transverse localization of the quinacrine binding site on the

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 9, Issue of March 25, pp. 1334845355,1993 Printed in U. S.A.

Transverse Localization of the Quinacrine Binding Site on the Torpedo Acetylcholine Receptor*

(Received for publication, June 29, 1992)

Hugo R. Arias, C. Fernando Valenzuela, and David A. Johnson$ From the Division of Biomedical Sciences and Department of Neuroscience, University of California, Riverside, Califom& 92521-0121

We demonstrated previously that a phencyclidine- displaceable quinacrine binding site exists at the lipid- protein interface of the Torpedo acetylcholine receptor (AcChR) (Valenzuela, C. F., Kerr, J. A., and Johnson, D. A. (1992) J. Biol. Chem. 267, 8238-8244). In this manuscript, we assess (1) the transverse position of this site in the lipid bilayer by examining the ability of a series of paramagnetic n-doxy1 stearates (n-SALS) and iodide to quench receptor-bound quinacrine and membrane-partitioned octadecyl rhodamine B (Cis- Rho) fluorescence and (2) the stoichiometry of histrionicotoxin- or phencyclidine-displaceable quinacrine binding. Initial experiments established what fraction of the n-doxy1 stearates partitioned into the membranes and that the n-doxy1 stearates do not interfere with quinacrine binding to the receptor at the concentrations used in the quenching studies. The n-doxy1 stearate quenching experiments indicated relatively small (<2) differences between the n-doxy1 stearates to quench receptor-bound quinacrine fluorescence, with a rank order of 7-SAL 2 5-SAL > 12-SAL > 16- SAL. This contrasts with the n-doxy1 stearate quenching of the membrane-partitioned Cis-Rho which showed as much as an 8.6-fold difference between the various isomers with a rank order of quenching efficiencies of 5-SAL > 7-SAL > 12-SAL 1 16-SAL. Iodide quenching measurements indicated significant solute accessibility to membrane-partitioned Cls-Rho but not to receptor-bound quinacrine. The ratios of the bimolecular quenching rate constants for free to bound quinacrine and for free rhodamine B to membrane- partitioned C18-Rho were 53.4 and 6.6, respectively. Direct titration of quinacrine into suspensions of a high concentration of AcChR-associated membranes yielded an upper limit to the binding stoichiometry of 1.4 HTX- or PCP-displaceable quinacrine binding sites/AcChR functional units. The results suggest that there is a single phencyclidine- or histrionicotoxin-displaceable quinacrine binding site located at or somewhat below the level of the Ca-C7 in the phospholipid acyl chains at the lipid-protein interface.

* This work was supported by National Science Foundation Grant IBN-9215105 and California Heart Association Grant 91-140 (to D. A. J.) and by a postdoctoral fellowship from the Argentine Consejo Nacional de Investigaciones Cientificas y TBcnicas (CONICET) (to H. R. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ T o whom correspondence and reprint requests should be ad- dressed Division of Biomedical Sciences, University of California, Riverside, CA 92521-0121. Tel.: 714-787-3831; Fax: 714-787-5504.

Acetylcholine (ACh)’ binds to a specific cell surface membrane protein, the acetylcholine receptor (AcChR), to initiate transient transmembrane cation movements. The AcChR is composed of four types of subunits, designated a, /3, y, and 6, in the stoichiometric ratio of 2:l:l:l (Galzi et al., 1991; Karlin, 1991). Based largely on amino acid hydropathy profiles, each of the five subunits is thought to snake back and forth across the membrane at least four times (Galzi et al., 1991). The four putative transmembrane segments of each subunit are se- quentially labeled M1, M2, M3, and M4. In addition to the binding sites for ACh and competitive antagonists, distinct binding sites exist for noncompetitive inhibitors (NCIs) that act by decreasing the probability of an agonist-induced open- channel state (Taylor et al., 1991). Two classes of NCIs have been described: low affinity and high affinity (Heidmann et al., 1983; Boyd and Cohen, 1984). The low-affinity NCIs are a diverse set of hydrophobic agents that bind to 10-30 unde- termined sites on the AcChR (Heidmann et al., 1983). The high-affinity NCIs are a group of agents that include quinacrine, meproadifen, histrionicotoxin (HTX), ethidium, local anesthetics, and phencyclidine (PCP) and are generally thought to bind with a ligand to AcChR stoichiometry of one/ receptor molecule (Galzi et al., 1991). Much effort has been directed toward the determination of the site of binding of high-affinity NCIs. Electrophysiological (Adams and Feltz, 1980a, 1980b), biochemical (Galzi et al., 1991; Taylor et ab, 1991), and molecular biological (Imoto et al., 1988; Leonard et al., 1988; Oiki et al., 1988; Charnet et al., 1990) data are consistent with the hypothesis that the high-affinity NCIs bind to and plug the lumen of the AcChR. However, it is unclear from affinity and photoaffinity labeling experiments that a single loci of the receptor forms the high-affinity NCI binding site. Photoactivated [3H]chlorpromazine (Giraudat et al., 1986, 1987, 1989; Revah et al., 1990) and [3H]triphenylmethylphosphonium (Hucho et al., 1986; Oberthur et al., 1986) covalently bind to a pair of homologous hydroxyl-containing amino acids located about 2 and 5 residues from the center of the putative lumen-forming M2 membrane-spanning segments, toward the intracellular side of the channel. On the other hand, a photoreactive derivative of quinacrine, [3H] quinacrine azide, covalently binds to amino acids near the beginning of the extracellular portion of the putative M1 membrane-spanning segment of the a-subunits (DiPaola et al., 1990; Karlin, 1991), and a reactive derivative of meproadifen, [3H]meproadifen mustard, attaches to the beginning of

The abbreviations and trivial names used are: ACh, acetylcholine; AcChR, acetylcholine receptor; NCI, noncompetitive inhibitor; HTX, histrionicotoxin; Cch, carbamylcholine hydrochloride; PCP, phencyclidine; dansyltrimetbylamine, dimethylaminonaphthalene-5-sulfon- amidoethyltrimethylammonium; 5-, 7-, 12-, and 16-SAL, 5-, 7-, 12-, and 16-doxy1 stearate positional isomers (n-SALS); Cls-Rho, octade- cy1 rhodamine B chloride.

6348

Transverse Localization of Quinacrine Binding Site 6349

the extracellular side of the a-subunit M2 segment (Dreyer et al., 1986; Pedersen et al., 1992).

