Downloaded from genesdev.cshlp.org on October 1, 2021 - Published by Cold Spring Harbor Laboratory Press A leucine zipper domain of the. suppressor of Hmry wing protein mediates its repressive effect on enhancer function Douglas A. Harrison, 1"3 David A. Gdula, ~ Robert S. Coyne, 2"4 and Victor G. Corces l's 1Department of Biology, The Johns Hopkins University, Baltimore, Maryland 21218 USA; 2Department of Biochemistry and Molecular Biology, Harvard University, Cambridge, Massachusetts 02138 USA The suppressor of Hairy-wing [su(Hw)] protein mediates the mutagenic effect of the gypsy retrotransposon by repressing the function of transcriptional enhancers controlling the expression of the mutant gene. A structural and functional analysis of su(Hw) was carried out to identify domains of the protein responsible for its negative effect on enhancer action. Sequence comparison among the su(Hw) proteins from three different species allows the identification of evolutionarily conserved domains with possible functional significance. An acidic domain located in the carboxy-terminal end of the Drosophila melanogaster protein is not present in su(Hw) from other species, suggesting a nonessential role for this part of the protein. A second acidic domain located in the amino-terminal region of su(Hw) is present in all species analyzed. This domain is dispensable in the D. melanogaster protein when the carboxy-terminal acidic domain is present, but the protein is nonfunctional when both regions are simultaneously deleted. Mutations in the zinc fingers result in su(Hw) protein unable to interact with DNA in vivo, indicating a functional role for this region of the protein in DNA binding. Finally, a region of su(Hw) homologous to the leucine zipper motif is necessary for the negative effect of this protein on enhancer function, suggesting that su(Hw) might exert this effect by interacting, directly or indirectly, with transcription factors bound to these enhancers. [Key Words: Leucine zipper domain; su(Hw); transcriptional enhancer function] Received May 13, 1993; revised version accepted July 28, 1993. Insertion of the gypsy retrotransposon into various controlled by tissue-specific transcriptional enhancers Drosophila genes results in mutations with phenotypes located in the intron and/or the 5' region of the respec- that can be reversed by second site mutations in the tive locus {Geyer and Corces 1987; Liu et al. 1991). In suppressor of Hairy-wing [su(Hw)] gene (Modolell et al. both cases, insertion of the gypsy element interferes 19831. This finding suggests a direct involvement of the with the expression of the gene in those tissues regulated su(Hw) protein in the generation of mutant phenotypes by enhancers located distally from the gypsy insertion by gypsy, because the lack of a functional su(Hw) protein site with respect to the promoter {lack et al. 1991; Geyer results in a reversion of the gypsy-induced phenotype. and Corces 1992}. In the case of yellow, the phenotypic su(Hw) is a zinc finger protein that binds to a specific effect of gypsy can be reproduced when the su(Hw)-bind- sequence, similar to the octamer motif, located in the ing sequences are present in the original gypsy insertion 5'-transcribed untranslated region of gypsy [Spana et al. site, suggesting that the su(Hw) protein alone is respon- 1988; Dorsett 1990; Spana and Corces 1990}. The neces- sible for the induction of the mutant phenotype (Spana sary and sufficient requirement of su(Hw) protein for and Corces 1990; Geyer and Corces 1992). This negative gypsy mutagenesis has been demonstrated in the case of effect of su(Hw) on transcription is not enhancer spe- hspTO, yellow, and cut alleles induced by this retrotrans- cific, because insertion of the su(Hw)-binding site in dif- poson (Holdridge and Dorsett 1991; lack et al. 1991; ferent regions of the yeIlow gene is able to inhibit the Geyer and Corces 1992; Smith and Corces 1992). The function of any enhancer located distally from the temporal and spatial expression of the two latter genes is su(Hw)-binding region with respect to the yellow pro- moter {Geyer and Corces 1992}. Furthermore, the pres- ence of the su(Hw)-binding region flanking the white Present addresses: 3Department of Genetics, Harvard Medical School, gene can buffer white expression from position effects of Boston, Massachusetts 02115 USA; 4Fred Hutchinson Cancer Research Center, Seattle, Washington 98104 USA. adjacent sequences independent of the location of the SCorresponding author. white gene in the genome (Roseman et al. 1993}. The 1966 GENES& DEVELOPMENT 7:1966-1978 9 1993by Cold Spring Harbor Laboratory Press ISSN 0890-9369/93 $5.00 Downloaded from genesdev.cshlp.org on October 1, 2021 - Published by Cold Spring Harbor Laboratory Press Functional domains of su(Hw) su(Hw) protein can also interfere with regulatory se- product. The