CRISPR/Cas9-engineered inducible gametocyte producer lines as a valuable tool for Plasmodium falciparum malaria transmission research

The malaria parasite Plasmodium falciparum replicates inside erythrocytes in the blood of infected humans. During each replication cycle, a small proportion of parasites commits to sexual development and differentiates into gametocytes, which are essential for parasite transmission via the mosquito vector. Detailed molecular investigation of gametocyte biology and transmission has been hampered by difficulties in generating large numbers of these highly specialised cells. Here, we engineer P. falciparum NF54 inducible gametocyte producer (iGP) lines for the routine mass production of synchronous gametocytes via conditional overexpression of the sexual commitment factor GDV1. NF54/iGP lines consistently achieve sexual commitment rates of 75% and produce viable gametocytes that are transmissible by mosquitoes. We also demonstrate that further genetic engineering of NF54/iGP parasites is a valuable tool for the targeted exploration of gametocyte biology. In summary, we believe the iGP approach developed here will greatly expedite basic and applied malaria transmission stage research.

M alaria is a vector-borne infectious disease caused by protozoan parasites of the genus Plasmodium. Infections with P. falciparum are responsible for the vast majority of all malaria-related morbidity and mortality in humans. Over the last two decades, concerted intervention efforts targeting both the insect vector and the parasite led to a remarkable decline in malaria cases worldwide 1 . However, progress has come to a standstill in the past few years and in 2019, malaria was still accountable for 229 million clinical cases and 409,000 deaths, primarily in sub-Saharan Africa 2 . To further reduce the spread of malaria, intervention strategies will not only have to overcome the widespread resistance of mosquitoes and parasites to insecticides and first-line antimalarial drugs, respectively, but will also have to include efficient tools that interrupt parasite transmission from the human host to the mosquito vector 3,4 .
People get infected with P. falciparum when infested female Anopheles spp. mosquitoes inject sporozoites into their skin. After reaching the liver via the bloodstream, sporozoites multiply within hepatocytes to release thousands of merozoites into circulation. Merozoites invade red blood cells (RBCs) and develop through the ring and trophozoite stage into a multinucleated schizont. After daughter cell formation, up to 32 merozoites egress from the infected RBC (iRBC) to invade and replicate inside new RBCs. Consecutive rounds of these intraerythrocytic developmental cycles (IDCs) are responsible for all disease symptoms and chronic infection. Importantly, however, during each replication cycle, a small proportion of schizonts produce sexually committed ring stage progeny that differentiate into either female or male gametocytes 5 . When taken up by a mosquito, terminally differentiated gametocytes egress from the iRBC and develop into gametes. The female gametocyte produces one macrogamete, while the male gametocyte undergoes three rapid rounds of genome replication followed by exflagellation of eight motile microgametes. After fertilisation, the zygote transforms into an ookinete that traverses the midgut epithelium to initiate sporogony, which ultimately renders the mosquito infectious to other humans. Hence, as the only forms of the parasite able to infect mosquitoes, gametocytes are an essential target for future antimalarial drugs aiming to prevent malaria transmision 4 .
Gametocytogenesis entails two major processes: sexual commitment/conversion and sexual differentiation. Sexual commitment occurs during the IDC preceding gametocyte differentiation. This understanding was gained from experiments showing that all ring stage descendants derived from a single schizont have the same fate; they either all undergo another round of intracellular replication or they all differentiate into either female or male gametocytes 6,7 . In addition to this "next cycle sexual conversion (NCC)" process, recent studies reported "same cycle sexual conversion (SCC)" where ring stages directly commit to sexual development 8,9 . Irrespective of the NCC or SCC routes, sexual conversion is triggered by an epigenetic switch that activates expression of the master transcription factor AP2-G [10][11][12][13] . In asexual parasites, heterochromatin-dependent silencing of ap2-g prevents AP2-G expression 10,[12][13][14] . In a small subset of trophozoites (NCC) or ring stages (SCC), however, the ap2-g locus gets activated by molecular mechanisms that are still largely unknown [9][10][11][12][13][15][16][17] . Gametocyte development 1 (GDV1), a nuclear protein essential for gametocytogenesis in P. falciparum 18 , plays a key role in the NCC process 16 . GDV1 is specifically expressed in sexually committed trophozoites and schizonts, where it displaces heterochromatin protein 1 (HP1) from the pfap2-g locus, thereby licensing PfAP2-G expression 16 . Consistent with the epigenetic control mechanisms regulating pfap2-g expression, sexual commitment rates vary in response to environmental changes 5 . In particular, depletion of the host serum lipid lysophosphatidylcholine (LysoPC) triggers sexual commitment in trophozoites and this response is channelled via induction of GDV1 and PfAP2-G expression 16,17 . Once expressed, PfAP2-G initiates a specific transcriptional program that drives sexual conversion and primes subsequent gametocyte differentiation 5 . While the PfAP2-G-dependent transcriptional changes in sexually committed schizonts are minor, a more pronounced gene expression signature emerges in the sexual ring stage progeny, where several dozen genes are specifically induced or repressed compared with asexual ring stages 10,12,13,15,16,19,20 .
