Mitochondrial transport of catalytic RNAs and targeting of the organellar transcriptome in human cells

Abstract Mutations in the small genome present in mitochondria often result in severe pathologies. Different genetic strategies have been explored, aiming to rescue such mutations. A number of these strategies were based on the capacity of human mitochondria to import RNAs from the cytosol and designed to repress the replication of the mutated genomes or to provide the organelles with wild-type versions of mutant transcripts. However, the mutant RNAs present in mitochondria turned out to be an obstacle to therapy and little attention has been devoted so far to their elimination. Here, we present the development of a strategy to knockdown mitochondrial RNAs in human cells using the transfer RNA-like structure of Brome mosaic virus or Tobacco mosaic virus as a shuttle to drive trans-cleaving ribozymes into the organelles in human cell lines. We obtained a specific knockdown of the targeted mitochondrial ATP6 mRNA, followed by a deep drop in ATP6 protein and a functional impairment of the oxidative phosphorylation chain. Our strategy provides a powerful approach to eliminate mutant organellar transcripts and to analyse the control and communication of the human organellar genetic system.


Induction of transgene expression
Transgenes in the pcDNA5/FRT/TO vector are under the control of the CMV/TetO2 tetracycline-inducible RNA polymerase II promoter.Tetracyline (from the K650001 core kit, ThermoFischer) was resuspended at 1 mg/ml in sterile water, filtered (0.22 μm filter) and stored at -20°C protected from light.Transgene expression was induced with tetracycline in transformed cells at 80-90% confluence.Tetracycline concentrations ranging from 0.5 to 10 µg per ml culture medium were tested.

Preparation of labeled RNA substrates for in vitro organellar import
For mitochondrial import assays, radiolabed BMV TLS and TMV TLS transcripts were generated by in vitro transcription of the above BstNI-linearized constructs in the presence of [α-32 P]UTP.Transcription reactions were performed using a MEGAshortscript T7 Kit (ThermoFischer) as described by the manufacturer.Reaction products were purified using the NucAway Spin Column kit (ThermoFischer) according to the recommended protocol.

Isolation of mitochondria
Tetracycline-induced transgenic cells or standard HepG2 cells grown to 90-95% confluence were harvested with a trypsin solution [0.025% (w/v) trypsin, 0.01% (w/v) EDTA in phosphate-buffered saline (PBS), ThermoFischer] and washed with PBS.Mitochondria were subsequently isolated using a Mitochondria Isolation Kit (Mitochondria MidiMACS Starting Kit, Miltenyl Biotec) according to the manufacturer's protocol.Mitoplasts were generated by hypotonic shock.The mitochondrial pellet was resuspended in 100 μL of breakage buffer [BB; 0.6 M mannitol, 1 mM EDTA, 10 mM Na-PIPES, pH 6.7, 0.3% (w/v) BSA] and diluted 10 times in 10 mM HEPES, pH 6.8.After 10 min incubation on ice, sucrose was added up to 0.25 M. The resulting mitoplasts were harvested by centrifugation for 5 min at 10 000 g and washed twice with BB.Residual cytosolic RNA contamination was removed by addition of a mixture of nucleases (10 units/ml of micrococcal nuclease, 100 μg/ml of RNase A and 25 units/ml of phosphodiesterase) and incubation for 15 min at 37°C.Nuclease treatment was followed by an incubation of the organelles at 20°C for 5 min in the presence of 1 mM CaCl2 and the mitochondrial RNA was finally isolated.
Alternatively, cells were resuspended in homogenization buffer [mannitol 0,6 M, Tris-HCl (pH 7,4) 10 mM, EGTA 10 mM, BSA 0.1% (w/v)] and lyzed in a Dounce homogenizer.The suspension was centrifuged successively at 600 g for 5 min at 4°C and 11 000 g for 10 min at 4°C.The final pellet was resuspended in 500 μL of homogenization buffer containing 50 μg RNase A and 375 units RNase T1.After 15 min incubation at room temperature, 500 µL of homogenization buffer, 1 μL of phenylmethylsulfonyl fluoride (PMSF, 0.1 M) and 1 μL of proteinase K (10 mg/ml) were added prior to 30 min incubation on ice.Mitochondria were subsequently re-isolated on a discontinuous Percoll gradient [13,5%, 21%, 45% (v/v)] and washed with homogenization buffer without BSA completed with 10 mM EGTA and 10 mM EDTA.To prepare mitoplasts, mitochondria were suspended in 500 μL of homogenization buffer and digitonine (170 μg/mg mitochondrial proteins) was added.After 15 min incubation on ice, mitoplasts were recovered by centrifugation at 11 000 g for 15 min at 4°C and washed with homogenization buffer without BSA before treatment with RNases A and T1.

