RNA binding proteins Smaug and Cup induce CCR4-NOT-dependent deadenylation of the nanos mRNA in a reconstituted system

Abstract

Posttranscriptional regulation of the maternal nanos mRNA is essential for the development of the anterior - posterior axis of the Drosophila embryo. The nanos RNA is regulated by the protein Smaug, which binds to Smaug recognition elements (SREs) in the nanos 3'-UTR and nucleates the assembly of a larger repressor complex including the eIF4E-T paralog Cup and five additional proteins. The Smaug-dependent complex represses translation of nanos and induces its deadenylation by the CCR4-NOT deadenylase. Here we report an in vitro reconstitution of the Drosophila CCR4-NOT complex and Smaug-dependent deadenylation. We find that Smaug by itself is sufficient to cause deadenylation by the Drosophila or human CCR4-NOT complexes in an SRE-dependent manner. CCR4-NOT subunits NOT10 and NOT11 are dispensable, but the NOT module, consisting of NOT2, NOT3 and the C-terminal part of NOT1, is required. Smaug interacts with the C-terminal domain of NOT3. Both catalytic subunits of CCR4-NOT contribute to Smaug-dependent deadenylation. Whereas the CCR4-NOT complex itself acts distributively, Smaug induces a processive behavior. The cytoplasmic poly(A) binding protein (PABPC) has a minor inhibitory effect on Smaug-dependent deadenylation. Among the additional constituents of the Smaug-dependent repressor complex, Cup also facilitates CCR4-NOT-dependent deadenylation, both independently and in cooperation with Smaug.

Figure 1.

Reconstitution of the Drosophila melanogaster CCR4–NOT complex. (A) Scheme of different variants of the Drosophila CCR4–NOT complex. Domains and interaction sites of NOT1 are denoted as follows: MIF4G, ‘middle of 4G’ domain; CN9BD, CNOT9 (= CAF40) binding domain; TTP denotes the interaction surface. NOT1_PE is a naturally occurring isoform. (B) Purification of the ^DmCCR4–NOT_MINI complex by gel filtration. The complex was first purified by FLAG affinity chromatography and then applied to a Superose 6 column as described in Materials and Methods. Left panel, UV profile of the column; right panel, analysis of relevant fractions by SDS polyacrylamide gel electrophoresis and Coomassie staining. Subunits of the CCR4–NOT complex are labeled with black dots and a contaminating band with an asterisk in the gel. (C) Purified proteins used in this study. Left panel, preparations of different variants of the ^DmCCR4–NOT complex. Prominent contaminants are indicated with asterisks. Right panel, purified components of the SRE-dependent repressor complex. Desired polypeptides are marked. Purified proteins were separated on SDS-polyacrylamide gels and stained with Coomassie. (D) Basal activity of different variants of the ^DmCCR4–NOT complex. Purified variants of the ^DmCCR4–NOT complex (6.25 nM each), were incubated with excess FAM-7mer-A₂₀ (50 nM), and aliquots were withdrawn as indicated. Numbers at the bottom represent average deadenylation rates in nt/min plus/minus standard deviation based on n = 3.

Figure 2.

Smaug is sufficient to induce SRE-dependent deadenylation by the CCR4–NOT complex. (A) Smaug induces deadenylation by the ^DmCCR4–NOT_MINI complex. Radioactively labelled SRE^WT-A₇₀ or SRE^MUT-A₇₀ RNAs (20 nM) were pre-incubated with 80 nM Smaug or deadenylation buffer. Reactions were started by the addition of 2 nM ^DmCCR4–NOT_MINI, and samples were taken and analyzed at the times indicated. Reactions containing only CCR4–NOT or only Smaug were included as controls. The graph at the bottom represents the time-dependent accumulation of fully deadenylated RNA (average plus/minus standard deviation based on n = 3). (B) Gel filtration reveals an association of Smaug with the ^DmCCR4–NOT_MINI complex. Smaug by itself or mixed with ^DmCCR4–NOT_MINI was analyzed by gel filtration as described in Material and Methods. Column fractions from the samples containing both CCR4–NOT and Smaug were analyzed by SDS polyacrylamide gel electrophoresis and Coomassie staining (top panel) and by western blotting with an antibody against Smaug (middle panel). Column fractions derived from the Smaug-only sample were only analyzed by western blotting (bottom panel). Elution positions of size markers are indicated above the respective fractions. In the Coomassie-stained gel, Smaug is labeled with a blue triangle, and subunits of the CCR4–NOT complex are labeled with red dots. (C) Smaug induces SRE-dependent deadenylation by the human CCR4–NOT complex. Reactions were carried out as in Figure 2A except that RNA was used at 5 nM, Smaug at 30 nM, and the human CCR4–NOT_FULL complex at 10 nM. The graph at the bottom represents the time-dependent accumulation of fully deadenylated RNA. A representative experiment of n = 2 is shown. The requirement for a higher concentration of CCR4–NOT compared to the experiment in Figure 2A is at least partially explained by the reaction temperature of 25°C, which is suboptimal for the human complex. These reactions were also carried out in the absence of BSA. (D) The CCR4–NOT ‘MINI’ complex is necessary and sufficient for Smaug-dependent deadenylation. 5 nM ³²P-SRE^WT-A₇₀ RNA was preincubated with 30 nM Smaug or buffer, then deadenylation was initiated by the addition of the different human CCR4–NOT complexes (10 nM of Full, Mini and Core; 50 nM of CCR4-Caf1). Samples were taken and analyzed at the times indicated. A representative experiment of n = 2 is shown.