P. falciparum sexual ring stage parasites differentiate over a period of 10-12 days into transmission-competent mature female or male gametocytes. Sexual ring stages (day 1) develop into spherical stage I gametocytes (day 2) that continuously elongate into lemon-shaped stage II (day 4), D-shaped stage III (day 6), and spindle-shaped stage IV cells (day 8), before falciform-mature stage V gametocytes are formed (day 10+) 21,22 . These morphological transitions are linked to the gradual expansion of the inner membrane complex, an endomembrane system underlying the parasite plasma membrane, and the microtubule and actin cytoskeleton networks underneath that are disassembled again at the stage IV to V transition [23][24][25][26][27] . mRNA and protein expression profiling studies conducted mainly on late-stage gametocytes [28][29][30][31][32] , but also on early stages 31,32 or across gametocyte maturation 33,34 , identified hundreds of genes and proteins differentially expressed between gametocytes and asexual blood stage parasites, or between female and male gametocytes [35][36][37][38] . Some of these genes are known or predicted to be involved in biological processes that are altered in gametocytes compared with asexual parasites, such as host cell remodelling, energy and lipid metabolism, cytoskeleton organisation, transcriptional regulation, chromatin structure and translational repression 5,39 . During infection, stage I to IV gametocytes are sequestered away from circulation, primarily in the parenchyma of the bone marrow and spleen 40,41 . The mechanisms underlying gametocyte homing to and sequestration in these extravascular niches are only poorly understood. However, the high rigidity of stage I-IV gametocyte-infected RBCs 23,42,43 , conferred by parasite-induced alterations of the RBC membrane and the underlying cytoskeletal networks 42,[44][45][46] , appears to play a primary role in gametocyte retention. Reversal of these modifications at the stage IV to V transition confers increased deformability 23,[42][43][44][45][46] , which is a likely prerequisite for the release of stage V gametocytes back into circulation 41,47 . Once in the bloodstream, stage V gametocytes circulate and remain competent for transmission to the mosquito for days/weeks 48 . The enormous transmission reservoir represented by the hundreds of millions of infected people in endemic areas, and the fact that almost all currently licensed antimalarial drugs, except primaquine, fail to kill mature gametocytes, poses major obstacles to malaria control and elimination efforts 4,49,50 .
The discovery and development of new transmission-blocking drugs and vaccines requires a detailed functional and mechanistic understanding of the molecular processes underlying essential transmission stage biology and depends on robust experimental tools for routine basic and applied research. Both prerequisites are currently not met due to the challenges associated with the in vitro cultivation of gametocytes, particularly the generation of large numbers of synchronous gametocyte stages for experimental studies: (i) P. falciparum parasites show low sexual commitment rates (<10%), resulting in low gametocyte yields 5,41 ; (ii) gametocytes are nonproliferative cells and thus are rapidly overgrown by asexual parasites; (iii) sexual commitment occurs during each consecutive IDC, which results in asynchronous gametocyte populations. Current approaches to increase sexual commitment rates in in vitro cultures are based on exposing parasites to poorly defined stress conditions such as high parasitemia and/or nutrient starvation (spent/conditioned medium) 22,51,52 . Several different protocols relying on this strategy exist and reach sexual commitment rates of 10-30% [53][54][55][56][57][58][59] . However, most of these protocols use cumbersome experimental workflows, rely on large culture volumes, are expensive, difficult to reproduce, or produce asynchronous gametocyte populations. Protocols employing LysoPCor choline-depleted minimal fatty acid medium 17 achieve sexual commitment rates in the range of 15-60%, depending on the strain used, but growth under these nutrient-restricted conditions reduces the number of sexually committed progeny 16,17,60,61 . A recent system based on an inducible promoter-swap approach at the pfap2-g locus triggers very high sexual conversion rates of up to 90%; however, gametocytes produced from this line are not infectious to mosquitoes and therefore unsuitable for research on late-stage gametocyte biology and gametocytocidal drug discovery 19 .
Here, we show that the insertion of conditional GDV1 expression cassettes into the genome of the transmissible NF54 strain transforms these parasites into inducible gametocyte producer (iGP) lines suitable for the routine mass production of infectious gametocytes. Upon targeted induction of GDV1 expression, NF54/iGP parasites consistently achieve sexual conversion rates of 75% and generate synchronous gametocyte populations at high yield across all stages of gametocytogenesis. Importantly, NF54/iGP stage V gametocytes retain their capacity to infect female Anopheles mosquitoes and produce sporozoites able to infect human hepatocytes. Furthermore, by tagging the nuclear pore protein PfNUP313, we demonstrate that further genetic engineering of NF54/iGP lines is straightforward and a promising approach to study gametocyte biology.

Results
CRISPR/Cas9-based engineering of P. falciparum NF54 inducible gametocyte producer lines based on the targeted overexpression of GDV1. Using the FKBP destabilization domain (DD) system for controllable protein expression 62 , we previously observed that conditional overexpression of plasmid-encoded GDV1-GFP-DD increases sexual conversion rates (SCRs) in the 3D7 reference strain 16 . Here, we exploited this finding in order to generate stable and marker-free inducible gametocyte producer (iGP) lines as tools to expedite basic and applied research on P. falciparum gametocyte biology and transmission.