Because the affinity labeling approach depends on the presence of reactive groups near the site of binding, the possibility that quinacrine and meproadifen bind to the same loci as chlorpromazine and triphenylmethylphosphonium has been raised (Karlin, 1991). Consequently, an alternative approach that does not require reactive groups near the binding site is needed. Toward filling this need, we have used quan- titative fluorescence spectroscopy to localize the high-affinity NCI binding site. We have determined that the distance between the two ACh sites and the ethidium binding site is between 20 and 40 A (Herz et al., 1989). Moreover, we have demonstrated that PCP-sensitive quinacrine fluorescence is readily quenched by a membrane-partitioning spin-labeled steroid and fluorescent lipid probes, suggesting that receptor- bound quinacrine is at a lipid-protein interface and not in the lumen of the AcChR (Valenzuela et al., 1992).

In this paper, we further localize the PCP-displaceable quinacrine binding site by comparing the efficiencies of a series of paramagnetic n-doxy1 stearate (n-SAL) positional isomers (Fig. 1) to quench PCP-sensitive quinacrine fluorescence. The n-SALS can be used to assess the membrane penetration depth, because they interdigitate between membrane phospholipids and the nitroxide in the doxy1 moiety quenches the emission of any fluorophore within its active sphere (London, 1982; Blatt and Sawyer, 1985). In addition, the solute accessibility to receptor-bound quinacrine was also determined by measuring the ability of the hydrophilic quencher iodide to quench PCP-sensitive quinacrine fluorescence, and the stoichiometry of HTX- and PCP-displaceable quinacrine binding was also assessed by direct stoichiometric titrations.

EXPERIMENTAL PROCEDURES

Materials-Quinacrine dihydrochloride, carbamylcholine hydrochloride (Cch), PCP, and suberyldicholine were obtained from Sigma, 1-dimethylaminonaphthalene-5-sulfonamidoethyltrimethylammon- ium perchlorate (dansyltrimethylamine) from Pierce Chemical Co., 5-, 7-, 12-, and 16-SAL, octadecyl rhodamine B chloride (C18-Rho) and fluorescein were from Molecular Probes (Eugene, OR), and rhodamine B (laser grade) was from Eastman Kodak Co. Potassium chloride and iodide were from Mallinckrodt, Inc. (Paris, KY) and G. Fredrick Smith Chemical Co. (Columbus, OH), respectively. Dimethyl sulfoxide (high-pressure liquid chromatography grade) and 1-butanol were from Aldrich. HTX was generously provided by John W. Daly of the National Institutes of Health and Torpedo californica electric rays were purchased from Marinus Inc. (Long Beach, CA).

Membrane-associated Receptor Isolation-AcChR-enriched membranes were isolated from T. californica electric organs by using differential sucrose gradient centrifugation (Johnson and Yguerabide, 1985). The specific activity of the AcChR preparations were determined by measurement of the decrease of dansyltrimethylamine (6.6 pM) fluorescence produced by the titration of suberyldicholine into suspensions of AcChR (0.3 mg of protein/ml) and PCP (100 p ~ ) following a modification of the method of Neubig and Cohen (1979). The specific activities ranged between 1.0 and 3.6 nmol of suberyldi-

5-SAL 7-SAL 12-SAL 1 &SAL FIG. 1. n-SAL structures.

choline binding sites/mg of protein. Because there are two ACh binding sites/receptor molecule, the AcChR concentration is one-half the concentration of suberyldicholine binding sites. Concentrations of AcChR reported below are in terms of receptor functional units.

Fluorescence Lifetime Measurements-Fluorescence lifetimes were determined by the time-correlated single-photon counting technique with an EEY Scientific nanosecond fluorometer equipped with a high-pressure hydrogen arc lamp. The instrumental arrangement and principles of data treatment have been discussed in detail elsewhere (Yguerabide, 1972).

K,, for Quinacrine Binding to a PCP-displaceable Binding Site- Direct titrations of quinacrine into suspensions of AcChR (0.35 pM) and Cch (1 mM) were used to determine the K d of quinacrine binding at 15 "C to the high-affinity NCI site. Steady-state fluorescence titrations were carried out in 0.5 X 0.5-cm cuvettes with a Perkin- Elmer Cetus MPF 66 spectrofluorometer. Excitation and emission wavelengths were 450 and 502 nm, respectively. A Corning 3-71 cut- off filter was placed in the path of the quinacrine emission beam to reduce stray-light effects. Fluorescence values were corrected for dilution resulting from added titrant and for the intrinsic fluorescence of the sample.

Specific or PCP-sensitive fluorescence associated with the binding of quinacrine to the high-affinity NCI site was defined by the change in quinacrine fluorescence produced by the addition of PCP (0.5 mM). PCP is a nonfluorescent high-affinity NCI that displaces AcChR- bound quinacrine and thus allows the measurement of NCI site- bound quinacrine fluorescence. Estimates of the Kd were made by fitting plots of the specific changes in quinacrine versus the free ligand concentration to the equation for a rectangular hyperbola with the Marquardt algorithm (Marquardt, 1959). All samples were suspended in 10 mM sodium phosphate buffer and 1 mM Cch, pH 7.4 (buffer I). The excess of Cch prevented quinacrine from binding to agonist AcChR binding sites and induced the receptor into a desen- sitization state which displays a higher affinity for quinacrine (Val- enzuela et al., 1992).

Effect of n-SALs on the K,, for Quinacrine-To quantify the competitive inhibition of quinacrine binding to the high-affinity NCI site, the apparent dissociation constant (Kd) of quinacrine in the absence and the presence of various concentrations of 5-SAL (I&') was determined as described above. These results were plotted using the Schild equation (Schild, 1949),

where Kl is the dissociation constant of 5-SAL for the quinacrine binding site and pAP is the negative common logarithm of the concentration of 5-SAL that reduces the apparent affinity of quinacrine by a factor of 2.

Fractional Partitioning of the n-SA& into AcChR-enriched Mem- branes-The various n-SAL (70 pM) in the presence and absence of membrane-associated AcChR (0.35 p ~ ) were suspended in 2 ml of buffer I, incubated for 30 min, and centrifuged for 30 min at 100,OOO X g all a t 15 "C. The amount of the n-SALS in the supernatants was determined by analytical high-pressure liquid chromatography following 2-bromoacetophenone derivatization using methods described elsewhere (Christie, 1987).

