Novel strategy for protein production using a peptide tag derived from Bacillus thuringiensis Cry4Aa Tohru Hayakawa 1 , Shinya Sato 1,2 , Shigehisa Iwamoto 2 , Shigeo Sudo 2 , Yoshiki Sakamoto 1 , Takaaki Yamashita 1 , Motoaki Uchida 1 , Kenji Matsushima 1 , Yohko Kashino 1 and Hiroshi Sakai 1 1 Graduate School of Natural Science and Technology, Okayama University, Japan 2 Department of Bioscience, Japan Lamb Co. Ltd, Development Division, Okayama, Japan Introduction Cry toxin, a specific insecticidal protein against insect larvae, has been used as an insect pest control agent throughout the world [1]. Cry toxin is usually pro- duced intensively during sporulation and accumulates in the form of protein crystals in Bacillus thuringiensis cells. Hyperexpression of Cry toxin may be supported by factors such as strong promoters, stable mRNAs and protein crystallization [2]. In particular, crystal formation is a characteristic of B. thuringiensis and may be useful as an aid in the overproduction of recombinant protein. Protein crystallization allows a large amount of protein to be packed into the limited intracellular space and protects protein from proteo- lytic degradation in the environment. Indeed, Cry toxin accumulates as a protein crystal that can account for up to 25% of the dry weight of the cell [2]. The mechanism of protein crystallization may vary with the type of Cry toxin. The larger-sized Cry pro- toxin ( 130 kDa) contains a C-terminal extension in addition to the insecticidal N-terminal region and probably crystallizes by self-assembly. When Cry1, a member of the large protoxin group, is expressed in Keywords 4AaCter peptide tag; Bacillus thuringiensis; Cry4Aa; Escherichia coli; TpN syphilis antigen Correspondence H. Sakai, Graduate School of Natural Science and Technology, Okayama University, 3-1-1 Tsushima-Naka, Okayama 700-8530, Japan Fax ⁄ Tel: +81 86 251 8203 E-mail: sakahrsh@biotech.okayama-u.ac.jp Database The nucleotide sequence of the synthetic gene cry4Aa-S2 is available in the EMBL ⁄ GenBank ⁄ DDBJ databases under accession number AB513706 (Received 26 February 2010, revised 11 April 2010, accepted 3 May 2010) doi:10.1111/j.1742-4658.2010.07704.x Numerous proteins cannot be sufficiently prepared by ordinary recombi- nant DNA techniques because they are unstable or have deleterious effects on the host cell. One idea to prepare such proteins is to produce them as protein inclusions. Here we developed a novel system to effectively prepare proteins by using peptide tags derived from the insecticidal Cry toxin of a soil bacterium, Bacillus thuringiensis. Fusion with this peptide tag, desig- nated 4AaCter, facilitates the formation of protein inclusions of glutathione S-transferase in Escherichia coli without losing the enzyme activity. Appli- cation of 4AaCter to the production of syphilis antigens TpN15, TpN17 and TpN47 from Treponema pallidum yielded excellent results, including a dramatic increase in the production level, simplification of the product purification and high reactivity with syphilis antibody. The use of 4AaCter may provide an innovational strategy for the efficient production of proteins. Abbreviations CBB, Coomassie brilliant blue; CDNB, 1-chloro-2,4-dinitrobenzene; GST, glutathione S-transferase; PBST, phosphate-buffered saline containing 0.05% Tween 20. FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS 2883 Escherichia coli, protein inclusions containing biologi- cally active toxin are formed [3,4]. Furthermore, the bipyramidal crystals denatured in 8 m urea revert to their original crystal shape when the urea is removed by dialysis [5]. The C-terminal extension is not directly related to the insecticidal activity and is usually removed upon proteolytic activation. The amino acid sequences of some segments in the C-terminal exten- sion are highly conserved among the large protoxin group and are believed to be important for protein crystallization. The smaller-sized Cry protoxin ( 70 kDa or less) consists of an insecticidal N-termi- nal region only and requires an accessory protein (such as P20) for crystallization [2]. The protein crystallization mechanism of the Cry toxin may prove beneficial for the expression of heterol- ogous proteins in general. Crystallization of the large protoxin seems to be dependent on some special protein structure in the C-terminal extension; therefore, this extension can be utilized as a peptide tag to facili- tate the crystallization of fused