Aic

MLSa

b

c

d

e

f

g

a

ARRAA

KTtS

1

utssNitRd

0d

Molecular & Biochemical Parasitology 171 (2010) 64–73

Contents lists available at ScienceDirect

Molecular & Biochemical Parasitology

population of tRNA-derived small RNAs is actively producedn Trypanosoma cruzi and recruited to specificytoplasmic granules

aria Rosa Garcia-Silvaa, Magali Frugierb, Juan Pablo Tosara, Alejandro Correa-Dominguezc,ysangela Ronalte-Alvesc, Adriana Parodi-Taliced, Carlos Rovirae, Carlos Robelloa,f,amuel Goldenbergc, Alfonso Cayotaa,g,∗

Functional Genomics Unit, Institut Pasteur de Montevideo, Mataojo 2020, CP11400 Montevideo, UruguayArchitecture et Réactivité de l’ARN, Université de Strasbourg, CNRS, IBMC, 15 rue René Descartes, 67084 Strasbourg Cedex, FranceInstituto Carlos Chagas, ICC-Fiocruz-Paraná, CIC 81350-010 Curitiba, PR, BrazilDepartment of Genetics, Faculty of Sciences, Iguá 4225, CP11400 Montevideo, UruguayDepartment of Oncology, Lund University, BMC C13, 221 84 Lund, SwedenDepartment of Biochemistry, Faculty of Medicine, Gral Flores 2125, CP11800 Montevideo, UruguayDepartment of Medicine, Faculty of Medicine, Av Italia s/n, CP11800 Montevideo, Uruguay

r t i c l e i n f o

rticle history:eceived 21 August 2009eceived in revised form 4 February 2010ccepted 5 February 2010vailable online 13 February 2010

eywords:rypanosoma cruzi

a b s t r a c t

Over the last years an expanding family of small RNAs (i.e. microRNAs, siRNAs and piRNAs) was recognizedas key players in diverse forms of gene silencing and chromatin organization. Effectors functions ofthese small RNAs are achieved through ribonucleoprotein (RNP) complexes containing at their centeran Argonaute/Piwi protein. Although these proteins and their small RNA-associated machinery can betraced back to the common ancestor of eukaryotes, this machinery seems to be entirely lost or extensivelysimplified in some unicellular organisms including Trypanosoma cruzi, which are unable to trigger RNAirelated phenomena. Speculating about the presence of alternate small RNA-mediated pathways in theseorganisms, we constructed and analyzed a size-fractionated cDNA library (20–35 nt) from epimastigotes

RNA

mall non-coding RNAs forms of T. cruzi. Our results showed the production of an abundant class of tRNA-derived small RNAspreferentially restricted to specific isoacceptors and whose production was more accentuated undernutritional stress. These small tRNAs derived preferentially from the 5′ halves of mature tRNAs and wererecruited to distinctive cytoplasmic granules. Our data favor the idea that tRNA cleavage is unlikely tobe the consequence of non-specific degradation but a controlled process, whose biological significanceremains to be elucidated.

. Introduction

Over the last decades, a novel and expanding family of reg-latory small RNAs (sRNAs) has been discovered [1,2]. Based onheir biogenesis and effectors functions the sRNA family was clas-ified into three main subfamilies known as microRNAs (miRNAs),mall-interfering RNAs (siRNAs), and Piwi-interacting RNAs (piR-As) [3]. The sRNAs participate in a wide range of cellular processes

ncluding gene silencing at the transcriptional and postrancrip-ional levels, genome stability and chromatin organization [4–7].ecently, a new subfamily of small RNAs derived from tRNAs wasescribed in plants, fungi and animals. Increasing experimental evi-

∗ Corresponding author. Tel.: +598 522 09 10; fax: +598 522 41 85.E-mail address: cayota@pasteur.edu.uy (A. Cayota).

166-6851/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.molbiopara.2010.02.003

© 2010 Elsevier B.V. All rights reserved.

dence revealed that tRNA-derived halves participate in processes oftranslational repression and are able to downregulate gene expres-sion under the direction of small-guide tRNA halves by a processcomparable to that of the RNAi machinery (see Section 4 for furtherdetails)

Effector functions of sRNAs are achieved through RNA-associated ribonucleoprotein complexes which contain at theircenter an Argonaute/Piwi (Ago/Piwi) protein family member [8].Ago/Piwi are present in bacteria, archaea and eukaryotes andwere classified into three main paralogous groups: Argonaute-like proteins, Piwi-like proteins and the group 3 of Argonautes

restricted to C. elegans [9,10]. The Argonaute-like proteins areubiquitously expressed and bind to siRNAs or miRNAs to guidepost-transcriptional gene silencing whereas the expression ofPiwi-like proteins is mostly restricted to the germ line [11]. TheAgo/Piwi proteins can be traced back to the common ancestor of

Bioche

eesmtaadsCevhpltopttatRem(mpihat3uiRs

2

2

oulhsptaaaKpwtg1wpwtwR

M.R. Garcia-Silva et al. / Molecular &

ukaryotes [12]. However, this machinery seems to be eitherntirely lost or extensively simplified in some unicellular organismsuch as Saccharomyces cerevisiae, Trypanosoma cruzi, Leishmaniaajor and Plasmodium falciparum which are unable to utilize dsRNA

o trigger degradation of target RNA [13,14]. In species “RNAi neg-tive” such as T. cruzi and L. major a genomic search revealedputative Argonaute/Piwi-like gene without a recognizable PAZ

omain by primary sequence analysis for which the biologicalignificance is uncertain [14,15]. T. cruzi, which is responsible ofhagas’ disease, has a complex life cycle including intracellular andxtracellular forms which alternate between invertebrate insectectors belonging to the subfamily Triatominae and mammalianosts including humans [16]. Gene expression regulation in thesearasites seems to be achieved essentially at the postranscriptional

