Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Dec 15.
Published in final edited form as: J Mol Biol. 2023 Nov 2;435(24):168340. doi: 10.1016/j.jmb.2023.168340

Solution structure of poly(UG) RNA

Cristian A Escobar 1,†, Riley J Petersen 1,†, Marco Tonelli 1,2, Lixin Fan 3, Katherine Henzler-Wildman 1,2, Samuel E Butcher 1,2
PMCID: PMC10841838  NIHMSID: NIHMS1944484  PMID: 37924862

Abstract

Poly(UG) or “pUG” RNAs are UG or GU dinucleotide repeat sequences which are highly abundant in eukaryotes. Post-transcriptional addition of pUGs to RNA 3’ ends marks mRNAs as vectors for gene silencing in C. elegans. We previously determined the crystal structure of pUG RNA bound to the ligand N-methyl mesoporphyrin IX (NMM), but the structure of free pUG RNA is unknown. Here we report the solution structure of the free pUG RNA (GU)12, as determined by nuclear magnetic resonance spectroscopy and small and wide-angle x-ray scattering (NMR-SAXS-WAXS). The low complexity sequence and 4-fold symmetry of the structure result in overlapped NMR signals that complicate chemical shift assignment. We therefore utilized single site-specific deoxyribose modifications which did not perturb the structure and introduced well-resolved methylene signals that are easily identified in NMR spectra. The solution structure ensemble has a root mean squared deviation (RMSD) of 0.62 Å and is a compact, left-handed quadruplex with a Z-form backbone, or “pUG fold.” Overall, the structure agrees with the crystal structure of (GU)12 bound to NMM, indicating the pUG fold is unaltered by docking of the NMM ligand. The solution structure reveals conformational details that could not be resolved by x-ray crystallography, which explain how the pUG fold can form within longer RNAs.

Keywords: RNA, NMR, poly(UG), Quadruplex, SAXS

Graphical Abstract

graphic file with name nihms-1944484-f0001.jpg

Introduction

Approximately 3% of the human genome is composed of simple sequence repeats, or microsatellites, thought to arise from slippage during DNA replication (1). The most frequently occurring microsatellites are TG and CA dinucleotide repeats (1–6). Although TG and CA repeats are equal in number in DNA, there is a strong bias for TG repeats to be downstream of transcription start sites and AC repeats to be upstream (4). Therefore, TG repeats are often transcribed into poly(UG) or “pUG” RNA sequences. The human transcriptome has ~20,000 pUGs of 12 or more dinucleotide repeats, which are predominately located in introns (7). (TG)n repeats in DNA genomes are often polymorphic, with n being a variable number, and genome wide association studies (GWAS) have linked the number of repeats to disease (8–18). Some of these diseases may arise from pUG sequences in RNA regulatory regions. For example, a variable pUG region within an intron of the CFTR gene has been correlated with splicing defects, leading to atypical cystic fibrosis with male infertility (11,12,19). We therefore hypothesized that disease correlations with (TG)n microsatellites may be due to pUG RNA structure (7). For example, in the case of CFTR, pUG structure may influence protein recognition of adjacent splice sites (11,19).

In C. elegans, pUGs can be added to RNAs post-transcriptionally by the ribonucleotidyltransferase enzyme Mut-2/Rde-3 (hereafter, Rde-3). Rde-3 adds poly(UG) or poly(GU) tails (“pUG tails”) to RNA 3’ ends, initiating with either G or U with equal frequency (20,21). pUG tails with 12 or more Gs convert otherwise inert RNAs into potent vectors for gene silencing (21). pUG tails function by recruiting RNA dependent RNA Polymerase (RdRP) to the ends of RNAs, marking them as templates for synthesis of small interfering RNAs (siRNAs) (21). Enzymatic cycling of sense strand pUG tail synthesis and antisense siRNA production amplifies the gene silencing pathway and allows it to persist for generations, a phenomenon known as transgenerational epigenetic inheritance (TEI) (21,22). pUG mediated TEI is important for transposon silencing and maintenance of genome stability in C. elegans (21).

We recently discovered the pUG fold, in which pUGs with 12 or more Gs fold into an atypical RNA quadruplex (G4) structure with a left-handed backbone that resembles Z RNA (7) (Figures 1A, B). The pUG fold G4 is required for gene silencing (7). We determined the crystal structure of the pUG fold bound to the G4-stabilizing ligand N-methyl-mesoporphyrin IX (NMM), which stacks on the structure (7) (Figure 1B). The pUG-NMM crystals were disordered due to 4-way twinning, owing to the near perfect 4-fold symmetry of the structure. The resulting electron density maps were insufficient to precisely locate some features of the RNA, including the positions of the 5’ and 3’ ends and the conformations of bulged uridines (7). Therefore, an improved understanding of the pUG fold structure is needed. Here we report the solution structure of the free pUG fold, determined by nuclear magnetic resonance (NMR) spectroscopy combined with small and wide-angle x-ray scattering (SAXS-WAXS). The low complexity and high symmetry of dinucleotide repeat sequences present challenges for NMR assignment and structure determination, and we successfully determined the structure by combining established strategies and new NMR methods. For reproducibility, the entire dataset, including raw and processed data (>30 GB), along with all software used in this study, have been made accessible within the NMRbox virtual machine resource (23) at www.nmrbox.org (public/pUGNMR). Overall, the solution structure of the pUG fold in the absence of NMM and the crystal structure bound to NMM are similar, with a few notable exceptions described below. The NMR data also show how the pUG fold can form within longer sequences, providing new insights into the mechanism by which pUG repeats can fold and function in cells.

Figure 1.

Figure 1.

