High-Throughput Growth Assay for Toxoplasma gondii Using Yellow Fluorescent Protein

Abstract

A high-throughput growth assay for the protozoan parasite Toxoplasma gondii was developed based on a highly fluorescent transgenic parasite line. These parasites are stably transfected with a tandem yellow fluorescent protein (YFP) and are 1,000 times more fluorescent than the wild type. Parasites were inoculated in optical-bottom 384-well culture plates containing a confluent monolayer of host cells, and growth was monitored by using a fluorescence plate reader. The signal was linearly correlated with parasite numbers over a wide array. Direct comparison of the YFP growth assay with the β-galactosidase growth assay by using parasites expressing both reporters demonstrated that the assays' sensitivities were comparable but that the accuracy of the YFP assay was higher, especially at higher numbers of parasites per well. Determination of the 50%-inhibitory concentrations of three known growth-inhibiting drugs (cytochalasin D, pyrimethamine, and clindamycin) resulted in values comparable to published data. The delayed parasite death kinetics of clindamycin could be measured without modification of the assay, making this assay very versatile. Additionally, the temperature-dependent effect of pyrimethamine was assayed in both wild-type and engineered drug-resistant parasites. Lastly, the development of mycophenolic acid resistance after transfection of a resistance gene in T. gondii was followed. In conclusion, the YFP growth assay limits pipetting steps to a minimum, is highly versatile and amendable to automation, and should enable rapid screening of compounds to fulfill the need for more efficient and less toxic antiparasitic drugs.

Toxoplasma gondii is a widespread apicomplexan parasite able to infect virtually all warm-blooded vertebrates (45). Twenty-two percent of the U.S. population is infected, but severe disease in adults is mainly limited to immunosuppressed patients. In patients with acquired immunodeficiency syndrome (AIDS), T. gondii causes a life-threatening opportunistic infection, with Toxoplasma encephalitis as its most severe manifestation (23, 24). T. gondii is also known to cause congenital infection and is among the pathogens with the highest incidence of complications in pregnancies (7, 31). Despite its clinical importance, only very few therapeutic drugs against T. gondii are available, all of which target the rapidly dividing tachyzoites, leaving the dormant encysted bradyzoite stage unaffected (15, 22, 42). As Toxoplasma encephalitis in AIDS patients results mostly from the reactivation of chronic stages (23), this is an important shortcoming, severely compromising the management of toxoplasmosis in these patients. Moreover, drug toxicity due to sulfadiazine hypersensitivity further complicates the life-long prophylactic treatment of immunosuppressed patients (15, 22) as well as treatment during pregnancy. New drugs with broader efficacy and lower toxicity are clearly needed.

Several microtiter plate-based growth assays for T. gondii drug screens have been developed in the past. Parasite growth can be measured by following the incorporation of radioactive uracil (30), by using T. gondii-specific antibodies in an enzyme-linked immunosorbent assay format (6, 26), or by using transgenic expression of the bacterial β-galactosidase reporter gene (25). A fluorescence activated cell sorting (FACS)-based assay to monitor parasite growth has also been developed (14). As with all assays, these have their inherent strengths and limitations. Some of the drawbacks are the use of radioactive compounds or the need for external factors to visualize the signal, requiring several pipetting steps. Moreover, most of these assays limit the readout to a single time point (a recent modification of the [³H]uracil incorporation assay now permits multiple sampling) (27).

In recent years, the use of green fluorescent protein (GFP) in drug screens has gained interest, since GFP does not require external factors and is stable throughout a wide spectrum of conditions (43). Transgenic microorganisms and viruses stably expressing GFP have been used in several drug discovery efforts targeting HIV (16) and Mycobacterium (46) and Leishmania (19) organisms. However, all of these assays employ FACS to assess drug activity and as such are limited in the number of compounds that can be tested during a given time period. Combining GFP with a microtiter format enhances the throughput of the screening system considerably and has recently been applied to determine the activities of anticancer drugs (37, 47), but to our knowledge it has not as yet been applied for antimicrobial drug screens.

