- Email: [email protected]
Investigating the role of the physiological isoform switch of cytochrome c oxidase subunits in reversible mitochondrial disease
Accepted Manuscript Title: Investigating the role of the physiological isoform switch of cytochrome c oxidase subunits in reversible mitochondrial disease Author: Veronika Boczonadi Michele Giunta Maria Lane Mar Tulinius Ulrike Schara Rita Horvath PII: DOI: Reference:
S1357-2725(15)00035-7 http://dx.doi.org/doi:10.1016/j.biocel.2015.01.025 BC 4550
To appear in:
The International Journal of Biochemistry & Cell Biology
Received date: Revised date: Accepted date:
6-11-2014 17-1-2015 29-1-2015
Please cite this article as: Boczonadi, V., Giunta, M., Lane, M., Tulinius, M., Schara, U., and Horvath, R.,Investigating the role of the physiological isoform switch of cytochrome c oxidase subunits in reversible mitochondrial disease, International Journal of Biochemistry and Cell Biology (2015), http://dx.doi.org/10.1016/j.biocel.2015.01.025 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Investigating the role of the physiological isoform switch of cytochrome c oxidase subunits in reversible mitochondrial disease Veronika Boczonadi1 , Michele Giunta1, Maria Lane1, Mar Tulinius2, Ulrike Schara3 and Rita
cr
Institute of Genetic Medicine, Wellcome Trust Mitochondrial Research Centre, Newcastle
us
1
ip t
Horvath1
University, Central Parkway NE1 3BZ Newcastle upon Tyne, UK;
Department of Paediatrics, The Sahlgrenska Academy, University of Gothenburg, Box 400,
an
2
Göteborg, SE-405 30 Sweden;
Department of Paediatric Neurology, University of Essen, Hufelandstraße 55, Essen, 45122
M
3
Word count: 3691 words
Ac ce p
Abstract: 229 words
te
d
Germany
Title: 127 characters including spaces and punctuation Keywords: reversible infantile respiratory chain deficiency, cytochrome c oxidase, isoform switch, mitochondrial tRNAGlu
Address correspondence to: Rita Horvath MD, PhD Institute of Genetic Medicine, Newcastle University, Central Parkway, Newcastle upon Tyne, NE1 3BZ, UK, Phone: +44 191 2418855, Fax: +44 191 2418666 Email: [email protected]
Page 1 of 32
2
HIGHLIGHTS We show a developmental isoform switch of COX6A/COX7A in muscle of humans
ip t
and mice.
cr
Liver isoforms are present at birth, heart/muscle isoforms increase ~3 months of
us
age.
The physiological isoform switch does not explain reversible mitochondrial disease.
Abstract
te
d
M
an
Developmental changes of COX isoforms may modify other mitochondrial diseases.
Ac ce p
Reversible infantile respiratory chain deficiency is characterised by spontaneous recovery of mitochondrial myopathy in infants. We studied, whether a physiological isoform switch of nuclear cytochrome c oxidase subunits contributes to the age-dependent manifestation and spontaneous recovery in reversible mitochondrial disease. Some nuclear-encoded subunits of cytochrome c oxidase are present as tissue specific isoforms. Isoforms of subunits COX6A and COX7A expressed in heart and skeletal muscle are different from isoforms expressed in liver, kidney and brain. Furthermore, in skeletal muscle both the heart and liver isoforms of subunit COX7A have been demonstrated with variable levels, indicating that the tissuespecific expression of nuclear-encoded subunits could provide a basis for the fine-tuning of cytochrome c oxidase activity to the specific metabolic needs of the different tissues.
Page 2 of 32
3
We demonstrate a developmental isoform switch of COX6A and COX7A subunits in human and mouse skeletal muscle. While the liver type isoforms are more present soon after birth, the heart/muscle isoforms gradually increase around 3 months of age in infants, 4 weeks of
ip t
age in mice, and these isoforms persist in muscle throughout life. Our data in follow-up biopsies of patients with reversible infantile respiratory chain deficiency indicate that the
cr
physiological isoform switch does not contribute to the clinical manifestation and to the
us
spontaneous recovery of this disease. However, understanding developmental changes of the different cytochrome c oxidase isoforms may have implications for other mitochondrial
Ac ce p
te
d
M
an
diseases.
Page 3 of 32
4
1. Introduction The mitochondrial oxidative phosphorylation (OXPHOS) machinery is composed of 5
ip t
multimeric complexes (complex I, II, III, IV and V) embedded in the inner mitochondrial membrane (Greaves et al., 2012). Under normal physiological conditions, mitochondrial
cr
oxidation is responsible for the majority of energy production in most cells and tissues. The
us
utilisation of oxygen by the mitochondria to generate ATP is regulated by the cytochrome c oxidase enzyme (COX, complex IV) that is the terminal enzyme of the mitochondrial electron
an
transport chain. The formation of the functional COX is under the control of two separate genetic systems, the nuclear genome (nDNA) and the mitochondrial genome (mtDNA)
M
coding for 10 and 3 subunits respectively (Shoubridge, 2001).
d
Within the OXPHOS system a unique feature of COX is the presence of subunits with tissue
te
specific isoforms. It was shown in early studies, that some, but not all nuclear-encoded subunits of COX are present as tissue specific isoforms (Taanman et al., 1993). At least five
Ac ce p
of the COX subunits have been reported with such isoforms suggesting a regulatory function in energy production (COX4I1/COX4I2, COX6A1/COX6A2, COX6B1/Cox6B2, COX7A1/COX7A2 and COX8A/COX8B/COX8C) (Huttemann et al., 2001; Kadenbach et al., 2000). In mammals three of these isoform pairs correspond to muscle and non-muscle specific forms: COX6A, COX7A and COX8 (Grossman and Lomax, 1997). Isoforms of COX4I and COX6B, represent specific expression in the lung and in the testes respectively (Huttemann et al., 2003). One isoform of subunit COX6A and COX7A is expressed in heart and skeletal muscle (heart/skeletal muscle isoform) while the other is expressed in liver, kidney and brain (liver or non-muscle type). Interestingly, in humans there is no evidence of a contractile muscle and non-muscle specific isoforms of subunit COX8A, as oppose to other mammals and birds
Page 4 of 32
5
(Rizzuto et al., 1989). It has been postulated in the past that the tissue-specific expression of nuclear-encoded subunits could provide a foundation for the fine-tuning of COX activity to the specific metabolic needs in different tissues (Taanman et al., 1993). Understanding basic
ip t
mechanisms of COX deficiencies raised interest in studying the development of nuclear COX
cr
subunits.
The potential relevance of the different tissue specific COX isoforms in manifestation of
us
mitochondrial disease has been hypothesized over 20 years ago (Tritschler et al., 1991). The clinical description of a unique, reversible mitochondrial disease, reversible infantile COX
an
deficiency myopathy (recently named as reversible infantile respiratory chain deficiency,
M
RIRCD) caused by a homoplasmic mt-tRNAGlu mutation (m.14674T>C/G), raised the possibility, that developmental factors may have a role in the clinical manifestation
d
(Horvath et al., 2009, Mimaki et al., 2010, Uusimaa et al., 2012). While most mitochondrial
te
diseases are severe, progressive conditions, RIRCD is one of the few hereditary diseases with a life-threatening onset showing a remarkable reversible disease course within the first year
Ac ce p
of life. Before the identification of the primary genetic cause of this condition immunohistochemistry was suggested to distinguish between the fatal and the benign forms of COX deficiency (Tritschler et al., 1991). It has been hypothesized previously that the nuclear-encoded COX6A and COX7A subunits undergo a developmental isoform switch from fetal to ubiquitous isoforms in early childhood and it may explain the disease recovery in reversible infantile myopathy (Tritschler et al., 1991; Taanman et al., 1993). Up to date no experimental evidence has been shown in patients carrying the m.14674T>C/G variant, that a switch between isoforms would underlie the recovery.
Page 5 of 32
6
We suggested that tissue-specific, developmentally timed processes play a role both in the age-dependent expression and in the reversibility in RIRCD. In this study we investigated the physiological isoform switch of nuclear COX subunits in normal human skeletal muscle
ip t
(infant, adult) and studied whether it contributes to the recovery in RIRCD.
cr
2.Materials and Methods
us
2.1. Immunofluorescence
an
OCT embedded mouse and human muscle tissues were sectioned. Dried sections were fixed with 4% PFA for immune-staining. The primary antibodies were: Anti-Cytochrome c
M
antibody (Abcam), COX6AL (Mitosciences), COX6AH (Abgent), COX7AL (Santa-Cruz), and COX7AH (Proteintech), MTCO1 (Abcam), Myosin Light chain (Millipore), Alexa fluor 488 and
d
596 conjugated secondary antibodies (Invitrogen) were used to detect the primary
te
antibodies. Cell nuclei were identified using DAPI. Immunofluorescence images were
Ac ce p
collected using a Zeiss Axioimager Z1 fluorescence microscope equipped with Zeiss Apotome 2 (Zeiss, Germany). Acquired images were processed with AxioVision Rel 4.9 software.
