Besnoitia besnoiti infection alters both endogenous cholesterol de novo synthesis and exogenous LDL uptake in host endothelial cells

Besnoitia besnoiti, an apicomplexan parasite of cattle being considered as emergent in Europe, replicates fast in host endothelial cells during acute infection and is in considerable need for energy, lipids and other building blocks for offspring formation. Apicomplexa are generally considered as defective in cholesterol synthesis and have to scavenge cholesterol from their host cells for successful replication. Therefore, we here analysed the influence of B. besnoiti on host cellular endogenous cholesterol synthesis and on sterol uptake from exogenous sources. GC-MS-based profiling of cholesterol-related sterols revealed enhanced cholesterol synthesis rates in B. besnoiti-infected cells. Accordingly, lovastatin and zaragozic acid treatments diminished tachyzoite production. Moreover, increased lipid droplet contents and enhanced cholesterol esterification was detected and inhibition of the latter significantly blocked parasite proliferation. Furthermore, artificial increase of host cellular lipid droplet disposability boosted parasite proliferation. Interestingly, lectin-like oxidized low density lipoprotein receptor 1 expression was upregulated in infected endothelial hostcells, whilst low density lipoproteins (LDL) receptor was not affected by parasite infection. However, exogenous supplementations with non-modified and acetylated LDL both boosted B. besnoiti proliferation. Overall, current data show that B. besnoiti simultaneously exploits both, endogenous cholesterol biosynthesis and cholesterol uptake from exogenous sources, during asexual replication.


exogenous cholesterol and desmosterol supplementation boosts parasite proliferation.
To control the role of exogenous cholesterol sources for parasite proliferation, we supplemented cholesterol and its precursor desmosterol by direct addition of the cell culture medium. As previously shown 38 , ethanol-dissolved cholesterol or desmosterol indeed directly acts on cells despite the hydrophobic characteristics of these molecules. In fact, supplementation with both desmosterol and cholesterol significantly enhanced B. besnoiti tachyzoite production in infected host cells (cholesterol p < 0.01; desmosterol p < 0.05; Fig. 1C). When comparing desmosterol and cholesterol for their effects, supplementation with cholesterol boosted parasite proliferation stronger than with desmosterol (p < 0.01). In addition, we supplied these molecules via MβCD complexes, in order to generate inclusion complexes that are able to donate these molecules to the host cell 39 . However, treating BUVEC with such complexes did not improve parasite proliferation (data not shown). In addition, treatments of tachyzoites with cholesterol-MβCD-complexes prior to infection also failed to influence parasite proliferation (data not shown).

Neutral lipids and lipid droplet formation are enhanced in B. besnoiti-infected host cells. Neutral lipids and lipid droplets could be visualized via Bodipy 493/503 and Nile Red staining in B.
besnoiti-infected BUVEC as well as in free tachyzoite stages. As estimated by fluorescence microscopy, infected cells showed increased numbers of lipid droplets in the cytoplasm of the host cell ( Fig. 2A1-4). These findings were verified on a quantitative level via a FACS-based approach. By applying two different MOIs (3:1 and 4:1) we could show that lipid droplets are significantly enhanced in B. besnoiti-infected endothelial host cells when compared to non-infected controls (all p < 0.01, Fig. 2B). However, comparing the different MOIs or time points of infection, no significant differences were detected.
Microscopic analyses showed a stronger Nile Red staining in B. besnoiti-infected host cells (Supplementary Fig. 2; white arrows) when compared to non-infected cells within the same cell layer (Supplementary Fig. 2; orange arrows). Single cell measurements (from 3 different BUVEC isolates) confirmed significantly enhanced levels of neutral lipids in B. besnoiti-infected host cells (p = 0.0181, Fig. 2C). Tachyzoite stages also showed strong reactions after Nile Red staining indicating the presence of neutral lipids. As also shown for filipin staining, the strongest reactions were apparent in the posterior part of the tachyzoites (Fig. 2D1-2, arrows).
whether an increase of lipid droplet formation is beneficial for optimal parasite proliferation, we here artificially enhanced host cellular lipid droplet numbers by oleic acid treatments prior to infection. As depicted in Fig. 2E, the artificial enhancement of lipid droplet disposability indeed significantly boosted B. besnoiti tachyzoite www.nature.com/scientificreports www.nature.com/scientificreports/ production. Thus both, the number of freshly released (=extracellular, Fig. 2E1, p = 0.0109) and still intracellular (Fig. 2E2, p = 0.0259) tachyzoites (both log-transformed in the comparison) was found upregulated in oleic acid-treated BUVEC resulting in a 9-fold and 2.5-fold increase of parasite proliferation within 48 h, respectively.
