An ultrasensitive NanoLuc-based luminescence system for monitoring Plasmodium berghei throughout its life cycle

Balb/c and C57BL/6 studies were carried out under the approval of the Animal Research Ethics Committee of the Canton Bern, Switzerland (Permit Number: 91/11 and 81/11); the University of Bern Animal Care and Use Committee, Switzerland; and the German Tierschutzgesetz (Animal Rights Laws). For all studies, females 5–8 weeks of age, weighing 20–30 g at the time of infection were used. Mice were purchased from Harlan or Charles River laboratories. All measurements involving in vivo bioluminescence were performed under isofluorane anaesthesia. Blood feeding was performed under ketavet/dorbene anaesthesia, and all efforts were made to minimize suffering.

The pL0017 vector (obtained through the Malaria Research and Reference Reagent Resource Center [(MR4; http://​www.​mr4.​org/​) (MR4 no., MRA-786)] allows constitutive expression of GFP in P. berghei parasites, under the control of the ef1α promoter, either by episomal expression, or by integration of the plasmid sequence into the cssu/dssu (c-type small subunit/d-type small subunit) ribosomal gene locus of the parasite. The NLuc sequence was amplified from the pNL1.1 (NLuc) vector (Promega) using Phusion ^® Taq High-Fidelity DNA polymerase, and the primers 5′-GGATCCATGGTCTTCACACTCGAAGATTTC-3′ and 5′-TCTAGATTACGCCAGAATGCGTTGCG-3′. The pL0017 plasmid was cut at BamHI and XbaI sites, and the GFP sequence excised. NLuc was introduced into the pJET vector as an intermediate step, and then introduced into the pL0017 vector using the BamHI and XbaI restriction sites to create plasmid pL0017-NLuc. Plasmid DNA was linearized with ApaI and SacII, and 5–10 μg of linearized plasmid DNA was used for transfection of either schizonts or detached cells.

Six to eight-week old Balb/c mice were infected with PbmCherry_Hsp70 [98] until the parasitaemia reached 5–6 %. Mice were bled by intracardiac puncture and schizont cultures generated. Schizonts were then isolated as previously published [99]. Parasites were transfected with pL0017-NLuc, and the transfectant injected into one Balb/c mouse by intravenous injection. One day post-infection, the mouse was treated with 70 mg/ml pyrimethamine in drinking water for positive selection of PbmCherry_Hsp70NLuc_ef1α.

An alternative method tested for the first time, was transfection of liver stage detached cells. HepG2 cells were seeded in a 24-well plate (Greiner BioOne) and the next day infected with PbmCherry_Hsp70 sporozoites. Infected HepG2 cells were maintained as described in further sections. Sixty-five hours after infection and upon completion of the liver stage of infection, detached cells or merosomes were collected from the supernatant and centrifuged for 10 s at 1600 rpm. Merosomes and detached cells were then transfected as previously published [99], with 5–10 µg of linearized pL0017-NLuc and injected into the tail vein of a naïve Balb/c mouse. One day post-infection, the mouse was treated with 70 mg/ml pyrimethamine in drinking water for positive selection of PbmCherry_Hsp70NLuc_ef1α.

For clone selection, Anopheles stephensi mosquito feeds were performed as described below, and at day 18 post-feed, mosquito salivary glands were dissected for infection of HepG2 cells (as described below). Upon completion of the liver stages, six individual detached cells were isolated using previously published methodology [100], and injected into separate mice. 10,000 infected red blood cells (iRBCs) were isolated from each mouse, and probed using the NanoGlo™ (Promega) assay as described below. The clones consistently displaying the brightest signal were selected for downstream analyses.

