Emendation of Rhodomonas marina (Cryptophyceae): insights from morphology, molecular phylogeny and water-soluble pigment in an Arctic isolate
Article information
Abstract
Rhodomonas (Cryptophyceae) and species assigned to this genus have undergone numerous taxonomic revisions. This also applies to R. marina studied here as it was originally assigned as a species of Cryptomonas and later considered a variation of R. baltica, the type species. Despite being described more than 130 years ago, R. marina still lacks a comprehensive characterization. Light and electron microscopy were employed to delineate a strain from western Greenland. The living cells were 18 μm long and 9 μm wide, elliptical in shape with a pointed to rounded posterior and truncated anterior in lateral view. Two sub-equal flagella emerged from a vestibulum, where also a furrow extended. In transmission electron microscopy, the furrow was associated with a tubular gullet and the pyrenoid embedded in a deeply lobed chloroplast. The chloroplast contained DNA in perforations and was surrounded by starch grains. A tubular nucleomorph was enclosed within the pyrenoid matrix. In scanning electron microscopy, the inner periplast consisted of rectangular plates with rounded edges and posteriorly these were replaced by a sheet-like structure. The water-soluble pigment was Crypto-Phycoerythrin type I (Cr-PE 545). A phylogenetic inference based on SSU rDNA confirmed the identity of strain S18 as a species of Rhodomonas as it clustered with congeners but also Rhinomonas, Storeatula, and Pyrenomonas. These genera formed a monophyletic clade separated from a diverse assemblage of other cryptophyte genera. To further explore the phylogeny of R. marina a concatenated phylogenetic analysis based on the SSU rDNA-ITS1-5.8S rDNA-ITS2-LSU rDNA region was performed but included only closely related species. The secondary structure of nuclear internal transcribed spacer 2 was predicted and compared to similar structures in related species. Using morphological and molecular signatures as diagnostic features the description of R. marina was emended.
INTRODUCTION
Most observations of the cryptophyte Rhodomonas marina (Dangeard) Lemmermann have been based on isolates (morphotypes) from temperate and subtropical waters, specifically the Atlantic coast of Europe (Dangeard 1892, Gameiro and Brotas 2010) including the Kattegat (Hill et al. 1992) and likely the Mediterranean Sea (Cerino and Zingone 2006). However, more recently R. marina has also been recorded in Arctic waters, more specifically the White Sea (Krasnova et al. 2014) and the Beaufort Sea (Ribeiro et al. 2020). The TARA expedition (Karsenti et al. 2011) has further expanded the biogeography of R. marina to also include the Southern Ocean, the Indian Ocean, and the Pacific (gbif.org visited on Jan 29, 2024). In summary, R. marina exhibits a global distribution, inhabiting both coastal regions and open oceans. Yet, a detailed morphological account has not been provided despite R. marina being recognized for more than 130 years. This is somewhat surprising as Rhodomonas Karsten emend. Hill & Wetherbee (including R. marina) is often used as feed in the aquaculture industry (Knuckey et al. 2005, Tremblay et al. 2007, Ohs et al. 2010, Zhang et al. 2013, Arndt and Sommer 2014, Thoisen et al. 2018, Oostlander et al. 2020, Nogueira et al. 2021, Fichtbauer et al. 2023). It could be argued that species with a desired fatty acid composition and other growth-enhancing compounds (Thoisen et al. 2018) calls for identifications based on well-circumscribed species.
The α- and β-taxonomy of cryptophytes has been studied since Ehrenberg (1831) discovered the algal group >190 years ago. At that time, only light microscopy was available to study morphology and chloroplast color of these single-celled bi-flagellates. When using light microscopy exclusively, species identification and subsequent cryptophyte taxonomy have proven difficult (e.g., Solarska et al. 2023). Electron microscopy and in particular molecular sequencing have for the past decades improved the taxonomy of Cryptophyceae (e.g., Klaveness 1989, Cavalier-Smith et al. 1996, Deane et al. 2002, Hoef-Emden et al. 2002, Novarino 2012, Majaneva et al. 2014, Daugbjerg et al. 2018, Altenburger et al. 2020, Khanaychenko et al. 2022). However, several taxonomic issues still need to be resolved and these range from the level of species to orders as was also seen in a recent phylogenomic study (Greenwold et al. 2023).
A brief historical account of the taxonomy of R. marina is provided below. It was first described by Dangeard (1892), although he described it as a member of the genus Cryptomonas. According to Dangeard (studying fewer than 10 cells), the color of the chloroplast in C. marina was yellow to brown-yellow and had a large oily-looking drop and two flagella arising from a vestibulum, which were approximately of the same length as the cell. Dangeard (1892) then compared the shape to Cryptomonas ovata, which was described by Ehrenberg (1832). Ehrenberg described C. ovata as being twice as long as wide, rounded on the dorsal side and compressed in the posterior end. Later, Karsten (1898) created the genus Rhodomonas with R. baltica as the type species. Lemmermann (1903) transferred R. marina into its present genus although no further morphological characteristics were added.
The taxonomy became even more complicated when Lohmann (1908) in a water sample from the Bay of Kiel discovered a cryptophyte, which he named Rhodomonas pelagica. Büttner (1910) discovered a C. marina similar to the one described by Dangeard (1892) and extended the list of characteristic features for this genus. However, Büttner (1910) did not take into consideration that Lemmermann (1903) had moved the species to a new genus. Zimmermann (1924) as cited in Hill and Wetherbee (1989), conducted a study of a cryptophyte species that he found in the Baltic Sea, which appeared to be R. baltica Karsten. He then considered R. marina (Dangeard 1892), R. pelagica (Lohmann 1908), and C. marina (Büttner 1910) to be forms (conspecifics) of R. baltica. Butcher (1967) questioned the validity of the genus Rhodomonas due to the only seemingly difference in Karsten’s (1898) description of Cryptomonas and Chroomonas was the floridian red color. He deemed this feature invalid for delineating genera. Thus, Butcher (1967) transferred species of Rhodomonas into either Chroomonas or Cryptomonas. In 1984, Santore erected the genus, Pyrenomonas, with characteristics resembling the original description of Rhodomonas. Two years later Santore (1986) transferred species from Butcher (1967), which were originally described as species belonging to Rhodomonas into this new genus. However, later Pyrenomonas was abandoned due to the validity of Karsten’s (1898) description of Rhodomonas, which had largely been overlooked (Hill and Wetherbee 1989). Today, the genus Rhodomonas is accepted taxonomically (Erata and Chihara 1989, Hill and Wetherbee 1989) and 20 species are currently assigned to Rhodomonas following Guiry and Guiry (2024) (AlgaeBase.org visited on May 20).
To ameliorate the lack of an integrative study of R. marina we studied a clonal culture (strain S18) established from a water sample collected in Disko Bay, western Greenland. Hence, the aims were to provide a detailed description of its morphology, phylogeny, and water-soluble pigment by applying advanced bio-imaging, scanning- and transmission electron microscopy, determinations of the nuclear-encoded small subunit ribosomal DNA (SSU rDNA), internal transcribed spacers and partial large subunit ribosomal DNA (LSU rDNA) as well as a spectrophotometric analysis.
MATERIALS AND METHODS
Sampling and cultured material
A clonal culture of R. marina (strain S18) was established from a sample collected in Disko Bay, western Greenland (July 2018) at 69°11.112′ N, 53°30.995′ W (Supplementary Fig. S1). The culture was kept at 4°C and grown in L1 medium (Guillard and Hargraves 1993) with a salinity of 30 at the Marine Biological Section, University of Copenhagen. The light : dark cycle was 16 : 8 h and LED light panels from Philips (33 W) provided a light intensity of 30–50 μmol photons m−2 s−1.
