Characteristics of Extracellular Vesicles Released by the Pathogenic Yeast-Like Fungi Candida glabrata, Candida parapsilosis and Candida tropicalis

Candida spp. yeast-like fungi are opportunistic pathogens in humans and have been recently found to release extracellular vesicles (EVs) that are involved in many vital biological processes in fungal cells. These include communication between microorganisms and host–pathogen interactions during infection. The production of EVs and their content have been significantly characterized in the most common candidal species Candida albicans, including the identification of numerous virulence factors and cytoplasmic proteins in the EV cargo. We have here conducted the isolation and proteomic characterization of EVs produced by the clinically important non-albicans Candida species C. glabrata, C. tropicalis and C. parapsilosis. With the use of ultracentrifugation of the cell-free culture supernatant, the candidal EVs were collected and found to be a heterogeneous population of particles for each species with sizes ranging from 60–280 nm. The proteinaceous contents of these vesicles were analyzed using LC-MS/MS, with particular attention paid to surface-expressed proteins that would come into immediate and direct contact with host cells. We thereby identified 42 extracellular and surface-connected proteins from C. glabrata, 33 from C. parapsilosis, and 34 from C. tropicalis, including membrane-associated transporters, glycoproteins and enzymes involved in the organization of the fungal cell wall, as well as several cytoplasmic proteins, including alcohol dehydrogenase, enolase, glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase and pyruvate kinase, for which the vesicular transport is a possible mechanism underlying their non-classical secretion.


Introduction
Polymorphous fungi of the Candida genus are still considered to be the main fungal infectious agents in humans, causing invasive mycoses and bloodstream infections that threaten the health and life of diverse groups of vulnerable patients. Groups that are particularly susceptible to poor outcomes from these infections include infants, the elderly, and individuals with impaired immunity or weakened defense mechanisms due to systemic diseases such as diabetes or cancer, or who have had medical procedures such as surgery or the application of parenteral nutrition or central venous catheters [1,2]. The wide distribution of these opportunistic pathogens in the environment, their relatively easy transmission between individuals, difficulties with making an accurate and rapid diagnosis, and the ineffective treatment of invasive candidiases related to acquired and naturally emerging pathogen resistance to antifungal drugs, result in high morbidity rates and significant mortality outcomes of up to 30-40% [3][4][5][6][7]. The burden of severe fungal infections worldwide and the related threats to the constantly increasing number of individuals susceptible to such infections, requires a comprehensive investigation of their pathogenesis. Although the best-known and the most common species of this genus causing infections in humans is still C. albicans, the global epidemiology of candidiasis is currently changing and other species, collectively called non-albicans Candida (NAC) species and primarily including C. glabrata, C. parapsilosis and C. tropicalis, are increasingly emerging as etiological agents of systemic diseases [8][9][10]. The biology and the virulence mechanisms of NAC species are still not thoroughly understood, which is a particularly important and pressing issue if one takes into account the growing clinical significance of these fungi and the increasing number of serious infections they cause [1].
One of the recently described mechanisms considered as essential not only in fungal cell physiology and intraspecies communication, but also in the virulence of fungal pathogens, is the production of extracellular vesicles (EVs). These are structures surrounded by a lipid bilayer, carrying various proteins, nucleic acids, carbohydrates, toxins, pigments and other molecules in their interior or on their surface, and are involved in the transport of this complex cargo from the inside of the cell to the surrounding environment across the fungal cell wall [11][12][13][14][15][16][17]. Fungal EVs can often be considered as tightly packed compartments, i.e., encapsulated sections of the cytoplasm additionally enriched with selected molecules as a result of external conditions, that deliver their content to the proper destination whilst protecting it from the degradation during transport and ensuring a sufficient concentration of specific co-transported molecules and the possibility of their cooperation at the target site [18][19][20][21]. The production of EVs has been reported for several species of fungal pathogens, including C. albicans, Cryptococcus neoformans, Histoplasma capsulatum, Paracoccidioides brasiliensis, Malassezia sympodialis and Alternaria infectoria, and their contents have been characterized, including the repeat identification of numerous virulence factors and molecules involved in cellular response to stress or adaptation to new environmental conditions [22][23][24][25][26][27][28]. The presence of proteins related to the pathogenesis and development of fungal infection, such as adhesins, immunogens, pigments-synthesizing enzymes, proteases and other hydrolases, in EV preparations undeniably indicates their significant importance in the interactions between invading microorganisms and their human host [29,30]. These relationships are associated with the possibility of the internalization of EVs by phagocytes and with the modulation of the inflammatory response resulting in the production of nitric oxide, IL-4, IL-12, IL-10, TGF-β and TNF-α by immune cells [31][32][33].