RNA extraction and RT-qPCR analyses
RNA was extracted from whole cells and mitochondrial preparations according to standard TRI Reagent protocols (Invitrogen).Remaining DNA was digested with DNase I using the DNA Free reagent kit (ThermoFischer) as described by the manufacturer.
Reverse transcription was carried out with the Transcriptor First Strand cDNA Synthesis Kit (Roche) according to the manufacturer's protocol.RNA sample quality and absence of remaining DNA was assessed through standard RT-PCR reactions and the amplification products were analyzed on agarose gel.RT-real-time PCR was subsequently run on a CFX Connect thermal cycler (Bio-Rad).Reaction mixtures were prepared with the LightCycler 480 SYBR Green I Master reagent Kit (Roche) according to manufacturer's protocols.As reference genes we used those for actin beta (ACTB), glyceraldehyde 3phosphate dehydrogenase (GAPDH), hypoxanthine-guanine phosphoribosyl transferase 1 (HPRT1) and porphobilinogen deaminase (PBDG).Analyses in relation to the reference genes was performed by the relative quantification method (ΔΔCt method).Three independent biological replicates were analyzed.

FLOE
For fluorescently labeled oligonucleotide extension assays (FLOE, Lloyd et al., 2005), reverse transcription was run with 5 µg of mitochondrial RNA and 5 nmol of the atp6_atp8_FAM reverse primer (Supplementary Table S1) labeled at the 5'-end with 6-Carboxyfluorescein (6-FAM) using the Transcriptor First Strand cDNA Synthesis Kit (Roche).After RNase A digestion, the cDNAs were ethanol precipitated, redissolved in water and completed with 0.5 µL of GeneScan 1200 LIZ dye Size Standard (Applied Biosystems).Samples (10 µL final volume) were finally analyzed with a capillary electrophoresis sequencer (Applied Biosystems).

Western blot analyses
Transgenic lines were grown in 6-well plates to a confluence of 85-90% and treated at Day 0 with 10 µg/ml tetracycline for induction of transgene expression.Cell samples were subsequently taken every day until Day 6 and lysed by sonication in 10 mM Tris-HCl, pH 7, containing a cocktail of protease inhibitors (cOmplete Mini, EDTA-free, Roche).The lysate was centrifuged at 15 000 g for 10 min and the supernatant was transferred to new tubes.Protein samples (40 µg) were denatured, separated by SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes following standard protocols.Membranes were blocked with 5% (w/v) skimmed milk.After incubation with primary and secondary antibodies, proteins of interest were detected with the SuperSignal West Pico PLUS Chemiluminescent Substrate (Life Technologies).The MT-ATP6 protein was detected with polyclonal antibodies (Life Technologies) at a 1:500 dilution.The MT-COX1 and MT-COX2 proteins were detected with monoclonal antibodies (Cell Signaling Technology) at a 1:1000 (COX1) or 1:750 (COX2) dilution.As a control, the GAPDH protein was detected with the GAPDH 0411 monoclonal antibody (Santa Cruz Biotechnology) at a 1:500 dilution.The peroxydase-conjugated antirabbit A6154 antibody (Sigma-Aldrich) was used as a secondary antibody at a 1:10000 dilution.Antibodies were diluted in 5% (w/v) skimmed milk.The intensity of individual bands was analyzed quantitatively with the Multi Gauge ver.2.0 software (Fujifilm).The relative level of MT-ATP6, MT-COX1 and MT-COX2 protein was normalized against the GAPDH level.

