Proc. Nati. Acad. Sci. USA Vol. 84, pp. 7987-7991, November 1987 Developmental Biology A short 5'-flanking DNA region is sufficient for developmentally correct expression of moth chorion genes in Drosophila (insect eggshell/P-element transformation/tissue specificity/gene regulatory elements/molecular evolution) S. ALEX MITSIALIS*, NIKOLAUS SPOEREL*, MICHAEL LEVITEN*, AND FOTIS C. KAFATOS*tt *Department of Cellular and Developmental Biology, Harvard University, 16 Divinity Avenue, Cambridge, MA 02138; and tInstitute of Molecular Biology and Biotechnology and Department of Biology, University of Crete, Heraklio, Crete, Greece 71110 Contributed by Fotis C. Kafatos, July 29, 1987

ABSTRACT Fusions with the bacterial gene for chloram- pSV2-cat (13), blunting the ends with T4 DNA polymerase, phenicol acetyltransferase followed by P-element-mediated and inserting the fragment into the Hpa I site in the polylinker germ-line transformation in Drosophila have permitted local- of pCar20 (14) through blunt-end ligation. The intergenic ization of the DNA sequence that confers a high degree of region of chorion gene pair A/B.L12 was isolated as a Bsm developmental specificity on a pair of silkmoth eggshell (chorion) I-Xmn I fragment from pB23 (12). The fragment was treated genes. The short, 272-base-pair, 5'-flanking region shared by with T4 DNA polymerase and blunt-end ligated into the the divergently transcribed genes is sufficient for developmen- CarCat-1 polylinker at a Sal I site that was filled in with DNA tally appropriate expression when placed upstream of the polymerase I. The resulting plasmids were named pAwt and chloramphenicol acetyltransferase gene, in either orientation. pBwt, depending on the orientation of the intergenic region A highly conserved motif within that region, TCACGT, is with respect to the chloramphenicol acetyltransferase (CAT) essential for chorion-specific expression. gene. In Vitro Mutagenesis. A 618-bp Sau3A-Fok I fragment from Our understanding of the mechanisms that regulate eukary- gene pair A/B.L12 (from position +78 to position +265 with otic gene expression has progressed rapidly, through in vitro respect to the A- and B-gene transcription initiation sites) was DNA manipulations coupled with in vivo transformation blunted and subcloned into the Sma I site ofM13mpl9 in both assays. Important cis-regulating elements have been defined orientations. Double-stranded replicative form of the appro- with some precision, many but not all of them in the 5'- priate M13 derivatives was linearized with Stu I (in the A- flanking DNA (e.g., refs. 1-3). For analysis of elements gene signal-peptide sequence) or Mlu I (in the B-gene intron). involved in developmental regulation, Drosophila melano- Two respective sets ofdeletion derivatives were then created gaster has proved a particularly favorable system, because of by digesting the linearized molecules with exonuclease BAL- the availability of a convenient germ-line transformation 31 for various periods of time, ligating the products to an procedure (4, 5) and a wealth of developmental, genetic, and 8-mer Bgl II linker (CAGATCTG), and recircularizing them. cytogenetic information. In addition to the investigation of The deletion derivatives of the intergenic region, AAT, AM1, promoter regions of individual Drosophila genes (6-9), stud- and AA2, were recovered as Bsm I-Bgl II fragments from ies have dealt with the control elements of genes that are M13 clones with the appropriate BAL-31 end points and organized in divergent, coordinately transcribed pairs (10). inserted into CarCat-1 as above. To generate linker substi- We have been studying the developmental regulation of tutions in the promoter region, matched end points offset by eggshell or chorion genes in the silkmoth Bombyx mori the length ofthe Bgl II linker (8 bp) were chosen from the two (reviewed in ref. 11). These genes are organized in divergent sets of deletions and combined by insertion into the plasmid pairs that are expressed coordinately, with very high sex, Bluescribe. The mutated promoter region in which a defined, tissue, and temporal specificity and that share an exceedingly short region had been substituted with the linker sequence short [300 ± 50 base pairs (bp)] common 5'-flanking DNA. CAGATCTG was then excised from the intermediate vector We have shown (12) that a pair of such genes, wholly with Bsm I and Xmn I and transferred into CarCat-1. contained within a 3.1-kilobase (kb) DNA fragment, is Generation of Transformant Lines. Germ-line transforma- normally regulated after it is integrated in the Drosophila tion ofDrosophila and selection, analysis, and propagation of genome, irrespective of the chromosomal integration site. independent transformant lines harboring single inserts were Thus, despite the evolutionary distance between these spe- performed essentially as described (12). cies, Drosophila