Keywords

1 Introduction

Diatoms occupy a wide variety of ecological niches in marine, freshwater, and terrestrial habitats. Their extraordinary ecological success suggests that diatoms possess the ability to rapidly adjust their cell division and growth to a multitude of cues from the external environment. Unlike most microalgae which exhibit haplontic or biphasic haplodiplontic life cycles, diatoms are diplontic, with gametes being the only haploid stage in their life cycle (von Dassow and Montresor 2011). Following mitotic division, each of the two diatom daughter cells inherits one of the parent cell wall valves that serves as an epitheca, and forms a new hypotheca. The daughter cell that inherits the original epitheca remains the same in size, but the daughter cell that inherits the original hypotheca becomes smaller (Fig. 1). As a result, diatom populations typically show a gradual reduction in mean cell size over successive mitotic divisions. Notable exceptions include the model diatoms Phaeodactylum tricornutum and Thalassiosira pseudonana, which do not show cell size decline (Lewis 1984; Chepurnov et al. 2008). The original cell size is typically restored through sexual reproduction, although vegetative cell size restitution has been observed in several species (Chepurnov et al. 2004). Estimates show that it can take natural diatom populations several years of vegetative reproduction before reaching sexual maturity through cell size decline (Mann 2011). Thus, the diatom life cycle predominantly consists of a long vegetative phase of mitotic cell divisions interrupted by only a short meiotic phase.

Fig. 1
figure 1

Simplified overview of cellular events during the diatom cell cycle. The phases of the mitotic cell cycle and cytokinesis/cell wall formation are indicated at the center of the figure, with G1: gap1 phase, S: DNA synthesis, G2: gap2 phase. Microscopic events in the cell during each phase are indicated around the cell cycle, with PP = polar plate, MC = microtubule center (MTOC), PC = polar complex, MT = microtubule, SDV = silica deposition vesicle

Full size image

The eukaryotic cell cycle consists of a tightly orchestrated succession of events, including DNA replication (S phase) and segregation of genetic material into daughter cells (M phase or simply “mitosis”) separated by two gap phases, G1 where metabolic changes necessary for S phase are performed, and G2 phase where preparation for mitosis takes place (Fig. 1). The M phase is ultimately followed by cell division (cytokinesis) (Morgan 2007). Although many elements of the mitotic cell cycle are conserved across eukaryotes, taxa display their own unique features. Traits that define the diatom cell cycle include an unusual spindle that is initially formed outside the nucleus, the biosynthesis of silica cell wall elements during cytokinesis, and the presence of multiple cell cycle checkpoints responding to nutrients, light, and other environmental conditions. Furthermore, several studies hint at the presence of species-specific molecular features in the diatom cell cycle, which might account for their adaptations to different circumstances.

Historically, diatoms have been one of the prime subjects in fundamental research on the cell cycle. Already in 1896, Robert Lauterborn observed cell division and karyokinesis in diatoms (Lauterborn 1896). For about a century, diatoms were the subject of intense study into the process of mitosis, first using light and electron microscopy, later with isolated spindles and biochemical approaches (De Martino et al. 2009). Several features make diatoms excellent models to study mitosis, such as the large cell size of some species facilitating microscopical observation, the well-organized spindles with spatially separated midzone and chromosome regions, and the possibility to study functional spindles in vitro (De Martino et al. 2009). Although several recent works have studied the molecular workings of the mitotic cell cycle, the structure and regulation of the diatom cell cycle is still largely unexplored. Nevertheless, genome sequencing of an increasing number of diatom species offers the possibility for the identification of the cell cycle machinery components across the major diatom clades. The establishment of new techniques for genome editing such as TALEN and CRISPR-Cas9 can be used for functional studies of genes involved in cell cycle progression. Such knowledge can subsequently be used to manipulate the growth of diatoms for industrial use.

In this chapter, we focus on the cytological events of the diatom mitotic cell cycle, followed by an overview of the diverse mechanisms of molecular and environmental regulation.

2 Cellular Events During the Cell Cycle

2.1 Organelle Division

Among diatoms, there is considerable variation in the number of chloroplasts per cell, ranging from monoplastidic to polyplastidic species (Mann 1996). While centric diatoms are generally polyplastidic, the number of chloroplasts in pennate diatoms is more restricted (mono-, di-, tetra- or oligoplastidic) (Mann 1996). Chloroplasts divide by constriction into two daughter chloroplasts. Two types of chloroplast division exist in diatoms: autonomous and imposed. In the former, constriction occurs without involvement of other organelles, while in the latter the chloroplast is divided by the cytoplasmatic cleavage furrow (Mann 1996). In mono-, di-, and tetraplastidic diatoms, the developmental cycle of chloroplasts (chloroplast ontogeny) is tightly linked with cell cycle progression, and division generally occurs just before or just after cytokinesis (Mann 1996). Recent studies showed that chloroplast division takes place between the start of S-phase and the end of mitosis in the pennate diatoms Seminavis robusta (Gillard et al. 2008) and P. tricornutum (Huysman et al. 2010). A mechanism that couples chloroplast division with successful progression through the S phase is suggested by the lack of chloroplast duplication when S. robusta cells were chemically prevented to enter S phase (Gillard et al. 2008). Furthermore, treatment with the cytostatic sex pheromone SIP which arrests the cell cycle in G1 phase led to a downregulation of the expression of key chloroplast division genes DRP5B and FtsZ (Bilcke et al. 2021a). Observations in other groups of algae additionally suggested the presence of a G2/M checkpoint that monitors the completion of chloroplast division before cells enter mitotic metaphase (Sumiya et al. 2016). The presence of a possible G2/M chloroplast division checkpoint in diatoms is so far unclear.

A key feature of chloroplast ontogeny in non-polyplastidic diatoms is a cell cycle-dependent rearrangement of chloroplasts prior to cell division. Chloroplasts move from the girdle region to the valves before or during mitosis and move back to the girdle after cytokinesis (Mann 1996; Gillard et al. 2008). In S. robusta, this repositioning coincides with the transcriptional induction of β-tubulin (Gillard et al. 2008). Combined with the observation that chloroplast movement in Cymatopleura solea is sensitive to the microtubule inhibitor colchicine (Pickett-Heaps 1991), this suggests that the microtubule cytoskeleton is involved in chloroplast rearrangements.

The inheritance of chloroplasts is often closely regulated during cytokinesis. While in polyplastidic diatoms the re-distribution of chloroplasts is stochastic, inheritance of chloroplasts is tightly controlled in non-polyplastidic diatoms, with species displaying either unique inheritance (daughter cell receives chloroplasts that are clonal) or dual inheritance (daughter cell receives the progeny of different chloroplasts) (Mann 1996). Dual inheritance, which is observed in species with imposed chloroplast division, leads to structural hybridity of chloroplasts in each of the clonal cells in the population (Mann 1996).

Recent work showed that in P. tricornutum, the Golgi apparatus duplicates just before cell division and positions itself at both sides of the cleavage furrow (Tanaka et al. 2015). Moreover, mitochondria appeared elongated and were aligned along the future division plane, probably to allow equal distribution of mitochondria to each daughter cell (Tanaka et al. 2015). Evidence has also emerged that next to chloroplasts, stramenopile mitochondria use ancestral bacterial-like division machinery such as FtsZ and Min proteins (Leger et al. 2015). While Thalassiosira pseudonana contains genes encoding this unique mitochondrial machinery, they were remarkably absent in P. tricornutum, where all bacterial division genes could be attributed to the chloroplast (De Martino et al. 2009). In S. robusta the three plastidic FtsZ homologs as well as a mitochondrial FtsZ show rhythmic expression during a diurnal regime, suggesting that both chloroplasts and mitochondria divide in synchrony with the diel cycle (Bilcke et al. 2021b).

