The unicellular alga Cyanidioschyzon merolae has a simple cellular structure; each cell has one nucleus, one mitochondrion, one chloroplast and one peroxisome. This simplicity offers unique advantages for investigating organellar proliferation and the cell cycle. Here, we describe CZON-cutter, an engineered clustered, regularly interspaced, short palindromic repeats (CRISPR)/CRISPR-associated nuclease 9 (Cas9) system for simultaneous genome editing and organellar visualization. We engineered a C. merolae strain expressing a nuclear-localized Cas9–Venus nuclease for targeted editing of any locus defined by a single-guide RNA (sgRNA). We then successfully edited the algal genome and visualized the mitochondrion and peroxisome in transformants using fluorescent protein reporters with different excitation wavelengths. Fluorescent protein labeling of organelles in living transformants allows us to validate phenotypes associated with organellar proliferation and the cell cycle, even when the edited gene is essential. Combined with the exceptional biological features of C. merolae, CZON-cutter will be instrumental for investigating cellular and organellar division in a high-throughput manner.
Eukaryotic cells contain membrane-bound organelles, functionally specialized units that respond to intercellular and intracellular signals for homeostasis and provide separate compartments for the biosynthesis of various biochemicals in varied metabolic pathways. Mitochondria and chloroplasts (plastids) are surrounded by a double membrane and are thought to have evolved from endosymbiotic bacteria (Gray, 1992; Martin and Kowallik, 1999; Mereschkowsky, 1905). Because of their evolutionary origin, they have their own genomes and only multiply via binary fission of pre-existing copies according to their own division cycles (Gillham et al., 1994; Kobayashi et al., 2011; Kuroiwa et al., 1998; Suzuki et al., 1994). Whether (and how) the cellular, nuclear, mitochondrial and plastid division cycles cooperate in the cells of photosynthetic eukaryotes is still largely unknown.
To fill this gap in our knowledge, researchers have recently focused on the simple unicellular alga Cyanidioschyzon merolae, commonly called ‘CZON’ (Matsuzaki et al., 2004; Nozaki et al., 2007). C. merolae cells possess very few membrane-bound organelles; each cell has one nucleus, one mitochondrion, one plastid, one peroxisome, one Golgi body with two cisternae, a few vacuoles and a simple-shaped endoplasmic reticulum (ER). All organelles divide shortly before cell division and are then inherited by the daughter cells (Imoto et al., 2010). Progression through the cell cycle can also be highly synchronized by entraining cultures to light–dark cycles (Suzuki et al., 1994). Together with its simple cell structure, these features provide C. merolae cells with unique advantages for investigating the molecular mechanisms related to organellar and cellular proliferation, as well as the cell cycle.
A series of studies using C. merolae have identified key factors for mitochondrial division [mitochondrial FtsZ, dynamin 1 (DNM1) and MITOCHONDRION-DIVIDING RING1 (MDR1)] and for plastid division [plastid FtsZ, dynamin 2 (DNM2), and PLASTID-DIVIDING RING1 (PDR1)] (Miyagishima et al., 2003; Nishida et al., 2003; Takahara et al., 2000; Yoshida et al., 2010, 2017). A subgroup of the kinesin super family participates in chromosome segregation (Yoshida et al., 2013). Furthermore, a recent transcriptome analysis using synchronized C. merolae cultures identified 454 genes with an expression pattern driven by the cell cycle, of which 181 genes have no known function (Fujiwara et al., 2020). These unknown genes might contribute to organellar and cellular division.
Gene targeting approaches are indispensable and powerful tools to study the function of a gene. An optimized gene targeting technique has been established for C. merolae using homologous recombination (HR) and a positive selection method for uracil auxotrophy or chloramphenicol resistance (Fujiwara et al., 2013, 2017; Imamura et al., 2009; Minoda et al., 2004). However, HR-mediated deletion of a locus typically entails assembling a targeting construct consisting of 1-kb fragments specific for the target gene flanking the URA5.3 selection marker before introduction into C. merolae cells (Fig. 1A). Even though C. merolae has a very small genome (16.5 Mb for the nuclear genome, encoding 4775 proteins; Matsuzaki et al., 2004; Nozaki et al., 2007), the time-consuming and labor-intensive process necessary for gene targeting has severely impeded genome-wide genetic analysis of C. merolae. The multi-step process of conventional gene targeting thus restricts investigation and is not efficient. Furthermore, organellar morphology should be characterized by live microscopy observations using fluorescently labeled organelles in the resulting knockout cells to determine whether the targeted gene is involved in cellular and/or organellar division. Therefore, a simpler and higher-throughput method to inactivate gene function in C. merolae is needed.
