Abstract
Rearrangements of mitochondrial DNA (mtDNA) are a well-recognized cause of human disease; deletions are more frequent, but duplications are more readily transmitted to offspring. In theory, partial duplications of mtDNA can be resolved to partially deleted and wild-type (WT) molecules, via homologous recombination. Therefore, the yeast CCE1 gene, encoding a Holliday junction resolvase, was introduced into cells carrying partially duplicated or partially triplicated mtDNA. Some cell lines carrying the CCE1 gene had substantial amounts of WT mtDNA suggesting that the enzyme can mediate intramolecular recombination in human mitochondria. However, high levels of expression of CCE1 frequently led to mtDNA loss, and so it is necessary to strictly regulate the expression of CCE1 in human cells to ensure the selection and maintenance of WT mtDNA.
INTRODUCTION
Recombination is the basis of sex, yet DNA, that is asexually transmitted may also undergo frequent recombination. Such is the case for mitochondrial DNA (mtDNA) of plants and fungi ( 1 , 2 ). DNA recombination may serve as a mechanism of replication in its own right ( 2 , 3 ), or to process the products of rolling circle replication ( 4 ). DNA recombination is also widely used in DNA repair ( 5 ) and can rescue stalled replication forks ( 6 ). Vertebrate mitochondria appear to replicate their DNA predominantly via theta modes in liver and placenta ( 7 , 8 ). In contrast, human heart mtDNA contains abundant Holliday (four-way) junctions ( 9 ); while such species represent recombination intermediates, it is not yet clear what role they play in mtDNA replication. Generally DNA recombination in mammalian mitochondria appears to be a rare event; nevertheless mtDNA mutants have provided evidence of both intramolecular ( 10 , 11 ) and intermolecular recombination ( 12 , 13 ).
Many mutations in mtDNA take the form of large-scale rearrangements, either partial deletions ( 14 ) or tandem duplications ( 15 ). Duplications can co-exist with deletions ( 16 , 17 ), and intramolecular homologous recombination (HR) of partially duplicated (PD) mtDNA can in theory give rise to a wild-type (WT) monomer and a partially deleted mtDNA molecule ( 17 , 18 ). Most patients with rearranged mtDNA have partial deletions, with few if any accompanying PD molecules ( 19 , 20 ); very rarely, PD mtDNA is observed without partially deleted mtDNA ( 18 ).
A key component of the DNA recombination machinery in yeast mitochondria is the Holliday junction resolvase (HJR) or cruciform cutting endonuclease, Cce1p ( 21 ). The enzyme works in concert with Mhr1p to maintain mtDNA, as the loss of CCE1 and MHR1 genes leads to frequent and rapid mtDNA depletion, whereas loss of either CCE1 or MHR1 alone has little effect on mtDNA stability ( 22 ). While the stability of WT (ρ + ) mtDNA is largely unaffected in Δcce1 yeast strains, the absence of the CCE1 gene has a pronounced effect on the transmission of hypersuppressive ρ − mtDNA, due to the complex unresolved recombination intermediates that form in such cells ( 1 ). While Cce1p is abundant in the mitochondria of budding yeast, HJR activity has not been detected in mammalian mitochondria, and there is no known homologue of CCE1 in the human genome. Nevertheless, some mammalian mitochondrial extracts can mediate HR ( 23 ). This activity must be insufficient or ineffective in cells with tandemly repeated mtDNA, such as the 6.3 kb partial duplication described previously ( 10 ). Therefore, we sought to boost the HR capacity of human osteosarcoma cells carrying rearranged mtDNA, via heterologous expression of CCE1 . This procedure led to the selection of lower order rearrangements and WT mtDNA, suggesting that the yeast enzyme is capable of mediating HR in human mitochondria. Although this approach was conceived as a potential therapeutic strategy for pathological partial duplications, the situation is complicated by the fact that sustained high levels of expression of Cce1p precipitated mtDNA depletion; and most pathological duplications will yield partially deleted mtDNAs that are replication competent, in addition to WT mtDNA.
