Microbes and Infection 12 (2010) 19e27www.elsevier.com/locate/micinf

Original article

Silencing of tachyzoite enolase 2 alters nuclear targeting of bradyzoiteenolase 1 in Toxoplasma gondii

Michael Holmes a, Urszula Liwak a, Ionut Pricop a, Xiang Wang a, Stanislas Tomavo b,Sirinart Ananvoranich a,*

a Department of Chemistry and Biochemistry, University of Windsor, 401 Sunset Avenue, Windsor, ON, N9B 3P4, Canadab CNRS UMR 8161, Institut Pasteur Lille, Batiment Guerin, France

Received 12 June 2009; accepted 5 September 2009

Available online 19 September 2009

Abstract

In Toxoplasma gondii, an intracellular parasite of the phylum Apicomplexa, two isoforms of enolase (ENO1 and ENO2) are expressed instage-specific manner. ENO2 is expressed only in rapidly growing tachyzoites, while ENO1 is in slowly growing bradyzoites. Interestingly, thelocalization of ENO1 and ENO2 in the nuclear compartment has suggested possible roles of the proteins in gene regulation and/or cell cycle. Tounderstand the physiological role of ENO2 in T. gondii, the expression of ENO2 was silenced using a homologous gene silencing procedure. Theintroduction or expression of ENO2 dsRNA successfully silenced the expression of ENO2 at the levels of transcripts and proteins. While therewas no change in the growth rate of both tachyzoites and bradyzoites, a subtle phenotypic change was observed in the localization of the ENO1gene product in the bradyzoite stage.� 2009 Elsevier Masson SAS. All rights reserved.

Keywords: Toxoplasma; Enolase; Gene silencing; Differentiation; Localization

1. Introduction

Toxoplasma gondii is an intracellular protozoan parasitethat currently infects over one third of the world’s population.Although usually asymptomatic in healthy individuals, as anopportunistic pathogen, this apicomplexan may cause severedisease and death in those with compromised immunesystems, such as AIDS patients. T. gondii is present in thehuman host in two asexual forms e the rapidly growingtachyzoite and the slowly growing encysted form known as thebradyzoite stage [1]. During initial infection, the parasitespreads quickly through the benefit of the rapidly growingtachyzoite stage. However, upon an immune response, tachy-zoites convert into bradyzoites and may lie dormant until the

* Corresponding author. Tel.: þ1 519 253 3000x3550; fax: þ1 519 973

7098.

E-mail address: anans@uwindsor.ca (S. Ananvoranich).

1286-4579/$ - see front matter � 2009 Elsevier Masson SAS. All rights reserved.

doi:10.1016/j.micinf.2009.09.010

host’s immune system weakens sufficiently. The re-emergenceof tachyzoites from encysted bradyzoites that may occur manyyears after initial infection is the cause of damage in thosewhose immune systems are depleted [1]. The effectors of thisinterconversion are currently unknown and the process ispoorly characterized. Therefore, it is of prime importance todetermine the identity of any of these effectors and how theyfunction. In order to understand this process, researchers havepaid special interest to developmentally regulated genes, asthese seem to be likely effectors of the interconversionprocess. Glycolytic enzymes in particular have proven to be ofinterest for several reasons, first of which is that most areencoded by at least two isoforms that are differentiallyexpressed in tachyzoites and bradyzoites [2,3]. These enzymeshave also attracted attention because of the difference inmetabolism between the two asexual stages. Slow growingbradyzoites lack functional mitochondria and therefore arelimited to glycolysis for their energy production whereastachyzoites appear to have the ability to perform fully

20 M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

functional cellular respiration. Due to these functional limi-tations, some of the glycolytic enzymes of T. gondii havedifferent biochemical properties [3].

One glycolytic enzyme that has proven to be of particularinterest as a possible effector of stage interconversion isenolase. T. gondii encodes two isoforms of enolase (ENO1 andENO2) that are expressed exclusively in bradyzoites andtachyzoites, respectively [3e5]. These two genes share highsequence similarity at both the nucleotide and protein levelsand also pose interest as potential drug targets as both ENO1and ENO2 share a similar pentapeptide insertion that separatesthem from mammalian enolases [5]. Although sharing similaraffinities for their substrate, the tachyzoite specific ENO2appears to be able to convert 2-phosphoglycerate to phos-phoenolpyruvate with a 3-fold greater reaction speed thanENO1. However, the bradyzoite specific ENO1 appears tohave a higher denaturation temperature that may help it adjustto increased stress conditions [3]. Most importantly, localiza-tion studies have determined that in T. gondii, both ENO1 andENO2 are found primarily in the nucleus [6]. Enolases havebeen implicated as regulators of gene transcription in a widevariety of organisms (reviewed in [7]) and therefore it is ofinterest to verify any potential role ENO1 and/or ENO2 mayhave as effectors of interconversion between the tachyzoiteand bradyzoite forms. Because ENO1 and ENO2 are locatedwithin 1.2 kb of each other on chromosome VIII [8,9], such aninvestigation could not be performed using a classical genereplacement methodology. Here we employed a homologousgene silencing technique for the study.

