- Email: [email protected]
Differential recognition of Toxoplasma gondii recombinant nucleoside triphosphate hydrolase isoforms by naturally infected human sera
International Journal for Parasitology 29 (1999) 1893±1905
Dierential recognition of Toxoplasma gondii recombinant nucleoside triphosphate hydrolase isoforms by naturally infected human sera M.S. Johnson *, K.W. Broady, A.M. Johnson Molecular Parasitology Unit, University of Technology, Sydney, Westbourne Street, Gore Hill, NSW 2065, Australia Received 15 June 1999; received in revised form 2 August 1999; accepted 2 August 1999
Abstract Toxoplasma gondii possesses a highly active nucleoside triphosphate hydrolase, which has been shown to be an immunodominant antigen in mice and humans. Two isoforms (I and II) which exhibit dierent activities with respect to hydrolysis of ATP exist. Past studies suggest that all strains of T. gondii contain the less active nucleoside triphosphate hydrolase II, whilst only virulent strains contain the nucleoside triphosphate hydrolase I isoform. In order to further investigate the correlation between nucleoside triphosphate hydrolase isoform and biological signi®cance, we cloned and expressed as glutathione S-transferase fusion proteins the full-length nucleoside triphosphate hydrolase I and II isoforms and two truncations of the nucleoside triphosphate hydrolase I isoform in Escherichia coli. We then used ELISAs with the full-length recombinant nucleoside triphosphate hydrolases as antigens to examine 188 naturally infected T. gondii-positive sera and 83 T. gondii-negative sera for antibody reactivity. All positive sera reacted to T. gondii whole tachyzoite lysate antigen, 31 sera reacted to both nucleoside triphosphate hydrolase isoforms, three sera reacted speci®cally to nucleoside triphosphate hydrolase I and two sera reacted to only nucleoside triphosphate hydrolase II. Immunoblot analysis of the ®ve sera reacting to either nucleoside triphosphate hydrolase I or II revealed both quantitative and qualitative dierences in reactivity to the two isoforms. Comparative immunoblot analysis using the truncations of the nucleoside triphosphate hydrolase I isoform, and one of these positive sera identi®ed a presumptive dierential epitope between the nucleoside triphosphate hydrolase I and II isoforms within an 81 amino acid region (amino acids 445±526) at the C-terminus of the nucleoside triphosphate hydrolase I isoform. This dierential reactivity was further localised to the 12-residue region of greatest variability between the two isoforms (residues 488±499) using synthetic peptides. This is the ®rst report where naturally infected human sera have been used to identify a dierential epitope. Because this region is essential for substrate binding, an antibody response to this region may play some role in inhibition of this highly active enzyme. # 1999 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. Keywords: ELISA; Epitope; Human infection; NTPase isoforms; Toxoplasma
* Corresponding author. Tel: 61-2-9514 4019; fax: 61-2-9514 4003 E-mail address: [email protected] (M.S. Johnson). 0020-7519/99/$20.00 # 1999 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 0 - 7 5 1 9 ( 9 9 ) 0 0 1 3 9 - 3
1894
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
1. Introduction Toxoplasma gondii is an important human pathogen with a worldwide distribution. Serological evidence indicates that 30±40% of Western populations have been exposed to the parasite [1]. Infection is generally asymptomatic and chronic, but in immunocompromised individuals (such as those suering from AIDS), reactivation of the acute phase of the disease can lead to toxoplasmic encephalitis. If the disease is ®rst acquired during pregnancy, abortion and foetal abnormalities may result [1]. A range of factors such as immune status and genetic susceptibility of the host in¯uence the severity of the disease [2], although the infecting strain clearly plays a role, as has been observed in virulence studies in mice [3]. Based on such studies, the genus Toxoplasma has been divided into mouse virulent strains and mouse avirulent strains, where virulent strains multiply rapidly in the host and cause acute infection, whilst avirulent strains form cysts leading to a chronic infection [4]. The division of strains into virulent and avirulent lineages has been strengthened by genotyping at numerous loci using a range of techniques [5]. Using Restriction Fragment Length Polymorphism (RFLP) analysis, virulent strains were revealed to have identical genotypes at the SAG1, SAG2 and 850 loci, and avirulent strains were found to be genetically dissimilar from virulent strains at these combined loci [6]. In addition, Guo and Johnson [7] and Guo et al. [8], using RAPD-PCR, were able to identify markers which correlate with mouse virulence or mouse avirulence, allowing the characterisation of strains into the virulent and avirulent lineages. Antigenic variation among strains has been revealed in the reactions of polyclonal immune sera to whole cell lysates [9, 10], and recently Meisel et al. [11] detected a dierential epitope that could distinguish virulent strains from avirulent strains by use of an mAb to the GRA4 antigen. An allele within the coding region of the nucleoside triphosphate hydrolase (NTPase) that correlates with virulence has been discovered
using RFLP analysis [12]. This potent NTPase has been shown to exhibit elevated levels of activity in the presence of dithiothreitol (DTT) [13]. Among the apicomplexan parasites, this elevated level of DTT-dependent activity has only been identi®ed in T. gondii [14] and the closely related Neospora caninum [15]. It is now known that T. gondii possesses two NTPase isoforms, termed `NTPaseI' and `NTPaseII', which are separable by two-dimensional PAGE [12] and are characterised by their eciency of substrate hydrolysis. In particular, NTPaseI has a speci®c activity of ATP hydrolysis which is 4.5-fold greater than NTPaseII. The rate of hydrolysis of ATP to ADP compared with ADP to AMP is also signi®cantly dierent between the two isoforms. NTPaseI hydrolyses ATP with 75 times more eciency than ADP, whilst ATP and ADP are hydrolysed with the same, albeit lower, eciency by NTPaseII (Asai, Sibley and Takeuchi. Behaviour of NTPase during parasitophorous vacuole formation in the Toxoplasma infected cell, in Abstracts of the IX International Congress of Parasitology, Chiba, Japan, 1998, pp. 203±207). It was also shown that the gene encoding the highly active NTPaseI isoform is present only in virulent strains, whilst the gene encoding the less active NTPaseII isoform is present in all T. gondii strains. Translated cDNA sequences of the two isoforms revealed that they dier by only 16 aa (residues 21, 38, 47, 104±105, 110, 152, 171, 399, 437, 488±489, 492±493, 497, 499) over the entire 628-aa sequence [12, 16]. Native NTPase is apparently also a dominant antigen of T. gondii, as detection of anti-NTPase antibody corresponds with detection of antiToxoplasma antibody by the Sabin±Feldman dye test [17]. Asai et al. [18] had shown previously that high levels of NTPase are detected as early as 1 day p.i. in the blood of mice experimentally infected with the virulent RH strain of T. gondii, whereas signi®cantly lower levels of NTPase antigen were detected 3 days p.i. with the avirulent Beverley strain of T. gondii. Because Toxoplasma NTPase isoforms are highly correlated with virulence in mice, and as antigenic dierences have been found amongst T. gondii strains, we studied antigenic recognition of the NTPase isoforms by
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
individuals naturally infected by T. gondii. These results are reported here. 2. Materials and methods 2.1. Tachyzoite preparation Tachyzoites were harvested from the peritoneal cavities of mice infected 3 days earlier by RH strain tachyzoites. The solution containing the parasites was forced through a 26-gauge needle and host cell debris removed by ®ltration through a 3-mm polycarbonate membrane ®lter. This was followed by three washes with PBS (pH 7.4). 2.2. Genomic DNA isolation Isolation of genomic DNA from the RH strain of T. gondii was performed as described previously [19]. 2.3. PCR The NTPase gene was ampli®ed using PCR primers based on the published sequence [12]. The two primers were NTPfor (5 0 -TGT AGA ATT CCC GGT GTA TGT GCC TCT; forward) and NTPrev (5 0 -GTG TAC GGA TAA TCA CT; reverse). The reaction mixture contained 20 ng RH genomic DNA, 1.5 mM MgCl2, 0.5 pmol of each primer, 200 mM dNTPs and 1 unit of Taq polymerase (Biotech International). Cycling conditions consisted of denaturation at 958C for 5 min, followed by 24 reaction cycles at 938C for 2 min, 558C for 2 min, 728C for 2 min, followed by an extension cycle at 728C for 2 min. 2.4. Cloning of PCR-generated nucleoside triphosphate hydrolase gene into pGEX-1lT The PCR product was puri®ed using a QIAquick puri®cation column (QIAGEN). The resulting puri®ed DNA was digested with EcoR1 restriction enzyme and cloned non-directionally into the EcoR1 site within the multicloning site of pGEX-1lT (Pharmacia). Forward orientation
1895
clones were determined by PCR using primer NTPfor and the pGEX 3 0 sequencing primer (Pharmacia). The forward clones were subjected to BamH1 digestion to determine the isoform, as described previously [12]. Here, the cloned NTPI isoform is termed NTPaseIpGEX-1lT, whilst the cloned NTPII isoform is termed NTPaseIIpGEX-1lT. Two shorter fragments of the NTPase gene were ampli®ed from the NTPaseIpGEX-1lT cloned gene and then subcloned into pGEX-1lT as described above. Firstly, a 246-bp 3 0 end fragment was subcloned following PCR ampli®cation using primers NTPfor (5 0 -ACT TCG AAT TCG TGA GCC AGG TAG AAA GC; forward) and pGEX 3 0 sequencing primer. Secondly, a 1452-bp fragment was ampli®ed from the 5 0 end of NTPaseI by PCR using primers NTPfor (forward, described previously) and NTPIctermRev (5 0 -ACT TCG AAT TCC ATT GGT GCT GTC GAG GA; reverse). These subclones were termed NTPaseIs-pGEX-1lT and NTPaseIctermDelpGEX1lT, respectively.
