expression of trypanosoma cruzi surface antigen fl-160 is controlled by elements in the 3′...

Molecular and Biochemical Parasitology 102 (1999) 53–66

Expression of Trypanosoma cruzi surface antigen FL-160 iscontrolled by elements in the 3% untranslated, the 3%

intergenic, and the coding regions�

David Weston 1, Anne C. La Flamme 2, Wesley C. Van Voorhis *Departments of Medicine and Pathobiology, Uni6ersity of Washington, Box 357185 Seattle, WA 98195, USA

Received 6 January 1999; accepted 8 April 1999

Abstract

The FL-160 surface antigen gene family of T. cruzi consists of hundreds of members of 160 kDa glycoproteinsexpressed in trypomastigotes, but not in epimastigotes. Steady-state levels of FL-160 mRNA were 80 to 100-foldhigher in trypomastigotes than in epimastigotes, yet transcription rates were equivalent between the lifecycle stages.Luciferase reporter constructs demonstrated that the 3% untranslated region (UTR) and intergenic region (IR)following the coding sequence of FL-160 was sufficient to generate 8-fold higher luciferase expression in trypomastig-otes compared with epimastigotes. Transfection of 3% UTR/IR deletion constructs revealed cis-acting elements whichconferred a trypomastigote-specific expression pattern similar to that of FL-160. Parasites treated with translation andtranscription inhibitors, cyclohexamide and Actinomycin D, respectively, displayed a stage-specific pattern of FL-160mRNA degradation. Epimastigotes, but not trypomastigotes, treated with the inhibitors accumulated a 1.4 KbFL-160 cleavage product. The cleavage site mapped to a 31 base poly-purine tract in the FL-160 coding region. Thefirst 526 aa of FL-160, containing the 31 base poly-purine tract and several smaller tracts, were fused to greenfluorescent protein (GFP) and expressed from the T. cruzi tubulin locus. Stable transformants expressed 4-fold moreFL-160:GFP fusion mRNA and 12-fold more fusion protein in the trypomastigote stage than in the epimastigote

www.elsevier.com/locate/parasitology

Abbre6iations: aa, amino acids; bp, base pairs; CAT, chloramphenicol acetyl-transferase gene; GFP, green fluorescent protein;HygTK, hygromycin resistance/thymidine kinase gene fusion; IR, intergenic region between two coding sequences not found inmature mRNA; Kb, kilobases; Kbp, kilobase pairs; kDa, kilodaltons; LUC, luciferase gene; ORF, open reading frame; UTR,untranslated region of mRNA.� Note : Nucleotide sequence data reported in this paper have been submitted to GenBank™ under accession numbers X70948,

AF080220, AF091835 and AF091836.* Corresponding author. Tel.: +1-206-5430821; fax: +1-206-6858681.E-mail address: [email protected] (W.C. Van Voorhis)1 Present address: Seattle Biomedical Research Institute, 4 Nickerson St., Seattle, WA 98109, USA.2 Present address: Cornell University, Deptartment of Microbiology and Immunology, C5-149 VMC, Ithaca, NY 14853, USA.

0166-6851/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved.

PII: S 0166 -6851 (99 )00079 -1

D. Weston et al. / Molecular and Biochemical Parasitology 102 (1999) 53–6654

stage suggesting post-transcriptional and translational control elements. These data reveal at least two distinct controlmechanisms for trypomastigote-specific expression of FL-160 surface glycoproteins, one involving the 3% UTR/IR andone involving the coding region of FL-160. © 1999 Elsevier Science B.V. All rights reserved.

Keywords: Trypanosoma cruzi ; Surface antigens; Gene expression; RNA stability

1. Introduction

The kinetoplastid parasite Trypanosoma cruzi,the causative agent of Chagas’ disease, is a ma-jor public health problem in Latin America.Millions of people are affected by this debilitat-ing and incurable disease [1,2]. Acute disease ismarked by parasite replication and, in nearly allcases, resolves spontaneously within the first fewweeks. Infection usually progresses undetecteduntil the chronic phase, when morbidity andmortality occur. During the lifecycle of T. cruzi,the parasite must adapt to very different envi-ronments requiring rapid morphological andbiochemical changes. In the intermediate hostreduviid bug, parasites multiply as extracellularepimastigotes in the midgut and subsequentlymigrate to the hindgut. In the hindgut, epi-mastigotes differentiate into metacyclic trypo-mastigotes and pass out of the vector with thefeces. The mammalian host is infected at a redu-viid bloodmeal when metacyclic trypomastigotesmake contact with mucous membrane or brokenskin. Once in the mammalian host, metacyclictrypomastigotes avoid the host’s many blood-stream defenses, locate and invade a wide vari-ety of host cell types. Within the cell,trypomastigotes escape the vacuole and enter thecytoplasm where they differentiate into replicat-ing amastigotes. After several rounds of divi-sion, amastigotes differentiate intotrypomastigotes, lyse the host cell and enter thebloodstream to infect new host cells.

The surface antigens of T. cruzi define theinterface between the parasite and the host’s im-mune system. Within the mammalian host, try-pomastigotes express a variety of surfaceantigens belonging to the sialidase/trans-sialidasesuperfamily [3]. These antigens include FL-160,[4], SA-85 [5], TSA [6], SAPA, [7], TCNA [8]

