Genetic transformation of Spizellomyces punctatus, a resource for studying chytrid biology and evolutionary cell biology

Chytrids are early-diverging fungi that share features with animals that have been lost in most other fungi. They hold promise as a system to study fungal and animal evolution, but we lack genetic tools for hypothesis testing. Here, we generated transgenic lines of the chytrid Spizellomyces punctatus, and used fluorescence microscopy to explore chytrid cell biology and development during its life cycle. We show that the chytrid undergoes multiple rounds of synchronous nuclear division, followed by cellularization, to create and release many daughter ‘zoospores’. The zoospores, akin to animal cells, crawl using actin-mediated cell migration. After forming a cell wall, polymerized actin reorganizes into fungal-like cortical patches and cables that extend into hyphal-like structures. Actin perinuclear shells form each cell cycle and polygonal territories emerge during cellularization. This work makes Spizellomyces a genetically tractable model for comparative cell biology and understanding the evolution of fungi and early eukaryotes.


Introduction
Zoosporic fungi, commonly referred to as 'chytrids', span some of the deepest fungal Phyla and comprise much of the undescribed environmental fungal DNA diversity in aquatic ecosystems (James et al., 2006;Richards et al., 2012;Powell and Letcher, 2014;Grossart et al., 2016). Many chytrids are saprophytes or parasites of photosynthetic organisms and actively shuttle carbon to higher trophic levels (Kagami et al., 2014;Grossart et al., 2016). Other chytrids are animal parasites, including the Batrachochytrium genus that includes the frog-killing B. dendrobatidis (Longcore et al., 1999) and salamander-killing B. salamandrivorans that are devastating global amphibian populations (Martel et al., 2013).
Chytrids are unique in that they have retained ancestral cellular features, shared by animal cells and amoebae, while also having fungal features. For example, chytrids begin their life as motile zoospores that lack a cell wall, swim with a single posterior cilium nucleated from a centriole, and crawl across surfaces (Fuller, 1976;Sparrow, 1960;Deacon and Saxena, 1997;Held, 1975;Fritz-Laylin et al., 2017b). Later life cycle stages exhibit fungal characteristics including the formation of chitinous cell walls, the growth of hyphal-like structures, and the development of a sporangium (sporangiogenesis); see Figure 1. Chytrid zoosporogenesis involves multiple rounds of mitosis without cytokinesis to create a multi-nuclear coenocyte, followed by cellularization to form zoospores with a single nucleus. The formation of a multinuclear compartment followed by cellularization is reminiscent of development in flies (e.g. Drosophila), amoeba (e.g. Physarum), and protozoa (e.g. Plasmodium). Although there are important differences from fly embryos (particularly the need for the chytrid sporangium to extract nutrients from the environment and coordinate growth with the cell cycle), determining the mechanisms controlling chytrid cellularization provides a comparative framework for understanding cellularization in animals and other eukaryotic lineages.
The major bottleneck to studying chytrids in molecular detail has been the lack of a model organism with tools for genetic transformation. Here, we describe the successful adaptation of Agrobacterium-mediated transformation to generate reliable and stable genetic transformation of the soil chytrid Spizellomyces punctatus. We expressed fluorescent proteins fused to histone and actin-binding proteins to characterize the development and cell biology of Spizellomyces throughout its life cycle using live-cell imaging. Below, we will show that Spizellomyces is a well-suited model system for uncovering molecular mechanisms of cell cycle regulation, cell motility, and development because it is fast-growing, displays both crawling and swimming motility, and possesses a characteristic chytrid developmental life cycle. These tools allow, for the first time, direct molecular probing to test new hypotheses about the evolution and regulation of the cell cycle (Medina et al., 2016), cell motility (Fritz-Laylin et al., 2017b), and development in chytrid fungi.  The uninucleate zoospore (nucleus in blue) has a cilium associated with a basal body. Swimming zoospores can also crawl on surfaces using amoeboid-like motion (polymerized actin in red). (B) The start of encystment (before 1 hr) occurs when the cilium retracts by a lash-around mechanism, followed by formation of cell wall (Koch, 1968). (C) The cyst then germinates and forms a single germ tube (at 1-3 hr) that later expands and branches into a rhizoidal system. The nucleus remains in the cyst during germ tube expansion as the cyst develops into a single reproductive structure called the sporangium. The first mitotic event (at 8-12 hr) usually correlates with the ramification of rhizoids from the germ tube. (D) Mitosis in the sporangium is coordinated with growth, as nuclei replicate and divide in a shared compartment. There can be a total of five to eight synchronous mitotic cycles as each sporangium develops a branched rhizoid system with subsporangial swelling in the main rhizoid. (E) Mitosis halts and zoospore formation begins in the sporangium. Ciliogenesis likely occurs before cellularization as in other chytrids (Renaud and Swift, 1964). (F) The nuclei cellularize and develop into zoospores while the sporangium develops discharge papillae. Once cellularization is complete and environmental conditions are favorable, the zoospores will escape the sporangium through the discharge papillae (at 20-30 hr). Diagram not drawn to scale. Times are relative to the start of microscopy after zoospore harvest.

