INTRODUCTION
Desmodesmus species (Scenedesmaceae, Chlorophyta) commonly inhabit freshwater ecosystems and have been reported in regions such as America, Georgia, New Zealand, and Korea (Johnson et al. 2007, Broady et al. 2012, Barinova and Kukhaleishvili 2014, Kim et al. 2020). The genus Desmodesmus is typically characterized by two- to four-celled coenobia with spines and diverse types of ultrastructures (e.g., ribs, warts, and rosettes) present on their cell surfaces (An et al. 1999, Lone et al. 2021). The coenobial cells in this genus are arranged linearly or alternately (i.e., in a zig-zag), which is generally observed in the family Scenedesmaceae (Cho and Lee 2024). Desmodesmus species typically have long and short spines, which distinguish them from morphologically similar species in the genus Scenedesmus (Scenedesmaceae; Trainor et al. 1976, Hegewald 1997, Shubert and Gärtner 2015). The ribs consist of a series of linearly arranged tubes on the cell surfaces, forming teeth-like structures. The warts show diverse types of cell surface ultrastructures, such as simple round, elongated, armed warts, and net-like structures (Hegewald and Schnepf 1991). Cell surfaces in the genus Desmodesmus show a range of rosettes, appearing as chimney-like, knob-like, covered with a net-like structure, and opening-like structure (Hegewald 1997). However, important ultrastructures such as short spines (e.g., the tooth) and teeth-like ribs in Desmodesmus species are often overlooked during light microscopic (LM) observation (Mai et al. 2023).
The genus Desmodesmus was reestablished through phylogenetic analysis using 18S ribosomal RNA (rRNA) and internal transcribed spacer 2 (ITS-2) sequences derived from the genus Scenedesmus (Hegewald 1997, Kessler et al. 1997, An et al. 1999, Hegewald 2000, Shubert and Gärtner 2015). In particular, the ITS regions show greater sequence variation for phylogenetic analysis, providing valuable information for comparing closely related Desmodesmus species (An et al. 1999, Van Hannen et al. 2002). However, 18S rRNA sequences of Desmodesmus species provide essential information for species identification (Kessler et al. 1997, Shubert and Gärtner 2015). Therefore, rRNA regions, including 18S rRNA sequences, a widely used molecular marker, play a crucial role in species identification and phylogenetic analysis of Scenedesmaceae species (Fawley et al. 2011, Mai et al. 2023, Cho and Lee 2024). However, many studies on the molecular characterization and phylogenetic analysis of Desmodesmus species rely solely on LM observations, despite significant differences in molecular sequences and phylogenetic relationship, as seen in D. communis-related taxa (Akgül et al. 2017). Therefore, reports of 18S rRNA sequences in Desmodesmus species with ultrastructures (e.g., cell surfaces, spines, and ribs) observed under electron microscopy are increasingly important, particularly for species identification (Fawley et al. 2011, Cho and Lee 2024, Demura 2024, Hegedűs et al. 2024).
This study aims to report two novel species (D. similis and D. complicatus) and four novel intraspecific taxa (D. communis var. columnaris, D. denticulatus var. denticuloides, D. denticulatus var. simplex, and D. denticulatus var. distinctus) within the genus Desmodesmus. In addition, we propose transferring Scenedesmus praetervisus to the genus Desmodesmus based on morphological characteristics and molecular evidence. This study could enhance our understanding of ultrastructures of Desmodesmus species, facilitating accurate species identification.
MATERIALS AND METHODS
Morphological observation of algal strains
The algal strains (CCAP 258/114, CCAP 258/122, FBCC-A0691, FBCC-A0718, FBCC-A0719, FBCC-A0724, FBCC-A0725, FBCC-A0726, and FBCC-A1393) were obtained from the Culture Collection of Algae and Protozoa (CCAP, UK) and the Freshwater Bioresources Culture Collection (FBCC) at the Nakdonggang National Institute of Biological Resources (NNIBR, Republic of Korea). They were cultured in Bold’s Basal Medium (Nichols and Bold 1965) under 12: 12 h light: dark photocycle, with a light intensity of 100 μmol m−2 s−1 in a 25°C incubation chamber. To accurately observe the morphological characteristics of green algal cells, we conducted microscopic observations of all algal strains within 2 weeks after subculturing, with sufficient nutrient availability (Cho and Lee 2024). We used a LM (ECLIPSE Ni-U; Nikon, Tokyo, Japan) and a scanning electron microscope (SEM; SU8220; Hitachi, Tokyo, Japan) to observe the morphology of the algal strains. Cell size measurements were conducted on 100 cells using the LM. For SEM observation, algal cells were fixed with 2% glutaraldehyde, mucilage was removed using 100% cetyltrimethylammonium bromide buffer (Tavera and Calderón 2013), and the cells were dehydrated through a series of ethanol concentrations (10, 20, 30, 40, 50, 60, 70, 80, 90, and 100%). The dehydrated samples were treated with critical point drying (HCP-2; Hitachi), followed by platinum coating (Cressington 208HR; CRESSINGTON, Watford, UK) for 90 s. We observed and described the conserved morphological features of the algal strains under LM and SEM.
DNA extraction and polymerase chain reaction amplification
We used the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) for DNA extraction from algal cells. The 18S rRNA–ITS-1–5.8S rRNA–ITS-2 regions were amplified using polymerase chain reactions (PCRs) with diverse primer pairs (Table 1, Supplementary Fig. S1). The PCR experiments were performed using AccuPower PCR PreMix (Bioneer, Daejeon, Korea) for 35 cycles (denaturation at 95°C for 30 s, annealing at 45°C for 30 s, and extension at 72°C for 1 min) with a pre-denaturation step at 95°C for 2 min and a final extension at 72°C for 15 min. Additional PCR experiments were conducted using KOD FX Neo (Toyobo, Osaka, Japan) for 35 cycles under the above conditions, with the annealing temperature modified to 55°C. The PCR products were purified using the LaboPass PCR Purification Kit and Gel extraction Kit (Cosmogenetech, Seoul, Korea). Sanger sequencing of the purified DNA was performed using an Applied Biosystems 3730xl DNA Analyzer (Applied Biosystems, Foster City, CA, USA) according to the instructions of the manufacturer at the Macrogen sequencing facility (Macrogen Inc., Seoul, Korea). We aligned and trimmed the DNA sequences obtained from Sanger sequencing using Geneious Prime (v2025.0.3; https://www.geneious.com).
