Full Text

Turn on search term navigation

ARTICLE

Received 26 Jul 2016 | Accepted 15 Dec 2016 | Published 6 Feb 2017

Annamaria Sgromo1, Tobias Raisch1, Praveen Bawankar1, Dipankar Bhandari1, Ying Chen1, Duygu Kuzuolu-zturk1, Oliver Weichenrieder1 & Elisa Izaurralde1

Human (Hs) Roquin1 and Roquin2 are RNA-binding proteins that promote mRNA target degradation through the recruitment of the CCR4-NOT deadenylase complex and are implicated in the prevention of autoimmunity. Roquin1 recruits CCR4-NOT via a C-terminal region that is not conserved in Roquin2 or in invertebrate Roquin. Here we show that Roquin2 and Drosophila melanogaster (Dm) Roquin also interact with the CCR4-NOT complex through their C-terminal regions. The C-terminal region of Dm Roquin contains multiple motifs that mediate CCR4-NOT binding. One motif binds to the CAF40 subunit of the CCR4-NOT complex. The crystal structure of the Dm Roquin CAF40-binding motif (CBM) bound to CAF40 reveals that the CBM adopts an a-helical conformation upon binding to a conserved surface of CAF40. Thus, despite the lack of sequence conservation, the C-terminal regions of

Roquin proteins act as an effector domain that represses the expression of mRNA targets via recruitment of the CCR4-NOT complex.

DOI: 10.1038/ncomms14307 OPEN

A CAF40-binding motif facilitates recruitment of the CCR4-NOT complex to mRNAs targeted by Drosophila Roquin

1 Department of Biochemistry, Max Planck Institute for Developmental Biology, Spemannstrasse 35, Tbingen 72076, Germany. Correspondence and requests for materials should be addressed to E.I. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

The CCR4-NOT deadenylase complex plays a central role in bulk mRNA degradation by catalysing the removal of mRNA poly(A) tails, which is the rst step in general

mRNA decay1. In addition to its role in global mRNA degradation, the CCR4-NOT complex regulates the expression of a large number of specic mRNAs, to which it is recruited via interactions with RNA-associated proteins. Consequently, CCR4-NOT functions as a major downstream effector complex in posttranscriptional mRNA regulation in eukaryotes.

The CCR4-NOT complex consists of several structurally and functionally distinct modules, which assemble around the NOT1 scaffold subunit1. NOT1 contains several a-helical domains that provide binding surfaces for the individual modules. A central domain of NOT1 that is structurally related to the middle portion of eIF4G (termed the NOT1 MIF4G domain) provides a binding site for the catalytic module, which comprises two deadenylases, namely CAF1 or its paralogue POP2 (also known as CNOT7 and CNOT8, respectively, in humans), and CCR4a or its paralogue CCR4b (also known as CNOT6 and CNOT6L, respectively, in humans). The NOT1 MIF4G domain also serves as a binding platform for the DEAD-box protein DDX6 (also known as RCK), which functions as a translational repressor and decapping activator2,3. C-terminal to the MIF4G domain, NOT1, contains a CAF40/NOT9-binding domain, CN9BD, that binds to the highly conserved CAF40 subunit, which is also known as CNOT9 (refs 2,3), followed by a NOT1 superfamily homology domain SHD, which interacts with NOT2-NOT3 heterodimers to form the NOT module4,5.

The CAF40 and NOT modules have no catalytic activity and have been implicated in mediating interactions with RNA-associated proteins that recruit the CCR4-NOT complex to their targets210. These proteins include the GW182 family, which is involved in miRNA-mediated gene silencing in animals2,3, tristetraprolin (TTP), a protein required for the degradation of mRNAs containing AU-rich elements6, the germline determinant Nanos79 and the human Roquin1 and Roquin2 proteins10.

The vertebrate Roquin proteins are negative regulators of T follicular helper cell differentiation and autoimmunity in vertebrates11,12. There are two partially redundant paralogues, Roquin1 and Roquin2 (initially named membrane-associated nucleic acid-binding protein), in vertebrates and only one family member in invertebrate species11,12. The proteins feature an N-terminal folded region followed by a C-terminal extension of variable length and low sequence complexity that is predicted to be predominantly unstructured13,14 (Fig. 1a). The N-terminal region contains a RING nger E3 ubiquitin ligase domain, a ROQ RNA-binding domain anked by a bilobed HEPN domain and a CCCH-type zinc nger domain, all of which are highly conserved in metazoans and dene the protein family1420. The RING domains of Hs Roquin2 and the Caenorhabditis elegans homologue of Roquin1, RLE-1, (regulation of longevity by E3) exhibit E3 ubiquitin ligase activity21,22.

The ROQ domain of Hs Roquin1 and Roquin2 recognizes specic stem-loop structures in the 30-untranslated region (UTR)

of target mRNAs. These targets include mRNAs encoding regulators of inammation such as the inducible T-cell costimulator, the costimulatory receptor Ox40, neuropilin-1, interleukin-6, interferon g (IFN-g) and the tumor necrosis factor-a (TNF-a)10,14,16,2331. Hs Roquin1 and Roquin2 downregulate these mRNA targets through interactions with the CCR4-NOT deadenylase complex and decapping factors10,32.

For Hs Roquin1, it has been shown that the interaction with the CCR4-NOT complex is mediated by the C-terminal region of the protein that is conserved only among vertebrate Roquin1 orthologues10. However, it is not known how Roquin2 recruits

the CCR4-NOT complex, because its C-terminal region shows no similarity with that of Roquin1. In addition, the C-terminal regions of the invertebrate Roquin proteins are highly divergent18,2022, and it is unclear whether Roquin proteins recruit the CCR4-NOT complex in invertebrates. However, the conservation of the ROQ domain indicates that Roquin proteins also bind RNA in invertebrates, although their specic RNA targets are currently unknown.

Despite extensive information regarding the mode of RNA recognition by Roquin proteins1519, a detailed molecular understanding of how the proteins interact with the CCR4-NOT complex is lacking, and it is not even known whether the interactions are direct. Here we investigate the molecular details of how Roquin proteins recruit the CCR4-NOT complex. First, we show that Hs Roquin2 and Dm Roquin interact with the CCR4-NOT complex and promote target mRNA degradation via their C-terminal regions, suggesting conserved functional principles among all Roquin proteins. Furthermore, we nd that the Dm Roquin C-terminal region contains multiple binding sites for the CCR4-NOT complex and that these sites act redundantly to promote mRNA degradation. Among these sites, we identify a short linear motif (SLiM) that is necessary and sufcient to mediate direct binding to the CAF40 subunit of the CCR4-NOT complex. This motif is termed the CAF40-binding motif (CBM), and we determine its crystal structure bound to CAF40. Structure-based mutations of the CAF40-CBM interface prevent binding of Dm Roquin to CAF40 and reduce the ability of the protein to degrade mRNA targets, indicating that CAF40 is an important mediator of the recruitment of the CCR4-NOT complex. Together with previous studies10, our results reveal a common role of the Roquin C-terminal region as an effector domain that regulates mRNA target expression through the recruitment of the CCR4-NOT complex despite the lack of sequence conservation.

ResultsRoquin C-terminal regions recruit the CCR4-NOT complex. Hs Roquin1 interacts with the CCR4-NOT deadenylase complex through a C-terminal region that shows very low sequence similarity to the C-terminal regions of the corresponding vertebrate Roquin2 paralogues and invertebrate Roquin10 (Fig. 1a). We therefore asked whether the C-terminal regions of Hs Roquin2 and Dm Roquin have the ability to interact with the CCR4-NOT complex. Hs Roquin2 and Dm Roquin expressed with a tag consisting of the V5 epitope followed by the streptavidin-binding peptide (V5-SBP) pulled down the endogenous CCR4-NOT complex in human HEK293T cells to a similar extent as Hs Roquin1 (Fig. 1b). Moreover, the C-terminal regions of Hs Roquin2 and Dm Roquin were necessary and sufcient for the interaction, as observed for Hs Roquin1 (Fig. 1c,d; Supplementary Fig. 1a)10. The observation that Dm Roquin interacts with the CCR4-NOT complex in human cells further suggests that the protein recognizes surfaces on the CCR4-NOT complex that are conserved across species.

Roquin C-terminal regions mediate mRNA degradation. We next investigated whether the C-terminal regions of Hs Roquin2 and Dm Roquin elicit the degradation of mRNA targets. To this end, we used an MS2-based tethering assay in human HEK293T cells. The full-length proteins and the corresponding N- and C-terminal fragments (Hs Roq2-N, Hs Roq2-C, Dm Roq-N and Dm Roq-C, respectively; Fig. 1a, Supplementary Table 1) were expressed with an MS2-HA tag that mediates binding to a b-globin reporter mRNA containing six MS2-binding sites in its 30-UTR (b-globin-6xMS2bs)33. Tethered Hs Roquin2 and Dm

2 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307 ARTICLE

Roquin reduced the level of the b-globin-6xMS2bs mRNA relative to the MS2-HA fusion protein, which was used as a negative control (Fig. 1ej). Furthermore, the C-terminal fragments retained the mRNA degradation activity of the full-length proteins, whereas the N-terminal fragments were inactive (Fig. 1ej). Similar results were obtained for Hs Roquin1 (Supplementary Fig. 1bd). The N- and C-terminal fragments were expressed at levels comparable to those of the full-length

proteins (Fig. 1g,j), and none of the proteins affected the expression of the control b-globin mRNA lacking MS2-binding sites (Fig. 1f,i; control). Thus we conclude that, despite the lack of sequence conservation, the C-terminal regions of Roquin proteins interact with the CCR4-NOT complex and promote the degradation of bound mRNAs.

As shown for Hs Roquin1 and Roquin2, Dm Roquin regulates the expression of a b-globin mRNA reporter containing the

75 324 686

324

128

117

326

399

413

440

501

533

788

Hs Roquin1

RING

1,133

N-HEPN C-HEPN

ROQ

ZnF

P-rich

Roq-N

Roq-C

125

174

396

411

437

501

575

Hs Roquin2

RING

1,191

N-HEPN C-HEPN

ROQ

ZnF

P-rich

Roq-N Roq-C

126

175

395

418

440

501

725

755

790

812

702

CN CC

819

Dm Roquin

RING NBM

Roq-C

Roq-N

N-HEPN

ROQ

C-HEPN

ZnF

Hs Roquin1

GFP-MBP

Hs Roquin2

Hs Roquin1

GFP-MBP

Dm Roquin

Hs Roquin2

GFP-MBP

Hs Roq2-N

Hs Roquin2

GFP-MBP

Hs Roq2-C

Hs Roq2-N

Dm Roquin

GFP-MBP

Dm Roq-N

Dm Roquin

GFP-MBP

Dm Roq-C

Dm Roq-N

Dm Roq-C

Hs cells

V5-SBP-

72 Anti-NOT2

130 Anti-NOT3

180

Anti-V5

130

95 72

Anti-V5

130

95 72

Anti-V5

130

Anti-NOT1

72 Anti-NOT2

130 Anti-NOT3

Anti-NOT1

180

Anti-NOT2

180

130

Anti-NOT3

2 8

Pulldown Pulldown

2 8

6 1 5

2 8

6 1 5

Input

Pulldown

Input

e -globin-6xMS2bs g Western blot

-6xMS2bs f

Relative -globin mRNA levels

h -globin-6xMS2bs

Relative -globin mRNA levels

100

0 MS2

Dm Roquin

Dm Roq-N

Hs cells Hs cells Hs cells

Hs cells

Hs Roquin2

Hs Roq2-C

Hs Roq2-N

Hs Roquin2

Hs Roq2-C

100

MS2-HA-

Control 6xMS2bs

Anti-HA

GFP-F-Luc

250 130

95 72

MS2

Hs Roquin2

0 Hs Roq2-N

Hs Roq2-C

Dm Roq-C

i Western blot

-globin-6xMS2bs

Control

6xMS2bs

-globin-TNF-

0 MS2

Hs Roquin1

-globin-TNF-

Hs cells Hs cells

Hs cells

Relative -globin mRNA levels

Dm Roq-C

Dm Roq-N

Dm Roquin

120

Dm Roquin

Hs Roquin1

Dm Roq-C

MS2-HA-

Dm Roq-C

Dm Roq-N

Dm Roquin

MS2-HA-

Dm Roq-N

130

95 72

Anti-HA

GFP-F-Luc

100

Control

TNF-

Hs cells

Dm Roquin

Dm Roq-N

Dm Roq-C

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

constitutive decay element (CDE) from the TNF-a mRNA in its 30-UTR (CDE-37; ref. 10) in human HEK293T cells (Fig. 1k,l), consistent with the conservation of the ROQ domain15,16,23. Regulation of the b-globin-TNF-a reporter by Dm Roquin was abolished by mutations in the CDE that disrupt the binding of the ROQ domain (Supplementary Fig. 1e,f; mutant MUT16; ref. 10). Furthermore, in contrast to the results obtained in the tethering assay, degradation of the b-globin-TNF-a reporter required both the N- and the C-terminal regions of Dm Roquin (Fig. 1k,l). Indeed, the C-terminal region alone was not sufcient to cause degradation of the b-globin-TNF-a reporter (Fig. 1k,l), most likely because it does not bind RNA.

