Genetic Manipulation of the Brassicaceae Smut Fungus Thecaphora thlaspeos

Investigation of plant–microbe interactions greatly benefit from genetically tractable partners to address, molecularly, the virulence and defense mechanisms. The smut fungus Ustilago maydis is a model pathogen in that sense: efficient homologous recombination and a small genome allow targeted modification. On the host side, maize is limiting with regard to rapid genetic alterations. By contrast, the model plant Arabidopsis thaliana is an excellent model with a vast amount of information and techniques as well as genetic resources. Here, we present a transformation protocol for the Brassicaceae smut fungus Thecaphora thlaspeos. Using the well-established methodology of protoplast transformation, we generated the first reporter strains expressing fluorescent proteins to follow mating. As a proof-of-principle for homologous recombination, we deleted the pheromone receptor pra1. As expected, this mutant cannot mate. Further analysis will contribute to our understanding of the role of mating for infection biology in this novel model fungus. From now on, the genetic manipulation of T. thlaspeos, which is able to colonize the model plant A. thaliana, provides us with a pathosystem in which both partners are genetically amenable to study smut infection biology.


Introduction
Smut fungi are important pathogens causing economic losses in crops such as barley, wheat, maize, and potato [1]. The dimorphic lifecycle of grass smut fungi is comprised of a yeast form [2] that, in contrast to many other biotrophic fungi, can be cultured, and is amenable to genetic manipulation [3]. In combination with the efficient homologous recombination, this has turned Ustilago maydis, the maize smut fungus [4], into an important model organism for fungal and infection biology [5,6].
U. maydis starts the infection by mating, resulting in the morphological switch to the infectious filamentous form. Mating in smut fungi is controlled by two mating loci. The a locus encodes for pheromones (mfa) and pheromone receptors (pra) that mediate recognition of compatible mating partners and trigger cell fusion [7]. The b locus contains two genes (bW, bE) encoding for subunits of a heterodimeric, homeodomain transcription factor. When an active transcription factor is assembled from different alleles in the common cytosol after plasmogamy, filamentous growth is initiated and thereby the fungus switches from saprophytic to pathogenic growth [8]. Notably, the infectious filaments are arrested in the cell cycle until successful penetration of the host plant [9]. This mating system is widely conserved in grass smut fungi and allows for genetic exchange between the mating partners [10].
In contrast to the well-characterized grass smut fungi infecting important monocot crop plants with U. maydis at the forefront, little is known about smut fungi infecting dicot plants. A reemerging example is Microbotryum that is regaining attention today [11].

Fungal Culture Conditions
T. thlaspeos LF1 and LF2 haploid strains [15] from our own collection were used in this study. Both were grown in YEPS light liquid medium (1% yeast extract w/v, 0.4% w/v Bacto TM-Peptone, and 0.4% w/v sucrose) or YEPS light solid medium with 0.6% w/v plant agar or 1% w/v phytagel at 18 • C. Cryostocks of T. thlaspeos were generated by mixing exponentially growing cultures with 30% glycerol in growth medium followed by immediate freezing at −80 • C. Filamentous cultures were started by resuspending the glycerol stock in growth medium or plating the cells on solid medium as described above.

Plasmid Construction
All plasmids used in this study were generated with the Golden Gate Cloning technique (Protocol S1) as described [27]. The hpt-egfp and hpt-mcherry sequences were codonoptimized for U. maydis. Promoter and terminator sequences from T. thlaspeos or U. maydis were amplified via PCR from genomic DNA.

Protoplasting
T. thlaspeos cultures were grown in YEPS light medium to an OD 600 of 0.5-0.8 for 3 to 4 days. Fungal mycelium was collected using cell strainers with 40 µm mesh size (VWR TM Darmstadt, Germany) and washed with protoplasting buffer (0.1 M sodium citrate, 0.01 M EDTA 1.2 M MgSO 4 , and pH 5.8) to remove residual culture medium. T. thlaspeos tissue was resuspended in 9 mL protoplasting buffer, supplemented with 10 mg/mL Yatalase (Takara Bio, Kusatsu, Japan) and 20 mg/mL Glucanex (Sigma-Aldrich, St. Luis, MI, USA) per 100 mL cell culture, and incubated for 30-60 min at room temperature. Protoplast formation was controlled microscopically. When protoplasting was finished, protoplasting buffer was added to a total volume of 24 mL. Aliquots of 6 mL crude protoplast solution were overlayed with 5 mL trapping buffer (0.6 M sorbitol, 0.1 M Tris/HCl pH 7.0) and centrifuged at 4863× g (5000 rpm) in a swing out rotor at 4 • C for 15 min. The interphase was collected from all tubes and diluted with 2 volumes of ice-cold STC buffer (0.01 M Tris/HCl pH 7.5, 0.1 M CaCl 2 , and 1.0 M sorbitol). Protoplasts were pelleted at 4863× g (5000 rpm) in a swing out rotor at 4 • C for 10 min and resuspended in 500 µL ice-cold STC buffer. 100 µL protoplast aliquots were used for transformation immediately. A bullet point version of the protocol is available with the supplementary files (Protocol S1).

