Báo cáo khoa học: Cytoglobin conformations and disulfide bond formation pdf

Cytoglobin conformations and disulfide bond formation Christophe Lechauve 1 ,Ce ´ dric Chauvierre 1 , Sylvia Dewilde 2 , Luc Moens 2 , Brian N. Green 3 , Michael C. Marden 1 , Chantal Ce ´ lier 1 and Laurent Kiger 1 1 Inserm U779, Universite ´ s Paris VI et XI, Le Kremlin-Bice ˆ tre, France 2 Department of Biomedical Sciences, University of Antwerp, Belgium 3 Waters MS Technologies Centre, Micromass UK Ltd., Altrincham, Cheshire, UK Introduction The vertebrate heme globin family has been extended to include two new members, neuroglobin (Ngb) and cytoglobin (Cygb), which differ in structure, tissue distribution and function [1–3]. Cygb is expressed at varying concentrations in many body tissues, such as connective tissue, fibroblasts and neurons [3–5]. The cellular localization of Cygb is atypical; it is present in the cytoplasm of almost all cell types; it is also found in the cytoplasm and nucleus of neuronal cells [4,6]. The physiological role of Cygb remains unknown, although several roles have been proposed, such as the protection of cells from oxida- tive-related damage during ischemic reperfusion injury following hypoxia [7,8], the metabolism of NO in tis- sues [9], and collagen synthesis [4,10]. It has also been reported that Cygb plays a role in cancer as a tumor suppressor gene; indeed, it was observed that the pro- moter region of Cygb is hypermethylated and under- expressed in tumors [11]. Cygb is considered to be in a clade with vertebrate myoglobin (Mb) and shares about 30% amino acid sequence identity with Mb, implying a common evolu- tionary ancestry [3]. Human Cygb is composed of 190 amino acids with the presence of extended N- and C-terminal regions of about 20 residues each. The crys- tal structure of this globin is characteristic of the clas- sical three-over-three a-helical globin fold in the core region of each subunit, and the asymmetric unit of the crystals contains two molecules of Cygb [12,13]. Cygb contains two exposed cysteine residues (Cys B2 and Cys E9), suggesting the possibility of inter or intra- molecular disulfide bridge formation. Spectroscopic studies have shown that Cygb belongs to the class of hexacoordinated globins, for which the Keywords disulfides, globins, kinetics, ligand binding, light scattering Correspondence L. Kiger, INSERM – U779, 78 rue du Ge ´ ne ´ ral Leclerc, Ho ˆ pital de Bice ˆ tre Bat. Broca, Niveau 3, 94275 Le Kremlin Bice ˆ tre, France Fax: (33) 1 49 59 56 61 Tel: (33) 1 49 59 56 64 E-mail: Laurent.Kiger@inserm.fr (Received 18 February 2010, revised 8 April 2010, accepted 13 April 2010) doi:10.1111/j.1742-4658.2010.07686.x The oligomeric state and kinetics of ligand binding were measured for wild-type cytoglobin. Cytoglobin has the classical globin fold, with an extension at each extremity of about 20 residues. The extended length of cytoglobin leads to an ambiguous interpretation of its oligomeric state. Although the hydrodynamic diameter corresponds to that of a dimer, it displays a mass of a single subunit, indicating a monomeric form. Thus, rather than displaying a compact globular form, cytoglobin behaves hydro- dynamically like a tightly packed globin with a greater flexibility of the N- and C-terminal regions. Cytoglobin displays biphasic kinetics after the photolysis of CO, as a result of competition with an internal protein ligand, the E7 distal histidine. An internal disulfide bond may form which modifies the rate of dissociation of the distal histidine and apparently leads to different cytoglobin conformations, which may affect the observed oxygen affinity by an order of magnitude. Abbreviations Cygb, cytoglobin; DLS, dynamic light scattering; Mb, myoglobin; Ngb, neuroglobin; SEC-MALLS, size exclusion chromatography with multi- angle laser light scattering. 