Figure 1
Click on the picture to view a high resolution version
PUF3 deletion affects the asymmetric localization of a specific subclass of MLR-mRNAs. A: Distribution of mRNA enrichment copurified with mitochondria for the 794 genes (786 nuclear + 8 mitochondrial genes) encoding mitochondrial proteins. The MLR value is expressed as a percentage mitochondrion-bound mRNA fraction, which is derived from the log2 values (see Materials and Methods). A cutoff value of 8% was statistically determined (see Figure S1), which is consistent with the quantitative assessments by RT-PCR of the localization of 100 mRNAs [12]. The number of genes with significantly asymmetrically localized mRNA is indicated on the right of the 8% threshold value. The 100% arrow indicates the median of the log(2) ratio of the 33 mitochondrial transcripts and represent 100% of mitochondrial localization. COX2 is indicated as a permanent mitochondrial resident (gene and mRNA). The 0% arrow indicates the median of the log(2) ratio of the 416 mRNAs not connected with mitochondria and represents the 0% of mitochondrial localization. The specific case of COX17 mRNA is worth considering as this mRNA is clearly delocalized in the absence of Puf3p. B: PUF3 motif : Puf3p target mRNAs(Class I) were defined as having a PUF3 motif in their 3'UTR and/or as interacting with Puf3p [17]. C: Scatter plot representation of the global asymmetric localization of class I (left panel) and class II (right panel) MLR transcripts in wild-type and Delta-puf3 strains. The set of mRNAs has been plotted in x-coordinate and sorted according to WT mitochondrial percentage of localization from 100% to 8 % represented by blue squares. For each single mRNA, the corresponding percentage of localization in Delta-puf3 mutant is represented by a red square. The number and percentage of class transcripts showing a decrease or an increase of the mitochondrial localization rate in absence of Puf3p compared to WT is indicated within the plot. A significant decrease of the mitochondrial localization rate is observed in absence of Puf3p for the class I MLR mRNAs (p value, p = 3.06 10-5). Conversely a significant increase of the mitochondrial localization rate is observed in absence of Puf3p for the class II MLR mRNAs (p value, p = 5.56 10-3). D: Table recapitulating the different classes of MLR mRNAs (Class I and II) and non MLR mRNAs (class III). The table shows that the presence of 3'UTR PUF3 motif and/or biochemical association with Puf3p (Puf3p targets column) are predominantly associated with MLR mRNAs (class I + class II) (student test p= 8.4 10-7).
Figure 2
Click on the picture to view a high resolution version
Mutation of the BCS1 Puf3p binding site affects its mRNA mitochondrial localization. A: Six of the ten nucleotides of the putative BCS1 Puf3p binding site have been deleted (red frame) or substituted (red letters) on pRS416 centromeric plasmid carrying BCS1 ORF and its own promoter (A generous gift of G. Dujardin). A bcs1 deleted strain has been transformed by either the plasmid carrying the wild type BCS1 gene (pRS416 + BCS1) or a plasmid containing the mutated bcs1 allele (pRS416 + bcs1-d1). B: FISH analysis of strain carrying WT or mutated BCS1 plasmids. Five fluorescent probes specific for mitochondrial ribosomal RNA delimit the mitochondrial compartment (red). Specific mRNAs (ATP4, ATP16, BCS1) were labeled in with specific sets of probes (green). The nucleus is stained by DAPI (blue). 410 cells (BCS1) and 417 cells (bcs1-d1) were examined using BCS1 specific FISH probes. At least 100 cells were examined for the ATP16 and ATP4 specific probes. Projections of 3D reconstructions are presented for the mutant and wild type cells and the 3 different FISH probes. C: Quantification of mitochondrial mRNA colocalization from the FISH experiment has been done using Corsen software (see Material and Methods) and presented in this histogram. The number of cell used for quantification is presented on top of each bar. Corsen allows the calculation of the percentage of cells showing significant mRNA asymmetric localization to the vicinity of mitochondria. Calculation was done in cells expressing WT (pRS416+BCS1) or mutant (pRS416+bcs1-d1) BSC1 alleles for ATP4, ATP16 and BCS1 mRNAs. For each measure a represented confident interval is calculated based on a binomial distribution. D: RT-quantitative PCR quantification of mRNAs associated with mitochondrial fraction in cells expressing WT (pRS416+BCS1) or mutant (pRS416+bcs1-d1) BCS1 alleles. Total and mitochondrial fraction RNAs were extracted as in Material and Methods. The real time quantification is done according to Garcia et al.(12). Each measure was made in triplicate and two independent quantitative PCR have been performed allowing the calculation of a standard deviation presented on the histogram.
