A 3-hydroxyacyl-CoA dehydratase (DEH) involved in very long chain fatty acid synthesis is crucial for oocyst mitotic division but not cell size during parasite transmission in Plasmodium

School of Life Sciences, Queens Medical Centre, University of Nottingham, Nottingham, UK The Leicester Cancer Research Centre, College of Life Sciences, University of Leicester, Leicester LE2 7LX, UK Translational Bioinformatics Group, International Centre for Genetic Engineering and Biotechnology, Aruna Asaf Ali Marg, New Delhi-110067, India Department of Biological and Medical Sciences, Faculty of Health and Life Science, Oxford Brookes University, Oxford, UK Wellcome Trust Centre for Anti-Infectives Research, School of Life Sciences, University of Dundee, Dundee DD1 5EH, UK The Francis Crick Institute, London NW1 1AT, UK +contributed equally Corresponding author E-Mail: dsg6@le.ac.uk E-mail: rita.tewari@nottingham.ac.uk


Introduction
Malaria remains one of the world's deadliest infectious diseases. Caused by apicomplexan parasites belonging to genus Plasmodium, malaria is responsible for great socio-economic loss to affected countries. According to WHO reports, there were 212 million clinical cases of malaria infection and 429000 deaths in 2015 (Organisation, 2018), and growing resistance against existing drugs has further intensified this problem. Hence, there is a growing need to identify new biological pathways and proteins essential for parasitic growth and development in human hosts, which could act as suitable drug targets. Plasmodium parasites have a complex life cycle and require two hosts to complete the life cycle: vertebrates (during asexual stages) and invertebrates (during sexual stages) . The disease is transmitted to vertebrate hosts by infected female Anopheles mosquitoes, which inject sporozoites into the dermis of the vertebrate host during a blood meal. The parasite enters the circulation, and once it invades the liver, and subsequently erythrocytes, undergoes several rounds of atypical closed mitotic cell division through multiple rounds of DNA replication and asynchronous nuclear division (termed schizogony) to produce merozoites that invade erythrocytes. During this period of cyclic asexual proliferation in the blood stream, a subpopulation undergoes gametocytogenesis to develop into male and female gametocytes, which are transmitted to a mosquito during its blood meal.
Gamete development, fertilization and zygote formation occur in the mosquito midgut, leading to the differentiation of an infective ookinete, which undergoes meiosis, and penetrates the midgut wall to develop into an oocyst on the basal surface of the midgut, where further rounds of closed mitotic cell division occur. Thousands of sporozoites develop within each oocyst, and then egress into the haemocoel to invade the salivary glands and begin a new life cycle.
Lipid metabolism includes essential cellular processes that use fatty acids (FAs) in membrane biosynthesis, energy storage and the generation of signaling molecules. FA elongation and very long chain fatty acid (VLCFA) synthesis occurs in two stages, both consisting of a four-step cyclic process that results in addition of two carbons to the chain with each cycle. In humans, the process involves condensation of acyl-CoA with malonyl-CoA to produce 3-ketoacyl-CoA (catalyzed by one of seven FA elongases), reduction of 3-ketoacyl-CoA by a 3-ketoacyl-CoA reductase (KAR) to 3-hydroxyacyl-CoA, dehydration of 3-hydroxyacyl-CoA to 2,3-trans-enoyl-CoA (catalysed by one of four 3-hydroxyacyl-CoA dehydratase isoenzymes: HACD1-4), and finally reduction to an acyl-CoA with two additional carbon chain units by 2,3-trans-enoyl-CoA reductase (TER) (Kihara, 2012). HACD1-4 were initially annotated as PTPLA, PTPLB, PTPLAD1 and PTPLAD2, respectively due to their similarities to the yeast Phs1 gene product (Ikeda et al., 2008). However, they were renamed to reflect their function as 3hydroxyacyl Co-A dehydratases and their relatedness. HACD1 and HACD2 genes restored growth of yeast SAY32 Phs1-defective cells, indicating that they are functional homologues of Phs1, i.e. 3-hydroxyacyl-CoA dehydratases. Further, studies have