GFP expression has been used in a number of studies investigating the cell biology of T. gondii (17, 18, 29, 38-40). In this report, we describe a T. gondii parasite line stably expressing a yellow version of GFP (YFP) (5) which results in exceptionally bright fluorescence. Using this parasite line, we have developed an in vitro growth assay. The capacity of this new assay was evaluated by direct comparison with a currently widely used growth assay using β-galactosidase (25). We show the utility of this fluorescent growth assay for drug testing with several known antiparasitic drugs. Finally, we explored the use of this assay to study drug resistance in T. gondii.

MATERIALS AND METHODS

Antimicrobial agents.

All drugs were purchased from Sigma (St. Louis, Mo.) with the exception of clindamycin, which was purchased from ICN (Irvine, Calif.). Drugs were diluted into culture medium from the following stocks: 10 mM pyrimethamine in ethanol, 46 mg of clindamycin per ml in ddH₂O, and 1 mM cytochalasin D in dimethyl sulfoxide (all stored at −20°C). Stocks of mycophenolic acid (25 mg/ml in ethanol) and xanthine (50 mg/ml in 0.5 M KOH) were stored at 4°C.

Plasmids.

The tandem YFP plasmid, tubYFP-YFP/sagCAT (Fig. 1A), was constructed by cloning the PCR-amplified coding sequence of YFP flanked by 5′ BglII and 3′ AvrII sites by using primers yfp-bgl2 (AGCTagatctAAAATGGTGAGCAAGGGCGAGGAGC; the BglII site is indicated in lowercase letters and the YFP open reading frame is underlined) and yfp-avr2 (CAGTcctaggCTTGTACAGCTCGTCCATGCCG) replacing the α-tubulin coding sequence of ptub-TUB-YFP/sagCAT (40). The resulting plasmid contains an α-tubulin promoter (20) separated from the first YFP coding sequence by a BglII site, an in-frame AvrII site separating both YFP coding sequences, a 3′ untranslated region derived from the T. gondii dihydrofolate reductase-thymidylate synthase (DHFR-TS) gene (32), and a NotI site separating these sequences from the pKS+ plasmid backbone (Stratagene, La Jolla, Calif.). A chloramphenicol acetyltransferase-selectable marker under the control of 5′ and 3′ sequences derived from the T. gondii SAG1/P30 gene (2, 20, 36) is present in front of the tubulin promoter. The plasmid is available upon request.

FIG. 1.

Expression of YFP-YFP yields highly fluorescent parasites. (A) Graphical representation of the ptubYFP-YFP/sagCAT plasmid. (B) Fluorescence microscopic image of tachyzoites stably transfected with plasmid ptubYFP-YFP/sagCAT (2F-1 YFP₂); the fluorescent marker evenly fills the parasites' cytoplasm. (C and D) Wild-type RH parasites (C) as well as 2F-1 YFP₂ transgenics (D) were analyzed by FACS. Cells were excited with the 488-nm line of an argon laser, and a 530/540-nm filter was used in the emission; 50,000 events are shown here. Transgenic parasites display a 1,000-fold-higher fluorescence.

The pHXGPRT-DHFR plasmid was constructed by filling in the ends of the 6.7-kb BamHI genomic DNA fragment containing the complete hypoxanthine-xanthine-guanine phosphoribosyltransferase (HXGPRT) gene (pTgHXGPRTg8 clone, a kind gift of D. Roos, University of Pennsylvania [10]) and cloning into the EcoRV site of the plasmid TscABP-RV (41). This plasmid contains the DHFR-TSm2m3 gene carrying two mutations conferring pyrimethamine resistance (9). Both HXGPRT and DHFR-TSm2m3 can be used to select for stable integration by using mycophenolic acid-xanthine and pyrimethamine, respectively.

Cells and parasites.

