Parkin-deficient Mice Exhibit Nigrostriatal Deficits but Not Loss of Dopaminergic Neurons*

Loss-of-function mutations in parkin are the major cause of early-onset familial Parkinson's disease. To investigate the pathogenic mechanism by which loss of parkin function causes Parkinson's disease, we generated a mouse model bearing a germline disruption in parkin. Parkin–/– mice are viable and exhibit grossly normal brain morphology. Quantitative in vivo microdialysis revealed an increase in extracellular dopamine concentration in the striatum of parkin–/– mice. Intracellular recordings of medium-sized striatal spiny neurons showed that greater currents are required to induce synaptic responses, suggesting a reduction in synaptic excitability in the absence of parkin. Furthermore, parkin–/– mice exhibit deficits in behavioral paradigms sensitive to dysfunction of the nigrostriatal pathway. The number of dopaminergic neurons in the substantia nigra of parkin–/– mice, however, is normal up to the age of 24 months, in contrast to the substantial loss of nigral neurons characteristic of Parkinson's disease. Steady-state levels of CDCrel-1, synphilin-1, and α-synuclein, which were identified previously as substrates of the E3 ubiquitin ligase activity of parkin, are unaltered in parkin–/– brains. Together these findings provide the first evidence for a novel role of parkin in dopamine regulation and nigrostriatal function, and a non-essential role of parkin in the survival of nigral neurons in mice.

Parkinson's disease (PD) 1 is an age-related movement disorder characterized by bradykinesia, rigidity, resting tremor, and postural instability. The neuropathologic hallmarks of PD are the loss of dopaminergic neurons in the substantia nigra (SN) and the presence of intraneuronal cytoplasmic inclusions known as Lewy bodies. The clinical manifestations of PD are due to progressive degeneration of dopaminergic neurons in the pars compacta of the SN that give rise to the nigrostriatal pathway, causing dopamine (DA) depletion in the striatum, where it is required for normal motor function. Little is known about the mechanisms of PD pathogenesis and nigral degeneration, although DA neurons have been shown to be susceptible to oxidative stress (1), mitochondrial defects (2), and environmental toxins (3).
The recent identification of genes linked to familial forms of PD (FPD) makes it possible to investigate the pathogenic mechanism by employing genetic approaches (4 -6). Over fifty recessively inherited mutations, including deletion, frameshift, nonsense, and missense mutations, have been identified in parkin in large numbers of families, making parkin the major gene responsible for early-onset FPD (7)(8)(9)(10). Although the first report linked parkin mutations to autosomal recessive juvenile parkinsonism (AR-JP) with atypical clinical features (5), many more cases identified subsequently were considered typical early-onset FPD with symptoms often indistinguishable from sporadic PD (9,11). Autopsies of limited numbers of patients showed selective loss of dopaminergic neurons in the SN either in the absence (12)(13)(14)(15) or in the presence (16) of Lewy bodies. The recessive inheritance mode and variety of parkin mutations indicate a loss-of-function pathogenic mechanism.
Parkin is widely expressed in most tissues including brain and heart (5). Although its transcripts are equally abundant in various brain sub-regions, parkin protein is enriched in the SN (5,(17)(18)(19)(20). In vitro studies have shown that parkin can function as an E3 ubiquitin ligase, mediating the covalent transfer of ubiquitin to protein substrates subject to proteasomal degradation (21)(22)(23). However, it is unclear how loss of parkin function leads to nigral degeneration and PD.
To investigate the pathogenic mechanism of PD in an animal model and to elucidate the normal physiological role of parkin in vivo, we created a mouse model bearing a targeted germline disruption of parkin. Molecular, histological, neurochemical, electrophysiological, behavioral, and biochemical analyses of parkinϪ/Ϫ mice reveal a novel role for parkin in dopamine regulation and nigrostriatal function in vivo, and a non-essential role of parkin in the survival of nigral neurons in mice.