Roquin proteins direct mRNAs to the 50-to-30 decay pathway. Given that Roquin proteins interact with the CCR4-NOT complex, we next investigated whether the proteins elicit degradation of mRNA targets via the 50-to-30 decay pathway, in which deadenylation is followed by decapping and 50-to-30 exonucleolytic degradation of the mRNA body. To this end, we performed tethering assays in HEK293T cells overexpressing a catalytically inactive DCP2 mutant (DCP2 E148Q), which inhibits decapping in a dominant-negative manner34. We observed that degradation of the b-globin-6xMS2bs mRNA by tethered Hs Roquin1, Hs Roquin2 and Dm Roquin was impaired in cells expressing the DCP2 mutant; in these cells, the reporter accumulated in a shorter deadenylated form (Fig. 2a,b, lanes 68). The expression of the tethered proteins was not affected by the expression of DCP2 mutant (Fig. 2c). Our results indicate that the three Roquin proteins direct mRNA targets to the 50-to-30 decay pathway.

Dm Roquin degrades bound mRNAs in Drosophila cells. To investigate whether the Dm Roquin protein has the ability to repress and degrade bound mRNAs in Dm cells, we used a lN-based tethering assay in Drosophila melanogaster Schneider

S2 cells35. Tethered Dm Roquin caused 10-fold repression of a rey luciferase (F-Luc) reporter containing ve binding sites for the lN-tag (BoxB hairpins) in its 30-UTR (Fig. 3ac). The reduction in F-Luc activity was accompanied by a corresponding decrease in mRNA abundance (Fig. 3ac) and a decrease in the half-life of the mRNA (Fig. 3d,e), indicating that Dm Roquin induces mRNA degradation in S2 cells. Dm Roquin did not affect the expression of an F-Luc reporter that lacked the BoxB hairpins (Supplementary Fig. 1g,h).

As observed in human cells, the Roq-C fragment retained the repressive activity of the full-length protein in the tethering assay and accelerated the degradation of the mRNA reporter (Fig. 3ae), whereas the Roq-N fragment was inactive (Fig. 3a,b).

Furthermore, the full-length Dm Roquin and the Roq-C fragment repressed the translation of a F-Luc mRNA reporter with a 30-end generated by a self-cleaving hammerhead ribozyme (F-Luc-5BoxB-A95C7-HhR; Fig. 3f,g). This reporter is resistant to deadenylation and is therefore not degraded in S2 cells36. Similar results have been reported for other RNA-associated proteins that recruit the CCR4-NOT complex8,9,36, although the involvement of other factors cannot be excluded.

To conrm that mRNA degradation caused by Dm Roquin is dependent on the CCR4-NOT complex, we depleted NOT1 in S2 cells. The ability of Dm Roquin to elicit the degradation of the F-Luc-5BoxB mRNA was partially suppressed in NOT1-depleted cells (Fig. 3h,i). Western blotting analysis indicated that the NOT1 levels were reduced to o25% of their control levels in depleted cells (Supplementary Fig. 1i). Thus Dm Roquin promotes mRNA degradation by recruiting the CCR4-NOT complex in Drosophila cells.

Dm Roquin interacts directly with CAF40 and the NOT module. To identify the subunits of the CCR4-NOT complex that interact with Dm Roquin, we expressed the GFP-tagged protein in S2 cells and determined whether it interacts with HA-tagged subunits of the CCR4-NOT complex using co-immunoprecipitation assays. We also tested for interactions of the protein with the PAN2-PAN3 deadenylase complex and with decapping factors. Dm Roquin interacted with NOT1, NOT2, NOT3, CAF40, NOT10, CAF1 and PAN3 (Supplementary Fig. 2aj). These interactions were observed in the presence of RNase A, suggesting that they are not mediated by RNA. Dm Roquin also interacted with the decapping factor HPat in an RNA-independent manner but not with other decapping factors (Supplementary Fig. 2ko). In particular, and in contrast to Hs Roquin1 (ref. 32), we observe no interaction of Dm Roquin with Dm EDC4 or Me31B, which is the Dm orthologue of DDX6/RCK (Supplementary Fig. 2n,o).

To discriminate between direct and indirect interactions with subunits of the CCR4-NOT complex, we performed pulldown assays in vitro using puried recombinant proteins expressed in Escherichia coli. Because Dm NOT1 is not expressed in a soluble form in bacteria, we expressed the human proteins, which interact with Dm Roquin (as shown in Fig. 1b) and tested them for interaction with Dm Roquin in vitro.

Initially, we used a puried human pentameric complex consisting of a NOT1 fragment comprising residues 10932376, CAF1, CAF40 and the C-terminal domains of NOT2 and NOT3 (Fig. 4a). The Dm Roq-C fragment carrying an N-terminal maltose-binding protein (MBP) tag pulled down the puried pentameric complex (Fig. 4b, lane 20), thus demonstrating its direct interaction with the complex.

Figure 1 | The C-terminal regions of Roquin proteins interact with the CCR4-NOTcomplex and induce degradation of bound mRNA. (a) Roquin proteins consist of a conserved N-terminal region containing a RING-nger E3 ubiquitin ligase domain, a ROQ RNA-binding domain anked by a bilobed HEPN domain and a CCCH-type zinc-nger (ZnF) domain. The N-terminal region is followed by a variable C-terminal extension (shown in grey) that often contains proline-rich sequences (P-rich). The positions of the CBM and the NBM in Dm Roquin are indicated. The numbers above the protein outline indicate the residues at domain/motif boundaries. (bd) SBP pulldown assays in HEK293Tcells expressing V5-SBP-tagged Hs Roquin1, Hs Roquin2 and Dm Roquin (full-length or N- and C-terminal fragments). A V5-SBP-tagged GFP-MBP fusion served as a negative control. The presence of endogenous NOT1, NOT2 and NOT3 in the bound fractions was analysed by western blotting using specic antibodies. The inputs (1.5% for the V5-SBP tagged proteins and 1% for NOT1, NOT2 and NOT3) and bound fractions (10% for the V5-SBP tagged proteins and 30% for NOT1, NOT2 and NOT3) were analysed by western blotting. (ej) Tethering assays using the b-globin-6xMS2bs reporter and MS2-HA-tagged Hs Roquin2 and Dm Roquin (full-length or the indicated fragments) in human HEK293Tcells. A plasmid expressing a b-globin mRNA reporter lacking MS2-binding sites (Control) served as a transfection control.

The b-globin-6xMS2bs mRNA level was normalized to that of the control mRNA and set to 100% in cells expressing MS2-HA. The mean valuess.d. from three independent experiments are shown in e,h. (f,i) show representative northern blottings. (g,j) show the equivalent expression of the MS2-HA-tagged proteins used in the corresponding tethering assays. (k,l) Effect of Dm Roquin on the expression of the b-globin-TNF-a mRNA reporter analysed as described in ej. Protein size markers (kDa) are shown on the right of the western blotting panels. Error bars represent s.d. from three independent experiments. Full images of western and northern blottings are shown in Supplementary Fig. 10.

4 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307 ARTICLE

To map the binding site more precisely, we tested interactions with individual CCR4-NOT subcomplexes, including the NOT1 MIF4G domain bound to CAF1, the NOT1 CN9BD bound to CAF40, a C-terminal connector domain of unknown function (CD) and the NOT module (comprising the NOT1 SHD and the C-terminal regions of NOT2 and NOT3). MBP-tagged Dm

Roq-C pulled down the CN9BD-CAF40 complex as well as the NOT module but not the NOT1 MIF4G-CAF1 complex or the CD (Fig. 4b, lanes 2124).

Further analysis indicated that Dm Roq-C interacted directly with both Hs and Dm CAF40 in the absence of the NOT1 CN9BD (Fig. 4c, lanes 11 and 12). We also investigated whether the interaction of Dm Roq-C with the NOT module was mediated by the NOT1 C-terminal SHD domain or by the NOT2-NOT3 dimer. However, splitting the NOT module resulted in severely reduced binding to both NOT1 and the NOT2-NOT3 dimer, demonstrating that only the assembled module is recognized efciently (Supplementary Fig. 3a). In summary, the Dm Roq-C fragment contains at least two distinct binding sites for the CCR4-NOT complex, one site that contacts CAF40 and a second site that contacts the NOT module.

Redundancy of CCR4-NOT-binding motifs in Dm Roq-C. To dene more precisely how Dm Roq-C interacts with CAF40 and the NOT module, we sought to identify conserved motifs within the primary sequence. Using only sequences from Drosophila species, it was possible to align the Dm Roq-C sequences across their entire length (Supplementary Fig. 4). The alignment revealed clusters of conserved residues dispersed throughout the sequence with a higher level of conservation evident at the C-terminal end of Roq-C (Supplementary Fig. 4). We therefore generated two fragments, which we termed Roq-CN and Roq-CC (Fig. 1a, Supplementary Fig. 4). Remarkably, each of these fragments in isolation exhibited repressive activity in tethering assays, indicating functional redundancy (Supplementary Fig. 3bh). However, only the Roq-CC fragment, comprising residues 702819, bound to the puried CAF40 protein in vitro as efciently as the entire Roq-C fragment, whereas binding of the Roq-CN fragment (residues 501702) was strongly impaired (Fig. 4d, lanes 23 and 20 versus lane 17). In contrast, both Roq-CN and Roq-CC retained binding activity for the NOT module, although their binding was reduced compared with that of the full Roq-C fragment (Fig. 4d, lanes 21 and 24 versus lane 18). These results indicate the presence of multiple binding sites for the NOT module within Roq-C.