Targeted genome editing via clustered, regularly interspaced, short palindromic repeats (CRISPR) and CRISPR-associated nuclease 9 (Cas9) has revolutionized genetic analysis (Adli, 2018; Jiang and Doudna, 2017; Sander and Joung, 2014). The CRISPR-Cas9 system relies on the formation of a complex between Cas9 (most often from Streptococcus pyogenes) and a single-guide RNA (sgRNA) whose sequence is complementary to 20 nucleotides (nt) of a sequence within the target gene, followed by a short DNA motif [protospacer-adjacent motif (PAM) sequence; NGG] (Nishimasu et al., 2014) that cleaves the genome at the target site. The resulting double-strand break can then be repaired via nonhomologous end-joining (NHEJ) or homology-directed repair (HDR) pathways in nearly all organisms. NHEJ-mediated repair frequently introduces insertion/deletion mutations (indels) of various lengths, which disrupt the open reading frame. When provided with exogenously supplied donor DNA templates, the HDR machinery can modify the genome by introducing specific point mutations or inserting desired sequences at the target site. The CRISPR-Cas9 system has therefore become a powerful tool for targeted genome editing in many established and emerging model organisms.
Here, to identify genes that are involved in organellar and cellular division, we describe an engineered CRISPR-Cas9 system for C. merolae, named CZON-cutter, that allows simultaneous site-selective genome editing and multiplexed organellar imaging. This CRISPR-based gene targeting system can be used in the engineered C. merolae strain YMT1, which expresses Cas9-Venus, upon transformation with a PCR amplicon containing a designed sgRNA (Fig. 1B). Gene specificity can be achieved by altering the 20-nt target sequence of the sgRNA in the plasmid. Furthermore, the nucleus, the single mitochondrion, and the single peroxisome of transformed cells can be visualized by fluorescent protein reporters with different excitation wavelengths. The ability to image living transformants makes it possible to validate phenotypes associated with organellar morphology swiftly and accurately without chemical staining in the context of the cell cycle by time-lapse microscopy. The CZON-cutter platform thus holds great promise as an efficient, versatile, and high-throughput approach to investigate the biological function of any gene at organellar resolution.
Engineering of the C. merolae YMT1 strain expressing Cas9-Venus and a universal plasmid DNA template containing a sgRNA and a mitochondrial reporter
To implement the CZON-cutter platform in C. merolae, we first fused the coding sequence for Cas9 from S. pyogenes to two copies of a nuclear localization sequence (NLS) and the yellow fluorescent protein gene Venus. Because the resulting NLS-Cas9-NLS-Venus construct is driven by the constitutive promoter from ELONGATION FACTOR 1 ALPHA (EF1A), Venus-tagged Cas9 (Cas9-Venus) should accumulate in the nucleus throughout the cell cycle. We then integrated a PCR amplicon consisting of the Cas9-Venus gene and a sequence encoding plastid-targeted chloramphenicol acetyltransferase (CAT) as a selectable marker into the region upstream of URA5.3 in the C. merolae uracil-auxotrophic M4 strain by HR (Fig. 1C). The upstream region of URA5.3 is one of the ‘safe harbor’ sites in the C. merolae genome, where integrations do not negatively affect the cell (Fujiwara et al., 2017; Kuroiwa et al., 2017). We selected transformants for resistance to chloramphenicol and named the resulting uracil-auxotrophic and chloramphenicol-resistant strain YMT1 (Fig. 1D). We confirmed that the YMT1 strain contains an inserted copy of the PCR amplicon and has a growth rate similar to that of the wild-type and M4 strains (Fig. 1E). Fluorescence microscopy of YMT1 cells specifically detected Venus fluorescence in the nucleus of dividing and non-dividing cells (Fig. 1F). We thus concluded that Cas9-Venus accumulates in the nucleus throughout the cell cycle.
Next, to edit the genome and visualize mitochondria simultaneously, we generated another construct with a sgRNA and a fluorescent mitochondrial marker in the region downstream of Cas9-Venus by HR. For greater versatility, we constructed a universal plasmid DNA template, pGuide-mitoScarlet, containing two homologous regions and gene cassettes for the sgRNA, the mitochondrion-targeted mScarlet marker gene (mitoScarlet) and the URA5.3 selection marker (Fig. 2A). The sgRNA consisted of three segments: a 20-nt target-specific complementary region, a 76-nt scaffold region, and a 6-nt transcription termination signal (Fig. 2B). We placed the expression of the sgRNA under the control of a 593-bp promoter fragment from the C. merolae non-coding small nuclear RNA U6, which should be recognized by RNA polymerase III and thus be highly expressed in C. merolae cells. The cassette encoding the mitochondrion-targeted fluorescent protein comprised the coding sequence for the red fluorescent protein mitoScarlet fused to a mitochondrial targeting sequence (MTS) driven by the constitutive CpcC promoter derived from the 5′ untranslated region (UTR) of CpcC encoding phycocyanin-associated rod linker protein, followed by the TUBB (encoding β-Tubulin) 3′ UTR. This cassette was inserted into the plasmid downstream of Venus and upstream of URA5.3.
