A hunt for OM45 synthetic petite interactions in Saccharomyces cerevisiae reveals a role for Miro GTPase Gem1p in cristae structure maintenance.

Om45 is a major protein of the yeast's outer mitochondrial membrane under respiratory conditions. However, the cellular role of the protein has remained obscure. Previously, deletion mutant phenotypes have not been found, and clear amino acid sequence similarities that would allow inferring its functional role are not available. In this work, we describe synthetic petite mutants of GEM1 and UGO1 that depend on the presence of OM45 for respiratory growth, as well as the identification of several multicopy suppressors of the synthetic petite phenotypes. In the analysis of our mutants, we demonstrate that Om45p and Gem1p have a collaborative role in the maintenance of mitochondrial morphology, cristae structure, and mitochondrial DNA maintenance. A group of multicopy suppressors rescuing the synthetic lethal phenotypes of the mutants on non-fermentable carbon sources additionally supports this result. Our results imply that the synthetic petite phenotypes we observed are due to the disturbance of the inner mitochondrial membrane and point to this mitochondrial sub-compartment as the main target of action of Om45p, Ugo1p, and the yeast Miro GTPase Gem1p.

INTRODUCTION

The outer mitochondrial membrane (OMM) is the cytosol‐exposed sub‐compartment that serves as an interface between the organelle and the remainder of the cell. It is the first physical barrier of the organelle, restricting access to the mitochondrial intermembrane space to solutes smaller than 5 kDa (Mannella, 1992; Vander Heiden et al., 2000). Om45p in the yeast Saccharomyces cerevisiae has been previously recognized as a major OMM protein in respiring yeast cells (Ohlmeier et al., 2004; Yaffe et al., 1989). The change in the expression of Om45p upon a shift from growth on the fermentable carbon source glucose to non‐fermentable glycerol is more than a magnitude higher than the general increase of mitochondrial proteins following the diauxic shift (Ohlmeier et al., 2004). Hence, Om45p is expected to have a role in mitochondrial functions. Surprisingly, an om45 knockout yeast strain showed no detectable phenotype under numerous conditions investigated (Yaffe et al., 1989). A C‐terminally tagged GFP Om45p is commonly used as a mitochondrial marker in studies of mitochondrial morphology and mitophagy (Sesaki and Jensen, 2001; Kanki et al., 2009).

Om45p is a membrane‐anchored protein consisting of a short hydrophobic N‐terminal transmembrane sequence of 22 amino acids. The remaining bulk of the protein constitutes a water‐soluble domain. For a long time, the latter had been described as facing the cytoplasm (Burri et al., 2006; Riezman et al., 1983; Waizenegger et al., 2003; Yaffe et al., 1989). Lauffer et al. first reported evidence for an intermembrane space orientation of the Om45p C‐terminus, along with data indicating the association of Om45p with yeast mitochondrial OMM proteins Por1p and Om14p. This high molecular weight complex interacts physically with the components of the IMM: transporters, adenine nucleotides carrier, the prohibitin complex, respiratory chain complexes III and IV, as well as ATP synthase (Lauffer et al., 2012). Evidence presented by both (Song et al., 2014; Wenz et al., 2014) indicates that Om45p insertion into the OMM is also dependent on protein factors and not only on the lipid composition of the OMM, as had been previously suggested.

Here, we describe the results of a screen for mutations interacting negatively with a deletion allele of OM45, revealing a role for its gene product in preserving inner mitochondrial membrane morphology. Using a colony color‐based synthetic petite screening approach in S. cerevisiae, we isolated three mutants that display a respiratory growth defect in the absence of Om45p. We tracked the source of the mutant phenotypes to point mutations to two distinct genes encoding the evolutionarily conserved OMM proteins, Ugo1p and Gem1p. Deletions of the genes encoding these factors have been previously reported to lead to disturbed mitochondrial morphology (Sesaki and Jensen, 2001; Frederick et al., 2004; Coonrod et al., 2007; Abrams et al., 2015). Our data presented here provide evidence for a joint function of Om45p and Gem1p in mitochondrial cristae structure maintenance and dynamics, and a requirement for OM45 for mitochondrial DNA (mtDNA) maintenance upon inactivation of its genetically interacting partner GEM1.

EXPERIMENTAL PROCEDURES

Escherichia coli strains and growth media

The E. coli strain TOP10 was used for plasmid cloning and propagation (Invitrogen, Carlsbad, CA, USA). Escherichia coli BL21 (DE3) (Studier and Moffatt, 1986; Studier et al., 1990) was used for protein production for antibody raising. Escherichia coli was cultivated on LB media/agar plates (10 g tryptone, 5 g yeast extract, and 10 g NaCl in 1 L) or LB media/agar plates supplemented with ampicillin 100 µg/ml or chloramphenicol 100 µg/ml. Super optimal broth with catabolite repression (SOC) (2% tryptone, 0.5% yeast extract, 10 mM NaCl, 2.5 mM KCl, 20 mM glucose, 10 mM MgCl2) was used during E. coli chemical transformation and electroporation.

Plasmids used in this study

For the plasmids used in the study, see Table 1. Plasmid construction was done using standard techniques (Ausubel et al., 1989). All new constructs were sequence‐verified.

TABLE 1

Yeast strains used in this study

Plasmid	References	Yeast marker
pAG32	Goldstein and McCusker (1999)	HPHMX
pET 15b	Novagen	E. coli expression vector
pET 15bOM45	This study	E. coli expression vector
pTSV30A	generated by J. Pringle and M. Longtine, previously described Kastaniotis et al. (2004)	LEU2
pTSV30OM45	This study	LEU2
pVT100U mtGFP	Westermann et al. (2000)	URA3
pYM8	Knop et al. (1999)	KanMX
YCp22FZO1	This study	TRP1
YCp22UGO1	This study	TRP1
YCp22ugo1‐1(P189L)	This study	TRP1
YCp33ADH1	This study	URA3
YCp33ADH1prom	This study	URA3
YCp33ADHpromOM45 (own terminator)	This study	URA3
YCp33AIM2	This study	URA3
YCp33AIM2	This study	URA3
YCp33DNM1	This study	URA3
YCp33FZO1	This study	URA3
YCp33GEM1	This study	URA3
YCp33gem1‐1 (R103K))	This study	URA3
YCp33gem1‐2 (S324N)	This study	URA3
YCp33MSC1	This study	URA3
YCp33OM45	This study	URA3
YCp33OM451GFP	This study	URA3
YCp33PET54	This study	URA3
YCp33PET9	This study	URA3
YCp33PRP3	This study	URA3
YCp33RPL27b	This study	URA3
YCp33RTT109	This study	URA3
YCp33SHY1	This study	URA3
YCp33TIM44	This study	URA3
YCp33UGO1	This study	URA3
YCp33YGR111	This study	URA3
YCp33YOR228c	This study	URA3
YCplac111	Gietz & Sugino (1988)	LEU2
YCplac22	Gietz & Sugino (1988)	LEU2
YCplac33	Gietz & Sugino (1988)	URA3
YEp112FZO1	This study	TRP1
YEp112OM45	This study	TRP1
YEp112UGO1	This study	TRP1
YEp112ugo1‐1 (P189L)	This study	TRP1
YEp181OM45‐proA	This study	URA3
YEp195ADH1promOM45	This study	URA3
YEp195AIM2	This study	URA3
YEp195AIM2	This study	URA3
YEp195DNM1	This study	URA3
YEp195FZO1	This study	URA3
YEp195FZO1	This study	URA3
YEp195GEM1	This study	URA3
YEp195OM45	This study	URA3
YEp195OM45GFP	This study	URA3
YEp195OM45‐GFP	This study	URA3
YEp195PET54	This study	URA3
YEp195PRP3	This study	URA3
YEp195RTT109	This study	URA3
YEp195SHY1	This study	URA3
YEp195TIM44	This study	URA3
YEp352(CTA1)OM45	This study	URA3
YEplac112	Gietz & Sugino (1988)	TRP1
YEplac181	Gietz & Sugino (1988)	LEU2
YEplac195	Gietz & Sugino (1988)	URA3

Yeast strains and growth conditions

For yeast strains used in this work see Table 2. All the strains were derivatives of W1536 8B (Kastaniotis et al., 2004; Zhao et al., 1998). Yeast strains were grown in either rich YPD medium (1% yeast extract, 2% peptone and 2% D‐glucose), YPG (3% glycerol), synthetic complete (SC), or synthetic complete drop‐out media Kaiser mix (Formedium™, Hunstanton, Norfolk, UK or Sigma‐Aldrich, St. Louis, MO, USA) SCD (2% D‐glucose), SCG (3% glycerol) or SC media lacking one or more nutrients. Media were solidified with 2% Bacto agar [Biokar – (Sigma‐Aldrich)]. Glycerol sectoring medium (GSM) was SCG medium supplemented with 0.05% glucose. The glucose supplementation is necessary for the red color development of the W1536 8B strain on glycerol (Kastaniotis et al., 2004).

