Structural and functional insights into Erwinia carotovora L-asparaginase Anastassios C. Papageorgiou 1 , Galina A. Posypanova 2 , Charlotta S. Andersson 1 , Nikolay N. Sokolov 3 and Julya Krasotkina 3 1 Turku Centre for Biotechnology, University of Turku and A ˚ bo Akademi University, Finland 2 Institute of Molecular Medicine, Russian Academy of Medical Sciences, Moscow, Russia 3 Institute of Biomedical Chemistry, Russian Academy of Medical Sciences, Moscow, Russia l-Asparaginases (EC 3.5.1.1) are enzymes that primar- ily catalyze the conversion of l-asparagine to l-Asn and ammonia, and to a lesser extent the hydrolysis of l-Gln to l-Glu. Two types of bacterial l-asparaginase have been identified, namely type I and type II [1]. Type I asparaginases are expressed constitutively in the cytoplasm and characterized by enzymatic activity for both l-Asn and l-Gln. Type II asparaginases are expressed under anaerobic conditions in the periplas- mic space of the bacterial membranes and display high specific activity against l-Asn. Type II asparagin- ases, in particular, display antitumor activity and are used as chemotherapeutics in acute lymphoblastic leu- kemia [2–4]. The antileukemic effect of l-asparaginas- es is believed to result from the depletion of circulating Asn. Certain tumors have decreased or absent activity of Asn synthase, so they are dependent on externally supplied l-Asn for growth [5]. Following administration of l-asparaginase, the Asn blood levels are reduced, leading to selectively induced inhibition and regression of tumors. A marked reduction of l-Asn concentration in blood and tissue fluids (< 10% of the normal level) is required for effective treatment. Keywords asparaginase; crystal structure; enzyme therapy; Erwinia; leukemia treatment Correspondence A. C. Papageorgiou, Turku Centre for Biotechnology, University of Turku and A ˚ bo Akademi University, Turku 20520, Finland Fax: +358 2 3338000 Tel: +358 2 3338012 E-mail: tassos.papageorgiou@btk.fi (Received 26 May 2008, revised 22 June 2008, accepted 26 June 2008) doi:10.1111/j.1742-4658.2008.06574.x Bacterial l-asparaginases are enzymes that catalyze the hydrolysis of l-asparagine to aspartic acid. For the past 30 years, these enzymes have been used as therapeutic agents in the treatment of acute childhood lymphoblastic leukemia. Their intrinsic low-rate glutaminase activity, however, causes serious side-effects, including neurotoxicity, hepatitis, coagulopathy, and other dysfunctions. Erwinia carotovora asparaginase shows decreased glutaminase activity, so it is believed to have fewer side- effects in leukemia therapy. To gain detailed insights into the properties of E. carotovora asparaginase, combined crystallographic, thermal stability and cytotoxic experiments were performed. The crystal structure of E. carotovora l-asparaginase in the presence of l-Asp was determined at 2.5 A ˚ resolution and refined to an R cryst of 19.2 (R free = 26.6%) with good stereochemistry. Cytotoxicity measurements revealed that E. carotovora asparaginase is 30 times less toxic than the Escherichia coli enzyme against human leukemia cell lines. Moreover, denaturing experiments showed that E. carotovora asparaginase has decreased thermodynamic stability as compared to the E. coli enzyme and is rapidly inactivated in the presence of urea. On the basis of these results, we propose that E. carotovora asparaginase has limited potential as an antileukemic drug, despite its promising low glutaminase activity. Our analysis may be applicable to the therapeutic evaluation of other asparaginases as well. Abbreviations ASA, solvent-accessible surface area; CS, cell survival; EcAII, Escherichia coli periplasmic asparaginase; ErA, Erwinia chrysanthemi asparaginase; EwA, Erwinia carotovora asparaginase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide. 4306 FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS Escherichia coli periplasmic l-asparaginase (EcAII) and Erwinia chrysanthemi l-asparaginase (ErA) have been successfully used in leukemia treatment. However, their l-glutaminase side activity limits their use and causes severe side-effects as a result of l-Gln depriva- tion [6]. Serious liver disorders, acute pancreatitis, hyperglycemia, immunosuppression and other dysfunc- tions are some of the side-effects in patients receiving l-asparaginase treatment. Also, differences between E. coli and Er. chrysanthemi asparaginase in respect to toxicity and efficacy have been found [2]. All known bacterial l-asparaginases show high simi- larity in their tertiary and quaternary structures. l-Asparaginases crystallize mostly as homotetramers with 222 symmetry. Structural information on several l-asparaginases, including EcAII [7], ErA [8], Wolinel- la succinogenes asparaginase [9] and related amido- hydrolases from Acinetobacter glutaminasificans [10] and Pseudomonas fluorescence [11], is available. Recently, an X-ray structure of ErA was refined to 1.0 A ˚ resolution [12]. The enzyme monomer consists of $ 330 amino acids arranged in two domains. The active site of l-asparaginase is located between the N-terminal and C-terminal domains of two adjacent monomers. Residues responsible for ligand binding form the rigid part of the active site. The flexible part of the active site (residues 14–33 in ErA) controls access to the binding pocket and carries the catalytic nucleophile Thr15, which is highly conserved for all l-asparaginases [8]. This