Balb/c mice were treated with phenylhydrazine three days prior to i.p infection with P. berghei expressing mCherry and either Firefly or NanoLuc luciferases (PbmCherry_Hsp70FLuc_ef1α or PbmCherry_Hsp70NLuc_ef1α). After 3 days of infection, exflagellation was assessed. The infected mice were then each used to feed 100–150 Anopheles stephensi female mosquitoes. Mice were anaesthetized with a combination of Ketasol/Dorbene anaesthesia, and euthanized with CO₂ after completion of the feed. Afterwards, mosquitoes were fed until use with 8 % fructose containing 0.2 % PABA.

Seven to nine days following the blood meal on infected mice, 100 mosquitoes from three different cages infected either with PbmCherry_Hsp70NLuc_ef1α (PbNLuc) or PbmCherry_Hsp70FLuc_ef1α (PbFLuc) were sorted using an epifluorescence binocular microscope (Olympus SZX10 using a U-HGLGPS UV lamp), and mCherry-positive mosquitoes selected, dissected and the midguts isolated to evaluate oocyst numbers and luciferase expression. For correlation between fluorescence and luminescence assays, 100 dissected midguts were transferred in 1 × PBS onto a coverslip, and imaged immediately using a 20× objective, of an upright epifluorescence microscope (Leica DM 5500 B). Four z-stacks from each midgut were obtained to allow for quantification of oocytsts over the entire midgut area. Upon finalizing image acquisition, midguts were immediately transferred to the 20 μl of 1 × PLB for downstream analysis. For prevalence assays, 200 mosquitoes were not sorted prior to dissection or homogenization, but randomly selected from 3 independent feed cages harbouring no more than 150 mosquitoes each. Bioluminescent measurements of all samples were performed in 96-well, flat-bottom, non-binding, black plates (Greiner BioOne). For consistent and unsaturated oocyst measurements, the optimal NanoGlo™ luciferase assay substrate dilution used for probing the midguts was 1:1000. The midguts of uninfected mosquitoes were used as negative controls.

Sixteen to 28 days post-feed, mosquitoes infected with PbmCherry_Hsp70NLuc_ef1α (PbNLuc) or PbmCherry_Hsp70FLuc_ef1α (PbFLuc) were dissected for extraction of their salivary glands. Sporozoites were counted using a Neubauer chamber, and serial dilutions of sporozoites made in 1 × PBS. The final diluted samples were centrifuged and the pellet (10 μl) transferred to 20 μl of 1 × PLB for downstream analysis. For imaging, a range between 1 and 100,000 sporozoites, different dilution sets from 20 separate mosquitoes were divided into independent 96-well plates to enable different imaging settings for maximum signal detection of samples at low densities (i.e. 1–100 sporozoites), while avoiding pixel saturation of samples at high densities (i.e., 10,000–100,000 sporozoites). For consistent and unsaturated sporozoite measurements, the optimal NanoGlo™ luciferase assay substrate dilution used was 1:250. Salivary glands of uninfected mosquitoes were used as negative controls.

Human hepatoma cells (HepG2 line) (European Collection of Cell Culture) were maintained in complete Minimum Essential Medium (cMEM) containing Earle’s Salts, supplemented with 10 % heat-inactivated fetal calf serum (FCS), 2 mM l-Glutamine, and 100 U Penicillin, 100 μg/ml Streptomycin (all from PAA laboratories, E15-024, A15-101, P11-010, M11-004). Cells were kept at 37 °C in a 5 % CO₂ cell incubator and were split every 4 days by treatment with accutase. For in vitro bioluminescence assays, 30,000 cells were seeded into wells of 24-well plates, and infected 18-24 h afterwards with 10,000 PbmCherry_Hsp70NLuc_ef1α (PbNLuc) or PbmCherry_Hsp70FLuc_ef1α (PbFLuc) sporozoites. Upon infection, sporozoites and HepG2 cells were incubated together in cMEM + 2.5 μg/ml amphothericin (AT-MEM) to avoid fungal contamination. At 6, 12, 24, 36, 48, and 56 h post-infection, total parasite numbers were quantified by fluorescence, and HepG2 cells detached using 100 μl of accutase. Cells were quantified using a Neubauer chamber, and specific parasite numbers isolated (ranging from 1 to 1000). The diluted samples containing the desired number of parasites were lysed in 20 μl of 1×PLB for downstream analysis. For isolation of single later stage parasites (36 h onwards), the cells detached by accutase treatment were imaged using an epifluorescence microscope to ensure capture of a single parasite (in a method analogous to that described by Stanway et al. [100]. Lysates were transferred to 96-well black plates, and imaged immediately at every time point using an IVIS Lumina II. For consistent and unsaturated measurements, the optimal NanoGlo™ luciferase assay substrate dilution used for probing liver stage parasites was 1:250. Uninfected HepG2 cells were used as negative controls.

