Immunogold labeling of extracellular vesicles from the fungal pathogen Fusarium graminearum reveals the presence of the protein marker Sur7

Abstract

Fusarium head blight, caused by Fusarium graminearum, is one of the most threatening fungal diseases of cereals worldwide. Current practices for control of F. graminearum are not always efficient, as epidemics still occur and there is low resistance in wheat varieties. Therefore, novel antifungal targets must be discovered by analyzing the molecular interaction between F. graminearum and its host. Fungal extracellular vesicles (EVs) are small membrane-bound compartments (30–1000 nm) that carry macromolecules and support fungal virulence, hence the disruption of EV production could lead to reduced fungal pathogenicity. However, EV study is limited by the lack of surface protein markers to aid in their characterization. Therefore, the aim of this report was to target a surface protein marker with an antibody, to unlock advanced EV characterization techniques. Using the list of potential EV markers for Candida albicans, we selected the tetraspanin-like Sur7 to perform immunogold microscopy, revealing that this protein is a surface marker of F. graminearum EVs. SUR7 is present on the surface of some but not all vesicles. EVs carrying SUR7 were larger than those without the marker, suggesting that there are subtypes of fungal EVs. The epitope recognized by the anti-Sur7 antibody is conserved in other Fusarium pathogens, making Sur7 a potential pan-Fusarium EV marker. Our results unlock techniques, such as immunoaffinity chromatography and antibody labeling, to track fungal EVs and understand their biogenesis, which may lead to the development of novel antifungals.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s40694-025-00206-8.

Background

Head blight of cereals is a devastating fungal disease caused by some species of the genus Fusarium. This disease causes billion-dollar losses during epidemics, low crop yields, and contamination of grains with mycotoxins [1]. Almost 300,000 metric tons of wheat were lost in the early century in the US [2], while 17 billion dollars were lost between 1993 and 2014 [3]. Around 20 species cause head blight, including F. avenaceum, F. culmorum, F. poae, and F. graminearum. These vary in incidence and abundance, but F. graminearum is considered the most damaging species [3]. This pathogen produces mycotoxins such as deoxynivalenol and zearalenone that harm humans and farm animals [4]. Toxin variants are being reported depending on the geographical location of the fungus [5, 6]. Integrated pest management combined with fungicide applications can control F. graminearum, but this is not effective with high humidity and moderate temperatures [7]. Therefore, an improved understanding of the Fusarium-host interaction is urgently needed to find new antifungal targets.

Fungal extracellular vesicles (EVs) mediate the interaction between pathogenic fungi and their hosts. These nano-sized vesicles (30–1000 nm in diameter) carry proteins, lipids, nucleic acids, metabolites, and induce changes in the recipient cell. At the time of this report, EVs from more than 50 species have been reported. The main genera include Aspergillus, Cryptococcus, Candida, Fusarium, Malassezia, Paracoccidioides, and Colletotrichum. Important plant pathogens such as Botrytis cinerea, Ustilago maydis, and Zymoseptoria tritici produce EVs [8–10], with comparable morphology and protein cargo. EVs are involved in cell wall remodeling, fungal adaptation, and phytotoxicity [11]. F. graminearum, B. cinerea, and Colletotrichum higginsianum produce EVs that carry effectors, cell wall-degrading enzymes, and RNA that alters the gene expression of the host [12–14]. Therefore, identifying the biosynthetic routes of EVs and the EV cargo that is linked to virulence is a huge opportunity for antifungal discovery. For instance, small molecules that disrupt the production of EVs may lead to decreased fungal virulence, and the pool of macromolecules transported by EVs may contain molecules with previously unknown roles in pathogenesis. However, the current tools available to study fungal EVs are substantially limited, leading to important gaps in this research field.

The main questions in the study of fungal EVs involve their biogenesis, movement across the fungal cell wall, EV subtype production, and cargo delivery into the host. These gaps are mostly underpinned by the lack of protein surface makers for fungal EVs. For instance, with the aid of mammalian EV markers such as CD9, CD63, Alix, or TSG101, we know that the biogenesis of mammalian EVs is regulated by secretory proteins, scramblases, and flippases [15–17]. However, there is very little knowledge on the biosynthetic routes of fungal EVs. Similarly, while cell wall polysaccharides influence EV production [18], EVs have not been imaged traversing the host-pathogen interface, making it unclear how fungal EVs reach the host and deliver their cargo. Finally, the different composition of EVs from Cryptococcus spp. and Exophiala spp. [19, 20] suggests that there are EV sub-types with different functions, but this has not been explored in other fungal species. Due to the large genetic variability of the fungi kingdom, it is likely that the initial discovery of protein markers will be species-specific.