Phylogenetic analysis
We collected homologous sequences (i.e., top 250 matches) for the 18S rRNA, ITS-2, rbcL, and tufA sequences of our Desmodesmus strains using the web-BLASTn search results (nt database; e-value cutoff = 1.e−10). To ensure more accurate taxon sampling, we also retrieved all available 18S rRNA, ITS-2, rbcL, and tufA sequences annotated as Desmodesmus species from the National Center for Biotechnology Information (NCBI) database. These sequences were manually curated and trimmed (e.g., removal of <500 bp 18S rRNA regions, introns and unverified data) based on alignments generated using MAFFT (v7.450) (Katoh and Standley 2013) with default options. From this dataset, we selected sequences with up to 10 nucleotide differences relative to each target to collect the most similar representatives. Redundant sequences were removed, and homologous datasets were merged. Each homologous dataset was aligned using MAFFT (v7.450) with default options. The intron sequences of the alignments were manually trimmed in Geneious Prime (v2025.0.3) using its alignment viewer function. A maximum-likelihood (ML) tree was constructed from the alignment using the IQ-tree program (v1.6.12) (Nguyen et al. 2015) with the following options: the model test (-m TEST), ultrafast bootstrapping with 1,000 replications (-bb 1,000), and SH-aLRT with 1,000 replications (-alrt 1,000). The best-fit models (‘TIM2 + F + I + G4’ for 18S rRNA, ‘SYM + I + G4’ for ITS-2, and ‘GTR + F + I + G4’ for rbcL and tufA alignments) were used for the phylogenetic analysis. The outgroup clade in each phylogeny was selected based on three Tetradesmus species (FBCC-A0708, SAG 5.95, and SAG 22.81) (Cho and Lee 2024) and other taxonomically distant groups.
Genetic distance analysis
Genetic distances were calculated using p-distance with pairwise deletion for gaps/missing in MEGA11 (v11.0.13) from trimmed alignments of the 18S rRNA and ITS-2 (Tamura et al. 2021).
Analysis of species hypotheses using the ASAP program
The Assemble Species by Automatic Partitioning (ASAP) program enables the construction of species partitions for target taxa based on genetic distances derived from the single locus sequence alignments (Puillandre et al. 2021). To test species hypothesis using ASAP, we selected phylogenetically clustered (ingroup) Desmodesmus taxa based on our phylogenetic analyses of the 18S rRNA, ITS-2, rbcL, and tufA sequences. Several partial sequences that were not comparable to our target taxa in each dataset were excluded. Each trimmed dataset was aligned using MAFFT (v7.450) with default options. We evaluated species hypotheses (i.e., species delimitations) for the target Desmodesmus taxa related to D. communis and D. denticulatus, based on the top five ranked partitions (two best and three alternatives) from the ASAP results (Puillandre et al. 2021). The species and intraspecific boundaries of the target Desmodesmus taxa were discussed based on both the ASAP results and conventional taxonomic considerations (e.g., ultrastructures, notable genetic variations, and phylogenetic relationships), as guided by Puillandre et al. (2021).
All information necessary to reanalyze the data (e.g., alignments) in this study is available upon request from the corresponding author (junmolee@knu.ac.kr).
RESULTS AND DISCUSSION
Morphological and molecular characterization of Desmodesmus species
The cells of FBCC-A0726 are arranged linearly in a two- to four-celled coenobium, and long spines are present on each terminal cell (Fig. 1A). The inner and outer cells show rounded and tapered apices, respectively. The coenobial cells of FBCC-A0726 frequently show an outermost cell wall-like structure (i.e., mucilage-like) under LM (Fig. 1A), although this feature was not always observed. Short tube-like structures composed of slightly elevated armed warts are present with a central hole on a net-like structure of the cell surface (Fig. 1B). The rosettes consist of bundles of column-like structures formed by five to six closed tubes at their apex (Fig. 1C). Most morphological features of FBCC-A0726 are similar to those described for Desmodesmus communis (Hegewald and Silva 1988, Hegewald 2000, Bica et al. 2012, Hegewald and Braband 2017), such as cell wall-like structures (mucilage-like) observed under LM and armed warts observed under SEM. However, the rosette features differ from those of D. communis, which shows rosettes composed of four to five perforated tubes as observed under SEM (Bica et al. 2012, Hegewald and Braband 2017) (Table 2). In addition, the ITS-2 sequence of the FBCC-A0726 differs by four to five bp from that of D. communis (Fig. 2) (Bica et al. 2012, Hegewald and Braband 2017). The 18S rRNA sequences of the FBCC-A0726 strain contain one nucleotide difference and an intron region (392 bp) compared to those of D. communis AICB 993 (Fig. 2). Although FBCC-A0726 shows distinct morphological and molecular characteristics (e.g., rosettes and an intron insertion) compared to D. communis (Hegewald 1977-170 and AICB 993), its phylogenetic position is not clearly separated from D. communis (see details in the following section). Based on the results, we suggest that FBCC-A0726 is D. communis var. columnaris, a novel variety of D. communis in the genus Desmodesmus.