In order to investigate detached cell formation by PbmCherry_Hsp70NLuc_ef1α (PbNLuc) parasites, HepG2 cells were infected as previously described, and at 63–65 h post-infection, the supernatant of the 24 well-plates, where detached cells (representing successfully completed liver stage development) are located, was collected for downstream analysis. For collection of single detached cells, the method previously published by Stanway et al. [100] was used. Prior to collection and lysis, individual detached cells were imaged by fluorescence using an inverted fluorescence microscope (Leica DMI 6000B with an incubation chamber), and a series of 3–4 z-stacks obtained for measurements. Upon collection, individual detached cells were lysed in 1xPLB, and assayed by luminescence. For consistent and unsaturated measurements, the optimal NanoGlo™ luciferase assay substrate dilution used for probing detached cells was determined to be 1:250. In order to determine the number of merozoites per individual detached cell, a previously published method [101] was used. Namely, from the fluorescence images, the radius of the detached cell was determined, and the volume of the detached cell (V _DC ) was calculated by the formula \(V_{DC} = \frac{4}{3}\pi r^{3}\). Sturm et al. previously determined the volume of individual merozoites (V _m ) to be equal to 2.66 μm³ [101]. An approximation of the number of merozoites (N _m ) per detached cell is, therefore, calculated as \(N_{m} = \frac{{V_{DC} }}{{V_{m} }}\).

To determine the pre-patency period detectable by NanoLuc, a total of 20 mice were infected with 5000 sporozoites from PbmCherry_Hsp70NLuc_ef1α-infected mosquitoes by tail-vein injection. 2 µl of blood were taken by tail puncture at 12 and 24 h post-infection and from 36 h onwards, 4-hourly measurements were made until 70 h post-infection. This blood was lysed in 20 µl of 1 × PLB, and for maximum sensitivity, probed with 1:50 NanoGlo™ luciferase assay substrate dilutions. To determine in vitro sensitivity, individual detached cells from infected HepG2 cultures (described above) were isolated as previously described [100], and imaged using an epifluorescence microscope for downstream volume calculations. Following imaging, each individual detached cell was transferred to 20 µl of 1 × PLB. Images of blood and detached cells were acquired immediately using an IVIS Lumina II equipment. For maximum sensitivity for appearance of parasites in the blood, a 1:50 dilution of NanoGlo™ luciferase assay substrate was used.

For blood stage measurement curves and assays for detecting the minimum sensitivity limit, parasites were diluted to ranges between 1 and 100,000. Each of the diluted samples were lysed in 1 × PLB, and imaged immediately. For calculation of the parasitaemia, three separate groups of five mice each (Balb/c and C57BL/6) were infected with 10⁶ schizonts of either PbmCherry_Hsp70NLuc_ef1α, or PbmCherry_Hsp70FLuc_ef1α. 2 µl of blood were obtained daily by tail vein prick, lysed in 20 µl of 1 × PLB, and probed immediately. Every day, thin smears were made on glass coverslips, and Giemsa stains performed for microscopic quantification of parasitaemia to assess correlations between luminescence values and parasitaemia in blood smears. For analysis of sequestration, mice were injected with either 10⁶ PbmCherry_Hsp70NLuc_ef1α schizonts, or with a mixture of PbmCherry_Hsp70NLuc_ef1α and PbmCherry_Hsp70FLuc_ef1α (5 × 10⁵ schizonts of each line), and a synchronous infection allowed to establish. 2 µl of blood were extracted bi-hourly within the first 28 h post-schizont injection via tail vein puncture, lysed in 20 µl of 1 × PLB, and imaged. For maximum sensitivity of single blood stages, a 1:50 dilution of NanoGlo™ luciferase assay substrate was used. In all other assays for in vivo parasitaemia measurements, a 1:250 dilution was used. Erythrocytes from uninfected mice were used as negative controls.

