Selective elimination of mutant mitochondrial genomes as therapeutic strategy for the treatment of NARP and MILS syndromes

Adenovirus construction

Our preliminary experiments confirmed the findings of Tanaka et al.¹⁴ that expression of the wild-type R.SmaI gene in human cells necessitates repeated transfections in order to achieve efficient elimination of mutant mtDNA from cybrid cells carrying a T8993G transversion. Presumably, this occurs because of low expression resulting from suboptimal codon usage (see Supplementary Figure S1). To circumvent this problem, a codon-optimized version of R.SmaI was generated and shown to be more efficient at the elimination of mutant mtDNA in transient transfection experiments (results not shown). However, some plasmids containing this synthetic gene proved to be toxic to Escherichia coli, resulting in failure to generate the desired constructs. Therefore, a codon-optimized version of the gene for R.XmaI, which is an isoschizomer of R.SmaI, was generated and used throughout this study.

Initial attempts to generate R.XmaI-encoding adenovirus have met with limited success, as infection of permissive HEK293 cells with viral particles appeared to be self-limiting and led to failure in the attempts to amplify virus. It appears plausible that productive infection with adenovirus, which results in the generation of up to 10 000 viral particles per cell,¹⁵ also results in massive overexpression of virus-encoded products, including RE. Such overexpression inevitably results in a failure of mitochondria to import an appreciable fraction of the RE produced, which may abort the infection through both destruction of viral genomes and nuclear DNA in infected cells. Therefore, the codon-optimized gene for M.SmaI, which cross-protects from R.XmaI digestion (see Supplementary Figure S2) was incorporated into the virus to protect against these effects.

The final bicistronic expression construct found in adenovirus Ad.1891 (Figure 1a) consists of four elements: (1) the cumate-inducible CMV promoter (CMV(CuO); MP Biomedicals, Solon, OH, USA); (2) codon-optimized M.SmaI fused to an SV40 nuclear localization sequence and a FLAG tag; (3) an internal ribosome entry site (IRES) of encephalomyocarditis virus¹⁶ and (4) the mitochondrial targeting sequence from human cytochrome C oxidase subunit VIII¹⁷ followed by a myc-tag and the codon-optimized R.XmaI gene.

Ad.1891 infection results in the selective elimination of mutant mtDNA

The infection of JCP239 cells (78% T8993G) with Ad1891 resulted in a dose-dependent expression of both M.SmaI and R.XmaI (Figure 1b). It is of note that while the expression of M.SmaI was detectable at the lowest multiplicity of infection (MOI) tested, the detection of R. XmaI expression required a twofold greater MOI. This is most likely a consequence of the generally lower expression attained by the second cistron in IRES-linked bicistronic constructs.¹⁸

The expressed R.XmaI was targeted to mitochondria as evidenced by immunocytochemistry with myc-tag specific monoclonal antibody (Figure 1c).

The N-terminal addition of the mitochondrial targeting sequence and myc-tag did not compromise the activity of R.XmaI. Indeed, mitochondrial lysates from cells infected with Ad.1891 specifically cleaved the PCR product containing the T8993G mutation in a dose-dependent manner with the formation of cleavage products indistinguishable in their migration pattern from those generated by a commercial preparation of R.SmaI. Importantly, these lysates failed to digest the wild-type (WT) PCR product (Figure 2a).

We next determined whether infection with Ad.1891 would affect the mutant mtDNA content in cells carrying the T8993G mutation. Two cell lines JCP213 (100% WT mtDNA) and JCP261 (100% T8993G mtDNA) were infected with Ad.1891, and Southern blotting was performed on total DNA isolated 48 h after infection. Mitochondrially targeted R.XmaI did not affect mtDNA content in JCP213, but completely destroyed T8993G-containing mtDNA in JCP261 cells (Figure 2b).

The elimination of the mutant DNA from JCP239 cells (78% T8993G) occurs in a dose-dependent manner, with the effect being noticeable at an MOI = 10, and with essentially complete elimination of mutant DNA at MOI ⩾50 (Figure 2c). Similarly, the effect of Ad.1891 infection is time dependent, and a decrease in mutant mtDNA content was detectable between 18 and 29 h after infection. This effect was essentially complete by 48 h after infection (Figure 2d).

