A Plasmodium berghei putative serine-threonine kinase 2 (PBANKA_0311400) is required for late liver stage development and timely initiation of blood stage infection

ABSTRACT In Plasmodium, protein kinases govern key biological processes of the parasite life cycle involved in the establishment of infection, dissemination and sexual reproduction. The rodent malaria model Plasmodium berghei encodes for 66 putative eukaryotic protein kinases (ePKs) as identified through modelling domain signatures and are highly conserved in Plasmodium falciparum. We report here the functional characterisation of a putative serine-threonine kinase PBANKA_0311400 identified in this kinome analysis and designate it as Pbstk2. To elucidate its role, we knocked out Pbstk2 locus and performed a detailed phenotypic analysis at different life cycle stages. The Pbstk2 knockout (KO) was not compromised in asexual blood stage propagation, transmission and development in the mosquito vector. The Pbstk2 KO produced viable salivary gland sporozoites that successfully transformed into exo-erythrocytic forms (EEFs) and were morphologically indistinguishable from wild-type GFP (WT GFP) with regard to size and shape until 48 h. An intravenous dose of 1×103 Pbstk2 KO sporozoites in C57BL/6 mice failed to establish blood stage infection and a higher dose of 5X103 showed a 2–3 day delay in prepatency as compared to WT GFP parasites. Consistent with such an observation, analysis of in vitro EEF development at 62 h revealed that the hepatic merozoite numbers were reduced to nearly 40% as compared to WT GFP and showed meagre expression of MSP1. Our studies provide evidence for the role of PbSTK2 in late liver stage development and for the successful establishment of a timely blood stage infection.


INTRODUCTION
Malaria is a mosquito-borne infectious disease caused by a protozoan parasite that belongs to the genus Plasmodium. The parasite kills nearly half a million people annually with deaths predominantly occurring in sub-Saharan Africa (WHO, 2017). The parasite infects a mammalian host via the bite of a female Anopheles mosquito that inoculates sporozoites into the skin during a blood meal (Sinnis and Zavala, 2008). The sporozoites make their way to the liver and develop into exoerythrocytic forms (EEFs) inside hepatocytes. After several rounds of asexual reproduction, the hepatic merozoites are released into bloodstream (Prudencio et al., 2006) to initiate an erythrocytic cycle, a phase that is responsible for all clinical manifestations of malaria. Gametocytes are the terminal stages of a parasite developing within erythrocytes and do not undergo further development in the mammalian host until they arrive in the mosquito gut. Within the mosquito midgut, the parasites undergo sexual reproduction, culminating in the production of thousands of infectious sporozoites. The sporozoites migrate to salivary glands and reside there to initiate new infection cycle in the mammalian host (Matuschewski, 2006).
Plasmodium parasites have evolved distinct kinase families with novel domain structures and biochemical features (Ward et al., 2004). These signalling molecules play a key role in the regulation of several physiological processes (Solyakov et al., 2011). In general, phosphorylation of specific amino acid residues like serine (Ser), threonine (Thr), tyrosine (Tyr), histidine (His), and aspartate (Asp) affects the activity of target proteins either by bringing a conformational change in its active site or regulating proteinprotein interactions (Pereira et al., 2011). The systematic functional investigation of Plasmodium berghei kinome by reverse genetic approach revealed that nearly two-thirds of the P. berghei kinases were essential (Tewari et al., 2010). While the possibility of targeting kinases essential for Plasmodium development in vector host may not be feasible, nonetheless several kinases seem to regulate the transmission of malaria to mosquitoes and the forms of parasite that are infective to hepatocytes can only be obtained from mosquito stage (Tewari et al., 2010). Thus it is imperative that an indepth functional investigation of kinase mutants be done at all life cycle stages for all 'possibly essential kinases' such that the importance of the same kinase playing a role at multiple life cycle stages of the parasite is not overlooked and those critical for establishment of malaria infection in a mammalian host is not undermined. To date, only a few protein kinases have been identified that are required for Plasmodium liver stage development. The lipid kinase, phosphatidylinositol-4-OH kinase [PI(4)K] is required for hypnozoite formation in a Plasmodium cynomolgi (McNamara