Broad active metabolic pathways, autophagy, and antagonistic hormones regulate dinoflagellate cyst dormancy in marine sediments

Abstract

This work aimed to reveal the molecular machinery regulating the dormancy of dinoflagellate resting cysts buried in marine sediments. Dinoflagellates play pivotal roles in marine ecosystems, particularly as major contributors of harmful algal blooms. Despite vital roles of cysts in blooming cycles and dinoflagellate ecology, the molecular processes controlling cyst dormancy have largely remained unexplored due to technological difficulties. Using DinoSL as a dinoflagellates-specific mRNA “hook” and SMRT sequencing, we analyzed metatranscriptomes of sediment-buried dinoflagellate cyst assemblages. The data show that most major metabolic and regulatory pathways, except photosynthesis, were transcriptionally active. This suggests the crucial importance of broad metabolic pathways in sustaining cyst viability and germination potential. Further expression analyses of 11 genes (relevant to autophagy and phytohormone gibberellin), lysosome/autolysosome staining, and germination experiments revealed vital roles of autophagy in energy generation, nutrient recycling, and of phytohormones abscisic acid/gibberellin in modulating dormancy/germination of resting cysts. Our findings lay a cornerstone for elucidating the molecular machinery regulating dinoflagellate cyst dormancy.

Major metabolic pathways and genes are not “resting” in dormant resting cysts of dinoflagellates buried in marine sediment.

INTRODUCTION

As a ubiquitous and ancient group of protists, dinoflagellates are pivotal contributors to coastal marine food webs, global carbon fixation, and O₂ production, as well as essential photosynthetic endosymbionts in coral reefs (1). However, dinoflagellates also include the highest number of species among marine phytoplankton that produce toxins (2), which poison animals and threaten human health through consumption of seafood. As the most common causative agents of harmful algal blooms (HABs) worldwide, nearly 200, out of more than 2900 (3), dinoflagellate species account for ~75% of HABs events (4) and ~40% of all HABs-causing algae (5). In addition to the toxicity to marine animals and humans, dinoflagellate HABs also cause other negative effects such as alterations of plankton community structure and reduction of the resource use efficiency of plankton community (6). Because of the increasingly observed frequency, intensity, temporospatial scale, and economic losses associated with HABs over the past few decades (7), as well as coral bleaching due to global warming (8), the biology and ecology of dinoflagellates have thus drawn great attention from the scientific community (1, 2, 5, 7).

Many organisms have dormant stages in their life cycles, which help them to cope with fluctuating environment and survive temporarily unsuitable conditions (9, 10). The dormant stage of dinoflagellates is well known as resting cysts, which are capable of maintaining dormancy in marine sediments for up to 1.5 centuries. This makes it possible for the resting population of phytoplankton to endure periods of darkness far exceeding those estimated for the Cretaceous-Paleogene extinction, allowing them to rapidly resurge primary production after prolonged catastrophic darkness (11, 12). Resting cysts play a central role in the ecology of dinoflagellates, in particular HABs-causative species (13–15), as encystment involves sex that creates genetic diversity (16), modulates initiation and termination of blooms (14, 15), and provides resistance to adverse environments (13, 16) and protection from pathogen and grazers attacks (16). Resting cysts can also maintain long-term survival in sediments until an internal clock or favorable environmental conditions trigger their germination (15, 17). Hence, the cyst plays vital roles on the geographic expansion of populations via natural or artificial vectors such as ships’ ballast water and sediment (16). Therefore, the abundance and distribution of resting cysts in marine sediments have been used to empirically predict the intensity and extent of forthcoming blooms (18–20). Although only about 10% of the 2900 described species of dinoflagellates (3) have been confirmed to form resting cysts, almost all of the notorious HABs-causing species have been discovered to produce resting cysts, and the number has been increasing over the past decade (13, 21–23). In addition, dinoflagellate cysts serve as excellent historical records and paleoecological indicators for environmental changes, and they have been used to retrospectively track changes in global climate, sea ice coverage, industrial pollution, coastal eutrophication, and human impacts on the environment (24).

Despite the obvious importance of resting cysts as outlined above, the molecular and physiological processes underlying cyst behaviors have hardly been explored. This is at least partly due to the lack of specific tools for in situ measurements of these processes or their corresponding molecular machinery, e.g., (25, 26) and difficulties in efficiently obtaining enough resting cysts in the laboratory or the field. Recently, on the basis of a transcriptomic and physiological study on Scrippsiella acuminata [=S. trochoidea; (27)], we observed a pause of photosynthesis and a possible involvement of the phytohormone abscisic acid (ABA) in the dormancy of newly formed resting cysts (28). Recently, a pioneering gene-knockdown study demonstrated the importance of cellulose synthesis for cellulosic thecal plate formation and efficient cyst-to-swarmer transition (29), providing a means to explore the functions of targeted genes in regulating life history stage transition and cyst dormancy in the laboratory. However, since the laboratory-controlled culture conditions differ from those in realistic marine sediments, and newly formed cysts are far different from those at the dormant stage in the field, the molecular mechanisms regulating dormancy, viability maintenance, and germination of dinoflagellate resting cysts in natural sediments remain a “black box.”

Dinoflagellates have huge genomes and exhibit many genomic and cytological peculiarities that make the whole-genome sequencing of most dinoflagellates highly challenging (30, 31). To date, high-quality dinoflagellate genome data are restricted to species with small genome sizes [0.12 to 4.8 gigabase pairs (Gbp)] species [(31) and the references therein]. As a result, our knowledge of the molecular processes or genetics of dinoflagellates has been largely generated from single-gene studies and transcriptomic analyses of individual species. On the other hand, metatranscriptomics has been proven advantageous, as it is free of culture-associated artifacts and useful for tracking metabolic activities and physiological responses of field populations (25, 32). To differentiate the transcriptomes of dinoflagellate cyst assemblages from those of potentially numerous other organisms in environmental samples, a dinoflagellate-specific, 22-nt-long spliced leader (DinoSL) at the 5′-end of mRNA provides a useful tool. DinoSL has been verified to be ubiquitous in the transcriptomes of all tested dinoflagellates (33, 34), offering a selective and efficient “hook” to track the nucleus-encoded gene transcripts specific to dinoflagellates, distinguishing them from those of other organisms (35, 36). Therefore, DinoSL affords a useful tool for unraveling the molecular processes within the resting cyst black box, enabling comprehensive sequencing of all actively transcribed genes in dinoflagellate cysts during dormancy, which are viewed as transcriptionally “active” genes in the context.

