Development of a genetic tool for functional screening of anti-malarial bioactive extracts in metagenomic libraries

Malaria is one of the most important and widely spread infectious diseases, especially in developing countries, constituting a major health threat to about 40 % of the world population [1]. The World Health Organization (WHO) estimated 207 million new cases of malaria worldwide in 2012, with over 600,000 deaths as a result of the severe complications associated with the disease. At present there is no fully effective vaccine against malaria, leaving chemotherapy as one of the most immediate options for fighting this disease [2]. Unfortunately, the number of available anti-malarial drugs with different chemotypes and novel mechanisms of action is low, and the number of compounds entering clinical trials is even lower [3], limiting the expected availability of new, approved compounds for human infections in the near future.

In addition, the rate at which drug resistance is emerging and expanding exceeds the current development of new drugs, in particular given the lengthy process from compound characterization to approval for use in humans [4]. This fact is of particular concern when taking into account the reports of artemisinin-resistant Plasmodium falciparum strains in Southeast Asia [5, 6]. Therefore, it is essential to keep searching for new anti-malarial chemotypes with novel mechanisms of action.

Efforts in recent years to find anti-malarial compounds have focused on the exploration of chemical libraries of public domain, using high-throughput screenings (HTS) on parasite culture [7‐9]. Natural products, which have been the basis of the majority of current anti-malarial medicines [10], are also an important source of compounds of medical importance as they may include different structures for optimizing therapeutic measures against malaria. For this reason, research for the identification and characterization of new anti-malarial agents over the past years has gone beyond the study of plants to include marine organisms, fungi and bacteria [11].

Nowadays it is widely accepted that the microbial world, besides being the largest fraction of biodiversity on the planet, is one of the greatest sources for drug discovery, with frequent applications in human health [12]. It is very likely that the wide diversity of unexplored microbial functions may harbour several uncharacterized molecules, such as metabolites or peptides with potential therapeutic activity. Efforts to elucidate the genomes from non-cultured microorganisms (about 99 % of the total diversity), coupled with the need to discover compounds with biotechnological potential, has promoted the development of metagenomics [13]. This approach involves the direct extraction of total DNA from a given environment, which is subsequently cloned and transferred into bacterial hosts (so far, mainly in Escherichia coli) allowing the construction of metagenomic libraries. These libraries are often subjected to functional analysis for phenotypic identification of a particular activity.

The identification of anti-malarial compounds in metagenomic libraries is a promising alternative, but it represents additional technical challenges. Despite the success in identifying novel drugs from natural sources, pharmaceutical companies still prefer to carry out massive screens of pure synthetic compounds [14], an approach that is difficult in metagenomic libraries. For instance, the screening of gene libraries requires assessing parasitic cultures in the presence of complex bacterial clone extracts on which is hard to identify potential antiparasitic activity. To overcome this limitation, this research focused on the design and development of a non-radioactive methodology to screen metagenomic libraries for anti-malarial activity using bacterial extracts. To do this, different extracts expressing the anti-malarial peptide Dermaseptin S4 (DS4) [15] were tested in order to optimize the experimental conditions that promote antiparasitic activity detection against a novel P. falciparum bioluminescent strain. With this purpose, a platform for the parasite growth consisting of 96-well microplates was established, which allowed for both the incubation of multiple bacterial extracts and the efficient detection of the parasite’s viability, achieved in a multiple, reproducible and composite manner. The results gathered here pave the way for future mid- to large-scale analysis of metagenomic library clones, in a novel strategy for the expanded exploration of microbial diversity for anti-malarial traits by culture independent means.

The development of a stable P. falciparum 3D7 transgenic line, capable of expressing the firefly luciferase gene under the control of the intergenic region of Plasmodium berghei elongation factor 1 (EF-1α) gene has been previously reported [16]. Moreover, a remarkable sensitivity of the luciferase assay using this transgenic parasite and a well-correlated linear relationship between luminescence and parasite density (r = 0.98) has been observed. Plasmodium falciparum 3D7/pHDEF1-luc was therefore used to develop a method for bioluminescence detection in 96-well microplates by assessing if parasite bioluminescence, as a determinant of parasite growth, was comparable to standard radioactive detection methods [18, 25]. Four anti-malarial drugs with different mechanisms of action were tested in vitro against P. falciparum 3D7/pHDEF1-luc cultures. Results demonstrated that IC50 values using the luciferase assay were not significantly different from the ones using the hypoxantine method (Table 1). The high variability observed in the triplicates of mefloquine is not uncommon in these assays [26]. Therefore, this transgenic parasite can be used to develop mid- to high-throughput tests for identifying potential anti-malarial compounds, as already shown for other bioluminescent transgenic lines similarly constructed [9, 27].

