The extracellular release of macromolecules is essential for a number of physiological events in fungal cells, including nutrition, intercellular communication, biofilm formation and, in pathogenic fungi, activation of immune cells1,2. Fungal cells are encased within a dense cell wall, so mechanisms are required for the transport of molecules across the wall for extracellular release in these organisms. In recent years, a number of studies have shown that the trafficking of molecules across the fungal cell wall requires extracellular vesicles (EVs)3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18.

Fungal EVs contain a complex combination of macromolecules, including many proteins, neutral lipids, glycans and pigments3,4,7,9,10,14,15,18. This complex molecular mixture, including many cytoplasmic components, is consistent with the proposed origin of fungal EVs as cytoplasmic subtractions19. This would suggest that the complexity of the composition of fungal EVs may be underestimated, because it is not possible to rule out the presence of other macromolecules present in the cytoplasm.

Fungal and mammalian EVs have similar morphological features and compositions20. Mammalian EVs efficiently transport mRNA and other macromolecules to the extracellular space1. Interestingly, mRNAs from mammalian EVs can be translated into proteins by target cells21,22. It has recently been shown that EVs also contain small noncoding RNA species22,23. EV-associated RNA produced by pathogens could affect the physiology of host cells, as suggested by experiments with mammalian EV RNA24,25.

The presence of RNA in fungal EVs was suggested by the findings of a single flow-cytometry study focusing on the human pathogen Cryptococcus neoformans6. However, the structural aspects of EV-associated fungal RNA remain to be explored. In this study, we selected four species of EV-producing fungal cells for an analysis of the presence of RNA sequences. Our results for the model yeast Saccharomyces cerevisiae and the human pathogens C. neoformans, Paracoccidioides brasiliensis and Candida albicans, demonstrated that fungal cells use EVs to export RNA molecules of different natures and with different biological functions.

Results and Discussion

Comparative transcriptomic analysis of fungal extracellular vesicles (EVs)

Total RNA was isolated from fungal EVs in three independent biological replicates. The RNA was then fractionated into samples enriched in small RNAs (less than 200 bp) and mRNAs. The small RNA-enriched fractions were then analyzed, to determine their molecular size distribution (Fig. 1A). EV small-RNA fractions consisted of molecules of various sizes, mostly fewer than 250 nucleotides long. One preparation obtained from P. brasiliensis Pb18 illustrate this profile (Fig. 1A) and similar observations were made for all the species studied. We conclude that fungal EVs carry small RNA molecules with similar size distributions.

Figure 1
figure 1

Electropherograms of the small-RNA content of EVs from the Pb18 strain of P. brasiliensis (A) and the control with RNAse treatment (B).

Retention time (s) and fluorescence intensity (FU) are indicated on the corresponding axes of the graphs generated from the profiles shown on the left. RNA length (bp) is shown for each peak.

We checked that RNA molecules were not loosely aggregated with fungal EV rather than being packaged within these vesicles, by subjecting control samples to RNAse treatment before RNA extraction (Fig. 1B). The small-RNA profile of RNAse-treated EVs was not modified, confirming that the small RNA molecules were within the vesicles, which protected them from hydrolysis. We also assessed RNase activity, by adding exogenous, total fungal RNA EV preparations before RNase treatment. The total RNA was degraded, consistent with functional RNase-mediated hydrolysis (Fig. S1A to D).

Genome RNA EV mapping statistics are listed in Table 1. For C. neoformans samples, about 90% of the RNA EV reads mapped to intronic regions, the remaining 10% mapping to exons, in mature mRNA. A similar profile was clearly observed in P. brasiliensis samples, in which 17% of the reads mapped to intronic regions and 82% mapped to exons. However, 21% of the reads in P. brasiliensis mapped to exon-intron regions, versus less than 1% of the mapped reads in C. neoformans. In S. cerevisiae and C. albicans, about 90% of the reads mapped to exons. This observation is consistent with the introns lower number in these species26,27, by contrast to P. brasiliensis and C. neoformans.

