Fig. 1. Replacing endogenous Math5 with Neurod1 or Math3. (A) Sequence relationships among bHLH domains of representative proneural bHLH genes [method described by Ledent et al. (Ledent et al., 2002)]. Mash1 was chosen as the outgroup. Branch lengths are proportional to the distance between the sequences. (B) Genome structure for Math5, the targeting construct and the predicted structure of the targeted Neurod1 and Math3 knock-in alleles. The single Math5 exon is depicted as a black box. The black bars underneath indicate the DNA fragments amplified from genomic DNA for the targeting construct. Red boxes depict FRT recombination sites. Blue boxes indicate SV40-pA sequence. Black arrows indicated the PCR primers used to amplify Math5 and Math5^Math3-KI for qRT-PCR analysis. The primer sequences are described in the Materials and methods. (C) Representative Southern blot analysis using the 5' probe to distinguish Math5 wild-type and Math5^Neurod1-KI alleles from genomic DNA of targeted ES cells. O indicates a targeted ES cell. (D) Representative Southern blot analysis using the Southern probe depicted in B to distinguish Math5 wild-type, Math5^Neurod1-KI and Math5^lacZ-KI alleles from tail genomic DNA of littermates resulting from a Math5^{Neurod1-KI/lacZ-KI} x Math5^lacZ-KI/+ cross. (E-H) Misexpression of Neurod1 and Math3 from the Math5 locus. Retinal sections from E12.5 wild-type (E,G) and Math5^{Neurod1-KI/Math3-KI} (F,H) embryos were immunostained with anti-Neurod1 antibody (E,F) or labeled with a Math3 antisense probe by in situ hybridization (G,H). The images in G and H have been enhanced using Photoshop. (I) qRT-PCR analysis of Math5 and Math5^Math3-KI alleles. Gene expression levels were normalized to the expression of endogenous GAPDH transcripts. Scale bar: 100 µm.