|Home | About | Journals | Submit | Contact Us | Français|
Malaria, most commonly caused by the parasite Plasmodium falciparum, is a devastating disease that remains a large global health burden. Lack of vaccines and drug resistance necessitate the continual development of new drugs and exploration of new drug targets. Due to their essential role in protein synthesis, aminoacyl-tRNA synthetases are potential anti-malaria drug targets. Here we report the crystal structures of P. falciparum cytosolic tryptophanyl-tRNA synthetase (Pf-cTrpRS) in its ligand-free state and tryptophanyl-adenylate (WAMP)-bound state at 2.34 Å and 2.40 Å resolutions, respectively. Large conformational changes are observed when the ligand-free protein is bound to WAMP. Multiple residues, completely surrounding the active site pocket, collapsed onto WAMP. Comparison of the structures to those of human cytosolic TrpRS (Hs-cTrpRS) provides information about the possibility of targeting Pf-cTrpRS for inhibitor development. There is a high degree of similarity between Pf-cTrpRS and Hs-cTrpRS within the active site. However, the large motion that Pf-cTrpRS undergoes during transitions between different functional states avails an opportunity to arrive at compounds which selectively perturb the motion, and may provide a starting point for the development of new anti-malaria therapeutics.
Malaria is one of the most serious diseases in the world, causing around 1.2 million of deaths in the year 2010 . The most common and serious human malaria parasite is Plasmodium falciparum. Currently, the near term prospects of arriving at a highly effective vaccines appear to be challenging , while drug resistant parasites remain one of the biggest threats in malaria control . Therefore, it is important that new drugs are developed to fill the pipeline of anti-plasmodium therapeutics.
The aminoacyl tRNA synthetases (aaRS) form a group of ubiquitous enzymes that perform the essential function of charging amino acids to their cognate tRNAs during protein synthesis . In general, the reaction is completed in two steps:
During this reaction, large conformational changes in aaRSs are often needed. Different conformations of the enzyme are required for binding of substrates, arranging the active site for catalysis and releasing products. Depending on the substrate bound, several structural states are observed in crystal structures of tRNA synthetases. Focusing here on tryptophanyl-tRNA synthetases (TrpRSs), extensive structural data are available on both the bacterial and the eukaryotic (cytoplasmic) TrpRS enzymes [5–10]. For bacterial TrpRS, at least three conformational states have been reported. In the open, ligand-free state (F-state), either Trp or ATP can bind to the enzyme without major changes in conformation. However, the simultaneous binding of Trp and ATP in the pre-transition state requires a compact active site where a conserved KMSKS loop closes onto the active site to interact with ATP, and its anticodon binding C-terminal domain moves towards the active site-containing Rossmann-fold domain. After the intermediate tryptophanyl-adenylate (WAMP) is formed, both the KMSKS loop and the C-terminal domain move slightly away from the catalytic core to allow turnover in the partially closed product-state (P-state) [5–7]. In contrast, in eukaryotic (cytoplasmic) TrpRS, the binding of Trp is accompanied by a change from an open F-state to a closed to Trp-bound state. Only the closed conformation has a Trp-binding pocket that complements the amino acid well. Structural elements that ‘closes’ onto the active site are mainly the N-terminus and a conserved AIDQ motif, but the domain motion observed in bacterial TrpRS does not occur. Subtle movements in the KMSKS loop subsequently assist ATP binding, amino acid activation and product release, but the overall closed conformation is maintained in the P-state . The N-terminus then needs to be displaced while the rest of the enzyme stays in the closed conformation to allow the acceptor arm of the tRNATrp to access the active site, resulting in another conformational state . Hence TrpRS, like many other aaRSs, undergo a series of complex conformational changes to perform their function.
Owing to its vital role in protein synthesis, this group of enzymes is in principle an important target for the development of anti-parasitic agents . Recently, there have been an increasing number of reports targeting aaRSs to treat infectious diseases [12–18]. Inhibitors of aaRSs typically bind to the active site or the editing site of these enzymes, preventing the binding or release of amino acid, ATP or tRNA. In order to be effective, these substrate binding site inhibitors need to have a high affinity for the pathogen but not for the homologous enzymes of the human host. High resolution three-dimensional structures should be helpful in identifying differences in the substrate binding sites and also for exploring the nature of conformational changes. Such changes might be prevented by small molecules, ultimately leading to new drugs.