EXPERIMENTAL PROCEDURES
Generation of parkinϪ/Ϫ Mice-A targeting vector was constructed using 1.8-and 3.5-kb DNA fragments as the 5Ј and 3Ј homologous sequences, respectively (Fig. 1A). A negative selection cassette, PGK-dt, which encodes the diphtheria toxin and has been shown to enhance * This work was supported by NINDS, National Institutes of Health and the Michael J. Fox Foundation for Parkinson's Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ ‡ To whom correspondence should be addressed: Center for Neurologic Diseases, Harvard Institutes of Medicine, 77 Avenue Louis Pasteur, Boston, MA 02115. Tel.: 617-525-5561; Fax: 617-525-5252; E-mail: jshen@rics.bwh.harvard.edu. 1 The abbreviations used are: PD, Parkinson's disease; DA, dopamine; SN, substantia nigra; FPD, familial Parkinson's disease; AR-JP, autosomal recessive juvenile parkinsonism; ACSF, artificial cerebrospinal fluid; RMP, resting membrane potential; AP, action potential; AHP, afterhyperpolarization; PSP, postsynaptic potential; DOPAC, dihydroxyphenylacetic acid; HVA, homovanillic acid; HPLC, high pressure liquid chromatography. screening efficiency as much as 75-fold (24), was also included. The linearized targeting vector was transfected into J1 (129/Sv) ES cells. After selection in G418, 200 clones were screened by Southern analysis for homologous recombination. Six clones were identified by the presence of the expected 3.7-kb band corresponding to the targeted allele. Using the 3Ј external probe and a probe specific for the neo sequence, two clones were confirmed to carry the desired homologous recombination events without random insertion. ES cells of both clones were injected into C57BL/6 and Balb/c blastocysts. Chimeric offspring were crossed with C57BL/6 mice to obtain germline transmission, which was confirmed by Southern analysis with the 5Ј probe shown in Fig. 1A. Heterozygous mice were then interbred to obtain homozygous knockout and wild-type control mice. Mice were subsequently genotyped by PCR using primers (5Ј-CCTACACAGAACTGTGACCTGG; 5Ј-GCAGAATTA-CAGCAGTTACCTGG; 5Ј-ATGTTGCCGTCCTCCTTGAAGTCG) specific for the wild-type or the targeted allele. The resulting 250 and 500 bp PCR products correspond to the wild-type and targeted alleles, respectively. All experimental procedures were carried out in accordance with the USPHS Guide for Care and Use of Laboratory Animals.
Histology and Neuron Count-Mouse brains were dissected, formalin fixed for 2 h, processed for paraffin embedding, and sectioned in the coronal plane at 16-m thickness. Each paraffin block contained 4 parkinϪ/Ϫ and 4 wild-type brains. Deparaffinized sections were stained with cresyl violet or tyrosine hydroxylase (TH) antibodies. The number of DA neurons in the SN was determined by counting THimmunoreactive neurons in coronal sections of four brains per genotype per age group using the fractionator and optical dissector methods of unbiased stereology (27) under a Leica DMRB microscope equipped with a CCD camera connected to a computer running Bioquant image analysis software. The same software was used to measure nigral DA neuron volumes from 4 wild-type and 4 parkinϪ/Ϫ brains at the age of 24 months. The experimenter was blind to the genotypes of the mice. Values are reported as means Ϯ S.E. Statistical differences were assessed by Student's t test.
Striatal DA Measurements-For striatal tissue DA measurement, striata were dissected and stored at Ϫ80°C. Frozen striata were sonicated in ice-cold 0.1 N perchloric acid, 0.2 mM sodium bisulfite and centrifuged 20 min at 20,000 ϫ g at 4°C to remove debris. The supernatant was filtered (0.2 m) and applied to a C18 reverse phase HPLC column linked to an ESA model 5200A electrochemical detector.
For no-net-flux microdialysis, parkinϪ/Ϫ and wild-type mice were implanted unilaterally under halothane anesthesia with a microdialysis probe (CMA 11, 2-mm membrane length, CMA/Microdialysis, Chelmsford, MA) in the striatum using the following stereotaxic coordinates measured from bregma and the skull surface in mm: rostral ϩ0.6, lateral ϩ1.8, ventral ϩ4.5. Probes were perfused with an artificial cerebrospinal fluid (ACSF: 0.2 mM ascorbic acid, 125 mM NaCl, 2.5 mM KCl, 0.9 mM NaH 2 PO 4 , 5 mM Na 2 HPO 4 , 1.2 mM CaCl 2 , 1 mM MgCl 2, pH 7.4) at a flow rate of 0.5 l/min. After 24 h, DA was incorporated in the microdialysis perfusion medium at 5 different concentrations (0, 5, 10, 20, 40 nM) each for 2 h in random order and dialysate was collected in 30 min. intervals into 1.5 l of 12.5 mM perchloric acid/250 M EDTA. Samples were frozen at Ϫ70°C for analysis by HPLC with electrochemical detection (Antec Leyden, Zoeterwoude, The Netherlands) as de-scribed previously (28). The difference in DA concentration between the perfusion medium flowing into the probe and that flowing out of the probe ([DA] in Ϫ [DA] out ) was plotted on the y-axis against [DA] in on the x-axis. A line of best fit was constructed by least-squares analysis. The intercept at the x-axis (DA concentration at the point of no net flux) and the slope of the line (the extraction fraction, a measure of DA reuptake) was determined for each animal and differences between genotypes were assessed by Students' t test.