2.2. Immunoblotting
Muscle tissues were lysed and the extracts were cleared by centrifugation and stored at -80 C until use. The extracts were boiled in 2X Laemmli sample buffer. Samples were than subjected to SDS-polyacrylamide gel electrophoresis followed by immunoblot analysis using specific antibodies raised against COX6AL (Mitosciences), COX6H (Abgent), COX7AL (SantaCruz), COX7AH (Proteintech), Porin (Abcam). Horseradish peroxidase-conjugated secondary
Page 6 of 32
7
antibodies (Jackson ImmunoResearch Laboratories) were used for detection applying the enhanced chemiluminescence method (GE Healthcare Bio-Sciences).
ip t
2.3. Isolation of RNA and cDNA Synthesis Total RNA was prepared from mouse muscle, liver, heart and human muscle by standard
cr
phenol/chlorophorm method. Reverse transcription was carried out using 1 ug of total RNA, SuperScript II (Invitrogen). For isoform-specific PCRs, primers were designed to be specific to
us
unique sequences of the isoform transcripts. cDNA input was standardized and RT-PCRs
an
were performed with SYBR Green PCR master mix (Applied Biosystems). Samples were normalized to Actin and Vimentin in mouse, Actin and GAPDH in humans, and fold change
M
was determined by the ΔΔCt method. We used the following primers to study gene expression in mouse tissues: COX6AL forward: 5’-GCTCAACGTGTTCCTCAA-3’, reverse: 5’-
d
CGGAAGTGGGTTCACATGAG-3’; COX6AH forward: 5’-TGCCTCTAAAGGTCCTGAGC-3’, reverse:
te
5’-GAAGAGCCAGCACAAAGGTC-3’; COX7AH forward: 5’-GACAATGACCTCCCAGTACAC-3’,
3’,
Ac ce p
reverse: 5’-CAGCCCAAGCAGTATAAGCA-3’; COX7AL forward: 5’- CAGTTCATCTGAAAGGCCGGreverse:
5’-ACGACATTAGTTCTGCTTCTTGG-3’;
Vimentin:
forward
5’-
CAGCAGTATGAAAGCGTGG-3’, reverse: 5’-GGAAGAAAAGGTTGGCAGAG-3’. The following human primers were used: MTRNR1 (12S rRNA) forward: 5’-AAACTGCTCGCCAGAACACT-3’, reverse: 5-‘CATGGGCTACACCTTGACCT-3’; COX6AH forward: 5’-CGCCCCGAGTTCCGTCCCTA3’,
reverse:
5’-GGGCAGAGGGTTCACGTGGC-3’;
COX6AL
forward:
5’-
GCTGAATGTGTACCTGAAGTC-3’ , reverse:5’-TGAGGGTTATGGAATAGAGTATGG-3’; COX7AH 5’-AATGACATCCCGTTGTACCT-3’,
reverse:
5’-GGCTTCTTGGTCTTAATTCCT-3’;
COX7AL:
forward: 5’-CCCTCCTGTATAGAGCCACC-3’, reverse: 5’-TGAATGAAACTGAACCAAGCGA-3’, Beta
Actin:
forward:
5’-
GATGCAGAAGGAGATCACTGC-3’,
reverse:
5’-
Page 7 of 32
8
ACATCTGCTGGAAGGTGGAC-3’; MT-ND1 forward: 5’-ATGTCCACCCTTATCACTCA-3’, reverse: 5’-TATATCTGAGGTGGTAGGC-3’; MT-COX2 forward: 5’-CCATCCCTACGCATCCTTTA-3’, reverse: mt-tRNATrp
5’-GCCGTAGTCGGTGTACTCGT-3’;
forward:
5’-
mt-tRNAGlu
forward:
TTCTCGCACGGACTACAACC-3’;
5’-ACAACGATGGTTTTTCATATCATT-3’, cyt-tRNALys
forward:
reverse:
5’-
5’-GTCGGTAGAGCATCAGACTT-3’,
cr
3’;
ip t
GAAATTTAGGTTAAATACAGACCAAGA-3’, reverse: 5’- GAAATTAAGTATTGCAACTTACTGAGG-
3',
reverse:
5'-TCTAGACGGCAGGTCAGGTC-3',
us
reverse: 5’-GAACAGGGACTTGAACCCTG-3’; GAPDH forward: 5'-ACCCAGAAGACTGTGGATGGRN5S1@
(5S
rRNA)
forward:
5’-
an
GTCTACGGCCATACCACCCTG-3’, reverse: 5’-AAAGCCTACAGCACCCGGTAT-3’.
M
2.4. Cell culture
Myoblast cell cultures of an RIRCD patient and controls were obtained from the Biobank of
d
the Medical Research Council, Centre for Neuromuscular Diseases, Newcastle and were
te
immortalized as described (Lochmüller et al., 1999). Informed consent was signed by all
Ac ce p
subjects and ethical approval was obtained from a local research ethics committee. Muscle cells were grown in skeletal muscle growth medium (PromoCell, Heidelberg, Germany), supplemented with 4mM L-glutamine and 10% foetal bovine serum and cultured as recommended by the supplier.
Thiamphenicol (TAP) was dissolved in ethanol, the concentration was calculated spectrophotometrically (ϵmM=0.64 at 273 nm) and administered at 50 tionl−1 final concentration. Control cells were supplemented with an equal volume of ethanol (Chrzanowska-Lightowlers et al., 1994). Control and patient immortalised myoblasts were washed and reseeded in media containing TAP. Cells were left to grow for 6 days. RNA was
Page 8 of 32
9
prepared from untreated and TAP-treated control and patient cells. Small RNA was isolated from the cells with miRNA Isolation kit (Ambion).
ip t
2.5. Statistical analysis The statistical package SigmaPlot 11.0 was used to perform all the statistics. Data are
cr
presented as mean ± standard deviation. The normal distribution was checked by the
us
Kolmogorov-Smirnov test and ANOVA tests were used to compare parameters. The HolmSidak method was used for pair-wise multiple comparison procedures. A p value of <0.05
an
was considered significant.
M
2.6 Biochemical Activity Assays
Three cycles of freeze-thaw were performed on each sample prior to each assay. Each assay
d
was performed on a Multiskan Ascent 96/384 Plate Reader (ThermoFisher Scientific, UK).
te
For the Citrate Synthase activity measurement 7.5 uL of sample was added to a 96 well plate
Ac ce p
in triplicate. 193.38 uL of citrate synthase buffer (0.1 mM Tris-HCl), 2 uL DTNB 10 mM, 2 uL Triton-X 100 10% w/v and 0.62 uL acetyl Co-A 32 mM were added to each well. 1 uL of oxaloacetate 50 mM was added to each well as the initiator and the assay run at 405 nm. The activity was calculated using the following equation: Activity = Absorbance/13.6 x Vol (total)/Vol (sample)
x 1000.
For the Complex IV activity measurements 7.5 uL of sample was added to a 96 well plate in triplicates. 195 uL of reduced cytochrome c 0.1% (freshly prepared) was added to each well as the initiator and the assay run at 550 nm. The activity was calculated using the following equation: Activity = Absorbance/18.7 x Vol (total)/Vol (sample) x 1000. In all cases triplicate assays were performed. Protein determination was performed according to Bradford.
Page 9 of 32
10
2.7 Protein structure analysis Protein sequence homology was analised with SIM-Alignment Tool (Huang et al., 1991). Predicted Protein 3D structure was completed on RaptorX structure prediction server
cr
ip t
(Kallberg et al., 2012).
us
3. Results
an
3.1. Protein expression of COX6A and COX7A isoforms through mouse development In murine skeletal muscle sections we conducted immune-staining and immunoblotting to
M
determine the level of protein expression of both liver and heart type isoforms of COX6A and COX7A through development. To confirm the mitochondrial localisation of all four
d
isoform specific antibodies, muscle sections were co-stained with cytochrome c oxidase
te
subunit 1 (MTCO1) as a mitochondrial marker (Fig. 1). Immuno-fluorescent assessment of
Ac ce p
the skeletal muscle sections confirmed the specificity of the antibodies against special isoforms as they co-localised with MTCO1 (Fig. 1 A-D). Frozen sections of normal skeletal muscle were collected from 1 day, 8 days, 2 weeks, 4 weeks old and adult mice. One day old mice indicated strong, distinct mitochondrial signals for the liver type isoforms of both COX6A and COX7A subunits (Fig. 2). From days 8, the liver type isoforms gradually decreased over 2 and 4 weeks and on the adult sections no staining was observed. On the other hand, the signal for the muscle/heart type isoforms seemed to be very low in 1 day old mice and increased through the developmental series, moreover the COX7AH isoform indicated higher expression in adult sections. To confirm the shift in the liver and heart type isoforms in mouse skeletal muscle, we resolved the whole protein
Page 10 of 32
11
lysate extracted from muscle tissue on SDS-PAGE. The liver type isoforms were only detectable before 2 weeks of age. Expression of both COX6AL and COX7AL was observed until 2 weeks of age then a dramatic drop in expression level was observed. In parallel
ip t
COX6AH and COX7AH expression appeared and remained highly expressed through adult
cr
age.
us
3.2. Protein expression pattern of COX6A and COX7A subunit isoforms in human muscle We have studied human skeletal muscle biopsies of children at 3 months of age and in
an
adults. Immuno-labelling detected that at 3 months of age both isoforms of COX6A and COX7A subunits were present however the muscle/heart specific isoforms were
M
predominantly expressed in the adult tissue (Suppl. Fig. 1). These data are fully in line with
the early months of life.
d
the protein expression pattern seen in mouse muscle, confirming the isoform switch within
te
3.3. Gene expression of COX6A and COX7A subunit isoforms during development in mice
Ac ce p
We analysed the changes in gene expression of COX6AL/COX6AH and COX7AL/COX7AH in a variety of mouse and human tissues by RT-PCR. As endogenous internal reference genes in mice we studied actin and vimentin. Total RNA was isolated from mouse muscle at different ages (1 day, 8 days, 2 weeks, 4 weeks and 1 year). The relative expression of COX6AL and COX6AH isoforms were not significantly different in the early ages (1 day, 8 day, 2 weeks) however, from the age of 4 weeks a dramatic difference was observed in gene expression of several genes. COX6AH expression increased ~13 fold and continued to rise at the age of 1 year (Fig. 3A). A very similar expression pattern was observed with COX7A subunit isoforms where the COX7AH isoform suddenly increased ~13 times fold at around 1 month of age and
Page 11 of 32
12
reached a 16 fold change at the age of 1 year (Fig. 3B). This dramatic shift in gene expression of tissue specific isoforms was not observed in liver or in heart tissues (Fig. 3B3C). Nevertheless, both liver type isoforms (COX6AL/COX7AL) showed significantly higher
ip t
levels in the liver tissues when compared to heart type isoforms (Fig. 3C-D), while the mouse heart tissues demonstrated that COX6AH and COX7AH are the prominent isoforms in
cr
heart also later in life (Fig. 3E-F). We also measured the biochemical activity of the citrate
us
synthase and complex IV in mouse tissues at different ages. No significant changes were detected in the activity of citrate synthase in the muscle and the heart while slightly lower
an
activity was detected in the liver around the age of 2 weeks (Fig. 3G). Interestingly, the complex IV activity demonstrated a slight increase in the muscle and heart tissues around 1
M
week of age after which the activity returned to similar levels measured at 0 day (Fig. 3H).
te
activity of the COX holoenzyme.
d
Based on these results the COX6A and COX7A isoform change does not affect the overall
3.4. Gene expression profiles of COX6A and COX7A isoforms in normal human skeletal
Ac ce p
muscle at different ages
We further investigated the skeletal muscle specific isoform shift in human muscle. Here we have collected muscle tissues from normal human controls at different ages (21 days, 3 months, 1 year 6 months, 5 years and adult). Our aim was to pinpoint the window of time when this shift occurs in humans (Fig. 3I-J). As endogenous internal reference genes in human muscle we used actin and GAPDH. The expression of both COX6AH and COX7AH mRNA increased significantly at 3 months of age ~9 fold and ~35 fold respectively, whereas the COX6AL and COX7AL isoforms did not show dramatic alteration by age.
