1 INTRODUCTION

Dinoflagellates constitute one of the main groups of marine phytoplankton in terms of their important contribution to the primary production, and comprise about 2 400 species belonging to 259 genera (Gómez, 2012), which is still growing as more new taxa are to be described (Gómez et al., 2015; Takahashi et al., 2015, 2017, 2019; Boutrup et al., 2017; Luo et al., 2018; Hu et al., 2020b, 2021; Ok et al., 2021; Gu et al., 2022). The appearance of dinoflagellates dates back more than 400 million years in the fossil record, and this group of protists has evolved diverse features to adapt to their environments, including diverse morphologies, multiple-membrane cell walls, different modes of nutrition, pigments, toxins, asexual and sexual reproduction, and other characters (Steidinger and Meave del Castillo, 2018). Among these features, the ability of about 10% of dinoflagellate species to produce resting cysts as a part of their life cycle is a vital one (Head, 1996; Bravo and Figueroa, 2014; Tang et al., 2016, 2021). It is now widely accepted that resting cysts play important roles in the biology and ecology of dinoflagellates (Anderson and Wall, 1978; Dale, 2001; Bravo and Figueroa, 2014; Tang et al., 2016, 2021; Ellegaard and Ribeiro, 2018; Figueroa et al., 2018).

In the past, field studies on dinoflagellate cysts are generally focused on the surface sediments for mapping the distribution and abundance of cysts of important species, investigating species diversity of cyst assemblage in a region of concern, and confirming the cyst presence for some species with particular importance (Luo et al., 2018; Limoges et al., 2020; Liu et al., 2020a, b, 2021; Mertens et al., 2020; Van Nieuwenhove et al., 2020; Hu et al., 2021, 2022), whereas sediment cores representing a valuable archive of phytoplankton communities are usually used to reconstruct past environmental changes (Keafer et al., 1992; Dai et al., 2012; Ellegaard et al., 2013, 2020; Bringué et al., 2016; García-Moreiras et al., 2018; Kim et al., 2018; Price et al., 2018; de Freitas et al., 2020; Li et al., 2021; Siano et al., 2021). Cultures established from reviving dinoflagellate resting cysts collected from sediment cores have been used to evaluate the impact of environmental changes on the physiology, genetic structure, and diversity in various species (Ribeiro et al., 2011; Klouch et al., 2016; Lundholm et al., 2017; Kremp et al., 2018; Delebecq et al., 2020; Ellegaard et al., 2020; Girault et al., 2021). However, the numbers of dinoflagellate cysts remaining viable in long-buried sediments and the mechanisms behind it still need to be explored. Therefore, mining cyst records of dinoflagellate species buried in sediment cores is important in many aspects of the ecology of dinoflagellates and in reconstruction of the history of marine environmental changes.

Recently, we successfully established seven clonal cultures of Biecheleria brevisulcata, Biecheleriopsis adriatica, and Scrippsiella donghaienis via cystgermination from the depth dated back to 1941±18 AD of a sediment core collected from the East China Sea, and further characterized their morphologies, pigment compositions, and the genetic diversity in their LSU rRNA gene and ITS sequences.

2 MATERIAL AND METHOD 2.1 Sediment core and surface sediment sampling and dating

One sediment core (S06-2, 120.417°E, 26.122°N) was collected from the East China Sea in September 19, 2018 during the public cruise of R/V Xiang Yang Hong 18 organized by the National NaturalScience Foundation of China and the First Institute of Oceanography, Ministry of Natural Resources, China, and one surface sediment sample (0‒2 cm; S01-1, 122.997°E, 31.000°N) was also collected from the East China Sea in September 17, 2019 during the public cruise of R/V Xiang Yang Hong 18. The sediment core was sliced into 2-cm layers (the top 20 cm) and 4-cm layers (20 cm to the bottom) by caution. A total of 16 subsamples were collected for the 210Pb and 137Cs measurements. The detailed measurement and age determination was reported in Liu et al. (2021), and the standard error of age was produced in regression uncertainties.

