Rediscovering Zygorhizidium affluens Canter: Molecular Taxonomy, Infectious Cycle, and Cryopreservation of a Chytrid Infecting the Bloom-Forming Diatom Asterionella formosa

Despite their ecological relevance, parasitic aquatic chytrids are understudied, especially due to the challenges associated with their isolation and maintenance in culture. Here we isolated and established a culture of a chytrid parasite infecting the bloom-forming freshwater diatom Asterionella formosa. The chytrid morphology suggests that it corresponds to the Asterionella parasite known as Zygorhizidium affluens. The phylogenetic reconstruction in the present study supports the hypothesis that our Z. affluens isolate belongs to the order Lobulomycetales and clusters within the novel clade SW-I. We also validate a cryopreservation method for stable and cost-effective long-term storage of parasitic chytrids of phytoplankton. The establishment of a monoclonal pathosystem in culture and its successful cryopreservation opens the way to further investigate this ecologically relevant parasitic interaction.

ABSTRACT

Parasitic Chytridiomycota (chytrids) are ecologically significant in various aquatic ecosystems, notably through their roles in controlling bloom-forming phytoplankton populations and in facilitating the transfer of nutrients from inedible algae to higher trophic levels. The diversity and study of these obligate parasites, while critical to understand the interactions between pathogens and their hosts in the environment, have been hindered by challenges inherent to their isolation and stable long-term maintenance under laboratory conditions. Here, we isolated an obligate chytrid parasite (CCAP 4086/1) on the freshwater bloom-forming diatom Asterionella formosa and characterized its infectious cycle under controlled conditions. Phylogenetic analyses based on 18S, 5.8S, and 28S ribosomal DNAs (rDNAs) revealed that this strain belongs to the recently described clade SW-I within the Lobulomycetales. All morphological features observed agree with the description of the known Asterionella parasite Zygorhizidium affluens Canter. We thus provide a phylogenetic placement for this chytrid and present a robust and simple assay that assesses both the infection success and the viability of the host. We also validate a cryopreservation method for stable and cost-effective long-term storage and demonstrate its recovery after thawing. All the above-mentioned tools establish a new gold standard for the isolation and long-term preservation of parasitic aquatic chytrids, thus opening new perspectives to investigate the diversity of these organisms and their physiology in a controlled laboratory environment.

IMPORTANCE Despite their ecological relevance, parasitic aquatic chytrids are understudied, especially due to the challenges associated with their isolation and maintenance in culture. Here we isolated and established a culture of a chytrid parasite infecting the bloom-forming freshwater diatom Asterionella formosa. The chytrid morphology suggests that it corresponds to the Asterionella parasite known as Zygorhizidium affluens. The phylogenetic reconstruction in the present study supports the hypothesis that our Z. affluens isolate belongs to the order Lobulomycetales and clusters within the novel clade SW-I. We also validate a cryopreservation method for stable and cost-effective long-term storage of parasitic chytrids of phytoplankton. The establishment of a monoclonal pathosystem in culture and its successful cryopreservation opens the way to further investigate this ecologically relevant parasitic interaction.

INTRODUCTION

Fungi belonging to the phylum Chytridiomycota (i.e., chytrids) are important parasites of freshwater phytoplankton (1). Their multifaceted contribution to food web dynamics is increasingly recognized (2, 3): chytrid infections have been shown to be one of the main factors controlling the density and genetic structure of their host population (4–6), with a huge impact on the succession of phytoplankton species and the entire trophic food web (7). In particular, chytrid infections of phytoplankton drive the mycoloop, a trophic shortcut that facilitates the transfer of organic carbon and key nutrients from inedible phytoplankton to higher trophic levels (2, 8–10). Large and heavily silicified diatom cells are a good example of inedible phytoplankton exploited by chytrids, the outbreaks of which can inhibit the development of blooms, as observed in Synedra (65.5% prevalence) and Asterionella (51.3% prevalence) from Lake Pavin and Lake Aydat (France) (11, 12). Studies on the interactions between chytrids and their hosts are rapidly moving from field-based observations to integrated “omics” and “metaomics” investigations, with the former requiring chytrid cultures and resource pooling. For this purpose however, establishing and maintaining pure laboratory cultures of obligate parasites of phytoplankton remain a bottleneck. Renewed efforts are currently being made, leading to the successful cultivation of chytrids parasitizing diatoms (13, 14), cyanobacteria (15), and green algae (16–18). Short life cycles, usually complete within a few days, require frequent medium transfer and the supply of fresh host to ensure the viability of cultures. This skill- and labor-intensive subculturing restricts the availability of isolates and makes them potentially subject to discontinued maintenance. To enable in-depth, long-term studies of this group of organisms, there is a need to ease the maintenance burden of cultures, while also guaranteeing their phenotypic and genotypic stability. Cryopreservation is thus an attractive option for long-term storage, reducing both the time employed in maintaining the culture and the risks associated with serial transfer and minimizing genetic drift, as well as the possibility of contamination and accidental loss (19, 20). Whereas several cryopreservation methods have been proposed for Chytridiomycota, further protocol optimization is needed to achieve quantitative recovery of infectivity postcryopreservation (21). It is widely known that cryopreservation protocols need to be adapted to individual species or even strains of the same species due to variable susceptibility to cryoinjury (22–24).