2696 FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works binding of an external gaseous ligand to heme requires the dissociation of an internal protein residue (His E7); both ligands are in competition for the distal heme bind- ing site. A biphasic pattern was observed, as expected for a hexacoordinated model of competition between two ligands for a single binding site [14]. In this study, we re-examined the structure–function relationships of Cygb, as several questions have been raised concerning the basic properties of Cygb, namely the role of the disulfide bond in protein function and its involvement in the tertiary and quaternary structures. Results and Discussion Quaternary structure analysis of human Cygb by size exclusion chromatography with multi-angle laser light scattering (SEC-MALLS) and DLS The theoretical molar mass of Cygb is 21 404 Da. Cygb has been characterized previously by SEC as a dimeric protein [12,14]. The elution volume measured by SEC is directly related to the hydrodynamic radius of the pro- teins, and is generally correlated with the logarithm of the molar mass. The Cygb multimeric state was initially determined using molar mass markers belonging to the globin protein family, but this established method depends on the reliability of the size of the markers and may not be accurate in the absolute molar mass deter- mination. For this reason, we used static light scatter- ing (MALLS) for the determination of the molar masses of the different Cygb protein species in solution. SEC-MALLS analysis gave a molar mass of 21 400 ± 100 Da for Cygb in solution (Fig. 1A). It should be noted that this method does not rely on reference proteins. We also estimated the hydrodynamic diameter using two independent instruments (Malvern Instruments, Malvern, Worcestershire, UK and Wyatt Technology, Santa Barbara, CA, USA); it was found to be indepen- dent of the redox and ligation states of Cygb. The hydrodynamic diameter versus molar mass was plotted for Cygb (5.2 nm), together with several globular pro- teins (Fig. 1B). It appears that this size parameter value for Cygb does not correlate well with the linear relationship found between the hydrodynamic diameter and molar mass for other proteins belonging to the globin family. SEC confirmed that the N- and C-termi- nal extensions of Cygb ( 20 residues each) are responsible for a large increase in hydrodynamic vol- ume for Cygb relative to a globular protein. A trun- cated Cygb of approximately 17 kDa, without the extensions, exhibits similar hydrodynamic properties to those of ‘classical’ globins (Fig. S1). These results explain the misinterpretation of the Cygb quaternary assembly arising from gel filtration experiments based only on a putative relationship between the hydrodynamic volume and molar mass [12,14]. It should also be noted that electrostatic repulsion (or attraction) forces between the proteins and column packing may contribute to a deviation from an ideal elution behavior based only on the hydrodynamic diameter, in this case for globular pro- teins. The N- and C-terminal residue segments of Cygb were not clearly resolved in two of the X-ray crystal structures [12,13]. In another structure display- ing covalently bound dimer through two intermolecu- lar disulfide bridges [15], the extensions formed an additional a-helix and an ordered loop, respectively, A B Fig. 1. (A) Molar mass determination of Cygb using SEC-MALLS: the thin full curve represents the relative UV absorbance at 280 nm; the bold line corresponds to the molar mass estimated from the light scattering profile and concentration measurement; both versus the elution time. (B) Correlation between the hydrody- namic diameters (measured by dynamic light scattering; Malvern Instruments and Wyatt Technology) and molar masses of six globu- lar proteins (monomeric a-globin, Ngb, Mb, tetrameric hemoglobin and octameric hemoglobin). The Cygb (red) hydrodynamic diameter deviates from the correlation line of other globin proteins. C. Lechauve et al. Cytoglobin conformations and disulfide bond formation FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works 2697 but not in a symmetrical way, as only one extremity is well defined in each subunit. Thus, the conforma- tional flexibility of both terminal regions may explain the increase in hydrodynamic diameter relative to other globins. Mass spectrometry assessment of Cygb disulfide bonds and general structure The results obtained by MALLS were complemented by electrospray ionization mass spectrometry (ESI- MS). When analyzed under conditions that should retain the noncovalent bonds for globin assembly [14], the dominant species was monomeric Cygb, together with approximately 5% dimeric Cygb. A solution of Cygb with dithiothreitol, a thiol reducing agent, ana- lyzed under denaturing