Molecular and Biochemical Parasitology 98 (1999) 105–116

Virulence in Trypanosoma cruzi infection correlates with theexpression of a distinct family of sialidase superfamily genes�

David Weston a, Bhavesha Patel b, Wesley C. Van Voorhis b,*a Seattle Biomedical Research Institute, 4 Nickerson St., Suite 200, Seattle, WA 98109, USAb Uni6ersity of Washington, Department of Medicine, Box 357185, Seattle, WA 98195, USA

Received 7 July 1998; received in revised form 4 October 1998; accepted 4 October 1998

Abstract

The overall success of Trypanosoma cruzi depends on its ability to invade the host and establish a long-terminfection. Little is known of the genetic factors responsible for observed differences in virulence from strain to strainin T. cruzi. A virulent T. cruzi line was derived from an attenuated parental line by two passages through mice. Toidentify virulence genes a subtraction library was constructed and screened for cDNA expressed exclusively in thevirulent line. One cDNA hybridized to 3.5 and 4.5 Kb RNA present in virulent trypomastigotes but absent inattenuated trypomastigotes. Sequence analysis showed the cDNA to encode an 85 kDa protein with homology tomembers of the sialidase/trans-sialidase superfamily and has been designated vp85.1. The highest amino acid sequencesimilarity was to a previously described T. cruzi sialidase-homologue pseudogene [Takle, G.B., O’Conner, J., Young,A.J. and Cross, G.A.M. (1992) Mol. Biochem. Parasitol. 56, 117–128]. The vp85.1 amino acid sequence has higherhomology to members of the 160 kDa flagellar-associated antigen family, FL-160, than to other 85 kDa expressedsialidase superfamily members. Southern blot analysis of virulent and attenuated lines demonstrated a complexhybridization pattern consistent with a multiple gene copy family that was identical in both lines. Antibody directedagainst recombinant vp85.1 peptide recognized proteins between 95 and 115 kDa in total virulent parasite lysateswhich were absent in attenuated lysates. Peptide N-glycosidase F treatment reduced the high molecular weight bandsto 85 kDa, indicating vp85 is an N-linked glycoprotein. Immunofluorescence with anti-vp85.1 demonstrated surfacelocalization of vp85.1 on virulent, but not attenuated, trypomastigotes. We postulate this new subfamily oftrans-sialidases may play a role in virulence. © 1999 Published by Elsevier Science B.V. All rights reserved.

Keywords: Trypanosoma cruzi; Virulence factors; Sialidase; SIRE; Surface antigens

Abbre6iations: aa, amino acids; Kb, kilobases; Kbp, kilobase pairs; kDa, kilo Daltons; ORF, open reading frame; UTR,untranslated region.

* Tel.: +1-206-543-0821; fax: +1-206-685-8681; e-mail: wesley@u.washington.edu.� Note : Nucleotide sequence data reported in this paper have been submitted to GenBank™ under accession numbers AF051695

and AF051696.

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

PII: S 0166 -6851 (98 )00152 -2

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116106

1. Introduction

Trypanosoma cruzi is a protozoan kinetoplastidparasite causing Chagas’ disease, a debilitatingand incurable disease affecting millions of peoplein Latin America [2,3]. Acute disease is marked byparasite replication and the infection is usuallynot detected until the chronic phase, when mor-bidity and mortality occur. An obligate intracellu-lar parasite in the mammalian host,trypomastigotes must survive the rigors of thebloodstream, attach to and enter host cells andescape the hostile vacuole environment into thecytoplasm. Upon entering the cytoplasm trypo-mastigotes differentiates into replicating amastig-otes, differentiate back to trypomastigotes andlyse the cell to escape and infect new host cells.The degree of success of the parasite in the mam-mal depends on its ability to invade the host andestablish a long-term infection. T. cruzi strainshave been isolated from human patients and in-sect vectors which demonstrate various degrees ofinvasiveness or virulence in the mammalian host.In these experiments using mouse models, therewas a direct correlation between parasitemia lev-els in the circulating blood and Chagas’ diseasepathology [4–6]. In addition, long-term serial pas-sage in axenic culture can lead to parasite strainswhich are attenuated relative to their parentallines [6,7]. However, little is known of the geneticfactors responsible for these observed differencesin virulence.

In an attempt to identify T. cruzi virulencefactors we have used subtraction hybridization toenrich for cDNA sequences expressed in a viru-lent parasite line but not an attenuated parasiteline. The attenuated T. cruzi line has been contin-uously passed in axenic culture for 6 years andhas lost the ability to generate measurable para-sitemia in mice yet can complete the life-cycle incell culture. A virulent phenotype was rederivedfrom this severely attenuated strain by two in vivopassages. The direct derivation of this virulent linefrom the parental attenuated line allows for sideby side genetic comparison by gene expressiondifference analysis. This subtraction library ap-proach identified a cDNA clone which was highlyexpressed in the virulent line, but weakly ex-

pressed in the attenuated line. This cDNA is amember of the sialidase/trans-sialidase superfam-ily [8], and is most similar to a previously de-scribed unexpressed sialidase-homologuepseudogene [1]. Evidence is presented suggestingthat this gene is a member of a distinct subfamilyof the sialidase/trans-sialidase superfamily. Ex-pression of this gene family by T. cruzi correlateswith a virulent phenotype.