Transformation
Transformation of T. thlaspeos protoplasts were carried out as described for U. maydis [28] with slight modifications. Transformation reactions were spread on YMPG-Reg (0.3% w/v Yeast extract, 0.3% w/v malt extract, 0.5% w/v Bacto-Peptone, 1% w/v glucose, 1 M sucrose, 0.6% w/v plant agar, Duchefa Biochemie, Haarlem, Netherlands) medium and incubated at 18 • C until colonies appeared (1-2 months). Selection was carried out on 10 µg/mL Hygromycin B (Roth, Karlsruhe Germany) following the layered-plate strategy used for U. maydis [28]. In this setup, hygromycin was provided in the bottom layer so that it took time to diffuse to the top before the selection took place. This gave protoplasts time to regenerate and express the resistance gene [28]. Colonies were singled-out on YEPS light solid medium supplemented with 10 µg/mL hygromycin. Single colonies were then used to inoculate YEPS light liquid cultures for molecular analysis.

Molecular Analysis of Transformants
Successful integration of the constructs were determined by PCR and Southern Blot analysis [28], and eGfp or mCherry fluorescence was used as a rapid indicator for expression of the constructs. Genomic DNA of T. thlaspeos was extracted using the NEB Monarch Genomic DNA Purification Kit (New England Biolabs, Frankfurt, Germany).

Mating Assay
For confrontation assays, liquid cultures of T. thlaspeos strains were spotted on YEPS light solid medium in close proximity and allowed to grow towards each other. When the hyphae of both strains were close enough to appear in the display window of the microscope, a time lapse experiment was conducted to monitor mating for 24-72 h.
For liquid mating assays, strains of compatible mating types were mixed in YEPS light liquid medium in equal amounts and incubated at 18 • C and 200 rpm. Medium was exchanged twice a week. After 8-14 days, plasmogamy was observed microscopically via eGfp and mCherry fluorescence.

Microscopy
Fluorescence microscopy as well time lapse experiments were performed on a Zeiss Axio Immager M1 according to [29]. Microscope control, image acquisition, and processing were done with the software package Meta-Morph (version 7; Molecular Devices).

Protoplast-Generation from Filamentous T. thlaspeos Cultures
Protoplast-mediated transformation is well-established for fungi [30]. The protocol for U. maydis [17] was successfully adapted for other smut fungi and therefore was also the starting point for T. thlaspeos protoplast generation. Critical factors besides cultivation conditions are the enzyme mixture, the buffer composition, and the osmotic stabilizer.
First, we compared different enzyme mixtures. Glucanex, a mix of lysing enzymes from Trichoderma harzianum including beta-1,3-glucanase activity, works well for U. maydis and the filamentous Basidiomycete S. indica [22]. Yatalase comprises of a mix of lysing enzymes from Corynebacterium, including chitinase-, chitobiase-, and beta-1,3-glucanase activity, for cell wall lysis of filamentous fungi. In combination with Glucanex, it is used for U. bromivora [21] and Agrocybe aegerita [31], or supplemented with chitinase for Aspergillus niger [32]. Novozyme 234, which worked very well for U. maydis [28], is no longer commercially available, so we did not include it in our study. In pilot studies, we compared enzyme and buffer combinations of published protoplasting protocols and found that the combination of Glucanex and Yatalase efficiently protoplasts the T. thlaspeos filaments (Table S1).
Therefore, to first optimize the osmotic stabilizer, we used this enzyme mix in the U. maydis protoplasting buffer. Typical osmotic stabilizers are inorganic salts, sugars, or sugar alcohols [33]. For example, sorbitol is used for U. maydis, and sucrose for U. esculenta. Thus, we tested sorbitol and sucrose, as well as MgSO 4 , which is used frequently in combination with Yatalase. Most protoplasts were obtained using MgSO 4 ( Figure 1). Notably, sorbitol and sucrose inhibited cell wall lysis, confirming early observations [34,35]. Subsequently, we tested various commonly used protoplasting buffers, in combination with MgSO 4 as the osmotic stabilizer. There were no significant differences between the four tested buffers ( Table 1), but we observed tendencies that citrate buffers work better for fast growing cultures that were sub-cultured bi-weekly ( Figure S1). Ultimately, we decided on the 0.1 M citrate buffer and 0.01 M EDTA, which is also the buffer used for An. flocculosa, the closest homolog of T. thlaspeos, and a bi-weekly splitting rhythm.

