Proteomic analysis of nuclei dissected from fixed rat brain tissue using expression microdissection.

Expression microdissection (xMD) is a high-throughput, operator-independent technology that enables the procurement of specific cell populations from tissue specimens. In this method, histological sections are first stained for cellular markers via either chemical or immuno-guided methods, placed in close contact with an ethylene vinyl acetate (EVA) film, and exposed to a light source. The focal, transient heating of the stained cells or subcellular structures melts the EVA film selectively to the targets for procurement. In this report, we introduce a custom-designed flashcube system that permits consistent and reproducible microdissection of nuclei across an FFPE rat brain tissue section in milliseconds. In addition, we present a method to efficiently recover and combine captured proteins from multiple xMD films. Both light and scanning electron microscopy demonstrated captured nuclear structures. Shotgun proteomic analysis of the samples showed a significant enrichment in nuclear localized proteins, with an average 25% of recovered proteins localized to the nucleus, versus 15% for whole tissue controls (p < 0.001). Targeted mass spectrometry using multiple reaction monitoring (MRM) showed more impressive data, with a 3-fold enrichment in histones, and a concurrent depletion of proteins localized to the cytoplasm, cytoskeleton, and mitochondria. These data demonstrate that the flashcube-xMD technology is applicable to the proteomic study of a broad range of targets in molecular pathology.

Effects of systemic and central nervous system localized inflammation on the contributions of metabolic precursors to the L-kynurenine and quinolinic acid pools in brain.

L-Kynurenine and quinolinic acid are neuroactive L-tryptophan-kynurenine pathway metabolites of potential importance in pathogenesis and treatment of neurologic disease. To identify precursors of these metabolites in brain, [(2)H(3) ]-L-kynurenine was infused subcutaneously by osmotic pump into three groups of gerbils: controls, CNS-localized immune-activated, and systemically immune-activated. The specific activity of L-kynurenine and quinolinate in blood, brain and systemic tissues at equilibrium was then quantified by mass spectrometry and the results applied to a model of metabolism to differentiate the relative contributions of various metabolic precursors. In control gerbils, 22% of L-kynurenine in brain was derived via local synthesis from L-tryptophan/formylkynurenine versus 78% from L-kynurenine from blood. Quinolinate in brain was derived from several sources, including: local tissue L-tryptophan/formylkynurenine (10%), blood L-kynurenine (35%), blood 3-hydroxykynurenine/3-hydroxyanthranilate (7%), and blood quinolinate (48%). After systemic immune-activation, however, L-kynurenine in brain was derived exclusively from blood, whereas quinolinate in brain was derived from three sources: blood L-kynurenine (52%), blood 3-hydroxykynurenine or 3-hydroxyanthranilate (8%), and blood quinolinate (40%). During CNS-localized immune activation, > 98% of both L-kynurenine and quinolinate were derived via local synthesis in brain. Thus, immune activation and its site determine the sources from which L-kynurenine and quinolinate are synthesized in brain. Successful therapeutic modulation of their concentrations must take into account the metabolic and compartment sources.

LC/MS analysis of NAD biosynthesis using stable isotope pyridine precursors.

A liquid chromatographic-electrospray ionization ion trap mass spectrometry (LC/MS) method has been developed to measure the biosynthetic incorporation of specific precursors into NAD. The stable isotope-labeled precursors tryptophan, quinolinic acid, nicotinic acid, and nicotinamide were added to the media of human liver tumor cells (SK-HEP) grown in culture. The cells were harvested, the NAD was extracted, and the ratio of labeled to unlabeled NAD was measured using the newly developed LC/MS assay. The quantity of NAD formed from each precursor relative to an internal standard (fully labeled 13C, 15N-labeled NAD prepared from baker's yeast) was measured. The detection limit (signal-to-noise ratio 5:1) of the LC/MS method was 37 fmol (25 pg) of NAD and was linear from 20.0 ng to 25 pg. All reported NAD levels were normalized relative to cellular protein measurements. At 50 microM precursor concentrations, nicotinamide was the dominant precursor and NAD levels in the cell rose well above normal levels. Other precursors were minimally incorporated. The same methods were applied to NAD biosynthesized by macrophages derived from peripheral blood monocytes. However, the NAD concentration in macrophages was about 5% of that in SK-HEP cells and the incorporation of stable isotope-labeled substrates remained below measurable levels.