Through a deletion analysis combined with binding, we then identied a motif comprising residues 790812 within Roq-CC that is necessary and sufcient for binding to the CAF40 armadillo repeat (ARM) domain (Figs 1a and 4e). Indeed, deletion of the CBM in Roq-CC abolished its binding to CAF40 (Figs 4e, lane 14). Conversely, the CBM is sufcient for binding to CAF40 (Fig. 4e, lane 16 versus lane 12). We also identied

Figure 2 | Roquin proteins degrade mRNAs through the 50-to-30 mRNA decay pathway. (a,b) A tethering assay was performed in HEK293T cells using the b-globin-6xMS2bs reporter as described in Fig. 1ej, except that different amounts of plasmid were transfected (see Methods section). The transfection mixture included plasmids expressing either GFP or a GFP-tagged catalytically inactive DCP2 mutant (E148Q). The b-globin-6xMS2bs mRNA levels were normalized to those of the control mRNA and set to 100% in the presence of MS2-HA for each condition. The mean valuess.d. from three independent experiments are shown in a. The white and black bars represent the normalized b-globin-6xMS2bs mRNA values in cells expressing GFP and the GFP-DCP2 mutant, respectively.

(b) shows a representative northern blotting. The positions of the polyadenylated (An) and deadenylated (A0, dashed red line) mRNA reporter are indicated on the right. (c) Western blotting analysis showing equivalent expression of MS2-HA-tagged proteins in cells expressing GFP or the DCP2 mutant. Error bars represent s.d. from three independent experiments. Full images of western and northern blottings are shown in Supplementary Fig. 11.

-globin-6xMS2bs Hs cells

100

-globin-6xMS2bs

Hs cells

MS2-HA-

GFPGFP-DCP2 E148Q

Relative -globin mRNA levels

Hs Roquin2

0 Dm Hs Roquin1

MS2 Roquin

DmRoquin

HsRoquin1

HsRoquin2

DmRoquin

HsRoquin1

HsRoquin2

Control

6xMS2bs

1 2 3 4

5 6 7 8

GFP GFP-DCP2 E148Q

Western blot Hs cells

Anti-HA

Anti-GFP

MS2-HA-

Hs Roquin2

Hs Roquin1

Dm Roquin

Hs Roquin2

Hs Roquin1

Dm Roquin

Anti-tubulin

1 2 3 4 5 6

GFP GFP-DCP2 E148Q

250

130

95 72

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

residues 725755 as being required for binding to the NOT module in the context of Roq-CC (this region was termed the NOT-module binding motif (NBM); Fig. 1a, Supplementary

Fig. 5a, lane 14); however, in isolation the NBM was not sufcient for binding, indicating a more complicated binding mode (Supplementary Fig. 5a, lane 16).

a b c

F-Luc-5BoxB

R-Luc

Western blot

Dm cells

N-HA-

Dm cells

140 Protein mRNA

Relative F-Luc / R-Luc levels

120

DmRoquin

DmRoq-N

DmRoq-C

DmRoquin

DmRoq-N

DmRoq-C

N-HA-

100

130 95 72 55

F-Luc

Anti-HA

R-Luc-V5

2 3

Dm Roquin

Dm Roq-N

Dm Roq-C

1 2 3

d e

F-Luc-5BoxB

Dm cells Dm cells

Dm cells Dm cells Dm cells Dm cells

N t1/2 146 45 min

Dm Roquin t1/2 18 1 min

0.2 min

240

Dm Roquin Dm Roq-C

18 1 min 9.7 0.2 min

100

F-Luc / rp49mRNA levels

t (min)

120

240

120

240

120

240

F-Luc

rp49

t1/2

146 45 min

0 0 200

Time (min)

120

160

f g h i

F-Luc-5BoxB-A95-C7-Hhr

Dm Roq-C

F-Luc-5BoxB

F-Luc

R-Luc

140

Relative F-Luc / R-Luc levels

F-Luc-5BoxB-A95-C7-Hhr

N-HA-

Protein mRNA

120

N-HA-

Dm Roquin

100

DmRoquin

DmRoq-N

DmRoq-C

F-Luc

R-Luc

Relative F-Luc / R-Luc levels

100

GST KD NOT1 KD

1 2 3 4

2 3 4

Dm Roq-N

0 Dm Roquin

GST

NOT1 KD

Dm Roquin

0 N

Dm Roquin

Figure 3 | Dm Roquin degrades bound mRNAs. (a,b) Results of tethering assays using the F-Luc-5BoxB reporter and lN-HA-tagged Dm Roquin (full-length or the indicated fragments) in Dm S2 cells. A plasmid expressing R-Luc served as a transfection control. F-Luc activity and mRNA levels were normalized to those of the R-Luc transfection control and set to 100% in cells expressing the lN-HA peptide. The mean valuess.d. from three independent experiments are shown in a. (b) shows a representative northern blotting. The corresponding control experiment with a F-Luc reporter lacking the BoxB sites is shown in Supplementary Fig. 1g,h. (c) Western blotting showing the equivalent expression of the lN-HA-tagged proteins used in a,b.

(d,e) Representative northern blotting showing the decay of the F-Luc-5BoxB mRNA in S2 cells expressing lN-HA or lN-HA-tagged Dm Roquin or the Roq-C fragment after inhibition of transcription by actinomycin D. F-Luc mRNA levels were normalized to those of the rp49 mRNA and plotted against time.

The mRNA half-life (t1/2)s.d. was calculated from the decay curve shown in e. (f,g) Results of the tethering assays using the F-Luc-5BoxB-A95-C7-HhR reporter and lN-HA-tagged Dm Roquin (full-length or the indicated fragments) in Dm S2 cells. The samples were analysed as described in a,b.

(h,i) Tethering assay using the F-Luc-5BoxB reporter and lN-HA-tagged Dm Roquin in Dm S2 cells depleted of NOT1 or in control cells (treated with a dsRNA targeting GST). The samples were analysed as described in a,b. The efcacy of the depletion is shown in Supplementary Fig. 1i. Error bars represent s.d. from three independent experiments. Full images of western and northern blottings are shown in Supplementary Fig. 12.

6 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307 ARTICLE

CAF1 (CNOT7) CAF40 (CNOT9)

NOT2 NOT3

N1SHD

Hs NOT1 HEAT HEAT MIF4G CN9BD

MBP MBP

Input MBP pulldown

1,610

262

1,000

685 840 1,093

1,317

1,356

1,588

1,813

1,847

2,361

2,376

CAF40 module

NOT module

Roq-C Roq-C

Pentameric

MIF4G-CAF1

CN9BD-CAF40

NOT1-CD

NOT1-2-3

Pentameric

MIF4G-CAF1

CN9BD-CAF40

NOT1-CD

NOT1-2-3

Pentameric

MIF4G-CAF1

CN9BD-CAF40

NOT1-CD

NOT1-2-3

Pentameric

MIF4G-CAF1

CN9BD-CAF40

NOT1-CD

NOT1-2-3

175

NOT1-1093-C

MBP-Roq-C

NOT1-SHD

MBP

CAF40-ARM

CAF1

NOT1-CN9BD

NOT1-MIF4G

NOT1-CD

NOT2-C

NOT3-C

1 2 3 4 5 6 7 12

13 14 15 16

17 18 19 24

MBP

HsCAF40

Roq-C

MBP Roq-C

MBP CBM CBM

Roq-CN Roq-CC

Roq-CC MBP CBM CBM

Roq-CC

DmCAF40

HsCAF40

DmCAF40

HsCAF40

DmCAF40

HsCAF40

DmCAF40

CAF40

56 43 35

30 MBP

MBP-Roq-CC

MBP-Roq-C

MBP-CBM

CAF40-ARM

MBP

Hs/Dm CAF40

1 2 3 4 5 6 7 12

8 1 2 3 4 5 6 7 8

9 13 14 15 16

Input MBP pulldown Input MBP pulldown

MBP Roq-C MBP Roq-C

Roq-CN Roq-CC

CAF40

NOT1-23

CAF40

NOT1-23

CAF40

NOT1-23

CAF40

NOT1-23

CAF40

NOT1-23

CAF40

NOT1-23

CAF40

NOT1-23

CAF40

NOT1-23

MBP-Roquin

NOT1-SHD

MBP

CAF40-ARM

NOT2-C

NOT3-C

Input MBP pulldown

1 2 3 4 5 6 7 12

8 13 14 15 16 17 18 19 24

Figure 4 | Dm Roq-C interacts directly with CAF40 and the NOT module. (a) Schematic representation of the Hs CCR4-NOT complex with the subunits used in this study. NOT1 contains two HEAT repeat domains (shown in blue and petrol), a MIF4G domain composed of HEAT repeats (shown in green), a three-helix bundle domain (CN9BD, yellow), a connector domain (CD, light blue) and a NOT1 superfamily homology domain (SHD, grey), which also consists of HEAT repeats. The positions of the other subunits indicate their binding sites on NOT1. (b) MBP pulldown assay showing the interaction of MBP-tagged Dm Roq-C with puried pentameric NOT1-2-3-7-9 complex, the indicated NOT1 domains and CCR4-NOT subcomplexes. MBP served as a negative control. The difference in the migration of the NOT3-C protein in different samples is due to the presence of a non-cleavable 6xHis-tag in the case of the pentameric NOT1-2-3-7-9 complex, whereas in the case of the NOT1-2-3 complex the HRV3C-cleavable 6xHis tag was removed during purication. (c) MBP pulldown assay showing the interaction of MBP-tagged Dm Roq-C with puried Hs and Dm CAF40 proteins. (d) MBP pulldown assay showing the interaction of MBP-tagged Dm Roquin fragments (Roq-C, Roq-CN and Roq-CC) with the Hs CAF40 proteins and the recombinant Hs NOTmodule (NOT1-2-3) containing the NOT1 SHD and the NOT2 and NOT3 C-terminal fragments. (e) MBP pulldown assay showing the interaction of MBP-tagged Dm Roquin fragments (Roq-CC, Roq-CC-DCBM or CBM alone) with the Hs CAF40-ARM domain. Protein size markers (kDa) are shown on the left in each panel. Full images of protein gels are shown in Supplementary Fig. 13.

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

We next tested whether the interaction with CAF40 is also observed for the human Roquin proteins. We observed that Hs CAF40 interacts directly with the C-terminal region of

Hs Roquin1 (Supplementary Fig. 5b); however, more detailed mapping to identify a single CBM was unsuccessful, thus suggesting the presence of multiple binding sites.