Table 1 Sequence mapping statistics

Gemome RNA EV mapping information is illustrated in Figure 2. Reads mapping to an entire mRNA (Fig. 2A), intergenic (Fig. 2B) and intronic regions (Fig. 2C), or aligning with specific positions in the mRNA (Fig. 2D) were observed for all species studied. This suggests that, in addition to the small RNA molecules previously observed, mRNAs are also present in fungal EVs. Most of the reads aligning with intergenic and intronic regions were found to be in the reverse orientation, complementary to the transcript (Fig. 2D). This pattern is characteristic of sequences processed by the small-RNA interference (RNAi) machinery28,29.

Figure 2
figure 2

Genome RNA EVs mapping (A), high-coverage exon; (B) intron; (C), intergenic regions; (D) low-coverage exon.

Gray bar, genomic region, with length in bp, shown above; black lines above the reads, consensus; vertical columns, number of sequences; yellow bars, exons (access codes shown); blue bars, introns. For the sequence colors: forward reads, red; reverse reads, green and non-specific match, yellow.

Small RNA characterization: miRNA-like sequences

For the identification of conserved miRNA sequences among the small RNAs identified in EVs from P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae, we used a database of mature miRNA sequences from all organisms characterized to date ( We identified a total of 1,246 conserved miRNA-like sequences, only 20 of which were common to all four fungal species studied (Table 2). The total numbers of miRNAs identified were 145 (1,953 mean reads) for P. brasiliensis, 344 (4,796 mean reads) for C. neoformans, 423 (2,477 mean reads) for C. albicans and 532 (4,349 mean reads) for S. cerevisiae (Table S1). We then used Baggerly's test30 to compare the levels of miRNA-like molecules in EVs among the four species. We found that 47 of these molecules were differentially distributed (Fig. 3). C. neoformans had the largest number of overrepresented sequences (16), followed by C. albicans (8), P. brasiliensis (12) and S. cerevisiae (1). We also identified five miRNA-like sequences exclusive to P. brasiliensis and one exclusive to S. cerevisiae. The distribution of miRNA-like molecules was more similar in C. neoformans and C. albicans EV, although the values for C. neoformans were considerably higher for most sequences (Fig. 3).

Table 2 miRNA-like common sequences from P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae EVs. Means of normalized values are shown
Figure 3
figure 3

Heat map of miRNA-like molecules showing relative levels in the EVs of P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae.

All sequences with differential distributions between these species are shown. The colors on the plot depend on the replicate means of normalized miRNA-like sequence levels in the EVs of each of the fungal species. The highest value in each row is shown in red and the other colors used change proportionally with the difference. *, the sequences most frequently represented in the species; arrows, unique sequences.

Emu-miR-8, egr-miR-8, tae-miR-156, osa-miR-156f and ata-miR-172f were more frequently represented in C. neoformans and C. albicans (Fig. 3). Members of these miRNA families are involved in the development of Drosophila31,32,33, zebrafish34, tumors35 and plants36. Sequences pvu-miR166a and ssl-miR166a, involved in plant biology37,38, were identified in P. brasiliensis, C. neoformans and S. cerevisiae.

Five miRNA-like sequences were exclusive to P. brasiliensis EVs (Fig. 3). One of them, dre-miR 125a-2, belongs to the miR 125 family, which is highly conserved in eukaryotes39 and is involved in many different cellular processes, including cell proliferation40,41, differentiation and apoptosis, through the targeting of different transcription factors40, matrix-metalloproteases42,43, growth factors44 and nonsense-mediated mRNA decay pathway45. In addition, miR-125 controls the differentiation of immune cells, thereby affecting responses to bacterial and viral infection39,46,47,48,49.

Many fungal species have RNAi machinery, which may be noncanonical in Schizosaccharomyces pombe, for e.g., or absent from others (e.g. S. cerevisiae) and functional in most filamentous fungi50,51,52,53. C. neoformans bears a fully functional RNAi machinery, including Argonaute proteins, Dicer and the RNA-dependent RNA polymerase54,55. In this organism, RNAi-dependent mechanisms are involved in the sex-induced silencing of transgenes during the sexual stage of the fungus55. In addition, C. neoformans miRNA and pre-miRNA may be similar to their mammalian counterparts54,55. In C. albicans, the RNAi machinery includes a noncanonical Dicer and small-RNA molecule preparations from this species are enriched in 22-mer sequences56. The occurrence of RNAi machineries in different fungal pathogens suggests a possible role for miRNA-like molecules during with the interaction of fungi with their host cells, possibly involving the mimicking of endogenous miRNA. This putative mechanism might regulate gene expression in host cells and modify sensitivity to infection. Further investigations, determining the roles of small RNAs in fungal EVs, are required to test this hypothesis.