DNA extraction, amplification, and sequencing of nuclear SSU rDNA
An exponentially growing culture (11 mL) was transferred to a Falcon tube and a cell pellet was made by centrifugation at 1,174 ×g for 10 min at 5°C. The pellet was transferred to an Eppendorf tube and frozen at -18°C. Two days later, extraction of total genomic DNA was performed using the PowerPlant Pro DNA isolation kit, following the manufacturer’s recommendations (MO Bio Laboratories Inc., Carlsbad, CA, USA). Extracted DNA was used as template for polymerase chain reaction (PCR) amplifications of SSU rDNA using forward primer ND1 and reverse primer ND6 (Ekelund et al. 2004). For amplification, the 5× Hot FIREPol Blend Master Mix from Solis BioDyne was used. The PCR temperature profile consisted of one initial cycle of denaturation at 95°C for 12 min, followed by 35 cycles each consisting of denaturation at 95°C for 30 s, annealing at 54°C for 30 s and extension at 72°C for 30 s. A final extension step at 72°C lasted 6 min. The length of the amplified products was confirmed by electrophoresis using an agarose gel (final concentration 1.5%). The PCR product was stained using GelRed and visualized using a gel documentation XR System (Bio-Rad, Hercules, CA, USA). For purification, the Nucleofast 96 PCR kit from Macherey-Nagel (GmbH & Co KG, Düren, Germany) was applied following the recommendations of the manufacturer. For sequence determination, the terminal primers (ND1F and ND6R) were used in addition to forward primer ND3F and reverse primers ND7R and ND8R (Ekelund et al. 2004). The service provided by Macrogen was used for sequence determination in both directions.
Amplification and sequencing of nuclear internal transcribed spacers and partial LSU rDNA
The DNA fragment corresponding to the internal transcribed spacers 1 and 2 (plus the 5.8S rDNA in between) and partial LSU rDNA was amplified using primers ITS1 and ITS4 (White et al. 1990). See Binzer et al. (2019) for PCR condition, purification, and sequence determination of this nucleotide fragment.
Alignment and phylogenetic analyses of SSU rDNA
To infer the phylogeny of R. marina, the SSU rDNA sequence was added to a data matrix previously compiled for studying another Artic cryptophyte (Daugbjerg et al. 2018). However, additional cryptophytes (particularly species / strains of Rhodomonas) were added prior to phylogenetic inferences using Bayesian analysis (BA) (Ronquist and Huelsenbeck 2003) and RAxML (Stamatakis 2014) as implemented in Geneious Prime (ver. 2023.1.2). The updated data matrix consisted of 1,962 base pairs (including introduced gaps) from 83 taxa (56 identified species and 20 genera). For BA 5 million generations were analyzed and a tree sampled every 1,000th generation. The burn-in was set to have occurred after 50,000 generations (10%) leaving 4,501 trees for a majority rule consensus tree. RAxML was used for bootstrap (BS) analysis with 1,000 replications and the nucleotide model used was GTR gamma. BS values were mapped onto the tree topology obtained from BA.
Alignment and phylogenetic analyses of concatenated data matrix
To further explore the phylogeny of strain S18, sequences of the internal transcribed spacer region (including 5.8S rDNA) and partial LSU rDNA (c. 200 base pairs) were retrieved from GenBank (May 3, 2023) for the most closely related cryptophytes available and together with ITS region of S18 added to the SSU rDNA alignment. This allowed for a phylogenetic analysis of a concatenated data matrix comprising 19 taxa and 2,800 base pairs including introduced gaps. Hemiselmis virescens (strain RCC3575) was used as outgroup. For phylogenetic inferences, BA and RAxML were used with similar settings as described above. Except for BA, where the genetic markers were divided into five data partitions, one partition for each marker. Thus, each of these regions could evolve under different models using the “unlink” option in MrBayes. The division into partitions was based on annotations of Rhodomonas sp. with GenBank accession No. KY095062, Rhodomonas sp. with GenBank accession No. AY095063 and Cryptomonas sp. with GenBank accession No. KF907360.
Sequence divergence estimates
PAUP* (ver. 4 3.99.169.0) (Swofford 2003) provided sequence divergence estimates based on the SSU rDNA matrix (1,610 base pairs including a few introduced gaps). Pairwise divergence estimates were calculated using the Kimura-2-parameter model but only included strains identified belonging to Rhodomonas or expected to belong to this genus (i.e., Rhinomonas reticulata and Pyrenomonas helgolandii). Sequence divergences were also estimated for strains S18, K-0332, and 08C1 using the concatenated matrix as input data (2,494 base pairs).
Predicted secondary structure of ITS2
The RNA folding program mFold was accessed through the web server at http://www.unafold.org/mfold/applications/rna-folding-form.php (Zuker 2003). Using default values, the secondary structure was predicted for helices I–IV of ITS2 in R. marina (strain S18). The structure of the four helices was drawn using VARNA (Darty et al. 2009).
Water-soluble pigment
Analysis of the water-soluble pigment in R. marina (strain S18) was based on Lawrenz et al. (2011). A volume of 450 mL of an exponentially growing culture (strain S18) was centrifuged for 30 min at 4,696 ×g and 4°C. The supernatant was removed, and 3 mL phosphate buffered saline buffer (pH 7.4) was added and mixed with the cell pellet. Cells were disrupted by exposing them to one freeze-thaw cycle (−18°C, 2 h) followed by centrifugation at 4,696 ×g for 20 min. Extraction was then completed at 4°C for 18 h. The absorbance of the purified extracts within the supernatant was analyzed by scanning the spectrum between 380 and 750 nm in a Shimadzu UV180 spectrophotometer (Holm & Halby A/S, Brøndby, Denmark). The phosphate buffer was used as the blank.
Light microscopy of living and fixed cells
Living cells of strain S18 were observed using a Carl Zeiss Axio Imager M2 equipped with Nomarksi interference contrast and a 63× oil immersion lens. Micrographs were taken with a Zeiss AxioCam HRc digital camera (Zeiss, Oberkochen, Germany). For epifluorescence microscopy of the chloroplast morphology, an exponentially growing culture was fixed in glutaraldehyde (final concentration 2%). The fixed cells were then filtered onto a white filter with a pore size of 0.22 μm (Whatman, GE Healthcare, Chicago, IL, USA). The filter was placed on a slide which already had received 20 μL of Vectashield. Another drop of Vectashield was placed on the upper side of the filter before placing a cover slip on top. Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA) contains DAPI (4′-6-diamidino-2-phenylindole), a DNA stain. Nuclear and chloroplast DNA was observed using the Zeiss filter set 49 (excitation band pass 365 nm and emission 445 nm). Chlorophyll autofluorescence of mounted specimens was recorded using a Zeiss Axiocam 506 mono digital camera and a Zeiss filter set 09 (excitation BP450–490, emission LP515). Recorded z-stacks comprised between 57 and 68 single images and were imported into the Imaris 9.2.1 software package (Bitplane Scientific Software, Zürich, Switzerland) for 3-dimensional studies. The snapshot option was used to export combined stacks. For measurements of cell dimensions a total of 40 images of randomly picked cells were taken. Length and width of these cells and length of flagella of five cells were measured using Zen image acquisition software from Zeiss. Means and standard deviations were calculated using Excel (Microsoft, Redmond, WA, USA).
