Plasmodium berghei kinesin-5 associates with the spindle apparatus during cell division but is dispensable for parasite proliferation

Kinesin-5 motors play essential roles in spindle apparatus assembly during cell division, by generating forces to establish and maintain the spindle bipolarity essential for proper chromosome segregation. Kinesin-5 is largely conserved structurally and functionally in model eukaryotes, but its role is unknown in the Plasmodium parasite, an evolutionarily divergent organism with several atypical features of both mitotic and meiotic cell division. We have investigated the function and subcellular location of kinesin-5 during cell division throughout the Plasmodium berghei life cycle. Deletion of kinesin-5 had little visible effect at any proliferative stage. The only significant effect was a decrease in the number of motile sporozoites in mosquito salivary glands, which were able to infect a new vertebrate host. Live-cell imaging showed kinesin-5-GFP located on the spindle and at spindle poles during both atypical mitosis and meiosis. Fixed-cell immunofluorescence assays revealed kinesin-5 co-localized with α-tubulin and centrin-2 and exhibited a partial overlap with kinetochore marker NDC80 during early blood stage schizogony. Dual-colour live-cell imaging during male gametogony showed that kinesin-5 is closely associated with NDC80, but not with kinesin-8B, a marker of the basal body and axonemes of the forming flagella. Treatment of gametocytes with microtubule-specific inhibitors confirmed kinesin-5 association with nuclear spindles and not cytoplasmic axonemal microtubules. Altogether, our results demonstrate that kinesin-5 is associated with the spindle apparatus and expressed in proliferating parasite stages, but it is not essential for parasite survival. Importance Kinesin-5 is a motor protein with important roles during cell proliferation and therefore considered a strong target for therapeutics development against many diseases. The role of kinesin-5 in Plasmodium is unknown, but in this organism there are atypical aspects of cell division that differ from those of model eukaryotes. For example, classical regulators such as polo-like kinase, classical cyclins and cyclin dependent kinases are absent from Plasmodium. Here, we show that in Plasmodium, unlike in most model eukaryotes, kinesin-5 is dispensable, suggesting that the parasite uses somewhat different machinery for cell proliferation. The study is important because it shows that kinesin-5 is not essential universally for cell division and suggests that there may be other non-canonical mechanisms and regulators. The general concept that kinesin-5 is essential for spindle pole formation and chromosome segregation during cell division is not true for Plasmodium, and this is of interest to a broad readership. Introduction Kinesin-5 proteins are a family of molecular motors that is structurally and functionally conserved throughout eukaryotes [1-3]. They are involved in spindle pole separation and are considered essential for mitosis in the vast majority of eukaryotes [4, 5], except Caenorhabditis elegans [6], Dictyostelium discoideum [7] and Candida albicans [8]. Kinesin5 contains an N-terminal kinesin motor domain, a central stalk and a C-terminal tail domain [1], and forms a bipolar homotetramer that cross-bridges and slides on parallel and antiparallel microtubules (MTs) [9]. The motor domain binds microtubules (MTs) and hydrolyses ATP, which are conserved functions, while the tail region can also bind MTs, regulates motor activity [10] and helps localization during mitosis [10, 11]. The stalk contains a coiled-coil region and neck linker that promote oligomerization and direction of movement, respectively [12]. The kinesin-5 motor domain is conserved across eukaryotes, but the tail region is highly variable except for a short, conserved region called the BimC box, which contains consensus sites for phosphorylation by cyclin dependent kinase 1 (Cdk1) in most eukaryotes [6, 13, 14]. Kinesin-5 is located at spindle MTs and spindle poles during cell division and is distributed diffusely in the cytoplasm during interphase in most eukaryotic cells [2, 5]. The essential roles of kinesin-5 in spindle assembly and spindle pole separation can be blocked by specific inhibitors or antibodies, resulting in collapse of bipolar spindles into monopoles [3, 5, 13, 15]. Kinesin-5 is required for maintenance of a bipolar spindle in fungi, Xenopus and Drosophila [13, 15, 16]. In budding yeast the kinesin-5 proteins, Cin8 and Kip1, are also present at the kinetochores and help in chromosome alignment during metaphase [17]. Malaria is a deadly vector-borne infectious disease, caused by a unicellular protozoan parasite of the genus Plasmodium, which infects many vertebrate hosts including humans, and is transmitted by female Anopheles mosquitoes [18]. During the complex life cycle two unique phases of atypical closed mitotic division occur. The first type of mitosis occurs during asexual proliferation, with multiple asynchronous nuclear divisions producing up to 32 nuclei (during schizogony in the blood of the vertebrate host) or more than 1000 nuclei (during schizogony in the liver of the vertebrate host and sporogony in the mosquito gut), with cytokinesis occurring only after nuclear division is complete, to produce haploid progeny cells [19-22]. The second type of mitosis is during male gametogony (part of the sexual stage in the mosquito gut) where there are three rapid rounds of DNA replication from 1N to 8N within 10 to 15 minutes, followed by karyokinesis and cytokinesis to produce 8 flagellate haploid male gametes [22, 23]. The first phase of meiotic division occurs following fertilization as the zygote differentiates into a motile ookinete in the mosquito gut. The 2N genome is duplicated and recombination occurs [21]; then the final reductive division likely occurs in the oocyst that is formed following ookinete penetration of the mosquito gut wall, leading to the formation of haploid sporozoites. Thus, Plasmodium has unusual ways to divide and survive under different physiological conditions in different hosts. As cell division in Plasmodium is atypical, so the molecules regulating cell division are also divergent from those of higher eukaryotes [24-27]. The large family of kinesins includes molecular motors that are essential for several processes during cell division [28, 29], and in Plasmodium berghei there are nine kinesin genes, including two kinesin-8 genes that are important in cell division and male gamete formation [23, 30]. There is a single Plasmodium kinesin-5, and since this protein plays an important role during cell proliferation in many eukaryotes, and is considered as a strong target for the development of therapeutics against many diseases including malaria [31], we decided to study its role during cell division in the malaria parasite. In the present study we examined the spatiotemporal dynamics and functional role of kinesin-5 during the proliferative stages of P. berghei. Unlike kinesin-5 in many other eukaryotes, P. berghei (Pb) kinesin-5 is dispensable for parasite growth and proliferation, although deletion of the gene results in a remarkable decrease in the number of mosquito salivary gland sporozoites. Live cell imaging shows that kinesin-5 is located both on the spindle and at spindle poles during mitotic and meiotic divisions in the parasite life cycle, colocalizing with centrosome marker centrin and kinetochore marker NDC80. Treatment of gametocytes with microtubule inhibitors confirmed the nuclear localization and association with spindle microtubules of kinesin-5 during male gametogony. Altogether, our results show that kinesin-5 is associated with spindle apparatus, which can be inhibited by microtubule inhibitors, but the protein is not essential for parasite survival.