Live cell 3D holotomographic microscopy. 3D holotomographic microscopy confirmed the presence of numerous lipid droplet-like structures in B. besnoiti-infected cells (Fig. 2A4, arrows). To prove the lipid Arithmetic mean of three biological with three technical replicates each, standard deviation; *p < 0.05. (B) Effects of CI976 (inhibitor of cholesterol esterification) treatment on tachyzoites proliferation: BUVEC were treated with CI976 (2.5, 5, 10 and 20 µM) 24 h before B. besnoiti infection. Non-treated host cells served as controls. 48 h after infection, the number of tachyzoites present in cell culture supernatants were measured. Bars represent arithmetic means of three biological replicates, standard deviation (**p < 0.01). (C) Effects of C75 (inhibitor of fatty acids synthesis) treatment on B. besnoiti replication. BUVEC were treated with C75 (2.5, 5, 10 and 20 µM) 24 h before B. besnoiti infection. Non-treated host cells served as controls. 48 h after infection, the number of tachyzoites present in cell culture supernatants was measured. Bars represent arithmetic means of three biological replicates, standard deviation (**p < 0.01).
www.nature.com/scientificreports www.nature.com/scientificreports/ droplet-nature of these structures, holographic tomography and epifluorescence analyses (Bodipy 493/503 staining) were performed in parallel on BUVEC which proved the precise matching of these two independent techniques (Fig. 3A). The mean refractive index (RI) of Bodipy stained lipid droplets was estimated (n = 50) and was 1.355 ± 0.00333. These parameters (RI > 1.355) were then applied to B. besnoiti-infected BUVEC (24 h p. i.) in 3D holotomographic microscopy and confirmed the presence of a high number of lipid droplets in infected cells (Fig. 3B). Moreover, Nile Red stained BUVEC also presented the same features (Fig. 3C).
Cholesterol esterification is essential for optimal parasite proliferation. Lipid droplets represent the main storage organelles for esterified cholesterol. Given that lipid droplets were found enhanced in B. besnoiti-infected BUVEC, we here analysed by biochemical means whether esterified cholesterol content was upregulated by B. besnoiti infection. Referring to total cholesterol content, B. besnoiti-infected BUVEC indeed showed a 1.5-fold, significant increase of esterified cholesterol levels when compared to non-infected controls (Fig. 4A, p = 0.045), thereby indicating infection-induced enhancement of free cholesterol conversion via esterification. To analyse further the role of cholesterol esterification for parasite proliferation we additionally performed functional inhibition experiments by the use of CI976, an inhibitor of the cholesterol-esterifying enzyme, sterol O-acyltransferase. Overall, CI976 treatments effectively inhibited B. besnoiti proliferation in a dose-dependent (p = 0.0038) manner (10 μM and 20 µM treatments: both p < 0.01; Fig. 4B). Thus, CI976 treatments led to a reduction of tachyzoite production of 4.6%, 14.1%, 64.7% and 82.9% when the cells were treated with 2.5, 5, 10 and 20 µM CI976 (Fig. 4B). Based on the inhibition of tachyzoite production, an IC 50 of 7.56 µM was calculated for CI976 treatments. Microscopic control revealed that the host cells themselves were not altered in their morphology by CI976 treatments.
For cholesteryl ester formation, the hydroxyl group of cholesterol is linked to the carboxylate group of a fatty acid. Therefore, we additionally tested whether the blockage of fatty acid synthesis by the compound C75 would also impair B. besnoiti proliferation. C75 treatments of B. besnoiti-infected host cells moderately inhibited tachyzoite formation in dose dependent reduction (p = 0.049). However, only at 20 µM concentration a significant reduction of the number of parasites was observed when compared to solvent control (DMSO 0.1%; p = 0.0032; Fig. 4C). www.nature.com/scientificreports www.nature.com/scientificreports/ endogenous cholesterol synthesis is enhanced in endothelial host cells and appears essential for optimal B. besnoiti proliferation. Endogenous de novo cholesterol synthesis is performed by a multi-step biochemical pathway being supported by numerous enzymatic reactions. Given that analyses on gene transcription or protein expression of certain involved molecules may not precisely reflect their true enzymatic activity, we here analysed the actual content of cholesterol-related sterols (e. g. cholesterol precursors, metabolites) via biochemical means in B. besnoiti-infected BUVEC and control cells. Therefore, the content of three groups of cholesterol-related sterols were measured: i) biosynthetic precursors of cholesterol in the endogenous synthesis pathway serving as indicators of cellular cholesterol de novo synthesis (lanosterol, dihydrolanosterol, lathosterol 7-dehydrocholesterol desmosterol); ii) downstream metabolites of de novo synthesis and indicators of cholesterol conversion [cholestanol, 24-hydroxycholesterol (24-OHC), 25-hydroxycholesterol (25-OHC), 27-hydroxycholesterol (27-OHC), 7α-hydroxycholesterol (7α-OHC), 7-ketocholesterol (7-ketoC), 4β-hydroxycholesterol (4β-OHC) and 7β-hydroxycholesterol (7β-OHC)] with some of these molecules (e. g. 7-ketoC, 7α-OHC and 7β-OHC) being recognized as indices of cellular oxidative stress; iii) phytosterols as indicators of cholesterol-uptake from the extracellular environment (campesterol, stigmasterol and sitosterol). We here calculated sterol:cholesterol ratios (Table 1) which are generally accepted as indicative for endogenous cholesterol synthesis rates. Notably, the individual BUVEC isolates differed considerably in their basic absolute levels of sterols as indicated by rather high standard deviations already present in non-infected samples.
