Engineering of a monomeric and low-glycosylated form of human butyrylcholinesterase Expression, puri®cation, characterization and crystallization Florian Nachon 1 , Yvain Nicolet 2 , Nathalie Viguie  1 , Patrick Masson 1 , Juan C. Fontecilla-Camps 2 and Oksana Lockridge 3 1 Centre de Recherches du Service de Sante  des Arme  es, Unite  d'Enzymologie, La Tronche, France; 2 Laboratoire de Cristallographie et Cristalloge  ne Á se des Prote  ines, Institut de biologie structurale ÔJ.P. EbelÕ, Grenoble, France; 3 University of Nebraska Medical Center, Eppley Research Institute, Omaha, NE, USA Human butyrylcholinesterase (BChE; EC 3.1.1.8) is of particular interest because i t hydrolyzes or scavenges a w ide range of toxic compounds including cocaine, organophos- phorus pesticides and nerve agents. The relative contribution of each N-linked glycan for the solubility, the stability and the secretion of the enzyme was investigated. A recombinan t monomeric BChE lacking four out of nine N-glycosylation sites and the C-terminal oligomerization domain was stably expressed as a monomer in CHO cells. The puri®ed recom- binant BChE showed catalytic p roperties similar t o those o f the native enzyme. Tetragonal crystals s uitable for X-ray crystallography studies were obtained; they were improved by recrystallization and found to diract t o 2.0 A Ê resolutio n using s ynchrotron radiation. The crystals belong to the tetragonal s pace group I422 with unit c ell dimensions a  b  154.7 A Ê ,c 124.9 A Ê , giving a V m of 2.73 A Ê 3 per Da (estimated 60% solvent) for a single molecule of recombinant BChE i n the asymmetric unit. The crystal structure of butyrylcholinesterase will h elp elucidate unsolved issues concerning cholinesterase mechanisms in general. Keywords: butyrylcholinesterase; crystallization; N-glycosy- lation; site-directed mutagenesis; X-ray diraction. Acetylcholinesterase (AChE; EC 3.1.1.7) and butyrylcho- linesterase (BChE; EC 3.1.1.8) are closely related serine hydrolases with different s ubstrate speci®city and inhibitor sensitivity. AChE terminates the action of the neurotrans- mitter acetylcholine at postsynaptic membranes and neuro- muscular junctions. Altho ugh BChE i s found in various vertebrate tissues (liver, intestine, lung, heart, muscle, brain, serum), its physiological role remains undetermined. How- ever, plasma BChE is o f pharmacological and t oxicological importance because it hydrolyzes ester-containing dru gs such as succinylcholine and cocaine. Consequently, puri®ed BChE has been used for treatment of succinylcholine- induced apnea in humans [1] and it is known to protect rodents from the toxic effects of cocaine [2,3]. To improve the rate of hydrolysis of cocaine, a mutated enzyme has been designed [4]. However, a h igher catalytic rate may be necessary if BC hE is to be used therapeutically in severe cocaine overdoses. Human BChE is also kno wn to be a good scavenger of organophosphorus (OP) pesticides and chemical warfare nerve agents [5]. For example, injections of puri®ed BChE as pretreatment against nerve agent poisoning in mice, rats and guinea pigs increased their survival with a higher ef®ciency than the c lassical pretreatment w ith pyridostigmine [ 6±8]. Similar observations have been reported for monkeys [9,10]. Mutants of human BChE (G117H) capable of hydrolyzing OP have also been designed [11]; however, their catalytic mechanism is unclear [12]. It is noteworthy that the equivalent human AChE mutant (G122H) did not acquire OP hydrolase activity ( Lockridge, O. & B artels, C.F. unpublished results). Thus, BChE could be used i n t he near future for OP decontamination, pretreatment and treatment of OP poisoning. Progress in engineering of BChE is currently limited by the l ack o f a three-dimensional structure. three-dimensional models of human BChE have been built by homology to the Torpedo californica acetylcholinesterase X-ray structure [13,14]. Although these models contributed to t he under- standing of some aspects of the difference in speci®city between AChE and B ChE, they are not satisfactory for enzyme engineering. The crystal structure of human BChE is expected to provide new insights into unsolved issues such as allosteric modulation o f cholinesterase activity (BChE presents substrate activation, whereas AChE has substrate inhibition) or the traf®c of s ubstrate, products, and water molecule s in and ou t of the act ive site gorg e [15,16]. Correspondence to F. Nachon, Centre de Recherches du Service de Sante  des A rme  es, Unite  d'enzymologie, 24 Avenue des Maquis du Gre  sivaudan, BP 87±38702 La Tronche Ce  dex, France. Fax:+33476636961,Tel.:+33476636988, E-mail: ¯orian@nachon.net Abbreviations: A C hE, acetylcholinesterase; BChE, butyrylcholinest- erase, CCD, charge coupled device; ChE, cholinesterase; C HO, Chi- nese hamster ovary; DMEM, Dulbecco's modi®ed Eagle's medium; Nbs 2 ,5,5¢-dithiobis-2-nitrobenzoic acid; HEK, human embryonic kidney cells; OP, organophosphorus ester. (Received 6 A ugust 2 001, revised 19 November 2 001, accepted 20 November 2001) Eur. J. Biochem. 