- Email: [email protected]
The role of Cys271 in conformational changes of arginine kinase
International Journal of Biological Macromolecules 49 (2011) 98–102
Contents lists available at ScienceDirect
International Journal of Biological Macromolecules journal homepage: www.elsevier.com/locate/ijbiomac
The role of Cys271 in conformational changes of arginine kinase Na Liu a,b , Jin-Song Wang a,b,∗ , Wei-Dong Wang a,b , Ji-Cheng Pan a,b,∗ a b
College of Life Science, Hubei Normal University, Huangshi, Hubei 435002, PR China Hubei Key Laboratory of Pollutant Analysis & Reuse Technology, PR China
a r t i c l e
i n f o
Article history: Received 3 March 2011 Received in revised form 2 April 2011 Accepted 4 April 2011 Available online 12 April 2011 Keywords: Arginine kinase Cys271 Conformational changes Site-directed mutagenesis
a b s t r a c t Arginine kinase (AK), a crucial enzyme in energy metabolism, buffers cellular ATP levels by catalyzing the reversible phosphoryl transfer between ATP and arginine. To better understand the role of Cys271 in conformational changes of AK from greasyback shrimp (Metapenaeus ensis), we replaced the residue with serine and alanine. A detailed comparison of the catalytic activity and conformation was made between wild-type AK and the mutants by means of activity analysis, ultraviolet (UV) difference, fluorescence spectrum and size exclusion chromatography (SEC). The results indicated that the catalytic activity of the two mutants was gone. The substrates, arginine–ADP–Mg2+ could induce conformational changes, and additional NO3 − could induce further changes in both the native enzyme and the variants. We speculated that Cys271 might be located in the hinge region between the two domains of AK and cause enzyme conformational changes upon addition of substrate. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Arginine kinase (AK, EC 2.7.3.3), a phosphagen kinase participating in cellular energy metabolism in invertebrates, buffers cellular ATP levels, which catalyzes the reversible phosphoryl transfer between ATP and arginine, leading to the production of phosphoarginine and Mg2+ ADP [1,2]. The latter is considered to be an energy reservoir, which is able to supply ATP, the primary energy source in bioenergetics, on demand [3]. The crystal structure of the transition state analog complex (TSAC) of monomeric AK from the horseshoe crab (Limulus) is determined at 1.86 A˚ [4] and refined at 1.2 A˚ [5], and it suggests that AKs can be divided into two domains, a smaller ␣-helical N-terminal domain and a larger C-terminal domain (residues 112–357). The catalytic center, where reversible transfer of phosphate occurs, is located in the C-terminal domain, but the N-terminal domain also contributes to arginine binding. The two domains undergo a hinge rotation of 13◦ upon formation of the TSAC. The substrate-free AK is considered to be in “open” form, changing to a “closed” form upon substrate binding [6]. The substrate-free structure of creatine kinase (CK) indicates that the induced fit observed in arginine
Abbreviations: WT-AK, wild-type arginine kinase; CK, creatine kinase; Cys271, cysteine of residue-271; SEC, size exclusion chromatography; TSAC, transition state analog complex; IPTG, isopropyl--d-thiogalactoside; ANS, 1-anilinonaphtalene-8sulfonate. ∗ Corresponding authors at: College of Life Science, Hubei Normal University, Huangshi, Hubei 435002, PR China. Tel.: +86 0714 6513163; fax: +86 0714 6513163. E-mail addresses: [email protected] (J.-S. Wang), [email protected] (J.-C. Pan). 0141-8130/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.ijbiomac.2011.04.002
kinase upon association of both substrates and the formation of TSAC (comprising Mg2+ ADP, nitrate and arginine) [7]. In the given examples, the conformational changes appear to be necessary for alignment of the two substrates for catalysis. The latest structural studies provide a detailed mechanism indicating that the binding of ADP–Mg2+ alone may trigger conformational changes in humanbrain-type-creatine-kinase (HbCK) [8]. During the reaction, critical elements of the active site are reconfigured and several loops and residues of key importance become ordered [7,9]. The roles of various amino acid residues in either substratebinding or catalytic activity of arginine kinase have been widely investigated. The substrate recognition region (GS region) of AK, which is composed of Ser63, Gly64, Val65 and Tyr68 is highly conserved [10]. Supporting evidence from enzyme kinetic and thermal stability experiments on wild-type and mutant AKs show that the activity of the enzyme is significantly reduced and that the substrate synergism is distinctly altered when the conserved sites are replaced [11–13]. The role of the amino acid residue P272 of AK is investigated via site-directed mutagenesis, which suggests that some residues near the active site may play a relatively important role in sustaining AK’s activity and structural stability [14]. Phosphagen kinases contain a single reactive cysteine residue (271 in AK, 278 in chicken mitochondrial creatine kinase) in each subunit which is highly conserved in all known phosphagen kinases sequences [15]. Comparison of the structure of the substrate-free and the transition-state of AK shows that, although Cys271 is located in the C-terminal domain, it moves towards the residues of the N-terminal domain during catalysis [6]. At present, whether the cysteine is essential for enzyme catalysis is still a controversial issue. The “essential” cysteine takes its name from early
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
chemical modifications [16,17] and mutagenesis experiments of the reactive cysteine that resulted in completely inactive enzymes [18,19]. However, chemical modification may lead to stability and structural changes in the enzymes [20,21]. The effect of a gross modification can be structurally rationalized by steric conflict with the substrate guanidinium [4]. More subtle effects of mutations [22] and chemical modifications of the cysteine [23,24] lead to several hypotheses that the cysteine may not be essential for activity, but may be involved in synergism upon the binding of both substrates or in mediating the substrate-induced hinge movement. It has been shown that the activity of uncharged reactive cysteine mutations could be partially rescued with chloride, leading to the proposal that the cysteine functions as a thiolate ion (AA–CH2 –S− ) [22]. The recent crystal structure for the TSAC of the C271A mutant at 2.3 A˚ resolution shows minor structural changes, and suggests that the cysteine enhances the catalytic rate through electrostatic stabilization of the transition state [25]. Experiments on wild-type and cysteine-substituted human muscle creatin kinase (HMCK) show that an ionized active site cysteine is required across the entire family of guanidino kinases [26]. This report focuses on the reactive Cys271, which is close to the catalytic center of AK. It examines the role of Cys271 in catalysis and the conformational change upon substrate binding by means of activity analysis, ultraviolet (UV) difference spectra, fluorescence spectra and size-exclusion chromatography (SEC). 2. Materials and methods 2.1. Cloning, and site-directed mutagenesis of the greasyback shrimp AK The pET-28a plasmid with the insert of the greasyback shrimp AK gene (WT-AK) [27] was used as a template for mutagenesis. The replacement of the amino acid Cys271 by alanine (Ala) and serine (Ser) were achieved by PCR with the KOD-Plus mutagenesis Kit (Toyobo, Tokyo, Japan), which are denoted C271A and C271S, respectively. The primers used for PCR amplification (mutated codons are underlined) were: C271A: 5 -GCTCCCACCAACCTTGGGCACCACTGTGC-3 3 -GTGCTGGCGGACCCGAAGGAGTGAAAG-5 ; C271S: 5 -AGTCCCACCAACCTTGGCACCACTGTGC-3 3 -GTGCTG-GCGGACCCGAAGGAGTGAAAG-5 . For both mutants, the fidelity of the PCR amplification was confirmed by sequencing.
