[3H]1‐[2‐(2‐thienyl)cyclohexyl]piperidine labels two high‐affinity binding sites in human cortex: Further evidence for phencyclidine binding sites associated with the biogenic amine reuptake complex

Previous work demonstrated two high‐affinity PCP binding sites in guinea pig brain labeled by [3H]TCP (1‐{1‐[2‐thienyl]cyclohexyl}piperidine): site 1 {N‐methyl‐D‐aspartate [NMDA]‐associated} and site 2 {dopamine‐reuptake complex associated}. The present study examined brain membranes prepared from various species, including human, for the presence of site 2, defined as binding in the presence of (+)‐5‐methyl‐10, 11‐dihydro‐5H‐dibenzo[a,d]cyclohepten‐5, 10‐imine maleate ((+)‐MK801) minus binding in the presence of 10 μM TCP (nonspecific binding). Studies were conducted in absence of sodium which was found to be inhibitory to [3H]TCP binding. The results demonstrated detectable levels of site 2 in brain membranes of guinea pig, rabbit, pig, mouse, sheep, and human but not in the rat or chicken. Using human cortical membranes, site 2 was the predominant binding site. Detailed studies conducted with human cortical tissue showed that high‐affinity dopamine {1‐[2‐]bis(4[fluorophenyl)[methoxy]ethyl]4‐(3‐phenylpropyl)piperazine (GBR12909)}, [1, 2]benzo(b)thiophenylcyclo‐hexylpiperidine (BTCP), and serotonin (fluoxetine) uptake inhibitors produced a wash‐resistant inhibition of [3H]TCP binding to site 2, but not site 1. Preincubation of guinea pig brain membranes with BTCP was shown to produce an increase in the dissociation rate of [3H]TCP from PCP site 2. Structure activity studies with various uptake inhibitors showed that GBR12909, benztropine, fluoxetine, and BTCP have higher affinity for site 2 than for site 1. (+)‐MK801, ketamine, and tiletamine were very selective for site 1, whereas dexoxadrol and TCP were moderately selective for site 1. These results suggest that human cortex possesses high‐affinity PCP binding sites associated with biogenic reuptake binding sites, and that guinea pig brain, but not rat brain, may be an appropriate animal model for studying PCP site 2 in human brain.

Some reports indicate that the pharmacologically relevant actions of PCP are mediated via receptors PUBLISHED 1991 WILEY-LISS, INC. associated with the NMDA receptor (Jarvis et al., 1987;Snell et al., 1988), and that interactions with other CNS receptorshinding sites occur at much higher doses (Jarvis et al., Koek and Woods, 1988a,b). However, in vivo effects of PCP which are mediated via the NMDA receptor such as its anticonvulsant (Leander et al., 1988) and neuroprotective (Sauer et al., 1988) effects, occur at the same dose ranges as effects of PCP which are probably mediated by blockade of dopamine reuptake, such as hypothermia (Pechnick et al., 1989a,b), elevation of prolactin levels (Pechnick et al., 19898, as well as increases in extracellular dopamine as measured by in vivo microdialysis (Carboni et al., 1989).
These data, and the good correlation between the affinity of arylcycloalkylamines for high-affinity PCP binding sites and their EDs0 values for inhibition of dopamine reuptake (Johnson and Snell, 1985;Vignon and Lazdunski, 19841, necessitates examining the hypothesis that at least two PCP binding sites can be detected in ligand binding studies. Published studies report either a single high-affinity binding site (Quirion et al., 1981;Zukin and Zukin, 1979) or multiple binding sites (Chicheportiche et al., 1988;Mendelsohn et al., 1984;Vignon et al., 1986, Vignon et al., 1989. In support of the two site model, a recent study from our laboratory using membranes prepared from guinea pig brain demonstrated a second high-affinity PCP binding site (site 2), which appeared to be associated with the dopamine reuptake carrier in caudate membranes (Rothman et al., 1989).
In the present study, we examined membranes prepared from the brains of various species for the occurrence of PCP site 2 and found detectable levels of site 2 in brain membranes of guinea pig, rabbit, pig, mouse, sheep, and human but not in the rat or chicken. Binding surfaces generated using [3HJTCI' to label the PCP binding site (Vignon et al., 1983) in conjunction with (+ )-MK801 to provide selectively for the NMDA-associated PCP binding site (Wong et al., 19861, permitted the resolution of two 13HlTCP binding sites. The observation that high-affinity biogenic amine reuptake inhibitors produce a wash-resistant inhibition of [3H]TCP binding to site 2, but not to site 1, provides additional evidence that PCP site 2 in human cortex is associated with biogenic amine reuptake complex.