Scanning electron microscopy
A total volume of 800 μL of strain S18 was fixed in 1,920 μL 2% OsO4 and 480 μL HgCl2 solution for 1 h. Cells were placed on a 5 μm polycarbonate filter and rinsed with Milli Q water five times. Then dehydration was executed in a series of ethanol (30, 50, 70, 90, and 99.9%) and each step lasted 20 min. This was followed by dehydration in absolute ethanol for 1 h. Subsequently, the sample was critical-point dried. The filter with fixed cells was mounted to an aluminum stub and sputter coated with gold / palladium in a JEOL JFC-2300HR sputter coater. The material was observed in a JEOL JSM-6335F field emission scanning electron microscope (JEOL Ltd., Tokyo, Japan) with a secondary electron detector and running at 12 kV.
Transmission electron microscopy
A fixation mixture was produced of L1 medium and a final concentration of 2% glutaraldehyde, 0.05 M cacodylate, and 0.5 M sucrose. The pH was adjusted to 7.2 using HCl. Approximately, 6 mL of strain S18 was added to the fixation mixture and centrifuged at 1,201 ×g for 10 min. After 20 min, the supernatant was removed and approx. 2 mL of medium containing 2% glutaraldehyde, 0.05 M cacodylate, and 0.5 M sucrose, was added. After 20 min of fixation, the supernatant was removed and 2 mL medium of 1% glutaraldehyde, 0.05 M cacodylate, and 0.25 M sucrose was added. After 20 min, the supernatant was removed once again and 2 mL of medium of 0.05 M cacodylate was added. After 20 min, 2 mL was added containing 0.05 M cacodylate and 1% osmium. After 60 min, the pellet was washed with 0.05 M cacodylate. Then the pellet was dehydrated in an ethanol series of 30, 50, 70, 96, and 99.9%. Each of these steps lasted 20 min and was followed by dehydration twice in absolute ethanol for 15 min. Next, the pellet was washed with 2 mL propylene oxide for 5 min, followed by adding a 1 : 1 mix of propylene oxide and EPON, which was left overnight allowing the propylene oxide to evaporate. Finally, EPON was removed, and cells transferred into new EPON and polymerized at 60°C for 3 h. Thin sections were cut with a diamond knife and stained first in uranyl acetate (10 min), then washed in MilliQ water and finally stained in lead citrate (10 min). Observations were made using a JEOL JEM-1010 transmission electron microscope (JEOL Ltd.). The accelerating voltage was 80 kV (JEOL Ltd.) and micrographs were taken using a Gatan Bioscan camera Model 792 (Gatan, Pleasanton, CA, USA).
RESULTS
Phylogeny based on SSU rDNA
A data matrix comprising nuclear SSU rDNA sequences from 20 cryptomonad genera with 56 identified species (83 strains) was compiled to infer the phylogeny of R. marina (strain S18). The single-gene phylogeny was shown in Fig. 1. The two heterotrophic genera (Hemiarma and Goniomonas) branched off first followed by a monophyletic clade comprising the cryptophytes (posterior probabilities [PP] = 1.0 and BS = 99%). The relationship between the deepest lineages within the cryptophyte clade was not well resolved. However, six major clades all received high branch support.
Phylogeny of Rhodomonas marina (strain S18, bold face) based on nuclear-encoded SSU rDNA sequences and inferred from Bayesian inference. The ingroup of Cryptophyceae with 9 families comprised 20 genera and 56 species. Roombia truncata (a katablepharid) formed the outgroup taxon. The robustness of the topology was evaluated from posterior probabilities (≥0.5) (numbers to the left of slashes) and bootstrap values (≥50%) from RAxML (numbers to the right of slashes). Strain numbers (when available) are given in parentheses and if listed they are followed by GenBank accession numbers. The branch lengths are proportional to the number of character changes. Clade designations in boxes were taken from a phylogenomic analysis by Greenwold et al. (2023).
With respect to Rhodomonas, which included strains of which 21 were identified to species, the phylogenetic analyses showed the genus to be polyphyletic. Rhodomonas minuta Skuja (GenBank accession No. MK828414) clustered with Plagioselmis prolonga Butcher ex Novarino, Lucas and Morrall (GenBank accession No. AF508272) and these two taxa formed a sister group to Teleaulax minuta, T. grailis, and Geminigera crypophila. Thus, R. minuta was distantly related to the core lineage of Rhodomonas spp. which included the type species (R. baltica). The lineage comprising the majority of Rhodomonas species (named the core hereafter) received maximum support from PP and BS value (PP = 1, BS = 100%). Considering the core only, the genus Rhodomonas was still polyphyletic as Rhinomonas, Storeatula, and Pyrenomonas clustered among species of Rhodomonas. Rhodomonas marina (strain S18) formed an unresolved relationship with R. cf. abbreviata, R. baltica and an unidentified species of Rhodomonas (strain 08C1). This clade formed a sister group to a lineage comprising Storeatula sp. and R. duplex (Fig. 1).
The phylogenetic analyses also suggested that the identity of strain CCAP979/15 (Rhinomonas reticulata var. reticulata) as well as R. marina and R. minuta with GenBank accession Nos. X81373 and MK828414, respectively, should be confirmed. The taxonomy of the genus Chroomonas also needs attention as species of this genus formed three separate lineages (Fig. 1).
Phylogeny based on concatenation of SSU rDNA-ITS1-5.8S rDNA-ITS2-partial LSU rDNA
The phylogenetic analysis based on the combined data set is shown in Fig. 2. Here R. marina clustered with R. baltica (strain K-332) (PP = 0.93 and BS = 98%), and these formed a sister group to Rhodomonas sp. (strain 08C1) (PP = 1.0 and BS = 100%).The short branch lengths for R. marina and R. baltica indicated highly similar sequences. Based on low statistical support values the tree topology for most of the deepest lineages was unresolved (PP ≤ 0.92 and BS < 50%). Hence, the closest sister group / taxon to R. marina / R. baltica / Rhodomonas sp. strain 08C1 remained unresolved. Another strain of R. baltica (NIES747) clustered with Rhodomonas sp. (CCMP768) and these taxa were united by a long branch and thus distantly related to strain K-0332 also identified as R. baltica. The majority of the Rhodomonas strains included were not identified to species. Obviously, this made it difficult to further explore the phylogeny of R. marina. In these analyses the genus Rhodomonas was also found to be polyphyletic.
Phylogeny of Rhodomonas marina (strain S18, bold face) based on concatenation of the SSU rDNA-ITS1-5.8S rRNA-ITS2-partial LSU rDNA region (2,800 base pairs including introduced gaps). Hemiselmis virescens formed the outgroup taxon. The robustness of the tree topology was evaluated from posterior probabilities (≥0.5) in Bayesian analysis (numbers to the left of slashes) and bootstrap values (≥50%) from RAxML (numbers to the right of slashes). Strain numbers were given in parentheses followed by GenBank accession numbers. The branch lengths were proportional to the number of character changes.
Sequence divergence estimates
To further evaluate the identity of Rhodomonas strains included in the phylogenetic analyses, divergence estimates for all pairwise comparisons (including Rhinomonas reticulata and Pyrenomonas helgolandii) were made using PAUP* (Swofford 2003). A total of 1,610 base pairs of the SSU rDNA gene were included. The distant relationship between R. minuta and the core of Rhodomonas was also reflected in the sequence divergence estimates as it ranged between 5.8–6.9%. This was significantly higher than any pairwise comparison between strains within the core Rhodomonas group, which ranged between 0 and 3.3%. The highest divergence was between R. storeatuloformis and R. duplex (data not shown). Comparing sequence divergences within the core group of Rhodomonas (Fig. 1), several strains were 100% identical. These have been marked in Fig. 1. Rhodomonas marina (X81373) differed by 0.51–0.53% to the 11 most closely related strains. Rhodomonas marina (strain S18) differed by 2.0% when compared to R. marina (X81373).