99
ous pH-spectrophotometric assay [28]. The enzyme concentration was estimated from the absorbance at 280 nm (absorbance 0.67 at 280 nm in 1 cm cuvette corresponds to 1 mg protein/mL) [29]. The enzymatic activity and concentration of the proteins were measured with an Ultrospec 4300 pro UV/visible spectrophotometer. 2.4. Spectroscopic measurements UV difference spectra of the enzyme complexes with substrates or transition-state analogues were measured relative to the substrate-free enzymes as described by Focant and Watts [30]. All absorption spectra were measured with an Ultrospec 4300 pro UV/visible spectrophotometer. Intrinsic fluorescence emission spectra were recorded on an F-4500 fluorescence spectrophotometer (Hitachi) with 1 cm pathlength cuvettes, excitation at 295 nm and emission wavelength in the range of 300–400 nm. ANS was used as an extrinsic probe [31], and a 20-fold molar excess of ANS was added to the samples for 30 min in the dark. An excitation wavelength of 380 nm was used to determine the ANS fluorescence intensity of the substratefree/bound enzymes in the range of 400–600 nm. All spectroscopy measurements were recorded after the enzymes were added to the substrate (ADP–Mg2+ ) for 0.5 h in the similar reaction condition described above, and NO3 ¯ was added to the reaction mixture above for another 0.5 h at 25 ◦ C. Appropriate buffers were used for baseline determinations. Stored spectra were corrected for dilution and baseline effects. 2.5. Size-exclusion chromatography Gel filtration was performed with a Superdex 200HR 10/30 column (fractionation range 1000–600,000) on a Pharmacia AKTA purifier apparatus at 25 ◦ C. All prepared solutions were passed through a filter and degassed. A 500 L sample of enzyme was added to ADP–Mg2+ for 0.5 h and ADP–Mg2+ –NO3 ¯ for 1 h and was then injected into a column equilibrated with buffer A (50 mM Tris/HCl, 500 mM NaCl, 0.57 mM arginine, 0.1 mM ADP, and 0.66 mM magnesium acetate, pH 8.0) and buffer B (0.5 mM NaNO3 added to buffer A), respectively. The flow rate was maintained at 0.5 mL min−1 , and the results were monitored at 280 nm. The distribution coefficient Kav [Kav = (Ve − V0 )/(Vt − V0 )], with Ve being the elution volume, V0 the void volume and Vt the column total volume was used to monitor the Stokes radius changes of the enzyme molecules during the enzyme association with the substrates. 3. Results
2.2. Expression and purification 3.1. Purification of the site-directed mutants To facilitate the expression of the mutagenic DNA, the recombinant plasmids (mutants with 6×His-tag) were transformed into the Escherichia coli (E. coli) strain Rosetta. The cells were grown at 37 ◦ C then induced with 0.8 mM IPTG at 22 ◦ C. The Wt-AK was expressed as described previously [27], but purified as mutant AKs using nickel affinity chromatography. The Ni2+ column was washed first with a washing buffer (50 mM Tris/HCl, 0.5 M NaCl, pH 8.0) and then washed gradually with the same buffer containing an imidazole gradient. The purified enzymes were confirmed to be homogeneous by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). The enzymes were placed on ice until use, and the enzymatic activity was determined within 12 h.
To elucidate the role of Cys271 in catalysis and the conformational change upon substrate binding, the Cys was replaced by Ala and Ser residues, respectively, resulting in the mutants C271A and C271S. The recombinant proteins were obtained from the E. coil strain Rosetta transformed with plasmids containing the genes for the wild-type and mutated AK. The purified enzymes were proven to be homogeneous by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 1). However, mutagenesis of the Cys271 resulted in completely inactive enzymes (data not shown). 3.2. Conformational changes of WT-AK and the mutants upon substrates binding
2.3. Enzyme concentration and activity The activity of wild-type and mutant AK in the forward reaction (arginine phosphate synthesis) was assayed using a direct continu-
The physico-chemical properties of the wild-type enzyme and mutants in the presence of either the substrates or the transitionstate analog were explored by intrinsic fluorescence (Fig. 2), ANS
100
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
Fig. 1. Analysis of the purified Wt-AK and mutated AK on 12% SDS-PAGE. M, the molecular mass standard. The variants are as indicated in the figure.