Purity and Relative Concentrations of the Stock n-SAL Solutwns- Each n-SAL was weighted separately and solubilized in dimethyl sulfoxide to a final concentration of 12.7 mM. The purity of the n- SAL stock solutions was checked by thin layer chromatography (TLC) using a mixture of cyc1ohexane:ethyl ether:acetic acid (3020:0.5 by volume) as the mobile phase through a stationaryphase of silica gel (Si1 Gel GF; Fisher). n-SAL spots were revealed by the ability of the nitroxide moiety to quench the fluorescence from the stationary phase which was impregnated with a fluorescent indicator. Only single spots of fluorescence quenching were detected from each probe used. Subsequent staining of the TLC plates with iodine vapor revealed the existence of trace (<5%) contaminants with larger RF values.

To assess the relative number of spins per n-SAL stock solution, quenching titrations were performed with solutions of fluorescein (1.0 p ~ ) in butanol for each n-SAL used. Excitation and emission wavelengths were 495 and 526 nm, respectively. The relative quenching efficiencies were determined from the slopes of Stern-Volmer-type plots, i.e. plots of the fluorescence in the absence divided by the fluorescence in the presence of quencher (WZ) versus quencher concentration (see Equation 3). These slopes yielded correction fac-

6350 Transverse Localization of Quinacrine Binding Site

tors of 1.0,0.92, 1.03, and 0.99, for the 5-, 7-, 12-, and 16-SAL isomer stock solutions, respectively. This method was used to assess the relative concentrations of the n-SAL stock solutions, because it was associated with less variability than the traditional double integration of the first derivative ESR spectra approach which can show as much as f 20% variation (Kosman and Bereman, 1981).

Fluorescence Quenching with n-SAG-Quinacrine (0.35 p ~ ) and membrane-associated AcChR (0.35 p ~ ) were suspended in 350 pl of buffer I and kept a t 15 "C 2 PCP (0.5 mM). The samples were then incubated for 1 h and placed in 0.5 X 0.5-cm cuvettes into which the n-SALS were titrated. For the quenching experiments with Cls-Rho, AcChR-enriched membranes containing approximately the same amount of total lipids that were used in the quinacrine quenching experiments were incubated with Cls-Rho (0.27 p ~ ) for 90 min at 15 "C. Excitation and emission wavelengths were 550 and 600 nm, respectively. An internal 610-nm Perkin-Elmer Cetus cut-off filter was placed in the path of the emission beam to reduce stray-light effects.

Each of the four n-SALS (5-, 7-, 12-, and 16-SAL) were dissolved in dimethyl sulfoxide and were added to the cuvettes with a Hamilton syringe. After each addition of titrant (0.5 pl), the samples were incubated for 5 min before fluorescence measurements were made. Initial experiments established that the titrants were at equilibrium with the membrane within 5 min. One set of control samples containing only AcChR-membrane suspension in buffer I monitored the background and titrant fluorescence ( I b b ) . A second set of control samples monitored time-dependent changes in the quinacrine samples.

Because only a fraction ( f = 0.54) of the total quinacrine is bound to the high-affinity NCI binding site on the AcChR, assessment of the extent of quenching produced by the different n-SAL isomers required the separation of the receptor-bound ( IB) from the free quinacrine fluorescence (Valenzuela et al., 1992). Equation 2 was used for this purpose,

IB = CtCd(z-PCP - Ibb) - Cd(1 - f ) ( Z + x P - lblk) (Eq. 2)

where I + w p and I - p ~ p are the magnitudes of fluorescence of samples that did or did not contain PCP, respectively. The c terms are correction factors for time-dependent changes in the receptor-bound quinacrine sample (c,) and for dilution effects of titrants ( 4 . No time-dependent corrections were made for the samples containing PCP, because the time-dependent changes of these samples were found to be insignificant. The fraction of quinacrine that was free in solution (1 - f ) was determined from the plot of the concentration versus PCP-dependent quinacrine fluorescence (Valenzuela et d., 1992). In some control experiments HTX (40 p ~ ) replaced the PCP (0.5 mM).

Iodide Quenching-Solute accessibility to AcChR-bound quinacrine was assessed as the magnitude of the bimolecular quenching rate constant, km, for iodide. A detailed description of the theory and method are provided elsewhere (Johnson and Yguerabide, 1985). Briefly, the fluorescence intensities of the fluorophores were monitored as a function of the iodide concentration. The intensities were then analyzed with the Stern-Volmer equation,

Id1 &[Q1 + 1 (Eq. 3)

where I. and I are fluorescence intensities observed in the absence and presence of a concentration [Q] of iodide. KQ, the Stern-Volmer quenching constant, is related to km by the follwing equation,

KQ = Tkm (Es. 4)

where 7 is the fluorescence lifetime. Because quinacrine binding to the AcChR is sensitive to the ionic

strength of the buffer, the measurement of iodide (KI) quenching of receptor-bound quinacrine fluorescence was performed under conditions of constant ionic strength. The ionic strength was held constant by adjusting the amounts of KI and KC1 added to two sets of cuvettes, one containing PCP (0.5 mM) and one not.

The receptor-bound quinacrine fluorescence (la) was calculated with Equation 2. The value o f f (the fraction of bound quinacrine) was calculated with Equation 5,

I-PCP - Iblk - 1

which requires the use of the previously determined value (2.26) for

the ratio of the quantum yield of receptor-bound to free quinacrine (QB/Qp) (Valenzuela et al., 1992) and determination of the ratios of the quantum yield of free quinacrine in the absence and the presence of either 40 or 80 mM KC1 (Q$/Q#"'). It was necessary to use two salt concentration correction factors (@/@"I), because the experiment was performed at a salt concentration of either 40 or 80 mM. The @/ Q#" for quinacrine in 40 mM KC1 was 1.22 and 1.40 in 80 mM KCl.

Equation 5 is derived from the fact that f equals the observed bound quinacrine fluorescence ((I-pcp - I-bk) - (1 - f) (I+pcp - zbb))

divided by what the fluorescence would be if all the quinacrine was bound ( I + p ~ p - (&/&) (@/e"). The multiplication of (a/ @'I) by (QB/@) corrects for the effects of ionic strength on quinacrine fluorescence.