Identification of non-coding RNAs (ncRNAs)

Our analysis of RNA sequences in fungal EVs included classes of non-coding RNA other than miRNAs. These classes were identified with the use of ncRNA data from the Saccharomyces Genome Database as a template. We identified a total of 114 different ncRNA sequences in EV preparations, many of which were specifically associated with C. neoformans (46), P. brasiliensis (38), C. albicans (68), or S. cerevisiae (106). Eleven ncRNAs were common to all four fungal species (Table 3 and Table S2): one tRNA, two snRNAs, two rRNAs and six snoRNAs. None of the roles of these ncRNAs had previously been characterized during host-parasite interactions.

Table 3 ncRNA sequences identified in EV preparations from C. neoformans (Cn), P. brasiliensis (Pb), S. cerevisiae (Sc) and C. albicans (Ca)

The ncRNA groups most frequently represented in all fungal EV preparations were small nucleolar RNAs (snoRNAs) and tRNAs, which accounted for 22 to 75% of all reads (Fig. 4). S. cerevisiae had a high proportion of snoRNA (75%), whereas tRNAs accounted for 60% of all reads in C. albicans. In C. neoformans, for the most part, reads were fairly evenly distributed between snoRNAs and tRNAs. P. brasiliensis had a slightly different profile, with long ncRNAs accounting for 8% of the sequences and rRNAs accounting for 13%.

Figure 4
figure 4

Distribution of noncoding RNA species (other than miRNA-like species) in fungal EVs.

The percentages of each RNA species indicated on the x-axis are expressed relative to total RKPM. Pb, P. brasiliensis; Cn, C. neoformans; Ca, C. albicans; Sc, S. cerevisiae.

Only eight long ncRNA species were identified in our samples (Table S2). None was common to all fungal EVs, but most were identified in at least two species. In mammalian cells, the transcription of several long ncRNAs is regulated by factors critical for mammalian development57,58,59,60. This suggests possible key roles in development for these ncRNAs, through the regulation of protein synthesis by mechanisms independent of the Dicer RNAi pathway57,60. Table S2 highlights the involvement of ICR157, a cis-interfering long intergenic ncRNA that regulates transcription of the FLO11 locus in S. cerevisiae. ICR1 and PWR1 form a circuit with a bidirectional toggle, including the upstream signaling pathway transcription factors Sfl1 and Flo8 and the chromatin remodeler Rpd3L, which is essential for phenotypic transitions in yeast57,60.

The ncRNA RPR1 was identified in S. cerevisiae EV samples. It interacts with nine other proteins to form the ribonuclease P (RNase P) complex, a ubiquitous endoribonuclease that cleaves precursor tRNAs to generate mature 5' termini61. RNase P is also involved in the turnover of normally unstable nuclear RNA62. The RNA component of mitochondrial RNase P, RPM163, was found in EVs from three fungal species. Another component of a mitochondrial RNAse – NME1 – was found in EV samples from C. albicans and S. cerevisiae. NME1 is a subunit of the essential ribonucleoprotein endoribonuclease, RNase MRP, which is required for the processing of pre-rRNA. In vitro, it promotes the cleavage of pre-5.8S rRNA64,65 and the degradation of specific mRNA sequences involved in cell cycle regulation66.

snRNAs constitute a small group of highly abundant, non-polyadenylated ncRNA transcripts present in the nucleoplasm67,68. They form the core of the spliceosome and catalyze the removal of introns from pre-mRNA67,68. Four different snRNAs were found in fungal EVs (Table S2). LSR1 and snR19 were present in all species, snR6 was absent from P. brasiliensis and snR14 was observed only in C. albicans and S. cerevisiae.