All data analysis of luminescence values was performed using the LivingImage 4.1 software. Analyses for Pearson correlation coefficients, Student’s t tests, and linear regressions were carried out using STATA 13.0 software. Other than the STATA-generated graphics, all other graphics were generated using the Prism 6.0 software. All parasite measurements including diameter, radius, and area, were performed using plugins freely available in Fiji. Significance in all assays was determined upon p values being equal to, or less than 0.05.

For the generation of the pNLuc plasmid, the GFP coding sequence from the pL0017 plasmid was excised, and replaced by the NanoLuc coding sequence, thereby allowing expression of NanoLuc under the control of the ef1α promoter. PbmCherry_Hsp70 parasites [98] were transfected with the pNLuc construct, and this resulted in single cross-over integration events. Individual Pb mCherry_Hsp70NLuc_ef1α clones were isolated by detached cell injection (89) into six mice. A drop of blood was isolated from all mice harbouring synchronous ring infections. Parasites were probed with NanoGlo™ according to manufacturer’s instructions, and luminescence measured with an IVIS Lumina II imager. Measurements were repeated three times at different days post-infection, isolating equal numbers of iRBCs by dilution. Although signal intensity between most clones did not differ significantly, the clone that consistently showed the highest signal was chosen for downstream analyses.

Microscopy-based imaging and quantification of PbmCherry_Hsp70, PbmCherry_Hsp70FLuc_ef1α [39] (henceforth referred to as PbFluc) and PbmCherry_Hsp70NLuc_ef1α (henceforth referred to as PbNLuc) in mosquito midguts showed that all three lines strongly express mCherry (Fig. 2a) (Diagram adapted from [102]) and have equivalent average numbers of oocysts per mosquito on days 7–9 post-feed (Fig. 2b). For comparison of PbNLuc and PbFLuc luminescence intensities, ten pools of five midguts with approximately equal numbers of oocysts from PbFLuc- and PbNLuc-infected mosquitoes were dissected, lysed, and their bioluminescence measured. Midguts infected with PbNLuc oocysts showed 160- to 180-fold higher signal (p < 0.05) than midguts infected with PbFLuc oocysts (Fig. 2c).

In order to determine the NanoLuc assay’s sensitivity in detecting changes in average numbers and sizes of oocysts throughout their development on midguts, 50 midguts from mosquitoes in three independent cages were dissected and measured at days 7–9 and 15 post-feed. Although detecting small variations (in the range of tens of oocysts) is challenging using firefly luciferase, NanoLuc-based luminescence in PbNLuc parasites allowed detection of a clear progressive increase in intensity from days 7–9 post-feed, which decreased again at day 15 coinciding with sporozoite egress and migration to the salivary glands (Fig. 2d).