RH-strain parasites were grown in human foreskin fibroblast (HFF) cells and transfected by electroporation as previously described (33). T. gondii 2F-1 YFP₂ was derived from the 2F-1 parasite stably expressing the β-galactosidase gene (2F-1 was a kind gift from V. B. Carruthers, John Hopkins University, Baltimore, Md.). This parasite was stably transfected with the tubYFP-YFP/sagCAT plasmid by chloramphenicol selection and cloned by limiting dilution. The parasites named RH-HX-KO-YFP₂-DHFR(m2m3) are derived from the HXGPRT knockout RH strain (10) and were successively stably transfected with tubYFP-YFP/sagCAT (chloramphenicol selection) and pHXGPRT-DHFR (mycophenolic acid-xanthine selection).

Growth monitoring by fluorescence.

Black, 384-well, tissue culture-treated plates with optical bottoms were purchased from Falcon/Becton-Dickinson (Franklin Lakes, N.J.). Cells were seeded by using an automatic liquid dispenser (Q-fill; Genetix, Dorset, United Kingdom) in a volume of 50 μl per well. Freshly lysed parasites were filtered through a 3-μm polycarbonate filter, centrifuged, and resuspended in parasite culture medium without phenol red (Gibco BRL Life Technologies, Rockville, Md.). Quadruple, twofold serial parasite dilutions (40 μl/well, starting from 2,000 parasites per well) in the presence of drugs were carried out by using a multichannel pipette. Plates were kept in a humidified incubator with 5% CO₂ at 37°C (or 35 or 40°C where indicated) and read daily in a BMG Fluostar fluorescent plate reader (Offenburg, Germany). To preserve sterility, the plates were read with covered lids, and both excitation (510 nm) and emission (540 nm) were read from the bottom. For excitation, a single flash from a Xenon lamp was used for each well, and emission signals were recorded with a gain setting of 60. A script written in Microsoft Excel (available upon request) automatically plotted time (in days) against averaged fluorescence (from quadruple data) expressed as percentage positivity (PP) to correct for day-to-day and plate-to-plate variations. A PP of 0% represents the background level, and 100% PP represents the maximum fluorescence signal observed in the particular experiment. Positive and negative controls were averaged from quadruple measurements performed simultaneously on the experimental plate. Values for fifty-percent inhibitory concentrations (IC₅₀) were extrapolated from the growth curve data by plotting the drug concentration against the growth rate. The growth rate was defined as 1 divided by the day when a plateau in the growth curve is reached (0 when no growth was observed). Transfection of the pHXGPRT-DHFR plasmid into RH-HX-KO-YFP₂ parasites was studied in the growth assay under increasing concentrations of mycophenolic acid and xanthine. Directly after electroporation, parasites were twofold serially diluted, starting with 8,000 parasites per well, in quadruplicate. Fluorescence was read daily, and data from the four identical wells were averaged.

Monitoring growth by using β-galactosidase.

Parasites were handled in the same manner as that used for parasites for the YFP growth monitoring until the desired readout day. Then chlorophenol red-β-galactopyranoside (CPRG; Boehringer Mannheim, Indianapolis, Ind.) was added to a final concentration of 0.5 mM by the addition of 4.5 μl of a 4.5 mM CPRG stock prepared in culture medium without phenol red. After substrate addition, the color development was measured by reading absorbance (570 nm) on a BMG Fluostar plate reader in absorbance mode at several time points (10, 30, 60, and 90 min). Absorbance data were expressed as PP values, essentially as described for the fluorescence data.

Microscopy.

Parasites in 384-well plates were directly studied by fluorescence microscopy by using a standard fluorescein isothiocyanate filter set (excitation band pass, 480/40 nm, and emission band pass, 527/30 nm) with a 10× objective on a Leica DM-IRBE microscope (Wetzlar, Germany). Images were recorded with a Hamamatsu C4742-95 cooled digital camera (Bridgewater, N.J.) and adjusted for contrast with Openlab software version 3.0.3 (Improvision, Quincy, Mass.). The composite images of whole wells shown in Fig. 2A were assembled by using Canvas (Daneba, Miami, Fla.) and originated from nine partly overlapping images covering a single well.