Introduction
Kinesin-5 proteins are a family of molecular motors that is structurally and functionally conserved throughout eukaryotes [1][2][3]. They are involved in spindle pole separation and are considered essential for mitosis in the vast majority of eukaryotes [4,5], except Caenorhabditis elegans [6], Dictyostelium discoideum [7] and Candida albicans [8].  contains an N-terminal kinesin motor domain, a central stalk and a C-terminal tail domain [1], and forms a bipolar homotetramer that cross-bridges and slides on parallel and antiparallel microtubules (MTs) [9]. The motor domain binds microtubules (MTs) and hydrolyses ATP, which are conserved functions, while the tail region can also bind MTs, regulates motor activity [10] and helps localization during mitosis [10,11]. The stalk contains a coiled-coil region and neck linker that promote oligomerization and direction of movement, respectively [12]. The kinesin-5 motor domain is conserved across eukaryotes, but the tail region is highly variable except for a short, conserved region called the BimC box, which contains consensus sites for phosphorylation by cyclin dependent kinase 1 (Cdk1) in most eukaryotes [6,13,14].
Kinesin-5 is located at spindle MTs and spindle poles during cell division and is distributed diffusely in the cytoplasm during interphase in most eukaryotic cells [2,5]. The essential roles of kinesin-5 in spindle assembly and spindle pole separation can be blocked by specific inhibitors or antibodies, resulting in collapse of bipolar spindles into monopoles [3,5,13,15].
Kinesin-5 is required for maintenance of a bipolar spindle in fungi, Xenopus and Drosophila [13,15,16]. In budding yeast the kinesin-5 proteins, Cin8 and Kip1, are also present at the kinetochores and help in chromosome alignment during metaphase [17].
Malaria is a deadly vector-borne infectious disease, caused by a unicellular protozoan parasite of the genus Plasmodium, which infects many vertebrate hosts including humans, and is transmitted by female Anopheles mosquitoes [18]. During the complex life cycle two unique phases of atypical closed mitotic division occur. The first type of mitosis occurs during asexual proliferation, with multiple asynchronous nuclear divisions producing up to 32 nuclei (during schizogony in the blood of the vertebrate host) or more than 1000 nuclei (during schizogony in the liver of the vertebrate host and sporogony in the mosquito gut), with cytokinesis occurring only after nuclear division is complete, to produce haploid progeny cells [19][20][21][22]. The second type of mitosis is during male gametogony (part of the sexual stage in the mosquito gut) where there are three rapid rounds of DNA replication from 1N to 8N within 10 to 15 minutes, followed by karyokinesis and cytokinesis to produce 8 flagellate haploid male gametes [22,23]. The first phase of meiotic division occurs following fertilization as the zygote differentiates into a motile ookinete in the mosquito gut. The 2N genome is duplicated and recombination occurs [21]; then the final reductive division likely occurs in the oocyst that is formed following ookinete penetration of the mosquito gut wall, leading to the formation of haploid sporozoites. Thus, Plasmodium has unusual ways to divide and survive under different physiological conditions in different hosts. As cell division in Plasmodium is atypical, so the molecules regulating cell division are also divergent from those of higher eukaryotes [24][25][26][27]. The large family of kinesins includes molecular motors that are essential for several processes during cell division [28,29], and in Plasmodium berghei there are nine kinesin genes, including two kinesin-8 genes that are important in cell division and male gamete formation [23,30]. There is a single Plasmodium kinesin-5, and since this protein plays an important role during cell proliferation in many eukaryotes, and is considered as a strong target for the development of therapeutics against many diseases including malaria [31], we decided to study its role during cell division in the malaria parasite.
In the present study we examined the spatiotemporal dynamics and functional role of kinesin-5 during the proliferative stages of P. berghei. Unlike kinesin-5 in many other eukaryotes, P. berghei (Pb) kinesin-5 is dispensable for parasite growth and proliferation, although deletion of the gene results in a remarkable decrease in the number of mosquito salivary gland sporozoites. Live cell imaging shows that kinesin-5 is located both on the spindle and at spindle poles during mitotic and meiotic divisions in the parasite life cycle, colocalizing with centrosome marker centrin and kinetochore marker NDC80. Treatment of gametocytes with microtubule inhibitors confirmed the nuclear localization and association with spindle microtubules of kinesin-5 during male gametogony. Altogether, our results show that kinesin-5 is associated with spindle apparatus, which can be inhibited by microtubule inhibitors, but the protein is not essential for parasite survival.