and vp85 [36]. Genes of the FL-160 gene familyencode 160 kDa glycoproteins which localize tothe trypomastigote flagellum and flagellarpocket. The FL-160 gene family consists ofabout 750 distinct genes, many of which aresimultaneously expressed [9], and found dis-persed throughout the genome on several differ-ent chromosomes [10]. High levels of steadystate FL-160 mRNA are observed in trypo-mastigotes, 80 to 100-fold more than the epi-mastigote stage. Little is known of themechanism of regulation of FL-160 gene expres-sion and how an entire gene family might becoordinately regulated. Understanding thismechanism is important as it addresses an im-portant biological function of T. cruzi and mayexpose novel, trypanosome-specific drug or vac-cine targets. Expression of other T. cruzi surfaceantigens, such as epimastigote-specific glyco-protein 72 (gp72) [11], trypomastigote-specificglycoprotein 85 (gp85) [12] and amastigote-spe-cific amastin [13], are all regulated post-tran-scriptionally. Sequences within the 5% and 3%untranslated regions (UTRs) and/or mRNAend-processing signals within the intergenic re-gions (IRs) play a role in regulating quantitativeand temporal expression of these genes [14,15].In other kinetoplastids, post-transcriptional reg-ulation of gene expression has been attributed toelements within the 5% UTR, 3% UTR, IRs andcoding region [16–19]. It is possible such se-quences play a similar role in regulating FL-160 expression. In this paper, we describe theparticipation of the FL-160 3% UTR/IR in thestage-regulation of FL-160 expression. In addi-tion, we present evidence that elements withinthe coding region of FL-160 also play a role intrypomastigote-specific expression of FL-160.Therefore, at least two distinct mechanisms maywork together to ensure proper trypomastigote-

D. Weston et al. / Molecular and Biochemical Parasitology 102 (1999) 53–66 55

specific expression of the FL-160 surface antigenfamily.

2. Materials and methods

2.1. Parasites

The T. cruzi CL strain [20] was used for allexperiments in this study. Trypomastigotes weregrown in vitro with murine BALB/c 3T3 fibrob-lasts in Dulbecco’s Minimal Essential Medium(Gibco BRL, Gaithersberg, MD) with 10% new-born calf serum or 10% cosmic calf serum (Hy-clone Laboratories, Inc., Logan, UT).Epimastigotes were grown in axenic cultures inliver-infusion/tryptone medium with 10% fetalcalf serum. Clones of transformed parasites wereobtained by limiting dilution.

2.2. Nuclear run-on transcription

Nuclei were isolated from epimastigotes andtrypomastigotes by 0.5% NP-40 lysis as describedelsewhere [21]. Run-on transcription was as fol-lows. Nuclei (2×107) were mixed with 150 mCia[32P]GTP (Amersham Life Sciences, ArlingtonHeights, IL), 10 mM unlabeled GTP, 2 mM ATP,CTP and UTP in buffer containing 1 U ml−1

RNAsin (Promega Corp., Madison, WI), 50 mMTris–HCl (pH 8.0), 25 mM NaCl, 2 mM dithio-threotol, 0.5 mM PMSF, 10 mM creatine phos-phate, and 20 mg ml−1 creatine phosphokinase.Reactions were incubated at 37°C for trypo-mastigote nuclei or 28°C for epimastigote nuclei.The labeled RNAs were allowed to hybridizewith 3 mg linearized and denatured plasmidDNAs applied by slot blot to Hybond-N (Amer-sham Life Sciences) nylon membranes. Targetplasmids included genomic clone pFL-160-2 [9],cDNA clones pSA85-1.1 and pSA85-1.2 [5] (giftof S. Kahn, Univ. of Washington), p154 trans-sialidase entire ORF (gift of D. Eichinger, NewYork Univ.) [22], pTC-UB1g calmodulin, poly-ubiquitin and FUS [23] (gift of J. Swindle, Univ.of Washington), and pBluescript (Stratagene).Blots were washed at high stringency and autora-diographed for 6 days.

2.3. Plasmid constructs

Luciferase reporter constructs were generatedfrom plasmid pGEM-luc (Promega Corp.), con-taining the firefly luciferase coding region. TheUTRs/IRs flanking the calmodulin A2 gene wereused to supply constitutive 5% and 3% mRNA endprocessing signals in the reporter constructs. The710 bp ‘BC’ Hind III-BamH I fragment was sub-cloned from pBS:CL-Neo-01/BC-x-10 [24], intothe respective sites of pGEM-luc. The 1240 bp ‘10’fragment was generated by PCR using theoligonucleotides UB31 and HSTOP [25], andcloned into the Sal I site of pGEM-luc. The 5%IR/UTR of FL-160 was generated by PCR, usingthe cDNA clone pFL-160-2 [9] as template(ACCc X70948). A gene-specific antisenseprimer, FL5% (5%-gaggatccgtcgggcctctgctgc-3%), anda vector-specific sense primer, complimentary tothe T7 promoter, were used to amplify the FL-1605% IR/UTR. This product and the pGEM-luc vec-tor were digested with Not I and made blunt withthe Klenow fragment of E. coli DNA polymerase(New England Biolabs, Beverly, MA), digestedwith BamH I and ligated together. The 3% UTR/IR of FL-160 was generated by subcloning the1,430 bp Sac II-EcoR I fragment from pFL-160-2(ACCc AF080220), to the respective sites ofpGEM-luc. The control plasmid, pBS:F-CAT-100[26] was a generous gift from Dr John Swindle.

For stable integration of foreign genes into thetubulin locus, plasmid pBS:THT-x-T was con-structed. In this pBluescript (Stratagene, La Jolla,CA) plasmid, the HygTK gene is flanked by btubulin UTRs/IRs and the gene of interest isflanked by a tubulin intergenic sequences. Theentire UTR/IR from the b tubulin ORF to the atubulin ORF (b–a UTR/IR) was generated bygenomic T. cruzi DNA PCR using the senseprimer SHB.S (5%-ggagcagtactaggtaggttggct-3%)and antisense primer BHB.AS (5%-cgcctcacgcatcct-gaaaagaagg-3%) (ACCc AF091835). The entireUTR/IR from the a tubulin ORF to the b tubulinORF (a-b UTR/IR) was generated by genomic T.cruzi DNA PCR using the sense primer KB2.S(5%-ggagtactagagtggcgctccc-3%) and antisenseprimer SB2.AS (5%-cacaatctcacgcattttcttgtcttg-3%)(ACCc AF091836). The Mlu I site in the b-a