Dopamine Receptor Binding Assays-D1 and D2 binding assays were performed with [ 3 H]spiperone and [ 3 H]SCH23390, essentially as previously detailed (29) using crude synaptic membrane fractions prepared from mouse striata (prepared as described in Ref. 30) with ketanserin (100 nM) included to inhibit binding to 5-HT2A/2C serotonin receptors. Protein determinations were performed using the BioRad kit with bovine serum albumin as a standard. Binding data were analyzed with Prism (GraphPad) as previously described (31). Data represent mean Ϯ S.E. of 3-6 separate determinations from striata isolated from wildtype and parkinϪ/Ϫ animals. Statistical differences were assessed by Student's t test.
Electrophysiology-Detailed electrophysiological procedures were previously described (32). Briefly, mice were anesthetized with halothane, decapitated, brains were placed in ice-cold oxygenated ACSF (in mM: NaCl 130, NaHCO 3 26, KCl 3, MgCl 2 5, NaH 2 PO 4 1.25, CaCl 2 1, glucose 10 (pH 7.2-7.4)) and coronal corticostriatal sections (ϳ350 m) were made so that striatal neurons could be studied both intracellularly and by activation of their main excitatory input from the cortex. Slices were transferred to a submersion recording chamber in which they were perfused constantly with oxygenated ACSF (31-32°C) in an atmosphere of warm, moist 95% O 2 , 5% CO 2 . Intracellular responses were recorded using sharp microelectrodes (60 -110 MOhms) filled with 3 M potassium acetate, 5 mM KCl, and 2% w/v biocytin to facilitate subsequent examination of the morphology of recorded neurons. Basic passive and active membrane properties (resting membrane potential (RMP), current-voltage relationships, input resistance, action potential (AP) parameters (amplitude and half-amplitude duration), after hyperpolarization (AHP) amplitude) were examined using intracellularly injected current pulses to determine neuronal excitability. Values are reported as means Ϯ S.E. Synaptic responses were evoked with a bipolar stimulating electrode placed in the corpus callosum to activate primarily corticostriatal afferents, the main excitatory pathway into the striatum, which release glutamate. Stimuli of increasing intensity (100-s duration) were delivered every 5 s and 5 traces at each intensity were averaged. Peak amplitudes of postsynaptic potentials (PSPs) were measured, and input-output relationships were plotted. From each cell, the averaged PSP whose peak amplitude was 50% of maximum on the input-output curve was further analyzed for between-group comparisons of peak amplitudes and half-amplitude durations. Paired-pulse facilitation was assessed by presenting two stimuli, which evoked responses at 50% maximum amplitude 50 ms apart and measuring the ratio of the peak amplitude of the second PSP divided by the first PSP. Biocytin labeling confirmed that all recovered neurons were mediumsized spiny neurons with similar appearance between genotypes.
Behavioral Tests-All tests were performed by investigators blind to the genotypes. For the open field test, individual mice were placed in 42 ϫ 42 cm acrylic animal cages for 15 min during which their horizontal and vertical movements were monitored by 3 arrays of 16 infrared light beam sensors (AccuScan Instruments) and analyzed using AccuScan VersaMax software.
For the rotarod test, mice were placed 4 at a time on an Economex accelerating rotarod (Columbus Instruments) equipped with individual timers for each mouse. Mice were initially trained to stay on the rod at a constant rotation speed of 5 rpm. After a 2-min rest, mice that would fall were repeatedly placed back on the rotarod until they were able to stay on the rotating rod for at least 2 min. Following training, mice were subsequently tested by placing them on the rod at a rotation speed of 5 rpm, as the rod accelerated by 0.2 rpm/sec, the latency to fall was measured. Mice were tested for a total of 3 trials.