2.2 Diatom Microtubule Organizing Centers

In general, the structural events of the mitotic cell cycle are mediated by two kinds of cytoskeletal proteins: the mitotic spindle consists of microtubule (MT) fibers, while the contractile ring formed during cytokinesis is actin-based (Scholey et al. 2003). Microtubule organizing centers (MTOCs) are structures found in eukaryotic cells from which microtubules (MTs) emanate (De Martino et al. 2009). They nucleate and localize MTs during interphase and in the mitotic spindle, while specialized MTOCs called basal bodies also occur at the basis of cilia and flagella. Genomic data indicate that typical MTOC-associated genes are conserved in diatoms, including members of the gamma-tubulin complex (gamma-TuRC (Scholey et al. 2003)) such as gamma-tubulin, GCP2 and GCP3 (De Martino et al. 2009). Diatom MTOCs exhibit several unique features. Firstly, diatom MTOCs lack centrioles. Centrioles are cylindrical structures consisting of nine MT triplet blades which were presumably present in the MTOC of the last eukaryotic common ancestor (LECA) (Marshall 2009). Although centrioles are retained in many eukaryotic lineages, they were secondarily lost in several lineages, including fungi, red algae, and angiosperms (Marshall 2009). While centrioles are present in the MTOCs of the related clade of brown algae, they are generally absent in diatoms, although they persist as basal bodies in the flagellated male gametes of centric diatoms (Round et al. 2007). Curiously, three putative centrins were identified in the genomes of the centric T. pseudonana and pennate P. tricornutum (De Martino et al. 2009) which are involved in the duplication of centrioles in animals (Salisbury et al. 2002). Another unusual feature of diatom MTOCs is its behavior during the mitotic cell cycle. In animals, an MTOC known as the centrosome is present in one copy during most of the interphase. The centrosome eventually duplicates to form the two poles of the mitotic spindle, which migrate to different sides of the cell. In diatoms, however, the MTOC—called the MT center (MC)—does not duplicate before mitosis. Instead, a second structure called the polar complex (PC) appears close to the MC (Fig. 1) (Lauterborn 1896). The PC is clearly a different structure as it is cubic in shape in contrast to the rounded MC. The MC disintegrates once the PC appears. Later, during cytokinesis, the MC reappears and the PC disappears (De Martino et al. 2009; Huysman et al. 2014a). Notably, molecular and biochemical research to investigate the nature of these temporally and spatially independent MTOCs in diatoms is completely lacking, presenting an interesting opportunity for future research.

2.3 Mitotic Prophase

During the prophase of mitosis, the PC separates into two structures called polar plates (Fig. 1). These flattened structures migrate to form the poles of the mitotic spindle outside the intact nuclear envelope (Fig. 1). Diatom spindle assembly is MTOC-directed, i.e. nucleated by the polar plates and not by the chromosomes (Tippit et al. 1980). The spindle is a dipolar structure consisting of two spindle poles from which MTs radiate with their plus ends towards the center of the spindle (Civelekoglu-Scholey and Scholey 2010). In diatoms, spindle MTs are organized in two sets: (i) the first set of MTs form an interdigitating region with MT from the other spindle pole called the central spindle or midzone, while (ii) the second set of MTs (“kinetochore MTs”) associate with chromosome kinetochores for segregation. The two groups of spindle MTs are spatially and functionally separated (McDonald et al. 1986). The kinetochore MTs play a role in chromosome attachment and segregation (prometaphase-anaphase A), while the central spindle is essential for spindle elongation (anaphase B). Spindle MTs serve as molecular ratchets. Polymerization and depolymerization at the ends of MTs exert compressive (push) and tensile (pull) forces. At the same time, they act as polarized tracks along which motor proteins can carry cargo (De Martino et al. 2009; Civelekoglu-Scholey and Scholey 2010). These motor proteins have various roles during mitosis, ranging from spindle assembly, chromosome segregation, spindle elongation, and cytokinesis (Shi et al. 2019).

2.4 Mitotic Prometaphase

The prometaphase is a vital stage in mitosis, which is characterized by the bipolar attachment of chromosomes to the mitotic spindle (Tippit et al. 1980) (Fig. 1). At the start of prometaphase, the mitotic spindle including the polar plates enters the nucleus which is regionally broken down. A complete nuclear envelope breakdown does not take place: the envelope stays partially intact throughout mitosis (Tippit et al. 1980). Thus, like in certain fungi, including Saccharomyces cerevisiae, but in contrast to animals and plants, diatoms exhibit closed mitosis, i.e. the chromosomes segregate inside an intact nuclear envelope (Boettcher and Barral 2013; Makarova et al. 2016).

After the spindle enters the nucleus, spindle MTs start binding chromosomes at their kinetochores. Kinetochores are protein complexes assembled at the centromere of each sister chromatid that attach the centromere to the kinetochore MTs (De Martino et al. 2009). Diatom kinetochores do not seem to nucleate MTs themselves, but attach existing MTs from the polar plates to the chromosomes (Tippit et al. 1980). De Martino and colleagues identified a homologue of the centromere-specific CENP-A in the genome of P. tricornutum (De Martino et al. 2009; Diner et al. 2017). This histone H3 variant replaces the default histone H3 at the site of kinetochore assembly during mitosis (Diner et al. 2017; Amor et al. 2004). Recently, 25 P. tricornutum centromere regions were identified using chromatin immunoprecipitation (ChIP) of the centromere-specific CENP-A. Centromere regions typically showed very low GC content (< 33%), but lacked sequence conservation. Notably, adding diatom centromeres to foreign DNA enabled a stable maintenance as an extrachromosomal episome (Diner et al. 2017).

Almost simultaneously with the introduction of the spindle inside the nucleus, diatom chromosomes show irregular oscillations, especially at their kinetochore region (Tippit et al. 1980). First, chromosomes move between the spindle center and one pole, presumably because the kinetochore is only attached to one pole of the spindle. Later, connection to the second pole is established and chromosome movement becomes more stable. Chromosomes move towards a central position to achieve metaphase configuration and a bidirectional tension builds up at the kinetochores, pulling the chromosome towards both poles simultaneously (Tippit et al. 1980). The oscillatory movement of diatom chromosomes is typical for the process of chromosome congression, which is the positioning of chromosomes near the spindle equator after attachment of the spindle to the kinetochores. It is thought to be necessary for the correct distribution of chromatids towards each pole during segregation (anaphase A) (Maiato et al. 2017). Across eukaryotes, chromosome congression is typically achieved by a combination of (de)polymerization of kinetochore MTs and the action of motor proteins (Scholey et al. 2003).

2.5 Mitotic Metaphase

During metaphase, chromosomes are attached to the spindle and the condensed chromosomes organize in a ring-formed “donut” around the spindle equator (Fig. 1) (De Martino et al. 2009). Chromosome condensation during mitosis and meiosis involves large condensin complexes, which contain core SMC (“Structural Maintenance of Chromosomes”) ATPases SMC2 and SMC4. These widely conserved condensin subunits could be identified in the gene sets of six diatom species (Patil et al. 2015).

The end of the metaphase is characterized by the inactivation of the spindle assembly checkpoint (SAC). This checkpoint prevents the metaphase-to-anaphase transition by inhibiting the APC/C if the spindle is not properly attached to the kinetochores (Kops 2008). Although SAC (MAD1, MAD2) and APC/C components were identified in diatom genomes (De Martino et al. 2009), the molecular basis for diatom SAC activation is unexplored.