2. Materials and methods

2.1. Parasite & culture conditions

All experiments were carried out using human foreskinfibroblasts (HFF, obtained from Dr. D. Roos, University ofPennsylvania) that were grown in DMEM media, supple-mented with 10% calf serum (Hyclone, Logan, UT), 5 mg/mlstreptomycin and 5 U/ml penicillin in a 5% CO2 atmosphere.Parasite strain RHDHX was maintained using an HFF mono-layer grown in MEM media supplemented with 1% dialyzedfetal bovine serum (Invitrogen Canada Inc., ON). Clonalparasites were selected and maintained in culture media con-taining 25 mg/ml mycophenolic acid and 50 mg/ml xanthine.To differentiate the tachyzoites into bradyzoites, parasiteswere allowed to invade HFF monolayers for 4 h. The mediawas subsequently replaced with RPMI media containing50 mM HEPES pH 8.2 and 5% dialyzed fetal bovine serumand grown in ambient CO2. The medium was replaced every 2days to maintain the alkaline pH [10].

2.2. Synthesis of dsRNA by in vitro transcription

The T7 RNA polymerase promoter (17 nt long) wasincorporated into the 50-end of the DNA templates by PCRusing T7 promoter containing oligonucleotides specific to thetargeted genes (Table 1). Plasmids, pCATGFP (gift from Dr.

D. Roos, University of Pennsylvania), pENO1, and pENO2 [5]were used as templates to produce T7 RNA polymerasepromoter-bearing DNA templates, which were then tran-scribed in vitro into their corresponding dsRNAs in in vitrotranscription reaction mixtures [11,12]. The dsRNAs homolo-gous to ENO1 (543 bps, Genbank accession no. AF051910),ENO2 (543 bps, Genbank accession no. AF123457), and GFP(714 nt) were all synthesized according to this method.

2.3. Construction of double-stranded RNA expressionplasmids

pTUB8mycHISGFP-HX, a Toxoplasma expressionplasmid, (obtained from Dr. D. Soldati, University of Geneva)was used as the parental plasmid for all constructs used in thisstudy. To create the dsRNA expression plasmids used in thisstudy, the modified tubulin promoter, referred to as TUB8 or5RT70, (590 bps) [13] was PCR amplified from and clonedinto pTUB8mycHISGFP-HX at PstI and BamHI site down-stream from the coding sequence of GFP. The resultantplasmid, named p(TUB8)2GFP-HX, carries two modifiedtubulin promoters arranged in a head-to-head fashion flankingthe GFP coding sequence. To construct p(TUB8)2ENO1-HXand p(TUB8)2ENO2-HX, the GFP coding sequence ofp(TUB8)2GFP-HX was removed by NsiI and PstI restrictiondigestion. Complementary DNA fragments corresponding tothe first 540 bps of the ORF of ENO1 and ENO2 wereamplified using the corresponding PCR primers listed in Table1. The DNA fragments were then individually cloned into theplasmid carrying two modified tubulin promoters that was pre-digested with NsiI and PstI. The resultant plasmids wereverified by restriction endonuclease analyses.

2.4. Generation of stably transgenic parasite lines

dsRNA expression plasmids, p(TUB8)2GFP-HX,p(TUB8)2ENO1-HX and p(TUB8)2ENO2-HX (Fig. 2A), wereindividually used in the transfection of RHDHX parasites byelectroporation [14]. Stable populations and clonal parasiteswere selected and maintained in culture media containing25 mg/ml mycophenolic acid and 50 mg/ml xanthine. Theparasites were named GFPkd, ENO1kd and ENO2kd corre-sponding to the gene that was targeted for silencing orknockdown of the gene expression (kd¼ knockdown).

