Background Autosomal recessive hereditary spastic paraplegias (AR-HSP) constitute a heterogeneous group of neurodegenerative diseases involving pyramidal tracts dysfunction. The genes responsible for many types of AR-HSPs remain unknown. We attempted to identify the gene responsible for AR-HSP with optic atrophy and neuropathy.
Methods The present study involved two patients in a consanguineous Japanese family. Neurologic examination and DNA analysis were performed for both patients, and a skin biopsy for one. We performed genome-wide linkage analysis involving single nucleotide polymorphism arrays, copy-number variation analysis, and exome sequencing. To clarify the mitochondrial functional alteration resulting from the identified mutation, we performed immunoblot analysis, mitochondrial protein synthesis assaying, blue native polyacrylamide gel electrophoresis (BN-PAGE) analysis, and respiratory enzyme activity assaying of cultured fibroblasts of the patient and a control.
Results We identified a homozygous nonsense mutation (c.394C>T, p.R132X) in C12orf65 in the two patients in this family. This C12orf65 mutation was not found in 74 Japanese AR-HSP index patients without any mutations in previously known HSP genes. This mutation resulted in marked reduction of mitochondrial protein synthesis, followed by functional and structural defects in respiratory complexes I and IV.
Conclusions This novel nonsense mutation in C12orf65 could cause AR-HSP with optic atrophy and neuropathy, resulting in a premature stop codon. The truncated C12orf65 protein must lead to a defect in mitochondrial protein synthesis and a reduction in the respiratory complex enzyme activity. Thus, dysfunction of mitochondrial translation could be one of the pathogenic mechanisms underlying HSPs.
- Clinical genetics
- Molecular genetics
- Movement disorders (other than Parkinsons)
Statistics from Altmetric.com
Hereditary spastic paraplegias (HSPs) comprise a large and heterogeneous group of genetic disorders mainly affecting the pyramidal tracts of the legs. The cardinal pathological findings in HSPs are the result of a dying back degeneration of the corticospinal tracts in the spinal cord. The longest fibres, innervating the lower extremities are mostly affected. HSPs are divided into two subtypes that comprise pure and complex forms. The pure form of HSP is characterised by progressive bilateral leg spasticity, weakness, exaggerated tendon reflexes and positive pathological reflexes, whereas the complex form of HSP shows the following additional symptoms: peripheral neuropathy, cerebellar atrophy, thin corpus callosum, optic atrophy, retinal degeneration, mental impairment, convulsions and extrapyramidal signs.1
HSPs can be inherited in an autosomal-dominant (AD), autosomal-recessive (AR) or X-linked recessive (XR) manner. To date, at least 52 spastic paraplegia gene (SPG) loci have been assigned, and approximately 30 genes have been identified. The pure form is usually transmitted as an AD trait, whereas the complex form is transmitted as an AR or XR one. The most common AD-HSP is SPG4 with the spastin gene mutation, accounting for 40–45% of AD-HSP.2 Meanwhile, the most frequent AR-HSP might be SPG11 with the spatacsin gene mutation, showing a complex phenotype including dementia and thin corpus callosum.3 The genes most responsible for AR-HSPs, however, remain unknown.
Several pathogenic mechanisms underlying HSPs have been suggested. HSPs might result from disruption of the axonal transport of molecules, organelles and other cargos, which mainly affects the distal parts of motor neurones.4 Axonal transport might be impaired by mutations of the SPAST gene5 and kinesin heavy chain KIF5A.6 An animal model of spastin deletion shows progressive axonal degeneration restricted to the central nervous system leading to a late and mild motor defect. The degenerative process is characterised by focal axonal swelling associated with abnormal accumulation of organelles and cytoskeletal components.7 The intracellular transport of molecules and organelles to and from nerve terminals also depends on the mitochondrial function.