In lakes, the freshwater diatom Asterionella formosa Hassall is one of the principal bloom-forming diatom species that are inedible to zooplankton (13, 25) and is known to be susceptible to chytrid parasitism (9, 26). A. formosa is infected by three well-described chytrid species, Rhizophydium planktonicum Canter emend., Zygorhizidium planktonicum Canter, and Zygorhizidium affluens Canter. The morphological similarities among these three species led to initial misidentification, which was later resolved by extensive morphological observation on sporangium operculation (26–28). Further studies on zoospore ultrastructure confirmed the existence of the three species, while also suggesting that the two Zygorhizidium species should be separated at a higher taxonomical level (29, 30).

Subsequently, molecular investigation resulted in placement of R. planktonicum and Z. planktonicum into the order Chytridiales and “novel clade II” (sensu Jobard et al., 2012 [31]), respectively (32). To the best of our knowledge, the phylogenetic position of Z. affluens remains to be ascertained by molecular methods, despite it being one of the major players in the epidemic outbreaks that control its host population (28).

Molecular ecology techniques have been applied to investigate chytridiomycosis outbreaks and, more generally, to explore fungal ecology in freshwater lakes (31, 33, 34) and in a range of other aquatic ecosystems spanning deep-sea hydrothermal vents (35), the Arctic Ocean and sea ice (36), and coastal marine habitats (37). Chytridiomycota are an important component of the fungal diversity in all aquatic ecosystems surveyed and are often the dominant fungal taxon, especially in the context of phytoplankton blooms (38). Despite their prevalence, the bulk of this environmental diversity remains unannotated both taxonomically and functionally (e.g., referred to as “dark matter fungi” [39]). Phylogenetic reconstructions show that novel chytrid lineages are composed almost entirely of uncultured organisms (40). Therefore, bringing chytrids into culture is needed not only to investigate their biology but also to establish a reference enabling the annotation of metagenomic data. Here, we isolated and molecularly characterized an obligate chytrid parasite on the freshwater bloom-forming diatom A. formosa and described its infectious cycle under controlled conditions. Furthermore, we developed a method for the cryopreservation of the chytrid that will allow us to continue the studies on this organism. Quantitative data on the chytrid life cycle pre- and postcryopreservation were obtained, including infection parameters (prevalence and intensity of infection [41]) and host viability over time. To investigate the putative conservation of the relationships within the host-parasite pairing as well as the infectivity of the chytrid after cryopreservation, we also propose a double-staining method based on a combination of two fluorochromes (calcofluor white [CFW] and carboxyfluorescein succinimidyl ester [CFSE]) coupled with epifluorescence microscopy.

RESULTS

Morphological characterization of chytrid strain CCAP 4086/1.

The progression of the chytrid through its life cycle is described in Fig. 1. Dissemination is ensured by a spherical zoospore (2 to 3.7 μm in diameter) bearing a posterior flagellum, a large lipid globule, and a crescent-shaped nuclear cap (Fig. 1A). The spore swims toward the host cell and encysts at its surface (Fig. 1B). The initial phase (stage I) (Fig. 1C) corresponds to the direct development of the zoospore into a young sporangium (endogenous development), characterized by the appearance of a germination tube which penetrates the diatom wall on the girdle region. It is followed by a maturation phase that comprises the development of the young sporangium and the growth of the germination tube into the host cell (Fig. 1D). This is followed by the differentiation of visible zoospores inside the sporangium, as well as the growth of a mainly unbranched (rarely laterally unibranched) rhizoid into the host cell (stage II) (Fig. 1E). Finally, a fully mature sporangium with numerous zoospores inside is produced (stage III) (Fig. 1F). During the dehiscence phase, the mature sporangium releases zoospores via a lateral (rarely basal or apical) operculum. The first sign of operculation can be observed with calcofluor white staining as a ring-shaped shade on the sporangium (Fig. 1H). The operculum is rarely seen attached to the empty sporangium and most often detaches completely from it. The sporangium keeps its shape after spore release (stage IV) (Fig. 1G). No sign of sexual reproduction or resting-spore formation was observed in the cultured strain CCAP 4086/1.

FIG 1.

Life cycle of Zygorhizidium affluens CCAP 4086/1. (A) Spherical zoospore with posterior flagellum (F), lipid globule (LG), and nuclear cap (NC). (B) Encystment of zoospores to diatom cells. (C) Development of a zoospore into a young sporangium (stage I) and appearance of a germination tube (GT). (D and E) Development of a young sporangium followed by further maturation with visible zoospores inside (stage II) and rhizoid (R) growth into the host cell. (F) Fully mature sporangium with numerous zoospores inside (stage III). (G) Empty sporangium after release of zoospores (stage IV). (H) Sporangia at stages I, II, and III stained with CFW. Note the ring-shaped shade (RS) on the mature (stage III) sporangium. (I) Operculate (OP) empty sporangium (stage IV) stained with CFW. Scale bars, 5 μm.

Molecular characterization of chytrid strain CCAP 4086/1.

In the concatenated maximum-likelihood (ML) tree of 18S, 5.8S, and 28S ribosomal DNA (rDNA) sequences, CCAP 4086/1 is firmly placed in the order Lobulomycetales (Fig. 2); it does not group with any other sequenced Zygorhizidium or Rhizophydium species, which all fall in the “novel clade II” (sensu Jobard et al., 2012 [31]) and in the order Chytridiales, respectively.

