Báo cáo Y học: Refolding of the Escherichia coli expressed extracellular domain of a7 nicotinic acetylcholine receptor Cys116 mutation diminishes aggregation and stabilizes the b structure pdf

Refolding of the Escherichia coli expressed extracellular domain of a7 nicotinic acetylcholine receptor Cys116 mutation diminishes aggregation and stabilizes the b structure Victor I. Tsetlin 1 , Natalia I. Dergousova 1 , Ekaterina A. Azeeva 1 , Elena V. Kryukova 1 , Irina A. Kudelina 1 , Elena D. Shibanova 1 , Igor E. Kasheverov 1 and Christoph Methfessel 2 1 Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, Moscow, Russia; 2 Central Research Biophysics, Bayer, Leverkusen, Germany Heterologous expression of the extracellular domains (ECDs) of the nicotinic acetylcholine receptor (AChR) subunits may give large amounts of proteins for studying the functional and spatial characteristics of their ligand-binding sites. The ECD of the a7 subunit of the homo-oligomeric a7 neuronal AChR appears to be a more suitable object than the ECDs of other heteromeric neuronal or muscle-type AChRs. The rat a7 ECDs (amino-acid residues  1–210) were recently expressed in Escherichia coli as fusion proteins with maltose-binding protein [Fischer, M., Corringer, P., Schott, K., Bacher, A. & Changeux, J. (2001) Proc. Natl Acad.Sci.USA98, 3567–3570] and glutathione S-trans- ferase (GST) [Utkin, Y., Kukhtina, V., Kryukova, E., Chiodini, F., Bertrand, D., Methfessel, C. & Tsetlin, V. (2001) J. Biol. Chem. 276, 15810–15815]. However, these proteins exist in solution mostly as high-molecular mass aggregates rather than monomers or oligomers. In the pre- sent work it is found that refolding of GST–a7-(1–208) protein in the presence of 0.1% SDS considerably decreases the formation of high-molecular mass aggregates. The C116S mutation in the a7moietywasfoundtofurther decrease the aggregation and to increase the stability of protein solutions. This mutation slightly increased the affinity of the protein for a-bungarotoxin (from K d  300 to 150 n M ). Gel-permeation HPLC was used to isolate the monomeric form of the GST–a7-(1–208) protein and its mutant almost devoid of SDS. CD spectra revealed that the C116S mutation considerably increased the content of b structure and made it more stable under different condi- tions. The monomeric C116S mutant appears promising both for further structural studies and as a starting material for preparing the a7 ECD in an oligomeric form. Keywords: a7 nicotinic acetylcholine receptor; extracellular domains; expression in Escherichia coli; Cys116 mutation; CD spectroscopy. Nicotinic acetylcholine receptors (AChRs), belonging to the family of ligand-gated ion channels, are divided into two major groups: muscle-type and neuronal receptors. The muscle-type AChR from the electric organ of the Torpedo ray has been studied most comprehensively. It is composed of five subunits, namely two a subunits, and one b, c,andd subunits. Mammalian muscle AChRs are similar, the only difference being the presence of e subunits in the mature receptors instead of c in the fetal versions. Neuronal AChRs are composed of a subunits (a2–a10) and b subunits (b2– b4), either as hetero-oligomers of a/b combinations or as homo-oligomers, like pentaoligomeric a7 AChRs (reviewed in [1–3]). The current ideas on the spatial organization of the whole family of nicotinic AChRs are based mainly on the electron microscopy data for the Torpedo AChR. The subunits are arranged pseudosymmetrically along the central axis form- ing a channel. A general shape of the molecule and its disposition in the membrane were established by electron microscopy two decades ago [4]. Recent cryo-electron microscopy data provided a better view of the channel structure and of the extracellular portions. In particular, a resolution of 4.6 A ˚ has been achieved for the extracellular domain (ECD) moiety of the membrane-bound Torpedo AChR [5]. This domain accommodates the binding sites for agonists and competitive antagonists. Biochemistry and molecular biology data also suggest that all other nicotinic AChRs should have a structural/functional organization similar to that of the Torpedo AChR (reviewed in [1–3]). In principle, heterologous expression of ECDs may provide sufficient amounts of proteins necessary for establishing their high-resolution spatial structure with the aid of X-ray analysis or NMR. The ECD (amino-acid residues  1–210) of the mouse muscle and Torpedo a subunits were obtained by heterolo- gous expression in mammalian cells [6] and in E. coli [7–9], respectively. The secondary structure of these proteins was determined by CD spectroscopy. However, these proteins apriorilack the contacts with other non-a subunits that also participate in forming the ligand-binding sites in the intact receptors. This drawback has been overcome by expressing the ECDs of all Torpedo subunits that assemble into a Correspondence to V. I. Tsetlin, Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, 16/10 Miklukho-Maklaya Str., V-437 Moscow GSP-7, 117997 Russia. Fax/Tel.: + 7095 335 57 33, E-mail: vits@ibch.ru Abbreviations: AChR, acetylcholine receptor; AChBP, acetylcholine- binding protein; ECD, extracellular domain; MBP, maltose-binding protein; GST, glutathione S-transferase; aBgt, a-bungarotoxin; IPTG, isopropyl thio-b- D -galactoside; GdnHCl, guanidine hydrochloride. (Received 6 December 2001, revised 8 April 2002, accepted 29 April 2002) Eur. J. Biochem. 269, 2801–2809 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02961.x pentaoligomer in a baculovirus system [10]. 