As a proof of concept, we inserted a single GDV1-GFP-DD expression cassette into the dispensable cg6 (glp3) locus (PF3D7_ 0709200) 63 in 3D7 parasites using a two-plasmid CRISPR/Cas9based gene editing approach 16 (Fig. 1a and Supplementary Fig. 1a). To this end, we cotransfected the pHF_gC-cg6 CRISPR/Cas9 plasmid (containing expression cassettes for the positive-negative selection marker human dihydrofolate reductase fused to yeast cytosine deaminase/uridyl phosphoribosyl transferase (hDHFR-yFCU), SpCas9, and a single-guide RNA targeting the cg6 locus) and the pD_cg6_cam-gdv1-gfp-dd donor plasmid (containing the gdv1-gfp-dd transgene cassette flanked by cg6 homology regions) into 3D7 wild-type (wt) parasites ( Supplementary Fig. 1a). Characterisation of the resulting transgenic 3D7/iGP mother line and three clonal populations demonstrated that (i) the gdv1-gfp-dd transgene cassette was successfully inserted into the cg6 locus (Fig. 1a, Supplementary Figs. 1b-d and Supplementary Note 1), and (ii) the addition of the stabilising ligand Shield-1 to the culture medium caused efficient induction of GDV1-GFP-DD expression in schizonts (Fig. 1b,  Supplementary Fig. 2a and Supplementary Note 1). Importantly, IFAs probing for the expression of the gametocyte-specific marker Pfs16 64 revealed that up to 75% of parasites in the progeny of Shield-1-treated parasites [36-44 h post invasion (hpi), generation 2, day 2 of gametocytogenesis] represented early stage I gametocytes, compared with only 3-5% in the control populations cultured in the absence of Shield-1 ( Fig. 1b- Based on the encouraging results, we wanted to transfer the iGP system to NF54 parasites, the strain most widely used for the study of gametocyte biology and mosquito infection 65 . Unexpectedly, however, multiple attempts to insert the GDV1-GFP-DD expression cassette into the cg6 locus in NF54 parasites were unsuccessful as we did not obtain viable parasites after multiple transfection experiments. While the reason for this failure is unknown, we suspected that the DD-dependent degradation of GDV1-GFP-DD may be less efficient in NF54 compared with 3D7 parasites, such that parasites harbouring the transgene cassette would commit to sexual differentiation at an increased rate and therefore display a slower proliferation rate. We therefore decided to employ the glmS riboswitch approach as an additional or alternative conditional expression system to regulate the ectopic expression of GDV1 in NF54 parasites. The glmS ribozyme element is usually placed within the untranslated region of mRNAs and mediates transcript degradation when activated by glucosamine (GlcN) 66,67 . We cloned two modified versions of the pD_cg6_cam-gdv1-gfp-dd donor plasmid by inserting a glmS ribozyme element directly downstream of the gdv1-gfp-dd (pD_cg6_cam-gdv1-gfp-dd-glmS) or gdv1-gfp coding sequence (pD_cg6_cam-gdv1-gfp-glmS), respectively. Cotransfection of the pBF_gC-cg6 CRISPR/Cas9 plasmid (which carries the blasticidin deaminase (bsd) gene as a positive selection marker instead of hdhfr) with either of the donor plasmids allowed us to successfully select for transgenic NF54/iGP parasites in both instances ( Fig. 2a and Supplementary Figs. 3a and 3b). PCRs on gDNA of the NF54/iGP1 line revealed complete disruption of the cg6 locus and insertion of a single GDV1-GFP-DD-glmS expression cassette ( Supplementary Fig. 3c). However, at least a subset of parasites in the population still contained the pBF_gC-cg6 CRISPR/Cas9 and/or integrated donor plasmid concatamers. Treatment with 5-fluorocytosine (5-FC) efficiently eliminated parasites harbouring the pBF_gC-cg6 plasmid but not those carrying integrated donor plasmid concatamers ( Supplementary  Fig. 3d). The NF54/iGP2 population was free of plasmids and showed complete disruption of the cg6 locus through insertion of a single GDV1-GFP-glmS expression cassette ( Fig. 2a and Supplementary Fig. 3e).
To confirm that induction of sexual commitment via ectopic GDV1 expression has no influence on gametocyte sex ratios, we performed IFAs on stage V gametocytes (day 10) using antibodies against the female-specific protein Pfg377 69 ( Supplementary  Fig. 5d). We observed no significant difference in female/male sex ratios between NF54/iGP2_E9 gametocytes obtained via GDV1-  Supplementary Fig. 6a). The same result was obtained from a single experiment performed using the NF54/iGP1_D8 clone ( Supplementary Fig. 6b). Together, these results show that GDV1 overexpression has no effect on gametocyte sex ratios, which in turn suggests that GDV1 is not involved in the hitherto unknown sex determination pathway in P. falciparum. Finally, as a proxy to assess gametocyte viability, we quantified male stage V gametocyte exflagellation on days 10, 13, and 14 of gametocytogenesis and demonstrate that both NF54/iGP1_D8 and NF54/iGP2_E9 gametocytes exflagellated as efficiently as NF54 wt gametocytes, and similar results were obtained for the corresponding NF54/iGP mother lines (Supplementary Figs. 6c and 6d).
NF54/iGP1 and NF54/iGP2 gametocytes are infectious to mosquitoes and produce viable sporozoites that infect human hepatocytes. To test if NF54/iGP gametocytes retained their capacity to undergo fertilisation and further life cycle progression in the mosquito vector, we fed stage V gametocytes to female Anopheles stephensi mosquitoes using Standard Membrane Ffeeding Aassays (SMFAs) 65,70 . To this end, we triggered sexual conversion in clones NF54/iGP1_D8 (−GlcN/+Shield-1) and NF54/iGP2_E9 (−GlcN) as outlined in Fig. 2c, and maintained the ring stage progeny (day 1 of gametocytogenesis) in culture for 14 days with daily medium changes to ensure complete differentiation into mature stage V gametocytes. For each of the two populations, separate SMFAs were performed on days 10, 13, and 14 of gametocytogenesis, and the entire experiment was repeated with a second batch of independently produced gametocytes. Eight days after the feeds, 20 mosquitoes each were dissected and midgut oocysts counted by microscopy. Both NF54/ iGP clones infected mosquitoes as efficiently as NF54 wt gametocytes on each of the three days of feeding, with 87.5-97.5% (NF54/iGP1_D8), 92.5-100% (NF54/iGP2_E9), and 90-100% (NF54 wt) of mosquitoes carrying midgut oocysts (Fig. 3a).
