Ibuprofen binding to secondary sites allosterically modulates the spectroscopic and catalytic properties of human serum heme–albumin Alessandra di Masi 1 , Francesca Gullotta 2,3 , Alessandro Bolli 1 , Gabriella Fanali 4 , Mauro Fasano 4 and Paolo Ascenzi 1 1 Department of Biology, and Interdepartmental Laboratory of Electron Microscopy, University Roma Tre, Italy 2 Department of Experimental Medicine and Biochemical Sciences, University of Rome ‘Tor Vergata’, Italy 3 Interuniversity Consortium for the Research on the Chemistry of Metals in Biological Systems, Bari, Italy 4 Department of Structural and Functional Biology, and Center of Neuroscience, University of Insubria, Busto Arsizio (VA), Italy Introduction Human serum albumin (HSA), the most abundant protein in plasma, provides a depot and carrier for many endogenous and exogenous compounds. Among other roles, HSA affects the pharmacokinetics of many drugs, holds some ligands in a strained orientation that results in their metabolic modification, renders poten- tial toxins harmless by transporting them to disposal sites, accounts for most of the antioxidant capacity of human serum, and displays (pseudo)enzymatic proper- ties [1–8]. HSA is a single nonglycosylated all-a-chain protein, composed of 585 amino acids, and containing three homologous domains (labeled I, II, and III). Each domain is composed of two separate subdo- mains (named A and B) connected by random coils. Interdomain helical regions link subdomain IB to Keywords allostery; human serum heme–albumin; ibuprofen binding; modulation of reactivity and spectroscopic properties; recombinant truncated human serum heme–albumin (Asp1–Glu382) Correspondence P. Ascenzi, Department of Biology, and Interdepartmental Laboratory of Electron Microscopy, University Roma Tre, Viale Guglielmo Marconi 446, I-00146 Rome, Italy Fax: +39 06 5733 6321 Tel: +39 06 5733 3200(2) E-mail: ascenzi@uniroma3.it (Received 14 July 2010, revised 24 November 2010, accepted 6 December 2010) doi:10.1111/j.1742-4658.2010.07986.x The ibuprofen primary binding site FA3–FA4 is located in domain III of human serum albumin (HSA), the secondary clefts FA2 and FA6 being sited in domains I and II. Here, the thermodynamics of ibuprofen binding to recombinant Asp1–Glu382 truncated HSA (tHSA)–heme-Fe(III) and nitrosylated tHSA–heme-Fe(II), encompassing domains I and II only, is reported. Moreover, the allosteric effect of ibuprofen on the kinetics of tHSA–heme-Fe(III)-mediated peroxynitrite isomerization and nitrosylated tHSA–heme-Fe(II) denitrosylation has been investigated. The present data indicate, for the first time, that the allosteric modulation of tHSA–heme and HSA–heme reactivity by ibuprofen depends mainly on drug binding to the FA2 and FA6 secondary sites rather than drug association with the FA3–FA4 primary cleft. Thus, tHSA is a valuable model with which to investigate the allosteric linkage between the heme cleft FA1 and the ligand-binding pockets FA2 and FA6, all located in domains I and II of (t)HSA. Abbreviations FA, fatty acid; HSA, human serum albumin; HSA–heme-Fe, human serum heme-albumin; HSA–heme-Fe(III), ferric HSA–heme-Fe; HSA–heme-Fe(II), ferrous HSA–heme-Fe; HSA–heme-Fe(II)-NO, nitrosylated HSA–heme-Fe(II); tHSA, truncated HSA; tHSA–heme-Fe(III), ferric heme tHSA; tHSA–heme-Fe(II), ferrous tHSA–heme-Fe; tHSA–heme-Fe(II)-NO, nitrosylated tHSA–heme-Fe(II). 