Gbetagamma affinity for bovine rhodopsin is determined by the carboxyl-terminal sequences of the gamma subunit.

Two native betagamma dimers, beta(1)gamma(1) and beta(1)gamma(2), display very different affinities for receptors. Since these gamma subunits differ in both primary structure and isoprenoid modification, we examined the relative contributions of each to Gbetagamma interaction with receptors. We constructed baculoviruses encoding gamma(1) and gamma(2) subunits with altered CAAX (where A is an aliphatic amino acid) motifs to direct alternate or no prenylation of the gamma chains and a set of gamma(1) and gamma(2) chimeras with the gamma(2) CAAX motif at the carboxyl terminus. All the gamma constructs coexpressed with beta(1) in Sf9 cells yielded beta(1)gamma dimers, which were purified to near homogeneity, and their affinities for receptors and Galpha were quantitatively determined. Whereas alteration of the isoprenoid of gamma(1) from farnesyl to geranylgeranyl and of gamma(2) from geranylgeranyl to farnesyl had no impact on the affinities of beta(1)gamma dimers for Galpha(t), the non-prenylated beta(1)gamma(2) dimer had significantly diminished affinity. Altered prenylation resulted in a <2-fold decrease in affinity of the beta(1)gamma(2) dimer for rhodopsin and a <3-fold change for the beta(1)gamma(1) dimer. In each case with identical isoprenylation, the beta(1)gamma(2) dimer displayed significantly greater affinity for rhodopsin compared with the beta(1)gamma(1) dimer. Furthermore, dimers containing chimeric Ggamma chains with identical geranylgeranyl modification displayed rhodopsin affinities largely determined by the carboxyl-terminal one-third of the protein. These results indicate that isoprenoid modification of the Ggamma subunit is essential for binding to both Galpha and receptors. The isoprenoid type influences the binding affinity for receptors, but not for Galpha. Finally, the primary structure of the Ggamma subunit provides a major contribution to receptor binding of Gbetagamma, with the carboxyl-terminal sequence conferring receptor selectivity.

Glutamate is a mediator of neurotoxicity in secretions of activated HIV-1-infected macrophages.

We sought to identify neurotoxin(s) secreted by HIV-1-infected mononuclear phagocytes that could contribute to the pathophysiology of HIV-1-associated dementia (HAD). Neurotoxic factors were characterized in batches of conditioned media (CM) from human monocyte-derived macrophages (MDM) infected with HIV-1(ADA) and/or activated with lipopolysaccharide (LPS). All of the neurotoxicity was: present in the <3000-Da fraction; blocked by 5 microM MK801; and not trypsin sensitive or extractable into polar organic solvents. Glutamate measured in CM accounted for all neurotoxic effects observed from HIV/LPS CM in astrocyte-poor neuronal cultures and may contribute to the pathophysiology of HIV-1-associated dementia.

Base- and acid-catalyzed interconversions of O-acyl- and N-acyl-ethanolamines: a cautionary note for lipid analyses.

The isolation and quantification of ethanolamine containing lipids from animal tissues may expose neutral lipid extracts to acidic or basic conditions during chromatographic separations or derivatization chemistry. While investigating the acid- and base-catalyzed production of anandamide in chromatographic fractions of rat brain extracts not containing anandamide, we observed that O,N-acyl migrations are facile. O,N-acyl migrations are well documented in synthetic organic chemistry literature, but are not well described or recognized with regard to methods in lipid isolation or lipid enzyme studies. We report here the synthesis and characterization of O- and N-acyl (palmitoyl- or arachidonoyl-)ethanolamines. Their rearrangements in base and acid are quantified by liquid chromatography;-electrospray ionization mass spectrometry. The rearrangements proceed through a cyclic intermediate that is also formed during chemical reactions commonly used for derivatization of acylethanolamines for gas chromatography-mass spectrometry. The isolation and characterization of N- or O-acylethanolamines and their enzymatic formation requires awareness and consideration of the proclivity of these compounds to chemically rearrange.

GC/MS analysis of anandamide and quantification of N-arachidonoylphosphatidylethanolamides in various brain regions, spinal cord, testis, and spleen of the rat.