1 The usefulness of the ÔdomainÕ approach became even more clear after solving the X-ray structure of the acetylcholine-binding protein (AChBP) from Lymneae stagnalis 2 [11,12], a major breakthrough in the field. This water-soluble protein has the size of the AChR subunit ECD and forms homopentamers. In view of this finding, it seems especially interesting to prepare for structural analysis the ECD of the a7 AChR, i.e. of the receptor known to function as a homopentamer [13]. In fact, there are already several publications on the a7ECD heterologous expression. The chicken a7 ECD (including also the transmembrane fragment M1) was isolated after expression in Xenopus oocytes [14]. It formed an oligomer, but the protein was obtained only in minute amounts. Large amounts of the rat a7 ECD (residues 1–196) were obtained recently in E. coli as a water-soluble fusion protein with the N-terminally attached maltose-binding protein (MBP) [15]. However, the major obstacle for structural studies appears to be the presence of aggregates with molecular masses considerably exceeding that of pentaoligomers. A similar problem was encountered in our laboratory with rat a7 ECD (residues 1–208) fused to glutathione S-transferase (GST) [16] or to MBP. Here we report optimization of refolding conditions, which for the GST–a7-(1–208) protein resulted in the diminished tendency for aggregation. The aggregation could be further decreased and, under certain conditions completely prevented, by blocking free SH group(s) in the refolded protein or by mutating the Cys116 residue. It allowed us to obtain the monomeric forms at relatively high concentrations. This more stable C116S monomer appears to be useful for structural studies and may serve as an intermediate on the way to an oligopentamer. As a first step, we determined its secondary structure by CD spectroscopy and found a predominance of b structure, similar to what was found in the extracellular moiety of the intact Torpedo AChR [5], muscle a subunits [6–9] and what is characteristic of AChBP [12]. EXPERIMENTAL PROCEDURES Construction of expression vectors The gene encoding the rat a7 AChR ECD was amplified by PCR using the a7 cDNA cloned in the pBS SK(–) (provided by H J. Kreienkamp) as template and primers 1/2 for GSTa7-(1–208) and 5/6 for MBP-a7-(1–208) (see Fig. 1). The PCR products were digested with appropriate restric- tion enzymes (BamHI and HindIII for GST fusion proteins; XmnIandBamHI for MBP fusion protein), gel-purified using the QIAquick gel extraction kit (Qiagen) and ligated with linearized pGEX-KG vector (Amersham Pharmacia Biotech) for GST fusion proteins and pMAL-c2 vector (New England Biolabs) for MBP-a7-(1–208). The ligation reaction mixtures were used to transform E. coli cells JM109 for GST fusion proteins and TB1 cells for MBP-a7- (1–208). Potential clones were screened with colony PCR, and the presence of the insert was confirmed by restriction analysis. To obtain the C116S mutant of the fusion protein GST– a7-(1–208), designated GST–a7m-(1–208), site-directed mutagenesis was performed in a two step PCR. The forward mutagenic primer 3 (see Fig. 1) contained an SspI restriction site. The codon CAG for Gln117 was exchanged for CAA and the codon CTC for Leu119 for TTG to generate an SspI restriction site. The reverse primer 4 also contained an SspI restriction site. Each construct was verified by DNA sequencing. Proteins expression, purification and refolding GST fusion proteins. JM109 cells carrying appropriate constructs were grown in Luria–Bertani medium with ampicillin (100 lgÆmL )1 )at37°C (about 3 h). When the D 600 reached a value of 0.4–0.6, isopropyl thio- b- D -galactoside (IPTG) was added to final concentration of 0.3 m M , and the bacteria were further cultured for 3 h at 37 °C. Cells were pelleted by centrifugation (15 min, 10 000 g) and stored at )20 °C. Both fusion proteins GST–a7-(1–208) and its C116S mutant GST–a7m-(1–208) were found in the inclusion bodies. The cells were suspended in 10 m M Tris/HCl, 150 m M NaCl, 1 m M EDTA, pH 7.5, and then disrupted by sonication (10 pulses by 30 s, 10 °C). After centrifugation (10 min, 14 000 g), the inclusion bodies were washed intensively three times with the above buffer containing in a series 0.1% Triton X-100, 2 M NaCl, and then 2 M urea. At each washing step the pellet was resuspended by sonication. Finally, the pellet was washed with 10 m M Tris/HCl, pH 7.5, harvested by centrifugation (15 min, 20 000 g) and stored at )20 °C. Partially purified inclusion bodies were solubilized in 50 m M Tris/HCl, 8 M urea or 6 M GdnHCl, 10 m M dithiothreitol at room temperature, with gentle stirring overnight. The concentra- tion of fusion protein at this step was no more than 1mgÆmL )1 . Denaturing and reducing agents were removed by dialysis against 20 m M Tris/HCl, pH 8.0, at 10 °C, 24 h. Three different conditions were tested for refolding of the GST fusion proteins: in the presence of 0.1% Chaps, 0.1% SDS, or without any detergent in the protein solution and in dialysis buffer. The concentration of each detergent was Fig. 1. Schematic representation of recombinant fusion proteins con- taining the rat a7ECD.The positions of the disulfide bridges in GST and a7 ECD, as well as of the unpaired Cys116 in a7, are indicated under the rectangles. 1–6, primers used for preparing the respective cDNAs (reconstituted restriction enzyme recognition sites are in bold, italics indicate the mutation sites). 2802 V. I. Tsetlin et al. (Eur. J. Biochem. 