region is often disordered in the crystals of the enzyme, indicating high mobility of the flexible loop. Suicide inhibitors bind covalently to the primary catalytic nucleophile Thr15 and Tyr29 of ErA, freezing the flexible loop in the ‘closed’ confor- mation [13]. Further experiments have shown that, indeed, substrate binding induces rapid closure of the loop, whereas in the absence of ligands, the loop stays predominantly open [14,15]. Desired properties for therapeutic l-asparaginases include high l-asparaginase activity, low l-glutamin- ase specificity, and a long half-time in the blood- stream [16]. These criteria have directed the search for an optimal therapeutic asparaginase that started more than 30 years ago [17] but has not led to any noticeable success. As a result, many asparaginases with promising catalytic properties did not pass pre- clinical trials, because of their therapeutic inefficiency [18]. In most of the studies, the catalytic properties of the enzymes were characterized in detail, but other important parameters, such as oligomerization and thermal stability, were overlooked. Hence, a clear understanding of the catalytic activity, conformational stability and structural properties is required to pro- vide adequate and efficient criteria for evaluation and prognosis of the therapeutic efficiency of new l-aspar- aginases. Recent studies have shown that Erwinia carotovora asparaginase (EwA), similarly to EcAII, exhibits a very low glutaminase activity – about 1.5% of its l-aspara- ginase activity [19,20]. On the other hand, ErA, which is more closely related to EwA at the amino acid sequence level than to EcAII, has a glutaminase activ- ity that is approximately 10% of its asparaginase activ- ity but is better for medication, owing to less severe immunorelated side-effects [21,22]. Consequently, EwA may exhibit better therapeutic potential by combining the advantages of both EcAII and ErA. To gain further insights into EwA and to assess its suitability as a drug candidate, we have determined the crystal structure of the enzyme in the presence of Asp and carried out stability and cytotoxicity measure- ments. Unlike in previously reported crystal structures of asparaginases, a dimer of homotetramers was found in the crystals. As compared to EcAII, EwA was found to have reduced stability with and without l-Asp. Furthermore, its cytotoxic efficiency was 30 times lower than that of EcAII. Results and Discussion Quality of the structure The structure of EwA was refined to an R cryst of 19.2% (R free = 26.6%). The final model consists of eight monomers, eight l-Asp molecules, and a total of 708 water molecules. The first three residues and a gap between residues 19 and 33 could not be fully modeled in any of the eight individual monomers, due to the lack of sufficient electron density. The Ramachandran plot showed 89.2% of the non-Gly and non-Pro resi- dues in the most preferable region and 0.4% in the generously allowed region. The overall G-factor is )0.05, which is better than expected for structures determined at the same resolution according to procheck. The average B-factor for all atoms in the structure is 34.6 A ˚ 2 , close to the B-factor of the col- lected dataset calculated by Wilson plot (39.6 A ˚ 2 ). The average B-factors for each monomer are: A, 29.6 A ˚ 2 ; B, 29.0 A ˚ 2 ; C, 30.2 A ˚ 2 ; D, 28.7 A ˚ 2 ; E, 39.4 A ˚ 2 ; F, 39.3 A ˚ 2 ; G, 39.5 A ˚ 2 ; and H, 39.8 A ˚ 2 . The map cor- relation is 0.95 (Fig. S1). Overall structure The eight monomers (indexed from A to H) are arranged in two tetramers (ABCD and EFGH) in the A. C. Papageorgiou et al. Erwinia carotovora L-asparaginase FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS 4307 asymmetric unit (Fig. 1A). The rmsd values between the monomers are no higher than 0.29 A ˚ (monomer A will be taken as reference unless otherwise stated). Each of the independent monomers has the charac- teristic two-domain architecture of l-asparaginases: a large N-terminal domain characterized by an eight- stranded antiparallel mixed b-sheet, and a smaller C-terminal domain that contains a parallel b-sheet (Fig. 1B) [23]. The two domains are connected by an approximately 20 residue flexible linker in the region around residue 200. Of all residues in the linker, Thr204 adopts a stereochemistry that lies in the disal- lowed region of the Ramachandran plot, in agreement with the corresponding Thr in EcAII and ErA. Comparison with other asparaginase monomers Structure-based sequence alignment was carried out against EcAII and ErA monomers using the coordi- nates of the 3ECA [7] and 1HG1 [8] Protein Data Bank entries, respectively (Fig. 2). An rmsd of 0.42 A ˚ (78% sequence identity) and 0.88 A ˚ (49% sequence identity) for Ca atoms was found for ErA and EcAII, respectively. The symmetry found in EwA tetramers matches that of EcAII and ErA, and the overall fold of the monomers is also identical (Fig. 2B). As expected, EwA and ErA keep a nearly identical fold, whereas there are more differences between EwA and EcAII. The largest differences are found in the flexible parts of the surface, but these are less important for the overall fold and active site orientation. One struc- turally important difference found between EwA and EcAII is the presence of a single disulfide bond in EcAII, involving Cys77 and Cys105. This disulfide bond is placed near the substrate entry, possibly pro- viding extra stabilization to