UBC Faculty Research and Publications

Small RNAs derived from tRNAs and rRNAs are highly enriched in exosomes from both old and new world Leishmania… Lambertz, Ulrike; Oviedo Ovando, Mariana E; Vasconcelos, Elton J; Unrau, Peter J; Myler, Peter J; Reiner, Neil E Mar 5, 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52383-12864_2015_Article_1260.pdf [ 2.4MB ]
Metadata
JSON: 52383-1.0223558.json
JSON-LD: 52383-1.0223558-ld.json
RDF/XML (Pretty): 52383-1.0223558-rdf.xml
RDF/JSON: 52383-1.0223558-rdf.json
Turtle: 52383-1.0223558-turtle.txt
N-Triples: 52383-1.0223558-rdf-ntriples.txt
Original Record: 52383-1.0223558-source.json
Full Text
52383-1.0223558-fulltext.txt
Citation
52383-1.0223558.ris

Full Text

RESEARCH ARTICLE Open AccessSmall RNAs derived from tRNAs and rRNAs aremes from both old andderived almost exclusively from non-coding RNAs. These exosomes are competent to deliver their cargo of novel,Lambertz et al. BMC Genomics  (2015) 16:151 DOI 10.1186/s12864-015-1260-7British Columbia, Vancouver, BC, CanadaFull list of author information is available at the end of the articlepotential small regulatory RNAs to macrophages where they may influence parasite-host cell interactions. Theremarkably high degree of congruence in exosomal RNA content between L. donovani and L. braziliensis, arguesfor the presence of a conserved mechanism for exosomal RNA packaging in leishmania. These findings open up anew avenue of research on non-canonical, small RNA pathways in this trypanosomatid, which may elucidatepathogenesis and identify novel therapeutic approaches.Keywords: Leishmania, Exosomes, Shuttle RNA, Small non-coding RNA, tRNA-derived small RNA* Correspondence: ethan@mail.ubc.ca†Equal contributors1Departments of Medicine, Microbiology and Immunology, University ofConclusions: These results show that leishmania exosomnew world Leishmania providing evidence forconserved exosomal RNA PackagingUlrike Lambertz1†, Mariana E Oviedo Ovando2†, Elton JR Vasconcelos3, Peter J Unrau2, Peter J Myler3,4and Neil E Reiner1*AbstractBackground: Leishmania use exosomes to communicate with their mammalian hosts and these secreted vesiclesappear to contribute to pathogenesis by delivering protein virulence factors to macrophages. In other eukaryotes,exosomes were found to carry RNA cargo, such as mRNAs and small non-coding RNAs, capable of altering recipientcell phenotype. Whether leishmania exosomes also contain RNAs which they are able to deliver to bystander cellsis not known. Here, we show that leishmania exosomes indeed contain RNAs and compare and contrast the RNAcontent of exosomes released by Leishmania donovani and Leishmania braziliensis.Results: We purified RNA from exosomes collected from axenic amastigote culture supernatant and found thatwhen compared with total leishmania RNA, exosomes mainly contained short RNA sequences. Exosomes withintact membranes were capable of protecting their RNA cargo from degradation by RNase. Moreover, exosomeRNA cargo was delivered to host cell cytoplasm in vitro. Sequencing of exosomal RNA indicated that the majorityof cargo sequences were derived from non-coding RNA species such as rRNA and tRNA. In depth analysis revealedthe presence of tRNA-derived small RNAs, a novel RNA type with suspected regulatory functions. Northern blottingconfirmed the specific and selective enrichment of tRNA-derived small RNAs in exosomes. We also identified anumber of novel transcripts, which appeared to be specifically enriched in exosomes compared to total cell RNA.In addition, we observed the presence of sequences mapping to siRNA-coding regions in L. braziliensis , but not inL. donovani exosomes.es are selectively and specifically enriched in small RNAshighly enriched in exoso© 2015 Lambertz et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly credited. The Creative Commons Public DomainDedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article,unless otherwise stated.Lambertz et al. BMC Genomics  (2015) 16:151 Page 2 of 26BackgroundProtozoan parasites of the genus Leishmania are highlyendemic to tropical and sub-tropical regions of theworld. They are transmitted to humans and other mam-mals by sandfly vectors that inject the flagellated, pro-mastigote life cycle stage of leishmania into the dermisof the host while taking a blood meal. After innocula-tion, promastigotes are engulfed by host mononuclearphagocytes either directly or indirectly as cargo of apop-totic neutrophils [1]. Following their ingestion by hostcells, promastigotes take up residence in the phagoly-sosome, where they transform into amastigotes andundergo cell proliferation. Depending on the infectingleishmania species, disease manifestations and symptomscan vary widely from mild self-healing cutaneous lesionsto lethal visceral disease. The two species that are thefocus of the present study, Leishmania donovani andLeishmania braziliensis, cause visceral and mucocutane-ous leishmaniasis, respectively. While the former isnaturally the more serious threat as it can lead to deathif left untreated, the latter can have an extremely highimpact on the affected individual due to debilitating anddisfiguring destruction of critical soft tissue structures.The current paucity of effective and well tolerateddrug treatments and even more so, the lack of highlyefficacious, well standardized and widely available vac-cination strategies can be attributed at least in part tothe gap of knowledge about the intricate interplay be-tween leishmania and host macrophages. Macrophagesare universal key players in both innate and adaptiveimmune responses. Their primary function is to engulfand digest prey, whether pathogen or debris from cellularturnover, which makes their intracellular environmentvery nutrient-rich. Leishmania exploits these macrophagecharacteristics in a very sophisticated manner: it lets thephagocyte ingest it, and then uses the cell as its safenursery, where it scavenges nutrients and replicateswhile remaining unrecognized by other immune cells.The mechanisms by which leishmania manages tosurvive within these potent immune cells are just startingto be elucidated. One key strategy employed by leishmaniaappears to be the prevention of macrophage activation,a step that is crucial to induce macrophage digestionand killing functions [2-4]. At the same time, leishmaniaare resistant to the harsh conditions of the acidifyingphagolysosome [5].In principle, there are two categories of molecules –surface associated and secreted- made available byleishmania to communicate with the host and turnon and off macrophage cellular functions. Regardingsecreted molecules, our group has recently discovered thatleishmania use a non-classical secretion mechanism to ex-port a majority of their secreted proteins, which involvesthe release of small vesicles called exosomes [6,7].Exosomes are 50–100 nanometre-sized membranevesicles secreted by a variety of single- as well as multi-cellular eukaryotic organisms. They are distinct frommembrane microvesicles, which are produced by bleb-bing, since their release occurs through fusion of multi-vesicular bodies from the endocytic/exocytic pathwaywith the plasma membrane of the cell [8]. Extracellularvesicles such as microvesicles and exosomes had longbeen considered to be simply cellular garbage bags. Onlyrecently has the release of specific cargo within vesicles,as well as their uptake and effects on recipient cells,been appreciated to represent important biologicalevents. Extracellular vesicle release has also been docu-mented in the context of infection, where the vesicleswere shown to contain both host and pathogen-derivedantigens and virulence factors (reviewed in [9]). Extracel-lular vesicles containing pathogen derived factors, maybe released either by infected cells, as has been shownfollowing infection with Eppstein-Barr virus, mycobac-teria, toxoplasma or plasmodia [10-13], or released bythe pathogen directly, e.g. mycobacteria, cryptococci,trypanosoma and leishmania [7,14-17].Importantly, in our studies, L. donovani exosomesand exosomal proteins were detected in the cytosoliccompartment of infected macrophages [7]. Moreover,we showed that L. donovani exosomes can modulatemononuclear cell phenotypes in vitro, rendering themanti-inflammatory by specifically inhibiting cytokineproduction. Studies with C57Bl/6 and Balb/c mice pro-vided evidence that treatment with exosomes fromL. donovani as well as Leishmania major prior to infec-tion exacerbated disease in vivo [18]. These findingshave fundamentally transformed our understanding ofhow leishmania are able to communicate with the host.Two other studies have since supported a role for exo-somes in leishmania pathogenesis. In the first one, theauthors showed that the metalloprotease GP63 deliv-ered by L. donovani exosomes cleaved the microRNA(miRNA) processing nuclease Dicer 1 in murine hepato-cytes, resulting in downregulation of microRNA-122 ex-pression, lowering of serum cholesterol and enhancementof murine liver infection [19]. In a second study, anothergroup looking at L. major exosomes reported that the vesi-cles globally affected macrophage gene expression, whichwas in part GP63-dependent [20]. In summary, these re-sults make a strong case for the importance of exosomesin leishmania pathogenesis.In addition to their protein cargo, exosomes andmicrovesicles were recently shown to be carriers ofnucleic acids in the form of RNA. This observation wasfirst made in mast cell exosomes, which were found tocontain mRNA as well as miRNA [21]. Surprisingly,these molecules were functional and could transducesignals in recipient cells. Since then, exosomal RNAsLambertz et al. BMC Genomics  (2015) 16:151 Page 3 of 26have been implicated in the pathogenesis of a variety ofimportant, chronic infections. For example, Eppstein-Barr virus-infected B-cells were shown to release exo-somes containing viral miRNAs which could regulategene expression in recipient cells [10]. Toxoplasma gon-dii-infected fibroblasts released exosomes containing aset of host mRNAs and miRNAs that was distinct fromthat of uninfected, serum-starved cells [12]. However, todate only two protozoan pathogens have been found torelease RNA-containing extracellular vesicles directly.Thus, Trichomonas vaginalis exosomes were reported tocontain RNA sequences, the biotype and function ofwhich still remain to be determined [22], and Trypano-soma cruzi was shown to release extracellular microvesi-cles containing a variety of non-coding RNAs includingtRNA-derived small RNAs, which have a suspectedregulatory nature [16,23].Based on the evidence that exosomes may serve asbiologically important shuttle vectors for RNAs, in thepresent study, we sought to investigate the RNA contentof leishmania exosomes. We indeed found that leish-mania exosomes contained RNA cargo which they werecapable of delivering to host cells in vitro. Using highthroughput sequencing and bioinformatics analyses, wefound that leishmania exosomes were enriched in smallRNAs derived from largely non-coding RNAs. Notably,we discovered that these vesicles contained a relativelyabundant and highly selective population of small RNAsderived from mature tRNAs. Furthermore, we found anumber of novel transcripts, some of which were highlyenriched in exosomes. Although exosomes released byboth L. donovani and L. braziliensis had largely similarRNA content, L. braziliensis exosomes specifically con-tained transcripts derived from genes that also code forsiRNAs.Taken together, these findings show for the first timethat leishmania exosomes are highly enriched in smallnon-coding RNAs, particularly tRNA-derived small RNAswith potential regulatory functions. This suggests thatthese RNAs may have functions in intercellular communi-cation. These findings hint at a previously unrecognizedpotential mechanism of leishmania pathogenesis, medi-ated through the exosomal delivery of small, principallynon-coding RNAs to mammalian host cells.ResultsL. donovani and L. braziliensis exosomes contain shortRNA sequences; and intact vesicles protect their RNAcargo from degradationWe have previously reported that leishmania use anexosome-based secretion mechanism in order to exportproteins with potential virulence properties [6,7]. Basedon a number of studies in mammalian systems demon-strating the presence of RNA in exosomes, we wereencouraged to expand on our findings and examine theRNA content of leishmania exosomes. We performed allexperiments for this study with exosomes purified fromsupernatants of L. donovani or L. braziliensis culturedin vitro under infection-like stressors (acidic pH and ele-vated temperature for 24 h, see Methods), which inducethe cells to transform into amastigotes. We had previ-ously observed that these “early” axenic amastigotesrelease increased quantities of exosomes enriched in spe-cific virulence factors [7]. Moreover, while undergoingtransformation into amastigotes, leishmania modulatecritical macrophage processes to allow for establishmentof chronic infection. Combined with the fact that amas-tigotes are literally the only life cycle stage found in vivoin the mammalian host once infection is established, wefelt that exosomes purified from this life cycle stage werethe most relevant to examine.Exosomes were purified from supernatants of earlyaxenic amastigotes and subjected to RNA extractionwith phenol/chloroform. Results depicted in Figure 1Aand B show that L. donovani axenic amastigote exo-somes contained significant amounts of RNA that weredetectable with the Agilent Bioanalyzer. Quantificationwith nanodrop revealed an average yield of 12.5 ng ofRNA per μg of exosomal protein (data not shown).Notably, the length profile of exosomal RNA wasdistinct from that of L. donovani total RNA, with thebulk of exosomal sequences being short (25–250 nt).Furthermore, we did not detect full length ribosomalRNA (rRNA) peaks in exosome RNA profiles. In con-trast, these full length rRNA peaks were prominent inthe total RNA profiles. To confirm that the purified nu-cleic acid was in fact RNA, we incubated exosome RNAwith DNase, RNase or KOH. As can be seen in Figure 1C,exosomal RNA was resistant to treatment with DNase,but was completely degraded upon exposure to eitherRNase or KOH.To exclude the possibility that RNA was merely co-purified during exosome isolation but was not directlyassociated with or internal to the vesicles, we treated in-tact exosomes with RNase in the presence or absence ofmembrane-permeabilizing detergent. The results inFigure 1D show that when treated with RNase alone,exosome RNA remained intact. In contrast, when exo-somes were treated with RNase and TritonX-100 simul-taneously, the RNA signal was greatly diminished. Thesefindings suggested that the RNA was confined withinthe exosomal membrane and thereby protected fromdegradation. The fact that we still saw a small residualsignal after detergent and RNase treatment could indi-cate that a fraction of the RNA was bound to RNA-binding proteins and was thereby protected.To investigate