Developing tools for genetic transformation
The plant pathogen Agrobacterium tumefaciens normally induces plant tumors by injecting and integrating a segment of transfer DNA (T-DNA) from a tumor-inducing plasmid (Ti-plasmid) into the plant genome. Researchers have exploited this feature to integrate foreign genes in plants by cloning them into the T-DNA region of the Ti-plasmid, inducing virulence genes for processing/transport of T-DNA, and co-culturing induced Agrobacterium with the desired plant strain. Because Agrobacterium-mediated transformation has been adapted for transformation of diverse animals and fungi (Bundock et al., 1995;Kunik et al., 2001;Covert et al., 2001;Ianiri et al., 2017;Vieira and Camilo, 2011), we chose to use this system in Spizellomyces punctatus. To this end, we modified an Agrobacterium plasmid to integrate and express a selectable marker (e.g. drug resistance) in Spizellomyces.
To determine a suitable selection marker for Spizellomyces, we tested the effects of drugs on the growth of the chytrid on agar plates. We spread zoospores on plates with various concentrations of drugs and assessed the cultures for cell growth, colony formation, and zoospore release using light microscopy. Although Geneticin (G418), Puromycin, and Phleomycin D10 (Zeocin) did not inhibit growth up to 800 mg/L, we determined that 200 mg/L Hygromycin and 800 mg/L Nourseothricin (CloNAT) resulted in complete absence of growth after 6 days of incubation at 30˚C. All remaining experiments were performed using Hygromycin (200 mg/L).
Next, we identified Spizellomyces promoters that can drive gene expression at sufficient levels to provide resistance to Hygromycin and measurable protein fluorescence. In the absence of a chytrid system to perform these tests, we reasoned that Spizellomyces promoters that express at high levels in yeast (Saccharomyces cerevisiae) would likely also work in chytrids. Therefore, we used an Agrobacterium plasmid  that propagates in yeast to first screen Spizellomyces promoters that successfully express a fusion of Hygromycin resistance (hph) and green fluorescent protein (GFP); see Materials and methods. We confirmed that Spizellomyces HSP70 and H2B promoters resulted in resistance to Hygromycin as well as measurable GFP fluorescence in yeast via flow cytometry (Figure 2-figure supplement 1A) and microscopy. All remaining experiments were performed using the stronger H2B promoter.
With active promoters and effective selection, we performed Agrobacterium-mediated transformation by co-culturing Spizellomyces zoospores with Agrobacterium carrying H2Bpr-hph-GFP; see Materials and methods. Although Hygromycin-resistant, none of the GFP transformants (Figure 2figure supplement 1B) exhibited green fluorescence above background. This has been seen in other emerging model systems (i.e. choanoflagellate Salpingoeca Booth et al., 2018) and is likely due to GFP misfolding. When GFP was replaced by tdTomato, we obtained transformants that exhibited both Hygromycin resistance ( . We then designed a construct with greater applicability, in which the selectable marker and fluorescent protein are expressed independently and where the fluorescent protein (tdTomato) is fused in-frame to the C-terminus of a protein of interest (POI). This design exploits the compact and divergent architecture of the Spizellomyces H2A/H2B promoters to express (POI)-tdTomato in an upstream direction (H2B promoter) while expressing hph in a downstream direction (H2A promoter). As a proof of concept and because we were interested in following nuclear dynamics to measure the timing and synchrony of mitotic events (e.g. DNA segregation, see next section), our first protein of interest was histone H2B; see Figure 2A.

