MAJOR ARTICLE

Serotyping of Toxoplasma gondii infections in Humans Using Synthetic Peptides Jiang-Ti Kong,1,a Michael E. Grigg,1,a Lyle Uyetake,1 Stephen Parmley,2,b and John C. Boothroyd1 1

Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, and 2Department of Immunology and Infectious Diseases, Research Institute, Palo Alto Medical Foundation, Palo Alto, California

To determine whether the characteristics of disease due to Toxoplasma gondii (toxoplasmosis) are dependent on the infecting strain, we have developed an enzyme-linked immunosorbent assay for typing strains that uses infection serum reacted against polymorphic peptides derived from Toxoplasma antigens SAG2A, GRA3, GRA6, and GRA7. Pilot studies with infected mice established the validity of the approach, which was then tested with human serum. In 8 patients who had Sabin-Feldman dye test titers 164 and for whom the infecting strain type was known, the peptides correctly distinguished type II from non–type II infections. ELISA analysis of a second group of 10 infected pregnant women from whom the parasite strain had not been isolated gave a clear prediction of the strain type causing infection. This method should allow statistically significant data to be obtained about whether different strain types cause disease with different characteristics. The zoonotic pathogen Toxoplasma gondii is highly prevalent in nature. Infection with this protozoan parasite results in a wide spectrum of disease states, ranging from a chronic, asymptomatic infection to lymphadenopathy, chorioretinitis, and even fatal encephalitis [1–3]. In addition, transplacental transmission can lead to hydrocephalus, blindness, deafness, and mental retardation in newborns, and essentially all children with congenital toxoplasmosis, including those who receive treatment for the disease, go on to develop retinal disease [4, 5]. Latent Toxoplasma infection can also reactivate in immunocompromised patients, leading to

Received 21 October 2002; accepted 11 December 2002; electronically published 15 April 2003. Financial support: National Institutes of Health (grant AI21423); County of Alameda District Attorney’s Office (CA); Medical Scholars Program Grant from the Stanford University School of Medicine (to J.-T.K.). a

J.-T.K. and M.E.G. contributed equally to this work.

b

Present affiliation: Maxygen, Redwood City, California.

Reprints or correspondence: Dr. John C. Boothroyd, Dept. of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA 94305 (john [email protected]). The Journal of Infectious Diseases 2003; 187:1484–95  2003 by the Infectious Diseases Society of America. All rights reserved. 0022-1899/2003/18709-0017$15.00

1484 • JID 2003:187 (1 May) • Kong et al.

focal and often fatal encephalitis in 25%–50% of untreated patients [6, 7]. Given the variable and sometimes severe consequences of Toxoplasma infection, it is of clinical and epidemiological importance to identify the factors that contribute to these different outcomes. Strain type is one of the key factors suspected to play a role in determining the outcome of Toxoplasma infection. Approximately 90% of Toxoplasma isolates that have been reported are classified into 1 of just 3 clonal lineages or strain types [8–10]. One of these lineages (type I) is highly virulent in murine infections [11], whereas type II is relatively avirulent and type III is of intermediate virulence in mice. It has recently been shown that polymorphism within the 3 dominant types, as well as many of the rarer types, ranges from 1% to 3% at the amino acid level and is typically limited to just 2 allelic classes, regardless of the genetic locus [12]. The 3 major genotypes, as well as many of the rarer types, thus appear to be distinct assortments of these 2 alleles. Although the “recombinant” strains are much rarer, representing only 5%–10% of isolates so far characterized, they were responsible for 140% of atypical, severe human ocular toxoplasmosis in a small series of 18 patients [13, 14]. Given that the vast majority of reported human isolates (194%) have been type I, II,

or III [10], the unusual abundance of “recombinant” strains in severe ocular disease strongly suggests that a connection exists between strain type and disease. This, together with the apparently simple population structure of Toxoplasma, the direct correlation between type I strains and virulence in mice, and the overrepresentation of type I parasites in congenital infections in some geographic regions [15], all suggest that strain type is a key determining factor in disease presentation and severity in humans. To further explore the correlation between strain type and human disease, it is necessary to determine the strains causing infections that span the full clinical spectrum of presentation. This has not previously been feasible because of the limitations of current strategies, which typically are focused on interstrain differences at either the genomic [16–18] or the protein [9] level. As a result, these strategies require either the isolation of sufficient parasite DNA or actual organisms to obtain protein. To date, parasites generally can be obtained only from symptomatic patients, and then with extreme difficulty and significant risk, through biopsy. Hence, the currently available assays have been able to identify the strains present in only a small proportion of human infections that principally represent extreme disease scenarios. What is needed is a rapid, highly sensitive, and relatively noninvasive means of identifying strain type in any disease state. Serologic testing using peptides was chosen as a promising method of accomplishing this for 2 reasons. First, Toxoplasma stimulates a strong and persistent humoral immune response in every host: antibodies to parasite proteins remain at high titers for the life of the host and are present in patients with essentially all clinical manifestations of Toxoplasma infection, ranging from ocular to congenital and, importantly, to asymptomatic cases. Second, evidence exists that the humoral response to Toxoplasma is strain specific. For example, 5 of 5 mouse monoclonal antibodies (MAbs) to SAG2A isolated after a natural infection were allele specific (i.e., the antibodies recognize the allele expressed by type I and III but not type II strains, even though all 3 types express approximately equal levels of the SAG2A protein) [19]. Hence, polymorphic epitopes appear to be highly immunogenic and possibly even immunodominant. We have exploited these facts to develop a serologic testing– based strain typing strategy that takes advantage of the differences in the antibody profiles generated by infection with different strains of parasites. Synthetic peptides from polymorphic antigens were screened serologically to identify those that discriminate between infection with different strain types. This method should make it possible to substantially expand the number of patients and the range of disease in which the strain type can be determined and should, eventually, lead to a de-

finitive answer to the question of how strain type influences disease outcome.