whether the release of RNA withinexosomes is conserved between leishmania species, weLambertz et al. BMC Genomics  (2015) 16:151 Page 4 of 26purified and analyzed RNA from exosomes released byL. braziliensis early axenic amastigotes using the sameprocedures as described for L. donovani. This analysisshowed that L. braziliensis exosomes also contain RNA,Figure 1 L. donovani exosomes contain RNA cargo. Exosomes were pudescribed in the Methods. RNA was extracted from exosomes or whole ceBioanalyzer RNA length profiles of exosome RNA alongside total RNA (~10Bioanalyzer measurement, C. Purified exosome RNA (~250 ng/sample) was eitradiolabelling with γ32P dATP and separation on a denaturing 15% polyacrylaRNA extraction, intact exosomes (purified from 400 mL culture supernatant) wAs a control for RNase A activity, 1 μL of the Agilent pico ladder was treated wRNA extraction and run on the Agilent Bioanalyzer. Arrowhead indicates inter3 independent experiments.with similar characteristics to that of L. donovani exo-some RNA (Additional file 1: Figure S1A and S1B).Taken together, these data represent the first descriptionof RNA released by leishmania within exosomes.rified from L. donovani axenic amastigote culture supernatant aslls by phenol-chloroform extraction and then analyzed. A. Agilent0 ng RNA were loaded for each), B. Gel-like image from Agilenther left untreated or treated with DNase I, RNase A or KOH followed bymide gel, D. RNA inside exosomes is resistant to degradation. Prior toere either left untreated, or treated with RNase A or TritonX-100 or both.ith the same concentration of RNase A. Samples were then subjected tonal 25 nt marker. nt, nucleotides. All images are representative of at leastLeishmania exosomes deliver RNA cargo to humanmacrophagesIn our previous studies, we observed that leishmaniaexosomes were released into infected macrophages andwere taken up by uninfected bystander cells, and thatexosomal proteins were delivered to host macrophagecytoplasm [7]. In order to investigate the potential deliv-ery of exosomal RNA cargo to host macrophages, welabelled exosomes purified from the supernatant ofL. donovani early amastigotes with an RNA specificfluorescent dye. Size and homogeneity of exosomes wasassessed by Nanosight analysis (see Figure 2A) and themedian size was determined to be 120 nm. Fluorescence oflabeled exosomes was confirmed by microscopy (Additionalfile 2: Figure S2). PMA-differentiated THP-1 cells wereincubated for 2 hours with fluorescently labelled exo-somes and uptake was assessed by flow cytometry andconfocal microscopy. As shown in Figure 2B, weobserved a dose-dependent increase in fluorescence ofcells, suggesting that macrophages readily take up exo-somes and their RNA cargo. In contrast, control cellsincubated at 4°C to inhibit phagocytosis showed onlybackground fluorescence. To exclude the possibility thatexosomes were just bound to the macrophage membranebut not internalized after incubation, we examinedexosome-treated cells by confocal microscopy. Figure 2Cshows that the fluorescence was localized to the cytoplasmof the macrophages and not to the membrane, indicatingthat the exosomes containing RNAs were indeed taken upby the cells. These results confirm that leishmania exo-somes and their RNA cargo can be internalized by hostcells and can access their cytoplasm.Characterization of leishmania exosome RNA cargo:Exosomes are enriched in small non-coding RNAs derivedfrom tRNAs and rRNAsIn order to assess the global transcriptome present inleishmania exosomes, we constructed complementaryDNA libraries for high-throughput sequencing. Wechose to compare RNA purified from exosomes releasedby early axenic amastigotes of L. donovani and L. brazi-liensis for three reasons: a) these two organisms causedistinct disease manifestations and hence can be ex-pected to differ in their mechanisms of pathogenesis; b)they are spread through different vectors: L. donovani istransmitted by sandflies of the genus Phlebotomus in theOld World, whereas L. braziliensis is transmitted by Lut-zomyia in the New World; and c) L. braziliensis wasesexencI) o. ClexLambertz et al. BMC Genomics  (2015) 16:151 Page 5 of 26Figure 2 Exosomal RNA cargo is delivered to macrophages. Exosomamastigotes, protein concentration was determined by Micro BCA, andPMA-differentiated THP-1 cells were incubated for 2 hours with labelledsize profile of purified exosomes, B. Cells were treated with different cocytometry. Histograms were drawn, median fluorescence intensity (MFwas calculated (error bars represent standard error of the mean, SEM). C(green) at 4°C (left) or 37°C (right). Cells were stained with phalloidin-Amicroscopy was done with a Leica DMIRE2 inverted microscope equippedmagnification objective. Images are representative of 3 independent expewere purified from 400–800 mL supernatant of L. donovani axenicosomes were stained with a green fluorescent RNA-specific dye.xosomes at either 37°C or 4°C to inhibit phagocytosis. A. Nanosightentrations of labelled exosomes as indicated and analysed by flowf cells was recorded, and the mean of 3 independent experimentsonfocal microscopy of cells incubated with 10 μg/mL exosomesa 594 to detect actin (red) and DAPI to detect nuclei (blue). Confocalwith a SP2 AOBS laser scanning head. Images were taken with a 63×riments.found to have a functional RNA interference pathway,which seems to be absent in L. donovani [24]. We hy-pothesized, therefore, that these two organisms coulddiffer in their composition of exosomal RNA and choseto examine this directly. We used a strategy for libraryconstruction that was optimized for sequencing ofsmall RNAs, as we had observed by gel electrophoresisLambertz et al. BMC Genomics  (2015) 16:151 Page 6 of 26that the exosomal RNA sequences were mainly short(Figure 1 and Additional file 1: Figure S1). We also in-corporated a series of enzymatic treatments includingdephosphorylation with calf intestinal alkaline phosphat-ase (CIP), 5′ cap removal with tobacco acid phosphatase(TAP) and 5′ re-phosphorylation with polynucleotidekinase (PNK) into the library construction procedure inorder to pick up all sequences present in the exosomaltranscriptome regardless of their 5′ modification (seeMethods). Sequencing of the libraries by paired end150 bp MiSeq Illumina sequencing resulted in ~1.4 millionpaired reads for L. donovani and ~1.1 million paired readsfor L. braziliensis (Table 1). After adapter trimming andadjustment of the orientation of all reads to correspondto that of the original RNA sequence, reads were col-lapsed into unique reads prior to further analysis. Asshown in the histograms in Figure 3A, read length dis-tributions of reads were clearly skewed towards shorterreads with the mean read length being 55 nt for L. dono-vani and 57 nt for L. braziliensis (medians 37 nt and49 nt, respectively).To get a general overview about what types of RNAtranscripts were represented in our libraries, we alignedthe reads of the L. donovani and the L. braziliensislibraries with reference genomes, respectively LdBPK(Leishmania donovani strain BPK282A1) and LbrM(Leishmania braziliensis MHOM/BR/75/M2904) usingBowtie 2 and the very-sensitive-local option which setsthe seed length to 20 nucleotides, allowing for only onemismatch within the seed alignment (see Methods). Wewere able to align 58.61% of reads from the L. donovanilibrary with the LdBPK reference genome and 22.87% ofreads from the L. braziliensis library with the LbrMreference genome (see Additional file 3: Table S1A andS1B for the full datasets). The comparatively low align-ment rate especially in case of the L. braziliensis libraryis likely a result of incomplete assembly of the referenceTable 1 Sequencing statisticsL. donovani library L. braziliensis libraryTotal paired reads 1435277 1062571Collapsed/unique reads 688524 538034Unique and single copy 574049 421086Numbers of reads for L. donovani and L. braziliensis exosome RNA libraries asobtained by high-throughput sequencing. Reads were combined into uniquereads by collapsing all identical reads to one read for downstream analysis.This also revealed the number of reads that were present in the dataset as asingle copy.genome or the fact that we used a different L. brazilien-sis strain (a clinical isolate from the Peruvian Amazonregion) than the strain used to generate the referencegenome. This is supported by the fact that we were ableto align significantly more reads (52.88%) from the L.braziliensis library with the L. major reference genome(Leishmania major MHOM/IL/81/Friedlin, LmjF, seeAdditional file 3: Table S1B). Other possible causes forlow alignment rates could be misinterpretation of modi-fied nucleosides by the sequencer or RNA editing ofsequences prior to packaging into exosomes, making itdifficult to compare our transcriptomic data with theavailable reference genomes derived from DNA sequen-cing. RNA editing is a well-described process in leish-mania and other trypanosomatids (e.g. [25,26]).In order to ensure that our libraries were not contami-nated with unrelated nucleic acids, we performed aBLAST search of all reads that failed to align with eitherthe LdBPK, the LbrM or the LmjF reference genomes,against the NCBI nucleotide collection database (NCBI-NT). The results of this analysis showed that 28.1% ofreads from the L. donovani exosome library and 36.3%of reads from the L. braziliensis exosome library alignedto sequences in the NCBI-NT database (see Additionalfile 4: Table S2A and S2B). Of these, 4.93% of L. dono-vani and 4.17% of L. braziliensis aligned with other leish-mania genomes. The rest aligned with a promiscuousgroup of >6000 different plant, fungi, helminth and bac-teria species, several of which were plant pathogens orsoil inhabitants. Based on the observation that there wasno enrichment of any particular species and that overall,the majority of reads from both libraries aligned withleishmania genomes (in total 63.54% of reads of theL. donovani library and 57.05% of the L. braziliensis li-brary, see summary of alignment statistics in Additionalfile 5: Table S3), we concluded that we did not have acontamination issue that would impugn our data. Wethink that many, if not all, of the reads mapping tobacteria or helminth genomes are likely false positivehits. Thus, even though our alignment rates were some-what lower than we might have expected, we think thatour datasets are valid and large enough to draw mean-ingful conclusions about the exosomal RNA content.When categorizing reads into RNA biotypes based onreference genome annotations, we saw that for bothlibraries, the majority of reads were aligning with rRNAand tRNA genes, in the sense orientation (Figure 3B). Inaddition, a large number of reads mapped to non-annotated (intergenic) regions of the reference genomes(42.47% for L. donovani and 34.46% for L. braziliensis),which could potentially be novel transcripts. Interest-ingly, we only saw less than 4% of reads mapping to pro-tein coding genes (CDS) or spliced leader (SL) RNAgenes. These results indicated that the majority ofLambertz et al. BMC Genomics  (2015) 16:151 Page 7 of 26sequences present in the leishmania exosome transcrip-tome are derived from non-coding RNAs and intergenicregions, whereas sequences derived from mRNAs areunderrepresented.When working with the LdBPK and LbrM referencegenomes, we had to take into account that both are lim-ited in their annotations. Hence, it was not surprisingthat we found a large number of reads in our librariesmapping to intergenic regions. Whereas the annotationsof CDS are thought to be comprehensive in these ge-nomes, the assignments of SL RNAs as well as structuralnon-coding RNAs such as rRNAs, tRNAs, snRNAs andsnoRNAs are clearly lacking in completeness. Conse-quently, it has to be considered that the large number ofreads mapping to intergenic regions may not necessarilyall be novel transcripts, but could also have resultedfrom incomplete annotation of non-coding RNA typesin these regions. Keeping this in mind and still trying todissect what types of RNA sequences are highly repre-sented in exosomes, we decided to inspect in greaterdetail the alignment of the most abundant exosomalFigure 3 Sequencing of leishmania exosomal RNA reveals conserved RNRNA. Exosome RNA from L. donovani and L. braziliensis was purified and pA. Sequence length distribution of reads obtained from sequencing L. donaccording to their alignment with genomic features annotated in the L. d*SL RNA, spliced leader RNA. Numbers for reads mapping to SL RNA geneas these genes have currently only been annotated in this genome.sequences manually using the Artemis genome browsersoftware [27]. For this purpose, reads were clustered intounique regions of alignment and the regions wereranked by abundance (number of reads found per re-gion). Considering that L. major is the species with thebest assembled genome to date and presents the mostcomplete annotation of non-coding RNAs, we also per-formed alignments of L. major annotated non-codingRNAs with the LdBPK and LbrM reference genomes, inorder to identify non-annotated, non-coding RNA lociin our target genomes. The results of the screening usingArtemis showed that the top 20 most abundant readsfrom both libraries mapped to three RNA classes in thesense orientation: rRNA, tRNA and snRNA (Table 2).The high abundance of reads mapping to rRNA genesobserved in both libraries is in compliance with other re-cent reports on RNA types found in exosomes. Uponcloser inspection we saw that the majority of reads map-ping to rRNA genes were shorter fragments (medianlength 39 nt for the L. donovani library and 52 nt forL. braziliensis, see Additional file 6: Figure S3). We thenA cargo composed mainly of sequences derived from non-codingrocessed for high-throughput sequencing as described in Methods.ovani and L. braziliensis exosome libraries, B. Categorization of readsonovani and L. braziliensis reference genomes. CDS, coding sequence;s were obtained from alignment with the L. major reference genome,Table 2 Top 20 most abundant clusters of transcripts present in leishmania exosomesL. donovani L. braziliensisChr Coordinates of genomiclocusAnnotation No. of reads RNA biotype Chr Coordinates of genomiclocusAnnotation No. of reads RNA biotype27 1014367 1019133 LdBPK_27rRNA3 344191 28S rRNA 6 334041 334897 LmjF.27.rRNA.31 58906 28S rRNALdBPK_27rRNA4 LmjF.27.rRNA.34LmjF.27.rRNA.13LmjF.27.rRNA.22LmjF.27.rRNA.29LmjF.27.rRNA.31LmjF.27.rRNA.33LmjF.27.rRNA.4227 1019947 1021495 LdBPK_27rRNA6 132730 18S rRNA 00 463646 464099 LbrM.27.rRNA1 11166 18S rRNA15 312758 313248 LmjF.15.TRNAASP.01 65737 tRNA-Asp 15 324587 325394 LbrM.15.tRNA1 18229 tRNA-AspLmjF.15.TRNAGLU.01 tRNA-Glu LbrM.15.tRNA2 tRNA-GluLmjF.09.5SrRNA.02 5S rRNA LbrM.15.rRNA1 5S rRNALmjF.05.5SrRNA.01LmjF.15.5SrRNA.0124 715730 715801 LdBPK_24tRNA5 43207 tRNA-Asp 24 659346 659417 LbrM.24.tRNA5 30583 tRNA-Asp17 328838 328909 LdBPK_17tRNA1 42601 tRNA-Asp 17 296223 296604 LbrM.17.tRNA1 30523 tRNA-AspLbrM.17.tRNA2 tRNA-SerLbrM.17.tRNA3 