Importantly, repeated application of the virus resulted in a progressive decrease in mutant mtDNA content in JCP239 cells. Thus, the mutant mtDNA content was reduced from approximately 60 to 36, 20 and 0.3% after the first round of infection, and to 20, 3.8 and 0% after the second round (MOI = 10, 20 and 50, respectively; Figures 2c and e).

Ad.1891 infection does not cause nuclear DNA damage

The intracellular expression of mitochondrially targeted RE is potentially associated with mistargeting due to either overexpression or an unfavorable combination of MTS and passenger polypeptide. We have recently shown that while the MTSs from mitochondrial superoxide dismutase and DNA polymerase-γ efficiently target their respective passenger polypeptides and the fluorescent protein DsRed-Express to mitochondria, they fail to do so for a closely related red fluorescent protein, mCherry.¹⁹ Therefore, the fusion of a protein to a mitochondrial targeting sequence cannot automatically be assumed to cause complete mitochondrial localization of this protein. A mistargeting of RE to the nucleus could cause severe side effects, including chromosomal aberrations and cell death. Therefore, we tested JCP239 cells for markers of double-stranded breaks in nuclear DNA upon infection with Ad.1891. While etoposide, an antitumor drug that blocks resealing of double-stranded breaks generated by DNA topoisomerase II in the nuclear DNA,²⁰ induced elevated levels of both phospho-P53 and γH2ax, infection with Ad.1891 failed to do so (Figure 3).

Ad.1891 infection improves the biochemical parameters compromised by T8993G mutation

In a medium containing galactose as the sole carbon source, mammalian cells exclusively rely on mitochondrial oxidative phosphorylation for ATP production. Therefore, growth in galactose-containing medium is a sensitive means for identifying the defects in oxidative phosphorylation.²¹ To test whether Ad.1891 infection produced an improvement in oxidative phosphorylation, the growth of the parental JCP239 cells in galactose-containing medium was compared to that of 2–50 cells (JCP239 cells, which underwent two rounds of Ad.1891 infection with MOI = 50). The growth rates of JCP239 and 2–50 cells were indistinguishable in glucose-containing medium (Figure 4a). In contrast, when forced to grow on galactose, JCP239 cells quickly lost their viability, while viability of 2–50 cells remained largely unaffected (Figure 4b).

The primary biochemical consequence of the T8993G mutation is a decrease in the rate of ATP synthesis.¹¹ Therefore, it was important to evaluate changes in both total cellular ATP content and in the rate of mitochondrial ATP synthesis following Ad.1891 infection. The improvement in the ability of JCP239 cells to utilize galactose as the sole carbon source upon Ad.1891 infection was accompanied by significant increases in total cellular ATP content and in the rate of ATP synthesis by mitochondria (Figures 5a and b). In accordance with the increased rate of mitochondrial ATP synthesis, the rate of oxygen consumption by 2–50 cells also was significantly higher than in JCP239 cells (Figure 5c).

Lactic acidosis is a common feature of mitochondrial diseases in general,²² and NAPR syndrome in particular.²³ Typically, the defect in oxidative phosphorylation leads to an inability of mitochondria to efficiently process pyruvate, which is produced during glycolysis. As a result, pyruvate is reduced to lactic acid, whose levels are consequently increased in blood. Our results indicate that lactic acid production by the 2–50 cells is diminished compared to the parental JCP239 cells (Figure 5d).

Cell lines and culture conditions

JCP213, JCP-239 and JCP261 are the derivatives of the human osteosarcoma cell line 143B TK-ρ°, which were generated by fusion of this line with platelets from patients with T8993G mutations.³¹ JCP213, JCP-239 and JCP261 contain 0, 78, and 100% of mtDNA with T8993G transversion, respectively, and were cultivated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 100 μg ml⁻¹ gentamycin, 1 mM sodium pyruvate and 2 mM glutamine at 37 °C in the atmosphere of 5% CO₂.