Page 12 of 32
13
Taken together these data illustrate that in infants COX6AL/COX6AH and COX7AL/COX7AH transcripts are co-expressed in skeletal muscle and around 3 months of age the gene expression of the muscle type isoforms (COX6AH and COX7AH) raises suddenly, and these
ip t
isoforms become the major isoforms in later age.
us
RIRCD in the early symptomatic phase and after recovery
cr
3.5. Expression pattern of COX6A and COX7A isoforms in skeletal muscle of patients with
To investigate the role of the isoform shift in RIRCD we studied how the transcription and
an
translation of these specific subunits changes after clinical recovery. We collected biopsies of RIRCD patients during the symptomatic phase (1 month) and after clinical recovery (> 2
M
years). Gene expression in muscle of an RIRCD patient (P11 in Horvath et al., 2009) during the symptomatic phase at the age of 1 month revealed significantly increased levels of
d
isoforms of both subunits, suggesting compensatory up-regulation. This change was very
te
pronounced with the contractile type of COX6AH (~23 fold) and even more significant with
Ac ce p
COX7AH (~90 fold) (Fig. 4A, 4B). After recovery however, the expression level of these transcripts returned to nearly normal level compared to age matched control muscle (Figure 4A, 4B). The gene expression profile of other mitochondrial encoded transcripts such as MTRNR1 (12S rRNA) or MT-ND1 did not increase and the level of MT-CO2 was significantly down-regulated in early RIRCD muscle explaining the observed lack of COX2 protein in RIRCD muscle tissue (Tritschler et al., 1991)(Fig. 4C). Tissue sections obtained from muscle biopsies of a patient with RIRCD were also examined by immune-labelling with isoform specific antibodies. Immunofluorescence analysis showed that during the early phase, none of the liver or the heart type isoforms of COX6A and COX7A were expressed normally in RIRCD skeletal muscle (Fig. 4D). Interestingly, only
Page 13 of 32
14
COX7AH was detectable which demonstrated a rather unusual appearance when compared to age matched control skeletal muscle sections. The expression of this isoform was decreased, but it showed distinct aggregates in some of the skeletal muscle fibres while no
ip t
abnormalities were observed with COX4 antibody (Fig. 4F). Double staining with myosin light chain antibody revealed that these clusters were mainly present in the oxidative type I
cr
fibres (Fig. 4G). Immuno-labelling of the follow-up muscle biopsy after recovery showed
us
normalised expression and localisation of these nuclear encoded COX subunits when compared to control muscle sections (Fig. 4E). Based on the data collected from muscle
an
biopsies of RIRCD patients we can conclude that the presented shift in isoforms through development cannot explain the reversibility of the disorder. Most likely the absence of the
M
mtDNA encoded COX subunits is caused by abnormal translation of mtDNA-encoded proteins triggered by the mt-tRNAGlu mutation. The reduced mitochondrial encoded COX
d
subunits further compromise COX assembly, which results in secondary impairment of the
te
nuclear encoded COX6A and COX7A subunits.
Ac ce p
3.6. What is the effect of blocking mitochondrial translation in RIRCD? In order to investigate whether the low steady–state level of mt-tRNAGlu in RIRCD happens during processing/maturation of the mt-tRNA or during protein synthesis, we blocked mitochondrial translation with TAP in myoblasts of an RIRCD patient (P14 in Horvath et al., 2009) and in a control cell line. We hypothesized that if the steady state level of mt-tRNAGlu remains low after blocking mitochondrial translation, it would indicate that the instability happens during processing/maturation of the mt-tRNA, however if the originally low tRNAGlu steady state returns to normal or increases, this would indicate that the mt-tRNA instability occurs during mitochondrial protein synthesis. Interestingly, blocking mitochondrial
Page 14 of 32
15
translation resulted in an increase of the tRNA steady state (Fig. 5), suggesting that the mutated mt-tRNA may cause a higher level of failed translation events, stimulating its decay, which may affect the function and assembly of mitoribosomes. The elevated level of mt-
ip t
tRNATrp, but no change in the expression of the cytoplasmic tRNALys further suggested a
cr
more general suppression of mitochondrial protein synthesis in RIRCD in vitro (Fig. 5).
us
4. Discussion
The main finding of the present study is that we define the specific time frame when the
an
COX6A and COX7A isoform switch occurs in mice and most importantly in human skeletal muscle during postnatal development. Previous studies in mice demonstrated that the liver
M
type isoforms are predominant in fetal tissues (Bonne et al., 1993), but these are gradually replaced by the heart type of isoforms in both mouse cardiac and skeletal muscle (Kim et al.,
d
1995). Here we show that the COX isoform switch in human muscle takes place by the age
te
of 3 months, as illustrated by the predominant expression of muscle or heart isoforms.
Ac ce p
It has been hypothesized previously that the developmental isoform switch of COX6A and COX7A isoforms may be the cause of spontaneous clinical recovery in RIRCD (Tritschler et al., 1991; Taanman et al., 1993). Before the discovery of the genetic cause of RIRCD, a hypothesis by Taanman (1993) has been put forward to explain how the reversibility may occur in RIRCD, suggesting a defect in the transcriptional regulatory mechanisms in the liver type subunits. However, based on our data the gene expression of both liver and heart type isoforms has dramatically increased during the symptomatic phase in patient’s muscle, possibly due to compensatory mechanisms, trying to correct reduced protein expression. Although we show a developmental switch in the studied nuclear COX isoforms around 3 months of age, just shortly prior to the spontaneous recovery in RIRCD (~6 months), our
Page 15 of 32
16
data suggest that it is most likely not causative for the improvement of clinical symptoms in RIRCD. Sequence analysis of the human COX6A and COX7A isoforms revealed 66.3% and 59.7% homology respectively (Ma et al., 2014, Peng and Xu, 2011a, 2011b, Suppl. Fig. 2.)
ip t
Discrepancies between the liver and muscle type of isoforms therefore could possibly indicate functional differences which might be important in mitochondrial diseases. The
cr
m.14674T>C/G mutation most likely impairs mitochondrial translation, as reflected by the
35
us
RRF/COX-negative fibres and the multiple RC defect in skeletal muscle. Previously however S-methionine pulse labelling in 2 patient cell lines carrying the m.14674T>C mutation
an
demonstrated normal or even slightly increased mitochondrial protein synthesis (Horvath et al., 2009). In parallel, immunoblotting showed normal or even higher mitochondrial protein
M
levels suggesting an in vitro compensatory mechanism. Northern blots for the expression of the mitochondrial encoded protein-coding genes also detected normal levels suggesting
d
that mitochondrial transcription remains intact (Horvath et al., 2009). Based on our data we
te
suggest that normal mitochondrial translation in RIRCD cells is kept by a significant effort
Ac ce p
through utilisation of cellular compensatory mechanisms. If an additional factor negatively alters mitochondrial translation, cellular compensation may not cope with the increased need which results in a severely impaired or stalled mitochondrial translation, probably leading to a chain of other events resulting in severe mitochondrial dysfunction. The increase in steady state levels of the mt-tRNAGlu after thiamphenicol treatment indicates that processing and maturation of the mt-tRNAGlu is stable and not altered by the m.14674T>C/G mutation. Interestingly, the role of COX6AH as an isoform required for the overall stability of the COX holoenzyme after assembly has also been recognised in the past and could further explain the predominant COX defect in patients with RIRCD (Fornuskova et al., 2010).
Page 16 of 32
17
Additional factors such as environmental contributions, epigenetic factors and the presence of other genetic polymorphisms may ultimately modify the nature of the disease phenotype and also have to be considered to explain reversibility in these conditions (Boczonadi et al.,
ip t
2013). Recently our group suggested that altered thio-modification of the mt-tRNAGlu due to low infantile cysteine concentration could be a potential explanation for the timed
cr
disease manifestation. Cysteine is an essential amino acid in infants, because the
us
cystathionine gamma-lyase (cystathionase) enzyme (rate-limiting enzyme for the synthesis of L-cysteine from L-methionine) appears to be low after birth and regulated at the post-
an
transcriptional level during development in the first months of life (Levonen et al., 2000). Supplementation of L-cysteine, that is essential for normal thio-modification of mt-tRNAGlu,
M
mt-tRNALys and mt-tRNAGln by TRMU, significantly improved most RC complex activities in TRMU and RIRCD patient cell lines. These data suggest that low dietary L-cysteine levels in
d
infants could act as a second hit on mitochondrial translation, explaining clinical
te
manifestations of RIRCD and TRMU deficiency, and an increase in L-cysteine due to
Ac ce p
cystathionase activation through development would contribute to the recovery in these patients (Boczonadi et al., 2013).
Conclusion
Here we show experimental evidence for an isoform switch of COX6A and COX7A in skeletal muscle, which happens around 3 month of age. However, in contrast to previously suggested hypotheses, we could not confirm a causative link between the isoform switch of COX subunits and the clinical recovery in RIRCD our data may have implication for severe infantile presentations of other mitochondrial myopathies.
Page 17 of 32
18
Acknowledgments The work is supported by the Medical Research Council (UK) (G1000848), the European Research Council (309548) and the Mitochondrial European Educational Training (MEET),
ip t
ITN MARIE CURIE PEOPLE, (317433). We are grateful to the Medical Research Council (MRC)
us
supporting this project and for providing primary human cells.
cr
Centre for Neuromuscular Diseases Biobank Newcastle and for the EuroBioBank for
M
The authors declare no conflict of interests
an
Conflict of interest statement
References
d
1. Boczonadi V, Smith PM, Pyle A, Gomez-Duran A, Schara U, Tulinius M, Chinnery PF,
te
Horvath R. Altered 2-thiouridylation impairs mitochondrial translation in reversible
Ac ce p
infantile respiratory chain deficiency. Hum Mol Genet 2013;22(22):4602-15.
2. Bonne G, Seibel P, Possekel S, Marsac C, Kadenbach B. Expression of human cytochrome c oxidase subunits during fetal development. Eur J Biochem 1993;217(3):1099-107.
3. Chrzanowska-Lightowlers ZM, Preiss T, Lightowlers RN. Inhibition of mitochondrial protein synthesis promotes increased stability of nuclear-encoded respiratory gene transcripts. J Biol Chem 1994;269(44):27322-8.
