Proc. Nati. Acad. Sci. USA Vol. 91, pp. 5647-5651, June 1994 Biochemistry Targeted mutation of the CREB : Compensation within the CREB/ATF family of transcription factors (CREM/gene targetIn/homogous recombination) EDITH HUMMLER*t, TIMOTHY J. COLE*, JULIE A. BLENDY*, RUTH GANSS*, ADRIANO AGUZZII, WOLFGANG SCHMID*, FRIEDRICH BEERMANN§, AND GONTHER SCHOTZ*¶ *Division of Molecular Biology of the Cell I, German Cancer Research Center, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany; and tInstitute of Neuropathology, Schmelzbergstrasse 12, CH-8091, Zfrich, Switzerland Communicated by Wolfgang Beermann, March 3, 1994

ABSTRACT The cAMP response element bnding of the CREB gene by homologous recombination in mouse (CREB) has been implicated as a key regulator in the tran- embryonic stem (ES) cells. scriptional control of many . To assess the functional importance of CREB in vivo and its role in development, we MATERIALS AND METHODS used gene targeting to generate mice with a disruption of the CREB gene. Homozygous mutant mice appeared healthy and Gene Targeting. The targeting vector consisted of7.1 kb of exhibited no impairment of growth or deveiopment. In this mouse CREB genomic DNA and a promoterless neomycin- report we demonstrate that CREB and two other members of resistance (Neo) gene inserted in-frame in exon 2 of the the CREB/ATF family, cAMP response element modulation mouse CREB gene via a unique Nco I site (ref. 18; Fig. 1A). protein (CREM) and activating factor 1 (ATF1), The construct (20 jg) was linearized with Not I and used to trnscription electroporate 1 x 107 D3 ES cells derived from 129Sv/J mice, appear to form a unique subgroup within this extensive class of which were cultured on mitomycin-treated embryonic fibro- transcription factors. Examination of CREM mRNA and pro- blast feeder layers (19). DNA from clones surviving G418 tein levels In CREB mutant mice demonstrated overexpresslon selection (200 .g/ml) were individually analyzed on Southern ofCREM in all tissues examined, but no change in ATF1 levels. blots and hybridized with probes located either 5' or 3' ofthe These data demonstrate that CREB is not the sole mediator of genomic sequence contained in the construct. Depending on cAMP-dependent transcriptional regulation and probably acts the probe used blots were prepared by digestion with Pvu II in concert with a specific subset of cAMP response dement- (3' probe) or Nco I (5' probe). bnding to transduce the cAMP signal and, in its Generation of CREB -/- Mice. ES cells from two clones absence, these same proteins can compensate for CREB func- were used for injection into blastocysts derived from tion in vivo. C57BL/6J mice. Blastocysts were transferred to pseudopreg- nant NMRI/Han females and chimeric offspring were de- Transcription of many genes is affected by changes in cAMP tected by the presence of agouti hairs (genotype Aw) on a levels in response to a variety of external signals and is nonagouti (a) background. Chimeric males were mated to mediated via a cAMP response element (CRE). This DNA females to produce ES-cell-derived offspring that were then sequence is recognized by a diverse family of DNA binding analyzed on Southern blots containing DNA isolated from which the has mouse tails (20). Mice heterozygous for the gene-targeting proteins (1-5), of CRE-binding protein (CREB) event were then used to generate homozygous mutant CREB been best characterized (6-12). Activation of the protein -/- mice. kinase A (PKA) pathway leads to phosphorylation of CREB RNA Analysis. Total RNA was isolated and prepared from at Ser 33, which is required for CREB to initiate transcription tissues ofwild-type (+/+), heterozygous (+/-), or homozy- of target genes (6, 13). Since the cloning of CREB, a large gous (-/-) CREB mutant adult mice as described (21). A number of CRE-binding proteins have been identified. They 32P-labele