RESULTS AND DISCUSSION
Targeting and import of yeast Cce1p to human mitochondria
Before introducing CCE1 into cells carrying rearranged mtDNA (as outlined in Fig. 1 A), we wished to confirm that the gene could be expressed and the protein product targeted to mitochondria of human cells and imported. Therefore, a haemagglutinin (HA) tagged version of the CCE1 gene was transfected into human 143B osteosarcoma cells. Fluorescin-labelled anti-HA antibody revealed a pattern of staining coincident with the mitochondrial specific dye, mitoTracker (Invitrogen), indicating exclusive mitochondrial targeting of the heterologously expressed protein (Fig. 1 B). In a parallel set of experiments, in vitro translated Cce1p was incubated with rat liver mitochondria to determine if the protein could be imported into mitochondria. Two forms were tested, one gene corresponded to the native yeast protein, which includes a short mitochondrial targeting signal (MTS) of 15 residues; the second carried, in addition, the MTS of human ATP synthase F 1 β L (F 1 β L -CCE1). Incubation with functional mitochondria produced a mobility shift in both proteins and the shorter species’ were protease protected, consistent with them being imported into mammalian mitochondria (Fig. 1 C).
Cloning and expression of the yeast resolvase Cce1p in human cells, and import into isolated rat liver mitochondria. ( A ) The CCE1 gene was cloned into vector pEF (Invitrogen) at the Nco I and Xba I restriction sites, with or without a HA tag. Empty vector or pEF-CCE1(-HA) was transfected into human 143B osteosarcoma and the levels of rearranged mtDNA monitored over time (see Figs. 2–4 ), The expression of HA tagged Cce1p was determined by immunoblotting with anti-HA antibody (Fig. 4 ). Initial confirmation that HA tagged Cce1p is targeted to human mitochondria was provided by immunocytochemistry of control 143B cells transfected with pEF-CCE1-HA and labelled with anti-HA antibody (green) and the mitochondrial specfic dye, MitoTracker red ( B ). ( C)35 S-methionine labelled translation products of native Cce1p or a version spliced to the leader sequence of the F 1 -β subunit of ATP synthase (F 1 βL-CCE1) yielded a protein of reduced molecular mass after sequential incubations with isolated rat liver mitochondria and protease, suggesting that Cce1p is imported into mammalian mitochondria without the requirement for a human MTS.
Changes in mtDNA effected by the introduction of the CCE1 gene into human osteosarcoma cells carrying rearranged mtDNA. Eco RI digested mtDNA of osteosarcoma cybrids carrying PD mtDNA, 143B.PD ( A and B ) or PT mtDNA, 143B.PT ( C and D ) was hybridized to radiolabelled probe spanning nt 16343–151 of human mtDNA ( 37 ). Each lane represents DNA derived from a distinct clonal cell line, approximately 1 month after transfection with empty vector, pEF (A and C) or vector containing the yeast CCE1 gene, pEF.CCE1 (B and D). Black asterisks denote samples with little or no mtDNA, grey asterisks denote samples with reduced mtDNA compared with controls. Nuclear (n) DNA was readily detectable in all the 143B.PT clonal cell lines irrespective of the level of mtDNA ( E and F ). Of the cell lines with little or no mtDNA only clones 8 and 9 (D) retained the CCE1 gene based on PCR analysis (not shown).
The introduction of CCE1 is associated with the appearance of WT mtDNA in cells hitherto stable for PD mtDNA
The results of the targeting and import experiments (Fig. 1 ) suggested that there would be no barrier to delivering yeast CCE1p to human mitochondria. Therefore, osteosarcoma cell lines carrying PD mtDNAs ∼23 kb in size, instead of the usual 16.6 kb WT mtDNA ( 10 ), were transfected with plasmid pEF (Invitrogen) containing or lacking the yeast CCE1 gene. Seventeen of 18 clones with pEF alone had little or no WT mtDNA and there was no reduction in mtDNA copy number; one clone had 49% PD mtDNA and just over half the usual mtDNA copy number of the other clones (Fig. 2 A). In the case of sister clones transfected with pEF.CCE1, five clones had a small amount of WT mtDNA (1.5–27.9%) and a further two clones had a substantial amount of WT mtDNA (73 and 87%), unlike any cybrid without the CCE1 gene. This suggested that yeast CCE1 could favour the selection of WT mtDNA over PD mtDNA, however, given the appearance of WT mtDNA in one control cell line it was not certain that selection was due to CCE1 ’s resolvase activity.
