Expression of the MSPDBL2 antigen in a discrete subset of Plasmodium falciparum schizonts is regulated by GDV1 but may not be linked to sexual commitment

To test the effect of GDV1 overexpression on the proportion of mature schizonts expressing MSPDBL2, we used the genetically engineered P. falciparum 3D7/GDV1-GFP-DD parasite line (3D7/iGP). In this line, GDV1 is overexpressed from an ectopic locus when parasites are cultured in the presence of the stabilizing reagent Shield-1 (23). As expected, in four independent experiments, live fluorescence microscopy showed that most trophozoites (a mean of 76.2%) expressed GDV1 in the presence of 1 µM Shield-1, compared to only 2.9% in the absence of Shield-1 (Fig. 1A). In each of these experiments, the proportion of schizonts that were MSPDBL2+ was also massively increased by GDV1 overexpression, the mean proportion being 38.5% with Shield-1 and only 1.1% without Shield-1 (Fig. 1A; Table S1). Additional tests were made using exposure to intermediate concentrations of Shield-1 reagent, which yielded a range of intermediate proportions of MSPDBL2+ schizonts (Fig. 1B; Table S2). The gametocyte conversion rates (GCRs) of these cultures had previously been assayed by examining parasites in the next cycle using Pfs16 staining [published data (24) tabulated here in Table S2]. There was a significantly positive correlation between these gametocyte conversion rates and the proportions of MSPDBL2+ schizonts (Spearman’s rho = 0.65, P = 0.0095)(Fig. 1C), showing that both responded in parallel to the overexpression of GDV1.

MSPDBL2 is normally only expressed in a small minority of schizonts in P. falciparum. To test whether GDV1 is required for this normal expression, parasites in which the gdv1 gene was disrupted were investigated. The P. falciparum clone NF54-GDV1Δ39 contains a premature stop codon that results in a truncated version of GDV1 which is inactive and prevents induction of sexual commitment (25). In three separate replicate experiments, the NF54-GDV1Δ39 clone was compared with the parental NF54 clone. No MSPDBL2+ schizonts were seen in the NF54-GDV1Δ39 line in any of the replicates, whereas the NF54 control parasites showed the typical minority proportion of positive mature schizonts (between 0.6% and 1.3% for each replicate culture) (Fig. 2; Table S3).

It was previously shown that the proportion of mature schizonts which express MSPDBL2 varies among P. falciparum lines (5). To gain a better view of MSPDBL2 variation in culture, multiple biological replicate preparations of each of 13 P. falciparum long-term culture adapted lines were analyzed (Fig. 3A), all lines being grown in identical serum-free medium supplemented with AlbuMAX II. Each of these replicates was sampled from a culture in which the gametocyte conversion rate in the subsequent cycle was measured and previously reported (24), with the slides of schizonts having been archived for immunofluorescence analysis of MSPDBL2 expression in the current study.

Most P. falciparum lines (D10, 3D7, NF54, 7G8, Palo Alto, RO33, T996, Dd2, and F12) consistently showed very low proportions of MSPDBL2+ schizonts (<1% positive as means in immunofluorescence using antibodies specific for a conserved region of MSPDBL2) (Fig. 3A; Table S4). It is notable that no parasite line expressed zero MSPDBL2+ schizonts, and that even a parasite line containing an ap2-g nonsense mutation (clone F12, previously derived from 3D7) expressed MSPDBL2 in 0.9% of schizonts. This indicates that MSPDBL2 expression is not dependent on AP2-G. Two of the parasite lines showed slightly higher proportions of MSPDBL2+ schizonts (means of 1.5% in FCC2 and 2.1% in D6), and two lines showed much higher proportions (means of 7.3% in GB4 and 13.2% in HB3) (Fig. 3A; Table S4), all of these being significantly higher than in those listed above (pairwise Mann-Whitney U-tests, Table S5).