269, 630±637 (2002) Ó FEBS 2002 During the past decade, the crystallization of puri®ed plasma BChE has not been successful, despite an exhaus- tive screening program in one of our laboratories. Human BChE is a heavily glycosylated homotetramer of 340 kDa with nine N-glycosylation sites per catalytic subunit representing almost 25% of its m ass [17,18]. It i s known that the glycan m oieties o ften perturb crystallization [19,20]. Human BChE o ligosaccharides, which are o f the complex biantennary type [21,22], could shield the protein surface and prevent o r reduce favorable crystal contacts. Therefore, several attempts to deglycosylate the n ative enzyme were made. Chemical deglycosylation with tri¯u- oroacetic acid, w hich was successfully used on horseradish peroxidase [23], as well as enzymatic partial deglycosyla- tion using neuraminidase and galactosidase (Masson, P. unpublished r esults) led to aggregation. Due t o the presence of fucose residues, enzymatic deglycosylation using large amounts of recombinant GST±N-glycosi- dase F fusion p rotein [24] was not ef®cient except under mild denaturing conditi ons. Thus, we decided to investi- gate the effects of the suppression of N-glycosylation sites to produce a low-glycosylated recombinant BChE suitable for crystallization. MATERIALS AND METHODS Mutagenesis 4sugOff 17/455/481/486 BChE D was obtained by PCR u sing Pfu polymerase. Carbohydrate attachment sites at N17, N455, N481, and N486 were deleted by mutating Asn residues to Gln residues. The t etramerization domain at t he C-terminus of BChE was deleted by placing a stop codon at position 530 [25,26]. The s top codon deleted 4 5 amino acids from t he C-terminus to yield a protein containing 529 amino acids and six carbohydrate chains. PCR fragments were cloned into the expression plasmid pGS and resequenced to con®rm that only the desired mutations were present. Plasmid p GS has the CMV p romoter and rat glutamine synthetase for selection. Other mutants from which carbohydrate a ttachment sites were deleted were also constucted by PCR. In each case, a codon for Asn was replaced by a c odon for Gln. The expression plasmid pGS was suitable for both transient and stable expression. Transient expression BChE mutants w ere t ransiently express ed i n human embryonic kidney cell line 293T/17, used with permission from D. Baltimore (Rockefeller University of New York; ATCC No CRL 11268). Cells were grown to 80±90% con¯uence i n 100 mm dishes and then transfected by calcium phosphate co-precipitation of 20 lg plasmid DNA per dish. Four days after transfection, the culture medium [5% f etal bovine serum in Dulbecco's modi®ed Eagle's medium (DMEM)] was harvested for a BChE activity assay. Each mutant BChE was transfected into ®ve dishes. Large scale production of recombinant human BChE 4sugOff 17/455/481/486 BChE D inpGSwasexpressedinCHO cells and stably transfected as previously described [11]. Selective p ressure to retain the plasmid was provided by 25 l M methionine sulfoximine. Secreted BChE was collected into serum-free and glutamine-free culture medium, Ultra- culture (BioWhittaker, Walkersville, MD, USA; catalogue no. 12±725B), thus avoiding contamination by AChE pre- sent in fetal bovine serum. No a ntibiotics were added to the culture medium. The cells were grown in 1-L roller bottles. The culture medium (150 mL per bottle) in th e roller bottles was changed every 2±4 days. A roller bottle yielded enzyme continuously for as long as 6 months. Each L of culture medium contained 3±5 mg of 4sugOff 17/455/481/486 BChE D . Puri®cation of 4sugOff 17/455/481/486 BChE D Units of activity are expressed as lmoles of substrate hydrolyzed per minute. Protein concentration was estimated from absorbance at 280 nm (E 1%  18). A speci®c activity of 720 Uámg )1 , measured at 25 °Cwith1m M butyrylthi- ocholine in 0.1 M potassium phosphate pH 7.0, was the standard for 100% pure native B ChE. All puri®cation steps were conducted at 4 °C. Serum-free culture medium was collected from roller bottles over a period o f 6 months. Twenty-six liters of culture medium containing 100 mg of 4sugOff 17/455/481/486 BChE D were loaded onto 400 mL o f procainamide± Sepharose p acked in a XK50/30 Pharmacia column (diameter, 5 cm; ¯ow rate of 1 Láh )1 ). The column was washed with 20 m M potassium phosphate, pH 7.0, 1 m M EDTA (until D 280  0)andthenwith0.1,0.2and0.3 M NaCl in buffer. The BChE activity was eluted with buffer containing 0.3 M NaCl and 0.1 M N(Me) 4 Br. The eluted enzyme was 21% pure as judged from speci®c activity. Then, the 4sugOff 17/455/481/486 BChE D was dialyzed against 20 m M Tris/HCl pH 7.4, and loaded onto 400 mL of DE52 anion exchanger (Whatman; catalog no. 4057200, purchased from Fisher Scienti®c) packed in Pharmacia C26/100 column. The column was washed with 20 m M Tris/HCl pH 7.4 until D 280  0. BChE was eluted with a NaCl gradient (0±0.5 M NaCl in 1 L buffer); 80% of the BChE activity was recovered. The cleanest fractions ( 80% pure) were load ed directly on to a 1 0-mL procainamide±Sepharose column packed in Pharmacia C10/20 (0.9 c m d iameter ´ 16 cm). The c olumn was washed with 2 L of 20 m M Tris/HCl pH 7.4. 