2.6 Mitotic Anaphase A

During anaphase A, chromosomes segregate, and the released chromatids start moving towards the spindle poles (Fig. 1). In contrast to anaphase B (spindle elongation, see below), relatively little is known about the molecular regulation of anaphase A. DNA replication (S-phase) and segregation of chromosomes (anaphase A of the M-phase) are temporally separated in eukaryotes. Therefore, it is essential that sister chromatids are held together to be released simultaneously at the start of anaphase A (Mehta et al. 2012). The inactivation of the SAC during the metaphase-anaphase transition leads to dissociation of the cohesin complex, the molecular lock that keeps the chromatids together. Cohesin complex subunits SMC1, SMC3, RAD21, and SCC3 were identified in six queried diatom genomes (Patil et al. 2015), indicating that the cohesin complex structure and function is likely also conserved in diatoms (Mehta et al. 2012). The chromatid-to-pole tension necessary for chromosome segregation in eukaryotes is based on both depolymerization of MT as well as various motor proteins at the kinetochores and/or the poles, although there is still ongoing debate about the relative contributions of each factor (Civelekoglu-Scholey and Scholey 2010).

2.7 Mitotic Anaphase B

In contrast to anaphase A, during which the chromatids are pulled towards the poles, the spindle itself elongates during anaphase B, pushing the spindle poles further apart (Fig. 1). Spindle elongation in diatoms has been investigated thoroughly in the 1970s and 1980s using in vitro and in vivo spindles (De Martino et al. 2009). A crucial role for outward MT sliding in the central region was suggested when spindle elongation was observed in isolated spindles supplemented with ATP, even after digestion of chromatin (Cande and McDonald 1985, 1986). Interestingly, when tubulin was added, elongation length increased markedly, suggesting that MT polymerization supports spindle elongation, although it is not strictly necessary (McDonald et al. 1986; Masuda and Cande 1987).

The sliding filament hypothesis, which states that central antiparallel MT moving in opposite directions generates outward forces on spindle poles, is currently well established across eukaryotic organisms, from budding yeast to vertebrates (Civelekoglu-Scholey and Scholey 2010). In these model organisms, an important role is set aside for bidirectional motor proteins such as kinesin-5 (Civelekoglu-Scholey and Scholey 2010; Singh et al. 2018), which are attached to antiparallel MT and move over the filaments to generate an outwards force on the spindle poles (Civelekoglu-Scholey and Scholey 2010). Also in diatoms, kinesins appear to be necessary for spindle elongation. In Cylindrotheca fusiformis, Wein and colleagues described Diatom Spindle Kinesin 1 (DSK1), a kinesin-like protein associated with the central spindle region. Inhibition of DSK1 function with anti-DSK1 antibodies abolishes in vitro spindle elongation (Wein et al. 1996). Later, immunofluorescence studies and tubulin extraction showed that DSK1 is associated with a non-microtubule scaffold in the central spindle of diatoms (Wein et al. 1998). Interestingly, the dynein inhibitor Vanadate can inhibit the elongation of spindles in vitro (Cande and McDonald 1985). This suggests that next to the plus-end directed kinesins, also minus-end directed dyneins are involved in spindle elongation. A role for dynein in anaphase B spindle dynamics was later also postulated in other eukaryotic taxa (Fink et al. 2006; Yeh et al. 1995).

2.8 Cytokinesis

Diatom cytokinesis occurs in two distinct stages: first, a cleavage furrow is formed that separates future daughter cells, followed by the formation of a new silica cell wall synthesized by specialized organelles called silica deposition vesicles (SDV) (Fig. 1). Diatoms exhibit a combination of plant and animal cytokinetic mechanisms: the centripetal (inwards) formation of a cleavage furrow closely resembles cleavage of animal cells, while the centrifugal (outwards) deposition of new cell wall material bears similarities to plant cell division (De Martino et al. 2009; Bowler et al. 2010). The diatom cell wall and its biosynthesis are covered in detail in Chaps. “Structure and Morphogenesis of the Frustule” and “Biomolecules Involved in Frustule Biogenesis and Function.” In this paragraph, we focus on the still largely unexplored molecular processes during diatom cell cleavage.

In the pennate diatom P. tricornutum, daughter cells are separated parallel to the existing valves by ingrowing cleavage furrows from both poles of the cell (Tanaka et al. 2015). In an elegant series of experiments on the centric Stephanopyxis turris, Wordeman and colleagues used a MT polymerization inhibitor to show that the position of cell separation is independent of the position of the mitotic spindle (Wordeman et al. 1986). When the mitotic spindle was displaced to one side of the cell, the cleavage furrow would nevertheless bisect the cell centrally, resulting in binucleate or anucleate cells. This indicates that the location of the division plane is determined before mitosis, unlike animal cells, where its location is determined by the location of the spindle (Gillies and Cabernard 2011). In many eukaryotic taxa, with the notable exception of higher plants, cell cleavage is mediated by an actomyosin-powered contractile ring (Cheffings et al. 2016). In the pennate diatom P. tricornutum (Tanaka et al. 2015) and the centric Rhizosolenia setigera (Van de Meene and Pickett-Heaps 2004), staining of actin filaments showed that a putative actomyosin contractile ring is present at the cleavage furrow. Intriguingly, recent evidence indicates that the centripetally progressing cleavage furrow bisects the intact nucleus in P. tricornutum (imposed karyokinesis) (Tanaka et al. 2015), implicating that karyokinesis is inherently linked to cytokinesis. In contrast, nuclear division was completed before initiation of cleavage furrow formation in the pennate diatom S. robusta (Gillard et al. 2008), and the centric S. turris, where karyokinesis is accompanied by a highly ordered spindle that nucleates astral microtubules (Wordeman and Cande 1990). Furthermore, YFP-labeling of marker proteins in P. tricornutum showed that vesicle trafficking is involved in the formation of the cleavage furrow (Tanaka et al. 2015). The authors hypothesized that Syntaxin (a t-SNARE complex subunit) is present on the newly formed cleavage furrow membrane to allow vesicle docking, while the vesicle trafficking protein SEC4 is localized in post-Golgi secretory vesicles that are already present in the future division plane (Tanaka et al. 2015).

3 Regulation of the Cell Cycle

3.1 Core Components of Cell Cycle Regulation in Diatoms

Progression through the cell cycle requires tight coordination of the timing of DNA replication, nuclear division and cytokinesis to allow accurate inheritance of the genetic information to the two daughter cells. The entry into the cell cycle is safeguarded by the G1/S phase checkpoint that ensures that cells start the process of cell division only under favorable internal and external conditions (Fig. 2). Similarly, a survey mechanism at the G2/M checkpoint prevents cells from entering the M phase before the completion of DNA replication, whereas the spindle checkpoint during anaphase (see above) prevents the onset of sister chromosome segregation until correct bipolar attachment to the spindle is achieved (Sullivan and Morgan 2007) (Fig. 2). These key characteristics of cell cycle progression are shared by all eukaryotes, but the duration of individual phases, the factors influencing the execution of the cell cycle checkpoints, and the components of the cell cycle machinery vary among species and cell types. In the following text, we will describe the key components found in diatoms controlling the molecular cell cycle machinery and discuss the differences to other eukaryotes.