trans-regulatory factors can recognize cor- RNA Analysis. Follicular RNA extraction and RNA gel blot rectly the cis-regulatory elements involved in the highly analysis using complementary RNA chorion probes were as precise developmental regulation of Bombyx chorion genes. reported (12). To generate a CAT-specific probe, a 0.66-kb We have pursued this trans-genic analysis and here report HindIII-Sca I fragment isolated from pSV2-cat was radiola- that regulatory elements sufficient for normal regulation beled using random primers and the Klenow fragment of reside within the shared 272-bp 5'-flanking (intergenic) DNA DNA polymerase I (15). of the gene pair. Within that DNA, an essential regulatory Extract Preparation. Drosophila tissues were homogenized element has been defined by introducing in vitro a small in 0.25 M Tris-HCI, pH 7.8/0.4 M sucrose/5 mM EDTA/0.5 cluster of base substitutions. mM leupeptin (TSEL). Use of TSEL improves the yield of CAT activity from transformed Drosophila tissues. Typical- ly, 10 ,ul ofTSEL was used per male, ovariectomized female, MATERIALS AND METHODS or pair of ovaries, and 3 ,ul of TSEL was used per follicle. Plasmid Construction. Construction ofthe CarCat-1 vector Homogenates were freeze-thawed twice and incubated for 10 involved isolating a 1.7-kb Bgl I-BamHI fragment from min at 65°C. The latter step removes from male and ovari- ectomized female extracts an activity (probably an acetyl- The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviation: CAT, chloramphenicol acetyltransferase. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 7987 Downloaded by guest on September 30, 2021 7988 Developmental Biology: Mitsialis et al. Proc. Natl. Acad. Sci. USA 84 (1987) CoA deacetylase) interfering with the CAT assay. Such an of protein content per homogenized, staged follicle (see also activity is not present in ovarian or follicular extracts. ref. 18). Extracts were centrifuged to pellet debris and denatured protein, and supernatants were assayed immediately or RESULTS AND DISCUSSION stored at -80'C until use. CAT Assays. A typical assay mixture contained 10-90 ,ul of Description ofthe Constructs. To analyze the cis-regulatory extract, 20 pl of freshly dissolved 4 mM acetyl-CoA, 2 ,Al of elements that permit developmentally correct expression of 0.1 mCi of [14C]chloramphenicol per ml (50-60 mCi/mmol; 1 Bombyx chorion genes in Drosophila, the bacterial gene for Ci = 37 GBq), 7.5 1.l of bovine serum albumin at 20 mg/ml, CAT was used as an expression marker. We constructed and 1.5 Al of 50 mM leupeptine. The final volume was vector CarCat-1 (Fig. 1), a hybrid derivative of pSV2-cat (13) adjusted to 150 pl with TSEL. Under these conditions CAT and the P-element transformation vector Carnegie 20 (14). In remains active for at least 15 hr, greatly improving the this construct, the 780-bp Tn9 sequence representing the sensitivity of the assay; standard incubations were for 5 hr at CAT structural gene is flanked by two simian virus 40 DNA 370C. Subsequently the reaction mixture was extracted with regions: a 69-bp segment derived from the 5'-untranslated ethyl acetate, and the extract was subjected to TLC on silica region of the early viral transcription unit and an 852-bp plates as described (16), separating unmodified chloramphen- segment that includes the tumor-antigen splice site and icol from its acetoxy derivatives. After exposure to x-ray polyadenylylation signals. DNA fragments to be assayed for film, radioactive spots of precursor and product were ex- promoter activity in Drosophila can be conveniently fused cised, and radioactivity was measured in a liquid scintillation upstream of this hybrid reporter gene through insertion into counter. CAT activity in each extract assayed was expressed a unique Sal I restriction endonuclease site. Further up- as the percentage of input chloramphenicol converted to the stream the vector contains the Drosophila rosy' gene that acetoxy forms during the standard incubation period. An serves as a transformation marker. The entire construct is activity standard (0.001 unit of commercial CAT enzyme, flanked by the P-element ends that permit transposition in the yielding 7.4 ± 1.4% conversion) was routinely included in presence of a helper plasmid that provides the transposase. each assay series. Values from different series ofassays were The origin of the Bombyx sequences analyzed here was compared after normalization with the activity standard. A clone pB23 that includes a 3.1-kb genomic DNA fragment minor correction for losses (typically, ±20%) was also made encompassing the divergently transcribed chorion gene pair by determining the protein content according to Bradford (17) A/B.L12 (19). We have shown (12) that when this clone is of each follicular extract and normalizing to a standard curve introduced into the Drosophila genome through P-element- A/B. L12 - I A -_: B B ser; ;es

