Freshwater samples were collected from various areas in Korea in 2018 and 2019 (Table 1). A single cell was isolated using the Pasteur pipette (L = 230 mm; Witeg, Wertheim, Germany) under an inverted microscope KI450 (Korea Lab Tech, Seongnam, Korea). For culture, single cells were plated into 96-well plates prefilled with 200 μL URO medium and then incubated for one or two weeks. With increasing cell density, we gradually transferred it to the 24-well plate and 50-mL cell culture flask (SPL, Pocheon, Korea). During cultivation, cells were under the 12 h : 12 h light and dark cycles, ~65 μmol photons m⁻² s⁻¹, and 16°C temperature.

The strains were harvested with a 0.2 μm pore-sized membrane filter (Millipore, Billerica, MA, USA) and treated with 800 μL cetyltrimethylammonium bromide (CTAB) buffer (100 mM Tris-HCl pH 8.0, 100 mM Na 2 EDTA, 100 mM sodium phosphate pH 8.0, 1.5 M NaCl, 1% CTAB). The sample was stored at −20°C until further experiment, and total genomic DNA was extracted following the CTAB method (Richards et al. 1994).

DNA barcoding loci were amplified with polymerase chain reaction (PCR) with each primer set (Table 2). The 20 μL PCR reaction mixture was composed of 12.8 μL sterile distilled water, 2 μL 10× Ex PCR buffer (Takara Shuzo, Kyoto, Japan), 2 μL dNTP mix (4 mM each), 1 μL of each primer (10 pmol), 0.2 μL Ex Taq polymerase (2.5 U), and 1 μL of template. The PCR performing conditions for nuclear regions were as follows: 94°C for 3 min; 40 cycles of 94°C for 30 s, 61°C for 40 s, and 72°C for 3 min; and 72°C for 10 min. For chloroplast regions, the PCR performing conditions were the same, except for the annealing temperature (61°C→55°C). The PCR product was confirmed with 1.5% agarose gel (Condalab, Madrid, Spain) stained with Midori Green under ultraviolet light. PCR products were purified with a QIAquick PCR Purification kit (Qiagen GmbH, Hilden, Germany).

DNA was sequenced by the Sanger sequencing method with the ABI3730 DNA sequencer (Applied Biosystems, Foster City, CA, USA). The obtained DNA fragments were assembled and edited with Sequencher ver. 5.1 (Gene Codes Cor., Ann Arbor, MI, USA).

Microscopic morphological identification was performed according to Komárková-Legnerová (1969), Hindák (1984), and Nygaard (1977). Morphology of the isolated strain were observed with a LM (Axioskop; Carl Zeiss, Oberkochen, Germany) under 400–1,000× magnification. Further identification was made by comparing DNA sequences in the National Center for Biotechnology Information (NCBI) database.

Basic Local Alignment Search Tool (BLAST) searches were performed to validate the sequences of rRNA molecules (18S, ITS + 5.8S, and 28S) and chloroplast genes (psbB, psbC, rbcL, and tufA) obtained from the present study. We aligned the sequences using MAFFT (Katoh and Standley 2013) and removed ambiguous regions using the Gblocks server (http://phylogeny.lirmm.fr/phylo_cgi/one_task.cgi?task_type=gblocks).

A phylogenetic tree of representative species was constructed using the neighbor-joining (NJ) algorithm (Saitou and Nei 1987) with the maximum composite likelihood model. The NJ topology was evaluated with 1,000 replicates, and the bootstrap proportions lower than 50 were discarded. Additionally, pairwise genetic distances were computed using the Kimura 2-parameter (K2P) model (Kimura 1980) with MEGA ver. 11 (Tamura et al. 2021). The genetic K2P distances were also utilized to define the barcode gaps at the species level, comparing the maximum intraspecific genetic distance with the minimum distance to the nearest neighbor.

NJ analysis was conducted with a selected model in MEGA ver. 11 (Tamura et al. 2021). The sequences of other Monoraphidium strains or species were obtained from the NCBI. The visualized tree was edited in Adobe Illustrator CC (Adobe Systems, San Jose, CA, USA).

In particular, ITS2 sequences were annotated using the Hidden Markov Models-based annotation tool present in the ITS2 database V (Ankenbrand et al. 2015). The tree construction was processed the same way as described above.