As mentioned above, 72 different snoRNA species were found in fungal EVs (Table S2). These RNAs can be classified into two groups on the basis of sequence motifs and secondary structures: C/D and H/ACA boxes, which are involved in the methylation and pseudouridylation of mammalian rRNA nucleotides, respectively69. They may also play a role in mRNA modification, participating in the control of expression70. The snoRNAs also guide rRNA nucleotide modifications and participate in pre-rRNA cleavage71. A few snoRNAs, such as snR10, have both functions. This molecule is required for normal cell growth, in addition to pre-rRNA processing in S. cerevisiae72. It was found in EVs from all species except C. neoformans. Other snoRNAs identified included U3 (SnR17a/b), U14 (SnR128) and snR30, which are essential for eukaryotic growth73. The collection of snoRNAs found in fungal EVs included snR191RNA, which has been reported to support cell growth in S. cerevisiae74, whereas the snR70, snR71, SnR65 and SNR68 mutants display growth defects75.

It is tempting to speculate that ncRNAs present in the P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae EV preparations may be involved in intercellular regulation, mostly through the regulation of protein synthesis.

The tRNAs constitute an interesting group of ncRNAs. The EVs of the fungal species studied containing 22 to 60% tRNAs, depending on the species analyzed (Fig. 5 and Table S2). In the protozoan parasite Trypanosoma cruzi, which has no RNAi machinery molecules, a PIWI-like protein has been characterized76. However, the cells of this parasite contain a homogeneous population of small RNAs derived from mature tRNAs, known as tsRNAs77. tsRNAs are produced by an alternative small RNA pathway77. Recent studies have shown that these molecules are packaged into vesicles that are shed by the parasite and can be transferred to other parasites and/or host cells24,25. The cargo of these vesicles can change the expression profile of the host cells, leading to changes in the cytoskeleton and extracellular matrix, increasing susceptibility to infection24,25. We detected fragments of mature tRNAs in the EVs of all the fungal species studied (Table 3 and Fig. 5). Fungal EVs were enriched in nuclear and mitochondrial tRNAs, which are not the most abundant population in the cell (Table 3). Specific enrichment in the 3′ region of the tsRNA was also observed (Fig. 5). This scenario is illustrated by the situation for arginine tRNA (CCU), which is present at a relatively low concentration in cellular compartments78,79. This tRNA, which was found in fungal EVs, is required for growth on nonfermentable carbon sources at high temperatures, for the synthesis of the heat shock protein Ssc1p and for Ty1 retrotransposition, through the regulation of translational frameshift78. Our observations indicate that fungal EVs contain tsRNAs that might modify and affect gene expression in host cells.

Figure 5
figure 5

Schematic graphs of reads aligning specifically with the 3′ end of the arginine tRNA(CCU)J identified in EVs from P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae.

Coverage is indicated on the y-axis and alignment position is shown on the x-axis.

mRNA identification

The sequencing of EV-associated small RNA-enriched fractions led to the identification of 253 different mRNAs in C. albicans, 84 in S. cerevisiae, 353 in C. neoformans and 73 in P. brasiliensis (Table S3). The ten most abundant mRNAs from each species found are listed in Table 4. These sequences were obtained from high-coverage sRNAs (Fig. 1A), suggesting that molecules of more than 200 nt were copurified despite the use of a strategy favoring the purification of small RNAs. We validated the mRNA sequencing data obtained for fungal EVs, by performing RT-qPCR on two representative sequences corresponding to exons with and without high coverage rates (Fig. S2). The RNA-seq data demonstrated that the EV mRNA molecules corresponded to 1–2% of those found in intracellular compartments26,27,80,81,82. The most abundant EV mRNAs differed from the most abundant cellular transcripts, consistent with regulated compartmentalization for extracellular export. This observation is exemplified by the ASH1 mRNA from S. cerevisiae, which had an RPKM value of more than 200 in vesicles. This mRNA is not particularly frequent in total cellular mRNA and is produced preferentially during cell budding83. Other examples include mRNAs required for the synthesis of the fatty acid desaturase Ole1, the heat-shock proteins Hsp104, Hsp82 and Hsp70, the regulator of endochitinase (CTS1) expression Nce2m, glyoxalate pathway regulators, a putative estradiol-binding protein and a putative sterol transporter from P. brasiliensis. All these mRNAs are upregulated in yeast cells81,84, but all were absent from EV preparations. The transcripts encoding urease and the cell division control protein CDC42, which may be involved in the pathogenicity of P. brasiliensis81,84, were identified in the vesicles. In C. albicans, the most abundant vesicular mRNAs were the CYB5 transcript (RPKM > 12,000), which encodes cytochrome b and the RTT109 transcript (RPKM > 3,500), encoding histone acetyltransferase. These transcripts are present at a relatively low abundance in C. albicans cells85.