fluorescence (Fig. 3), ultraviolet (UV) difference (Fig. 4), and sizeexclusion chromatography (SEC) (Fig. 5). The intrinsic fluorescence (Fig. 2), which was excited at 295 nm to minimize the contributions of residues other than Trp, indicated that the maximum emission wavelengths (Em ) were similar, but the fluorescence intensity decreased sharply upon substrate binding
(shown as curve 2) and then slightly decreased after NO3 ¯ addition as the TSAC formed (shown as curve 3). This suggested that the substrates arginine–ADP–Mg2+ and additional NO3 ¯ might induce microenvironmental changes of the Trp residues upon association with the enzymes and affect the conformational changes of the WTAK and mutant AK. The ANS fluorescence intensity of WT-AK was higher than substrate-bound WT-AK or the TSAC at the same concentration (Fig. 3A), which reflected that the “closed” form of AK had less hydrophobic exposure than did the “open” form to allow the binding of the ANS molecules. Similar to the WT-AK’s, a significant decrease of the ANS fluorescence intensity was observed for the two mutants (Fig. 3B and C). Fig. 4A shows the overall change (240–320 nm) to the UV spectrum of wild-type AK when excess substrates and NO3 ¯ were added to the free protein. Upon substrate binding and TSAC formation, the magnitude of the induced difference spectrum increased between 240 and 270 nm with the peak center at approx. 260 nm. This is characteristic of the phenylalanine spectrum. There was a decreased in tyrosine absorbance as shown by negative peaks centering around 286 nm. A decrease in tyrosine absorbance in proteins is typically associated with the exposure of tyrosine to a more polar
Fig. 2. Intrinsic fluorescence spectra for wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. Final concentrations were 14.2 M enzyme in 50 mM Tris/HCl buffer, pH 8.0, 0.57 mM arginine, 0.1 mM ADP, and 0.66 mM magnesium acetate. NaNO3 (0.5 mM) was also added in addition to the substrates for the measurement of the intrinsic fluorescence of the complex with the transition-state analog. In subparts (A)–(C) the curves represent different state of enzyme as follows: (1) represents spectra of apo-enzyme, (2) enzyme–substrate complex and (3) the complex with the transition-state analog.
Fig. 3. ANS fluorescence spectra for wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. The conditions were the same as for Fig. 2. In subparts (A)–(C) the curves represent different states of enzyme as follows: (1) represents the emission fluorescence spectra of apo-enzyme, (2) enzyme–substrate complex and (3) the complex with the transition-state analog.
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
101
Fig. 4. Spectra for wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. The conditions were the same as for Fig. 2. In subparts (A)–(C) the curves represent different states of enzyme as follows: (1) represents the difference spectrum for apo-enzyme as contrast, (2) enzyme–substrate complex and (3) the complex with the transition-state analog.
Fig. 5. Size exclusion chromatography of wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. The parameters of the column (HR10/30, Pharmacia) were V0 = 7.77 mL and Vt = 23.562 mL. The other conditions were the same as for Fig. 2. In subparts (A)–(C) the curves represent different states of enzyme as follows: (1) represents apo-enzyme, (2) enzyme–substrates complex and (3) the complex with the transition-state analog.
environment. A similar comparison of arginine–ADP–Mg2+ and additional NO3 ¯ induced difference spectrum for the two mutants (C271A and C271S) showed that the conformational changes of C271A (Fig. 4B) were nearly the same as WT-AK. C271S had a different UV difference spectrum (Fig. 4C), which may be due to the structural changes after cysteine was replaced by serine. The retention times of size exclusion chromatography had been obtained as shown in Table 1, and mean SD values had also been calculated. The elution volume of substrate-free AK was different from the substrate-bound form and TSAC (Fig. 5A), and the two mutants had the same phenomena (Fig. 5B and C). Although the change was subtle, it also suggested that the conformation of the wildtype AK and the two mutants was changed upon substrate binding and TSAC formation, which is consistent with the above conclusion from the measurement of fluorescence spectra (Figs. 2 and 3) and UV difference spectra (Fig. 4). 4. Discussion In a previous study, the gene encoding arginine kinase from greasyback shrimp (Metapenaeus ensis) was suc-
cessfully cloned, sequenced, and over-expressed in E. coli [27]. Modified arginine kinase reactivated by dithiothreitol suggests that the reactive cysteine plays an important role in the Table 1 Retention volumes of wild-type and mutation-type AKs in SEC. Proteins WT-AK
C271A
C271S
Mean ± SD
Retention volume (mL) a
17.166 17.277b 17.452c 17.256a 17.438b 17.589c 17.237a 17.421b 17.335c
a
17.158 17.265b 17.435c 17.235a 17.437b 17.574c 17.254a 17.402b 17.301c
a
17.175 17.287b 17.470c 17.215a 17.461b 17.593c 17.246a 17.431b 17.345c
17.163 17.276 17.452 17.235 17.445 17.585 17.246 17.418 17.327
± ± ± ± ± ± ± ± ±
0.008505 0.011015 0.017502 0.020502 0.013577 0.010017 0.008505 0.014731 0.023065
Samples (WT-AK, C271A, C271S) were added to calibrated column of Superdex 200 (HR10/30, Pharmacia), and its retention volumes were monitored and recorded, respectively. Every sample was detected three times, and SD value was calculated. a apo-enzyme. b Enzyme–substrates complex. c The complex with the transition-state analog.
