Engineering of monomeric FK506-binding protein 22 with peptidyl prolyl cis-trans isomerase Importance of a V-shaped dimeric structure for binding to protein substrate Cahyo Budiman 1 , Keisuke Bando 1 , Clement Angkawidjaja 1 , Yuichi Koga 1 , Kazufumi Takano 1,2 and Shigenori Kanaya 1 1 Department of Material and Life Science, Graduate School of Engineering, Osaka University, Yamadaoka, Suita, Osaka, Japan 2 CRESTO, JST, Yamadaoka, Suita, Osaka, Japan Keywords FKBP22; homodimer; peptidyl-prolyl cis-trans isomerase (PPIase); protein engineering; substrate binding Correspondence S. Kanaya, Department of Material and Life Science, Graduate School of Engineering, Osaka University, 2-1, Yamadaoka, Suita, Osaka 565-0871, Japan Tel ⁄ Fax: +81 6 6879 7938 E-mail: kanaya@mls.eng.osaka-u.ac.jp (Received 1 May 2009, revised 24 May 2009, accepted 28 May 2009) doi:10.1111/j.1742-4658.2009.07116.x FK506-binding protein 22 (FKBP22) from the psychrotrophic bacterium Shewanella sp. SIB1 is a homodimeric protein with peptidyl prolyl cis–trans isomerase (PPIase) (EC 5.2.1.8) activity. Each monomer consists of 205 amino acid residues. According to a tertiary model, SIB1 FKBP22 assumes a V-shaped structure, in which two monomers interact with each other at their N-termini. Each monomer consists of an N-terminal domain with a dimerization core and a C-terminal catalytic domain, which are separated by a 40-residue-long a-helix. To clarify the role of this V-shaped structure, we constructed a mutant protein, in which the N-domain is tandemly repeated through a flexible linker. This protein, termed NNC-FKBP22, is designed such that two repetitive N-domains are folded into a structure similar to that of the Shewanella sp. SIB1 FKBP22 wild-type protein (WT). NNC-FKBP22 was overproduced in Escherichia coli in a His-tagged form, purified and biochemically characterized. Gel-filtration chromatography and ultracentrifugation analyses indicate that NNC-FKBP22 exists as a monomer. Analysis of thermal denaturation using differential scanning cal- orimetry indicates that NNC-FKBP22 unfolds with two transitions, as does the WT protein. NNC-FKBP22 exhibited PPIase activity for both peptide and protein substrates. However, in contrast to its activity for peptide sub- strate, which was comparable to that of the WT protein, its activity for protein substrate was reduced by five- to six-fold, compared to that of the WT. Surface plasmon resonance analyses indicate that NNC-FKBP22 binds to a reduced form of a-lactalbumin with a six-fold weaker affinity than that of WT. These results suggest that a V-shaped structure of SIB1 FKBP22 is important for efficient binding to a protein substrate. Structured digital abstract l MINT-7136140: FKBP22 (uniprotkb:Q765B0) binds (MI:0407)toAlpha-lactalbumin (uni- protkb: P00711)bysurface plasmon resonance (MI:0107) Abbreviations DSC, differential scanning calorimetry; FKBP, FK506-binding protein; MIP, macrophage-infectivity potentiator; pNA, p-nitroanilide; PPIase, peptidyl prolyl cis–trans isomerase; RCM, reduced and carboxymethylated; suc-ALPF-pNA, N-succinyl-Ala-Leu-Pro-Phe-p-nitroanilide; WT, Shewanella sp. SIB1 FKBP22 wild-type protein. FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS 4091 Introduction Peptidyl prolyl cis-trans isomerase (PPIase) (EC 5.2.1.8) catalyzes cis–trans isomerization of the Xaa– Pro peptide bonds of proteins [1]. Because the peptide bond in the cis conformation is energetically unfavor- able compared with that in the trans conformation [2,3], and cis–trans isomerization of this peptide bond is intrinsically very slow [4], prolyl isomerization is regarded as a rate-limiting step of the folding reaction of proteins that contain cis prolines in a folded state [5]. PPIases are divided into four structurally unrelated families: FK506-binding proteins (FKBPs), cyclophi- lines, parvulins and the Ser ⁄ Thr phosphatase 2A activator, PTPA [6]. FKBP22 is a member of the FKBP family and pres- ent in various Gram-negative bacteria [7–9]. It is a homodimer, in which each subunit consists of an N-terminal domain (N-domain) and a C-terminal PPIase domain (C-domain). Based on its similarities to the macrophage-infectivity potentiator (MIP) protein from Legionella pneumophila in amino acid sequence, FKBP22 has been classified as a member of the MIP- like FKBP subfamily [7]. Escherichia coli FkpA is also a member of this subfamily [10]. Of the members of this subfamily, L. pneumophila MIP [11] and E. coli FkpA [12] are the only ones for which the crystal structures have been determined. These structures strongly resemble each other, having an rmsd of 0.8 A ˚ for all Ca atoms. According to these structures, these proteins assume a nonglobular V-shaped homodimeric structure, in which two monomers interact with each other at their N-domains. Each monomer assumes a dumbbell-like structure, in which the N-domain (con- sisting of a1 and a2 helices) and the C-domain [con- sisting of six b strands (b1–b6) and an a4 helix] are linked by a 40-residue-long a3 helix. As a result, the C-domains, which are located at both ends of a V-shaped structure, face each other across the cleft of this structure. The interface of two monomers, which is located at the bottom of the V-shaped structure, is stabilized by the hydrophobic interactions between the a1 helix of one monomer and the a2 helix of the other. However, the role of a V-shaped structure of MIP-like FKBP subfamily proteins remains to be understood. We have previously shown that FKBP22 from the psychrotrophic bacterium Shewanella sp. SIB1 exists as a homodimer and exhibits PPIase activity for both peptide and protein substrates [8]. SIB1 FKBP22 shows amino acid sequence identities of 56% to E. coli FKBP22 [7], of 43% to E. coli FkpA [10] and of 41% to L. pneumophila MIP [13]. Construction of the mutant proteins N-domain + and C-domain + , which lack the C- and N-domains of SIB1 FKBP22, respectively, followed by biochemical characterization of these proteins, suggest