UTR/IR was converted to a unique Sal I site byligating a self-annealing adapter (5%-cgcgacgtcgacgt -3%) into the Mlu I site. Antisenseprimer FLAS526 (5%-gagaattccggcacagtctgcctcat-gtcg-3%) and sense primer FLSTART (5%-gagtc-tagacatgtcccgtcgtgtt-3%) were used to amplify theDNA encoding the first 526 aa of FL-160-2 byPCR. Antisense primer TUB-STOP (5%-acta-gaattcgtactcctccacatcctcctc-3%) and sense primerTUB-START (5%-atagcgcgcatgcgtgaggcgatttg-cattc-3%) were used to amplify the DNA encodingthe entire ORF of a tubulin by PCR. Antisenseprimer GFP-STOP (5%-cttagcgcgctttgtatagttcatcc-3%) and sense primer GFP-START (5%-ctgaattcat-gagtaaaggagaagaact-3%) were used to amplify thegreen fluorescent protein (GFP) coding region,using the plasmid pKEN-GFP3 [27] as template, agenerous gift from Dr Stanley Falkow (StanfordUniversity School of Medicine, Stanford, CA).The FL-160 coding region fragment and GFPfragment were ligated in-frame to form the genefusion, FL-1601–526:GFP, which was subsequentlyligated into the Xba I site of pBS:THT-x-T togenerate plasmid pBS:THT-FL1–526:GFP-T. Thea tubulin coding region and GFP fragment wereligated in-frame to form the gene fusion,TUB:GFP which was subsequently ligated intothe Xba I site of pBS:THT-x-T to generate controlplasmid pBS:THT-TUB:GFP-T.

2.4. Transfections

Mid-log phase epimastigotes were harvested bycentrifugation (1000×g) and washed three timesin ice-cold epimastigote electroporation buffer(1×PBS with 0.5 mM MgCl2 and 0.1 mMCaCl2). Epimastigotes were resuspended at a finalconcentration of 1.5×10−9 ml in ice-cold epi-mastigote electroporation buffer. Culture derivedtrypomastigotes were harvested by centrifugation(1200×g) and washed three times in room tem-perature Dulbecco’s Minimal Essential Medium.Trypomastigotes were resuspended at a final con-centration of 1.2×109/ml in Dulbecco’s MinimalEssential Medium. For each transient transfectionexperiment 100 mg LUC reporter construct and 50mg CAT control construct were added to 0.3 mlparasites and incubated 5 min on ice or at room

temperature for epimastigotes and trypomastig-otes, respectively. Electroporations were per-formed using an Electro Cell Manipulator 600(BTX, San Diego, CA) and 0.2 cm cuvettes (BTX,Inc.). The settings used for epimastigotes were 340V/1500 mF/13V, and for trypomastigotes theywere 330 V/1500 mF/13 V. Protein lysates wereprepared 48 h after electroporation and measuredfor levels of CAT and LUC expression. Levels ofCAT activity were measured by a published pro-cedure [28]. Radiolabeled (14C) Chloramphenicolwas obtained from American Radiolabeled Chem-icals, Inc. (St. Louis, MO) and butryl-Co-enzymeA from Sigma (St. Louis, MO). Luciferase assayswere performed using the Luciferase Assay Sys-tem (Promega Corp., Madison, WI) and mea-sured over a 15 s period with a Monolight 2001luminometer (Analytical Luminescence Labora-tory, San Diego, CA).

For stable transfections, pTHT-FL1–526:GFP-Tand pBS:THT-TUB:GFP-T were linearized withSal I and 50 mg used to transfect epimastigotestage parasites. Transformants were selected forresistance to hygromycin (Boehringer–Mannheim, Mannheim, Germany) at 500 mgml−1, cloned and screened for correct integrationof the plasmid into the tubulin locus. GFP expres-sion was tested by western blot of SDS-PAGEfractionated parasite lysates by standard tech-niques [28]. Transfer membranes were PVDF Im-mobilon P (Millipore Corp., Bedford, MA).Polyclonal rabbit anti-GFP (Clontech Laborato-ries, Inc., Palo Alto, CA) was used at a 1:100dilution. Immunological detection was carried outwith alkaline phosphatase-labeled secondary anti-bodies and detection reagents (Promega Corp.,Madison, WI).

2.5. RNA and DNA analysis

Total parasite RNA was harvested using Ultra-spec RNA reagent (Biotecx Laboratories, Inc.,Houston, TX). Rapid amplification of cDNAends (RACE) was used to determine thepolyadenylation site for FL-160 mRNA. Trypo-mastigote RNA was reverse transcribed to firststrand cDNA by standard procedures [28] usingan oligo(dT) anchor primer (5%-cctctgaaggttcacg-


gatccacatctaga(t)18vn-3%), where 6=a, c or g andn=any base. First round nested PCR was per-formed using sense primer S1 (5%-ggttttttactcgt-gaagggaact-3%) and antisense primer B1(5%-cctctgaaggttcacgga-3%). Second round PCRused sense primer S2 (5%-catgtggtgaagcctgaaccca-3%) and antisense primer B2 (5%-cacggatccacatc-taga-3%) to amplify the 3% end.

Epimastigotes and trypomastigotes were treatedwith cyclohexamide (20 mg/ml) and ActinomycinD (10 mg/ml) and RNA harvested at timepointsfor northern blot analysis. Trypomastigotes wereremoved from 3T3 co-culture and placed undercell free conditions during exposure to the aboveinhibitors. All northern blots were performed us-ing BrightStar Plus (Ambion, Inc., Austin, TX)nylon membranes. Genomic DNA purificationand Southern blot analysis was performed usingstandard methodology [28]. Southern blots wereperformed using Hybond-N (Amersham Life Sci-ences) nylon membranes. The FL-160, GFP andHygTK coding regions were radiolabeled with 32Pby standard methodology. All blots were washedto high stringency and subjected to autoradiogra-phy. Signal quantification was achieved by phos-phoimaging analysis (Phosphoimager SP,Molecular Dynamics, Sunnyvale, CA) and theNIH Image Quant. software (Bethesda, MD).

