JOURNAL OF CLINICAL MICROBIOLOGY,0095-1137/00/$04.0010

Apr. 2000, p. 1453–1460 Vol. 38, No. 4

Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Fractionation of Membrane Components from Tachyzoite Forms ofToxoplasma gondii: Differential Recognition by Immunoglobulin M

(IgM) and IgG Present in Sera from Patients withAcute or Chronic Toxoplasmosis

MONICA GIRALDO,1,2 HELIA CANNIZZARO,1,2 MICHAEL A. J. FERGUSON,3 IGOR C. ALMEIDA,3†AND RICARDO T. GAZZINELLI1,2*

Department of Biochemistry and Immunology, UFMG, 31270-910 Belo Horizonte, MG,1 and Laboratory of Chagas Disease,Centro de Pesquisas Rene Rachou, FIOCRUZ, 30190-002 Belo Horizonte, MG,2 Brazil, and Department

of Biochemistry, University of Dundee, DD1 5EH Dundee, Scotland, United Kingdom3

Received 22 June 1999/Returned for modification 1 September 1999/Accepted 8 December 1999

Tachyzoite forms of Toxoplasma gondii were subjected to a sequential organic solvent extraction, which allowsfractionation of membrane components according to their degrees of hydrophobicity, yielding three fractionsnamed F1 (most hydrophobic) to F3 (least hydrophobic). Fractions F2 (80.85% specificity and 86.95% sensitiv-ity) and F3 (89.36% specificity and 93.61% sensitivity) gave the best results, being preferentially recognized byimmunoglobulin M (IgM) and IgG in sera from patients with acute and chronic toxoplasmosis, respectively. Im-proved scores of specificity (100%) and sensitivity (100%) were achieved when a secondary antibody againsthuman IgG1 instead of total IgG was employed to measure the reactivity of IgG antibodies with the F3 fraction.To purify tachyzoite antigens recognized by human IgM or IgG antibodies, the F2 or F3 fraction was loadedonto an octyl-Sepharose column and eluted with a propan-1-ol gradient. The main antigen(s) recognized by IgM orIgG eluted in a single peak from the octyl-Sepharose resin loaded with either F2 (30 to 50% propan-1-ol) or F3 (15to 35% propan-1-ol), respectively. These semipurified fractions gave improved scores when used to detect T. gon-dii-specific IgM (95.7% specificity and 81.8% sensitivity) or IgG (100% specificity and 93.75% sensitivity) in anenzyme-linked immunosorbent assay. Further biochemical and immunological analyses of antigens partiallypurified from F2 and F3 indicate that glycoinositolphospholipids are preferentially recognized by IgM, whereasproteins of approximately 30 to 40 kDa are recognized by IgG, elicited during T. gondii infection in humans.

Toxoplasma gondii is widespread throughout the world, withno geographic or zoological boundaries, so that human popu-lations are constantly exposed to and infected with this par-asite (7). It is estimated that toxoplasmosis exists in a chronic,asymptomatic form in 500 million to 1 billion of the world’shuman population (17). Whereas infection with T. gondii isusually innocuous or asymptomatic in most individuals, itcauses serious morbidity and mortality in fetuses of primarilyinfected pregnant women (19) and in immunocompromisedindividuals (4). The simultaneous infection with T. gondii andhuman immunodeficiency virus type 1 is of increasing concern,since it is reported that this parasite is the major infectiouscause of encephalitis in AIDS patients, being among the top 10opportunistic infections which are more often encountered asAIDS-defining illness (22).

Therefore, there are at least two major situations in whichthe diagnosis of T. gondii infection, leading to therapeuticintervention, is of medical importance. The first one is thedetection of T. gondii-specific immunoglobulin M (IgM) in serafrom pregnant women, who, if not treated with specific che-motherapy, may have serious fetal problems, including malfor-mation or abortion (19). Second, different studies indicate thatup to 15% of AIDS patients who have positive serological tests

for T. gondii may develop toxoplasmic encephalitis. Toxoplas-mic encephalitis is often difficult to diagnose and has to betreated immediately after the initial symptoms to avoid fatality(2, 16).

Different studies have defined the major targets for T. gon-dii-specific IgM or IgG antibodies found in sera from acutely orchronically infected individuals (6, 19). However, most sero-logical tests used in the laboratory employ parasite extractsrather than purified or recombinant antigens. This is especiallytrue in the case of tests to detect T. gondii-specific IgM thattarget complex glycolipids that are difficult to synthesize in thelaboratory. In addition, false-positive and false-negative re-sults, using commercial kits for parasite-specific IgM detection,are often reported (15). Even in tests for detection of tachy-zoite-specific IgG, the vast majority of which recognize parasiteproteins, the use of recombinant protein or synthetic peptideshas been problematic (23), also yielding dubious results.

In the present study, we used a methodology that employs asequential organic solvent extraction, which allows the frac-tionation of membrane components according to their degreesof hydrophobicity (1, 10). This methodology yielded two dis-tinct fractions, named F2 and F3, which were preferentiallyrecognized by IgM and IgG present in sera from patients withacute and chronic toxoplasmosis, respectively. Because the ma-jor targets for either IgM or IgG have been defined as a specificsubset of glycoinositolphospholipids (GIPLs) (21, 24) or glyco-sylphosphatidylinositol (GPI)-anchored proteins (14, 25), re-spectively, we used hydrophobic interaction chromatographyto further purify the parasite molecules which are major targetsfor human antibodies. The antigens recognized by IgM or IgG

* Corresponding author. Mailing address: Laboratory of Chagas’Disease, Centro de Pesquisas Rene Rachou, FIOCRUZ, Av. Augustode Lima 1715, 30190-002 Belo Horizonte, MG, Brazil. Phone: (031)295-3566. Fax: (031) 295-3115. E-mail: ritoga@dedalus.lcc.ufmg.br.

† Present address: Department of Parasitology, Biomedical ScienceInstitute, University of Sao Paulo, Sao Paulo 05508-900, Brazil.