269) Ó FEBS 2002 below critical micellar concentration. Oxidation of the fully reduced protein was performed on air 3 ,at7°Candwas monitored by titration of free SH groups [17]. If a precipitate was formed, it was removed by centrifugation (15 min, 20 000 g), and the protein solution was stored at 4 °C. GST was expressed in E. coli strain BL21 as a soluble protein, and was purified by chromatography on glutathi- one–agarose as recommended by manufacturer (Amersham Pharmacia Biotech). The purified protein was denatured in 50 m M Tris/HCl, 8 M urea, 10 m M dithiothreitol and refolded under the conditions described above. MBP fusion protein. Different E. coli strains were used to optimize the MBP-a7-(1–208) expression. The best results were obtained with a protease-deficient strain CAG597. The cells with recombinant plasmid were grown in Luria– Bertani medium with 0.2% glucose and ampicillin (100 lgÆmL )1 )at37°C (about 6 h). IPTG was added to final concentration of 0.3 m M (D 600 ¼ 0.4–0.6), and the bacteria were further cultured overnight at 30 °C. Cells were pelleted by centrifugation (10 min, 10 000 g) and stored at )20 °C. MBP-a7-(1–208) was expressed mostly as a soluble protein, smaller amounts being detected by SDS/PAGE in the inclusion bodies. MBP-a7-(1–208) was purified by chromatography on amylose resin as recommended by manufacturer (New England Biolabs), using 20 m M maltose for eluting the fusion protein. After purification, the protein obtained was unstable. However, it could be stored for over a month at 4 °C in the following buffer: 20 m M Tris/HCl, pH 7.5, 200 m M NaCl, 1 m M EDTA, 0.05% octyl-b- D -glucoside, 1m M sodium azide, 1 l M pepstatin A, 10 l M leupeptin, 10 l M chymostatin, 10 l M antipain, 10 l M bestatin, 1 m M phenylmethanesulfonyl fluoride. MBP was expressed and purified under the same conditions. With the same protocol, the expression of the C116S mutant of MBP-a7-(1–208) was carried out to give mostly a soluble protein. However, in contrast to all other expressed products, this mutated protein was found considerably more toxic for E. coli and the level of expression was very low; insufficient amounts of the purified MBP-a7m-(1–208) protein were obtained to perform further studies. Determination of protein concentration Protein concentrations were determined using the Bradford Protein assay (Bio-Rad) with bovine serum albumin as reference, and by UV spectra. Extinction coefficients (k 278 nm) for each fusion protein were determined as a sum of all the extinction coefficients of the protein aromatic amino acids. SDS/PAGE and Western blotting All protein samples were analyzed by SDS/PAGE according to Laemmli [18] in a 12% gel. Samples were prepared under denaturating reducing conditions (by boiling in a sample buffer containing 1% 2-mercaptoethanol and 2% SDS) or under nonreducing conditions (without boiling and with no 2-mercaptoethanol in a sample buffer containing 0.1% SDS). For Western blotting, the proteins were transferred from unstained SDS gels to a nitrocellulose membrane using a TransBlot SemiDry Electrophoretic Transfer (Amersham Pharmacia Biotech) at 30 V in 2 h. After blocking with 1% bovine serum albumin in phosphate buffered saline, pH 8.2, membranes were incubated with 125 I-labelled a-cobratoxin from the Naja kaouthia cobra venom at 10 °Covernight. Unbound toxin was removed by washing, and labeled proteins were detected by autoradiography. Gel-permeation HPLC Gel-permeation HPLC was carried out on a Super- dex 200 HR column (10 · 300 m M ; Amersham Pharmacia Biotech) by isocratic elution at a flow rate 0.5 mLÆmin )1 in 20 m M Tris/HCl buffer, pH 8.0, containing 150 m M NaCl, either in the presence of 0.1% detergent (SDS or Chaps) or in the absence of any detergent. The column was calibrated with standard proteins dissolved in the same buffers as the expressed proteins under examination. Binding experiments 125 I-Labelling of a-bungarotoxin (aBgt) and a-cobratoxin (for Western blotting experiments) was carried out with the chloramine T method followed by desalting on a G15 Sephadex column in 50 m M Tris/HCl buffer, pH 8.0, as described previously [19]. Equilibrium binding was analyzed on anion-exchange filters DE81 [20] in a fast filtration modification [21]. Various amounts of 125 I-labelled aBgt (from 1.5 to 70 pmol, specific radioactivity  25 CiÆmmol )1 ) were incubated with 20 pmol of different proteins for 2 h at room temperature in 50 lLof50m M Tris/HCl buffer, pH 8.0. The final concentration of detergents in the reaction mixtures was 0.004 or 0.06% for the proteins refolded in 0.1% SDS or 0.1% Chaps, respectively. The SDS concen- tration was determined in reaction with p-rosaniline chloride [22]. Nonspecific binding was determined by preincubation for 1 h of the expressed products with a 100-fold molar excess of a-cobratoxin isolated from the Naja kaouthia cobra venom. As the additional controls, binding experi- ments with the expressed GST or MBP were carried out under the same conditions. Incubation mixtures were applied on DE81 filters presoaked in 50 m M Tris/HCl buffer, pH 8.0, containing 0.1% Triton X-100, and quickly washed under vacuum with 5 mL of the same buffer. The analysis of binding experiments was carried out using ORIGIN v5.0 (MicroCal Software, Inc.). GST–a7-(1–208) fusion protein modification with N -ethylmaleimide A 0.5-mL volume of refolded fusion protein GST–a7- (1–208) solution (0.2 mgÆmL )1 )in20 m M Tris/HCl, pH 8.0, was incubated with 10 m M dithiothreitol for 3 h at room temperature. The reduced protein was dialyzed against the same buffer without dithiothreitol for 20 h at 4 °Cand oxidized by air. A decrease in the free SH content was monitored by the Ellman’s method. The protein solution containing 1.0 ± 0.1 SH per mol protein was obtained. N-ethylmaleimide was added to a final concentration of 1m M , incubated for 15 min at room temperature, and then the excess of N-ethylmaleimide was removed by dialysis. The residual content of SH groups was found to be less than 0.15 mol SH per mol protein. Ó FEBS 2002 Mutation and refolding of a7 extracellular domain (Eur. J. Biochem. 