evel through modulation of mRNA stability and translational con-rol mechanisms [17] allowing rapid changes in gene expression inrder to cope with changing environments. The absence of sRNAathways in this and other related organisms has raised the ideahat probably theses pathways are either not essential for life orhey developed alternative pathways. Therefore, this work wasimed to identify small RNA pathways which could contribute tohe gene expression control in T. cruzi where a canonical smallNA pathway appears to be absent. We reported here the pres-nce of a homogeneous population of small RNAs derived fromature tRNAs representing about 25% of the small RNA population

herein named small tRNAs or tRNA halves). Additional experi-ents demonstrated that production of tRNA halves were more

ronounced in parasites undergoing nutritional stress. Interest-ngly, more than 98% of cloned tRNA halves derived from the 5′

alves of tRNAAsp and tRNAGlu. However, northern blot assayslso revealed a minor fraction of tRNA halves derived from otherRNA precursors. Surprisingly, small tRNAs derived from 5′ and′ halves were found associated to different cytoplasmic gran-les in T. cruzi epimastigotes. Whereas the biological significance

s currently unknown they could represent a subfamily of smallNAs with putative biological relevance in these unicellular para-ites.

. Material and methods

.1. T. cruzi epimastigote cultures and total RNA extraction

T. cruzi epimastigotes from the Dm28c clone were used through-ut this work and cultured either in normal growth conditions ornder nutritional stress. Unstressed epimastigotes were grown in

iver infusion tryptose (LIT) medium supplemented with 10% ofeat-inactivated fetal calf serum at 28 ◦C. Parasite cultures weretarted at 1 × 106 ml−1 and harvested in late logarithmic growthhase at a cell density of ∼50 × 106 ml−1 parasites and washedwice in phosphate-buffered saline supplemented with d-glucoset 2 g/l. When indicated Dm28c epimastigotes were submitted tonutritional stress by incubation for 3 h in TAU medium (Tri-

tome artificial urine medium) containing 190 mM NaCl, 17 mMCl, 2 mM MgCl2, 2 mM CaCl2, 0.035% sodium bicarbonate, 8 mMhosphate, pH 6.9 at room temperature. When nutritional stressas prolonged beyond 3 h the TAU medium was substituted by

he TAU3AAG medium (TAU medium supplemented with 10 mMlucose; aspartic acid 2 mM, glutamic acid 50 mM and proline0 mM). The TAU/TAU3AAG media are chemically defined mediahich reproduce the biological environment found in the most

osterior portion of the digestive tract of the invertebrate hosthere metacyclogenesis takes place as an adaptative response

o nutritional stress [18]. Stressed parasites were harvested andashed twice in PBS previously to total RNA extraction. TotalNA was then extracted from cell pellets by adding the Trizol

mical Parasitology 171 (2010) 64–73 65

reagent (Invitrogen) and processed according to the manufacturer’sinstructions.

2.2. Cloning of small RNAs

Cloning was performed as described previously [19,20] withminor modifications from 300 �g of total RNA. Briefly, small RNAsbetween 18 and 40 nt were size fractionated on a denaturing15% PAGE. Unlabeled DNA size markers were used to identifybands at the desired sizes which were excised according to themobility of the DNA markers as assessed by ethidium bromidestaining. After the ligation of 5′ and 3′ adaptors containing BanI restriction sites and amplification by PCR, Ban I-digested PCRproducts were concatamerized and cloned into pGEM T-easy(Promega Corp., Buenos Aires, Argentina) according to manu-facturer’s instructions. Colonies were screened using PCR andM13 sequencing primers. PCR products were directly submittedto custom sequencing. Sequence analysis and blast search wereperformed using information obtained from the public databasesGenBank (http://www.ncbi.nlm.nih.gov/Genbank/index.html),GeneDB (http://www.genedb.org/genedb/tcruzi/index.jsp) andTcruziDB (http://tcruzidb.org/tcruzidb/).

2.3. Immunofluorescence and fluorescence in situ hybridization(FISH)

For standard FISH assays epimastigotes were cultured either innormal growing conditions or submitted to a nutritional stress inTAU medium for 3 h. After cultures, parasites were washed twicein PBS and then fixed with 4% paraformaldehyde in PBS for 10 minat room temperature, washed twice with PBS and further incu-bated in 25 mM NH4Cl for 10 min. Parasites were allowed to adhereto polylysine-coated microscope slides for 20 min at room tem-perature and after washing twice in PBS cells were permeabilizedwith 0.2% Triton X-100 in PBS for 5 min. Slides were then blockedand prehybridized for 2 h at room temperature in bovine serumalbumine 2%, 5× Denhardt, 4× SSC and 35% deionized formamide(hybridization solution). Hybridization was performed overnight atroom temperature in a humid chamber at 37 ◦C in the presence of1 ng ml−1 of the indicated oligonucleotide conjugated either to flu-orescein isothiocyanate (FITC) or to the cyanine 3 dye (Cy3). Whenindicated FISH assays were performed under denaturing conditionsby heating slides at 75 ◦C for 5 min just prior to the hybridizationstep. After hybridization slides were washed once in 2× SSC plus50% deionized formamide, once in 2× SSC, once in 1× SSC plus 4,6-diamidino-2-phenylindole (DAPI) at 1 mg ml−1, and twice in 0.5×SSC. Slides were mounted with 5 ml of Fluor Save Reagent (DAKO).Analyses were performed in an Olympus IX 81 microscope coupledto a Hamamatsu Orca-ER camera (Diagnostic Instruments). Mergedimages were obtained by superimposing the indicated images files.

