Campomelic dysplasia (CD), a semilethal human skeletal malformation syndrome with XY sex reversal, is caused by heterozygous mutations in the SRY-related transcription factor gene SOX9 (Foster et al., 1994; Wagner et al., 1994). The mouse Sox9 gene, located on distal chromosome 11, is expressed during chondrogenesis (Wright et al., 1995), and is a key regulator of cartilage differentiation (Bi et al., 1999) and testis development (Vidal et al., 2001). Sox9 is expressed in a variety of embryonic tissues (Ng et al., 1997), suggesting further functions for Sox9/SOX9 during organogenesis. Indeed, CD patients show malformations in nonskeletal organs such as brain, heart, and kidneys (Houston et al., 1983; Mansour et al., 1995). Recently, Sox9 / mutant mice have been generated by a classical gene targeting approach and were shown to phenocopy the skeletal malformations observed in CD (Bi et al., 2001). However, because Sox9 / mice die shortly after birth, it was not possible to generate a stable Sox9 mutant mouse line. To circumvent this problem and to provide a tool for an in-depth analysis of Sox9 functions in skeletal development and the formation of other organs, we have generated a conditional Sox9 mutant allele in mice using the Cre/loxP system (Nagy, 2000; Le and Sauer, 2001) (Fig.1). Five correctly targeted embryonic stem (ES) cell clones were identified, and, following successful deletion of the neo cassette by transient in vitro expression of Cre recombinase (Taniguchi et al., 1998), two clones were microinjected into C57BL/6 blastocysts. Germline transmission was obtained for both clones and two Sox9-flox mouse lines have been established by backcrossing to C57BL/6 mice. Both Sox9 homozygous male and female mice are viable and fertile and do not show any obvious phenotype. To test for conditional inactivation of Sox9 in vivo, Sox9 /flox mice were crossed to Cre deleter mice which ubiquitously express the Cre recombinase (Schwenk et al., 1995) (Fig 2a). In Sox9 ; Cre–transgenic animals, designated as Sox9 / , exon 2 encoding half of the DNA-binding high-mobility group (HMG) domain and exon 3 encoding the entire transactivation domain are deleted (Fig. 2a, b). Because the transactivation domain is essential for SOX9 function (Sudbeck et al., 1996), and because truncated HMG domains cannot bind to DNA (Giese et al., 1991), our strategy most likely results in a functional null allele of the Sox9 gene. Complete recombination of the loxP-flanked Sox9 allele was demonstrated by allele-specific PCR analysis of Sox9 / mutants (Fig. 2c). Sox9 / mice died shortly after birth, presumably because of severe respiratory failure caused by a cleft secondary palate. Skeletal staining of newborn Sox9 / mutants revealed bone malformations typically observed in CD patients (Fig. 2 and data not shown). For example, the scapula was hypoplastic and malformed and the deltoid tuberosity of the humerus was missing (Fig. 2e). In addition, premature mineralization occurred in many skeletal elements, including the tail vertebrae (Fig. 2g). This is in agreement with the recently proposed role for Sox9 in regulating the rate of hypertrophic chondrocyte differentiation (Bi et al., 2001). Taken together, our analysis revealed that the skeletal phenotype of conditional Sox9 / mutants is very similar to that described for Sox9 / mutants (Bi et al., 2001). The different genetic background of mice used in the two studies may account for minor differences we observed in the severity of the skeletal phenotypes. The availability of tissue-specific or inducible Cre-transgenic mouse strains will make the conditional Sox9 allele a valuable tool not only to understand cartilage differentiation and skeletogenesis but also to investigate Sox9 functions during the development of a variety of other organs during mammalian development.