Among the Candida fungi, most of information available on vesicle production is for C. albicans [26,27,30,34]. For other species of this genus with great epidemiological and clinical significance, including C. glabrata, C. tropicalis and C. parapsilosis, there are no data on these structures in spite of their potential biological and pathological importance. Hence, the aim of our present study was the isolation and characterization of EVs produced by C. glabrata, C. tropicalis and C. parapsilosis, with a particular focus on proteomic analysis and the proteins exposed at the surface of EVs that come into immediate and direct contact with host cells.

Isolation of Extracellular Vesicles
To isolate EVs, supernatants from aerobic cultures of C. glabrata, C. tropicalis and C. parapsilosis grown in RPMI 1640 medium were collected by centrifugation at 4000 g for 15 min at 4 • C, repeated twice, and the pellet containing cells and their remnants was discarded each time. The supernatants were then concentrated 400-fold using an Amicon Ultra-15 Centrifugal Filter Unit with a 100 kDa cut off (Merck, Darmstadt, Germany) with the addition of cOmplete Protease Inhibitor Cocktail (Roche, Basel, Switzerland). The samples were centrifuged for 5 min at 5000 g and after discarding the pellet were filtered using an Ultrafree-CL Centrifugal Filter with pore size of 0.65 µm (Merck). After this step, the aliquots were spread onto YPD agar plates to verify the absence of any remaining fungal cells. Concentrated supernatants were then ultracentrifuged at 4 • C for 1 h at a rotor speed of 40,000 rpm, which corresponds to relative centrifugal field of 146,000 g (k factor 108), using a fixed-angle type 50.2 Ti Rotor and polycarbonate thick wall centrifuge tubes (13 × 64 mm) with a 13 mm diameter Delrin tube adapters in an Optima TM L-90K Ultracentrifuge (all from Beckman Coulter, Brea, CA, USA). The obtained EV pellets were gently washed with 400 µL of 0.22 µm-filtered phosphate buffered saline (PBS) buffer, pH 7.4 and subjected to another ultracentrifugation step in 400 µL of PBS under the same conditions. The supernatant was then discarded and the EVs were transferred in 200 µL of PBS to Eppendorf tubes and frozen at −80 • C until further use.

EV Size and Concentration Measurements
The size and concentration of the EVs prepared from C. glabrata, C. tropicalis and C. parapsilosis were measured using the nanoparticle tracking analysis (NTA) and NanoSight NS300 system with camera type sCMOS, laser Blue488 and NTA software Version 3.4 (Malvern Instruments, Malvern, UK). The samples from two different isolations were recorded three times for 60 s with camera level of 13 and the threshold parameter set on 2. Representative histograms of size distribution were selected for presentation. Measurements were conducted at 25 • C in PBS buffer filtered through a 0.22 µm filter.

Protein and Phospholipid Concentration Measurements
Protein concentrations in the EV-containing samples was measured in five biological replicates for each species with o-phthalaldehyde (OPA; Sigma) according to previously described methods [35,36]. In brief, 10 µL of sample and 300 µL of OPA reagent were applied to the wells of a black 96-well microplate (Greiner Bio-One, Monroe, NC, USA) and the fluorescence intensity was measured with excitation and emission wavelengths of 340 and 455 nm, respectively, using a Synergy H1 Microplate Reader (BioTek Instruments, Winooski, VT, USA). The phospholipid concentrations in the EV preparations were measured in six biological replicates for each species using the Phospholipid Assay Kit (Sigma) strictly in accordance with the manufacturer's instructions.