2.2 Culture establishment

Cyst assemblage in the subsample (42‒44 cm) of core sediment (S06-2) and surface sediment sample (S01-1) was concentrated using sodium polytungstate solution (SPT) (Bolch, 1997). For subsample of core sediment, single cysts were washed at least three times using sterile seawater (with a salinity of 31) enriched with f/2-Si medium (Guillard, 1975), and then micropipetted to a 24-well culture plate with each well containing 2.5-mL fresh medium and 2% antibiotic solution (a mixture of 10 000-IU penicillin and 10 000-μg/mL streptomycin; Solarbio, Beijing, China). The rest of cyst assemblage was transferred to a 6-well culture plate with each well containing 10-mL fresh medium and 2% antibiotic solution. The plates were incubated at 21 ℃, 12-h꞉12-h light꞉dark cycle, and ~100 μmol photons/(m2·s). Cysts and new germlings were observed every day or every other day with an inverted microscope (IX73, Olympus, Japan) and photographed by a DP80 digital camera (Olympus, Japan). Five strains of B. brevisulcata (S1, S2, S3, S4, and S5) were established from the cyst assemblage germination experiment, but the resting and empty cysts of B. brevisulcata were not observed. One strain of Bps. adriatica (S21) and one strain of S. donghaienis (S23) were established from singlecysts germination experiments, their resting and empty cysts were clearly recorded using an inverted microscope (IX73, Olympus, Japan) equipped with a DP80 digital camera (Olympus, Japan). All cultures were routinely maintained in the same condition mentioned above.

2.3 Light microscopic observation

Live cells of B. brevisulcata (strain S1), Bps. adriatica (strain S21), and S. donghaienis (strain S23) were observed and photographed using a Zeiss Imager Z2 (Carl Zeiss, Gottingen, Germany) equipped with differential interference contrast (DIC), or an inverted microscope (IX73, Olympus, Japan) equipped with a digital camera (DP80, Olympus, Japan). For observation of thecal plates of S. donghaienis, live cells were stained with Calcofluor White (Sigma-Aldrich, St. Louis, MO, USA) and examined using an epifluorescence microscope (BX53, Olympus, Japan) with a UV filter set (Fritz and Triemer, 1985). Cells sizes of B. brevisulcata, Bps. adriatica, and S. donghaienis for 50 live cells at the mid-exponential growth phase were measured at ×400 (for B. brevisulcata and Bps. adriatica), and ×200 (for S. donghaienis) magnification using a DP80 digitalcamera (Olympus, Tokyo, Japan).

2.4 SEM observation

For SEM observation, vegetative cells of B. brevisulcata and Bps. adriatica at mid-exponentialgrowth stage were fixed with OsO₄ (2% final concentration), and S.donghaieniswith glutaraldehyde (2.5% final concentration) for 40–50 min. Fixed cells were gently filtered onto 5-μm (B. brevisulcata and Bps. adriatica) and 11-μm (S. donghaienis) pore sizeMillipore nylon membranes, dehydrated in an acetone series (10%, 30%, 50%, 70%, 90%, and three times in 100%, 15 min for each step), and critical point-dried with liquid CO₂ (EM CPD300, Leica, Austria). They were sputter-coated with platinum-palladium (EM ACE200, Leica, Austria), and observed using an S-3400N SEM (Hitachi, Japan).

2.5 Pigment analyses

Fifty milliliter of each culture of B. brevisulcata (strain S1), Bps. adriatica (strain S21), and S. donghaienis (strain S23) in exponential growth(in cell densities of ca. 102 150, 95 900, and 980 cells/mL, respectively) were filtered through 25-mm diameter glass fiber filter (Whatman, Maidstone, UK) and immediately frozen at -80 ℃ for later analyses. Pigments were analyzed on an Alliance HPLC (e2695, Waters, Milford, Massachusetts, USA) using a 100-μL sample injection according to Kong et al. (2012) and Hu et al. (2020a). Pigments were identified and quantified using Shimadzu Class-VP software and by comparing pigment spectra and retention times with those of 26 standard pigments (DHI Water and Environment, H rsholm, Denmark; Kong et al., 2012; Hu et al., 2020a).

