2025 年 101 巻 1 号 p. 41-53
Cell proliferation is a fundamental characteristic of organisms, driven by the holistic functions of multiple proteins encoded in the genome. However, the individual contributions of thousands of genes and the millions of protein molecules they express to cell proliferation are still not fully understood, even in simple eukaryotes. Here, we present a genome-wide translation map of cells during proliferation in the unicellular alga Cyanidioschyzon merolae, based on the sequencing of ribosome-protected messenger RNA fragments. Ribosome profiling has revealed both qualitative and quantitative changes in protein translation for each gene during cell division, driven by the large-scale reallocation of ribosomes. Comparisons of ribosome footprints from non-dividing and dividing cells allowed the identification of proteins involved in cell proliferation. Given that in vivo experiments on two selected candidate proteins identified a division-phase-specific mitochondrial nucleoid protein and a mitochondrial division protein, further analysis of the candidate proteins may offer key insights into the comprehensive mechanism that facilitate cell and organelle proliferation.
Mitochondria and chloroplasts are thought to have originated from free-living prokaryotes through endosymbiotic events.1)-4) The birth of endosymbiotic organelles brought significant bioenergetic benefits to eukaryotes.5)-7) However, innovations in cellular systems and the resultant complexity of cell structures necessitated the establishment of a more highly regulated mechanism for proliferation in eukaryotic cells than in prokaryotic cells. To perform cell proliferation, parental DNA must be copied and segregated into the two daughter cells to pass on genomic information, and this is required for not only nuclear DNA but also endosymbiotic organelles such as mitochondria and chloroplasts.8),9) As with DNA, endosymbiotic organelles themselves must divide to be inherited by daughter cells. Division of these endosymbiotic organelles is carried out by specific protein complexes called mitochondrial and chloroplast division machineries.10)-15) The division machineries are constructed along the mid-region of these organelles just before organelle division. Although some components have been lost and replaced by others, the division machineries are composed of multi-layered ring structures and are well conserved throughout eukaryotes. An inner ring structure, the FtsZ ring, derived from the bacterial division system,16)-18) is placed in the matrix or stromal region.19)-21) An outer ring structure, the MD/PD ring and dynamin ring, evolved in eukaryotes,22) is assembled on the outer organelle membranes.10),23)-27) Importantly, genes encoding the responsible proteins of the division rings are not in organelle DNAs but in nuclear DNA, and the expression of these genes is under the control of the cell division cycle.28),29) As imperfection or failure of endosymbiotic organelle proliferations results in the loss of these organelles from the cell, a series of cycle of organelle DNA replication, DNA segregation and organelle division is strictly regulated in the cells of photosynthetic eukaryotes.30) The cell proliferation mechanism in photosynthetic eukaryotes is driven by an ensemble of concerted functions of multiple proteins, but the whole picture remains unclear. In eukaryotic cells, which contain multiple mitochondria and chloroplasts, the timings of endosymbiotic organelle divisions are random and not synchronized with the host cell division. On the other hand, the unicellular red alga Cyanidioschyzon merolae, which has a simple cell structure with the minimum set of organelles, can highly synchronize its cell division cycle among cells.31) Combined with its simple genome structure, the study of C. merolae enables us to explore the ensemble regulation mechanism of the eukaryotic cell proliferation system.
In this study, to elucidate the complete picture of gene expression during cell proliferation in C. merolae, we conducted genome-wide analysis of translation in non-dividing and dividing cells using ribosome profiling. This technique allows us to understand genome-wide translation activity in a cell at a given time through the deep sequencing of ribosome-protected messenger RNA (mRNA) fragments.32),33) Although general RNA transcriptome techniques illustrate mRNA abundance corresponding to each gene, mRNA levels are an imperfect reflection of protein translation levels due to extensive regulation. In contrast, ribosome profiling enables us to directly monitor protein synthesis in each gene with subcodon resolution.
