D4 Dopamine Receptor-Mediated Phospholipid Methylation and Its

ORIGINAL RESEARCH ARTICLE D4 dopamine receptor-mediated phospholipid methylation and its implications for mental illnesses such as schizophrenia A Sharma1*, ML Kramer1,2,*, PF Wick1, D Liu1, S Chari1, S Shim1, W Tan1, D Ouellette1, M Nagata1, CJ DuRand3, M Kotb4 and RC Deth1

1Department of Pharmaceutical Sciences, Northeastern University, Boston, Massachusetts 02115; 3ENRM VA Medical Center, Bedford, Massachusetts 01730; 4University of Tennessee School of Medicine, Memphis, Tennessee 38163, USA

Previous studies have shown D2-like dopamine receptor involvement in the regulation of phospholipid methylation (PLM), while others have documented impaired methionine and folate metabolism in schizophrenia. Utilizing [14C]formate labeling in cultured neuroblastoma cell lines, we now show that D4 dopamine receptors (D4R) mediate the stimulatory effect of dopamine (DA) on PLM. The effect of DA was potently blocked by highly D4R-selective antagonists and stimulated by the D4R-selective agonist CP-226269. DA-stimulated PLM was dependent upon the activity of methionine cycle enzymes, but DA failed to increase PLM in [3H]methionine labeling studies, indicating that a methionine residue in the D4R might be involved in mediating PLM. A direct role for MET313, located on transmembrane helix No. 6 immediately adjacent to phospholipid headgroups, was further suggested from adenosylation, site- directed mutagenesis and GTP-binding results. A comparison of PLM in lymphocytes from schizophrenia patients vs control samples showed a four-fold lower activity in the schizophrenia group. These ﬁndings reveal a novel mechanism by which the D4R can regulate membrane composition. Abnormalities in D4R-mediated PLM may be important in psychiatric illnesses such as schizophrenia. Keywords: receptors; G protein-coupled; phospholipid methylation; methionine; folic acid; membrane ﬂuidity; purines; de novo synthesis

Schizophrenia is a complex psychiatric disorder affect- dopamine receptor (D4R) has received particular atten- ing approximately 1% of the population worldwide tion, since clozapine, a neuroleptic drug with remark- with a typical onset between the late teens and mid- able clinical efficacy, binds preferentially to the D4R thirties, often in the absence of significant prodromal subtype.6 In addition, several studies have reported psychiatric symptoms.1 Decades of research have that D4R density is increased in post-mortem brain yielded many theories on the origin of schizophrenia, samples from persons with schizophrenia,7,8 although although none provides a satisfying mechanistic expla- other studies failed to replicate this finding.9 While the nation. Logically, dysfunctional neurotransmission is clinical trial of a highly D4R-selective ligand (L- the most widely held theory, and specific abnormali- 745870) did not reveal any benefit,10 recent findings ties involving dopaminergic,2 glutamatergic,3 GABA- suggest this compound may in fact be an agonist, rather ergic4 and nicotinic cholinergic5 signaling have been than an antagonist.11 proposed. Of these, the ‘Dopamine Hypothesis’ is Structural features of the D4R suggest that it could dominant, based largely on the recognized therapeutic play a unique role in neuropsychiatric disorders. Thus efficacy of drugs which block D2-like dopamine recep- in humans and primates the D4R has a polymorphic tors (ie D2, D3 and D4 receptors), although their spe- 16 amino acid repeat in its third intracellular loop cific mechanism of benefit and the importance of indi- (Figure 1) and humans can exhibit anywhere from two vidual receptor subtypes remains unclear. The D4 to ten repeats,12,13 making the D4R one of the most hypervariable proteins in man. Several studies have found a correlation between the presence of a higher number Correspondence: Dr RC Deth, Department of Pharmaceutical of repeats (eg seven) with personality traits such as Sciences, 312 Mugar Hall, Northeastern University, 360 Hunt- novelty seeking,14,15 alcoholism and drug-taking ington Avenue, Boston, MA 02115, USA. E-mail: behavior in adults,16 as well as with attention deficit r.dethȰnunet.neu.edu hyperactivity disorder (ADHD) in children,17 although * The first two authors contributed equally to this work 18 2 Present address: ArQule Inc, Waltham, MA 02154, USA others have not found such correlations. Received 6 July 1998; revised 10 September and 2 December There is an extensive literature documenting 1998; accepted 11 December 1998 impaired methionine metabolism in schizophrenia. D4 dopamine receptor-mediated PLM A Sharma et al 236 medications at the VA Medical Center while control subjects were recruited from both Northeastern Univer- sity and the VA Medical Center. An earlier study in lymphocytes found that such medications did not appear to influence methionine metabolism.23 All patients had firm diagnoses of schizophrenia or schi- zoaffective disorder (DSM-IV criteria), based upon multiple diagnostic clinical interviews by several psy- chiatrists, long-term clinical observation of their course of illness as well as a semi-structured interview and record review at the time blood samples were obtained. As previously described by LeFur et al,29 lymphocytes (4 × 106) were resuspended in Hank’s balanced salt solution containing 8 ␮Ci ml−1 [3H]-methyl-methionine for 60 min. DA (10 ␮M) and GTP (1 mM) with or without haloperidol (1 ␮M) were added to additional groups. After 60 min, 1.5 ml of ice-cold 10% trichlo-

Figure 1 Schematic diagram of the D4 dopamine receptor roacetic acid (TCA) was added and cells were har- indicating the location of the agonist binding site, methionine vested by centrifugation. The pellet was washed with No. 313, the proline-rich putative SH3 recognition domain 2.5 ml of TCA followed by extraction with 15 ml of a

and the 16 amino acid repeat segment which exhibits hyper- 6:3:1 mixture of CHCl3/CH3OH/HCl. The CHCl3 phase variability in man and other primates. was washed twice with 4 ml of 0.1 N KCl in 50%