Anandamide [N-arachidonoylethanolamide (NAE)] was initially isolated from porcine brain and proposed as an endogenous ligand for cannabinoid receptors in 1992. Accumulating evidence has now suggested that, in the tissue, NAE is generated from N-arachidonoylphosphatidylethanolamides (N-ArPEs) by phosphodiesterase. In this study a sensitive and specific procedure was developed to quantify NAE and N-ArPE, including organic solvent extraction, reverse-phase C-18 cartridge separation, derivatization, and gas chromatography/mass spectrometry (GC/MS) analysis. NAE is converted by a two-step derivatization procedure to a pentafluorobenzoyl ester followed by pentafluoropropionyl acylation. Quantification was performed by isotope dilution GC/MS using deuterium-labeled NAE (NAE-2H8) as an internal standard. The same chemical derivatization was applicable to N-ArPE quantification. The separated N-ArPE fractions were converted by a two-step cleavage/derivatization procedure into the pentafluorobenzoyl ester of NAE and then to its pentafluoropropionyl amide. The derivative was quantified by GC/MS using deuterium-labeled 1,2-[2H8]dioleoyl-sn-glycero-3-phospho(arachidonoyl)ethanolamid e as an internal standard. Using these methods, we have found that endogenous NAE levels in rat brain, spleen, testis, liver, lung, and heart were below the level of quantification achievable (0.1 pmol/mg of protein) but that N-ArPE is readily quantifiable and is widely distributed in the rat CNS with the highest level in the spinal cord. The striatum, hippocampus, and accumbens contain intermediate concentrations of N-ArPE, whereas the value is lowest in the cerebellum.

Labeled kynurenine pharmacokinetic modeling studies in gerbils. Nonequilibrium between infused and endogenous kynurenine.

In order to complete pharmacokinetic studies on the central vs. peripheral origin of several tryptophan metabolites, we infused gerbils with labelled kynurenine (2H4 or 15N2). Osmotic minipumps charged with kynurenine solutions were surgically implanted subcutaneously in adult female gerbils (50-60 g). After a variable number of hours, the gerbils were sacrificed and organs taken for determination of labelled/unlabelled kynurenine ratios using mass spectrometric assay of a pentafluorobenzyl derivative as described previously. Surprisingly high ratios of 2H to 1H-kynurenine were measured in the kidney (0.25-0.40) and urine (4.0-8.0), although the ratio of deuterium labelled to endogenous kynurenine remained below detection limits (< 0.05) in serum and other tissues. Infusion of greater quantities of 2H4-kynurenine confirmed these observations in gerbils in which ratios of 2H4-to-1H kynurenine were measurable in serum and tissues. Synthesis and infusion of 15N2-kynurenine demonstrated that these effects were not due to deuterium isotope substitution. The data demonstrate a non-equilibrium between infused and endogenous kynurenine, which is related to differential rates of protein binding and the rapid clearance of free, infused kynurenine by kidney.

Sources of the neurotoxin quinolinic acid in the brain of HIV-1-infected patients and retrovirus-infected macaques.

This study investigated the sources of quinolinic acid, a neurotoxic tryptophan-kynurenine pathway metabolite, in the brain and blood of HIV-infected patients and retrovirus-infected macaques. In brain, quinolinic acid concentrations in HIV-infected patients were elevated by > 300-fold to concentrations that exceeded cerebrospinal fluid (CSF) by 8.9-fold. There were no significant correlations between elevated serum quinolinic acid levels with those in CSF and brain parenchyma. Because nonretrovirus-induced encephalitis confounds the interpretation of human postmortem data, rhesus macaques infected with retrovirus were used to examine the mechanisms of increased quinolinic acid accumulations and determine the relationships of quinolinic acid to encephalitits and systemic responses. The largest kynurenine pathway responses in brain were associated with encephalitis and were independent of systemic responses. CSF quinolinic acid levels were also elevated in all infected macaques, but particularly those with retrovirus-induced encephalitis. In contrast to the brain changes, there was no difference in any systemic measure between macaques with encephalitis vs. those without. Direct measures of the amount of quinolinic acid in brain derived from blood in a macaque with encephalitis showed that almost all quinolinic acid (>98%) was synthesized locally within the brain. These results demonstrate a role for induction of indoleamine-2,3dioxygenase in accelerating the local formation of quinolinic acid within the brain tissue, particularly in areas of encephalitis, rather than entry of quinolinic acid into the brain from the meninges or blood. Strategies to reduce QUIN production, targeted at intracerebral sites, are potential approaches to therapy.