Using this technique, we unveiled protein translation levels and credible start codons corresponding to each gene encoded in nuclear, mitochondrial, and chloroplast DNAs. Comparison of ribosome footprint profiles in the two states showed that 241 genes are translated in dividing cells at levels more than five-fold higher than in non-dividing cells. Among these, 47% of the genes were classified as proteins of unknown function. In vivo experiments to verify the ribosome profiling results showed that two selected candidate proteins function as a mitochondrial division-phase-specific nucleoid protein, MDN, and a mitochondrial division protein, MDR2. Thus, this study has identified factors that contribute to the cell proliferation mechanism, and further studies of these factors will provide insight into the fundamental principles of how eukaryotic cells proliferate.
The Cyanidioschyzon merolae 10D strain (NIES-3377)31),34) was used for ribosome profiling and for in vivo analyses as the wild-type in this study. Synchronization of cultures was performed as described previously,35) and the synchronized cells at 6 hr into the second light phase and 2 hr into the second dark phase (30 hr and 38 hr from the start of synchronous cultivation, respectively) were used as non-dividing cells and dividing cells, respectively.
Library preparation was performed in accordance with the method described previously36),37) with the following modifications. Cells at 30 and 38 hr after synchronization were rapidly harvested by vacuum filtration using a cellulose nitrate membrane with a pore size of 0.1 μm (Whatman). The cells were immediately scraped from the membrane using a metal spatula and rapidly frozen in liquid nitrogen. The frozen cell pellet was resuspended in lysis buffer [20 mM Tris, pH 7.5, 150 mM NaCl, 5 mM MgCl2, 1 mM dithiothreitol (DTT), 1% (v/v) Nonidet P-40, 100 μg/mL cycloheximide, 100 μg/mL anisomycin] to prepare the whole-cell lysate. Then, cell lysate containing 12 μg of total RNA was treated with 15 units of RNase I (Biosearch Technologies) at 25°C for 45 min. To isolate ribosomes, the RNase-digested cell lysate was centrifuged through a sucrose cushion at 100,000 rpm for 1 h at 4°C in a TLA110 rotor. The ribosomal pellet was resuspended in splitting buffer (20 mM Tris, pH 7.5, 300 mM NaCl, 5 mM EDTA, 1 mM DTT, 1% (v/v) Triton X100, 20 U/mL SUPERase In) in which a low concentration of magnesium causes dissociation of the ribosomal subunits. The suspension was cleaned using a Amicon Ultra Molecular Weight Cut Off 100 kDa centrifugal filter (Merck Millipore) and the flow-through containing RNA was collected. The ribosome footprints (19-34 nt long) were gel purified with a 15% polyacrylamide TBE-Urea gel, dephosphorylated using T4 polynucleotide kinase (New England Biolabs) and ligated to preadenylylated linkers (Table S1) using T4 RNA Ligase 2, truncated K227Q (New England Biolabs). Ribosomal RNA was depleted using the biotinylated rRNA depletion oligos for yeast (Table S2) and streptavidin magnetic beads (New England Biolabs). For reverse transcription, an oligo 5′-(Phos)NNAGATCGGAAGAGCGTCGTGTAGGGAAAGAG(iSp18)GTGACTGGAGTTCAGACGTGTGCTC-3′ was used, where Phos indicates 5′ phosphorylation, iSP18 indicates the 18-atom hexa-ethyleneglycol spacer and N indicates a random nucleotide. Reverse transcription products were circularized with CircLigase II (Biosearch Technologies). PCR was performed with oligonucleotides 5′-AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTC-3′ and 5′-CAAGCAGAAGACGGCATACGAGATJJJJJJGTGACTGGAGTTCAGACGTGTG-3′, where J indicates the reverse complement of the index sequence discovered during Illumina sequencing. The libraries were sequenced on a HiSeq-X (Illumina).