Pbkinesin-5 is expressed at multiple proliferative stages in the parasite life cycle
To quantify the expression of kinesin-5 at different stages of the parasite life cycle, we isolated RNA and performed qRT-PCR. Kinesin-5 is expressed constitutively throughout the blood and mosquito stages of parasite development, with the highest level in gametocytes, followed by schizonts and ookinetes (Fig. 1A).
Pbkinesin-5 is dispensable for cell division during the different proliferative stages of the Plasmodium life cycle.
The Plasmodium life cycle has two unusual mitotic processes that occur during schizogony/sporogony and male gametogony, and a single meiotic stage during zygote to ookinete transformation [32]. To examine any functional role of kinesin-5 during these processes, we deleted the gene from the P. berghei genome using a double crossover homologous recombination strategy in a parasite line constitutively expressing green fluorescent protein (GFP) at all stages of the parasite life cycle (Fig. S1A) [33]. Diagnostic PCR, to show successful integration of the targeting construct at the kinesin-5 locus (Fig.   S1B), and quantitative real time PCR (qRT-PCR), to show lack of kinesin-5 expression in gametocytes, confirmed the complete deletion of the kinesin-5 gene (Fig. 1B). Successful creation of this transgenic parasite indicated that the gene is not essential for mitosis during asexual blood stage schizogony. Further phenotypic analysis of the Δ kinesin-5 parasite was carried out at other stages of the life cycle, comparing the parental parasite (WT-GFP) with two independent gene-knockout parasite clones (clones 3 and 5) generated by two independent transfections. Both gene-knockout clones had the same phenotype and data presented here are the combined results from both clones. Since Δ kinesin-5 parasites underwent asexual blood stage development, exhibiting no change in morphology, number of progeny merozoites or parasitemia, an essential role for kinesin-5 is unlikely during these stages, which cause the disease in the mammalian host.
The transgenic parasites also produced gametocytes in mice, and therefore next we analysed male and female gametocyte differentiation following activation in the exflagellation/ookinete medium that mimics the mosquito gut environment. Male and female gametes emerge from the infected erythrocyte, and in the case of male gamete development this is preceded by three rapid rounds of genome duplication, resulting in eight flagellate male gametes [22,23]. There was no defect in male gamete exflagellation for either of the Δ kinesin-5 parasite clones, with the same exflagellation frequency as WT-GFP parasites ( Fig. 1C). Fertilization, zygote formation and ookinete differentiation, when meiosis occurs [21], were not significantly different in Δ kinesin-5 parasites from these processes in parental parasites (Fig. 1D).
To investigate the role of kinesin-5 in oocyst development and sporogony, Anopheles stephensi mosquitoes were fed on mice infected with  (Fig. 1E), and the size of the oocysts was similar for both parasites (Fig. 1F). However, we observed a 40 to 50% decrease in the number of sporozoites in each oocyst at days 14 and 21 post-infection in Δ kinesin-5 parasites compared to WT-GFP parasites (Fig. 1G). In comparison to the WT-GFP parasite, a significant decrease in the number of Δ kinesin-5 sporozoites in salivary glands was also observed (Fig. 1H), but the shape, size, and motility of these sporozoites were indistinguishable from WT-GFP parasites (Fig. 1I, J). Furthermore, western blot analysis of circumsporozoite protein (CSP) showed that proteolytic processing of CSP in Δ kinesin-5 sporozoites, an indicator of normal sporozoite maturation, was also not affected (Fig. 1K).
The infected mosquitoes were used for bite back experiments to ascertain the infectivity of Δ kinesin-5 sporozoites in mice; a blood stage infection was observed after 4 days with both Δ kinesin-5 and WT-GFP sporozoites (Fig. 1L).
The loss of kinesin-5 may have been compensated for by the over-expression of another member of the kinesin family, therefore we analysed the transcript level of kinesin-8B, kinesin-8X and kinesin-13 in gametocytes of Δ kinesin-5 parasites. These kinesins are highly expressed in gametocytes and have important role during male gametogony [23,30]. We found that the kinesin-13 transcript level was significantly upregulated (Fig. 1M), suggesting that higher levels of kinesin-13 may -in some way -compensate for the kinesin-5 deletion.