Introduction of CCE1 into cells with PT mtDNA frequently precipitates mtDNA loss
As a further test of the idea that CCE1 might mediate HR in human cells, a cell line with partially triplicated (PT) mtDNA and no detectable WT mtDNA was transfected with pEF or pEF.CCE1 and the mtDNA of the clones analysed 5–6 weeks later. All 18 clones carrying empty vector retained PT mtDNA and there was no detectable WT mtDNA, nor any hint of mtDNA depletion in these clones. In contrast, 10 of 18 PT clones transfected with pEF.CCE1 had no detectable mtDNA (mutant or WT) and one other clone had 23% of the control level of mtDNA, the remaining seven clones were indistinguishable from the parental cell line or the clones carrying empty vector (Fig. 2 D). Although complete loss of mtDNA was strictly associated with the introduction of vector containing the CCE1 gene, there was no detectable CCE1 DNA in seven of nine clones that lost mtDNA, based on PCR analysis (data not shown). All the cell lines remained resistant to G418, and so the CCE1 gene was presumably incorporated into nuclear DNA after transfection, and later lost due to negative selection, yet the mtDNA depletion precipitated by CCE1 must have been complete, or irreversible, by the time the CCE1 gene was lost.
The level of Cce1p expression determines whether lower order rearrangements are selected or mtDNA is lost
In theory, cybrids created with 143B osteosarcoma cells are immortal, nevertheless they can deteriorate with time. For instance, the yield of colonies following transfection of cybrids appears to correlate inversely with the length of time the cybrid cells have been maintained in culture (unpublished observations). Therefore, a ‘younger’ version of the same 143B.PT cybrid was transfected with pEF.CCE1, with and without an HA tag. On this occasion, only one of 24 clones had no detectable mtDNA at first screening and five clones had some WT mtDNA, unlike any of the 12 clones carrying empty vector (Fig. 3 ).
Introduction of CCE1 into human cells with PT mtDNA leads to the appearance of PD and WT mtDNA in some clones. 143B.PT cells were transfected with empty vector, pEF, or CCE1 (pEF.CCE1) with or without a HA tag. DNA was harvested from clonal cell lines 36–45 days (t 1 ) after selection in G418, digested with Eco RI, separated by agarose gel electrophoresis and hybridized to the same probe as Fig. 2 , A – D . The boxed clones carrying pEF.CCE1-HA were re-examined at later time points (see Fig. 4 ).
Twelve clones carried pEF.CCE1-HA, and so protein expression was followed in parallel with mtDNA structure, for a period of 3–8 months (Fig. 4 ). Persistent high expression of Cce1p was associated with mtDNA depletion in the case of clone 7, corroborating the earlier conclusion that Cce1p can cause mtDNA loss. A gradual increase in expression of Cce1p was also associated with mtDNA depletion, in clone 3 (Fig. 4 A and B). Cell clone 4 initially had a substantial amount of WT mtDNA and a high level of Cce1p expression (t 2 in Fig. 4 A), but there was a dramatic return to a high proportion of PT mtDNA four weeks later and a concomitant reduction in Cce1p (t 3 , Fig. 4 A), still later a slightly higher level of expression correlated with the appearance of PD mtDNA (t 4 in Fig. 4 A). A similar level of Cce1p-HA expression was also associated with the appearance of PD mtDNA in clone 2 (t 4 in Fig. 4 A). The only clone to convert to WT mtDNA completely, clone 5, had a low level of Cce1p at t 2 and this fell below the limits of detection at later time points (Fig. 4 A).
Cce1p expression and changes in mtDNA structure and amount in 143B.PT cells. pEF.CCE1-HA clones 2–8 (see Fig. 3 ) were maintained for several months; at intervals (t 2 –t 5 ) DNA and protein were extracted from cells. mtDNA was analysed by Southern blotting (upper gel image of each pair), and Cce1 protein was revealed by western blotting using an anti-HA antibody ( A ). To aid interpretation a box encloses all the clone 4 samples with arrows indicating the temporal order of events. mtDNA was quantified additionally by real-time PCR, using primer pairs detecting COXII of mtDNA (located in the non-duplicated region), and the nuclear APP gene ( B ). The level of mtDNA was arbitrarily set at 1 for 143B osteosarcoma cells.