1453

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from

were eluted as a single peak from octyl-Sepharose resin loadedwith either F2 (30 to 50% propan-1-ol) or F3 (15 to 35%propan-1-ol) and highly enriched. The fractions obtained fromoctyl-Sepharose loaded with F2 and F3, when used in an en-zyme-linked immunosorbent assay (ELISA), resulted in anassay of much higher specificity and approximately the samesensitivity to detect T. gondii-specific IgM and IgG, respec-tively.

MATERIALS AND METHODS

Population studied. Coded serum samples were obtained from 23 patientswith acute T. gondii infection (IgG positive and IgM positive) and 47 patientswith chronic T. gondii infection (IgG positive and IgM negative). Sera from 47uninfected individuals (IgG and IgM negative) were used as controls. The pa-tients with acute infection were further classified as high (n 5 5; average serumtiter, 1:960) and low (n 5 18; average serum titer, 1:126) IgM producers, asindicative of early and late acute toxoplasmosis, respectively. Toxoplasma-specificIgM serological testing was performed with a commercially available immuno-fluorescence assay (IFA) kit with fixed tachyzoites (Imunotoxo; bioMerieux,Marcy l’Etoile, France), using the anti-human IgM (whole-molecule) fluoresceinisothiocyanate conjugate (Fluoline H; bioMerieux) (3). Toxoplasma-specific IgGserological testing was performed using a ELISA kit employing a tachyzoiteextract (Toxonostika IgG; Organon, Boxtel, The Netherlands) (9). All of thepatients with chronic toxoplasmosis were asymptomatic. In contrast, patientswith acute infection presented variable clinical symptoms, ranging from no symp-toms to fever, headache, lymphoadenopathy, and/or pneumonia.

Parasites. Tachyzoites of RH strains of T. gondii were maintained by in vitropassage in human foreskin fibroblasts at 37°C (12). Tachyzoites were harvestedat 4 to 5 days postinfection, centrifuged at 70 3 g for 10 min in order to removecell debris, and then pelleted at 590 3 g for 10 min. The parasite pellet waswashed twice by resuspension in cold phosphate-buffered saline (PBS) and cen-trifugation at 590 3 g for 10 min. The final pellet was stored at 270°C until usedfor sequential organic solvent extraction.

Sequential organic solvent extraction of tachyzoite membrane components.The tachyzoite pellet frozen at 270°C was lyophilized and subjected to extractionwith chloroform-methanol-water (5/10/4, vol/vol) (Fig. 1A) (1, 10). Ten volumesof chloroform-methanol-water was added to the parasite pellet and sonicated for15 min, followed by centrifugation at 5,000 3 g for 15 min at 10°C. The resultingpellet was subjected to same protocol twice more, and the supernatants werepooled, dried in a speed vacuum (Savant Instruments Inc., Farmingdale, N.Y.),and subjected to a butan-1-ol–water (1/1, vol/vol) partition. The butanolic andaqueous phases generated by the butan-1-ol–water partition were named F1 andF2, respectively. The pellet obtained after the chloroform-methanol-water ex-tractions was dried in a speed vacuum and extracted three times with 10 volumesof 9% butan-1-ol for 3 h with shaking at room temperature, followed by centrif-ugation at 5,000 3 g for 15 min at 10°C. The 9% butan-1-ol supernatants werepooled and named F3. The resulting pellet (cell debris) and fractions F1 to F3were all dried and resuspended in water, and their protein concentrations weredetermined by the Bradford method (Bio-Rad Laboratories, Richmond, Calif.)using bovine serum albumin as a standard. Cell debris and F1 to F3 samples werethen stored at 270°C until used in the ELISA and Western blotting assay.

Octyl-Sepharose chromatography. Frozen F2 and F3 fractions were resus-pended in 100 mM ammonium acetate containing 5% propan-1-ol and subjectedto hydrophobic interaction chromatography using octyl-Sepharose resin (Phar-macia Biotech, Uppsala, Sweden) elution with a propan-1-ol (5 to 60%) gradient.Two-milliliter fractions were collected and assayed for myo-inositol content,protein concentration, and the ability to bind to IgM and IgG present in serafrom patients with acute and chronic toxoplasmosis, respectively.

myo-Inositol measurements. Briefly, samples were preincubated with 40 pmolof deuterated myo-inositol, dried in a SpeedVac centrifuge (Savant Instruments),resuspended in 50 ml of deionized water, and transferred to glass capillary tubes.Samples were dried again, and 50 ml of 6 N HCl was added. The capillary tubeswere then sealed under vacuum and subjected to hydrolysis at 110°C for 16 to18 h. Samples were dried under vacuum, and the residual HCl was removed byevaporation after addition of 50 ml of water. For dehydration, 50 ml of methanolwas added to each sample and dried under vacuum. The samples were thenincubated with fresh trimethylsilyl (TMS) reagent for 15 to 30 min at roomtemperature. TMS derivatives were analyzed (1 ml per sample) in an SE-54 (0.25mm by 30 m; Alltech) capillary column using a temperature gradient of 140°C for1 min, 140 to 250°C for 7.3 min (15°C/min), and 250°C for 5 min. Selective ionmonitoring was carried out for TMS derivatives of d6-myo-inositol at 307 and 321m/z and of myo-inositol at 305 and 318 m/z (5).

