Introduction

Parkinson's disease (PD) is the second most common neurodegenerative disorder and is characterized by the loss of dopaminergic cells in the substantia nigra pars compacta and the presence of Lewy bodies. Alpha-synuclein, parkin, UCH-L1 (Ubiquitin-C terminal hydrolase-L1), DJ-1, PINK1, and LRRK2 (leucine-rich repeat kinase 2 mutations have been found in familial forms and/or early onset PD [1–8]. Polymorphic variants of genes involved in xenobiotism or gene-environmental interaction may also modulate the risk of PD [9–11].

The parkin gene has 12 exons, encoding 465 amino acids, spanning 1.3 Mb of genomic DNA. Parkin functions as a E3-ubiquitin ligase, with a UBL domain N′ terminus and a two-RING finger domains [7, 8]. Mutations in the parkin gene is the most common cause of familial and early onset PD, accounting for 50% of the familial cases and more than 70% in those sporadic cases with age of onset <20 years [12]. It has been thought that parkin-linked PD is recessively inherited, where a deleterious alteration is presumed on both alleles and heterozygous carriers are unaffected. However, some studies suggest that more than half of the total number of cases identified with a parkin mutation had only a single heterozygous mutation [13, 14]. Patients having a loss of one parkin allele may suffer from haploinsufficiency as a consequence of a reduced expression and enzymatic activity, resulting in a greater risk for disease [15, 16]. Despite the heterogeneity of parkin mutations (missense, truncating, duplication, exonic rearrangement, etc.), there is a lack of genotype–phenotype correlation, suggesting presence of other genetic or epigenetic modulation. The role of parkin in sporadic PD has not been clarified, although variability of the promoter region and, consequently, altered gene expression has been implicated [16].

Alternative RNA splicing may be used as a mechanism of gene regulation [17]. The primary transcript can be processed in different ways by alternative usage of exons. Altered regulation and splicing of genes either under oxidative stress or due to mutations in the splicing sites could lead to significant changes in gene expression and altered protein activity [17]. There is currently little information on the prevalence, type, and significance of splicing of the parkin gene in the literature. A parkin exons 3-4-5 (447 bp; 149 AA) splice variant (E3-4-5SV) was found in human leukocytes [18], and a parkin exon 5 (84 bp; 28 AA) splice variant (E5SV) was found in brain [4]. The reading frame was maintained over the junction of exons 2–6 or exons 4–6. The translated protein would have the two-RING functional domain.

In this study, we identified a parkin splice variant in the substantia nigra and leukocytes of sporadic PD patients, and using a case control methodology, investigated the exon 4 SV (E4SV)/wild-type parkin expression in the leukocytes of sporadic PD patients and healthy individuals.

Materials and methods

Subjects

The PD patients and healthy controls were recruited from our movement disorder clinics and health screening unit as previously described [19]. All subjects were examined by the authors and gave informed consent for their blood samples for genetic research. The work received approval from the ethics committee of our institution. There were 80 PD and 80 controls with mean age of 56.0±11.0 and 53.0±8.0 years comprising 60 and 55% men, respectively. All study subjects were ethnic Chinese.

Extraction of total RNA and DNase treatment

Total RNA from human peripheral leukocytes was isolated with erythrocyte lysis buffer (Qiagen) and RNAgents total RNA isolation system (Promega). Total RNA from HEK 293 and neuroblastoma SH-SY5Y cells was isolated directly with RNAgents total RNA isolation system (Promega). Total RNA from human substantia nigra was purchased from Ambion. To remove DNA contamination from the total RNA preparation, DNase digestion was performed following the manual instructions (Invitrogen).

RT-PCR analysis

For cDNA synthesis, 500 ng of total RNA was reverse transcribed using oligo-dT primers following the manual instructions of the “First Strand cDNA Synthesis Kit” (Invitrogen).