CH3OH and an aliquot of the phospholipid-containing Studies by Kety and his colleagues found that methion- CHCl3 was dried and counted. ine supplementation could precipitate an acute psychotic episode in many schizophrenic patients, while Cell culture conditions it had no effect in normal individuals,19,20 a finding SK-N-MC cells were grown in six-well plates in Ham’s which has been replicated many times (for a review see F-10 medium supplemented with 12% fetal bovine reference 21). Moreover, the rate of whole body meth- serum and 1% PSF. Chinese hamster ovary (CHO) cells ionine metabolism in people with schizophrenia were grown in ␣-modified minimal essential medium (including untreated individuals) is reduced to less (␣-MEM) supplemented with 2.5% donor horse serum than one-third that of controls.22 This deficit has also and 2.5% fetal bovine serum and 1% PSF. CHO cells been demonstrated in vitro with patient lymphocytes.23 transfected with either the wild-type D4R or with the Studies of individual methionine metabolism-related M313A mutant D4R (kindly provided by Dr H Van Tol) enzymes in schizophrenia have found deficits in the were grown in the presence of 400 mg L−1 G418. activity of methionine adenosyltransferase (MAT),24,25 5,10-methylenetetrahydrofolate reductase (5,10- [14C]Formate-based phospholipid methylation methyleneTHFR) 26–28 and serine hydroxymethyl- After a brief wash with Hank’s balanced salt solution, transferase (SHMT).25 cells were incubated for 30 min or the desired time in In 1981, LeFur et al29 reported that DA could aug- 600 ␮l of Hank’s solution containing 1 ␮Ci ml−1 ment [3H]methionine-based PLM in murine lympho- [14C]formate, with or without drugs as indicated. The cytes, but only when GTP was present. They further reaction was terminated by an initial wash with ice- showed that neuroleptic drugs such as haloperidol cold unlabeled Hank’s solution followed by the blocked this effect, indicating involvement of a D2-like addition of 1 ml of ice-cold 10% TCA. After scraping receptor. Subsequently, Leprohon et al30 showed that the cells and transferring them to an Eppendorf tube, dopamine could stimulate PLM in rat brain-derived the sample was sonicated and an aliquot removed for synaptosomes and this effect was also blocked by halo- protein assay. Following centrifugation (15 min × peridol. We therefore undertook a more detailed study 10 000 rpm), the pellet was dispersed in 1.5 ml of 2

of dopamine receptor involvement in PLM. N HCl/MeOH/CHCl3 (1:3:6), vortexed and allowed to separate. The lower CHCl3 layer was washed twice with 400 ␮l of 0.1 N KCl in 50% MeOH and an aliquot Materials and methods counted for radioactivity after evaporation. Results are [3H]Methionine-based phospholipid methylation in expressed as the mean ± SEM of six replicates. human lymphocytes This clinical research project was carried out with the [35S]GTP␥S binding studies approval of institutional review boards at both North- Stably transfected CHO cells were grown in ␣-MEM eastern University and the Edith Nourse Rogers Mem- without nucleosides, supplemented with 2.5% fetal orial Veterans Administration Medical Center bovine serum, 2.5% horse serum, 1% PSF and 400 ␮g (Bedford, MA, USA), and informed consent was ml−1 G418. Cells were collected in phosphate-buffered

obtained from all participants. Patients were under saline (137 mM NaCl, 2.6 mM KCl, 10 mM Na2HPO4, ° × treatment with standard or atypical antipsychotic 1.8 mM KH2PO4) and pelleted at 4 C at 200 g. The D4 dopamine receptor-mediated PLM A Sharma et al 237 pellet was resuspended in lysis buffer (5 mM Tris (pH course over a 2-h period of observation. In the presence 7.5), 5 mM EDTA, 5 mM EGTA, 0.1 mM phenylmethyl- of DA (1 ␮M) an increased rate of PLM was evident sulfonyl fluoride) and homogenized with a Dounce starting at 30 min (Figure 3a). As shown in Figure 3b, homogenizer. The cell lysate was centrifuged at 34 000 DA-stimulated PLM exhibited a sigmoidal log dose- × g for 15 min and the pellet resuspended in membrane response relationship, reaching a maximum of almost ␮ buffer (50 mM Tris (pH 7.5), 0.6 mM EDTA, 5 mM three times the basal level with an EC50 of 2.3 M. MgCl2, 0.1 mM phenylmethylsulfonyl fluoride). The potency of various antagonists for inhibition of [35S]GTP␥S binding studies were conducted in a final DA-stimulated PLM was examined in order to deter- volume of 100 ␮l containing 5–10 ␮g membrane pro- mine which receptor subtype mediated this effect. SK- tein, 2–3 nM [35S]GTP␥S, 2 ␮M GDP, 1 mM EDTA, N-MC cells were treated with 1 ␮M DA in the absence

5 mM MgCl2, 100 mM NaCl, 1 mM dithiothreitol and or presence of increasing concentrations of two highly 50 mM Tris (pH 7.5). Non-speciﬁc binding was deter- D4R-selective antagonists, U-101958 and L-745870, as mined in the presence of 10 ␮M GTP␥S. well as the moderately D4R-selective antagonist clozapine. As shown in Figure 4, each agent produced a Adenosylation studies dose-dependent reduction of the DA stimulation.

MAT activity was extracted from peripheral blood Using the EC50 value of DA and the IC50 value of each mononuclear cells and assayed as previously antagonist, respective Ki estimates of 1.6, 1.5 and described.31 D4R-transfected CHO cells were harvested 66 nM were calculated, in good agreement with their and pelleted in ice-cold modiﬁed Kreb’s buffer and expected potency at D4Rs.33–35 It should be noted that washed with MAT assay buffer containing 50 mM KCl, in this and other studies, U-101958 displayed partial

15 mM MgCl2, 0.3 mM Na2EDTA, 4 mM dithiothreitol, agonist activity (relative efﬁcacy of 0.26 compared to 50 mM HEPES (pH 7.7 with KOH). Cells were dis- dopamine) accounting for its partial inhibition of DA- rupted by three rapid freeze thaw cycles and aliquots stimulated PLM. In contrast to these D4R antagonists, containing 100–300 ␮g protein were used for each the presence of either the D2/D3-selective antagonist assay tube. Adenosylation was carried out for 60 min raclopride or the D1-receptor antagonist SCH23390 had at 37°C in MAT assay buffer (total vol. = 100 ␮l) sup- no effect on DA-stimulated PLM at 1 ␮M (Figure 4). plemented with 0.1 mg ml−1 BSA and [3H]ATP. Reac- Thus the ability of DA to stimulate PLM appears to be tion was terminated by the addition of 200 ␮lof1M mediated exclusively by the D4R. NaCl followed by centrifugation. The pellet was resus- Comparisons of agonist potency also indicated D4R pended in 500 ␮l of buffer containing 0.2% SDS and mediation of DA-stimulated PLM, as shown in Figure sonicated to solubilize membrane proteins. After pre- 5a. Thus the highly D4R-selective agonist CP-226269

clearing the sample with non-immune rabbit serum stimulated PLM with an EC50 of 2.9 nM, comparable and Pansorben treatment followed by centrifugation, to its reported EC50 of 5.8 nM for inhibition of cAMP anti-D4R receptor antiserum was added (1:1000) for formation via D4Rs.36 Notably, CP-226269 was a partial 30 min at 25°C followed by Pansorben treatment and agonist with a relative efﬁcacy of 0.63 as compared to centrifugation. After a wash the pellet was resus- DA. The non-selective D2-like receptor agonist quinpi- pended and counted. role was also a partial agonist (relative efﬁcacy = 0.63; = EC50 21 nM), while the D1-selective agonist SKF- 38393 stimulated PLM only at supra-micromolar con- Results centrations. As noted earlier, U-101958 exhibited weak