2. Materials and methods

2.1. Parasites and animals

The T. cruzi CL strain [9] 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, Logan, UT). Epimastigoteswere grown in axenic cultures in liver-infusion/tryptone medium with 10% fetal calf serum. Forin vivo studies T. cruzi were passaged through6–10 week old, female C3H/HeJ mice by way ofintraperitoneal inoculation. Parasitemia levelswere determined by tail-bleeding infected miceand quantified as previously described [10]. Im-munizations for antibody production were per-formed in 6–10 week old outbred ICR mice.Clones of attenuated and virulent T. cruzi CLstrains were obtained by limiting dilution.

2.2. Library construction, RNA/DNA analysis

Poly(A)+ RNA was purified from attenuatedand virulent T. cruzi CL trypomastigotes usingoligo(dT) spin columns (Invitrogen, Carlsbad,CA). Using 2 mg poly(A)+ RNA from each para-site line, a subtracted library was constructedusing PCR-Select™ (Clonetech, Palo Alto, CA)according to the manufacturer’s recommenda-tions. The methodology used was based on sup-pression subtractive hybridization [11]. Insertsfrom individual subtracted clones were radioac-tively labeled for use as probes in northern andSouthern blot analysis by standard procedures

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116 107

[12]. All blots in this study were performed usingHybond-N (Amersham Life Sciences, ArlingtonHeights, IL) nylon and were subjected to a finalwash at 65°C in 0.1×SSC [1×SSC is 150 mMsodium chloride, 15 mM sodium citrate, pH 7.0].Total parasite RNA was harvested using Ultra-spec RNA (Biotecx Laboratories, Houston, TX).Rapid amplification of cDNA ends (RACE) wasused to clone two full length vp85 cDNA se-quences. For 3% RACE total virulent clone 3trypomastigote RNA was reverse transcribed tofirst strand cDNA by standard procedures [12]using an oligo(dT) anchor primer (5%-cctctgaag-gttcacggatccacatctaga(t)18vn-3%), where v=a, c org and n=any base. First round nested PCR wasperformed using sense primer S1 (5%-ggttttttactcgt-gaagggaact-3%) and antisense primer B1 (5%-cctct-gaaggttcacgga-3%). Second round PCR used senseprimer S2 (5%-catgtggtgaagcctgaaccca-3%) and anti-sense primer B2 (5%-cacggatccacatctaga-3%) to am-plify the 3% end. For 5% RACE total RNA wasprimed with AS5 (5%-ctccatgcgaagtggacttg-3%) andused as template for nested PCR. First roundPCR was performed using sense primer SL1 (5%-aacgctattattgatacagtt-3%) and antisense primerAS5. For second round PCR sense primer SL2(5%-atacagtttctgtactataatg-3%) and antisense primerAS6 (5%-ccacaatggatggatccc-3%) were used to clonethe 5% end of the gene. Primers SL1 and SL2 arecomplimentary to the spliced leader sequence. Forsequence analysis plasmid DNA was preparedusing column chromatography (Qiagen, Valencia,CA) and sequenced by Dye Terminator CycleSequencing kit (PE Applied Biosystems, FosterCity, CA).

2.3. Protein analysis, antibodies andimmunofluorescence

Oligonucleotide primers (Gibco BRL) wereconstructed for PCR amplification of a portion ofthe vp85.1 open reading frame for expression inE. coli (strain XL1-Blue). The sense primer, S12(5%-atggatccaacgaaggcgacgaaagacc-3%), defined theN-terminal end of the recombinant peptide atamino acid 188, while the antisense primer, AS7(5%-ttgatttcaaggacatcagccaccgt-3%), defined the C-terminus at amino acid 634. The PCR product