269) 2803 CD spectroscopy CD spectra were recorded on a JASCO J-500A spectro- polarimeter (Japan). The results were expressed as molar ellipticity, [Q](degÆcm 2 Ædmol )1 ), with the average mean amino-acid residue weight (MRW) of 115. The molar ellipticity was determined as [Q] ¼ Q·100MRWÆc )1 Æl )1 , where c is the protein concentration in mgÆmL )1 , l is the light path length in centimeters, and Q is the measured ellipticity in degrees at a wavelength k. The instrument was calibrated with (+)-10-camphorsulfonic acid, assuming [Q] 291 ¼ 7820 degÆcm 2 Ædmol )1 [23]. Secondary structure was calculated according to the program CONTIN for globular proteins [24]. RESULTS Expression in E. coli and refolding of the GST– a7-(1–208) protein, its C116S mutant, and MBP-a7-(1–208) protein These fusion proteins are schematically depicted in Fig. 1. Figure 2 shows the SDS/PAGE analyses in reducing conditions of the (GST)-a7-(1–208) (lane 1) and of the respective C116S mutant (lane 2) after refolding from inclusion bodies in the presence of 0.1% SDS. A similar picture was observed when these proteins were refolded either in the presence of 0.1% Chaps or in aqueous buffer without any detergents (data not shown). The 1–208 fragment was also successfully expressed as a soluble MBP-a7-(1–208) protein and isolated with the aid of affinity chromatography on an amylose resin (lane 3). A compar- ison with the standards (lane 4) shows that all the proteins obtained have apparent molecular masses in the expected range. On storage of the GST–a7-(1–208) protein refolded in the absence of any detergents, precipitation usually occurred. This protein could be kept in solution in concentrations up to  1.2 mgÆmL )1 if supplemented with 0.1% Chaps [16]. Somewhat higher concentrations ( 1.5–2.0 mgÆmL )1 ) were achieved with refolding in the presence of 0.1% SDS. The solutions in 0.1% SDS-containing buffers could be stored at 4 °C for over a month, the C116S mutant being especially stable. Although the MBP-a7-(1–208) was pro- duced as a water-soluble protein, after purification it could be kept for prolonged period in solution at concentrations below 0.2 mgÆmL )1 only if supplemented with 0.05% octyl- b- D -glucoside and a cocktail of protease inhibitors. Analysis of a-bungarotoxin binding As seen from Fig. 3A, the expressed proteins specifically bind 125 I-labelled a-cobratoxin on blots. Data on the 125 I-labelled aBgt equilibrium binding in solution are compiled in Table 1 and illustrated in Fig. 3B for GST– a7-(1–208) and its C116S mutant. The K d values are in the range of 100–820 n M anddependontherefoldingcondi- tions. When refolded in the presence of 0.1% SDS, the GST–a7-(1–208) and its C116S mutant have very similar K d values (310 and 160 n M , respectively). The difference was larger (180 and 820 n M ) when refolding was carried out in 0.1% Chaps. For MBP-a7-(1–208) protein, K d values were in the low micromolar range (data not shown). It is known that aBgt binds to the full-size a7AChRin the low nanomolar range [13,25]. A K d of 1.6–2.0 n M was reported for the extended chicken a7 ECD expressed in the Xenopus oocytes [14], whereas a water-soluble protein MBP-a7-(1–196) produced in E. coli had a K d of 2.5 l M [15]. The highest affinity for GST–a7-(1–208) proteins was  150 n M , which is still considerably weaker than for the full-size a7AChR 4 . However, the GST–a7-(1–208) refolded in 0.1% Chaps was able to distinguish the long-chain a-neurotoxins from the short ones, as well as the a7AChR targeting a-conotoxin ImI from the muscle-type targeting a-conotoxin GI [16]. These properties of the intact a7 AChRs are preserved in the GST–a7-(1–208) refolded in 0.1% SDS (data not shown). Analysis of aggregation state The aggregation state of the fusion proteins containing the rat a7 ECD was examined by SDS/PAGE (Fig. 4) and gel- permeation HPLC (Fig. 5). In the presence of 2-mercapto- ethanol, the GST–a7-(1–208) protein refolded in 0.1% SDS is predominantly a monomer, whereas in the absence of the reducing agent it contains also a considerable portion of oligomers and high-molecular mass aggregates (cf. lanes 1 and 2 in Fig. 4). For the chick a7 AChR heterologously expressed in a mammalian cell line, it was recently shown that cysteines of the ECD are involved in aggregation, which can be considerably decreased by reducing agents or by treatment with N-ethylmaleimide [26]. We decided to check whether Fig. 2. SDS/PAGE of recombinant fusion proteins containing the rat a7 ECD. (1 and 2) GST–a7-(1–208) and GST–a7m-(1–208) fusion pro- teins, respectively, after refolding from the inclusion bodies in the presence of 0.1% SDS; (3) MBP-a7-(1–208) fusion protein after affinity purification on amylose resin; (4) molecular mass markers (in kDa). 2804 V. I. Tsetlin et al. (Eur. J. Biochem. 269) Ó FEBS 2002 modification with N-ethylmaleimide gave a similar effect for the rat a7ECDexpressedinE. coli and the consequences of deleting the SH group of the specified residue, Cys116, by mutating the latter into a Ser. After reduction–reoxidation of the GST–a7-(1–208) and dialysis, only 1 SH group per mol was detected by Ellman’s reagent. Because of the absence of free SH groups in GST (see Fig. 1), there were reasons to believe that it was that of Cys116. The N-ethylmaleimide treatment of the fusion protein GST–a7-(1–208) refolded in 0.1% SDS leads to an clear decrease in aggregates and oligomers on nonreducing gels (cf. lanes 2 and 6 in Fig. 4). The effect of mutating Cys116 is even more dramatic: there is almost no difference between the SDS/PAGE under reducing and nonreducing conditions (lanes 3 and 4 in Fig. 4), the monomer prevailing in both cases. This result explains why the solutions of the C116S mutant in 0.1% SDS are much less prone to aggregation and precipitation on prolonged storage as compared to the GST–a7-(1–208) protein itself. The dependence of the aggregation state on the refolding conditions and on the introduced mutation was examined by gel-permeation HPLC on a Superdex 200 column (Fig. 5). The GST–a7-(1–208) proteins, refolded in the presence of 0.1% SDS, were subjected to chromatography in the presence of 0.1% SDS, which is similar to SDS/ PAGE under nonreducing conditions. It was found that the fraction of aggregates (marked with an asterisk) is consid- erably smaller for the C116S mutant (Fig. 5A). The major broad peak centred at  70 kDa originates mainly from a monomeric fraction as indicated from its position on SDS/ PAGE under nonreducing conditions (data not shown; the shoulders at < 43 kDa are the concomitant E. coli pro- teins). The consequences of mutation are very pronounced for the GST–a7-(1–208) proteins refolded in 0.1% Chaps. Chromatography in a