Insect Biochem. Molec. Biol. Vol. 23, No. 1, pp. 105 114, 1993 0965-1748/93 $6.00 + 0.00 Printed in Great Britain. All rights reserved Copyright © 1993 Pergamon Press Ltd

Structural Features Critical to the Activity of Ecdysone Receptor Binding Site CHRISTOPHE ANTONIEWSKI,* MONIQUE LAVAL,* JEAN-ANTOINE LEPESANT*t

a n

Two ecdysone-response elements from the hsp27 (hsp27 EcRE) and the Fbp 1 (D EcRE) genes of Drosophila melanogaster were used as probes in a gel shift assay to investigate the interactions of the ecdysone receptor (EcR) with its cognate DNA response element. The source of EcR was a nuclear extract from the late third-larval instar fat body. The hsp27 and D EcREs share a sequence similarity at 12 positions over a 15bp region including an imperfect palindromic structure consisting of two pentamer half-sites separated by a single intervening nucleotide. We have shown that a short oligonucleotide containing this l lbp imperfect palindrome of the hsp27 EcRE and three flanking bp on each side is an efficient EcR binding site. Mutational analysis confirms that the integrity of both these half-sites as well as their lbp spacing are critical for binding of the ecdysone receptor. The D EcRE behaved as a much weaker EcR binding site than the hsp27 EcRE but a single bp substitution was sufficient to confer upon it a binding capacity equivalent to that of the hsp27 EcRE. These results have led us to propose the sequence PuG(G[T)T(C/G)A(N)TG(C/A)(C/A)(C[t)Py as a revised version of a previously proposed EcRE consensus sequence.

Ecdysone receptor Ecdysone-response element D. melanogaster Fat body nuclear extract Gel shift assay

INTRODUCTION

Over the past few years, a wealth of information has been gathered on the structure and function of the various members of the vertebrate nuclear receptor superfamily, as well as on the definition of the DNA sequences which act as hormone response elements (HRE) (for reviews see Evans, 1988; Green and Cham- bon, 1988; Beato, 1989; Parker, 1991). By comparison the mechanism of action of steroid receptors in invert- ebrates is much less well documented. The cloning of a Drosophila melanogaster gene coding for an ecdysone receptor (EcR) belonging to the nuclear receptor super- family has been reported only recently (Koelle et al., 1991). A large number of ecdysone-inducible genes have been cloned and characterized, but analysis of cis-acting regulatory sequences has been performed on relatively few of them (Segraves and Richards, 1990; Andres and Thummel, 1992). The first identification of an ecdysone- response element (EcRE) was provided by Riddihough and Pelham (1987) who demonstrated that a 23bp sequence from the hsp27 gene promoter was necessary and sufficient to confer ecdysone inducibility on an heterologous reporter gene in a Drosophila cell transfec- tion assay. Since this finding, several authors have confirmed that this 23bp element acts as a functional EcRE in similar transfection assays (Str/ihle et al., 1988;

*Institut Jacques Monod, CNRS et Universit6 Paris 7, 2, place Jussieu, 75251 Paris Cedex 05, France.

Cherbas et al., 1991; Dobens et al., 1991). More recently, biochemical purification of an ecdysone receptor was carried out by oligonucleotide affinity chromatography using a 33bp DNA fragment which included this 23bp EcRE from hsp27 (Luo et al., 1991).

Using a cell transfection assay to test the effect of a series of base pair changes, Martinez et al. (1991) demonstrated that within the 23bp hsp27 EcRE the imperfect palindromic structure, GGTTCAATGCACT, with a 1 bp spacing between the half-sites, is critical for the ecdysone response (Fig. 1). The importance of this structure for in vitro EcR binding has also been demon- strated by Ozyhar et al. (1991) using a partially purified embryonic extract as a source of EcR. The use of high specific activity radiolabelled Ponasterone A allowed Cherbas et al. (1991) to devise a new technique to detect binding of the hormone-labelled receptor to filter-bound DNA fragments. This led them to identify three EcR binding sites within the 5' and 3' flanking sequences of the ecdysone-inducible Eip28/29 gene and to derive a consensus sequence, RG(G/T)TCANTGA(C/A)CPy, for an EcR binding site (Cherbas et al., 1991). More degenerate but similar consensus sequences have been proposed by comparing the hsp27 EcRE with putative'EcREs from other ecdysone-inducible genes (Dobens et aL, 1991; Luo et al., 1991). Martinez et al. (1991) derived an idealized EcRE sequence (A/C)GGTCANTGACCT from a mutational analysis of the hsp27 EcRE. These results strongly suggest that EcREs present a high structural similarity to the

105

106 CHRISTOPHE ANTONIEWSKI et al.

-85 cgatt IG G G T T G A A T G A tgctg -1o9 Fbpl EcRE

-7 -6 -5 -4 -3 -2 -I 0 +i +2 +3 +4 +5 +6 +7

FIGURE 1. Sequence comparison of the hsp27 and D EcREs. The nucleotide sequences of hsp27 and D EcRE are numbered relative to the transcription start site (+ 1) of the hsp27 (Riddihough and Pelham, 1986) and Fbp 1 gene (Maschat et al., 1990), respectively. The positions of nucleotides relative to the imperfect palindromic structure of the hsp27 EcRE (arrows) are given

below the sequences.

vertebrate HR E sequences and are in agreement with the finding that their respective cognate receptors belong to a family of transcription factors that have evolved from a common ancestor (for a review see Amero et al., 1992). However, experimental analysis of the in vivo and in vitro functional properties of an EcRE has only been performed on the hsp27 EcRE taken as a canonical EcRE.

