A Widely Distributed Thraustochytrid Parasite of Diatoms Isolated from the Arctic Represents a gen. and sp. nov.

A unicellular, heterotrophic, eukaryotic parasite was isolated from nearshore Arctic marine sediment in association with the diatom Pleurosigma sp. The parasite possessed ectoplasmic threads that could penetrate diatom frustules. Healthy and reproducing Pleurosigma cultures would begin to collapse within a week following the introduction of this parasite. The parasite (2–10 μm diameter) could reproduce epibiotically with biflagellate zoospores, as well as binary division inside and outside the diatom host. While the parasite grew, diatom intracellular content disappeared. Evaluation of electron micrographs from co‐cultures revealed the presence of hollow tubular processes and amorphic cells that could transcend the diatom frustule, generally at the girdle band, as well as typical thraustochytrid ultrastructure, such as the presence of bothrosomes. After nucleotide extraction, amplification, and cloning, database queries of DNA revealed closest molecular affinity to environmental thraustochytrid clone sequences. Testing of phylogenetic hypotheses consistently grouped this unknown parasite within the Thraustochytriidae on a distinct branch within the environmental sequence clade Lab19. Reclassification of Arctic high‐throughput sequencing data, with appended reference datasets that included this diatom parasite, indicated that the majority of thraustochytrid sequences, previously binned as unclassifiable stramenopiles, are allied to this new isolate. Based on the combined information acquired from electron microscopy, life history, and phylogenetic testing, this unknown isolate is described as a novel species and genus.

THE Labyrinthulea are heterotrophic eukaryotic stramenopiles historically considered morphologically distinct by the presence of an ectoplasmic net (Tsui et al. 2009) that is used to interface nutrient acquisition and motility (Bennett et al. 2017).The Labyrinthulea are generally divided into three morphologically distinct groups, based on the utility and function of the ectoplasmic net: the labyrinthulids, thraustochytrids, and the aplanochytrids (Leander et al. 2004). The most recent molecular phylogenies suggest that Labyrinthulea is comprised of at least four orders: Amphitremida, Labyrinthulida, Oblongichytrida, and Thraustochytrida (Pan et al. 2017), with the possibility of several additional orders that remain phylogenetically unresolved (Bennett et al. 2017). Within Thraustochytrida, there are at least two families: the Thraustochytriidae and Althorniidae (containing one genus, Althornia). Thraustochytriidae is the most diverse taxonomic family within the Thraustochytrida that circumscribes numerous genera: Aurantiochytrium, Botryochytrium, Hondaea, Labyrinthulochytrium, Monorhizochytrium, Parietichytrium, Schizochytrium, Sicyoidochytrium, Thraustochytrium, and Ulkenia (Bahnweg and Sparrow 1974;Dellero et al. 2018;Doi and Honda 2017;Hassett and Gradinger 2018;, with additional genera not yet represented in sequence databases (Pan et al. 2017).
Labyrinthulea have been cultured and detected by molecular methods throughout the global marine realm (Bai et al. 2019;Bochdansky et al. 2017;Pan et al. 2017). Ecologically, Labyrinthulea behave as bacteriovores, saprotrophs, and symbionts in marine ecosystems (Bennett et al. 2017) and are implicated as important degraders of coastal detritus (Raghukumar 2002;Ueda et al. 2015) and marine snow (Bochdansky et al. 2017). Members of Labyrinthulea can associate with photosynthetic organisms as substrate for development (Scholz et al. 2016). However, parasitism of photosynthetic organisms by this group is atypically reported. Consequently, the ecological role of Labyrinthulea as algal parasites is nebulous. Ambiguity surrounding the parasitic nature of Labyrinthulea is heightened by observations of these organisms at the tips of filamentous algae (Raghukumar 2006) and on senescent and moribund diatoms (Raghukumar 1986), suggesting these organisms are opportunists, as opposed to pathogens that have co-evolved to parasitize. Select Labyrinthulea members within the aplanochytrids and Labyrinthula can consume (Popova et al. 2020) and parasitize diatoms with their ectoplasmic nets (Hamamoto and Honda 2019); however, few observations (Gaertner 1979) have ever reported thraustochytrids parasitizing healthy diatom cells. As a result, the relevance of Labyrinthulea algal parasites to marine ecosystems remains unknown.