EcAII. Hence, site-directed mutagenesis to engineer a similar disulfide bond into EwA may alter the properties of the enzyme and improve its stability. Active site The catalytic site of the enzyme is located between the N-terminal and C-terminal domains of adjacent mono- mers (A and C; B and D). Sequence comparison indi- cates that the active site is highly conserved between EwA, ErA, and EcAII (Fig. 2A). Strong electron den- sity in the active site region was found and, due to its shape, it was assigned to a bound l-Asp. The low B-factors, around 30 A ˚ 2 , for the active site residues in monomer A indicated that there is relatively stable binding of the ligand, as also suggested by the B-factor (36.1 A ˚ 2 )ofl-Asp itself. Higher B-factors for l-Asp were found in other monomers, possibly due to less tight binding. The catalytic residues are believed to be Thr15 and Thr95 according to previous structures of l-asparaginases. Thr15 is part of the active site flexible loop comprising residues 14–33. Evaluation of contacts within the structure shows that the surrounding residues Ser62, Glu63, Asp96 and Ala120 also have significant contacts with the ligand (Table 1). Ala120, in particular, is approximately 0.3 A ˚ closer to the OD2 of l-Asp, and it may provide additional stabilization for ligand binding (Fig. 3). Hence, a bulkier residue could further reduce the size of the binding pocket, possibly leading to an additional reduction of the glu- taminase activity. In EcAII, Glu283 from monomer C is involved in a strong hydrogen bond with the nitro- gen atom of l-Asp. In contrast, only Ser254 OG from monomer C of the AC EwA dimer makes a weak con- tact (4.25 A ˚ ) with the nitrogen atom of l-Asp, and no structural equivalent of EcAII Glu283 is found in EwA and ErA, due to a deletion in their primary structure (Fig. 2A). This might explain the higher affinity of EcAII for l-Asn (K m =15lm) [15] as com- pared with EwA (K m =98lm) [19] or ErA (K m =55lm) [24]. Close inspection of this region in the crystal structures of EwA and EcAII reveals that Asp287 of EwA monomer C is approximately 5.8 A ˚ away from EcAII Glu283 and hence unable to make any contacts with l-Asp. Structural comparison with the active site of ErA shows strong conservation of the participating residues. One notable exception is EwA AB Fig. 1. (A) Ribbon diagram of EwA dimer of homotetramers. Each monomer is shown in a different color. (B) Ribbon diagram of EwA monomer. The coloring changes from blue (N-terminal) to red (C- terminal). The active site location is indicated by the bound L-Asp (sticks). The figure was drawn with PYMOL 0.99 (DeLano Scientific, Palo Alto, CA, USA). Secondary structure elements were assigned by DSSP [45]. Erwinia carotovora L-asparaginase A. C. Papageorgiou et al. 4308 FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS Gly61, which corresponds to Ala61 in ErA. Given that EcAII has also a Gly in the same position, the presence of Ala in ErA may contribute to the increased glutaminase activity of the latter as compared to EwA and EcAII, because of small conformational changes or different flexibility of the surrounding residues. Different conformations of the active site flexible loop may also contribute to the variations in substrate speci- ficity among l-asparaginases [20]. Hence, the active site flexible loop could be another good starting point for protein engineering efforts to produce l-asparaginase variants with altered substrate specificity. The EwA tetramer As the tetramer is known to be the catalytically active form of asparaginases [25], its preservation is impor- tant for the enzymatic and, consequently, therapeutic properties of the enzyme. Spatial and energetic properties of monomer interfaces in EwA and EcAII tetramers are summarized in Tables 2 and 3. The largest contact area is between monomers A and C, which participate in the active site formation. This interface corresponds to 24–28% of the dimer solvent- accessible surface area (ASA). The asparaginase monomers that do not contribute to the catalytic center formation are less tightly bound. Their pairwise contact area involves no more than 10–12% of the dimer ASA. The strength of the intersubunit binding correlates well with the interaction energy, which is also significantly higher for the AC dimers than for other dimer combinations. The comparison of EwA crystal structures in complex with l-Asp in the active site and in the absence of any ligand (Protein Data Bank code 1ZCF) reveals that l-Asp binding tightens the interaction of each monomer; for example, the A B Fig. 2. (A) Structure-based alignment of EwA, EcAII and ErA. Secondary structure elements are shown in blue, and active site residues are marked with black arrows. The figure was drawn with ESPRIPT [46]. Solvent accessibility for EwA is also shown at the bottom from dark blue (solvent-accessible) to white (buried). (B) Ca-trace structural superposition (in wall-eye stereo mode) of EwA, ErA and EcAII in gray, yellow, and red, respectively. Every 20th residue of EwA is labeled. The orientation is similar to that in Fig. 1B. A. C. Papageorgiou et al. Erwinia carotovora L-asparaginase FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS 4309 number of hydrogen bonds increases and the electro- static part of the interaction energy is reduced to half, with the exception of the EwA AD dimer. This finding implies that the thermodynamic stability of asparagin- ase in solution