Protoplast-Generation From Filamentous T. thlaspeos Cultures
Protoplast-mediated transformation is well-established for fungi [30]. The protocol for U. maydis [17] was successfully adapted for other smut fungi and therefore was also the starting point for T. thlaspeos protoplast generation. Critical factors besides cultivation conditions are the enzyme mixture, the buffer composition, and the osmotic stabilizer. First, we compared different enzyme mixtures. Glucanex, a mix of lysing enzymes from Trichoderma harzianum including beta-1,3-glucanase activity, works well for U. maydis and the filamentous Basidiomycete S. indica [22]. Yatalase comprises of a mix of lysing enzymes from Corynebacterium, including chitinase-, chitobiase-, and beta-1,3-glucanase activity, for cell wall lysis of filamentous fungi. In combination with Glucanex, it is used for U. bromivora [21] and Agrocybe aegerita [31], or supplemented with chitinase for Aspergillus niger [32]. Novozyme 234, which worked very well for U. maydis [28], is no longer commercially available, so we did not include it in our study. In pilot studies, we compared enzyme and buffer combinations of published protoplasting protocols and found that the combination of Glucanex and Yatalase efficiently protoplasts the T. thlaspeos filaments (Table S1).
Therefore, to first optimize the osmotic stabilizer, we used this enzyme mix in the U. maydis protoplasting buffer. Typical osmotic stabilizers are inorganic salts, sugars, or sugar alcohols [33]. For example, sorbitol is used for U. maydis, and sucrose for U. esculenta. Thus, we tested sorbitol and sucrose, as well as MgSO4, which is used frequently in combination with Yatalase. Most protoplasts were obtained using MgSO4 ( Figure 1). Notably, sorbitol and sucrose inhibited cell wall lysis, confirming early observations [34,35]. Subsequently, we tested various commonly used protoplasting buffers, in combination with MgSO4 as the osmotic stabilizer. There were no significant differences between the four tested buffers (Table 1), but we observed tendencies that citrate buffers work better for fast growing cultures that were sub-cultured bi-weekly ( Figure S1). Ultimately, we decided on the 0.1 M citrate buffer and 0.01 M EDTA, which is also the buffer used for An. flocculosa, the closest homolog of T. thlaspeos, and a bi-weekly splitting rhythm. Figure 1. Identification of an osmotic stabilizer. Thecaphora thlaspeos LF1 culture was grown to an OD600 = 0.4-0.8. To optimize protoplasting of T. thlaspeos hyphae by Yatalase and Glucanex, the osmotic stabilizers MgSO4, sorbitol, and sucrose were tested. With MgSO4 as osmotic stabilizer, all filaments were digested; while in sorbitol and sucrose, no protoplasts were obtained. Black arrowheads: filaments; white arrowheads: protoplast; scale bar: 50 µm. Table 1. Optimizing the protoplasting buffer. To identify the optimal osmotic stabilizer, fungal hyphae were filtered and incubated in 0.02 M citrate buffer, supplemented with different osmotic stabilizers and 10 mg/mL Yatalase + 20 mg/mL Glucanex, for 60 min at RT. Protoplasting worked only if MgSO4 was used as osmotic stabilizer. To optimize the buffer for the use of MgSO4, hyphae were filtered and incubated in different buffers, supplemented with 1.2 M MgSO4 and 10 mg/mL Yatalase + 20 mg/mL Glucanex, for 60 min at RT. There was no significant difference between the indicated buffers, but a tendency towards higher yields with citrate buffers.
An advantage of MgSO 4 as osmotic stabilizer is that the majority of intact protoplasts in the presence of MgSO 4 have large vacuoles [38], which enables collection and purification, and floating in a trapping buffer [36]. Intact protoplasts accumulate in a sharp band at the interphase, and debris pellet at the bottom ( Figure 2). After washing with STC buffer, up to 10 8 protoplasts/g fresh weight can be recovered.
Upon determination of the optimal buffer and osmotic stabilizer, we reevaluated the composition and concentration of the enzyme mix ( Figure 3). As the pilot study had indicated, for efficient degradation of the T. thlaspeos cell wall, the combined activity of Yatalse and Glucanex is necessary. As expected, individually the enzymes are poorly active, leaving filaments behind ( Figure 3A). This emphasizes the importance of testing various lysing enzymes, alone and in combination, in different buffers to find a mix suitable for the organism of choice and its individual cell wall composition [30].
Finally, we aimed to decrease the enzyme concentrations to save costs. However, lowering the concentration to half resulted in incomplete digestion of the fungal cell wall after 30 min ( Figure 3B). Since it was described earlier that a shorter incubation time is preferable, compared to a low enzyme concentration, regarding protoplast viability [39,40], we did not reduce the enzyme concentrations. Table 1. Optimizing the protoplasting buffer. To identify the optimal osmotic stabilizer, fungal hyphae were filtered and incubated in 0.02 M citrate buffer, supplemented with different osmotic stabilizers and 10 mg/mL Yatalase + 20 mg/mL Glucanex, for 60 min at RT. Protoplasting worked only if MgSO 4 was used as osmotic stabilizer. To optimize the buffer for the use of MgSO 4 , hyphae were filtered and incubated in different buffers, supplemented with 1.2 M MgSO 4 and 10 mg/mL Yatalase + 20 mg/mL Glucanex, for 60 min at RT. There was no significant difference between the indicated buffers, but a tendency towards higher yields with citrate buffers.

