2.2. Drosophila cell culture and conditions for RNAi

Drosophila embryonic D.Mel-2 cells were maintained at 28 °C in HyQ SFM (HyClone) supplemented with 16 mM l-glutamine, 50 units/ml penicillin and 50 μg/ml streptomycin. For RNAi treatment, cells were diluted to a concentration of 1.0 × 10⁶ cells/ml in 10 ml of complete HyQ SFM in 75-cm² flasks. dsRNA (15 μg per 10⁶ cells) was added directly to the medium. Flasks were swirled by hand and cells were incubated at 28 °C for 1 h. Then, 10 ml of medium was added to obtain a cell density of 0.5 × 10⁶ cells/ml, an additional incubation at 28 °C for 72 h followed. D-MTERF5 and Opa1 dsRNAs were produced using the MEGAscript T7 transcription kit (Ambion). Templates were PCR-derived fragments carrying the T7 promoter sequence at both ends. The gene-specific sequences of the primers were as follows: D-MTERF5 (GenBank ID: NM_142436) forward-primer nt 269–284, reverse-primer nt 1054–1040; Opa1 (GenBank ID: NM_079504) forward-primer nt 303–317, reverse-primer nt 1017–1004. DREF RNAi was performed as reported by Fernández-Moreno et al. (2009).

2.3. Subcellular localization of D-MTERF5 by confocal microscopy

D.Mel-2 cells were diluted to a final concentration of 2.0 × 10⁶ cells/ml in 5 ml of antibiotic-free HyQ SFM in 75-cm² flasks 16 h prior to transfection. Ten micrograms of plasmid pMt/Hy containing the entire coding sequence of D-MTERF5 fused with the FLAG-tag sequence was added in the presence of 50 μg of Cellfectin (Invitrogen) in fresh antibiotic-free medium. Cells were incubated at 28 °C for 24 h and then induced with 500 μM CuSO₄. Twenty hours after induction, cells were settled down on poly l-lysine coated glass coverslips and mitochondria were stained with 250 nM MitoTracker Red CMXRos (Invitrogen-Molecular Probes) at 20 °C for 1 h in Drosophila SFM medium. Then cells were washed with phosphate buffered saline (PBS), fixed with 3.7% formaldehyde/PBS for 20 min and permeabilized with 0.1% Triton X-100 for 10 min. After blocking with 0.1% gelatin in PBS for 45 min, coverslips were incubated with mouse monoclonal anti-FLAG M2 antibody (Sigma, 1:500 dilution in 0.1% gelatin in PBS) at room temperature for 2 h, washed three times for 5 min each with 0.1% gelatin in PBS and incubated with an Alexa Fluor 488 conjugated anti-mouse antibody (Invitrogen-Molecular Probes, 1:1000 dilution in 0.1% gelatin in PBS) at room temperature for 1 h. Finally, coverslips were washed three times with PBS, rapidly rinsed in distilled water, and then mounted on glass slides with glycerol/PBS. Confocal images were captured by means of a Leica TCS SP5 confocal microscope (Leica Microsystems, Mannheim, Germany) using a 63 × oil immersion objective (N.A. = 1.32) and a 100 mW Argon laser (488 nm and 543 nm lines).

2.4. RT-PCR and real-time RT-PCR assays

Total cellular RNA was extracted from treated and untreated D.Mel-2 cells by the RNeasy Midi Kit (Qiagen). Semiquantitative RT-PCR assay was performed using the Enhanced Avian HS RT-PCR Kit (Sigma). Reactions contained 300–500 ng of Drosophila total RNA as template, and 50 pmol of sequence specific primers in a final volume of 50 μl.

For real-time RT-PCR assay, RNA was reverse-transcribed using the Enhanced AMV Reverse Transcriptase (Sigma). Each reaction was carried out in a final volume of 25 μl using 250 ng of total RNA and 25 pmol of mtDNA gene-specific primer. Real-time PCR was performed using the SYBR® Green PCR MasterMix (Applied Biosystems). Primers were designed using the Primer Express 2.0 software (Applied Biosystems) and sequence positions on D. melanogaster mtDNA (GenBank ID: NC_001709) were as follows: ND2-for nt 558–573, ND2-rev nt 651–632; COI-for nt 1838–1862, COI-rev nt 1932–1904; COII-for nt 3569–3593, COII-rev nt 3652–3631; COIII-for nt 5304–5328, COIII-rev nt 5379–5355; ND5-for nt 7040–7061, ND5-rev nt 7114–7095; ND4/4L-for nt 8561–8584, ND4/4L-rev nt 8722–8696; cyt b-for nt 10697–10719, cyt b-rev nt 10766–10748; ND1-for nt 11984–12002, ND1-rev nt 12056–12034;lrRNA-for nt 12827–12845, lrRNA-rev nt 12906–12887. RP49 rRNA (GenBank ID: NM_170461) and 28S rRNA (GenBank ID: M21017) were used as an endogenous control; primer sequence positions were: RP-for nt 189–208, RP-rev nt 319–301 and 28S-for nt 1407–1429, 28S-rev nt 1480–1462, respectively. Real-time RT-PCR reactions and analysis were conducted as described by Roberti et al. (2006).