Overall, concerning indicators of endogenous cholesterol synthesis, B. besnoiti infections indeed led to a shift of sterol:cholesterol ratios by that way that several cholesterol precursors were found at enhanced contents in parasite-infected BUVEC. Overall these reactions proved significant for lanosterol (p < 0.0001), dihydrolanosterol (p = 0.0071), lathosterol (p = 0.0007) and 7-dehydrocholesterol (p = 0.0008) and indicated that host cellular de novo synthesis of cholesterol is upregulated by B. besnoiti infection (Table 1). In contrast to indicators of endogenous cholesterol synthesis, most oxysterols representing downstream metabolites were not found changed in their sterol:cholesterol ratios. Thus, oxysterol:cholesterol ratios denied any positive shift of these ratios in infected cells, but even confirmed a slight decrease of the 27-OHC:cholesterol ratio (p = 0.037) in infected cells ( Table 1). Given that especially 7β-OHC, 7-ketoC and 7α-OHC upregulation may indicate oxidative cell stress reactions since these molecules represent autoxidatives that are mainly formed by radical oxidative species, B. besnoiti infections do not appear to cause considerable oxidative stress in bovine endothelial host cells. Along with enhanced endogenous cholesterol de novo synthesis, cholesterol-related needs can also be satisfied via an enhanced sterol uptake from the extracellular environment. Since phytosterols exclusively are of plant origin and are submitted to the cells via the FCS fraction, intracellular phytosterols levels are often used as indices of sterol uptake. Although the overall absolute levels of sitosterol, stigmasterol and campesterol were found slightly increased in B. besnoiti-infected host cells (sitosterol: 1.2-fold, stigmasterol: 1.4-fold and campesterol: 1.5-fold), phytosterol:cholesterol ratios did not confirm enhanced levels and even showed slightly reduced values for stigmasterol (p = 0.0105) and sitosterol (p = 0.0118) (Table 1) in B. besnoitia-infected cells.
Given that cholesterol-related sterol profiling indicated enhanced endogenous cholesterol synthesis rates, we here additionally performed functional inhibition experiments using blockers of the mevalonate biosynthesis pathway. Therefore, statin treatments were here applied. Statins represent a class of drugs widely used to lower plasma cholesterol levels 40 . We used lovastatin, which affects the total cellular isoprenoid/steroid synthesis and thus interferes at a very early step of de novo synthesis by blocking HMG-CoA-reductase (HMGCR). Overall, lovastatin treatments exhibited dose-dependent significant effects on parasite proliferation (p = 0.02). Thus, tachyzoite production was reduced for 21% and 66% when infected cells were treated with 10 µM (p < 0.05) and www.nature.com/scientificreports www.nature.com/scientificreports/ 20 µM (p < 0.01) lovastatin, respectively (Fig. 5A). Based on the inhibition of tachyzoite production an IC 50 of 11.31 µM was calculated for lovastatin treatments.
Besides lovastatin, we also used zaragozic acid (syn. squalestatin) for treatments of B. besnoiti-infected BUVEC. Zaragozic acid is a squalene synthase inhibitor, which directly targets sterol synthesis. Zaragozic acid treatments resulted in a reduction of parasite replication (Fig. 5B), but effects were less prominent than those observed with lovastatin and required higher inhibitor concentrations. However, 60 µM zaragozic acid treatment induced a significant reduction of B. besnoiti replication (48.8%, p < 0.05). These data underline the key role of cellular de novo synthesis for successful parasite replication.
B. besnoiti infections induce LOX-1 but not LDLR expression and profit from exogenous sterol uptake. Besides being synthesized endogenously, cholesterol may also be taken-up from extracellular sources via receptor-mediated LDL incorporation. The endothelial cell type is well-known to internalize different LDL species, such as non-modified LDL (LDL), acetylated LDL (acLDL) or oxidized LDL (oxLDL). LDL uptake is preferentially mediated by the classical LDL receptor (LDLR) whilst a series of non-classical, so-called scavenger receptors (e. g. LOX-1, SRB1) preferentially promote modified LDL incorporation. In this study we focused on two key receptors, LDLR and LOX-1, and analysed whether the gene transcription and protein expression of these receptor was influenced by B. besnoiti infections in BUVEC. Gene transcriptional profiling revealed that LDLR and LOX-1 were differentially altered by parasite infection. Whilst LDLR gene transcription and protein expression was not altered in B. besnoiti-infected BUVEC (Fig. 6A,C), LOX-1 gene transcripts were found upregulated throughout B. besnoiti in vitro infection peaking at 12 h p. i. (Fig. 6B). In agreement, LOX-1 protein expression was also found upregulated in infected cells: this was confirmed by two methods, a commercial LOX-1-specific ELISA (Fig. 5D Given that LDLR and LOX-1differentially promotes LDL species uptake, we here analysed whether supplementation of different LDL species (LDL, acLDL, oxLDL) is beneficial for parasite proliferation. Thus, LDL and acLDL was supplemented to B. besnoiti-infected BUVEC cultures at 10 µM final concentration. In case of oxLDL, the final concentration had to be reduced to 2.5 µM since this LDL variant revealed as toxic for BUVEC at higher concentrations. As depicted in Fig. 5F, both LDL and acLDL significantly boosted B. besnoiti tachyzoite production when compared to non-supplemented controls (LDL: p < 0.01; acLDL: p < 0.001). Moreover, acLDL induced a significantly stronger parasite proliferation than LDL supplementation (p < 0.05). In contrast, oxLDL failed to improve parasite replication at 2.5 µM supplementation. These data revealed that optimal B. besnoiti proliferation depends on the supply of exogenous cholesterol or other lipids and additionally proves that infected host cells may profit from different LDL variants, i. e. from modified and non-modified LDL.