2.5. Reverse transcription and polymerase chainreaction (RT-PCR)

Reverse transcription (RT) reactions were carried out aspreviously described [15]. The sequences of primers are listedin Table 1. For the detection of antisense RNA, an oligonu-cleotide specific to the sense strand of each gene target wasused for the cDNA synthesis. In semi-quantitative RT-PCRanalyses of the steady state mRNA levels, an oligo dT(dT36A/G/C, Bio Basic Inc.) was used as the RT primer. Equalamounts of cDNA products were subsequently used for PCRwith gene specific primers. The PCR was performed using Taq

Table 1

Oligonucleotides used in the study. Sequences corresponding to the promoter of T7 RNA polymerase are bold, while the recognition sites of restriction endo-

nucleases underlined.

Oligonucleotides Sequences

For the synthesis of the dsRNA DNA templates

pT7on5’GFP TAATACGACTCACTATAGGTGAGCAAGGGCGAGGAG

pT7on3’GFP TAATACGACTCACTATAGGTACAGCTCGTCCATGCC

T7on5’ENO1 TAATACGACTCACTATAGGATGGTGGTTATCAAGG

T7on3’ENO1 TAATACGACTCACTATAGGCCAACGGGAGCGATC

T7on5’ENO2 TAATACGACTCACTATAGGATGGTGGCCATCAAGG

T7on3’ENO2 TAATACGACTCACTATAGGCCGACGGGGGCGATC

For the construction of dsRNA expression plasmids

PstI5’ENO1 CAGCTGCAGATGGTGGTTATCAAGGAC

PstI3’ENO1 CAGCTGCAGCCAACGGGAGCGATC

PstI5’ENO2 CAGCTGCAGATGGTGGCCATCAAGG

PstI3’ENO2 CAGCTGCAGCCGACGGGGGCGATC

For RT-PCR analysis

ENO1_sense ATGGTGGTTATCAAGG

ENO1_antisense CCAACGGGAGCGATC

ENO1_sense ATGGTGGCCATCAAGG

ENO1_antisense CCGACGGGGGCGATC

ENOx ATGGTGG(C/T)(C/T)ATCAAGG

ROP_ sense GGAACATGGGCCACAGG

ROP_ antisense CGCCGAAAGCGTCTCTG

21M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

DNA polymerase. To rule out the presence of genomic DNAcontamination in the extracted and treated RNA samples,control PCRs were conducted by directly subjecting RNApreparations to PCR.

2.6. Enolase enzymatic assays

Freshly released parasites were harvested and lyzed ina buffer containing 100 mM Tris HCl, pH 7.5, and 10 mMPMSF. Ten mg of lysates were the incubated in a mixturecontaining 50 mM imidazole, 1.1 mM MgCl2, 0.1 mM EDTA,250 mM KCl, pH 7.2 with 0.2 mM to 2 mM of 2-phospho-d-glycerate (2-PGA sigma) as substrate for 2.5 min at 25 �C.Phosphoenolpyruvic acid formation was detected bymeasuring the absorbance at 240 nm [5].

2.7. Immunolocalization assays

Cultures containing tachyzoites were then allowed to growfor 2e3 days while intracellular bradyzoites were cultured for5 days before analysis. All experiments were set up in dupli-cate and repeated at least twice. Slides were fixed and pro-cessed as previously described [15]. Slides were thenincubated for 1 h with a primary antibody either the rabbitENO1 or ENO2 antiserum (diluted 1:500) or the mouse anti-SAG4 (T84A121C3, 1:100), and subsequently with a corre-sponding secondary antibody labeled with Rhodamine orFITC (1:600). For cyst staining, Dolichos biflorus agglutininconjugated to FITC (diluted 1:300, Sigma) was used. Stainingof the nuclei was carried out by incubation in the presence ofHoechst (Sigma). Slides were mounted and examined witha Leica DMIRB microscope. Images were taken with a cooledQ-Imaging CCD camera using the Northern Eclipse software.

In addition, to visualize the localization of ENO1 signals,slides were visualized using an Olympus IX81 invertedmicroscope outfitted with the FluoView 1000 Confocalimaging system.

2.8. Experimental infections in mice

Purified tachyzoites of each tested parasite strain wereinoculated into groups of 5 female 8e9 week-old BALB/cmice at 103 or 104 parasites per mouse and monitored untildeath or survival for 2 weeks.