indicated that HACD1 has an essential role in myoblast proliferation and differentiation (Lin et al., 2012), with HACD1-deficient cell lines displaying S-phase arrest, compromised G2/M transition and retarded cell growth. Studies of the Arabidopsis thaliana Phs1 homologue PASTICCINO2 or PAS2 showed the protein has an essential role in VLCFA synthesis (Bach et al., 2008), as well as being essential during cell division, proliferation and differentiation (Bellec et al., 2002). Further, Arabidopsis PAS2 complements Phs1 function in a yeast mutant defective for FA elongation (Morineau et al., 2016). PAS2 interacted with FA elongase subunits in the endoplasmic reticulum (ER) and in its absence 3-hydroxyacyl-CoA accumulates, as expected from loss of a dehydratase involved in FA elongation. Similarly, in the yeast Saccharomyces cerevisiae VLCFA synthesis is also catalyzed in the ER by a multi-protein elongase complex, following a similar reaction pathway as mitochondrial or cytosolic fatty acid synthesis (Tehlivets et al., 2007).
In Apicomplexans, the process of fatty acid synthesis and assembly into more complex molecules is critical for their growth and development, while also determining their ability to colonize the host and to cause disease. They acquire lipids through de novo synthesis and through scavenging from the host (Mazumdar et al., 2007), and simple components like mosquito-derived lipids determine within-host Plasmodium virulence by shaping sporogony and metabolic activity, affecting the quantity and quality of sporozoites, respectively (Costa et al., 2018). Fatty acid synthesis (FAS) occurs in the apicoplast via the type II FAS (FASII) pathway, followed by fatty acid elongation (FAE) on the cytoplasmic face of the ER through the elongase (ELO) pathway (Ramakrishnan et al., 2012, Ramakrishnan et al., 2013. Studies on whether FAS is essential in Plasmodium are contradictory. In Plasmodium yoelii the FASII enzymes are only essential during liver stages (Yu et al., 2008; whereas in Plasmodium falciparum, genetic disruption of the FASII enzymes FabI and FabB/F results in complete abolition of sporogony (van Schaijk et al., 2014). Specifically, day 17 to day 23 after mosquito feeding, FabB/F mutant oocysts appeared to degenerate, and expression of the marker GFP was barely detectable, confirming that VLCFA synthesis is crucial for commencement of sporogony. The enzymatic steps of the ELO process are similar to those in the FASII pathway in the apicoplast ; however, the growing chain is held by CoA instead of acyl carrier protein (ACP). In Toxoplasma gondii, the activity of the ELO-pathway is considered an alternative route to FASIIindependent 14 C-acetate incorporation (Bisanz et al., 2006) and is engaged in conventional elongation rather than de novo synthesis (Mazumdar et al., 2007). Indeed, P. falciparum parasites with no functional FASII pathway can still elongate fatty acids; possibly because of the activity of the ELO pathway (Yu et al., 2008).
In a genome-wide study of Plasmodium berghei (Pb) protein phosphatases, we identified 30 phosphatase genes together with one for a predicted protein tyrosine phosphatase-like protein, PbPTPLA, which was shown to be essential for sporozoite formation and completion of the parasite life cycle, but not fully characterized (Guttery et al., 2014). However, despite the original annotation as an inactive PTP-like protein (Andreeva et al., 2008, Wilkes et al., 2008, Guttery et al., 2014, Pandey et al., 2014, more recent functional studies indicate that it is a key component of the VLCFA elongation cycle -more specifically the ELO pathway as a 3-hydroxyacyl-CoA dehydratase (DEH) (Stanway et al., 2019). Therefore, to investigate further the role of DEH in Plasmodium development, we performed an in-depth genotypic and phenotypic analysis of the protein, using in-silico, genetic manipulation and cell biological techniques. We show evolutionary conservation of DEH in the model organisms examined here. Furthermore, we show that PbDEH is located at the ER and is essential for cell division and parasite budding within oocyst but not cell growth of oocysts, thereby blocking parasite transmission.