will benefit from the presence of l-Asp. Although EcAII and EwA do not differ significantly with respect to the geometry of the dimer interface, EwA is characterized by a considerably higher electro- static energy of monomer–monomer interactions. Although the electrostatic energy is decreased upon binding of l-Asp to EwA, it still remains approxi- mately three times higher than that of the correspond- ing EcAII monomer pairs. This observation suggests reduced thermodynamic stability of EwA and, conse- quently, a possible decreased ability of this enzyme to inhibit tumor cell growth. The hydrophobic inter- actions were not taken into consideration, although they are very important in evaluating the stability of protein–protein complexes. However, the calculation of the hydrophobic constituent in a vacuum using crystal structure is not accurate, as it essentially depends on the protein environment. Table 1. Protein–L-Asp contacts at the active site; distances from 2.3 to 3.5 A ˚ . L-Asp atoms EwA EcAII a N E63 OE1 (2.90), D96 OD2 (2.53) Q59 OE1 (2.96), D90 OD2 (3.00), N48 b ND2 (3.47), E283 b OE2 (2.46), E283 b OE1 (3.05), E283 b CD (3.12) CA T15 OG1 (3.25), D96 OD2 (3.36) T12 OG1 (3.27), V27 CG2 (3.36), E283 b OE2 (3.30) CB T15 OG1 (3.19) T12 OG1 (3.11), D90 OD1 (3.37) CG T15 CB (3.27), T15 OG1 (2.96), T95 OG1 (3.13), T N (3.31) T12 CB (3.40), T12 OG1 (2.92), T89 OG1 (2.86) OD1 T15 OG1 (3.31), T15 CB (3.37), T15 N (3.01), T95 N (2.88), G94 CA (3.00), G94 C (3.39) T12 CB (3.46), T12 OG1 (3.18), T12 N (3.19), G88 CA (2.98), G88 N (3.19), T89 N (2.56), T89 OG1 (3.40) OD2 A120 O (2.82), T15 CB (3.24), T15 OG1 (3.21), T95 OG1 (2.88) A114 O (3.14), T12 CB (3.42), T12 OG1 (3.29), T89 CB (3.39), T89 OG1 (2.42) C E63 OE1 (3.05), S62 N (3.39) S58 OG 3.36, G88 CA 3.41, S58 N (3.42) O T95 N (3.35), D96 N (3.11), S62 OG (2.65), E63 OE1 (3.32), D96 CG (3.25), D96 OD2 (3.19), D96 CB (3.21), S62 CB (3.42), G94 (3.33), S62 N (3.34) G11 CA (3.31), G88 CA (3.18), G57 CA (3.22), G57 C (3.42), S58 N (2.75) OXT G14 CA (3.27), G61 C (3.45), G61 CA (3.28), S62 N (2.75), G63 OE1 (3.15), G94 CA (3.32) Ser58 OG (2.47), Gly88 CA (3.35), G88 C (3.36), T89 N (3.25), D90 N (2.96), D90 CB (3.05), D90 CG (3.21) a Active site of monomer A in 3ECA crystal structure. b Monomer C. A B Fig. 3. Stereodiagrams (wall-eye mode) of the active site of (A) EwA and (B) EcAII (Protein Data Bank code 3ECA) [7] with bound L-Asp (in green). The active site residues involved in interactions with the substrate and major hydrogen bonds are shown (a full list of contacts is given in Table 1). Glu283 from monomer C of the EcAII homotetramer is colored yellow. Erwinia carotovora L-asparaginase A. C. Papageorgiou et al. 4310 FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS Crystal packing in the asymmetric unit EwA was found to form a dimer of homotetramers in the asymmetric unit. The intermolecular interactions between the two tetramers are mediated mainly by salt bridges found between B Asp215 and G Arg148, B Asp215 and G Lys147, and C Arg148 and F Asp215. Calculation of the ASA between the two EwA tetramers gives an estimated contact area of $ 3700 A ˚ 2 , which is in the upper range of contact areas (> 2000 A ˚ 2 ) found in protein–protein complexes [26]. The shape comple- mentarity S c [27] of the two tetramers is 0.517, which corresponds to the lower limit of S c found in protein– protein interfaces. Thus, the possibility that EwA may be able to form higher-order oligomers under physio- logical conditions cannot be ruled out according to structural considerations. Changes in the oligomeric form of l-asparaginase could be clinically important and may require further investigation. Earlier studies have shown that EcAII is able to form oligomers of dif- ferent molecular masses, depending on the enzyme con- centration and the influence of various chemicals [28]. Cytotoxicity tests EwA showed an inhibitory effect on leukemia cell growth (Fig. 4). Three human tumor cell lines com- monly used to test asparaginase cytotoxicity were employed. The enzyme concentrations used in the experiments (from 0.01 to 15 IUÆmL )1 ) covered the range of asparaginase concentration in blood after injection of a standard dose of this drug (0.5– 2IUÆmL )1 ) [29]. The viabilities of acute lymphoblastic leukemia and Burkitt’s lymphoma cells were signifi- cantly decreased after 72 h of incubation with EwA. Chronic myeloid leukemia cells were rather resistant to asparaginase treatment [18]. A significant decrease in cell viability began at an EwA concentration of 0.5 IUÆmL )1 for K562 and MOLT-4 cells, and at 0.3 IUÆmL )1 for Raji cells, and this was followed by a further linear decay of the number of surviving cells up to 0% of control. The LC 50 values were determined to be 3.26 ± 0.29 and 7.33 ± 0.42 IUÆmL )1 for Raji and MOLT-4 cells, respectively. For K562 cells, the LC 50 value was not determined, as a 50% decrease in the number of viable cells was not reached even with the highest concentration of 15 IUÆmL )1 used in the experi- ment. The concentration of 5 IUÆmL )1 marked a pla- teau on the dose-dependence curve, suggesting that the LC 50 level may not be reached at all for K562 cells. Sim- ilar results were obtained with EcAII, which was used as a reference enzyme. Only a quantitative differ- ence between EcAII and EwA antiproliferative prop- erties was found. The EcAII LC 50 values for both Raji (0.11 ± 0.005 IUÆmL )1 ) and MOLT-4 (0.27 ± 0.03 