Gene synthesis

The synthetic, codon-optimized genes for the XmoI RE (R.XmaI) and SmaI methyltransferase (M.SmaI) were assembled as described previously.³² Briefly, amino acid sequences for the R.XmaI (GenBank AF051091) and M.SmaI (GenBank X16458) were back translated using the codons most frequently employed by human cells with the aid of the VectorNTI software package (Invitrogen, Carlsbad, CA, USA). Overlapping 60-mer oligonucleotides covering the whole coding sequence were synthesized by Integrated DNA Technologies (Ames, IA, USA) and used without further purification. The gene assembly was achieved by two sequential rounds of PCR. In the first round, a mixture of oligonucleotides (2 pmol each) was subjected to 25 rounds of PCR in 100 μl (95 °C, 1 min; 55 °C, 1 min; 72 °C, 1 min) using Taq polymerase (New England Biolabs, Ipswich, MA, USA). In the second round, 3 μl of the product of the first round were subjected to 30 rounds of PCR with Vent DNA polymerase in a final volume of 100 μl (New England Biolabs) using the same cycling parameters and a pair of primers flanking the coding sequence. The product of the second round was cloned into pBluescriptII SK+ vector and the fidelity of the assembly was verified by sequencing at the DNA Sequencing Facility at Iowa State University.

Adenovirus generation

This construct was generated with the help of standard molecular biology procedures³³ using the adenovirus transfer vector pAdenoVatorCMV(CuO) (MP Biomedicals) as a backbone. To generate viral particles, the 1891 construct was first recombined with the adenovirus genomic plasmid pAdenoVator ΔE1/E3 (MP Biomedicals) in E. coli BJ5183 cells, and the resulting construct was introduced into QBI-HEK293CymR after linearization with PacI RE. The resulting virus was plaque purified, amplified and stocks of purified virus were generated by CsCl banding. The physical titer of the virus was determined by A₂₆₀ method.³⁴

Adenovirus infections

Adenovirus infections were performed in 800 μl of DMEM plus supplements per 35-mm dish at the appropriate MOIs. MOIs were calculated assuming that infectious viral particles constitute 1% in our viral preparations.

Immunocytochemistry

For immunocytochemistry, the cells were grown on 13-mm glass coverslips, infected with adenovirus, stained with 400 μM MitoTracker Red (Invitrogen), fixed at −20 °C in methanol and permeabilized with 0.05% saponin in phosphate buffered saline (PBS). After rehydration with PBS, the anti-myc antibody (Cell Signalling Technology, Danvers, MA, USA) was applied for 1 h followed by fluorescein isothiocyanate-conjugated goat anti-mouse IGG (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The coverslips were mounted and observed using a Leica DM RXE microscope and a TCS SP2 confocal system in combination with a × 63 oil immersion objective.

Western blotting

Protein lysates were prepared using buffer containing 10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.5% SDS, 1 mM NaVaO₄, 30 mM NaF and EDTA-free protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN, USA). Following sonication and lysate clearing by centrifugation (13 000 g, 5 min) protein concentrations were determined using a DC Protein Assay (BioRad, Hercules, CA, USA). Proteins were separated on 10% polyacrylamide gel, and transferred to PVDF membrane. The membrane was probed with appropriate primary monoclonal antibodies raised against β-actin, FLAG-tag (Sigma-Aldrich, St Louis, MO, USA), total P53, S15 phospho-P53, myc-Tag (Cell Signalling Technology) or γ-H2AX (Affinity Bioreagents, Golden, CO, USA) followed by HRP-conjugated goat anti-mouse secondary antibody (Santa Cruz Biotechnology). The detection was performed using a SuperSignal West Pico kit (Pierce, Rockford, IL, USA).

Detection of T8993G mutation

The T8993G mutation in cybrid cells was analyzed by PCR-RFLP. Briefly, cells were lyzed in 10 mM Tris, pH 8.0, 1 mM EDTA, 0.5% SDS and 300 μg ml⁻¹ proteinase K. The lysate was incubated overnight at 37 °C, adjusted to 1 M with NaCl and subjected to three consecutive extractions with an equal volume of chloroform:isoamyl alcohol (24:1). The total DNA was precipitated with 2 volumes of 100% ethanol, redissolved in water, and subjected to PCR with primers flanking the site of mutation (cctctattgatccccacctcc and ggctggagtggtaaaa ggctc). The 915 bp PCR product was purified with a Qiaquick PCR purification kit and digested with AvaI or SmaI REs. Neither of these enzymes digests the WT PCR product, while the digestion of the PCR product containing the T8993G mutation with each enzyme resulted in 530 and 385 bp products.