2.5. Protein extract preparation and Western blotting analysis

Preparation of mitochondrial protein extracts from D.Mel-2 cells, as well as Western blotting detection of DmTTF, was carried out as described by Roberti et al. (2006). For Western blotting analysis of D-MTERF5, mitochondrial proteins were electrophoresed on 12% SDS-polyacrylamide gel and electro-transferred to PVDF (Hybond-P, GE Healthcare). Membranes were pre-incubated in 1 × PBS, 3% BSA for 1 h, and then incubated in the same solution for 16 h at 4 °C in the presence of polyclonal antibodies raised in rabbit against recombinant D-MTERF5 (Eurogentec). Membranes were then washed twice in 1 × PBS, 3% BSA and twice in 1 × PBS and incubated for 1 h with horseradish peroxidase-conjugated anti-rabbit IgG (Santa Cruz Biotechnology) and washed with 1 × PBS. Protein bands were visualized using the ECL Plus Western blotting reagents (GE Healthcare).

3.2. D-MTERF5 influences mitochondrial transcription

To obtain information on the role of D-MTERF5, a protein knock-down phenotype was produced by treating D.Mel-2 cells with a dsRNA encompassing 787 nt of the D-MTERF5 coding sequence. Depletion of D-MTERF5 mRNA was monitored by semiquantitative RT-PCR assay using primers delimiting the entire coding sequence (about 1680 bp). The level of D-MTERF5 mRNA in treated cells was almost undetectable up to 40 amplification cycles (Fig. 2A); the decrease was specific since no effect was obtained in mock control cells treated with dsRNA containing the sequence of the Opa1 gene (not shown).

Open in a separate window

Fig. 2

D-MTERF5 depletion affects mitochondrial transcription. (A) D-MTERF5-targeted RNAi in D.Mel-2 cells was monitored by RT-PCR. Total RNA (800 ng) extracted from cells untreated (cont.), treated with D-MTERF5 dsRNA (RNAi) was used as template in RT-PCR reactions. Samples were run on a 1.5% agarose gel and stained with ethidium bromide. Nuclear-encoded RP49 mRNA was used as endogenous control. (B) Relative quantification of mitochondrial transcripts by real-time RT-PCR. The mtDNA molecule is represented as a linear map. The directions of transcription of the (−) and (+) strands are indicated by black arrows; the two DmTTF binding sites are marked by vertical dotted lines and are indicated as T1 (the site at the boundary between the tRNA^Glu and tRNA^Phe genes) and T2 (the site at the boundary between the tRNA^Ser(UCN) and ND1 genes). Bars reported above the (−) strand and below the (+) strand, in correspondence to the position of the analyzed genes, indicate the relative content (R) of transcript, normalized to RP49 mRNA (endogenous control), in D-MTERF5 depleted (RNAi) with respect to control cells fixed as 1-value. The relative quantification was performed according to the equation R = (E_T)^ΔCt,T / (E_C)^ΔCt,C (Pfaffl, 2001). For each amplicon, the mean ratio value was obtained from at least four real-time RT-PCR assays, using RNA obtained from independent RNAi experiments; standard deviations are indicated on the top of the bars. Statistical analysis, performed using paired two-tailed Student's t-test showed that the differences between RNAi and control are all significant (P < 0.05).

Since D-MTERF5 belongs to the MTERF family, whose members include factors controlling mtDNA transcription and replication, we decided to test the effect of a D-MTERF5 knock-down on mtDNA content and mitochondrial transcript levels. MtDNA copy number, determined as a mtDNA/nDNA ratio by means of real-time PCR, was found not to vary significantly between control and RNAi cells (data not shown).

