Killer yeasts exert anti-plasmodial activities against the malaria parasite Plasmodium berghei in the vector mosquito Anopheles stephensi and in mice

Background Wickerhamomyces anomalus is a yeast associated with different insects including mosquitoes, where it is proposed to be involved in symbiotic relationships with hosts. Different symbiotic strains of W. anomalus display a killer phenotype mediated by protein toxins with broad-spectrum antimicrobial activities. In particular, a killer toxin purified from a W. anomalus strain (WaF17.12), previously isolated from the malaria vector mosquito Anopheles stephensi, has shown strong in vitro anti-plasmodial activity against early sporogonic stages of the murine malaria parasite Plasmodium berghei. Results Here, we provide evidence that WaF17.12 cultures, properly stimulated to induce the expression of the killer toxin, can directly affect in vitro P. berghei early sporogonic stages, causing membrane damage and parasite death. Moreover, we demonstrated by in vivo studies that mosquito dietary supplementation with activated WaF17.12 cells interfere with ookinete development in the midgut of An. stephensi. Besides the anti-sporogonic action of WaF17.12, an inhibitory effect of purified WaF17.12-killer toxin was observed on erythrocytic stages of P. berghei, with a consequent reduction of parasitaemia in mice. The preliminary safety tests on murine cell lines showed no side effects. Conclusions Our findings demonstrate the anti-plasmodial activity of WaF17.12 against different developmental stages of P. berghei. New studies on P. falciparum are needed to evaluate the use of killer yeasts as innovative tools in the symbiotic control of malaria. Electronic supplementary material The online version of this article (10.1186/s13071-019-3587-4) contains supplementary material, which is available to authorized users.


Background
Malaria is a mosquito-borne disease that kills about half a million people each year, predominantly children under five years of age. Plasmodium, the etiological agent of malaria, is transmitted to humans (vertebrate host) through the bite of a female Anopheles mosquito (vector and invertebrate host). Efforts to reduce morbidity and mortality have been impaired by drug and pesticide resistance and the lack of effective vaccines. To ameliorate the efficacy of malaria control efforts, innovative strategies are urgently needed [1,2].
Plasmodium has a complex life-cycle including sexual and asexual developmental stages. It alternates a sporogonic phase in the mosquito with differentiation of zygotes, ookinetes, oocysts and sporozoites, and hepatic/erythrocytic phases in the vertebrate host where it multiplies different asexual stages, micro-and macro-gametocytes. This multifaceted phenotypic variability is advantageous for the parasite but hampers

Open Access
Parasites & Vectors *Correspondence: irene.ricci@unicam.it 1 School of Biosciences and Veterinary Medicine, University of Camerino, Camerino, Italy Full list of author information is available at the end of the article the development of new therapeutic drugs and/or prophylaxis, thus facilitating the spread of malaria.
In recent years, symbiotic control (SC) has been proposed as a possible strategy for preventing malaria [3]. SC is intended to block malaria transmission by reducing anopheline vector competence through symbiotic microbes that affect the development of the parasite in the mosquito midgut [4,5]. Some bacteria, including Asaia, Wolbachia and Pantoea, and fungi such as Metarhizium, have been proposed for the SC of malaria [3,[6][7][8]. In this frame, the identification of killer yeasts that exert symbiotic functions in mosquitoes is very promising [9,10].
Among the different isolated symbiotic killer strains, Wickerhamomyces anomalus offers a series of highly appealing features [11]. In fact, W. anomalus displays a wide antimicrobial property mediated by killer toxins (KTs) that exert an enzymatic activity targeting the cell-wall glucan components of bacteria, yeasts and protozoa [12]. Interestingly, W. anomalus is considered safe [13] and some environmental strains are used against spoilage microbes for food and beverage biopreservation [14,15].
Symbiotic strains of W. anomalus producing KTs have been identified in beetles, mosquitoes and sand flies where they are proposed to be involved in nutritional and/or protective functions [16][17][18]. In particular, the W. anomalus strain (WaF17.12) isolated from the malaria vector Anopheles stephensi, produces a KT (WaF17.12-KT) that has shown, as a purified product, a strong anti-plasmodial activity in vitro against sporogonic stages of the murine malaria parasite Plasmodium berghei, causing a surface membrane damage mediated by a β-glucanase activity [19]. WaF17.12-KT seems to be poorly expressed in the mosquito, although its production can be stimulated by selective media conditions, and yeast cultures can be reintroduced into the mosquito through diet [20]. Generally, the killer phenotype is activated or enhanced under stress conditions, as the competition with other microorganisms for environmental resources, whereas the host body is a niche favorable to the proliferation of yeast with good availability of food [21].
In the present study, in vitro and in vivo studies have been carried out to unveil interactions between the killer strain WaF17.12 and the murine malaria parasite P. berghei. The results demonstrated that the yeast affects different stages of the parasite development by WaF17.12-KT mediated mechanisms, both in mosquitoes and mice (vertebrate host). These outcomes validate WaF17.12 as an effective tool against P. berghei.