For the beam traversal task, a Plexiglas beam (Plastics Zone Inc., Woodland Hills, CA) consisting of four sections (25 cm each, 1 m total length) of varying width (3.5, 2.5, 1.5, and 0.5 cm) was used. To increase the difficulty of the test, a wire mesh cover (1 cm 2 ) of corresponding width was placed on the beam surface. Mice were trained for 2 days to traverse the beam without the wire mesh to their home cages. On the day of the test, mice were trained further with two trials without the grid overlay and two trials with the wire grid placed on the beam. Mice were then tested for 3 trials by traversing the grid-surfaced beam, and their performance was videotaped. The numbers of steps and slips (a limb slipped through the wire grid during a forward movement) were counted by viewing the videotapes in slow motion. Fisher's LSD was used for planned comparisons between genotypes.
For the adhesive removal test, small adhesive stimuli of five increasing sizes were placed on the forehead, out of view for the mice. The stimuli consisted of 0.25 and 0.5 inch Avery labels cut in half or whole or combined. To remove the stimulus, mice raised both forelimbs toward their head and swiped off the stimulus with both forepaws within a 60-s trial, after which the experimenter removed the adhesive. Each mouse was given a score equal to the largest size adhesive it was unable to sense and remove, averaged over two trials. Scores were compared between genotypes using a Mann-Whitney U test. All mice were able to sense and remove the largest size adhesive, but none could sense and remove the smallest adhesive.

RESULTS
Generation of Parkin-deficient Mice-parkin is a large gene (ϳ2 Mb), which contains 12 exons and encodes a protein of 465 amino acid residues (5,33). The exon 3 deletion mutation is one of the most common mutations in AR-JP and results in absence of parkin protein (8,18,20). Exon 3 contains a non-integral number of codons, thus, deletion of exon 3 results in a frameshift after amino acid 57 and premature termination at a stop codon in exon 4 following 39 additional out-of-frame amino acid residues in humans (8,18,20). We therefore chose to target exon 3 to generate a parkin-null mutant mouse. A targeting vector was constructed in which most of exon 3 was replaced in-frame by the coding sequence of EGFP, followed by translation and transcription termination sequences and the PGK-neo cassette (Fig. 1A). The protein sequences predicted to result from wild-type and mutant transcripts are depicted in Fig. 1B. Two clones of ES cells carrying the proper homologous recombination events without random integration of the targeting vector were injected into blastocysts. Germline transmission of the targeted allele was confirmed by Southern analysis (Fig.  1C). Interbreeding of heterozygous mice gave rise to wild-type, heterozygous, and homozygous knockout (parkinϪ/Ϫ) mice at the expected Mendelian ratio.
To determine whether our targeted mutation causes skipping of exon 3, we performed Northern and RT-PCR analyses. Northern analysis of total RNA using a probe specific for exons 4 -12 showed a smaller parkin transcript in parkinϪ/Ϫ brains (Fig. 1D). RT-PCR analysis using primers specific for exons 2 and 5 followed by sequencing confirmed that in parkinϪ/Ϫ brains exon 2 was spliced to exon 4, skipping exon 3 entirely (Fig. 1E). Exon 3 skipping causes a reading frameshift after amino acid 57 and premature termination at a stop codon in exon 5 following 49 additional out-of-frame amino acid residues in mice (Fig. 1B). The sensitivity of RT-PCR confirmed the complete absence of intact parkin transcripts in parkinϪ/Ϫ mice. Sequencing also revealed an aberrant splice product, which results from the use of a cryptic splice acceptor site 3 bases into exon 4, leading to addition of 48 rather than 49 out-of-frame amino acid residues (Fig. 1B). Although these truncated parkin transcripts are present in parkinϪ/Ϫ mice, it is unlikely that functional parkin fragments can be produced from these truncated transcripts. Western analysis using an antiserum raised against the C-terminal region of parkin confirmed the absence of parkin in parkinϪ/Ϫ mice (Fig. 1F), and ruled out the presence of possible parkin fragments initiated from in-frame ATGs downstream of exon 3, consistent with the notion that reinitiation of translation following a sizable open reading frame is highly unlikely (34).