As proof of principle for the use of PbNLuc in assessment of infection prevalence in mosquitoes, a total of 200 mosquitoes were sorted by fluorescence into two groups consisting either of mCherry-positive (infected), or mCherry-negative (non-infected) mosquitoes. By microscopy alone, the prevalence of infection was calculated to be 70 % (SD, 10.5 %). These mosquitoes were then dissected and their midguts imaged by bioluminescence. The prevalence estimate detected by luminescence of this same group, was 88 % (SD, 12 %) (Fig. 2e), emphasizing the enhanced sensitivity of the assay for determining presence or absence of midgut oocysts by NanoLuc luminescence rather than fluorescence. In addition, the assay avoids observer bias in terms of assessment of fluorescence at low oocyst densities. While fluorescence-based sorting is to an extent subjective, sensitive bioluminescence imaging enables non-subjective, quantitative assessment of infection prevalence.

Having proven the enhanced sensitivity of PbNLuc over fluorescence for assessment of prevalence, a next step was to explore the correlation of PbNLuc luminescence with oocyst numbers in individual midguts. At days 7–9 post-feed, 100 infected midguts of PbFLuc or PbNLuc-infected mosquitoes were dissected, imaged by fluorescence for manual or semi-automated quantification of oocysts, and then immediately lysed for imaging by luminescence. Oocyst numbers were quantified by two different methods, namely, a previously published semi-quantitative method based on fluorescence [103]; and light emission values based on NanoLuc or firefly luciferase bioluminescence. Figure 2f, g show fitted correlations and 95 % confidence intervals of bioluminescence versus automated fluorescence counts (each dot corresponds to the measurement of one individual midgut). PbNLuc shows a Pearson correlation coefficient of 0.74 (Fig. 2f), while PbFLuc shows a correlation coefficient of 0.61 (Fig. 2g). This suggests that the increased brightness of NanoLuc luciferase in PbNLuc parasites results in more sensitive measurements, allowing differences of even ten to a hundred oocysts to be distinguished. Nevertheless, a perfect correlation between counts and luminescence is not expected, as individual oocysts may greatly vary in their sporozoite load. This phenomenon has previously been reported in luminescence-based assays for transmission assessment [50].

In the Plasmodium life cycle, a critical step following oocyst formation is the production of sporozoites, as these latter forms ultimately enable mosquito-to-human transmission. Having explored the suitability and sensitivity of PbNLuc to assess oocyst formation and development, this section explores the sensitivity and limit of detection of PbNLuc to quantify sporozoites. Fluorescence-based images of sporozoites within salivary glands of mosquitoes infected with PbFLuc or PbNLuc were obtained (Fig. 3a) (Diagram adapted from [102]), and the number of sporozoites in both luminescent lines were quantified at days 16, 18, 20, 22, 24, 26 and 28 days post-feed. PbFLuc and PbNLuc show equivalent sporozoite counts at all days post-feed, with maximum sporozoite loads detected between days 20 and 24 post-feed (Fig. 3b) suggesting no differences in egress and development between the two parasite lines. Next, equal numbers of PbFLuc and PbNLuc sporozoites (four separate pools of 50,000) were lysed, and their luminescence measured for comparison of their luminescence intensities. Luminescence values of PbNLuc sporozoites were 170-fold higher than luminescence values for equal numbers of PbFLuc sporozoites (Fig. 3c), confirming the results presented in Fig. 2c. To determine the detection limit of NanoLuc for sporozoites, serial dilutions of PbNLuc sporozoites were generated in triplicate, and lysed for measurement. While the minimum number of detectable PbFLuc sporozoites was between the range of 50 and 100, single PbNLuc sporozoites were detectable (subject to a sporozoite actually being present in the well) (Fig. 3d, e). The reason why 5 parasites per well are presented was the experimental setup itself: using serial dilutions, the probability of one single sporozoite being present in each well was 10 %. Nevertheless, every time there was a single sporozoite per well, signal was detected. Sporozoite numbers showed high correlation values with the log values of luminescence (R₂ = 0.913 for PbNLuc, and R₂ = 0.921 for PbFLuc) (Fig. 3d). This highly significant correlation shows the success of the assay in terms of both detection of very low sporozoite densities, and detection of small variations in sporozoite numbers. The detection limit of PbNLuc, however, is higher than that of PbFLuc. As a proof of principle for its application, salivary glands of 100 PbNLuc-infected mosquitoes were dissected, and the correlation between sporozoite numbers (as assessed by manual counting) and luminescence was calculated. Each dot represents the sporozoites of a single mosquito, as assessed by luminescence (x-axis) and manual counts (y-axis). On one hand, the test confirmed a high correlation between sporozoite numbers and luminescence values (R₂ = 0.95) when applied to single mosquitoes (Fig. 3f). On the other hand, the assay also shows that given this high correlation, visual inspection of plates to estimate variability in the overall burden of sporozoites per mosquito is possible (Fig. 3g).