The cells of FBCC-A0691 are linearly arranged in coenobia composed of two to four cells, with long spines present on each terminal cell (Fig. 1D). The inner and outer cells have rounded and tapered apices, respectively. The coenobial cells of FBCC-A0691 show an outermost cell wall-like structure (i.e., mucilage-like) as observed under LM (Fig. 1D). On a net-like structure of the cell surface, short tube-like structures composed of slightly elevated armed warts are present, including a central hole (Fig. 1E). The rosettes are located on the cell surface and are composed of four to six tubes (Fig. 1F). Most morphological features of FBCC-A0691 are similar to those described for D. communis such as cell wall-like structures (mucilage-like) observed under LM, and armed warts and four to six tube rosettes observed under SEM (Hegewald and Silva 1988, Hegewald 2000, Bica et al. 2012, Hegewald and Braband 2017) (Table 2). However, molecular markers (e.g., ITS-2 and 18S rRNA) of FBCC-A0691 are distinct from those of D. communis strains (Hegewald 1977-170 and AICB 993), which have been validated by molecular and morphological features (Bica et al. 2012, Hegewald and Braband 2017). For example, the ITS-2 sequence of the FBCC-A0691 strain differs from that of D. communis (Hegewald 1977-170 and AICB 993) by five to six nucleotides, and the 18S rRNA sequence differs from that of D. communis AICB 993 by nine nucleotides (Fig. 2). In addition, FBCC-A0691 shows a distinct phylogenetic position from D. communis (Hegewald 1977-170 and AICB 993; see details in the following section). Therefore, although FBCC-A0691 is morphologically similar to D. communis, it is suggested as a novel species, D. similis, in the genus Desmodesmus, based on distinct molecular differences and its phylogenetic position.
The coenobial cells of FBCC-A1393 are arranged linearly in a two- to four-celled coenobium, and long spines are present on each terminal cell (Fig. 1G). The inner cells have slightly tapered apices, while the outer cells show tapered apices. The coenobial cells of FBCC-A1393 show an outermost cell wall-like structure (i.e., mucilage-like) as observed under LM (Fig. 1G). Morphological features of FBCC-A1393 observed under LM are also similar to those of D. communis, FBCC-A0691, and FBCC-A0726. However, the warts on the cell surfaces of FBCC-A1393 were not prominent, and rib-like structures are formed by slightly folded cell surface structures arranged linearly (Fig. 1H). The rosettes are located at the margins of the terminal cells and near the rib-like structures of the inner cells, which is composed of four to six tubes (Table 2, Fig. 1I). In addition, the ITS-2 and 18S rRNA sequences of FBCC-A1393 differ significantly by four and six bp from those of D. communis Hegewald 1977-170 and AICB 993, respectively (Fig. 2). Moreover, FBCC-A1393 also shows a distinct phylogenetic position from D. communis (Hegewald 1977-170 and AICB 993) as well as from D. similis FBCC-A0691 (see details in the following section). Therefore, we suggest that FBCC-A1393 is D. complicatus as a novel species in the genus Desmodesmus.
Although type strains of D. communis are currently unavailable, Hegewald, who first reported D. communis (Hegewald 1977), later re-described the species using ITS-2 sequences (Hegewald 1977-170 strain; KU359282.1) and transmission electron microscopy images (Hegewald 1975-135 strain) in a more recent study (these strains are also not available; Hegewald and Braband 2017). However, the outermost cell wall-like structure (i.e., mucilage-like) in Desmodesmus species is commonly observed under LM, which leads to frequent classification of these taxa as D. communis. As a result, many DNA sequences from morphologically unverified D. communis have been reported in the NCBI database (see details in the following section). In contrast to these unverified taxa, we found that three FBCC strains (FBCC-A0691, FBCC-A0726, and FBCC-A1393) show distinct ultrastructural morphologies as observed under SEM compared to D. communis (Bica et al. 2012, Hegewald and Braband 2017) (Fig. 1). Furthermore, while most ultrastructural features of FBCC-A0691 are similar to those of D. communis, its rRNA sequences are significantly different. Therefore, we suggested that SEM observation and molecular evidence are essential to classify and distinguish D. communis-like species recognized under LM observation. As a result, we suggested FBCC-A0726 as a novel intraspecific taxon of D. communis, while FBCC-A0691 and FBCC-A1393 as novel species in the genus Desmodesmus.
The coenobial cells of CCAP 258/114, as observed under LM, are alternately arranged (i.e., in a zig-zag pattern) and typically consist of two to four cells with short spines. These cells have an outermost cell wall-like structure (i.e., mucilage-like) (Fig. 3A). Two types of warts (large round warts and long elongate tube warts) are frequently observed on the cell surface (Fig. 3B). Short spines are present around the cell apex (Fig. 3B). The rosettes are surrounded by a simple wall with a net-like structure (Fig. 3C). The alternately arranged coenobial cells with short spines under LM observation are typical morphological features of Desmodesmus denticulatus (Hegewald and Silva 1988, An et al. 1999, Hegewald 2000). However, D. denticulatus has a well-developed net-like structure with armed warts on the cell surface, which are recognizable only under SEM (An 1996-12 strain) (Hegewald 1997, An et al. 1999) (Table 2). These features are distinctly different from those observed in CCAP 258/114 (Fig. 3B). In addition, the ITS-2 sequences of D. denticulatus An 1996-12 contain a 19 bp intron and four nucleotide differences compared to those of CCAP 258/114 (Fig. 4). Although the morphological and molecular features between CCAP 258/114 and D. denticulatus An 1996-12 are clearly distinct, their separation of phylogenetic relationship is not well resolved enough to distinguish them at the species level (see details in the following section). Based on the results, we suggest that CCAP 258/114 is D. denticulatus var. denticuloides, a novel variety of D. denticulatus in the genus Desmodesmus.
The coenobial cells of FBCC-A0719 are arranged alternately (i.e., in a zig-zag pattern) and typically consist of two to four cells, which have an outermost cell wall-like structure (i.e., mucilage-like) as observed under LM (Fig. 3D). Short spines are arranged along the cell edges and small simple round (granule-like) warts with a net-like structure are present on the cell surface (Fig. 3E). The short spines consist of multiple tubes, observable only under SEM. The rosettes are surrounded by a simple wall (Fig. 3F). The morphological features observed under LM are similar to those of D. denticulatus, but the simple net-like structure observed under SEM in FBCC-A0719 is significantly different from that of D. denticulatus (Hegewald 1997, An et al. 1999) (Table 2). In addition, the ITS-2 region of FBCC-A0719 lacks an intron, similar to D. denticulatus var. denticuloides CCAP 258/114, but unlike D. denticulatus (Fig. 4). Interestingly, FBCC-A0719 contains a 358 bp intron in the 18S rRNA region and three nucleotide differences in the ITS-1 region compared to D. denticulatus var. denticuloides CCAP 258/114, but they share identical 18S rRNA and ITS-2 sequences, excluding the intron sequence (Fig. 4). In addition, the phylogenetic position of FBCC-A0719 is also not clearly separated from D. denticulatus An 1996-12 (see details in the following section). Therefore, we suggest that FBCC-A0719 is D. denticulatus var. simplex, a novel variety of D. denticulatus, based on the distinct morphological and molecular characteristics (Figs 3 & 4). FBCC-A0718 shares identical morphological and molecular features (i.e., 18S rRNA, ITS, rbcL, and tufA sequences) with D. denticulatus var. simplex FBCC-A0719. In addition, the collection sites and dates of FBCC-A0718 (Jeju-si, Jeju-do, Republic of Korea; May 19, 2017) and FBCC-A0719 (Jangheung-gun, Jeollanam-do, Republic of Korea; Aug 7, 2017) are different. Therefore, these strains (FBCC-A0718 and FBCC-A0719) are regarded as different populations, and their intron insertions are conserved at the population level.