IUÆmL )1 ) cells were approximately 30 times lower than the EwA LC 50 value, suggesting that EcAII is $ 30 times more toxic than EwA. Table 3. Interaction energy (kJÆmol )1 ) of monomers in asparaginase tetramer calculated using GROMOS96 forcefield. Monomer partitioning for EcAII and EwA in complex with L-Asp corresponds to that of the 3ECA structure. Monomers A, B, C and D for EwA correspond to monomers A, B, E and F, respectively, in 1ZCF. Monomer EcAII in complex with L-Asp (3ECA) EwA in complex with L-Asp EwA (Protein Data Bank code 1ZCF) Nonbonded Electrostatic Total Nonbonded Electrostatic Total Nonbonded Electrostatic Total AB )340 )41 )381 )206 )15 )221 )343 )8 )351 AC )1037 )252 )1289 )939 )86 )1025 )999 )40 )1039 AD )434 )6 )440 )513 )2 )515 )297 )25 )322 Table 2. Geometric characterization of interfaces in asparaginase tetramers. EcAII in complex with L-Asp (3ECA) EwA in complex with L-Asp EwA free (1ZCF) ASA (A 2 ) Contact area (A 2 ) (%) Hydrogen bonds ASA (A 2 ) Contact area (A 2 ) (%) Hydrogen bonds ASA (A 2 ) Contact area (A 2 ) (%) Hydrogen bonds AB 24 267 3689 (15.2) 10 24 762 3628 (14.6) 12 26 961 3276 (12.1) 7 AC 21 945 6153 (28.0) 25 22 257 6421 (28.8) 27 25 065 6078 (24.2) 19 AD 24 508 3900 (15.9) 15 24 374 3215 (13.1) 12 27 191 2877 (10.6) 12 A. C. Papageorgiou et al. Erwinia carotovora L-asparaginase FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS 4311 Stability measurements Comparative studies of thermodynamic stabilities revealed that EwA, in contrast to EcAII, rapidly lost its activity under denaturating conditions (Fig. 5). Incuba- tion for 3 min at 35 °C led to a 60% irreversible decrease of EwA activity, whereas EcAII retained most of its initial activity after incubation under the same conditions. EwA also displayed low stability in urea solutions (Fig. 5B), and its enzymatic activity was com- pletely lost after 1 h of incubation with 2 m urea. In contrast, no decrease in activity was observed for EcAII with up to 4 m urea. Furthermore, the presence of l-Asp increased the stability of both asparaginases with respect to the denaturing effect of temperature and urea, in agreement with the structural analysis of the tetramer. Thus, the stability of EwA remained much lower than that of EcAII even in the presence of l-Asp. With all of these findings, the drastic difference between the cytotoxic activities of EcAII and EwA may be attributed to the ability of these asparaginases to maintain their catalytically active oligomeric con- formation. Indeed, the rapid loss of EwA activity 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 120 A B Residual activity (%)Residual activity (%) T (°C) 012345678 0 20 40 60 80 100 120 Urea (M) Fig. 5. Stabilities of EwA ( ) and EcAII (d), (A) at different temper- atures and (B) in urea solution. Data points represent the average value from a triplicate experiment. Data obtained in the presence of L-Asp are indicated by empty symbols. 120 100 80 60 40 20 0 120 100 80 60 40 20 0 120 100 80 60 40 20 0 0.01 Cell survival (%) Cell survival (%) Cell survival (%) 0.1 0.5 1 2 5 10 20 0.01 0.1 0.5 1 2 51020 0.01 0.1 Aspara g inase (IU.mL –1 ) Asparaginase (IU.mL –1 ) Asparaginase (IU.mL –1 ) 0.5 1 2 5 10 20 A B C Fig. 4. Antiproliferative effects of EwA (d) and EcAII ( ) on human leukemia cell growth following a 72 h incubation. (A) MOLT-4. (B) Raji. (C) K562. Data points represent the average value from a duplicated experiment calculated from at least three replicate wells. Erwinia carotovora L-asparaginase A. C. Papageorgiou et al. 4312 FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS below 30–40 °C is due probably to disruption of the quaternary tetrameric structure rather than to the unfolding of individual monomers. The same is expected for asparaginase inactivation in urea solution. Shifrin et al. [25] have shown in sedimentation velocity experiments, confirmed by fluorescent studies, that the inactivation of asparaginase in urea proceeds through tetramer dissociation. Implications The results presented here have unveiled a potential correlation between the thermodynamic stability of EwA and its cytotoxicity. Thus, two important conclu- sions can be drawn from our studies. First, thermody- namic stability appears to provide a new criterion that has been previously overlooked during the search for new therapeutic candidates in the l-asparaginase family. Second, our analysis may help to interpret the variations in the clinical profiles of asparaginase ther- apy that so far are not fully understood. For example, the recently published results of a comparative study of therapeutic l-asparaginases conducted for 7 years by the Dana-Farber Cancer Institute showed that ErA is less toxic and less efficacious than EcAII [30,31]. It is possible that ErA may dissociate rapidly and lose its activity after injection, so becoming unable to induce either a therapeutic effect or toxicity as compared to EcAII. This assumption agrees well with the essentially different half-times of these enzymes in blood (6–15 h for ErA and 30–34 h for EcAII) [18]. Further investi- gations are clearly necessary to obtain a complete understanding of the interplay of therapeutic potential and protein stability in l-asparaginases. Experimental procedures Cloning, expression and purification of EwA A plasmid pACY177 ⁄ ECAR-LANS containing a 1126 bp EwA genome fragment including the EwA gene (ansB1, Gene ID 2884179) was kindly provided by