2.5. Preparation of recombinant antigens Plasmids were transfected in CaCl2 competent Escherichia coli (BL21 strain) and were expressed by the addition of IPTG (0.1 mM ®nal concentration) to the culture media. The full-length recombinant NTPase isoforms were puri®ed from inclusion bodies as described previously [20]. Following initial puri®cation, the antigens were further puri®ed by anion exchange chromatography using DEAE Sephacel (Pharmacia) equilibrated in 50 mM Tris (pH 8), eluted with an 80 mM NaCl step gradient, and detected by immmunoblotting with goat anti-glutathione Stransferase (anti-GST; Pharmacia). The puri®ed antigens were then dialysed against 50 mM Tris (pH 8)/10% glycerol. The two shorter antigens were expressed and puri®ed from the soluble fraction by anity chromatography using glutathione-bound agarose (Sigma). Protein concentrations were determined by a Bradford dye binding assay.
1896
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
2.6. Preparation of Toxoplasma tachyzoite antigen Tachyzoites of T. gondii were prepared and the soluble fraction was retained after disruption by sonication, as described previously [21]. Protein concentrations were determined by a Bradford dye binding assay. 2.7. Sera The 188 T. gondii-positive sera and 81 T. gondii-negative sera were obtained from 269 patients who were investigated at the Flinders Medical Centre (Adelaide) in 1995 for a variety of conditions in which toxoplasmosis may be important. Discrimination of positive and negative sera was based on a routine diagnostic ELISA [22]. Positive reference sample was pooled from six positive sera and a negative reference sample was pooled from six negative sera. 2.8. ELISA Antigens were diluted in 0.1 M carbonate/bicarbonate buer (pH 9.6). Flat-bottom ELISA plates (Costar) were coated with 100 ml of antigen solution per well, incubated overnight at 48C and then washed four times with PBS/0.03% Tween20 (PBST). The sensitised wells were then blocked with 2.5% w/v skim milk powder in PBST and incubated at 378C for 1 h, followed by four washes with PBST. Serum samples were diluted in 1% w/v skim milk powder in PBST and 100 ml added to each respective well. The plates were incubated and washed as before. Anti-human Ig conjugated to alkaline phosphatase (Dako) was diluted in 1% w/v skim milk powder and 100 ml added to each well. The plates were incubated and washed as before. One hundred microlitres of 1 mg/ml p-nitrophenyl phosphate substrate (Sigma) dissolved in 10% diethylamine buer was added to each well and incubated at 228C in the dark. The O.D. values were measured at 405 nm against a control well that had not received serum or conjugate.
2.9. Optimisation Chequerboard titrations of antigen, primary pooled sera and conjugate were performed as outlined previously [22], to determine the optimum conditions which discriminate between positive and negative sera. To account for any possible reactivity to the GST portion of the recombinant NTPase antigens, a GST antigen well containing the quantity of GST attached to the NTPase antigen was included for every sample. The absorbance of this GST well was subtracted from the recombinant NTPase absorbance for every sample tested. 2.10. PAGE and immunoblot analysis Immunoblot analysis of sera A±E was performed twice. Recombinant proteins were electrophoresed by SDS±PAGE [23] using 4±20% Tris glycine gels under denaturing conditions. Two micrograms of each isoform were loaded and after electrophoresis proteins were transferred to polyvinylidene ¯uoride membrane by electroblotting. The membranes were stained with Ponceau S to ensure that equal amounts of each isoform had been transferred. Membranes were then blocked for 1 h in PBST/5% w/v skim milk powder. After washing with PBST, membranes were incubated with primary sera diluted 1:400 for 1 h. The membranes were washed with PBST, followed by incubation with the antihuman Ig conjugated to alkaline phosphatase (Dako) diluted 1:1000. The membranes were developed using nitroblue tetrazolium/BCIP (Sigma) dissolved in alkaline phosphatase buer (100 mM Tris, pH 9.5, 5 mM MgCl2, 100 mM NaCl). 2.11. Densitometry Membranes were scanned with a Microtek ScanmakerIII ¯atbed scanner. The band intensities (O.D.mm2) were measured using Molecular Analyst version 2.01 (Biorad). After subtraction of backgound, the band intensity for each isoform was expressed as percentage of the total
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
band intensity (O.D.mm2 2 O.D.mm NTPaseII).
1897
NTPaseI +
2.12. Synthetic peptides Synthetic peptides (length = 20 aa; NTPaseI: H±CKAPMIVTGGGMLAAINTLK±OH; NTPaseII: H±CKAPMFITGREMLASIDTLK± OH) encompassing the 12-aa region of variability were manufactured commercially (Chiron Mimotopes). For the ELISA studies, 0.25 mg of the synthetic peptides was coated to each well, and the ELISA performed as described above. 3. Results 3.1. PCR and cloning An NTPase gene fragment was ampli®ed from T. gondii RH strain genomic DNA, using primers NTPfor and NTPrev. A 1603-bp fragment resulted (data not shown) which, after puri®cation, was digested with EcoR1 to liberate a 1571-bp DNA fragment, representing aa 3±526, or 84% of the protein. After cloning this fragment and determining correct orientation, the isoforms were selected by digesting representative clones with BamH1. NTPaseI clones (NTPaseIpGEX1lT) failed to cut, whilst NTPaseII clones (NTPaseIIpGEX1lT) were digested to reveal two bands of 5408 and 1135 bp (Fig. 1). A 239-bp fragment representing aa 445± 526 and a 1452-bp fragment representing aa 3±
Fig. 1. Identi®cation of nucleoside triphosphate hydrolase isoform by BamH1 digestion of recombinant pGEX/nucleoside triphosphate hydrolase clones. Lane 1: 100-bp DNA ladder. Lane 2: BamH1 digested recombinant pGEX clone containing the nucleoside triphosphate hydrolase I isoform (NTPaseIpGEX1lT). Lane 3: BamHI digested pGEX recombinant clone containing the nucleoside triphosphate hydrolase II isoform (NTPaseIIpGEX1lT).
487 were subcloned from NTPaseIpGEX1lT, to yield recombinant plasmids NTPaseIs-pGEX1lT and NTPaseIctermDel-pGEX1lT, respectively.
Table 1 Absorbances of ®ve sera reacting positively to either nucleoside triphosphate hydrolase I or II in the ELISA Isoform NTPaseI
NTPaseII
Sample
Absorbance
Result
Absorbance
Result
A B C D E
0.33620.042 0.44120.097 0.18020.048 0.15920.031 0.16020.008
Positive Positive Positive Negative Negative
0.11720.016 0.20820.016 0.17920.012 0.26320.016 0.23020.016
Negative Negative Negative Positive Positive
1898
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
3.2. Expression The NTPaseI and NTPaseII clones were expressed in E. coli. The resulting fusion proteins were always found in the insoluble fraction in the form of inclusion bodies after cell lysis (data not shown). This was found to be independent of temperature of expression or IPTG concentration. Furthermore, it was noted that the E. coli heat shock protein `DNA K' was also present in the insoluble fraction (data not shown). The insoluble inclusion bodies were able to be puri®ed by centrifugation, and after solubilisation in 8 M urea (pH 8), were further puri®ed by anion exchange chromatography under denaturing conditions to remove DNA K. After renaturation, both puri®ed NTPase isoforms were determined to have an apparent Mr of 88 kDa (data not shown). Expression of NTPaseIs and NTPaseIctermDel truncations yielded proteins of 36 and 84 kDa, respectively (data not shown). 3.3. Optimisation of ELISA A working dilution of 1:50 of primary sera and 1:500 of the conjugate was determined by chequerboard assay with the T. gondii soluble fraction antigen (1 mg per well). The two NTPase isoforms (0.5 mg per well) were determined to have a working dilution of 1:50 for the primary sera and 1:250 for the conjugate. 3.4. Test sera The mean + 3 S.D. of the absorbances from the 83 negative sera were used to determine the value where discrimination between positive and negative samples could be determined. The
mean + 3 S.D. was 0.381 for the T. gondii lysate (positive control antigen), 0.178 for recombinant NTPaseI and 0.222 for recombinant NTPaseII. Samples with an absorbance value greater than the discrimination value were determined as positive. Eighteen percent (34/188) of the known positive sera were determined as positive for the NTPaseI isoform, and 17.5% (33/188) of known positive samples were determined as positive with the NTPaseII isoform. One hundred percent of known positive sera were determined to be positive with T. gondii lysate (positive control antigen). No false positives were detected among the known negative sera for any of the antigens. Interestingly, three samples were exclusively positive to NTPaseI and two samples were exclusively positive for NTPaseII. These ®ve sera were tested twice in duplicate and the results are summarised in Table 1. Of the 36 sera which were determined to be positive by ELISA for at least one isoform, eight sera displayed dierences in antibody reactivity to the isoforms (Fig. 2). 3.5. Immunoblot analysis To con®rm that the dierences detected in the ®ve sera were speci®c to NTPase, immunoblots were performed (Fig. 3). Quantitative dierences were observed in sera D and E, whilst a qualitative dierence was observed in serum A. No signi®cant dierence (P > 0.05) was observed in sera B and C. Because serum A reacted strongly with the NTPaseI isoform and no reaction with the NTPaseII isoform was detected, immunoblots were performed with the truncated forms NTPaseIs and NTPaseIctermDel to isolate the
Fig. 2. Dierential antibody response to nucleoside triphosphate hydrolase isoforms. (A) The 36 sera (X-axis) which were determined to be positive to at least one of the nucleoside triphosphate hydrolase isoforms by ELISA were plotted against the dierence in absorbance at 405 nm (nucleoside triphosphate hydrolase Iÿnucleoside triphosphate hydrolase II). The response types are: ``A'', equal response to each isoform; ``B'', response to nucleoside triphosphate hydrolase I>response to nucleoside triphosphate hydrolase II; ``C'', response to nucleoside triphosphate hydrolase I < response to nucleoside triphosphate hydrolase II. (B) The same 36 sera (X-axis) were plotted against the standard deviation (n = 2) of the absorbances of each sera to the two isoforms. The mean S.D. is 0.050. LCL = lower con®dence limit; UCL = upper con®dence limit. Eight sera, corresponding to the response type B and C sera, are signi®cantly greater than the mean S.D. Twenty-eight sera corresponding to the type A response have standard deviations less than or equal to the mean S.D. (®gure opposite)
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
Fig. 2.