2.6 DNA extraction, PCR amplification, and rDNA-based phylogenetic analyses

Genomic DNA of B. brevisulcata, Bps. adriatica, and S. donghaienis were extracted using a plant DNA extraction kit (Tiangen, Beijing, China) according to the manufacturer's protocol. For B. brevisulcata, Bps. adriatica, and S. donghaienis, about 1 400 bpof LSU rDNA were amplified using primers of D1R (forward, 5′-ACCCGCTGAATTTAAGCATA-3′) (Scholin et al., 1994) and 28-1483R (reverse, 5′-GCTACTACCACCAAGATCTGC-3′) (Daugbjerg et al., 2000), and for S. donghaienis, about 660 bp of ITS was amplified using primers of ITS1 (forward, 5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (reverse, 5′-GCATATCAATAAGCGGAGGA-3′) (White et al., 1990). Polymerase chain reaction (PCR) reactions were conducted using a PCR Master Cycler nexus gradient (Eppendorf, Hamburg, Germany), and performed with a final volume of 25 μL, containing 9.5-μL ddH₂O, 12.5-μL 2×Taq PCR MasterMix, 1 μL of each PCR primer (10 mmol/L), and 1 μL of the DNA template. The following cycling conditions were used: an initial denaturation at 94 ℃ for 5 min, 35 cycles at 94 ℃ for 20 s, 55 ℃ for 30 s, and 72 ℃ for 2 min, and a final elongation step of 10 min at 72 ℃. The PCR products were confirmed using 1% agarose gel electrophoresis and visualized with ultraviolet light. Targeted bands were purified using an agarose gel DNA fragment recovery kit (GENEray Biotechnology, Shanghai, China), ligated with pMD18-T cloning vector (TaKaRa, Tokyo, Japan), and then sequenced (Sangon, Shanghai, China). Sequences were deposited in GenBank with accession numbers OL355144‒OL355148 (LSU, B. brevisulcata), OL355142 (LSU, Bps. adriatica), OL314541 (LSU, S. donghaienis), and OL314542 (ITS, S. donghaienis).

For surface sediment sample (S01-1), the single cysts were individually micropipetted onto a glass slide, then photographed with an inverted microscope (IX73, Olympus, Japan) equipped with a DP80 digital camera (Olympus, Japan). Subsequently, individual cyst was micropipetted and transferred onto a sterile slide, then broken by another coverslip. The crushed cyst and the coverslip pieces were transferred into a 250-μL centrifuge tube, as the template to amplify about 1 400 bp of the LSU rDNA using the primer set, D1R (forward, 5′-ACCCGCTGAATTTAAGCATA-3′) (Scholin et al., 1994) and 28-1483R (reverse, 5′-GCTACTACCACCAAGATCTGC-3′) (Daugbjerg et al., 2000). The following procedure was according to Shang et al. (2019).

For phylogenetic analyses of the LSU rDNA regions of B. brevisulcata and Bps. adriatica, and LSU and ITS rDNA regions of S. donghaienis, newly obtained LSU rDNA and ITS sequences were incorporated into those of closely related species available in the GenBank and that of outgroup taxa were first aligned using MAFFT v7.475 (Katoh et al., 2002) online program (http://mafft.cbrc.jp/alignment/server/) with default settings, and alignments were manually checked with BioEdit v7.2.5 (Hall, 1999). The final alignments of the LSU and ITS rDNA sequences of S. donghaienis consisted of 97 and 41 taxa and contained 750 and 600 positions (including gaps introduced from alignment), and the sequences of Cryptoperidiniopsis brodyi (DQ991374) and Pentapharsodinium dalei (JX262496) were used asthe outgroup, respectively. LSU rDNA sequences of Bps. adriatica and B. brevisulcata consisted of 66taxa with 1 590 positions (including gaps introduced from alignment), and the sequence of Alexandrium margalefii (AY154957) was used as outgroup. Theprogram jModelTest 2.1.4 was used to select the most appropriate model of molecular evolution with Akaike information criterion (AICc) (Posada, 2008), models GTR+G+I and TrN+I+G were selected as the best-fit model for the LSU rDNA and ITS datasets of S. donghaienis, and TrN+I+G for the LSU rDNA dataset of Bps. adriatica and B. brevisulcata. Phylogenetic trees were constructed using Bayesian inference (BI) and maximum likelihood (ML) analyses. Bayesian inference (BI) was performed with MrBayes 3.2.6 (Ronquist and Huelsenbeck, 2003) with the best-fitting substitution models (GTR+I+G for LSU rDNA dataset of S. donghaienis, TrN+I+G for ITS dataset of S. donghaienis, and TrN+I+G for LSU rDNA dataset of Bps. adriatica and B. brevisulcata). Four independent Markov chain Monte Carlo simulations were run simultaneously for 5 000 000 generations and trees were sampled every 1 000 generations. The first 10% trees were discarded as burn-in. The convergence was judged based on the average standard deviation of split frequencies (all less than 0.01). The remaining trees were used to generate a consensus tree and calculate the posterior probabilities of all branches using a majority-rule consensus approach. Maximum likelihood (ML) analyses were conducted with raxmlGUI v1.3.1 (Silvestro and Michalak, 2012; Stamatakis, 2014) using the models GTR+I+G (for LSU rDNA dataset of S. donghaienis), GTR+G (for ITS dataset of S. donghaienis; the model GTR+G ranked the third, andthe score of this model was close to model TrN+I+G), and GTR+I+G (for LSU rDNA dataset of Bps. adriatica and B. brevisulcata; the model GTR+I+Granked the second, and the score of this model was close to model TrN+I+G). Node support was assessed with 1 000 bootstrap replicates. FigTree (v1.4.4) was used to view and edit trees for publication.