A ' % I P-el ryt Ac/%8I~' ///CAT SV40 P-el a - /-- I, %V I I ,'I ATG TAA An

CarCat-1 I % I I I A A series B

gA LgctATTACTGCACAGCGGCAATCCTTTTATACGAATAAGTGTATTCCAATAAGGGCGATTACGACRAAAGTTT'(;AAT(GTAGAA LC CAGATCTGC > m4 TTCTGAAAAAAACTTATTTTTCCTTGATTGACACGTATCATTTACCTGTCTCCAAAACCCCCCCCACTACTACCCTAATCTCAATCAAACAAAATTTC AC AA

Li CAGATCTG GGTACAGTGATGTAGGTTAAAC TCACGT TTTTGAGATAGGACACTTGAGAAATATTCCGTATAAAACAGACACAACTGACATCGAACTca BE: AA2 AM1 A/\T

FIG. 1. (Upper) Description of constructs. The divergently transcribed chorion gene pairA/B.L12 is contained in the 3.1-kb B. mori genomic DNA fragment present in transformants of the B23 series (12). Exons, boxes (thick for coding and thin for untranslated regions); angled lines, introns; straight lines, 5'- and 3'-flanking sequences; arrows, direction of transcription. CarCat-1 is the vector we constructed for Drosophila transformation. Open boxes, P-element sequences; thin lines, ry+ gene; solid boxes, simian virus 40 sequences; hatched box, the CAT gene. The translational initiation and termination codons, the intron, and the polyadenylylation signals (An) are indicated. The 272-bp B. mori S'-flanking (intergenic) region was inserted in the polylinker ofCarCat-1 in both orientations, as indicated for the A and B series oftransformants. (Lower) Sequence of the intergenic region of gene pair A/B.L12. The sequence shown is from +1 to +1, with nucleotides absent from our constructs indicated in lower case. The chorion-specific hexanucleotide TCACGT is boxed, and the TATA elements for the two transcriptional directions are underlined. Thick arrows indicate elements with dyad symmetry, including the sequence TTGTGAAA, a motif highly conserved in intergenic regions of B. mori chorion gene pairs (19). Sequences with partial homology to this element are indicated with thin arrows, AA&T, A1M, and AA2 indicate the 3' end points of corresponding deletion mutants. Brackets enclose the regions substituted in mutants Li and LC, and the introduced sequence is shown above the wild-type one. Downloaded by guest on September 30, 2021 Developmental Biology: Mitsialis et al. Proc. Natl. Acad. Sci. USA 84 (1987) 7989 mediated transformation, correctly initiated transcripts from assay, the following reconstruction experiment was per- both genes of the pair accumulate abundantly in ovaries of formed: ovarian extract from a line of the Awt series, in an transformed females and specifically in follicles at the late amount 0.2% of that routinely used, was assayed in the stages of choriogenesis. A 272-bp fragment representing the presence of a 500-fold excess of extract from males, ovari- shared intergenic region of the A/B.L12 gene pair from ectomized females, or ovaries from a control CC line (Fig. 2, positions -4 to -2 (relative to the transcriptional initiation lanes R). No difference in activity was observed among the sites of the A and B genes, respectively) was inserted into various assay conditions. CarCat-1 in both possible orientations. The resulting plas- The formal possibility that the absence of detectable mids pAwt and pBwt were used to transform Drosophila activity in nonovarian tissues might be due to tissue-specific yielding lines of the A or B series. Lines harboring single instability of CAT mRNA and/or protein cannot be excluded inserts were selected for further analysis. but is rather unlikely. Expression of the CAT gene under the Sex and Tissue Specificity of Marker Gene Expression. In a control of certain Drosophila promoters results in accumu- preliminary series of experiments, deletion of each of the lation of stable CAT-specific transcripts and the appearance chorion genes of clone pB23 at the border of the intergenic of detectable enzymatic activity in the fat body (20) as well region was shown not to interfere with normal expression of as in certain tissues of male transformants (B. Bienz-Tadmor, the remaining gene (data not shown). Subsequent analysis of personal communication). Furthermore, we have verified transformant lines harboring CAT fusions with the intergenic that CAT-specific transcripts are absent in RNA from non- region alone, devoid of all other regions of the A/B.L12 pair, ovarian tissues of the Awt and Bwt transformant series demonstrated exclusively ovarian expression of the reporter (results not shown). gene, monitored through assays of CAT enzymatic activity Temporal Specificity of Marker Gene Expression. The (16). Thus, neither the chorion structural genes nor their temporal specificity of ovarian expression of the CAT gene 3'-flanking sequences are essential for the correct tissue and under the control of the Bombyx chorion gene promoter was sex specificity of expression. Assays on lines representative directly