The ribosomal protein Rpl22 controls ribosome composition by directly repressing expression of its own paralog, Rpl22l1

Abstract

Most yeast ribosomal protein genes are duplicated and their characterization has led to hypotheses regarding the existence of specialized ribosomes with different subunit composition or specifically-tailored functions. In yeast, ribosomal protein genes are generally duplicated and evidence has emerged that paralogs might have specific roles. Unlike yeast, most mammalian ribosomal proteins are thought to be encoded by a single gene copy, raising the possibility that heterogenous populations of ribosomes are unique to yeast. Here, we examine the roles of the mammalian Rpl22, finding that Rpl22(-/-) mice have only subtle phenotypes with no significant translation defects. We find that in the Rpl22(-/-) mouse there is a compensatory increase in Rpl22-like1 (Rpl22l1) expression and incorporation into ribosomes. Consistent with the hypothesis that either ribosomal protein can support translation, knockdown of Rpl22l1 impairs growth of cells lacking Rpl22. Mechanistically, Rpl22 regulates Rpl22l1 directly by binding to an internal hairpin structure and repressing its expression. We propose that ribosome specificity may exist in mammals, providing evidence that one ribosomal protein can influence composition of the ribosome by regulating its own paralog.

Figure 1. Mouse Rpl22 has an expressed paralog, Rpl22l1.

(A) Alignment of Rpl22 and Rpl22l1 protein sequences. Lung, liver, spleen, kidney and pancreas, harvested from Rpl22^−/− mice and their littermate controls, were analyzed for relative Rpl22 and Rpl22l1 mRNA levels by qRT-PCR (B) or protein expression by Western blot analysis (C). Results are representative of 3 independent experiments.

Figure 2. Both mouse Rpl22 and Rpl22l1 proteins can be incorporated into ribosomes.

Liver tissue was isolated from Rpl22^+/+ (A, C) and Rpl22^−/− (B, D) mice then, after sedimentation of the lysates on sucrose gradients, fractions were collected and loaded onto an SDS-page gel for western blot analysis (C and D, respectively). Images are representative of 3 independent experiments. Multiple Reaction Monitoring Mass spectrometry (MRM-MS) analysis was performed on free 60S subunits and 80S monosomes from actively translating polysomes. Liver lysates from Rpl22^+/+ and Rpl22^−/− mice were subjected to a brief treatment with low amounts of RNase A to degrade mRNA between ribosomes in polysomes and release the ribosomes as 80S monomers. After inhibiting the RNase with KCl and heparin, the samples were fractionated on 10–30% sucrose gradients containing 800 mM KCl to disrupt any “nonproductive couples” of 40S and 60S subunits . (E, F) Representative gradient profiles for Rpl22+/+ and Rpl22−/−samples, respectively. Fraction 6 was used to isolate 60S subunits for MRM-MS, while fraction 7 was used to isolate 80S monosomes. Summation of the integrated MRM peak areas for all transitions from all observed peptides for Rpl22 (G) and Rpl22l1 (H) proteins yielded the total MRM peak areas plotted for each of the four tested samples (WT60S, WT80S, KO60S, KO80S), indicating the relative amounts of Rpl22 and Rpl22l1 in these samples. The height of each bar represents the average of the three technical replicates performed for the given sample, and each error bar represents +/−1 standard deviation. p-values>3E-3 in both cases by paired student's t-test. Rpl22 peptides: AGNLGGGVVTIER; ITVTSEVPFSK; YFQINQDEEEEEDED. Rpl22l1 peptides: TGNLGNVVHIER; ITVVSEK.

Figure 3. Acute knockdown of Rpl22 enhances Rpl22l1 expression.

3T9 cells were transduced with doxycycline-inducible shRNA lentiviral constructs directed at Rpl22 (shRNA 1 and shRNA 2) or a non-specific (ns) shRNA construct and treated with doxycycline for 3 days to induce shRNA expression. NS, shRNA 1, or shRNA 2 expressing cells were analyzed for relative Rpl22 and Rpl22l1 mRNA levels by qRT-PCR (A) or protein expression by Western blot analysis (B). Results are the average ± SEM of 4 independent experiments (A) or representative of 3 independent experiments (B). Statistical significance is indicated (*; p<0.001 compared to untreated control).