Although tetraspanins and ESCRT proteins are excellent markers of mammalian EVs, these are absent or in low abundance in fungal EVs [21]. While a pan-fungal marker has not been reported, species-specific protein markers are now available and will improve the understanding of these vesicles. Markers with confirmation at the Western blot level include the tetraspanin PLS1 from Botrytis cinerea [13], three proteins from Colletotrichum higginsianum: ChSso2, ChSnc, ChBmh1 [14], and RsTsp2 from Rhizoctonia solani [22]. In C. albicans, a list of 22 putative EV markers is available [23], and one promising marker is the tetraspanin-like Sur7. This protein influences EV morphology, cargo, and fungal virulence, and was enriched in F. graminearum EVs [12, 23, 24]. Still, no fungal EV marker has been used yet to investigate EV subtypes. Achieving EV recognition using antibodies will not only enhance the characterization techniques available for fungal EVs but will also improve the understanding of fungal-host interactions.

In this report, we used previously published proteomics data from F. graminearum [12] with the aim of expanding the availability of surface markers for fungal EVs. We identified four potential markers with multiple extracellular loops that were suitable for antibody recognition in silico: 1,3-β-glucanosyltransferase (Uniprot ID: I1RZX4), Yop1 (I1RTB8), Fet3 (I1RMG9), and Sur7 (I1RWL7). We obtained a polyclonal antibody that targeted the large extracellular loop of Sur7, revealing that the antibody reacted with EV-enriched samples via Western blot. Immunogold electron microscopy revealed Sur7 on the surface of F. graminearum EVs, confirming that this protein is a marker. Sequence analysis suggests that Sur7 could be a marker for other Fusarium species. The ability to track fungal EVs using antibodies unlocks several techniques to accelerate their characterization and understand their function.

Methods

Fungal cultures

F. graminearum isolate PH-1 was a gift from Dr. Kim Hammond-Kosack (Rothamsted Research, Harpenden, Herts., UK). For spore production, a plate of 7-day old mycelium grown on potato dextrose agar (Sigma) was used to inoculate 500 mL of Czapek Dox medium (BD Difco) at 25 °C in agitation, with a photoperiod of 12 h for light and darkness. After 48 h, the mycelium was removed using Miracloth (Merck) and discarded. Spores were harvested by centrifugation at 4000×g and 4 °C for 4 min and resuspended in ultrapure water. Using a hemocytometer, spores were adjusted to a concentration of 5 × 10⁶ spores/mL in water and preserved in 20% glycerol at − 80 °C until use.

For EV production, a medium named “YNB+” (6.7 g/L yeast nitrogen base with (NH₄)₂SO₄, no amino acids, no carbohydrates, US Biological Life Sciences), with added -Leu dropout supplement (0.69 g/L, Takara), L-leucine (0.076 g/L, Sigma), and L-glutamic acid (0.5 g/L, Sigma). The medium was filtered using 0.22 μm cellulose membranes (HAWP, Merck) prior to inoculation. Five hundred mL of YNB + were inoculated with 10⁴ spores/mL, followed by incubation for 5 days at 25 °C and 50 rpm agitation in a 2 L plastic baffled flask (Corning). Next, the mycelium was removed using sterile Miracloth (Merck). The culture was filtered using Whatman filter paper #1, followed by 0.45 μm cellulose membranes (HAWP, Merck). The remaining filtered culture was concentrated down to 50 mL using a Vivaflow Hydrosart 50R tangential flow filtration cassette with a MWCO of 100 KDa (Sartorius). This sample was concentrated again down to 500 µL using a Vivaflow 20 with a MWCO of 300 kDa (Sartorius). A cell lysate was prepared for each fungal culture as described before [12]. Samples were kept on ice for all steps or preserved at − 80 °C until used. S. cerevisiae BY4741 and C. albicans ATCC90028 were maintained on YPD agar.

EV separation

EVs were separated by size-exclusion chromatography (SEC) as described previously with modifications [25]. Briefly, the concentrated supernatant was loaded onto a 20 mL gel filtration column (Takara) containing 10 mL of Sepharose CL 2B (GE) equilibrated with Dulbecco’s phosphate buffered saline (DPBS, Thermo Fisher). At least 24 fractions (approximately 300 µL each) were eluted with DPBS and collected in a clear microtiter plate with flat wells (Greiner BioOne). The protein concentration of the samples was determined with a Qubit3 (Thermo Fisher) or with a NanoDrop Eight (Thermo Fisher). All samples were frozen in liquid nitrogen and preserved at − 80 °C until further use. For EV separation, S. cerevisiae and C. albicans were grown as described in [18] and [23], respectively. EVs from these organisms were separated using the size-exclusion chromatography method described above.