Mitotic cycles are fast and highly synchronous during sporangiogenesis
Chytrid transformation with H2B-tdTomato resulted in bright nuclear localization of fluorescence when compared to cytoplasmic tdTomato ( Figure 2B). The presence of the T-DNA (total size of 4280 bp) in the transformants was confirmed using PCR for hph and H2B-tdTomato (   Figure 2. Genomic integration of H2B-tdTomato using Agrobacterium-mediated transformation. (A) Plasmid GI3EM20C takes advantage of the divergent architecture of H2A/H2B to express an H2B-tdTomato fusion in an upstream direction (H2B promoter) while expressing hph in a downstream direction (H2A promoter). (B) Representative images from wild type (left), and transformants expressing cytoplasmic hph-tdTomato (plasmid GI3EM18) (center) and nuclear-localized H2B-tdTomato (right). Top row shows DIC and the middle row shows fluorescence microscopy at 561 nm with overlaid images on the bottom row. For comparable results, all strains are presented at the same intensity levels used for H2B-tdTomato fluorescence image. Scale bar indicates ten microns. Image acquisition conditions: POL: transmittance 32%, exposure 0.15 s; TRITC filter, maximal projection, transmittance 32%, Exposure 0.2 s, 0.3 micrometers slice thickness. (C) Southern blot of four transformants, in which genomic DNA was digested either with XbaI or KpnI and probed using the Hygromycin resistance gene (hph). We used plasmid GI3EM20C as a positive control (+). (D) Four independent transformants were transferred, every two days, in both selective and non-selective medium at 30˚C for a total of 27 days (minimum of 23 life cycles or 116-185 mitotic cycles), followed by a challenge on selective medium. These strains were spotted in a twofold dilution series on non-selective and selective (+Hygromycin) plates, and incubated for 2 days at 30˚C. The online version of this article includes the following video, source data, and figure supplement(s) for figure 2:       In three of the four transformants (EM20C-1, 2, 3), the T-DNA LB was located within 200 bp upstream of the transcription start site (TSS) of a gene (SPPG_04375 an M48 peptidase, SPPG_03425 an adenine nucleotide hydrolase, and SPPG_02523 a PHO-like cyclin, respectively). For strain EM20C-3, two different DNA:genome junctions were detected in the 5'UTR of the gene SPPG_02523, suggesting an irregular T-DNA insertion. Last, for strain EM20C-4, the LB T-DNA was inserted 844 bp from the closest TSS (SPPG_08788 a hypothetical protein). As observed in Arabidopsis, we might expect variation in gene expression based on the genomic locus of integration of the T-DNA fragment. To quantify this variation, we measured H2B-tdTomato expression in our EM20C transformants using flow cytometry (Figure 2figure supplement 4). The data show that strains EM20C-2, 3, 4 showed similar and unimodal levels of tdTomato fluorescence at the population level, despite the different sites of genomic integration. The exception is EM20C-1, which exhibits bimodal gene expression: the top mode is identical to the other transformants, but the bottom mode is half the intensity. Last, we established that the transformants had transgenerational stability by passaging them in non-selective medium for several weeks, followed by a challenge in selective medium ( Figure 2D).
To quantify the timing and synchrony of the Spizellomyces cell cycle, we used live cell epi-fluorescence imaging of H2B-tdTomato strains EM20C-1 and EM20C-2 at 2 min intervals (Figure 2-figure supplement 5). Our results show that zoospores have a single nucleus, they retract their flagellum and encyst in less than 1 hr, the germ tube emerges at~1À3 hr, the first mitotic event (i.e. one nucleus to two nuclei) occurs at~8À12 hr, and sporangia develop and undergo 5-8 mitotic cycles in less than 30 hr before completing their life cycles and releasing 32-256 zoospores; see Figure 2video 1. To measure all nuclei within a sporangium with better temporal and z-resolution, we followed nuclear dynamics of EM20C-1 at 1 min time intervals using live-cell confocal microscopy of a H2B-tdTomato strain ( Figure 3-video 1). We measured the number of nuclei over time per sporangium ( Figure 3A) to estimate the synchrony of nuclear division waves and the period of time between waves of nuclear division. Wave time (Dt) is the time for a wave of synchronous nuclear divisions to propagate across the sporangium. The cell cycle period (t ) is the interval of time between nuclear division waves. We found that the average cell cycle period was~150 min and that each wave of nuclear division was completed within 1 min ( Figure 3B). In addition, by following the compaction and localization of H2B-tdTomato we show that all measurable mitotic events occurred within 5 min, or less than 3.3% of the cell cycle period ( Figure 3B and C). Altogether, these results show that Spizellomyces is mitotically inactive in its early life cycle (zoospore and germination stages) but, once committed, the cell cycle is fast and nuclear divisions within each sporangium are highly synchronous; see Table 1.