FIG 2.

Maximum-likelihood reconstruction (1,000 bootstraps) of chytrid fungal phylogeny based on three concatenated rRNA-encoding gene sequences (18S, 5.8S, and 28S). Symbols near the species name indicate the presence (*) or absence (−) of genes encoding 18S, 5.8S, and 28S rRNAs, respectively, in the alignment. Species names in bold indicate known parasites of the diatom Asterionella formosa.

Together with “uncultured Chytridiomycota Ay2007E7” from Lake Aydat in France (31), CCAP 4086/1 defines a novel clade (100% support). This clade is a sister to a second well-supported (100%) clade that contains uncultured chytrids from the Baltic Sea (3c-D9, VM3-110, and 5-C10) (42). In turn, these two clades are sisters to the recently described chytrid parasite Algomyces stechlinensis within the robust group (100% bootstrap support) named SW-I (Fig. 2) (18).

Analysis of the environmental molecular diversity surrounding CCAP 4086/1 confirms the existence of a well-defined SW-I clade (100% support) within the Lobulomycetales and reveals substantial diversity within it (Fig. 3, shaded clade).

FIG 3.

Maximum-likelihood reconstruction (1,000 bootstraps) of the phylogeny of the order Lobulomycetales based on the 18S rRNA gene sequences. Aligned reference sequences of OTUs from environmental barcode surveys in aquatic ecosystems highlight the hidden diversity of close relatives of Zygorhizidium affluens CCAP 4086/1 (bold) within the novel clade SW-I (gray background).

Quantitative data on the life cycle of the chytrid strain CCAP 4086/1 in culture.

In order to obtain a clear understanding of the chytrid development and the diatom viability upon infection, we developed a double-staining method combining the vital cytosolic stain CFSE with the chytrid stain CFW. CFSE is a lipophilic molecule that easily permeates the cell membrane, and it is intracellularly sequestered after hydrolysis by esterases and covalent conjugation with cytoplasmic amino groups (43). Therefore, no interaction can occur with calcofluor white, which stains extracellular N-acetylglucosamine in the cell wall (44, 45). This allowed us to follow live and dead diatom cells in A. formosa colonies, the number of chytrids per host cell, and the developmental stage of the chytrid (Fig. 4) (see details in Materials and Methods). Under optimal temperature conditions for the chytrid (15°C), the maturation of young sporangia occurred in 24 h, as judged from the maturation of stage I and II at day 1 into mature sporangia (stage III) at day 2. At day 3, 75.6% of the chytrid population was back to stage I, demonstrating that the entire development cycle (transformation of stage I into stage IV and release of new infectious spores) was completed in less than 3 days (Fig. 5A).

FIG 4.

Infected A. formosa stained by CFW and CFSE in bright-field (A) and UV (B) illumination as observed by optical microscopy. D, dead cell; L, living cell; C, chytrid. Scale bars, 10 μm.

FIG 5.

Chytrid infection development in culture (A and B), under control DMSO conditions (C and D), and after cryopreservation (E and F). Chytrid sporangium development (A, C, and E) and prevalence (Pr) and intensity (I) of infection (B, D, and F) were assessed over 6 days (A, B, C, and D) or 28 days (E and F). dx/y indicates x days after the yth event with introduction of fresh A. formosa cells (see Fig. 6 for details). Panels B, D, and F show means of replicates ± standard deviation.

The prevalence of infection rose steadily from 12.5% ± 1% to 26.4% ± 0.8% for the first 3 days, increasing significantly (Kruskal-Wallis test, P < 0.05) from day 4 to day 6 to reach a maximum of 72.7% ± 3.6% at the end of the experiment (Fig. 5B). Accordingly, the number of stage I sporangia increased significantly between day 2 and day 3 (from 37.66 ± 1.5 to 107.33 ± 3.05 sporangia · ml⁻¹, respectively) and then again between day 4 and day 5 (from 133.3 ± 6.1 to 378 ± 45.4 sporangia · ml⁻¹) (not shown). The intensity of infection stayed stable for the first 4 days (1.33 ± 0.07 to 1.34 ± 0.06 at day 1 and day 4, respectively) before increasing significantly (P < 0.05) to reach 2.79 ± 0.26 at day 6 (Fig. 5B).

Quantitative assessment of chytrid infectivity postcryopreservation.

Several methods for the cryopreservation of CCAP 4086/1 were tested (Table 1). The optimal protocol involved dimethyl sulfoxide (DMSO) (10% [vol/vol]) as a cryoprotectant and a two-step cooling approach. However, it was apparent that chytrid zoospores lost their ability to swim very rapidly and died within minutes during incubation in 10% DMSO. Cultures with a majority of stage I sporangia, as well as 1-week-old cultures, with a high intensity of infection and a range of life cycle stages all failed to survive the process. After several attempts to find the optimal developmental stage of the chytrid, it was concluded that a 3-day-old culture with high prevalence but low intensity of infection and with a majority of mature sporangia (stage III) was most suitable to ensure successful cryopreservation. The diatom host did not survive cryopreservation, regardless of the conditions used.

TABLE 1.