All pairwise comparisons based on the SSU rDNA-ITS1-5.8S rDNA-ITS2-partial LSU rDNA region for strains S18, K-0332, and 081C revealed highly similar divergence estimates. The difference between S18 and K-0332 was 0.09% and these two strains varied between 0.25–0.32% when compared to strain 081C.
Predicted secondary structure of ITS2
Predicted secondary structures of nuclear ITS2 helices have been included as diagnostic molecular markers in descriptions of new or emended species belonging to for example Cryptomonas (Hoef-Emden 2007, Gusev et al. 2022), Chroomonas (Hoef-Emden 2018), Rhinomonas (Majaneva et al. 2014), and Rhodomonas (Khanaychenko et al. 2022). Here the predicted secondary structure of ITS2 helices I-IV in strain S18 were shown in Fig. 3. Helix I comprised 13 base pairs (including a A A mismatch and a single-nucleotide [C] bulge) and a terminal loop of 4 bases. In helix II, the stem comprised 17 base pairs (including three mismatches G G, U U, and A C). The terminal loop had 6 base pairs. The highly conserved U U mismatch in helix II (Hoef-Emden 2007) was also present in strain S18. Helix III was by far the longest of the four with a total of 45 base pairs (including 5 non-canonical base pairs). Helix III was also characterized by the presence of a large internal loop (12 bases) and a bulge comprising 5 bases (UUAAG). The terminal loop had 5 bases. Helix IV had 11 base pairs in the stem (no internal loops or bulges) and 4 bases (UUUA) in the terminal loop.
Predicted secondary structures of helices I–IV of ITS2 in Rhodomonas marina (strain S18). Mfold (ver. 2.3) (Zuker 2003) was used to predict the structures using default settings and VARNA (ver. 3.93) (Darty et al. 2009) to draw them.
Pigment profile
Analysis of the water-soluble pigment in R. marina (S18) revealed an absorption maximum at 545–548 nm (Fig. 4). This indicated the presence of the phycobilin Crypto-Phycoerythrin type I (Cr-PE 545) (Greenwold et al. 2019).
Absorption spectrum of the water-soluble pigment in Rhodomonas marina (strain S18). The arrow indicating the peak in the graph corresponds to Crypto-Phycoerythrin 545 (Cr-PE 545).
Morphology, chloroplast color, and swimming behavior
Under the light microscope, living cells measured 18.0 ± 1.9 μm in length (min. 14 μm, max. 22 μm) and 9.0 ± 1.5 μm in width (min. 6 μm, max. 12 μm) (n = 40). Cells appeared elliptical with a rounded (Fig. 5A & B) to pointed posterior (Fig. 5C & D). The anterior end of cells was rounded in ventral view (Fig. 5E & F) and truncated in lateral view (Fig. 5A & B). Two unequally long flagella were inserted in the upper part of the vestibulum (Fig. 5A & B). The length of the flagella was 11.2 ± 1.9 μm (min. 9 μm, max. 14 μm) (n = 5). On the ventral side rows of large ejectisomes were seen near the gullet (Fig. 5A–F & H). A contractile vacuole was observed in the apical end (Fig. 5A, C & E) and in a contracted stage in Fig. 5D & F. Based on a single video recording (not shown), a full contractile vacuole period from diastole to systole lasted approximately 12 s. The nucleus was located at the posterior end (Fig. 5A & I–L). The single chloroplast was almost the same length as the cell (Fig. 5A–D). The upper most dorsal part of cells seemed devoid of this organelle (Fig. 5A–D). The single chloroplast was deeply lobed (H-shaped) and perforated but some variation in the outline was noted in epifluorescence microscopy (Fig. 5I–K). DAPI-stained cells confirmed the posterior position of the nucleus but also DNA in the chloroplast associated with the perforations (Fig. 5L). A pyrenoid was observed in the mid-dorsal part of the cell, associated with the chloroplast (Fig. 5C & D). Storage material (starch grains) encircled the pyrenoid (Fig. 5C & D). Inner periplast plates were aligned in longitudinal rows (Fig. 5G & H).
Light microscopy of Rhodomonas marina (strain S18) grown at 30–50 μmol photons m−2 s−1. (A–H) Nomarski interference contrast. (I–L) Stacked images from epifluorescence microscopy. (A & B) Same cell seen from a right lateral view. Two unequally long flagella (f) emerge subapically from within the depression / vestibulum. A contractile vacuole is indicated by an arrow. The nucleus (N) is located in the posterior end. Note rounded posterior of cell. (C & D) Same cell seen from a left lateral view. Note rows of large ejectisomes (e), chloroplast (c) and a centrally located pyrenoid (p) surrounded by starch grains (s). Note pointed posterior end of cell. (E & F) Same cell seen from the ventral side. Note also rows of large ejectisomes and a contractile vacuole in (E) (arrow) which has contracted in (F) (arrow). (G & H) Peripheral view from the ventral side illustrating the longitudinal rows of periplast plates in (G) and rows of ejectisomes aligning the gullet in (H). (I & K) 3D reconstruction of chloroplast and DAPI-stained nucleus in three cells with different orientation. Note variations of the deeply lobed (H-shaped) chloroplast. I, 63 stacked images; J, 68 stacked images; K, 57 stacked images. (L) Single frame image (from a series of 53 images) revealing chloroplast DNA in the perforations of the single chloroplast. Scale bars represent: A–L, 5 μm.
The color of the chloroplast was yellow-brown and it did not change under the light conditions provided here (30–50 μmol photons m−2 s−1).
Cells continued to swim in a nearly straight path for long periods of time while rotating around their longitudinal axis. When they stopped swimming, cells recoiled due to the discharge of ejectisomes.
External morphology, scanning electron microscope
Cells appeared elliptical with a pointed (Fig. 6A & B) or rounded (Fig. 6D) posterior and a truncate anterior in lateral view (Fig. 6D & F). An apical depression formed a vestibulum from which the furrow was protruding (Fig. 6A & D–F). The furrow extended approx. 1/6 of the cell length (Fig. 6A). Two unequally long flagella emerged from the vestibulum. The longer flagellum was situated towards the dorsal side, whereas the shorter flagellum was situated towards the ventral side (Fig. 6E). The periplast plates (0.5 μm × 0.4 μm) were arranged in longitudinal rows and appeared rectangular with rounded edges and prominent ridges (Fig. 6A–D). The plates became sheet-like in the posterior end (Fig. 6A–D, G & H). In the posterior end, a mid-ventral band was protruding from the dorsal side of the cell (1/10 of cell length) onto the ventral side extending 2/5 of the cell length (Fig. 6G & H). Numerous discharged ejectisomes were observed on most cells (Fig. 6A–F).