Immunocytochemistry of mtDNA and HA tagged Cce1p revealed heterogeneity within clonal cell lines. Clones 2 and 4 both comprised mixtures of cells that did and did not express Cce1p, at t 4 (Fig. 5 A and B). These cells were also heterogeneous in terms of PD and PT mtDNA (clone 2: 25% PD, 74% PT mtDNA; clone 4: 47% PD, 51% PT mtDNA, Fig. 4 C), implying that Cce1p expression correlated directly with the appearance of PD mtDNA. In support of this view, the population of cells comprising clone 4 contained a higher proportion of Cce1p expressing cells than clone 2, based on visual inspection of large numbers of anti-HA labelled cells. Crucially, the proportion of cells with PD mtDNA increased further, such that at t 5 , when clones 2 and 4 had 86 and 93% PD mtDNA, respectively, the vast majority of cells were labelled with anti-HA antibody (Fig. 5 C and E). Thus, in the case of clones 2 and 4, expression of Cce1p produced an almost complete replacement of PT with PD mtDNA. In the case of clone 3, expression of Cce1p was associated with loss of mtDNA, the few cells retaining mtDNA did not express Cce1p (Fig. 5 D).
Distribution and expression of Cce1p in 143B.PT cells. 143B.PT cell lines (clones 2, 3 and 4) expressing Cce1.HAp were plated on coverslips, permeabilsed and incubated with primary and secondary antibodies to the HA tag (red) and DNA (green). Not all the clone 2 and clone 4 cells expressed Cce1.HAp at t 4 ( A and B ), and these cells were had substantial amounts of both PT and PD mtDNA (Fig. 4 A). By t 5 , < 6% of the mtDNA of clone 4 was in the form of PT mtDNA, and almost all the cells clearly expressed Cce1p ( C ). A higher level of Cce1 protein was associated with mtDNA loss in clone 3 cells (Fig. 4 A), only cells lacking Cce1p retained mtDNA ( D ). The cells retaining mtDNA and lacking Cce1p-HA are presumed to have excised both the CCE1 gene and the neo cassette, although contamination cannot be excluded absolutely.
The appearance of substantial amounts of WT mtDNA, in cells with PD or PT mtDNA carrying the CCE1 gene, suggests that the yeast mitochondrial HJR can mediate intramolecular recombination in human mitochondria. However, it is clear that high levels of expression of Cce1p precipitate catastrophic and irreversible mtDNA depletion in human cells carrying PT mtDNA. This was implied by the loss of mtDNA in 10 of 18 clones in the first experiment (Fig. 2 D) and supported by the consistent high level of expression Cce1p-HA in clone 7, which lacked mtDNA (Fig. 4 ). The most compelling case for a link between high expression of Cce1p and mtDNA depletion was provided by clone 3, where increasing expression of the protein accompanied mtDNA loss (Fig. 4 ), and immunocytochemistry revealed two populations of cells, the majority expressing the protein and lacking mtDNA, and a minor population with normal mtDNA copy number and no detectable Cce1p (Fig. 5 D). An excess of Cce1p must therefore compromise essential mitochondrial functions, not merely cause OXPHOS failure, if mtDNA loss is preferable to maintaining mtDNA in the presence of Cce1p. This would be the case if large amounts of Cce1p distort the structure of mitochondrial nucleoids in the inner mitochondrial membrane, thereby causing proton leak and depolarization. Clone 4 suggests that high levels of Cce1p are also deleterious to cells with WT mtDNA, as cells with PT mtDNA were resurgent at t 3 , despite the culture being dominated by cells with WT mtDNA at t 2 (Fig. 4 ). The best outcome was clone 5 where most of the rearranged mtDNA had been ablated by t 2 and expression of Cce1p was in decline, which led to a cell line with ∼100% WT mtDNA and no detectable Cce1p (t 4 in Fig. 4 ).
