Mol Gen Genet (1995) 249 : 545-556 © Springer-Verlag 1995

Christophe Antoniewski • Michael S. O'Grady Ronda G. Edmondson • Suzanne M. Lassieur Helen Bene~

Characterization of an EcR/USP heterodimer target site that mediates ecdysone responsiveness of the Drosophila I.sp.2 gene

Received: 15 May 1995 / Accepted: 24 July 1995

Abstract The Larval serum protein-2 gene (Lsp-2) of Drosophila melanogaster is uniquely expressed in the fat body tissue from the beginning of the third instar to the end of adult life. Accumulation of the larval Lsp-2 transcript is enhanced by 20-hydroxyecdysone. To study the molecular basis for ecdysone regulated Lsp-2 activity, deletion mutants of the Lsp-2 Y-flanking region were constructed by fusion to either the Escherichia coli chloramphenicol acetyltransferase (CAT) gene or to an hsp70-lacZ hybrid gene encoding /Lgalactosidase. Constructs transfected into Drosophila $2/M3 cells were shown to confer transient ecdysone inducibility on the reporter genes. A single functional ecdysone response element (EcRE) was localized at position - 7 5 relative to the Lsp-2 transcription initia- tion site. In gel mobility shift assays using fat body nuclear extracts or nuclear receptors synthesized in vitro, a 27-bp sequence harboring the EcRE bound both the Drosophila ecdysone receptor and the Drosophila retinoid-X homologue, Ultraspiracle, in a cooperative manner. Competition experiments indi- cate that the affinity of the Lsp-2 EcRE for the ecdysone receptor complex is comparable to that of the canoni- cal EcRE of the hsp27 gene and is at least 4-fold greater than that of Fbpl, another fat body-specific Drosophila

Communicated by J. A. Campos-Ortega

C. Antoniewski Institut Jacques Monod, Centre National de la Recherche Scientifique and Universit6 de Paris 7 - Denis Diderot, 75251 Paris Cedex 05, France

M. S. O'Grady 1 - R. G. Edmondson • S. M. Lassieur • H. Beneg (5:~) Department of Biochemistry and Molecular Biology, University of Arkansas for Medical Sciences, Slot 516, 4301 West Markham St., Little Rock, Arkansas 72205, USA

Present address: 1 Department of Zoology, University of British Columbia, 2354-6270 University Blvd., Vancouver, B.C. V6T 1Z4, Canada

gene. Our results suggest that structural features of this EcRE determine its ability to induce ecdysone respon- siveness at a lower ligand concentration and may form the basis for differential hormone responsiveness within the fat body.

Key words Drosophila melanogaster • Larval serum protein-2 gene. Ecdysone response element. Ecdysone receptor • USP/Drosophila retinoid-X receptor

Introduction

The arthropod steroid hormone 20-hydroxyecdysone (20E) is the main orchestrator of the complex sequence of morphogenetic events underlying metamorphosis in holometabolous insects such as Drosophila melanogas- ter (reviewed in Riddiford 1993). Ecdysone action, like that of vertebrate steroids, is mediated by specific nu- clear receptor binding to target sites of regulated genes. The ecdysone receptor (EcR) was shown to bind to specific ecdysone response elements (EcRE) only as a heterodimer comprising the EcR protein encoded by the ecdysone receptor gene (Koelle et al. 1991) and the Drosophila retinoid-X receptor homologue, USP, a product of the ultraspiracle (usp) locus (Thomas et al. 1993; Yao et al. 1992, 1993). Functional EcREs consist of imperfect palindromes, similar to the sequences that bind vertebrate steroid hormone receptors (Antoniewski et al. 1993; Cherbas et al. 1991; Luo et al. 1991; Martinez et al. 1991; Riddihough and Pelham 1987; Fig. 2).

In vitro analysis of the EcR binding properties of mutated EcREs has demonstrated the requirement for single-base spacing between the half palindromes and the importance within the palindrome of specific positions dictating binding affinity (Antoniewski et al. 1993; Martinez et al. 1991; Ozyhar and Pongs 1993). However, to date there has been no analysis of the role of EcRE structure in determining a tissue-specific

546

h o r m o n e response in the context of the varying concen- t rat ions of the receptor ligand, eedysone, during Drosophila development . Unders tand ing this role re- quires the character isat ion of na tura l EcREs in the cis-regulatory sequences of ecdysone-responsive genes. Examples of such Drosophila genes are the Fat body protein 1 (Fbpl) gene and the Larval serum protein-2 gene (Lsp-2) which encodes the basic subunit of one of the ma jo r hexameric serum proteins, LSP-2, of third- instar larvae (Akam et al. 1978). Fbpl and Lsp-2 were originally identified as two of six genes uniquely ex- pressed in the Drosophila fat body during the third larval instar (Lepesant et al. 1982). Lsp-2 is ecdysone- inducible well before Fbpl, which responds directly to ecdysone during the second half of the third larval instar and has a well character ized EcRE (Andres et al. 1993; Antoniewski et al. 1993; Nakan i sh i and Ga ren 1983). Fol lowing induct ion at the onset of the third instar, Lsp-2 m R N A accumulation peaks at mid-third instar and then declines significantly by the time of pupar- ium format ion (Andres et al. 1993; Lepesant et al. 1986). A low-level phase of Lsp-2 expression then sets in by the head-evers ion stage of pupa l deve lopment and is main- tained th roughou t adult life (Beneg et al. 1990; Shirras and Bownes 1989). In the ecdysone-deficient m u t a n t l(3)ecd its or ecd 1 (Garen et al. 1977; Lepesant et al. 1982) the Lsp-2 R N A level is s t rongly reduced at the restrict- ive temperature; addi t ion of ecdysone to m u t a n t larvae results in a 10-fold induct ion ofLsp-2 m R N A synthesis (Nakanishi and Ga ren 1983; Lepesant et al. 1986).

Recently Andres et al. (1993) observed during the third larval instar four coordinate changes in activity of multiple genes, which m a y reflect regulation by small hormone peaks in prepara t ion for metamorphos is in Drosophila. As one of only a few cloned Drosophila genes with a unique tissue specificity and a k n o w n ecdysone response early in the third instar well before the onset of metamorphos i s , Lsp-2 offers an ideal oppor tun i ty to characterize the molecular basis for stage- and tissue- specific responses to varying h o r m o n e levels. Here we repor t the m a p p i n g and character izat ion of a func- t ional EcRE within the Y-flanking region of Lsp-2. We show that the EcR isoform found in the larval fat body binds to the Lsp-2 EcRE and forms a nucleoprote in complex including the U S P protein. The Lsp-2 EcRE appears as a new functional E c R / U S P target site whose high affinity for E c R / U S P m a y provide an explanat ion for the ecdysone responsiveness of Lsp-2 at a low con- centra t ion of l igand during the early third larval instar.

Materials and methods

Plasmid constructs

Transfection control plasmids

The actin-fl-gal expression vector includes 298 bp of the Drosophila cytoplasmic actin gene promoter plus untranslated RNA leader

sequences linked to the E. coli lacZ reporter gene in the pCaSpeR vector (Thummel et al. 1988). The X-188-cc-cat expression vector includes the basal Eip28/29 promoter (-188 to + 11) linked to the cat gene (Cherbas et al. 1991).

