trp Repressorltrp Operator Interaction - Semantic Scholar

Vol. 267, No. 24, Issue of August 25. pp. 16783-16789,1992 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

trp Repressorltrp Operator Interaction EQUILIBRIUM AND KINETIC ANALYSIS OFCOMPLEX

FORMATION AND STABILITY* (Received for publication, March 6, 1992)

Barry K. HurlburtS and Charles Yanofskys From the Department of Biological Sciences, Stanford University, Stanford, California94305

The trp repressor of Escherichia coli regulates tran- shown to bind to three of its target operators approximately scription initiation in the trp operon by binding at an equally well and with high affinity (8).Equilibrium dissociaoperator located within the trp promoter region. We tion constantsfor the trp repressorltrpoperon operator comhave used a filter binding assay to analyze the inter- plex have been determined using a variety of assay procedures; action between purified trp repressor and a synthetic 2 X 10”’ M (Ref. 9, this work), 3 X 10”’ M (lo), 5 X 10”’ M 43-base pair DNA fragment containing the natural trp (11, 12), 2 X lo-’ M (13) and 6 X lo-’ M (14). Estimated KO promoter-operator region. In equilibrium binding ex- values for the weaker aporepressor-operator interactionrange periments, the KO of high affinity binding of trp re- from 100 nM (11) and 600 nM (9) to 1.6 p M (12). Thus, pressor to this DNA fragment was determined to be 2 tryptophan binding to theaporepressor results inan apparent X 10”’ M. Low affinity binding was observed at reincrease in affinityof 2-3 orders of magnitude. Two molecules pressor concentrations above 10 nM. In kinetic experiments with various input ratios of repressor to oper- of tryptophan arebound by trp aporepressor noncooperatively ator, trp repressor-operator complexes dissociated with a dissociation constant of approximately 20 p M (6, 10, with equivalent, first-order kinetics. Instantaneous re- 15, 16). trp repressor binds to non-operator DNA with apduction of the tryptophan concentration resulted in proximately 1/200th the affinity of the repressor-operator increased rates of complex dissociation, indicating that interaction (11). X-ray crystallographic and NMRstudies have revealed loss of one or both tryptophan molecules from the repressor-operator complex destabilizes the complex. A many of the structural and functional features of this reguheterodimeric repressor with a single tryptophan bind- latory protein (7, 17-21). The secondary structure of the trp ing site was constructed and its affinity for operator repressor polypeptides is mostly a-helical, with each polypepwas compared with that of ligand free aporepressor tideconsisting of six helices, designated A-F (17, 20, 21). and tryptophan saturated repressor. The heterodi- There is a single, central domain comprised of helices A, B, meric repressor had a 20-25-fold higher affinity for C, and F from both polypeptides chains (17,21). Thisdomain operator than did the aporepressor, and it had a 20- has been termed the “central core” (17). Helices A, B, and C 25-fold lower affinity for operator than did the tryp- are interlocked in the central core by interchain contacts. tophan-saturated repressor. Certain residues in the centralcore are involved in maintaining the normal DNA-binding characteristics of the wild type repressor (9). Two “helix-turn-helix” domains, helices D and E, are extended from the central core appropriately to allow The trp repressor of Escherichia coli regulates expression interactions with symmetrical operator sequences, approxiof at least four related operons, trpEDCBA, aroH, trpR, and mately 30 A apart (17,19). The helix-turn-helix is a structural mtr (1-5). The 107-amino acid residue product of the repres- motif common to many DNA-binding regulatory proteins sor structural gene, trpR, dimerizes to form the inactive trp (reviewed in Ref. 22). From NMR studies, the D-Edomain is aporepressor (6). Whenthe aporepressor binds two molecules known to be quite mobile in solution (21). Intercalation of a of the corepressor, tryptophan, it is converted to the fully molecule of L-tryptophan in the binding site located between active species, the trp repressor (7). trp repressor forms spe- the central core and each helix-turn-helix motif restricts the cific, tight complexes with operator sequences within the movement of helices D and E, positioning them properly for promoters of its regulated operons and acts by inhibiting intimate association with operator (7, 18, 19). The two NH2transcription initiation (8). Under conditions in which tryp- terminal arms, 14-16 residues in length, were unresolved in tophan becomes growth-limiting, the aporepressor is the pre- the three-dimensional structures deduced from NMR and dominant form of the trpR gene product, and expression of crystallographic studies, including that for the repressor-opthe operons it controls increases (8).Thus, in vivo trp represerator complex (7,17-21,23). However, these arms have been sor activity is sensitive to the level of free tryptophan in the shown to contribute to repression in vivo (24) and operator cell. binding in vitro (24, 25). Using three in vitro DNA-binding assays, trp repressor was In the classical description of the trp repressor-operator * These studies were supported by National Science Foundation interaction one dimeric repressor molecule binds to the trp Grant DMB-8703685. The costs of publication of this article were operon operator (8). Two alternative models describing trp defrayed in part by the payment of page charges. This article must repressor interaction with the trp promoter-operator region therefore be hereby marked “advertisement” in accordance with 18 have been proposed (26, 27). These models are based on U.S.C. Section 1734 solelyto indicate this fact. footprinting (26) and gel mobility-shift analyses (27). In these $ Supported by National Institutes of Health-National Research Service Award GM11980. Present address: Dept. of Biochemistry and models, repression of the trpoperon results from binding of 2 Molecular Biology, University of Arkansas for Medical Sciences, or 3 molecules of repressorltrp promoter-operator. Additionally, Staake et al. (27) have challenged the validity of the Little Rock, AR 72205. Career Investigator of the American Heart Association. x-ray crystallographic structure of the repressor-operator