Regeneration of viable cultures of CCAP 4086/1 using various cryopreservation procedures and a range of developmental stages of chytrid culture

Stage (culture age, days)	Regeneration obtained with the following conditions and thawing temp (°C):
	Nalgene				Controlled cooler
	10% Glycerol		10% DMSO		10% Glycerol		10% DMSO
	30	40	30	40	30	40	30	40
Zoospores only	No	No	No	No	No	No	No	No
I (1)	No	No	No	No	No	No	No	No
I, II, III, and IV (6)	No	No	No	No	No	No	No	No
III (3)	No	No	No	No	No	No	Yes	Yes

The capability of the chytrid to infect its host after cryopreservation was compared with quantitative data gathered from the culture and a control containing 10% DMSO in order to check for the stability of infectivity postcryopreservation and optimize propagation/recovery (Fig. 5). The use of the vital staining CFSE together with CFW allowed witnessing of the propagation of the chytrid to freshly added diatom cells that had not undergone cryopreservation.

Initially during the first 2 days in the DMSO control, the prevalence of infection on live cells appeared to be stable (Fig. 5D); however, as the prevalence on live cells increased at day 3 to reach 32.9% ± 6.7%, the number of live cells started to decline from 1.78 · 10⁴ to 1.05 · 10⁴ cells/ml at day 3 and day 6, respectively (not shown). This occurred as the development of the chytrid life cycle progressed through a majority of young sporangia (stage I) at day 4 exactly when the intensity started to dramatically rise from 1.47 ± 0.05 at day 4 to 2.28 ± 0.91 at day 6 (Fig. 5D). After this, the young sporangia developed into stage II (Fig. 5C) at day 5, followed by further maturation over the course of a day into mature and then dehiscent sporangia releasing the zoospores and generating a majority of young sporangia again at day 6 (Fig. 5C).

Initially, 6 days after the samples were thawed (Fig. 5F, d6/1 [i.e., at day 6 and after one addition of fresh A. formosa host]), the prevalence of infection was significantly lower (14.8% ± 3.13%; P < 0.05) than in the DMSO control at day 6 (54.6% ± 28.1%) (Fig. 5D), showing that the infectivity of the chytrid was reduced by the cryopreservation process. After addition of a second batch of fresh host (Fig. 5F), the prevalence of infection remained low (9.8% ± 3.2%); however, it rose steadily after the addition of a third batch of diatoms. After two further additions of bait (Fig. 5F, d6/5), the prevalence of infection in the cryopreserved culture finally reached 67.7% ± 10.9%, a level comparable to that in both the noncryopreserved DMSO-containing control (Fig. 5D) and the reference culture (Fig. 5B) at day 6 (P > 0.05; 54.6% ± 28.1% and 72.7% ± 3.6%, respectively). At that point, a high proportion of the chytrid was at young or maturing developmental stages (stages I and II), demonstrating the dynamism of infection. We also verified that the life cycle was completed in less than 3 days.

DISCUSSION

Identification of CCAP 4086/1 as Zygorhizidium affluens and its taxonomic implications.

The establishment of a coculture from a single sporangium of the chytrid CCAP 4086/1 on a monoclonal culture of A. formosa revealed that the morphological features and the development stages of this parasite are extremely similar to those described for Rhizophydium planktonicum (26, 27): spherical zoospores with a posterior flagellum and a single large lipid globule attach (encyst) to the host frustule, where they develop directly (endogenous growth) into a eucarpic monocentric sporangium, and the rhizoidal system is long and thread-like (rarely with a single lateral branch), occupying nearly the entire infected cell. However, careful observation revealed the presence of an operculum detaching after zoospore release. Coupled with the robust and firm nature of the sporangium, which does not collapse after spore release, this observation ruled out R. planktonicum as a plausible affiliation for CCAP 4086/1 (26–28). We also ruled out rarely reported chytrid parasites of A. formosa such as species 4 and species 5 (observed only once in Tarns, Cumbria, United Kingdom [26]), because they are characterized by an obovoid or irregularly shaped sporangium derived from the asymmetrical swelling of the encysted zoospore, coupled with a long, laterally branched rhizoid, in sharp disagreement with the morphology of CCAP 4086/1. The peculiar sporangial morphology of Rhizophydium tetragenum, i.e., originally a tetroid evolving into a Sarcina-like zoosporangium (71), made it easy to dismiss this species as a possible candidate for our organism. Of the remaining described chytrid parasites infecting A. formosa, only Zygorhizidium planktonicum and Zygorhizidium affluens are operculated (26, 28). Z. planktonicum is distinct from both R. planktonicum and Z. affluens on the basis of (i) an obpyriform sporangium that is taller than it is broad, (ii) a short and heavily branched rhizoidal system, and (iii) the presence of an apical papilla where the operculum is formed (26, 46). This operculum tends to remain hinged to the empty sporangium (26), in contrast to our observations, where the operculum had a tendency to be cast off after dehiscence. Similar to the case for CCAP 4086/1, the operculum of Z. affluens tends to detach, leaving behind only a long-lasting spherical sporangium with a broad lateral opening. The development of the operculum starts with a circular thinning usually on one side of the sporangium, observed in transmission electron microscopy (TEM) and identified as an “opercular rim” by Beakes et al. (46). A similar feature was observable in mature CFW-stained sporangia of CCAP 4086/1 (Fig. 1H, ring-shaped shade), consistent with a degradation of the chitinous cell wall around the opening of the forming operculum. We were also able to detect the presence of a “nuclear cap” in zoospores (Fig. 1A), which was already reported as a character specific to Z. affluens as seen by bright-field optical microscopy (28) and TEM observations (29, 30). In summary, our observations match in all respects the description of Zygorhizidium affluens Canter (28) and distinguish the isolate CCAP 4086/1 from Z. planktonicum, R. planktonicum, and all other known species of chytrid parasites infecting A. formosa. On these bases, CCAP 4086/1 is here named Zygorhizidium affluens CCAP 4086/1. Recent multigene phylogenetic analysis of the rDNAs of chytrids morphologically identified as Z. planktonicum and Zygorhizidium melosirae Canter emend. confirmed their close relationship (14) and placed these two species in the so-called “novel clade II” (sensu Jobard et al., 2012 [31]), which is otherwise composed only of environmental sequences. In the same study, a multigene rDNA sequence of R. planktonicum was produced, confirming its affiliation to the Chytridiales (14, 47). Our data show that Z. affluens CCAP 4086/1 belongs to the order Lobulomycetales (48) and is thus distantly related to both R. planktonicum and the Zygorhizidium species already sequenced.