Primers at exons 2 and 5 (forward: 5′-gaggaatgactggactgtgc-3′; reverse: 5′-ctagtcccagggcagtgtg-3′) were used for amplification of parkin normal fragment and parkin splice variant. A 20-μl polymerase chain reaction (PCR) contained 0.2 μl cDNA, 1 μl DMSO, 2 μl 10× PCR buffer (Qiagen), 1.5 mM magnesium chloride, 0.25 μM of each primer, 0.4 μM dNTP mix, and 0.5 U HotstarTaq DNA polymerase (Qiagen). Thermal cycler parameters were 35 cycles at 95°C for 30 s, 58°C for 1 min, and 72°C for 1 min, with preheat at 95°C for 15 min and last extension at 72°C for 5 min. Electrophoretical separation of the reverse transcriptase (RT) PCR fragments was performed on ethidium bromide-stained 2% agarose gels running at 100 mA for 40 min.

Nucleotide sequence analysis of all 12 exons of parkin including the intron–exon boundary regions

The 12 exons of parkin including the intron–exon boundary region were amplified following the conditions previously described [4]. The PCR products were visualized on ethidium bromide-stained 2% agarose gel. The PCR products were subjected to direct nucleotide sequence analysis using automated DNA sequencers (Applied Biosystems).

Gene dosage analysis of parkin exons 3 and 4

Gene dosage analysis was performed using the ABI Prism 7700 sequence detections systems (Applied Biosystems). β-globin was used as internal control, and the primers and TaqMan probe were purchased from Applied Biosystems. Primers and TaqMan probe for exon 3 (forward primer: 5′-GGATCAGCAGAGCATTGTTCAC-3′, TaqMan probe: 5′-TTGTGCAGAGACCGTGGAGAAAAGGT-3′, reverse primer: 5′-GCCTCCAGTTGCATTCATTTCT-3′) and primers and Taq Man probe for exon 4 (forward primer: 5′-CAAAGGCCCCTGTCAAAGA-3′, TaqMan probe: 5′-CTCAGGGTACAGTGCAGCACCTGC-3′, reverse primer: 5′-GGTCAAGGTGAGCGTTGC-3′) were synthesized according to previous published literature [20]. The cycle number at which each PCR reached a significant threshold (CT) during the log phase of the amplification was used as a quantitative measure of the target product. The value calculated from CT for exons 3 and 4 was normalized using those for the β-globin gene.

Quantitative PCR using real-time TaqMan–PCR methods

Quantitative PCR was performed using the real-time TaqMan–PCR method. In this method, AmpliTaqMan DNA polymerase extends the primer and displaces the TaqMan probe through its 5′- to 3′- exonuclease activity. The probes were labeled with a reporter fluorescent dye [6-carboxy-fluorescein dye (FAM)] at the 5′ end and a quencher fluorescent dye (TAMARA) at the 3′ end. The expression level of β-actin gene was used as an endogenous control, and a TaqMan probe for β-actin was procured from Applied Biosystems.

Primers and TaqMan probes for real-time PCR

The primers and TaqMan probes were designed according to published sequence of the parkin mRNA(AB009973). Specific primers were used to amplify the wild-type (forward: 5′-cactgacagcaggaaggactcac-3′ and reverse: 5′-ctttgcaatacacataaaagctgttgta-3′) and splice variant (forward: 5′-cactgacagcaggaaggactcac-3′ and reverse: 5′-ccgtttggaattaaaacatcatc-3′) alleles. TaqMan probes were designed at the junction of exons 3–4 (wild-type parkin: 5′-tctacctgctggacttc-3′) or exons 3–5 (E4SV: 5′-atggaccctggacttc-3′). The reaction mixture contained 25×TaqMan mixture (Applied Biosystems), the appropriate primer set, the TaqMan probe, and the cDNA template in a total volume of 50 μl. The thermal cycling protocol consisted of a minute at 95°C and 10 min at 95°C, followed by 45 cycles at 95°C for 15 s and 60°C for 1 min. TaqMan–PCR was analyzed using a Model 7700 Sequence Detector (Applied Biosystem). The quantity of the target mRNA was calculated using the Sequence Detector software package version 1.6.3 (Applied Biosystem). The starting quantity of each target mRNA was normalized to that of the β-actin endogenous control mRNA, allowing for measurement of the relative expression level of each RNA species with the comparative CT method (ABI user's manual). For control TaqMan studies, we prepared constructs for both wild-type and splice variant alleles in the plasmid vector DNA3.1 (Qiagen). The specificity of the primers and TaqMan probes for the wild-type and splice variant alleles was confirmed using three templates: the wild-type allele only, a mixture of wild-type and splice variant alleles at a 1:1 ratio, and the splice variant allele only.