proteins. The crystalli- zation may protect the protein from degradation by in vivo proteinase and may also simplify the purification steps. Cry4Aa is a dipteran-specific Cry toxin produced by B. thuringiensis subspecies israelensis, and is a member of the larger-sized protoxin group. In the midgut of a sus- ceptible insect, Cry4Aa is processed into protease-resis- tant segments of 20 and 45 kDa via a 60 kDa intermediate generated by the removal of the C-terminal extension [6]. In the present study, we developed a Cry4Aa-derived peptide tag that facilitates the formation of protein inclusions of fused proteins. We demonstrated the usefulness of the peptide tag by efficiently producing TpN antigens from Treponema pallidum, which is a spi- rochetal bacterium causing syphilis. In general, because T. pallidum cannot be cultured continuously in vitro, establishment of the efficient production system for TpN antigens that are used to diagnose syphilis is eagerly desired. Our newly isolated peptide tag may contribute to the establishment of novel and efficient expression sys- tems for proteins. Results Glutathione S-transferase (GST) fused with Cry4Aa peptides To determine the peptide stretch involved in protein crystallization, nine polypeptides from the C-terminal region of Cry4Aa were fused, as peptide tags, to GST. The resulting fusion proteins were named GST–4AaCt- ers 852–1180, 696–851, 696–799, 801–851, 801–834, 801–829, 807–824, 807–819 and 812–824, based on the amino acid number from the N-terminal end of the Cry4Aa protoxin (Fig. 1). All GST fusion proteins were successfully expressed in E. coli, and their sizes estimated by SDS ⁄ PAGE were 64 kDa (GST–4AaCter 852–1180), 44 kDa (GST–4AaCter 696–851), 39 kDa (GST–4AaCter 696–799), 32 kDa (GST–4AaCter 801– 851), 31 kDa (GST–4AaCter 801–834), 30 kDa (GST– 4AaCter 801–829), 29 kDa (GST–4AaCter 807–824), 28 kDa (GST–4AaCter 807–819) and 28 kDa (GST– 4AaCter 812–824) (Fig. 2A). These sizes correspond to those predicted from the deduced amino acid sequences of each GST fusion protein. More than 95% of GST–4AaCters 696–851, 801– 851, 801–834 and 801–829, which contain the highly C ry 4Aa Ins e cticidal N-te r minal h a lf G 58 Q 695 I 6 96 E 1 1 80 C- ter m in al h alf Mu ta nts GST A 85 2 E 1 1 80 GST-4 Aa Cter 852–1180 GST- 4 AaCte r 696–851 GST I 69 6 P 851 B l ock s12 3 456 7 8 GST-4AaCter 696–799 GST I 696 G 79 9 G S T-4AaCter 801–851 G S T I 8 0 1 P 851 I 80 1 FPT YI FQK ID ES K LKP Y TRYL V R G FV GS SK D VEL- - - P 851 GST I 8 0 1 E 834 G S T I 8 01 S 8 29 GST- 4A a C te r 8 0 1–834 GS T - 4A a C ter 8 0 1–829 B l o c k7 GST-4A a Ct e r8 0 7–819 GST F 8 07 R 82 4 GS T GS T - 4 A aC t e r 812–824 T 81 9 E 8 1 2 GST-4AaCter 807–824 GS T F 8 0 7 R 8 2 4 Fig. 1. Design of GST–4AaCters. The schematic structure of GST–4AaCters is shown. Nine peptide segments from the Cry4Aa C-terminal region are fused with GST. The amino acid sequence of Cry4Aa block 7 is underlined. Efficient protein production using 4AaCter T. Hayakawa et al. 2884 FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS conserved Block 7 sequence, were found in the insolu- ble fraction (Fig. 2B). In contrast, GST–4AaCters 852–1180 and 696–799 were mostly found in the soluble fraction (< 10% insoluble). In the case of GST–4AaCters 807–824, 807–819 and 812–824, 20– 40% of the proteins were insoluble (Fig. 2B). Morphological observation of the inclusions of GST Spherical inclusion in cells expressing GST–4AaCters were visible by light microscopy (Fig. 3A). Most E. coli cells expressing GST–4AaCters 696–851, 801– 851, 801–834 and 801–829 contained inclusions, but the percentage of cells with inclusions was decreased in cells expressing GST–4AaCters 807–824, 807–819 and 812–824 (Fig. 3A). Almost no inclusions were observed in cells expressing GST–4AaCters 852–1180 and 696– 799 (Fig. 3A). The percentage of protein expressed as the insoluble form seemed to be correlated with the percentage of cells containing inclusions. Scanning electron micrographs revealed that the inclusions of GST–4AaCters 696–851, 801–834 and 801–829 were approximately spherical, with a diameter of 0.5– 0.7 lm, but the surface of the inclusions appeared to be rugged (Fig. 3B). Contrary to our expectation, the morphology of the GST–4AaCter inclusions looked different from protein crystal as imagined in general. Enzyme activity of GST–4AaCters derived from inclusions Upon incubation of the inclusions in alkali buffer (50 mm NaHCO 3 , NaOH, pH 