Because ionic strength has little influence on rhodamine fluorescence, the Cls-Rho iodide quenching measurements were performed by titrating KI into suspensions of membrane-associated AcChR (0.3 p ~ ) that had been preincuhated with 0.27 pM Cls-Rho for 2 h at 15 "C. Because the low solubility of Cls-Rho in buffer, rhodamine B was used at the same final concentration (0.27 pM), in 10 mM sodium phosphate, pH 7.4, to assess iodide accessibility to the fluorophore free in solution. Stock solutions of KI (3.57 M) were prepared in 10 mM sodium phosphate (pH 7.4) and sodium thiosulfate (0.1 mM) was added to minimize the formation of triiodide anion.

Stoichiometry of Quinacrine Binding-To estimate the stoichiometry of quinacrine binding sites on the AcChR, fluorescence titrations f either HTX (40 p ~ ) or PCP (0.5 mM) were performed under conditions where the concentration of quinacrine binding sites was in excess of the & for quinacrine. For these experiments the specific activities of the native membranes were determined by both suberyldicholine back titration as described previously (Valenzuela et d., 1992) and by adsorption of the laI-a-bungarotoxin/AcChR complexes onto DE81 ion exchange filters (Schmidt and Raftery, 1973). The results of both assays were within 5% of one another.

RESULTS

Effect of n-SALs on the Kd for Quinacrine-Because fatty acids can act as noncompetitive inhibitors of the AcChR (Andreasen and McNamee, 1980), we initially sought to find what concentration of n-doxy1 stearates would produce little or no effect on quinacrine binding to the AcChR. Conse- quently, the apparent Kd of quinacrine toward the AcChR was determined in the presence of various concentrations of 5- SAL. Plotted in Fig. 2 is the apparent Kd of quinacrine for the AcChR versus 5-SAL concentration. At and below a concentration of 0.07 mM, &SAL had no significant effect on the apparent & of quinacrine; the values were 0.30 +- 0.11 FM in the presence of 5-SAL (0.07 mM) compared with 0.23 f

2.5

u 2.0

21.5 a

Y w c (d

$1 .o t I 1

0

0

0.0 0.0 0.1 0.2 0.3

[!%SAL], mM FIG. 2. Effect of 5-SAL on the apparent K d of quinacrine

for the PCP-displaceable binding site on the AcChR. The arrow indicates the maximum final concentration that was used in experiments illustrated in Figs. 3-5.

Tramverse Localization of Quinacrine Binding Site 6351

0.13 p~ in its absence (Table I). A Schild plot of these results (inset, Fig. 2) yielded a KI of 160 p~ for 5-SAL. To verify that the other n-SALS used in the quenching experiments had no significant effect on quinacrine binding, the K d of quinacrine for the AcChR was determined in the presence of 0.07 mM 7-, 12-, and 16-SAL. The Kd values were found to be 0.32 f 0.06, 0.20 f 0.08, and 0.27 f 0.05, respectively (Table I). These experiments demonstrated that up to 0.07 mM the n- SALS had no significant effect on quinacrine binding to the high-affinity NCI site.

Fractional Partitioning of n-SALS into AcChR-enriched Membranes-Because n-doxy1 stearates partition differen- tially into lipid bilayers (Blatt et al., 1984), the fractional partitioning of the n-SALS into the AcChR-enriched membranes was determined under the same conditions used for the fluorescence quenching experiments. For 5-, 7-, 12-, and 16-SAL, the fraction of each probe concentration that partitioned into the membranes was 0.96, 0.95, 0.79, and 0.84, respectively. These values were multiplied by the total concentration of n-SAL in each sample cuvette to determine the relative concentration of each spin label in the membranes. This allowed the quenching efficiencies of the n-SALS to be compared.

n-SAL Quenching of AcChR-bound Quinacrine and Mem- brane-partitioned Cls-Rho-Fig. 3 illustrates the uncorrected results of the titration of 5-SAL into suspensions of AcChR

TABLE I Effect of n-SAL positional isomers on the apparent Kd of quinacrine

for the AcChR ~~~ ~ ~

Condition Ib f S.D.” Apparent

Control (no SAL) 0.23 f 0.13 (8) 5-SALb 0.30 f 0.11 (3) I-SALb 0.32 +- 0.06 (3) 12-SALb 0.20 +- 0.08 (3) 16-SALb 0.27 +- 0.05 (3)

‘Apparent dissociation constant of quinacrine for the PCP-displaceable site on the AcChR. Values in parentheses indicate number of determinations.

The total n-SAL concentration was 70 NM added to each cuvette.

TIME (mid 0 4 8 1 2 16 20

90

FIG. 3. Time-dependent and 6-SAL-induced changes in quinacrine fluorescence in the presence of AcChR-associated membranes. 5-SAL was titrated into a suspension of membrane- associated AcChR (0.35 p ~ ) , quinacrine (0.35 p ~ ) , and Cch (1 mM) in the absence (0) or presence (0) of PCP (0.5 mM). A control sample (W) that shows the apparent fluorescence from membrane-associated AcChR and titrant. Time-dependent changes in the AcChR, quinacrine, and Cch suspension without (- - - -) and with PCP (. . . .). For details see “Experimental Procedures.”

and Cch in the presence and absence of quinacrine and excess PCP. The quinacrine fluorescence was -2.3 times greater when it was not prevented from binding to the AcChR by PCP (0.5 mM). Without making any corrections, titration of &SAL produced concentration-dependent quenching that was greater in the absence of PCP (-57%, filled circles). In the presence of PCP (open circles), &SAL produced only an -30% reduction in quinacrine fluorescence. In the absence of quinacrine, 5-SAL had no effect at any concentration tested on background fluorescence (filled squares). In the same figure, time-dependent effects on free (dotted lines) and AcChR- bound (broken lines) quinacrine fluorescence are shown. Only a small (-11%) time-dependent decrease in AcChR-bound quinacrine fluorescence was observed. No significant time- dependent change in quinacrine fluorescence in the presence of PCP was observed. This time-dependent decrease in AcChR-bound quinacrine fluorescence was previously shown not to be due to photobleaching (Valenzuela et aL, 1992).