102
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
conformational changes caused by the transition-state analog [24]. However, chemical modification may lead to stability and structural changes of the enzymes [20,21]. In this research, the two mutants, C271S and C271A, were successfully cloned, and their expression level was lower than wild-type AK (WT-AK) (Fig. 1). Kinetic analysis showed that C271S and C271A had no activity (data not shown). The results of spectroscopic measurements (Figs. 2–5) illustrated that the substrates, arginine–ADP–Mg2+ , could induce conformational changes and that additional NO3 ¯ could induce further changes in both the wild-type enzyme and the variants. This suggested that Cys271 was not located in the ATP/ADP binding site but very close to it, such that the mutants could bind to the substrate even though there was no enzyme activity. This was in accord with the reported crystal structure of AK from horseshoe crab [4], which shows that the reactive Cys is near the ␥ phosphate of ATP. Fluorescence spectroscopy is an effective method of studying protein conformation in aqueous solution. Proteins autofluorescence (Fig. 2), and the fluorescence of the fluorogenic probe associated with the proteins in specific areas (Fig. 3) was detected in this paper. As shown in the figures, the microenvironment of the Trp residues was changed upon substrate binding (Fig. 2) and the “closed” form of AK had less hydrophobic exposure than did the “open” form to allow the binding of the ANS molecules (Fig. 3). This illustrated that the substrates arginine–ADP–Mg2+ and additional NO3 ¯ might affect the conformational changes of WT-AK and mutations after association with enzymes. It is well known that, in the presence of the substrates (arginine–ADP–Mg2+ ) or the transition-state analog (MgADP–arginine–NO3 − ), the enzyme undergoes conformational changes as indicated by absorbance changes in the UV region [30]. The substrates, arginine–ADP–Mg2+ , could also induce a change of the absorbance of the mutated arginine kinase in the UV region (Fig. 4B and C), showing that MgADP could bind to the mutated enzyme. The transition-state analog was certified by NMR to mimic the binding of substrate in the transition state during catalysis, involving probable conformational changes of the enzyme molecule [32]. The present work showed that both arginine–ADP–Mg2+ and additional NO3 ¯ could induce fluorescence intensity and UV absorbance changed in the two mutants similar to those observed in the wild-type arginine kinase. Interestedly, we had found that compared to WT, the conformational changes of the mutants could be induced by substrate or NO3 ¯ whereas these changes were not fit the enzyme catalysis, as C271A and C271S had no detectable activity. Analyzing the phenomenon, we speculated that there might be two reasons: (1) the conformational changes trend of the mutants might not be same as WT; (2) although the trend was similar to WT, the two domains underwent a hinge rotation that might not be 13◦ , which happens upon formation of the TSAC of WT-AK [6]. The results differed somewhat from our previous study, in which NO3 ¯ did not induce further absorbance changes of the modified enzyme [24]. There
may be steric effect due to the structural changes of chemical modification of AK. In conclusion, Cys271 may play an important role in the conformational changes caused by the substrates and the transition-state analog, and it may be located in the hinge region of the two domains of AK, guiding conformational changes of the enzyme upon addition of substrates. Acknowledgements The present investigation was supported by grant 30570412, 2005ABA192, 2008CDA066 from National Natural Science Foundation of China and Hubei Province’s Natural Science Foundation of China. References [1] E.A. Newsholme, I. Beis, A.R. Leech, V.A. Zammit, Biochem. J. 172 (1978) 533–537. [2] W.R. Ellington, J. Exp. Biol. 143 (1989) 177–194. [3] W.R. Ellington, Ann. Rev. Physiol. 63 (2001) 289–325. [4] G. Zhou, T. Somasundaram, E. Blanc, G. Parthasarathy, W.R. Ellington, M.S. Chapman, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 8449–8454. [5] M.S. Yousef, F. Fabiola, J.L. Gattis, T. Somasundaram, M.S. Chapman, Acta Crystallogr. D Biol. Crystallogr. 58 (2002) 2009–2017. [6] M.S. Yousef, S.A. Clark, P.K. Pruett, T. Somasundaram, W.R. Ellington, M.S. Chapman, Protein Sci. 12 (2003) 103–111. [7] M. Forstner, M. Kriechbaum, P. Laggner, T. Wallimann, Biophys. J. 75 (1998) 1016–1023. [8] S.M. Bong, J.H. Moon, K.H. Nam, K.S. Lee, Y.M. Chi, K.Y. Hwang, FEBS Lett. 582 (2008) 3959–3965. [9] G. Zhou, W.R. Ellington, M.S. Chapman, Biophys. J. 78 (2000) 1541–1550. [10] T. Suzuki, H. Fukuta, H. Nagato, M. Umekawa, J. Biol. Chem. 275 (2000) 23884–23890. [11] K. Uda, T. Suzuki, Protein J 23 (2004) 53–64. [12] N. Fujimoto, K. Tanaka, T. Suzuki, FEBS Lett. 579 (2005) 1688–1692. [13] Q.Y. Wu, F. Li, X.Y. Wang, Int. J. Biol. Macromol. 45 (2009) 393–398. [14] Q.Y. Wu, F. Li, X.Y. Wang, Int. J. Biol. Macromol. 43 (2008) 367–372. [15] R. Virden, D.C. Watts, Biochem. J. 99 (1966) 162–172. [16] C. Roustan, L.A. Pradel, R. Kassab, N. Van Thoai, Biochim. Biophys. Acta 250 (1971) 103–116. [17] D.C. Watts, E.O. Anosike, B. Moreland, R.J. Pollitt, C.R. Lee, Biochem. J. 185 (1980) 593–599. [18] L. Lin, M.B. Perryman, D. Friedman, R. Roberts, T.S. Ma, Biochim. Biophys. Acta 1206 (1994) 97–104. [19] S.J. Strong, W.R. Ellington, Comp. Biochem. Physiol. B Biochem. Mol. Biol. 113 (1996) 809–816. [20] T. Hatakeyama, Y. Matsuyama, T. Funada, S. Fukuyama, H. Kuwahara, H. Aoyagi, N. Yamasaki, Biosci. Biotechnol. Biochem. 62 (1998) 1185–1189. [21] T. Okajima, Y. Kawata, K. Hamaguchi, Biochemistry 29 (1990) 9168–9175. [22] R. Furter, E.M. Furter-Graves, T. Wallimann, Biochemistry 32 (1993) 7022–7029. [23] Y. Yang, H.M. Zhou, Biochim. Biophys. Acta 1388 (1998) 190–198. [24] J.C. Pan, Y. Cheng, E.F. Hui, H.M. Zhou, Biochem. Biophys. Res. Commun. 317 (2004) 539–544. [25] J.L. Gattis, E. Ruben, M.O. Fenley, W.R. Ellington, M.S. Chapman, Biochemistry 43 (2004) 8680–8689. [26] P.F. Wang, A.J. Flynn, M.M. Naor, J.H. Jensen, G. Cui, K.M. Merz Jr., G.L. Kenyon, M.J. McLeish, Biochemistry 45 (2006) 11464–11472. [27] J.S. Wang, Z.L. Zheng, J. Lei, J.C. Pan, G.L. Zou, Comp. Biochem. Physiol. B Biochem. Mol. Biol. 153 (2009) 268–274. [28] Z. Yu, J. Pan, H.M. Zhou, Protein Pept. Lett. 9 (2002) 545–552. [29] R. Virden, D.C. Watts, E. Baldwin, Biochem. J. 94 (1965) 536–544. [30] B. Focant, D.C. Watts, Biochem. J. 135 (1973) 265–276. [31] L. Brand, J.R. Gohlke, Ann. Rev. Biochem. 41 (1972) 843–868. [32] D.H. Buttlaire, M. Cohn, J. Biol. Chem. 249 (1974) 5733–5740.
