Computationally driven, quantitative experiments discover genes required for mitochondrial biogenesis - PubMed (original) (raw)
. 2009 Mar;5(3):e1000407.
doi: 10.1371/journal.pgen.1000407. Epub 2009 Mar 20.
Chad L Myers, Curtis Huttenhower, Matthew A Hibbs, Alicia P Hayes, Jadine Paw, John J Clore, Rosa M Mendoza, Bryan San Luis, Corey Nislow, Guri Giaever, Michael Costanzo, Olga G Troyanskaya, Amy A Caudy
Affiliations
- PMID: 19300474
- PMCID: PMC2648979
- DOI: 10.1371/journal.pgen.1000407
Computationally driven, quantitative experiments discover genes required for mitochondrial biogenesis
David C Hess et al. PLoS Genet. 2009 Mar.
Abstract
Mitochondria are central to many cellular processes including respiration, ion homeostasis, and apoptosis. Using computational predictions combined with traditional quantitative experiments, we have identified 100 proteins whose deficiency alters mitochondrial biogenesis and inheritance in Saccharomyces cerevisiae. In addition, we used computational predictions to perform targeted double-mutant analysis detecting another nine genes with synthetic defects in mitochondrial biogenesis. This represents an increase of about 25% over previously known participants. Nearly half of these newly characterized proteins are conserved in mammals, including several orthologs known to be involved in human disease. Mutations in many of these genes demonstrate statistically significant mitochondrial transmission phenotypes more subtle than could be detected by traditional genetic screens or high-throughput techniques, and 47 have not been previously localized to mitochondria. We further characterized a subset of these genes using growth profiling and dual immunofluorescence, which identified genes specifically required for aerobic respiration and an uncharacterized cytoplasmic protein required for normal mitochondrial motility. Our results demonstrate that by leveraging computational analysis to direct quantitative experimental assays, we have characterized mutants with subtle mitochondrial defects whose phenotypes were undetected by high-throughput methods.
Conflict of interest statement
The authors have declared that no competing interests exist.
Figures
Figure 1. An overview of our iterative framework integrating computational and experimental methodologies for discovery of gene function.
Our study uses an ensemble of computational gene function prediction methods (bioPIXIE , MEFIT , and SPELL [29]), each of which predicts new genes involved in mitochondrial function based on high-throughput data and examples of known mitochondrial proteins (the gold standard, Table S3). Complete lists of predictions are provided (Tables S4 and S5). We selected test candidates by integrating these approaches based on estimated precision of each method and tested these predictions experimentally using three biological assays (see Methods for details). Upon evaluating these experimental results, the proteins newly discovered to be involved in mitochondrial function were added to the known examples, and the process was iterated to comprehensively characterize additional mitochondrial proteins. See Table 1 for an overview of our results, and Table S2 for a full listing of results.
Figure 2. Schematic overview of the petite frequency assay.
(A) Initially, strains were grown in a non-fermentable carbon source (liquid YP-Glycerol) for 48 hours. All cells growing under this condition must be respiratory competent and contain functional mitochondria. Any strain with no viable cells after this step was deemed respiratory deficient and did not continue in the assay. (B) Cell cultures were serially diluted and plated on a fermentable medium (YPD) and grown for 48 hours to form colonies founded from a single cell. At this point, the requirement for respiratory competency is lifted, so that daughter cells can survive while losing respiratory function. (C) A single colony is picked and briefly re-suspended in water. (D) The suspension is diluted and plated on YPD and grown to form colonies founded from single cells. After 48 hours, agar containing tetrazolium is overlaid on the plates. Colonies founded by respiratory competent cells will take up the tetrazolium and appear large and red. Colonies founded from respiratory deficient cells (“petite” colonies) appear smaller and white.
Figure 3. The combination of computational predictions and quantitative assays discovers novel genes involved in mitochondrial function.