Scanning electron microscopy of Rhodomonas marina (strain S18). (A) Ventral view. Note numerous discharged ejectisomes surrounding the cell. Subapical depression / vestibulum is also visible. The furrow (fu) extends 1/6 of the cell length. Two flagella emerge from the vestibulum. (B) Dorsal view. Periplast plates appear rectangular with rounded edges and prominent ridges. The periplast plates are organized in longitudinal rows and in the posterior part of the cell they are replaced by a sheet-like structure. Note pointed posterior. (C) Right side of the cell. (D) Left side of the cell. (E) Ventral side of a cell illustrating the furrow and two subapically inserted flagella. (F) High magnification of a cell from the right side showing furrow and numerous discharged ejectisomes. (G) High magnification of antapical end seen from the ventral side. Note mid-ventral band (arrow) and the change from periplast plates into a sheet-like structure. (H) High magnification of dorsal side showing the mid-ventral band and the sheet-like structure of the periplast. Scale bars represent: A–G, 1 μm; H, 0.5 μm.
Ultrastructure, transmission electron microscope
Longitudinal sections revealed the general disposition of major cell organelles (Fig. 7A & B). A pyrenoid took up a large part of the cell center (Fig. 7B), whereas the nucleus was seen in the posterior end, below the pyrenoid (Fig. 7A & B). The pyrenoid was surrounded by starch grains and the pyrenoid matrix itself was devoid of penetrating thylakoids (Fig. 7B & D). In sectioned material the chloroplast extended from the anterior to the posterior end of the cell only allowing a single layer of ejectisomes between the cell membrane and chloroplast lobes (Fig. 7B). The tubular gullet was protruding from the subapical vestibulum (Fig. 7A & F) and several rows of large ejectisomes (each measuring 0.5 μm × 0.5 μm) surrounded this invagination (Fig. 7A, C, D & F). A microbody was present in the anterior part of the cell (Fig. 7E). The full extension of this structure was not followed. Rows of small ejectisomes (0.1 μm × 0.1 μm) with identical orientation were seen in the cell perimeter (Fig. 7G).
Transmission electron microscopy of Rhodomonas marina (strain S18). (A) Middle (right) view. The gullet (gu) forms a depression in the anterior part. Large ejectisomes (le) are adjoining the gullet. The nucleus (N) is dispositioned in the lower part. Numerous parts of the lobed chloroplast (c) containing paired thylakoids are also shown. (B) Cell in lateral view showing a large central pyrenoid (p) surrounded by starch grains (s). Notice that the peripheral chloroplast extends from anterior to posterior part of the cell. (C) Cross section showing 11 vesicles containing large ejectisomes (le) surrounding the tubular gullet. Small ejectisomes (se) lie close to the cell membrane. (D) Cross section at the level of the pyrenoid (p). The chloroplast forms a horseshoe shape. Parts of the nucleomorph (Nm) penetrate the ventral part of the pyrenoid near the gullet (gu). (E) Microbody (mb) in anterior part of the cell. (F) High magnification of gullet surrounded by several large ejectisomes. A mitochondrion (m) with flattened cristae is also visible. (G) Surface view of posterior part of the cell with numerous small ejectisomes arranged in rows. The mid-ventral band (mvb) is also visible. Scale bars represent: A–D, 2 μm; E, 0.5 μm; F & G, 1 μm.
The structure and orientation of the nucleomorph were studied in a series of sections (Fig. 8A–E). These revealed the nucleomorph to penetrate through the pyrenoid matrix in a tubular, slightly bend arrangement. The nucleomorph was oriented in an anterior-posterior direction and positioned slightly towards the ventral side of the cell (Fig. 8A). It was surrounded by two nuclear membranes and two chloroplast membranes (Fig. 8E).
Transmission electron microscopy of Rhodomonas marina (strain S18). (A) Lateral view displaying the outermost margin of the nucleomorph within the pyrenoid. Nucleomorph reaches from anterior to posterior part of the pyrenoid. (B) High magnification of A. (C) Tubular invagination of nucleomorph reach into the pyrenoid matrix. (D) Nucleomorph forms a complete tube through the pyrenoid matrix. (E) High magnification of the nucleomorph seen in cross section. The nucleomorph has two associated membranes and is surrounded by two chloroplast membranes. Scale bars represent: A, 2 μm; B–D, 1 μm; E, 200 nm.
In the apical end and next to the gullet, a contractile vacuole complex was observed to comprise the typical spongiome surrounding the contractile vacuole itself (Fig. 9). Filling of the contractile vacuole involved fusion and thus emptying of small membranous vesicles (Fig. 9). This fusion involved a rather complex folding of the membrane of the contractile vacuole.
Transmission electron microscopy of Rhodomonas marina (strain S18). High magnification of anterior part showing contractile vacuole (cv) and a large number of adjoining small-sized vacuoles (sv). Some of the smaller vacuoles are seen to fuse with a complex system of membranes enclosing the contractile vacuole. A single mitochondrion (m), the Golgi body (G) and small ejectisomes (se) are also shown. Scale bar represents: 0.5 μm.
DISCUSSION
Phylogeny based on SSU rDNA
The phylogenetic analyses conducted here corroborate several other studies (e.g., Marin et al. 1998, Deane et al. 2002, Shalchian-Tabrizi et al. 2008). If excluding R. minuta (strain CPCC344), the core Rhodomonas / Storeatula / Rhinomonas / Pyrenomonas clade was monophyletic (highly supported by posterior probabilities and BS). The taxonomy of the rarely recorded R. minuta is complicated as it is considered a synonym of Chroomonas minuta (Bourrelly 1970). However, strain CPCC344 did not cluster with any of the other Chroomonas species included and its identity should be reassessed. Instead, R. minuta was part of a highly supported clade with Geminigera-Teleaulax-Plagioselmis (Fig. 1). Hill and Wetherbee (1989) speculated that freshwater species of Rhodomonas (e.g., R. minuta) might belong to a different genus due to their seemingly different morphology. Unfortunately, the morphological study of R. minuta conducted by Hulburt (1965) was insufficient to determine if his material was a true member of Rhodomonas. Features typically used are shape of the inner periplast plates and morphology of the furrow. None of these morphological features were included in the study by Hulburt (1965).
The phylogeny of Rhodomonas based on SSU rDNA revealed it to be polyphyletic as Rhinomonas nottbecki, Rhinomonas pauca, Pyrenomonas helgolandii, and Storeatula major clustered within the core group of Rhodomonas (Fig. 1). This branching pattern has also been seen in other studies (e.g., Cavalier-Smith et al. 1996, Marin et al. 1998, Deane et al. 2002, Hoef-Emden et al. 2002, Shalchian-Tabrizi et al. 2008, Majaneva et al. 2014). Additionally, the branching pattern for many of the Rhodomonas species was not well resolved as indicated by the short branch lengths and identical rDNA sequences. Hence, strains and potentially the taxonomy of Rhodomonas salina, R. maculata, R. atrorosea, R. chrysoidea, R. falcata, Rhinomonas reticulata var. reticulata, and Pyrenomonas helgolandii should be confirmed. This was already suggested by Marin et al. (1998) and Deane et al. (2002) more than two decades ago.
Another strain identified as R. marina (X81373; with no information on strain number or origin in the GenBank accession file) was also included in the phylogenetic analyses. This strain did not form a sister taxon to strain S18 but grouped with a large unresolved clade of Rhodomonas spp., Rhinomonas reticulata, and Pyrenomonas helgolandii. Due to lack of available information, a morphological comparison between this isolate and strain S18 was not possible. This in many ways illustrates the challenges revising the taxonomy of Rhodomonas.
