Functional tagging of regulatory elements in the plant genome

JENNIFER F. TOPPING 0 WENBIN WEI 0 KEITH LINDSEY 0 0 Leicester Biocentre, University of Leicester , Leicester LEI 7RH , UK *To whom correspondence should be addressed - In comparison with animals, relatively few plant genes have been identified that have been shown to be under organ-, tissue- or cell-type-specific regulation. In this paper, we describe how the /J-glucuronidase (GUS) reporter gene (gusA or uidA), fused to a weak promoter (a truncated ( - 9 0 bp) CaMV35S promoter), can be used to identify tissue-specific markers in transgenic tobacco plants. The rationale was that the expression of gusA would be determined primarily by position effect. Quantitative analysis revealed that, of 184 90-gus transgenic plants, 73 % exhibited gusA gene activation in leaf tissue, and the level of GUS enzyme activity varied over a 300-fold range within the population. In comparison, transformation with a promoterless gusA gene resulted in GUS expression in 78 % of all plants analyzed (in leaf and/or root) and expression levels were The advances that have been made in techniques for the stable transformation of plants have created new opportunities for investigating the regulatory sequences of specific genes (Kuhlemeier et al. 1987) and the control of plant differentiation and development (Goldberg, 1988). Significant future progress in each of these areas in turn requires further advances in the identification of new plant genes, and an improved understanding of their regulation in situ and the factors that influence the stable expression of foreign genes when introduced into the plant genome. A direct approach to the detection of cell- and tissuespecific enhancer sequences, and the generation of visual markers of development, has been described recently for Drosophila (O'Kane and Gehring, 1987; Bellen et al. 1989) and mouse (Allen et al. 1988). The reporter gene lacZ was stably introduced under the transcriptional control of a weak promoter, and the qualitative and quantitative expression of the gene was determined predominantly by the activity of nearby native enhancer sequences associated with specific genes (Wilson et al. 1989). Transgenic lines were generated which exhibited unique and restricted patthree-fold or more lower. Qualitative GUS analysis of single locus 90-gus transformants revealed differential expression in diverse tissues. The spatial pattern of GUS activity was unique to individual transformants, was a reflection of differential gusA gene transcription, and was stably transmissible to progeny. Evidence for preferential expression in roots not only of the -90-gus, but also the promoterless gusA gene is presented. The value of the -90 bp promoter-gusA sequence, which is termed an 'interposon', as a tool both to identify native enhancer sequences in situ and to investigate position effects in plants, is discussed. terns of /J-galactosidase activity, reflecting the regulatory properties of nearby sequences. In view of the non-homologous recombination of foreign genes with the genomes of plants and animals (Wallroth etal. 1986), it would be expected that a range of enhancer and other sequences could be detected, by virtue of the apparently random site of insertion of the 'detector sequences'. The approach also provides a means of investigating the effects of the local chromatin environment on foreign gene expression, for which very little information is available in plants. In animal systems such effects have been described, particularly for genes under the transcriptional control of weak promoters (Palmiter and Brinster, 1986). The identification of factors that determine these so-called 'position effects' is a desirable objective both to identify regulatory elements that can be exploited as cell-typespecific markers and ultimately to increase experimental control over foreign gene expression. In this paper, we describe a strategy to identify tissuespecific markers and to investigate the mechanisms that determine position effects in plants. The rationale is to introduce into a population of plants a reporter gene (gusA, uidA encoding ^-glucuronidase) fused to a weak promoter (the -90 bp region of the Cauliflower Mosaic Virus (CaMV) 35S RNA gene promoter), such that its expression is determined primarily by its local chromatin environment. Transgenic plants can be screened for quantitative and qualitative aspects of gus expression, and the observed pattern therefore represents a functional 'tag' of a regulatory region close to or at the site of integration. The inserted sequence, which we term an 'interposon', is expected to be physically linked to the putative regulatory sequences influencing its expression, and will facilitate the isolation and cloning of those sequences in a way analagous to transposon tagging (Doring and Starlinger, 1986). However, the generation of a visual change in plant phenotype is not a prerequisite for mutant gene isolation. The - 9 0 bp region of the CaMV35S promoter contains the CAATand TATA boxes but excludes some of the upstream transcription enhancer sequences (Fang et al. 1989). This truncated promoter has been used previously to detect enhancer elements within the upstream regions of specific, isolated genes by linking 5' flanking sequences to the 90CaMV35S promoter which in turn was linked to a reporter (e.g. Fluhr et al. 1986; Nagy et al. 1987), and also as part of an enhancer cloning vehicle (Ott et al. 1990; Ott and Chua, 1990). Fang et al. (1989) have demonstrated that the -90 bp to 46 bp region of the promoter by itself has little activity, but plays an accessory role in modulating the transcriptional activity of two upstream regions of the CaMV35S promoter (-343 bp to -208bp and -208bp to -90 bp) which themselves can enhance the activities of other promoters such as the nopaline synthase (nos) promoter of Agrobacterium tumefaciens (Odell et al. 1988). Poulsen and Chua (1988) and Benfey et al. (1989) have provided evidence that the -90CaMV35S promoter directs gene expression predominantly in the roots of transgenic plants. This is probably due to binding the transcriptional activator ASF-1 (Katagiri et al. 1989). Our results demonstrate that this minimal promoter can be activated in situ in diverse tissues in a quantitatively and qualitatively distinct manner in individual lines of transgenic tobacco, and the expression pattern is inherited by the Fj (T2) progeny. The implications for the facilitated cloning of tissuespecific genes, enhancers and other regulatory sequences, and for the study of position effect mechanisms, are discussed. Materials and methods Plant material Tobacco shoot cultures and transformation Shoot cultures of tobacco (Nicotiana tabacum cv. Petit Havana SRI) were maintained on agar-solidified Murashige and Skoog (1962) medium (MS; Sigma) supplemented with 30 g I"1 sucrose (MS30), as described by Topping and Lindsey (1991). These provided leaf material for Agrobacterium tumefaciens-mediated tr (...truncated)