METHODS Selection of antigens and peptide sequences. Nucleotide sequences from archetypal type I (RH), II (Me49 and Prugniaud), and III (VEG and CEP) strains were obtained from GenBank, from Toxoplasma expressed sequence tags, or by direct sequencing for the following 14 Toxoplasma immunogens: dense granule proteins GRA1 [20], GRA3 [21], GRA4 [22], GRA6 [23], and GRA7 [24, 25]; NTPase I and III [26, 27]; surface antigens SAG1, SAG2, SAG3, SAG4, BSR4, and SRS2 [13, 16, 28]; and the rhoptry protein ROP1. Polymorphic regions were identified after alignment of all synthesized alleles and allele-specific peptides. Polymorphic peptides were selected on the basis of the following criteria: substantial content of charged and/or hydrophilic amino acids, a predicted a-helical structure, the presence of proline residues, and the presence of short regions of typically 8–12 aa that are bounded by 2 cysteine residues. Peptides. A total of 60 peptides were synthesized. Ten were determined empirically to detect strain-specific antibodies in serum from T. gondii–infected animals and/or humans. All peptide sequences are presented in tables 1–3. Two peptides, ctrl1 (CEVVHDYRLFNP) and ctrl-2 (CENFSPHFVGLD), were synthesized as negative controls. They were derived by randomizing the sequences of the 6-I/III strain–specific peptides. All peptides were made by solid-phase methods with an automated peptide synthesizer, and their purity was examined by mass spectroscopy at the Stanford Protein and Nucleotide facility (Stanford, CA). Conjugation to carrier protein. A cysteine residue was added to either the C or the N terminus of each peptide to enable coupling to the carrier protein keyhole limpet hemocyanin (KLH; Pierce) by maleimide chemistry, according to specifications supplied by the manufacturer. Peptides were diluted in 0.1% acetic acid, water, 10% acetonitrile, or 20% dimethyl sulfoxide to a final concentration of 10 mg/mL immediately before coupling. The dissolution conditions for all 60 peptides can be supplied on request. For conjugation, 50 mL of peptide was diluted in 150 mL of 0.1 mM EDTA in PBS. Fifty microliters of water-reconstituted KLH (10 mg/mL) was then added to the mixture, and the mixture was incubated at room temperature for 3 h. The product was dialyzed extensively in PBS (3 buffer exchanges). Dialyzed, coupled peptide was aliquoted and stored at ⫺20C. Undiluted, coupled peptide stored at 4C was stable for at least 3 months. Serum from T. gondii–infected mice. Thirty outbred Swiss Webster mice (Simenson Laboratories) were orally inoculated with ∼10 bradyzoite cysts from the following T. gondii strains:

T. gondii Serotyping with Synthetic Peptides • JID 2003:187 (1 May) • 1485

Table 1. Allele-specific peptides, derived from Toxoplasma immunogens, that are strongly recognized in an allele-specific manner by most (if not all) serum samples from Toxoplasma gondii–infected animals and/or humans. Genetic locus

Name

Peptide sequence

GRA6-I/III-220

6-I/III

CLHPERVNVFDY

dGRAS6-I/111-220

(d)6-1/111

CLHPERVNVFD

dGRAS6-III-220(9)

—

GRA6-II-214

6-II

dGRA6-II-214

(d)6-11

CLHPGSVNEFDF

dGRA6-II-216(9)

—

dGRA6-II-214(9)

—

GRA3-I/III-28

CLHPERVNV CLHPGSVNEFD C——PGSVNEFDF CLHPGSVNE

3-I/III

ADQPENHQALAEPVC

GRA3-II-28

3-II

ADQPGNHQALAEPVC

GRA7-II-225

7-II

CVPESGKDGEDARQ

dGRA7-II-225

(d)7-II

CVPESGKDGEDA

GRA7-III-225

7-III

CVPESGEDREDARQ

dGRA7-III-225

(d)7-III

CVPESGEDREDA

SAG2A-I/III-131

—

PAGRNND—GSSAPTPKC

dSAG2A-I/III-131(13)

—

PAGRNND—GSSAPC

dSAG2A-I/III-134(10)

2-I/III

SAG2A-II-131 dSAG2A-II-134(11)

— 2-II

RNND—GSSAPC PAGRNNDGGSSAPTPKC RNNDGGSSAPC

NOTE. “I,” “II,” and “III” refer to the archetypal strain that includes the peptide sequence, and the nos. following these indicate position in the coding sequence of the first amino acid of each peptide. The cysteine (C) at the N or C terminus was present for coupling purposes and was not present in the protein. Bold type indicates polymorphic sites.

type I, CAST; type II, Me49 and FORT; and type III, C56 and VEG. Because the RH strain does not readily form cysts, mice inoculated with this strain were injected intraperitoneally with 100 tachyzoites. Mice inoculated with RH, CAST, C56, or VEG were given sulfadiazine (200 mg/kg) in their water from postinfection day 3 or 4 through day 10, to prevent death during the acute stages of infection. Two months after infection, mice were bled out. Serum samples from 5 uninfected mice were used as negative controls. Serum from T. gondii–infected humans. Typed serum samples (samples PA1–PA10, PA21, and JSR14) from 12 patients for whom the T. gondii strain responsible for infection had been determined by polymerase chain reaction (PCR)/ restriction fragment–length polymorphism analysis (RFLP) were obtained from Dr. Jack Remington at the Palo Alto Medical Foundation (Palo Alto, CA). Seven of the serum samples were from patients with AIDS who had focal encephalitis. Three were from T. gondii–infected pregnant women carrying infected fetuses, 1 was from a patient with chorioretinitis, and 1 was from a patient with both lymphadenopathy and chorioretinitis (table 4). The authors were blinded to strain genotype during the course of experimentation. Untyped serum samples (samples PA11–PA20) from 10 preg1486 • JID 2003:187 (1 May) • Kong et al.

nant women who were human immunodeficiency virus (HIV) negative and who had positive Sabin-Feldman dye test titers against Toxoplasma lysates were provided by the Palo Alto Medical Foundation. No attempt had been made to isolate or directly type the strains responsible for these infections, and this was not possible with the material available at the time of these experiments (serum only). Serum samples from 25 subjects who had negative dye test titers were obtained from the Palo Alto Medical Foundation and used as negative controls. Each serum sample was tested against the same peptide panel as were samples PA1–PA21, to establish the full range of background values for each of our test peptides. ELISA protocol. Peptides coupled to KLH were diluted to 10 mg/mL in 0.1 M carbonate buffer at pH 8.5. As a positive control for each serum sample, 1% Nonidet P-40 (Calbiochem; NP-40) lysates of Toxoplasma were prepared and diluted to ∼1000 parasite equivalents/mL. Fifty microliters (i.e., 0.5 mg) of each peptide or of the total Toxoplasma lysate (i.e., ∼50 parasites) was loaded into each well of a polystyrene ELISA plate (Falcon flat-bottom ELISA plate) for coating overnight at 4C. Each well was then blocked with 200 mL of a 2% casein solution in PBS with 0.01% thimerosal, and plates were held for 2 h at room temperature. Next, 50 mL of diluted infection serum was added to each well, and plates were incubated at Table 2. Allele-specific peptides, derived from Toxoplasma immunogens, that are recognized by some, but not all, serum samples from Toxoplasma gondii–infected animals and/or humans. Genetic locus

Peptide sequence

GRA6-I/III-175

CGRRSPPERSGDGG

GRA6-II-175

CGRRSPQERSGGGG

GRA6-I/III-199

CGNEGRGYGGRGEG

GRA6-I-207

CGRGEGGAEDDRRP

GRA6-II-202

CGRGEGG—EDDRRPL

ROP1-I-85

PVRGPDQVPAC

ROP1-II/III-85

PVRDPRQVPGRGEC

ROP1-II/111-359

CTRVRGALR—GRGR

GRA7-I-163

ELTEEQQRGDEPLC

dGRA7-I-162

CPELTEEQQRG

dGRA7-I-164

C——LTEEQQRG

GRA7-III-163

ELTEQQQTGDEPLC

GRA7-II/III-162

CPELTEQQQTG

GRA7-I/II-215

CSRQPALEQEVPES

GRA7-III-215

CSRQPAPEHEVPES

GRA7II-225

CVPESGEDGEDARQ

SAG3-II-49

GNSRRKITYC

NOTE. “I,” “II,” and “III” refer to the archetypal strain that includes the peptide sequence, and the nos. following these indicate position in the coding sequence of the first amino acid of each peptide. The cysteine (C) at the N or C terminus was present for coupling purposes and was not present in the protein. Bold type indicates polymorphic sites.

