Trypanosomes are protozoan organisms that cause important veterinary diseases and two of them are responsible for significant human diseases: Trypanosoma brucei is the agent of sleeping sickness (African trypanosomiasis) and T. cruzi causes Chagas disease in the American continent. Trypanosomes diverged early in evolution from other eukaryotes and posses unique features in terms of energetic metabolism. A peroxisome-like structure, the glycosome, contains most of the glycolytic and pentose phosphate pathway enzymes. In addition, the trypanosome mitochondria present unusual structure and functional properties that strongly depend on the different environments encountered by parasites.1
Arginine kinase (AK, E.C. 188.8.131.52) is a phosphotransferase that catalyzes the interconversion between phosphoarginine and ATP. This enzyme is present in some invertebrates, including the trypanosomatids, and represents an analogous system to the creatine kinases in vertebrates. Phosphagens such as phosphoarginine and phosphocreatine play a critical role as “energy storage” molecules because the high-energy phosphate can be rapidly transferred to ADP when the renewal of ATP is needed.2 Phosphagens support bursts of cellular activity until catabolic events such as glycogenolysis, glycolysis, and oxidative phosphorylation are switched on. In T. cruzi, AK overexpression was shown to increase the parasite survival capability under pH and nutritional stress conditions3 and seems to play a critical role as a regulator of energetic reserves and cell growth. A similar situation occurs when a heterologous AK is expressed in organisms lacking phosphagen kinases, such as yeast and bacteria.4, 5 More recently, it was observed that T. cruzi epimastigotes treated with hydrogen peroxide showed a significant increase in both AK expression and survival capability during exposure, suggesting the participation of AK in the parasite response to oxidative stress.6
Although the crystal structures of several vertebrate creatine kinases are currently available, only a single AK has been structurally characterized so far, that of Limulus polyphemus (Horseshoe crab) AK.7 We have previously characterized the molecular and biochemical characterization of the T. cruzi AK (TcAK)8, 9 and we report here the crystal structure of the ligand-free (open form) of the enzyme at 1.9 Å resolution.
MATERIALS AND METHODS
A fragment carrying the TcAK gene was obtained by PCR and cloned into the pRSet A vector (Invitrogen) to obtain pRSetA-AK. Escherichia coli BL21(DE3)pLysS bacteria transformed with pRSetA-AK were grown at 37°C until late log phase in Luria-Bertani (LB) medium with 50 μg/mL ampicillin. Induction of expression was conducted for 3 h at 30°C after addition of 1 mM IPTG. The bacterial pellet was resuspended in 50 mM Hepes buffer, pH 7, 0.2M NaCl, in the presence of protease inhibitors, and sonicated. After lysate centrifugation, the supernatant was applied to a Ni-column (Pharmacia) using an FPLC system and eluted by an imidazole gradient (0–0.5M). An additional gel filtration step (Superdex-75) allowed to separate the purified protein from aggregated material, as well as to remove imidazole and most of the Ni+2-cations. The protein was subsequently concentrated to 20 mg/mL and tested for biochemical activity as described.8
Crystals of TcAK were obtained by the hanging drop method, mixing 1.5 μL of protein solution and 1.5 μL of the reservoir solution containing 2.5M ammonium sulfate, 0.1M Tris HCl, pH 7.5. Crystals grew within a week at 18°C to a maximum size of 0.2 mm × 0.2 mm × 0.3 mm. For freezing, crystals were transferred to a cryoprotectant solution containing 2.5M ammonium sulfate and 20% (v/v) glycerol. A native data set at 1.9 Å resolution was collected on beamline ID29 at the European Synchrotron Radiation Facility, and reduced using the programs MOSFLM/SCALA from the CCP4 program package10 (Table I). The 3D structure was determined by molecular replacement methods using the program AMoRe11 and the atomic coordinates of Limulus polyphemus arginine kinase (LpAK, PDB code 1M8012). Crystallographic refinement was carried out with the program REFMAC13 using TLS matrices, alternated with manual cycles of model reconstruction using COOT.14 Water molecules were added with programs Arp/Warp.15 The parameters after the final refinement cycle are given in Table I. Atomic coordinates and structure factors are available from the PDB under accession code 2J1Q.
Table I. Data Collection and Refinement Statistics
Values in parentheses apply to the high resolution shell.
