Structure and function of pseudoknots involved in gene expression control

Abstract

Natural RNA molecules can have a high degree of structural complexity but even the most complexly folded RNAs are assembled from simple structural building blocks. Among the simplest RNA elements are double-stranded helices that participate in the formation of different folding topologies and constitute the major fraction of RNA structures. One common folding motif of RNA is a pseudoknot, defined as a bipartite helical structure formed by base-pairing of the apical loop in the stem-loop structure with an outside sequence. Pseudoknots constitute integral parts of the RNA structures essential for various cellular activities. Among many functions of pseudoknotted RNAs is feedback regulation of gene expression, carried out through specific recognition of various molecules. Pseudoknotted RNAs autoregulate ribosomal and phage protein genes in response to downstream encoded proteins, while many metabolic and transport genes are controlled by cellular metabolites interacting with pseudoknotted RNA elements from the riboswitch family. Modulation of some genes also depends on metabolite-induced messenger RNA (mRNA) cleavage performed by pseudoknotted ribozymes. Several regulatory pseudoknots have been characterized biochemically and structurally in great detail. These studies have demonstrated a plethora of pseudoknot-based folds and have begun uncovering diverse molecular principles of the ligand-dependent gene expression control. The pseudoknot-mediated mechanisms of gene control and many unexpected and interesting features of the regulatory pseudoknots have significantly advanced our understanding of the genetic circuits and laid the foundation for modulation of their outcomes.

FIGURE 1

Pseudoknot schematics. (a) Linear representation of base-pairing (dashed lines) in the H-type RNA pseudoknot. The color code is used throughout all figures. Nucleotides are depicted by small open circles. (b) Two-dimensional representation of the H-type pseudoknot formation. The schematic highlights the base-pairing between an apical loop of a hairpin and an external region resulting in the formation of the pseudocontinuous S1-S2 helix. (c-e) The secondary structure of the H-type pseudoknot as a result of coaxial stacking when loop L2, L1 or L3 is eliminated, respectively. (f) A hypothetical secondary structure presentation depicting various types of pseudoknots formed by different types of loops. (Adapted with permission from [12]. Copyright 1990 Elsevier Science Publishers). (g) Circular representation of two adjacent hairpins. RNA is shown as a semi-circle with nucleotides in small circles and base-pairing depicted by lines. (Adapted with permission from [13]. Copyright 1992 American Chemical Society). (h) Circular representation of a pseudoknot. Note crossing of the chords. (i) Planar representation of the geometric types of pseudoknots that delineates connectivity of complex pseudoknotted structures. (Adapted with permission from [14]. Copyright 2003 Oxford University Press). H depicts a hairpin and L designates a bulge, interior or multiple loop.

FIGURE 2

Pseudoknot in the phage T4 gp32 mRNA and the mechanism of gp32 autorepression. (a) gp32 (light orange ovals) binds single-stranded DNA as the replication fork advances. If gp32 is overproduced, one molecule of the protein binds to the pseudoknot located in the 5′-UTR of gp32 mRNA. Once the gp32-mRNA interaction has been nucleated, it is extended in a cooperative fashion towards the initiation codon, blocking the Shine-Dalgarno sequence and preventing translation of the mRNA. PK, pseudoknot. SD, Shine-Dalgarno sequence. ORF, open reading frame. Pol, DNA polymerase complex. (b) Secondary structure of the gp32 mRNA pseudoknot. (c) NMR structure of the gp32 mRNA pseudoknot (PDB code 2TPK). (d) Crystal structure of gp32 shown with electrostatic surface potential (PDB code 1GPC).