V. V. Bogush (Institute of Gene Biology, Moscow). EwA was PCR ampli- fied from this plasmid using the following forward and reverse primers: 5¢-CGT CATATGAAAAGGATGTTTAA GG-3¢ and 5¢-CCT CTCGAGATAAGCGTGGAAGTAA TCC-3¢. The NdeI and XhoI sites (underlined) were intro- duced at the 5¢-terminus and the 3¢-terminus, respectively, of the EwA gene for cloning into the pET22b vector (Novagen). The resulting pET22 ⁄ EwA plasmid was inserted into BL21 (DE3) E. coli cells after verification of construct fidelity by sequencing. Cultures were grown in LB medium supplemented with 100 lgÆmL )1 ampicillin at 37 °Cto A 600 nm = 0.8–0.9, and EwA expression was induced with 0.5 mm isopropyl thio-b-d-galactoside for 5 h. Cells were collected by centrifugation at 4000 g for 15 min and stored at )20 °C until use. All purification steps were performed at 4 °C. Cells were resuspended in water and disrupted by sonication. Then, the pH of the cell-free extract was decreased to 6.0 with 1 m NaH 2 PO 4 . Cell debris and precip- itated proteins were removed by centrifugation at 10 000 g for 20 min. The soluble fraction of the cell-free extract was loaded onto an SP-Sepharose column (1 · 8 cm) equili- brated with 20 m m sodium phosphate buffer (pH 6.0), and unbound proteins were washed off with the same buffer. EwA was eluted with 100 mL of a 0–0.8 m NaCl linear gra- dient in column buffer, desalted on a PD10 column, and concentrated to 10 mgÆmL )1 . Crystallization and data collection EwA was crystallized in the presence of l-Asp as previously described [32]. Briefly, a 10 mgÆ mL )1 protein solution con- taining 10 mml-Asp was mixed with a reservoir solution containing 16–18% poly(ethylene glycol) 4000 and 0.2 m NaF. Crystals usually appear after $ 5 days. Data were collected on station X13 at EMBL Hamburg (c ⁄ o DESY) from a single crystal at 100 K, using 20% glycerol as cryo- protectant. A MARCCD detector and a fixed wavelength (0.8043 A ˚ ) were used during data collection. In total, 900 images with an oscillation range of 0.25° and exposure time $ 30 s were collected, processed and scaled using xds [33]. Details of data processing are given in Table 4. Structure determination and refinement Initial phases were calculated with phaser [34], using a poly-alanine model of ErA tetramer (Protein Data Bank code 1HG1). Refinement was first carried out with cns [35], and at the later stages with refmac5 [36]. Refinement statistics are summarized in Table 4. Rebuilding and man- ual fitting into 2|F o |–|F c | and |F o |–|F c | difference electron density maps were performed in o [37] and coot [38]. Non- crystallographic symmetry restraints were used in the initial stages of refinement, and were gradually released at the final stages. Water molecules were added either manually or automatically by arp ⁄ warp [39]. Only water molecules with B-factors < 65 A ˚ 2 after a round of refinement were kept for subsequent refinement. Bond distances and con- tacts were analyzed by the program contact from the ccp4i program suite [40]. The geometric quality of the refined model was assessed with procheck [41]. Monomer interfaces were analyzed by cache 6.1 (http://www.fqs.pl/ chemistry/cache), and interaction energies were calculated using swiss-pdbviewer 3.7 with the gromos96 forcefield (http://www.expasy.org/spdbv). Structure-based super- position was performed with ssm [42] and structural com- parison between monomers with lsqkab [40]. A. C. Papageorgiou et al. Erwinia carotovora L-asparaginase FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS 4313 Enzymatic assay The rate of l-Asn hydrolysis by EwA was measured in a dis- continuous assay following the formation of ammonia by the Nessler method, as described elsewhere [43]. The aspara- ginase activity was expressed in International Units (IU), with 1 IU defined as 1 lmol of ammonia released per 1 min. Antiproliferative assay Human leukemia cell lines MOLT-4 (acute T-lymphoblastic leukemia), Raji (Burkitt’s lymphoma) and K-562 (human chronic myeloid leukemia) were used to test asparaginase cytotoxicity. Cultures were maintained at 37 °C in a 5% CO 2 incubator and propagated in RPMI-1640 medium (HyClone, Logan, UT, USA) supplemented with 10% (v ⁄ v) fetal bovine serum (HyClone) and 50 lgÆmL )1 gentamicin as antibiotic solution. Cells were plated in 96-well plates (Corning-Costar, Cambridge, MA, USA) at a density of 2.5 · 10 4 cellsÆmL )1 in fresh medium. The range of the final l-asparaginase con- centrations in each well was from 0.01 IUÆmL )1 to 15 IUÆmL )1 . Each test was performed in triplicate. Commer- cially produced EcAII l-asparaginase (Medac, Hamburg, Germany) was used as a reference. After 72 h of incubation under standard conditions, a colorimetric assay using the metabolizable salt 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl- tetrazolium bromide (MTT) (Sigma-Aldrich, St Louis, MO, USA) was employed to measure cell survival [44]. Two hours before the end of incubation, 50 lL of MTT (1 mgÆmL )1 )in the culture medium were added to each well. After staining, the culture medium was removed, formazan crystals pro- duced by reduction of MTT were dissolved in 100 lLof dimethylsulfoxide, and staining intensity was measured by absorption at 540 nm on an ELISA reader (Labsystems, Helsinki, Finland). Cell survival (CS) was calculated by the equation: CS = (D treated well ⁄ mean D control wells ) · 100%. The enzyme concentration lethal to 50% of cells (LC 50 ) was