3. Results

3.1. Enolase 2 expression can specifically be down-regulated by homologous double-stranded RNA

To evaluate the efficiency and specificity of homologousdsRNA induced silencing of ENO2, transient down-regulationassays were performed using a previously described procedure[12]. In vitro synthesized dsRNAs (5 mg) homologous to GFP,ENO1 and ENO2 were electroporated into the parasite strainRH. GFP gene was used as a control to evaluate the effect ofexogenous and unrelated dsRNA in T. gondii. It should benoted that 69� 10% of electroporated parasites were able touptake dsRNA, when fluorescein-labeled dsRNAs were used(Al-anouti and Ananvoranich, unpublished data). For theassessment of ENO2 expression, which is constitutivelyexpressed in tachyzoites, total RNAs were isolated from newlyreleased parasites at 3e5 days post-electroporation. Usingequal amounts (1 mg) of cDNA, semi-quantitative RT-PCRanalyses were performed with primers specific to either ENO2,rhoptry protein 1 (ROP1), or Argonaute (AGO). ROP1,

22 M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

a constitutively expressed gene, was used as internal control toensure the RNA quality and to allow general assessment ofnon-target genes. AGO, a putative element responsible inhomologous gene silencing machinery [16], was used as anindicator that the gene silencing operation was not impaired intested samples. A lowered level of ENO2 was clearly observedwhen the homologous ENO2 dsRNA was used (Fig. 1A, upperpanel). Despite the 65.8% nucleotide sequence identitybetween the ORFs of ENO1 and ENO2, the presence of ENO1dsRNA did not affect the level of ENO2 expression nor did theGFP dsRNA (Fig. 1A). Equivalent levels of ROP1 and AGORT-PCR products were observed in all tested samples(Fig. 1A, middle and lower panels), indicating that the pres-ence of ENO1, ENO2, and GFP dsRNAs did not disturb thesteady state levels of ROP1 and AGO. The intensities of theENO2 RT-PCR products were then quantified against internalcontrol genes (ROP1 and AGO) and against those of mocktransfections (Fig. 1A). Electroporation with ENO2 dsRNAlowered the level of ENO2 mRNA to approximately 10% of itsnormal level, as compared to mock, ENO1, and GFP dsRNAelectroporation. To determine whether the transient geneknockdown affected the levels of gene products, western blotanalysis was performed using polyclonal antibodies raisedagainst ENO2 and LDH1. Initially we speculated that thealteration of ENO2 expression might influence the expressionof LDH1, as the latter is an enzyme involved downstream ofenolase in the glycolytic pathway. However, LDH1 proteinlevels remained unchanged among samples tested, whileENO2 was decreased following the ENO2 dsRNA electro-poration (Fig. 1B). As LDH1 protein levels were constant inall samples, the level of ENO2 was normalized against that ofLDH1 for the purpose of comparison. The ENO2 protein waslowered only to approximately 40%, as compared to levelsproduced by mock and the electroporation with dsRNA nothomologous (ENO1) or unrelated (GFP) to ENO2 (Fig. 1B).

3.2. Transgenic and stable ENO2-knockdown parasites

To understand their physiological functions, we createdtransgenic parasite lines that stably express the dsRNAhomologous to the ORFs of ENO1, ENO2, or GFP, referred toas ENO1kd, ENO2kd and GFPkd, respectively, (Fig. 2A). Thetransgenic GFPkd strain served as a control for the study. Thetransgenic parasites were selected and monitored for theirexpression of the dsRNA homologous to ENO1 and ENO2 byRT-PCR of the antisense RNA strand (Fig. 2B, asRNA). ENO1and ENO2 antisense transcripts from the transgenic ENO1kdand ENO2kd strains were detected and confirmed by SacI andHindIII digestion (Supplementary Fig. S1A). Since there wereno antisense transcripts detected in the RHDHX or GFPkdstrains, this confirmed that no natural antisense transcriptscorresponding to enolases are naturally produced. To ensurethat the transgenic GFPkd parasite harboured the plasmidconstruct and produced dsGFP, we carried out the RT-PCRusing the primer set specific to GFP and showed that the GFPantisense transcripts were produced (SupplementaryFig. S1B).

We anticipated that the expressed dsRNA would invokehomologous gene silencing and cause a decrease in the steadystate level of the target mRNA. Consequently we monitoredthe steady state level of ENO2 mRNA in comparison to thelevel of ROP1 mRNA. In the transgenic ENO2kd strain, thesteady state level of ENO2 mRNA was approximately 88%lower than those of the parental and control strains, indicatingthat the presence of ENO2 dsRNA caused homologous mRNAdegradation (Fig. 2B, middle panel). Interestingly, an attenu-ated decrease in ENO2 mRNA was observed in the ENO1kdstrain (w35% lower than those of the parental strains), whilethe transgenic GFPkd control strain, the expression of GFPdsRNA did not cause any silencing effect on the ENO2 geneexpression. It thus indicated that the presence of unrelatedGFP dsRNA did not cause off-target gene silencing. Unex-pectedly, an elevated level of ENO2 mRNA was detected in thetransgenic GFPkd strain, suggesting that the over-expressionof GFP dsRNA could interfere with the overall expressionprofile of the parasite and/or cause cellular stresses resulting inthe elevated ENO2 mRNA.