conditions, showed predomi- nantly a monomer with a mass of 21 404.7 ± 0.3 Da (sequence mass 21 404.7 Da) and no dimer (Fig. 2A). When this solution was incubated at 37 °C and pH 8 after dithiothreitol removal, the mass of Cygb progres- sively decreased over 7 h to 21 402.6 ± 0.3 Da (Fig. 2B), implying the formation of an intramolecular disulfide bond (calculated mass 21 402.7 Da). Also pres- ent was a minor component (not shown) of mass 42 805.3 ± 5 Da ( 5% of Cygb), which is consistent with a Cygb dimer formed by one or two intermolecular disulfide bonds between two Cygb monomers (calculated masses of 42 807.3 and 42 805.3 Da, respec- tively). As expected, when the solution of Cygb used to produce Fig. 2B was reduced with dithiothreitol, the mass of the Cygb monomer reverted to 21 404.7 ± 0.3 Da (Fig. 2C), and the putative dimer disappeared. These results confirm the occurrence of mainly mono- meric Cygb with an intramolecular disulfide bridge. Crystallographic structure The initial crystallographic structures of Cygb were described as dimers [12,13]. There are major differences between these two structures. One was obtained from a mutated protein without cysteines [Cys38(B2)Ser and Cys83(E9)Ser mutant], whereas the other protein con- tained the native cysteines. An asymmetric unit includ- ing two globins is observed for Cygb, and does not necessarily imply a dimeric form in dilute solutions. However, Sugimoto’s dimeric structure (pdb 1V5H [13]), with two intermolecular disulfide bridges, is not present significantly in our results, which show mono- meric Cygb with an intramolecular disulfide bridge at protein concentrations of 0.1–20 lm. Structure com- parison shows a difference at the position of Cys B2, with the distance between Ca atoms for residues B2 and E9 of 7.35 A ˚ in de Sanctis’ structure (pdb 1UM0 Reduced Cygb Oxidized Cygb Reduced Cygb A B C Fig. 2. ESI-MS analyzed under denaturing conditions of: (A) a solution of reduced Cygb; (B) solution in (A) after 7 h of incubation at 37 °C under aerobic conditions, pH 8; (C) solution in (B) after reduction with dithiothreitol. Cytoglobin conformations and disulfide bond formation C. Lechauve et al. 2698 FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works [12]) and 12.0 A ˚ in Sugimoto’s structure (Fig. 3). There is a loss of one turn of the B-helix in Sugimoto’s struc- ture, with Cys B2 moving away from the Cys E9 posi- tion, which would not allow an intramolecular disulfide bridge (Fig. 3). The buried interface surface of Sugimoto’s structure [13] is half that (640 A ˚ 2 ) found for another Cygb dimeric structure (2DC3 [15]) exhib- iting the same intermolecular disulfide bonds. In this latter crystal structure, in addition to the fact that only one terminal sequence was ordered in each monomer (one a-helix and one ordered loop motif), structural differences were also found in the conformation of certain residues in the vicinity of heme. Ligand binding of human Cygb Cygb shows the characteristic absorption spectra (Fig. S2) and ligand binding for the hexacoordinated globins. The kinetics of ligand binding to Cygb after the photodissociation of CO shows the form expected for hexacoordinated globins (Fig. 4A). The rapid bimolecular phase corresponds to the competitive binding between CO and the internal histidine residue (His E7). For the fraction binding histidine, the return to the final (CO-bound) state involves the slow dissociation of histidine. A previous study [16] has reported a single exponential decay, probably because the observation time was not sufficiently long for histidine dissociation and replacement by CO. E9 B2 Fig. 3. Two Cygb crystallographic structures, superimposed by the full sequence method of DS-Visualizer (Accelrys), which takes into account all common atoms of the aligned globins. The residues dis- played correspond to the position for the cysteine pair (B2 and E9) and the distal and proximal histidines (E7 and F8) in Cygb. The structure in red of human Cygb was obtained by Sugimoto et al. [13] (1V5H) and the structure in green (B2 and E9 in purple) was obtained by De Sanctis et al. [12] with a mutated protein (1UMO) without cysteines: Cys38(B2)Ser and Cys83(E9)Ser. B A Fig. 4. Flash photolysis kinetics for Cygb at 25 °C. (A) Recombina- tion kinetics at different CO concentrations [from top