1899
1900
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
Fig. 3. Immunoblot analysis of the ®ve sera which were positive for one isoform only by ELISA. A±E represent the ®ve sera A±E. Recombinant nucleoside triphosphate hydrolase isoforms were run in triplicate on 4±20% Tris±glycine gels under reducing conditions and transferred to polyvinylidene ¯uoride membrane. Lane I: recombinant nucleoside triphosphate hydrolase I (2 mg). Lane II: recombinant nucleoside triphosphate hydrolase II (2 mg). O.D. values (data not shown) were determined by densitometry, as described above.
region of this speci®c reactivity. Reactivity was observed in the NTPaseIs truncated protein (Fig. 4), which represents aa 445±526. No reactivity was observed in the NTPaseIctermDel (aa 3±487) truncated protein or the GST control, thus con®rming that the only area of antibody
Fig. 4. Localisation of reactivity to nucleoside triphosphate hydrolase immunoblotted with serum A. Lane 1: nucleoside triphosphate hydrolase I (aa 3±526). Lane 2: NTPaseIctermDel (aa 3±487). Lane 3: NTPaseIs (aa 445± 526). Lane 4: glutathione S-transferase control.
recognition on the NTPaseI isoform occurs within 38 C-terminal amino acids (aa 488±526).
3.6. Antibody reactivity to synthetic peptides The mean + 3 S.D. of the absorbances from 12 T. gondii-negative samples diluted 1:50 was determined for each peptide. The mean + 3 S.D. for the peptides KCAPMIVTGGGMLAAINTLK and KCAPMFITGREMLASIDTLK were 0.260 and 0.255, respectively. The samples which were determined previously to be NTPase positive were then assayed for reactivity to the synthetic peptides. Five of the 36 positive samples were identi®ed as having absorbances greater than the mean + 3 S.D. for both of the peptides. These ®ve sera, as well as the negative reference sera, were serially diluted to con®rm speci®c antibody reactivity (Fig. 5). Four of these sera (panels 2± 5) showed equal reactivity to both peptides, whilst one serum (panel 1, serum A described previously) displayed dierential reactivity, where the antibody response to the peptide CKAPMIVTGGGMLAAINTLK was signi®cantly greater (P < 0.05) than the response to the corresponding peptide CKAPMFITGREMLASIDTLK. A Toxoplasma pooled negative sample (described previously) showed no response over the dilution range.
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
1901
Fig. 5. Antibody response to synthetic peptides. Panels 1±5: serial dilution of sera which were found to have absorbances greater than the mean + 3 S.D. of 12 Toxoplasma-negative sera against peptides (0.25 mg per well). CKAPMIVTGGGMLAAINTLK (w) and KCAPMFITGREMLASIDTLK (t). Panel 6: serial dilution of pooled Toxoplasma-negative reference sera against the peptides.
1902
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
4. Discussion The NTPase of T. gondii has DTT-dependent hydrolytic properties which were thought to be unique to the genus. The ®nding that T. gondii possessed two NTPase isoforms, and the classi®cation of these isoforms as NTPaseI and NTPaseII based upon varying substrate hydrolysis, was made more interesting by the fact that only virulent strains possessed the highly active NTPaseI [12]. Recently, Asai et al. [15] have demonstrated that the closely related genus Neospora also shares this unique DTT-dependent NTPase activity. It is interesting to note that N. caninum (mouse virulent) was found to possess NTPaseI activity. The reason for the existence of two NTPase isoforms and explanations for a correlation of NTPaseI with mouse virulent strains are unclear. The role of antibody response to the two isoforms is also unknown and we believe that our study reported here is the ®rst to examine this interesting biological question. Both qualitative and quantitative dierences in antibody reactivity to the two NTPase isoforms of T. gondii were found to exist in sera from naturally infected humans. From a diagnostic point of view, the recombinant NTPase ELISA was able to identify only 18% of sera with anti-NTPase antibody in patient sera known to contain anti-Toxoplasma antibody. Asai et al. [17] demonstrated that antiNTPase antibodies detected by the use of native NTPase in an indirect ELISA correlated with detection of anti-Toxoplasma antibody. Tenter and Johnson [22] showed previously that a 50 kDa recombinant NTPase ELISA could detect anti-NTPase antibody in only 7% of known T. gondii-positive sera. The longer recombinant NTPase isoforms used here achieved a greater sensitivity of detection, probably due to a greater representation of epitopes. However, the sensitivities achieved by recombinant NTPase antigens here and in the previous study by Tenter and Johnson [22] are substantially less than those achieved in the study of native NTPase ELISA of Asai et al. [17]. There are 103 C-terminal amino acids in the native NTPase isoforms which
are not represented in the recombinant isoforms studied here. It is possible that there may be epitopes in this region that may also contribute to antigenicity. Problems with expression should also be considered, as conformationally correct recombinant proteins can be dicult to express in the reducing environment of the E. coli cytoplasm, especially if the recombinant protein contains disulphide bonds. Silverman et al. [24] suggested that only about 5% of the NTPase in the vacuole space is in its reduced (insoluble) active form, and perhaps NTPase is most antigenic in its oxidised (soluble) inactive form. This hypothesis is supported by the fact that the native NTPase antigen used by Asai et al. [17] was extracted from the soluble fraction of T. gondii tachyzoites. Alternatively, recombinant NTPase expressed in E. coli may not be conformationally correct due to incorrect post-translational modi®cations such as incorrect glycosylation, since the NTPase isoforms have one potential N-linked glycosylation site at aa 432 [16]. Determination of enzyme activity could con®rm whether conformationally correct recombinants have been expressed, whilst the use of eukaryotic expression systems such as yeast or baculovirus may be useful in the expression of NTPase with correct conformation for diagnostic purposes. Although the assay identi®ed anti-NTPase antibody in only about 18% of patients known to possess anti-Toxoplasma antibody, ®ve sera were identi®ed which were positive for one of the isoforms only. Immunoblotting was used to determine the speci®city of these reactions, and both quantitative and qualitative dierences in antibody response to the isoforms were noted (Fig. 3). The qualitative dierence in reactivity to NTPase observed in serum A has been isolated to a region of 81 aa (aa 445±526) at the C-terminus of NTPase. Deletion of this C-terminus region (aa 487±526) resulted in the complete loss of antibody reactivity (Fig. 4), con®rming that the only area of antibody recognition resides in this 38-aa C-terminus region. It is interesting to note from the previously published sequences [12, 16] that there are only six amino acids in a stretch of 12 that vary between the two
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
isoforms in this region. This 12-aa stretch (residues 488±499, IVTGGGMLAAIN: NTPaseI; FITGREMLASID: NTPaseII), which contains three charged amino acids in NTPaseI but is uncharged in NTPaseII, represents the region of greatest variability between the two isoforms, and may form part of a dierential epitope. This ®nding is made more interesting by a recent study which has identi®ed this region as being responsible for substrate binding and for the control of substrate speci®city [25]. Antibody reactivity to this region may possibly play some role in regulating NTPase activity by inhibiting the binding of substrate. Because our immunoblot analysis was performed under denaturing conditions, the epitopes recognised are probably linear rather than conformational. Epitope mapping of this region by the use of overlapping synthetic peptides should con®rm the exact epitope sequence, and DNA sequence analysis of this region in a number of virulent and avirulent strains may further strengthen the correlation between the NTPase isoform and virulence. Although serum A was the only serum to demonstrate qualitative dierences in response to the NTPase isoforms, four other sera were identi®ed in which a positive ELISA reaction was exclusive to one isoform only. Sera D and E were found to react exclusively to NTPaseII in ELISA. These results were con®rmed by the signi®cantly greater band intensities (P < 0.05) observed for the NTPaseII isoform by immunoblotting. However, unlike serum A, the dierences observed by immunoblotting were quantitative. The presence of the GST fusion partner may explain why the dierences were quantitative rather than qualitative; however, the possibility that a dierential epitope exists in NTPaseII cannot be excluded. It was surprising to note that immunoblotting of sera B and C did not con®rm the ELISA results. For example, serum B reacted to NTPaseI with greater than twice the absorbance it did to NTPaseII by ELISA, yet no signi®cant dierence between the isoforms was detected by immunoblotting. This would suggest that reactivity to conformational epitopes has occurred in the ELISA, but this was