The intron insertion variability, coupled with distinct morphological features, could occur among intraspecific taxa in the family Scenedesmaceae (Cho and Lee 2024). For example, a previous study reported two morphologically distinguishable algal taxa that shared identical 18S rRNA sequences, but showed additional genetic variations, including different intron insertion patterns in rRNA regions and distinct exon-intron boundaries in organelle genes (i.e., psaB and psbA). These were suggested as intraspecific taxa (Tetradesmus obliquus f. obliquus UTEX 3031 and T. obliquus f. rectilineare UTEX 393) (Cho and Lee 2024). We also verified that this type of genetic variation in 18S rRNA and rbcL sequences (see details in the following section) is present in Desmodesmus taxa.
The coenobial cells of CCAP 258/122 are alternately arranged and typically consist of two to four cells. The outermost cell is often flat, and mucilage-like is absent (Fig. 3G). Short spines are arranged along the cell edges on the cell surface (Fig. 3H). Large round warts are generally present on the cell surface, and some include ‘stiffenings’ (i.e., a pointed end; Fig. 3H). The rosettes are surrounded by a simple wall with a net-like structure, similar to D. denticulatus var. denticuloides CCAP 258/114 (Fig. 3I). CCAP 258/122 also shows morphological features (e.g., coenobial arrangement and short spines) similar to those of D. denticulatus under LM observation, but their cell surface structures (e.g., well-developed net-like structure warts) differ under SEM (Table 2, Fig. 3H). In addition, the ITS-2 region of CCAP 258/122 lacks an intron, similar to CCAP 258/114 and FBCC-A0719, but unlike D. denticulatus (Fig. 4). However, CCAP 258/122 shows three nucleotide differences in the ITS-1 region and single nucleotide differences in both the 18S rRNA and ITS-2 sequences compared to those of D. denticulatus var. denticuloides CCAP 258/114 (Fig. 4). The phylogenetic position of CCAP 258/122 is also not clearly distinct from D. denticulatus An 1996-12 and its intraspecific taxa (CCAP 258/114 and FBCC-A0719; see details in the following section). Based on these results, we suggest that CCAP 258/122 is D. denticulatus var. distinctus as a novel variety of D. denticulatus.
D. denticulatus var. denticuloides CCAP 258/114 and D. denticulatus var. distinctus CCAP 258/122 were previously misidentified as D. denticulatus, likely due to type strains for D. denticulatus are currently unavailable. However, the original description (Lagerheim 1883) and ultrastructural images of D. denticulatus (An 1996-12 strain) (An et al. 1999) allow for the recognition of key morphological characteristics of the species, such as alternately arranged cells with short spines under LM and a well-developed net-like structure with armed warts on the cell surface under SEM. However, the alternately arranged coenobial cells with short spines in Desmodesmus species are commonly observed under LM, which leads to frequent misidentification as D. denticulatus. Moreover, authentic molecular data for D. denticulatus are limited to ITS-2 sequences (An 1996-12 strain) (An et al. 1999), but the presence of a 19 bp intron in this region provides a distinctive feature. Therefore, SEM observation and molecular evidence are essential for the accurate classification of Desmodesmus species, including alternately arranged coenobial cells and short spines. Based on these features, we suggested CCAP 258/114, CCAP 258/122, and FBCC-A0719 as novel intraspecific taxa of D. denticulatus.
The cells of FBCC-A0724 are oval and elongated, with slightly acute apices, and are arranged linearly in a four-celled coenobium (Fig. 5A). The ribs on the cell surface of FBCC-A0724 are visible under LM (Fig. 5B). The ribs consist of linearly arranged tubes (Hegewald 1997) along the cell edges, forming teeth-like structures that extended to each cell apex as short spine-like formations (Fig. 5C & D). Simple rounded warts are present on the cell surface (Fig. 5E). Rosette-like (unclear rosette) structures surrounded by tubes are observed along the ribs (Fig. 5F). The spine-like structures of FBCC-A0724 are challenging to recognize under LM, thus its oval-shaped cells with short spines can frequently be mistaken for the morphological characteristics of Scenedesmus species under LM. Desmodesmus was previously considered a subgenus of the genus Scenedesmus based on the morphological features (Trainor et al. 1976, Hegewald and Schnepf 1991). Therefore, original descriptions of Scenedesmus species may still include taxa that are now classified as Desmodesmus. We propose that the morphological features of FBCC-A0724 observed under LM correspond to the original description of Scenedesmus praetervisus Chodat (Wołoszyńska 1912, Chodat 1926, Hegewald and Silva 1988). However, the morphological features and phylogenetic position (details in the following section) of FBCC-A0724 indicate that this species belongs to the genus Desmodesmus (Hegewald and Silva 1988). Therefore, we suggest transferring Scenedesmus praetervisus to the genus Desmodesmus based on our results. FBCC-A0725 shares identical morphological and molecular features (e.g., 18S rRNA, ITS, rbcL, and tufA sequences) with D. praetervisus FBCC-A0724. In addition, the collection sites and dates of FBCC-A0724 (Hwaseong-si, Gyeonggi-do, Republic of Korea; Mar 24, 2017) and FBCC-A0725 (Jangheung-gun, Jeollanam-do, Republic of Korea; Aug 7, 2017) are different. Therefore, these strains are regarded as different populations, and their intron insertions are conserved at the population level.