The order Lobulomycetales has been described on the basis of genetic markers and zoospore ultrastructure (48) and so far contains eight characterized species (17, 31, 45–48), to which metabarcoding surveys added a high richness of environmental sequences from various habitats, including corn rhizosphere (C. Hussels, unpublished data), salt marshes (49), abyssal hydrothermal vents (35), alpine snow (50), and the Arctic Ocean/sea ice (36). The brown and red seaweed obligate parasite Algochytrops polysiphoniae (51, 52) and the recently described parasite of volvocacean algae Algomyces stechlinensis (18) are so far the only known parasitic members of the order. In the phylogenetic reconstruction presented here, both Z. affluens CCAP 4086/1 and Algomyces stechlinensis belong to the well-supported novel clade SW-I (18). Within SW-I, Z. affluens CCAP 4086/1 groups together with “uncultured Chytridiomycota Ay2007E7,” which was retrieved from the eutrophic freshwater Lake Aydat (France), close to Lake Pavin where our strain was isolated (31) (Fig. 2). These two organisms are sisters to a robust clade composed of three environmental sequences from the Baltic Sea (42; L. Montonen, V. Kinnunen, and K. T. Steffen, unpublished data), while Algomyces stechlinensis (the basal taxon within SW-I in our phylogenetic reconstruction) infects Eudorina elegans in the oligotrophic Lake Stechlin (Germany). Within SW-I, Zygorhizidium affluens CCAP 4086/1 and Algomyces stechlinensis are the only two species for which the ecological function has been ascertained, both being parasites of microalgae in freshwater habitats. It is worth mentioning that environmental 28S sequences from the Arctic Ocean which have been coupled with observations of chytrid parasitism on diatoms (36) clustered as a sister taxon to the above-described group within clade SW-I in the phylogenetic tree presented in reference 18. This suggests that algal parasitism could be a conserved or widespread ecological strategy within SW-I. Our phylogeny agrees with the work of Beakes et al. (29), who hypothesized that Z. planktonicum and Z. affluens belonged to different genera on the basis of zoospore ultrastructure. In particular, Beakes et al. observed that Z. affluens zoospores lack microtubule roots, a Golgi apparatus, and a rumposome (fenestrated cisterna) (29, 30, 46). The lack of these features together with the presence of an opaque flagellar plug bearing extensions are the principal ultrastructural characteristics used to define the order Lobulomycetales (48). On this basis, Simmons et al. (48) hypothesized the possibility of an inclusion of Z. affluens into the Lobulomycetales, in agreement with our phylogenetic conclusions. TEM analysis of Z. affluens CCAP 4086/1 zoospore ultrastructure is under way. Operational taxonomic units (OTUs) retrieved using the Z. affluens CCAP 4086/1 18S rDNA as a query to screen metabarcoding data sets from different aquatic ecosystems support the evidence of a high diversity within SW-I, which is confirmed as a well-defined subclade in the order Lobulomycetales. Our results do not allow us to speculate on the habitat preferences within SW-I, since we did not investigate ecosystem others than aquatic ones. However, our results suggest that SW-I can potentially be a high-rank taxon within the Lobulomycetales, although this hypothesis will need further molecular and ultrastructural data to be confirmed. Overall, our data add weight to the hypothesis that the genus Zygorhizidium is polyphyletic and therefore will need revision. For this purpose, ultrastructural data for CCAP 4086/1 would be required, together with a better resolution of the position of Z. affluens CCAP 4086/1 within the SW-I subclade and, ideally, the ultrastructural characterization of the closely related species Algomyces stechlinensis in order to identify morphological synapomorphies defining the clade. Finally, molecular investigation of the type species for Zygorhizidium (i.e., Zygorhizidium willei Löwenthal, which is parasitic on the green alga Cylindrocystis brebissonii) is required to determine whether clade SW-I or novel clade II (sensu Jobard et al., 2012 [31]), if either, should retain the name Zygorhizidium.

Cryopreservation.