4sugOff 17/455/481/486 BChE D (9.3 mg, 6740 U; 98% pure) was eluted with 400 mL of 0.6 M NaCl in 20 m M Tris/HCl pH 7.4, then dialyzed against 5 m M Mes pH 6.5 and concentrated to 10 mg ámL )1 (7200 UámL )1 ) in an Amicon Dia¯o appa- ratus with a PM10 membrane. The dialyzed, concentrated sample was ®ltered through a 0.2- lm®lterandstoredat4 °C. Determination of kinetic parameters Hydrolysis of butyrylthiocholine iodide at 25 °Cwas measured at concentrations ranging from 0.010 to 50 m M according to the method of Ellman [27]. The buffer w as 0.1 M sodium phosphate at pH 7.0 and contained 0.1 mgámL )1 Nbs 2 and 0.1% BSA. The active sites were titrated by the method of residual activity u sing diisopropyl phosphoro ¯uoridate (DFP) as titrant [28]. Kinetic parameters (k cat , K m , K ss , b factor) were deter- mined by nonlinear ®tting of the apparent rate vs. [S] using the equation described by Radic et al.[29]. Ó FEBS 2002 Butyrylcholinesterase designed for crystallization (Eur. J. Biochem. 269) 631 Crystallization A home-made sparse m atrix kit similar to the one described by Jancarik & Kim [30] was u sed to s creen for initial crystallization conditions in a hanging drop system [20]. BChE crystallized at a conce ntration of 6 .6 mgámL )1 from a 0.1- M Mes buffer solution, pH 6.5 at 20 °C, containing 2.05±2.15 M (NH 4 ) 2 SO 4 (Fluka) and using drops of 3 lL andaproteintoreservoirratioof1:2(v/v).Crystalsgrew in about 1 week. Their quality was improved using the recrystallization procedure described by Kryger [31]. Catalytic activity in the crystals A recombinant BChE crystal grown at pH 6.5 was washed twice for 5 min in a 100 lLdropof0.1 M Mes pH 6.5 buffer containing 2.4 M (NH 4 ) 2 SO 4 . Then the crystal was soaked in a 20-lL drop of the same buffer containing 0.1 mgámL )1 Nbs 2 and 5 m M butyrylthiocholine iodide (Sigma). The change in crystal coloration (turning yellow) was followed under a binocular magnifying glass. No spontaneous hydrolysis of the substrate i n t he soaking liquor was observed when monitored by spectrophotometry at 412 nm. Data collection Diffraction data were collected at k  0.932 A Ê wavelength to 2.0 A Ê resolution at the ID14-eh2 beamline of the European Synchrotron Radiation Facility with a MAR- Research CCD detector. To prevent ice formation, crystals were soaked for a few minutes in a 2.4- M (NH 4 ) 2 SO 4 , 15% glycerol, 0.1 M Mes pH 6 .5 buffer just before ¯ash-cooling at 100 K in a nitrogen stream. Collected data were indexed, integrated and reduced using MOSFLM and SCALA from the CCP4 suite [32]. RESULTS AND DISCUSSION Engineering of a low-glycosylated truncated BChE Glycosylation in¯uences the folding, secretion, stability, and solubility of ChE as well as the clearance of their plasmatic forms [22,33±35]. Heavy glycosylation of BChE contributes to its long residence time in blood circulation and protects it against proteolysis. For example, the glycosylation patterns may change with the tissue localiza- tion, but do not seem to play a critical role in the catalytic properties of the enzyme [36]. Our goal was to favor the crystallization of BChE by designing an enzyme with the fewest possible glycosylation sites, while preserving its solubility, stability a nd functional properties. Amino-acid sequences of AChE and BChE from different species were aligned to pinpoint the conserved N-glycosylation sites (Table 1). BChEs are generally more glycosylated than AChEs. AChEs from different species contain three to six N-glycosylation s ites, t hree of which are conserved in BChE. Therefore, our ®rst attempt w as to construct a recombinant BChE containing only these three glycosyla- tion sites (positions 256, 341 and 455). This was achieved by mutating six Asn residues in Asn-X-Ser/Thr recognition sites into Gln residues. These studies overlooked the possibility that a muta- tion of Asn486 might unmask a glycosylation site at Asn485. Three gly cosylation r ecognition s ites are present in the sequence N 481 ETQNNSTS 489 , but the peptide sequencing of human BChE showed that positions 481 and 486 were glyc osylated, and