Evaluation of anti-malaria potency of wild and genetically modified Enterobacter cloacae expressing effector proteins in Anopheles stephensi

Malaria is one of the most lethal infectious diseases in tropical and subtropical areas of the world. Paratransgenesis using symbiotic bacteria offers a sustainable and environmentally friendly strategy to combat this disease. In the study reported here, we evaluated the disruption of malaria transmission in the Anopheles stephensi-Plasmodium berghei assemblage using the wild-type (WT) and three modified strains of the insect gut bacterium, Enterobacter cloacae. The assay was carried out using the E. cloacae dissolvens WT and three engineered strains (expressing green fluorescent protein-defensin (GFP-D), scorpine-HasA (S-HasA) and HasA only, respectively). Cotton wool soaked in a solution of 5% (wt/vol) fructose + red dye (1/50 ml) laced with one of the bacterial strains (1 × 109cells/ml) was placed overnight in cages containing female An. stephensi mosquitoes (age: 3–5 days). Each group of sugar-fed mosquitoes was then starved for 4–6 h, following which time they were allowed to blood-feed on P. berghei–infected mice for 20 min in the dark at 17–20 °C. The blood-fed mosquitoes were kept at 19 ± 1 °C and 80 ± 5% relative humidity, and parasite infection was measured by midgut dissection and oocyst counting 10 days post-infection (dpi). Exposure to both WT and genetically modified E. cloacae dissolvens strains significantly (P < 0.0001) disrupted P. berghei development in the midgut of An. stephensi, in comparison with the control group. The mean parasite inhibition of E. cloacaeWT, E. cloacaeHasA, E. cloacaeS−HasA and E. cloacaeGFP−D was measured as 72, 86, 92.5 and 92.8 respectively. The WT and modified strains of E. cloacae have the potential to abolish oocyst development by providing a physical barrier or through the excretion of intrinsic effector molecules. These findings reinforce the case for the use of either WT or genetically modified strains of E. cloacae bacteria as a powerful tool to combat malaria.