Dm cells Dm cells Dm cells

Dm cells

CBM+NBM

NBM

CBM

Dm Roquin

F-Luc

CBM+NBM

NBM

CBM

Dm Roquin

F-Luc

Dm Roquin

F-Luc

CBM

CBM+NBM

NBM

Dm Roquin

F-Luc

CBM+NBM

NBM

CBM

CBM+NBM

NBM

CBM

GFP-

Dm Roquin

F-Luc

Dm Roquin

F-Luc

CBM+NBM

NBM

CBM

GFP-

130

Anti-GFP

130

Anti-GFP

HA-NOT2

130 Anti-GFP

HA-NOT3

HA-CAF40 HA-CN9BD

2 3 4

6 7 8 9 10 3 4 5 6 7 8 9 10 3 4 5 6 7 8 9 10

2 3 4

1 5 6 7 8 9 10

1 2

Input

IP(GFP)

Input IP(GFP)

Input

IP(GFP)

F-Luc-5BoxB

Dm cells

Dm Roquin

F-Luc-GFP

CBM+NBM

NBM

CBM

CBM+NBM

NBM

CBM

Dm Roquin

F-Luc-GFP

Hs cells

C-CBM+NBM

C-NBM

C-CBM

Dm Roq-C

V5-SBP-

Protein

Relative F-Luc / R-Luc levels

100

130

mRNA

N-HA-

Anti-V5

F-Luc

CAF40-HA-MBP

CN9BD-HA

R-Luc

2 3 4

1 5

C-CBM

Dm Roq-C

0 N

C-CBM+NBM

C-NBM

Input Pulldown

-globin-6xMS2bs

MBP-F-Luc-GFP

Hs cells

Anti-NOT1

CBM+NBM

NBM

CBM

Roquin MBP-F-Luc-GFP

Hs cells Hs cells

V5-SBP-

Dm Roquin

CBM

CBM+NBM

NBM

CBM+NBM

NBM

CBM

Dm Roquin

CBM+NBM

NBM

CBM

Dm Roquin

100

MS2-HA-

130

Relative -globin mRNA levels

Control

6xMS2bs

Anti-V5

1 2 3 4 5

Anti-NOT2

Anti-NOT3

180

130

CBM Dm Roquin

0 MS2

CBM+NBM

NBM

3 4 5 6 7 8 9 10

Input

Pulldown

j Western blot

-globin-TNF-

100

-globin-TNF-

Hs cells Hs cells Hs cells

Relative -globin mRNA levels

Dm Roquin

Hs Roquin1

MS2-HA-

CBM

NBM

CBM+NBM

MS2-HA-

250

Control

TNF

130

Anti-HA

GFP-F-Luc

2 3 4 5

NBM

CBM Dm Roquin

0 MS2

Hs Roquin1

CBM+NBM

1 2 3 4 5

8 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307 ARTICLE

The CBM contributes to the activity of Dm Roquin. To assess the relative contribution of the CBM and the NBM to the binding of Dm Roquin to the CCR4-NOT complex and to the repressive activity of Dm Roquin, we deleted these motifs from the protein and performed co-immunoprecipitations and tethering assays in Dm S2 cells and human HEK293T cells. Deletion of the CBM abolished the interaction of full-length Dm Roquin with the Dm CAF40-CN9BD module in S2 cells (Fig. 5a, lane 8). However, this deletion did not affect the binding of Dm Roquin to NOT2 or NOT3 (Fig. 5b,c, lane 8). Deletion of the NBM had no effect on CAF40, NOT2 or NOT3 binding (Fig. 5ac, lane 9), consistent with the observation that Dm Roquin harbours multiple binding

sites for the NOT module. Accordingly, deletion of the CBM and NBM in the context of Roq-C reduced but did not abolish the activity of this fragment in tethering assays (Fig. 5d,e, Supplementary Fig. 5ce).

Similarly, in human cells, deletion of the CBM abolished the interaction of full-length Dm Roquin with the CAF40-CN9BD module (Fig. 5f, lane 8), indicating that the CBM represents the only binding site for CAF40 in Dm Roquin. Importantly, deletion of the CBM also abolished the interaction of full-length Dm Roquin with the endogenous CCR4-NOT complex in human cells (Fig. 5g, lane 8). In agreement with these results, deletion of the CBM reduced the ability of Dm Roquin to degrade b-globin-6xMS2bs and b-globin-TNF-a mRNAs (Fig. 5hl). Furthermore, depletion of CAF40 partially suppressed the activity of Hs Roquin1, Hs Roquin2 and Dm Roquin in tethering assays (Supplementary Fig. 6ac), indicating that CAF40 is indeed an important recruitment factor, but other, redundant interactions compensate for the lack of CAF40 in Dm and human cells.

In summary, the interaction of the CBM with CAF40 contributes to the recruitment of the CCR4-NOT complex by Dm Roquin. However, the observation that deletion of the NBM and CBM have little effect on the activity of full-length Dm Roquin in S2 cells probably indicates that Dm Roquin establishes additional uncharacterized interactions with the CCR4-NOT complex or additional binding partners to regulate mRNA targets.

Crystal structure of the Dm Roquin CBM bound to CAF40. To elucidate the molecular principles underlying the recruitment of CAF40 by Dm Roquin, we sought to determine the crystal structure of the CBM peptide bound to the Dm and Hs CAF40 ARM domains, which exhibit 81% identity (Supplementary Fig. 7). However, only the complex containing Hs CAF40 yielded well-diffracting crystals, from which we obtained a structure at 2.15 resolution (Table 1). Two copies of the complex, which are structurally highly similar, are present in the crystal asymmetric unit (Supplementary Fig. 8a,b; root-mean-square deviation (r.m.s.d.) 0.24 over 254 Ca atoms).

As previously described, the CAF40 ARM domain consists of 17 a-helices arranged into six armadillo (ARM) repeats. These repeats adopt the typical crescent-like shape of ARM domains (Fig. 6ac)2,3,37, with the concave surface that accommodates the CBM peptide (residues 790810). The CBM peptide folds into an amphipathic helix (residues 795810) that runs centrally across the concave surface of CAF40 and binds to a conserved hydrophobic patch close to the previously postulated nucleic acid-binding groove (Fig. 6df)37.

Table 1 | Data collection and renement statistics.

CAF40DmCBM

Data collection

Space group P21 No. of reections 34,975

Cell dimensionsa, b, c () 56.9, 103.7, 60.8 a, b, g () 90.0, 113.0, 90.0

Wavelength () 1.00001 Resolution () 46.82.15 (2.202.15)* Rsym 0.048 (0.52)

I/sI 13.6 (2.2) Completeness (%) 98.7 (98.8)

Redundancy 2.9 (2.7)

RenementResolution () 46.72.15 No. of reections 34,964 Rwork/Rfree 18.4%/22.6%

No. of atoms 4,905

Protein 4,678 Water 191 Other solvent molecules 36 B-factors (2) 60.0

Protein 60.0 Water 52.4 Other solvent molecules 102.5 Ramachandran plot

Favoured regions (%) 98.8 Disallowed regions (%) 0.0 r.m.s. deviations

Bond lengths () 0.004 Bond angles () 0.597

*Values in parentheses are for the highest-resolution shell.

Figure 5 | The CBM contributes to the mRNA degradation activity of Dm Roquin. (ac) Immunoprecipitation assays showing the interaction of GFP-tagged Dm Roquin (wild-type or the indicated deletion mutants) with HA-tagged CAF40, NOT2 and NOT3 in Dm S2 cells. In (a) the interaction was tested in the presence of HA-tagged CN9BD. GFP-tagged rey luciferase (F-Luc) served as a negative control. Input and immunoprecipitates were analysed using anti-GFP and anti-HA antibodies. For the GFP-tagged proteins, 3% of the input and 10% of the immunoprecipitates were loaded. For the HA-tagged proteins, 1% of the input and 30% of the immunoprecipitates were analysed. In all panels, the cell lysates were treated with RNase A prior to immunoprecipitation. (d,e) Tethering assay using the F-Luc-5BoxB reporter and lN-HA-tagged Dm Roq-C or the indicated deletion mutants in Dm S2 cells.

The samples were analysed as described in Fig. 3a,b. The corresponding control experiment with a F-Luc reporter lacking the BoxB sites and a western blotting showing the equivalent expression of the tethered proteins are shown in Supplementary Fig. 5ce. (f,g) Interaction of V5-SBP-tagged Dm Roquin (full-length or the indicated deletion mutants) with HA-tagged CAF40 (in the presence of the HA-tagged CN9BD) and with endogenous NOT1, NOT2 and NOT3 in HEK293T cells. A V5-SBP-tagged MBP-F-Luc-GFP fusion served as a negative control. The inputs (0.75% for V5-SBP-tagged proteins and 1% for NOT1, 2, 3) and bound fractions (5% for SBP-V5-tagged proteins and 30% for NOT1, NOT2 and NOT3) were analysed by western blotting. (h,i) Tethering assay using the b-globin-6xMS2bs reporter and the indicated MS2-HA-tagged proteins in HEK293T cells. The samples were analysed as described in

Fig. 1ej. (j,k) The effect of full-length Dm Roquin or the indicated deletion mutants on the expression of the b-globin-TNF-a mRNA reporter was analysed as described in Fig. 1k,l. (l) Western blotting analysis showing comparable expression of the MS2-HA-tagged proteins used in hk. The asterisk indicates cross-reactivity with the anti-HA antibody. Error bars represent s.d. from three independent experiments. Full images of western and northern blottings are shown in Supplementary Fig. 14.

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

17 N

CBM

CBM CAF40

NOT1

5 8

8 9

W W

CAF40

d Surface conservation

CBM

Identical residues between Hs and Dm

Not identical

e Electrostatic potential

CBM

Hydropathic index

CBM

CAF40 CAF40

CAF40

Nucleic acid binding groove

Low

High +5 kT/e

V809

L805

CBM

Y134

N804

L805

T180

L801

M798 I793

M798

I793

L801

F184

N88

I808

V181

N88

Y134

T180

G791

G141

5 8

CAF40

P131 A84

T138

L137

Q98

G141

A145

L137

T138

Figure 6 | Structure of the Dm Roquin CBM bound to CAF40. (a) The Dm CBM peptide (green; backbone shown in ribbon representation) bound to Hs CAF40 (light brown). The helices of CAF40 are depicted as tubes and are numbered in black. The orange semicircle marks the predominantly hydrophobic interface between the CBM peptide and CAF40. (b) Cartoon representation of the Dm CBM peptide bound to Hs CAF40. Secondary structure elements of CAF40 are labelled in black. (c) Structural model of the CBM peptide bound to the CAF40 module (consisting of CAF40 bound to the NOT1 CN9BD domain). The model was obtained by superimposing the structure shown in (b) with the structure of CAF40 bound to the NOT1 CN9BD domain and free tryptophan (PDB 4CRU)2. (d) Conservation of the CBM-binding surface on CAF40. CAF40 is shown in surface representation. Surface residues that are identical between Hs and Dm are shown in light brown; all other residues are shown in white. (e) Surface representation of CAF40 with residues coloured in a gradient from white to yellow with increasing hydophobicity57. The CBM is shown in green. (f) Surface representation of CAF40 with residues coloured in a gradient from red over white to blue according to the electrostatic potential from 5 to 5 kTe 1. The electrostatic potential was calculated

using the APBS tools plugin within PyMOL (http://www.pymol.org

Web End =http://www.pymol.org). The proposed nucleic acid-binding groove37 is indicated by an orange circle. (g,h) Close-up views of the CAF40CBM-binding interface in two orientations. Selected residues of CAF40 and Roquin are shown as orange and green sticks, respectively. Hydrogen bonds are indicated by red dashed lines. The residues that were mutated in this study are underlined.

Superposition of the CAF40-CBM complex with the structure of the CAF40 dimer37 (r.m.s.d. of 0.87 over 260 Ca atoms;

PDB 2FV2) or with that of CAF40 bound to the NOT1 CN9BD (r.m.s.d. of 0.59 over 254 Ca atoms, PDB 4CRV)2 shows that binding of the CBM peptide does not induce major conformational changes in the CAF40 ARM domain (Supplementary Fig. 8c,d). Importantly, binding of the CBM does not interfere with NOT1 binding (Fig. 6c), suggesting that Dm Roquin can interact with CAF40 in the context of the CCR4-NOT complex. Finally, binding of the CBM does not block access to the tryptophan-binding pockets on the convex surface of CAF40 that serve as binding sites for the GW182/TNRC6 proteins involved in miRNA-mediated gene silencing (Fig. 6c)2,3.

The amphipathic helix of the CBM peptide lies almost parallel to helices a5, a8 and a11 and it uses residues M798, L801, L805,

I808 and V809 on its hydrophobic side to interact with CAF40 residues Y134, L137, G141, G144 and A145 (helix a8) and L177,

T180, V181 and F184 (helix a11; Fig. 6g,h and Supplementary

Fig. 8e,f). Furthermore, I793 anchors the N-terminal extension of the CBM helix between CAF40 residues A84, R130, P131 and Y134, resulting in a total buried surface of 1903 2 that does not include any water molecules (Fig. 6h). Finally, the CBM peptide is xed by two hydrogen bonds between N804 and the CAF40 residue Q98 (helix a5) and by a hydrogen bond from CAF40 N88 (helix a5) to the carbonyl oxygen of G791 (Fig. 6g). The CAF40 residues R130 and K148 may have additional roles in anchoring the CBM peptide, but they have distinct orientations in the two copies of the complex.