16783

16784

trp Repressor-Operator Complex Formation Stability and

complex, proposing that an incorrect operator fragment was used. Recent gel mobility-shift analyses support binding at the classical operator sequence (28). In this study, we used a filter binding assay to analyze the formation and stability of complexes between trp repressor and a synthetic operator DNA. The operator fragment used is identical tothenaturaltrp promoter-operator (9)and contains all sequences thought to be involved in repressor binding in the three models. We observed a biphasic increase in operator retention on filters in equilibrium binding experiments, indicative of at least one high affinity binding site (KO = 2 X 10"' M) and at least one low affinity binding site ( K D= 5-7 X lo-' M) in the promoter-operator fragment. trp repressor-operator complexes, formed with saturatingand subsaturating levels of repressor dissociated with equivalent, first-order kinetics (tllP= 2.5 min), suggesting that there isa single, high affinity binding site for trp repressor in the trp promoter-operator region. We also analyzed the effects of ligand concentration on trp repressor-operator complex formation and the rate of complex dissociation. In the presence of saturating levels of either of the corepressing ligands, tryptophan or 5-methyltryptophan, the KDfor the trprepressor-operator interaction was 2 x 10"' M. However, a lower concentration of 5-methyltryptophan than tryptophan was required to activate the trpaporepressor for operator binding, consistent with the reported affinities of these ligands for aporepressor (29). trp repressor-operator complexes were found to have 2.5-min half-lives in the presence of saturating levels of either tryptophan or 5-methyltryptophan. The rate of complex dissociation increased appreciably as theconcentration of free tryptophan was reduced demonstrating that under such conditions loss of bound tryptophan decreases the stability of the complex. To determine the stability of repressor-operator complexes containing one bound tryptophan we examined the operator affinity of a heterodimer composed of wild type and mutant polypeptide chains. Thisheterodimeric repressor, which we refer to as the "hemirepressor," is believed to have a single tryptophan binding site. The hemirepressor's affinity for the operator was found to be intermediate between that of the aporepressor and repressor suggesting that repressor with a single bound tryptophan forms a less stable repressor-operator complex. EXPERIMENTAL PROCEDURES

Repressor Purification-Wild type and GR85 trp repressors were produced and purified as described (6,30). Thecharacteristics of the GR85 repressor have been examined (16,31). Thepurity of repressor preparations was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Following Coomassie Brilliant Blue or silver staining, these repressor preparations were visually judged to be greater than 95% pure. Repressor concentration was determined spectrophotometrically using an extinction coefficient of 1.2 cm" mg" ml at 280 nm (6). Heterodimer (Hemirepressor) Formation-Heterodimers of GR85 and wild type polypeptide subunits were prepared by heat treatment essentially as described (16). Purified G B 5 and wild type aporepressors were mixed a t molar ratios of 5 0 1 and 25:1, respectively, in 10 mM sodium phosphate, pH 7.6,150 mM sodium chloride, 0.1 mM EDTA, heated to 65 "C for 4 min and slowly cooledto room temperature. The efficiency of heterodimer formation was monitored using radiolabeled wild type repressor. Following native polyacrylamide gel electrophoresis, all of the radiolabeled protein was found to have a mobility intermediate between that of the GR85 and wild type homodimers. Hemirepressor preparations were stored in small aliquots at -20 "C. Fresh samples were used in each experiment. Operator DNA Used for DNA-binding Assays-The 35S-labeled43base pair operator DNA used in filter binding assays was prepared as described (9). This sequence corresponds exactly to base pairs -31 to