Pbkinesin-5 is located at the spindle apparatus during mitotic stages of asexual blood stage schizogony
To examine expression at the protein level and study the real-time dynamic location of kinesin-5 during cell division, we generated a kinesin-5-GFP transgenic P. berghei line expressing kinesin-5 with a C-terminal GFP tag, by inserting an in-frame gfp coding sequence at the 3' end of the endogenous kinesin-5 locus using single homologous recombination (Fig. S2A). Successful insertion was confirmed by diagnostic PCR (Fig.   S2B). Western blot analysis of a schizont protein extract using an anti-GFP antibody revealed kinesin-5-GFP protein at the expected size of 198 kDa compared to the 29 kDa GFP (Fig. S2C). This kinesin-5-GFP transgenic line was used to examine the spatiotemporal profile of kinesin-5-GFP protein expression and location by live cell imaging during the whole parasite life cycle, initially during asexual blood stage development in erythrocytes.
After haploid merozoite invasion of an erythrocyte, the initial ring and trophozoite stages are followed by schizogony, which results in the formation of further merozoites that invade fresh erythrocytes. During schizogony there are several independent asynchronous rounds of closed mitosis to produce a multi-nucleate syncytium, followed by cytokinesis and egress of mature merozoites from the infected erythrocyte. Kinesin-5 expression was not detected in the ring and early trophozoite stages that are considered as interphase or G 0 in the cell cycle [34]. A very low and diffuse expression of kinesin-5 throughout the cytoplasm was observed in older trophozoites, a stage similar to G1 phase in higher eukaryotes [34,35] when preparation for DNA replication begins. Late trophozoites mark the transition into early S phase when DNA synthesis starts [35] and schizogony is marked by the presence of two or more nuclei. Kinesin-5 was observed as strong foci adjacent to the Hoechst-stained DNA in early schizonts when nuclear division had commenced, and representing the first M phase of the cell cycle [34,35] (Fig. 2A). Each kinesin-5-GFP focus elongated and split into two foci that migrated away from each other, remaining adjacent to the nuclear DNA that then separated into two nuclear masses ( Fig. 2A). Alternating repeated S/M phases followed the division of individual nuclei, accompanied by repeated elongation and duplication into multiple points of these kinesin-5-GFP foci, showing the asynchronous pattern of nuclear division. Following completion of schizogony, kinesin-5 expression was almost undetectable in merozoites ( Fig. 2A).
To compare the location of kinesin-5 with that of other mitotic protein markers, including  2B). Similarly, kinesin-5 co-localised with centrin-2, confirming its location close to or at the putative MTOC (Fig. 2C). Using anti-GFP with anti-NDC80 antibodies revealed that kinesin-5-GFP is located in close proximity to, but not overlapping, the kinetochores (Fig. 2D); this location was confirmed by live cell imaging with a dual colour parasite line expressing kinesin-5-GFP and NDC80-mCherry (Fig. 2E, F).