A coculture of a pennate diatom within the genus Pleurosigma was established with an unknown parasite that was isolated from Arctic marine sediment in Tromsø, Norway. Analysis of this parasite revealed closest molecular affinity to environmental clone sequences generated in New York, USA, and the recently described Norwegian isolate, Labyrinthulochytrium arktikum. Microscopic examination of this isolate revealed a unique life history, primarily the ability to parasitize diatoms by penetrating their frustules with ectoplasmic threads. Phylogenetic analyses using deoxyribonucleic acid (DNA)-encoding regions of the 18S small ribosomal subunit (18S rRNA) locus place this unknown isolate among the diverse thraustochytrids in the family Thraustochytriida. Based on phylogenetic testing, contextualized with reference trees (Pan et al. 2017), ultrastructure, and life history, this unknown isolate is described as a new genus, Phycophthorum (gen. nov.) and species Phycophthorum isakeiti (sp. nov.).

Isolation and culture
A surface sediment sample was collected from the nearshore marine environment outside of Tromsø, Norway (N69.632 E18.906), on January 21, 2019, at the end of the polar night. Sediment was aliquoted into Petri dishes and diluted with unfiltered natural seawater. Diluted sediment was explored for the presence of parasitized Pleurosigma sp. diatoms with an inverted microscope (Leica DM IL, Wetzlar, Germany). Diatoms identified as Pleurosigma sp. that were hosting parasites were harvested from sediment with a pipette, re-suspended in sterile seawater, and incubated in 96-well plates with a Pleurosigma sp. isolated in Helgoland, Germany (Buaya et al. 2019). After stable cocultures were established, initially free of other foreign or contaminating species, isolates were maintained at 12°C with a light:dark cycle of 14 h:10 h in f/2 medium. The temperature and light-dark cycle used in this study was adopted from the original conditions used to cultivate and maintain the diatom in Helgoland, Germany.
Cocultures were transferred to PmTG (peptonized milk, tryptone, and glucose) agar media, as well as f/2 nutrient media, and explored for the presence of diatom-independent growth. In addition, a supplemental growth analysis was conducted using various media (Daniel Powers, unpublished thesis): GPY, Honda, KMV, and MC (Rosa et al. 2011).

Scanning electron microscopy
Samples were fixed with 4% formaldehyde and 2.5% glutaraldehyde in PHEM-buffered f/2. Samples were then rinsed three times with PHEM buffer and subsequently postfixed with 1% OsO 4 in ddH 2 O for 1 h. After postfixing, samples were rinsed three times with PHEM buffer and then dehydrated in a graded series of ethanol (30%, 60%, 90%, and 96% for 15 min and then in absolute ethanol three times for 15 min). Samples were dried using an EM CPD300 (Leica). The samples were then mounted on scanning electron microscopy stubs with quick-drying silver paint and coated with gold/palladium in a Polaron Sputter Coater (Quorum Technologies, Lewes, UK). Imaging was conducted with a Zeiss Sigma (Jena, Germany) with an in-lens detector at 2kV.

Phylogenetic analysis
Basic Local Alignment Search Tool (BLAST) queries were conducted with clone sequences against the National Center for Biotechnology Information (NCBI)'s nucleotide database and subsequently sorted as diatom or thraustochytrid. Sequences allied to the thraustochytrids were parsed, oriented, and aligned. Multiple sequence alignments (MSAs) were created de novo from sequences (minimum 1,200 base pairs length) retrieved from the NCBI nucleotide database. These sequences were aligned in MEGA7 using MUSCLE (À400 gap penalty). Unalignable regions were eliminated from the MSA. The MSA was then end-trimmed. Bayesian posterior probabilities were generated in MrBayes v3.2.7a (Ronquist et al. 2012) using the GTR + I + G model with 10,000,000 generations, with sampling events occurring once every 100 generations and a "burnin" value of 2,500,000 generations, as previously described (FioRito et al. 2016). The consensus tree from this analysis and corresponding alignment were imported into MEGA7 and analyzed with Maximum Likelihood methods with 1,000 pseudoreplications.