We have recently identified an EcRE in the proximal upstream region of the ecdysone-inducible Fbpl gene whose transcription is restricted to the fat body, of late third-instar larvae of D. melanogaster (Lepesant et al., 1982, 1986). Using a germ-line transformation assay we have demonstrated that Fbpl sequences between - 1 3 8 and - 6 8 relative to the transcription start behave as a hormone-inducible enhancer element in that they can drive specifically the ecdysone-dependent expression of a heterologous reporter gene in the fat body of late third-instar larvae (Laval et al., in preparation). We have shown that several factors bind to this 70bp enhancer using fat body nuclear extracts of third-instar larvae in a gel retardation assay (Antoniewski et al., in prep- aration). One of these factors binds to the central region ( - 8 5 to - 109 ) of this enhancer, hereafter referred to as the D region. The same factor also binds to the hsp27 EcRE (Fig. 1). Using anti-EcR antibodies (Koelle et al., 1991) we have demonstrated that this factor contains at least an ecdysone receptor (Antoniewski et al., in prep- aration). The D region thus exhibits all the in vivo and in vitro characteristics of an EcRE (D EcRE).

Sequence comparison shows that the hsp27 and D EcREs share an extensive sequence similarity at 12 positions over a 15bp region including the imperfect palindromic structure described above (Fig. 1).

These results prompted us to study the binding inter- actions of the EcR with both the hsp27 and D EcREs and to perform a mutational analysis in order to further investigate the structural features of an EcR binding site.

M A T E R I A L S A N D M E T H O D S

Mass preparation of fa t body from third-&star larvae

Larvae emerging from a 6 h egg collection of the D. melanogaster Canton S stock were reared in M medium (100 g/1 dried yeast, 120 gl sucrose, 17.5 g/1 agar, 8.3 g/1 K2HPO4, 9 g/1 KH2PO4, 1.76 g/1 methyl hydroxy-4 ben- zoate, 0.8% propionic acid) at 25°C. When larvae reached the wandering stage, 120 + 8 h after egg laying,

they were collected in water, floated away from the contaminating growth medium by suspension in 2 M NaC1 solution and washed again with water through a 600#m mesh. From this point all operations were carried out at 4°C with media chilled to this temperature. For one standard fat body preparation, 200 g of third- instar larvae were ground with a motor-driven grinding mill, under a continuous flow of FB extraction medium (20 mM sodium glycerophosphate, 10 mM KHzPO4, 30 mM KC1, 0.17 M sucrose) from a squeezable bottle. The grindate was passed through a 600 # m mesh. Tissue fragments in suspension in the filtrate were collected on a 100/~ m mesh and transferred to centrifuge tubes with the aid of a squirt of FB extraction medium. After centrifugation (4mn, 3000g) floating fat body pieces were collected with a truncated pipetteman tip and put into 1.5 ml Eppendorf tubes. After centrifugation (1 mn, 12,000g), the excess of FB medium was carefully re- moved with a Pasteur pipette. Using this procedure, Ca 2 4 ml of purified fat bodies were obtained per 200 g of larvae.

Fat body nuclear extract

All steps were performed at 4°C. Fat bodies (usually 2 4 ml) were homogenized in a Dounce homogenizer in 2 vol of buffer A (10 mM Hepes pH 7.9, 10 mM KC1, 1.5mM MgC12, 0.5raM DTT, 0.1 mM EGTA pH 8) by 20-30 strokes of pestle B. 1 vol of buffer A supplemented with 1.2 M sucrose was added and 10 additional strokes were applied. The homogenate was centrifuged (15 min, 5000 rpm Sorvall HB4 rotor). The lipid layer was re- moved with a tissue, and the supernatant was decanted with a pipette. The pellet was resuspended first in 3 vol of buffer A with 0.3 M sucrose and 3 vol of buffer A with 1.7 M sucrose were added. The homogenate was trans- ferred to a 1.5ml Eppendorf tube and centrifuged (15 mn, 12,000g). The pellet was suspended in 3 vol of buffer B (10mM Hepes pH 7.9, 0.6 M NaC1, 1.5mM MgC12, 5% glycerol, 0.5 mM DTT, 0.1 mM EGTA pH 8) and gently mixed on a rotary wheel for 30 min. After centrifugation (30min 12,000g), the supernatant was dialysed overnight against 1 litre of buffer C (20 mM Hepes pH 7.9, 75 mM NaC1, 20% glycerol, 0.5 mM DTT, 0.1 mM EDTA pH 8, 0.5 mM PMSF). Buffers A, A-0.3 M sucrose, A-1.2 M sucrose, A-1.7 M sucrose and B were supplemented with 0.5 mM PMSF and 0.5 mg/ml of leupeptine, chymostatin, aprotinin, antipain and pep- statin just before use.