disease. Anopheles stephensi is the main malaria vector throughout its range from Asia to the Horn of Africa [2][3][4][5]. Therefore, programs aimed at controlling A. stephensi populations and limiting the ability of these mosquitoes to transmit Plasmodium (refractory mosquitoes) have the potential to reduce the burden of malaria disease [6].
Currently, the most common methods of mosquito control are indoor residual spraying (IRS) and insecticide-treated nets (ITNs) [1,7]. However, novel and innovative control measures are urgently needed due to emerging insecticide resistance in malaria vectors, particularly in A. stephensi [8][9][10], and increasing ecoenvironmental concerns about the off-target effects of insecticide use [11][12][13][14][15]. Transmission-blocking strategies (TBS) have recently been proposed as a potential means of malaria control, with an increased emphasis on inhibiting the development of the Plasmodium parasite in the vector mosquito [16][17][18][19]. Gametocytocidal drugs, transmission-blocking vaccines and the replacement of wild mosquitoes with refractory mosquitoes are currently the most important methods used in TBS. The last method consists of genetically manipulating Anopheles mosquitoes to render them refractory to Plasmodium parasite development. This is accomplished using anti-plasmodium molecules (transgenesis) [20], naturally refractory mosquitoes [21], artificial gene-drive mechanisms [22,23] and/or micro-symbionts genetically modified by effector molecules that are reintroduced into the wild mosquito population (paratransgenesis) [6,12,24,25].
During its development in its invertebrate hosts (vectors), the Plasmodium parasite undergoes a decreasing population trend, from 10 3 -10 4 gametocytes to 10 2 -10 3 motile ookinetes and, finally, to ≤ 5 oocysts [26,27]. This bottleneck could be considered a prime target for intervention and the blocking of parasite transmission [28,29]. The main factors creating the bottleneck include gut digestive enzymes, the mosquito's immune responses and intestinal microbial flora. The intestinal microbial flora plays a vital role in blocking parasite development in the Anopheles midgut. This effect is exerted directly by the proliferation of bacteria after a blood meal, simultaneously with the development of the ookinete stage, and indirectly via the expression of antimicrobial genes [30][31][32][33][34][35][36].
To date, a number of different symbiotic bacteria have been suggested for use in a paratransgenesis strategy for combating malaria. Serratia AS1 (isolated from Anopheles spp.), Asaia sp. (isolated from Anopheles gambiae, An. stephensi, Aedes albopictus and Aedes aegypti) and Pantoea agglomerans (isolated from An. stephensi, An. gambiae and Anopheles funestus) are the bacterial species most commonly used for the paratransgenic control of malaria [11,25,27,31,[37][38][39][40][41]. These genetically modified bacteria could potentially eliminate the development of the Plasmodium parasite in the Anopheles midgut by expressing anti-Plasmodium molecules. However, due to convergent evolution, the wild-type (WT) bacteria have shown limited intrinsic antiparasitic activities in the mosquito midgut [27,42].
Enterobacter cloacae, a Gram-negative, facultative anaerobic, rod-shaped bacterium, has been found to be a component of the microflora of An. stephensi [43,44], Anopheles albimanus [45], Ae. albopictus and Ae. aegypti [46], as well as of other medically important insects [47,48]. This bacterium has been shown to have an innate blocking effect on Plasmodium development and could potentially limit P. berghei and P. falciparum development in An. stephensi by markedly increasing the population, thereby leading to stimulation of the mosquito immune system and expression of immune response compounds, such as serine protease inhibitors (SRPN6) [49]. These innate features suggest that E. cloacae is a suitable candidate for paratransgenesis studies against the malaria parasite.
Scorpine is a small cationic antimicrobial peptide (AMP) found in the venom of the scorpion Pandinus imperator with anti-plasmodial and anti-bacterial activity, and also a strong inhibition of dengue 2 virus (DENV-2) infection [50][51][52]. Defensins are small cysteine-rich cationic proteins that are found in plants, vertebrates and invertebrates [53][54][55]. Scorpine has been used as effector molecule against malaria parasites in paratransgenic mosquitoes carrying scorpine-secreting bacteria or fungi [11,27,56]. The aim of the present study was to evaluate the transmission-blocking potential of WT and engineered strains of E. cloacae expressing defensin and scorpine effector molecules to block P. berghei development in An. stephensi.

Mosquito rearing
Anopheles stephensi, Beech strain, was used in this study. The strain was originally collected in Pakistan as an additional type of the SDA500 strain and was kindly provided in 2005 by Professor P.F. Billingsley, Sanaria, Inc. [57]. Mosquitoes were maintained on a 5% (wt/vol) fructose solution at 27 ± 1 °C and 65 ± 5% relative humidity (RH) under a 12:12 dark:light (D:L) photoperiod. All mosquito-rearing facilities were provided by Tehran University of Medical Sciences, School of Public Health.

Maintenance of parasite life-cycle
BALB/c mice aged 6 weeks (weight 8-20 g) were used in this study. The mouse colonies were maintained in an animal house at 40-50% RH and 24 ± 1 °C. The P. berghei ANKA strain was used, specifically clone 2.34 (a gift from Prof. Marcelo Jacob-Lorena, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA). Parasites were maintained using previously described protocols [58,59]. Briefly, the parasites were maintained in female BALB/c mice by serial mechanical passages (3 or 4 passages). To maintain gametocyte infectivity to mosquitoes, in direct passages, hyper-reticulocytosis was induced 3 days before infection by treating each mouse with 100 μl 1% phenylhydrazinium chloride (PH) (Sigma-Aldrich, St. Louis, MO, USA) administered intraperitoneally (ip; 10 mg/ml in phosphate-buffered saline [PBS]) per 10 g mouse body weight. Parasitemia was monitored in Giemsa-stained tail-blood smears. Exflagellation was examined at 5-6 days post-infection (dpi) (dose of injected parasite: approx. 10 4 ) by mixing a drop of infected blood (4-5 µl) with 20-25 μl ookinete culture medium (RPMI 1640 medium containing L-glutamine and 25 mM HEPES, 2 g/l NaHCO, 50 mg/l hypoxanthine, 50,000 U/l penicillin and 50 mg/l streptomycin; pH 8.3, filter sterilized). Mice with approximately three exflagellation centers in each field of 40× microscopic magnification were used in the transmission blocking assay (Fig. 1).