that the C-domain is required for PPIase activity, and the N-domain is required for dimerization and efficient binding to a protein substrate of PPIase [14,15]. However, it remains to be determined whether a V-shaped struc- ture is required for efficient binding of SIB1 FKBP22 to a protein substrate, because the N-domain + , which retains the ability to bind to a protein substrate, con- tains the entire a3 helix and therefore may be able to assume a V-shaped homodimeric structure. Attempts to construct the N-domain without the a3 helix have so far been unsuccessful because of the instability of the protein. In this report, we constructed the mutant protein, NNC-FKBP22, in which the N-terminal region (Met8– Ala60) is repeated twice within a single polypeptide chain, and characterized it biochemically. This mutant protein was designed such that the repetitive N-termi- nal region is folded into a structure similar to that of the Shewanella sp. SIB1 FKBP22 wild-type protein (WT), which has a homodimeric structure. Based on these results, we discuss the role of a V-shaped struc- ture of FKBP22. Results Design of monomeric mutant protein The monomeric mutant protein (NNC-FKBP22) was designed based on a model of the 3D structure of WT (SIB1 FKBP22) (Fig. 1A), which has previously been reported [14]. According to this model, WT assumes a V-shaped homodimeric structure, like those of L. pneu- mophila MIP [11] and E. coli FkpA [12]. In this struc- ture, the Ala60 of one monomer is located in close proximity to the Met8 of the other monomer. Both residues are located close to the bottom of the cleft of the V-shaped structure. Therefore, it is strongly expected that the mutant protein, termed NNC- FKBP22, in which Met1–Ala60 of SIB1 FKBP22 is attached to Met8–Ile205 of SIB1 FKBP22 through three glycine residues, is monomeric and folded into a structure similar to that of WT without one arm of the ‘V’. A model of the 3D structure of NNC-FKBP22 is shown in Fig. 1A. Its primary structure is also sche- matically shown in Fig. 1B in comparison with that of WT. NNC-FKBP22 and SIB1 FKBP22 (WT) in a His- tagged form will be designated as NNC-FKBP22 and SIB1 FKBP22 (WT), respectively, hereafter. Engineering of monomeric FKBP22 C. Budiman et al. 4092 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS NNC-FKBP22 was overproduced in E. coli at 10 °C, as previously reported for WT [8]. The protein accumulated in the E. coli cells in a soluble form and was purified to give a single band on SDS–PAGE (see, Fig. S1). WT was also overproduced and purified as reported previously [8]. Determination of oligomeric state The molecular mass values of NNC-FKBP22 and WT were estimated to be 34 and 28 kDa, respectively, from SDS–PAGE. These values are considerably higher than those calculated from their amino acid sequences (27 kDa for NNC-FKBP22 and 21 kDa for WT). It has been reported for WT that the molecular mass of the protein determined by ESI-MS (23 kDa) is compa- rable to the calculated molecular mass and therefore the molecular mass of the protein estimated from SDS–PAGE is considerably higher than the calculated molecular mass as a result of its unusual behavior on SDS–PAGE [8]. Because the difference in molecular mass values of NNC-FKBP22 and WT, estimated from SDS–PAGE, is comparable to the difference in their molecular mass values calculated from amino acid sequences, the molecular mass of NNC-FKBP22 is considerably higher than the calculated molecular mass, probably as a result of its unusual behavior on SDS–PAGE, like WT. The molecular mass values of native forms of NNC-FKBP22 and WT were estimated to be 58 and 99 kDa, respectively, from gel-filtration column chromatography. These values are 4.5- and 2.1-fold higher than those calculated from their amino acid sequences. However, it has been reported for WT that the molecular mass of the protein, as determined by sedimentation equilibrium analytical ultracentrifu- gation (44 kDa), is two-fold higher than the calculated molecular mass, and the discrepancy between the molecular mass values estimated from gel filtration and analytical ultracentrifugation is a result of the unusual behavior of WT on gel filtration. The unusual behavior of the protein on gel filtration has also been reported for L. pneumophila MIP. The molecular mass of this protein, as estimated from gel filtration, is higher than the calculated molecular mass by 2.7-fold, instead of by two-folds, because it is cylindrical rather than globular [16]. Therefore, the molecular mass 9 23 34 45 52 93 95 99 102 111 121 130 134 139 146 150 155 163 167 174 195 205 α1 α2 α3 β1 β2 β3 β4a β4b β5 β6 α4 1 N-domain B C-domain 65 79 90 101 108 149 151 155 158 167 177 186 190 195 202 206 211 219 223 230 251 261 α4 α5 α6 β1 β2 β3 β4a β4b β5 β6 α7 9 23 34 45 52 α1 α2 α3 1 60 GGG C-domain N-domain SIB1 FKBP22 NNC-FKBP22 C-domain Ala60 (Gly) 3 Met64 N-domain N-domain C-domain C-domain Ala60 Met8 SIB1 FKBP22 NNC-FKBP22 α 1 α 5 α 4 α 2 α 2 α 1 α 2 α 1 α 3 α 6 α 3 N-domain A Fig. 1. (A) 3D structure models of SIB1 FKBP22 and NNC-FKBP22. For the SIB1 FKBP22 structure, one monomer is deeply- colored, while the other is lightly-colored. The N- and C-domains and the a1-3 helices are indicated. The side chain of Met8 of one monomer and that of Ala60 of the other monomer are indicated by stick models. For the NNC-FKBP22 structure, the correspond- ing domains, helices and side chains of the amino acid residues are indicated. A loop consisting of three glycine residues, which connects Ala60 and Met64 (corresponding to Met8 of SIB1 FKBP22), is schematically shown in cyan. (B) Schematic representa- tions of the primary structures of SIB1 FKBP22 and NNC-FKBP22. A His-tag attached to the N-termini of the proteins is represented by a shaded box. The a-helices and b-strands are represented by cylinders and arrows, respectively. These secondary structures are arranged based on tertiary models of SIB1 FKBP22 and NNC-FKBP22. Numbers indicate the positions of the