Visualization of EVs
To visualize the EVs obtained from C. glabrata, C. tropicalis and C. parapsilosis, negative stained transmission electron microscopy (TEM) was used with formvar coated, 300 mesh copper grids prepared for each EV preparation using 2% uranyl acetate (Chemapol, Prague, Czech Republic). The JEOL JEM2100 HT transmission electron microscope was then used to observe the samples (JEOL, Tokyo, Japan).

Liquid Chromatography-Coupled Tandem Mass Spectrometry (LC-MS/MS) Identification of Proteins Localized Within and at the Surface of the EVs
To identify proteins localized at the surface of fungal EVs, a procedure consisting of the hydrolysis of surface-exposed proteins with trypsin ("surface shaving") was used with a similar methodology to a previously described technique [37] but with some significant modifications. Briefly, C. glabrata, C. parapsilosis and C. tropicalis EVs were suspended in 400 µL of 25 mM ammonium bicarbonate buffer (NH 4 HCO 3 ) with 1 µg of sequencing-grade trypsin (Promega, Madison, WI, USA) and incubated for 10 min at 37 • C. The EVs were then pelleted by ultracentrifugation (as described above) and the collected supernatant was further incubated for 24 h at 37 • C to enhance tryptic digestion. The obtained supernatants were then further centrifuged (15 min, 12,000 g), dried in a Speed-Vac (Martin Christ, Osterode am Harz, Germany). The obtained peptides were identified using LC-MS/MS analysis as described earlier [38], being first separated on a 100 mm × 2.1 mm Aeris 3.6 µm PEPTIDE XB-C18 column (Phenomenex, Torrance, CA, USA) with an ultra-high-performance liquid chromatography (LC) Dionex Ultimate 3000 system (Dionex, Sunnyvale, CA, USA), and directly analyzed with LC-coupled HCTUltra ETDII ion-trap mass spectrometer equipped with an electrospray ionization ion source (Bruker, Bremen, Germany). Four independent biological replicates were prepared from the EVs of each investigated Candida species.
To also identify the proteins present inside the EVs, the vesicles were suspended in 400 µL of ultrapure water and sonicated twice for 30 s with UP50H Compact Lab Homogenizer (amplitude 80%, cycle 0.5, 50 watts, 30 kHz; Hielscher Ultrasonics, Teltow, Germany). After ultracentrifugation, the protein-containing supernatants were collected and lyophilized in an Alpha 1-2 lyophilizer (Martin Christ). Three independent biological replicates were performed for each Candida species. Additionally, membranes that remained from the EVs after sonication and ultracentrifugation were subjected to further protein extraction to enrich the analysis by identifying proteins located within or near the membrane internal or external surface. These pellets were therefore incubated in 200 µL of 25 mM NH 4 HCO 3 with 0.1 µg of trypsin for 30 min at 37 • C or lyophilized and then further incubated for 30 min with 15 µL of 50 mM Tris-HCl, pH 6.8 supplemented with 0.4% sodium dodecyl sulphate (SDS) and 4% 2-mercaptoethanol (2-ME). After trypsinization, ultracentrifugation was carried out as described above to remove debris and collect peptide-containing supernatant that was subjected to further overnight incubation at 37 • C to enhance tryptic digestion and enable for further LC-MS/MS analysis.
Proteins obtained after lyophilization were separated by SDS-PAGE electrophoresis and stained with Coomassie Brilliant Blue R-250. After obtaining the proteins by EVs sonication, the average amount of protein of 10 µg for C. glabrata and C. tropicalis and 15 µg for C. parapsilosis was applied to the gel per lane. All noticeable protein bands were manually excised; however, if they were smudged, the entire lane was cut into several fragments, then subjected to tryptic digestion in accordance with a procedure described elsewhere [39] and analyzed using LC-MS/MS as described above. The lists of obtained peaks were searched against the nonredundant protein database of the NCBI with taxonomies restricted to fungi (26,490,256 sequences for all entries, 1,924,810 sequences for fungi) using an in-house Mascot server (v.2.3.0; Matrix Science, London, UK). The following search parameters were applied: fixed modification, carbamidomethylation (C); variable modifications, oxidation (M); protein mass, Cells 2020, 9, 1722 5 of 22 unrestricted; number of missed cleavages, 2; peptide mass tolerance of ± 0.3 Da and fragment mass tolerance of ± 0.5 Da. Proteins with at least one unique peptide and for which the Mascot protein score was higher than 53 were taken into account in further analyzes, and of these proteins, only those that appeared at least twice in all replicates performed were considered the most abundant in the samples tested.

