Insight into picophytoplankton diversity of the subarctic White Sea—The first recording of Pedinophyceae in environmental DNA

Abstract Operational taxonomic units 94%–95% similar to the known Pedinophyceae were found as a result of high‐through sequencing of 18S rDNA V4 amplicons of environmental DNA from the summer picophytoplankton samples from the White Sea. Partial sequence of a ribosomal operon (the 5,298 bp includes partial 18S and 28S rDNA, complete 5.8S rDNA, ITS1, and ITS2 sequences) and a partial 2,112 bp chloroplast 23S rDNA sequence White Sea Pedinophyceae was amplified from metagenomic DNA by specific primers and sequenced. A new phylotype was designated as uncultured Pedinophyceae WS. On Chlorophyta phylogenetic trees the discovered phylotype occupies a basal position in the Marsupiomonadales clade. The synapomorphic base substitutions in rRNA hairpins confirm the relationship of Pedinophyceae WS to Marsupiomonadales and its difference from known genera of the order. The obtained results extend knowledge of picophytoplankton diversity in subarctic waters.

scarce. One reason for the lack of information is probably a pico-size of Marsupiomonadales, and their small size mostly hinders microscopic identification on lower taxonomic levels (Vaulot et al., 2008).
Precise taxonomical identification and the accurate phylogenetic affiliation of pico-sized Pedinophyceae require the use of molecular methods. Three incomplete sequences of 18S rDNA uncultured eukaryotes deposited in the NCBI GenBank were referred to as Marsupiomonadales by BLAST analysis. The first one is a eukaryote that was found in a sample collected at Long Island, New York (FJ221481). The second uncultured eukaryote (KC879111) from the order of Marsupiomonadales has detected in winter ice-covered picoplankton from the alkaline Zab-szék shallow pan in Hungary (Pálffy et al., 2014). The third eukaryote (KC539447) was discovered in the Dapeng Bay, Taiwan in a coastal lagoon (Kuo et al., 2014).
Most molecular phylogeny studies of Pedinophyceae used the sequences of nuclear and plastid ribosomal genes (Marin, 2012;Sym, 2015;Wang et al., 2016). The data of the nuclear ribosomal Thus, at present, few species of the Marsupiomonadales are known.
The White Sea, a marginal subpolar shelf region basin adjoins the Barents Sea to the south of the Kola Peninsula and has features similar to those of the Arctic shelf seas (Berger et al., 2001). The sea is usually covered with ice for 5-6 months, from December to May. In summer, the temperature of the surface layer is similar to that in temperate waters. Phytoplankton of the White Sea is studied in detail during last 30 years (Ilyash, Belevich, Zhitina, Radchenko, & Ratkova, 2018). The species composition of nano-and microphytoplankton has been studied by microscopy; the list of algae contains 449 species.  Medlin, Elwood, Stickel, and Sogin (1988). c Data base "Primers for Eukaryotic Nuclear LSU rRNA" (http://bio.cug.

TA B L E 3 Sequencing primers
edu.cn/rRNAp rimer s/NL_lst.html). c Primer constructed in this work.
Primers of the first round of PCR are 23dir1 and 23rev3 and of the second round are 23dir2 and 23rev1c (Table 2). To sequence the amplified fragment, the internal primer 23rev2c, complementary to the conserved region of the gene, was also used ( Table 3).
The length of the partial sequence of the chloroplast 23S rRNA gene was 2,112 bp. PCR products of both nuclear and chloroplastic ribosomal genes in agarose gel electrophoresis moved as single bands.
PCR was performed in a 25 µl reaction mix. Cycling conditions were as follows: initial denaturation at 95°C for 3 min; in each cycle, denaturation was carried out for 20 s at 95°C. The annealing temperature was determined by the primer melting temperature, with the annealing time of 20 s.

