Conditional mutagenesis of CamKIV

Abstract

Ca /Calmodulin-dependent kinase IV (CamKIV) has been suggested to be involved in multiple physiological processes including male and female fertility, T-cell maturation, and long-term memory (Corcoran and Means, 2001). It is activated in response to Ca /calmodulin and by phosphorylation of Ca /Calmodulin-dependent kinase. CREB (cAMP response element binding protein) is one of the best defined substrates of this kinase. CamKIV activates CREB-dependent transcription via phosphorylation of serine 133 in the CREB protein (Corcoran and Means, 2001). Inactivation of the CamKIV (Camk4) gene has been reported by two independent groups. Interestingly, in one study the inactivation of CamKIV shows a severe phenotype consisting of impaired viability, male and female infertility, and neurological disorders (Ribar et al., 2000; Wu et al., 2000; Wu et al., 2000), whereas only a mild neurological phenotype was observed in an independent study (Ho et al., 2000). The discrepancy in the phenotypes of the two knockouts could be due to the different placement of the PGKneo cassette into the targeting constructs. It has been shown that the presence of the PGKneo cassette can influence the phenotype of knockout mice (Scacheri et al., 2001). Thus, the generation of CamKIV conditional knockout mice using the Cre/loxP system is an attractive alternative since it allows the generation of a null mutation that lacks the PGKneo cassette (Gu et al., 1994). A fragment of DNA containing the ATP binding domain of the CamKIV gene (exon III) was isolated from a 129-PAC genomic library. Since it has previously been shown that mutation of lysine 71 in the ATP binding domain abolishes kinase activity (Chatila et al., 1996), we chose to flank exon III, which contains this residue with two loxP sites. In addition, deletion of this exon should create a frame shift that prevents translation of the rest of the protein. The targeting construct was generated using ET-cloning (Angrand et al., 1999; Zhang et al., 1998). In the first round of ET-cloning, a loxP site was introduced downstream of exon III of the targeting construct (Fig. 1A). In the second round, the loxP●FRT●PGKTn5neo●FRT cassette was introduced upstream of exon III (Fig. 1A). Finally, the TK gene was introduced in the targeting construct by standard cloning procedures to generate a TK/neo fusion gene (PGKTK/neo). A total of 129 ES cells were electroporated with 20 g of linearized construct. ES cell clones were analysed for the correct homologous recombination event by Southern blot analyses using a 5 external probe (Fig. 1B), a 3 external probe (Fig. 1C) and an internal probe (data not shown). The FRT●PGKTn5TK/neo●FRT cassette was removed from correctly targeted ES cell clones by electroporation of the FLPe-expressing plasmid pCAGGS-FLPe (Fig. 1A) (Schaft et al., 2001). From 200 gancyclovirresistant clones, two were positive for excision of the FRT●PGKTn5TK/neo●FRT cassette as determined by Southern blot analysis. Chimeric animals were generated by blastocyst injection of positive ES cell clones and mated to C57BL/6 animals for germ-line transmission. In order to validate the functionality of the loxP sites in vivo, CamKIV flox animals (CamKIV) were mated with transgenic mice expressing the Cre recombinase under the control of a CamKII BAC (CamKII iCre) (Casanova et al., 2001). Double-transgenic animals CamKIV, CamKII iCre (CamKIV ) showed selective loss of the CamKIV protein only in those regions where the Cre recombinase is expressed. Specifically, CamKIV expression was lost from hippocampus, striatum, and cortex but not cerebellum (Fig. 2). In summary, we have generated mice with a conditional CamKIV allele using a combination of ET-cloning and the Cre/loxP and FLP/FRT systems, which allows the rapid generation of targeting constructs. Furthermore, the ET-cloning technique allows the introduction of loxP and FRT sites into DNA which is independent of restric-