Discussion
Apicomplexan parasites are generally considered as defective in cholesterol synthesis. Given that apicomplexan species are obligate intracellular parasites with highly proliferative capacities, this metabolic characteristic renders these pathogens as highly dependent on their respective host cells in terms of cholesterol supply. In principle, host cells support two major pathways of cholesterol resourcing: endogenous de novo synthesis and exogenous sterol uptake. Most reports indicate that different apicomplexan species may utilize diverse pathways of cholesterol acquisition in a species-or even host cell type-dependent manner.
The current data demonstrate that B. besnoiti uses both pathways of cholesterol acquisition when infecting bovine endothelial cells, which correspond to host cells to be infected during the acute phase of cattle besnoitiosis 2,41 . This is in agreement to recent data on a slow proliferating apicomplexan species, Eimeria bovis 27,36,37 but differs from findings on T. gondii 17 . Since we here measured the cellular content of cholesterol biosynthetic precursors and metabolites via GC-MS-based approaches, the current data should directly reflect the actual biochemical situation in B. besnoiti-infected endothelial host cells. Sterol:cholesterol ratios of cholesterol biosynthetic precursors, which are generally accepted as reflecting the activity of the endogenous cholesterol biosynthetic pathway 42 , showed a significant enhancement in B. besnoiti-infected host cells. Accordingly, an infection-driven increase of lanosterol-, dihydrolanosterol-, 7-dehydrocholesterol-and lathosterol-related ratios was detected. In line, a pivotal role of cholesterol de novo synthesis was also suggested for E. bovis or T. gondii infections 27,35,36,43 and in case of E. bovis these data also relied on GC-MS-based analyses 37 . These biochemical data were furthermore supported by a significant reduction of B. besnoiti proliferation triggered by statin treatments, which interfere with endogenous cholesterol synthesis. Statins represent a class of drugs widely used to lower plasma cholesterol levels 40 . Statin treatments also proved effective in other apicomplexan-related infection systems, such as T. gondii-infected macrophages 35,43 , Plasmodium-and Babesia-infected erythrocytes 44 , C. parvum-infected epithelial cells 33 or E. bovis-infected BUVEC 36 . The current data revealed a higher efficacy of lovastatin, which interferes with the total cellular isoprenoid/steroid synthesis, when compared to zaragozic acid that directly targets sterol synthesis. Noteworthy, the blockage of merozoite production in other apicomplexan also depends on the choice of statin. As such, treatments with rosuvastatin and atorvastatin failed to influence parasite proliferation, whilst the application of lovastatin in higher concentrations (~72-96 µM) reduced tachyzoite production for more than 50% in T. gondii-infected macrophages 43 . Zaragozic acid is an inhibitor of squalene synthase and for this reason considered as more specific for cholesterol blockage than other statin treatments 45 . In B. besnoiti-infected host cells, 60 µM zaragozic acid treatments resulted in a significant blockage of tachyzoites production (reduction of 48.8%) confirming that B. besnoiti replication depends on host cell de novo synthesis. In contrast to E. bovis with 70.2% reduction rate at 5 µM treatment 36 , 15 µM zaragozic acid treatments in C. parvum-infected epithelial cells only induced moderate effects, as indicated by 25% growth delay 33 . Comparable rates of reduction were described in T. gondii-infected macrophages with 1-10 µM zaragozic acid 35 . Squalene synthase-defective CHO cells revealed no significant anti-proliferative effects on T. gondii development compared to non-defective controls 17 . However, (2019) 9:6650 | https://doi.org/10.1038/s41598-019-43153-2 www.nature.com/scientificreports www.nature.com/scientificreports/ other authors applied two quinuclidine-based inhibitors of squalene synthase in T. gondii-infected epithelial cells and described anti-proliferative effects of both compounds achieving a similar percentage of reduction of tachyzoite replication as in B. besnoiti-infections (48-58% reduction) but with much lower doses of zaragozic acid (3 µM) 46 . Overall, the data on cholesterol biosynthetic precursors and inhibitor treatments indicate that successful B. besnoiti replication depends on the host cell cholesterol de novo synthesis. Nevertheless, the fact that replication is not entirely blocked by zaragozic acid treatment may argue for additional sources of cholesterol besides de novo synthesis.
As a general feature, sterol up-take from the extracellular compartment seems to be exploited by apicomplexan parasites in a species-specific and host cell-specific manner. Whilst this pathway appeared of major importance in the case of T. gondii in CHO cells 17 , C. parvum in epithelial cells 33 , N. caninum infections 34 or E. bovis-infected endothelial cells 37 , a minor or even absent role of LDL-mediated cholesterol supply was reported www.nature.com/scientificreports www.nature.com/scientificreports/ for T. gondii infections in macrophages 35 or for Plasmodium spp. infections in hepatocytes 32 . In the case of B. besnoiti infections, the current data on LDL supplementation confirm a pivotal role of exogenous sterol uptake for parasite proliferation. For the first time, we here applied different LDL species (LDL, acLDL, oxLDL), which are present in blood or lymph and may therefore all serve as exogenous sources in vivo. Interestingly, the supplementation with both, acLDL and LDL, revealed as beneficial for B. besnoiti proliferation and boosted tachyzoite production. LDL supplementation was also beneficial for T. gondii proliferation in CHO cells 17 whilst such a treatment had no stimulatory effects on hepatic Plasmodium spp. and C. parvum proliferation 32,33 or on T. gondii growth in macrophages 35 indicating parasite-and cell type-specific mechanisms. However, to our best knowledge this signifies the first report on acLDL-mediated effects on intracellular apicomplexan proliferation.