3. Results and discussion

3.1. The 3 % UTR/IR contributes to thestage-specific expression of FL-160

Transcriptional regulation of FL-160 gene ex-pression was tested by nuclear run-on analysis.Three independent experiments confirmed thatequal rates of FL-160 transcription occurred intrypomastigotes and epimastigotes (Fig. 1). Thesedata were not surprising as other T. cruzi genestested [13,14], and most other kinetoplastid genestested, are primarily controlled by post-transcrip-tional mechanisms [29,30]. In these reports, theUTRs and IRs flanking the coding sequencesplayed a role in regulating levels of gene expres-sion. Therefore, the sequences flanking the FL-160-2 [9] coding region were tested for the ability

to regulate stage-specific expression in LUC re-porter constructs (Fig. 2). For clarity, the 3% UTRis the region of a gene beginning at the firstnucleotide downstream of the translational stopcodon and ending downstream at the polyadeny-lation site. The 5% UTR is the region of a genebeginning at the first nucleotide downstream ofthe splice leader acceptor site and ending at thefirst nucleotide upstream of the translational startcodon. The IR, not included in the maturemRNA, is the sequence beginning downstream ofthe polyadenylation site and ending immediatelyupstream of the splice leader acceptor site of adownstream gene. When LUC was flanked by theUTRs/IRs of the calmodulin A2 gene, a constitu-tively expressed gene, luciferase expression washigh in both trypomastigote and epimastigotestages (Fig. 2, construct BC-Luc-10). FlankingLUC with the UTRs/IRs of FL-160-2, resulted instage-specific expression of luciferase in trypo-mastigotes but not in epimastigotes (constructFL5%-Luc-FL3%). Next, the FL-160-2 5% and 3%UTRs/IRs were individually tested for the abilityto confer stage-regulated luciferase expression.

Fig. 1. Run-on analysis of transcription rates in epimastigotes(Epis) and trypomastigotes (Trypos). Nuclei were isolatedfrom Epis and Trypos and nascent chain RNA was labeled.This labeled RNA was used as a probe for plasmid DNAdenatured and bound to nylon membranes using a slot blotapparatus. The plasmids contained the coding regions of:FL-160-2 (a); SA85-1.1 (b); SA85-1.2 (c); trans-sialidase (d);and calmodulin, poly-ubiquitin, and FUS (e). A control plas-mid, pBluescript without a T. cruzi coding region, was boundto lane f.


Fig. 2. Luciferase reporter constructs and luciferase activity in epimastigote and trypomastigote stage parasites. The LUC gene wasflanked by UTRs/IRs from the calmodulin A2 gene (cross-hatched boxes), and UTRs/IRs from the FL-160-2 gene (gray boxes). Thearrowheads points to the general location of the FL-160 polyadenylation site. To control for electroporation variability, alltransfections included a CAT expression plasmid and all lysates were normalized for CAT activity before being measured for LUCactivity. Each construct was tested in each lifecycle stage at least five times.

When the 5% UTR/IR of FL-160-2 was tested,high levels of luciferase expression were found inboth epimastigote and trypomastigote stages(construct FL5%-Luc-10), suggesting the 5% UTR/IR alone does not regulate FL-160 expression.However, when the 3% UTR/IR of FL-160-2 wastested, luciferase expression was stage-specificallyexpressed in trypomastigotes but not in epimastig-otes (construct BC-Luc-FL3%). Therefore, thesedata suggest the FL-160-2 3% UTR/IR was suffi-cient to confer trypomastigote-specific expressionof LUC in transient transfection assays. However,expression levels of FL5%-Luc-FL3% and BC-Luc-FL3% were only 8-fold higher in trypomastigotesthan epimastigotes, well below the 80 to 100-folddifference observed with FL-160 mRNA levels[4,9].

Deletion analysis was undertaken in an attemptto identify regulatory elements within the FL-1603% UTR/IR. First, the wildtype polyadenylationsite was determined to establish the boundarybetween the 3% UTR and the IR. TrypomastigoteRNA was reversed transcribed and used as tem-plate for 3% RACE. Eight independent cDNAclones were sequenced to identify the polyadeny-lation site. FL-160 polyadenylation occurred in allclones at the three adenosine residues 640 basesdownstream of the translation termination codon(Fig. 3).

Deletion mutants were generated progressivelyinto the 3% end of the FL-160-2 UTR/IR. Con-structs BC-Luc-FL990, BC-Luc-FL760 and BC-Luc-FL570, contained the first 990 bp, 760 bp and570 bp of the FL-160 UTR/IR, respectively (Fig.


2). Construct BC-Luc-FL990 displayed statisti-cally similar trypomastigote-specific luciferase ex-pression as the full length 3% UTR/IR construct,BC-Luc-FL3%. However, testing deletion constructBC-Luc-FL760, which includes the complete 3%UTR and 115 bp of the IR, resulted in 3-foldhigher luciferase expression in epimastigotes thanthe BC-Luc-FL3% construct, and approx. 30% de-crease in trypomastigote expression. Deletion intothe FL-160 3% UTR, 70 bp upstream of the wild-type polyadenylation site (construct BC-Luc-FL570), resulted in 8-fold higher luciferaseactivity than the BC-Luc-FL3% construct in epi-mastigotes, with a corresponding decrease ofabout 40% in trypomastigote expression. Similarresults were obtained by inserting the three dele-tion constructs into the calmodulin locus as stablegene replacements ([24]; unpublished data).

The 3-fold loss of regulated luciferase expres-sion with construct BC-Luc-FL760 suggests thepresence of stage-regulatory sequences in the IRbetween positions 760 and 990 bp downstream ofthe translation stop. In other kinetoplastid sys-tems, polyadenylation and trans-splicing are cou-pled mechanisms, suggesting the IRs regulate thematuration of upstream genes [31–34]. DNA se-quence analysis revealed an ORF 3% to FL-160-2and several possible upstream AG dinucleotidesplice acceptor sites (Fig. 3). However, northern

analysis failed to identify a steady-state transcriptfrom this downstream region (not shown), sug-gesting it may represent a pseudogene. This puta-tive pseudogene could supply the trans-splicingsignals that couple with FL-160 polyadenylation,but more experimental data are required to testthis speculation.

