Targeting Protein Synthesis in the Apicoplast and Cytoplasm of Plasmodium

Informations projet

MEPHITIS

N° de convention de subvention: 223024

Projet clôturé

Date de début 1 Janvier 2009

Date de fin 30 Juin 2012

Financé au titre de

• Specific Programme "Cooperation": Health

Coût total € 3 253 380,21

Contribution de l’UE € 2 145 151,00

Coordonné par

FUNDACIO INSTITUT DE RECERCA BIOMEDICA (IRB BARCELONA)
Spain

Final Report Summary - MEPHITIS (Targeting Protein Synthesis in the Apicoplast and Cytoplasm of Plasmodium)

Executive Summary:
The protein synthesis machinery represents one of the most useful targets for the development of new anti-infectives. MEPHITIS has, for the first time, coordinated an effort to study tRNA biology in Plasmodium falciparum. It has implemented specific schemes for the development of new pharmacological screens, several initiatives for the selection of new potential anti-malarial drugs, and projects designed to answer fundamental questions regarding protein synthesis in Plasmodium. As a result, the laboratories in MEPHITIS do now accumulate a large body of experience in the biology of protein synthesis in this parasite, which will form the basis for future research. The aims and results of the project can be summarized as follows:
a) To study the protein translation machinery of Plasmodium.
Mephitis has unveiled a new tRNA binding protein that can transport exogenous tRNAs into the parasite’s cell. Within MEPHITIS we have characterized the cellular distribution and the biological activity of several components of the protein synthesis apparatus, clarifying the cellular distribution, determining their ability to recognize exogenous tRNAs, and studying the effect of metabolic cogactors upon their activity.

b) To investigate the effects of know translation inhibitors in the malaria parasite.
We have studied a battery of known ARS inhibitors, exploring their ability to specifically shut down apicoplast translation, and the possibility of their use in combination. As a result of this initial characterization we have been able to focus in a new family of molecules that shows great promise in malarial chemotherapy. We have used zebrafish as a reporting model to test the toxicity levels of the apicoplast-specific molecules detected in this project and found them to be innocuous.

c) To identify and characterize novel drug targets in Plasmodium falciparum .
We have solved and published two different protein structures of Plasmodium proteins linked to tRNA biology: D-amino acid deacylase and TyrRS. In addition two additional structures (LysRS and TrpRS) are solved and submitted for publication. Two different reports have been published on the evolution of Plasmodium ARS, one of them as a general review article signed by all participants in the project published in Trends in Parasitology, a widely read journal in the field.

d) To generate new lead compounds, and improve the methods to select them.
We have attempted to implement a new selection screen for anti-malarial compounds in human cells. Unfortunately, solubility and folding problems encountered when expressing the Plasmodium ARS selected in human cells have prevented us from reaching that goal. We have successfully used our data to generate at least two new families of potential inhibitors. One is based in the structure of lysyl adenylate to target LysRS, and the second is based in the structure of borrelidin to target ThrRS. Several hundreds of new compounds have been synthesized and tested against the parasite. The results of this work are being currently prepared for publication.

Project Context and Objectives:
The specific aims of the project were:
- To study the protein translation machinery of Plasmodium.
This aim included the investigation of the dynamics of tRNA in Plasmodium, including the characterization of a novel tRNA binding protein discovered by the Frugier lab (partner 4). It also contained the characterization of the tRNA recognition mechanisms by Plasmodium ARS. This goal has required the purification of pure and active forms of these enzymes. Finally, we aimed at studying gene translation in the apicoplast and cytosol of Plasmodium, unraveling differences that could provide ways to rationally design new translation inhibitors. The progress made in the molecular biology aspects of this goals allow us to be optimistic about fully reaching them before the end of the project.

- To investigate the effects of know translation inhibitors in the malaria parasite.
As part of this aim we have characterized the activities of a large set of known and new ARS inhibitors both in terms of inhibitory activity and inhibitory nature. We have accomplished this analysis and identified compounds with low nanomolar activity that could be taken into further tests in animals. We have thus proceeded to test their relative activity and toxicity in a vertebrate model. This has allowed us to conclude that borrelidin, a ThrRS inhibitor, displays a powerful anti-malarial activity that is coupled to an interesting induction of immunity. We are pursuing these studies further in collaboration with the laboratory of Dr. Bautista (UCM, Madrid). In addition, we had initially planned to investigate the different effects of these molecules upon the parasite by array techniques. This goal was amended however (and approved by the commission), and we will instead apply the same transcriptomic techniques to the study of the biological role of the protein being investigated by partner 4.

- To identify and characterize novel drug targets in Plasmodium falciparum.
We have obtained structural data for several Plasmodium aminoacyl-tRNA synthetases, and we have published the first crystal structure of one of these enzymes. We have also obtained structural data for other components of the translation machinery, namely elongation and initiation factors active in the apicoplast, and crystallization test of these proteins are underway. Finally, we have determined and published the evolutionary history and proximity to human orthologs for all Plasmodium ARS, and essential set of data for future analysis of these enzymes.

- To generate new lead compounds, and improve the methods to select them.
We continue to develop a new method of screening based on a positive selection scheme. Currently we are trying to solve problems of expression and solubility of Plasmodium ARS when expressed by human cells. In parallel, we have used our data to design novel potential inhibitory scaffolds directed against two different enzymes: lysyl- and glutaminyl-tRNA synthetases. Efforts to construct chemical libraries against two additional components of the translational apparatus are also under way. All these libraries are being, or will be, tested for their anti-malarial activity and promising compounds will be selected for further development.

Project Results:
MEPHITIS started in January 2009, as a consortium of eight laboratories that are closely interlinked. The three main areas covered by its components are: molecular biology, structural and computational analysis, and drug discovery.
The molecular biology efforts within MEPHITIS have focused mainly in the analysis of two different enzyme families: the aminoacyl-tRNA synthetases (ARS) and the elongation factors. In both cases we have studied the function of these proteins in the different biological compartments of the parasite. Of particular importance to us is the study of these proteins in the apicoplast, a Plasmodium-specific cellular compartment that constitutes an excellent target for the development of new drugs.
The computational and structural analysis of protein synthesis has focused mainly in the study of the evolutionary relationships among aminoacyl-tRNA synthetases, because this is an obligate step to the identification of the best candidates for future drug development. In parallel the crystallographic analysis of these enzymes has been undertaken with excellent results. Drug development, on the other hand, has been the main focus of the project. Intense collaboration between different partners has resulted in the discovery of new inhibitory molecules, and the re-visiting of previously known drugs that have shown promising results in animal tests.

Results achieved per individual Work packages :

WP 1-2: CHARACTERIZING THE POTENTIAL OF KNOWN ARS INHIBITORS AS ANTI-MALARIAL TARGETS
WP Leader: Dr. Lluis Ribas de Pouplana, IRB Barcelona

Summary of progress
This Work Package requires the establishment of reliable assays for the analysis of ARS inhibition in vitroWe have been able to optimize two different assays to test inhibitors in vivo and in vitro, and we have extended these studies to animal models.

Most significant results
1. We have successfully obtained a soluble and active form of apicoplastic PfKRS, which is now available for the in vitro analysis of new inhibitors.
2. Testing of a complete battery of known ARS inhibitors is finished. We have now also tested combinations of the two molecules that showed promising activities (borrelidin (Bo) and pseudomonic acid (PMA)). The characterization of combinations of Bo and PMA has been extended to an animal model of Malaria, in collaboration with Dr. Bautista (UC Madrid). Very promising results have been obtained with borrelidin. This molecule shows therapeutic activity comparable to chloroquine.
3. We have designed a bioinformatic model of PfKRS to select new potential Plasmodium ARS inhibitors. We have generated a new library of 50 compounds, in collaboration with Partner 6.
4. Moreover, the latest generation of compounds designed in collaboration with Partner 6 has yielded compounds that display selectivity towards apicoplastic PfKRS. These new compounds represent the first example of drugs that come out of MEPHITIS and can specifically attack the translation machinery of this organelle.
WP 1-3: CHARACTERIZING tRNA RECOGNITION MECHANISMS.
WP Leader: Dr. Lluis Ribas de Pouplana, IRB Barcelona (Spain)

Summary of progress
1. First we have studied the cellular distribution of ARS in the parasite, to try to understand if Plasmodium ARS exist in the mitochondria. This requires the construction of Plasmodium falciparum stable transgenic parasite lines.
2. We have generated the constructs for the integration of chimeric forms of PfKRS and PfIRS predicted to be active in the apicoplast and citosol of Plasmodium. We are using two different vectors, encoding different fluorescent proteins (GFP, Cherry) to localize these proteins in the parasite (by using of in vivo confocal microscopy techniques, like FRAP). We expect that in the coming months we should be able to obtain the transgenic parasites. Initial experiments indicate that the constructs have integrated. We are currently characterizing the new strains of the parasite.
3. We included PfQRS in this set of experiments. This decision is based on bioinformatic predictions that suggest that this protein may be transported to the mitochondria. This prediction is rendered even more interesting by the unique evolutionary history of this enzyme. The localization or not of this protein in the mitochondria could be crucial to determine the translationally active compartments in Plasmodium.