position 485 was not [18]. Because Asn485 and Asn486 are adjacent, the nongly- cosylation of Asn485 may be due to steric hindrance. Therefore, w e assume that all of th e constructs with the double m utation N 481Q/N486Q should be glycosylated at position 485. Table 1. C omparison of the N-glycosylation positions for various cholinesterases. Enzyme Potential N-glycosylation sites (human BChE numbering) 17 57 106 241 256 341 455 481 486 Others BChE human X X X XXXXXX monkey X X X XXXXXX cat XXXXXXXX tiger X X XXXXXX rabbit X X XXXXXX mouse X X X XXXX horse X X XXXXXX rat XXX XXXX1 AChE human X X X cat X X X rabbit X X X mouse X X X cow X X X X Torpedo ®sh X X 2 rat X X X Bungarus X X X 2 eel X X X X 2 zebra®sh X X X X 2 632 F. Nachon et al. (Eur. J. Biochem. 269) Ó FEBS 2002 A recombinant BChE containing the three conserved glycosylation sites, N256, N341 and N455 plus the one unmasked at position 485, was transiently expressed in 293T cells. Unfortunately, the expression level in the culture medium was  10-fold lower than f or the n ative enzyme (Table 2, s ix sites off). T he suppression of seven o r nine sites yielded poor expression levels as well (Table 2, seven and nine sites off) due to retention of the protein inside the cell, as shown by Western blotting. As the expression level of these c lones w as not high enough to p roduce large amounts of BChE, new constructs were tested in which the N-glycosylation sites were suppressed empirically. Suppression of sites N481 and N486 (Table 2, two s ites off) led to 45% higher expression levels than native BChE. Suppression of sites N455, N481 and 486 led to a 15% greater expression level than the native enzyme (Table 2, three s ites off). When an additional site was suppressed at position N256, the expression level was similar to t hat of the native enzyme (Table 2, four sites off; oligomeric domain: Ôyes Õ). The additional N 341Q mutation resulted in a ®vefold lower active enzyme (Tables 2, ®ve sites off). Consequently, the ® ve glycosylation sites mutant was not used any further. Interestingly, the N341 s ite is also conserved i n Candida rugosa lipase, where it plays an important role in the stabilization of the open con formation of the enzyme [37]. Such a role has not yet been observed in cholinesterases. The tetramerization domain is located at the C-termini of AChE and B ChE. In human BChE, this domain comprises 40 amino acids, e ncoded by exon 4. I ts deletion leads to higher levels of secretion into the culture medium and expression of monomers [ 25]. Crystallization o f monomeric cholinesterases is more favorable t han for oligomeric forms, even if they form a noncovalent dimer by association of a four-helix bundle ( helices 383±372 and 526±543; human Table 2. I n¯uence of the number and position of N-glycosylation sites on the expression level of secreted human BChE. The presence or absence of the oligomerization domain at the C-terminus is indicated by yes or no. Transient transfection in 293T cells was repeated in ®ve dishes. The r elative expression unit corresponds to 0 .2 lmol butyrylthiocholine hydrolyzed p er minute. Number sites o Oligomeric domain Potential N-glycosylation sites 17 57 106 241 256 341 455 481 485 486 Relative expression level 0 a Yes XXXXXX X1 a 2 Yes X X X X X X X X 1.45 3 Yes X X X X X X X 1.15 4YesXXXXX X1 4 No XXXX X X 7.2 4 b No XXXXXX X 6 b 5 Yes X X X X 0.2 6 Yes X X X X X X 0.1 7 Yes X X X 0.1 9 Yes X 0.025 a Wild-type human BchE. b Clone chosen for crystallization trials. Fig. 1. Alignment of t he amino-acid sequences of 4sugO 17/455/481/486 BChE D , human BChE, and crystallized forms of human AChE, mouse AChE and Torpedo c alifornica AChE. 4sugO 17/455/481/486 BChE D (Rec. B ChE), human B ChE [18], human AChE [44], mouse AChE [38] an d T. californica AChE (Torc a AChE) [45] w ere aligned using CLUSTALW . Asterisks denote identity, and full s tops show high similarity. Ó FEBS 2002 Butyrylcholinesterase designed for crystallization (Eur. J. Biochem. 269) 633 AChE numbering) under the protein concentrations used for crystallization [31,38]. Therefore, a truncated BChE lacking both the tetramerization domain and the N256, N455, N481 and N 486 N-glycosylation sites was c on- structed. Deletion of the tetramerization domain was achieved by introducing a stop codon at position 530 according to Blong et al.