+14 of the natural trp promoter-operator. +1

AATTAATCATCGAACTAGTTAACTAGTACGC~GTTCACGT~ TTAATTAGTAGCTTGATCAATTGATCATGCGTTCAAGTGCATTTT

For the HpaI restriction-site protection assay (32), the same 43-mer (unlabeled) was cloned into the unique SmaI site of pUCll8 (33) by standard procedures. The resulting plasmid (pBKH26) was linearized by digestion with A d , which cleaves 600 nucleotides from the inserted operator, and radiolabeled with [32P]dCTP(3000 Ci/mmol) using the large fragment of E. coli DNA polymerase I in the presence of unlabeled dTTP, dGTP, and dATP. Unincorporated nucleotides were removed by gelfiltration chromatography. DNA concentrations were determined spectrophotometrically a t 260 nm. In Vitro DNA-binding Assays-The filter binding assay was performed as described (9). The 5-methyltryptophan used (Sigma) was a racemic mixture. The concentration of L-5-methyltryptophan was taken as one-half the concentration of DL-5-methyltryptophan.The HpaI restriction site protection assay was modified from that described (32). The central 6 base pairs of the trp operator, GTTAAC, is the recognition site for HpaI. Thus, incubation of a repressoroperator mixture with an appropriate amount of HpaI endonuclease will result in cleavage of unbound operator DNA. 0.6 unit of HpaI was found to be the minimum amount of enzyme required to digest 1 pM pBKH26 completely in 10 min in a 2 0 4 reaction volume. The reaction buffer was 25 mM Hepes-KOH,' pH 7.5, 100 mM potassium glutamate, 10 mM magnesium acetate, 0.5 mM 8-mercaptoethanol. 300 p M tryptophan was present when appropriate. To measure repressor binding, serially diluted repressor was mixed with a template solution to give a final concentrationof 1p~ 3ZP-labeledpBKH26 in 18 pl. Following equilibration, 0.6 units of HpaI were added in 2 pl and the mixtures were incubated for 10 min at 37 "C. The reaction mixtures were immediately loaded onto a 0.7% agarose gel in Trisborate buffer, pH 7.6, and electrophoresis was carried out for 250 Vh. The resolved DNA species were transferred to Nytran membranes (Schleicher & Schuell) electrophoretically, according to themanufacturer's instructions. The dried membranes were exposedto x-ray film (Kodak XAR) at -80 "C using intensifying screens (Du Pont Cronex). The concentration of repressor required for approximately 50% protection was estimated by visual inspection. RESULTS

Equilibrium Binding of trp Repressor to Synthetic trp Operator-Filter binding experiments were performed in which a limiting amount of operator (1 X IO-" M) was equilibrated with increasing amounts of trp repressor, followed by filtration.Thistitration experiment gave a two-phase binding curve (Fig. 1).The first plateau occurred when approximately 60% of the available operator DNA was retained on the filters. The estimated K D for the repressor-operator interaction in the range of 0 and 60%operator retention is 2 X 10"' M. This value agrees with our previous estimate, aswell as with those of others (9-12). In our experience with the filter binding assay with trp repressor and trp operator less than 100% of high affinity repressor-operator complexes were retained on filters during sample application and washing. Increasing the volume of either the sample or the wash reduced the fraction of repressor-operator complexes retained on filters, and also reduced the amount of operator bound to filters nonspecifically (in the absence of repressor).' When the repressor concentration was raised above 10 nM (Fig. l ) , we observed an additional increase in bound operator, up to 100% retention. The estimated KD for this second phase of operator retention is M. We interpret the second phase approximately 5-7 x of operator retention as low affinity binding of additional repressor molecules to DNA molecules that already have The abbreviations used are: HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; WT, wild type; FBB, filter binding solution B. * B. K. Hurlburt and C. Yanofsky, unpublished data.

Repressor-Operator trp

1

0.1

0.01

10

Complex Formation and Stability

100 1000

[Repressorn ]M FIG. 1. Equilibrium binding of limiting trp operator by increasing concentrations of trp repressor. '%-Labeled operator

(1 X lo-" M ) was incubated with trp repressor in filter binding solution B (FBB)plus 300 PM tryptophan. trp repressor concentration was varied from 0 to 600 nM. % Operator Bound was the fraction of the totalradioactivity in the mixtures that was retained on the filters following washing (background was not subtracted).

a

d 1

100

0

F9 a

0

c, (d

r 8

0"

*

10

0

1

2

3

4

5

6

T i m e( m i n u t e s ) FIG. 2. Determination of the dissociation rates of trp repressor-operator complexes. Purified wild type trp repressor was incubated with 35S-labeled,43-base pair operator DNA in FBB plus 300 P M tryptophan. The operatorconcentration was 1 nM. The 1:1, 1 nM repressor concentrations were as follows: 1:10, 0.1 nM (a), (A),and 1O:l) 10 nM (+). At time zero, unlabeled 43-base pair operator fragment was added to a final concentration of 500 nM. Aliquots were then removed at various times, filtered, and washed. 100% Operator Bound is defined as theamount of labeled operator fragment retained on the filter prior to the addition of unlabeled operator.