Full-length transcriptome sequencing offers a practical solution to decipher the physiological mechanisms without the reference genome. The single-molecule real-time (SMRT) sequencing, a third-generation sequencing technique developed by Pacific Biosciences (PacBio), allows direct sequencing of full-length transcripts (FLTs) with high accuracy (37, 38). After evaluating the feasibility of different technologies, we hypothesized that combining DinoSL, SMRT sequencing, and metatranscriptomic analyses could yield fundamental insights into the molecular and physiological processes underlying cyst behaviors. In this study, we applied this integrated approach to investigate the overall patterns of gene expression in resting cyst assemblages isolated from three marine sediment samples. Informed by the metatranscriptomic data thus obtained, we used S. acuminata as a representative of cyst-forming and HABs-causing dinoflagellate to generate full-length cDNA sequences of eight autophagy-related genes (ATGs) and three genes involved in the biosynthesis and catabolism of the phytohormone gibberellins (GAs) to gain further insights into the molecular and physiological processes in dinoflagellate resting cysts. The results of quantitative polymerase chain reaction (qPCR), along with direct fluorescent staining of lysosomes/autolysosomes, revealed potentially vital roles played by autophagy and GA during resting cyst dormancy. The findings and the dataset obtained in this study provide a key steppingstone toward an explicit understanding of the molecular mechanisms governing life cycle transitions and cyst dormancy in dinoflagellates, one of the most fundamental processes determining their ecology and role in HABs.

RESULTS

Retrieval of a large transcriptome for sediment-buried resting cysts

SMRT transcriptome sequencing generated 705,433 raw reads, yielding 11.87-Gb raw data [data S1; deposited in the Sequence Read Archive (SRA) database at National Center for Biotechnology Information (NCBI) with the accession number PRJNA699390]. After quality filtering, 45,907, 77,720, and 35,804 high-quality FLTs were obtained from the three samples processed (JZB, JZBC, and TJ, respectively) (data S1). Combining the three field-derived DinoSL-based cDNA libraries and collapsing redundancies led to 159,431 nonredundant (Nr) high-quality FLTs, with a N50 length of 1539 bp and an average size of 1417 bp. For convenience and clarity, these FLTs will be termed “SMRT genes” hereafter.

Functional annotation indicated matches of 159,421 of the SMRT genes with known genes in the databases (data S1). BLAST against the NCBI Nr protein database showed that more than half of the SMRT genes were best aligned to genes of dinoflagellates or recently sequenced microalgae (fig. S1). The top-hit species included the dinoflagellates Symbiodinium microadriaticum (15.44%), Karlodinium veneficum (4.39%), Heterocapsa triquetra (3.62%), Pfiesteria piscicida (2.51%), Heterocapsa rotundata (1.85%), Prorocentrum minimum (1.81%), Lingulodinium polyedra (1.56%), Crypthecodinium cohnii (1.49%), Alexandrium fundyense (=A. catenella; 1.49%), and Amphidinium carterae (1.48%), which together contributed up to 35.64% of the entire Nr assignments (fig. S1). Another 15.46% matched genes from Aureococcus anophagefferens (4.96%), Emiliania huxleyi (2.64%), Guillardia theta (2.54%), Volvox sp. (2.07%), Ectocarpus siliculosus (1.89%), and Thalassiosira pseudonana (1.36%), because these genes have not been characterized in dinoflagellates and hence are not represented in the Nr database. When we aligned the SMART genes against full-length transcriptome data from cultured S. acuminata vegetative cells and newly formed resting cysts (NCBI accession number: PRJNA1127753), nearly 95% of the SMRT genes had significant matches (E values of <10⁻⁵). These analysis results together indicate that the DinoSL-based approach has allowed us to specifically retrieve dinoflagellate gene transcripts from the sediment-buried cysts.

A total of 3292 SMRT genes were shared among all three libraries sequenced (fig. S2): JZB from Jiaozhou Bay, using mRNA extracted immediately; JZBC from Jiaozhou Bay, using mRNA extracted after samples having been stored at 4°C in darkness for 2 years; and TJ from Tianjin, using mRNA extracted immediately. The common genes were identified as sequences matching the same entry (according to accession number) in the Nr database. The two JZB libraries shared 6011 common SMRT genes but only 4790 (JZB) or 4707 (JZBC) with TJ (fig. S2), indicating more similar cyst compositions between the two JZB samples from the same ecosystem. The low number of common SMRT genes (i.e., 3292) shared by all three samples were most likely due to differences in cyst species composition and their respective abundances and the transcript levels of genes among samples.

Most major metabolic and regulatory pathways represented in the metatranscriptome

To further characterize the metabolic pathways in resting cysts buried in sediments, we mapped the SMRT genes to the respective pathways of our full-length transcriptome sequencing data of the dinoflagellate S. acuminata (NCBI accession number: PRJNA1127753), which were generated from vegetative cells and newly formed resting cysts. The metatranscriptome of the resting cyst assemblages covered all major metabolic and regulatory pathways (Fig. 1), except for photosynthesis (fig. S3). These included the metabolism of carbohydrates, lipids, nucleotides, amino acids, and energy-related components (table S1). Six of the major carbohydrate metabolic processes (>80% of their components), including glycolysis/gluconeogenesis, tricarboxylic acid (TCA) cycle, pentose phosphate pathway, fructose and mannose metabolism, starch and sucrose metabolism, and glyoxylate and dicarboxylate metabolism, were among the expressed pathways (table S1). We also retrieved over 80% of the gene repertoires contributing to fatty acid metabolism (table S1) and over 90% of the genes involved in biosynthetic and catabolic pathways of purine and pyrimidine nucleotides (table S1). Amino acid metabolism was also transcriptionally active. Our SMRT genes covered the pathways for cysteine and methionine metabolism; glycine, serine, and threonine metabolism; valine, leucine, and isoleucine degradation; arginine and proline metabolism; lysine degradation; alanine, aspartate, and glutamate metabolism; tryptophan metabolism; tyrosine metabolism; phenylalanine metabolism; histidine metabolism; d-arginine and d-ornithine metabolism; taurine and hypotaurine metabolism; selenocompound metabolism; cyanoamino acid metabolism; β-alanine metabolism; and glutathione metabolism (table S1). Moreover, we also found 3480 SMRT genes encoding energy-related components, including subunits of multisubunit membrane-protein complexes, cytochrome c oxidase, ubiquinol-cytochrome c reductase, and reduced form nicotinamide adenine dinucleotide dehydrogenase (table S1).