TABLE 2

Yeast strains used in this study

Name	Genotype	Reference of the source
W1536 5B/8B	MAT a (5B)/MATα (8B), ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1,ura3‐1	Kastaniotis et al. (2004)
W 1536 8B Δom45	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, yil136w::KanMX4	This study
W 1536 8B Δgem1	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, yal048c::HPHMX	This study
W 1536 8B Δgem1Δom45	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, yil136w::KanMX4, yal048c::HPHMX	This study
W1536 8B ∆om45 ∆ugo1	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, yil136w::KanMX4, Δydr470c::HPH MX4	This study
W1536 8B Δpsd2	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1;psd2::ClonNat	This study
W1536 8B Δpsd1Δpsd2	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1; psd2::ClonNat; ynl169c::HPH MX	This study
W 1536 8B Δgem1Δom45Δpsd2	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, yil136w::KanMX4, yal048c::HPHMX, psd2::ClonNat	This study
W 1536 8B Δom45 Δpsd2	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, yil136w::KanMX4, psd2::ClonNat	This study
W 1536 8B Δgem1 Δpsd2	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, yal048c::HPHMX, psd2::ClonNat	This study
W1536 5B rho0 (Δetr1)	MAT α ade2Δ, ade3Δ, can1‐100, his3‐1115, leu2‐3112, trp1‐1, ura3‐1, rho0, etr1Δ	Kursu et al. (2013)

Yeast transformation

Yeast one‐step transformation was carried out as described earlier (Chen et al., 1992) with minor modifications. Briefly, yeast cells from 300 µl of overnight culture were sedimented by a short spin in a table centrifuge (10 s, 11,000  ). The supernatant was removed and the cells resuspended in 100 µl of the transformation mix [0.17 M lithium acetate, 34% polyethylene glycol 3350 (PEG), 100 mM DTT, 0.3 µg/µl single‐stranded salmon or herring sperm DNA (ssDNA), 1 µl plasmid DNA (0.1–2 µg)]. The cells were incubated in the transformation mix for 30 min at 42°C, plated on the appropriate selective medium, and incubated at 30°C for 3–7 days.

Yeast high‐efficiency transformation was performed following the method of Gietz and Woods (Daniel Gietz et al., 2002). Yeast cells were reinoculated from an overnight culture to fresh YPD (or appropriate selective medium) to obtain 5 × 106 cells/ml and incubated for 3–5 h in a shaker while completing 2 doublings. Cells were harvested by centrifuging for 5 min × 3000  , washed once with sterile distilled H2O and once with 100 mM lithium acetate. To each 4 × 106 cells 360 µl of the transformation mix [240 µl PEG (50% w/v), 36 µl 1.0 M. lithium acetate 2 µl ssDNA (10 mg/ml), 0.1–10 µg plus H2O to a volume of 34 µl] were added. Afterward, cells were incubated for 50 min at 42°C. Transformed cells were incubated in YPD liquid medium for 2–4 h and plated onto the appropriate selective medium.

PCR‐mediated gene replacement

All yeast knockout strains were created by PCR‐targeting with short flanking homology (PCR‐mediated gene replacement). The homologous ends of the transferring DNA were introduced by primers with a short region (25 bases) homologous to the knockout cassette (KANMX or HPHMX) and a region (40–45 bases) homologous to the gene to be replaced (5′ or 3′ to the stop and start codons of the ORF of the gene) (Goldstein and McCusker, 1999; Knop et al., 1999). The resulting PCR product, therefore, consists of the KANMX (or HPHMX) dominant antibiotic resistance gene and the target gene homologous regions. This DNA product was introduced to the yeast cell by high‐efficiency transformation. KANMX (or HPHMX) integrates at the target locus, resulting in geneticin (or hygromycin) resistant transformants. The success of the knockout procedure was verified by PCR.

Droplet digital PCR (ddPCR)

Yeast genomic DNA (gDNA) from each sample was isolated using the method described previously (Hoffman and Winston, 1987). The concentration and purity of the DNA were determined using the NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA) and restriction digested with the Sac1‐HF (New England BioLabs Inc., Ipswich, MA, USA). A final concentration of 0.1 ng digested DNA was used for the Droplet Digital Polymerase Chain Reaction (ddPCR). Unless otherwise stated, all the instruments, software, consumables, primers, and probes related to the ddPCR experiments were purchased from Bio‐Rad Inc., Hercules, CA, USA. The mtDNA encoded cytochrome c oxidase subunit 3 (COX3) (SGD ID S000007283) gene, which encodes one of the three mtDNA subunits of complex IV of the mitochondrial electron transport chain, was used as a marker gene for the yeast mtDNA genome maintenance. A ddPCR reaction probing the nuclear‐encoded gene ACT1 (SGD ID S000001855) (encoding actin) was used as a standard. COX3 and actin gene‐specific primers were labeled with HEX and FAM probes respectively. The COX3 primers sequences were: forward primer 5´‐ CATTCAGCTATGAGTCCTG‐3′, reverse primer 5′‐CCTGCGATTAAGGCATG‐3′ and Probe sequence: 6HEX 5′‐AGGTGCATGTTGACCACCCGTAGG‐3′‐Iowa Black FQ. The ACT1 primer sequences were: forward primer 5′‐CAAACCGCTGCTCAAT‐3′, reverse primer 5′‐TACCTGGGAACATGGTG‐3′, and the probe 6FAM‐5′‐TGGTAACGAAAGATTCAGAGCCCCAGAAG‐3′‐Iowa Black FQ. A T100™ Thermal Cycler and the QX 200™ droplet reader were used and a count of the positive and negative droplets (events) utilizing the HEX/FAM channels was performed. The wells displaying ≥15,000 events were chosen for the analysis. QuantaSoft 1.7.4 software was used to analyze the copy number variation (CNV). The reference gene was ACT1 (one copy/haploid cell). CNV was expressed as copies/cell. Three biological replicates were performed in the analysis.

Mutagenesis and sectoring screen

Ethyl methanesulfonate (EMS) mutagenesis was performed as described previously (Amberg et al., 2005). Yeast cultures were grown to an early stationary phase in SC selective medium supplemented with 2% glucose. Two separate 1 ml samples (one as a control) were pelleted in a table centrifuge (10 s, 5000 g). The obtained pellets were washed once with distilled sterile water and resuspended in 0.1 M sodium phosphate buffer (pH 7.0). 30 µl of EMS were added to one of the two tubes and mixed by vortexing. Both tubes were incubated for 50 min at 30°C. Cells were pelleted, resuspended in 200 µl of 5% sodium thiosulfate, transferred to fresh tubes, and washed twice with 200 µl of 5% sodium thiosulfate, resuspended in 1 ml of sterile water, and plated. 1:10–1:10,000 dilutions from both the sample and the control (the cells incubated without EMS) were prepared for later calculation of the killing rate, which ranged between 50–80%. Petri dishes with cells plated on GSM medium were incubated at 30°C for 10–15 days. The sectoring screen was performed manually as described earlier (Kastaniotis et al., 2004; Kursu et al., 2013).

Cloning of the synthetic mutations

Upon loss of the OM45 containing plasmid, it was not possible to complement the synthetic petite mutants with any plasmid encoding the intact copies of OM45 or the mutated genes (our observation). For this reason, cloning by complementation had to be accomplished via two plasmid shuffles. First, the synthetic petite mutants carrying the pTSV30OM45 plasmid were transformed with multicopy YEp112OM45 (TRP1 selective marker). Sectoring colonies obtained after the transformation were streaked out on SCG‐TRP media containing 0.05% glucose. The white colonies collected from this restreaking were further tested on SCD‐LEU plates to confirm the loss of the pTSV30OM45 plasmid. The TRP1 marker allowed us to perform a counterselection against the OM45 plasmid using 5‐fluoroanthranilic acid (5‐FAA) in the following step. For cloning of the mutations by complementation two multicopy libraries were used. pRS426‐based (referred in the text as HeAl) (Kastaniotis et al., 2004) was made in the Heitman laboratory by Clara Alarcon and contains yeast genomic sequences inserted in the Sal1 site of the vector and Lacroute library constructed by F. Lacroute in pFL44L (2µ plasmid with a URA3 marker) (Bonneaud et al., 1991) with genomic DNA of strain FL100 (Harington et al., 1993).

Mutant strains carrying YEp112OM45 were grown overnight on SCD‐TRP media, transformed with the Lacroute or the HeAl library according to the high‐efficiency lithium acetate/SS‐DNA/PEG protocol (Gietz & Woods, 2002), and plated on SCG‐URA plates containing 250 mg/L 5‐FAA, low tryptophan (10 mg/L) and 0.05% glucose. Glucose was added in this case to enhance the recovery of library transformants. The used concentration is low enough to allow for efficient selection of the respiratory competent colonies.

Test for petite colonies generation rate

Strains from SCG plates grown overnight in liquid medium (YPD) were plated onto SCD to yield approximately 50–100 colonies per plate. After 7 days of incubation at 30°C, the colonies were replica plated onto YPD and SCG and incubated for 3–7 days at 30°C. The number of colonies on corresponding SCG (respiratory competent) and YPD (total number of colonies) plates were determined. The number of the petite colonies = number of colonies on YPD plate minus number of colonies on SCG plate.

Construction of His6‐Om45p, protein expression, purification, and antibody production

DNA encoding the soluble domain of Om45p without the first 22 amino acids (predicted N‐terminal transmembrane domain) was cloned onto the expression vector pET15b (Novagen‐Addgene, Cambridge, MA, USA) and expressed in E. coli BL21 (DE3). The expressed chimeric protein containing a His6 tag on its N‐terminus, a thrombin cleavage site, and N‐terminally truncated Om45p was affinity purified with Ni2+ chelating Sepharose (GE Healthcare, Piscataway, NJ, USA), concentrated 10 times with a Millipore Amicon Ultra‐15 10000 NMWL concentrator (Carrigtwohill, Cork, Ireland) in a table centrifuge at 4°C and subjected to anion‐exchange chromatography in Fractogel EMD DEAE (M) (Merck KGaA, Darmstadt, Germany), in a 10 × 150 mm column (20 mM Tris/HCl pH 7.4, linear gradient elution with NaCl from 50 to 200 mM at the flow rate 1 ml/min in 25 min). The purity of the peak fractions was controlled by SDS‐PAGE and mass spectrometry. Peak fractions were concentrated to 1.69 mg of protein per ml. 500 µl of the protein were used for the generation of polyclonal antisera fin two rabbits by Davids Biotechnologie (Regensburg, Germany).