Cytotoxicity assays
For cytotoxicity assays, cells were grown in 96-well plates to a confluence of 85-90% and treated with 0.5 to 10 µg/ml tetracycline.After 24 h, the culture medium was withdrawn and 100 µL of MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] at 5 mg/ml in EMEM medium were added to each well.The plates were incubated for 90 min at 37°C.The MTT solution was then withdrawn and 100 µL of DMSO (dimethyl sulfoxide) were added per well, followed by shaking for 10 min at 230 rpm to evenly dissolve the formazan generated.The amount of MTT reduced to the purple formazan was finally measured spectrophotometrically at a wavelength of 590 nm using a BioTek Synergy Microplate Reader.

Mitochondrial membrane potential (ΔΨm) analysis
The JC-1 cationic carbocyanine dye was used as a probe to evaluate the mitochondrial membrane potential status in whole cells.In a regular physiological state, the JC-1 dye is driven into mitochondria along the membrane potential and aggregates in the organelles.The so-called J-aggregate form can be quantified by reading the corresponding emission of red fluorescence at 590 nm.In cells with an impaired mitochondrial membrane potential, the JC-1 dye remains in its monomeric form in the cytosol and emits green fluorescence at 529 nm.Fluorescence of both the aggregated and the monomeric form can be triggered by light with a wavelength of 514 nm.Carbonyl cyanide m-chlorophenyl hydrazine (CCCP) was used as a positive control.CCCP acts as a ionophore and directly dissipates the potential of the mitochondrial membrane.Cells treated with CCCP should thus fluoresce only in the range of 529 nm.
Stably transformed cells were grown in 96-well plates to a confluence of 85-90% and treated at Day 0 with 10 µg/ml tetracycline for transgene induction.Test samples were subsequently generated every day.Day 0, before transgene expression, was considered as the regular physiological state.For each time point, the culture medium was withdrawn and replaced by 100 µL of DMEM (for Flp-In T-REx cells) or EMEM (for Flp-In HepG2 cells) medium complemented with 2 μM of JC-1 dye (ThermoFischer).The plates were subsequently incubated for 30 min at 37°C and the medium was replaced by 100 µL of warm PBS per well.Total fluorescence at 529±5 nm and 590±5 nm was measured using a fluorescence reader.Then, a representative sample of the measured wells was imaged using a confocal microscope in the same wavelength range.Double reading was used to eliminate false readings caused by poor condition of the cells.

Analysis of energy metabolism
Activity of the OXPHOS chain was evaluated with an Agilent Seahorse XFp analyzer using a Seahorse XF Cell Mito Stress Test Kit following the instructions of the manufacturer.Built-in injection ports on the XFp sensor cartridges allow to add OXPHOS modulators into the wells carrying the cells to reveal the key parameters of mitochondrial function.The modulators included in the kit are Oligomycin, Carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (FCCP), Rotenone and Antimycin A. They are injected into the wells in this order.
To run the assays, stably transformed cell lines were seeded in Agilent XFp 8-well cell culture microplates (7.5 × 10 3 cells in 100 µl medium per well).At 80-90% confluence, transgene expression was induced with 10 µg/ml tetracycline and the oxygen consumption rate (OCR) of the cells was tested in the absence and presence of the above mentioned effectors.Briefly, the culture medium was removed from each well and replaced with 180 µL of Seahorse XF Base Medium containing 1 mM pyruvate and 2 mM glutamine, supplemented with 10 mM glucose and prewarmed to 37°C.The cells were incubated in a CO2-free incubator at 37°C for 1 h.Prior to measurements, the XFp analyzer gently mixed the assay media in each well for 10 min to allow the oxygen partial pressure to reach equilibrium.The OCR and the extracellular acidification rate (ECAR) were measured simultaneously three times to establish a baseline rate.The OCR of the samples was tested over 7 min time intervals.The Agilent Seahorse XF Cell Mito Stress Test Report Generator automatically calculates the Agilent Seahorse XFp Cell Mito Stress Test parameters from Wave data that have been exported into MS Excel.