CRISPR-based genome editing in C. merolae is performed by the HDR pathway
To validate the CRISPR-Cas9 gene editing system described above with the YMT1 strain and the pGuide-mitoScarlet plasmid, we selected the putative cryptochrome gene CRY (also called PHR1, C. merolae Genome Project accession number, CMO348C) as a target for genome editing. Cryptochromes in the fruit fly Drosophila melanogaster and other insects act as blue light circadian photoreceptors and, in animals, they act as an integral component of the circadian machinery (Chaves et al., 2011; Öztürk et al., 2007), so we hypothesized that inactivation of C. merolae CRY, which clustered with animal and insect cryptochromes (Fig. S1), might disturb synchronization of cell division under light–dark cycles. Although the proteins encoded by the C. merolae CRY gene family exhibit photolyase activity, which can repair ultraviolet radiation-induced DNA damage, it remains unclear whether C. merolae CRYs behave as blue light photoreceptors (Asimgil and Kavakli, 2012).
We designed the sgRNA target sequence against CRY with the web tool CRISPRdirect (https://crispr.dbcls.jp/) (Naito et al., 2015) by selecting a unique target sequence with a very low possibility of off-target effects. We then attempted to integrate a PCR amplicon, using the resulting pGuide-CRY232-254-mitoScarlet as template, into the genome of YMT1 cells by HR, but we failed to obtain a single positive genome-edited colony. This result suggests that C. merolae might not employ NHEJ to repair DNA double-strand breaks. In fact, the main components of the NHEJ pathway, Ku70 and Ku80 (Fell and Schild-Poulter, 2015), appear to be missing from the C. merolae genome.
Therefore, we next tested Cas9-mediated genome editing via the HDR pathway. We used single-stranded oligodeoxynucleotides (ssODNs) as a donor template, consisting of two homology arms of 30 nt each flanking the 20-nt CRY target sequence, which contained three single-base substitutions, one introducing a stop codon (CRY R79*) and the other two eliminating the PAM sequence in the CRY target region (Fig. 2C). We then transfected the YMT1 strain with a PCR amplicon corresponding to pGuide-mitoScarlet-CRY together with the ssODN to modify the target region of the CRY gene. We obtained 16 colonies by the uracil-autotrophic selection and 13 positive clones by sequencing (∼11.1 transformants per 108 cells). As each positive clone contained several nucleotide substitutions in the CRY target region, we verified that Cas9-Venus-mediated mutations occurred in the target region of the 13 positive clones to evaluate accuracy and variation caused by unexpected mutations during genome editing (Fig. 2D–G). Ten clones (76.9%) harbored the desired mutations (Fig. 2D). The remaining three clones (23.1%) had imperfect and/or unexpected mutations at the CRY target site (Fig. 2E–G). Furthermore, we detected red fluorescence in mitochondria from the accumulation of mScarlet and green fluorescence from Cas9-Venus in all 13 colonies by fluorescence microscopy (Fig. 3A). These results demonstrated successful genome editing and simultaneous visualization of the nucleus and mitochondrion in C. merolae; we named this CRISPR-based genome editing system CZON-cutter.
Phenotypic validation of the CRISPR-mediated CRY knockout strain by synchronization cultivation under distinct light conditions
To investigate whether C. merolae CRY affects cell cycle progression, we next tested the synchronization of CRISPR-generated CRY loss-of-function mutants (cry) under light–dark cycles using one strain as a representative. For this purpose, we selected the wild-type 10D strain rather than the uracil-auxotrophic YMT1 strain for control experiments to compare growth kinetics using culture conditions in non-uracil-supplemented medium. We cultured one cry strain and the wild-type 10D strain under light–dark cycles with white light (photon flux of 30 μmol m−2 s−1). Given that about half (53.8% in this study) of the wild-type 10D cells are in the dividing phase 38 h after the initiation of synchronization, we compared the percentage of dividing cells in the knockout strain to that of the wild-type strain at this time point (Fig. 3B left; Table S1). Fewer cells appeared to be dividing in the cry strain (28.2%, P<0.01) relative to the wild-type strain, suggesting that inactivation of CRY may alter cell cycle progression, possibly by disrupting circadian rhythms in C. merolae.
Using the cry strain, we next explored the effect of light quality on entrainment of the circadian clock in C. merolae. We cultured cry and the wild-type strain under light–dark cycles consisting of a combination of blue and red light irradiation (Fig. 3B, middle; Table S1). To maintain photosynthetic activity, we adjusted the total photon flux to provide ∼30 μmol m−2 s−1 (blue, 15.4 μmol m−2 s−1, red, 15.4 μmol m−2 s−1). We observed that wild-type cells exhibit a comparable synchronization rate to that seen in white light, with a dividing rate of 53.6%. The percentage of dividing cells in the cry strain was significantly lower at 42.3% (P<0.01), although it was higher than that of the cry strain grown in white light.
As both the wild-type and the cry strains hardly proliferated under pure blue light, we could not confirm the synchronization properties of these strains under pure blue light conditions. Therefore, we also evaluated whether C. merolae cell division was entrained by blue light by culturing cells under pure red light (Fig. 3B, right; Table S1). Under these conditions, the percentages of dividing cells in the wild-type and cry strains were identical at 22.4% and 22.5%, respectively, indicating that the cell cycle of C. merolae is entrained by blue light irradiation. These findings suggest that CRY is involved in entrainment of the cell cycle by blue light in C. merolae. Although further experimental corroboration is required, these preliminary results suggest that C. merolae CRY may contribute to circadian entrainment, as the cell cycle is gated by the circadian clock (Yang et al., 2010).