et al., 2013). Two mitogen-activated protein kinases (MAPKs) have also been identified in P. berghei and are designated as PbMAPK1 and PbMAPK2. While both the PbMAPK genes are transcribed during P. berghei liver stage development, and PbMAPK1 localises to the cytomere stage, depleting its locus did not affect the parasite viability in the liver stages (Wierk et al., 2013). In P. berghei, the cGMP-dependent protein kinase (PKG) have been implicated in liver stage schizont development, in addition to its role in ookinete differentiation and motility. The Plasmodium falciparum orthologue of PKG was shown to be required for gametogenesis and rupture of asexual blood stage schizonts (Hopp et al., 2012). Small molecule inhibitors active against liver-stage expressed kinases may offer more realistic chemotherapy as it may block the onset of clinical disease. Indeed, studies in this direction demonstrated that both genetic ablation (Falae et al., 2010) and target based drug delivery (Panchal and Bhanot, 2010) against Plasmodium kinases uniquely expressed in liver stages can inactivate pre-erythrocytic stages (Panchal and Bhanot, 2010;McNamara et al., 2013). For example, conditional depletion of cGMP dependent protein kinases (PKG) in sporozoite stage resulted in arresting the parasite at late liver stages that suffered from an inability to generate infectious merosomes, and mice infected with PKG mutants developed immunity that conferred protection against subsequent sporozoite challenge (Falae et al., 2010). Further PKG inhibitors effectively diminished sporozoite infectivity demonstrating the exciting feasibility of using kinase inhibitors as pre-erythrocytic antimalarials (Panchal and Bhanot, 2010). Also, a recent study demonstrated effective inhibition of P. cynomolgi hypnozoites by imidazopyrazines (McNamara et al., 2013).
In order to ascertain function to other kinases uniquely expressed in the pre-erythrocytic stages, we selected a putative serinethreonine kinase PBANKA_031140 for our investigation. Previous findings have shown that the P. falciparum orthologue of PBANKA_031140 was detected in the proteomic analysis of salivary gland sporozoites (Lasonder et al., 2008). Since salivary gland sporozoites are infective forms of the parasite to the mammalian hepatocytes, we wanted to investigate if sporozoite specific expression of PBANKA_031140 was linked to a hepatocyte infection or subsequent intrahepatic EEF development. By using a reverse genetics approach, we demonstrate the role of PBANKA_031140 in late liver stage development and initiation of a timely blood stage infection. We designated this kinase as PbSTK2 owing to the previous description a P. falciparum STK (Kuang et al., 2017).
The Ser/Thr protein kinases interact with diverse substrates like enzymes, other kinases, transcription factors, receptors and regulatory proteins (Goldsmith et al., 2007). Fig. 1D depicts the STRING interaction network of PbSTK and Fig. 1E details its possible interacting partners based on co-expression of genes. These included several putative candidates like calmodulin (PBANKA_ 101060), a protein kinase (PBANKA_141450), mitogen-activated protein kinase 1 (PBANKA_101330), mitogen-activated protein kinase 2 (PBNAKA_093370), a protein phosphatase 2b regulatory subunit, (PBANKA_131540), a Cdc2-like protein kinase (PBANKA_123020), guanylate cyclase beta (PBANKA_113670), centrin-2, putative (PBANKA_131040), centrin-3, putative (PBANKA_051180) and centrin-4, putative (PBANKA_094140). The STRING prediction of PbSTK2 interaction with calmodulin is consistent with its inclusion under the CaMK group (Tewari et al., 2010). Calmodulin binding motifs type 1-10 and 1-16 are found in CaM-dependent PKs and CaM-dependent KKs, respectively. From sequence analysis and local similarity search, both motif types were found by matching the calmodulin binding peptide sequence obtained from PDB structures of calmodulin complexes with the PbSTK2 amino acid sequence. Motif type 1-10, similar to calmodulin binding peptide of Human CaMKII (PDB ID: 3GP2), occurred in PbSTK2 at 1505-1511 (SFKKRRK) and motif type 1-16, similar to calmodulin binding peptide of Rat CDPK (PDB ID: ICKK), occurred in PbSTK2 at 1200-1217 (FMSLVPGSLHIYEIKKIMKNIF) (Fig. 1C). The prediction of PbSTK2 1200-1217 as calmodulin-binding motifs is more reliable as the site contains two phenylalanine (F) residues separated by 20 residues. The secondary structure prediction of this peptide reveals it to be mostly helical, a requirement for calmodulin binding. The prediction of PbSTK2 1505-1511 (1-10) is not very reliable as it does not contain the conserved F residues at proper positions (Osawa et al., 1999).