Identification of Fusarium graminearum orthologs of protein markers

We retrieved the sequences of 22 candidate protein markers for EVs from C. albicans [23] from the Candida genome database [26]. The F. graminearum EV proteome from Garcia-Ceron and colleagues [12] was scanned for the 22 sequences using NCBI’s BLASTp, with default settings. The log2-fold change (log2FC) from the previous analysis of F. graminearum EVs, which compares the protein abundance between the F. graminearum EVs and the cell lysate [12], was also retrieved to determine if the putative markers were enriched in F. graminearum EVs.

Protein predictions and generation of polyclonal antibodies

Whole protein sequences were analyzed using TOPCONS2 [27], and Protter 1.0 [28] to predict their membrane topology and potential extracellular regions, using default settings. Bepipred 3.0 [29] was used to detect amino acid epitopes that were potentially antigenic using default settings. The epitope sequence in the larger extracellular loop of the F. graminearum Sur7 sequence (Uniprot ID: I1RWL7) was provided to Mimotopes Australia Pty Ltd to manufacture a synthetic peptide linked to keyhole-limpet hemocyanin (KLH). The peptide was injected into two rabbits, followed by purification of an anti-Fusarium graminearum Sur7 (anti-FgSur7) polyclonal antibody. The full protocol was performed by Mimotopes Australia. The antibodies were obtained freeze-dried, followed by resuspension to a final concentration of 1 mg/mL using ultrapure sterile water. The solution was centrifuged briefly at 4000×g, aliquoted, and stored at − 20 °C until used.

Proteinase K treatment of EV samples

Proteinase K (produced in-house recombinantly) was added at a concentration of 0.5 mg/mL to 15 ug of EVs without additional detergents. The sample was incubated at 37 °C for 45 min, followed by inactivation at 95 °C for 10 min.

SDS-PAGE and Western blot

Twenty-five µL of EVs or cell lysate (25 µg of protein per gel lane) was mixed with 5 µL of NuPAGE LDS sample buffer (Thermo Fisher) and with 1 µL bond breaker (Thermo Fisher). Size-exclusion chromatography fractions (300 µL) were combined in pairs and concentrated down to 25 µL (≈ 2.5 µg of protein per lane for fractions 8–17). Samples were denatured at 95 °C for 5 min and then resolved using 4–12% Bis-Tris polyacrylamide gels (Thermo Fisher), using MES running buffer (Thermo Fisher) with 500 µL of NuPAGE antioxidant (Thermo Fisher). Five µL of size marker PageRuler or SeeBlue Plus (Thermo Fisher) were included for reference. Samples were resolved for 15 min running at 180 volts. Gels were then rinsed with reverse osmosis (RO) water and transferred onto a mini PVDF transfer pack (Bio-Rad, CA, USA) using a Trans-blot Turbo transfer system (Bio-Rad) at 25 volts for 7 min. The membrane was blocked with 3% BSA + 0.1% Tween 20 for 1 h at RT, and washed three times in TBS buffer (8 g/L NaCl, 20 mL of 1 M Tris HCl/L, pH 7.6, 0.1%Tween) for 10 min. The membrane was probed with anti-FgSur7 antibody diluted 1:1000 in TBS for 1 h at RT, washed, and probed with a goat anti-rabbit HRP-IgG secondary antibody (1/50,000; Pierce) for 1 h at RT. The membrane was revealed by adding 7 mL of ECL Western blotting detection solution (Promega). Images were taken using a ChemiDoc (BioRad).

Immunogold labeling of EVs

Five µL of EVs at a protein concentration of 0.05 µg/µL were placed on carbon film coated copper grids (400 mesh, ProSciTech, Australia) and incubated for 20 min. All steps were performed at RT. The excess EV buffer was removed using filter paper. All the following steps were performed by inverting the grid, EV side down, on top of a liquid droplet and transporting the grids along a series of droplet solutions using tweezers. The grids containing EVs were blocked with 1% BSA (Sigma) in DPBS for 30 min at room temperature. Excess solution was immediately absorbed using filter paper before the grid was moved to the next droplet. After blocking, the grids were incubated in primary antibody (anti-FgSur7) in DPBS (1:100) for 1 h, followed by three DPBS washes of 5 min each. The grids were then incubated in DPBS for 30 min, followed by 1 h in secondary antibody (goat anti-rabbit IgG conjugated to 5 nm colloidal gold; Sigma) in DPBS (1:20). Grids were washed three times in DPBS, followed by uranyl acetate staining as reported previously [12]. Samples were imaged using a Jeol JEM-2100 electron microscope under an accelerating voltage of 80 kV and images captured on a Gatan Orius SC 200 CCD camera (Scitek, Australia).