ELISA. Immulon-2 plates (Dynatech Laboratories, McLean, Va.) were coatedwith 100 ml of either F1, F2, or F3 at a protein concentration of 10 mg/ml in 0.05M carbonate-bicarbonate buffer, pH 9.6. Alternatively, 0.5 pmol of F2-derivedeluate F, or 1.0 pmol of F3-derived eluates E and F, per well in 50 ml of 0.05 Mcarbonate-bicarbonate buffer (pH 9.6) was used to coat the Immulon-2 plates.Plates were incubated overnight at 4°C, blocked with 2% casein (Calbiochem, LaJolla, Calif.) for 2 h at 37°C, and then washed four times with 0.15 M PBS (pH

7.2)–0.05% Tween 20 (Sigma, St. Louis, Mo.) (PBS-T). One-hundred-microliterportions of sera at dilutions of 1:50 to 1:200 in PBS-T–1% bovine serum albumin(Biobras, Montes Claros, Brazil) were added and incubated for 1 h at 37°C.Plates were then incubated with biotinylated conjugates of anti-human IgG,IgG1, IgG2, IgG3, IgG4, or IgM (Sigma) at 1:20,000 in PBS-T for additional 1 hat 37°C and washed with PBS-T. Streptavidin-peroxidase conjugate (Sigma) at a1:1,000 dilution was added and incubated for 30 min at 37°C. The plates werethen washed with PBS-T and developed using ABTS [2,29-azinobis(3-ethylbenz-thiazolinesulfonic acid)] as a substrate. The reaction was terminated by theaddition of 50 ml of 1% sodium dodecyl sulfate (SDS) solution, and results wereread at 405 nm.

SDS-PAGE. Different tachyzoite antigen preparations were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on 10 or 15% gels under re-ducing conditions as previously described (8). Gels were silver stained afterfixation for 30 min in 50% methanol–10% acetic acid and for 30 min in 5%methanol–7% acetic acid, followed by 30 min of incubation with 20 mM dithio-threitol and 0.1% AgNO3. The gels were developed in 3% Na2CO3–0.01% form-aldehyde.

Immunoblotting. Proteins separated by SDS-PAGE were transferred to nitro-cellulose paper (27), using the Mini-Protean system (Bio-Rad). Alternatively, 5ml of parasite antigen suspension in PBS was added to a nitrocellulose paper,which was left drying for 10 min. Blots or dot blots were first soaked in 2%casein–PBS-T for 1 h at room temperature to block free binding sites. The blotswere then incubated overnight at 4°C with a pool of sera, at a 1:200 dilution, frompatients with either acute or chronic toxoplasmosis or from individuals who didnot have any evidence of T. gondii infection. The nitrocellulose sheets were thenincubated with biotin-conjugated goat anti-human IgM or IgG antibody (Sigma)for 1 h at room temperature and then reincubated for 30 min at room temper-ature with streptavidin-peroxidase conjugate (Sigma) at a 1:1,000 dilution. Aftereach incubation, the membranes were washed three times with PBS-T. Finally,after being rinsed with 0.05 M carbonate-bicarbonate buffer (pH 9.6), blots were

FIG. 1. (A) Strategy used for fractionating components from tachyzoite par-asites based on their hydrophobicity-hydrophilicity properties. (B) Ten micro-grams of F1, F2, F3, or cell debris (F4) was run on an SDS–15% polyacrylamidegel and silver stained. The numbers on the left indicate the molecular masses ofproteins used as standard markers.

1454 GIRALDO ET AL. J. CLIN. MICROBIOL.

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from

incubated with ECL reagent (Amersham, Little Chalfont, England) and exposedto X-ray films.

ES-MS analysis. Electrospray-mass spectrometry (ES-MS) analysis was car-ried out on a Quattro apparatus (Micromass, Manchester, United Kingdom) innegative mode. Samples diluted in 50% propan-1-ol–0.2% formic acid wereintroduced into the ES source at 5 ml/min using a Harvard syringe pump. Thecapillary voltage was kept at 2.3 kV, the cone voltage was kept at 40 V, and thecone/skimmer offset was kept at 5 V.

Statistical analysis. The antigen concentration and serum dilution were de-fined by analysis of variance with 10 samples from individuals with either acuteor chronic toxoplasmosis. The positive-negative borderline was calculated by Zdistribution. For IgM assays, 23 acute, 24 chronic, and 23 unreactive sampleswere used; for IgG assays, 47 chronic and 47 negative samples were used. Thesensitivity and specificity of our assays with each antigen were calculated by thechi-square test. Analyses were performed using the Statistic software (version4.5).

RESULTS

Preparation of tachyzoite extracts according to their hydro-phobic-hydrophilic properties. As shown in Fig. 1, we used astrategy that yields three different fractions (F1, F2, and F3)based on their degrees of hydrophobicity (Fig. 1A), where F1and F3 were the most and least hydrophobic fractions, respec-tively. F4 was considered cell debris; it was not solubilized bythe solvent system used. The results presented in Fig. 1B showthat, except for F1, the fractions generated by this protocol stillpresented a complex protein profile when analyzed by SDS-PAGE. When analyzed for their ability to be recognized byhuman sera, F2 or F3 was preferentially recognized by IgM orIgG antibodies present in sera from acutely or chronicallyinfected individuals, respectively (see below).

Identification of tachyzoite antigens that are preferentiallyrecognized by IgM antibodies from sera of patients acutelyinfected with T. gondii. The results presented in Table 1 showthe ability of the F2 fraction (80.85 specificity and 86.95%sensitivity) to detect specific IgM antibodies present in sera ofpatients with acute toxoplasmosis. However, a high number offalse-positive results were observed in the experiments usingthe F2 fraction, as indicated by the relatively low (80.85%)specificity of the assay.

The pattern of antigen complexity of this fraction was fur-ther analyzed by immunoblotting analysis, using sera from pa-tients with acute toxoplasmosis. Figure 2A shows that the mainantigen recognized in the F2 fraction by the IgM antibodiespresent in sera from acutely infected patients was an antigenwith a diffuse pattern of migration and an apparent molecularmass of below 14 kDa. This antigen was also recognized by amonoclonal antibody (MAb), T33F12, against GIPLs from T.gondii tachyzoites (Fig. 2C) (23). A less diffuse band of approx-imately 30 kDa and a more defined band at 70 kDa present inthe F2 fraction were also recognized by IgM antibodies presentin sera from acutely infected patients.