An abnormal mitochondrial function also leads to several HSPs: SPG7 with the paraplegin gene mutation and SPG13 with the heat-shock protein 60 (HSPD1) one. For instance, paraplegin is a part of the metalo-protease AAA (ATPases associated with diverse cellular activities) complex with AFG3L2,8 an ATP-dependent proteolytic complex located at the mitochondrial inner membrane, which controls protein quality and regulates ribosome assembly.9 A homozygous mutation of AFG3L2 also leads to spastic ataxia-neuropathy syndrome (SPAX-5 in OMIM).10 Paraplegin-deficient mice are affected by distal axonopathy of spinal and peripheral axons, characterised by axonal swelling and degeneration caused by massive accumulation of organelles and neurofilaments, similar to those observed in the animal model of spastin deletion.11
Here, we report a novel homozygous nonsense mutation in the chromosome 12 open reading frame 65 (C12orf65) gene in patients with AR-HSP with optic atrophy and neuropathy found on linkage analysis involving single nucleotide polymorphism (SNP) and exome sequencing. Furthermore, we revealed that this mutation led to a mitochondrial translation dysfunction in a patient (proband). This is the first report that a C12orf65 mutation causes HSP.
Patients and methods
The present study involved two patients from a family with spastic paraplegia, optic atrophy and peripheral neuropathy described elsewhere.12 The family pedigree is shown in figure 1A. The parents were first cousins. Two affected (IV-3 and 6) and three unaffected members (III-1, 2 and IV-4) of the family underwent neurological examinations and nerve conduction studies, except for III-2 and IV-4.12 Presently, the unaffected members are all deceased.
Genomic DNA was extracted from blood samples from the two affected individuals (IV-3 and 6) with written informed consent (figure 1A), and then multipoint parametric linkage analysis involving a SNP high-throughput linkage analysis system (SNP HiTLink) was performed.13 With this system, SNP chip data for the Mapping 100 k/500 k array set and Genome-Wide Human SNP array 6.0 (Affymetrix, Santa Clara, California, USA) can be directly imported and passed to a multipoint parametric linkage analysis programme Allegro.14 Parametric LOD scores were calculated using Allegro V.2 with the parameter setting of an AR model with 100% penetrance.
Copy-number variation detection
We performed array-based comparative genomic hybridisation (aCGH) analysis for copy-number alteration detection in the candidate gene areas. We developed custom aCGH arrays against the candidate areas in the four chromosomes using a Human Genome CGH Microarray Kit 244K (Agilent Technologies, Santa Clara, California, USA) according to the manufacturer's protocol.15 ,16 We made one CGH probe every 160 bp on average against the candidate gene areas.
We collected a blood sample from one affected individual (IV-6) and performed massively parallel sequencing. Genomic DNA was extracted from leukocytes from the case and then sheared. An adaptor-ligated library was prepared and clustered on the cBOT system (Illumina, San Diego, California, USA). Exon capture was performed with a SureSelect Human All Exon kit (Agilent). Paired-end sequencing was carried out on an Illumina Hiseq 2000 that generated 91-bp reads. For sequence alignment, variant calling and annotation, the sequences were aligned with the human genome reference sequence (hg19 build) using a Burrows-Wheeler Aligner. Substitution calling was carried out with a Genome analysis tool kit (GATK). SNP calls were made with a GATK Unified Genotyper, and indel calls were made with a GATK IndelGenotyper V2. SNP calling was performed with reference to dbSNP131 and dbSNP134. All variants were annotated with reference to consensus coding sequences (CCDS) (NCBI release 20090902) and RefSeq (UCSC dumped 20101004).
The coding exons and flanking intronic sequences of C12orf65 were amplified using the genomic DNA of the patients, and using an Marshall Scientific (MS) Research Thermal Cycler. The primer sequences were as follows: Ex2-F: aac atg gca gac agt gca ag, Ex2-R: ggc tga tcc cat tca cac tt, Ex3-F: ttc tga ggt cct gtc cat ttt t, and Ex3-R: gcc cag ccg agt ttt att ct. Sanger sequencing was performed according to an established standard protocol on an Applied Biosystems (ABI) 3730 capillary sequencer (Applied Biosystems, Carlsbad, California, USA). We sought additional C12orf65 mutations by Sanger sequencing of the coding regions of two exons and their flanking sequences of C12orf65 in 74 index Japanese AR-HSP patients without known HSP gene (SPG1/2/3A/4/5/6/7/8/10/11/13/17/20/21/31/33) substitutions established by the Japan Spastic Paraplegia Research Consortium, and one case in a family with Charcot–Marie–Tooth disease (CMT) and optic atrophy with AR transmission. Genomic DNA samples from 200 Japanese subjects without apparent neurologic disorders were also analysed as controls.