Image and statistical analyses

Electron microscopy micrographs were processed using Fiji (ImageJ 1.54; downloaded July 2025) to measure particle size. Data for particle sizes were analyzed using the Shapiro-Wilk test for normality, and Levene’s test for homogeneity of variance. The Wilcoxon test was used for pair-wise comparison of particle size between the Sur7(+) and Sur7(−) classes. Kernel density estimation (KDE) was employed to calculate the particle distribution density between EV classes. Data and plots were processed using RStudio 2025.051 + 513. Micrographs were cropped and the scale bar was re-drawn. All images were created using Adobe Illustrator (version 29.61). No image enhancements were performed.

Sequence analysis and phylogenetic trees

Putative Sur7 ortholog sequences were retrieved from Uniprot and aligned using its built-in tool. Phylogenetic trees were generated with Uniprot and drawn using Adobe Illustrator.

Results

EVs from F. graminearum contain orthologs of putative C. albicans EV markers

The study of C. albicans EVs yielded a list of 22 potential markers for fungal EVs [23]. Thirteen of these proteins were identified in the F. graminearum EV proteome from Garcia-Ceron and colleagues [12], indicating that they are potential orthologs (Table 1). The level of the putative protein markers in F. graminearum EVs relative to the cell lysate (previously published in [12]) revealed that three proteins were exclusively in the F. graminearum EVs, six were enriched in EVs, and four had no difference between EVs and cell lysate (Table 1). Proteins that were exclusive or enriched in F. graminearum EVs, and had transmembrane domains were selected for further analysis. They were Fet3, 1,3-β-glucanosyltransferase, Sur7, and Yop1.

Table 1.

Putative orthologs of C. albicans EV markers present in F. graminearum EVs

C. albicans marker	F. graminearum putative ortholog	E-value	Sequence identity	Gene name	Enrichment (log2FC)
Cdc42	CDC42 homolog	3 × 10–112	79.4%	FGSG_05447	EV exclusive
Rho3	RHO3-like	9 × 10− 198	66.8%	FGSG_00170	EV exclusive
Ykt6	YKT6-like	9 × 10− 82	55.0%	FGSG_06394	EV exclusive
Fet34	Fet3	0.0	51.8%	FGSG_05159	EV enriched
Phr1	1,3-beta-glucanosyltransferase	3 × 10− 152	48.8%	FGSG_09980	EV enriched
Rac1	Uncharacterized protein	2 × 10− 88	65.5%	FGSG_08857	EV enriched
Sur7	Sur7-like	3 × 10− 20	32.3%	FGSG_08692	EV enriched
Yck2	Kinase domain-containing protein	1 × 10− 166	65.2%	FGSG_10066	EV enriched
orf19.2168.3	Yop1	2 × 10− 46	46.4%	FGSG_07419	EV enriched
Mts1	Sphingolipid C9-methyltransferase 2	0.0	63.3%	FGSG_05593	No difference
orf19.1054	Uncharacterized protein	2 × 10− 25	28.0%	FGSG_04983	No difference
Sec 4	GTP-binding protein YPT2-like	1 × 10− 90	61.9%	FGSG_06209	No difference
Vac8	Vacuolar protein 8	0.0	67.2%	FGSG_08997	No difference

Fusarium graminearum EV markers with predicted extracellular domains

The amino acid sequences of Fet3, 1,3-β-glucanosyltransferase, Sur7 and Yop1, were submitted to Protter 1.0 (Fig. 1A) and TOPCONS 2.0 (Fig. 1B) to predict their membrane topology. These tools returned inconsistent numbers of transmembrane domains and extracellular residues for Fet3 and Yop1, thus these two candidates were removed from further analysis. The predicted topology of 1,3-β-glucanosyltransferase and Sur7 were consistent between the two tools but given that Sur7 had already been identified as a potential marker of C. albicans EVs, and that Sur7 has similar topology to the mammalian tetraspanins, we selected Sur7 for polyclonal antibody production.

Fig. 1.

Selection of a putative Fusarium graminearum EV surface marker. A Proteins from F. graminearum EV proteome that are potential orthologs with C. albicans EV markers. The membrane topology of the marker candidates identified were obtained with (B) Protter and (C) TOPCONS [27, 28]. The prediction for FgSur7 was consistent between the two tools and thus prioritized for further analysis. C The amino acid sequence of FgSur7 was analyzed using Bepipred 3.0 [29] to predict epitopes for polyclonal antibody production. A peptide of 24 amino acids (DGARAAPVIGKAWDSNPRNAPSS) located in the large extracellular loop had an antigenicity score above the threshold of 0.2 and was selected for antibody production. D The epitope was located in the extracellular region of FgSur7, indicating potential surface availability (3D structure from AlphaFold; AF-I1RWL7-F1-v4)

FgSur7 contains epitopes with predicted antigenicity

The complete FgSur7 amino acid sequence was submitted to Bepipred 3.0 to identify antigenic regions (Fig. 1C). Extracellular amino acid regions with a score above 0.2 were considered for further analyses. A potential epitope was selected, and a cysteine was added to the N-terminus (CDGARAAPVIGKAWDSNPRNAPSS) for linkage to the carrier protein. The location of the potential epitope was highlighted in the predicted 3D structure of F. graminearum Sur7 (AlphaFold [30]) (Fig. 1D, highlighted in red) confirming potential surface availability.