. Rcryst and Rfree were calculated from the working and test reflection sets, respectively.
The 3D structure of ligand-free TcAK was determined by molecular replacement methods and refined at 1.9 Å resolution (Table I). Three regions of the protein (residues 289–293, 310–320 and the two C-terminal residues, 356/7) are disordered in the crystal and were not modelled. Also, a few solvent exposed side chains, mainly in the loop 294–299, are missing in the model. The overall structure of TcAK [Fig. 1(A,B)] is very similar to that of ligand-free Limulus polyphemus (Horseshoe crab) arginine kinase12 (LpAK, 72% sequence identity, rms deviation of 0.9–1 Å for 334 aligned residues) and to those of several eukaryotic creatine kinases, such as human type B or muscle creatine kinases (sequence identities 45–47%, rms deviations of 1.2–1.3 Å for 310–320 aligned residues).
The catalytic mechanism of phosphagen kinases has been thoroughly studied by classical methods. Despite their different substrate specificities, these enzymes are thought to share a common mechanism of direct, associative in-line γ-phosphoryl transfer,16 and this notion is reinforced by a significant conservation of AK amino acid sequences and overall structures. In particular, the sequence and length of the “specificity loop” (residues 61–68), which has been postulated to mediate substrate specificity for different guanidine kinases,7, 17 has a nearly identical sequence in TcAK (LDSGIGVY) and LpAK (LDSGVGIY), and the loop also retains a similar conformation in the two structures. Two other well-conserved residues in guanidine kinases are Cys 271, which plays an important role in substrate binding but is not essential for catalysis,18 and Glu 225, postulated as one of the two bases involved in the catalytic mechanism.19 The corresponding structural regions are also well conserved between the two structures of ligand-free AKs. On the other hand, the second proposed catalytic base, Glu 314, belongs to a disordered loop in TcAK and is missing from the model of the open form of the enzyme.
The structure of TcAK corresponds to the open form of the enzyme, in the absence of ligands. All our attempts to obtain crystals of TcAK in complex with substrates or analogs were unsuccessful. However, most residues that were seen to undergo the largest changes in LpAK upon substrate binding16 are strictly conserved in TcAK, strongly suggesting that the induced-fit conformational changes may be similar in both enzymes. Thus, the conserved residues Ser122, His185 and His 284 (involved in interactions with the nucleotide ring) and the positively charged cluster Arg124, Arg126, Arg229 and Arg280 (which attract ADP/ATP to the active site) all occupy the same conformational space in both ligand-free TcAK and LpAK. Furthermore, the loop 310–320, which closes the active site of LpAK when the enzyme forms a complex with substrate analogs7 [Fig. 1(C)], is equally flexible in both unliganded structures.
Significant conformational changes take place upon substrate binding, as illustrated by the superposition of ligand-free TcAK with the transition state analog complex of LpAK (PDB code 1BG07), which shows a rms deviation of 2.1 Å for 316 aligned residues. While AK may be considered a two domain protein, the structural comparison of ligand-free and ligand-bound AKs strongly suggests that the differences upon substrate binding can be attributed to the motion of three rigid groups with respect to a fourth, fixed, subdomain12 [Fig. 1(C)]. Since these changes occur near the phosphoryl and nucleotide acceptors, both ligands appear to be responsible for the induction of the conformational changes that close the active site. Indeed, such a mechanism could be useful to prevent wasteful hydrolysis of ATP in the absence of phosphagen.
The recently finished “Tritryp” genome projects (T. cruzi, T. brucei, and Leishmania major) confirmed that AK is present in the first two genomes but not in L. major.20–22T. cruzi has a single AK of 357 amino acids length, whereas T. brucei has three isoforms of 356 (the T. cruzi ortholog), 370, and 404 amino acids. Considering the large-scale synteny between the trypanosomatid genomes, the TcAK gene could have been acquired by horizontal gene transfer, with a subsequent gene loss event in Leishmania spp. Since AK is absent in mammals but is well conserved in the human pathogens T. cruzi and T. brucei (pairwise sequence identities of 82% between T. brucei and T. cruzi AKs), the crystal structure of the trypanosomal enzyme reported here could provide a template for the rational development of new drugs against African and American trypanosomiases.