Actin polymerization drives zoospore motility
Like their pre-fungal ancestors, chytrids swim with a motile cilium. Some chytrids can also crawl across and between solid substrates, much like amoeba and animal immune cells (Fuller, 1976;Sparrow, 1960;Whisler et al., 1975;Couch, 1945;Deacon and Saxena, 1997;Held, 1973;Dorward and Powell, 1983;Fritz-Laylin et al., 2017b). Eukaryotes employ multiple strategies to crawl (filopodia, pseudopodia and blebs) that depend on distinct molecular mechanisms (Fritz-Laylin et al., 2017a). One form of crawling, the pseudopod-based a-motility, relies on the expansion of branched-actin filament networks that are assembled by the Arp2/3 complex and allow cells to navigate complex environments at speeds exceeding 20 mm/min. The activators of branched-actin assembly WASP and SCAR/WAVE have been described as a molecular signature of the capacity for a-motility (Fritz-Laylin et al., 2017b). Spizellomyces has homologs of WASP (SPPG_00537), SCAR/ WAVE (SPPG_02302), and its zoospores are proficient crawlers.
To test whether Spizellomyces zoospores crawl using a-motility, we expressed a LifeAct-tdTomato fusion. LifeAct is a 17 amino acid peptide that binds specifically to polymerized actin in a wide variety of cell types, such as actin patches and cables in yeast and actin-filled protrusions of crawling animal cells (Riedl et al., 2008;Belin et al., 2014). To confirm that our LifeAct-tdTomato fusion binds specifically to polymerized actin in Spizellomyces, we first fixed and stained actin in zoospores ( Figure 4) and sporangia ( Figure 5) with fluorescent phalloidin. We then compared the fluorescent images of cells expressing LifeAct-tdTomato and those expressing hph-tdTomato (negative control) relative to phalloidin-staining.
In contrast to the relatively homogeneous distribution of fluorescence in hph-tdTomato zoospores, both fixed ( Figure 4A-C) and living zoospores ( Figure 4D-E, and Figure 4-video 1) of the LifeAct-tdTomato strain showed a thin layer of fluorescence at the cell cortex, and high levels of fluorescence in the pseudopods at the leading edge. Because the LifeAct fusion localization nearly perfectly correlated with phalloidin intensities in fixed cells, but the hph fusion did not, we presume that this fluorescence represents polymerized actin. The minor deviations between Lifeact and phalloidin can be explained by intrinsic biases of the probes. All live-cell probes of filamentous actin (e.g. Lifeact, F-tractin, Utrophin actin-binding domain) have different biases in their patterns of F-actin  (1) leakage from the nucleus likely due to fenestration of nuclear envelope by the mitotic spindle (Heath, 1980;Fuller, 1976), followed by chromosome condensation, and (2) chromosome separation during anaphase. Dotted line highlights the cell wall of the sporangium. This particular example shows a mitosis duration of 4 min (t M =time from nuclear leakage to anaphase). Time in hr:min. Scale 2.5 micrometers. Distributions are from one time-lapse movie of EM20C-1 (6 cells). The online version of this article includes the following video and source data for figure 3: Source data 1. Number of nuclei per cell as a function of time (min). Figure 3-video 1. Time-lapse microscopy of nuclear divisions in developing H2B-tdTomato sporangia taken over the course of 24 hr with images captured every 1 min. https://elifesciences.org/articles/52741#fig3video1 localization and dynamics (Belin et al., 2014). None of them can fully recapitulate the patterns seen with phalloidin, but they do provide new insights into the live cell dynamics of actin. The distribution and dynamics of actin that we see in zoospores is in agreement with a-motility and actin localization observed for fixed cells during zoospore crawling of Neocallimastix (Li and Heath, 1994) and Batrachochytrium (Fritz-Laylin et al., 2017b).