To investigate the effect of D-MTERF5 knock-down on mitochondrial transcription, we measured the steady-state level of nine RNAs distributed on both mtDNA strands. We found that the level of all analyzed transcripts varied significantly in depleted cells (Fig. 2B). In particular, the (−) strand transcripts ND2, COI, COII and COIII increased to about 30–50%, whereas cyt b mRNA decreased to about 40%. The (+) strand transcripts lrRNA and ND1 showed an increased level (40% and 50%, respectively); ND4/4L and ND6 mRNAs showed a 30% and 50% decrease, respectively.

To rule out the possibility that D-MTERF5 knock-down could have altered the levels of other proteins involved in mtDNA transcription and in turn, the abundance of mitochondrial transcripts, we measured the mRNA level of POLRMT, the initiation factors TFAM and TFB2M, the repressor MTERF3 and the termination factor DmTTF. In all cases we found no variations in RNAi cells with respect to control cells (data not shown); hence, we conclude that the observed changes in mitochondrial RNA levels are likely due to a direct consequence of the D-MTERF5 knock-down. Altogether these data strongly support the involvement of D-MTERF5 in Drosophila mitochondrial transcription.

Importantly, a clear correlation can be observed between the trend in the variation of transcript levels in DMTERF5 depleted cells and the location of DmTTF binding sites (see Fig. 2B and Roberti et al, 2003). Both the levels of the ND2, COI, COII and COIII transcripts in the (−) strand and the IrRNA and ND1 transcripts in the (+) strand, which, according to transcription direction, map upstream of the T1 and T2 DmTTF binding sites, respectively, increased upon knock-down of D-MTERF5. Conversely, levels of the cyt b mRNA in the (−) strand and the ND4/4L and ND5 mRNAs in the (+) strand, which map downstream of T1 and T2, respectively, decreased when D-MTERF5 is knocked down. Since, as we previously reported (Roberti et al, 2006) in DmTTF depleted cells, on both mtDNA strands, transcripts mapping upstream of the T1 and T2 binding sites were decreased, whereas transcripts mapping downstream of the protein binding sites were increased, the effect of D-MTERF5 RNAi on transcript levels is exactly the opposite to that produced by DmTTF knock-down. Therefore, the role of D-MTERF5 in transcription appears to be somehow antithetical to that of the transcription termination factor.

In the light of the possible involvement of D-MTERF5 in the DmTTF mediated termination process, we measured the in vivo levels of both proteins in D.Mel-2 cells by Western blot (Supplementary Fig. 1) and we found that the contents of D-MTERF5 and DmTTF were 0.8 ± 0.17 pmol and 2.3 ± 0.25 pmol per mg of total mitochondrial proteins, respectively. Therefore the amount of D-MTERF5 appears to be in the same order of magnitude of that of the transcription termination factor.

3.3. D-MTERF5 interacts with DmTTF binding sites on mtDNA

The above data prompted us to investigate the possibility that D-MTERF5 could interact with mtDNA at the same sites bound by DmTTF. We expressed a His-tagged mature form of D-MTERF5 in bacteria, and purified it at almost homogeneity by means of metal-affinity chromatography (Supplementary Fig. 2A). We first tested the capacity of recombinant D-MTERF5 to interact in an EMSA assay with a 139 bp fragment, which spans the tRNA^Glu and tRNA^Phe genes almost completely, and contains the DmTTF T1 binding site. As reported in Fig. 3A, no evident retarded complexes were obtained in the presence of a large excess of protein, thus arguing against the capacity of D-MTERF5 to bind this mtDNA region per se. On the other hand, the possible involvement of DmTTF and D-MTERF5 in the same process suggested that a physical interaction could occur between the two proteins. On this basis, and considering the nearly stoichiometric level of the two proteins observed in vivo, we investigated the binding of D-MTERF5 to the EMSA probe in the presence of recombinant purified DmTTF (see Supplementary Fig. 2B, for DmTTF purification). The results of the mobility shift experiment are shown in Fig. 3B. In the presence of DmTTF only (lane 2), we observed the typical DmTTF–DNA complexes with the faster migrating band (A) representing the most abundant complex formed by one DmTTF molecule, and the slower migrating band (B) consisting in the less abundant complex likely formed by two DmTTF molecules. When D-MTERF5 was added to the binding reaction, we observed additional retarded bands migrating more slowly than the DmTTF/DNA complex A; the abundance of these bands was dependent on the amount of D-MTERF5 (lanes 3 and 4, complexes I–III). Electrophoretic mobility suggests that complex I is generated by D-MTERF5 alone and complexes II and III contain both proteins, probably in various stoichiometries. Differently from complexes I and II, which resulted as poorly reproducible (see Fig. 3C) probably due to their low stability, complex III was the most abundant and always reproducible. The presence of D-MTERF5 in complex III was confirmed by using anti-His antibodies (Fig. 3C). Increasing amounts of antibodies added to the reaction mixture interfered with the binding of His-tagged-D-MTERF5 to DmTTF–DNA complex and caused the progressive decrease of complex III formation.

