Fig. 5

Two single-stranded regions on pre-mRNA are associated with polyadenylation. a, b SHAPE reactivity profiles across poly(A) sites for in vivo (a) and deproteinized (b) conditions. The X-axis represents the relative position to the poly(A) site. Average SHAPE reactivities in two regions (gray shaded) were significantly higher compared to flanking regions for both in vivo (purple) and deproteinized (black) conditions (Fisher’s exact test, P value = 3.1e−14 and 3.6e−6 for in vivo; P values = 4.7e−12 and 1.4e−3 for deproteinized). The numbers of control sites are 4575 and 983 for in vivo and deproteinized conditions, respectively. The numbers of poly(A) sites are 3077 and 551 for in vivo and deproteinized conditions, respectively. c Heatmap showing in vivo SHAPE reactivity profiles across the PAS motif “AAUAAA” and poly(A) site. The pre-mRNAs are sorted by the distance between PAS and poly(A) site. The gradient color from light to dark red represents SHAPE reactivity from low to high. The SHAPE reactivities are much higher at both the PAS and poly(A) sites compared to flanking regions. d, e Base-pairing probability (BPP) profiles across poly(A) sites for in vivo (d) and deproteinized (e) conditions. Average BPPs in two regions (gray shaded) were significantly lower compared to flanking regions for both in vivo (purple) and deproteinized (black) conditions (Fisher’s exact test, P value = 5.0e−12 and 1.4e−7 for in vivo; P values = 1.3e−8 and 2.8e−6 for deproteinized). f Heatmap showing in vivo BPPs across the conventional PAS motif “AAUAAA” and poly(A) site. g Classification of RNA structure elements across the PAS and poly(A) sites. The three different single-stranded types (multiple loop, hairpin loop, and internal loop) and the double-stranded stem type were assessed for all the PAS and poly(A) sites. The percentage of each type is shown. Most of the PAS and poly(A) sites are located in the single-stranded loop regions. h Illustrations of two individual pre-mRNA structures with both the PAS and poly(A) sites located in single-stranded loop regions. Nucleotides were color-coded according to the in vivo SHAPE reactivity values