Having proven the use and sensitivity of NanoLuc in mosquito stages, next liver stages of infection were assessed in vivo and in vitro. HepG2 cells seeded in 24 well-plates were infected with 10,000 PbNLuc sporozoites per well, and imaged by fluorescence microscopy at 6, 12, 24, 36, 48 and 56 h post-infection (Fig. 4a shows representative images and a schematic (adapted from [104]) of liver stage development). To exclude differences in luminescence due to differences in overall development or sizes, PbmCherry, PbFLuc and PbNLuc parasite sizes were measured at 6, 12, 24, 36, 48 and 56 h post-infection, and infection success and survival (i.e. total number of HepG2 cells infected at each time point) were quantified by fluorescence. No differences were detected between PbmCherry, PbFLuc and PbNLuc, neither in terms of size, nor successful liver stage completion (detached cell formation).

In order to compare luminescence intensities of PbFLuc and PbNLuc, HepG2 cells in 6 wells of a 24-well plate were infected with an equal number of sporozoites of either one of the strains. At 40 h post-infection these cells were lysed and their luminescence quantified. PbNLuc showed 140-fold higher luminescence than PbFLuc (Fig. 4b). Next, in order to determine the detection limit for liver stages, PbNLuc-infected HepG2 cells were detached, total parasite numbers estimated, and serial dilutions generated for each time post-infection, ranging from 1 to 1000 parasites for each time point (i.e., 6, 12, 24, 36, 48 and 56 h post-infection). The assay was successful in detection of single liver stage parasites at all times post-infection, and enabled differences (Fig. 4c) in the range of tens to hundreds to be detected. This level of sensitivity has not been reported for firefly luciferase, nor was it ascertained in the present work.

Next, kinetics of liver stage infection in vitro throughout pre-erythrocytic development in terms of growth and survival, were determined by luminescence (Fig. 4d). HepG2 cells were infected with an initial inoculum of 10,000 sporozoites per well. Fluorescence images of at least 100 infected cells per time point were acquired for size measurement. Ten wells of a 24 well-plate were assessed by fluorescence imaging for quantification of parasites for each time point (shown in Fig. 4e). Finally cells were lysed at 6, 12, 24, 36, 48 and 56 h post-infection, and bioluminescence measured (Fig. 4f). Each time point was repeated in 10 wells in three independent experiments. Total parasite sizes and numbers at 6 h post-infection were set as 100 % (Fig. 4e, f). While parasite numbers decreased significantly already at 12 h post-infection (50 % reduction), a dramatic decrease in numbers (of about 80 % compared to 6 h post-infection) was observed at 24 h post-infection. Parasite number remained relatively constant at from 24 to 36 and 48 h, and showed only a 2 % further decrease at 56 h post-infection, potentially explained by a small number of detached cells occurring at this relatively early egress time point (Fig. 4e, black line). Conversely, parasite sizes at 6 and 12 h post-infection were almost equal, while size increased 3-, 6-, and 10-fold at 24, 36 and 48 h respectively, compared to parasite sizes at 6 h post-infection. The parasite size increase between 48 and 56 h was only minimal (Fig. 4e, blue line). The overall result of these changes in parasite numbers and sizes, in addition to potential changes in promoter activity, resulted in variations in luminescence intensity through time, with total luminescence being lowest at 24 h post-infection, after which these values increased by 48 h to almost 300 % of the level measured at 6 h post-infection (Fig. 4f).