Identification of tachyzoite antigens that are preferentiallyrecognized by IgG antibodies from sera of patients chronicallyinfected with T. gondii. Our experiments also indicate that F3gave the best results for IgG detection, with higher and loweraverages for infected and uninfected individuals as well as thebest sensitivity (93.61%) and specificity (89.36%) scores (Table2).

Figure 2B shows that the pattern of recognition of differentfractions by IgG was more complex than the pattern generatedby IgM antibodies. A band of approximately 30 kDa was themajor antigen recognized by T. gondii-specific IgG from sera ofpatients with chronic toxoplasmosis. This antigen was presentin both F2 and F3. In addition to the 30-kDa antigen, manyother antigens with molecular masses of above 30 kDa werealso recognized by IgG antibodies in sera from chronicallyinfected individuals. The same 30-kDa antigen appears to berecognized by IgM antibodies, but with much lower intensity(Fig. 2A and B) than IgG antibodies, from chronically infectedpatients. Accordingly, in the ELISA the total tachyzoite soni-cate and F3 fraction were poorly recognized by T. gondii-specific IgM antibodies compared to the F2 fraction.

FIG. 2. Western blotting analysis of F2 and F3 fractions developed with apool of sera (dilution, 1:200) from patients with acute toxoplasmosis (A), a poolof sera (dilution, 1:100) from patients with chronic toxoplasmosis and (B), MAbT33F12, specific for tachyzoite-derived GIPLs (C). Ten micrograms of F2 or F3was run on an SDS–15% polyacrylamide gel, transferred to a nitrocellulosesheet, incubated with specific antibodies, and then developed using an ECL kit.The numbers on the left indicate the molecular masses of proteins used asstandard markers.

TABLE 1. Specificity and sensitivity of the ELISA using F2tachyzoite extract to identify patients with acute toxoplasmosis

and high levels of anti-T. gondii IgMa

Patient group (n)OD405

Median SD

Uninfected (24) 0.129 0.06Acutely infected (23) 0.503 0.17Chronically infected (23) 0.214 0.09

a The ELISA was developed as described in Materials and Methods. Theoptical densities at 405 nm (OD405) were obtained from individual sera frompatients from the same group. The sensitivity (86.95%) and specificity (80.85%)were calculated using the chi-square test to compare the ELISA using the F2fraction and anti-IgM conjugate with IFA (P , 0.001).

TABLE 2. Specificity and sensitivity of the ELISA using the F3fraction to identify patients with chronic toxoplasmosis

and high levels of anti-T. gondii IgGa

IgGisotype

OD405 for:

Sensitivity(%)

Specificity(%) PUninfected

patients

Chronicallyinfectedpatients

Median SD Median SD

Total IgG 0.209b 0.12 1.036b 0.25 93.61 89.36 ,0.001IgG1 0.165 0.04 1.418 0.82 100 100 ,0.001IgG2 0.171 0.06 0.263 0.25 75.8 80 0.06IgG3 0.218 0.04 0.354 0.14 75.8 83.3 ,0.001IgG4 0.158 0.03 0.349 0.45 82.7 73.3 0.012

a The ELISA was developed as described in Materials and Methods. Theoptical densities at 405 nm (OD405) were obtained from individual sera from thesame group. The sensitivity and specificity were calculated using the chi-squaretest to compare ELISA using the F3 fraction and either anti-total IgG, anti-IgG1,anti-IgG2, anti-IgG3, or anti-IgG4 conjugates with a commercially available testto detect T. gondii-specific IgG antibodies. Unless otherwise indicated, n 5 30 foruninfected patients and n 5 29 for chronically infected patients.

b n 5 47.

VOL. 38, 2000 TACHYZOITE MEMBRANE TARGETS FOR IgG AND IgM 1455

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from

We also determined the main IgG isotype present in sera ofpatients with chronic toxoplasmosis that recognized the tachy-zoite antigens present in F3. The results presented in Table 2show a clear dominance of the IgG1 isotype among IgG anti-bodies specific for T. gondii antigens. Importantly, the use ofanti-IgG1 instead of anti-total IgG secondary antibody also re-sulted in an increased specificity (100%) and sensitivity (100%)to discriminate infected from uninfected individuals.

Purification and partial characterization of tachyzoite mol-ecules recognized by human IgM and IgG from sera of pa-tients with acute or chronic toxoplasmosis. In order to im-prove the scores of our ELISA test, we decided to furtherpurify components of the tachyzoite membrane by hydropho-bic interaction chromatography using octyl-Sepharose. In fact,different studies suggest that GIPLs and GPI-linked proteinsare the main targets of IgM (19, 20, 22) and IgG (13, 23)antibodies present in sera from humans infected with T. gondii.

Figure 3A shows protein concentrations (absorbance at 280nm) and myo-inositol concentrations of fractions A to H re-leased during the propan-1-ol gradient treatment used to re-lease the tachyzoite molecules from octyl-Sepharose columnsloaded with F2 and F3. Major protein and myo-inositol peakswere observed in eluate B (5% propan-1-ol) for the columnloaded with either F2 or F3, and these correspond to unboundmaterial. Two additional major myo-inositol peaks were de-tected in eluates E (30% propan-1-ol) and F (40% propan-1-ol) from the octyl-Sepharose column loaded with F2. Minormyo-inositol peaks were also observed in eluates E (15% pro-pan-1-ol), F (25% propan-1-ol), and G (35% propan-1-ol)from the octyl-Sepharose column loaded with F3.

Each of the eluates obtained from octyl-Sepharose columnsloaded with F2 and F3 were also characterized for their ability

to be recognized by IgM and IgG from sera of patients withacute or chronic toxoplasmosis (Fig. 3B). These studies wereperformed using the dot immunoblotting analysis. Our resultsshow that the eluate F and, to a lesser extent, eluate E elutedfrom octyl-Sepharose loaded with F2 were preferentially rec-ognized by IgM antibodies present in sera from patientsacutely infected with T. gondii. This same eluate F was recog-nized specifically by MAb T33F12. In contrast, eluate F ob-tained from the octyl-Sepharose column loaded with F3 re-acted preferentially with IgG antibodies from sera of patientschronically infected with T. gondii. No reactivity with any of theeluates was observed when we used sera from uninfected in-dividuals.