compared with the temporal specificity ofexpression of the various transformant series are shown in Fig. 2. Ex- of the intact moth gene pair within the same tissues. A tracts were prepared from males, ovariectomized females, transformant line of the Awt series, harboring the CAT and ovaries. Invariably, all control extracts prepared from construct in the X chromosome, was crossed to a transform- nine lines transformed with the CarCat-1 vector alone were ant line that had an insert of the complete 3.1-kb A/B.L12 devoid of activity (Fig. 2, lanes CC). In contrast, all lines fragment in the X chromosome (B23 series, ref. 12). Heter- ozygous females resulting from this cross were dissected, and harboring the intact wild-type 272-bp moth intergenic region RNA as well as total cellular extracts were prepared from in either the A or B orientation (14 and 9 independent lines, staged follicles representing the various stages ofDrosophila respectively) yielded enzymatic activity, exclusively in the oogenesis (21). The RNA was subjected to RNA gel blot ovary (Fig. 2, lanes Awt and Bwt). The sensitivity of the analysis using probes specific for either the CAT gene or the assay was sufficient to establish the absence of activity in moth B chorion gene. As has been reported (12), B-gene males and ovariectomized females at a level 4000 times lower specific transcripts accumulate predominantly in the late than in ovarian extracts. To eliminate the possibility that the stages of Drosophila choriogenesis (Fig. 3 Upper, arrow B). absence of detectable activity in nonovarian extracts is due Transcripts ofthe CAT gene accumulate coordinately, reach- to the presence of factors interfering with the enzymatic ing maximal levels during the later part of stage 14 (Fig. 3 Upper, arrow CAT). Only trace amounts of CAT and B gene transcripts are detectable in earlier follicular stages, in a manner reminiscent of the Drosophila siS chorion gene (12, 22). The size of CAT transcripts corresponds to the predicted polyadenylylated message spanning the CAT structural gene and the simian virus 40 sequences. Low amounts ofa smaller RNA species, with homology to simian virus 40 but not CAT sequences, appear transiently during stages 9 to 11 (results not shown). The temporal profile of CAT activity in follicular extracts is in general agreement with the temporal profile of the transcripts: in all independent lines, ofboth the Awt and Bwt series, maximal levels ofactivity are observed in late stage 14 (Fig. 3 Lower). A low but significant level of enzymatic 'I ^ 0 6 6 ** 0 * * 0 o To @ a 0 ~.!.s activity is also evident in earlier stages of oogenesis. Its prevalence could be explained by low-level transcription and

> > > > > > > >-0 > "o > > a significantly higher stability ofCAT protein relative to CAT s"o 0 0>so00 0 slo 0 0>no00 0 so 0 0 :> 0 0 0 0 mRNA. It is pertinent that early follicular stages last sub- 0+ 0+ 0+ G+- 0+- 0-* stantially longer than the late stages 11 stages: whereas 1-8 are L o-f 1 1 I awt Bw l a total of 63 hr long, stages 9-10 are a total of 11 hr long, and CC Awt Bwt ALI B L1 BLC R stages 11-13 are a total of 3 hr long, stage 14 lasts <2 hr, and FIG. 2. Regulated expression of the CAT gene in transformant its later part even less (23). The "early" CAT activity, which Drosophila. CAT activity assays were performed on extracts derived is one to two orders of magnitude lower than in late stage 14, from males (d), ovariectomized females (9-ov), or ovaries (ov) from is apparently specific for follicles, since no enzymatic activity lines representing the various transformant series. Lanes: CC, line is detectable in other tissues, and tends to be enhanced in harboring the CarCat-1 vector alone; Awt and Bwt, lines in which the early to mid-choriogenic versus prechoriogenic stages. CAT gene is under the control of the wild-type intergenic region in Quantitation of CAT activity in staged follicles reveals the two respective orientations; ALl, BL1, and BLC, lines in which identical the CAT gene is under the control of the respective substitution temporal specificities for transformants of the Awt mutants; R, reconstruction experiment showing absence of CAT- and the Bwt series (Fig. 4). Thus, the intergenic region acts inhibitory activities in the various extracts. The lower spots in the bidirectionally as two coordinate promoters that can devel- autoradiogram represent input chloramphenicol, and the upper spots opmentally regulate a marker gene, just as it does for the its acetoxy forms. intact A and B chorion genes. We conclude that the 272-bp Downloaded by guest on September 30, 2021 7990 Developmental Biology: Mitsialis et al. Proc. Natl. Acad. Sci. USA 84 (1987) B23 Awt 1.5 Awt 5 Bwt kb _ 1.0 I.C 0) .2U