Sequencing reads were de-multiplexed and stripped of 3′ linker sequences using Cutadapt v2.8. UMI, which can remove PCR duplications generated during library preparation, was extracted by UMI-tools v1.0.1.38) The reads were first filtered by mapping to the Bowtie Index, composed of rRNA and tRNA genes of C. merolae, using Bowtie2 v2.2.5.39) Reads were mapped to the C. merolae genome (ASM9120v1) using STAR v2.7.8.40) Only uniquely mapping reads from the final genomic alignment were used for subsequent analysis. The position of the A site from the 5′-end of the reads at the initiation codon was estimated based on the length of each footprint using plastid v0.4.7.41) The mapped read counts were calculated using plastid v0.4.7, with 5′-end assignment for A site positioning of translating ribosomes. The 20-22 nt and 27-29 nt long reads for dividing cells, and the 20-22 nt and 29-31 nt long reads for non-dividing cells were regarded as ribosome-protected mRNA fragments, and the offsets used for 5′-end assignment were 14 for 27 nt, 15 for 20-21 nt and 28-30 nt, and 16 for 22 nt and 31 nt.
To address genome-wide protein translation during cell and organelle proliferation in C. merolae, we conducted ribosome profiling on non-dividing and dividing C. merolae cells (Fig. 1A). Ribosome profiling has not been established for C. merolae cells. Therefore, we adapted the method from yeast Saccharomyces cerevisiae.36) Cells from synchronized cultures were quickly harvested by filtration and lysed using non-ionic detergent in the presence of cycloheximide and anisomycin to preserve the protein translational state in each ribosome. After separating a portion of each sample for RNA sequencing, we treated the samples with RNase and subsequently isolated the ribosomes, thereby obtaining ribosome-protected mRNA fragments from each sample. Because membrane-bound organelles were also lysed by the buffer, mRNA fragments were derived from mitochondrial and chloroplast ribosomes, in addition to cytosolic ribosomes. After a series of molecular biological processes, we finally prepared DNA libraries and performed deep sequencing for these. From 19,411,358 and 15,708,666 footprint sequences for non-dividing cells and 17,767,453 and 17,111,286 footprint sequences for dividing cells, we measured the translation of 4695 nuclear DNA encoded genes, 34 mitochondrial DNA encoded genes, and 207 chloroplast DNA encoded genes. Reproducibility of footprint data sets between technical replicates was reliable (Pearson’s r = 0.981258 for non-dividing cells and 0.99968 for dividing cells) (Fig. 1B and C). The relationship between ribosome footprint densities and the codon adaptation index for each gene42),43) was not significant (Pearson’s r = 0.237631). The highest number of ribosome footprints was detected in CMK024C, which encodes an uncharacterized protein, and the lowest number of ribosome footprints was detected in CMN333C, which encodes phosphatidylinositol-4-phosphate 5-kinase (PIP5K), among the nuclear DNA-encoded genes. Additionally, chloroplast Ycf23, which encodes an uncharacterized protein, and mitochondrial RPS12 had the highest number of ribosome footprints among chloroplast and mitochondrial DNA-encoded genes, respectively. There were no protein-coding genes that were transcribed but translationally inactive. In both non-dividing and dividing cell samples, 28 protein-coding genes encoded by nuclear DNA were not detected at either the transcriptional or translational levels (Supplementary Data 1). Given that some of these non-expressed genes corresponded to well-established proteins, such as ribosomal proteins and the Rieske iron-sulfur protein in the cytochrome bc1 complex, these genes are likely to be pseudogenes, genes expressed only at highly specific times or under particular conditions, and/or the result of insufficient gene structural prediction.
Ribosome profiling in Cyanidioschyzon merolae. (A) A schematic of the ribosome profiling protocol in C. merolae cells. RNAs with ribosomes were isolated from non-dividing and dividing cells. RNA-containing ribosomes were derived not only from the cytosol but also from the mitochondrion and the chloroplast. Through sample preparation, a DNA library for ribosome-protected mRNA fragments derived from the cytosol, mitochondria, and chloroplasts were obtained and analyzed using deep sequencing. See Materials and methods for further details. MD, mitochondrial division machinery; PD, plastid division machinery. (B and C) Ribosome footprint densities of nuclear DNA-encoded genes (B) and organelle DNA-encoded genes (C) in two technical replicates for non-dividing cells.
