Decoding positional information: regulation of the pair-rule gene hairy

KENNETH R. HOWARD 0 GARY STRUHL 0 0 Howard Hughes Medical Institute, Columbia University College of Physicians and Surgeons , 722 West 168th Street, New York, NY 10032 , USA - In the series of local gene activations that occur during early Drosophila development, the striped expression patterns of the pair-rule genes provide the first indication of segmental periodicity. The experiments that we report here address the question of how these patterns arise, by studying the regulation of one of these genes, hairy. We show that each of the seven stripes of hairy expression is controlled by a distinct subset of cisacting regulatory elements, some mediating transcriptional activation and others transcriptional repression. In general, elements necessary and sufficient for triggering a particular stripe response are clustered on the DNA and appear to overlap or be interspersed with During Drosophila oogenesis, the primary determinants of anteroposterior body pattern are deposited at the poles of the maturing egg (reviewed in NussleinVolhard et al. 1987). Following fertilization, these determinants trigger the expression of the gap gene products, including hunchback (hb), Kruppel (Kr) and knirps (kni), in a series of overlapping domains (Tautz, 1988; Nauber et al. 1988; Stanojevic et al. 1989; Gaul and Jackie, 1987, 1989; Pankratz et al. 1989; Fig. 5A). Subsequently, pair-rule genes are activated in repeating 'zebra stripe' patterns, which depend on the gap gene products (Carroll and Scott, 1986; Ingham et al. 1986; Frasch and Levine, 1987; Carroll et al. 1988). These observations together with analyses of partial loss- or gain-of-function mutations of the gap genes (Wieschaus et al. 1984; Lehmann and Nusslein-Volhard, 1987; Lehmann, 1988; Howard, 1988; Hulskamp et al. 1989; Struhl, 1989a,b) suggest that the overlapping distributions of gap gene proteins provide the spatial cues responsible for pair-rule gene activation. These findings raise the question of how pair-rule genes decode this spatial information into a periodic expression pattern. There are at least eight pair-rule genes and studies of their interactions suggest that only a few, the 'primary' pair-rule genes, respond directly to the positional elements involved in at least one other stripe response. Our results extend previous findings suggesting that periodic hairy expression arises by a decoding process in which each stripe is triggered by particular combinations or concentrations of regulatory factors. These regulatory factors are likely to include the products of the gap class of segmentation genes that are required for activating or positioning particular subsets of hairy stripes and are expressed with overlapping distributions during early embryogenesis. signals from the gap gene products (Carroll and Scott, 1986; Howard and Ingham, 1986; Frasch and Levine, 1987; Ingham and Gergen, 1988; Carroll et al. 1988). Deletion analyses of two of these primary pair-rule genes, hairy and eve, suggest that both contain large regulatory domains responsible for generating their periodic patterns of expression which can be subdivided into smaller regions necessary for triggering transcription in specific subsets of stripes (Howard et al. 1988; Goto et al. 1989; Harding et al. 1989; Pankratz et al. 1990). These findings are consistent with the simple hypothesis (Howard et al. 1988) that the control regions of both genes are modular, containing clusters of exacting regulatory elements which activate transcription in a single stripe when certain combinations or concentrations of particular gap gene products or maternal determinants are present. The fact that the three gap proteins hb, Kr and kni bind to parts of the control regions of these genes (Stanojevic et al. 1989; Pankratz et al. 1990; Rushlow, personal communication) suggest that this control may be achieved by the direct action of gap proteins. Here we use gene fusion analysis to assay systematically hairy for regulatory elements governing the individual stripe responses. Like the similar, although less exhaustive, analysis of Pankratz et al. (1990) we find that most of the individual' stripe responses are associated with discrete enhancer-like elements. Moreover, we have also been able to dissect some of these regulatory elements into two component parts: one that mediates transcriptional activation in the general vicinity of the stripe, and another that confers local repression of and serves to refine the broad response into a stripe. Thus, our results support the proposal that periodic hairy expression arises piecemeal and suggest that each stripe response reflects the integration of distinct activating and repressing functions operating directly at the level of DNA binding and transcriptional activation. Materials and methods Plasmid cloning and transformation were by standard techniques. After establishing individual lines all the flies containing a particular construct were pooled and embryos collected and stained. Constructs that showed staining at the blastoderm stage were identified and several lines were tested individually. Generally, we did not establish homozygous lines and assume that the embryos showing the strongest staining are homozygous for the construct in question. Fixation was in heptane phase partition with 4% formaldehyde in PEM buffer (100 mM Pipes, pH6.8; 2mM EGTA; lmM MgSO4). Embryos were devitellinised in 90% methanol, 50 mM EGTA and rinsed 5x in 100% methanol before being rehydrated in TBT (10 mM Tris pH8.0; 0.25 M NaCl; 0.1% Triton X-100). Immunochemical staining was performed using rabbit anti-beta-galactosidase antibodies (two lots of sera were used, one a generous gift from Paul Macdonald, another a commercial lot from Cappel). Antihairy sera was raised in rats using a full-length protein produced in bacteria using the T7 system (Studier and Moffaatt, 1986). Primary sera were preadsorbed to fixed embryos and used at final dilutions of 1:200 to 1:1000. Signal detection was either by immunofluorescence, in which case fluorochrome-conjugated secondary antisera from either Tago or Jackson were used at final dilutions of 1:200 to 1:400, or by immunochemistry in which case HRP conjugates of goat anti-rabbit (Bio Rad) or AP goat anti-rat (Jackson) were used at 1:2000. All specimens were examined in a Zeiss Axiophot Microscope. Double label fluorescent samples were analyzed using double exposure color, or pairs of single exposure black and white, photomicrographs on Fujichrome 400D at 400 ASA or Ilford XP1 at 400 ASA, respectively. Pairs of images were matched up by printing the entire frame including the borders which were then used to align the two. Immunochemical stains were recorded using Fujichrome 64T at 50 ASA or Technical Pan at 50 ASA developed with HC110. During cellularization of the blastoderm, the hairy gene is transcribed in a periodic pattern of seven evenly spaced stripes in the middle two thirds of the body (numbered 1- (...truncated)