Table 3. Allele-specific peptides, derived from Toxoplasma immunogens, that are not recognized by serum samples from Toxoplasma gondii–infected animals and/or humans. Genetic locus

Name

Peptide sequence

NTP3-I-99

—

SIQLIGAGKRFAGLRC

NTP1-II/III-99

—

SIRLIREGKRFTGLRC

NTP1-I/II/III-485

—

CAPMFITGREMLASIDT

NTP3-I-485

—

CAPMIVTGGGMLAAINT

BSR4-I/II-155

—

KVNEQREESNKSQKC

BSR4-I/II-336

—

PKKDKESGTETGAPC

SAG1-174

—

SYGADSTLGPVKLC

SAG1-I-244

—

SDKGATLTIKKEAFPC

SAG2A-I/III-88

—

PGAVLTAKVQQPAKGPC

SAG3-II-120

—

CHIDAKDQDD

SAG4A-I/II-84

—

CKDEPVELAAL

SRS1-I-50

—

SMTSPLLTWDGNKVTC

SRS2-I-53

—

GPPYRYEPEKFTC

GRA1-I-92

—

CSYSEVGNVNVEE

GRA1-III-92

—

CSYSEVGDVNVEE

GRA1-I/III-159

—

CQDEMKVIDDVQQ

GRA1-II-159

—

CQDEMNVIDDVQQ

GRA1-II-159

—

SAAIGGRMVSRTLRDNIPGC

GRA3-I/III-189

—

RRKPKDEGAGVDKAC

GRA4-I-232

—

CTEDSGLTGVKDSSS

ROP1-I-131

—

NSEDDDTFHDAC

ROP1-II/III-131

—

NSEDD—TFHDAC

ROP1-II/III-181

—

QELPPPNAQELC

(r: 6-I/III)

ctrl-1

CEVVHDYRLFNP

(r: 6-II)

ctrl-2

CENFSPHFVGLD

NOTE. “I,” “II,” and “III” refer to the archetypal strain that includes the peptide sequence, and the nos. following these indicate position in the coding sequence of the first amino acid of each peptide. The cysteine (C) at the N or C terminus was present for coupling purposes and was not present in the protein. Bold type indicates polymorphic sites.

room temperature for 3 h. Each serum sample from an infected patient was titered into the linear range against Toxoplasma lysate to determine the appropriate dilution for screening against the KLH-coupled peptides. Each ELISA plate was washed 4 times with 0.1% Tween 20 in PBS before incubation with horseradish peroxidase–conjugated detection antibody, consisting of either an MAb against human IgG (Pharmingen) or a total goat IgG preparation against mouse IgG (Sigma). After incubation at room temperature for 1–2 h, the secondary antibodies were aspirated, and the plates were washed with 0.1% Tween 20 4 times and with PBS twice. Plates were developed with 150 mL of the ABTS reagent (Kirkegaard and Perry Laboratories), and the absorbance was measured at 405 nm after 120 min. Serum samples from healthy mice and a nonimmunogenic peptide derived from the signal sequence of the Toxoplasma SAG4 gene and coupled to KLH served as neg-

ative controls to establish background reactivity in developing this ELISA.

RESULTS The majority of T. gondii isolates are genotypically classified into 1 of 3 clonal lineages, referred to as “type I,” “type II,” or “type III.” Each of these 3 archetypes can be discriminated serologically by strain-specific MAbs derived from natural infections that recognize the polymorphic antigens BSR4, GRA3, GRA4, SAG2A, and SAG4 (data not shown). One plausible explanation for this is that a selective pressure operates at the immunological level during infection and that polymorphic regions are immunogenic and perhaps even immunodominant. Because synthetic peptides can be recognized by serum samples obtained during acute infection [29], we investigated whether these strain-specific MAbs recognized denatured Toxoplasma antigen and, thus, were likely to recognize polymorphic linear epitopes within their cognate antigen. Western blot analysis against denatured lysate derived from each of the 3 archetypal lineages showed that 2 MAbs, 5A6 (specific for SAG2A) and 2H11 (specific for GRA3), readily recognized polymorphic linear epitopes in a strain-specific manner (figure 1A). In effect, types I and III shared a common epitope that was absent from type II strains. Western blot analysis with the SAG1 MAb DG52 served as a loading control to ensure that equivalent amounts of parasite lysate from each type were present (figure 1A). Polymorphism at SAG2A and GRA3 is between 1% and 5% and is clustered at discrete sites within each coding sequence. Sequence alignments of archetypal alleles showed that the majority of polymorphic sites were contained within 3 short regions (∼15 aa in length) for both loci. Allele-specific peptides spanning these polymorphic regions were synthesized and screened by ELISA for MAb reactivity. Two polymorphic peptides were identified, 2-I/III and 3-I/III, as specific for MAbs 5A6 (SAG2A) and 2H11 (GRA3), respectively, and coupling of these peptides to the carrier protein KLH was determined to be optimal for their recognition (figure 1C; data not shown). Each peptide possessed a unique polymorphism (figure 1C), and neither MAb recognized the other allele peptide by ELISA (data not shown). To determine whether the 2 polymorphic peptides were sufficiently antigenic to be recognized by strain-specific antibodies in serum from experimentally infected animals, outbred Swiss Webster mice were infected naturally by the oral route with archetypal type I, II, and III strains to generate the infection serum required to answer this question (see Methods). This mimics at least 1 route of infection and ensures that the immune response is as close to the natural situation as possible. All serum samples were titered into the linear range against Toxoplasma antigen derived from the 3 archetypal lineages to ensure maximum sensitivity for the detection of polymorphic T. gondii Serotyping with Synthetic Peptides • JID 2003:187 (1 May) • 1487

Table 4.