MATERIALS AND METHODS
Preparation of membranes Frozen brains or tissues were thawed and dissected as necessary into cortex, caudate, hippocampus, or kept as whole brain. Tissues were homogenized using a polytron for I minute (setting #5) in 5 ml ice cold buffer (50 mM Tris HCI, pH 8.0) per gram wet weight of tissue. The homogenates were then centrifuged at 39,OOOg for 10 minutes using an SA-600 rotor (Sorvall Instruments). The supernatant was discarded and the pellet was resuspended by brief poIytroning in the same voIume of buffer. This was centrifuged again at 39,OOOg for 10 minutes and the supernatant was discarded. The pellet was resuspended in 8 ml buffer per gram wet weight and incubated a t 0°C for 15 minutes. The membranes were then washed four times by 10 minute centrifugations at 30,00Og, with resuspensions in the same volume of buffer.
Following the final centrifugation, pellets were resuspended in 0.5 ml ice-cold 5 mM Tris-HC1, pH 8.0, per gram wet weight. One milliliter aliquots were distributed into microcentrifuge tubes, which were stored at -80°C. Rat lysed P2 membranes were prepared as previously reported from our laboratory (Rothman et al., 1988a).
For the wash-resistant inhibition studies, human cortical membrane pellets stored at -80°C (see above) were thawed and resuspended in 55.2 mM sodium phosphate, pH 7.4 (0.5-1 mg/ml protein), which was chosen to optimize binding to the biogenic amine transporter. Similar buffers are commonly used in binding assays for labeling biogenic amine transporters (Reith et al., 1984). Membranes were then incubated for 60 minutes at 0°C with various drugs, under conditions which should optimize their binding to the DA or serotonin transporters, after which they were centrifuged at 9,5009 for 10 minutes. Membranes were then washed twice by centrifugation with the same volume of ice-cold 5 mM Tris-HC1 (pH 8.0) at 9,500g for 10 minutes. The pellets were used immediately for assay following resuspensions in same volume of ice-cold 5 mM Tris-HC1, pH 8.0.
In related wash-resistant inhibition experiments, termed "supernatant experiments," guinea pig brain membranes were preincubated with concentrations of BTCP, washed as described above for human brain membranes, and resuspended immediately for assay. However, aliquots of the membrane suspension were set aside, and, after a 3 hour incubation at O°C, they were centrifuged a t 9,500g for 10 minutes. Aliquots (800 pl) of the supernatant were later assayed for inhibitory activity against [3H]TCP binding, as described in section 2.2.
The protease inhibitor cocktail was composed of25 g/ ml leupeptin, 25 pg/ml chymostatin, 0.1 mM EDTA, and 0.1 m M ethyleneglycol-bis-(beta-aminoethylether) N,N,N',N'-tetraacetic acid (EGTA). Incubation time was for 3 hours at 0°C (steady state) in a final volume of 1 ml. In the "supernatant experiments," 800 pl of the supernatant was used, and membranes were added in 100 p1 so as to deliver the usual protein concentration. In this way, the inhibitory activity present in the supernatant of a typical 800 pl aliquot of membrane suspension was measured. Triplicate samples were filtered with an MR24 Brandel Cell Harvester and washed with two 5 ml aliquots of ice-cold buffer. Whatman G F B filters were presoaked in buffer containing 2% polyethylenimine. The tritium retained on the filters were determined using a Taurus scintillation counter with 44% efficiency. Nonspecific binding was determined using 10 pM of TCP. Protein was determined using the method of Lowry et al. (1951).