Transfection vectors

Lsp-2-cat constructs. The Lsp-2-cat plasmids were constructed by inserting blunt-ended fragments from the 5'-flanking region of the Drosophila Lsp-2 gene (Lepesant et al. 1982; Fig. 1A) into the SmaI site of the promoterless plasmid vector pSV0-CATsm a (Benyajati and Dray 1984). 3.1Lsp-2-cat includes an EcoRI-XhoI fragment of Lsp-2 (Lsp-2 sequences -3100 to + 16); 1.5Lsp-2-cat, an XhoI fragment (Lsp-2 sequences -1500 to + 16). 0.68Lsp-2-cat contains a ClaI- XhoI fragment (Lsp-2 sequences -667 to + 16); and 0.28Lsp-2-cat, a PvuII-XhoI fragment (sequences -267 to + 16 ofLsp-2). O.08Lsp- 2-cat was constructed using an HaeIII-XhoI fragment (sequences - 6 5 to + 16). All constructs were verified by restriction endonuc-

lease analysis. Lsp-2-hsp-lacZfusion 9enes. To study regulatory sequences both

in cell transfection assays and whole-animal, in vivo transformation assays, an hsp-lacZ shuttle vector (pJen) was derived from pHZ50PL (Hiromi and Gehring 1987) by elimination of the rosy gene and P element sequences, pJen includes the Drosophila hsp70 basal promoter ( -50 to +271), the lacZ reporter gene and appropriate hsp70 polyadenylation signal sequences, inserted in pUC8 (Fig. 3A); the original polylinker of pHZ50PL was retained. Blunt-ended frag- ments of the Lsp-2 Y-flanking region were inserted upstream of the hsp70 promoter into the unique, blunt-ended NotI site of pJen. 0.23Lsp-2-hsp70-IacZ contains a PvulI-NciI fragment (Lsp-2 sequences - 267 to - 37). 0.126Lsp-2-hsp70-IacZ carries a HaeIII fragment (Lsp-2 sequences -190 to -66), and O.076Lsp-2-hsp70-lacZ, a PvuII-HaeIII fragment ( - 267 to - 191). The 0.23Lsp-2*-hsp70-lacZ construct was generated by mutagenesis of the 0.23Lsp-2-hsp70-lacZ fusion gene using the primer 5'-GGATGCGACCGGTTGTCTGC- CGGATTGC-Y and the Unique-Site-Elimination kit (Pharmacia). All constructs were verified by DNA sequence analysis.

Receptor expression vectors

The Mt-EcR expression vector was constructed by inserting a blunt- ended XspI-HindlII fragment containing the entire coding region of the EcR-B1 cDNA (pMK1; Talbot et al. 1993) into the HinclI site of pRmHa-1 (Bunch et al. 1988). The Mt-USP expression vector con- tains the entire insert from the cDNA clone pZ7-1 (Henrich 1995), excised with SmaI and EcoRV, and inserted into the HincII site of pRmHa-1.

Cell culture and transfection

Cultured Drosophila $2/M3 cells (Cherbas et al. 1991) were grown at 23°C in M3 medium (Shields and Sang 1977) supplemented with 12.5% heat-inactivated fetal calf serum; Kc167/M3 cells (Cherbas et al. 1991) were grown at 23°C in M3 medium modified after Lindquist et al. (1982), supplemented with 5% heat-denatured fetal calf serum. Cells were transfected either with a calcium phosphate- DNA co-precipitate (DiNocera and Dawid 1983) or with liposomes formed in the presence of N-[1-(2,3-dioleoyloxy)propyl]-N,N,N- trimethylammonium methylsulfate (DOTAP, Boehringer Mann- heim) according to the manufacturer's instructions. In the C a P O 4 t r ans fec t ions , 10 7 cells were exposed to a test construct (20 gg of Lsp-2-cat or Lsp-2-hsp-lacZ DNA) and a control for transfection efficiency [1 gg of actin fl-gal DNA, described below, or 1.8 gg of X-188-cc-cat DNA (Cherbas et al. 1991)] in a final 10-ml volume of medium. Following a 3 h incubation with the CaPO4-DNA

co-precipitate, cells were cultured for 48 h in M3 medium with or without 1 FM 20-hydroxyecdysone (20E, Calbiochem). In the lipofection experiments, 3-5 x 106 cells were incubated for 24 h in the presence of 20 gg of DNA, including 10 gg of reporter plasmid and 5 gg each of ecdysone receptor and USP expression vectors and then cultured for an additional 24 h with or without 1 gM 20E. The ecdysone receptor/USP expression vectors were used in the absence of copper, following preliminary experiments showing high basal expression level of their metallothionein promoters. Transfections with all constructs were repeated at least three times in different experiments; each set of transfections included a negative or positive control for ecdysone-inducible reporter gene expression.

547

the beginning of each binding reaction to a final volume of 16 gl. For synthesis of EcR protein in vitro, a BamHI-HindIII fragment of the cDNA encoding the EcR-B1 isoform was subcloned in Bluescript KS + to generate the EcR-Bluescript KS + plasmid. USP protein was synthesized from a full-length cDNA clone, pZ7-1 (a gift of Vincent Henrich). After linearization with HindIII for EcR-Blue- script KS + and with ClaI for pZ7-1, RNA was transcribed in vitro (Stratagene) and translated in vitro in a rabbit reticulocyte lysate (Promega) according to the manufacturers' instructions. Translation products were used directly as reticulocyte lysates without further purification. Lysates incubated in the absence of RNA template were used as negative controls for non-specific binding.

Transient expression assays

Following treatment, the cells were collected by centrifugation, washed in phosphate-buffered saline, and a cell extract was prepared as described by Gorman et al. (1982). Protein concentrations were determined by the Bio-Rad protein assay. Depending on the gene construct tested, the activities of a given reporter gene were nor- malized for transfection efficiency using the activity of a co-transfec- ted standard: actin/~-gal for the Lsp-2-cat constructs and X-188-cc- cat for the Lsp-2-hsp-lacZ constructs. ~-Galactosidase assay. Cell extracts from individual transfected plates were assayed either according to Norton and Coffin (1985) or using a chorophenolred-/?-D-galactopyranoside (CPRG) substrate according to Simon and Lis (1987). Relative ]~-galactosidase activ- ities were derived from a standard curve of known/~-galactosidase activity using purified enzyme (Sigma) and normalized for protein concentration of each lysate. In the initial experiments to test the Lsp-2-hsp-lacZ constructs, /~-galactosidase activities were also cor- rected for transfection efficiency as determined by CAT activity from co-transfected X-188-cc-cat. ChloramphenicoI acetyltransferase (CAT) assay. Aliquots of cell extracts were assayed for CAT activity using [-14C]chloramphenicol as a substrate according to two methods: a chromatographic assay (Gorman et al. 1982) was used for the Lsp-2-cat constructs and a phase-extraction assay (Seed and Sheen 1988) for measuring X188-cc-cat expression as a control for transfection efficiency. Following product analysis by thin-layer chromatography and autoradiography, relative CAT activity was measured by excising [14C]chloramphenicol-containing spots and assaying radioactivity by scintillation counting. Activity values were normalized for protein content and relative transfection efficiency determined from expression of co-transfected actin/~-gal. In the phase-extrac- tion assay, product formation was measured by direct scintillation counting.

Gel mobility shift assays

Larval fat-body nuclear extract was prepared and mobility shift assays were performed as described in Antoniewski et al. (1993). Reaction mixtures contained 6-7 gg of fat-body extract and 12.5 fmol (80-100000 cpm) of probe in binding buffer (25 mM HEPES, pH 7.6, 60 mM KC1, 5% glycerol, 5 mM MgC12, 0.1 mM EDTA, pH 8, 0.75 mM dithiothreitol). Oligonucleotides were synthesized with a DNA synthesizer (Pharmacia or ABI) and purified as recom- mended by the manufacturer. Double-stranded probes and compet- ing oligonucleotides were obtained by annealing of complementary strands. Competing oligonucleotides were used in molar excess as indicated in the Figure legends. The intensity of the autoradio- graphic signals produced by protein/DNA probe complexes was quantified using a Molecular Dynamics Phosphorimager. The fol- lowing antibodies were used: anti-EcR or monoclonal DDA2.7 (Koelle et al. 1991) and anti-USP or monoclonal ABl l (Christian- son et al. 1992). One microliter of undiluted antibody was added at

Results

Ecdysone stimulation of the basal Lsp-2 promoter

A strategy involving deletion mutagenesis of cis-regula- tory regions and transient expression of fusion genes in Drosophila cell lines has been applied successfully to map DNA sequences underlying the ecdysone respon- siveness of a number of Drosophila genes, including hsp27, hsp23, Eip28/29, ~4ubulin (Cherbas et al. 1991; Dobens et al. 1991; Luo et al. 1991; Sobrier et al. 1989). In a preliminary survey of available Drosophila cul- tured cell lines, we used Northern blot analysis to assay for endogenous Lsp-2 expression. We observed no basal or ecdysone-inducible Lsp-2 expression in seven different well-characterized Drosophila cell lines ($2/M3, $3, Kc167/M3, D20, $2, Ox, or D12), treated with 1 gM 20-hydroxyecdysone (20E) (data not shown). Southern blot analysis of genomic DNA indicated that the Lsp-2 gene is indeed present in the different cell lines. Although expression of the endogenous Drosophila alcohol dehydrogenase gene (Adh) is sim- ilarly absent in specific Drosophila cell lines, Benyajati and Dray (1984) have induced transient ex- pression of both Adh transcription units in the same cultured cells. Given the successful use of cultured cells to characterize the transcriptional apparatus of Adh (Benyajati et al. 1992), we were encouraged to under- take a similar approach in the study of ecdysone-regu- lated Lsp-2 expression.