To further evaluate the extent of the homologous genesilencing, their cell-free extracts were analyzed for the level ofENO2 (Fig. 2C). Densitometric measurements of ENO2protein were then obtained and normalized against LDH1 forcomparison. The parental (RHDHX) and control (GFPkd)strains expressed relatively similar ENO2/LDH1 ratios indi-cating that although ENO2 mRNA appears to be up-regulatedby the introduction of unrelated dsRNA (herein GFP dsRNA),ENO2 expression appeared to be unaffected at the proteinlevel. The transgenic ENO2kd strain showed the lowest levelof ENO2 at w35% of parental, indicating a homologous genesilencing in the presence of the expressed ENO2 dsRNA. Inthe transgenic ENO1kd strain, the attenuated ENO2 mRNAlevels corresponded to a moderate ENO2 down-regulation(ENO2 55% of parental). However, the lowered level of ENO2protein in either ENO1kd or ENO2kd strains was not trans-lated into a decrease in overall ENO2 enzymatic activity. Allparasites tested exhibited w14� 0.9 mM min�1 mg�1.

3.3. Effect of ENO2 on the cell biology of the parasite

In order to determine how the decreased level of ENO2expression affected the parasite, the growth of the transgenicparasites in comparison to the parental strains were examinedusing infected monolayers grown under either tachyzoite orbradyzoite culture conditions. Under tachyzoite cultureconditions, all clonal strains tested multiplied in similarpatterns. At 24 h post-infection, the majority (w70%) of thevacuoles contained 8 parasites; at 32 h post-infection, themajority (w50%) of the vacuoles contained 16 parasites(Fig. 3A&B). When the parental and transgenic parasites weregrown under alkali conditions to induce differentiation, w15%of the population converted to bradyzoites, determined by thepresence of cyst structures stained by Dolichos biflorusagglutinin conjugated to FITC. There was no significantdifference detected among these strains. In order to betterunderstand how the transgenic ENO2kd parasites propagated

Fig. 1. Transient gene silencing by dsRNAs in the tachyzoites. (A) Semi-quantitative RT-PCR analyses were performed using an equal amount of RNA isolated

from tachyzoites that were electroporated with dsRNA homologous to ENO1, ENO2, or GFP genes. A representative gel image is shown, and the sizes of the

expected PCR products are indicated. Normalized ENO2 expression was calculated and obtained from three independent experiments from which the intensities of

the RT-PCR products (ENO2/ROP1 or ENO2/AGO) were quantified. (B) Western blot analysis was performed to evaluate the extent of the ENO2 gene silencing.

The presence of ENO2 and LDH1 were revealed using an anti-ENO2 and an anti-LDH1 antibody. Normalized values of ENO2/LDH1 proteins were calculated and

obtained following the densitometric measurement of revealed ENO2 and LDH1 bands from three independent analyses.

23M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

and affected the hosts in vivo, the parental (RHDHX) and twoindependent B9 and WA clones of the transgenic ENO2kdwere tested in a murine model. As shown in Fig. 3C, nosurvival was found in all mice infected with either transgenicor and parental parasite strains at 14 days post-infection,indicating that the virulence of the transgenic parasites was notaffected by the lowered level of ENO2 expression. Therefore,there was no significant difference in the growth and virulenceof transgenic parasites, both in vivo and in vitro, in comparisonto that of the parental parasite. We speculated that the effectof ENO2 knockdown was very subtle and did not allow sucha phenotypic display.

In order to better understand the effect of the ENO2knockdown in the parasites’ cell biology, clonal ENO2kdstrains were grown under alkaline conditions to induce in vitrodifferentiation. While the localization of ENO2 in the trans-genic ENO2kd and parental tachyzoites is mainly in thenucleus (Fig. 4A); the localization of ENO1 in the transgenicENO2kd bradyzoites was erratic in comparison to the parentalRHDHX bradyzoites. Among positively a-ENO1 stainedvacuoles, 83� 2% of vacuoles harboured ENO2kd brady-zoites with ENO1 mainly localized in the nucleus, similar tothat of the parental bradyzoites (Fig. 4B). The subpopulation(w17%) of these vacuoles carried bradyzoites with reducednuclear localization (Fig. 4B). Notably, these vacuoles with

aberrant ENO1 localization were prominently stained by theDolichos biflorus agglutinin conjugated to FITC, indicatingencystations. In addition, these transgenic ENO2kd brady-zoites were positively stained by antibody raised againsta bradyzoite-specific antigen (SAG4), a marker of in vitro cystformation. It was noted that within these encysted vacuoles,the erratically reduced nuclear localization was not uniform(Fig. 4C), suggesting that the reduction of ENO1 nuclearlocalization might correlate with the bradyzoite’s growth andmaturity in the encysted vacuoles. The aberrant localizationwas not observed in the transgenic GFPkd and parental strains.We therefore speculated that ENO2 expression might beimportant for the correct localization of ENO1.