Fig. 1. Transcriptionally active pathways in dinoflagellate resting cyst assemblages from marine sediments.

(A) Metabolic circuit map constructed from the metatranscriptome of dinoflagellate resting cysts persistence in marine sediments. (B) Schematic summary of transcriptionally active pathways revealed by our metatranscriptomic analyses. Resting cysts buried in natural scenarios persist until either germination or die due to energy exhaustion, aging, predation, or decay. Energy generation, adjustment of gene transcription and translation, repair, and protection of proteins and nucleic acids are therefore key endogenous determinants of resting cysts that contribute to persistence. The arrow indicates induction and the blunt end stands for repression. In green/blue are factors that positively/negatively influence resting stage persistence.

The SMRT genes were also highly enriched in key regulatory pathways, including gene transcription, translation, DNA replication and repair, protein folding, sorting, and degradation (table S2). A set of genes associated with transcription and translation in the nucleolus, nucleus, nuclear pore, cytoplasm, and chloroplast/mitochondrion were identified, covering all members in the categories of basal transcription factors and spliceosome, and a large majority of the components associated with the machinery assembly of RNA polymerase, RNA transport, mRNA surveillance pathway, ribosome, ribosome biogenesis in eukaryotes, and aminoacyl-tRNA biosynthesis (table S2). A total of 1833 genes relevant to replication and repair were identified, with functions in DNA replication, nucleotide excision repair, base excision repair, mismatch repair, homologous recombination, and nonhomologous end-joining (table S2). We also detected 12,219 genes encoding proteins or enzymes involved in cellular processes such as folding, sorting, and degradation. The most notable processes included protein processing in endoplasmic reticulum, ubiquitin-mediated proteolysis, proteasome, RNA degradation, protein export, soluble N-ethylmaleimide–sensitive factor attachment protein receptor interactions in vesicular transport, and sulfur relay system (table S2).

Furthermore, our SMRT genes also covered intracellular transport and catabolism, including at least 90% of the genes involved in phagosome, endocytosis, peroxisome and autophagy (table S3). A total of 4238 SMRT genes exhibited functions in environmental adaptation, including stress response, antioxidative response, pathogens resistance, reactive oxygen species scavenging, and the glutathione-ascorbate cycle (table S4). We also identified 511 SMRT genes involved in meiosis [including six meiosis-specific and 17 meiosis-related genes; (39)], 97 SMRT genes encoding cyclins, and 80 SMRT genes encoding cyclin-dependent kinases (table S5).

Autophagy transcriptionally active in sediment-buried resting cysts

From our SMRT dataset, we identified 599 SMRT genes involved in autophagy process, 193 of which were annotated as ATGs (table S3). These genes covered the core machinery of autophagy that is conserved from yeasts to human (40). Fourteen of these genes, ATG1, ATG17, ATG18, ATG13, ATG27 of the ATG9 cycling system, ATG6 and ATG14 in the phosphatidylinositol 3-kinase complex, and ATG3, ATG7, ATG4, ATG8, ATG12, ATG16, and ATG5 in the ubiquitin-like protein conjugation systems, matched those in the classical ATGs in yeast (41, 42). In addition, we also identified genes encoding lysosomal marker enzymes, including 51 acid phosphatase, four β-glucuronidase, and five N-acetyl-β-hexosaminidase (table S3).

Transcriptionally active autophagy confirmed in laboratory-induced cysts in the representative cyst-producing dinoflagellate species

To confirm and further explore autophagy activity during dormancy of dinoflagellate resting cysts, we conducted a series of investigations using the tractable cyst-producing dinoflagellate, S. acuminata. First, we obtained the full-length cDNA sequences of eight ATG genes from S. acuminata based on our previous transcriptomic library of this species (28), including ATG1 in the ATG9 cycling system; ATG6 in the PI3K complex; and ATG3, ATG4, ATG5, ATG7, ATG8, and ATG12 in the ubiquitin-like protein conjugation systems (refer to table S6 for the details of these eight genes and their GenBank accession numbers). Seven phylogenetic trees based on the deduced amino acid sequences of ATG12, ATG3, ATG4, ATG5, ATG6, ATG7, and ATG8 obtained in our study, as well as those from alveolates, algae, higher plants, animals, and fungi, demonstrated a tight clustering between the ATG genes from S. acuminata and those from our metatranscriptomes, which formed a sister group to the clades of other alveolate species (figs. S4 to S10). The ATG1 tree is not included here due to a limited number of ATG1 sequences in GenBank, but blast analysis yielded the top hit from Symbiodinium natans ATG1 (CAE7228241.1). These results verify that the ATG genes recovered from the sediment originated from dinoflagellates.

Second, we tracked the transcription profiles of the eight generated ATGs from S. acuminata via qPCR in vegetative cells at different growth stages and in resting cysts with different durations of dormancy and different storage temperatures (data S2). Regardless of whether the cysts were maintained under routine culturing conditions (21°C with a light:dark cycle of 12:12 hours) or stored at 4 ± 1°C in darkness, all the eight ATGs showed significantly higher transcription in resting cysts (newly formed and time series–stored cysts) than in vegetative cells (at exponential and stationary growth) [analysis of variance (ANOVA), P < 0.01 in Fig. 2A; fig. S11). Their transcription in resting cysts consistently displayed an “increase first then decrease” trend within 6 months (Fig. 2A and fig. S11). Compared to cysts stored under routine culturing conditions, the expression of all the eight ATGs in cysts stored at 4°C in darkness exhibited slower declining trends within the 6 months of storage and maintained relatively higher transcription levels up to the end of experiment (Fig. 2A), suggesting active autophagy in resting cysts, with its intensity influenced by the length of dormancy and temperature.