Deletion of the 3% FL-160-2 UTR/IR seventybases 5% to the polyadenylation site (constructBC-Luc-FL570) resulted in even higher expression(8-fold) in epimastigotes relative to BC-Luc-FL990, but 40% less expression in trypomastig-otes. Interestingly, deletion of this portion of theFL-160 3% UTR/IR has a much more dramaticeffect on the stage-specific expression of FL-160in epimastigotes than in trypomastigotes. Thesedata suggest that the wildtype polyadenylationsite, and/or surrounding sequences, may be im-portant for stage-specific, and quantitative, ex-pression of FL-160. The UTR/IR of theprocyclin, VSG and hexose transporter genes ofT. brucei confer post-transcriptional, stage-specificgene expression [16,19,35]. In these reports, com-plex interactions of positive and negative regula-tory elements contribute to mRNA stability andtranslational efficiency. Preliminary data suggestthe 3% UTR and IR of FL-160 also containscomplex regulatory elements controlling stage-de-pendent, and stage independent, FL-160 expres-

Fig. 3. Nucleotide sequence of the FL-160-2 3% untranslated region and intergenic region (ACCc AF080220). The three adenosinenucleotides, where polyadenylation occurs are shown under the diamonds. The start ATG (double underline) and upstream AGsplice acceptor sites (bold type) of the putative downstream pseudogene are shown. The 3%-most termini of the deletion constructsare labeled and defined by the nucleotide under the double crosses.


sion (unpublished data). However, the highestlevels of stage-regulated expression conferred byany fragment of the 3% UTR/IR were between 6-and 9-fold differences between trypomastigoteand epimastigote stages. These observations sug-gest there may be other layers of post-transcrip-tional regulation to account for the 80 to 100-folddifference of FL-160 mRNA in the two stages.

3.2. Analysis of FL-160 mRNA half-life andstability

To address the stability of FL-160 transcripts,epimastigote and trypomastigote stage parasiteswere treated with the transcription inhibitor Acti-nomycin D and the translation inhibitor cyclohex-amide. Treatment of trypomastigote parasiteswith Actinomycin D showed a steady decrease inFL-160 mRNA levels (Fig. 4A). FL-160 mRNAhalf-life, in the trypomastigote stage, was esti-mated at 5 h. The limited quantity of FL-160mRNA in epimastigotes, detectable only by over-loading the gel, decreased upon treatment withActinomycin D (Fig. 4B). Determination of FL-160 half-life was not possible in epimastigotes.

Cyclohexamide treated epimastigotes showedan increase in FL-160 mRNA abundance whichpeaked at 2.5-fold over normal levels after 9 hexposure (Fig. 4C). Trypomastigotes also showedan increase in FL-160 mRNA abundance follow-ing exposure to cyclohexamide which peaked at2.2-fold over normal levels after 9 h (not shown).Surprisingly, following exposure to ActinomycinD (Fig. 4B) or cyclohexamide (Fig. 4C and D),epimastigotes showed an accumulation of a 1.4Kb FL-160 degradation product and an accumu-lation of high molecular weight FL-160-hybridiz-ing RNA, possibly representing polycistronicunprocessed RNA. However, trypomastigotes didnot show an accumulation of the 1.4 Kb degrada-tion product, nor did they accumulate the highmolecular weight RNA following exposure to ei-ther inhibitor (Fig. 4A and D). The probe used inthis experiment corresponds to the 5% end of FL-160, which suggests the FL-160 transcripts arecleaved 1.4 Kb into the 5% end of the transcript,within the coding region. In some experiments,the 1.4 Kb fragment was less abundant but joined

by the presence of smaller distinct degradationproducts (Fig. 4D). These data have several inter-pretations. First, the stage-specific accumulationof the FL-160 degradation products suggest thepresence of an epimastigote-specific RNase withendonuclease activity (Fig. 4B, C and D). Second,since these degradation products are accumulatingduring exposure to the inhibitors there may exist alabile RNase with exonuclease activity that wouldnormally clear the degradation products. Third,treatment of parasites with the inhibitors appearsto have resulted in the stage-specific accumulationof unprocessed FL-160 polycistronic RNA in epi-mastigotes (Fig. 4B, C and D, see arrowheads).This suggests the presence of a labile factor, de-pleted in epimastigotes but not trypomastigotesfollowing treatment with either inhibitor, involvedin processing of FL-160 mRNA.

3.3. The coding region contributes tostage-regulated expression of FL-160

A shorter fragment representing the FL-160 5%UTR was used to probe RNA from cyclohexam-ide treated parasites resulting in the same epi-mastigote-specific 1.4 Kb degradation product asdemonstrated with the longer FL-160 probe (Fig.4C). This strongly suggests the fragment containsthe 5% end of FL-160-2 and is generated by cleav-age of epimastigote FL-160 mRNA 1.4 Kb intothe coding region of the transcript. The cleavageof FL-160-2 transcript into the 1.4 Kb fragmentmaps to a 31 base poly-purine tract within theFL-160 coding region (Fig. 5). DNA sequenceanalysis of FL-160-2 revealed eight poly-purinetracts, consisting of at least 15 purines in lengthand not interrupted by two or more adjacentpyrimidines, in the 5% end of the ORF (Fig. 5).Therefore, it is possible one or more cis-actingpoly-purine tracts within the FL-160 coding re-gion may play a role in mRNA stability or trans-lational efficiency. To test this hypothesis theORF encoding the first 526 aa of FL-160, con-taining eight of ten FL-160 poly-purine tracts(Fig. 5), was fused in-frame to the ORF of GFP(FL1–526:GFP) for expression studies. If FL-160coding region elements could independently con-tribute to stage-regulated gene expression, then