Effect of intermittent urethral catheterization and oxybutynin bladder instillation on urinary continence status and quality of life in a selected group of spinal cord injury patients with neuropathic bladder dysfunction.

OBJECTIVES: A comparative assessment of (i) urinary continence status, (ii) quality of life, and (iii) sexuality in spinal cord injury patients prior to, and during intermittent catheterization with adjunctive intravesical oxybutynin therapy (Cystin: manufactured by Leiras Oy. Helsinki, Finland). SETTING: A hospital, and community-based study of selected adult, male, spinal cord injury patients registered with the Regional Spinal Injuries Centre, Southport. PATIENTS: Seven patients (mean age: 44.3 years) suffering from neuropathic bladder due to suprasacral spinal cord lesion of traumatic aetiology, and well settled in the community in the north-west of England were the subjects of this study. Before commencing the intermittent catheterization regime, these patients were on penile sheath drainage. INTERVENTION: Intermittent urethral catheterization was performed with sterile, single-use Nelaton catheters 5-6 times a day with intravesical instillation of oxybutynin 5 mg in 30 ml. 1-3 times a day for periods ranging from 14 to 30 months. OUTCOME MEASURES: Assessment of urinary continence, sexuality, and quality of life was made (i) at the outset before any intervention, (ii) during intermittent catheterization regime, and (iii) when the patients were using the oxybutynin bladder instillation along with intermittent catheterization. RESULTS: Initially all the seven patients were constantly wearing penile sheaths and leg bags. When these patients performed intermittent catheterization 5-6 times in 24 h, they attempted to discard the penile sheath during the day but they were experiencing mild to moderate urine leak between catheterization. They were compelled to wear penile sheaths during night. Subsequently, five patients took oxybutynin by mouth, but developed an unacceptable degree of side-effects necessitating discontinuation of the medication. Following commencement of intravesical oxybutynin therapy, all of them were able to discard the penile sheaths and leg bags during the day as well as during the night. However, on waking-up after a full night's sleep, three patients found dampness of their undergarments 1-2 times per week. None of the patients experienced side-effects attributable either to the intermittent catheterization procedure, or to the intravesical oxybutynin therapy. The number of episodes of urinary infection requiring antibiotic therapy was 0.08/patient/month. All the seven patients noticed a remarkable improvement in the quality of life because they had achieved a high degree of continence. All the seven patients commented on the improved sense of their own sexuality which was attributed to (i) absence of incontinence episodes, (ii) improved self-image, and (iii) not wearing penile sheaths and leg bags. CONCLUSION: These seven spinal cord injury patients achieved socially acceptable continence with improved quality of life, and enhanced sexuality with the intermittent urethral catheterization regime and intravesical oxybutynin therapy.

Endozepine stupor. Recurring stupor linked to endozepine-4 accumulation.

Recurring stupor can be caused by repeated metabolic, toxic or structural brain disturbances. Recently, cases of recurring stupor, with fast EEG activity were shown to display increased endogenous benzodiazepine-like activity during the episodes of stupor. Patients with recurring stupor underwent extensive metabolic and toxicologic screening, EEG and brain imaging. Endozepines and exogenously administered benzodiazepines were assayed in plasma and CSF by means of mass spectrometry. Flumazenil, a benzodiazepine antagonist was administered and the behavioural and EEG responses monitored. Treatment with oral flumazenil was attempted in selected cases. Twenty patients were found with recurring stupor. Episodes had begun between ages 18 and 67 years, and in nine patients, had disappeared spontaneously after 4-6 years with symptoms. Stupor lasted hours or days. Onset of the episodes and frequency were unpredictable. Patients were normal between attacks. Stupor was characterized by initial drowsiness, staggering and behavioural changes, followed by deep sleep and spontaneous recovery with post-ictal amnesia. Biochemical screening and brain imaging were always normal. Ictal EEG showed fast background activity, and flumazenil transiently awoke the patients and normalized the EEG. In the nine cases examined, endozepine-4 levels were increased during the stupor. Oral flumazenil reduced the frequency of the attacks in three of these nine patients. Recurring episodes of stupor may be due to increased endozepine-4. We propose the term 'endozepine stupor' for such episodes. Endozepine-4 is an endogenous ligand for the benzodiazepine recognition site at the GABAA receptor, with unknown molecular structure.