Mosquitoes infected with NF54/iGP1_D8 gametocytes consistently developed about half as many oocysts compared with the NF54 wt control, and for both of these lines the highest oocyst numbers were observed in mosquitoes fed with day 13 gametocytes (median=13 oocysts/mosquito for NF54/iGP1_D8; med-ian=27 oocysts/mosquito for NF54 wt) (Fig. 3a). Interestingly, NF54/iGP2_E9 gametocytes achieved even higher oocyst intensities compared with NF54 wt gametocytes and were most infectious on day 13 (median = 37 oocysts/mosquito) and day 14 (median = 47 oocysts/mosquito) (Fig. 3a). Gametocytes of the NF54/iGP1 and NF54/iGP2 mother lines were also infectious to mosquitoes. While infections with NF54/iGP1 gametocytes resulted in infection rates and oocyst intensities similar to those obtained with the clonal NF54/iGP1_D8 line, NF54/iGP2 gametocytes infected fewer mosquitoes and developed fewer oocysts on all three feeding days ( Supplementary Fig. 7a).
Together, these results demonstrate that the NF54/iGP1_D8 and NF54/iGP2_E9 clones, and the NF54/iGP1 mother line, complete their life cycle in the mosquito vector as efficiently as NF54 wt parasites and produce sporozoites able to infect and mature within human liver cells.
Further genetic manipulation of NF54/iGP parasites for the targeted investigation of gametocyte biology. To explore NF54/ iGP parasites as a tool to study specific aspects of gametocyte biology, we engineered NF54/iGP2 parasites expressing a mScarlet-tagged version of the nuclear pore protein PfNUP313 ( Supplementary Fig. 8). Nuclear pores are large macromolecular complexes that are embedded in the nuclear envelope and consist of approximately 30 different nucleoporins (NUPs) (each present in multiple copies) 71 . Nuclear pore complexes (NPCs) act as essential gateways for molecular transport into and out of the nucleus and play additional crucial roles in the regulation of gene expression and genome organisation 71,72 .
In malaria parasites, nuclear pores have hardly been studied and only six NUPs have been identified so far [73][74][75][76][77] , of which four have been localised in P. falciparum asexual blood stage parasites [PfNUP116 (PF3D7_1473700) 74 , PfNUP221/PfNUP100  77 , PfNUP313 (PF3D7_1446500) 73 , and PfSEC13 (PF3D7_1230700) 76 ]. We were interested in visualising nuclear pores in gametocytes for two reasons. First, the number and distribution of NPCs within the nuclear envelope of asexual parasites changes dramatically as they progress through the IDC 74,78,79 . The functional relevance of these dynamic NPC expression patterns is currently unknown, but may be linked to regulatory strategies of stage-specific gene expression and/or the coordination of nuclear segregation during schizogony 78 . We therefore wished to learn how NPC abundance and localisation compares in differentiating gametocytes. Second, early studies investigating gametocyte morphology at the ultrastructural level suggested a pronounced nuclear dimorphism between female and male gametocytes 26,80 , but this has not been further explored in any great detail. To revisit this intriguing observation, we wanted to visualise the outer confines of the parasite nucleus throughout gametocytogenesis by fluorescence microscopy. Since P. falciparum lacks homologues of known nuclear envelope-associated proteins such as lamin, we used the NPC as a surrogate marker for the nuclear membrane.
To obtain NF54/iGP2_NUP313-mSc parasites, we cotransfected the pBF_gC-nup313 CRISPR/Cas9 plasmid and the pD_nup313-mScarlet donor plasmid into the NF54/iGP2 line ( Supplementary Fig. 8). Transgenic parasites were successfully selected on BSD-S-HCl and PCRs on gDNA confirmed correct editing of the nup313 gene and absence of parasites carrying the wt locus ( Supplementary Fig. 8). Validation of the NF54/ iGP2_NUP313-mSc line by live-cell fluorescence imaging verified the expected localisation of nuclear pores across the different IDC stages as determined in previous studies 74,78,79 (Fig. 4a). PfNUP313-mScarlet localised to a single region adjacent to the DAPI-stained area in merozoites and early ring stages. The increased number and even distribution of NPCs within the nuclear envelope in trophozoites and early schizonts was reflected by a circular perinuclear pattern of PfNUP313-mScarlet foci directly adjoining the genetic material. In late schizonts, the number of PfNUP313-mScarlet signals decreased again to one or two per nucleus. To assess PfNUP313-mScarlet localisation in gametocytes, we induced sexual commitment in NF54/ iGP2_NUP313-mSc parasites according to the protocol outlined in Fig. 2c and performed live-cell fluorescence imaging for all five gametocyte stages. As shown in Fig. 4b, PfNUP313-mScarlet signals were abundant and surrounded the often elongated area of nuclear DNA in a dot-like fashion in all five gametocyte stages, similar to the localisation pattern observed in trophozoites (Fig. 4a). However, in stage II to V gametocytes, the NPC signals frequently stretched away from bulk chromatin, suggestive of rounded expansions (pink arrowheads) and narrow lateral extensions (yellow arrowheads) of the nuclear envelope that sometimes reached close to the cellular poles (Figs. 4b and 5, Supplementary Fig. 9). Primarily in late-stage gametocytes, we also observed nuclei with two clearly separate PfNUP313delineated regions, of which either both (white arrowheads) or only one (blue arrowhead) contained genetic material detectable by Hoechst staining (Fig. 5). In summary, our findings show that gametocyte nuclei contain a similar or even higher density of NPCs compared with trophozoites, undergo profound changes in nuclear shape and size, and often contain regions of low chromatin density.

Discussion
Here, we engineered robust inducible gametocyte producer (iGP) lines that will allow investigating gametocyte biology with a similar level of routine, depth, and detail hitherto reserved only for the study of asexual blood stage parasites. We achieved this by inserting a single conditional GDV1 expression cassette into the genomes of the 3D7 and NF54 strains. In these parasites, a single pulse of ectopic GDV1 overexpression in trophozoites and schizonts is sufficient to induce sexual commitment in a high proportion of parasites. Importantly, this approach mimics the natural process of GDV1-dependent sexual conversion. GDV1 is a specific and essential activator of PfAP2-G expression, and induction of endogenous GDV1 expression in trophozoites and schizonts is part of the inherent pathway of sexual commitment 16,18,81 . Furthermore, we have previously shown that the temporary activation of ectopic GDV1 expression triggers the typical PfAP2-G-dependent transcriptional cascade of sexual conversion and early gametocyte differentiation and has no effect on the transcription of other genes 16 .