Contents lists available at ScienceDirect
International Journal of Biological Macromolecules journal homepage: www.elsevier.com/locate/ijbiomac
The role of Cys271 in conformational changes of arginine kinase Na Liu a,b , Jin-Song Wang a,b,∗ , Wei-Dong Wang a,b , Ji-Cheng Pan a,b,∗ a b
College of Life Science, Hubei Normal University, Huangshi, Hubei 435002, PR China Hubei Key Laboratory of Pollutant Analysis & Reuse Technology, PR China
a r t i c l e
i n f o
Article history: Received 3 March 2011 Received in revised form 2 April 2011 Accepted 4 April 2011 Available online 12 April 2011 Keywords: Arginine kinase Cys271 Conformational changes Site-directed mutagenesis
a b s t r a c t Arginine kinase (AK), a crucial enzyme in energy metabolism, buffers cellular ATP levels by catalyzing the reversible phosphoryl transfer between ATP and arginine. To better understand the role of Cys271 in conformational changes of AK from greasyback shrimp (Metapenaeus ensis), we replaced the residue with serine and alanine. A detailed comparison of the catalytic activity and conformation was made between wild-type AK and the mutants by means of activity analysis, ultraviolet (UV) difference, fluorescence spectrum and size exclusion chromatography (SEC). The results indicated that the catalytic activity of the two mutants was gone. The substrates, arginine–ADP–Mg2+ could induce conformational changes, and additional NO3 − could induce further changes in both the native enzyme and the variants. We speculated that Cys271 might be located in the hinge region between the two domains of AK and cause enzyme conformational changes upon addition of substrate. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Arginine kinase (AK, EC 2.7.3.3), a phosphagen kinase participating in cellular energy metabolism in invertebrates, buffers cellular ATP levels, which catalyzes the reversible phosphoryl transfer between ATP and arginine, leading to the production of phosphoarginine and Mg2+ ADP [1,2]. The latter is considered to be an energy reservoir, which is able to supply ATP, the primary energy source in bioenergetics, on demand [3]. The crystal structure of the transition state analog complex (TSAC) of monomeric AK from the horseshoe crab (Limulus) is determined at 1.86 A˚ [4] and refined at 1.2 A˚ [5], and it suggests that AKs can be divided into two domains, a smaller ␣-helical N-terminal domain and a larger C-terminal domain (residues 112–357). The catalytic center, where reversible transfer of phosphate occurs, is located in the C-terminal domain, but the N-terminal domain also contributes to arginine binding. The two domains undergo a hinge rotation of 13◦ upon formation of the TSAC. The substrate-free AK is considered to be in “open” form, changing to a “closed” form upon substrate binding [6]. The substrate-free structure of creatine kinase (CK) indicates that the induced fit observed in arginine
Abbreviations: WT-AK, wild-type arginine kinase; CK, creatine kinase; Cys271, cysteine of residue-271; SEC, size exclusion chromatography; TSAC, transition state analog complex; IPTG, isopropyl--d-thiogalactoside; ANS, 1-anilinonaphtalene-8sulfonate. ∗ Corresponding authors at: College of Life Science, Hubei Normal University, Huangshi, Hubei 435002, PR China. Tel.: +86 0714 6513163; fax: +86 0714 6513163. E-mail addresses: [email protected] (J.-S. Wang), [email protected] (J.-C. Pan). 0141-8130/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.ijbiomac.2011.04.002
kinase upon association of both substrates and the formation of TSAC (comprising Mg2+ ADP, nitrate and arginine) [7]. In the given examples, the conformational changes appear to be necessary for alignment of the two substrates for catalysis. The latest structural studies provide a detailed mechanism indicating that the binding of ADP–Mg2+ alone may trigger conformational changes in humanbrain-type-creatine-kinase (HbCK) [8]. During the reaction, critical elements of the active site are reconfigured and several loops and residues of key importance become ordered [7,9]. The roles of various amino acid residues in either substratebinding or catalytic activity of arginine kinase have been widely investigated. The substrate recognition region (GS region) of AK, which is composed of Ser63, Gly64, Val65 and Tyr68 is highly conserved [10]. Supporting evidence from enzyme kinetic and thermal stability experiments on wild-type and mutant AKs show that the activity of the enzyme is significantly reduced and that the substrate synergism is distinctly altered when the conserved sites are replaced [11–13]. The role of the amino acid residue P272 of AK is investigated via site-directed mutagenesis, which suggests that some residues near the active site may play a relatively important role in sustaining AK’s activity and structural stability [14]. Phosphagen kinases contain a single reactive cysteine residue (271 in AK, 278 in chicken mitochondrial creatine kinase) in each subunit which is highly conserved in all known phosphagen kinases sequences [15]. Comparison of the structure of the substrate-free and the transition-state of AK shows that, although Cys271 is located in the C-terminal domain, it moves towards the residues of the N-terminal domain during catalysis [6]. At present, whether the cysteine is essential for enzyme catalysis is still a controversial issue. The “essential” cysteine takes its name from early
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
chemical modifications [16,17] and mutagenesis experiments of the reactive cysteine that resulted in completely inactive enzymes [18,19]. However, chemical modification may lead to stability and structural changes in the enzymes [20,21]. The effect of a gross modification can be structurally rationalized by steric conflict with the substrate guanidinium [4]. More subtle effects of mutations [22] and chemical modifications of the cysteine [23,24] lead to several hypotheses that the cysteine may not be essential for activity, but may be involved in synergism upon the binding of both substrates or in mediating the substrate-induced hinge movement. It has been shown that the activity of uncharged reactive cysteine mutations could be partially rescued with chloride, leading to the proposal that the cysteine functions as a thiolate ion (AA–CH2 –S− ) [22]. The recent crystal structure for the TSAC of the C271A mutant at 2.3 A˚ resolution shows minor structural changes, and suggests that the cysteine enhances the catalytic rate through electrostatic stabilization of the transition state [25]. Experiments on wild-type and cysteine-substituted human muscle creatin kinase (HMCK) show that an ionized active site cysteine is required across the entire family of guanidino kinases [26]. This report focuses on the reactive Cys271, which is close to the catalytic center of AK. It examines the role of Cys271 in catalysis and the conformational change upon substrate binding by means of activity analysis, ultraviolet (UV) difference spectra, fluorescence spectra and size-exclusion chromatography (SEC). 2. Materials and methods 2.1. Cloning, and site-directed mutagenesis of the greasyback shrimp AK The pET-28a plasmid with the insert of the greasyback shrimp AK gene (WT-AK) [27] was used as a template for mutagenesis. The replacement of the amino acid Cys271 by alanine (Ala) and serine (Ser) were achieved by PCR with the KOD-Plus mutagenesis Kit (Toyobo, Tokyo, Japan), which are denoted C271A and C271S, respectively. The primers used for PCR amplification (mutated codons are underlined) were: C271A: 5 -GCTCCCACCAACCTTGGGCACCACTGTGC-3 3 -GTGCTGGCGGACCCGAAGGAGTGAAAG-5 ; C271S: 5 -AGTCCCACCAACCTTGGCACCACTGTGC-3 3 -GTGCTG-GCGGACCCGAAGGAGTGAAAG-5 . For both mutants, the fidelity of the PCR amplification was confirmed by sequencing.