Spatiotemporal dynamics of Pbkinesin-5 reveal its location on the spindle apparatus during male gametogony
In order to study the dynamics of kinesin-5 during the rapid genome replication in male gametogony we examined its expression by live-cell imaging during the 15-minute period following activation of male gametocytes. Both male and female gametocytes express kinesin-5, with a diffuse nuclear location. At the start of male gametogony, kinesin-5 accumulated at one end of the nucleus at a single focal point 1-minute post-activation (mpa) (Fig. 3A). By 2 mpa, this focal point extended to form a bridge across one side of the nucleus, followed by the separation of the two halves of the bridge to produce shorter linear rods that then contracted to two clear single foci by 3 mpa (Fig. 3A). This process repeated twice, resulting in 8 discrete kinesin-5-GFP foci. These discrete kinesin-5 foci then dispersed just before cytokinesis and exflagellation and the protein remained diffused in the nucleus of the remnant gametocyte (Fig. 3A). A schematic diagram for this process is shown in the upper panel of Fig. 3A.
To study the association of kinesin-5 with the mitotic spindle we used immunofluorescencebased co-localization assays with anti-GFP antibodies to stain kinesin-5, and anti-α-tubulin antibodies for MT staining. This analysis showed clear co-localisation of kinesin-5-GFP with the MT marker, both on the bridge-like structure and the foci, representing the spindle and spindle pole body, respectively (Fig. 3B). However, kinesin-5 was not present in mature male gametes following their egress (Fig. 3A, B).
To further investigate whether the location of kinesin-5 is cytoplasmic or nuclear, the kinesin-5-GFP parasite line was genetically crossed with the NDC80-mCherry kinetochore marker line [32] and kinesin-8B-mCherry cytoplasmic axoneme marker line [23], and the crosses were used for live cell imaging of both markers to establish their spatiotemporal relationship.
We found that both kinesin-5 and NDC80 were located next to the nuclear DNA, with colocalization on both spindle and spindle poles during different stages of male gametogony (Fig. 4A, B). In contrast, kinesin-5 did not co-localise with kinesin-8B which is located on the cytoplasmic basal bodies in early stages of male gametogony and later distributes across the axonemes (Fig. 4C, D).
To study further the association of kinesin-5 with nuclear or cytoplasmic MTs, we examined the effects of tubulin inhibitors specific to nuclear (taxol) and cytoplasmic (DDD01028076) MTs on kinesin-5 organisation during male gametogony [23]. Addition of taxol at 1 mpa blocked the dynamic redistribution of both kinesin-5 and NDC80 in more than 70% of male gametocytes, whereas DMSO-treated gametocytes showed normal mitotic progression and kinesin-5 distribution (Fig. 4D). This result showed that kinesin-5 distribution and localization is associated with spindle dynamics, similar to the behaviour of NDC80 [32], and can be blocked by taxol treatment, which binds tubulin and stabilises MTs by preventing depolymerisation (Fig. 4D). In contrast, treatment with DDD01028076 had no effect on either kinesin-5 or NDC80 location, consistent with the specificity of this inhibitor for cytoplasmic MT (axonemes) (Fig. 4D), as shown previously [23].