The alignment previously used to generate reference trees (Pan et al. 2017) was imported into MEGA7. The assembled DNA clone sequence from the diatom parasite was appended to this MSA and manually aligned (File S1 and Table S1). This MSA was analyzed with RAxML using the rapid hill climbing algorithm and GTRCATI evolutionary model, as previously described (Pan et al. 2017) with the T-Rex web server (Boc et al. 2012). Trees were then imported into FigTree (v1.4.4) for formatting and presentation (Rambaut et al. 2018).
Widely distributed. Symbiotic with the marine diatom Pleurosigma. Zoospores, sporangia, and vegetative cells hyaline. Cells epibiotic and endobiotic. Epibiotic and endobiotic reproduction by binary division. Vegetative cells spheroidal (2-10 lm diameter) and infrequently amorphic. Motile biflagellate zoospores slightly oval (2 lm diameter) and appear subfusiform in lateral view and oblong with scanning electron microscopy. Motile cell settles and penetrates frustule with tubular process. Cell contains bothrosomes, mitochondria with tubular cristae, vacuoles, and lipid inclusions. Cell wall comprised of lamellae that can form a contiguous tubular process.

Microscopy analysis
Light microscopic analysis of environmental sediment samples revealed the presence of numerous spherical hyaline cells in association with Pleurosigma sp. diatoms that had a noticeable loss of chlorophyll content. After successive rounds of sub-culturing, cocultures of Pleurosigma cells from Helgoland and this unknown parasite were established. Repeated co-culturing revealed that viable cultures of chlorophyll-rich diatoms would be decimated with the introduction of this parasite. Three days after the parasite was introduced into diatom cultures, biflagellate zoospores (2 lm in diameter) migrated and clustered around diatoms. Zoospores then interacted with the outer surface of the diatom frustule with ectoplasmic threads (Video S1, Fig. 1A) that frequently exceeded 80 lm in length. After several hours, the zoospore appeared to contact the diatom cell surface (Fig. 1B). After five days, maturing hyaline cells appeared on the outside and inside of the diatom frustule (Fig. 1C). Endobiotic cells were capable of forming larger amorphoric cells (Fig. 1D). Endobiotic P. isakeiti cells possessed ectoplasmic threads that facilitated slow movement and were capable of reproducing by binary division, eventually resulting in numerous cells dispersed throughout the inside of diatom frustules (Fig. 1D, E).Repeated rounds of binary division of P. isakeiti resulted in diatom cells bursting, releasing Phycophthorum cells, and leaving a broken exterior flap of frustule nearly perpendicular to the diatom (Fig. 1F). Epibiotic Phycophthorum cells were observed reproducing by binary division. Further differentiation resulted in the formation of a sporangium with subsequent zoospore (8-20) release (Fig. 1G). Some immature epibiotic cells began forming an amorphic appendage that could transcend the diatom frustule through a narrow break (Fig. 1H). Cellular content of parasitized diatoms first turned brown and then retracted toward the center (Fig. 1D) and in some instances toward the tips of the diatom, ultimately leading to the demise of diatom cells (Fig. 1I). Presumed Phycophthorum cells (~1 lm diameter) were frequently observed rapidly swarming and darting among moribund diatoms. However, the relevance of these cells to the infection process was unsuccessfully determined with light microscopy. Repeated attempts to grow the isolate on various media were unsuccessful.