Identification of putative components in the regulation mechanism of circadian rhythms in C. merolae by genome-wide transcriptome analysis using the cry strain
To explore the molecular links between cellular and organellar division and circadian rhythms in C. merolae, we compared gene expression profiles of the wild-type and the cry strains under continuous white light conditions by transcriptome deep sequencing (RNA-seq) (Fig. 3C). We identified the 50 most highly upregulated genes and the 50 most strongly downregulated genes between the two strains using a dataset from a single replicate (Table S2). CRY itself was among the upregulated genes in the cry strain, suggesting that an uncharacterized transcriptional activator may induce CRY expression to compensate for its loss. The expression level of the gene encoding the CRY regulatory protein, CONSTITUTIVE PHOTOMORPHOGENIC1 (COP1), was lower in the cry strain. COP1 was originally identified as a key factor repressing light-mediated development and growth in plants by degrading photomorphogenic transcription factors in the dark (Lau and Deng, 2012). In addition, CRYs antagonize COP1 activity to regulate circadian rhythms in mammals (Rizzini et al., 2019). As the loss of CRY function affected cell synchronization and the expression levels of COP1, C. merolae CRY may also play a pivotal role in the regulation of circadian rhythms in this alga.
RNA-seq analysis indicated that 40% of the differentially expressed genes have unknown biological functions in the C. merolae genome database, hinting that some of these genes might act as nodes linking the circadian clock with the cellular and organellar division cycle. We independently validated the expression pattern of two of the most highly upregulated genes in our dataset, CMB087C and CMB081C, over a diurnal time course, with a peak in expression in the middle of the light period (CMB087C) or dark period (CMB081C), respectively (Fig. 3D). This diurnal pattern remained unchanged in knockout strains of either the CYCLIN-DEPENDENT KINASE A (CDKA) or the RETINOBLASTOMA-RELATED (RBR) genes. Thus, further exploration of the function of these genes may reveal the regulatory system underlying cellular/organellar division and circadian rhythms.
Gene cassette knock-in for knocking out a target gene and fluorescent protein labeling of the peroxisome by CZON-cutter
We next used CZON-cutter to insert a sequence of interest into the genome (knock-in) while also knocking out a target gene and marking the single peroxisome with another fluorescent protein. We selected ACTIN as a target for gene knock-in. Although the C. merolae genome contains a single ACTIN gene (Matsuzaki et al., 2004; Takahashi et al., 1995), cytokinesis in C. merolae does not involve actin (Yagisawa et al., 2020). As actin plays a central role in cell division in contemporary eukaryotes, direct evidence of actin-free cell division in C. merolae would illustrate a new framework for cytokinesis in simpler eukaryotes. To generate a construct for CRISPR-mediated gene knock-in, we modified the sgRNA cloned into the pGuide-mitoScarlet plasmid to target ACTIN. We also constructed a plasmid that would act as template for HDR, pCer3-PTS1, in which the constitutive ApcC promoter derived from the 5′ UTR of ApcC encoding allophycocyanin-associated rod linker protein drives the expression of the gene encoding cyan fluorescence protein mCerulean3 fused to a peroxisome targeting signal (mCerulean3-PTS1, hereafter perCerulean3), followed by the TUBB 3′ UTR (Fig. 4A). Target site specificity was provided by the PCR primers, which included 50-nt overhangs with homology to the target locus of interest (Fig. 4B,C). We transformed the YMT1 strain with a pool of these two PCR products (one derived from pGuide-mitoScarlet-ACTIN and one from pCer3-PTS1) and obtained three uracil-autotrophic colonies expressing both mitoScarlet and perCerulean3 (∼2.6 transformants per 108 cells).
We chose one ACTIN knockout (actin) strain as a representative for further analysis and confirmed the presence of perCerulean3 at the ACTIN locus (Fig. 4D; Fig. S2). We then used fluorescence microscopy to image the peroxisome, nucleus, mitochondrion and chloroplast in the same cell. We noticed no significant differences in the growth curves or cell shapes of the wild-type and actin strains over the cell cycle (Fig. 4E,F), and observations via fluorescence microscopy demonstrated that cells lacking actin undergo cytokinesis successfully and divide into two daughter cells, each with its own nucleus, mitochondrion, chloroplast, and peroxisome (Fig. 4G). We conclude that actin is not required for cell growth or proliferation in C. merolae. In addition, we confirmed that the perCerulean3 cassette can be inserted into the CRY locus using the same approach, resulting in transformants that accumulate the fluorescent reporters as seen in the actin strain (Fig. S3), suggesting that the CZON-cutter platform can insert the perCerulean3 cassette with 50-nt short homology arms into any locus of interest.