Figure 4. Rpl22 directly binds Rpl22l1 mRNA to regulate its expression levels.

(A) In the absence of Rpl22, Rpl22l1 mRNA levels are more stable in the presence of Actinomycin D. Rpl22+/+ or Rpl22−/− 3T9 cells were treated with Actinomycin D (1 µM final concentration) and total RNA was harvested at the time points shown. Levels of Rpl22l1 mRNA were quantitated by qRT-PCR. Results are the average ± SEM of 3 independent experiments and the statistical significance indicated is (*, p<0.01, compared to Rpl22+/+ untreated; ** p<0.001, compared to Rpl22+/+ at each time point). (B) M-fold analysis of zRpl22l1 mRNA reveals the presence of a consensus Rpl22 RNA-binding motif. In green are the residues deleted to remove the hairpin (zRpl22l1Δhp). In blue are the residues known to be essential for Rpl22 binding. (C) Autoradiogram of ribonuclease protection assay reveals Rpl22 protein binds to Rpl22l1 mRNA and this binding is abrogated upon removal of the hairpin. 32P labeled EBER1 (positive control), EBER 2 (negative control), zRpl22l1 or zRpl22l1Δhp RNAs were incubated in the absence or presence of GST-Rpl22 (41.7 kDa), GST (27 kDa) or m88, a GST-Rpl22 RNA binding mutant (41.6 kDa), as indicated, then UV-cross-linked, digested with RNase A, and run on a SDS protein gel. GST-Rpl22 was detected, hence, bound to EBER1 and zRpl22l1 RNAs but not Rpl22l1Δhp RNA, indicating Rpl22 binds to Rpl22l1 mRNA and this binding is abrogated upon removal of the hairpin. Numbers indicate molecular weight protein ladder in kDa.

Figure 5. Regulation of Rpl22l1 mRNA expression is mediated by a hairpin structure.

(A) Schematic representation of the biosensor quantification assay. (B–D) Stereoimages of zebrafish embryos illustrate that co-injection of zRpl22 repressed fluorescence derived from an EGFP-Rpl22l1 fusion protein upon injecting mRNA for both and assessing fluorescence at 6 hours post fertilization. Rpl22, Rpl22l1 or mutated Rpl22l1 (Rpl22l1mt) coding sequence was fused to EGFP mRNA and co-injected with mCherry mRNA (injection control) along with the corresponding inhibitor mRNAs (zRpl22 or zRpl22l1) into 1-cell stage zebrafish embryos. (F) Schematic representation of the experimental procedure. A zRpl22l1-150h-EGFP heterologous reporter mRNA, containing the minimal sequence identified by mFOLD to form the hairpin structure, was co-injected with mCherry mRNA (injection control) and (G) Rpl22 mRNA or (I) Rpl22-Morpholino (Rpl22-MO) into 1-cell stage zebrafish embryos. (E, H, J) At 10 hpf, the relative fluorescence intensity was calculated and normalized to control injections (n = 3, each group). Data are shown as mean ± standard deviation (s.d.).

Figure 6. Acute knockdown of Rpl22l1 expression impairs cellular growth.

(A) Lysates isolated Rpl22^+/+ and Rpl22^−/− 3T9 cells treated with or without doxycycline were collected to assess levels of Rpl22 or Rpl22L1 by Western blot analysis. GAPDH was used as a loading control. (B) Growth of Rpl22^+/+ and Rpl22^−/− 3T9 cells was compared. Levels of Rpl22l1 were analyzed by Western blot analysis to confirm that the Rpl22l1-shRNA knocked down levels of Rpl22l1 in doxycycline-treated Rpl22^+/+ (C) and Rpl22^−/− (E) 3T9 cells transduced with the shRNA construct. Growth of Rpl22^+/+ (D) and Rpl22^−/− (F) 3T9 cells transduced with each shRNA construct was determined. Results are representative of 2 independent experiments with error bars representative of ±SD. Statistical significance is indicated (*, p<0.05 compared to untreated control).