Distance-dependent Interactions between Basal ... - Semantic Scholar

4 downloads 0 Views 2MB Size Report
interactions between Pit-1 and T3R in directing rGH pro- moter activity have ..... for 24 h results in additive effects, whereas treatment for 48 h results in strong ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 268, No. 20, Issue of July 15,pp. 1490614911, 1993 Printed in U.S.A.

Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Distance-dependent Interactions between Basal,Cyclic AMP, and Thyroid Hormone Response Elementsin the Rat Growth Hormone Promoter* (Received for publication, January 20, 1993, and in revised form, March 23, 1993)

William P. TanseyS, Fred Schaufele, Margaret Heslewood, Cheryl Handford, Timothy L. Reudelhuberg, and Daniel F. Catanzaroli From the School of Biological Sciences, The University of Sydney, New South Wales, Allstralia 2006 and the Metabolic Research Unit, University of California, SanFrancisco, California 94143-0540

Developmental stage- and tissue-specific expression bp’ upstream of the transcription start site. Withinthis of the rat growthhormone (rGH) geneis conferred by region, binding of the pituitarytranscription factor Pit-1 DNA sequences within 237 base pairs of the transcrip- directs cell specificity (2, 5, 6, 9-12) and responsiveness to tion start site. Although binding of a number of tran- CAMP and phorbol esters (13-161, while a number of other scription factors includingPit-1, S p l , GHF3, and thy- more ubiquitous transcription factors including S p l (17), roid hormone receptor (T3R) stimulates rGH expres- GHF-3 (18,191, and thyroid hormone receptor (T3R) (14,20sion, several studies have suggested that interactions 22) contribute to basal activity and responsiveness to thyroid between these factors are important in determining hormone (T3).Although the interrelationships between these cell specificity and responsiveness to extracellular sig- factors are not well understood, the more distal of two Pit-1 nals. We have directly tested this hypothesis by creating a set of nested insertional mutations at two posi- binding sites overlaps a single Spl site in both rat (17) and tions inthe rGH promoter. Sequenceswere inserted at human (23) GH promoters in such a way as to prevent the either position -148, separating GHF-3 and T3R bind- simultaneous binding of these two factors. Pit-1 binding also ing sites from the downstreamPit-1 and S p l binding appears to be required for thyroid hormone responsiveness of sites, or at -51, separating the above elements from both rat (18) and human (24) GH promoters, and synergistic the TATA box.All insertions weremade in the context interactions between Pit-1 and T3R in directing rGH proof the rGH gene -237/+8 5”flanking DNA, linked to moter activity have recently been reported (25). However, a chloramphenicol acetyltransferase reportergene and little is known about the spatial requirements for these interthyroid hormone response elements tested for activity by transient transfection in GC pi- actions. Synthetic tuitary tumor cells. Insertions at both -148 and -51 (TREs) have been shown to function up to 700 bp upstream caused sharp distance-dependent reductions in serum- of the rGH promoter (26), although the extent of TBinduction stimulated expression such that insertions of 23 base appears to be reduced when the TREis movedfrom its native pairs at -61 or 44 base pairs at -148 weresufficient position. to isolate the effects of sequences upstream of the inTo examine the functional relationships between rGH regsertion point. Insertions at -148 reduced T3 respon- ulatory elements, we carried out an insertional mutagenesis siveness severalfold but hadlittle orno effect on stim- of the rGH 5”flanking DNA, introducing sequences of various ulation by forskolin, whereas insertions at -51 re- lengths either between the TATA box and the Pit-1binding duced bothT3 andforskolin responsiveness. Our sites (-51 insertions) or between the Pit-1 binding sites and results are consistent with thehypothesis that expres- upstream sequences, which contain the principal TREs and sion and regulation of the rGH gene is dependent on GHF-3 binding site (-148 insertions). Our results show that are spacing between the TATA box and more distal elements is short-rangeprotein-proteininteractions,which more critically dependenton spacing than the relative critical for activity. Generally, insertions at -51 had a more orientation of the transcription factor binding sites. deleterious effect on activity than insertions at -148. However, insertion of either 23 bp at -51 or 44 bp at -148 effectively abolished the effects of sequences upstream of the insertion point on serum-stimulated activity. Transfections Studies employing both cell cultures (1-6) and transgenic employing cells grown in stripped serum showed that insermice (7, 8) have shown that pituitary specificity of growth tions at -148 reduced T3responsiveness severalfold, but had hormone gene expression is mediated by sequences within 237 little or no effect on stimulation by forskolin, whereas insertions at -51 reduced both TSand forskolin responsiveness, * This work was supported by a grant from the National Health although the effect on forskolin responsiveness appeared to and Medical Research Council of Australia and by National Institutes be slightly greater. Taken together, these results suggest a of Health Grant DK39998. The costs of publication of this article model for rGH promoter function in which spacing between were defrayed in part by the payment of page charges. This article the component elements is critical for activity. must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. 4 Present address: Cold Spring Harbor Laboratory, P.O. Box 100, Cold Spring Harbor, NY 11724. §Present address: IRCM, 110 Ave. de Pins, Montreal, Quebec H2WIR7, Canada. n To whom correspondence and requests for reprints should be addressed Dept. of Physiology and Biophysics, Cornel1 University Medical College, 1300 York Ave.,New York, NY 10021. Tel.: 212746-6497;Fax: 212-746-8451.