654 FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS subdomain IIA, and subdomain IIB to subdo- main IIIA (Fig. 1) [8,9]. The structural organization of HSA provides several ligand-binding sites (Fig. 1). HSA has seven binding clefts hosting chemically diverse ligands, including fatty acids (FAs), that are labeled FA1–FA7 (Fig. 1). In particular, FA1 (located in subdomain IB) has evolved to specifically bind heme, FA3 and FA4 make up the so-called Sudlow site II (located in subdo- main IIIA), which preferentially recognizes aromatic carboxylates with an extended conformation, and FA7 is the so-called Sudlow site I (located in subdo- main IIA), which binds, in particular, bulky heterocy- clic anions. Remarkably, warfarin, a coumarinic anticoagulant drug, and ibuprofen, a nonsteroidal anti-inflammatory drug, are considered to be stereotyp- ical ligands for Sudlow sites I and II, respectively. In contrast to warfarin, ibuprofen has been reported to also bind to secondary sites in HSA domains I and II that have been characterized by X-ray crystallographic and solution spectroscopic studies [1,2,6,8–20]. Recently, the recombinant truncated form of HSA (tHSA), encompassing residues Asp1–Glu382 (corre- sponding to domains I and II, containing only the ibu- profen secondary binding sites), has been preliminarily characterized [19]. Here, ibuprofen binding to tHSA– heme-Fe(III) and tHSA–heme-Fe(II) is reported. Moreover, the allosteric effect of ibuprofen on tHSA– heme-Fe(III)-mediated peroxynitrite isomerization and denitrosylation of tHSA–heme-Fe(II)-NO has been investigated. The present data indicate, for the first time, that the allosteric modulation of tHSA–heme and HSA–heme reactivity by ibuprofen depends mainly on drug binding to the secondary sites FA2 and FA6 rather than drug association with the FA3– FA4 primary cleft. Results Ibuprofen binding to tHSA–heme-Fe(III) and HSA–heme-Fe(III)] Figure 2A shows the binding isotherm for ibuprofen binding to tHSA–heme-Fe(III) and HSA–heme-Fe(III). FA1 FA2 FA7 FA6 Glu382 FA3-FA4 FA5 Fig. 1. HSA structure. The six subdomains of HSA are colored as follows: IA, blue; IB, cyan; IIA, dark green; IIB, light green; IIIA, orange; IIIB, red. The heme (red) fits the primary cleft in subdo- main IB, corresponding to FA1. Sudlow site I (in subdomain IIA, corresponding to FA7) is occupied by warfarin (purple). Glu382 is highlighted. Sudlow site II (in subdomain IIIA, corresponding to FA3–FA4) and FA6 (in subdomain IIB) are occupied by ibuprofen (magenta). Sites FA2 (at the subdomain I–IIA interface) and FA5 (in subdomain IIIB) are occupied by myristate (orange). Atomic coordi- nates were taken from Protein Data Bank entries 1O9X [40], 2BXD, and 2BXG [6]. For details, see text. Fig. 2. Ibuprofen binding to tHSA–heme-Fe(III) and HSA–heme Fe(III). (A) Thermodynamics of ibuprofen binding to tHSA–heme- Fe(III). The continuous line was calculated according to Eqn (1) by non-linear regression curve fitting with the following parameters: K 2 = (1.7 ± 0.2) · 10 )5 M; K 3 = (8.9 ± 0.9) · 10 )4 M; a = 0.18 ± 0.03; and 1 – a = 0.82 ± 0.04. [tHSA–heme-Fe(III)] = 1.9 · 10 )6 M. (B) Thermodynamics of ibuprofen binding to HSA–heme-Fe(III). The continuous line was calculated according to Eqn (1) by non-linear regression curve fitting with the following parameters: K 2 = (6.9 ± 0.7) · 10 )6 M; K 3 = (8.1 ± 0.7) · 10 )4 M; a = 0.14 ± 0.03; and 1 ) a = 0.86 ± 0.04. [HSA–heme-Fe(III)] = 2.2 · 10 )6 M. The ibupro- fen concentration corresponds to that of the free ligand. Where not shown, the standard deviation is smaller than the symbol. All data were obtained at pH 7.0 and 20.0 °C. For further details, see text. A. di Masi et al. Ibuprofen binding to truncated HSA secondary sites FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS 655 The analysis of the data given in Fig. 2A,B, according to Eqn (1), allowed the determination of K 2 and K 3 values for ibuprofen binding to tHSA–heme-Fe(III) (1.7 · 10 )5 and 8.9 · 10 )4 m, respectively) and to HSA–heme-Fe(III) (6.9 · 10 )6 and 8.1 · 10 )4 m, respectively). The spectroscopic contributions of ibu- profen binding to the high-affinity and low-affinity sites of tHSA–heme-Fe(III) and HSA–heme-Fe(III) [represented