As part of our ongoing efforts to explore aaRSs as anti-parasitic drug targets [14, 19–25], we have determined the crystal structure of one of the two tryptophanyl-tRNA synthetases from P. falciparum. Like many other higher eukaryotes, P. falciparum has two genes encoding for TrpRS. The two gene products are predicted to be localized in the cytosol (PlasmoDB PF3D7_1336900) and in the apicoplast (PlasmoDB PF3D7_1251700) . Here we name the two proteins Pf-cTrpRS and Pf-aTrpRS, respectively, to denote their likely localization compartments. The structures reported here are of Pf-cTrpRS. Based on sequence comparison, Pf-cTrpRS is more closely related to human cytosolic TrpRS (Hs-cTrpRS, ~ 44 % identity) than to human mitochondrial TrpRS (~ 16 % identity). In addition to the Rossmann-fold and the C-terminal anticodon-binding α-helical domains common to all Class I aaRS, Pf-cTrpRS has an N-terminal extension of 300 residues. The first 228 residues appear to be apicomplexan-specific, with weak homology to the editing domain of archaeal and eukaryotic AlaRS, although the function of these residues in P. falciparum is unknown. The next 72 residues of the N-terminal extension can be aligned with the eukaryote-specific extension (ESE) of Hs-cTrpRS (Fig. 1).
We crystallized Pf-cTrpRS in its F-state and solved its structure to 2.34 Å resolution. Subsequently, tryptophan (Trp), ATP and Mg2+ were soaked into crystals which resulted in the 2.40 Å structure of the WAMP-bound P-state of this malaria parasite enzyme. Based on these two structures, which reveal considerable conformational differences, we examine the potential of this enzyme as a drug target.
Attempts to express soluble protein from the full-length Pf-cTrpRS sequence (PlasmoDB PF3D7_1336900) were unsuccessful. However, soluble expression of a protein from a construct represent residues 229 – 632 of the protein was achieved. The sequence represents the complete ESE, Rossmann-fold domain and C-terminal helical domain (Fig. 1A). While not tested for Pf-cTrpRS, the deletion of the first part of N-terminal extension did not affect the aminoacylation activity of the homologous apicomplexan Cryptosporidium parvum TrpRS . The construct was cloned into the AVA0421 vector for expression in E. coli . Protein was purified by a Ni-NTA affinity column followed by overnight cleavage of the N-terminal hexa-histidine tag using N-terminally histidine tagged 3C protease at 4°C. Cleaved protein was purified by a second Ni-NTA step and then size-exclusion chromatography on a Superdex 75 column (Amersham Pharmacia Biotech) using a buffer containing 25 mM HEPES, 500 mM NaCl, 2 mM DTT, 5% glycerol, and 0.025% NaN3 at pH 7.0. Purified protein retained five residues of the 3C protease cleavage site (GPGSM) at the N-terminus.
Crystals of Pf-cTrpRS were obtained by vapor diffusion using sitting drops equilibrated at room temperature against 50 μL of a reservoir containing 0.2 M sodium citrate and 20% PEG 3350. The drop consisted of 0.15 μL protein at 21.8 mg/mL plus 0.15 μL of the reservoir solution. Crystals grew in 1–2 days at room temperature. Crystals were cryo-protected with 20% of glycerol before freezing under liquid nitrogen for data collection. In order to determine the intermediate product WAMP-bound structure, the crystals were soaked in the cryo-protecting solution in the presence of 5 mM L-Trp, ATP and MgCl2 for 5 min before freezing.
Data for the ligand-free structure was collected from a single crystal using a wavelength of 0.98 Å at synchrotron beamline 5.0.2 of the Advanced Light Source in Berkeley, CA. Data was processed with HKL2000  (Table 1). The coordinates of TrpRS from C. parvum were used as search model for phase determination by molecular replacement using the program Phaser . The solution was fed to the model building software ARP/wARP  for automatic model rebuilding. Subsequent iterated manual building/rebuilding and refinement of models were performed using Coot  and Refmac5 , respectively. In the final cycles of refinement, protein structures were refined with translational/libration/screw (TLS) groups identified by the TLS motion determination server  before restrained refinement in Refmac5. The structure validation server MolProbity  was used throughout the process to monitor the progress of structure determination. The final crystallographic refinement statistics are given in the Table 1. Figures were created and rendered with Pymol . Superposition of structures are carried out with Pymol, which first aligns proteins by sequence followed by structural alignment using five cycles of refinement to improve the fit by discarding pairs with high relative variability .