bound trp repressor at a high affinity site(s). The range of low affinity binding was similar to that observed previously for repressor binding to non-operator DNA using the filter binding assay (9). From this experiment we cannot determine the repress0r:operator stoichiometry in high or low affinity trp repressor/DNA complexes. Dissociation of trp Repressor-Operator Complexes-The possible existence of multiple, high affinity repressor binding sites within the trp promoter-operator region was examined by measuring the kinetics of trp repressorltrp operator complex dissociation at different repress0r:operator ratios (Fig, 2). Purified repressor and radiolabeled operator (1 nM) were mixed at ratios of k10, 1:1, and 10:1, respectively, and dissociation measured in the presence of a 500-fold molar excess of unlabeled operator. When the repressor:operator ratio is low (l:lO), assuming there is no cooperative binding, all of the complexes formed should have one bound repressor molecule. (Using the filter binding assay, we have not observed cooperativity of trp repressor-operator binding, data not shown.) At a high repress0r:operator fragment ratio (1O:l) and an operator fragment concentration of 1 nM complexes of 2 or 3 repressors/operator fragment should form if multiple,

16785

high affinity binding sites arepresent. In this experiment the amount of operator fragment retained on the filters prior to addition of the unlabeled competitor operator was taken as 100% binding for each mixture. The actual percentage of the total available operator bound in repressor-operator complexes at t = 0 was 60% for the high ratio of repress0r:operator (lOl), 44% for the equimolar ratio (l:l), and 6% for the low ratio (1:lO). It is important to note that the shape of the equilibrium binding curve obtained using increasing concentrations of remessor with 1 nM oDerator was essentiallv the same as that shown in Fig. 1 (data not shown). At 1 nM operator fragment concentration and a KO = 2 X 10"' M for each (1,2, or 3) potential high affinity site, most of the binding sites would be occupiedwith bound repressor at thehigh ratio (1O:l) of repress0r:operator. From the results presented in Fig. 2 it is evident that repressor-operator complexes formed in the three mixtures dissociate with equivalent, first-order kinetics, indicating that at theratios of repressor to operator examined, there are no differences in the stability of the high affinity repressor-operator complexes that are formed. The simplest interpretation of this result is that there is one high affinity binding site for trp repressor within the trp promoteroperator region. Equilibrium Binding of trp Repressor to trp Operator in the Presence of Saturating Tryptophan and Corepressor AnalogsT o determine the KOfor the trp repressor-operator interaction in the presence of various ligands, filter binding experiments were performed in which a limiting amount of operator (1 x lo-" M) was equilibrated with increasing amounts of trp repressor, followed by filtration (Fig. 3). The mixtures contained a saturating concentration of either tryptophan (300 pM), 5-methyltryptophan (50 p M ) , or indoleacrylic acid (50 p ~ )The . K D of trp repressor for the operator fragment was 2 X 10"' M in the presence of either tryptophan or 5-methyltryptophan. No repressor-operator complexes were detected when the mixtures containedindoleacrylic acid. Both 5-methyltryptophan and indoleacrylic acid bind with higher affinity to trp aporepressor than does tryptophan (29). These results are similar to those reported by Marmorstein et al. (14),who used a different assay method. However, in that study, under different conditions, a lower KDwas observed for the formation of repressor-operator complexes in the presence of 5methyltryptophan (KO= 1.3 nM) than in tryptophan (Kn= 5.9 nM).

100 1

40

20 0 0.001 0.01

0.1

1

10

100

[Repressor] n M FIG. 3. Determination of KO of trp repressor-operator interaction in the presence of tryptophan and tryptophan analogs using filter binding. Wild type trp repressor and 35S-labeled

operator DNA (1 X lo-" M) were incubated in FBB containing either 50 P M 5-methyltryptophan (A),or 50 P M 300 PM tryptophan (a), indoleacrylic acid (+). The repressor concentration was varied from 0 to 10 nM. Background retention was subtracted, and the datawere normalized to 100%.