All the changes in rearranged mtDNA were seen with both tagged and untagged forms of Cce1p (Fig. 3 ), therefore none of the results is attributable to the HA tag. Cce1p has been shown in vitro to resolve four-way junctions without the need for additional proteins ( 24 , 25 ), and several non-mtDNA modifying proteins have been shown to function in mammalian mitochondria ( 26–28 ) so there is little doubt that the natural mitochondrial protein Cce1p will be capable of resolving four-way junctions in human mitochondria. All the steps of Holliday junction formation, nicking, strand exchange and branch migration can occur spontaneously, albeit at low efficiency without proteins analogous to RecA ( 29 ). DNA lzigase is required in vivo to repair four-way junctions processed by Cce1p, but this is not an obstacle as DNA ligases are essential for replicaiton, and DNA ligase III is known to be present in mammalian mitochondria ( 30 ).
While there are idiosyncrasies within the results, several conclusions can be drawn from the data. First, maintenance of mtDNA is unsustainable in human 143B cells expressing high levels of Cce1p. At t 2 (Fig. 4 A), three clonal cell lines (clones 3, 4 and 7) expressed a high level of Cce1p; one (clone 7) had already undergone mtDNA depletion, another was in the process of losing its mtDNA (clone 3) and the third (clone 4) did not maintain a high level of Cce1p expression (t 3 in Fig. 4 A). Although clone 3 had more Cce1p than clone 7, and was less severely depleted at t 2 (Fig. 4 A), this may not have been true of earlier time points; alternatively or additionally there may be considerable variation in the amount of Cce1p that can be tolerated by different clones. Second, low to moderate expression of Cce1p is (more) tolerable to cybrids and selects PD over PT mtDNA, as exemplified by clones 2 and 4, which were able to maintain their mtDNA in the presence of Cce1p, and indeed gained PD mtDNA at the expense of PT mtDNA, over a period of several months (Fig. 4 A). Whether this would have been sustainable indefinitely is a moot point, as clone 2 was showing evidence of mtDNA depletion at t 5 (Fig. 4 B), and expression of Cce1p in clone 4 had, by this time, reached a level that was incompatible with mtDNA maintenance in other clones (Fig. 4 A). Third, the level of expression required for selection of WT mtDNA overlaps that which causes mtDNA depletion and so only occurs as a chance event; i.e. for (WT) mtDNA to be maintained, sufficient mtDNA must be converted to WT before the deleterious effects of Cce1p become overwhelming, and this can only be achieved if it is timed to coincide with a decline in Cce1p expression. Hence, the competing forces of selection of (PD and WT) mtDNA and counter-selection of mtDNA (depletion), unleashed by Cce1p, serve to make expression of the protein unstable.
The mechanism by which yeast Cce1p causes mtDNA depletion is uncertain, although it is well recognized that over-expression of a protein often proves deleterious. A DNA binding protein could inhibit a variety of DNA transactions, including replication; as has been shown for the mitochondrial transcription factor and DNA packaging protein Tfam ( 31 ). If Holliday junction resolution by yeast Cce1p is inefficient in human mitochondria then an excess of Cce1p could cause mtDNA depletion by stabilizing four-way junctions, such structures would block the advance of a replisome and so impede mtDNA replication. Furthermore, fork reversal at replication pause sites results in the formation of four-way junctions and these might be stabilized or processed by Cce1p, which would at the very least delay the resumption of replication. These problems may well be more acute in cells with PT mtDNA, because of a greater propensity for PT mtDNA to form four-way junctions than PD or WT mtDNA.