2.7 Genetic diversity analyses

The pairwise distances were computed among all sequences that were newly obtained in the present work for B. brevisulcata, Bps. adriatica, and S. donghaienis and that retrieved from the NCBIdatabase for these three species together with other reference sequences. Sequences were aligned using the MAFFT v7.475 with the default settings (Katoh et al., 2002) (http://mafft.cbrc.jp/alignment/server/) and modified manually using BioEdit v7.2.5 (Hall, 1999). Pairwise evolutionary distances were then computed using Jukes and Cantor algorithm implemented in the MEGA X (Tamura et al., 2004; Kumar et al., 2018).

3 RESULT 3.1 Morphological observations of the resting cysts

Resting cyst of Bps. adriatica was sub-spherical to spherical and light brown, with a diameter of ~7.3 μm, full of small granules and had 2–3 red accumulation bodies (Fig. 1a). After four days' incubation, the cyst was germinated. The cyst wall was thick and smooth, and the archeopyle was tremic (Fig. 1b). Resting cyst of S. donghaienis was noncalcareous, spherical and brown, full of different sizes of granules (Fig. 1c). The diameter of the cyst was ~10.7 μm, and it contained several red accumulation bodies (Fig. 1c). After four days' incubation, the cyst was germinated. The cyst wall was thick, and the archeopyle was unclear (Fig. 1d). As the five clonal cultures of B. brevisulcata were established from cyst assemblage germination, the resting and empty cysts of this species were not observed.

3.2 Morphological observations for the vegetative cells of Biecheleria brevisulcata, Biecheleriopsis adriatica, and Scrippsiella donghaienis

The vegetative cells of B. brevisulcata were spherical to ellipsoidal, 7.2–9.2 μm long (average 8.2±0.6 μm; n=50) and 5.8–8.9 μm wide (average 7.3±0.7 μm; n=50) (Figs. 2–3). The epicone was slightly wider and longer than the hypocone, and mushroom-shaped (Figs. 2–3). The hypocone was bilobed (Figs. 2–3). The descending cingulum was deep and wide, and displaced by 1.5 times its own width (Figs. 2a–b, 3a–b). The sulcus was in the form of a sigmoid curve (Figs. 2a, b, d, & 3a). The nucleus was round and located in the middle or slightly upper part of the hypocone (Fig. 2f). Numerous yellow-brownish and reticulated or granulated chloroplasts were distributed peripherally (Fig. 2). The arrangement of polygonal amphiesmal vesicles (AVs) was shown in Fig. 3. Latitudinal rows of AVs were in four series on the epicone (E1‒E4), three series on the hypocone (H1‒H3), and three series in the cingular area (C1‒ C3; Fig. 3b, d & e). A narrow elongate apical vesicle (EAV) was present in the apical area (Fig. 3e–f), which was surrounded by five AVs (a small four-sided AV (X) and four elongated quadrangular AVs) and eight irregular pentagonal AVs (E1-1‒E1-8; Fig. 3e–f).

The vegetative cells of Bps. adriatica was spherical to ellipsoid, 5.8–10.1 μm in length (average 7.8±1.2 μm; n=50) and 4.1–8.0 μm in width (average 6.0±0.9 μm; n=50). The epicone was rounded, and almost equal to the hypocone (Figs. 4–5). The cingulum was deeper and wider than the sulcus (Figs. 4–5). The cingulum was median and displaced by 1.5 times its own width (Figs. 4–5). The sulcus was slightly sigmoid (Figs. 4a–b, e–f, & 5a). Chloroplasts were yellowish-brown, and reticulated or granulated, which were distributed peripherally (Fig. 4). The round refractive bodies were commonly observed (Fig. 4). The nucleus was located in the center or slightly upper part of the hypocone (Fig. 4c). Many pentagonal or hexagonal AVs were observed on the cell surface (Fig. 5). Four epiconal, three cingular, and four hypoconal AV series were formed the latitudinal series (Fig. 5a & c). The apical furrow was composed of an EAV and several surrounding AVs (Fig. 5a–b, d, & f).

The vegetative cells of S. donghaienis was 12.8–20.2 μm long (average 16.4±2.1 μm; n=50) and 9.6–16.0 μm wide (average 12.9±1.5 μm; n=50). The epitheca was conical and longer than the hypotheca, and the hypotheca was rounded and bilobed (Figs. 6–7). The plate formula is Po, x, 4′, 3a, 7″, 6c, 6s, 5‴, 2 (Fig. 7). The cingulum was wide and deep (Fig. 7a–c). The upper part of sulcus was narrower than its lower part (Figs. 6a–b & 7a). The apical pore complex (APC) comprised a round apical pore plate and a long canal plate (Fig. 7a & e–f). The nucleus was rounded and located centrally (Fig. 6d). The yellowish-brown, and granulated chloroplasts were distributed peripherally (Fig. 6d–e).