1903
not observed in the immunoblot analysis due to the denaturing conditions used. It should also be noted that, although serum C is exclusively positive to NTPaseI, the absorbance values of serum C for each isoform by ELISA are almost identical and there is no signi®cant dierence in reactivity to the isoforms observed by immunoblotting. This highlights a problem associated with classifying a serum as ELISA positive when it displays an absorbance close to the positive cut-o value. When looking at the dierences in absorbance between the two isoforms detected by ELISA (Fig. 2) for the 36 samples which were positive for at least one isoform, it becomes clear that the patient sera can be classi®ed into three types of response. Firstly, sera which show equal response for each isoform (`type A' response). Secondly, sera which show a signi®cantly greater response to the NTPaseI isoform (`type B' response) and, ®nally, sera which show a signi®cantly greater response to the NTPaseII isoform (`type C' response). The response for most sera was found to fall into the type A category, where absorbances were very similar for both isoforms. This type of response can be explained by the high degree of homology between the two isoforms. However, even when looking at this type A response, sera tend to exhibit a greater reactivity to the NTPaseII isoform. This may be indicative of a higher prevalence of avirulent strains causing infection among these patients. Support for this observation comes from a recent study which employed a nested PCR to type clinical Toxoplasma isolates and found that 81% were of the avirulent (type II) lineage [26]. The bias toward higher reactivity to the NTPaseI isoform observed in the type B response may be due to infection by virulent Toxoplasma strains where the increased absorbance may be due to antibody binding to the dierential epitope described previously. Conversely, the higher reactivity to the NTPaseII isoform observed in the type C response may be due to infection by avirulent strains, where once again the increased absorbance may be caused by a dierential epitope. Although dierences were observed in antibody reactivity to the two recombinant isoforms,
1904
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
it is unclear what regions of the NTPase isoforms contribute to this dierential recognition by human sera. The data from immunoblotting with serum A suggest that dierential recognition occurs within 38 C-terminal amino acids. In order to test the possibility that sera other than serum A display dierential reactivity in this region, we used 20-residue synthetic peptides encompassing this region of variability to study the 36 NTPase-positive sera identi®ed previously by the recombinant ELISA. Five sera were identi®ed with an absorbance greater than the mean + 3 S.D. of 12 Toxoplasma-negative samples. When these sera were serially diluted against the peptides representing each isoform, four sera reacted equally to both isoforms, whilst one serum sample (serum A, described previously) displayed dierential reactivity (Fig. 5). Although the antibody response of serum A to the NTPaseI peptide was signi®cantly greater than to the NTPaseII peptide, some cross-reactivity was observed. The results con®rm that the dierential reactivity in serum A occurs at the 12-aa region of greatest variability. However, this region does not appear to be involved in dierential antibody reactivity for any of the other 35 NTPase-positive sera, although it is clearly a recognised epitope for ®ve of these sera. It may be possible that the IVTGGGMLAAIN sequence (NTPaseI) is the homologous epitope stimulating the antibody reponse, whereas the FITGREMLASID sequence (NTPaseII) is a heterologous epitope, and is thus the cross-reactive epitope. Further studies of this epitope region using sera from virulent and avirulent experimental infections will determine if there is a correlation between reactivity to the IVTGGGMLAAIN epitope and infection by virulent Toxoplasma strains. A correlation of congenital toxoplasmosis with infection by virulent strains has been observed previously [27], and therefore antibodies from naturally infected humans that recognise an epitope that correlates with infection by virulent T. gondii strains would have important diagnostic and epidemiological value.
References [1] Remington JS, Desmonts G. In: Remington JS, Klein JO, editors. Infectious diseases of the fetus and newborn infant. Philadelphia. W.B. Saunders, 1990; 89±195. [2] Johnson AM. In: Long PL, editor. Coccidiosis of man and domestic animals. Boca Raton, FL: CRC Press, 1990; 121±135. [3] Johnson AM. Strain-dependent route of challenge-dependent, murine susceptibility to toxoplasmosis. Z Parasitenk 1984;70:303±9. [4] Dubey JP. Advances in the life cycle of Toxoplasma gondii. Int J Parasitol 1998;28:1019±24. [5] Johnson AM. Speculation on possible lifecycles for the clonal lineages in the genus Toxoplasma. Parasitol Today 1997;13:393±7. [6] Parmley SF, Gross U, Sucharczuk A, Windeck T, Sgariato GD, Remington JS. Two alleles of the gene encoding surface antigen P22 in 25 strains of Toxoplasma gondii. J Parasitol 1994;80:293±301. [7] Guo ZG, Johnson AM. Genetic characterization of Toxoplasma gondii strains by random ampli®ed polymorphic DNA polymerase chain reaction. Parasitology 1995;111:127±32. [8] Guo ZG, Gross U, Johnson AM. Toxoplasma gondii virulence markers identi®ed by RAPD PCR. Parasitol Res 1997;83:453±8. [9] Ware PL, Kasper LH. Strain-speci®c antigens of Toxoplasma gondii. Infect Immun 1987;55:778±83. [10] 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 1987;157:7±13. [11] Meisel R, Stachelhaus S, Mevelec MN, Reichmann G, Dubremetz JF, Fischer HG. Identi®cation of two alleles in the GRA4 locus of Toxoplasma gondii determining a dierential epitope which allows discrimination of type I versus type II and III strains. Mol Biochem Parasitol 1996;81:259±63. [12] Asai T, Sibley LD, Okabayashi H, Takeuchi T. Biochemical and molecular characterization of nucleoside triphosphate hydrolase isozymes from the parasitic protozoan Toxoplasma gondii. J Biol Chem 1995;270:11391± 7. [13] Asai T, O'Sullivan WJ, Tatibana M. A potent nucleoside triphosphate hydrolase from the parasitic protozoan Toxoplasma gondii. J Biol Chem 1983;258:6816±22. [14] Asai T, Kanawaza T, Kobayashi S, Takeuchi T, Kim T. Do protozoa conceal a high potency of nucleoside triphosphate hydrolysis in Toxoplasma gondii?. Comp Biochem Physiol 1986;85B:365±7. [15] Asai T, Howe DK, Nakajima K, Nozaki T, Takeuchi T, Sibley LD. Neospora caninum: tachyzoites express a potent type-1 nucleoside triphosphate hydrolase activity. Exp Parasitol 1998;90:277±85. [16] Bermudes D, Peck KR, A®® MA, Beckers CJM, Joiner KA. Tandemly repeated genes encode nucleoside tripho-
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
[17]
[18]
[19] [20] [21]
sphate hydrolase isoforms secreted into the parasitophorous vacuole of Toxoplasma gondii. J Biol Chem 1994;269:29252±60. Asai T, Mizuno F, Kojima S, Takeuchi T, Kobayashi A, Suzuki Y. High correlation in antibody titers between Sabin±Feldman dye test and an enzyme linked immunosorbent assay detecting immunoglobulin G antibodies to the nucleoside triphosphate hydrolase of Toxoplasma gondii. J Clin Microbiol 1992;30:1291±3. Asai T, Kim T, Kobayashi M, Kojima S. Detection of nucleoside triphosphate hydrolase as a circulating antigen in the sera of mice infected with Toxoplasma gondii. Infect Immun 1987;55:1332±55. Johnson AM, Dubey JP, Dame JB. Puri®cation and characterization of Toxoplasma gondii tachyzoite DNA. Aust J Exp Biol Med Sci 1986;64:351±5. Lin KH, Cheng SY. An ecient method to purify active eukaryotic proteins from inclusion bodies in Escherichia coli. Biotechniques 1991;11:748±52. Johnson AM, Gu QM, Roberts H. Antibody patterns in the serological diagnosis of acute lymphadenopathic toxoplasmosis. Aust NZ J Med 1987;17:430±4.
1905
[22] Tenter AM, Johnson AM. Human antibody response to the nucleoside triphosphate hydrolase of Toxoplasma gondii. J Immunoassay 1990;11:579±90. [23] Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680±5. [24] Silverman JA, Huilin Q, Riehl A, Beckers C, Nakaar V, Joiner KA. Induced activation of the Toxoplasma gondii nucleoside triphosphate hydrolase leads to depletion of host cell ATP levels and rapid exit of intracellular parasites from infected cells. J Biol Chem 1998;273:12352±9. [25] Nakaar V, Beckers CJM, Polotsky V, Joiner KA. Basis for substrate speci®city of the Toxoplasma gondii nucleoside triphosphate hydrolase. Mol Biochem Parasitol 1998;97:209±20. [26] 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. [27] Howe DK, Sibley LD. Toxoplasma gondii comprises three clonal lineages: Correlation of parasite genotype with human disease. J Infect Dis 1995;172:1561±6.