Fluorescence microscopy

Fluorescence microscopy was performed with either live or fixed cells using a Zeiss LSM700 confocal fluorescence microscope at 100x magnification. Mitochondria were visualized as described previously (Westermann and Neupert, 2000) with pYX142mtGFP, pVT100UmtGFP, or using MitoTracker®Red CMX‐ROS (Molecular Probes, Eugene, OR, USA). DAPI (Thermo Scientific, Rockford, IL, USA) staining of mtDNA in methanol‐fixed cells was performed as described previously (Jones and Fangman, 1992).

Yeast cell fixation

For fixation, yeast cells were incubated in the growth media containing formaldehyde 4.4% (v/v) final concentration for 30 min at 30°C under shaking, then washed with 1.2 M sorbitol in 0.1 M potassium phosphate (KPi) buffer (pH 6.5), harvested at 700 g for 1 min and stored at 4°C (Lee et al., 1996).

Transmission electron microscopy

Transmission electron microscopy was performed according to the Tokuyasu method for yeast described previously (Griffith et al., 2008). Cells were fixed in 2% paraformaldehyde, 0.2% glutaraldehyde in 0.1 M PHEM buffer (60 mM PIPES, 25 mM HEPES, 10 mM EGTA, 2 mM MgCl2, pH 6.9) at room temperature for 3 h. Cells were washed with 0.1 M PHEM buffer, resuspended in 1% periodic acid in 0.1 M PHEM buffer, and incubated at room temperature for 1 h. After washing in 0.1 M PHEM buffer, cells were infiltrated with 12% gelatin dissolved in 0.1 M PHEM buffer at 37°C for 10 min. After solidification at 4°C, small gelatin blocks were immersed in 2.3 M sucrose and frozen in liquid nitrogen. Ultrathin (75 nm) sections were cut at −100°C using UC7 ultramicrotome with cryo attachment (Leica Microsystems) and picked up on grids with a drop of 1% methylcellulose, 1.1 M sucrose mixture in PHEM buffer. Sections on a grid were contrasted with 2% neutral uranyl acetate for 5 min, embedded in 2% methylcellulose, 0.4% uranyl acetate mixture, and examined using Tecnai G2 Spirit transmission electron microscope (FEI Europe, Eindhoven, Netherlands). Images were captured by a Quemesa CCD camera and analyzed using iTEM software (Olympus Soft Imaging Solutions, Münster, Germany). Yeast cell fixations and electron microscopy were performed by the staff of the Biocenter Oulu Electron microscopy core facility.

Lipid analysis

Preparation of the cells for lipid extraction

To prevent DNA loss during culture growth, yeast cell lawns were prepared on SCG plates supplemented with 2 mM ethanolamine (Etn) (Sigma‐Aldrich), by incubation at 30°C for 4 days. The Etn supplementation was necessary since the negative control W1536 8B Δpsd1Δpsd2 strain is an auxotroph for Etn. The cells were scraped with 2 × 2 ml SCG containing Etn (or SCD containing Etn) media, inoculated in 400 ml SCG containing Etn (or SCD containing Etn) to OD595 0.4, and grown for 43 h (or 15 h in case of SCD containing Etn).

For the phospholipidome analysis yeast cells were harvested from a 400 ml growth culture by centrifugation. The cell mass was adjusted to 1 g wet weight, washed with sterile water. Cells were pelleted by centrifugation (5 min, 3000  ), the pellets were frozen and stored at −70°C until lipid isolation.

Mitochondrial isolation for lipid extraction

Highly purified mitochondria were prepared as described before by Meisinger et al. (2000). Briefly, yeast cultures grown as described above were harvested by centrifugation (Beckman J‐6‐Mi) for 35 min at 5000  , washed with deionized water, and incubated for 20 min in DTT buffer [100 mM Tris/H2SO4 (pH 9.4), 10 mM DTT], 2 ml/g of wet cells, washed with zymolyase buffer [1.2 M sorbitol, 20 mM potassium phosphate buffer (pH 7.4)] and then digested with zymolyase 20T (5 mg/g of wet cells) in zymolyase buffer for 30 min. Spheroplasts were lysed in a Teflon potter (500 rpm) using 6.5 ml of homogenization buffer [(0.6 M sorbitol, 10 mM Tris/HCl, pH 7.4, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF), 0.2% (w/v)] per gram of wet cells. The homogenate was centrifuged at 1500  for 5 min and the supernatant was further cleared by centrifugation at 3000  for 5 min. The crude mitochondria from the supernatant were collected by centrifugation in the SS‐34 rotor at 12,000  for 15 min. The mitochondrial pellet was resuspended in cold SEM buffer (250 mM sucrose, 1 mM EDTA, 10 mM Mops, pH 7.2), manually homogenized in a small Teflon potter with 20 strokes, and centrifuged at 12,000  for 15 min. The pellets were resuspended in 5 ml of cold SEM buffer, applied to the sucrose gradient, and ultracentrifuged at 100,000  for 30 min at 4°C. Mitochondria were collected with a syringe needle from the interface between the 32 and 64% sucrose layers. Mitochondria were washed with SEM buffer, weighed, flash‐frozen in liquid nitrogen in 200 µl of the SEM buffer, and stored at −70°C.

The yeast samples for the lipid analysis were grown on synthetic complete media containing either glucose or glycerol, supplemented with 2 mM Etn.

Lipid extraction

Frozen cells (1 g) were thawed at room temperature and subjected to ethanol extraction (95% ethanol, 4 ml) in acid pre‐washed glass tubes with a Teflon cover (Pyrex England) at 22°C for 1 h under constant shaking (Vari‐Mix, Barnstead/Thermolyne, Ashville, NC, USA). The standard lipids (1,2‐diheptadecanoyl‐sn‐glycerol‐3‐phosphatidylethanolamine m/z (−) 718.5387; 1,2‐diheptadecanoyl‐sn‐glycerol‐3‐phosphatidyl‐L‐serine sodium salt m/z (−) 762.5292; 1,1,2,2‐tetramyristol‐cardiolipin (ammonium salt) m/z (−) 1239.841; L‐α‐lysophosphatidylcholine‐palmitoyl‐D3 (methyl‐D3) m/z (−) 557.36, each 0.2 µg/sample, were added to control the efficiency of the extraction. After 5 min spinning at 4100 g the supernatant was collected in a glass evaporation tube (vial no.1; Verex™. Phenomenex®, Torrance, CA, USA) and evaporated in a speed vac (ThermoElectron, SPD 2010, Ashville, NC, USA) at 45°C under vacuum. The pellets were further extracted twice for 20 min with 2 ml chloroform: methanol (2:1 v:v), centrifuged at 4000  for 5 min after each extraction, and finally extracted twice in acidic solvent (20 min with 2 ml of chloroform: methanol: HCl (124:65:1 v:v:v) followed by centrifugation at 4000  for 5 min after the acidity was adjusted to pH 4.0 with 0.5 M NaOH. Extracts from all steps were redissolved in 1 ml CHCl3: methanol (2:1 v:v), combined and washed with 1 ml 0.9% NaCl of pH 3.0 (Christie, 2003).

Experiments with the lipids obtained from whole cells were performed in the following replicates: three independent experimental repeats were performed for each yeast strain; one to three independent repeats were taken from each sample culture. Two to three independent mass spectrometric measurements were performed from each lipid extraction. Overall, 13 measurements per yeast strain were performed for the determination of the phospholipid content in whole cells and six for the analysis of the mitochondrial phospholipidome.

LC‐MS analysis

The separation of the phospholipids was performed on a UPLC® HSS T3 1.8 μm column (2.1 × 100 mm) at 55°C using an ultra‐performance LC system (Waters Acquity™, Milford, MA, USA). The binary gradient was made with A: 10 mM ammonium acetate in acetonitrile and deionized water (60:40); B: 10 mM ammonium acetate in isopropanol/acetonitrile (90:10). The column was eluted at 0.35 ml/min with a linear change from 60 to 100% B over 15 min.

Eluted lipids were analyzed by positive and negative electrospray ionization in separate runs on a Synapt G‐2 Quadrupole/time of flight mass spectrometer in MSE mode with lock mass correction over a mass range from 100 to 1500 Da.

Data were processed with Mass Lynx and its Marker Lynx option to obtain automatic quantification of large data sets. The ion counts were normalized to internal standards; lipids were identified by their accurate masses, MS/MS fragmentation, and their chromatographic behavior. The analysis was focused on the most abundant (C 34:2, C 34:1, C 32:2 and C 32:1) phosphatidylserine (PS), phosphatidylethanolamine (PE), phosphatidylcholine (PC) and cardiolipin (CL) (CL 64:4; CL 66:4; CL 68:4; CL 70:4 and CL 72:4) species.

Protein analysis

Protein extraction for western blot analysis

Proteins were extracted from the cultured yeast cells or directly from the plate lawns as described previously by Platta et al. (2004). In brief, cells were harvested by centrifugation, washed with sterile water. 30 mg of cell pellets and mixed with 15 µl of 1 M potassium phosphate buffer, pH 7.4 and 100 µL of trichloroacetic acid were incubated for 30 min at −70°C, thawed, and pelleted. Proteins were further precipitated in 500 µl of ice‐cold 80% acetone, pelleted, redissolved in 60 µl of 1% SDS/0.1 M NaOH and 20 µl of 4 × SDS loading buffer [500 mM Tris/Cl (pH 6.8), 10% SDS, 30% glycerol, bromphenol blue, 10% β‐mercaptoethanol], heated 5 min at 95°C.