tRNA-Ala24 658796 658976 LdBPK_24tRNA2 35611 tRNA-Gln 24 600448 600615 LbrM.24.tRNA2 8749 tRNA-Gln09 429809 430355 LdBPK_09tRNA6 29246 tRNA-Glu 09 395278 395779 LbrM.09.tRNA3 4467 tRNA-ValLmjF.09.TRNAARG.01 tRNA-Arg LbrM.09.tRNA4 tRNA-HisLmjF.09.TRNAVAL.02 tRNA-Val LbrM.09.rRNA1 5S rRNALmjF.05.5SrRNA.01 5S rRNA LmjF.05.5SrRNA.01LmjF.11.5SrRNA.03 LmjF.11.5SRRNA.03LmjF.21.5SrRNA.01 LmjF.21.5SrRNA.0209 403494 403565 LbrM.09.tRNA5 5226 tRNA-Glu31 495812 496115 LdBPK_31tRNA3 18528 tRNA-Glu 31 582437 582738 LbrM.31.tRNA2 4719 tRNA-GlyLbrM.31.tRNA3 tRNA-Glu27 1019543 1019804 LdBPK_27rRNA5 18473 5.8S rRNA11 156707 157038 LmjF.11.TRNAALA.01 15493 tRNA-Ala 11 63421 63678 LmjF.33.TRNAALA.01 3041 tRNA-AlaLmjF.36.TRNALEU.01 tRNA-Leu LmjF.11.TRNALEU.02 tRNA-LeuLambertzetal.BMCGenomics (2015) 16:151 Page8of26Table 2 Top 20 most abundant clusters of transcripts present in leishmania exosomes (Continued)27 1014054 1014340 LmjF.27.rRNA.47 14260 28S rRNALmjF.27.rRNA.4823 230438 230509 LdBPK_23tRNA9 13814 tRNA-Gly 23 216842 216916 LbrM.23.tRNA9 3679 tRNA-Gly36 1630332 1630403 LdBPK_36tRNA2 11529 tRNA-Gln05 360707 361335 LdBPK_05snRNA1 9971 snRNA 05 349991 350587 LbrM.05.rRNA1-1 6306 5S rRNALbrM.05.ncRNA1-1 ncRNA33 104560 104930 LdBPK_33tRNA1 9352 tRNA-Ala 33 105787 105859 LbrM.33.tRNA1 5730 tRNA-ArgLdBPK_33tRNA2 tRNA-ArgLdBPK_33tRNA323 229645 229857 LdBPK_23tRNA5 8487 tRNA-Leu 35 2472707 2472788 LbrM.35.tRNA4 2592 tRNA-LeuLdBPK_23tRNA6 tRNA-Thr16 445957 446028 LdBPK_161tRNA1 7916 tRNA-Gln 16 442089 442160 LbrM.16.tRNA1 8695 tRNA-Gln23 230585 230656 LdBPK_23tRNA10 7804 tRNA-Trp 23 216992 217063 LbrM.23.tRNA10 3657 tRNA-Trp21 430678 430798 LbrM.21.rRNA1 3273 5S rRNAReads were clustered into genomic loci based on Bowtie 2 alignments with reference genomes (as described in Methods) to identify the RNA biotypes that were most abundant in exosomes. The details of the top 20clusters with the highest numbers of reads falling into them are listed. Clusters of reads in the L. donovani library are listed in descending order of abundance, with the homologous cluster of reads in the L. braziliensislibrary given in the same row. Chr = chromosome number, annotation = annotation in reference genomes (LdBPK = L. donovani, LbrM = L. braziliensis or LmjF = L. major) followed by the gene name, No. of reads = number ofreads from the respective library falling into this cluster, RNA biotype = type(s) of RNA that is annotated in the reference genomes in this region.Lambertzetal.BMCGenomics (2015) 16:151 Page9of26looked for enrichment of specific rRNA genes withinour pool and saw that the majority of reads from both li-braries mapped to 28S and 18S rRNA genes (>90%,Additional file 7: Table S4). Furthermore, we investigatedthe position of alignment of reads within the variousrRNA genes, and found that for both libraries, readsaligned along the entire length of these genes (Additionalfile 7: Table S4). It was of particular interest to find a largenumber of reads mapping to tRNAs in both libraries, astRNA-derived small RNAs have recently been discoveredin T. cruzi [28,29], and these novel small RNAs arethought to participate in regulation of gene expression[23,30] (see below for a more detailed analysis of tRNA-derived small RNAs).Notably, to our surprise, the overlap of the RNA pro-files for L. donovani and L. braziliensis was striking.Thus, these parallel and independent RNA-seq replicatesprovide direct evidence for the reproducibility of ourdata.These findings suggested that the sequences mapping tothese regions corresponded to bona fide novel transcripts.Notably, we found homologous novel transcripts in bothlibraries, providing evidence that they are both conservedbetween species as well as packaged into exosomes. Whenoverlaying our sequencing data from the L. donovani exo-some RNA library with a recently sequenced L. donovanispliced leader (SL) RNA library (P. Myler, unpublisheddata), we observed that the genomic loci giving rise to ouridentified novel transcripts had SL sites in the 5′ regionupstream of them (see Figure 4 for 2 examples). This indi-cates that they might be processed by trans-splicing andare hence likely to be functional mature transcripts ratherthan promiscuous transcriptional by-products.We also searched for open reading frames (ORF)within the sequences of the novel transcripts to seewhether they have the potential to code for a protein orpeptide and found potential ORFs for the majority ofthem (Additional file 8: Table S5). However, when weriplistNamLambertz et al. BMC Genomics  (2015) 16:151 Page 10 of 26Exosomes carry putative novel transcriptsTo make sure we did not miss any important informa-tion amongst the group of less abundant reads, we ran-domly selected a number of less abundant reads fromboth libraries and inspected their alignment with the ref-erence genomes manually. Interestingly, we discoveredseveral reads mapping to intergenic regions at differentgenomic loci (Table 3). These intergenic regions wereneither annotated at those loci in any of the sequencedleishmania or trypanosome genomes, nor did they sharehomology to any known trypanosomatid gene (as assessedby performing BLAST searches on TriTrypDB and NCBI).Table 3 Intergenic regions coding for putative novel transcL. donovaniName Coordinates No. of reads No. of ORFLdBPK_301180_leftof 379397 380435 363 7LdBPK_291610_leftof 706385 707360 181 9LdBPK_360420_leftof 109068 109733 139 3LdBPK_363000_leftof 1183324 1183522 132 1LdBPK_362290_leftof 872006 872406 89 5LdBPK_313190_leftof 1452630 1452719 83 1LdBPK_040550_leftof 225336 230737 80 8LdBPK_131560_leftof 555895 556192 75 2LdBPK_364270_leftof 1570740 1570991 57 1LdBPK_366120_leftof 2270195 2272110 49 13LdBPK_330560_leftof 173136 173362 40 0LdBPK_366590_leftof 1903364 1906720 0 N.A.List of 12 intergenic regions identified with numbers of reads mapping to themhomologous genomic region in the L. braziliensis library given in the same row.plus the designation “_leftof”, indicating that the intergenic region is on the left siddirection. There is 75% overlap of intergenic regions coding for novel transcripts innot applicable.translated the ORFs and looked for homologies toknown proteins in the NCBI database using Blastp, wedid not obtain any hits.Based on the hypothesis that these novel transcriptscould have a role in regulation of gene expression ineither the mammalian or insect host or both, weperformed Bowtie2 alignments to the human and thevector (Lutzomyia and Phlebotomous) genomes tosearch for potential targets in these genomes, lookingfor complementarity. We obtained 60 hits for all of the1288 reads representing novel transcripts in the L. dono-vani library when searching against the human genomeand 15 hits when searching against the Phlebotomusts in exosomesL. braziliensisName Coordinates No. of reads No. of ORFLbrM.30.1240_leftof 394563 397304 317 13LbrM.29.1600_leftof 663702 664543 103 7LbrM.35.0480_leftof 132138 133299 185 5LbrM.35.3080_leftof 1176690 1176833 36 1LbrM.35.2400_leftof 892711 895894 29 28LbrM.31.3490_leftof 1508257 1510948 0 N.A.LbrM.04.0610_leftof 229925 233378 111 25LbrM.13.1200_leftof 433509 433481 0 N.A.LbrM.35.4310_leftof 1563654 1563940 29 2LbrM.35.6160_leftof 2242345 2244214 153 14LbrM.33.0550_leftof 184438 184751 19 1LbrM.35.6630_leftof 2438694 2438825 155 0ed by descending order of abundance in the L. donovani library, with thees are derived from the annotated genes adjacent to the intergenic regione of the annotated gene on the same strand, regardless of transcriptionalthe L. donovani and L. braziliensis libraries. ORF, open reading frame. N.A.,Figure 4 (See legend on next page.)Lambertz et al. BMC Genomics  (2015) 16:151 Page 11 of 26tem) wPKned fonoosicaLambertz et al. BMC Genomics  (2015) 16:151 Page 12 of 26genome (Additional file 9: Table S6A). For the 1137 readsthat comprise the novel transcripts in the L. braziliensislibrary, 25 hits were observed when searching against thehuman genome and 6 hits when searching against theLutzomyia genome (Additional file 9: Table S6B). How-ever, nearly all of these hits were determined to be innon-annotated regions of the respective genomes, im-plying – at least for the human genome - that there areno genes that could be regulated by the novel tran-scripts, based on the parameters applied in our analyses(perfect complementarity). Of note, the annotation ofthe vector genomes is a very recent effort and far fromcomplete. It is quite possible, therefore, that there are asyet non-annotated protein coding genes in the genomicregions where the novel exosomal transcripts aligned,which would have been missed, resulting in false nega-tive findings. Moreover, in most animals, regulatoryRNAs such as miRNAs have incomplete homology withtheir target sequences [31] and, therefore, our predic-tions based on perfect complementarity may havemissed some potential targets in the host genomes,again leading to false negative results. Unfortunately, asthe tools to predict RNA-RNA interactions at the levelof potential regulatory RNA-mRNA target pairs arefairly limited (generating a massive amount of ambigu-ous results when analysing large datasets), we were un-able to carry out a comprehensive host mRNA targetprediction with the novel transcripts that was inform-ative. We also performed an alignment search of thenovel transcripts against the databases of human andmouse miRNAs (mirbase.org), but failed to detect hom-ologous sequences. All of the novel transcripts we identi-(See figure on previous page.)Figure 4 Novel transcripts are found in leishmania exosomes. A. Arand a L. donovani spliced leader sequence library (P. Myler, unpublishedreads mapping to them. Top: intergenic region on chromosome 30 (LdB(LdBPK_291610_leftof). Light blue boxes on grey tracks are annotated gereads from the exosome libraries. “P” designates the regions that were usedesigned for novel transcripts found in exosomes, corresponding to the geadditional probe against a 5.8S rRNA (27rRNA5). L. donovani total (T) and exRNA (3 μg) were loaded in each lane. Sizes of bands on membranes as indand 21 nt size markers.fied were present in the sense orientation of transcriptionin leishmania, which implies that they are unlikely to bepresent in exosomes as double strands, which is a charac-teristic of canonical siRNAs and miRNAs.To validate the presence of the identified novel tran-scripts in exosomes and compare their expression in exo-somes with total leishmania RNA, we designed probes forNorthern blotting. We selected the two most abundantnovel transcripts identified in the L. donovani exosomelibrary, one of which was positioned in between the genes1170 and 1180 on chromosome 30 (LdBPK_301180_lef-tof) and the other in between the genes 1600 and 1610 onchromosome 29 (LdBPK_291610_leftof ) for probe de-sign. The regions that are complementary to the probesused are indicated in Figure 4A. We loaded equalamounts of L. donovani total and exosome RNA into apolyacrylamide gel, along with 21 nt and 150 nt sizemarkers. The results from the Northern blots showedthat both novel transcripts could be detected in totaland exosome RNA (Figure 4B). Of interest, we detectedbands of larger size that were present in both total andexosome RNA (which likely represent the primarytranscript), but where the signal appeared to be strongerin the exosomal RNA lane. In addition, we detectedbands of smaller size in the exosome RNA, that werecompletely absent in the total RNA. For comparison, wealso incubated a blot with a probe for 5.8S rRNA(LdBPK_27rRNA5, Ref ), which appeared to be muchmore abundant in total than in exosome RNA. Theseresults confirm the presence of the novel transcriptsidentified by sequencing in exosomes and indicate thatfragments of these transcripts with specific lengths areuniquely present in exosomes, consistent with selectivepackaging.L. braziliensis exosomes carry a low abundance of sequencesderived from siRNA-coding regionsAs discussed above, L. braziliensis can regulate gene ex-pression through the RNAi pathway and produce smallinterfering RNAs (siRNAs) [32]. We were interested inexploring the possibility to find siRNAs as part of L. bra-ziliensis exosomal RNA cargo. The main classes of L.braziliensis siRNAs are derived from the spliced leader-associated conserved sequence (SLACS) retroposon [33],is genome browser alignments of the L. donovani exosome libraryith the L. donovani reference genome. Shown are two regions with_301180_leftof), bottom: intergenic region on chromosome 29s. In dark blue are the reads from the spliced leader library, in red ther designing probes for Northern blotting. B. Northern blots with probesmic regions shown in panel A (301180_leftof and 291610_leftof), plus anome (E) RNA were probed on the same membrane. Equal amounts ofted are in nucleotides (nt) and were calculated based on 262 nt, 150 ntthe telomere-associated transposable element (TATE)[33], the L. braziliensis-specific telomere-associated se-quence (TAS) [34] and the chromosomal internal re-peats, 74-nucleotide long (CIR74) [32]. We generated aBLAST database from a FASTA file with 41 SLACS/TATEs extracted from TriTrypDB-5.0_LbraziliensisM-HOMBR75M2904_ AnnotatedTranscripts.fasta compris-ing the nucleotide sequences of the SLACS and TATEsgenetic elements and performed a BLAST search withour libraries against this database. In the L. braziliensislibrary, we found 4471 reads mapping to these elements(Additional file 10: Table S7). Interestingly, about 50% ofLambertz et al. BMC Genomics  (2015) 16:151 Page 13 of 26these reads were both sense and antisense, suggestingthat the sequences were present in exosomes as double-stranded RNAs. The lengths of reads were somewhatheterogeneous, ranging from 20 nt to 70 nt, whereasL. braziliensis mature siRNAs (LbrAGO1-bound) are be-lieved to be 20–25 nt in length [32]. For comparison, wealso performed the same BLAST analysis with the L.donovani library, where we only found 353 reads map-ping to the siRNA-coding genetic elements (Additionalfile 10: Table S7). The fact that we found some reads inthe L. donovani library mapping to these elements couldbe due to settings used for the BLAST search that werenot stringent enough (cut off 80% identity and 70%query coverage), possibly resulting in false positive align-ments. Despite this, it is clear that there were >10 timesmore reads in the L. braziliensis library mapping tosiRNA-coding genetic elements, indicating that ourresults are specific and providing evidence that L. brazi-liensis may export siRNAs or their precursors withinexosomes. However, we cannot rule out the possibilitythat the sequences we found in exosomes may originatefrom regions of the SLACS/TATEs genes other than theones giving rise to siRNAs.Leishmania exosomes contain an abundance of specifictRNA-derived fragmentsRemarkably, we found a large number of reads in boththe L. donovani and L. braziliensis exosome RNA librar-ies that mapped to tRNA genes. A few recent studiescharacterizing the RNA content of mammalian exo-somes had reported the presence of tRNAs or their frag-ments