by a and (1 – a), respectively, in Eqn (1)] are 0.18 and 0.82, and 0.14 and 0.86, respectively (Table 1). Effect of ibuprofen on peroxynitrite isomerization by tHSA–heme-Fe(III) In the absence and presence of tHSA–heme-Fe(III), the kinetic data of peroxynitrite isomerization were fit- ted to a single-exponential decay for more than 95% of their course (Eqn 2). According to the literature [21], this indicates that the formation of the transient tHSA–heme-Fe(III)-OONO species represents the rate-limiting step in catalysis, the conversion of the tHSA–heme-Fe(III)-OONO complex to tHSA–heme- Fe(III) and NO À 3 being faster by at least one order of magnitude. In the absence and presence of ibuprofen, the observed rate constants for tHSA–heme-Fe(III)-cata- lyzed isomerization of peroxynitrite (l obs ) increased linearly with the tHSA–heme-Fe(III) concentration (Fig. 3A). The analysis of data reported in Fig. 3A, according to Eqn (3), allowed the determination of val- ues of the second-order rate constant for peroxynitrite isomerization by tHSA–heme-Fe(III) (l on = 4.3 · 10 5 m )1 Æs )1 , corresponding to the slope of the linear plots) and of the first-order rate constant for peroxynitrite isomerization in the absence of tHSA–heme-Fe(III) (l 0 = 2.4 · 10 )1 s )1 , corresponding to the y-intercept of the linear plots). Ibuprofen dose-dependently impairs tHSA–heme- Fe(III)-mediated isomerization of peroxynitrite (Fig. 3A,B). Indeed, values of l on for tHSA–heme- Fe(III)-catalyzed isomerization of peroxynitrite decreased from 4.3 · 10 5 m )1 Æs )1 in the absence of ibu- profen [l on(top) ] to 5.8 · 10 4 m )1 Æs )1 at [ibupro- fen] = 1.0 · 10 )2 m (Fig. 3A,B). On the other hand, values of l 0 were unaffected by ibuprofen, being depen- dent only on CO 2 (Fig. 3A,C). The values of l on and l 0 for tHSA–heme-Fe(III)-catalyzed isomerization of peroxynitrite determined here are very similar to those previously reported for HSA–heme-Fe(III)-mediated peroxynitrite isomerization [21]. The analysis of the dependence of l on on ibuprofen concentration for tHSA–heme-Fe(III)-catalyzed isom- erization of peroxynitrite (Fig. 3B), according to Eqn (4), allowed determination of the value of the dissociation equilibrium constant for ibuprofen binding to tHSA–heme-Fe(III) (K 3 = 9.3 · 10 )4 m). Under conditions where [ibuprofen] >> K 3 , tHSA–heme- Fe(III) did not catalyze the isomerization of peroxy- nitrite. Indeed, the value of l obs (2.6 · 10 )1 s )1 ) was independent of the tHSA–heme-Fe(III) concentration, and corresponded to that obtained in the absence of tHSA–heme-Fe(III) (l 0 = 2.4 · 10 )1 s )1 ). Ibuprofen binding to tHSA–heme-Fe(II)-NO Figure 4 shows the binding isotherm for ibuprofen binding to tHSA–heme-Fe(II)-NO. Analysis of the dependence of the molar fraction of the ibuprofen- bound tHSA–heme-Fe(II)-NO (Y) on the ibuprofen concentration (Fig. 4), according to Eqn (5), allowed to determine the value of the dissociation equilibrium Table 1. Values of the dissociation equilibrium constants for ibuprofen binding to tHSA–heme and HSA–heme derivatives obtained by differ- ent experimental methods. For consistency with previous studies [20,21], equilibrium dissociation constants and Hill coefficients are indi- cated as K 1 , K 2 , and K 3 , and as n 1 , n 2 , and n 3 , respectively. Protein Experimental method K 1 (M) n 1 K 2 (M) n 2 K 3 (M) n 3 tHSA–heme-Fe(III) a Absorbance spectroscopy – b – b 1.7 · 10 –5 0.98 8.9 · 10 –4 1.01 tHSA–heme-Fe(III) a Peroxynitrite isomerization – b – b – b – b 9.3 · 10 –4 0.99 tHSA–heme-Fe(II)-NO a Absorbance spectroscopy – b – b – b – b 2.1 · 10 –3 0.99 tHSA–heme-Fe(II)-NO a Denitrosylation kinetics – b – b 1.3 · 10 –4 0.99 2.5 · 10 –3 1.01 HSA–heme-Fe(III) a Absorbance spectroscopy – b – b 6.9 · 10 –6 1.00 8.1 · 10 –4 0.99 