It is not entirely clear why cells with PT mtDNA expressing intermediate levels of Cce1p should form PD mtDNA (Fig. 4 , clones 2 and 4, t 4 and t 5 ) in preference to WT mtDNA. In theory, intramolecular recombination can remove both tandem repeats in a single step; however, topological constraints imposed on mtDNA within mitochondrial nucleoids could hinder the process. It is possible to generate WT mtDNA from PD or PT mtDNA, as WT mtDNA was present in several PT cybrids after transformation with CCE1 (Figs 3 and 4 ). The preference for PD mtDNA in clones 2 and 4 (Fig. 4 ) could reflect counter-selection, as PD mtDNA has previously been shown to possess a replicative advantage over WT mtDNA in 143B cells ( 10 ). Although PT should exert the same replicative advantage over PD mtDNA, the OXPHOS phenotype associated with PT mtDNA is considerably more severe than PD mtDNA ( 10 ). Thus, an equilibrium may be reached where in the presence of a moderate amount of Cce1p, the beneficial effects of PD mtDNA outweigh the replicative advantage of PT mtDNA, whereas the smaller biochemical benefit of WT mtDNA is insufficient to counter the replicative advantage of PD over WT mtDNA. Increasing the amount of Cce1p could favour resolution of PD mtDNA, however, this advantage is likely to be outweighed by the demonstrable negative effects of excessive amounts of Cce1p.
Previous results implied that biased segregation of rearranged human mtDNA and mtDNA depletion are irreversible once set in train (e.g 10 , 32 ); for instance we were able to generate numerous clones with 0 and 100% PD mtDNA, but stable clones with intermediate levels of PD mtDNA were rare ( 10 ). Recent studies in plants indicate that a single-strand binding protein (OSB1) plays a critical role in preventing the amplification of ubiquitous, rearranged mtDNAs known as sublimons ( 33 ). Interruption of the OSB1 gene leads to an increase in sublimons, which is reversible; next a particular rearranged mtDNA is selected and comes to dominate the genotype (despite being deleterious); the second stage is irreversible ( 33 ). While there does not appear to be a direct homologue of OSB1 in animals, we nevertheless predict that a human analogue exists. A clearer understanding of the role of OSB1 in plant mtDNA replication could facilitate identification of its animal counterpart, thereby revealing a key component of the segregation apparatus. It is conceivable that the recently identified D-loop binding protein, ATAD3p, plays such a role. ATAD3A appears to be ubiquitously expressed in human cell lines, whereas ATAD3B is only expressed in a subset of cells ( 34 ) and unpublished data). Intriguingly, 143B cells that can sustain and select PD mtDNA express ATAD3B , whereas A549 cells that invariably gain WT mtDNA in place of PD mtDNA ( 10 ) have little or no ATAD3Bp ( 34 ).
Deletions versus duplications
Partially deleted mtDNA is a predicted product of HR on this or any other PD mtDNA ( 35 ). Mixtures of deleted and PD mtDNA have been seen in some cases of mitochondrial disease ( 17 ); yet they were never seen in patient-derived tissues or cell lines in the case of the particular rearrangement studied here ( 10 , 18 ). The absence of detectable deleted mtDNA in cells expressing Cce1p reported here strengthens the view that this particular partly deleted molecule is replication incompetent.
PD mtDNA is transmitted through the germ-line ( 15 , 18 ), whereas the vast majority of patients with partially deleted mtDNA are sporadic cases ( 19 , 20 ). As alluded to above, HJR activity or HR by other means will convert PD to partially deleted mtDNA (plus WT). Hence, the resolution of PD mtDNA could prevent transmission of rearranged mtDNAs and therefore disease. This may explain why partial deletions are the more frequent cause of disease, despite the fact that partial deletions of mtDNA are more deleterious than duplications, in terms of their effect on oxidative phosphorylation ( 10 , 36 ). Thus, the danger to the individual of resolving duplications to deletions may well be outweighed by the advantage to the species of preventing the transmission of rearranged mtDNA. If so, DNA recombination must be sufficiently active in human mitochondria to ensure that most individuals convert PD to partially deleted mtDNA, at least in some tissues or at particular stages of development. A further corollary is that patients with PD mtDNA are deficient in some part of the mitochondrial recombination apparatus. A fuller understanding of the process of DNA recombination in mammalian mitochondria may therefore provide a more practical solution to the problem of PD mtDNA than the introduction of a foreign gene such as CCE1 , that, unless tightly regulated, risks precipitating mtDNA loss.