Characteristics and results of serotyping, using synthetic peptides, for Toxoplasma gondii strains from 3 groups of patients. b

ELISA result Serum Predicted Dye PCR-RFLP Clinical Date of a sample 6-I/III (d)6-I/III 6-II (d)6-II 7-II (d)7-II 7-III (d)7-III type test titer typing results Strain characteristics presentation PAC1

0.9

1.0

0.8

0.9

0.8

0.8

0.9

0.8

NA

PA1

1.7

1.0

1.0

1.0

1.8

0.8

1.6

0.9

I/III

NA 32

NA I

PA2

0.9

1.0

3.2

2.7

0.9

3.3

1.0

0.8

II

512

II

CS

LAN, EYE

Apr 1994

PA3

1.0

1.0

1.2

1.2

1.0

0.9

1.0

0.9

?

32

II

DAG

AIDS, ENC

Dec 1982

PA4

0.9

0.9

1.5

1.6

1.5

0.9

1.3

0.8

II

128

II

FORT

AIDS

May 1985

PA5

0.9

1.0

1.0

1.0

1.2

1.0

1.1

1.0

?

32

II

MOY

AIDS, ENC

Mar 1988

PA6

2.2

1.2

4.4

2.4

4.3

1.0

1.0

0.8

II

8000

II

PED

PREG

Oct 1986

PA7

1.1

1.2

1.2

1.2

2.0

0.8

2.7

1.2

?

64

III

POE

AIDS, ABS

Dec 1987

PA8

1.0

1.1

1.0

1.2

1.2

0.9

1.2

0.9

?

64

II/III

SOU

AIDS, ENC

Aug 1985

PA9

1.1

1.1

8.0

2.2

7.5

0.9

2.2

1.1

II

4096

II

TIM

PREG

Jan 1993

PA10

1.7

1.0

1.1

1.0

1.1

0.8

1.3

0.9

I/III

128

III

VEG

AIDS

Nov 1988

PA21

1.7

1.0

1.0

1.0

1.8

0.8

1.6

0.9

I/III

2048

I

—

CONG

Mar 1995

PA11

1.3

1.1

6.6

3.5

10.1

0.8

2.3

0.8

II

8000

ND

—

—

—

PA12

10.1

1.1

1.0

1.1

1.3

1.2

1.4

1.1

I/III

4096

ND

—

—

—

PA13

6.4

3.1

6.4

6.4

7.8

1.3

2.1

1.1

Mixed?

4096

ND

—

—

—

PA14

6.2

2.3

1.3

1.0

1.3

0.9

1.4

0.9

I/III

16,000

ND

—

—

—

PA15

0.9

1.1

1.4

1.3

1.7

1.0

1.1

0.9

II

4096

ND

—

—

—

PA16

1.4

1.0

2.9

1.9

4.1

0.9

1.1

0.9

II

8000

ND

—

—

—

PA17

1.1

1.2

2.5

1.5

3.0

1.2

1.3

0.9

II

8000

ND

—

—

—

PA18

1.4

1.0

3.8

1.1

3.9

1.2

1.0

1.2

II

8000

ND

—

—

—

PA19

1.4

1.1

1.0

1.0

1.2

1.1

1.2

0.9

I/III

4096

ND

—

—

—

PA20

6.8

1.0

0.9

1.0

1.2

0.8

1.4

0.8

I/III

32,000

ND

—

—

—

JSR02

0.9

1.1

0.8

1.0

0.8

1.0

0.9

0.9

?

128

IV

NA

EYE

Oct 1999

JSR05

1.4

1.3

1.2

1.3

1.2

1.1

1.1

1.1

?

2048

IV

NA

EYE

Aug 1999

JSR14

5.6

1.0

1.1

0.9

1.0

0.9

1.8

1.6

I/III

512

I

NA

EYE

Aug 2000

CAST

AIDS

Jul 1988

NOTE. All serum samples were provided by Dr. J. Remington (Palo Alto Medical Foundation, Palo Alto, CA) as a blinded set. ABS, brain abscess; CONG, congenital toxoplasmosis; ENC, encephalitis; EYE, chorioretinitis; HIV, human immunodeficiency virus; LAN, lymphadenopathy; NA, not applicable; ND, not done; PCR, polymerase chain reaction; PREG, pregnancy; RFLP, restriction fragment–length polymorphism analysis; ?, no clear prediction. a

Samples PA1–PA10 and PA21 are from patients (principally patients with AIDS) from whom parasites were recovered and genotyped by PCR-RFLP. Samples PA11–PA20 are from HIV-negative women infected with T. gondii of unknown strain type. Samples JSR02, JSR05, and JSR14 are from patients with ocular toxoplasmosis. Sample PAC1 is from an uninfected individual and is representative of the profiles of 24 other serum samples from uninfected individuals. b Serum samples were tested against a panel of 8 strain-typing peptides. Data are the means of normalized data from either 3 (JSR02, JSR05, and JSR14) or 9 (PA1–PA21) independent experiments. A value of ⭓1.4 was considered to be significant; based on this, a clear prediction of the strain type responsible for the infection was ascertained for 17 of 24 serum samples from patients infected with T. gondii.

epitopes by ELISA. Significantly, no strain-specific differences against total Toxoplasma antigen of one type were readily detectable with any of the serum samples, which indicates that strain-specific antibodies do not predominate (figure 1B). Serum samples were then tested against the SAG2A and GRA3 polymorphic peptides identified by the MAbs. Figure 1C shows that both peptides were readily recognized by infection serum in an allele-specific manner. Recognition was specific; none of the mice antisera reacted with the type II version of the SAG2A and GRA3 peptides (data not shown). These results gave us the first indication that serologic strain typing of toxoplasmosis using synthetic peptides is feasible. To develop additional peptides for screening, only Toxoplasma antigens determined empirically to be immunogenic in all hu1488 • JID 2003:187 (1 May) • Kong et al.

mans were targeted. This list was based on our own preliminary data and on those in the literature [30] (see Methods). In total, 60 allele-specific peptides were synthesized after alignment of sequences from the 3 archetypal lineages identified all polymorphic sites in these 14 Toxoplasma immunogens (tables 1–3). To screen for candidate strain-typing reagents, each peptide was tested with the panel of serum samples from infected mice described above. Any peptide detected by the majority of serum samples from infected mice was selected for further analysis. The principle was to identify all peptides recognized serologically, regardless of their allele specificity, although a special emphasis was given to peptides that reacted specifically with serum from mice infected by any given type (i.e., types I, II, and/or III). After the first screen, 6 peptides were selected for further analysis: 2