Deletion mutants of the Lsp-2 5'-flanking region, extending from position -3100 to + 16 relative to the transcription start site (Fig. 1), were fused to the protein coding sequence of the E. coIi chloramphenicol acetyl transferase (CAT) gene. The nucleotide sequence of the Lsp-2 gene, including a 660-bp 5'-flanking region, the transcription start site (+ 1) and a TATA box ( -31) , had been determined previously (Beneg et al. 1990; Mousseron, Lepesant-Kejzlarova, Chihara, Pictet and Lepesant, unpublished data). To include the putative Lsp-2 promoter in all 5' deletion mutants, all fusions were made at the + 16 position within the untranslated region of the Lsp-2 transcription unit. Transfection of Lsp-2-cat constructs carrying 3.1, 1.5, 0.68, or 0.28 kb of the Lsp-2 5'-flanking/promoter region (Fig. 1B) resulted in a reproducible 2 to 3-fold enhancement of CAT

548

expression following 20E treatment of $2/M3 cells (Fig. 1C). Shortening the Y-flanking region to 0.080 kb (O.08Lsp-2-cat), still produced strong CAT activity but the positive response to 20E was lost (Fig. 1C). These results suggested that one or more cis-acting DNA

A -3100 +16

E X C X C S

"" "'.,.,..

B .. ",, "'., Lsp-2 "',.. Cat -3100 +16

C E [

3.1Lsp-2-Cat r / / / / / / / / / / / / / / / / ~ I

-1500 +16 1.5Lsp-2-Cat v / / / / / / A I

-667 +15 0.68Lsp-2-Cat r / / / ~ = ,:'

-267 +16 0.28Lsp-2-Cat rA =

-65 +16 O. 08Lsp.2-Cat ~ =

SVO-Cat i I Sma

pBR322

elements between Lsp-2 positions -267 and - 6 6 could confer hormone regulation on a reporter gene in ecdysone-responsive Drosophila cells. Sequences be- tween positions - 2 6 8 and -3100 appear to have no effect on the ecdysone response of the Lsp-2 promoter in these cultured cells. Furthermore, Lsp-2 sequences - 6 5 to + 16 in the 0.08Lsp-2-cat construct contain

elements sufficient for a basal level of transcription.

Identification of a functional EcRE

Using the Genetics Computer Group (Madison, Wis., USA) software programs for DNA sequence analysis, we examined positions -267 to + 16 of the Lsp-2 5'-flanking region for potential EcREs by compari- son with a consensus (Fig. 2) derived from published

- EcRE Consensus Sequence:

- hsp27:

-6 -5 -4 -3 -2 -1 +1 +2 +3 +4 *5 +6

R G G T C A N T G A A A Y C T T G C C C

G G T T C A A T G C A C T a

C Deletion Construct Relative CATActivity Fold Induction

.1. i

3.1Lsp-2-Cat 185 100 1,9

l.SLsp-2-Cat 146 79 1.9

0.68Lsp-2-Cat 184 90 2.1

O.28Lsp-2-Cat 191 61 3.1

O.08Lsp-2-Cat 70 82 0.9

- Fbpl: G G T T G A A T G A A T T b

- S g s - 4 :

- Lsp-2:

A G T T C G A G G C A C C (D c

G G T T A A G T A A A C T (ID c

G A T T C G A A T C G T A A C G C T G C A A T G A T C G G T -3 .15

SVO-Cat 1.5 1.5 1.0 SmQ

Fig. 1A-C Structure and transient expression of Lsp-2-eat genes. A Linear map of the wild-type Lsp-2 gene showing the Y-flanking and Lsp-2 promoter region examined ( - 3 1 0 0 to +16), the Lsp-2 coding region (filled box) and the transcription initiation site (ar- row). Restriction sites used in construction of some of the fusion genes are indicated: E, EcoRI; X, XhoI; C, ClaI; S, SalI. B Deletion mutants of the Lsp-2 Y-flanking and promoter region ( -3100 to +16) fused to the cat gene. Fusion genes were constructed by

ligation of blunted Lsp-2 fragments into the Sinai site of the pSV0- CATsma vector as described in the text. Numbers above the maps refer to nucleotide positions relative to the Lsp-2 transcription start site at + 1 (arrow). The filled box represents the SV40 polyadenyla- tion signal sequences. C Effect of ecdysone on transient Lsp-2-cat expression in $2/M3 cells. Levels of CAT activity were determined in cell extracts 48 h after transfection with a particular Lsp-2-cat con- struct B and treatment with (+ ) or without ( - ) 1 gM 20E. Relative CAT activities are the average of at least three independent transfec- tion experiments; the normalized activity of the 3.1Lsp-2-cat gene was designated as 100. 'Fold induction' is the ratio of averaged CAT activities _+ ecdysone.

G A T T G C A A T C C G G C C A A T G A C C G G -63 -75

- Mutated sequence: C G G C a g A c a A C C G G

Fig. 2 Comparison of Lsp-2 Y-flanking DNA sequences to known EcREs. A more generalized EcRE consensus was derived from published consensus sequences. Lsp-2 sequences - 2 6 7 t o +16were analyzed for potential EcREs by comparison to a consensus se- quence derived from previously published consensus sequences (An- toniewski et al. 1993; Cherbas et al. 1991; Lehmann and Korge 1995; Luo et al. 1991; Martinez et al. 1991) and specifically to the hsp27 and Fbpl-D and Sgs-4 EcREs. Base positions with regard to the palindromic structure (lines) are indicated above the consensus. The nucleotide sequences of potential Lsp-2 EcREs are numbered rela- tive to the transcription start site of the Lsp-2 gene. Y indicates pyrimidine residues; R, purines; and N, any nucleotide. Additional palindromes flanking the potential Lsp-2 EcREs are marked by dashed lines. Mutations in the potential Lsp-2 EcRE at - 75, intro- duced in the context of the 0.23Lsp-2*-hsp-IacZ construct (see Fig. 3), are indicated at the bottom by small letters, a Riddihough and Pelham (1987); b Antoniewski et al. (1993); e Lehmann and Korge (1995)

549

consensus sequences (Antoniewski et al. 1993; Cherbas et al. 1991; Lehmann and Korge 1995; Luo et al. 1991; Martinez et al. 1991). The EcREs from three other Drosophila genes, hsp27, a prototypic ecdysone-in- ducible gene, Fbpl, and Sgs-4, a salivary gland-specific gene, were included in the comparison. Two putative EcRE palindromes from the Lsp-2 5'-flanking region, located at - 7 5 and -15 , respectively, are depicted in Fig. 2 and referred to by the 5' nucleotide of the coding strand. The similarity between both palindromes ex- tends beyond the 13 nucleotides of a typical EcRE but the - 7 5 sequence exhibits much stronger comp- lementarity between the two half-palindromes.