Fig. 4. Analysis of FL-160 mRNA stability and half-life. Northern blots of (A) 12 mg RNA from trypomastigotes treated withActinomycin D (10 mg ml−1), (B) 60 mg RNA from epimastigotes treated with Actinomycin D and (C) 60 mg RNA fromepimastigotes treated with cyclohexamide (20 mg ml−1), fractionated on a 1.2% agarose gel. (D) Northern analysis of epimastigote-specific FL-160 degradation products. RNA from cyclohexamide treated epimastigotes (60 mg) and trypomastigotes (12 mg) washarvested at 33 and 48 h and fractionated on a 1.5% agarose gel. The probe used for blots A and B was the 1050 bp Pst I-EcoRV restriction fragment from plasmid pFL-160-2, representing the 5% two-thirds of the FL-160-2 ORF [9]. The probe used for blotsC and D was the FL-160-2 5% UTR/IR fragment (see Section 2). Lanes are numbered according to the hours of treatment withinhibitors. The arrowheads point to hybridizing high molecular weight RNAs at the compression zone of the gels. Size of RNA ismarked in kilobases.

levels of the FL1–526:GFP fusion mRNA and/orprotein would reflect this mechanism. As a con-trol, a tubulin which is expressed at similar levelsin epimastigotes and trypomastigotes, was fusedin-frame to GFP (TUB:GFP).

The FL1–526:GFP fusion gene and theTUB:GFP fusion gene were cloned into a bi-cistronic, tubulin-based expression plasmid, con-

taining the HygTK fusion for hygromycinresistance selection, to generate plasmid pTHT-FL1–526:GFP-T (Fig. 6) and pTHT-TUB:GFP-T(not shown). The Mlu I restriction site within theb–a UTR/IR of both plasmids were converted toa unique Sal I restriction site to allow for vectorlinearization and integration into the T. cruzitubulin locus, organized in tandem array (Fig. 6).


Hygromycin resistant clone D6 (FL1–526:GFP)was obtained and analyzed by Southern blot todetermine correct integration within the tubulinlocus. If the pTHT-FL1–526:GFP-T vector inte-grated correctly, it would be located between anupstream b tubulin and downstream a tubulingene. Restriction digestion of clone D6 genomicDNA with Bgl II generated the predicted 12.4Kbp band when probed with GFP and HygTK

(Fig. 7). Digestion of D6 genomic DNA with SacII generated the predicted 3.7 Kbp GFP band and3.2 Kbp and 1.4 Kbp HygTK bands. In addition,Sph I digested D6 genomic DNA yielded thepredicted 6.3 Kbp GFP band and 5.0 Kbp and2.6 Kbp HygTK bands (Fig. 7). These digestsdemonstrated that Sal I linearized pTHT-FL1–

526:GFP-T integrated in the tubulin locus asshown in Fig. 7. Directional PCR on D6 DNA,

Fig. 5. Nucleotide sequence of the first 526 aa of the FL-160-2 coding region. The sequence begins at the start methionine codon(dark-underlined) and ends at the FLAS526 primer sequence (underlined). The poly-purine tracts are shown in bold uppercase andnumbered to show relative position on the lower schematic diagram. The ratios of purines making up each tract are listed at theright.


Fig. 6. Schematic diagram of the FL-160:GFP fusion expression strategy. (A) Diagram of plasmid pTHT-FL1–526:GFP-T. Regionsmarked ‘a–b’ represent the UTR/IR between an upstream a tubulin and a downstream b tubulin coding region, while the regionsmarked ‘b–a’ represent the UTR/IR between an upstream b tubulin and downstream a tubulin coding region. (B) Diagram andrestriction map of clone D6 integration into the tubulin locus. Enzymes marked are Bgl II (B), Sac II (S) and Sph I (Sp).

using a tubulin and b tubulin specific primers withGFP and HygTK primers, added further supportthe integration went as predicted (not shown).The linearized control construct, pTHT-TUB:GFP-T, correctly integrated into the tubulinlocus as well (not shown).

To test for stage-dependent regulation ofmRNA, northern blots of D6 trypomastigotesand epimastigotes were probed with the GFPcoding region. The GFP probe did not hybridizeto untransfected parasite mRNA (not shown).Fig. 8 demonstrates that FL1–526:GFP mRNAwas expressed at higher levels in trypomastigotesthan epimastigotes. Quantitation of hybridizationresulted in a 4-fold increase in FL1–526:GFPmRNA in trypomastigotes over epimastigotes.The northern blots were stripped and reprobedwith a-tubulin ORF and equal amounts of a-tubulin mRNA were found in epimastigotes andtrypomastigotes (Fig. 8). Since the FL1–526:GFPwas expressed flanked by a-tubulin UTRs/IRs,the influence of the a-tubulin UTRs/IRs on differ-ential levels of FL1–526:GFP mRNA in epimastig-otes compared to trypomastigotes should benegligible. Parasites stably transfected with con-

trol construct pTHT-TUB:GFP-T expressed equalquantities of TUB:GFP fusion mRNA in bothtrypomastigotes and epimastigotes (not shown).

Fig. 7. Southern blot analysis of the pTHT-FL1–526:GFP-Tintegration into the tubulin locus. Genomic DNA (10 mg),from clone D6 was digested with restriction enzymes Bgl II(B), Sac II (S) and Sph I (Sp) and fractionated on a 0.7%agarose gel. Blots were probed with the entire coding region ofGFP and subsequently stripped and probed with the entirecoding region of the HygTK gene. DNA sizes are marked inkilobase pairs.