Table 4 mRNA sequences identified in P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae EVs

Gene ontology analysis revealed enrichment in RNA molecules relating to specific cellular processes in fungal EVs. Some of these processes, such as vesicle-mediated transport and metabolic pathways, were common to all four species analyzed in this study (Fig. 6). Other processes were limited to two species. For example, mycelium development and filamentous growth were restricted to C. albicans and C. neoformans (Fig. 6). Similarly, EV samples from C. neoformans and P. brasiliensis were enriched in mRNAs relating to cellular responses to stress, whereas those from C. neoformans and S. cerevisiae were rich in mRNAs involved in transcriptional regulation and those from P. brasiliensis and S. cerevisiae were enriched in mRNAs involved in cell cycle control (Fig. 6). These observations indicate that fungal EVs carry mRNAs common to different species, but they may also carry molecules restricted to one or two species. The biological significance of this finding is unknown.

Figure 6
figure 6

Gene Ontology (GO) function profile of proteins corresponding to the high-coverage RNA sequences identified in the sRNA-enriched fractions isolated from EVs from P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae.

The number of hits identified for each term is indicated on the x-axis.

We compared vesicular mRNA sequences from the four species studied, to identify sequences that were species-specific and characteristics common to EV preparations from S. cerevisiae, C. neoformans, P. brasiliensis and C. albicans. In C. neoformans, the most abundant mRNAs with sequences corresponding to known functions (RPKM > 1,000) were those encoding citrate synthase, ubiquitin-activating enzyme, succinate dehydrogenase, VpsA (vacuolar protein sorting A), L7 ribosomal protein, ATP-binding cassette transporter, cellulase and the AGC/AKT protein kinase associated with the TOR complex. In P. brasiliensis, the most abundant transcripts with annotated functions corresponded to cell division control proteins, translation initiation factor RLI 1 and nucleotide-binding proteins. Most of the mRNAs identified in C. albicans encoded hypothetical proteins, but those encoding cytochrome b5, histone acetyltransferase, cell division control protein (CDC43) and arginine biosynthetic process 1 were also identified. Only three of the transcripts in S. cerevisiae samples had a RPKM value greater than 1,000; the most abundant mRNA were those encoding Myb-related transcription factor, the Dam-1 complex associated with cell division, a cyclin involved in cell cycle progression, a pheromone-regulated protein associated with mating, a component of autophagosomes, pre-mRNA splicing and inhibitor of HO transcription.

Many of the mRNA sequences found in EV preparations had a higher abundance than expected from genomic data (Table 5). For instance, in S. cerevisiae, vesicular mRNAs from the cellular protein modification process (2.18-fold) and small molecule metabolic process (1.82-fold) categories were more frequent than would be expected given the relative distribution of the corresponding genes in the genome (Table 1). Similar findings were obtained for the categories chromatin repression by histone regulators (5-fold) and mitosis regulator (2-fold) in P. brasiliensis. In C. albicans, vesicle samples were enriched in sequences required for vesicle-mediated transport (2.43-fold), catabolic processes (2.1-fold) and transport (2.09-fold). The EVs of C. neoformans were enriched in mRNAs corresponding to the transport (3.61-fold), cytoskeleton organization (3.06-fold), signal transduction (2.37-fold) and vesicle-mediated transport (1.94-fold) categories (Table 5).

Table 5 Gene Ontology (GO) function profile from P. brasiliensis, C. neoformans, C. albicans and S. cerevisiae EVs. * p-value < 0.05

We then compared the RNA content of extracellular vesicles with that of other vesicles from the species studied. The analysis was performed with OrthoMCL software86, which uses the Markov Clustering (MCL) algorithm designed to group orthologous sequences on a genome-wide scale. It identifies molecules orthologous in different species and functionally redundant or “recent”, paralogs within species. Most of the RNAs identified were species-specific: 70% were exclusive to S. cerevisiae, 89% to P. brasiliensis, 91% to C. neoformans and 92% to C. albicans (Fig. 7). None of the mRNA sequences were common to all four fungi, but we identified orthologs in two or three species. The largest numbers of orthologs were found in C. neoformans and S. cerevisiae (16) and in C. neoformans and C. albicans (11 RNAs). The small numbers of orthologs common to several species may reflect the characteristics of each species and of the RNA cargo, which plays a specific role in each fungus. This result differs from the observations made during proteomic comparisons of the vesicles, for which a core set of proteins was found to be common to different fungal species3.