During meiosis in zygote to ookinete development, Pbkinesin-5 location follows spindle dynamics
We studied the location of kinesin-5 in the meiotic stage during zygote differentiation to ookinete over the 24-hour period after fertilisation.
Kinesin-5-GFP fluorescence was initially diffuse within the zygote nucleus, but after 1.5 to 2 hours post fertilization the GFP signal coalesced to a single focal point adjacent to the DNA (Fig. 5A). As ookinete development proceeds, with a small apical protrusion (stage I), the intensity of kinesin-5 increased and by stage II to III it was observed on spindles and more prominently on spindle poles (Fig. 5A). Later, in development stage IV, the polar localization of kinesin-5-GFP was lost and by stage V to VI, it again became diffuse in the nucleus with some less-prominent foci (Fig. 5A).

Pbkinesin-5-GFP exhibits multiple nuclear foci during oocyst development and in liver stage schizogony
Mitosis during oocyst development (sporogony) resembles that of schizogony within the mammalian host, and, as in exoerythrocytic schizogony (in hepatocytes), with many nuclei.
Each oocyst contains multiple lobes and produces thousands of sporozoites [32]. Kinesin-5-GFP fluorescence was observed as multiple foci representing a location at the putative MTOC/nuclear poles, from very early in development (day 7) to late stages (at day 14) of oocyst maturation (Fig. 6B), similar to the pattern for PbCEN-4 as described previously [36].
Many arc-shaped GFP signals were also observed that may represent the distribution of kinesin-5 on mitotic spindles (Fig. 5B). This mirrors what has been seen in electron microscopy studies: nuclear spindles radiate from the nuclear poles, with attached kinetochores within an intact nuclear membrane during oocyst development [30,32].
Interestingly, in mature oocysts (day 21), kinesin-5-GFP fluorescence was restricted to the residual body of the oocyst and was absent from mature sporozoites, suggesting that once nuclear division is completed kinesin-5-GFP is degraded (Fig. 5B).
Sporozoites produced in oocysts move to the mosquito salivary glands and, when transmitted by mosquito bite, infect the new host and migrate to the liver. As a model to study expression and location of kinesin-5 during mitosis in liver cells, we infected HeLa cells with sporozoites in vitro. The pattern of kinesin-5 distribution in these cells was similar to that of other asexual proliferative stages showing multiple foci of kinesin-5GFP next to DNA staining (Fig. 5C).