The secondary structures of the rRNA ITS2 region were modeled using the Fold Web Server (https://rna.urmc.rochester.edu/RNAstructureWeb/Servers/Fold/Fold.html). It folded the nucleic acids into lowest free energy conformation. The construction of the model considered the presence of the pyrimidine-pyrimidine unpaired section at the ninth position in the second helix (Caisová et al. 2013) and the length and nucleotide composition of the spacers in the core of the model delimiting the helix boundaries (Schultz et al. 2005). The resulting secondary structures were visualized in PseudoViewer ver. 2.5 (Byun and Han 2009).

In addition, ITS2 sequences were aligned with the ClustalW algorithm (Thompson et al. 1994), and the sequence-structure alignment was performed in 4SALE ver. 1.7.1 (Seibel et al. 2006). Species delimitation according to ITS2 secondary structure within Monoraphidium species was defined using compensatory base changes (CBCs) according to Coleman (2000) and Müller et al. (2007). The ITS2 structure was further edited in Adobe Illustrator CC (Adobe Systems) to indicate the CBCs on the secondary structure of ITS2 of the strain.

In this study, we examined morphological key characteristics (e.g., cell length, width, shape, and tip shape) of 16 strains, and identified them as nine distinct Monoraphidium species (Table 1). The DNA sequences of nuclear rRNA (18S, ITS, and 28S) and chloroplast genes (psbB, psbC, rbcL, and tufA) were separately obtained from 16 Monoraphidium strains (Table 3). The average lengths of the seven gene sequences were 2,003 (18S), 700 (ITS), 1,339 (28S), 1,447 (rbcL), 1,277 (psbB), 1,190 (psbC), and 719 (tufA), respectively. Introns were found in three regions, 18S and 28S nuclear RNAs, and rbcL in chloroplast genes. Eight of the 16 individuals contained at least one and up to three introns within the genetic regions. Intron lengths ranged from about 400 bp to about 1.7 kb.

In addition, a total of 10 species, including 9 Monoraphidium species and the outgroup Kirchneriella aperta, were used for phylogenetic analysis to compare the degree of resolution. Each phylogenetic tree was resized to fit a shared distance scale bar, maintaining the original rate of length among branches (Fig. 1A & B). The branch lengths of 18S and 28S were shorter than other trees, and tufA was the longest. The branch lengths of all the trees showed distinct branches for each species, except Monoraphidium minutum and Monoraphidium nanum in psbC.

The p-distances of each sequence were visualized in bar graphs (Fig. 2). Bar heights of sequences were almost matched to the length of branches. Overall, nuclear regions showed lower nucleotide divergence than chloroplast genes; however, the ITS region was as variable as much as chloroplast genes. The least diverse region was 18S, and the most diverse region was tufA. The average of p-distances was 7.78 for seven selected sequences.

The pairwise distance distributions were analyzed with a box plot graph comparing the interspecific and intraspecific distances (Fig. 3). The box plots represent medians and interquartile ranges for intraspecific and interspecific K2P genetic distances of seven DNA regions.

No DNA barcode region was classified as ‘poor’, while two regions, 28S and psbC, were evaluated as ‘intermediate’ grade. Five regions—18S, ITS, psbB, rbcL, and tufA—were assessed as ‘good’ DNA barcodes, showing clear interspecies gaps. In addition, we evaluated the regions using several criteria, such as PCR success percentage on the first try, the presence of introns, and sequence length used in the analysis (Table 4). PCR success percentage was higher in nuclear regions than in chloroplast genes. In particular, one sequence of psbB (M. nanum C-253) was not amplified and failed to obtain the sequence. Another factor, the introns, was detected at both nuclear (18S and 28S) and chloroplast (rbcL) genes. In 18S, Monoraphidium komarkovae Ga022, M. minutum Ga027, M. nanum Ga031, and Monoraphidium subclavatum Ga035 included intron regions, and in 28S, Monoraphidium dybowskii Ga011, M. dybowskii Ga012, M. minutum Ga027, M. nanum Ga031, and M. nanum C-253 included intron regions. In rbcL genes, introns were detected from Monoraphidium contortum Ga008, M. dybowskii Ga012, and M. nanum Ga031. The average length of rbcL without introns was calculated to be about 1,160 bp, but the rbcL length of M. nanum Ga031 was more than 4,000 bp. Because of its long length, partial introns were not fully determined, and the incompletely determined region was represented by multiple Ns representing a blank space and an estimated length. We also calculated the sequence length and selected the three shortest regions (ITS, psbC, and tufA). Only ITS met all the criteria required for DNA barcode evaluation, and the second-best marker was tufA.