[25].Asexpected,activity measured in culture media was about sevenfold higher for the monomeric form (BChE D ) than for the o ligomeric form (Tables 2, four sites off; oligomeric domain: ÔnoÕ). In another effort, a second truncated clone also lacking four N-glycosylation sites (N17, N455, N481 and N486) was constructed. The activity level of this enzyme was slightly lower than t hat of t he previous clone but suf®ciently h igh to produce signi®cant amounts of enzyme. Consequently, this clone (4sugOff 17/455/481/486 BChE D ) was chosen for large scale expression. Figure 1 shows how the amino-acid sequence of 4sugOff 17/455/481/ 486 BChE D compares to the native human BChE enzyme and to th e crystallized forms of Torpedo californica, human and mouse AChEs. According to this alignment, the X- ray structure of Torpedo californica AChE should provid e a good probe model to solve the structure of 4sugOff 17/455/481/486 BChE D by a molecular- replacement procedure. Preparation of 4sugOff 17/455/481/486 BChE D The mutated BChE D cloned into the pGS e xpression vector, that expresses Gln-synthethase for selection pu rposes, was transfected into CHO cells. Stable clones secreting high levels of recombinant BCh E D were selected for large-scale production. Puri®cation was carried out by anion-exchange and af®nity chromatography. Axelsen et al. reported that decamethonium, used during the last af®nity chromatogra- phy step of T. californica AChE, was present in t he crystals despite extensive dialysis of the puri®ed enzyme [39]. Thus, to avoid contamination by a ligand, NaCl was used for elution of BChE f rom af®nity chromatography gels. T he purity of the ®nal enzyme preparation was estimated to be greater than 98% based on its speci®c activity and the presence of a single band on SDS/PAGE. Characterization of 4sugOff 17/455/481/486 BChE D The kinetics of butyrylthiocholine hydrolysis by recombi- nant BChE under standard conditions (0.1 M phosphate buffer, pH 7.0) can b e d escribed by the model of Radic [29]. The kinetic parameters are very close to the values reported previously for the native BChE [40], with k cat  28 000 min )1 and K m  25.6  0.4 l M (n  3). The native enzyme and recombinant BChE display similar substrate activation with K ss  510  35 l M (n  3) and b factor  2.85  0.15 (n  3). Thus, t he catalytic properties of the recombinant enzyme can be considered to be the same as t he plasma enzyme. SDS/PAGE analysis of the puri®ed recombinant BChE monomer displayed a single broad band in the 70±75 kDa molecular mass range. In contrast, the puri®ed plasma BChE showed a faint band at 170 kDa (nonreducible dimer) and a major broad band at 85 kDa (monomer) under r educing conditions (Fig. 2A). The apparent molec- ular mass of the r ecombinant monomer is consistent with the expected molecular mass for the truncated BChE after the d eletion of 45 residues a t t he C-terminal sequence and Fig. 2. Ge l e lectrophoresis analysis of 4sugO 17/455/481/486 BChE D (Rec) and human native BChE (Nat). (A) SD S/PAGE (4.5% stacking/10% separating) was carried out under reducing conditions according to Laemmli [46] using the Biorad MiniProtean II gel system and Coomassie blue staining. (B) Isoelec trofocusing g el was carried out on a Pharmacia Phast System using Phast gel (4±6.5; pH r ange) and silver staining [4 7]. 634 F. Nachon et al. (Eur. J. Biochem. 269) Ó FEBS 2002 four N-glycosylation sites. The broadness of the band suggests t hat t he puri®ed 4sugOff 17/455/481/486 BChE D still displays a signi®cant glycosylation-related heterogeneity. This issue was addressed using IEF analysis. The carbohy- drate chains of BChE are partly capped by sialic acids [22], which directly in¯uence the pI of the enzyme. Whereas plasma BChE displayed a continuous smear e xtending from pH 4.0 t o 5.7, thus re¯ecting h igh sialylation h eterogeneity, 4sugOff 17/455/481/486 BChE D displayed 10 well-resolved bands between p H 5.0 and 6.5 (Fig. 2B). T his was a de®nite improvement of the enzyme homogeneity, and encouraged us to start BChE crystallization trials. Crystallization of 4sugOff 17/455/481/486 BChE D and data collection Initial crystallization conditions were screened according to Jancarick & Kim [30] using the hanging drop method. Tetragonal crystals appeared within 1 week in a pH 6.5 0.1 M Mes buffer solution containing 2.1 M (NH 4 ) 2 SO 4 (Fig. 3A). Interestingly, these crystals appear morphologi- cally similar to t he crystals of fully glycosylate d equine serum BChE obtained in 1944 [41]. However these BChE crystals were not f urther characterized due to technical limitations at that time. To check whe ther the crystallized recombinant BChE was still active, one crystal w as soaked in Ellman's b uffer containing 5 m M butyrylthiocholine and 2.4 M (NH 4 ) 2 SO 4 . This higher concentration of precipitant was necessary to avoid the dissolution of the crystal. After a few minutes, the colorless crystal turned yellow, the color of the product of the Ellman's reaction (Fig. 3B). The crystalline enzyme seems to be s uf®ciently ¯exible to d isplay an observable catalytic activity, and small molecules such as butyrylthi- ocholine, Nbs 2 and the product of the Ellman's reaction may easily