Figure 1. Serologic typing of Toxoplasma gondii using synthetic peptides identified by polymorphic strain-specific monoclonal antibodies (MAbs). A, Western blot analysis against denatured (boiled and reduced) Toxoplasma lysate derived from either RH (type I), Me49 (type II), or CEP (type III) parasites and separated by SDS-PAGE. MAbs 5A6, 2H11, and DG52 (the latter MAb against boiled, not reduced, lysate) were used to specifically identify SAG2A, GRA3, and SAG1, respectively. B, Serum samples from T. gondii–infected mice were tested at various dilutions to determine that the assay was in the linear range, using type II strain (Me49) parasite lysate coated onto an ELISA plate at 50 parasite equivalents/well. NC, negative control (serum from healthy, uninfected mice). C, ELISA identifying the polymorphic residues responsible for allele specificity of the 2 strain-specific MAbs 5A6 (SAG2A) and 2H11 (GRA3). Polymorphisms at SAG2A (2-I/III vs. 2-II) and GRA3 (3-1/III vs. 3-II) are in bold type. Bar graphs show specific recognition of the type I/III peptides by murine antibodies present in serum samples from patients infected with type I and III strains (CAST, RH, and VEG and C56) but not type II strains (ME49 and FORT). An OD ⭓0.5 was considered to be specific.

from GRA6, 3 from GRA7, and 1 from ROP1 (data not shown; see below). For each of these 6 immunogenic sites, peptides corresponding to both alleles were synthesized and subjected to delimitation or truncation in an effort to eliminate potential cross-reactive epitopes and were rescreened for their ability to detect strain-specific antibodies in infection serum. Thirteen serum samples from mice, 3 of which came from type I infections, 4 from type II infections, and 6 from type III infections, were tested against the peptides. As shown in figure 2A, the GRA6 peptide at the C terminus of the protein (designated “6-I/III,” because type I and III strains share an identical peptide sequence in this region, whereas type II strains possess a substantially different allele) was readily detected by serum from infected mice in an allele-specific manner (i.e., by serum samples from mice infected with type I or III but not II). This

peptide exhibited a specificity of 100%, with 89% sensitivity (there was only 1 false-negative result). The possibility that the sample with the false-negative result had low antibody titers was excluded, because that sample reacted normally to total Toxoplasma lysate (data not shown). Serum samples from mice infected with type II strains were essentially indistinguishable from uninfected mouse serum, in terms of their reactivity to this peptide. No serum samples reacted to a KLH-coupled control peptide derived by randomizing the 6-I/III peptide sequence (figure 2A). When the serum samples were tested against the 6-II peptide, which represents the other allele of GRA6, all serum samples from mice infected with type II strains yielded a strongly positive signal (figure 2B). Unlike the 6-I/III peptide, however, the 6-II peptide was less specific, and 1 sample from a type I inT. gondii Serotyping with Synthetic Peptides • JID 2003:187 (1 May) • 1489

Figure 2. Serologic discrimination of type II from non–type II Toxoplasma gondii infections, using polymorphic allele-specific peptides at GRA6. A, Allele-specific recognition of a polymorphic GRA6 peptide (6-I/III) by antibodies present in serum from T. gondii–infected mice that was derived from 2 type I (RH and CAST) and 2 type III (C56 and VEG) but not from 2 type II (Me49 and FORT) infections. Each pair of values represents an individual mouse (2 for RH and 3 each for ME49, C56, and VEG). B, Recognition of the GRA6 peptide (6-II) by antibodies present in serum from T. gondii–infected serum. C, Delimitation of 6-II peptide ([d]6-II) to eliminate false-positive findings of cross-reactivity by antibodies present in serum from mice infected with type I and III strains. An OD of ⭓0.5 was considered to be specific.

fection and 2 samples from type III infections also recognized it. One possible explanation for this would be that there is a common epitope present in the nonpolymorphic amino acids of this peptide and that this epitope either is not naturally 1490 • JID 2003:187 (1 May) • Kong et al.

immunogenic in infections with type II strains or is masked in the type I/III peptide. To address this possibility, delimited versions of the 6-II peptide were synthesized. Although truncations at the N terminus did not improve the specificity of the original

peptide (data not shown), deletion of the C-terminal amino acid eliminated all false-positive results, as well as one of the true-positive results (figure 2C). The pair of allele-specific complementary peptides at the C terminus of GRA7 likewise exhibited high sensitivity (80%) and high allele specificity (for 7-II, 90%; for 7-III, 70%) for the 13 samples from infected mice (data not shown). Delimitation of these 2 peptides, in which the 2 C-terminal amino acids were removed, resulted in no false-positive results but a significant number of false-negative results (sensitivity was only 10% and 30% for [d]7-II and [d]7-III, respectively). Therefore, to increase the sensitivity of the assay, all 8 GRA6/7 peptides were included. Together with the allele-specific peptides identified at SAG2A and GRA3, the ELISA reliably discriminated type II from non–type II infections in mice. To determine whether the strain-specific peptides can serologically identify the parasite genotype causing infection in humans as well as in animals, a panel of 12 serum samples from T. gondii–infected humans (samples PA1–PA10, PA21, and JSR14) were obtained for which the parasite strain causing infection, or at least present at the time the sample was obtained, was known. Throughout these analyses, we were blinded to information about the strain types, clinical presentations, and dye test titers of these 12 samples. As with the mouse serum, to achieve maximal sensitivity, the human serum samples were titered into the linear range against total Toxoplasma lysate. To measure the reliability and reproducibility of the serotyping assay, 25 serum samples from uninfected individuals were tested against the panel of 10 strain-specific peptides and 3 control peptides (2 of which were derived by randomizing the sequence of peptide 6-I/III) to determine the mean and SD of what is empirically background reactivity. This was done to establish the level of response that was true negative. Twelve typed infection serum samples were likewise tested against the control peptides, and, for each sample, the mean optical density (OD) for a given control peptide, when divided by the mean OD for the 25 samples from uninfected individuals tested against the same control peptide, yielded a value of ∼1 (data not shown). This result established that the data could be normalized for each infection serum sample by dividing the OD value obtained for any of the 10 serotyping peptides by the mean of the OD readings for the 3 control peptides. For those infection samples that did not react with any given serotyping peptide, the expected value after normalization should be close to 1. Thus, the reproducibility of results with the serotyping reagents across many experiments can be assessed. The 12 typed serum samples from T. gondii–infected individuals were tested against the panel of 10 serotyping peptides. Unfortunately, the GRA3 and SAG2A peptides did not react with sufficient sensitivity to be useful for serotyping in human serum (data not shown). The GRA6 and GRA7 peptides, however,