Phylogenetic analysis using ITS-2, 18S rRNA, rbcL, and tufA
We constructed phylogenetic trees using several representative marker genes (ITS-2, 18S rRNA, rbcL, and tufA) from our Desmodesmus taxa. In the phylogenetic tree based on ITS-2 sequences of Desmodesmus species (Supplementary Fig. S2), D. communis var. columnaris FBCC-A0726 clustered with D. communis-like species, including D. communis Hegewald 1977-170 and AICB 993, but their phylogenetic support was weakly resolved. D. similis FBCC-A0691 and D. complicatus FBCC-A1393 clustered with D. communis-like species, but distantly from the authentic D. communis Hegewald 1977-170. However, the ITS-2 sequences are highly conserved (or identical) among morphological distinct Desmodesmus taxa (e.g., AICB 954, AICB 989, and AICB 1007) (Bica et al. 2012), particularly within D. communis-like species, as noted by Hegewald and Braband (2017). We confirmed this observation that ITS-2 sequences among genetically similar Desmodesmus taxa could not provide sufficient information for distinguishing and identifying them. As a result, we suggest that ITS-2 sequences of Desmodesmus taxa are more suitable for comparing nucleotide differences between taxa than for studying their phylogenetic relationships.
D. denticulatus An 1996-12 and its varieties (CCAP 258/114, FBCC-A0718, FBCC-A0719, and CCAP 258/122) show a monophyletic clade in the ITS-2 tree (86%) (Supplementary Fig. S2). Although they are not phylogenetically distinct in the ITS-2 tree, they show different ultrastructures (e.g., warts and rosettes) and molecular features (e.g., an intron insertion and sequence identity) (Figs 3 & 4). This clade also includes strains (SAG 19.81, dSgKDesOc1, and dSgDesBig12-2) labeled as D. denticulatus, whose morphologies have not been reported (i.e., unverified taxa). These strains lack the intron of ITS-2 sequences found in D. denticulatus An 1996-12 and are therefore not identified as that species (genetic distances: 0.0283–0.0331) (Supplementary Fig. S3). D. serratus ACKU_Y_81 is also included in this clade, although its morphological characteristics remain undescribed. In general, D. serratus shows distinct ultrastructural features, particularly in warts (e.g., large warts) (Fawley et al. 2011), compared to D. denticulatus (An et al. 1999) and its varieties (this study). Among other unidentified Desmodesmus strains in this clade, the morphological features of the Tow 8/18 T-23W and Tow 8/18 P-25W strains have been reported under LM, but their coenobial cells are not alternately arranged (Johnson et al. 2007). Based on these results, we confirmed that only An 1996-12 currently represents D. denticulatus, while the others labeled as D. denticulatus are unverified taxa.
D. praetervisus strains (FBCC-A0724 and FBCC-A0725) clustered with unidentified Desmodesmus strains (NDem 6/3 P-3d and Tow 6/16 T-10w) in the ITS-2 tree (Supplementary Fig. S2) and D. lamellatus and they share identical ITS-2 sequences. However, the unidentified Desmodesmus strains have previously been suggested as D. hystrix-like species based on sequence similarity (Johnson et al. 2007), but D. hystrix contains numerous spine-like structures on the cell surface (Lagerheim 1883). In addition, D. lamellatus shows large fold-like structures along cell ridges under SEM (Demura 2024). Therefore, we conclude that ITS-2 sequences are insufficient to resolve taxonomic discrimination among D. praetervisus and these morphologically divergent taxa.
We constructed a ML tree using the 18S rRNA sequences of Desmodesmus species (Fig. 6). D. communis var. columnaris FBCC-A0726 clustered with D. communis-like species, including D. communis AICB 993, but its phylogenetic position was still weakly resolved, similar to that observed in the ITS-2 phylogenetic tree (Supplementary Fig. S2). Nevertheless, the 18S rRNA sequence of D. communis AICB 993, which has been validated by morphological features (Bica et al. 2012), is different from that of FBCC-A0726 (1 bp; genetic distance = 0.0006). In addition, the 18S rRNA region of FBCC-A0726 contains 392 bp of intron sequences (Fig. 2). The notable genetic variation between D. communis (Hegewald 1977-170 and AICB 993) and FBCC-A0726 supports its designation as a novel intraspecific taxon of D. communis, namely D. communis var. columnaris FBCC-A0726. In contrast, D. similis FBCC-A0691, and D. complicatus FBCC-A1393 belong to distinct phylogenetic clades from D. communis AICB 993 (Fig. 6). In addition, they are morphologically distinct from D. communis (Bica et al. 2012, Hegewald and Braband 2017) and show genetic variation in 18S rRNA sequences (9 bp in FBCC-A0691, genetic distance = 0.0051; 6 bp in FBCC-A1393, genetic distance = 0.0034) (Figs 1 & 2). These results support the conclusion that these taxa are clearly distinct from D. communis and should be regarded as novel species within the genus Desmodesmus. However, the D. communis-like species (i.e., labeled as D. communis) that clustered with D. similis FBCC-A0691, and D. complicatus FBCC-A1393 are phylogenetically distinct from D. communis AICB 993, thus these D. communis-like strains may require reclassification. Although D. communis-like species show nucleotide variations in the 18S rRNA regions and distinct phylogenetic positions from D. communis (Hegewald 1977-170 and AICB 993), most of them were identified as D. communis (Fig. 6). In addition, several 18S rRNA sequences of D. communis-like species in the public database were reported from morphologically unverified strains. We confirmed that phylogenetic relationships based on morphologically unverified or misidentified taxa can obscure species classification and phylogenetic interpretation (Fig. 6). Although several morphologically unverified D. communis-like strains remain (Fig. 6), we have successfully reestablished the morphological and molecular characterization of D. communis and related taxa, which include available data.