Fig. 8. Analysis of the stage-specific expression of the FL1–

526:GFP fusion gene. (A) Northern blot analysis of RNAharvested from clone D6 epimastigotes (E) and trypomastig-otes (T). The blot was probed with the coding region of GFPand subsequently stripped and probed with the coding regionof a tubulin as a loading control. (B) Western blot analysis ofthe FL1–526:GFP fusion protein from clone D6, epimastigote(E) and trypomastigote (T), protein lysates. Protein lysateconcentrations were determined by the Bradford assay and 50mg of each lysate fractionated by 10% SDS-PAGE. Equal laneloadings were confirmed by Ponceau S staining of the fraction-ated lysates after transfer to the nitrocellulose membranes [25].The numbers under each band represent the intensity of thebands.

other FL-160 cDNA and genomic clones revealedthat any nucleotide changes in the poly-purinetracts were maintained as purine substitutions(not shown). Similar poly-purine tracts can beidentified in the 5% end of other members of thestage-regulated trans-sialidase superfamily, suchas SA-85 [5], TSA1 [6], gp85 [12] and vp85 [36](not shown). Poly-purine tracts have been shownto be involved in post-transcriptional regulationof mRNA in another system. The half-life ofc-myc mRNA is regulated by a factor, whichbinds to a purine rich sequence in the codingregion and protects it from endonucleolytic attack[37]. The involvement of poly-purine tracts instage-dependent regulation needs to be tested bytransfection with further constructs that retain theORF of FL-160 but disrupt the poly-purinetracts.

To test for stage-dependent regulation ofprotein levels, western blots of clone D6 trypo-mastigote and epimastigote lysates were probedwith anti-GFP antibody. The expected band forthe fusion protein of FL1–526:GFP was observedat 90 kDa in lysates of transfected parasites (Fig.8), but was absent in lysates of untransfectedparasites (not shown). Trypomastigotes had a 12-fold increase in FL1–526:GFP protein levels com-pared with epimastigotes (Fig. 8). The ratio oftrypomastigote to epimastigote expression ofFL1–526:GFP protein was 3-fold greater than FL1–

526:GFP mRNA (12-fold vs. 4-fold). This sug-gests that the FL1–526 coding region contributes tostage-dependent regulation of FL-160 protein lev-els at the level of translational control. This isconsistent with the observation of different levelsof expression of FL-160 mRNA compared toprotein, as there are trace amounts of FL-160mRNA detectable in cultures of epimastigotes butprotein is not detectable in epimastigotes [4,9].The effect of the FL1–526 coding region could beon stage-dependent regulation of translation orprotein stability. If there is stage-dependent regu-lation of translation rates, this could be associatedwith the RNA degradation rates as RNA degra-dation machinery associates with polysomes ineukaryotes.

In summary, this report documents two inde-pendent mechanisms for stage-dependent regula-

This result suggests that GFP, expressed as afusion in the tubulin locus, itself is not under theinfluence of any stage-specific expression mecha-nisms. Taken together, these data suggest that thestage-specific increase in trypomastigote FL-160mRNA levels is partially conferred by cis-ele-ments within the FL-1601–526 sequence.

The observations that the FL1–526 region con-tains poly-purine tracts, that epimastigote-specificdegradation fragments are cleaved at, or near,these poly-purine tracts, and that this region con-fers increased expression of mRNA in trypo-mastigotes suggests that these poly-purine tractsare partially responsible for the stage dependentregulation of the FL-160 mRNA. Analysis of


tion of FL-160 mRNA expression, the 3% UTR/IRand the 5% coding region. These two mechanismstogether could explain about 30-fold difference inmRNA in trypomastigotes compared to epi-mastigotes though we cannot make precise quan-titative conclusions. Nonetheless, the effect of thetwo mechanisms together is closer to the observed80 to 100-fold difference in FL-160 mRNA levelsin trypomastigotes as compared to epimastigotesthan the effect of either mechanism alone. Thus,the stage-dependent regulation of FL-160 expres-sion is multifactorial and may involve additionalmechanisms than have been described here.

Acknowledgements

The authors would like to thank Lynn Barrettand Bhavesha Patel for their technical assistanceand Dr John Swindle for many helpful discus-sions. This work was supported by NIAID grantsR21 AI38924 and F32 AI09235.

References

[1] Brener Z. The biology of Trypanosoma cruzi. Ann RevMicro 1973;27:349–81.

[2] World Health Organization. Control of Chagas’ Disease.WHO technical series c811. World Health Organization,Geneva, 1991.

[3] Takle GB, Cross GAM. An 85 kDa surface antigen genefamily of Trypanosoma cruzi encodes polypeptides ho-mologous to bacterial neuraminidases. Mol Biochem Par-asitol 1991;48:185–98.

[4] Van Voorhis WC, Eisen H. FL-160, a surface antigen ofTrypanosoma cruzi that mimics mammalian nervous tis-sue. J Exp Med 1989;169:641–52.

[5] Kahn S, Van Voorhis WC, Eisen H. The major 85-kDsurface antigen of the mammalian form of Trypanosomacruzi is encoded by a large heterogeneous family of simul-taneously expressed genes. J Exp Med 1990;172:589–97.

[6] Fouts DL, Ruef BJ, Ridley PT, et al. Nucleotide se-quences and transcription of a trypomastigote surfaceantigen gene of Trypanosoma cruzi. Mol Biochem Para-sitol 1997;46:189–200.

[7] Pollevick GD, Affranchino JL, Frasch ACC, SanchezDO. The complete coding sequence of a shed acute-phaseantigen of Trypanosoma cruzi. Mol Biochem Parasitol1991;47:247–50.

[8] Pereira MEA, Mejia JS, Ortega-Barria E, Matzilevich D,Prioli RP. The Trypanosoma cruzi neuraminidase contains

sequences similar to bacterial neuraminidases, YWTDrepeats of the low density lipoprotein receptor, and typeIII modules of fibronectin. J Exp Med 1991;174:179–91.

[9] Van Voorhis WC, Barrett L, Koelling R, Farr AG.FL-160 proteins of Trypanosoma cruzi are expressed froma multigene family and contain two distinct epitopes thatmimic nervous tissues. J Exp Med 1993;178:681–94.

[10] Van Voorhis WC, Pham VT, De Vos T. The FL-160genes of Trypanosoma cruzi are dispersed on multiplechromosomes. Exp Parasitol 1995;80:578–82.

[11] Cooper R, Inverso JA, Espinoza M, Nogueira N, CrossGAM. Characterization of a candidate gene for GP72, aninsect stage-specific antigen of Trypanosoma cruzi. MolBiochem Parasitol 1991;49:45–60.