Actin polymerization during sporangiogenesis
Once the chytrid zoospore encysts, it builds a sporangium by growing both radially and in polarized fashion during germ tube extension and rhizoid formation (Figure 1). During early sporangiogenesis, the nuclei are very dynamic while replicating and dividing but then slow down in late sporangiogenesis, presumably during cellularization and zoospore formation (Figure 2-video1). Actin has been reported to play fundamental roles during cellularization in another chytrid, Allomyces macrogynus (Lowry et al., 1998;Lowry et al., 2004). Thus, we expect polymerized actin to play a role in the nuclear dynamics and cellularization during sporangiogenesis in Spizellomyces.
Our experiments revealed an actin cytoskeleton distributed primarily between cortical patches and linear structures that resemble the actin cables of other fungal species ( Figure 5A). Each of these structures was visible in both fixed phalloidin stained cells and living cells expressing the Life-Act-tdTomato gene. We also detected transient perinuclear polymerized actin shells with a period similar to mitotic events ( Figure 5B and Figure 5-video 1), which suggests a role for polymerized actin with nuclei during the cell cycle. In late sporangiogenesis, polymerized actin delineated polygonal zoospore territories ( Figure 5B-C). This is reminiscent of epithelial cellularization seen during Drosophila embryogenesis when syncytial nuclei are encapsulated by cell membrane.

Discussion
Here, we report stable and robust genetic transformation of the chytrid Spizellomyces punctatus. We identified and tested native Spizellomyces promoters, including a divergent H2A/H2B promoter that can simultaneously express Hygromycin resistance (hph) and a gene of interest throughout the chytrid life cycle. This design was used to express fluorescent proteins and resistance genes in Dikaryotic marine fungi (data not shown), which suggests they should be useful for transforming other chytrids, such as the pathogenic Batrachochytrium. Agrobacterium was previously used to transform the aquatic chytrid Blastocladiella emersonii (Vieira and Camilo, 2011). Unfortunately, genetic transformation of Blastocladiella displayed a narrow window of gene expression and could only be detected at the zoosporic stage. The advances described in this study should be useful for random and possibly targeted gene integration in other chytrids. For example, Agrobacterium T-DNA is often randomly integrated into the host chromosome as a single copy and, if inserted within a gene, it will disrupt gene function and act as a mutagen (Gelvin, 2017). Random insertional mutagenesis by Agrobacterium has been exploited for forward genetics in plants and fungi (Alonso et al., 2003;Michielse et al., 2005;Idnurm et al., 2017). In Fungi, Agrobacterium-mediated transformation has also been used for targeted gene knock-outs by including 1-1.5 kb long,  The online version of this article includes the following video for figure 5: Figure 5-video 1. Time-lapse microscopy of developing LifeAct-tdTomato sporangia taken over the course of 19 hr with images captured every 5 min. https://elifesciences.org/articles/52741#fig5video1 allocates a narrow window (~5 min) of time to the mitotic process during each cell cycle with a highly synchronous wave of nuclear divisions. The level of synchrony is similar to plasmodial nuclei in the amoeba Physarum polycephalum or syncytial nuclei during Drosophila development, where all nuclei divide within 2 min (Deneke et al., 2016) (see Table 1). These mitotic dynamics make Spizellomyces an interesting comparative model for exploring the physical and molecular determinants of cell division synchrony and its evolution in fungi, animals, and amoeba. By visualizing actin dynamics throughout the chytrid life cycle, we found that Spizellomyces zoospores assemble thick actin cortices and build actin-filled protrusions during motility, similar to other chytrid species (Fritz-Laylin et al., 2017b), animal cells, and amoebae. Once the zoospores encyst, there is a drastic shift in actin cytoskeleton organization, in which the cortical shell of polymerized actin was replaced by dynamic puncta (actin patches), some of which are associated with actin cables that often extended into the germ tube or rhizoids. This architecture is typical of fungi, where actin patches are associated with endocytosis and cell wall deposition, whereas actin cables are pathways for targeted delivery of exocytic vesicles (Lichius et al., 2011). This biphasic actin distribution -an actin cortex and actin-filled protrusions in zoospores, and actin patches and cables in sporangiaindicates that, like its cell cycle regulatory network (Medina et al., 2016), the actin cytoskeleton of Spizellomyces displays features that resemble those of both animal and fungal cells.