Dierential recognition of Toxoplasma gondii recombinant nucleoside triphosphate hydrolase isoforms by naturally infected human sera M.S. Johnson *, K.W. Broady, A.M. Johnson Molecular Parasitology Unit, University of Technology, Sydney, Westbourne Street, Gore Hill, NSW 2065, Australia Received 15 June 1999; received in revised form 2 August 1999; accepted 2 August 1999
Abstract Toxoplasma gondii possesses a highly active nucleoside triphosphate hydrolase, which has been shown to be an immunodominant antigen in mice and humans. Two isoforms (I and II) which exhibit dierent activities with respect to hydrolysis of ATP exist. Past studies suggest that all strains of T. gondii contain the less active nucleoside triphosphate hydrolase II, whilst only virulent strains contain the nucleoside triphosphate hydrolase I isoform. In order to further investigate the correlation between nucleoside triphosphate hydrolase isoform and biological signi®cance, we cloned and expressed as glutathione S-transferase fusion proteins the full-length nucleoside triphosphate hydrolase I and II isoforms and two truncations of the nucleoside triphosphate hydrolase I isoform in Escherichia coli. We then used ELISAs with the full-length recombinant nucleoside triphosphate hydrolases as antigens to examine 188 naturally infected T. gondii-positive sera and 83 T. gondii-negative sera for antibody reactivity. All positive sera reacted to T. gondii whole tachyzoite lysate antigen, 31 sera reacted to both nucleoside triphosphate hydrolase isoforms, three sera reacted speci®cally to nucleoside triphosphate hydrolase I and two sera reacted to only nucleoside triphosphate hydrolase II. Immunoblot analysis of the ®ve sera reacting to either nucleoside triphosphate hydrolase I or II revealed both quantitative and qualitative dierences in reactivity to the two isoforms. Comparative immunoblot analysis using the truncations of the nucleoside triphosphate hydrolase I isoform, and one of these positive sera identi®ed a presumptive dierential epitope between the nucleoside triphosphate hydrolase I and II isoforms within an 81 amino acid region (amino acids 445±526) at the C-terminus of the nucleoside triphosphate hydrolase I isoform. This dierential reactivity was further localised to the 12-residue region of greatest variability between the two isoforms (residues 488±499) using synthetic peptides. This is the ®rst report where naturally infected human sera have been used to identify a dierential epitope. Because this region is essential for substrate binding, an antibody response to this region may play some role in inhibition of this highly active enzyme. # 1999 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. Keywords: ELISA; Epitope; Human infection; NTPase isoforms; Toxoplasma
* Corresponding author. Tel: 61-2-9514 4019; fax: 61-2-9514 4003 E-mail address: [email protected] (M.S. Johnson). 0020-7519/99/$20.00 # 1999 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 0 - 7 5 1 9 ( 9 9 ) 0 0 1 3 9 - 3
1894
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
1. Introduction Toxoplasma gondii is an important human pathogen with a worldwide distribution. Serological evidence indicates that 30±40% of Western populations have been exposed to the parasite [1]. Infection is generally asymptomatic and chronic, but in immunocompromised individuals (such as those suering from AIDS), reactivation of the acute phase of the disease can lead to toxoplasmic encephalitis. If the disease is ®rst acquired during pregnancy, abortion and foetal abnormalities may result [1]. A range of factors such as immune status and genetic susceptibility of the host in¯uence the severity of the disease [2], although the infecting strain clearly plays a role, as has been observed in virulence studies in mice [3]. Based on such studies, the genus Toxoplasma has been divided into mouse virulent strains and mouse avirulent strains, where virulent strains multiply rapidly in the host and cause acute infection, whilst avirulent strains form cysts leading to a chronic infection [4]. The division of strains into virulent and avirulent lineages has been strengthened by genotyping at numerous loci using a range of techniques [5]. Using Restriction Fragment Length Polymorphism (RFLP) analysis, virulent strains were revealed to have identical genotypes at the SAG1, SAG2 and 850 loci, and avirulent strains were found to be genetically dissimilar from virulent strains at these combined loci [6]. In addition, Guo and Johnson [7] and Guo et al. [8], using RAPD-PCR, were able to identify markers which correlate with mouse virulence or mouse avirulence, allowing the characterisation of strains into the virulent and avirulent lineages. Antigenic variation among strains has been revealed in the reactions of polyclonal immune sera to whole cell lysates [9, 10], and recently Meisel et al. [11] detected a dierential epitope that could distinguish virulent strains from avirulent strains by use of an mAb to the GRA4 antigen. An allele within the coding region of the nucleoside triphosphate hydrolase (NTPase) that correlates with virulence has been discovered
using RFLP analysis [12]. This potent NTPase has been shown to exhibit elevated levels of activity in the presence of dithiothreitol (DTT) [13]. Among the apicomplexan parasites, this elevated level of DTT-dependent activity has only been identi®ed in T. gondii [14] and the closely related Neospora caninum [15]. It is now known that T. gondii possesses two NTPase isoforms, termed `NTPaseI' and `NTPaseII', which are separable by two-dimensional PAGE [12] and are characterised by their eciency of substrate hydrolysis. In particular, NTPaseI has a speci®c activity of ATP hydrolysis which is 4.5-fold greater than NTPaseII. The rate of hydrolysis of ATP to ADP compared with ADP to AMP is also signi®cantly dierent between the two isoforms. NTPaseI hydrolyses ATP with 75 times more eciency than ADP, whilst ATP and ADP are hydrolysed with the same, albeit lower, eciency by NTPaseII (Asai, Sibley and Takeuchi. Behaviour of NTPase during parasitophorous vacuole formation in the Toxoplasma infected cell, in Abstracts of the IX International Congress of Parasitology, Chiba, Japan, 1998, pp. 203±207). It was also shown that the gene encoding the highly active NTPaseI isoform is present only in virulent strains, whilst the gene encoding the less active NTPaseII isoform is present in all T. gondii strains. Translated cDNA sequences of the two isoforms revealed that they dier by only 16 aa (residues 21, 38, 47, 104±105, 110, 152, 171, 399, 437, 488±489, 492±493, 497, 499) over the entire 628-aa sequence [12, 16]. Native NTPase is apparently also a dominant antigen of T. gondii, as detection of anti-NTPase antibody corresponds with detection of antiToxoplasma antibody by the Sabin±Feldman dye test [17]. Asai et al. [18] had shown previously that high levels of NTPase are detected as early as 1 day p.i. in the blood of mice experimentally infected with the virulent RH strain of T. gondii, whereas signi®cantly lower levels of NTPase antigen were detected 3 days p.i. with the avirulent Beverley strain of T. gondii. Because Toxoplasma NTPase isoforms are highly correlated with virulence in mice, and as antigenic dierences have been found amongst T. gondii strains, we studied antigenic recognition of the NTPase isoforms by
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
individuals naturally infected by T. gondii. These results are reported here. 2. Materials and methods 2.1. Tachyzoite preparation Tachyzoites were harvested from the peritoneal cavities of mice infected 3 days earlier by RH strain tachyzoites. The solution containing the parasites was forced through a 26-gauge needle and host cell debris removed by ®ltration through a 3-mm polycarbonate membrane ®lter. This was followed by three washes with PBS (pH 7.4). 2.2. Genomic DNA isolation Isolation of genomic DNA from the RH strain of T. gondii was performed as described previously [19]. 2.3. PCR The NTPase gene was ampli®ed using PCR primers based on the published sequence [12]. The two primers were NTPfor (5 0 -TGT AGA ATT CCC GGT GTA TGT GCC TCT; forward) and NTPrev (5 0 -GTG TAC GGA TAA TCA CT; reverse). The reaction mixture contained 20 ng RH genomic DNA, 1.5 mM MgCl2, 0.5 pmol of each primer, 200 mM dNTPs and 1 unit of Taq polymerase (Biotech International). Cycling conditions consisted of denaturation at 958C for 5 min, followed by 24 reaction cycles at 938C for 2 min, 558C for 2 min, 728C for 2 min, followed by an extension cycle at 728C for 2 min. 2.4. Cloning of PCR-generated nucleoside triphosphate hydrolase gene into pGEX-1lT The PCR product was puri®ed using a QIAquick puri®cation column (QIAGEN). The resulting puri®ed DNA was digested with EcoR1 restriction enzyme and cloned non-directionally into the EcoR1 site within the multicloning site of pGEX-1lT (Pharmacia). Forward orientation
1895
clones were determined by PCR using primer NTPfor and the pGEX 3 0 sequencing primer (Pharmacia). The forward clones were subjected to BamH1 digestion to determine the isoform, as described previously [12]. Here, the cloned NTPI isoform is termed NTPaseIpGEX-1lT, whilst the cloned NTPII isoform is termed NTPaseIIpGEX-1lT. Two shorter fragments of the NTPase gene were ampli®ed from the NTPaseIpGEX-1lT cloned gene and then subcloned into pGEX-1lT as described above. Firstly, a 246-bp 3 0 end fragment was subcloned following PCR ampli®cation using primers NTPfor (5 0 -ACT TCG AAT TCG TGA GCC AGG TAG AAA GC; forward) and pGEX 3 0 sequencing primer. Secondly, a 1452-bp fragment was ampli®ed from the 5 0 end of NTPaseI by PCR using primers NTPfor (forward, described previously) and NTPIctermRev (5 0 -ACT TCG AAT TCC ATT GGT GCT GTC GAG GA; reverse). These subclones were termed NTPaseIs-pGEX-1lT and NTPaseIctermDelpGEX1lT, respectively.