Ethics statement
All animal work was performed following ethical approval and was carried out under

Identification of conserved domains and evolutionary lineage
The deduced amino acid sequence of PBANKA_134650 (PbPTPLA) now classified as DEH in the manuscript, was retrieved from PlasmoDB (release 27) (Aurrecoechea et al., 2009). Research Tool (SMART) (Schultz et al., 1998) and Protein Family Database (PFAM) (Finn et al., 2008). The deduced amino acid sequence and individual conserved domains were used as BLAST (BLASTP) queries to identify orthologues in PlasmoDB and NCBI protein databases. OrthoMCL database (version 5) was used to identify and retrieve P. berghei orthologues (Table S1) (Li et al., 2003). Multiple sequence alignment was performed for the retrieved sequences using ClustalW (Larkin et al., 2007).
ClustalW alignment parameters included gap opening penalty (GOP) 10 and gap extension penalty (GOE) 0.1 for the pairwise sequence alignment; GOP 10 and GOE 0.2 was used for multiple sequence alignment. A gap separation distance cut-off of 4 and Gonnet protein weight matrix was used for the alignments. Residue-specific penalty and hydrophobic penalties were used, whereas end gap separation and negative matrix were excluded in the ClustalW alignments. The phylogenetic tree was inferred using the neighbor-joining method, computing the evolutionary distances using the Jones-Taylor-Thornton (JTT) model for amino acid substitution with the Molecular Evolutionary Genetics Analysis software (MEGA 6.0) (Tamura et al., 2013). Gaps and missing data were treated using a partial deletion method with 95% site-coverage cut-off and 1000 bootstrap replicates to generate a phylogenetic tree. iTOL was used to visualize the phylogenetic tree (Letunic et al., 2019). The STRING database was used to identify PTPLA interacting proteins (Franceschini et al.). For STRING search, we used a cut-off of 0.70 for the parameters of neighborhood, gene fusion, co-occurrence, co-expression, experiments, databases and text mining results. Predotar (Small et al., 2004) was used for inferring Pf and PbDEH subcellular localization.

Generation of transgenic parasites and genotype analysis
Details of GFP-tagged PTPLA (termed DEH-GFP in this study) and deh (PBANKA_1346500) knockout (KO) parasite lines (Δdeh in this study) are given in (Guttery et al., 2014). For this study, the KO construct was transfected into the GFPCON wild-type line (Janse et al., 2006), with 3 clones produced by serial dilution.

Parasite development in the mosquito
Anopheles stephensi mosquitoes (3-6 days old) were allowed to feed on anaesthetized mice infected with either wild type or mutant parasites at comparable gametocytemia as assessed by blood smears. Mosquitoes were dissected post-blood meal, on the days indicated. For midgut and salivary gland sporozoites, organs from 10-20 mosquitoes were pooled and homogenized, and released sporozoites were counted using a haemocytometer. For oocyst counts, midguts taken at day 7, 14 and 21 post-infection were harvested, mounted on a slide and oocysts counted using phase or fluorescence microscopy. To quantify sporozoites per oocyst (the ratio of number of sporozoites to number of oocysts), an equal number of mosquitoes from the same cage were used to count the number of oocysts and sporozoites. This number varied among experiments but at least 20 mosquitoes were used for each count. For light microscopy analysis of developing oocysts, at least 20 midguts were dissected from mosquitoes on the indicated days and mounted under Vaseline-rimmed cover slips. ER tracker (ThermoFisher) was used to perform co-localization studies according to manufacturer's instructions. Images were collected with an AxioCam ICc1 digital camera fitted to a Zeiss AxioImager M2 microscope using a 63x oil immersion objective. Statistical analyses were performed using Graphpad prism software.

Electron Microscopy
The guts from mosquitoes harvested at 14 and 21 days post-infection were dissected and fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer and processed for electron microscopy (Guttery et al., 2012).