4. Discussion

To better understand the significance of ENO1 and ENO2genes in the parasite’s biology, an ideal genetic analysis viaa loss-of-function study would be accomplished by generatinga null-mutation. However, such effort has been improbablebecause of the tight genetic arrangement of these two gene locion chromosome VIII [8,9]. Previously we have shown that theloss-of-function analysis can also be carried out by homolo-gous gene silencing using either in vitro synthesized or over-expression of homologous dsRNA [11,12,15]. Here we

Fig. 2. Effect of constitutive dsRNA expression in stably transgenic parasites. (A) Schematic representation of plasmid constructs used in the generation of

transgenic parasites as described in Materials and method. (B) RT-PCR analyses for dsRNA expression and steady state mRNA levels. Total RNA was isolated from

tachyzoites of the parental parasite (RHDHX) and the transgenic parasite (GFPkd, ENO1kd, and ENO2kd) lines. To confirm dsRNA expression, the oligonu-

cleotide primer specific for either enolase isoform, named ENOx, was used for the synthesis of cDNA derived from the antisense RNA strand of ENO1 or ENO2.

The PCR products are shown in the upper panel and their identity was confirmed by a restriction endonuclease digestion (Supplementary Fig. S1A). The steady

state mRNA levels of ENO2 and ROP1 in these parasites were semi-quantitatively analyzed using equal amounts of cDNA reverse-transcribed with an oligo dT

primer. (C) Western blot analyses were performed to evaluate the level of ENO2 and LDH1 proteins in the parental and transgenic parasites. Normalized values of

ENO2/LDH1 proteins were calculated and obtained following the densitometric measurement of revealed ENO2 and LDH1 bands from three independent

analyses.

24 M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

showed that such a method can be used to lower the expressionof ENO2. Evidently, the presence of ENO2 dsRNA, viaexpression or electroporation, gave rise to a detectable andspecific gene silencing (Fig. 1 and 2). Although a transientknockdown effect was seen in both mRNA and protein levels,the level of ENO2 mRNA was lowered to greater extent thatthose of the ENO2 protein. This could be due to the stabilityand turnover of the ENO2 protein. Also, as these composite

data were taken from three independent electroporations, withvaried transfection efficiencies the amounts of dsRNA maythus have been introduced into the parasites and may result inmixed levels of down-regulation.

It was also noted that the transient and stable gene silencingof ENO2 results in a slightly different expression profile.While the electroporation of ENO1 dsRNA did not give rise tothe down-regulation of ENO2 expression (Fig. 1), the over-

Fig. 3. Parasite growths. The percentage distribution of vacuole size (number

of parasites/vacuole) under tachyzoite culture conditions was determined at

24 h (A) and 32 h (B) after infection with the parental parasite (RHDHX) and

two clonal transgenic lines of ENO2kd, named B9 and WA. The number of

parasites per vacuole was counted for 100 vacuoles from two different

experiments [-A-, RHDHX; ---, ENO2kd_B9; -:-, ENO2kd_WA]. (C)

Comparative mortality of RHDHX (-A-), ENO2kd_B9 (---) and

ENO2kd_WA (-:-). Female BALB/c mice (groups of five) were inoculated

with tachyzoites and mortality was monitored over 14 days.

25M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

expression ENO1 dsRNA in the ENO1kd strain gave rise to anattenuated decrease in ENO2 mRNA. This may possibly beexplained by the perfectly matched and mismatched nucleo-tide sequence spans between the ENO1 dsRNA and the ENO2mRNA (Supplementary Fig. S2). Moreover, a criticalconcentration is probably needed to generate a knockdowneffect. In the transgenic ENO1kd strain the concentrations ofeffectors were kept steady and an enhanced effect maytherefore have been observed. On the contrary, by transientlytransfecting the ENO1 dsRNA, the concentrations of effectors

were not maintained at a high enough level to give anobservable knockdown effect of ENO2. Unexpectedly, thetachyzoites of the parental and transgenic strains exhibitedsimilar enolase enzymatic activity. A plausible explanation isthat the parasite maintains the physiological level of ENO2proteins, but not all of ENO2 proteins are enzymaticallyactive, at a given time. Such an enzymatic regulation wouldallow the parasite to maintain the requisite level of enolaseactivity to keep the glycolytic pathway unperturbed, eventhough the level of enolase protein was altered.