FIG. 2.

Parasite fluorescence can be used to accurately measure parasite numbers in microtiter plates. Twofold serial dilutions of 2F-1 YFP₂ parasites were used to infect 384-well plates containing HFF cells grown to confluence. Fluorescence (YFP) was read daily with a plate reader. β-Galactosidase activity was detected by measuring the change in absorbance at 540 nm due to CPRG substrate conversion on the particular readout day. (A) Correlation between fluorescence intensity and infection dose after 7 days of incubation. The inserts show a fluorescence micrograph of the well corresponding to the indicated plate reader data point; rfu denotes relative fluorescence units. (B) Direct comparison of results of the β-galactosidase assay (after 5 days of incubation and 60 min of CPRG staining) and the YFP assay (after 6 days of incubation). Fluorescence and absorbance data from quadruple measurements were automatically averaged and converted into PP data and plotted against parasite inoculum numbers per well. PP for both assays was calculated by first subtracting the averages of the background signals read from four independent wells without parasites, which were subsequently set to 0% PP, followed by setting the maximum measured signal to 100% PP and expressing all other values relative to these two reference points.

FACS analysis.

Freshly lysed parasites were filtered through a 3-μm filter, spun down, and resuspended in phosphate-buffered saline at 1 × 10⁶ parasites/ml. Fifty thousand data points were collected on a MO-FLO instrument (Cytomation, Fort Collins, Colo.) with a 488-nm argon laser for excitation and an emission filter with a band pass of 530/540 nm. The data were managed with Summit version 3.1 (Cytomation) and prepared for publication with Flowjo version 4.0.2 (Treestar, San Carlos, Calif.).

RESULTS

YFP-YFP-expressing parasites are highly fluorescent.

T. gondii tachyzoites were transfected with plasmid tubYFP-YFP/sagCAT, and stable clonal lines were selected. This construct encodes a tandem YFP-YFP protein that is expressed under the control of the strong α-tubulin promoter as shown in Fig. 1A. When analyzed by FACS, the transgenic line 2F-1 YFP₂ (Fig. 1D) displayed a 1,000-fold-higher fluorescence relative to that of the wild-type RH parasites (Fig. 1C). It must be noted that the excitation was not optimal in this experiment, as the 488-nm line of the argon laser that was used is considerably lower than the excitation maximum of YFP at 510 nm (28). The true fluorescence intensity of the 2F-l YFP₂ parasites is therefore likely to be even higher than shown in Fig. 1C. By using fluorescence microscopy, it was determined that this strong fluorescence was located in the cytoplasm of the parasite, as is to be expected in the absence of any specific targeting information (Fig. 1B).

Calibration of the fluorescence growth assay.

We reasoned that this new line might be sufficiently fluorescent to be detectable with a fluorescence plate reader. We tested several combinations of microtiter plates, fluorescence readers, and filter sets. By using plates with a thin special optics bottom and a plate reader with fiberglass coupling (excitation, 510 nm, and emission, 540 nm), a minimum number of 5,000 parasites can be detected unambiguously above the background of host-cell autofluorescence (data not shown). To test the correlation between the relative fluorescence intensities and parasite numbers, multiwell cultures were infected with a twofold serial dilution of freshly lysed tachyzoites (starting with a parasite inoculum of 2,000 organisms per well). The relative fluorescence for each well was measured daily by using a plate reader. After 4 days, infections that had been initiated by a single parasite per well had resulted in replications of organisms to numbers high enough to be detected by the fluorescence reader. Seven days were required for these cultures to reach a plateau of maximum fluorescence signal (data not shown). Figure 2A shows a twofold serial dilution ranging from 1,000 to 0 parasites that were used to infect each single well. To estimate the true number of viable parasites that initiated the infection in a particular well, we used fluorescence microscopy. Each well was imaged by fluorescence microcopy, and composite images of selected wells are shown as inserts. We were unable to detect any parasites in wells with background readings in the plate reader. The first well with a positive fluorescence readout shows two fluorescent spots. These spots represent two independent parasite plaques resulting from an initial infection by two parasites (33). Moving up along the curve, a steady increase in the number of plaques can be observed until the host cell monolayer is completely lysed and the well is filled with extracellular parasites. The constant slope of the curve indicates a linear relationship between parasite number and measured fluorescence up to complete lysis of the host cells. We noticed a robust linear readout across the plate with the exception of the two outermost rows of wells. Medium evaporation is higher in these wells and seems to interfere with growth (data not shown). This feature has also been observed when using 96-well plates for T. gondii culture (21) and is likely unavoidable. The outermost wells were therefore routinely excluded from analysis in further assays.