Figure 7
figure 7

Venn diagram of orthologous RNA sequences identified in the sRNA-enriched fractions isolated from the extracellular vesicles of P. brasiliensis (n = 73), C. neoformans (n = 353), C. albicans (n = 253) and S. cerevisiae (n = 84).


We characterized EVs as biological carriers of fungal RNA. The implications of these findings for fungal physiology and pathogenesis remain unclear, but the presence of RNA in fungal EVs is consistent with the hypothesis that the vesicle-mediated export of nucleic acids interferes with gene expression in host and fungal cells. The presence of different RNA classes in fungal EVs opens up new possibilities for investigating the ways in which fungal cells communicate and respond to external stimuli. Our results suggest that the RNA exported by fungi may interfere with the physiology of host tissues.


Fungal strains and growth conditions

The P. brasiliensis isolate used in our study was Pb18, which was recovered from the lungs of BALB/c mice after experimental infection. Eight-week-old male Balb/c were obtained from the Center for Development of Experimental Models (CEDEME), at Universidade Federal de São Paulo (UNIFESP). Animals were kept in ventilated isolators at 22 ± 2 °C and 55 ± 10% humidity. All animal manipulations were made in compliance with the protocols approved by Federal University of São Paulo Ethics Committee for Animal Experimentation (project approval number 366/07). Pb18 cells were maintained in the yeast phase at 36 °C, in solid, modified YPD medium (0.5% yeast extract, 0.5% casein peptone, 1.5% glucose, pH 6.5). Yeast cells were transferred from seven-day-old slant cultures to Erlenmeyer flasks containing Ham's F12 medium (Invitrogen) supplemented with 1.5% glucose, in which they were cultured for four days at 36 °C, with shaking (pre-inoculum). Cells from four pre-inocula were then collected by gravitational sedimentation. As much supernatant as possible was removed and the cells were used to inoculate fresh medium (500 ml). They were cultured at high density for another 48 h, for vesicle purification. The C. neoformans isolate was the standard strain H99 (serotype A), which was maintained at 30 °C in Sabouraud dextrose plates (1% dextrose, 4% peptone). Yeast cells recovered from the plates were used to inoculate minimal medium composed of dextrose (15 mM), MgSO4 (10 mM), KH2PO4 (29.4 mM), glycine (13 mM) and thiamine-HCl (3 μM) and were cultured for three days at 30 °C, with shaking. S. cerevisiae (strain SEY6210) cells were maintained at 25 °C for 24 h, on Sabouraud dextrose plates. For EV isolation, yeast cells were transferred to Sabouraud dextrose broth and cultured for an additional 24 h at 25 °C, with shaking. Strain 11 of C. albicans was isolated from a male patient87 and kindly provided by Dr. Dornelas from the Institute of Hematology Arthur Siqueira Cavalcanti (Rio de Janeiro, Brazil). C. albicans cells were maintained at 36°C on Sabouraud dextrose plates. For EV isolation, yeast cells were grown in Sabouraud dextrose broth, with shaking, at 36°C for 48 h.

Vesicle isolation

EVs were isolated from fungal culture supernatants, as previously described10,16. Briefly, cell-free culture supernatants were recovered by centrifugation at 4,000 × g for 15 minutes at 4°C and the resulting supernatants were recentrifuged at 15,000 × g for 30 minutes, to remove smaller debris. The final supernatants were concentrated by a factor of 20 with an Amicon ultrafiltration system (100-kDa cutoff, Millipore). Concentrated supernatants were centrifuged at 15,000 × g for 30 minutes, to ensure the removal of aggregates and the resulting supernatant was then ultracentrifuged at 100,000 × g for 1 h to precipitate vesicles. Vesicle pellets were washed once in phosphate-buffered saline (PBS) and the final pellets were suspended in PBS and lyophilized for RNA isolation.