only phenotypic effect of null mutations in quences located in genes for which gypsy-induced muta- the su(Hw) gene is female sterility. Females carrying the tions have not been found. For example, the insertion of su(Hw) v null allele (Harrison et al. 1992) are unable to the su(Hw)-binding region in the 5' end of the hsp70 lay eggs because the egg chambers degenerate before gene, between the heat shock element and the promoter, completion of oogenesis. Egg chambers from su(Hw) v interferes with proper heat shock induction of transcrip- females appear less regular in shape and spaced more tion (Holdridge and Dorsett 1991). closely on the ovarioles than wild-type chambers (data These results indicate that the su(Hw)-mediated mu- not shown). In stages 7, 8, and 9, nurse cells of su(Hw) tagenic effect of gypsy is attributable to an inhibitory mutants can often be seen to shrink away from one an- effect of the su(Hw) protein on enhancer elements that other, leaving gaps in the anterior portions of the egg control the expression of the mutated gene. The lack of chambers. In addition, wild-type oocytes begin accumu- specificity in the nature of the affected enhancers and lating yolk at stage 8. The stage 8 and 9 egg chambers of the directionality of the inhibitory effect suggest a few su(Hw) mutant females show a considerable reduction altematives to explain the mechanism underlying this in deposition of yolk in the oocyte. Egg chambers of the phenomenon. One possibility is that binding of the mutant cease to grow and eventually degenerate before su(Hw) protein causes changes in the conformation of stage 10. DAPI staining shows that wild-type ovaries the adjacent chromatin that spread distally with respect contain nurse cells with homogeneously staining chro- to the promoter and interfere with the binding of tran- mosomes within the nucleus after very early stages of scription factors to distal enhancers. The requirement oogenesis (King et al. 1956). The DNA appears to be de- for a directional (away from the promoter) spreading of condensed and spread throughout the nucleus (Fig. 1A). chromatin changes can be explained if binding of su(Hw) In su(Hw) mutant nurse cells, the chromosomes look to DNA creates boundaries between higher order do- normal until stage 3 or 4, but afterward become aggre- mains of gene activity (Roseman et al. 1993). A second gated and condensed (Fig. 1B). The bulbous nature of the possibility is that the inhibitory effect of su(Hw) on en- chromosomes from su(Hw) mutants is reminiscent of hancer function is attributable to the interaction of this that seen fleetingly in normal egg chambers only at stage protein with enhancer-bound transcription factors as 4. It seems that su(Hw) v nurse cell chromosomes gener- they try to reach the transcription complex by tracking ally retain the bulbous morphology throughout later along the DNA or looping the intervening sequences stages of oogenesis, suggesting that the su(Hw) gene (Geyer and Corces 1992). This model explains the direc- product might be required for decondensation of the tionality of the su(Hw) effect better than that based on nurse cell chromosomes. As su(Hw) is a DNA-binding chromatin changes, but the direct contact of su(Hw) protein capable of interacting with many sites through- with transcription factors might require a degree of spec- out the genome (Spana et al. 1988), it could be hypothe- ificity in the interaction that is at odds with the apparent sized that the binding of su(Hw) directly to these bul- universal effect of su(Hw) on all enhancers tested. bous chromosomes is responsible for decondensation. To further the understanding of the mechanism by The chromosomes of follicle cells appear to be unaf- which su(Hw) affects enhancer function, we have sought fected in the mutant. to determine whether specific domains of su(Hw) are An additional defect seen in some of the su(Hw) alleles involved in the negative effect of this protein on tissue- is that many egg chambers contain more than the nor- specific transcription. Here, we report a functional study mal 15 nurse cells. This defect is likely to be the result of of the su(Hw) protein carried out by analysis of extant the fusion of egg chambers rather than the overprolifer- su(Hw) mutants, as well as new mutations induced in ation of nurse cells, because most of the aberrant cham- vitro and introduced into the fly by P element-mediated bers contain -30 nurse cells, twice the normal number. transformation. These studies suggest that an extended In addition, the anterior nurse cells are often smaller amphipathic s-helix of the su(Hw) protein with struc- than the posterior, as might be expected because the tural characteristics similar to the leucine zipper motif younger chambers are located more anteriorly in the ova- might be responsible for the inhibitory effect of su(Hw) riole.