resi- dues relative to the initiator methionine resi- due of the proteins without a His-tag. The ranges of the N- and C-domains are also shown. C. Budiman et al. Engineering of monomeric FKBP22 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS 4093 values of native forms of NNC-FKBP22 and WT were also determined using sedimentation equilibrium analytical ultracentrifugation. The data fitted well to a single-species model, and showed no evidence of aggregation, and the molecular mass values of NNC- FKBP22 and WT were determined as 30.4 and 43.9 kDa, respectively. These values are 1.1- and 1.9-fold larger than the calculated molecular mass values, indicating that NNC-FKBP22 and WT exist as a monomer and a dimer in solution, respectively. CD spectra The far- and near-UV CD spectra of NNC-FKBP22 and WT were measured at 10 °C. The far-UV CD spectrum of NNC-FKBP22 is similar to that of WT (Fig. 2A). However, the depth of its trough is slightly larger than that of WT. From these spectra, the helical contents of NNC-FKBP22 and WT are estimated to be 44 and 38%, respectively, using the method of Wu et al. [17], which are comparable to those calculated from their tertiary models (54% for NNC-FKBP22 and 49% for WT). The near-UV CD spectrum of NNC-FKBP22 is also similar to that of WT, although the height of its peak is slightly larger than that of WT (Fig. 2B). The near-UV CD spectra reflect the 3D environments of aromatic residues, such as Tyr and Trp. WT contains one Trp residue and seven Tyr resi- dues in each monomer. This tryptophan residue (Trp157) is conserved in the C-domain of MIP-like FKBP subfamily proteins and is required for PPIase activity. In addition, six of the seven Tyr residues are located in the C-domain. Therefore, the near-UV CD spectra of these proteins mainly reflect the conforma- tion of the C-domain. These results suggest that the 3D structure of NNC-FKBP22 is similar to that of WT, except that the a3 helix and C-domain of one monomer are removed. Thermodynamics of unfolding SIB1 FKBP22 is thermally denatured with two transi- tions, the first and the second ones for denaturation of the C- and N-domains, respectively [14]. To examine whether NNC-FKBP22 gives a similar thermal-dena- turation curve, heat-induced unfolding of NNC-FKBP and WT was analyzed by differential scanning calorim- etry (DSC). Thermal unfolding of these proteins was highly reversible, as indicated by repeating thermal scans to reproduce DSC curves. Both proteins gave denaturation curves with two well-separated transitions (Fig. 3), suggesting that both domains of NNC- FKBP22 are folded into structures similar to those of WT. However, both domains of NNC-FKBP22 are apparently more stable than those of WT. Deconvolu- tion of the thermogram of NNC-FKBP22, according to a non two-state denaturation model, gives melting temperature (T m ) values of 36.9 and 50.3 °C for the (θ) (deg cm 2 ·dmol –1 ) Wavelength (nm) –20 000 –15 000 –10 000 –5000 0 5000 A B 200 220 240 260 Wavelength (nm) –400 –200 0 200 400 600 260 280 300 320 Fig. 2. CD spectra of NNC-FKBP22. The far-UV (A) and near-UV (B) CD spectra of NNC-FKBP22 (thick line) are shown in comparison with those of SIB1 FKBP22 (thin line). Both spectra were measured at 10 °C as described in the Experimental procedures. Temperature (°C) 0 5 10 Cp (kJ·mol –1 ·K –1 ) 15 20 25 20 30 40 50 60 70 Fig. 3. DSC curve of NNC-FKBP22. The DSC curve of NNC- FKBP22 (thick line) is shown in comparison with that of SIB1 FKBP22 (thin line). These curves were measured at a scan rate of 1°CÆmin )1 . Both proteins were dissolved in 20 mM sodium phos- phate (pH 8.0) at approximately 1 mgÆmL )1 . Engineering of monomeric FKBP22 C. Budiman et al. 4094 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS first and second transitions, which probably reflect denaturation of the C- and N-domains, respectively. These values are higher than the corresponding values of WT by approximately 4 °C (Table 1). PPIase activity The PPIase activitiy for peptide substrate was deter- mined using N-succinyl-Ala-Leu-Pro-Phe-p-nitroanilide (Suc-ALPF-pNA) as a substrate. The catalytic efficien- cies (k cat ⁄ K m ) of NNC-FKBP22 and for each mono- mer of WT were 1.08 ± 0.07 and 0.80 ± 0.04 lm )1 Æs )1 , respectively, indicating that the catalytic efficiency of NNC-FKBP22 is slightly higher than, but comparable to, that of WT. The temperature depen- dence of the PPIase activity of NNC-FKBP22 was almost identical to that of WT (Fig. 4A). By contrast, when the PPIase activity was determined by a refold- ing assay using RNase T 1 as a protein substrate, NNC-FKBP22 exhibited much lower activity than that of WT. RNase T 1 has been widely used as a protein substrate for PPIase activity, because cis–trans isomeri- zation of two peptidyl prolyl bonds (Tyr38–Pro39 and Ser54–Pro55) of RNase T 1 is a rate-limiting step of its folding [18–20]. The refolding of RNase T 1 was not seriously accelerated in the presence of 10 nm NNC- FKBP22 (Fig. 4B), while it was significantly acceler- ated in the presence of 75 nm NNC-FKBP22 to a level similar to that observed in the presence of 10 nm WT (data not shown). The k cat ⁄ K m values of NNC- FKBP22 and WT were estimated to be 0.08 ± 0.005 and 0.53 ± 0.03 lm )1 Æs )1 , respectively. Binding to reduced a-lactalbumin a-Lactalbumin is stabilized by four disulfide bonds and a single Ca 2+ ion [21,22], and therefore reduction of these disulfide bonds produces a protein with a par- tially folded molten globule-like structure [23]. Reduced and carboxymethylated (RCM) a-lactalbumin has been used as a folding intermediate of proteins to analyze the chaperone functions of GroEL [24,25] and FKBP family proteins [15,26,27]. RCM a-lactalbumin has been shown to compete with the protein substrate of PPIase for binding to the trigger factor [26] and E. coli FkpA [27], suggesting that RCM a-lactalbumin and a protein substrate of PPIase share a common binding site of FKBP family proteins. In order to examine whether