3.3 Pigment composition

Based on available standards, four photosynthetic pigments were identified in B. brevisulcata (strain S1), Bps. adriatica (strain S21), and S. donghaienis (strain 23), including one kind of chlorophyll (Chl a) and three carotenoids (peridinin, diadinoxanthin, and diatoxanthin; Fig. 8). Chl-a contents of B. brevisulcata (strain S1), Bps. adriatica (strain S21), and S. donghaienis (S23) were 1.02, 0.25, and 10.89 pg/cell, respectively. Peridinin (3.51, 4.91, and 79.11 pg/ cell) was the most abundant carotenoid for the three species, then diadinoxanthin, and diatoxanthin. There were 10, 11, and 11 unidentified small peaks (either new pigments or known pigments but without standards) for B. brevisulcata strain S1, Bps. adriatica strain S21, and S. donghaienis strain S23.

3.4 Molecular phylogeny

Seven partial LSU rRNA gene sequences of B. brevisulcata (strains S1, S2, S3, S4, and S5;OL355144‒OL355148), Bps. adriatica (strain S21; OL355142), and S. donghaienis (strain 23; OL314541), and one ITS sequence of S. donghaienis (strain 23; OL314542) were obtained from the clonal cultures. One partial LSU rRNA gene sequence of Bps. adriatica (ON350794) was obtained using single-cell PCR sequencing for the cyst from surface sediment sample of S01-1.

The partial LSU rRNA gene sequence of B. brevisulcata strain S1 (1 374 bp; OL355144) was 99.93% (1 373 bp/1 374 bp) identical to the entity of the type material (AB858351), 99.88% (852 bp/853 bp) to 99.93% (1 359 bp/1 360 bp) identical to four entities (AB858352, AB858353, LC068842, and OL699922) deposited as B. brevisulcata, and 98.60% (845 bp/857 bp), 99.41%(509 bp/512 bp), 99.30% (853 bp/859 bp), and 99.32% (877 bp/883 bp) identical to B. pseudopalustris (syn. Woloszynskia pseudopalustris; AF260402), B. baltica (syn. Woloszynskia halophila sensu Kremp et al. (2005); AY628430), B. cincta (syn. Woloszynskia cincta; FJ024705), and B. tirezensis (LT601379).

The partial LSU rRNA gene sequence of Bps. adriatica strain S21 (1 425 bp; OL355142) was 99.69% (1 267 bp/1 271 bp) identical to the entity of the type material deposited at GenBank as Gymnodinium pygmaeum strain K-0968, 99.24% (1 172 bp/1 181 bp)to 99.79% (1 422 bp/1 425 bp) identical to 12 entities (AB858354‒AB858356, LC068843, LC413947‒ LC413950, LM992904‒LM992906, and OL691545) deposited as Bps. adriatica, 99.78% identical to 11 entities (KM603188‒KM603198) deposited as Bps. cf. adriatica, and 99.13% (794 bp/801 bp) to99.30% (1 415 bp/1 425 bp) identical to four entities (JN558103‒JN558105, KM603185) deposited as Protodinium simplex. Among all these entities inGenBank, Bps. adriatica also corresponded to the cultures established from vegetative cells isolated from the same area where the cultures from cysts were established (i.e., East China Sea; Luo et al., 2015).

The partial LSU rRNA gene sequence of S. donghaienis strain 23 (1 433 bp; OL314541) was 98.51% (796 bp/808 bp) to 99.51% (812 bp/816 bp) identical to 69 entities deposited as S. donghaienis in GenBank. The ITS sequence of S. donghaienis strain 23 (663 bp; OL314542) was 93.80% (454 bp/484 bp) to 99.64% (558 bp/560 bp) identical to five entities (AY685008, HQ729492, HQ729502, JN982374, and MG914024) deposited as S. donghaienis, and 99.28% (550 bp/554 bp) to 99.68% (617 bp/619 bp) identical to seven entities (AY499533, AY676151, AY67615, AY788357, and FJ823594‒FJ823596) deposited as Scrippsiella sp. in GenBank.