buffer containing the same detergent showed in the GST–a7-(1–208) protein a large peak of aggregates overlapping the oligomeric and monomeric fractions, while for the C116S mutant a broad peak of oligomers dominated (Fig. 5B). The proteins obtained were also analyzed on a Super- dex 200 column equilibrated in purely aqueous buffers. For the GST–a7-(1–208) protein refolded in Tris/HCl buffer containing 1 m M dithiothreitol and dialysed against the same buffer without dithiothreitol, the intense peak of high- molecular mass aggregates (m > 600 kDa) and a broad peak with a centre at  450 kDa corresponding to oligo- mers and monomers are present (Fig. 5C). It should be noted that the resolution of the column is better here than in the presence of 0.1% SDS, the standards on average elute later (cf. Fig. 5A,C). For the GST–a7m-(1–208) protein, also refolded in the absence of 0.1% SDS, the peak of aggregates is decreased by  25%, the centre of broad peak of oligomers shifts from 450 to 300 kDa (corresponding to oligomers of five or six units). When the MBP-a7-(1–208) protein, produced in a water-soluble form and not subjected to action of such denaturants as 8 M urea or GdnHCl, was chromatographed under the conditions of Fig. 5C, only one sharp peak of high-molecular mass aggregates was obtained (Fig. 5D) showing that this protein is not promising for further physicochemical studies. On the other hand, the chromatography on Superdex 200 in the absence of 0.1% SDS revealed a dramatic difference between the GST–a7- (1–208) protein and its C116S mutant when they were refolded in the presence of 0.1% SDS (Fig. 5E). Whereas for GST–a7-(1–208) the peak of aggregates is much larger than the broad peak centred at  70 kDa, the latter is predominant for GST–a7m 5 -(1–208). Thus, the presented SDS/PAGE and gel-permeation HPLC data show that the problem of aggregation can be partly solved by chemical modification of accessible sul- fhydryl groups or by mutating the Cys116 residue. How- ever, formation of intermolecular disulfides with the participation of the Cys116 sulfhydryl is one of the factors leading to aggregation, but not the sole one. A further decrease in the aggregation extent depends only weakly on the state of SH groups and requires the addition of 0.1% SDS. The monomeric fractions 2 (Fig. 5E) were collected for further analysis. The K d values characterizing their inter- Fig. 3. Interaction of fusion proteins containing a7 ECD with iodinated a-neurotoxins. (A) Autoradiography of 125 I-labelled a-cobratoxin binding by MBP-a7-(1–208) (lane1) and GST–a7-(1–208) (lane 2) on blots. Lane 3, protein standards as detected by Coomassie staining of the respective portion of the gel. (B) 125 I-Labelled aBgt binding curves for GST–a7-(1–208) protein (filled circles) and GST–a7m-(1–208) protein (open circles) refolded in 0.1% SDS. Before the binding assay, the SDS concentration was decreased to 0.004% by dilution with aqueous buffer. The indicated K d values are averaged from two inde- pendent experiments and calculated using ORIGIN 5.0. Ó FEBS 2002 Mutation and refolding of a7 extracellular domain (Eur. J. Biochem. 269) 2805 action with 125 I-labelled aBgt (see Table 1) are very close to those of the starting GST–a7-(1–208) and GST–a7m- (1–208) proteins, confirming a somewhat higher affinity of the C116S mutant. The SDS content in the pooled fractions in the reaction with p-rosaniline chloride [22] was estimated to be 0.0001%. When these fractions were reapplied to the same column equilibrated without 0.1% SDS, only the peaks of monomers were observed, virtually without traces of aggregates (Fig. 5F, fraction 2). For comparison, rechromatography of the pooled aggregates peaks is shown (Fig. 5F, fraction 1). CD spectra analysis CD spectra of the GST–a7-(1–208) protein and its C116S mutant look similar (Fig. 6), having the contributions from both a helices and b structure 6 . Analysis of the CD curves (Table 2) revealed that the content of the secondary structure varies depending on the protein and the refolding conditions. In general, these variations are smaller for the C116S mutant: the secondary structure is very similar for the protein refolded either in purely aqueous solution, or refolded and analyzed in 0.1% SDS, or isolated from the latter by HPLC in the absence of detergent. The average content of the a helices, b structure and unordered confor- mation is estimated to be 22, 45, and 33%, respectively. It is also clear from the Table 2 that under all similar conditions, the C116S mutant is characterized by almost a twofold higher content of b structure than the starting GST–a7- (1–208) protein. In contrast, the latter has a 1.5-fold higher content of a helices and a somewhat higher percentage of unordered structure. In view of the more stable secondary structure of the mutant and its decreased tendency to aggregation, we assume that the C116S mutant is a better model of the monomeric a7 ECD than the starting GST– a7-(1–208) protein. If we further assume that GST, which does not change its secondary structure under different conditions (Table 2), also preserves it in the a7ECDfusion proteins, a rough estimate of the secondary structure can be made for the a7 ECD moiety. Because GST and a7ECD have very close molecular masses, their contributions to CD curves should be almost equal. Using the averaged values for the C116S mutant, we obtained the following values for the a7 ECD: 17% a helix, 56% b structure and 27% of unordered structure. DISCUSSION The results obtained, along with other recently reported data [15,16], show that ECD of the rat a7AChRcanbe heterologously expressed in E. coli as different fusion proteins soluble, either in purely aqueous solutions or in the presence of detergents. The respective proteins bind aBgt and other long-chain a-neurotoxins with an affinity which is not strongly dependent on the chosen variant of the a7 ECD, production of proteins in water-soluble form or their recovery from the inclusion