Our results revealed further actin dynamics and organization during the chytrid development. This includes the formation of perinuclear actin shells that are fleetingly detected by live imaging. Previous observations suggested that perinuclear actin shells of anaerobic chytrids were a fixation artifact (Li and Heath, 1994); however, our live cell data indicate that these are real, dynamic cellular structures that likely occur in many chytrids. Similar perinuclear actin shells are associated with nuclear lamina in animal cells, where they play a role in changing nuclear shape before and after mitosis (Clubb and Locke, 1998;Clubb and Locke, 1996 or when squeezing through narrow channels (Thiam et al., 2016). Although chytrids and other fungi have lost nuclear lamins, it seems likely that Spizellomyces perinuclear actin shells are associated with changes in nuclear shape during the cell cycle.
Finally, we showed that polymerized actin is likely involved in the formation of zoospore polygonal territories ( Figure 5B) before the formation of the cleavage planes during cellularization. In contrast to cellularization in Drosophila, which occurs at the surface of the embryo (i.e. two-dimensional cellularization), cellularization in Spizellomyces happens in a three-dimensional context within 3 hr. The chytrid cellularization process is reminiscent of the polarized epithelium of the social amoeba Dictyostelium discoideum (Dickinson et al., 2011;Dickinson et al., 2012) and the membrane invagination dynamics during cellularization in the ichthyosporean Sphaeroforma arctica, an unicellular relative of animals . To what extent is the emergence of multicellularity, which appears to have evolved independently in amoeba, animals, and fungi, dependent upon a shared ancestral toolkit? Establishing shared and unique traits between chytrid fungi and other key lineages will provide a powerful cross-lineage experimental system to test core hypotheses on the evolution of multicellularity and derived fungal features. Based on our tools and findings, we expect that Spizellomyces will be a useful model system to study the evolution of key animal and fungal traits.

Strains and growth conditions
We used Spizellomyces punctatus (Koch type isolate NG-3) Barr (ATCC 48900) for all chytrid experiments. Unless otherwise stated, Spizellomyces were grown at 30˚C in Koch's K1 medium (1L; 0.6 g peptone, 0.4 g yeast extract, 1.2 g glucose, 15 g agar if plates; Koch, 1957). Two days prior to harvesting zoospores, we aliquoted and spread 1 mL of active, liquid culture pregrown in K1 medium onto K1 plates and incubated them to allow zoospores to encyst, mature, and colonize the agar surface. We flooded each active Spizellomyces plate with 1 mL of dilute salt (DS) solution (Machlis, 1953) and incubated at room temperature. After 1 hr, released zoospores were retrieved by harvesting the DS medium and purified by slowly filtering the harvest in Luer-Lok syringe through an autoclaved syringe filter holder (Advantec REF:43303010) preloaded with a 25 mm Whatman Grade one filter paper (CAT No. 1001-325). Strains listed in Table 3 are available from the Buchler lab upon request.

Plasmids
We initially tried to transform Spizellomyces via zoospore electroporation using a protocol developed in zoosporic protists, such as Phytophthora (Ah-Fong et al., 2018). This was unsuccessful and we turned to Agrobacterium-mediated transformation because it has worked in other fungi. We used the pGI3 plasmid backbone for Agrobacterium-mediated transformation , which contains the Saccharomyces cerevisiae 2m replication origin and the URA3 selectable marker. This allows pGI3 and its derivatives to replicate in E. coli, A. tumefaciens and S. cerevisiae. Complete details of primers and plasmid construction are in Table 2 and the Appendix. All plasmids are available from Addgene and their RRIDs are listed in the Appendix.

Agrobacterium-mediated transformation of Spizellomyces
We prepared competent Agrobacterium EHA105 strains following the protocol of Weigel and Glazebrook (2006). Plasmids were transformed into competent Agrobacterium using 0.2 cm cuvettes in a Gene Pulser electroporator (Bio-Rad, USA) at 25 mF, 200 W , 2.5kV. Single colonies were streaked on selective plates (Kanamycin). A colony of transformed Agrobacterium containing pGI3derived plasmid was grown overnight at 30˚C in 5 mL of Luria-Bertani broth supplemented with Kanamycin (50 mg/L). After centrifugation, the cell pellet was resuspended in 5 mL of IM (Bundock et al., 1995), diluted to an OD 660 of 0.1 and grown under agitation at 30˚C until achieving a final OD 660 of 0.6, at which point the culture was ready for co-culturing with the chytrid (300 mL per transformation). IM is composed of MM salts (Hooykaas et al., 1979) and 40M 2-(N-morpholino)ethanesulfonic acid (MES) pH 5.3, 10 mM glucose, 0.5% (w/v) glycerol and 200mM acetosyringone.