in these vesicles. For example, reads mapping totRNAs were found in sequencing libraries made withRNA from exosomes released from neuronal cells(13.5%) [35], immune cells (~7%) [36] and plasma exo-somes (1.24%) [37]. Strikingly, in our datasets, 351,919reads (36.4 %) and 135,149 reads (21.1%) from L. dono-vani and L. braziliensis, respectively, mapped to tRNAgenes when aligned to the Leishmania major MHOM/IL/81/Friedlin (LmjF) reference genome (which has thebest curation on tRNA annotation amongst leishmaniaspecies). These frequencies exceeded by some measurethose reported for mammalian exosomes in the studiescited above. Close inspection of the genome alignmentsrevealed that a high percentage of these sequenceswere covering only parts of the respective tRNA genes(Figure 5), consistent with the occurrence of tRNA-derived small RNAs (tsRNAs), which has recently beenrecognized as a specific process. In light of these findingswe decided to characterize the reads mapping to tRNAs inmore detail. In case of both libraries, the vast majority(99.8%) of reads were in the sense direction of transcrip-tion. Looking at their length profiles, we found the meanread length to be slightly different between the twolibraries, 38 nt for L. donovani and 46 nt for L. braziliensis,however, the median read length was similar (33 nt and34 nt, respectively) (Figure 5A). For both leishmania li-braries, tsRNAs derived from tRNA-Asp, tRNA-Gln,tRNA-Glu and tRNA-Leu were most abundantlypresent (Figure 5B and Table 4). To make a case thatthese tsRNAs were specific cleavage products that areselectively packaged into exosomes and not just a by-product of tRNA turnover that is disposed by the cell,we calculated the Pearson’s correlation of the predictedcellular amino acid usage and the relative expression ofour tsRNAs as assessed by sequencing. The resultsshowed that there was no correlation (r = 0.163 for L.donovani and r = 0.114 for L. braziliensis), indicatingthat the tsRNAs were unlikely to be random degrad-ation products. Strikingly, we observed the same rankorder frequency of tRNA isotypes as origins of tsRNAsin both libraries (Figure 5B and Table 4), indicating thatthe formation of specific tsRNAs and their appearance asexosomal cargo is an evolutionary conserved phenomenonin leishmania.Next, we investigated whether the tsRNAs were de-rived from the 3′ or 5′ end of mature tRNAs, and foundthat the most abundant tsRNAAsp and tsRNAGln werederived from the 5′end or mid-5′ end in both libraries(Table 4). Notably, this was not the case for all of thetsRNAs, as some appeared to be derived from the 3′end, and two were not derived from the same end in thetwo different species (tsRNAGlu and tsRNAGly) (Table 4).Furthermore, we saw tsRNAs of different lengths, someof them corresponding to the length of tRNA-halves(~30 nt, e.g. tsRNALeu), and others to the length of tRNA-derived RNA fragments (25 nt, e.g. tsRNAIle) (Table 4).Based on the hypothesis that leishmania tsRNAs mayact as miRNAs or siRNAs in the mammalian or inverte-brate host we performed additional Bowtie2 alignmentswith all reads mapping to leishmania tRNAs against thehuman and vector genomes looking for complementarityto find potential targets for these potential regulatoryRNAs. This search yielded a very large number of hits(~20,000), the majority of which fell into non-annotatedregions of the host genomes (Additional file 9: Table S6Aand S6B). Analogous to our target prediction analyses withthe novel transcripts described above, we were unable toperform a more comprehensive (based on more complexRNA-RNA interactions rather than simple complemen-tarity alone) analysis of potential leishmania tsRNA-host target mRNA interactions due to the lack of appro-priate tools for large datasets. We also performed amiRNA homology search against the human and mousemiRNA database, and found only one miRNA (miR-135b-5p) that shared 88% identity with one of thetsRNAs present in both libraries (tsRNAArg) (Additionalfile 11: Table S8). These results indicate that leishmaniaFigure 5 tRNA-derived fragments are cargo of leishmania exosomes. A. Length distributions of reads mapping to tRNAs in L. donovani andL. braziliensis exosome RNA sequencing libraries. B. Bar graph showing percentages of reads from L. donovani (white bars) and L. braziliensis (greybars) mapping to the respective tRNA isoacceptors. C. tRNA secondary structures for leishmania tRNA-Asp and tRNA-Leu (downloaded from [38]).Arrowheads indicate major cleavage products as observed in the sequenced libraries. D. Northern blots with probes designed against tRNA-Asp(Asp1) and tRNA-Leu (Leu1 and Leu2). L. donovani total (T) and exosome (E) RNA were probed on the same membrane. Equal amounts of RNA (3 μg)were loaded in each lane. Asp1 and Leu1 probes were designed to specifically detect full length tRNA as well as t-RNA-derived small RNAs seen insequencing libraries, whereas Leu2 was designed against the mid region (anticodon loop) of tRNA-Leu and hence only detects the full length tRNA.Lambertz et al. BMC Genomics  (2015) 16:151 Page 14 of 26on′′′′Lambertz et al. BMC Genomics  (2015) 16:151 Page 15 of 26Table 4 Reads mapping to tRNAsL. donovani librarytRNA % of total tRNA reads Length [nt]A PositiAsp 34.91 47 5′Gln 16.00 32 mid-5Glu 11.48 38 3′Leu 8.80 29 5′Gly 8.02 37 3′Arg 5.40 31 mid-5Ala 4.03 34 3′Trp 2.38 37 mid-3Val 2.24 36 5′Thr 1.33 31 3′His 1.08 39 mid-3Tyr 1.06 31 3′Ser 1.02 34 3′Pro 0.80 31 3′Ile 0.44 25 3′tsRNAs are not highly similar to canonical mammalianmiRNAs.In order to validate the presence of the identifiedtsRNAs in exosomes and to compare their abundance inexosomes with total RNA, we performed Northern blot-ting with probes specific for tsRNAAsp (Asp1) andtsRNALeu (Leu1). The probes were designed to be com-plementary to the 5′ end of each tRNA and hencerecognize both full length tRNA and 5′ tsRNA. We alsoincluded a probe that was designed against the middleregion (anticodon loop) of tRNA-Leu for comparison(Leu2). When hybridizing blots of RNA isolated fromL. donovani total cells and versus exosomes with theseprobes, we detected a common band corresponding tothe full length tRNA in total and exosome RNA in caseof both the probes Asp1 and Leu1 (72 nt and 82 nt re-spectively) (Figure 5D). In addition, we detected twosmaller bands in the exosome RNA lane of the blotprobed with Asp1 which were absent in the total RNAlane. Similarly, we detected a smaller band in the exo-some RNA lane of the blot probed with Leu1 which wasabsent in the total RNA lane. These results demonstratethat 5′ tsRNAs are produced from tRNA-Asp andCys 0.34 25 midPhe 0.32 36 3′Lys 0.24 36 3′Met 0.10 29 3′Asn 0.01 33 mid-5′AAverage read length.BMost abundant read position (of all tRNA reads).Distribution of reads from L. donovani and L. braziliensis libraries over different tRNAlibrary, with the equivalent reads from the L. braziliensis library in the same row. ntL. braziliensis libraryB % of total tRNA reads Length [nt]A PositionB43.51 58 5′13.73 36 mid-5′9.65 46 5′7.66 31 5′3.11 43 5′9.62 28 mid-5′2.21 34 3′2.42 40 mid-3′1.28 41 5′1.49 35 3′1.99 42 mid-3′0.38 28 3′0.82 39 mid-3′1.02 34 3′0.31 31 3′tRNA-Leu and that these tsRNAs are highly enriched inexosomes. In the blot probed with Leu2 we onlydetected a band corresponding to the full length tRNAin both total and exosome RNA, confirming the absenceof fragments that are derived from the anticodon loop ofthis tRNA. In summary, these findings are the first toshow that leishmania produce tRNA-derived smallRNAs and that tsRNAs are specifically enriched inexosomes.DiscussionLeishmania exosomes contain specific RNA cargoIt has been firmly established that exosomes released byvarious mammalian cell types can serve as shuttle vehi-cles to deliver RNA molecules to recipient cells, therebyinfluencing gene expression. However, to date no proto-zoan pathogen has been shown to release bona fideexosomes containing RNAs with gene regulatory orother sequence-specific properties. Leishmania have re-cently been shown to secrete exosomes that contain aplethora of protein virulence factors capable of affecting thephenotype of host mononuclear phagocytes [7,18]. Consid-ering the enormous potential impact of exosome-mediated0.48 26 mid0.07 36 3′0.09 37 3′0.15 33 3′0.01 38 mid-3′isoacceptors. These are sorted by descending abundance in the L. donovani= nucleotide. N.A. = not applicable.Lambertz et al. BMC Genomics  (2015) 16:151 Page 16 of 26delivery of regulatory RNAs to either recipient leish-mania or mammalian host cells or both, we sought toinvestigate whether leishmania exosomes carry RNAcargo. Here, we provide unambiguous evidence thatleishmania parasites of two distinct species, namely L.donovani and L. braziliensis, release exosomes contain-ing RNA sequences. These RNA sequences were hetero-geneous, but overall short in length (25–250 nt). Thus,despite the abundance of longer sequences in total cellRNA, we were unable to detect them in exosomes. Theenrichment of short RNA sequences in leishmania exo-somes is concordant with the majority of reports onexosome RNA in other organisms published thus far(e.g. [22,36]). Although, there have been a few reports ofthe presence of longer RNAs such as full length riboso-mal RNAs [39] and mRNAs [21] as well. The RNAcargo of exosomes is largely dependent on the cell oforigin and appears to be affected by environmentalconditions such as infection or nutritional stress [12],which likely explains the observed differences.One important property of exosomes is their capacityto act as both short and long distance messengers. RNA-containing exosomes have been detected in a variety ofhuman body fluids such as plasma, saliva and semen[40-42], which supports a role in long distance commu-nication. As RNases are ubiquitously present in all or-ganisms, RNAs travelling long distances need to beprotected from degradation. In our in vitro experiments,we were able to show that leishmania exosomal RNAcargo is protected from degradation by exogenousRNase. When we incubated PMA-differentiated THP-1cells in vitro with exosomes purified from axenic amasti-gotes of L. donovani, we saw that the exosome RNAcargo was readily taken up by host cells. This findingsuggests that it should be possible for leishmania-derived RNAs to gain access to host cells through exo-somes in vivo.Numerous studies on exosome RNA have reported thepresence of small regulatory RNAs such as micro RNAs(miRNAs) in these vesicles. It was of interest, therefore,to examine the leishmania exosome RNA content in de-tail by high throughput sequencing. It is important tomention here that leishmania are a special case withregard to small regulatory RNA pathways: L. braziliensisand other species from the new world Leishmania(Viannia) subgenus have been shown to have a func-tional RNAi pathway and actively produce siRNAs[24,32]. Conversely, this pathway appears to have beenevolutionarily lost in old world Leishmania (Leishmania)species, such as L. major and L. donovani [24]. With thiscontrast in mind, we elected to sequence exosome RNAfrom L. braziliensis and L. donovani in parallel in orderto compare the exosome RNA trascriptome of an RNAi-competent organism with an RNAi-deficient one.When sequencing exosomal RNAs from these twoleishmania species, we found that they both contained avariety of small non-coding RNA species, the majority ofwhich appeared to be cleavage products derived fromlonger known non-coding RNAs such as rRNA, tRNA,snoRNA and snRNA. We also saw a small number ofreads mapping to protein coding genes. In addition, wediscovered a number of novel transcripts that were con-served in both libraries, and L. braziliensis exosomesuniquely contained transcripts derived from siRNA-coding putative mobile elements and repeats, such asSLACS and TATEs [32]. Other studies looking at mam-malian exosome RNA content by deep sequencing havereported a similar composition of the exosomal tran-scriptome, with a dominant fraction of sequences being de-rived from rRNA and other non-coding RNA [36,37,39,43].Conspicuously, sequences derived from protein codinggenes seem to be underrepresented in exosomes. Thus,it appears that exosomes from many diverse organismsselectively package non-coding RNAs, the exact functionof which still needs to be determined.Importantly, our study is the first to purify bona fideexosomes from a protozoan parasite and provide a com-prehensive analysis of high-throughput sequencing dataof exosomal RNA cargo. By virtue of comparing two dis-tinct (old and new world) species of leishmania, we havemade the serendipitous discovery that the packaging ofspecific RNA sequences into exosomes appears to be aconserved phenomenon in leishmania. At the presenttime it remains unclear whether our findings are illustra-tive of what happens in other eukaryotic pathogens;however, there is some evidence indicating that the re-lease of RNA within vesicles might occur in other para-sitic organisms as well. In particular, there have beentwo articles published by independent groups that dem-onstrate the release of tRNA-derived small RNAs andother types of RNA within extracellular vesicles shed bythe protozoan T. cruzi [17,44]; however, these vesicleswere not characterized or classified as bona fide exo-somes. The distinction between exosomes and otherextracellular vesicles is important, as their origin, mech-anism of biogenesis and thus loading of cargo differssubstantially [45]. Three other articles have been pub-lished looking at the RNA cargo of exosomes released byparasitic pathogens; one of them a protozoan (Trichomonasvaginalis) and the other two helminths (Heligmosomoidespolygyrus and Dicrocoelium dendriticum). However, all ofthese studies have significant limitations in their experi-mental design and RNA analysis when compared with thepresent study. The study on T. vaginalis only shows asize profile of RNA purified from exosomes measuredby Bioanalyzer, but no sequencing data on exosomalRNA [22]. The article on D. dendriticum describes theanalysis of exosomes by high-throughput sequencing;Lambertz et al. BMC Genomics  (2015) 16:151 Page 17 of 26however, data analysis was focused on looking at microRNA and does not report on other types of RNA in theexosomes [46]. Lastly, a very recent report on H. poly-gyrus reports sequencing data of libraries that have beengenerated with RNA obtained from parasite secretionsand a vesicular fraction collected by ultracentrifugation,but not RNA from bona fide exosomes that were puri-fied with specific exosome purification protocols [47].The limitations of these studies do not allow for a directcomparison with our data and do not definitely answerthe question whether the release of specific types ofRNA within exosomes is a widespread phenomenonamong parasites. While the data available suggest thatthis may certainly well be the case, further research willbe needed to confidently answer this question.Results from a number of studies [48-51] led to thesuggestion that fragments derived from non-coding RNAspecies