bodies, or on refolding with or without detergents. The earlier reported affinities for MBP-a7-(1–196) in aqueous solution [15] and for GST–a7- Table 1. Radioligand assay data of 125 I-labelled iodinated a-bungarotoxin binding by the proteins refolded under different conditions. Protein Detergent content in refolding buffer (%) 125 I-Labelled aBgt binding parameters K d (n M ) B max (%) b GST–a7-(1–208) – 260 ± 170 6.6 ± 3.8 SDS, 0.1 310 ± 70 12.5 ± 3 Chaps, 0.1 180 ± 75 3.0 ± 0.5 Monomeric form a SDS, 0.0005 240 ± 160 1.7 ± 0.4 GST–a7m-(1–208) – 130 ± 75 3.3 ± 0.6 (C116S mutant) SDS, 0.1 160 ± 40 10 ± 1 Chaps, 0.1 820 ± 175 3.0 ± 0.4 Monomeric form a SDS, 0.0005 100 ± 60 1.6 ± 0.3 a The monomeric forms were prepared by ultrafiltration of the HPLC purified monomeric fractions (see Fig. 5E,F) using an Amicon 8010 ultrafiltration membrane YM10. SDS content in the concentrated samples was determined in reaction with p-rosaniline chloride [22]. b The B max values are presented as a percentage ratio of the calculated B max in n M to the concentration of the respective protein (in n M ) in the incubation mixture. The data presented were calculated with the use of ORIGIN 5.0. Fig. 4. Analysis of the a7 ECDs aggregation state by SDS/PAGE in the presence (+) or absence (–) of 1% 2-mercaptoethanol (ME). Lanes 1 and 2, GST–a7-(1–208); lanes 3 and 4, GST–a7m-(1–208) mutant; lanes 5 and 6, GST–a7-(1–208) modified with N-ethylmaleimide, lane 7, molecular mass markers, kDa. 2806 V. I. Tsetlin et al. (Eur. J. Biochem. 269) Ó FEBS 2002 (1–208) in 0.1% Chaps [16] are in the low micromolar range. The K d values obtained in this work are in the range of 100– 850 n M , the highest affinity being shown by the C116S mutant of GST–a7-(1–208) protein refolded from the inclusion bodies in the presence of 0.1% SDS (Table 1). Although even this affinity is much weaker than that of the intact a7 AchR, which binds aBgt with K d  1–5 n M [13,25], such important features of the a7 AChR selectivity as the capacity to discriminate long- and short-chain a-neurotoxins and various a-conotoxins, are retained in the expressed a7 ECD (see Results). High molecular mass aggregates, whose molecular mas- ses are considerably larger that of a pentamer, are the major obstacle to obtainaing soluble and correctly folded hetero- logously expressed ECDs of AChRs. Aggregation usually leads to insoluble precipitates. However, a7ECDsmay contain high molecular mass aggregates even in solution (Fig. 5). The major cause of poor solubility might have been the necessity of isolating the proteins from inclusion bodies under denaturing conditions and then to refold them. Therefore, much effort has been made to produce the AChR ECDs or their large portions as soluble fusion proteins [15,27 7 ]. In particular, this goal was achieved for the rat a7 ECD in [15] and in the present work using similar fusion proteins with MBP. However, these soluble proteins contained various amounts of aggregates. Aggregation and precipitation of the proteins obtained from the inclusion bodies could be partly overcome by optimizing the refolding conditions. If 0.1% SDS was present from the stage of dissolving the inclusion bodies, the GST–a7-(1–208) protein could be kept in solution (Tris/ HCl buffer, pH 8.0, 0.1% SDS) for a long time at concentrations about 1 mgÆmL )1 . Under these conditions the major fraction of the protein is a monomer, but the fraction of high-molecular mass aggregates is also quite large (Fig. 5A). It is known that SDS is capable of inhibiting the aggregation of bacterially expressed or denatured proteins and therefore is widely used in refolding studies [28–30]. It is assumed that the masking of hydrophobic protein interfaces by detergent molecules results in 8 this reduction in aggregation. There are also indications that SDS diminishes the aggregation by inhibiting the formation Fig. 5. Gel-permeation HPLC of the a7 ECD-containing fusion proteins on a Superdex 200 HR column. (A) GST–a7-(1–208) (solid line) and GST–a7m-(1–208) (dashed line) refolded in the presence of 0.1% SDS are analyzed on the column equilibrated in 20 m M Tris/HCl, pH 8.0, containing 150 m M NaCl (elution buffer) supplemented with 0.1% SDS. Proteins (30–50 lg) were eluted at a flow rate of 0.5 mLÆmin )1 . The same conditions, with the exception of presence or absence of detergents in the elution buffers, and designations apply to (B), (C) and (E). (B) GST–a7-(1–208) and GST–a7m-(1–208) refolded in the pres- ence of 0.1% Chaps, the column is equilibrated and run in the elution buffer containing 0.1% Chaps. (C) GST–a7-(1–208) and GST–a7m- (1–208) proteins refolded in the absence of detergents, the elution buffer is without detergents. (D) MBP-a7-(1–208) protein analyzed in the conditions of (C) (E) GST–a7-(1–208) and GST–a7m-(1–208) proteins refolded in the presence of 0.1% SDS are analyzed as in (C) and (D). (F) Rechromatography of the fractions marked with bars 1 (aggregates) and 2 (monomer) collected on chromatography of GST– a7-(1–208) (E). In this figure solid and dashed lines correspond to chromatography of the fractions 1 and 2 from (E), respectively. In all figures, vertical short lines correspond to the elution times of protein standards (with masses in kDa) under the chosen chromatographic conditions, the asterisk marking the exclusion volume. Fig. 6. CD spectra of fusion protein GST–a7-(1–208) (solid line) and GST–a7m-(1–208) mutant (dashed line) refolded in 20 m M Tris/HCl, pH 8.0, containing 0.1% SDS. Ó FEBS 2002 Mutation and refolding of a7 extracellular domain (Eur. J. Biochem. 