The transfection experiments reported above showed that the potential EcRE at - 1 5 by itself, in the O.08Lsp-2-cat construct, was not capable of conferring ecdysone regulation on the Lsp-2 promoter (Fig. 1). This prompted us to test the functional significance of the other putative EcRE at -75 . We made a series of constructs (Fig. 3A) with Lsp-2 sequences between - 2 6 7 and - 3 7 inserted in an hsp-lacZ shuttle vector

carrying no previously detected EcRE (see Materials and methods). In repeated experiments, transfection of 0.23Lsp-hsp-lacZ resulted in 3-fold stimulation of re- porter gene expression following 20E treatment (data not shown). Since sequences downstream of - 65 in the O.08Lsp-2-cat construct did not allow ecdysone induc- tion of reporter gene activity, Lsp-2 sequences - 66 to -267 were examined further by dissection into two

constructs: O.076Lsp-2-hsp-lacZ and 0.126Lsp-2-hsp- lacZ (Fig. 3A). Neither construct resulted in/?-galacto- sidase activity significantly different from the hsp-lacZ vector; and neither was ecdysone inducible following transfection into $2/M3 cells (data not shown). These experiments suggested that one or more EcREs might lie at positions -191 and/or - 6 6 of Lsp-2, the latter including part of the putative EcRE at - 7 5 (see Fig. 2).

To distinguish between these possibilities and to test specifically the role of the - 7 5 sequence in ecdysone induction of hsp-lacZ expression, we constructed the 0.23Lsp-2*-hsp-lacZ fusion gene (Fig. 3A) by introducing

four point mutations in the - 2 , - 1 , +1 and +2 residues of the putative EcRE (Fig. 2) at --75 within the 0.23Lsp-2-hsp-IacZ gene. The specific changes in the Lsp-2* sequence were in accordance with previous studies showing these residues to be critical for EcR/USP binding (Antoniewski et al. 1993; Ozyhar and Pongs 1993). For example, mutation of the C in the - 2 position of hsp2 7 sequence to an A prevents EcR

binding (Ozyhar and Pongs 1993), whereas mutation to a G had no effect (Antoniewski et al. 1993). Ecdysone inducibility of both wild-type and mutant Lsp-2-hsp- lacZ constructs was examined following transfection of $2/M3 cells (Fig. 3B). In contrast to earlier experi- ments, more reproducible levels of reporter gene activ- ity were obtained by co-transfection of the cultured cells with expression vectors for the Drosophila EcR and USP proteins. Under these conditions, the wild- type Lsp-2 sequences -267 to - 3 7 produced the same 3- to 4-fold stimulation of expression previously seen in re- sponse to ecdysone, whereas mutation of the putative EcRE at - 7 5 resulted in a complete loss of inducibil- ity. These results indicate not only that the putative EcRE at - 7 5 is functional in ecdysone-responsive cells, but that no other Lsp-2 sequence between posi- tions -267 and - 3 7 can mediate an ecdysone re- sponse in cultured cells. Furthermore, irrespective of hormone treatment, the mutant construct produced

Fig. 3A, B Effect of ecdysone on Lsp-2-hsp-lacZ expression. A Struc- ture of the Lsp-2-hsp-lacZ genes. Blunt-ended DNA fragments from the 5'-flanking region immediately upstream of the Lsp-2 "TATA box" were fused to the basal hsp70 promoter in the hsp-lacZ vector (pJen) as described in the text. 0.23Lsp-2*-hsp-lacZ was generated by introducing four point mutations in the EcRE_75 sequence (Fig. 2) of 0.23Lsp-2-hsp-IacZ. Numbers above the maps refer to nucleotide positions relative to the Lsp-2 transcription start site at + 1. B Effect of ecdysone on transient Lsp-2-hsp-lacZ expression in $2/M3 cells. Levels of/~-galactosidase activity were determined in cell extracts 48 h after co-transfection with the indicated hsp-lacZ gene and the Mt-EcR and Mt-USP expression vectors and a 24-h treatment with (+ ) or without ( - ) 1 laM 20E. /?-Galactosidase activities are the average of three independent transfection experiments and are ex- pressed in O.D. units per tag protein per rain; bars indicate standard errors of the mean

A

O. 23 L sp- 2- h sp -/a c Z V / / / / / / / / / / / / / / / A I t

-267 -191

O.076Lsp.2-hsp-lacZ

-190 -66

O. 126Lsp.2-hsp-lacZ

-267 -37

O. 23L sp -2 *-11 s,o -la c Z ~ / / / / / / ' / / / / / ~ / A

hsp-/acZ (pJen)

Lsp-2 lacZ pUC8 -267 -37

hsD70'TATA' ~ hsDTO'DolvAd'

B 0.025 -

C

0.020 - C

O S_ o.o15.

0.010 ,t-

d d 0.005

- +

hsp- lacZ

T

iiiiiiiiiii!iiil

T iiil I I - -I-

Lsp-2-hsp- lacZ

ITl - +

Lsp-2*-hsp- lacZ

550

Fig. 4A-C Binding of a putative Lsp-2 EcRE to the Drosophila EcR/USP complex. A Cooperative binding of EcR and USP proteins to an Lsp-2 EcRE. Nuclear receptor proteins were synthesized by in vitro translation of cDNA clones as described in the text. A series of common bands (indicated by the dot) due to non-specific binding of reticulocyte lysate proteins to the probe is observed in all lanes. B Binding of an Lsp-2 EcRE to EcR and USP proteins in a larval fat-body nuclear extract. Gel shift assays were performed as described in the text with the indicated 32P_labelled probes; the hsp27 EcRE probe was used unlabelled as a competitor in 400-fold excess. Antibodies were added at the beginning of each binding reaction. Complex B (arrow) remains after competition with the hsp27 EcRE probe. C Oligonucleotide probes used in gel shift assays. EcRE palindromes are underlined. Locations of the EcREs relative to + 1, the transcription start sites, are marked. The Lsp-2* probe was derived from the Lsp-2 EcRE probe by mutation of the indicated residues (lowercase letters) to prevent EcR/USP binding. Asterisks mark a motif common to the palindromic C/EBP and the Lsp-2 probes. a Antoniewski et al. (1992) b Rorth and Montell (1992)

A Translation product

Antibody

a two-fold higher level of reporter gene activity than the hsp-lacZ reporter gene alone (Fig. 3B). This observa- tion suggests that in the absence of hormone the Lsp-2 EcRE may actually mediate repression of promoter activity by the EcR/USP complex or interfere with a transcriptional activator.

1 2 3 /-., 5 6 USP

/ EcR EcR EcR EcR ,,Ik + .f.

USPUSP USP EcR USP

B 1 2 3 4 5 Probe Lsp2 Lsp2

Lsp2 Lsp2 hsp27

Competitor / hsp27 / / / Antibody / / EcR USP /

B - -

C hsp 2 7 EcRE probe:

5' AGACAAGGGTTCAATGCACTTGTCCAA 3' a -552 -526

Lsp-2 EcRE probe:

5' TGCAATCCGGCCAA.TGACCGGTCGCAT 3' -56 -82

Lsp-2* mutant probe:

5' TGCAATCCGGCagAcaACCGGTCGCAT 3'

Binding of nuclear factors to the Lsp-2 EcRE

Recently several investigators have shown that the Drosophila EcR must form a heterodimer with the USP protein to bind and regulate transcription through functional ecdysone response elements (Thomas et al. 1993; Yao et al. 1993). Hence, we tested the binding of the Lsp-2 palindrome at - 75 to the EcR/USP hetero- dimer by a gel mobility shift assay. An Lsp-2 EcRE probe (Fig. 4C) was synthesized, placing the -75 pal- indrome centrally within a 27-bp oligonucleotide in- cluding upstream and downstream Lsp-2 sequences. The subunits of the Drosophila EcR/USP heterodimer were prepared by in vitro transcription/translation of

Fbpl probe: 5' CGATTGGGTTGAATGAATTTTGCTG 3' a -85 -109

C/EBP probe: 5' TGCAGATTGCGCAATCTGCA 3'

cDNA clones of the EcR and usp genes. As shown in Fig. 4A, a retarded complex including the Lsp-2 EcRE probe is formed only in a mixture of both the EcR and the USP proteins (lanes 1-4). The presence of both receptor partners in the Lsp-2/protein complex was confirmed by supershifting of the complex by both EcR and USP antisera (lanes 5 and 6). In accordance with

551

these and the results shown in Fig. 3B, we shall refer hereafter to the palindromic sequence at position - 7 5 of the Lsp-2 gene as EcRE_75.

