Conservation of brown Gene tmmhactivation in Drosophila Linda

The mechanism underlying transinactivation associated with dominant position effect variegation (PEV) of the Drosophila melanogaster brown gene has been addressed by a comparison with its D. uinlis homologue. This comparison revealed 86% identity between conceptual translation products of the h o r n gene from these two species, functional homology, as the D. virilis gene rescues a D. melanogaster null brm mutation, and conservation of the sequences required for transinactivation, as the D. uirilis gene in D. melanogasteris subject to dominant PEV. An extended region of sequence similarity upstream of the open reading frame is observed. As the D. Virilis homologue is functionally interchangeable with the D. melanogastergene, these genes must share regulatory sequences as well as protein coding homology. These results support a model in which transinactivation is mediated by a heterochromatin-sensitive transcription factor. P OSITION effect variegation (PEV) is a gene-silencing phenomenon in which a novel junction is created between euchromatin and heterochromatin resulting in the variable inactivation of genes located in the vicinity of this junction. The Drosophilu mhnogaster browz gene is one of many genes for which variegating alleles have been studied. brown PEV alleles are remarkable in comparison with other PEV mutations. Unlike almost all other variegating alleles tudied to date, variegating brown alleles are consistently dominant over wildtype alleles (MULLER 1930; GLASS 1933). That is, brown is affected both in cis and in trans to the heterochromatic junction. Like the cis effect on PEV alleles, the trans effect is sensitive to modification by suppressors and enhancers of PEV. At the molecular level, the trans effect is associated with reduced mRNA accumulation, or " transinactivation" of the wild-type gene (HENIKOFF and DREESEN 1989). transinactivation of brown (bw) can show remarkable strength; the eyes of bwD/bw+ flies have 298% brownommatidia (HENIKOFF and DREESEN 1989). The strength of gene silencing is lessened by disruptions in pairing between the variegating and wild-type brown alleles (DREESEN et al. 1991). Thus, two requirements for tran*inactivation are known: a novel euchromatinheterochromatin junction in the vicinity of brown and pairing between alleles. In addition, transinactivation has a third requirement-sequences located close to or within the trans copy of brm but not the cis copy. To explain this asymmetry, DREESEN et al. (1991) proposed that sequences regulating brown gene expression are responsible for transinactivation. These sequences Corresponding authoc Steven Henikoff, Fred Hutchinson Cancer Research Center A1-162, 1124 Columbia St., Seattle, WA 98104. E-mail: [email protected] Genetics 140: 193-199 (May, 1995) would correspond to a transcriptional enhancer, to which a heterochromatin-sensitive transcription factor binds. This notion was supported by deletion experiments that delineated this sensitive region to a 3.8-kb interval that includes the brm open reading frame, 1.2 kb of 5' flanking sequence, and no 3' nontranscribed sequence (MARTIN-MOWS et al. 1993). In no case was transinactivation eliminated without also eliminating gene expression. Thus the sequence element that confers sensitivity to heterochromatin in trans must reside within or just upstream of the brown gene. Because deletion of a necessary enhancer element would cripple the reporter of transinactivation, conclusive evidence for the model could not be obtained by further deletion mapping. Therefore, we have undertaken a species comparison to analyze the regulatory regions of the brm gene and test the ability of these sequences to bring about transinactivation. A species comparison is informative when the function of interest is conserved and when the evolutionary distance between species is great enough to allow for divergence of sequences that are not functionally constrained. Both requirements appear to be met by comparing the putative D. virilis brown cognate with D. melanogaster brm. The requirement for shared function is probably met because of the existence of a brown-eyed null mutation in D. virilis (MORI 1937) located in the chromosome arm homologous to D. mlunogaster 2R (ALEXANDER 1976). The requirement for sufficient divergence is met because the most recent common ancestor of D. virilis and D. mlunogasteris thought to have existed -60 mya (BEVERLEY and WILSON 1984). This evolutionary distance is sufficient for meaningful, noncoding sequence elements to be asily sifted from such alignments (e.g., BRAY and HIRSH 1986; MACDONALD 1990; WALLRATH and FRIEDMAN 1991; YAO and WHITE 1991). 194 L. E. Martin-Morris and S. Henikoff Analysis of this putative D. uirilis homologue is described in the context of the D. melanogarter genome. In so doing, we can ask whether there are cisacting sequences in the D. uirilis gene that can mediate tram inactivation in the presence of D. melanogaster tramacting factors. This might be the case if a D. melanogaster transcription factor is able to bind to the D. virilis gene and mediate trans-inactivation. Here we show that the brown gene cognate of D. uirilis is homologous to the D. melanogastergene in sequence and function. At the amino acid level, there is 86% identity between the proteins encoded by these two genes. Immediately 5' to the proteincoding region, we find stretches of nucleotide sequence conservation suggestive of regulatory function. We show that the D. virilis gene is capable of substituting for the D. melanogaster gene in transgenic assays for function. Finally, we report that a variegating D. virilis brown transgene is capable of trans-inactivating a paired D. virilis brown