The iGP lines described in this study overcome the limitations of current protocols used for the bulk preparation of gametocytes. First, induction of sexual commitment is entirely independent of cumbersome and unreliable culture-handling protocols and only requires adding Shield-1 and/or removing GlcN from the culture medium to induce ectopic GDV1 expression. Second, the targeted induction of ectopic GDV1 expression triggers consistent sexual commitment rates of up to 75%, which allows generating large numbers of gametocytes even from small culture volumes. Third, sexual commitment is induced in a synchronous population of asexual blood stage parasites, which leads to synchronous sexual differentiation in the progeny and therefore facilitates preparing stage-specific gametocyte populations across gametocyte maturation. Fourth, gametocyte synchronicity and yield are highly reproducible between experiments; starting with approximately 0.75 × 10 7 ring-stage parasites in the commitment cycle (5% haematocrit, 1.5% parasitemia) routinely delivers up to 2 × 10 8 stage V gametocytes per 10 ml of culture in a single induction experiment (Fig. 6), which increases gametocyte yields by at least one order of magnitude compared with previous reports 53,58,82 . Because gametocyte production can be induced once every second day from a synchronous asexual feeder culture, the iGP lines Fig. 3 NF54/iGP1_D8 and NF54/iGP2_E9 gametocytes complete their life cycle in the mosquito vector and produce infectious sporozoites. a NF54/ iGP1_D8 (orange), NF54/iGP2_E9 (blue), and NF54 wt control stage V gametocytes (green) were fed to female Anopheles stephensi mosquitoes on day 10, 13, and 14 of gametocytogenesis in two independent SMFA experiments. The violin plots show the distribution of the number of oocysts detected in each of the 20 mosquitoes dissected per feed, with open circles and triangles representing data from SMFA replicates 1 and 2, respectively (left y axis). The median (thick red line) and upper and lower quartiles (thin red lines) are indicated. Closed red circles represent the mean oocyst prevalence (number of infected mosquitoes) determined for each of the two replicate feeds (right y axis). NF54 wt day 10 and day 13 gametocytes were only included in SMFA replicate 1, and for SMFA replicate 2, only 10 mosquitoes infected with NF54 wt day 14 gametocytes have been dissected. b Mean number of salivary gland sporozoites per oocyst (left y axis) and per mosquito (closed red circles; right y axis) 17 days after infection with NF54/iGP1_D8 (orange), NF54/iGP2_E9 (blue), and NF54 wt control day 14 gametocytes (green) (SMFA replicate 2 data). Values represent the results from a single experiment (≥26 mosquitoes dissected per infected batch). c Confocal microscopy IFA images showing intracellular parasites after infection of primary human hepatocytes with NF54/ iGP1_D8, NF54/iGP2_E9, and NF54 wt control sporozoites. Parasites were stained with α-PfHSP70 (cytosol; red) and α-PfEXP2 antibodies (parasitophorous vacuolar membrane; purple). α-GFP antibodies were used to test for potential ectopic expression of GDV1 in liver stages. Nuclei were stained with DAPI. Images are representative of a single experiment. Scale bar, 18 µm.
provide a constant rich source of synchronous gametocyte populations for experimental studies.
Similar to the recently developed 3D7/E5ind line that uses a conditional promoter-swap approach to induce PfAP2-G expression 19 , the 3D7/iGP line produces exflagellation-defective male stage V gametocytes and is therefore of limited value for the investigation of late-stage gametocyte biology and gametocytocidal drug discovery. However, 3D7/iGP parasites still provide a useful tool to dissect the molecular events linked to sexual conversion and early gametocyte differentiation. The 3D7/E5ind line is superior in this regard as it displays virtually zero background sexual commitment and achieves sexual conversion rates of up to 90% upon induction of PfAP2-G expression, which allows identifying molecular signatures of sexually committed parasites with high precision 19 . 3D7/iGP parasites will be useful to complement such studies and to investigate the mechanisms involved in GDV1-dependent activation of pfap2-g.  In contrast to the above-mentioned cell lines, NF54/iGP gametocytes are infectious to mosquitoes. Our comprehensive characterisation of NF54/iGP1_D8 and NF54/iGP2_E9 validated these clonal lines as most promising tools for future research. Both clones infect mosquitoes and produce oocysts and infectious salivary gland sporozoites as efficiently as NF54 wt parasites. Furthermore, the mosquito infection rates and oocyst/sporozoite intensities achieved in this study are consistent with SMFA outcomes usually obtained with NF54 wt gametocytes [83][84][85] . Moreover, NF54/iGP1_D8 and NF54/iGP2_E9 parasites are confirmed plasmid-and marker-free and therefore allow further rounds of genetic manipulation using any of the three drug resistance markers routinely used for selection of transgenic parasites (hdhfr, bsd, yeast dihydroorotate dehydrogenase). Manipulating genes in NF54/iGP parasites provides the great advantage that the study of gene and protein function by phenotypic, cell biological, biochemical, or structural analyses can routinely be performed not only in asexual blood stage parasites but readily also in a large number of isogenic gametocytes. Owing to these favourable properties, NF54/iGP1_D8 and NF54/iGP2_E9 parasites offer numerous opportunities for basic and applied malaria transmission stage research ( Fig. 6; see also Supplementary Table 2 for a comparison of both clones and recommendations for their use in future research). For instance, NF54/iGP parasites will facilitate the targeted dissection of molecular mechanisms underlying the specific biology of gametocytes. They will also lend themselves for time-resolved highthroughput profiling experiments to generate comprehensive transcriptomics, (phospho-)proteomics, epigenomics and metabolomics reference and experimental datasets for each stage of gametocyte development. Indeed, NF54/iGP2 parasites have recently been used for in-depth comparative complexome profiling experiments, revealing crucial insight into the differential abundance and composition of mitochondrial respiratory chain complexes between asexual blood stage parasites and gametocytes 86 . Importantly, NF54/ iGP parasites will also simplify and streamline high-throughput drug screening campaigns to identify gametocytocidal compounds and support the preclinical development of transmission-blocking drugs and vaccines. NF54/iGP gametocytes may further provide a reliable resource for research on mosquito and liver stage parasites and for the optimisation of protocols aiming to produce zygotes, ookinetes, oocysts, and sporozoites in vitro [87][88][89] (Fig. 6). Transferring the iGP approach to other P. falciparum strains could facilitate the systematic analysis of potential strain-dependent differences in gametocyte biology, sensitivity to antimalarial drugs, infectiousness to mosquitoes, or the capacity to undergo cross-fertilisation.
