CORRECTION: Gmnc Is a Master Regulator of the Multiciliated Cell Differentiation Program

The above report analyzed the requirement of gmnc in the formation of multiciliated cells (MCCs) in zebrafish embryos. We used CRISPR/Cas9 technology to generate a gmnc mutant fish line. It has come to our attention that the gmnc locus deletion characterization in our paper was incorrect. We had used a combination of two guide RNAs to make the deletion at the gmnc locus. Even though the F0 fish showed the expected deletion, however, only the lesion mediated by a single cut due to the guide RNA sgRNAex1, instead of the double cuts described in the paper, was transmitted to the F1 generation. We have now characterized the lesion as a 25-bp deletion at the target site, which is predicted to result in a severely truncated (and likely non-functional) Gmnc polypeptide upon transcription and translation. As a result, several words on page 3267 of the main text, as well as Figures S1D and S1E and the corresponding sections in the Supplemental Experimental Procedures, have been corrected in the version of the article available online. These corrections do not modify the conclusions of the paper that gmnc is required for the formation of MCCs in zebrafish embryos.
Furthermore, our paper studied the interaction between GMNC and the E2F4/DP1 complex. We performed co-immunoprecipitation experiments between GMNC-HA and GFP-E2F4/GFP-DP1 and concluded that GMNC does not interact with E2F4/DP1. We discovered subsequently that the human GMNC-HA construct we used had a mutation in the GMNC open reading frame that resulted in a premature STOP codon. We obtained an error-free GMNC-HA construct, and additionally we used the Myc epitope to tag E2F4 and DP1 instead of GFP. We repeated the interaction, and the results showed that there is no interaction between GMNC and E2F4/DP1, the same conclusion we had reached originally. We have corrected Figure S3F and the corresponding sections in the Supplemental Experimental Procedures accordingly.
The corrected panels Figures S1D, S1E, and S3F are shown here, along with the original panels for comparison. The authors apologize for these oversights.