Scanning electron microscopy revealed biflagellate zoospores 2 lm in diameter. One of these flagella possessed mastigonemes ( Fig. 2A) that were able to interact with the diatom frustule (Fig. 2B). Many ectoplasmic threads exceeded 80 lm (Fig. 2C). Mature epibiotic cells were observed in various stages of binary division (Fig. 2D), some appearing fused with dorsally protruding daughter cells (~1 lm diameter). Epibiotic cells were observed producing ectoplasmic threads that could penetrate the diatom frustule at the girdle band (Fig. 2E). Ectoplasmic threads frequently emanated from the top of thraustochytrid cells (Fig. 2E). Epibiotic cells formed large aggregates that were intermixed with extruded diatom content (Fig. 2F).
Transmission electron microscopy revealed cells comprised of multilayered lamellar cell walls. Cell walls could form a contiguous electron-dense ectoplasmic thread that bridged dividing (Fig. 3A) and recently divided cells (Fig. 3B). This ectoplasmic thread connected cell walls to the interior of other cells (Fig. 3A). Cells contained a single Golgi body in association with a single nucleus (Fig. 3B, C), lipid inclusions (Fig. 3B), and vacuoles (Fig. 3B). Bothrosomes were visible at the base of ectoplasmic threads (Fig. 3D). Mitochondria possessed tubular cristae (Fig. 3E). Zoospores possessed a nucleus adjacent to a mitochondrion, as well as a single vacuole and lipid inclusions (Fig. 3F). Cells were frequently observed, many with vesicles protruding from the cell membrane (Fig. 3G, H). Closer examination of Pleurosigma diatoms revealed endogenous thraustochytrid cells (Fig. 3I), as well as ectoplasmic threads crossing the diatom frustule (Fig. 3J). Many thraustochytrid cells were observed in close association with the outside of diatom frustules, frequently with membrane interfacing cells (Fig. 3K). Ectoplasmic threads were observed inside the diatom (Fig. 3L).

Molecular analysis
After successful transformation, 10 clones for each primer set were Sanger sequenced. Analysis of BLAST results identified that all clone sequences generated with the NS1-NS4 primer set were allied to thraustochytrids. When all NS1-NS4 clones were aligned, oriented, and trimmed, there were three high-quality (i.e. well-defined chromatogram peaks) SNPs observed. Analysis of BLAST results identified that eight clone sequences generated with the NS5-NS8 primer set were allied to diatoms and two clones (one sequenced 5 0 -3 0 orientation and the second sequenced 3 0 -5 0 orientation) to thraustochytrids. When the thraustochytridallied NS5-NS8 sequences were aligned and oriented, there were no SNPs observed. BLAST queries of the isolate's 18S gene revealed closest molecular affinity to a clone sequence (NCBI accession FJ800589) generated in Long Island, New York, USA (98%, 77% query coverage). The most molecularly similar described species allied to this unknown isolate was Labyrinthulochytrium arktikum (also isolated in Tromsø, Norway, Hassett and Gradinger 2018) with 90% identity and 100% query coverage.
To uniformly test evolutionary hypotheses within the Labyrinthulea, phylogenetic trees were inferred with Maximum Likelihood methods generated with the GTR model (FioRito et al. 2016). The final MSA alignment included 1,163 positions. Maximum Likelihood inferences consistently produced tree topologies that grouped P. isakeiti within the Thraustochytriidae, allied phylogenetically to the clone sequence generated in New York, USA, supported with high posterior node probabilities (100% support). Several clades from this de novo analysis formed polytomies (Fig. 4). To  supplement this de novo analysis, the diatom parasite was manually aligned against a prealigned database used to generate reference trees (Pan et al. 2017). In this expanded analysis, the diatom parasite grouped within Thraustochytriidae and formed a long branch adjacent to "Lab19" sequences ( Fig. 5) that contains the New York clone, as well as a sequence from the Baltic Sea (96% support).