Mitochondria were isolated from primary fibloblasts from the patient and two controls. Mitochondrial protein (10 μg per lane) was fractionated by 12% NuPAGE Gel (Life Technologies) and transferred to polyvinylidene fluoride (PVDF) membranes. The filters were preincubated for 1 h with Block One (Nacalai), followed by incubation for 1 h with Can Get Signal solution 1 (Toyobo) containing C12orf65 antibodies (Abcam) or voltage dependent anion channel (VDAC) antibodies (MitoSciences). The filters were washed four times with phosphate-buffered saline (PBS) containing 0.1% Tween 20, incubated for 1 h with Can Get Signal solution 2 (Toyobo) containing horseradish peroxidase-conjugated antirabbit or mouse immunoglobuline G (igG) (GE Healthcare), and then washed with PBS containing 0.1% Tween 20. Protein bands were visualised using enhanced chemiluminescence (ECL) prime western blotting reagents (GE Healthcare).
Mitochondrial protein synthesis assaying17
The patient's (IV-6) skin fibroblasts were obtained by skin biopsy with informed consent. Fibroblasts derived from two non-mitochondrial disease cases were used as controls. (35S)Methionine, cysteine incorporation into mitochondrially encoded proteins was analysed essentially as described previously,18 ,19 using the patient's and control fibroblasts in culture. Semiconfluent cells in 6-well plates were labelled with (35S) EXPRESS protein labelling mix (Perkin Elmer) for 30 min, in the presence of emetine (0.1 mg/ml) in methionine-free DMEM supplemented with dialysed 10% fetal bovine serum (FBS). The cells were lysed in 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, and 1% sodium dodecyl sulfate (SDS). Total cellular proteins (10 μg/lane) were fractionated by 15–20% gradient SDS-PAGE. The gel was stained with Quick-coomassie brilliant blue (CBB) PLUS (Wako) and dried. The gel was exposed to an imaging plate, and the labelled polypeptides were located with a bioimaging analyser (BAS3000, Fuji Photo Film).
Blue native polyacrylamide gel electrophoresis (BN-PAGE) and western blotting for immunodetection20
Mitochondrial proteins were isolated from cultured fibroblasts from patient IV-6.21 A mixture of mitochondrial proteins from 10 non-mitochondrial disease cases was used as a control. For the detection of individual complexes, the mitochondrial fraction (20 μg protein) was solubilised with 0.5% (w/v) n-dodecyl-β-D-maltoside, and for the detection of supercomplexes, the mitochondrial fraction (30 μg protein) was solubilised with 1% (w/v) digitonin, respectively. Electrophoresis was performed on 3–12% gradient polyacrylamide gels (Invitrogen).21 ,22 Following BN-PAGE, the gels were blotted onto polyvinylidene fluoride membranes using an iBlot transfer system (Invitrogen). Subunit-specific primary antibodies were used to immunodetect protein complexes. The cocktail of primary antibodies comprised NDUFA9 (complex I, Invitrogen) (2.0 μg/ml), SDHA (complex II, Invitrogen) (0.02 μg/ml), UQCRC2 (complex III, Abcam) (0.2 μg/ml), MTCO1 (complex IV, Invitrogen) (0.2 μg/ml), and ATP5B (complex V, Invitrogen) (2.0 μg/ml). After removing the cocktail of primary antibodies, alkaline phosphatase-conjugated secondary antibodies were added, and then chemiluminescent detection was performed with a bioimaging system (LAS4000 mini, GE Healthcare).
Enzymatic activity of respiratory chain complexes20
The enzymatic activity of individual mitochondrial respiratory complexes was determined using the mitochondrial fractions (1 μg protein) isolated from cultured fibroblasts from patient IV-6 and 10 controls according to Trounce et al23 with modifications. The activities of complexes I, II, III and IV were measured using a multiwell plate reader spectrophotometric system (SPECTROstar Nano, BMG Labtech). Citrate synthase (CS) activity was used for normalisation. All measurements were performed in triplicate and averaged.
