Dual nature of translational control by regulatory BC RNAs

Abstract

In higher eukaryotes, increasing evidence suggests, gene expression is to a large degree controlled by RNA. Regulatory RNAs have been implicated in the management of neuronal function and plasticity in mammalian brains. However, much of the molecular-mechanistic framework that enables neuronal regulatory RNAs to control gene expression remains poorly understood. Here, we establish molecular mechanisms that underlie the regulatory capacity of neuronal BC RNAs in the translational control of gene expression. We report that regulatory BC RNAs employ a two-pronged approach in translational control. One of two distinct repression mechanisms is mediated by C-loop motifs in BC RNA 3' stem-loop domains. These C-loops bind to eIF4B and prevent the factor's interaction with 18S rRNA of the small ribosomal subunit. In the second mechanism, the central A-rich domains of BC RNAs target eIF4A, specifically inhibiting its RNA helicase activity. Thus, BC RNAs repress translation initiation in a bimodal mechanistic approach. As BC RNA functionality has evolved independently in rodent and primate lineages, our data suggest that BC RNA translational control was necessitated and implemented during mammalian phylogenetic development of complex neural systems.

Fig. 1.

Rodent BC1 RNA and primate BC200 RNA feature identical C-loop motifs in their 3′ stem-loop domains. (A) C-loop motifs are boxed in secondary structure representations of the two BC RNAs. The structure of BC1 RNA has previously been ascertained by chemical and enzymatic probing (34). (B) BC RNA C-loop motifs are aligned with a previously described eukaryotic C-loop motif (18). C-loop motifs are boxed in the upper panels. Canonical and noncanonical nucleotide interactions are highlighted in the lower panels. Base pairings are typified using the Leontis and Westhof scheme of symbolic representations (–18). Base pairings are color coded as follows (18): green, cis-WC/WC; blue, cis-WC/SE; purple, trans-WC/H; and yellow, cis-WC/WC. (C) Three types of mutations were introduced into BC RNA C-loop motifs. (i) Mutant WC: conversion to WC base pairings forces the 3′ stem-loops into standard A-form helices, abolishing the C-loop motifs. (ii) Mutant A-U: U-A pairs were changed to A-U pairs. Although this inversion does not alter cis-WC/WC interactions (green and yellow), C-loop motif architecture is intolerant of this change as U cannot substitute for A in motif-essential noncanonical interactions (blue and purple). (iii) Mutant Loop: the long-strand C-loop nucleotides were exchanged in a manner that continues to allow noncanonical interactions (18).

Fig. 2.

BC1 and BC200 RNA C-loop motifs bind to eIF4B. EMSA experiments were performed with ³²P-labeled BC1 RNA, BC200 RNA, and derivatives. Labeled RNAs were incubated with eIF4B (412 nM). (A) WT BC1 RNA produced a band shift with eIF4B (2nd lane) relative to the position of BC1 RNA in the absence of protein (1st lane). The intensities of the RNA-protein complexes were substantially lower with BC1 RNA C-loop mutant WC (3rd lane) than with the WT. C-loop mutant Loop produced an intensity of the shifted band similar to that produced by WT BC1 RNA (4th lane). C-loop mutant A-U produced a significantly lower intensity of the shifted band (5th lane). In contrast, central domain mutant U⁺ showed no difference from the WT (6th lane). U6 RNA failed to produce a band shift (7th lane; note that free U6 RNA exited the gel). (B) RNA-protein complex band intensities were quantified, and results were plotted as binding to eIF4B relative to the binding of WT BC1 RNA. Error bars represent standard errors of the mean (SEM); n = 4, one-way analysis of variance (ANOVA), P < 0.001; Tukey post hoc analysis, comparison with WT BC1 RNA, P < 0.001 for WC, P < 0.01 for A-U, P = 0.979 for Loop, and P = 0.744 for U⁺. (C) WT BC200 RNA produced a shift with eIF4B (2nd lane) relative to the position of BC200 RNA without eIF4B (1st lane). The intensities of the RNA-protein complexes were significantly lower with BC200 RNA C-loop mutants WC and A-U (3rd and 5th lanes, respectively) than with the WT. C-loop mutant Loop produced an intensity of the shifted band similar to that produced by WT BC200 RNA (4th lane). The intensity of the band produced by mutant U⁺ did not differ significantly differ from that of the WT (6th lane). U6 RNA did not produce a band shift (7th lane). Raw data were as follows (percentage of bound relative to total): WT, 11.5; WC, 2.6; Loop, 11.8; AU, 3.2; and U⁺, 11.0. (D) Results were plotted as binding to eIF4B relative to that of WT BC200 RNA. Error bars represent SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with WT BC200 RNA, P < 0.001 for WC, P < 0.001 for A-U, P = 0.890 for Loop, and P = 0.997 for U⁺. (E) The addition of an antibody (Ab) against eIF4B to the BC1 RNA-eIF4B complex reaction produced a supershift (2nd lane, 2 μl antibody, and 3rd lane, 3 μl antibody; compare with 1st lane in which no antibody was added). No supershift was observed following the addition of an irrelevant antibody (directed against β3-tubulin; 4th and 5th lanes). (F) In assays using labeled BC1 RNA in RRL, supershifts were observed with antibodies (3 μl each lane) against eIF4B (1st lane) and PABP (3rd lane) but not with an antibody against GST (2nd lane).

