MitoLoc: A method for the simultaneous quantification of mitochondrial network morphology and membrane potential in single cells.

Yeast strains, media and plasmids All experiments involving wild-type yeast strains were carried out using YSBN1, a prototrophic diploid variant of S. cerevisiae S288c ( Canelas et al., 2010 ). Yeast strains deleted for proteins involved in mitochondrial morphology were obtained from the yeast gene deletion collection ( Winzeler et al., 1999 ). Cells were cultured in YPD medium (1% yeast extract, 2% peptone, 2% glucose) or synthetic complete (SC) medium with 100 μg/mL Nourseothricin (Werner BioAgents) for a minimum of 4 h to ensure log-phase growth. Where indicated, cells were treated with 1 mM H 2 O 2 (Sigma) 45 min prior to collection, or with CCCP (Sigma) at the time and concentration indicated. Yeast cells depleted of mtDNA (ρ 0 ) were generated by plating onto YPD agar containing 0.1 g/L ethidium bromide and incubation at 30 °C for 2 d. Absence of mtDNA was confirmed in isolated surviving clones by DAPI staining and growth assays using the non-fermentable carbon sources ethanol and glycerol. Staining with DiOC 6 (Invitrogen) was performed according to manufacturer's instructions, and cells were mounted on agarose pads containing the respective culture conditions. pMitoLoc was constructed by first replacing the URA3 marker gene of pUG35 ( Niedenthal et al., 1996 ) with the nouresothricin marker (NAT) from pAG25 using homologous recombination cloning. Then, the preSU9 localisation sequence of pYES_mtGFP (a gift from B. Westermann ( Westermann and Neupert, 2000 )) was inserted 5′ of the yEGFP gene using Bam HI and Eco RV sites. Subsequently, we used the plasmid's Sac I and Bst BI sites to introduce the preCOX4-mCherry gene of pHS12-mCherry (a gift from C. Dunn ( Sesaki and Jensen, 1999 )), resulting in a dual-reporter CEN6 plasmid we termed pMitoLoc. The plasmid is made available through Addgene ( www.addgene.org ), plasmid ID #71207.

Fluorescence microscopy

For microscopy, ~ 6 × 10 6 yeast cells were collected by centrifugation, washed twice in PBS and re-suspended in formaldehyde solution (4 g/L PFA, 3.6% sucrose) to preserve mitochondrial morphology. After 15 min, cells were washed in PBS, and where indicated 2.5 μg/mL DAPI or 5 μg/mL Calcofluor White (Sigma) was added. After one more washing step with PBS, cells were resuspended in 20 μL Vectashield mounting medium (Vector Labs). 2 μL of this mixture was applied to poly- l -lysine coated microscope slides. For live-cell microscopy, cells were mounted on agarose pads. Super-resolution fluorescence microscopy for morphological analysis was carried out using a Deltavision 3D-SIM OMX system (GE Healthcare) equipped with a 100 × 1.4NA oil objective (Olympus), 405 nm, 488 nm and 594 nm laser lines, and the OMX Standard filter set drawer. Images were acquired in structured illumination mode using a Z-spacing of 125 nm, and reconstructed using Softworx software (GE Healthcare). Conventional widefield fluorescence microscopy was carried out using an Olympus IX81 wide field microscope (Deltavision, GE Healthcare) equipped with a 60 × 1.42NA PlanApoN oil objective (Olympus) and an LED light source capable of delivering 405 nm, 488 nm and 594 nm excitation wavelengths. The filter sets used were FITC (490/20ex 528/38em), TRITC (555/28ex 617/73em) and DAPI (360/40ex 457/50em), and images with a Z-spacing of 200 nm were recorded with a CoolSNAP HQ2 CCD camera. Deconvolution was performed using Softworx software. For standard fluorescence microscopy, cells were examined under an Olympus BX51 microscope using filters YGFP (GFP) and HcRed1 (mCherry). Images were recorded with the help of QImage software.

Mitochondrial morphology analysis

Super-resolution images were analysed by the software plugin Yeast_MitoMap (available via web supplement and from http://www.gurdon.cam.ac.uk/stafflinks/downloadspublic/imaging-plugins ) in ImageJ ( Schneider et al., 2012 ). MitoMap automates the process of defining GFP-labelled mitochondria in a selected region of interest and calculates their volume, surface area and shape descriptors using the formulae listed in Table 1 . Documentation on how to use the plugin is included in the supplementary material. 32-bit OMX image stacks are converted to 16-bit and Otsu thresholding ( Otsu, 1979 ) is used to extract the labelled volume. From this volume, surface voxels are defined as those having at least one exposed face and assigned to the classes defined by Mullikin and Verbeek ( Mullikin and Verbeek, 1993 ) extended with additional classes to allow for different dimensions in xy and z. This gives a total of 15 different possible surface voxel configurations, each with a weighting factor used to estimate their contribution to the 3D object surface area. This method was validated by comparing the estimated surface areas of binary voxel representations of spheres to the calculated volumes of spheres with the same radii. For each single or dividing cell, a ROI was chosen that contained no other cells. Generally, cells with the highest absolute intensities were chosen where more than 30 cells had been acquired. To exclude artefacts, objects with a volume smaller than 0.1 μm 3 were excluded. For geometric analysis, each object's features were weighted by the respective surface area to minimize over-representation of small objects. For fragmentation analysis, relative volumes V s of each cell's objects were calculated. Then, relative volumes from 30 cells per genotype were added using a V s binning of 10 considering objects with V s ≤ 20 as fragmented ( Table 1 ). In the case of H 2 O 2 -treated yeast, we observed a heterogeneous population of ~ 40% cells with wild type mitochondria, while the remaining cells presented with a heavily fragmented mitochondrial network. This observation was reproducible and occurred in spite of vigorous mixing after H 2 O 2 addition. We therefore excluded non-responding cells from the analysis. Data was plotted using R ‘ggplot2’ and ‘ggbiplot’ packages. Imaris software (Oxford Instruments) was used to generate 3D renderings of yeast mitochondria.

Mitochondrial protein import analysis

Colocalisation of preSU9-GFP and preCOX4-mCherry was quantified using images acquired with conventional resolution by first cropping each image to contain one single or dividing yeast cell. Cropping was based on Calcofluor White staining in order to avoid artefacts. Image stacks were subjected to colocalisation analysis in Volocity software (Perkin Elmer) without defining further ROIs, as automatic cell shape definition using brightfield images or Calcofluor staining proved unreliable. Alternatively, image stacks were analysed using ImageJ plugin yeast_correlation (available via web supplement and from http://www.gurdon.cam.ac.uk/stafflinks/downloadspublic/imaging-plugins ), where cell areas for PCC analysis were defined by applying the Otsu thresholding method ( Otsu, 1979 ) to images convolved with a Gaussian blur (σ = 5).

Oxidant tolerance tests

Exponentially growing yeast cells were collected and spotted in 1/5 serial dilutions onto SC (6.8 g/L YNB (Sigma), 0.59 g/L CSM (MP Biomedicals)) containing 2% glucose and 1.25 mM diamide (Sigma), or 3% glycerol where indicated. Growth was documented after 3 days incubation at 30 °C.

Appendix A Supplementary data

Supplementary material: ImageJ plugins. Supplementary figures.