Dissociation experiments were performed with minor modifications of published procedures (Rothman et a]., 1989). Briefly, membranes pretreated with BTCP were incubated with 5 nM f3H1TCP under two conditions: 1) in the presence of 500 nM (+ )-MK801 to block binding to PCP site 1, and 2) in the presence of 10 FM TCP to define nonspecific binding. After a 3 hour incubation at 0"C, a time 0 was filtered, and TCP was added to condition 1 to give a final concentration of 10 pM. Aliquots of both conditions were then filtered 15,30,60, 120, and 180 min later. Since our previous work demonstrated that the dissociation of i3H1TCP from PCP site 2 is well described by a single exponential component, the data from the experiments conducted here were fit to the following equation using MLAB: BJ3, = 100e-kt, where Bo is the binding at time 0, B, is the binding at a specified time, and k is the dissociation off rate.
Experimental design Experiments were designed according to the method of binding surface analysis, described in detail elsewhere (Rothman, 1986;Rothman, et al., 1988a). Briefly, binding surface analysis is a method of experimental design and data analysis which permits accurate determination of binding parameters. Variations of this method are in use by other laboratories which perform quantitative ligand binding studies (McGonigle et al., 1986;Rovati et al., 1990). In the first set of experiments, two concentrations of [3H]TCP (1 and 5 nM) were each displaced by eight concentrations of TCP and by test drugs thought to be selective for PCP site 1 {(+)-MK801} or site 2 (benztropine). These data were fit to a two site model, in order to determine the approximate "blocking" concentrations of (+)-MK801 and benztropine. The "blocking" concentrations were estimated to be 1 pM and 80 pM, respectively. In the second set of "cross competition" experiments, two concentrations of [3HlTCP (1 and 5nM) were each displaced by TCP, (t)-MK801, and benztropine in the presence of these "blocking concentrations'' of (+)-MK801 (1 pM) and benztropine (80 pM). Each data point was the mean of two separate experiments, each measured in triplicate. The combined data, termed the primary data set, were fit to the one or two site binding model, using MLAB (Knott and Reece, 1972), which uses a weighted nonlinear least squares curve fitting algorithm. For the structure activity studies, a single concentration of c3H1TCP (5 nM) was displaced by concentrations of test drugs in the absence and presence of the blocking concentrations Statistics Each experiment was repeated on the same day or on different days with triplicate observations. Data were analyzed using one-way analysis of variance with the post-hoc Scheffe F-test, with significance set at P < .05.
The F-test (Munson and Rodbard, 1980) was used to determine the best fitting binding model. Materials T3H1TCP (48.9 CUmrnol) was purchased from New England Nuclear (Boston, M). GBR12909 and (+I-MK801 were synthesized as described elsewhere (Monn et al., 1990;Van der Zee et al., 1980). The synthesis of BTCP will be described elsewhere. Bupropion, xylamine, benztropine, and ketamine were purchased from Research Biochemicals Inc. (-)-Cocaine was supplied by Dr. A. Pert (NIMH). Paroxetine was kindly provided by Dr. Reith  Frozen brains of rat, guinea pig, mouse, chicken, rabbit, pig, and sheep were purchased from Pel Freeze Laboratories, (Rogers, AR). Post mortem human cortex was obtained from Dr. W. Tourtellote of the National Neurological Research Bank (Los Angeles, CAI. In each case, tissues were frozen at -80°C until used for the preparation of membranes. of (+)-MK801. Figure 1 reports the TCP, (+)-MK801, and benztropine binding surfaces. These experiments were carried out as described in "Experimental design" above. In the first set of experiments, two concentrations of [3H]TCP were displaced by TCP, (+)-MK801, and benztropine. The latter two agents were thought to be selective for site 1 and site 2, respectively. Preliminary analysis of these data yielded initial estimates of the blocking concentrations for (+)-MK801(1 FM) and benztropine (80 pM), which were used in the subsequent "cross competition studies." The data of both experiments were combined, and are shown in Figure 1. The displacement of two concentrations of f3H1TCP by TCP, in the absence or presence of 1 pM (+)-MK801 and 80 pM benztropine (panel A) provides saturation binding information. Panel B reports the benztropine binding surface in the absence or presence of 1 pM (+)-MK801, which will block [3HlTCP binding to site 1. Panel C reports the (+)-MK801 binding surface in the absence or presence of 80 pM benztropine, which will partially block binding to site 2. points) to a one site binding model resulted in a poor fit whereas fitting to a two site model resulted in a highly significant reduction in the sum-of-squares (SS). Consideration of a third site did not significantly improve the fit. The densities of sites 1 and 2 were 492 and 2,038 fmol/mg protein, respectively (Table I) The results are reported in Table TI. (t)-MK8Ol, tiletamine, and ketamine bound very selectively to site 1. TCP, PCP, and dexoxadrol were moderately selective  ]Binding surfaces were generated by displacing 5 nM I3HVCP by eight concentrations of the test drug in the absence of a blacker, in the presence of 80 WM benzt.ropine or 1 rM (+)-MK801. The data were fit to the two site binding model for thebest-fitvaluesof theKi,withtheKdvaluesofTCP, theKivaluesof(+)-MKSOL, the Ki values of benztropine, and the Bmax values fixed to the values reported in Table I The best-fit parameter estimates (k SEM) of various drugs for the two PCP binding sites are reported above. The r2 for the two site fits ranged from 0.89 to 0.99. for site 1. The serotonin reuptake inhibitor fluoxetine was almost 100-fold selective for site 2. The dopamine reuptake inhibitors GBR12909, benztropine, and BTCP were also relatively selective for site 2. Bupropion and (-)-cocaine had very low affinity for site 2 and were slightly selective for site 1.