Subcellular fractionation of parasite lysates and detection of DEH
Immunoprecipitation and subcellular fractionation of lysates containing GFP tagged DEH was performed as described previously (Guttery et al., 2014). WT-GFP was used as the control protein in all experiments. In summary, cells from mouse blood infected with the DEH-GFP-expressing parasite were pelleted and then lysed in hypotonic buffer (10 mM Tris-HCl pH 8.4, 5 mM EDTA) containing protease inhibitors (Roche), freeze/thawed twice, incubated for 1 hr at 4°C and then centrifuged at 100,000 g for 30 min. The supernatant was collected as the soluble protein fraction (cytosol). The pellet was resuspended and washed in carbonate solution (0.1M Na 2 CO 3 , pH 11.0) containing protease inhibitors (Roche), and after incubation for 30 min at 4°C the sample was centrifuged again at 100,000 g for 30 min. The supernatant was saved as the peripheral membrane protein fraction (PMF). The residual pellet was solubilized in 4% SDS and 0.5% Triton X-100 in PBS, to form the integral membrane protein fraction (IMF).
Samples from these three fractions, containing equal amounts of protein, were then analyzed by western blot using anti-GFP antibody.

Phylogenetic analysis reveals that DEH is highly conserved among eukaryotes
Genome-wide analysis showed DEH is present in all the eukaryotic organisms studied here, which includes apicomplexans, yeast, fungi, plants, nematodes, insects, birds and mammals. The number of encoded DEH proteins was shown to vary from one to three in the studied organisms, with Plasmodium spp coding for a single DEH. Intriguingly, both Arabidopsis thaliana and Oryza sativa encode three DEHs (PAS2 and 2 HACD isozymes) each, as compared to two (HACD1 and HACD2) in Homo sapiens (plus two sharing relatively weak similarity -HACD3 and HACD4) and two in Mus musculus.
Phylogenetic analysis using the neighbor joining method, clustered organisms based on their evolutionary relatedness ( Figure S1, Table S1). In addition, the phylogenetic analysis suggests that gene duplication in non-chordata, chordata and plants where there are multiple DEH copies, may have happened from a single DEH gene independently to perform specific functions after divergence during evolution, based on the clustering of all DEH isoforms in the same cluster.

Plasmodium DEH does not contain the canonical PTPLA CXXGXXP motif and is
predicted in silico to interact with factors associated with FAE P. berghei (PBANKA_1346500) and P. falciparum (Pf; PF3D7_1331600) DEH genes are annotated as PTPLA (pfam04387) (Andreeva et al., 2008, Wilkes et al., 2008, Guttery et al., 2014, Pandey et al., 2014, the criterion for PTPLA being the presence of a PTP active site motif (CXXGXXR) but with arginine replaced by proline (CXXGXXP).

DEH is expressed throughout the Plasmodium life-cycle stages and localized to the ER
To determine the expression profile and location of PbDEH, we used a single homologous recombination strategy to tag the 3' end of the endogenous deh locus with sequence coding for GFP (Guttery et al., 2014), and then analyzed blood and mosquito stages of the life-cycle for GFP. Strong GFP fluorescence was observed throughout the life-cycle, with areas of localized expression in the cytoplasm and a circular ring formation around the nucleus (Figure 1). Predotar analysis (Small et al., 2004) predicted an ER localization for both PbDEH and PfDEH. Colocalization with ER tracker confirmed the DEH-GFP location at the ER, in all parasite stages analyzed (Figure 2A), with subcellular fractionation of blood stage parasites confirming its integral membrane location ( Figure 2B).