Southern blotting

Total cellular DNA was extracted as described above for the detection of the T8993G mutation and digested with XbaI. Restriction fragments were separated in a 1% agarose gel, transferred to a nylon membrane and probed with a chimeric 1515 bp nuclear/mitochondrial probe (no. 1717) which consisted of 752 bp of the D-loop region of human mtDNA and 763 bp of the nuclear 18S rRNA gene. The probe was labeled by the random priming method³⁵ and hybridized according to standard protocols.³³

Oxygen consumption

Cells were dissociated by trysinization and collected by centrifugation at 800 g for 3 min at 25 °C. The cell pellet was washed in cold PBS (138 mM NaCl, 2.7 mM KCl, 12 mM Na₂HPO₄ · 7H₂O, 1.5 mM KH₂PO₄, pH 7.3), and resuspended in PBS buffer to 1 × 10⁷ cells per milliliter. A 0.3 ml portion of this suspension was used to measure the rate of oxygen consumption at 37 °C for 5 min. The oxygen consumption rate was measured polarographically using a Clark-type electrode (Oxygraph, Hansatech, UK) connected to a computer, which gave an on-line display of the rate value. Oxygen consumption was expressed as nmol O2 per minute per cell.

ATP content

The ATP content was determined using an ATP Bioluminescence Assay Kit HS II (Roche Diagnostics GmbH, Indianapolis, IN, USA) according to the manufacturer’s recommendations. Briefly, cells were collected by trypsinization, pelleted by centrifugation at 800 g for 3 min at 4 °C, resuspended in 0.5 ml of cell lysis reagent and incubated for 15 min at 4 °C. The chemiluminescence was measured after diluting samples 1:10. The ATP concentration was determined by comparing the chemiluminescence of the samples to chemiluminescence of a set of ATP standards and expressed in nmol mg⁻¹ protein.

ATP synthesis rate

The ATP synthesis rate was determined as described previously.³⁶ Briefly, cells were pelleted by centrifugation at 800 g for 3 min at 4 °C, resuspended to 1 × 10⁷ cell ml⁻¹ in buffer A (150 mM KCl, 25 mM Tris-HCl, 2 mM EDTA, 0.1% bovine serum albumin, 10 mM KH₂PO₄, 0.1 mM MgCl₂, pH 7.4). This suspension can be kept on ice for up to 45 min without loss of ATP synthesis activity. A 160 μl portion of suspension was incubated with 8 μl of digitonin (1 mg ml⁻¹) for 1 min at room temperature with gentle agitation. Cells were washed with 1 ml of buffer A and pelleted at 800 g for 3 min. Then the cell pellet was resusupended in 160 μl of buffer A containing 0.15 mM P¹, P⁵ Di(adenosine-5′) penthaphosphate pentasodium salt, 5 mM pyruvate, 2.5 mM malate, 0.2 mM ADP, 0.8 mM luciferin/20 μg ml⁻¹ luciferase/0.5 M Tris-acetate, pH 7.75. For each sample, one replicate tube was prepared containing the above components plus 10 μg ml⁻¹ oligomycin. The ATP sythase inhibitor olygomycin was added to obtain baseline luminescence corresponding to nonmitochondrial ATP production. The cells were placed in the luminometer, and the light emission was recorded. The ATP synthesis rate was expressed as nanomole per minute per 1.6 × 10⁶ cells.

Proliferation assays

Cells were cultured in six-well plates containing DMEM supplemented with either glucose (1 g l⁻¹) or galactose (1 g l⁻¹) as the sole carbon source. The initial plating density was 10 000 cells per well in glucose-containing medium and 50 000 cells per well in galactose-containing DMEM. Every 24 h cells were collected by trypsinization of triplicate wells per experimental condition, and cell count was determined using a particle counter (Beckman-Coulter, Fullerton, CA, USA). Results are presented as means ± s.e.m. (n = 3).

Lactate production

The lactate production was measured using Lactate Reagent according to the manufacturer’s recommendations (Trinity Biotech PLC, Co Wicklow, Ireland). For lactate assays 1–1.5 × 10⁵ cells were plated in 35 mm dishes in 2 ml DMEM without sodium pyruvate. Lactate levels were measured in the medium 24 h after plating and normalized to protein content of the cells collected by trypsinization of the same wells.