A species of Pyrenomonas (viz. P. helgolandii U. Santore) also clustered within the core group of Rhodomonas species. As already mentioned, Pyrenomonas was abandoned based on Rhodomonas being a valid genus (Hill and Wetherbee 1989). Thus, P. helgolandii should be considered a species of Rhodomonas. Except for the species studied here we are reluctant to conduct any taxonomic changes before additional pheno- and genotypic characters are provided for many of the currently recognized Rhodomonas species. Such work should start with the type species, R. baltica.
Zimmermann (1924), as cited in Hill and Wetherbee (1989), proposed that R. baltica and R. marina were conspecific. Should we accept the identification of strains NIES700 and S18 as representative of R. baltica and R. marina, respectively, then the phylogenetic inference suggests that Zimmermann’s proposal should be disregarded.
In a recent phylogenomic study, Greenwold et al. (2023) examined the phylogeny of 89 cryptophyte strains and two species of Goniomonas. This resulted in a phylogenetic tree characterized by 3 monophyletic clades designated Clade 1–3 (fig. 1 in Greenwold et al. 2023). Clades 1 and 2 were further subdivided into 4 and 7 subclades, respectively; all highly supported. For reasons of comparison, their designated clades including subclades were mapped onto Fig. 1. Clade 1 and subclades 1A–AD were also obtained in the SSU rDNA-based phylogeny, but the branching pattern differed. Monophyly of Clade 2 was not resolved in the SSU rDNA-based tree but individual subclades 2A–2F all formed highly supported lineages (Fig. 1). Comparing the two trees the suggested relationship between subclades 2A–2F differed markedly. It should be noted that in the study by Greenwold et al. (2023), the topology for Clade 2 was not statistically different from a polytomy, hence not strictly bifurcating. Clade 3 comprised species of Goniomonas only. It was paraphyletic in the SSU rDNA tree but monophyletic in the phylogenomic study.
Interestingly, the phylogenetic and evolutionary history of cryptophytes as proposed by the phylogenomic approach taken by Greenwold et al. (2023) still proposes the need for a revision of cryptophyte taxonomy. Such an undertaken is not a trivial matter and should start by characterizing type species using diagnostic features from both electron microscopical observations and molecular analyses.
Phylogeny based on concatenated data matrix
The phylogenetic analyses revealed Rhodomonas to be polyphyletic as Rhinomonas nottbecki clustered within species of Rhodomonas. Due to high statistical branch support, the inference based on concatenated nucleotide sequences (2,800 base pairs) supported a close relationship between strains S18, 08C1, and K-0332. This was substantiated by the highly similar sequence divergence estimates. Typically, internal transcribed spacer sequences are highly variable between species (populations) (e.g., Binzer et al. 2019). So, when strains that are considered to belong to different species exhibit identical ITS sequences, it warrants caution. Here strain K-0332 should be re-identified. This analysis also strongly indicated that strains CCAP979/15, CCAP978/6A, and CCAP979/14 belong to the same taxon, either a species of Rhodomonas or Rhinomonas. Overall, this analysis was hampered by a reduced taxon-sampling as the number of Rhodomonas species / strains for which ITS sequences are available was limited.
Predicted secondary structure of ITS2
Inclusions of ITS2 secondary structures in taxonomical classification of cryptophytes have been pioneered by Hoef-Emden (2007, 2018) and Hoef-Emden and Melkonian (2003). This approach has been followed by others, for example, Majaneva et al. 2014, Martynenko et al. 2020, Gusev et al. 2022, and Khanaychenko et al. 2022. Likely, the inclusion of secondary structure models has arisen due to challenges of using morphological features to delineate species and thus obtain a solid taxonomy. The predicted secondary structures of helices I–IV in strain S18 (Fig. 3) revealed significant differences to those available of R. storeatuloformis, R. lens, and Rhodomonas strains CCMP743, CCMP760, and CCMP763 (Khanaychenko et al. 2022) but also Rhinomonas nottbeckii, R. reticulata (CCAP979/15), and Rhinomonas sp. (CCMP740) (Majaneva et al. 2014).
In strain S18, helix I did not possess an internal bulge as seen in R. storeauloformis and Rhodomonas strains CCMP743 and CCMP763. As in R. lens it had a single-nucleotide bulge after the first 6 base pairs (a C in S18 and an A in R. lens [CCMP739]). Except for R. lens, helix II was similar in strains S18, IBSS-59 (R. storeauloformis), CCMP743 and CCMP763. Here 9 out of the first 11 base pairs were identical. The first difference was a non-canonical G G in S18 and CCMP760 and a G-C in R. storeauloformis, CCMP743 and CCMP763 (the fourth base pair). The second difference was a C-G base pair in S18 and CCMP760 and a U-A base pair in R. storeauloformis, CCMP743 and CCMP763 (the fifth base pair). Some of the most noticeable differences were observed for helix III. Though the total number of base pair (including non-canonical base pairings but excluding single-nucleotide bulges) was similar (ranged between 41 and 47; with 45 in strain S18), the composition of the base pairs differed markedly. However, the hairpin was somewhat similar between this species. It had 3–5 base pairs in the stem and 3–7 bases in the terminal loop. The stem of the hairpin always started with three identical base pairs (C-G, C-G, and U-A). Except for the first base pair in R. storeauloformis and CCMP763, the nucleotides in the stem of helix IV all formed base pairs. The length of the stem varied between 9 and 11 base pairs as did the composition of these for most of the strains. However, 8 out of the first 10 base pairs were identical comparing S18 and R. lens (base pairs 6 and 8 of the stem varied with a single nucleotide in each).
Based on this comparison strain S18 comprised several unique molecular signatures and these together with V4 of the SSU rDNA were used in the emendation of R. marina, see below.
Sequence divergence
The sequence divergence comparison supported the view that identification of cryptophyte species can be difficult when based solely on light microscopy. The identical SSU rDNA sequences for the clade with 11 strains of Rhodomonas (Pyrenomonas, Rhinomonas and six species of Rhodomonas) implied that these may only represent a single taxon. If future studies show this to be the case, the species diversity of Rhodomonas will be markedly reduced and a taxonomic revision necessary. This is likely also needed for the cryptophyte strains with identical SSU rDNA sequences as marked in Fig. 1.
It should be noted here that strain K-0435 (available in GenBank with accession No. HF952572) and identified as R. marina was 100% identical to strain S18 (R. marina) and R. cf. abbreviata (CCMP1178). This explains why it was not included in phylogenetic analyses shown in Fig. 1 and the pairwise comparisons of sequence divergences.
Periplast, small ejectisomes, and mid-ventral band
The periplast plates appeared rectangular with rounded edges and surrounding prominent ridges (Fig. 6). The ridges in the periplast appeared to accommodate the ejectisomes, which was in accordance with observations made by Hill and Wetherbee (1989) when studying R. maculata and R. duplex D. R. A. Hill and Wetherbee. The morphology of the periplast differed in the lower most part of the cell. Here the periplast and its ridges appeared to form a sheet. The arrangement of ejectisomes also changed (Fig. 6H). Smaller periplast plates in the posterior end have also been observed in R. salina and R. maculata (Hill and Wetherbee 1989) as well as in Rhodomonas sp. 3 (Cerino and Zingone 2006). The explanation for the change in size and morphology of the periplast plates needs further study.