Wash-resistant inhibition experiments
Human cortical membranes were incubated for 60 minutes at 0°C with varying concentrations of fluoxetine, GBR12909, and BTCP, and then washed by centrifugation as described in "Materials and Methods." [3HlTCP binding was measured in the absence and presence of 1 JLM (+)-MK801, permitting calculation of specific binding to site 1 and site 2. As reported in Figure  2, these agents produced a dose-dependent wash-resistant inhibition of E3HlTCP binding to site 2. There was no significant effect on site 1 (data not shown).
A trivial explanation for wash-resistant inhibition is the presence of residual drug which is not removed by the washing procedures. To examine this possibility, guinea pig membranes were pretreated with 0,500, and 5,000 nM BTCP. As described in "Materials and Meth-~d~, " t h e membranes were assayed for [3H]TCP binding to site 1 and site 2. Aliquots of the membranes were retained, and after a 3 hour incubation these were centrifuged, and the ability of an 800 pl of the supernatant (the amount present in a 800 pl aliquot of the membrane suspension) to inhibit L3H1TCP binding was measured. The results demonstrated that 5,000 nM, but not 500 nh4 BTCP, produced a wash-resistant inhibition of PCP site 2, and that there was no significant inhibitory activity in the corresponding supernatants (Fig. 3). BTCP had no effect on [3HlTCP binding to PCP site 1.
In related experiments, the dissociation of I3HITCP from PCP site 2 was measured in guinea pig membranes pretreated with 0 or 5,000 nM BTCP. These results demonstrated that preincubation with 5,000 nM BTCP increased the dissociation rate from 0.002 to 0.004 min-' ( Table 111).
Occurrence of PCP site 1 and PCP site 2 in membrane preparations of various species Membranes were prepared from selected brain regions ofvarious species, and the occurrence of PCP site 1 and PCP site 2 determined using 0.5 KM (+ )-MK801 to block the binding of [3HITCP to site 1. Site 2 was considered undetected if the binding in the presence of 0.5 pM (+)-MK801 was not significantly different from the nonspecific binding. We chose to decrease the blocking concentration of (+)-MK801 from 1 pM to 0.5 pM, in light of the species variation of the Ki of (+)-MK801 for the PCP site, For example, in human brain the Ki is 11,367nM (Table I), but in guinea pig brain it is 1,331 nM (Table V). Thus a concentration of 1 p, M would not significantly inhibit [3H]TCP binding to site 2 using human brain membranes, but would if guinea pig mem- t3H]TCP to site 1 and site 2 was measured. There was no significant effect on site 1 (data not shown). Each point is the mean ? SEM (n = 6). *P < .05 when compared to control.  'Guinea pig brain membranes were prepared as described in "Materials and Methods" and were pretreated with 0 or 5,000 nM BTCP. Following a 3 hour incubation with 5 nM [3H]TCPin the presence of 500 nM (+)-MK801 (to block binding to site l), 10 pM TCP was added to block the association of 13H]TCP with PCP site 2, and aliquots were subsequently filtered at various time points. The dissociation rate constant, K -,, was determined by fitting the dissociation data to the equation describing a single component dissociation model (see "Materials and Methods"). Data represent mean i SEM for three experiments each with triplicate observations. *P < .05 when compared to control. branes were used. In view of the generally high affinity of (+)-MK801 for PCP site 1, and our ignorance of its Ki value for PCP site 2 of the different species which were examined, we felt that 500 nM (+)-MK801 was a concentration which would block binding to site 1, without significantly inhibiting 13H]TCP binding to site 2. As reported in Table