11) for 1 h at 37 °C, 852–1180 Insoluble (%) 4 100 10 100 100 100 35 20 25 5 210 119 90 65 49 37 29 21 9 (kDa) SIS I SISI S ISIS I SISIS I 696–851 696–799 801–851 801–834 801–829 807–824 807–819 812–824 GST GST-4AaCter 210 119 90 65 49 37 29 21 9 (kDa) 852–1180 696–851 696–799 801–851 801–834 801–829 807–819 812–824 807–824 GST GST-4AaCter B A Fig. 2. Expression of GST–4AaCters in E. coli. (A) GST–4AaCters expressed in E. coli were separated by SDS ⁄ PAGE (15%) and stained with CBB. GST–4AaCters are indicated by arrowheads. (B) Cells expressing GST–4AaCters were disrupted by sonication and separated into solu- ble and insoluble proteins by centrifugation. Proteins were analysed by SDS ⁄ PAGE (15%) and visualized with CBB. GST–4AaCters are indi- cated by arrowheads. The percentages of GST–4AaCters in the insoluble fraction are shown. S, soluble proteins; I, insoluble proteins. 852–1180 696–851 696–799 801–851 801–834 801–829 807–819 812–824 807–824 GST 696–851 801–834 801–829 A B Fig. 3. Micrograph of the inclusions of GST–4AaCters. (A) Micrograph of the E. coli cells expressing GST–4AaCters. Spherical inclusions of GST–4AaCters are indicated by arrows. Bar, 1 lm. (B) Scanning electron micrograph of the inclusion bodies of GST–4AaCters 696–851, 801–834 and 801–829. Bar, 600 nm. T. Hayakawa et al. Efficient protein production using 4AaCter FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS 2885 GST–4AaCters were solubilized almost completely. This suggested that the inclusion of GST–4AaCters shared some characteristics with the protein crystal produced by B. thuringiensis and was different from the inclusion of denatured proteins frequently observed in heterologous protein expression systems. GST- 4AaCters recovered in soluble fraction were subjected to a 1-chloro-2,4-dinitrobenzene (CDNB) assay to measure their GST activities. GST–4AaCter 696–851, which formed inclusions with almost 100% efficiency, had the highest GST activity (459 ± 103 lmolÆmin )1 Æ nmol )1 ) (Figs 3B and 4). This value was a bit higher than that of purified GST (319 ± 63 lmolÆmin )1 Æ nmol )1 ). Although GST–4AaCters 801–851, 801–834 and 807–824 formed inclusions with almost 100% effi- ciency, GST–4AaCters 801–851 and 801–834 had decreased GST activities (135 ± 26 and 141 ± 59 lmolÆmin )1 Ænmol )1 , respectively), and GST–4AaCter 807–824 had a very low GST activity (24 ± 18 lmolÆ min )1 Ænmol )1 ) (Fig. 4). The GST fusion proteins that formed inclusions with lower efficiencies had moderate GST activity ranging from 65 to 110 lmolÆmin )1 Æ nmol )1 (Fig. 4). Thus, because relatively high GST activity was observed, the inclusion formed by 4AaCt- ers seemed to be different from the inclusion formed by denatured proteins. Production of TpN antigens using 4AaCter GST fused with the 4AaCter 696–851 efficiently formed protein inclusions in E. coli, and the fusion protein solubilized from the inclusions had the highest GST activity among the tested samples. Therefore, it was most probable that 4AaCter 696–851 could be used as a peptide tag for the efficient production of heterologous proteins in E. coli. TpNs are surface anti- gens of T. pallidum and are highly demanded to diag- nose syphilis. We anticipated that the efficient production system of TpN antigens would be con- structed by using 4AaCter as a peptide tag. We used modified 4AaCter 696–851 that was fused with a 6 · His oligopeptide at the N-terminal end and an oligopeptide containing the cleavage site for Pre- Scission protease at the C-terminal end (Fig. 5A). The fusions of TpNs and modified 4AaCter tag were expressed using pGEX–DGST vector. 0 100 200 300 400 500 600 852–1180 696–851 696–799 801–851 801–834 801–829 807–819 812–824 807–824 GST GST activity (µmol·min –1 ·nmol –1 ) GST–4AaCters Fig. 4. Enzyme activities of GST–4AaCters. Alkali-solubilized GST– 4AaCters from the inclusions were subjected to CDNB assay. Error bars represent the standard deviation of three or more replicate experiments. 