While it was possible to investigate overall liver stage growth of PbNLuc parasites by bioluminescence, a marker of fully successful liver stage development is the successful production of merosomes and detached cells in vitro. Detached cell formation in vitro, like merosome formation and parasite egress from the liver in vivo, occurs upon formation of merozoites [101]. Therefore whether a PbNLuc-based luminescence is sensitive enough to allow estimation of numbers of merozoites in individual detached cells, was investigated next. While individual merozoites were not quantified, their number was estimated by measuring the detached cell volume as previously published [101], and as shown in Fig. 5a. Figure 5b shows representative bioluminescence images of individual detached cells of various sizes (and therefore, different numbers of merozoites). Each well has a single detached cell. While the median calculated number of merozoites in 100 detached cells was 4500 (with the average detached cell volume being 11,780 μm³), Fig. 5b shows clear outliers (marked) representing both larger detached cells or those with much higher merozoite numbers (marked in red outline), and also those extremely small or with very reduced merozoite numbers (marked in blue outline). Figure 5c shows representative fluorescence images of these extreme size variations between merosomes (M) and large detached cells (L-DC). Figure 5d shows the correlation of detached cell luminescence with the volume calculated by fluorescence imaging, and the number of merozoites calculated on the basis of the volume estimated.

Given the high sensitivity of NanoLuc detection, liver stage egress in vivo was re-investigated. While the most sensitive assay reported to date states that the first hepatocyte-derived blood stage parasites can be detected earliest at 60 h post-sporozoite injection [15], previously reported in vivo luminescence assays [39], as well as in vivo measurements performed here using PbFLuc (Fig. 6a) suggested a significant decrease in liver loads from 42 h post-infection onwards in untreated mice. This suggests that egress has potentially already begun occurring at this time point. Whether the previously reported lack of detection of parasites in peripheral blood between 42 and 60 h post-sporozoite injection is due to very low parasite densities (and therefore undetectable by firefly luciferase), or due to migration of merosomes to specific tissues (away from the periphery), was not known. Following infection of C57BL/6 (permissive) and Balb/c (refractive) mice with 5000 PbNLuc sporozoites, luminescence in peripheral blood was measured at 12, 24, 30 h post-infection, and every 4 from 36 h post-infection onwards, until 70 h post-infection. Using the maximum sensitivity setup for NanoLuc, luminescence signal was detected from 36 h post-infection in some mice, and by 40 h post-infection and at all subsequent time points, in all mice. Consistent with the known permissiveness of C57BL/6 mice for liver stages, and the refractoriness associated with liver stage infections in Balb/c mice, the detectable levels of luminescence in peripheral blood were on average 100-fold lower in Balb/c mice, but nevertheless the earliest signal suggesting egress, occurred at 36 h in both mouse strains (Fig. 6b). In parallel to PbNLuc infections, a separate group of C57BL/6 mice was infected with PbFLuc. As published in previous work [15], firefly-luciferase parasites were only detected at 60 h post-infection, but not earlier (Fig. 6c). Dynamics of egress displayed by PbNLuc and PbFLuc are shown in Fig. 6c.

In order to determine whether parasites (merosomes, free merozoites or early infected RBCs) present in peripheral blood from 36 h onwards were infectious to naïve mice thereby confirming egress, 20 μl of blood were extracted from the sporozoite-infected C57BL/6 mice at 36, 40, 42, 44, 52, 56 and 60 h post-infection, and injected intravenously into three naive C57BL/6 mice for each time point. While blood extracted at 36, 40 and 42 h did not lead to blood stage infections in the naïve mice, blood extracted at 44, 52, 56 and 60 h resulted in infection in all three recipient mice (Fig. 6d). Careful analysis by fluorescence microscopy allowed observation of amorphous, relatively large structures (potentially merosomes) at 36 to 42 h post-infection, which nevertheless did not lead to infection in the recipient mice.