to bottom: 0.01, 0.1 and 1 atm CO (760 Torr)]. After flash photolysis of CO, the first phase represents competitive binding between CO and histidine to the heme sites. The second phase is a slow replace- ment reaction of histidine by CO to return to the preflash state. The broken lines are simulations using the model for competitive ligand binding. (B) Kinetic curves of the slow phase. Samples with dithiothreitol (DTT)-reduced Cygb show a single form (also observed for the mutant Cygb without cysteines). The cysteine oxi- dized form, which can form an internal S–S bond, displays a faster replacement reaction; however, heterogeneity is observed, indicat- ing the presence of two protein conformations. The rapid phase arising from the competition between CO and histidine for heme rebinding (with an observed rate equal to the sum k CO on + k His on ) is the same for samples with or without DTT. C. Lechauve et al. Cytoglobin conformations and disulfide bond formation FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works 2699 By analyzing the data from experiments performed at different CO concentrations, one can extract the rates for CO and histidine association, as well as those for histidine dissociation [14,17]. Results from experi- ments with a mixed CO–O 2 atmosphere allow a deter- mination of the O 2 binding rates. The intrinsic O 2 or histidine dissociation is quite slow, requiring about 1 s (Fig. 4A). Overall, the kinetics show the same form as for Ngb, although the ligand association rates to Cygb are slower than those to Ngb. As this occurs for both oxy- gen and histidine, the overall oxygen affinity remains on the order of 0.1–10 Torr at 25 °C, which is a common range of O 2 affinity measured for the hexaco- ordinated globins. For both Cygb and Ngb, the slow oxygen and histidine dissociation rates compensate for the effect on the overall oxygen affinity. In the case of Ngb, there is a shift in the observed oxygen affinity from about 1 to 10 Torr on reduction of the cysteines; the internal mechanism is mainly caused by a change in the histidine dissociation rate. Cygb shows a similar effect, with the oxygen affinity changing from about 0.2 to 2 Torr (Fig. 5); however, the transition only occurs for about one-half of the molecules. The major change in kinetics of Cygb after the addi- tion of dithiothreitol, known to reduce cysteines and therefore to break S–S bonds, is shown in Fig. 4B. Although the cysteine-reduced form shows homoge- neous kinetics, there is heterogeneity of the histidine dissociation rate for the protein solution exposed to O 2 after purification. Two histidine dissociation rates are measurable, each representing about one-half of the sample, and differ by a factor of 10 (Fig. 4B, Table 1), whereas the other binding parameters are little changed. This heterogeneity could not be eliminated by vari- ous preparations from different laboratories or by changing the temperature and pH conditions. It thus appears to be a basic property of Cygb samples. One can then question whether there is an interconversion of the two conformations. Any interconversion of the conformations is apparently slow, as the ligand bind- ing and Fe 2+ oxidation kinetics (which occur on the order of several minutes) display distinct phases for the two forms. Indeed, a partial autoxidation under air of the heme iron results in an enhancement of the higher oxygen affinity fraction (the more slowly oxidiz- ing form) as assessed by flash photolysis (data not shown). The stopped-flow data for the replacement reaction indicate a further complexity. For samples with and without dithiothreitol, at least two rates were observed (this work) [18]. This could indicate several possible conformations for the distal histidine, and only certain of these forms are apparent when the sample has been incubated in the His-Fe-His state. As seen in the P 50 (Torr) 0.1 1 10 100 Cygb S-S Oxygen affinity Cygb T-state HbA Mb R-state HbA Ngb S-S Ngb Fig. 5. Synthesis of the oxygen affinity for the various allosteric states of members of the globin family. For Ngb and Cygb, the oxy- gen affinity, conserved for various species, falls within the physio- logical range, supporting a role for oxygen delivery. Table 1. Kinetic parameters determined from flash photolysis. DTT, dithiothreitol; WT, wild-type. Protein k