MATERIALS AND METHODS

Cell Culture-GC cells were grown in monolayer culture in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% ‘The abbreviations used are: bp, base pair(s); rGH, rat growth hormone, CAT, chloramphenicol acetyltransferase; TRE, thyroid hormone response element; DMEM, Dulbecco’s modified Eagle’s medium; FCS, fetal calf serum.

14906

Distance-dependent Interactions in the rGH Promoter

14907

not encompass a critical sequence, substitution mutations in this region were also created. Constructions containing insertions at the-148 position were also prepared so as to retain a common sequence at the 3’ end. Although the 15-bp insertion differedat the3‘ side, this entiresequence was contained within the 35- and 44-bp insertions. Inspection of the insertedsequences and thenew sequences created at their junctions with rGH sequences revealed no significant identities to binding sites for any known transcription factors. In addition, analysisof these sequences for their potential flexibility and curvature showed no strong predisposition toward the formationof structures that might influence promoter activity(31, 32). Restriction fragments spanning the rGH sequences -237/ previously (17). +7 from either the nativegene or each insertion mutantwere Transfections and CATAssays-GC cells were transfected by electroporation as described (23) using a Bio-Rad gene pulser apparatus. linked to a CAT reporter gene (23). Additional CAT conTen pg of each construct were transfected into 3 X lo7 cells along structs containing the rGH native sequences -148/+7 or -51/ with 5 pg of either pRSVPGal(28) or pRSV. CAT (17) to standardize +7 were also prepared. All constructs were transfected by for transfection efficiency, dependingon whether cells were assayed electroporation into the rat pituitary tumor cell line,GC, for CAT activity or mRNA. Conditions for electroporation were 960 together with a hybrid RSV-p-galactosidase gene cotransmicrofarads, 300 V, 0.4 cm electrode gap in a total volume of 0.4 ml in Dulbecco’s phosphate-buffered saline.Transfected cells were fected to normalize for variations in transfection efficiency. plated into 4 X 10-cm dishes for CAT analysis or 1 X 175 cm2 tissue In these and previous studies (23), we found the activity of culture flasks for analysis of RNA. Cultures were maintained post- the RSV-@gal construct to be stimulatedless than 2-fold by transfection in DMEM supplemented with 10% FCS or, where the either T3or forskolin. effects of case of TSand forskolin were to be tested, DMEM suppleEffects of Insertions at -148 and -51 on Serum-stimulated mented with either 10%stripped FCS and 10 nM T, (in ethanol), 10 Activity of the rGH Promoter-To determine the effect of p~ forskolin (in dimethyl sulfoxide), both, or 10% stripped FCS and insertions on rGH promoter activity, transfectionswere peran equivalent volume of solvent(s)as control. ForCAT analysis, cells were harvested by scraping 20 h after formed on cells grown in the presence of 10% fetal bovine transfection. CAT and P-galactosidase assays were as described pre- serum. The results of these experiments are summarized in viously (23). CAT activities for each dish were normalized for differ- Fig. 2. Activities were normalized to the -237 rGH.CAT ences in transfection efficiency by dividing by the @-galactosidase construct, which yielded 23,289 k 5,476 cpm in six separate activity determined for the corresponding dish. As CAT activities between transfections differed up to 5-fold,CAT activities were transfections. Generally, insertions at -148 (Fig. 2 A ) had a normalized to the native -237 rGH. CAT construction. less deleterious effect than insertions at-51 (Fig. 2B). InserRNA Preparation and Analysis-Total cellular RNA was prepared tion of 15 bp at -148 decreased activity approximately 40%, by lysis of the cellsin situ with 4 M guanidinium isothiocyanate, followed bypurification through a CsCl step gradient (30).Transcrip- while longer insertions progressivelyreduced activity such tion initiation sites were mapped using a ribonuclease protection thatinsertion of 44 bp abolished anyapparent effect of sequences -237/-148. At the -51 position, insertion of as assay as described previously (17). little as 5 bp caused a 2-%fold reduction in CAT expression, RESULTS while insertion of 23 bp was sufficient