Open in a new tab

The pUG fold has four-fold symmetry of in solution. (A) Sequence of (GU)12, one of the RNAs used in this study. (B) Crystal structure of (GU)11.5 bound to N-methyl mesoporphyrin IX (NMM)(red), PDB ID 7MKT. Purple spheres are K+ ions. (C) Short-hand description (50) of the pUG-fold observed in the crystal structure bound to NMM. Upper case nucleotides are Anti, lower case are Syn. Bold nucleotides are 2’-endo and italicized nucleotides are 3’ endo. Inverted nucleotides indicate inverted strand polarity relative to the 5’ nucleotide. < indicates bulged uridines. Circled nucleotides are single nucleotide loops and the numbers shown refer to the nucleotide number in the sequence rather than the length of the loops (50). (D) 31P 1D NMR spectrum showing six main peaks for the 24-nucleotide (GU)12 RNA. Resonance assignments are indicated above the peaks. Labels are color coded to match the RNA structure with G quartet nucleotides in green, U nucleotides in blue. The two small unassigned peaks are likely due to the presence of a small amount of RNA that is longer by one nucleotide (n+1), a common byproduct of run-off transcription by T7 RNA polymerase. Note there is no assignment for G1 because the 5’ end has a hydroxyl group. (E) 2D 1H,13C HSQC of aromatic resonances. The spectrum shown is an overlay of two experiments, one performed on a G-labeled sample (green) and one from a U-labeled sample (blue). The terminal U24 nucleotide has two chemical shifts (24a and 24b) owing to a minor population of n+1 transcription product. (F). 2D 1H,13C HSQC of ribose resonances for the G-labeled sample and (G) U-labeled sample.

Results

Repetitive low complexity sequences present unique challenges for assignment of NMR resonances, because the sequence diversity of neighboring nucleotides is a major determinant of RNA chemical shifts (24,25). Since all nucleotides in pUG RNA have the same neighbors except for the 5’ and 3’ ends, extensive chemical shift overlap is expected. Additionally, the pUG fold has 4-fold symmetry (Figures 1A, B and C), which results in further overlap of chemical shifts. We therefore utilized pUG fold sequence variants to assign all 1H, 13C, 15N and 31P resonances. These variants included the minimal 23-nucleotide pUG fold sequence of 11.5 GU repeats or (GU)11.5 (7), (GU)12 which is longer by one 3’ uridine, two nucleotide-type labeled samples of (GU)12 with 13C,15N-labeled G or Us, and six variants harboring single 2’ deoxyribose modifications at different positions.

In the pUG-NMM crystal structure, six nucleotides adopt a unique conformation that is repeated four times within the intramolecular quadruplex (Figures 1B and C). Consistent with this 4-fold symmetry, a 31P 1D NMR spectrum reveals 6 disperse peaks for the 24-nucleotide (GU)12 sequence (Figure 1D). 2D 1H,13C HSQC spectra also show 4-fold symmetry except for the unique 5′ and 3’ ends which break the symmetry of the molecule (Figure 1E–G). For example, the aromatic resonances of the 12 guanosine and 12 uridine residues are each in 3 major peaks groups, with the exception of the terminal G1 and U24 nucleotides which have distinct chemical shifts (Figure 1E). Nucleotides near the 3’ end can be unambiguously assigned by comparing spectra of (GU)11.5 to (GU)12 which terminate at G23 and U24, respectively, but otherwise adopt the same pUG fold (Supplemental Figure 1A–D). Ribose spin systems were assigned by 3D HCCH TOCSY and COSY experiments using the 13C,15N samples labeled at G or U. Triple resonance HCP experiments were used to connect adjacent ribose groups via scalar couplings across the phosphodiester bond (Supplemental Figure 2). Triple resonance HCN experiments (26) were then used to connect ribose groups to their aromatic nucleobase resonances across the glycosidic bond (Figure 2).

Figure 2.

Figure 2.

Open in a new tab

Ribose to nucleobase correlations via through-bond triple resonance HCN experiments. (A) G-labeled ribose to aromatic correlations via the N9 glycosidic nitrogen. (B). U-labeled ribose to aromatic correlations via the N1 glycosidic nitrogen. Connectivity to aromatic resonances for U4,10,16,22 and U24b are observed only at lower contour levels due to low signal intensities.

Sequence specific resonance assignments were confirmed by analyzing NMR spectra of pUG RNA variants harboring single deoxyribose modifications at different positions. In the pUG-NMM crystal structure (7), there are no intramolecular hydrogen bonds involving 2’ hydroxyl groups, and most of the ribose sugars adopt C2’ endo puckers. Since the C2’ endo pucker is energetically favored by deoxyribose, we reasoned that single deoxyribose modifications should induce little if any structural perturbation. Deoxy substitutions were incorporated at G1, U2, G3, U4, G5, and U6. These six nucleotides were chosen as they adopt a single unit of the quadruplex that is repeated four times in the crystal structure, a symmetry which is also observed by NMR (Figure 1). The single deoxyribose modifications break this symmetry, induce local chemical shift perturbations, and introduce new, well-resolved methylene signals between 2.5 to 3.5 ppm which are readily identifiable because the RNA has no NMR signals in this region (Figure 3). The deoxyribose methylene groups also give rise to NOEs that could be used to further verify assignments but were not used for structure determination. In all cases, the chemical shifts of the deoxy substituted nucleotides matched most closely to the chemical shifts of their corresponding unmodified nucleotides; for example, the resonances for deoxy G1 (dG1) variant correspond most closely to the G1 resonances of the unmodified RNA (Figure 3). All single deoxyribose variants have NOE patterns indicative of the same overall structure as the unmodified RNA.

Figure 3.

Figure 3.

Open in a new tab

2D 1H,1H-NOESY of (GU)12 single deoxyribose variants. Top rows show intra-residue NOE cross-peaks between aromatic and H2’ and H2” methylene protons from 2’-deoxyribose modifications at positions G1, U2, G3, U4, G5 and U6. Bottom row shows overlay with NOESY spectrum of wild type (GU)12 (black), with red arrows showing chemical shift perturbation by the deoxy substitutions.

