INTRODUCTION Many genes important for early development have first been identified in Drosophila due to the relative ease of mutational analysis. Segmentation of the Drosophila embryo is one of the earliest morphological manifestations of regional tissue specialization. A cascade of gene products has been identified that is responsible for the proper regulation of this process (for reviews see Scott and Carroll, 1987; Kornberg and Tabata, 1993). These genes are activated at several hierarchical levels, starting with maternal genes regulating the expression of the gap genes which in turn control the pair-rule genes that act on the segment polarity genes. Finally, the homeotic genes supply positional information and determine segment identity. The discovery of these gene classes in Drosophila has led to a breakthrough in vertebrate developmental biology by the realization that many of them have their vertebrate counterparts with a remarkable correspondence in structure and function. A most striking finding was the collinear and homologous relationship between the Drosophila HOM-C and the vertebrate Hox homeobox gene complexes (Krumlauf, 1992). The paired genes are also well conserved between Drosophila and vertebrates (Noll, 1993). A particularly striking example of this gene class is pax-6, a master gene which controls Drosophila, mouse and human eye development (Quiring et al., 1994). Finally the segment polarity genes hedgehog (Ingham, 1994) and wingless and their signaling mechanisms (Orsulic and Peifer, 1996) are now known to be of fundamental importance for the regulation of both invertebrate and vertebrate development. The Drosophila gene ten-m has been discovered in two laboratories by different experimental approaches. Baumgartner et al. (1994) searched for Drosophila homologues of the vertebrate extracellular matrix proteins tenascins (for review see Chiquet-Ehrismann, 1995), whereas Levine et al. (1994) cloned a partial cDNA for a protein named odd oz which was recognized by a monoclonal antibody raised against proteins phosphorylated on tyrosines. Baumgartner et al. (1994) showed that the Ten-m protein is expressed in seven stripes during the blastoderm stage in early embryos before gastrulation. Furthermore, Ten-m could be immunoprecipitated from the conditioned medium of Schneider cells and thus occurs as a secreted protein (Baumgartner et al., 1994). Although the interpretation of the protein sequences obtained in the two laboratories were quite contradictory, the tissue distribution of the protein and the transcripts during Drosophila development were consistent and in both cases mutations were identified to result in a pair-rule phenotype. This is highly 2019 Journal of Cell Science 112, 2019-2032 (1999) Printed in Great Britain © The Company of Biologists Limited 1999 JCS0332 The Drosophila gene ten-m is the first pair-rule gene not encoding a transcription factor, but an extracellular protein. We have characterized a highly conserved chicken homologue that we call teneurin-1. The C-terminal part harbors 26 repetitive sequence motifs termed YD-repeats. The YD-repeats are most similar to the core of the rhs elements of Escherichia coli. Related repeats in toxin A of Clostridium difficile are known to bind specific carbohydrates. We show that recombinantly expressed proteins containing the YD-repeats of teneurin-1 bind to heparin. Furthermore, heparin lyase treatment of extracts of cells expressing recombinant YD-repeat protein releases this protein from high molecular mass aggregates. In situ hybridization and immunostaining reveals teneurin-1 expression in neurons of the developing visual system of chicken and Drosophila. This phylogenetic conservation of neuronal expression from flies to birds implies fundamental roles for teneurin-1 in neurogenesis. This is supported by the neurite outgrowth occurring on substrates made of recombinant YD-repeat proteins, which can be inhibited by heparin. Database searches resulted in the identification of ESTs encoding at least three further members of the teneurin family of proteins. Furthermore, the human teneurin-1 gene could be identified on chromosome Xq24/25, a region implied in an X-linked mental retardation syndrome. Key words: Teneurin-1, Development, Nervous system, Carbohydrate-binding, Heparin SUMMARY Teneurin-1, a vertebrate homologue of the Drosophila pair-rule gene Ten-m, is a neuronal protein with a novel type of heparin-binding domain Ariane D. Minet 1 , Beatrix P. Rubin 1 , Richard P. Tucker 2 , Stefan Baumgartner 1, * and Ruth Chiquet-Ehrismann 1,‡ 1 Friedrich Miescher Institute, PO Box 2543, CH-4002 Basel, Switzerland 2 Department of Cell Biology and Human Anatomy, School of Medicine, University of California, Davis, CA 95616-8643, USA *Present address: Department of Cell and Molecular Biology, Section of Developmental Biology, University of Lund, Box 94, S-22100 Lund, Sweden ‡ Author for correspondence (e-mail: chiquet@fmi.ch) Accepted 22 April; published on WWW 26 May 1999