Octarepeat changes of prion protein in Parkinson's disease

1. Introduction 

Parkinson's disease (PD), with dopaminergic neuron loss in substantia nigra, has been regarded as the second most common neurodegenerative disease probably attributed to both environmental and genetic etiologies. Clinically the symptoms of PD consists of tremor, bradykinesia, rigidity, and postural instability, mostly with asymmetric onset. Some familial PD cases have mutations in α-synuclein, parkin, UCH-L1, LRRK2, or PINK1 genes [1]. Suspicious genes have been evaluated in patients with sporadic PD, for example, the beta-glucocerebrosidase [2]. In addition to the pathological findings of dopaminergic neuronal loss and cytoplasmic Lewy body, there are evidences of microglial activation, oxidative stress and inflammation in the dopaminergic neurons of substantia nigra in PD [3], [4], [5], and protection of midbrain cells by indomethacin and naloxone from the lipopolysaccharide-induced lesion [6].

Inflammatory microglial activation was also seen in another neurodegenerative disease, Creutzfeldt–Jakob disease (CJD), which is one of the transmissible spongiform encephalopathies (TSE) [7]. Some parkinsonian features in addition to myoclonus and ataxia may occur in CJD. The causative agent has been attributed to prion protein (PrP), an inducible protein [8], which involves diverse cell functions, including neural synaptic transmission, neuroplasticity, signal transduction, cell survival, anti-oxidation and superoxide dismutase activity, and even infectious disease. It contains 253 amino acids, encoded by the PRNP gene on the chromosome 20. Some familial TSE cases are related to mutations of PRNP gene, and polymorphisms of PRNP gene have also been reported to be associated with several human diseases, such as Met129Met with the new variant CJD [9] and early-onset memory disorder [10], 178Asn and 129Val with CJD, 178Asn and 129Met with familial fatal insomnia [11], and Asn171Ser with the surgical outcome of mesial temporal seizure [12]. Another domain of PrP, octarepeat region of PrP through the codon 51–91, abbreviated as R1–R2–R2–R3–R4 on the amino acid sequences, is related to the copper ion binding [13] and the anti-oxidative effect [14]. There were reports about octarepeat insertion associated with one genetic PrP disease [15], but few reports about the disease-associated deletion variants. Conformational change of PrP might be an immediate result from insertion or deletion in PrP, no matter whether gain or loss of function occurred.

On searching the new genetic factors in PD, the objective of the present study is to explore whether PrP polymorphism may play a role in the pathogenesis of PD, with evidences of some clinical features and pathological presence of microglial activation in both of PD and CJD, especially the codon 129 and the octarepeat region. This report not only gives information about the genetic distribution of the codon 129 of PrP among the parkinsonian subjects in Taiwan, but also is the first one about octarepeat changes of PrP among those with PD. A hypothesis about the possibly disorder of RNA synthesis from polymorphic DNA strands is proposed, and may be regarded as one of the pathomechanisms in neurodegeneration, such as in PD.

3. Results 

The DNA samples from our subjects were subjected to PCR and re-confirmatory sequencing of all the products after restriction digestion, which revealed only one Val/Val homozygote at codon 129 in the control group. Met/Val heterozygotes were found as 9 among PD subjects (2.7%) and 6 among controls (3.0%). The Val-harboring rate was 1.4% among PD subjects and 2.0% in controls. The rest of PD patients and controls were all Met/Met homozygotes at codon 129. This genetic polymorphic distribution of Met and Val allele disclosed no difference between the PD and the control groups (χ2=0.65, df=1, p=0.42).

All the control subjects exhibited the homozygotic R1–2–2–3–4/R1–2–2–3–4 pattern on octarepeat analysis. But four PD subjects had changes in the octarepeat region. Their brief histories were listed in Table 1. Subject 9361 had an unusual initial presentation of orobuccolingual dyskinesia but later developed rigidity and bradykinesia, with clinical diagnosis of PD supported by reduced and asymmetric putaminal uptake on 99mTc-TRODAT-1 SPECT. One patient had the nucleotide substitution (NS) heterozygosity (R1–2–3g–3–4/R1–2–2–3–4, R3g sequence: CCC CAT GGT GGT GGC TGG GGg CAG) in the second R2 copy (Fig. 1A). This silent change encodes proline (CCC or CCT) and glycine (GGG or GGA) on protein synthesis. We also found three heterozygous PD subjects with a R2 repeat deletion (CD pattern) in one allele, resulting in R1–2–3–4/R1–2–2–3–4 pattern (Fig. 1B,C). The same results were obtained repeatedly in triplicate. The shorter and longer bands detected on electrophoresis were further eluted from the agarose gel and the sequential sequencing confirmed the CD phenomenon. All these four patients with octarepeat changes exhibit Met/Met allele at codon 129. No other octarepeat changes occur among our subjects, such as copy insertion, other copy deletion or NS variants (e.g. R2a or R2c).