SDS‐PAGE and western blot analysis

Proteins were analyzed by SDS‐PAGE (12%) (Laemmli, 1970), stained with Coomassie brilliant blue, or subjected to western blotting using nitrocellulose membranes (Bio‐Rad) and the enhanced chemiluminescence system (Bio‐Rad).

RESULTS

Synthetic petite screen

It has been previously reported that Om45p is not essential for yeast growth or mitochondrial functions (Yaffe et al., 1989). Our W1536 8B om45 deletion strain, a derivative of W303, grew equally well as the corresponding wild‐type strain under the growth conditions investigated (Figure A1). We hypothesized that the absence of any growth phenotype for the major protein of the OMM may be due to the presence of (a) protein(s) with redundant functions. This possibility was addressed in a screen for synthetic petite mutants (SPM) (Kursu et al., 2013) (Figure 1a). W1536 5B/8B strains carrying an ade2, ade3 double deletion, as well as leu2, ura3, and trp1 marker mutations are deleted for om45 (Δom45). When transformed with the pTSV30OM45 plasmid carrying OM45 and ADE3 genes (as well as the LEU2 gene as a transformation selection marker), the cells become genotypically Δade2 and develop the characteristic red colony color phenotype. Without selection for the plasmid, it is lost at high frequency, resulting in a “sectored” colony color appearance. The cells are exposed to the mutagen ethyl methanesulfonate (EMS) and plated on media containing only a non‐fermentable carbon source. Synthetic petite mutants are identified by their inability to lose the OM45 plasmid on non‐fermentable media, which can be monitored by the formation of completely red, non‐sectoring colonies on synthetic complete glycerol (SCG) media. The ability to lose the OM45 plasmid on synthetic complete dextrose (glucose) (SCD) media indicates that the mutation only affects mitochondrial function, but not respiratory growth. The screen was designed to identify mutations that in absence of OM45 would lead to a synthetic growth defect specifically on non‐fermentable media containing glycerol, but not on fermentable media containing glucose. We screened over 50,000 colonies and obtained three stable synthetic petite mutants in a Δom45 strain background that were dependent on an OM45 plasmid for growth on SCG after chemical mutagenesis. These mutants readily lost the OM45 plasmid on SCD, upon which they became respiratory deficient, and were viable on fermentable carbon sources in absence of OM45 (Figure 1b and c). Upon loss of the OM45‐containing plasmid, it was not possible to complement the synthetic petite mutants (SPMs) with any plasmid encoding the intact copies of OM45. Hence, all the complementation experiments and cloning of the mutations by library complementation were done via plasmid shuffle. The obtained synthetic petite mutants were recessive, as diploids of the Δom45 parentals with the SPMs are respiratory competent in absence of the OM45 plasmid.

FIGURE A1

Growth phenotype of W1536 8B wild type and om45 deletion strain on different growth media. Yeast strains were grown overnight in SC liquid medium supplemented with glucose. Cultures were normalized to OD595 = 1.0 and serial 10‐fold dilutions were spotted onto corresponding solid media. Photographs were taken after 3−6 days of growth

FIGURE 1

Synthetic petite screen in yeast hunting for genetic interactions of OM45. (a) Schematic outline of the screen. See text for further description. SPM synthetic petite mutant. The X symbolizes the mutation in an interacting gene. (b) and (c) Test for a dependence of SPMs on the OM45 plasmid for respiratory growth and demonstration of respiratory deficiency in absence of OM45. (b) The synthetically petite mutant gem1(R103K) and (c) ugo1(P189L) obtained via the genetic screens are respiratory deficient upon loss of the OM45 plasmid. All the complementation studies were done via plasmid shuffle, loss of the original pTSV30OM45 plasmid was monitored via colony sectoring and confirmed by re‐streaking on SC‐LEU drop‐out media supplemented with glucose. Only strains that carry a copy of OM45 on a plasmid are viable on respiratory media (SCG). All strains can grow on media containing the fermentable carbon source glucose (SCD, lowest panel). (d) Schematic representation of the location of the obtained synthetic mutations in the Gem1p and Ugo1p translation products based on published data on Ugo1p (Hoppins et al., 2009) and Gem1p (Frederick et al., 2004) Black boxes designate transmembrane domains (TMDs). MIRO, GTPase domains; EF1/2, EF‐hand domains. (e) Nucleotide and amino acid sequences of obtained synthetic petite mutations. The nucleotide sequence excerpts displaying the gem1(R103K) and gem1(S324N) mutations show nucleotides 301–324 and 961–984 of the GEM1 ORF, respectively. The UGO1 ORF sequence excerpt stretches from 553 to 576. Mutated amino acid residues and corresponding wild‐type residues are marked in red

Synthetic mutations in the GEM1 and UGO1 genes

The gene alleles with acquired point mutations in the three synthetic petite mutant yeast lines were identified by cloning based on genomic DNA library complementation. For this purpose, two yeast multicopy genomic DNA libraries: HeAl (Kastaniotis et al., 2004) and Lacroute (Bonneaud et al., 1991; Harington et al., 1993) were used for transformation of the synthetic petite mutants carrying the YEp112OM45 plasmid. The latter construct carries a TRP1 marker gene that can be selected against on 5′ fluoroanthranilic acid (5′FAA)‐containing media. From about 50,000 independent transformants able to grow on selective media supplemented with glycerol upon loss of the YEp112OM45 plasmid, several independent clones containing OM45 were isolated during our library complementation screen, confirming the good quality of the libraries and the validity of the cloning strategy. We tracked the cause of the mutant phenotype of the synthetic petite strains to point mutations in two distinct genes encoding OMM proteins Ugo1p, a participant in the mitochondrial fusion machinery (Sesaki and Jensen, 2001), and Gem1p, the mitochondrial Rho (Miro) GTPase (Frederick et al., 2004) (Figure 1d and e). Single copy plasmids carrying wild‐type alleles of these plasmids complemented, as well as the original multicopy (2µ) library plasmids. Sequencing of the mutant alleles revealed a transition mutation in ugo1(c.566C>T), translating to p.P189L in the mutant Ugo1 protein. This mutation is located in a segment of the Ugo1p amino acid sequence that is exposed to the intermembrane space (IMS) (Hoppins et al., 2009). We mapped the gem1‐1 (c.308G>A/ p.R103K) mutation to the first GTPase domain of Gem1p and gem1‐2 (c.971G>A/p.S324N) to the region between the two EF‐hand domains of the protein (Frederick et al., 2004) (Figure 1d and e). The OM45 genetic interactions with both GEM1 and UGO1 found in our screen are consistent with high throughput yeast synthetic lethality data reported before (Hoppins et al., 2011). The Gem1 and Ugo1 proteins have previously been shown to play a role in yeast mitochondrial morphology maintenance (Frederick et al., 2004; Sesaki and Jensen, 2004).

Mitochondrial morphology defects in synthetic petite mutants devoid of OM45 plasmid

In addition to the growth defects, SPM demonstrated changes in mitochondrial structure. We examined mitochondrial morphology in SPMs in the presence and absence of an OM45 plasmid (Figure 2). Mitochondria of the ugo1(P189L) SPM strain grown on glucose media displayed many fragmented, collapsed mitochondria close to each other, nonetheless apparently unable to fuse and form the long tubular structures that can be found in the WT control. The presence of the plasmid‐borne copy of OM45 slightly ameliorated the phenotype and some tubular‐like mitochondria were present. SPMs of gem1(R103K) and gem1(S324N) (not shown) also contained many fragmented and collapsed mitochondria. Some tubular mitochondrial structures were present in these mutant cells, although the diameter of such tubules was irregular, they appeared collapsed and reminiscent of a “beads on a string” arrangement. The appearance of mitochondria in the SPMs clearly differs from the morphology observed in rho cells, which maintain a tubular structure (Sesaki & Jensen, 2001). Thus, observed changes in SPM mitochondrial morphology cannot be explained as a consequence of mtDNA loss. The presence of plasmid‐borne OM45 improved the SPM phenotype, resulting in mitochondria with a more tubular structure and more regular in diameter (Figures 2, 3).

FIGURE 2

Synthetic lethal mutants show mitochondrial morphology defects in the absence of the complementing OM45 plasmid. Yeast cells from wild‐type (WT), SPMs, and SPMs with the complementing plasmid were grown in selective liquid synthetic complete medium supplemented with glucose (2%) to mid‐log phase, and mitochondria (Mito) were visualized by mitochondria‐targeted GFP. The om45 deletion does not visibly affect mitochondrial morphology (second row from the top). The gem1 mutants still maintain some tubular structure but display mitochondrial fragmentation and collapsed structure (rows three and four from the top). This phenotype is exacerbated in the ugo1 SPM, which nearly completely lacks long tubular mitochondria. Mitochondrial morphology of the SPM cells is rescued by the presence of a plasmid‐borne copy of OM45, here shown for the Δom45,ugo1(P198L) SPM. BR, brightfield image; Mito, mitochondrially targeted GFP. Projection images of mitochondria were generated using the Zeiss software (see Experimental procedures). Size bar: 5 µm

FIGURE 3

Quantitative analysis of mitochondrial morphology types in SPM cells in the absence or presence of complementing OM45. (a) Representative images of morphology types observed in W1536 8B Δgem1Δom45 cells with and without the complementing 2µ OM45 plasmid. Mitochondria were visualized using matrix targeted mtGFP expressed from the pVT100 U mtGFP (URA3) plasmid. Cells were grown overnight in a selective liquid synthetic complete medium supplemented with glucose (2%). (b) Quantification of mitochondrial morphology phenotypes as exemplified in panel A. Shown are % ratios of cells from 2 experimental replicates observing at least 100 cells in each experiment, except for data of mutant Δom45, gem1(R103K) for which results of one set of 100 examined cells are shown

Electron microscopy analysis of the SPMs with and without the plasmid copy of OM45 revealed defects in the folding of the inner mitochondrial membrane (IMM) (Figure 4). The double membranes of mitochondria were clearly observed, but the IMM was completely devoid of regular cristae organization (compare the figure to wild‐type examples in Figure A2, A3, A4, A5, A6, A7, A8), was found either adjacent to the OMM or forming a stretch along the mitochondria approaching the OMM at both ends or displaying small circles and exhibited a multi‐layered appearance. In contrast, the cristae organization of W303 strain background cells carrying the Δgem1 mutation alone has been reported to be unaffected (Frederick et al., 2004) and the IMM of W1536 8B Δom45 strain is indistinguishable from wild type under regular growth conditions. The membrane morphology of the SPMs also differs from the membrane defects observed in rho cells, which lack properly formed cristae but do neither exhibit the multi‐layered membrane phenotype nor the dramatic clustering that can be observed in SPM mitochondria (Figures A2, A3, A4, A5, A6, A7, A8). These results indicated a synthetic negative effect of om45 and gem1 mutations on IMM structure. The presence of a multicopy OM45‐carrying plasmid visibly relieved the severity of the phenotype, largely restoring the classical cristae organization of most mitochondria.

