Complex regulatory element within the gE- and gF-crystallin enhancers mediates Pax6 regulation and is required for induction by retinoic acid Jarmila Kra ´lova ´ a , Thomas Czerny b , Hana S ˇ panielova ´ c , Veronika Ratajova ´ a , Zbynek Kozmik a, * a Institute of Molecular Genetics, Flemingovo 2, 16637 Prague 6, Czech Republic b Institute of Animal Breeding and Genetics, University of Veterinary Medicine, Veterinaerplatz 1, A-1210 Vienna, Austria c Department of Genetics and Microbiology, Vinicna 5, 128 44 Prague 2, Czech Republic Received 6 August 2001; received in revised form 11 December 2001; accepted 21 January 2002 Received by A.J. van Wijnen Abstract The paired domain, DNA-binding domain of Pax6 and other Pax transcription factors, is composed of two subdomains (PAI and RED), each recognizing distinct half-sites of the bipartite binding site in adjacent major grooves of the DNA helix. The alternatively spliced Pax6(5a) isoform containing 14 extra amino acids within the PAI domain recognizes the 5aCON sequence consisting of four interdigitated 5 0 half-sites of the bipartite consensus sequence. A genome database search for similar tetrameric Pax6(A) recognition sequences led to the identification of a Pax6-binding site in the lens-specific enhancer of the mouse gE- and gF-crystallin genes. This binding site combines the properties of bipartite and tetrameric recognition sequences and, by mutational analysis, is shown to mediate Pax6-dependent regulation of the gE- and gF-crystallin promoter constructs both in primary chicken lens cells and in chicken embryo fibroblasts. The Pax6-binding site is adjacent to a previously identified retinoic acid response element and is itself required for retinoic acid induction of the gF- and gE-crystallin genes, suggesting that Pax proteins and retinoic acid receptors cooperate in transcriptional regulation. In summary, our protein–DNA binding and transactivation studies suggest that g-crystallin genes are under the control of a multifunctional enhancer element that mediates Pax6 regulation as well as retinoic acid-mediated induction. q 2002 Elsevier Science B.V. All rights reserved. Keywords: Paired box; DNA binding; Transcription; Retinoic acid receptor; Lens 1. Introduction Pax proteins are transcription factors defined by the presence of a highly conserved DNA-binding motif of 128 amino acids, the so-called paired domain (Bopp et al., 1986). The paired domain is composed of two subdomains (PAI and RED), each recognizing distinct half-sites of the bipartite binding site in adjacent major grooves of the DNA helix (Czerny et al., 1993; Jun and Desplan, 1996; Xu et al., 1999). Alternative splicing of Pax6 transcripts modifies the PAI domain by an insertion of 14 amino acids encoded by the additional exon 5a, thus generating the Pax6(5a) isoform. In addition, both human and mouse Pax8 are subject to an alternative splicing, which results in the inser- tion of a single amino acid (serine) in the Pax8 PAI domain. Both exon 5a (in Pax6) and serine (in Pax8) insertions have been shown to alter the DNA-binding specificity of the paired domain (Epstein et al., 1994; Kozmik et al., 1997), which now recognizes a new consensus sequence, 5aCON. The 5aCON sequence consists of four interdigitated 5 0 half- sites of the bipartite consensus sequence and is thus bound by four molecules via the intact RED domain (Kozmik et al., 1997). In the developing and adult chicken and mouse lens, Pax6 mRNA is found mainly in the epithelium (Li et al., 1994; Koroma et al., 1997). However, significant amounts of the Pax6 protein are also present in the primary fiber cells of the embryonic lens (Koroma et al., 1997; Duncan et al., 1998). As fiber cells differentiate, they begin to accumulate water- soluble proteins, named collectively as crystallins. In mammals, there are three major classes of crystallins (a, b and g), which provide the lens with its specific refractive Gene 286 (2002) 271–282 0378-1119/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved. PII: S0378-1119(02)00425-0 www.elsevier.com/locate/gene Abbreviations: PD, paired domain; PAI, N-terminal subdomain of paired domain; RED, C-terminal subdomain of paired domain; RA, retinoic acid; RAR, retinoic acid A receptor; RXR, retinoic acid X receptor; RARE, retinoic acid response element; Grg-4, a vertebrate Groucho family member; LSR, lens specific regulatory region; EMSA, electrophoretic mobility shift assay; CEF(s), chicken embryo fibroblasts; PLE, embryonic chicken primary lens epithelial cells; PCR, polymerase chain reaction; PBS, phosphate-buffered saline * Corresponding author. NIH, Building 6, Room 203, Bethesda, MD 20892, USA. Tel.: 11-301-496-3234; fax: 11-301-435-7678. E-mail address: kozmikz@nei.nih.gov (Z. Kozmik).