The gametocyte conversion rates were previously assayed in each of these individual replicate cultures, using Pfs16 staining to discriminate early gametocytes in the following cycle (24), so these were compared with the proportions of MSPDBL2+ schizonts. All of the parasite lines except F12 produce some early gametocytes, although some of the lines produce very few (Fig. 3B; Table S4). Notably, proportions of MSPDBL2+ schizonts were not significantly correlated with the gametocyte conversion rates. For example, there are very low proportions of MSPDBL2+ schizonts in several of the parasite lines that show relatively high rates of gametocyte conversion (lines 7G8, RO33, Dd2, NF54, and FCC2) (Fig. 3A and B). In a pooled analysis across all parasite lines, there was only a very weak positive correlation between proportions of MSPDBL2+ schizonts and gametocyte conversion rates (Spearman’s rho = 0.18, P = 0.10, Fig. 3C). Considering inter-replicate variation in rates for individual lines, separate analyses of each of the three parasite lines with most inter-replicate variation in proportions of MSPDBL2+ schizonts (FCC2, D6, and GB4) showed no significant correlations with the gametocyte conversion rates of the same culture replicates (Fig. 3C; Table S4). As routine estimates of parasitemia by Giemsa staining of slides in culture maintenance had been recorded for 42 of these cultures in both cycles (Table S4), the possibility of correlations between proportions of MSPDBL2+ schizonts and the parasitemia levels or fold changes between cycles was also explored. This revealed no significant correlation and no directional trend for variation within each of the parasite lines that had at least five replicates (mean rho = 0.21, P-values >0.56 for MSPDBL2 expression in schizonts versus parasitemia in the same cycle; mean rho = −0.09, all P values >0.26 for MSPDBL2 expression in schizonts and parasitemia fold change) (Table S6).

Choline supplementation of serum-free medium has been previously shown to suppress parasite GDV1 expression (22) and sexual commitment rates (22, 24, 26, 27). Given that GDV1 expression increases MSPDBL2+ schizont proportions in the inducible parasite line 3D7/iGP, it is relevant to test for a potentially suppressive effect of choline supplementation on proportions of MSPDBL2+ schizonts. The HB3 and NF54 lines were analyzed here, as they have contrasting proportions of MSPDBL2+ schizonts (relatively high in HB3 and very low in NF54) (Fig. 3). Multiple biological assay replicates were performed on each of these lines to test the effect of culture medium to which either 2 mM choline or no choline was added, with parasites exposed from the ring stage of development onward and allowed to develop to schizonts for analysis of MSPDBL2 expression.

The addition of 2 mM choline significantly suppressed the proportion of MSPDBL2+ schizonts in four of seven replicates performed on the HB3 parasite line (P < 0.001 for each of the four replicates, Fig. 4A; Table S7). All four replicates with significant effects had >10% MSPDBL2+ schizonts in the controls without choline treatment, whereas the three replicates that showed no effect had significantly lower MSPDBL2+ schizont proportions in the controls (Fisher’s Exact P = 0.029 in test for association between initial proportions positive and effect of choline). For NF54, which had very low MSPDBL2+ schizont proportions in all controls, there was no significant effect of choline in reducing these proportions in any of the six replicate assays performed. These results indicate that adding choline suppresses proportions of MSPDBL2+ schizonts under conditions in which these would otherwise be high. The effect of choline in suppressing gametocyte conversion rates was previously reported for these same culture replicates (24), showing significant effects in most replicates of both parasite lines (Fig. 4B).

Given that GDV1 induces sexual commitment through de-repression of the epigenetically silenced ap2-g locus (22, 28), it was tested whether AP2-G is expressed in the same individual schizonts as MSPDBL2. Using an engineered parasite line with hemagglutinin (HA)-tagged AP2-G (3D7/AP2G-HA) (29), schizonts were examined by two-color immunofluorescence, using anti-HA antibody to detect AP2-G alongside antibodies to a conserved region of MSPDBL2. It was expected that proportions of schizonts positive for MSPDBL2 would be very low, as data presented above show that <1% of 3D7 schizonts express the protein. Out of a total of 5,571 mature schizonts examined (each having at least eight nuclei), only two schizonts were positive for both proteins, whereas 30 schizonts were positive for MSPDBL2 only (these were negative for AP2-G), and seven were positive for AP2-G only (these were negative for MSPDBL2) (Fig. 5). This finding at the protein level is consistent with results from a single-cell transcriptomic analysis of parasite line HB3, which indicated ap2-g and mspdbl2 transcripts occurred mostly in different schizonts (17). The results indicate that, although AP2-G and MSPDBL2 expression are both upregulated by GDV1, they are not usually co-expressed, so that MSPDBL2 is not a marker of sexually committed schizonts.