The phylogenetic tree constructed using 18S rRNA sequences shows a monophyletic clade of D. denticulatus varieties, which also includes SAG 19.81 and CCAP 258/139, both labeled as D. denticulatus (Fig. 6). However, the ITS-2 region of SAG 19.81 lacks the intron (MK975487) (Supplementary Fig. S3), which indicates that it is not D. denticulatus. In addition, the morphological features of SAG 19.81 under LM (as available from the SAG culture collection; https://sagdb.uni-goettingen.de/detailedList.php?str_number=19.81) do not show alternately arranged coenobial cells, which are a typical characteristic of D. denticulatus, suggesting that SAG 19.81 may have been misidentified. Nevertheless, SAG 19.81 clustered closely with D. denticulatus An 1996-12 in the ITS-2 tree (Supplementary Fig. S2). Therefore, we confirm that CCAP 258/114, CCAP 258/122, FBCC-A0718, and FBCC-A0719 remain phylogenetically indistinguishable from D. denticulatus (or a closely related taxon, SAG 19.81) even though these four strains and D. denticulatus show clearly distinguishable morphological and molecular characteristics (Figs 3 & 4). We postulate that these notable differences, despite the close phylogenetic relationship, indicate that these strains (CCAP 258/114, CCAP 258/122, FBCC-A0718, and FBCC-A0719) should be regarded as intraspecific taxa of D. denticulatus. The morphological features of CCAP 258/139, based on its LM images (as available from the CCAP culture collection; https://www.ccap.ac.uk/catalogue/strain-258-139), may not correspond to those of D. denticulatus, although its ultrastructural characteristics and ITS-2 region require validation. Moreover, further verification of this strain is necessary due to variability in the data (five nucleotide differences in the comparable 18S rRNA region) compared to another dataset of SAG 19.81 (KF673382) (Supplementary Fig. S3). NIES-4265 shared identical 18S rRNA sequences with D. denticulatus var. denticuloides CCAP 258/114, but only NIES-4265 contains a 407 bp intron insertion, which is located in a different region compared to D. denticulatus var. simplex FBCC-A0719 (Supplementary Fig. S3). Therefore, NIES-4265 is also regarded as D. denticulatus-related taxon, but validation of its ultrastructural features is also required. Two KMMCC strains (KMMCC 1258 and KMMCC 1297) are misidentified as Scenedesmus, thus further validation of their morphological features is also necessary.
The 18S rRNA region of D. praetervisus (FBCC-A0724 and FBCC-A0725) includes two intron regions (467 bp and 388 bp). The phylogenetic branch of D. praetervisus could not be compared to that of other Desmodesmus species due to unclear topology and low bootstrap support (Fig. 6). Nevertheless, we verified that this species is classified within the genus Desmodesmus based on its monophyletic relationship with other Desmodesmus species (Fig. 6). D. pannonicus GM4n shows a monophyletic relationship with D. praetervisus FBCC strains (86% in Fig. 6), but the 18S rRNA region of D. pannonicus GM4n lacks introns and differs from D. praetervisus by one nucleotide, appearing as a gap. In addition, D. pannonicus GM4n has long spines observed under LM (Hoshina 2014) and teeth-like structures observed under SEM are not prominent in the internal cells of this species (Staehelin and Pickett-Heaps 1975). These morphological features of D. pannonicus are distinct from D. praetervisus (Fig. 5).
In the phylogenetic trees based on rbcL and tufA sequences (Supplementary Figs S4 & S5), our target strains show monophyletic relationships with previously misidentified or mislabeled taxa, frequently without consideration of their ultrastructural features, even among strains from the same FBCC repository. For example, the rbcL sequence of FBCC-A0726 (OR566816) was previously deposited as D. communis in the NCBI database (Supplementary Fig. S4), but we have newly identified this strain as D. communis var. columnaris, a novel intraspecific taxon described in this study (Figs 1 & 2). Morphologically unverified FBCC strains, FBCC-A0406 and FBCC-A0053, were labeled as D. communis, but these strains are significantly different and phylogenetically distinct from D. communis AICB 993 (Fig. 6). In contrast, these strains are phylogenetically clustered with D. similis FBCC-A0691 in the phylogenetic analyses using 18S rRNA, rbcL, and tufA sequences. Therefore, FBCC-A0406 and FBCC-A0053 are considered misidentified strains. Similarly, tufA sequences of FBCC-A0718 (OR566875.1) and FBCC-A0719 (OR566876.1) were previously misidentified as D. denticulatus (Supplementary Fig. S5), but we now suggest them as D. denticulatus var. simplex (Figs 3 & 4). Based on the rbcL and tufA sequences of FBCC-A0724 and FBCC-A0725, these strains were previously regarded as unidentified Desmodesmus taxa, but we newly suggest them as D. praetervisus (Fig. 5). We argue that additional marker gene data from a broader range of Desmodesmus taxa should be required for advancing molecular taxonomic analysis in this genus. Nevertheless, in the rbcL tree, we confirm that D. communis var. columnaris FBCC-A0726, D. similis FBCC-A0691, and D. complicatus FBCC-A1393 show distinct phylogenetic clades (Supplementary Fig. S4).
In the phylogenetic trees, our targets in this study and their related taxa show very close phylogenetic relationships, but they show distinct ultrastructural features and notable genetic variations (e.g., different intron insertions). Therefore, we conclude that ultrastructural analysis and molecular characterization are essential for distinguishing and identifying genetically similar Desmodesmus taxa, particularly when phylogenetic analysis alone cannot resolve their taxonomic discrimination.
Analysis of species hypotheses using the ASAP program
To test the species and intraspecific boundaries of Desmodesmus taxa, we additionally used the ASAP program (Puillandre et al. 2021), which allows target taxa to be distinguished at the species level based on genetic distances from single locus sequence alignments and to test their species hypotheses (e.g., species delimitation). Puillandre et al. (2021) suggest that the two best partitions (i.e., ASAP 1st and 2nd) provide the closest to the truth among the 10 (two best and eight alternative) species partitions, although these should also be evaluated with other lines of evidence (e.g., additional genetic markers and morphology).