Transformation of bacteria
The E. cloacae dissolvens subspecies used in this study was originally isolated by sampling the midgut microflora of the sand fly Phlebotomus papatasi, found in the main zoonotic cutaneous leishmaniasis foci in central Iran. The procedures used for isolating, characterizing and identifying this bacterium have been outlined previously [47]. We genetically engineered two different E. cloacae strains to produce: (1) a strain containing defensin (a peptide isolated from radish seeds, Rs-AFP [60]) plus green fluorescent protein (GFP), referred to here as E. cloacae GFP−D ), and (2) a strain containing Scorpine-HasA (scorpion Pandinus imperator venom) plus HasA, a heme-binding protein as an exporting system [61], referred to here as E. cloacae S−HasA . Defensin and scorpine proteins are anti-malarial effector molecules with different killing mechanisms. As a control, we also genetically engineered a strain of E. cloacae to produce only HasA, referred to here as E. cloacae HasA . The transgenic E. cloacae GFP−D strain with the originally manipulated pBR322 plasmid was used in this study. The engineered pBR322 plasmid containing the defensin gene as effector molecule plus a GFP marker and tetracycline resistance genes, called the pBR/DG plasmid, is maintained in Tehran University of Medical Sciences (Fig. 2). The β-lactamase gene of the plasmid was replaced with the DG construct. A detailed description of the construct is provided in Additional file 1: Dataset S1. The recombinant plasmid was electro-transformed first to the recipient strain of Escherichia coli-DH5α and then to E. cloacae dissolvens cells. This strain is distinguishable from other similar bacteria colonies its green fluorescence under microscopy ( Fig. 3). For construct verification, it was amplified from the pBR/ DG plasmid using the forward primer DGFP1 (5′-GGA ATT CAA ATA CAT TCA AAT ATG TAT CCG-3′) and the reverse primer DGFP7 (5′-TTC TGC AGT TAT TAT TTG TAT AGT TCA TCC ATG-3′). The pBR/DG and pBR322 plasmids also were digested at the EcoRI and Pst1 restriction sites which were incorporated at the 5′ and 3′ ends of the construct, respectively. The digested plasmids were electrophoresed in 1% agarose gel (Additional file 2: Fig. S1).
The proteins were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) to detect the secreted defensin molecules from the E. cloacae dissolvens strain carrying the pBR/DG plasmid [62]. For this purpose, the recombinant strain was grown overnight at 37 °C in LB liquid medium containing the tetracycline (12.5 µg/ml) antibiotics. Bacterial cultures were centrifuged at 4000 g for 15 min at 4 °C, and the supernatant was collected and separated by 13% gradient SDS-PAGE. The supernatant from WT E. cloacae dissolvens was used as a control.
We used two plasmids, including PDB47-Scorpine-HasA and PDB47-HasA hosted in Serratia AS1 (a gift from Prof. Marcelo Jacob-Lorena, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA) [11]. The plasmids were extracted and transferred into the WT E. cloacae dissolvens.