269) 2807 of intermolecular disulfide bridges [31]. Note that aggrega- tion, oligomerization and the acquisition of aBgt binding capacity by the a7andTorpedo AChRs were shown by Green and coworkers to depend on the state of cysteine residues in the ECDs of the a7anda subunits, respectively [26,32]. The major role of the redox state of the disulfide Cys128–Cys142 was assumed for the a7AChR[26], whereas the involvement of the disulfide Cys192–Cys193 was demonstrated for the Torpedo a subunit [32]. The a7 ECD has the Cys128–Cys142 disulfide bond 9 characteristic for the whole family of ligand-gated ion channels, as well as the vicinal disulfide bond 10 Cys192–Cys193 (see Fig. 1) present also in all other a subunits of neuronal and muscle-type AChRs. On the other hand, Cys116 is present only in a7anda8 subunits. We thought that the presence of a free SH group of Cys116 might be one of the factors leading to formation of intermolecular disulfides and aggregates in the bacterially expressed a7ECDs. Indeed, the GST–a7m-(1–208) protein on SDS/PAGE under nonreducing conditions was present mainly as a monomer, whereas the starting protein contained a large proportion of aggregates (Fig. 4). A similar effect could be achieved by blocking free sulfhydryl groups in GST–a7- (1–208) protein by N-ethylmaleimide (Fig. 4). The consid- erable decrease in aggregation caused by the C116S mutation with the proteins refolded and analyzed in the presence or absence of various detergents is illustrated by HPLC data (Fig. 5). For the proteins refolded in the presence of 0.1% SDS, chromatography on a column equilibrated in a purely aqueous buffer allowed us to obtain the monomeric fractions (Fig. 5E,F). These fractions bind 125 I-labelled aBgt with the affinities similar to those of the starting proteins (Table 1). Because the mutant revealed even higher affinity, the presence of the Cys116 free SH group is not essential for aBgt binding. CD spectra (Fig. 6 and Table 2) indicate that the GST– a7-(1–208) and GST–a7m-(1–208) proteins refolded under different conditions contain varying amounts of regular secondary structure. Interestingly, under all similar condi- tions the C116S mutant has about a twofold higher content of b structure than the GST–a7-(1–208) protein. The C116S mutant is less prone to aggregation (see Figs 4 and 5), binds aBgt even better than the starting protein (Table 1), and its CD curves are not very sensitive to the conditions of refolding or measuring the spectra. Therefore, the confor- mation of the a7 moiety in the mutant may be more similar to the ECD conformation of the intact a7 AChR. Although calculation of the secondary structure of the 1–208 fragment from the CD data for the fusion protein GST–a7-(1–208) and GST might seem arbitrary, we believe that it gives a qualitatively correct conclusion: the measured relatively high content of b structure is not largely due to the contribution of GST. A high content of b structure found experimentally for the C116S mutant and the even higher content calculated for the a7 ECD moiety is important. It is presumed that the ECDs in different, hetero-oligomeric or homo-oligomeric AChRs, have a similar spatial organization. CD analyses of the mouse muscle a subunit ECD heterologously expressed in mammalian cells [6] and of the Torpedo a subunit ECD expressed in E. coli [7–9] gave an estimate of  50% for b structure. These results are in good agreement with the subsequently published data of cryo-electron microscopy, which revealed a high content of b structure in the ECD of the intact Torpedo californica AChR [5]. Our present results suggest that b structure is an important element of spatial organization of the a7ECD. When our work was in progress, isolation of acetylcho- line-binding protein (AChBP) from a mollusc Lymneae stagnalis and X-ray analysis of the respective protein heterologously expressed in yeast have been published [11,12]. This protein, with 24% homology to a7ECD, contains practically all amino-acid residues of the AChR ligand-binding sites, has two invariant disulfides and lacks extra Cys116 of a7. It has an immunoglobulin-like topology (rich in b structure) and forms homopentamers [12]. There- fore, at least in terms of secondary structure, the monomeric form of the GST–a7m-(1–208) protein resembles the protomer of the AChBP. This would justify further efforts to prepare the oligomeric a7 ECD using the obtained monomers as the starting material. ACKNOWLEDGEMENTS The authors are grateful to Dr H J. Kreienkamp for a7 cDNA clone, to O. Ustinova for help with CD measurements and to Dr J. Freigang for fruitful discussion. The work was supported by grants (to V. T.) from Bayer AG (Leverkusen) and Russian Foundation for Basic Research. Table 2. CD data of the different expressed proteins refolded under various conditions. Protein Detergent content in refolding buffer Concentration (mgÆmL )1 ) Calculated secondary structure (%) abR GST–a7-(1–208) – 0.20 43 12 45 – 0.10 44 27 29 0.1% SDS 0.16 29 36 35 Monomeric form a 0.0005% SDS 0.13 39 23 38 GST–a7m-(1–208) (C116S mutant) – 0.21 21 44 35 – 0.84 20 47 33 0.1% SDS 0.20 19 47 34 Monomeric form a 0.0005% SDS 0.12 29 40 31 GST – 0.20 28 35 37 0.1% SDS 0.90 27 31 42 a See the respective note in the Table 1. 2808 V. I. Tsetlin et al. (Eur. J. Biochem. 269) Ó FEBS 2002 REFERENCES 1. Changeux, J.P. & Edelstein, S.J. (1998) Allosteric receptors after 30 years. Neuron 21, 959–980. 2. Lindstrom, J. (2000) The structures of neuronal nicotinic recep- tors. In Handbook of Experimental Pharmacology (Clementi, F., Fornasari, D. & Gotti, C., eds), 144, pp. 101–162. Springer-Verlag, Berlin-Heidelberg. 3. Hucho, F. & Weise, C. (2001) Ligand-gated ion channels. Angew. Chem.Int.Ed.40, 3100–3116. 4. Klymkowsky, M.W. & Stroud, R.M. (1979) Immunospecific identification and three-dimensional structure of a membrane- bound acetylcholine receptor from Torpedo californica. J. Mol. Biol. 128, 319–334. 5. Miyazawa, A., Fujiyoshi, Y., Stowell, M. & Unwin, N. (1999) Nicotinic acetylcholine receptor at. 4.6 A ˚ resolution: transverse tunnels in the channel wall. J. Mol. Biol. 288, 765–786. 6. West, A.P. Jr, Bjorkman, P.J., Dougherty, D.A. & Lester, H.A. (1997) Expression and circular dichroism studies of the extra- cellular domain of the a-subunit of the nicotinic acetylcholine receptor. J. Biol. Chem. 272, 25468–25473. 