The cryopreservation method proposed here uses the standard cryoprotectant (DMSO) and cooling rate already proposed for other fungi (53). However, previous studies have highlighted the need of incorporating the physiological state of the organism and the analysis of infectivity posttreatment (21). In the present study, the biological condition of the organism precryopreservation was assessed and the comparison on infectivity pre- and posttreatment investigated to ensure the stability of the culture, allowing the long-term study of the organism.

Samples with a majority of just-encysted sporangia (stage I) did not survive the cryopreservation process, which is probably related to the fact that they are trophically dependent on their host cell (54), which also did not survive cryopreservation. Likewise, we assume that 1-week-old cultures with a high intensity of sporangia in all development stages did not survive because the host was unable to support the further development of the chytrid after thawing.

As previously described (23), motile zoospores did not survive cryopreservation; therefore, we hypothesized and experimentally verified that the best results would be obtained with stage III chytrids that no longer trophically rely on the host, i.e., mature sporangia full of zoospores. Once thawed, the zoospores are released and continue the infection upon addition of fresh host cells. This addition of fresh host is imperative because A. formosa did not survive the cryopreservation process. Repeated inoculation of new host after each life cycle (∼3 days) allowed the propagation of the chytrid.

The decrease in the prevalence and intensity of infection observed immediately after cryopreservation was likely due to increased stress levels resulting from tissue damage due to ice formation and other cryopreservation-related drawbacks (55–57). However, after 6 life cycles (∼18 days) and 5 inoculums of the host, both the prevalence and intensity of infection steadily rose to become comparable to those pretreatment (Fig. 5B to D).

Although cryopreservation protocols must be tailored to the species level with specific cryoprotectants and cooling rates (21, 22, 24, 53), the tools described here are the basis for the appropriate study of infectivity pre- and postcryopreservation. Thus, they illustrate the reliability of the method. We hope that the novel protocols established here will ease the maintenance burden for obligate chytrid parasites and therefore stimulate efforts on the isolation of novel strains, the investigation of their physiology and phenotypic plasticity, and the generalization of our results. We also hope that these protocols may inspire future research on other parasites, for example, obligate biotrophic plant pathogens.

MATERIALS AND METHODS

Sample collection.

Samples were collected in Lake Pavin (45°29′41″N, 2°53′12″E), an oligotrophic deep volcanic mountain lake (maximum depth [Z_max] = 92 m) characterized by small surface (44 ha) and small drainage basin (50 ha) areas. A weekly sampling mission was undertaken from March to April 2013 during the annual diatom bloom, near the center of the lake at the point of maximum depth. Twenty liters was sampled using an 8-liter Van Dorn bottle at the middle of the euphotic layer (estimated from Secchi depth). To eliminate the metazoan zooplankton, collected samples were immediately prefiltered through a 150-μm-pore-size nylon filter. The filtrate was then concentrated on a 25-μm-pore-size nylon filter, collected by washing the filter with 0.2-μm-pore-size-filtered lake water, poured into sterilized transparent recipient flasks, and then transferred immediately to the laboratory for processing.

Strain isolation, purification, and culture conditions.

A. formosa was isolated by micropipetting using a 20-μl glass capillary (Brauband; intraMARK, Germany). Single colonies of A. formosa were picked and transferred into 6-well plates containing fresh sterile diatom medium (DM) (29). The diatom was maintained at 20°C under a 12/12-h light/dark regime (irradiance, ∼64 μmol · m⁻² · s⁻¹). Likewise, colonies of A. formosa infected with one sporangium of the chytrid were isolated following the same method and incubated at 15°C using the same light conditions as described above.

A number of strategies were combined to purify these clonal cultures. Initially, serial dilution was used, by micropipetting single colonies and inoculating them into fresh sterile medium; A. formosa infected with chytrids was then filtered on 50-μm and 20-μm filter units (Celltrics; Partec, Germany); the diatom colonies were retained by the filter and then washed into sterile medium. This was repeated until bacteria were the only other organisms present in the culture. Once established in culture, the chytrid-A. formosa pairing and the uninfected A. formosa strain were maintained by serial transfer every 6 days. Specifically, a 180-ml 1-week-old A. formosa culture was infected with 20 ml of a 1-week-old chytrid culture.

All strains used in this study are freely available from the Culture Collection of Algae and Protozoa (CCAP) under the following accession numbers: A. formosa, CCAP 1005/23; and chytrid, CCAP 4086/1.

Strain characterization and culture synchronization.

In order to characterize the chytrid strain and quantify the infection process under optimal conditions (as described above), a synchronized chytrid culture was studied over 6 days. To obtain a synchronized chytrid culture, a 1-week-old chytrid culture was filtered successively through 25-μm, 10-μm, and 5-μm nylon meshes to obtain a suspension of fungal zoospores, free of host cells. As previously observed, the filtration process does not affect the zoospore swimming activity (13). After 1 week, the chytrid culture was subcultured as described above (10% [vol/vol]) into 1-week-old A. formosa previously grown at 20°C under a 12/12-h light/dark regime (irradiance, ∼64 μmol · m⁻² · s⁻¹). Daily, the host density, the chytrid life cycle, and the prevalence (percentage of infected cells in a host population) and intensity (number of sporangia per infected cell) of infection, two classical parameters used to study this group of organisms (41), were studied. Diatom concentrations were determined with a Sedgwick-Rafter counting chamber (Hausser Scientific, Horsham, PA, USA). To determine the prevalence and intensity, 1 ml of the chytrid culture was stained with the fluorochrome calcofluor white (CFW) (final concentration, 2.5% [vol/vol]) and examined using UV excitation (405 nm) under an inverted Zeiss Axioskop 2 epifluorescence microscope (Carl Zeiss, Germany). Systematically, 100 colonies, representing at least 400 cells, were examined, and both the prevalence and intensity of infection, as well as the chytrid life stage and morphological characteristics, were recorded.