99
ous pH-spectrophotometric assay [28]. The enzyme concentration was estimated from the absorbance at 280 nm (absorbance 0.67 at 280 nm in 1 cm cuvette corresponds to 1 mg protein/mL) [29]. The enzymatic activity and concentration of the proteins were measured with an Ultrospec 4300 pro UV/visible spectrophotometer. 2.4. Spectroscopic measurements UV difference spectra of the enzyme complexes with substrates or transition-state analogues were measured relative to the substrate-free enzymes as described by Focant and Watts [30]. All absorption spectra were measured with an Ultrospec 4300 pro UV/visible spectrophotometer. Intrinsic fluorescence emission spectra were recorded on an F-4500 fluorescence spectrophotometer (Hitachi) with 1 cm pathlength cuvettes, excitation at 295 nm and emission wavelength in the range of 300–400 nm. ANS was used as an extrinsic probe [31], and a 20-fold molar excess of ANS was added to the samples for 30 min in the dark. An excitation wavelength of 380 nm was used to determine the ANS fluorescence intensity of the substratefree/bound enzymes in the range of 400–600 nm. All spectroscopy measurements were recorded after the enzymes were added to the substrate (ADP–Mg2+ ) for 0.5 h in the similar reaction condition described above, and NO3 ¯ was added to the reaction mixture above for another 0.5 h at 25 ◦ C. Appropriate buffers were used for baseline determinations. Stored spectra were corrected for dilution and baseline effects. 2.5. Size-exclusion chromatography Gel filtration was performed with a Superdex 200HR 10/30 column (fractionation range 1000–600,000) on a Pharmacia AKTA purifier apparatus at 25 ◦ C. All prepared solutions were passed through a filter and degassed. A 500 L sample of enzyme was added to ADP–Mg2+ for 0.5 h and ADP–Mg2+ –NO3 ¯ for 1 h and was then injected into a column equilibrated with buffer A (50 mM Tris/HCl, 500 mM NaCl, 0.57 mM arginine, 0.1 mM ADP, and 0.66 mM magnesium acetate, pH 8.0) and buffer B (0.5 mM NaNO3 added to buffer A), respectively. The flow rate was maintained at 0.5 mL min−1 , and the results were monitored at 280 nm. The distribution coefficient Kav [Kav = (Ve − V0 )/(Vt − V0 )], with Ve being the elution volume, V0 the void volume and Vt the column total volume was used to monitor the Stokes radius changes of the enzyme molecules during the enzyme association with the substrates. 3. Results
2.2. Expression and purification 3.1. Purification of the site-directed mutants To facilitate the expression of the mutagenic DNA, the recombinant plasmids (mutants with 6×His-tag) were transformed into the Escherichia coli (E. coli) strain Rosetta. The cells were grown at 37 ◦ C then induced with 0.8 mM IPTG at 22 ◦ C. The Wt-AK was expressed as described previously [27], but purified as mutant AKs using nickel affinity chromatography. The Ni2+ column was washed first with a washing buffer (50 mM Tris/HCl, 0.5 M NaCl, pH 8.0) and then washed gradually with the same buffer containing an imidazole gradient. The purified enzymes were confirmed to be homogeneous by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). The enzymes were placed on ice until use, and the enzymatic activity was determined within 12 h.
To elucidate the role of Cys271 in catalysis and the conformational change upon substrate binding, the Cys was replaced by Ala and Ser residues, respectively, resulting in the mutants C271A and C271S. The recombinant proteins were obtained from the E. coil strain Rosetta transformed with plasmids containing the genes for the wild-type and mutated AK. The purified enzymes were proven to be homogeneous by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 1). However, mutagenesis of the Cys271 resulted in completely inactive enzymes (data not shown). 3.2. Conformational changes of WT-AK and the mutants upon substrates binding
2.3. Enzyme concentration and activity The activity of wild-type and mutant AK in the forward reaction (arginine phosphate synthesis) was assayed using a direct continu-
The physico-chemical properties of the wild-type enzyme and mutants in the presence of either the substrates or the transitionstate analog were explored by intrinsic fluorescence (Fig. 2), ANS
100
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
Fig. 1. Analysis of the purified Wt-AK and mutated AK on 12% SDS-PAGE. M, the molecular mass standard. The variants are as indicated in the figure.