D4 receptor involvement in phospholipid methylation partial agonism with an EC50 of 0.85 nM, similar to its 3 Previous studies of PLM have utilized either [ H] Ki for inhibition of DA response. methyl-methionine or [3H]methyl-S-adenosylmethion- Lanau et al37 reported that epinephrine and norepi- ine as a source of labeled methyl groups. In order to nephrine are only slightly less potent agonists than include the possibility that folate-dependent processes dopamine at the D4R (6-fold and 13-fold respectively), might be important in PLM, we employed a novel whereas at D2 and D3 receptors they are about 100-fold [14C]formate-based labeling approach, since formate is less potent. As shown in Figure 5b, epinephrine and readily converted to 10-formylTHF. As illustrated in norepinephrine are full agonists at stimulating PLM. 14 Figure 2, [ C] label then distributes to other single-car- Epinephrine was almost equipotent as dopamine (EC50 bon folates including 5-methylTHF, a required co-fac- values of 2.3 and 2.4 ␮M respectively), whereas norepi- tor for the remethylation of homocysteine to methion- nephrine was 4.6-fold less potent than DA and their ine. S-adenosylation of methionine by the action of effect was blocked by L-745870. Serotonin also methionine adenosyltransferase (MAT), facilitates appeared to be a full agonist, but was 43-fold less transfer of the labeled methyl group to phosphatidyl- potent than DA. Again, these potencies are consistent ethanolamine (PE) to form N-methylPE (Figure 2; with D4R mediation of DA-stimulated PLM. lower circle). SK-N-MC neuroblastoma cells, which express D1 Role of methionine cycle enzymes in DA-stimulated and D4 dopamine receptors,6,32 were used as a test sys- PLM tem for the study of PLM. After a delay attributable to We pretreated SK-N-MC cells with various inhibitors the time required for movement of [14C] label through to probe the involvement of methionine cycle enzymes the folate pathway, basal PLM displayed a linear time- in DA-stimulated PLM. Involvement of MAT was con- D4 dopamine receptor-mediated PLM A Sharma et al 238

Figure 2 Proposed pathways for D4-receptor-mediated and non-receptor-mediated phospholipid methylation. The N-methylation of phosphatidylethanolamine (PE) can result from methyl donation by either S-adenosylmethionine (SAM) or by the S-

adenosylated D4-receptor D*4 SAM). Both receptor (upper) and non-receptor (lower) pathways are represented as cyclic processes (although available methionine can sustain the lower cycle) which are dependent upon 5-methyltetrahydrofolate (5-methylTHF) for remethylation. The pathway by which [14C]formate can enter the methylation cycle is illustrated in bold print. Receptor and non-receptor pathways are dependent upon the same four enzymes: methionine adenosyltransferase (MAT), phospholipid methyltransferase (PLMT), S-adenosylhomocysteine hydrolase (AdoHCYase) and methionine synthase (Met Syn). Non-methion-

ine forms of the D4-receptor are designated with an asterisk to indicate their enhanced potential for contributing to spontaneous G-protein activation. As indicated, antagonist occupation of the methionine form of the D4-receptor would reduce receptor participation in phospholipid methylation, while agonist binding would increase its participation. Enhanced N-methylation of PE could decrease phospholipid packing density and consequently alter the conformational dynamics of other membrane

proteins, especially those in close proximity to the D4-receptor. Other abbreviations: phosphatidylserine (PS), serine hydroxyme- thyltransferase (SHMT).

firmed by the inhibitory effect of the selective inhibitor (26%) and DA-stimulated (40%) PLM (Figure 6b), indi- cycloleucine,38 which reduced DA-stimulated PLM by cating that methionine synthase plays an important 90% (Figure 6a). As expected, basal PLM was also role in both components. TLC separation of [14C]-lab- reduced, but to a lesser extent. Similarly, involvement eled phospholipids showed that D4R-mediated PLM of adenosylhomocysteine hydrolase (AdoHCYase), was primarily reflected N-methylated derivatives of PE evident by a significant (P Ͻ 0.01) reduction in DA- (data not shown), as is the case for SAM-mediated stimulated methylation caused by the inhibitor 3- PLM.41 deaza-adenosine,39 although the level of inhibition Clearly, DA can stimulate PLM in intact SK-N-MC (36%) was less than that produced by cycloleucine. cells when [14C] formate labeling is employed, how- This result is consistent with a direct, obligatory role ever, earlier studies in murine lymphocytes using [3H]- for MAT in the PLM pathway, while AdoHCYase is methyl-methionine labeling found that when DA was involved only in the recycling of homocysteine. Meth- applied alone it had no stimulatory effect.29 As shown ionine synthase utilizes the co-factor 5-methylTHF to in Figure 6c, even though basal labeling of phospho- convert homocysteine to methionine in a Vit B12- lipids was robust, 1 ␮M DA had no effect on PLM in dependent reaction which can be inhibited by nitrous SK-N-MC cells using [3H]-methyl-methionine labeling, oxide,40 the general anesthetic. Exposure of SK-N-MC nor did L-745870 affect PLM. This indicates that the cells to nitrous oxide during [14C] formate labeling D4R-activated mechanism of PLM does not utilize caused a significant reduction (P Ͻ 0.05) in both basal SAM-derived methyl groups in SK-N-MC cells. Since D4 dopamine receptor-mediated PLM A Sharma et al 239

Figure 4 Inhibition of DA-induced phospholipid methylation by dopaminergic receptor antagonists. SK-N-MC cells were incubated in [14C]formate Hank’s buffer containing 1 ␮M DA for 30 min in the presence or absence of antagonists at the indicated concentrations. Results are normalized to the level of DA-stimulated PLM in the absence of antagonist. Data shown are the mean ± SEM of six replicates.

[14C] formate readily labels single-carbon folates whereas [3H]-methyl-methionine does not, this suggests a novel mechanism of D4R involvement in PLM in which the receptor promotes the transfer of a methyl group from 5-methylTHF to PE.