Most significant results
1. We have managed to determine the cellular localization of PfQRS, and demonstrated that this enzyme does not reside in the mitochondria, at least during the asexual erythrocytic cycle of the parasite.
2. We have been able to identify apicoplast-specific inhibitors of translation. This was one of the major goals of MEPHITIS, and shall be published soon.
3. We have also developed a new assay for the analysis of PfWRS activity, in a collaboration with the Sharma laboratory that has been made possible by MEPHITIS, and which will allow us to monitor the activity of this enzyme as a function of the anemic state of the host. A previously unexplored question in the biology of gene translation in Plasmodium.













WP 1-4: DEVELOPMENT OF NEW SELECTION SCHEMES FOR INHIBITORS OF TRANSLATION ENZYMES.
WP Leader: Dr. Lluis Ribas de Pouplana, IRB Barcelona (Spain)

Summary of progress
The generation of new screening mechanisms for the selection of ARS compounds requires the successful expression of Plasmodium ARS in human cells. This task has proven to be extremely difficult with current expression technology. To try to advance in this direction, and screen as large a number of expression constructs as possible, we have generated a battery of fusion constructs to express Plasmodium falciparum ARS in different E. coli strains. We are taking advantage of the Protein Expression Core Facility from IRB to ensure the expression of these engineered proteins in E. coli, but also to be sure that these proteins will be soluble and active.
As described above, we have been successful in obtaining a pure, soluble, and active form of apicoplastic KRS. This represents a major step towards the characterization of this protein and for the design of apicoplast-specific translation inhibitors. However all attempts to express chimeric forms of Plasmodium ARS in human cells have been, so far, unsuccessful.
Despite the initial negative results we will continue to try expressing Plasmodium ARS in human cells. A possible solution to the detected solubility problems may be the co-expression of cognate tRNAs which may be able to help in the proper folding of the enzymes. In addition, having the orthologous ARS-tRNA pair expressed in the human cell would greatly facilitate the construction of the positive selection scheme proposed in this Work Package.

Most significant results
1. Codon optimized PfKRS1, PfKRS2, and a truncated form of PfKRS2 have been cloned in a battery of different bacterial vectors. Successful purification of pure PfKRS2 has been achieved in April 2010.
2. Unfortunately we must report that the generation of human cell cultures expressing active forms of PfKRS has not been possible. The multiple control experiments ran point at a folding problem of this enzyme in the context of the human cell that rends it completely inactive. Although this result does not necessarily detract from the initial proposal, it is clear that a different enzyme target will need to be used to continue this approach for drug-dicovery against malaria.


WP 3-1: STRUCTURAL AND FUNCTIONAL STUDIES ON tyr-tRNA SYNTHETASE (TyrRS) FROM Plasmodium Falciparum
WP Leader: Dr. Amit Sharma, International Centre for Genetic Engineering, (New Delhi, India)

Summary of progress
Pf-TyrRS gene has been PCR amplified and cloned into suitable vector which expresses soluble protein in E. coli. The protein has been purified to homogeneity and its activity tested in an aminoacylation assay (in collaboration with partner 1, IRB Barcelona) expressed and purified. A number of active and other site mutants have also been produced and activity assays on several of them are on-going within this consortia. The protein has also been crystallized and its structure resolved.

Most significant results
1. Soluble, active protein has been produced. This is a major achievement as production of soluble, functional P. falciparum proteins is not always routine. The protein is produced in sufficient amounts for experimental work and its purification has been standardized to give reproducible batches of protein.
2. Crystals of good quality have been obtained. Again, this is the first report of P. falciparum trna synthetase crystallization. This has been very tricky to resolve and thousands of conditions were screened for crystallization. We are now in a position to routinely make crystals although their quality varies somewhat and we are therefore still optimizing the system.


WP 3-2: STRUCTURAL AND FUNCTIONAL STUDIES ON D-tyr-tRNA DEACYLASE (DTD) FROM Plasmodium Falciparium
WP Leader: Dr. Amit Sharma, International Centre for Genetic Engineering, (New Delhi, India)

Summary of progress
D-tyr-tRNA deacylase gene has cloned, expressed and purified. D-tyr-tRNA deacylase gene has been cloned and expressed in modified pET28a vector. Protein was purified to homogeneity using different chromatography techniques. Structure determination has been done and several complexes of DTD with adenine and amino acids have been resolved crystallographically.

Most significant results
Once again, soluble, active protein DTD has been produced in E. coli in sufficient amounts for all experimental work. DTD purification has been standardized to give reproducible batches of protein and antibodies against the protein have been produced. Localization studies on DTD have also been completed.

WP 3-3: STRCTURE-BASED INHIBITOR DISCOVERY
WP Leader: Dr. Amit Sharma, International Centre for Genetic Engineering, (New Delhi, India)

Summary of progress
For both DTD and YRS, structural information has helped in identification of several potential inhibitors of these enzymes. In case of DTD, few inhibitors have been tested experimentally and these data are part of the paper published on DTD from this work and grant.

Most significant results

1. DTD: The structural analysis of DTD allowed us to do docking studies using various libraries in order to identity potential inhibitors. Several potential hits were tested and we then were able to study in greater detail 4 inhibitors. Part of this exercise is already published in 2010.

2. YRS: a recent report has identified 2 inhibitors of YRS in P. falciparum. Clearly, the parasite has two versions of YRS (in different organelles presumably) and therefore the ability of the identified inhibitors against either or both YRS needs to be addressed. It has been very difficult to produce the presumed apicoplastic YRS from P. falciparum so far, but the inhibitors may be testable directly against the other YRS from P. falciparum. We have performed screening of inhibitors based on structure of YRS.

WP 4-1: aaRS AND tRNA IDIOSYNCRASIES IN Plasmodium
WP Leader: Dr. Magali Frugier, Centre National de la Recherche Cientifique (France)

Summary of progress
The objective of this Work Package 4.1 is to characterize aminoacylation systems from Plasmodium falciparum and compare them to their homologues in Homo sapiens. Three systems are studied, cytosolic Tyrosyl-tRNA synthetase (TyrRS) and Asparaginyl-tRNA synthetase (AsnRS) and the apicoplastic TyrRS; they were expressed and tested in vitro for their capacity (i) to aminoacylate not only their own tRNA but also (ii) to perform cross-species aminoacylations. Indeed, the ability for the Plasmodial enzymes to recognize and aminoacylate the corresponding human tRNA would support the hypothesis of human tRNA import in the parasite (see WP 4.2).

Most significant results
In vitro characterization of cytosolic TyrRS
Plasmodium falciparum (Pf) and Homo sapiens (Hs) cytosolic TyrRS have been cloned, expressed and characterized in vitro. Our results indicate that cross-species aminoacylations are very efficient (Hs TyrRS aminoacylates Pf tRNATyr and Pf TyrRS amlinoacylates Hs tRNATyr). This first observation is in agreement with our hypothesis where P. falciparum could use efficiently human tRNAs to support its own translation.

In vitro characterization of cytosolic AsnRS
The study of Pf and Hs cytosolic AsnRSs has been also initiated. In this case, enzymes are much more difficult to handle (than cytosolic TyrRSs), certainly due to the presence of long extensions at the N-terminus of both enzymes and Asn-rich insertions in the plasmodial one. Both AsnRSs are active in amino acid activation (first step of the reaction). However, the detection of in vitro asparaginylation necessitated some adjustment (radioactive asparagine is not commercialy available). We were able to produce radio-labeled Asn, by converting 14C aspartic acid into 14C asparagine with the Pyrococcus abissi Asparagine synthetase (gift from D. Kern). Both the human and the plasmodial systems were then able to catalyze in vitro asparaginylation, kinetic parameters have been determined and show that:
- tRNA transcripts (without post-transcriptional modifications) are sub-optimal substrates, especially in the case of the human system.
- Moreover, unlike AspRS and TyrRS, the plasmodial AsnRS does not charge any human tRNAAsn: neither the tRNAAsn present in the total RNA from HeLa cells nor the in vitro transcribed human tRNAAsn sequences (several human isodecoders have been tested), indicating that cross aminoacylation reactions are inefficient. This result is not in contradiction with our work hypothesis. Indeed, the absence of any cross-aminoacylation between the parasite’s AsnRS and the human tRNAAsn would keep the concentration of Asn-tRNAAsn limiting in the parasite and would not destroy the hypothetical co-translational folding mechanism of plasmodial proteins (Frugier et al., 2010).
Post-transcriptional modifications, a quosine at position 34, could also play an important role in asparaginylation in human. In order to test if this modification is beneficiary or not for asparaginylation in human and in the parasite, wild type tRNA sequences (human and Plasmodium) were cloned and expressed in a E. coli strain expressing or not the corresponding modifying enzyme. Purification conditions need to be optimized to recover pure tRNAs (with or without the modification) from total E. coli tRNAs.