We analyzed ASAP using four representative genetic markers (i.e., ITS-2, 18S rRNA, rbcL, and tufA) used in our phylogenetic analyses (Supplementary Fig. S6). Identical subset numbers in the species partition potentially indicate the same species group, whereas different subset numbers indicate the distinction of taxa at different species levels. We used the top five partitions to test the species hypotheses (see Methods for details). The two best partitions show a consistent species hypothesis: D. denticulatus var. denticuloides CCAP 258/114 and D. denticulatus var. simplex FBCC-A0719 are clustered as the same species level (i.e., possibly intraspecific taxa), while the remaining Desmodesmus taxa are distinguished at different species levels. The alternative partitions (3rd–5th) also support these results. However, we postulate that the results are insufficient to distinguish all these Desmodesmus taxa at the species level due to their close phylogenetic relationships. Moreover, the species delimitations using D. denticulatus and their related taxa was tested only based on the ITS-2 region (Supplementary Fig. S6). Therefore, these taxa should still be regarded as varieties of D. denticulatus based on their phylogenetic relationships, as described above.
D. communis (Hegewald 1977-170 and AICB 993), D. communis var. columnaris FBCC-A0726, D. complicatus FBCC-A1393, and D. similis FBCC-A0691 are also indicated as distinct species in the two best partitions as well as in the 3rd-ranked alternative partition, except in the analyses based on the 18S rRNA region (Supplementary Fig. S6). The 4th- and 5th-ranked alternative partitions also support the distinction of our target taxa (D. communis var. columnaris FBCC-A0726, D. complicatus FBCC-A1393, and D. similis FBCC-A0691) from D. communis (Hegewald 1977-170 and AICB 993), at least (Supplementary Fig. S6). However, the phylogenetic position of D. communis var. columnaris FBCC-A0726 was weakly resolved (i.e., indistinct) from D. communis (Fig. 6, Supplementary Fig. S2), thus this taxon should still be considered an intraspecific taxon of D. communis. In addition, the ASAP species partitions derived from the 18S rRNA sequences may not adequately represent species delimitations in Desmodesmus because the 18S rRNA region of Desmodesmus taxa generally shows relatively few nucleotide substitutions despite its long alignment length compared to other genetic markers, thus suggesting that their resolution for species delimitation is lower than that of the phylogenetic analysis (Fig. 6).
As we tested species boundaries of Desmodesmus taxa based on ASAP, almost species hypotheses of them are rejected based on their phylogenetic relationships. Therefore, we could conclude that this method is not appropriate to test species hypothesis of genetically closely related taxa. We suggest that ultrastructural features, notable genetic variations, phylogenetic relationships, and their comprehensive understanding are more appropriate for determining species (or intraspecific) boundaries among genetically similar but morphological divergent microalgal taxa (e.g., the genus Desmodesmus).
Descriptions of novel Desmodesmus taxa
Desmodesmus communis var. columnaris Y. J. Jang et al., var. nov
Description
Cells are arranged linearly in a two- to four-celled coenobium, and long spines are present on each terminal cell. The inner and outer cells show rounded and tapered apices, respectively. The coenobial cells frequently show an outermost cell wall-like structure (i.e., mucilage-like) under LM, though this is not always observed. Short tube-like structures, composed of slightly elevated armed warts, are present with a central hole on a net-like structure of the cell surface. The rosettes consist of bundles of column-like structures constructed by five or six closed tubes at the apex (Fig. 1A–C). Cell length ranges from 6.33 to 14.46 μm and cell width ranges from 3.28 to 8.81 μm.
Holotype
Permanent slide (NNIBRCL24127) of strain FBCC-A0726 deposited in the FBCC (http://fbp.nnibr.re.kr/fbcc/) at the NNIBR, Sangju, Gyeongsangbuk-do, Republic of Korea.
Etymology
The varietal epithet “columnaris” is derived from the Latin word “columnaris” (columnar) and is based on the rosette shape formed by column-like structures on the cell surface.
Desmodesmus similis Y. J. Jang et al., sp. nov
Description
Cells are arranged linearly in a two- to four-celled coenobium, and long spines are present on each terminal cell. The inner and outer cells show rounded and tapered apices, respectively. The coenobial cells frequently show an outermost cell wall-like structure (i.e., mucilage-like) under LM, though this is not always observed. Short tube-like structures, composed of slightly elevated armed warts, are present with a central hole on a net-like structure of the cell surface. The rosettes are located on the cell surface and are composed of four to six tubes (Fig. 1D–F). Cell length ranges from 10.48 to 23.81 μm and cell width ranges from 4.21 to 10.89 μm.
Holotype
Permanent slide (NNIBRCL28325) of strain FBCC-A0691 deposited in the FBCC (http://fbp.nnibr.re.kr/fbcc/) at the NNIBR, Sangju, Gyeongsangbuk-do, Republic of Korea.
Etymology
The specific epithet “similis” is derived from the Latin word “similis” (similar) and refers to the resemblance to the original D. communis caused by ultrastructures on the cell surface.
Desmodesmus complicatus Y. J. Jang et al., sp. nov
Description
Cells are arranged linearly in a two- to four-celled coenobium, and long spines are present on each terminal cell. The inner cells have slightly tapered apices, while the outer cells show tapered apices. The coenobial cells show an outermost cell wall-like structure (i.e., mucilage-like) under LM. The warts on the cell surfaces are unapparent. The rib-like structures are formed by slightly folded cell surface structures arranged in a line. The rosettes are located at the margins of the terminal cells and near the rib-like structures of the inner cells, which is composed of four to six tubes (Fig. 1G–I). Cell length ranges from 10.04 to 16.9 μm and cell width ranges from 3.24 to 8.38 μm.
Holotype
Permanent slide (NNIBRCL24128) of strain FBCC-A1393 deposited in the FBCC (http://fbp.nnibr.re.kr/fbcc/) at the NNIBR, Sangju, Gyeongsangbuk-do, Republic of Korea.
Etymology
The specific epithet “complicatus” is derived from the Latin word “complicatus” (folded) and is based on the rib-like structures composed of folded cell surface structures.
Desmodesmus denticulatus var. denticuloides Y. J. Jang et al., var. nov
Description
Cells are arranged alternately (i.e., in a zig-zag pattern) in a two- to four-celled coenobium covered with mucilage-like along with structures that resemble short spines. Two types of warts (large round warts and long elongate tube warts) are frequently observed on the cell surface. The rosettes are surrounded by a simple wall with a net-like structure (Fig. 3A–C). Cell length ranges from 4.0 to 20.6 μm and cell width ranges from 2.0 to 15.7 μm.