Fig. 2
figure 2

Regulation of the diatom cell cycle: Regulatory pathways controlling cell cycle progression are shown, with cell cycle checkpoints indicated by a yellow arrow. At the G1-S and G2-M checkpoints, environmental cues marked with a vertical red line impact cell cycle progression at these transition points in specific species, being Fragilariopsis cylindrus (Fcy), Thalassiosira weissflogii (Twe), Cylindrotheca fusiformis (Cfu), Thalassiosira pseudonana (Tps), Seminavis robusta (Sro), Phaeodactylum tricornutum (Ptr), and Pseudo-nitzschia multistriata (Pmu)

Full size image

3.1.1 CDK Complexes

In all eukaryotes, the progression through the cell cycle requires the activity of the so-called CDK complexes, composed of a cyclin-dependent kinase (CDK) and a cyclin. Active CDK complexes drive the progression through the cell cycle through phosphorylation of target molecules (Morgan 2007). Conversely, the exit from mitosis and entry into a new G1 phase requires dephosphorylation of CDK target proteins, which is achieved through a sharp decrease in CDK complex activity through proteolytic removal of all M-phase cyclins, marking the end of the cell cycle (Sullivan and Morgan 2007; Minshull 1989; Murray et al. 1989).

3.1.2 Cyclin-Dependent Kinases

In general, CDKs serve as effector subunits that phosphorylate downstream substrates, whereas cyclins are the regulatory subunits that activate CDK activity by inducing a conformational change in the CDK subunit. Additionally, cyclins contribute to the substrate specificity of the CDK (Morgan 2007; Pines and Hunter 1989). However, the activity of CDK complexes is regulated at many more levels. The cyclin-dependent kinases belong to a family of serine/threonine protein kinases and are inactive as monomers. Their activation includes a phosphorylation event at the so-called T-loop of the CDK by the CDK activating kinase (CAK) (Kaldis 1999). Without this phosphorylation, the T-loop obscures the catalytic cleft of the CDK, preventing substrates to dock on the CDK/cyclin complex. Conversely, CDK activity can be negatively regulated through phosphorylation by the Wee1 or Myt1 kinases and can additionally be inhibited by binding of mostly small sized proteins that either block the catalytic cleft or through inducing a conformational change in the CDK subunit by which it loses interaction with the cyclin subunit. It is important to note that besides the CDKs directly involved in driving cell cycle progression, eukaryotic genomes contain a varying number of additional CDK genes with diverse functions in RNA splicing, metabolism, transcription, or epigenetic regulation (Gressel et al. 2017; Li et al. 2014; Lim and Kaldis 2013).

Based on sequence data only, seven CDK-like proteins encoded in P. tricornutum genome cannot be phylogenetically assigned to a known CDK class due to their highly divergent sequences and were denoted as hypothetical or hCDKs. Five other CDKs do show homology to known CDKs (Huysman et al. 2010; Huysman et al. 2015). Two of them encode C-type CDKs that in metazoans and plants play a role in the phosphorylation of the C-terminal domain of the largest subunit of RNA polymerase II and in the release of the RNA polymerase II, activating transcription (Gressel et al. 2017; Li et al. 2014; Pinhero et al. 2004). One gene encodes a putative CDK activating kinase, but to this date its function was not yet investigated. The two other CDK genes, denominated CDKA1 and CDKA2, encode CDKs with PSTAIRE and PSTALRE cyclin binding motifs respectively. The presence of two distinct motives indirectly suggests that the two CDKs bind to different types of cyclins. PSTAIRE-containing CDKs have been found in all eukaryotic model systems (Malumbres 2014). Conversely, CDKs with the PSTALRE motif are found only in a few organisms, including Ostreococcus tauri, Dictyostelium discoideum, Ectocarpus siliculosus, and Cyanidioschyzon merolae (Bisova et al. 2005; Bothwell et al. 2010; Cízková et al. 2008; Corellou et al. 2005; Michaelis and Weeks 1992). This motif appears to be situated in between the prototypic PSTAIRE motif and the PPTALRE and PPTTLRE motifs found in B1- and B2-type CDKs of the plant lineage, where they have been shown to play a role during the G2/M phase and DNA repair by homologous recombination (Bisova et al. 2005; Tulin and Cross 2014; Weimer et al. 2016). The P. tricornutum CDKA1 was found to be cell cycle phase-dependently transcribed throughout the cell cycle with maximal mRNA levels during G1/S phase (Huysman et al. 2010; Huysman et al. 2015; Annunziata et al. 2019). In accordance with this observation, CDKA1 was demonstrated to bind the dsCYC2 cyclin that is a driver of cell cycle onset after light exposure (Huysman et al. 2013) (Fig. 2). In contrast, the transcription of CDKA2 peaks at the G2/M boundary (Huysman et al. 2010, 2015) and the overproduction of the CDKA2 protein delayed cell cycle progression by causing a prolonged M phase (Fig. 2). This could be explained by delayed anaphase onset due to increased CDK activity. In line with the suggested role of CDKA2 in G2/M phase transition, the CDKA2 protein localizes in the nucleus in interphase cells, but after chloroplast division and translocation, the protein is re-localized to the cell division plane between the separated chloroplasts.

3.1.3 Cyclins

The number of cyclin genes varies widely among eukaryotic species. The fission yeast Schizosaccharomyces pombe encodes only five cyclin-like genes, and it was shown that a single construct of the CDK protein Cdc2 fused with the Cdc13 cyclin is sufficient to drive the progression through the mitotic as well meiotic cell divisions (Gutiérrez-Escribano and Nurse 2015). In contrast, the human genome encodes at least 29 cyclins, whereas 49 genes with cyclin domain(s) were found in the genome of the angiosperm plant Arabidopsis thaliana and 59 in maize (Hu et al. 2010; Wang et al. 2004). It was proposed that such gene family expansion allows functional specialization during evolution (Bloom and Cross 2007). In silico examination of the genomes of P. tricornutum, S. robusta, and T. pseudonana showed that they encode 28, 56, and 57 cyclin genes, respectively (Huysman et al. 2010; Bilcke et al. 2021a). Current evidence from transcriptional studies and functional characterization supports the idea of functional specialization of cyclin family members in diatoms in response to various external and physiological conditions.

Based on their sequence, cyclins are classified into different types. In mammals the A-, B-, D-, and E-type cyclins are directly involved in cell cycle progression. Similar to plants, no E-type cyclin was discovered in the genomes of the three diatoms. As expected, A- and two B-type cyclins were expressed at the G2/M phase of cell cycle (Huysman et al. 2010; Annunziata et al. 2019). D-type cyclins are known for driving the G1/S phase transition in response to mitogenic stimuli. However, the only P. tricornutum CYCD1 was found to be rather expressed in the G2/M phase. It remains to be determined whether the CYCD1 is responsive to endogenous or exogenous factors and what role it plays in G2/M transition. Meanwhile, there are several other cyclins such as P-type cyclins and diatom-specific cyclins that are potential candidates for the integration of mitogenic stimuli at the cell cycle progression.

Similar to the CDKs, some cyclins rather link with metabolic and transcriptional processes. From the families of these cyclins that are conserved among eukaryotes, one H-type, one L-type cyclin, and six P-type cyclins were found in P. tricornutum. Cyclin H1 is a putative partner of CAK protein. Correspondingly, both genes display co-expression during the P. tricornutum cell cycle. L-type cyclins are involved in the regulation of transcription by RNA polymerase II and pre-mRNA splicing (Chen et al. 2007; Dickinson et al. 2002). P- or the Pho-80 type cyclin was first discovered in S. cerevisiae as a cyclin that mediates the cell division response to environmental phosphate availability (Toh-e et al. 1988). Related family members were later identified in other protists and plants where they play roles in nutritional sensing to cell cycle progression, cell division, development, and differentiation (Hu et al. 2010; Carroll and O’Shea 2002; Hammarton et al. 2004; Peng et al. 2014; Roques et al. 2015; Torres Acosta et al. 2004; Van Hellemond et al. 2000). Five P-type cyclins can be found in P. tricornutum (Huysman et al. 2010). Expression of all members of this group is present at the beginning of the cell cycle, with CYCP3/5 and CYCP6 peaking just after the transition from dark to light (Huysman et al. 2010, 2013; Smith et al. 2016). This suggests a possible role of diatom P-type cyclins in the transduction of environmental signaling on the cell cycle progression, but their exact roles remain to be determined.