Since we introduced the EGFP cDNA fused in-frame into parkin exon 3, which was intended for a reporter system for the parkin promoter activity, we also performed Northern analysis using an EGFP-specific probe and confirmed the presence of EGFP transcripts in parkinϪ/Ϫ mice (data not shown). RT-PCR followed by sequencing confirmed that the EGFP coding sequence is intact and fused in-frame to parkin exon 3. How-ever, the parkin-EGFP fusion protein was barely detectable by Western analysis (data not shown), perhaps due to the presence of the ubiquitin-like domain of parkin (35).
parkinϪ/Ϫ mice are viable and fertile without obvious abnormalities. Open field tests of parkinϪ/Ϫ mice revealed no significant alterations in their general behavior and exploratory anxiety (Fig. 2). Nissl staining revealed normal brain morphology in parkinϪ/Ϫ mice (Fig. 3, A and B). Immunohis- representation of the structure of the expected protein products derived from the wild-type and the targeted alleles. Amino acid residue numbers are shown above relevant domains (shaded gray) or frame-shifted sequences (hatched). Numbers in boxes depict the exons coding for each region. C, Southern analysis of tail DNA digested with EcoRV and hybridized with the 5Ј probe. D, Northern analysis of total RNA hybridized with a probe specific for exons 4 -12 shows a smaller mRNA species in parkinϪ/Ϫ mice. E, RT-PCR analysis of parkin transcripts using primers specific for exons 2 and 5. F, Western analysis of total brain homogenates confirms the reduction and absence of parkin in heterozygous and homozygous mutant mice, respectively. The same blot was incubated with a ␤-tubulin antibody to normalize loading. tochemical analysis of parkinϪ/Ϫ brains using antibodies specific for synaptophysin, Munc-18 and calbindin showed grossly normal synaptic staining and striatum formation (data not shown). No inclusions were observed in any brain sub-regions, including the SN, using antibodies specific for ␣-synuclein and ubiquitin (data not shown).
Normal Neuroanatomy of DA Neurons in parkinϪ/Ϫ Mice-It has been proposed that the reduction in DA neurons in AR-JP could be due to impaired generation and maturation of these neurons, based on the observation that the remaining DA neurons in autopsies of AR-JP patients appear immature (18). Recently, parkin-null flies have also been shown to exhibit smaller dopaminergic neurons (36). Immunohistochemical analysis of parkinϪ/Ϫ mice using an antibody specific for TH, however, revealed normal morphology of DA neurons in the SN (Fig. 3, C-F) and noradrenaline neurons in the locus ceruleus (Fig. 3, G and H).
Increased Extracellular DA in the Striatum of parkinϪ/Ϫ Mice-We then looked for alterations in dopamine neurotransmission, which may occur prior to frank loss of dopaminergic neurons. Dopaminergic projections revealed by TH immunoreactivity appear normal in the striatum of parkinϪ/Ϫ mice (Fig.  3, I and J). Striatal levels of DA and its major metabolites, dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA) are similar between parkinϪ/Ϫ and control mice at 6, 12, 18, and 24 months (p Ͼ 0.05) (Fig. 4A). To determine whether there are alterations in dopamine release or reuptake, we performed no-net-flux microdialysis (37,38) in the striatum of parkinϪ/Ϫ and wild-type mice at 8 -9 months. When perfused with ACSF containing no DA, the dialysate DA concentration in parkinϪ/Ϫ mice (n ϭ 10) (10.8 Ϯ 1.2 nM) was significantly higher than that in wild-type mice (n ϭ 10) (7.7 Ϯ 0.4 nM; p Ͻ 0.02) (Fig. 4B).
The concentration of DA at the interpolated point of no-netflux, a measure of extracellular DA concentration, was also significantly higher in parkinϪ/Ϫ mice (n ϭ 10) (24.2 Ϯ 1.2 nM) relative to wild-type controls (n ϭ 9) (20.2 Ϯ 0.9 nM; p Ͻ 0.02) (Fig. 4, C and D). The data from one wild-type mouse (indicated by an arrow in Fig. 4C), which met the statistical standard as an outlier (39), was excluded. The extraction fraction, a measure of DA reuptake, was not significantly different between wild-type (0.39 Ϯ 0.04) and parkinϪ/Ϫ (0.46 Ϯ 0.05; p Ͼ 0.05) mice (Fig. 4E). The elevated level of extracellular DA in par-kinϪ/Ϫ mice, therefore, likely results from increased release of dopamine.