The presence of a mid-ventral band has also been observed in other species of the genus (e.g., R. maculata, R. duplex, and R. stigmatica) (Hill and Wetherbee 1989, Hill 1991) but it lacks in R. baltica and R. salina (Hill and Wetherbee 1989) and in the recently described R. storeatuloformis (Khanaychenko et al. 2022). The mid-ventral band may be used in separating species. However, it should not be considered a synapomorphy of Rhodomonas as a very similar structure has also been observed in a few distantly related cryptophytes, e.g., Baffinella frigidus (Daugbjerg et al. 2018), Plagioselmis prolonga (now called the Plagioselmis stage of Teleaulax amphioxeia) (Altenburger et al. 2020), and Plagioselmis sulcata (Cerino and Zingone 2006). When present in Rhodomonas, the length of the mid-ventral band varies. In R. maculata it extends ca. 1/4 of the cell length (fig. 24 in Hill and Wetherbee 1989) whereas in Rhodomonas sp. 1 from the Gulf of Naples (Mediterranean Sea), it runs from the bottom of the furrow to the antapical end of the cell (or 3/5 of the cell length) (fig. 33 in Cerino and Zingone 2006). The function of the mid-ventral band has yet to be fully elucidated, but it has been speculated that it together with ejectisomes in the posterior end might play a role in cell fusion (Altenburger et al. 2020). This is because the mid-ventral band is clearly expressed in the Plagioselmis stage of Teleaulax amphioxeia (the haploid stage) and either short or absent in the diploid stage of Teleaulax (Altenburger et al. 2020).
Furrow-gullet complex
A tubular gullet (Fig. 7A & F) and a furrow (Fig. 6A & E) extended towards the posterior end from a subapical depression (Figs 5A, B, 6A, D–F & 7A).
Though a characteristic feature of many Rhodomonas species (Table 1), the furrow-gullet complex is not unique to the genus (and even absent in R. storeatuloformis, Khanaychenko et al. 2022). The same complex has also been shown in Baffinella (Daugbjerg et al. 2018), Geminigera (Hill 1991), and Cryptomonas (Hoef-Emden and Melkonian 2003). Following the phylogenetic analysis conducted here (Fig. 1) and elsewhere (e.g., Deane et al. 2002, Hoef-Emden et al. 2002) these cryptophytes do not share a single common ancestor. Hence, the furrow-gullet complex has evolved multiple times as indicated by the phylogenetic analysis shown here and likely lost in R. storeatuloformis. When aiming for a natural classification system based on common ancestry, the furrow-gullet complex is therefore of limited taxonomic value for defining cryptophytes at the genus level. Other morphological and chemical traits (e.g., shape of inner periplast components, position of nucleomorph, type of phycobiliprotein) similarly do not reflect a derived (synapomorphic) status. The multiple origins (or losses) of these phenotypic traits illustrate the complex evolutionary history of cryptophytes and thus also explain why their taxonomy has been and still is notoriously difficult to establish when based solely on morphological characters (e.g., Solarska et al. 2023).
Comparison of 19 species of Rhodomonas scored for nine characters
Contractile vacuole
Small, contributing vacuoles which seemed to originate from the Golgi-body fused into the single large contractile vacuole in the apical end (Fig. 9). Contractile vacuoles have been observed in all species of Rhodomonas (Erata and Chihara 1989), but the function remains unknown in marine members which are considered isotonic. Typically, contractile vacuoles are found in freshwater flagellates where they have an osmoregulatory function. In marine species, contractile vacuoles might be seen as an adaptation to cope with salinity changes (Santore 1984). In the case of R. marina, which is most often observed in coastal surface waters characterized by salinity fluctuations due to rain fall, freshwater discharge from land, and sea ice melt particularly in Arctic regions, a contractile vacuole will thus help to maintain an osmotic balance of the cell. Contractile vacuoles have also been speculated to function as ways to transport products to the cell surface and for elimination of waste substances (Hausmann et al. 2003).
Chloroplast, pyrenoid, and nucleomorph
The series of stacked images revealed a deeply lobed (H-shaped) chloroplast (Fig. 5I–L). To the best of our knowledge, such an outline of a chloroplast in cryptophytes has not been documented before. However, the drawing of Storeatula major included as fig. 5 in Hill 1991 illustrated a deeply lobed chloroplast but the lobes were not easy to see in the single light micrograph included as fig. 34. Future observations applying advanced bio-imaging techniques will reveal whether chloroplasts in other cryptophytes are lobed similarly. Obviously, observations based on single images from conventional light microscopy may not provide the true outline of chloroplasts compared to 3D imaging. The series of stacked images also revealed DAPI-stained dots situated more or less in the perforations in the chloroplast (Fig. 5L). These dots, considered chloroplast nucleoids, are known to exist in Cryptophyceae (Kuroiwa 1991), where they are dispersed between the thylakoid membranes in the stromal regions of the chloroplast. Chloroplast nucleoids have also been observed in Rhodomonas helgolandii (as Pyrenomonas helgolandii) by Sato et al. (2014).
The pyrenoid occupied a major part of the cell center towards the dorsal side (Fig. 7B & D). It was not traversed by thylakoids but instead by a nucleomorph which formed a tubed penetration through the pyrenoid matrix (Fig. 8D). This was in accordance with the arrangement of the nucleomorph in Rhodomonas ovalis (Kugrens et al. 1999) as well as in R. maculata and R. baltica (Hill and Wetherbee 1989). This arrangement has also seen in the closely related Storeatula major D. R. A. Hill (Hill 1991), which clustered within the Rhodomonas clade (Fig. 1).
Evaluating the identity of Rhodomonas marina
The original description of Cryptomonas marina (currently accepted as Rhodomonas marina) by Dangeard (1892) was sparse. He described it to possess a single chloroplast with a yellow to yellow-brown color. This agrees with our observations of R. marina. Dangeard (1892) also noted a large centrally located oily drop now considered to be the pyrenoid surrounded by a starch sheet. The cell shape was also similar compared to our Arctic isolate of R. marina (i.e., twice as long as wide and with a pointed posterior at least in some cells). Dangeard also observed two flagella which emanated from a vestibulum, and these were approximately the same length as the cell. In this study, some variation was noted regarding the length of the flagella as they ranged from half the cell length to equal the cell length (data not shown). These differences may reflect observations of different asexual division stages in the culture studied. Thus, the description of C. marina (Dangeard 1892) agreed with our observations. However, this description fits many species of Rhodomonas when examined under the light microscope. Dangeard (1892) did not measure the length of the cell but in the work by Büttner (1910) cell dimensions were given as 18–22 μm in length and 6–10 μm in width. These dimensions agreed with the morphometric measurements of R. marina made here. Other species of Rhodomonas with the same size are the marine species R. baltica and R. stigmatica as well as the freshwater species R. tenuis and R. lens (Hill and Wetherbee 1989, Hill 1991, Moustaka-Gouni 1996, Javornický 2001) (Table 1). Since Dangeard (1892) and Büttner (1910) both observed R. marina in marine waters, the freshwater species are not considered further. Rhodomonas stigmatica possesses two pyrenoids (Butcher 1967, Hill 1991), which were not observed by Dangeard (1892) and though Büttner (1910) described C. marina (now R. marina) as having two chloroplasts, he only described a single pyrenoid. It seems fair to reason that observations of chloroplast number were a challenging task in the early 1900 as a single deeply lobed chloroplast may easily appear as two. However, as only one pyrenoid was observed we find it likely that only one chloroplast was in fact present in the material studied by Büttner (1910). In all currently described Rhodomonas species only one pyrenoid is associated with a single chloroplast (Table 1). Rhodomonas baltica and R. marina have been considered to be conspecific (Zimmermann 1924, as cited in Hill and Wetherbee 1989). However, R. baltica has only been observed having a rounded posterior (Karsten 1898, Hill and Wetherbee 1989, Hill et al. 1992), while R. marina possessed a pointed posterior (Dangeard 1892, Büttner 1910, Hill et al. 1992) although this may vary as shown here. The observations made by Büttner (1910) therefore also concur with our study. Admittedly, R. marina and R. baltica are morphologically very similar but they may be distinguished by the difference in the length of the furrow (1/6 of the cell length in R. marina and 1/8 of the cell length in R. baltica) and the absence of a mid-ventral band in R. baltica (Table 1). Based on these morphological differences and the different nuclear-encoded SSU rDNA sequences it was concluded that strain S18 studied here represented R. marina and therefore separate from R. baltica.