Retrieving CCAP 4086/1 rDNAs.

A transcriptomic database generated from the pathosystem involving the chytrid CCAP 4086/1 and A. formosa CCAP 1005/23 was queried for the presence of the parasite and its host. Briefly, reads for each sample were quality checked with FastQC (58), trimmed using Trimmomatic (59), and quality checked a second time (FastQC), and then each sample was lane-wise assembled via Trinity (60). Using BLAST and NCBI E-utilities, de novo-assembled contigs from a heavily infected sample of A. formosa and CCAP 4086/1 were queried to obtain rDNA belonging to the host diatom and the chytrid parasite using GenBank 18S rDNA sequences of A. formosa (HQ912633) and “uncultured Chytridiomycota Ay2007E7” (JQ689413). Contigs whose similarity to the query sequences was above 95% were subjected to further BLAST analysis against GenBank, and via this procedure contig Trinity_DN14199_c0_g7_i3 (4,962 bp containing 18S rDNA, internal transcribed spacer 1 [ITS1], 5.8S rDNA, ITS2, and 28S rDNA) was identified as belonging to the chytrid parasite (with 99.48% identity to JQ689413 over 1,162 bp) and subsequently chosen as a genetic marker for further phylogenetic analysis.

Phylogenetic analysis.

We assembled a data set of the 18S, 5.8S, and 28S rDNA sequences of chytrids based on the work of Seto et al. (14) in order to encompass all the known molecular diversity of chytrid parasites of A. formosa. Since preliminary findings pointed toward inclusion of CCAP 4086/1 in the recently established order Lobulomycetales (48), long and informative rDNA sequences within this order were included in the tree (18, 32, 61). Finally, “uncultured Chytridiomycota Ay2007E7” was included as the best GenBank match against CCAP 4086/1, and three 18S rDNA sequences from environmental surveys (uncultured fungi 3c-D9, 5-C10, and VM3-110) matching the query sequence with identities of >95% were also added to our data set (42). Sequences were aligned in Geneious 6.1.8 (62) using MAFFT (63), manually checked, and trimmed for the presence of introns. With IQ-TREE 1.5.5 (64), substitution models best fitting the data were assessed for each gene separately via ModelFinder (65), resulting in TIM2+R5 (18S and 28S) and TPM2+I+G4 (5.8S). A concatenated alignment was analyzed with the same software using a partitioned model (66) under the −spp option, i.e., allowing each gene to evolve at its own speed, and a maximum-likelihood tree was inferred using a bootstrap test of phylogeny with 1,000 replicates.

Diversity assessment in environmental barcode data sets.

An in-house script combining EDirect and SRA Toolkit utilities, referred to as MOULINETTE (67), was used to screen ∼19,000 metabarcode data sets from projects deposited in the NCBI Sequence Read Archive (SRA). Those were selected using the keywords “freshwater,” “lake,” “wastewater,” and “aquatic” (see reference 67 for details). Briefly reads are retained when they are at least 97% identical over 80% of the length of the query sequence (in this case CCAP 4086/1), extracted, paired, and filtered (expected error, over 1.0). Paired reads that survived the filtering process were then clustered into OTUs using usearch (v9.1.13) (68). All OTUs were then aligned to reference Lobulomycetales sequences, including the parasite CCAP 4086/1, using MAFFT (63). An ML tree (1,000 bootstraps) was inferred using IQ-TREE 1.5.5 (64) and ModelFinder (65) to assess the best-fitting model of molecular evolution (i.e., HKY+R3).

Cryopreservation.

A range of cryoprotectants and cooling and thawing procedures were tested on cultures at different life stages (Table 1). After this initial screen, the following optimal protocol was used for all subsequent experiments. Cultures were cryopreserved in triplicate, using dimethyl sulfoxide (DMSO) (10% [vol/vol]) as a cryoprotectant and following a two-step cooling approach involving initial controlled-rate cooling followed by plunging into liquid nitrogen. Three-day-old infected cultures were employed since they had both a high prevalence and a low intensity of infection. Aliquots of the chytrid-A. formosa pairing (0.5 ml) were dispensed into cryovials (Greiner Bio-One GmbH, Germany). DMSO (Sigma-Aldrich Ltd., UK) was filter sterilized in sterile DM to a final concentration of 20% (vol/vol) using a 0.20-μm sterile syringe filter (Anachem, UK). An aliquot (0.5 ml) of the 20% (vol/vol) DMSO solution was added to the harvested cells to give a final DMSO concentration of 10% (vol/vol). This was then incubated at room temperature (∼20°C) for 20 min prior to cryopreservation. The cryovials were then transferred to a controlled-rate cooler (Kryo 360 3.3; Planer plc, UK) and cooled at 1°C · min⁻¹ between 20°C and −40°C. After being held for a further 15 min at −40°C, the cryovials were rapidly removed, plunged into liquid N₂, and then transferred to a cryostorage container filled with liquid N₂.