was digested with EcoRI and BamHI and clonedinto the respective sites of the prokaryotic expres-sion vector pGEX2T (Pharmacia Biotech, Pascat-away, NJ). Protein coding sequences cloned intopGEX2T are fused in-frame with Schistosomajaponicum glutathione-S-transferase. The vast ma-jority of the vp85.1 recombinant peptide was ex-pressed as insoluble inclusion bodies. High yieldof purified vp85.1 recombinant protein was ob-tained by isolating and purifying the inclusionsusing standard techniques [12]. To generatevp85.1-specific antiserum, ten ICR mice were ini-tially immunized with 100 mg of recombinantvp85.1 in Freund’s complete adjuvant (Difco Lab-oratories, Detroit, MI). Three boost immuniza-tions were delivered in 50 mg doses in Freund’sincomplete adjuvant at 3 week intervals. Pooledimmune serum was prepared three weeks after thefinal boost. Cross reacting antibodies were ad-sorbed by incubation of the serum with a soni-cated antigen preparation from XL1-Blueexpressing pGEX2T only. For total proteinpreparations, parasites were harvested andwashed before being boiled in 0.5% SDS and 1%b-mercaptoethanol. SDS-PAGE and western blotanalysis were performed by standard methods[12]. Removal of N-linked glycosylation groupsfrom parasite lysates was performed with peptideN-glycosidase F (PNGase F) (New England Bio-labs, Beverly, MA). The membrane used for west-ern analysis was Hybond ECL nitrocellulose(Amersham Life Sciences). Immunological detec-tion was carried out with alkaline phosphatase-la-beled secondary antibodies (Promega, Madison,WI). Virulent and attenuated trypomastigoteswere tested for surface immunofluorescence usingthe vp85.1 polyclonal antiserum. Anti-vp85.1 anti-serum was adsorbed with a freeze/thaw lysate of107 ml−1 attenuated epimastigotes (clone C11)prior to use in immunofluorescence. Virulent andattenuated trypomastigotes were washed threetimes in DMEM before being applied to poly-L-lysine (1 mg ml−1) coated microscope slides. Para-sites were fixed with 4% paraformaldehyde in PBSfor 30 min at room temperature. The parasiteswere incubated in a 1:200 dilution of anti-vp85.1serum for 1 h at room temperature. The parasiteswere then incubated in a 1:200 dilution of fluores-

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116108

cein-conjugated rat anti-mouse F(ab%)2 (AccurateBiochemicals, Westbury, NY) for 30 min at roomtemperature. Trypomastigotes were viewed byfluorescence microscopy (Carl Zeiss, Thornwood,NY). Controls run but not shown include (1)normal mouse serum as the primary antibody and(2) secondary antibody only. Both controls gavelittle or no background fluorescence.

3. Results and discussion

3.1. In 6i6o selection for 6irulent parasites

In this paper an attenuated T. cruzi CL line wasused which has been continuously passed as epi-mastigotes in axenic culture for 6 years and haslost the ability to generate microscopically de-tectable parasitemia in mice. Mice chronically in-fected with this attenuated line failed to exhibitany signs of chronic tissue pathology characteris-tic of T. cruzi infection. Although severely attenu-ated in vivo, the attenuated line epimastigotes cantransform to trypomastigotes in cell culture, repli-cate normally as amastigotes in macrophages andfibroblasts, and complete the lifecycle in vitro.The ability of the attenuated line to thrive in vitroyet struggle in vivo led us to test the reversibilityof this phenotype. A dose of 1×108 lab attenu-ated epimastigotes was inoculated into mice. Al-though no parasites were detected by microscopy,plasma was harvested from infected mice at day15 post-infection (normal height of parasitemia)and co-cultured with 3T3 fibroblasts. After :3weeks of growth in vitro, attenuated trypomastig-otes expanded to detectable numbers. A dose of5×106 trypomastigotes, passed once throughmice and expanded in vitro, was inoculated asecond time into mice. Unlike the first passage invivo, parasitemia was easily detected by mi-croscopy by day 6 post-infection and peaked byday 14. Trypomastigotes were harvested from theserum of the mice, expanded one time in vitro,stored as frozen stocks and named the rederivedvirulent line. We believe the above experimentselected a sub-population of parasites within theattenuated line which retained a virulent pheno-type because we were unable to restore virulence

from re-cloned attenuated parasites (see below).Certain virulence genes, which convey no advan-tage in vitro but convey an advantage in vivo,could have been silenced during long term axenicpassage of the attenuated line. The close geneticbackground shared between the attenuated lineand the rederived virulent line makes an excellentmodel to study virulence by gene expression dif-ference analysis.

3.2. Parasitemia experiments

Both attenuated and virulent lines, as well asclones derived from each line, were quantitativelytested for the ability to cause parasitemia in vivo.Mice were inoculated with a dose of 1×104 try-pomastigotes and parasitemia was measured dur-ing acute infection. The attenuated line andattenuated clone (clone C11) failed to demon-strate any parasitemia detectable by microscopy,whereas the virulent line and virulent clone (clone3) gave high parasitemia (Fig. 1A). Mice were alsoinoculated with a high dose of trypomastigotes(5×106) and, while the virulent line and clone 3gave high parasitemia, both the attenuated lineand clone C11 failed to generate any (not shown).In mice given the high dose of clone 3, 3 of the 5mice inoculated with 5×106 virulent clone 3 try-pomastigotes died at day 8 post-infection, whileone of the remaining two died at day 15 post-in-fection. In mice given the high dose of the virulentline, one died at day 15 post-infection (Fig. 1B).None of the mice inoculated with the attenuatedline or clone C11 died within 60 days after infec-tion. This indicates virulent parasites not onlygained the ability to cause high parasitemia butalso caused higher mortality rates than the attenu-ated parasites.