Detection of FgSur7 in EVs by Western blot

The anti-FgSur7 antibody reacted with the EV sample, revealing a strong band below the 28 kDa mark, consistent with the theoretical molecular mass of Sur7 (27 kDa; Fig. 2A). The cell lysate band was below the level of detection of the Western blot. However, previous data from our group confirmed that Sur7 is present in the F. graminearum cell lysate, and that compared to this sample, Sur7 is enriched in EVs [12]. Size-exclusion chromatography fractions revealed a similar signal between the 15–25 kDa mark in fractions 8–13 (Fig. 2B). EVs and cell lysate both contained a non-specific band of ~ 50 kDa that was also observed using both the pre- and post-immunization sera (Figure S1). This was deemed non-specific binding. The purified epitope peptide (~ 2600 Da) without carrier protein was included as positive control (Figure S2). When the EV aliquot was pretreated with proteinase K, the FgSur7 signal at about 28 kDa was not detected (Fig. 2A). To check for cross-reactivity, EVs from S. cerevisiae and C. albicans were also analyzed by Western blot. These EVs did not react with the anti-FgSur7 antibody (Fig. 2A).

Fig. 2.

Identification of Sur7 on Fusarium graminearum (Fg) EVs. A The Fg EV sample (25 µg of protein) revealed the presence of a band below the 28 kDa mark (labeled “Sur7”), similar to its theoretical MW (27 kDa). The same bands was not detected in the cell lysate. Non-specific bands (labeled “NS”, Figure S1) from the polyclonal anti-FgSur7 antibody were present near the 50–70 kDa marks. Fg EVs (15 µg of protein) were treated with proteinase K resulting in the degradation of the signal. Non-specific bands correspond to the approximate MW of proteinase K. EVs from S. cerevisiae (Sc) and C. albicans (Ca) did not cross-react with the anti-FgSur7 antibody. Blots are representative of three biological replicates (Figure S2). B Sur7-positive EVs eluted mostly in the size exclusion chromatography fractions 8–13, which coincided with the elution of EVs as reported previously [12]. Protein per lane: ≈2.5 µg. C, D Immunogold labels (yellow pointers) concentrated in some particles (Sur7(+)). E Not all captured particles were labeled (Sur7(−)). (F) EVs stained with uranyl acetate and without antibodies revealed typical cup-shape morphology. G The average diameter of Sur7(+) vesicles (orange dots) was significantly larger than the Sur7(−) particles (blue dots) (p = 0.0039, Wilcoxon test; error bars = SEM). H The average, median, and mode diameter of the two EV classes was compared using Kernel density estimation. I Representative low magnification micrograph of primary + secondary antibody labeling. Sur7(+) and Sur7(−) particles are shown in blue and orange, respectively. Other controls revealed no unspecific labeling: no primary antibody, 1% BSA, blocked primary antibody, and pre-immunization rabbit serum. Gold nanoparticles did not concentrate away from vesicular structures or in the grid’s background (Figure S3). Scale bars: E, F, H: 200 nm; G: 100 nm

FgSur7 is present on the surface of extracellular vesicles

EVs were subjected to immunogold staining to determine whether EVs display FgSur7 on the membrane surface. Gold nanoparticles were observed in EV structures when vesicles were labeled using the anti-FgSur7 and secondary antibodies (Fig. 2C-D). Not all EVs were labeled (Fig. 2E). Micrographs of EVs stained only with uranyl acetate revealed typical structures (Fig. 2F). A Wilcoxon pair-wise test between the SUR7 positive (Sur7(+)) and Sur7 negative (Sur7(−)) vesicles revealed significant differences in their diameter (p = 0.0039; Fig. 2G; blue and orange dots, respectively). Sur7(+) vesicles had an average diameter of 209 nm (n = 40), while Sur7(-) had an average diameter of 138 nm (n = 16). The median and mode diameters of the two classes were visualized using the Kernel estimation density (Fig. 2H). Low magnification micrographs revealed the presence of labeled and unlabeled vesicular structures and were used to measure the diameters of EVs (Fig. 2I).

