Cyanidioschyzon merolae aurora kinase phosphorylates evolutionarily conserved sites on its target to regulate mitochondrial division

Abstract

The mitochondrion is an organelle that was derived from an endosymbiosis. Although regulation of mitochondrial growth by the host cell is necessary for the maintenance of mitochondria, it is unclear how this regulatory mechanism was acquired. To address this, we studied the primitive unicellular red alga Cyanidioschyzon merolae, which has the simplest eukaryotic genome and a single mitochondrion. Here we show that the C. merolae Aurora kinase ortholog CmAUR regulates mitochondrial division through phosphorylation of mitochondrial division ring components. One of the components, the Drp1 ortholog CmDnm1, has at least four sites phosphorylated by CmAUR. Depletion of the phosphorylation site conserved among eukaryotes induced defects such as mitochondrial distribution on one side of the cell. Taken together with the observation that human Aurora kinase phosphorylates Drp1 in vitro, we suggest that the phosphoregulation is conserved from the simplest eukaryotes to mammals, and was acquired at the primitive stage of endosymbiosis.

Keywords: Cell biology; Plant sciences.

Fig. 1. The Cyanidioschyzon merolae Aurora kinase CmAUR has conserved properties of Aurora kinase orthologs.

a Localization of CmAUR by immunofluorescence analysis. CmAUR was stained with CmAUR antibody. Mitochondria were stained by Ef-Tu antiserum. DNA was stained with DAPI. White arrowheads indicate signals localized to the spindle or spindle pole. Black arrowheads indicate plastid autofluorescence. Arrows indicate plastid DNA. The Pearson correlation coefficient (PCC) in each cell was calculated using areas without plastids. We observed 24 cells and present representative images in this figure. b Autophosphorylation assay of wild-type and mutant CmAUR. ATP-γS was used as a substrate for autophosphorylation, and the phosphorylation was detected by western blotting using thiophosphate ester antibody. c Mitotic inhibition in a dominant-negative mutant of CmAUR. CmAUR^K208R is a kinase-dead mutant; GFP was used as a negative control. Relevant genes were transiently overexpressed in C. merolae and mitotic cells among transfectants were counted. The error bars indicate standard error of the mean. d Immunofluorescence image of histone H3 Ser10 phosphorylation in mitosis. More than 18 cells were observed. e In vitro phosphorylation assay of histone H3 Ser10 with recombinant CmAUR. Glutathione-S-transferase (GST) was used as a negative control. Distilled water was used as a blank sample. f In vitro phosphorylation assay of histone H3 Ser10 treated with hesperadin. The concentration of hesperadin was 10 µm. The concentration of dimethyl sulfoxide in the final solution was 0.1% (v/v). Bars: 1 µm a, 2  μm c.

Fig. 2. CmAUR involvement in mitochondrial division in mitosis.

a Localization of activated CmAUR. Activated CmAUR was visualized using antibody to phosphorylated Aurora kinase. Mt indicates mitochondria, which were stained by Ef-Tu antiserum. Arrows indicate speckles localized to the mitochondrial division site. White arrowheads indicate signals that seemed to be localized to the spindle or spindle pole. Asterisks indicate speckles, which are not predicted to be localized to a specific organelle. Black arrowheads indicate plastid autofluorescence signals. Inter, interphase; Pt/Mt division, plastid/mitochondrial dividing phase; PC, phase contrast image. The PCC of each cell was calculated using areas without plastids. More than 30 cells were observed. b Colocalization of activated CmAUR and mitochondrial ring protein Mda1. More than 30 cells were observed. c Intensity of activated CmAUR speckles localized to the mitochondrial division ring. The relative intensity of human phospho-Aurora antibody (phAUR) and Mda1 in each cell is indicated on the right and left, respectively. **p = 0.0045 (two-sided Student’s t test). d Scheme of mitochondrial division in C. merolae. e, f Effect of kinase-dead mutant e and overexpression f of CmAUR on mitochondrial division in C. merolae. GFP was used as a negative control. GFP, CmAUR^K208R-GFP, and CmAUR-GFP were transiently overexpressed in C. merolae. According to the classification in d, transfectants were counted. g Co-staining of CmAUR and with ProQ Diamond. CmAUR was stained with α-CmAUR antibody. The two images are of representative M-phase cells. The red autofluorescence signal indicates plastids. h Colocalization analysis of ProQ Diamond and CmAUR. Co-stained M-phase cells were analyzed. Each point indicates the PCC of individual C. merolae cells. The area in each cell without plastids was used for analysis. n = 135. Bars: 1 μm (a, b, left), 250 nm (b, right), 2 µm g.

Fig. 3. CmAUR phosphorylated mitochondrial division-related proteins in vitro.

a–c In vitro kinase assays of recombinant CmAUR with recombinant mitochondrial division ring proteins. CBB staining indicates Coomassie Brilliant Blue staining.

Fig. 4. Effects of amino-acid substitutions of phosphorylation sites of Dnm1 by Aurora kinase in Dnm1 on mitochondrial division.

a In vitro kinase assay of recombinant CmAUR and site-directed Dnm1 mutants and those with mutations of 4, 7, 8, and 9 residues (see Supplementary Table 3). b Quantification of in vitro kinase assays of CmDnm1 single amino-acid variants. Signals from different membranes were normalized against the luminosity of wild-type CmDnm1. The relative value of wild-type CmDnm1 was 1. n = 3 for T139A, n = 4 for S726A. The error bars indicate standard error of the mean. c Frequency of transformants expressing CmDnm1 variants. Alanine and phosphomimetic glutamic acid mutants of CmDnm1^T139 and CmDnm1^S726 were overexpressed in C. merolae cells. According to classification Fig. 2d, transfectants were counted. Transfectants with an aberrant number of plastids (three or more) were not counted. d Normal phenotype of dividing mitochondrion in cell overexpressing wild-type CmDnm1. e Phenotype of single mitochondrion without division in cell overexpressing CmDnm1^S570E. f Phenotype of multiple chloroplasts in CmDnm1^T139A overexpression mutant. g Frequency of cells with multiple chloroplasts among Dnm1-mutant transfectants. Total countable transfectants were counted. h In vitro kinase assay of human Aurora kinases with recombinant GST-human Drp1. Bars: 1 μm.