September 15, 2015;
A novel role for Celf1 in vegetal RNA localization during Xenopus oogenesis.
The localization of certain mRNAs to the vegetal cortex of Xenopus oocytes is of crucial importance for germ cell development and early embryonic patterning. Vegetal RNA localization is mediated by cis-acting RNA localization elements (LE). Several proteins assemble on the RNA LE and direct transport to the vegetal cortex. Although a number of localization RNP components have been identified, their full composition is unknown. In an RNA affinity purification approach, using the dead end 1 (dnd1) RNA LE, we identified Xenopus Celf1
as a novel component of vegetal localization RNP complexes. Celf1
is part of an RNP complex together with known vegetal localization factors and shows specific interactions with LEs from several but not all vegetally localizing RNAs. Immunostaining experiments reveal co-localization of Celf1
with vegetally localizing RNA and with known localization factors. Inhibition of Celf1
protein binding by localization element mutagenesis as well as Celf1
overexpression interfere with vegetal RNA localization. These results argue for a role of Celf1
in vegetal RNA localization during Xenopus oogenesis.
Elavl Ab 2
[+] show captions
Fig. 1. Celf1 is a novel component of the vegetal RNA localization complex in Xenopus oocytes. (A) Gel-electrophoretic analysis of the total protein composition of RNA-affinity-purified complexes. Proteins specifically co-precipitating with dnd1-LE and/or velo1-LE but not, or only to a lesser extent, with the β-globin 3′-UTR control RNA, are marked by an asterisk. Protein bands corresponding to TEV-protease and PP7CP are indicated. Co-purified RNAs appear as yellow bands in the silver stained gel. (B) Western blot analysis for co-purification of known localization proteins with tagged dnd1-LE, velo1-LE and a control RNA. (C) Comparative co-immunoprecipitation analysis with Flag-tagged versions of Celf1 and Ptbp1 in the absence or presence of RNase A. Co-precipitating proteins from total oocyte extract were detected by Western blot as indicated. Immunoprecipitation with extract from uninjected oocytes served as negative control.
Fig. 2. Celf1 is a cytoplasmic protein that co-localizes with vegetal RNAs and known localization factors in Xenopus oocytes. (A) Temporal analysis of Celf1 expression during Xenopus oogenesis and embryogenesis. Western blot analysis of Celf1 with equivalent amounts of oocyte and embryonic extract; stages of oogenesis and embryogenesis were as indicated. (B) Western blot analysis of Celf1 with cytoplasmic (C) and nuclear (N) fractions from staged oocytes. α-Stau1 was utilized as control for a cytoplasmic ( Allison et al., 2004) and α-Hnrnpab for a predominantly nuclear oocyte protein ( Czaplinski et al., 2005). (C) Co-localization analysis for endogenous Celf1 protein and vegetally localizing, microinjected Cy3-dnd1-LE RNA in Xenopus oocytes. α-Celf1 immunostaining was performed on Cy3-dnd1-LE RNA injected stage III oocytes. (D) α-Celf1/α-Igf2bp3 co-immunostaining was performed with stage III oocytes that were injected with Cy3-dnd1-LE. Single confocal sections are shown. Scale bars indicate 100 µm (whole oocyte) and 20 µm (magnification).
Fig. 3. Celf1 interacts with different vegetally localizing RNAs from Xenopus oocytes. (A) In vitro interaction of Celf1 with different LE RNAs. Cy3-labeled LE RNAs and β-globin 3′UTR control RNA were co-immunoprecipitated with Flag-tagged, in vitro translated Celf1 or Ptbp1. Unprogrammed reticulocyte lysate served as a negative control. Supernatant and bound RNAs were separated by urea-PAGE and detected by fluorescence imaging. (B) Enrichment of endogenous localizing RNAs in Celf1-containing RNPs. Either α-Celf1 or α-Myc (control) antibodies were used for immunoprecipitation reactions from stage III/IV oocyte extracts; co-precipitating RNAs were detected by qPCR. Relative RNA enrichment in the IP-fractions versus the input and control IP-fractions is depicted; average enrichment factors from three independent experiments are shown. The error bars indicate standard error of the mean.
Fig. 4. In vitro binding of Celf1 to dnd1-LE involves U/G-rich elements. (A) Co-immunoprecipitation reactions were performed with Cy3-labeled full length (FL) and 5′- or 3′-deleted fragments of the dnd1-LE RNA and in vitro translated Flag-tagged proteins. The region critical for Celf1 binding is marked in light gray (nt 86-145). (B) Binding affinities of different bacterially expressed known localization factors to wild-type (wt) and a mutant (mut) version of dnd1-LE RNA that carries multiple U/C to A point mutations in the U/G-rich Celf1 binding region, were analyzed by electrophoretic mobility shift assays. The relative amount of complexed RNA is plotted against the individual protein concentrations using non-linear curve fitting; KD values for wild-type and mutant dnd1-LE RNAs are indicated.
Fig. 5. Efficient binding of Celf1 to dnd1-LE RNA is critical for vegetal RNA localization. Wild-type and the mutant dnd1-LE with reduced affinity for Celf1 were analyzed for their ability to localize to the vegetal cortex in Xenopus oocytes. (A) Nuclei of stage III oocytes were injected with Cy3-labeled wild-type (wt) or mutant (mut) dnd1-LE RNA, fixed after 0–48 h, and analyzed by confocal imaging. Single confocal sections are shown. The scale bar indicates 100 µm. (B) Regional RNA quantification was done by measuring the ratio of the mean fluorescence along a line at the vegetal cortex (Vg) versus a line in the animal cytoplasm (An). Kinetics of localization in a time span of 5 to 48 h after microinjection of mutant (mut) or wild-type (wt) dnd1-LE RNA. Error bars indicate standard deviation of approx. 10 oocytes. (C) Quantification of localization efficiency for wild-type (wt) and mutant (mut) dnd1-LE. Stage III oocytes were co-injected with wild-type (wt) or mutant (mut) Cy3- or Atto633-labeled dnd1-LE RNAs. Average localization efficiencies of three independent experiments are shown. The error bars indicate standard error of the mean. p<0.05 is indicated (*).
Fig. 6. Overexpression of Celf1 reduces vegetal localization of dnd1-LE in a dose dependent manner. (A) Stage III–IV oocytes pre-injected with indicated amounts of RNA encoding Flag-tagged Celf1 or eGFP. Localization of lacZ-dnd1-LE was assayed by whole mount in situ hybridization against the lacZ-tag. (B) Quantification of localization capability. Variations in oocyte localization capability were equalized by normalization to localization activity in eGFP injected oocytes, which was set to 100%. Average results of three independent experiments are shown. The error bars indicate standard error of the mean. p<0.05 is indicated (*).