FIGURE 3

Pseudoknots in the autorepression of ribosomal proteins genes. (a), Secondary structure of the double pseudoknot specifically recognized by the E. coli ribosomal protein S4. (b) Secondary structure of the pseudoknot in the E. coli rpsO mRNA encoding the ribosomal protein S15. Ribosomal protein S15 (oval in wheat color) is positioned along S2 of the pseudoknot according to its location in the S15-mRNA complex. Translation initiation codon (AUG) is in light blue color. (c) An atomic model of the S15-pseudoknot complex derived from the crystal structure of the S15-rRNA complex and molecular modeling, supported by biochemical^, and cryo-EM data (PDB code 2VAZ). (d) Schematic representation of the S15-mRNA interactions with the ribosome and the entrapment mechanism of repression. If S15 is overproduced in the cell, the protein binds and stabilizes the pseudoknot in its own mRNA. The S15-mRNA complex is then loaded onto the ribosome and makes SD/anti-SD interactions with the 16S rRNA. However, the ribosome cannot melt the pseudoknot structure so that the initiation codon, located in L3 of the pseudoknot in vicinity of S2, cannot reach the ribosome P site and interact with the initiator tRNA. As a consequence, the 3′ end of the mRNA rests on the surface of the ribosome rather than inside the mRNA channel of the ribosome. The resulting ‘entrapped’ complex is stalled in the inactive pre-initiation state. In the absence of S15, the rpsO mRNA either unfolds on the ribosome or binds the ribosome in the unfolded conformation so that mRNA enters the mRNA channel in the single-stranded conformation (dashed line). In this position, the initiation codon can be placed into the P site and interact with the initiator tRNA, thereby initiating translation of the mRNA.

FIGURE 4

H-type pseudoknot-based riboswitches. Arrows indicate rotation of zoomed-in structures relative to global views. Secondary structure (a), three-dimensional X-ray structure (b), A-minor groove triples (c), and fluoride binding (d), in the fluoride riboswitch structure (PDB code 4ENC). Fluoride anion is shown by a red sphere. Magnesium cations (green spheres) are shown with coordination bonds (green sticks) and coordinated water molecules (light red spheres). Putative hydrogen bonds are depicted by black dashed lines. Secondary structure (e), three-dimensional NMR structure (f), A-amino kissing motif (g), and preQ₁ base-pairing arrangement from the top view (h) in the preQ₁ riboswitch structure (PDB code 2L1V). Secondary structure (i), three-dimensional X-ray structure (j), and preQ₁ recognition from the side (k) and top (l) views in the preQ₁–II riboswitch structure (PDB code 4JF2). Secondary structure (m), three-dimensional X-ray structure (n), inclined A-minor motif (o), and SAM binding (p) in the SAM-II riboswitch structure (PDB code 2QWY). Electrostatic interactions between sulfur atom (yellow) and oxygen atoms of RNA are shown by yellow dashed lines.

FIGURE 5

Infrequent pseudoknot arrangements. Secondary structure (a), three-dimensional X-ray structure (b), and SAH binding (c) in the SAH riboswitch structure (PDB code 3NPN). Secondary structure (d), three-dimensional X-ray structure (e), major groove triples (f), and c-di-GMP binding (d) in the c-di-GMP riboswitch structure (PDB code 3Q3Z).

FIGURE 6

Pseudoknots in the glmS riboswitch-ribozyme. (a) Secondary structure schematic of the RNA showing formation of hairpins and pseudoknots. Secondary structures, depicted from the crystal structure of the RNA,^, highlighting pseudoknot PK1 (b), PK2 (c), and PK3 (d). The three-dimensional structure of the RNA (PDB code 2GCV) highlighting pseudoknots PK1 (e), PK2 (f), and PK3 (g). (h) Zoomed-in view of PK3 showing interactions of L3 with the pseudoknot stem. (i) Formation of the active site by pseudoknots PK1 and PK2. The color code corresponds to PK1 except S1.2 from PK2 depicted in cyan. (j) Zoomed-in view of the active site, showing recognition of GlcN6P (red) and position of the scissile phosphate (SP, gray sphere).

FIGURE 7

Summary of typical features in the regulatory pseudoknots. The pseudoknot schematic is consistent with the average length of the pseudoknot elements and preferred nucleotide composition on L1 and L3 loops. Nucleotides are depicted by circles; nucleotide preference is indicated within large circles. The length range is indicated in the schematic. Red oval shows the location of the ligand binding sites.

FIGURE 8

A model of the SAM-II riboswitch folding and translation repression. Mg²⁺ cations stabilize the prefolded stem-loop bearing S1 (left panel) by triple interactions with L3 (top middle panel) and facilitate the formation of the transient pseudoknot-like conformations, capable of ligand binding (bottom middle panel). SAM captures such conformers and induces adaptive changes, resulting in the extension of initial pairing in S2 that sequesters the Shine-Dalgarno sequence (right panel). (Adapted with permission from [77]. Copyright 2011 Nature Publishing Group).