Other negative controls that included no primary antibody, blocked primary antibody (epitope incubated with primary antibody), 1% BSA, and the pre-immunization rabbit serum, revealed no significant non-specific gold labeling (Figure S3). Wide amplification micrographs confirmed that gold labeling was concentrated around the vesicles (“Background”, Figure S3).

Sur7 is likely a pan-Fusarium EV protein marker

The similarity of Sur7 between F. graminearum and C. albicans was 30.96%, and 25.42% when compared to Sur7 in S. cerevisiae (Clustal Omega). The anti-FgSur7 antibody did not react with EVs from C. albicans or S. cerevisiae on a Western blot, hence we searched for other organisms in which the Sur7 epitope is conserved. Aligning the extracellular loop in Sur7 in several Fusarium species revealed that the epitope is fully conserved, besides a single-base mutation (Fig. 3A). Furthermore, the epitope was at least 85% identical in 65 Fusarium species (Table S1). Fungi from the genre of Cylindrodendrum, Ilyonectria, Clonostachys, Neonectria, and Stachybotrys shared 70% identity with the Sur7 epitope (Table S2). The amino acid sequence of the extracellular loop of C. albicans Sur7 was also retrieved to analyze differences in the amino acid sequence in the region where the anti-FgSur7 epitope is located, revealing no significant sequence conservations between the two species (Fig. 3A). Potential Sur7 orthologs from Trichoderma harzianum, Magnaporthe grisea, Zymoseptoria tritici, Fusarium pseudograminearum, Candida albicans, Saccharomyces cerevisiae, Paracoccidioides lutzii, Botrytis cinerea, Fusarium oxysporum, Colletotrichum higginsianum, Fusarium oxysporum f. sp. cubense, and Fusarium graminearum were aligned and presented in a phylogenetic tree (Fig. 3B) revealing that Fusarium pathogens have greater sequence similarity compared to non-Fusarium fungi.

Fig. 3.

Sequence analysis of Sur7 in different fungal pathogens. A Multiple sequence alignment (Clustal Omega) of the large extracellular loop of F. graminearum Sur7 (Uniprot ID: I1RWL7). The amino acids in the large extracellular loop were identified using Protter [28]. The sequence was submitted to BLASTp using default settings, and the top Fusarium hits were aligned using Clustal Omega. The alignment revealed that the epitope used to create the anti-FgSur7 antibody (highlighted in red) is highly conserved in other Fusarium plant pathogens with minimal amino acid changes. The amino acid sequence of the extracellular loop of C. albicans and S. cerevisiae SUR7 was included in the alignment to understand the lack of cross-reactivity of the anti-FgSur7 antibody with C. albicans and S. cerevisiae EVs. This alignment revealed no significant sequence conservation. B Phylogenetic analysis of Sur7 orthologs in fungal pathogens. Sur7 potential orthologs were identified in major plant and human pathogens to assess sequence similarities. The tree revealed differences between Fusarium and non-Fusarium pathogens

Discussion

EVs enhance fungal virulence by transporting protein effectors, toxins, and other molecules that support infections. For instance, EVs from F. graminearum and Rhizoctonia solani transport protein effectors that cause damage to the host, Cryptococcus gattii EVs trigger a “division of labor” mechanism that enables fungal virulence of cells over long distances, and Botrytis cinerea might transport RNA in EVs to silence host defense genes [11, 13, 20, 22, 31]. These virulence-associated molecules represent excellent targets for the development of new antifungal molecules. Furthermore, disrupting EV production altogether could lead to attenuated fungal virulence, but this has not been attempted because the EV biosynthetic routes, as well as the subclasses of fungal EVs, are poorly understood. The first step towards a better understanding of EV biosynthesis and their role in fungal virulence is the establishment of tools that assist in the separation of different classes of EVs to help in dissecting their roles. This relies heavily on protein surface markers [32, 33]. Our group had previously identified the tetraspanin-like Sur7 protein as a potential marker of EVs in C. albicans [23]. This protein later revealed important roles in EV morphology and virulence [24], and was identified in EVs from other fungal species, including F. graminearum [9, 12, 20]. Therefore, the aim of this study was to generate a polyclonal antibody to target a suitable surface marker on EVs from F. graminearum. Using immunogold microscopy, we revealed that Sur7 is present in EVs from this fungal pathogen. This antibody did not cross-react with EVs from C. albicans or S. cerevisiae but was highly conserved in other Fusarium pathogens. Sur7-positive vesicles are larger than those that do not exhibit the protein and thus could have different biological roles.