PbDEH is essential for mitotic cell division but not cell size of Plasmodium during oocyst development
Previous comparison of Δ deh and WT parasite lines highlighted the non-essential role of this gene for blood stage development (Guttery et al., 2014). In this study we confirmed it is not essential during asexual blood stages, or for zygote development ( Figure 3A). However, while the overall number of oocysts observed in Δ deh and WT lines was not significantly different (Guttery et al., 2014), there was a significant reduction in Δ deh GFP-expressing oocysts beginning at day 7 and continuing through to day 21 post-infection ( Figure 3B, C), with many appearing to be degenerating. Analysis of oocyst size revealed a small decrease from day 10 onwards in Δ deh lines compared to WT ( Figure 3D), and by day 21 the vast majority of Δ deh oocysts expressed GFP no longer, and in the few that did GFP was present at very low levels or in fragmented patterns. However, it is important to note that Δ deh oocysts that continued to express GFP and showed faint DAPI staining of DNA, were similar in size to WT oocysts; whereas the vast majority of oocysts that reduced in size did not express GFP or stain with DAPI, suggesting they were dead. Analysis of salivary glands from mosquitoes  (Figure 4i, j). In contrast, the majority of Δ deh oocysts exhibited advanced stages of degeneration with increased cytoplasmic vacuolation, dilitated nuclear membranes, and evidence of mitochondrial abnormalities (Fig 4c, d). There was little evidence of retraction of the plasmalemma from the oocyst wall and no evidence that sporozoite inner membrane complex (IMC) formation had been initiated in any of the oocysts. At day 21 post-infection, all Δ deh oocysts were in an advanced stage of degeneration -almost completely vacuolated with a few nuclei appearing to have undergone apoptotic-like nuclear chromatin condensation (Fig 4e, f), in contrast to WT oocysts, which were mostly mature with numerous fully formed and free sporozoites (Fig 4k, l) although a few degenerate oocysts were observed.