Page 18 of 32
19
4. Fornuskova D, Stiburek L, Wenchich L, Vinsova K, Hansikova H, Zeman J. Novel insights into the assembly and function of human nuclear-encoded cytochrome c oxidase subunits 4, 5a, 6a, 7a and 7b. Biochem J 2010; 428(3):363-74
ip t
5. Greaves LC, Reeve AK, Taylor RW, Turnbull DM. Mitochondrial DNA and disease. J
cr
Pathol 2012;226(2):274-86.
us
6. Grossman LI, Lomax MI. Nuclear genes for cytochrome c oxidase. Biochim Biophys Acta 1997;1352(2):174-92.
an
7. Horvath R, Kemp JP, Tuppen HA, Hudson G, Oldfors A, Marie SK, Moslemi AR, Servidei S, Holme E, Shanske S, et al. Molecular basis of infantile reversible
M
cytochrome c oxidase deficiency myopathy. Brain 2009;132,3165-3174.
d
8. Huang X. and Miller W. A Time-Efficient, Linear-Space Local Similarity Algorithm.
te
Advances in Applied Mathematics, vol. 12 (1991), pp. 337-357.
Ac ce p
9. Hüttemann M, Kadenbach B, Grossman LI. Mammalian subunit IV isoforms of cytochrome c oxidase. Gene 2001;267(1):111-23.
10. Hüttemann M, Schmidt TR, Grossman LI. A third isoform of cytochrome c oxidase subunit VIII is present in mammals. Gene 2003;312:95-102.
11. Kadenbach B, Hüttemann M, Arnold S, Lee I, Bender E. Mitochondrial energy metabolism is regulated via nuclear-coded subunits of cytochrome c oxidase. Free Radic Biol Med 2000;29(3-4):211-21.
Page 19 of 32
20
12. Kim K, Lecordier A, Bowman LH. Both nuclear and mitochondrial cytochrome c oxidase mRNA levels increase dramatically during mouse postnatal development. Biochem J 1995;306(Pt 2):353-8.
ip t
13. Källberg M, Wang H, Wang S, Peng J, Wang Z, Lu H, Xu J. Template-based protein
cr
structure modelling using the RaptorX web server. Nat Protoc. 2012;7(8):1511-22.
us
14. Levonen AL, Lapatto R, Saksela M, Raivio KO. Human cystathionine gamma-lyase: developmental and in vitro expression of two isoforms. Biochem J 2000;347(Pt
an
1):291-5.
15. Lochmüller H, Johns T, Shoubridge EA. Expression of the E6 and E7 genes of human
M
papillomavirus (HPV16) extends the life span of human myoblasts. Exp Cell Res
d
1999;248(1):186-93.
te
16. Ma J, Wang S, Zhao F, and Xu J. Protein threading using context-specific alignment potential. Bioinformatics (Proceedings of ISMB) 2013;29(13):i257-i265.
Ac ce p
17. Mimaki M, Hatakeyama H, Komaki H, Yokoyama M, Arai H, Kirino Y, Suzuki T, Nishino I, Nonaka I, Goto Y. Reversible infantile respiratory chain deficiency: a clinical and molecular study. Ann Neurol 2010;68(6):845-54.
18. Peng J. and Xu J. A multiple-template approach to protein threading. Proteins, 2011;79(6):1930-9.
19. Peng J. and Xu J. RaptorX: exploiting structure information for protein alignment by statistical inference. Proteins, 2011;79(S10):161-71.
Page 20 of 32
21
20. Rizzuto R, Nakase H, Darras B, Francke U, Fabrizi GM, Mengel T, Walsh F, Kadenbach B, DiMauro S, Schon EA. A gene specifying subunit VIII of human cytochrome c oxidase is localized to chromosome 11 and is expressed in both muscle and non-
ip t
muscle tissues. J Biol Chem 1989;25;264(18):10595-600.
cr
21. Shoubridge EA. Cytochrome c oxidase deficiencies. Am J Med Genet. 2001
us
Spring;106(1):46-52. Review.
22. Taanman JW, Hall RE, Tang C, Marusich MF, Kennaway NG, Capaldi RA. Tissue
an
distribution of cytochrome c oxidase isoforms in mammals. Characterization with monoclonal and polyclonal antibodies. Biochim Biophys Acta 1993;1225:95-100.
M
23. Tritschler HJ, Bonilla E, Lombes A, Andreetta F, Servidei S, Schneyder B, Miranda AF,
d
Schon EA, Kadenbach B, DiMauro S. Differential diagnosis of fatal and benign
te
cytochrome c oxidase-deficient myopathies of infancy: an immunohistochemical
Ac ce p
approach. Neurology 1991;41(2 ( Pt 1)):300-5. 24. Uusimaa J, Jungbluth H, Fratter C, Crisponi G, Feng L, Zeviani M, Hughes I, Treacy EP, Birks J, Brown GK, et al. Reversible infantile respiratory chain deficiency is a unique, genetically heterogeneous mitochondrial disease. J Med Genet 2011;48:660-668.
Page 21 of 32
22
Figure legends Figure 1. COX6A and COX7A subunit isoforms colocalise with MTCO1 in mouse skeletal muscle. Double immune-labelling of muscle sections. Merged images are shown in the last
ip t
column including nuclear staining (DAPI blue channel). (A). COX6AL (green channel)
cr
colocalisation with MTCO1 (red channel). (B). COX6AH (green channel) colocalisation with MTCO1 (red channel). (C). COX7AL (green channel) colocalisation with MTCO1 (red channel).
us
(D). COX7AH (green channel) colocalisation with MTCO1 (red channel). Scale bar represents:
an
20 micron.
M
Figure 2. Expression of COX6A and COX7A isoforms during mouse development. (A). Immunofluorescence micrographs of mouse skeletal muscle sections stained with
d
COX6AL, COX6AH, COX7AL or COX7AH (green channel) with DAPI (blue channel).
te
Developmental series are shown from the top to bottom (1 day, 8 days, 2 weeks, 4 weeks and adult). Scale bar represents 20 micron. (B). Western blot analysis of COX6A and COX7A
Ac ce p
subunit isoforms through development. Porin was used as a loading control.
Figure 3. Gene expression of COX6A and COX7A subunit isoforms through development in mice and human. Analysis of gene expression was performed by quantitative RT-PCR at different ages. The graphs represent the average of triplicate wells in three independent experiments with standard deviation shown as error bars. (A) Expression pattern of the liver and contractile type of COX6A in mouse skeletal muscle. (B) Gene expression of the liver and muscle type of COX7A in mouse skeletal muscle. (C-D) Expression profiles of COX6A and COX7A subunit isoforms in murine liver and (E-F) heart. (G-H) Biochemical measurement of
Page 22 of 32
23
citrate sythase and complex IV activity in the mouse muscle, heart and liver at different ages. Measurements were carried out as triplicates (n=3). (I-J) Relative quantification of the COX6A and COX7A isoforms in human skeletal muscle at different ages. Mouse samples
ip t
were normalised to actin and vimentin; human samples were normalised to actin and GAPDH. Data are presented as mean ± standard deviation. A p value of <0.05 was
us
cr
considered significant.
Figure 4. Expression pattern of COX6A and COX7A isoforms in skeletal muscle of patients
an
with RIRCD in the early symptomatic phase and after recovery. (A,B) Graphs are representing COX6A and COX7A isoform gene expression in control human samples at
M
different ages (3 months, 1 year 6 months, adult) and in RIRCD muscle before (1 month) and after recovery (8 years 9 months) (marked by a red line). (C) Gene expression of MT-CO2,
d
MTRNR1 (12S rRNA) and MT-ND1 in control and RIRCD muscle. The graphs represent the
te
average of triplicate wells in three independent experiments with standard deviation shown
Ac ce p
as error bars. (D) Immunofluorescence labelling of 1 month old human control muscle section with antibodies against COX6A and COX7A subunit isoforms (first column). 1 month old RIRCD muscle sections stained for COX6A and COX7A subunit isoforms (second column). (E) Adult human control muscle section (first column) and RIRCD muscle sections (second column) after recovery, labelled with COX6A and COX7A subunit isoforms. (F) 1 month old control and 1 month old RIRCD muscle sections labelled with COX7AH (green channel) and COX4 (red channel). White arrow heads indicate fibres with COX7AH clusters. (G) Control and RIRCD muscle sections at 1 month of age were co-labelled with COX7AH (black and white) and Myosin light chain Type I. (green channel). Scale bar represents 20 micron.
Page 23 of 32
24
Figure 5. Gene expression analysis of mt-tRNAGlu, mt-tRNATrp and cyt-tRNALys in control and patient myoblasts after blocking mitochondrial translational with Thiamphenicol
ip t
Analysis of relative gene expression was performed by quantitative RT-PCR. The graphs
cr
represent the average of triplicate wells in three independent experiments with standard deviation shown as error bars. CNTR: control immortalised myoblasts, P.: patient
us
immortalised myoblasts.
Ac ce p
te
d
M
an
Table. Laboratory investigations and muscle biopsy findings of the patient with RIRCD.