Fig. 3.

BC1 RNA and BC200 RNA compete with 18S rRNA for binding to eIF4B. (A) Formation of RNA-protein complexes between BC1 RNA and eIF4B was significantly inhibited by increasing amounts of unlabeled BC1 RNA (3rd and 4th lanes). Unlabeled U6 RNA did not inhibit the formation of BC1 RNA-eIF4B complexes (5th and 6th lanes). In contrast, unlabeled 18S rRNA inhibited the formation of BC1 RNA-eIF4B complexes (7th and 8th lanes). The 1st lane shows labeled BC 1 RNA in the absence of protein, and the 2nd lane shows the formation of BC1 RNA-eIF4B complexes in the absence of competitor RNA. (B) Quantitative analysis confirmed competition of BC1 RNA with 18S rRNA for eIF4B binding. Competitor RNA concentrations were 1.7 nM (light gray) or 3.4 nM (dark gray). Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for BC1 RNA and 18S rRNA. (C) Formation of BC200 RNA-eIF4B complexes was reduced by unlabeled competitor BC200 RNA (3rd and 4th lanes) and by unlabeled 18S rRNA (7th and 8th lanes). Unlabeled U6 RNA did not inhibit complex formation (5th and 6th lanes). The 1st lane shows labeled BC200 RNA, and the 2nd lane shows the formation of BC200 RNA-eIF4B complexes in the absence of competitor RNA. (D) Quantitative analysis confirmed competition of BC200 RNA with 18S rRNA for eIF4B binding. Competitor RNA concentrations were 1.7 nM (light gray) or 3.4 nM (dark gray). Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for BC200 RNA and 18S rRNA.

Fig. 4.

The central A-rich domain of BC1 RNA mediates binding to eIF4A. Labeled RNAs were incubated with eIF4A (820 nM). WT BC1 RNA produced a band shift with eIF4A (2nd lane) relative to the position of BC1 RNA in the absence of protein (1st lane). RNA-protein complexes were also observed with BC1 RNA C-loop mutants WC (3rd lane), Loop (4th lane), and A-U (5th lane). In contrast, central domain mutant U⁺ failed to produce a band shift (6th lane). Similarly, U6 RNA failed to interact with eIF4A (7th lane).

Fig. 5.

The BC RNA A-rich central domains mediate repression of eIF4A helicase activity. RNA duplexes (12/44 nt, ³²P-labeled on the 12-nt short strand) were used as helicase substrates in the presence of 1 mM ATP. (A and B) The BC1 RNA central domain mediates repression of eIF4A helicase activity. Labeled RNA monomer and RNA duplexes in the absence of eIF4A were run in the 1st and 2nd lanes, and eIF4A-unwound duplexes were run in the 3rd lane. U6 RNA did not significantly inhibit unwinding (4th lane). WT BC1 RNA inhibited RNA duplex unwinding (5th lane). BC1 RNA C-loop mutants WC, Loop, and A-U (6th to 8th lanes) all inhibited unwinding in a manner indistinguishable from the results for WT BC1 RNA. In contrast, central domain mutant U⁺ did not inhibit unwinding (9th lane). (B) Quantitative analysis confirmed the above-described results. Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for WT BC1 RNA and mutants WC, Loop, and A-U and P = 0.970 for mutant U⁺; comparison with WT BC1 RNA, P < 0.001 for U6 RNA and mutant U⁺, P = 0.803 for WC, P = 0.891 for Loop, and P = 0.969 for A-U. (C and D) The BC200 RNA central domain mediates repression of eIF4A helicase activity. WT BC200 RNA significantly repressed eIF4A helicase activity (2nd lane) while U6 RNA did not (1st lane). BC200 RNA C-loop mutants WC, Loop, and A-U (3rd to 5th lanes) inhibited eIF4A helicase activity in a manner indistinguishable from the results for WT BC200 RNA. In contrast, central domain mutant U⁺ (6th lane) did not significantly inhibit duplex unwinding (indistinguishable from the results for U6 RNA in the 1st lane). (D) The above-described results were confirmed by quantitative analysis. Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for WT BC200 RNA and mutants WC, Loop, and A-U and P = 0.976 for mutant U⁺; comparison with WT BC200 RNA, P < 0.001 for U6 RNA and mutant U⁺ and P = 1.0 for mutants WC, Loop, and A-U. (E and F) eIF4B-stimulated eIF4A helicase activity is repressed by BC1 RNA via its central A-rich domain. Labeled RNA duplexes in the absence of protein were run in the 1st lane. eIF4A (2nd lane) or eIF4A and eIF4B in combination (3rd lane) were added. WT BC1 RNA and BC1 RNA C-loop mutants WC, Loop, and A-U inhibited RNA duplex unwinding (5th to 8th lanes) in a manner indistinguishable from each other. In contrast, neither BC1 RNA central domain mutant U⁺ (9th lane) nor U6 RNA (4th lane) inhibited unwinding. (F) Quantitative analysis confirmed the above-described results. Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for WT BC1 RNA and C-loop mutants WC, Loop, and A-U and P = 0.998 for mutant U⁺; comparison with WT BC1 RNA, P < 0.001 for U6 RNA and mutant U⁺ and P = 1.0 for mutants WC, Loop, and A-U.