Three attenuated clones were tested for theability to convert to the virulent phenotype aftersubjecting them to multiple in vivo passages. Al-though each clone failed to produce parasitemiadetectable by microscopy, in vivo passed attenu-ated clones could be recovered and expanded bypassing the infected mouse serum into cell culture.Nevertheless, four in vivo passages of the attenu-ated clones failed to convert them to a virulentphenotype. This is in contrast to the attenuated

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116 109

line, where virulence was re-established by onlytwo in vivo passages. This suggests the virulentparasites formed a sub-population of the originalattenuated line that could be selected for by invivo passage.

Our hypothesis is that certain infection or viru-lence genes had been silenced or lost during long-term passage in axenic culture in the attenuatedparasites. Two passages of the attenuated linethrough mice selected for parasites within this linewith increased ability to invade the mammalianhost. Therefore, it is likely the rederived T. cruzi

Fig. 2. Northern analysis of attenuated and virulent trypo-mastigotes. Total RNA was isolated from attenuated line andvirulent line culture-derived trypomastigotes. Twenty five mi-crograms of total RNA was fractionated on a 1% formalde-hyde agarose gel, transferred to nylon membrane and probedwith labeled TCVF11 insert. Lane ‘A’ is attenuated trypo-mastigote RNA and lane ‘V’ is virulent trypomastigote RNA.The blot was stripped and re-probed with a calmodulin probe(CalA2) to confirm equivalent gel loading. Transcript sizes aremarked in kilobases.

Fig. 1. Measurement of parasitemia and mortality in C3H/HeJmice. A. The mean parasites per milliliter of blood measuredon various days after inoculating 5 mice with 104 trypomastig-otes intraperitoneally. Parasites used in this experiment werethe virulent line (), virulent clone c3 ("), the attenuatedline (�) and attenuated clone C11 (D). B. Mortality wasrecorded for each group of infected mice infected with 5×106

trypomastigotes. The group of mice infected with clone C11lacked mortality in the acute stage and are not included in thisgraph as the data superimposes the attenuated line data.

parasites express some infection genes or virulencegenes that the attenuated parasites do not express.

3.3. Cloning of 6p85, a distinct member of thesialidase superfamily

To begin analyzing the gene expression differ-ences between the two lines a subtracted librarywas constructed using suppression subtractive hy-bridization [11]. The resulting library should beenriched for cDNA clones expressed at higherlevels in virulent trypomastigotes. Twelve cDNAclones, ranging in size from 200 to 600 bp, weretested for virulent-specific expression by northernblot analysis. A total of 6 of the 12 clones hy-bridized to virulent trypomastigote RNA but notto attenuated trypomastigote RNA. A total of 4of the 12 cDNA clones hybridized to neither, and2 of the 12 clones hybridized to both. One clone,TCVF11, hybridized to two transcripts 3.5 and4.5 Kb in size in the RNA from virulent parasitesthat were absent in the RNA from attenuated

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116110

Fig

.3.N

ucle

otid

ese

quen

ceal

ignm

ent

ofth

eor

igin

alcD

NA

clon

eT

CV

F11

,gen

omic

clon

ec1

821

[1],

vp85

.1an

dF

L-1

60[1

4].T

hefu

llnu

cleo

tide

sequ

ence

ofT

CV

F11

issh

own

onth

efir

stlin

e,w

ith

the

rem

aini

ngse

quen

ces

alig

ned

belo

wit

.T

hedo

tsre

pres

ent

iden

tica

lba

ses,

dash

esre

pres

ent

gaps

inth

ese

quen

ce.

Hom

olog

ysc

ores

wit

hre

spec

tto

TC

VF

11ar

esh

own

inbo

ld.

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116 111

Fig. 4. Schematic diagram of the full length vp85.1 cDNA. The large boxed region represents the 789 aa ORF. Within the ORF thegray shaded area represents the region sharing sequence homology with other sialidase superfamily members, while the open areasrepresent regions lacking significant homology with other sialidases. The lines extending from the ORF represent the 5% UTR (238bp) and the 3% UTR (525 bp). The smaller boxed region above the vp85.1 schematic represents the original TCVF11 cDNA. Thevp85.1 sub-terminal sialidase motif (VTVxNVfLYNR) is shown. Below the vp85.1 schematic are the alignments of the twoSIRE-associated sequences SZ10 and SZ12; the dark portion of the lines represent sequence homology with vp85.1 while the dashedlines represents SIRE sequences. Also, below the vp85.1 schematic are the alignments of the genomic clones Tt21 and c1821 [1].