Figure 3
Click on the picture to view a high resolution version
Effects of cycloheximide and puromycin on the asymmetric localization (MLR) of nuclear-encoded mitochondrion-linked mRNAs. The distributions focus on the 794 genes encoding mitochondrially localized proteins (A) in presence of cycloheximide (B) or puromycin (C). The threshold of 8% was determined as in figure 1. A few mRNAs are indicated as markers of different subcellular translation sites: mitochondrial mRNAs (COX2), free cytoplasmic polysomes (ATP16), mitochondrion-bound cytoplasmic mRNAs delocalized to free polysomes by cycloheximide (COX17, YAH1), or mitochondrion-bound mRNA whose mitochondrial enrichment is increased by cycloheximide (ATP2). All represent different classes of mRNAs, the properties of which are described in detail in the supporting information (Table S1). D: Scatter plot representation of the Mitochondrial Localization Rate of the 794 transcripts extracted from cells treated in presence of cycloheximide (left panel) or puromycin (right panel) and compared to untreated cell. The set of mRNAs has been plotted in x-coordinate and sorted according to the untreated cell mitochondrial percentage of localization (MLR) and represented by blue squares. For each mRNA, the corresponding percentage of localization calculated from extract of cell treated with translation inhibitor is represented by a yellow square (cycloheximide) or green square (puromycin). E: Box-plot representation of the asymmetric localization of mRNAs encoding mitochondrial proteins. The line through the center of each box represents the median value of the distribution; the size of the box represents the median absolute deviation of the distribution. Three sets of experimental conditions were analyzed in triplicate: no translation inhibitor (left), with cycloheximide (centre), with puromycin (right).
Figure 4
Click on the picture to view a high resolution version
PUF3 deletion affects the asymmetric localization of different functional modules differently. A: Distribution of Delta-puf3 effects on some of the mitochondrial functional modules described by Perocchi [5]. Only the functional modules for which average MLR value was significantly affected in the Delta-puf3 strain have been represented. The mean RNA localization value of each functional module is either decreased (class A) or increased (class B). Asterisks denote the p values (Student test) of each wild type versus mutant differences: * = p < 0.05, ** = p < 0.01, *** = p < 0.001. All A-type functional modules are involved in translation or are assembly factors: thus RCCasm modules (1 to 5) include 29 genes (ATP11, ATP12, SHY1, COX17, etc...); translation module includes 15 genes mainly involved in mitochondrial translation control (CBS2, CBP6, PET494, etc...); MRPL and MRPS contain 48 and 37 genes coding for mitochondrial ribosomal proteins respectively ; B-type functional modules concern metabolic pathways: Branched aa (10 genes), TCA (13 genes), isoTCA (6 genes), Glutamate (6 genes), Ketoglutarate (13 genes), Folate (10 genes), Arginine (5 genes), Alcohol (10 genes). Detailed for all these modules can be found in [5] and in Table S1. The significant increased mean MLR value after PUF3 deletion might be the consequence of the modification of the global PUF proteins due to the absence of Puf3p. B. The specific example of cytochrome c-oxidase complex biogenesis (RCC4). Concerning the nuclear encoded subunits, the 18 assembly factors necessary for the synthesis of the complex [4] are translated on polysomes linked to mitochondria whereas the 8 structural proteins are translated on cytoplasmically-free polysomes (Table S4 for gene details)
Figure 5
Click on the picture to view a high resolution version
The 480 mRNAs localized to the vicinity of mitochondria (MLR-RNAs) can be classified according to their Puf3p-dependence Two main classes of MLR-mRNAs can be identified: class I and class II are distinguished by the presence or absence of at least one PUF3 motif in their 3'UTR sequence respectively. In addition classI mRNAs can be, in the absence of PUF3, totally (class I-1) or partially (class I-2) delocalized. This property seems to be correlated with the high sensitivity to cycloheximide for class I-1 mRNAs. We propose that class I-1 Puf3p target mRNAs requires mainly Puf3p for their mitochondrial localization whereas class I-2 mRNAs localization also depends on other unknown trans-acting products. Class II mRNAs are localized to the vicinity of mitochondria under the control of unidentified trans-acting factors.