calculated from the fit of experimental data to the dose– response equation CSðEÞ¼A 1 þðA 2 À A 1 Þ=½1 þ 10 ðLC 50 ÀEÞ p; where A 1 and A 2 are the bottom and top asymptotes, E the enzyme concentration, and p the Hill slope. Data were processed with origin 7.0 (http://www.originlab.com). Conformational stability measurements To test asparaginase thermal stability, 10 lL of enzyme stock solution (150 IUÆmL )1 ) were added to 500 lL of pre- heated 50 mm phosphate buffer (pH 7.2), supplemented with 50 mm NaCl, mixed vigorously for a few seconds, and incubated at a given temperature for 3 min. After that, the assay mixture was immediately put on ice for 1 h for refolding. Then, the residual asparaginase activity was mea- sured under physiological temperature conditions (37 °C). The reaction was started by addition of 40 lL of 200 mm l-Asn solution in NaCl ⁄ P i , and quenched with 250 lLof 20 mm trichloroacetic acid. Empirical melting temperatures were read at the inflection points of the resulting activity versus temperature profiles. The asparaginase stability was also evaluated by denatur- ation experiments in urea solutions. In this case, 2 lLof 1000 IUÆmL )1 enzyme stock solution was mixed with 18 lL of increasing urea concentrations (from 0 to 7.2 m)in 50 mm phosphate buffer (pH 7.2). Following incubation for 1 h at room temperature (23 ° C), all 20 lL of the assay mixture were added to 530 lLof50mm phosphate buffer (pH 7.2), containing 100 mm NaCl and 15 mml-Asn. The reaction was stopped with 250 l Lof20mm trichloroacetic acid. Denaturation experiments in the presence of 2.7 mm l-Asp were also conducted. Data deposition Atomic coordinates and structure factors have been depos- ited in the Protein Data Bank (ID code 2jk0). Table 4. Data collection and refinement statistics. Data collection Space group P2 1 2 1 2 1 Cell dimensions (A ˚ ) 73.65 · 135.65 · 250.10 No. of protein molecules in the ASU 8 Resolution range (A ˚ ) 20.0–2.50 (2.60–2.50) a Temperature (K) 100 No. of observations 368 260 (23 099) No. of unique reflections 83 599 (7463) Completeness (%) 95.8 (78.0) R merge (%) 7.8 (43.8) I ⁄ sigma(I) 13.9 (3.2) Refinement statistics Resolution range (A ˚ ) 20.0–2.5 No. of reflections (working ⁄ test) 79 398 ⁄ 4197 R cryst ⁄ R free (%) 19.2 ⁄ 26.6 Water molecules 708 rmsd Bond lengths (A ˚ ) 0.009 Bond angles (º) 1.29 Average B-factors (A ˚ 2 ) Main chain 34.4 Side chain 34.5 Waters 39.4 L-Asp 50.7 (A, 36.1; B, 37.8; C, 46.6; D, 54.9; E, 62.1; F, 55.0; G, 49.7; H, 56.8) Wilson B-factor (A ˚ 2 ) 39.6 Ramachandran statistics (%) Most favored regions 89.2 Additional allowed regions 9.9 Generously allowed regions 0.5 Disallowed regions 0.4 a Numbers in parentheses correspond to the highest resolution shell. ASU, asymmetric unit. Erwinia carotovora L-asparaginase A. C. Papageorgiou et al. 4314 FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS Acknowledgements This work was supported by the Sigrid Juse ´ lius Founda- tion and the Academy of Finland (Grant No. 121278). We thank the staff at EMBL Hamburg for help with data collection. Access to EMBL Hamburg (c ⁄ o DESY) was provided by the European Community (Access to Research Infrastructure Action of the Improving Human Potential Programme to the EMBL Hamburg Outstation, contract number HPRI-CT-1999-00017). Funding from the Russian Fund for Basic Research (grant 06-04-49792) is also acknowledged. References 1 Campbell HA, Mashburn LT, Boyse EA & Old LJ (1967) Two L-asparaginases from Escherichia coli B. Their separation, purification, and antitumor activity. Biochemistry 6, 721–730. 2 Duval M, Suciu S, Ferster A, Rialland X, Nelken B, Lutz P, Benoit Y, Robert A, Manel AM, Vilmer E et al. (2002) Comparison of Escherichia coli-asparagin- ase with Erwinia-asparaginase in the treatment of child- hood lymphoid malignancies: results of a randomized European Organisation for Research and Treatment of Cancer – Children’s Leukemia Group phase 3 trial. Blood 99, 2734–2739. 3Go ¨ kbuget N & Hoelzer D (2006) Treatment of adult acute lymphoblastic leukemia. Hematol Am Soc Hema- tol Educ Program 1, 133–141. 4 Verma N, Kumar K, Kaur G & Anand S (2007) L-as- paraginase: a promising chemotherapeutic agent. Crit Rev Biotechnol 27, 45–62. 5 Holcenberg J (2005) New insights on asparaginase. J Pediatr Hematol Oncol 27, 246–247. 6 Ollenschlager G, Roth E, Linkesch W, Jansen S, Simmel A & Modder B (1988) Asparaginase-induced derangements of glutamine metabolism: the pathoge- netic basis for some drug-related side-effects. Eur J Clin Invest 18, 512–516. 7 Swain AL, Jasko ´ lski M, Housset D, Rao JK & Wloda- wer A (1993) Crystal structure of Escherichia coli L-as- paraginase, an enzyme used in cancer therapy. Proc Natl Acad Sci USA 90, 1474–1478. 8 Aghaiypour K, Wlodawer A & Lubkowski J (2001) Structural basis for the activity and substrate specificity of Erwinia chrysanthemi L-asparaginase. Biochemistry 40, 5655–5664. 9 Lubkowski J, Palm GJ, Gilliland GL, Derst C, Ro ¨ hm KH & Wlodawer A (1996) Crystal structure and amino acid sequence of Wolinella succinogenes L-asparaginase. Eur J Biochem 241, 201–207. 10 Lubkowski J, Wlodawer A, Housset D, Weber IT, Ammon HL, Murphy KC & Swain AL (1994) Refined crystal structure of Acinetobacter glutaminasificans glutaminase-asparaginase. Acta Crystallogr D Biol Crystallogr 50, 826–832. 11 Lubkowski J, Wlodawer A, Ammon HL, Copeland TD & Swain AL (1994) Structural characterization of Pseu- domonas 7A glutaminase-asparaginase. Biochemistry 33, 10257–10265. 