parasites (Fig. 2). However, the attenuated laneshowed a hybridizing high molecular weight bandmigrating at the compression zone of the gel. Thisband did not occur in every experiment and ispresumed to be an unprocessed polycistronicRNA or an artifact. Hybridization of these blotswith calmodulin (Fig. 2), tubulin, SA85 [13] orFL-160 [14] showed approximately equal hy-bridization for both, demonstrating the attenu-ated parasites had intact mRNA (not shown). Inaddition, the TCVF11 probe failed to hybridize toRNA from epimastigote stage parasites (notshown). Sequence analysis of the 450 bp TCVF11cDNA revealed sequence homology with the siali-dase/trans-sialidase superfamily [8]. Highest ho-mology (83%) was shared with c1821, a previouslydescribed unexpressed gp85-homologue pseudo-gene [1] (Fig. 3).

Rapid amplification of cDNA ends (RACE)was used to clone a full length cDNA. Flankingoligonucleotide primers complementary to theends of TCVF11 were designed and used to am-plify virulent clone 3 trypomastigote cDNA. Am-plification resulted in 1.1 Kbp 5% RACE and 2.3Kbp 3% RACE fragments which were subsequentlycloned into plasmids for sequencing. A total of 405% and 3% RACE products were tested for highstringency hybridization using an internalTCVF11 probe and 13 strongly hybridizing cloneswere chosen for full or partial sequencing. DNAsequence analysis revealed 5% and 3% RACE prod-ucts which were 93–100% identical amongstthemselves. Two full length cDNA clones were

constructed in this way and were designatedvp85.1 and vp85.2. Clone vp85.1 was 100% identi-cal within the overlap between the 5% end and the3% end fragments. The other, vp85.2, was 99.7%identical in the overlapping region. These datasuggest that vp85 is a multi-copy gene family (seebelow) with many members being simultaneouslyexpressed. Although there is no guarantee the twoends of vp85.1 came from the same gene, thefull-length vp85.1 cDNA is likely to be represen-tational (Fig. 4). Without the poly-adenylatedtails the vp85.1 and vp85.2 cDNAs are 3150 and3160 bp in length, respectively, and are 97.2%identical at the nucleic acid level. Both vp85.1 andvp85.2 share 87% nucleic acid sequence identitywith the original cDNA clone TCVF11 (Fig. 3).The ORFs of vp85.1 and vp85.2 are 790 and 789amino acids in length, respectively, and encode apredicted protein of about 85 kDa. The aminoacids encoded by the two open reading frames are95.4% identical and encode the conserved subter-minal sialidase motif VTVxNVfLYNR. For thepurpose of this report attention will focus onclone vp85.1.

Southern analysis of restriction enzyme digestedgenomic DNA from both attenuated and virulentparasites, probed with the TCVF11 (vp85) cDNA,revealed vp85 to be a member of a multigenefamily (Fig. 5). The TCVF11 probe hybridizedwith an 850 bp BamHI fragment apparently con-served among many of the vp85 sub-family mem-bers. This BamHI fragment starts 60 bp upstreamand ends 800 bp downstream of the vp85.1 trans-

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116112

lation start. Therefore, the Southern analysis cor-relates with the DNA sequence data for the vp85clones. There were no gross differences or restric-tion polymorphisms between the attenuated andvirulent clones to indicate any significant changein genomic organization. Several additional re-striction enzymes were tested, and in each case nodifferences were seen between virulent and attenu-ated TCVF11-hybridizing bands (not shown).Therefore, the loss of expression of the vp85 genesin the attenuated line does not appear to be dueto large deletions or genomic rearrangements.

The vp85.1 gene shares similarity with othermembers of the T. cruzi sialidase superfamily[13,15–17]. The highest overall similarity was togenomic clone c1821, an unexpressed sialidase-ho-

mologue pseudogene [1]. In their report Takle etal. could find no c1821-hybridizing RNA bynorthern analysis or PCR on reverse transcribedmRNA from their Y strain trypomastigotes [1].Interestingly, like the CL strain used in this re-port, the Y strain Takle et al. used was severelyattenuated in vivo (personal communication). Themost significant nucleotide homology, other thanwith c1821, is shared between all sialidase mem-bers in the 5% UTR, where identity is as high as90% over 196 bp between vp85.1 and the 160 kDaflagellum associated antigen, FL-160 [17]. Thesimilarities within the coding region range from37 to 49% identity at the amino acid level, thehighest score again being shared with FL-160.Multiple sequence alignment of the sialidase fam-ily members with vp85.1 revealed the coding re-gion homology to be restricted to the middle 450residues of the protein; from approximatelyamino acid 200 to the subterminal sialidase motif(Fig. 4). Outside of this ORF region vp85.1 sharesonly signal sequence and GPI anchor sequencehomology with the other sialidase superfamilymembers. Besides the 5% UTR and the central 450aa of the ORF, the similarity vp85.1 shares withFL-160 can also be found in the 3% UTR. Thevp85.1 3% UTR is 75% identical to the 3% UTR ofFL-160 over 476 bp but vp85.1 shares no homol-ogy with the 3% UTR of other sialidase superfam-ily members. This suggests the FL-160 and vp85gene sub-families share a closer evolutionary rela-tionship than with other members of the sialidase/trans-sialidase superfamily.