To investigate the potential effect of the DMSO on the infectivity pattern of the chytrid, the same procedure (except for the cryopreservation/cooling of the samples) was followed to establish 3 control replicates. After incubation with DMSO, the control samples were inoculated into 9 ml sterile DM to dilute the DMSO 10-fold, and an inoculum of the diatom host (3-fold more than the density in the initial sample) was added. The control replicates were then incubated under the same light and temperature conditions used for the cryopreserved samples (see “Thawing and recovery” below) and were sampled daily for 6 days by removing 1 ml and inspecting at least 100 A. formosa colonies (Fig. 6). For each colony encountered, the number of A. formosa cells, their viability, the number of chytrids in each cell, and the chytrid life cycle stage were recorded.

FIG 6.

Schematic representation of the cryopreservation procedure (left, gray flasks) and control conditions (right, black flasks). An inoculum of a three-day-old infected culture (A) was mixed with the cryoprotectant DMSO (B) in a cryovial. The chytrid culture was then cryopreserved using a two-step cooling approach involving initial controlled-rate cooling followed by plunging into liquid nitrogen. Immediately after thawing, the samples were inoculated into cell culture flasks containing 3-fold more fresh A. formosa CCAP 1005/23 (C) to allow the development of chytrid CCAP 4086/1. Finally, DM (D) was added to dilute the DMSO 10-fold, avoiding possible toxicity. Viability was assessed by cell counts and subsequent determination of prevalence and intensity of infection.

Thawing and recovery.

After 1 week, the samples were transferred in liquid N₂ from the cryostorage facility to the lab. They were then thawed by direct immersion in a preheated water bath at 30°C and were removed as soon as all visible ice had melted. Immediately after thawing, the samples were aseptically inoculated into 50-ml cell culture flasks containing fresh A. formosa CCAP 1005/23 to allow the development of chytrids, in the proportion of 1 thawed cell to 3 fresh A. formosa cells. The flask was topped up with 9 ml sterile DM to dilute the DMSO 10-fold. The flasks were then transferred at 15°C under reduced light intensity (irradiance of ca. 12 μmol photons · m⁻² · s⁻¹) and a 12/12-h light/dark regime for the first 24 h to reduce potential light-induced stress (19). The samples were then incubated at an irradiance of ca. 64 μmol photons · m⁻² · s⁻¹ for another 6 days to generate sufficient material to undertake postpreservation functional stability assessment.

Postcryopreservation viability assay.

To assess postcryopreservation viability, we developed a double-staining method with carboxyfluorescein diacetate succinimidyl ester (CFSE) used as a vital stain (69) together with the fluorochrome calcofluor white (CFW), allowing us to simultaneously assess the viability of the diatom host and stain the chytrid cell wall (Fig. 4). CFSE in the form of a 10 mM stock solution in DMSO was freshly prepared and added to the samples to give a final concentration of 2 μM, and then the cells were incubated at room temperature for 15 min (70). Following this, CFW was added to the samples at a 2.5% (vol/vol) final concentration following a protocol described previously (4) and incubated for a further 10 min before examination under a Zeiss Axioskop 2 epifluorescence microscope (Carl Zeiss, Germany) fitted with 100-W UV illumination and two filter sets, i.e., a type 09 filter (excitation, 450 to 490 nm; dichroic mirror, 510 nm; emission, long path, 515 nm) and a type 02 filter (excitation, 365 nm; dichroic mirror, 395 nm; emission, long path, 420 nm). Chitin walls stained with CFW were examined using UV excitation (405 nm), and the viability of diatoms stained with CFSE was explored under blue light illumination (488 nm) using UV light excitation. Micrographs were taken with an AxioCam HRc camera (Carl Zeiss, Germany) using the AxioVision software, version 4.7.1 (Carl Zeiss, Germany).

The viability of thawed samples was estimated by systematically inspecting at least 100 A. formosa colonies. For each colony encountered, the number of A. formosa cells, their viability, the number of chytrids on each cell, and the chytrid life cycle stage were recorded to elucidate the prevalence and intensity of infection (41). These parameters were recorded 6 days (∼2 life cycles) after thawing to allow the culture to recover from the stress induced by the cryopreservation method. After this, a new inoculum of the diatom host (3-fold cell/cell) was added to allow the development of the infection, and the same counts were repeated after periods of 3 days (∼1 life cycle) for 3 times, adding a new inoculum of the host (3-fold) each time. Following this, the samples were left for 3 days under normal conditions and a new inoculum of the host (3-fold) was added; the samples were then left to develop a normal infection, and the counts were repeated to be able to compare the infection levels after 2 life cycles (∼6 days) postcryopreservation with the control. The entire procedure is summarized in Fig. 6.

Statistical analysis.

Due to the nonnormal data, differences in the prevalence and intensity of infection among time points and experimental conditions were tested with the nonparametric Kruskal-Wallis test, followed by a Mann-Whitney pairwise comparison with Bonferroni correction. All statistical analyses were conducted using PAST 3.08 (http://palaeo-electronica.org/2001_1/past/issue1_01.htm).

Accession number(s).

The rDNA sequence of CCAP 4086/1 is available in GenBank under accession number MH626496.