To further examine the identity of R. marina it was also compared to 3 other species where detailed studies have been conducted (Table 1). Rhodomonas marina is similar to R. duplex in having a similar length of the furrow (1/6 and 1/5 of the cell length, respectively) and a mid-ventral band. However, R. duplex is smaller than R. marina and it has two chloroplasts each associated with a pyrenoid. Additionally, in R. marina the inner periplast plates near the mid-ventral band become a sheet, which is not the case for R. duplex (Hill and Wetherbee 1989). Rhodomonas maculata also appears similar to R. marina due to similar length of the furrow (1/6 and 1/5 of the cell length, respectively) and the mid-ventral band. Their cell dimensions were similar, and both species possessed just one chloroplast. However, the chloroplast in R. maculata is small (Hill and Wetherbee 1989) and does not reach the full length of the cell as in R. marina. These two species are further separated by cell shape and the inner periplast. The cell shape of R. maculata is cylindrical, whereas in R. marina it is elliptical. The inner periplast plates become a sheet towards the posterior in R. marina, which is not the case in R. maculata (Hill and Wetherbee 1989). The morphological distinctions between R. marina and R. duplex / R. maculata was also supported by the phylogenetic inference (Fig. 1). Here R. duplex and R. maculata were distantly related to R. marina (strain S18). The lack of a mid-ventral band and the somewhat shorter furrow relative to the cell length in R. salina clearly separates this species from R. marina.
Comparison of Rhodomonas to related genera
Storeatula and Rhinomonas represent the two most closely related genera to Rhodomonas (Fig. 1). They possess an ellipsoid to oval cell shape (with some variations), an intra-pyrenoidal nucleomorph and a pyrenoid not traversed by thylakoids (Santore 1984, Clay et al. 1999, Kugrens et al. 1999, Hoef-Emden et al. 2002, Novarino 2012). Due to differences in arrangement of the inner periplast this character has been used to delineate these genera (Majaneva et al. 2014). The inner periplast formed a sheet in Storeatula (Hill 1991), hexagonal plates in Rhinomonas (Hill and Wetherbee 1988), and rectangular plates in Rhodomonas (Hill and Wetherbee 1989, this study). Except for Rhodomonas storeatuloformis all other species in this genus possesses a furrow-gullet complex. However, the furrow-gullet complex is not present in Storeatula and Rhinomonas. Storeatula and Rhinomonas only possess a gullet (Hill and Wetherbee 1988, Hill 1991, Kugrens et al. 1999, Majaneva et al. 2014). Nonetheless, the grouping of these three genera was supported by the pigment profile, and is in accordance with the pigment profile found in current study. All possess the water-soluble pigment Cr-PE 545 (Greenwold et al. 2019). If future studies reveal that Rhinomonas and Storeatula are part of the same life cycle (dimorphism as suggested by Majaneva et al. 2014), then Rhinomonas has priority, and it will be the only known genus closely related to Rhodomonas.
Dimorphism
At least four examples of dimorphism have been described in the Cryptophyceae. The most recent example was provided by Altenburger et al. (2020) when showing dimorphism in Teleaulax amphioxeia and Plagioselmis prolonga. Of relevance to this study, Altenburger et al. (2020) and Majaneva et al. (2014) suggested that Rhinomonas and Storeatula (Pyrenomonadaceae) represent the cryptomorph (haploid) and campylomorph (diploid) stages, respectively, of the same species. The shared morphological feature that unites them is the lack of a furrow, which is otherwise a characteristic feature of Rhodomonas. If correct, this leaves only two genera in the Pyrenomonadaceae (Rhinomonas [valid publication of R. pauca by Hill and Wetherbee 1988] and Rhodomonas). To the best of our knowledge, dimorphism has not yet been detected in Rhodomonas.
Emendation of Rhodomonas marina
Based on the morphological and molecular data presented here an emendation is proposed.
Rhodomonas marina (P. A. Dangeard) Lemmermann emend. Daugbjerg & Devantier. Fig. 5A & B. Dangeard [1892, p. 32, Pl. II, fig. 20, original drawing].
Epitype
A stub with material of strain S18 prepared for scanning electron microscopy has been deposited at the Natural History Museum of Denmark and given the reference number: C-A-99713.
Lectotype
Designated here, Pl. II, fig. 20 in Dangeard 1892.
Emended diagnosis
Elliptical cells 14–22 μm long and 6–12 μm wide. Truncated anterior in lateral view and rounded to pointed posterior. Two flagella unequal in length emanate from vestibulum. With furrow-gullet system lined with ejectisomes. Single deeply lobed chloroplast yellow-brown in color, containing central pyrenoid encircled by starch grains. Tubular nucleomorph penetrates pyrenoid matrix. Inner periplast composed of rectangular plates with rounded edges. Nucleus in posterior end and contractile vacuole in anterior end. Contains Cr-PE 545.
Species differs from others of the genus by nuclear- encoded SSU rDNA and V4 has the following sequence V4:5′-GTCGGGCTCGGGGAGATTGTCGGCCTTTGGTCGGATAGTTTCTCCGGGCCTTTCTGCCTGGGAACCCTATTC<+141 bp upstream>-3′ (synapomorphic characters are underlined).
Diagnostic molecular characters of ITS2′s helix I: [5′-CAUCACCUAAAACCUCGGUUUUGGCGUGGAG-3′] [(.(((((((((((....))))))).)))).)].
Diagnostic molecular characters of ITS2′s helix II: 5′-UGAGCGUCAUGGCUG-CCUUUGUUGGUAGCCAC GUCGGUCG-3′] [(((.((...((((((((......))))))))...)).)))].
Diagnostic molecular characters of ITS2′s helix III: [5′-GGUGACGCCUCACGACC-GAAAGGUAUGUGGUG UGCCGUGAUGUAGCAGCCUUCUCGUAAGGAGC AACAUCAUCGUACAUUAAGCCAUAUGCCAUCAAG UAGGCGCGUGGUUUGCACU-3′][(((((.(((.(.((.((....((((((((((((((.(((((((.((..((((.....)))).)).))))))).))))).....)))))))))........)))).).))).).))))].
Diagnostic molecular characters of ITS2’s helix IV: [5′-UUUAGGAGUAUUUAUACUCUUAGG-3′][((((((((((.... ))))))))))].
ACKNOWLEDGEMENTS
We thank Øjvind Moestrup for help with the TEM fixation and access to his library on microalgae. Lis Munk Frederiksen is thanked for thin sectioning of the material. ND thanks Carlsbergfondet (2012-01-0509, 2013-01-0259) and Brødrene Hartmanns Fond (A22920) for equipment grants. We thank two anonymous reviewers for their comments on an earlier version of the manuscript.
Notes
The authors declare that they have no potential conflicts of interest.
SUPPLEMENTARY MATERIALS
Map of sampling location (■) of Rhodomonas marina in Disko Bay, western Greenland (https://www.e-algae.org).