The vp85.1 gene also shares high sequence iden-tity with two short interspersed repetitive element(SIRE)-associated genomic sequences [18], SZ10and SZ12 (Fig. 4). The first 380 bp of genomicclone SZ10 shares 80% nucleic acid identity withinthe C-terminus of the vp85.1 coding region andextends to the first polypyrimidine tract 100 bpinto the 3% UTR (Fig. 6). Genomic clone SZ12also shares 80% nucleic acid identity over 700 bpwithin the C-terminus of the vp85.1 coding regionand extends through the 3% UTR to the lastpolypyrimidine tract, just prior to the polyadeny-lation site (Fig. 6). The high homology the SZ10and SZ12 genomic sequences share with vp85.1suggest they are members of the vp85 sub-family.

Fig. 5. Southern analysis of attenuated and virulent parasite’sgenomic DNA. Genomic DNA was restriction digested tocompletion and fractionated on a 0.8% agarose gel (EcoRI,BamHI) and a 1.2% agarose gel (SacII). Blots were probedwith the TCVF11 insert (1×106 CPM ml−1), washed at highstringency and exposed to autoradiography. Attenuated para-site (A) and virulent parasite (V) genomic DNA was com-pared. DNA fragment sizes are marked in kilobase pairs.

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116 113

Fig

.6.

Nuc

leot

ide

sequ

ence

alig

nmen

tof

vp85

.1w

ith

the

SIR

Eas

soci

ated

sequ

ence

s.T

hedo

tsre

pres

ent

perf

ect

mat

ches

,da

shes

repr

esen

tse

quen

cega

psan

dth

epo

siti

onof

the

term

inat

ion

codo

nis

unde

rlin

ed.

Whe

rese

quen

ceho

mol

ogy

wit

hvp

85.1

ends

and

the

SIR

Eel

emen

tsbe

gin

inSZ

10an

dSZ

12(G

enB

ank™

acce

ssio

nX

8920

0,X

8369

8,un

publ

ishe

dda

ta)

are

mar

ked

inbo

ld.

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116114

Interestingly, the last 400 bp of the SZ10 andSZ12 sequences contain SIRE sequences, suggest-ing that some members of the vp85 sialidasesub-family are linked to SIRE sequences. Se-quence analysis of an FL-160 genomic clone iden-tified a SIRE sequence 600 bp downstream of thepolyadenylation site, adding further evidence of aclose evolutionary relationship with the vp85 genefamily (not shown). The SIRE sequences arefound within intergenic regions and have beenfound to supply a functional 3% spliced leaderacceptor site for a downstream gene [18]. In addi-tion, the SIRE sequences may alter polyadenyla-tion site selection by interrupting the nativepolyadenylation site and providing its own func-tional polyadenylation signal [18]. As more T.cruzi genes are sequenced more examples ofSIRE-linked gene sequences are demonstrated, in-cluding histone 2A [19], hexose transporter [20],flagellar antigen FL-160 [14], a 24S-a ribosomalpseudogene [21] and the two vp85 sub-familymembers, SZ10 and SZ12, mentioned above.

Transcription of the vp85 genes results in atleast two transcripts :3.5 and 4.5 Kb in sizespecific to the trypomastigote stage. It is not yetknown how the two transcripts differ from eachother. The size of the vp85.1 and vp85.2 cDNAswould suggest they represent the 3.5 Kb tran-script. With the 5% UTR of vp85 so highly con-served with other sialidase superfamily membersone could speculate the difference in transcriptsizes resides within the 3% UTR. Whether this isdue to differing SIRE sequence insertions and/orthe use of alternative polyadenylation sites re-mains to be investigated. Transcripts for the re-lated genes FL-160 and SA85 were found in theattenuated and virulent trypomastigotes (notshown), but transcripts for vp85 were found invirulent trypomastigotes only. Since multiplegenes for vp85 are found in both attenuated andvirulent parasites, this suggests that defective tran-scriptional or post-transcriptional control led toloss of vp85 expression in attenuated parasites.These defects appear specific to the vp85 familybecause there was normal mRNA content forFL-160, SA85 and calmodulin in attenuatedtrypomastigotes.