NNC-FKBP22 binds to a protein substrate with similar affinity as that of WT, the binding affinities of NNC-FKBP22 and WT to reduced a-lactalbumin were analyzed using surface plasmon resonance (Biacore). Reduced a-lactalbumin was injected onto the sensor chip, on which NNC-FKBP22 or WT was immobilized. The amount of protein Table 1. Thermodynamic parameters for heat-induced unfolding of the protein. The melting temperature (T m ), calorimetric enthalpy (DH cal ) and van’t Hoff enthalpy (DH vH ) of SIB1 FKBK22 and NNC- FKBP22 were obtained from the DSC curves shown in Fig. 3 using ORIGIN software (Microcal Inc.) RNase H Transition T m (°C) DH cal (kJÆmol )1 ) DH vH (kJÆmol )1 ) SIB1 FKBP22 First 32.5 ± 0.13 83 ± 2.2 404 ± 4.1 Second 46.4 ± 0.07 195 ± 2.1 304 ± 2.0 NNC-FKBP22 First 36.9 ± 0.04 143 ± 1.4 305 ± 3.5 Second 50.3 ± 0.02 250 ± 1.5 327 ± 2.1 Temperature (°C) 0 0.5 K cat /K m (μ M –1 ·s –1 ) 1 1.5 A B 0 5 10 15 20 Time (s) 40 60 80 100 0 1000 2000 3000 Relative fluorescence (%) Fig. 4. PPIase activities of NNC-FKBP22. (A) The temperature dependence of the PPIase activity of NNC-FKBP22 (closed circle), which was determined by a protease-coupling assay using Suc- ALPF-pNA as a substrate, is shown in comparison with that of SIB1 FKBP22 (open circle). The catalytic efficiency was calculated according to Harrison & Stein [46]. The experiment was carried out in duplicate. Each plot represents the average value, and errors from the average values are shown. (B) The increase of tryptophan fluorescence at 323 nm during the refolding of RNase T 1 (0.2 lM) is shown as a function of the refolding time. Refolding was carried cout at 10 °C in the absence (broken line), or presence of 10 n M of NNC-FKBP22 (thick solid line) or SIB1 FKBP22 (thin solid line). C. Budiman et al. Engineering of monomeric FKBP22 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS 4095 immobilized on the sensor chip was equivalent to 1200 resonance units for NNC-FKBP22 and 4000 resonance units for WT. The sensorgrams obtained by injecting 100 lm of reduced a-lactalbumin onto these sensor chips are shown in Fig. 5A as a representative. Because the association and dissociation of reduced a-lactalbumin were too fast to determine the kinetic constants, such as k on and k off , accurately, the dissoci- ation constant, K D , was determined by measuring equi- librium-binding responses at various concentrations of reduced a-lactalbumin. The plots of the equilibrium- binding responses as a function of the concentration of a-lactalbumin gave a saturation curve, as shown in Fig. 5B. These plots showed a good fit to a single binding-affinity model and the K D value for binding of reduced a-lactalbumin to NNC-FKBP22 was deter- mined to be 42.5 ± 2.1 lm. This value is higher than that of WT (6.5 ± 0.38 lm) by 6.5-fold, indicating that the binding affinity of NNC-FKBP22 to a folding inter- mediate of proteins, and probably to a protein substrate, is greatly reduced compared with that of WT. Discussion Role of a V-shaped structure In this report, we showed that NNC-FKBP22 is mono- meric and that its binding affinity to reduced a-lactal- bumin and its PPIase activity for protein substrate are reduced by five- to six-fold compared with those of WT. These results strongly suggest that a V-shaped structure of the SIB1 FKBP22 homodimer is impor- tant for binding to a folding intermediate of proteins and therefore for PPIase activity for a protein sub- strate. Neither the NNC-FKBP22 nor the WT struc- ture has been determined. However, because of the high amino acid sequence similarity between SIB1 FKBP22 and E. coli FkpA [10] or L. pneumophila MIP [13], SIB1 FKBP22 might assume a V-shaped homodi- meric structure such as E. coli FkpA [12] and L. pneu- mophila MIP [11]. Hu et al. [28] have proposed a ‘Mother’s arm’ model for the substrate-binding mecha- nism of E. coli FkpA, which exhibits both PPIase and chaperone activities, based on the observation that the a3 helix is rather flexible and controls plasticity of a V-shaped structure. According to this model, two long a3 helices act as flexible ‘arms’, which can bend at the ‘elbows’ (presumably located at the middle of the a3 helix). Two catalytic domains act as ‘hands’ and the active-site residues act as ‘fingers’ for protein sub- strates. As ‘mother’ holds her ‘baby’ by bending both of her arms, a dimer form of E. coli FkpA holds a protein substrate by bending its two long a3 helices. A V-shaped structure of SIB1 FKBP22 may also be required to hold a protein substrate with a similar mechanism. The plasticity of a V-shaped structure may lead to a conformational flexibility to adopt various types of protein substrates. The importance of a V-shaped dimeric structure for binding various types of protein substrates has also been reported for a protein disulfide isomerase, DsbC, from E. coli [29]. E. coli DsbC is a homodimer of the 23-kDa protein and assumes a V-shaped structure. The structural arrangement of E. coli DsbC is similar to those of E. coli FkpA and L. pneumophila MIP, and Response units Time (s) 0 200 400 600 A B 0 20 40 60 Response units 0 100 200 300 400 500 600 0 2040608 0 100 (Reduced α-lactalbumin) (μM) Fig. 5. Binding of reduced a-lactalbumin to NNC-FKBP22 and to SIB1 FKBP22. (A) Sensorgrams from Biacore X showing the binding of reduced a-lactalbumin (100 l M) to immobilized NNC-FKBP22 (thick line) and SIB1 FKBP22 (thin line). The sensorgram showing the binding of nonreduced a-lactalbumin (100 l M) to NNC-FKBP22, which is similar to that to SIB1 FKBP22, is also shown (broken line). Injections were performed at time zero for 60 s. (B) Relation- ships between the equilibrium-binding response and the concentra- tion of reduced a-lactalbumin. The equilibrium-binding responses of NNC-FKBP22 (closed circle) and of SIB1 FKBP22 (open circle) are shown as a function of the concentration of reduced a-lactalbumin. The solid line represents the fitting curve of a single binding-site affinity model using the BIAevaluation