In a second experiment, the Lsp-2 EcRE probe was incubated with a nuclear fat body extract from third- instar larvae to test if EcRE_75 can bind the ecdysone receptor in the appropriate tissue during the last larval instar. As a control for EcR/USP binding, a parallel reaction was carried out with an hsp27 EcRE probe used in previous experiments (Antoniewski et al. 1993). A major retarded complex formed with the Lsp-2 probe co-migrated with a similar complex binding the hsp27 EcRE probe (compare lanes 1 and 5, Fig. 4B). The nature of the Lsp-2 complex was revealed by inclusion of the hsp27 EcRE probe (lane 2) or antibody against either the EcR or USP protein in the binding reactions (lanes 3 and 4). A 400-fold molar excess of hsp27 EcRE resulted in 100% competition or complete removal of the major Lsp-2 complex. Supershifted protein/DNA complexes in lanes 3 and 4 demonstrate that both EcR and USP proteins are present in the major Lsp-2 com- plex observed in lane 1. These results confirm that the Lsp-2 EcRE probe does indeed contain an ecdysone receptor binding site which binds the specific isoform of the EcR heterodimer in the fat body tissue where Lsp-2 is expressed.

To characterize the relative affinity of the Lsp-2 EcRE_75 for the Drosophila EcR/USP complex, we performed a competition experiment using the larval

hsp27 EcRE Lsp-2 EcRE Fbpl-D EcRE

nocomp X13 X66 X133 X13 X66 ~ '33 X13 X66 X133

~ - 1 0 ( " ~ E . ~ o 8¢ zmO c E o 6C

O9 O9 r r r r

,.63 2C ~ o 0

uJ no comp X13 X66 X133 X13 X66 X133 X13 X66 X133

Fig. 5 Relative affinities of three EcREs for the EcR/USP complex of the larval fat body. Binding of the EcR complex to the 32p. labelled hsp27 EcRE probe in the presence of varying amounts of cold competing probes (hsp27 EcRE, Lsp-2 EcRE 75, or Fbpl-D EcRE) was used to compare the binding affinities of these EcREs. The probes depicted in Fig. 4C were incubated with a fat body nuclear extract as indicated above each lane and described in the text. The relative affinities were measured as a percentage of hsp27 EcRE/receptor complex remaining and detected by autoradiogra- phy at different molar ratios (13, 66 or 133 : 1) of competitor to hsp27 probe. The amount of complex remaining was expressed as a per- centage of complex formed in the absence of competitor (no com- petitor = 100%)

fat body extract and varying concentrations of three cold oligonucleotide EcRE probes as homologous and heterologous competitors against the radiolabelled hsp27 EcRE probe (Fig. 5). The Lsp-2 EcRE_ 75 probe clearly served as a much better competitor than a pre- viously characterized EcRE (D EcRE) from another fat body-specific Drosophila gene, Fbpl (Antoniewski et al. 1993). While over 90% of the labeled hsp27 EcRE/EcR/ USP complex remained in the presence of a 13-fold excess of the cold Fbpl competitor, only 21% and 9% remained with the same excesses of Lsp-2 and hsp27 competitors, respectively. We conclude that under these conditions the Lsp-2 EcRE_75 exhibited about half the affinity of the hsp27 EcRE prototype but at least a 4-fold greater affinity than the EcRE of the Fbpl gene.

While competition with the hsp27 EcRE probe (Fig. 4B, lane 2) removed the major Lsp-2/EcR/USP complex, there remained a second complex, Complex B, formed with one or more proteins of the fat body nuclear extract. This Complex B is unique to the Lsp-2 EcRE probe (lanes 1-4) and is clearly not formed with the hsp27 probe (lane 5). In order to investigate the nature of EcR/USP complex binding to EcRE_ 7s and its effect on the formation of Complex B, further com- petition experiments were performed using the fat-body nuclear extract. A mutated Lsp-2 EcRE probe (Lsp-2*, Fig. 4C) was synthesized to correspond to the four point mutations introduced into the 0.23Lsp-2*- hsp-lacZ construct which is incapable of hormone induction.

The results shown in Fig. 6 demonstrate the specifi- city of EcR/USP complex binding to the EcRE_75 of Lsp-2. Cross-competition with the Lsp-2* probe had no effect on formation of the major complex (compare lanes 1, 4 and 5), whereas 100% competition was again observed with the hsp27 EcRE probe (lanes 6 and 7). The lack of competition with the Lsp-2* probe confirms the central region or the specific EcRE_ 7s palindrome in the Lsp-2 probe as the binding site for the EcR/USP complex. Furthermore, this EcR/USP binding site does not appear to overlap with the site used in formation of Complex B. Only cross-competition with the Lsp-2* probe completely and specifically removed Complex B (compare lanes 1, 4-7), indicating that Complex B forms with Lsp-2 sequences flanking EcRE _ 7 s. When the radiolabeled Lsp-2* probe itself is used for binding to the fat body nuclear extract in the absence of any competing oligonucleotide, only Complex B is formed (Fig. 6, lane 8). While these experiments suggested that Complex B does not include the EcR/USP dimer or nuclear proteins capable of specific binding to the hsp27 EcRE, it is possible that either the EcR or USP protein might bind the Lsp-2* probe in associ- ation with other nuclear receptors as partners. How- ever, the same antibodies used to reveal EcR and USP binding to EcRE_ 75 produced no supershifting of fat- body nuclear proteins bound to the Lsp-2* probe

552

Lsp2 Lsp2* Probe

' x50X200X50X200X50X200 ' ' Competitor Lsp2 Lsp2* hsp27

/ L s p 2 Lsp2* hsp27 / 1 2 3 4 5 6 7 8

anti anti / / / / / / / Antibody EcR USP x50X200X50X200X50X200 Competitor

Lsp2* hsp27 e/EBP / / /Lsp2* hsp27 c/EBP 1 2 3 4 5 6 7 8 9

Lsp2* probe

Fig. 6 An Lsp-2 binding activity that is independent of the EcRE. Gel shift assays were performed with a fat body nuclear extract and the indicated 32P-labelled probes as in Fig. 4C. The competitors were the same unlabelled probes (Fig. 4C) used in 50- and 200-fold molar excess as indicated. Complex B, remaining after competition with the hsp27 EcRE probe, is marked with an arrow

Fig. 7 Characterization of Complex B. Gel shift assays were per- formed with a fat body nuclear extract as described in the text and the 32p-labelled Lsp-2* probe (see Fig. 4C). The competitors were the same unlabelled probes described in Fig. 4C and used in 50- and 200-fold molar excess as indicated

(Fig. 7, lanes 1-3). We concluded that Complex B con- tains neither subunit of the EcR/USP heterodimer.

Examination of DNA sequences flanking the Lsp-2 EcRE_ 75 had revealed a 10-bp perfect palindrome dir- ectly downstream (Fig. 2). A strikingly related palin- drome is also found 3 nucleotides downstream of the -15 EcRE-like sequence. The palindrome specifically

flanking EcRE_75 exhibits strong homology with CCAAT/enhancer-binding protein- or C/EBP-like sites, previously mapped to several Drosophila fat body-specific and liver-specific enhancers (Fig. 2; An and Wensink 1995; Falb and Maniatis 1992). Further- more, in vertebrate systems C/EBP-like binding sites can be involved in the acquisition of steroid hormone responsiveness (Baniahmad et al. 1995; Ben-Or and Okret 1993). The palindrome flanking EcRE-75 (Fig. 2) most closely resembles one of the bZip sites (ATGTTGCAAT) of the YP enhancer, bound in vitro by the Drosophila C/EBP homologue DmC/EBP (Montell et al. 1992; An and Wensink 1995).