16786

Operator Binding by trp Repressor as a Function of Ligand Concentration-From the results presented in the previous section, tryptophan and 5-methyltryptophan function equivalently in stabilizing the complex when present at high concentrations, even though the affinities of these ligands for the aporepressor are not the same. To examine the dependence of repressor-operator complex formation and/or stability on the concentration of available corepressor, filter binding experiments were performed in which the concentrations of trp repressor and operator fragment were held constant and the corepressor concentration was varied (Fig. 4). In agreement with the &fold higher affinity of these ligands for the aporepressor (29), a5-fold higher concentration of tryptophan (20 p M ) than 5-methyltryptophan (4p M ) was required to achieve 50%retention of the radiolabeled operator on the filters. There was no concentration of indoleacrylic acid tested that resulted in the retention of trp repressor-operator complexes on the filters. Determination of Complex Dissociation Rate as a Function of Instantaneous Decreases in the TryptophanConcentration-To determine the effect of decreasing the tryptophan concentration on the rate of repressor-operator dissociation, we used dilution to instantaneously lower the tryptophan concentration of a solution containing trp repressor-operator complexes. There are two possible outcomes for this experiment. 1) Decreasing the tryptophan concentration leads to a more rapid dissociation of the complex, which would indicate that either one or two molecules of tryptophan have dissociated from the complex, resulting in aless stable complex. 2) Decreasing the tryptophan concentrationdoes not change the repressor-operator dissociation rate. This result would imply that repressor-operator dissociation does not depend on tryptophan dissociation. The concentration of tryptophan in the solutions prior to dilution was 60 pM, a concentration shown to be subsaturating (Fig. 4).The half-life of the complex was observed to be 2.5 min when the mixture was diluted into a solution containing the original concentration of tryptophan (plus an excess, 500 nM unlabeled operator). This result is similar to the previously determined complex half-life of 3 min in saturating tryptophan (9).Upon dilution into increasingly lower concentrations of tryptophan, the half-lives of repressor-operator complexes decreased (Fig. 5). Dilution into a solution containing no tryptophan resulted in a half-life of less than 30 s (note in Fig. 6 that dissociation is even more rapid in the presence of indoleacrylic acid, as discussed below).

c

0

! /!I'

60

c.'

40 6)

0" 8

20

-

0 0 1

I1 100

1

10

1000

[Ligand] uM FIG.4. Effect of ligand concentration on the trp repressoroperator interaction using filter binding. Wild type trp repressor (1 X lo-' M) and 35S-labeledoperator DNA (1 X M) were incubated in FBB. The concentration of ligand L-tryptophan (O),5methyltryptophan (A),or indoleacrylic acid (+), was varied from 0 to 300 PM. Background retention (absence of ligand) was subtracted, and the datawere normalized to 100%.

100

T i m e( m i n u t e s ) FIG.5. Half-life determination of the trp repressor-operator complex in response to instantaneous decreases in tryptophan concentration. Wild type trp repressor (10 nM) and 35Slabeled operator DNA (1 nM) were incubated in FBB containing 60 p~ tryptophan. At time zero, this mixture wasmixed with nine volumes of FBB containing 500 nM unlabeled operator and various concentrations of tryptophan. The final concentrationsof tryptophan were: 60 p M (A),35 p M (a),24 p M (+), 12 p M (m), and 6 p M (v). Samples were filtered at various times and theamount of radioactivity retained was normalized to 100%at time zero.

0

1

2

3

4

5

6

T i m e( m i n u t e s ) FIG. 6. Half-life determination of the trp repressor-operator complex in response to dilution into solutions containing the corepressor analogs 6-methyltryptophan and indoleacrylic acid. Wild type trp repressor (10 nM) and 35S-labeledoperator DNA (1 nM) were incubated in FBB containing 60 p M tryptophan. At time zero, this mixture was mixed with nine volumes of FBB containing 500 nM unlabeled operator and either 60 pM tryptophan (O),60 p~ 5-methyltryptophan (A),or 60 p~ tryptophan plus 60 p M indoleacrylic acid (V). Samples were filtered at various times. The amount of radioactivity retained was determined and normalized to 100% at time zero.

These results indicate that when bound tryptophan dissociates from the repressor-operator complex, the stability of the complex is reduced.From these datawe cannot determine whether the principal repressor species that dissociates from the operator upon ligand dilution has one bound tryptophan or no bound tryptophan.

Analogs of Tryptophan Can Exchange with Tryptophan i n the trpRepressor-Operator Complex-To analyze exchange of tryptophan bound in the repressor-operator complex, we determined the rate of complex dissociation following addition of various tryptophan analogs to repressor-operator complexes (Fig. 6). The half-life of repressor-operator complexes was unchanged upon dilution into 5-methyltryptophan.However, dilution of the repressor-operator mixture into asolution of tryptophan and indoleacrylic acid resulted in rapid dissociation of the complex. The initial subsaturating concentration of tryptophan was included with indoleacrylic acid in order to observe specifically the effects of indoleacrylic acid