21 0 119 90 65 4 9 3 7 2 9 ( kDa ) 21 9 S I S ISISI S I S I 4 AaCter TpN15 TpN17 T pN47 4 A aCt e r - TpN15 ATG 4 A aCte r TpN15 T pN17 TpN47 4 Aa C ter- T pN 1 7 4 AaCt e r-TpN47 6xH is PP 4 A aC te r 4 A a Ct e r 4AaCter 4AaCter pG E X- Δ GST P t ac AB Fig. 5. Design and expression of 4AaCter–TpNs. (A) Schematic structure of 4AaCter–TpNs. TpN15, 17 and 47 were fused with a 6 · His tag, 4AaCter 696–851 and the recognition site of PreScission protease. The 4AaCter–TpNs were expressed using pGEX–DGST. 4AaCter, 4AaCter 696–851 (peptide tag); PP, recognition site of PreScission protease; Ptac, tac promoter. (B) Expression of 4AaCter–TpNs in E. coli. TpNs with or without the 4AaCter peptide tag were expressed in E. coli. Upon disruption by sonication, soluble and insoluble proteins were separated by SDS ⁄ PAGE (15%) and visualized with CBB. The 4AaCter–TpNs are indicated by arrowheads. S, soluble proteins; I, insoluble proteins. Efficient protein production using 4AaCter T. Hayakawa et al. 2886 FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS SDS ⁄ PAGE analysis of TpN antigens fused with 4AaCter tag at the N-terminal end revealed proteins with the expected size in the insoluble fraction (Fig. 5B). The expression levels of 4AaCter–TpNs were significantly higher than the levels of TpNs without 4AaCter, as estimated by densitometric scanning of the protein bands. On the other hand, when 4AaCter tag was fused at the C-terminal end of TpNs, no enhance- ment was observed in the expression level compared with that of TpNs without 4AaCter (data not shown). Cells were disrupted by sonication, and the inclusions of 4AaCter–TpNs were isolated by centrifugation. The yields of 4AaCter–TpN15, –TpN17 and –TpN47 were 0.31, 0.31 and 0.45 mgÆmL )1 culture, respectively. 4AaCter–TpNs solubilized in alkali buffer were already relatively pure, and their purities were estimated as 70%, 78% and 85%, respectively (Fig. 5B). 4AaCter– TpNs were further purified using a Ni-NTA column and were then treated with PreScission protease. The released 4AaCter tag and uncleaved 4AaCter–TpNs were removed using the Ni-NTA column. The purities of the final products of TpN15, TpN17 and TpN47 were 88%, 94% and 85%, respectively. Thus, to obtain highly purified TpNs, complicated purification steps were not required in this system. The estimated final yields of purified TpN15, TpN17 and TpN47 were 0.09, 0.07 and 0.12 mgÆmL )1 culture. Reactivity of recombinant TpN antigens The reactivity of the 4AaCter–TpNs against human serums was evaluated by ELISA and compared with that of native TpN antigens. The TpNs produced by removing 4AaCter peptides from the corresponding 4AaCter–TpNs were also used in this assay. The recombinant TpNs prepared by our method showed similar reactivity to that of the native TpN antigens (Tables 1 and 2). Among the recombinant TpNs, 4AaCter–TpN17 and the TpN17 were the most reac- tive with Tp-positive human serum (Table 1). A similar observation was reported for a recombinant TpN17 constructed in another experiment [7]. On the other hand, the reactivity of recombinant TpN15 and -47 varied among different human serum samples (Tables 1 and 2). There was no significant difference in reactivity between the 4AaCter–TpNs and the TpNs in which the 4AaCter peptide wa sremoved(Tables1and2). This result suggests that the human serum did not Table 1. Reactivity of recombinant TpNs against Tp-positive human serums. Sample name 4AaCter–TpN15 TpN15 4AaCter–TpN17 TpN17 4AaCter–TpN47 TpN47 Native Tp antigen 10175535 0.20 0.22 1.73 1.79 0.14 0.12 0.38 10175539 0.14 0.15 0.14 0.14 0.06 0.06 0.14 10175550 0.22 0.28 1.50 1.41 0.09 0.08 0.54 10175556 0.44 0.48 1.90 1.88 0.87 0.79 1.17 10175572 0.07 0.05 0.78 0.78 0.08 0.06 0.13 10175575 1.20 1.27 2.65 2.69 2.32 2.24 2.65 10175582 0.26 0.35 3.48 3.50 0.78 0.66 1.74 10175586 0.21 0.18 1.36 1.35 0.23 0.20 0.71 Infectrol D-00 0.20 0.21 1.07 0.99 0.40 0.36 0.60 Table 2. Reactivity of recombinant TpNs against Tp-negative human serums. Sample name 4AaCter–TpN15 TpN15 4AaCter–TpN17 TpN17 4AaCter–TpN47 TpN47 Native Tp antigen FH15079 0.08 0.06 0.07 0.07 0.08 0.06 0.11 FH15082 0.11 0.08 0.09 0.07 0.08 0.07 0.06 FH15772 0.09 0.05 0.08 0.05 0.09 0.05 0.06 FH15784 0.15 0.05 0.16 0.05 0.09 0.05 0.07 FH15788 0.07 0.05 0.08 0.05 0.09 0.05 0.06 FH15808 0.10 0.06 0.08 0.08 0.10 0.07 0.08 FH15821 0.08 0.05 0.05 0.06 0.05 0.04 0.23 FH15829 0.08 0.07 0.07 0.06 0.09 0.06 0.19 FH15833 0.31 0.28 0.25 0.27 0.21 0.31 0.17 FH15834 0.06 0.06 0.11 0.21 0.08 0.07 0.09 T. Hayakawa et al. Efficient protein production using 4AaCter FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS 2887 contain antibodies against the 4AaCter peptide and that the protein produced by our method can be used for diagnostic purposes without removing the 4AaCter peptide. Discussion The crystallization mechanism of Cry toxins is not fully understood. Although 3D structures of the N-terminal toxic region consisting of domains I, II and III have been