We constructed an NJ phylogenetic tree for 35 Monoraphidium strains using the ITS marker with the outgroup K. aperta (Fig. 4). It showed distinct clades for each species and close distances between the same species. Clades of each species were supported with bootstrap values above 50 except M. minutum Ga027 and M. contortum Ga008. On the other hand, most of the bootstrap values supporting the relationships between species were lower than 50. It suggested that ITS regions are suitable for species distinction but not for phylogenetic analysis.

In addition to the ITS phylogenetic tree, the ITS2 secondary structure combined with CBCs was constructed. The CBCs mapped on ITS2 are a tool to distinguish species, proposed by Coleman (2000) and Müller et al. (2007). According to Coleman (2000), even a single CBC in the conserved regions of Helix II and Helix III could discriminate two species. Müller et al. (2007) proposed that any CBC in the entire ITS2 could be enough to distinguish the species. Our study showed the representative ITS2 structure of M. dybowskii GA011 and marked the location of CBCs of each species (Fig. 5A). Secondary structures of each strain were constructed, and detailed nucleotide divergence within branches is shown in Fig. 5B–F.

Monoraphidium was once classified into the family Chlorellaceae (Chlorococcales) and the Ankistrodesmaceae (Chlorellales) because of its coccoid form (Hindák 1977, Komárek and Fott 1983). It is now classified under the family Selenastraceae, with genera differentiated primarily based on morphological characteristics. These include the shape of cells and colonies, the arrangement of autospores inside the mother cells, the development of mucilage and encrustations on the cell wall, and the presence, number, and type of pyrenoids (Komárek and Fott 1983, Marvan et al. 1984). Several genera, such as Selenastrum, Ankistrodesmus, and Chlorolobion, differ from each other by only one of the above characters (Krienitz et al. 2001). However, Monoraphidium has similar morphologies with Kirchneriella, Ankistrodesmus, and Chlorolobion, which have the same features in the arrangement of autospores, pyrenoid structures, and the presence of starch envelope, respectively (Krienitz et al. 2001). Within the genus Monoraphidium, species are categorized according to colony formation, length and width, the degree of curvation, shape of the ends, shape, and existence of chloroplast and pyrenoid, the position of the nucleus, and characteristics of autospores. These are distinguishable under LM: however, some Monoraphidium species have almost identical morphology in size, cell shape, and the structural features of cell walls and pyrenoids (Krienitz et al. 2001).

As mentioned above, Monoraphidium poses significant difficulties in distinguishing among species due to their limited morphological variability. This challenge suggests the importance of DNA barcoding as a molecular tool that uses specific genetic markers for accurate species identification. Recent study has shown that even with standard DNA barcoding markers like 18S, species-level identification can be problematic (Zou et al. 2016). To solve this, our study takes a comprehensive approach by evaluating seven genetic markers—18S, ITS, 28S, psbB, psbC, rbcL, and tufA—all derived from the same isolates. This method ensures data consistency, allowing us to directly compare the performance of each marker and minimize any biases that might arise from using sequences from different strains. With these comparisons, we could identify the most effective marker for distinguishing among closely related Monoraphidium species, contributing to more effective species identification.

The effectiveness of DNA barcoding for species identification relies heavily on the performance of genetic markers, which must be evaluated using standardized criteria to ensure accuracy and reliability. For example, barcode gap analysis suggested as critical for assessing a marker’s ability to distinguish individual species (Badotti et al. 2017). Barcode gap analysis examines whether there is a distinct separation between intraspecific and interspecific genetic distances. This criterion is crucial because it offers a visual and intuitive way to assess whether a marker can consistently separate species (Fig. 3). A clear gap reduces the risk of misidentification, which is particularly important in groups with closely related or cryptic microalgal species. By highlighting cases where overlap occurs, barcode gap analysis ensures researchers can identify when a marker may fail and consider alternatives. In the present study, we applied these criteria to evaluate seven gene markers for the green algal genus Monoraphidium, and succeeded to identify the most effective marker for species delineation.

To overcome the limitation of morphological characters, molecular analyses using 18S rRNA sequences have been investigated. The results verified the taxonomical lineage of Monoraphidium in the family Selenastraceae, however, revealed the unrelativeness between morphological criterion and genetic analyses and polyphyletic lineages (Krienitz et al. 2001, Fawley et al. 2006). Although 18S is one of the most phylogenetically typical and preferred markers, the phylogeny of Monoraphidium does not appear to be well described. In addition, these results show that one marker cannot perform the same function for all taxa.