The 12 guanosines in the pUG fold form three G quartets with stable, long-lived hydrogen bonds that we directly detected by NMR (7). In the pUG-NMM crystal structure, all Us are bulged out except for U4, U10, U16 and U22 which form a symmetric “U4 quartet” at a solvent exposed edge of the structure with imino-O4 hydrogen bonds (7). 2D 1H,15N HSQC spectra reveal dispersed guanosine imino peaks but only a single, exchange-broadened U imino peak for the U4 quartet (Figures 4A and B). NOEs from the U4 quartet are observable in 2D 15N-HSQC filtered 1H,1H NOESY (Figure 4C). The only observed inter-nucleotide NOE from the U4 quartet is a weak NOE to the G5 quartet, which is consistent with the non-sequential G1-G3-G5-U4 stacking observed in the crystal structure (7) (Figure 4C). For non-exchangeable resonances, intra- and inter-residue NOEs were distinguished using 3D NOESY-1H,13C-HSQC and 13C-filtered-NOESY-1H,13C-HSQC experiments (27) (Figures 4D and E). A sequential walk through a 2D NOESY is shown (Supplemental Figure 3). In total, 549 NOEs were measured for the 24 nucleotide (GU)12 pUG RNA (Table 1). Ambiguous restraints were implemented for all NOEs that were overlapped due to 4-fold symmetry. Sugar puckers were measured using 2D 1H,1H-TOCSY spectra (Supplemental Figure 4, Supplementary Table 1). 21 residual dipolar couplings (RDCs), including 6 1H-15N and 15 1H-13C RDCs, were measured from non-overlapped signals for the G and U labeled samples and used as restraints for structure determination (Supplemental Figure 5). All restraints have been deposited into the PDB.

Figure 4.

Figure 4.

Open in a new tab

(A) 2D 1H,15N-HSQC of imino signals from G-labelled and (B) U-labeled (GU)12 RNA. (C) 2D 1H, 1H 15N-HSQC-NOESY of U-labeled sample. (D and E). 2D planes of 3D 13C-filtered HSQC-NOESY spectra showing inter-residue NOEs between U4 and G5 observed using the G-labeled (D) and U-labeled (E) samples.

Table 1.

(GU)12 20 lowest-energy structures
NMR distance and dihedral constraints
Distance restraints
Total NOE 549
Intra-residue 237
Inter-residue 312
Sequential 142
Non-sequential 170
Hydrogen bonds 72
Total dihedral angle restraints 143
RDCs 21
Structure statistics
rmsd (deviation)
 Distance constraints (Å) 0.081 ± 0.002
 Dihedral angle constraints (°) 1.452 ± 0.067
 RDC Rfree (%)
  G sample 18.1 ± 2.8
  U sample 30.4 ± 7.3
Deviations from idealized geometry
 Bond lengths (Å) 0.007±0.0
 Bond angles (°) 0.894±0.03
 Impropers (°) 0.701±0.03
Average rmsd (Å)
 All atom to average structure 0.62
Open in a new tab

Small and wide-angle x-ray scattering (SAXS-WAXS) data were measured in solution for (GU)12 and used to jointly refine the structure against NMR and SAXS-WAXS data (Supplementary table 2). The 20 lowest energy models out of 100 refined NMR-SAXS-WAXS structures are shown (Figure 5A–C), and structure determination statistics are listed in Table 1. The pairwise RMSD with respect to the average structure is 0.62 Å over all nucleotides, with most of the variation coming from the bulged uridines (U2, U8, U14 and U20), single uridine propeller loops (U6, U12 and U18) and 3’ end uridine (U24). The core of the structure formed by the three G quartets and U quartet is well defined with an overall RMSD of 0.16 Å. The experimental RDC restraints agree well with the predicted values from the structure (R2=0.99, (Supplemental Figure 5, Table 1). A comparison between the experimental and back-calculated SAXS-WAXS data is shown (Figure 5D–F, Supplementary Table 3). Independently of the NMR data, an electron density map was calculated ab initio from the SAXS-WAXS data using the DENSS algorithm (28). The corresponding electron density map has a resolution of ~9.9 Å (Supplementary table 2) and agrees well with the structure ensemble (Figures 5E, F). The RMSD between the average energy minimized NMR structure and the pUG-NMM crystal structure is 1.82 Å. These data indicate that the free pUG RNA structure in solution is very similar to the 1.97 Å crystal structure bound to NMM. The quartet interactions in the solution structure ensemble are shown in Supplemental Figure 6.

Figure 5.

Figure 5.

Open in a new tab

(GU)12 structure ensemble. (A) The 20 lowest energy structure models for (GU)12 are shown. (B) Energy minimized average structure, as viewed from above the G1-G7-G13-G19 quartet. (C) Same as (B) but rotated by 90°. (D) SAXS-WAXS profiles. Experimental x-ray scattering intensities as a function of the scattering vector q are shown in black. Predicted X-ray scattering profiles back-calculated from the (GU)12 NMR-SAXS-WAXS structures in (A) were computed using Crysol (red) and Xplor-NIH software (blue). The predicted scattering profiles from the crystal structure 7MKT, minus the NMM ligand, were calculated using Crysol (green). (E) (GU)12 electron density model calculated from the SAXS-WAXS data using DENSS algorithm with 4-fold symmetry applied. Image was prepared using Pymol using the default DENSS color gradient (28) corresponding to standard units of electron density (sigma) where blue is 2.5, cyan is 5, green is 7.5, yellow is 10 and red is 15–200. (F) same as (E) but rotated 90°.

The unique chemical shifts of the pUG fold should enable the identification of the fold within larger RNAs. To test this idea, we transcribed an oligonucleotide corresponding to the 3’ end of an oma-1 mRNA fragment previously shown to induce RNA silencing in vivo upon addition of a pUG tail (7,21). We compared the 31P NMR spectra of the 27-nucleotide oma-1 fragment with and without the addition of the 24-nucleotide pUG fold sequence (UG)12 (Figure 6). The 31P peaks of the oma-1 fragment are mostly overlapped and centered near −1 ppm, as expected for an unstructured single stranded RNA. The 51-nucleotide oma-1-(UG)12 spectrum clearly shows distinct pUG fold peaks at 0.5 and −3 ppm, which correspond to the unique backbone inversions in the pUG fold at positions U2, U8, U14, and U20 and G5, G11, G17 and G23. Thus, the pUG fold can be readily detected in a longer RNA by 1D 31P NMR.

Figure 6.

Figure 6.

Open in a new tab

1D 31P NMR spectra comparing (GU)12, the final 27 nucleotides of the 3’ end of the oma-1 mRNA (oma-1), and the same oma-1 sequence followed by a (UG)12 pUG tail (oma-1-pUG). The 51-nucleotide oma-1-pUG spectrum shows two distinct pUG fold peaks at 0.5 and −3 ppm which arise from U2, U8, U14, U20 and G5, G11, G17, G23 as indicated.