J. Fungi 2021, 7, x FOR PEER REVIEW 6 of 16
Optimizing the Osmotic Stabilizer 0.02M citrate, pH 5.8 [28] 0.4 M sucrose [20] no protoplasts 1.2 M MgSO4 [36] 5.52 ± 1.22 1.0 M sorbitol [28] no protoplasts Optimizing the buffer composition 0.1M citrate, 0.01 M EDTA, pH 5.8 [23] 1.2 M MgSO4 [36] 7.46 ± 2.02 0.02M citrate, pH 5.8 [28] 7.49 ± 1.51 0.02M MES, pH 5.8 [21] 5.28 ± 0.66 0.01 M phosphate, pH 5.8 [37] 5.01 ± 0.51 An advantage of MgSO4 as osmotic stabilizer is that the majority of intact protoplasts in the presence of MgSO4 have large vacuoles [38], which enables collection and purification, and floating in a trapping buffer [36]. Intact protoplasts accumulate in a sharp band at the interphase, and debris pellet at the bottom ( Figure 2). After washing with STC buffer, up to 10 8 protoplasts/g fresh weight can be recovered. Upon determination of the optimal buffer and osmotic stabilizer, we reevaluated the composition and concentration of the enzyme mix ( Figure 3). As the pilot study had indicated, for efficient degradation of the T. thlaspeos cell wall, the combined activity of Yatalse and Glucanex is necessary. As expected, individually the enzymes are poorly active, leaving filaments behind ( Figure 3A). This emphasizes the importance of testing various lysing enzymes, alone and in combination, in different buffers to find a mix suitable for the organism of choice and its individual cell wall composition [30]. In contrast to other applications of protoplasts, for genetic manipulation, the protoplasts have to be viable and able to regenerate their cell wall. Thus, we next investigated the influence of the osmotic stabilizer on regeneration of the protoplasts after transformation. This allowed us to separate optimization of both steps. Using protoplasts generated in the presence of MgSO 4 , we assessed regeneration media containing different osmotic stabilizers such as sucrose, glucose, sorbitol, and KCl ( Figure 4). MgSO 4 was excluded due to incompatibility with the gelling agent. Maximal regeneration was obtained with 1 M sucrose as osmotic stabilizer, followed by sorbitol and glucose, while 1 M KCl completely inhibited regeneration and fungal growth ( Figure 4A). In support, T. thlaspeos LF1 cell cultures do not grow on 1 M KCl, indicating that it is toxic at this concentration, while sucrose, glucose, and sorbitol only reduce the growth rate ( Figure 4B). Without an osmotic stabilizer, cells were not able to regenerate, indicating that any residual filaments were efficiently removed during purification of the protoplasts ( Figure 4A). U. maydis protoplasts regenerate into yeast cells and form colonies within three days [17,41]. For T. thlaspeos, we expected filaments to emerge from the protoplast, since we had never observed yeast-cells for this fungus. Furthermore, the slow growth rate of T. thlaspeos cultures suggests a longer regeneration time. To confirm our expectation, we described the regeneration process and its timing for T. thlaspeos protoplasts on regeneration medium with 1 M sucrose. After one day, protoplasts turned dark, which is indicative of cell wall regeneration. Three to eight days later, a filament emerges from the protoplast that starts branching after 7-13 days, finally resulting in a micro-colony after 11-18 days ( Figure 4C). Further proliferation leads to filamentous colonies after four to five weeks, which are indistinguishable from the original culture. Finally, we aimed to decrease the enzyme concentrations to save costs. However, lowering the concentration to half resulted in incomplete digestion of the fungal cell wall after 30 min ( Figure 3B). Since it was described earlier that a shorter incubation time is preferable, compared to a low enzyme concentration, regarding protoplast viability [39,40], we did not reduce the enzyme concentrations.
In contrast to other applications of protoplasts, for genetic manipulation, the protoplasts have to be viable and able to regenerate their cell wall. Thus, we next investigated the influence of the osmotic stabilizer on regeneration of the protoplasts after transformation. This allowed us to separate optimization of both steps. Using protoplasts generated in the presence of MgSO4, we assessed regeneration media containing different osmotic stabilizers such as sucrose, glucose, sorbitol, and KCl ( Figure 4). MgSO4 was excluded due to incompatibility with the gelling agent. Maximal regeneration was obtained with 1 M sucrose as osmotic stabilizer, followed by sorbitol and glucose, while 1 M KCl completely inhibited regeneration and fungal growth ( Figure 4A). In support, T. thlaspeos LF1 cell cultures do not grow on 1 M KCl, indicating that it is toxic at this concentration, while sucrose, glucose, and sorbitol only reduce the growth rate ( Figure  4B). Without an osmotic stabilizer, cells were not able to regenerate, indicating that any residual filaments were efficiently removed during purification of the protoplasts ( Figure  4A). U. maydis protoplasts regenerate into yeast cells and form colonies within three days [17,41]. For T. thlaspeos, we expected filaments to emerge from the protoplast, since we had never observed yeast-cells for this fungus. Furthermore, the slow growth rate of T. thlaspeos cultures suggests a longer regeneration time. To confirm our expectation, we described the regeneration process and its timing for T. thlaspeos protoplasts on regeneration medium with 1 M sucrose. After one day, protoplasts turned dark, which is indicative of cell wall regeneration. Three to eight days later, a filament emerges from the