Fig. 2. Transcription profiles of ATGs and autophagy activity in vegetative cells and resting cysts of S. acuminata.

Pink bars/dots show vegetative cells, newly formed cysts and resting cysts maintained at routine culturing conditions for different periods of incubation time (1 to 6 month). Blue bars/dots show vegetative cells, newly formed cysts and resting cysts stored at 4° ± 1°C in darkness for different periods of time (1 to 6 month). (A) Transcription profiles of ATGs. The Y axis shows the transcription levels of targeted gene relative to the reference gene. Error bars represent the standard deviations from triplicate replicates. Asterisk indicates a significant difference between the two samples according to T test in an ANOVA at P < 0.05. (B) Epifluorescence micrographs of vegetative cells/resting cysts those were stained by the LysoTracker and LysoSensor for showing the intracellular location of lysosomes/autolysosomes. (a and b) The same vegetative cell in bright-field and stained by LysoSensor. Note the vegetative cell deformed during the observation. (c and d) The same resting cyst in bright-field and stained by LysoSensor. (e and f) The same resting cyst in bright-field and stained by LysoSensor. (g and h) The same resting cyst in bright-field and stained by LysoTracker. The lysosomes/autolysosomes are indicated by arrows. Scale bars, 10 μm for (a) to (f) and 50 μm for (g) and (h). (C) The percent of positively stained cells/cysts (PPC%) of S. acuminata at different life cycle stages. Error bars indicate ± SDs of n = 6. VC1, vegetative cells at exponential growth stage; VC2, vegetative cells at stationary growth stage; NFC, newly formed cysts.

Active autophagy during cyst dormancy visualized in cysts with fluorescent staining

Autophagy occurs in autophagosome, which fuses with the lysosome/vacuole to form the autolysosome, where the autophagic degradation is promoted by resident acid hydrolases (40). Hence, labeling lysosomes/autolysosomes with fluorescent probes (or sensors) such as LysoTracker and LysoSensor serves as a proxy of autophagy activity. The labeled lysosomes/autolysosomes in vegetative cells or cysts emitted blue fluorescence, which was observed in the stained cells (Fig. 2B, arrows). Using the percentage of positively stained cells/cysts (PPC%) to assess the level of autophagic activity in the population, the labeling results indicated markedly higher autophagy activity in resting cysts of S. acuminata (up to 23%) than in the vegetative cells at both exponential and stationary growth stages (~1%) (Fig. 2C). The PPC% of cysts maintained under routine culturing conditions (21°C with a light:dark cycle of 12:12 hours) continually increased with the duration of storage, reaching the peak at 3 months (51%) before significantly declining to ~30% at 4 to 6 months, which was still slightly higher than that in newly formed cysts (Fig. 2C). However, the PPC% of cysts stored at 4° ± 1°C in darkness was significantly lower than those at 21°C at all time points from 1 to 6 months. It started at 31% at 1 month and decreased gradually, although slowly, over time, reaching its lowest level by the end of experiment at 6 months (Fig. 2C). A two-way ANOVA showed that temperature, time of dormancy, and the interaction between temperature and time all significantly affected PPC% of cysts (P < 0.0001 for all three factors). These staining results together provide independent evidence for active autophagy in resting cysts, with the intensity of activity varying with the duration of dormancy and temperature in a manner similar to that observed in the transcription of ATGs described above.

Biosynthesis, catabolism, and signal transduction of phytohormones transcriptionally active in resting cyst assemblages

From the metatranscriptome obtained from resting cyst assemblages of marine sediments, we identified 947 SMRT genes that are associated with biosynthesis, catabolism, and signal transduction of nine classes of phytohormones, including the five classical hormones: auxin, ABA, cytokinin (CK), ethylene (ET), and GA, as well as four other well-known hormones: brassinosteroid (BR), jasmonic acid (JA), salicylic acid (SA), and melatonin (MEL) (table S7). Among these, 420 SMRT genes were homologous to essential components in signaling pathways (table S7). The remaining 527 genes were related to biosynthetic or catabolic pathways of these phytohormones, covering key genes in the biosynthetic pathways of auxin (flavin monooxygenase, catalyzing the rate-determining step), ABA (zeaxanthin epoxidase, responsible for the first committed step), GA (gibberellin 20-oxidase, catalyzing the rate-determining step; gibberellin 3-oxidase, responsible for the final step), CK (cytokinin hydroxylase, responsible for a key step), ET (ACC synthase, catalyzing the rate-determining step; ACC oxidase, responsible for the final step), JA (12-oxophytodienoate reductase 3, responsible for a key step), SA (isochorismate synthase, catalyzing the first and rate-limiting step), BlR (steroid 22-alpha-hydroxylase, catalyzing the rate-determining step), and MEL (caffeic acid O-methyltransferase, responsible for the final step) [table S7; (43–45)]. In addition, we identified the foremost genes in the catabolic pathways of ABA (ABA 8′-hydroxylase), GA (gibberellin 2-oxidase), MEL (indoleamine 2, 3-dioxygenase), and CK (cytokinin glycosyltransferase) [table S7; (43–44)].

Exogenous ABA and GA influence germination of natural resting cysts and GA biosynthesis and catabolism are transcriptionally active in deep-dormant cysts of S. Acuminata

Because of the well-established antagonistic interaction between ABA and GA in higher plants (43–45), we added different doses of ABA, GA3, and their combinations to resting cysts assemblages concentrated from marine sediment to test the possible effect of exogenous ABA and GA on the germination of resting cysts (data S3). We focused on the total germination rate (%) of resting cysts without identifying individual cyst species. While ABA at concentrations above 1000 μM exhibited a significant inhibitory effect on cyst germination (data S3), GA3 exhibited a significant enhancing effect on the total germination rate in a dose-dependent manner within the concentration range of 1.0 and 10 μM (data S3). The two phytohormones also exhibited an antagonistic interactive effect on cyst germination, as GA3 alleviated the ABA-induced inhibitory effect on cyst germination, and ABA also counteracted the GA3-induced enhancing effects (see data S3 for more details).