Discussion

The similarities between the free pUG fold solution structure and the pUG-NMM crystal structure indicate NMM docks to the pUG fold without significantly perturbing the RNA structure. These solution and crystal structures also highlight the strengths and limitations inherent to each method. For example, 6 coordinated potassium ions are well-defined in the crystal structure (Figure 1B) (7) but are not visible by NMR. Conversely, disorder in the crystal due to 4-way twinning resulted in an electron density that could not reveal the positions of the terminal nucleotides, while the unique environment of the termini results in distinct chemical shifts that identify these residues in the NMR spectra. In cells, pUG tails can extend beyond 100 nucleotides in length (20), and CD data indicate that the pUG fold can form when embedded within the middle of an RNA in vitro (7). However, it was unclear from the crystal structure as to how the RNA chain could extend beyond the penultimate G23 nucleotide, which is stacked in the middle of the quadruplex (Figure 1B), and no density could be observed for the terminal U24 nucleotide (7). The NMR-SAXS-WAXS structure shows the 3’ terminal U24 nucleotide extends away from G23 and the rest of the pUG fold to produce a solvent accessible 3’ end, creating room for additional nucleotides (Figure 5B).

The bulged nucleotides U2, U8, U14 and U20 have anti glycosidic torsion angles (Supplementary Table 1). In the crystal, these nucleotides were modeled in the syn conformation (7), but can be repositioned in the electron density map in the anti conformation in a manner that is consistent with both the solution and x-ray data (Supplemental Figure 7). Ambiguities in determining syn vs anti conformations of purines in crystal structures have been previously noted (29), and this problem is even more challenging for smaller pyrimidine nucleobases. However, the orientation of the uridines in the pUG fold are well-defined by the NMR data. The dynamic bulged uridines (U2, U8, U14 and U20) and looped uridines (U6, U12, U18) (Figure 5A, Supplemental Figure 6) explain why sequence variants with di-adenosine (AA) insertions can still adopt the pUG fold and are functional for gene silencing in vivo (7). Thus, the pUG fold does not have to be entirely poly(UG).

The free pUG fold is an unusual left-handed quadruplex comprised of 3 G quartets and a U quartet stacked non-sequentially (G1-G3-G5-U4), with a fold that mirrors the conformation of Z RNA (7,30) (Figure 7). The similarity between the pUG fold and Z RNA include: 1) an overall left-handed topology, 2) alternating C2’ endo and C3’ endo sugar puckers of stacked nucleotides, 3) alternating inverted nucleotides, and 4) syn and anti nucleotide conformations at analogous positions (Figure 7A and B). Owing to these extensive similarities, we describe the pUG fold as predominately Z-form. On the other hand, there are notable differences between the pUG fold and duplex Z RNA structures. For example, in the pUG fold the guanosine imidazole rings of G1 and G3 are stacked, whereas the equivalent CpG step in Z RNA is unstacked (Figure 7C and D). Unstacking of CpG in the Z RNA duplex allows formation of a lone pair-π interaction (31), which is not observed in the pUG fold (Figure 7C and D). These data indicate the lone pair-π interaction is not strictly required for formation of a Z-form backbone conformation in RNA. Due to differences in stacking and the presence of bulged nucleotides which are not found in Z RNA, the phosphates in the pUG fold create a series of “backwards S” shapes (Figure 1B) instead of a Z-shape.

Figure 7.

Figure 7.

Open in a new tab

Comparison of pUG fold to Z RNA. (A) A segment of (GU)12. (B) A single strand segment of duplex Z-RNA (PDB 2GXB). Ribose orientations are indicated with black arrows. Conformational similarities are listed in green text beside the corresponding nucleotides. (C and D) Top and side views of the first two stacked nucleotides of each structure. A lone pair-pi interaction in Z-RNA is shown as a dashed line (D); this interaction is replaced by base stacking in the pUG fold.

The pUG fold can be readily identified by NMR despite its low complexity sequence and symmetry. For example, the aromatic H6 chemical shifts of the inverted U quartet (U4-U10-U16-U22) are 5 standard deviations from the mean value in the BioMagResBank (BMRB) data base (Figure 1C and E), and the 31P chemical shift dispersion allows detection by 1D NMR in larger RNAs (Figure 6). Unusual chemical shifts for the pUG fold are expected since it is the first Z-form G4 structure to be investigated by NMR. The deoxyribose nucleotide substitution approach described here may be generally useful for other RNAs with low complexity sequences and/or symmetry.

In future studies, it will be interesting to investigate how proteins recognize the pUG fold. One such recently identified protein is the human methyltransferase DNMT1, which specifically binds to, and is inhibited by, the pUG fold (32). These data, along with the numerous numbers of pUG repeat sequences in the human genome (7), suggest the pUG fold occurs in human RNA. The human Polycomb Repressive Complex 2 also binds to pUG fold sequences (33), and to G4 RNA structures in general (34). Other pUG fold binding proteins have been identified from C. elegans (7,21).

The pUG fold reveals two new principles of nucleic acid folding. First, stable G4 structures can be formed from RNA sequences with no consecutive guanosines. Second, Z-form nucleic acids are not limited to duplex conformations (35,36). The fact that a simple dinucleotide repeat sequence adopts a unique fold with four structurally distinct quartets, multiple backbone inversions, 4 bulged nucleotides and 4 propeller loops at specific positions seems paradoxical and leads to the question, why does the pUG tail curl (37)? The answer is the pUG fold is the lowest free energy structure; it maximizes base stacking and hydrogen bonding while minimizing electrostatic repulsion. The left-handed Z form backbone allows extensive base stacking (Figure 6A and C), which is not possible in the alternative A-form RNA structure due to unstacking of the right-handed pyrimidine-purine steps and thermodynamically weak nearest neighbor interactions of UG and GU wobble pairs (38). The pUG fold is also the first example of an independently stable, Z-form nucleic acid structure that can form in solution under physiological salt concentrations. Z-form nucleic acid duplexes are high energy structures that are relatively unstable due to closely positioned (~8 Å) phosphates on opposite strands; however, Z-duplexes can be stabilized by high salt concentrations, bound proteins, or supercoiling forces (30,35,36). The pUG fold, on the other hand, solves the electrostatic repulsion problem by maximizing interphosphate distances (~17 Å) while also chelating a central core of potassium ions. The 3 potassium ions are dehydrated and sandwiched between the 4 quartets with each ion coordinated to 8 carbonyl oxygens (7). An additional 3 potassium ions are partially hydrated and bind peripherally at the backbone (7). The hydrogen bonding potential of the pUG fold is maximized with each guanosine participating in 4 hydrogen bonds and each uridine in the U quartet participating in 2 hydrogen bonds. Therefore, at physiological salt concentrations, the pUG fold is more energetically favorable than alternative structures, including helical A, B or Z form duplex conformations. It is surprising that an apparently simple dinucleotide repeat sequence can encode a complex and stable RNA fold, which in turn reveals new information about the range of conformations that are possible for RNA.