The proband (IV-6) was a 32-year-old man who was admitted to our hospital for evaluation of slowly progressive weakness of the lower extremities and decreased visual acuity. He noticed the reduced visual acuity at age 7 years. At about age 10 years, his leg weakness and drop feet led to a steppage gait. He could not execute fine finger movement at age 16 years. On ophthalmologic examination, he exhibited 20/100 vision in the right eye and 20/200 in the left one. Fundoscopic examination demonstrated bilateral optic atrophy with central scotoma. On neurologic examination, bilateral leg spasticity, and anterior tibial muscle weakness and atrophy were noted. Mild distal arm weakness without atrophy was observed. Tendon reflexes, except for normal ankle jerks, were exaggerated in all extremities. Bilateral planter responses were flexion. Superficial and vibratory sensations were diminished in the distal legs with a glove and stocking distribution and preserved position sense. The results of blood and cerebrospinal fluid examinations were within normal ranges. The urine organic acid pattern was normal. EEG, ECG and brain CT showed no remarkable findings. A nerve conduction study disclosed mild decreases of motor and sensory nerve conduction velocities in the upper extremities, but they were not evoked in the lower extremities. High-amplitude NMUs, and a reduced number of NMUs, were observed in the extremities on needle electromyography (EMG). Motor-evoked potential examination revealed prolongation of the central motor conduction times in the corticospinal tracts. Microscopic examination of a muscle biopsy specimen showed grouped atrophy in the tibialis anterior muscle. A sural nerve biopsy revealed a decreased number of nerve fibres of large diameter, the formation of an onion bulb-like structure, and endoneural fibrosis (figure 1B).
The second case (IV-3), a brother of the proband, exhibited essentially the same clinical phenotype. He noticed the visual difficulty at age 7 years, the onset of slowly progressive muscle atrophy in the lower extremities at age 10 years, and pes equinovarus deformities at age 12 years. Neurologic examination at age 42 years showed bilateral optic atrophy with central scotomas and diminished visual acuity (10/200 on right and 16/200 on left). In the lower extremities, marked muscular atrophy, spasticity and pes equinovarus were noted. Deep tendon reflexes were exaggerated in all extremities with left patellar clonus and extensor planter reflexes bilaterally. Superficial sensation and vibration were diminished in a glove and stocking distribution. Position sense was normal. Needle EMG showed neurogenic patterns. A nerve-conduction study disclosed slight decreases in motor nerve conduction velocities and normal sensory nerve conduction velocities in the upper extremities, but they were not evoked in the lower extremities.
Identification of candidate chromosome areas
We found linkages, that were not statistically significant, to chromosomes 2 (rs116837089-rs118552355), 6 (rs131535042-rs138306536), 12 (rs119856515-rs125250432) and 13 (rs67990237-rs70707721), with maximum cumulative logarithm of the odds (LOD) scores of 1.8 (figure 2A). Array CGH analysis revealed no pathological copy-number alterations in the four candidate chromosome areas. These four areas contained neither previously identified HSP loci nor CMT loci.
Exome sequencing allowed identification of the candidate gene substitutions
The presence of consanguinity, and the fact that the parents appeared asymptomatic, suggested that the patients had homozygous disease-causing mutations, and therefore, a homozygous autosomal-recessive model was applied. Exome sequencing covered 98.65% of the target region, and the average sequence depth on target was 41.47.
We identified three homozygous, non-synonymous single nucleotide variants (c.394C>T (p.R132X) in C12orf65, c.136G>A (p.E46K) in COQ5, and c.6599T>G (p.I2200S) in KNTC1) in the chromosome 12 candidate area with reference to dbSNP131. No such variations were detected in the chromosome 2, 6 and 13 areas. We could subsequently exclude the two candidate gene mutations in COQ5 and KNTC1 as benign polymorphisms, because the two variants had been registered as SNPs to the dbSNP134 and 1000 genomes (rs139585780: c.136G>A in COQ5, A allele frequency=0.0112 and rs140880563: c.6599T>G in KNTC1, G allele frequency=0.0056 in Japanese). The remaining candidate gene, C12orf65, was confirmed to have a homozygous nonsense mutation (c.394 C>T, p.R132X) on Sanger sequencing in the two patients, IV-3 and 6 (figure 2B). This nonsense mutation was not found in 200 Japanese control genomic DNAs.