Figure 3.

Smaug associates with CCR4–NOT via NOT3. (A) Domain structure of NOT3. NAR, NOT1 anchor region. The NOT box mediates the interaction with NOT2. Borders of fragments used in the interaction assays are indicated at the bottom. (B) Smaug interacts with NOT3 in a relocalization assay. Drosophila proteins were fused with PH-mEGFP or mCherry as indicated and transiently coexpressed in Drosophila S2R+ cells. After two days, subcellular protein localization was examined by confocal live fluorescence microscopy. Smaug relocalization was detected only with the NOT3 subunit. ‘Control’ indicates a plasmid expressing EGFP only. Scale bar is 10 μm. Additional images supporting the Smaug – NOT3 interaction are shown in Supplementary Figure 3A. (C) The Smaug interacting surface is in the C-terminal domain of NOT3 as determined with the ReLo assay. NOT3 fragments were used as bait fusions as indicated. Fluorescent tags were swapped in comparison to (B). ‘Control’ indicates a plasmid expressing mCherry only. Scale bar is 10 μm. Additional images supporting the interaction of Smaug with the C-terminal fragment of NOT3 are shown in Supplementary Figure 3B. (D) Smaug interacts with NOT3 in a yeast two-hybrid assay. Split-ubiquitin yeast two-hybrid assays were performed with bait and prey constructs containing Drosophila proteins as indicated or no insertion (–). Three 10-fold dilutions of the cells were spotted and imaged after three days of incubation. Selection medium lacked histidine and adenine. (E) Yeast two hybrid assay confirms interaction of Smaug with the C-terminal domain of NOT3. The same fragments as in (C) were used as prey fusions in the two-hybrid assay.

Figure 4.

CCR4 and CAF1 make similar contributions to the activity of the ^DmCCR4–NOT_MINI complex. (A) Inactivation of either catalytic subunit has a similar effect on the basal activity of ^DmCCR4–NOT_MINI. Point mutations in the active sites of CCR4 and CAF1 are depicted in the cartoons. 50 nM FAM 7mer-A₂₀ RNA was incubated with the respective enzyme complexes (5 nM) for the times indicated. Numbers at the bottom represent average deadenylation rates in nt/min plus/minus standard deviation, based on n = 3. (B) Inactivation of either catalytic subunit has a similar effect on Smaug-dependent deadenylation. 40 nM Smaug was pre-incubated with 10 nM SRE^WT-A₇₀ substrate RNA, deadenylation was started by the addition of 2 nM ^DmCCR4–NOT_MINI complex and stopped at the times indicated. The right panel shows a quantification of the fully deadenylated product (average plus/minus standard deviation based on n = 3).

Figure 5.