Statistical analyses
Results represent mean values of three independent biological replicates ± standard deviation (SD).Statistical significance was evaluated using the GraphPad Prism ver.5.1 (GraphPad) software.Differences between the mean values of the test and the control samples were evaluated using ANOVA variance extended by Tukey or Bonferroni post hoc tests.Statistically significant results were assigned as: * for p<0.05; ** for p<0.01; *** for p<0.001; no statistical significance for p≥0.05.

The TYMV PKTLS tRNA-like shuttle is unable to translocate into the mitochondrial matrix in human cells
In line with our previous success in plants (Val et al., 2011;Sultan et al., 2016;Niazi et al., 2019), we tested the possiblity to translocate the TYMV PKTLS (i.e. the last 120 nucleotides at the 3'-end of the viral genomic RNA) into mitochondria in human cells.The corresponding sequence was assembled with the HDV cis-ribozyme sequence (Perrotta and Been, 2007) (cHDV) at its 3'-end in the pcDNA5-135-PKTLS-cHDV plasmid under the RNA polymerase II promoter of the pcDNA5/FRT/TO vector (see Materials and Methods).Flp-In T-REx 293 human kidney cells stably transformed with this plasmid were generated.Upon expression of a PKTLS-cHDV transcript mediated by RNA polymerase II, self-cleaving of the HDV cisribozyme was expected to eliminate all downstream vector sequences together with the polyA tail.Conversely, due to the location of the polymerase II transcription start site with respect to the cloning site of the transgene in the construct, 135 nucleotides of vector sequence were added as a cargo to the 5'-end of the PKTLS in the transcripts generated (Supplementary Figure S1A).This cargo sequence however was not coding and carried no function in the cells.
The possibility to express the TYMV PKTLS, i.e. a plant viral sequence, in human cells was assessed by RT-PCR upon transgene induction with tetracycline.Mitochondria were subsequently isolated from cells at 24 h after induction.Half of the organelles was directly used for further processing, while the other half was submitted to digitonine treatment, so as to generate mitoplasts by breaking the outer membrane and exposing RNAs that would be stuck in the intermembrane space.Both organelle fractions were submitted to RNase treatment prior to lysis and RNA extraction.RT-PCR analysis with specific primers (Supplementary Table S1) highlighted the presence of the PKTLS RNA in intact mitochondria, but not in mitoplasts (Supplementary Figure S1B).Conversely, the mitochondrial ND3 control mRNA was present in both fractions, assessing that the matrix material was not lost during mitoplast preparation.The TYMV PKTLS thus appeared to be able to translocate through the mitochondrial outer membrane, but unable to cross the inner membrane in human cells.
Supplementary Figure S1 Mitochondrial translocation of the TYMV PKTLS in Flp-In T-REx 293 human kidney cells stably transformed with the pcDNA5-135-PKTLS-cHDV construct.A Schematic representation of the primary transcript deriving from the transgene.The final product carries 135 nucleotides of upstream vector sequence, while self-cleaving of the cHDV ribozyme eliminates downstream vector sequences and releases the PKTLS 3'-CCA end.B RT-PCR analysis of RNAs extracted from mitochondria (Mito) or mitoplasts (Mpl).RNAs were probed for the TYMV PKTLS (PKTLS), the cytosol-specific tRNA Pro (Cyto tRNA Pro ) and the mitochondrial ND3 mRNA (Mito ND3).RT-PCR products were fractionated on agarose gel.
Following these observations, we tested two adaptations of our strategy.The first idea was to test whether the presence of the 135 nucleotide cargo sequence at the 5'-end of the PKTLS moiety was preventing final translocation through the inner membrane.In this respect, we included in a second construct called pcDNA5-cHh-PKTLS-cHDV, a cis-cleaving hammerhead ribozyme (cHh) upstream of the PKTLS sequence, in addition to the downstream cHDV, so that after transcription and self-cleavage of the upstream and downstream ribozymes the final product was expected to be the PKTLS moiety alone (Supplementary Figure S2A).The cis-cleaving hammerhead ribozyme was designed according to Fechter et al. (1998).As mentioned above, little information is available on the set of tRNAs capable of translocation into human mitochondria.The second option was thus to change the aminoacylation specificity of the TYMV PKTLS moiety.A TYMV PKTLS variant with methionine specificity has been described earlier, upon mutations in the acceptor stem pseudoknot and anticodon switch from CAC to CAU (Dreher et al., 1996).This PKTLSmet variant served as negative organellar import control in our previous studies in plants (Val et al., 2011), as cytosolic tRNA Met is not found in plant mitochondria (Salinas et al. 2008).Whether methionine specificity can provide importability into human mitochondria has not been documented.We tested such a possibility with the pcDNA5-cHh-PKTLSmet-cHDV construct encoding the TYMV PKTLSmet variant flanked with an upstream hammerhead cisribozyme and with the downstream HDV cis-ribozyme (Supplementary Figure S2B).The pcDNA5-cHh-PKTLS-cHDV construct and the pcDNA5-cHh-PKTLSmet-cHDV construct were both properly expressed in stably transformed Flp-In T-REx 293 cells and their products were recovered in intact mitochondria, but neither the valine-specific PKTLS without a 5' cargo nor the PKTLSmet variant was detected in mitoplast fractions.