In parallel, chytrid zoospores were harvested and pelleted by centrifugation at 800 g for 10 min. Zoospores were then gently resuspended in 300mL of induction medium (IM). We found that one K1 plate provides enough zoospores for one transformation. For every transformation, zoospores and Agrobacterium were combined at four different ratios: 1:1, 1:0.25, 0.25:1, and 0.25:0.25 in a total volume of 200mL. To guarantee tight contact between A. tumefaciens-S. punctatus cells, the surface of the IM plate was rubbed with the bottom of a sterile glass culture tube to generate slightly concave depressions in each quadrant of the plate, wherein each 200mL co-incubation mixture was spotted. Plates were incubated unsealed for 4 days at room temperature. Mock transformations with empty Agrobacterium (no binary plasmid; grown in the absence of plasmid selective medium) were included as a negative control.
After co-incubation, we added 1 mL of DS solution and gently scraped the plate with a razor blade, pooling the different cell ratios into a single 50 mL centrifuge tube, raising the volume to 20 mL with DS solution, and re-suspending clumps by inversion. The mixture was centrifuged at 1000 g for 10 min and the liquid phase was discarded. The remaining pellet was carefully resuspended with DS solution, plated on K1 plates containing Ampicillin (50 mg/L) and Tetracycline (50 mg/L) to select against Agrobacterium and Hygromycin (200 mg/L) to select for transformed Spizellomyces. Spizellomyces survival controls were performed by plating transformations after co-culture in non-selective K1 media (Ampicillin (50 mg/L), Tetracycline (50 mg/L)). Transformation plates and controls were incubated at 30˚C until colonies were observed (5-6 days). All plates were sealed with parafilm to prevent desiccation. Single colony isolates were retrieved with a sterile needle, resuspended in DS solution and re-plated on a selective Hygromycin plate. Chytrid strains listed in Table 3 are available from the Buchler lab upon request.

Nucleic acid manipulation
High-molecular-weight genomic DNA extraction was performed using CTAB/Chloroform protocol (CTAB/PVP buffer: 100 mM Tris-HCl pH7.5; 1.4M NaCl; 10 mM EDTA; 1% CTAB; 1% PVP; 1% b-Mercaptoethanol (added just before use)). Briefly, 10 plates of the selected strain were grown at 30C for 2 days and zoospores were harvested and purified as described before. The pellet of zoospores was resuspended directly in 900mL of CTAB/PVP buffer pre-warmed to 65˚C and transferred to 2 mL centrifuge tubes. Tubes were centrifuged briefly and incubated at room temperature for 1 hr in a nutating mixer. After 5 min incubation on ice, DNA was extracted with Chloroform (Sigma-Aldrich REF:288306) twice followed by treatment of supernatant with 100 ng RNase A (Biobasic; 60 U/mg Ref:XRB0473) for 30 min at room temperature. DNA was then precipitated by adding 0.2 vol  Identification of T-DNA insertion sites by inverse PCR. 2.5 mg of genomic DNA of wild type (WT) and the four transformed strains of Spizellomyces (EM20C-1,2,3,4) was digested in a final volume of  50mL with 100U of EcoRI-HF (NEB R3101S) or HindIII-HF (NEB R3104S) for 24 hr at 37˚C. After assessing the quality of the digestion by gel electrophoresis, the reaction was heat inactivated and 48mL of the digested DNA was incubated with 1mL of T4 ligase (400u/mL) for 48 hr at 4˚C. The ligation was purified by chloroform extraction twice, followed by DNA precipitation and resuspended in 30mL of nuclease-free water. 100 ng of ligated product was used for touch-down Inverse PCR reactions in a final volume of 50mL using NEB Phusion high-fidelity DNA polymerase and 3% DMSO following manufacturer instructions and primers ai77_F and 4_LB_R. Touch-down PCR amplification protocol included an initial denaturation step at 98˚C for 3 min followed by 10 cycles of amplification in which the annealing temperature was decreased 1C/cycle until an annealing temperature of 62˚C was achieved, followed by 20 amplification cycles at 62˚C. Annealing and elongation time during all these cycles was 30 s and 6 min, respectively. Amplification was assessed by gel electrophoresis and bands were retrieved using a razor blade, purified using the Promega Wizard SV Gel and PCR Clean-Up System, and Sanger-sequenced using the primer ai77_F.