Phylogenetic analyses of B. brevisulcata, Bps. adriatica, and S. donghaienis using maximumlikelihood (ML) and Bayesian inference (BI) generated similar trees based on LSU rRNA gene and ITS sequences but differed at a few internal nodes (Figs. 9‒11). For B. brevisulcata, our sequences and other sequences (AB858351‒AB858351, LC068842) formed a coherent clade with strong support (0.71/100), which is sistering to the clade including Woloszynskia halophila (EF205019, AY628430), B. cincta (FJ024705), Woloszynskia pseudopalustris (AF260402), and Gymnodinium sp. (AY318248) with strong support (0.95/100; Fig. 9). For Bps. adriatica, our sequence and other sequences (LC068843, LC413947, and LC413948), Bps. cf. adriatica (KM603188‒KM603198), and G. corii (GU477610) formed a coherent clade withstrong support (0.93/100; Fig. 9), and formed a well-supported sister clade (0.98/100; Fig. 9) including the type material of Bps. adriatica ("G. pygmaeum", EU857537), and G. corii (EU165298, AF318226). For the phylogenetic analysis of S. donghaienis based on LSU rRNA gene sequences, our sequence (OL31454), 69 sequences deposited as S.donghaienis, and Scrippsiella sp. (AY685011) formed a coherent clade with maximal support (1/100; Fig. 10), and formed sister groups with other Scrippsiella species (Fig. 10). For the phylogenetic analysis of S. donghaienis based on ITS sequences, our sequence (OL314542), four sequences deposited as S. donghaienis (HQ729492, HQ729502, JN982374, and MG914024), and seven sequences deposited as Scrippsiella sp. (AY499533, AY676151, AY676155, AY788357, FJ823594‒FJ823596) formed a coherent clade with medium support (0.51/88; Fig. 11), and formed a sister group with one sequence deposited as S. donghaienis (AY685008) with maximal support (1/100; Fig. 11).

3.5 The genetic diversity reflected in partial rDNA and ITS sequences

The pairwise distances computed using Jukes-Cantor model showed that the sequence divergence among the LSU rRNA gene sequences of Bps. adriatica from the present study, type material (deposited in GenBank as G. pygmaeum with the accession No. EU857537), and sequences deposited as Bps. adriatica (AB858354‒AB858356, LC068843, LC413947‒LC413950, and LM992904‒LM992906) ranged from 0.000 to 0.006, the distances between our sequence and Bps. cf. adriatica (KM603188‒ KM603198) were 0.002, and among our sequence, and the sequences deposited as P. simplex (AF060901), Gymnodinium sp. (EF205006), G. corii (AF318226, EU165298, GU477610, KT389967, KT389942), P. simplex (FJ024704, JN558103‒JN558105, KM603185‒KM603187) ranged from 0.000 to 0.007 (Supplementary Table S1). However, the distances between our sequence and B. brevisulcata (AB858351), Ansanella granifera (HG529980), and A. margalefii (AY154957) were 0.039, 0.029, and 0.244, respectively (Supplementary Table S1). Sequence differences among our strain (S21), the other stains from water samples or surface sediments formed a well-supported clade with ours (Fig. 9), and the one we found in the surface sediment sample in the same sea area were found mainly at three stable positions (3 bp/1 438 bp (D1–D6 regions); Fig. 12).

The genetic distance based on LSU rRNA gene of B. brevisulcata obtained in the present study and previous works, and other related species were compared (Supplementary Table S2). The sequence divergence among the LSU rRNA gene sequences of B. brevisulcata obtained in our work (OL355144‒OL355148) and other strains previously deposited in GenBank (AB858351‒AB858353, LC068842) ranged within 0.000‒0.007 (Supplementary Table S2), the sequence divergence among B. brevisulcata (OL355144‒OL355148) and other Biecheleria species (B. baltica, B. cincta, B. pseudopalustris, and B. tirezensis) ranged 0.002‒0.015 (SupplementaryTable S2), but the sequence divergence among B. brevisulcata (OL355144‒OL355148) and otherdistant species Pelagodinium beii (DQ195347), Polarella glacialis (FJ939578), and A. margalefii (AY154957) were 0.055‒0.057, 0.077‒0.079, and 0.808‒0.875, respectively (Supplementary Table S2).

