Identification of endoplasmic reticulum-shaping proteins in Plasmodium parasites

In eukaryotic cells, the endoplasmic reticulum (ER) is a continuous membrane system involved in many critical cellular processes, including protein synthesis, lipid synthesis, and calcium storage. Morphologically, the ER is composed of cistern-like sheet structures and a reticular network of tubules. Classes of integral membrane proteins that shape the ER have been identified: the reticulons and DP1/Yop1p proteins generate ER tubules by inducing high curvature in the membrane (Voeltz et al., 2006), the atlastin GTPases and Sey1p/RHD3 proteins mediate fusion of ER membranes, forming a tubular network (Hu et al., 2009; Orso et al., 2009), and Climp63, Kinectin, and p180 play a role in stabilizing ER sheets (Shibata et al., 2010). Mutations in the determinants of ER tubules cause growth defects, short root hairs, and a neurodegenerative disease called hereditary spastic paraplegia (HSP) (Hu et al., 2009), and sheet-formation proteins are strongly upregulated in professional secretory cells when ER sheet expansion is needed, suggesting that ER morphology is tightly associated with its physiological functions (Shibata et al., 2010). ER morphology has rarely been studied in protozoan parasites, likely due to difficulties posed by the lack of genetic manipulation and relatively small scale of these cells. In Entamoeba histolytica, the intestinal protozoan that causes invasiveamebiasis, theERwasfirst thought tobecomposedof vesicles of varying size but recently shown to be a continuous network (Teixeira and Huston, 2008). In Toxoplasma gondii, the infectious agent resulting in toxoplasmosis, expansion and partitioning of the ER has been followed during the cell cycle (Nishi et al., 2008). The ER in Plasmodium parasites, the causative agents of malaria, has been visualized during the erythrocytic cycle (vanDoorenet al., 2005); it transforms froma perinuclear structure with no distinctive morphological characters into a reticular network throughout the cytoplasm. Plasmodium develops first in the Anopheles mosquito, and then invades the liver upon injection into mammalian hosts. Whether the ER adopts its characteristic shapes in these stages is yet to bedetermined, and themolecular determinants of the ER in Plasmodium parasites remain unclear. To identify Plasmodium falciparum orthologs of ER-shaping proteins, we conducted Blast searches of the P. falciparum genomic database (PlasmoDB, www.plasmoDB.org, v. 6.3, released December 22, 2009) using Saccharomyces cerevisiaeYop1p (ScYop1p) and humanDP1protein sequences as queries.Bothsearches revealed35%–42% identity (55%–60% homology) with PFC0730w (Gene ID PF3D7_0316700 in PlasmoDB v.26, released October 15, 2015). The PFC0730w sequence was further used to query P. berghei proteins using the tblastn tool. This search returned PBANKA_0414500 and PBANKA_1135000 as the closest P. berghei homologs (Fig. S1). PBANKA_0414500 was recently annotated as a putative HVA22/TB2/DP1 family protein and PBANKA_1135000 as a HVA22-like protein. Because HVA22 is the Yop1p homolog in plants and TB2 is a previously used alias forDP1, our homologous searches confirm the annotation of the genomic database. We renamed PBANKA_0414500 as PbYOP1 and PBANKA_1135000 as PbYOP1L (Fig. 1A and 1B).When searching for the reticulon homolog in theP. berghei genome, we found PBANKA_1139900 and renamed it PbRTN1 (Fig. 1B). According toPlasmoDBdatabase, the three potential Plasmodium tubule-forming proteins have distinct expression profiles during the asexual cycle: the expression of PbYOP1 gradually increases and peaks in trophozoites and schizonts, the levels of PbYOP1L peak in rings and trophozoites but decrease in schizonts, and levels of PbRTN1 are relatively constant. These data suggest non-redundant functions of these proteins. Thus, three potential ER tubule-forming proteins were identified in P. berghei. A common feature of ER tubuleforming proteins is a reticulon-homology domain (RHD) consisting of two tandem transmembrane hairpins (TMH) (Fig. 1B). The sequences of all three candidates for ER tubule formation in P. berghei exhibited characteristics of the RHD domain. Notably, the intervening loop between the two TMHs in PbRTN1 is much longer than that of Yop1 family proteins, including PbYOP1 and PbYOP1L (Fig. 1B). In contrast, PbYOP1 bears a much longer N-terminus than PbYOP1L and PbRTN1. The length of the TMHs is also variable, with the first TMH of PbYOP1L and the second TMH of PbRTN1 being shorter than other TMHs (Figs. 1B and S2), suggesting that it partially traverses the lipid bilayer. Importantly, the primary structure of PbYOP1 is highly conserved among Plasmodium species (Fig. S2), implying a fundamental role of the protein.