2.4. Nucleic acid probes and antibodies

Immunofluorescence assays aimed to detect specific cyto-plasmic structures using fluorescent labeled antibodies wereperformed simultaneously with FISH by co-incubating oligonu-cleotides probes in the presence of the respective antibody.Glycosomes were detected by using an anti-malate dehydroge-nase rabbit polyclonal antibody diluted at 1/500 (kindly giftedby Dr. Cristina Nowicki, Buenos Aires, Argentina). Reservosomes

were detected by using an anti-cruzipain rabbit polyclonal anti-body diluted 1/200 (kindly gifted by Dr. Juan Cazzullo, Buenos Aires,Argentina). After extensive washing a FITC-labeled goat anti-rabbitsecondary antibody was used to develop images in both cases. Syn-thetic oligonucleotide probes to detect 5′ and 3′ halves for different

66 M.R. Garcia-Silva et al. / Molecular & Biochemical Parasitology 171 (2010) 64–73

Fig. 1. Frequencies and properties of small RNAs and tRNA halves cloned from T. cruzi epimastigotes growing in optimal culture conditions. (A) Agarose gel electrophoresis(3%) showing small RNA populations after amplification of cloned RNAs ranging from 18 to 40 nt by using 5′ and 3′ adaptor specific primers. After electrophoresis agaroseg ution( age sF rsor t

tU1i

2

atlStapl7wcbp2ow22Mwrferr

els were ethidium bromide-stained and the negative image is depicted. (B) DistribC) Schematic representation of small tRNA sequences cloned showing the 5′ cleavrequency of different tRNA halves grouped according to isoacceptor types of precu

RNAs were purchased from Integrated DNA Technologies Inc. (IA,SA) and were labeled with FITC or with Cy3 (Supplementary Table). A synthetic poly-(dT)30 probe conjugated to Cy3 was used to

dentify poly-(A) + mRNAs.

.5. Northern blot analysis

A total of 30 �g of total RNA was run on 15% denaturing poly-crylamide gel in 0.5× Tris–Borate-EDTA running buffer. Prior toransfer, gels were stained with ethidium bromide to verify equaloading. The RNA was electrophoretically transferred to Gene-creen Plus Membranes (Schleicher and Schuell, NH, USA). Afterransfer, membranes were UV cross-linked at 120 mJ and bakedt 80 ◦C for 30 min. Oligonucleotide probes were generated by T4olynucleotide kinase-mediated end-labeling (New England Bio-

abs, Beverly, MA, USA) using [32P] �-ATP (end-labeling grade,000 Ci mmol−1; MP Biomedicals, Illkirch, Strasbourg, France) andere purified from unincorporated label with G-25 microspin

olumns. Filters were prehybridized at 40 ◦C in modified Church’suffer (5% sodium dodecyl sulfate (SDS), EDTA 1 mM in sodiumhosphate buffer 0.25 M, pH 7.2). After prehybridization for at leasth, radiolabeled specific probes were added and further incubatedvernight at 42 ◦C. Membranes were then washed twice at 42 ◦Cith 2× sodium chloride/sodium citrate buffer (SSC) 0.1% SDS for

0 min, 1× SSC 0.1% SDS for 20 min and 0.5× SSC and 0.1% SDS for0 min. The hybridized membranes were exposed to Kodak BiomaxS X-ray films. Hybridized probes were stripped from blots byashing twice in hot 0.1% SDS in deionized water for 15 min and

eprobed up to three times. The relative amounts of small tRNAragments were calculated using a densitometric analysis of north-rn blot signals. Values were normalized by using the respective 5SRNA signal as reference gene and expressed as a percentage of theespective precursor tRNA.

of 348 cDNA clones from T. cruzi epimastigotes according to functional attributes.ites on mature tRNAs and the number of times each one was cloned (arrows). (D)RNAs.

3. Results

3.1. Small RNA cloning from T. cruzi epimastigotes reveals ahomogenous population of tRNA 5′ halves restricted to specifictRNA isoacceptors

Total RNA was size fractionated by denaturating PAGE elec-trophoresis. Fragments from 20 to 35 nt were ligated to specific5′ and 3′ adaptors as described. After amplification using 5′ and3′ adaptor specific primers, small RNA fragments were distributedin three discrete bands corresponding to fragments of ∼22, ∼30and ∼40 nt as revealed by 3% agarose gel electrophoresis (Fig. 1A).Cloning and sequence analysis of the ∼40 nt population revealed ahigh number of random fragments corresponding to the mini-exondonor RNA (data not shown). Bands corresponding to ∼22 and ∼30nt were pooled and subjected to cDNA cloning. A total of 348 cloneswere obtained from the epimastigote small RNA library growingunder normal culture conditions. The clones were analyzed bypublic database search and classified according to their assignedfunction based on sequence identity (Fig. 1B). Four main small RNApopulations were identified of which two were assigned to ran-domly breakdown products of rRNA and mRNA. Most of sequencesidentified as products of mRNA degradation were within the ∼22nt population of small RNAs cloned. Sequences not matching thetrypanosome genome represented about 3% which could resultfrom either unexpected contaminating nucleic acids or genomicdifferences between the reference CL Brener and Dm28c clones.Eighty-seven small RNAs of ∼30 nt in length were assigned to the

tRNA category, representing ∼25% of all sequenced small RNAs. Allof these small tRNA-derived fragments, aligned only with the 5′

halves of mature tRNAs with a common 3′ end located from 1 to3 nucleotides 5′ just before the anticodon triplet (Fig. 1C). Morethan 90% of tRNA cleavage products were derived from only three

M.R. Garcia-Silva et al. / Molecular & Biochemical Parasitology 171 (2010) 64–73 67

Fig. 2. Northern blot analysis of tRNA-derived fragments from T. cruzi epimastigotes. Northern blots were probed using oligonucleotides recognizing either the 5′- or 3′-halvesof respective tRNAs. Total RNA was obtained from epimastigotes growing under optimal culture conditions (−) or undergoing nutritional stress (+) in TAU medium for atleast 3 h. A total of 30 �g was loaded per line and submitted to 15% denaturing PAGE. Gels were electrophoretically transferred to nylon membranes. Stressed and unstressedp the re( re shor ountss norm

tg(3AttwfS

3u

teccw3iiati

arasites were run side by side on the same gels and hybridized individually with∼72 nt) and the respective 5′ or 3′ cleavage products (∼30 nt). 5S rRNA levels aadiolabeled probes on the same strips used for the respective tRNA. The relative amtrip and expressed as a percentage of the respective mature tRNA (%). Values were

RNA isoacceptors (tRNAAspGUC, tRNAGlu

CUC and tRNAGluUUC), sug-

esting that small tRNAs are the product of a specific cleavageFig. 1D). The analysis of sequenced tRNAs halves reveals a mean of0 nt in length ranging from 29 to 33 nt (Supplementary Table 2).ll small tRNAs cloned derived from the 5′ halves of tRNAAsp

GUC,RNAGlu

CUC, tRNAGluUUC, tRNAGln

CUG, tRNAGlnUUG, tRNAAla

CGC andRNAGly

UCC exhibited a slight heterogeneity at their 3′ and 5′ endshich started at 1 to 3 nucleotides from de 5′ end and ended

rom 1 to 3 nucleotides before the anticodon triplet (Fig. 1C andupplementary Table 2).