fluorescence (Fig. 3), ultraviolet (UV) difference (Fig. 4), and sizeexclusion chromatography (SEC) (Fig. 5). The intrinsic fluorescence (Fig. 2), which was excited at 295 nm to minimize the contributions of residues other than Trp, indicated that the maximum emission wavelengths (Em ) were similar, but the fluorescence intensity decreased sharply upon substrate binding
(shown as curve 2) and then slightly decreased after NO3 ¯ addition as the TSAC formed (shown as curve 3). This suggested that the substrates arginine–ADP–Mg2+ and additional NO3 ¯ might induce microenvironmental changes of the Trp residues upon association with the enzymes and affect the conformational changes of the WTAK and mutant AK. The ANS fluorescence intensity of WT-AK was higher than substrate-bound WT-AK or the TSAC at the same concentration (Fig. 3A), which reflected that the “closed” form of AK had less hydrophobic exposure than did the “open” form to allow the binding of the ANS molecules. Similar to the WT-AK’s, a significant decrease of the ANS fluorescence intensity was observed for the two mutants (Fig. 3B and C). Fig. 4A shows the overall change (240–320 nm) to the UV spectrum of wild-type AK when excess substrates and NO3 ¯ were added to the free protein. Upon substrate binding and TSAC formation, the magnitude of the induced difference spectrum increased between 240 and 270 nm with the peak center at approx. 260 nm. This is characteristic of the phenylalanine spectrum. There was a decreased in tyrosine absorbance as shown by negative peaks centering around 286 nm. A decrease in tyrosine absorbance in proteins is typically associated with the exposure of tyrosine to a more polar
Fig. 2. Intrinsic fluorescence spectra for wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. Final concentrations were 14.2 M enzyme in 50 mM Tris/HCl buffer, pH 8.0, 0.57 mM arginine, 0.1 mM ADP, and 0.66 mM magnesium acetate. NaNO3 (0.5 mM) was also added in addition to the substrates for the measurement of the intrinsic fluorescence of the complex with the transition-state analog. In subparts (A)–(C) the curves represent different state of enzyme as follows: (1) represents spectra of apo-enzyme, (2) enzyme–substrate complex and (3) the complex with the transition-state analog.
Fig. 3. ANS fluorescence spectra for wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. The conditions were the same as for Fig. 2. In subparts (A)–(C) the curves represent different states of enzyme as follows: (1) represents the emission fluorescence spectra of apo-enzyme, (2) enzyme–substrate complex and (3) the complex with the transition-state analog.
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
101
Fig. 4. Spectra for wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. The conditions were the same as for Fig. 2. In subparts (A)–(C) the curves represent different states of enzyme as follows: (1) represents the difference spectrum for apo-enzyme as contrast, (2) enzyme–substrate complex and (3) the complex with the transition-state analog.
Fig. 5. Size exclusion chromatography of wild-type arginine kinase (A), C271A (B) and C271S (C) with the substrates and the transition-state analog. The parameters of the column (HR10/30, Pharmacia) were V0 = 7.77 mL and Vt = 23.562 mL. The other conditions were the same as for Fig. 2. In subparts (A)–(C) the curves represent different states of enzyme as follows: (1) represents apo-enzyme, (2) enzyme–substrates complex and (3) the complex with the transition-state analog.
environment. A similar comparison of arginine–ADP–Mg2+ and additional NO3 ¯ induced difference spectrum for the two mutants (C271A and C271S) showed that the conformational changes of C271A (Fig. 4B) were nearly the same as WT-AK. C271S had a different UV difference spectrum (Fig. 4C), which may be due to the structural changes after cysteine was replaced by serine. The retention times of size exclusion chromatography had been obtained as shown in Table 1, and mean SD values had also been calculated. The elution volume of substrate-free AK was different from the substrate-bound form and TSAC (Fig. 5A), and the two mutants had the same phenomena (Fig. 5B and C). Although the change was subtle, it also suggested that the conformation of the wildtype AK and the two mutants was changed upon substrate binding and TSAC formation, which is consistent with the above conclusion from the measurement of fluorescence spectra (Figs. 2 and 3) and UV difference spectra (Fig. 4). 4. Discussion In a previous study, the gene encoding arginine kinase from greasyback shrimp (Metapenaeus ensis) was suc-
cessfully cloned, sequenced, and over-expressed in E. coli [27]. Modified arginine kinase reactivated by dithiothreitol suggests that the reactive cysteine plays an important role in the Table 1 Retention volumes of wild-type and mutation-type AKs in SEC. Proteins WT-AK
C271A
C271S
Mean ± SD
Retention volume (mL) a
17.166 17.277b 17.452c 17.256a 17.438b 17.589c 17.237a 17.421b 17.335c
a
17.158 17.265b 17.435c 17.235a 17.437b 17.574c 17.254a 17.402b 17.301c
a
17.175 17.287b 17.470c 17.215a 17.461b 17.593c 17.246a 17.431b 17.345c
17.163 17.276 17.452 17.235 17.445 17.585 17.246 17.418 17.327
± ± ± ± ± ± ± ± ±
0.008505 0.011015 0.017502 0.020502 0.013577 0.010017 0.008505 0.014731 0.023065
Samples (WT-AK, C271A, C271S) were added to calibrated column of Superdex 200 (HR10/30, Pharmacia), and its retention volumes were monitored and recorded, respectively. Every sample was detected three times, and SD value was calculated. a apo-enzyme. b Enzyme–substrates complex. c The complex with the transition-state analog.