FIGURE 4

Inner mitochondrial membrane folding is distorted in synthetic petite mutants upon loss of the OM45 plasmid. Electron micrographs of the synthetic petite mutants (a, c, e) with pTSV30OM45 and (b, d, f) without pTSV30OM45. The gem1 synthetic petites (b and d) show an improvement of cristae structure in the presence of the OM45 carrying plasmid (a and c). The ugo1 mutant, which has a highly disturbed mitochondrial membrane structure (f), also shows mitochondria with a more regular shape in the presence of OM45 (e). Arrows point at mitochondria. Cells were grown o/n in liquid synthetic complete media supplemented with glucose (2%). Arrows indicate mitochondria. Scale bar 200 nm

FIGURE A2

Electron micrographs of W1536 5B‐derived WT and rho cells. Panels a–d: W1536 5B cells. Panels c and d are enlarged sections of panels a and b, respectively, as indicated by the boxes in a and b. Panels e–h: W1536 5B rho (Δetr1) cells. Panels g and h are enlarged sections of panels e and f, respectively, as indicated by the boxes in g and h

FIGURE A3

Tetrad dissection of SPMs. Cells were initially dissected on YPD or SCD media and then replica plated on the indicated selective media (SCG and SCLac or only SCG). The Δom45, ugo1(P198L) examples show the dissection master plate before the transfer; the other examples, the SCD master plate AFTER replica plating/transfer. Ratios (respiratory competent: respiratory deficient): a 0:4, b 4:0, c 1:3, d 3:1, e 2:2

FIGURE A4

Growth of the reconstructed of Δom45/ugo1 SPM. Strains were constructed similarly to the reconstructed gem1 SPMs (BY series background), initially generating the Δom45, Δugo1 double deletion mutant. However, a plasmid harboring WT UGO1 was maintained throughout to prevent mitochondrial DNA loss. The reconstructed SPM has a mild growth defect in this strain background, indicating a partial function of the ugo1(P189L) – encoded protein. The presence of UGO1 in the Δom45, Δugo1mutant completely rescued growth, and the presence of both plasmids carrying the ugo1(P189L) mutant allele and OM45 improved growth compared to the ugo1(P189L) construct being present alone, albeit not completely to WT levels on glycerol. When the Δom45, Δugo1 mutant strain only carried empty plasmids, the cells were completely respiratory deficient

FIGURE A5

Mitochondrial morphology changes and mtDNA loss in the SPM. Representative images of morphology types observed in W1536 8B WT, W1536 8B Δom45, and SPM cells with and without the complementing 2µ OM45 plasmid. Mitochondria visualization was done with matrix targeted mtGFP expressed from plasmid pVT100 U mtGFP(URA3). Cells were grown overnight in a selective liquid synthetic complete medium supplemented with glucose (2%). mtDNA was visualized with a DAPI stain. The concentration of mtDNA was determined with ddPCR in the mutants with and without 2µ OM45 plasmid grown overnight on SC or SC selective medium supplemented with 2% glucose. (a) Original images. (b) same images as in a, but the original blue color of the DAPI stain was replaced with a light turquoise color to enhance the visibility of the DNA stain. (Corel Photo‐Paint 2020, 64‐bit version, “Replace colors” function in the “Adjust” menu)

FIGURE A6

Growth assay of the W1536 8B Δom45, ugo1(P189L) mutant transformed with multicopy suppressor plasmids

FIGURE A7

Growth assay of the W1536 8BΔom45, gem1(R103K) mutant transformed with multicopy suppressor plasmids

FIGURE A8

Whole cell and mitochondrial phospholipidome analysis of W1536 8B Δpsd2, W1536 8B Δpsd2Δom45, W1536 8B Δpsd2Δgem1, W1536 8BΔpsd2Δom45Δgem1 and W1536 8B Δpsd2Δpsd1 strains. (a) Cellular content of indicated phospholipids from total cells phospholipid preparations in the indicated yeast strains. (b) The ratio of phosphatidylserine over phosphatidylethanolamine from total cell phospholipids. (c) Mitochondrial content of total cells phospholipid preparations in the indicated yeast strains. (d) The ratio of phosphatidylserine over phosphatidylethanolamine from mitochondrial phospholipids preparations. CL, cardiolipin; PE, phosphatidylethanolamine; PS, phosphatidylserine; PC, phosphatidylcholine. Phospholipids were extracted either from whole cells grown on SCG or from mitochondria purified from cells grown on SCG. The pictures present % ratios with the total sample phospholipids set as100%. The statistically supported changes as indicated with * between the samples and W 1536 8B Δpsd2 control (p < 0.05), **(p < 0.01) were determined with a t‐test

Reconstruction of the mutants

To better characterize the SPMs, we performed a tetrad analysis of Δom45/ Δom45‐SMP diploids. The dissection pattern of the mutants during the tetrad analysis (Figure A3) did not follow Mendelian segregation. If only one mutation was involved we would expect a clear 2:2 (viable:non‐viable) segregation pattern on glycerol. In case two mutations were required to get the respiratory deficient phenotype in the Δom45 background, we would expect: 2:2 for the parental ditype, 4:0 for the non‐parental ditype, and 3:1 for the tetratype. There is no scenario following Mendelian genetics that would ever result in a 0:4 (respiratory competent:respiratory deficient) ratio, which we frequently encountered in the tetrad dissection analysis of the gem1 SPMs. Tetrad analysis of the ugo1 SPM yielded many 4:0 (respiratory competent:respiratory deficient) and a few 3:1 tetrads, not consistent with a simple relationship between the Δom45 genetic background and the ugo1 SPM.

In our interpretation, the erratic inheritance pattern of the gem1 SPMs was caused by disturbed mitochondrial DNA inheritance. Because this feature of our mutants interfered with our attempts to breed out possible background mutations and to investigate or completely rule out the possibility of a complex combination of several nuclear mutations as the root of the non‐Mendelian segregation pattern, we reconstructed the mutant genotype in cells that had not been exposed to EMS mutagenesis.

The W1536 8B Δgem1Δom45 double deletion strain was created by introducing the HPHMX KO cassette (conferring hygromycin resistance) into the GEM1 locus the W1536 8B Δom45 strain, thereby removing GEM1. None of the single deletion strains (Δgem1 or Δom45) are respiratory deficient (Frederick et al., 2004; Yaffe et al., 1989). The obtained Δgem1Δom45 double deletion strain, in contrast to the gem1 point mutation SPMs, was initially not respiratory deficient but a very slow grower on glycerol‐containing media. The Δom45gem1(R103K) and Δom45gem1(S324N) point mutants were reconstructed by the introduction of the plasmid‐borne gem1(R103K) and gem1(S324N) alleles into the W1536 8B Δgem1Δom45 double deletion strain. The behavior of these three reconstructed strains under different growth conditions was tested (Figure 5a). A plasmid harboring wild‐type GEM1 (YCp33GEM1) completely rescued the growth of the W1536 8B Δgem1Δom45 strain, whereas YCp33gem1(R103K) and YCp33gem1(S324N) rescued the growth only partially, (Figure 5a).

FIGURE 5

Growth phenotypes of reconstructed SPMs, mitochondrial morphology changes, and mtDNA loss of the W1536 8B ∆om45∆gem1 strain. (a) The ∆om45∆gem1 strain exhibits a synthetic petite phenotype, already apparent from the slow growth on SCD media. The SPM‐derived gem1 mutant alleles only partially rescue growth in comparison to WT GEM1. Yeast strains were grown overnight in appropriate SC or SC selective liquid medium supplemented with glycerol. Cultures were normalized to OD595 = 1.0 and serial 10‐fold dilutions were spotted onto SC solid media supplemented with glucose (2%) or glycerol (3%). Photographs were taken after 3 and 5 days of growth at 30°C for yeast growing in glucose and glycerol media, respectively. (b) Electron micrograph of W1536 8B ∆om45∆gem1 with similarly disturbed cristae structure as the SPMs in Figure 4. (c) The same strain carrying the 2µ OM45 plasmid, with improved cristae structure. Cells were grown for 4 days in liquid synthetic complete media supplemented with glycerol (3%). EM Scale bar: 500 nm. (d) Electron micrograph of WT strain W1536 8B mitochondria. Scale bar indicates 500 nm. (e) Scheme for testing the rate of petite formation. After prolonged growth on fermentable media, the Δgem1 strain exhibits a mild increase of the respiratory deficient fraction of cells, while the ∆om45∆gem1 double mutant massively accumulates respiratory deficient cells. Strains were initially grown on SCG media. They were then inoculated in liquid SC media supplemented with glucose (2%), grown overnight, and plated on solid rich YP media supplemented with glucose (2%). After 7 days of incubation at 30°C colonies were counted, and replicas plated on solid SC media supplemented with glycerol (3%). The number of the petite colonies = number of colonies on YPD plate minus number of colonies on SCG plate (N = 3). (f) A statistical representation of experiment described in (e). (g) ddPCR analysis of mtDNA content in cells grown on glucose or glycerol. All the yeast strains were transferred from an SCG plate on a YPD (2% glucose) plate (mother plate). Individual colonies from the mother plate were re‐streaked on YPD plates several times and checked on the YPG plate for respiration competence. For genomic DNA isolation, all the strains were grown on YPD (2% glucose) liquid media overnight. Negative control: BY4741Δmtf2 (rho0 strain).