such as rRNA, snoRNA, vault RNA (vRNA) andtRNA can act as regulatory RNAs similar to miRNAs inRNAi. This hypothesis was based upon the finding thatthese fragments were shown to bind to Argonaute (AGO)proteins and formed RNA-induced silencing complexes(RISCs) which regulate expression of target mRNAs. L.donovani does not have the canonical proteins that arerequired for functional RNAi including AGO. However,an AGO/PIWI-like protein homolog, has been found inRNAi-deficient leishmania and other trypanosomes [52].The function of this AGO homolog is currently unknown;however, one study suggested that it is not involved in thebiogenesis or stability of siRNAs [53]. The presence of analternative pathway of regulation of gene expression inRNAi-deficient trypanosomatids is likely, since theseorganisms perform their transcription polycistronicallyand hence regulation of expression of individual genes canonly take place at the post-transcriptional level. A numberof studies have indicated that post-transcriptonal regula-tion of gene expression in trypanosomes may involve cis-acting regulatory motifs within the 3′ UTRs of mRNAsand trans-acting RNA-binding proteins [54-56]. Other evi-dence for the presence of an alternative RNAi pathway inRNAi-deficient trypanosomatids comes from a recentseries of studies in T. cruzi. The authors identified aunique AGO/PIWI protein termed TcPIWI-tryp that isexpressed in all life cycle stages of the parasite and local-izes to the cytoplasm [57]. Interestingly, they found thatTcPIWI-tryp bound to a repertoire of RNAs distinct fromsiRNAs, namely small RNAs derived from rRNAs andtRNAs [23]. However, while these findings are intriguing,it remains to be established whether these complexesfunction in regulation of gene expression.A large portion of reads in both our libraries mappedto rRNA genes in the reference genomes and theyappeared to be shorter fragments. The presence of rRNA-derived sequences in leishmania exosomes is in agreementwith other recent reports on exosome RNA cargo [39,58].Sequences mapping to rRNAs have also been found inother types of extracellular vesicles, for example shedvesicles released by T. cruzi [17,44]. At this point it isunknown whether rRNA fragments have any specificfunction. Some limited evidence has been presented toshow that rRNA fragments are produced by specificcleavage rather than random degradation in humansand mice [59]. These specific products were character-ized by termini specific processing and asymmetricstabilization. However, in our data, no bias for either 5′ or3′ processing was observed (Additional file 7: Table S4),but mapping of reads was rather scattered along the entirelength of the rRNA gene. Moreover, we did not see enrich-ment of specific rRNA fragments derived from a subset ofgenes. Further study will be needed to elucidate whetherrRNA fragments in leishmania exosomes are specificallyenriched or selectively packaged.Our finding that leishmania exosomes are overallenriched in non-coding RNA fragments which are takenup by host mononuclear cells, raises the interestingpossibility that these RNA fragments may interfere withgene expression in the host, possibly by binding to hostAGO. This type of epigenetic regulation of gene expres-sion across kingdoms has been proposed, but so farlittle consistent and conclusive evidence has been pre-sented. One elegant study recently showed that smallRNAs from the plant fungal pathogen Botrytis cineracould silence Arabidopsis and tomato genes involved inplant immunity by binding host AGO [60]. This was thefirst time that naturally occurring cross-kingdom RNAiwas shown to be a potential, novel virulence mechan-ism. Regarding human pathogens, some evidence hasbeen presented that T. cruzi produces tRNA-derivedsmall RNAs (tsRNAs) that may be delivered to suscep-tible mammalian cells [17]. Moreover, it was shown thattransfection of host HeLa cells with synthetic T. cruzitsRNAs can modify the expression of specific genes asassessed by microarray [61]. It remains to be estab-lished, how these tsRNA-induced changes of host geneexpression were brought about and if this process in-volves hijacking of host RNAi pathways. In what followsbelow, we discuss potential roles of the non-coding,small RNA species found in leishmania exosomes whichwe believe are most likely to have regulatory functions,namely novel transcripts, siRNAs and tRNA-derivedsmall RNAs.Novel transcriptsWhen examining reads that were less abundant in thelibraries, we discovered 12 distinct genomic loci (Table 3and Figure 4) that apparently gave rise to transcriptswhich have not been previously described. None of thesetranscripts had homology to any annotated gene inLambertz et al. BMC Genomics  (2015) 16:151 Page 18 of 26TriTrypDB or GenBank. BLAST search, however, con-firmed that these non-annotated genomic sequenceswere conserved amongst most leishmania species. Inconsiderable interest, we found that all of these noveltranscripts had a spliced leader site upstream of their5′ end (see Figure 4 for two examples) implying thatthey are processed alongside other transcripts duringtrans-splicing. The fact that the majority of novel tran-scripts contained one or more ORFs suggests that theymay be translated into peptides or proteins. However,we were not able to find any homologous protein inany other species.Two of the twelve novel transcripts that were mostabundant from the group were further examined byquantifying their expression in exosomes in comparisonto total cells by Northern analysis (Figure 4B). We foundthat both of these transcripts produced shorter process-ing products that were uniquely present in exosomes.This indicates that cleavage products of these transcriptsmay be specifically targeted for packaging into exosomesfor release from the cell. The lack of a signal for theseshorter products in Northerns of leishmania total RNAmay also explain why these transcripts have not been re-ported in any of the previous studies on sequencing theleishmania transcriptome. Another possibility for whythey have not previously been found is that they couldbe differentially expressed in the different life cyclestages (as we focussed only on axenic amastigotes).One question that remains to be answered is what typeof RNA these novel transcripts represent (protein cod-ing, structural, regulatory) or whether they representnovel type(s) of RNA. Further studies will be needed toproperly address the functions of these novel transcripts.siRNAsWe detected a number of sequences in the L. braziliensisexosome RNA library that mapped to siRNA coding locisuch as SLACS and TATEs in the L. braziliensis genome.Even though the functional significance of endogenoussiRNAs in L. braziliensis is still unclear, they are thoughtto be a genome defence mechanism to control thespread of potentially harmful nucleic acids, such as mo-bile elements, repeats and viruses [32]. Although thesesequences were detectable in our library, they were gen-erally in low abundance when compared to fragments ofrRNA or tRNA. The fact that half of these sequenceswere each sense and antisense supported their tentativeidentity as siRNAs, given that one cardinal feature ofsiRNAs is that they are double stranded. Moreover, thesewere the only type of sequences in our libraries wherethis phenomenon was observed, as the vast majority ofthe other sequences (rRNA and tRNA fragments) werepresent only in the sense direction. The lengths of theseputative siRNA sequences did not correspond exactly towhat has been reported for L. braziliensis mature siR-NAs (LbrAGO1-bound) [32]. This might be due todifferent library construction strategies (size selection),it may be that what we detected were siRNA precursors,or that the sequences we detected were derived fromdistinct regions within the SLACS and TATEs loci. Thefinding that L. braziliensis packages these putativesiRNA sequences as cargo of exosomes is of significantinterest. It implies that these RNAs may function notonly in the cell where they originated, but that they mayalso act in intercellular communication when taken upby other leishmania or by host cells or both. To ourknowledge, no other parasite has been shown to releasepathogen-derived siRNAs within vesicles directly. Fur-ther studies will be needed to confirm the identity ofthe sequences as siRNAs and delineate their function inparasite biology and in host-pathogen interaction.tRNA-derived small RNAsA striking finding of the present study was the abun-dance of tRNA fragments principally originating from asmall subset of tRNA isoacceptors (Figure 4B) that werehighly conserved in the L. donovani and L. braziliensisexosome transcriptomes. Only recently tRNA fragmen-tation has been recognized as a specific process. tRNAfragments have been found in all domains of life and canbe divided into several categories, depending on thecleavage site: cleavage within the anticodon loop givesrise to 5′ and 3′ tRNA halves (30–35 nt), and cleavagewithin the D-arm (5′) or T-arm (3′) gives rise to smallertRNA-derived RNA-fragments (tRFs) (13–20 nt) [62].Each of these fragments appears to be generated throughdistinct pathways. tRNA halves are known to be pro-duced in response to stress [62], whereas the smallertRFs, on the other hand, can be generated at any pointof tRNA processing, by Dicer-dependent or –independ-ent mechanisms [62,63]. Together, tRNA halves andtRFs are referred to as tRNA-derived small RNAs(tsRNAs). tsRNAs have recently been described inhigher as well as lower eukaryotes Their physiologicalfunction is not well understood, but they have com-manded increasing interest due to their suspected regu-latory nature. Notably, it appears that tRNA halves andtRFs have distinct biological functions. In human cells,tRNA halves were found to inhibit protein translationby specifically targeting the translation initiation ma-chinery and displacing elongation initiation factors [64].tRFs, on the other hand, were shown to be involved inregulation of translation by directly binding to the smallribosomal subunit and interfering with peptidyl trans-ferase activity in archae [65]. Furthermore, a similarmechanism was observed in human cells, where a tRFwas shown to inhibit translation by affecting peptidebond formation [66]. In addition to these effects onLambertz et al. BMC Genomics  (2015) 16:151 Page 19 of 26translation, tRFs have also been shown to function ingene silencing. Haussecker et al. showed that tRFs canassociate with Argonaute proteins, however, they asso-ciated more effeciently with the non-silencing AGO3and AGO4 [50]. Furthermore, they saw that tRFs canaffect the silencing activities of miRNAs and siRNAs,indicating a potential broad based role in regulatingRNA silencing. In another study it was found that tRFscan function like miRNAs in RNAi and inhibit the ex-pression of RPA1 (a protein involved in DNA repair) bybinding to the 3′UTR of its mRNA [67].A small number of studies have looked at the presenceof tsRNAs in protozoan parasites. In Plasmodium ber-ghei and Toxoplasma gondii, tRNA-halves were detected,and a relation between tRNA-half production andgrowth rate was observed [68]. However, the precisefunction of these parasite-derived tRNA-halves remainsunknown. Interestingly, tsRNAs were recently discov-ered in leishmania’s close relative T. cruzi. Despite thefact that the lengths and origins of tsRNAs differedslightly from study to study, their production was con-vincingly demonstrated in both trypomastigotes and epi-mastigotes in a number of reports [28,29,69]. Importantly,it was found that T. cruzi tsRNAs were bound to TcPIWI-tryp, a distinct Argonaute protein that has been describedin this RNAi-deficient organism [23,57]. The majority ofthese TcPIWI-tryp bound tsRNAs corresponded to the 5’halves of tRNA-Glu. Importantly, TcPIWI-tryp-tsRNAcomplexes were also found in vesicles shed from T. cruzi.The authors proposed that these vesicles may have a rolein life cycle transition from epimastigote to trypomastigoteas well as contribute to infection susceptibility of mamma-lian cells [17]. These findings provide evidence thattsRNAs in T. cruzi may participate in non-canonical regu-latory pathways and raise the question as to whether asimilar phenomenon may be operative in leishmania.In the present study, we provide evidence that leish-mania also produces specific tsRNAs, and that thesepotential regulatory RNAs are released by the intracellu-lar amastigote stage within bona fide exosomes, compe-tent for delivery to mammalian cells. The major fractionof tsRNAs found in both L. donovani and L. braziliensisexosomes were 5′ tRNA halves, however, we also foundshorter tsRNAs derived from the D-arm or T-arm of thetRNA, corresponding to 5′tRFs and 3′tRFs. The produc-tion and presence of tRNA halves in exosomes fromleishmania amastigotes might be a result of the elevatedtemperature and acidic pH the parasites were exposed to.We found that the vast majority of tsRNAs in leishmaniaexosomes were derived from tRNA-Asp, tRNA-Gln,tRNA-Leu, tRNA-Glu and tRNA-Gly (Table 4 andFigure 5) and this was highly conserved between L.donovani and L. braziliensis. Although we did not carryout a comprehensive and quantitative analysis of thefrequencies of all tsRNAs in leishmania whole cells,strikingly, we found by Northern blotting that tsRNAsfrom tRNA-Asp and tRNA-Leu were highly enriched inexosomes, with no detectable amounts in leishmaniatotal RNA. This indicates that these tsRNAs are prefer-entially and quantitatively packaged into exosomes tobe released from the cell rather than being retained inthe whole cell (minus exosomal) RNA pool.At this point, the mechanism of biogenesis and functionof tsRNAs in leishmania is unknown. We are cannot,therefore, be 100% certain that the same classification oftRNA halves and tRFs as recently proposed by severalgroups [62,63,70] applies to our data. However, as many ofthe characteristics (length, cleavage site, isoacceptororigin) correspond to what has been reported in otherorganisms, we conclude that the phenomenon of spe-cific tsRNA generation is evolutionarily conserved inleishmania as well. Based upon our initial functionalpredictions it appears clear that leishmania tsRNAs arenot highly similar to canonical mammalian miRNAs orsiRNAs. Further detailed investigations will be neededto delineate the functions of tsRNAs in leishmaniabiology, what roles they play in parasite-parasite,parasite-vector or parasite-host interactions, whether thisinvolves their association with the host RNAi machineryand how they are targeted for exosomal packaging.ConclusionsIn summary, this report provides evidence that leish-mania exosomes are enriched in short sequences derivedfrom non-coding RNAs such as rRNAs and tRNAs.Moreover, exosomes contain a number of novel tran-scripts, albeit in relatively low abundance. The RNAi-proficient L. braziliensis appears to package putative siR-NAs or their precursors into exosomes, whereas RNAideficient L. donovani does not. Based on Northern ana-lyses, our data indicate that specific RNA