program. Engineering of monomeric FKBP22 C. Budiman et al. 4096 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS its N- and C-domains are connected by a hinged three- turn linker helix. A characteristic common to these proteins is that the cleft of a V-shaped structure is more hydrophobic than it is externally. The cleft of a V-shaped structure of SIB1 FKBP22 is also more hydrophobic than it is externally, suggesting that SIB1 FKBP22 binds to a protein substrate mainly through hydrophobic interactions. It has been reported that slyD proteins [30–32], an archaeal FKBP17 [33,34] and FKBP12 with a heterol- ogous chaperone domain [35], all of which are FKBP family proteins, exhibit PPIase activities for both pep- tide and protein substrates when in a monomeric form. The findings from the present study, that NNC- FKBP22 with a monomeric structure exhibits PPIase activities for these substrates, is consistent with these results. However, the role of a dimeric structure of MIP-like FKBP subfamily proteins has not so far been analyzed. It has been reported that the PPIase activity of the E. coli FkpA mutant, which lacks the N-domain and therefore loses the ability to form a dimeric struc- ture, is indistinguishable from that of the wild-type protein [12,36]. However, the C-domain of E. coli FkpA tends to oligomerize [36] and therefore the role of a V-shaped dimeric structure of E. coli FkpA can- not be clearly understood by analyzing this mutant protein. Thus, this is the first report which shows the importance of a V-shaped dimeric structure of an MIP-like FKBP subfamily protein for binding to a protein substrate. However, it remains to be deter- mined whether SIB1 FKBP22 exhibits a chaperone function and, if so, whether its dimeric structure is responsible for this function. PPIase activities of NNC-FKBP22 NNC-FKBP22 exhibits 1.4-fold higher activity than WT (per monomer) for a peptide substrate. It has pre- viously been shown that the mutant protein of SIB1 FKBP22 (C-domain + ), which lacks the N-domain and exists as a monomer, exhibits 1.6-fold higher activity than WT (per monomer) [14]. These results suggest that the C-domain is sufficient for the binding and catalysis of a peptide substrate. These results also sug- gest that a V-shaped structure is not favorable for binding a peptide substrate. In this structure, freedom of each catalytic domain is probably restricted and therefore the opportunity of this domain to contact with the substrate decreases. By contrast, in a mono- meric structure, the freedom of the catalytic domain increases and therefore the opportunity of this domain to contact with the substrate increases. By contrast, for a protein substrate, the activity of NNC-FKBP22 is six-folds lower than that of WT (per monomer). In this case, only one of the two catalytic domains of WT may serve as a catalytic site because the space between them seems to be too small to accommodate two pro- tein substrates simultaneously. We have previously shown that C-domain + exhibits activity 30-fold lower than WT (per monomer) for a protein substrate [14]. Therefore, the PPIase activities of SIB1 FKBP22 and its derivatives for a protein substrate increase as fol- lows: C-domain + < NNC-FKBP22 < WT. Likewise, the binding affinities of these proteins to a folding intermediate of protein increase in this order. These results suggest that a monomeric form of FKBP22 with N- and C-domains is sufficient for PPIase activity for a protein substrate, but a V-shaped structure is required to increase it to the maximal activity. Stability of NNC-FKBP22 DSC analyses indicate that both domains of NNC- FKBP22 are more stable than the corresponding domains of WT by approximately 4 °C (Fig. 3). The repetitive N-domains of NNC-FKBP22 are presumably folded into a structure similar to that of the N-domains of WT with a homodimeric structure. This structure is more stable than that of WT, probably because a dimeric structure of the repetitive N-domains of NNC-FKBP22 is stabilized not only by hydrophobic interactions but also by covalent linkage through three glycine residues. The covalent bond is known as the strongest chemical bond contributing to protein stability [37–40]. In WT, a dimeric structure of the N-domains is stabilized only by noncovalent, mainly hydrophobic, interactions. According to the crystal structures of L. pneumophila MIP [11] and E. coli FkpA [12], the C-domain is completely sepa- rated from the N-domain. Nevertheless, the C-domain of NNC-FKBP22 is stabilized in parallel with its N-domain compared with the corresponding domains of WT. The C-domain is linked to the N-domain through the a3 helix. Therefore, the C-domain of NNC-FKBP22 is probably indirectly stabilized when the N-domain is stabilized. It is noted that the optimum temperature for the activity of NNC-FKBP22 (10 °C) is identical to that of WT, despite the fact that the catalytic domain of NNC-FKBP22 is more stable than that of WT by approximately 4 °C. According to the thermal-denatur- ation curves of NNC-FKBP22 and WT, their C-domains start to unfold at temperatures that are considerably higher than the optimum temperatures for the activities of these proteins. These results suggest that the local conformation around the active C. Budiman et al. Engineering of monomeric FKBP22 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS 4097 site is more sensitive to thermal denaturation than the entire domain structure. The stability of the local conformation around the active site of NNC-FKBP22 may not be seriously changed compared with that of WT. Experimental procedures Plasmid construction Plasmid pSIB1-NNC, used to overproduce a His-tagged form of NNC-FKBP22, was constructed using the PCR overlap extension method [41]. Plasmid pSIB1, used to overproduce a His-tagged form of SIB1 FKBP22 [8], was used as a template. The sequences of the PCR primers used are as follows: 5¢-AGAGAGAATT CATATGTCAGATT TGTTCAG-3¢ for primer 1; 5¢-TTCCATACCACCACCT GCAACTTGAAGCTC-3¢ for primer 2; 5¢-GTTGCAGGT GGTGGTATGGAACAGCATGCT-3¢ for primer 3; and 5¢-GGCCACT GGATCCAACTACAGCAATTCTCA-3¢ for primer 4 [the NdeI (primer 1) and BamHI (primer 4) sites are underlined]. Primers 1 (forward) and 2 (reverse) were used to amplify the gene encoding Met1–Ala60 of SIB1 FKBP22, with three additional glycine residues at the C-terminus. Primers 3 (forward) and 4 (reverse) were used to amplify the gene encoding Met8–Ile205, with three addi- tional glycine residues at the N-terminus. The resultant two PCR fragments were combined and amplified by PCR using primers 1 and 4. The PCR product was ligated into the NdeI–BamHI sites of pET28a (Novagen, WI, USA) to pro- duce pSIB1-NNC. PCR was performed with the GeneAmp PCR system 2400 (Applied Biosystems, Tokyo, Japan) using KOD polymerase (Toyobo Co., Ltd., Kyoto, Japan). The nucleotide sequence was confirmed using the Prism 310 DNA sequencer (Applied Biosystems). All oligonucleotides were synthesized by Hokkaido System Science (Sapporo, Japan). Overproduction and purification E. coli BL21(DE3) [F ) ompT hsdS B (r B ) m B ) ) gal (kcI857 ind1 Sam7 nin5 lacUV5-T7gene1) dcm (DE3)] (Novagen) was used as a host strain for the overproduction of His- tagged SIB1 FKBP22 and NNC-FKBP22. Transformation of the E. coli cells with plasmid pSIB1 or pSIB1-NNC, and overproduction and purification of the recombinant proteins, were carried out as described previously for His-tagged SIB1 FKBP22 [8]. The production levels of the recombinant proteins in the E. coli cells, and their purities, were analyzed by SDS–PAGE [42] using a 15% polyacryl- amide gel, followed by staining with Coomassie Brilliant Blue. Protein concentrations were determined from the UV absorption on the basis that the absorbance at 280 nm of a 0.1% (1 mgÆml )1 ) solution is 0.68 for SIB1 FKBP22 and 0.61 for NNC-FKBP22. These values were calculated by using e = 1576 m )1 Æcm )1 for Tyr and 5225 m )1 Æcm )1 for Trp, at 280 nm [43]. Molecular mass Sedimentation equilibrium analytical ultracentrifugation was performed at 4 °C for 20 h with a Beckman Optima XL-A Analytical Ultracentrifuge (Beckman, Tokyo, Japan) using an An-60 Ti rotor at 140 000 g. Before measurement, the protein solution was dialyzed overnight against 20 mm sodium phosphate (pH 8.0) at 4 °C. The concentration of the protein for initial loading was 2 mgÆmL )1 . Distribution of the protein within the cell was analyzed by monitoring the absorbance at 280 nm. Analysis of the sedimentation equilibrium was performed using the program xlavel, ver- sion 2 (Beckman). Gel-filtration column chromatography was carried out using HPLC with a TSK-GEL G2000SWXL column (Tos- oh Co., Tokyo, Japan) equilibrated with 50 mm Tris–HCl (pH 8.0) containing 50 mm NaCl. Elution was performed at a flow rate of 0.5 mLÆmin )1 . Bovine tyroglobulin (670 kDa), bovine c-globulin (158 kDa), chicken ovalbumin (44 kDa), horse myoglobin (17 kDa) and vitamin B12 (1.3 kDa) were used as standard proteins. Enzymatic activity The PPIase activity was determined using a protease-cou- pling assay [44] and an RNase T1 refolding assay [45]. For the protease-coupling assay, chymotrypsin was used as a protease and Suc-ALPF-pNA (Wako Pure Chemical Indus- tries, Ltd., Osaka, Japan) was used as a substrate. The reaction mixture (2 mL) contained 35 mm HEPES (pH 7.8), 25 lm Suc-ALPF-pNA and an appropriate amount of the enzyme. The reaction mixture was incubated at the reaction temperature (4, 10, 15 or 20 °C) for 3 min before the addition of chymotrypsin. The reaction was initiated by the addition of 30 lL of 0.76 mm chymotrypsin. The isom- erization of the Leu–Pro bond, catalyzed by PPIase, was measured by monitoring the change in the concentration of p-nitroanilide (pNA), because pNA is released from the substrate only when this peptide bond is in a trans confor- mation. The increase in the rate of isomerization is implicit in the increased rate of pNA release, because catalysis of isomerization produces a trans substrate with increased fre- quency. The concentration of pNA was determined from the absorption at 390 nm, with the molar absorption coeffi- cient value of 8900 M )1 cm )1 , using a Hitachi U-2010 UV ⁄ VIS spectrophotometer (Hitachi High-Technologies Co., Tokyo, Japan). The catalytic efficiency (k cat ⁄ K m ) was calculated from the relationship k cat ⁄ K m =(k p –k n ) ⁄ E, where E represents the concentration of the enzyme, and k p and k n represent the first-order rate constants for the Engineering of monomeric FKBP22 C. Budiman et al. 4098 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS release of pNA from the substrate in the presence and absence of the enzyme, respectively [46]. For the RNase T 1 refolding assay, RNase T 1 (16 lm) (Funakoshi Co., Ltd., Tokyo, Japan) was first unfolded by incubation in 20 mm sodium phosphate (pH 8.0), contain- ing 0.1 mm EDTA and 6.2 m guanidine hydrochloride, at 10 °C overnight. Refolding was then initiated by diluting this solution 80-fold with 20 mm sodium phosphate (pH 8.0) containing 100 mm NaCl in the presence or absence of the enzyme. The final concentrations of RNase T 1 and the enzyme were 0.2 lm and 10 nm, respectively. The refolding reaction was monitored by measuring the increase in tryp- tophan fluorescence using an F-2000 spectrofluorometer (Hitachi High-Technologies Co.). The excitation and emis- sion wavelengths were 295 and 323 nm, respectively, and the band width was 10 nm. The refolding curves were ana- lyzed with double exponential fit [36]. The k cat ⁄ K m values were calculated from the relationship mentioned above, where k p and k n represent first-order rate constants for the faster refolding phase of RNase T 1 in the presence and absence of the enzyme, respectively. CD The CD spectra were measured at 10 °C on a J-725 auto- matic spectropolarimeter (JASCO Co., Tokyo, Japan). The protein was dissolved in 20 mm sodium phosphate (pH 8.0) and incubated at 10 °C for 30 min before the measurement was made. For measurement of the far-UV