worked extremely well. Figure 3 shows the SD of the means of normalized data for 3 independent experiments for 2 of the 12 serum samples, samples JSR14 and PA2, which are from patients apparently infected with type I and II strains, respectively. JSR14 showed strong reactivity to the 6-I/III and both the 7-III and (d)7-III peptides but did not react to any other peptide, as indicated by a normalized value of ∼1. This reactivity pattern is consistent with infection by a type I or III parasite (type I and III strains encode nearly identical GRA6 and GRA7 proteins) but not a type II strain. In contrast, PA2 showed strong reactivity with the 6-II and (d)6-II peptides, as well as the (d)7-II peptide, which indicates a type II infection. The fact that both JSR14 and PA2 reacted to allele-specific peptides derived from GRA6 and GRA7 implies that both loci are immunogenic in the individuals from whom these samples were obtained. Table 4 lists the mean of the normalized data from 9 independent ELISA experiments performed on serum samples PA1–PA10 and PA21 with the 8 serotyping peptides and predicts the strain type responsible for infection. A value of 1.0 indicates a lack of specific reactivity with the test peptide; each OD reading was normalized to the mean reactivity of each serum sample to, at a minimum, 2 irrelevant control peptides. Results with the control serum sample PAC1, from an uninfected individual, which was included in all 9 data sets, ranged in value from 0.8 to 1.0 for any given serotyping peptide, which validates the normalization approach. During the interpretation of the data, we were blind to any identifying or clinical data available for the serum samples from T. gondii–infected individuals. The results show that samples PA2, PA4, PA6, and PA9 had OD profiles that were consistent with infection by type II parasites, whereas PA1, PA10, and PA21 showed infection by non–type II strains, possibly type I or III strains. Specifically, PA2 and PA9 reacted positively to 3 of the 4 type II peptides and none of the type I/III peptides. Although PA6 reacted positively to 3 of the 4 type II peptides, it also reacted with 6-I/III. However, its reactivity to 6-I/III was weak, compared with its reactivity to 6-II, and no reaction was seen to the delimited version of this peptide. PA4 was more difficult to judge, but because (d)6-II has high predictive value in mice (100% specificity) and PA4 reacted more strongly with 7-II than with 7-III, we felt confident in our prediction that the patient from whom sample PA4 was obtained was infected by a type II parasite. The patients from whom samples PA1, PA10, and PA21 were obtained were determined to be infected by non–type II strains, primarily on the basis of positive reactivity to 6-I/III and lack of reactivity to 6-II. In murine infections in which strain type was known, a positive reaction to 6-I/III accompanied by a corresponding negative reaction to 6-II was absolutely predictive. Unfortunately, these 3 serum samples showed no discriminating reactivity against the GRA7 peptides, which indicates that these peptides need further optimization. Samples PA3, PA5, and PA8 T. gondii Serotyping with Synthetic Peptides • JID 2003:187 (1 May) • 1491

Figure 3. Serologic discrimination of Toxoplasma gondii strains causing human infection, using a panel of allele-specific peptides. ELISA showing SD of the means of normalized data for 3 independent experiments with 2 serum samples from T. gondii–infected individuals (samples JSR14 and PA2) reacted against a panel of 8 allele-specific peptides. JSR14 reacted specifically with 6-I/III and both 7-III and (d)7-III, which indicates that this patient was infected with a non–type II strain. PA2 reacted specifically with 6-II, (d)6-II, and (d)7-II, which is highly indicative of a type II strain infection.

gave no clear indication of the type responsible for infection; reactivity against all of the peptides tested was weak. The reaction of sample PA7 to 7-III appears to indicate that the patient was infected with a type III strain, but the high titer against 7II and lack of reactivity to the 6-I/III peptides preclude making a definitive judgment. Table 4 lists the dye test titers, clinical presentation, date of serum collection, and strain name and type assigned on the basis of PCR-RFLP genotyping for samples PA1–PA10 and PA21. In cases in which sufficient anti-peptide reactivity was detected, our serotyping prediction agreed with PCR-RFLP genotype designation, which supports the validity of this serologic testing–based genotyping assay. It is worth noting that the majority of these typed serum samples were from severely immunocompromised patients with extremely low dye test titers. It is perhaps not surprising, then, that the data do not allow a prediction of strain type for patients with dye test titers that just exceeded the 1:16 threshold for positivity (samples PA3, PA5, PA7, and PA8), which is well below the titers seen in patients who did not have AIDS. The majority of individuals infected with T. gondii, however, are healthy adults with intact immune systems. To evaluate the sensitivity of the assay further, 10 additional serum samples were obtained (PA11–PA20) from patients who were infected during pregnancy, were HIV negative, and had more usual dye test titers 1492 • JID 2003:187 (1 May) • Kong et al.

of ⭓4096. Analysis of this patient group, using only the GRA6 peptides, yielded a type prediction in most of the samples (5 type II infections and 3 non–type II infections). Because PA19 showed no reactivity at 6-II but had weak, highly reproducible reactivity with the 6-I/III peptide (SD, 0.1), it likely was from a non–type II infection (table 4). Only PA13 was difficult to predict, because it reacted strongly to all 4 GRA6 peptides. Even the inclusion of the GRA7 peptides did not resolve the genotype responsible for infection; instead, it raised the possibility that this patient was multiply infected with a mix of 2 different strains. Unfortunately, because no parasites were isolated from any of these 10 patients, it was not possible to assess the genotypes by any other means to confirm or refute our results. In a previous study, the parasite genotypes responsible for causing severe ocular toxoplasmosis were identified by PCRRFLP analysis of parasite DNA present in the vitreous fluid [13]. Serum samples from 3 of these patients (samples JSR02, JSR05, and JSR14) were obtained to further evaluate the panel of serotyping peptides. Only JSR14 had a clearly positive reaction that predicted infection by a non–type II genotype (table 4). Interestingly, JSR02 and JSR05, which were identified as having the relatively rare genotype IV, did not react with the panel of peptides. One possible explanation for this lack of reactivity could be that genotype IV strains have altogether different alleles at these genetic loci and therefore do not gen-

erate antibodies during infection that are capable of recognizing the GRA6 and GRA7 type II or I/III allele-specific peptides. Unfortunately, sufficient vitreous from JSR02 and JSR05 was not available to perform studies that would verify this.

DISCUSSION Despite its simple population structure, Toxoplasma is known to cause a wide spectrum of disease states in humans. To investigate why these 3 genotypes are so successful and to identify which of these 3 and/or whether less prevalent lines are associated with particular disease states in susceptible hosts, a serologic testing–based strain-typing ELISA was developed to distinguish infection caused by the different strains of Toxoplasma. Using synthetic polymorphic peptides derived from Toxoplasma immunogens, strain-specific antibodies present in serum from T. gondii–infected humans and animals were detected that distinguished type II from non–type II infections. These polymorphic peptides, derived from the Toxoplasma antigens SAG2A, GRA3, GRA6, and GRA7, accurately predicted the strain responsible for infection in mice. When screened against serum samples from infected humans, peptides from GRA6, in particular, correctly distinguished type II from non–type II infections, if sufficient antibodies to Toxoplasma were present (i.e., dye test titer 164). In a group of patients for whom no conventional strain typing methods were possible, serotyping at GRA6 suggested that the majority of infections had been caused by type II strains. This preponderance of type II strains is in agreement with other studies performed on human infections [8, 28]. It should now be possible to undertake retroactive and/or prospective studies on serum collected from infected, asymptomatic individuals, from patients involved in outbreaks of infection, or from patients with atypical clinical presentations (e.g., lymphadenopathy and ocular disease) to ascertain whether strain type is associated with disease outcome. Serologic strain typing using synthetic peptides is possible because of the limited number of strains that cause infection in most regions of the world. Recent studies have shown that the predominant or archetypal lineages are recently evolved, highly successful recombinants from a genetic mixing between 2 distinct ancestral lines [12]. Hence, for any given genetic locus, there are generally only 2 allelic types represented, and each archetypal line possesses a unique combination of those alleles. This means that, for any given immunogen, only 2 discriminatory peptides derived from polymorphic, immunodominant sites are required to determine which allele is carried by the infecting strain. Finding these antigens is difficult, however, because antibodies to strain- and/or stage-specific proteins do not apparently predominate in hosts infected with Toxoplasma, although several studies have successfully, albeit crudely, identified a number of such antigens by immunoblot