.2. Production of tRNA-derived halves is extended to otherncloned tRNAs and is accentuated under nutritional stress

As a consequence of their extensive base modifications andheir highly stable tertiary structure, eukaryotic tRNAs are gen-rally refractory to reverse transcription and conventional cDNAloning [21]. Thus, we speculated about the potential bias in theloning procedure. For this, we designed a Northern blot assayith radiolabeled probes hybridizing separately to the 5′ and the

′ halves derived from cloned and uncloned representative tRNAs

soacceptors. Since recent reports have described tRNA cleavagento halves during nutritional and oxidative stress in plants, fungind mammals [22–27] the levels of mature tRNA and their respec-ive fragments were examined in the case of parasites growingn optimal culture conditions and compared to those undergo-

spective radiolabeled probes. Northern blot signals correspond to full-length tRNAwed for each experimental condition as reference loading controls using specificof small tRNA fragments calculated obtained by densitometry were added for eachalized by using the respective 5S rRNA signal as reference gene.

ing nutritional stress (depicted as (−) and (+) respectively inFig. 2).

Results clearly showed that a population of 5′ halves wasdetectable not exclusively for cloned tRNAAsp

GUC, tRNAGluCUC,

tRNAGluUUC and tRNAGly

UCC but also for the unrelated and unclonedtRNAHis

CGC, tRNAPheGAA, tRNAAsn

GUU and tRNAAlaCUC. However, lev-

els of tRNA-derived halves were significantly higher for clonedsmall tRNAs. We also detected a population of ∼30 nt sized RNAsoriginated from the 3′ arm of most tRNA tested (Fig. 2), but inall cases their levels were significantly lower than the respective5′ halves (Fig. 2 and Supplementary Table 3). These results sug-gest that tRNA cleavage is a more global phenomenon which alsoinvolves the generation of 3′ tRNA-derived halves. Nevertheless,the generation of tRNA halves appears to be preferentially directedtoward the generation of 5′-derived tRNA halves from a restrictedgroup of precursors tRNAs. Additionally, the higher levels of 5′-derived halves suggest that after endonucleolytic cleavage of theprecursor tRNAs the respective 3′-derived halves could be prefer-entially degraded.

As depicted in Fig. 2 and Supplementary Table 2, nutritionalstress induced a significant increase in most tRNA-derived halves

analyzed. As revealed by densitometric analysis, nutritional stressinduced a 2- to 4-fold increase in tRNA halves, irrespectively of their5′ or 3′ origin. Changes induced by nutritional stress in the gener-ation of tRNA halves for tRNAHis, tRNAPhe and tRNAAsp were eitherless pronounced or null.

68 M.R. Garcia-Silva et al. / Molecular & Biochemical Parasitology 171 (2010) 64–73

Fig. 3. Subcellular localization of 5′ and 3′ tRNA-derived fragments by FISH. Subcellular distribution of 5′ and 3′ halves derived from representative mature tRNAs (upperpanel tRNAGlu

UUC, middle panel tRNAAspGUC, bottom panel tRNAAla

CGC) performed in normal growing T. cruzi epimastigotes. Images were obtained using FITC-labeled probesrecognizing both the 5′ (5′ half probe) and the 3′ (3′ half probe) halves of each tRNA isoacceptor. Counterstaining with DAPI was used as a reference to identify nuclei andkinetoplasts. The respective differential interference contrast images are depicted (DIC) in order to identify the parasite cellular body and the respective flagellum origin(scale bars = 5 �m).

Fig. 4. Specificity of probes used in FISH assays. Representative experiment concerning probes used for 5′ halves derived from tRNAAspGUC performed in both non-denaturing

and denaturing conditions. Denaturing conditions were achieved as described in Section 2 by heating slides at 75 ◦C for 5 min just prior to probe addition. Anticodon bridgingprobes (middle panel) were designed to anneal with the anticodon loop which recognize only the anticodon loop of whole tRNA avoiding significant hybridization signalswith 5′ or 3′ halves. Shuffled probes harbor the same number and type of nucleotides present in the specific 5′ probe (upper panel) but in an unordered array (bottom panel).Scale bars = 5 �m.

Biochemical Parasitology 171 (2010) 64–73 69

3c

uflppdth

iarpls5ao(s(nc(bnhtttsnhc

Fig. 5. Validation of “shuffled” and “bridging” control probes by northern blot. The

Faoifl

M.R. Garcia-Silva et al. / Molecular &

.3. tRNA-derived 5′ and 3′ halves accumulate in specificytoplasmic granules

To determine the subcellular distribution of small tRNA pop-lations in intact T. cruzi epimastigotes, we carried out a set ofuorescent in situ hybridizations with probes designed to inde-endently recognize the 5′ and the 3′ halves of each representativerecursor tRNA. Experimental conditions used in FISH assays wereesigned to avoid denaturation of highly stable mature tRNAs andherefore to avoid preferential hybridization of the probes usedere with the respective full-length tRNAs.