In addition to LDL-related data, the direct exogenous supply of excess cholesterol or desmosterol also boosted B. besnoiti proliferation. This is in line with other data that proved this method of cholesterol supplementation as effective in the case of T. gondii in CHO cells 17 or endothelial E. bovis infections 27 . So far, it remains unclear why the supplementation of cholesterol via MβCD complexes failed to influence parasite proliferation which contrasts data of other studies 47 . Referring to exogenous sterol supplementation, the current data on phytosterols measurements appear somewhat conflicting. Phytosterols signify cholesterol analogues that exclusively originate from plant dietary intake 48 . In cell cultures, these molecules are derived from FCS present in the cell medium and an enhanced cellular content of plant sterols is commonly accepted as an indicator of cellular sterol uptake from the extracellular environment 42,48 . However, current biochemical measurements data failed to show enhanced phytosterol:cholesterol ratios in B. besnoiti-infected BUVEC but even revealed slightly reduced values for stigmaand sitosterol. So far, we have no plausible explanation for these findings. While cell type-specific differences in individual phytosterol species uptake were reported, the capacity of BUVEC to incorporate all three phytosterols here detected was previously proven since a significant enhancement of sitosterol, stigmasterol and campesterol contents was reported in E. bovis-infected endothelial cells 37 .
Given that LDL species are internalized via a receptor-mediated uptake, we here furthermore analysed whether B. besnoiti infections influence the gene transcription and protein expression of two classical endothelial LDL-related receptors. In contrast to reports on T. gondii infections in CHO cells 17 and E. bovis infections in BUVEC 27 , LDLR expression was not altered by B. besnoiti infections. These data were in accordance to the fact that reduced LDLR expression did not affect the liver stage burden in the case of Plasmodium spp. 32 . Contrary to LDLR, an infection-driven upregulation of the scavenger receptor LOX-1 was here found in B. besnoitia-infected BUVEC on both, gene transcriptional and protein expression level. Recent microarray data on B. besnoiti-infected BUVEC confirmed a significant upregulation of LOX-1 (Silva L.M.R., unpublished data). LOX-1 is considered as an important receptor for ox-LDL internalization in vascular endothelial cells [49][50][51] but is also able to bind acLDL at a comparable affinity 52 . This receptor was recently also reported to be upregulated in E. bovis infections of bovine endothelial cells 27,53 . Given that in both cases of LOX-1 induction, endothelial host cells served as host cells may indicate a cell type-specific mechanism.
Excess cellular levels of free cholesterol demand for cholesterol efflux or conversion since too high concentrations are toxic for cells. Two main conversion routes in endothelial cells involve cholesterol esterification and oxidation. In the current study, a significantly enhanced level of esterified cholesterol was detected in B. besnoiti-infected cells when compared with non-infected controls. This corresponds to recent findings in E. bovis-infected BUVEC 37 . In line, the key role of cholesterol esterification was here additionally confirmed by its chemical blockage via CI976 treatments leading to a significant inhibition of B. besnoiti replication. Thus, tachyzoite production was reduced by 64.7 and 82.9%, at 10-and 20-µM treatments, respectively. These data are in agreement with reports on T. gondii and E. bovis documenting the essential role of cholesterol esterification for optimal parasite proliferation 19,28 . Concerning relevant inhibitor concentrations, it has to be noted that antiproliferative effects occurred at a comparable level in T. gondii (merozoite reduction rates of approximately 60 and 70% induced by 4-and 10-μM CI976 treatments) and B. besnoiti infections. Even stronger effects were reported for E. bovis infections since 5 µM treatments almost entirely inhibited merozoite I production (99.6% reduction) as mirrored by a low IC 50 of 0.34 µM 36 . The higher sensitivity of E. bovis to CI976 treatments may be due to a stronger need for cholesteryl ester formation during macromeront formation (>120.000 merozoites I) when compared to non-macromeront-forming parasites. Interestingly, T. gondii appears to be able to synthesize and store cholesteryl esters by itself, if host cell cholesterol is available 19,28 . In fact, two SOAT-like molecules were identified in T. gondii stages and proved sensitive to SOAT inhibitor treatments 28,54 . So far, no data are available with this respect for B. besnoiti. Nevertheless, it remains to be elucidated whether CI976-driven, detrimental effects on B. besnoiti proliferation accounted only to the host cell compartment or were also brought about by direct antiparasitic effects, as in the case of T. gondii.
Cholesteryl esters play a pivotal role in the development of several apicomplexan parasites and cholesterol and fatty acids are needed for cholesteryl ester formation, though the effects of fatty acid synthase blockage was additionally here investigated. The synthetic α-methylene-γ-butyrolactone compound C75 inhibits fatty acid synthase activity and has been studied for its anti-inflammatory and anti-tumoral activities [55][56][57] . C75 treatments induced a significant reduction of B. besnoiti replication at 20 µM concentration. However, in the case of E. bovis, lower concentrations of C75 were needed for anti-proliferative effects indicating a high relevance of fatty acids for this parasite 36 . Moreover, C75 inhibition was ineffective on Trypanosoma cruzi growth in infected macrophages 58 .