ECDYSONE RECEPTOR BINDING SITE 107

Gel shift assay

3-5 #g of protein extract, 2/xg of poly(dl-dC) and specific competitor DNA, if appropriate, were mixed 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 DTT) in a final volume of 10 #1 and preincubated for 15 min on ice. After addition of 2 fmol of 32p-5'-end- labelled probe (2.5-7.5 x 104 cpm/fmol), incubation was continued for 15 min at 4°C. Free and complexed DNAs were separated at 4°C in a low-ionic strength 4% polyacrylamide gel (39:1 cross linking ratio) containing 25mM Tris-base, 190mM glycine, 1 mM EDTA and 2.5% glycerol which had been prerun (25 mA, 1 h). After electrophoresis (25 mA, 3 h), the gel was dried and autoradiographed.

Oligonucleotides were synthesized with a DNA syn- thesizer (Pharmacia) and purified as recommended by the manufacturer. Probes and competitor oligonucle- otides as listed in Table 1 were obtained by annealing of complementary strands.

All gel-shift experiments were repeated once. Quanti- tative analysis of complexes was performed by scanning autoradiographs with an Imstar (Paris, France) image analyser (512 x 512 pixels, 256 grey levels).

RESULTS

The ecdysone receptor has a lower binding affinity for the D EeRE than for the hsp27 EcRE

The gel retardation assay was used to analyse the interactions of the ecdysone receptor (EcR) with the hsp27 and D EcREs (Fig. 1). Using anti-EcR antibodies in separate experiments (Antoniewski et al., in prep- aration) we have shown that the two complexes detected with these EcREs as radioactive probes and a third- larval instar fat body nuclear extract are EcR-EcRE complexes (EcRCs) (Fig. 2, lanes 1 and 8). As will be discussed, the reason for the formation of these two different complexes is unclear, but competition exper- iments using various concentrations of the correspond- ing cold oligonucleotide as a homologous competitor show that both complexes are sequence-specific (Fig. 2, lanes 2 and 12). It is notable that in these experiments a 90% inhibition of EcRC formation is obtained with a 50-fold excess of cold hsp27 EcRE (Fig. 2, lane 2) whilst a 200-400-fold molar excess of cold D EcRE is required to achieve the same level of competition (Fig. 2, lane 12). This may reflect a higher binding affinity of the EcR for the hsp27 as opposed to the D

TABLE 1. Structure of the hsp27 EcRE, D EcRE and synthetic variant oligonucleotides. Bold-face small letters represent mutant positions. Sequences are aligned with the imperfect palindromic structure of the hsp27 EcRE. Base

positions with regard to this structure are indicated at the top

HSP27 EcRE

HSP(-2)

HSP(-5,-6)

HSP(+2)

HSP(17bp)

D EcRE

D(+2)

D(+5)

D(+3,+4,+5,+6)

- 6 - 5 - 4 - 3 - 2 - 1 ; ~1 ",2 ',3~'4 ~'5",6

AGACAAC~~~CAA i i

AGACAAC933Tf gA~CCACITGTCCAA I i

~E~CA AG aa T I C_A ATC_~_AC T l~%v_~/<A i

AGACAAC~T cCACTII3%I3C3iA i i

i i i i o

c c 4 x ~ ~ ~ i

i

CC~~%~AAc ~YG i

CGA~TGccgg ~

D 3

D 3 ( 0 , - 1 , - 2 , - 3 )

D 3 ( 0 )

D 3 ( i n s )

TIE~CTCCCGATIE<~TTGA/~TGAATrTIV~TG

TIC_~CTCCCC~~ gt ce ~ . ~ A ~ i

i

TIV:ACTCCCGA~TIC_~AATITIUCTG , + i i

TIC~CTCCCGATICP~TIC4~tg ~AATITICCTG o

i

i

108 CHRISTOPHE A N T O N I E W S K I e t al.

E 0 0

0 (..-

UA Ct" W 0 LLI W I.IJ t'r" HJ W W

uJ w O~ n- ~ o uJ O~ n" n ~ n - o o o w n- o o o

c~. ~ w w w ~ . o LU W W W n w a r7 t'7

~- a o o o o o ° r~ o o o 0 0 0 0 0 0 0 0 0 0 LO LO .,-- 0,1 ~ "~ 0 LO "~-- C,,I ~.. X X X X X CS X X X X

2 3 4 5 6 7 8 9 1 0 1 1 12

100 ~ 100 90 ~ 90

x 80 ~ 80 oE~- 70 ~ 70

~ o o 60 ~ oo 60

t~ o 50 r ~ 50 ~. 40 a. ea ~ . ~ 40 ~ 30 ~ ~ 30

20 ~ 20 10 '~ 10 8 O=

-'= 0 ~ 0

1 2 3 4 5 6 7 8 9 10 11 12

F I G U R E 2. Analysis of the binding of the ecdysone receptor to the hsp27 and D EcREs by gel retardation assay. The hsp27 EcRE (lanes 1-7) and the D EcRE (lanes 8 12) were used as radioactive probes (see Table 1 and Fig. 1 for sequences). Competitors and molar ratios of competitor to probe are indicated above each lane. Quantitative analysis of the complexes is shown at the bot tom as a percentage of the total quantity of complexes formed in the presence of either the hsp27 (1-7)

or the D EcRE (8 12) probe in the absence of competitor (lanes 1 and 8).