Transformation
Five antibiotic resistance markers directed against phleomycin, hygromycin, nourseothricin, geneticin, and carboxin are routinely used in U. maydis [3]. To develop markers for T. thlaspeos, we tested culture growth on four of these antibiotics. Phleomycin is a mutagen and therefore was not considered [42]. T. thlaspeos cells were efficiently killed by the four antibiotics ( Figure S2). Concentration gradients with hygromycin, nourseothricin, and carboxin revealed that T. thlaspeos was more sensitive towards these antibiotics than U. maydis. 10 µg/ hygromycin mL and 50 µg/mL nourseothricin efficiently killed T. thlaspeos hyphae. This is 20 times and three time less than the standard concentration used for U. maydis, respectively. By contrast, cells are less sensitive towards carboxin and remained resistant at 2 µg/mL, the standard concentration used for U. maydis, but were sensitive at 100 µg/mL. (Figure S2). Carboxin inhibits the mitochondrial succinate dehydrogenase (SDH2), and a point mutation, H253L, leads to a resistant form in U. maydis [43]. The T. thlaspeos SDH2 was highly conserved, with 82% amino acid similarity, and contained an arginine instead of the histidine at this position ( Figure S3). This might explain the reduced sensitivity.
Due to the high hygromycin sensitivity, the bacterial hygromycin-phospho-transferase (hpt) [17] was used as the first resistance marker. We expressed it as a fusion protein hpt-eGfp [44] under the control of T. thlaspeos and U. maydis promoters and terminators (Table S2). Promoter activity was verified in U. maydis. Both P Tthsp70 and P Ttrps27 were active in U. maydis, and all five constructs resulted in hygromycin-resistant transformants with eGfp-fluorescence ( Figure 5 and Figure S4), suggesting the constructs are functional and can be used to transform T. thlaspeos.