Transmission-blocking assay
The transmission-blocking assay was carried out using WT/engineered bacteria, including the ampicillinresistant E. cloacae WT, the ampicillin-and tetracyclineresistant E. cloacae GFP−D strain and the ampicillin-and apramycin-resistant E. cloacae HasA and E. cloacae S−HasA strains. The bacterial strains were cultured in BHI broth at 37 °C, and antibiotics, including ampicillin (100 μg/ ml), tetracycline (12.5 μg/ml) and apramycin (80 μg/mL) were added to the media based on their antibiotic resistance patterns. After overnight growth, the bacteria were harvested by centrifugation (3000 g, 10 min), washed twice in sterile PBS and resuspended (5% wt/vol)in a sterile fructose and red dye (1/50 ml) solution (Nilgoon ® , Tehran, Iran) to obtain a concentration of 10 9 cells/ml. Female mosquitoes aged 3-5 days were placed in five groups, each containing 25 females, and fed on sterile cotton wool soaked in the fructose + dye solution, with or without bacterial cells, for 24 h. As each mosquito became satiated with the sugar solution, as identified by the red dye, they were separated via a sucking tube and transferred into another cage where they were starved for 4-6 h prior to being allowed to feed on infected blood. Each of the five groups of mosquitoes (each comprising 15-20 females) were fed on the same P. berghei-infected mouse. The infected blood-feeding process was carried out at 17-20 °C for 20 min, and the infected blood-fed mosquitoes were maintained on 5% fructose/0.05% paraaminobenzoic acid (PABA) at 19 ± 1 °C and 75 ± 5% RH under a 12:12 D:L photoperiod. Data were pooled from four biological replicates.
A subset of the sugar-fed females infected with bacteria, as well as the blood-fed specimens, were tested for the presence and proliferation of the bacteria by midgut dissection at 0, 12, 18, 24, and 36 h post-blood meal (Fig. 4). The population dynamics of the engineered bacteria colony-forming units (CFUs) were defined by plating serially diluted homogenates of midguts on LB agar plates containing 100 μg/ml of the appropriate antibiotics.
All experiments on the rodents were performed in accordance with the guidelines of the Ethical Board of Tehran University of Medical Sciences (TUMS), Iran.

Statistical analyses
The normality of data was checked using the Shapiro-Wilks test. Significant differences in oocyst intensities between two samples were analyzed using the Mann-Whitney test. Multiple-sample comparisons were analyzed using the nonparametric Kruskal-Wallis test, and means and medians were compared using Dunn's test. All statistics were performed using GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, CA, USA). A P-value < 0.05 was considered to be statistically significant.

Proliferation of bacteria in mosquito midgut
The dynamics of the engineered bacteria showed that, in general, the bacteria were easily established in the mosquito midgut through feeding the mosquitoes on sugar meals laced with bacteria. The bacteria proliferated strongly in the mosquito midgut after ingestion of a blood meal and became the dominant microflora of the midgut, based on the number of CFUs in the plates. Enterobacter cloacae dynamics were monitored at different times post-blood meal by dissection of the midgut and plating of the midgut homogenates on selective antibiotic-containing plates. The number of engineered bacteria increased dramatically, by more than approximately 10,000-fold at 24 h after ingestion of a blood meal (Fig. 4a, b), and the rapid propagation of transgenic E. cloacae occurred simultaneously with the development of the ookinete stage of the Plasmodium parasite in the mosquito gut (Fig. 4c, 5).

Transmission blocking assay
The oocyst numbers in the mosquito midguts were counted on 10 dpi. The median values of all strains of WT or transgenic bacteria significantly (P < 0.0001) impaired the development of P. berghei in the An. stephensi midgut, in comparison with the control group ( Table 1).
The WT strain of E. cloacae was found to inhibit oocyst formation by approximately 72%, in comparison with the control group (without bacteria). The transgenic E. cloacae strain expressing the HasA protein alone inhibited oocyst formation by 86%, while the transgenic E. cloacae strain expressing both the HasA and scorpine proteins inhibited oocyst formation by 92.5%. The transgenic E. cloacae strain expressing the defensin protein had the greatest inhibitory effect (92.8%) on oocyst formation. Importantly, the infection prevalence (the percentage of mosquitoes that had ≥ 1 oocysts) was 86.3% in the control group, which was reduced to 47.1, 25 and 20% in paratransgenic mosquitoes with the WT type, GFP-D and S-HasA strains, respectively. The transmission-blocking potential (TBP) index was determined in paratransgenic mosquitoes with the WT, GFP-D and S-HasA strains was determined to be 45.4, 71 and 76.8, respectively (Fig. 6).