ECDYSONE RECEPTOR BINDING SITE 109

EcRE. Two lines of evidence strongly support this hypothesis. First the relative amount of EcRC formed with the hsp27 probe is much higher than that formed with the D probe (Fig. 2, lanes 1 and 8). Furthermore the D EcRE behaves as a much weaker competitor (Fig. 2, lanes 3-7) than the homologous hsp27 sequence (Fig. 2, lane 2) for EcRC formation with the hsp27 probe. It can be noted that contrary to what one might have expected, the EcRCs formed with the 25bp D probe migrate slightly more slowly than the ones formed with the 27bp hsp27 probe. However, we have observed that EcRCs formed with the 33bp D3 probe (Table 1) migrate faster than those formed with the D probe (data not shown). This suggests that adding flanking sequences to the EcR binding site may influence the conformation of the complexes and make them migrate faster.

Structural features of the hsp27 and D EcREs critical to the formation of EeRCs

In order to investigate the critical bases required for EcR binding, we performed competition experiments with either the hsp27 or D EcRE as a labelled probe and various oligonucleotides corresponding to wild-type and mutated sequences of either EcRE as competitors (Table 1). Mutations of the D EcRE were also introduced into

the 33bp D3 oligonucleotide which includes the D EcRE and can hence be considered as a wild-type competitor. Each half-palindromic site was mutated in the hsp27 [ h s p 2 7 ( - 5 , - 6 ) , h sp27( -2 ) and hsp27(+2) ] and D [ D 3 ( - 3 , - 2 , - 1 , 0 ) , D ( + 2 ) and D ( + 4 , +5 , +6, +7)] EcREs. The importance of the identity of the central base pair [D3(0)] and the spacing between the half-sites [D3(ins)] were also examined. A shorter version of the hsp27 EcRE [hsp(17bp)] including only the l lbp imper- fect palindrome and three flanking bp on each side was also tested. In all these experiments the concentration of competitors was adjusted in order to take into account the different affinities of the EcR for each EcRE type (Figs 3 and 4).

Results obtained with the labelled hsp27 EcRE as a probe (Fig. 3) can be summarized as follows. (i) A 50-fold molar excess of hsp27 EcRE, hsp(17bp) or hsp27( -2 ) results in 100% competition whilst the same excess of hsp27( - 5, - 6) and hsp27(+ 2) results only in a 30-40% competition. (ii) A 400-fold molar excess of the wild-type oligonucleotides D and D3 results in 80-90% competition but the mutated forms D (+4 , +5, +6, +7), D(+2) , D 3 ( - 3 , - 2 , - 1 , 0 ) and D3(ins) give only 20-30% competition. (iii) A 400-fold molar excess of the mutated form D3(0) results in an intermediate value of 60% competition.

uJ +

1 2 3 4 5 6 7 8 9 10 11 12 13

l o o

90

8 8O

70

60 "10 ~= 5 o o

40

3O

2o

10

o 1 2 3 4 5 6 7 8 9 10 11 12 13

FIGURE 3. Binding specificity of the ecdysone receptor to the hsp27 EcRE. Gel retardation assays were performed in the presence of the hsp27 EcRE radioactive probe and competitor DNAs as indicated above each lane (see Table 1 for sequences). The molar ratio of competitor to probe was 50 for lanes 1-5 and 400 for lanes 6-12. The free probe is not shown. Quantitative analysis of the complexes is shown at the bottom as a percentage of the total quantity of complex formed with the hsp27 EcRE

probe in the absence of competitor (lane 13).

I10

kkl

O Q-

A2~

1 2

CHRISTOPHE ANTONIEWSKI et al.

3

G" i

_c: E3

4 5 6

i

+ ,m

o o

d d 8 ~ 8 B o

7 8 9 10 11 12 13

1 0 0

90

8O

8 7O

60

50

= 40 o

30

20 c',

lO

0

1 2 3 4 5 6 7 8 9 10 11 12 13

FIGURE 4. Binding specificity of the ecdysone receptor to the D EcRE. Gel retardation assays were performed in the presence of the radioactive D EcRE probe and competitor DNAs as indicated above each lane (see Table 1 for sequences). The molar ratios of competitor to probe were 25 for lanes I-5 and 400 for lanes 6-12. The free probe is not shown. Quantitative analysis of the complexes is shown at the bottom as a percentage of the total quantity of complexes formed with the D EcRE probe

in the absence of competitor (lane 13).

Competit ion data obtained with the D EcRE as a labelled probe are shown in Fig. 4. They parallel those obtained with the hsp27 EcRE (Fig. 3) as expected from the fact that the D and hsp27 probes bind the same factor.

In sum, these results demonstrate that a 17bp sequence is an efficient EcR binding site whose palin- dromic structural integrity is required for full activity.