The genetic distance based on LSU rRNA gene and ITS of S. donghaienis obtained in the present study and previous work, and other related species were compared (Supplementary Tables S3 & S4). The sequence divergence among the LSU rRNA gene sequence of S. donghaienis obtained in our work (OL314541) andother strains previously deposited in GenBank ranged 0.000‒0.009 (Supplementary Table S3), 0.013‒0.139, and the sequence divergence among S. donghaienis (OL314541) and other Scrippsiella species (S. acuminate, S. erinaceus, S. sweeneyae, S. spinifera, S. plana, S. bicarinata, S. kirschiae, S. trifida, S. infula, S. rotunda, S. lachrymose, S. enormis, S. masanensis, S. precaria, and S. ramonii) ranged 0.002‒0.015(Supplementary Table S3). The pairwise distances computed using Jukes-Cantor model showed that the sequence divergence between the ITS sequences of S. donghaienis obtained in the present study, and sequences deposited as S. donghaienis (AY685008, HQ729492, HQ729502, JN982374, and MG914024) ranged from 0.000 to 0.015 (Supplementary Table S4), the sequence divergence between the ITS sequences of S. donghaienis obtained in the present study, and the other Scrippsiella species ranged 0.175–0.360 (Supplementary Table S4). Sequence differences between our strain (S23) and the other stains from water samples or surface sediments formed a well-supported clade with ours (Fig. 11) were found mainly at two stable positions (2 bp/599 bp; Fig. 13).

4 DISCUSSION 4.1 Identification of Biecheleriopsis adriatica resting cyst presented in the field sediment

Biecheleriopsis adriatica was described by Moestrup et al. (2009) who differentiated it from the species of Biecheleria based on the presence of a nuclear connector and a 51-bases long fragment of D2 domain of LSU rRNA gene. Moestrup et al. (2009) also found resting cysts in the culture of Bps. adriatica, described as having ellipsoidal to ovoid shapes and numerous spines, a size of 7‒8-μm length and 5‒6-μm width. Benico et al. (2019) found resting cyst-like cells in unialgal culture being morphologically different from vegetative cell, and they called them as resting cyst-like cells, which were ~16 μm, spherical, covered by a transparent thick wall, and contained a red accumulation body. Kang et al. (2009), Wang et al. (2013), and Kang and Wang (2018) germinated Bps. adriatica from cyst assemblages (identified as G. corii in Kang et al. (2009) and Wang et al. (2013))from Southern Chinese coastal sediments, but did not observe the genuine resting cyst of this organism. Luo et al. (2015) also germinated Bps. cf. adriatica (very possible conspecific with our strain S21) from the sediment collected from the Yellow Sea and the South China Sea, but also without observation on the morphology of Bps. cf. adriatica cyst. The above-mentioned works on Bps. adriatica have confirmed that it could form resting cyst, but none of them reported the morphology of cyst from the field. In this work, we germinated Bps. adriatica from a single cyst collected from the East China Sea, and the species was identified with morphological and molecular characterization allowing us to confidently confirm the cyst-motile stage relationship of this small-sized species. Resting cyst of Bps. adriatica was spherical and sub-spherical, which is similar to the so-called "resting cyst-like cells" as observed by Benico et al. (2019), but different from the cyst from culture having an ellipsoidal to ovoid morphology (Moestrup et al., 2009). The diameter of Bps. adriatica in our work (~7.3 μm) is very close to the cyst (7‒8 μm long and 5‒6 μm wide) observed by Moestrup et al. (2009), but much smaller than the resting cyst-like cells (~16 μm) observed by Benico et al. (2019). The sieves with the pore size larger than 20 μm were routinely used to concentrate cysts, and then followed by cyst assemblage germination, the cysts with smaller size (< 20 μm) would be lost during processing, therefore, the efforts in discovering more small sized cysts should focus on the cyst assemblage being smaller than 20 μm.

4.2 Viability of dinoflagellates cysts stored in long-buried sediments

In the past, changes of species composition in sediments have been used to assess environmental changes including eutrophication, changes in salinity, or oxygen concentration (Dale et al., 1999; Ellegaard et al., 2013, 2020; Bringué et al., 2016; García-Moreiras et al., 2018; Li et al., 2021; Siano et al., 2021). The adaptive responses of dinoflagellate species might be influenced by anthropogenic activities or climate changes, and revealing the evolutionary processes in a species is very important to understanding the genetic structure of populations. However, the length of live resting cyst preserved in the sediment cores (e.g., decades to centuries) or the molecular information (rRNA genes or other gene markers) for the buried cysts or residual fragments have not been well explored before 2010. After that, the impacts of environmental changes on the physiology, genetic structure, and diversity in various dinoflagellate species (Ribeiro et al., 2011; Klouch et al., 2016; Lundholm et al., 2017; Kremp et al., 2018; Delebecq et al., 2020; Ellegaard et al., 2020; Girault et al., 2021) have been investigated. In this study, we successfully revived B. brevisulcata, Bps. adriatica, and S. donghaienis from a sediment coredated back to 1941±18 AD from the East China Sea, which indicates that these cysts are viable for at least 70 years. Klouch et al. (2016) detected molecular signal of S. donghaienis in a sediment core sample dated back to 1866±7 AD from the Bay of Brest, France, and successfully germinated this organism only from the layers of 2–17 cm corresponding to the years of 2010±1 AD to 1978±2 AD, much more recent than 1941±18 AD. In older sediments (70‒100 years), Protoceratium reticulatum, Lingulodinium polyedrum, and P. dalei were successfully germinated(Lundholm et al., 2011; Ribeiro et al., 2011). Girault et al. (2021) successfully germinated A. minutum and S. acuminata from the sediment corresponding to1947±11 AD. Kremp et al. (2018) found Apocalathium malmogiense is viable in the sediment layers of 106-year old. Recently, Delebecq et al. (2020) germinated A. minutum, Heterocapsa minima, Margalefidinium polykrikoides, Protoperidinium spp., and Diplopsalis group in sediments dated back to 117±21 years ago, and S. acuminata beyond 156±27 years ago. Our work increases the diversity of species (B. brevisulcata and Bps. adriatica) potentially revivable from more long-buried sediments, which will promote studies in the field of resurrection ecology.