Crystallization assays of aaRS
Diffusion of Light Scattering (a technique which indicate if the molecule is homogeneous or not) have been performed and they indicate that P. falciparum AsnRS is homogenous, but the human enzyme is not. P. falciparum AsnRS was used in crystallization assays and we obtain one small crystal, however, these conditions need to be optimized to grow better crystals. In the case of the human enzyme, we need to change expression and purification conditions to produce a more soluble protein, suitable for crystallization.
Ion the case of P. falciparum Aspartyl-tRNA synthetase, in collaboration with partner 5, the codon usage in the gene encoding AspRS was “harmonized”. This new sequence, when expressed in E. coli did not produce a more active protein. However, we still need to test if its solubility is improved at concentrations that are needed for crystallization.

WP 4-2: A UNIQUE tRNA BINDING PROTEIN IN Plasmodium
WP Leader: Dr. Magali Frugier, Centre National de la Recherche Cientifique (France)

Summary of progress
Our objective was to study tRBP (tRNA binding protein), a protein identified by our team and found only in Apicomplexa. Theoretically, this protein is built around a putative trans-membrane domain and displays a small internal domain at its N-terminus and an external tRNA-binding domain at its C-terminus. Our work consisted in characterizing in vitro and in vivo this protein and the identifying its function.

Most significant results
In vitro characterization of tRBP. We have shown that tRBP binds tRNAs (crude human tRNAs) in vitro with a high affinity. This binding is specific only for tRNAs. As anticipated, the tRNA-binding capacity of tRBP is concentrated in its only C-terminal domain. Moreover, directed mutagenesis, in this domain, allowed also to point out at least 2 amino-acids specifically involved in this interaction. Foot-printing experiments indicated the protection pattern left by tRBP on the tRNA is “complete” suggesting that the structure of the tRNA molecule is modified upon binding tRBP. Moreover, despite numerous attempts, it was impossible to observe the complex by DLS or purify it by gel exclusion chromatography. All together, these results suggests that tRBP binds transiently the tRNA and unfolds it .
Expression and localization of tRBP in vivo. Western blot experiments show that tRBP is present during the hepatic and the erythrocytic stages of the mammalian host but also during the different development stages in the mosquito vector. Only the stage corresponding to oocysts in mosquito seems not to express tRBP. Moreover, as anticipated, immuno-localization experiments show clearly that when tRBP is expressed, it is localized at the surface of the parasite (the C-terminal domain being exposed outside Plasmodium).
Human tRNA import. Because of tRBP biding capacity and localization, we tested the possibility for Plasmodium sporozoites (the infective parasitic form found in mosquito salivary glands and which infects host hepatocytes) import exogenous tRNAs. Our results show that indeed, exogenous tRNAs enters alive parasites and can be specifically detected by FISH. Import happens only with tRNAs and with in vitro transcribed tRNA (unmodified). It is inhibited when anti-tRBP antibodies are added, suggesting that tRBP is involved in this process.
Expression of tRBP in HeLa cells. Two constructs of tRBP (full-legth and the C-terminal domain) were expressed in HeLa cells to observe the putative toxicity of this protein, however, no effect was observed on the growth or the translational efficiency of HeLa cells.
In vivo deletion of tRBP. The design of the KO strains for tRBP has been discussed with partner 1 and we have produced several constructs. These constructs have been used by partner 1 to replace or delete the tRBP gene in P. falciparum. Insertions were verified by Southern Blots by partner 1 and all recombinant strains grow and does not seem to be affected by the absence of tRBP. It indicates that the protein is not essential for the parasite survival, at least at the erythrocytic stage.

Structural study of tRBP. Seven different constructs have been cloned, all of them can be expressed in E. coli and only three are purified to homogeneity (as measured by DLS experiments). Numerous crystallization conditions have been tested, with the different constructs (full-length tRBP, N-terminal domain or C-temrinal domain), with or without tRNAs, yet without success, until now.
Study of Plasmodium CSP binding to human ribosome. Plasmodium falciparum CSP (Circum Sporozoite Protein) is secreted when the parasite traverses hepatocytes (at the beginning of the hepatic stage) and was previously shown to inhibit the human translation. We want to study the capacity of CSP to bind human ribosomes in vitro and determine (i) if this binding is specific or not, (ii) which domain of the protein is involved in this process and (iii) how it functionally inhibits translation.
The CSP was cloned, expressed successfully in E. coli and purified as a soluble protein, yet only when diluted. We have also purified large amount of human ribosomes from Hela cells and both 40S and 60S subunits have been prepared. Our first tests indicate that, indeed the CSP binds specifically eukaryotic ribosomes (human and yeast) but not bacterial ribosomes. It binds the 40S ribosome sub-unit but not the 60S sub-unit or the 80S ribosome). In vitro translation systems allowed us to show that the CSP inhibits efficiently translation at the initiation stage. We are completing this work by designing CSP mutants in order to determine the protein domains involved in this inhibitory action.
Determination of CSP binding site on Human Ribosome. We planned to use foot-printing experiments to determine the CSP binding site on the human ribosome. However, naked risosomal RNA (18S) is not recognized by the CSP anymore (ribosomal proteins may be involved in the binding directly or indirectly), a situation that does not allow to perform any footprinting experiments.






WP 5-2: TOXICOGENOMICS
WP Leader: Dr. Manuel Santos, Universidade de Aveiro (Portugal)

Summary of progress
The animal model of zebrafish has been used to evaluate the toxicity of compounds generated by partners 1 and 6.

Most significant results
Several compounds displaying specific inhibition of apicoplastic enzymes and good anti-malarial activity in vitro has been shown to be innocuous in the vertebrate model used in this work package.

WP 5-3: CODON USAGE AND CONTEXT
WP Leader: Dr. Manuel Santos, Universidade de Aveiro (Portugal)

Summary of progress
Using Anaconda, an in house developed bioinformatics tool, the relative synonymous codon usage, the codon context and the codon adaptation index (CAI) of Plasmodium falciparum, Escherichia coli and Saccharomyces cerevisiae were determined. Anaconda was also used to study regions of rare codons in different tRNA synthetases and the elongation factor Tu (EF-Tu). In the lysyl-tRNA synthetase gene, nine sites of potentially relevant rare codons were identified. Thus far, eight of these sites could be modified by site directed mutagenesis and were tested for their effects on expression and protein solubility (see task 5-3.2.B).

A novel bioinformatics tool for gene optimisation was also developed (Eugene, papers published in Bioinformatics and Current Bioinformatics). Main features of this tool include: a) standard optimization algorithms merged into a single tool, b) context optimization (taken over from ANACONDA), c) harmonization of the codon usage profiles (novel algorithm) and c) visualization of rare codon rich regions through automated alignments of orthologous genes and protein structural features.

Most significant results
Using Anaconda, the P. falciparum LysRS was optimized for expression in E. coli based on CAI, context and a combination of CAI and context. Further, rare codon rich regions in the LysRS were studied and synonymous genes with preserved rare codons were designed. Additionally, a working version of the new gene optimisation tool has been developed, which allows optimisations based on e.g. codon harmonisation, codon usage, GC content and combinations of different design parameters. Our experimental data show that codon context, codon usage and are codons have a significant impact on translational accuracy. We were able to design, synthesize and express in E. coli different versions of the P. falciparum LysRS whose translational error was lower than that of the wild type gene (See below), thus validating our in silico methodologies.

WP 6-1: DESIGN AND SYNTHESIS OF NEW ANTI-MALARIAL AGENTS BASED ON A RATIONAL TARGET PRIORIZATION APPROACH AND KNOWN ARS INHIBITORS
WP Leader: Dr. Miriam Royo, Parc Cientific de Barcelona, Spain

Summary of progress
This WP is dedicated to the design and synthesis of new anti-malarial based on rational target priorization approach and known ARS inhibitors using combinatorial synthesis tools. The validation target process and the study of the action of the known compounds was carried out by partner 1(Ribas) and 7 (Ralph), and the library design and synthesis by partner 6 (PCB). During this period of time (M19-M42) the work was mainly focused on target validated or known compounds studied and selected by partner 1 (Ribas).

Task 6.1-A consisted on the design and synthesis of a new library of inhibitors of Lys RS based on compounds previously described with relevant inhibitory capacity. In collaboration with partner 1 we have designed the first library based on the intermediate complex that is formed in the aminacylation reaction. We have built a virtual library with 1764 members based on the lysine intermediate. All compounds were docked by partner 1 against PfLysRS and HsLysRS, in order to assess not only specificity but also selectivity between both species. The GlideScore was used to rank the capacity of inhibition of these ligands. The cut-off of this value to consider a ligand with a positive prediction of inhibition was set at -10.5. Also two minimum requirements were established to consider a ligand as a positive ”hit”: 1) a minimum GlideScore of -10,5 when docked against PfLysRs, and 2) a minimum difference of 1.5 points between GlideScores when docked to PfLysRS and HsLysRS.
Applying these restrictions 36 compounds were preselected, 70% of them were hydroxamates and 65% had S,S quirality on the 4-aminoproline residue. This library is composed by two sub-libraries, one with lysine as recognition moiety and the other with thiolysine (a lysine mimetic). To synthesize both sub-libraries protected lysine or thiolisine was needed. Protected lysine was commercially available but was necessary to synthesize protected thiolysine. Both sub-libraries were synthesized by solid-phase methodology and all compounds were purified and characterized.
Primary in vivo Plasmodium growth inhibition assays (pLDH method, 3D7A strain) with Lys sub-library compounds at 75-150 ?M gave four potential hits. On three of them the inhibition effect was detected at 96h means that probably means these compounds are inhibiting LysRS at the apicoplast. Primary in vivo Plasmodium growth inhibition assays (pLDH method, 3D7A strain) with thiolysine sub-library showed 3 potential hits that also showed a delay inhibition effect (96h). All potential hits were also evaluated by visual inspection of smears.
Compounds that showed in vivo growth inhibition activity were resynthesized to achieve better final purities (?95%). These seven compounds were resynthesized in solution phase to improve the final yield of the synthesis with the aim to obtain the final compounds easily. We have detected some problems in the synthesis process: the thiolysine synthesis and the purification of final compounds from a secondary product (POPh3). Purification problem was solved with the use to Gold columns (Isco Teledyne) and finally thiolysine was achieved. Two of the compounds with an ester bond were chemically unstable and consequently discarded.