CCR4-NOT is recruited via the concave surface of CAF40. To validate the interfaces determined from the crystal structure, we introduced mutations in CBM and CAF40 and tested them in MBP pulldown assays in vitro. Substitution of Dm Roquin interface residues L805 and V809 by glutamic acid (mutant M2) and the further introduction of I793E and L801E substitutions to create a quadruple mutant (mutant M4) abolished the interaction

10 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307 ARTICLE

a c

MBP M2

CAF40

Roq-CC M4 MBP M2

Roq-CC M4

Dm cells

CAF40

F-Luc

Dm Roquin

F-Luc

GFP-

MBP-Roq-CC

1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12

13 14 15 16

Anti-GFP

130

MBP

CAF40-ARM

HA-CAF40 HA-CN9BD

Input MBP pulldown

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Input MBP pulldown

Input

IP (GFP)

b d

Y139D+G146W

V186E

MBP

Roq-CC MBP Roq-CC Dm cells

Y139D+G146W

V186E

DmRoquin

Y134D+G141W

V181E

Y134D+G141W

V181E

Y134D+G141W

V181E

Y134D+G141W

V181E

CAF40

F-Luc

DmRoquin

F-Luc

GFP-

MBP-Roq-CC

Anti-GFP

130

MBP

HA-CAF40 HA-CN9BD

CAF40-ARM

1 2 3 4 5 6 7 8 9 10 11 12

Input IP (GFP)

e f g

F-Luc-5BoxB

Y139D+G146W

Dm cells

N-HA-

1 2 3 4 5 6 7

Dm cells

V186E

CAF40 wt

F-Luc

Y139D+G146W

V186E

CAF40 wt

F-Luc

140

Relative F-Luc / R-Luc levels

GFP-

Dm Roq-C

Dm Roquin M4

Dm Roquin

Dm Roq-C M4

Dm GST- Roq-CC

Dm GST- Roq-CC M4

120

100

F-Luc

Anti-GFP

HA-CN9BD 36 27

R-Luc

1 2 3 4 5 6 7 8

IP (GFP)

Dm Roq-C M4

Dm GST- Roq-CC

Dm GST- Roq-CC M4

Input

Dm Roq-C Dm Roquin M4

0 Dm Roquin

Figure 7 | The CBM is the only CAF40-binding site in Dm Roquin. (a) In vitro MBP pulldown assay showing the interaction of MBP-tagged Dm Roq-CC and the indicated mutants (M2 and M4; see Supplementary Table 1) with the puried Hs CAF40-ARM domain. MBP served as a negative control.(b) In vitro MBP pulldown assay showing the interaction of MBP-tagged Dm Roq-CC with the Hs CAF40-ARM domain (wild-type or the indicated mutants). (c) Interaction of GFP-tagged Dm Roquin (wild-type or the M1, M2, M3 and M4 mutants; see Supplementary Table 1) with HA-tagged CAF40 in the presence of the HA-tagged CN9BD in Dm S2 cells. F-Luc-GFP served as a negative control. (d) Interaction of GFP-tagged Dm Roquin wild-type with HA-tagged CAF40 (wild-type or the indicated mutants) in the presence of the HA-tagged CN9BD in Dm S2 cells. (e) Interaction of GFP-tagged Dm CAF40 (wild-type or mutants) with HA-tagged CN9BD in Dm S2 cells. (f,g) A tethering assay using the F-Luc-5BoxB reporter and lN-HA-tagged Dm Roquin,

Roq-C and GST-Roq-CC (full-length or mutant M4) was performed in Dm S2 cells as described in Fig. 3a,b. (f) shows mean valuess.d. for normalized F-Luc mRNA levels from three independent experiments. (g) shows a northern blotting of representative RNA samples corresponding to the experiment shown in f. Error bars represent s.d. from three independent experiments. Full images of western and northern blottings are shown in Supplementary Fig. 15.

of the MBP-tagged Roq-CC with puried Hs CAF40 in vitro (Fig. 7a, Supplementary Table 1). Conversely, a single V181E substitution in Hs CAF40 or substitution of residues Y134 and G141 by aspartic acid and tryptophan, respectively, abolished the interaction of Hs CAF40 with MBP-Roq-CC (Fig. 7b,

Supplementary Table 1). In contrast, mutation of the concave surface of CAF40 was not sufcient to disrupt the interaction with Hs Roq1-C (Supplementary Fig. 5b). Thus, although Hs Roq1-C binds to CAF40 directly, it must contact additional and/or alternative CAF40-binding surfaces.

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

a b

Hs cells

MBP-F-Luc-GFP

Dm Roquin

MBP-F-Luc-GFP

Dm Roquin

MBP-F-Luc-GFP

Dm Roquin

MBP-F-Luc-GFP

130

180

130

V5-SBP-

Anti-V5

72 Anti-NOT2

Anti-NOT3

Anti-V5

CAF40-HA-MBP

CN9BD-HA

Anti-NOT1

1 3

Pulldown

130

4 5 6

Input

1 3 4 5 6

Input Pulldown

c d e

-globin-TNF-

MS2

0 Dm Roquin

-globin-TNF-

Western blot

Hs cells Hs cells Hs cells

Relative -globin mRNA levels

100

Dm Roquin

MS2-HA-

Dm Roquin

MS2-HA-

Anti-HA

GFP-F-Luc

TNF-

130 95 72

95 72

Control

Figure 8 | The CBM mediates the activity of Dm Roquin in human cells. (a) SBP pulldown assays in HEK293T cell lysates showing the interaction of V5-SBP-tagged Dm Roquin (full-length or the M4 mutant) with HA-tagged CAF40-MBP in the presence of CN9BD-HA. A V5-SBP-tagged MBP-F-Luc-GFP fusion served as a negative control. The presence of the HA-tagged proteins in the bound fractions was tested by western blotting using anti-HA antibodies. The samples were analysed by western blotting as described in Fig. 5f. (b) SBP pulldown assays in HEK293T cell lysates showing the interaction of V5-SBP-tagged Dm Roquin (full-length or the M4 mutant) with endogenous NOT1, NOT2 and NOT3. A V5-SBP-tagged MBP-F-Luc-GFP fusion served as a negative control. The samples were analysed by western blotting as described in Fig. 1bd. (c,d) The effectof Dm Roquin full-length or the M4 mutant on the expression of the b-globin-TNF-a mRNA reporter was analysed in HEK293T cells as described in Fig. 1k,l. Error bars represent s.d. from three independent experiments. (e) Western blotting analysis showing the equivalent expression of the proteins used in the tethering assays shown in c,d.

Full images of western and northern blottings are shown in Supplementary Fig. 16.

The residues in CAF40 that interact with the Dm Roquin CBM peptide are conserved in Hs and Dm (Fig. 6d, Supplementary Fig. 7), suggesting that the CBM interacts with Dm CAF40 via a similar binding mode. Nevertheless, it was important to test whether Dm Roquin binds to the concave surface of Dm CAF40. All tested mutations in the CBM (M1, M2, M3 and M4; Supplementary Table 1) were sufcient to disrupt the interaction of full-length Dm Roquin with Dm CAF40 in co-immunoprecipitation assays in Dm S2 cells (Fig. 7c, lanes 912), conrming that the CBM is the only motif in Dm Roquin that mediates binding to CAF40. Conversely, substitutions in Dm CAF40 corresponding to the mutations of Hs CAF40 shown in Fig. 7b abolished the interaction of

full-length Dm CAF40 with full-length Dm Roquin (Fig. 7d). The mutations in CAF40 did not abolish its binding to NOT1 CN9BD, indicating that they do not disrupt the CAF40 fold (Fig. 7e).

Notably, mutations in the CBM did not disrupt the interaction of full-length Dm Roquin with NOT2 or NOT3 in Dm S2 cells (Supplementary Fig. 9a,b), in agreement with the observation that Dm Roquin contains multiple sequences that mediate binding to the CCR4-NOT complex in a redundant manner. Consistent with this redundancy, mutations in the CBM abolished the activity of the Roq-CC fragment in tethering assays in Dm S2 cells but impaired the activity of full-length Dm Roquin only slightly (Fig. 7f,g, Supplementary Fig. 9ce). In contrast, when it was tested in human cells, mutations in the Dm Roquin CBM not only abolished interaction with CAF40 but strongly reduced the interaction with the endogenous CCR4-NOT complex (Fig. 8a,b) indicating that the CBM provides a major contribution to the interaction of Dm Roquin with the CCR4-NOT complex in human cells. Accordingly, the CBM mutants strongly reduced the ability of Dm Roquin to degrade the b-globin-TNF-a reporter containing the CDE-37 element in the 30-UTR in human HEK293T cells (Fig. 8ce).

DiscussionIn this study, we show that Roquin proteins (Hs Roquin1, Hs Roquin2, Dm Roquin) use their C-terminal extensions to directly recruit the CCR4-NOT complex to mRNA targets, promoting the degradation of these mRNAs. We show that this function is conserved among Roquin proteins despite the fact that the sequences of the unstructured C-terminal regions of these proteins are highly divergent and do not share similar motifs.

In Dm Roquin, the interaction with the CCR4-NOT complex is mediated by multiple and partially redundant motifs that include a CBM. We further elucidated the structural basis of the interaction of the CBM with CAF40 and identied the concave surface of CAF40 as a target for amphipathic helices to recruit the CCR4-NOT complex.

The nding that Roquin proteins use a combination of SLiMs (or eukaryotic linear motifs)38 to recruit the CCR4-NOT complex has important functional implications. First, although SLiMs mediate relatively low-afnity interactions, these interactions can be highly specic, and stable binding can be achieved through avidity effects generated by contributions from the anking disordered regions that extend the interaction interface38,39. In the case of Dm Roquin, the sequences anking the CBM provide binding sites for the NOT module, indicating that the C-terminal region of Dm Roquin contacts multiple subunits within the CCR4-NOT complex.

A second consequence of the nature of SLiMs is their evolutionary plasticity38,39. Owing to their short length and lack of sequence constraints in the absence of a protein fold, even single point mutations can render an existing motif nonfunctional or generate a new motif in another location of the same protein. In particular, the CBM is present in Drosophila species, but sequence analysis of Roquin proteins from other insects, worms and vertebrates does not reveal a detectable CBM. Nevertheless, Hs Roquin1 and Roquin2 also interact with the CCR4-NOT complex via their unstructured C-terminal regions (ref. 10 and this study), indicating that the overall principle of CCR4-NOT complex recruitment and mRNA target repression is evolutionarily conserved even though the interaction details have diverged across species.

In addition to the previously identied tryptophan-binding pockets on the CAF40 convex surface2,3, our crystal structure demonstrates that the concave surface of CAF40 is also used to

12 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307 ARTICLE

recruit the CCR4-NOT complex by RNA-associated proteins. The hydrophobic properties of the concave surface are highly conserved, making it an ideal partner for amphipathic helices or disordered, short hydrophobic peptides that are frequently present in RNA-binding proteins.

Therefore, the C-terminal regions of Hs Roquin1 and Roquin2 may also target this surface of CAF40, although no clear CBM motifs are detectable by sequence analysis. However, the proteins may contain cryptic CBMs that are likely discontinuous and probably also target neighbouring surfaces on CAF40 because mutations affecting the concave surface of Hs CAF40 did not disrupt its interaction with the Hs Roquin1 C-terminal fragment.