The most diverse group of cyclins in diatom genomes are the diatom-specific cyclins with 11 members in P. tricornutum and 45 in T. pseudonana. These cyclins form a distinct group that does not contain cyclins from non-diatoms in phylogenetic analysis. The function of these cyclins is mostly unclear, the only diatom cyclin functionally characterized to date being the P. tricornutum diatom-specific dsCYC2 (Huysman et al. 2013). DsCYC2 transcription is instantly induced upon transition from dark growth conditions to light, with wavelengths corresponding to blue light eliciting the strongest response. This expression pattern is under the control of the transcription factor bZIP10 and AUREOCHROME1a, a member of stramenopile-specific group of photoreceptors that combines a light-responsive photoreceptor domain with a transcription factor domain (Heintz and Schlichting 2016; Kerruth et al. 2014; Takahashi et al. 2007; Depauw et al. 2012). Consistent with its presumed role at the onset of the cell cycle, dsCYC2 binds the CDKA1 kinase and was able to complement a yeast G1 cyclin mutant. Correspondingly, its silencing prolonged the duration of the cell cycle by extending the G1 phase. This phenotype was absent when cells were cultivated under continuous light, further confirming a role for dsCYC2 in a light-dependent G1 cell cycle checkpoint. How the cell cycle arrest in darkness is established and which targets become phosphorylated by dsCYC2/CDKA1 upon light exposure remains to be determined.

Another member of diatom-specific cyclins, Sro299_g111470, was recently shown to be expressed specifically during the early stages of sexual reproduction in S. robusta. The expression of this cyclin was strongly upregulated in cells exposed to the sex-inducing pheromone of the opposite mating type, making it the first known cyclin putatively involved in diatom sexual reproduction (Bilcke et al. 2021a) (See Chap. “Life Cycle Regulation”).

Several studies followed the transcriptional responses of diatoms to various treatments and conditions. Experiments focused on the response and recovery from silica starvation in T. pseudonana showed differences in the expression pattern between the closest homologs of three T. pseudonana cyclins and one hypothetical CDK compared to P. tricornutum (Shrestha et al. 2012). While profiling the transcriptome of P. tricornutum in response to limited iron supply over diurnal cycles, Smith and colleagues found several cyclins to be expressed in a different phase of the cell cycle than previously described by Huysman and colleagues (Huysman et al. 2010; Smith et al. 2016). These discrepancies in expression profiles may be explained by differences in synchronization procedures used between experiments. Moreover, the activity of CDK complexes is dependent not only on cyclin transcription but also on the rate of mRNA translation and the stability of cyclins at the protein level (Kronja and Orr-Weaver 2011; Peters 2002). Therefore, the combined use of translational reporter fusion systems for cyclins, observation of the dynamics of cyclin localization during distinct stages of cell cycle in individual cells, and functional characterization and identification of CDK targets should elucidate the participation of individual CDK and cyclin genes in cell cycle progression in diatoms.

3.1.4 Modulators of the Cell Cycle

As highlighted above, CDK activity can be modulated in many different ways, including inhibitory phosphorylation and proteolytic turnover of cyclins. Although there are several families of CDK inhibitors (CKIs) known in various eukaryotes, no clear homologous genes were found during the examination of the P. tricornutum genome (Huysman et al. 2010). The Cdc28 kinase subunit 1 or CKS1/SUC1 family proteins were found to be directly interacting with the CDK complexes and to act as a docking factor, promoting the interaction of CDK complexes with primed substrates (Kõivomägi et al. 2013; McGrath et al. 2013). One CKS1 gene was found in P. tricornutum and according to yeast two-hybrid experimental results, this protein binds specifically to CDKA2 but not to CDKA1 (Huysman et al. 2015). In accordance with this observation, both genes show the highest expression in the G2/M phase. The inhibitory phosphorylation of CDK complexes is carried out by kinases from the WEE1/MYT1/MIK1 family. These kinases were shown in other species to prevent premature entry into mitosis and act at the G2/M transition by phosphorylation of CDK-cyclin B complexes (Enoch and Nurse 1990; Kellogg 2003; Nurse 1990). Their activity is counteracted by the CDC25 phosphatase (Katayama et al. 2005; Li et al. 2010; Moura and Conde 2019; Okamoto and Sagata 2007; Ovejero et al. 2012; Watanabe et al. 1995). One gene with homology to Myt1 kinase was found in the P. tricornutum genome but no sequence homolog of CDC25 could be found (Huysman et al. 2010). Thus the role and regulation of the CDK inhibitory phosphorylation in diatom cell cycle progression remains an open question for future research.

3.1.5 Components of the Proteolytic Control of Cell Cycle Progression

The oscillation of CDK activity that drives cell division is ensured by several mechanisms, but timely proteolysis is essential to ensure the proper one-way sequence of events (Clarke 2002). Two conserved E3 ubiquitin ligase complexes are key for the targeting of cyclins, CDK inhibitors and CDK targets, including transcriptional regulators and DNA replication factors for proteasomal degradation. One is the Skp1/Cullin/F-box protein or SCF complex that operates mainly at the G1/S and G2/M transition (Cardozo and Pagano 2004; Willems et al. 2004). The second one is the anaphase-promoting complex or cyclosome (APC/C). As its name suggests, the APC is predominantly required for anaphase onset and progression. All SCF complexes consist of three core subunits; Skp1, Cullin (CUL1), and the RBX1 RING finger domain protein that associate with an F-box protein. In P. tricornutum, components of the cell cycle proteolytic machinery include three Skp1, three CUL1, and one Rbx1 homolog (Huysman et al. 2014b). Most eukaryotic species contain multiple F-box genes, since they encode the variable components that mediate the interaction with the substrate. The number of F-box genes differs greatly among species, with 11 members in budding yeast to 779 putative genes in rice (Kipreos and Pagano 2000; Xu et al. 2009). Seventeen F-box proteins were reported for P. tricornutum (Huysman et al. 2014b). Expression profiling of individual components of the SCF complex showed that they are slightly upregulated at the G1/S phase transition. This may suggest an increased requirement for the SCF complex at this stage of the cell cycle. However, as posttranslational modifications are key regulators of SCF complexes in other eukaryotes, it can be assumed that this will also be the case in diatoms.

APC/C activity is essential for the separation of sister chromatids in anaphase, exit from mitosis, cytokinesis and establishment of G1 phase (Peters 2006) (Fig. 2). The exact regulation and targets of APC/C in diatom mitotic and meiotic cell divisions and its role in the response to environmental signals are currently unknown. The APC/C complex is the largest member of the ubiquitin ligase family with 14 core subunits in metazoans and 13 in yeast. Its activity is triggered by binding of either CDC20 or CDH1, which have different timing during the cell cycle. Part of APC/C with CDC20 (APC/CCDC20) is activated at the end of mitotic prophase and targets A-type cyclins for degradation (Di Fiore and Pines 2010; den Elzen and Pines 2001). However, APC/C activity is under the control of the spindle assembly checkpoint (SAC) and is only fully activated when all chromosomes are properly attached to the spindle. Upon SAC release, APC/CCDC20 destroys two main targets: securin whose degradation allows activation of separase, subsequent cleavage of cohesin and thus sister chromatid separation, and B-type cyclins. B-type cyclin degradation results in a rapid decrease of mitotic CDK activity and allows dephosphorylation of CDK targets necessary for progression through anaphase, cytokinesis, exit from mitosis, and assembly of pre-replicative complexes at the origins of replication. The dephosphorylation of CDH1 allows its association with APC/C (APC/CCDH1) during anaphase and further targeting of mitotic cyclins and additional proteins, including CDC20 (Kramer et al. 2000; Li and Zhang 2009). In G1 phase, APC/CCDH1 prevents premature entry into S phase by targeting mitotic cyclins (Sigrist and Lehner 1997).