4.3 Genetic diversity among dinoflagellates revived from cysts stored in long-buried sediments and current condition

Genetic diversity is an important aspect of biodiversity, which is defined as measurements that determine the changes of genetic variability within any level of a taxon but more often within a species or even population, e.g., allelic diversity or richness, mutational diversity, and effective population size (Hughes et al., 2008; Ebenezer et al., 2012; Ellegren and Galtier, 2016). Previous phylogeographic studies have revealed high levels of genetic diversity in a number of dinoflagellate species, e.g., A. fundyense (Erdner et al., 2011), Amphidinium spp. (Murray et al., 2012), Gambierdiscus spp. (Nishimura et al., 2013), Margalefidinium fulvescens (Lin et al., 2020), Ostreopsis spp. (Lee and Park, 2020), and Pseudocochlodinium profundisulcus (Hu et al., 2021).However, all above-mentioned works examined the genetic diversity among contemporary populations or even within a population rather than that between historical and contemporary populations as done in the present work, although the history here is less than a century (~70 years). From our phylogenetic tree, our five strains of B. brevisulcata branched together with the other strains of this species, but their evolution distance varied. Our strain of Bps. adriatica and three strains of Bps. adriatica, eleven strains of Bps. cf. adriatica, and one strain of G. corii formeda coherent clade, and a sister group with the type material of this species, but differed at three stable positions. Our strain of S. donghaienis, the other strains of this species, and one strain of Scrippsiella sp. formed a coherent clade. Within this clade, there were four branches, our strain grouped together with those historical and contemporary populations. We also compared the rRNA gene based genetic distances of the seven strains of B. brevisulcata, Bps. adriatica, and S. donghaienis revived from a long-buried sediment dated back to 1941±18 AD and those from water samples and surface sediments (the contemporary age), and found the genetic distance between our strain of B. brevisulcata and other strains range from 0.002 to 0.006, but among other strains, the genetic distances are 0.000‒0.005, most 0.001, and similar trends were also found in Bps. adriatica, and S. donghaienis. We also found differences in three stable positions of LSU rRNA gene sequences between the population of Bps. adriatica we found 70 years ago and the contemporary or present-day population (including the cyst of this species we found in the surface sediment and vegetative cells found by Luo et al. (2015) in the same sea area), and ITS sequences in two stable positions between the population of S. donghaienis we found 70 years ago and the contemporary or present-day population (including the cyst of this species found by Gu et al. (2008) in the surface sediment in the same sea area). It seemed these current populations are genetically different from those that existed in the area 70 years ago, which suggests that there has been a shift in the populations of Bps. adriatica and S. donghaienis. Due to a limited size of dataset for the present work, the possible historical succession of populations of Bps. adriatica, S. donghaienis and even other speciesin the area requires a more intensive and extensive investigation, as these possible shifts may be highly indicative of the environmental changes and anthropological activities that occurred in the area during the past.

5 DATA AVAILABILITY STATEMENT

All data generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

6 ACKNOWLEDGMENT

We are highly grateful of the two anonymous reviewers for their patience, critical comments, and generous suggestions, which helped greatly the improvement of the manuscript. Sediment samples' collections were supported by NSFC Open Research Cruise, funded by Shiptime Sharing Project of NSFC. We appreciate for the cruises conducted by R/V Xiang Yang Hong 18 and staff from First Institute ofOceanography, Ministry of Natural Resources, China.

Electronic supplementary material

Supplementary material (Supplementary Tables S1–S4) is available in the online version of this article at https://doi.org/10.1007/s00343-022-2122-1.