Following the calculation, using Equation 2, of the receptor- bound quinacrine fluorescence in the absence (Io) and presence of various concentrations of the n-SALS ( I ) , the results were plotted with the Stern-Volmer equation in Fig. 4A. The datum points for the 12-SAL and 16-SAL were practically linear over the concentration range examined. The plots of the effect of 5- and 7-SAL appeared to deviate somewhat from linearity, but this apparent nonlinearity probably results from the variability of the technique. In agreement with this is the linearity observed for 5-SAL in parallel experiments illustrated in Fig. 5. The linear slopes of these plots, which are

0 20 40 60 80

[SAL], pM FIG. 4. Stern-Volmer plots of n-SAL quenching of AcChR-

bound quinacrine and rhodamine fluorescence. The results represent the average of three determinations of the fluorescence in the absence (Io) divided by in the presence ( I ) of various concentrations of the n-SALS ([SAL]). The n-SAL concentrations were corrected for their fractional partitioning into AcChR-enriched membranes. A, receptor-bound quinacrine fluorescence determined by

AcChR (0.35 p ~ ) containing both quinacrine (0.35 PM) and Cch (1.0 measuring the fluorescence of a suspension of membrane-associated

mM) * PCP (0.5 mM) and then calculated with Equation 2. E , fluorescence from the Cls-Rho (0.27 @M) partitioned into AcChR- enriched membranes (0.35 pM in receptors). The various n-SAL positional isomers are 5-SAL (O), 7-SAL (0) 12-SAL (W), and 16- SAL 0.


1

1

+I

1

1

1

0 25 50 75

[5-SALl, pM FIG. 5. Comparison of 6-SAL quenching of HTX- and PCP-

sensitive quinacrine fluorescence in the presence of AcChR. HTX- or PCP-sensitive quinacrine fluorescence determined by measuring the fluorescence from native membrane-associated AcChR (0.35 pM in functional units) containing both quinacrine (0.35 p ~ ) and Cch (1.0 mM) were measured in the absence and the presence of either HTX (40 PM) (0) or PCP (0.5 mM) (0). The results represent the average of three determinations of the calculated receptor-bound quinacrine fluorescence in the absence (I.,) divided by in the presence ( I ) of various concentrations of the 5-SAL. For details see “Experi- mental Procedures.”

TABLE I1 Paramagnetic quunching with n-SAL positional isomers of AcChR- bound quinacrine and membrane-partitioned Cts-Rho fluorescence

AcChR-bound Membrane-partitioned n-SAL quinacrine Cle-Rho

Slope” llslowb Slope” l/slopeb

rnM’ mM rnN’ mM 5-SAL 8.8 f 0.8 0.11 6.0 f 0.2 0.17 7-SAL 9.9 f 0.3 0.10 5.0 f 0.3 0.20 12-SAL 5.9 k 0.2 0.17 0.9 k 0.1 1.1 16-SAL 5.1 k 0.3 0.20 0.7 -+ 0.1 1.3

The slope from plots of I,/I versus [SAL] in Fig. 3 which is equal to the apparent Stern-Volmer quenching constant in Equation 4.

The n-SAL concentration at which the initial fluorescence inten- sity is quenched 50%. Statistical tests of parallelism of the quinacrine quenching data indicated that the 5- and 7-SAL lines are significantly different from 12- and 16-SAL lines ( p < 0.01). The 12- and 16-SAL lines were somewhat less significantly different from one another (p < 0.05). The 5- and the 7-SAL were not significantly different from one another ( p < 0.25).

approximately proportional to the quenching efficacy of each n-SAL, are shown in Table 11. The rank order of the quenching efficiencies of the n-SALS was 7-SAL ? 5-SAL > 12-SAL > 16-SAL. The 7-SAL displayed about two times the quenching efficiency of 16-SAL. Statistical tests of parallelism of the plots indicated that the 5- and 7-SAL lines are significantly different from 12- and 16-SAL lines ( p < 0.01). The 12- and 16-SAL lines are somewhat less significantly different form one another ( p < 0.05). The 5- and the 7-SAL were not significantly different form one another ( p < 0.25).

To partially validate this approach and to get a sense of the meaning of the relative quenching efficiencies of the various n-doxy1 stearates, the efficiencies of the n-SAL isomers to quench the fluorescence of membrane-partitioned ClS-Rho were also measured. The C18-Rho is composed of a positively charged fluorophore and an 18-carbon acyl chain. The charged fluorophore should be expected to position at or near the lipid-water interface and the acyl chain in the hydrocarbon core of the lipid membrane. The Z,/Z versus n-SAL concentration plots are shown in Fig. 4B. As with the receptor-bound quinacrine quenching results, the plots from each n-SALS titration were linear over the concentration range examined.

The slopes of these plots (Table 11) indicate that the rank order of the quenching efficiencies of the n-SALS examined was 5-SAL > 7-SAL > 12-SAL 2 16-SAL. The 5-SAL displayed about 9 times the quenching efficacy of 16-SAL isomers. These results are consistent with the positioned of fluorescent moiety of the ClS-Rho at or near the lipid-water interface.

Taken together the quinacrine and ClS-Rho results suggest that the fluorophore (acridine moiety) of receptor-bound quinacrine is located deeper into the lipid membrane than the rhodamine moiety of C18-Rho, because 12- and 16-SAL are more effective at quenching receptor-bound quinacrine than membrane-partitioned C18-Rho. Similarly, the acridine moiety of receptor-bound quinacrine is not positioned at the level of C12-C16 in the phospholipid chains, because 5- and 7- SAL are about two times more effective at quenching receptor-bound quinacrine fluorescence than 12- and 16-SAL. The position of the acridine moiety would thus appear to be at or somewhat below the level of C6-C7 in the phospholipid chain.

Specificity of PCP Displacement of Receptor-bound Quina- crine-PCP binds to both the high- and low-affinity NCI sites on the AcChR (Heidmann et al., 1983); consequently, it is possible that the above results examining PCP-sensitive fluorescence reflect quinacrine being displaced from low-affinity NCI sites. Because HTX has been previously shown to bind more selectively than PCP to the high-affinity NCI sites, we compared the ability of PCP (0.5 mM) and HTX (40 FM) to displace receptor-bound quinacrine by comparing the 5-SAL quenching results obtained using either PCP or HTX as the displacing agent. Stern-Volmer plots of 5-SAL quenching of receptor-bound quinacrine fluorescence measured with either 0.5 mM PCP or 40 PM HTX as the displacing agent (Fig. 5) indicated no significant difference between either displacing agent. Although PCP treatment produced slightly greater (9%) quenching, this value is within the intrinsic error of this method.

As an aside, it is worth noting that the magnitude of the 5- SAL-induced quenching is slightly lower in Fig. 5 than in Fig. 4. This reflects the fact that an AcChR preparation with a lower specific activity was utilized for the data illustrated in Fig. 5. Lower preparation specific activities are associated with a higher ratio of lipid to AcChR molecules and, therefore, a lower concentration of 5-SAL in the lipid membrane and thus less quenching.