[12] Takle GB, Cross GAM. An 85 kDa surface antigen genefamily of Trypanosoma cruzi encodes polypeptides ho-mologous to bacterial neuraminidases. Biochem Mol Par-asitol 1991;48:185–98.

[13] Teixeira SMR, Russell DG, Kirchoff LV, Donelson JE. Adifferentially expressed gene family encoding ‘amastin’, asurface protein of Trypanosoma cruzi amastigotes. J BiolChem 1994;269:20509–16.

[14] Nozaki T, Cross GAM. Effects of 3% untranslated andintergenic regions on gene expression in Trypanosomacruzi. Mol Biochem Parasitol 1995;75:55–67.

[15] Teixeira SMR, Kirchhoff LV, Donelson JE. Post-tran-scriptional elements regulating expression of mRNAsfrom the amastin/tuzin gene cluster of Trypanosoma cruzi.J Biol Chem 1995;270:22586–94.

[16] Berberof M, Vanhamme L, Tebabi P, et al. The 3% termi-nal region of the mRNAs for VSG and procyclin canconfer stage specificity to gene expression in Trypanosomabrucei. EMBO J 1995;14:2925–34.

[17] Ramamoorthy R, Swihart KG, McCoy JJ, Wilson ME,Donelson JE. Intergenic regions between tandem gp63genes influence the differential expression of gp63 RNAsin Leishmania chagasi promastigotes. J Biol Chem1995;270:12133–9.

[18] Aly R, Argaman M, Halman S, Sharpira M. A regulatoryrole for the 5% and 3% untranslated regions in differentialexpression of hsp83 in Leishmania. Nucleic Acids Res1994;22:2922–9.

[19] Schurch N, Furger A, Kurath U, Roditi I. Contributionsof the procylin 3% untranslated region and coding regionto the regulation of expression in bloodstream forms ofTrypanosoma brucei. Mol Biochem Parasitol1997;89:109–21.

[20] Plata F, Pons FG, Eisen H. Antigenic polymorphism ofTrypanosoma cruzi : clonal analysis of trypomastigotessurface antigens. Eur J Immun 1984;14:392–8.

[21] Sriastava KK, Cable EE, Bonkovsky HL. Purifyingnascent RNA from run-on assays using guanidiniumisothiocyanate. Biotechniques 1993;15:226–7.

[22] Uemura H, Schenkman S, Nussenzweig V, Eichinger D.Only some members of a gene family in Trypanosomacruzi encode proteins that express both trans-sialidase andneuraminidase activities. EMBO J 1992;11:3837–44.


[23] Swindle J, Ajioka J, Eisen H, Sanwal B, Jacquemot C,Browder Z, Buck G. The genomic organization and tran-scription of the ubiquitin genes of Trypanosoma cruzi.EMBO J 1998;7:1121–7.

[24] Buckner FS, Verlinde CLMJ, La Flamme AC, VanVoorhis WC. Efficient technique for screening drugs foractivity against Trypanosoma cruzi using parasites express-ing b-galactosidase. Antimicrob Agents Chemother1996;40:2592–7.

[25] Chung SH, Gillespie RD, Swindle J. Analyzing expressionof the calmodulin and ubiquitin-fusion genes of Try-panosoma cruzi using simultaneous, independent dual genereplacements. Mol Biochem Parasitol 1994;63:95–107.

[26] Gillespie RD, Ajioka J, Swindle J. Using simultaneoustandem gene replacements to study expression of themulticopy ubiquitin-fusion (FUS) gene family of Try-panosoma cruzi. Mol Biochem Parasitol 1993;60:281–92.

[27] Cormack BP, Valdivia RH, Falkow S. FACS-optimizedmutants of the green fluorescent protein (GFP). Gene1996;173:33–8.

[28] Sambrook J, Fritsch E, Maniatis T. In Molecular Cloning:A Laboratory Manual. Cold Spring Harbor, NY: ColdSpring Harbor Laboratory Press, 1989.

[29] Graham SV. Mechanisms of stage-regulated gene expres-sion in kinetoplastida. Parasitol Today 1995;11:217–23.

[30] Clayton C, Hotz HR. Post-transcriptional control ofPARP gene expression. Mol Biochem Parasitol1996;77:1–6.

[31] LeBowitz JH, Smith HQ, Rushe L, Beverley SM. Cou-pling of poly(A) site selection and trans-splicing in Leish-mania. Genes Dev 1993;7:996–1007.

[32] Schurch N, Hehl A, Vassella E, Braun R, RoditiI. Accurate polyadenylation of procyclin mRNAs inTrypanosoma brucei is determined by pyrimidine-rich ele-ments in the intergenic regions. Mol Cell Biol1994;14:3668–75.

[33] Hug M, Hotz HR, Hartmann C, Clayton C. Heirarchiesof RNA-processing signals in a trypanosome surface anti-gen mRNA precursor. Mol Cell Biol 1994;14:7428–35.

[34] Mathews KR, Tschudi C, Ullu E. A common pyrimidine-rich motif governs trans-splicing and polyadenylation oftubulin polycistronic pre-mRNA in trypanosomes. GenesDev 1994;8:491–501.

[35] Hotz HR, Lorenz P, Fischer R, Krieger S, Clayton C.Role of 3% untranslated regions in the regulation of hexosetransporter mRNAs in Trypanosoma brucei. Mol BiochemParasitol 1995;75:1–14.

[36] Weston D, Patel B, Van Voorhis WC. Virulence in Try-panosoma cruzi infection correlates with the expression ofa distinct family of sialidase superfamily genes. MolBiochem Parasitol 1999;98:105–16.

[37] Bernstein PL, Herrick DJ, Prokipcak RD, Ross J. Controlof c-myc mRNA half-life in vitro by a protein capable ofbinding to a coding region stability determinant. GenesDev 1992;6:642–54.

.

.

expression of trypanosoma cruzi surface antigen fl-160 is controlled by elements in the 3′...

Documents