Characterization of C. glabrata, C. tropicalis and C. parapsilosis EVs with High-Resolution Flow Cytometry Analysis
High-resolution flow cytometry analysis with an Apogee A60-Micro-PLUS flow cytometer (Apogee Flow Systems, Hemel Hempstead, UK) was applied for the detection of EV surface-localized α-mannopyranosyl residues with the use of fluorescein-conjugated concanavalin A (Con A; Thermo Fisher Scientific). Prior to use, Con A was centrifuged at 21,000 g for 20 min at 4 • C to remove any potential lectin aggregates. Fungal EVs were then suspended in 0.22 µm-filtered PBS (Lonza, Basel, Switzerland) and incubated with Con A (100 µg/mL) for 30 min at room temperature and immediately acquired. Apogee Calibration Beads #1493 and #1517, containing a mixture of green-fluorescent polystyrene (PS) and non-fluorescent silica (Si) beads, were used to calibrate the flow cytometer. The obtained data were analyzed using Apogee Histogram Software and representative plots from two biological replicates for EVs isolated from each Candida species were selected for presentation.

Results
For the preparation of fungal EV samples, Candida species were cultured in RPMI 1640 medium. The cultures that were used further for isolation contained approximately 4 × 10 10 (C. glabrata and C. tropicalis) or 10 × 10 10 (C. parapsilosis) living cells, which were verified by CFU counting on YPD agar plates. No fungal cells were found in the filtered concentrated supernatants that were subsequently subjected to ultracentrifugation. Nanoparticle tracking analysis (NTA) was used to estimate the number of particles contained in a single EV sample obtained after concentration of the total harvested volume of supernatant from the growth of particular Candida species. For C. glabrata, the particle number was determined as 2.55 ± 0.17 × 10 10 of produced EVs, whereas for C. parapsilosis this was 2.3 ± 0.08 × 10 10 particles and for C. tropicalis was 8.88 ± 0.63 × 10 9 particles. The measured contents of proteins and phospholipids per 1 × 10 10 EVs for each Candida species are listed in Table 1. As recommended by the International Society for Extracellular Vesicles [40], the Candida EV sizes were measured by NTA and the heterogeneity of the preparations, i.e., the presence of several EV populations of different sizes, was demonstrated. The measurement data are provided in Table 2 and representative histograms for the EV samples from the three Candida species under analysis are presented in Figure 1.   Each of the tested Candida species produced EVs with sizes ranging from 60-280 nm (90% of total population). The NTA results were confirmed by TEM analysis which revealed spherical structures of different sizes within this size range ( Figure 2). Each of the tested Candida species produced EVs with sizes ranging from 60-280 nm (90% of total population). The NTA results were confirmed by TEM analysis which revealed spherical structures of different sizes within this size range ( Figure 2).  To characterize the protein composition of the EVs, protein identification was performed with LC-MS/MS and the topology of individual molecules was assessed in accordance with ISEV recommendations [40]. EV surface shaving with trypsin was conducted as previously described [41], which is a similar method to that previously performed for whole fungal cells [37,38]. Proteins exposed at the surfaces of C. glabrata, C. parapsilosis and C. tropicalis EVs were then identified. The concentration of trypsin and the reaction time used were selected carefully to avoid breakage of the vesicles and uncontrolled release of their contents [42]. The fungal EVs were also sonicated to disrupt their membrane integrity and enrich for molecules contained within the vesicles. However, several proteins associated with the