To determine if Complex B might contain a DmC/ EBP-related protein, we performed a cross-competi- tion experiment between the Lsp-2* probe and a DmC/EBP-binding site (Rorth and Montell 1992). As shown in Fig. 7 (lanes 8 and 9), the fat body protein(s) in Complex B are not competed away by the DmC/EBP

probe. Furthermore, a polyclonal antibody C138 (Rorth and Montell 1992) recognizing a specific motif of DmC/EBP outside the conserved bZip domain pro- duced no supershift of Complex B formed on the Lsp-2* probe (results not shown). These results indicate that the Drosophila homologue of C/EBP characterized by Montell et al. (1992) is not involved in the formation of Complex B.

Discussion

Using transfection of Lsp-2-fusion genes into cultured cells and gel mobility shift assays to characterize bind- ing to the ecdysone receptor complex, we have identi- fied a functional EcRE at position -75 of the fat body-specific Drosophila Lsp-2 gene. This EcRE-Ts represents one of a limited number of Drosophila EcREs which requires binding to the EcR/USP hetero- dimeric complex to mediate a hormone response (Cher- bas et al. 1991; Laval et al. 1993; Luo et al. 1991; Thomas et al. 1993). EcRE_ 75 lies just upstream of the Lsp-2 TATA box within a 230-bp region ( - 2 6 7 to -37), or larval enhancer, recently shown to be sufficient for unique larval fat-body expression of an

hsp-lacZ reporter gene in vivo (Beneg, Neal, Willis, Gadde, Castleberry and Korochkina, in press. No other potential EcREs fitting published consensus se- quences (Fig. 2) were observed in an extended analysis of over 3100 bp of Lsp-2 DNA, including a 667-bp Y-flanking region, the 2300-bp coding region and 160 bp of 3'-flanking region. Given that EcRE_ 75 binds the EcR and USP proteins in a fat-body nuclear extract from third-instar larvae, we propose that this EcRE and the surrounding sequences of the larval enhancer are responsible for the ecdysone regulation of Lsp-2 during the last larval instar.

Recently we tested the Lsp-2 sequences between -267 to - 3 7 by fusion, in either direct or reverse

orientation, upstream of the basal Eip28/29 promoter and the CAT reporter gene. Expression of the (direct- Lsp-2)-Eip-cat gene was enhanced 3-fold by 20E treat- ment, whereas the (reverse-Lsp-2)-Eip-cat construct allowed only a negligible (1.34-fold) ecdysone enhance- ment (data not shown). While this set of experiments confirmed that ecdysone-responsive Lsp-2 sequences lie between positions - 3 7 and -267, upstream of the Lsp-2 TATA box, the position of such sequences rela- tive to the basal promoter appeared critical to the ecdysone response.

Numerous studies of steroid hormone response ele- ments (HREs) have shown synergistic effects of mul- tiple elements as palindromes, half-palindromes or combinations of both (Beato 1988; Jantzen et al. 1987; Kato et al. 1992; Luo et al. 1991). Hence it is conceiv- able that the -15 EcRE-like palindrome of Lsp-2 (Fig. 2) might synergize the effects of EcRE_ 75. Indeed, the similarities between the -15 and -75 palin- dromes and their Y-flanking sequences (Fig. 2) suggest that they may have evolved as sequence duplications. However, the transfection assays reported here indicate that the - 15 palindrome alone is not sufficient for an ecdysone response in cultured cells.

Conservation of specific positions within each half- site of a EcRE palindrome has been shown to be critical to the binding of the EcR/USP complex. Previous mutational analysis of the EcREs of the hsp27 and Fbpl genes pointed to the importance of guanines at positions -6 , - 5 and + 2 for EcR binding in vitro (Antoniewski et al. 1993; Ozyhar and Pongs 1993). In contrast, divergence from the consensus at positions -4 , - 3 , + 3 or + 4 does not have a significant effect

on binding affinity for the EcR/USP complex. Both the -15 and - 7 5 Lsp-2 sequences have the requisite

G residues at +2 but only the -75 palindrome has a G at the - 5 position. This may explain why the - 15 palindrome does not act as an EcRE in cell transfection assays. However, given that the EcRE of the Drosophila hsp23 gene functions with a T in the same position (Luo et al. 1991), further studies of the tissue-specific ecdysone response of Lsp-2 in the whole animal are necessary to determine the significance of the - 1 5 palindrome.

553

Further comparison of the EcREs from the hsp27, Fbpl, Sgs-4 and Lsp-2 genes reveals the importance of specific residues for binding affinity to the EcR/USP heterodimer. Using the Lsp-2 and Fbpl EcREs as non- radioactive competitors against the labeled hsp27 EcRE probe, we show that the Lsp-2 EcRE binds the EcR/USP dimer with a four-fold greater affinity than the EcRE of the Fbpl gene. Previous mutational analy- sis of the Fbpl EcRE pointed to a C residue at position +5 as an important determinant of high-affinity

EcR/USP binding (Antoniewski et al. 1993). Conserva- tion of this C in both the Lsp-2 and the hsp27 EcREs may be responsible for the higher affinities of these elements for the EcR/USP heterodimer as compared to the Fbpl element. Similarly, the hsp23 EcRE (CTTTCAATGGCAG), with marked differences at the

- 5 and + 5 positions, is significantly weaker than that of hsp27 (Luo et al. 1991). A recent comparison of the binding affinities of two Sgs-4 EcREs with that of the hsp27 gene revealed one weak and one strong element (Lehmann and Korge 1995). Analysis of our results and those of Lehmann and Korge suggest that the strongest elements are those with conservation of three specific positions, - 5, + 2 and + 5. Conservation of only two of the three residues, as in element II of Sgs-4 or in the Fbpl element, results in weaker binding affinity. In contrast, positions ( - 6 and + 6, at which the Lsp-2 EcRE diverges most significantly from the consensus sequences as well as from the hsp27, Fbpl and Sgs-4 EcREs (Fig. 2), do not appear to be as critical to EcR/USP binding.

The 27-bp Lsp-2 probe for binding to EcRE_ 75 sup- ports the formation of at least two DNA-protein com- plexes with fat-body nuclear extracts from third-instar larvae. In contrast, the hsp27 EcRE probe of the same size forms only a single complex (Fig. 4). No other DNA- binding activity so closely associated with an EcR binding site has been observed in any previously characterized EcRE. In vitro neither the EcR/USP het- erodimer nor the proteins in Complex B interfere with the formation of the other complex (Fig. 6, lanes 1, 4-7). Studies of vertebrate steroid receptor activities have demonstrated positive or negative effects of nu- clear factors that are not members of the nuclear recep- tor superfamily (reviewed in Truss and Beato 1993). The term 'composite hormone response element' is used to describe those DNA sequences which bind both the hormone receptor and the accessory factor(s) to bring about the specific hormone response (Miner and Yamamoto 1991). The close proximity of the Complex B binding site to EcRE_ 75 suggests that the two bind- ing activities might cooperate in mediating the stimula- tory effect of ecdysone on Lsp-2 expression. Thus the 27-bp Lsp-2 EcRE probe may represent only part of a composite hormone element.

Our preliminary characterization of Complex B indi- cates that neither the EcR/USP subunits nor DmC/EBP, a specific Drosophila homologue of the

554

transcriptional activator C/EBP, are involved in its formation. Recently An and Wensink (1995) completed a detailed molecular analysis of functional transcrip- tion factor binding sites in the fat body-specific enhan- cer of the Drosophila yolk protein gene. This study suggests that another, as yet unidentified, member of the bZip family of proteins is present in the Drosophila fat body. However, at present, identification of such a binding protein in fat body nuclear extracts is parti- cularly intractable, given that bZip proteins may bind very similar DNA sequences.