IV, rabbit, sheep, chicken, guinea pig, mouse, and pig brains had varying levels of PCP site 2. It is important to note that the values reported in Table IV are not Bmax values, but are the specific binding observed with 2 nM [3H]TCP. Rat brain Pig cortex 185 f 17 9 f 0.9 8 i 0.5 176 0 had no detectable levels of PCP site 2. TCP and (+)-MK801 binding surfaces in guinea pig hippocampus were well described by a two site model (Table V), with best-fit parameter estimates similar to those observed using whole guinea pig brain (Rothman et al., 1989). Various preparations of rat brain were examined in greater detail for the occurrence of' PCP site 2. Binding surfaces were generated by displacing 1 nM and 5 nM [3H]TCP each by eight concentrations of TCP, (+)-MK801, and (-)-cyclazocine (data not shown). Using membranes prepared from whole brain and caudate, the data were well described by .one site binding models. Fitting to a two site model did not improve the goodness of fit. The best-fit parameter estimates of the one site binding models are reported in Table VI.
The effect of NaCl on PCP site 1 and site 2 in guinea pig brain Sodium ion is well known to inhibit L3H]TCP binding (Vignon et al., 1982). To examine the effect of NaCl on PCP site 2, the binding of [3HlTCP in the absence and presence of 500 nM (+)-MK801 was determined in the presence of varying concentrations of NaCl (1 to 128 mM). The results, shown in

DISCUSSION
Phencyclidine and phencyclidine-like compounds produce various dose-dependent behavioral effects which are likely mediated through different molecular sites of actions. Single-and multiple-affinity PCP binding sites have been reported in the CNS. For example, Jarvis et al. (19871, Vincent et al. (1979), Zukin and Zukin (19791, and Quirion et al. (1981) reported the    presence of a single high-affinity binding site for 13H]TCP, which is now known to exist as a component of the NMDNionophore complex . Multiple PCP binding sites have also been reported in the rat (Haring et al., 1987;Itzhak, 1988;Mendelsohn et al., 1984;Vignon et al., 1986Vignon et al., ,1989 as well as human brain tissue (Vignon et al., 1989). Moreover, recent work from our laboratory demonstrated a second high-affinity PCP binding site (site 2) in guinea pig brain which appeared to be associated with the dopamine reuptake carrier using striatal membranes (Rothman et al., 1989). Unfortunately, there is currently insufficient data to permit an unambiguous identification of PCP site 2 with the other low-affinity PCP binding sites reported by other investigators (see the Discussion in Rothman et al., 1989). Among the structurally diverse group of agents which act as noncompetitive NMDA antagonists, only PCP and its congeners of the arylcycloalkylamine group potently inhibit the reuptake of biogenic amines (Johnson and Snell, 1985;Vignon and Lazdunski, 1984). Indeed, a recent study of  demonstrated a good correlation between inhibition of [3H]PCP binding and inhibition of [3HlDA reuptake only for arylcyclohexylamines bearing an unmodified phenyl group. Nevertheless, several lines of evidence support the hypothesis that the noncompetitive NMDA antagonist effect of PCP is sufficient to explain its pharmacological effects: 1) the selective noncompetitive NMDA antagonist, (+)-MK801, produces the same behavioral effects as PCP in laboratory animals (Koek and Woods, 1988b), yet it is a very weak reuptake blocker (Snell et al., 1988). 2) The anatomical distribution of [3H]TCP binding sites in rat brain is almost identical with that of the NMDA receptor (Jarvis et al., 1987). 3) The general inability, as noted above, to detect [3H]TCP binding sites in rat brain which are not associated with the NMDA receptor.