membrane and cell wall were also identified in this group, indicating that they may become fragmented and detached from the surface during the procedure. Nevertheless, given that transmembrane proteins or proteins anchored in the membrane and thus located in close proximity to it could be difficult to identify using these two abovementioned methods, pellets containing vesicle membrane remnants were collected by ultracentrifugation and subjected to additional hydrolysis with trypsin or protein extraction using detergent and reducing reagent. The lists of the most abundant identified proteins, i.e., those that were identified at least twice within all of the biological replicates analyzed, are presented in Table 3 for C. glabrata, Table 4 for C. parapsilosis and Table 5 for C. tropicalis EVs. Additionally, all proteins that were identified in all biological replicates and those that were identified only once in a particular fraction are listed in Supplementary Tables S1-S3, together with the raw data and mass spectrometry identification parameters. Table 3. Mass spectrometry identification of abundant proteins in C. glabrata EVs after vesicle surface shaving with trypsin (T) or vesicle sonication (S), and in EV membrane protein-enriched fractions (M). The peptides obtained were analyzed using the Dionex Ultimate 3000 UHPLC system coupled to an HCTUltra ETDII mass spectrometer and the obtained lists of peaks were searched against the NCBI protein database using an in-house Mascot server.    We identified 42 highly abundant EV proteins from C. glabrata, 33 from C. parapsilosis, and 34 from C. tropicalis. We found that 32, 22 and 30 of these abundant proteins in C. glabrata, C. parapsilosis and C. tropicalis EVs, respectively, were surface-exposed proteins. In addition, we identified three new proteins for C. glabrata and C. tropicalis, and 10 for C. parapsilosis after sonication. Extending the procedure by further additional step allowed for identification of seven membrane proteins from C. glabrata EVs and one cell wall protein for the other two species, i.e., a C. parapsilosis putative GPI-anchored protein (Pga17) and a C. tropicalis putative adhesin-like cell wall mannoprotein CTRG_02188 (Flo9). This number of identified proteins was lower than in some other proteomic studies on fungal EVs probably due to the differences in the equipment used and detailed methodology that allowed to identify only those proteins that were most abundant in the sample. However, some contamination of individual fractions with proteins from other fractions, which may have various causes, cannot be excluded; therefore particular caution is needed in the interpretation of results that are essentially a preliminary indication of the potential location of these proteins.

NCBI
The Venn diagram shown in Figure 3 indicates that although most of the abundant EV proteins were specific to that Candida species, several EV-derived proteins were shared between species, and two of these proteins-cell wall protein Scw4 and alcohol dehydrogenase Adh1-were common to C. glabrata, C. parapsilosis and C. tropicalis EVs. The proteins found in the EVs from C. glabrata and C. parapsilosis included exo-1,3-beta-glucanase (Exg1/Xog1) and GPI-anchored cell wall protein (Ecm33). Seven proteins were shared for C. tropicalis and C. parapsilosis EVs, including two glycosidases (Phr1 and 2), 1,3-beta-glucosyltransferase (Bgl2) and cell wall mannoprotein (Mp65). C. tropicalis and C. glabrata EVs shared six proteins, including enolase (Eno1), glyceraldehyde-3-phosphate dehydrogenase (Tdh3), phosphoglycerate kinase (Pgk1) and pyruvate kinase (Cdc19), which are highly evolutionarily conserved and common proteins between different cells, hence the most likely their sharing between less related species.