Since the ugo1 null mutant is respiratory deficient as a result of mtDNA loss after the loss of UGO1 (Sesaki and Jensen, 2001), reconstruction of the strain was achieved via plasmid shuffle. The UGO1 gene was knocked out in the presence of a plasmid carrying a wild‐type UGO1 allele and the ugo1 (P189L) allele was then introduced with a second plasmid, selecting against the initial construct carrying the wild‐type copy. On the non‐fermentable media SCG the growth of the Δom45, Δugo1/ugo1(P189L) reconstructed synthetic petite mutant was strongly reduced. The pTSV30OM45 construct as well as the plasmid harboring an intact copy of UGO1 rescued the respiratory growth of the reconstructed mutant (Figure A4).

The reconstruction of the mutants confirmed the negative synthetic interaction of the om45 deletion with the ugo1 and gem1 mutations. As Ugo1p is poorly defined and the synthetic petite interaction was more robust for the reconstructed ∆om45Δgem1 mutant strains, we decided to focus on the gem1/om45 interaction for our further analyses.

As analysis of the initial SPMs indicated severe mitochondrial ultrastructure morphology changes, we investigated the Δgem1Δom45 double deletion mutant using electron microscopy. Because our previous analyses of the SPMs suggested increased mtDNA loss of the mutants upon non‐selective growth on fermentable carbon sources, all cells used for the ultrastructure analysis were maintained and grow on media containing glycerol as the sole carbon source. The Δgem1Δom45 double mutant cells are viable but extremely slow growers on this media (see also below for mtDNA content). Electron microscopy analysis of the ∆om45∆gem1 double mutant revealed mitochondrial membrane defects (Figure 5b) similar to the phenotype we observed in the SPMs (Figure 4b,d). The cristae structure of the IMM was severely disturbed or completely absent. However, when these cells were transformed with a plasmid carrying OM45, cristae were restored (Figure 5c) and the mitochondria in the transformed strain were much closer to the appearance of WT organelles (Figure 5d), again mimicking the behavior of the SPM strains (Figure 4a,c). Overall, the reconstructed mutants, as well as the ∆om45∆gem1 double deletion mutant, very closely resembled the original SPMs in their phenotypes.

The ∆om45∆gem1 strain generates petites at a high frequency when grown on fermentable media

While W1536 8B ∆gem1∆om45 cells can be maintained indefinitely growing, albeit slowly, on glycerol media, a large fraction of the W1536 8B ∆gem1∆om45 cells similarly to original SPM became completely respiratory deficient after just a brief growth phase on glucose media and could not be rescued anymore with plasmids carrying OM45 or the GEM1. Similarly, we observed that the original synthetic petite mutants, once they had lost the initial OM45 plasmid upon maintenance on glucose‐containing medium, could not be rescued again by the retransformation with a 2µ OM45 plasmid.

We compared the frequency of generation of petites between W1536 8B ∆om45, W1536 8B ∆gem1, and W1536 8B ∆om45∆gem1 strains, starting from cells growing on glycerol. After transfer to liquid YPD with a fermentable carbon source and overnight growth, cells were plated on a YPD plate, grown 4–5 days until proper colonies were formed, and subsequently replica plated onto non‐fermentable SCG media (Figure 5e). W1536 8B ∆om45 formed only very few respiratory deficient petite colonies. The W1536 8B Δgem1 mutant, which had previously been reported as losing mtDNA at a slightly elevated rate (Frederick et al., 2004), generated up to 15% of respiratory deficient colonies. The Δgem1Δom45 petite generation rate under the same conditions approached 85%. Therefore, the percentage of generation of respiratory deficient progeny was strongly increased in the ∆om45∆gem1 double mutant compared to the individual deletion strains (Figure 5f).

We hypothesized that this high‐frequency generation of petites is caused by mtDNA loss, which would be consistent with our inability to complement the SMPs with plasmids carrying wild‐type copies of OM45 (or GEM1/UGO1) once the OM45 plasmid was lost.

Synthetic petite mutants and mtDNA maintenance

Respiratory competence of the SPMs upon loss of the OM45 plasmid could not be restored upon retransformation. To investigate DNA copy number in wild type, Δom45, and Δgem1 mutants as well as the Δom45Δgem1 double mutant, we performed droplet digital PCR (ddPCR) probing for the mtDNA encoded COX3 (cytochrome c oxidase subunit 3) gene and nuclear‐encoded ACT1 as reference. We found that the Δom45 mutant harbored mtDNA to wild‐type levels, while the Δgem1 strain exhibited a slight reduction of mtDNA when grown on non‐fermentable glycerol. Intriguingly, mtDNA content of the Δom45Δgem1 double deletion mutant was equivalent to, or perhaps even slightly increased when grown on glycerol (Figure 5g), as judged by the COX3 copy number signal. This is in stark contrast to the situation after the Δom45Δgem1 strain was subjected to overnight growth on fermentable YPD (Figure 5g) and mtDNA content was massively reduced (Figure 5g). The irreversible respiratory defect can be caused by either distortion of mitochondrial genome integrity (rho −) or by its complete loss (rho). A DAPI stain of SPM mutant cells showed a severe reduction of fluorescent spots co‐localizing with mitochondrial matrix targeted GFP protein in absence of an OM45‐carrying plasmid in the original Δom45gem1 SPMs (but not in the Δom45ugo1 SPM). We were able to confirm mtDNA phenotypes in the SPMs by droplet digital PCR quantitation using glucose‐grown cells (Figure A5a and b). Interestingly, the gem1 SPMs maintain some mtDNA, unlike the Δom45Δgem1 mutant that completely loses mtDNA after growth on glucose. These results are more consistent with a reduction of mtDNA copy number in the SPMs, quickly leading to complete loss of mtDNA (rho cells) than with deletions within the mitochondrial DNA (rho − mutations) as cause for secondary respiratory deficiency. They also clearly indicate that the morphology defect in the Δom45ugo1(P189L) SPM cannot be caused simply by mtDNA loss.

Suppression of SPM phenotypes

In addition to UGO1, GEM1, and OM45, several other genes from the two rescuing library plasmids were identified and found to be suppressors of the mutations in gem1 and ugo1 genes (Table A1, Figures A6 and A7). Among these suppressors are genes encoding proteins of both the IMM (TIM44, SHY1, RTT109, and PET9) and the OMM (FZO1, MCP1, and PET54). The MCP1, PET54, and YPL109c partially rescued the growth of all three mutants, also from a single copy plasmid, but no mutation was detected in the genomic copies of the genes. Respiratory growth of Δom45 ugo1 (P189L) SPMs was restored by multicopy plasmids carrying TIM44, SHY1, DNM1, and FZO1.

TABLE A1

Library plasmids rescuing the respiratory deficient phenotype of the SPMs

	Chromosome			Genes on the plasmid	Growth rescue
	Number	From	To		Δom45, gem1 (S324N)	Δom45, gem1 (R103K)	Δom45, ugo1 (P189L)
Isolated from synthetic petite mutant Δom45, gem1 (S324N)
L	X	204,547	209,,917	MDV1,CCT7	3/3 weak
L	III	175,407 175,411	178,862 178,866	RPS14A,SNR65SNR189
L	X	39,879	41,655	NUC1	1/3 weak	1/3 weak
H	VII	937,532	942,532	PET54	3/3 weak	0/3
	VII	937,396	940,318	PET54	3/3 weak	0/3
L	IV	1,320,450	1,325,454	RPN9,YDR426C,BNA7,TIF35	3/3 weak	3/4 weak
L	Not sequenced				3/3 weak	3/3 weak
L	I	50,217 52,050	55,793 55,793	AIM2, GEM1	3/3	3/3	3/3
Isolated from synthetic petite mutant Δom45, ugo1 (P189L)
	II	575,000	595,000	FZO1	0/3	0/3	3/3
	IV	1,399,442	1,407,105	UGO1	0/3	0/3	3/3
L	II	162,308	167,700	PET9	3/3	3/3	3/3
	IX	79,000	99,000	OM45
	VII	715,554	719,176	YGR111, SHY1	3/3	3/3	3/3
	XIV	386,997	392,422	EXBP6,SPC98
L	IX	309,455	313,798	TIM44 , RPB3	3/3 weak	0/3	no
	IX	82,000	102,000	OM45
L		145,655	150,076	RTT109 , DNM1	NO	3/3 weak	(0/6)
Isolated from synthetic petite mutant Δom45, gem1 (R103K)
L	VII	932,885	940,245	PET54 others
L	VII	937,760	946,051	PET54 others	3/3	3/3 weak
L	IX	93,574	97,460	OM45	3/3	3/3
L	Not sequenced				3/3 weak	4/4 weak
L	XV	102,337	102,577	Truncation of PAP2	3/3 weak	3/3 weak
L					3/3 weak	0/3
L	Not sequenced				3/3 weak	3/3 weak
H	V	116,587	117,048	URA3	3/3 weak
H	Not sequenced				3/3 weak	3/3 weak
H	XV	766,540	767,312	WTM2, YOR228C	3/3	2/3
H	Not sequenced				3/3 weak	1/3 weak
H	XVI	830,340	834,690	YPR150W,SUE1,URN1,YPR153W	3/3 weak	0/3
L	V	91,739	97,260	(ECM10) SSC3	3/3 weak	7/7 weak
L	I	52,050	56,140	GEM1	3/3	3/3
H	IX	94,170	94,363	OM45	2/2	3/3
L	VI	52,051	56,163	GEM1	3/3	3/3
H	XVI	341,371	341,810	Truncation of GDE1
L	XVI	341,370	348,068	YPL109C , GDE1	3/3	3/3
L	V	116,676	117,010	URA3	3/3 weak	0/3
L	IX	307,795	313,463	TIM44	3/3	3/3	3/3
L	XI	384,109	390,299	YKL027w, TFA1	0/3	0/3
L	Not sequenced				3/3 weak	3/3 weak