Data for the WAMP-bound structure were collected from a single crystal soaked for 5 min with 5 mM of L-Trp, ATP and MgCl2 using a MicroMax-007 HF rotating anode (Rigaku) equipped with VariMax HF (Osmic) monochromator and a Saturn 994 (Rigaku) CCD detector at a wavelength of 1.54 Å. Data were integrated using XDS  and scaled with Scala . Coordinates of the ligand-free structure of Pf-cTrpRS were used as search model for molecular replacement in Phaser . Subsequent model building and validation process are the same as described above for the ligand-free structure.
The construct used for the structure determination contains the ESE (Ser229 to His300), the Rossmann-fold (Lys301 to Thr520 & Thr611 to Met632) and the C-terminal helical domain (Asp521 to Leu610) (Fig. 1). Pf-cTrpRS crystallized in space group P21212 with a dimer in the asymmetric unit (Fig. 2A). The structure of the ligand-free F-state was refined to 2.34 Å with good statistics (Table 1). Part of the ESE (Ser229 to Glu243), along with an insertion loop near the KMSKS motif (KMSST in Pf-cTrpRS), are disordered in both subunits. In addition, residues Gly538 to Gly550 are disordered in subunit B only. Subunits A and B superimpose with an rmsd of 0.23 Å for 299 Cα atoms and are essentially identical in structure. A search on the DALI server  showed that Pf-cTrpRS is structurally closest to Hs-cTrpRS (PDB: 2QUK) . The dimer of Pf-cTrpRS superimposes onto Hs-cTrpRS with an rmsd of 1.11 Å for 586 equivalent Cα atoms. Soaking of the crystals in Trp, ATP and Mg2+ resulted in a 2.40 Å structure with clear and strong density of WAMP in both active sites representing the P-state of the Pf-cTrpRS dimer (Fig. 2B). In their P-state, subunit A can be superimposed onto subunit B with an rmsd of 0.25 Å for 310 Cα atoms. Since in both the F structure and the P structure the two subunits A and B per asymmetric unit are essentially identical to each other, the analysis below will focus on subunits A only.
Comparing the P-state with the F-state of Pf-cTrpRS reveals large structural differences of the protein (Fig. 3A). Subunit As of the F- and P-states superimpose with an rmsd of 0.42 Å based on 283 out of 370 Cα atoms. A total of 87 residues exhibit considerable changes in conformation. Many of these residues are located in secondary structure elements spread over two domains, the Rossmann-fold and the ESE, and completely surrounding the active site. This contributes to an overall ‘collapse’ of the active site pocket in the P-state (Fig. 3A). On average, the shift of the Cα atoms of the 87 residues is 1.9 Å. The overall volume of the active site decreases from 1658 Å3 to 838 Å3 as calculated using the CASTp server .
Conformational changes of both bacterial and eukaryotic (cytoplasmic) TrpRS enzymes upon ligand binding are well documented as described above. Interestingly, structural features involved in the transitions are different. Bacterial TrpRS is postulated to move from an open, F-state to the closed, P-state via a closed, pre-transition state. The ‘closing’ of the active site pocket in this case is mainly due to an inter-domain twisting motion of the Rossmann-fold relative to the C-terminal domain, instead of an intra-domain collapsing of the active site pocket as observed in Pf-cTrpRS. This collapsing of the active site pocket of Pf-cTrpRS is replicated, albeit not completely, in eukaryote cytosolic TrpRS, specifically Hs-cTrpRS, as discussed below.
The most complete model of Hs-cTrpRS available (PDB: 1R6T)  contains in the asymmetric unit one subunit in the F-state and another subunit in the P-state. Superpositions of the F-state on the P-state of Hs-cTrpRS results in an RMSD of 0.56 Å for 323 Cα atoms, with only 39 residues (out of 362) differing significantly in position (Fig. 3B). These structures are used for comparison with Pf-cTrpRS. Many of these shifts in the human and the parasite enzyme are similar:
Another difference between WAMP binding by the parasite and the human enzyme is that two features remain disordered up WAMP binding by the parasite enzyme while these become ordered in the human enzyme [8, 40–42]:
Pf-cTrpRS has a unique insertion of 15 residues in the loop after the KMSST motif (Fig. 1), which is equivalent to the KMSAS motif in the human enzyme, and this insertion may contribute to the flexibility of the KMSST motif-containing loop in Pf-cTrpRS, in both the F- and P-states.