Iodide Quenching-To provide evidence supporting the location of the quinacrine binding site below the lipid membrane surface, the solute accessibility to receptor-bound quinacrine was assessed for AcChR-bound and free quinacrine. For the purpose of comparison the solute accessibility to an AcChR membrane-partitioned fluorescent probe (C18-Rho) and a water-soluble analog (rhodamine B) that was free in solution were also assessed. The bimolecular quenching rate constants, kFB values, were calculated using Equation 4. The fluorescence decays of free and bound quinacrine as well as membrane- partitioned C18-Rho were adequately fitted to double exponential functions (data not shown); consequently, geometric average fluorescence lifetimes (<T>) were used in the calculation of k F ~ (Table 111).

Over the concentration range examined, all the Stern- Volmer quenching plots were basically linear (Fig. 6) and consistent with a dynamic quenching mechanism. The bimolecular quenching rate constants are summarized in Table 111. When quinacrine is bound to the PCP-displaceable binding site, the kFQ is 53.4-fold less than when free in solution and nonspecifically in the membrane (Table 111). If we assume that quinacrine binds to a nonluminal site at the lipid-protein

Transverse Localization of Quinacrine Binding Site

TABLE 111

6353

Iodide quenching of free and bound quinacrine, rhodamine B, and membrane-partitioned Cla-Rhn fluorescence Fluorophore State KQ ( 7 Y kre X lo-' ' k z . / k v '

hf" n.s " .Q-1

Quinacrine Freed 11.6 f 0.6 3.8 & 0.1 3.1 Quinacrine AcChR-bound 0.5 f 0.7 8.6 f 0.5 0.058 53.4 Rhodamine B Free 4.8 f 0.1 0.8 f 0.1 6.0 Cla-Rho Membrane-partitioned 1.0 & 0.1 1.5 f 0.2 0.91 6.6

a With the exception of rhodamine B in buffer, the fluorescence decays were adequately fitted to biexponential functions; consequently, the fluorescence lifetimes are given as the geometric average of the two lifetimes. The rhodamine B decay in buffer was adequately fitted to only a single exponential function.

Biomolecular association rate constant calculated using Equation 4.

Free here means not bound to the high-affinity NCI site. PCP (0.5 mM) was added to suspension of AcChR (0.35 p ~ ) , Cch (1 mM), and quinacrine (0.35 p ~ ) to prevent quinacrine binding to the high-affinity NCI site. The fluorescence quenching in the presence of PCP reflects an average signal from quinacrine in solution and nonspecifically in membranes.

e Ratio of kw of fluorophores in the free over bound (or membrane-partitioned) states.

2.01 ' I ' ' ' ' ' 'A

' ;::IB 1.2 ,..? l.l 1 .o P 0.9

0 20 40 60 80 [KII, mM

FIG. 6. Stern-Volmer plots of iodide quenching of quinacrine and rhodamine fluorescence. The results represent the average of two determinations of the fluorescence in the absence (Z,) divided by in the presence ( I ) of various concentrations of the KI. For details see "Experimental Procedures." A, quinacrine free in solution (m) and bound to the PCP-displaceable NCI binding site on the AcChR (0). B, rhodamine B free in solution (m) and Cla-Rho membrane-partitioned in the lipid membrane (0).

interface and that this site does not exclude anions, our findings are consistent with the AcChR-bound quinacrine being completely inaccessible to the buffer. The km of the fluorescent moiety of the membrane-partitioned Cls-Rho, on the other hand, is only 6.6-fold less than its water soluble derivative rhodamine B in solution. Perhaps half of this relatively small reduction can be explained by the loss of rotational freedom of a fluorophore when it goes from a highly mobile state in solution to a more rotationally constrained state partitioned into the membrane (Johnson and Yguera- bide, 1985). Consequently, the results are consistent with the fluorescent moiety of Cls-Rho being partially accessible to solutes.

Stoichiometry of the High-affinity Quinacrine Binding-To provide further support for the specificity of the HTX and PCP displacement of receptor-bound quinacrine, the number of quinacrine binding sites/receptor was estimated by direct titration of quinacrine into suspensions of membrane-associated AcChR which was at a concentration that would ensure near stoichiometric binding. We previously estimated the Kd of PCP-sensitive quinacrine binding to be 0.14 p~ (Valenzuela et al., 1992); consequently, quinacrine was titrated into a

QUINACRINE lAcChR FUNCTIONAL UNITS

FIG. 7. Stoichiometry of quinacrine binding. Quinacrine was titrated into a suspension of membrane-associated AcChR (1.4 p~ in functional units) and CCh (1 mM) f HTX (40 p ~ ) as described under "Experimental Procedures." The solid lines represent the extrapola- tion of the linear increase in HTX-sensitive fluorescence to the level reached at saturation and yield an estimate of the number of quinacrine binding sites/AcChR functional units to be 1.4 & 0.1 (average of three determinations).

suspension of Torpedo membranes that was 1.4 p~ in functional units, i.e. 210-fold greater than its apparent K d . Fig. 7 illustrates the HTX-sensitive changes in quinacrine fluorescence upon titration into this concentration of receptor. Ex- trapolation of the linear increase in HTX-sensitive fluorescence to the level reached at saturation, yielded a cyinacrine concentration at saturation of 1.4 f 0.1 quinacrine binding sites/AcChR functional units (three determinations). Similar titrations (data not shown) were performed using 0.5 mM PCP to displace quinacrine from the high-affinity NCI sites and a value of 1.4 k 0.2 quinacrine binding sites/AcChR functional units (two determinations) was obtained.

DISCUSSION

In this paper, we took advantage of the ability of paramagnetic n-SALS to interdigitate between the membrane phospholipids and sense the local environment at different depths in the membrane. Specifically, the transverse position of the high-affinity NCI binding site for quinacrine on the AcChR was examined by comparing the efficiencies of a series of n- doxy1 stearates to quench PCP-sensitive quinacrine fluorescence. We found, after correction for differences in the membrane partition coefficients of the n-SALs, that the rank order of the quenching efficiencies was 7-SAL 2 5-SAL > 12-SAL > 16-SAL and that the difference between the quenching efficiencies of the various n-SALS is relatively small (<2).