The antimalarial activity of the re-synthesized compounds was evaluatedby visual analysis of P. falciparum smears and Fluorescence-assisted cell sorting (FACS) was used to calculate the IC50 of the most active compounds (three compounds) that some of them also showed a clear delayed effect (two compounds) . For FACSanalysis, Syto-11 was used to discriminate parasitized from non-parasitized RBCs.
In order to investigate the selectivity and specificity the two compounds that showed inhibition delay effect (M-26 and M-37), their ability to inhibit HsKRS, PfKRS-1, and PfKRS-2 was explored. In vitro aminoacylation assays were performed using radiolabelled lysine and in vitro transcribed tRNALys, and the effect of the compounds upon these aminoacylation reactions was quantified. Both M-26 and M-37 were found to inhibit PfKRS-2 (Apicoplastic pfKRS) , but were not active against HsKRS or PfKRS-1 (cytoplamatic pfKRS), which is in accordance with the docking predictions. Thus, we can conclude that M-26 and M-37 are selective inhibitors of apicoplastic PfKRS-2.
Compounds M-26 and M-37 were tested for toxicity on zebrafish embryos to anticipate possible undesired toxic effects. M-26 and M-37 were added at their respective IC50 concentrations to zebrafish embryos (20 embryos per compound, independent duplicates), and the toxicity was evaluated at 24, 48 and 72h. None of these two compounds showed any type of genotoxicity.
In collaboration with the group of Prof. Jose M. Bautista (UCM) in vivo antimalarial activity of both compounds in mice was studied using the rodent malaria parasite Plasmodium yoelii 17XL (Py17XL) MRA-267 as animal model. Both compounds were not enough active to reduce the hyperparasithemia of the mice.
A manuscript have been prepared and submitted to Journal of Medicinal Chemistry with all these results.

Together with partner 1 we have established a planning to design these potential inhibitors based on 6 points: 1) Identification of well-known inhibitors of IleRS and LysRS; 2) Docking of all these inhibitors against apicoplast PfIleRS and PfLysRS; 3) Ranking inhibitors establishment vs. PfIleRS and PfLysRS; 4) Design of the dual compounds based on the results of the ranking; 5) Docking of these dual compound vs. PfIleRS and PfLysRS; and 6) Ranking of these dual compounds and synthesis of the best qualified.
We have identified 72 LysRS inhibitors and 829 IleRS inhibitors from literature. These structures were transferred to partner 1 in sdf format for docking against PfIleRS and PfLysRS. Partner 1 also docked compounds from different commercial sources (FDA drugs ?3.000 molecules; Prestwick Chemical Library ?1.800 molecules and ZINC library ? 30.000 molecules) against PfIleRS and PfLysRS. Docking results with PfIleRS gave small number of compounds (16 hits) vs. hits obtained for PfLysRS (141). All compounds also were docked against PfGlnRS and 161 hits were obtained.
In parallel to virtual screening we have searched other well known aaRS inhibitors from the synthetic point of view to develop a new family of dual inhibitors. We have searched for structures that can be easily modified and we have designed potential single and dual inhibitors. As first choice we have selected cispentacin (ProRS inhibitor) and icofungipen (IleRS inhibitor) as structures easily modified by simple synthetic modifications.
The synthesis of protected cispentacin and icofungipen in large quantities was established and protected and free cispentacin was obtained. In the case of Icofungipen two syntheses were explored and both presented diverse problems in intermediate steps.

Tetracycline derivatives: These compounds were designed as potential dual inhibitors of PfGlnRS-PfLysRS. Together with partner 1 docking studies were performed to establish the points on tetracyline structure that can be modified to introduce derivatizations that promote the interaction with the active center of both enzymes. Two potential points were detected: the tertiary amine and the amide. Diverse reactions that can introduce modifications at these points were explored without any results. Only one compound with a modification in the tertiary amine was synthesized and explored its Plasmodium growth inhibitory capacity. This compound was less active than tetracycline and showed a remarkable delay effect at 150 ?M. This effect can be attributed to a potential action on the apicoplast. The impossibility to obtain other different derivatives forced us to leave the synthesis of this compound type.
Benzimidazole derivatives: The compounds were designed as potential inhibitors of PfGlnRS. As a result of the virtual screening two of them were synthesized (M2-01 and M2-02) and their Plasmodium growth inhibitory capacity was tested at different concentrations. Both compounds showed inhibitory properties and one of them (M2-01) also a delay effect (potential action on the apicoplast). Considering these results a library of 50 compounds based on benzimidazole scaffold was designed and synthesized. All these compounds were easily obtained by two synthesis step process with good yields and high purity (?95%).
The Plasmodium growth inhibitory properties of the whole library were screened by FACS at 75?M at 48 and 96h using P. falciparum 3D7A strain. Only those compounds that showed inhibition percentage over 70% were selected as potential hits (6 compounds). The IC50 of these compounds at 48h and 96h was determined given values between 10-40?M. The values of IC50 of these compounds at 48 and 96h showed a delay effect but not highly remarkable. The hemolytic activity of the selected compounds was also determined. None of them showed hemolytic properties. In order to establish preliminary results related to the selectivity of these compounds to P. falciparum, the capacity of these compounds to inhibit E. coli growth at 100?M was studied. None of them showed E. coli inhibition properties. Their potential cytotoxicity to human cells was explored by MTT assays with HEK 293T cells incubated with 100?M of each compound during 24, 48, 72h. None of these compounds showed remarkable cytoxicity vs. HEK 293T cells.
The six compounds were also tested for toxicity (zebrafish embryos; Danio rerio (Zebrafish), AB strain) at concentrations near the IC50 obtained for the parasite by partner 5 (Manuel Santos) to anticipate possible undesired toxic effects in vertebrates.
All experiments were performed according to the European law for animal experiments (2010/63/EU). No specific ethics approval under EU guidelines was required for this project, as all zebrafish used in this study were between 0 and 4 days old. This is within the European law (Council Directive 86/609/EEC), which excludes foetal and embryonic forms. The toxic effect of the selected compounds on the zebrafish embryos was evaluated following diverse parameters: accumulative mortality and morphological and physiological abnormalities.
Besides mortality, the presence of morphological and physiological abnormalities is also a good indicator of the compounds toxicity. The embryonic development was monitored until 96h in the presence of the different benzimidazol derivatives and registered specific alterations, namely developmental delay (24h) edema (48h), pigmentation loss (48h) and hatching delay (72h).
Considering the high toxicity of compounds M2-26, M2-46 and M2-48 (mortality, developmental delay and pericardial edema incidence), at doses lower or within the range of the IC50’s calculated for Plasmodium, these compounds are not suitable to be used in vertebrates. On the other hand, compounds M2-17 and M2-47 show low toxicity levels as they do not induce mortality even at the higher concentrations and do not induce developmental delay or malformations and can be considered are promising compounds for the treatment of Plasmodium infection in vertebrates.
Taking all results account compounds M2-17 and M2-47 were re-synthesized (around 100mg each) to be evaluated their anti-malarial efficacy in mice (these experiments will be carried out by Prof. Jose M. Bautista group, UCM). Part of the re-synthesized compounds will be send to GlaxoSmithKline (Diseases for Developing World Department) to determinate the IC50 by other methodology and the mode-of action. All results will be collected in a publication.

Cispentacin and Icofungipen derivatives: Together with partner 1 diverse dipeptides and tripeptides that contain icofungipen and cispentacin on the second position were docked vs. catalytic center of PfLysRS, Pf IleRS and PfGlnRS, obtaining the best potential modifications build dual ligands based on these to scaffolds, Icofungipen and cispentacin. Due to difficulties in the synthesis of icofungipen, this was abandoned and we focused on the synthesis of cispentacin derivatives. To build the derivatives based on cispentacin we have considered other configurations of both chiral centers. The capacity of these compounds to inhibit P. falciparum growth was evaluated by flow cytometry using 3D7A strain. No inhibition activity was detected at 100?M, probably due to these compounds don’t reach their target.