sequences areselectively, and in some cases quantitatively packagedinto exosomes. This conclusion is supported further bythe highly biased distribution of sequences detected inexosomes over only a subset of genes in the leishmaniagenome combined with a striking paucity of transcriptsderived from protein coding genes which are otherwiseabundant in total cellular RNA.Importantly, our findings provide at least three lines ofevidence arguing for the presence of an evolutionarilyconserved mechanism for packaging small non-codingRNAs into exosomes in leishmania: 1) the high degreeoverlap between the top 20 most abundant sequencesfound in L. donovani and L. braziliensis exosomes, 2)the vast majority of identified novel transcripts waspresent in exosomes from both species, and 3) the mostabundant tsRNAs found in exosomes were derived froma highly biased subset of the same tRNA isoacceptors inLambertz et al. BMC Genomics  (2015) 16:151 Page 20 of 26both species. Taken together, the data argue stronglythat leishmania exosomal RNA sequences are specific-ally produced and packaged into exosomes for release,likely with the purpose of modifying host cell pheno-type to support chronic infection. The investigation ofthe precise functions of these small, non-coding, leish-mania RNAs should contribute significantly to ourunderstanding of parasite biology and mechanisms ofpathogenesis.MethodsCell cultureL. donovani Sudan strain S2 promastigotes were rou-tinely cultured in M199 (Sigma-Aldrich) with 10% heatinactivated fetal bovine serum (FBS, Gibco), 20 mMHEPES (Stemcell), 6 μg/mL hemin (Sigma-Aldrich),10 μg/mL folic acid (Sigma-Aldrich), 2 mM L-glutamine(Stemcell), 100 U/mL penicillin/streptomycin (Stemcell)and 100 uM adenosine (Sigma-Aldrich) at 26°C. Every3 days the organisms were subcultured 1:10 in freshmedium and were kept in culture for a maximum of20–25 passages. Fresh parasites were obtained by purifi-cation of amastigotes from spleens of infected SyrianGolden hamsters followed by in vitro transformationinto promastigotes by culturing for 7 days at 26°C inpromastigote media.L. braziliensis (clinical isolate from the Peruvian Amazonregion) promastigotes were routinely cultured in the samemedia as above except for supplementation with 20% FBS.L. braziliensis promastigotes were subcultured 1:5 every3 days in fresh media and kept at 26°C.Purification of exosomesExosomes were purified from L. donovani and L. brazi-liensis axenic amastigote culture supernatant as de-scribed previously [7,18]. Briefly, 400–800 mL of day 5promastigotes (at a concentration of 5×10E7 cells/mL)were washed 2× with Hank’s buffered salt solution(HBSS, Sigma-Aldrich) followed by incubation in serum-free buffered exosome collection media at pH = 5.5,RPMI1640 supplemented with 1% D-glucose, 20 mMHEPES, 2 mM L-glutamine, 100 U/mL penicillin/streptomycin and 25 mM MES (all from Sigma-Aldrich),at 34°C for L. braziliensis and 37°C for L. donovani. After24 hours of incubation, exosomes were purified from the400–800 mL culture supernatant under endotoxin-freeconditions by a series of centrifugation and filtrationsteps, followed by flotation on a sucrose cushion, as de-scribed in [18]. After a final pelleting step at 100,000×gfor 1 hour, purified exosomes were resuspended in 50–100 μL of PBS and processed immediately (in case ofRNA extractions) or stored at 4°C for a maximum of5 days (for macrophage uptake experiment and Nanosightanalysis).Nanosight particle tracking analysisThe size and concentration of the isolated exosomeswere analysed using the NanoSight™ LM10-HS10 system(Malvern Instruments). For analysis, a monochromaticlaser beam (405 nm) was applied to the diluted exosomesolution (1:100 in 0.02 μm filtered PBS) that was injectedinto a LM12 viewing unit using a computer controlledsyringe pump. NanoSight™ tracking analysis (NTA)software version 2.3 was used to produce the mean andmedian vesicle size together with an estimate of particleconcentration. Samples were measured 3 times toconfirm reproducibility.Extraction and biochemical characterization of RNARNA was purified from leishmania exosomes by phenol/chloroform extraction using all RNA-grade reagents. Forthis purpose, 150 uL of LETS buffer (0.1 M LiCl, 0.01 MNa2EDTA, 0.01 M Tris-Cl pH = 7.4, 0.2% SDS, all Sigma-Aldrich) was added to 50 μL of exosomes resuspended inPBS followed by addition of 200 μL Ultra-Pure buffer-saturated phenol pH = 7.4 (Life Technologies). Themixture was vortexed vigorously and centrifuged for2 minutes at 13,000×g in a microcentrifuge at roomtemperature. The upper aqueous phase was collectedand the phenol extraction was repeated once morefollowed by two extractions over 200 μL chloroform each(Fisher Scientific). RNA was precipitated by addition of0.3 M NaCl, 2 μg/mL glycogen (Ambion) and 75% EtOH,and incubation at −20°C over night. RNA was pelleted bycentrifugation at 13,000×g for 30 minutes at 4°C. RNApellets were washed with ice-cold 75% ethanol and re-suspended in 10–20 μL ddH2O. RNA concentration wasdetermined by measuring the OD260 with the nanodrop(Thermo).To look at length profiles of exosome-derived RNA,2 μg of purified leishmania total RNA and 1 μg of exo-some RNA were first treated with 5 units DNase I(Thermo) to remove potential DNA contamination.After incubation for 30 minutes at 37°C, DNase wasinactivated by addition of 2.5 mM EDTA and incubationat 65°C for 10 minutes followed by phenol-chloroformextraction and ethanol precipitation as above. RNA wasresuspended in 4 μL ddH2O and RNA length profileswere obtained with the Agilent Bioanalyzer using theRNA 6000 Pico kit according to the manufacturer’s in-structions (Agilent). Alternatively, DNase-treated RNAwas run on a 15% polyacrylamide gel, stained with SYBRgreen (Life Technologies) and imaged with UV-imaging.To confirm identity of nucleic acid purified from exo-somes as RNA, 1–2 μg of phenol/chloroform extractedRNA was treated with DNase (as above), followed bytreatment with either 0.4 mg/mL RNase A (Thermo) for15 mins at 37°C or hydrolysis with 50 mM KOH(Sigma-Aldrich) for 15 min at 95°C. Samples were thenLambertz et al. BMC Genomics  (2015) 16:151 Page 21 of 265′ end labeled according to the manufacturer’s instruc-tions using polynucleotide kinase (PNK) (New EnglandBiolabs, NEB) and γ32PdATP (Life Technologies) andrun on 15% polyacrylamide gels followed by imagingwith a Typhoon phosphor-imager (GE Healthcare).To assess whether the exosomal membrane was pro-tecting the vesicular RNA content from degradation byexogenous RNases, intact exosomes resuspended in PBS(from 400 mL culture supernatant, split into 4 samples)were treated with 0.4 mg/mL RNAse A for 15 mins at37°C in the presence or absence of 0.1% Triton X-100(Sigma-Aldrich). As a control for RNase activity, 1 μL ofprepared RNA pico ladder (Agilent) was treated withRNase A under the same conditions. After incubation,samples were extracted with phenol/chloroform 2×each and RNA was precipitated with ethanol as above.Samples were then treated with DNase, again phenol/chloroform extracted and ethanol precipitated, resus-pended in 4 μL ddH2O and run on the Agilent Bioanalyzerto determine whether or not RNA had undergonedegradation.Vesicle delivery of RNA cargo to macrophagesExosomes were purified from 400 mL culture super-natant of L. donovani axenic amastigotes as describedabove. Pelleted exosomes were resuspended in 100 μLPBS. Protein concentration in the exosome preparationwas determined using the Micro BCA Protein Assay kit(Pierce). Exosomes were then stained with the membrane-permeant, RNA-specific dye SYTO RNASelect (LifeTechnologies) according to the manufacturer’s recom-mendations. For this purpose, the SYTO dye was dilutedin DMSO and added to the resuspended exosomes at afinal concentration of 10 μM, followed by 20 minutes in-cubation at 37°C. Excess unbound dye was removed bywashing twice with 1 mL PBS, pelleting the exosomes at100,000×g for 1 hour at 4°C. Exosomes were then resus-pended in the original volume of PBS (100 μL). Labellingefficiency was assessed by fluorescence microscopy usingan Axioplan II epifluorescence microscope equipped with63×/1.4 Plan-Apochromat objective (Carl Zeiss Inc). Im-ages were recorded using an AxioCam MRm Cameracoupled to the AxioVision software Version 4.8.2 (CarlZeiss Inc.).To investigate the exosome-mediated delivery of RNAto host macrophages, THP-1 cells were differentiated overnight with 10 ng/mL phorbol-12-myristate 13-acetate(PMA), followed by washing and resting cells for 24 hours.Differentiated cells were then treated with labeled exo-somes for 2 hours at 37°C. As a negative control, cellswere treated with labelled exosomes and incubated at 4°Cfor 2 hours, preventing phagocytosis. For quantificationof exosome RNA uptake, exosome-treated THP-1 cellswere washed 3× with PBS to remove non-internalizedexosomes. Cells were then fixed with 2% paraformal-dehyde (Sigma) in PBS for 15 minutes at roomtemperature. After fixation, cells were again washedwith PBS and then analyzed by flow cytometry (FACSCalibur, BD). To verify that exosomes were in fact inter-nalized and not just bound to the cell membrane, thesame experiment was performed with THP-1 cellsgrown on coverslips to be analysed by confocal micros-copy. After incubation, cells on coverslips were washedand fixed as above, permabilized with 0.1% Triton X-100 in PBS for 5 minutes, and stained with Alexa Fluor594 phalloidin (Life Technologies) for 1 hour at roomtemperature in the dark. After 3× washing with PBS,coverslips were mounted with Prolong Gold antifademounting media containing DAPI (Life Technologies)to detect macrophage nuclei. Confocal microscopy wasdone with a Leica DMIRE2 inverted microscopeequipped with a SP2 AOBS laser scanning head. This isa filter-free spectral confocal and multiphoton micro-scope, and all imaging operations which include selec-tions of laser, detection channels and other functionsare fully automated and computer controlled. Pictureswere taken with a 63× magnification oil immersionobjective.Library construction and sequencingWe used 1–2 μg of RNA extracted by phenol/chloro-form extraction from L. donovani and L. braziliensisexosomes (from one individual exosome preparationeach, from 800 mL supernatant) as starting material.RNA was first treated with DNase I (as described above)to remove potentially contaminating DNA. To remove5′ phosphates on the RNA, we first performed a calf in-testinal alkaline phosphatase (CIP) treatment using 1unit of CIP (Roche) per 10 μL reaction and incubationfor one hour at 37°C. Once incubation was completed,samples were phenol-chloroform extracted twice andethanol precipitated as described above. In order tomonitor the efficiency of the CIP treatment, a parallelreaction was spiked with a 24 nt long radiolabelledRNA, and pre and post incubation with CIP were loadedin a 20% denaturing polyacrylamide gel to follow thedisappearance of counts. Next, the CIP treated RNAsample (resuspended in 10 μL ddH2O) was treated withtobacco acid phosphatase (TAP) to remove 5′ caps. Halfof the CIP treated RNA sample was combined with 2.5units of TAP (Epicentre), 1X TAP buffer, brought to afinal volume of 10 μL with ddH2O and incubated forone hour at 37°C. In order to control for the efficiencyof 5′ cap removal, a parallel reaction was spiked withγ32PdATP and pre-and post ligation samples wereloaded in a 20% plyacrylamide gel to monitor the dis-appearance of counts. RNA was ethanol-precipitatedand resuspended in 10 μL ddH2O. The CIP and TAPLambertz et al. BMC Genomics  (2015) 16:151 Page 22 of 26treated RNA were then labeled with polynucleotidekinase (PNK) to have the same 5′ phosphate in all RNAmolecules about to be ligated. Ten units of PNK (NEB)were used along with 1X PNK buffer (NEB), andγ32PdATP in a 10 μL reaction and incubation for onehour at 37°C. Next, ~ 10% of the sample was loadedonto a denaturing 15% polyacrylamide gel. The remain-der of the PNK reaction was taken to a 30 μL volumewith ddH2O and run through a dye terminator removal(DTR) cartridge (EdgeBio) following manufacture’s indi-cations in order to remove ions and unincorporatedγ32PdATP. The sample was ethanol precipitated andresuspended in 5 μL of ddH2O. The next step was toligate a custom adenylated AppDNA adaptor (5′ App-GAAGAGCCTACGACGA) to the 3′ end of RNA mole-cules. This adaptor was slightly modified so that the 3′end was blocked in order to prevent self ligation. Halfof the pre-treated exosomal RNA sample was combinedwith T4 RNA ligase buffer (50 mM HEPES, pH 8.3,10 mM MgCl2, 3.3 mM DTT, 10 g/ml BSA and 8.3%glycerol; [71], 2.5 U of T4 RNA Ligase (Epicentre), and20 μM AppDNA adaptor. Reactions were incubated atroom temperature for one hour, and then the enzymewas denatured by heating at 65°C for 20 minutes.Samples were gel purified on 10% polyacrylamide gelsto remove un-ligated adaptor. A second ligation reactionwas set up to attach an RNA adaptor (5′ rAUCGUAGG-CACCUGAAA) to the 5′ end of the RNA-DNA hybrid.Conditions were the same as described above for the firstligation reaction with the only difference being thatγ32PdATP (final concentration of 0.4 mM) was added. Theligation reaction was gel purified from a denaturing 10%polyacrylamide gel and the recovered material was used asa template in a reverse transcription (RT) reaction. Forthis purpose, half of the recovered sample was combinedwith 100 μM RT primer (5′ TCGTCGTAGGCTCTTC),ddH2O and incubated at 80°C for two minutes. Aftercooling samples down slowly, 1X First Strand Buffer(Life Technologies), 0.8 μM dNTP and 200 U of Super-script II Reverse Transcriptase (Life Technologies) wereadded. Controls with no enzyme and no template in thereaction were prepared in parallel. Reactions were incu-bated for 1 hour at 48°C. The RNA template was hydro-lyzed by heating in the presence of 100 mM KOHfollowed by neutralization with 1 M Tris–HCl pH = 5(to a final pH = 8), and the resulting cDNA was isolatedon a 10% denaturing polyacrylamide gel. 20 cycles ofPCR amplification were performed in the presence of5 mM MgCl2, 100 μM dNTPs, 1 μM each forwardprimer (5′ ATCGTAGGCACCTGAAA) and reverse pri-mer (same as RT primer), 1X Taq buffer and 2.5 Units Taqpolymerase (UBI). PCR products were then gel purifiedand quantified by Qubit (Life Technologies) and used asinput material for ligation of TruSeq adapters (Illumina)according to the manufacturer’s recommendations. 