insertion. This indicates that trans-inactivation is mediated by a diffusible factor that can recognize the D. virilis brown cognate in D. melanogaster eye pigment cells. MATERIALS AND METHODS Isolation and characterization of D. Yiri1i.s brown homologue: D. uirilis lambda phage libraries from J. TAMKUN and R. BLACKMAN were screened using a 3.8-kb BglII fragment from the D. melanogaster brown gene as a probe. The BglII fragment corresponds with the entire coding region (exons 1-8) and 1.2-kb of 5' untranslated region. Plaques were screened by hybridization of probe (prepared by random priming) in 30% formamide, 5X SSPE (sodium chloride/ sodium phosphate/EDTA), 5X Denhardt's, 0.5% SDS, 100 pg/ml salmon sperm DNA at 42" (SAMBROOK et al. 1989). The final hybridization wash was at 55" in 1 X SSPE/O.l% SDS. Positive plaques were selected and purified by sequential replating. DNA was prepared using the Qiagen phage DNA kit. D. virilis SalI fragments were subcloned into the pVZl vector (HENIKOFF and ECHTEDARZADEH 1987). Sequence determination and analysis: Each SalI subclone that hybridized to the D. melanogaster probe was sequenced. ExonucleaseIII-generated deletions (CWUC and HENIKOFF 1993) were obtained using the Promega Erase-a-Base kit. Deletion derivatives were prepared as double-stranded DNA templates using Qiagen mini-prep columns. Sequencing reactions were performed using Sequenase (US Biochemicals) as described in MARTIN-MORRIS et al. (1993). The sequence was read from only the 5'-3' direction by this method. Following characterization of the gene organization, oligonucleotide primers were designed for sequencing of the reverse orientation of the upstream region. Sequence was entered using DNA editor software, and initial manipulations were performed with GENEPRO (Riverside Scientific). The protein alignment was modified from the output of the ESEE program (CABOT and BECKENBACH 1989). Analysis of sequence similarity was carried out by performing a search for significant nucleotide sequence alignments within defined genomic regions, and the analysis was visualized as a matrix plot (SCHWARTZ et al. 1991). Dyad analysis was performed with a program fromJoHN W. KELLER (personal communication) using default parameters. The genomic sequence of the D. uirilis gene is available from GenBank, accession No. L37035. Fly cultures and P-element transformation: All stocks were kept at room temperature on cornmeal/molasses food unless otherwise indicated. The mutations bwD, speck (sp), and scarkt (st) are described by LINDSLEY and GRELL (1968). To readily observe brown' pigment encoded on P transposons, the background genotype was bwD/+st (whiteeyed) unless otherwise indicated. Crosses for determining transinactivation were performed at 18", as lowered temperatures enhance PEV (SPOFFORD 1976). Other crosses were performed at 25". P-transposase induced transformation was performed as described by SPRADLINC (1986). A 9-kb D. uirilis fragment was cloned into the pDM24 vector (MISMER and RUBIN 1987) (P[uirilis-bw+]) and was injected into bw";st D. melanogaster embryos at a molar ratio of 3:l with "helper" transposaseencoding plasmid derived from P[A2-3(99B)] (ROBERTSON et al. 1988). Transformants were selected by their resistance to 300 pg/ml G418. G418 resistant flies also showed a brown+ phenotype. The P[virilis-bw+] transposon was mobilized to different sites in the D. melanogaster genome by first mating females bearing an X-linked p[uirilis-bw+] with P[A2-3(99B)] males (ROBERTSON et al. 1988). Transposase supplied by P[A23(99B)] catalyzed the transposition of the pIuirilzs-bw+] to other chromosomes in the germ lines of the F1 male progeny which were crossed to bwD;st females. F2 males with brown' eyes were selected. X-ray screen for PEV: Unmated red-eyed male flies homozygous for a D. uirilis bw+-transgene on chromosome 3 (bwD;stP[uirilisbw+]) were irradiated with 4000 rad and crossed to whiteeyed bwDsp;st females. F1 progeny (6000) were screened for variegated eyes. One female with variegated eyes was found and crossed to bupsp;st males. We determined that the variegating third chromosome insert was linked to the second and third chromosomes by segregation analysis using Cy0 and TM3 balancer chromosomes, respectively. After outcrossing, the bromvariegated phenotype always segregated away from either balancer. The variegating insert on this T(2;3) translocation chromosome is referred to as P[ uirilis-bw""7 . For analysis of PEV modification of the P [ uirilzsbw""7 allele, we chose a strain bearing a second chromosome suppressor mutation generated by TALBERT et al. (1994); Su(uar)2-S29D. Su(var)Z-S29D was classified as a general dominant suppressor as it suppressed two different brown variegating alleles as well as In(l)g4. We crossed &(war) bup +/+ bwD sp; st males by bwD sp;st P[uirilisbw"aT]/st females at 18" and compared the sibling flies with pigmented eyes from the progeny. Flies with an axillary speck ( sp) are thus Su(var)+, and speck+ flies are Su(uar). RESULTS AND DISCUSSION Cloning and sequence analysis of D. viri'lis brown: To isolate the D. virilis brown gene, we screened two lambda phage libraries prepared from D. virilis genomic DNA with the D. melanogaster brown gene at moderate stringency. Five candidate D. uirilis brown clones were isolated. SalI fragments from each overlapping phage clone were subcloned. Each D. virilis subclone hybridized to a subset of D. uirilis genomic restriction fragments that were detected by the D. melanogaster probe on genomic Southern blots (data not shown), indicating that each subclone probably represented part of the D. uirilis gene. As further analysis confirmed that we had identified the D. uirilis homologue of the D. trunsInactivation in Drosophila 195