In addition to the Roquin proteins, a large number of RNA-associated proteins have been shown to recruit the CCR4-NOT complex to their mRNA targets, thereby repressing translation and/or promoting mRNA degradation. These include GW182 proteins, TTP, Nanos and the Dm proteins CUP, Bicaudal C, Smaug and Pumilio2,3,6,8,9,4044. For most of these proteins, it has been shown that interaction with the CCR4-NOT complex is mediated by SLiMs embedded in peptide regions of predicted disorder. However, detailed characterization of the binding mode on the structural level is only available for TTP, GW182, vertebrate and Dm Nanos and Dm Roquin (refs 2,3,6,8,9 and this study). Similar to Roquin, Dm Nanos, GW182 proteins and TTP all contain multiple binding sites for different subunits of the CCR4-NOT complex that act redundantly to recruit the complex to mRNA targets. This modular recruitment mode likely enhances the binding afnity and confers redundancy and robustness to the repression mechanism.

As observed for Roquin, the motifs in TTP and vertebrate and Dm Nanos, which have been structurally characterized, adopt an a-helical conformation that possibly forms only upon binding (refs 6,8,9 and this study). In contrast, GW182 peptides likely bind to the CCR4-NOT complex in an extended conformation and insert tryptophan residues into tandem hydrophobic pockets exposed at the convex surface of CAF40 and into additional pockets in NOT1 that remain to be identied2,3.

Because Dm Roquin and GW182 proteins can bind CAF40 simultaneously (Fig. 6c), it is possible that the proteins cooperate to recruit the CCR4-NOT complex to mRNAs. The presence of multiple CCR4-NOT-binding partners on an mRNA likely enhances the efciency of recruitment and the extent of the regulation.

In summary, together with previous studies2,3,6,8,9, our results indicate that SLiMs in unstructured and poorly conserved regions of RNA-associated proteins represent a common and widespread mode of recruitment of the CCR4-NOT complex to mRNA targets, resulting in a common downstream repressive mechanism that has a major role in posttranscriptional mRNA regulation in eukaryotic cells.

Methods

DNA constructs. Luciferase reporters and plasmids for the expression ofGFP- and HA-tagged subunits of the CCR4-NOT and PAN2-PAN3 deadenylase complexes and decapping factors in S2 cells were previously described35,45,46. Dm Roquin cDNA was purchased from the Drosophila Genomics Resource Center, amplied by PCR and inserted into the HindIII and NotI restriction sites of the pAc5.1-lN-HA or pAc5.1-GFP vectors. For expression in HEK293T cells, the cDNA encoding Dm Roquin was inserted into the NotI and ApaI restriction sites of the pcDNA3.1-MS2-HA vector33 and between the HindIII and KpnI sites of the pT7-V5-SBP-C1 vector47.

A cDNA sequence encoding Hs Roquin1 open-reading frame was amplied by PCR from human HEK293T cell total cDNA and inserted between the EcoRI and SacII sites of the pT7-V5-SBP-C1 vector and between the NotI and ApaI sites of the pcDNA3.1-MS2-HA vector. A cDNA encoding Hs Roquin2 open-reading frame was amplied by PCR from human HeLa cell total cDNA and inserted between the BamHI and NotI restriction sites of the pcDNA3.1-MS2-HA vector and between the SalI and BamHI restriction sites of the pT7-V5-SBP-C1 vector.

b-Globin reporters containing the wild-type or mutant (MUT16) TNF-a CDE were obtained by replacing the 6xMS2-binding sites in vector b-globin-6xMS2bs33 with the CDE-37 (TTGGCTCAGACATGTTTTCCGTGAAAACGGAGCTGAA) or CDE-37-MUT16 (TTGGCTCAGACATGTTTTCCGTGAAATGGGAGCT GAA) sequences10.

For expression of recombinant proteins in E.coli, cDNAs encoding Hs Roq1-C and Dm Roquin fragments were inserted between the AII and AvrII and the AII and XbaI restriction sites of the pnYC-pM plasmid48, respectively, resulting in Roquin fusion proteins carrying N-terminal MBP tags that are cleavable by HRV3C protease. The Roq-C and Roq-CC cDNAs and all of the constructs derived from them also contain a C-terminal GB1 tag49 followed by a MGSS linker and a 6xHis tag.

A cDNA encoding Dm CAF40 was inserted between the NdeI and XbaI restriction sites of the pnEA-pM plasmid48, generating a fusion protein containing an N-terminal MBP tag cleavable by the HRV3C protease.

For purication of the pentameric NOT1-2-3-7-9 complex, a cDNA encoding human NOT1 (residues 10932371) was inserted between the XhoI and BamHI restriction sites of the pnYC-pM vector, resulting in a fusion protein containing an N-terminal MBP-tag. A multicistronic plasmid was generated by inserting an expression cassette containing 6xHis-NOT3-C, MBP-NOT2-C, 6xHis-CAF40-ARM and GST-NOT7 (all tags except the CNOT3 6xHis tag are cleavable by HRV3C protease) into the pnEA vector.

A cDNA encoding NOT1-CD (residues 16071815) was inserted between the XhoI and BamHI restriction sites of the pnYC-pM plasmid48, generating a fusion protein containing an N-terminal MBP tag cleavable by HRV3C protease.

Plasmids for the expression of NOT1-MIF4G, NOT1-CN9BD, NOT1-SHD, NOT2-C, NOT3-C, NOT7 and CAF40-ARM have been previously described2,5,50. The DNA constructs used in this study are listed in Supplementary Table 1.

Co-immunoprecipitation and SBP-pulldown assays. For co-immunoprecipitation assays in S2 cells (ATCC), 2.5 106 cells were seeded per well in six-well plates

and transfected using Effectene transfection reagent (Qiagen). The transfection mixtures contained 1 mg of plasmid expressing HA-tagged deadenylase or decapping factors and 1.5 mg of GFP-tagged Roquin (either full length or fragments). S2 cells were harvested 3 days after transfection and co-immunoprecipitation assays were performed as previously described51.

For SBP pulldown assays in human cells, HEK293T cells (ATCC) were grown in 10-cm dishes and transfected using TurboFect transfection reagent (Thermo Fisher Scientic). The transfection mixtures contained 6, 30 and 2 mg of plasmids expressing Hs Roquin1, Hs Roquin2 and Dm Roquin, respectively. In the experiment shown in Fig. 1c, the transfection mixture contained 25 mg of plasmids expressing Hs Roquin2 and Roq2-N and 12 mg of a plasmid expressing Roq2-C. In the experiment shown in Fig. 1d, the transfection mixtures contained 15, 20 and 25 mg of plasmids expressing Dm Roquin, Roq-N and Roq-C, respectively. In the experiments shown in Fig. 5f,g, 10 mg of plasmids expressing Dm Roquin, DCBM and DNBM and 15 mg of plasmid expressing Dm Roquin DCBM NBM were

included. In the experiments shown in Fig. 8a,b, 10 mg of the Dm Roquin and Dm Roquin M4 plasmids was transfected. In the experiments shown in Figs 5f and 8a, the transfection mixtures contained 7.5 mg of a plasmid expressing

CAF40-HA-MBP and 5 mg of a plasmid expressing CN9BD-HA. In the experiments shown in Supplementary Fig. 1a, the transfection mixtures contained 1 mg of plasmids expressing Hs Roquin1 and Roq1-C and 10 mg of a plasmid expressing Roq1-N. Human cells were harvested 2 days after transfection, and co-immunoprecipitation assays were performed as previously described8. Western blottings were developed using the ECL Western Blotting Detection System(GE Healthcare) according to the manufacturers recommendations. The antibodies used in this study are listed in Supplementary Table 2.

Tethering assays in human and S2 cells and RNA interference. Tethering assays in human HEK293T cells using the b-globin reporter containing six MS2-binding sites (6xMS2bs)33 were performed as previously described8. Briey, cells were seeded in six-well plates (0.8 106 cells per well) and transfected using

Lipofectamine 2000 (Thermo Fisher Scientic).

The transfection mixtures contained 0.5 mg of the control plasmid (containing the b-globin gene fused to a fragment of the GAPDH gene but lacking MS2-binding sites)33, 0.5 mg of the b-globin-6xMS2bs reporter and varying amounts of pcDNA3.1-MS2-HA plasmids expressing MS2-HA-tagged proteins. The plasmid amounts were as follows: for Hs Roquin1, 0.6 mg full length and 0.5 mg each Roq1-N and Roq1-C; for Hs Roquin2, 1.5 mg full length, 0.6 mg Roq2-N and0.3 mg Roq2-C; for Dm Roquin, 0.3 mg full length, 0.1 mg each Roq-N and Roq-C,0.2 mg each Roq-DCBM and DNBM and 0.3 mg Roq-DCBM NBM. In the

experiment shown in Fig. 2a,b, 0.175 mg of Dm Roquin, 0.4 mg of Hs Roquin1 and 1 mg of Hs Roquin2 were transfected. The transfection mixtures also contained plasmids expressing GFP (0.2 mg) or a GFP-tagged catalytically inactive DCP2 mutant (1.5 mg), as indicated.

When the b-globin-TNF-a reporter was used, cells were co-transfected with0.5 mg b-globin-TNF-a reporter (CDE37 wild-type or MUT16; ref. 10) and plasmids expressing the HA-MS2 tagged proteins (Hs Roquin1: 0.5 mg full-length,0.4 mg each Roq1-N and Roq1-C; Dm Roquin: 0.2 mg full-length, 0.1 mg each Roq-N, Roq-C, DCBM and DNBM and 0.2 mg DCBM DNBM). In the

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 13

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

experiments shown in Fig. 8ce, the cells were transfected with 0.05 mg of Dm Roquin and Dm Roquin M4. Cells were harvested 2 days after transfection.

For the lN-tethering assay in Dm S2 cells, 2.5 106 cells per well were seeded

in six-well plates and transfected using Effectene (Qiagen). The transfection mixtures contained 0.1 mg rey luciferase reporter (F-Luc-5BoxB, F-Luc-V5 or

F-Luc-5BoxB-A95C7-HhR), 0.4 mg Renilla luciferase transfection control and various amounts of plasmids expressing lN-HA-tagged Dm Roquin constructs(0.01 mg full-length, 0.003 mg Roq-N, 0.025 mg Roq-C, 0.05 mg GST-Roq-CN, and0.02 mg GST-Roq-CC). In the experiment described in Fig. 7f,g and Supplementary Fig. 9ce, half of these amounts were transfected. In the experiment described in Supplementary Fig. 3g,h, the GST tags were replaced by GFP tags and 0.02 mg

GFP-Roq-CN, and 0.005 mg GFP-Roq-CC were transfected. Cells were harvested 3 days after transfection.

NOT1 knockdowns using dsRNA were performed as previously described35. To measure the mRNA half-life, cells were treated with actinomycin D (5 mg ml 1 nal concentration) 3 days after transfection and collected at the indicated time points. RNA samples were analysed by northern blotting. The level of reporter mRNA was normalized to the levels of endogenous rp49 mRNA in three independent experiments, averaged and plotted against time. The data were tted to a double exponential decay function prior to averaging. The reported half-lives (t1/2) correspond to 50% decay with respect to the initial amount of reporter RNA.

Half-life errors are calculated from the standard tting error.

Knockdowns in HeLa cells (provided by O. Mhlemann) were performed as described previously2. The 19 nt target sequences are as follows: control 50-ATT

CTCCGAACGTGTCACG-30, CAF40 50-GATCTATCAGTGGATCAAT-30. Cells were transfected in six-well plates using Lipofectamine 2000 according to the manufacturers protocol. Transfection mixtures contained 0.2 mg of Dm Roq-C;

0.3 mg of Hs Roq1-C and 0.4 mg of Hs Roq2-C.