2.5. Preparation of recombinant antigens Plasmids were transfected in CaCl2 competent Escherichia coli (BL21 strain) and were expressed by the addition of IPTG (0.1 mM ®nal concentration) to the culture media. The full-length recombinant NTPase isoforms were puri®ed from inclusion bodies as described previously [20]. Following initial puri®cation, the antigens were further puri®ed by anion exchange chromatography using DEAE Sephacel (Pharmacia) equilibrated in 50 mM Tris (pH 8), eluted with an 80 mM NaCl step gradient, and detected by immmunoblotting with goat anti-glutathione Stransferase (anti-GST; Pharmacia). The puri®ed antigens were then dialysed against 50 mM Tris (pH 8)/10% glycerol. The two shorter antigens were expressed and puri®ed from the soluble fraction by anity chromatography using glutathione-bound agarose (Sigma). Protein concentrations were determined by a Bradford dye binding assay.
1896
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
2.6. Preparation of Toxoplasma tachyzoite antigen Tachyzoites of T. gondii were prepared and the soluble fraction was retained after disruption by sonication, as described previously [21]. Protein concentrations were determined by a Bradford dye binding assay. 2.7. Sera The 188 T. gondii-positive sera and 81 T. gondii-negative sera were obtained from 269 patients who were investigated at the Flinders Medical Centre (Adelaide) in 1995 for a variety of conditions in which toxoplasmosis may be important. Discrimination of positive and negative sera was based on a routine diagnostic ELISA [22]. Positive reference sample was pooled from six positive sera and a negative reference sample was pooled from six negative sera. 2.8. ELISA Antigens were diluted in 0.1 M carbonate/bicarbonate buer (pH 9.6). Flat-bottom ELISA plates (Costar) were coated with 100 ml of antigen solution per well, incubated overnight at 48C and then washed four times with PBS/0.03% Tween20 (PBST). The sensitised wells were then blocked with 2.5% w/v skim milk powder in PBST and incubated at 378C for 1 h, followed by four washes with PBST. Serum samples were diluted in 1% w/v skim milk powder in PBST and 100 ml added to each respective well. The plates were incubated and washed as before. Anti-human Ig conjugated to alkaline phosphatase (Dako) was diluted in 1% w/v skim milk powder and 100 ml added to each well. The plates were incubated and washed as before. One hundred microlitres of 1 mg/ml p-nitrophenyl phosphate substrate (Sigma) dissolved in 10% diethylamine buer was added to each well and incubated at 228C in the dark. The O.D. values were measured at 405 nm against a control well that had not received serum or conjugate.
2.9. Optimisation Chequerboard titrations of antigen, primary pooled sera and conjugate were performed as outlined previously [22], to determine the optimum conditions which discriminate between positive and negative sera. To account for any possible reactivity to the GST portion of the recombinant NTPase antigens, a GST antigen well containing the quantity of GST attached to the NTPase antigen was included for every sample. The absorbance of this GST well was subtracted from the recombinant NTPase absorbance for every sample tested. 2.10. PAGE and immunoblot analysis Immunoblot analysis of sera A±E was performed twice. Recombinant proteins were electrophoresed by SDS±PAGE [23] using 4±20% Tris glycine gels under denaturing conditions. Two micrograms of each isoform were loaded and after electrophoresis proteins were transferred to polyvinylidene ¯uoride membrane by electroblotting. The membranes were stained with Ponceau S to ensure that equal amounts of each isoform had been transferred. Membranes were then blocked for 1 h in PBST/5% w/v skim milk powder. After washing with PBST, membranes were incubated with primary sera diluted 1:400 for 1 h. The membranes were washed with PBST, followed by incubation with the antihuman Ig conjugated to alkaline phosphatase (Dako) diluted 1:1000. The membranes were developed using nitroblue tetrazolium/BCIP (Sigma) dissolved in alkaline phosphatase buer (100 mM Tris, pH 9.5, 5 mM MgCl2, 100 mM NaCl). 2.11. Densitometry Membranes were scanned with a Microtek ScanmakerIII ¯atbed scanner. The band intensities (O.D.mm2) were measured using Molecular Analyst version 2.01 (Biorad). After subtraction of backgound, the band intensity for each isoform was expressed as percentage of the total
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
band intensity (O.D.mm2 2 O.D.mm NTPaseII).
1897
NTPaseI +
2.12. Synthetic peptides Synthetic peptides (length = 20 aa; NTPaseI: H±CKAPMIVTGGGMLAAINTLK±OH; NTPaseII: H±CKAPMFITGREMLASIDTLK± OH) encompassing the 12-aa region of variability were manufactured commercially (Chiron Mimotopes). For the ELISA studies, 0.25 mg of the synthetic peptides was coated to each well, and the ELISA performed as described above. 3. Results 3.1. PCR and cloning An NTPase gene fragment was ampli®ed from T. gondii RH strain genomic DNA, using primers NTPfor and NTPrev. A 1603-bp fragment resulted (data not shown) which, after puri®cation, was digested with EcoR1 to liberate a 1571-bp DNA fragment, representing aa 3±526, or 84% of the protein. After cloning this fragment and determining correct orientation, the isoforms were selected by digesting representative clones with BamH1. NTPaseI clones (NTPaseIpGEX1lT) failed to cut, whilst NTPaseII clones (NTPaseIIpGEX1lT) were digested to reveal two bands of 5408 and 1135 bp (Fig. 1). A 239-bp fragment representing aa 445± 526 and a 1452-bp fragment representing aa 3±
Fig. 1. Identi®cation of nucleoside triphosphate hydrolase isoform by BamH1 digestion of recombinant pGEX/nucleoside triphosphate hydrolase clones. Lane 1: 100-bp DNA ladder. Lane 2: BamH1 digested recombinant pGEX clone containing the nucleoside triphosphate hydrolase I isoform (NTPaseIpGEX1lT). Lane 3: BamHI digested pGEX recombinant clone containing the nucleoside triphosphate hydrolase II isoform (NTPaseIIpGEX1lT).
487 were subcloned from NTPaseIpGEX1lT, to yield recombinant plasmids NTPaseIs-pGEX1lT and NTPaseIctermDel-pGEX1lT, respectively.
Table 1 Absorbances of ®ve sera reacting positively to either nucleoside triphosphate hydrolase I or II in the ELISA Isoform NTPaseI
NTPaseII
Sample
Absorbance
Result
Absorbance
Result
A B C D E
0.33620.042 0.44120.097 0.18020.048 0.15920.031 0.16020.008
Positive Positive Positive Negative Negative
0.11720.016 0.20820.016 0.17920.012 0.26320.016 0.23020.016
Negative Negative Negative Positive Positive
1898
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
3.2. Expression The NTPaseI and NTPaseII clones were expressed in E. coli. The resulting fusion proteins were always found in the insoluble fraction in the form of inclusion bodies after cell lysis (data not shown). This was found to be independent of temperature of expression or IPTG concentration. Furthermore, it was noted that the E. coli heat shock protein `DNA K' was also present in the insoluble fraction (data not shown). The insoluble inclusion bodies were able to be puri®ed by centrifugation, and after solubilisation in 8 M urea (pH 8), were further puri®ed by anion exchange chromatography under denaturing conditions to remove DNA K. After renaturation, both puri®ed NTPase isoforms were determined to have an apparent Mr of 88 kDa (data not shown). Expression of NTPaseIs and NTPaseIctermDel truncations yielded proteins of 36 and 84 kDa, respectively (data not shown). 3.3. Optimisation of ELISA A working dilution of 1:50 of primary sera and 1:500 of the conjugate was determined by chequerboard assay with the T. gondii soluble fraction antigen (1 mg per well). The two NTPase isoforms (0.5 mg per well) were determined to have a working dilution of 1:50 for the primary sera and 1:250 for the conjugate. 3.4. Test sera The mean + 3 S.D. of the absorbances from the 83 negative sera were used to determine the value where discrimination between positive and negative samples could be determined. The
mean + 3 S.D. was 0.381 for the T. gondii lysate (positive control antigen), 0.178 for recombinant NTPaseI and 0.222 for recombinant NTPaseII. Samples with an absorbance value greater than the discrimination value were determined as positive. Eighteen percent (34/188) of the known positive sera were determined as positive for the NTPaseI isoform, and 17.5% (33/188) of known positive samples were determined as positive with the NTPaseII isoform. One hundred percent of known positive sera were determined to be positive with T. gondii lysate (positive control antigen). No false positives were detected among the known negative sera for any of the antigens. Interestingly, three samples were exclusively positive to NTPaseI and two samples were exclusively positive for NTPaseII. These ®ve sera were tested twice in duplicate and the results are summarised in Table 1. Of the 36 sera which were determined to be positive by ELISA for at least one isoform, eight sera displayed dierences in antibody reactivity to the isoforms (Fig. 2). 3.5. Immunoblot analysis To con®rm that the dierences detected in the ®ve sera were speci®c to NTPase, immunoblots were performed (Fig. 3). Quantitative dierences were observed in sera D and E, whilst a qualitative dierence was observed in serum A. No signi®cant dierence (P > 0.05) was observed in sera B and C. Because serum A reacted strongly with the NTPaseI isoform and no reaction with the NTPaseII isoform was detected, immunoblots were performed with the truncated forms NTPaseIs and NTPaseIctermDel to isolate the
Fig. 2. Dierential antibody response to nucleoside triphosphate hydrolase isoforms. (A) The 36 sera (X-axis) which were determined to be positive to at least one of the nucleoside triphosphate hydrolase isoforms by ELISA were plotted against the dierence in absorbance at 405 nm (nucleoside triphosphate hydrolase Iÿnucleoside triphosphate hydrolase II). The response types are: ``A'', equal response to each isoform; ``B'', response to nucleoside triphosphate hydrolase I>response to nucleoside triphosphate hydrolase II; ``C'', response to nucleoside triphosphate hydrolase I < response to nucleoside triphosphate hydrolase II. (B) The same 36 sera (X-axis) were plotted against the standard deviation (n = 2) of the absorbances of each sera to the two isoforms. The mean S.D. is 0.050. LCL = lower con®dence limit; UCL = upper con®dence limit. Eight sera, corresponding to the response type B and C sera, are signi®cantly greater than the mean S.D. Twenty-eight sera corresponding to the type A response have standard deviations less than or equal to the mean S.D. (®gure opposite)
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
Fig. 2.