HSA–heme-Fe(III) c Peroxynitrite isomerization – b – b – b – b 9.7 · 10 –4 – d HSA–heme-Fe(II)-NO e Absorbance spectroscopy – b – b – b – b 2.6 · 10 –3 – d HSA–heme-Fe(II)-NO e Denitrosylation kinetics 3.1 · 10 –7 – d 1.7 · 10 –4 – d 2.2 · 10 –3 – d a pH 7.0 and 20.0 °C. Present study. b Not applicable; see text. c pH 7.2 and 22.0 °C. From [21]. d Although values of n 1 , n 2 and n 3 have not been reported [20,21], the data fit to simple equilibria, implying n 1 =1,n 2 = 1, and n 3 =1. e pH 7.0 and 10.0 °C. From [20]. Ibuprofen binding to truncated HSA secondary sites A. di Masi et al. 656 FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS constant for drug binding to tHSA–heme-Fe(II)-NO (K 3 = 2.1 · 10 )3 m) (Fig. 4 and Table 1). Effect of ibuprofen on tHSA–heme-Fe(II)-NO denitrosylation In the absence and presence of ibuprofen, the time course for NO dissociation from tHSA–heme-Fe(II)- NO conformed to a single-exponential decay for more than 95% of its course (Eqn 6). According to the literature [20], this indicates that the formation of the transient tHSA–heme-Fe(II) species represents the rate-limiting step in tHSA–heme-Fe(II)-CO formation. Values of the first-order rate constant for NO dissocia- tion from tHSA–heme-Fe(II)-NO (k off ) were indepen- dent of wavelength and [CO] in the presence of excess dithionite. Values of k off for tHSA–heme-Fe(II)-NO denitrosy- lation increased from 1.5 · 10 )4 s )1 in the absence of ibuprofen (k þ off in Eqn 7) to 8.5 · 10 )3 s )1 in the pres- ence of 1.0 · 10 )2 m ibuprofen (Fig. 5). The k off value for tHSA–heme-Fe(II)-NO denitrosylation determined here in the absence of ibuprofen is very similar to that previously obtained for HSA–heme-Fe(II)-NO denitro- sylation [20,22]. Analysis of the dependence of k off for tHSA–heme- Fe(II)-NO denitrosylation on the ibuprofen concentra- tion (Fig. 5), according to Eqn (7), allowed us to determine the values of the dissociation equilibrium constants for drug binding to tHSA–heme-Fe(II)-NO (K 2 = 1.3 · 10 )4 m and K 3 = 2.5 · 10 )3 m) (Fig. 5). The values of k off(2) and k off(3) for tHSA–heme-Fe(II)- NO denitrosylation determined here (1.1 · 10 )3 and 9.1 · 10 )3 s )1 , respectively) are very similar to those previously obtained for HSA–heme-Fe(II)-NO denitro- sylation [20]. Fig. 4. Ibuprofen binding to tHSA–heme-Fe(II)-NO. The continuous line was calculated according to Eqn (5) with K 3 = (2.1 ± 0.3) · 10 )3 M). The tHSA–heme-Fe(II)-NO concentration was 4.8 · 10 )6 M. Where not shown, the standard deviation is smaller than the symbol. All data were obtained at pH 7.0 (1.0 · 10 )1 M phosphate buffer) and 20.0 °C. For further details, see text. Fig. 3. Ibuprofen inhibits peroxynitrite isomerization by tHSA– heme-Fe(III). (A) Dependence of the pseudo-first-order rate constant for peroxynitrite isomerization (l obs ) on the tHSA–heme- Fe(III) concentration. The peroxynitrite concentration was 2.5 · 10 )4 M. The ibuprofen concentrations were 0.0 M (trace a), 1.5 · 10 )3 M (trace b), and 7.5 · 10 )3 M (trace c). The continuous lines in (A) were calculated according to Eqn (3) with the following parameters: trace a, l on = (4.3 ± 0.4) · 10 5 M )1 Æs )1 and l 0 = (2.4 ± 0.3) · 10 )1 s )1 ; trace b, l on = (1.8 ± 0.2) · 10 5 M )1 Æs )1 and l 0 = (2.8 ± 0.3) · 10 )1 s )1 ; and trace c, l on = (6.8 ± 0.4) · 10 4 M )1 Æs )1 and l 0 = (2.7 ± 0.3) · 10 )1 s )1 . (B) Effect of ibuprofen concentration on the second-order rate constant for tHSA–heme-Fe(III)-catalyzed isomerization of peroxynitrite (l on ). The filled symbol on the ordinate indicates the l on value obtained in the absence of ibuprofen [l on(- top) = (4.3 ± 0.4) · 10 5 M )1 Æs )1 ]. The continuous line was calculated according to Eqn (4) with l on(top) = (4.3 ± 0.4) · 10 5 M )1 Æs )1 and K 3 = (9.3 ± 1.0) · 10 )4 M. (C) Effect