The key difference in the conformational changes of ligand-free parasite and human cTrpRS upon WAMP binding is that in the Pf-cTrpRS the active site is completely surrounded by shifting residues, while in Hs-cTrpRS this is only partially so (Fig. 3A vs. Fig. 3B). The motions of Hs-cTrpRS upon WAMP binding are less extensive in that movements listed above in (F), (G) and (H) for Pf-cTrpRS, do not occur in Hs-cTrpRS.
Residues that are directly in contact with WAMP are highly conserved between Pf-cTrpRS and Hs-cTrpRS. Based on a 4.5 Å radius around WAMP in Pf-cTrpRS P-state structure, 28 contact residues can be identified (Fig. 3C). Twenty one of these 28 residues are identical in Pf and Hs-cTrpRS. The remaining 7 residues are mostly conserved substitutions (Fig. 1B). Overall, the positions of contact residues of the adenine binding site appear to be slightly more variable between Pf-cTrpRS and Hs-cTrpRS but the size of the differences is limited. In several cases contact atoms are conserved even though there is a difference in the identity of the residues (e.g. Gln452 vs. Cys309, Val481 vs. Thr338). There are a few subtle differences, however, likely to decrease the size of the pocket in Pf-cTrpRS (Fig. 3C). Examples are: (i) the phenolic oxygen of Tyr425 has no equivalent in human (Phe280 is at this position); (ii) the positions of the side chain amino groups of Pf-cTrpRS Lys347 and equivalent Hs-cTrpRS Lys200 differ by 3.5 Å, but this difference is quite far from the ligand; and, (iii) Phe482 in Pf-cTrpRS is ~ 1.0 Å closer to the adenine ring of WAMP than Phe339 in Hs-cTrpRS. Overall, no striking features can be readily identified and exploited for the design of selective inhibitors based solely on the structure of the active site pockets in human and parasite enzymes.
When looking at the ligand binding region only, it seems that there are limited opportunities to arrive at selective active site inhibitors for Pf-cTrpRS. However, instead of targeting the active site, conformational changes occurring during the transition from F- to P-state as observed in the current structures (Fig. 3A,B) could well be a basis of selective inhibition by small molecules using an alternative mode of action. As discussed above, there are major differences in the transitions of states in Pf-cTrpRS and Hs-cTrpRS. The need for the conformational changes during the aminoacylation reaction could make Pf-cTrpRS susceptible to inhibitors that prevent or perturb the motion. Specific inhibitors could possibly be developed to target residues that are involved in facilitating the motions of the protein while performing its function. Away from the substrate binding sites, the conservation of residues and conformations decreases and hence the chance of specificity increases. In this case, Pf-cTrpRS and Hs-cTrpRS have 75 % identity regarding WAMP-contacting residues, but the identity decreases to 55 % among residues involved in the motion of state transitions. More importantly, as mentioned earlier, many shifts in residues located in Pf-cTrpRS are not observed in Hs-cTrpRS, forming a patch of surface that could be targeted to arrive at specific inhibitors (Fig. 3B). Admittedly, it could be challenging to design such inhibitors de novo. However, high-throughput screening of small molecule libraries or fragment-based screening using Pf-cTrpRS are powerful tools to discover inhibitory small molecules which prevent conformational changes, discoveries which could be followed up by structure-based approaches as in the case of HIV reverse transcriptase [43, 44], protein tyrosine phosphatase 1B  and fructose 1,6-bisphosphatase . The enzyme•ligand structures reported here reveal differences in conformational changes upon ligand binding by parasite and human enzyme which might inspire screening compound libraries to arrive at hits for further development as anti-malaria agents.
We thank Stewart Turley, Frank Zucker and Jonathan Kay for providing support for the X-ray data collection, database management and computing environment at the Biomolecular Structure Center of the University of Washington. We also thank the research scientist staff of MSGPP for their assistance in cloning and expression of protein. We also thank the staff of Advanced Light Source (ALS) for assistance during data collection. This work is funded by the National Institutes of Health (grants P01AI067921, (Medical Structural Genomics of Pathogenic Protozoa (MSGPP)), R56AI084004, and RO1AI084004.).
Coordinates and structure factors for Pf-cTrpRS F- and P-states are deposited in the Protein Data Bank under accession codes 4J76 and 4J75, respectively.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.