| Phylogenetic and rRNA secondary structure analysis
Three data sets were used for phylogenetic trees reconstruction: nuclear-encoded 18S rDNA, 18S + 5.8S + 28S rDNA, and chloroplast 23S rDNA. Alignments of nucleotide sequences were performed by .09 software (Katoh & Standley, 2013) and adjusted by eye. Intron sequences, intergenic spacers, and ambiguously aligned regions were excluded. The final alignment lengths were 1,806 bp for 18S rDNA, 5,103 bp for 18S + 5.8S + 28S rDNA, and 2,266 bp for 23S rDNA. The phylogenetic trees were inferred by the maximum likelihood method using RAxML 8.2.10 program (Stamatakis, 2014) with default options. The bootstrap replicates numbers were set by bootstrapping criterion implemented in RAxML.

| Sequence analysis
We

| Phylogenetic trees reconstruction
The phylogenetic tree of nuclear-encoded 18S rRNA from 32 Chlorophyta taxa with 15 Pedinophyceae phylotypes is shown in Figure 1. The uncultured Pedinophyceae WS is embedded in Marsupiomonadales clade where it occupies a basal position. The mean similarity percentages of Pedinophyceae 18S rRNA are indicated in Table 5.
Available data for the nuclear-encoded rRNA operon are more limited and includes only four Pedinomonas and one Marsupiomonas taxa. In the phylogenetic tree for these sequences, Pedinophyceae WS clustered with M. pelliculata with maximum bootstrap support in a clade sister to the clade of Pedinomonas (data not shown). On the chloroplast 23S rDNA tree (Figure 2), this taxon is a sister to Marsupiomonas species as well.

| 18S rRNA secondary structure analysis
In the secondary structure of Pedinophyceae 18S rRNA, there is a number of compensatory base changes (CBC) differentiating the
Thus, hairpin H8 (hairpin numbering is given according to the model of the secondary structure of the red algae Palmaria palmata 18S rRNA, Wuyts et al., 2000) in Marsupiomonadales, including the Pedinophyceae WS, has two CBC which are different from those in Pedinomonadales-CG > UA and GC > AU (Figure 3a). A number of CBC in V4 variable region hairpins of the 18S rRNA confirms the affinity of Pedinophyceae WS to Marsupiomonadales.
So, the apical part of the hairpin E23-1,2 from Pedinomonadales has no paired G-C found in Marsupiomonas, P. noctilucae, and Pedinophyceae WS (Figure 3b). In the hairpin E23-1, the fourth nucleotide pair in the Pedinomonadales is U-A, whereas in the known of Marsupiomonadales and Pedinophyceae WS it is C-G.

| 23S rRNA secondary structure analysis
The secondary structure of the chloroplast 23S rRNA gene of the Pedinophyceae WS also has features confirming its affinity to Marsupiomonadales. As an example, we use a conservative hairpin, which is formed by nucleotides 1,295-1,308 and 1,621-1,645 (Gutell & Fox, 1988)   TA B L E 5 The mean similarity (in %) of Pedinophyceae 18S rDNA of phytoplankton, including picoforms, due to penetration of algae from temperate waters into the Arctic and disappearance of Arctic endemics (Lovejoy et al., 2007). Besides, the greater involvement of picoforms in primary production and more significant contribution of the smallest photoautotrophs in total phytoplankton abundance are predicted (Kilias, Wolf, Nöthig, Peeken, & Metfies, 2013;Li, McLaughlin, Lovejoy, & Carmack, 2009).

| Affiliation Pedinophyceae WS to order of Marsupiomonadales
NGS results of the V4 region of 18S rRNA show that only one taxon of Pedinophyceae presents in two studied plankton samples of the White Sea. Therefore, we assume that the sequences of nuclear and plastid ribosomal genes belong to the same taxon of the uncultured Pedinophyceae.
Data on the secondary structure of ribosomal genes of Pedinophyceae WS allowed clarifying some previous conclusions made regarding synapomorphies of the orders of Pedinomonadales (Marin, 2012). Thus, according to this author, the second pair in the apical part of hairpin H46 of the V8 variable region is G-C in all Pedinomonadales; however, the G-C pair is also observed in this po-