determined for several Cry toxins [8–13], the structure of the C-terminal nontoxic region has not yet been determined. Therefore, the crystal structure of Cry protoxin is still unknown. The formation of the crystals and their solubility characteristics presumably depend on a variety of factors, but almost none of them has been identified. The primary amino acid sequences of the three segments designated as blocks 6, 7 and 8 in the C-terminal extension are highly con- served among members of the large protoxin group [1]. Therefore, the C-terminal extension is thought to con- tain important elements responsible for protein crystal- lization. In addition, the cysteines located in the C-terminal extension may form intermolecular disulfide bridges in the crystal lattice [14], which may result in the solubilization in the presence of reducing agents at the high pH that is typical for the midgut juice of lepi- dopteran and dipteran larvae [15]. In lepidopteran- specific Cry1Ab, 14 of the 16 cysteines are localized to the C-terminal extension, and the replacement of one cysteine to a heterogeneous amino acid affects the for- mation of the crystal [16]. These observations suggest that the cysteine-rich C-terminal extension is involved in the crystallization and stabilization of Cry toxin. The C-terminal extension of Cry4Aa is structurally similar to that of Cry1 [17]. Eight of the 13 cysteines of Cry4Aa are located in the C-terminal extension. These observations suggest that Cry1 and Cry4Aa have similar crystallization mechanisms. In the present study, we discovered that polypeptides containing the conserved block 7 sequence of Cry4Aa facilitated the formation of protein inclusions of fused heterologous protein in E. coli. Therefore, the block 7 sequence may be one of the factors responsible for the crystallization of Cry toxins. Because no cysteines are located in block 7 of Cry4Aa, the mechanism mediated by block 7 was independent of intermolecular disulfide bridges. GST fused with 4AaCter 696–851, which contains block 6 in addition to block 7, formed crystal-like inclusion bodies with almost 100% efficiency and had much greater enzyme activity than GST fused with 4AaCter 801–834, which contains only block 7. This result suggests that some other sequence stretch in the C-terminal extension, such as block 6, also had a role in maintaining the structural stability of GST incorpo- rated into the inclusions. The toxicity of the Cry1Ab C-terminal extension against E. coli and Agrobacte- rium tumefaciens has been reported previously [18]. The toxicity is reduced or neutralized when the C-terminal region is fused with domain III of the insecticidal N-terminal region. However, 4AaCter 696– 851 by itself was expressed and formed inclusions in E. coli as efficiently as GST–4AaCter 696–851 (data not shown). This observation suggests that the C-terminal extensions of each Cry toxin may have dif- ferent functions. We used the segment 696–851 of the C-terminal half of Cry4Aa as a peptide tag for the efficient pro- duction of T. pallidum antigens (TpNs). This experi- ment was a touchstone to evaluate the usability of 4AaCter 696–851 as a peptide tag for efficient protein production. The use of this peptide tag resulted in increased expression efficiency and simplified purifica- tion steps. The reactivity of the recombinant TpNs against human serum was similar to the reactivity of native TpNs, as estimated by western blotting and ELISA. Although fusion of 4AaCter 696–851 at the C-terminal end of TpNs conferred almost no positive effect, fusion at the N-terminal end caused a dramatic increase in the expression level and the formation of the inclusions that were easily solubilized in alkali buf- fer (pH 11). The reason why such differences were observed between fusions at the N- and C-terminal ends of TpNs remains to be resolved. The toxic domain of TpNs may be in the N-terminal region and 4AaCter at the N-terminal end may neutralize the toxicity of TpNs, but 4AaCter at the C-terminal end may not. In general, because some TpNs are hard to express without the appropriate tag, TpNs have been produced using a tag such as GST derived from Schistosoma ja- ponicum [7]. However, S. japonicum is one of the human parasites and there is a possibility that the prod- ucts prepared using the GST tag react with the human serum and cause objective background. On the other hand, 4AaCter 696–851 is derived from Cry toxin produced by the soil bacterium