In a functional assay using ecdysone-responsive cul- tured Drosophila cells, we show that EcRE_75 may exhibit functional properties not previously observed for other well-characterized EcREs. While EcRE_ 75 of Lsp-2 clearly mediates a three-fold stimulation of repor- ter gene expression in response to hormone, mutation of the element to prevent EcR-complex binding results in a reproducible two-fold activation of heterologous promoter activity irrespective of the presence or ab- sence of ecdysone (Fig. 3B). Inhibition of basal tran- scription in the absence of hormone has been observed for the EcREs of the Drosophila Eip28/29 and hsp27 genes (Cherbas et al. 1991; Dobens et al. 1991). In these cases binding of the EcR without ligand presumably caused the hormone response elements to repress re- porter gene activity below the basal level of the minimal promoter. Similar silencing activities (reviewed in Burcin et al. 1995) have been observed for unliganded vertebrate receptors, such as the thyroid hormone re- ceptor (Baniahmad et al. 1990; Graupner et al. 1989) which binds a corepressor protein in the absence of ligand (Baniahmad et al. 1995).

Our transfection experiments with the mutant 0.23Lsp-2*-hsp-lacZ suggest that the effects of the EcRE of Lsp-2 may be quite different. In the absence of ligand, EcRE_75 may also block the full effect of a second transcriptional activating sequence located within the 230-bp larval-specific enhancer. Treatment with hormone presumably results in a conformational change in the EcR-complex, such that either it alone becomes the primary activator of tran- scription or it cooperates with the second activator to bring about the full induction in response to ec- dysone. When the EcRE is mutated, the blocking on the second activator is relieved but so is the ec- dysone inducibility. While both the hsp2 7 and Eip28/29 EcREs may act as repressors, neither has been shown to interfere with a second activator function. Although the activating sequences revealed by the 0.23Lsp-2*- hsp-lacZ construct may lie anywhere within the 230-bp larval enhancer of Lsp-2, the bZip-like site just down- stream of EcRE-75 is an intriguing candidate for just such a transcriptional activating element. Indeed, binding of the EcR complex in the absence of hormone might partially block the ability of the bZip-like factor to activate a basal promoter. Mutation of the EcRE to prevent EcR binding removes the block,

revealing what may be the full activating potential of the bZip-like factor.

Little is known about the molecular basis for tissue- specific, ecdysone-responsive gene activity, in particu- lar early in the third instar. Recently the Drosophila LSP genes were shown to be part of the first coordinate response to a small ecdysone peak following the molt to the third instar (Andres et al. 1993). During the last larval instar, LSPs are resorbed from the hemolymph by the larval fat body and incorporated into character- istic protein granules thought to have important roles in protein metabolism during insect metamorphosis (Teller and Kunkel 1991). Ecdysone induces LSP re- sorption in Drosophila and may stimulate expression of a mature receptor for LSP uptake (Lepesant et al. 1978; Thomasson and Mitchell 1972). The product of the ecdysone-regulated Drosophila Fbpl gene may be part of a fat-body LSP receptor (Lepesant et al. 1978; Jean- Antoine Lepesant, personal communication). As Fbpl exhibits a rapid, direct response to a later peak of ecdysone, the difference in structure between the Lsp-2 and Fbpl EcREs may reflect different abilities to re- spond. It is interesting to speculate that the Lsp-2 promoter and its EcRE may be exquisitely sensitive to a low level of ecdysone. Indeed, the specific temporal patterns of Lsp-2 and Fbpl expression necessary for the completion of metamorphosis may be coordinated by fluctuating titers of ecdysone within the same tissue.

Acknowledgements We thank Lucy Cherbas for the gifts of the $2/M3 and Kc167/M3 cell lines, the X-188-cc-CAT vector, and for advice on cell culture and transfection experiments; Cheeptip Benyjati for the SVO-CATsm" vector; Vincenzo Pirrotta for the actin-fl-gal vector; Marilena Furia for the Mt-EcR and Mt-USP vectors; David S. Hogness for the EcR-B1 cDNA and EcR antibody DDA2.7; Vincent Henrich for the USP cDNA; Fotis C. Kafatos for the anti-USP monoclonal antibody AB11 and P. Rorth for the C/EBP polyclonal antibody. We are grateful to David Spivey and Mary Elizabeth Huffine for the construction of certain fusion genes; to Rebecca Willis for in vitro mutagenesis of EcRE_75; and to Jean-Antoine Lepesant for critical reading of the manuscript. C. Antoniewsld is a postdoctoral fellow of the Association pour la Recherche sur le Cancer. This work was supported by grants to HB from the National Science Foundation (DCB-8615894) and the Uni- versity of Arkansas for Medical Sciences and to J. A. Lepesant from the European Economic Community (SCI*-0123-C), the Association pour la Recherche sur le Cancer, the Ligue Nationale contre le Cancer (grant no. 6294) and the Centre National de la Recherche Scientifique.

References

Akam ME, Roberts DB, Richards GP, Ashburner M (1978) Drosophila: the genetics of two major larval serum proteins. Cell 13 : 215-225

An W, Wensink PC (1995) Three protein binding sites form an enhancer that regulates sex and fat body-specific transcription of Drosophila yolk protein genes. EMBO J 14:101-110

Andres AJ, Fletcher JC, Karim FD, Thummel CS (1993) Molecular analysis of the initiation of insect metamorphosis: a comparative study of Drosophila ecdysteroid-regulated transcription. Devel Biol 160: 388-404

555

Antoniewski C, Laval M, Lepesant J-A (1993) Structural features critical to the activity of an ecdysone receptor binding site. Insect Biochem Molec Biol 23:105-114

Antoniewski C, Laval M, Dahan A, Lepesant J-A (1994) The ec- dysone-response enhancer of the Fbpl gene olD. melanogaster is a direct target for the EcR/USP nuclear receptor. Mol Cell Biol 14 : 4465-4474

Baniahmad A, Steiner C, Kohne AC, Renkawitz R (1990) Modular structure of a chicken lysozyme silencer: involvement of an un- usual thyroid receptor binding site. Cell 61 : 505-514

Baniahmad A, Leng X, Burris TP, Tsai SY, Tsai MJ, O'Malley BW (1995) The tau 4 activation domain of the thyroid hormone receptor is required for release of a putative corepressor(s) neces- sary for transcriptional silencing. Mol Cell Biol 15 : 76-86

Beato M (1988) Gene regulation by steroid hormones. Cell 56 : 335-344

Ben-Or S, Okret S (1993) Involvement of a C/EBP-like protein in the acquistion of responsiveness to glucocorticoid hormones during chick neural retinal development. Mol Cell Biol 13:331-340

Bene~ H, Edmondson RG, Fink P, Kejzlarova-Lepesant J, Lepesant J-A, Miles JP, Spivey DW (1990) Adult expression of the Drosophila Lsp-2 gene. Devel Biol 142 : 138-146

Benyajati C, Dray J (1984) Cloned Drosophila alcohol dehydrogen- ase genes are correctly expressed after transfection into Drosophila cells in culture. Proc Natl Acad Sci USA 81 : 1701-1705

Benyajati C, Ewel A, McKeon J, Chovav M, Juan E (1992) Charac- terization and purification of Adh distal promoter factor 2, Adf-2, a cell-specific and promoter-specific repressor in Drosophila. Nucleic Acids Res 20:4481-4489

Bunch TA, Grinblat Y, Goldstein LS (1988) Characterization and use of the Drosophila metallothionein promoter in cul- tured Drosophila melanogaster cells. Nucleic Acids Res 16:1043-1061

Burcin M, Kohne AC, Runge D, Steiner C, Renkawitz R (1995) Factors influencing nuclear receptors in transcriptional repres- sion. Sem in Cancer Biol 5 : 337-346

Cherbas L, Lee K, Cherbas P (1991) Identification of ecdysone response elements by analysis of the Drosophila Eip28/29 gene. Genes Dev 5 : 120-131