Table 1. The brief histories of the PD subjects with octarepeat changes of PrP

	Patient	Sex	Age at onset	Onset region	Presenting symptom	Octarepeat genotype
	9361	M	77	Orobuccal	Dyskinesia	R1–2–3–4/R1–2–2–3–4
	9805	M	70	Left limb	Tremor	R1–2–3–4/R1–2–2–3–4
	9917	F	50	Left limb	Rigidity	R1–2–3–4/R1–2–2–3–4
	0051	M	76	Left limb	Tremor	R1–2–3g–3–4/R1–2–2–3–4

View Large Image Download to PowerPoint

Fig. 1 The octarepeat changes of PrP in the PD subjects. (A) The genomic sequencing of the PD patient with R1,2,2,3,4/R1,2,3g,3,4 heterozygosity. R2 sequence: CCT CAT GGT GGT GGC TGG GGG CAG, and R3g sequence: CCC CAT GGT GGT GGC TGG GGg CAG. (B) The electrophoretic finding of the PCR product of the heterozygote PD subject with octarepeat CD. (C) The sequencing of the shorter gel fragment revealed CD and presented as R1–2–3–4.

To examine the expression of transcripts in those with octarepeat changes, we performed cDNA analysis by recruiting five subjects, two controls and three PD individuals. The individual genomic DNA patterns of the three PD subjects were NS heterozygosity of R1–2–3g–3–4/R1–2–2–3–4, CD heterozygosity of the single R2, and wild-type homozygosity without octarepeat changes. All these three RT-PCR products had single band on electrophoresis. The cDNA of the two controls and the PD subject with wild-type genome were homozygotes of R1–2–2–3–4/R1–2–2–3–4. The cDNA sequencing disclosed consistent R1–2–3g–3–4/R1–2–2–3–4 in the heterozygote with genomic NS but discordant in the one with genomic CD (Fig. 2). The PCR results of the genomic DNA and cDNA were conducted in triplicate and duplicate, respectively. Besides, we also used the FailSafeTM PCR PreMix Selection Kit (Biotechnologies) to amplify this GC-rich region, as the already known fact that GC-rich template may form secondary structure leading to truncated products. The sequences of the products were the same as the prior studies.

View Large Image Download to PowerPoint

Fig. 2 The PrP cDNA sequence of the PD subject with genomic CD heterozygosity showed the R4–3–3g–2–1 pattern (antisense reading) rather than R4–3–2–2–1.

4. Discussion 

Polymorphic distribution of PrP may differ in populations, and the most known polymorphic locus is the codon 129. There was a lower Val-carrying rate in Asian population but, in Europe, the Val/Val allele was found in about 7–9% population and Met/Val allele in 44–50% population [16]. In this study, we found only one Val/Val homozygote who is a control subject. The main population in this study is Met/Met homozygotes in both groups, while the frequency of Met/Val allele is 2.7% in the PD subjects and 3.0% in controls. It presented a result of Met129Val not associated with PD, as the report in the districts outside Asia [17]. The polymorphism at codon 129 of PrP was not regarded as a susceptible locus for Alzheimer's disease [18], but we surveyed the genotypes of codon 129 in our subjects due to the prior report about the association with hypokinetic-rigid parkinsonism [19]. Val allele has been reported to have protective effect from neurodegeneration, e.g., the Val/Val genotypes with less expression of the vCJD [20], the Met/Val heterozygotes with longer survival in sporadic CJD [21] and one Met/Val heterozygote remaining asymptomatic from vCJD 5 years after transfusion of the contaminated blood [22]. From the molecular viewpoint, rapid β-sheet-rich oligomerization of the Met variant of PrP may be an explanation for the increased disease susceptibility in CJD [23]. It has once been proposed that an alternative initiation site of translation at codon 129 in Met variant may be present, which did not get confirmed by the cDNA analysis from our five subjects who all carry Met/Met homozygosity. Although there is no association of codon 129 polymorphism with PD in this study, we did not exclude the possibility that the PrP might contribute to the vulnerability to the diseases with functional element in the promoter [24] or in 3′-untranslated region rather than in the reading frame.