In a previous study, we found and investigated ABA in both vegetative cells and resting cysts of S. acuminata (28). Here, we focused on the genes regulating GA biosynthesis and catabolism (table S6 and data S2). The qPCR results indicated down-regulated GA biosynthesis (GA20ox, gibberellin 20-oxidase, regulating the rate-limiting step of GA synthesis; and GA3ox, gibberellin 3-oxidase, responsible for the final step of GA synthesis) during cyst formation and dormancy, and an up-regulated GA catabolism (GA2ox, gibberellin 2-oxidase, catalyzing the rate-determining step of GA catabolism) during dormancy maintenance (Fig. 3 and fig. S12), with both trends intensified at lower temperature (4 ± 1°C) in the darkness (Fig. 3).

Fig. 3. Transcriptional profiles of genes involved in GA biosynthesis and catabolism in vegetative cells and resting cysts of S. acuminata.

Orange bars show vegetative cells, newly formed cysts, and resting cysts maintained at routine culturing conditions for different periods of incubation time (1 to 6 month). Green bars show vegetative cells, newly formed cysts, and resting cysts stored at 4° ± 1°C in darkness for different periods of time (1 to 6 months). The Y axis shows the transcription levels of targeted gene relative to the reference gene. Error bars represent the SDs from triplicate replicates. Asterisk indicates a significant difference between the two samples according to t test in an ANOVA at P < 0.05.

DISCUSSION

Transcriptional evidence of broad active metabolic landscape during dormancy of dinoflagellate resting cysts

With DinoSL as a “hook” of dinoflagellate mRNAs and third-generation sequencing, we successfully created a full-length transcriptomic dataset for sediment-buried dinoflagellate resting cyst assemblages. Until now, dinoflagellate cyst transcriptomes have only been generated for laboratory-induced cysts (28, 46, 47). The metabolic processes represented in the metatranscriptomes are notably extensive for all three field-collected resting cyst assemblages. Notably, one of the sediment samples had been stored at 4°C for 2 years after collection before RNA extraction. As energy stored in the macromolecules in a cyst is limited, cysts must efficiently budget their metabolism to sustain the long duration of dormancy, and conceivably, no unnecessary metabolic activity would be turned on. Therefore, the broad active metabolic landscape detected in the sediment-buried cyst assemblages indicates a vital aspect of the molecular machinery needed for the dormancy maintenance of cysts.

In higher plants, metabolic reduction occurs during seed dormancy, especially at the dormancy maintenance phase (48). In hibernating animals, a general metabolic arrest occurs during hibernation, in which metabolic rate can be maximally reduced to only 1% of the minimum euthermic level (49). Broad lineages of phytoplankton, such as dinoflagellates (resting cyst), diatoms (resting spore), cyanobacteria (akinete), green algae (akinete or oospore), chrysophytes (stomatocyst), raphidophytes, haptophytes, cryptophytes, and euglenophytes, have a resting stage in their life cycles and their metabolism level in the early resting stage is approximately 10% of that in the active growth stage [as reviewed in (9)]. Transcriptomic analysis also indicates that the cells at resting stage typically exhibit different degrees of metabolic suppression [e.g., dinoflagellates (28), diatoms (50), green algae (51), and cyanobacteria (52)]. Dinoflagellates are metabolically similar to those “dormant” stages of the abovementioned different organisms. Almost all previous studies on cysts of dinoflagellates revealed that the respiration rate is extremely low or even below detection limit (28, 53–55). The respiration rate of S. acuminata was reported to be ~10% of that in vegetative cells during the initial dormancy period and ~1.5% in quiescent cysts (53) and that in temporary cysts of Scrippsiella hangoei under darkness was almost undetectable (54). The energy generation pathways, such as photosynthesis, glycolysis, and TCA cycle, were also found to be largely suppressed in resting and/or pellicle cysts of S. acuminata than those in vegetative cells (28, 46, 47, 55). Consistently, a much lower cellular adenosine triphosphate content was measured in pellicles (47) and resting cysts (55), respectively, than that in vegetative cells. These studies together implied a general reduction of metabolic activity in the resting stage of dinoflagellates, and most of the genes seemingly should be “at rest” or “dormant” in resting cysts, as the term “resting cyst” implies. Therefore, the most conspicuous and unexpected finding of this study is that the vast majority of SMRT genes encode for all major metabolic and regulatory pathways, except for photosynthesis, and maintained active (expression at the transcriptional level) in the natural assemblage of dinoflagellate resting cysts. In the future, we will further test what have been obtained in this work with an approach combining transcriptomic, proteomic, and metabolomic investigation and more cyst-producing dinoflagellates, with a focus on the pathways relevant to energetic metabolism and phytohormone regulation.

Although it is generally recognized that the persistence of resting stages of phytoplankton is influenced by both inherent characteristics and external environment (9), our current knowledge about dormant stage maintenance of dinoflagellates is very limited. The resting cysts of dinoflagellates have many structural, functional, and adaptive features analogous to the seeds of higher plants, including a lengthy viability (9, 13, 56). On the basis of our results and by referring to seed persistence in terrestrial environments, we hypothesized that there might be a set of “active” genes and pathways that are vital or even essential to maintain viability and germination potency of all species of dinoflagellate resting cysts in the field, at least including those encoding for energy generation, the adjustment of gene transcription and translation, and the repairment and protection of proteins and nucleic acids. As seen from the number of SMRT genes above, this set of “active” genes was rather large, notably contrasting to an intuitional estimate as implied in the terms “resting” and “dormancy.” Given the dominance of post-transcriptional regulation in dinoflagellates (30, 57), the large number of SMRT genes and their inclusion of all major metabolisms and pathways may simply imply a mechanism adopted by resting cysts to maintain viability and save energy consumption: constantly maintaining all transcripts needed for survival and initiation of germination at a level as low as possible in general. This may be made possible through enhancing the stability and half-life of mRNA. Dinoflagellate mRNA has been shown to have longer half-life than other organisms (57). This approach may keep energy cost at a minimal level. Nevertheless, this needs further verification with more intensive (e.g., multiomics), extensive (e.g., more cyst species and dormant conditions), and definitive [e.g., knockdown of cyst-relevant gene(s); (29)] investigations.