Christianson AMK, King DL, Hatzivassiliou E, Casas JE, Hallen- beck PL, Nikodem VM, Mitsialis SA, Kafatos FC (1992) DNA binding and heterodimerization of the Drosophila transcription factor chorion factor 1/ultraspiracle. Proc Natl Acad Sci USA 89 : 11503-11507

DiNocera PP, Dawid IB (1983) Transient expression of genes intro- duced into cultured cells of Drosophila. Proc Natl Acad Sci USA 80 : 7095-7098

Dobens L, Rudolph K, Berger EM (1991) Ecdysterone regulatory elements function as both transcriptional activators and repres- sors. Mol Cell Biol 11 : 1846-1853

Falb D, Maniatis T (1992) A conserved regulatory unit implicated in tissue-specific gene expression in Drosophila and man. Genes Dev 6: 454-465

Garen A, Kauvar L, Lepesant JA (1977) Roles of ecdysone in Drosophila development. Proc Natl Acad Sci USA 74:5099-5103

Gorman CM, Moffatt LP, Howard BH (1982) Recombinant genomes which express chloramphenical acetyltransferase in mammalian cells. Mol Cell Biol 2:1044-1051

Graupner G, Wills KN, Tzukerman M, Zhang XK, Pfahl M (1989) Dual regulatory role for thyroid-hormone receptors allows con- trol of retinoic-acid receptor activity. Nature 340 : 653-656

Henrich VC (1995) A steroid/thyroid hormone receptor superfamily member in Drosophila melanogaster shares extensive sequence similarity with a mammalian homologue. Nucleic Acids Res 18:4143 4148

Hiromi Y, Gehring WJ (1987) Regulation and function of the Drosophila segmentation gene fashi tarazu. Cell 50:963-974

Jantzen HM, Strahle U, Glass B, Stewart F, Schmid W, Boshart M, Miksicek R, Schutz G (1987) Cooperativity of glucocorticoid

response elements located far upstream of the tyrosine amino- transferase gene. Cell 49 : 29-38

Kato S, Tora J, Yamauchi J, Masushige S, Bellard M, Ghambon P (1992) A far upstream estrogen response element of the oval- bumin gene contains several half-palindromic 5'- TGACC-3' motifs acting synergistically. Cell 68 : 731-742

Koelle MR, Talbot WS, Segaves WA, Bender MT, Cherbas P, Hogness DS (1991) The Drosophila EcR gene encodes an ec- dysone receptor, a new member of the steroid receptor super- family. Cell 67 : 59-77

Laval M, Pourrain F, Deutsch J, Lepesant J-A (1993) In vivo functional characterization of an ecdysone response enhancer in the proximal upstream region of the Fbp-1 gene of D. meIanogas- ter. Mech Dev 44 : 123-138

Lehmann M, Korge G (1995) Ecdysone regulation of the Drosophila Sgs-4 gene is mediated by synergistic action of ecdysone receptor and SEBP 3. EMBO J 14:716-726

Lepesant JA, Kejzlarova-Lepesant J, Garen A (1978) Ecdysone- inducible functions of larval fat bodies in Drosophila. Proc Natl Acad Sci USA 75:5570-5574

Lepesant JA, Levine M, Garen A, Kejzlarova-Lepesant J, Rat L, Somme-Martin G (1982) Developmentally regulated gene ex- pression in Drosophila larval fat bodies. J Mol Appl Genet 1 : 371-383

Lepesant J-A, Maschat F, Kejztarova-Lepesant J, Benes H, Yanicos- tas C (1986) Developmental and ecdysteroid regulation of gene expression in the larval fat body of Drosophila melanogaster. Arch Insect Biochem Physiol sl : 133-141

Lindquist SL, Sonada S, Cox T, Slusser K (1982) Instant medium for Drosophila tissue culture cells. Dros Inf Serv 58 : 163

Luo Y, Amin J, Voellmy R (1991) Ecdysterone receptor is a sequence-specific transcription factor involved in the develop- mental regulation of heat shock genes. Mol Cell Biol 11:3660-3675

Martinez E, Givel F, Wahli W (1991) A common ancestor DNA motif for invertebrate and vertebrate hormone response ele- ments. EMBO J 10:263 268

Maschat F, Dubertret ML, Therond P, Claverie JM, Lepesant JA (1990) Structure of the ecdysone-inducible P1 gene of Drosophila melanogaster. J Mol Biol 214 : 359-372

Miner JN, Yamamoto KR (1991) Regulatory crosstalk at composite response elements. Trends Biochem Sci 16:423M31

Montell D, Rorth P, Spradling AC (1992) slow border cells, a locus required for a developmentally regulated cell migration during oogenesis, encodes Drosophila C/EBP. Cell 71 : 51 62

Nakanishi Y, Garen A (1983) Selective gene expression induced by ecdysterone in cultured fat bodies of Drosophila. Proc Natl Acad Sci USA 80:2971~975

Norton PA, Coffin JM (1985) Bacterial/Lgalactosidase as a marker of Rous sarcoma virus gene expression and replication. Mol Cell Biol 5:281 290

Ozyhar A, Pongs O (1993) Mutational analysis of the interaction between ecdysteroid receptor and its response element. J Steroid Biochem Mol Biol 46 : 135-145

Riddiford LM (1993) Hormones and Drosophila development. In: Bate M, Martinez-Arias A (eds) The Development of Drosophila melanogaster Cold Spring Harbor, Laboratory Press, Cold Spring Harbor, New York pp 899-939

Riddihough G, Pelham HRB (1987) An ecdysone response element in the Drosophila hsp27 promoter. EMBO J 6 : 3729-3734

Rorth P, Montell DJ (1992) Drosophila C/EBP: a tissue-specific DNA-binding protein required from embryonic development. Genes Dev 6:2299-2311

Seed B, Sheen J-Y (1988) A simple phase-extraction assay for chloramphenicol acetyltransferase activity. Gene 67 : 271 277

Shields G, Sang JH (1977) Improved medium for culture of Drosophila embryonic cells. Dros Inf Serv 52 : 161

Shirras AD, Bownes M (1989) CrickIet: a locus regulating a number of adult functions of Drosophila melanogaster. Proc Natl Acad Sci USA 86 : 4559~4563

556

Simon JA, Lis JT (1987) A germline transformation analysis reveals flexibility in the organization of heat shock consensus elements. Nucleic Acids Res 15:2971-2988

Sobrier ML, Chapel S, Couderc JL, Micard D, Lecher P, Somme- Martin G, Dastuge B (1989) 20-OH-ecdysone regulates 60C beta tubulin gene expression in Kc cells and during Drosophila devel- opment. Exp Cell Res 184:241-249

Talbot WS, Swyryd EA, Hogness DS (1993) Drosophila tissues with different metamorphic responses to ecdysone express different ecdysone receptor isoforms. Cell 73 : 1323 1337

Teller WH, Kunkel JG (1991) The function and evolution of insect storage hexamers. Annu Rev Entomol 36:205 228

Thomas HE, Stunnenberg HG, Stewart AF (1993) Heterodimeriz- ation of the Drosophila ecdysone receptor with retinoid X recep- tor and ultraspiracle. Nature (London) 362:471-475

Thomasson WA, Mitchell HK (1972) Hormonal control of protein granule accumulation in fat bodies of Drosophila melanogaster larvae. J Insect Physiol 18:1885-1889

Thummel CS, Boulet AM, Lipshitz HD (1988) Vectors for Drosophila P-element-mediated transformation and tissue cul- ture transfection. Gene 74:445-456

Truss M, Beato M (1993) Steroid hormone receptors: interaction with deoxyribonucleic acid and transcription factors. Endocrinc Rev 14 : 459-479

Yao T-P, Segraves WA, Oro AE, McKeown M, Evans RM (1992) Drosophila uItraspiracIe modulates ecdysone receptor function via heterodimer formation. Cell 71:63-72

Yao T-P, Forman BM, Jiang Z, Cherbas L, Chen J-D, McKeown M, Cherbas P, Evans RM (1993) Functional ecdysone receptor is the product of EcR and Ultraspiracle genes. Nature 366 : 476-479