12 Lubkowski J, Dauter M, Aghaiypour K, Wlodawer A & Dauter Z (2003) Atomic resolution structure of Erwi- nia chrysanthemi L-asparaginase. Acta Crystallogr D Biol Crystallogr 59, 84–92. 13 Ortlund E, Lacount MW, Lewinski K & Lebioda L (2000) Reactions of Pseudomonas 7A glutaminase-aspar- aginase with diazo analogues of glutamine and aspara- gine result in unexpected covalent inhibitions and suggests an unusual catalytic triad Thr-Tyr-Glu. Biochemistry 39, 1199–1204. 14 Aung HP, Bocola M, Schleper S & Rohm KH (2000) Dynamics of a mobile loop at the active site of Escheri- chia coli asparaginase. Biochim Biophys Acta 1481, 349– 359. 15 Derst C, Henseling J & Ro ¨ hm KH (2000) Engineering the substrate specificity of Escherichia coli asparaginase. II. Selective reduction of glutaminase activity by amino acid replacements at position 248. Protein Sci 9, 2009– 2017. 16 Avramis VI & Tiwari PN (2006) Asparaginase (native ASNase or pegylated ASNase) in the treatment of acute lymphoblastic leukemia. Int J Nanomedicine 1, 241–254. 17 Distasio JA, Niederman RA, Kafkewitz D & Goodman D (1976) Purification and characterization of L-aspara- ginase with anti-lymphoma activity from Vibrio succin- ogenes. J Biol Chem 251, 6929–6933. 18 Narta UK, Kanwar SS & Azmi W (2007) Pharmacolog- ical and clinical evaluation of L-asparaginase in the treatment of leukemia. Crit Rev Oncol Hematol 61, 208–221. 19 Krasotkina J, Borisova AA, Gervaziev YV & Sokolov NN (2004) One-step purification and kinetic properties of the recombinant L-asparaginase from Erwinia caroto- vora. Biotechnol Appl Biochem 39, 215–221. 20 Kotzia GA & Labrou NE (2005) Cloning, expression and characterisation of Erwinia carotovora L-asparagin- ase. J Biotechnol 119, 309–323. 21 Carlsson H, Stockelberg D, Tengborn L, Braide I, Carne- skog J & Kutti J (1995) Effects of Erwinia-asparaginase on the coagulation system. Eur J Haematol 55, 289–293. 22 Eden OB, Shaw MP, Lilleyman JS & Richards S (1990) Non-randomised study comparing toxicity of Escheri- chia coli and Erwinia asparaginase in children with leukaemia. Med Pediatr Oncol 18, 497–502. 23 Miller M, Rao JK, Wlodawer A & Gribskov MR (1993) A left-handed crossover involved in amidohydro- lase catalysis. Crystal structure of Erwinia chrysanthemi L-asparaginase with bound L-aspartate. FEBS Lett 328, 275–279. A. C. Papageorgiou et al. Erwinia carotovora L-asparaginase FEBS Journal 275 (2008) 4306–4316 ª 2008 The Authors Journal compilation ª 2008 FEBS 4315 [...].. .Erwinia carotovora L-asparaginase A C Papageorgiou et al 24 Kotzia GA & Labrou NE (2007) L-Asparaginase from Erwinia Chrysanthemi 3937: cloning, expression and characterization J Biotechnol 127, 657–669 25 Shifrin S, Luborsky SW & Grochowski BJ (1971) L-Asparaginase from Escherichia coli B Physicochemical studies of the dissociation... Krasotkina J, Kuchumova A, Sokolov NN & Papageorgiou AC (2005) Crystallization and preliminary crystallographic analysis of L-asparaginase from Erwinia carotovora Acta Crystallogr F Struct Biol Cryst Commun 61, 407–409 33 Kabsch W (1993) Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants J Appl Crystallogr 26, 795–800 34 McCoy AJ, Grosse-Kunstleve... 210, 37–43 44 Denizot F & Lang R (1986) Rapid colorimetric assay for cell growth and survival Modifications to the tetrazolium dye procedure giving improved sensitivity and reliability J Immunol Methods 89, 271–277 45 Kabsch W & Sander C (1983) Dictionary of protein secondary structure: pattern recognition of hydrogenbonded and geometrical features Biopolymers 22, 2577– 2637 46 Gouet P, Courcelle E, Stuart... (1994) Improved methods for building protein models in electron density maps and the location of errors in these models Acta Crystallogr A 42, 140–149 38 Emsley P & Cowtan K (2004) Coot: model-building tools for molecular graphics Acta Crystallogr D Biol Crystallogr 60, 2126–2132 39 Morris RJ, Perrakis A & Lamzin VS (2003) ARP ⁄ wARP and automatic interpretation of protein electron density maps Methods Enzymol... Escherichia coli B Biochim Biophys Acta Protein Struct 194, 161– 169 29 Vieira Pinheiro JP, Wenner K, Escherich G, LanversKaminsky C, Wurthwein G, Janka-Schaub G & Boos J (2006) Serum asparaginase activities and asparagine concentrations in the cerebrospinal fluid after a single infusion of 2,500 IU ⁄ m(2) PEG asparaginase in children with ALL treated according to protocol COALL-06-97 Pediatr Blood Cancer 46,... at 1 sigma level The Figure was drawn with pymol This supporting information can be found in the online version of this article Please note: Blackwell Publishing are not responsible for the content or functionality of any supporting information supplied by the authors Any queries (other than missing material) should be directed to the corresponding author for the article FEBS Journal 275 (2008) 4306–4316 . Structural and functional insights into Erwinia carotovora L-asparaginase Anastassios C. Papageorgiou 1 , Galina A. Posypanova 2 , Charlotta S. Andersson 1 , Nikolay N. Sokolov 3 and Julya. both EcAII and ErA. To gain further insights into EwA and to assess its suitability as a drug candidate, we have determined the crystal structure of the enzyme in the presence of Asp and carried. periplasmic l-asparaginase (EcAII) and Erwinia chrysanthemi l-asparaginase (ErA) have been successfully used in leukemia treatment. However, their l-glutaminase side activity limits their use and causes