Prp43 bound at different sites on the pre-rRNA performs distinct functions in ribosome synthesis

Abstract

Yeast ribosome synthesis requires 19 different RNA helicases, but none of their pre-rRNA-binding sites were previously known, making their precise functions difficult to determine. Here we identify multiple binding sites for the helicase Prp43 in the 18S and 25S rRNA regions of pre-rRNAs, using UV crosslinking. Binding in 18S was predominantly within helix 44, close to the site of 18S 3' cleavage, in which Prp43 is functionally implicated. Four major binding sites were identified in 25S, including helix 34. In strains depleted of Prp43 or expressing only catalytic point mutants, six snoRNAs that guide modifications close to helix 34 accumulated on preribosomes, implicating Prp43 in their release, whereas other snoRNAs showed reduced preribosome association. Prp43 was crosslinked to snoRNAs that target sequences close to its binding sites, indicating direct interactions. We propose that Prp43 acts on preribosomal regions surrounding each binding site, with distinct functions at different locations.

Figure 1

Identification of Putative Binding Sites of Prp43 on Ribosomal RNA Control cells (C) or Prp43-HTP-expressing cells (A) were UV crosslinked in vivo or crosslinking was performed in vitro after cell lysis and enrichment of Prp43-HTP-containing complexes on IgG Sepharose (B). Crosslinked RNAs were trimmed and ligated to linkers followed by RT-PCR and Solexa sequencing. The obtained sequences were aligned with the rDNA encoding 35S pre-rRNA (nucleotides 1–6880), and the number of hits for each individual nucleotide is plotted. The position of the mature 18S, 5.8S, and 25S rRNAs is indicated by bars below, and processing site D is shown. Numbers at peaks indicate the corresponding helices of 18S or 25S, while 25S peaks labeled with “C” represent background peaks also found in the control. In vivo crosslinking of Prp43-HTP-expressing cells resulted in two major peaks, whereas in vitro crosslinking of purified complexes yielded three sequence clusters. Thus, CRAC experiments have identified several potential binding sites for Prp43 on pre-rRNA.

Figure 2

Prp43 Has Distinct Binding Sites on Preribosomal RNA Shown is predicted secondary structure of regions of yeast 18S (A) and 25S rRNA (B–E), modified from Piekna-Przybylska et al. (2007). The binding sites of Prp43 are marked in gray. A boxed “M” represents a site of 2′-O-methylation in the rRNA, and large numbers indicate the corresponding snoRNA. Small numbers represent the position of the indicated residue in 18S or 25S. In (C), part of the 25S structure is excised at “a” and shown below the main model, as it projects into the background. The major binding site for Prp43 in 18S is located in helix 44 (A), close to the dimethylation sites of Dim1 (${\mathrm{m}}_2^6$A${\mathrm{m}}_2^6$A) and cleavage site D (the 3′ end of the 18S rRNA). In 25S, Prp43 binds between helices 39 and 40 (B), between helices 23 and 24 (C), over helix 34 (D), and along helix 83 (E).

Figure 3

Prp43 Crosslinking to snoRNAs and snRNAs Protein-RNA complexes from Prp43-HTP-expressing cells and control cells were crosslinked and analyzed as described for Figure 1. The percentage of sequences found for individual snoRNAs (A) or snRNAs (B) and the crosslinking sites of Prp43 on the most highly enriched snoRNAs snR51 (C), snR72 (D), snR60 (E), and the splicing snRNA U6 (F) are presented as total number of hits. The total number of hits (red line) and the positions of deletions (dashed green line) and nucleotide substitutions (dashed blue line) in the obtained sequences are shown. The regions of the guide sequences and functional elements of the snoRNAs are indicated below.

Figure 4

Prp43 Is Required for Release of snoRNAs from Preribosomes WT and P_tet-prp43 strains were pregrown in YPD medium and doxycycline was added for 6 hr to allow Prp43 depletion. Soluble cellular extracts were prepared and separated on 10%–45% sucrose gradients. RNA was extracted from gradient fractions, fractions containing either free or preribosome-bound snoRNAs were pooled, and the levels of all 75 yeast snoRNAs were determined by quantitative RT-PCR. Data from depletions were normalized to WT samples processed in parallel and average ratios of preribosome-bound to unbound snoRNAs are shown. Error bars represent standard error. snoRNAs showing an accumulation outside the 95% confidence interval (lines) are marked in gray. Depletion of Prp43 results in a shift of snR39, snR39b, snR50, snR55, snR59, snR60, and snR72 into fractions containing preribosomes, while the preribosomal levels of snR64 and snR67 are reduced.

Figure 5

snoRNAs Are Retained on Preribosomes Following Depletion of Prp43 Strains were grown and samples prepared as in Figure 4. The distribution of a set of snoRNAs was analyzed by northern hybridization for each gradient fraction. The fractions pooled for Figures 4 and 6 as the free snoRNAs (pool 1) and preribosomes (pool 2) are indicated. Bars indicate fractions containing pre-40S, pre-60, and 90S preribosomes.

Figure 6

Analysis of snoRNA Distribution in Prp43 Mutant Strains The distribution of snoRNAs dependent on Prp43 for their release from preribosomes was analyzed in strains depleted of genomic PRP43 but expressing WT or mutant Prp43 from a plasmid (see Figure S3). Samples were processed for qPCR as described in Figure 4, and averages of at least four independent experiments are shown. The line at 2.1 indicates the 95% confidence interval. Error bars represent standard error. (A) Depletion of genomic PRP43 was performed for 10 hr in medium containing 20 μg/ml doxycycline. While U3, U14, snR64, and snR65 showed no effect, only expression of WT Prp43, but not the mutants or empty vector, could prevent accumulation of snR39, snR39b, snR50, and snR59 in fractions containing preribosomes. (B) Shown is comparison of effects on snR59 distribution after 7 or 10 hr of genomic PRP43 depletion. After 7 hr depletion, cells expressing the Prp43 mutants T123A or S247A show a dominant-negative effect on snoRNA accumulation as compared to cells carrying the empty vector. After 10 hr depletion, cells carrying the empty vector showed similar effects as cells expressing one of the mutants, while expression of the WT protein resulted in the WT distribution of snR59.