Fig. 7. Western blot analysis of the vp85 proteins. Totalprotein lysate from virulent clone c3 trypomastigotes wassubjected to western analysis before (lane 1) and after (lane 2)PNGase F treatment. Samples were fractionated by 10% SDS-PAGE, transferred to nitrocellulose and probed with anti-vp85.1 antibody (1:1000). Protein concentrations weredetermined by the Bradford assay and equal lane loadingconfirmed by Ponceau S staining of blots prior to immunos-taining.

3.4. Analysis of the 6p85.1 expressed product

A partial fragment of vp85.1 was over-ex-pressed in E. coli. Clone pGEX85.1 expressedamino acids 188–634 of the vp85.1 protein fusedto the C-terminus of Schistosoma mansoni glu-tathione S-transferase. The vp85.1 fusion proteinwas purified and used to immunize mice. Fourimmunizations yielded pooled mouse serum witha high antibody titer against the recombinantpGEX85.1 peptide (not shown). Denaturing SDS-PAGE and western blot analysis were undertakento characterize the vp85.1 gene product. The poly-clonal antiserum recognized multiple bands mi-grating between 95 and 115 kDa and a 71 kDaband in the virulent trypomastigote lysate (Fig. 7,lane 1). These bands were completely absent inpooled attenuated clone lysates (not shown). Themass of the reactive bands was greater than the 85kDa predicted from the amino acid sequence sug-gesting vp85 proteins are post-translationallymodified. The identity of the 71 kDa band is notknown but could represent a degradation productor a cross-reactive molecule. No immunologicallycross-reactive bands were detected in attenuatedor virulent epimastigote lysates (not shown).

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116 115

Other members of the sialidase superfamily areknown to have N-linked glycosylation moieties.Therefore, virulent clone 3 trypomastigote lysatewas treated with peptide-N-glycosidase F (PN-Gase F) to remove all N-linked carbohydrategroups. PNGase F treatment reduced the 95–115kDa reactive bands to a single band of the pre-dicted molecular weight of 85 kDa (Fig. 7, lane 2).This suggests there is heterogeneous glycosylationamong the native vp85 proteins. It is not knownwhy the 71 kDa band (Fig. 7, lane 1) disappearedupon treatment with PNGase F (Fig. 7, lane 2).Several lines of evidence suggest vp85 proteins aresurface antigens. First, vp85 shares sequence ho-mology with sialidase superfamily members whichare known T. cruzi surface antigens. Second, thelast twenty amino acids of vp85.1 encode theconserved GPI anchor sequence shared with otherT. cruzi surface antigens [8]. Finally, vp85proteins are N-linked glycosylated glycoproteins.Therefore, virulent and attenuated trypomastig-otes were tested for surface immunofluorescenceusing the vp85.1 polyclonal antiserum. Virulentclone 3 trypomastigotes showed positive stainingover the entire surface, consistent with surfacelocalized molecules (Fig. 8B). As expected fromwestern analysis, attenuated clone C11 trypo-mastigotes failed to stain with the vp85.1 antibod-ies (Fig. 8F). As a side note, the virulent clone 3population consisted of approximately 10% slen-der forms and 90% stumpy forms, previouslydescribed in the literature [22]. The long slenderforms demonstrated little if any vp85.1 immunos-taining while the short and stumpy forms stainedpositively (Fig. 8D). These results suggest vp85proteins are a distinct sub-family of sialidase/trans-sialidase homologues associated with thesurface of virulent trypomastigotes but not atten-uated trypomastigotes.

It is not yet known if the vp85 proteins havesialidase/trans-sialidase activity. Although vp85.1and vp85.2 contain the sub-terminal VTVxN-VfLYNR motif, both lack the N-terminalSXDXGXTW box motifs found in bacterial andtrypanosomal sialidases [8], suggesting they arenot active sialidases. We compared virulent trypo-mastigote and attenuated trypomastigote lysatesfor sialidase activity and both had equivalentlevels of activity (not shown).

Thus we have shown that vp85.1 is expressed asmRNA and the corresponding protein is found onthe surface of virulent trypomastigotes but notattenuated trypomastigotes. The fact that manygenomic restriction fragments hybridized to theTCVF11 probe and that RACE cloning yieldedmany different but highly related vp85 cDNAsdemonstrates that a large family of vp85 genes isexpressed in virulent but not attenuated trypo-mastigotes. Since the vp85 genes are identical inSouthern blots of DNA from virulent and attenu-ated parasites, it seems likely that there is aregulatory abnormality that leads to the absenceof vp85 mRNA in attenuated trypomastigotes.Furthermore, these observations suggest the vp85

Fig. 8. Surface immunofluorescence of virulent and attenuatedtrypomastigotes. Virulent clone 3 trypomastigotes (A–D) andattenuated clone C11 trypomastigotes (E–F) were tested andcompared for surface immunofluorescence with the vp85.1polyclonal antiserum. Immunofluorescence images are shownin panels B, D and F while the corresponding phase-contrastimages are shown in panels A, C and E. Magnification:(630× ). Normal mouse serum and secondary antibody onlycontrols failed to show any background fluorescence (notshown).