Sur7 may be involved in the biosynthesis of distinct EV subclasses

Sur7 is a component of the “membrane compartment of Can1” (MCC) and the eisosome, which work in plasma membrane organization, cell wall integrity, and virulence [34–36]. Sur7 has been detected in EVs from other fungi [9, 16, 18], but its role in these particles has not been elucidated. The F. graminearum EV proteome from Garcia-Ceron and colleagues [12] did not contain all the MCC/eisosome proteins known in C. albicans, hence this compartment might work differently between yeast and filamentous fungi, or the sorting of the MCC/eisosome proteins into EVs is selective. In the filamentous fungus Beauveria bassiana, Sur7 is located on vacuoles and in plasma membranes, and might be involved in vacuolar protein sorting [37] suggesting roles beyond plasma membrane reorganization. A Sur7 deletion in C. albicans led to different EV proteome and attenuated virulence in a Galleria mellonella model, which was restored in the KO by addition of WT EVs. Also, Sur7 KOs produced a similar number of EVs but these were smaller compared to WT EVs [24]. Our results confirm this also for F. graminearum, since EVs without Sur7 were significantly smaller than those with SUR7. Differences in EV size based on protein marker presence have also been reported for mammalian EVs. For instance, medium EVs carry a higher proportion of markers CD9 and CD81 [38], while small EVs carry more CD63 [39]. This correlates with tetraspanin cellular location: CD9 and CD81 are located in the plasma membrane and are mostly associated with medium EVs, while CD63 is found in endosomes and often decorates small EVs [40, 41]. Since SUR7 is predicted to be a plasma membrane protein, it is reasonable to speculate that SUR7 is involved in the biosynthesis of medium fungal EVs that bud off directly from the plasma membrane, although the presence of Sur7 in EVs of other sizes cannot be ruled out.

Tetraspanins are EV markers across different fungal species

In addition to F. graminearum Sur7, the fungal EV markers confirmed to date, at least by Western blot, are: PLS1 from Botrytis cinerea [13], ChSso2, ChSnc1, and ChBmh1 from Colletotrichum higginsianum [14], and RsTsp2 from Rhizoctonia solani [22].

PLS1 is a gene that encodes a tetraspanin that is essential for penetration of the leaf epidermis by the rice blast fungus Magnaporthe grisea [42]. An ortholog of PLS1 (BcPLS1) was detected in EV samples from B. cinerea [13]. The PLS1-positive vesicles were essential for fungal small RNA delivery into Arabidopsis plants, while B. cinerea knockouts that did not produce BcPLS1 had decreased virulence [13]. Although BcPLS1 shares low amino acid similarity with F. graminearum SUR7 (19.5%, ClustalO), both proteins have similar 3D structures comprised of four transmembrane domains, and two extracellular loops. An ortholog of PLS1 has been identified in F. graminearum (63), but this protein was not identified in the EV proteome from Garcia-Ceron and colleagues [12]. Other pathogens in which a PLS1 ortholog has been identified include B. cinerea, M. grisea, and Colletotrichum lindemuthianum, all of which produce an appressorium [54]. Since F. graminearum does not produce an appressorium, it is possible to speculate that the lack of PLS1 in F. graminearum EVs could be explained by differences in the formation of infection structures.

ChSso2, ChSnc1, and ChBmh1 from C. higginsianum are two SNARE proteins and a 14-3-3 family protein, respectively, that were detected in the lumen of EVs from C. higginsianum [14]. ChSnc1 was enriched in EVs, while ChBmh1 was enriched in the cell lysate. We found putative orthologs of these proteins in the F. graminearum EV proteome with similar abundance levels: the ortholog of ChBmh1 (Uniprot ID: I1RCD4) was enriched in the cell lysate, and the ortholog of ChSnc1 (Uniprot ID I1RW76) was exclusive to F. graminearum EVs [12]. This suggests that the ChBmh1 and ChSnc1 might be EV markers conserved across fungal species.

RsTsp2 is yet another tetraspanin recently identified as a fungal EV marker in R. solani [22]. RsTsp2 was enriched in EVs, and dsRNA-based silencing of the gene RsTsp2 led to decreased EV production and reduced disease in rice [22]. However, the size of the EVs was consistent between the untreated and the dsRNA-treated samples. An ortholog of RsTsp2 was not detected in the F. graminearum EV proteome [12]. Taken together, there are now six reported markers of fungal EVs, and three of these are tetraspanin-like proteins. Thus, it is possible to speculate that other tetraspanins are markers of fungal EVs in other organisms.