toreduce CAT expreslevel of the -51 rGH promoter construct. For both T o alter the spacing and alignmentbetween promoter ele- sion to the -51 insertions, activitywas decreased in proportion -148 and ments in the rGH gene, DNA sequences of varying length, derived from the pUC18 polylinker, were inserted at either to the length of insertion. However, there was no apparent turn dependencebetween even (10 and20 bp) andodd (5,15, position -148 or -51. Insertions of 10-11 bp or multiples thereof, which represent integral turns of DNA helix, alter and 35 bp) turn insertions, suggesting that spacing of prorelative the distance between sequences either side of an insertion moter elements is more critical for activity than their orientations. point but not their alignment and are referred to as “even” rGH Promoter Activity IsNot Affected by Mutations Spaninsertions. Nonintegral turns of DNA helix alter both the distance and alignmentof sequences either side of the inser- ning the -51 Insertion Point-To determine whether theloss tion point and are referred to“ oasd d insertions. Fig. 1shows of activity associated with insertion of sequence at the -51 the location andsequence of the insertion mutations preparedposition might bedue to disruption of a factor binding site,a in this studyrelative to known promoter elements in the rGHscanning mutagenesis, introducingblocks of four nonconservative transversions from -53 to -38 (Fig. l),was carried gene. Insertionpoints were chosentoseparateeitherthe TATA box from sequences upstream (-51 insertions) or out. Theeffects of these mutationswere assessed in an RNase sequences containing cell-specific elements from more distal protection assay using an RSV. CAT construct as a cotransregions containing the principal TREs and GHF-3 binding fected control.Theresults of theseexperiments (Fig. 3) site (-148 insertions). The specific sites selectedfor these demonstrate that base substitutions in this region have either insertions span no known factor binding sites (seebelow). no effect on activity (-45/-42) or increase activity less than Insertions were created so as to retain as much common %fold (-53/-50 and -41/-38). This would suggest that the sequence as possible between the various constructions. At decreased activities caused by insertions at the -51 position the -51 position, the 5-bp insertion duplicated the native are not due to disruption of the site at which the insertions sequence adjacent to the insertion point. Longer insertions were made. Moreover, mutations downstream of the insertion were created by adding sequences between this duplication. point had no appreciable effect on activity. Gel mobility shift As a consequence, sequences on both sides of the insertion assays carried out using restriction fragments spanning this point resemble the native sequence. This reduces, although region showed that the mutations had no effect on factor does not eliminate, thepossibility that theeffect of insertion binding either in thevicinity of the insertion point(GHF-6)’ may be due to the disruption of some previously undetected factor binding site. To confirm that the insertion pointdoes F. Schaufele, unpublished observations.

fetal calfserum (FCS) (Cytosystems) (23). TO test the effects of thyroid hormone (T3)and forskolin,cells were grown for 24 h before transfection in DMEM supplementedwith 10% charcoal/resinstripped FCS (27). Generation of &H.CAT Hybrid Genes and Insertion MutantsSequences containingthe rGH 5”flanking DNA -237/+8, -148/+8, and -51/+8 were inserted into pPCAT (23) to generate -237 rGH. CAT, -148 rGH. CAT, and -51 rGH .CAT. All insertion constructs were generated inthe context of -237 rGH.CAT. Insertions at -148 were created by first inserting a 44-bp polylinker fragment at an FnuDII site at -148 and then using this construction to deletevarious length fragmentsfrom the polylinker segment to generate a series of shorter insertions. Insertions at -51 were generated by subcloning rGHsequences -237/-51 between the BamHI and SmaI sites of pUC18 and ligating these to -51 rGH. CAT with different lengths of polylinker. Site-directed mutagenesis was carried out as described

14908

Distance-dependent Interactions in the rGH Promoter GHF3 (-219/-239)

-

TRE (-1 66/-190)

TATA (-30)

-Pit-I -i (-1 OW-1 43) (-W-94)

I

1 1 - 1 -

-

WT

..

...