A single base-pair substitution confers a high EcR binding affinity upon the D EcRE

Since the hsp27 and D EcREs diverge at positions - 2, + 3 and + 5 (Fig. 1) any of these nucleotide changes could be responsible for their large differences in EcR binding affinity. The divergence at position - 2 can be excluded because substitution of a C by a G at this position in the hsp27 EcRE [hsp( -2) ] has no effect on EcR binding (Figs 3 and 4). Divergence at position ÷ 5 was eliminated by substituting a C for the wild-type T in the D EcRE. The mutated sequence D ( + 5) was used as a competitor in a gel-shift experiment conducted with the hsp27 EcRE as a probe. While a 400-fold molar excess of D EcRE was necessary to obtain a 90% competition, a 50-fold molar excess of D ( + 5) EcRE is now sufficient to achieve > 95% competition, indicating

that the EcR has a similar binding affinity for the D ( + 5) and hsp27 EcREs [Fig. 5(A)]. This is confirmed by the fact that the amount of EcRC formed with the D ( + 5) EcRE is similar to that formed with the hsp27 EcRE if differences in the specific activities of the probes are taken into account [Fig. 5(B)]. Preliminary results from competition experiments using a range of concentrations of D ( + 5 ) or hsp27 EcRE indicate that the binding affinity of the EcR for the D ( + 5) oligonucleotide is slightly higher than that for the hsp27 oligonuclcotide (data not shown). The effect of the divergence at position + 3 was not tested.

DISCUSSION

The palindromic structure of the EcRE

Extensive studies of mammalian systems have shown that most of the D N A hormone-response elements (HREs) consist of two half-sites arranged as direct or inverted repeats separated by variable spacing. Both the sequence of the half-sites and their spacing determine the specificity of the receptor binding site [for a review see Martinez and Wahli (1991)].

ECDYSONE RECEPTOR BINDING SITE 111

A

hsp 27 EcRE probe ! I

LLI

LLI LLI ".~ LLI n" Q- OJ cO I l l LQ E " t u ,2. + c - ' l ° ~ a o 0 0

0 0 0 0 0 t ~ ~'0 ~ " LO C X X X X

1 2 3 4 5

..Q o

rr e U_I U.I r - - n -

Q . UA

r - C~

0) ..Q

e

+

D

B

100

90 ~ ~°

70

~ 60 ~ so

m.

40

30

20

10

1 2 3 4 5

FIGURE 5. Binding affinity of the ecdysone receptor to the oligonucleotide D(÷5) . (A) Formation of ecdysone receptor complexes in the presence of the hsp27 EcRE (lanes 1~6), D EcRE (lane 7) or D ( + 5 ) (lane 8) probe was analysed by gel retardation assay. Where competitor was added, the molar ratios of competitor to probe are indicated above each lane. Specific activities of the probes were 3.9 × 104 cpm/fmol (hsp27 EcRE), 7.3 × 104 cpm/fmol (D EcRE) and 7.7 x 104 cpm/fmol [D(+ 5)], respectively. (B) Quantification of complexes formed in lanes 1-5 as a percentage of the total quantity of complexes formed

with the hsp27 EcRE in the absence of competitor (lane 1).

112 CHRISTOPHE ANTONIEWSKI et al.

The 27bp hsp27 and the 25bp D EcREs include an imperfect palindromic structure with a lbp spacing of the two half-sites (Fig. 1). Our finding that the oligonu- cleotide hsp(17bp) competes as efficiently as the 27bp hsp27 EcRE for EcRC formation provides evidence that a strong in vitro EcR binding site requires no more than the central 1 l bp imperfect palindromic structure plus three flanking bp. This conclusion is in agreement with results obtained in vivo by Cherbas et al. (1991) showing that a 15bp fragment from the hsp27 promoter including this imperfect palindrome is sufficient to direct ecdysone- inducible expression of a CAT reporter gene in a cell transfection assay. It is worth noting that while three flanking bp are sufficient for in vitro efficient EcR binding, optimal binding of the glucocorticoid receptor (GR) to its 15bp response element (GRE) requires 8-10 flanking bp (Chalepakis et al., 1990).

The lbp spacing between the two inverted repeats appears to be critical for binding of the EcR since its extension to 3bp in oligonucleotide D3(ins) strongly reduces EcR binding. In addition the identity of the lbp spacing has some importance as we showed that oligonu- cleotide D3(0), in which the wild-type central nucleotide A has been substituted by a G, competes for only 60% of the EcRCs formation. This result is comparable to that obtained by Nordeen et al. (1990) showing that a change at position 0 in the glucocorticoid-response element weakens its activity.

In this study, we have mutated each of the two half-sites in both the hsp27 and D EcREs. Mutations of the wild-type hsp27 sequence at positions (+2) or ( - 5 , - 6 ) strongly reduce EcR binding. Mutations that affect either positions ( - 3 , - 2 , - 1 , 0 ) , (+2) or (÷ 4, + 5, ÷ 6, ÷ 7) in the D EcRE have the same effect. Moreover, we have shown that a mutated 70bp Fbp 1 enhancer containing a D EcRE altered at positions ( - 4 , - 5 , - 6 , - 7 ) no longer binds the EcR (An- toniewski et al., in preparation). These results demon- strate that the integrity of each half-site is necessary for EcR binding. They are in agreement with those of Martinez et al. (1991) and Ozyhar et al. (1991) and provide additional evidence that an EcRE consists of inverted repeats separated by a lbp spacing, a structure highly similar to that of vertebrates HREs.

