Cell, Vol. 64, 169-200, January 11, 1991, Copyright 0 1991 by Cell Press The Recombination Activating Gene-l (RAG-I) Transcript Is Present in the Murine Central Nervous System Jet-old J. M. Chun: David G. Schatr: Marjorie A. Oettinger,’ Rudolf Jaenisch: and David Baltimore*t l Whitehead Institute for Biomedical Research Nine Cambridge Center Cambridge, Massachusetts 02142 and Department of Biology Massachusetts Institute of Technology Cambridge, Massachusetts 02139 tThe Rockefeller University 1230 York Avenue New York, NY 10021 Summary The recombination activrting genes, RAG-1 and RAG-2, are likely to encode components of the V(D)J site- specific recombination machinery. We report here the detection of low levels of the RAG-1 transcript in the murlne central nervous system by polymerase chain reaction, in situ hybridization, and Northern blot anal- yses. In conWas& an authentic RAG-2 transcript could not be detected reproducibly in the central nervous system. The RAG-1 transcript was found to be wide- spread in embryonic and postnatal neurons, with tmn- scription being most apparent in regions of the post- natal brain with a high neuronal cell density (the cerebellum and the hippocampal formation). The m- suits suggest that RAG-l functions in neurons, where its role might be to recombine elements of the neu- ronal genome site-specifically, or to prevent detri- mental alterations of the genome in these long-lived cells. Introduction Site-specific recombination plays a critical role in the life cycle of a variety of organisms, where it is used to mediate genetic mobility and gene assembly and to regulate gene expression and plasmid copy number. In vertebrates, the only site-specific recombination system that has been identified functions in developing lymphocytes to assem- ble the variable portions of the genes that encode immu- noglobulin and T cell receptor molecules using compo- nent V (variable), D (diversity), and J (joining) gene segments. V(D)J recombination is largely responsible for generating the highly diverse array of immunoglobulins and T cell receptors that is characteristic of the vertebrate immune system. While the enzymatic machinery that carries out V(D)J recombination is poorly understood, recent findings sug- gest that it consists of components whose expression is limited to recombinationally active ceils, as well as an ac- cessory factor(s) whose expression is more widespread. We recently reported the isolation of two recombination activating genes, RAG-1 and RAG-2, whose expression activates V(D)J recombination in fibroblastoid cells (Oet- tinger et al., 1990; Schatz et al., 1989). Using Northern blot hybridization methods, RAG-1 and RAG-2 expression was found only in precursor B and T cells and not in cells from other stages of lymphoid development or in nonlym- phoid tissues or cell lines examined; expression of these genes thus correlated precisely with expression of V(D)J recombination activity (Lieber et al., 1987). The two genes are immediately adjacent to one another in the mouse ge- nome and activate V(D)J recombination synergistically (Oettinger et al., 1990). RAG-1 and RAG-2 are likely to en- code developmental stage-specific components of the recombinational machinery (see below). The one acces- sory factor thus far identified that is likely to be essential for V(D)J recombination is the product of the gene that is mutant in the severe combined immunodeficient (SC@ mouse (Bosma et al., 1983). The scid mutation severely disrupts V(D)J recombination (for review, see Schiiler, 1990), and the gene appears to be ubiquitously expressed since nonlymphoid scid cell lines and tissues are radia- tion hypersensitive (Fulop and Phillips, 1990; E. Hendrick- son, X. Qin, E. Bump, and D. Weaver, unpublished data). A number of observations had previously led us to sug- gest that RAG-1 and RAG-2 encode components of the V(D)J recombinational machinery rather than indirectly activating other factors required for recombination (Oet- tinger et al., 1990; Schatz et al., 1989). First, RAG-l and RAG-2 are sufficient to activate the V(D)J recombinational machinery in fibroblastoid cells, and they appear to do so specifically, since other lymphoid-specific properties have not been detected in fibroblasts expressing the two genes. Second, their coexpression correlates precisely with V(D)J recombination. Finally, both genes have been con- served through vertebrate evolution, as have the cis- acting DNA elements required for V(D)J recombination. Two recent observations strengthen this argument and provide connections between RAG-1 and RAG-2 and other recombinational processes. First, the carboxy-ter- minal portion of the RAG-1 protein has homology with the product of the yeast gene HfFll (Wang el al., 1990), which itself has homology to yeast topoisomerase I and func- tions at least in part to suppress intrachromosomal exci- sion recombination (Aguilera and Klein, 1990). Second, RAG-2 expression in the chicken bursa and in bursal cell lines correlates exactly with the gene conversion process that diversifies chicken immunoglobulin variable regions (Carlson el al., 1991; see accompanying article). The superficial similarities between the immune system and the central nervous system (CNS)-extreme com- plexity, the capacity for memory, and extensive develop- mental cell death-have been extended by recent results demonstrating numerous molecular interrelationships be- tween the two systems. A variety of polypeptide signaling molecules, transcription factors, cell surface antigens, and receptors are common to both the immune and ner- vous systems, and an ever-increasing number of immuno- globulin supergene family members are being discovered