Such a pattern of quenching would place the quinacrine binding site at or somewhat below the level of Cs-C7 in the phospholipid acyl chains (Fig. 8). The ineffectiveness of iodide to quench receptor-bound quinacrine fluorescence is consistent with quinacrine binding below the hydrocarbon-polar head group interface. Although quinacrine can bind to both the agonist and the high-affinity NCI binding sites (Tsai et al., 1979), sufficient Cch was added to all samples to prevent binding to the ACh binding sites, so it is unlikely that the results reflect quinacrine binding to the agonist binding sites. Also, the prolonged presence of Cch stabilized the receptor in the slow-onset desensitized state (Sugiyama et al., 1976; Wei- land et al., 1977), consequently, the present conclusions refer to only this functional conformation.

In addition to assessing the transverse position of the quinacrine binding site, an upper limit of the quinacrine- binding stoichiometry was also estimated to be 1.4 sites/ AcChR functional units. The approach to measure this stoichiometry involved the direct titration of quinacrine into a high concentration of membrane-associated AcChR (1.4 p ~ ) +. HTX or +- PCP. The 'inflection point at which added quinacrine produced no significant increase in HTX- or PCP- sensitive fluorescence defined the maximum number of quinacrine binding sites in the AcChR suspension. This approach assumes quinacrine binds stoichiometrically to the high-affinity NCI binding sites. If there is significant nonspecific partitioning of quinacrine into the lipid membrane along with specific binding, then there would be less quinacrine available to bind to the high-affinity NCI sites (ligand depletion), and, consequently, the number of binding sites would be over estimated. Because it is reasonable that an amphipathic molecule like quinacrine nonspecifically partitions into membranes, it is likely that our stoichiometric estimate is high. Therefore, there is probably only one quinacrine site/AcChR functional unit, which would make quinacrine binding stoichiometry the same as other high-affinity NCIs (Galzi et al., 1991).

The above conclusions are based on several assumptions that must be weighed in the interpretation of the results. First, the n-SALS are assumed to be oriented with their polar carboxylate moieties at the water-lipid interface and the hydrocarbon acyl chains aligned normal to the membrane surface. The existence of a flexibility gradient associated with the position of the doxy1 group along the acyl chains of the Torpedo membrane-incorporated n-SALS is consistent with

Transverse Cross-Section n

FIG. 8. Schematic representation of a cross-section of the AcChR and the surrounding lipids, illustrating the transverse location of the HTX- or PCP-displaceable quinacrine binding site.

this assumption (Marsh and Barrantes, 1978; Andreasen and McNamee, 1980). Also, we assumed that the n-SALS maintain this orientation when associated with the AcChR. We are unaware of any evidence which supports this assumption other than the fact that no positively charged residues exist in the putative transmembrane segments of the receptor (Noda et al., 1982, 1983; Claudio et al., 1983) which could attract and thereby alter the orientation of the negatively charged n-SALS.

Additionally, we assumed that the n-SALS are inaccessible to the lumen of the ion channel. Evidence supporting this assumption comes from unpublished observations comparing quinacrine and ethidum, another fluorescent NCI (Herz et al., 1987). Specifically, PCP-sensitive ethidium fluorescence, un- like quinacrine, is not quenched by 5-SAL or spin-labeled androstane. Moreover, 5-SAL inhibits PCP-sensitive quinacrine but not ethidium binding. These results point to the existence of both luminal and nonluminal high-affinity NCI binding sites that are mutually coupled. n-SALS and spin- labeled androstane do not appear to be accessible to luminal NCI binding sites.

Another issue is the specificity of the PCP displacement of quinacrine from the receptor. PCP binds to both high and low-affinity NCI sites. The K d and stoichiometry of PCP binding to the high-affinity site are 0.4-0.8 p~ and one site/ receptor, respectively (Heidmann et al., 1983; Herz et al., 1987). The K d of PCP for the low-affinity sites is 170 pM (Heidmann et al., 1983). Clearly, the concentration of PCP added in the various experiments described here may have displaced some quinacrine bound to low-affinity NCI sites. However, given the quinacrine concentration used (0.35 p ~ ) and assuming that quinacrine and PCP display the same K d

for the low-affinity sites (170 pM), the equation for simple competitive antagonism would predict that 0.5 mM PCP would displace less than 1% of the quinacrine bound to the low- affinity sites. Moreover, essentially the same magnitude of 5- SAL quenching of receptor-bound quinacrine fluorescence was obtained when either PCP or HTX was used as a quinacrine-displacing agent (Fig. 5). Similarly, the estimated stoichiometry of quinacrine binding to the high-affinity NCI site was the same when either HTX or PCP was used as a displacing agent, further demonstrating the specificity of PCP (0.5 mM) displacement of quinacrine binding to the high- affinity NCI site.

Other evidence supports the existence of a high-affinity NCI binding site at the lipid-protein interface. First, the close proximity between fluorescent lipid probes and spin-labeled androstane to AcChR-bound quinacrine indicates the quinacrine binding site is at a lipid-protein interface (Valenzuela et al., 1992). Moreover, [3H]quinacrine azide covalently binds toward the synaptic side, at the beginning of the a-subunit M1 transmembrane domains (Karlin, 1991). The M1 domain appears to interact with the lipid domain, because it is pho- tolabeled by hydrophobic 3-trifluoromethyl-3-(n-['251]iodo- pheny1)diazirine and 1-azidopyrene (Blanton and Cohen, 1992). Thus, the results with [3H]quinacrine azide are consistent with a site of high-affinity NCI binding at the lipid- protein interface.

The observation that ['Hlquinacrine azide labels the beginning of the M1 domain and Pro211) of the a-subunits (Karlin, 1991) suggests that the quinacrine binding site is positioned a bit closer to the water-lipid interface than sug- gested by the present results. This difference may be due a lack of reactive groups at the quinacrine binding site. Argm and Pro211 may simply be in the migration path to reach the real quinacrine binding site which is below the hydrocarbon-

Transverse Localization of Quinacrine Binding Site 6355

polar head group interface. This discrepancy can also be explained by the separation between the reactive azido and fluorescent acridine moieties. If quinacrine azide is in an extended conformation, the center of the acridine can be as much as 5 A from the azido moiety. Alternatively, the quinacrine binding site may shift relative to the water-lipid interface depending on the receptor conformational state. The present results deal only with the slow-onset desensitized state of the receptor, whereas optimal [3H]quinacrine azide labeling occurs to the open-channel state of the AcChR (DiPaola et al., 1990; Karlin, 1991).