The dyad symmetry of the EcREs suggests that the EcR binds as a homodimer, with each receptor molecule recognizing an arm of the EcRE, as is the case for other members of the nuclear receptor superfamily (Wrange et a!., 1986; Kumar and Chambon, 1988; Tsai et al., 1988; H/ird et al., 1990; Schwabe et al., 1990; Luisi et al., 1991). Using a chromatographic assay, Ozyhar et al. (1991) provided evidence that a partially purified EcR binds to the hsp27 EcRE as a dimer. Binding of a homodimer is not, however, an absolute rule among members of the nuclear receptor family. For example, receptors for thyroid hormone (T3R) and retinoic acid (RAR) bind to half-sites as monomers and, depending on the orien- tation and distance between the half-sites, bind either as T3D and RAR homodimers or as T3R-RAR het-

erodimers (Lazar et al., 1991; Forman et al., 1992). In our retardation assay, both the hsp27 and D EcREs give rise to two distinct migrating EcRCs. Apart from a specific proteolysis of the slow migrating EcRC resulting in a faster migrating one, there are a nmnber of other possible explanations. The slow migrating complex could correspond to a dimeric form of the EcR bound to both half-sites and the fast migrating complex to a monomeric form of the EcR bound to only one of the two half-sites. We can eliminate this hypothesis because mutations altering either half-site of the hsp27 or D EcRE inhibit, partially or completely (but always to the same extent), the formation of both the EcRCs.

On the other hand, the EcR gene has been shown to encode three EcR isoforms, A, B1 and B2, sharing the same DNA binding domain (M. Koelle and D. Hogness, pers. commun.). The two EcRCs could thus be due either to the binding of two different homodimers or to two heterodimers of EcR isoforms present in the fat body nuclear extracts we used. Alternatively, they might also result from interactions between EcR(s) and other fat body nuclear factors.

Critical nucleotides for EcR binding affinity and the definition o f an E c R E sequence

Do some positions within the half-sites need to be strictly conserved for EcR binding? Guanines at pos- itions - 6, - 5, and + 2 are conserved between the hsp27 and D EcREs. Mutations affecting positions (+2) or ( - 5, - 6) in both the hsp27 and D EcREs result in a loss of EcR binding suggesting that a G is important at these positions. Further evidence supports this conclusion. First, methylation interference experiments (An- toniewski et al., in preparation) indicate that methyl- ation of these three guanines interferes with EcR binding. Secondly, using a gel retardation assay with a partially purified EcR from embryonic extracts, Ozyhar and Pongs (1992) have found that point mutations at positions - 6, - 5 or + 2 of the hsp27 EcRE inhibit EcR binding. Positions - 1 , - 3 , - 4 , +1, +4 and +6 are also conserved between the hsp27 and D EcREs. Oligonucleotides D3(-3 , - 2 , - 1, 0) and D(+4, +5, +6, +7) failed to compete for EcRC for- mation, suggesting that the nature of some of these nucleotides is also critical for EcR binding. The individ- ual importance of these positions cannot be assessed yet because we did not perform a systematic point mutation analysis of the two EcREs.

Our results show that the EcR binding affinities for the hsp27 and D EcREs are very different. The imperfect hsp27 and D palindromes diverge at positions - 2 , + 3 and + 5. Ozyhar and Pongs (1992) report that changing the C to an A at position - 2 of the hsp27 EcRE results in a loss of EcR binding. By contrast, our results clearly show that substitution of this C by a G, has no effect. Altogether, these results indicate that the palindromic EcRE sequence can possess at position --2 either a C or a G but not an A, without any change of EcR binding affinity. Since the D EcRE contains a G at position -2 ,

ECDYSONE RECEPTOR BINDING SITE 113

we can conc lude tha t d ivergence at this pos i t ion does not expla in the difference o f E c R affinity for the two EcREs . The E c R E s also diverge at pos i t ion + 3 wi th a C in the hsp27 E c R E and an A in the D EcRE. Ozyha r and Pongs (1992) r epor t tha t changing this C into an A in the hsp27 E c R E does no t affect E c R binding. On the o ther hand this pos i t ion is given as an A in the consensus E e R E sequences deduced by different au thor s (Cherbas et al.,

1991; D o b e n s et al., 1991; Mar t i nez et al., 1991). The divergence at pos i t ion + 3 is thus p r o b a b l y no t respon- sible for the weaker b ind ing affinity o f the E c R for the D EcRE. By cont ras t , the last d ivergent pos i t ion , + 5, appea r s to be crit ical . R e m a r k a b l y the subs t i tu t ion in the D E c R E of the + 5 T by a C which el iminates this d ivergence and increases the c o m p l e m e n t a r i t y o f its half-si tes results in an E c R b ind ing site as s t rong as the hsp27 EcRE.