Surprisingly, fluorescent signals obtained with probes recogniz-ng only the 5′ halves of representative tRNAs revealed a patternssociated to intense granular signals located in the posterioregion of the cell body (Fig. 3). The images obtained with fluorescentrobes specifically recognizing the 3′ halves also displayed a granu-

ar pattern though with lower signal intensities and a more diffuseubcellular distribution when compared to those associated with′-derived halves. To confirm the specificity of hybridization probesseries of experiments were performed under either denaturing

r non-denaturing conditions by using additional control probesFig. 4). For instance, a negative control probe was designed foreveral tRNA isoacceptors using randomly “shuffled” oligoprobesshuffled probes) which preserved the same nucleotide species andumber as well as the same hybridization temperature of the spe-ific probe. A second control probe bridging the 5′ and 3′ regionsbridging probes) spanning nucleotides from 7–9 nucleotides justefore and after the anticodon were designed in order to recog-ize only intact tRNA with minimal complementarity for 5′ and 3′

alves. As depicted in Fig. 4, randomly shuffled oligoprobes failedo hybridize with the respective tRNA halves when compared tohe original probe in both denaturing and non-denaturing condi-

ions. Additionally, bridging oligoprobes showed strong fluorescentignals under denaturing FISH while no signals were observed inon-denaturing conditions. These results prove that tRNA-derivedalves and not mature tRNAs are specifically recognized with spe-ific probes used in our experimental conditions. The specificity of

ig. 6. FISH analysis of representative 5′ tRNA-derived halves in epimastigotes submittedspartic acid (upper panel, tRNAAsp

GUC) and glutamic acid (bottom panel, tRNAGluUUC). T. c

r in the presence of TAU medium for 3 h (stressed epimastigotes). FISH assays were thdentify nuclei and kinetoplasts. The respective differential interference contrast imagesagellum origin (scale bars = 5 �m).

same membranes used for northern blot analysis of different tRNAs and their frag-ments were stripped and hybridized separately with each control probe and probesrecognizing the 5S rRNA as a loading control. Bands corresponding to whole tRNAsand 5S rRNA are indicated by arrows.

5′- and 3′-halves probes was further evaluated by northern blot(Fig. 5) and results were in agreement with FISH assays, thus vali-dating probe specificity in non-denaturing conditions.

As depicted in Fig. 6 the fluorescence intensity and number offluorescent foci obtained with representative 5′ probes were morepronounced when parasites were grown in conditions of nutritionalstress. These results were in agreement with northern blots (Fig. 2)which indicated that although endonucleolytic cleavage of tRNAwas present in unstressed T. cruzi epimastigotes, this process wasclearly accentuated when parasites were submitted to a nutritional

stress.

To distinguish between signals originated from 5′- and 3′-derived halves we designed probes labeled with fluorescent dyesdiffering in their emission wavelength. Epimastigotes under nor-mal growth conditions were fixed and stained simultaneously with

to a nutritional stress. Representative images obtained for 5′ halves derived fromruzi epimastigotes were cultured either in LIT medium (unstressed epimastigotes)en performed using FITC-labeled probes for 5′ halves with DAPI as a reference toare depicted (DIC) in order to identify the parasite cellular body and the respective

70 M.R. Garcia-Silva et al. / Molecular & Biochemical Parasitology 171 (2010) 64–73

Fig. 7. Immunofluorescence coupled to FISH for co-localization assays. Upper panel: Simultaneous detection of 5′ halves (Cy3) and 3′ halves (FITC) derived from therepresentative tRNAAsp

GUC. Middle panel: Immunofluorescence coupled to FISH depicting the relationship of tRNA 5′ halves (Cy3) with FITC-labeled anti-cruzipain antibodiesrecognizing reservosomes. Bottom panel: dual fluorescence assays showing the relationship of glycosomes (green fluorescence) with the representative 5′-derived half fromtRNAAsp

GUC (red fluorescence). The respective differential interference contrast images are depicted (DIC) in order to identify the parasite cellular body and the flagellumorigin (scale bars = 5 �m). Merged images were counterstained with DAPI to identify the nuclei and kinetoplasts. The embedded scatter grams in the bottom right corners ofmerged images are depicted to estimate the degree of overlapping signals of FITC (green, y-axis) and Cy3 (red, x-axis). Overlapping pixels of yellow color are depicted alongthe diagonal of the scatter gram. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig. 8. Subcellular localization and relationship to poly (A) + mRNA of 5′ tRNAAspGUC halves from normal and stressed T. cruzi epimastigotes. Subcellular localization of mRNA

(Poly (A) + mRNA) showed as red fluorescence (Cy3) and their relationship to 5′ tRNA halves in normal (upper panel, unstressed parasites) and stressed parasites for 24 h(bottom panel, stressed parasites) depicted as green (FITC) fluorescence. Merged images show superimposed images of selected tRNA halves and poly(A) + mRNA. DNAwas stained with DAPI. The respective differential interference contrast images are depicted (DIC) in order to identify the parasite cellular body and flagellum origin (scalebars = 5 �m). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Bioche

5fthmwpAtdpneba

[epnt(acmp

tltemttct

tcrs

4

kwagaeppmcPi[

paTpn(

M.R. Garcia-Silva et al. / Molecular &

′ Cy3-labeled (red label) and 3′ FITC-labeled (green label) probesor tRNAAsp

GUC used as a representative experiment. Interestingly,hese results clearly showed that most 5′- and 3′-derived tRNAalves did not co-localize and were recruited to distinct cytoplas-ic structures (upper panel in Fig. 7). The eventual co-localizationith morphologically similar organelles classically found in thesearasites such as reservosomes or glycosomes was also analyzed.s depicted in Fig. 7 (middle panel) the 5′-derived as well as

he 3′-derived granules appear with a weak association (embed-ed scatter gram) suggesting that although we cannot exclude aartial co-localization most small tRNA-containing granules didot co-localize with reservosomes. Additionally, there were novidences for significant co-localization with glycosomes (Fig. 7ottom panel). Similarly we did not find co-localization with eithercidocalcisomes or polysomes (data not shown).