Each of the fractions that showed reactivity with either hu-man IgM or IgG was further analyzed by SDS-PAGE andimmunoblotting analysis. Our results demonstrate that eluatesE and F did not present a single protein band when silverstained. Only a major diffuse band with molecular mass ofbelow 14 kDa was prominent in eluate F obtained from thecolumn loaded with F2 (Fig. 4A, left panel). This low-molec-ular-mass diffuse band was recognized by IgM antibodies fromsera of patients with acute toxoplasmosis (Fig. 4A, middlepanel). After a further butan-1-ol–water partition, the butan-olic phase of F2-derived eluate F showed by ES-MS (negativemode) a group of doubly charged [(M 2 2H)22] pseudomo-lecular ions at m/z 900 to 1150. At least four major species atm/z 905, 959, 1015, and 1028 were observed. Interestingly, twoof the less abundant species, at m/z 905 and 986, have esti-mated molecular masses (1,812 and 1,974 Da, respectively)consistent with two major GIPL structures previously report-ed (Fig. 4A, right panel) (23). These structures correspond to(i) (ethanolamine-PO4)-Mana1-2Mana1-6(GalNAcb1-4)

FIG. 3. Purification of tachyzoite molecules recognized by human IgM and IgG present in sera of patients with acute and chronic toxoplasmosis. (A) Fraction F2or F3 was loaded into an octyl-Sepharose column and eluted in a gradient of propan-1-ol. The octyl-Sepharose eluates were pooled in eight distinct fractions (A to H),and the propan-1-ol, myo-inositol, and protein (absorbance at 280 nm) concentrations in each were measured. (B) Two microliters of each eluate (A to H),unfractionated F2 or F3, or total parasite extract was dotted on a nitrocellulose sheet that was then incubated with specific antibodies, i.e., a pool of acute sera (dilution,1:200), a pool of chronic sera (dilution, 1:100), a pool of sera from noninfected individuals (dilution, 1:100), or MAb T33F12 (dilution, 1:100). Dot immunoblottingwas performed with an ECL kit.

1456 GIRALDO ET AL. J. CLIN. MICROBIOL.

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from

Mana1-4GlcNa-inositol-PO4-diacyl(C16:0/C18:0)-glycerol (mo-lecular mass, 1,812 Da) and (ii) (ethanolamine-PO4)-Mana1-2Mana1-6(Glca1-4GalNAcb1-4)Mana1-4GlcNa-inositol-PO4-diacyl(C16:0/C18:0)-glycerol (molecular mass, 1,974 Da). Infact, most of the major doubly charged species observed (m/z959, 1015, 1028, 1040, 1096, 1109, and 1121) could be derivedfrom the species at m/z 905 and 986, as indicated in Fig. 4A(right panel) and Table 3.

Eluates E and F obtained from the column loaded with F3showed an intense reactivity with sera from chronically in-fected individuals (Fig. 4B). Analysis of silver-stained poly-acrylamide gels showed that whereas eluate E had multiplebands with apparent molecular masses of 20 to 33 kDa, eluateF had a major protein band of approximately 40 kDa. Asshown by immunoblotting analysis, a band with average mo-lecular mass of approximately 30 kDa was the main IgG target

FIG. 4. Gel electrophoresis, immunoblot, and mass spectrometry analyses of F2-derived eluates E and F and F3-derived eluates E and F. (A) Ten micrograms ofF2-derived eluate E or F was run on an SDS–15% polyacrylamide gel and silver stained (left panel) or transferred to a nitrocellulose sheet for immunoblotting analysisusing a pool of sera from patients with acute toxoplasmosis (dilution, 1:200) and an anti-human IgM secondary antibody (middle panel). The right panel shows theES-MS profile of a GIPL preparation from tachyzoite membranes, which is highly reactive with human IgM MAbs and MAb T33F12. (B) Ten micrograms of F3-derivedeluate E or F was run on an SDS–15% polyacrylamide gel and silver stained (left panel) or transferred to a nitrocellulose sheet for immunoblotting analysis using apool of sera from patients with chronic toxoplasmosis (dilution, 1:100) and a secondary anti-human IgG antibody (right panel). The numbers on the left indicate themolecular masses of proteins used as standard markers.

TABLE 3. Proposed assignments for Toxoplasma GIPL species observed by ES-MS

GIPL series [(M 2 2H)22] (m/z) Mass (Da) Proposed assignmenta

A 905 1,812 (Hex3HexNAc)(EtNP)-HexN-InsP-(C16:0/C18:0)DAGb

986 1,974 (Hex4HexNAc)(EtNP)-HexN-InsP-(C16:0/C18:0)DAGb

B 959 1,920 (Hex3HexNAc)(EtNP)(AEP)-HexN-InsP-(C16:0/C18:0)DAG1,040 2,082 (Hex4HexNAc)(EtNP)(AEP)-HexN-InsP-(C16:0/C18:0)DAG1,121 2,244 (Hex5HexNAc)(EtNP)(AEP)-HexN-InsP-(C16:0/C18:0)DAG

C 1,015 2,032 NDc

1,096 2,194 ND

D 1,028 2,058 (Hex4HexNAc)(EtNP)-HexN-InsP-(C16:0/22:0)DAG1,109 2,220 (Hex5HexNAc)(EtNP)-HexN-InsP-(C16:0/22:0)DAG

a Hex, hexose; HexNAc, N-acetylhexosamine; EtNP, ethanolaminephosphate; HexN, hexosamine; InsP, myo-inositol-phosphate; DAG, diacylglycerol; AEP, 2-ami-noethylphosphonate.

b Based on structures described by Striepen et al. (21).c ND, not determined.

VOL. 38, 2000 TACHYZOITE MEMBRANE TARGETS FOR IgG AND IgM 1457

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from

present in eluate E (Fig. 4B, right panel). In eluate F, weobserved two bands, at 33 and 40 kDa, with strong reactivitywith sera of chronically infected individuals.