Monitoring living transformants to evaluate the effect of knocking out essential genes
Next, we assessed the applicability of the CZON-cutter platform to test the potential contribution of a given gene during organellar and cellular division, even when the gene is essential for growth and/or proliferation. For this purpose, we modified the sgRNA in the pGuide-mitoScarlet plasmid to target MDR1 or TUBG (encoding γ-Tubulin). MDR1 participates in the assembly of the mitochondrion-dividing (MD) ring, the molecular machinery that severs the mitochondrion for mitochondrial proliferation (Yoshida et al., 2017), whereas γ-Tubulin plays a critical role in spindle assembly and chromosome segregation during M phase (Wiese and Zheng, 2006). We predicted that inactivation of MDR1 or γ-Tubulin would result in severe, abnormal mitochondrial division and cell division phenotypes.
To evaluate the effect of knocking out these genes, we first investigated the morphology of transformants 2 days after transformation with PCR amplicons for the sgRNA targeting MDR1 and the perCerulean3 fluorescence reporter. We identified cells with fluorescence from mScarlet in the mitochondrion and mCerulean3 in the peroxisome, which comprised ∼0.1% of all cells, indicative of our transformation efficiency (Fig. 5A). Since we detected fluorescence from both reporters (mitoScarlet for sgRNA and perCerulean3 for gene knock-in) by fluorescence microscopy, we hypothesized that the perCerulean3 cassette is integrated at the MDR1 locus, thereby knocking it out.
A morphological examination of transformants showed an obstruction of mitochondrial division and mitochondrion overgrowth. Furthermore, we also noticed nuclear division defects, peroxisomal division defects, and an enlargement of cells. The fluorescence derived from Cas9-Venus allowed us to identify the nucleus in each cell; together with the presence of an overgrown mitochondrion, these observations suggested that the nuclear envelope is intact and that the transformant is arrested in prophase. The observed phenotype of mitochondrial division defects is very similar to that previously reported for antisense suppression of MDR1 (Yoshida et al., 2017), suggesting that the gene knockout approach employing CZON-cutter will be an effective method for studying genes related to organellar and cellular division.
Finally, we knocked out TUBG and characterized the resulting phenotypes with the same strategy. Two days after transformation, we identified cells exhibiting mitoScarlet and perCerulean3 fluorescence at a similar frequency (∼0.1%) as with MDR1 (Fig. 5B). Although we detected strong fluorescence for Cas9-Venus in the putative nuclear region, we also observed some fluorescence in the cytosol, indicating that the nuclear envelope is partially disassembled and that the transformant is arrested in prometaphase or metaphase. In addition, the chloroplast and mitochondrion had already undergone division, while we observed a single peroxisome in one of the two dividing cells. The results indicate that the division of each organelle – chloroplast, mitochondrion, and peroxisome – occurs at different times during the cell cycle (Fig. 5C). The CZON-cutter platform will be instrumental in exploring the function of genes involved in organellar and cellular division, even if the gene is essential.
CRISPR-mediated genome editing with multiplexed organelle visualization in C. merolae
The unique cell structure and availability of the complete genome sequence of C. merolae offers a new avenue to study intracellular mechanisms of eukaryotic cells. Here, we describe an efficient and accurate gene-targeting method to take advantage of the unique features of this unicellular alga. Conventional HR techniques in C. merolae involve the introduction of a linear DNA fragment consisting of a gene cassette and selection marker flanked by over 1 kb of sequence from the locus of interest on either side (Fujiwara et al., 2017). The generation of a transformation construct therefore entails multiple steps, and the resulting plasmid is rather large and complex (Fig. 1A). Although C. merolae boasts the smallest genome of all photosynthetic eukaryotes, cumbersome gene targeting has hindered genome-wide analyses of gene function in this alga. Here, we established the CRISPR-mediated genome-editing platform CZON-cutter, which can disrupt or modify any gene in the genome with the added benefit of multiplexed visualization of organelles in a simple procedure (Fig. 1B; Figs S4 and S5). The engineered strain used here, YMT1, showed no abnormal growth phenotype (Fig. 1E). Given the small size of the C. merolae genome, potential off-target effects by CZON-cutter should be minimal. Furthermore, multiplexed visualization of organelles with bright fluorescence proteins allows the prompt characterization of the phenotypes associated with a gene knockout in terms of organellar division and morphology.
The CZON-cutter platform provides an efficient, versatile and high-throughput approach to investigate the biological function of any gene at organellar resolution
Mitochondria and chloroplasts proliferate via binary division of preexisting organelles by using specialized ring-shaped complexes called mitochondrial- and plastid-division machineries (Yoshida and Mogi, 2019; Yoshida et al., 2010, 2017). Peroxisomes divide by using a homologous division machinery (Imoto et al., 2013). However, the individual components and the underlying regulatory mechanisms of these division machineries are largely unknown. A recent transcriptome study identified 181 uncharacterized genes that are specifically expressed during mitochondrial and chloroplast division (Fujiwara et al., 2020), which may be candidate genes for exploring the mechanisms behind organellar division.