diffuse in a short period of time inside and outside the crystals. However w e cannot rule out the possibility that t he substrate might have been hydrolyzed by the protein located in the crystal surface, which is likely to solubilize during the soaking experiment. The crystals that m easured up to 0 .3 mm in their longest dimension diffracted to 2.2±2.3 A Ê resolution at 100 K, using 15% glycerol (v/v) as a cryoprotectant, and synchrotron radiation at the ESRF ID14-eh1 beamline. As recrystalli- zation improved the quality of human AChE crystals [31], we reproduced the procedure by transferring crystal- containing drops over reservoirs of water until the crystals dissolved. The drops were then placed over the original reservoir solution, or a solution with slightly lower precip- itant concentration, for r ecrystallization. As reported for human AChE, these new crystals were fewer but larger with longest dimensions of up to 0.6 mm. They diffracted to 2.0 A Ê at 100 K, using 15% glycerol (v/v) as a cryoprotec- tant, and synchrotron radiation at the E SRF ID14-eh2 beamline. Analysis of the c ollected data (Table 3) indicated that BChE crystals belong to the tetragonal space group I422 with unit cell dimensions a  b  154.7 A Ê , c  127.9 A Ê , giving a V m of 2.73 A Ê 3 per Da (estimated 60% solvent) for a crystal containing a single molecule of recombinant BChE (  70 kDa) per asymmetric unit [ 42]. A total of 371 832 observations were obtained at 2.0 A Ê resolution giving  49 298 unique re¯ections (98.6% com- plete, R sym  0.073). The structure has been successfully solved by molecular replacement s tarting from the model of native T. californica AChE, PDB code 2ace [43]. T he re®nement of the model is underway. In summary, a recombinant human butyrylcholinesterase suitable for crystallization has been constructed by sup- pressing four out of nine N-glycosylation sites and deleting its oligomerization domain. Large amounts of pure recom- binant enzyme were obtained by expression in CHO cells and puri®cation by anion-exchange and af®nity chroma- tographies. The recombinant enzyme s howed less heteroge- neity than the natural f orm while conserving identical catalytical properties. Crystals were grown at pH 6.5 using (NH 4 ) 2 SO 4 as the p recipitant. A fter their quality was improved by recrystallization, they diffracted to 2.0 A Ê resolution. The ®rst three-dimensional structure of a butyrylcholinesterase is expected to improve our knowledge regarding ChE mechanism, such as allosteric modulation, product clearance outside th e active site go rge and motion of water molecules. Moreover, the three-dimensional struc- ture of human BChE sho uld provide a template for the design of new mutants capable of hydrolyzing nerve agents and drugs such as cocaine with increased ef®ciency. ACKNOWLEDGEMENTS This work was supported by the US Army Medical Research and Materiel Command under contract DAMD 17-97-1-7349 to O. L. and Fig. 3. Te tragonal crystals of 4sugO 17/455/481/486 BChE D . (A) The larger crystal has dimension of 0.5 ´ 0.5 ´ 0.3 mm 3 . (B) Crystal after a 10-min soaking in Ellman's buer with precipitant and 5 mM butyryl- thiocholine. Table 3. D ata collection and processing. Values fo r the high est reso- lution shell a re giv en in parentheses. Space group I422 Unit-cell parameters a  b  154.66 A Ê , c  127.89 A Ê a  b  c  90° X-ray source ESRF Beamline ID14-eh2 Wavelength 0.933 A Ê Diraction limit 2.0 A Ê No. of measured re¯ections 371 832 No. of unique re¯ections 49 298 Highest resolution shell 2.1 ® 2.0 A Ê Completeness 98.6% (99.1%) Multiplicity 7.1 (6.4) R sym (on I) 0.073% (0.431%) I/r 6.5 (1.7) b Factor average 30.65 A Ê 2 Ó FEBS 2002 Butyrylcholinesterase designed for crystallization (Eur. J. Biochem. 269) 635 the De  le  gation Ge  ne  rale de l'Armement under contract DGA/DSP/ STTC-PEA 990802/99 CO 029 (ODCA, Washington, DC, 00-2-032-0- 00) to P. M. W e thank, respectively, Hassan Belrhali and Jo anne McCarthy for the opportunity to collect data at the ID14-eh1 and ID14-eh2 be amline at the ESRF in G renoble. REFERENCES 1. Viby-Mogensen, J. (1981) Succinylcholine neuromuscular block- ade in subjects h eterozygous for abnormal plasma cholinesterase. Anesthesiology 55, 231±235. 2. Homan, R.S., Morasco, R. & Goldfrank, L.R. (1996) Admin- istration of puri®ed human plasma cholinesterase protects against cocaine toxicity in mice. J. Toxicol. Clin. Toxicol. 34 , 2 59±266. 3. Lynch, T.J., Mattes, C.E., Singh, A ., Bradley, R.M., Brady, R.O. & Dret chen, K.L. (1997) Cocaine detoxi®cation by human plasma butyrylcholinesterase. Toxicol. Appl. P harmacol. 145, 3 63±371. 4. Xie, W., Altamirano, C.V., Bartels, C.F., Speirs, R.J., Cashman, J.R. & Lockridge, O. (1999) An im proved cocaine hydrolase: the A328Y mutant o f human butyrylcholinesterase is 4 -fold more ecient. M ol. Pharmacol. 