Methionine 313 and D4R-mediated PLM The above findings raise the intriguing possibility that methionine cycle enzymes such as MAT might act on a methionine residue in the D4R to provide for DA- stimulated PLM. The human D4R contains only six methionine residues:6 the N-terminal residue, three residues located within transmembrane helix No. 3, one located within transmembrane helix No. 5 and one which is located in the cytoplasmic extension of helix No. 6 (MET313 in the human D4.2R; Figure 1). The first five residues are not accessible to cytoplasmic enzymes, so MET313 is the only methionine residue in the entire D4R which conceivably could play a role in Figure 3 DA stimulation of phospholipid methylation. SK- PLM. Moreover, MET313 is located immediately adjac- N-MC cells were incubated in Hank’s balanced salt solution ent to the headgroups of phospholipids in the cytoplas- containing [14C]formate for different time periods in the pres- mic leaflet of the membrane, a highly favorable locus ence or absence of 1 ␮M DA (a) or for 30 min in the presence of different concentrations of DA (b). Phospholipids were for possible involvement in PLM. extracted and their content of radioactivity normalized to pro- The MET313 locus is part of a cytoplasmic extension tein content (CPM ␮g−1). Data shown are the mean ± SEM of of transmembrane helix No. 6 which has been shown six replicates. to undergo a large rotational movement upon receptor activation from the inactive R state to the active R* state in another receptor.42 In a number of other receptors, mutation of the wild-type residue at the MET313 D4 dopamine receptor-mediated PLM A Sharma et al 240

Figure 6 Effects of methionine cycle enzyme inhibitors on DA-stimulated PLM. (a) SK-N-MC cells were pretreated for 1 h with either cycloleucine (2 mM) or 3-deaza-adenosine (10 ␮M) prior to 30 min of [14C]formate labeling in the presence or absence of DA (10 ␮M). (b) Cells were incubated in 14 100% N2O during the 30 min of [ C]formate labeling in the presence or absence of DA (10 ␮M). (c) PLM was measured after 30 min incubation in [3H]-methyl-methionine ± DA (30 ␮M) in the presence or absence of L-745 870 (100 nm). ** or *** indicate a signiﬁcant difference (P Ͻ 0.01 or 0.001) vs non-DA-treated group. Data shown are the mean ± SEM of six replicates.

Figure 5 Stimulation of PLM by dopaminergic agonists. SK- we compared the level of constitutive [35S]GTP␥S bind- N-MC cells were incubated for 30 min in [14C]formate Hank’s ing in membranes from non-transfected CHO cells with buffer containing agonists at the indicated concentrations of membranes from stably-transfected CHO cells express- non-biogenic amine agonists (a) or biogenic amines (b). ing the human D4.4R. As shown in Figure 7a, a large Results are normalized to the maximum level of DA stimu- increase of basal GTP binding was evident in the D4R- lation. Comparison of data shown are the mean ± SEM of containing membranes, indicative of constitutive six replicates. receptor activity. Pretreatment with the MAT inhibitor cycloleucine caused GTP binding to almost completely return to normal levels (Figure 7b), while cycloleucine locus to any other residue has been shown to cause had no effect on GTP binding in non-transfected cells an increase in the level of R* resulting in constitutive (Figure 7c). These results indicate that D4Rs can exhi- receptor activation of G proteins.42,43 Post-translational bit constitutive activity toward G protein activation. modiﬁcation of the sidechain of MET313, such as Since this activity can be eliminated by inhibition of adenosylation by MAT, is equivalent to a mutation and MAT, it strongly suggests that adenosylation of would therefore be expected to likewise result in MET313 is responsible for constitutive activity. constitutive D4R activity. To examine this possibility To more directly investigate whether the D4R might D4 dopamine receptor-mediated PLM A Sharma et al 241

Figure 7 MAT-dependent spontaneous G protein activation in membranes from CHO cells overexpressing the D4-receptor. 35 ␥ (a) Basal [ S]GTP S binding in CHO/D4 cell membranes as compared to binding in membranes from non-transfected Figure 8 Adenosylation of the D4 dopamine receptor and a CHO cells. (b) Inhibition of basal [35S]GTP␥S binding in receptor-derived peptide. (a) Adenosylation of the D4-receptor CHO/D4 membranes from cells pretreated for 110 min with was measured by immunoprecipitation of the receptor after either cycloleucine (0.1 and 1 mM) or vehicle. (c) Basal incubation of disrupted CHO/D4 cells (subconfluent or [35S]GTP␥S binding in non-transfected CHO cell membranes confluent) with increasing concentrations of [3H]-labeled pretreated for 110 min with 1 mM cycloleucine. ATP. (b) Adenosylation of the D4 receptor was significantly Ͻ (P 0.01) decreased by the D4-selective antagonist clozapine and by the MAT inhibitor cycloleucine. (c) A decapeptide be adenosylated by MAT, we carried out a broken cell consisting of residues 305–314 from the human D4 receptor was incubated with methionine adenosyltransferase (MAT) in MAT assay with D4R-expressing CHO cells and meas- 3 ured the incorporation of radiolabel from [3H]-ATP into the presence of increasing concentrations of [ H]ATP. Incor- poration of radiolabel into the peptide (10 ␮M) was compared a D4R immunoprecipitate. As shown in Figure 8a, with incorporation into equimolar levels of methionine. incorporation of label into the receptor-enriched frac- tion was dependent upon the concentration of [3H]ATP and a significantly higher level of incorporation was cine inhibited D4R adenosylation (Figure 8b), to an observed with subconfluent cells vs fully confluent extent roughly similar to its inhibition of constitutive cells, consistent with the reported higher activity of D4R-dependent [35S]GTP␥S binding, confirming the MAT in actively dividing cells.31 Addition of cycloleu- role of MAT in this process and suggesting that adeno- D4 dopamine receptor-mediated PLM A Sharma et al 242 sylation and spontaneous D4R activity are related. Fur- As shown in Figure 10a, basal incorporation of radio- thermore, the addition of clozapine caused a large label into phospholipids during a 1-h incubation was reduction in the incorporation of radiolabel (Figure markedly reduced in patient samples (P Ͻ 0.005), 8b). Since antagonists stabilize receptors in their inac- reaching a level only one-fourth that of controls. For tive (R) conformation,44 this suggests that adenosyl- both groups, the addition of DA (10 ␮M) and GTP 3 ation of the D4-receptor primarily occurs when the (1 mM) caused an increase in [ H]methyl incorporation receptor is in its active (R*) conformation. in both groups, although the percentage increase was In order to test whether MET313 might be competent greater in the patient (100%) vs control (35%) samples, to serve as a site of adenosylation, we prepared a deca- due to lower basal levels in the former. Addition of the peptide corresponding to residues Nos 305–314 of the non-selective D2-like antagonist haloperidol (1 ␮M) in receptor and examined its ability to serve as a substrate the presence of DA and GTP caused a significant for partially purified lymphocyte MAT. As shown in decrease of PLM (P Ͻ 0.05) to a level less than that of Figure 8c, the receptor-derived peptide was adenosyl- untreated cells suggesting that D2-like receptors con- ated to approximately the same extent as free methion- tribute to the basal level of PLM in lymphocytes. While ine when compared on an equimolar basis. preliminary, these results provide suggestive, albeit To determine the importance of MET313 in DA- significant, evidence that PLM is impaired in lympho- stimulated PLM, we compared [14C] formate PLM in cytes from individuals with schizophrenia. CHO cells overexpressing either wild-type D4Rs or a mutant D4R in which MET313 was replaced with an alanine residue. As expected, overexpression of the wild-type receptor resulted in a four-fold increase in DA-stimulated PLM, accompanied by a small decrease in basal PLM (Figure 9). In contrast, overexpression of the M313A mutant D4R failed to augment DA-stimulated PLM and increased basal PLM. This result con- firms the importance of MET313 for DA-stimulated PLM.

