0 INSTITUT PASTEURIELSEVIER Paris 1996 Res. Immunol. 1996, 147, 261-266 Why is CTX all the RAGE? L. Du Pasquier (*) and I. Chretien Base1 Institute for Immunology, PO Box, CH 4005, Base1 (Switzerland) From several contributions to the Forum published in this issue (Medzhitov and Janeway, Ohno, Makody), it seems that the origin of T-cell receptors (TCR) and immunoglobulins (Ig) of vertebrates would be best explained by the exis- tence of a primitive ancestral receptor probably resembling or being a non-dimeric adhesion molecule. Made of Ig domains without somatic rearrangement and therefore not clonally expressed, this molecule would transduce, through its own cytoplasmic tail, some signals to the inside of a primitive lymphocyte. Other contributions (Litman and Rast, Du Pasquier and Chretien) point out candidates that would par- tially fulfill these criteria. We proposed that CTX (for cortical thymocyte marker in Xenopus) might resemble the early lymphocyte receptor, but indeed it seems to be already a differentiated type of receptor. Its restricted tissue distribution is certainly a sign of specialization, and being part of a dimer, which CTX is likely to be (Robert er uZ., submitted), is already an advanced character. For more primitive genes with ances- tral characteristics, data base search did not yield many other candidates. We therefore took a dif- ferent approach, neglecting primary sequence homology at least in a first step. We recently found that CTX segregates with the Xenopus major histocompatibility complex (MHC) in crosses (unpublished). We then looked to see whether other Ig superfamily (IgSF) members had been reported to segregate with MHC. B-G (Kaufman et al., 1991) (already mentioned in our paper), MOG (the myelin/oligodendrocyte glycoprotein of Gardi- nier and Matthieu, 1995), and butyrophilin (Vernet et al., 1993) are all obviously homolo- gous in the V region of CTX, but they are otherwise very different. RAGE (receptor for advanced glycosylation end products) (Neeper et al., 1992) is a member of the IgSF; its appa- rently monomeric molecule consists of one V, domain (according to our nomenclature) and two C, domains. Although it resembles the molecule named “Amalgam” from Drosophila (already picked up as homologous to CTX, see our contribution), RAGE’s external domains, unlike Amalgam, are associated with a trans- membrane region and a long cytoplasmic tail. RAGE is also known in rat and cattle and is located within the MHC (3’ to DR), at least in human beings (Sugaya et al., 1994). Aside from the highly conserved V and C2 specific resi- dues, the homology of RAGE with CTX is poor but its exon-intron organization bears striking similarities to CTX. Its V domain is encoded by 2 half-domain exons (fig. 1). RAGE Vb and CTX Vb have the same length, glycosylation sites are in the same position, and splicing bet- ween Va and Vb is, in both cases, of type 0. The splicing between transmembrane and cytoplas- mic domains is of type 2 (rarely encountered in Submitted May 17, 1996, accepted May 30, 1996. (*) For correspondence.