The comparison of ribosome profiling data for alpha-tubulin mRNA, specifically expressed during the cell division phase in C. merolae, between non-dividing and dividing cells, clearly illustrated differences in ribosome footprint density across cell phases (Fig. 2A and 2B). These results confirmed that ribosome profiling can accurately reflect the translatome landscape in both non-dividing and dividing cells. Moreover, the ribosome footprint density for each gene identified a reliable start codon within the reading frame. Although translation initiation sites provide crucial information for annotating translating reading frames, the identification of start codons typically depends on gene prediction methods, which include inherent uncertainties. As ribosome footprints accumulate along the RNA reading frame, this allowed us to verify the start codons of all nuclear-encoded genes, enabling a comprehensive mapping of protein structures in the C. merolae genome. By analyzing ribosome footprints near putative start codons, we identified more reliable start codons for 161 genes than those previously reported (Supplementary Data 2). As a representative result, we presented a revised reading frame for the plastid division protein PDR126) (Fig. 2C). The newly found start codon is located 128 amino acids downstream from the previous start codon, resulting in a PDR1 protein that has been revised to 555 amino acids in length. We also discovered that 6 protein-coding genes lack a distinct start codon (Supplementary Data 3). As a representative, we showed the ribosome footprint density of CML004C, which encodes a putative iron permease (Fig. 2D). Ribosome footprints were only confirmed in the middle region of their ORFs, and notably, there is no methionine residue in these regions. Furthermore, given that 11 putative transmembrane domains were identified through the putative protein-coding region, the start codon for the iron permease is presumed to be the first amino acid. However, the underlying reason for the scarce ribosome footprints around the start codon remains unclear.
Ribosome footprint densities in several representative protein-coding genes. (A and B) Cell division phase-specific translation in alpha-tubulin. Although the gene of alpha-tubulin (CMT504C, TUBA) is not expressed in non-dividing cells (A), it is specifically expressed in dividing cells (B) in C. merolae. (C) A revised start codon position (arrow) for the plastid division gene PDR1 based on ribosome footprint data. The PDR1 gene length was fixed from 2,049 nt (683 aa) to 1,665 nt (555 aa) (D) A representative for an uncertain start codon position. Methionine positions and putative transmembrane (TM) domains in the gene encoding iron permease-like protein (CML004C) are shown with arrows and brown boxes, respectively.
Through genome-wide measurements of translation, we concluded that ribosome profiling can serve as an index of protein translation activity at a given time in C. merolae. Thus, to reveal the comprehensive translational landscape during cell proliferation, we compared the translatomes of non-dividing and dividing cells using ribosome profiling data (Fig. 3A). As a result, one-quarter of the nuclear-encoded genes showed changes in relative protein translation levels during cell proliferation. Variation in protein translation levels was observed in various genes (Fig. 3B and 3C), even in housekeeping proteins (Fig. 3D). Although the translation of 657 genes increased by more than two-fold (Supplementary Data 4), the translation of 611 genes decreased to less than half (Supplementary Data 5). Among these, 241 genes were strongly upregulated (by more than fivefold) and 104 genes were strongly downregulated (to less than one-fifth). The ribosome footprint data showed that the genome-wide change in protein translation levels during cell proliferation corresponded to rearrangement of the allocation of cytosolic ribosomes (Table 1). In dividing cells, cytosolic ribosomes are allocated to protein synthesis for DNA replication (0.92%), nucleosome assembly (15.74%), chromosome segregation (3.64%), and organelle division (0.80%). Nascent proteins from the 657 upregulated genes in dividing cells were synthesized from 40.77% of available cytosolic ribosomes. Although these numbers provide evidence that the processes of cell and organelle proliferation have an extremely large impact on the cell, they also prompt us to investigate how the cell manages to mobilize such a large portion of cytosolic ribosomes during the proliferation phase. The most significant trade-off was the decreased production of ribosomes themselves (from 14.30% to 10.35%) and chloroplast-targeting proteins (from 19.31% to 12.58%) during the dividing phase. Furthermore, many types of metabolic processes such as glycolysis, amino acid synthesis, and oxidative phosphorylation (OXPHOS) also reduced the allocation of cytosolic ribosomes. It is possible that the dynamic transition of translation levels for ribosomal genes and chloroplast-targeting proteins is related not to cell division cycle-dependent expression patterns, but to light-dark cycle-dependent patterns. To clarify this, we analyzed the time-course gene expression profiles of these genes using C. merolae transcriptome data (Fig. S1, Supplementary Data 6 and 7). As a result of clustering in the transcriptome data, we found that 93 genes encoding chloroplast-targeting proteins showed light-dependent expression patterns, whereas 7 genes encoding ribosomal proteins and 44 genes encoding chloroplast-targeted proteins showed dark-dependent expression patterns. However, dynamic changes in transcription levels for the remaining ribosomal protein genes and chloroplast-targeting protein genes during the light or dark phases indicate that these genes are likely regulated by the cell-division cycle, not the light-dark cycle. Thus, these observations suggest that a genome-wide rearrangement of protein translation levels, achieved through the reallocation of finite intracellular resources, has been performed to facilitate cell and organelle proliferation.
Genome-wide translational changes during cell division. (A) Changes in ribosome footprints for each nuclear-DNA-encoded gene from non-dividing to dividing cells. (B) Comparisons of genes related to cell and organelle division. Histones and ribosomal protein genes are also shown in yellow and green, respectively. (C) Changes in genes with unknown functions. Conserved genes and non-conserved genes are shown in magenta and blue, respectively. (D) Genes with no change in translation level during cell division. Ubiquitin C, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), glucose-6-phosphate dehydrogenase (G6PD), and hypoxanthine phosphoribosyltransferase (HPRT) are shown in black. The gene expression level of the protein-coding gene CMK024C was the highest among all nuclear-DNA-encoded genes in non-dividing cells, and 1.63-1.70% of cytosolic ribosomes were allocated for its protein translation throughout the cell cycle.
Ribosome allocation rate for each functional class during cell division
| Functional classification | Number of genes | Rates in non-dividing cell | Rates in dividing cell |
|---|---|---|---|
| Segregation and division | |||
| DNA segregation | 13 | 2.03 | 3.64 |
| Mitochondrial division | 7 | 0.05 | 0.23 |
| Chloroplast division | 4 | 0.02 | 0.57 |
| Gene expression | 31 | 0.13 | 0.92 |
| DNA replication | 31 | 0.13 | 0.92 |
| RNA transcription | 44 | 0.54 | 0.48 |
| Protein synthesis (ribosome) | 102 | 14.30 | 10.35 |
| Metabolism | |||
| Nucleotide metabolism | 61 | 0.85 | 0.87 |
| Amino acid synthesis | 141 | 4.09 | 2.20 |
| Fatty acid synthesis | 19 | 0.39 | 0.27 |
| Glycolysis | 43 | 2.11 | 1.08 |
| Oxidative phosphorylation | 59 | 2.42 | 1.73 |
| Specific protein group | |||
| Histone | 7 | 4.00 | 15.74 |
| Mitochondrial targeting protein* | 289 | 5.76 | 4.32 |
| Chloroplast targeting protein* | 328 | 19.31 | 12.58 |
Unknown function protein (conserved) |
784 | 11.11 | 9.68 |
Unknown function protein (non-conserved) |
1132 | 20.24 | 20.96 |
* Computationally predicted proteins with a putative mitochondrial or chloroplast targeting peptide58) are shown.