GGAGCCTTGGGGTCGAGGAAAACA CCTCEAACCCCA$CTCCTTTTGT

5bp

-

T-TCGAGGAAAACA ACCCCSCTCAGCTCCTTTTGT

1~ b p TGGGZ&XTAGWKGAGGAAAACA ACCCCMGSATCTCAGCTCCTTTTGT

1%

23bQ

-

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

-

T G T C G A G G A M i A C A A C C C A G C T C C T T T T G T

43/40

mutations -41 1-38

"ICGAGGAAAACA -GCTCCTTTTGT TGGGGTCGATTCCA?ICA ACCCCAGCTAAM3TTGT TGGGGTCGAGGACCCCAGCTCCT-

FIG. 1. Factor binding sites inthe rGH promoter and the location and sequences of insertions. Coordinates of binding sites for factors described in the text are numbered backwards from the transcription start site. Sequences were inserted either at position -148 or -51. The sequence of the 15-bp insertion at -148, which is contained within the 35- and 44-bp insertions, is underlined. Sequences duplicated a t the 5' and 3' ends of the -51 insertions are ouerlined. Although the restriction enzyme TqI cleaves at position -48, duplication of part of this site in each of the insertions resulted in a shift of the apparent insertion point to -51. The location of a T3R binding site immediately adjacent to the -51 insertion point is shown as the residues making contact with receptor ( e ) identified by methylation interference (22). Substitution mutations spanningthe -51 insertion point are indicated. WT, wild type. (see below) or at other sites more distant (Pit-l/GHF-l or Spl) (data not shown). Effects of Insertions a t -148 and -51 on T3and Forskolin Stimulation of the rGH Promoter-In addition to the pituitary-specific expression conferred by Pit-l/GHF-l, rGH promoter activity is stimulated by both T3 and CAMP (14, 18). To determine whether T3- and CAMP-responsive elements were affected by spacing and toexamine their contributionto changes observed in serum-stimulated activity, transfections were carried out on cells grown in hormone-depleted media. In theabsence of hormones, basal activities from constructs containing native sequences and insertion or deletion mutations were low and differed less than %fold (Figs. 4 and 5). It is noteworthy, however, that atboth the -148 and -51 sites, even turn insertions reproducibly had agreater effect on basal activities than odd turn insertions. Although this effect was small, it affected the apparenthormone inductions, especially of the -51 insertions (see below). These results indicate that the activity of the rGH promoter is largely hormone-dependent and suggest that the effects of insertions observed in the presence of whole serum may be due to the disruption of interactions involving hormone-responsive elements. Individually, both Tsand forskolin-stimulated expression of the wildtype and insertion constructs and their effects when applied together were additive (Figs. 4 and 5). Maximally stimulated activities of the native rGH promoter (Fig. 4, 40,914 f 457 cpm, n = 2; and Fig. 5, 27,325 f 7,161 cpm, n = 4) were similar to the serum-stimulated activities observed in Fig. 2,

suggesting that addition of T3 plus forskolin mimics the effect of serum. The relative activities of -148 and -51 insertion mutants, after treatmentwith T3, forskolin, or both,however, were reduced up to &fold. With insertionsof increasing length at -148, T3induction was reduced 2-3-fold, whereas forskolin induction was relatively unaffected. Deletion of sequences to -148 virtually abolished induction by T3but had no significant effect on forskolin responsiveness. As induction by T3 and forskolin were of a similar magnitude, when the two agents were added together, stimulation by forskolin diminished the effect of insertions on T3 induction. Conversely, insertions at -51 reduced induction by both T3 and forskolin, although the effects on induction by T3 alone were smaller. Although insertions reduced forskolin induction to the 3-5fold range, this was still significantly greater than -51 rGH . CAT (2-fold), suggesting that the CREs are contained upstream of -51. Moreover, the effects of forskolin were most evident when applied in combination with T3. Under these conditions, the effect of T3 plus forskolin was significantly greater than the effect of T3 alone for native and insertion constructs but not-51 rGH. CAT. The weak turn dependence observed in the basal activities of insertion mutants was also evident in the stimulated activities. This resulted in similar inductions for each construct even though the induced activity was significantly reduced by longer insertions. For example, while the hormone-induced activities of all -51 insertion constructs differed significantly from the native control,only some of the differences in inductions were significant. More-