Transformation
Five antibiotic resistance markers directed against phleomycin, hygromycin, nourseothricin, geneticin, and carboxin are routinely used in U. maydis [3]. To develop markers for T. thlaspeos, we tested culture growth on four of these antibiotics. Phleomycin is a mutagen and therefore was not considered [42]. T. thlaspeos cells were efficiently killed by the four antibiotics ( Figure S2). Concentration gradients with hygromycin, nourseothricin, and carboxin revealed that T. thlaspeos was more sensitive towards these antibiotics than U. maydis. 10 µg/ hygromycin mL and 50 µg/mL nourseothricin efficiently killed T. thlaspeos hyphae. This is 20 times and three time less than the standard concentration used for U. maydis, respectively. By contrast, cells are less sensitive towards carboxin and remained resistant at 2 µg/mL, the standard concentration used for U. maydis, but were sensitive at 100 µg/mL. (Figure S2). Carboxin inhibits the mitochondrial similarity, and contained an arginine instead of the histidine at this position ( Figure S3). This might explain the reduced sensitivity.
Due to the high hygromycin sensitivity, the bacterial hygromycin-phosphotransferase (hpt) [17] was used as the first resistance marker. We expressed it as a fusion protein hpt-eGfp [44] under the control of T. thlaspeos and U. maydis promoters and terminators (Table S2). Promoter activity was verified in U. maydis. Both PTthsp70 and PTtrps27 were active in U. maydis, and all five constructs resulted in hygromycin-resistant transformants with eGfp-fluorescence ( Figure 5 and Figure S4), suggesting the constructs are functional and can be used to transform T. thlaspeos. Figure 5. Verification of resistance-reporter constructs in U. maydis. Reporter constructs containing a fusion of hygromycinphospho-transferase gene (hpt) and the fluorescent marker (egfp or mcherry) under the control of hsp70 promoter and terminator regions derived from the T. thlaspeos genome were tested in U. maydis. Upon transformation of the linearized construct, it randomly integrates into the genome. Protein accumulation was visualized by the green/red fluorescence. The eGfp expression under the promoter region of T. thlaspeos was stronger than compared to the stably integrated construct under the control of a strong, synthetic promoter (Potef). This confirms that the fusion protein is active. In Figure 5. Verification of resistance-reporter constructs in U. maydis. Reporter constructs containing a fusion of hygromycinphospho-transferase gene (hpt) and the fluorescent marker (egfp or mcherry) under the control of hsp70 promoter and terminator regions derived from the T. thlaspeos genome were tested in U. maydis. Upon transformation of the linearized construct, it randomly integrates into the genome. Protein accumulation was visualized by the green/red fluorescence. The eGfp expression under the promoter region of T. thlaspeos was stronger than compared to the stably integrated construct under the control of a strong, synthetic promoter (Potef ). This confirms that the fusion protein is active. In comparison, mcherry-fluorescence in the strain carrying the Tthsp70 promoter was weaker than the stably integrated construct under the control of the Potef promoter. Scale bar: 10 µm.
First, we needed to define which plasmid to use. Therefore, we transformed an equimolar mixture of five hpt-eGfp plasmids with different promoters (Table S2) into T. thlaspeos LF1 protoplasts generated with the optimized method using the standard U. maydis conditions for transformation [28]. This resulted in a single transformant which had stably integrated the P Tthsp70 ::hpt-egfp:T Tthsp70 into the genome ( Figure S5, Figure 6). Now, transformations using this plasmid regularly result in fluorescent transformants. Based on this successful transformation, a P Tthsp70 ::hpt-mcherry:T Tthsp70 construct was generated (Table S2), tested in U. maydis (Figure 5), and transformed into T. thlaspeos LF2 (Figure 6), showing that compatible mating partners of T. thlaspeos can be tagged with different reporters to follow the mating process. had stably integrated the PTthsp70::hpt-egfp:TTthsp70 into the genome ( Figure S5, Figure 6). Now, transformations using this plasmid regularly result in fluorescent transformants. Based on this successful transformation, a PTthsp70::hpt-mcherry:TTthsp70 construct was generated (Table S2), tested in U. maydis (Figure 5), and transformed into T. thlaspeos LF2 (Figure 6), showing that compatible mating partners of T. thlaspeos can be tagged with different reporters to follow the mating process. One key aspect for gene targeted manipulations is efficient homologous recombination. To test whether T. thlaspeos reaches the same high rates of up to 50% as U. maydis, we targeted the pheromone receptor gene pra1 in the T. thlaspeos LF1 background for deletion. The construct design was based on U. maydis with 1 kb flanking sequences [27]. Transformation of the construct resulted in 122 candidates on the transformation plates. Reselection of 19 candidates on fresh hygromycin plates led to only nine candidates that remained resistant. The other candidates were either false positives, or they only transiently expressed the resistance protein. These are not interesting for stable integration. In subsequent analysis of the nine candidates, successful deletion of the pra1 locus was confirmed for two transformants (Figure S5), giving a homologous recombination rate of 22%.
In summary, we have now adapted the protoplast-mediated transformation for the filamentously growing Brassicaceae smut fungus T. thlaspeos. Together, with its ability for efficient homologous recombination, this gives us a tool to study plant-microbe interactions of smut fungi in the model plant A. thaliana with two genetically tractable partners.