Among the strongly upregulated genes during cell and organelle proliferation, 114 genes were classified as uncharacterized protein-coding genes in the C. merolae genome data. Additionally, homologous genes corresponding to 27 of these genes were found in other organisms. Given their specific gene expression during cell proliferation, these genes were thought to have functions related to the proliferation of cells or organelles. We investigated the spatiotemporal distributions of two candidate proteins, CMO005C and CMT564C, by fluorescence microscopy to confirm the timing of protein expression and to infer their intracellular functions related to cell proliferation. Because these genes were also found in a previous proteomic analysis of isolated mitochondrial division machineries, we assumed that these genes are related to mitochondrial proliferation.
CMO005C encodes 648 amino acids and includes a putative guanosine tetra- and pentaphosphate [(p)ppGpp] hydrolase domain in the mid-region. (p)ppGpp is well known as an intracellular signaling molecule that reprograms bacterial cell physiology, facilitating stress adaptation in harsh environmental conditions.44)-47) Furthermore, (p)ppGpp molecules have been identified in eukaryotic cells, and (p)ppGpp causes metabolic changes linked to the stringent response.48),49) In plants, (p)ppGpp molecules are synthesized in chloroplasts by calcium-activated (p)ppGpp synthetases.50) Genes possibly relating to (p)ppGpp synthesis, RSH4a and RSH4b, have also been identified in C. merolae.51) To assess the function of CMO005C in cell proliferation, we introduced CMO005C protein tagged with the yellow fluorescent protein Venus, driven by the native CMO005C promoter, into C. merolae cells via homologous recombination. Fluorescence microscopy showed that CMO005C-Venus was expressed during mitochondrial division phase and localized in mitochondria (Fig. 4A). Given that fluorescent signals of CMO005C-Venus were detected at two loci during the mitochondrial division phase, even while daughter mitochondria were still connected, it was thought that CMO005C localizes to mitochondrial nucleoids. This was confirmed by fluorescence imaging of isolated mitochondrial nucleoids from cells (Fig. 4B). The fluorescence intensities of CMO005C-Venus and DAPI-stained DNA correlated well (Fig. 4C and 4D), suggesting that CMO005C is a component of mitochondrial nucleoids. Based on these results, CMO005C was named MDN (mitochondrial division-phase-specific nucleoid protein).
A mitochondrial division phase-specific nucleoid protein MDN. (A) Intracellular localization of Venus-fused MDN in non-dividing cells and dividing cells. Mitochondria are visualized with MitoTracker Red. (B) Fluorescence images of isolated mitochondrial nucleoids. Mitochondrial nucleoids were isolated from cells expressing MDN-Venus. (C) Histograms of fluorescence intensity from Venus or DAPI. (D) A scatterplot comparison of Venus fluorescence intensity and DAPI fluorescence intensity in each isolated mitochondrial nucleoid. Scale bars, 2 μm (A) and 1 μm (B).
Using the same approach, we investigated the intracellular function of CMT564C in cell proliferation. Protein structural prediction and homology search suggested that CMT564C encodes 503 amino acids and consists of a coiled-coil domain, and no distinct homologous proteins of CMT564C have been found in other organisms. Venus-fused CMT564C was specifically expressed during the mitochondrial division phase and was localized to the mitochondrial division site (Fig. 5A, left). Comparisons with Mda1 and Dnm1 (Fig. 5A, middle and right), which are components of the mitochondrial division machinery, showed that the protein distributions and behaviors of CMT564C were more similar to those of Mda1 than Dnm1. This was further clarified by fluorescence imaging of cell-cycle arrested cells at the G2-M phase after camptothecin treatment (Fig. 5B). Although Dnm1 localized not only at the mitochondrial division site but also in the cytosolic space, Mda1 and CMT564C specifically localized and accumulated at the mitochondrial division site. Additionally, these fluorescence images inferred that the width of the CMT564C ring was slightly thicker than that of Mda152) and similar to MDR1,27) which is also a component of the mitochondrial division machinery. Based on these results, we concluded that CMT564C is a novel mitochondrial division protein and named it MDR2 (mitochondrial division ring 2). Through in cellulo experiments on two selected genes, we showed that one of them corresponded to a division phase-specific mitochondrial nucleoid protein, whereas the other is a novel component of the mitochondrial division machinery. These findings indicated that the candidate proteins, which were identified by comparisons of ribosome profiling of non-dividing and dividing cells, may be key players in the cell proliferation mechanism in eukaryotic cells.