Phospholipid methylation in schizophrenia The status of PLM in lymphocytes from controls and individuals with schizophrenia was investigated using [3H]-methyl-methionine labeling methodology as previously described by LeFur et al for murine lymphocytes.29 Patient subjects were recruited from the inpa- tient population at the ENRM Veterans Administration Medical Center (Bedford, MA, USA) and were on standard antipsychotic medication. All were male, and their mean age was 46.4 ± 11.2 yrs (range: 27–70). Con- trol subjects were recruited from Northeastern Univer- sity and the VA Medical Center. Ten were male and two female, and their mean age was 40.2 ± 8.3 yrs (range: 26–52).

Figure 10 Phospholipid methylation in lymphocytes from persons with schizophrenia vs normal controls. Studies were carried out using the [3H]methionine-based method previously described by LeFur et al.29 (a) Individual basal levels Figure 9 Role of MET313 in DA-stimulated PLM. Non-trans- of phospholipid methylation and their mean ± SEM. (b) fected CHO cells (left) and CHO cells transfected with either Influence of DA (10 ␮M) and GTP (1 ␮M) ± haloperidol (1 ␮M) wild-type (middle) or M313A mutant D -receptors (right) 4 on phospholipid methylation in control and schizophrenia were incubated for 30 min in [14C]formate-containing Hank’s samples. Differences between control and patient values are buffer ± DA (30 ␮M). Data shown are the mean ± SEM of significant (P Ͻ 0.005) for all treatment groups. six replicates. D4 dopamine receptor-mediated PLM A Sharma et al 243 Discussion nition domain (Figure 1), could play a role in modulat- While it has been proposed that D4Rs have a unique ing SH3 domain interactions, although this possibility role in novelty-seeking and attention-related remains to be tested. behavior14–17 as well as in the etiology of schizo- Ligand-gated ion channels are highly sensitive to phrenia,6,7 the particular signaling pathway of D4Rs perturbances in the lipid environment. Recent studies which might provide this role has yet to be identified. suggest that GABAA and glycine receptors may be the At the same time, there is an increasing recognition pharmacological target of ethanol and general anaes- 46–47 48 that changes in the membrane environment, intro- thetics, while the activity of NMDA and nicotinic 52 duced either by drugs such as ethanol or general anes- cholinergic receptors are also modulated. Interest- thetics,45–48 can have profound effects on the signaling ingly, inhibitory receptor activity is enhanced by etha- properties of membrane receptor systems, such as nol, while excitatory activity is suppressed. Intermol- ligand-gated ion channels. In the current report we pro- ecular contacts at the protein/phospholipid interface vide the first evidence that D4Rs are intimately may be the source of this modulation, and the energy involved in modulating the membrane environment barrier between inactive and active conformational via promotion of PLM, an activity previously thought states may be affected by changes in membrane compo- to rely only upon SAM as the methyl donor.41 sition or packing density. Clearly D4R-mediated Increased PLM may thus provide a ‘missing link’ to changes in PLM could provide a physiological counter- explain D4R-mediated effects of dopamine on part to these pharmacological actions. behavior. Although it is too early to comprehend the DA modulation of NMDA receptor function via D2- full significance of this finding, these observations like receptors has been reported by a number of investi- afford a novel perspective on the role of D4Rs in both gators. For example, D2-like receptor antagonists normal and pathological settings. caused up to a four-fold increase of NMDA currents in In SK-N-MC cells, from which the D4R was first a rat frontal cortex slice preparation, with clozapine 53 cloned,6 DA-stimulated PLM was potently inhibited by exerting the most potent activity. Conversely, neuro- several highly D4R-selective antagonists but was not toxicity induced by the NMDA antagonist MK-801 is inhibited by the D2/D3-selective antagonist raclopride prevented by neuroleptic agents, with clozapine again 54 or the DIR antagonist SCH23390 (Figure 4). The potent being the most potent. Motor and psychotomimetic effect of the D4R-selective agonist CP-226,269 and the responses to the non-competitive NMDA receptor 55 comparatively high potency of norepinephrine and antagonist ketamine are blocked by clozapine as well 55,56 epinephrine are also indicative of D4R receptor as other neuroleptics. DA inhibits NMDA receptor- involvement (Figure 5). mediated responses in the CA1 hippocampal region, Adenosylation of MET313 in the D4R is suggested by and clozapine is more effective in blocking this effect 57 several observations: (1) A receptor segment containing than the non-D4R-selective antagonist eticlopride. MET313 can be adenosylated by MAT (Figure 8c). (2) Moreover, NMDA receptors have been shown to form Incorporation of adenosyl-labeled [3H]ATP into a D4R complexes with several SH3 domain-containing pro- 58 59 immunoprecipitate was reduced by the MAT inhibitor teins, including Src and PSD-95. Thus DA may cycloleucine (Figure 8b). (3) Constitutive activity of the modulate the responsiveness of excitatory (and poss- D4R toward G proteins was abolished by cycloleucine ibly inhibitory) receptor systems by D4R-mediated (Figure 7b). Previous studies in a number of different alterations in their local membrane microenvironment. receptors have found that mutation of the MET313 locus results in constitutive receptor activity toward G Folate metabolism, D4R-mediated PLM and proteins.43,44 Thus in the case of the D4R, adenosyl- schizophrenia ation of the receptor in its R* state may interfere with Our results indicate that cyclic D4R-mediated PLM its return to the inactive R state. depends upon single-carbon units from the folate pool. Thus [14C]-formate labeling of the folate pool facilitated Functional significance of D4R-mediated the observation of DA-stimulated PLM whereas [3H]- phospholipid methylation methyl-methionine did not (Figure 6c). In contrast to Phospholipid methylation affects the fluid properties SAM-mediated PLM, it is clear that the D4R mech- of membranes, secondary to a reduction in headgroup anism is primarily dependent upon 5-methylTHF for packing density.49 Under these conditions fatty acid cyclic function, whereas the dietary intake of methion- tail segments have greater mobility with possible ine can supply methyl groups