Pbstk2 is maximally expressed in salivary gland sporozoite stage Stage-specific gene expression of Pbstk2 was analysed by standard quantitative real-time PCR (qRT-PCR) using cDNA samples generated from mixed blood stages, midgut sporozoites, salivary gland sporozoites and liver stages at 17 h, 25 h, 38 h, 48 h and 65 h. We noted maximal expression in salivary gland sporozoite stage followed by midgut sporozoite stage and at 65 h liver stages ( Fig. 2A).

Pbstk2 locus is not refractory to genetic manipulation
The strategy for the generation of Pbstk2 KO is shown in Fig. 2B. The 5′ fragment and 3′ fragment of Pbstk2 were amplified using primer sets FP1-RP1 and FP2-RP2 that yielded, respectively, products of 725 bp and 619 bp. The stable integration of the pBC-GFP-hDHFR carrying the 5′ and 3′ region of Pbstk2 at the desired locus was confirmed by diagnostic PCR. The primer sets FP3-RP3 and FP4-RP4 amplified products of 1102 bp and 924 bp, respectively, from both Pbstk2 KO clones Cl1 and Cl2, but not from WT GFP indicating the correct genomic integration of the Pbstk2 KO construct (Fig. 2C). Additionally, we also sequenced the Pbstk2 KO locus and confirmed the presence of GFP-hDHFR cassette (Fig. S1). Further, a primer set FP5-RP5 designed within Pbstk2 ORF amplified a product of 710 bp from WT GFP line but not from Cl1 and Cl2 (Fig. 2D).

Pbstk2 KO parasites develop normally during asexual blood stages and in mosquito stages
With the cloned lines of Pbstk2 KO, we analysed the asexual blood stage propagation as compared to WT GFP. Two groups of Swiss mice received intraperitoneal injection of 200 µl infected blood (having 0.2% parasitemia) from both cloned lines and WT GFP parasites. Parasitemia was monitored daily, by making Giemsa stained blood smears. We observed no difference in the blood stage propagation of Pbstk2 KO parasites as compared to WT GFP thus confirming the dispensable role of PbSTK2 in asexual blood stage propagation (Fig. 3A).
We next investigated the role of Pbstk2 KO in sexual stage development. In vitro cultures were set up from gametocyte positive blood of Pbstk2 KO Cl1, Cl2 and WT GFP. The cultures were harvested at 2 h, 8 h and 20 h and stained with anti-p28 antibody, a marker for zygote, retort and ookinete stage. We observed normal development of both Pbstk2 KO clones (data is shown for Pbstk2 KO Cl1) during sexual stages that were comparable to WT GFP ( in both Pbstk2 KO clonal lines were comparable to WT GFP parasites. We conclude that PbSTK2 has no role in Plasmodium development that occurs within the mosquito vector. Pbstk2 KO sporozoites exhibit normal gliding motility, cell traversal activity, and infectivity Next, we studied activities of Pbstk2 KO sporozoite required for hepatocyte infection. Sporozoites rely on actin-based gliding motility that facilitates host cell invasion (Kappe et al., 2004) and crossing of cellular barriers in vivo. The gliding pattern of Pbstk2 KO sporozoite was similar to WT GFP (Fig. 4A, data is shown for Pbstk2 KO Cl1). We next tested the cell traversal activity of sporozoite. The Pbstk2 KO were able to traverse through monolayers of HepG2 cells as visualised by deposition of rhodamine-labelled dextran (data is shown for Pbstk KO Cl1) as noted for WT GFP. Lack of labelling in HepG2 cultures in the absence of sporozoites or following addition cytochalasin D treated sporozoites (inhibitor of actin polymerisation) confirmed the specificity of labelling (Fig. 4B). Quantification of Pbstk2 KO sporozoite infectivity in HepG2 cells by inside out assay revealed no apparent defect in both Pbstk2 KO clones in colonising hepatocytes (Fig. 4C).