For origin of the plasmid: L‐Lacroute library, H‐HeAl library. The plasmids are presented in the order of mutants found complemented, although eventually suppression was tested in all three mutants. Chromosome number and location are indicated. Since multiple ORFs were present on most of the library plasmids, the underlined genes were confirmed as the rescuing component after subcloning to separate multiple and single copy vectors. Rescue x/y numbers indicate the number of growing clones (x) out of tested (y). “weak” indicate very weak growth of the rescued colonies. Empty fields: Not tested.

The functionality of Om45p fusion constructs

Om45p‐GFP constructs have been used for mitochondrial morphology as well as autophagy studies due to Om45p abundance in the OMM during respiration and a previously reported lack of a phenotype for the W1536 8B Δom45 strain (Sesaki and Jensen, 2001; Kanki et al., 2009). The reconstructed synthetic petite mutant W1536 8B Δom45Δgem1 strain serves as a unique working tool for testing the functionality of Om45p and its fusion constructs.

First, we performed the complementation analysis of our mutants with an Om45p‐GFP construct. The GFP‐tagged OM45 single and multicopy plasmids (YCp33OM45‐GFP and YEp195OM45‐GFP) rescued the respiratory growth of the W1536 8B Δom45Δgem1 mutants much weaker than the plasmids carrying native OM45 (Figure 6a). This effect could be observed already in the sole W1536 8B ∆om45 mutant, although much less pronounced (Figure 6b). Western blot analysis indicated that the levels of expression of Om45p and Om45‐GFP were similar (Figure 6c).

FIGURE 6

The complementation potency of Om45 is diminished by C‐terminal GFP tagging. Spotting assay with the (a) W1536 8B ∆om45∆gem1 and (b) W1536 8B ∆om45 strain expressing Om45 or Om45‐GFP from CEN (URA3) or 2µ (URA3) (multicopy) plasmids. Yeast strains carrying an indicated plasmid were grown overnight in a synthetic complete medium lacking uracil and supplemented with glycerol (3%). Cultures were normalized to OD595 = 1.0 and serial 10‐fold dilutions were spotted onto synthetic complete solid media lacking uracil and supplemented with either glucose (2%) or glycerol (3%). Photographs were taken after 3 and 5 days of growth at 30°C for yeast growing in glucose and glycerol media, respectively. (c) Representative western blot image. Om45‐GFP levels expressed in the ∆om45 strain are comparable to Om45p levels expressed from the corresponding vector. Cells were grown at 30°C. Western blot probed with anti‐OM45p rabbit serum and anti‐rabbit antibodies conjugated to HRP. Imaging with BioRad ChemiDoc™. (d) SPM complementation with CEN and 2µ OM45 plasmids. Cells carrying a 2µ OM45(TRP1) plasmid and either of the test plasmids were plated on SCG‐URA low TRP medium supplemented with 5′FAA for TRP1 plasmid counterselection. OM45‐GFP (CEN) (low copy plasmid); OM45‐GFP (2µ) (high copy plasmid); YCp33OM45 (+); YCplac33 (−) (empty plasmid)

In the case of the Δom45ugo1(P189L) SPM, the OM45‐GFP is not rescuing the respiratory deficient phenotype of the mutant strain at all (Figure 6d).

Phospholipid analysis

The results of the synthetic lethal screen implied that Gem1p and Om45p are likely to have some redundancy in their physiological roles. It has been suggested that Gem1p may regulate ER‐mitochondria tethering, although the gem1 deletion alone is not sufficient to impair the phospholipid (PL) trafficking between ER and mitochondria (Nguyen et al., 2012). We decided to investigate whether GEM1 loss in combination with the om45 deletion leads to impaired ER‐mitochondria interaction. Phosphatidylcholine (PC) synthesis from phosphatidylserine (PS) in yeast requires proper tethering of ER with OMM (Daum and Vance, 1997; Carman and Henry, 1999; Stone and Vance, 2000; Kornmann et al., 2011; Nguyen et al., 2012; Horvath and Daum, 2013). To develop an ER‐mitochondria interaction assay we sought to take advantage of this dependence of the phospholipid biosynthesis process on a tight association of the two organelles.

In addition to the mitochondrial PS decarboxylation, there is a minor vacuolar/Golgi pathway in yeast involving Psd2p, which does not depend on mitochondria‐ER interaction. To study exclusively the contribution of mitochondrial PS‐synthesis, all strains were generated in a Δpsd2 background (Trotter et al., 1993; Trotter and Voelker, 1995; Voelker, 1997). If Om45p and Gem1p had a redundant function in mediating ER‐mitochondria interaction, the Δpsd2Δgem1Δom45 triple knockout strain would be predicted to show an accumulation of PS and possibly depletion of phosphatidylethanolamine (PE) or PC due to impaired trafficking between ER and mitochondria. A suitable control is the isogenic W1536 8B Δpsd1Δpsd2 strain, where deletion of both PS decarboxylase enzymes severely compromises phospholipid metabolism, leading to accumulation of unprocessed PS. As additional controls, phospholipid profiles of W1536 8B Δgem1 or W1536 8B Δom45 deletion strains also lacking psd2 were investigated to assess the effect of the individual deletions of these OMM proteins on phospholipid turnover.

UPLC‐Q‐Tof‐MS was used to analyze the lipidomes of whole cells and of highly purified mitochondria from five different deletion strains grown on glycerol‐containing media: W1536 8B Δom45Δpsd2; W1536 8B Δgem1Δpsd2; W1536 8B Δgem1Δom45Δpsd2; W1536 8B Δpsd2 and Δpsd1Δpsd2 (Figure A6). The W1536 8B Δpsd1Δpsd2 control strain lacking both PS decarboxylases is an auxotroph for ethanolamine required to produce the essential PE and PC phospholipids via the Kennedy pathway (McMaster, 2018). Hence, the growth media of all strains were supplemented with 2 mM ethanolamine. The growth of strains on glycerol is cumbersome especially for the very slow‐growing Δgem1Δom45Δpsd2 variant, but as we wanted to avoid lesions secondary to the original mutations, it was necessary to produce cell material under conditions that prevented mtDNA loss.

Our mass spectrometric PL analysis was consistent with the results of Vance et al., who demonstrated that PC is the most abundant PL in eukaryotic cells. PE and PI are also present in ample amounts, while PS, a precursor for PE and PC is quantitatively a minor phospholipid (Vance and Steenbergen, 2005) (Figure A8) The data are presented in the % ratios, with the total sample phospholipids set as100% as well as PS/PE ratio.

As expected, the PS/PE ratio showed a clear increase in the W1536 8B Δpsd1Δpsd2 strain, where deletion of both PS decarboxylase enzymes severely compromised phospholipid metabolism and caused PS accumulation. W1536 8B Δgem1Δom45Δpsd2 did not show any increase in PS, and the PS/PE ratio in this strain was similar to the control W1536 8B Δom45Δpsd2, W1536 8B Δgem1Δpsd2 and W1536 8B Δpsd2 strains both in whole cell lipids extracts and in the highly pure mitochondria. Thus, neither the om45 deletion alone nor in combination with the gem1 deletion appear to affect the PS to PE conversion in yeast cells.

DISCUSSION

The cellular role of Om45, one of the major proteins of the yeast OMM, has remained obscure for several decades. The protein is fixed to the OMM via an N‐terminal α‐helical membrane anchor, and for a long time, it was believed that the major solvent‐exposed C‐terminal domain of Om45 was facing the cytosol. Hence the place of action of Om45 was suspected to be on the mitochondrial outer surface. In 2012, Lauffer et al. first reported evidence for the IMS orientation of Om45p (Lauffer et al., 2012). Data obtained by us in this present study are consistent with the configuration proposed in the latter work.

The discovery of yeast Miro GTPase in the early 2000s initially caused a stir in the mitochondrial community (Frederick et al., 2004). Reports of a role of Gem1 as a partner of the ERMES complex, regulating mitochondria‐ER interactions and phospholipid exchange (Kornmann et al., 2011) could not be unequivocally confirmed (Nguyen et al., 2012). Here, we present evidence for a collaborative role of Om45p and Gem1p in mitochondrial inner membrane organization. We were unable to find any support for a function of these proteins, either alone or in combination, in phospholipid trafficking. Our results imply that the action of Gem1p and Om45p is oriented toward the IMM and required for cristae and mtDNA maintenance.