needed for the SAM- consequences for their interaction with transmembrane dependent process. Deficits in the availability of 5- proteins, especially for conformational changes which methylTHF would therefore have greater consequences occur within the lipid bilayer. For DA-stimulated PLM, for D4R-mediated PLM than for the non-receptor these effects will be most prominent in the immediate pathway. environment around the D4R, and the recently Folate and methionine metabolism are altered in described ability of D4Rs to form SH3 domain-depen- schizophrenia22–28 and in other neuropsychiatric dis- dent complexes with other proteins50,51 can provide a turbances,24,27 raising the possibility that dysfunctional targeting mechanism to increase the specificity of DA- D4R-mediated PLM may play a role in these disorders. stimulated PLM. The proline-rich hypervariable repeat Our finding of reduced PLM in lymphocytes of persons segment, located adjacent to the putative SH3 recog- with schizophrenia (Figure 10) is suggestive of this D4 dopamine receptor-mediated PLM A Sharma et al 244 possibility, although the relevance of this observation previously disparate areas of investigation. Most to brain function remains to be established. The exist- importantly, we hope that our findings may illuminate ence of D4Rs or even D2-like receptors in lymphocytes new approaches for the treatment of schizophrenia and has been the subject of much debate.60–62 However, perhaps other neuropsychiatric disorders. using PCR methodology, D4R mRNA has been demonstrated in both B- and T-lymphocytes,62 raising the Acknowledgements possibility that the inhibitory effect of haloperidol on PLM (Figure 10) might reflect D4R involvement. Our The authors wish to thank Dr Hubert van Tol for his preliminary results indicate the need for further lym- generosity in providing CHO cells expressing wild- phocyte studies using D4R-selective antagonists, in type or M313A mutant D4 dopamine receptors, to Dr order to determine whether D4Rs are indeed involved. Martha Teeter for her assistance in molecular modeling Schizophrenia is a disorder with a complex etiology, of dopamine receptors and to Marion Avtges RN and resulting from multiple interacting factors which Drs Godehard Oepen, James Koster and Lawrence Herz impinge upon a common process. A significant genetic for their assistance in clinical studies. Funding was contribution is likely, and several putative loci have provided by a grant from GlaxoWellcome (to RCD) and been proposed from linkage studies,63,64 although there from the schizophrenia research fund at the ENRM VA have been no reports of schizophrenia-associated D4R Medical Center (to CJD). mutations. The D4R-mediated PLM pathway outlined in Figure 2 incorporates several elements which are consistent with previous findings in schizophrenia. References These include: (1) involvement of a D2-like (or D4) 1 First MB (ed). Diagnostic and Statistical Manual of Mental Dis- dopamine receptor; (2) reduced activity of methionine orders (4th edn). Am Psychiatric Assoc: Washington, DC, 1994. 2 Matthysse S. Antipsychotic drugs: a clue to the neuropathology of cycle (MAT) and folate pathway (5,10-methylene schizophrenia? Fed Proc 1973; 32: 200–205. THFR and SHMT) enzymes; (3) reduced levels of phos- 3 Deutsch SI, Mastripaolo J, Schwartz BL, Rosse RB, Morihisa JM. A phatidylethanolamine (PE);65 (4) altered activity of ‘glutamatergic hypothesis’ of schizophrenia: rationale for pharma- membrane proteins such as NMDA receptors. Together cology with glycine. Clin Neuropharmacol 1989; 12: 1–13. these deficits illustrate a potential multifactorial basis 4 Benes FM. Is there a neuroanatomic basis for schizophrenia? An old question revisited. The Neuroscientist 1995; 1: 104–115. for schizophrenia. 5 Freedman R, Coon H, Myles-Worsley M, Orr-Urtreger A, Olincy A, Since they are antagonists at the D4R, neuroleptic Davis R. Linkage of a neurological deficit in schizophrenia to a drugs would be expected to acutely reduce D4R- chromosome 15 locus. Proc Natl Acad Sci USA 1997; 94: 587–592. mediated PLM and therefore worsen rather than help 6 Van Tol HHM, Bunzow JR, Guan H-C, Sunahara RK, Seeman P, Niznik HB et al. Cloning of the gene for a human dopamine D4 any schizophrenia-associated deficit. However, the receptor with high affinity for the antipsychotic clozapine. Nature clinical benefits of neuroleptics are better correlated 1991; 350: 610–614. with the timecourse of receptor up-regulation which 7 Seeman P, Guan H-G, Van Tol HHM. Dopamine D4 receptors are accompanies receptor blockade.66,67 An increased num- elevated in schizophrenia. Nature 1993; 365: 441–445. ber of receptors might partially compensate for the 8 Murray AM, Hyde TM, Knable MB, Herman MM, Bigelow LB, Carter JM et al. Distribution of putative D4 dopamine receptors in deficits which adversely affect D4R-mediated PLM. postmortem striatum from patients with schizophrenia. J Neurosci- The typical onset of schizophrenia, in late-teens and ence 1995; 15: 2186–2191. early-twenties, suggests an age-dependent triggering 9 Reynolds GP, Mason SL. Absence of detectable striatal dopamine event which may be superimposed on a previously D4 receptors in drug-treated schizophrenia. Eur J Pharmacol 1995; benign background of risk. Certainly the cessation of 281: R5–R6. 10 Kramer MS, Last B, Getson A, Reines SA & the D4 dopamine antag-

linear growth and changes in metabolism which onist group. The effects of a selective D4 dopamine receptor antag- accompany the termination of adolescence should be onist (L-745 870) in acutely psychotic inpatients with schizo- considered as potential triggering events. Indeed, the phrenia. Arch Gen Psychiatry 1997; 54: 567–572. uptake of methionine by the cerebral cortex shows a 11 Gazi L, Bobirnac I, Danzeisen M, Schu¨ pbach E, Bruinvels AT, Geisse S et al. The agonist activities of the putative antipsychotic remarkable age-dependent decrease up until the start of agents, L-745 870 and U-101958 in HEK293 cells expressing the 68 the third decade of life. Insofar as they might reﬂect human dopamine D4.4 receptor. Br J Pharmacol 1998; 124: 889–896. a reduced demand for de novo purine and thymidine 12 Van Tol HHM, Wu CM, Guan H-C, Ohara K, Bunzow JR, Civelli O synthesis, age-dependent metabolic changes could et al. Multiple dopamine D4 receptor variants in the human population. Nature 1992; 358: 149–152. alter the dynamics of folate metabolism, with possible 13 Lichter JB, Barr CL, Kennedy JL, Van Tol HHM, Kidd K, Livak KJ. A