Seven proteins were shared for C. tropicalis and C. parapsilosis EVs, including two glycosidases (Phr1 and 2), 1,3-beta-glucosyltransferase (Bgl2) and cell wall mannoprotein (Mp65). C. tropicalis and C. glabrata EVs shared six proteins, including enolase (Eno1), glyceraldehyde-3-phosphate dehydrogenase (Tdh3), phosphoglycerate kinase (Pgk1) and pyruvate kinase (Cdc19), which are highly evolutionarily conserved and common proteins between different cells, hence the most likely their sharing between less related species. The protein descriptions and gene ontology (GO) annotations for each of C. glabrata, C. parapsilosis or C. tropicalis orthologous genes were assigned based on the information from the Candida Genome Database (CGD, http://www.candidagenome.org) [43], Saccharomyces Genome Database (SGD, https://www.yeastgenome.org) [44] and UniProtKB Database (https://www.uniprot.org) [45]. Proteins identified in the EVs of individual non-albicans Candida species were classified into functional groups according to the appropriate GO annotations and with regard to their principal biological functions. Those groups were as follows: (i) fungal-type cell wall organization, (ii) membrane transport and organization, (iii) stress response, (iv) pathogenesis, (v) cell metabolism, (vi) transcription and translation, and (vii) proteins of unknown function. A graphic representation of this functional division is presented in Figure 4 and indicates that the largest functional group of EV proteins, other than proteins of unknown function, was cell metabolism for C. glabrata (31%) and C. tropicalis (32%) and cell wall organization for C. parapsilosis (37%). The second largest group included proteins involved in membrane transport and organization for C. glabrata (24%), cell metabolism for C. parapsilosis (18%) and cell wall organization for C. tropicalis (14%). In the case of C. glabrata, EV proteins involved in translation (14%) were significantly represented. The protein descriptions and gene ontology (GO) annotations for each of C. glabrata, C. parapsilosis or C. tropicalis orthologous genes were assigned based on the information from the Candida Genome Database (CGD, http://www.candidagenome.org) [43], Saccharomyces Genome Database (SGD, https://www.yeastgenome.org) [44] and UniProtKB Database (https://www.uniprot.org) [45]. Proteins identified in the EVs of individual non-albicans Candida species were classified into functional groups according to the appropriate GO annotations and with regard to their principal biological functions. Those groups were as follows: (i) fungal-type cell wall organization, (ii) membrane transport and organization, (iii) stress response, (iv) pathogenesis, (v) cell metabolism, (vi) transcription and translation, and (vii) proteins of unknown function. A graphic representation of this functional division is presented in Figure 4 and indicates that the largest functional group of EV proteins, other than proteins of unknown function, was cell metabolism for C. glabrata (31%) and C. tropicalis (32%) and cell wall organization for C. parapsilosis (37%). The second largest group included proteins involved in membrane transport and organization for C. glabrata (24%), cell metabolism for C. parapsilosis (18%) and cell wall organization for C. tropicalis (14%). In the case of C. glabrata, EV proteins involved in translation (14%) were significantly represented.
High-resolution flow cytometry was next used for the detection of α-mannopyranosyl residues within EV surface exposed glycoproteins with the use of fluorescein-conjugated Con A, a lectin that recognizes mannose-containing glycoconjugates ( Figure 5). Calibration of the flow cytometer using appropriately sized beads was performed and it was verified that the PBS buffer used for the dilution of EV preparations did not contain significant numbers of particles. Upon comparing the fluorescence intensity of unstained vesicles with that of Con A-fluorescein-bound vesicles, the exposure of mannoproteins on the EV surface was confirmed for each of the Candida species tested. The percentage of positive objects in this regard was close to 90% in the case of C. glabrata, and greater than 90% for C. parapsilosis and C. tropicalis.  High-resolution flow cytometry was next used for the detection of α-mannopyranosyl residues within EV surface exposed glycoproteins with the use of fluorescein-conjugated Con A, a lectin that recognizes mannose-containing glycoconjugates ( Figure 5). Calibration of the flow cytometer using appropriately sized beads was performed and it was verified that the PBS buffer used for the dilution of EV preparations did not contain significant numbers of particles. Upon comparing the fluorescence intensity of unstained vesicles with that of Con A-fluorescein-bound vesicles, the exposure of mannoproteins on the EV surface was confirmed for each of the Candida species tested. The percentage of positive objects in this regard was close to 90% in the case of C. glabrata, and greater than 90% for C. parapsilosis and C. tropicalis.

Discussion
EVs are ubiquitously produced by organisms of all kingdoms and take part in important biological processes in cells, in intercellular communication, and also in interactions between pathogens and their host during infection [46]. These bodies are richly loaded with variety of cargos that are protected by a lipid bilayer and transported from the cell to the external milieu through different mechanisms which are not yet fully recognized. Vesicle populations are often heterogenous in size and composition, depending on the secretion pathway, the physiological state of the cells, or the influence of the extracellular environment, and this is reflected in the large variety of their functions [47]. The crucial role of EVs in health disorders and infectious diseases has led to a continuous and increasing interest in these structures [46]. The composition and role of EVs have been characterized to some extent for several species of fungi that are pathogenic to humans [23][24][25]28]. Among Candida species, however, such data are currently available only for the EVs produced by C. albicans [26,27,30,34,48]. Imaging results for these extracellular structures have been published for C. parapsilosis [22].