A strategy based on the selection-linked integration (SLI) system (30) was employed to generate parasite lines expressing a C-terminal TdTOM-3×HA-tagged full-length version of MSPDBL2 or 3×HA-tagged truncated version of MSPDBL2. The 3D7/iGP parasite line was chosen as the background for engineering, as it allows conditional overexpression of GDV1 and induction of elevated gametocyte conversion rates (23). The SLI system was used to enable modification at the endogenous mspdbl2 locus to allow selection of schizonts expressing either an intact full-length MSPDBL2 (referred to as MSPDBL2-TAG) or a disrupted version of MSPDBL2 (referred as MSPDBL2 DEL) lacking two-thirds of the protein that covers the DBL domain (Fig. 6A; Fig. S1 and S2). It is important to note that no Shield-1 reagent was added at the selection stage, so there was no manipulation of GDV1 expression in that part of the process, but the engineered MSPDBL2-expressing parasites were selected for neomycin resistance.

Polyclonal antibodies against the conserved N-terminal region of MSPDBL2 (5) were used to detect either the intact or truncated protein, while polyclonal antibodies against the conserved C-terminal region of MSPDBL2 (5) were used to detect the intact protein (Fig. 6A). After repeated rounds of G418 selection for parasites with the modified locus, a majority of schizonts in the MSPDBL2-TAG line expressed intact MSPDBL2 [staining with either the anti-N-terminal MSPDBL2 Ab or the anti-C-terminal MSPDBL2 Ab (Fig. 6B and C; Table S8)]. In comparison, the MSPDBL2 DEL line expressed a truncated version of MSPDBL2 in most schizonts (positive with the anti-N-terminal but not with the anti-C-terminal MSPDBL2 Ab) (Fig. 6B and C; Table S8). The proportions of schizonts expressing the respective intact or truncated MSPDBL2 proteins in these parasite lines remained similarly very high 6 days after removal of the G418 selection (Table S9), showing the artificially engineered expression to be maintained over at least three cycles.

To test whether there was any direct effect of expressing intact or disrupted MSPDBL2 on the rates of gametocyte conversion per cycle, assays involving overexpression of GDV1 were conducted with the mspdbl2-engineered lines as described above. Each line was tested with multiple experimental replicates, in the absence and presence of Shield-1 reagent to stabilize the overexpression of GDV1, which drives sexual commitment. Gametocyte conversion rates were assayed using a method previously described to detect Pfs16-positive parasites early in the following cycle (24). All parasite lines showed massively increased gametocyte conversion rates in response to overexpression of GDV1 under Shield-1 (Fig. 7). There were no significant differences in the rates between MSPDBL2-TAG and MSPDBL2 DEL lines in which most parasites respectively express intact or disrupted MSPDBL2 (as shown in Fig. 6), and both showed similar rates to the 3D7/iGP parental line (Fig. 7; Table S10). This confirms that intact MSPDBL2 is not required for sexual conversion or subsequent sexual differentiation.

To test whether there was an effect of expressing intact or disrupted MSPDBL2 on asexual growth, the asexual blood stage multiplication rates of parasites were analyzed, comparing the engineered lines in which most schizonts expressed the full-length or disrupted version of MSPDBL2, as well as parasites in which most schizonts are MSPDBL2-negative. Each of the parasite lines was tested in three separate 6-day exponential growth assays, each separate assay involving triplicate flasks with erythrocytes from different donors using a protocol described previously (31, 32). This showed that the multiplication rates of 3D7/iGP, MSPDBL2-TAG, and MSPDBL2 DEL were not significantly different (Fig. 8), all being approximately eightfold per 48 hours. These rates were similar to those of the 3D7 parasite line that was incorporated alongside as a control for assay quality (Table S11), which was consistent with its previously established multiplication rate using this assay method (31, 32).