Improvement of ELISA scores using eluates from octyl-Sepharose columns loaded with either fraction F2 or F3. Asshown in Fig. 5A, each of the eluates from the octyl-Sepharosecolumn loaded with F2 was tested for the ability to react withsera from infected as well as uninfected individuals. Our resultsshow that for discriminating acutely infected from chronicallyinfected and uninfected individuals, eluate F had the best per-formance (Fig. 5A). Importantly, even the sera from individ-uals producing low levels of tachyzoite-specific IgM were read-ily detected in our ELISA using the F2 eluate F (data notshown). Comparing the total F2 fraction with eluate F, thesensitivity of the assay persisted in the range of 81.8%; how-ever, the specificity of the assay improved from 80.85 to 95.7%(Fig. 5A). Periodate treatment destroyed most of the reactivityof F2 eluate F with IgM from sera of patients with acutetoxoplasmosis (data not shown), indicating the carbohydratenature of these epitopes.

An improvement was also observed when we compared theeluates obtained from the octyl-Sepharose column loaded withF3 in regard to their abilities to be recognized by sera fromchronically infected but not from uninfected individuals. Theresults presented in Fig. 5B demonstrate that eluates E and Fwere highly effective in discriminating sera from chronically

infected and uninfected individuals, as seen by their high spec-ificity (100%) and sensitivity (93.75%) scores. Treatment withproteinase K destroyed most reactivity of F3 eluates E and Fwith IgG from sera of patients with chronic toxoplasmosis(data not shown), indicating the proteinaceous nature of theseepitopes.

The results presented in Fig. 6 show the individual values ofparasite-specific IgM (Fig. 6A) or IgG (Fig. 6B) in ELISAusing different antigen preparations as well as sera from pa-tients with acute toxoplasmosis, patients with chronic toxoplas-mosis, and uninfected controls. These results show an alreadysubstantial improvement after the sequential organic extrac-tion when comparing the serology results using F2 (IgM) andF3 (IgG) with total tachyzoite extracts. The data were furtherimproved when eluate F from F2 and eluates E and F from F3were used to measure IgM and IgG specific for tachyzoiteantigens, respectively. The latter improvement was mainly dueto an increase in the specificity of the assay, i.e., a decrease inthe number of false-positive results with sera from chronicallyinfected patients in the ELISA to measure T. gondii-specificIgM as well as a reduction in the number of false-positiveresults with sera from uninfected controls in the assay used tomeasure parasite-specific IgG.

DISCUSSION

Despite major advances in the field of DNA technology,most serological tests used for diagnosis of T. gondii infectionstill employ paraformaldehyde-fixed parasites or crude extractsfrom tachyzoites instead of parasite recombinant antigens.Thus, in addition to the Sabin-Feldmen test (17), which isconsidered the standard serological test for toxoplasmosis, im-munofluorescence of fixed parasites is used for detection ofIgM present in sera of acutely infected patients. For detectionof IgG present in sera of chronically infected patients, anELISA using total tachyzoite extracts is the most usual methodemployed. The failure of recombinant antigens to provide atest with high specificity and sensitivity scores may be in partattributed to the facts that (i) carbohydrates instead of pep-tides are the major targets for IgM antibodies elicited duringthe acute infection with T. gondii and (ii) improper folding ofrecombinant antigens may result in a dramatic reduction in thebinding of a considerable amount of anti-Toxoplasma IgG an-tibodies, which may recognize tertiary rather than primarypeptide structures.

As previously established, patients in the early stages ofacute toxoplasmosis produce high levels of parasite-specificIgM (3). Therefore, our acutely infected patients were dividedinto those producing high and low levels of T. gondii-specificIgM, independent of the levels of parasite-specific IgG. Thesera from uninfected controls were all negative for T. gondii-specific IgM and IgG, whereas sera from patients with chronictoxoplasmosis were all IgM negative and IgG positive as de-termined by parasite-specific IFA and ELISA, respectively. Inthe present study we compared different extracts preparedfrom tachyzoite antigens in regard to their ability to discrimi-nate sera from patients acutely or chronically infected withT. gondii from those from uninfected individuals.

Several studies suggest that the main targets for antibodyproduction during the acute and chronic phases of infectionare the surface antigens present in the tachyzoite membrane.More precisely, in humans most of the IgM responses againstT. gondii are directed against the carbohydrates (11), whichwere recently shown to be a branch derived from the glycancore of a unique GIPL structure (21). In addition, the surfaceantigens of approximately 20 (SAG-2), 30 (SAG-1), and 40

FIG. 5. Human IgG and IgM recognition of eluates derived from octyl-Sepharose columns loaded with F2 and F3 fractions. (A) Fraction F2 was loadedonto an octyl-Sepharose column and eluted in a gradient of propan-1-ol, asdescribed for Fig. 3A. The octyl-Sepharose eluates were tested for their ability tobe bound by human IgM present in sera (dilution, 1:200) from patients withacute toxoplasmosis. The results represent the means and standard deviations forsera from 23 uninfected controls (M), 23 patients with acute toxoplasmosis(f), and 24 patients with chronic toxoplasmosis (d). (B) Fraction F3 wasloaded onto an octyl-Sepharose column and eluted in a gradient of propan-1-olas described for Fig. 3A. The octyl-Sepharose eluates were tested for their abilityto be bound by human IgG present in sera (dilution, 1:100) from patients withchronic toxoplasmosis. The results represent the means and standard deviationsfor sera from 29 uninfected controls and 32 patients with chronic toxoplasmosis.The ELISA was developed as described in Materials and Methods. The meansand standard deviations were obtained from optical densities (OD) at 405 nmobtained from individual sera from patients of the same group. Asterisks indicateeluates with higher performance in discriminating sera from patients at differentstages of infection with T. gondii, as determined by the chi-square test.