Heightened autophagy suggesting an alternate energy generation and cellular resources recycling as a crucial strategy for cysts persistence

Autophagosome is the central organelle where the major processes of autophagy take place, while a set of ATG proteins is hierarchically recruited to form the initial membrane template of autophagosome (40, 42). Since initially identified in budding yeast, over 40 ATG members have been revealed in yeast, filamentous fungi, mammals, and plants (41). A total of 14 ATGs (accounting for 193 SMRT genes) were found to be transcriptionally activated in assemblages of dinoflagellate resting cysts. Our subsequent transcriptional profiling for eight ATGs via qPCR in vegetative cells and resting cysts of S. acuminata provided more convincing evidence for an active autophagy in resting cysts (higher than that in vegetative cells), which was further visually and quantitatively supported by organelle staining. Lysosome, as the major catabolic factory in eukaryotic cells (40, 41), has been previously confirmed to be present in dinoflagellates and termed as PAS/accumulation bodies (58). Once autophagosome fused with lysosome membrane, the interior single-membrane vesicles are released into lysosome lumen to form autolysosome, which is subsequently consumed by the resident acid hydrolases (40, 41). Therefore, the presence and the intensity of blue fluorescence in a cyst stained with the sensors indicate the activity and intensity of autophagy in the cyst (but not phagocytosis). The identification of 51 SMRT genes encoding lysosomal marker enzymes from our metatranscriptomes also supported the sensor-labeling results. Our results showed that the intensity of autophagy in resting cysts of S. acuminata was significantly higher than that in vegetative cells, and the autophagy intensity increased at higher temperatures and during the first few months after encystment.

Autophagy is a highly conserved eukaryotic pathway that contributes to nutrient recycling and cellular homeostasis through degradation of cytoplasmic constituents (40). It has been well established that autophagy functions in a broad range of cellular events and plays a central role in plant responses to biotic and abiotic stresses (40, 41). Autophagy was previously found to be up-regulated during a dinoflagellate bloom caused by Prorocentrum shikokuense under phosphorus depletion and hypothesized to facilitate internal nutrient recycling as an adaptive response strategy (59). This was corroborated by subsequent culture-based study showing autophagy as a cellular response to phosphorus deficiency in (60). In addition, autophagy was shown to be involved in N limitation-induced accumulation of neutral lipid and starch in dinoflagellate P. lima (61). However, compared to yeast and mammals, in silico and experimental investigations into autophagy activity in microalgae remain rare and are completely unexplored in dinoflagellate resting cysts. In this study, we provided a line of evidence for the markedly higher autophagy activity during the resting stage of dinoflagellates, collectively suggesting a vital role of autophagy in maintaining viability and dormancy of dinoflagellate resting cysts by possibly releasing energy or/and recycling intermediate components from macromolecules accumulated before cyst formation.

Potential roles of phytohormones in maintaining dormancy of resting cysts and the antagonistic functions of ABA and GA

Although the bioactive forms of some well-known phytohormones have been found in a broad spectrum of algal lineages, their functional roles in algae, especially in microalgae, have remained elusive [(62, 63) and the references therein]. In terrestrial plants, hormones and their interactions have been well documented as one of the vital internal factors modulating seed dormancy (43, 45). Among them, ABA and GA are widely recognized as two primary hormones that antagonistically regulate this process, while other hormones (e.g., auxin, CK, ET, BR, JA, SA, and MEL) exert their effects by influencing ABA and/or GA (43–45). Although dinoflagellate resting cysts are analogous to seeds of higher plants in having many structural, functional, and adaptive parallels (9, 13, 56), study on the functions of phytohormones in the cyst dormancy of dinoflagellates is still at their infancy (11, 28, 46). On the basis of a transcriptomic profiling and direct measurements of intracellular ABA of the resting cysts and vegetative cells of S. acuminata, we found direct evidence suggesting a possible involvement of ABA in the dormancy maintenance of cysts (28). A recent work on ancient cysts preserved in century-old sediments showed that exogenous addition of MEL and GA could induce the dormancy release (cyst germination) in 11 dinoflagellate taxa, which did not occur under normal culturing conditions. This suggests their roles in promoting dormancy release (germination) of dinoflagellate cysts (11). Given that phytohormones act as endogenous signaling molecules at very low concentrations (43, 63), the enzymes and their encoding genes responsible for biosynthesis and catabolism of hormones are expected to be expressed at relatively low levels. Therefore, the presence of such a variety of phytohormones-related transcripts indicated that the dormancy of resting cysts is regulated by a complex interacting network of many hormones.

Endogenous hormone levels are the result of dynamic balance between synthesis and catabolism (43). Our qPCR detections showed continuous repression of GA biosynthesis and elevated catabolism during dormancy maintenance, implying that endogenous GA levels were much lower in resting cysts than that in vegetative cells, and continuously decreased as dormancy progressed. We previously confirmed that endogenous ABA content remained significantly high during cyst formation and dormancy maintenance (relative to vegetative cells) in S. acuminata (28). Combined with our work on ABA (28), an antagonistic regulation of GA and ABA metabolic genes at transcriptional level was revealed during encystment and cyst dormancy. Low temperature and darkness, two environmental cues typically present in marine sediments, also invoked opposite effects on both biosynthesis and catabolism of GA and ABA at the transcriptional level [(28) and the current work]. These results collectively suggest that high ABA accumulation and low GA levels are associated with deep dormancy of dinoflagellate cysts buried in natural environment and vice versa. Furthermore, experiments with exogenous addition provided strong evidence for the functions of GA and ABA, as well as their antagonistic effects on dormancy promotion and release. Together, we postulate that ABA and GA are two important, but antagonistic, regulatory factors involved in the maintenance of dormancy in dinoflagellate resting cysts.