Holotype
Permanent slide (NNIBRCL28326) of strain CCAP 258/114 deposited in the FBCC (http://fbp.nnibr.re.kr/fbcc/) at NNIBR, Sangju, Gyeongsangbuk-do, Republic of Korea.
Etymology
The varietal epithet “denticuloides” is derived from a compound of the Latin word “denticulatus” and the suffix “-oides” (-like), reflecting its similarity to, yet distinction from, D. denticulatus based on the ultrastructure of the cell surface and molecular evidence.
Desmodesmus denticulatus var. simplex Y. J. Jang et al., var. nov
Description
Cells are arranged alternately (i.e., in a zig-zag pattern) and typically consist of two to four cells, which have an outermost cell wall-like structure (i.e., mucilage-like) as observed under LM (Fig. 3D). Short spines are arranged along the cell edges and small simple round (granule-like) warts with a net-like structure are present on the cell surface. The short spines consist of multiple tubes that are only visible under SEM. The rosettes are surrounded by a simple wall (Fig. 3D–F). Cell length ranges from 8.2 to 15.4 μm and cell width ranges from 4.2 to 8.7 μm.
Holotype
Permanent slide (NNIBRCL24105) of strain FBCC-A0719 deposited in the FBCC (http://fbp.nnibr.re.kr/fbcc/) at NNIBR, Sangju, Gyeongsangbuk-do, Republic of Korea.
Etymology
The varietal epithet “simplex” is derived from the Latin word “simplex” (simple) based on the simple ultrastructure of the cell surface.
Desmodesmus denticulatus var. distinctus Y. J. Jang et al., var. nov
Description
Cells are arranged alternately (i.e., in a zig-zag pattern) in a two- to four-celled coenobium. The outermost cell is often flat, and mucilage-like is absent. Short spines are arranged along the cell edge. Two types of warts (long round warts and long round warts with stiffenings) are present on the cell surface. The rosettes are surrounded by a simple wall with a net-like structure (Fig. 3G–I). Cell length ranges from 11.4 to 26 μm and cell width ranges from 4.4 to 16.3 μm.
Holotype
Permanent slide (NNIBRCL28327) of strain CCAP 258/122 deposited in the FBCC (http://fbp.nnibr.re.kr/fbcc/) at NNIBR, Sangju, Gyeongsangbuk-do, Republic of Korea.
Desmodesmus praetervisus (Chodat) Y. J. Jang, H. S. Cho, J. H. Kim, C. S. Lee, S. W. Nam, and J. M. Lee, comb. nov
Epitype (designated here)
Permanent slide NNIBRCL24126 prepared from strain FBCC-A0724, deposited in NNIBR (Sangju, Republic of Korea; acronym NNIBR). The original type is non-material (illustration only) and inadequate for analysis of light-microscopic and ultrastructural characters; epitypification enables robust comparison with modern molecular and morphological data.
CONCLUSION
We report two novel species and four novel intraspecific taxa within the genus Desmodesmus based on morphological features and molecular data. In particular, we focused on several important ultrastructures in the genus Desmodesmus, such as cell surfaces (e.g., warts, ribs, and rosettes) and spine-like structures (e.g., extended teeth-like ribs). To accurately identify the morphological characteristics of Desmodesmus species, we recommend observations of the ultrastructures under SEM. In addition, a mucilage removal step is necessary for the clear observation of morphological characteristics under SEM (see Methods).
To classify species, conserved morphological and molecular characteristics are required and should be compared with those of other species. However, the intraspecific boundaries (e.g., variety- and forma-level) in algal taxonomy have traditionally been distinguished largely based on morphological characteristics, guided by the empirical and philosophical opinions of expert taxonomists. Therefore, we think that intraspecific boundaries can now be discussed using DNA sequence data (Supplementary Text S1). We believe that intraspecific relationships can be identified not only by subtle morphological differences (i.e., the traditional criterion) but also by highly conserved 18S rRNA sequences between morphologically divergent taxa. In addition, notable genetic variations, such as differences in intron insertion patterns and exon-intron boundaries, in rRNA regions or organelle genomes could occur between intraspecific taxa (e.g., D. communis var. columnaris FBCC-A0726 and D. denticulatus var. simplex FBCC-A0719 in this study; T. obliquus f. rectilineare UTEX 393 and T. dissociatus var. oviformis SAG 5.95 in Cho and Lee 2024). In contrast, these intraspecific taxa generally show indistinct phylogenetic relationships, which are not sufficient to distinguish them at the species level, despite their notable (unchangeable) morphological and genetic differences, thus we suggest classifying them as intraspecific taxa. Although there are several exceptions to the phylogenetic species concept (e.g., hybridization and introgression) (Leliaert et al. 2014), our results are based on several representative markers, which are generally sufficient to resolve their phylogenetic and taxonomic relationships. We postulate that such hidden diversity is frequently overlooked, especially in microalgal taxonomy. Therefore, we suggest that ultrastructural observation, molecular characterization, phylogenetic analyses using multiple marker genes, and their integrated interpretation are essential for studying intraspecific relationships in microalgae. However, our criteria for identifying intraspecific taxa may be applicable only to the family Scenedesmaceae, which often includes genetically very closely related taxa that show distinct morphological features.
In this study, we suggest intraspecific taxa of D. communis and D. denticulatus, which show divergent morphological features under SEM observation (Figs 1 & 3) and different intron insertions in the ITS-2 and 18S rRNA regions (Figs 2 & 4). In particular, we postulate that different intron insertion patterns among genetically similar taxa could reflect genomic diversification and potentially represent an early stage of speciation (e.g., intraspecific divergence). In addition, the insertion of mobile DNA elements such as introns and plasmid genes is inheritable at the species level, as validated by several different algal populations (D. denticulatus var. simplex and D. praetervisus in this study) (Lee et al. 2016). Therefore, to study intraspecific and genetically similar Desmodesmus taxa, future studies should analyze informative molecular data (e.g., 18S-ITS1-5.8S-ITS2 regions and organelle genes) and notable genetic variations (e.g., intron insertions) from morphologically verified Desmodesmus strains (or specimens).
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation 
Supplement
Print 