The inducible gametocyte producer P. falciparum clone 3D7/iGP_D9 was studied (23) (referred to as 3D7/iGP for brevity in the current paper) to assess the impact of GDV1 overexpression on the proportion of MSPDBL2+ schizonts using methods described in a separate subsection below. To test the effect of gdv1 gene disruption on expression of MSPDBL2, the P. falciparum line NF54-GDV1Δ39 was studied (this has a truncated gdv1 coding sequence previously shown to be non-functional for promoting sexual commitment) (25) and compared with its parental line NF54 that has an intact gdv1 gene. To analyze variation in proportions of MSPDBL2+ schizonts among isolates, and test for correlations with gametocyte conversion rates, 13 other P. falciparum lines were also studied, with putative origins given here in parentheses: NF54 (an “airport malaria” isolate from Netherlands with genetic similarities to African isolates), 3D7 (cloned from NF54), F12 (derived from 3D7 with a non-functional mutated ap2-g), D10 (Papua New Guinea), D6 (Sierra Leone), Dd2 (Southeast Asia), FCC2 (China), HB3 (Honduras), R033 (Africa), Palo Alto (Uganda), 7G8 (Brazil), and GB4 (Ghana). All lines were thawed from pre-existing frozen stocks of cultures at the London School of Hygiene and Tropical Medicine and were not re-cloned prior to this study. To examine whether MSPDBL2 was expressed in the same schizonts as AP2-G, the parasite line 3D7/AP2G-HA (with a 3×HA epitope-tagged AP2-G engineered into the E5 clone of 3D7) (29) was used. Additional parasite modifications conducted in the present study are described in a separate sub-section below.

The P. falciparum lines were cultured as described previously (24), being maintained in RPMI 1640 medium supplemented with AlbuMAX II in human erythrocytes at 3% hematocrit and incubated at 37°C in atmospheric air with 5% CO₂. Late-stage parasites containing hemozoin were positively selected using MACS LD magnetic columns (46). Within 24 hours, this was followed by a second MACS purification, from which the flow through containing ring-stage parasites was collected, and returned into culture, following which parasites were allowed to develop until a large proportion was at schizont stages (24). Cells from these cultures were harvested by centrifugation and washed in phosphate-buffered saline (PBS) with 3% bovine-serum albumen (BSA), resuspended to 2% hematocrit and spotted onto multiwell slides (Hendley, Essex, UK) which were then air-dried and stored at −80°C. Mature schizonts (with at least eight nuclei) were analyzed by immunofluorescence microscopy using murine polyclonal serum specific for an N-terminal conserved region of MSPDBL2 (5, 17) to determine the proportions that were MSPDBL2+, as this gives clear positive and negative discrimination following a protocol previously described (5, 17).

Briefly, slides were fixed for 30 min with PBS/4% paraformaldehyde and permeabilized for 10 min with PBS/0.1% Triton X-100. The polyclonal anti N-terminal MSPDBL2 serum was diluted 1:200 in PBS/3% BSA and incubated for 30 min at room temperature. Subsequently, Alexa Fluor 594-conjugated anti-mouse IgG (Invitrogen, A11032) was used as secondary antibody, diluted 1:1,000 in PBS/3% BSA and incubated for 30 min at room temperature. Finally, 4′,6-diamidino-2-phenylindole (DAPI)-containing ProLong Diamond antifade mountant (ThermoFisher Scientific) was used as slides mounting media. Parasite counting was performed using a Zeiss CCD fluorescence microscope, with identical settings for each experiment. Approximately 1,000 parasites were counted and used to determine the proportions of mature schizonts expressing MSPDBL2 in each preparation, although occasionally lower numbers were counted if parasitemia was low.

For the parasite line 3D7/AP2G-HA in which the AP2-G is 3×HA epitope-tagged (29), two-color immunofluorescence was performed to discriminate schizonts expressing MSPDBL2 and AP2-G. In this case, the polyclonal murine anti N-terminal MSPDBL2 serum and a rat monoclonal α-HA antibody (Roche 3F10) were used, with fluorescine-conjugated anti-mouse IgG and Alexa Fluor 594-conjugated anti-rat IgG secondary antibodies used along with DAPI staining.