We illustrated the feasibility of subjecting NF54/iGP parasites to a second round of genetic engineering by tagging the nuclear pore component PfNUP313 in the NF54/iGP2 line. Our fluorescence microscopy data demonstrate marked differences in NPC abundance and distribution in the nuclear envelope, as well as in nuclear morphology between asexual blood stage parasites and gametocytes. The results obtained from asexual parasites recapitulate previous findings obtained by high-resolution microscopy 74,78,79 . These studies showed that merozoites and ring stage parasites possess only three to seven nuclear pores that are closely clustered in one region of the nuclear membrane. As ring stage parasites develop into trophozoites, the number of NPCs increases and up to 60 NPCs are evenly distributed throughout the nuclear envelope. During schizogony, the number of pores per nucleus decreases with increasing numbers of nuclei formed, such that in late schizonts, each nucleus of the developing merozoites again possesses only a small number of clustered nuclear pores 74,78,79 . In contrast, we found that gametocyte nuclei contain a stable and relatively high number of evenly distributed NPCs throughout all stages of gametocyte development, similar to what is observed for trophozoite nuclei. These observations are consistent with the open chromatin conformation and high transcriptional activity observed in both life cycle stages [90][91][92] and are suggestive of a high demand for nucleocytoplasmic shuttling during periods of growth. Importantly, because NPCs serve as a surrogate marker delineating the nuclear envelope, our results also allow us to conclude that the nucleus in stage II to V gametocytes (i) undergoes marked morphological transformations reflected in narrow lateral extensions along the longitudinal axis of the gametocyte or rounded expansions of the nuclear envelope; and (ii) is frequently considerably larger than what might be extrapolated from the Hoechst-stained bulk of genetic material, in line with observations made in P. berghei gametocytes 75 . In some latestage gametocytes, we even observed two distinct NPCdemarcated regions, of which one or both stained positive with Hoechst. Together, these data suggest that nuclear reorganisation and functional genome compartmentalisation plays an important role in transcriptional reprogramming during gametocytogenesis and/or in the preparation for genome endoreplication and fertilisation in male and female stage V gametocytes, respectively. Interestingly, the results obtained from transmission electron microscopy studies suggested that the nuclei of female and male gametocytes differ markedly both in size and shape, with the female nucleus being generally oval-shaped and comparably small and the male nucleus appearing substantially larger with a lobular shape and distending toward the poles of the gametocyte 26,80 . It is therefore tempting to speculate that the cells containing enlarged and irregularly shaped nuclei detected in our study may primarily represent male gametocytes. However, although differential haemozoin crystal distribution has been proposed as a diagnostic feature to distinguish between the two sexes (dispersed in males, clustered in females) 21,26 , we observed ambiguous pigment patterns and irregular nuclear shapes in gametocytes with clustered or dispersed haemozoin distribution alike (Figs. 4b and 5 and Supplementary Fig. 9), suggesting that the nucleus undergoes marked transformations in both male and female gametocytes. Notwithstanding this uncertainty, our analysis of NF54/iGP2_NUP313-mSc parasites revealed new insight into the intriguing morphological features of gametocyte nuclei based on standard fluorescence microscopy and provides an excellent starting point for further detailed cytological and functional investigations of nuclear biology in P. falciparum transmission stages. Fig. 4 Visualisation of nuclear pore distribution in NF54/iGP2_NUP313-mSc asexual blood stage parasites and gametocytes. a Schematic maps of the disrupted cg6 locus carrying a single inducible GDV1-GFP-glmS expression cassette and the tagged nup313 locus in the double-transgenic NF54/ iGP2_NUP313-mSc line are shown on top. The 5′ and 3′ homology regions used for CRISPR/Cas9-based genome editing are shown in orange. Live-cell fluorescence microscopy images showing the localisation of NUP313-mScarlet (red) in asexual blood stage parasites. ER/LR, early/late ring stage; T, trophozoite; ES/LS, early/late schizont; M, merozoite. DIC, differential interference contrast. Nuclei were stained with DAPI. Images are representative of three biologically independent experiments. Scale bar, 5 µm. White frames refer to the magnified view presented in the rightmost images (scale bar, 2 µm). b Live-cell fluorescence microscopy images showing the localisation of NUP313-mScarlet (red) in stage I to V gametocytes. Lateral extensions of the nucleus away from Hoechst-stained bulk chromatin are highlighted by yellow arrowheads. I-V, stage I to V gametocytes. DIC, differential interference contrast. Nuclei were stained with Hoechst (blue). Images are representative of four biologically independent experiments. Scale bar, 5 µm. White frames refer to the magnified view presented in the rightmost images (scale bar, 2 µm).  In summary, we engineered marker-free inducible gametocyte producer lines that facilitate the routine mass production of synchronous gametocyte populations using a simple and reproducible experimental setup. We demonstrate that NF54/iGP1_D8 and NF54/iGP2_E9 parasites complete their life cycle in the mosquito vector and produce sporozoites that are infectious to human hepatocytes. We also showed that further genetic engineering of NF54/iGP parasites is straightforward. Hence, we believe the iGP approach developed in this study will become an invaluable and broadly applicable tool for fundamental, applied and translational research in the field of malaria-transmission biology.