Pbstk2 KO manifest a delay in the pre-patent period To analyse the ability of Pbstk2 KO sporozoites to induce infection in a mammalian host, we intravenously infected C57BL/6 mouse with two different doses viz., 5×10 3 or 1×10 3 of WT GFP and Pbstk2 KO sporozoites and analysed the pre-patent period. At 5×10 3 dose, the Pbstk2 KO sporozoites became patent on day 5-7 as compared to day 3 for WT GFP. A dose of 1×10 3 Pbstk2 KO sporozoites was unable to initiate blood stage infection as judged by  PBANKA_031140) locus by double cross-over homologous recombination. A 725 bp and 619 bp of 5′ and 3′ region of Pbstk2 was amplified using primers FP1-RP1 and FP2-RP2, respectively. Both fragments were cloned into a pBC-GFP-hDHFR vector using XhoI-ClaI for the 5′ region and NotI-AscI for the 3′ region. (C) Diagnostic PCR showing the correct integration of targeting construct at Pbstk2 locus in Cl1 and Cl2 KO lines. The correct integration of the 5′ region was confirmed by a product of 1102 bp amplified using primers FP3 and RP3. The correct integration of the 3′ region was confirmed by a product of 924 bp amplified using primers FP4 and RP4. No product was amplified from genomic DNA of WT parasites using the aforementioned primer sets. (D) Amplification of a 710 bp fragment within Pbstk2 ORF using primer sets FP5 and RP5 from WT parasite genomic DNA but not from Cl1 and Cl2 confirmed successful establishment of Cl1 and Cl2 lines.
the complete absence of pre-patent period till day 30 (Table 2). We conclude that low doses of Pbstk KO sporozoites fail to cause infection while high doses cause occasional breakthrough infection.
Pbstk KO late liver stages show reduced hepatic schizogony and MSP1 expression A delay in the pre-patent period implicated an obvious growth defect in EEF development in the Pbstk2 KO. In order to identify precise time point of the defect, Pbstk2 KO sporozoites were added to monolayers of HepG2 cultures and their development was monitored at various time points by staining with anti-UIS4 antibody, a marker for parasitophorous vacuolar membrane (PVM). HepG2 cell and parasite nuclei were stained with Hoechst 33342. We observed no difference in the growth of EEFs at all time points tested in WT GFP (Fig. 5A) and Pbstk2 KO clone 1 (Fig. 5B).
To account for the delay in prepatency of Pbstk2 KO, we analysed 62 h liver stage culture by staining with monoclonal antibody 25.1 specific for merozoite surface protein 1 (MSP1). Hoechst 33342 was used for nuclear staining. We noted dramatic decrease in MSP1  Equal numbers (2×10 4 ) sporozoites of WT GFP and Pbstk2 KO clones Cl1 and Cl2 were added to monolayers of HepG2 cells. One hour later, the cultures were fixed and the extracellular sporozoites were stained with anti-CSP monoclonal antibody 3D11 and the immunoreactivity was revealed using anti-mouse secondary antibody conjugated to Alexa Fluor. The slides were observed under fluorescent microscope Ni-AR and in each field, the total number of sporozoites (GFP positive) and a number of extracellular sporozoites (red fluorescence) were counted and the invasion efficiency related to WT GFP was calculated. Scale bar: 20 μm. expression in Pbstk2 KO EEFs (Fig. 5C) as compared to WT GFP. The nuclei of hepatic merozoites were quantified using ImageJ software (https://imagej.nih.gov/ij). We noted that the hepatic merozoite numbers were reduced to nearly 40% as compared to WT GFP (Fig. 5D). We conclude that loss of PbSTK2 expression led to significant reduction in schizogony that affected both hepatic merozoite maturation and MSP1 expression thus preventing or delaying blood stage infection.

DISCUSSION
Identifying protein kinases that play a role during pre-erythrocytic stages of Plasmodium life cycle holds the potential to develop inhibitors that may delay or prevent blood stage infection. In this direction, we prioritised investigating the role of PbSTK2 owing to its detection in the proteomic analysis of salivary gland sporozoites (Lasonder et al., 2008). Consistent with the high levels of Pbstk2 transcripts and occurrence of protein in salivary gland sporozoites, our study demonstrates its requirement at 65 h liver stage for successful completion of hepatic schizogony and initiation of blood stage infection.