In summary, contrary to the intuitive notion implied by the term resting cyst, which suggests a physiologically inactive state, this study found that the vast majority of genes involved in major metabolic and regulatory pathways, except for photosynthesis, were transcriptionally active in resting cyst assemblages collected from the field. This suggests a mechanism used by dinoflagellates during dormancy persistence: constantly maintaining all transcripts necessary for survival and germination initiation at minimal levels, unless otherwise required to cope with environmental changes or manage energy consumption. Markedly higher autophagy activity was revealed during the resting stage of dinoflagellates, especially at higher temperatures and during the first few months after cyst formation, highlighting its vital role in releasing energy and recycling cellular resources for cyst persistence. The two classical phytohormones, ABA and GA, played antagonistic roles in modulating cysts dormancy maintenance and release. Low temperature and darkness, the two environmental cues typically observed in marine sediments, elicited opposite effects on both the biosynthesis and catabolism of GA and ABA, leading to high ABA accumulation and low GA levels corresponding to deep dormancy in dinoflagellate cysts buried in natural environments. The findings from this study and the accompanying dataset provide a highly valuable steppingstone toward an explicit understanding of the molecular machinery regulating dinoflagellate resting cyst dormancy, a fundamental process determining the ecology of dinoflagellates and HABs.

MATERIALS AND METHODS

Full-length transcriptome SMRT sequencing of S. acuminata vegetative cells and newly formed resting cysts

The calcareous dinoflagellate S. acuminata [formerly S. trochoidea; (27)] is a cosmopolitan (64) and toxic (65) species that commonly forms dense blooms worldwide [as reviewed in (65)]. Because of its propensity for readily producing resting cysts, S. acuminata has been adopted as representative species for studies on dinoflagellates life history (28, 46, 47, 52, 66–68). In this study, we used the species as a representative of cyst-producing dinoflagellates to obtain its full-length transcriptome of vegetative cells and newly formed resting cysts through SMRT sequencing on the PacBio Sequel platform (NCBI accession number: PRJNA1127753; refer to Supplementary Materials and Methods for more details). This full-length transcriptome of S. acuminata was used as a reference transcriptome to provide a whole transcriptomic background in vegetative cells and newly formed resting cysts of dinoflagellate.

Sediments collection and resting cysts concentration

Two sediment samples were collected from Jiaozhou Bay, Qingdao, Shandong Province, China (36.159°N, 120.229°E; table S8) on 29 April 2014 (JZBC) and 24 May 2016 (JZB), respectively. The third sediment sample was collected from coastal Tianjin, China (38.929°N, 117.939°E; table S8) on 8 June 2016 (TJ). The upper 10 cm of sediment was taken using a grab sampler, transferred to sterile plastic bags, and kept on ice in a cooler to maintain a condition similar to the in situ environment (darkness and low temperature). Upon arrival at the laboratory, the samples tagged as JZB and TJ were immediately processed to concentrate resting cysts. Meanwhile, the sample tagged as JZBC was first preserved at 4° ± 1°C in darkness in a refrigerator for more than 2 years (until May 2016) to allow the cysts to reach deep dormancy and avoid possible interference of the residual vegetative cells and then subjected to cyst concentration.

For cyst concentration, approximately 200 g (wet weight) subsurface sediment (the layer of 3 to 6 cm depth) was aseptically subsampled from each sediment sample and immediately subjected to the resting cyst concentration process following the protocol of sodium polytungstate density-gradient centrifugation as described in (69). All samples were prepared in biological triplicates (labeled as -A, -B, and -C). The concentrated resting cyst assemblages (together with other taxa’s resting spores, eggs, and pollens) were washed several times with fresh sterile filtered (0.22-μm pore-sized filters) seawater, concentrated by centrifugation, and immediately used for total RNA extraction.

Preparation of dinoflagellate-specific e-cDNA libraries and SMRT sequencing

Total RNA was isolated using RNeasy Plant Mini Kit (QIAGEN, Germany) and was treated with RNase-Free DNase Set (QIAGEN, Germany) to remove residual genomic DNA. The quantity and quality of total RNA were assessed with NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, USA) and Agilent Technologies 2100 Bioanalyzer (Agilent Technologies, CA, USA). The dinoflagellate-specific e-cDNA library was generally prepared according to the isoform sequencing protocol using the Clontech SMARTer PCR cDNA Synthesis Kit and the BluePippin Size Selection System protocol as described by PacBio (PN 100-092-800-03) with the following modifications. The first-strand cDNA was prepared from total RNA with the anchor primer (table S9) by using Clontech SMARTer PCR cDNA synthesis kit. Then the first strand cDNA was used as template for PCR with a pair of primers DinoSL-Racer3 (table S9) to amplify the dinoflagellate-specific cDNAs. PCR was performed using AmpliTaq Gold360 Master Mix (Invitrogen, Karlsruhe, Germany) under the following program: 94°C for 3 min, followed by 95°C 15 s, 68°C 3 min 30 s for 5 cycles; and 95°C 15 s, 62°C 30 s, 72°C 3 min for 15 cycles, and a final extension at 72°C for 10 min. Size fractionation and selection were performed using the BluePippin with the following two bins for each sample: 0.5 to 1.8 kb and > 1.8 kb. A total of 18 e-cDNA libraries [three samples (JZBC, JZB, and TJ) × 3 biological replicates × 2 bins] were constructed. Each library underwent SMRT sequencing using one SMRT cell on the PacBio Sequel platform.

The metatranscriptomic data processing and functional annotation of transcripts

The raw reads of SMRT sequencing in the metatranscriptome were filtered and clustered into Circular Consensus Sequences (CCS) by using the SMRT Analysis Server 2.3.0 (PacBio). The CCS reads containing both primers and a poly(A) tail, as well as those not containing any additional copies of the adapter sequences within the DNA fragment, were recognized as full-length, nonchimeric (FLNC) reads. The obtained FLNC reads were clustered using the iterative clustering for error correction algorithm and subsequently polished to yield full-length, high-quality consensus sequences (accuracy ≥99%). The final Nr genes were filtered by removing redundant sequences for subsequent analysis. Functional annotation of each transcript was performed using the BLASTX algorithm against public databases (see Supplementary Materials and Methods for more details).