Firey and Renilla luciferase activities were measured using a Dual-Luciferase

Reporter Assay system (Promega). Northern blotting was performed as previously described35.

Protein expression and purication. All recombinant proteins were expressed inE. coli BL21 (DE3) Star cells (Invitrogen) in LB medium at 20 C overnight. Dm Roquin fragments were expressed as fusion proteins containing N-terminal MBP tags cleavable by HRV3C protease48. In addition, the Dm Roq-C, -CN and -CC fragments carried an HRV3C-cleavable C-terminal GB1-6xHis tag49. The cells were lysed using an Avestin Emulsiex-C3 homogenizer in lysis buffer (50 mM HEPESNaOH (pH 7.5), 200 mM NaCl, 20 mM imidazole and 2 mM b-mercaptoethanol)

supplemented with complete EDTA-free protease inhibitors (Roche), 5 mg ml 1 DNaseI and 1 mg ml 1 lysozyme. The proteins were separated from the crude lysate using amylose resin (New England Biolabs) and subsequently eluted from the resin in lysis buffer containing 25 mM D-( )-maltose. The proteins containing

a GB1-6xHis tag were further puried by nickel afnity chromatography using a HiTrap IMAC column (GE Healthcare). Proteins without GB1-6xHis tags were further puried by anion exchange chromatography using a HiTrapQ column (GE Healthcare). The nal purication step for all proteins was size exclusion chromatography using a Superdex 200 16/600 column (GE Healthcare) in a buffer containing 10 mM HEPES-NaOH (pH 7.5), 200 mM NaCl and 2 mM DTT.

Hs CAF40 (residues 19285) was expressed with a 6xHis tag cleavable by the HRV3C protease. Lysis was carried out in lysis buffer containing 50 mM potassium phosphate (pH 7.5), 500 mM NaCl, 10% glycerol, 20 mM imidazole and 2 mM b-mercaptoethanol supplemented with complete EDTA-free protease inhibitors,

DNaseI and lysozyme. The Hs 6xHis-CAF40 ARM domain was isolated from the lysate using a HiTrap IMAC column (GE Healthcare). The 6xHis tag was removed by cleavage using HRV3C protease during dialysis in low salt buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl and 1 mM DTT. Subsequently, Hs CAF40 was further puried using a HiTrap Heparin column (GE Healthcare) followed by gel ltration on a Superdex 200 26/600 column (GE Healthcare) in gel ltration buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl and1 mM DTT.

Dm CAF40 (residues 25291) was expressed with an N-terminal MBP tag. The cells were lysed in a buffer containing 50 mM HEPES-NaOH (pH 7.5), 500 mM NaCl and 2 mM DTT supplemented with complete EDTA-free protease inhibitors, DNaseI and lysozyme. The protein was isolated from the crude lysate using amylose resin followed by anion exchange chromatography using a HiTrapQ column (GE Healthcare). The MBP tag was removed by cleavage with HRV3C protease during dialysis in a buffer containing 10 mM potassium phosphate(pH 8.0), 400 mM NaCl, 10% glycerol, 50 mM ammonium sulfate and 2 mM DTT. The nal purication step was size exclusion chromatography using a Superdex 200 26/600 column equilibrated with the same buffer.

The pentameric human NOT1-2-3-7-9 complex was obtained by co-expression of MBP-NOT1 (residues 10932371), MBP-NOT2-C, 6xHis-NOT3-C, GST-NOT7 and 6xHis-CAF40-ARM. The cells were lysed in buffer containing 50 mM potassium phosphate (pH 7.5), 300 mM NaCl and 2 mM DTT supplemented with complete EDTA-free protease inhibitors, DNaseI and lysozyme. The complex was puried using amylose resin, and the tags were removed by cleavage with HRV3C protease during dialysis in buffer containing 20 mM HEPES-NaOH (pH 7.5), 100 mM NaCl, 5% glycerol and 2 mM DTT. The complex was further puried using a HiTrap Heparin column (GE Healthcare) followed by size exclusion

chromatography on a Superdex 200 26/600 column equilibrated in 20 mM HEPES-NaOH (pH 7.5), 300 mM NaCl, 5% glycerol and 2 mM DTT.

To purify the complex containing 6xHis-NOT1-MIF4G (residues 10931317) bound to MBP-CAF1, cells coexpressing the proteins were lysed in a buffer containing 50 mM HEPES-NaOH (pH 7.5), 300 mM NaCl, 20 mM imidazole and 2 mM b-mercaptoethanol supplemented with complete EDTA-free protease inhibitors, DNaseI and lysozyme. The complex was isolated from the crude lysate using amylose resin followed by chromatography on a HiTrap IMAC column. The afnity tags were removed by cleavage using the HRV3C protease, followed by nal size exclusion chromatography on a Superdex 200 26/600 column in a buffer containing 10 mM HEPES-NaOH (pH 7.5), 200 mM NaCl and 2 mM DTT.

The NOT1 CD (residues 16071815) was expressed with an N-terminal MBP tag. The cells were lysed in a buffer containing 50 mM HEPES (pH 7.5), 300 mM NaCl and 2 mM DTT supplemented with complete EDTA-free protease inhibitors, DNaseI and lysozyme. The protein was isolated from the lysate using amylose resin followed by a HiTrapQ column. The MBP tag was removed by cleavage with HRV3C protease, followed by nal size exclusion chromatography on a Superdex 200 26/600 column using a buffer containing 10 mM HEPES-NaOH (pH 7.5), 200 mM NaCl and 2 mM DTT.

Purication of the Hs NOT module comprising NOT1-SHD (residues 18332361), NOT2-C (residues 350540) and NOT3-C (residues 607748) and the complex comprising NOT1-CN9BD (residues 13561588) bound to the CAF40-ARM has been previously described2,9.

The Dm Roquin CBM peptide (residues 790810) used for crystallization was synthesized by EMC Microcollections. The peptide was dissolved in 20 mM Tris-HCl (pH 7.5), 150 mM NaCl and 1 mM DTT.

Crystallization. Dm Roquin CBM was mixed with the puried Hs CAF40 ARM domain in an equimolar ratio. Initial crystallization screens were carried out using the sitting drop vapor diffusion method at 22 C by mixing 200 nl of the CAF40-Roquin complex solution (at 6.75 mg ml 1) with 200 nl of reservoir solution. Crystals appeared within 1 day under several conditions. Optimized crystals grew at 18 C after 1 day using hanging drops after mixing 0.8 ml of the

CAF40-Roquin complex solution (at 2.25 mg ml 1) with 0.8 ml of reservoir solution containing 0.1 M NaOAC (pH 5.0), 17.5% (w v 1) PEG 4000 and 0.1 M

AmSO4. Crystals were cryoprotected using reservoir solution supplemented with 15% glycerol and ash-frozen in liquid nitrogen.

Data collection and structure determination. Diffraction data were collected on a PILATUS 6M detector (Dectris) at the PXII beamline of the Swiss Light Source. The best data set extending to a resolution of 2.15 was recorded at a wavelength of 1.000 and processed in space group P21 using XDS and XSCALE52. Two copies of the CAF40 ARM domain (PDB-ID 2FV2, chain A) were found in the asymmetric unit by molecular replacement using PHASER53 from the CCP4 package54. This initial model was improved by iterative cycles of model building in COOT55 and renement using PHENIX56. As the nal step, two Roquin CBM peptides were modeled into the density and improved by further renement cycles.

The nal model was rened with excellent stereochemistry to Rwork 18.4%

and Rfree 22.6% and includes all residues of the two CAF40 molecules (residues

19285 plus six residues from the HRV3C cleavage site and linker sequences; chains A and C) and all residues (790810) of the Roquin peptides for both chains (B and D). For the C-terminal two pairs of a-helices in CAF40, the nal atomic

B-factors are clearly above average, pointing to an elevated mobility and/or statistical disorder in this part of the molecule, which is not involved in binding the CBM.

In vitro MBP-pulldown assays. Puried MBP (20 mg) or MBP-tagged Roquin fragments (40 mg) were incubated with equimolar amounts of puried CCR4-NOT subcomplexes and 50 ml of the amylose resin slurry (New England Biolabs) in 1 ml of pulldown buffer (50 mM HEPES-NaOH (pH 7.5), 200 mM NaCl, 2 mM DTT). After a 1-h incubation, the beads were washed ve times with pulldown buffer and the proteins were eluted with pulldown buffer supplemented with 25 mM

D-( )-maltose. The eluted proteins were precipitated with trichloroacetic acid

and analysed by SDS-PAGE followed by Coomassie blue staining.

Data availability. The coordinates for the structure of the Dm Roquin CBM peptide bound to CAF40 were deposited in the Protein Data Bank (PDB) under ID code 5LSW. The authors declare that the data supporting the ndings of this study and relevant source data are available within the article and its Supplementary Information le. Other data and materials are available from the authors upon request.

References

1. Wahle, E. & Winkler, G. S. RNA decay machines: deadenylation by the Ccr4Not and Pan2Pan3 complexes. Biochim. Biophys. Acta 1829, 561570 (2013).

2. Chen, Y. et al. A DDX6-CNOT1 complex and W-binding pockets in CNOT9 reveal direct links between miRNA target recognition and silencing. Mol. Cell 54, 737750 (2014).

14 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307 ARTICLE

3. Mathys, H. et al. Structural and biochemical insights to the role of the CCR4-NOT complex and DDX6 ATPase in MicroRNA repression. Mol. Cell 54, 751765 (2014).

4. Bhaskar, V. et al. Structure and RNA-binding properties of the Not1Not2 Not5 module of the yeast Ccr4Not complex. Nat. Struct. Mol. Biol. 20, 12811288 (2013).

5. Boland, A. et al. Structure and assembly of the NOT module of the human CCR4NOT complex. Nat. Struct. Mol. Biol. 20, 12891297 (2013).

6. Fabian, M. R. et al. Structural basis for the recruitment of the human CCR4NOT deadenylase complex by Tristetraprolin. Nat. Struct. Mol. Biol. 20, 735739 (2013).

7. Suzuki, A., Saba, R., Miyoshi, K., Morita, Y. & Saga, Y. Interaction between NANOS2 and the CCR4-NOT deadenylation complex is essential for male germ cell development in mouse. PLoS ONE 7, e33558 (2012).

8. Bhandari, D., Raisch, T., Weichenrieder, O., Jonas, S. & Izaurralde, E. Structural basis for the Nanos-mediated recruitment of the CCR4NOT complex and translational repression. Genes Dev. 28, 888901 (2014).

9. Raisch, T. et al. Distinct modes of recruitment of the CCR4-NOT complex by Drosophila and vertebrate Nanos. EMBO J. 35, 974990 (2016).

10. Leppek, K. et al. Roquin promotes constitutive mRNA decay via a conserved class of stem-loop recognition motifs. Cell 153, 869881 (2013).

11. Heissmeyer, V. & Vogel, K. U. Molecular control of Tfh-cell differentiation by Roquin family proteins. Immunol. Rev. 253, 273289 (2013).

12. Schaefer, J. S. & Klein, J. R. Roquin a multifunctional regulator of immune homeostasis. Genes Immun. 17, 7984 (2016).

13. Newman, J. R. S. & Keating, A. E. Comprehensive identication of human bZIP interactions with coiled-coil arrays. Science 300, 20972101 (2003).

14. Vinuesa, C. G. et al. A RING-type ubiquitin ligase family member required to repress follicular helper T cells and autoimmunity. Nature 435, 452458 (2005).

15. Schlundt, A. et al. Structural basis for RNA recognition in roquin-mediated post-transcriptional gene regulation. Nat. Struct. Mol. Biol. 21, 671678 (2014).