Page 24 of 32
25
Table. Laboratory investigations and muscle biopsy findings of the patient with RIRCD. Laboratory results
Muscle histology RRF/COX- Lipid Glycogen Other +
-
abnormal mt
COX ↓↓↓
8y 9m: +
-
-
mild myopathy
norm
CK
norm
te
d
M
an
us
normal
CI ↓↓↓
ip t
1m: +++
cr
Lactate
Ac ce p
Patient
Biochemical activity of the respiratory chain enzymes
Page 25 of 32
cr
i
Graphical Abstract (for review)
COX6AL
COX6AH
COX6AH
M an
NEW BORN
us
COX6AL
COX7AH
ce pt
COX7AL
ed
SKELETAL MUSCLE DEVELOPMENT
ADULT
COX7AL
COX7AH
Ac
~ 3 months
L-isoform ubiquitously expressed (embryonic isoform)
H-isoform expressed in mature contractile tissues (muscle/heart isoform)
Page 26 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 1
Page 27 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 2
Page 28 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 3
Page 29 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 4
Page 30 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 5
Page 31 of 32
Table
Table. Laboratory investigations and muscle biopsy findings of the patient with RIRCD. Biochemical activity of the respiratory chain enzymes
Lactate CK normal
1m: +++
+
-
COX ↓↓↓
CI ↓↓↓
8y 9m: +
-
-
norm
norm
ce pt
ed
M
an
us
cr
abnormal mt mild myopathy
ip t
Muscle histology RRF/COX- Lipid Glycogen Other
Ac
Patient
Laboratory results
Page 32 of 32
S1357-2725(15)00035-7 http://dx.doi.org/doi:10.1016/j.biocel.2015.01.025 BC 4550
To appear in:
The International Journal of Biochemistry & Cell Biology
Received date: Revised date: Accepted date:
6-11-2014 17-1-2015 29-1-2015
Please cite this article as: Boczonadi, V., Giunta, M., Lane, M., Tulinius, M., Schara, U., and Horvath, R.,Investigating the role of the physiological isoform switch of cytochrome c oxidase subunits in reversible mitochondrial disease, International Journal of Biochemistry and Cell Biology (2015), http://dx.doi.org/10.1016/j.biocel.2015.01.025 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Investigating the role of the physiological isoform switch of cytochrome c oxidase subunits in reversible mitochondrial disease Veronika Boczonadi1 , Michele Giunta1, Maria Lane1, Mar Tulinius2, Ulrike Schara3 and Rita
cr
Institute of Genetic Medicine, Wellcome Trust Mitochondrial Research Centre, Newcastle
us
1
ip t
Horvath1
University, Central Parkway NE1 3BZ Newcastle upon Tyne, UK;
Department of Paediatrics, The Sahlgrenska Academy, University of Gothenburg, Box 400,
an
2
Göteborg, SE-405 30 Sweden;
Department of Paediatric Neurology, University of Essen, Hufelandstraße 55, Essen, 45122
M
3
Word count: 3691 words
Ac ce p
Abstract: 229 words
te
d
Germany
Title: 127 characters including spaces and punctuation Keywords: reversible infantile respiratory chain deficiency, cytochrome c oxidase, isoform switch, mitochondrial tRNAGlu
Address correspondence to: Rita Horvath MD, PhD Institute of Genetic Medicine, Newcastle University, Central Parkway, Newcastle upon Tyne, NE1 3BZ, UK, Phone: +44 191 2418855, Fax: +44 191 2418666 Email: [email protected]
Page 1 of 32
2
HIGHLIGHTS We show a developmental isoform switch of COX6A/COX7A in muscle of humans
ip t
and mice.
cr
Liver isoforms are present at birth, heart/muscle isoforms increase ~3 months of
us
age.
The physiological isoform switch does not explain reversible mitochondrial disease.
Abstract
te
d
M
an
Developmental changes of COX isoforms may modify other mitochondrial diseases.
Ac ce p
Reversible infantile respiratory chain deficiency is characterised by spontaneous recovery of mitochondrial myopathy in infants. We studied, whether a physiological isoform switch of nuclear cytochrome c oxidase subunits contributes to the age-dependent manifestation and spontaneous recovery in reversible mitochondrial disease. Some nuclear-encoded subunits of cytochrome c oxidase are present as tissue specific isoforms. Isoforms of subunits COX6A and COX7A expressed in heart and skeletal muscle are different from isoforms expressed in liver, kidney and brain. Furthermore, in skeletal muscle both the heart and liver isoforms of subunit COX7A have been demonstrated with variable levels, indicating that the tissuespecific expression of nuclear-encoded subunits could provide a basis for the fine-tuning of cytochrome c oxidase activity to the specific metabolic needs of the different tissues.
Page 2 of 32
3
We demonstrate a developmental isoform switch of COX6A and COX7A subunits in human and mouse skeletal muscle. While the liver type isoforms are more present soon after birth, the heart/muscle isoforms gradually increase around 3 months of age in infants, 4 weeks of
ip t
age in mice, and these isoforms persist in muscle throughout life. Our data in follow-up biopsies of patients with reversible infantile respiratory chain deficiency indicate that the
cr
physiological isoform switch does not contribute to the clinical manifestation and to the
us
spontaneous recovery of this disease. However, understanding developmental changes of the different cytochrome c oxidase isoforms may have implications for other mitochondrial
Ac ce p
te
d
M
an
diseases.
Page 3 of 32
4
1. Introduction The mitochondrial oxidative phosphorylation (OXPHOS) machinery is composed of 5
ip t
multimeric complexes (complex I, II, III, IV and V) embedded in the inner mitochondrial membrane (Greaves et al., 2012). Under normal physiological conditions, mitochondrial
cr
oxidation is responsible for the majority of energy production in most cells and tissues. The
us
utilisation of oxygen by the mitochondria to generate ATP is regulated by the cytochrome c oxidase enzyme (COX, complex IV) that is the terminal enzyme of the mitochondrial electron
an
transport chain. The formation of the functional COX is under the control of two separate genetic systems, the nuclear genome (nDNA) and the mitochondrial genome (mtDNA)
M
coding for 10 and 3 subunits respectively (Shoubridge, 2001).
d
Within the OXPHOS system a unique feature of COX is the presence of subunits with tissue
te
specific isoforms. It was shown in early studies, that some, but not all nuclear-encoded subunits of COX are present as tissue specific isoforms (Taanman et al., 1993). At least five
Ac ce p
of the COX subunits have been reported with such isoforms suggesting a regulatory function in energy production (COX4I1/COX4I2, COX6A1/COX6A2, COX6B1/Cox6B2, COX7A1/COX7A2 and COX8A/COX8B/COX8C) (Huttemann et al., 2001; Kadenbach et al., 2000). In mammals three of these isoform pairs correspond to muscle and non-muscle specific forms: COX6A, COX7A and COX8 (Grossman and Lomax, 1997). Isoforms of COX4I and COX6B, represent specific expression in the lung and in the testes respectively (Huttemann et al., 2003). One isoform of subunit COX6A and COX7A is expressed in heart and skeletal muscle (heart/skeletal muscle isoform) while the other is expressed in liver, kidney and brain (liver or non-muscle type). Interestingly, in humans there is no evidence of a contractile muscle and non-muscle specific isoforms of subunit COX8A, as oppose to other mammals and birds
Page 4 of 32
5
(Rizzuto et al., 1989). It has been postulated in the past that the tissue-specific expression of nuclear-encoded subunits could provide a foundation for the fine-tuning of COX activity to the specific metabolic needs in different tissues (Taanman et al., 1993). Understanding basic
ip t
mechanisms of COX deficiencies raised interest in studying the development of nuclear COX
cr
subunits.
The potential relevance of the different tissue specific COX isoforms in manifestation of
us
mitochondrial disease has been hypothesized over 20 years ago (Tritschler et al., 1991). The clinical description of a unique, reversible mitochondrial disease, reversible infantile COX
an
deficiency myopathy (recently named as reversible infantile respiratory chain deficiency,
M
RIRCD) caused by a homoplasmic mt-tRNAGlu mutation (m.14674T>C/G), raised the possibility, that developmental factors may have a role in the clinical manifestation
d
(Horvath et al., 2009, Mimaki et al., 2010, Uusimaa et al., 2012). While most mitochondrial
te
diseases are severe, progressive conditions, RIRCD is one of the few hereditary diseases with a life-threatening onset showing a remarkable reversible disease course within the first year
Ac ce p
of life. Before the identification of the primary genetic cause of this condition immunohistochemistry was suggested to distinguish between the fatal and the benign forms of COX deficiency (Tritschler et al., 1991). It has been hypothesized previously that the nuclear-encoded COX6A and COX7A subunits undergo a developmental isoform switch from fetal to ubiquitous isoforms in early childhood and it may explain the disease recovery in reversible infantile myopathy (Tritschler et al., 1991; Taanman et al., 1993). Up to date no experimental evidence has been shown in patients carrying the m.14674T>C/G variant, that a switch between isoforms would underlie the recovery.
Page 5 of 32
6
We suggested that tissue-specific, developmentally timed processes play a role both in the age-dependent expression and in the reversibility in RIRCD. In this study we investigated the physiological isoform switch of nuclear COX subunits in normal human skeletal muscle
ip t
(infant, adult) and studied whether it contributes to the recovery in RIRCD.
cr
2.Materials and Methods
us
2.1. Immunofluorescence
an
OCT embedded mouse and human muscle tissues were sectioned. Dried sections were fixed with 4% PFA for immune-staining. The primary antibodies were: Anti-Cytochrome c
M
antibody (Abcam), COX6AL (Mitosciences), COX6AH (Abgent), COX7AL (Santa-Cruz), and COX7AH (Proteintech), MTCO1 (Abcam), Myosin Light chain (Millipore), Alexa fluor 488 and
d
596 conjugated secondary antibodies (Invitrogen) were used to detect the primary
te
antibodies. Cell nuclei were identified using DAPI. Immunofluorescence images were
Ac ce p
collected using a Zeiss Axioimager Z1 fluorescence microscope equipped with Zeiss Apotome 2 (Zeiss, Germany). Acquired images were processed with AxioVision Rel 4.9 software.
2.2. Immunoblotting
Muscle tissues were lysed and the extracts were cleared by centrifugation and stored at -80 C until use. The extracts were boiled in 2X Laemmli sample buffer. Samples were than subjected to SDS-polyacrylamide gel electrophoresis followed by immunoblot analysis using specific antibodies raised against COX6AL (Mitosciences), COX6H (Abgent), COX7AL (SantaCruz), COX7AH (Proteintech), Porin (Abcam). Horseradish peroxidase-conjugated secondary
Page 6 of 32
7
antibodies (Jackson ImmunoResearch Laboratories) were used for detection applying the enhanced chemiluminescence method (GE Healthcare Bio-Sciences).
