LETTERS During limb outgrowth, signaling by bone morphogenetic proteins (BMPs) must be moderated to maintain the signaling loop between the zone of polarizing activity (ZPA) and the apical ectodermal ridge (AER). Gremlin, an extracellular BMP antagonist, has been proposed to fulfill this function and therefore be important in limb patterning. We tested this model directly by mutating the mouse gene encoding gremlin (Cktsf1b1, herein called gremlin). In the mutant limb, the feedback loop between the ZPA and the AER is interrupted, resulting in abnormal skeletal pattern. We also show that the gremlin mutation is allelic to the limb deformity mutation (ld). Although BMPs and their antagonists have multiple roles in limb development, these experiments show that gremlin is the principal BMP antagonist required for early limb outgrowth and patterning. During development of the vertebrate limb, signals from the AER and the ZPA direct outgrowth and patterning of the limb skeletal elements 1–3 . The AER is a strip of columnar cells at the distal edge of the limb bud that secretes multiple fibroblast growth factors 1 Department of Molecular and Cell Biology, University of California-Berkeley, 401 Barker Hall, Berkeley, California 94720, USA. 2 Present addresses: Renovis, 270 Littlefield Avenue, South San Francisco, California 94080, USA (D.H.) and Department of Microbiology and Immunology, Stanford University, California 94305, USA (M.S.D.). 3 These authors contributed equally to this work. Correspondence should be addressed to R.M.H. (harland@socrates.berkeley.edu). Gremlin is the BMP antagonist required for maintenance of Shh and Fgf signals during limb patterning Mustafa K Khokha 1,3 , David Hsu 1–3 , Lisa J Brunet 1 , Marc S Dionne 1,2 & Richard M Harland 1 NATURE GENETICS VOLUME 34 | NUMBER 3 | JULY 2003 303 NheI lacZ neo DTA neo lacZ 5' Probe 3' Probe FRT FRT FRT FRT Targeting cassette Genomic region Null allele XbaI EcoRV XmnI EcoRV NheI XbaI NheI XmnI EcoRV gremlin XmnI EcoRV NheI EcoRV NheI XbaI XmnI EcoRV XmnI XmnI EcoRV/5' Probe XmnI/3' Probe 10 kb 4 kb 8 kb 6 kb 11 kb +/+ +/+ +/– +/+ +/– +/– ∆neo gremlin mRNA LacZ +/– +/– XmnI a b c 1 2 3 4 5 6 Figure 1 Targeting of the gremlin locus. (a) Map of the targeting vector, gremlin locus and the mutated allele produced by homologous recombination in ES cells. DTA, diptheria toxin cassette for negative selection; neo, the PGK- neo R cassette for positive selection flanked by FRT sites. An EcoRV–NheI fragment was used as a Southern-blot probe to confirm correct 5′ targeting. An EcoRV–XmnI fragment was used to confirm 3′ targeting. (b) Southern-blot analysis of genomic DNA to identify heterozygotes. The 5′ probe detects a 10-kb band in the wild-type allele (lanes 1,2) and a 4-kb fragment in the null allele (lane 3) when genomic DNA is digested with EcoRV. A small band cross-hybridizes and is detected in all lanes. For the 3′ end, the 3′ probe detects an 8-kb band in the wildtype allele (lane 5) and a 6-kb band in the null allele (lane 6) when genomic DNA is digested with XmnI. When the PGK-neo R cassette is removed by Flp recombinase, the 3′ probe detects a longer (11-kb) fragment (lane 4). (c) Expression of gremlin in the limbs, somites and flank of the embryo. LacZ staining of heterozygous embryos shows a similar pattern. Embryo heads were removed for genotyping. © 2003 Nature Publishing Group http://www.nature.com/naturegenetics