High-throughput sequencing
Prior to expanding the reference database with P. isakeiti and Labyrinthulochytrium arktikum, only 4,223 18S rRNA V9 sequences from the Arctic Ocean environmental sequence datasets were classified as Labyrinthulea. However, upon expansion of the database with P. isakeiti and L. arktikum, the total number of sequences classified as Labyrinthulea increased to 344,431, corresponding to a reduction in "unclassifiable stramenopiles" by 474,312 sequences. From all classified Labyrinthulea sequences in this appended database, 339,827 (98.6%) were allied to P. isakeiti. Reprocessing of V3-V4 databases with the expanded reference dataset, primarily from Utqiagvik (Barrow), Alaska, revealed that 9,449 sequences were classified as Labyrinthulea. Of these, 7,840 sequences were classified as P. isakeiti. Sequences classified as P. isakeiti were detected in 20 m sediment trap samples, at depths exceeding 200 m, in sea ice, and in surface seawater.

DISCUSSION
In this study, a thraustochytrid with atypical life history and morphology was isolated from the Arctic marine realm. Thraustochytrids are more abundant in water temperatures > 5°C (Pan et al. 2017), but have been frequently isolated from high-latitude seas (Hassett and Gradinger 2018;Moro et al. 2003;Naganuma et al. 2006). Ecologically, this isolate is currently unique among the thraustochytrids, as it parasitizes healthy diatoms. Diatom parasitism by thraustochytrids is seldom reported.
Consequently, thraustochytrids are almost exclusively considered saprobes of detritus and opportunistic parasites of marine animals (Liu et al. 2009). In this study, a combination of electron microscopy, light microscopy, and repeated co-culturing observations underscore that this isolate is capable of parasitizing viable, nonmoribund diatoms by penetrating the frustule at the girdle band, resulting in the demise of diatom cells. In the environment, thraustochytrid abundances marginally regress with chlorophyll a (correlation coefficient = 0.06-0.32) (Kimura et al. 2001;Raghukumar et al. 2001;Ueda et al. 2015;   2017), generally ascribed to the metabolism of algaederived organic molecules (Damare and Raghukumar 2008). Ultimately, if algal parasitism is a widely distributed phenomenon, the marginal relationship between chlorophyll a and thraustochytrid abundances could be partially explained by previously uncharacterized algal parasitism.
The specific ecological relevance of P. isakeiti and the environmental factors that regulate its population size remain to be elucidated. The detection of P. isakeiti-classified sequences at > 200 m depth, as well as in sediment trap material, indicates diatom-associated vertical flux, as reported elsewhere in the Arctic (Rapp et al. 2018). It is challenging to discern if its detection in the deeper ocean can be accurately ascribed to simply vertical flux or alternatively to sustained reproduction in the aphotic zone. This isolate was cocultured at the end of the polar night when primary production is arrested or marginal (Berge et al. 2015). The absence of new photosynthates to sustain saprotrophic growth under aphotic conditions might induce opportunistic or facultative parasitism. However, diatom parasitism was readily observed in the presence of artificial light, under controlled laboratory conditions. Moreover, attempts to culture this parasite in the presence of known carbon sources on conventional thraustochytrid agar media were unsuccessful. Consequently, this isolate appears to have a nutritional affinity for Pleurosigma sp. diatoms.
Ectoplasmic threads frequently emanated from the top of diatom-associated thraustochytrid cells. This morphological feature (i.e. dorsal as opposed to basal emanation), in addition to long migrating ectoplasmic threads, inhibited the ability to generate a micrograph of a contiguous thraustochytrid structure entering the diatom from an epibiotic thraustochytrid cell. The general mechanism of frustule penetration with an ectoplasmic thread is conserved between related Labyrinthulea algal parasites (Hamamoto and Honda 2019), suggesting that other Labyrinthulea lineages might be diatom parasites. TEM cross sections of P. isakeiti revealed numerous membrane-bound vesicles. The type and utility of these membrane-bound vesicles is unknown. However, it is possible that these vesicles contain degradative enzymes used to facilitate the osmotrophic life style of Labyrinthulea (Bongiorni et al. 2005;Taoka et al. 2009). The small cells observed with various types of microscopy, in addition to mature epibiotic cells appearing fused (Fig. 2D), suggest a possible sexual cycle, as previously reported (Ganuza et al. 2019). If true, these small cells could help explain why Labyrinthulea-allied sequences have been recovered and reported from among the picoplankton (de Vargas et al. 2015).