Yeasts and killer toxin production
Two W. anomalus (Wa) strains were used in the present study: WaF17.12, isolated from An. stephensi mosquitoes (KT-producer) [17]; and WaUM3, environmental strain (KT non-producer) [22]. Both strains were cultured in selective growth conditions to stimulate KT production: YPD broth (20 g/l peptone, 20 g/l glucose, 10 g/l yeast extract), buffered at 4.5 pH with 0.1 M citric acid and 0.2 M K 2 HPO 4 [23]. Cells were incubated at 26 °C for 36 h at 70× rpm. To verify the activation of WaF17.12 and confirm KT expression, the supernatants of both yeast cultures were separated by chromatography and the obtained fractions analysed by western blot using the monoclonal mAbKT4 antibody specific for yeast KTs, as described in Valzano et al. [19] and Cappelli et al. [20]. The positive fraction obtained from WaF17.12 was named as WaF17.12-KT + and the corresponding fraction isolated from WaUM3 (WaF17.12-KT -) was used as KT negative control.

Malaria parasites
Two P. berghei (murine malaria parasite) transgenic strains were used in the present work: PbCTRPp.GFP that expresses green fluorescent protein (GFP) during the early sporogonic phase of the parasite cycle [24], and PbGFP CON that expresses GFP throughout the whole parasite cycles both in the mosquito and vertebrate host [25]. The two parasites were differently used, depending on the type of assay. PbCTRPp.GFP was used for cultivating early sporogonic stages in vitro, whereas PbGFP CON, which is detectable during all the developmental stages, was used during in vivo experimentations. Both parasite strains were visualized using a fluorescence microscope (Axio Observer.Z1; Carl Zeiss, Milan, Italy).

Mice
BALB/c mice were maintained in the Animal Facilities of the University of Camerino at 24 °C, fed on standard laboratory mice pellets (Mucedola S.r.l., Milano, Italy) and provided with tap water ad libitum. Mice were used as P. berghei vertebrate hosts. Eight-week-old female mice (18-25 g) were infected with PbCTRPp.GFP or PbGFP-CON directly by an intraperitoneal injection of 10% parasitaemic blood obtained from the tail of a donor mouse (acyclic passage). About 10 6 infected erythrocytes of the donor mouse were diluted in 200 μl of PBS (7.2 pH) and parasitaemia in injected mice was constantly monitored by optical microscope using a 10% Giemsa stained blood smear and fluorescence microscope (Axio Observer.Z1). Recipient mice were used for experimental purposes when showing 5% parasitaemia. All animal rearing and handling was carried out according to the Italian Legislative Decree (116 of 10/27/92) on the "use and protection of laboratory animals" and in agreement with the European Directive 2010/63/UE. The experimentation was approved by the Ethical Committee of University of Camerino.