Due to the recent description of poly(A) + mRNA granules28,29], additional experiments were designed to exclude theventual recruitment of tRNA-derived fragments in mRNA cyto-lasmic granules. For this, epimastigotes were grown either inormal conditions (unstressed parasites) or submitted to a nutri-ional stress by incubating parasites for 3 h in TAUAAG3 mediumstressed parasites). As depicted in Fig. 8, poly(A) + mRNA granulesssociated to nutritional stress did not co-localize with granulesontaining representative tRNA-derived 5′ halves. Further experi-ents also revealed that 3′-derived halves did not co-localize with

oly(A) + mRNA (data not shown).To determine whether both 5′- and 3′-halves from different

RNA precursors were recruited to the same respective granules, co-ocalization assays were performed combining oligoprobes for twoRNAs isoacceptors (tRNAAsp

GUC and tRNAGluUUC) as representative

xperiments. Results depicted in Supplementary Fig. 1, showed thatost 5′-derived halves from these two tRNAs strongly co-localized

o the same granules (panel A). Similar results were observed forhe respective 3′ halves (panel B). These results showed that pro-essed 5′ and 3′ halves from different isoacceptors were recruitedo the same respective granules.

Overall these results suggested that tRNA halves are recruitedo specific cytoplasmic granules which differ from other wellharacterized T. cruzi cytosplasmic granular structures. Similaresults were obtained with other tRNA isoacceptors (data nothown).

. Discussion

During the past decade, small non-coding RNAs have emerged asey players in diverse forms of gene regulation by a variety of path-ays. The widespread taxonomic distribution of the small RNAs

nd their associated machinery in all studied eukaryotic super-roups studied [11,12] suggested that these components werelready present in the last common ancestor of eukaryotes. How-ver, several unicellular organisms including S. cerevisiae and somearasitic protozoan such as T. cruzi and L. major appear having inde-endently lost their Argonaute genes together with the entire RNAiachinery [30]. The finding in these organisms of genomic ORFs

orresponding to Argonaute/Piwi-like proteins without a classicalAZ domain raised the intriguing question about the dispensabil-ty of small RNA-mediated pathways to accomplish their vital cycle15].

To examine whether alternate or ancestral small RNA-mediatedathways could be present in these organisms, we constructed and

nalyzed a size-fractionated cDNA library (from 20 to 35 nt) from. cruzi epimastigotes. Sequence analysis revealed a homogeneousopulation of small RNAs represented mainly by 5′ halves ∼30t in length derived from a restricted group of tRNA isoacceptorstRNAAsp

GUC, tRNAGluCUC and tRNAGlu

UUC) and probably resulting

mical Parasitology 171 (2010) 64–73 71

from a process of endonucleolytic cleavage in the anticodon loopnear or just before the 5′ side of the anticodon triplet.

Although cloned sequences were restricted to few tRNA pre-cursors, northern blot assays demonstrated that tRNA cleavage inT. cruzi was extended to other tRNAs such as tRNAAla and tRNAHis

although in lower levels. Also lower levels of the respective 3′ halvescould be detected. Thus, northern blots assays suggest that tRNAcleavage in T. cruzi epimastigotes can be characterized by a pref-erential generation of 5′-halves from at least certain full-lengthtRNAs such as tRNAAsp

GUC, tRNAGluCUC and tRNAGlu

UUC. The exis-tence of a highly homogeneous population of 5′ halves of ∼30 ntfavors the view that these molecules result from a highly controlledtRNA processing rather than from a non-specific degradation.

Recent studies reported endonucleolytic tRNA cleavage duringamino acid starvation in Tetrahymena thermophila and Aspergillusfumigatus [23,24], in the encystation process of Giardia lamblia[25], during development in the bacterium Streptomyces coelicolor[31], and in human cell lines, plants, and yeast systems duringoxidative stress [22,26,27]. Additionally, large-scale sequencing ofsmall RNAs revealed an abundant family of tRNA-derived halvesin human cell lines [32]. Several experimental evidences supportthe idea that production of tRNA-derived fragments are likely tobe specifically processed in plants, fungi and animals and do notappear as products of random tRNA cleavage [33]. Reported tRNAfragments from different species differ in terms of length, the ratioof halves originated from 5′ versus those from 3′ segments andthe frequency of processed tRNA isoacceptors acting as precur-sors which could reflect species-specific characteristics of tRNAprocessing.

As reported tRNA cleavage appears as a conserved pathwayin prokaryotes as well as in eukaryotes frequently initiated afternutritional, biological or physicochemical stress. Although, tRNAhalves described in the present work have been identified in par-asites growing in the absence of any form of physicochemical ornutritional stress, when epimastigotes were grown under nutri-tional stress a pronounced increase in small tRNAs productionwas observed. Similarly, recent reports demonstrated that humancell lines produced a variety of tRNA-derived fragments in normalgrowing conditions as well as under oxidative stress [26,32]. How-ever, we cannot completely exclude that at least a fraction of tRNAfragments could be by-products of tRNA turnover which in turncould increase on starvation conditions.

Subcellular localization of tRNA halves by FISH revealed thatthese molecules were recruited to specific granular structures inthe cytoplasm. Our results showed that 5′- and 3′-derived halveswere recruited to different granules, with 5′ halves granules locatedpreferentially in the posterior region of the cellular body. Thestructure and composition of these granules is unknown and itselucidation will require specific experimental designs.

A major and intriguing issue is the biological significanceof tRNA-derived fragments which remain to be completely elu-cidated. In eukaryotes, tRNA cleavage has been proposed asa mechanism to downregulate or inhibits protein synthesisunder conditions of differentiation into relatively resting states[23,25,31], oxidative stress [26] as well as under nutrient starva-tion [24]. However, two recent reports have provided experimentalevidence that tRNA fragments might have a role in translationrepression [27,34]. Thus, tRNA cleavage into halves emerges as anevolutionary conserved pathway which is unlikely to be the con-sequence of a meaningless random degradation phenomenon witha putative role in the regulation of gene expression under certain

survival pressures. Of note, in Drosophila, analogous tRNA-derivedfragments comprise a small subset of Piwi-associated RNAs (piR-NAs), which suggest that tRNA anticodon cleavage could lead tothe assembly of specific ribonucleoprotein complexes involved inpiRNA related pathways [35].