The pHF_gC-cg6 CRISPR/Cas9 plasmid was generated by T4 DNA ligasedependent insertion of annealed complementary oligonucleotides (11 F, 11 R) encoding the single-guide RNA (sgRNA) target sequence sgt_cg6 along with compatible single-stranded overhangs into BsaI-digested pHF_gC 16 . The sgt_cg6 target sequence (gcacaaatataaattaaatt) is positioned 24-44 bp downstream of the start codon of the cg6 (glp3) gene and has been designed using CHOPCHOP 94 . pBF_gC-cg6 was generated in the same manner using the pBF_gC vector 16 .
To tag the nucleoporin NUP313 (PF3D7_1446500) with mScarlet 96 in NF54/ iGP2 parasites, the following CRISPR/Cas9 and donor plasmids were cloned. The pBF_gC-nup313 vector was generated by T4 DNA ligase-dependent insertion of annealed complementary oligonucleotides (18 F, 18 R) encoding the sgRNA target sequence sgt-nup313 along with compatible single-stranded overhangs into BsaIdigested pBF_gC 16 Fig. 6 Scheme depicting the simple induction protocol for the routine mass production of NF54/iGP gametocytes and potential applications for future research. Addition of Shield-1/removal of GlcN (NF54/iGP1) or removal of GlcN (NF54/iGP2) from a synchronous ring stage culture triggers sexual commitment in trophozoites and produces progeny consisting of up to 75% sexual ring stage parasites. Addition of 50 mM GlcNAc to the culture medium for the next six days eliminates the remaining asexual parasites. Expected numbers of parasite-infected RBCs (#) and percent parasitaemia (% P) of total (black letters), asexual (grey letters), and sexual (red letters) parasites in the progeny routinely obtained from a 10 ml culture at 5% haematocrit (HC) and 1.5% starting parasitaemia are indicated. Asexual parasites are depicted in grey, sexually committed parasites and gametocytes are depicted in purple (females) and green (males). asS/scS asexual/sexually committed schizont, asR/scR asexual/sexually committed ring stage, T trophozoite, I-V gametocyte stages I-V, D1-D10 days 1-10 of gametocyte maturation, G gametes, Z zygote, Ok ookinete, Oc oocyst, Sp sporozoites, hepS intrahepatic schizont. Possible applications of NF54/iGP lines for basic, applied, and translational research on P. falciparum gametocytes and mosquito-stage parasites are listed below the schematic.
positioned at 91-110 bp downstream of the nup313 coding sequence. The pD_nup313-mScarlet donor vector is a result of Gibson assembly of five PCR fragments encompassing (1) a 941-bp nup313 5′ HR spanning bps 8063-9003 of the nup313 coding sequence (amplified from NF54 gDNA using primers 13 F and 13 R), (2) the mScarlet gene (amplified from a P. falciparum codon-adjusted synthetic sequence (Genscript) (Supplementary Fig. 10) using primers 14 F and 14 R), (3) the P. falciparum histidine-rich protein 2 (hrp2) terminator sequence (amplified from pBF_gC with primers 15 F and 15 R), (4) a 1000 bp nup313 3′ HR spanning the region 90-1089 bp downstream of the nup313 coding sequence (amplified from NF54 gDNA using primers 16 F and 16 R), and (5) the vector backbone amplified from pBF_gC using primers 17 F and 17 R. All oligonucleotide sequences used for cloning and Sanger sequencing are provided in Supplementary  Table 1.
Parasite transfection and selection of transgenic lines. Parasites were transfected and gene-edited parasites selected as described by Filarsky and colleagues 16 . Briefly, RBCs derived from 5 ml of a synchronous ring stage culture (5-10% parasitaemia) were co-transfected with the matching pair of CRISPR/Cas9 and donor plasmids (50 µg each) using a Bio-Rad Gene Pulser Xcell Electroporation System (single pulse, 310 V, 250 µF). Twenty-four hours after transfection, 3D7/ iGP parasites were selected with 5 nM WR99210 (WR) for the following six days and NF54/iGP1, NF54/iGP2, and NF54/iGP2_NUP313-mSc parasites were selected with 2.5 µg/ml blasticidin-S-HCl (BSD) for the following eleven days. Thereafter, all transfected parasites were cultured in the absence of drug pressure, until stably propagating populations were established. PCR on gDNA was performed to confirm proper genome editing and absence of plasmid DNA. The 3D7/iGP and NF54/iGP1 lines, which still retained the pHF_gC or pBF_gC CRISPR/Cas9 plasmids expressing the negative selection marker yFCU fused to hDHFR or BSD deaminase, respectively, were treated with 40 µM 5-fluorocytosine (5-FC) to obtain plasmid-free populations. 3D7/iGP, NF54/iGP1, and NF54/iGP2 lines were cloned out by limiting dilution using plaque assays 97 . All oligonucleotide sequences used for PCR on parasite gDNA and plasmid DNA are provided in Supplementary  Table 1.
Live-cell fluorescence microscopy. To perform live-cell fluorescence imaging of NUP313-mScarlet presented in Fig. 4 and Supplementary Fig. 9, asexual blood stages were stained with 1 µg/mL DAPI (Biomol) in RPMI and gametocytes were stained with 4.5 µg/mL Hoechst33342 (Chemodex) in PBS at 37°C for 15 min. Stained cells were pelleted at 1000 g for 1 min and the pellet resuspended in an equal volume of supernatant. About 10 µL of the sample were placed on a microscopy slide and covered with a cover slip. Fluorescence microscopy images were acquired with a Leica DM6 B microscope equipped with Leica DFC9000 GT camera using a 100x immersion oil objective (total magnification = 1000x). Filter block settings: mScarlet (Ex. 542-585 nm; Em. 604-644 nm), DAPI/Hoechst (Ex. 325-375; Em. 435-485). All images were acquired with the Leica Application Suite X software (LAS X) and processed using Adobe Photoshop CS2 with identical settings.