Discussion
Lipid metabolism is essential for cellular function, and includes critical pathways for FA synthesis and elongation. DEH is a 3-hydroxyacyl-CoA dehydratase involved in VLCFA synthesis, which interacts with several elongase units, is located at the ER (Beaudoin et al., 2009, Morineau et al., 2016 and has an essential role during development, differentiation, and maintenance of a number of tissue types (Li et al., 2000, Bellec et al., 2002, Pele et al., 2005. In this study, we examined the location and function of Plasmodium DEH using in silico, genetic manipulation and cell biological techniques.
In our previous phosphatome study, a putative, catalytically inactive, PTP-like protein with an essential role during sporogony was identified (Guttery et al., 2014), which had been classified as a putative PTPLA by others (Andreeva et al., 2008, Wilkes et al., 2008, Pandey et al., 2014 based on high sequence similarity and e-score values. However, a recent genome-wide functional screen in P. berghei showed that PbPTPLA has an essential role in lipid metabolism, specifically during the ELO pathway as a 3hydroxyacyl dehydratase (DEH) (Stanway et al., 2019). The specific criterion for a PTPlike protein is the presence of a CXXGXXP motif (i.e. the CXXGXXR motif of PTPs, but with the arginine replaced by proline). However, we show here that this motif is not present in either P. falciparum or P. berghei protein and this, along with its proven function in lipid metabolism (Stanway et al., 2019), suggests that the classification as a phosphatase-like protein is erroneous. Our in silico interactome analysis suggests that PfDEH interacts with a number of proteins involved in lipid metabolism, confirming previous functional findings (Stanway et al., 2019) and adding further weight to its annotation as a 3-hydroxyacyl-CoA dehydratase.
Studies in mammalian systems have suggested that the ER-bound DEH catalyzes the third of four reactions in the long-chain FA elongation cycle (Ikeda et al., 2008). Our detailed GFP-based localization analyses showed that the protein is expressed strongly throughout the life-cycle, with protein expression at localized areas in the cytoplasm and as a circular ring-like structure around the nucleus. In-silico analysis using Predotar and microscopy-based co-localization using ER tracker confirmed the ER location, consistent with previous studies suggesting a role in FAS in Plasmodium (Stanway et al., 2019). Phenotypic analysis of DEH function throughout the life cycle confirmed the results of our previous study (Guttery et al., 2014), highlighting that it is essential for oocyst maturation and sporozoite development, but dispensable for asexual blood stage development (Bushell et al., 2017). Time-course analysis at days 7, 14 and 21 after mosquito infection showed that while early-stage Δ deh oocysts were comparable in size to WT oocysts, they begin to degenerate at an early stage of development, with a significant decrease in GFP-expressing oocysts even at day 7 post-infection, and as seen previously in other FAE-critical mutant parasites (Stanway et al., 2019).
Ultrastructure analysis confirmed that at 14 days post-infection, Δ deh oocysts were at an advanced state of degeneration, with no evidence of sporozoite development.
Retraction of the oocyst plasmalemma (the parasite plasma membrane) from the oocyst capsule is a crucial first stage in sporozoite development, where sporoblast formation is followed by thousands of sporozoites budding off into the space delineated by the capsule . A model of this process is detailed in (Burda et al., 2017). Our study suggests that initiation of mitosis, which results in sporozoite development, does not even commence in Δ deh oocysts, since retraction of the plasmalemma and initiation of daughter IMC formation is ablated. The phenotype is similar to that of a cyclin-3 mutant (Roques et al., 2015), with defects leading to abnormalities in membrane formation, vacuolation and subsequent cell death during the later stages of sporogony.
However, in contrast to the cyclin-3 mutant, oocyst growth was not affected but sporogony was completely ablated in Δ deh parasites, and no transmission was observed in bite-back experiments. This suggests that the parasite cannot scavenge VLCFA from its mosquito host environment, and that DEH (and therefore the ELO pathway) is critical for oocyst mitotic maturation and differentiation. The cells were unable to progress further to form additional lobes and start sporozoite budding in its absence, although the oocyst size was not grossly affected, suggesting that two independent processes drive oocyst formation and sporogony, respectively.
While FASII activity is exclusively in the apicoplast (Shears et al., 2015), our study showed that DEH-GFP is located at the ER, suggesting that it is an active component of the ELO pathway. The genes involved in the ELO pathway include members of the ELO family (1, 2 and 3), of which ELO2 and ELO3 are involved in keto-and enoyl-reduction (Kohlwein et al., 2001). In yeast, the dehydratase step is carried out by the DEH-like homologue, Phs1, which has also been characterized as a cell cycle protein with mutants defective in the G2/M phase (Yu et al., 2006). Gene knockout studies for any ELO proteins are few, with a single genome-wide functional analysis showing that the P. berghei homologue of PF3D7_0605900 (a putative long chain polyunsaturated fatty acid elongation enzyme) is dispensable during the asexual blood stages (Bushell et al., 2017). In addition, the comprehensive analysis of FAE in Plasmodium by Stanway et al. (Stanway et al., 2019) showed that mutants of a ketoacyl-CoA reductase (KCR) have an identical phenotype to our DEH gene knockout lines, with normal development of ookinetes and oocysts gradually disappearing over the course of development, resulting in the complete ablation of sporogony. In contrast, ELO-A (stage 1 of VLCFA synthesis) mutants were critical for liver stage development. This suggests that reduction of ketoacyl-CoA to hydroxyacyl-CoA and subsequent dehydration of hydroxyacyl-CoA to enoyl-CoA (i.e. stages 2 and 3 of VLCFA synthesis) are the most crucial stages for oocyst maturation and sporogony; whereas the lack of a phenotype during sporogony of the ELO-A deletion may suggest functional redundancy and/or a compensatory mechanism such as an overlapping specificity for condensation of malonyl-CoA by either ELO-B or ELO-C, as suggested in Trypanosoma brucei (Lee et al., 2006) and Toxoplasma gondii (Ramakrishnan et al., 2012).
In conclusion, our PbDEH analysis using various in-silico, in-vitro and in-vivo approaches provides important insights into the crucial role DEH plays during VLCFA synthesis, and how disruption of the gene can affect parasite development in the mosquito. Future studies will elucidate further how lipid metabolism in Plasmodium can be explored as a viable target for therapeutic intervention.