Searches in the P. tricornutum genome revealed the presence of all core APC/C subunits with the exception of APC4 (Huysman et al. 2014b). It was proposed that the absence of APC4 might be compensated by APC5, whose overexpression rescues the APC4 mutants in yeast. The APC/C adaptor proteins CDC20 and CDH1 are each encoded by one gene. The recognition motifs of these proteins are known and conserved in most species. Profiling of P. tricornutum cyclins for the presence of such degron sequences, revealed that the destruction box or D-box, which is recognized by both CDC20 and CDH1, is present not only in mitotic cyclins as expected but also in cyclins that were found to be expressed at the G1/S transition (Huysman et al. 2010, 2014b). The dsCYC2 involved in the light-dependent cell cycle onset in diatoms is one of the G1/S cyclins containing a putative D-box, which raises questions about the APC/C involvement in maintaining cell cycle arrest in the darkness. CDH1 has a broader range of targets compared to CDC20, as it targets proteins containing both D-box, KEN-box, and a few less-well-known motifs such as A-box, CRY-box, GxEN-box, and LxExxxN motif (Castro et al. 2002, 2003; Littlepage and Ruderman 2002; Pfleger and Kirschner 2000; Reis et al. 2006). However, no KEN box was detected in P. tricornutum cyclins. The KEN box was also absent in CDC20, a known target of APC/CCDH1. However, this protein contains both a putative D-box and a putative GxEn box, which are recognized by APC/CCDH1. As the D-box is also recognized by CDC20 this points to possible self-regulation of APC/CCDC20.

The transcription of APC/C components was profiled in synchronized P. tricornutum cells (Huysman et al. 2014b). Most APC/C core components show a rather constitutive expression throughout the cell cycle, although APC8 shows transcript levels which rise after light exposure and then remain constant for the following 12 h. As the APC/C activating subunits CDC20 and CDH1 are regulated by protein degradation themselves, their levels need to be replenished each cycle. Correspondingly, transcription of the CDC20 gene was found to be upregulated during the G2/M transition in accordance with its role in the regulation of mitotic progression. Interestingly, the transcript levels of CDH1 were also upregulated during the G2 and M phase time period. This upregulation was inhibited in nocodazole-treated cells that are arrested at metaphase due to activation of the spindle checkpoint. These data suggest that most of the CDH1 transcription occurs only after the anaphase onset and that in P. tricornutum, CDH1 might be required only after anaphase, during cytokinesis and in the establishment of G1 phase.

3.2 Internal and External Factors Influencing Cell Cycle Progression in Diatoms

The exposure of diatom cells to the variable external environment has an important influence on their cell cycle progression, the rate of cell division, and consequently the potential of populations to increase in cell numbers. Here, we summarize the current evidence about how diatoms integrate such environmental changes in their cycle progression through the use of cell cycle checkpoints.

3.2.1 Cell Cycle Checkpoints

Cell cycle progression is continuously monitored for the correct order of events, the timing of the phase transitions and integrity of the genome to ensure precise genome duplication and chromosome segregation to daughter cells. Several checkpoints safeguard this process and halt cell cycle progression until all internal and external conditions are favorable (Elledge 1996; Murray 1994). Three major checkpoints are present in most eukaryotes, including the G1/S checkpoint that controls the decision of cells to transition from the G1 phase to the S phase; the G2/M checkpoint that prevents cells with unrepaired DNA to enter mitosis; and the spindle assembly checkpoint (SAC) that controls the bipolar attachment of mitotic chromosomes (Fig. 2).

The decision to exit the cell cycle to quiescence or to enter a new cycle takes place in G1 phase. Species identity, life history and cell type determine which factors are crucial in this decision, e.g. nutrient availability and the presence of pheromones are known to be of importance in yeast, while interaction with surrounding cells and the presence or absence of growth factors play a key role in animal cells (Cooper 2000). In diatoms, light and silica play a key role in the decision to proceed to cell division (Huysman et al. 2013; Brzezinski et al. 1990). Other factors known to influence the G1 checkpoint in diatoms include the availability of nutrients such as phosphate and iron, but also the presence of sex pheromones (see Chap. “Life Cycle Regulation”) and interaction with other species. In adverse conditions, diatoms can form resting stages that include resting spores, resting cells, and winter stages of polar diatoms (Mcquoid and Hobson 1995, 1996). These stages have in general decreased metabolism and can survive for several years and, as recently suggested for Chaetoceros muelleri, up to millennia (Sanyal et al. 2019) (see Chap. “Life Cycle Regulation”). Mitochondrial maintenance metabolism through the Entner−Doudoroff pathway, glycolysis, and the TCA cycle allows polar diatoms to survive the period of darkness during winter and flexible control of activation of photosynthetic activity and entry into cell division upon exposure to light (Walter et al. 2017; Morin et al. 2020; Kennedy et al. 2019).

In all eukaryotes with the exception of yeast, the G1 checkpoint is secured by the Retinoblastoma (RB) protein (Giacinti and Giordano 2006). In a simplified general model, the RB protein binds the E2F/DP transcription factor(s) to keep them inactive. Once the external and internal conditions for progression to the next round of cell division are met, the G1 CDK complexes are activated by the destruction of their inhibitors and phosphorylate the RB protein. This phosphorylation releases the E2F/DP factor transcription factors from transcriptional repression, by which they activate the transcription of genes required for DNA synthesis and chromosomal replication. Presence of RB, E2F and DP genes was confirmed in diatoms S. robusta (Bilcke et al. 2021a) and P. tricornutum (Marie Huysman et al., personal communication). Cell cycle arrest induced by treatment of sexually competent S. robusta cells with sex inducing pheromone (SIP) caused a significant decrease of two E2F and one DP encoding genes. However, one gene encoding a putative DP gene was upregulated in arrested cells in comparison with cell progressing through cell cycle, pointing at specialized roles of the DP factors in S. robusta’s cell cycle.

The G2/M checkpoint surveys the completion of the DNA replication process or presence of damaged DNA to prevent the entry into mitosis until all DNA is replicated and the damage repaired (Stark and Taylor 2006). The DNA damage is signaled to a kinase network that in most organisms comprises the ATM and ATR kinases, which target and activate the CHK1 and CHK2 kinases among other proteins. CHK1 and CHK2 prevent the entry into M-phase through activation of the WEE1/MYT1/MIK1 kinases, while restricting the activity of the counteracting CDC25 phosphatase. As a consequence, M-phase CDK complex activity is inhibited. Once the DNA is fully replicated and the integrity of the DNA is restored, CDK complexes are phosphorylated by CAK kinases at the activation loop and translocated to the nucleus, where CDC25 phosphatase removes the inhibitory phosphorylation group. A positive feedback loop involving activation of CDC25 by M-phase CDK complexes leads to a steep increase in the M-phase CDK activity and the induction of entry into mitosis.