150base pair, paired-end sequencing was performed using anIllumina MiSeq instrument (Illumina) at the Epigenomicscore of Albert Einstein College of Medicine, NY.Sequencing data analysisAfter completion of Illumina paired-end sequencing andread quality control checking by FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), both L.donovani and L. braziliensis exosomal RNA reads hadtheir adapters trimmed by cutadapt version 1.0 (http://journal.embnet.org/index.php/embnetjournal/article/view/200/479). For each library, the output files from thetrimming were separated into RNA adapter-trimmedreads and DNA adapter-trimmed reads, and the formerwas used to guide the assignment of correct orientationfor all reads sequenced. We ran FLASh (settings: −M100 –x 0.2) [72] and FASTX-Collapser (http://hannon-lab.cshl.edu/fastx_toolkit) to respectively combine themates, for the cases where DNA inserts were shorterthan twice the length of reads, and then collapsed iden-tical sequences into single ones to facilitate handlingthe data in subsequent specific analyses. Bowtie2version 2.1.0 (settings: very-sensitive-local –N1) [73]was used to align the collapsed reads (cReads) fromboth libraries against their respective reference genomes(LdBPK (Leishmania donovani strain BPK282A1) andLbrM (Leishmania braziliensis MHOM/BR/75/M2904)from TriTrypDB version 6.0), as well as against the spe-cies with the best assembled and annotated genome(LmjF, L. major MHOM/IL/81/Friedlin, TriTrypDBversion 6.0). The very-sensitive-local setting of Bowtie 2uses a seed length of 20 nt for the alignment, and the –N1 command allows for only one mismatch on thatseed alignment. The alignments with LdBPK and LbrMwere used to categorize the exosomal RNAs for therespective species, relying on htseq-count script [74]and the GFF files provided by TriTrypDB v6.0. Thealignment with LmjF was done mainly to refine the ana-lyses of reads mapping onto tRNAs. Of note, right afterbowtie2 execution, samtools version 0.1.18 [75] wasapplied to generate sorted bam files, which were then usedas input to cufflinks (settings: −u –min-intron-length3 –3-overhang-tolerance 25 –overlap-radius 10 –min-frags-per-transfrag 1) [76] for the assembly of reads map-ping on the same locus into individual “transcripts” orclusters. Artemis genome browser software [27] wasused to manually inspect in greater detail and visualizethe alignment of exosomal sequences with the referencegenomes.As mentioned above, we used reads mapping to L.major tRNAs for a better categorization of potentialtRNA-derived small RNAs present within the exosomes.An ad-hoc PERL script was written to calculate theLambertz et al. BMC Genomics  (2015) 16:151 Page 23 of 26cReads position within each tRNA feature they mappedonto: 5′end (cReads mapping entirely on the 5′end halfof the tRNA gene), mid-5′ (cReads starting on the first 1/3and ending before the last 1/3 of the tRNA gene length),3′end (cReads mapping entirely on the 3′end half of thetRNA gene), mid-3′end (cReads starting after the first 1/3and ending within the last 1/3 of tRNA gene length), mid(cRead overlaps both halves of the tRNA gene and notwithin the 1/3 extremity regions). The same method wasused to calculate the cReads position within each rRNAfeature for rRNA fragments found in exosomes.The discovered 12 novel transcribed loci had their nu-cleotide sequences translated by the getorf programfrom the EMBOSS package [77] with the following para-maters: −minsize 33 -find 1 –noreverse, which sets a 10amino acids minimum ORF length, translates solelyfrom ATG to STOP codons and only on the three pos-sible frames from the same strand where the exosomeRNA reads mapped to, respectively. In order to checkwhether the putative ORFs outputted by the getorf pro-gram have similarity to any already known protein, weran sensitive blastp against nr-NCBI (−word_size 2-num_descriptions 5 -num_alignments 5 -evalue 1e-3)and no hits were found for any of them.To determine whether there were transposableelements-derived RNA fragments within leishmaniaexosomes and also discard any possibility of cross-contamination between the libraries, we performed aBlastN search [78] for L. braziliensis-specific SLACS/TATEs elements (extracted from TriTrypDB-6.0_Lbra-ziliensisMHOMBR75M2904_ AnnotatedTranscripts.fastadownloadable file at tritrypdb.org). The following thresh-olds were applied during this screen: e-Value < = 1,identity > = 80% and query (cRead) coverage > = 70%.To search for any sequences homologous to mam-malian miRNAs within the leishmania exosomal RNAlibraries, we ran blastn from the BLAST Plus package[79] version 2.2.29+ querying the top thousand mostabundant cReads on each library against the wholehuman and mouse miRNA dataset (hairpin and mature)available at miRBase (mirbase.org). The blastn + parame-ters were the following: −dust no -word_size 4 -evalue 1-outfmt 6, and we also established a cutoff of 70% iden-tity and 70% sequence coverage (ad-hoc PERL script). Ina parallel approach to identify host genes that couldpotentially be targeted by putative regulatory RNAs inleishmania exosomes, we aligned the cReads from bothlibraries against human (hg19, NCBI) and vector (Lutzo-myia longipalpis and Phlebotomus papatasi, https://www.vectorbase.org/) reference genomes. Bowtie2 ver-sion 2.1.0 (settings: −-very-sensitive-local –N1) [73] andhtseq-count script (http://www-huber.embl.de/users/an-ders/HTSeq/doc/overview.html) (using the option –sreverse, which reports reads mapping to annotatedfeatures on a reverse complement fashion) were usedfor this purpose.Northern blottingAliquots of ~3 μg RNA per lane (L. donovani axenicamastigote total RNA from a single culture, or exosomeRNA pooled from 4 separate exosome preparations)were loaded onto 8% denaturing polyacrylamide gels.Gels were stained with SYBR Green and visualized, thenthe samples were blotted onto Hybond N+ nylon mem-brane (GE Healthcare). The membranes were UVcross-linked using a Stratalinker (1200 μJ for 30 sec-onds), blocked, probed and washed according to [80].Twenty one and 150 nt long in vitro transcribed RNAswere used as size markers. For hybridization, 5′end la-belled DNA probes were used (LdBPK_291610_leftofprobe 5′ AAGGCGTCCCCATGATAACG, LdBPK_301180_leftof probe 5′ GACCTCAAGTATCTACGGGAGA, tRNA-Asp probe ASP1 5′ GGCGGGTATACTAACCACTATAC, tRNA-Leu probe LEU1 5′ AGACCACTCGACCATCTCA, tRNA-Leu probe LEU2 5′TGGAACCTTAATCCAACGTCTT, 5.8S rRNA probesequence was taken from [81]). For 5′ end labelling,the PNK labelling reaction was carried out as sug-gested by the manufacturer (NEB). The efficiency ofγ32PdATP incorporation was determined by running asmall fraction of the PNK reaction on a native 20%polyacrylamide/urea gel and typically ranged between80-95%, resulting in probes with high specific activity.The blotted membranes were placed in glass bottlescontaining a minimal amount of hybridization buffer(6X SSPE, 1% SDS, 2X Denhart’s solution, 100 μg/mLof salmon sperm DNA; [82]) in a Hybaid™ mini ovenMKII and pre-hybridized with constant rotation for4 hours at 37°C. After pre-hybridization, approximately10 μCi of labelled probe was added and the membranewas hybridized for at least 18 hours at 37°C. The nextday, the membrane was washed twice with a high strin-gency solution (2X SSPE and 0.1% SDS), and twicewith a low stringency solution (0.2X SSPE and 0.1%SDS) for 15 minutes each at room temperature. Theradioactive signal from the membranes was detectedusing a Storm 820 phosphorimager. Quantification ofsignals was performed in Imagequant.Statistical analysis and graphsR environment version 3.0 was used to generate readlength distribution histograms for each library (calcu-lating their mean and median values), as well as to per-form the Pearson’s Correlation analysis regarding thetRNA-derived small RNA reads abundance and theamino acid usage frequency of the respective predictedproteomes. Other graphs were generated with Graph-Pad Prism 4.0 and EXCEL.Lambertz et al. BMC Genomics  (2015) 16:151 Page 24 of 26Additional filesAdditional file 1: Figure S1. Leishmania braziliensis exosomes containRNA. Exosomes were purified from L. braziliensis axenic amastigote culturesupernatant as described in the Methods. RNA was extracted from exosomeswith phenol-chloroform and then analyzed. A. Agilent Bioanalyzer RNA lengthprofiles of exosome RNA alongside total RNA, B. RNA inside exosomes isresistant to degradation. Prior to RNA extraction, intact exosomes wereleft untreated or treated with either RNase A or TritonX-100 or both.Samples were then subjected to RNA extraction and run on the AgilentBioanalyzer. Arrowhead indicates internal 25 nt marker. nt = nucleotides.Additional file 2: Figure S2. Leishmania donovani exosomes can beefficiently stained with green fluorescent SYTO RNASelect dye. Exosomeswere purified from 400 mL supernatant of L. donovani axenic amastigotesand stained with a membrane permeant, green fluorescent RNA-specificdye (as described in Methods). A sample each of stained and unstained L.donovani exosomes were then examined by microscopy using an AxioplanII epifluorescence microscope equipped with 63×/1.4 Plan-Apochromatobjective (Carl Zeiss Inc). Images were recorded using an AxioCam MRmCamera coupled to the AxioVision software Version 4.8.2 (Carl Zeiss Inc.).Additional file 3: Table S1. Sequences found in L. donovani and L.braziliensis exosomes. File 1 ‘LD_Collapsed_reads.tab’: Collapsed Read IDsincluding full nucleotide sequences, File 2 ‘LD_vsLdoB-bowtie2-HTseq.tab’: Results of the Bowtie2 alignment of the L. donovani exosomesequencing library against the LdBPK reference genome, File 3‘LD_vsLmjF-bowtie2-HTseq.tab’: Results of the Bowtie2 alignment of theL. donovani exosome sequencing library against the LmjF referencegenome. File 4 ‘LB_Collapsed_reads.tab’: Collapsed Reads including fullnucleotide sequences, File 5 ‘LB_vsLbrM-bowtie2-HTseq.tab’: Results ofthe Bowtie2 alignment of the L. braziliensis exosome sequencing libraryagainst the LbrM reference genome, File 6 ‘LB_vsLmjF-bowtie2-HTseq.tab’: Results of the Bowtie2 alignment of the L. braziliensis exosomesequencing library against the LmjF reference genome. “cRead ID” refersto the identifier of the collapsed read, which is composed of the cReadnumber, followed by the number of copies. “Strand” refers to the strandthe reads were aligning with, plus = top strand, minus = bottom strand.Additional file 4: Table S2. Other hits found in L. donovani and L.braziliensis exosomes. File 1 ‘LD_blastN-results-vsNT.tab’: Results of theblastN search of the unaligned reads (i.e. not aligned to LdoB or LmjF)from the L. donovani exosome sequencing library against the NCBInucleotide collection database. File 2 ‘LD_GBhits_description.tab’:GenBank description of hits. File 3 ‘LD_reads_per_species.tab’: Summaryof reads per species. File 4 ‘LB_blastN-results-vsNT.tab’: Results of theblastN search of the unaligned reads (i.e. not aligned to LbrM or LmjF)from the L. braziliensis exosome sequencing library against the NCBInucleotide collection database. File 5 ‘LB_GBhits_description.tab’:GenBank description of hits. File 6 ‘LB_reads_per_species.tab’: Summaryof reads per species.Additional file 5: Table S3. Overall alignment statistics.Additional file 6: Figure S3. Length histograms of reads mapping torRNA genes.Additional file 7: Table S4. Mapping of reads to rRNA genes based onBowtie 2 alignments with the LmjF reference genome.Additional file 8: Table S5. Putative ORFs found in the novel transcripts.Additional file 9: Table S6A. Results of blast search of all reads from L.donovani exosome library against human and vector genomes. Table S6BResults of blast search of all reads from L. braziliensis exosome libraryagainst human and vector genomes.Additional file 10: Table S7. Results of blast search of all reads from L.donovani and L. braziliensis exosome library against SLACS and TATEs database.Additional file 11: Table S8. Hits against miRbase: Top 1000 most abundantreads in both libraries were blast searched against human and mouse miRNAs.AbbreviationsAGO: Argonaute; BLAST: Basic local alignment search tool;cDNA: Complementary DNA; CDS: (protein) coding sequence; CIP: Calfintestinal phosphatase; CIR74: Chromosomal internal repeats, 74-nucleotidelong; DNA: Deoxyribonucleic acid; FBS: Fetal bovine serum; HBSS: Hank’sbuffered salt solution; LINE: Long interspersed elements; miRNA: micro RNA;mRNA: Messenger RNA; nt: Nucleotide; PAGE: Polyacrylamide gel electrophoresis;PBS: Phosphate buffered saline; PNK: Polynucleotide kinase; RNA: Ribonucleicacid; RNAi: RNA interference; rRNA: Ribosomal RNA; RT: Reverse transcription;scRNA: Small cytoplasmic RNA; SINE: Short interspersed elements; siRNA: Smallinterfering RNA; SL: Spliced leader; SLACS: Spliced leader-associatedconserved sequence; snoRNA: Small nucleolar RNA; snRNA: Small nuclearRNA; srpRNA: Signal recognition particle RNA; TAP: Tobacco acid phosphatase;TAS: Telomere-associated sequence; TATE: Telomere-associated transposableelement; tRF: tRNA-derived RNA fragment; tRNA: Transfer RNA; tsRNA:tRNA-derived small RNAvRNA, vault RNA; U: Unit; UTR: Untranslated region.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsUL and MOO contributed equally to this manuscript. UL and MOO designedthe methods and carried out the experiments, analyzed the data andinterpreted the results. UL and NER conceived of the research theme,reviewed and interpreted all data and drafted the manuscript. EJRV did thebioinformatics analyses of sequencing data. PJU and PJM co-worked onexperimental design and interpretation of data and contributed reagents/analysis tools. All authors have contributed to, read and approved the finalmanuscript.AcknowledgementsThe authors would like to thank Dr. Gowthaman Ramasamy (Seattle Biomed)for his support with the bioinformatic analyses, the staff of the CellularImaging and Biophysics Core at the Centre for Heart Lung Innovation (UBCand St. Paul’s Hospital) for their support with confocal microscopy, and Dr.Geetanjali Kharmate and Dr. Emma Guns (Vancouver Prostate Centre, UBC)for their support with the Nanosight analysis. This work was funded by GrantMOP – 125879 from the Canadian Institutes for Health Research (to NER),Grant RGPIN - 238948 from the National Sciences and Engineering Council ofCanada (to PJU) and National Institute of Health Award Number 1R01AI103858 (toPJM). EJRV is a post-doctoral fellow from Conselho Nacional de DesenvolvimentoCientífico e Tecnológico – CNPq, Brazil [PDE 202223/2012-4].The funding agencies had no role in design, collection, analysis, orinterpretation of data; neither were they involved in writing the manuscript;or in the decision to submit the manuscript for publication.Author details1Departments of Medicine, Microbiology and Immunology, University ofBritish Columbia, Vancouver, BC, Canada. 2Department of Molecular Biologyand Biochemistry, Simon Fraser University, Burnaby, BC, Canada. 3SeattleBiomedical Research Institute, Seattle, WA, USA. 4Departments of GlobalHealth and Biomedical Informatics & Medical Education, University ofWashington, Washington, WA, USA.Received: 19 January 2015 Accepted: 21 January 2015References1. van Zandbergen G, Klinger M, Mueller A, Dannenberg S, Gebert A, SolbachW, et al. Cutting edge: neutrophil granulocyte serves as a vector forLeishmania entry into macrophages. J Immunol. 2004;173:6521–5.2. Nandan D, Yi T, Lopez M, Lai C, Reiner NE. Leishmania EF-1alpha activatesthe Src homology 2 domain containing tyrosine phosphatase SHP-1 leadingto macrophage deactivation. J Biol Chem. 2002;277:50190–7.3. Nandan D, Lo R, Reiner NE. Activation of phosphotyrosine phosphataseactivity attenuates mitogen-activated protein kinase signaling and inhibitsc-FOS and nitric oxide synthase expression in macrophages infected withLeishmania donovani. Infect Immun. 