Our in vitro bind ing studies a l low us to p rop ose the fo l lowing revised version, P u G ( G / T ) T ( C / G ) A ( N ) T G - (C /A) (C/A) (C/ t )Py , o f the E c R E consensus sequence, R G ( G / T ) T C A N T G A ( C / A ) C P y , first p r o p o s e d by Che rbas et al. (1991). This new consensus takes into account var ia t ions at pos i t ions - 4, - 2, + 3, + 4, which have no de tec tab le effect on E c R E activity. The small t at pos i t ion + 5 indicates a change to a weaker EcR b ind ing site. Requ i r emen t for the centra l bp awai ts fur ther exper iments because only the effect o f the pres- ence o f an A or a G in the context o f the D E c R E has been tested at this pos i t ion .

The sequence flexibili ty reflected in this consensus implies a r a the r large degree o f divergence a m o n g the E c R E s and an equal ly large range o f affinities o f the ecdysone recep tor for these elements. These as sumpt ions which are still based on the charac te r i za t ion o f the in vivo

and in vitro proper t ies o f a relat ively small n u m b e r o f EcREs , m a y have i m p o r t a n t func t iona l impl ica t ions . T h r o u g h o u t the life cycle o f the fly and mos t impor t an t l y dur ing tha t pe r iod o f the th i rd larval ins tar preceding me tamorphos i s , large var ia t ions in the ecdysone t i tre and level o f the ecdysone receptor are observed; in th i rd - ins ta r la rvae such var ia t ions have been demon- s t ra ted to be associa ted with a sequent ia l induc t ion o f a n u m b e r o f ecdysone- induc ib le genes (Ashburne r et al.,

1974; Thummel , 1990; Andre s and Thummel , 1992). The existence o f EcREs with very different EcR-b ind ing affinities cou ld p rov ide a s imple mechan i sm by which the ac tua l level o f a ho rmone -ac t i va t ed recep tor species in a given cell de te rmines the stage- and tissue-specific gene response to a cons tan t ly changing ecdysone titre. In add i t i on synergist ic p r o t e i n - p r o t e i n in terac t ions o f the E c R with t issue-specific and stage-specific t r ansc r ip t ion fac tors m a y also be involved in the fine ad jus tmen t o f this response in a dynamic way.

In this respect , it is in teres t ing to note tha t the D E c R E we have charac te r ized in this s tudy as a weak E c R b ind ing site is inc luded in the 70bp Fbp 1 enhancer which is involved in the ecdysone-dependen t and tissue- specific t r ansc r ip t ion o f this gene at the end o f the th i rd larval ins ta r (Laval et al., in p repara t ion) . W e have

shown tha t several o ther fat b o d y nuclear fac tors b ind to this enhancer (Anton iewski et al., in p r epa ra t i on ) and exper iments are in progress to invest igate their pu ta t ive regu la to ry role and re la t ionship with the ecdysone receptor .

REFERENCES

Amero S. A., Kretsinger R. H., Moncrief N. D., Yamamoto K. R. and Pearson W. R. (1992) The origin of nuclear receptor proteins: a single precursor distinct from other transcription factors. Molec. Endocr. 6, 3-7.

Andres A. J. and Thummel C. S. (1992) Hormones, puffs and flies: the molecular control of metamorphosis by ecdysone. Trends Genet. 8, 132-138.

Ashburner M., Chihara C., Meltzer P. and Richards G. (1974) Temporal control of puffing activity in polytene chromosomes. Cold Spring Harbor symp. quant. Biol. 38, 655~62.

Beato M. (1989) Gene regulation by steroid hormones. Cell 56, 335 344.

Chalepakis G., Shauer M., Cao X. and Beato M. (1990) Efficient binding of glucocorticoid receptor to its responsive element re- quires a dimer and DNA flanking sequences. DNA cell Biol. 9, 355-368.

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

Dobens L., Rudolph K. and Berger E. (1991) Ecdysterone regulatory elements function as both transcriptional activators and repressors. MoL cell. Biol. 11, 1846-1853.

Evans R. M. (1988) The steroid and thyroid hormone receptor superfamily. Science 240, 88%895.

Forman B. M., Casanova J., Raaka B. M., Ghysdael J. and Samuels H. H. (1992) Half-site spacing and orientation determine whether thyroid hormone and retinoic acid receptors and related factors bind to DNA response elements as monomers, homodimers, or het- erodimers. Molec. Endoer. 6, 429442.

Green S. and Chambon P. (1988) Nuclear receptors enhance our understanding of transcription regulation. Trends Genet. 4, 309-313.

H~ird T., Kellenbach E., Boelens R., Maler B. A., Dahlman K., Freedman L, P., Carlstedt-Duke J., Yamamoto K. R., Gustafsson J.-A. and Kaptein R. (1990) Solution structure of the glucocorticoid receptor DNA-binding domain. Science 249, 157-160.

Koelle M. R., Talbot W. S., Segraves W. A., Bender M. T., Cherbas P. and Hogness D. S. (1991) The Drosophila EcR gene encodes an ecdysone receptor, a new member of the steroid receptor superfamily. Cell 67, 59-77.

