Beta1D is a skeletal muscle-specific splice variant of the beta1 integrin subunit, while beta1A integrin subunit has a wide tissue distribution. We have previously shown that replacement of beta1A by beta1D by homologous recombination (knockin) in all mouse tissues was embryonic lethal. Through two successive rounds of homologous recombination, we have now produced embryonic stem (ES) cells expressing beta1D instead of beta1A, and analyzed the ability of beta1D to support ES cell differentiation in vitro and in teratomas in vivo. Beta1D knockin (KI) ES cells grew at a similar rate but as more compact colonies than the beta1A-expressing cells. Increased cell cohesiveness, however, did not appear to involve changes in cadherin activity. Although in both beta1A and beta1D-KI ES cells only one beta1 allele is active; the expression of beta1 integrins in the beta1D-KI ES cells was reduced by 50%, compared with that in the beta1A-expressing cells; this correlated with impaired adhesive and migratory capacities. It appeared that during in vitro cardiac differentiation, in spite of a slight delay in the induction of two cardiac-specific transcripts, the alpha- and beta-myosin heavy chains, contracting cardiomyocytes were detected in similar numbers and at the same time in embryoid bodies (EB) derived from beta1D-KI and from beta1A cells. Furthermore, replacement of beta1A by beta1D in ES cells did not affect neurite differentiation in embryoid bodies in the presence of retinoic acid suggesting that beta1D supports neurogenesis. However, the impaired migration of other cells from the EB, including endodermal cells, prevented the normal outgrowth of neurites in beta1D-KI EB. Finally, injection of beta1D-KI ES cells in the flank of syngeneic mice gave rise to fully developed teratomas containing simple and pluristratified epithelia, muscle, cartilage, blood vessels, and tissues from the neural lineage. These results show that the muscle-specific splice variant beta1D, in spite of its specific cytoplasmic domain, supports the differentiation of many cell types. This further suggests that the embryonic lethality in the beta1D-KI embryos was mainly due to the different ability of beta1 A and beta1D to mediate cell adhesion and migration.