Production of a Mouse Line with a Conditional Crim1 Mutant Allele

Crim1 encodes a developmentally expressed transmembrane protein that contains six von Willebrand Factor-C (vWFC)-like cysteine-rich repeats (CRRs) similar to the BMP-regulating protein, Chordin (Georgas et al., 2000; Kolle et al., 2000; Pennisi et al., 2007). Crim1 can bind a broad range of cystine-knot growth factors, including TGFβ, BMP, VEGF, and PDGF family members, and such binding occurs only when Crim1 is co-expressed in the same cell as the growth factor (Wilkinson et al., 2003). Recent studies have further supported the idea that Crim1 has a role in antagonizing BMP function, including the Drosophila homolog of Crim1, Crimpy (James and Broihier, 2011; Zhang et al., 2011). Our analysis of a Crim1 gene-trap line mutant, Crim1KST264, found that homozygous Crim1KST264/KST264 mice died perinatally on a C57Bl6 genetic background and displayed abnormal kidney, eye, limb and placental development (Pennisi et al., 2012; Pennisi et al., 2007). When bred onto a mixed C57Bl6/CD1 background, Crim1KST264/KST264 mice display a similar array of phenotypes, but can survive to adulthood (with reduced viability; (Wilkinson et al., 2007)). Exon 2-minus transcripts are produced at low levels from the Crim1KST264 locus and are predicted to produce in-frame transcripts, suggesting the Crim1KST264 may be a hypomorphic or gain-of-function mutation (Pennisi et al., 2007). Our previous work showed that Crim1 binds and regulates VEGF-A activity in vivo, with Crim1KST264/KST264 displaying excessive VEGF-A diffusion away from the podocytes of the renal glomerulus (where Crim1 and VEGF-A are co-expressed), resulting in increased activation of VEGFR-2 in adjacent vascular endothelial cells (Wilkinson et al., 2007). Crim1 is expressed in numerous cell types during development and homeostasis, and the activity of numerous growth factors derived from multiple cell types is likely to be perturbed upon mutation of Crim1 (Georgas et al., 2000; Kolle et al., 2000; Pennisi et al., 2007; Wilkinson et al., 2007; Wilkinson et al., 2003). Thus, analysis of the pleiomorphic phenotypes observed in Crim1KST264/KST264 mice is problematic. To overcome some of these difficulties, and to facilitate investigations on tissue-specific and postnatal roles for Crim1 in embryonic development and disease, we generated a Crim1 conditional knock-out mouse line by flanking exons 3 and 4 with unidirectional LoxP sites (Crim1FLOX allele, Figure 1). Figure 1 Generation of a conditional mutant Crim1 allele. (a) Ideogram of the targeting strategy to produce the Crim1FLOX conditional allele. Shown are the Crim1 gene (exons not to scale), the region including exon 3 and exon 4 of the wild-type Crim1 locus to ... PCR analysis of genomic DNA was used to confirm the production of the mutant Crim1 alleles and to establish a convenient method for genotyping (Figure 2 a–c). Crim1FLOX/FLOX mice on C57Bl6 and CD1 genetic backgrounds were viable, fertile, and displayed no anomalous phenotypes relative to Crim1+/FLOX or Crim1+/+ littermates (not shown). Colonies of Crim1Δflox mice were established after breeding Crim1FLOX mice with a CMV-Cre line. We then confirmed that exons 3 and 4 were deleted in Crim1Δflox allele. PCR was performed on the genomic DNA (gDNA) of embryonic samples of various genotypes using primer pairs specific for Crim1 exons 1 and 2 (5´ to the FLOXed region), exon 3 (within the FLOXed region), and exon 11 (3´ to the FLOXed region). Amplicons for all exons tested were observed for Crim1+/+ and Crim1+/Δflox samples. However, only exons 1, 2 and 11 were observed in Crim1Δflox/Δflox samples (Figure 2 d). Figure 2 Genotyping and genomic DNA analysis of mutant Crim1 alleles. (a) Ideogram of Crim1FLOX conditional and Crim1Δflox mutant alleles showing the primers used for genotyping (arrows) and LoxP sites (triangles). (b) PCR genotyping for the Crim1FLOX ... Characterization of the transcripts from the Crim1Δflox locus was performed using RT-PCR on mRNA from embryonic kidney samples. Full-length transcripts were observed in Crim1+/+ samples, however we only found a shorter transcript lacking the regions encoded by exons 3 and 4 in Crim1Δflox/Δflox samples (Figure 3 a). In addition, no transcripts from exon 3 could be detected by real-time PCR with exon-specific primer pairs on whole kidney cDNA (qRT-PCR), further confirming the Cre-mediated deletion of the genomic region in Crim1Δflox/Δflox mice (Figure 3 b). The Crim1Δflox transcript lacks