Cripto protein is a member of the ''EGF family'' of growth factors pre
sent in colon tumors and in human and mouse undifferentiated teratocar
cinoma cells. During gastrulation in the mouse, cripto-encoding transc
ripts are expressed in the forming mesoderm and later in the truncus a
rteriosus of the developing heart. As a necessary step prior to invest
igating the in vivo role of cripto through gene disruption, we have is
olated all the genomic cripto-related sequences in the mouse. One gene
(Tdgf1) and two pseudogenes (Tdgf2 and Tdgf3) have been isolated and
characterized. The mouse Tdgf1 (coding for cripto), like the human gen
e, is divided into six exons. Comparison of the human and mouse genomi
c sequences reveals that mouse exons 1 and 3 are shorter than the corr
esponding human exons. The pseudogene Tdgf2 corresponds to about 1 kb
of the mRNA and contains five base substitutions in the coding region
that represent both silent and replacement substitutions. The pseudoge
ne Tdgf3 corresponds only to the coding portion of Tdgf. Many mutation
s have been introduced in this pseudogene, suggesting its early origin
. Alignments of the Tdgf3, human and mouse mRNA sequences, shows that
this pseudogene has retained the 33 nucleotides of the human exon 3 th
at are missed in the Tdgf1 gene. Taken together, these data suggest th
at Tdgf3 is derived from an ancestral gene and that the human and mous
e genes are probably evolving separately.