1999;67:4055–63.4. Junghae M, Raynes JG. Activation of p38 mitogen-activated protein kinaseattenuates Leishmania donovani infection in macrophages. Infect Immun.2002;70:5026–35.5. McConville MJ, de Souza D, Saunders E, Likic VA, Naderer T. Living in aphagolysosome; metabolism of Leishmania amastigotes. Trends Parasitol.2007;23:368–75.Lambertz et al. BMC Genomics  (2015) 16:151 Page 25 of 266. Silverman JM, Chan SK, Robinson DP, Dwyer DM, Nandan D, Foster LJ, et al.Proteomic analysis of the secretome of Leishmania donovani. Genome Biol.2008;9:R35.7. Silverman JM, Clos J, de’Oliveira CC, Shirvani O, Fang Y, Wang C, et al. Anexosome-based secretion pathway is responsible for protein export fromLeishmania and communication with macrophages. J Cell Sci.2010;123:842–52.8. Kowal J, Tkach M, Thery C. Biogenesis and secretion of exosomes. Curr OpinCell Biol. 2014;29C:116–25.9. Silverman JM, Reiner NE. Exosomes and other microvesicles in infectionbiology: organelles with unanticipated phenotypes. Cell Microbiol.2011;13:1–9.10. Pegtel DM, Cosmopoulos K, Thorley-Lawson DA, van Eijndhoven MA,Hopmans ES, Lindenberg JL, et al. Functional delivery of viral miRNAs viaexosomes. Proc Natl Acad Sci U S A. 2010;107:6328–33.11. Giri PK, Kruh NA, Dobos KM, Schorey JS. Proteomic analysis identifies highlyantigenic proteins in exosomes from M. tuberculosis-infected and culturefiltrate protein-treated macrophages. Proteomics. 2010;10:1–13.12. Pope SM, Lasser C. Toxoplasma gondii infection of fibroblasts causes theproduction of exosome-like vesicles containing a unique array of mRNAand miRNA transcripts compared to serum starvation. J Extracell Vesicles.2013;2:22484.13. Mantel PY, Hoang AN, Goldowitz I, Potashnikova D, Hamza B, Vorobjev I,et al. Malaria-infected erythrocyte-derived microvesicles mediate cellularcommunication within the parasite population and with the host immunesystem. Cell Host Microbe. 2013;13:521–34.14. Prados-Rosales R, Baena A, Martinez LR, Luque-Garcia J, Kalscheuer R,Veeraraghavan U, et al. Mycobacteria release active membrane vesicles thatmodulate immune responses in a TLR2-dependent manner in mice. J ClinInvest. 2011;121:1471–83.15. Panepinto J, Komperda K, Frases S, Park YD, Djordjevic JT, Casadevall A,et al. Sec6-dependent sorting of fungal extracellular exosomes and laccaseof Cryptococcus neoformans. Mol Microbiol. 2009;71:1165–76.16. Bayer-Santos E, Guilar-Bonavides C, Rodrigues SP, Cordero EM, Marques AF,Varela-Ramirez A, et al. Proteomic analysis of Trypanosoma cruzi secretome:characterization of two populations of extracellular vesicles and solubleproteins. J Proteome Res. 2013;12:883–97.17. Garcia-Silva MR, das Neves RF, Cabrera-Cabrera F, Sanguinetti J, MedeirosLC, Robello C, et al. Extracellular vesicles shed by Trypanosoma cruzi arelinked to small RNA pathways, life cycle regulation, and susceptibility toinfection of mammalian cells. Parasitol Res. 2014;113:285–304.18. Silverman JM, Clos J, Horakova E, Wang AY, Wiesgigl M, Kelly I, et al.Leishmania exosomes modulate innate and adaptive immune responsesthrough effects on monocytes and dendritic cells. J Immunol.2010;185:5011–22.19. Ghosh J, Bose M, Roy S, Bhattacharyya SN. Leishmania donovani targetsDicer1 to downregulate miR-122, lower serum cholesterol, and facilitatemurine liver infection. Cell Host Microbe. 2013;13:277–88.20. Hassani K, Shio MT, Martel C, Faubert D, Olivier M. Absence of metalloproteaseGP63 alters the protein content of Leishmania exosomes. PLoS One.2014;9:e95007.21. Valadi H, Ekstrom K, Bossios A, Sjostrand M, Lee JJ, Lotvall JO.Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanismof genetic exchange between cells. Nat Cell Biol. 2007;9:654–9.22. Twu O, de Miguel N, Lustig G, Stevens GC, Vashisht AA, Wohlschlegel JA,et al. Trichomonas vaginalis exosomes deliver cargo to host cells andmediate hostratioparasite interactions. PLoS Pathog. 2013;9:e1003482.23. Garcia-Silva MR, Sanguinetti J, Cabrera-Cabrera F, Franzen O, Cayota A. Aparticular set of small non-coding RNAs is bound to the distinctive Argonauteprotein of Trypanosoma cruzi: insights from RNA-interference deficientorganisms. Gene. 2014;538:379–84.24. Lye LF, Owens K, Shi H, Murta SM, Vieira AC, Turco SJ, et al. Retention andloss of RNA interference pathways in trypanosomatid protozoans. PLoSPathog. 2010;6:e1001161.25. Ramirez C, Puerta C, Requena JM. Evidence of RNA editing in Leishmaniabraziliensis promastigotes. Parasitol Res. 2011;108:731–9.26. Maslov DA. Complete set of mitochondrial pan-edited mRNAs in Leishmaniamexicana amazonensis LV78. Mol Biochem Parasitol. 2010;173:107–14.27. Carver T, Harris SR, Berriman M, Parkhill J, McQuillan JA. Artemis: anintegrated platform for visualization and analysis of high-throughputsequence-based experimental data. Bioinformatics. 2012;28:464–9.28. Franzen O, Arner E, Ferella M, Nilsson D, Respuela P, Carninci P, et al. Theshort non-coding transcriptome of the protozoan parasite Trypanosomacruzi. PLoS Negl Trop Dis. 2011;5:e1283.29. Garcia-Silva MR, Frugier M, Tosar JP, Correa-Dominguez A, Ronalte-Alves L,Parodi-Talice A, et al. A population of tRNA-derived small RNAs is activelyproduced in Trypanosoma cruzi and recruited to specific cytoplasmicgranules. Mol Biochem Parasitol. 2010;171:64–73.30. Sobala A, Hutvagner G. Transfer RNA-derived fragments: origins, processing,and functions. Wiley Interdiscip Rev RNA. 2011;2:853–62.31. Didiano D, Hobert O. Perfect seed pairing is not a generally reliablepredictor for miRNA-target interactions. Nat Struct Mol Biol. 2006;13:849–51.32. Atayde VD, Shi H, Franklin JB, Carriero N, Notton T, Lye LF, et al. Thestructure and repertoire of small interfering RNAs in Leishmania (Viannia)braziliensis reveal diversification in the trypanosomatid RNAi pathway. MolMicrobiol. 2013;87:580–93.33. Peacock CS, Seeger K, Harris D, Murphy L, Ruiz JC, Quail MA, et al.Comparative genomic analysis of three Leishmania species that causediverse human disease. Nat Genet. 2007;39:839–47.34. Fu G, Barker DC. Characterisation of Leishmania telomeres reveals unusualtelomeric repeats and conserved telomere-associated sequence. NucleicAcids Res. 1998;26:2161–7.35. Bellingham SA, Coleman BM, Hill AF. Small RNA deep sequencing reveals adistinct miRNA signature released in exosomes from prion-infected neuronalcells. Nucleic Acids Res. 2012;40:10937–49.36. Nolte-‘t Hoen EN, Buermans HP, Waasdorp M, Stoorvogel W, Wauben MH,‘t Hoen PA, et al. Deep sequencing of RNA from immune cell-derived vesiclesuncovers the selective incorporation of small non-coding RNA biotypes withpotential regulatory functions. Nucleic Acids Res. 2012;40:9272–85.37. Huang X, Yuan T, Tschannen M, Sun Z, Jacob H, Du M, et al. Characterization ofhuman plasma-derived exosomal RNAs by deep sequencing. BMC Genomics.2013;14:319.38. Chan PP, Lowe TM. GtRNAdb: a database of transfer RNA genes detected ingenomic sequence. Nucleic Acids Res. 2009;37:D93–7.39. Miranda KC, Bond DT, Levin JZ, Adiconis X, Sivachenko A, Russ C, et al.Massively parallel sequencing of human urinary exosome/microvesicle RNAreveals a predominance of non-coding RNA. PLoS One. 2014;9:e96094.40. Rabinowits G, Gercel-Taylor C, Day JM, Taylor DD, Kloecker GH. ExosomalmicroRNA: a diagnostic marker for lung cancer. Clin Lung Cancer.2009;10:42–6.41. Michael A, Bajracharya SD, Yuen PS, Zhou H, Star RA, Illei GG, et al. Exosomesfrom human saliva as a source of microRNA biomarkers. Oral Dis. 2010;16:34–8.42. Vojtech L, Woo S, Hughes S, Levy C, Ballweber L, Sauteraud RP, et al.Exosomes in human semen carry a distinctive repertoire of small non-codingRNAs with potential regulatory functions. Nucleic Acids Res. 2014;42:7290–304.43. Schageman J, Zeringer E, Li M, Barta T, Lea K, Gu J, et al. The completeexosome workflow solution: from isolation to characterization of RNA cargo.Biomed Res Int. 2013;2013:253957.44. Bayer-Santos E, Lima FM, Ruiz JC, Almeida IC, da Silveira JF. Characterizationof the small RNA content of Trypanosoma cruzi extracellular vesicles. MolBiochem Parasitol. 2014;193:71–4.45. Colombo M, Raposo G, Thery C. Biogenesis, secretion, and intercellularinteractions of exosomes and other extracellular vesicles. Annu Rev Cell DevBiol. 2014;30:255–89.46. Bernal D, Trelis M, Montaner S, Cantalapiedra F, Galiano A, Hackenberg M,et al. Surface analysis of Dicrocoelium dendriticum. The molecularcharacterization of exosomes reveals the presence of miRNAs. J Proteomics.2014;105:232–41.47. Buck AH, Coakley G, Simbari F, McSorley HJ, Quintana JF, Le BT, et al. Exosomessecreted by nematode parasites transfer small RNAs to mammalian cells andmodulate innate immunity. Nat Commun. 2014;5:5488.48. Ender C, Krek A, Friedlander MR, Beitzinger M, Weinmann L, Chen W, et al. Ahuman snoRNA with microRNA-like functions. Mol Cell. 2008;32:519–28.49. Persson H, Kvist A, Vallon-Christersson J, Medstrand P, Borg A, Rovira C. Thenon-coding RNA of the multidrug resistance-linked vault particle encodesmultiple regulatory small RNAs. Nat Cell Biol. 2009;11:1268–71.50. Haussecker D, Huang Y, Lau A, Parameswaran P, Fire AZ, Kay MA. HumantRNA-derived small RNAs in the global regulation of RNA silencing. RNA.2010;16:673–95.51. Falaleeva M, Stamm S. Processing of snoRNAs as a new source of regulatorynon-coding RNAs: snoRNA fragments form a new class of functional RNAs.Bioessays. 2013;35:46–54.82. Rose M, Winston F, Hieter P. Methods in yeast genetics: a laboratory coursemanual. NY: Cold Spring Harbor Laboratory Press, Cold Spring Harbor; 1990.Lambertz et al. BMC Genomics  (2015) 16:151 Page 26 of 2652. Ullu E, Tschudi C, Chakraborty T. RNA interference in protozoan parasites.Cell Microbiol. 2004;6:509–19.53. Padmanabhan PK, Dumas C, Samant M, Rochette A, Simard MJ,Papadopoulou B. Novel features of a PIWI-like protein homolog in theparasitic protozoan Leishmania. PLoS One. 2012;7:e52612.54. De Gaudenzi JG, Carmona SJ, Aguero F, Frasch AC. Genome-wide analysisof 3′-untranslated regions supports the existence of post-transcriptionalregulons controlling gene expression in trypanosomes. Peer J. 2013;1:e118.55. D’Orso I, Frasch AC. TcUBP-1, a developmentally regulated U-rich RNA-bindingprotein involved in selective mRNA destabilization in trypanosomes. J BiolChem. 2001;276:34801–9.56. Walrad P, Paterou A, Costa-Serrano A, Matthews KR. Differential trypanosomesurface coat regulation by a CCCH protein that co-associates with procyclinmRNA cis-elements. PLoS Pathog. 2009;5:e1000317.57. Garcia Silva MR, Tosar JP, Frugier M, Pantano S, Bonilla B, Esteban L, et al.Cloning, characterization and subcellular localization of a Trypanosoma cruziargonaute protein defining a new subfamily distinctive of trypanosomatids.Gene. 2010;466:26–35.58. Jenjaroenpun P, Kremenska Y, Nair VM, Kremenskoy M, Joseph B, KurochkinIV. Characterization of RNA in exosomes secreted by human breast cancercell lines using next-generation sequencing. Peer J. 2013;1:e201.59. Li Z, Ender C, Meister G, Moore PS, Chang Y, John B. Extensive terminal andasymmetric processing of small RNAs from rRNAs, snoRNAs, snRNAs, andtRNAs. Nucleic Acids Res. 2012;40:6787–99.60. Weiberg A, Wang M, Lin FM, Zhao H, Zhang Z, Kaloshian I, et al. Fungalsmall RNAs suppress plant immunity by hijacking host RNA interferencepathways. Science. 2013;342:118–23.61. Garcia-Silva MR, Cabrera-Cabrera F, das Neves RF, Souto-Padron T, de SouzaW, Cayota A. Gene expression changes induced by Trypanosoma cruzi shedmicrovesicles in mammalian host cells: relevance of tRNA-derived halves.Biomed Res Int. 2014;2014:305239.62. Raina M, Ibba M. tRNAs as regulators of biological processes. Front Genet.2014;5:171.63. Kumar P, Anaya J, Mudunuri SB, Dutta A. Meta-analysis of tRNA derived RNAfragments reveals that they are evolutionarily conserved and associate withAGO proteins to recognize specific RNA targets. BMC Biol. 2014;12:78.64. Ivanov P, Emara MM, Villen J, Gygi SP, Anderson P. Angiogenin-inducedtRNA fragments inhibit translation initiation. Mol Cell. 2011;43:613–23.65. Gebetsberger J, Zywicki M, Kunzi A, Polacek N. tRNA-derived fragments targetthe ribosome and function as regulatory non-coding RNA in Haloferax volcanii.Archaea. 2012;2012:260909.66. Sobala A, Hutvagner G. Small RNAs derived from the 5′ end of tRNA caninhibit protein translation in human cells. RNA Biol. 2013;10:553–63.67. Maute RL, Schneider C, Sumazin P, Holmes A, Califano A, Basso K, et al.tRNA-derived microRNA modulates proliferation and the DNA damageresponse and is down-regulated in B cell lymphoma. Proc Natl Acad Sci U S A.2013;110:1404–9.68. Galizi R, Spano F, Giubilei MA, Capuccini B, Magini A, Urbanelli L, et al.Evidence of tRNA cleavage in apicomplexan parasites: Half-tRNAs as newpotential regulatory molecules of Toxoplasma gondii and Plasmodiumberghei. Mol Biochem Parasitol. 2013;188:99–108.69. Reifur L, Garcia-Silva MR, Poubel SB, Alves LR, Arauco P, Buiar DK, et al.Distinct subcellular localization of tRNA-derived fragments in the infectivemetacyclic forms of Trypanosoma cruzi. Mem Inst Oswaldo Cruz.2012;107:816–9.70. Gebetsberger J, Polacek N. Slicing tRNAs to boost functional ncRNAdiversity. RNA Biol. 2013;10:1798–806.71. Ebhardt HA, Thi EP, Wang MB, Unrau PJ. Extensive 3′ modification of plantsmall RNAs is modulated by helper component-proteinase expression. ProcNatl Acad Sci U S A. 2005;102:13398–403.72. Magoc T, Salzberg SL. FLASH: fast length adjustment of short reads toimprove genome assemblies. Bioinformatics. 2011;27:2957–63.73. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. NatMethods. 2012;9:357–9.74. Anders S, Pyl PT, Huber W. HTSeq – A Python framework to work withhigh-throughput sequencing data. Bioinformatics. 2015;31:166–9.75. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The SequenceAlignment/Map format and SAMtools. Bioinformatics. 2009;25:2078–9.76. Roberts A, Pimentel H, Trapnell C, Pachter L. Identification of noveltranscripts in annotated genomes using RNA-Seq. Bioinformatics.2011;27:2325–9.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistribution77. Rice P, Longden I, Bleasby A. EMBOSS: the European Molecular BiologyOpen Software Suite. Trends Genet. 2000;16:276–7.78. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al.Gapped BLAST and PSI-BLAST: a new generation of protein database searchprograms. Nucleic Acids Res. 1997;25:3389–402.79. Zhang Z, Schwartz S, Wagner L, Miller W. A greedy algorithm for aligningDNA sequences. J Comput Biol. 2000;7:203–14.80. Krieg P. A Laboratory guide to RNA: isolation, analysis, and synthesis. NewYork, NY: Wiley-Liss Inc.; 1996.81. Dinhopl N, Mostegl MM, Richter B, Nedorost N, Maderner A, Fragner K, et al.In situ hybridisation for the detection of Leishmania species in paraffinwax-embedded canine tissues using a digoxigenin-labelled oligonucleotideprobe. Vet Rec. 2011;169:525.Submit your manuscript at www.biomedcentral.com/submit

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52383.1-0223558/manifest

Comment

Related Items