-Synthesis of a new antimalarial agent inhibitor library based on ARS mediated by partner 7 (task 6.1C):
Due intellectual property issues was not possible to work with the structures that partner 7 (S. Ralph, U. Melbourne) previously determined as potential inhibitors of divers ARS. To overcome this problem we decided to work in new family of potential antimalarial compounds, based on a scaffold (?-peptide) that our laboratory developed previously. This library was previously screened by Swiss Tropical and Public Health Institute and found 4 compunds with clear antimalarial activity and two of them with IC50 similar to chloroquine. We selected these two compounds and re-checked their antimalarial activity and evaluated their chemical stability. From these two compounds one was not chemically stable and was discarded and the other gave % P. falciparum growth inhibition similar to those from Swiss Tropical and Public Health Institute. We have considered this compound as potential validated hit and generated a small library of compounds based on its structure. In collaboration with Prof. Jose M. Bautista the antimalarial efficiency in vivo was established. This compound was not capable to inhibit the parasitemia in infected mice, but was capable to delay the death of the mice two days. This fact can be caused by the poor solubility of the compound. Based on these results a library of 20 compounds based on ?-peptide scaffold was designed and 14 compounds were successfully synthesized and purified. The inhibitory capacity of these compounds was evaluated at 48 and 96h by flow cytometry using P.falciparum 3D7A strain at 20?M. In general all compounds didn’t show a delay effect in the parasite growth inhibition process. The IC50 at 48h of some compounds is being estimated but some of them are around 1?M being promising to be assayed in animal experiments. We have established a contact with GlaxoSmithKline (Diseases for Developing World Department) to determinate the IC50 by other methodology and the mode-of action. The antimalarial efficacy in vivo (mice) will be estimated by the Prof. Bautista, UCM.






WP 6-2: DESIGN AND SYNTHESIS OF APICOPLAST-TARGETED INHIBITORS OF TRANSLATION
WP Leader: Dr. Miriam Royo, Parc Cientific de Barcelona, Spain

Summary of progress
We have designed and synthesized a family apicoplast PfLysRS inhibitors in collaboration with partner 1 (Ll. Ribas). The members of this library were specifically designed to be inhibitors of PfLysRS-2 that is located in the apicoplast. Malaria parasites possess two distinct lysyl-tRNA synthetases, PfKRS-1 (PF13_0262), and PfKRS-2 (PF14_0166). Based on subcellular localization prediction software, PfKRS-1 is expected to be cytosolic, whereas PfKRS-2 is expected to be targeted to the apicoplast. In general, apicoplastic-targeted enzymes tend to be of bacterial origin. To confirm the bacterial origin of PfKRS-2 and evaluate its evolutionary distance with its human homologue, Ribas group performed a structure-based phylogenetic analysis on class II lysyl-tRNA synthetases from all kingdoms. The results show that apicoplastic lysyl-tRNA synthetases cluster with bacterial enzymes and are distantly related to HsLysRS. Using a manually curated homology model of PfLysRS-2 we noticed that those residues involved in the recognition of lysine in the bacterial, human, and P. falciparum enzymes are conserved across species. Importantly, however, other residues in the active site cavity that are not involved in substrate recognition are not so well conserved, and the sizes of the catalytic cavities are significantly different. This suggests that PfLysRS-2 might be able to accommodate ligands in the active site that may not be able to bind in the HsLysRS cavity due to sterical restrictions. Altogether, these analyses suggested that specific design of inhibitors targeting the active site of PfLysRS-2 is feasible. We have designed a compound virtual library to identify molecules that may mimic the lysyl-adenylate intermediate. To construct the library, a proline derivative was used as a ribose mimetic, and an heterocycle as an adenylate substitute, as previously described. In addition, four more points of chemical diversity were explored: i) both lysine and thialysine derivatives were used as lysine analogues; ii) the phosphate linker was replaced by other types of chemical linkers; iii) heterocyclic groups were used as substituents of the adenylate moiety, and iv) the stereochemistry of the proline and the lysine derivatives was varied. With this approach a library of 1764 compounds was designed and evaluated by docking the molecules against the 3D structures of both PfLysRS-2 and HsLysRS.
All 1764 compounds were docked both to PfLysRS-2 and HsLysRS, and the different ligand poses obtained were ranked using GlideScore. The compounds to be synthesized for experimental testing were selected on the basis of their selectivity towards the PfLysRS-2 enzyme. By this criterion we selected 36 compounds for further analysis. From these 36 compounds we have established the criteria to design and synthesize new library (50 compounds). As described in task 6.1-A the library was synthesized employing a strategy based on solid-phase synthesis. All synthesized compounds were initially tested for their ability to kill P. falciparum parasites using the pLDH assay. Inhibitors of apicoplastic protein synthesis kill the parasite in a retarded manner, and therefore, we used the "delayed death" phenotype as an initial indication that a compound in our library may be preferentially targeting apicoplastic lysyl-tRNA synthetase. Our initial screening allowed us to select five compounds that presented a clear delayed inhibitory effect. The activity of these compounds was further confirmed by visual inspection of smears.

The five most active compounds from the library (M-12, M-24, M-26, M-33 and M-37) were re-synthesized. The antimalarial activity of the re-synthesized compounds was evaluated by visual analysis of P. falciparum smears. Highest inhibition rates were observed for compounds M-12, M-33 and M-37. In order to select specific inhibitors of the apicoplastic translation machinery it was decided to focus on those compounds causing a clear delayed effect phenomenon. Thus, M-26 and M-37 were chosen as drug candidates for further in vitro and in vivo analyses, given that these compounds show maximal difference between the inhibitory rates of these compounds at 48 and 96h.
In order to investigate the selectivity and specificity of M-26 and M-37, we first verified their ability to inhibit HsLysRS, PfLysRS-1, and PfLysRS-2. In vitro aminoacylation assays were performed using radiolabelled lysine and in vitro transcribed tRNALys, and the effect of the compounds upon these aminoacylation reactions was quantified. Both M-26 and M-37 were found to inhibit PfLysRS-2, but were not active against HsLysRS or PfLysRS-1, which is in accordance with the docking predictions. Thus, we can conclude that M-26 and M-37 are selective inhibitors of apicoplastic PfLysRS-2.
From other screening carried out in task 6.1-B in collaboration with partner 1 we have detected tetracycline as potential scaffold that can act into an apicoplast target, due to can be can accommodate in the catalytic site of PfLysRS-2 being recognized at the binding site. From docking studies and analyzing the predicted binding mode we have predicted which tetracycline scaffold positions can be modified to be PfLysRS-2 inhibitors. Concretely to generate inhibitors for apicoplastic PfLysRS the amide of the tetracycline should be modified. Assays tetracycline showed a delay effect in their inhibition mode of action that can be attributed to activity vs apicoplastic targets. We have tried to introduce diverse modifications that mimetize lysine side chain in the amide, but we weren’t successful. Modifications in other positions (i.e.: tertiary amine) that were predicted favorable to bind to PfGlnRS catalytic binding site were also attempted, but only one compound was obtained. Evaluation of the Plasmodium growth inhibition by this compound at different concentrations showed a remarkable delay effect. This delay effect can be attributed to the action of tetracycline to other apicoplast targets than PfLysRS-2. Due to the generation of derivatives of tetracycline by simple modifications was not possible and more complicated synthetic processes should be applied this research line was abandoned.

WP 6-3: ORGANIZATIONOF MEPHITIS PROJECT KICK-OFF MEETING
Organization: Dr. Miriam Royo, Parc Cientific de Barcelona, Spain

First meeting of MEPHITIS partners was held in Barcelona, between 29 and 30 of March, 2009. External invited experts attended (Didier Leroy, Gerardo Turcatti and José M Bautista) the meeting and gave positive inputs to the consortium. The meeting was divided in three scientific sessions: 1) Molecular Biology of Plasmodium; 2) Computational and structural biology of Malaria and 3) Drug Discovery. All partners presented their work related to the project and all collaborations where discussed and arranged.
Decisions related to internal organization and coordination of the consortium were made and established the bases for the collaboration.





WP 7-1: GENETIC VALIDATION OF ARS DRUG TARGETS
WP Leader: Dr. Stuart Ralph, University of Melbourne (Melbourne, Australia)

Summary of progress
This work package involves the localisation and disruption of potential ARS drug targets in Plasmodium falciparum based on parasite transfection. These localisations were initially planned to be conducted using a series of Gateway vectors. These vectors have been finished, but have been founded to be unstable and unreliable for the desired purpose, so we changed to a different vector system based on the pGLUX vector that allows episomal expression. We have applied these vectors to the localisation of 10 different enzymes involved in Plasmodium translation.

So far we have used this to localise and publish localisations for 3 ARS enzymes and 1 apicoplast elongation factor. (Biswas et al 2011, Jackson et al 2012) Localisations have been completed for an additional ARS and for 4 other putative translation factors. Publications are in preparation for those additional proteins.