B. thurigiensis, and may not react with the human serum. As expected, the peptide of 4AaCter 696–851 did not show specific reactivity against the human serum. This was a great advantage of using the recombinant TpNs prepared by our method for diagnostic purposes. Thus, the use of peptide tag 4AaCter 696–851 is very simple and applicable for the efficient production of heterologous proteins in E. coli. Efficient protein production using 4AaCter T. Hayakawa et al. 2888 FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS Materials and methods Construction of GST expression vectors Nine peptide sequences derived from the Cry4Aa C-termi- nal region were fused with GST and expressed in E. coli strain BL21. Briefly, the DNA fragment encoding the poly- peptide from I696 to P851 of Cry4Aa was amplified by PCR. A synthetic cry4Aa gene, cry4Aa-S2 (AB513706), was used as a template. The amplified fragment was inserted in frame into the XhoI site of the expression vector pGEX- 6P-1 (GE Healthcare, Little Chalfont, UK) to generate pGST–4AaCter 696–851. Similarly, to construct pGST– 4AaCter 852–1180, the DNA fragment encoding the poly- peptide from A852 to E1180 of Cry4Aa was excised from cry4Aa-S2 by NaeI cleavage and inserted into the SmaI site of pGEX-6P-1. The recombinant plasmids pGST– 4AaCter 801–851 and 696–799 were constructed by removing the EcoRI–KpnI and KpnI–NaeI segments, respectively, from pGST–4AaCter 696–851. pGST– 4AaCter 801–834 and 801–829 were constructed through a 1 day mutagenesis procedure [19] using pGST–4AaCter 801–851 as a template. To construct pGST–4AaCter 807– 824, 807–819 and 812–824, the DNA fragments encoding the polypeptides F807 to R824, F807 to T819 and E812 to R824 of Cry4Aa, respectively, were synthesized by using a recursive PCR procedure [20,21], in which over- lapping oligonucleotides matching parts of the sense and antisense strands of the DNA sequence were used. The synthesized fragment was inserted between the BamHI and XhoI sites of pGEX-6P-1. Nucleotide sequences of the recombinant plasmids were confirmed with an ABI PRISMÔ 310 genetic analyser (Applied Biosystems, Foster City, CA, USA). Characterization of GST–4AaCters Expression of the GST–4AaCters was induced with 0.06 mm isopropyl b-d-1-thiogalactopyranoside at 37 °C for 3 h. The E. coli cells were disrupted by sonication, and sol- uble and insoluble proteins were separated by centrifuga- tion (12 000 g, 5 min, 4 °C). Proteins were separated by SDS ⁄ PAGE (15%) and then visualized by Coomassie bril- liant blue (CBB) staining. Protein concentrations were esti- mated by using a protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA) with bovine serum albumin as a stan- dard or by using the densitometric scanning method with a cooled CCD camera system (Ez-Capture; ATTO, Tokyo, Japan) and image analysis software (CS analyser ver. 3.0; ATTO, Tokyo, Japan). The insoluble pellet was washed twice with water and sol- ubilized in alkali buffer (50 mm NaHCO 3 , NaOH, pH 11) for 1 h at 37 °C. GST activities of GST–4AaCters were analysed by using glutathione and CDNB, as described pre- viously [22]. GST purified by using glutathione sepharose, as described previously [21], was used as a positive control for the assay. The reaction was performed in a well of a 96-well microtitre plate in a final volume of 200 lL, which contained 100 mm phosphate buffer (pH 7.4), 1 mm gluta- thione, 1 mm CDNB and an appropriate amount of alkali- soluble recombinant protein. After incubation for 1 min at room temperature, the absorbance was read once every minute at 340 nm for 5 min. The changes in absorbance per minute were converted into mol substrate conju- gatedÆmin )1 Ænmol )1 protein by using the molar extinction coefficient e 340 = 9.8 mm )1 Æcm )1 . Microscopic observation Escherichia coli cells expressing GST–4AaCters were observed under a light microscope (Axio Observer A1; Carl Zeiss, Go ¨ ttingen, Germany). In addition, the bacterial cells were disrupted, and crystal-like inclusion bodies were har- vested by centrifugation and fixed with phosphate-buffered saline containing 2% glutaraldehyde and 2% formaldehyde for 8 h. The samples were sequentially immersed in 50– 99.5% ethanol and then soaked in 100% t-butanol. The samples were allowed to dry and then coated with OsO 4 by using an osmium coater HPC-1S (Vacuum Device Inc., Mito, Japan). The scanning electron micrographs were taken on a S-900 scanning electron microscope (Hitachi Kyowa Engineering