Fig. 6.

Translational repression by BC RNAs is dualistic and mediated both by their 3′ domain C-loop motifs, interacting with eIF4B, and by their central A-rich domains, interacting with eIF4A. (A) WT BC1 RNA inhibited translation of CAT mRNA in RRL, whereas U6 RNA did not. The translation repression competence of BC1 RNA C-loop mutants was diminished to various degrees (compared to that of WT BC1 RNA), whereas BC1 RNA central domain mutant U⁺ (lane 6) was completely ineffective in repressing translation, in a manner that was indistinguishable from the results for U6 RNA. (B) For quantitative analysis, the intensities of the protein product bands were normalized against the one obtained in the presence of U6 RNA, as described previously (19). Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for WT BC1 RNA and mutants WC, Loop, and A-U and P = 0.965 for mutant U⁺; comparison with WT BC1 RNA, P < 0.001 for U6 RNA and mutants WC, A-U, and U⁺ and P = 0.485 for mutant Loop. (C) Regression analysis showed that binding to eIF4B and translational repression by BC1 RNA are tightly correlated, indicating that complete binding inability results in complete inability to mediate repression. Relative quantitative repression competence (from the experiments whose results are shown in panels A and B) was plotted against relative quantitative binding to eIF4B (from the experiment whose results are shown in Fig. 2) for the RNAs indicated. A correlation coefficient of R² = 0.96075 was obtained. Error bars show SEM. (D) Analogous to BC1 RNA, BC200 RNA inhibited translation of CAT mRNA in RRL. The repression competence of BC200 RNA C-loop and central domain mutants was diminished to degrees similar to the results for the respective BC1 RNA mutants. (E) BC200 RNA quantitative analysis results are as follows. Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for WT BC200 RNA and mutants WC, Loop, and A-U and P = 0.996 for mutant U⁺; comparison with WT BC200 RNA, P < 0.001 for U6 RNA and mutant U⁺, P < 0.01 for mutants WC and A-U, and P = 1.0 for mutant Loop. (F) Recombinant eIFs 4A and 4B restored in vitro translation in RRL. No protein was added in the 1st and 2nd lanes. Recombinant eIF4A was added in the 3rd lane, recombinant eIF4B protein in the 4th lane, and both recombinant factors in the 5th lane. (G) Quantitative analysis results are as follows. Error bars show SEM; n = 4, one-way ANOVA, P < 0.001; Tukey post hoc analysis, comparison with U6 RNA, P < 0.001 for WT BC1 RNA, addition of eIF4A, addition of eIF4B, and addition of eIFs 4A and 4B in combination.

Fig. 7.

Conservation of BC1 C-loop motifs in rodents and BC200 C-loop motifs in primates (Anthropoidea). The 3′ domains of BC1 RNA (top) and BC200 RNA (bottom) are given for all published sequences. C-loop motif nucleotides, shaded in light blue, are invariant across all species shown with the exception of a U-to-C transition in four primate species. This transition results in the replacement of standard WC U-A pairs with wobble WC C · A pairs. This pair (U165-A185 in human BC200 RNA) is the basal C-loop WC pair. While U-A pairs are most frequently observed in this position in eukaryotic C-loops, C · A pairs are not uncommon (16). A second U-A to C · A transition occurs at position 170 to 181, i.e., outside the C-loop motif. (Sequence data are from references and .) Species: Rattus norvegicus (Rno), Mus musculus (Mmu), Mesocrietus auratus (Mau), Cavia porcellus (Cpo), Homo sapiens (Hsa), Pan paniscus (Ppa), Gorilla gorilla (Ggo), Pongo pygmaeus (Ppy), Hylobates lar (Hla), Papio hamadryas (Pha), Chlorocebus aethiops (Cae), Macaca mulatta (Mmu), Macaca fascicularis (Mfa), Aotus trivirgatus (Atr), Saguinus imperator (Sim), and Saguinus oedipus (Soe). Nucleotide numbering is given for Rno (BC1 RNA) and Hsa (BC200 RNA).