Unchanged Levels and Binding Affinities of D1 and D2 Receptors in the parkinϪ/Ϫ Striatum-We then examined whether the increase in extracellular DA results in an alteration of striatal DA receptors. Similar total receptor binding was found for both D1 and D2 receptors in the striatum of Loss of parkin Decreases Synaptic Excitability of Striatal Neurons-The significant increase in extracellular DA in the striatum of parkinϪ/Ϫ mice prompted us to examine the electrophysiological responses of striatal neurons, as previous studies have demonstrated that the excitability of these neurons is strongly influenced by DA levels (40,41). We focused our analysis on the medium-sized spiny neurons, which are the major neuronal subtype in the striatum. Major inputs received by these neurons include glutamatergic excitatory inputs from the cortex and dopaminergic inputs from the SN. Acute coronal corticostriatal slices were prepared from parkinϪ/Ϫ mice (n ϭ 7) and wild-type controls (n ϭ 6) at 6 -9 months of age and striatal neurons (Ϫ/Ϫ: n ϭ 13, ϩ/ϩ: n ϭ 10) were examined by injecting current pulses intracellularly into each neuron (Fig. 5A). No significant differences in passive and active membrane properties were observed between the genotypes and all values were in the range of those reported in the literature for medium-sized spiny neurons (RMP: Ϫ78 Ϯ 3 versus Ϫ80 Ϯ 2 mV; input resistance: 29 Ϯ 5 versus 37 Ϯ 5 MOhms; AP amplitude: 66 Ϯ 3 versus 68 Ϯ 3 mV; AP width at 1/2 amplitude: 0.76 Ϯ 0.03 versus 0.75 Ϯ 0.07 ms; AHP amplitude: 11.7 Ϯ 1.4 versus 12.1 Ϯ 1.0 mV in ϩ/ϩ versus Ϫ/Ϫ neurons, respectively) (Fig. 5A).
Behavioral Impairments in parkinϪ/Ϫ Mice-We further tested parkinϪ/Ϫ mice for possible alterations in motor activities using a beam traversal task, which is sensitive to impairment in the nigrostriatal pathway (42)(43)(44)(45). Planned comparisons revealed that parkinϪ/Ϫ mice performed significantly worse than wild-type controls with higher numbers of slips and slips per step at all age groups (p Ͻ 0.05), while the number of steps was similar between the genotypic groups (Fig. 6, A-C). We further evaluated these mice in the rotarod task (46), which is commonly used to score severe motor impairments in rodents. parkinϪ/Ϫ and control mice exhibited similar latencies for remaining on the rotating rod (Fig. 6D).
Another test sensitive to nigrostriatal dysfunction in rodent models is the adhesive removal test, which assesses somatosensory abilities (47-49). parkinϪ/Ϫ mice at 2-4 and 7 months performed significantly worse compared with wild-type controls (p Ͻ 0.05), though the difference at 18 months was not as marked, likely due to the age-dependent decline of wild-type mice (Fig. 6E). Only 63% of wild-type mice at 18 months removed the middle size stimulus, compared with 100% at 4 months. This result is consistent with a previous report showing poorer performance by older animals (50).
Unchanged Levels of parkin Substrates-Recent in vitro studies have shown that parkin functions as an E3 ubiquitin ligase (21)(22)(23)26,(51)(52)(53)(54). Several proteins, such as CDCrel-1, synphilin-1, and glycosylated ␣-synuclein, have been shown to be ubiquitinylated by parkin in vitro. We therefore examined whether loss of parkin function results in accumulation of these proteins in parkinϪ/Ϫ mice. Western analysis showed similar levels of CDCrel-1, synphilin-1, and ␣-synuclein in parkinϪ/Ϫ and wild-type control brains (Fig. 7, A-C). Immunoprecipitation of ␣-synuclein followed by Western blotting using a second ␣-synuclein-specific antibody revealed the presence of an additional low abundance isoform at ϳ22 kDa, consistent with the apparent molecular weight of the glycosylated isoform of ␣-synuclein previously reported as a parkin substrate (26). No differences were observed in the steady state level of this higher molecular weight species between the genotypes (Fig.  7D). These results indicate that loss of parkin function does not affect the stead state level of any of these proteins in the mouse brain.