Co., Ltd., Hitachi, Japan) at a voltage of 20 kV. Construction of TpNs expression vector The synthetic TpN15, TpN17 and TpN47 genes were PCR tailed with BamHI and XhoI cleavage sequences at the upstream and downstream termini, respectively. The DNA fragment encoding 4AaCter 696–851 was PCR amplified and tailed with the BamHI or XhoI cleavage sequence at both termini. In addition, the amplified DNA fragment was designed to add a short peptide encoding 6 · His and the recognition site of PreScission protease (GE Healthcare, Little Chalfont, UK) upstream and downstream ends of 4AaCter, respectively. pGEX–DGST was used for the expression of TpN anti- gens. pGEX–DGST was constructed by removing the GST coding region, except for the first ATG, from the expression vector pGEX–4T-3 (GE Healthcare, Little Chalfont, UK) using a 1 day mutagenesis procedure [19]. Specific primer pairs, pDGST–4T-3MB-f (GGATCCCCGAATTCCCGG) and pDGST–4T-3MB-r (CATGAATACTGTTTCCTG), were used for PCR. The synthetic TpN genes were inserted between the BamHI and XhoI sites of pGEX–DGST to gener- ate pDGST–TpNs, and then the DNA fragment encoding 4AaCter was inserted into the BamHI or XhoI site of pDGST–TpNs to generate pDGST–4AaCter–TpNs. T. Hayakawa et al. Efficient protein production using 4AaCter FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS 2889 Preparation of 4AaCter–TpNs 4AaCter–TpNs were expressed in E. coli BL21. The crystal- like inclusion bodies were harvested by centrifugation and solubilized in alkali buffer (50 mm NaHCO 3 , NaOH, pH 11) for 1 h at 37 °C. The 4AaCter–TpNs were purified using a Ni-NTA column (Invitrogen, Carlsbad, CA, USA) and incubated with PreScission protease at a concentration of 1 U Æ100 lg )1 protein to remove the 6 · His–4AaCter tag. The released 6 · His–4AaCter tag and undigested 4AaCter–TpNs were removed using a Ni-NTA column. The 4AaCter–TpNs and the purified recombinant TpNs were analysed by western blotting with anti-TpN human antiserum (ProMedDx, Norton, MA, USA). ELISA In total, 100 lL recombinant TpNs (1 lgÆ mL )1 ) per well was incubated overnight in a 96-well microplate at 4 °C. Native T. pallidum was prepared from rabbit testis, as described pre- viously [23]. Native TpNs used as a positive control was extracted in the extraction buffer (50 mm Tris ⁄ HCl, 5 mm EDTA, 3% n-octyl-b-d-thioglucoside) and purified using Superdex 200 (GE Healthcare, Little Chalfont, UK). The wells were washed with phosphate-buffered saline containing 0.05% Tween 20 (PBST) and then treated with blocking solution (PBST containing 4% Block Ace; DS Pharma Bio- medical, Osaka, Japan) for 2 h at room temperature. The reactivity of TpN antigens was evaluated with eight different human serum samples that were reactive against T. pallidum antigens (ProMedDx, Norton, MA, USA) and 10 healthy donor serum samples (Uniglobe Research, Reseda, CA, USA). Accurun Ò Series Infectrol Ò D-00, which is of human serum origin and is reactive against T. pallidum antigens (SeraCare Life Science, Norton, MA, USA), was also used for the assay. Native T. pallidum antigens prepared from rabbit testis were used as a positive control. After washing the wells with PBST, 100 lL human antiserum diluted 200- fold in blocking solution was added to each well for 1 h at room temperature. The wells were washed with PBST and diluted (·5000) goat anti-human Ig conjugated with horse- radish peroxidase (Biosource, Camarillo, CA, USA) was added to the wells and incubated for another 1 h at room temperature. After incubation with phosphate citrate buffer containing o-phenylenediamine (0.4 mgÆmL )1 ) and H 2 O 2 (2 lLÆmL )1 ) for 20 min at room temperature, the reaction was stopped by adding 100 lL1m sulfuric acid, and the absorbance of each well at 490 nm was measured. Acknowledgement This work was supported in part by a grant from the Kurata Memorial Hitachi Science and Technology Foundation. References 1 Schnepf E, Crickmore N, Van RieJ, Lereclus D, Baum J, Feitelson J, Zeigler DR & Dean DH (1998) Bacillus thuringiensis and its pesticidal crystal proteins. Microbiol Mol Biol Rev 62, 775–806. 2 Agaisse H & Lereclus D (1995) How Bacillus thuringien- sis produce so much insecticidal crystal protein? J Bacteriol 177, 6027–6032. 3 Chak K-F & Ellar DJ (1987) Cloning and expression in Escherichia coli of an insecticidal crystal protein gene from Bacillus thuringiensis var. aizawai HD-133. J Gen Microbiol 133, 2921–2931. 