The DS4 peptide was chosen for the design of a gene capable of promoting anti-malarial activity when expressed in E. coli. This linear peptide, isolated from the skin glands of Phyllomedusa sauvagii frogs, has shown anti-malarial effect (evidenced by radiolabeled hypoxanthine parasite incorporation) on both chloroquine-sensitive and resistant P. falciparum strains, with IC50 values ranging from 0.27 to 2.2 μM [15]. The coding sequence for DS4 peptide was designed based on the most robust E. coli codons, different from the common approach of using the most frequent codons or those which optimize the codon adaptation index (CAI) [28]. This has been shown to be a determinant for efficient expression of some recombinant proteins, primarily determined by the availability of the amino-acylated (charged) tRNAs and not by the total levels of tRNAs [29]. Transcription of this coding sequence is controlled by an upstream ribosome-binding site (RBS) and the arabinose promoter (P_BAD), and a downstream transcriptional terminator region (Fig. 1). The in silico identification of secondary structures in the region −4 to +37 (from the first ATG codon) showed a single secondary structure with a free energy value of −7.5 kcal/mol, much higher than the values reported as detrimental for an efficient initiation of recombinant protein synthesis in E. coli [30].

Once the ds4 synthetic gene was obtained, it was amplified and inserted into a unique restriction site of the metagenomic insert of pCC2FOS_F076. After transforming both episomal DNAs harbouring the ds4 gene (pUC57_DS4 and F076_DS4) in E. coli Epi300, gene transcription was validated by independent induction of bacterial gene expression with L-arabinose, compared with the same DNAs lacking the ds4 gene (Fig. 2a). The P_BAD was able to both promote ds4 gene transcription from the F076_DS4 construct (in the context of metagenomic DNA), and to tightly repress the ds4 transcription by incubating the bacterial culture with D-glucose (Fig. 2b). Since there are no previous reports of recombinant DS4 peptide expression it is important to mention that the regulatory system used in this research, based on the interaction of P_BAD with the AraC protein, has proven to be quite useful for the expression of potentially toxic proteins [22]. Although the AraC protein is essential for transcription from P_BAD [31], its encoding gene was not included in the synthetic ds4 gene. This is due to the fact that the E. coli Epi300 strain used has a genomic copy of the araC gene that supplies that regulatory function [32]. Therefore, DS4 peptide expression from the designed synthetic gene proposed in this research is restricted to bacterial hosts that express the AraC protein.

Parasitic growth inhibition assays were performed using the cell pellet extracts (concentrated samples) from E. coli Epi300 pUC57_DS4 and the metagenomic clone E. coli Epi300 F076_DS4, compared with the controls not expressing the DS4 peptide. The DS4 peptide is lipophilic and has high electrostatic affinity for cell membranes [33], a partition that would result in a relative decrease of its concentration in the soluble fraction, after using the non-denaturing extraction protocol proposed. After standardizations, significant anti-malarial activity was observed using bacteria expressing the DS4 peptide from plasmid and fosmid DNAs (Fig. 3), demonstrating the feasibility of this approach. It is important to note that the bacterial pellet extracts from E. coli harbouring either one of the episomal DNAs lacking the antimalarial DS4 peptide coding gene (pUC57, pCC2FOS or pCC2FOS_F076) had per se an inhibitory growth effect on the parasite proportional to the amount of bacterial extract used. Thus, it is necessary to normalize bacterial extract concentration when testing multiple metagenomic clones, in order to observe antiparasitic activity over the inhibitory effect of the extract itself, avoiding false positives. These results showed that bacterial extracts expressing the recombinant anti-malarial DS4 peptide exerted overall significantly higher parasite mortalities when compared to the effect of the extracts alone. This system can be used to screen bacterial extracts with metagenomic-encoded information for anti-malarial activities using the tested and optimized bioluminescent assay for P. falciparum viability in a 96-well format.