BMV and TMV tRNA-like structures are taken up by isolated human mitochondria in vitro
Failure to obtain complete translocation of the valine-specific TYMV PKTLS, or of its methionine accepting mutant derivative, into mitochondria in human cells prompted us to test other plant viral TLSs with further aminoacylation specificities.Apart from valine, two other native aminoacylation specificities have been documented for plant viral TLSs, i.e. histidine and tyrosine (Mans et al., 1991).The BMV TLS is a representative of the tyrosine-accepting group, while the TMV TLS is a histidine-accepting representative and we decided to evaluate the possibility to use one or the other of them as an organellar shuttle in human cells.
As a first approach, we tested whether the BMV or TMV TLS can be taken up by isolated human mitochondria.To that end, the last 200 nucleotides from the 3'-end of each genomic RNA were amplified by PCR from BMV and TMV cDNAs (a gift from David Gilmer and Salah Bouzoubaa, IBMP, Strasbourg) and cloned under the control of the T7 RNA polymerase promoter.A BstNI restriction site was placed at the 3' end of the TLS sequences.The resulting templates linearized with BstNI were transcribed with T7 RNA polymerase in the presence of [α-32 P]UTP, yielding radiolabeled BMV and TMV TLS with a regular CCA end.Both were used for in vitro import assays with mitochondria isolated from standard HepG2 human cells.Importing tRNAs into human mitochondria in vitro needs, among other factors, the presence of import-directing proteins (Entelis et al., 2001).Human import directing proteins (HmIDPs) were prepared, also from HepG2 cells, and separated into fractions precipitating at 30, 60 and 90% ammonium sulfate saturation.Following the import step, one part of the mitochondria was analyzed directly, while mitoplasts were generated from the other part.In these assays, the BMV TLS and the TMV TLS were both recovered in the final RNA fractions from mitochondria and from mitoplasts (Supplementary Figure S3), implying that they can both be taken up by isolated human organelles.All three HmIDP fractions stimulated the import (Supplementary Figure S3).

SupplementaryFigure S2
Schematic representation of the primary transcript deriving from the transgene in Flp-In T-REx 293 human kidney cells stably transformed with the pcDNA5-cHh-PKTLS-cHDV construct (A) or the pcDNA5-cHh-PKTLSmet-cHDV construct (B).The final products are restricted to the PKTLS (A) or the PKTLSmet (B) sequence upon self-cleaving of the 5' cHh ribozyme and the 3' cHDV ribozyme.Nucleotide mutations switching the aminoacylation specificity from valine to methionine are indicated in blue on a white background (B).