4 Oeda K, Oshie K, Shimizu M, Nakamura K, Yamam- oto H, Nakayama I & Ohkawa H (1987) Nucleotide sequence of the insecticidal protein gene of Bacillus thuringiensis strain aizawai IPL-7 and its high-level expression in Escherichia coli. Gene 53, 113–119. 5 Lecadet M-M (1967) Action compare ´ e de l’ure ´ eetdu thioglycolate sur la toxine figure ´ edeBacillus thuringien- sis. C R Acad Sci Hebd Seances Acad Sci D 264, 2847– 2850. 6 Yamagiwa M, Esaki M, Otake K, Inagaki M, Komano T, Amachi T & Sakai H (1999) Activation process of dipteran-specific insecticidal protein produced by Bacillus thuiringiensis subsp. israelensis. Appl Environ Microbiol 65, 3464–3469. 7 Fujimura K, Ise N, Ueno E, Hori T, Fujii N & Okada M (1997) Reactivity of recombinant Treponema palli- dum (r-Tp) antigens with anti-Tp antibodies in human syphilitic sera evaluated by ELISA. J Clin Lab Anal 11, 315–322. 8 Boonserm P, Davis P, Ellar DJ & Li J (2005) Crystal structure of the mosquito-larvicidal toxin Cry4Ba and its biological implications. J Mol Biol 348, 363–382. 9 Boonserm P, Mo M, Angsuthanasombat C & Lescar J (2006) Structure of the functional form of the mosquito larvicidal Cry4Aa toxin from Bacillus thuringiensis at a 2.8-angstrom resolution. J Bacteriol 188, 3391–3401. 10 Galitsky N, Cody V, Wojtczak A, Ghosh D, Luft JR, Pangborn W & English L (2001) Structure of the insec- ticidal bacterial delta-endotoxin Cry3Bb1 of Bacillus thuringiensis. Acta Crystallogr D Biol Crystallogr 57, 1101–1109. 11 Grochulski P, Masson L, Borisova S, Pusztai-Carey M, Schwartz JL, Brousseau R & Cygler M (1995) Bacillus thuringiensis CryIA(a) insecticidal toxin: crystal struc- ture and channel formation. J Mol Biol 254, 447–464. 12 Li J, Carroll J & Ellar DJ (1991) Crystal structure of insecticidal delta-endotoxin from Bacillus thuringiensis at 2.5 A ˚ resolution. Nature 353, 815–821. 13 Morse RJ, Yamamoto T & Stroud RM (2001) Structure of Cry2Aa suggests an unexpected receptor binding epi- tope. Structure 9, 409–417. Efficient protein production using 4AaCter T. Hayakawa et al. 2890 FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS 14 Nickerson KW (1980) Structure and function of the Bacillus thuringiensis protein crystal. Biotechnol Bioeng 22, 1305–1333. 15 Bietlot HP, Vishnubhatla I, Carey PR, Pozsgay M & Kaplan H (1990) Characterization of the cysteine residues and disulphide linkages in the protein crystal of Bacillus thuringiensis. Biochem J 267, 309–315. 16 Yu J, Xie R, Tan L, Xu W, Zeng S, Chen J, Tang M & Pang Y (2002) Expression of the full-length and 3¢-spliced cry1Ab gene in the 135-kDa crystal protein minus derivative of Bacillus thuringiensis subsp. kyushu- ensis. Curr Microbiol 45, 133–138. 17 Ho ¨ fte H & Whiteley HR (1989) Insecticidal crystal pro- teins of Bacillus thuringiensis. Microbiol Rev 53, 242– 255. 18 Vazquez-Padron RI, de la Riva G, Agu ¨ ero G, Silva Y, Pham SM, Sobero ´ n M, Bravo A & Aı ¨ touche A (2004) Cryptic endotoxic nature of Bacillus thuringiensis Cry1Ab insecticidal crystal protein. FEBS Lett 570, 30–36. 19 Imai Y (1999) Wan dei mutajeneshisu. Jikkenigaku (in Japanese) 17, 771–775. 20 Prodromou C & Pearl LH (1992) Recursive PCR: A novel technique for total gene synthesis. Protein Eng 5, 827–829. 21 Hayakawa T, Howlader MTH, Yamagiwa M & Sakai H (2008) Design and construction of a synthetic Bacillus thuringiensis Cry4Aa gene: Hyper-expression in Escherichia coli. Appl Microbiol Biotechnol 80, 1033–1037. 22 Habig WH, Pabst MJ & Jakoby WB (1974) Glutathione S-transferases. The first enzymatic step in mercapturic acid formation. J Biol Chem 249, 7130– 7139. 23 Hanff PA, Norris SJ, Lovett MA & Miller JN (1984) Purification of Treponema pallidum, Nichols strain, by Percoll density gradient centrifugation. Sex Transm Dis 11, 275–286. T. Hayakawa et al. Efficient protein production using 4AaCter FEBS Journal 277 (2010) 2883–2891 ª 2010 The Authors Journal compilation ª 2010 FEBS 2891 . Novel strategy for protein production using a peptide tag derived from Bacillus thuringiensis Cry4Aa Tohru Hayakawa 1 , Shinya Sato 1,2 , Shigehisa Iwamoto 2 , Shigeo Sudo 2 , Yoshiki Sakamoto 1 , Takaaki. Sakamoto 1 , Takaaki Yamashita 1 , Motoaki Uchida 1 , Kenji Matsushima 1 , Yohko Kashino 1 and Hiroshi Sakai 1 1 Graduate School of Natural Science and Technology, Okayama University, Japan 2 Department. 3-1-1 Tsushima-Naka, Okayama 700-8530, Japan Fax ⁄ Tel: +81 86 251 8203 E-mail: sakahrsh@biotech.okayama-u.ac.jp Database The nucleotide sequence of the synthetic gene cry4Aa- S2 is available in the EMBL