Materials and methods

RNA preparation

Unlabeled RNAs used for NMR and SAXS-WAXS experiments were prepared by chemical synthesis (Integrated DNA technologies, Inc.). 13C,15N G and U uniformly labeled (GU)12 RNA samples were prepared by in vitro transcription and 5’-triphosphates removed with calf intestinal alkaline phosphatase (Invitrogen) as previously described (7). For measurement of RDCs, RNA samples were partially aligned in 20 mg/ml Pf1 phage (ASLA Biotech) as previously described (7). Oma-1 and oma-1-pUG samples were prepared via in vitro transcription. All RNAs were purified by denaturant PAGE (15% Acrylamide in 7M urea and TBE buffer) followed by anion exchange chromatography in Hi-trap Q column (Cytiva). After purification, RNAs were exchanged into pure water using an Amicon Ultra 3-kDa filter, followed by addition of folding buffer. RNA folding buffer contained 20 mM potassium phosphate pH 7.0 and 100 mM KCl in either 10% D2O or 99.99% D2O with the exception of one unlabeled (GU)12 sample that was prepared in 50 mM HEPES at pH 7.0 containing 150 mM KCl. The latter sample had identical chemical shifts to the other RNA samples. Samples were folded by incubation at 90 °C followed by slow cooling to room temperature over the course of 5 hours or more.

Oligonucleotide sequences

(GU)11.5: 5’-GUGUGUGUGUGUGUGUGUGUG-3’

(GU)12: 5’-GUGUGUGUGUGUGUGUGUGUGU-3’

(dG1-GU)12: 5’-dGUGUGUGUGUGUGUGUGUGUGU-3’ (dG= deoxyguanosine)

(dU2-GU)12: 5’-GdUGUGUGUGUGUGUGUGUGUGU-3’ (dU= deoxyuridine)

(dG3-GU)12: 5’-GUdGUGUGUGUGUGUGUGUGUGU-3’ (dG= deoxyguanosine)

(dU4-GU)12: 5’-GUGdUGUGUGUGUGUGUGUGUGU-3’ (dU= deoxyuridine)

(dG5-GU)12: 5’-GUGUdGUGUGUGUGUGUGUGUGU-3’ (dG= deoxyguanosine)

(dU6-GU)12: 5’-GUGUGdUGUGUGUGUGUGUGUGU-3’ (dU= deoxyuridine)

Oma-1 : 5’-GGAAACGGUGCCUUUUCAUUCAUCCCG-3’

Oma-1-pUG : 5’-GGAAACGGUGCCUUUUCAUUCAUCCCGUGUGUGUGUGUGUGUGUGUGUGUG −3’

NMR experiments

NMR data was collected at the National Magnetic Resonance Facility using a temperature of 3 °C or 20 °C on Bruker Avance III HD 600 MHz and 750 MHz spectrometers and Varian VNMR 600 MHz and 800 MHz spectrometers. All spectrometers were equipped with cryogenic probes. A 2D 1H,1H-TOCSY was used to define sugar puckers, in which nucleotides with observable H1’-H2’ cross peaks were designated as C2’ endo, and all others were restrained as C3’ endo. NOEs used for distance restraints were obtained from a 2D 1H,1H-NOESY recorded on an unlabeled (GU)12 sample and 3D NOESY-1H,13C-HSQC experiments recorded on U-labeled and G-labeled samples. 1H, 13C experiments included HSQC, 3D HC(C)H-COSY and 3D HC(C)H-TOCSY. 1H, 15N experiments included 2D HSQC and 2D 1H, 1H planes of NOESY-1H,15N-HSQC spectra. Triple resonance 2D HCN experiments (26) were used to correlate ribose groups to their corresponding nucleobases. Triple resonance 3D HCP experiments (39) were used to connect adjacent bases across the phosphodiester bond. Resonance assignments were corroborated using 2D 1H,1H-NOESY of 2’ deoxyribose variants of (GU)12 at positions G1, U2, G3, U4, G5 and U6. Through-hydrogen bond experiments for G quartets, as well as the methods used for extracting RDCs were previously described (7). Non-uniform sampling (NUS) was used for all 3D experiments to increase signal-to-noise ratios (40). NMR experiments used in this study, along with acquisition parameters, are listed in Supplementary Tables 4, 5 and 6. NMR data were processed in NMRbox (23) using NMRPipe (41) and NUS reconstructions were performed using SMILE (42). NMRFAM-SPARKY (43) was used to analyze and assign NMR spectra. Proton chemical shifts were referenced via the internal standard sodium 4,4-dimethyl-4-silapentane-1-sulfonate (DSS) at 0 ppm. 13C, 15N and 31P chemical shifts were referenced indirectly to 1H (DSS) as described (44).