ip t
2.3. Isolation of RNA and cDNA Synthesis Total RNA was prepared from mouse muscle, liver, heart and human muscle by standard
cr
phenol/chlorophorm method. Reverse transcription was carried out using 1 ug of total RNA, SuperScript II (Invitrogen). For isoform-specific PCRs, primers were designed to be specific to
us
unique sequences of the isoform transcripts. cDNA input was standardized and RT-PCRs
an
were performed with SYBR Green PCR master mix (Applied Biosystems). Samples were normalized to Actin and Vimentin in mouse, Actin and GAPDH in humans, and fold change
M
was determined by the ΔΔCt method. We used the following primers to study gene expression in mouse tissues: COX6AL forward: 5’-GCTCAACGTGTTCCTCAA-3’, reverse: 5’-
d
CGGAAGTGGGTTCACATGAG-3’; COX6AH forward: 5’-TGCCTCTAAAGGTCCTGAGC-3’, reverse:
te
5’-GAAGAGCCAGCACAAAGGTC-3’; COX7AH forward: 5’-GACAATGACCTCCCAGTACAC-3’,
3’,
Ac ce p
reverse: 5’-CAGCCCAAGCAGTATAAGCA-3’; COX7AL forward: 5’- CAGTTCATCTGAAAGGCCGGreverse:
5’-ACGACATTAGTTCTGCTTCTTGG-3’;
Vimentin:
forward
5’-
CAGCAGTATGAAAGCGTGG-3’, reverse: 5’-GGAAGAAAAGGTTGGCAGAG-3’. The following human primers were used: MTRNR1 (12S rRNA) forward: 5’-AAACTGCTCGCCAGAACACT-3’, reverse: 5-‘CATGGGCTACACCTTGACCT-3’; COX6AH forward: 5’-CGCCCCGAGTTCCGTCCCTA3’,
reverse:
5’-GGGCAGAGGGTTCACGTGGC-3’;
COX6AL
forward:
5’-
GCTGAATGTGTACCTGAAGTC-3’ , reverse:5’-TGAGGGTTATGGAATAGAGTATGG-3’; COX7AH 5’-AATGACATCCCGTTGTACCT-3’,
reverse:
5’-GGCTTCTTGGTCTTAATTCCT-3’;
COX7AL:
forward: 5’-CCCTCCTGTATAGAGCCACC-3’, reverse: 5’-TGAATGAAACTGAACCAAGCGA-3’, Beta
Actin:
forward:
5’-
GATGCAGAAGGAGATCACTGC-3’,
reverse:
5’-
Page 7 of 32
8
ACATCTGCTGGAAGGTGGAC-3’; MT-ND1 forward: 5’-ATGTCCACCCTTATCACTCA-3’, reverse: 5’-TATATCTGAGGTGGTAGGC-3’; MT-COX2 forward: 5’-CCATCCCTACGCATCCTTTA-3’, reverse: mt-tRNATrp
5’-GCCGTAGTCGGTGTACTCGT-3’;
forward:
5’-
mt-tRNAGlu
forward:
TTCTCGCACGGACTACAACC-3’;
5’-ACAACGATGGTTTTTCATATCATT-3’, cyt-tRNALys
forward:
reverse:
5’-
5’-GTCGGTAGAGCATCAGACTT-3’,
cr
3’;
ip t
GAAATTTAGGTTAAATACAGACCAAGA-3’, reverse: 5’- GAAATTAAGTATTGCAACTTACTGAGG-
3',
reverse:
5'-TCTAGACGGCAGGTCAGGTC-3',
us
reverse: 5’-GAACAGGGACTTGAACCCTG-3’; GAPDH forward: 5'-ACCCAGAAGACTGTGGATGGRN5S1@
(5S
rRNA)
forward:
5’-
an
GTCTACGGCCATACCACCCTG-3’, reverse: 5’-AAAGCCTACAGCACCCGGTAT-3’.
M
2.4. Cell culture
Myoblast cell cultures of an RIRCD patient and controls were obtained from the Biobank of
d
the Medical Research Council, Centre for Neuromuscular Diseases, Newcastle and were
te
immortalized as described (Lochmüller et al., 1999). Informed consent was signed by all
Ac ce p
subjects and ethical approval was obtained from a local research ethics committee. Muscle cells were grown in skeletal muscle growth medium (PromoCell, Heidelberg, Germany), supplemented with 4mM L-glutamine and 10% foetal bovine serum and cultured as recommended by the supplier.
Thiamphenicol (TAP) was dissolved in ethanol, the concentration was calculated spectrophotometrically (ϵmM=0.64 at 273 nm) and administered at 50 tionl−1 final concentration. Control cells were supplemented with an equal volume of ethanol (Chrzanowska-Lightowlers et al., 1994). Control and patient immortalised myoblasts were washed and reseeded in media containing TAP. Cells were left to grow for 6 days. RNA was
Page 8 of 32
9
prepared from untreated and TAP-treated control and patient cells. Small RNA was isolated from the cells with miRNA Isolation kit (Ambion).
ip t
2.5. Statistical analysis The statistical package SigmaPlot 11.0 was used to perform all the statistics. Data are
cr
presented as mean ± standard deviation. The normal distribution was checked by the
us
Kolmogorov-Smirnov test and ANOVA tests were used to compare parameters. The HolmSidak method was used for pair-wise multiple comparison procedures. A p value of <0.05
an
was considered significant.
M
2.6 Biochemical Activity Assays
Three cycles of freeze-thaw were performed on each sample prior to each assay. Each assay
d
was performed on a Multiskan Ascent 96/384 Plate Reader (ThermoFisher Scientific, UK).
te
For the Citrate Synthase activity measurement 7.5 uL of sample was added to a 96 well plate
Ac ce p
in triplicate. 193.38 uL of citrate synthase buffer (0.1 mM Tris-HCl), 2 uL DTNB 10 mM, 2 uL Triton-X 100 10% w/v and 0.62 uL acetyl Co-A 32 mM were added to each well. 1 uL of oxaloacetate 50 mM was added to each well as the initiator and the assay run at 405 nm. The activity was calculated using the following equation: Activity = Absorbance/13.6 x Vol (total)/Vol (sample)
x 1000.
For the Complex IV activity measurements 7.5 uL of sample was added to a 96 well plate in triplicates. 195 uL of reduced cytochrome c 0.1% (freshly prepared) was added to each well as the initiator and the assay run at 550 nm. The activity was calculated using the following equation: Activity = Absorbance/18.7 x Vol (total)/Vol (sample) x 1000. In all cases triplicate assays were performed. Protein determination was performed according to Bradford.
Page 9 of 32
10
2.7 Protein structure analysis Protein sequence homology was analised with SIM-Alignment Tool (Huang et al., 1991). Predicted Protein 3D structure was completed on RaptorX structure prediction server
cr
ip t
(Kallberg et al., 2012).
us
3. Results
an
3.1. Protein expression of COX6A and COX7A isoforms through mouse development In murine skeletal muscle sections we conducted immune-staining and immunoblotting to
M
determine the level of protein expression of both liver and heart type isoforms of COX6A and COX7A through development. To confirm the mitochondrial localisation of all four
d
isoform specific antibodies, muscle sections were co-stained with cytochrome c oxidase
te
subunit 1 (MTCO1) as a mitochondrial marker (Fig. 1). Immuno-fluorescent assessment of
Ac ce p
the skeletal muscle sections confirmed the specificity of the antibodies against special isoforms as they co-localised with MTCO1 (Fig. 1 A-D). Frozen sections of normal skeletal muscle were collected from 1 day, 8 days, 2 weeks, 4 weeks old and adult mice. One day old mice indicated strong, distinct mitochondrial signals for the liver type isoforms of both COX6A and COX7A subunits (Fig. 2). From days 8, the liver type isoforms gradually decreased over 2 and 4 weeks and on the adult sections no staining was observed. On the other hand, the signal for the muscle/heart type isoforms seemed to be very low in 1 day old mice and increased through the developmental series, moreover the COX7AH isoform indicated higher expression in adult sections. To confirm the shift in the liver and heart type isoforms in mouse skeletal muscle, we resolved the whole protein
Page 10 of 32
11
lysate extracted from muscle tissue on SDS-PAGE. The liver type isoforms were only detectable before 2 weeks of age. Expression of both COX6AL and COX7AL was observed until 2 weeks of age then a dramatic drop in expression level was observed. In parallel
ip t
COX6AH and COX7AH expression appeared and remained highly expressed through adult
cr
age.
us
3.2. Protein expression pattern of COX6A and COX7A subunit isoforms in human muscle We have studied human skeletal muscle biopsies of children at 3 months of age and in
an
adults. Immuno-labelling detected that at 3 months of age both isoforms of COX6A and COX7A subunits were present however the muscle/heart specific isoforms were
M
predominantly expressed in the adult tissue (Suppl. Fig. 1). These data are fully in line with
the early months of life.
d
the protein expression pattern seen in mouse muscle, confirming the isoform switch within
te
3.3. Gene expression of COX6A and COX7A subunit isoforms during development in mice
Ac ce p
We analysed the changes in gene expression of COX6AL/COX6AH and COX7AL/COX7AH in a variety of mouse and human tissues by RT-PCR. As endogenous internal reference genes in mice we studied actin and vimentin. Total RNA was isolated from mouse muscle at different ages (1 day, 8 days, 2 weeks, 4 weeks and 1 year). The relative expression of COX6AL and COX6AH isoforms were not significantly different in the early ages (1 day, 8 day, 2 weeks) however, from the age of 4 weeks a dramatic difference was observed in gene expression of several genes. COX6AH expression increased ~13 fold and continued to rise at the age of 1 year (Fig. 3A). A very similar expression pattern was observed with COX7A subunit isoforms where the COX7AH isoform suddenly increased ~13 times fold at around 1 month of age and
Page 11 of 32
12
reached a 16 fold change at the age of 1 year (Fig. 3B). This dramatic shift in gene expression of tissue specific isoforms was not observed in liver or in heart tissues (Fig. 3B3C). Nevertheless, both liver type isoforms (COX6AL/COX7AL) showed significantly higher
ip t
levels in the liver tissues when compared to heart type isoforms (Fig. 3C-D), while the mouse heart tissues demonstrated that COX6AH and COX7AH are the prominent isoforms in
cr
heart also later in life (Fig. 3E-F). We also measured the biochemical activity of the citrate
us
synthase and complex IV in mouse tissues at different ages. No significant changes were detected in the activity of citrate synthase in the muscle and the heart while slightly lower
an
activity was detected in the liver around the age of 2 weeks (Fig. 3G). Interestingly, the complex IV activity demonstrated a slight increase in the muscle and heart tissues around 1
M
week of age after which the activity returned to similar levels measured at 0 day (Fig. 3H).
te
activity of the COX holoenzyme.
d
Based on these results the COX6A and COX7A isoform change does not affect the overall
3.4. Gene expression profiles of COX6A and COX7A isoforms in normal human skeletal
Ac ce p
muscle at different ages
We further investigated the skeletal muscle specific isoform shift in human muscle. Here we have collected muscle tissues from normal human controls at different ages (21 days, 3 months, 1 year 6 months, 5 years and adult). Our aim was to pinpoint the window of time when this shift occurs in humans (Fig. 3I-J). As endogenous internal reference genes in human muscle we used actin and GAPDH. The expression of both COX6AH and COX7AH mRNA increased significantly at 3 months of age ~9 fold and ~35 fold respectively, whereas the COX6AL and COX7AL isoforms did not show dramatic alteration by age.