Nevertheless, other data indicate that more complex interpretations may be necessary to explain the behavioral and pharmacological effects of PCP. For example, animals trained to discriminate PCP from saline do not completely generalize to competitive NMDA antagonists (Jackson and Sanger, 1988;Willets and Balster, 1988). Rhesus monkeys with a recent history of exposure to cocaine will not learn to self-administer MK801 (Beardsley et al., 1990), though these rhesus monkeys readily learn to self-administer PCP. Unlike PCP, MK801 does not inhibit DA reuptake (Snell et al., 1988). Moreover, effects of PCP which can be ascribed to reuptake blockade such as hypothermia (Pechnick et al., 1989b1, decrease in prolactin levels (Pechnick et al., 1989c,d), and elevation of extracellular levels of DA (Carboni et al., 1989) as measured by in vivo microdialysis occur in the same dose range as effects of PCP which can be ascribed to antagonism of the NMDA receptor, such as its anticonvulsant (Leander et al., 1988) and neuroprotective (Sauer et al., 1988) effects. The recent work by Rao et al. (1990) also support the notion that pharmacologically relevant effects of PCP occur via activation of non-NMDA-associated PCP receptors.
These considerations prompted us to examine the hypothesis that at least two high-affinity PCP binding sites should be detectable in ligand binding studies: the first associated with the NMDA receptors, and the second associated with the biogenic amine reuptake carrier. Therefore, since prior work from our laboratory indicated the presence of PCP site 2 binding in guinea pig brain but not in rat brain (Rothman et al., 1988b), in the present study we sought to determine the occurrence of PCP site 2 in the brains of various species, with an emphasis on the human brain.
The data under the present experimental conditions clearly indicate that most species, with the exception of the rat, have detectable levels of PCP site 2 (Table IV). The inability to detect PCP site 2 in rat brain prompted the authors to use membrane preparations which would be enriched with reuptake binding sites, such as lysed-P2 membranes (prepared from nerve terminals), and caudate membranes, which are enriched with the DA reuptake complex. However, even with these membrane preparations PCP site 2 was not detectable. Although PCP clearly inhibits the reuptake of biogenic amines in rat brain, the reason why this site cannot be labeled in vitro with [3H]TCP remains enigmatic. It is possible that the absence of sodium in this assay condition may contribute to the weak presence of PCP site 2, which is thought to be associated with the reuptake system (a process that requires the presence of sodium ion). However, it is necessary to exclude NaCl from the assay, since, as reported by other investigators (Vignon et al., 1982(Vignon et al., ,1983, [3H]TCP binding to both site 1 and site 2 is inhibited by sodium ion (Table VII). Nevertheless, our inability to detect PCP site 2 in rat brain does not mean that it does not exist.
Quantitative examination of 13H]TCP binding to human brain membranes yielded qualitatively the same data as previously obtained with guinea pig brain membranes (Rothman et al., 1989): a two site model with (+)-MK801 being highly selective for site 1. There were interesting quantitative differences. The Kd values of TCP for site 2 were higher in human brain (197 nM) than for guinea pig brain (46 nM), and the Ki value of (+)-MK801 for site 1 was lower in guinea pig brain (3.2 nM) than in human brain (8.9 nM). In guinea pig whole brain membranes, PCP sites 1 and 2 are present at about equal densities (Rothman et al., 1989); in human brain cortex membranes, site 2 is the more abundant binding site.
As reported for guinea pig brain (Rothman et al., 1989), several lines of evidence support the hypothesis that PCP site 2 in human brain is associated with biogenic amine reuptake carriers. First, high-affinity serotonergic (fluoxetine) and dopaminergic (GBR12909 and BTCP) reuptake inhibitors bind relatively selectively to site 2. The reason for their low-affinity interaction with site 2 probably relates to the absence of NaCl in the assay medium, since NaCl is required to detect high-affinity binding sites for conventional reuptake blockers. Second, PCP, which inhibits the reuptake of DA and serotonin (Smith et al., 1977), has high affinity for site 2, whereas (+ )-MK801, which is a very weak DA reuptake blocker (Snell et al., 1988>, binds selectively to site 1. Third, fluoxetine, GBR12909 and BTCP produce wash-resistant inhibition of [3H]TCP binding to site 2 (Fig. 2).