2,0 . ---- CAT >- 5

0 i 5 - - 1.5 AM 1 .5 BLC 1.1 4 o 0 1.0 1.c 0.8- 0.- -W-- B 0.5 i -l

- - - , andr-

OF ! .- ... 1-8 .9 101OA 8 111213 14L .fi1i8 9 10A 1 1213Itz i4L i4E 14E Stages of Oogenesis Stages of Oogenesis ,I.. I , ...... 0 65 70 75 0' 65 70 75 DEVELOPMENTAL TIME (hrs.) W su ~0 A - FIG. 4. Temporal specificity of CAT gene expression in various transformant lines. Extracts from an equal number of follicles from F- each oogenetic stage were assayed. CAT activity is expressed as the (-) percentage of input chloramphenicol converted to the acetoxy forms under standard conditions. Values for corresponding stages in the different lines were corrected for recovery by reference to the 0 0@ 0 0 00 0 0 average protein content of each stage. The duration of each stage is indicated relative to the total time required for oocyte development 1I (23). CD < CO W1 G 0 .,- N .'r cap site sequences that could exert influence on the rate of STAGES OF OOGENESIS transcriptional initiation. Also, since we do not know the relative stabilities of CAT and moth chorion transcripts in the FIG. 3. Correlation of chorion gene and CAT gene expression in follicle cells, we cannot compare directly the rates of tran- staged follicles. Follicles representing the various stages ofDrosoph- CAT constructs and with the ila oogenesis were dissected from heterozygous females harboring scription achieved with the both an Awt and a B23 transposon. Stage 14 of King (21) was intact chorion gene pair. We note, however, that only in the subdivided into early (14E) and late (14L) parts, according to highest expressing Awt line is the level of CAT transcripts completion of respiratory appendage formation. (Upper) RNA gel comparable to the level of transcripts accumulated in trans- blot analysis ofequal amounts ofstaged follicular RNA, using probes specific for the CAT gene or the B chorion gene. Arrows point to the respective major transcripts. The migration of RNA and single- . 0 stranded DNA markers of known size (in kb) is indicated. (Lower) S CAT activity levels in total cellular extracts prepared from an equal number of follicles from each respective stage. 100

intergenic region contains elements sufficient for temporal as Cl: -I 4 z , 0 well as sex- and tissue-specific regulation. -