55 , 83±91. 5. Lockridge, O. & Masson, P. (2000) Pesticides and susceptible populations: p eople with b utyrylcholinesterase genetic variants may b e at risk. Neurotoxicology. 21, 113±126. 6. Raveh,L.,Grunwald,J.,Marcus,D.,Papier,Y.,Cohen,E.& Ashani, Y. (1993) Human butyrylcholinesterase as a general prophylactic antidote for nerve agen t toxicity. In vitro and in vivo quantitative characterization. Biochem. Pharmacol. 45, 2465±2474. 7. Ashani, Y., Shapira, S., Levy, D., Wolfe, A.D., Doctor, B.P. & Raveh, L. (1991) Butyrylcholinesterase and acetylcholinesterase prophylaxis against soman poisoning in mice. Biochem. Pharma- col. 41, 37±41. 8. Allon, N., Raveh, L., Gilat, E., Cohen, E., Grunwald, J. & Ashani, Y. (1998) Prophylaxis against soman i nhalation toxicity i n guinea pigs by pretreatment alone with human serum butyrylcholinest- erase. Toxicol. Sci. 43, 1 21±128. 9. Broom®eld, C.A., Maxwell, D.M., Solana, R.P., Castro, C.A., Finger, A.V. & Lenz, D.E. (1991) Protection by butyrylcholin- esterase against organophosphorus poisoning in nonhuman pri- mates. J. Pharmacol. Exp. Ther. 259, 6 33±638. 10. Wolfe, A.D., Blick, D.W., Murphy, M.R., Miller, S.A., Gentry, M.K., Hartgraves, S.L. & Doctor, B.P. (1992) Use of cholinest- erases as pretreatment drugs for the protection of r hesus monkeys against soman toxicity. Toxicol. Appl. Pharmacol. 11 7, 1 89±193. 11. Lockridge, O., Blong, R.M., Masson, P., Froment, M.T., Millard, C.B. & Broom®eld, C.A. ( 1997) A s ingle amino acid substitution, Gly117His, confers phosphotriesterase (organophosphorus acid anhydride hydrolase) activity on h uman butyrylcholinesterase. Biochemistry. 36 , 7 86±795. 12. Albaret,C.,Masson,P.,Broom®eld,C.A.,ElKaim,L.&Fortier, P.L. (1998) Mechanical a spects of t he phosphotriesterase activity of human b utyrylcholinesterase G117H mutant. In Structure and Function of Cholinesterases and Related Proteins (Doctor, B.P., ed.), pp. 3 39±405. Plenum Press, New Y ork, NY. 13. Millard, C.B. & Broom®eld, C.A. (1992) A computer model of glycosylated human butyrylcholinesterase. Biochem. Biophys. Res. Commun. 189, 1 280±1286. 14.Harel,M.,Sussman,J.L.,Krejci,E.,Bon,S.,Chanal,P., Massoulie, J. & Silman, I. (1992) Conversion of acetylcholinest- erase to butyrylcholinesterase: modeling and mutagenesis. Proc. Natl Aca d. Sci. USA 89 , 1082 7±10831. 15. Gilson, M.K., Straatsma, T.P., Mccammon, J.A., Ripoll, D.R., Faerman, C.H ., Axelsen, P.H., Silman, I. & Sussman, J .L. (1994) Open back door in a m olecular dynamics simulation of ac etyl- cholinesterase. Science 263, 1276±1278. 16. Henchman,R.,Tai,K.,Shen,T.B.,Bo È rjesson, U., Philippopou- los, M. & Mccammon, J.A. (2001) A Structural and Dynamical Analysis of Water in Acetylcholinesterase (Paper presented at the American Chemical Society National Meeting 2001). The Amer- ican Chemical Society, Sa n D iego, C A, U SA. 17. Lockridge, O., Eckerson, H.W. & Du La, B.N. (1979) Interchain disul®de bonds and subunit organization in human serum chol- inestera se . J. Biol. Chem. 254, 8324±8330. 18. Lockridge, O., Bartels, C.F., Vaughan, T.A., Wong, C.K., Norton, S.E. & Johnson, L.L. (1987) Complete amino acid seque- nce of human serum cholinesterase. J. Biol. Chem. 262, 549 ±557. 19. Davis, S.J., Puklavec, M.J., Ashford, D.A., Harlos, K., Jones, E.Y., Stuart, D.I. & Williams, A.F. (1993) Expression of soluble recombinant glycoproteins with prede®ned g lycosylation: appli- cation to the crystallization of the T-cell glycoprot ein CD2. Protein Eng. 6, 229±232. 20. Mcpherson, A. (1982) Preparation and Analysis of Protein Crys- tals. Jo hn W iley & Sons, N ew York, NY. 21. Ohkura, T., Hada, T., Higashino, K., Ohue, T., Kochibe, N., Koide, N. & Y amashita, K. ( 1994) Increase of fucosylated serum cholinesterase in relation to high risk groups for hepatocellular carcinomas. Canc er Res. 54, 55±61. 22. Saxena, A., Raveh, L., Ashani, Y. & Doctor, B.P. (1997) Structure of glycan moieties responsible for the extended circulatory life time of fetal bovine serum acetylcholinesterase and equine serum butyrylcholinesterase. Biochemistry 36, 7 481±7489. 23. Tams, J.W. & Welinder, K.G. (19 95) Mild chemical deglycosyla- tion o f horseradish p eroxidase yields a fully active , homogeneous enzyme. Anal. Biochem. 228, 48±55. 24. Grueninger-Leitch, F., D'Arcy, A., D'Arcy, B. & Chene, C. (1996) Deglycosylation of proteins for crystallization u sing recombinant fusion protein glycosidases. Protein Sci. 5, 2617±2622. 25. Blong, R.M., Bedows, E. & Lockridge, O. (1997) Tetramerization domain of human butyrylcholinesterase is at the C-terminus. Biochem. J. 32 7, 7 47±757. 26. Altamirano, C.V. & Lockridge, O. (1999) Conserved aromatic residues of the C-terminus of hu man butyrylcholinesterase mediate t he association of tetramers. Biochemistry 38, 13414± 13422. 