Open in a separate window

Fig. 3

Analysis of the DNA-binding properties of D-MTERF5. EMSA and footprinting assays were employed to test the interaction of D-MTERF5 with mtDNA at the tRNA^Glu/tRNA^Phe boundary region containing the T1 DmTTF binding site. A DNA fragment spanning position nt 6261–6399 of D. melanogaster mtDNA (GenBank ID: NC_001709) labeled with cyanine5 at the 5′-ends was used as a probe in the binding reactions. (A) EMSA was conducted by incubating the DNA probe (25 fmol) with 50 ng (about 0.8 pmol) of recombinant purified D-MTERF5. (B) EMSA was conducted by incubating the DNA probe (25 fmol) with 25 or 12.5 ng (about 0.4 and 0.2 pmol, respectively) of recombinant purified D-MTERF5 in the presence of 50 ng (about 1.2 pmol) of recombinant purified DmTTF. Complexes formed by DmTTF are indicated by capital letters, and complexes obtained following the addition of D-MTERF5 are indicated by Roman numerals. (C) The indicated amounts of anti-6xHis antibodies were added to the EMSA reaction mixture containing 25 fmol of DNA and 50 ng of both D-MTERF5 and DmTTF (about 0.8 and 1.2 pmol, respectively). (D) DNase I protection analysis was conducted on protein–DNA complexes obtained by adding to 100 fmol of DNA fragment labeled at the 5′-end of either (+) or (−) strand, 500 ng (about 8.2 pmol) of purified DmTTF or 500 ng of DmTTF (about 8.2 pmol) plus 250 ng of D-MTERF5 (about 4.0 pmol). Protected regions are indicated by brackets; the asterisk indicates the DNase I-hypersensitive site.

On the basis of the EMSA results we investigated the capacity of D-MTERF5 to directly contact mtDNA in the presence of DmTTF by means of DNaseI footprinting experiments. When the EMSA probe labeled in the (+) strand was incubated with both DmTTF and D-MTERF5 (Fig. 3D, lane 3), the typical protection by DmTTF (6314–6339) was extended by about 10 nucleotides in the direction of the tRNA^Phe gene. When the DNA fragment was labeled in the (−) strand, addition of D-MTERF5 (Fig. 3D, lane 6) caused a weak protection extending by about 15 nucleotides in the direction of the tRNA^Glu gene. A clear hypersensitivity site at the boundary of the DmTTF footprint was also observed. The obtained data indicate that D-MTERF5 modifies the DNaseI protection pattern produced by DmTTF.

3.4. Homology modeling of D-MTERF5

Recently the crystal structure of mTERF and its binding mode to mtDNA have been clarified (Jiménez-Menéndez et al., 2010; Yakubovskaya et al., 2010). The protein shows a modular architecture due to the repetition of eight-nine 35-residue mterf motifs, consisting of three short alpha-helices forming a left handed triangular superhelix folded around a hydrophobic core. The motifs form a solenoid-like structure, which twists to the right acquiring a convex surface and a positively charged concave surface that accommodates the DNA duplex. Sequence specific interactions with DNA cause protein and DNA conformational changes that are necessary for promoting transcription termination. The crystal structure of human MTERF3 has been also solved (Spåhr et al., 2010). It is very similar to that of mTERF and consists in the repetition of seven mterf motifs; however, MTERF3 lacks the residues that in mTERF are involved in sequence-specific DNA binding.