We were able to utilize homologous gene silencing to down-regulate the expression of enolase isoforms. The down-regula-tion revealed that ENO2 has little effect on the doubling time oftachyzoites (w8 h, Fig. 3) and the ability to differentiate intobradyzoites (w15%), when the transgenic and parental strainswere grown under comparable conditions. This finding thussuggests that when compromised, other genes allow the para-sites to maintain viability within the host cells and couldcompensate for the non-enzymatic the function(s) of ENO2.Although a similar level of gene silencing of LDH1 and LDH2allowed the phenotypic displays of the gene functions [15], thelack of a robust phenotype following the ENO2 gene silencingcould not be used to counter-propose the putative function(s) ofenolase in gene regulation. Particularly we observed that underalkali conditions a subpopulation of ENO2kd parasites showedreduced levels of nuclear localization. Although such a pheno-type is not commonly found in an in vitro culture, this phenotypehas been reported in mature cysts found in mice’s brains [6]. Ithas been suggested that the strong nuclear signal of ENO1localization is indicative of its important function during theearly stages of intracellular proliferation and development,where nuclear activity is required [6]. The reduced nuclearlocalization of ENO1 protein is, retrospectively, indicative oflower nuclear activity and mature cyst formation. Here weobserved that all ENO2kd parasites with aberrant ENO1localization had prominent cystewall structures, suggestingthat by altering the expression of ENO2, natural stages ofintracellular proliferation and development are affected ina manner that might mimic a complex hosteparasite interactionas those found in a murine system.

Acknowledgements

This work was supported by the Natural Sciences andEngineering Research Council of Canada and CanadianFoundation for AIDS Research (Canada, SA) and the CentreNational de la Recherche Scientifique (France, ST). We thankDr. Barbara Zielinski for her kind assistance inimmunocytochemistry.

Appendix. Supplementary data

Fig. S1 (A) The RT-PCR products specific to the enolaseantisense RNA, which were obtained from ENO1kd andENO2kd strain, were subjected to SacI (S) and HindIII (H)digestions. A unique HindIII restriction site in the ORF of

Fig. 4. Differentiation of stable parasite lines. (A) The tachyzoites (Tz) of parental (RHDHX) and ENO2kd parasite lines were grown in HFF monolayers under

tachyzoite culture conditions. Nuclei were stained with Hoechst (Htz, left most images). The tachyzoites were subjected to immunofluorescence assays using the

anti-ENO2 (E2) antiserum and then with a rhodamine conjugated secondary antibody. (B) The bradyzoites (Bz) of parental (RHDHX) and ENO2kd parasite lines

were subjected to immunofluorescence assays using the anti-ENO1 (E1) antiserum, and then with a rhodamine conjugated secondary antibody. In addition the

bradyzoites (Bz) were stained with FITC-labeled Dolichos biflorus agglutinin (Dol) and anti-SAG4 antibody in conjunction with FITC labeled conjugated

secondary antibody (SAG4). (C) Confocal image of ENO2kd bradyzoites stained with the anti-ENO1 (E1) antiserum, and subsequently with a rhodamine

conjugated secondary antibody. Arrow heads indicated bradyzoites with nuclear localization. Scales bars represent 5 mm.

26 M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

ENO1 was used to confirm that the RT-PCR fragment wasderived from the ENO1 asRNA; while SacI for ENO2. (B)Constitutive dsRNA expression of GFP dsRNA in the trans-genic GFPkd parasites. The quality of the tachyzoites’ isolatedtotal RNA was verified prior to RT-PCR analyses, and thepresence of dsRNA was validated by RT-PCR of its antisensestrand (asRNA).

Fig. S2 Nucleotide sequence alignment. Partial nucleotidesequences, corresponding to the dsRNA used in the study, of

TgENO1 (542 bps, accession number AF051910) andTgENO2 (542 bps, AF123457) were aligned using Align Xprogram (VNTI version 10) to show various regions whichmight confer crossed gene silencing between the two genetargets; for example, nucleotides 1e22, 295e316 and 356e374.

Supplementary data associated with this article can befound in the online version, at doi: 10.1016/j.micinf.2009.09.010.