To assess how the fluorescence assay performs relative to other growth assays, a direct comparison was made with the β-galactosidase growth assay (25). For this purpose, we established a parasite line carrying both the tandem YFP and the β-galactosidase transgene. After reading the YFP fluorescence, the β-galactosidase substrate CPRG was added to the wells and the reaction development was read at 10, 30, 60, and 90 min. The assays were performed in parallel on the same plate, and an optimal data series was chosen for each assay so that the comparison represents the optimal performance of each assay (Fig. 2B). To allow reliable comparison between both assays, data of both assays were expressed as PP and plotted in the same graph. Data expression in PP values also corrects for small day-to-day and plate-to-plate variations, resulting in more-standardized data. For this reason, all further data are presented in PP. Both assays showed very similar results. We noted that in our experiments, the standard deviations between individual measurements were consistently smaller with the YFP assay (see the error bars in Fig. 2B), suggesting a higher accuracy for this assay. This is especially the case at higher parasite numbers, as the CPRG signal becomes saturated as it reaches an optical density at 570 nm of greater than 2.5.

Drugs with different parasite inhibition kinetics can be evaluated by using the YFP assay.

In this assay, culture plates are read through the bottom, leaving the lid on; the sterility of the culture is not compromised. By measuring fluorescence daily, we could plot precise growth curves for each 50-μl well. To challenge this capability of the assay, three drugs with different growth-inhibiting kinetics were chosen, namely cytochalasin D, which prevents host cell invasion (8), pyrimethamine, which prevents parasite division (9, 13), and clindamycin, which delays parasite death due to interference with translation in the plastid organelle (1, 4, 11). Growth in the presence of cytochalasin D shows a threshold at a drug concentration of 5.0 μM (IC₅₀, 2.8 μM), above which no fluorescence increase can be observed (Fig. 3A). Some growth is seen in the presence of a 2.5-μM concentration, but the growth does not reach the maximum level that is seen under lower or no drug pressure. As for the invasion-preventing drug cytochalasin D, blocking parasite division by pyrimethamine results in similar repression kinetics (Fig. 3B): no growth occurs with concentrations above 0.6 μM, intermediate growth occurs at 0.3-μM concentrations, and growth is unaffected at lower drug concentrations or in the absence of drug (IC₅₀, 0.52 μM). Clindamycin, however, allows several parasite divisions until it causes lethality, and this is reflected in the growth curves, which are characterized by equal signal increases at all drug concentrations for the first 2 days, after which growth slows down and halts at concentrations of 0.6 ng/ml and above (Fig. 3C) (IC₅₀, 0.45 ng/ml).

FIG. 3.

Fluorescence growth assays reveal distinct growth kinetics for different drugs. One thousand 2F-1 YFP₂ parasites were used to infect each well of a confluent 384-well plate culture. Parasites were incubated for 7 days in the presence of cytochalasin D (A), pyrimethamine (B), and clindamycin (C) at the indicated concentrations, and fluorescence was measured daily by using a plate reader. Growth curves were constructed from fluorescence measurements, automatically averaged from four identical wells, converted into PP data, and plotted on a time scale. Note the robust initial growth in the presence of high concentrations of clindamycin.

Characterization of drug resistance.