trp Repressor-Operator Complex Formation and Stability

16787

Wild Type Homodimers [nhl] t L-Trp on thecomplexes. Under these conditions, the dissociation of repressor-operator complexes was morerapid than when the 10 5 2.5 1.25 0.63 0.32 0.16 0.ox 0.M 0 mixture was diluted with a solution containing no tryptophan (Fig. 5). These results indicate that the bound tryptophan 1 molecules inthe repressor-operator complex dissociate rapidly A and are replaced by indoleacrylic acid molecules. Bound indoleacrylic acid molecules then destabilize the repressor-operator complex. Determination of the Affinity of a Hemirepressor for Oper0 39 ator-In order to understand complex destabilization as the first and second molecules of tryptophan dissociate from the repressor-operator complex, we determined the relative affin- B ity of the repressor, a specific hemirepressor,and aporepressor for operator. Since it is impossible to maintain a wild type hemirepressor preparation for biochemicalmeasurements, an unnatural hemirepressor composedof one mutant polypeptide 0 30 chain and one wild type polypeptide chain (presumably containing only one tryptophan binding site) was prepared. The Gly-85 Arg change eliminates tryptophan binding (16) and Gly-85 does not make contact with the operator in the crystalline complex (19). Purified GR85 repressor and wild type repressor weremixed at ratios of 25:l and 5 0 1 and the mixtures treated appropriately to form heterodimers. Under 3Y the conditions used we observed that all of the wild type polypeptide chains were present in WT/GR85 heterodimers. From structural considerations, we assume that the hemirepressor can only bind a single tryptophan molecule, although this has not been demonstrated directly. The two molecules of tryptophan that bind to the wild type trp repressor bind non-cooperatively (6, 15). The wild type and GR85 homodimericrepressors and WT/ 0 GR85 hemirepressor were assayed foroperator binding using the filter binding assay, but both repressors containing GR85 chains bound too weakly to allow accurate measurements (data not shown). Difficulty inmeasuring weak affinity using filter binding has been observed previously(9,34, 35). FIG. 7. Determination of the relative affinity of the trp To measure the relative affinities of the repressor, hemire- repressor, thetrp aporepressor, and a wild type/GRSBhemipressor, and aporepressor for the trp operator we used the repressor for operator using a HpuI protection assay. AutoraHpaI protection assay (32). Competition between trp repres- diograms of HpaI digested mixtures of “*P-labeled pBKH26 DNA sor and a restriction enzyme for the operator has been used with repressor, aporepressor, and hemirepressor. 300 PM tryptophan previously to measure the relative affinities of repressor- was included in the mixtures indicated. A , wild type plus tryptophan; B, wild type without tryptophan; C, GR85 homodimers plus tryptooperator complexes (6,8, 32). We measured relative operator phan; D, GR85 homodimers without tryptophan; E, wild type/GR85 binding affinities in the presence and absence of tryptophan hemirepressor plus tryptophan. (Fig. 7). The concentration of repressor required to give 50% protection was taken as the intrinsic protection value. 50% to dissociation of hemirepressor and formation of wild type protection was observedwith 1nM wild type repressor in the homodimers, becausewild type homodimers wouldhave been presence of saturating tryptophan (Fig. 7 A ) .500 nM wild type readily detected in the filter binding assay. In addition, the aporepressor was required for 50% protection (Fig. 7B). For 251 and 5 0 1 mixtures of GR85 and wild type repressors used the GR85 homodimer, 50% protection was observed at 500 to form hemirepressors behaved identically in this assay. nM protein in the presence or absence of saturating tryptoDISCUSSION phan (Fig. 7 , C and D).This is in accordance with the results of Graddis et al. (26), who found that GR85 repressor does The trp repressor of E. coli regulates transcription initiation not bind tryptophan appreciably and, therefore, that operator in at least four operons concerned with tryptophan metabobinding by GR85 repressor should not be affected by core- lism. Repression results from binding of the trp repressor to pressor. It is important to note that wild type aporepressor operator sequences located within the promoters of the regu(Fig. 7B) and GR85 repressor in the presence or absence of lated operons. We have characterized the formation and statryptophan behaved identically in this assay (Fig. 7, C and bility of trp repressor-operator complexes using the filter D).In the presence of excess tryptophan, the concentration binding assay primarily. The trp operator was originally deof GR85/WT hemirepressor required for 50%protection was scribed as an approximately 20-base pair sequence with an 20-25 nM (Fig. 7E). Theconcentration of hemirepressor used axis of symmetry between A-ll/T-12 of the promoter (see in this experiment was taken as theconcentration of the wild “Experimental Procedures”) (32).It was assumed that binding type polypeptide chain, or two times the original wild type of one repressor dimer to thissequence resulted in repression dimer concentration. Clearly the WT/GR85 hemirepressor (8). Two alternative modelshavebeen presented (26,27) has greater affinity for operator than does the wild type proposing multiple, functional binding sites in the trp operaaporepressor, but has a lower affinity than the wild type tor region and the interaction of trp repressor at these sites. repressor. The hemirepressor binding activity cannot be due Both modelsinvoke higher order repressor-operator com-