To compare the proportion of mature schizonts expressing MSPDBL2 in medium with and without added choline, multiple replicate cultures of HB3 and NF54 parasites were prepared similarly as described above and then split after the second MACS column purification (used to isolate the ring stages contained in the flow through) into separate wells with a minimal culture medium (RPMI 1640 with 25 mM HEPES, 100 mM hypoxanthine, 1 mM L-glutamine, 0.39% fatty acid-free BSA, 30 mM palmitic acid, 30 mM oleic acid) either supplemented with 2 mM choline or lacking choline. Parasites were exposed to this treatment from the ring stage of development onward and allowed to grow until the schizont stage to be analyzed by immunofluorescence assay (IFA) as described above.

The P. falciparum 3D7/iGP line was used to test the effect of GDV1 overexpression on the proportion of MSPDBL2+ schizonts. When parasites were at the ring stage, specified concentrations of Shield-1 reagent (from 0 to 1 µM, Tables S1 and S2) were added to culture media, as Shield-1 stabilizes the overexpressed GDV1-GFP-DD (the destabilization domain DD tagged to a GFP motif having been fused to GDV1 and integrated into the dispensable cg6 locus PF3D7_0709200) (22). Parasites were allowed to grow until most were at trophozoite stages (30–35 hours post invasion), and the GDV1+ proportion was analyzed using an EVOS ThermoFisher fluorescence microscope. A portion of the culture was harvested and stained with DAPI, and the GDV1+ proportion determined by differential counting of GFP-positive trophozoites as a proportion of all DAPI-positive trophozoites. Then, the parasites were allowed to grow until the schizont stage, when a portion of the culture was harvested and analyzed to determine proportions of MSPDBL2+ schizonts among mature schizonts with at least eight nuclei, as described above. The remaining parasite culture was allowed to re-invade erythrocytes, and after 38–46 hours, parasites were analyzed for gametocyte conversion using Pfs16/DAPI staining as described previously (24). Briefly, a monoclonal α-Pfs16 murine antibody 93A3A2 (47) was diluted 1:2,000 in PBS/3% BSA and incubated for 30 min at room temperature. Alexa Fluor 594-conjugated anti-mouse IgG (Invitrogen, A11032) was used as secondary antibody, diluted 1:1,000 in PBS/3% BSA, and incubated for 30 min at room temperature, and DAPI-containing ProLong Diamond antifade mountant (ThermoFisher Scientific) was used. Parasite counting was performed using a Zeiss CCD fluorescence microscope, with identical settings for each experiment. A minimum of 300 parasites were counted to determine the proportion of Pfs16-positive stage I gametocytes.

Tagging of MSPDBL2 with C-terminal TdTOM-3×HA for expression in a majority of schizonts (MSPDBL2-TAG line) was performed by modifying the endogenous mspdbl2 locus through single homologous recombination using the SLI T2A/G418 strategy (30). To construct the transfection vector, a 1,002 bp region of mspdbl2 starting 1,284 bp downstream from the start codon and lacking the stop codon was amplified (primers P3 and P4, Table S12) and inserted into a Xho1/BamH1-digested pL7M2_TdTom_T2A/G418 plasmid vector. This vector corresponds to a modified version of a pL7M2_T2A/G418 vector (48) with TdTomato sequence amplified from a pRHOPH3-RAP1-tdTom plasmid (49) inserted in the phase of the N-terminal of the 3ha_t2a_neoR (neomycin resistance) locus of the pL7M2_T2A/G418 vector previously digested with BamH1 and Kpn1 (primers P1 and P2, Table S12; Fig. S1). To generate mspdbl2-disrupted parasites expressing a short truncated version of the protein (MSPDBL2 DEL line), the SLI T2A/G418 strategy was employed, using a transfection plasmid with a truncated mspdbl2 sequence (up to 505 bp from the start codon, amplified with primers P8 and P9, Table S12) inserted in frame with a 3×HA tag and the N-terminal of the 3ha_t2a_neoR (neomycin resistance) locus of the pL7M2_T2A/G418 vector previously digested with BamH1 and Kpn1 (48) (Fig. S2). The inserted modified mspdbl2 sequences in the plasmid vector constructs are shown (Table S13).