CO on (lMÆs )1 ) k O 2 on (lMÆs )1 ) k O 2 off (s )1 ) K O 2 (nM) k His on (s )1 ) k His off (s )1 ) K His P O 2 50 (Torr) Cytoglobin WT with DTT 0.65 2.5 0.9 35 140 1.5 90 1.8 Doubly mutated Cygb without Cys 0.65 2.5 0.9 35 180 1.5 120 2.4 Cytoglobin WT with S–S bridge Conformation 1 0.65 2.5 0.9 35 130 0.9 140 2.8 Conformation 2 0.65 2.5 0.9 35 130 12 11 0.2 Experimental conditions: 50 m M phosphate buffer at pH 7.0, 25 °C. The O 2 solubility coefficient was taken as 1.82 lMÆTorr )1 . Note the distinction between the intrinsic affinity K O 2 = k off ⁄ k on (as for pentacoordinated forms) and the overall affinity observed for the ligand compe- tition of the hexacoordinated globins K O 2 =(k off ⁄ k on ) ⁄ (1 + K His ). Cytoglobin conformations and disulfide bond formation C. Lechauve et al. 2700 FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works structural images (Fig. 3), at least two distinct posi- tions for Cys(B2)38 can be observed for Cygb without the internal disulfide bond; this residue may thus adopt various intermediate positions which may influence the ligand binding kinetics. A slow relaxation process might explain the two conformations as, in the stopped-flow experiments, the protein is incubated in the His-Fe-His state for at least several minutes. It should be noted that no change in the proportion of the kinetic phases was observed after repeated photolysis of the hexacoordinated form over 5 s by a series of laser pulses at 10 Hz to remove the CO ligand (for samples equilibrated under a 5% CO atmosphere), meaning that slow relaxation between the Cygb conformations does not take place within a time- scale of seconds. We also measured by stopped flow the histidine replacement by CO, starting from the hexacoordinated ferric form. Using a double mixing sequence, one can first reduce the iron and then incu- bate Cygb in the ferrous hexacoordinated state for var- ious delay times before mixing with CO to probe the histidine to CO replacement reaction. No difference was observed for incubation times ranging from milli- seconds to seconds, meaning that the redox state of iron does not influence the Cygb conformation in the bis-histidyl bound state. As the rapid phase of the his- tidine to CO replacement reaction seems to be most characteristic of the flash photolysis method, this spe- cial state could depend on the recent history of binding of an external ligand. For both Cygb and Ngb [14], the disulfide bond influences the final position of the E-helix, which, in turn, modifies the affinity for the distal histidine. As the overall O 2 affinity depends on histidine binding, this leads to a modification in O 2 affinity. The mecha- nism in Ngb involves a disulfide bond in the CD loop region which influences the position of the E-helix. For Cygb, a more direct mechanism can be envisaged from the structure: the Cys E9 and the distal His E7 are on opposite sides of the E-helix (Fig. 3); thus formation of the E9–B2 bond should pull directly on His E9. This change in conformation could also perturb the local molecular dynamics, which may influence ligand binding. Cygb presents the same features as Ngb; however, the sample with S–S bonds is not homogeneous. Only about one-half of the signal corresponds to a high dis- sociation rate for histidine and, consequently, one-half to a low dissociation rate, the slow rate being similar to that measured in the absence of the disulfide bridge (Fig. 4B). For both Ngb and Cygb, this transition is reversible: reduction of the protein by dithiothreitol leads to a single kinetic species, and the re-oxidation of cysteines (to re-form the disulfide bridge) again exhibits heterogeneous kinetics. This reduction after the addition of 2 mm dithiothre- itol takes a few minutes at 37 °C based on the CO binding kinetics, whereas S–S bridge formation under air is achieved after 7 h as measured by mass spectros- copy (Fig. 2). Therefore, a faster formation of the disulfide bond, necessary for a physiological role, would probably require the presence of a catalyst. A catalyst of disulfide bridge formation that acts with a coupling reaction between both redox centers, namely heme and cysteines, could also be considered. Whatever the putative partner responsible for cysteine oxidation, the reverse reduction is probably performed by glutathione. The