CD spectra (200–260 nm), the protein concentration was approximately 0.2 mgÆmL )1 and a cell with an optical path length of 2 mm was used. For measurement of the near-UV CD spectra (250–320 nm), the protein concentration was approximately 0.7 mgÆmL )1 and a cell with an optical path length of 10 mm was used. The mean residue ellipticity, h, which has units of degÆcm 2 Ædmol )1 , was calculated by using an average amino acid relative molecular mass of 110. DSC The DSC measurement was performed on a high-sensitivity VP-DSC controlled by the VPVIEWERTM software pack- age (Microcal Inc., Northampton, MA, USA) at a scan rate of 1 °CÆmin )1 . The protein was dissolved in 20 mm sodium phosphate (pH 8.0) at approximately 1.0 mgÆmL )1 . Before performing the measurement, the protein solution was fil- tered through 0.22-lm pore-size membranes and then degassed in a vacuum. The reversibility of thermal denatur- ation was verified by reheating the sample. Surface plasmon resonance The interaction between SIB1 FKBP22 or NNC-FKBP22 with reduced a-lactalbumins was monitored by surface plas- mon resonance using the Biacore X instrument (Biacore, Uppsala, Sweden). Immobilization of the His-tagged pro- tein to a Ni 2+ -chelated nitrilotriacetic acid sensor chip (Biacore) was carried out as described previously [15]. Reduced a-lactalbumin, which was dissolved at a concen- tration of 100 lm in 20 mm sodium phosphate (pH 8.0) containing 2 mm dithiothreitol, 100 mm NaCl and 1 mm EDTA, was then injected at 10 °C, with a flow rate of 10 lLÆmin )1 , onto the surface of the sensor chip on which the His-tagged protein was immobilized. Binding surfaces were regenerated by washing with 0.5 m EDTA. To determine the dissociation constant, K D , the concen- tration of reduced a-lactalbumin injected onto the sensor chip was varied from 0.5 to 100 lm. From the plot of the equilibrium-binding responses as a function of the concen- trations of reduced a-lactalbumin, the K D values were determined using steady-state affinity program of BIAevalu- ation Software (Biacore). Acknowledgements We thank Dr T. Tadokoro for helpful discussions. This work was supported, in part, by a grant (21380065) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and by an Industrial Technology Research Grant Program from the New Energy and Industrial Technology Develop- ment Organization (NEDO) of Japan. References 1 Lang K, Schmid FX & Fischer G (1987) Catalysis of protein folding by prolyl isomerase. Nature 329, 268– 270. 2 Grathwohl C & Wuthrich K (1981) NMR studies of the rates of proline cis-trans isomerization in oligopeptides. Biopolymers 20, 2623–2633. 3 Cheng HN & Bovey FA (1997) Cis-trans equilibrium and kinetic studies of acetyl-L-proline and glycil-L-pro- line. Biopolymers 16, 1465–1475. 4 Jacob RP & Schmid FX (2008) Energetic coupling between native-state prolyl isomerization and conforma- tional protein folding. J Mol Biol 377, 1560–1575. 5 Schmid FX, Mayr LM, Mucke M & Schonbrunner ER (1993) Prolyl isomerase : role in protein folding. Adv Protein Chem 44, 25–66. 6 Lu KP, Finn G, Lee TH & Nicholson LK (2007) Prolyl cis-trans isomerization as molecular timer. Nat Chem Biol 3, 619–629. 7 Rahfeld JU, Rucknagel KP, Stoller G, Horne SM, Schierhorn A, Young KD & Fischer G (1996) Isolation and amino acid sequence of a new 22-kDa FKBP-like peptidyl-prolyl cis ⁄ trans-isomerase of Escherichia coli. C. Budiman et al. Engineering of monomeric FKBP22 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS 4099 Similarity to Mip-like proteins of pathogenic bacteria. J Biol Chem 271, 22130–22138. 8 Suzuki Y, Haruki M, Takano K, Morikawa M & Kanaya S (2004) Possible involvement of an FKBP family member protein from a psychrotrophic bacte- rium Shewanella sp. SIB1 in cold-adaptation. Eur J Bio- chem 271, 1372–1381. 9 Tremmel D & Tropschung M (2007) Neurospora crassa FKBP22 is a novel ER chaperone and functionally cooperates with BiP. J Mol Biol 369, 55–68. 10 Horne SM & Young KD (1995) Escherichia coli and other species of the Enterobacteriaceae encode a protein similar to the family of Mip-like FK506-binding pro- teins. Arch Microbiol 165, 357–365. 11 Riboldi-Tunnicliffe A, Konig B, Jessen S, Weiss MS, Rahfeld J, Hacker J, Fischer G & Hilgenfeld R (2001) Crystal structure of Mip, a prolylisomerase from Legio- nella pneumophila. Nat Struct Biol 8, 779–783. 12 Saul FA, Arie JP, Vulliez-le Normand B, Kahn R, Betto NJM & Bentley GA (2004) Structure and func- tional studies of FkpA from Escherichia coli,acis ⁄ trans peptidyl-prolyl isomerase with chaperone activity. J Mol Biol 335, 595–608. 13 Engleberg NC, Carter C, Weber DR, Cianciotto NP & Eisenstein BI (1989) DNA sequence of mip,aLegionella pneumophila gene associated with macrophage infectiv- ity. Infect Immun 57, 1263–1270. 14 Suzuki Y, Takano K & Kanaya S (2005) Stabilities and activities of the N- and C-domains of FKBP22 from a psychrotrophic bacterium overproduced in E. coli. FEBS J 272, 632–642. 15 Suzuki Y, Win OY, Koga Y, Takano K & Kanaya S (2005) Binding analysis of a psychrotrophic FKBP22 to a folding intermediate of protein using surface plasmon resonance. FEBS Lett 579, 5781–5784. 16 Schmidt B, Rahfeld J, Schieron A, Ludwig B, Hacker J & Fischer G (1994) A homodimer represents an active species of the peptidyl-prolyl cis ⁄ trans isomerase FKBP25mem from Legionella pneumophilla. FEMS Microbiol Lett 118, 23–30. 17 Wu CS, Ikeda K & Yang JT (1981) Ordered confor- mation of polypeptides and proteins in acidic dodecyl sulfate solution. Biochemistry 20, 566–570. 18 Kiefhaber T, Quaas R, Hahn U & Schmid FX (1990) Folding of ribonuclease T 1 . 1. Existence of multiple unfolded states created by proline isomerization. Biochemistry 29, 3051–3061. 19 Kiefhaber T, Quaas R, Hahn U & Schmid FX (1990) Folding of ribonuclease T 1 . 2. Kinetic models for the folding and unfolding reactions. Biochemistry 29, 3061–3070. 20 Schindler T, Mayr LM, Landt O, Hahn U & Schmid FX (1996) The role of a trans-proline in the folding mechanism of ribonuclease T 1 . Eur J Biochem 241, 516–524. 