Bioinformatic analysis predicted one or more calmodulinbinding domains in PbSTK2 suggesting that it belongs to the family of CaMK and concurs with previous studies (Tewari et al., 2010). CaMKs are activated by an increase in intracellular concentration of Ca 2+ ions leading to the transfer of phosphates from ATP to defined serine or threonine residues of the substrate proteins. In view of our observation that PbSTK2 is required to complete hepatic schizogony, it is likely possible that this kinase may be activated by intracellular Ca 2+ that raises transiently in hepatic merozoites to prevent apoptosis (Sturm et al., 2006). This may explain partially as to why the mutants that lack PbSTK2 activity are impeded in completion of liver stage development. The fact that Pbstk2 KO showed an overall reduced hepatic schizogony may also point to its likely role in the regulation of the cell cycle, given that CaMKs influence the activity of cell division cycle (CDC) regulators (Sur and Agrawal, 2016). Further investigations are required to ascertain such hypotheses.
CaMKs have also been shown to play a role in other life cycle stages. In Plasmodium gallinaceum morphologic differential from zygote to ookinete stage was shown to be dependent on CaMK. Cellular extracts of zygote and ookinete phosphorylated autocamtide-2, a classic CaMK substrate that was blocked by both calmodulin antagonists W-7 and CaMK inhibitor KN-93 (Silva-Neto et al., 2002). While this study did not precisely identify the parasite-specific CaM kinase that was sensitive to the aforementioned inhibition, the fact that Pbstk2 mutants did not manifest any phenotype in the zygote and ookinete stage implied the presence of multiple CaM kinases that may have stage-specific functions.
Of the 66 P. berghei ePKs characterised previously (Tewari et al., 2010), 43 kinases were refractory to genetic manipulation and Pbstk2 was inclusive in this group. More recent studies of Bushell et al. (Bushell et al., 2017) also report this candidate to likely be essential for blood stages. However, we were able to successfully disrupt Pbstk2 locus. A likely explanation for this discrepancy, as reiterated previously (Tewari et al., 2010), may be the limitations of working with a rodent model, where gene targeting may not be effective every time and hence the inability to disrupt a gene does not necessarily point to its essential nature. However, piggyback screen performed in P. falciparum indicated that stk2 (PF3D7_0214600) is dispensable in the blood stages (Zhang et al., 2018). The other kinases classified as 'possibly essential' therefore merits a thorough functional investigation in order to unravel their unique stage-specific functions.
Our studies provide a possibility of targeting PbSTK2 to prevent blood stage infections for two important reasons. Firstly, because of its conserved nature across all Plasmodium species and its distinctness from the human counterpart, a PbSTK2 inhibitor may likely block the parasite's ability to cause break through infection. Secondly, owing to its requirement for late liver stage development, Pbstk2 locus can be an additional target while considering a multiple attenuated mutant solely based on late liver stage genes. Such mutants may have the dual advantage of conferring cross-stage immunity as well as reducing the risk of breakthrough infection. Taken together, our studies provide evidence for the first time of a role for PbSTK2 in late liver stage development that may have important therapeutic implications for preventing clinical malaria.

Ethics statement
All animal experiments performed in this study were approved by the Institutional Animal Ethics Committee at CSIR-Central Drug Research Institute, India (approval no: IAEC/2013/83) and Institutional Animal Ethics Committee at University of Hyderabad (approval no: UH/SLS/IAEC/ 2014-1/9b and UH/SLS/IAEC/2014-1/9c).

In silico analysis of P. berghei STK2
The amino sequence of P. berghei STK2 (PBANKA_031140) was downloaded from PlasmoDB (www.plasmodb.org), and phylogenetic data was obtained using BLAST and analysed for its closest orthologues in various groups such as bacteria, plants, fungi, animals and apicomplexans. The amino acid sequences of all the orthologues were aligned and constructed into a phylogenetic tree by ClustalX (Larkin et al., 2007). The tree was visualized and illustrated by the Interactive Tree of Life web service (Letunic and Bork, 2016). Multiple sequence alignment (MSA) of the kinase domain of PbSTK was performed with other species of Plasmodium to study its residue conservation. ClustalX was used to create the alignment, and it was visualized and illustrated by Unipro UGENE (Okonechnikov et al., 2012). STRING interactions database was used to draw an interaction network for PbSTK (Szklarczyk et al., 2015). The network does not necessarily point out direct protein-protein interactions but it indicates the relation of a protein with others by co-expression, co-evolution, direct interaction evidence, etc. This network helps to study the function of the protein in a broader genomic context.