highly concentrated glutathione redox couple (GSH ⁄ GSSG) is known to be displaced towards the reduced form in the cytosol, whereas, in the endoplasmic reticulum, most of the constitutive disulfide bridges are usually formed in secretory pro- teins. Therefore, the equilibrium for disulfide bridge formation in Cygb will be influenced by the redox state of the cytosol, which may also be transiently influ- enced by reactive oxygen species or other oxidative species derived from nitric oxide. It should be noted that intracellular compartments may provide favorable local environments in which the thiol redox state may be different from that of the bulk cytoplasm, but also that Cygb can be present elsewhere in the nucleus [6]. Conclusion In conclusion, human Cygb is monomeric in solution, as measured by MALLS, at micromolar concentrations of heme, the range expected for Cygb in vivo. The molar mass of Cygb samples measured by SEC- MALLS, mass spectrometry and flash photolysis con- firms the possibility of formation of an intramolecular disulfide bridge. The formation of an intermolecular disulfide bridge was observed in mass spectrometry, but represents only a small fraction of the Cygb prepa- ration and is negligible in the kinetic analysis. A con- formational transition in Cygb between an oxidized (intramolecular disulfide bond) form and a reduced (disulfide free) Cygb form is evident in the kinetics; such a transition affects the E-helix position, allowing a fine modulation of the endogenous His E7 affinity for heme binding. Disulfide bridge formation probably creates a significant stress on the E-helix of Cygb, resulting in a change in the distal histidine end-position of the liganded and unliganded states, as well as in its movement for heme binding. A constraint provided by the disulfide bridge creates two possible positions of distal histidine after photodissociation, leading to C. Lechauve et al. Cytoglobin conformations and disulfide bond formation FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works 2701 heterogeneity of kinetics. After breaking the S–S bond, the histidine relaxes towards a stable conformation, and kinetic curves show a clean biphasic form, as expected for the competition model of two ligands with one affinity for histidine. It can be proposed, as for Ngb [14], that, on oxidative stress, a change in the local redox state in response to physiological stimuli may induce the rupture and ⁄ or formation of the intramolec- ular B2–E9 disulfide bond in Cygb, thus initiating a conformational change that affects the overall oxygen affinity (Fig. 5) and other sensing functions as well. Materials and methods Expression and purification of recombinant Cygb The expression of wild-type human Cygb and the doubly mutated Cygb C38S ⁄ C83S mutant was performed as described previously [19]. The truncated CYGB differs from the wild-type by the removal of the amino-terminal residues 1–17 and the carboxyl-terminal residues 165–190, as described previously [14]. Inclusion bodies were solubilized in 6 m guanidinium hydrochloride and, after elimination of the insoluble material, Cygb was reconstructed by adding free hemin and dialyzed overnight. The samples were then purified with an Akta purifier system (GE Healthcare, Life- sciences, Uppsala, Sweden) on a Hitrap DEAE Sepharose column (GE Healthcare, Lifesciences, Uppsala, Sweden). The concentrated material was loaded onto a Superose 12 HR 16 ⁄ 50 column (GE Healthcare). The final purity of the pooled Cygb was checked by absorbance spectra and SDS- PAGE. Dynamic light scattering The particle size was measured with a Zetasizer Nano-ZS (Malvern Instruments), based on dynamic light scattering. Size distribution by volume was used for data interpreta- tion. Measurements were performed at 20 °C in 100 mm NaCl, 30 mm phosphate buffer at pH 7.5, in triplicate on each sample, and the average was taken for the diameter. Size exclusion by fast protein liquid chromatography and MALLS The Cygb shape and molar mass in solution were deter- mined using online SEC-MALLS. The gel filtration separa- tion was carried out using an EttanÔ LC liquid chromatography system (GE Healthcare) equipped with a SuperoseÔ 12 HR 10 ⁄ 300 GL column (GE Healthcare). Isocratic elution was performed at a flow rate of 0.39 mLÆmin )1 using a mobile phase of 30 mm NaCl ⁄ P i (pH 7.5), 150 mm NaCl and 0.03% sodium azide at 25 °C. It should be noted that NaCl was added to the elution buf- fer to avoid unwanted interactions between proteins and the solid phase. Light scattering analysis was performed using an EttanÔ LC HPLC system with automatic degasser and thermostatically controlled autosampler, connected inline to a DAWN Ò HELEOSÔ II 18-angle static light scattering detector, equipped with a QELS (quasi-elastic light scattering) instrument (Wyatt Technology) and an Optilab Ò rEX differential refractometer, equipped with a Peltier temperature-regulated flow cell maintained at 25 °C (Wyatt Technology). Calibration of the light scattering detector was subsequently verified using an albumin mono- mer standard (Sigma-Aldrich, Copenhagen, Denmark), recombinant Ngb, Mb (Sigma-Aldrich, Copenhagen, Den- mark), Mb dimer, diaspirine cross-linked (DCL) hemoglo- bin tetramer (Baxter Healthcare Corporation, Deerfield, IL, USA) and an octameric recombinant hemoglobin based on the natural variant hemoglobin Ta-Li [20]. The molar mass for the protein was calculated from the light scattering data using a specific refractive index increment (dn ⁄ dc) value of 0.183 mLÆg )1 . The light scattering on the different detectors was analyzed using astra v software (Wyatt Technology, version 5.3.4.13) to finally obtain the absolute molar mass and the hydrodynamic diameter for the eluted fractions. Sample preparation and ESI-MS Native samples were infused (5 LÆmin )1 ) into the mass spec- trometer (Quattro Ultima, Micromass Ltd., Wythenshawe, UK) at a concentration of approximately 5 lm in 1 : 1 ace- tonitrile–water containing 0.2% formic acid. Data were acquired over the mass-to-charge ratio (m ⁄ z) range 600– 2000 (5 min) and deconvoluted to present the spectra on a molar mass scale using the maximum entropy (maxent)- based software supplied with the spectrometer. Mass scale calibration employed the series of ions with multiple charges from separate introductions of Mb (sequence mass 16 951.5 Da). Spectra and ligand binding ⁄ dissociation kinetics Spectral measurements were made with a Varian Cary 400 (Varian, Inc., Palo Alto, CA, USA) or an HP 8453 diode array spectrophotometer (Hewlett Packard, Bracknell, UK). All ligand binding experiments were performed in 30 mm phosphate buffer at pH 7.5 with (10 mm) or without dithiothreitol to reduce the cysteine residues. The binding kinetics after heme ligand photolysis were performed using an Nd:YAG laser CFR-300 (Quantel, Les Ulis, France) generating 8 ns pulses of 160 mJ at 532 nm. The laser beam, as well as the monochromatic detection light, were brought to the sample cuvette by an optical fiber. The methods used to assess the hexacoordination and bimolecu- lar CO and O 2 rate constants have been described previously [17]. Samples from 1 to 10 lm on a heme basis were placed in 4mm·10 mm quartz cuvettes. We also used a SFM-3 Cytoglobin conformations and disulfide bond formation C. Lechauve et al. 2702 FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works stopped-flow rapid mixing equipment (Bio-logic SAS, Claix, France) to study ligand replacement. Experiments were repeated at least three times for each sample condition. Acknowledgements We thank Professors T. Burmester (University of Hamburg) and T. Hankeln (University of Mainz) for the Cygb plasmid, and Veronique Baudin-Creuza (Inserm U779) for the octameric recombinant hemo- globin sample. This work was supported by Inserm, DGA (De ´ le ´ gation Ge ´ ne ´ rale pour l’Armement) contract N° 07.34.004, and Universite ´ s Paris VI et XI. References 1 Burmester T, Weich B, Reinhardt S & Hankeln T (2000) A vertebrate globin expressed in the brain. Nat- ure 407, 520–523. 2 Kawada N, Kristensen DB, Asahina K, Nakatani K, Minamiyama Y, Seki S & Yoshizato K (2001) Charac- terization of a stellate cell activation-associated protein (STAP) with peroxidase activity found in rat hepatic stellate cells. J Biol Chem 276, 25318–25323. 3 Burmester T, Ebner B, Weich B & Hankeln T (2002) Cytoglobin: a novel globin type ubiquitously expressed in vertebrate tissues. Mol Biol Evol 19, 416–421. 4 Schmidt M, Gerlach F, Avivi A, Laufs T, Wystub S, Simpson JC, Nevo E, Saaler-Reinhardt S, Reuss S, Hankeln T et al. (2004) Cytoglobin is a respiratory protein in connective tissue and neurons, which is up-regulated by hypoxia. J Biol Chem 279, 8063– 8069. 