1899
1900
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
Fig. 3. Immunoblot analysis of the ®ve sera which were positive for one isoform only by ELISA. A±E represent the ®ve sera A±E. Recombinant nucleoside triphosphate hydrolase isoforms were run in triplicate on 4±20% Tris±glycine gels under reducing conditions and transferred to polyvinylidene ¯uoride membrane. Lane I: recombinant nucleoside triphosphate hydrolase I (2 mg). Lane II: recombinant nucleoside triphosphate hydrolase II (2 mg). O.D. values (data not shown) were determined by densitometry, as described above.
region of this speci®c reactivity. Reactivity was observed in the NTPaseIs truncated protein (Fig. 4), which represents aa 445±526. No reactivity was observed in the NTPaseIctermDel (aa 3±487) truncated protein or the GST control, thus con®rming that the only area of antibody
Fig. 4. Localisation of reactivity to nucleoside triphosphate hydrolase immunoblotted with serum A. Lane 1: nucleoside triphosphate hydrolase I (aa 3±526). Lane 2: NTPaseIctermDel (aa 3±487). Lane 3: NTPaseIs (aa 445± 526). Lane 4: glutathione S-transferase control.
recognition on the NTPaseI isoform occurs within 38 C-terminal amino acids (aa 488±526).
3.6. Antibody reactivity to synthetic peptides The mean + 3 S.D. of the absorbances from 12 T. gondii-negative samples diluted 1:50 was determined for each peptide. The mean + 3 S.D. for the peptides KCAPMIVTGGGMLAAINTLK and KCAPMFITGREMLASIDTLK were 0.260 and 0.255, respectively. The samples which were determined previously to be NTPase positive were then assayed for reactivity to the synthetic peptides. Five of the 36 positive samples were identi®ed as having absorbances greater than the mean + 3 S.D. for both of the peptides. These ®ve sera, as well as the negative reference sera, were serially diluted to con®rm speci®c antibody reactivity (Fig. 5). Four of these sera (panels 2± 5) showed equal reactivity to both peptides, whilst one serum (panel 1, serum A described previously) displayed dierential reactivity, where the antibody response to the peptide CKAPMIVTGGGMLAAINTLK was signi®cantly greater (P < 0.05) than the response to the corresponding peptide CKAPMFITGREMLASIDTLK. A Toxoplasma pooled negative sample (described previously) showed no response over the dilution range.
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
1901
Fig. 5. Antibody response to synthetic peptides. Panels 1±5: serial dilution of sera which were found to have absorbances greater than the mean + 3 S.D. of 12 Toxoplasma-negative sera against peptides (0.25 mg per well). CKAPMIVTGGGMLAAINTLK (w) and KCAPMFITGREMLASIDTLK (t). Panel 6: serial dilution of pooled Toxoplasma-negative reference sera against the peptides.
1902
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
4. Discussion The NTPase of T. gondii has DTT-dependent hydrolytic properties which were thought to be unique to the genus. The ®nding that T. gondii possessed two NTPase isoforms, and the classi®cation of these isoforms as NTPaseI and NTPaseII based upon varying substrate hydrolysis, was made more interesting by the fact that only virulent strains possessed the highly active NTPaseI [12]. Recently, Asai et al. [15] have demonstrated that the closely related genus Neospora also shares this unique DTT-dependent NTPase activity. It is interesting to note that N. caninum (mouse virulent) was found to possess NTPaseI activity. The reason for the existence of two NTPase isoforms and explanations for a correlation of NTPaseI with mouse virulent strains are unclear. The role of antibody response to the two isoforms is also unknown and we believe that our study reported here is the ®rst to examine this interesting biological question. Both qualitative and quantitative dierences in antibody reactivity to the two NTPase isoforms of T. gondii were found to exist in sera from naturally infected humans. From a diagnostic point of view, the recombinant NTPase ELISA was able to identify only 18% of sera with anti-NTPase antibody in patient sera known to contain anti-Toxoplasma antibody. Asai et al. [17] demonstrated that antiNTPase antibodies detected by the use of native NTPase in an indirect ELISA correlated with detection of anti-Toxoplasma antibody. Tenter and Johnson [22] showed previously that a 50 kDa recombinant NTPase ELISA could detect anti-NTPase antibody in only 7% of known T. gondii-positive sera. The longer recombinant NTPase isoforms used here achieved a greater sensitivity of detection, probably due to a greater representation of epitopes. However, the sensitivities achieved by recombinant NTPase antigens here and in the previous study by Tenter and Johnson [22] are substantially less than those achieved in the study of native NTPase ELISA of Asai et al. [17]. There are 103 C-terminal amino acids in the native NTPase isoforms which
are not represented in the recombinant isoforms studied here. It is possible that there may be epitopes in this region that may also contribute to antigenicity. Problems with expression should also be considered, as conformationally correct recombinant proteins can be dicult to express in the reducing environment of the E. coli cytoplasm, especially if the recombinant protein contains disulphide bonds. Silverman et al. [24] suggested that only about 5% of the NTPase in the vacuole space is in its reduced (insoluble) active form, and perhaps NTPase is most antigenic in its oxidised (soluble) inactive form. This hypothesis is supported by the fact that the native NTPase antigen used by Asai et al. [17] was extracted from the soluble fraction of T. gondii tachyzoites. Alternatively, recombinant NTPase expressed in E. coli may not be conformationally correct due to incorrect post-translational modi®cations such as incorrect glycosylation, since the NTPase isoforms have one potential N-linked glycosylation site at aa 432 [16]. Determination of enzyme activity could con®rm whether conformationally correct recombinants have been expressed, whilst the use of eukaryotic expression systems such as yeast or baculovirus may be useful in the expression of NTPase with correct conformation for diagnostic purposes. Although the assay identi®ed anti-NTPase antibody in only about 18% of patients known to possess anti-Toxoplasma antibody, ®ve sera were identi®ed which were positive for one of the isoforms only. Immunoblotting was used to determine the speci®city of these reactions, and both quantitative and qualitative dierences in antibody response to the isoforms were noted (Fig. 3). The qualitative dierence in reactivity to NTPase observed in serum A has been isolated to a region of 81 aa (aa 445±526) at the C-terminus of NTPase. Deletion of this C-terminus region (aa 487±526) resulted in the complete loss of antibody reactivity (Fig. 4), con®rming that the only area of antibody recognition resides in this 38-aa C-terminus region. It is interesting to note from the previously published sequences [12, 16] that there are only six amino acids in a stretch of 12 that vary between the two
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
isoforms in this region. This 12-aa stretch (residues 488±499, IVTGGGMLAAIN: NTPaseI; FITGREMLASID: NTPaseII), which contains three charged amino acids in NTPaseI but is uncharged in NTPaseII, represents the region of greatest variability between the two isoforms, and may form part of a dierential epitope. This ®nding is made more interesting by a recent study which has identi®ed this region as being responsible for substrate binding and for the control of substrate speci®city [25]. Antibody reactivity to this region may possibly play some role in regulating NTPase activity by inhibiting the binding of substrate. Because our immunoblot analysis was performed under denaturing conditions, the epitopes recognised are probably linear rather than conformational. Epitope mapping of this region by the use of overlapping synthetic peptides should con®rm the exact epitope sequence, and DNA sequence analysis of this region in a number of virulent and avirulent strains may further strengthen the correlation between the NTPase isoform and virulence. Although serum A was the only serum to demonstrate qualitative dierences in response to the NTPase isoforms, four other sera were identi®ed in which a positive ELISA reaction was exclusive to one isoform only. Sera D and E were found to react exclusively to NTPaseII in ELISA. These results were con®rmed by the signi®cantly greater band intensities (P < 0.05) observed for the NTPaseII isoform by immunoblotting. However, unlike serum A, the dierences observed by immunoblotting were quantitative. The presence of the GST fusion partner may explain why the dierences were quantitative rather than qualitative; however, the possibility that a dierential epitope exists in NTPaseII cannot be excluded. It was surprising to note that immunoblotting of sera B and C did not con®rm the ELISA results. For example, serum B reacted to NTPaseI with greater than twice the absorbance it did to NTPaseII by ELISA, yet no signi®cant dierence between the isoforms was detected by immunoblotting. This would suggest that reactivity to conformational epitopes has occurred in the ELISA, but this was