Discussion
The Plasmodium parasite is very vulnerable in the mosquito midgut and, consequently, components of the midgut microbiome could negatively affect parasite development in several ways, such as by impairing its development through secreting anti-parasitic compounds, by activating the host immune system and/or by competing with the parasite for available space in the midgut [6,27]. Enterbacter cloacae is a known symbiont of the gut microflora of most Anopheles species; as such, it has been suggested as a good candidate for the paratransgenic control of malaria [63]. In our study, transgenic E. cloacae showed rapid propagation by 18-24 h after the mosquito consumed the blood meal, suggesting it can block parasite development by competing for the same space as the parasite (Figs. 4c, 5), by a greatly increased expression of anti-plasmodium molecules that lyse the Plasmodium parasite in the Anopheles midgut (Fig. 6) and by activating the mosquito  immune system against the bacteria, which also leads to parasite control [49].
The present study was designed to investigate the efficacy of different strains of E. cloacae in disrupting P. berghei development, while previous studies have investigated different aspects of E. cloacae [45,49,63,64]. In our study, we showed that E. cloacae multiplied rapidly in the mosquito midgut, and that by 18-24 h after ingestion of a blood meal it was the dominant species in the midgut microflora. This was shown by the GFP marker and by culturing the mosquito's midgut contents at different times after the blood meal. Similarly, Pumpuni et al. [31] showed that the midgut bacterial load of An. gambiae and An. stephensi increased by 11-40 fold by 24 h after blood-feeding. Demaio et al. [65] also obtained similar results in Aedes triseriatus, Culex pipiens and Psorophora columbiae, and Wang et al. [27] reported that the bacterial load of Plasmodium agglomerans increased by 200-fold in the An. stephensi midgut by 24-48 h after ingestion of a blood The frequency of E. cloacae GFP-D in the An. stephensi midgut at different times (T) after ingestion of a blood meal. c Schematic development of the Plasmodium parasite in Anopheles midgut shows simultaneous bacterial proliferation and ookinete formation at 18-24 h after a blood meal meal. The finding of Dehghan et al. [63], that E. cloacae was highly stable in a sugar solution, suggested that using sugar bait stations to introduce the transgenic bacteria in the field could be a feasible paratransgenic approach.
It is known that the development of the Plasmodium parasite could be affected by the presence of certain bacteria in the microflora of the mosquito midgut. In this study, the interaction of E. cloacae and P. berghei in vivo led to a significant inhibition of oocyst formation, relative to the control group (P-value < 0.0001). This correlates well with the findings of Pumpuni et al. [30] who reported that the presence of 100,000 Ewingella americana cells in the mosquito midgut reduced the P. falciparum infection rate to zero and those of Gonzalez-Ceron et al. [45] who reported a reduction in the P. vivax infection rate in An. albimanus in the presence Serratia marcescens, E. cloacae and Enterobacter amnigenus. In this regard, the coincidence of bacterial multiplication with the ookinete stage in the Anopheles gut will affect the bacteria-parasite interaction both directly and indirectly. We showed that transgenic bacteria could overcome the harsh environment and barriers in the Anopheles midgut, such as digestive enzymes, to become the dominant component of the gut microflora, leading to an increase in Fig. 5 The presence or absence of motile ookinetes (barrel-shaped, shown by arrows) of P. berghei in the midgut of An. stephensi mosquitoes in the control and E. cloacae GFP-D (Test) group at 20-24 h after ingestion of an infected blood meal. Upper panels: Presence of P. berghei ookinetes are seen in the remains of digested red blood cells in the mosquito midgut in the absence of E. cloacae bacteria. Lower panels: Strong proliferation of E. cloacae GFP-D bacteria correlates with a lack of ookinetes 20 h after ingestion of an infected blood meal with 10 4 -10 5 parasites/µl in the midgut. The thick smear was stained by Giemsa. The bacteria in the dissected An. stephensi midgut (squares) were transferred to BHI agar plates and found to express GFP under the fluorescent (Fig. 3b) microscope  the expression of antiparasitic molecules. This correlates well with the findings of Dong et al. [32], who showed that when the Chryseobacterium meningosepticum bacterium entered the An. gambiae midgut, it rapidly became a dominant species, indicating the competitive nature of this bacterium in the midgut environment.