DISCUSSION
Despite a large body of evidence linking loss-of-function mutations of parkin to FPD, the normal physiological role of parkin and the mechanism by which parkin mutations cause nigral degeneration are unknown. Here we present three independent lines of evidence supporting a novel role of parkin in the nigrostriatal pathway. First, the extracellular DA concentration is significantly increased in the striatum of par-kinϪ/Ϫ mice, which is likely due to an increase in dopamine release from nigral neurons. Second, the synaptic excitability of medium-sized spiny striatal neurons is reduced, as evidenced by increased currents required to evoke synaptic responses and action potentials. Lastly, parkinϪ/Ϫ mice exhibit deficits in behavioral tasks previously shown to be sensitive to nigrostriatal dysfunction (42-45, 47-49, 55, 56), although we cannot exclude the possibility that alterations in other neural circuits may also contribute to the observed behavioral deficits.
Together, these results indicate that parkin loss-of-function mutations cause defects in both DA release from nigral neurons and synaptic excitability of medium-sized spiny striatal neurons, which are the major target of nigral dopaminergic projections. Previous studies have shown that DA decreases gluta- FIG. 5. Electrophysiological deficits in parkin؊/؊ mice. A, current-voltage relationship of medium-sized striatal neurons recorded in acute slices, which was generated by injecting current pulses into each neuron and recording the change in induced voltages. No significant differences are found in current-voltage relationships or average passive and active membrane properties between parkinϪ/Ϫ and control neurons. Numbers to the left are RMPs for the two cells shown. B, depolarizing synaptic responses evoked by stimulation of corticostriatal afferents (arrows indicate the time of stimulation). Traces show no differences between the genotypes in superimposed representative response curves at ϳ20, 50, and 80% of the maximum amplitude (each curve is the average of at least 5 traces). C, normalized input-output relationship. Filled circles represent the intensity of the stimulation current (mean Ϯ S.E.) needed to evoke responses at 20, 50, and 80% of the maximum postsynaptic potential in parkinϪ/Ϫ (n ϭ 13) and wild-type (n ϭ 10) neurons. Group differences at 50% (p Ͻ 0.03) and 80% (p Ͻ 0.003) were statistically significant (analysis of variance followed by post-hoc Student's t tests).
The cardinal neuropathological feature of PD, profound loss of dopaminergic neurons in the SN, is absent in parkinϪ/Ϫ mice. Thus, the mechanism underlying the mild motor deficits observed in parkinϪ/Ϫ mice may differ from the movement disorder characteristic of PD, which results primarily from substantial loss of DA neurons. Our results, however, are consistent with other mutant mice that exhibit decreased performance in beam traversal and increased concentrations of striatal DA (43). Mouse models of neurodegenerative diseases have previously been observed to recapitulate some aspects of the disease in the absence of substantial neuronal loss in the affected brain subregions. Transgenic mice overexpressing wildtype and FPD-linked mutant human ␣-synuclein exhibit motor deficits in the absence of loss of DA neurons (59 -63). Similarly, genetically engineered mouse models of Alzheimer's (64) and Huntington's diseases (65) have successfully reproduced behavioral and neuropathological aspects of these diseases largely without recapitulating the cortical and striatal neuronal loss, respectively.
The fact that loss of parkin function alone in mice is insufficient to cause substantial loss of dopamine neurons suggests that other factors also contribute to nigral degeneration. One possibility is that parkin protects neurons from various insults, which has been observed by overexpressing parkin in transfected cells (66,67) and transgenic flies (68). However, no decrease in DA neuron survival has been observed in Drosophila in which endogenous parkin has been inactivated (36) or partially inactivated by RNAi (68), consistent with our findings in parkinϪ/Ϫ mice. Together, these findings suggest that additional events may be required for neuronal degeneration. This notion is also consistent with the fact that the age of onset for FPD patients carrying parkin mutations ranges from juvenile to elderly and varies by as much as 20 years even within single families (9). Other possibilities, including higher tolerance to neuronal degeneration in the mouse DA neuron, shorter life span of mice, and the well-controlled environment in which these mice are housed, may also contribute to the absence of the profound nigral degeneration characteristic of PD brains.
In summary, our findings provide important insights into the normal physiological role of parkin in dopamine regulation and nigrostriatal function, which will facilitate the exploration of the mechanisms of PD pathogenesis and may assist the identification of novel preventative and therapeutic strategies. , and ␣-synuclein (C) shows similar steady-state levels of these proteins in brain homogenates from par-kinϪ/Ϫ and wild-type mice. Immunoprecipitation of ␣-synuclein, followed by Western blotting reveals equivalent levels of a higher molecular size species at ϳ22 kDa that is immunoreactive for ␣-synuclein (D).