Species heterogeneity between gerbils and rats: quinolinate production by microglia and astrocytes and accumulations in response to ischemic brain injury and systemic immune activation.

Quinolinic acid is an excitotoxic kynurenine pathway metabolite, the concentration of which increases in human brain during immune activation. The present study compared quinolinate responses to systemic and brain immune activation in gerbils and rats. Global cerebral ischemia in gerbils, but not rats, increased hippocampus indoleamine-2,3-dioxygenase activity and quinolinate levels 4 days postinjury. In a rat focal ischemia model, small increases in quinolinate concentrations occurred in infarcted regions on days 1, 3, and 7, although concentrations remained below serum values. In gerbils, systemic immune activation by an intraperitoneal injection of endotoxin (1 mg/kg of body weight) increased quinolinate levels in brain, blood, lung, liver, and spleen, with proportional increases in lung indoleamine-2,3-dioxygenase activity at 24 h postinjection. In rats, however, no significant quinolinate content changes occurred, whereas lung indoleamine-2,3-dioxygenase activity increased slightly. Gerbil, but not rat, brain microglia and peritoneal monocytes produced large quantities of [13C(6)]-quinolinate from L-[13C(6)]tryptophan. Gerbil astrocytes produced relatively small quantities of quinolinate, whereas rat astrocytes produced no detectable amounts. These results demonstrate that the limited capacity of rats to replicate elevations in brain and blood quinolinic acid levels in response to immune activation is attributable to blunted increases in local indoleamine-2,3-dioxygenase activity and a low capacity of microglia, astrocytes, and macrophages to convert L-tryptophan to quinolinate.

Pentoxifylline decreases brain levels of platelet activating factor in murine AIDS.

Tumor necrosis factor-alpha (TNF-alpha) and platelet-activating factor (PAF) have been implicated in the pathogenesis of human immunodeficiency virus (HIV)-associated encephalopathy. The effects of pentoxifylline on brain PAF levels were examined in mice infected with the LP-BM5 murine leukemia virus (MuLV). Seven weeks after viral inoculation, significant increases in serum TNF-alpha and brain PAF levels were observed. One week of treatment with pentoxifylline initiated 6 weeks postinfection significantly reduced both serum TNF-alpha and brain PAF levels. A significant positive correlation was observed between the levels of these substances (r = 0.62; P < 0.01). This study demonstrates that pentoxifylline treatment was effective in decreasing the levels of TNF-alpha in the serum and PAF levels in the brain of mice infected with the LP-BM5 MuLV.

Human microglia convert l-tryptophan into the neurotoxin quinolinic acid.

Immune activation leads to accumulations of the neurotoxin and kynurenine pathway metabolite quinolinic acid within the central nervous system of human patients. Whereas macrophages can convert L-tryptophan to quinolinic acid, it is not known whether human brain microglia can synthesize quinolinic acid. Human microglia, peripheral blood macrophages and cultures of human fetal brain cells (astrocytes and neurons) were incubated with [13C6]L-tryptophan in the absence or presence of interferon gamma. [13C6]Quinolinic acid was identified and quantified by gas chromatography and electron-capture negative-chemical ionization mass spectrometry. Both L-kynurenine and [13C6]quinolinic acid were produced by unstimulated cultures of microglia and macrophages. Interferon gamma, an inducer of indoleamine 2,3-dioxygenase, increased the accumulation of L-kynurenine by all three cell types (to more than 40 microM). Whereas large quantities of [13C6]quinolinic acid were produced by microglia and macrophages (to 438 and 1410 nM respectively), minute quantities of [13C6]quinolinic acid were produced in human fetal brain cultures (not more than 2 nM). Activated microglia and macrophage infiltrates into the brain might be an important source of accelerated conversion of L-tryptophan into quinolinic acid within the central nervous system in inflammatory diseases.

Assessment of DNA oxidative damage by quantification of thymidine glycol residues using gas chromatography/electron capture negative ionization mass spectrometry.