To further characterize the metabolic pathways in resting cysts buried in sediments, we mapped the metatranscriptomic data to the respective pathways of our full-length transcriptomic sequencing data of S. acuminata (NCBI accession number: PRJNA1127753). The entire metatranscriptomic dataset was BLAST searched against the S. acuminata full-length transcriptomic sequences with a 10⁻⁵ E value cutoff. The metatranscriptomic SMRT genes associated with the major pathways and functions elaborated in this paper were manually reanalyzed using BLAST searches against the GenBank annotated database to verify the functional predictions.

Algal culture of S. acuminata and preparation of vegetative cells and resting cysts

The S. acuminata strain IOCAS-St-1 isolated from the Yellow Sea of China was routinely maintained as previously described in (28). The samples of vegetative cells at different growth stages and resting cysts were prepared also described previously (28, 67) (see Supplementary Materials and Methods for more details). For cloning the target genes, fresh vegetative cells were used for total RNA extraction and template preparation. For qPCR detections and lysosome/autolysosome fluorescence staining (see below), vegetative cells at different growth stages, newly formed resting cysts, and cysts maintained in dormancy for different durations were used in this study (refer to Supplementary Materials and Methods for more details).

Cloning of ATGs and genes involved in GAs biosynthesis and catabolism from S. Acuminata

The full-length cDNA sequences of eight ATGs (ATG1, ATG12, ATG3, ATG4, ATG5, ATG6, ATG7, and ATG8) and three genes involved in GA biosynthesis [GA20ox (gibberellin 20-oxidase) and GA3ox (gibberellin 3-oxidase)] or catabolism [GA2ox (gibberellin 2-oxidase)], from S. acuminata [designated as SaATG1, SaATG12, SaATG3, SaATG4, SaATG5, SaATG6, SaATG7, SaATG8, SaGA20ox, SaGA3ox, and SaGA2ox, respectively] were isolated using specific primers (table S9) based on partial sequences obtained from a previous transcriptomic library of S. acuminata in our laboratory [GenBank accession no. SRP058465; (28)] and using RACE (rapid amplification of cDNA ends) procedure (refer to Supplementary Materials and Methods for more details). For the obtained ATG genes, phylogenetic trees were inferred from amino acid sequence alignments using the Bayesian method [(68); Supplementary Materials and Methods].

Transcriptional profiles of targeted genes at different stages of the life cycle with real-time qPCR

To investigate transcriptional profiles of targeted genes (ATGs and genes involved in GA biosynthesis and catabolism) in response to the alteration of life cycle stages and different treatment conditions of cysts, two arrays of experiments were performed as described before (28) and briefly summarized as follows: The first experiment involved cells harvested at different life cycle stages, including vegetative cells, newly formed resting cysts, and resting cysts that were maintained under culture conditions for 1, 2, 3, 4, 5, and 6 months. The second experiment involved in resting cysts that were kept at 4 ± 1°C in darkness for 0, 1, 2, 3, 4, 5, and 6 months, respectively (see Supplementary Materials and Methods). All samples were prepared in biological triplicates.

Relative standard curves for gene transcripts were generated using serial 10-fold dilutions of cDNA. The qPCR efficiency (E) was determined from the slope of a linear regression model (70) and calculated according to the equation: E = (10^[−1/slope] − 1) × 100 (71). On the basis of the reference gene validation described before (28), the gene UBC (encoding for ubiquitin conjugating enzyme) was used as reference gene in the subsequent qPCR analyses to assess transcription levels of target genes at different life cycle stages and under various treatment conditions. Relative expression levels were analyzed using the 2^−△△Ct method (72) and expressed as dimensionless units normalized with the expression of the reference gene (see Supplementary Materials and Methods for more details). Data were presented in graphs as mean ± SD, subjected to one-way ANOVA and followed by Tukey’s post hoc test. Significance was accepted when P ≤ 0.05.

Lysosome/autolysosome fluorescence staining with sensors

Vegetative cells at exponential and stationary stages, newly formed resting cysts and cysts maintained in dormancy for different durations were used for their lysosome/autolysosome fluorescence staining. The probes LysoTracker (Blue DND-22; wavelengths of absorption and emission, 373 and 422 nm) and LysoSensor (Blue DND-167; wavelengths of absorption and emission, 373 and 425 nm) (Molecular Probes, Invitrogen, Thermo Fisher Scientific, USA) were used to label lysosomes/autolysosomes in the live cells and/or cysts of S. acuminata, based on the principle that after the probes entered the acidic milieu of lysosomes/autolysosome (pH < 6), they emit blue fluorescence (73). This fluorescence is easily distinguishable from the red autofluorescence of chlorophyll and green autofluorescence present in most microalgae (74). Cells/cysts that emit recognizable blue fluorescence were counted under an inverted epifluorescence microscope (IX73, Olympus, Japan). The average value of the counts from the two fluorescent dyes was used to calculate the final PPC%. Data were presented in graphs as mean ± SD and subjected to two-way ANOVAs for significance evaluation of variations (Supplementary Materials and Methods).

Effects of exogenous ABA and GA on germination of resting cysts assemblages concentrated from marine sediment

A sediment sample collected from Jiaozhou Bay, Qingdao, Shandong Province, China (36.159°N, 120.229°E) on September 2020 was used for resting cysts concentration for germination experiments. The sample was first preserved at 4° ± 1°C in darkness in a refrigerator for more than 1 year and then subjected to cyst concentration (performed on December 2021). A total of four arrays of treatments were designed: (i) only ABA; (ii) only GA3 (the major bioactive form GA in plants); (iii) 1500 μM ABA (determined to inhibit cysts germination) combined with a gradient of GA3; (iv) 10 μM GA3 (determined to promote cysts germination) combined with a gradient of ABA (see Supplementary Materials and Methods for more details). All arrays were prepared with biological triplicates. The experiments lasted 20 days, and cyst germination was carefully monitored under an inverted microscope. The germination rate was calculated by dividing the number of germinated empty cysts by the number of initial cysts. Data were presented in graphs as mean ± SD and subjected to two-way ANOVAs for the significance evaluation of variations.

Supplementary Materials

The PDF file includes:

Supplementary Materials and Methods

Figs. S1 to S12

Legends for tables S1 to S9

Data S1 to S3

References

Other Supplementary Material for this manuscript includes the following:

Tables S1 to S9