Distance-dependent Interactions

o

o

15

5

20

I O

35

15

the in

rGH Promoter

14909

44 -148rGH

23

-51 rGH

FIG. 2. Effects of insertions on serum-stimulated, transient expression following transfection into GC cells. Insertion and deletion mutations described in Fig. 1 and under “Materials and Methods” were transiently transfected by electroporation into GC cells grown in DMEM containing 10% FCS. Following transfection under serum-free conditions, cells were replated in serum-containing media, harvested, and lysed for assay after 24h. A, activities of constructions containing insertions at -148 (length of insertion shown for each construct) and -148 rGH.CAT. R, activitiesof constructions containing insertions at -51 (length of insertion shown for each construct) and -51 rGH. CAT. Each construction was tested in three to six separate experiments.

over, differences in activities between the 5-bp insertion and longer insertions differed significantly for a number of treatments, whereas noneof the differences in inductions between different length insertionswere significant. Ribonuclease protectionexperiments revealed that correctly initiated transcripts from the various constructions accumulated to levels corresponding to CAT activities in response to hormone ments (data not shown). DISCUSSION

Spatial requirementsfor interactions between promoter and enhancer elements havebeen most extensively characterized for the SV40 regulatory region (33). The latter studyshowed that spacing and alignment was far more criticalfor promoter than enhancer elements. Thus, insertions between promoter and enhancer sequences were found to reduce activity 2-5fold in a turn-dependent manner such that the insertion of whole turns of DNA helix had a less deleterious effect than insertion of nonintegral turns. Insertions between the TATA box and sequences upstream reduced TATA-dependent transcripts 4-10-fold, while TATA-independent transcripts increased up to 4-fold, both in a relatively turn-independent manner. Studies on the SV4Q enhancer (34-37) havealso demonstrated interactions betweena number of functionally redundant modules, which may be interchanged or separated up to 50 bp. Each module may be further subdivided into a

rGH

-

RSV

-

FIG. 3. Effects of mutations in the -51 region of the rGH promoter on serum-stimulated, transient expression following transfection intoGC cells. Each mutation shown in Fig. 1 was introduced into the context of-237 rGH. CAT and transiently transfected into GC cells by electroporation together with RSV. CAT as a cotransfected control. RNA was prepared and analyzed as described under “Materials and Methods.” Transcripts originating from either of these plasmids are indicated. WT, wild type.

number of smaller segments (38) thatmost likely correspond to binding sites for individual transcription factors(39). However, unlike the enhancer modules they form, the arrangement of these sites ishighly sensitive tospacing. The results of the present study show critical distancedependent relationships betweensequence elements in the rGH promoter. Insertionsa t -148, which separated the GHF3 site and TRE from elements downstream, reduced activity and Ts responsiveness but had little effect on forskolin responsiveness, while insertions a t -51, which separated the treat- box from elementsupstream, reduced activityand TATA both Ts and forskolin responsiveness. These observations are consistent with thelocalization of a predominant TRE to the region -190/-166 (20,22) and cAMPresponsive elements to the region -104/+11(14). Ourdatafurtherindicatethat cAMP response elements are contained upstream of-51. Takentogether,theseobservations localize a CRE to the region -104/-51, which contains the proximal Pit-1 site. A number of recent studies indicate that cAMP responsiveness may be mediated by Pit-1 (16, 40, 41). Although insertions at -51 attenuated forskolinresponsiveness, this effect was most apparent when forskolin was added in combination with T3. However, unlike the strong synergism reported by Copp and Samuels (14), under the conditions of the present study, the effects of forskolin and TBwere at best only slightly more than additive. This is most likely due todifferences in the timecourse of transfection and hormonetreatment.Inthepresentstudy, hormones were applied for 24 h post-transfection, whereas Copp and Samuels (14) treated thecells for 48 h. We have found that treatment