21 Acharya KR, Stuart DI, Walker NP, Lewis M & Phil- lips DC (1989) Refined structure of baboon a-lactalbu- min at 1.7 A ˚ resolution. Comparison with C-type lysozyme. J Mol Biol 1, 99–127. 22 Acharya KR, Ren JS, Stuart DI, Phillips D & Fenna RE (1991) Crystal structure of human a-lactalbumin at 1.7 A ˚ resolution. J Mol Biol 2, 571–581. 23 Kuwajima K (1989) The molten globule state as a clue for understanding the folding and cooperativity of glob- ular-protein structure. Proteins 6, 87–103. 24 Hayer-Hartl MK, Ewbank JJ, Creighton TE & Hartl FU (1994) Conformational specificity of the chaperonin GroEL for the compact folding intermediates of a-lact- albumin. EMBO J 13, 3192–3202. 25 Okazaki A, Ikura T, Nikaido K & Kuwajima K (1994) The chaperonin GroEL does not recognize apo-a-lactal- bumin in the molten globule state. Nat Struct Biol 1, 439–446. 26 Scholz C, Stoller G, Zarnt T, Fischer G & Schmid FX (1997) Cooperation of enzymatic and chaperone func- tions of trigger factor in the catalysis of protein folding. EMBO J 16, 54–58. 27 Ramm K & Pluckthun A (2001) High enzymatic activ- ity and chaperone function are mechanistically related features to the dimeric E. coli peptidyl-prolyl-isomerase FkpA. J Mol Biol 310, 485–498. 28 Hu K, Galius V & Pervushin K (2006) Structural plas- ticity of peptidyl)prolyl isomerase sFkpA is a key to its chaperone function as revealed by solution NMR. Bio- chemistry 45, 11983–11991. 29 McCarthy AA, Haebel P, Torronen A, Rybin V, Baker EN & Metcalf P (2000) Crystal structure of the protein disulfide bond isomerase, DsbC, from Escherichia coli. Nature 7, 196–199. 30 Hottenrott S, Schumann T, Pluckhthun A, Fischer G & Rahfeld JU (1997) The Escherichia coli SlyD is a metal ion-regulated peptidyl-prolyl cis ⁄ trans-isomerase. J Biol Chem 272, 15697–15701. 31 Scholz C, Eckert B, Hagn F, Schaarschmidt P, Balbach J & Schmid FX (2006) SlyD proteins from different spe- cies exhibit high prolyl isomerase and chaperone activi- ties. Biochemistry 45, 20–33. 32 Han KY, Song JA, Ahn KY, Park JS, Seo HS & Lee J (2007) Solubilization of aggregation-prone heterologous protein by covalent fusion of stress-responsive Escheri- chia coli protein, SlyD. Protein Eng Des & Sel 20, 543– 549. 33 Furutani M, Ideno A, Iida T & Maruyama T (2000) FK506 binding protein from a thermophilic archeon, Methanococcus thermolithotropicus, has chaperone-like activity in vitro. Biochemistry 39, 453–462. 34 Suzuki R, Nagata K, Yumoto F, Kawakami M, Nemoto N, Furutani M, Adachi K, Maruyama T & Tanokura M (2003) Three-dimensional solution structure of an archaeal FKBP with a dual function of Engineering of monomeric FKBP22 C. Budiman et al. 4100 FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS [...]... Cleavage of structural proteins during the assembly of the head of bacteriophage Nature 227 , 680–685 Engineering of monomeric FKBP22 43 Goodwin TW & Morton RA (1946) The spectrophotometric determination of tyrosine and tryptophan in proteins Biochem J 40, 628–632 44 Fischer G, Wittmann-Liebold B, Lang K, Kiefhaber T & Schmid FX (1989) Cyclophilin and peptidyl- prolyl cis-trans isomerase are probably... following supplementary material is available: Fig S1 SDS–PAGE of purified NNC-FKBP22 and SIB1 FKBP22 proteins This supplementary material can be found in the online version of this article Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited... Budiman et al 35 36 37 38 39 40 41 42 peptidyl prolyl cis-trans isomerase and chaperone-like activities J Mol Biol 328, 1149–1160 Knappe TA, Eckert B, Scaarscmidt P, Scholz C & Schmid FX (2007) Insertion of a chaperone domain converts FKBP12 into powerful catalyst of protein folding J Mol Biol 368, 1458–1468 Ramm K & Pluckthun A (2000) The periplasmic Escherichia coli peptidylprolyl cis,trans -isomerase. .. cis,trans -isomerase FkpA II Isomerase- independent chaperone activity in vitro J Biol Chem 275, 17106–17113 Grandbois M, Beyer M, Rief M, Clausen-Schaumann H & Gaub HE (1999) How strong is a covalent bond? Science 288, 1727–1730 Zaccai G (2000) How soft is protein? A protein dynamics force constant measured by neutron scattering Science 288, 1604–1607 Thrope MF, Lei M, Rader AJ, Jacobs DJ & Kuhn LA (2001) Protein. .. probably identical proteins Nature 337, 476–478 45 Schonbrunner ER, Mayer S, Tropschug M, Fischer G, Takahashi N & Schmid FX (1991) Catalysis of protein folding by cyclophilins from different species J Biol Chem 266, 3630–3635 46 Harrison RK & Stein RL (1990) Mechanistic studies of peptidyl prolyl cis-trans isomerase: evidence for catalysis by distortion Biochemistry 29, 1684–1689 Supporting information The... flexibility and dynamic using constraint theory J Mol Graph Model 19, 60–69 Jacob D & Dallakyan S (2005) Elucidating protein thermodynamics from the three-dimensional structure of the native state using network rigidity Biophys J 88, 903–915 Horton RM, Cai ZL, Ho SN & Pease LR (1990) Gene splicing by overlap extension: tailor-made genes using the polymerase chain reaction BioTechniques 8, 528–535 Laemmli... materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset Technical support issues arising from supporting information (other than missing files) should be addressed to the authors FEBS Journal 276 (2009) 4091–4101 ª 2009 The Authors Journal compilation ª 2009 FEBS 4101 . 5¢-AGAGAGAATT CATATGTCAGATT TGTTCAG-3¢ for primer 1; 5¢-TTCCATACCACCACCT GCAACTTGAAGCTC-3¢ for primer 2; 5¢-GTTGCAGGT GGTGGTATGGAACAGCATGCT-3¢ for primer 3; and 5¢-GGCCACT GGATCCAACTACAGCAATTCTCA-3¢ for primer. University, Yamadaoka, Suita, Osaka, Japan 2 CRESTO, JST, Yamadaoka, Suita, Osaka, Japan Keywords FKBP22; homodimer; peptidyl- prolyl cis-trans isomerase (PPIase); protein engineering; substrate binding Correspondence S FKBP22 may also be required to hold a protein substrate with a similar mechanism. The plasticity of a V-shaped structure may lead to a conformational flexibility to adopt various types of protein substrates. The