Results

Identification of a parkin splice variant from the human peripheral leukocytes and brain by RT-PCR

During the screening of parkin gene mutations in our patients, we identified a parkin SV, E4SV. RT-PCR amplified a major 470 bp and a minor fragment around 340 bp from the leukocytes and substantia nigra of PD patients with primers in exons 2 and 5 (Fig. 1). DNA sequence analysis confirmed that the former (470 bp) corresponded to exons 3–5 and the latter (340 bp) contained the exon 4-spliced variant transcript. A 122-bp deletion corresponding to exon 4 (nucleotide, 514–635 bp; residues, 139–178 AA) resulted in exon 3-derived sequence being directly connected to exon 5-derived sequence (Fig. 2). The reading frame was shifted over the junction and a stop codon (tga) appeared just 17 bp downstream from exon 3. A similar splicing was also found from leukocytes in the controls without neurological disease, suggesting that this is a common splice variant.

Fig. 1
figure 1

Abnormal splicing of parkin: cDNA derived from the human leukocytes (normal control, lane 1; PD, lane 3) and substantia nigra (normal control, lane 2; PD, lane 4) was amplified with primers in exons 2 and 5. Six microliters of product underwent electrophoresis on a 1.5% agarose gel. Lane M is a 100-bp DNA ladder (MBI Fermantas). A duplex banding pattern is observed in lanes 1, 2, 3, and 4. The normal size is about 470 bp, the size of parkin spliced variant is about 340 bp. The PCR products from cDNA clones of wild-type parkin (lane 5) and E4SV (lane 6) were used as positive control. Lane 7 is a negative control without cDNA template

Full size image
Fig. 2
figure 2

Sequence analysis of the RT-PCR products. The 470-bp fragment corresponds to exons 3–5 (a). The 340-bp fragment corresponds to exons 3–5 with exon-4 (122 bp) deletion (b)

Full size image

Quantitative PCR using real-time TaqMan–PCR method

The expression of wild-type parkin and parkin SV was discriminated using the protocol described in Materials and methods.

Expression of wild-type parkin and parkin SV in leukocytes from the controls and PD patients

Expression analysis of the relative levels of parkin SV and wild-type parkin was performed using TaqMan real-time–PCR method. Total RNA samples were extracted from leukocytes of 80 PD patients and 80 controls. All 80 PD patients were previously screened negative for parkin gene mutations based on sequence analysis (including exon 4 and its exon–intron boundaries). Although E4SV is a common splice variant, E4SV expression may be influenced by a small likelihood of gene dosage multiplication in exons 3 and 4. However, no abnormalities on gene dosage studies of exons 3 and 4 were found in our study subjects.

The results demonstrated that expression level of wild-type parkin was lower in PD patients (p<0.05), whereas the expression level of E4SV increased modestly (Fig. 3a,b). These changes in the expression levels of wild-type parkin and E4SV in PD patients led to a significantly increased ratio of E4SV to wild-type parkin expression (p<0.0005) (Fig. 3c). To examine for confounding or predictive factors of the observed difference, a logistic regression analysis of the following variables: age, gender, disease group (PD or control), interaction of age and disease group, and interaction of gender with disease group, was carried out with ratio of parkin SV to wild-type parkin expression as the outcome measure. None of the variables confounded our findings except the interaction of age with the disease group. This meant that for any given age, there was a significant difference in the ratio of E4SV/wild-type parkin between PD and controls (p<0.0005). The ratio of E4SV/wild-type parkin increased with age in PD patients, but this was not observed in the controls (Fig. 4).