7 ioche

tdt[ZZahrbpip

oRwtZpfttsnto

ipncien

otvchtsrocspip

cbiR

A

Mt

ty

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

2 M.R. Garcia-Silva et al. / Molecular & B

Surprisingly, recent experimental evidences demonstrated thathe long form of the tRNA-maturing enzyme tRNase ZL was able toownregulate expression of human genes by degrading the respec-ive mRNAs under the direction of small-guide tRNA-derived halves36,37]. In this respect, the efficacy of the small-guide RNA/tRNaseL was comparable to that of the RNAi machinery. The tRNasebelongs to the metallo-�-lactamase family of proteins which

re present in all three kindoms [38]. Two forms of the enzymeave been described, a short form (tRNase ZS) encoded by bacte-ial and archaeal genomes and the long form (tRNase ZL) encodedy eukaryote genomes. Since databank searches revealed that try-anosomatid genomes encode for at least 4 metallo-�-lactamases

t is tempting to speculate whether a similar tRNase Z-dependentathway involved in gene silencing could exist in these parasites.

Interestingly, several experimental evidences suggested vari-us functional links between the canonical Dicer-dependent smallNAs and the small-guide RNA/tRNase ZL pathways. In agreementith this, a subset of human microRNAs forming hook-like struc-

ures could guide target mRNA cleavage by human cytosolic tRNaseL in vitro [39]. On the other hand eukaryotic cells were reported toroduce a homogeneous population of ∼19 nt small RNAs derivedrom tRNAs by means of a Dicer-dependent process [40]. Although,rypanosomatids lack of Dicer homologues it is noteworthy thatRNase Z enzymes may participate in gene silencing mechanismsimilarly to the more elaborated pathways involved in RNAi phe-omena but using tRNAs halves as small-guide RNAs. Undoubtedly,his is an attractive hypothesis to explore the biological significancef tRNA-derived small RNAs in these and other unicellular parasites.

Another important concern is related to the mechanismsnvolved in this site specific cleavage of tRNAs. In bacteria, thelasmid-encoded colicins D and E5 [41,42] and the anticodonuclease PrrC [43] target some tRNAs for cleavage within the anti-odon loop in response to stress-induced signals and T4 phagenfection respectively resulting in translational inhibition. How-ver, to the best of our knowledge there are no descriptions of suchucleases in eukaryotes.

A recent report demonstrated that angiogenin, a ribonucleasef the RNase A superfamily was responsible for the cleavage ofRNA into tRNA halves in mammalian cells both in vivo and initro [22]. However, the search for putative genes bearing ribonu-lease A motifs in public databases (http://tcruzidb.org/tcruzidb/;ttp://www.genedb.org/genedb/tcruzi/) failed to reveal hits forrypanosomatids. More recently, Thompson et al. [44] demon-trated that in yeast the T2 RNase family member Rny1p wasesponsible of rRNA and tRNA cleavage at the anticodon loop duringxidative stress. Thus, T2 RNase enzymes emerge also as relevantandidates for tRNA cleavage in trypanosomatids. Although, theearch for the corresponding orthologs in trypanosomatids fromublic databases do not reveal ORFs with significant sequence sim-

larity we cannot exclude a member of this family involved in tRNArocessing with a significant degree of divergence.

Whatever the case, more studies are needed to completely elu-idate the mechanisms and pathways of tRNA cleavage and theiriological function in trypanosomatids which could provide new

nsights in regulatory networks mediated by small non-codingNAs.

cknowledgments

This work was supported partially by the “Institut Pasteur de

ontevideo” and Conselho Nacional de Desenvolvimento Cien-

ifico e Tecnologico (CNPq) Brazil.MRGC is a research fellow from the “Agencia Nacional de Inves-

igación e Innovacion (ANII)” del Ministerio de Educacion y CulturaPresidencia de la Republica, Montevideo Uruguay.

[

[

mical Parasitology 171 (2010) 64–73

CR, AC and APT are researchers from the Sistema Nacional Inves-tigadores (ANII), Montevideo Uruguay.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.molbiopara.2010.02.003.

References

[1] Carninci P, Hayashizaki Y. Noncoding RNA transcription beyond annotatedgenes. Curr Opin Genet Dev 2007;17:139–44.

[2] Eddy SR. Non-coding RNA genes and the modern RNA world. Nat Rev Genet2001;2:919–29.

[3] Chu CY, Rana TM. Small RNAs: regulators and guardians of the genome. J CellPhysiol 2007;213:412–9.

[4] Ambros V. The functions of animal microRNAs. Nature 2004;431:350–5.[5] Elbashir SM, Lendeckel W, Tuschl T. RNA interference is mediated by 21- and

22-nucleotide RNAs. Genes Dev 2001;15:188–200.[6] Klattenhoff C, Theurkauf W. Biogenesis and germline functions of piRNAs.

Development 2008;135:3–9.[7] Zamore PD, Tuschl T, Sharp PA, Bartel DP. RNAi: double-stranded RNA directs

the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell2000;101:25–33.

[8] Farazi TA, Juranek SA, Tuschl T. The growing catalog of small RNAs andtheir association with distinct Argonaute/Piwi family members. Development2008;135:1201–14.

[9] Carmell MA, Xuan Z, Zhang MQ, Hannon GJ. The Argonaute family: tenta-cles that reach into RNAi, developmental control, stem cell maintenance, andtumorigenesis. Genes Dev 2002;16:2733–42.

10] Yigit E, Batista PJ, Bei Y, et al. Analysis of the C. elegans Argonaute family revealsthat distinct Argonautes act sequentially during RNAi. Cell 2006;127:747–57.

11] Hutvagner G, Simard MJ. Argonaute proteins: key players in RNA silencing. NatRev Mol Cell Biol 2008;9:22–32.

12] Cerutti H, Casas-Mollano JA. On the origin and functions of RNA-mediatedsilencing: from protists to man. Curr Genet 2006;50:81–99.

13] Robinson KA, Beverley SM. Improvements in transfection efficiency and testsof RNA interference (RNAi) approaches in the protozoan parasite Leishmania.Mol Biochem Parasitol 2003;128:217–28.

14] Ullu E, Tschudi C, Chakraborty T. RNA interference in protozoan parasites. CellMicrobiol 2004;6:509–19.