Pyrimethamine targets the parasite's DHFR-TS and is the drug of choice for the treatment of toxoplasmosis (13). Moreover, it is also a widely used antimalarial agent (44). In an unrelated genetic experiment, we noticed that the level of pyrimethamine inhibition seems to be temperature dependent. Since this is an important point for genetic experiments with pyrimethamine selection and might have clinical implications, we designed an experiment with the YFP growth assay to determine the exact kinetics of this phenomenon in vitro. Increasing pyrimethamine concentrations were tested on T. gondii parasites carrying the sensitive wild-type DHFR-TS as well as on parasites carrying an engineered drug-resistant DHFR-TSm2m3 gene based on resistant Plasmodium alleles (9). These experiments were performed in parallel at three different temperatures (35, 37, and 40°C). For each growth curve, the growth rate was deduced, and subsequently the IC₅₀ was determined. The tables inserted in Fig. 4 show the deduced IC₅₀ values for each parasite at each temperature. It is clear that pyrimethamine has a temperature-dependent effect in both sensitive and resistant parasite lines. Raising the temperature from 35 to 40°C increases the IC₅₀ from 0.32 to 1.25 μM in the pyrimethamine-sensitive line, whereas the resistant line has an IC₅₀ of 11.4 μM at 35°C and an IC₅₀ of over 20 μM at 37 and 40°C, since growth was hardly affected at the highest concentration (20 μM) tested (Fig. 4).

FIG. 4.

The antiparasitic effect of pyrimethamine is temperature dependent. One thousand parasites were inoculated into each well of an optical-bottom 384-well plate containing a confluent layer of HFF cells. Cultures were incubated at 35, 37, or 40°C. Pyrimethamine was added at day 0 at increasing concentrations, and the plates were read daily with a plate reader. Growth curves were constructed from fluorescence measurements, automatically averaged from four identical wells, converted into PP data, and plotted on a time scale (data not shown). These growth curves have been used to establish growth rates for each drug concentration and temperature. Here, the pyrimethamine concentration is plotted against these growth rates, which are defined as 1 divided by the day when a plateau of maximum fluorescence was reached in the representative growth curve. (A) Wild-type DHFR-TS 2F-1 YFP₂ parasites; (B) pyrimethamine-resistant RH-HX-KO-YFP₂-DHFRm2m3 parasites. The deduced IC₅₀ values for each temperature are shown in the inserted tables.

The capacity of the assay to read the same well at different time points was also exploited to monitor the growth kinetics of parasites carrying a plasmid-encoded drug resistance gene directly after electroporation. A plasmid carrying the HXGPRT gene was electroporated into HXGPRT knockout parasites that express the tandem YFP protein (RH-HX-KO YFP₂). HXGPRT knockout parasites are unable to synthesize purines in the presence of mycophenolic acid and require complementation with the plasmid-derived HXGPRT gene for growth (10). After transfection, growth was monitored in 384-well plates over a 10-day period under different mycophenolic acid concentrations (Fig. 5A). A control was run in parallel and consisted of organisms with a plasmid lacking the HXGPRT gene (TscABP-RV); these parasites were sensitive to all mycophenolic acid concentrations tested (Fig. 5B). The HXGPRT plasmid resulted in a gradually lowered sensitivity upon increasing mycophenolic acid concentrations, but growth was observed at all mycophenolic acid concentrations, reaching a plateau after 10 days. Apparently, the growth rate is slightly lowered by the presence of mycophenolic acid, since untreated parasites reach a maximum after 7 days, as observed in all other experiments.

FIG. 5.

The fluorescence growth assay can be used to track drug resistance. RH-HX-KO-YFP₂ parasites were transfected with plasmids pHXGPRT-DHFR (A) or TscABP-RV (B) by electroporation. Electroporated parasites were directly inoculated into confluent 384-well plate HFF cell cultures. Fluorescence was measure daily, and measurements were converted into PP data and plotted into growth curves. Mycophenolic acid and xanthine were added at the indicated final concentrations on day 0. After initial growth under drug pressure, a plateau is reached, suggesting only transient and not stable drug resistance, upon which continued growth would have been expected (continued growth is not seen when monitored for as long as 3 weeks).