C

D

E

16788


plexes. One model proposes that repressor binds at operator sequences other than the symmetrical sequences of the classical operator. Recent gel mobility-shift analyses using the classically defined operator and the alternative postulated operator (27) have provided support for the classical representation of the operator (28). We have used the filter binding assay with pure trp repressor and a 43-base pair synthetic trp promoter-operator (9) to analyze repressor-operator interactions. Saturation of a limiting amountof radiolabeled operator with trp repressor (-10 nM)wasobservedwhen approximately 60% of the input operator was retained on filters. This plateau in operator retention reflected the conditions used high salt binding buffer, 200 pl of sample, and 200 pl of wash. In light of proposed models of trp repressor-operator binding with stoichiometries of 2:l or 3:1,we reexamined the formation and stability of repressor-operator complexes. A biphasic operator binding curve wasobserved when the repressor concentration was increased to approximately 600 nM (Fig. 1). Biphasic binding clearly indicates that there are at least two modes of repressor binding to the operator containing fragment. The apparent K D for high affinity binding was estimated to be 2 X 10"' M. This is the value reported previously (9). The second phase of operator retention on filters occurred with much lower binding affinity (estimated KO = 50-70nM). Similar biphasic binding has been observed in gel mobilityshift analyses (11, 28). It is possible that high affinity binding involves formation of high affinity complexes with two or more repressor dimers per operator fragment. To examine this possibility, dissociation rates were determined with different repressor:operator ratios. Equivalent, first-order dissociation kinetics were observed in all cases (Fig. 2). These results suggest that under our filter binding conditions, trp repressor binds with high affinity to a unique site located within the natural trp promoter-operator region. Subsequent to occupancy of the high affinity site, secondary binding can be observed with much lower affinity. The tryptophan analog 5-methyltryptophanhas been shown to bind more avidly to the trp aporepressor than the natural corepressor, tryptophan (29). Consistent with this result, we observed that a lower concentration of 5-methyltryptophan than tryptophan was required to activate the aporepressor for operator binding (Fig. 4). The KO for the repressor-operator complex was found to be identical, 2 x M, in the presence of a saturating concentration of either tryptophan or 5-methyltryptophan(Fig. 3). The rates of complex dissociation in the presence of excess tryptophan or 5methyltryptophan also were identical (Fig. 6). These results indicate that repressor-operator complexes formed in the presence of a saturating concentration of either corepressing ligand are functionally equivalent. We used dilution experiments to show that tryptophan bound to repressor in the repressor-operator complex dissociates rapidly and is in equilibrium with free tryptophan (Fig. 5). Complexes that lose bound tryptophan dissociate readily. In addition, replacement of tryptophan in the corepressor binding site by indoleacrylic acid resulted in exceptionally rapid dissociation of the complex(Fig. 6), suggesting that bound indoleacrylic acid destabilizes the aporepressor-operator complex. To determine the contribution of each bound tryptophan molecule to repressor-operator complez. stability, we analyzed the operator binding of a heterodimeric repressor (a hemirepressor) with a single tryptophan binding site. The hemirepressor consisted of one wild type polypeptide chain and one

mutant polypeptide chain from a repressor that is unable to bind tryptophan. Using the HpaI protection assay, we found that the affinity of the hemirepressor for operator DNA was intermediate between that of the repressor and the aporepressor. We used the relative affinity values obtained in the HpaI protection assay to calculate the apparent KO values for operator binding by the hemirepressor and the aporepressor. The calculations were normalized to theKO for the repressoroperator interaction (2 X 10"' M) determined using the filter binding assay. The hemirepressor was approximately 22.5fold less avid in operator binding than the repressor; this corresponds to a calculated KO of approximately 4.4 nM. The calculated KO for the aporepressor-operator complex was 100 nM. This value agrees with the value for aporepressor reported by Carey using a gel mobility-shift assay (11). Our results indicate that trp repressor binds at a single, high affinity site within the trp promoter-operator region. Stability of the repressor-operator complex depends on the presence of tryptophan, since free tryptophan exchanges rapidly with tryptophan bound in the complex. The hemirepressor cannot bind operator as avidly as repressor, probably due to a more rapid rate of complex dissociation. In the presence of a subsaturating concentrationof tryptophan, complex dissociation may be due to loss of one of the bound tryptophan molecules andthe formation of a hemirepressor-operator complex. These observations indicate that changes in the intracellular tryptophan concentration can have a profound effect on the ratio of active repressor to inactive aporepressor, as well as altering the stability of already formed repressoroperator complexes. Acknowledgments-We thank Andrzej Joachimiak, Paul Sigler, Jannette Carey, Paul Gollnick, Bob Matthews, Cheryl Arrowsmith, and Oleg Jardetzky for critical reading and comments on previous versions of this manuscript. REFERENCES 1. Squires, C. L., Lee, F. D., and Yanofsky, C. (1975) J. Mol. Biol. 92,93-111 2. Gunsalus, R. P., and Yanofsky, C. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,7117-7121 3. Zurawski, G., Gunsalus, R. P., Brown, K. D., and Yanofsky, C. (1981) J. Mol. Biol. 146, 47-53 4. Heatwole, V. M., and Somerville, R. L. (1991) J. Bacteriol. 1 7 3 , 108-115 5. Sarsero, J. P., Wookey P. J., and Pittmd, A. J. (1991) J. Bacteriol. 1 7 3 , 4133-4143 6. Joachimiak, A., Kelley, R. L.,.Gunsalus, R. P., Yanofsky, C., and Sigler, P. (1983) Proc. Natl. Acad. Scz. U. S. A. 80,668-672 7. Zhane. R.-e.. Joachimiak. A.. Lawson. C. L.. Schevitz. R. W.. Otwinowski. Z.,ind S'ler, P.B. (1987) Nature 327,591-597 8. Klig, L. S., tarey, J., and Yanofsky, C. (1988) J. Mol. Biol. 202,769-777 9. Hurlburt. B. K.. and Yanofskv. C. (1990) J. Biol. Chem. 266.7853-7858 10. Chou, Wi-Y., Bieber, C., and-Matthews, K. S. (1989) J. Biol Chem. 2 6 4 , 18309-18313 11. Carey, J. (1988) Proc. Natl. Acad. Sci. U. S. A. 8 6 , 975-979 12. He, J.-j., and Matthews, K. S. (1990) J. Biol. Chem. 2 6 6 , 731-737 13. Klig, L. S., Crawford, I. P., andYanofsky, C. (1987) Nucleic Acids Res. 16, '