Page 12 of 32
13
Taken together these data illustrate that in infants COX6AL/COX6AH and COX7AL/COX7AH transcripts are co-expressed in skeletal muscle and around 3 months of age the gene expression of the muscle type isoforms (COX6AH and COX7AH) raises suddenly, and these
ip t
isoforms become the major isoforms in later age.
us
RIRCD in the early symptomatic phase and after recovery
cr
3.5. Expression pattern of COX6A and COX7A isoforms in skeletal muscle of patients with
To investigate the role of the isoform shift in RIRCD we studied how the transcription and
an
translation of these specific subunits changes after clinical recovery. We collected biopsies of RIRCD patients during the symptomatic phase (1 month) and after clinical recovery (> 2
M
years). Gene expression in muscle of an RIRCD patient (P11 in Horvath et al., 2009) during the symptomatic phase at the age of 1 month revealed significantly increased levels of
d
isoforms of both subunits, suggesting compensatory up-regulation. This change was very
te
pronounced with the contractile type of COX6AH (~23 fold) and even more significant with
Ac ce p
COX7AH (~90 fold) (Fig. 4A, 4B). After recovery however, the expression level of these transcripts returned to nearly normal level compared to age matched control muscle (Figure 4A, 4B). The gene expression profile of other mitochondrial encoded transcripts such as MTRNR1 (12S rRNA) or MT-ND1 did not increase and the level of MT-CO2 was significantly down-regulated in early RIRCD muscle explaining the observed lack of COX2 protein in RIRCD muscle tissue (Tritschler et al., 1991)(Fig. 4C). Tissue sections obtained from muscle biopsies of a patient with RIRCD were also examined by immune-labelling with isoform specific antibodies. Immunofluorescence analysis showed that during the early phase, none of the liver or the heart type isoforms of COX6A and COX7A were expressed normally in RIRCD skeletal muscle (Fig. 4D). Interestingly, only
Page 13 of 32
14
COX7AH was detectable which demonstrated a rather unusual appearance when compared to age matched control skeletal muscle sections. The expression of this isoform was decreased, but it showed distinct aggregates in some of the skeletal muscle fibres while no
ip t
abnormalities were observed with COX4 antibody (Fig. 4F). Double staining with myosin light chain antibody revealed that these clusters were mainly present in the oxidative type I
cr
fibres (Fig. 4G). Immuno-labelling of the follow-up muscle biopsy after recovery showed
us
normalised expression and localisation of these nuclear encoded COX subunits when compared to control muscle sections (Fig. 4E). Based on the data collected from muscle
an
biopsies of RIRCD patients we can conclude that the presented shift in isoforms through development cannot explain the reversibility of the disorder. Most likely the absence of the
M
mtDNA encoded COX subunits is caused by abnormal translation of mtDNA-encoded proteins triggered by the mt-tRNAGlu mutation. The reduced mitochondrial encoded COX
d
subunits further compromise COX assembly, which results in secondary impairment of the
te
nuclear encoded COX6A and COX7A subunits.
Ac ce p
3.6. What is the effect of blocking mitochondrial translation in RIRCD? In order to investigate whether the low steady–state level of mt-tRNAGlu in RIRCD happens during processing/maturation of the mt-tRNA or during protein synthesis, we blocked mitochondrial translation with TAP in myoblasts of an RIRCD patient (P14 in Horvath et al., 2009) and in a control cell line. We hypothesized that if the steady state level of mt-tRNAGlu remains low after blocking mitochondrial translation, it would indicate that the instability happens during processing/maturation of the mt-tRNA, however if the originally low tRNAGlu steady state returns to normal or increases, this would indicate that the mt-tRNA instability occurs during mitochondrial protein synthesis. Interestingly, blocking mitochondrial
Page 14 of 32
15
translation resulted in an increase of the tRNA steady state (Fig. 5), suggesting that the mutated mt-tRNA may cause a higher level of failed translation events, stimulating its decay, which may affect the function and assembly of mitoribosomes. The elevated level of mt-
ip t
tRNATrp, but no change in the expression of the cytoplasmic tRNALys further suggested a
cr
more general suppression of mitochondrial protein synthesis in RIRCD in vitro (Fig. 5).
us
4. Discussion
The main finding of the present study is that we define the specific time frame when the
an
COX6A and COX7A isoform switch occurs in mice and most importantly in human skeletal muscle during postnatal development. Previous studies in mice demonstrated that the liver
M
type isoforms are predominant in fetal tissues (Bonne et al., 1993), but these are gradually replaced by the heart type of isoforms in both mouse cardiac and skeletal muscle (Kim et al.,
d
1995). Here we show that the COX isoform switch in human muscle takes place by the age
te
of 3 months, as illustrated by the predominant expression of muscle or heart isoforms.
Ac ce p
It has been hypothesized previously that the developmental isoform switch of COX6A and COX7A isoforms may be the cause of spontaneous clinical recovery in RIRCD (Tritschler et al., 1991; Taanman et al., 1993). Before the discovery of the genetic cause of RIRCD, a hypothesis by Taanman (1993) has been put forward to explain how the reversibility may occur in RIRCD, suggesting a defect in the transcriptional regulatory mechanisms in the liver type subunits. However, based on our data the gene expression of both liver and heart type isoforms has dramatically increased during the symptomatic phase in patient’s muscle, possibly due to compensatory mechanisms, trying to correct reduced protein expression. Although we show a developmental switch in the studied nuclear COX isoforms around 3 months of age, just shortly prior to the spontaneous recovery in RIRCD (~6 months), our
Page 15 of 32
16
data suggest that it is most likely not causative for the improvement of clinical symptoms in RIRCD. Sequence analysis of the human COX6A and COX7A isoforms revealed 66.3% and 59.7% homology respectively (Ma et al., 2014, Peng and Xu, 2011a, 2011b, Suppl. Fig. 2.)
ip t
Discrepancies between the liver and muscle type of isoforms therefore could possibly indicate functional differences which might be important in mitochondrial diseases. The
cr
m.14674T>C/G mutation most likely impairs mitochondrial translation, as reflected by the
35
us
RRF/COX-negative fibres and the multiple RC defect in skeletal muscle. Previously however S-methionine pulse labelling in 2 patient cell lines carrying the m.14674T>C mutation
an
demonstrated normal or even slightly increased mitochondrial protein synthesis (Horvath et al., 2009). In parallel, immunoblotting showed normal or even higher mitochondrial protein
M
levels suggesting an in vitro compensatory mechanism. Northern blots for the expression of the mitochondrial encoded protein-coding genes also detected normal levels suggesting
d
that mitochondrial transcription remains intact (Horvath et al., 2009). Based on our data we
te
suggest that normal mitochondrial translation in RIRCD cells is kept by a significant effort
Ac ce p
through utilisation of cellular compensatory mechanisms. If an additional factor negatively alters mitochondrial translation, cellular compensation may not cope with the increased need which results in a severely impaired or stalled mitochondrial translation, probably leading to a chain of other events resulting in severe mitochondrial dysfunction. The increase in steady state levels of the mt-tRNAGlu after thiamphenicol treatment indicates that processing and maturation of the mt-tRNAGlu is stable and not altered by the m.14674T>C/G mutation. Interestingly, the role of COX6AH as an isoform required for the overall stability of the COX holoenzyme after assembly has also been recognised in the past and could further explain the predominant COX defect in patients with RIRCD (Fornuskova et al., 2010).
Page 16 of 32
17
Additional factors such as environmental contributions, epigenetic factors and the presence of other genetic polymorphisms may ultimately modify the nature of the disease phenotype and also have to be considered to explain reversibility in these conditions (Boczonadi et al.,
ip t
2013). Recently our group suggested that altered thio-modification of the mt-tRNAGlu due to low infantile cysteine concentration could be a potential explanation for the timed
cr
disease manifestation. Cysteine is an essential amino acid in infants, because the
us
cystathionine gamma-lyase (cystathionase) enzyme (rate-limiting enzyme for the synthesis of L-cysteine from L-methionine) appears to be low after birth and regulated at the post-
an
transcriptional level during development in the first months of life (Levonen et al., 2000). Supplementation of L-cysteine, that is essential for normal thio-modification of mt-tRNAGlu,
M
mt-tRNALys and mt-tRNAGln by TRMU, significantly improved most RC complex activities in TRMU and RIRCD patient cell lines. These data suggest that low dietary L-cysteine levels in
d
infants could act as a second hit on mitochondrial translation, explaining clinical
te
manifestations of RIRCD and TRMU deficiency, and an increase in L-cysteine due to
Ac ce p
cystathionase activation through development would contribute to the recovery in these patients (Boczonadi et al., 2013).
Conclusion
Here we show experimental evidence for an isoform switch of COX6A and COX7A in skeletal muscle, which happens around 3 month of age. However, in contrast to previously suggested hypotheses, we could not confirm a causative link between the isoform switch of COX subunits and the clinical recovery in RIRCD our data may have implication for severe infantile presentations of other mitochondrial myopathies.
Page 17 of 32
18
Acknowledgments The work is supported by the Medical Research Council (UK) (G1000848), the European Research Council (309548) and the Mitochondrial European Educational Training (MEET),
ip t
ITN MARIE CURIE PEOPLE, (317433). We are grateful to the Medical Research Council (MRC)
us
supporting this project and for providing primary human cells.
cr
Centre for Neuromuscular Diseases Biobank Newcastle and for the EuroBioBank for
M
The authors declare no conflict of interests
an
Conflict of interest statement
References
d
1. Boczonadi V, Smith PM, Pyle A, Gomez-Duran A, Schara U, Tulinius M, Chinnery PF,
te
Horvath R. Altered 2-thiouridylation impairs mitochondrial translation in reversible
Ac ce p
infantile respiratory chain deficiency. Hum Mol Genet 2013;22(22):4602-15.
2. Bonne G, Seibel P, Possekel S, Marsac C, Kadenbach B. Expression of human cytochrome c oxidase subunits during fetal development. Eur J Biochem 1993;217(3):1099-107.
3. Chrzanowska-Lightowlers ZM, Preiss T, Lightowlers RN. Inhibition of mitochondrial protein synthesis promotes increased stability of nuclear-encoded respiratory gene transcripts. J Biol Chem 1994;269(44):27322-8.