Mosquitoes
Anopheles stephensi (Liston strain) mosquitoes were reared at 29 °C and 85-90% relative humidity with a 12:12 light-dark photoperiod in the insectarium of the University of Camerino. Immediately after emergence of adult mosquitoes, three cages, each containing 250 females, were set up for two trials. Newly emerged mosquitoes were fed ad libitum on a cotton pad soaked with 5% (w/v) sterile sucrose solution (control group), or 5% sucrose solution plus WaF17.12, or 5% sucrose solution plus WaUM3.
Killing activity assay of WaF17.12 against P. bergheiPb CTRPp.GFP sporogonic stages in vitro Activated WaF17. 12 and WaUM3 cultures were pelleted at 3000× g for 10 min and suspended in PBS (7.4 pH), whereas the supernatants were processed as described above. PbCTRPp.GFP cultures were obtained as follows: early sporogonic stages were developed in vitro incubating gametocytemic blood (20 µl) from a donor mouse showing 5% parasitaemia with ookinete-medium (180 µl) on microtiter plates. Ookinete-medium was prepared as described in Valzano et al. [19]. For the killing activity assay, parasite cultures were incubated with activated WaF17.12 (20 μl) or WaUM3 (20 μl) (yeasts final concentration 10 8 cell/ml) for 24 h at 19 °C in the dark (20 μl of PBS 7.4 pH was added to control wells). During incubation, gametocytes develop into fluorescent zygotes and ookinetes. The parasite number in each well was evaluated using a fluorescence microscope with a 40× objective (Axio Observer.Z1). Experiments were performed in triplicate.
WaUM3-KT − was tested to evaluate possible effects due to other co-purified molecules, whereas control wells were prepared by adding PBS (7.4 pH). After incubation, parasites were treated with 20 µg/ml of the red-fluorescent dye propidium iodide (PI) (Sigma-Aldrich, Saint Louis, USA) in DNAse-free PBS, for 30 min at 20 °C, and visualized by fluorescence microscope (Axio Observer. Z1; Carl Zeiss, Milan, Italy). PI staining is used for identifying dead or damaged cells, because the dye penetrates cells with altered membranes where it intercalates DNA double strand, but it is effectively excluded from viable cells.
Killing activity assay of WaF17.12 against P. berghei PbGFP CON sporogonic stages in vivo Three cages, each containing newly emerged female mosquitoes (n = 250), were set up with a cotton pad soaked with 5% sterile sucrose solution (control group), 5% sucrose solution plus WaF17.12 or 5% sucrose solution plus WaUM3. Yeasts were grown under stimulating conditions to produce KT as described above. Yeast cultures were centrifuged at 3000× g for 10 min at 4 °C, washed three times in 0.9% (w/v) NaCl solution, and suspended in 5% sucrose solution at a final concentration of 10 8 cells/ml. Food preparations were provided ad libitum to mosquitoes and the cotton pad was refreshed every two days for the duration of the experiment (21 days). To verify the presence of activated yeasts in the mosquito gut, an immunofluorescent assay (IFA) assay at 24 h, and 10 and 21 days post-infection was performed using a monoclonal anti-KT antibody (mAbKT4) obtained at a concentration of 2 mg/ml, as described by Cappelli et al. [20]. After six days, mosquitoes had a blood meal on a 5% parasitaemic mouse infected with PbGFP CON . Unfed mosquitoes were removed and cages were laid in a chamber at 19 °C and 95 ± 5% relative humidity for P. berghei development in the mosquito. Twenty-four hours after the blood meal (before ookinetes migrate across the midgut epithelium [26]), mosquito guts (n = 24 per cage) were dissected and individually homogenised in sterile PBS. Gut preparations were analysed through fluorescence microscopy using a 100× objective (Axio Observer. Z1) to detect PbGFP CON early sporogonic stages. The number of oocysts and sporozoites was evaluated, analysing guts and salivary glands (n = 10 per cage) 10 and 21 days post-infection, respectively. Experiments were performed twice.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 5 software. The differences between the groups in each experiment were compared using a Mann-Whitney U-test. A P-value < 0.05 was considered statistically significant.