analyses [31–33]. At present, our assay is capable of distinguishing only between type II and non–type II strains. This is because type I and III strains possess identical alleles at SAG2A and GRA3 and are nearly identical at GRA6 and GRA7, and therefore our assay cannot discriminate between infection by these 2 archetypes. What is needed are more peptides at polymorphic antigens that possess different allele segregation patterns among the archetypal lines. Peptides specific for SAG1 and ROP1 would, for example (where types II and III share 1 allele that type I does not share), discriminate between infections caused by type I and III strains and perhaps also nonarchetypal lines. Our attempts to identify such peptides for these 2 proteins have not yet been successful. Thus, by using allele-specific peptides derived from multiple loci, the possibility of developing “signature” serologic profiles should facilitate the identification of strains associated with a particular disease in humans and animals and quite possibly help to identify new recombinant strains with nonarchetypal allelic combinations that cause these diseases. This will be particularly relevant in the analysis of serum samples from patients afflicted with ocular toxoplasmosis, among whom there is a high incidence of rare, recombinant strains [13]. In fact, serum samples from 2 such patients that possess the rare genotype IV strain (samples JSR02 and JSR05) did not react to any significant degree with either the GRA6 or the GRA7 peptide, which may indicate that other peptides will be needed for serologic identification of type IV infections (table 4). There are some limitations to the serologic assay, however. First, it is based on the assumption that the host has an intact humoral immune system. Patients with end-stage AIDS, for example, may not be able to produce sufficient strain-specific antibodies to reach the threshold of detection. Second, peptide sequences are most likely to be derived from either of the 2 predominant allelic types, which are present in the archetypal, as well as the majority of recombinant, strains so far investigated. There are, however, rare isolates with alleles that are significantly different from the 2 allelic types [12]. Without peptides specific for these isolates, the serologic assay will not be able to identify these rare, exotic strains. In addition, infections with these rare genotypes may induce entirely different humoral responses in susceptible hosts that are focused on sets of antigens that would not be detectable with our panel of peptides. However, these infections can be identified by incorporating total Toxoplasma lysate and/or recombinant antigens, such as the universal immunogen SAG1, in the panel. If a serum sample from an infected individual yields a very strong positive reaction to both Toxoplasma lysate and SAG1 but does not react to a large panel of peptides, these results could represent infection by a nonarchetypal and quite possibly exotic strain. The biological significance of the existence of these immunodominant and polymorphic antigenic sites is intriguing. Two T. gondii Serotyping with Synthetic Peptides • JID 2003:187 (1 May) • 1493

facts suggest that a powerful selection must be operating at the immune level at discrete sites within the Toxoplasma genome. First, 9 MAbs raised from natural infections recognize polymorphic epitopes [34, 35], and, second, a reciprocal pattern of allele-specific antibodies is present at highly immunogenic sites within GRA6 and GRA7. Such polymorphism may have been selected for and could exist as a critical parameter for the successful colonization of intermediate hosts infected with multiple Toxoplasma strains with different genotypes [18, 36, 37]. This would promote sexual reproduction, a process already demonstrated in Toxoplasma to rapidly generate recombinant genotypes capable of expanding into new ecological niches or to cause new disease [12]. The development of many diagnostic peptides to multiple antigens is therefore critical in discriminating between mixed-genotype infections (in which, for some peptides, both allelic forms would be recognized) and infections with new recombinant, nonarchetypal strains, in which a new “signature” reactivity with the panel of peptides would be identified. Ultimately, we envisage that the panel of peptides will reveal information beyond parasite genotype, such as host susceptibility to particular disease presentations or whether asexual or sexual transmission is responsible for the majority of Toxoplasma infections. In the latter case, it should be possible to identify an oocyst-specific immunogenic peptide or whole protein that would identify the stage of the parasite responsible for causing infection. Because serologic testing is fast, inexpensive, relatively noninvasive, and likely to be more sensitive than the conventional assays already in use (there is no need to isolate parasites, which often are not obtainable), this technique has the potential to become the method of choice for identifying the strain type causing Toxoplasma infections. This method also is not subject to biases resulting from animal passage or PCR that may decrease the accuracy of the conventional assays, and it can reveal cases in which one infection follows another (antibodies to both infecting strains should be present, whereas it may be possible to isolate parasites only from the second infection). Hence, this method should allow researchers to build large data sets and rapidly explore the degree to which strain type affects disease outcome. Ultimately, this will allow physicians to make a more accurate prognosis and implement the most appropriate treatment, on the basis of the genotype responsible for causing infection.

Acknowledgements

We thank Jack Remington and Gary Holland for much encouragement and provision of key materials used in the execution of this study. We also thank Kevin DeWaalt, Raymond Ramirez, Rima McLeod, and Marie Laure Darde´, for crucial advice and reagents. 1494 • JID 2003:187 (1 May) • Kong et al.