We have additionally adopted a second localisation strategy is underway based on the integration of an haemagglutinin (HA) tag into the chromosomal copy through 3’ replacement. We have now successfully validated this approach and have used this to confirm localisations for seven of proteins involved in Plasmodium translation. This approach also allowed us to characterise mechanisms of dual targeting for proteins sent to both the apicoplast and the cytosol.
The HA tagging vectors also allow us to test whether we can target the loci of these ARSs for disruption. Confirmation of integration into the chromosomal locus for four CysRS loci now allows us to directly test whether failure to disrupt ARS genes is due to growth defects, rather than refractoriness of the locus to genetic manipulation.
Generation of parasites with fluorescent proteins targeted to the apicoplast and mitochondria allowed us to test specificity of action of inhibitors. We demonstrated specific morphological defects in apicoplast for parasites treated with mupirocin, confirming the specificity of this compound for apicoplast translation.

Most significant results
We have so far localised the EFTS protein through use of GFP-fusion approach. This work was performed in collaboration between Dr Ralph and Dr Habib and has been published (Biswas et al 2011)
We have established a cell biological assay for demonstrating organellar specificity of ARS inhibitors, and used this to confirm an apicoplast target for mupirocin (Jackson et al 2012)
We have established localisation for four ARS enzymes, showing each of the four unique ARSs in Plasmodium to be dual targeted to the apicoplast and Cytosol (Jackson et al 2012, Pham et al manuscript in preparation).





WP 7-2: RATIONAL IDENTIFICATION OF ARS INHIBITORS
WP Leader: Dr. Stuart Ralph, University of Melbourne (Melbourne, Australia)

Summary of progress
This work package involves the rational identification of ARS inhibitors. A bioinformatic tool has been generated in collaboration with the TDRTarargets network and used to prioritise the most promising drug targets. We have used the TDRTargets database to prioritise the most promising ARS enzymes from P. falciparum on the basis of druggability, potential for selectivity, quality of structural model, and desirability of chemical starting material. From this subset, druggable pockets were sought. Structural models of ARSs were developed based on homology modeling, and these were used to search for druggable sites. We have selected the dual targeted cysteinyl-tRNA synthetase, cytosolic glutaminyl-tRNA synthetase and apicoplast methionine tRNA synthetase as promising targets for inhibitor development

The cysteinyl tRNA synthetase was selected as the first starting target with attractive druggable pockets and was used to perform in silico docking. Libraries of several million compounds selected for lead and drug –like molecules have been docked against the cys-tRNA synthetase and 50 molecules have been selected as promising putative. All 50 compounds have been tested for parasiticidal activity using a SYBR Green parasite growth assay and 11 were identified as having low micromolar IC50 inhibition. In vivo protein assay tests have now been performed on all these 11 positive compounds, demonstrating different levels of direct translational inhibition.

These active compounds were used to identify two promising chemical scaffolds, and a further 94 compounds were ordered based on these scaffolds. 50 of these compounds had low µM inhibition of parasite growth and will now be tested for specificity against CysRS and in toxicity assays. The cysteinyl tRNA synthetase has been successfully overexpressed in E. coli and purified, and we have now established an assay to measure aminoacylation based on previously published assays for the corresponding enzyme in E. coli. This assay was used to show that the PfCysRS can charge both apicoplast and cytosolic tRNA-cys, consistent with the dual targeting of the enzyme to both the cytosol and apicoplast.

The Plasmodium falciparum cysteinyl tRNA synthetase was used to complement an E. coli temperature sensitive mutation in the EcCysRS, further confirming the activity of this Plasmodium enzyme. Both the complemented bacterial strain and the in vitro assay will now be used to monitor specificity of inhibition for the compounds identified as putative CysRS inhibitors. In collaboration with Dr Claire Simons, we tested compounds that were synthesized based on their predicted inhibition of the MetRS. Although several of these killed Plasmodium parasites grown in culture, none were sufficiently potent to pursue as drug leads.

Most significant results
65 Compounds that kill Plasmodium falciparum cultured in blood at low micromolar concentrations have been identified based on a rational screen against the PfCysRS. These inhibitors will be used as a starting point for further compound improvement, but may already be appropriate to serve as useful tools for investigating ARS biology.

WP 8-1: COMPARATIVE GENOMICS OF AARs INHIBITORS
WP Leader: Dr. Elisabetta Pizzi, Istituto Superiore di Sanittà (Italy)

Summary of progress
Hidden Markov Models (HMMs) for each of the 20 AARs have been constructed on the basis of the AARs Data Base at http://www.man.poznan.pl/aars/ and then have been used to identify AARs in P. falciparum proteome. We identified 34 AARs which have been characterized in terms of functional domains, architectures and evolutionary relationships with other species (see Deliverable Report no. 8-1-A on “Identification, characterization and evolution of aars in P. falciparum” submitted on 31 December 2009). The prediction of signal sequences for cellular localization in P. falciparum was performed using various available online web-servers – MITOPROT , PredictNLS and PATS for mitochondria, nucleus and apicoplast respectively. Apicoplast proteins have been predicted. Furthermore amino acid sequences of AARs have been aligned with those from reference organisms to identify Parasite Specific Domains (PSDs) (see Deliverable Report 8-2-A on “Identification and characterization of PSDs” submitted on 6 August, 2010).

Most significant results
Identification of cytosolic and apicoplast AARSs in P. falciparum (see report on Deliverable 8-1-A): First of all we exploited current annotation available in PlasmoDB to identify the repertoire of AARSs in P. falciparum genome. According to Enzyme Commission (EC) 38 proteins in PlasmoDB are annotated as belonging to the EC group 6.1.1.- which is the EC number provided for AARSs. Although for the majority of those proteins current annotations allow an assignment to Class I and II of AARSs, for some of them the annotation is still preliminary in that they are annotated as hypothetical proteins or as generic tRNA ligases. For these reasons we decided to use Hidden Markov Models (HMMs). For each AARSs a set of known sequences was utilized to construct 20 HMMs which are then used to carry out data base searches vs P. falciparum proteins. For each data base search a score distribution is obtained and 4 cutoffs were considered to identify AARSs. We observed that 2 proteins annotated as belonging to EC group 6.1.1.- in PlasmoDB are not found by HMMs. PF14_0401 annotated as MetRs is instead a generic tRNA binding proteins as elucidated in the genome reannotation process, while the second one (PFC0470w) is still annotated as ValRs. A total of 18 AARSs can be classified within the 10 AARSs that define class I. All members of this class are represented in P. falciparum proteome. 34 AARSs have been identified in P. falciparum proteome. We found that 23 P. falciparum aaRSs have signal peptides, possibly for directing them to different cellular organelles. 20 AARSs bearing apicoplast targeting signals and 13 AARSs may be exclusive to this organelle. Others are predicted to be shared between apicoplast and mitochondria.
Domain architecture of P. falciparum AARSs: Careful examination of identified AARSs P. falciparum revealed that 31/36 of them has a generic modular architecture that adheres to prototypical AARs. The remaining P. falciparum AARSs have complex domain architectures with concatenation with unusual domains such as Ybak (PF04073), GST (PF00043), Ser-Thr kinase (PF00069), and DNA binding domains (PF03483). The functional relevance of these additional domains fused to typical AARSs in P. falciparum needs to be experimentally addressed. Clearly, the presence of unusual domain fusions in P. falciparum AARSs suggests multiple functional roles for many of these P. falciparum enzymes.
Phylogenetics (see report on Deliverable 8-1-A): In order to study evolutionary relationships of P. falciparum aaRSs with other species, phylogenetic trees were developed in PHYML using maximum likelihood method. For each type of P. falciparum aaRS a separate tree was constructed. AARSs sequences from 102 different species were used for multiple sequence alignments. Based on distance matrices, several P. falciparum AARs sequences clustered as being closer to plants (A. thaliana) or to bacteria (E. coli). It is already known that apicomplexan parasites like P. falciparum house a secondary endosymbiotic plastid, possibly hijacked by lateral genetic transfer from an alga. Therefore, the P. falciparum AARS sequences which are evolutionary close to bacteria and plants are likely to be the outcome of horizontal gene transfer from the plastid. P. falciparum contains 12 such AARS sequences which cluster with bacterial or plant sequences. Functional and structural characterization of these bacterial/plant-like AARS may be relevant in focusing efforts at using AARs as drug targets.
Parasite Specific Domains in AARSs (see report on Deliverable 8-2-A): We have used our knowledge of AARSs to identify and localize PSDs in different AARS polypeptide chains. Structurally, the abundance of crystal structures and functional studies make these multi-domain proteins an ideal model for the search of a putative structural and/or functional role for PSD insertions. Because Plasmodia belong to the Apicomplexa, which are unicellular protists closely-related to fungi and plants, P. falciparum cytosolic AARSs were aligned with their S. cerevisiae homologues. Since apicoplast pathways are essentially prokaryotic, we chose to align putative apicoplast AARSs with their bacterial homologues. In this case, bacterial sequences were chosen largely based on crystal structure availability. In the cytosolic AARSs, 35 insertions were identified and they were found in both class I and class II proteins (19 and 16, respectively). Less than 25% of these PSDs are present at the junction between functional domains, while the majority is inserted between structural modules within these functional domains. They vary in both size (up to 300 amino acids long) and the extent of amino acid repetition. Runs of 6 consecutive Asn are frequent and the longest repetition is found in the first insertion of the cytosolic GluRS with 23 consecutive Asn. Long extensions are found at the N-termini of certain proteins (TrpRS, AlaRS, GlyRS, HisRS and ThrRS), but an “alternative” initiator ATG is consistently present, suggesting that translation may begin in closer proximity to the conserved domains: a scenario that has been verified for the expression of the cytosolic AspRS.