While a putative MYT1 was found in P. tricornutum and T. pseudonana, CDC25 is missing (De Martino et al. 2009; Huysman et al. 2010). A putative CHK2 homolog was found in both P. tricornutum and T. pseudonana (De Martino et al. 2009), but there is no functional analysis of the ATM/ATR kinase network in diatoms so far and the exact details of the G2/M checkpoint await further characterization. However, it is clear that in contrast to animal or fungal cells where the G1/S phase checkpoint integrates both internal and external signals and the G2/M checkpoint is responsive to the internal cues, external factors influence both G1/S and G2/M checkpoints in many diatoms. Silica depletion was shown to arrest or lengthen both G1 and G2 phase duration in F. cylindrus, Thalassiosira weissflogii, and C. fusiformis, while T. pseudonana arrests predominantly in G1 (Brzezinski et al. 1990; Coombs et al. 1967; Darley and Volcani 1969; Oh et al. 2018; Thamatrakoln and Hildebrand 2007; Vaulot et al. 1986). All these diatoms also arrest their cell cycle at both G1/S and G2/M in response to light deprivation. No cell cycle arrest response to silica depletion was observed in P. tricornutum, probably because this is a rare example of a diatom with a facultative silica cell wall (Brzezinski et al. 1990). The light-dependent cell cycle arrest in G2 phase was also missing in this species. Other external cues, such as sex pheromones appear to arrest cells only when they are in the G1 phase of the cell cycle in Pseudo-nitzschia multistriata and S. robusta, a situation similar to yeast (Basu et al. 2017; Fields 1990; Moeys et al. 2016).

In most eukaryotes, the G1/S and G2/M checkpoint mechanisms also survey the growth of the cell (Barnum and O’Connell 2014). Individual diatom cells increase their volume in the course of the cell cycle (Olson et al. 1986). While several hypotheses have been proposed as to how living cells sense their own cell volume (Amodeo and Skotheim 2016), cell size sensing in the diatom cell cycle is likely a more complex process considering the gradual decline in cell size of a clonally propagating population following auxosporulation.

The mitotic or spindle checkpoint ensures correct segregation of chromosomes during mitosis. Whereas the two previous checkpoints operate through the suppression of CDK activity, high M-phase CDK complex activity prevents progression from metaphase to anaphase at the spindle checkpoint (Musacchio 2015). Chromosomes need to be properly attached to the mitotic spindle in order to segregate each chromatid to the opposite spindle pole. This so-called bipolar attachment of kinetochores is sensed by the mitotic checkpoint complex or MCC, supposedly acting as a pseudosubstrate inhibitor of the APC/CCDC20. Inactivation of APC/CCDC20 prevents degradation of securin and mitotic B-type cyclins. Once all chromosomes are correctly attached, the APC/CCDC20 is activated and securin and B-type cyclin(s) are destroyed. The degradation of securin allows a release of separase, a protease that cuts the cohesin subunit Scc1 and thus promotes the separation of sister chromatids (Ross and Cohen-Fix 2002). The degradation of B-type cyclin(s) permits the dephosphorylation of CDK targets necessary for progression through anaphase, cytokinesis and the establishment of the next G1 phase. As already described above, the components of the APC/C are conserved in diatoms (Huysman et al. 2014b) and homologs of MAD1, MAD2 and separase were identified in P. tricornutum and T. pseudonana (De Martino et al. 2009). Although no functional analysis of the spindle checkpoint was done in diatoms so far, there is indirect evidence that the spindle checkpoint functions in a manner similar to other eukaryotes. The treatment of cells with the microtubule depolymerizing drug nocodazole induces a cell cycle arrest was at the M-phase, consistent with spindle destabilization triggering spindle checkpoint signaling (Huysman et al. 2010; Wordeman et al. 1986).

3.2.2 Factors Influencing the Progression Through the Cell Cycle in Diatoms

3.2.2.1 Light

As diatoms are mostly photosynthetic organisms, light is an essential condition for their survival (Depauw et al. 2012). However, both long-term lack of and excess light can be damaging to cells. Several studies linked light perception to cell cycle progression. A cell cycle arrest was demonstrated in both G1 and G2 phase by light deprivation in six diatoms species (Brzezinski et al. 1990). The only species that lacked the G2 light-dependent checkpoint in their analysis was P. tricornutum. Examination of physiological differences between the two cell cycle arrest phases revealed that T. weissflogii arrested in G1 resumed cycling at a much faster rate than cells arrested in G2 phase (Vaulot et al. 1986). The mechanism behind the cell cycle arrest in G2 phase was not elucidated, but the key components involved in the restart of the cell cycle in G1 arrested cells in P. tricornutum were discovered (Huysman et al. 2013). As described above, it was shown that in dark arrested cells, light exposure induces binding of AUREOCHROME1a to the promoter region of diatom specific cyclin dsCYC2 and transcription factor bZIP10. This promotes the expression of dsCYC2 which forms a complex with CDKA1 and presumably drives progression through G1 phase by phosphorylation of target proteins. The bZIP10 transcription factor was shown to both homodimerize and heterodimerize with AUREOCHROME1a, providing another putative regulatory loop. Aureochromes are one of the groups of proteins that are involved in light perception and signaling (Depauw et al. 2012; Kroth et al. 2017). These proteins found in stramenopiles and oomycetes have a specific domain organization: they combine a LOV blue light-sensing domain with a bZIP transcription factor DNA binding domain. The LOV domain changes conformation upon exposure to blue light and allows its dimerization and subsequent association of the dimer with DNA through the bZIP DNA binding domain (Heintz and Schlichting 2016; Kerruth et al. 2014; Banerjee et al. 2016). Proteins from other groups of light receptors such as cryptochrome/photolyase or red/far-red light sensing phytochrome were found in diatoms, but their link to cell cycle progression was not investigated yet (Annunziata et al. 2019; Coesel et al. 2009; Fortunato et al. 2016). Although the shift from darkness to blue light-induced an immediate onset of the cell cycle, it was reported that the shift from red to blue light caused a temporal inhibition of the cell division (Jungandreas et al. 2014). Thus, the relationship between light spectral composition and intensity and diatom cell division appears to be complex. Even more complicated might be the integration of light signaling, cell cycle progression, and cellular metabolism. Inhibition of CDK activity by inhibitor NU 2058 and resulting G1 phase arrest did not influence the rate of photosynthetic activity but resulted in the remodeling of intermediate metabolism and consequently in the accumulation of intracellular storage of lipids (Kim et al. 2017). Further research is needed to clarify the adaptation of various diatom species to diverse light conditions, such as extended periods of near-complete darkness in polar regions and its influence on cell cycle and metabolism.

3.2.2.2 Silicon

Most diatoms display a strong dependence of their cell division on the availability of silicic acid, resulting from the tight co-regulation of silica cell wall synthesis gene expression with cell cycle progression (Hildebrand et al. 2018). Following silicon depletion, a cell cycle arrest in G1, G2/M, or both phases was observed in several diatom species (Brzezinski et al. 1990; Coombs et al. 1967; Darley and Volcani 1969; Olson et al. 1986). Silicon deprivation was found to be an effective method for cell cycle arrest and synchronization after silicon repletion (Thamatrakoln and Hildebrand 2007).

Analysis of genes co-regulated with a T. pseudonana Silaffin (TpSil3), encoding a protein involved in silica cell wall formation, revealed upregulation of three cyclins and one cyclin-dependent kinase (Shrestha et al. 2012). Despite the fact that a cell cycle arrest was not observed in response to a lack of silicon in P. tricornutum, silicon depleted cells grew more slowly under green light and in low temperature (Zhao et al. 2014). Moreover, changes in gene expression were found when comparing cultures with or without silicon, including upregulation of the dsCYC9 cyclin (Huysman et al. 2010; Sapriel et al. 2009). P. tricornutum occurs in several morphotypes, with only the oval morphotype being extensively silicified (De Martino et al. 2007, 2011; Francius et al. 2008; He et al. 2014), raising the question of how silicon depletion influences the growth rates of different morphotypes and the transitions between them.