SAXS-WAXS experiments

SAXS and simultaneous WAXS data were collected using beam line 12-ID-B at the Advance Photon Source, Argonne National Laboratory. Photon energy was 13.3-KeV and sample-to-detector distance was set in a way that q ranges in SAXS and WAXS are overlapping and achieve a total range of 0.004< q < 2.69 Å−1, where q = (4π/λ)sinθ, and 2θ is the scattering angle. The 2D scattering patterns were recorded with Pilatus 2M detector for SAXS and Pilatus 300K detector for WAXS. Samples of (GU)12 were prepared in 20 mM HEPES, 150 mM KCl, pH 7.0 at concentrations of 0.26, 1, 1.4, and 1.6 mg/ml. Concentration series measurements were carried out for removing the scattering contribution due to interparticle interactions and extrapolating the data to infinite dilution in data analysis. The sample solution and matching buffer were measured using a flow cell to minimize radiation damage. A total of 60 sequential data frames were recorded with the exposure time of 0.5 second for each sample. The 2D intensity maps were corrected and reduced to 1D scattering profiles and then averaged after eliminating outliers using software developed by 12ID-B beamline. The buffer background subtraction and intensity extrapolation to infinite dilution were carried out using NCI SAXS Core in-house software (NCI-SAXS). Data analysis was carried out using ATSAS software (45). CRYSOL (46) was used for calculation of scattering profiles from structures. Modeling of electron density from SAXS-WAXS data was performed using DENSS algorithm (28) implemented on the BioXTAS RAW software package (47). 4-fold symmetry consistent with the NMR data was imposed during the DENSS electron density calculation. SAXS-WAXS profiles, DENSS model and fitting to average NMR structure were deposited in SASBDB with accession code SASDSW7.

(GU)12 structure calculations

Unique chemical shifts were observed for the 5’ G1, U2 and G3 nucleotides and the 3’ G23 and U24 nucleotides. All other chemical shifts could be assigned to nucleotides related through 4-fold symmetry. When NOEs could be connected to unique chemical shifts they were restrained to specific nucleotides, while NOEs between symmetry-related positions were ambiguously restrained. Hydrogen bonds were also restrained in this manner. Sugar puckers were constrained to C2’ endo if J couplings were observed between H1’ and H2’ proton resonances in 2D 1H,1H-TOCSY spectra and to C3’ endo if no couplings were observed. Structure calculations were performed using Xplor-NIH software version 3.5 (48). The initial folding protocol used randomization of initial coordinates, followed by high temperature dynamics stage at 3000 K, simulated annealing from 3000 K to 25 K and a final energy minimization step. 200 models were initially calculated. Force constants used to enforce experimental restraints during calculations were defined as follows: Distance restrains for NOEs and hydrogen bonds were ramped from 2 to 50 kcal/mol/Å2, dihedral angles were set to 200 kcal/mol/rad2 to define base conformation and sugar puckers, planarity restraints for bases on each quartet were set to 300 kcal/mol/Å2. The lowest energy model from this step was used as the initial starting structure for refinement as previously described (49), where RDCs and SAXS-WAXS data were used as restraints with the same force constants described above. The RDC restraint force constant was ramped from 0.01 to 1 kcal/mol/rad2, and the axial (DA) and rhombic (Rh) component values of the alignment tensor were allowed to vary. RDC R-free values were obtained by performing ten different structure calculations in which 5 randomly chosen RDC values (out of 21) were removed. R-free was calculated as the average for the R factors calculated for each structure using the respective RDCs removed from the set. The SAXS-WAXS restraint force was set to 400 kcal/mol. Out of 100 models generated, the 20 lowest energy models with no distance and dihedral angle restraint violations were used to represent the final ensemble of the (GU)12 structure. Pymol was used for data analysis and structure visualization.

Supplementary Material

1
2

Highlights.

  • Poly(UG) or “pUG” RNA is abundant in humans and directs gene silencing in C. elegans.

  • The solution structure of pUG RNA reveals a Z-form quadruplex, or “pUG-fold.”

  • The pUG fold has three well-defined guanosine quartets and one uridine quartet.

  • The pUG fold can form and be detected in longer RNAs.

Acknowledgements

This study made use of the National Magnetic Resonance Facility at Madison, which is supported by NIH grant R24GM141526 and P41GM103399. This study made use of NMRbox: National Center for Biomolecular NMR Data Processing and Analysis, a Biomedical Technology Research Resource (BTRR), which is supported by NIH grant P41GM111135 (NIGMS). This study was supported by NIH grant R35 GM118131 to S.E.B. We acknowledge the use of the SAXS Core facility of the Center for Cancer Research (CCR), National Cancer Institute (NCI) of National Institutes of Health (NIH). The facility is supported by the intramural research program of the NIH, NCI, CCR and Frederick National laboratory for Cancer Research under contract 75N91019D00024. Use of the Advanced Photon Source, an Office of Science User Facility operated for the U.S. Department of Energy (DOE) Office of Science by Argonne National Laboratory, was supported by U.S. Department of Energy, Office of Biological and Environmental Research under contract DE-AC02– 06CH11357. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. The authors thank Dr. Yuichiro Nomura for helpful comments and assistance with visualization.

Footnotes

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT Author Statement

Cristian Escobar: Investigation, Resources, Formal analysis, methodology, data curation, writing, visualization. Riley Petersen: Investigation, Resources, Formal analysis, methodology, data curation, writing, visualization. Marco Tonelli: Methodology, Investigation, data curation. Lixin Fan: Methodology, Investigation, data curation. Katherine Henzler-Wildman: Writing. Samuel E. Butcher: Conceptualization, investigation, writing, visualization.

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 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.

Data Accessibility

NMR chemical shifts have been deposited into BioMagResBank, accession code 52045.

Structure coordinates have been deposited into the PDB, ID 8TNS.

SAXS-WAXS data was deposited in SASBDB with accession code SASDSW7.

All NMR and SAXS data (~30 GB, including raw, processed and annotated spectra), input files, structure calculation scripts and software used in this study can be accessed, analyzed and reproduced within the NMRbox virtual machine resource (www.nmrbox.org) in the folder public/pUGRNA.