Mitochondrial respiratory function is impaired by the mutation in C12orf65
Very recently, two homozygous C12orf65 1 bp deletion mutations were identified in patients with Leigh syndrome, optic atrophy and ophthalmoplegia.24 The C12orf65 protein shows high sequence similarity to mitochondrial class I peptide release factors (RFs), and it has been reported that mutations in C12orf65 cause a mitochondrial translation defect. Analysis of the assembly of mitochondrial phosphorylation complexes showed decreases of complexes I, III, IV and V. This disease entity is called combined oxidative phosphorylation deficiency 7 (COXPD7) (online Mendelian inheritance in man (OMIM) #613559). To determine how this nonsense mutation affects the mitochondrial function, we first performed immunoblot analysis and mitochondrial protein synthesis assaying of patient fibroblasts. Immunoblot analysis showed that a smaller C12orf65 protein was generated in the patient's fibroblasts than the wild-type protein in controls (figure 3A). Marked reductions of the synthesised polypeptides were observed in patient IV-6 compared with those in controls (16% for COX I and COX II, figure 3B), implying the aberrant statuses of respiratory complexes. Actually, complexes I and IV were severely impaired in their enzymatic functions (29% and 13%, respectively, figure 4A), and holoenzyme structures (the average value for two experiments, 17% and 23%, respectively, figure 4B) in patient IV-6, probably due to the mitochondrial translation defect, as shown in figure 3B. Furthermore, the amounts of respiratory supercomplexes that consist of complexes I, III and IV were also significantly reduced in patient IV-6 (about 30%, figure 4B). Thus, the disrupted protein integrity (function and structure) of complexes I and IV, which is closely associated with the aberrant mitochondrial bioenergetics in patient IV-6, would constitute the molecular pathogenicity in a patient carrying a mutation in C12orf65.
In the present study, we identified a novel C12orf65 nonsense mutation (c.394C>T, p.R132X) in patients with spastic paraplegia with optic atrophy and neuropathy with AR inheritance (SPG55). The possibility remains that there is another undetected mutation in the four candidate chromosome areas because the exome sequencing did not cover all exons in the areas. However, our patients carried the nonsense mutation of C12orf65 that had been identified as the causative gene for the neurological disorder COXPD7. In COXPD7, it had been reported that deletion mutations of C12orf65 lead to a decrease in mitochondrial translation and combined oxidative phosphorylation (OXPHOS) deficiencies. In the present study, we clearly demonstrated that a nonsense mutation of C12orf65 led to a mitochondrial protein synthesis defect and respiratory complex enzyme activity reduction, similarly. Thus, the spastic paraplegia phenotype in our patients could be a novel and distinct clinical entity based on the C12orf65 mutation.
The C12orf65 protein is considered to belong to the mitochondrial class I peptide RFs.24 This protein is a soluble matrix, one that is not coprecipitated with mitochondrial ribosomes,25 and that has no peptidyl-tRNA hydrolase activity.24 Therefore, although the actual function of C12orf65 is still not clear, it might play a role in recycling abortive peptidyl-tRNAs that have been prematurely released from ribosomes during polypeptide elongation. The nuclear C12orf65 gene comprises three exons, and its protein-coding region comprises 501 bp derived from its exons 2 and 3. The C12orf65 protein consists of 166 amino acids that include a RF-1 domain (amino acid numbers 53–146) and a glycine-glycine-glutamine (GGQ) motif (amino acid numbers 71–73). Our patients had a homozygous nonsense mutation (c.394 C>T) in exon 3, that is predicted to result in a stop codon at 132. Previously reported COXPD7 patients had two homozygous deletion mutations (c.210delA and c.248delT), and both mutations resulted in a premature stop codon at 84 in exon 2.24
The RT-PCR product derived from the C12orf65 mRNA was not significantly reduced in COXPD7 patients24 or our patients compared with that in normal controls (data not shown). We suggest that truncated C12orf65 proteins were generated in these COXPD7 and our patients (figure 3A). COXPD7 and our patients share a few clinical symptoms, that is, optic atrophy and peripheral neuropathy. We consider that our patients could be included in COXPD7 with a different clinical phenotype. However, the cardinal clinical feature in COXPD7 patients is Leigh syndrome, while that in our patients is spastic paraplegia. The former shows a more severe phenotype than the latter. We found that our patient has a relatively larger C12orf65 protein (131 amino acids) than that in COXPD7 cases (83 amino acids). Moreover, COXPD7 patient's fibroblasts have been reported to exhibit severe decreases in complexes I, IV and V. Meanwhile, our patient's fibroblasts showed decreases in complexes I and IV, and a milder decrease in complex V. In several reports, patients with mtDNA translation defects were found to be associated with an unaffected amount of complex III or a milder decrease than of other complexes.26–29 In the case of COXPD7 patient's fibroblasts, there was a milder decrease in complex III, while our patient's fibroblasts had an unaffected amount of complex III. Therefore, our patients might have a preserved C12orf65 protein function compared with that in COXPD7.