7. Krivoshein, A.V., Kudelina, I.A., Alexeev, T.A., Shevalier, A.F., Utkin, YuN. & Tsetlin, V.I. (1998) The secondary structure of some a-subunit fragments of Torpedo californica acetylcholine receptor. Bioorgan. Khim. 24, 825–827. 8. Schrattenholz, A., Pfeiffer, S., Pejovic, V., Rudolph, R., Godovac- Zimmermann,J.&Maelicke,A.(1998)Expressionandcircular dichroism studies of the extracellular domain of the a-subunit of the nicotinic acetylcholine receptor. J. Biol. Chem. 273, 32393– 32399. 9. Alexeev, T., Krivoshein, A., Shevalier, A., Kudelina, I., Telyak- ova, O., Vincent, A., Utkin, Y., Hucho, F. & Tsetlin, V. (1999) Physico-chemical and immunological studies of the N-terminal domain of the Torpedo acetylcholine receptor a-subunit expressed in Escherichia coli. Eur. J. Biochem. 259, 310–319. 10. Tierney, M.L. & Unwin, N. (2000) Electron microscopic evidence for the assembly of soluble pentameric extracellular domains of the nicotinic acetylcholine receptor. J. Mol. Biol. 303, 185–196. 11. Smit, A.B., Syed, N.I., Schaap, D., van Minnen, J., Klumperman, J., Kits, K.S., Lodder, H., van Der Schors, R.C., van Elk, R., Sorgedrager, B., Brejc, K., Sixma, T.K. & Geraerts, W.P. (2001) A glia-derived acetylcholine-binding protein that modulates synaptic transmission. Nature 411, 261–268. 12. Brejc, K., van Dijk, W.J., Klaassen, R.V., Schuurmans, M., van Der Oost, J., Smit, A.B. & Sixma, T.K. (2001) Crystal structure of an ACh-binding protein reveals the ligand-binding domain of nicotinic receptors. Nature 411, 269–276. 13.Couturier,S.,Bertrand,D.,Matter,J.M.,Hernandez,M.C., Bertrand, S., Millar, N., Valera, S., Barkas, T. & Ballivet, M. (1990) A neuronal nicotinic acetylcholine receptor subunit (a7) is developmentally regulated and forms a homo-oligomeric channel blocked by a-BTX. Neuron 5, 847–856. 14. Wells, G.B., Anand, R., Wang, F. & Lindstrom, J. (1998) Water- soluble nicotinic acetylcholine receptor formed by a7 subunit extracellular domains. J. Biol. Chem. 273, 964–973. 15. Fischer, M., Corringer, P.J., Schott, K., Bacher, A. & Changeux, J.P. (2001) A method for soluble overexpression of the a7 nicotinic acetylcholine receptor extracellular domain. Proc. Natl Acad. Sci. USA 98, 3567–3570. 16. Utkin, Y.N., Kukhtina, V.V., Kryukova, E.V., Chiodini, F., Bertrand,D.,Methfessel,C.&Tsetlin,V.I.(2001)ÔWeak toxinÕ from Naja kaouthia is a nontoxic antagonist of a7andmuscle– type nicotinic acetylcholine receptors. J. Biol. Chem. 276, 15810– 15815. 17. Ellman, G.L. (1959) Tissue sulfhydryl groups. Arch. Biochem. Biophys. 82, 70–77. 18. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 19. Klukas, O., Peshenko, I.A., Rodionov, I.L., Telyakova, O.V., Utkin, Y.N. & Tsetlin, V.I. (1995) Fragments 183–198 and 125– 145 of the a-subunit of the nicotinic acetylcholine receptor from Torpedo californica bind a-bungarotoxin and neurotoxin II from Naja naja oxiana. Bioorgan. Khim. 21, 152–155. 20. Schmidt, J. & Raftery, M.A. (1974) The cation sensitivity of the acetylcholine receptor from Torpedo californica. J. Neurochem. 23, 617–623. 21. Kachalsky, S.G., Jensen, B.S., Barchan, D. & Fuchs, S. (1995) Two subsites in the binding domain of the acetylcholine receptor: an aromatic subsite and a proline subsite. Proc. Natl Acad. Sci. USA 92, 10801–10805. 22. Amons, R. & Schrier, P.I. (1981) Removal of sodium dodecyl sulfate from proteins and peptides by gel-filtration. Anal. Biochem. 116, 439–443. 23. Schippers, P.H. & Dekkers, H.P.J.M. (1981) Direct determination of absolute circular dichroism data and calibration of commercial instruments. Anal. Chem. 53, 778–788. 24. Provencher, S.W. (1982) A constrained regularization method for inverting data represented by linear algebraic or integral equa- tions. Comput. Phys. Commun. 27, 229–242. 25. Gerzanich, V., Anand, R. & Lindstrom, J. (1994) Homomers of a8anda7 subunits of nicotinic receptors exhibit similar channel but contrasting binding site properties. Mol. Pharmacol. 45, 212–220. 26. Rakhilin, S., Drisdel, R.C., Sagher, D., McGehee, D.S., Vallejo, Y. & Green, W.N. (1999) a-Bungarotoxin receptors contain a7 subunits in two different disulfide-bonded conformations. J. Cell Biol. 146, 203–218. 27. Grant, M.A., Gentile, L.N., Shi, Q.L., Pellegrini, M. & Hawrot, E. (1999) Expression and spectroscopic analysis of soluble nicotinic acetylcholine receptor fragments derived from the extracellular domain of the a-subunit. Biochemistry 38, 10730–10742. 28. Couthon, F., Clottes, E. & Vial, C. (1996) Refolding of SDS–and thermally denatured MM–creatine kinase using cyclodextrins. Biochem. Biophys. Res. Commun. 227, 854–860. 29. Shao, H., Jao, S., Ma, K. & Zagorski, M.G. (1999) Solution structures of micelle–bound amyloid b–(1–40) and b–(1–42) pep- tides of Alzheimer’s disease. J. Mol. Biol. 285, 755–773. 30. MacRaild, C.A., Hatters, D.M., Howlett, G.J. & Gooley, P.R. (2001) NMR structure of human apolipoprotein C–II in the presence of sodium dodecyl sulfate. Biochemistry 40, 5414–5421. 31. Schrooyen, P.M., Dijkstra, P.J., Oberthur, R.C., Bantjes, A. & Feijen, J. (2001) Stabilization of solutions of feather keratins by sodium dodecyl sulfate. J. Colloid. Interface Sci. 240, 30–39. 32. Mitra, M., Wanamaker, C.P. & Green, W.N. (2001) Rearrange- ment of nicotinic receptor a–subunits during formation of the ligand binding sites. J. Neurosci. 21, 3000–3008. Ó FEBS 2002 Mutation and refolding of a7 extracellular domain (Eur. J. Biochem. 269) 2809 . Refolding of the Escherichia coli expressed extracellular domain of a7 nicotinic acetylcholine receptor Cys116 mutation diminishes aggregation and stabilizes the b structure Victor. form. Keywords: a7 nicotinic acetylcholine receptor; extracellular domains; expression in Escherichia coli; Cys116 mutation; CD spectroscopy. Nicotinic acetylcholine receptors (AChRs), belonging to the family. studying the functional and spatial characteristics of their ligand-binding sites. The ECD of the a7 subunit of the homo-oligomeric a7 neuronal AChR appears to be a more suitable object than the