14910

Distance-dependent Interactionsin the rGH Promoter

A

A

0 352015

B

44

-148 rGH

5

0

10

15

rGH pPCAT

-51 23

B 30

I

I

T

0T3

0 352015 44 -148 rGH 0 5 10 -51 2315 rGH pPCAT FIG. 4. Effects of -148 insertions on basal, Ts-, and forskoFIG. 5. Effects of -51 insertions on basal.. Ts-. - . and forskl lin-stimulated transient expressionfollowing transfection lin-stimulated transient expressionfollowing transfection into GC cells. Cells were grown inDMEM containing 10% charcoal/ into GC cells. Transfection conditions, treatments, relative activiresin stripped-FCS for 48 h prior to transfection and replated post- ties, and -fold inductions (mean & S.E., for four separate transfectransfection in fresh media containing vehicle (-), T, (2'3) (10 nM), tions) were as for Fig. 4. Constructions are labeled as shown in Fig. forskolin ( F ) (1 p M ) , or T3 plusforskolin (T3 + F ) , as indicated. 2B. In A , all differences between the activities of insertions and the Constructions are labeled as shown in Fig. 2 4 . A , activity normalized native controlwere significant at p < 0.05 for each treatment. Asterfor the -237 rGH.CAT construct stimulated with T3 + F (mean ? isks indicate significant differences between the activitiesof specific S.E., for two separate transfections);B, fold inductions on activities constructs andthe 5-bp insertionconstruct for eachtreatment. In B , shown in A relative to no treatment control (-). Differences in -fold asterisks indicatesignificantdifferences ( p < 0.05) between the inductions (relative tothe native rGH sequence)were evaluated by a activities of insertionconstructsand the nativecontrolforeach Scheffe F-test. Asterisks indicate differences significantat p < 0.05. treatment.

for 24 h results in additive effects, whereas treatment for 48 of a h results in strong apparent synergism that may be due, at tant factor binding site but rather to the perturbation critical spacingof sites either side of the insertion point. least in part, to inductionof Pit-1 by CAMP (42). Turn effects havebeen demonstratedfor insertions between Of particular importance is the possibility that the inser(18) and tions created in the rGH promoter might disrupt the bindingthe two Pit-1 binding sites in the rGH promoter site(s) for some previouslyunidentified factor(s). Sinceat the between the two most proximal Pit-1 binding sites in the rPRL promoter (43), where odd turn insertions appeared to -148 position only longer insertions appeared to have any significant effect, it is unlikely that this position occurs withinbe more deleterious than even turn insertions. In the present the binding site for any important factor(s). However, the study, a weak turn effect resulted in slight increases in activity the or -51 sites. sharp reductions in activity observed with all insertions a t among the odd turn insertions at both -148 -51 suggest that this sitemay bind some important factor(s). This could be due either to creating a more favorable alignof a n Although the -51 site is adjacent to a known T3R binding ment between positive control elements or to disruption optimal alignment of negative elements. However, the major site, oligonucleotides containing sequences immediately upstream of the insertion point have been shownto bind receptor effect of insertions at either site in the rGH promoter was (22). This region also appears to bind another factor (GHF- distance- rather than turn-dependent. Insertions at -51 re6).* However, mutations in thisregion had no apparent effect sulted in marked reductions in activity similar to those observed in theSV40 control region when insertions were made on GHF-6binding. These same mutations had either effect no on rGH promoter activity or increased activity slightly. Taken between the TATAbox and promoter and enhancer elements together, these data suggest that the effect of insertions at further upstream (33).Although insertions at this position in the -51 position is not due to the disruptionof some impor- the SV40 control region exhibited some turn dependence, the

