologically equivalent despite sequence and structural differences. The sequence network resulting from the structural relationships shared by these RNA motifs can be used as a proto-language for assisting prediction and rational design of RNA tertiary structures.

Figure 1.

Type I and type II UA_handle signatures and conformers. (A) The UA_handle is typically formed of a U(2):A(3) (WC:HG) trans bp (in blue) stacked on a classic WC bp, X(1):X(5) [purine in green, pyrimidine in yellow), with one or more bulging nucleotides located 3′ of A(3), (N(4.n) in pink]. For bp annotations, see legend of Figure 2. (B) Sequence variation of X(1):X(5) in function of N(4.n). The table is based on the sequence analysis of 99 atomic structures of UA_h with canonical U(2):A(3) (WC:HG) trans bp. The occurrence of each bp is indicated between brackets. (+) These three G(1):G(5) bps are all from one of the UA_hs of the double T-loop motif. Despite a 2nt bulge, this non-canonical UA_h is of type I. (C) Superposition of canonical UA_handle motifs identified in X-ray structures and listed Table S1. The color code corresponds to the one in (A). The conserved atomic positions in X(1):X(5) and U(2):A(3) were used for the superposition. (D) Sequence signature of type I UA_h: an H-bond can be formed between N7-G(5) and 2′OH-A(3) when the sugar pucker of A(3) is in C2′ endo. (E) Sequence signature of type II UA_h: an H-bond can form between N4-C(5) and the O1 or O2 of P(O)4-N(4.2). As seen in the stereographic image, an additional H-bond can potentially form between 2′OH-A(3) and O2P-N(4.2). For more examples of type I UA_h, see also Figure S1 from Supplementary Data.

Figure 2.

Hierarchical organizational network of RNA structure motifs built from the UA_h submotif. Motifs are organized from the left to the right according to their increase in structural complexity. Submotifs are minimally sized recurrent set of nucleotides with conserved conformation (circled in bold lines). They are not expected to be stable by themselves as they are almost always associated to other submotifs or nts to create full-fledge stable motifs (circled in blue or gray for secondary and tertiary motifs, respectively.) The most significant structural characteristics of UA_h motifs are indicated on their respective sequence signatures according to the annotation of Leontis and Westhof (74). Visual of the 3D structures of some of these motifs can be found in the figures indicated in blue below the motif name. For bps symbols, see the legend in inset: WC, Watson–Crick edge (circle); HG, Hoogsteen edge (square); SG, shallow groove edge (triangle). For example, the WC:HG trans bp is symbolized by an open circle associated to an open square. The WC:HG cis bp is represented by a plain circle combined to a plain square. Capital letters indicate that the nucleotide is conserved in more than 95% of the cases: small letters indicate that the nucleotide position is conserved in more than 85% of the cases. X, any nucleotide (A, U, C or G) paired to another one through classic WC cis bp; N, any nucleotide (A, U, C or G); R or r, purine; Y or y, pyrimidine; N_n, a sequence of n nucleotides; U or G-clamp, WC edge of a U or G in interaction with a phosphate; syn or S, indicates that a single or paired nucleotide is in syn; Hl and Hs stand for the long and short stems of a pseudo-knot, respectively. 5′ and the arrow symbol indicate 5′ and 3′ ends, respectively. See also Figure S3 (Supplementary Data).

Figure 3.

The UA_handle motif is involved in various modes of tertiary interactions. (A) 2D diagram summarizing the interactions that the different nt positions of the UA_handle can form. For the legend, see Figures 2 and S3. Interactions are colored according to the nucleotide position involved in the UA_h. (B–K) Examples of various interactions involving the U(2):A(3) bp. Bps and tertiary interactions are indicated according to the color code and annotations in (A). Bps indicated between brackets are not shown on the stereo images. (B) GAAA/11 nt motif, A(s.1) is in violet (see Figure 4 and text); (C) UAA:GAN motif (23S; Ec U1578), A(s.1) and A(s.2) are in violet; (D) UA_handle turn (23S; Tt U1621), A(s.1) is in violet; (E) Nested double T-loop (23S; Hm U481); (F) UAh_PK motif (23S; Hm U2069); (G) UA_h_PK motif bound to DielsAlder reaction product (Diels-Alderase ribozyme); (H) T-loop motif (tRNA^Phe), notice the U-clamp instead of G-clamp (Figure S3); (I) T-loop like motif bound to AMP (AMP/ATP aptamer); (J) T-loop motif (23S; Hm U1388); (K) T-loop motif bound to TPP (TPP riboswitch). Motifs boxed in blue share a WC:WC trans bp involving A(3). Motifs boxed in orange share a G-clamp or U-clamp. On the right side of the red dotted line are motifs from riboswitches and ribozyme recognizing ligands. Notice the similar modality of nt (or ligand) recognition between the tRNA^Phe T-loop and ATP aptamer T-loop-like and between the TPP riboswitch T-loop and the 23S RNA T-loop.

Figure 4.