sRNA isolation

RNA was isolated with the RNeasy mini kit (Qiagen) and was then treated with the RNeasy MinElute Cleanup Kit (Qiagen) according to the manufacturer's protocol, to obtain small RNA-enriched fractions. We used an Agilent 2100 Bioanalyzer (Agilent Technologies) to detect small RNA (sRNA) molecules. For the confirmation of sRNA extraction, some extracellular vesicle preparations were treated with 30 U DNase I (Qiagen) and characterized with the Bioanalyzer. We checked that the RNA was confined within the extracellular vesicles, as previously described22. We added 5 μg cellular RNA or no cellular RNA to the Pb18 preparation, which was then treated with 0.4 μg μl−1 RNase (Promega) for 10 minutes at 37 °C. RNase was inactivated by incubation for 10 minutes at 65 °C, after which, RNA was extracted immediately.

RNA sequencing

We used 100 ng of purified sRNA for RNA-seq from three independent biological replicates. The RNA-seq was performed in a SOLiD 4 platform using RNA-Seq kit according to the manufacturer's recommendations.

In silico data analysis

The sequencing data obtained were analyzed with CLC Genomics Workbench© v 5.5.1. The reads were trimmed on the basis of quality, with a threshold Phred score of 15. The reference genomes used for mapping were obtained from the NCBI database (C. neoformans - GCA_000149245.3, C. albicans - GCA_000149445.2, P. brasiliensis - GCA_000150735.1 and S. cerevisiae - GCA_000146045.2). The alignment was performed as follows: additional 100-base upstream and downstream sequences; minimum number of reads, 10; maximum number of mismatches, 2; nonspecific match limit, −2 and minimum length fraction of 0.9 for the genome mapping and 1.0 for the miRNA mapping. The minimum similarity of the reads mapped on the reference genome was 80%. Only uniquely mapped reads were considered in the analysis. The libraries were normalized per million and the expression values for the transcripts are presented in RPKM (Reads Per Kilobase per Million).

Quantitative PCR (qPCR)

DNA and RNA were extracted from P. brasiliensis yeast cells recovered from EV preparations, as previously described88. Due to the absence of control genes in EV preparations, a standard curve based on amplified DNA fragments was constructed, for rough quantification of the target sequences in P. brasiliensis RNA isolated from yeast cells and EVs. Sequences were initially amplified from isolated genomic DNA with primers for PADG_00645 (F:5′-GGATCTGAAAGGCACCCGAA-3′ and R:5′-ACCCATACGCTCTACCCCTT-3′) and PADG_00901 (R:5′-CCGATGCCAGCACTACTACC-3′; F:5′-CCATGACATCGACACCGGAA-3′). The reaction mixture (50 μl) contained 1U of Taq High-Fidelity Platinum polymerase (Invitrogen), 2 mM MgSO4, 0.4 mM of each primer and 0.2 mM of dNTP in High-Fidelity PCR buffer. Amplifications were performed as follows: 5 minutes at 94 °C, followed by 35 cycles of 94 °C for 1 minutes, 60 °C for 1 minute and 70 °C for 1 minute, followed by an extension phase of 7 minutes at 72°C. The final product was purified with gel electrophoresis and the QIAEX II Gel Extraction kit (Qiagen) after the excision of amplicons from the 2.5% agarose gel. Amounts were estimated in a Nanodrop device and the number of molecules was estimated with a tool available from site. We carried out qPCR with the Fast SYBR Green Master Mix (Applied Biosystems), according to the manufacturer's instructions, in a 7500 Fast Real-Time PCR System (Applied Biosystems). Cycling conditions were as follows: 2 minutes at 50 °C, 1 minute at 95 °C, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 s. An additional cycle of 95 °C for 15 s, 60 °C for 20 s and 95 °C for 15 s was performed, to determine the dissociation curve at the end of the reaction. The standard curve for absolute quantification from genomic DNA, based on decreasing known numbers of copies (106, 105, 104, 103 and 102 theoretical copies), was constructed in parallel, from total cell and EV RNA reactions. Measurements were made in triplicate.

Data access

The data have been submitted to the Sequence Read Archive (SRA) database under study accession number (SRA: SRP022849).