Bicoid associates with the 5'-cap-bound complex of caudal mRNA and represses translation

Abstract

Translational control plays a key role in many biological processes including pattern formation during early Drosophila embryogenesis. In this process, the anterior determinant Bicoid (BCD) acts not only as a transcriptional activator of segmentation genes but also causes specific translational repression of ubiquitously distributed caudal (cad) mRNA in the anterior region of the embryo. We show that translational repression of cad mRNA is dependent on a functional eIF4E-binding motif. The results suggest a novel mode of translational repression, which combines the strategy of target-specific binding to 3'-untranslated sequences and interference with 5'-cap-dependent translation initiation in one protein.

Figure 1

BCD copurifies with 5′-cap-bound proteins. Cytoplasmic protein extracts of young embryos were affinity-purified using a cap-analog m⁷GTP-sepharose resin (Edery et al. 1988). (a) Silver-stained SDS-PAGE of affinity-purified proteins contained within cytoplasmic extracts of wild-type and bcd mutant Drosophila embryos. Arrowhead marks the protein band that was subsequently identified as BCD (left, see also lanes 1–4 in b). Note that eIF4E, the most abundant 30-kD component among the purified proteins, is run off the gel (left) to obtain maximum resolution of the relevant range of protein bands between 50 and 100 kD. (Lanes 1–4) Relevant portions of silver-stained SDS-PAGE (dotted box) including a ∼70-kD protein (lane 1, arrowhead) that is recognized by monospecific anti-BCD antibodies on Western blots (lane 2). This protein is not present in silver-stained SDS-PAGE (lane 3), and anti-BCD antibodies stained Western blots (lane 4) of corresponding extracts from embryos that derived from homozygous bcd mutant females. (b) Relevant portions of silver-stained SDS-PAGE showing that the GFP–BCD fusion protein (lane 1, 95 kD) is absent in wild-type embryos (lane 2). Corresponding Western blots show that the 95-kD fusion protein reacts with both anti-BCD (lane 3; lane 4 is a wild-type control lacking the GFP–BCD transgene) and anti-GFP antibodies (lane 5; lane 6 is a wild-type control lacking the GFP–BCD transgene; asterisk marks cross-reacting protein of unknown identity). (c) Schematic diagram of BCD (positions refer to sequence according to Berleth et al. 1988) showing the homeodomain (HD, gray box; position 91–154), the PEST domain (PEST, black box; position 170–203), and a YIRPYL motif (green box; position 68–73). The eIF4E-binding properties of this motif have been recently analyzed in great detail in the context of human BP1 (Marcotrigiano et al. 1999; for review, see Raught et al. 2000; Sachs and Varani 2000; Miron et al. 2001). (d) Western blots showing that BCD bound to m⁷GTP-sepharose coupled recombinant eIF4E (lanes 1,4) can be competed for with 100 nM (lane 2) and 1 mM (lanes 3,5) of a peptide containing the YDRKFL motif of human BP1, whereas 1 mM of a mutated human BP1 peptide (mutated motif is ADRKFR) did not interfere with the binding of BCD to eIF4E (lane 6). (e,f) Autoradiogram showing that in vitro translated ³⁵S-labeled BCD (input; a schematic representation of the protein is shown at the bottom of e) is capable of interacting with m⁷GTP-sepharose-bound recombinant eIF4E in the presence of in vitro transcribed cad 3′-UTR (e), whereas in vitro translated ³⁵S-labeled BCD^1–489AR (input; a schematic representation of the protein is shown at the bottom of f) showed no significant binding to eIF4E (f). Note that there is no unspecific binding of in vitro translated protein to m⁷GTP-beads without precoupled eIF4E. For details, see Materials and Methods.

Figure 2

Functional analysis of mutant BCD by transgene-dependent expression in bcd mutant embryos. (a) Schematic representation of the BCD-deletion mutants (cf. Fig. 1c, wild-type BCD). (b–d) Wild-type embryos are characterized by translational repression of cad mRNA (absence of anti-CAD antibody staining (green) in the anterior region of preblastoderm embryos (b; Dubnau and Struhl 1996; Rivera-Pomar et al. 1996), by activation of zygotic hb transcription as revealed by whole mount in situ hybridization with an hb cDNA probe (c; Klingler and Gergen 1993; reviewed in Martinez Arias 1993), and by the wild-type cuticle pattern (d). (e–g) Embryos derived from homozygous bcd^E1 mutant females fail to repress cad mRNA translation (e), lack the anterior hb expression domain (which is replaced by a duplication of the posterior, BCD-independent hb expression domain; f), and show a bcd mutant cuticle phenotype (g; Frohnhöfer and Nüsslein-Volhard 1986). (h–j) Transgene-derived BCD^1–489 expression restores all aspects of the bcd mutant phenotype including translational repression of cad mRNA (h), anterior hb expression (i), and the larval cuticle phenotype (j). (k–m) cad mRNA translation is not repressed in response to transgene-derived BCD^77–202 (k) or BCD^89–202 (data not shown). In both cases, the transgene-expressed BCD deletion mutants also fail to restore anterior hb expression (l) and develop a bcd mutant cuticle phenotype (m) because of the absence of the C-terminal transactivation domains (Sauer et al. 1995; Schaeffer et al. 1999). Orientation of embryos is anterior left and dorsal up.

Figure 3

Translational repression of cad mRNA, transcriptional activation of anterior hb expression, and rescue of segmentation defects in response to transgene-expressed BCD replacement mutants. (a) Schematic representation of the BCD replacement mutants BCD^1–202AR and BCD^1–489AR in which the YIRPYL motif is changed into AIRPYR (green box; for other details, see Fig. 1c). (b,d,f) Embryos from homozygous bcd^E1 females expressing transgene-derived BCD^1–202AR fail to repress cad mRNA translation (b) or to activate anterior hb expression (d), and develop a bcd mutant cuticle phenotype (f). (c,e,g) Embryos from homozygous bcd^E1 females expressing transgene-derived BCD^1–489AR fail to repress cad mRNA translation (c) but activate anterior hb expression (e) and develop the normal head and thoracic pattern elements (g). For further details, see text.

Figure 4

Different modes of cap-dependent translational repression by interference with the assembly of the eIF4E ∷ eIF4G interaction complex. (a) Binding of BP to eIF4E blocks eIF4G-binding and modulates translation efficiency of 5′-capped mRNAs in an insulin signaling-dependent manner (for review, see Raught et al. 2000; Sachs and Varani 2000). (b) Translational repression of mRNAs, which anchor CPEB through a CP element in their 3′-UTR. CPEB is able to associate with Maskin, which successively blocks the eIF4E ∷ eIF4G interaction by binding to eIF4E (for review, see Richter 2000; Richter and Theurkauf 2001). (c) BCD uses a similar strategy of repression by combining the binding properties of both CPEB and Maskin. BCD binds directly to the BRE in the 3′-UTR and blocks the eIF4E ∷ eIF4G interaction at the 5′-end. In each case (a–c), the eIF4E interaction involves the YxxxxL motif of the translational repressors.