Overall, B. besnoiti infections failed to trigger the synthesis of oxysterols in host endothelial cells. Interestingly, enzymatically synthesized side-chain oxysterols (e. g. 24-OHC, 25-OHC, 27-OHC), which are known as key regulators of cholesterol homoeostasis and as effector molecules of cellular innate immunity [59][60][61][62][63] were not affected by B. besnoiti infections. In contrast, E. bovis infections were recently proven to selectively upregulate the synthesis of 25-OHC, a molecule that was proven to bear antiparasitic properties 37 . Moreover, a lack of 7β-OHC, 7-ketoC and 7α-OHC upregulation, which is generally mediated via autoxidative processes, revealed that BUVEC do not experience considerable oxidative cell stress triggered by B. besnoiti infections.
In summary, this investigation adds further data on the modulatory capacity of the fast proliferating apicomplexan parasite B. besnoiti. The current data strengthen the assumption that the modulation of distinct pathways of cholesterol acquisition is dependent on both, the parasite species and the host cell type. Thus, we here show that successful B. besnoiti infections in primary bovine endothelial host cells rely on both, endogenous cholesterol synthesis and on sterol uptake from exogenous sources by using different LDL species leading to selective LOX-1 upregulation.

Cholesterol visualization and quantification. For staining intracellular stages for cholesterol, BUVEC
were grown on coverslips (n = 3) and infected with B. besnoiti tachyzoites (MOI = 5:1). At 24 h p. i. (hours post infection), the samples were washed with PBS, fixed in 4% paraformaldehyde (10 min), washed three times with PBS. To detect free cholesterol, the samples were stained by filipin III (35 µg/mL in PBS, 15 min, in the dark, RT; Sigma-Aldrich). All samples were washed with PBS, mounted in Fluoromount-G mounting medium (Invitrogen). Cells were analysed using an inverted fluorescence microscope (IX81, Olympus) applying the UV filter set (340-380 nm excitation, 430 nm pass filter) and/or by using confocal microscopy analysis (Confocal LSM 710; Zeiss; 63X magnification, numerical aperture 1.2 µm). Single cells (n = 20) fluorescence intensity measurements were performed using ImageJ 37 (mean grey value) and expressed as mean ± standard deviation. Image processing was carried out with ImageJ using merged channels plugins and restricted to minor adjustment of brightness and contrast.
Total cholesterol quantification was performed according to previous studies 27 . Therefore, total lipid extractions from B. besnoiti-infected (12, 24 and 48 h p. i.) and control BUVEC (n = 6) were executed in hexane:isopropanol 66 . The cells were washed twice with ice-cold PBS, trypsinized, washed again (400 × g, 10 min) and total cell numbers were counted using a Neubauer chamber. Hexane:isopropanol (3:2, v/v) was added to the cell pellet. Cells were disrupted for 10 min in Tissue Lyser (Qiagen) using stainless steel beads. After centrifugation (8,000 × g, 1 min) the supernatants were collected. The extraction was repeated once for each sample. www.nature.com/scientificreports www.nature.com/scientificreports/ fluorescence intensity measurements were performed using ImageJ 37 (mean grey value) and expressed as mean ± standard deviation. Image processing was carried out with ImageJ using merged channels plugins and restricted to minor adjustment of brightness and contrast.
For neutral lipid quantification, infected BUVEC layers (MOI = 3:1) were stained with Nile Red. The mean fluorescence intensities per area of single infected (n = 20) and non-infected single cells (n = 20) were estimated applying identical experimental conditions using an inverted fluorescence microscope (IX81, Olympus). Fluorescence intensity measurements were performed using ImageJ 37 (mean grey value) and expressed as mean ± standard deviation. For lipid droplet quantification, B. besnoiti-infected (MOI = 3:1 and 4:1) and control BUVEC (n = 3) were trypsinized at 12 and 24 h p. i. and pelleted (400 × g, 3 min, 4 °C). Resuspended cells were stained with Bodipy 493/503 (10 min, on ice) and washed twice with 1 ml PBS (400 × g, 3 min, 4 °C). The cells were transferred to 5-ml FACS tubes (BD Biosciences) containing 200 µl PBS and were processed in a FACSCalibur flow cytometer (BD Biosciences) by laser excitation at 488 nm (FL1-H channel). Flow cytometry data were acquired by the BD Cell Quest Pro software as previously reported (Hamid et al. 27 ). To estimate the effect of supplementations on parasite proliferation, tachyzoite numbers were estimated 48 h p. i. by quantitative PCR.

Lipid droplet enrichment. To artificially enhance lipid droplet formation in host cells, oleic acid (OA;
Sigma-Aldrich) was supplemented in BSA formulation complexes to the cell culture medium according to Martin and Parton 68 . Direct conjugation was performed by mixing oleic acid-free BSA (fraction V, Roth) with oleic acid at the molar ratio of 6:1 (oleic acid:BSA). Prior to infection, an induction step was performed with culture medium suplemented with 50 µM OA (1 h, 37 °C, 5% CO 2 ). Parasites were allowed to infect BUVEC in non suplemented cell culture medium for 4 h. Then, medium was changed and OA was suplemented in a final concentration of 2.5 µM.