Kumar V. and Chambon P. (1988) The estrogen receptor binds tightly to its responsive element as a ligand-induced homodimer. Cell 55, 145-156.

Lazar M. A., Berrodin T. J. and Harding H. P. (1991) Differential DNA binding by monomeric, homodimeric, and potentially het- eromeric forms of the thyroid hormone receptor. Mol. cell. Biol. 11, 5005-5015.

Lepesant J. A., Maschat F., Kejzlarovfi-Lepesant J., Benes H. and Yanicostas C. (1986) Developmental and ecdysteroid regulation of gene expression in the larval fat body of Drosophila melanogaster. Archs Insect. Biochem. Physiol. I, 133-141.

Lepesant J. A., Levine M., Garen A., Lepesant-Kejzlarov~ J., Rat L. and Somm6-Martin G. (1982) Developmentally regulated gene expression in Drosophila larval fat bodies. J. molec, appl. Genet. 1, 371-383.

Luisi B. F., Xu W. X., Otwinowski Z., Freedman L. P., Yamamoto K. R. and Sigler P. B. (1991) Crystallographic analysis of the interaction of the glucocorticoid receptor with DNA. Nature 352, 497-505.

Luo Y., Amin J. and Voellmy R. (1991) Ecdysterone receptor is a sequence-specific transcriptionfactor involved in the developmental regulation of heat shock genes. Mol. cell. Biol. 11, 3660-3675.

114 CHRISTOPHE ANTONIEWSKI et al.

Martinez E. and Wahli W. (1991) Characterization of hormone response elements. In Nuclear Hormone Receptor, Molecular Mech- anisms, Cellular Functions, Clinical Abnormalities (Edited by Parker M. G.), pp. 125-153. Academic Press, San Diego.

Martinez E., Givel F. and Wahli W. (1991) A common ancestor DNA motif for invertebrate and vertebrate hormone response elements. EMBO J. 10, 263568.

Maschat F., Dubertret M. L., Th6rond P., Claverie J. M. and Lepesant J. A. (1990) Structure of the ecdysone-inducible P1 gene of Drosophila melanogaster J. molec, biol. 214, 359072.

Nordeen S. K., Sub B. J., Kiihnel B. and Hutchison III C. A. (1990) Structural determinants of a glucocorticoid receptor recognition element. Molec. Endocr. 4, 1866-1873.

Ozyhar A. and Pongs O. (1992) Mutational analysis of the inter- action between ecdysteroid receptor and its response-element, p. 63. Presented at the Xth Ecdysone Workshop, Liverpool, April 1992.

Ozyhar A., Strangemann-Diekmann M., Kiltz H. H. and Pongs O. (1991) Characterization of a specific ecdysteroid DNA complex reveals common properties for invertebrate and vertebrate hormone- receptor DNA interactions. Eur. J. Biochem. 200, 329 336.

Parker M. G. (1991) Nuclear Hormone Receptor, Molecular Mechan- isms, Cellular Functions, Clinical Abnormalities. Academic Press, San Diego.

Riddihough G. and Pelham H. R. B. (1986) Activation of the Drosophila hsp27 promoter by heat shock and by ecdysone involves independent and remote regulatory sequences. EMBO J, 5, 1653-1658.

Riddihough G. and Pelham H. R. B. (1987) An ecdysone re- sponse element in the Drosophila hsp27 promoter. EMBO J. 6, 3729-3734.

Schwabe J. W. R., Neuhaus D. and Rhodes D. (1990) Solution structure of the DNA-binding domain of the oestrogen receptor. Nature 348, 458-461.

Segraves W. A. and Richards G. (1990) Regulatory and developmental aspects of ecdysone-regulated gene expression. Invert. reprod. Dev. 18, 67-76.

Str/ihle U., Munsterberg A., Mestril R., Klock G., Ankerbauer W., Schmid W., Schutz G. (1988) Cooperative action of the glucocorti- cold receptor and transcription factors. Cold Spring Harbor symp. quant. Biol. 7, 835-841.

Thummel C. S. (1990) Puffs and Gene regulation: molecular insights into the Drosophila ecdysone regulatory hierarchy. BioEssays 12, 561 568.

Tsai S. Y., Carlsted-Duke J., Weigel N. L., Dahlman K., Gustafsson J.-A., Tai M.-J, and O'Malley B. W. (1988) Molecular interactions of steroid hormone receptor with its enhancer eiement: evidence for receptor dimer formation. Cell 55, 361-369.

Wrange O., Eriksson P. and Perlmann T. (1986) The purified activated glucocorticoid receptor is a homodimer. J. biol. Chem. 264, 5253 5259.

Acknowledgements--We wish to thank Dr Andrzej Ozyhar for discus- sion. We acknowledge the expert advice and help of Dr Clotilde Terry in scanning autoradiographs. We thank Drs Helen Benes and Jeremy Garwood for critical reading of the manuscript. C. Antoniewski is a predoctoral fellow of the Minist+re de la Recherche et de la Technolo- gic. This work was supported by grants from the European Economic Community (grant No. SCI*-0123-C), the Association pour la Recherche sur le Cancer and the Centre National de la Recherche Scientifique to J. A. Lepesant.