3.2.2.3 Other Nutrients

Together with light and silica, nutrient availability is an indispensable condition for diatom growth. Nutrients include macronutrients such as nitrogen and phosphorus, micronutrients such as boron, cobalt, copper, iron and manganese, as well as vitamins that they cannot synthesize such as biotin, cobalamin and thiamin. While light and silica arrest the diatom cell cycle at both G1 and G/M phase in most species, reports on cell cycle arrest during nitrogen deprivation were discordant (Brzezinski et al. 1990; Vaulot et al. 1986; Olson et al. 1986; Claquin et al. 2002). Olson and colleagues reported a G1 phase arrest in response to nitrogen deprivation in T. weissflogii, whereas Claquin and colleagues reported a prolonged G2/M phase in cells deprived of nitrogen and phosphorus in T. pseudonana (Olson et al. 1986; Claquin et al. 2002). Profiling of the transcriptomic response of cell cycle-related genes to repletion of nitrogen, phosphate, iron and trace elements in P. tricornutum cultures (Huysman et al. 2010) showed that nitrogen repletion induced reentry into cell division and expression of the early cell cycle genes CYCP6, CYCH1 and hCDK5. As shown by nitrate repletion in darkness, these genes were not directly responding to nitrate but are generally expressed at the beginning of the cell cycle. Phosphate repletion did not have such a positive effect on cell cycle reentry, but specifically induced expression of dsCYC5, dsCYC7 and dsCYC10. Downregulation of two of these cyclins, dsCYC7 and dsCYC10 was confirmed in an independent study on phosphate-depleted cultures (Valenzuela et al. 2012). In contrast, it was found that cyclin B1 is upregulated in cells arrested by nutrient depletion. The dsCYC10 cyclin was also upregulated at low iron conditions (Smith et al. 2016). Comparison of gene expression in iron-limited and iron-replete cultures showed that the growth rate was reduced in iron-limited cultures in comparison with iron-replete cultures, being accompanied by later and longer expression of cell cycle marker genes, possibly indicating that these cultures were entering their cell cycle later and in a less synchronous fashion compared to their iron-replete counterparts. The signaling cascade and components that prevent CDK activity in response to nutrient deprivation are currently unknown. On the other hand, cell cycle progression was shown to influence the expression of genes involved in nutrient uptake. In conditions with sufficient nitrogen concentration in the growth medium, the RNA levels of nitrate transporter gene were cell cycle-regulated in synchronized C. fusiformis cells (Hildebrand and Dahlin 2000), with major peaks in early G1 and S/G2 phases. This suggests that the progression through the cell cycle is responsive to nutrient availability and at the same time genes required for nutrient uptake are regulated in a cell cycle-dependent manner.

3.2.2.4 Biotic Agents

Biotic agents that influence diatom growth and subsequently the regulation of cell cycle include the interaction between bacteria and diatoms, reactions to grazers and other predators, interactions among different diatoms species, as well as within species interaction such as during mating. Although many examples of positive and negative interactions were described in the literature, only a few examples of molecular components in the signaling are known, let alone the link with the cell cycle.

Bacteria and diatoms are ubiquitously present in the waters on earth. Their coexistence results in complex trophic relationships (Amin et al. 2012). Several bacteria secrete compounds that positively enhance diatom cell division activity. A strain-specific interaction was shown between the diatom Pseudo-nitzschia multiseries and a strain of Sulfitobacter (Amin et al. 2015). P. multiseries growth was enhanced by 19–35% in the presence of this bacterial strain. Based on transcriptomics and targeted metabolite analyses, it was shown that P. multiseries produced taurine that was utilized by the Sulfitobacter, and in turn metabolized ammonium produced by bacteria. This interaction was enhanced by tryptophan and hormone indole-3-acetic acid exchange. Treatment of diatom with indole-3-acetic acid alone enhanced cell growth, but to a lesser extent than during co-cultivation with bacteria. This suggests that indole-3-acetic acid can directly enhance P. multiseries growth, but there are presumably other compounds with the same effect produced by Sulfitobacter. The transcriptomic analysis showed enhanced expression of two cyclins in co-cultivation experiments that are possibly responsible for the enhanced growth—CYC2 and CYC8.

Besides positive effects, many negative interactions between bacteria and diatoms were reported. Croceibacter atlanticus was found to have an inhibitory effect on several diatoms species (van Tol et al. 2017). Addition of C. atlanticus to T. pseudonana resulted in a decrease in culture growth and after 6 days of co-cultivation diatom cells with multiple plastids and nuclei were found, as well as cells without a nucleus. This suggests that C. atlanticus induces chloroplast division, polyploidy and eventually nuclear disintegration in T. pseudonana, indicating that the bacteria impacts cell cycle components to induce chloroplast divisions and karyokinesis in absence of cytokinesis. Several other studies reported beneficial and antagonistic effects of co-cultivation with bacteria on diatom growth, but whether it impacts the cycle components directly or it is an indirect effect through changes in metabolism is not clear (Amin et al. 2009; Cirri et al. 2018; Durham et al. 2015; Foster et al. 2011; Haines and Guillard 1974; Paul and Pohnert 2011; Stock et al. 2019).

Interactions among diatoms species are similarly complex and involve interaction with environment and competition for resources. Examples of both positive and negative influences of diatoms species on growth were found (Jong et al. 1984; Metaxas and Lewis 1991). The best understood intra-specific interaction is the influence of mating pheromones on cells of opposite mating type within the same species in two pennate diatoms: P. multistriata and S. robusta (Bilcke et al. 2021a; Basu et al. 2017; Moeys et al. 2016) (See Chap. “Life Cycle Regulation”). Treatment of sexually competent cells with a spent medium of the other mating type induced a G1 cell cycle arrest in most cells in culture, accompanied by a downregulation of several cyclin genes and upregulation of meiotic genes, including a putative sexual cyclin in S. robusta (Bilcke et al. 2021a; Moeys et al. 2016). This allows cells to switch to meiotic cell division if a mating partner is found, but leaves open the possibility to continue mitotic growth if no contact with mating partner occurs. The identity of the molecular components of the intracellular pheromone signaling pathway and the transduction of the signal to cell cycle components were not yet identified.

4 Future Research

Over the course of this chapter, it has become clear that although specific peculiarities of the diatom cell cycle have been characterized, much of the cell cycle and its regulation has not been studied in detail from a molecular perspective. Nevertheless, a considerable set of tools is available to researchers aspiring to start investigating the diatom cell cycle. These include techniques that have played an important role in historical endeavors into the diatom cell cycle which can be found all throughout this chapter. These methods include cell cycle synchronization, spindle isolation, flow cytometry, and the use of chemical inhibitors and cell wall stains. Additionally, the dawn of new techniques in the last decade—such as large-scale expression profiling and the creation of transgenic fluorescent or knock-out lines – should allow for an accelerated pace of discovery. Promising research topics include the plethora of diatom cyclins and their functional diversification and the response of cell cycle checkpoints to external abiotic and biotic cues. Furthermore, unique cellular and biochemical features of diatom mitosis are awaiting further investigation, such as the behavior of the mitotic spindle and unique MTOCs. To aid in the elucidation of genetic cell cycle pathways, we have listed the major cell cycle genes mentioned throughout this chapter in Table 1. For each candidate gene, the Phatr3 gene identifier is shown alongside the PLAZA Diatoms gene ID. The latter can be used to trace each gene of interest in the PLAZA Diatoms platform for comparative genomics (https://bioinformatics.psb.ugent.be/plaza/versions/plaza_diatoms_01/, Osuna-Cruz et al. 2020) where homologs, gene families, and phylogenetic trees can be retrieved for nine other diatom species.

Table 1 Reference list of P. tricornutum genes with a presumed role in the mitotic cell cycle. Gene identifiers are given in Phatr3 and PLAZA Diatoms 1.0 format
Full size table