WP 8-2: MOLECULAR MODELLING OF AARs
WP Leader: Dr. Elisabetta Pizzi, Istituto Superiore di Sanittà (Italy)

Summary of progress
To date, no crystal structures have been obtained for any AARSs from P. falciparum. Hence, we performed homology modeling of several P. falciparum AARs using homologous structures available in Protein Structure Data Base (PDB). Known structural templates (? 40% identity) were used for molecular modeling of several P. falciparum AARs including the two TyrRSs (PF11_0181, MAL8P1.125) the PheRS (PFA0480w), ThrRS (PF11_0270), LysRS (PF13_0262), MetRS (PF10_0340), TrpRS (PF13_0205) and GluRs (MAL13P1.281). When PSDs are present both homology modeling and ab initio methods have been utilized. 3D models have been analysed and compared with counterparts in other organisms.



Most significant results
Homology modeling and structure comparisons (Deliverable 8-2-B): Many PSDs were observed for P. falciparum enzymes when compared to their homologous. Location of PSDs in P. falciparum TyrRS between well-conserved secondary structures suggests ability of TyrRS anticodon binding core to accommodate larger sequence inserts with minimum disruption to the catalytic domain. Direct comparison of modeled P. falciparum AARSs with human AARSs revealed several other important structural differences. In particular we examined 3D models for cytosolic TyrRs and putative apicoplast GluRs . Numerous insertions are present in the loop regions linking various ?-helices (?10 to ?13) in anticodon binding domain of P. falciparum TyrRSs (PF11_0181 and MAL8p1.125) when compared to its human homologous (2PID and 1N3L) respectively. The structural changes that manifest as partial conservation of important motifs in P. falciparum AARSs reflect evolutionary divergence, and may be useful for exploitation of parasite-specific features as drug targets. Then we analyzed the 3D model obtained for putative apicoplast GluRs (MAL13P1.281). It is known that in bacteria, archea and organelles tRNAglu and tRNAgln are both charged by a non-discriminating enzyme Glu/GlnRs. The structural bases of the anticodon recognition by discriminating (DgluRs) and non-discriminating (NDGLuRs) GluRs have been elucidated. In DGluRs the anticodon binding site is large enough to accommodate the third base of the anticodon for tRNAgln (G), while in DgluRs the long side chain of an Arginine makes possible to form hydrogen bonds only with the third base of the anticodon tRNAglu (C). When we compared the 3D-model for MAL13P1.281 we found that the anticodon binding site resembles that of a NDGluRs supporting the prediction that this protein is localised in the apicoplast.

WP 9-1: CHARACTERIZATION OF APICOPLAST ENCODED EF-Tu
WP Leader: Dr. Saman Habib, Central Drug Research Institute (Lucknow, India)

Summary of progress
Translation elongation factor Tu is encoded by the tufA gene of the Plasmodium apicoplast and is one of the best conserved proteins encoded by the malarial plastid genome. The protein was expressed as a recombinant fusion protein with MBP at the N-terminus (technical details of expression issues with other tags were described in the report for milestone 2). Although MBP-EFTu (~89 kDa) was obtained primarily as an insoluble protein in E.coli we were able to solubilize and refold the protein into a functional form. The refolded, purified protein was used in assays for GTP hydrolysis and analysis of nucleotide (GDP) release mediated by nuclear-encoded EF-Ts. Apicoplast EF-Tu was also capable of acting as a disulfide reductase, a chaperone-related activity that may explain the retention of the tufA gene in the apicoplast.

Most significant results
EF-Tu gunaine nucleotide binding and GTPase activity : Nucleotide binding of apicoplast EF-Tu was studied by Fluorescence resonance energy transfer (FRET) that measured the excitation of mant-GDP by the single intrinsic trytophan in the nucleotode binding pocket of EF-Tu. Enhanced excitation of mant was observed with increasing MBP-EFTu concentrations. MBP alone and the MBP-EFTu W196G mutant, used as negative controls, did not excite Mant-GDP. Comparison of the affinity of EF-Tu for GDP and a non-hydrolyzable form of GTP (GMPPCP) revealed that the PfEF-Tu had ~10-fold higher affinity for GDP.

Nucleotide exchange on EF-Tu: The greater affinity of EF-Tu for GDP compared to GTP necessitates the requirement of a nucleotide release factor (EF-Ts) that releases GDP from EF-Tu thus letting GTP access the protein. This allows recycled EF-Tu.GTP to participate in the formation of a ternary complex with the next aa-tRNA. The single EF-Ts predicted on the Plasmodium nuclear genome carries a predicted bipartite apicoplast targeting sequence. The predicted processed EF-Ts was cloned and expressed in E.coli (WP 9-2). Increasing concentrations of purified PfEF-Ts were then used to monitor release of mant-GDP from the PfEF-TU.mant-GDP complex. Dose-dependent enhancement of mant-GDP release (indicated by decrease in FRET) from PfEF-Tu was observed indicating that PfEF-Ts is capable of mediating nucleotide exchange on apicoplast EF-Tu. The nucleotide exchange reaction mediated by EF-Ts follows rapid kinetics. The kinetics of GDP release from EF-Tu mediated by EF-Ts was studied by fluorescence measurements on a stopped-flow CD spectro-polarimeter. The apparent rate constant (Ki) of the reaction was determined to be ~2.2 sec-1, a value 55 times higher than the rate of spontaneous release of GDP from EF-Tu (0.04 sec-1).

WP 9-2: CHARACTERIZATION OF EF-Ts
WP Leader: Dr. Saman Habib, Central Drug Research Institute (Lucknow, India)

Summary of progress
The P. falciparum apicoplast genome encodes EF-Tu while its interacting partner, EF-Ts, is predicted to be encoded by the nuclear genome. PfEF-Ts is predicted to carry the bipartite element required for import of proteins into the apicoplast and results obtained in collaboration with Ralph (partner 7) (reported as milestone 3) have shown that the bipartite leader targets PfEF-Ts to the P. falciparum plastid. Although PfEF-Ts has only 19% amino acid identity with E.coli EF-Ts, it carries the conserved DFVA motif required for nucleotide exchange on EF-Tu. The protein was expressed in E. coli as a glutathione transferase (GST)-fusion protein (reported as deliverable 9-2-A), affinity purified and analysed for its ability to mediate nucleotide exchange on Plasmodium apicoplast EF-Tu as well as E. coli EF-Tu. Antibodies were generated againt the recombinant protein; these detected predicted unprocessed and processed forms of the protein in P. falciparum lysate.

Most significant results
Ability of the N-terminal bipartite leader sequence to target nuclear-encoded EF-Ts to the apicoplast was established in collaboration with Ralph (partner 7). Ralph’s laboratory generated parasite transfectants and studied localization of the EF-Tsleader-GFP fusion.

Apicoplast-targeted EF-Ts mediated nucleotide exchange on apicoplast EF-Tu in P. falciparum. (described in report for deliverable 9-2-A). Functional interaction between two proteins encoded by different cellular compartments in Plasmodium was demonstrated.
Plasmodium EF-Ts was unable to catalyze GDP release from Escherichia coli EF-Tu even at high Tu:Ts molar ratios. The lack of conservation of some interacting residues on PfEF-Ts as well as the change in the orientation of the PfEF-Ts N-terminal domain compared to EcEF-Ts, predicted by homology modeling of the Plasmodium factor, may compromise its interaction with helix D of EcEF-Tu.
Molecular dynamics of the PfEF-Tu.PfEF-Ts complex suggested that intercalation of Phe80 of the conserved DFVA motif of PfEF-Ts between the His85 and His119 of PfEF-Tu results in a combination of changes that cause distancing of residues critical for GDP binding in PfEF-Tu. Significant alterations in the GDP-binding pocket of the PfEF-Tu structure upon interaction with PfEF-Ts were observed. As in E.coli Phe80 of the conserved DFVA motif of PfEF-Ts intercalated between the His85 and His119 residues of helix B and C of PfEF-Tu, respectively and caused movement of the two helices. Movement of the P-loop away from GDP was also seen although the flip in the backbone of the P-loop that breaks the H-bond to the ?-phosphate as a result of movement of Val20 in E. coli was not observed. PfEF-Ts also induced a prominent shift of the N136 and D139 residues, which form H-bonds and stabilize the guanine base, thereby relaxing the interactions of these residues with the nucleotide.