In our present study, we have for the first time characterized the EVs produced by C. glabrata, C. parapsilosis and C. tropicalis, with a particular emphasis on the vesicular proteome. The EVs produced by planktonic cells were isolated in our experiments but the fungi were not grown in the YPD medium routinely used for yeast culture but in buffered RPMI 1640 medium that provides all the components necessary for growth, but without proteinaceous content. The number of vesicles produced by fungal cells under these conditions was then determined, although it was lower than the number of living cells in each case which was probably due to the general difficulties in multi-stage EV isolation procedures using culture supernatants. The phospholipid and protein contents were also measured and clearly revealed major vesicle-building components in the obtained preparations. The differences between the protein concentration for EVs of individual species were not statistically significant, whereas the highest lipid concentration was demonstrated for C. parapsilosis EVs, with the lowest found for C. glabrata.
Analysis of the sizes of the isolated C. glabrata, C. parapsilosis or C. tropicalis EVs using NTA and TEM revealed heterogeneous populations for each tested species. This is also the case in the vesicles produced by the planktonic form of C. albicans in which fractions of EVs of 50-100 nm and larger than 200 nm in size were previously described [26,27].
Mass spectrometry identification of proteins from C. glabrata, C. parapsilosis and C. tropicalis vesicles revealed the presence of numerous proteins involved in the membrane transport of different molecules including glucose, siderophores, copper and hydrogen ions, suggesting a plasma membrane origin of the EVs. This group of proteins was particularly represented in the case of the C. glabrata EVs, where only in the fraction of membrane-associated proteins Hxt5 and Hxt6, Put4, Ctr1, Arn1 and Tna1 were identified. In addition, the proteins involved in the building, remodelling and integrity of the fungal cell wall were abundant in EVs from all three tested species. Cell wall glucanases, including Eng1, Exg1/Xog1, Bgl2 and Scw4 were indicated to be EV-derived proteins shared between these investigated NAC organisms. This last protein was identified for all three investigated species both at the surface and in the inside of the EVs, similarly to the location of Bgl2 for C. parapsilosis and C. tropicalis. The activity of these factors may be important in the process of vesicle passage through the cell wall to the external environment, but also may play a role in the remodelling of surface polysaccharides, enabling fungi to modulate their interactions with host immune cells by interfering with the recognition of pathogens via specific receptors [49]. The typical candidal cell wall proteins are most often highly glycosylated and equipped with branched O-and N-linked mannosides [50]. Their presence at the surface of C. glabrata, C. parapsilosis and C. tropicalis vesicles was also confirmed in our current experiments by high resolution flow cytometry after their staining with the ConA lectin that recognizes mannose residues, thus corroborating the identification of several mannoproteins in EVs using LC-MS/MS.
Regarding the postulated significant role of EVs in the virulence of Candida fungi, which has been previously suggested for C. albicans [27,34], it was particularly important in our present analyses to identify the proteinaceous surface components of EVs that would be capable of mediating their immediate and direct contact with host proteins and cells during infection. Some of the vesicular proteins identified in our present analysis have a reported role in adhesion and pathogenesis that was directly attributed previously based on the functions of their orthologous proteins. These include the C. parapsilosis yeast cell wall protein Ywp1 involved in biofilm dispersal and β-1,3-glucan masking in yeast cells [51,52], the C. parapsilosis and C. tropicalis cell wall mannoprotein Mp65 that exhibits adhesive properties and contributes to fungal virulence and the stimulation of dendritic cells [53,54] and inducible acid phosphatase Pho100 [55], which has also been identified at the EV surface in these two Candida species. Mp65 was also identified inside the EVs of these two species. In addition, two aspartic proteinases belonging to larger families of proteins involved in proteolysis and virulence-C. glabrata yapsin Yap3 and C. tropicalis Sapt4-were also identified in our analysis, although their specific role in infections requires further study [56][57][58]. C. glabrata proteinase was identified in all tested fractions isolated during EVs analysis, whereas the enzyme of C. tropicalis only at their surface. For some of the typical proteins associated with the fungal cell wall and identified in the EVs from our current NAC species, and additional involvement in pathogenesis apart from cell wall organization was also indicated in prior reports for their orthologs, such as cell wall 1,3-beta-glucosyltransferase Bgl2 [59] and glycosidase Phr 2 [60] that we here identified from the C. parapsilosis and C. tropicalis vesicles.
