Experimental relocation of the mitochondrial ATP9 gene to the nucleus reveals forces underlying mitochondrial genome evolution

Abstract

Only a few genes remain in the mitochondrial genome retained by every eukaryotic organism that carry out essential functions and are implicated in severe diseases. Experimentally relocating these few genes to the nucleus therefore has both therapeutic and evolutionary implications. Numerous unproductive attempts have been made to do so, with a total of only 5 successes across all organisms. We have taken a novel approach to relocating mitochondrial genes that utilizes naturally nuclear versions from other organisms. We demonstrate this approach on subunit 9/c of ATP synthase, successfully relocating this gene for the first time in any organism by expressing the ATP9 genes from Podospora anserina in Saccharomyces cerevisiae. This study substantiates the role of protein structure in mitochondrial gene transfer: expression of chimeric constructs reveals that the P. anserina proteins can be correctly imported into mitochondria due to reduced hydrophobicity of the first transmembrane segment. Nuclear expression of ATP9, while permitting almost fully functional oxidative phosphorylation, perturbs many cellular properties, including cellular morphology, and activates the heat shock response. Altogether, our study establishes a novel strategy for allotopic expression of mitochondrial genes, demonstrates the complex adaptations required to relocate ATP9, and indicates a reason that this gene was only transferred to the nucleus during the evolution of multicellular organisms.

Figure 1. Deletion of the yeast mitochondrial ATP9 gene and resulting phenotypes.

A) The mitochondrial ATP9 gene was deleted and replaced with ARG8^m (see Figure S1 for details) in a wild-type strain lacking the nuclear ARG8 gene. As a result, the Δatp9 yeast grow on glucose (Glu) media lacking arginine (Arg) whereas the parental strain (WT) does not; in addition, Δatp9 yeast cannot grow on glycerol (Gly). B) ATP synthase levels in WT and Δatp9. Isolated mitochondria were separated by BN-PAGE and western blotted with antibodies against Atp4p; V ₁ and V _n respectively indicate monomeric and oligomeric forms of ATP synthase. C) Pulse labelling of proteins translated in mitochondria. Total proteins were prepared from cells incubated in the presence of ³⁵S methionine and cysteine as well as cycloheximide to inhibit cytosolic protein synthesis. Proteins (40,000 cpm per lane) were separated on 12% (Cox3p and Atp6p) or 17% (Atp9p and Atp8p) SDS-PAGE containing 6 M urea. D) Electron microscopy of WT (a) and Δatp9 (b–d) cells grown in galactose (80 nm-thin sections); m, mitochondria; Cr, cristae; Ib, inclusion bodies; arrowheads in (a) point to Cr, to outer mitochondrial membrane in (c), and to septae in (d); bars, 0.2 µm.

Figure 2. A nuclear version of the yeast mitochondrial ATP9 gene fails to complement the Δatp9 yeast.

We engineered a nuclear version of the yeast ATP9 gene (yAtp9-Nuc) by adding a mitochondrial targeting sequence (derived from the P. anserina Atp9-7 gene) and adjusting the genetic code for nuclear expression of the endogenous gene (see Figure S2 and S3A for amino acid and nucleotide sequences). yAtp9-Nuc was tested for its capacity to complement Δatp9 yeast with respect to respiratory capacity. A) Growth on rich glucose (YPGA) and glycerol (N3) media of serial dilutions of WT, Δatp9, and Δatp9 transformed with yAtp9-Nuc. B) Total cellular (T), mitochondrial (M) and post-mitochondrial supernatant (C) protein extracts were prepared from WT and Δatp9+yAtp9-Nuc strains. Samples were separated via SDS-PAGE and probed with antibodies against yeast Atp9p and the cytosolic protein Pgk1p (phosphoglycerate kinase). C) Western blot of total proteins prepared from WT and Δatp9 yeast transformed with yAtp9-Nuc with or without the YME1 gene (Δyme1) reveals that the yAtp9-Nuc protein is degraded by the i-AAA protease (an oligomer of Yme1p).

Figure 3. The P. anserina Atp9 genes restore respiratory competence in Δatp9 yeast.