In this study, we generated a knockout strain for CRY with CZON-cutter, which revealed that CRY may be a blue light photoreceptor with critical roles in the regulation of cell cycle progression in C. merolae. CZON-cutter also paves the way for studies of the physiological and molecular mechanisms of the cell cycle. Furthermore, the successful knockout of MDR1 and TUBG also demonstrated that CZON-cutter makes it possible to evaluate the function of essential genes (Fig. 5). In addition, our results hint at the links between organellar division and cellular division. A defect in mitochondrial division caused by genome editing of MDR1 arrested cell cycle progression at prophase, indicating that mitochondrial division may participate in a prophase–prometaphase checkpoint. Another notable illustration of the power of CZON-cutter is evident in the gene knockout of TUBG, which not only blocked chromosome segregation but also caused a peroxisomal division defect. These results indicate that chloroplast, mitochondrial and peroxisomal divisions are executed in a specific sequence during the cell cycle and that completion of each organellar division is likely to serve as a checkpoint in the cell cycle. Further studies will aim to explore the molecular mechanisms of this highly interdependent system between division of organelles and the cell. In conclusion, CZON-cutter enables the systematic analysis of gene function by genome editing and multiplexed organellar visualization. The originality and versatility of CZON-cutter will help accelerate gene discovery to reveal biological principles of the cell.
MATERIALS AND METHODS
C. merolae YMT1 strain
To create a C. merolae strain constitutively accumulating nuclear-localized Cas9 fused to Venus, a construct containing Cas9-Venus and chloroplast-targeted CAT was assembled as follows. The EF1A promoter (1 kb) was PCR amplified using genomic DNA from C. merolae strain 10D as template. A version of S. pyogenes Cas9, codon-optimized for mammalian expression and fused to the sequence of three copies of the FLAG tag and two copies of a nuclear localization signal from simian virus 40 (SV40), was PCR-amplified from plasmid pKIR1.1 (Addgene #85758). Venus codon-optimized for expression in C. merolae (Nagai et al., 2002) and fused to the sequence of three copies of the HA tag, the UBQ3 (UBIQUITIN 3) 3′ UTR, the ApcC promoter and CAT fused to a chloroplast transit peptide (CTP) derived from ApcC, and the TUBB (β-Tubulin) 3′ UTR were amplified by PCR using plasmid pVenus-CAT (gift from Takayuki Fujiwara, National Institute of Genetics, Japan) as template. The resulting PCR amplicons were cloned into a linear vector amplified by PCR using pNIRp::sfGFP (gift from T. Fujiwara) (Fujiwara et al., 2015) as template. The resulting construct, pCas9-Venus-CAT, contained the gene cassettes Cas9-Venus and CTP-CAT ∼0.9 kb upstream of URA5.3. For the construction of pCas9-Venus-CAT, PCR amplification and assembly of DNA fragments were performed with Platinum SuperFi II DNA polymerase (Thermo Fisher Scientific) and the NEBuilder HiFi DNA assembly cloning kit (New England Biolabs), respectively. Finally, using pCas9-Venus-CAT as template, a PCR amplicon was amplified that contained ∼1400 bp of sequence upstream of URA5.3 (from −2300 to −898 bp), the EF1A promoter, Cas9-Venus, the UBQ3 3′ UTR, the ApcC promoter, CTP-CAT, the TUBB 3′ UTR and ∼900 bp of sequence upstream of URA5.3 (−897 to −1 bp). The PCR amplicon was introduced into the uracil-auxotrophic mutant M4 cells by polyethylene glycol (PEG)–mediated transformation as described by Ohnuma et al. (2008), Imamura et al. (2009) and Fujiwara et al. (2015). Chloramphenicol-resistant transformants were selected by cultivation on chloramphenicol-containing medium at 150 μg ml−1 for 15 days as described by Fujiwara et al. (2017). After chloramphenicol selection, transformants were spread on modified Allen's (MA) medium supplemented with uracil (0.5 mg ml−1) to isolate single colonies. Accumulation of Cas9-Venus was examined by fluorescence microscopy, and the presence of the introduced DNA was confirmed by direct colony PCR and sequencing. The full amino acid sequence of Cas9-Venus is given in Table S5.