A mitochondrial division protein MDR2. (A) Intracellular protein distribution of MDR2 in cells according to the mitochondrial division process. The images, taken from different cells, show representative localization of Venus-fused MDR2. The fluorescence images for Mda1 and Dnm1, which are components of the mitochondrial division machinery, are also shown. (B) Protein localization of MDR2, Mda1, and Dnm1 in mitochondrial division-arrested cells by camptothecin treatment. Scale bars, 2 μm.
Resultant translatome maps of non-dividing and dividing C. merolae cells from ribosome profiling unveiled a genome-wide rearrangement of protein synthesis during proliferation. The ribosome footprint results provided information to identify candidate proteins corresponding to cell or organelle divisions. In 241 strongly and specifically upregulated genes in the dividing phase, while 125 genes were well-studied cell and organelle division genes, 114 genes (approximately 47%) were classified as genes with unknown functions in this study. The common translational characteristics suggested that these uncharacterized genes are likely related to the proliferation process, similar to known cell and organelle division proteins. In addition, 3.98% of cytosolic ribosomes are allocated to protein synthesis for the uncharacterized proteins, suggesting that the contribution and functional importance of these genes in cell and organelle division will have an undeniable impact. Indeed, in cellulo analyses for two selected genes showed that these genes were involved in mitochondrial nucleoid and mitochondrial division (Fig. 4 and 5). The protein encoded by CMO005C contains a putative (p)ppGpp hydrolysis domain similar to that of Mesh1,48) although the catalytic His-Asp motif responsible for (p)ppGpp hydrolysis was not conserved (Fig. S2). As (p)ppGpp molecules can suppress the protein synthesis activity of ribosomes, intracellular concentrations of (p)ppGpp affect various physiological aspects, such as stress responses related to amino acid starvation and the rearrangement of metabolism.44)-49) Given the putative function of the (p)ppGpp hydrolase-like domain of MDN, MDN can sense mitochondrial (p)ppGpp concentration; however, the exact function of MDN is still unknown.
4.2. A coiled-coil protein MDR2 is a component for the mitochondrial division machinery.Using these analyses, we also revealed that another candidate gene corresponds to mitochondrial division. MDR2 contains only a coiled-coil region as a putative functional domain and is not conserved in other organisms; therefore, its function in mitochondrial division was unable to be presumed from the sequence information. However, fluorescence microscopy illuminated several functional aspects of MDR2. Fluorescent protein-tagged MDR2 was clearly localized at the mitochondrial division site, and MDR2 proteins appeared from the beginning to the end of mitochondrial division. This protein distribution pattern of MDR2 was like those of Mda1 and MDR1. Given that the WD40-repeat protein Mda1 is a component of the mitochondrial division machinery and the glycosyltransferase protein MDR1 is presumed to synthesize the skeletal structure of the mitochondrial division machinery,27) MDR2 might contribute to the protein-protein interactions of components to assemble the division machinery via its coiled-coil domain. Notably, some coiled-coil domain-containing proteins are identified as mitochondrial division proteins in animals and fungi, such as MFF in humans53),54) and Mdv1 in yeast.55) Except for the dynamin-related protein Dnm1,23),24) other mitochondrial division proteins are not conserved throughout eukaryotes. However, the functional necessity of determining the mitochondrial division site and guiding Dnm1 to the site should be common. Given that MDR2 protein was involved in the proteome of isolated mitochondrial division machinery,27) the idea of a functional role for MDR2 as a connector of other component proteins is strongly supported. Taken together, these in cellulo experiments showed that the 114 uncharacterized components are very likely to be involved in cell or organelle proliferation. Further investigation of the candidate proteins will help illustrate a comprehensive picture of the cell proliferation mechanism in eukaryotic cells.