the region encoded by exons 3 and 4, that produces an out-of-frame transcript resulting in a premature stop codon, and is predicted to be non-functional. Section immunohistochemistry was then performed with an anti-Crim1 antibody on kidneys from Crim1+/Δflox and Crim1Δflox/Δflox 16dpc embryos. As expected, we observed immunostaining signal in Crim1+/Δflox samples. However, no signal was observed in Crim1Δflox/Δflox samples, consistent with the lack of a functional protein translated from Crim1Δflox transcripts (Figure 3 c–e). Figure 3 Confirmation of the mutant nature of the Crim1Δflox allele. (a) Qualitative RT-PCR analysis of Crim1 transcripts from total 15.5dpc kidney mRNA from Crim1+/+ and Crim1Δflox/Δflox embryos. Primers were designed to amplify transcripts ... We intercrossed Crim1+/Δflox mice to examine the nature of the conditional Crim1 mutation. From matings with C57Bl6 background, we did not observe Crim1Δflox/Δflox pups (Table 1). By contrast, from matings of Crim1+/Δflox mice with a CD1 background, Crim1Δflox/Δflox pups were observed at a frequency less than would be expected based on Mendelian ratios (approximately 62.5% of expected values; Table 1). The frequency of the genotypes of embryos at different stages was then analyzed after intercrosses of Crim1+/Δflox mice with a C57Bl6 background (Table 2). Although there may be a trend of reduced viability of Crim1Δflox/Δflox embryos, a statistically significant difference in the numbers of Crim1Δflox/Δflox mice was not observed until the postnatal period. Crim1Δflox/Δflox embryos displayed a range of phenotypes, including peridermal blebbing evident at 12.5dpc and 13.5dpc (Figure 4 a, b); mild digit syndactyly from 13.5dpc (Figure 4 c–f); eye hypoplasia from 13.5dpc (Figure 4 g, h); renal hypoplasia (Figure i, j) and glomerular dysgenesis (Figure 4 k–n); and a proportion displayed widespread edema (Figure 4 o, p). Crim1Δflox/Δflox mice, like the previously described Crim1KST264/KST264 mice, are characterized by a pleiomorphic phenotype affecting numerous organ systems in development. These include the perinatal lethality on a C57Bl6 background, renal, eye, and digit dysgenesis, and peridermal blebbing, are consistent with those described in Crim1KST264/KST264 embryos (Pennisi et al., 2007). However, a widespread edema was not observed in Crim1KST264/KST264 embryos. Like Crim1KST264/KST264 mice, Crim1Δflox/Δflox mice on a predominantly CD1 genetic background can survive postnatally, albeit at reduced viability (Wilkinson et al., 2007). Figure 4 The phenotype of Crim1Δflox/Δflox embryos resembles that of Crim1KST264/KST264 embryos. (a, b) Micrographs of Crim1+/+ (a) and Crim1Δflox/Δflox (b) 13.5dpc embryos viewed in whole-mount. Note the peridermal blebbing in ... Table 1 The frequency of the genotypes of pups from intercrosses of Crim1+/Δflox mice with a C57Bl6 or CD1 genetic background. Table 2 The frequency of the genotypes of embryos at different stages from intercrosses of Crim1+/Δflox mice with a C57Bl6 background. It is noteworthy that, although similar phenotypes were observed among Crim1Δflox/Δflox embryos and mice, relative to that described for the Crim1KST264/KST264 mutation, there appeared to be some difference in the severity and/or the penetrance of phenotypes. This likely reflects the differences in the Crim1Δflox and Crim1KST264 mutations. Transcripts from the Crim1KST264 allele include a fusion of the coding region of Crim1 exon 1 and the β-Geo from the gene-trap and an in-frame exon 2-minus splice variant (Pennisi et al., 2007). This latter transcript is also detected in wild-type embryos as a minor splice variant in numerous developing organs (Pennisi et al., 2007). Importantly, RT-PCR analyses on transcription from the Crim1Δflox allele confirm an out-of-frame transcript lacking exon 3 and exon 4. Although unlikely, an alternative explanation remains a formal possibility; as there have been less than ten generations of backcrossing, it is possible that some residual genetic material from the chimeric founders may contribute sufficient genetic variability. The production of a mouse line with a conditional Crim1FLOX allele will be an invaluable tool in dissecting the tissue-specific contribution of Crim1 to the development of different organ systems. Furthermore, by the use of inducible Cre-expressing lines, a conditional null Crim1 allele will allow the examination of the role of Crim1 in postnatal development or disease by circumventing confounding embryonic defects.