Sensitivity of CarCat-1 Derivatives to Position Effects. H- 10 Quantitation of CAT activity in late stage 14 follicles of - 0~~~~~~~~~~~~0 independent transformant lines of the Awt and Bwt series J revealed remarkable position effects on expression: whereas -. 0 .- the sex, tissue, and temporal specificity is invariant; within 4 0~~~~~~~~~ each series, level ofexpression can vary by >50-fold (Fig. 5). Lii The magnitude of this quantitative variation is higher than in 0 most other studies involving Drosophila transformation 0~~~~~~~ (6-10). It is also substantially higher than the variation r-J observed with the complete 3.1-kb moth chorion DNA O.1 i.. II-- - .£ fragment (12) or with truncated derivatives bearing the intergenic region and one of the two genes (data not shown). Awl AAAT" AWAIAA'2 'AL1 'BL1' 'BLC''Bwt' The median level of activity seems lower with Bwt than with Awt lines (28 ± 10 versus 149 ± 33 arbitrary units, TRANSFORMANT LINES respectively). Analysis of the two distributions through the two-sample Kolmogorov-Smirnov test (24) indicates that this FIG. 5. CAT activity levels in various transformant lines. Assays < on an equal number of stage 14L follicles from each transformant line difference is highly significant (P 0.005). However, these were performed, and the relative CAT activity values were corrected results do not necessarily suggest an inherent asymmetry in for recovery by reference to total protein content. Solid circles, promoter strength between the A and B directions, since the independent transformant lines. Median values for each series of fusions of the intergenic region with the CAT gene in each of lines are indicated by horizontal bars and standard deviations by the two respective constructs create nonidentical putative vertical bars. Downloaded by guest on September 30, 2021 Developmental Biology: Mitsialis et A Proc. Natl. Acad. Sci. USA 84 (1987) 7991 formants bearing the intact chorion gene pair. It is possible the quantitative reduction of expression caused by the AAT, that the chorion structural genes or the 3'-flanking DNA the AAM, and possibly the LC mutations. It is noteworthy that within the 3.1-kb fragment contain additional element(s) with the intergenic region includes multiple short repeats, some of enhancing effects on the rate oftranscription (although not on them in a dyad symmetrical pattern. The 272-bp intergenic its developmental specificity). When the intergenic region is region may also act as a tissue-specific enhancer, since it tested in isolation, substitute enhancing elements might be leads to late follicle cell-specific activation ifplaced upstream occasionally provided by host sequences proximal to the of an enhancerless Adh gene (N .S., unpublished data). In any insertion site, accounting for extensive expression in certain case, the results establish that one necessary, positively lines. In this context it is interesting to note that deletion of acting cis-regulatory element encompasses the chorion- 3'-flanking sequences from the fushi tarazu gene results in specific hexanucleotide TCACGT. That element acts bidi- increased sensitivity to position effects on expression (25) rectionally, since it is necessary for expression from both the and that within this deleted region reside specific nuclear A and the B gene promoters. scaffold attachment sites (26). Deletion and Substitution Mutations: The Hexanucleotide We thank Dr. R. C. Lewontin foradvice on statistical analysis; Dr. TCACGT Is Essential for Chorion Gene Promoter Function. C. Gorman for providing plasmid pSV2-cat; Drs. H. T. Nguyen and As a first step toward identifying the cis-regulatory elements B. Bienz-Tadmor for helpful discussions; M. Youk-See for graphics; the a series in vitro-mutated B. Klumpar for photography; and T. Dahill for secretarial assistance. of intergenic region, of deriva- The work was supported by a grant from the National Science tives of the region were constructed and functionally ana- Foundation to F.C.K. S.A.M. and N.S. were Fellows ofthe Charles lyzed as above. A deletion derivative of the Awt series, A. King Trust. lacking the distal 29 nucleotides of the intergenic region, including the B-gene "TATA" element (see AAT in Fig. 1), 1. Chambon, P., Dierick, A., Gaub, M. P., Jakowlew, S., retains the ability to activate the marker gene in the ovaries Jonstra, J., Kurst, A., Lepeynec, J. P., Oudet, P. & (see AAT in Fig. 5), but at a considerably lower level (24 + Reudelhyber, T. (1984) Rec. Prog. Hormone Res. 40, 1-42. 