Mating of Filaments
When T. thlaspeos teliospores germinate, they give rise to an infectious filament that can directly penetrate the plant. On the other hand, these filaments also can give rise to One key aspect for gene targeted manipulations is efficient homologous recombination. To test whether T. thlaspeos reaches the same high rates of up to 50% as U. maydis, we targeted the pheromone receptor gene pra1 in the T. thlaspeos LF1 background for deletion. The construct design was based on U. maydis with 1 kb flanking sequences [27]. Transformation of the construct resulted in 122 candidates on the transformation plates. Reselection of 19 candidates on fresh hygromycin plates led to only nine candidates that remained resistant. The other candidates were either false positives, or they only transiently expressed the resistance protein. These are not interesting for stable integration. In subsequent analysis of the nine candidates, successful deletion of the pra1 locus was confirmed for two transformants (Figure S5), giving a homologous recombination rate of 22%.
In summary, we have now adapted the protoplast-mediated transformation for the filamentously growing Brassicaceae smut fungus T. thlaspeos. Together, with its ability for efficient homologous recombination, this gives us a tool to study plant-microbe interactions of smut fungi in the model plant A. thaliana with two genetically tractable partners.

Mating of Filaments
When T. thlaspeos teliospores germinate, they give rise to an infectious filament that can directly penetrate the plant. On the other hand, these filaments also can give rise to haploid culture. Our haploid cultures T. thlaspeos LF1 and LF2 have compatible mating types. They can fuse at the tip and form a new filament [15]. To visualize directional growth of compatible LF1 and LF2 hyphae towards each other during mating, we carried out confrontation experiments. In close proximity, LF1 and LF2 hyphae sense each other, and reorient their growth to meet ( Figure 7A). In some cases, some hyphae return their growth in direction towards the compatible filament after initial passage. Upon contact, they fuse and result in a new filament ( Figure 7A,B, Video S1). To prove that fused hyphae really share a common cytoplasm, mating was also observed in cocultivation experiments of compatible strains expressing eGfp and mCherry ( Figure 7C). After hyphal fusion, eGfp and mCherry fluorescence could be observed in one cytoplasmic segment indicative of plasmogamy. On the other hand, if the pheromone receptor Pra1 is deleted, hyphae of compatible strains grow directly past each other without hyphal fusion ( Figure 7A,B and Video S2). These findings confirm that the pheromone-receptor system in T. thlaspeos [15] is active and initiates mating. In the future, the generation of nuclei-reporter-strains with NLS-fusion-constructs will allow tracking of the nuclei and thereby the investigation of karyogamy during mating. and reorient their growth to meet ( Figure 7A). In some cases, some hyphae return their growth in direction towards the compatible filament after initial passage. Upon contact, they fuse and result in a new filament ( Figure 7A, B, Video S1). To prove that fused hyphae really share a common cytoplasm, mating was also observed in cocultivation experiments of compatible strains expressing eGfp and mCherry ( Figure 7C). After hyphal fusion, eGfp and mCherry fluorescence could be observed in one cytoplasmic segment indicative of plasmogamy. On the other hand, if the pheromone receptor Pra1 is deleted, hyphae of compatible strains grow directly past each other without hyphal fusion ( Figure 7A,B and Video S2). These findings confirm that the pheromone-receptor system in T. thlaspeos [15] is active and initiates mating. In the future, the generation of nuclei-reporter-strains with NLS-fusion-constructs will allow tracking of the nuclei and thereby the investigation of karyogamy during mating.