METHODS
DNA modification, separation and hybridization
Restriction enzyme digestions were performed under conditions recommended by the manufacturer (New England Biolabs). DNA fragments were separated at 5 V/cm for 3 h on slab gels of between 0.55–1.0% (w/v) agarose, at room temperature. Southern blots were hybridized to specific regions of human mtDNA by overnight incubation at 65°C in 7% SDS, 0.25 m sodium phosphate (pH 7.2). Post-hybridization washes were 1 × SSC followed by 1 × SSC, 0.1% SDS, twice for 30 min at 65°C. Filters were exposed to X-ray film and developed after 0.5–10 days.
Radiolabelled probes were generated from human mtDNA fragments amplified from purified placental mtDNA using pairs of oligonucleotides (Sigma-Genosys). Briefly, 50 µCi of [α- 32 P] dCTP (3000 Ci/mmol, Amersham Biosciences) was incubated with 50 ng of heat denatured DNA and DNA labelling beads (GE Healthcare) for 30 min at 37°C. Oligonucleotide primers based on the revised Cambridge reference sequence (rCRS ( 37 ), were as follows: 5′-TTACAGTCAAATCCCTTCTCGTCC-3′, nt 16 341–16 364 and 5′-GGATGAGGCAGGAATCAAAGACAG-3′ nt 151–128. Oligonucleotide primers for an 18S rRNA nuclear DNA probe were 5′- CAGTTATGGTTCCTTTGGTCGCTCG-3′, nt 16 341–16 364 and 5′-TCCTTGGATGTGGTAGCCGTTTCTC-3′.
Confocal microscopy
Transfected cells were washed and stained with 250 n m of MitoTracker orange (Invitrogen) as described previously ( 38 ), then fixed and stained with anti-DNA antibody (PROGEN Biotechnik) and anti-HA antibody (3F10, Roche) as primary antibodies. A confocal microscopy system (Radiance 2000; BioRad Laboratories) was used for cell imaging, and images were edited using Photoshop Element (Adobe).
Real-time PCR estimation of mtDNA copy number in human osteosarcoma cells
Total DNA samples were prepared from cells lysed in 500 µl 75 m m NaCl, 50 m m EDTA, 0.2% SDS (pH 8.0). Cell lysates were incubated at 50°C for 2 h with 400 µg/ml proteinase K. DNA was precipitated by addition of an equal volume of isopropanol, pelleted by centrifugation at 8 500 gmax for 30 min, and dissolved in 100 µl of TE buffer (pH 8.0). mtDNA copy number was estimated by amplifying a portion of the cytochrome c oxidase ( COXII ) gene of mtDNA and comparing it to the amplification profile of a nuclear single copy gene, amyloid precursor protein or (APP). Primers for COXII were forward 5′-CGTCTGAACTATCCTGCCCG-3′, reverse 5′-TGGTAAGGGA GGGATCGTTG-3′ and probe 5′-CGCCCTCCCATCCCTACGCAT-3′ (FAM/TAMRA label). APP primers were forward 5′-TTTTTGTGTGCTCTCCCA GGTCT-3′, reverse 5′-TGGTCACTGGTTGGTTGGC-3′ and probe 5′-CCCTGAA CTGCAGATCACCAATGTGGTAG-3′ (FAM/TAMRA label). Other reagents, software and hardware were as described previously ( 38 ).
Western blotting
Transfected cells were lysed in 0.6% SDS in PBS and measured total protein as A 280. Cell lysates of 50 µg total protein in sample buffer (10% glycerol, 2% SDS, 0.1 M DTT, 0.0005% bromophenol blue and 0.1 m Tris–Cl, pH 6.8), were heated to 95°C for 5 min and separated by SDS-PAGE on a 12% gel at a constant 20 mA for 90 min. After separation, proteins were transferred to nitrocellulose membrane (Whatman) using a Trans-Blot SD blotter (BioRad) at 3 m A/cm 2 for 30 min. Membranes were blocked by incubation with 3% skimmed milk at room temperature for 1 h and incubated with 1:5000 diluted anti-HA-peroxidase high affinity antibody (3F10, Roche) at room temperature for 1 h. CCE1 with HA tag proteins were detected with ECL western blotting detection reagents (GE Healthcare) and medical X-ray film (FUJIFILM).
ACKNOWLEDGEMENTS
The study was funded by the UK Medical Research Council.
Conflict of Interest statement . None declared.
