We report here that Capn2-/- mice, which lack the m-calpain catalytic subunit, die at the preimplantation stage of development. This observation allows us to now conclude that m-calpain is specifically required during embryogenesis. Since homozygous disruption of the Capn4 gene was also expected to abolish m-calpain activity, this result is in agreement with the phenotype presented by one of the two Capn4 targeted lines in which preimplantation lethality of null embryos was also observed [32]. The survival of Capn4-/- embryos from the original targeted line reported by Arthur and colleagues to mid-gestation is more difficult to reconcile [31]. In retrospect, it seems likely that the latter line represents a hypomorphic state, rather than a true null. The Capn4 targeting strategy employed by Arthur and colleagues involved disrupting the C-terminus of the calpain small subunit by insertion of the PGK-Neo cassette into the middle of exon 9, which caused truncation of the protein [31]. This strategy was based upon previous structure/function studies showing that excision of the C-terminal 25 amino acid residues of the small subunit abolished calpain activity when co-expressed with the m-calpain large subunit in E. coli [34]. In contrast, the Capn4 targeting strategy employed by Zimmerman and colleagues involved a much more extensive deletion of exons 4 through 8 [32]. It now seems probable that the difference in the time of lethality of these two targeted Capn4 lines can be explained by different extents of disruption. The Zimmerman allele probably represents a true null genotype while the Arthur allele is likely a hypomorphic mutation. Alternate targeting strategies have been shown to yield different phenotypes in gene disruption studies. For example, three different targeting strategies were independently used to disrupt the mixed lineage leukemia (Mll) gene. In all three studies, homozygous null embryos perished during embryogenesis, but at different stages (E0.5, E10.5, E14.5) [35]. The variation in phenotype was attributed to the differences in degree of function of the truncated proteins produced from the mutant alleles. A similar effect might be at work in the two Capn4 transgenic lines. Efforts were made to detect transcripts or calpain activities derived from the Arthur Capn4 allele. This allele gave rise to multiple mRNA species, detectable by RT-PCR, reading through from the first half of exon 9 to at least two different cryptic splice sites in the PGK promoter sequence [31]. These transcripts could give rise to defective calpain small subunits with 10–30 inappropriate C-terminal acids, which might be sufficient to support a low level of calpain activity. However, expression of calpains with these modified small subunits did not give rise to any detectable calpain activity when expressed in E. coli, although their functionality in mammalian cells has yet to be determined (J.S. Elce, unpublished work) It has also been suggested that calpain large subunits alone might provide some activity in eukaryotic cells, although the Zimmerman et al. knockout appears to exclude that possibility, and no calpain activity was observed in our hands when calpain large subunits were expressed alone either in E. coli or in mammalian cells [36]. The different timing of lethality in the Capn4 knockouts might also be a consequence of the different genetic backgrounds of the two transgenic lines, which has been observed to influence the phenotype of transgenic mice on a number of occasions [37].