Marker-specific antibodies will expand the study of fungal EVs

The targeting of EVs from F. graminearum the anti-Sur7 antibody in this study led to the identification of a strong band via Western Blot. This band was not detected in the cell lysate, and was degraded in the presence of proteinase K, suggesting that a section of Sur7 is not protected from degradation. The strong band corresponding to FgSur7 in the Western blot could in fact represent two closely spaced bands. Potential explanations for this include protein post-translation modifications, partial degradation or incomplete protein reduction, or the presence of protein isoforms. Western blot revealed the presence of Sur7 in fractions 8–13 from size-exclusion chromatography. This is comparable with our previous data that EVs from F. graminearum elute mostly on fractions 7–15 [12, 43], suggesting that the anti-FgSur7 antibody allows to track EVs during the size-exclusion chromatography separation. Labeled and unlabeled vesicles were observed using immunogold electron microscopy. We hypothesize that these two classes of EVs have different cargo and biological functions, as reported for mammalian EVs [44]. EVs released by different cell types, such as planktonic cells or biofilms, differ in size and proteome [45] but the separation of EV subpopulations from the same cell class has not been investigated thoroughly. This is likely due to the absence of EV marker antibodies, and because density gradients are sometimes unable to separate EV subclasses [13, 46, 47]. To date, most of the information of EV subclasses comes from their size analysis using techniques such as nanoparticle tracking analysis (NTA), but, to our knowledge, no study has fractionated EVs from the same WT source. Using the anti-FgSur7 antibody, EV subpopulations may be separated by immunoaffinity chromatography using inexpensive and rapid protocols [48], or by more sophisticated methods such as phosphoproteomics [49].

Sur7 will also allow the tracking of EVs from the fungal cell into the host. Using techniques such as immunogold microscopy, F. graminearum EVs can be labeled using Sur7, while plant EVs can be labeled using homologs of the markers PEN1 and TET8 from A. thaliana [50]. The labeling of Sur7 will also reveal potential sites for EV biogenesis inside the fungal cell. This would confirm if, as hypothesized, fungal EVs originate from multivesicular bodies and direct plasma membrane budding [51]. Finally, tracking Sur7 will provide valuable information on how EVs traverse the fungal cell wall. This could also be achieved by generating SUR7-GFP transgenic lines.

Understanding the EV surface will facilitate their application as potential crop protectants. Cryptococcus EVs carry immunomodulatory proteins that are vaccine candidates against cryptococcoses in mice models [20], while mammalian EVs are also candidates to treat SARS-CoV-2 and other viruses [52]. This suggests that EVs from plant pathogens might trigger plant defense mechanisms and enhance disease resistance when applied in the absence of the pathogen. We can now test this hypothesis by using the anti-FgSur7 antibody to target Sur7 in the EV surface and track the EV interaction with the host.

Sur7 is a potential fungal EV marker in other organisms

The epitope used to produce the anti-FgSur7 antibody is conserved in several Fusarium pathogens, indicating that Sur7 is likely a pan-Fusarium EV marker. This antibody could become a platform for the qualitative detection of Fusarium pathogens. Most of the diagnostic tools for F. graminearum involve the production of mycotoxins [53], but there are isolates that do not produce toxins, yet they cause crop loss. A protein-based diagnosis would be complementary to currently employed tools. Although Sur7 is present in other fungi, our epitope does not return significant sequence matches in other fungal genera. This high difference in protein sequence explains why the anti-FgSur7 antibody did not react with EVs from C. albicans or S. cerevisiae. However, our group has also produced an anti-Sur7 antibody for C. albicans that can identify Sur7 using Western blot (Dr. James McKenna, pers. comm.). Finally, other putative markers that were exclusive to F. graminearum EVs and thus have promising potential are Cdc42, Rho3, and Ykt6-like. Similarly, Yop1 is a promising candidate and more information on its protein structure will be helpful for the future development of antibodies to track fungal EVs.

Conclusion

We provide visual evidence of a surface marker for fungal EVs. The tetraspanin-like protein Sur7 was detected by Western blot and immunogold microscopy in some but not all vesicles from Fusarium graminearum isolate PH-1. Sur7-labeled vesicles were significantly larger than unlabeled EVs, indicating that there are subpopulations of EVs. We speculate that these subpopulations have different cargo and biological roles. The availability of a fungal surface marker unlocks advanced EV characterization techniques.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

DGC, MRB, and MAA conceived the project and acquired the funding. DGC performed the experiments and wrote the manuscript. SW prepared and imaged the samples for immunogold labeling. All authors edited the manuscript.

Funding

DGC is funded by the Australian Government Linkage Grant Scheme through the Australian Research Council Research Hub for Sustainable Crop Protection (Project Number IH190100022). This project was also funded by Australian research council DP DP160100309 and DP200103393 to MAA, and a La Trobe University ABC grant to DGC.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All co-authors read the manuscript and agreed its publication.

Competing interests

The authors declare no competing interests.