To perform live-cell fluorescence imaging of NUP313-mScarlet presented in Fig. 5, gametocytes were stained with 4.5 µg/mL Hoechst33342 (Sigma-Aldrich) in PBS at 37°C for 20 min. Stained cells were pelleted at 300 g for 1 min and the pellet resuspended in an equal volume of the supernatant. About 3 µL of the sample were placed on a microscopy slide, mixed with 3 µL Vectashield, and covered with a cover slip. Images were acquired with a Leica DM 5000B microscope and a Leica DFC 345 FX camera using a 100x immersion oil objective (total magnification =1250x). Filter block settings: mScarlet (Ex. 543/22 nm; Em. 593/40 nm), Hoechst (Ex. 377/50 nm; Em. 447/60). All images were acquired using the Leica Application Suite software (LAS 4.9.0) and processed using ImageJ (version 1.52n) using identical settings.
For gametocyte maturation assays, parasites were induced for sexual conversion as describe above. Culture medium containing 50 mM GlcNAc was added to the ring stage progeny (generation 2; day 1 of gametocytogenesis) and changed daily for six days to eliminate asexual parasites 53,68 . From day seven onward, gametocytes were cultured with normal culture medium. Gametocyte stages and gametocytaemia were assessed by visual inspection of Giemsa-stained thin blood smears prepared until day 12 of gametocytogenesis using a 100x immersion oil objective (total magnification = 1000x).
To quantify male gametocyte exflagellation, 10 μL of stage V gametocyte cultures was added to 10 μL of 50 μM xanthurenic acid in incomplete RPMI 160 medium, placed into a Neubauer slide with rhodium-coated chamber bottom, and incubated for 10-15 min at room temperature. The exflagellation events were counted and the average per µl and 1% parasitaemia calculated.
Standard membrane feeding assays (SMFAs). On day 10, 13, and 14 of gametocytogenesis, aliquots of stage-V gametocyte cultures were tested in the Standard Membrane Feeding Assay (SMFA) 65 according to an established workflow 70 . Briefly, after a final change of culture medium on the day of mosquito feeding, a small volume of the gametocyte culture was used to prepare the blood meal and to make blood smears. About 300 µL of the gametocyte culture (approximately 2.5% hamatocrit and 3-10% gametocytemia) was mixed with prewarmed 180 µL of packed RBCs and the cells were pelleted for 20 sec. The supernatant was carefully removed and the RBC pellet resuspended in 150 µL of prewarmed human serum. Small membrane feeders were used to feed female Anopheles stephensi mosquitoes separately with each of the prepared gametocyte populations (approximately 0.15-0.5% gametocytemia and 50% hematocrit). For sporozoite production experiments, midi membrane feeders were used with three times the above volumes with the ratios consistent to feed 100 mosquitoes 65 . On day 8 after feeding, midguts of 20 mosquitoes per feed were dissected and stained for oocysts using 3% mercurochrome. The numbers of oocysts per midgut were counted by microscopy (1000x magnification). For sporozoite collection on day 17, at least 25 mosquitoes were dissected with salivary glands collected and crushed in complete Williams B media at room temperature. Sporozoites were counted using a Neubauer slide with rhodium-coated chamber bottom and used directly for hepatocyte infection assays.
Hepatocyte infection assay. Fresh human hepatocytes were isolated from patients undergoing a partial hepatectomy and seeded (6 × 10 4 hepatocytes/well) onto collagen-coated 96-well black/clear flat-bottom TC-treated imaging microplates (Corning, #353219) in complete Williams B medium and incubated at 37°C in 5% CO 2 99 . On day 3 after seeding, sporozoites were added (6 × 10 4 sporozoites/well), spun down at 500 g for 5 min, and allowed to invade hepatocytes for three hours. The wells were washed in complete Williams B medium and incubated at 37°C in 5% CO 2, with the medium changed every day 99 . Plates were fixed with ice-cold methanol on days 3 and 5 and used for IFA analysis as described above.
Ethical statement. Primary human liver cells were freshly isolated from remnant surgical material. The samples were anonymized, and general approval for use of remnant surgical material was granted in accordance with the Dutch ethical legislation as described in the Medical Research (Human Subjects) Act and confirmed by the Committee on Research involving Human Subjects, in the region of Arnhem-Nijmegen, the Netherlands.
Statistics and reproducibility. All data from assays quantifying sexual commitment rates, gametocyte staging, gametocyte sex ratios, exflagellation rates, mosquito infection rates, and oocyst and sporozoite production are represented in graphs showing individual data points and the mean with error bars defining the standard deviation, or the median and interquartile ranges as indicated in the figure legends and the Source data file. To test for statistical significance of the difference in gametocyte sex ratios, a paired two-tailed Student's t test was used as indicated in the figure legend. The exact number of biological replicates performed per experiment and the number of cells analysed per sample are indicated in the figure legends and the Source data file. Data were analysed using Excel 2016 and plotted using GraphPad Prism (version 8.2.1).
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
All data generated or analysed during this study are included in this published article and its associated Supplementary Information and Source data files. The associated Source data file contains the raw data underlying the graphs presented in Figs. 1d, 2b, 2e, 2g, 3a, 3b and Supplementary Figs. 2b, 5a, 6b-d, 7a-b. Correspondence and requests for materials should be addressed to T.S.V. Source data are provided with this paper.