102
N. Liu et al. / International Journal of Biological Macromolecules 49 (2011) 98–102
conformational changes caused by the transition-state analog [24]. However, chemical modification may lead to stability and structural changes of the enzymes [20,21]. In this research, the two mutants, C271S and C271A, were successfully cloned, and their expression level was lower than wild-type AK (WT-AK) (Fig. 1). Kinetic analysis showed that C271S and C271A had no activity (data not shown). The results of spectroscopic measurements (Figs. 2–5) illustrated that the substrates, arginine–ADP–Mg2+ , could induce conformational changes and that additional NO3 ¯ could induce further changes in both the wild-type enzyme and the variants. This suggested that Cys271 was not located in the ATP/ADP binding site but very close to it, such that the mutants could bind to the substrate even though there was no enzyme activity. This was in accord with the reported crystal structure of AK from horseshoe crab [4], which shows that the reactive Cys is near the ␥ phosphate of ATP. Fluorescence spectroscopy is an effective method of studying protein conformation in aqueous solution. Proteins autofluorescence (Fig. 2), and the fluorescence of the fluorogenic probe associated with the proteins in specific areas (Fig. 3) was detected in this paper. As shown in the figures, the microenvironment of the Trp residues was changed upon substrate binding (Fig. 2) and the “closed” form of AK had less hydrophobic exposure than did the “open” form to allow the binding of the ANS molecules (Fig. 3). This illustrated that the substrates arginine–ADP–Mg2+ and additional NO3 ¯ might affect the conformational changes of WT-AK and mutations after association with enzymes. It is well known that, in the presence of the substrates (arginine–ADP–Mg2+ ) or the transition-state analog (MgADP–arginine–NO3 − ), the enzyme undergoes conformational changes as indicated by absorbance changes in the UV region [30]. The substrates, arginine–ADP–Mg2+ , could also induce a change of the absorbance of the mutated arginine kinase in the UV region (Fig. 4B and C), showing that MgADP could bind to the mutated enzyme. The transition-state analog was certified by NMR to mimic the binding of substrate in the transition state during catalysis, involving probable conformational changes of the enzyme molecule [32]. The present work showed that both arginine–ADP–Mg2+ and additional NO3 ¯ could induce fluorescence intensity and UV absorbance changed in the two mutants similar to those observed in the wild-type arginine kinase. Interestedly, we had found that compared to WT, the conformational changes of the mutants could be induced by substrate or NO3 ¯ whereas these changes were not fit the enzyme catalysis, as C271A and C271S had no detectable activity. Analyzing the phenomenon, we speculated that there might be two reasons: (1) the conformational changes trend of the mutants might not be same as WT; (2) although the trend was similar to WT, the two domains underwent a hinge rotation that might not be 13◦ , which happens upon formation of the TSAC of WT-AK [6]. The results differed somewhat from our previous study, in which NO3 ¯ did not induce further absorbance changes of the modified enzyme [24]. There
may be steric effect due to the structural changes of chemical modification of AK. In conclusion, Cys271 may play an important role in the conformational changes caused by the substrates and the transition-state analog, and it may be located in the hinge region of the two domains of AK, guiding conformational changes of the enzyme upon addition of substrates. Acknowledgements The present investigation was supported by grant 30570412, 2005ABA192, 2008CDA066 from National Natural Science Foundation of China and Hubei Province’s Natural Science Foundation of China. References [1] E.A. Newsholme, I. Beis, A.R. Leech, V.A. Zammit, Biochem. J. 172 (1978) 533–537. [2] W.R. Ellington, J. Exp. Biol. 143 (1989) 177–194. [3] W.R. Ellington, Ann. Rev. Physiol. 63 (2001) 289–325. [4] G. Zhou, T. Somasundaram, E. Blanc, G. Parthasarathy, W.R. Ellington, M.S. Chapman, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 8449–8454. [5] M.S. Yousef, F. Fabiola, J.L. Gattis, T. Somasundaram, M.S. Chapman, Acta Crystallogr. D Biol. Crystallogr. 58 (2002) 2009–2017. [6] M.S. Yousef, S.A. Clark, P.K. Pruett, T. Somasundaram, W.R. Ellington, M.S. Chapman, Protein Sci. 12 (2003) 103–111. [7] M. Forstner, M. Kriechbaum, P. Laggner, T. Wallimann, Biophys. J. 75 (1998) 1016–1023. [8] S.M. Bong, J.H. Moon, K.H. Nam, K.S. Lee, Y.M. Chi, K.Y. Hwang, FEBS Lett. 582 (2008) 3959–3965. [9] G. Zhou, W.R. Ellington, M.S. Chapman, Biophys. J. 78 (2000) 1541–1550. [10] T. Suzuki, H. Fukuta, H. Nagato, M. Umekawa, J. Biol. Chem. 275 (2000) 23884–23890. [11] K. Uda, T. Suzuki, Protein J 23 (2004) 53–64. [12] N. Fujimoto, K. Tanaka, T. Suzuki, FEBS Lett. 579 (2005) 1688–1692. [13] Q.Y. Wu, F. Li, X.Y. Wang, Int. J. Biol. Macromol. 45 (2009) 393–398. [14] Q.Y. Wu, F. Li, X.Y. Wang, Int. J. Biol. Macromol. 43 (2008) 367–372. [15] R. Virden, D.C. Watts, Biochem. J. 99 (1966) 162–172. [16] C. Roustan, L.A. Pradel, R. Kassab, N. Van Thoai, Biochim. Biophys. Acta 250 (1971) 103–116. [17] D.C. Watts, E.O. Anosike, B. Moreland, R.J. Pollitt, C.R. Lee, Biochem. J. 185 (1980) 593–599. [18] L. Lin, M.B. Perryman, D. Friedman, R. Roberts, T.S. Ma, Biochim. Biophys. Acta 1206 (1994) 97–104. [19] S.J. Strong, W.R. Ellington, Comp. Biochem. Physiol. B Biochem. Mol. Biol. 113 (1996) 809–816. [20] T. Hatakeyama, Y. Matsuyama, T. Funada, S. Fukuyama, H. Kuwahara, H. Aoyagi, N. Yamasaki, Biosci. Biotechnol. Biochem. 62 (1998) 1185–1189. [21] T. Okajima, Y. Kawata, K. Hamaguchi, Biochemistry 29 (1990) 9168–9175. [22] R. Furter, E.M. Furter-Graves, T. Wallimann, Biochemistry 32 (1993) 7022–7029. [23] Y. Yang, H.M. Zhou, Biochim. Biophys. Acta 1388 (1998) 190–198. [24] J.C. Pan, Y. Cheng, E.F. Hui, H.M. Zhou, Biochem. Biophys. Res. Commun. 317 (2004) 539–544. [25] J.L. Gattis, E. Ruben, M.O. Fenley, W.R. Ellington, M.S. Chapman, Biochemistry 43 (2004) 8680–8689. [26] P.F. Wang, A.J. Flynn, M.M. Naor, J.H. Jensen, G. Cui, K.M. Merz Jr., G.L. Kenyon, M.J. McLeish, Biochemistry 45 (2006) 11464–11472. [27] J.S. Wang, Z.L. Zheng, J. Lei, J.C. Pan, G.L. Zou, Comp. Biochem. Physiol. B Biochem. Mol. Biol. 153 (2009) 268–274. [28] Z. Yu, J. Pan, H.M. Zhou, Protein Pept. Lett. 9 (2002) 545–552. [29] R. Virden, D.C. Watts, E. Baldwin, Biochem. J. 94 (1965) 536–544. [30] B. Focant, D.C. Watts, Biochem. J. 135 (1973) 265–276. [31] L. Brand, J.R. Gohlke, Ann. Rev. Biochem. 41 (1972) 843–868. [32] D.H. Buttlaire, M. Cohn, J. Biol. Chem. 249 (1974) 5733–5740.