A technique to assess DNA oxidative damage by quantification of thymidine glycol residues is described. 2-Methylglycerate was released from thymidine glycol in DNA by alkaline cleavage/borodeuteride reduction, then derivatized to form a combined pentafluorobenzyl-tertbutyldimethylsilyl (PFB-TBDMS) derivative and analyzed by gas chromatography/electron capture negative ionization mass spectrometry. [2H4]Thymine glycol was used as an internal standard. The derivatization chemistry was assessed by using [14C-methyl]glycerate. Successful esterification was achieved with 75-80% yield using tetrabutylammonium sulfate-assisted anhydrous pentafluorobenzylation. The PFB-TBDMS derivative exhibits excellent chromatographic and detection properties with a detection limit of 41 amol injected on column. Freshly dissolved calf thymus DNA was used to test the method performance. The background level of thymidine glycol detected in this DNA was 11.7 +/- 0.3 x 10(-6) mol thymidine glycol per 1 mol thymidine. The thymidine glycol background in undamaged DNA establishes a lower limit of oxidative damage below which biological oxidation events would not be measured by this method. The method was linear for 4-20 micrograms DNA added per tube. The minimum measurable amount of thymidine glycol in DNA sample was 36 fmol. An increased level of thymidine glycol was measured in salmon sperm DNA which had autoxidized during storage in a refrigerated aqueous solution, 71.2 +/- 14.3 x 10(-6) mol thymidine glycol per 1 mol thymidine.

Nicardipine and MK-801 attenuate platelet-activating factor increases following cerebral ischemia-reperfusion in gerbils.

The effects of pretreatment with nicardipine (dihydropyridine Ca2+ channel antagonist), Bay K8644 (dihydropyridine Ca2+ channel agonist), and MK-801 (N-methyl-D-aspartate-receptor antagonist) on changes of platelet-activating factor (PAF) concentrations in transient ischemic brain are reported. The tissue concentration of PAF increases significantly in hippocampus, cortex and thalamus by 210%, 169% and 168% of controls without ischemia-reperfusion, respectively after 1 h of reperfusion. Nicardipine (5 mg/kg) reduces the accumulation of PAF, the remaining increases in hippocampus, cortex and thalamus being 151%, 138% and 145% of the controls, respectively. In contrast, Bay K8644 (2.5 mg/kg) enhances the accumulation of PAF, its concentrations in hippocampus, cortex and thalamus being 376%, 233% and 204% of the controls, respectively. The Bay K8644 enhancement in hippocampus is completely inhibited by pretreatment of nicardipine (5 mg/kg). MK-801 (10 mg/kg) reduces the accumulation of PAF, the remaining increases in hippocampus, cortex and thalamus being 152%, 147% and 144% of the controls, respectively. Moreover, brain tissue from animals subjected to the combined pretreatment with nicardipine (5 mg/kg) and MK-301 (10 mg/kg) indicates there is greater inhibition of ischemia-induced PAF increases than with either drug alone. These results indicate that PAF production in the ischemic brain may be regulated by Ca2+ influx through voltage-sensitive Ca2+ channels which are antagonized and agonized by nicardipine and Bay K8644, respectively and receptor-operated Ca2+ channels which are antagonized by MK-801. Because it is known that increases of intracellular Ca2+ in the brain accompany ischemia and early periods of reperfusion and that PAF exhibits neurotoxicity, the present findings support the role of PAF as a mediator in ischemia-induced brain damage at early stages of reperfusion.

Facile oxidative decarboxylation of 3,4-dihydroxyphenylacetic acid catalyzed by copper and manganese ions.

Under physiological conditions, we observed the rapid, pH- and temperature-dependent, oxidative decarboxylation and hydration of 3,4-dihydroxyphenylacetic acid (DOPAC) to form 3,4-dihydroxybenzyl alcohol (DBAlc). This product was oxidized and underwent tautomerization to form 3,4-dihydroxybenzaldehyde (DBAld). This reaction did not occur in the presence of EDTA, was catalyzed by copper (CuI, CuII) and manganese (MnII) and was oxygen dependent. A variety of mono- and dihydroxyphenyl carboxylic acids were tested and the reaction producing DBAlc as an intermediate was observed to be unique to DOPAC. 3.4-Dihydroxymandelic acid (DOMA) was rapidly oxidatively decarboxylated to form DBAld directly. The substrate and catalyst selectivity of this reaction suggest that this may have physiological relevance in the neurotoxic consequences of manganese and copper to the dopaminergic system in man.