The Δatp9 strain was transformed with CEN or 2 µ plasmids harbouring PaAtp9-7 or PaAtp9-5, yielding strains AMY8 (Δatp9+PaAtp9-7, CEN), AMY11 (Δatp9+PaAtp9-7, 2 µ), AMY7 (Δatp9+PaAtp9-5, CEN), and AMY10 (Δatp9+PaAtp9-5, 2 µ). A) Growth curves of all strains in rich glycerol/ethanol medium at 28°C (subsequent panels use the same growth conditions). B) ATP synthase levels in WT and AMY10 revealed by separation of isolated mitochondria by BN-PAGE. The ATP synthase complexes (V₁, monomer; V_n, oligomers) and free F₁ were revealed in-gel by their ATPase activity (right). C) Total cellular (T) and mitochondrial (M) protein extracts were prepared from WT, AMY7 and AMY10 grown in YPEG. Samples (20 µg) were separated by SDS-PAGE and probed with antibodies against yeast Atp6p and γ-F₁. D) Differential interference contrast microscopy (left, ‘Nomarski’) and DAPI staining/fluorescence microscopy (middle) of WT and AMY10 cells. Right (‘Ultrastructure’) are electron micrographs of AMY10 cells (80 nm-thin sections). V, vacuole; n, nucleus; m, mitochondria.

Figure 4. The P. anserina Atp9 proteins are less hydrophobic than yeast Atp9p.

A) Hydropathy profiles of the PaAtp9-7 and PaAtp9-5 proteins and yeast Atp9p, generated according to with a window size of 13. B) P. anserina strains expressing exclusively either PaAtp9-7 (PaΔAtp9-5) or PaAtp9-5 (PaΔAtp9-7) were constructed and ATP synthase was enriched from their mitochondrial extracts, separated by SDS-PAGE and silver-stained along with WT yeast ATP synthase. Positions of some ATP synthase subunits are indicated. The PaAtp9-5 protein is stained much more strongly than the PaAtp9-7 protein, which may be due to the differences in their amino acid sequences.

Figure 5. Reducing the hydrophobicity of the first transmembrane segment of yeast Atp9p improves its import into mitochondria.

The Δatp9 strain was transformed with a hybrid Atp9 gene (Atp9-Hyb) encoding the mitochondrial targeting sequence (MTS) and first transmembrane segment (TMH1) of the PaAtp9-7 protein, followed by the connecting loop and second transmembrane segment (TMH2) of yeast Atp9p (see Figure 2A for amino acid sequence and Figure S3E for nucleotide sequence). Total cellular (T), mitochondrial (M) and post-mitochondrial supernatant (C) protein extracts were prepared from WT and Δatp9+Atp9-Hyb strains. Samples were separated by SDS-PAGE and probed with antibodies against yeast Atp9p and the cytosolic protein Pgk1p (phosphoglycerate kinase).

Figure 6. Transcriptome profiles of yeast strains expressing P. anserina Atp9 genes indicate functional OXPHOS and regulatory responses to the nuclear relocation of ATP9.

For all genes in the yeast genome, the expression levels in AMY11 (expressing PaAtp9-7) are plotted against those of AMY10 (PaAtp9-5), both displayed as log₂ ratios to WT expression levels; differentially expressed genes in the main functionally relevant categories are indicated by colours. The square formed by the grey lines delineates the boundaries of statistically significant expression differences (see Text S1) between either strain and the WT; genes beyond the diagonal grey lines are differentially expressed between AMY10 and AMY11. For clarity, genes in the categories listed are only indicated if they were differentially expressed relative to WT in at least one strain. Categories were defined as follows: OXPHOS pathway - subunits (1/35 differentially expressed) and biogenesis factors (1/42); Retrograde pathway – transcriptional targets of the factors Gcn4p (29/126) and Rtg3p (6/31), plus CIT2 and CIT3; Heat response - Gene Ontology (GO)-annotated “response to heat” genes (28/199); Morphology - Phd1p targets (23/81), plus GO “cell-cell adhesion” (2/4) and “cytokinesis, completion of separation” genes (6/11). All categories except OXPHOS were significantly enriched among differentially expressed genes (according to Fisher's exact test with multiple hypothesis testing correction, or to Model Gene Set Analysis (MGSA); see Text S1).

Figure 7. Model of how a reduction in the hydrophobicity of subunit 9 permits its functional expression from nuclear DNA.

When the hydrophobicity of subunit 9 is too high, the protein cannot cross the inner mitochondrial membrane (IM) and is degraded in the intermembrane space by the i-AAA protease. With reduced hydrophobicity, subunit 9 can cross the IM and is processed by the matrix processing peptidase (MPP), properly inserted into the IM, and assembled into ATP synthase (see text for details). OM, outer mitochondrial membrane; MTS, mitochondrial targeting sequence, TMH, transmembrane segment; TOM, translocase of the OM; TIM, translocase of the IM.