Microscopy
H2B-tdTomato-expressing chytrids were harvested from plates, placed in a glass-bottom dish (Mattek), and covered with a 1.5% K1 agarose pad to keep cells healthy and in the plane of focus (Young et al., 2012). We harvested zoospores from plates and re-suspended them in Leu/Lys paralyzing solution (Dill and Fuller, 1971 ) before putting them in glass-bottom dishes, as above. Livecell epifluorescence was performed on a temperature-controlled Deltavision Elite inverted microscope equipped with 60x/1.42 oil objective and Evolve-512 EMCCD camera using optical axis integration. Optical axis integration (OAI) is a setting on the Deltavision microscope that opens the shutter and continuously excites and measures fluorescence emission while sweeping from top (+20 microns) to bottom of the z-axis (À20 microns) in 0.5 s, integrating directly the intensities onto the CCD chip. OAI has the advantage that total exposure time is reduced relative to a traditional z-stack. Excitation light was 542/27 nm (7 Color InsightSSI) and emission to the camera was filtered by 594/45 nm (TRITC). Epifluorescence live-cell imaging was done at 30˚C. Live-cell confocal microscopy was performed on a Nikon Eclipse Ti inverted microscope equipped with a 100x/N.A. 1.49 CFI Apo TIRF oil objective and fitted with a Yokogawa CSU-X1 spinning disk and Andor iXon 897 EMCCD camera. Excitation light was via 488 nm laser. Confocal live-cell imaging was done at 27˚C. LifeAct-expressing zoospores were collected in DS solution and transferred to cover-glass bottom dishes. For phalloidin staining, glass coverslips were plasma cleaned and immediately coated with 0.1% polyethyleneimine for 5 min, washed thrice with water, then overlaid with zoospores or sporangia suspended in DS solution. Cells were allowed to adhere for 5 min before fixation by adding four volumes of 4% paraformaldehyde in 50 mM Cacodylate buffer (pH 7.2). Cells were fixed for 20 min on ice, washed once with PEM buffer (100 mM PIPES, 1 mM EGTA, 0.1 mM MgSO 4 ), permeabilized and stained with 0.1% Triton X in PEM with 1:1000 Alexa Fluor 488 Phalloidin (66 nM in DMSO, Sigma D2660) for 10 min at room temperature, washed with PEM, then mounted onto glass slides using Prolong Gold with DAPI (Invitrogen P36931). Zoospores were imaged on a Nikon Ti2-E inverted microscope equipped with 100x oil PlanApo objective and sCMOS 4mp camera (PCO Panda). Excitation light was via epi fluorescence illuminator at 405 nm, 488 nm, and 561 nm. Sporangia were imaged on a Nikon Ti-E inverted microscope equipped with 100X oil objective and fitted with a Yokogawa X1 spinning disk (CSU-W1) with 50 mm pinholes and Andor xIon EMCCD camera. Excitation light was via 405 nm laser, 488 nm laser, and 561 nm laser. Z-stack fluorescent images were deconvolved with NIS Elements v5.11 using 20 iterations of the Richardson-Lucy algorithm. Image analysis was performed with the ImageJ bundle Fiji (Schindelin et al., 2012). All imagings were done at room temperature.

Flow cytometry
Zoospores from wild-type and EM20C-(1-4) transformants were harvested from plates and put on ice until fluorescence measurement. Flow cytometry was performed on a MACSQuant VYB using the yellow laser (561 nm) to measure forward scatter (FSC), side scatter (SSC), and tdTomato fluorescence (Y2, 615/20 nm filter). The FSC, SSC, Y2 voltage gain settings for each PMT were 350 V, 350 V, and 510 V, respectively. We recorded the pulse height, width, area for a minimum of 100,000 events per strain. Singlets were identified using FSC/SSC width versus height, and a histogram of singlet tdTomato fluorescence (Y2 height) was plotted using the FlowJo software package.