Construction of sgRNA expression vectors
To generate the sgRNA expression vector (pGuide-mitoScarlet), the C. merolae U6 promoter and a sgRNA scaffold with a termination signal were synthesized. The synthetic DNA fragments were assembled with a linear vector DNA containing 2300 bp of URA5.3 upstream sequence and the URA5.3 open reading frame, PCR amplified using pNIRp::sfGFP (Fujiwara et al., 2015) as a template. Then, the UBQ3 3′ UTR, the CpcC promoter, the MTS derived from mitochondrial EF-Tu, mScarlet (Bindels et al., 2016) codon-optimized for C. merolae expression and the TUBB 3′ UTR were introduced into the vector. Finally, the URA5.3 upstream sequence (from −2300 to −898 bp) was replaced by Venus and the UBQ3 3′ UTR to create a HR site to introduce constructs into the YMT1 strain. Two types of procedures for construction of the pGuide-mitoScarlet plasmid targeting a target locus are shown in Fig. S4. Potential Cas9 target sites in the CRY, ACTIN, MDR1 and TUBG loci were searched using the CRISPRdirect web server (http://crispr.dbcls.jp/) using the genome of C. merolae ASM9120v1 (Table S3). To create the sgRNA expression vector targeting CRY (pGuide-CRY232-254-mitoScarlet), a 497-bp synthesized DNA fragment containing a sgRNA spacer sequence targeting nucleotides 232–254 of the CRY locus was assembled with a linear vector amplified by PCR with primer sets #1 and #2 using pGuide-mitoScarlet as the template. Similarly, the sgRNA expression vectors targeting ACTIN (pGuide-ACTIN239-261-mitoScarlet) and MDR1 (pGuide-MDR1254-276-mitoScarlet) were constructed from synthesized DNA fragments containing a sgRNA spacer sequence targeting the ACTIN or MDR1 locus, respectively. The complete nucleotide sequences of synthetic DNAs for sgRNAs targeting the CRY, ACTIN, and MDR1 loci are given in Table S4 as #16 to #18. To create the sgRNA expression vector targeting TUBG (pGuide-TUBG863-885-mitoScarlet), a 62-bp ssODN (#9) containing a sgRNA spacer sequence targeting nucleotides 863–885 bp of the TUBG locus was assembled with a linear vector amplified by PCR with primer sets #10 and #11 using pGuide-mitoScarlet as template. One guanine nucleobase was added at the 5′ side of the target sequence for CRY and TUBG to stimulate transcription of sgRNA in vivo.
The plasmid containing peroxisome-targeted mCerulean3 codon-optimized for C. merolae expression was constructed as follows. A synthesized DNA fragment containing mCerulean3 was cloned into a vector pUC57. mCerulean3 (Markwardt et al., 2011) fused to peroxisomal targeting signal 1 (PTS1) was amplified by PCR with primer sets #19 and #20 using the above plasmid as template to generate peroxisome-targeted mCerulean3 (perCerulean3). The ApcC promoter, perCerulean3, the TUBB 3′ UTR and the pUC57 vector were amplified by PCR using primer sets #19 to #26 and assembled into a new vector named pCer3-PTS1. The full nucleotide sequence of perCerulean3 is given in Table S6.
For all plasmids, PCR amplification and assembly of DNA fragments were performed using Platinum SuperFi II DNA polymerase (Thermo Fisher Scientific) and the NEBuilder HiFi DNA assembly cloning kit (New England Biolabs), respectively. Primer sequences are given in Table S4.
Genome editing experiments
For genome editing of the CRY locus, a PCR amplicon was amplified from pGuide-CRY232-254-mitoScarlet as template and contained the sgRNA targeting nucleotides 232–254 of the CRY locus, mitochondrion-targeted red fluorescent protein mScarlet (mitoScarlet) and the selection marker URA5.3, flanked by HR sites. The PCR amplicon was then mixed with an 80-nt ssODN (Fig. 2C) consisting of a 20-nt target sequence and two 30-nt flanking upstream and downstream sequences of the target sequence. The mixture was then transformed into YMT1 cells as described by Fujiwara et al. (2015). For gene knock-in at the target locus, PCR amplicons were amplified with primer sets #3 and #4 from pGuide-ACTIN239-261-mitoScarlet, pGuide-MDR1254-276-mitoScarlet, or pGuide-TUBG863-885-mitoScarlet as templates; each PCR product consisted of the respective sgRNA, mitoScarlet and URA5.3 flanked with HR sites. In addition, the perCerulean3 cassette was PCR amplified with short homology arms with primer sets #5 and #6 (for ACTIN), #27 and #28 (for CRY), #12 and #13 (for MDR1), or #14 and #15 (for TUBG) using pCer3-PTS1 as template. The resulting PCR amplicons (from pGuide-mitoScarlet and pCer3-PTS1) were then mixed and introduced into YMT1 cells as described by Fujiwara et al. (2015).
Fluorescence observations were conducted on an Olympus IX83 inverted microscope with a 1.45 NA, 100× oil immersion objective. Illumination was provided by a fluorescent light source (U-HGLGPS; Olympus), and the samples were observed through excitation filters [490-500HQ (Olympus) for Venus, FF01-549/12-25 (Semrock) for mScarlet, FF01-427/10-25 (Semrock) for mCerulean3, and FF01-405/10-25 (Semrock) for chloroplasts], custom dichroic mirrors [Di03-R514-t1-25×36 (Semrock) for Venus, Di03-R561-t1-25×36 (Semrock) for mScarlet, FF458-Di02-25×36 (Semrock) for mCerulean3, and T455lp (Chroma) for chloroplasts], and emission filters [FF02-531/22-25 (Semrock) for Venus, FF02-585/29-25 for mScarlet (Semrock), FF01-474/27-25 for mCerulean3, and FF02-617/73-25 for chloroplasts (Semrock)]. Images were acquired with a Zyla 4.2 sCMOS camera (Andor) controlled by MetaMorph software (Molecular Devices). The effective pixel size was 65.2 nm×65.2 nm.