9 arbitrary units versus 149 ± 33 arbitrary units in the Awt 2. Schaffner, W. (1985) in Current Communications in Molecular series). That difference is highly significant according to the Biology: Eukaryotic Transcription, ed. Gluzman, Y. (Cold Kolmogorov-Smirnov test (P < 0.005). Derivatives lacking Spring Harbor Laboratory, Cold Spring Harbor, NY), pp. an additional 15 AA1 in retain 1-17. bp (see Fig. 1) also the ability 3. Kelly, J. H. & Darlington, G. J. (1985) Annu. Rev. Genet. 19, to support marker gene expression, at a similarly reduced 273-296. level (Fig. 5). The median CAT activity level for 11 lines in 4. Spradling, A. C. & Rubin, G. M. (1982) Science 218, 341-347. this series (17 ± 6 arbitrary units) is also significantly lower 5. Rubin, G. M. & Spradling, A. C. (1982) Science 218, 348-353. than in the Awt series (P < 0.005). Despite this nearly one 6. Cohen, R. S. & Meselson, M. (1985) Cell 43, 737-746. order of magnitude quantitative difference the wild-type 7. Fisher, J. A. & Maniatis, T. (1986) EMBO J. 5, 1275-1289. temporal regulation of expression is preserved (see AA1 in 8. Bourouis, M. & Richards, G. (1985) Cell 40, 349-357. Fig. 4). Removal of an additional 22 nucleotides (see AA2 in 9. Raghavan, K. V., Crosby, M. A., Mathers, P. H. & Meyero- Fig. 1), totally eliminates CAT gene expression in all seven witz, E. M. (1986) EMBO J. 5, 3321-3326. lines AA2 in 10. Garabedian, M. J., Shepherd, B. M. & Wensink, P. C. (1986) analyzed (see Fig. 5). Cell 45, 859-867. The AA2 deletion removes the hexanucleotide TCACGT 11. Kafatos, F. C., Mitsialis, S. A., Nguyen, H. T., Spoerel, N., (Fig. 1), an element which is specifically found in equivalent Tsitilou, S. G. & Mazur, G. D. (1987) in Development as an positions in all sequenced Drosophila chorion gene promot- Evolutionary Process, eds. Raff, R. & Raff, E. C. (Liss, New ers (27) and in the intergenic region of most moth chorion York), pp. 161-178. gene pairs (19). Therefore, we tested the possibility that 12. Mitsialis, S. A. & Kafatos, F. C. (1985) Nature (London) 317, inactivation of the promoter in the AA2 construct was due to 453-456. deletion of the chorion-specific TCACGT element. A clus- 13. Gorman, C. M., Moffat, L. F. & Howard, B. H. (1982) Mol. tered substitution mutant of the intergenic region was con- Cell. Biol. 2, 1044-1051. in which 8 nucleotides were 14. Rubin, G. M. & Spradling, A. C. (1983) Nucleic Acids Res. 11, structed, including TCACGT 6341-6351. replaced by an unrelated sequence (Fig. 1, sequence Li). 15. Feinberg, A. P. & Vogelstein, B. (1983) Anal. Biochem. 132, This "linker-substituted" mutation indeed completely pre- 6-13. vented activation of the marker gene in the follicles. The 16. Shaw, W. & Brodsky, R. (1968) J. Bacteriol. 95, 28-36. effect was evident in both orientations of the intergenic 17. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. region (Figs. 2 and 5; 5 and 6 independent lines for ALl and 18. Ruddel, A. & Jacobs-Lorena, M. (1983) Wilhelm Roux's Arch. BL1, respectively). In contrast, all 13 lines bearing a control Dev. Biol. 192, 189-195. linker substitution, within a part of the intergenic region that 19. Spoerel, N., Nguyen, H. T. & Kafatos, F. C. (1986) J. Mol. does not show a comparable evolutionary conservation (Fig. Biol. 190, 23-35. had detectable 20. Davies, J. A., Addison, C. F., Delaney, S. J., Sunkel, C. & 1, sequence LC), ovarian-specific expression Glover, D. M. (1986) J. Mol. Biol. 189, 13-24. (Figs. 2 and 5, lanes BLC). This mutation did not affect the 21. King, R. C. (1970) Ovarian Development in Drosophila temporal specificity of expression (Fig. 4, BLC series), melanogaster (Academic, New York). although it did appear to reduce the median CAT activity 22. Thireos, G., Griffin-Shea, R. & Kafatos, F. C. (1980) Proc. level of this series in comparison to the Bwt series (2.2 ± 1 Natl. Acad. Sci. USA 77, 5789-5793. versus 28 ± 10 arbitrary units, respectively; 0.025 < P < 23. David, J. & Merle, J. (1968) Drosoph. Inf. Serv. 43, 122-123. 0.05). 24. Sokal, R. R. & Rohif, F. J. (1981) Biometry: The Principles Concluding Remarks. These studies show that DNA ele- and Practice of Statistics in Biological Research (Freeman, ments sufficient for developmentally appropriate expression San Frarncisco). are a 25. Hiromi, Y., Kuroiwa, A. & Gehring, W. J. (1985) Cell 43, contained within 272-bp intergenic region ofthe B. mori 603-613. chorion gene pair, A/B.L12. We suspect that multiple regu- 26. Gasser, S. M. & Laemmli, U. K. (1986) Cell 46, 521-530. latory elements are present in that region, because of the 27. Wong, Y.-C., Pustell, J., Spoerel, N. & Kafatos, F. C. (1985) highly specific developmental expression of the genes, and Chromosoma 92, 124-135. Downloaded by guest on September 30, 2021