^DmCCR4–NOT-dependent deadenylation is moderately affected by PABPC. (A) PABPC stimulates the deadenylation of FAM 7mer-A₂₀. Substrate RNA (25 nM) was incubated with 200 nM PABPC in the presence of tRNA, and deadenylation was started by the addition of 2.5 nM ^DmCCR4–NOT_MINI. A representative experiment of n = 3 is shown. (B) Both CCR4 and CAF1 can degrade a poly(A) tail bound by PABPC. Deadenylation time courses were carried out with wild-type ^DmCCR4–NOT_MINI or mutant variants as indicated. Reaction conditions were as in (A). PABPC was present at 200 nM. Numbers at the bottom report deadenylation rates (nt/min). A representative experiment of n = 2 is shown. (C) PABPC modestly stimulates deadenylation of SREonly-A₇₀. SRE^WTonly-A₇₀ RNA (5 nM) was deadenylated in the presence of the indicated concentrations of PABPC. ^DmCCR4–NOT_MINI was used at 0.5 nM. A negative control (S) contained 80 nM PABPC in the absence of CCR4–NOT. The graph shows the accumulation of completely deadenylated product (average plus/minus standard deviation based on (n = 3). The curves for 40 and 80 nM PABPC lie on top of each other. (D) PABPC inhibits SRE-dependent deadenylation. SRE^WTonly-A₇₀ RNA (5 nM) was first pre-incubated with Smaug (30 nM) or buffer for 20 min, then the indicated amounts of PABPC or buffer were added, and the incubation was continued for another 20 min. Finally, deadenylation was started by the addition of ^DmCCR4–NOT_MINI (0.5 nM). Quantification was as in (C).

Figure 6.

Smaug makes the CCR4–NOT complex processive. (A) The ^DmCCR4–NOT_MINI complex is distributive on its own. A constant concentration of the FAM 7mer-A₂₀ RNA (50 nM) was incubated with 6.25, 12.5, 25 or 50 nM ^DmCCR4–NOT_MINI, resulting in the molar ratios indicated. Aliquots were withdrawn at the time points indicated. Numbers at the bottom represent average deadenylation rates in nt/min plus/minus standard deviation, based on n = 3. (B) The ^DmCCR4–NOT_MINI complex acts processively in Smaug-dependent deadenylation. The substrate RNA in this assay was fluorescently labelled TCE^WT-A₇₀ RNA, which was used at 50 nM. The RNA was pre-incubated with or without 300 nM Smaug as indicated and the reaction started by the addition of ^DmCCR4–NOT_MINI (5 nM). Aliquots were withdrawn as indicated. Note that the time scale is in seconds. The asterisk indicates an unknown RNA species that we have not been able to remove. The graph on the right shows the accumulation of fully deadenylated product in the reaction containing Smaug (average plus/minus standard deviation based on n = 3)

Figure 7.

Cup also induces deadenylation. (A) Cup induces deadenylation by the ^DmCCR4–NOT_MINI complex. SRE^WT-A₇₀ RNA (20 nM) was pre-incubated with 80 nM Cup, and deadenylation was initiated by the addition of 2 nM ^DmCCR4–NOT_MINI. Aliquots were withdrawn at different time points as indicated. Controls included incubations in the absence of Cup, with Cup only, or no protein added. The graph shows the accumulation of completely deadenylated product in the Cup + CCR4–NOT reaction (average plus/minus standard deviation based on n = 3). (B) A scheme of Cup showing its division into an N-terminal part harbouring the eIF4E binding motifs, a middle part with the Cup homology domain (CHD) binding Me31B, and a C-terminal part rich in glutamine residues. (C) The ability of Cup to stimulate the CCR4–NOT complex is distributed over the protein. The Cup fragments shown in (A) were fused to His-λN-MBP and purified (Supplementary Figure 7B). Proteins (40 nM each, except M, which was 80 nM) were pre-incubated for 15 min either with the nLuc-2xBoxB-A₇₀ RNA (2 nM; left panel) or with the nLuc-BRE^MUT-A₇₀ RNA (2 nM, right panel). Deadenylation was started by the addition of ^DmCCR4–NOT_MINI (1 nM) and allowed to proceed for 30 min. Controls included incubations in the absence of CCR4–NOT, with CCR4–NOT only or with CCR4–NOT plus His₈-λN-MBP. A representative experiment of n = 2 is shown. (D) Cup and all its fragments can be UV-crosslinked to RNA. Left panel: Coomassie-stained SDS gel showing MBP-Cup fragments lacking λN. Right panel: Cup fragments shown in the left panel were UV-cross-linked to radiolabeled RNA. Cross-linking products were analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography. (E) Smaug and Cup jointly stimulate deadenylation. Reactions were carried out in the absence of PEG. The SRE^WT-A₇₀ RNA (10 nM) was pre-incubated with Smaug (80 nM) or with buffer, then Cup (80 nM) or buffer was added for an additional 20 min. Deadenylation was started by the addition of ^DmCCR4–NOT_MINI (1 nM). Left panel: Analysis of reaction products by denaturing gel electrophoresis. Right panel: Completely deadenylated products were quantified (average of n = 2). The broken yellow line indicates theoretical product accumulation predicted by additive behavior of Smaug and Cup.