DISCUSSION

We have developed and evaluated a straightforward growth assay for T. gondii by employing parasites stably expressing a highly fluorescent tandem YFP protein (Fig. 1). The relation between parasite growth and the increase in fluorescence intensity was shown to be linear over a wide range (Fig. 2A). Direct comparison with a previously described growth assay based on the β-galactosidase expression (25) in the same parasites showed excellent correlation and showed that the YFP assay might be slightly more accurate, especially at higher parasite numbers (Fig. 2B). The YFP growth assay is as apt for monitoring growth as the β-galactosidase-based assay and the radioactive uracil incorporation assay (25, 30), but it requires no additional reagents or pipetting, thus considerably reducing cost and effort.

The identification of growth retardation for several proven growth-inhibiting compounds that interfere at different stages in the parasite's life cycle demonstrated the versatility of the approach for large-scale drug testing (Fig. 3). IC₅₀ values for pyrimethamine and clindamycin in this miniature assay were in the same range as values previously reported. For pyrimethamine, we found an IC₅₀ of 0.52 μM where values between 0.15 and 1.6 μM have been reported previously (12, 25, 30). For clindamycin, we determined the IC₅₀ to be 0.45 ng/ml, whereas previous studies state values between 2.3 and 11 ng/ml (12, 25). Drugs affecting the parasite's plastid organelle result in a delayed death phenotype and are difficult to measure with other growth assays, since initial growth has to be differentiated from unaffected growth. The best-studied compound for this type of action is clindamycin (1, 4, 11, 12), and its delayed pharmacological kinetics were readily visible as stagnation in the fluorescence increase after a period of uninhibited growth for the first 3 days of culture. Based on the results shown in Fig. 3, a single readout at day 7 after the beginning of the experiment would reliably identify drugs interfering at all levels of the parasite's tachyzoite life cycle. Reading plates daily for the same period will provide detailed kinetic information for each well and each compound tested.

Pyrimethamine is a drug widely used to treat toxoplasmosis and malaria, and we encountered a temperature-dependent growth-inhibiting effect that was explored in more detail with the YFP growth assay. The results shown in Fig. 4 indicate that at higher temperatures, higher concentrations of pyrimethamine are required to inhibit parasite growth in both the wild-type and drug-resistant lines. Consequently, when pyrimethamine is used as a selection drug in genetic experiments employing temperature-sensitive mutants, the dosage needs to be modified to the specific needs. Similar temperature dependency has been previously reported with other drugs (3, 35), and indications for a temperature-dependent effect of pyrimethamine on DHFR-TS comes from modeling studies indicating changes in the binding site (34). From the clinical perspective, these data suggest that fever might influence treatment efficacy.

Finally, the time-resolved capability of the YFP growth assay was used to study the transfection of a plasmid containing the HXGPRT gene under mycophenolic acid selection in HXGPRT knockout parasites (Fig. 5). It is clearly visible that transfecting the HXGPRT gene into a mycophenolic-sensitive parasite line (HXGPRT knockout) results in the growth of drug-resistant parasites as soon as 4 days after transfection. At a low drug concentration (10 μg/ml), no growth impairment is seen relative to the drug-free control, but increasing the drug concentration resulted in less parasite growth.

In conclusion, the YFP-based T. gondii growth assay combines the advantages of using an endogenous reporter, which requires no additional factors for signal measurement, with the high throughput of a microtiter plate-based assay. The strongest asset of the YFP assay is that it can be read at multiple time points, since measurement does not interfere with parasite development. This feature permits not only the titration of the end points of drug inhibition but also the recording of detailed growth curves containing important transient details, like the delayed-death effect of clindamycin. The use of 384-well plates makes this assay highly amendable to automation and should permit the screening of large compound libraries for novel antiparasitic drugs for T. gondii and related apicomplexans.