Our synthetic lethal screen based on the colony color‐sectoring assay revealed two evolutionarily conserved genes, GEM1 and UGO1, as interacting genetically with OM45. The SPMs obtained in our screen as well as the mutant reconstruction variants showed impaired IMM organization in the absence of Om45 where they also become respiratory deficient due to progressive mtDNA loss. The data speak in favor of at least partial redundancy of Om45p with Gem1p and Ugo1p in terms of their physiological function. Unlike Gem1 and Ugo1, which are conserved in higher eukaryotes, no homolog of Om45 could be identified in mammals. Deletions of either GEM1 or UGO1 have been shown to affect mitochondria morphology and mtDNA maintenance in yeast. While Δgem1 strains experience only a mild increase of mtDNA loss (Frederick et al., 2004), the ugo1 deletion results in a rho phenotype (Sesaki and Jensen, 2001). Our ∆om45, ∆gem1 double deletion strain, and gem1 SPMs display a similar loss of mtDNA.

The Ugo1 protein has been reported to bring together OMM and IMM by acting as a scaffold between Mgm1p and Fzo1p proteins in the IMM and the OMM, respectively (Sesaki and Jensen, 2004; Coonrod et al., 2007; Anton et al., 2011). There have been speculations about a possible lipid mixing role of Ugo1 required for mitochondrial fusion (Hoppins et al., 2009). Cells lacking Ugo1p fail to form a tubular mitochondrial network, shifting the mitochondrial dynamics balance toward fragmentation, which ultimately leads to complete loss of mtDNA (Sesaki and Jensen, 2001; Sesaki and Jensen, 2004; Itoh et al., 2013). Mutations in the Ugo1p‐like protein SLC25A46 in humans were reported to cause an optic atrophy spectrum disorder. In contrast to the phenotype of UGO1 deletions in yeast, the knockdown of SLC25A46 was found to cause mitochondrial hyperfusion, while overexpression of the mammalian UGO1 homolog led to mitochondrial fragmentation (Abrams et al., 2015). While we did not pursue any deeper analysis of the ∆om45, ugo1 SPM, our ability to identify such an interaction in general supports our conclusions on the functions of Om45p and Gem1p.

Gem1p is the yeast homolog of human Miro GTPase (Frederick et al., 2004; Tang, 2015). In the absence of Gem1p, yeast mitochondria exhibit a pronounced defect in mitochondrial cellular distribution and morphology, forming collapsed tubular and globular mitochondria with an irregular diameter (Frederick et al., 2004). Unlike its mammalian homolog, the role of Gem1p in interaction with the cytoskeleton has not been observed in yeast. Instead, deletion of GEM1 has been reported to affect the size and the number of mitochondria‐ER encounter structures (ERMES) in S. cerevisiae. The role of Gem1p in ERMES function has been controversial (Nguyen et al., 2012). In our eyes, the synthetic defect in ∆om45/gem1(R103K), ∆om45/gem1(S324N), and ∆om45 /ugo1(P189L) can be only explained by (partially) redundant function(s) of the proteins acting in parallel. Inspired by experiments described previously for Δgem1 strains by the Shaw group (Nguyen et al., 2012), we examined the phospholipid composition of the Δom45 and Δgem1 strains to determine if the redundant functions of Om45p and Gem1p may be obscuring an effect of the Δgem1 mutation on phospholipid levels and, by extension, on ERMES function. Like in the Nguyen et al. report, PS conversion to PE and further to PC in the yeast carrying single deletions of Δom45, or Δgem1 as well as in the Δom45Δgem1 double deletion was not distinguishable from the control. Hence, we found no evidence for a possible common role of the two proteins in mediating the tethering of mitochondria to the ER.

Genetic interaction of OM45 with GEM1 was previously reported in a high throughput study describing components of the mitochondrial contact site (MICOS) complex (Hoppins et al., 2011), a scaffold‐like structure required for mitochondrial IMM organization (Alkhaja et al., 2012; Harner et al., 2011; van der Laan et al., 2012; Pfanner et al., 2014; Rabl et al., 2009; Zerbes et al., 2012). Our SPMs are characterized by strong distortion of IMM folding, which may be a reason for progressive mtDNA loss. We argue for the IMM folding disturbance as the cause of the mtDNA loss in the SPMs rather than a result thereof based on three observations: First, mitochondrial morphology in the mutants is not consistent with the morphology observed in rho cells (compare Figure 2/Figure A5 and (Sesaki & Jensen, 2001, Figure 3)), and mitochondrial ultrastructural changes in rho0 cells differ from the changes observed in the SPMs (Figure 4, Figure A2). Second, and even more compelling, we can show that the IMM disturbance is observable already in cells that still harbor mtDNA (Figure 5b, representative micrograph of a Δom45Δgem1 cell grown on glycerol, Figure 5g, mtDNA content of Δom45Δgem1 cells grown on glycerol). Lastly, at least the ugo1(P189L) SPM clearly maintains mtDNA despite severe mitochondrial morphology and ultrastructural defects (Figure A5).

Intriguingly, the combination of mutations in OMM proteins causes the strongest phenotype in the IMM folding. The reported IMS orientation of Om45p (Lauffer et al., 2012; Song et al., 2014; Wenz et al., 2014), however, makes a role of the protein in the mediation of OMM and IMM contacts plausible.

In addition to the GEM1 and UGO1 isolates obtained by library complementation, we found several multicopy suppressors that can rescue the synthetic petite phenotypes. Among them are FZO1, MCP1 encoding OMM proteins, as well as TIM44, SHY1, PET9, and PET54 encoding IMM proteins. The function of Mcp1p is not well understood, but it appears to be involved in lipid homeostasis (Tan et al., 2013). Physical interaction of Fzo1 with Ugo1 is required for mitochondrial fusion (Sesaki and Jensen, 2004; Coonrod et al., 2007; Anton et al., 2011). Shy1 and Pet54 play roles in respiratory complex IV component expression and assembly. Shy1p is necessary for full expression of mitochondrial COX1 in the yeast model of Leigh's syndrome (Barrientos et al., 2002). The PET54 gene of Saccharomyces cerevisiae: characterization of a nuclear gene encoding a mitochondrial translational activator and subcellular localization of its product (Costanzo et al., 1989). Pet9 (Aac2) is an ADP/ATP carrier of the IMM, affecting the level of ATP in the IMS and also found to be associated with MICOS and Om45‐Om14‐Por1 protein complexes (Lauffer et al., 2012; Linden et al., 2020). Tim44 is an essential component of the TIM23 complex (Blom et al., 1993; Pfanner et al., 1996; Popov‐Čeleketić et al., 2008), physically interacting with Ssc1 (D'Silva et al., 2004; Liu et al., 2001; Wadhwa et al., 2002). While we do not clearly understand the mechanisms of multicopy suppression by these factors as a whole they may point toward membrane fusion and protein import into the mitochondrial matrix as well as respiratory chain assembly functions.

Although Om45p and Gem1p (as well as Ugo1p) reside on the OMM, our results firmly indicate the IMM is most strongly affected by their loss of function. Our investigation of the role of Om45 under growth conditions that have not been tested before as well as the genetic and physical interactions of Om45p that were identified during this work revealed several cellular processes affected by Om45p. In particular, mitochondrial morphology maintenance, IMM folding, and mtDNA inheritance are affected by the absence of Om45p in a gem1 (or ugo1) mutant strain background. Taken together, the mutant morphology, growth phenotypes, multicopy suppressors, and phospholipidome studies we present here point to a possible role of Om45 and Gem1 in IMM homeostasis and mediation of interaction between the two mitochondrial membranes. We find no direct evidence for the role of these proteins in ER/mitochondrial phospholipid exchange processes.

An additional aspect of our work is that the phenotype of the synthetic mutant strains is strongly dependent on the presence of Om45p, which makes these synthetic petites a unique system for testing the functionality of Om45p constructs. Our results indicate that the C‐terminal tagging of Om45p interferes with its functions. GFP‐tagged Om45p did not complement the synthetic mutants and displayed a growth defect in the Δgem1Δom45 and Δom45 strain backgrounds. In light of our observation of the role of Om45p in mitochondrial morphology maintenance, such constructs may not be the best choice as a marker for mitochondrial morphology studies for which it has been used previously (Sesaki and Jensen, 2001; Sesaki et al., 2003).

ETHICS STATEMENT

None required.

CONFLICT OF INTEREST

None declared.

AUTHOR CONTRIBUTIONS

Antonina Shvetsova: Conceptualization (equal); Formal analysis (lead); Investigation (lead); Methodology (equal); Project administration (equal); Validation (lead); Visualization (lead); Writing‐original draft (lead); Writing‐review & editing (lead). Ali Julfiker Masud: Conceptualization (supporting); Formal analysis (supporting); Investigation (supporting); Methodology (supporting); Project administration (supporting); Writing‐review & editing (supporting). Laura Schneider: Investigation (supporting); Methodology (supporting); Writing‐original draft (supporting); Writing‐review & editing (supporting). Geoffray Monteuuis: Investigation (supporting); Methodology (supporting); Writing‐original draft (supporting); Writing‐review & editing (supporting). Ulrich Bergmann: Conceptualization (supporting); Formal analysis (equal); Investigation (supporting); Methodology (equal); Project administration (supporting); Supervision (equal); Validation (lead); Writing‐original draft (supporting); Writing‐review & editing (supporting). Ilkka Miinalainen: Formal analysis (equal); Investigation (supporting); Methodology (equal); Writing‐original draft (supporting); Writing‐review & editing (supporting). J. Kalervo Hiltunen: Conceptualization (supporting); Formal analysis (supporting); Methodology (supporting); Project administration (supporting); Supervision (equal); Writing‐original draft (supporting); Writing‐review & editing (supporting). Alexander Kastaniotis: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Funding acquisition (equal); Investigation (equal); Methodology (lead); Project administration (lead); Resources (equal); Software (equal); Supervision (lead); Validation (equal); Visualization (equal); Writing‐original draft (equal); Writing‐review & editing (equal).