Discussion
Spindle apparatus assembly and chromosome segregation are key processes of nuclear division that require forces generated by MT-based motor proteins [28,37]. In most eukaryotes kinesin-5 is the major mitotic motor protein that crosslinks anti-parallel spindle MTs and drives bipolar spindle formation [5,38]. Any defect in kinesin-5 function results in the failure of spindle pole separation and prevents successful nuclear division [3,13,15]. In the present study we show by deletion of kinesin-5 that this protein is not essential for either mitotic or meiotic division during P. berghei parasite proliferation. This mirrors the situation in some other species such as C. elegans, D. discoidium and C. albicans, where kinesin-5 is not essential for survival [5][6][7][8].
The biology of malaria parasite development is very different from that of many model eukaryotes, from which it is evolutionarily distinct. This protozoan parasite has a complex life cycle and there are differences in the modes of cell division at different stages. Asexual proliferation is by asynchronous closed mitosis to produce a multinucleate syncytium that undergoes cytokinesis at the end of the cell cycle to produce haploid extracellular progeny.
This cell division lacks several classical regulators such as polo-like kinases, group 1 cyclins and many components of the anaphase promoting complex [24][25][26][27]39]. The classical cell division kinase, cyclin dependent kinase 1 (CDK1) is not essential for cell division in Plasmodium [24], but the parasite does possess other divergent and apicomplexan specific kinases that may be important in this process [24,40]. Spatiotemporal dynamic localization of the protein during blood stage schizogony revealed that kinesin-5 starts to coalesce adjacent to the nuclear DNA in late trophozoite stage. This is the time when the centriolar plaque or spindle pole body appears for the first time before the start of mitosis, and serves as a putative MTOC [35,36]. The location of kinesin-5 on this putative MTOC was confirmed by co-localization with centrin-2 and α -tubulin, which have been shown to track MTOC [20,35,36]. With the progression of nuclear division during schizogony, kinesin-5 showed its characteristic location on spindles, similar to the situation in other eukaryotes, where it helps in spindle assembly and chromosome segregation [5,35]. Kinesin-5 has the same pattern of localization as α -tubulin, showing its association with spindles and MTOCs in consecutive nuclear divisions [5,35]. A similar localization of kinesin-5 was observed during the other asexual mitotic stages, liver schizogony and sporogony in the mosquito gut. Kinesin-5 expression was not detected in mature and extracellular merozoites and sporozoites, indicating that once its role during mitosis is over it is degraded or discarded. The location of kinesin-5 in the residual body of mature oocysts following release of sporozoites, suggests that kinesin-5 is actively involved during mitosis in oocysts and then it is discarded at the end of endomitotic cell division. A similar fate was also observed for another molecular motor, myosin J (MyoJ) that also accumulates in the residual body during sporogony [42].
The residual bodies play important roles in Toxoplasma during organization of developing progeny inside the parasitophorus vacuole and promote their orderly and efficient externalization after maturation [43]. A defect in Toxoplasma MyoF molecular motor function results in enlarged residual bodies with accumulation of intact organelles [44], but Pbkinesin-5 deletion did not show any such phenotype. Although the number of Kinesin-5 was also found to be associated with kinetochore dynamics as shown by its colocalization with NDC80 and the synchronised movement of both proteins during male gametogony, as shown by the live-cell imaging of dual colour-tagged parasites expressing kinesin-5-GFP and NDC80-mCherry. This dynamic location of Plasmodium kinesin-5 and NDC80 is consistent with a study showing that in yeast, kinesin-5 is recruited to the kinetochore and plays an important role in its organisation [17]. The correct organisation of kinetochores during mitosis is very important as they are assembled at centromeres and facilitate the attachment of chromosomes to spindles before chromosome segregation [45].
In conclusion, this is the first study to explore the real-time dynamics and functional role of the kinesin-5 molecular motor in the mitotic and meiotic cell division cycles of the different stages of Plasmodium development.

Ethics statement
The animal work performed in this study has passed an ethical review process and was approved by the United Kingdom Home Office. Work was carried out in accordance with the

Generation of transgenic parasites
The gene-deletion targeting vector for Pbkinesin-5 (PBANKA_0807700) was constructed using the pBS-DHFR plasmid, which contains polylinker sites flanking a T. gondii dhfr/ts expression cassette conferring resistance to pyrimethamine, as described previously [24].
PCR primers N1061 and N1062 were used to generate a 995 bp fragment of kinesin-5 5′ upstream sequence from genomic DNA, which was inserted into ApaI and HindIII restriction sites upstream of the dhfr/ts cassette of pBS-DHFR. A 1008 bp fragment generated with primers N1063 and N1064 from the 3′ flanking region of kinesin-5 was then inserted downstream of the dhfr/ts cassette using EcoRI and XbaI restriction sites. The linear targeting sequence was released using ApaI/XbaI. A schematic representation of the endogenous Pbkinesin-5 locus, the construct and the recombined kinesin-8B locus can be found in Fig. S1.
To generate kinesin-5-GFP, a region of kinesin-5 gene downstream of the ATG start codon was amplified using primers T1921 and T1922 and ligated to p277 vector, and transfected as described previously [46]. The p277 vector contains the human dhfr cassette, conveying resistance to pyrimethamine. Pbkinesin-5 was tagged with GFP at the C-terminus by single crossover homologous recombination. A schematic representation of the endogenous gene locus, the constructs and the recombined gene locus can be found in Fig. S2. The oligonucleotides used to generate the mutant parasite lines can be found in Supplementary   Table S1. P. berghei ANKA line 2.34 (for GFP-tagging) or ANKA line 507cl1 expressing GFP (for gene deletion) parasites were transfected by electroporation [33].