References

  • 1.Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W et al. (2001) Initial sequencing and analysis of the human genome. Nature, 409, 860–921. [DOI] [PubMed] [Google Scholar]
  • 2.Toth G, Gaspari Z and Jurka J (2000) Microsatellites in different eukaryotic genomes: survey and analysis. Genome Res, 10, 967–981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Weber JL and May PE (1989) Abundant class of human DNA polymorphisms which can be typed using the polymerase chain reaction. Am J Hum Genet, 44, 388–396. [PMC free article] [PubMed] [Google Scholar]
  • 4.Sawaya S, Bagshaw A, Buschiazzo E, Kumar P, Chowdhury S, Black MA and Gemmell N (2013) Microsatellite tandem repeats are abundant in human promoters and are associated with regulatory elements. PLoS One, 8, e54710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Bagshaw ATM (2017) Functional Mechanisms of Microsatellite DNA in Eukaryotic Genomes. Genome Biol Evol, 9, 2428–2443. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Kelkar YD, Strubczewski N, Hile SE, Chiaromonte F, Eckert KA and Makova KD (2010) What is a microsatellite: a computational and experimental definition based upon repeat mutational behavior at A/T and GT/AC repeats. Genome Biol Evol, 2, 620–635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Roschdi S, Yan J, Nomura Y, Escobar CA, Petersen RJ, Bingman CA, Tonelli M, Vivek R, Montemayor EJ, Wickens M et al. (2022) An atypical RNA quadruplex marks RNAs as vectors for gene silencing. Nat Struct Mol Biol, 29, 1113–1121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Noriega V, Martinez-Laperche C, Buces E, Pion M, Sanchez-Hernandez N, Martin-Antonio B, Guillem V, Bosch-Vizcaya A, Bento L, Gonzalez-Rivera M et al. (2015) The Genotype of the Donor for the (GT)n Polymorphism in the Promoter/Enhancer of FOXP3 Is Associated with the Development of Severe Acute GVHD but Does Not Affect the GVL Effect after Myeloablative HLA-Identical Allogeneic Stem Cell Transplantation. PLoS One, 10, e0140454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Lei SF, Deng FY, Dvornyk V, Liu MY, Xiao SM, Jiang DK and Deng HW (2005) The (GT)n polymorphism and haplotype of the COL1A2 gene, but not the (AAAG)n polymorphism of the PTHR1 gene, are associated with bone mineral density in Chinese. Hum Genet, 116, 200–207. [DOI] [PubMed] [Google Scholar]
  • 10.Keneme B and Sembene M (2019) GTn Repeat Microsatellite Instability in Uterine Fibroids. Front Genet, 10, 810. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Hefferon TW, Groman JD, Yurk CE and Cutting GR (2004) A variable dinucleotide repeat in the CFTR gene contributes to phenotype diversity by forming RNA secondary structures that alter splicing. Proc Natl Acad Sci U S A, 101, 3504–3509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Groman JD, Hefferon TW, Casals T, Bassas L, Estivill X, Des Georges M, Guittard C, Koudova M, Fallin MD, Nemeth K et al. (2004) Variation in a repeat sequence determines whether a common variant of the cystic fibrosis transmembrane conductance regulator gene is pathogenic or benign. Am J Hum Genet, 74, 176–179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Gill AJ, Garza R, Ambegaokar SS, Gelman BB and Kolson DL (2018) Heme oxygenase-1 promoter region (GT)n polymorphism associates with increased neuroimmune activation and risk for encephalitis in HIV infection. J Neuroinflammation, 15, 70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Gao PS, Heller NM, Walker W, Chen CH, Moller M, Plunkett B, Roberts MH, Schleimer RP, Hopkin JM and Huang SK (2004) Variation in dinucleotide (GT) repeat sequence in the first exon of the STAT6 gene is associated with atopic asthma and differentially regulates the promoter activity in vitro. J Med Genet, 41, 535–539. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Dias C, Elzein S, Sladek R and Goodyer CG (2019) Sex-specific effects of a microsatellite polymorphism on human growth hormone receptor gene expression. Mol Cell Endocrinol, 492, 110442. [DOI] [PubMed] [Google Scholar]
  • 16.Daenen KE, Martens P and Bammens B (2016) Association of HO-1 (GT)n Promoter Polymorphism and Cardiovascular Disease: A Reanalysis of the Literature. Can J Cardiol, 32, 160–168. [DOI] [PubMed] [Google Scholar]
  • 17.Cai Q, Gao YT, Wen W, Shu XO, Jin F, Smith JR and Zheng W (2003) Association of breast cancer risk with a GT dinucleotide repeat polymorphism upstream of the estrogen receptor-alpha gene. Cancer Res, 63, 5727–5730. [PubMed] [Google Scholar]
  • 18.Boraska Jelavic T, Barisic M, Drmic Hofman I, Boraska V, Vrdoljak E, Peruzovic M, Hozo I, Puljiz Z and Terzic J (2006) Microsatelite GT polymorphism in the toll-like receptor 2 is associated with colorectal cancer. Clin Genet, 70, 156–160. [DOI] [PubMed] [Google Scholar]
  • 19.Buratti E, Brindisi A, Pagani F and Baralle FE (2004) Nuclear factor TDP-43 binds to the polymorphic TG repeats in CFTR intron 8 and causes skipping of exon 9: a functional link with disease penetrance. Am J Hum Genet, 74, 1322–1325. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Preston MA, Porter DF, Chen F, Buter N, Lapointe CP, Keles S, Kimble J and Wickens M (2019) Unbiased screen of RNA tailing activities reveals a poly(UG) polymerase. Nat Methods, 16, 437–445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Shukla A, Yan J, Pagano DJ, Dodson AE, Fei Y, Gorham J, Seidman JG, Wickens M and Kennedy S (2020) poly(UG)-tailed RNAs in genome protection and epigenetic inheritance. Nature, 582, 283–288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Shukla A, Perales R and Kennedy S (2021) piRNAs coordinate poly(UG) tailing to prevent aberrant and perpetual gene silencing. Curr Biol, 31, 4473–4485 e4473. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Maciejewski MW, Schuyler AD, Gryk MR, Moraru II, Romero PR, Ulrich EL, Eghbalnia HR, Livny M, Delaglio F and Hoch JC (2017) NMRbox: A Resource for Biomolecular NMR Computation. Biophys J, 112, 1529–1534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Barton S, Heng X, Johnson BA and Summers MF (2013) Database proton NMR chemical shifts for RNA signal assignment and validation. J Biomol NMR, 55, 33–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Wang Y, Han G, Jiang X, Yuwen T and Xue Y (2021) Chemical shift prediction of RNA imino groups: application toward characterizing RNA excited states. Nat Commun, 12, 1595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Sklenar V, Dieckmann T, Butcher SE and Feigon J (1998) Optimization of triple-resonance HCN experiments for application to larger RNA oligonucleotides. J Magn Reson, 130, 119–124. [DOI] [PubMed] [Google Scholar]