As described in recent publications (Reid et al., 1990;Xuet al., 19911, there are several mechanisms, shown in Table VIII, which can explain the phenomenon of washresistant inhibition. The trivial explanation is that residual drug not removed by the wash procedure acts as a competitive inhibitor of binding. Since a membrane suspension is composed of an aqueous phase (the buffer) and a lipid phase (the membranes), the residual drug may be present in either the aqueous phase (mechanism 1) or the lipid phase (mechanism 21, or both. Alternatively, the wash-resistant inhibitor might be tightly bound to the binding site of interest over the time course of the experiment. In the simplest of the tight-binder scenarios, occupation of the binding site by the washresistant inhibitor precludes labeling of the site by the [3Hlligand (mechanism 3). This would produce a decrease in the Bmax of the target binding site. In a more complex scenario, the wash-resistant inhibitor binds to a different binding site, or to a domain of the binding site, producing a conformational change in the site labeled by the 13H]ligand (mechanism 4). This would typically produce an increase in the Kd, without altering the Bmax. Since the increase in the Kd reflects a conformational change in the binding site, an expected effect would be an increase in the dissociation rate of the 13H]ligand for that site.
As described in Table VIII, it is possible to distinguish these various models by measuring the effect of the wash-resistant inhibitor on 1) the residual drug in the aqueous phase, 2) the Bmax and Kd of the target binding site, and 3) the dissociation rate of the [3H]ligand. To measure residual drug in the aqueous phase, membranes are centrifuged, and the supernatant is assayed for inhibitory activity. Saturation binding studies permit determination of the Bmax and Kd, while dissociation studies permit measurement of the dissociation rate, kp1.
Since C3H]TCP binding to PCP site 2 is more readily measured with guinea pig membranes, the authors chose to use this tissue in preference to human brain membranes. Moreover, since previous work demonstrated that 5,OOOnM BTCP produced about a 50% decrease in [3HlTCP binding to guinea pig brain membranes (Rothman et al., 1989), which is similar to the data observed here with human brain membranes, this agent was chosen for further study. After pretreatment with 5,000 nM BTCP, the level of 13H]TCP binding to PCP site 2 was diminished by about 50% (Fig. 3), yet there was not detectable residual drug (Fig. 3). Due to the poor signal, it was not possible to conduct saturation binding studies. The dissociation experiments demonstrated that pretreatment with BTCP increased the dissociation rate of [3HlTCP at PCP site 2 (Table 111). This observation is incompatible with the residual drug hypothesis, since residual drug would increase the apparent Kd by acting as a competitive inhibitor, but not alter the intrinsic dissociation rate.
These data collectively suggest that the reuptake inhibitors bind to a site from which they slowly dissociate (the serotonin and dopamine reuptake carriers) and as a result of this, the binding of [3H]TCP to site 2 decreases. The observation that neither GBR12909 nor fluoxetine produced a complete wash-resistant inhibition of I3HITCP binding to site 2 suggests that in human cortex PCP site 2 may be a mixture of both serotonergic and dopaminergic reuptake carriers.
In a recent study, Vignon et al. (1988) reported labeling the dopamine reuptake site in the rat striatum with the PCP analog L3H1BTCP. Since the binding of L3H1BTCP and other ligands such as [3H]cocaine (Reith et al., 1985) and l3H1GBR12935 (Janowsky et al., 1986) to the DA transporter requires the presence of Na+, our inability to detect PCP site 2 in rat caudate with [3HlTCP might potentially reflect the absence of Na+ in the e3H1TCP assay. However, the observation that Na+ inhibits the binding of [3HlTCP to site 2 (Table VII) argues against this point of view. There are several alternative explanations. Assuming that [3HlTCP and the standard reuptake inhibitors label the same site on the transporter, then it is possible that L3H]TCP binds to the site with high affinity in the absence of Na+, and with low affinity in the presence of Na'. Similar reasoning would also apply if I3H3TCP and the reuptake inhibitors label different sites on the transporter. It is misleading to expect i3H1TCP to exhibit the same binding properties as i3H1BTCP. Although BTCP is chemically related to PCP, its neurochemicaI (Vignon et aI., 1988) and behavioral actions (Koek et al., 1989) indicate that it is best described as a cocaine-like drug, which like [3H]cocaine, binds to the DA transporter in a Na+dependent manner.
In summary, the major findings of this study are 1) that 13H]TCP labels two binding sites in human cortex, and that PCP site 2 is the predominant binding site; 2) PCP site 2 is associated with biogenic amine reuptake carriers and is optimally labeled by 13H]TCP in the absence of NaCl; and 3) guinea pig brain, but not rat brain, is an appropriate model system for studying PCP site 2.