A study involving cultured neurones showed a complex I deficiency could increase mitochondrial reactive oxygen species (ROS) production, and increase neuronal death attenuated by ROS scavengers.30 Accumulated intracellular oxidative damage to neurones could be a causative mechanism for neurodegeneration caused by mitochondrial respiratory deficiencies.
To date, several HSP genes related to mitochondria have been identified: paraplegin and HSPD1. The SPG31 protein, receptor expression-enhancing protein 1 (REEP1), coordinates ER shaping and microtubule dynamics,31 although another report has proposed that REEP1 is a mitochondrial protein.32 SPG7 is an AR-complicated HSP resulting from paraplegin gene mutations. The clinical features of SPG7 are varied with associated symptoms: cerebellar signs, optic atrophy, distal amyotrophy and peripheral neuropathy. Our patients and SPG7 ones share the symptoms of optic atrophy and neuropathy as well as spastic paraplegia, although the paraplegin gene mutation was not identified in this family.
Paraplegin is a nuclear-encoded mitochondrial protein. SPG7 fibroblasts have been demonstrated to exhibit reduced respiratory chain complex I activity.33 Complex I activity was also reduced in our patient's fibroblasts. Considering the above findings, the present type of HSP has some clinical and biochemical features in common with SPG7. Furthermore, there has been one report of a defect of complex I activity in HSP patients without SPG7 gene mutations,34 and another one with significant decreases in mitochondrial respiratory chain complexes I and IV in non-SPG4/SPG7 HSP families.35 These families might have a mutation in the C12orf65 genes responsible for reduced mitochondrial respiratory chain complex activities.
It may be noteworthy that in Friedreich's ataxia (FA), another nuclear-encoded mitochondrial defect, there is corticospinal degeneration. In rare cases, FA may present clinically as spastic paraparesis.36 ,37
In conclusion, the present study allowed identification of a C12orf65 gene mutation in AR-HSP with optic atrophy and neuropathy, and revealed a mitochondrial translation dysfunction resulting in reduced enzyme activities of respiratory chain complexes. Our study should provide additional insights into the pathogeneses of HSPs or other neurodegenerative diseases involving mitochondrial gene mutations.
We wish to thank Dr Mayumi Komine (Department of Dermatology, Jichi Medical University) for performing the skin biopsy on the patient. We also thank Dr Satoko Kumada (Department of Neuropediatrics, Tokyo Metropolitan Neurological Hospital) and Dr Kazuma Sugie (Department of Neurology, Nara Medical University) for sending the patients’ samples.
Contributors HS, YT, JG, ST and YG designed the study; HS, HI, CS, YM, HH, JH, KS, TN, MN, YF and YT performed the experiments; HS, YT, HI, CS, YM, HH, JG, ST and YG collected and analysed the data; HS provided the DNA samples and clinical information; HS, YT, ST and YG wrote the manuscript; and YT, JG, ST, YG and IH provided technical support, conceptual advice and project coordination.
Funding This work was supported by a grant from the Research Committee for Ataxic Diseases (YT and HS) of the Ministry of Health, Labour and Welfare, Japan. This work was also supported by Grants-in-Aid from the Research Committee of CNS Degenerative Diseases (IN and YT), and the Ministry of Health, Labour and Welfare of Japan, and by a Grant-in-Aid for Scientific Research (C) (23591253 to HS) from The Ministry of Education, Culture, Sports, Science and Technology in Japan.
Competing interests None.
Patient consent Obtained.
Ethics approval This study was approved by the institutional review board of the Jichi Medical University, University of Yamanashi, University of Tokyo, and the National Institute of Neuroscience, National Centre of Neurology and Psychiatry.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Exome sequencing data are available (DDBJ Sequence Read Archive (DRA) accession number DRA000534).
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.