27. Ellman, G.L., Courtney, K .D., Andres, J.V. & Featherstone, R.M. (1961) A new and rapid colorimetric determination of acetylcholinestera se activity. Biochem. Pharm acol. 7, 88±95. 28. Leuzinger, W. (1971) The number of catalytic sites in acetyl- cholinesterase. Biochem. J. 23 6 , 139±141. 29. Radic, Z., P ickering, N .A., Vellom, D.C., Camp, S. & Taylor, P. (1993) Three distinct domains in the cholinesterase molecule confer selectivity for acetyl- and butyrylcholinesterase inhibitors. Biochemistry 32, 12074±12084. 30. Jancarik, J. & Kim, S.H. (1991) Sparse matrix sampling: a screening method for crystallization of proteins. J. Appl. Crystal- logr. 24, 409±411. 31. Kryger, G., Harel, M., Giles, K., Toker, L., Velan, B., Lazar, A., Kronman, C., Barak, D., Ariel, N., Shaerman, A., Silman, I. & Sussman, J.L. (2000) Structures of r ecombinan t native and E202Q mutant h uman acetylcholinesterase comp lexed with the snake-venom toxin fasciculin-II. Acta Crystallogr. D56, 1385±1394. 32. Collaborative Computational Computer Project 4. . (1994) The CCP4 suite: programs for protein crystallography. Acta Crystal- logr. D50, 760±763. 33. Kronman, C., Chitlaru, T., Elhanany, E., Velan, B. & Shaerman, A. (2000) Hierarchy of post-translational modi®cations involved in the circulatory longevity of glycoproteins. Demonstration of concerted contributions of glycan sialylation and subunit assembly to the pharmacokinetic behavior of bovine acetylcholinesterase. J. Biol. C hem. 275 , 2 9488±29502. 34. Wang, C., Eufemi, M., Turano, C. & Giartosio, A. ( 1996) In¯u- ence of the carbohydrate m oiety on the stability o f glycoproteins. Biochemistry 35, 7299±7307. 636 F. Nachon et al. (Eur. J. Biochem. 269) Ó FEBS 2002 35. Kronman, C., Velan, B., Marcus, D., Ordentlich, A., Reuveny, S. & Shaerman, A. (1995) Involvement of oligomerization, N-glycosylation a nd sialylation in the clearance o f cholinesterases from the circulation. Bioc h em. J. 311, 9 59±967. 36. Liao, J., Heider, H., Sun, M.C. & Brodbeck, U. (1992) Dierent glycosylation i n acetylcholinesterases f rom mammalian brain and erythrocytes. J. Neurochem. 58 , 123 0±1238. 37. Brocca, S., Persson, M., Wehtje, E., A dlercreutz, P., Alberghina, L. & Lotti, M. (2000) Mutants provide evidence of the importance of glycosydic chains in the activation of lipase 1 from Candida rugosa. Pr otein Sci. 9, 985±990. 38. Marchot, P., Ravelli, R.B., Raves, M.L., Bourne, Y., Vellom, D.C., Kanter, J., C amp, S., Sussman, J.L. & Taylor, P. (1996) Soluble monomeric acetylc holinesterase from mouse: expression, puri®cation, and crystallization in complex with fasciculin. Protein Sci. 5, 672±679. 39. Axelsen, P.H., Harel, M., Silman, I. & Sussman, J.L. ( 1994) Structure and dynamics of the ac tive site gorge of acetylcholin- esterase: synergistic use of molecular dynamics simulation and X-ray crystallography. Protein S ci. 3, 188±197. 40. Masson, P., F roment, M.T., Bartels, C.F. & L ockridge, O. (1996) Asp70 in t he peripheral anionic site of human butyrylcho linest- erase. Eur. J. Biochem. 235, 36±48. 41. Baden, R., Schutz, F. & Stacey, M. (1944) A crystalline serum mucoprotein w ith h igh c holinesterase a ctivity. Natu re 154, 183±184. 42. Matthews, B.W. (1968) Solvent content of protein crystals. J. Mol. Biol. 33, 491±497. 43. Raves, M.L., Harel, M., Pang, Y.P., Silman, I., Kozikowski , A.P. & Sussman, J.L. (1997) Structure of acetylcholinesterase com- plexed with the nootropic alkaloid, (±)-huperzin e A. Nat. Struct. Biol. 4, 57±63. 44. Soreq,H.,Ben-Aziz,R.,Prody,C.A.,Seidman,S.,Gnatt,A., Neville, L., Lieman -Hurwitz, J., Lev- Lehma n, E., Ginzb erg, D. & Lipidot-Lifson, Y. (1990) Mo lecular cloning a nd construction of the coding region for human acetylcholinesterase reveals a G + C-rich attenuating structure. Proc. Natl Acad. Sci. USA 87 , 9688± 9692. 45. Schumacher, M., Camp, S., Maulet, Y., Newton, M., Macphee- Quigley, K., T aylor, S.S., Friedmann, T. & Taylor, P. ( 1986) Primary structure of Torpedo californica acetylcholinesterase deduced f rom its cDNA sequence. Nature 319 , 407 ±409. 46. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of t he h ead of bacteriophage T4. Nature 227, 680±685. 47. Morrissey, J.H. (198 1) Silve r stain for proteins in polyacrylamide gels: a modi®ed procedure with enhanced uniform sen sitivity. Anal. Biochem. 117, 307±310. Ó FEBS 2002 Butyrylcholinesterase designed for crystallization (Eur. J. Biochem. 269) 637 . Engineering of a monomeric and low-glycosylated form of human butyrylcholinesterase Expression, puri®cation, characterization and crystallization Florian. that positions 481 and 486 were glyc osylated, and position 485 was not [18]. Because Asn485 and Asn486 are adjacent, the nongly- cosylation of Asn485 may