Next it was determined whether differences exist in egress times of PbNLuc parasites in mice injected with sporozoites by intravenous injection, or mice on which infected mosquitoes fed. No differences in egress times were observed, despite mice infected by intravenous injection of 5000 sporozoites displaying at least tenfold higher signal than mice infected by feeds with five mosquitoes (Fig. 6e).

As for other stages, the sensitivity of NanoLuc for low parasite densities was tested in blood stages of infection with PbNLuc. Limiting dilution ranging from 1 to 10,000 parasites showed that even single parasites can be detected, and that parasite number and luminescence are highly correlated (R₂ = 0.956) (Fig. 7a). While infections with PbFLuc hold equally high correlation (R₂ = 0.921), its signal intensity is at least 100-fold lower, and does not allow detection of single parasites (Fig. 7a). To determine differences in sensitivity of PbNLuc compared to PbFLuc in blood stage infections, a drop of blood was obtained by tail vein puncture, and 1000 parasites were isolated by dilution in 1xPBS. Figure 7b shows that PbNLuc has 150-fold higher signal than PbFLuc, supporting the results with the other parasite stages. To determine whether differences in parasite development exist in Balb/c and C57BL/6 mice, parasitaemias arising from infection with 10⁶ schizonts by tail-vein injection were followed by luminescence and Giemsa stains for 7 days. No significant differences concerning blood stage infection were found between the two mouse strains (Fig. 7c, d). Figure 7c shows representative images from 5 different mice at days 1–4 post-infection.

While assessment of various sequential parasite dilutions (Fig. 7a) showed high correlation with luminescence, whether in vivo parasitemia correlates with luminescence was further explored. Figure 7e shows the correlation of luminescence with parasitaemia calculations assessed by manual counting of Giemsa-stained thin blood smears. Parasitaemia and luminescence values showed a strong correlation (R² = 0.907) (Fig. 7e).

The sensitivity of the NanoLuc assay allows determination and comparison of multiplication rates between mice and parasite strains. Previous work has explored the use of luminescence to detect schizont sequestration in synchronous infections in relation to the day light-cycle [16, 17]. In this work, blood stage parasites from a donor mouse were cultured in order to purify schizonts. Schizonts were injected into naïve mice to establish synchronous infections that were followed bi-hourly over 27 h, when specific times would correspond to synchronous ring, trophozoite, and schizont stages. Figure 7f shows results of these bi-hourly luminescence measurements of peripheral blood parasitaemia of mice infected with PbNLuc schizonts. The drop in parasitaemia from 18 h onwards, coinciding with the beginning of schizont sequestration in the vasculature, is clearly identified.

In order to test the use of PbNLuc in multiplex assays using two different luciferases, mice were co-infected with PbNLuc and PbFLuc. Figure 7g shows the comparative values of luminescence of both luciferases in the same mice, clearly showing sequestration of schizonts from both strains at around 18 h after synchronous infection.

Finally, to test further potential applications of the NanoLuc-expressing parasites, since even single PbNLuc parasites can be identified due to the strong luminescence intensity of NanoLuc, and limiting dilution is a routinely used method for clonal selection of genetically-modified parasites, it was determined whether identification of single parasites expressing NanoLuc, followed by their injection into naïve mice could lead to effective infection. By limiting dilution, the estimated volume containing 1 parasite was isolated and deposited into a 96-well plate. NanoLuc was diluted in 1 × PBS rather than lysis buffer to ensure the viability of the parasites, and 4 of the 96 (i.e. 5 %) wells were found to be positive (Fig. 7h). The full volume of all four wells was isolated and injected into recipient mice. All caused effective infections in the recipient animals, suggesting the usefulness of the method to avoid large animal usage in limiting dilution experiments.