Quantitative gene expression analysis of Pbstk2 across all the life cycle stages of P. berghei
The expression of the Pbstk2 was measured by quantitative real-time using absolute quantification method. RNA was isolated from mixed blood stages, midgut sporozoites, salivary gland sporozoites and in vitro liver stages, at time points 17 h, 25 h, 38 h, 48 h and 65 h post-infection using Trizol (Invitrogen), and purified using RNA isolation kit (Life Technologies) according to the manufacturer's instructions. Nearly 2 µg of RNA from aforementioned stages were used for reverse transcription in a reaction mixture containing 1× PCR buffer, 2.5 mM dNTPs, 5 mM MgCl2, 1.5 units RNase inhibitor, 2.5 mM random hexamers and 1.5 units reverse transcriptase (Applied Biosystems). Gene-specific standards were generated for Pbstk2 by amplifying 120 bp product from P. berghei genomic DNA using forward primer-Pbstk2 TA FP and reverse primer Pbstk2 TA RP and ligated into pTZ57R/T vector. Pb18S rRNA standard was used as an internal control (Kumar et al., 2004). The Pbstk2 plasmid standards were generated in a log fashion ranging from 10 8 to 10 2 copies per μl for both Pbstk2 and Pb18S rRNA. The cDNA samples were run alongside with both standards. Quantitative real-time PCR (qRTPCR) was carried out in a 10 μl reaction containing SYBR Green PCR Master Mix (Bio-Rad) and 0.25 μm gene-specific primers. qRTPCR was performed using the EFFENDORF REALPLEX 2 qPCR machine. A ratio of transcript numbers of Pbstk2 and Pb18S rRNA was obtained to normalize the gene expression data (Vaughan et al., 2009).

Generation of Pbstk2 (PBANKA_031140) KO construct
Double homologous recombination strategy was used to delete Pbstk2 locus. To achieve this, Pbstk2 KO targeting construct was generated using pBC-GFP-hDHFR vector. Plasmodium berghei genomic DNA was used as a template to amplify 725 bp of 5′ upstream region of Pbstk2 using Pbstk2 5′ forward primer-FP1 and Pbstk2 5′ reverse primer-RP1. A 619 bp of 3′ downstream region of Pbstk2 was also amplified using Pbstk2 3′ forward primer-FP2 and Pbstk2 3′ reverse primer-RP2. PCR was set up with 1× PCR buffer, 0.25 µm FP, 0.25 µm RP, 1 mM dNTPs each (Thermo Fisher Scientific, Cat No. R72501), 2.5 mM MgCl 2 , 30-50 ng genomic DNA and 2.5 units of Taq DNA polymerase (Thermo Fisher Scientific, Cat No. 11615010). Thermal cycling was performed in Eppendorf Mastercycler PCR machine, and the cycling conditions were as follows: initial denaturation at 94°C for 2 min and 94°C for 30 s, annealing at 54°C 30 s followed by synthesis at 72°C for 1 min. The cycles were repeated 35 times except for initial denaturation and final extension was done at 72°C for 10 min. The PCR product corresponding to 5′ and 3′ regions were confirmed by sequencing. The 5′ and 3′ PCR amplified products were cloned into the pBC-GFP-hDHFR vector at XhoI/ClaI and NotI/AscI sites, respectively. The knockout plasmid was digested with XhoI/AscI and gel purified and its concentration was determined using NanoDrop 2000.
Plasmodium berghei transfection and confirmation of stable integration of targeting construct at Pbstk2 locus Electroporation of the targeting construct was done essentially as described previously (Janse et al., 2006). In brief, blood was collected from mice having 4-5% parasitemia and an overnight culture was set up in RMPI medium containing 20% FBS (Gibco, South American origin) and 0.35 mg/ ml gentamycin (Gibco). The next day the cultures were subjected to density gradient centrifugation using 70% nycodenz (Sigma-Aldrich). The purified schizonts were used for electroporation of the Pbstk2 KO construct using Amaxa nucleofector device, program U033 (Janse et al., 2006). Two independent transfections were performed, and the schizonts were immediately injected intravenously into Swiss mice. The following day, blood smears were made from mouse harbouring transfected parasites and, after confirming parasitemia, the mice were subjected to an antimalarial pyrimethamine that was administered orally through drinking water. After 7 days of stringent drug selection, when the parasitemia reached around 5%, genomic DNA was isolated using a genomic DNA isolation kit (Genetix), following the manufacturer's instructions. To confirm the site-specific integration, diagnostic primers were designed within the targeting construct and beyond the site of integration at both the 5′ and 3′ ends of recombination using primer sets FP3, RP3 and FP4, RP4. Limiting dilution was done to isolate single clonal population and to eliminate non transfectants. Single clones were further confirmed for the absence of wild-type contamination by diagnostic PCR using a forward primer-FP5 and reverse primer-RP5 within the Pbstk2 ORF. The information of all primers used in this study is indicated in Table 1.