A large proportion of the proteins we identified in the EVs produced by each of the three NAC species in our current study are involved in basic cell metabolism. This may indicate that these vesicles are derived from the cytoplasm and thus carry many intracellular enzymes [18]. As such proteins are often found outside of fungal cells or at their surface [37] but do not possess a signal sequence directing them to a classical secretory pathway, there is likely to be another unconventional pathway leading to their secretion, i.e., EV transport [61]. Numerous cytoplasmic proteins have already been identified in C. albicans vesicles, including enolase (Eno1), alcohol dehydrogenase (Adh1), phosphoglycerate kinase (Pgk1) and glyceraldehyde-3-phosphate dehydrogenase (Tdh3) [62]. In the case of the EVs from our current NAC species, the cytoplasmic proteins we identified included Adh1, Eno1, Pgk1, Tdh3, pyruvate kinase Cdc19, pyruvate decarboxylase Pdc1 and 6-phosphogluconate dehydrogenase Gnd1. In the case of all these proteins identified for C. glabrata, they were found in all three analyzed fractions, both at the surface of the vesicles and inside them, whereas in the case of C. tropicalis, the surface location predominated for most of them among the proteins identified for this species, except for Eno1 and Pgk1 also found inside of the vesicles. Although these proteins have a cytoplasmic origin and are involved in basic metabolism in accordance with their primary function, important roles in host-pathogen interactions have been repeatedly attributed to these factors or their orthologs in C. albicans, thus defining them as moonlighting proteins and candidal molecules contributing to virulence [63][64][65]. C. tropicalis enolase was identified previously as a protein that binds human high-molecular-mass kininogen, a component of the plasma kinin-generating system involved in innate immunity, and human vitronectin, fibronectin and laminin [66,67]. It was also demonstrated in our current study that Eno1 is present at the surface of C. tropicalis and C. glabrata EVs. Adh1 was identified in the EVs of all three investigated NAC species in our present study. Its ortholog from C. albicans is known to be immunogenic during infection and is responsible for binding of extracellular matrix proteins and for the differentiation of human monocytes to macrophages enhancing their adhesion, phagocytosis, and killing capacities [68][69][70]. In addition, other proteins identified in fungal vesicles might be immunogenic during infection, including C. parapsilosis Adh1 [71] and C. tropicalis Eno1 and Tdh3 [72], thus being potential candidates for future immunodiagnostic tests or vaccines.
Of the EV proteins identified for Candida species, many of them were also indicated in the extracellular vesicles of other fungi, both S. cerevisiae and other species pathogenic to humans [21]. Such intracellular proteins as Pgk1, Tdh3 and Gnd1 were also found in EVs of P. brasiliensis, H. capsulatum, C. neoformans or S. cerevisiae, Eno1 was identified for C. neoformans, P. brasiliensis and S. cerevisiae, and Fba1 for the last two species and H. capsulatum [12,22,23,73]. This indicates a fairly common phenomenon associated with the presence of these proteins in the EVs among different fungal organisms, which may be associated with their unusual secretion. In addition, among the proteins involved in transmembrane transport, two, i.e., ADP, ATP carrier protein and plasma membrane ATPase identified for C. glabrata or C. tropicalis were also found in EVs of C. neoformans, P. brasiliensis, H. capsulatum and S. cerevisiae [12,22,23,73] and enzymes responsible for cell wall remodelling such as Exg1 and Scw4 or cell wall associated protein Ecm33 were also found in S. cerevisiae vesicles [73].
The characteristics of EVs produced by C. glabrata, C. parapsilosis and C. tropicalis presented in our current analysis significantly contribute to our understanding of the biology of these pathogens, and the importance of these vesicles in the pathogenesis of infections caused by NAC species. It is particularly noteworthy in this regard that these structures are involved in the transport of a number of fungal virulence factors.