Both constructs were transfected into purified schizonts of the 3D7/iGP P. falciparum line (23) using an AMAXA nucleofector 4D (Lonza) and P3 reagent and selected as previously described with some modifications (30). Briefly, parasites were transfected with 40 µg of purified plasmid. In addition 5 nM of WR99210 (WR) was used to positively select the transfectants. Then, the transfected parasites were subjected to selection using 1 mg/mL G418 (Cambridge Bioscience, 1557-1G) for 6 days to select integrants able to produce the tagged or truncated version of MSPDBL2. To avoid any possible reversion of the single homologous recombinaton, cultures were then routinely subjected to additional rounds of G418 selection, with one round of selection every 4 weeks. Integration was monitored by PCR using specific primers to detect the 5´ integration of the TdTOM-3×HA tag (Fig. S1) and detect the truncation in the disrupted line (Fig. S2) within the endogenous gene locus (Fig S1 and S2; Table S13). Additionally, G418 treatment for at least 6 days was performed on all cultures immediately prior to phenotyping assays, which ensured that a majority of schizonts expressed the engineered MSPDBL2.

The expression of MSPDBL2 in the tagged/disrupted lines was confirmed by immunofluorescence microscopy. To distinguish the intact and truncated MSPDBL2 expression, murine sera raised against the conserved N-terminal or C-terminal sequences region of MSPDBL2 were used as alternative primary antibodies for comparisons (1:200 in PBS/3% BSA) and incubated for 1 hour at room temperature. Following washing of slides, Alexa Fluor 594-conjugated anti-mouse IgG [1:1,000 in PBS/3% BSA (Invitrogen, A11032)] was used as secondary antibody and incubated for 30 at room temperature. DAPI-containing ProLong Diamond antifade mountant (ThermoFisher Scientific) was used as mounting medium and parasite imaging was performed on an Inverted Nikon Ti Eclipse fluorescence microscope.

Parasite multiplication rate assays under exponential growth conditions were performed using a previously described protocol (31, 32). Erythrocytes were obtained from three anonymous donors, stored at 4°C until use within 2 days, and each assay was performed in triplicate by culturing with erythrocytes from each donor in a separate flask. Asynchronous parasite cultures were diluted to 0.02% parasitemia and 3% hematocrit in 5 mL culture volumes, and grown over 6 days with 300 µL culture being collected for DNA extraction and quantitative PCR (qPCR) every 48 hours (days 0, 2, 4, and 6), culture media being replaced on days 2, 4, and 5. To assess the numbers of parasite genome copies at each assay timepoint, qPCR analysis was performed targeting the highly conserved single-locus Pfs25 gene (31). Quality control filtering was performed to exclude any point with less than 1,000 estimated genome copy numbers, any point either lower or more than 20-fold higher than measured for the same culture at the previous timepoint 2 days earlier, and any point among the triplicates with more than twofold difference from the others. An overall parasite multiplication rate (per 48 hours) was calculated for each assay with 95% confidence intervals using a standard linear model on the Log₁₀-transformed genome copy numbers, using GraphPad PRISM. The P. falciparum clone 3D7, previously determined to have an exponential multiplication rate of approximately eightfold per 48 hours (31), was assayed in parallel as a control in all assays.

Testing for the rank correlation between the proportion of MSPDBL2+ schizonts and gametocyte conversion rate was performed using Spearman’s rho coefficient. Comparisons of MSPDBL2+ schizont proportions between different lines were analyzed for significance using the non-parametric Mann-Whitney U-test, being performed on the multiple replicate assay measurements made for each line. For each individual biological replicate data point, the 95% confidence intervals in these proportions were calculated from the exact numbers of cells counted. Statistical analyses were performed using Prism version 9. All count data and other numbers are tabulated in the Supplementary Information files.