Discussion
When we first set out to work with T. thlaspeos, our aim was to establish a genetically tractable smut fungus in a model host plant such as A. thaliana [15]. An important aim for reaching this goal was genetic manipulation. Here, we show that like other smut fungi, T. thlaspeos is amenable to protoplast-mediated transformation. We have generated a hygromycin resistance cassette, where expression of the hygromycin-phospho-transferase, hpt, is controlled by the T. thlaspeos hsp70 promoter sequence, similar to the cassettes used in U. maydis [3]. Interestingly, promoter sequences seem to be exchangeable between smut fungi, since the T. thlaspeos promoters were active in U. maydis and several groups have successfully used U. maydis constructs [23,45]. This now enables us to generate reporter strains for a broad range of scientific questions.
Most important to establishing a successful protoplasting protocol is the choice of the lytic enzyme(s). The fungal cell wall is a multilayered, chemically complex structure consisting mainly of polysaccharides and varying amounts of lipids, proteins, and polyphosphates [46]. Its composition is not only variable between species [30], but also highly dependent on the culture conditions [47] and morphology [48]. In our case, the combined activity of Yatalase and Glucanex was necessary for efficient digestion of the T. thlaspeos cell wall, although they appear to have overlapping enzymatic properties. Similar additive effects have recently been shown for the ascomycete Hirsutella sinensis [49] and Ag. aegerita [31]; while in Cordyceps militaris, the enzymes mix is less active than Glucanex alone [50]. Hence, during the establishment of conditions for protoplasting, various lytic enzymes and combinations should be tested to reach optimal cell wall degradation [30]. Moreover, commercial manufacturing of enzymes can be stopped, with the broadly used Novozyme 234 being a recent example. Hence, the identification of suitable enzymes can be a reoccurring problem even for established protocols.
The second important factor is the osmotic stabilizer, because it depends on the choice of the protoplasting enzyme. For example, the enzymatic activity of Yatalase is inhibited by sorbitol and sucrose. In the 1970s, similar observations were made for helicase [35] and snail enzyme [34]. Protoplasting protocols with Yatalase use inorganic salts as osmotic stabilizer [21,31,37,40,51] and similar to these old reports, for T. thlaspeos, we now use MgSO 4 to enable cell wall degradation.
Together with other factors influencing the protoplast formation, such as growth conditions of the culture, buffer composition, pH, temperature, or protoplasting time, establishing new transformation protocols quickly becomes a multi-factorial challenge, and testing full-factorial replicates is time-consuming and costly. For T. thlaspeos, we designed pilot studies covering selected combinations in single replicates based on existing transformation protocols, and used the most promising buffer, osmotic stabilizer, and enzyme combination for further optimization. While this approach does not cover all combinations, it allowed us to establish a good transformation protocol with reasonable effort.
As a proof-of-principle for our transformation protocol, we investigated the wellcharacterized smut fungal mating process in T. thlaspeos. In the first step, we generated reporter strains expressing cytosolic eGfp or mCherry to visualize the fusion of hyphae. The resulting filaments express both eGfp and mCherry, indicative of a common cytoplasm, as typical for the dikaryotic smut fungi [52,53]. Next, we looked into dependency on the pheromone receptor. To this end, we generated a deletion mutant of the pheromone receptor pra1 [15] based on the strategy of U. maydis [28]. Notably, homologous recombination also takes place in T. thlaspeos, so we can modify genes in the haploid culture background easily.
T. thlaspeos pra1 deletion mutants cannot mate anymore. This finding is especially interesting since it is not yet known whether mating is required for T. thlaspeos to fulfil its life cycle. Infectious filaments emerge directly from germinating T. thlaspeos teliospores. By contrast, teliospore germination of grass smut fungi gives rise to yeast-like sporidia. Subsequent pathogenic development depends on the morphological switch from yeast to filamentous growth brought about by mating [52]. However, the functional conservation of mating genes in T. thlaspeos suggests an evolutionary-conserved, and therefore important, role of mating also in this fungus [15]. This raises several questions. Is mating necessary for the lifecycle of T. thlaspeos? Where and when does mating occur? When do T. thlaspeos hyphae undergo meiosis? Is the filament emerging from the teliospore diploid or dikaryotic? Is the transition to haploid hyphae also occurring naturally in this state of the lifecycle? With the established transformation protocol, we will be able to further address these questions. This will shed light not only onto the mating process of T. thlaspeos, but also on the role of RNA communication in virulence, perennial persistence of the fungus in planta, and nutrition of a smut fungus during biotrophic growth.

Conclusions
Establishing the genetic manipulation of the Brassicaceae smut fungus T. thlaspeos now allows us to generate reporter strains as well as targeted deletions or modifications of fungal genes. Combined with the fungal colonization of the model plant A. thaliana, we thereby provide a pathosystem, in which both partners have a small, genetically tractable genome for addressing the current and future questions of plant-microbe interactions.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available in this manuscript and constructs can be requested from the corresponding author.