Page 18 of 32
19
4. Fornuskova D, Stiburek L, Wenchich L, Vinsova K, Hansikova H, Zeman J. Novel insights into the assembly and function of human nuclear-encoded cytochrome c oxidase subunits 4, 5a, 6a, 7a and 7b. Biochem J 2010; 428(3):363-74
ip t
5. Greaves LC, Reeve AK, Taylor RW, Turnbull DM. Mitochondrial DNA and disease. J
cr
Pathol 2012;226(2):274-86.
us
6. Grossman LI, Lomax MI. Nuclear genes for cytochrome c oxidase. Biochim Biophys Acta 1997;1352(2):174-92.
an
7. Horvath R, Kemp JP, Tuppen HA, Hudson G, Oldfors A, Marie SK, Moslemi AR, Servidei S, Holme E, Shanske S, et al. Molecular basis of infantile reversible
M
cytochrome c oxidase deficiency myopathy. Brain 2009;132,3165-3174.
d
8. Huang X. and Miller W. A Time-Efficient, Linear-Space Local Similarity Algorithm.
te
Advances in Applied Mathematics, vol. 12 (1991), pp. 337-357.
Ac ce p
9. Hüttemann M, Kadenbach B, Grossman LI. Mammalian subunit IV isoforms of cytochrome c oxidase. Gene 2001;267(1):111-23.
10. Hüttemann M, Schmidt TR, Grossman LI. A third isoform of cytochrome c oxidase subunit VIII is present in mammals. Gene 2003;312:95-102.
11. Kadenbach B, Hüttemann M, Arnold S, Lee I, Bender E. Mitochondrial energy metabolism is regulated via nuclear-coded subunits of cytochrome c oxidase. Free Radic Biol Med 2000;29(3-4):211-21.
Page 19 of 32
20
12. Kim K, Lecordier A, Bowman LH. Both nuclear and mitochondrial cytochrome c oxidase mRNA levels increase dramatically during mouse postnatal development. Biochem J 1995;306(Pt 2):353-8.
ip t
13. Källberg M, Wang H, Wang S, Peng J, Wang Z, Lu H, Xu J. Template-based protein
cr
structure modelling using the RaptorX web server. Nat Protoc. 2012;7(8):1511-22.
us
14. Levonen AL, Lapatto R, Saksela M, Raivio KO. Human cystathionine gamma-lyase: developmental and in vitro expression of two isoforms. Biochem J 2000;347(Pt
an
1):291-5.
15. Lochmüller H, Johns T, Shoubridge EA. Expression of the E6 and E7 genes of human
M
papillomavirus (HPV16) extends the life span of human myoblasts. Exp Cell Res
d
1999;248(1):186-93.
te
16. Ma J, Wang S, Zhao F, and Xu J. Protein threading using context-specific alignment potential. Bioinformatics (Proceedings of ISMB) 2013;29(13):i257-i265.
Ac ce p
17. Mimaki M, Hatakeyama H, Komaki H, Yokoyama M, Arai H, Kirino Y, Suzuki T, Nishino I, Nonaka I, Goto Y. Reversible infantile respiratory chain deficiency: a clinical and molecular study. Ann Neurol 2010;68(6):845-54.
18. Peng J. and Xu J. A multiple-template approach to protein threading. Proteins, 2011;79(6):1930-9.
19. Peng J. and Xu J. RaptorX: exploiting structure information for protein alignment by statistical inference. Proteins, 2011;79(S10):161-71.
Page 20 of 32
21
20. Rizzuto R, Nakase H, Darras B, Francke U, Fabrizi GM, Mengel T, Walsh F, Kadenbach B, DiMauro S, Schon EA. A gene specifying subunit VIII of human cytochrome c oxidase is localized to chromosome 11 and is expressed in both muscle and non-
ip t
muscle tissues. J Biol Chem 1989;25;264(18):10595-600.
cr
21. Shoubridge EA. Cytochrome c oxidase deficiencies. Am J Med Genet. 2001
us
Spring;106(1):46-52. Review.
22. Taanman JW, Hall RE, Tang C, Marusich MF, Kennaway NG, Capaldi RA. Tissue
an
distribution of cytochrome c oxidase isoforms in mammals. Characterization with monoclonal and polyclonal antibodies. Biochim Biophys Acta 1993;1225:95-100.
M
23. Tritschler HJ, Bonilla E, Lombes A, Andreetta F, Servidei S, Schneyder B, Miranda AF,
d
Schon EA, Kadenbach B, DiMauro S. Differential diagnosis of fatal and benign
te
cytochrome c oxidase-deficient myopathies of infancy: an immunohistochemical
Ac ce p
approach. Neurology 1991;41(2 ( Pt 1)):300-5. 24. Uusimaa J, Jungbluth H, Fratter C, Crisponi G, Feng L, Zeviani M, Hughes I, Treacy EP, Birks J, Brown GK, et al. Reversible infantile respiratory chain deficiency is a unique, genetically heterogeneous mitochondrial disease. J Med Genet 2011;48:660-668.
Page 21 of 32
22
Figure legends Figure 1. COX6A and COX7A subunit isoforms colocalise with MTCO1 in mouse skeletal muscle. Double immune-labelling of muscle sections. Merged images are shown in the last
ip t
column including nuclear staining (DAPI blue channel). (A). COX6AL (green channel)
cr
colocalisation with MTCO1 (red channel). (B). COX6AH (green channel) colocalisation with MTCO1 (red channel). (C). COX7AL (green channel) colocalisation with MTCO1 (red channel).
us
(D). COX7AH (green channel) colocalisation with MTCO1 (red channel). Scale bar represents:
an
20 micron.
M
Figure 2. Expression of COX6A and COX7A isoforms during mouse development. (A). Immunofluorescence micrographs of mouse skeletal muscle sections stained with
d
COX6AL, COX6AH, COX7AL or COX7AH (green channel) with DAPI (blue channel).
te
Developmental series are shown from the top to bottom (1 day, 8 days, 2 weeks, 4 weeks and adult). Scale bar represents 20 micron. (B). Western blot analysis of COX6A and COX7A
Ac ce p
subunit isoforms through development. Porin was used as a loading control.
Figure 3. Gene expression of COX6A and COX7A subunit isoforms through development in mice and human. Analysis of gene expression was performed by quantitative RT-PCR at different ages. The graphs represent the average of triplicate wells in three independent experiments with standard deviation shown as error bars. (A) Expression pattern of the liver and contractile type of COX6A in mouse skeletal muscle. (B) Gene expression of the liver and muscle type of COX7A in mouse skeletal muscle. (C-D) Expression profiles of COX6A and COX7A subunit isoforms in murine liver and (E-F) heart. (G-H) Biochemical measurement of
Page 22 of 32
23
citrate sythase and complex IV activity in the mouse muscle, heart and liver at different ages. Measurements were carried out as triplicates (n=3). (I-J) Relative quantification of the COX6A and COX7A isoforms in human skeletal muscle at different ages. Mouse samples
ip t
were normalised to actin and vimentin; human samples were normalised to actin and GAPDH. Data are presented as mean ± standard deviation. A p value of <0.05 was
us
cr
considered significant.
Figure 4. Expression pattern of COX6A and COX7A isoforms in skeletal muscle of patients
an
with RIRCD in the early symptomatic phase and after recovery. (A,B) Graphs are representing COX6A and COX7A isoform gene expression in control human samples at
M
different ages (3 months, 1 year 6 months, adult) and in RIRCD muscle before (1 month) and after recovery (8 years 9 months) (marked by a red line). (C) Gene expression of MT-CO2,
d
MTRNR1 (12S rRNA) and MT-ND1 in control and RIRCD muscle. The graphs represent the
te
average of triplicate wells in three independent experiments with standard deviation shown
Ac ce p
as error bars. (D) Immunofluorescence labelling of 1 month old human control muscle section with antibodies against COX6A and COX7A subunit isoforms (first column). 1 month old RIRCD muscle sections stained for COX6A and COX7A subunit isoforms (second column). (E) Adult human control muscle section (first column) and RIRCD muscle sections (second column) after recovery, labelled with COX6A and COX7A subunit isoforms. (F) 1 month old control and 1 month old RIRCD muscle sections labelled with COX7AH (green channel) and COX4 (red channel). White arrow heads indicate fibres with COX7AH clusters. (G) Control and RIRCD muscle sections at 1 month of age were co-labelled with COX7AH (black and white) and Myosin light chain Type I. (green channel). Scale bar represents 20 micron.
Page 23 of 32
24
Figure 5. Gene expression analysis of mt-tRNAGlu, mt-tRNATrp and cyt-tRNALys in control and patient myoblasts after blocking mitochondrial translational with Thiamphenicol
ip t
Analysis of relative gene expression was performed by quantitative RT-PCR. The graphs
cr
represent the average of triplicate wells in three independent experiments with standard deviation shown as error bars. CNTR: control immortalised myoblasts, P.: patient
us
immortalised myoblasts.
Ac ce p
te
d
M
an
Table. Laboratory investigations and muscle biopsy findings of the patient with RIRCD.
Page 24 of 32
25
Table. Laboratory investigations and muscle biopsy findings of the patient with RIRCD. Laboratory results
Muscle histology RRF/COX- Lipid Glycogen Other +
-
abnormal mt
COX ↓↓↓
8y 9m: +
-
-
mild myopathy
norm
CK
norm
te
d
M
an
us
normal
CI ↓↓↓
ip t
1m: +++
cr
Lactate
Ac ce p
Patient
Biochemical activity of the respiratory chain enzymes
Page 25 of 32
cr
i
Graphical Abstract (for review)
COX6AL
COX6AH
COX6AH
M an
NEW BORN
us
COX6AL
COX7AH
ce pt
COX7AL
ed
SKELETAL MUSCLE DEVELOPMENT
ADULT
COX7AL
COX7AH
Ac
~ 3 months
L-isoform ubiquitously expressed (embryonic isoform)
H-isoform expressed in mature contractile tissues (muscle/heart isoform)
Page 26 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 1
Page 27 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 2
Page 28 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 3
Page 29 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 4
Page 30 of 32
Ac ce p
te
d
M
an
us
cr
ip t
Figure 5
Page 31 of 32
Table
Table. Laboratory investigations and muscle biopsy findings of the patient with RIRCD. Biochemical activity of the respiratory chain enzymes
Lactate CK normal
1m: +++
+
-
COX ↓↓↓
CI ↓↓↓
8y 9m: +
-
-
norm
norm
ce pt
ed
M
an
us
cr
abnormal mt mild myopathy
ip t
Muscle histology RRF/COX- Lipid Glycogen Other
Ac
Patient
Laboratory results
Page 32 of 32