Distance-dependent Interactions in the rGH Promoter

14911

major effect was distance-dependent with insertions of 5,10, especially between the Pit-1 sitesand the TATA box, is and 15 bp reducing activity to approximately 15, 20, and 5% critical for their optimal interactions. of the wild type, respectively. Turn effects have been described REFERENCES for insertions between the glucocorticoid response element 1. Casanova, J., Copp, R. P., Janocko, L., and Samuels, H. H. (1985) J. Biol. Chern. 260,11744-11748 and CACCC box of the tryptophan oxygenase promoter (44). 2. Catanzaro, D. F., West, B. L., Baxter, J. D., and Reudelhuber, T. L. (1987) Our failure to observe turn-dependent effects on activity and Mol. Endocrinol. 1,90-96 3. Crew, M. D., and Spindler, S. R. (1986) J. Biol. Chem. 261,5018-5022 hormone responsiveness may be due to the distance of the 4. Nelson, C., Crenshaw E., 111, Franco, R., Lira, S. A,, Albert, V. R., Evans, GHF-3 site and TRE from the insertion point (>35 bp). This R. M., and Rosenfeid, M. G. (1986) Nature 322,557-562 5. Ye, Z. S., and Samuels, H. H. (1987) J. Biol. Chem. 262,6313-6317 could permit sufficient twisting and bending in the DNA to 6. West, B. L., Catanzaro, D. F., Mellon, S. H., Cattini, P. A., Baxter, J. D., allow contacts between transcription factors or transcription and Reudelhuber, T. L. (1987) Mol. Cell. Biol. 7 , 1193-1197 7. Behringer, R. R., Mathews, L. S., Palmiter, R. D., and Brinster, R. L. factor complexes, which would be placed on opposite sides of (1988) Genes & Deu. 2,453-461 the DNA helix by odd insertions and thusexperience greater 8. Lira S. A,, Crenshaw, E., 111, Glass, C. K., Swanson, L. W., and Rosenfeld, M . G. (1988) Proc. NQtl.Acad. Sci. U. S. A . 85,4755-4759 difficulty in interacting than configurations where shorter 9. Bodner, M., Castrillo, J., Theill, L. E., Deerinck, T., Ellisman, M., and Karin, M. (1988) Cell 55,505-518 sequences separated their binding sites. Alternatively, turn H. A., Chen, R., Man alam, H. J., Elsholtz, H. P., Flynn, S. E., effects could be masked by the effects of other transcription 10. Ingraham, Lin. C. R.. Simmons. D. M.. lwanson., L.., and Rosenfeld. M. G. (1988) . . Celi55,519-529 factors whose suboptimal alignment might be improved by 11. Mangalam, H. J., Albert, V. R., Ingraham, H. A,, Kapiloff, M., Wilson, L., insertions. However, even if such a phenomenon were to Nelson, C., Elsholtz, H., and Rosenfeld, M. G. (1989) Genes & Deu. 3, 946-9.58 ~ . ~ . .. obscure turn dependence, it could not account for the predom12. Nelson, C., Albert, V. R., Elsholtz, H. P., Lu, L. I., and Rosenfeld, M. G. inant distance effects we observed. (1988) Science 239,1400-1405 G. A,, Harney, J. W., Moore, D. D., and Larsen, R. P. (1988) Mol. Distance-dependent interactions between steroid hormone 13. Brent, Endocrinol. 2 , 792-798 receptors andtranscription complexes assembled atthe 14. Copp, R. P., and Samuels, H. H. (1989) Mol. Endocrinol. 3, 790-796 Dana, S., and Karin, M. (1989) Mol. Endocrinol. 3,815-821 TATA box (29) have also been demonstrated in a study that 15. 16. Ka iloff, M. S., Farkash, Y., Wegner, M., and Rosenfeld, M. G. (1991) showed that a glucocorticoid response element could stimulate 8cience 253,786-789 F., West, B. L., and Reudelhuber, T. L. (1990) J. Biol. Chern. transcription from a minimal TATA box promoter when 17. Schaufele, 266,17189-17196 placed close but not when placed several hundred base pairs 18. Ye, 2. S., Forman, B. M., Aranda, A., Pascual, A., Park, H. Y., Casanova, J., and Samuels, H. H. (1988) J. Biol. Chem. 263,7821-7829 upstream (45). However, in the more distal position, dupli- 19. Schaufele, F., Cassill, J. A., West, B. L., and Reudelhuber, T. (1990) J. Biol. Chem. 2 6 5 , 14592-14598 cation of the glucocorticoid response element or addition of a G. A., Harney, J. W., Chen, Y., Warne R.L., Moore, D.D., and binding site for one of several common transcription factors 20. Brent, Larsen. P.R. (1989) ~ o lEndocrinol. . ~3.199d-2004 ~.~ ~,- ~ ~ . ,~ yielded a transcriptionally active unit (45). A similar obser- 21. Koenig, R . J., Brent, G. A,, Warne, R. L., Larsen, P. R., and Moore, D. D. