Structural and functional studies of Bud23-Trm112 reveal 18S rRNA N7-G1575 methylation occurs on late 40S precursor ribosomes

Abstract

The eukaryotic small ribosomal subunit carries only four ribosomal (r) RNA methylated bases, all close to important functional sites. N(7)-methylguanosine (m(7)G) introduced at position 1575 on 18S rRNA by Bud23-Trm112 is at a ridge forming a steric block between P- and E-site tRNAs. Here we report atomic resolution structures of Bud23-Trm112 in the apo and S-adenosyl-L-methionine (SAM)-bound forms. Bud23 and Trm112 interact through formation of a β-zipper involving main-chain atoms, burying an important hydrophobic surface and stabilizing the complex. The structures revealed that the coactivator Trm112 undergoes an induced fit to accommodate its methyltransferase (MTase) partner. We report important structural similarity between the active sites of Bud23 and Coffea canephora xanthosine MTase, leading us to propose and validate experimentally a model for G1575 coordination. We identify Bud23 residues important for Bud23-Trm112 complex formation and recruitment to pre-ribosomes. We report that though Bud23-Trm112 binds precursor ribosomes at an early nucleolar stage, m(7)G methylation occurs at a late step of small subunit biogenesis, implying specifically delayed catalytic activation. Finally, we show that Bud23-Trm112 interacts directly with the box C/D snoRNA U3-associated DEAH RNA helicase Dhr1 supposedly involved in central pseudoknot formation; this suggests that Bud23-Trm112 might also contribute to controlling formation of this irreversible and dramatic structural reorganization essential to overall folding of small subunit rRNA. Our study contributes important new elements to our understanding of key molecular aspects of human ribosomopathy syndromes associated with WBSCR22 (human Bud23) malfunction.

Keywords: S-adenosyl-l-methionine; methyltransferase; rRNA modifying enzyme; ribosome synthesis; translation.

Fig. 1.

Crystal structures of S. cerevisiae Bud23–Trm112 MTase complexes. (A) Ribbon representation of the complex between Trm112 (green) and Bud23 (blue). The zinc atom bound to Trm112 is depicted as a dark gray sphere, and the side chains of the cysteine residues coordinating this zinc atom are shown as sticks. The SAM cofactor bound to Bud23 is shown as sticks with carbon atoms shown in light gray. The SAM methyl group (CH₃) transferred during the methylation reaction is shown as a gray sphere. Secondary structure elements of Trm112 are indicated in italics. Bud23 secondary structure elements are annotated according to the nomenclature proposed in ref. . (B) Conformational changes induced in Bud23 upon SAM binding. Residues I12–G24 from Bud23 are shown in orange as they become ordered (I12–H20) or rearrange to fold as the C-terminal extremity of helix αY (K21–G24) upon SAM binding. The conformation adopted by Bud23 residues K21–G24 in the absence of SAM is shown in magenta. The Bud23 loop encompassing residues A126–N131, which is only visible in the SAM-bound form, is also shown in orange. The side chain of W122, which adopts different conformations in the apo- (magenta) and SAM-bound (blue) structures, is shown as sticks. Bud23 residues involved in SAM binding are shown as sticks, and hydrogen bonds formed between Bud23 residues and SAM are depicted by black dashed lines. Residues mutated in this study are underlined. (C and D) Detailed views of the interface between Bud23 and Trm112. Residues involved in the interaction are shown as sticks. Electrostatic interactions (hydrogen bonds and salt bridges) are depicted as black dashed lines. Color code as in A.

Fig. 2.

Bud23 catalytic pocket and interaction of Bud23–Trm112 with pre-ribosomes. (A) Superimposition of Bud23 (blue) onto the structure of C. canephora xanthosine MTase (purple) (35) bound to xanthosine (Xant; green sticks) and comparison of their active sites. The SAM cofactor bound to Bud23 is shown as sticks with carbon atoms shown in light gray. The SAM methyl group (CH₃) transferred during the methylation reaction is shown as a gray sphere and the recipient N⁷ atom by a blue sphere. Important residues of the xanthosine MTase active site and equivalent residues of Bud23 are shown as sticks. (B) Model of a GMP molecule (yellow) bound to the Bud23 active site. Model based on the superimposition shown in A and generated by superimposing the guanine ring of GMP onto xanthosine. Conserved residues of the Bud23 active site that could be involved in GMP binding are shown as sticks, and hydrogen bonds that these residues could form with GMP are depicted by black dashed lines. The distance between the SAM methyl group (gray sphere) and the guanine N⁷ atom (blue sphere) is 2.6 Å and indicated by a red dashed line. Residues mutated in this study are underlined. Water molecules are depicted by red spheres. (C) Distribution of Bud23–Trm112 in velocity gradients. Identical amounts of total extract from the indicated strains were resolved on 10–50% sucrose gradients. Total protein extracted from each fraction was collected and analyzed by Western blotting with antibodies specific to Bud23 and Trm112. (D) Quantification of panels shown in C. Signals captured with the BioRad ChemiDoc MP imaging system. x axis, fractions 1–24; y axis, percent signal intensity for the indicated protein in each fraction.