  • 27.Peterson RD, Theimer CA, Wu H and Feigon J (2004) New applications of 2D filtered/edited NOESY for assignment and structure elucidation of RNA and RNA-protein complexes. J Biomol NMR, 28, 59–67. [DOI] [PubMed] [Google Scholar]
  • 28.Grant TD (2018) Ab initio electron density determination directly from solution scattering data. Nat Methods, 15, 191–193. [DOI] [PubMed] [Google Scholar]
  • 29.Shi H, Kimsey IJ, Gu S, Liu HF, Pham U, Schumacher MA and Al-Hashimi HM (2021) Revealing A-T and G-C Hoogsteen base pairs in stressed protein-bound duplex DNA. Nucleic Acids Res, 49, 12540–12555. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Nichols PJ, Krall JB, Henen MA, Vogeli B and Vicens Q (2023) Z-RNA biology: A central role in the innate immune response? RNA. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Zirbel CL and Auffinger P (2022) Lone Pair...pi Contacts and Structure Signatures of r(UNCG) Tetraloops, Z-Turns, and Z-Steps: A WebFR3D Survey. Molecules, 27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Jansson-Fritzberg LI, Sousa CI, Smallegan MJ, Song JJ, Gooding AR, Kasinath V, Rinn JL and Cech TR (2022) DNMT1 inhibition by pUG-fold quadruplex RNA. RNA. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Wang X, Goodrich KJ, Gooding AR, Naeem H, Archer S, Paucek RD, Youmans DT, Cech TR and Davidovich C (2017) Targeting of Polycomb Repressive Complex 2 to RNA by Short Repeats of Consecutive Guanines. Mol Cell, 65, 1056–1067 e1055. [DOI] [PubMed] [Google Scholar]
  • 34.Song J, Gooding AR, Hemphill WO, Kasinath V and Cech TR (2023) Structural basis for inactivation of PRC2 by G-quadruplex RNA. bioRxiv. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Tang Q (2022) Z-nucleic acids: Uncovering the functions from past to present. Eur J Immunol, 52, 1700–1711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Krall JB, Nichols PJ, Henen MA, Vicens Q and Vogeli B (2023) Structure and Formation of Z-DNA and Z-RNA. Molecules, 28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Monchaud D (2023) Why does the pUG tail curl? Mol Cell, 83, 330–331. [DOI] [PubMed] [Google Scholar]
  • 38.Zuber J, Schroeder SJ, Sun H, Turner DH and Mathews DH (2022) Nearest neighbor rules for RNA helix folding thermodynamics: improved end effects. Nucleic Acids Res, 50, 5251–5262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Marino JP, Schwalbe H, Anklin C, Bermel W, Crothers DM and Griesinger C (1994) Three-Dimensional Triple-Resonance 1H, 13C, 31P Experiment: Sequential Through-Bond Correlation of Ribose Protons and Intervening Phosphorus along the RNA Oligonucleotide Backbone. Journal of the American Chemical Society, 116, 6472–6473. [Google Scholar]
  • 40.Palmer MR, Suiter CL, Henry GE, Rovnyak J, Hoch JC, Polenova T and Rovnyak D (2015) Sensitivity of nonuniform sampling NMR. J Phys Chem B, 119, 6502–6515. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J and Bax A (1995) NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J Biomol NMR, 6, 277–293. [DOI] [PubMed] [Google Scholar]
  • 42.Ying J, Delaglio F, Torchia DA and Bax A (2017) Sparse multidimensional iterative lineshape-enhanced (SMILE) reconstruction of both non-uniformly sampled and conventional NMR data. J Biomol NMR, 68, 101–118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Lee W, Tonelli M and Markley JL (2015) NMRFAM-SPARKY: enhanced software for biomolecular NMR spectroscopy. Bioinformatics, 31, 1325–1327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Markley JL, Bax A, Arata Y, Hilbers CW, Kaptein R, Sykes BD, Wright PE and Wuthrich K (1998) Recommendations for the presentation of NMR structures of proteins and nucleic acids--IUPAC-IUBMB-IUPAB Inter-Union Task Group on the standardization of data bases of protein and nucleic acid structures determined by NMR spectroscopy. Eur J Biochem, 256, 1–15. [DOI] [PubMed] [Google Scholar]
  • 45.Manalastas-Cantos K, Konarev PV, Hajizadeh NR, Kikhney AG, Petoukhov MV, Molodenskiy DS, Panjkovich A, Mertens HDT, Gruzinov A, Borges C et al. (2021) ATSAS 3.0: expanded functionality and new tools for small-angle scattering data analysis. J Appl Crystallogr, 54, 343–355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Svergun D, Barberato C and Koch MHJ (1995) CRYSOL - A program to evaluate x-ray solution scattering of biological macromolecules from atomic coordinates. Journal of Applied Crystallography, 28, 768–773. [Google Scholar]
  • 47.Hopkins J, Gillilan R and Skou S (2020) BioXTAS RAW 2.0: The latest in SAXS data analysis. Acta Crystallogr A, 76, A27–A27. [Google Scholar]
  • 48.Schwieters CD, Kuszewski JJ and Clore GM (2006) Using Xplor-NIH for NMR molecular structure determination. Prog Nucl Mag Res Sp, 48, 47–62. [DOI] [PubMed] [Google Scholar]
  • 49.Cornilescu G, Didychuk AL, Rodgers ML, Michael LA, Burke JE, Montemayor EJ, Hoskins AA and Butcher SE (2016) Structural Analysis of Multi-Helical RNAs by NMR-SAXS/WAXS: Application to the U4/U6 di-snRNA. J Mol Biol, 428, 777–789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Banco MT and Ferre-D’Amare AR (2021) The emerging structural complexity of G-quadruplex RNAs. RNA, 27, 390–402. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS1944484-supplement-1.pdf (12.1MB, pdf)
2
NIHMS1944484-supplement-2.pse (4.2MB, pse)

Data Availability Statement

NMR chemical shifts have been deposited into BioMagResBank, accession code 52045.

Structure coordinates have been deposited into the PDB, ID 8TNS.

SAXS-WAXS data was deposited in SASBDB with accession code SASDSW7.

All NMR and SAXS data (~30 GB, including raw, processed and annotated spectra), input files, structure calculation scripts and software used in this study can be accessed, analyzed and reproduced within the NMRbox virtual machine resource (www.nmrbox.org) in the folder public/pUGRNA.

RESOURCES