D. Weston et al. / Molecular and Biochemical Parasitology 98 (1999) 105–116116

genes are coordinately regulated in a manner in-dependent from other surface proteins in the siali-dase superfamily such as SA85 and FL-160, whichhad equal levels of mRNA in virulent and attenu-ated trypomastigotes. Loss of vp85 expression inthe attenuated parasites could be at the transcrip-tional level, at the level of mRNA processing, orat the level of mRNA stability. Further experi-ments are required to understand how the entirevp85 gene family is specifically silenced in theattenuated line. Finally, the role vp85 proteinsplay in virulence is unclear. However, expressionof this gene family in virulent, but not attenuated,trypomastigotes is suggestive that vp85 has a rolein virulence.

Acknowledgements

The authors would like to thank Lynn Barrettand Manami Nishi for their technical assistance,Dr. Caroline Cameron for editorial assistance andDr. Stuart Kahn and Dr. John Swindle for helpfuldiscussions. Special thanks go to Dr. FrederickBuckner for his many helpful discussions and forgenerating the CL clone 3. This work was sup-ported by NIAID grants R21 AI38924 and F32AI09235.

References

[1] Takle GB, O’Conner J, Young AJ, Cross GAM. Sequencehomology and absence of mRNA defines a possible pseu-dogene member of the Trypanosoma cruzi gp85/sialidasemultigene family. Mol Biochem Parasitol 1992;56:117–28.

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

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

[4] Lauria-Peres L, Teixeira AR. Virulence and pathogenicityassociated with diversity of Trypanosoma cruzi stocks andclones derived from Chagas’ disease patients. Am J TropMed Hyg 1996;55:304–10.

[5] Rowland E, Luo H, McCormick T. Infection characteris-tics of an Ecuadorian Trypanosoma cruzi strain withreduced virulence. J Parasitol 1995;81:123–6.

[6] Contreras VT, Araque W, Delgado VS. Trypanosomacruzi : metacyclogenesis in vitro I: changes in the proper-ties of metacyclic trypomastigotes maintained in the labo-

ratory by different methods. Mem Inst Oswaldo Cruz1994;89:253–9.

[7] Alves AMB, De Almeida DF, Von Kruger WMA.Changes in Trypanosoma cruzi kinetoplast DNA minicir-cles induced by environmental conditions and subcloning.J Euk Microbiol 1994;41:415–9.

[8] Cross GAM, Takle GB. The surface trans-sialidase familyof Trypanosoma cruzi. Ann Rev Microbiol 1993;47:385–411.

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

[10] Buckner FS, Wilson AJ, Van Voorhis WC. Trypanosomacruzi : use of a herpes simplex virus-thymidine kinase as anegative selectable marker. Exp Parasitol 1997;86:171–80.

[11] Diatchenko L, Lau YC, Campbell AP. Suppression sub-tractive hybridization: a method for generating differen-tially regulated or tissue-specific cDNA probes andlibraries. Proc Natl Acad Sci USA 1996;93:6025–30.

[12] Sambrook J, Fritsch E, Maniatis T. In: MolecularCloning: A Laboratory Manual. Cold Spring Harbor,NY: Cold Spring Harbor Laboratory Press, 1989.

[13] 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.

[14] 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.

[15] Fouts DL, Ruef BJ, Ridley PT, Wrightsman RA, PetersonDS, Manning JE. Nucleotide sequence and transcriptionof a trypomastigote surface antigen gene of Trypanosomacruzi. Mol Biochem Parasitol 1991;46:189–200.

[16] 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.

[17] 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.

[18] Vasquez MP, Schijman AG, Levin MJ. A short inter-spersed repetitive element provides a new 3% acceptor sitefor trans-splicing in certain ribosomal P2b genes of Try-panosoma cruzi. Mol Biochem Parasitol 1994;64:327–36.

[19] Puerta C, Martin J, Alonso C, Lopez MC. Isolation andcharacterization of a gene encoding histone H2A fromTrypanosoma cruzi. Mol Biochem Parasitol 1994;64:1–10.

[20] Tetaud E, Bringand F, Chabas S, Barret M, Baltz T.Characterization of glucose transport and cloning of ahexose transporter gene in Trypanosoma cruzi. Proc NatlAcad Sci USA 1994;91:8278–82.

[21] Vieira de Arruda M, Reinach FC, Colli W, Zingales B.Sequence of the 24S-a ribosomal RNA gene and charac-terization of a corresponding pseudogene from Try-panosoma cruzi. Mol Biochem Parasitol 1990;40:35–42.

[22] Brener Z. The behavior of slender and stout forms ofTrypanosoma cruzi in the bloodstream of normal andimmune mice. Ann Trop Med Parasitol 1969;63:215–20.