PCR-based quantification of B. besnoiti tachyzoites. Tachyzoite numbers were estimated via a quantitative PCR according to Cortes, et al. 69 . Therefore, biological triplicates with technical duplicates were processed. Firstly, cell culture supernatants containing free-released tachyzoites (=extracellular tachyzoites) were collected at 48 h p. i. and pelleted (600 × g, 15 min). In addition, the remaining host cells carrying not yet released tachyzoites (=intracellular tachyzoites) were trypsinized and pelleted (600 × g, 15 min). All cell/parasite pellets were treated with 200 μL of cell lysis buffer containing 0.32 M sucrose, 1% Triton X-100, 0.01 M Tris-HCl (pH 7.5), 5 mM MgCl 2 and incubated in 100 μL 1X PCR buffer (Quanta) containing 20 μL proteinase K (20 mg/mL; Qiagen) at 56 °C for 1 h. Proteinase K was heat-inactivated by heating the samples (95 °C, 10 min) and the DNA-containing samples were frozen at −20 °C until further use. Real-time PCR was performed in a total volume of 20 μL containing 2 μL DNA of test samples, 400 nM of each primer, 200 nM probe and 10 μL PerfeCTa MasterMix (Quanta) at the following cycling conditions: 95 °C for 10 min; 40 cycles at 95 °C for 10 s, 60 °C for 15 s and 72 °C for 30 s.

Biochemical estimation of cholesterol esterification and of cholesterol-related sterols.
Pellets of B. besnoiti-infected BUVEC and of non-infected control cells were dried in a Savant SpeedVac concentrator (Thermo Fisher Scientific) for 24 h. Cholesterol, non-cholesterol sterols and oxysterols were extracted from dry weight aliquots using Folch reagent (chloroform/methanol; 2:1 (v/v); with 0.25 mg BHT added per mL solvent) per 10 mg dried cell pellets. Extraction was performed for 12 h at 4 °C in a dark cold room. The extracts were kept at −20 °C until further use. One mL of the Folch was submitted to alkaline hydrolysis, extraction of the free sterols and oxysterols, silylation to their corresponding (di)trimethylsilyl ethers prior to gas chromatographic separation and detection by mass selective detection (for non-cholesterol sterols or oxysterols using epicoprostanol and the corresponding deuterium labelled oxysterols as internal standards, respectively) as described in detail previously 70,71 . The degree of esterification of cholesterol was calculated from total (alkaline hydrolysis) and free cholesterol (without alkaline hydrolysis) concentrations using D6-cholesterol as internal standard. besnoiti-infected and control BUVEC were equally processed for total RNA isolation at different time points of parasite proliferation (3, 6, 12, 24 h p. i.). For total RNA isolation, the RNeasy kit (Qiagen) was used according to manufacturer's instructions. Therefore, BUVEC were lysed within the cell culture flasks with RLT lysis buffer (600 µL/25 cm 2 flask) and processed as proposed by the manufacturer including an on-column DNase treatment. Total RNAs were stored at −20 °C until further use. The quality of total RNA samples was controlled on 1% agarose gels. In order to remove any genomic DNA leftover, a second DNA digestion step was performed. Therefore, 1 µg of total RNA was treated with 10 U DNase I (Thermo Scientific) in 1x DNase reaction buffer (37 °C, 1 h). DNase was inactivated by heating the samples (65 °C, 10 min). The efficiency of genomic DNA digestion was verified by including no-RT-controls in each RT-qPCR experiment. cDNA synthesis was performed using the SuperScript III First-Strand Synthesis System (Thermo Fisher Scientific) according to manufacturer's instructions with slight modifications. 1 μg of DNase-treated total RNA was added to 0.5 µL of 50 μM oligo d(T), 1 µL of 50 ng/μl random hexamer primer, 1 µL of 10 mM dNTP mix and DEPC-treated water was adjusted to 10 µL total volume. The samples were incubated at 65 °C for 5 min and then immediately cooled on ice. For first strand cDNA synthesis, 2 µL of 10x RT buffer, 4 µL 25 mM MgCl 2, 2 µL 0.1 M DTT, 1 µL RNaseOUT (40 U/µL, Thermo Fisher Scientific), 0.5 µL SuperScript III enzyme (200 U/µL) and 0.5 µL DEPC-treated water were added and the samples were incubated at 25 °C for 10 min followed by 50 °C for 60 min and a 85 °C-inactivation step for 15 min.