WaF17.12 affects sporogonic stages of P. berghei in vitro and in vivo
The anti-plasmodial activity of WaF17. 12-KT has been demonstrated in vitro against cultivated P. berghei sporogonic stages, but any direct effect of such KT-producing yeasts on the parasite development still required validation. Therefore, in the present study, in vitro and in vivo experiments were carried out to evaluate direct interactions between WaF17.12 and the murine malaria parasite P. berghei.
During in vitro tests, the number of developed parasites (zygotes and ookinetes) was counted and the mean values ± SEM from triplicate experiments are reported for each group (Fig. 1a). The treatment with WaF17.12 decreased the number of parasites by approximately 37.1%, whereas no significant differences between WaUM3 and control were detected. Such data suggests that the effect of W. anomalus on P. berghei sporogonic stages development was strain-dependent and attributable to the presence of the KT in the medium. To assess the activation of the WaF17.12 strain and to confirm KT expression, the supernatants from both strains were separated by ion-exchange chromatography and obtained fractions were analysed by western blot as described in the material and methods section. The KT containing fraction (WaF17.12-KT + ) and the corresponding KT − (WaUM3-KT − ) were collected and tested against cultured early P. berghei sporogonic stages (Fig. 1b). A cell viability assay was performed using the propidium iodide (PI) dye, to evaluate the effect of individual purified fractions on parasite development. Treatment with WaF17.12-KT + induced an alteration of ookinetes morphology, characterized by bigger and more jagged shapes, and PI penetration in damaged and non-viable parasites. Conversely, WaUM3-KT − did not affect ookinetes, which showed a normal shape and no PI accumulation, similar to the control group, demonstrating that the anti-plasmodial activity of the WaF17.12-KT + fraction was specific and not due to the presence of other co-purified molecules.
These findings prompted additional in vivo studies on the evaluation of the WaF17.12 strain anti-plasmodial activity in An. stephensi. The anti-plasmodial activity results confirmed the ability of WaF17.12 to abundantly colonize the An. stephensi gut, whereas no KT signal was detected in mosquitoes treated with WaUM3 and in the control group. PbGFP CON early sporogonic stages in the midgut lumen of mosquitoes colonized with activated yeast cultures and in control groups, were counted 24 h post-infection and the median values (m) obtained from two independent experiments are reported (Fig. 2). Mosquitoes colonized with WaF17.12 developed 65.2% fewer parasites than the control group (m = 710 and m = 247, respectively) and 43.9% fewer parasites than the WaUM3 group (m = 440). Differently from the in vitro analysis, a significant decrease in parasite number of 38.0% with respect to the control was also detected in mosquitoes colonized with WaUM3, indicating a non-KT related anti-plasmodial activity of this strain.
Further analyses were then performed to evaluate possible inhibition effects of P. berghei (PbGFP CON ) in the later stages of the sporogonic cycle. Specifically, guts and salivary glands of 10 mosquitoes for each group were analysed at 10 and 21 days post-infection, respectively. The number of oocysts and sporozoites was not significantly different in the three groups (data not shown). Nevertheless, the huge number of oocysts developing during experimental infections could have impaired the detection of anti-plasmodial effects in the later sporogonic stages.

WaF17.12-KT + anti-plasmodial activity against P. berghei PbGFP CON erythrocytic stages and preliminary safety tests
PbGFP CON infected donor mice were used to test the effects of WaF17.12-KT on erythrocytic stages. Infected blood samples were individually incubated for 90 min at 25 °C with WaF17.12-KT + (100 µg/ml KT), or WaF17.12-KT + (100 µg/ml KT) plus the specific anti-KT antibody mAbKT4 (100 µg/ml) [22], or PBS at 7.2 pH (control). The resulting mixtures were inoculated in healthy mice (n = 5 for each group), and parasitaemia was evaluated five days post-infection. The mean values ± SEM of two independent experiments are reported for each group (Fig. 3). A significant reduction in the parasitemia of both WaF17.12-KT + (55.4%) and WaF17.12-KT + plus Ab (40.1%) groups was detected indicating that WaF17.12-KT + interferes with PbGFP CON erythrocytic stages and impairs the ability of the parasite to infect recipient mice. Interestingly, treatment with the monoclonal antibody mAbKT4 in part inhibited the effects of WaF17.12-KT + showing a reduction percentage that is lower with respect to the KT group. It is reasonable that a saturating concentration of the antibody could cause a complete block of WaF17.12-KT activity.
To evaluate the potential therapeutically use of WaF17.12-KT, a preliminary test was performed using the murine cell line HEPA 1-6. A viability assay was carried out through a colorimetric MTT test. Cells were incubated with different doses of WaF17.12-KT + for 24 h (Fig. 4). Our data demonstrated that the treatment slightly affects cell viability only at the highest doses (60, 80 and 100 µg/ml KT), suggesting that KT can be considered safe on vertebrate cells. We carried out additional analysis using PI on HEPA 1-6 cells treated and not treated with KT, showing no cell damage at a dose of 100 μg/ml at 24 h of treatment (Additional file 1: Figure S1).