15] Durand-Dubief M, Bastin P. TbAGO1, an argonaute protein required for RNAinterference, is involved in mitosis and chromosome segregation in Try-panosoma brucei. BMC Biol 2003;1:2.

16] Barrett MP, Burchmore RJ, Stich A, et al. The trypanosomiases. Lancet2003;362:1469–80.

17] Clayton C, Shapira M. Post-transcriptional regulation of gene expression intrypanosomes and leishmanias. Mol Biochem Parasitol 2007;156:93–101.

18] Contreras VT, Araujo-Jorge TC, Bonaldo MC, et al. Biological aspects of the Dm28c clone of Trypanosoma cruzi after metacyclogenesis in chemically definedmedia. Mem Inst Oswaldo Cruz 1988;83:123–33.

19] Marton S, Garcia MR, Robello C, et al. Small RNAs analysis in CLL reveals aderegulation of miRNA expression and novel miRNA candidates of putativerelevance in CLL pathogenesis. Leukemia 2008;22:330–8.

20] Pfeffer S, Lagos-Quintana M, Tuschl T. Cloning of small RNA molecules. In:Ausubel FM, Brent R, Kingston RE, et al, editors. Current Protocols in MolecularBiology. New York: Wiley and Sons; 2003. pp. 26.24.21–16.

21] Huttenhofer A, Kiefmann M, Meier-Ewert S, et al. RNomics: an experimentalapproach that identifies 201 candidates for novel, small, non-messenger RNAsin mouse. Embo J 2001;20:2943–53.

22] Fu H, Feng J, Liu Q, et al. Stress induces tRNA cleavage by angiogenin in mam-malian cells. FEBS Lett 2009;583:437–42.

23] Jochl C, Rederstorff M, Hertel J, et al. Small ncRNA transcriptome analysis fromAspergillus fumigatus suggests a novel mechanism for regulation of proteinsynthesis. Nucleic Acids Res 2008;36:2677–89.

24] Lee SR, Collins K. Starvation-induced cleavage of the tRNA anticodon loop inTetrahymena thermophila. J Biol Chem 2005;280:42744–9.

25] Li Y, Luo J, Zhou H, et al. Stress-induced tRNA-derived RNAs: a novel class ofsmall RNAs in the primitive eukaryote Giardia lamblia. Nucleic Acids Res 2008.

26] Thompson DM, Lu C, Green PJ, Parker R. tRNA cleavage is a conserved responseto oxidative stress in eukaryotes. RNA 2008;14:2095–103.

27] Zhang S, Sun L, Kragler F. The phloem-delivered RNA pool contains small non-coding RNAs and interferes with translation. Plant Physiol 2009;150:378–87.

28] Cassola A, De Gaudenzi JG, Frasch AC. Recruitment of mRNAs to cytoplasmicribonucleoprotein granules in trypanosomes. Mol Microbiol 2007;65:655–70.

29] Holetz FB, Correa A, Avila AR, Nakamura CV, Krieger MA, Goldenberg S. Evidenceof P-body-like structures in Trypanosoma cruzi. Biochem Biophys Res Commun

2007;356:1062–7.

30] Shabalina SA, Koonin EV. Origins and evolution of eukaryotic RNA interference.Trends Ecol Evol 2008;23:578–87.

31] Haiser HJ, Karginov FV, Hannon GJ, Elliot MA. Developmentally regulated cleav-age of tRNAs in the bacterium Streptomyces coelicolor. Nucleic Acids Res2008;36:732–41.

Bioche

[

[

[

[

[

[

[

[

[

[

[

M.R. Garcia-Silva et al. / Molecular &

32] Kawaji H, Nakamura M, Takahashi Y, et al. Hidden layers of human small RNAs.BMC Genomics 2008;9:157.

33] Li Y, Zhou H. tRNAs as regulators in gene expression. Sci China C Life Sci2009;52:245–52.

34] Yamasaki S, Ivanov P, Hu GF, Anderson P. Angiogenin cleaves tRNA and pro-motes stress-induced translational repression. J Cell Biol 2009;185:35–42.

35] Brennecke J, Aravin AA, Stark A, et al. Discrete small RNA-generating loci as mas-ter regulators of transposon activity in Drosophila. Cell 2007;128:1089–103.

36] Elbarbary RA, Takaku H, Uchiumi N, et al. Modulation of gene expression by

human cytosolic tRNase Z(L) through 5′-half-tRNA. PLoS One 2009;4:e5908.

37] Nakashima A, Takaku H, Shibata HS, et al. Gene silencing by the tRNA maturasetRNase ZL under the direction of small-guide RNA. Gene Ther 2007;14:78–85.

38] Vogel A, Schilling O, Spath B, Marchfelder A. The tRNase Z family of proteins:physiological functions, substrate specificity and structural properties. BiolChem 2005;386:1253–64.

[[

mical Parasitology 171 (2010) 64–73 73

39] Elbarbary RA, Takaku H, Uchiumi N, et al. Human cytosolic tRNase ZL can down-regulate gene expression through miRNA. FEBS Lett 2009;583:3241–6.

40] Cole C, Sobala A, Lu C, et al. Filtering of deep sequencing data reveals theexistence of abundant Dicer-dependent small RNAs derived from tRNAs. RNA2009.

41] Ogawa T, Tomita K, Ueda T, Watanabe K, Uozumi T, Masaki H. A cyto-toxic ribonuclease targeting specific transfer RNA anticodons. Science1999;283:2097–100.

42] Tomita K, Ogawa T, Uozumi T, Watanabe K, Masaki H. A cytotoxic ribonuclease

which specifically cleaves four isoaccepting arginine tRNAs at their anticodonloops. Proc Natl Acad Sci USA 2000;97:8278–83.

43] Kaufmann G. Anticodon nucleases. Trends Biochem Sci 2000;25:70–4.44] Thompson DM, Parker R. The RNase Rny1p cleaves tRNAs and promotes

cell death during oxidative stress in Saccharomyces cerevisiae. J Cell Biol2009;185:43–50.