5 Nakatani K, Okuyama H, Shimahara Y, Saeki S, Kim DH, Nakajima Y, Seki S, Kawada N & Yoshizato K (2004) Cytoglobin ⁄ STAP, its unique localization in splanchnic fibroblast-like cells and function in organ fibrogenesis. Lab Invest 84, 91–101. 6 Geuens E, Brouns I, Flamez D, Dewilde S, Timmer- mans JP & Moens L (2003) A globin in the nucleus! J Biol Chem 278, 30417–30420. 7 Fordel E, Thijs L, Martinet W, Lenjou M, Laufs T, Van Bockstaele D, Moens L & Dewilde S (2006) Neuroglobin and cytoglobin overexpression protects human SH-SY5Y neuroblastoma cells against oxida- tive stress-induced cell death. Neurosci Lett 410, 146– 151. 8 Hodges NJ, Innocent N, Dhanda S & Graham M (2008) Cellular protection from oxidative DNA damage by over-expression of the novel globin cytoglobin in vitro. Mutagenesis 4, 293–298. 9 Halligan KE, Jourd’heuil FL & Jourd’heuil D (2009) Cytoglobin is expressed in the vasculature and regulates cell respiration and proliferation via nitric oxide dioxy- genation. J Biol Chem 284, 8539–8547. 10 Hankeln T, Ebner B, Fuchs C, Gerlach F, Haberkamp M, Laufs TL, Roesner A, Schmidt M, Weich B, Wystub S et al. (2005) Neuroglobin and cytoglobin in search of their role in the vertebrate globin family. J Inorg Biochem 99, 110–119. 11 Shivapurkar N, Stastny V, Okumura N, Girard L, Xie Y, Prinsen C, Thunnissen FB, Wistuba II, Czerniak B, Frenkel E et al. (2008) Cytoglobin, the newest member of the globin family, functions as a tumor suppressor gene. Cancer Res 68, 7448–7456. 12 De Sanctis D, Dewilde S, Pesce A, Moens L, Ascenzi P, Hankel T, Burmester T & Bolognesi M (2004) Crystal structure of cytoglobin: the fourth globin type discov- ered in man displays heme hexa-coordination. J Mol Biol 336, 917–927. 13 Sugimoto H, Makino M, Sawai H, Kawada N, Yoshizato K & Shiro Y (2004) Structural basis of human cytoglobin for ligand binding. J Mol Biol 339, 873–885. 14 Hamdane D, Kiger L, Dewilde S, Green BN, Pesce A, Uzan J, Burmester T, Hankeln T, Bolognesi M, Moens L et al. (2003) The redox state of the cell regulates the ligand binding affinity of human neuroglobin and cyto- globin. J Biol Chem 278, 51713–51721. 15 Makino M, Sugimoto H, Sawai H, Kawada N, Yoshizato K & Shiro Y (2006) High-resolution struc- ture of human cytoglobin: identification of extra N- and C-termini and a new dimerization mode. Acta Crystal- logr D62, 671–677. 16 Trent JT III & Hargrove MS (2002) A ubiquitously expressed human hexacoordinate hemoglobin. J Biol Chem 277, 19538–19545. 17 Uzan J, Dewilde S, Burmester T, Hankeln T, Moens L, Hamdane D, Marden MC & Kiger L (2004) Neuroglo- bin and other hexacoordinated hemoglobins show a weak temperature dependence of oxygen binding. Biophys J 87, 1196–1204. 18 Smagghe BJ, Sarath G, Ross E, Hilbert JL & Hargrove MS (2006) Slow ligand binding kinetics dominate ferrous hexacoordinate hemoglobin reactivities and reveal differences between plants and other species. Biochemistry 45, 561–570. 19 Dewilde S, Mees K, Kiger L, Lechauve C, Marden MC, Pesce A, Bolognesi M & Moens L (2008) Expres- sion, purification, and crystallization of neuro- and cytoglobin. Methods Enzymol 436, 341–357. 20 Baudin-Creuza V, Chauvierre C, Domingues E, Kiger L, Leclerc L, Vasseur C, Ce ´ lier C & Marden MC (2008) Octamers and nanoparticles as hemoglobin based blood substitutes. Biochim Biophys Acta 1784, 1448–1453. Supporting information The following supplementary material is available: Fig. S1. Partition coefficients measured by gel filtration. C. Lechauve et al. Cytoglobin conformations and disulfide bond formation FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works 2703 Fig. S2. Absorbance spectra of Cygb. 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 or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Cytoglobin conformations and disulfide bond formation C. Lechauve et al. 2704 FEBS Journal 277 (2010) 2696–2704 Journal compilation ª 2010 FEBS. No claim to original US government works . arising from supporting information (other than missing files) should be addressed to the authors. Cytoglobin conformations and disulfide bond formation C. Lechauve. correlation line of other globin proteins. C. Lechauve et al. Cytoglobin conformations and disulfide bond formation FEBS Journal 277 (2010) 2696–2704 Journal compilation