Primers (MWG Biotech) and probes used for qPCR are shown in Table 2 27 . Probes were labelled at the 5′-end with a reporter dye FAM (6-carboxyfluorescein) and at the 3′-end with the quencher dye TAMRA (6-carboxytetramethyl-rhodamine). qPCR amplification was performed on a Rotor-Gene Q Thermocycler (Qiagen) in duplicates in a 10 µL total volume containing 400 nM forward and reverse primers, 200 nM probe, 10 ng cDNA and 5 µL 2x PerfeCTa qPCR FastMix (Quanta Biosciences). The reaction conditions for all systems were as follows: 95 °C for 10 min, 40 cycles at 95 °C for 10 s, 60 °C for 15 s and 72 °C for 30 s. No-template controls (NTC) and no-RT reactions were included in each experiment. As reference gene GAPDH was used as previously reported 27,65,73 . Analyses of the qPCR data used the comparative ΔΔC T method 74 and reported as n-fold differences comparing B. besnoiti-infected BUVEC with non-infected controls after normalizing the samples by the GAPDH reference gene. protein extraction. For protein extraction, B. besnoitia-infected and non-infected BUVEC isolates (n = 3) were washed in PBS to remove any medium traces, trypsinized and pelleted (600 × g, 10 min). Proteins were extracted by homogenizing the cell pellets in RIPA buffer [50 mM Tris-HCl, pH 7.4; 1% NP-40; 0.5% Na-deoxycholate; 0.1% SDS; 150 mM NaCl; 2 mM EDTA; 50 mM NaF (all Roth)] in the presence of a protease inhibitor cocktail (Sigma-Aldrich). The homogenates were centrifuged at 10,000 × g for 10 min at 4 °C to sediment intact cells and nuclei. The supernatants were stored at −80 °C until further use. The protein content was quantified via Coomassie Plus (Bradford) Assay Kit (Thermo Scientific) following the manufacturer instructions.
sDs-pAGe and Western Blotting. The samples were denatured using 6 M urea loading buffer and heated for 5 min at 95 °C. Fifty µg of total protein were loaded per slot and run in 12% polyacrylamide gels (120 V, 1.5 h) 75 . After electrophoretic separation, the proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore) (200 mA, 2 h). Blots were blocked with 3% BSA in TBS containing 0.1% Tween (Sigma) for 1 h at RT and then incubated overnight in primary antibody solutions (see Table 3). Vinculin expression was used as a reference for the normalization of the samples. After washing in TBS containing 0.1% Tween (thrice,  www.nature.com/scientificreports www.nature.com/scientificreports/ 5 min), blots were incubated in secondary antibody solutions (30 min, RT). Signal development was accomplished with an enhanced chemiluminescence detection system (ECL plus kit, GE Healthcare) and signal strength was determined in a ChemoCam Imager (Intas Science Imaging). Protein sizes were controlled by a protein ladder (PageRuler Plus Prestained Protein Ladder ~10-250 kDa, Thermo Fisher Scientific).
Additionally, the expression of LOX-1 in B. besnoiti-infected BUVEC and non-infected control cells was determined by a commercially available bovine ELISA kit (DL-Develop). Therefore, BUVEC (n = 3) were grown to subconfluency in 75 cm 2 cell culture flasks (Greiner) and infected at an MOI of 5:1 with freshly collected B. besnoiti tachyzoites. Cells were harvested at different time points p. i. (12,24,36 h p. i.) according to manufacturer's instructions. Briefly, after washing with PBS, the cells were trypsinized and pelleted (400 × g, 12 min). Cell pellets were washed thrice with PBS and the number of cells per sample was determined microscopically before ultrasonication treatment (3 times for 20 s, on ice). Samples were centrifuged at 1,000 × g (15 min, 2-8 °C) to remove cell debris and the supernatants were stored at −20 °C until being processed by the ELISA kit.
Live cell 3D holotomographic microscopy. BUVEC (n = 3) were seeded into 35 mm tissue culture µ-dishes (Ibidi ® ), grown overnight and infected with freshly released B. besnoiti tachyzoites (MOI 3:1). At 24 h p. i., holotomographic images were obtained by using 3D cell-explorer microscope (Nanolive 3D) equipped with a 60x magnification (λ = 520 nm, sample exposure 0.2 mW/mm 2 ) and a depth of field of 30 µm 76 . For lipid droplet visualization, cells were stained with Bodipy 493/503 (1 µg/mL, Life Technologies; 3 h, 37 °C, in the dark). After incubation, medium was changed. Live cell 3D holotomographic microscopy and analysis of Bodipy 493/503-based fluorescence was performed in parallel to prove the identity of lipid droplets. A total of 50 lipid droplets were measured for their refractive index to obtain marker values for these organelles in BUVEC. Images were analysed using STEVE software (Nanolive) to obtain a refractive index-based z-stack 77 and digital staining was applied according to the refractive index of the lipid droplets. statistical analysis. Statistical analyses were performed with the statistical program package BMDP 78 or with GraphPad Prism. In all cases, data revealed as normally or log-normally distributed (verified by residual analysis) and parametric statistical methods could be applied. Data description was performed by presenting arithmetic mean ± standard deviation for not log-transformed data, by geometric mean and standard deviation for log-transformed data or as n-fold changes relative to the controls. Depending on the design of the experiment, some data were analysed by one-or two-way analysis of variance (ANOVA) with repeated measures (program BMDP2V) to test the effects of infection and/or incubation time, dose of the inhibitor or cell ratio. If only one time point was considered, the analysis could be reduced to a t-test for dependent samples (program BMDP3D). In the case of a hierarchical design of the experiments incorporating more than one random factor (e. g. BUVEC and replication) a general mixed model analysis (glmm) with equal sample size (program BMDP8V) was applied. Occasionally, post hoc pairwise multiple comparison tests succeeded the global comparison by ANOVA, being performed either by Student-Newman-Keuls method (SNK-test) or by Bonferroni-Holm method controlling the family-wise error rate, or even byt-test. The outcomes of the statistical tests were considered to indicate significant differences when p ≤ 0.05 (significance level).  Table 3. List of antibodies used for Western blot and surface expression assays.