1458 GIRALDO ET AL. J. CLIN. MICROBIOL.

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from

(SAG-3) kDa have also been shown to be major targets for IgGresponses during chronic infection with T. gondii in humans;several studies suggest a dominant response to SAG-1 (6). It isnoteworthy that most of the surface molecules are linked to thetachyzoite surface through GPI anchors (13, 18, 25, 26). Thestrategy used to prepare tachyzoite extracts was the adaptationof a protocol first used for fractionation of Leishmania dono-vani (10) and Trypanosoma cruzi (1) membrane componentsbased on their hydrophobicities. As described in Materials andMethods, this protocol generates three fractions, F1 to F3,which consist of highly hydrophobic molecules (F1) (e.g., phos-pholipids), amphipathic components (F2) (e.g., GIPLs), andhydrophilic molecules (F3) (e.g., GPI-linked glycoproteins).

This study shows that by using sequential organic solventextraction, we were able to produce a tachyzoite extract,named F2, which was highly enriched for GIPLs and displayeda pronounced ability to identify sera from patients with highToxoplasma-specific IgM titers. However, this fraction stillgave a high number of false-positive results and therefore lowscore for specificity (80.85%). In contrast, the F3 extract gaveexcellent results in discriminating sera from T. gondii-infectedindividuals from those from uninfected individuals, with spec-ificity and sensitivity scores in the ranges of 89.36 and 93.61%,respectively; both are within the values required for standardserological methods for toxoplasmosis. Further improvementin discriminating sera from chronically infected individualsfrom those from uninfected controls in the ELISA was ob-tained from the use of the F3 fraction and a secondary anti-

body against IgG1, instead of total IgG, which resulted in a testwith 100% sensitivity as well as 100% specificity.

Further improvements of our ELISA scores were obtainedafter fractionation of the F2 and F3 extracts using an octyl-Sepharose column. Thus, the specificity and sensitivity of ourELISA for IgM, employing eluate F, were 95.7 and 81.8%,respectively. The biochemical and immunochemical data areconsistent with the fact that eluate F, obtained from the octyl-Sepharose column loaded with F2, consisted mainly of GIPLsderived from tachyzoite membranes. Furthermore, this is inagreement with previous studies showing that the IgM anti-bodies from acutely infected patients recognize mainly carbo-hydrate epitopes (11).

We also observed a small increase in the specificity scorewhen F3-derived eluates E (100%) and F (100%) were usedinstead of the F3 extract to discriminate sera of chronicallyinfected individuals from those of uninfected individuals. Theseeluates E and F consisted mainly of protein of approximately30 and 40 kDa, respectively. In contrast to the antibodies of theIgM isotype, the IgG antibodies were directed mainly againstproteinaceous epitopes, as previously suggested by Hadman etal. (6) and Noat et al. (14).

Thus, our study shows that by using a simple biochemicalprocedure we can fractionate the major membrane compo-nents of the tachyzoite membrane. The use of these proteins of30 and 40 kDa (eluates E and F from F3) leads to an improve-ment of the specificity and sensitivity scores of the ELISA fordetecting sera from patients with chronic toxoplasmosis. In

FIG. 6. Individual serological tests for parasite-specific IgM and IgG using tachyzoite-derived preparations after different steps of purification. (A) Sera (dilution,1:200) from patients with acute (n 5 23) and chronic (n 5 24) toxoplasmosis and from uninfected individuals (n 5 23) were tested for sonicated tachyzoites, F2, andthe F2 eluate F. (B) Sera (dilution, 1:100) from patients with chronic toxoplasmosis (n 5 32) and uninfected individuals (n 5 29) were tested for sonicated tachyzoites,F3, and F3 eluate E or F. Results for F3 eluate F are not shown but were identical to those obtained with F3 eluate E [indicated as F3(E/F)]. The ELISAs weredeveloped as described in Materials and Methods. The means and standard deviations were obtained from optical densities (OD) at 405 nm obtained from individualsera from patients of the same group.

VOL. 38, 2000 TACHYZOITE MEMBRANE TARGETS FOR IgG AND IgM 1459

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from

addition, false-negative results are common finding in IFAused to detect tachyzoite-specific IgMs. The main reason forthe false-negative results is the saturation of IgM binding sitesby IgG antibodies. In order to avoid this problem, the use ofIgM capture assays to measure T. gondii-specific antibodies hasbeen recommended. Our data indicate that the direct recog-nition of fraction F2 eluate F by IgM is minimally affected bytachyzoite-specific IgG. Therefore, the chemical isolation of afraction highly enriched for tachyzoite-derived GIPLs that arepreferentially recognized by IgM, but not IgG, antibodies mayhelp in the development of a simpler direct ELISA for detect-ing T. gondii-specific IgM antibodies, with high specificity andfewer problems with false-negative results.

ACKNOWLEDGMENTS

We thank Jean Francois Drubemetz and Striepen Boris for provid-ing MAb T33F12 and Leonides Resende, Jr., for providing human seratested for T. gondii-specific IgG and/or IgM.

This work was supported in part by CNPq/PADCT SBIO (62.0106/95-6). R.T.G. is a research fellow of the CNPq. M.G. and H.C. aregraduate students with scholarships from COLCIENCIAS and CAPES,respectively. I.C.A. was a postdoctoral fellow with a fellowship (no.96/04260-0) from FAPESP.

REFERENCES

1. Almeida, I. C., M. A. J. Ferguson, S. Schenkman, and L. R. Travassos. 1994.Lytic anti-a-galactosyl antibodies from patients with chronic Chagas’ diseaserecognize novel O-linked oligosaccharides on mucin-like glycosyl-phospha-tidylinositol-anchored glycoproteins of Trypanosoma cruzi. Biochem. J. 304:793–802.

2. Ammassari, A., R. Murri, A. Cingolani, A. DeLuca, and A. Antinori. 1996.AIDS-associated cerebral toxoplasmosis: an update on diagnosis and treat-ment. Curr. Topics Microbiol. Immunol. 219:209–222.