Foscholine-12. The elution was collected and concentrated to 500 µL and 30U thrombin added to remove the His-tag overnight. The protein was then further purified by gel filtration (Superdex 200,GE healthcare) in TSG containing 0.1% Foscholine-12. Selected fractions were concentrated to 2 mg/mL for reconstitution.

Reconstitution
For SUV production, POPC and DOPS (Avanti Polar Lipid) were mixed at a molar ratio of 85:15 (10 mM of total lipid). The mixture was then dried under a stream of N2 and rehydrated in TSG1 buffer (50 mM Tris pH 8.0, 150 mM NaCl, 10% glycerol).
After 10 freeze-thaw cycles, the vesicles were then extruded 11 times through the extruder (Avanti Polar Lipid) with 100 nm or 400 nm filter membranes. Purified PbYOP1 (final concentration 0.4 mg/mL) in Foscholine-12 was then mixed with the pre-formed SUV (final concentration 0.4 mM) for 30 min at room temperature. The detergent was removed by addition of SM-2 Bio-beads four times at room temperature. Finally, the proteoliposomes were cleared by centrifugation at 14,000 rpm for 10 min.

Electron microscopy
Negative staining was performed with 2% uranyl acetate. First, 5 µL of the 5X diluted proteoliposome sample was placed onto a carbon-coated copper grid for 1 min, the excessive sample dried by filter paper, and the grid washed with deionized water. A total of 5 µL of filtered 2% uranyl acetate was placed on the grid and excessive stain dried by filter paper. Images were collected at room temperature using a Hitachi TEM system operated at an acceleration voltage of 100 kV. Images were recorded at a magnification of 30,000 and a defocus value of 1.5 µm. All images were recorded on a 2k x 2k CCD camera.
Proteoliposomes treated with 20 µL 1% digitonin, Triton X-100, or SDS were loaded onto the top layer of the gradient. The samples were then centrifuged at 174,000 rpm in a TLS-55 rotor (Beckman Coulter) for 2 h at 4˚C. Each fraction sample was analyzed by Western blot with HA-tag antibody.

Circular dichroism
Synthesized peptide (50 µM) in 10 mM potassium phosphate (pH 7.5) was mixed with TSG buffer and 100 nm SUV (1 mM lipids) or 400 nm SUV (1 mM lipids) separately. Circular dichroism was performed on a Biologic MOS-450 instrument at room temperature. Spectra were collected from 190-260 nm at a bandwidth of 1 nm.

COS-7 cells (American Type Culture Collection) were maintained in complete
Dulbecco's minimum essential medium supplemented with 10% fetal bovine serum in 5% CO2 at 37°C. Transfections were performed using Turbo (Thermo) according to the manufacturer's instructions. To generate ATL-deleted cells by CRISPR/Cas9 genome editing, guide RNA (gRNA) sequences were designed using the CRISPR design tool as following, with protospacer adjacent motifs (PAMs) underlined: gRNA containing oligonucleotides were introduced into the pX330 vector.
CRISPR/Cas9 plasmids were transiently transfected into COS-7 cells along with pLKO.1-puro at 1:1:1 ratio using TurboFect transfection reagent (Thermo). 24 hours later, transfected cells were selected with 1 μg/ml puromycin for 1 week. The cells were then sorted for single cell into a 96 well plate by a BD FACS Aria-II sorter.
Different single clones were verified by immunoblotting and sequencing.
All images were captured on an OLYMPUS FV1200 confocal microscope with a 60×/1.40 NA Plan Apochromat oil immersion objective lens using the Olympus Fluoview Version 2.0b Viewer. Brightness and contrast were adjusted across the entire image using Adobe Photoshop. PbYOP1L ETKLWLTYWVVFSLFFFIEYLI-DIILFWIPFYYVIKLLFLLYLYMPQVRGAETVYNYII HVA22 DDEQWLAYWILYSFITLLEMVA-EPVLYWIPVWYPVKLLFVAWLALPQFKGASFIYDKVV .