The nuclear and chloroplast sequences are the most broadly used DNA barcodes for algae (Saunders and McDevit 2012). The 18S rRNA gene has been used widely for DNA barcoding, but it does not have sufficient resolution at the species level and even shows the lowest divergence among seven markers within Monoraphidium (Figs 1 & 2). Alternatively, the nuclear 28S rRNA D2/D3 region is considered one of the primary barcodes for green algae (Kress and Erickson 2007, 2012, Minicante et al. 2014). However, the 28S regions of the Monoraphidium were classified as ‘intermediate’ due to the absence of barcode gaps and suggested to be unsuitable (Fig. 3). More variable markers, such as chloroplast genes or the ITS regions, were suitable for species delineation (Schultz and Wolf 2009).

On the other hand, chloroplast genes psbB and psbC contribute to light harvesting in photosynthesis in living cells and have been evaluated as DNA barcodes in plants and green algae (Cameron and Carmen Molina 2006, Shetty et al. 2021). In Monoraphidium, both genes exhibited challenges, including low PCR success rates, sequence degradation, presence of introns, and lack of clear barcode gaps. In addition, the rbcL is one of the standard DNA barcodes established among plant groups (CBOL Plant Working Group et al. 2009). However, green algae frequently retain introns; thus, introns have been found in rbcL of some green algae (Hanyuda et al. 2000), including Monoraphidium. Such introns can negatively impact the universality of barcode markers, as they hinder the amplification and sequencing of large fragments with a single bidirectional read. Lastly, the plastid elongation factor tufA is the primary barcode for green algae, and it is known to lack introns for the variety of green taxa (Famà et al. 2002, Verbruggen et al. 2009). Monoraphidium also showed no introns of tufA from all tested strains. It also showed a barcoding gap and was classified as ‘good’, but had lower PCR success rates on the first try than nuclear ITS. Our results suggest that the ITS region showed the best performances as DNA barcodes for Monoraphidium and the tufA may be second.

The ITS region is considered one of the most useful markers within the Sphaeropleales. Especially, the ITS2 region, a part of the ITS, was recognized by Coleman and Müller for its ability to distinguish species based on secondary structure and is still used today for discovering new species. This finding further supports the suitability of ITS as a DNA barcode. Beyond Selenastraceae, which includes Monoraphidium (Garcia da Silva et al. 2017, Liu et al. 2024), ITS regions are widely used for species distinction across various families, such as Hydrodictyaceae (McManus and Lewis 2005), Mychonastaceae (Krienitz et al. 2011, Martynenko et al. 2022), and Scenedesmaceae (Nguyen et al. 2023). These findings and our results suggest that the species discrimination capabilities of ITS sequences may be applicable across the Sphaeropleales. On the other hand, other DNA barcoding studies in green algae suggested tufA as a suitable marker (Saunders and Kucera 2010, Vieira et al. 2016). Moreover, Vieira et al. (2016) also reported that the ITS region showed a low amplification and sequencing success rate for freshwater Chlorophyceae. These conflicting results imply that more DNA barcode studies for green algae are needed across various levels of taxonomy.

Our current understanding suggests that Monoraphidium is polyphyletic, sharing morphological similarities with genera like Ankistrodesmus within Selenastraceae. Consequently, it is essential to identify barcodes capable of discriminating between these genera. Sequencing the ITS region of Ankistrodesmus will be necessary to confirm its potential for species differentiation in future studies.

In the present study, we assessed several promising gene markers for the species identification of the genus Monoraphidium and selected the ITS region as the most suitable marker. Its efficiency and capability were evaluated through genetic diversity and pairwise distance comparison. Moreover, these were verified by constructing an NJ tree and analyzing the secondary structure of ITS2 with CBCs. The tree showed distinctly separated groups of species, and the structure indicates nucleotide variations among species. In contrast, tufA, another promising marker among green algae, was also evaluated but ranked second for Monoraphidium. With these results, we can conclude that the ITS may be the best marker for Monoraphidium and a promising marker among green algae. Furthermore, we suggest the ITS region to be an effective marker for the order Sphaeropleales (Chlorophyta), and this study could be a basis for future DNA barcoding studies for green algae.