Examples of tertiary interactions involving the X(1):X(5) WC bp. (A) the A-minor tilted interaction (type-I/IIT A-minor) from the GAAA/11 nt motif (19): notice the canonical ribose zipper (75) that involves the second and third adenines of the GAAA, in type-I and type-IIT A-minor contacts, respectively (Hm 23S; U_1457). Typical H-bonds pattern for type-I/IIT is indicated. (B) UA_h A-minor twist interaction (type-II/IIST A-minor): this pattern of H-bonds presents the interacting nucleotides in a super-tilted configuration. Notice the cis ribose zipper that involves the second and third adenines, in type-II and type-IIST A-minor contacts, respectively (Bs RNase P; U147). The second adenine in type-IIST allows formation of H-bond, N3(A2): N2(G5), instead type-IIT H-bond, N6(A2):2′OH(G5). Typical H-bond pattern for type-II/IIST is indicated. (C–D) G-ribose (G-ribo) interactions (36): example of ‘closed” and ‘open’ G-ribose interacting motif from UA_h turns. (C) In the ‘closed’ conformation (Hm 23S: U2330), the ‘G ribose’ G:C bp makes three H-bond contacts with the UA_h motif: note the 2′OH-N3(A3) H-bond. (C) In the ‘open’ conformation (Ec 23S: U1019), only two contacts are formed with X(5).

Figure 5.

Examples of typical UA_handle interactions involving the bulging nucleotides N(4.n). (A) Schematic highlighting the various bulge mediated tertiary interactions in UA_handle motifs. (B) UA_h_turn motifs: (top) Hm 23S rRNA U1116 (PDB_ID:1JJ2), (middle) Tt 23S rRNA U1621 (PDB_ID:2J01), (bottom) Hm 23S rRNA U1696 (PDB_ID:1JJ2). (C) UA_h A-minor junction motifs: (top) Tt 23S rRNA U2562 (PDB_ID:2J01), (bottom) from TPP riboswitch (PDB_ID:2GDI). (D) UA_h 3WJ motifs: (top) Ec 16S rRNA U652 (PDB_ID:2AW7), (bottom) Ec 23S rRNA U1991 (PDB_ID:2AWB). (E) UA_h_PK motifs: (top) Hm 23S rRNA U1676 (PDB_ID:1JJ2), (bottom) from Diels-alderase (PDB_ID:1YKV). (F) UAA/GAN motif: Hm 23S rRNA U1457 (PDB_ID:1JJ2). (G) UGA/gAN motif; Tt 23S rRNA U1621 (PDB_ID:2J01). (H) Nested double T-loop: Tt 23S rRNA U475 (PDB_ID:2J01). (I) T-loop motif: Hm 23S rRNA U1388 (PDB_ID:1JJ2). For each examples displayed, the type of interactions involved is indicated in the upper left corner. Conserved and semi-conserved interactions are circled with continuous or dashed lines, respectively. For annotations, see Figures 2 and S3.

Figure 6.

Principle of functional/topological equivalence. The kink turn [Right: in H5-7 domain (PDB_ID: 1JJ2)] and UA_h turn [Left: in H82–87 domain (PDB_ID:2J01)] can both be components of the doughnut-PK domain identified in the 23S rRNA. Albeit of different sequence signatures and local structures, they are topologically equivalent as they both contribute to the folding of the same topological domain. The doughnut-PK domain can act as scaffolding for structural expansion by accommodating sequence insertions in the UA_h turn (left). For annotations see Figures 2 and S3.

Figure 7.

The T-loop PK domain as a prevalent structural scaffold in natural RNA molecules. (A) Summary of the sequence insertions within the 3D structure of the T-loop PK domain. The constituent sub-motifs are in blue (UA_h), yellow (U-turn) and magenta (A-minor triple). The T-loop PK domain shown is from the 23S rRNA (Hm:301–350; Ec:296–342). (B) Detailed secondary structure diagram of the consensual structural core of the bacterial and Archaea 23S rRNAs deduced from crystallographic structures. Note that the T-loop PK domain (boxed in blue) brings together the various structural domains (I, II, III, IV, V, VI and VII) of the ribosome. (C) Prediction of the architecture of the core of the natural AdoCbl aptamer (62). (D) Prediction of the structural core of the FMN aptamer (62). Positioning of elements P1–P2, P4 and P6 are not well defined compared to P3–P5. The T-loop in P2 should form an intercalating nt interaction with a distant nt that is most likely one of the conserved purines in L6. The GA shared motif in P6 as well as the base pairs in P2 might be involved in one of the riboswitch structural states. (B,C,D) The tertiary interaction (T.i) of the T-loop PK domain (boxed in blue) corresponds to a class of topological equivalence. The conserved T-loop can either recognize another terminal loop (B, D), an UGA/gAN motif (C) or a helix (Figure 2), leading to structurally different but functionally equivalent interactions. Red stars indicate new tertiary interaction predictions [in comparison to (62)]. Thin lines are for zero nt length connectors. Thick lines are for variable length sequence insertions. For annotations, see Figures 2 and S3.