The phylogenetic position of P. isakeiti is within "Lab19," which contains one sequence from New York and one from the Baltic Sea. Within this larger clade, P. isakeiti forms a well-supported, moderately long branch. Longer, divergent phylogenetic branches are not uncommon for organisms under positive selection arising from pathogen-host arms races (Abbott 2014;Bromham et al. 2013). The moderate length of P. isakeiti's branch length suggests a degree of positive selection or alternatively, divergence due to sexual recombination. Tests of phylogenetic hypotheses using formally described taxa produced tree topologies with well-supported terminal clades, several forming a polytomy. These poytomies were generally resolved by expanding the total number of sequences used in the analysis with reference tree sequences that are more representative of actual Labyrinthulea diversity. The placement of P. isakeiti among several environmental sequence clades underscores that this isolate represents novel recovered diversity. Future characterizations of cultured isolates and molecular phylogenies expanded with additional environmental sequence data will elucidate evolutionary relationships among Phycophthorum, Labyrinthulochytrium, and their nearest branching relatives. Differential colony morphology on agar media and lipid composition is used to distinguish and discern select Labyrinthulea clades (Dellero et al. 2018;FioRito et al. 2016). However, morphological observations that require agar media are difficult to ascertain for putative biotrophs and other fastidious organisms, thereby necessitating supplemental methodologies to assess diversity and robustly interpret evolutionary patterns. As thraustochytrid diversity continues to be described, molecular phylogenies based on nucleotide information will guide inferences of evolution among the Labyrinthulea (Damare and Raghukumar 2006;). If diatom parasitism is conserved among members of Phycophthorum, media-based culturing efforts (Rosa et al. 2011) would likely not capture diversity within this new genus, possibly explaining why members in this group, and potentially other environmental sequence clades, have evaded cultivation, to-date.
Reprocessing of 18S rRNA high-throughput sequencing databases identified that the majority of Labyrinthulea detected in the HTS eukaryotic microbial survey were allied to P. isakeiti. Sequences allied to P. isakeiti were detected throughout the Arctic Ocean in various environments (sediment, sea ice, open ocean), underscoring the potential wide distribution of this new isolate, while supporting broad dispersal suppositions that were informed by close molecular affinity to a New York clone. The approximate doubling of Labyrinthulea-classified HTS reads, while concurrently approximately halving "unclassifiable stramenopiles," with the addition of two reference sequences suggests that thraustochytrids are underrepresented in HTS-based surveys. Consequently, the relevance of this group to larger ecological process, such as biogeochemical cycling, and the regulation of algal bloom dynamics are likely underreported and consequently underappreciated.
Ultimately, this new isolate expands the known ecology and biodiversity of the thraustochytrids. The life history, sequence identity, and phylogenetic placement among numerous undescribed clades suggest this isolate represents a gen. et sp. nov.

ACKNOWLEDGMENTS
This research has been jointly funded by UiT the Arctic university of Norway and the Tromsø Research Foundation under the project "Arctic Seasonal Ice Zone Ecology," project number 01vm/h15. I would like to acknowledge Marti A. Arumi for his technical laboratory support and his efforts to maintain cocultures. I would also like to acknowledge Augusta Aspar and Randi Olsen at UiT for their assistance with electron microscopy.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article.
File S1. Alignment of Pan et al. (2017) used to generate Figure 5, expanded with P. isakeiti, used to create Figure 5. Table S1. Table of NCBI accessions and associated taxa from Pan et al. (2017), expanded with P. isakeiti, used to create Figure 5. Video S1. Zoospores using ectoplasmic threads to interact with surface of diatom frustule.