The primary structure of D-MTERF5 also contains several repetitions of the mterf motif. Therefore, it is highly probable that the protein adopts a three-dimensional shape similar to that determined for human mTERF and MTERF3. Homology modeling of D-MTERF5 performed using the Swiss-Model server confirmed this hypothesis, showing that the predicted structure of the protein is arranged as the mTERF-like series of left handed super-helices forming a right twisted solenoid structure (Fig. 4A). D-MTERF5 would also possess the positively charged concave surface able to accommodate the negatively charged double helix, being therefore, in principle, able to bind DNA (Fig. 4B). Interestingly, the predicted structure shows the presence, in both the N- and C-terminal halves of the protein, of a beta–alpha–beta domain protruding from the convex surface, which interrupts the series of super-helices. The position of the predicted mterf motifs and beta-alpha-beta domains is depicted on the D-MTERF5 sequence reported in Supplementary Fig. 3.

Open in a separate window

Fig. 4

Homology modeling of D-MTERF5. (A) Overall structure of D-MTERF5 (residues 62–443 of the mature protein) obtained by comparison with human mTERF, by using the Swiss-Model Automated Protein Modeling Server, in which predicted alpha-helices are in red and beta-sheets are in yellow. (B) Molecular surface of D-MTERF5 colored according to the electrostatic surface potential (blue + 12 kT, red − 12 kT).

It is known that beta-sheet domains can be used by proteins (as homo- or hetero-dimerization domains) to establish contacts with interacting partners (Remaut and Waksman, 2006); therefore, it can be supposed that D-MTERF5, at the termination sites, could use the beta–alpha–beta domains to establish interactions with other proteins, such as POLRMT or the termination factor. Interestingly, structure modeling suggests that DmTTF also contains two short beta-sheet domains (not shown).

3.5. D-MTERF5 expression is controlled by DREF

We analyzed the control of D-MTERF5 gene expression by DREF (DNA Replication-related Element binding Factor). DREF is a nuclear factor that activates genes involved in nuclear DNA replication and cell cycle control (Matsukage et al., 2007), and is also able to activate genes involved in mtDNA replication and maintenance (Lefai et al., 2000; Ruiz De Mena et al., 2000; Takata et al., 2001), as well as the DmTTF gene (Fernández-Moreno et al., 2009). We found DRE sequences in the D-MTERF5 promoter at positions-11 and -37; they showed 8/8 and 7/8 matches with respect to the DRE consensus TATCGATA and were highly conserved in other Drosophila species (Fig. 5A). We investigated the effect of DREF depletion on the level of D-MTERF5 transcript. DREF knock-down (more than a 90% decrease of the polypeptide level) was obtained by RNAi treatment in Drosophila S2 cultured cells, as already reported by Fernández-Moreno et al. (2009). The results from three independent RNAi experiments indicated that in DREF-depleted cells the steady-state level of D-MTERF5 mRNA, measured by real-time RT-PCR, is diminished to about 70% with respect to non-depleted cells (Fig. 5B), thus demonstrating a positive control by DREF on D-MTERF5 expression. Co-regulation by DREF of DmTTF and D-MTERF5 suggests that expression of the two factors needs to be co-ordinated to ensure modulation of Drosophila mitochondrial transcription.

Open in a separate window

Fig. 5

Regulation of D-MTERF5 expression by DREF. (A) Multialignment of D-MTERF5 proximal promoters from different Drosophila species showing the location and conservation of DRE sequences. Promoter nucleotide sequences were obtained from FlyBase Database. ClustalW alignment was performed at NPS@ web server of the PBIL. The arrow indicates the transcription start point according to the D. melanogaster + 1 site as in FlyBase. DRE sequences within the promoters are indicated by solid bars; the DRE consensus sequence is also shown above bars. (B) Effect of DREF knock-down on D-MTERF5 mRNA levels. DREF protein was knocked down in S2 cells by means of RNAi as reported by Fernández-Moreno et al. (2009). Total RNA was extracted from Drosophila S2 cells, either untreated (control) or treated with LacZ (mock) or DREF dsRNA; relative quantification of D-MTEF5 mRNAs was carried out by real-time RT-PCR. Bars indicate the relative content of transcript, normalized to 28S rRNA (endogenous control), with that in control cells fixed as 1-value. Relative quantification and statistical analysis were performed as described in the Fig. 2 legend (* indicates P < 0.05).

This work was supported by grants from: Università di Bari, Progetto di Ricerca di Ateneo, M.I.U.R. COFIN-PRIN 2008 and Telethon — Italy (grant GGP06233). Authors thank T. De Filippis (Dept. of Biosciences, Biotech. and Pharmacol. Sciences, Univ. of Bari) for her precious technical assistance and F. Fracasso (Dept. of Biosciences, Biotech. and Pharmacol. Sciences, Univ. of Bari) for his help in preparing figures and manuscript.