3. Camargo, M. E., and P. G. Leser. 1976. Diagnostic information from sero-logical tests in human toxoplasmosis. Rev. Inst. Med. Trop. Sao Paulo 18:227–238.

4. Dubey, J. P. 1998. Advances in the life cycle of Toxoplasma gondii. Int. J.Parasitol. 28:1019–1024.

5. Ferguson, M. A. J. 1992. Chemical and enzymatic analysis of glicosyl-phos-phatidylinositol anchors, p. 196–230. In N. M. Hooper and A. J. Turner (ed.),The chemical and enzymatic analysis of GPI fine structure. Lipid modifica-tion of proteins: a practical approach. IRL Press, Oxford, United Kingdom.

6. Hadman, E., J. W. Goding, and J. S. Remington. 1980. Detection andcharacterization of membrane antigens of Toxoplasma gondii. J. Immunol.124:2578–2583.

7. Joiner, K. A., and J. F. Dubremetz. 1993. Toxoplasma gondii: a parasite forthe nineties. Infect. Immun. 61:1169–1172.

8. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227:680–685.

9. McCabe, R. E., and J. S. Remington. 1983. The diagnosis and treatment oftoxoplasmosis. Eur. J. Clin. Microbiol. 2:95–104.

10. McConville, M. J., and J. M. Blackwell. 1991. Developmental changes in the

glycosylated phosphatidylinositols of Leishamania donovani. Characteriza-tion of the promastigote and amastigote glycolipids. J. Biol. Chem. 266:15170–15179.

11. Mineo, J. R., M. E. Camargo, and A. W. Ferreira. 1980. Enzyme-linkedimmunosorbent assay for antibodies to Toxoplasma gondii polysaccharides inhuman toxoplasmosis. Infect. Immun. 27:283–287.

12. Nagel, S. D., and J. C. Boothroyd. 1988. The alpha and beta tubulins ofToxoplasma gondii are encoded by single copy genes containing multipleintrons. Mol. Biochem. Parasitol. 29:261–273.

13. Nagel, S. D., and J. C. Boothroyd. 1989. The major surface antigen, P30, ofToxoplasma gondii is anchored by a glycolipid. J. Biol. Chem. 264:5569–5574.

14. Noat, Y., D. R. Guptill, J. Mullenax, and J. S. Remington. 1983. Character-ization of Toxoplasma gondii antigens that react with human immunoglobulinM and immunoglobulin G antibodies. Infect. Immun. 41:331–338.

15. Noat, Y., and J. S. Remington. 1980. An enzyme-linked immunosorbentassay for detection of IgM antibodies of Toxoplasma gondii: use for diagnosisof acute acquired toxoplasmosis. J. Infect. Dis. 142:757–766.

16. Porter, S. B., and M. A. Sande. 1992. Toxoplasmosis of the central nervoussystem in the acquired immunodeficiency syndrome. N. Engl. J. Med. 327:1643–1648.

17. Savva, D. 1992. Toxoplasma, p. 163–185. In S. Myint and A. Cann (ed.),Molecular and cell biology of opportunistic infections in AIDS. Chapman &Hall, London, United Kingdom.

18. Schwarz, R. T., and S. Tomavo. 1993. The current status of the glycobiologyof Toxoplasma gondii: glycosylphosphatidylinositols, N- and O-linked gly-cans. Res. Immunol. 144:24–31.

19. Sharma, S. D. 1990. Immunology of toxoplasmosis, p. 184–199. In D. J.Wyler (ed.), Modern parasite biology cellular, immunological and molecularaspects. W. H. Freeman and Co., New York, N.Y.

20. Sharma, S. D., J. Mullenax, F. G. Araujo, H. A. Erlich, and J. S. Remington.1983. Western blot analysis of the antigens of Toxoplasma gondii recognizedby human IgM and IgG antibodies. J. Immunol. 131:977–983.

21. Striepen, B., C. F. Zinecker, J. B. L. Damm, P. A. T. Melgers, G. J. Gerwig,M. Koolen, J. F. G. Vliegenthart, J. F. Dubremetz, and R. T. Schwarz. 1997.Molecular structure of the “low molecular weight antigen” of Toxoplasmagondii: a glucose a1-4 N-acetylgalactosamine makes free glycosyl-phosphati-dylinositols highly immunogenic. J. Mol. Biol. 266:797–813.

22. Suzuki, Y., and J. S. Remington. 1993. Toxoplasmic encephalitis in AIDSpatients and experimental models for study of the disease and its treatment.Res. Immunol. 144:66–67.

23. Tenter, A. M., and A. M. Johnson. 1991. Recognition of recombinant Tox-oplasma gondii antigens by human sera in an ELISA. Parasitol. Res. 77:197–203.

24. Tomavo, S., G. Couvreur, M. A. Leriche, A. Sadak, A. Achbarou, B. Fortier,and J. F. Dubremetz. 1994. Immunolocalization and characterization of thelow molecular weight antigen (4-5 kDa) of Toxoplasma gondii that elicits anearly IgM response upon primary infection. Parasitology 108:139–145.

25. Tomavo, S., J. F. Dubremetz, and R. T. Schwarz. 1992. A family of glycolipidsfrom Toxoplasma gondii. Identification of candidate glycolipid precursor(s)for Toxoplasma gondii glycosylphosphatidylinositol membrane anchors.J. Biol. Chem. 267:11721–11728.

26. Tomavo, S., R. T. Schwarz, and J. F. Dubremetz. 1989. Evidence for glycosyl-phosphatidylinositol anchoring of Toxoplasma gondii major surface antigens.Mol. Cell. Biol. 9:4576–4580.

27. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer ofproteins from polyacrylamide gels to nitrocellulose sheets: procedure andsome applications. Proc. Natl. Acad. Sci. USA 76:4350–4354.

1460 GIRALDO ET AL. J. CLIN. MICROBIOL.

on January 17, 2021 by guesthttp://jcm

.asm.org/

Dow

nloaded from