14. Marmorstein, R. Q., and Sigler, P. B. (1989) J. Biol.Chem. 2 6 4 , 9149915A

15. A&ids& D. N., Bruce, C., and Gunsalus, R. P. (1986) J. Biol. Chem. 2 6 1 , 238-243 16. Graddis, T.,Klig, L. S., Yanofsky, C., and Oxender, D. L. (1988) Proteins 4, 173-181 17. Schevitz, R.,Otwinowski, Z., Joachimiak, A,, Lawson, C. L., and Sigler, P. B. (1985) Nature 317,782-786 18. Lawson, C. L., Zhang, R.-g., Schevitz, R. W., Otwinowski, Z.,Joachimiak, A,, and Sigler, P. B. (1988) Proteins 3 , 18-31 19. Otwinowski, Z., Schevitz, R., Zhang, R.-g., Lawson, C. L., Joachimiak, A., Marmorstein, R. Q., Luisi, B. F., and Sigler, P. B. (1988) Nature 3 3 6 , 931-R')Cl V I A V " I

20. Arrowsmith, C. H., Pachter, R., Altman, R., Iyer, S. B., and Jardetzky, 0. (1990) Biochemistry 29,6332-6341 (1990) 21. Arrowsmith, C. H., Pachter, R., Altman, R., and Jardetzky, 0.(1991) Eur. J. Biochem. 202,53-66 22. Harrsion, S. C., and Aggarwal, A. K. (1990) Annu. Reu. Biochem. 69,933o m YI"

23. Arrowsmith, C. H., Carey, J., Treat-Clemons, L., and Jardetzky, 0. (1989) Biochemistry 28,3875-3879 24. Hurlburt, B. K., and Yanofsky, C. (1992) Nucleic Acids Res. 20,337-341 25. Carey, J. (1989) J. Biol. Chem. 264,1941-1945 26. Kumamoto, A. A,, Miller, W. G., and Gunsalus, R. P. (1987) Genes Deu. 1, 556-564 27. Staake, D., Walter, B., Kisters-Woike, B., Wilcken-Bergmann, B. v., and

trp Repressor-Operator Complex Formation Stability and Muller-Hill, B. (1990) EMBO J. 9, 1963-1967 28. Carey, J., Lewis, D.E. A., Lavoie, T.A., and Yaw, J. (1991) J. Bid. C k m . 2 6 6 , 24509-24513 29. Marmorstein, R. Q., Joachimiak, A,, Sprinzl, M., and Sigler, P. B. (1987) J. Bid. Chern. 262,4922-4927 30. Paluh, J. L., and Yanofsky, C. (1986) Nucleic Acids Res. 16, 7851-7860

31. 32, 33. 34, 35.

16789 Kelle , R. L., and Yanofsky, C. (1985) Proc. Natl. Acad. Sci. L'. S. A. 82, 481-487G, N, and Yanofsk Bennett (1978) J , Mol, Bioi, 121,179-192 Vieira j.and hessing, J. (1&7).Methods Enzyymol. 153,3-11 ~ ~ . j.~(1975) ~ ~~t~~~ k , 256 672-674 Takeda, Y., Folkmanis. A., and Echois, H.(1977)J. B i d Chern. 262,61776183