Discussion
SC strategies propose the use of symbiotic microorganisms that interfere with the vector's capability of blocking the transmission of malaria. In recent years, several bacteria and fungi have been indicated as good candidates for SC applications against malaria since they can inhibit the development of the parasite in wild mosquitoes. Direct or indirect anti-plasmodial mechanisms can be exerted by wild-type or recombinant microbial strains. In this context, the use of natural killer strains limits the introduction of genetically manipulated microorganisms into the environment. This aspect makes symbiotic strains of W. anomalus highly appealing for SC strategies, particularly after the discovery of their ability to produce KTs with anti-parasitic effects in the Anopheles midgut. In the present study, we demonstrated that a diet supplemented with activated WaF17.12 cultures inhibits the development of the early sporogonic stages of P. berghei in An. stephensi (− 65.21%). Interestingly, the KT non-producer strain WaUM3 also affected parasites (− 38.02%), likely by host immune system stimulation due to the massive presence of the yeast in the gut [28]. Nevertheless, the significantly higher effect of WaF17.12 highlighted a direct mechanism due to the action of KT. We have used only the KT non-producer strain WaUM3 as a negative control since a non-stimulated WaF17.12 control group is not obtainable. In fact, it is possible to induce higher or lower KT production by cultural methods using growth media at different pH values but not KT free cultures [20].
These outcomes are consistent with the demonstration that, in vitro, only WaF17.12 is able to prevent ookinete development (− 40%), supporting the hypothesis that the surface membrane damage is associated to WaF17.12-KT β-glucanase activity, as previously reported by Valzano et al. [19].
Although in vitro and in vivo inhibition rates cannot be compared because of substantial differences between the two experimental systems, we could speculate that the interaction of WaF17.12-KT with the parasite cellwall carbohydrates may be more efficient in the midgut lumen. We also investigated the development of later sporogonic stages in the mosquito fed with WaF17.12, finding that there is no antagonist effect on oocysts and sporozoites. In fact, Cappelli et al. [20] reported that WaF17.12-KT is released in the midgut lumen surrounding the epithelium, thus oocysts and sporozoites, differently from ookinetes, are not exposed to the toxin. Additionally, the 100-fold higher number of oocysts produced upon experimental infections with respect to natural conditions, could mask the anti-plasmodial effect of killer yeasts at the level of later sporogonic stages. In fact, the considered murine malaria system (P. berghei/An. stephensi/BALBc) is an experimental model optimized to have abundant infections in mosquitoes [29,30]. To address this issue, additional tests using killer yeasts against human malaria parasites in naturally infected mosquitoes that develop no more than 5-10 oocysts [31], could be advisable. In fact, under these conditions WaF17.12-KT could break down the vector capacity. Interestingly, yeasts are already used in traps because the smell produced by their fermentation is able to attract mosquitoes [32]. Thus, sugary preparations supplemented with WaF17.12 cultures could be placed in traps to collect wild mosquitoes to be used for testing the anti-plasmodial effects on human malaria parasites. Further studies will assess symbiotic killer yeasts of anophelines as new candidates for innovative and safe SC-strategies of malaria.
The present study also provides evidence for the antiplasmodial activity of WaF17.12-KT on erythrocytic stages, as shown by the KT-mediated reduction of parasitatemia observed in experimentally infected mice.
Interestingly, WaF17.12-KT was also active against parasite stages that develop in the vertebrate host, likely according to an antimicrobial mechanism of action based on glucanase activity. In fact, β-1,3-glucans are microbial cell wall components in charge of maintaining cell morphology and osmotic integrity [33]. Nevertheless, KT action could involve also different pathways, as a recent study reported the apoptotic activity of peptides mimicking KTs against the protozoan parasite Toxoplasma gondii [34]. However, further investigations on human malaria parasites are requested to assess the mechanism of action of KTs.
Regardless of the mechanisms of action, tests against a murine cell line showed that WaF17.12-KT does not affect vertebrate cells.

Conclusions
Overall, these results demonstrate the anti-plasmodial activity of WaF17.12 on P. berghei sporogonic stages, showing that killer yeasts can potentially be used as safe and innovative tools against malaria. WaF17.12 could be applicable for symbiotic control in the field, taking advantage of the fact that killer yeasts are easily releasable in mosquito feeding sites. Furthermore, the effect of WaF17.12-KT against the erythrocytic stages suggest that yeast killer toxins can be used as novel antiparasitic drugs.

Additional file
Additional file 1: Figure S1. KT treatment does not induce cell damage. PI staining on HEPA 1-6 cells treated and not treated with KT (100 μg/ ml) analysed by cytofluorimeter. Abbreviation: MFI, mean fluorescence intensity.