1903
not observed in the immunoblot analysis due to the denaturing conditions used. It should also be noted that, although serum C is exclusively positive to NTPaseI, the absorbance values of serum C for each isoform by ELISA are almost identical and there is no signi®cant dierence in reactivity to the isoforms observed by immunoblotting. This highlights a problem associated with classifying a serum as ELISA positive when it displays an absorbance close to the positive cut-o value. When looking at the dierences in absorbance between the two isoforms detected by ELISA (Fig. 2) for the 36 samples which were positive for at least one isoform, it becomes clear that the patient sera can be classi®ed into three types of response. Firstly, sera which show equal response for each isoform (`type A' response). Secondly, sera which show a signi®cantly greater response to the NTPaseI isoform (`type B' response) and, ®nally, sera which show a signi®cantly greater response to the NTPaseII isoform (`type C' response). The response for most sera was found to fall into the type A category, where absorbances were very similar for both isoforms. This type of response can be explained by the high degree of homology between the two isoforms. However, even when looking at this type A response, sera tend to exhibit a greater reactivity to the NTPaseII isoform. This may be indicative of a higher prevalence of avirulent strains causing infection among these patients. Support for this observation comes from a recent study which employed a nested PCR to type clinical Toxoplasma isolates and found that 81% were of the avirulent (type II) lineage [26]. The bias toward higher reactivity to the NTPaseI isoform observed in the type B response may be due to infection by virulent Toxoplasma strains where the increased absorbance may be due to antibody binding to the dierential epitope described previously. Conversely, the higher reactivity to the NTPaseII isoform observed in the type C response may be due to infection by avirulent strains, where once again the increased absorbance may be caused by a dierential epitope. Although dierences were observed in antibody reactivity to the two recombinant isoforms,
1904
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
it is unclear what regions of the NTPase isoforms contribute to this dierential recognition by human sera. The data from immunoblotting with serum A suggest that dierential recognition occurs within 38 C-terminal amino acids. In order to test the possibility that sera other than serum A display dierential reactivity in this region, we used 20-residue synthetic peptides encompassing this region of variability to study the 36 NTPase-positive sera identi®ed previously by the recombinant ELISA. Five sera were identi®ed with an absorbance greater than the mean + 3 S.D. of 12 Toxoplasma-negative samples. When these sera were serially diluted against the peptides representing each isoform, four sera reacted equally to both isoforms, whilst one serum sample (serum A, described previously) displayed dierential reactivity (Fig. 5). Although the antibody response of serum A to the NTPaseI peptide was signi®cantly greater than to the NTPaseII peptide, some cross-reactivity was observed. The results con®rm that the dierential reactivity in serum A occurs at the 12-aa region of greatest variability. However, this region does not appear to be involved in dierential antibody reactivity for any of the other 35 NTPase-positive sera, although it is clearly a recognised epitope for ®ve of these sera. It may be possible that the IVTGGGMLAAIN sequence (NTPaseI) is the homologous epitope stimulating the antibody reponse, whereas the FITGREMLASID sequence (NTPaseII) is a heterologous epitope, and is thus the cross-reactive epitope. Further studies of this epitope region using sera from virulent and avirulent experimental infections will determine if there is a correlation between reactivity to the IVTGGGMLAAIN epitope and infection by virulent Toxoplasma strains. A correlation of congenital toxoplasmosis with infection by virulent strains has been observed previously [27], and therefore antibodies from naturally infected humans that recognise an epitope that correlates with infection by virulent T. gondii strains would have important diagnostic and epidemiological value.
References [1] Remington JS, Desmonts G. In: Remington JS, Klein JO, editors. Infectious diseases of the fetus and newborn infant. Philadelphia. W.B. Saunders, 1990; 89±195. [2] Johnson AM. In: Long PL, editor. Coccidiosis of man and domestic animals. Boca Raton, FL: CRC Press, 1990; 121±135. [3] Johnson AM. Strain-dependent route of challenge-dependent, murine susceptibility to toxoplasmosis. Z Parasitenk 1984;70:303±9. [4] Dubey JP. Advances in the life cycle of Toxoplasma gondii. Int J Parasitol 1998;28:1019±24. [5] Johnson AM. Speculation on possible lifecycles for the clonal lineages in the genus Toxoplasma. Parasitol Today 1997;13:393±7. [6] Parmley SF, Gross U, Sucharczuk A, Windeck T, Sgariato GD, Remington JS. Two alleles of the gene encoding surface antigen P22 in 25 strains of Toxoplasma gondii. J Parasitol 1994;80:293±301. [7] Guo ZG, Johnson AM. Genetic characterization of Toxoplasma gondii strains by random ampli®ed polymorphic DNA polymerase chain reaction. Parasitology 1995;111:127±32. [8] Guo ZG, Gross U, Johnson AM. Toxoplasma gondii virulence markers identi®ed by RAPD PCR. Parasitol Res 1997;83:453±8. [9] Ware PL, Kasper LH. Strain-speci®c antigens of Toxoplasma gondii. Infect Immun 1987;55:778±83. [10] 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 1987;157:7±13. [11] Meisel R, Stachelhaus S, Mevelec MN, Reichmann G, Dubremetz JF, Fischer HG. Identi®cation of two alleles in the GRA4 locus of Toxoplasma gondii determining a dierential epitope which allows discrimination of type I versus type II and III strains. Mol Biochem Parasitol 1996;81:259±63. [12] Asai T, Sibley LD, Okabayashi H, Takeuchi T. Biochemical and molecular characterization of nucleoside triphosphate hydrolase isozymes from the parasitic protozoan Toxoplasma gondii. J Biol Chem 1995;270:11391± 7. [13] Asai T, O'Sullivan WJ, Tatibana M. A potent nucleoside triphosphate hydrolase from the parasitic protozoan Toxoplasma gondii. J Biol Chem 1983;258:6816±22. [14] Asai T, Kanawaza T, Kobayashi S, Takeuchi T, Kim T. Do protozoa conceal a high potency of nucleoside triphosphate hydrolysis in Toxoplasma gondii?. Comp Biochem Physiol 1986;85B:365±7. [15] Asai T, Howe DK, Nakajima K, Nozaki T, Takeuchi T, Sibley LD. Neospora caninum: tachyzoites express a potent type-1 nucleoside triphosphate hydrolase activity. Exp Parasitol 1998;90:277±85. [16] Bermudes D, Peck KR, A®® MA, Beckers CJM, Joiner KA. Tandemly repeated genes encode nucleoside tripho-
M.S. Johnson et al. / International Journal for Parasitology 29 (1999) 1893±1905
[17]
[18]
[19] [20] [21]
sphate hydrolase isoforms secreted into the parasitophorous vacuole of Toxoplasma gondii. J Biol Chem 1994;269:29252±60. Asai T, Mizuno F, Kojima S, Takeuchi T, Kobayashi A, Suzuki Y. High correlation in antibody titers between Sabin±Feldman dye test and an enzyme linked immunosorbent assay detecting immunoglobulin G antibodies to the nucleoside triphosphate hydrolase of Toxoplasma gondii. J Clin Microbiol 1992;30:1291±3. Asai T, Kim T, Kobayashi M, Kojima S. Detection of nucleoside triphosphate hydrolase as a circulating antigen in the sera of mice infected with Toxoplasma gondii. Infect Immun 1987;55:1332±55. Johnson AM, Dubey JP, Dame JB. Puri®cation and characterization of Toxoplasma gondii tachyzoite DNA. Aust J Exp Biol Med Sci 1986;64:351±5. Lin KH, Cheng SY. An ecient method to purify active eukaryotic proteins from inclusion bodies in Escherichia coli. Biotechniques 1991;11:748±52. Johnson AM, Gu QM, Roberts H. Antibody patterns in the serological diagnosis of acute lymphadenopathic toxoplasmosis. Aust NZ J Med 1987;17:430±4.
1905
[22] Tenter AM, Johnson AM. Human antibody response to the nucleoside triphosphate hydrolase of Toxoplasma gondii. J Immunoassay 1990;11:579±90. [23] Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680±5. [24] Silverman JA, Huilin Q, Riehl A, Beckers C, Nakaar V, Joiner KA. Induced activation of the Toxoplasma gondii nucleoside triphosphate hydrolase leads to depletion of host cell ATP levels and rapid exit of intracellular parasites from infected cells. J Biol Chem 1998;273:12352±9. [25] Nakaar V, Beckers CJM, Polotsky V, Joiner KA. Basis for substrate speci®city of the Toxoplasma gondii nucleoside triphosphate hydrolase. Mol Biochem Parasitol 1998;97:209±20. [26] 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. [27] Howe DK, Sibley LD. Toxoplasma gondii comprises three clonal lineages: Correlation of parasite genotype with human disease. J Infect Dis 1995;172:1561±6.