The results of this study showed that all of the bacterial strains tested disrupted the development of P. berghei to a significant degree, compared with the control group (P < 0.0001). Even the E. cloacae WT led to significantly impaired parasite development (P < 0.0001), indicating the inherent effect of these bacteria in parasite control. The transgenic E. cloacae GFP−D strain, expressing defensin, inhibited parasite development still further compared with the WT (P = 0.003), indicating the suppressive effect of defensin, which lyses the parasite inside the mosquito gut [66][67][68]. The inhibitory effect of scorpine was very similar to that of defensin, and we saw no significant differences in inhibition of oocyst formation between E. cloacae S−HasA and E. cloacae GFP−D (P = 0.051). Similarly, Kokoza et al. [66] expressed cecropin A and defensin A in Ae. aegypti mosquitoes to control P. gallinaceum, and reported that Plasmodium transmission was completely blocked.
Scorpine is an anti-malarial peptide from the venom of the Pandinus imperator scorpion and its amino acid sequence is very similar to those of cecropin and defensin, which has led to the suggestion that scorpine might have a similarly inhibitory effect on the P. berghei [50]. Indeed, Conde et al. [50] found that it completely inhibited P. berghei fertilization and oocyst formation. Wang et al. [11,27] reported that the symbiotic bacteria, P. agglomerans and Serratia AS1, transgenically expressing scorpine, could inhibit P. falciparum development in An. gambiae by 98% and 93%, respectively. The additional expression of HasA protein in the E. cloacae S−HasA strain was found to enhance the anti-Plasmodiun effectiveness of scorpine. It is possible that HasA could create a membrane pore in the E. cloacae wall to allow the direct export of scorpine protein from the bacterial cytoplasm into the mosquito midgut.
Three bacterial species have previously been proposed as candidates for paratransgenetic malaria control: Serratia AS1, Asaia sp. and P. agglomerans; these species transgenically express anti-Plasmodium proteins and have been demonstrated to be suitable micro-symbionts in the mosquito midgut [27,42]. Here, we evaluated a new candidate bacterium, E. cloacae, and showed that it has a strong innate control effect on the Plasmodium parasite in the mosquito midgut, and that this effect could be enhanced by the transgenic expression of anti-Plasmodium proteins.
The symbiotic bacterium Asaia, transgenically expressing scorpine, has previously been shown to inhibit P. berghei development by 63% in the An. stephensi midgut [42]; in comparison, we found that, when expressed in E. cloacae, scorpine caused a 92.5% inhibition of oocyst formation. This remarkable difference is most likely due to the inherent antiparasitic activity of the E. cloacae bacterium. In addition, Wang et al. [27] reported that the expression in P. agglomerans of HlyA protein (which, like HasA, causes pore formation in the bacterial wall) had a negligible effect (21.2%) on parasite development. Therefore, E. cloacae, owing to its intrinsic antiparasitic properties could be preferred to other paratransgenesis candidates such as Asaia sp. and P. agglomerans. This advantage can be attributed to the stimulation of the mosquito's immune system and the secretion of serine protease inhibitors, which are produced by mosquitoes to control bacteria, but which are not specific to the target organism and are suppressed if the Plasmodium parasite is present in the midgut [49]. The E. cloacae bacterium is found in the normal gastrointestinal microflora of humans and many other animals and is generally reported to be widespread in insect midguts [43,45,64,69,70], thus alleviating any potential safety concerns concerning its release in the field.

Conclusions
In conclusion, we consider an alternative strategy for control of the Plasmodium parasite that involves the use of bacterial symbionts of the mosquito, genetically engineered to express anti-Plasmodium effector molecules. The paratransgenesis strategy converts the proven mosquito vector into an ineffective disease vector. This approach could be effective for multiple mosquito and parasite species concomitantly. The findings of the present study provide the foundation for the use of either WT or genetically modified E. cloacae bacteria as a powerful tool to combat malaria. However, further studies are needed to determine how effectively these bacterial strains can be established in the field and the conditions required to do this.