Fig. 3
figure 3

Real-time PCR assays of parkin mRNA levels in the human leukocytes from PD patients and normal controls. Expression of wild-type parkin (a) and E4SV (b), p<0.05; ANOVA. Ratio of E4SV/wild-type parkin expression higher in PD, p<0.0005; ANOVA (c). Raw data were normalized to beta-actin. Bars correspond to mean±SEM

Full size image
Fig. 4
figure 4

The ratio of E4SV vs. wild-type parkin expression increased with age in PD patients (p<0.0005)

Full size image

Validation of mRNA analysis

Following our previous protocol [21], the gene expression analysis was repeated for a random sample of 36 PD and 36 controls over a 2-month interval. The median fold expression of wild-type parkin for PD (original vs duplicate) was 0.2438 vs 0.2121 (p>0.05), and for controls 0.3527 vs 0.3184 (p>0.05). The median fold expression of E4SV for PD (original vs duplicate) was 0.0237 vs 0.0205 (p>0.05), and for controls 0.0183 vs 0.0178 (p>0.05). The median fold expression of E4SV vs wild-type parkin for PD (original vs duplicate) was 0.0612 vs 0.0598 (p>0.05), and for controls 0.0383 vs 0.0326 (p>0.05).

Discussion

There is little published information on the prevalence, type, and role of parkin splicing in PD. One parkin splice variant lacking 84 bases was found during the cloning of human parkin [4] This variant corresponds to exon 5 encoding amino acid residues 179–206 that reside within the unique segment. It is postulated that this splicing could alter parkin function as deletion of exon 5 in the parkin gene has been found in autosomal recessive PD [22]. A parkin SV involving splicing of exons 3–5 was found in leukocytes, but not in brain tissue [18]. Other workers have also isolated a novel parkin SV in mouse that lacks the C-terminal RING box with a predicted molecular mass of 28 kD [23]. Using antibodies directed against Parkin UBL domain, a predominantly 50 kD parkin in mouse tissues and another product similar to the predicted size of the mouse parkin variant described by Kitada [23] has been isolated [24]. This 22-kD N-terminal parkin fragment, containing the UBL domain, has some semblance to the fragment released by caspase cleavage [25]. More recently, other splice variants have also been found in developing mouse and humans [26, 27].

In this study, we identified a parkin E4SV in the SNc and leukocytes in sporadic patients and healthy controls. The E4SV appears to be ubiquitous inasmuch as it could be found in both brain and leukocytes. There was an increase of E4SV and a significant reduction in wild-type parkin expression, with a net result of increased E4SV/wild-type expression ratio in PD patients compared to healthy controls. The E4SV we have identified corresponds theoretically to a genomic heterozygous exon 4 deletion, leading to a truncated protein devoid of the RING-functional domains, hence with total loss of enzymatic activity. Homozygous or heterozygous exon 4 deletions have been described in PD patients, and the heterozygous exon 4 deletion was associated with the disease [12, 15]. At any given age, there was a significant difference in the ratio of E4SV/wild-type parkin between PD and controls. The ratio of E4SV/wild type increased with age in PD patients, but this was not observed in the controls. This is interesting in light of the increased risk of PD with age, and the expression of E4SV/wild type may predispose to PD in certain at-risk individuals. It has been demonstrated that there is a shift in parkin solubility with aging, suggesting an age-dependent modification of parkin or its substrates [28]. It is possible that E4SV may also be a factor in these interactions. The relative amounts of the E4SV and wild type parkin could modulate the steady-state dynamics of parkin substrates and proteasomal functioning. Alternatively, the relative expression of E4SV/wild type may act via the mechanism of haploinsufficiency in risk association. This may provide an explanation why some parkin mutations (e.g., a 40-bp deletion in exon 3) manifest a varied mode of inheritance (both recessive and dominant) in familial cases [29, 30].

Although we demonstrated different E4SV/WT parkin expression in PD compared to controls, the exact mechanism governing skipping of exon 4 has not been clarified. We sequenced the exon–intron boundaries (between 36 to 50 bp) in our PD patients but did not find any mutations, particularly in the 5′ splice site. It is possible that the altered splice site choice is enriched in PD and with advancing age, but whether this is modulated by specific splicing factors is currently unclear. The stop codon may cause nonsense decay of the mRNA transcript. However, as all the PD patients have been treated with levodopa and other dopaminergic agents, we could not conduct a subset analysis of the effect of drugs on the potential nonsense decay.

In conclusion, we identified a parkin E4SV in the substantia nigra and leukocytes of sporadic PD and controls. The relative expression of E4SV/wild type parkin was increased in sporadic PD patients compared to controls. Based on our observations, further functional studies to determine the pathophysiologic role of E4SV in sporadic PD patients are warranted.