Cell cultures and synchronization cultivation
C. merolae 10D strain cells (NIES-3377) were used as the wild type in this study. Wild-type cells and CRISPR-generated mutants were maintained on 2× Allen's medium (Allen, 1959). The uracil-auxotrophic M4 strain and the uracil-auxotrophic/chloramphenicol-resistant YMT1 strain were maintained on MA2 medium (Ohnuma et al., 2008) supplemented with uracil (0.5 mg ml−1) and 5-fluoroorotic acid monohydrate (0.8 mg ml−1). All strains were cultured in flasks with agitation at 120 rpm under continuous white light (22 μmol m−2 s−1) at 38°C.
For synchronization of cultivation, cells were subcultured to <107 cells ml−1 in a 100-ml flask and bubbled with filtered clean and humidified air through a tube connected to an aquarium pump. Cells were then incubated under a 12-h-light–12-h-dark cycle at 42°C in a SLI-700 incubator (EYELA). Blue and red light irradiation were supplied using 470 nm±20 nm LEDs (ISL-150×150-HBB; CCS Inc.) and 660 nm±20 nm LEDs (ISL-150×150-RR; CCS Inc.), respectively.
The GenBank accession numbers used were as follows: Cyanidioschyzon merolae CRY/PHR1 (XP_005537706), C. merolae PHR2 (BAM80259), C. merolae PHR3 (BAM80957), C. merolae PHR4 (BAM79915), C. merolae PHR5 (BAM78760), C. merolae PHR7 (BAM82280), Arabidopsis thaliana CRY1 (AEE82696), A. thaliana CRY2 (AEE27692), A. thaliana CRY-DASH (Q84KJ5), Homo sapiens CRY1 (NP_004066), H. sapiens CRY2 (Q49AN0), Mus musculus CRY1 (AAD39548), M. musculus CRY2 (AAD46561), Gallus gallus CRY1 (AAK61385), G. gallus CRY2 (AAK61386), G. gallus CRY4 (NP_001034685), Drosophila melanogaster CRY (NP_732407), D. melanogaster 6-4 photolyase (BAA12067), Escherichia coli DNA photolyase (WP_062883603), and Synechocystis sp. PCC 6803 DNA photolyase (Q55081).
The phylogenic tree was generated with MEGA X (Kumar et al., 2018). Amino acid sequences were aligned with the ClustalW algorithm with default settings in MEGA X using the maximum likelihood method with the LG+G+I model. The local probability of each branch was calculated using the neighbor-joining method with 1000 replications.
RNA sequencing and analysis
Cells for the wild-type (10D) strain and a single cry strain were grown in 2× Allen's medium under continuous white light for 3 days at 40°C and harvested from a 15-ml culture [optical density at 750 nm (OD750) of 0.6] by centrifugation at 1500 g for 5 min at room temperature. Total RNA was purified using the Trizol/RNeasy hybrid protocol (Trizol, Life Technologies; RNeasy Mini Kit, Qiagen). Polyadenylated [poly(A)] RNA was then purified with the NEBNext Poly(A) mRNA Magnetic Isolation Module (NEB, E7490). Sequencing libraries were constructed with the NEBNext Ultra II Directional RNA Library Prep kit for Illumina (NEB, E7760), amplified with custom oligonucleotides, and sequenced as 150-bp paired-end reads on an Illumina NovaSeq sequencer at GENEWIZ Inc. After adaptor trimming and removing low-quality reads by Trimmomatic-0.39 (Bolger et al., 2014), clean reads were mapped to the C. merolae genome (ASM9120v1) using bowtie2 and counted by featureCounts (Langmead and Salzberg, 2012; Liao et al., 2014). The RNA-seq experiment was performed using a single replicate. Genes with over 100 reads were selected and their expression normalized to transcripts per kilobase million (TPM). We identified the top upregulated and downregulated genes between the wild-type and the cry strains from the Log2 (fold-change) of TPM-normalized read counts. The genes are listed in Table S2.
We thank our lab colleagues for their support and advice during this project.
Conceptualization: Y.Y.; Methodology: Y.M., T.F., Y.Y.; Validation: N.T., Y.M., K.Y., Y.T., T.H., Y.Y.; Formal analysis: N.T., Y.M., K.Y., Y.T., Y.Y.; Investigation: N.T., Y.M., T.F., Y.Y.; Resources: Y.M., T.F.; Data curation: N.T., Y.M., Y.Y.; Writing - original draft: N.T., Y.M., T.H., Y.Y.; Visualization: N.T., Y.M., Y.Y.; Supervision: Y.Y.; Project administration: Y.Y.; Funding acquisition: T.F., T.H., Y.Y.
This work was supported by PRESTO from the Japan Science and Technology Agency (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. JP18K06325 to Y.Y. and 18K06300 to T.F.); the Sumitomo Foundation (no. 180705 to Y.Y.); the Institute for Fermentation, Osaka (L-2020-2-008 to Y.Y.); the Grant-in-Aid for Scientific Research on Innovative Areas from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (no. 16H06465 to T.H. and Y.Y.); and CREST from the Japan Science and Technology Agency (JPMJCR20E5 to T.H.).
Raw RNA sequencing data have been deposited into the Sequence Read Archive (SRA) with project ID PRJNA773182.
The authors declare no competing or financial interests.