Fig. 3.

Bud23–Trm112 complex interacts directly with Dhr1. (A, Left) Superposition of three different gel filtration elution profiles: Bud23–Trm112 complex alone (OD signal at 280 nm, black dotted line), Dhr1[58–270] alone (OD signal at 220 nm, gray dashed line), and the Bud23–Trm112–Dhr1[58–270] ternary complex after in vitro reconstitution (addition of Dhr1 to Bud23–Trm112, OD signal at 280 nm, black solid line). Black solid line, the minor peak observed (volume 15–16 mL) corresponds to the excess of Dhr1 [58–270] used to reconstitute the ternary complex. (Right) Analysis of the content of the major elution peak observed for each experiment on a 15% SDS/PAGE. MW, molecular weight marker. (B) The distribution of Bud23–Trm112 in velocity gradients is affected upon alteration of Dhr1 expression. (i and ii) Sucrose gradient analysis (Fig. 2) of total extracts of pMET::dhr1 cells grown in the absence or presence of methionine (met). The same amount of total extract, according to the OD₂₆₀, was loaded on each gradient. All Western blot membranes were exposed to film under the same conditions (i), or the luminescent signal was quantitated with a ChemiDoc (BioRad) and expressed as a percentage detected in each fraction (ii). (iii) Depletion of Dhr1 leads to inhibition of pre-rRNA processing at sites A₁ and A₂, with concomitant accumulation of aberrant 22S RNA. Total RNA extracted from pMET::dhr1 cells grown in the absence or presence of methionine for time points indicated was analyzed by Northern blotting with a probe specific to ITS1 (oligo b; Fig. 4B). (iv) Drop assay (1x and 10x dilution) on synthetic medium containing methionine or not. Cells were grown for 3 d at 30 °C.

Fig. 4.

m⁷G methylation is a late maturation event occurring on 3′-extended 18S rRNA precursors. (A) Mature yeast 40S ribosomal subunit (PDB ID codes 3U5B and 3U5C) with G1575 highlighted in red; ribosomal proteins in gray; 18S rRNA in orange. (Inset) Superimposition of Bud23–Trm112–GMP model (blue) on G1575 (yellow sticks) in 40S reveals major steric clashes. (B) Simplified 18S rRNA synthesis pathway showing pre-rRNA intermediates (35S, 33S, 32S and 20S), cleavage sites (A₁, A₂, A₃, and D), oligonucleotides a and b, and the positions and timing of the m⁷G₁₅₇₅ and $m_2^6{A}_{1781}{m}_2^6{A}_{1782}$ modifications. (C and D) Primer extension mapping of m⁷G₁₅₇₅ and $m_2^6{A}_{1781}{m}_2^6{A}_{1782}$. Total RNA extracted from the strains indicated was processed for primer extension analysis with oligonucleotide a or b. For m⁷G detection, total RNA was first cleaved at the site of modification by NaBH₄/aniline treatment (17). For $m_2^{6}A$ detection, total RNA was directly subjected to primer extension (7). Lanes 1–2 and 3–4 in D show two independent biological replicates. Structural stops (*) indicate that gel loading was even. (C) m⁷G₁₅₇₅ is strictly dependent on Bud23 and Trm112. Dim1 dimethylation at A1781/A1782 requires neither Bud23 nor Trm112. (D) In wild-type cells (lanes 1 and 4), the presence of $m_2^6{A}_{1781}{m}_2^6{A}_{1782}$ on 20S pre-rRNA precludes detection of m⁷G₁₅₇₅ from primer b because the reverse transcriptase cannot get across position 1781/1782. The m⁷G modification is detected only in catalytically deficient dim1 cells (dim1-Y131G, lanes 2 and 3). Pre-rRNAs selected by primer b on total RNA include 20S, 32S, 33S, and 35S. Because 32S, 33S, and 35S are naturally nondimethylated and because no m⁷G signal is detected in lanes 1 and 4, we conclude that m⁷G takes place at the level of 20S pre-rRNAs.