WP 9-3: ANALYSIS OF APLICOPLAST-TARGETED INHIBITORS OF TRANSLATION
WP Leader: Dr. Saman Habib, Central Drug Research Institute (Lucknow, India)

Summary of progress
Antibiotics known to target prokaryotic translation elongation factors are being analysed for their inhibition of apicoplast factors. The antibiotic kirromycin (a complex linear polyketide) inhibits protein synthesis by inhibiting the release of bacterial EF-Tu.GDP from the ribosome. The EF-Tu.GDP.kirromycin complex remains on the ribosome preventing the peptidyl transferase reaction and subsequent translocation. In addition, the antibiotic alters the behaviour of E.coli EF-Tu such that the activities elicited by cellular effectors are mimicked. Kirromycin is capable of enhancing GDP release from bacterial EF-Tu in the absence of EF-Ts and it also stimulates GTP hydrolysis in the absence of ribosome and aa-tRNA by lowering the Km and increasing the Kcat of the reaction. The drug has been shown to exhibit anti-malarial activity in P.falciparum blood culture with an IC50 of 50 µM and can bind to P.falciparum apicoplast EF-Tu.GDP [Clough et al. (1999) Protist 150-189-195].

We investigated the effect of kirromycin on GTP hydrolysis by PfEF-Tu as well GDP release from PfEF-Tu in the absence of EF-Ts and observed differences with its effect on bacterial EF-Tu (reported as milestone 2). Assessment of inhibitory effects of two other antibiotics, GE2270A and pulvomycin, on apicoplast EF-Tu was also carried out. GE2270A (a cyclic thiazolylpeptide) and pulvomycin were purified from the Actinomycetes Planobispora rosea and Streptoverticillium netropsis, respectively. The IC50 of pulvomycin for P. falciparum in vitro growth inhibition was calculated to be 0.2 µM.

Most significant results
GTPase activity of PfEF-Tu was not significantly affected by kirromycin. There was only a minor enhancement of GTPase activity with increasing concentrations of kirromycin (0.1 to 100 µM). This is at variance with its effect on bacterial EF-Tu. On the other hand, kirromycin clearly stimulated the release of GDP from the EF-Tu.GDP complex as measured by FRET using mant-GDP. The fraction of non-dissociated mant-GDP consistently fell with increasing concentrations of kirromycin indicating enhancement of GDP dissociation by the drug.
Pulvomycin and GE2270A inhibit parasite growth in vitro. The former enhances the GTPase activity of apicoplast EF-Tu whil the latter has no effect on GTPase activity of the factor, an observation similar to that made for bacterial EF-Tu.
Apicoplast- and mitochondria-targeted EF-G were expressed as recombinant proteins and the effect of fusidic acid on locking of the EF-G.GDP complex onto the ribosome as well as on GTP hydrolysis was assayed. FA exhibited differential inhibition of the apicoplast and mitochondrial factor and inhibited apicoplast EF-G to a greater extent.

WP 9-4: PROTEOMIC STUDIES ON APICOPLAST TRANSLATION PROTEINS
WP Leader: Dr. Saman Habib, Central Drug Research Institute (Lucknow, India)

Summary of progress
All translation factors of the apicoplast, apart from apicoplast-encoded EF-Tu, are encoded by the parasite nucleus. Since the identity and localization of these factors remained to be established, analysis of the apicoplast proteome was envisaged. Earlier deliverables analysed apicoplast EF-Tu, EF-Ts and EF-G. We were now concerned with identification and characterization of initiation factors (IFs), release factors (RFs) and ribosome recycling factor (RRF). In order to characterize the apicoplast proteome, we began by preparation of protein extracts from organelle-enriched parasite fractions that were isolated by cell disruption and differential centrifugation. Since mitochondrial contamination was consistently found in these fractions, we did not proceed with this strategy. The following methods were subsequently used:

Translation factors putatively targeted to the apicoplast were identified from PlasmoDB by sequence identity with prokaryotic factors followed by prediction of targeting to the apicoplast or mitochondria. Putative N-terminal targeting sequences were analysed by PATS, PlasmoAP, SignalP, PlasMIT and MitoProtII. Genes encoding translation factors predicted to function in the apicoplast were PCR-amplified from the parasite genome, cloned in appropriate expression vectors for recombinant expression in E. coli. Purified recombinant proteins were used in functional assays and for generating antibodies for detection in the parasite. Organellar localisation was investigated by construction of N-terminal leader-GFP fusion and transfection.

Most significant results
Putative factors required for translation initiation, peptide release, and ribosome recycling in apicoplast and mitochondria of P. falciparum were identified. Initiation factor IF-1 has a strong prediction for apicoplast targeting. Antibodies against recombinant IF-1 detected transit peptide containing unprocessed form and a transit-cleaved processed form indicating organellar transport of the factor. The putative IF-2 (PF18_0018) exhibited GTPase activity and its tRNA binding domain interacted with IF-1 in vitro. Putative apicoplast peptide release factors RF1 and RF2 have been expressed in E. coli and purified. Antibodies have been generated against RF1 and are to be used for localization of the protein in the parasite. Stop-codon recognition assays are being set up.
Function and localization of RRF1 (PFB0390w) demostrates that it is the apicoplast ribosome recycling factor that functions in conjunction with apicoplast-targeted EF-G to split post-translocation ribosomes. Structural modeling of PfRRF1 with the E. coli RRF as template reveals that RRF1 has a long insertion in the hinge region between the head and tail domains and evaluation of the RRF leader-GFP cleavage products suggests that RRF1 also has a large N-terminal extension that remains after cleavage of the signal and transit sequences. A possible mitochondrial RRF (PFD0990w) is also being investigated.



WP 9-5: CRYSTALLOGRAPHIC ANALYSIS OF EF-Tu and EF-Ts
WP Leader: Dr. Saman Habib, Central Drug Research Institute (Lucknow, India)

Summary of progress
PfEF-Tu has been expressed as a fusion protein with MBP. We were unable to obtain expression of stable protein with other tags (His, GST and S-tag). Co-expression with EF-Ts using the PET-Duet vector also failed to stabilize PfEF-Tu. The insoluble MBP-EFTu fusion protein has been solubilised and refolded to generate a functional form. The amount of purified protein obtained is low and would be insufficient for crystallization. PfEF-Ts was expressed as a recombinant protein in E. coli and purified. The protein purified predominantly as soluble oligomers. A very minor component of the protein was monomeric and this was further purified and used for functional assays with PfEF-Tu. Again the low levels of expression and the still lower levels of monomeric EF-Ts were insufficient for setting up crystallization trays.

Deviations from Technical annex:
Although we got enough refolded recombinant EF-Tu and EF-Ts for functional and activity assays, the yield was low and was insufficient for crystallization and structural studies. This is an independent WP and will not impact other tasks or work of other partners.

WP 9-6: MOLECULAR BIOLOGY AND BIOCHEMISTRY MEETING ORGANIZATION
WP Leader: Dr. Saman Habib, Central Drug Research Institute (Lucknow, India)
Type of the meeting: progress meeting
Date: 1-3 April, 2010
Place: New Delhi, India
Host institution: Dr. Saman Habib, CDRI (CSIR), India (Partner 9)

Attendees: 33 people from all groups.

Notes on the meeting schedule and order of the talks
The meeting started with a short introduction by Dr. Philip de Taxis du Poet, Head of S&T, EU Delegation to India. Next, the external scientific experts appointed in MEPHITIS Scientific Advisory Board and invited to attend this coordination meeting gave their inputs to the consortia on their corresponding expertise areas. After discussions and interactions between partners and the external experts, the project partners started their presentations following the order established by the agenda.
Potential Impact:
MEPHITIS has generated an impact and has been disseminated in the following areas of society:
- Academic knowledge. The main impact in this area has been in the form of numerous scientific publications and collaborations that have resulted from the project. At least one of these collaborations has already been continued in the form of a new competitive grant between the labs of Dr Ribas (partner 1) and Dr. Sharma (partner 4), which extends the scientific collaborations between Europe and India. Many other interactions remain in place that will result in a growing number of reports in the coming years. More importantly, MEPHITIS has opened several new avenues of research in three main areas: the cross-talk between the protein synthesis apparatuses of host and pathogen, the regulatory role of ARS in the infection of Plasmodium, and the development of new families of inhibitors against Plasmodium ThrRS.
- Societal awareness. The main impact in this area has been in the form of public dissemination, which has focused, mainly in Catalonia and India. Apart from the web of the project MEPHITIS has been featured in many mainstream newspapers, educational videos in Internet, radio programs, and TV interviews. We have also participated in Science fairs in Barcelona to explain to young students the importance of the problem of malaria, and the efforts being carried out to eradicate the disease.
- Industry interactions. Mephitis has engaged in contacts with private companies (GSK, Biotica Ltd, and Omnia Molecular SL), with private-public partnerships (MMV), and with other public consortia (BIOMALPAR, CRIMALDDI). In terms of a continuing collaboration we want to emphasize the potential of the discovery that some borrelidin analogs generated by Biotica display an excellent profile of Plasmodium inhibition and safety both in vitro and in vivo. We regard this as one of the most exciting results of this consortium. Borrelidin had long been considered a powerful inhibitor with a poor safety profile. Thanks to MEPHITIS new variants of this molecule with much improved selectivity towards the pathogen have been discovered.

List of Websites:
http://www.mephitis.eu/