(A) Mitochondrial transmission rates of single gene knockouts were determined for 193 genes predicted to be involved in mitochondrial function and for 48 control genes known to be involved (raw data in Table S6). A box plot is shown for each deletion strain tested; red indicates the inability to grow on a non-fermentable carbon source (glycerol), yellow indicates a mitochondrial transmission rate significantly altered from wild type, and gray indicates no significant difference from wild type. Significance was determined using a Mann-Whitney U-test comparing at least 12 independent measurements of wild type to at least 8 independent measurements of each mutant strain. The green shaded region indicates one quartile above and below the median rate for all 358 wild type replicates. A total of 100 of the 193 prediction candidates were confirmed (an additional 9 genes were confirmed through double knockout analysis, see Figure 4). (B) Distribution of petite frequency phenotypes among positive controls (left), first iteration predictions (center), and second iteration predictions (right) with colors as in (A). Severe phenotypes (red) were more prevalent among positive controls, while the majority of confirmed predictions exhibited an intermediate phenotype (yellow). We hypothesize that this difference is due to a bias towards detection of extreme phenotypes in classical genetic screens and high throughput methodologies.
Figure 4. Double mutant petite frequency phenotypes.
Based on their persistence as strongly predicted candidates during our second iteration, we selected 26 genes unconfirmed by single mutant analysis for investigation of synthetic phenotypes. The single mutant petite frequency is shown for each of these strains on the left. Each of the 26 strains was crossed with 4 genes known to be involved in mitochondrial transmission (_aim17_Δ, _tom6_Δ, _rvs167_Δ, and _ehd3_Δ) to create ∼100 double mutant strains. Results are shown in blue for each of the 4 strains crossed into, followed by all 26 double mutants constructed against that strain (raw data Table S7). The order of the double mutants is the same as in the 26 single mutants shown on the left. Colors are as in Figure 3. Significantly altered double mutant strains are marked with numbers, corresponding to the key above the box plots.
Figure 5. YIR003W (AIM21) is required for mitochondrial motility.
(A)–(B). Dual immunofluorescence of mitochondria (outer membrane protein porin stained in red) and actin (total actin, stained in green) in the indicated yeast strains (scale bar 2 µm). (C) Mitochondrial motility was measured in strains carrying an integrated mitochondrially-targeted GFP (methods) by tracking the movement of the tip of a mitochondrion within a budding cell every second for two minutes. A sustained mitochondrial movement is defined as movement in the same direction for at least three consecutive seconds. PUF3 is a gene with known involvement in mitochondrial motility . To determine the frequency of sustained mitochondrial movement resulting from Brownian motion or other passive processes (methods), sustained mitochondrial movement was measured in the presence of the metabolic inhibitors sodium azide (NaN3) and sodium fluoride (NaF). 10 mM concentrations of these inhibitors were compared to a control of 10 mM sodium chloride (NaCl). Raw data are available in Table S8. Due to its lack of static actin or mitochondrial phenotypes, the motility defect in AIM21 mutants would be difficult to find without integrative computational predictions driving specific experimental assays.
Figure 6. Respiratory growth phenotypes.
(A) Scatter plot of growth rate and saturating density measured from growth curves in minimal non-fermentable media (raw data in Dataset S1). The vertical axis indicates the maximum (saturating) optical density achieved by the strain, and the horizontal axis represents the estimated doubling time based on an exponential fit to the growth curve (methods). Green shading indicates the distribution of all 536 wild type measurements. Triangles represent strains with saturation density and/or doubling time significantly altered on glucose, while squares represent strains that showed normal growth on glucose. Each point is colored by the strength of its respiratory growth phenotype (see part C). (B) Example growth curves for wild type and strains representing each of the three phenotypic classes: weak, moderate, and severe respiratory growth defects. (C) Determination of respiratory growth phenotype. Each growth parameter (saturation density and doubling time) was statistically scored as no effect (+), intermediate effect (+/−), or extreme effect (−) (methods). The combination of saturation density and doubling time results produces a final respiratory growth phenotype, with maroon representing a severe defect, purple a moderate defect, blue a weak defect, and gray no defect. Respiratory growth is not strongly correlated with petite frequency (Figure S3).
References
- Murali TM, Wu CJ, Kasif S. The art of gene function prediction. Nat Biotechnol. 2006:1474–1475; author reply 1475–1476. - PubMed
- Schaefer AM, Taylor RW, Turnbull DM, Chinnery PF. The epidemiology of mitochondrial disorders–past, present and future. Biochim Biophys Acta. 2004;1659:115–120. - PubMed
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Molecular Biology Databases