Although there is now significant support for the existence of a high-affinity NCI site at the lipid-protein interface, all high-affinity NCIs, as discussed above, may not be binding to this site. Consequently, high-affinity NCIs appear to bind in a mutually exclusive manner to distinct luminal or nonluminal sites on the receptor. An area for further research will be to determine how these high-affinity NCI binding sites interact with each other and what can be deduced from this analysis about the opening of the channel by agonists.

Acknowledgment-We thank Thu Doan for the preparation of AcChR associated membranes.

REFERENCES Adams, P. R., and Feltz, A. (1980a) J. PhysioL ( L o n d . ) 306,261-281 Adams, P. R., and Feltz, A. (1980b) J. Physiol. ( L o n d . ) 306,283-306 Andreasen, T. J., and McNamee, M. G. (1980) Biochemistry 19,4719-4726 Blanton, M. P., and Cohen, J. B. (1992) Biochemistry 31,3738-3750 Blatt, E., and Sawyer, W. H. (1985) Biochim. Biophys. Acta 8 2 2 , 43-62. Blatt, E., Chatelier, R., and Sawyer, W. H. (1984) Photochem. Photobwl. 39 ,

Boyd, N. D., and Cohen, J. B. (1984) Biochemistry 23,4023-4033 Chamet, P., Labarca, C., Leonard, R. J., Vogelaar, N. J., Czyzyk, L., Gouin, A,,

Christie, W. W. (1987) HPLC and Ltplds, Pergamon Press, Oxford Claude, T., Ballivet, M., Patnck, P., and Hememann, S. (1983) Nature SO,

DiPaola, M., Kao, P. N., and Karlin, A. (1990) J. Biol. Chem. 266, 11017-

Dreyer, E. B., Hasan, F., Cohen, S. G., and Cohen, J. B. (1986) J. Biol. Chem.

477-482

Davldson, N., and Lester, H. A. (1990) Neuron 2,87-95

1111-1115

11029

261,13727-13734

Galzi, J.-L., Revah, F., Bessia, A., and Changeux, J.-P. (1991) Annu. Rev.

Giraudat, J., Dennis, M., Heidmann, T., Chang, J. Y., and Changeux, J.-P.

Giraudat, J., Dennis, M., Heidmann, T., Haumont, P.-Y., Lederer, F., and

Giraudat, J., Galzi, J.-L., Revah, F., Changeux, J.-P., Haumont, P.-Y., and

Heidmann, T., Oswald, R. E., and Changeux, J.-P. (1983) Biochemistry 2 2 ,

Hen, J. M., Johnson, D. A., and Taylor, P. (1987) J. Biol. Chem. 262 , 7238-

Pharmacol. 31,37-72

(1986) Proc. Natl. Acad. Scz. U. S. A. 8 3 , 2719-2723

Changeux, J.-P. (1987) Biochemistry 26,2410-2418

Lederer, F. (1989) FEBS Lett. 263,190-198

3112-3127

7347 He4;j . M., Johnson, D. A., and Taylor, P. (1989) J. Biol. Chem. 264,12439-

Hucho, F. L., Oberthiir, W., and Lottspeich, F. (1986) FEES Lett. 206 , 137-

Imoto K. Busch C., Sakmann, B., Mishina, M., Konno, T., Nakai, J. M., Bujo,

Johnson, D. A., and Yguerabide, J. (1985) Bzophys. J . 48,949-955

Kosman, D. J., and Bereman, R. D. (1981) in Spectrosco in Biochemistry Karlin, A. (1991) Haruey Lect. 8 6 , 71-107

Leonard, R. J., Labarc;, 8 G.,Charnet, P., DAvidson, N., and dester, H. A.

London, E. (1982) Mol. Cell. Biochem. 46,181-188 Marquardt, D. W. (1959) Chem. Eng. Prog. 66,65-70 Marsh, D., and Barrantes, F. J. (1978) Proc. Natl. Acd. Sci. U. S. A. 76,4329-

Neubig, R. R., and Cohen, J . B. (1979) Biochemistry 18,5464-5475 Noda, M., Takahashi, H., Tanabe, T., Toyosato, M., Furutani, Y., Hirose, T.,

Asai, M., Inayama, S., Miyata, T., and Numa, S. (1982) Nature 299 , 793-

12448

142

H., Moii, Y., Fukuda, M., and Numa, S. (1988) Nature 336,645-648

(Bell, J. E., ed) Vol. 2 p 57 108, CRC Press Inc., Bocaxton FC

(1988) Science 242,1578-1581

4333

797 NO& M., Takahashi, H., Tanabe, T., Toyosato, M., Kikyotani S., Hirose, T.

Asai, M., Takashima, H., Inayama, S., Miyata, T., and N b a , S. (19833

Oberthiir, W., Miihn, P., Baumann, H., Lottspeich, F., Wittmann-Liebold, B., Nature 301,251-255

Oiki, S., Danko, W., Modison. V.. and Montal. M. (1988) Proc. Natl. Acad. Sci. and Hucho, F. (1986) EMBO J. 6,1815-1819

[I. S. A. 86.8703-8707 PedeGen, S. E., Sharp, S. D., Liu, W.-S., and Cohen, J. B. (1992) J. Biol. Chem. 267.10489-10499

Reiah F., Galzi, J. I.., Giraudat, J., Haumont P., Lederer, F., and Changeux,

Schild, H. 0. (1949) Brzt. J . Pharmacol. 4.277-280

~, -~ "~ " - - -

J.-P: (1990) Proc. Natl. Acad. Sci. U. S. A. S i , 4675-4679

Schmidt. d. . and Raftnry, M. A. (1973) A d . Biochem. 62,349-355 J. L., and Chaneeux. J.-P. (1976) J. Mol. Biol. 106.485-

T a ~ lor P. Abramson S. Johnson D. A. Valenzuela, C. F., and Herz, J. M.

Tsai, M.-C., Oliveira, A. C., Albuquerque, E. X., Eldefrawi, M. E., andEldefrawi,

Valenzuela, C. F., Kerr, J. A., and Johnson, D. A. (1992) J. Biol. Chem. 2 6 7 ,

899;) Ann. N. Y. Acd. Sei. 266,568-687

A. T. (1979) Mol. P h a r m o l . 16,382-392

823R-8244 Wei!andG., Georgia, B., Lappi, S., Chignell, C. F., and Taylor, P. (1977) J.

Yguerablde, J. (1972) Methods Enzymol. 26,498-578 Bwl C k m . 262,7648-7656

transverse localization of the quinacrine binding site on the

Documents