References 1. Remington JS. Toxoplasmosis in the adult. Bull NY Acad Med 1974; 50:211–27. 2. Suzuki Y, Wong SY, Grumet FC, et al. Evidence for genetic regulation of susceptibility to toxoplasmic encephalitis in AIDS patients. J Infect Dis 1996; 173:265–8. 3. Holland GN. Reconsidering the pathogenesis of ocular toxoplasmosis. Am J Ophthalmol 1999; 128:502–5. 4. Remington JS, McLeod R, Thulliez P, Desmonts G. Toxoplasmosis. In: Remington JS, Klein JO, eds. Infectious diseases of the fetus and newborn infant. 5th ed. Philadelphia: WB Saunders, 2001:205–346. 5. McLeod R, Boyer K, Roizen N, et al. The child with congenital toxoplasmosis. Curr Clin Top Infect Dis 2000; 20:189–208. 6. Luft BJ, Hafner R, Korzun AH, et al. Toxoplasmic encephalitis in patients with the acquired immunodeficiency syndrome. Members of the ACTG 077p/ANRS 009 Study Team. N Engl J Med 1993; 329: 995–1000. 7. Grant IH, Gold JW, Rosenblum M, Niedzwiecki D, Armstrong D. Toxoplasma gondii serology in HIV-infected patients: the development of central nervous system toxoplasmosis in AIDS. AIDS 1990; 4:519–21. 8. Darde ML, Bouteille B, Pestre-Alexandre M. Isoenzyme analysis of 35 Toxoplasma gondii isolates and the biological and epidemiological implications. J Parasitol 1992; 78:786–94. 9. Darde ML. Biodiversity in Toxoplasma gondii. Curr Top Microbiol Immunol 1996; 219:27–41. 10. Howe DK, Sibley LD. Toxoplasma gondii comprises three clonal lineages: correlation of parasite genotype with human disease. J Infect Dis 1995; 172:1561–6. 11. Sibley LD, Boothroyd JC. Virulent strains of Toxoplasma gondii comprise a single clonal lineage. Nature 1992; 359:82–5. 12. Grigg ME, Bonnefoy S, Hehl AB, Suzuki Y, Boothroyd JC. Success and virulence in Toxoplasma as the result of sexual recombination between two distinct ancestries. Science 2001; 294:161–5. 13. Grigg ME, Ganatra J, Boothroyd JC, Margolis TP. Unusual abundance of atypical strains associated with human ocular toxoplasmosis. J Infect Dis 2001; 184:633–9. 14. Boothroyd JC, Grigg ME. Population biology of Toxoplasma gondii and its relevance to human infection: do different strains cause different disease? Curr Opin Microbiol 2002; 5:438–42. 15. Fuentes I, Rubio JM, Ramirez C, Alvar J. Genotypic characterization of Toxoplasma gondii strains associated with human toxoplasmosis in Spain: direct analysis from clinical samples. J Clin Microbiol 2001; 39: 1566–70. 16. Grigg ME, Boothroyd JC. Rapid identification of virulent type I strains of the protozoan pathogen Toxoplasma gondii by PCR–restriction fragment length polymorphism analysis at the B1 gene. J Clin Microbiol 2001; 39:398–400. 17. Pastinen T, Raitio M, Lindroos K, Tainola P, Peltonen L, Syvanen AC. A system for specific, high-throughput genotyping by allele-specific primer extension on microarrays. Genome Res 2000; 10:1031–42. 18. Ajzenberg D, Banuls AL, Tibayrenc M, Darde ML. Microsatellite analysis of Toxoplasma gondii shows considerable polymorphism structured into two main clonal groups. Int J Parasitol 2002; 32:27–38. 19. Parmley SF, Gross U, Sucharczuk A, Windeck T, Sgarlato GD, Remington JS. Two alleles of the gene encoding surface antigen P22 in 25 strains of Toxoplasma gondii. J Parasitol 1994; 80:293–301. 20. Beghetto E, Pucci A, Minenkova O, et al. Identification of a human immunodominant B-cell epitope within the GRA1 antigen of Toxoplasma gondii by phage display of cDNA libraries. Int J Parasitol 2001; 31:1659–68. 21. Robben J, Hertveldt K, Bosmans E, Volckaert G. Selection and identification of dense granule antigen GRA3 by Toxoplasma gondii whole genome phage display. J Biol Chem 2002; 277:17544–7. 22. Mevelec MN, Mercereau-Puijalon O, Buzoni-Gatel D, et al. Mapping of B epitopes in GRA4, a dense granule antigen of Toxoplasma gondii

23.

24.

25.

26.

27.

28.

29.

and protection studies using recombinant proteins administered by the oral route. Parasite Immunol 1998; 20:183–95. Fazaeli A, Carter PE, Darde ML, Pennington TH. Molecular typing of Toxoplasma gondii strains by GRA6 gene sequence analysis. Int J Parasitol 2000; 30:637–42. Jacobs D, Vercammen M, Saman E. Evaluation of recombinant dense granule antigen 7 (GRA7) of Toxoplasma gondii for detection of immunoglobulin G antibodies and analysis of a major antigenic domain. Clin Diagn Lab Immunol 1999; 6:24–9. Neudeck A, Stachelhaus S, Nischik N, Striepen B, Reichmann G, Fischer HG. Expression variance, biochemical and immunological properties of Toxoplasma gondii dense granule protein GRA7. Microbes Infect 2002; 4:581–90. Bermudes D, Peck KR, Afifi MA, Beckers CJ, Joiner KA. Tandemly repeated genes encode nucleoside triphosphate hydrolase isoforms secreted into the parasitophorous vacuole of Toxoplasma gondii. J Biol Chem 1994; 269:29252–60. Johnson MS, Broady KW, Johnson AM. Differential recognition of Toxoplasma gondii recombinant nucleoside triphosphate hydrolase isoforms by naturally infected human sera. Int J Parasitol 1999; 29:1893–905. Howe DK, Honore S, Derouin F, Sibley LD. Determination of genotypes of Toxoplasma gondii strains isolated from patients with toxoplasmosis. J Clin Microbiol 1997; 35:1411–4. Godard I, Estaquier J, Zenner L, et al. Antigenicity and immunogenicity of P30-derived peptides in experimental models of toxoplasmosis. Mol Immunol 1994; 31:1353–63.

30. Aubert D, Maine GT, Villena I, et al. Recombinant antigens to detect Toxoplasma gondii–specific immunoglobulin G and immunoglobulin M in human sera by enzyme immunoassay. J Clin Microbiol 2000; 38: 1144–50. 31. Kasper LH, Ware PL. Recognition and characterization of stage-specific oocyst/sporozoite antigens of Toxoplasma gondii by human antisera. J Clin Invest 1985; 75:1570–7. 32. Ware PL, Kasper LH. Strain-specific antigens of Toxoplasma gondii. Infect Immun 1987; 55:778–83. 33. Weiss LM, Udem SA, Tanowitz H, Wittner M. Western blot analysis of the antibody response of patients with AIDS and toxoplasma encephalitis: antigenic diversity among Toxoplasma strains. J Infect Dis 1988; 157:7–13. 34. Handman E, Goding JW, Remington JS. Detection and characterization of membrane antigens of Toxoplasma gondii. J Immunol 1980; 124: 2578–83. 35. Couvreur G, Sadak A, Fortier B, Dubremetz JF. Surface antigens of Toxoplasma gondii. Parasitology 1988; 97:1–10. 36. Araujo F, Slifer T, Kim S. Chronic infection with Toxoplasma gondii does not prevent acute disease or colonization of the brain with tissue cysts following reinfection with different strains of the parasite. J Parasitol 1997; 83:521–2. 37. Dao A, Fortier B, Soete M, Plenat F, Dubremetz JF. Successful reinfection of chronically infected mice by a different Toxoplasma gondii genotype. Int J Parasitol 2001; 31:63–5.

T. gondii Serotyping with Synthetic Peptides • JID 2003:187 (1 May) • 1495