consequences for D4R-mediated PLM. In particular, hypervariable segment in the human dopamine receptor D4 (DRD4) age-dependent changes in folate metabolism could gene. Human Molec Genet 1993; 2: 767–773. serve as a triggering event for the onset of schizo- 14 Ebstein RP, Novick O, Umansky R, Priel B, Osher Y, Blaine D et phrenia. al. Dopamine D4 receptor (D4DR) exon III polymorphism associated with the human personality trait of Novelty Seeking. Nature In summary, we have identiﬁed a novel, single-car- Genet 1996; 12: 78–80. bon folate-dependent activity of the D4 subtype dopa- 15 Benjamin J, Li L, Patterson C, Greenberg BD, Murphy DL, Hamer mine receptor, which allows the receptor to regulate DH. Population and familial association between the D4 dopamine membrane composition in its microenvironment. Rec- receptor gene and measures of Novelty Seeking. Nature Genet 1996; 12: 81–84. ognition of this activity affords a new perspective on 16 Muramatsu T, Higuchi S, Murayama M, Matsushita S, Hayashida the possible role of D4Rs in psychiatric illnesses such M. Association between alcoholism and the dopamine D4 receptor as schizophrenia, a perspective which encompasses gene. J Med Genet 1996; 33: 113–115. D4 dopamine receptor-mediated PLM A Sharma et al 245 17 LaHoste GJ, Swanson JM, Wigal SB, Glabe C, Wigal T, King N et analogues of S-adenosyl-L-homocysteine as potential inhibitors. al. Dopamine D4 receptor gene polymorphism is associated with Mol Pharmacol 1977; 13: 939–947. attention deﬁcit hyperactivity disorder. Mol Psychiatry 1996; 1: 40 Deacon R, Lumb M, Perry J, Chanarin I, Minty B, Halsey MJ et al.

121–124. Selective inactivation of vitamin B12 in rats by nitrous oxide. Lan- 18 Jo¨nsson EG, No¨then MM, Gustavsson JP, Neidt H, Brene´ S, Tylec cet 1978; ii: 1023–1024. A et al. Lack of evidence for allelic association between personality 41 Hirata F, Axelrod J. Phospholipid methylation and biological signal