16. Tan, D., Zhou, M., Kiledjian, M. & Tong, L. The ROQ domain of Roquin recognizes mRNA constitutive decay element and double-stranded RNA. Nat. Struct. Mol. Biol. 21, 679685 (2014).

17. Murakawa, Y., Rosenbaum, E., Landthaler, M. & Heinemann, U. Roquin binding to target mRNAs involves a winged helix-turn-helix motif. Nat. Commun. 5, 5701 (2014).

18. Zhang, Q. et al. New insights into the RNA-binding and E3 ubiquitin ligase activities of roquins. Sci. Rep. 5, 15660 (2015).

19. Sakurai, S., Ohto, U. & Shimizu, T. Structure of human Roquin-2 and its complex with constitutive-decay element RNA. Acta Crystallogr. Sect. F Struct. Biol. Commun. 71, 10481054 (2015).

20. Schlundt, A., Niessing, D., Heissmeyer, V. & Sattler, M. RNA recognition by Roquin in posttranscriptional gene regulation. Wiley Interdiscip. Rev. RNA 7, 455469 (2016).

21. Li, W., Gao, B., Lee, S.-M., Bennett, K. & Fang, D. RLE-1, an E3 ubiquitin ligase, regulates C. elegans aging by catalyzing DAF-16 polyubiquitination. Dev. Cell 12, 235246 (2007).

22. Maruyama, T. et al. Roquin-2 promotes ubiquitin-mediated degradation of ASK1 to regulate stress responses. Sci. Signal. 7, ra8ra8 (2014).

23. Yu, D. et al. Roquin represses autoimmunity by limiting inducible T-cell co-stimulator messenger RNA. Nature 450, 299303 (2007).

24. Linterman, M. A. et al. Roquin differentiates the specialized functions of duplicated T cell costimulatory receptor genes Cd28 and Icos. Immunity 30, 228241 (2009).

25. Vogel, K. U. et al. Roquin paralogs 1 and 2 redundantly repress the Icos and Ox40 costimulator mRNAs and control follicular helper T cell differentiation. Immunity 38, 655668 (2013).

26. Schuetz, A., Murakawa, Y., Rosenbaum, E., Landthaler, M. & Heinemann, U. Roquin binding to target mRNAs involves a winged helix-turn-helix motif. Nat. Commun. 5, 5701 (2014).

27. Schaefer, J. S., Montufar-Solis, D., Nakra, N., Vigneswaran, N. & Klein, J. R. Small intestine inammation in Roquin-mutant and Roquin-decient mice. PLoS ONE 8, e56436 (2013).

28. Murakawa, Y. et al. RC3H1 post-transcriptionally regulates A20 mRNA and modulates the activity of the IKK/NF-kB pathway. Nat. Commun. 6, 7367 (2015).

29. Mino, T. et al. Regnase-1 and Roquin regulate a common element in inammatory mRNAs by spatiotemporally distinct mechanisms. Cell 161, 10581073 (2015).

30. Janowski, R. et al. Roquin recognizes a non-canonical hexaloop structure in the 30-UTR of Ox40. Nat. Commun. 7, 11032 (2016).

31. Jeltsch, K. M. et al. Cleavage of roquin and regnase-1 by the paracaspase MALT1 releases their cooperatively repressed targets to promote TH17 differentiation. Nat. Immunol. 15, 10791089 (2014).

32. Glasmacher, E. et al. Roquin binds inducible costimulator mRNA and effectors of mRNA decay to induce microRNA-independent post-transcriptional repression. Nat. Immunol. 11, 725733 (2010).

33. Lykke-Andersen, J., Shu, M.-D. & Steitz, J. A. Human Upf proteins target an mRNA for nonsense-mediated decay when bound downstream of a termination codon. Cell 103, 11211131 (2000).

34. Chang, C.-T., Bercovich, N., Loh, B., Jonas, S. & Izaurralde, E. The activation of the decapping enzyme DCP2 by DCP1 occurs on the EDC4 scaffold and involves a conserved loop in DCP1. Nucleic Acids Res. 42, 52175233 (2014).

35. Behm-Ansmant, I. et al. mRNA degradation by miRNAs and GW182 requires both CCR4:NOT deadenylase and DCP1:DCP2 decapping complexes. Genes Dev. 20, 18851898 (2006).

36. Zekri, L., Kuzuolu-ztrk, D. & Izaurralde, E. GW182 proteins cause PABP dissociation from silenced miRNA targets in the absence of deadenylation. EMBO J. 32, 10521065 (2013).

37. Garces, R. G., Gillon, W. & Pai, E. F. Atomic model of human Rcd-1 reveals an armadillo-like-repeat protein with in vitro nucleic acid binding properties. Protein Sci. Publ. Protein Soc. 16, 176188 (2007).

38. Davey, N. E. et al. SLiMPrints: conservation-based discovery of functional motif ngerprints in intrinsically disordered protein regions. Nucleic Acids Res. 40, 1062810641 (2012).

39. Tompa, P. Intrinsically disordered proteins: a 10-year recap. Trends Biochem. Sci. 37, 509516 (2012).

40. Semotok, J. L. et al. Smaug recruits the CCR4/POP2/NOT deadenylase complex to trigger maternal transcript localization in the early drosophila embryo. Curr. Biol. 15, 284294 (2005).

41. Goldstrohm, A. C., Hook, B. A., Seay, D. J. & Wickens, M. PUF proteins bind Pop2p to regulate messenger RNAs. Nat. Struct. Mol. Biol. 13, 533539 (2006).

42. Chicoine, J. et al. Bicaudal-C recruits CCR4-NOT deadenylase to target mRNAs and regulates oogenesis, cytoskeletal organization, and its own expression. Dev. Cell 13, 691704 (2007).

43. Igreja, C. & Izaurralde, E. CUP promotes deadenylation and inhibits decapping of mRNA targets. Genes Dev. 25, 19551967 (2011).

44. Van Etten, J. et al. Human Pumilio proteins recruit multiple deadenylasesto efciently repress messenger RNAs. J. Biol. Chem. 287, 3637036383 (2012).

45. Tritschler, F. et al. Similar modes of interaction enable trailer Hitch and EDC3 to associate with DCP1 and Me31B in distinct protein complexes. Mol. Cell Biol. 28, 66956708 (2008).

46. Haas, G. et al. HPat provides a link between deadenylation and decapping in metazoa. J. Cell Biol. 189, 289302 (2010).

47. Jonas, S., Weichenrieder, O. & Izaurralde, E. An unusual arrangement of two 14-3-3-like domains in the SMG5SMG7 heterodimer is required for efcient nonsense-mediated mRNA decay. Genes Dev. 27, 211225 (2013).

48. Diebold, M.-L., Fribourg, S., Koch, M., Metzger, T. & Romier, C. Deciphering correct strategies for multiprotein complex assembly by co-expression: application to complexes as large as the histone octamer. J. Struct. Biol. 175, 178188 (2011).

49. Cheng, Y. & Patel, D. J. An efcient system for small protein expression and refolding. Biochem. Biophys. Res. Commun. 317, 401405 (2004).

50. Petit, A.-P. et al. The structural basis for the interaction between the CAF1 nuclease and the NOT1 scaffold of the human CCR4NOT deadenylase complex. Nucleic Acids Res. 40, 1105811072 (2012).

51. Braun, J. E., Huntzinger, E., Fauser, M. & Izaurralde, E. GW182 proteins directly recruit cytoplasmic deadenylase complexes to miRNA targets. Mol. Cell 44, 120133 (2011).

52. Kabsch, W. XDS. Acta Crystallogr. D Biol. Crystallogr. 66, 125132 (2010).53. McCoy, A. J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 40,

658674 (2007).

54. Winn, M. D. et al. Overview of the CCP4 suite and current developments. Acta Crystallogr. D Biol. Crystallogr. 67, 235242 (2011).

55. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486501 (2010).

56. Afonine, P. V. et al. Towards automated crystallographic structure renement with phenix.rene. Acta Crystallogr. D Biol. Crystallogr. 68, 352367 (2012).

57. Kyte, J. & Doolittle, R. F. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157, 105132 (1982).

Acknowledgements

We thank Dr Oliver Mhlemann for the kind gift of the HeLa cell line used in knockdown experiments and Dr Jens Lykke-Andersen for providing the plasmids for performing MS2 tethering assays in human cells. We are grateful to Heike Budde, Sigrun Helms, Maria Fauser and Catrin Weiler for excellent technical support. We thank the staff at the PX beamlines of the Swiss Light Source for assistance with data collection. This work was supported by the Max Planck Society.

Author contributions

A.S. and P.B. cloned proteins for expression in human and S2 cells. A.S. conducted tethering assays in human and Drosophila cells and co-immunoprecipitation assays in Drosophila cells with help from D.K.-O. and D.B. P.B. conducted pulldown assays in

NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 15

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14307

human cells. T.R. and A.S. expressed and puried proteins from E. coli for crystallization and pulldown assays, crystallized the CAF40-CBM complex, conducted in vitro pulldown assays and analysed the data. T.R. solved and rened the crystal structure of the CAF40-CBM complex. Y.C. cloned and puried the pentameric complex. O.W. supervised the structural part of the study. E.I. was the principal investigator and conceived and supervised the project. A.S., T.R., O.W. and E.I. wrote the manuscript with contributions from P.B. and D.K.-O.

How to cite this article: Sgromo, A. et al. A CAF40-binding motif facilitates recruitment of the CCR4-NOT complex to mRNAs targeted by Drosophila Roquin. Nat. Commun. 8, 14307 doi: 10.1038/ncomms14307 (2017).

Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2017

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

16 NATURE COMMUNICATIONS | 8:14307 | DOI: 10.1038/ncomms14307 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

Word count: 13625

Show less

Abstract

Translate

Human (Hs) Roquin1 and Roquin2 are RNA-binding proteins that promote mRNA target degradation through the recruitment of the CCR4-NOT deadenylase complex and are implicated in the prevention of autoimmunity. Roquin1 recruits CCR4-NOT via a C-terminal region that is not conserved in Roquin2 or in invertebrate Roquin. Here we show that Roquin2 and Drosophila melanogaster (Dm) Roquin also interact with the CCR4-NOT complex through their C-terminal regions. The C-terminal region of Dm Roquin contains multiple motifs that mediate CCR4-NOT binding. One motif binds to the CAF40 subunit of the CCR4-NOT complex. The crystal structure of the Dm Roquin CAF40-binding motif (CBM) bound to CAF40 reveals that the CBM adopts an α-helical conformation upon binding to a conserved surface of CAF40. Thus, despite the lack of sequence conservation, the C-terminal regions of Roquin proteins act as an effector domain that represses the expression of mRNA targets via recruitment of the CCR4-NOT complex.

Details

Title

A CAF40-binding motif facilitates recruitment of the CCR4-NOT complex to mRNAs targeted by Drosophila Roquin

Author

Sgromo, Annamaria; Raisch, Tobias; Bawankar, Praveen; Bhandari, Dipankar; Chen, Ying; Kuzuoglu-Öztürk, Duygu; Weichenrieder, Oliver; Izaurralde, Elisa

Pages

14307

Publication year

2017

Publication date

Feb 2017

Publisher

Nature Publishing Group

e-ISSN

20411723

Source type

Scholarly Journal

Language of publication

English

DOI

https://doi.org/10.1038/ncomms14307

ProQuest document ID

1865238851

A CAF40-binding motif facilitates recruitment of the CCR4-NOT complex to mRNAs targeted by Drosophila Roquin

Jump to:

Full Text

Abstract

Details

Suggested sources