| Ecology of the new Pedinophyceae
In the White Sea the eDNA was studied in different areas (Kandalaksha and Onega bays) and various biotopes-ice, under-ice water, and summer plankton-total 17 samples (Belevich et al., 2015(Belevich et al., , 2017a(Belevich et al., , 2017b. Pedinophyceae WS was found only in two summer plankton samples and was not found in the samples of ice and under-ice water studied by metagenomic analysis. The water temperature in summer varied from 11 to 15°C. Other representatives of Marsupiomonadales were noted both in the under-ice plankton temperate alkaline shallow pan at mean water temperature 0.5°C (Pálffy et al., 2014), and in the subtropical lagoon at mean water temperature 27°C (Kuo et al., 2014). In the White Sea in summer F I G U R E 3 The secondary structure of 18S rRNA hairpins from Pedinophyceae. (a) H8 hairpin, (b) E23-1,2 hairpins (c) E23-4,7 hairpins salinity varies significantly (Berger et al., 2001). In Onega Bay where Pedinophyceae WS was registered salinity varies from 8 to 27.6 psu (Belevich et al., 2016). The habitat of the Pedinophyceae WS confirms the conclusion that the order of Marsupiomonadales includes marine and brackish species (Marin, 2012).
Pedinophyceae WS was revealed in weakly stratified waters due to tidal mixing with a photic layer length of 6-26 m. There was no nutrient limitation, and the chlorophyll "a" concentration varied from 0.3 to 2 mg/m 3 (Belevich et al., 2017b(Belevich et al., , 2016. The contribution of Pedinophyceae WS to the total number of NGS reads of photosynthetic picoeukaryotes did not exceed 1.35%. This indicates a low abundance of Pedinophyceae WS since the total amount of photosynthetic picoeukaryotes was low (0-36.9 × 10 4 cells/L), and the molecular and microscopic signals are generally correlated (Giner et al., 2016). Similarly, the frequency of occurrence of Pedinophyceae WS is also low, since it was revealed in only two of seven metagenomic summer samples. Thus, in the summer phytoplankton of the subarctic White Sea, the Pedinophyceae WS is a rare taxon.
The Pedinophyceae WS is not the only picophytoplanktonic phylotype that was first discovered not only in the White Sea but also in subarctic waters. Our early studies revealed previously unknown phylotypes Micromonas, Mantoniella, and Bolidophyceae in the environmental DNA of the White Sea plankton (Belevich et al., , 2017a. The identification of a new Pedinophyceae phylotype broadens the current understanding of the picoforms biodiversity in subarctic waters and the biogeography of this poorly studied group of photosynthetic plankton picoforms. Considering the ongoing changes in the White Sea by global warming and their implications (Pozdnyakov et al., 2007), we can expect a change in the structure of phytoplankton and, in particular, an increase of the role of rare taxa.

ACK N OWLED G M ENTS
This work was performed in the framework of MSU project "Noah's Ark" and supported by the Russian Foundation for Basic Research grants # 16-05-00502 and 19-05-00026. DNA sequencing was provided by Center of Collective Use "Genome" at Engelhard Institute of Molecular Biology, Russian Academy of Science (http://www. eimb.ru/ru/ckp/ccu_genome_c.php).

CO N FLI C T O F I NTE R E S T S
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N S
IAM and TAB were involved in project administration. TAB and LVI were involved in sample collection. IAM was involved in performing the experiments. IAM, TAB, and AVT were involved in data analysis.
TAB was involved in original draft preparation. IAM, TAB, LVI, and AVT were involved in review and editing.

E TH I C S S TATEM ENT
None required.

DATA ACCE SS I B I LIT Y
The DNA sequences have been deposited in GenBank under accession numbers MK030604, MK030605, and MK550895.

R E FE R E N C E S
TA B L E 6 Compensatory base changes in rRNAs hairpins of Marsupiomonadales and Pedinomonadales taxa (Figures 3 and 4)

Hairpin numbers Marsupiomonadales Pedinomonadales
H8 of 18S rRNA G-C A-U