traits and the dopamine D4 receptor gene polymorphisms. Am J transmission. Science 1980; 209: 1082–1090. Psychiatry 1997; 154: 697–699. 42 Farrens DL, Altenbach C, Yang K, Hubbell WL, Khorana G. Require- 19 Park LC, Baldessarini RJ, Kety SS. Methionine effects on chronic ment for rigid-body motion of transmembrane helices for light acti- schizophrenics. Arch Gen Psychiatry 1965; 12: 346–351. vation of rhodopsin. Science 1996; 74: 768–769. 20 Kety SS. Biochemical theories of schizophrenia. Science 1959; 129: 43 Kjelsberg MA, Cotecchia S, Ostrowski J, Caron MG, Lefkowitz RJ. ␣ 1528–1532. Constitutive activation of the 1B adrenergic receptor by all amino 21 Cohen SM, Nichols A, Wyatt R, Poulin W. The administration of acid substitutions at a single site: evidence for a region which con- methionine to chronic schizophrenic patients: a review of ten stud- strains receptor activation. J Biol Chem 1992; 267: 1430–1433. ies. Biol Psychiatry 1974; 8: 209–225. 44 Samama P, Cotecchia S, Costa T, Lefkowitz RJ. A mutation-induced ␤ 22 Sargent T, Kusubov N, Taylor SE, Budinger TF. Tracer kinetic evi- activated state of the 2-adrenergic receptor. J Biol Chem 1993; 268: dence for abnormal methyl metabolism in schizophrenia. Biol Psy- 4625–4636. chiatry 1992; 32: 1078–1090. 45 Harris RA, Schroeder F. Ethanol and the physical properties of 23 Ismail L, Sargent T, Dobson TEL, Pollycove M. Altered metabolism brain membranes. Mol Pharmacol 1981; 20: 128–137. of the methionine methyl group in the leukocytes of patients with 46 Mihic SJ, Ye Q, Wick MJ, Koltchine VV, Krasowski MD, Finn SE schizophrenia. Biol Psychiatry 1978; 12: 649–660. et al. Sites of alcohol and volatile anaesthetic action on GABAA and glycine receptors. Nature 1997; 389: 385–389. 24 Smythies JR, Alarcon RD, Morere DA, Monti JA, Steele M, Tolbert 47 Lovinger DM, White G, Weight FF. NMDA receptor-mediated syn- LC et al. Abnormalities of one-carbon metabolism in psychiatric aptic excitation selectively inhibited by ethanol in hippocampal disorders: study of methionine adenosyltransferase kinetics and slice from adult rat. J Neurosci 1990; 10: 1372–1379. lipid composition of erythrocyte membranes. Biol Psychiatry 1986; 48 Franks NP, Lieb WR. Molecular and cellular mechanisms of general 21: 1391–1398. anaesthesia. Nature 1994; 367: 607–614. 25 Carl GF, Crews EL, Carmichael SM, Benesh FC, Smythies JR. Four 49 Casal HL, Mantsch HH. The thermotropic phase behavior of N- enzymes of one-carbon metabolism in blood cells of schizo- methylated dipalmitoylphosphatidylethanolamines. Biochim phrenics. Biol Psychiatry 1978; 13: 773–776. Biophys Acta 1983; 735: 387–396. 26 Freeman JM, Finkelstein JD, Mudd SH. Folate-responsive homocys- 50 Oak JN, Oldenhof J, Schoots O, Van Tol HHM. Screening for novel tinuria and ‘schizophrenia’. New Engl J Med 1975; 292: 491–496. interactions with the polymorphic region of the human dopamine 27 Arinami T, Yamada N, Yamakawa K, Kobayashi N, Hamaguchi H, D4 receptor. Biochemistry 1998; 37: 15726–15736. Toru M. Methylenetetrahydrofolate reductase variant and 51 Liu Y-F, Deth RC. Association of the dopamine D4 receptor with schizophrenia/depression. Am J Med Genet 1997; 74: 526–528. SH3 domain-containing proteins. J Biol Chem (under review). 28 Regland B, Germgard T, Gottfries CG, Grenfeldt B, Koch-Schmidt 52 Mancillas J, Siggins GR, Bloom FE. Systemic ethanol: selective AC. Homozygous thermolabile methylenetetrahydrofolate enhancement of responses to acetylcholine and somatostatin in the reductase in schizophrenia-like psychosis. J Neural Trans 1997; rat hippocampus. Science 1986; 231: 161–163. 104: 931–941. 53 Arvanov VL, Liang X, Schwartz J, Grossman S, Wang RY. Clozapine 29 Le Fur G, Phan T, Canton T, Tur C, Uzan A. Evidence for a coupling and haloperidol modulate N-methyl-D-aspartate- and non-N- between dopaminergic receptors and phospholipid methylation in methyl-D-aspartate receptor-mediated neurotransmission in rat mouse B lymphocytes. Life Sci 1981; 29: 2737–2749. prefrontal cortical neurons in vitro. J Pharmacol Exp Ther 1997; 30 Leprohon CE, Blusztajn JK, Wurtman RJ. Dopamine stimulation of 283: 226–234. phosphatidylcholine (lecithin) biosynthesis in rat brain neurons. 54 Olney JW, Farber NB. Efficacy of clozapine compared with other Proc Natl Acad Sci USA 1983; 80: 2063–2066. antipsychotics in preventing NMDA-antagonist neurotoxicity. J 31 De La Rosa J, Le Gros HL Jr, Geller AM, Kotb M. Changes in the Clin Psychiatr 1994; 55 (Suppl B): 43–46. relative amount of subunits of methionine adenosyltransferase in 55 Malhotra AK, Adler CM, Kennison SD, Elman I, Pikar D, Breier A. human lymphocytes upon stimulation with a polyclonal T cell Clozapine blunts N-methyl-D-aspartate antagonist-induced psy- mitogen. J Biol Chem 1992; 267: 10699–10704. chosis: a study with ketamine. Biol Psychiatry 1997; 42: 664–668. 32 Sidhu A, Fishman PH. Identification and characterization of func- 56 Yamamoto M, Mizuki Y, Suetsugi M, Ozawa Y, Ooyama M, Suzuki tional D1 dopamine receptors in a human neuroblastoma cell line. M. Effects of dopamine antagonists on changes in spontaneous EEG Biochem Biophys Res Comm 1990; 166: 574–579. and locomotor activity in ketamine-treated rats. Pharmacol 33 Kulagowski JJ, Broughton HB, Curtis NR, Mawer IM, Ridgill MP, Biochem Behav 1997; 57: 361–365. Baker R et al. 3-[[4-(4- Chlorophenyl)piperazin-1-yl]-methyl]-1H- 57 Otmakhova NA, Lisman JE. Dopamine selectively inhibits the pyrrolo[2,3-b]pyridine: an antagonist with high affinity and selec- direct cortical pathway to the CA1 hippocampal region. J Neurosci tivity for the human dopamine D4 receptor. J Med Chem 1996; 39: 1998; 19: 1437–1445. 1941–1942. 58 Yu X-M, Askalan R, Keil GJ II, Salter M. NMDA channel regulation 34 Schlacter SK, Poel TJ, Lawson CF, Dinh DM, Lajiness ME, Romero by channel-associated protein tyrosine kinase Src. Science 1997; AG et al. Substituted 4-aminopiperidines having high in vitro 275: 674–678. affinity and selectivity for the human dopamine D4 receptor. Eur J 59 Kornau H-C, Schenker LT, Kennedy MB, Seeburg PH. Domain Pharmacol 1997; 322: 283–286. interaction between NMDA receptor subunits and the postsynaptic 35 Chabert C, Cavegn C, Bernard A, Mills A. Characterization of the density protein PSD-95. Science 1995; 269: 1737–1740. functional activity of dopamine ligands at human recombinant 60 Bondy B, Ackenheil M, Bizle W, Elbers R, Fro¨hler M. Catecholam- dopamine D4 receptors. J Neurochem 1994; 63: 62–65. ines and their receptors in blood: evidence for alterations in schizo- 36 Zorn SH, Jackson ER, Johnson CG, Lewis JM, Fliri A. CP-226,269 phrenia. Biol Psychiatry 1984; 19: 1377–1393. is a selective dopamine D4 receptor agonist. Soc Neurosci Abs 61 Vile JM, Strange PG. D2-like dopamine receptors are not detectable 1997; 23: 685. on human peripheral blood lymphocytes. Biol Psychiatry 1996; 40: 37 Lanau F, Zenner M-T, Civelli O, Hartman D. Epinephrine and nor- 881–885. epinephrine act as potent agonists at the recombinant human dopa- 62 Bondy B, de Jonge S, Pander S, Primbs J, Ackenheil M. Identifi- mine D4 receptor. J Neurochem 1997; 68: 804–812. cation of dopamine D4 receptor mRNA in circulating human lym- 38 Lombardini JB, Talalay P. Analogues of methionine as substrates phocytes using nested polymerase chain reaction. J Immunol 1996; and inhibitors of the methionine adenosyltransferase reaction. Mol 71: 139–144. Pharmacol 1970; 6: 481–499. 63 Karayiorgou M, Morris MA, Morrow B, Shprintzen RJ, Goldberg R, 39 Chiang PK, Richards HH, Cantoni GL. Homocysteine hydrolase: Borrow J et al. Schizophrenia susceptibility associated with inter- D4 dopamine receptor-mediated PLM A Sharma et al 246 stitial deletions of chromosome 22q11. Proc Natl Acad Sci USA neostriatum: effects of chronic treatment with typical and atypical 1995; 92: 7612–7616. antipsychotic drugs. J Pharmacol Exp Ther 1997; 283: 939–946. 64 Wang S, Sun C, Walczak CA, Ziegle JS, Kipps BR, Goldin LR et al. 67 Tarazi FI, Yeghiayan SK, Baldessarini RJ, Kula NS, Neumeyer JL. Evidence for a susceptibility locus for schizophrenia on chromo- Long-term effects of S(+)N-n-propylnorapomorphine compared some 6pter-p22. Nature Genet 1995; 10: 41–46. with typical and atypical antipsychotics: differential increases of

65 Rotrosen J, Wolkin A. Phospholipid and prostaglandin hypotheses cerebrocortical D2-like and striatolimbic D4-like dopamine recep- of schizophrenia. In: Meltzer H (ed). Psychopharmacology: The tors. Neuropsychopharmacology 1997; 17: 186–196. Third Generation of Progress. Raven Press: New York, 1987, 68 O’Tauma LA, Phillips PC, Smith QR, Uno Y, Dannals RF, Wilson pp 759–764. AA et al. L-Methionine uptake by human cerebral cortex: matu- 66 Lidow MS, Goldman-Rakic PS. Differential regulation of D2 and ration from infancy to old age. J Nucl Med 1991; 32: 16–22. D4 dopamine receptor mRNAs in the primate cerebral cortex vs