The role of the Schizosaccharomyces pombe gar2 protein in nucleolar structure and function depends on the concerted action of its highly charged N terminus and its RNA-binding domains

Abstract

Nonribosomal nucleolar protein gar2 is required for 18S rRNA and 40S ribosomal subunit production in Schizosaccharomyces pombe. We have investigated the consequences of the absence of each structural domain of gar2 on cell growth, 18S rRNA production, and nucleolar structure. Deletion of gar2 RNA-binding domains (RBDs) causes stronger inhibition of growth and 18S rRNA accumulation than the absence of the whole protein, suggesting that other factors may be titrated by its remaining N-terminal basic/acidic serine-rich domain. These drastic functional defects correlate with striking nucleolar hypertrophy. Point mutations in the conserved RNP1 motifs of gar2 RBDs supposed to inhibit RNA-protein interactions are sufficient to induce severe nucleolar modifications but only in the presence of the N-terminal domain of the protein. Gar2 and its mutants also distribute differently in glycerol gradients: gar2 lacking its RBDs is found either free or assembled into significantly larger complexes than the wild-type protein. We propose that gar2 helps the assembly on rRNA of factors necessary for 40S subunit synthesis by providing a physical link between them. These factors may be recruited by the N-terminal domain of gar2 and may not be released if interaction of gar2 with rRNA is impaired.

Figure 1

(A) Schematic representation of gar2 protein and its derivatives used in this study. Hatched box, N-terminal basic/acidic serine-rich domain; light gray boxes, RBDs containing conserved RNP1 motifs (checkered boxes); vertical-hatched box, GAR domain; black box, HA epitope; *, Y306, Y308, F409, and Y411 residues have been replaced with leucines. (B) Effects of gar2 mutations on cell growth. Growth comparisons were done by spotting each strain onto solid minimal medium after growing for 24 h in liquid medium at 30°C under derepressive conditions. The plates were incubated for 8 d at 19°C. From left to right, 25000, 5000, 1000, 200, and 40 cells were spotted.

Figure 2

Effects of gar2 mutations on 18S rRNA accumulation. 25-to-18S ratios were measured on three samples of RNAs from each strain, after extraction, resolution on agarose gel under denaturing conditions, ethidium bromide staining, and observation under UV light to evaluate the deficit in 18S rRNA. Overexpression of gar2 either without its N-terminal or GAR domain restores normal 18S rRNA accumulation in the Sp91 (gar2-null) strain. Overexpression of gar2 lacking its RBDs increases the deficit in 18S rRNA, and point mutations in the conserved RNP1 motifs of the RBDs prevent total restoration of 18S level. The Sp108 strain expressing the gar2ΔRBDs protein under the control of the gar2 promoter also exhibits stronger deficit in 18S rRNA than the gar2-null strain.

Figure 3

Electron microscopic transversal sections of actively growing S. pombe cells at 30°C. (No, nucleolus; Np, nucleoplasm). (A) A wild-type (Sp15) strain. (B) A gar2-null strain (Sp91) with a slightly larger but structurally well-organized nucleolus. (C) Sp91 expressing gar2ΔRBDs protein presents an unusual hypertrophied nucleolus. (D) Sp15 expressing gar2ΔRBDs protein exhibits the same nucleolar hypertrophy. (E) Sp91 overexpressing gar2 mutated in the RNP1 motifs (gar2RNP1*) presents local dense fibrillar structures. (F) The peculiar fibrillar structures are absent from the nucleolus of Sp91 expressing gar2RNP1* lacking the N-terminal end of the protein, indicating that these modifications are dependent on the presence of the N-terminal domain of gar2 when binding to RNA is impaired. Bar, 500 nm.

Figure 4

Immunostaining of various nucleolar antigens on S. pombe cells. (A) Immunodetection with anti-HA antibodies on Sp91 (gar2-null) overexpressing gar2–HA. The nucleolus is normal, and the protein is exclusively nucleolar. (B) Immunodetection with anti-HA antibodies of gar2ΔRBDs overexpressed in Sp91. The truncated protein is both in the large fibrillar nucleolus and in the nucleoplasm. (C and D) Anti-HA (C) and anti-Spgar1 (D) antibodies on Sp91 overexpressing gar2RNP1*. gar2 precisely colocalizes with the dense structures in the nucleolus, which appears larger than these dense zones according to gar1 localization. Bar, 500 nm.

Figure 5

gar2 and gar2ΔRBDs migrations in glycerol velocity gradients. Equivalent amounts of total extracts from Sp91 (gar2-null) overexpressing gar2–HA or gar2ΔRBDs–HA were loaded onto 10–30% glycerol gradients and subjected to centrifugation. Both proteins were detected with anti-HA antibodies. Whereas gar2 belongs to 40–60S sedimentation coefficient complexes, gar2ΔRBDs protein is found either free at the top of the gradient or associated with structures larger than those with which the wild-type protein is associated.

Figure 6

Far-Western blots on S. pombe ribosomal proteins from the small (1) or the large (2) subunit. Blots were incubated with 5 μg/ml recombinant gar2–T7 (A), gar2ΔNterm–T7 (B), gar2ΔRBDs–T7 (C), or gar2RNP1*–T7 (D) proteins. Protein–protein interactions were revealed by immunostaining with anti-T7 antibodies and ECL. The full-length gar2 protein interacts in vitro with ribosomal proteins from both subunits (A), and affinity of gar2ΔNterm for ribosomal proteins is identical to that of wild-type gar2 (B). The affinity of gar2ΔRBDs for ribosomal proteins in vitro is generally low, suggesting strong alteration of gar2 structure after the deletion of the RBDs. Only a few interactions are significantly lost (C). gar2RNP1* has the same affinity for ribosomal proteins in vitro as wild-type gar2 (D).