Octarepeat changes in our four PD subjects, one with NS and three with CD, occurred with a prevalence of 1.2%. The NS pattern, R3g, in one of the PD subjects (R1–2–3g–3–4/R1–2–2–3–4) is considered as a degenerated codon with synonymous translation. A son of this patient possesses homozygotic R1–2–2–3–4/R1–2–2–3–4. R3g may be considered as the oligomer hybrid of the R3 (5′-half) and R2 (3′-half), but how R3g develops in evolution and what outcome it goes is still elusive. Two possibilities rise: (1) modification of cytidine methylation in the Watson or Crick strands in R3g as well as the other NS variants (such as R2a or R2c) with the GC-rich content; or (2) the alternative splicing by the “GC-AG” [25] or “AT-AC” classes [26] rather than the canonical “GT-AG” rule after nucleotide substitutions in R2 copies and R3g in R3 copy, resulting in in-frame changes of PrP on translation. However, the RT-PCR results from our PD subject with NS phenomenon of octarepeat of PrP did not show evidence of alternative splicing of PrP.

Another change of octarepeat region in our PD subjects is CD. PrP octarepeats may prevent formation of PrP amyloid fibrils [27] and a possible sacrificial role during oxidative stress [28]. R2 deletion occurred in one subject of the Korean population [29] and among our parkinsonian subjects, but is absent among our controls. The CD phenomenon of PrP octarepeats has been reported in two subjects with the iatrogenic CJD in the past [30] but the relationship between susceptibility of CJD and CD phenomenon is still not known. CD may change the tertiary structure of PrP, resulting in the reduced binding affinity of copper ion, a loss of anti-oxidative function of PrP, different epitope expression of PrP and aberrant cell signaling. But it is difficult to explain the discrepancies of findings of the cDNA (R1–2–3g–3–4/R1–2–2–3–4) and genomic DNA in our PD heterozygote with R2 deletion, R1–2–3–4/R1–2–2–3–4. We propose one hypothesis here (Fig. 3), including the longer DNA strand as the template, hairpin structure in the transcription bubble, and transcript shift to continue the synthesis of this RNA motif. This consequence of RNA synthesis resulted in the R4–3–3g–2–1 (Fig. 2). The shorter coding strand with CD might play a key role while forming the hairpin structure due to unequal length of DNA in the transcription bubble. The R2 fragment just near the R3 might contribute to the RNA synthesis and form the secondary structure. But the “shift-backward” synthesis on the R3 template in this hypothesis may be a speculative phenomenon without evidences in our study. If the hypothesis works, it may explain the generation of the degenerate codons R2a, R2c and R3g based on the presence of CD of one allele, but is still not able to know how CD formed while chromosomal recombination and segregation of mitosis take place. Although absence of the cultures of fibroblasts or lymphocytes from the PD subject with CD to prove this hypothesis, we still considered the changed secondary structure of DNA template strand as a possible mechanism related to PD reported here. Neurodegenerative signal might begin from the RNA level.

View Large Image Download to PowerPoint

Fig. 3 The hypothesis about the synthesis of the RNA sequence of R1–2–3g–3–4 in the genomic CD heterozygotes. (A) The longer DNA strand (R1–2–2–3–4) acted as the template. Transcription took place from the R4 fragment, followed by the R3, and then the partial R2 (symbolized as r2). (B) The hairpin structure formed by the two tandem R2 repeats was attributed to the unequal length of the DNA strands in the transcription bubble. The structure caused by GGGGG in one R2 motif and CCTC in the other tandem R2 in the transcription bubble rendered the transcript “shifting backward” to R3 motif which has the high homology to R2 sequence. The backward shift was for the purpose to release the hairpin structure, resulting in the tandem partial R3 synthesis (symbolized as r3) and thus the oligomer hybrid of the R3 and R2 formed. (C) Recovery of the secondary structure occurred on synthesizing the fourth copy via the template R2 and the R2–R1 synthesis continued.

In this study, we found no significant polymorphic difference at codon 129 of PrP between PD and control subjects, but a minor group of PD subjects having octarepeat changes. All of the subjects with CD or NS patterns of octarepeat displayed Met/Met homozygosity at codon 129. These changes may be regarded as silent polymorphic variants as the RNA evaluation by RT-PCR. However, it may give hints about octarepeat polymorphism during degeneration of dopaminergic neurons, possibly not only attributed to the protein molecule but to the RNA or other molecules taking part in the neurodegeneration.