4.3. Genome-wide rearrangement of protein translation for cell and organelle proliferation in a simple unicellular alga.Through the present study, we revealed comprehensive protein synthesis during cell proliferation and other processes using ribosome profiling. This is the first evidence of a genome-wide translatome dynamics, involving the rearrangement of ribosome allocation, to optimize protein synthesis rates and perform cell proliferation in the simple unicellular alga C. merolae. A series of new findings presented in this study using ribosome footprint data demonstrated that identifying protein synthesis at a given time is essential for understanding biological phenomena. It also raises concerns about biological studies that rely solely on transcriptome data. Given the presence of several genes corresponding to key physiological functions with very high translation level, such as Ubiquitin C (Fig. 3D), RbcL and RbcS (Fig. 1C), it is implied that uncharacterized genes with similarly high translation level, such as CMK024C (Fig. 1B and 3D) and Ycf23 (Fig. 1C), may have crucial, but not yet recognized, roles in cells. In contrast, the fact that there were no protein-coding genes that are transcribed but not translated suggests that newly emerged ORFs, which obtained a start codon and a stop codon via mutation or gene reconstruction, are also very likely to be translated into proteins as long as they are transcribed into mRNA. Indeed, our results showed that approximately 20% of protein synthesis throughout the cell cycle is for genes encoding non-conserved proteins (Table 1), which are likely to be emerging genes evolved by a de novo process.56),57) When proteins are expressed from an emerging gene, they display unpredictable biochemical properties, which can have both positive and negative effects on the cell. Despite these risks, the study suggested that eukaryotic cells tend to prioritize the potential benefits of innovative proteins with new biochemical functions over the disadvantages, such as toxicity from protein aggregation. Therefore, further studies on this permissive protein translation activity may be crucial for understanding the driving forces behind evolution or enhancing the evolvability of eukaryotic cells. Findings from protein translatome dynamics in C. merolae provide insights not only into the mechanisms of cell and organelle proliferation but also into the fundamental principles of unicellular eukaryotes and the diversification of eukaryotes.
This work was supported by PRESTO from the Japan Science and Technology Agency (no. JPMJPR21EE to Y.Ma., no. JPMJPR20EE to Y.Y.); the Human Frontier Science Program Career Development Award (no. CDA00049/2018-C to Y.Y.); Japan Society for the Promotion of Science KAKENHI (no. JP22H02653 to Y.Y.); and the Institute for Fermentation, Osaka (no. L-2020-2-008 to Y.Y.).
Supplementary materials are available at https://doi.org/10.2183/pjab.101.002.
The authors declare no competing interests.
The sequencing data are deposited at NCBI Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE273424. The data that support the findings of this study are available in the main and Supplementary materials.
Conceptualization: Y.Mo., Y.Y.; Methodology: Y.Mo., Y.Ma., T.H., T.I., Y.Y.; Investigation: Y.Mo., Y.Ma., Y.K., Y.Y.; Writing - original draft: Y.Mo., Y.Ma., Y.Y.; Writing - review & editing: Y.Mo., Y.Ma., Y.K., T.H., T.I., Y.Y.; Visualization: Y.Mo., Y.Y.; Supervision: Y.Y.; Project administration: Y.Y.; Funding acquisition: Y.Ma., Y.Y.
Edited by Shigekazu NAGATA, M.J.A.
Correspondence should be addressed to: Y. Yoshida, Department of Biological Sciences, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan (e-mail: yamato.yoshida@bs.s.u-tokyo.ac.jp).
4′,6-diamidino-2-phenylindole
MDmitochondrial division
ORFopen-reading frame
PDplastid division