Genotypic analysis of parasites
For the gene knockout parasites, diagnostic PCR was used with primer 1 (IntN106) and primer 2 (ol248) to confirm integration of the targeting construct, and primer 3 (N106 KO1) and primer 4 (N106 KO2) were used to confirm deletion of the kinesin-5 gene (Fig. S1). For the parasites expressing a C-terminal GFP-tagged kinesin-5 protein, diagnostic PCR was used with primer 1 (IntT192) and primer 2 (ol492) to confirm integration of the GFP targeting construct (Fig. S2).

Phenotypic analyses
To initiate infections, blood containing approximately 60,000 parasites of Δ kinesin-5 line was injected intraperitoneally (i.p) into mice. Asexual stages and gametocyte production were monitored on Giemsa-stained thin smears. Four to five days post infection, exflagellation and ookinete conversion were examined as described previously [46] with a Zeiss AxioImager Purification of gametocytes was achieved using a protocol as described previously [47] with some modifications. Briefly, parasites were injected into phenylhydrazine treated mice and

Generation of dual tagged parasite lines
The kinesin-5-GFP parasites were mixed with either NDC80-cherry or kinesin-8B-cherry parasites in equal numbers and injected into a mouse. Mosquitoes were fed on this mouse 4 to 5 days after infection when gametocyte parasitemia was high. These mosquitoes were checked for oocyst development and sporozoite formation at day 14 and day 21 after feeding. Infected mosquitoes were then allowed to feed on naïve mice and after 4 -5 days these mice were examined for blood stage parasitemia by microscopy with Giemsa-stained blood smears. In this way, some parasites expressed both kinesin-5-GFP and NDC80-cherry or kinesin-5-GFP and kinesin-8B-cherry in the resultant gametocytes, and these were purified, and fluorescence microscopy images were collected as described above.

Inhibitor studies
Gametocytes from the parasites expressing kinesin-5-GFP and NDC80-cherry were purified as above and treated with Taxol (Sigma) and an antimalarial molecule (DDD01028076) [23] at 1 mpa and then fixed with 4% paraformaldehyde (PFA, Sigma) at 8 min after activation.
Dimethyl sulfoxide (DMSO) was used as a control treatment. These fixed gametocytes were then examined on a Zeiss AxioImager M2 microscope.

Fixed Immunofluorescence Assay
The purified schizonts from PbKinesin-5-GFP parasites were fixed in 2% PFA and smeared on poly-L-lysine coated slides. Purified gametocytes were activated in ookinete medium then

Sporozoite motility assays
Sporozoites were isolated from salivary glands of mosquitoes infected with WT-GFP and Δ kinesin-5 parasites on day 21 post infection. Isolated sporozoites in RPMI 1640 containing 3% bovine serum albumin (Fisher Scientific) were pelleted (5 min, 5,000 rpm, 4°C) and used for motility assays as described previously [42]. Briefly, a drop (6 μ l) of sporozoites was transferred onto a microscope glass slide with a cover slip. Time-lapse videos of sporozoites (one frame every 1 s for 100 cycles) were taken using the differential interference contrast settings with a 63x objective lens on a Zeiss AxioImager M2 microscope and analysed with the AxioVision 4.8.2 software. The motility was also analysed by using Matrigel. A small volume (20 μ l) of sporozoites, isolated as above was mixed with Matrigel (Corning). The mixture (6 μ l) was transferred on a microscope slide with a cover slip and sealed with nail polish. After identifying a field containing sporozoites, time-lapse videos (one frame every 2 s for 100 cycles) were taken using the differential interference contrast settings with a 63x objective lens.

Liver stage parasite imaging
For P. berghei liver stage parasites, 100,000 HeLa cells were seeded in glass-bottomed imaging dishes. Salivary glands of female A. stephensi mosquitoes infected with kinesin-5-GFP parasites were isolated and disrupted using a pestle to release sporozoites, which were arginyl-t RNA synthetase (PBANKA_143420) genes were used as endogenous control reference genes. The primers used for qPCR can be found in Supplementary Table1.

Statistical analysis
All statistical analyses were performed using GraphPad Prism 5 (GraphPad Software). For qRT-PCR, an unpaired t-test was conducted to examine significant differences between wild-type and mutant strains.