Gliding motility
Gliding motility assays were performed as described previously (Stewart and Vanderberg, 1988). Briefly, Lab-Tek wells were coated overnight with 10 µg/ml of mAb 3D11 diluted in PBS and kept at 25°C. WT GFP or Pbstk2 KO salivary gland sporozoites were dissected and added to Lab-Tek wells and incubated for 1 h at 37°C. The sample was subsequently fixed in 4% PFA and stained with biotinylated mAb 3D11 followed by streptavidin-FITC (Invitrogen) to visualise the CS protein-containing trails.

Sporozoite cell traversal assay
To examine the cell traversal activity, WT GFP or Pbstk2 KO sporozoites were added in Lab-Tek wells seeded with HepG2 cells. The HepG2 cells were procured from NCCS, Pune, India and were confirmed negative for mycoplasma contamination. The sporozoite invasion was performed in the presence of 1 mg/ml rhodamine-labelled dextran (10,000 MW, neutral; Invitrogen). Cells traversed by sporozoites captures dye-labeled dextran. Dextran-positive cells were enumerated manually under the fluorescent microscope (Mota et al., 2001).

Sporozoite invasion assay
HepG2 cells were plated in eight-chambered Lab-Tek wells and were maintained in DMEM (Gibco, supplemented with L-glutamine and glucose) medium containing 10% FBS (Gibco) and allowed to grow until they reached subconfluency. Nearly 2×10 4 sporozoites of WT GFP or Pbstk2 KO were added in each well and following 1 h incubation at 37°C, the cells were washed and fixed with 4% paraformaldehyde. Extracellular sporozoites were in stained without permeabilising the cultures using 3D11 monoclonal antibody specific for the repeat region of CSP (Yoshida et al., 1980) and the immunoreactivity was revealed using anti-mouse secondary antibody conjugated to Alexa Fluor 594. To enumerate the sporozoite infectivity, the number of GFP expressing sporozoites (total) were counted per field and the number of sporozoites positive for CSP expression (extracellular sporozoites) was enumerated per field (Renia et al., 1988). The percentage sporozoite infection was calculated by using the formula: the number of GFP expressing sporozoites per field -number of sporozoites stained red per field/number of GFP expressing sporozoites per field × 100.
Determination of pre-patent periods C57BL/6 mice were infected with WT GFP or Pbstk2 KO sporozoites through intravenous injection of sporozoites. Parasitaemia of infected mice was monitored daily by Giemsa-stained blood smear.
In vitro EEF culture and Immunofluorescence assays Human liver hepatocellular carcinoma (HepG2) cells (5.0×10 4 per well) were seeded in a 48-well plate, 24 h before prior to addition of sporozoites. The cultures were maintained in DMEM containing 10% FCS as described in the earlier section, in addition to 1× antibiotic and antimycotic. Nearly 5×10 3 sporozoites of either WT GFP or Pbstk2 KO were added per well and centrifuged at 320 g for 4 min. To quantify EEF development, sporozoites were added to HepG2 cells and maintained for different time points and fixed with 4% paraformaldehyde. The EEFs in different stages of development were stained with 1:1000 dilution of anti-UIS4 antibody (Mueller et al., 2005a) specific for the parasitophorous vacuolar membrane (PVM) and revealed with 1:300 dilution of anti-rabbit Alexa Fluor 594 antibody (Thermo Fisher Scientific, Cat No. A-11012). To stain hepatic merozoites, a monoclonal antibody 25.1 specific for merozoite surface protein 1 (MSP1) was used at 1:5000 dilution (Holder and Freeman, 1981), and the immunoreactivity was revealed using 1:300 dilution of anti-mouse Alexa Flour 594 (Thermo Fisher Scientific, Cat No. A21203). The host and parasite nuclei were stained with Hoechst 33342 (Thermo Fisher Scientific, Cat No. H1399).