27M. Holmes et al. / Microbes and Infection 12 (2010) 19e27

References

[1] L.M. Weiss, K. Kim (Eds.), Toxoplasma gondii, the Model Apicom-

plexan: Perspectives and Methods, Academic Press, Elsevier, 2007.

[2] T. Fleige, K. Fischer, D.J. Ferguson, U. Gross, W. Bohne, Carbohy-

drate metabolism in the Toxoplasma gondii apicoplast: localization of

three glycolytic isoenzymes, the single pyruvate dehydrogenase

complex, and a plastid phosphate translocator. Eukaryot. Cell. 6 (2007)

984e996.

[3] S. Tomavo, The differential expression of multiple isoenzyme forms

during stage conversion of Toxoplasma gondii: an adaptive develop-

mental strategy. Int. J. Parasitol. 31 (2001) 1023e1031.

[4] F. Dzierszinski, O. Popescu, C. Toursel, C. Slomianny, B. Yahiaoui, S.

Tomavo, The protozoan parasite Toxoplasma gondii expresses two

functional plant-like glycolytic enzymes. Implications for evolutionary

origin of apicomplexans. J. Biol. Chem. 274 (1999) 24888e24895.

[5] F. Dzierszinski, M. Mortuaire, N. Dendouga, O. Popescu, S. Tomavo,

Differential expression of two plant-like enolases with distinct enzymatic

and antigenic properties during stage conversion of the protozoan para-

site Toxoplasma gondii. J. Mol. Biol. 309 (2001) 1017e1027.

[6] D.J. Ferguson, S.F. Parmley, S. Tomavo, Evidence for nuclear localisation of

two stage-specific isoenzymes of enolase in Toxoplasma gondii correlates

with active parasite replication. Int. J. Parasitol. 32 (2002) 1399e1410.

[7] J.W. Kim, C.V. Dang, Multifaceted roles of glycolytic enzymes. Trends.

Biochem. Sci. 30 (2005) 142e150.

[8] B. Gajria, A. Bahl, J. Brestelli, J. Dommer, S. Fischer, X. Gao, M.

Heiges, J. Iodice, J.C. Kissinger, A.J. Mackey, D.F. Pinney, D.S. Roos,

C.J. Stoeckert Jr., H. Wang, B.P. Brunk, ToxoDB: an integrated

Toxoplasma gondii database resource. Nucleic Acids Res. 36 (2008)

D553e556.

[9] M.K. Kibe, A. Coppin, N. Dendouga, G. Oria, E. Meurice, M. Mortuaire,

E. Madec, S. Tomavo, Transcriptional regulation of two stage-specifi-

cally expressed genes in the protozoan parasite Toxoplasma gondii.Nucleic Acids Res. 33 (2005) 1722e1736.

[10] M. Soete, D. Camus, J.F. Dubremetz, Experimental induction of brady-

zoite-specific antigen expression and cyst formation by the RH strain of

Toxoplasma gondii in vitro. Exp. Parasitol. 78 (1994) 361e370.

[11] F. Al-Anouti, T. Quach, S. Ananvoranich, Double-stranded RNA can

mediate the suppression of uracil phosphoribosyltransferase expression

in Toxoplasma gondii. Biochem. Biophys. Res. Commun. 302 (2003)

316e323.

[12] F. Al-Anouti, S. Ananvoranich, Comparative analysis of antisense RNA,

double-stranded RNA, and delta ribozyme-mediated gene regulation in

Toxoplasma gondii. Antisense Nucleic Acid Drug. Dev. 12 (2002)

275e281.

[13] D. Soldati, J.C. Boothroyd, A selector of transcription initiation in the

protozoan parasite Toxoplasma gondii. Mol. Cell Biol. 15 (1995) 87e93.

[14] D.S. Roos, R.G. Donald, N.S. Morrissette, A.L. Moulton, Molecular

tools for genetic dissection of the protozoan parasite Toxoplasma gondii.

Methods Cell Biol. 45 (1994) 27e63.

[15] F. Al-Anouti, S. Tomavo, S. Parmley, S. Ananvoranich, The expression of

lactate dehydrogenase is important for the cell cycle of Toxoplasma

gondii. J. Biol. Chem. 279 (2004) 52300e52311.

[16] A.A. Riyahi, F. Al-Anouti, M. Al-Rayes, S. Ananvoranich, Single

argonaute protein from Toxoplasma gondii is involved in the double-

stranded RNA induced gene silencing. Int. J. Parasitol. 36 (2006) 1003e

1014.