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 5670-5674 vation was made for thyroid hormone receptor activation of 22. Norman, M. F.. Lavin. T. N.. Baxter.. J. D... and West. B. L. (1989) . , J. Biol. Chem.' 264,12063-i2073 rGH expression (26). Separating sequences containing the 23. Tansey, W. P., and Catanzaro, D. F. (1991) J. Biol. Chem. 266,9805-9813 rGH wild-type TRE or an up-mutantbased on this TRE from 24. Voz, M. L., Peers, B., Belayew, A., and Martial, J. A. (1991) J. Biol. Chem. 2 6 6 , 13397-13404 a truncated 137-bp wild-type rGH promoter resulted in only 25. Schaufele, F., West, B. L., and Baxter, J. D. (1992) Mol. Endocrinol. 6 , modest levels of T S responsiveness, unless present as two or 656-665 Brent, G. A., Williams, G. R., Harney, J. W., Forman, B. M., Samuels H. 26. three tandem copies. These findings are consistent with our H., Moore, D. D., and Larsen, P. R. (1991) Mol. Endocrinol. 5 , 542-848 H. H., Stanley, F., and Casanova, J. (1979) Endocrinology 1 0 5 , observation that separating the rGH TRE by as little as 15 27 . Samuels, 011 Q F bp reduced T3responsiveness %fold. Our findings of distance 28. Fluvv-uv F , Copp, R. P. Casanova J. Horowitz, 2. D. Janocko L., Plotnick, hf., and Samuels, H. H. (198i) j.Biol. Chem. 262,6373-6382 dependence of T3 responsiveness support the concept that T3 29. Sawadogo, M., and Roeder, R. G. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 , responsiveness of the rGH promoter involves interactions 4394-4398 between T3R and some factor downstream of the TRE, pos- 30. Chirgwin, J. M.,Przybyla, A. E., MacDonald, R. J., and Rutter, W. J. (1979) Biochemistry 18,5294-5299 sibly Pit-1 as has previously been suggested (18, 25). 31. Bracco, L., Kotlarz, D., Kolb, A., Diekmann, S., and Buc, H. (1989) EMBO J. The findings of the present study suggest that expression 32. Collis,8,4289-4296 C. M., Molloy, P. L., Both, G. W., and Drew, H.R. (1989) Nucleic Acids Res. 17,9447-9468 and regulation of the rGH gene is dependent on relatively 33. Takahashi, K., Vigneron M. Matthes, H. Wildeman, A., Zenke, M., and short-range protein-protein interactions,which are more critChambon, P. (1986) Nhture 319,121-156 ically dependent on spacing than alignment of the transcrip- 34. Schaffner, G., Schirm, S., Muller-Baden, B., Weber, F., and Schaffner, W. (1988) J. Mol. Biol. 2 0 1 , 81-90 tion factor binding sites. This wouldimply that the exact 35. Zenke, M., Grundstrom, T., Matthes, H., Wintzerith, M., Schatz, C., Wildeman, A., and Chambon, P. (1986) EMBO J. 5 , 387-397 distance between binding sites plays an important role in the 36. Ondek, B., Shepard, A., and Herr, W. (1987) EMBO J. 6,1017-1025 assembly of transcription factor complexes. DNA bending by 37. Fromental, C., Kanno, M., Nomiyama, H., and Chambon, P. (1988) Cell R A QA.?-QK? ~ _--,Pit-1 (46) would tend to curve the DNA between the T3R 38. Ondek B. Gloss L. and Herr, W. (1988) Nature 333,40-45 binding site and the TATA box, bringing these sites closer 39. Dynan: W'. S. (l689j Cell 5 8 , 1-4 40. Yan, G. Z., and Bancroft C. (1991) Mol. Endocrinol. 5 , 1488-1497 together. In addition, we have recently shown that a T3R- 41. Yan,.G. Z., Pan, W. T., i n d Bancroft, C. (1991) Mol. Endocrinol. 5 , 535341 accessory factor complex binding to the upstream TRE in42. Chen R. P. In aham H. A. Treacy M. N. Albert, V. R., Wilson, L., and duces a sharp75" bend in the rGH5"flanking DNA centered Rohenfelci, $G. (1690) Nhture 346,583i586 C., Jackson, S. M. Siddiqui, S. K., and Gutierez-Hartmann, A. at position -181 (47). This could result in the T3R-accessory 43. Harvey, (1991) Mol. Endocrinol. 5; 836-843 factor complex being drawn closer to Pit-1bound downstream 44. Schule, R., Muller, M., Otsuka, M. H., and Renkawitz, R. (1988) Nature 332,87-90 and transcription factor complexes assembled at the TATA 45. Strahle, U., Schmid, W., and Schutz, G. (1988) EMBO J. 7,3389-3395 46. Verrijzer C. P. van Oosterhout J. A. van Weperen, W. W., and van der box. As DNA bending has also been described for TFIID Vliet P. C. ( i m ) EMBO J. io, 3067-3014 binding at the TATA box(481, the entire array of factors 47. Kin I.'N., de Soyza, T., Catanzaro, D. F., and Lavin, T.N. (1993) J. Biol. C k m . 268,495-501 bound to therGH promoter may be drawn closely together in 48. Horikoshi, M., Bertuccioli, C., Takada, R., Wan , J , Yamamoto, T., and such a way that the spacing between factor binding sites, Roeder, R. G. (1992) Proc. Natl. Acad. Sei. U. ,!?A: 8 9 , 1060-1064 '

~I

1 "

~

~

~~

~

~