Fig. 1

rG4-seq reveals the global landscape of G-rich regions with the potential to fold into RG4s in Arabidopsis. a Schematic of an RNA G-quadruplex (RG4). The schematic depicts an RG4 with three layers of G-quartets (G3 RG4, guanine (G) coloured in orange), with the loop length of any nucleotide (N, coloured in grey), potassium ions (K⁺, grey ball) coordinated within the G-quartets stabilize RG4s. b Workflow of rG4-seq. Poly(A)-enriched RNAs were subjected to reverse transcription under the buffers with Li⁺ (non-stabilizing condition), K⁺ (stabilizing condition) or K⁺+PDS (stronger stabilizing condition), respectively. The G-rich region sites with folding potential were identified by comparing the coverage of reads between the rG4-seq libraries with different cations as described above. c rG4-seq profiles of AtSMXL5 displayed the read coverage of reverse transcription (RT) with Li⁺ (top), K⁺ (middle) and K⁺+PDS (bottom), respectively. The 3′end of the G-rich region is indicated by a red triangle. A (blue), C (light grey), G (yellow) and U (green). d Residue distribution around RTS sites. Guanine (G) was strongly enriched in the upstream sequences of RT stalling (RTS) identified under both K⁺ and K⁺+PDS conditions, but not in the transcriptome and the downstream sequences of RTS. A (blue), C (light grey), G (yellow) and U (green). e Classification of G-rich regions with folding potential. G-rich regions with folding potential identified in K⁺ (dark blue) and K⁺+PDS conditions (black) were classified into six categories according to the number of G-quartets (G2 with two G-quartets or G3 with three G-quartets), loop length (L, 1–15 nt) and bulge size (non-canonical G3 RG4s with a guanine vacancy: G3VL1-9, or a bulge: G3bulge). f, g rG4-seq profiles of G2 G-rich region on AT4G30460 (f) and G3 G-rich region on AT3G23450 (g), otherwise in c. h The prevalence of both detected and computational-predicted G-rich regions in different genic regions. Computational-predicted G-rich regions were obtained by searching the sequence feature of GxLnGxLnGxLnGx in Arabidopsis transcriptome. i Comparison of base-pairing probability (BPP) of alternative secondary structure in G-rich regions that are detected with K⁺(blue) and undetected (grey) using rG4-seq. The Gs in the G-rich region were classed into 8 bins; flanking sequences (100 nt on both sides) were classed into 20 bins, with 15 bins close to G-rich regions shown. Differences of BPPs between G-rich regions and flanking regions, detected by rG4 with K⁺, P = 0.444; undetected regions, P < 10⁻¹⁶; P values, paired Student’s t test. j Secondary structure of G-rich region detected by rG4-seq (in g) and flanking sequences on AT3G23450, predicted using Vienna RNAfold. The filling colours of orange, green and blue indicate the base-pairing probability of below 0.3, 0.3–0.7 and above 0.7, respectively. Red stars indicate the guanines comprising the G3 region. Numbers indicate nucleotide positions on the transcript