of ibuprofen concentration on the first-order rate constant for tHSA–heme-Fe(III)-catalyzed isomer- ization of peroxynitrite (l 0 ). The filled symbol on the ordinate indi- cates the l 0 value obtained in the absence of ibuprofen [(2.4 ± 0.3) · 10 )1 s )1 ]. Values of l 0 are independent of ibuprofen concentration; the average l 0 value is (2.6 ± 0.3) · 10 )1 s )1 . Where not shown, the standard deviation is smaller than the symbol. All data were obtained at pH 7.0 (1.0 · 10 )1 M phosphate buffer) and 20.0 °C. For further details, see text. A. di Masi et al. Ibuprofen binding to truncated HSA secondary sites FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS 657 Discussion The present results highlight the rol e of ibuprofen second- ary sites in modulating HSA–heme spectroscopic p roper- ties and reactivity. Indeed, values of the diss ociation equilibrium constants for ibuprofen b inding to secon dary sites present in tHSA–heme (present study) are in excel- lent agreement with those for drug binding to full- length HSA–heme (Table 1). However, the abse nce of domain III in tH SA–heme precludes the binding of ibuprofen to its primary clef t [19–21] (Table 1). The analysis of thermodynamic parameters reported in Table 1 allows the following conclusions to be drawn. l Ibuprofen binds to two secondary sites of tHSA– heme-Fe(III) and HSA–heme-Fe(III) (present study), affecting the absorption spectra. However, only ibupro- fen binding to the (t)HSA–heme-Fe(III) secondary site showing the lowest drug affinity allosterically modulates peroxynitrite isomerization (present study and [21]). l Ibuprofen affects the absorption spectra of tHSA– heme-Fe(II)-NO (present study) and HSA–heme- Fe(II)-NO [20] by binding to only one secondary site. However, the allosteric modulation of tHSA–heme- Fe(II)-NO denitrosylation (present study) reflects ibu- profen binding to two secondary sites. By contrast, HSA–heme-Fe(II)-NO denitrosylation is allosterically modulated by ibuprofen binding not only to both secondary sites located in domains I and II, but also to the primary cleft sited in domain III [20]. l Values of K 2 and K 3 for ibuprofen binding to tHSA–heme-Fe(III) and HSA–heme-Fe(III) (present study) are lower than those for drug binding to tHSA– heme-Fe(II)-NO (present study) and HSA–heme- Fe(II)-NO [20]. This indicates that the redox and the (un)ligated state of the heme Fe atom allosterically affects ibuprofen binding to (t) HSA–heme. l Values of K 2 and K 3 for ibuprofen binding to tHSA–heme-Fe (present study) are in excellent agree- ment with those reported for drug binding to HSA– heme species (present study and [20,21]). Also, the kinetic parameters for ibuprofen-mediated peroxyni- trite isomerization by (t)HSA–heme-Fe(III) and for (t)HSA–heme-Fe(II)-NO denitrosylation are very simi- lar (present study and [20,21]). This indicates that the removal of domain I from HSA does not significantly affect the functional and structural properties of domains I and II [i.e. of the 1–382 region of (t)HSA]. l The values of the Hill coefficients (n 1 , n 2 , and n 3 ) for ibuprofen binding to tHSA–heme and HSA–heme derivatives range between 0.98 and 1.01 (present study Table 1), indicating that drug association with (t)HSA– heme is a non-cooperative event, under all the experi- mental conditions. l The ibuprofen-dependent (t)HSA–heme-Fe(III)- mediated peroxynitrite isomerization and (t)HSA– heme-Fe(II)-NO denitrosylation reflect drug-dependent structural changes occurring at the heme-binding pocket (FA1). Indeed, ibuprofen binding to HSA–heme has been reported to induce the hexa-coordination of the heme Fe atom [18,21,23]. Thus, peroxynitrite can- not bind to the heme Fe(III) atom of hexa-coordinated HSA–heme-Fe(III), and therefore cannot undergo facil- itated isomerization [21]. Moreover, the increase in k off for NO dissociation from HSA–heme-Fe(II)-NO upon stabilization of the hexa-coordinated heme-Fe(II)-NO atom is reminiscent of what has been reported for abacavir-induced and warfarin-induced hexa- coordination of HSA–heme-Fe(II)-NO [22], and for 1-methyl-imidazole-mediated hexa-coordination of the heme-Fe(II)-NO model compound [24]. The ibuprofen primary binding cleft (i.e. Sudlow site II, formed by the FA3 and FA4 sites) and the drug secondary pocket (i.e. the FA6 region) have been substantiated by X-ray crystallography [6]. The third low-affinity ibuprofen-binding site has been tentatively identified with the FA2 site. Indeed, it acts as the modulatory site that controls the FA-induced conformational switch [18,21,25,26]. Ibuprofen binding to the third site induces the stabilization of the hexa-coordinate derivative of the HSA–heme-Fe(III) and HSA–heme-Fe(II)-NO species, which are instead predominantly penta-coordinated in the absence of allosteric effectors. Indeed, high (> 1.0 · 10 )3 m) ibu- profen concentration clearly induces the coordination of a histidine nitrogen as the sixth coordination posi- Fig. 5. Effect of ibuprofen on tHSA–heme-Fe(II)-NO denitrosylation. The continuous line was calculated according to Eqn (7) with k off(2) = (1.1 ± 0.1) · 10 )3 s )1 , K 2 = (1.3 ± 0.2) · 10 )4 M, k off(3) = (9.1 ± 1.0) · 10 )3 s )1 , K 3 = (2.5 ± 0.3) · 10 )3 M), and k þ off = (1.5 ± 0.2) · 10 )4 s )1 . The tHSA–heme-Fe(II)-NO concentration was 2.9 · 10 )6 M. Where not shown, the standard deviation is smaller than the symbol. All data were obtained at pH 7.0 (1.0 · 10 )1 M phosphate buffer) and 20.0 °C. For further details, see text. Ibuprofen binding to truncated HSA secondary sites A. di Masi et al. 658 FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS tion of the heme Fe, as the result of the conforma- tional transition following ligand binding to FA2 [18]. The present considerations appear to apply only to ibuprofen, as other Sudlow site II ligands, such as FAs, diazepam, and diflunisal, display different binding properties. FAs bind to multiple sites with different affinities and functional effects. In particular, the FA6 (corresponding to the highest-affinity site for ibuprofen in tHSA) and the FA7 (also named Sudlow site I and corresponding to the warfarin-binding pocket) sites show low affinity for FAs, whereas the FA1 (also named the heme site) and the FA2 (corresponding to the regulatory site of the neutral-to-basic conformational transition) pockets display high affinity for FAs [17]. Therefore, the effect of FAs is opposite to that of ibuprofen [8]. Moreover, FA binding to the FA1 site induces heme dissociation from HSA–heme, thus impairing HSA–heme reactivity [25]. Diazepam has been reported to bind only to Sudlow site II [17], so it does not bind to tHSA, which is deprived of domain III (Fig. 1). Finally, the FA6 site is also the secondary bind- ing cleft for diflunisal, another Sudlow site II ligand. However, diflunisal binds not only to the FA6 site (corresponding to the highest-affinity site for ibuprofen in tHSA), but also to the FA7 cleft (corresponding to the warfarin pocket) [17], thus inducing mixed effects, combining the actions of ibuprofen and warfarin [8]. It is worth noting that the modular architecture of HSA allows the removal of domain III without this affecting the conformational stability of the remaining protein scaffold. Moreover, lost contacts between par- alogous HSA domains do not impair the correct fold- ing of single domains, as shown by the agreement between thermodynamic and kinetic parameters in HSA and tHSA derivatives. As a whole, the present data: (a) demonstrate unequivocally, for the first time, that ibuprofen allosterically modulates the spectro- scopic and reactivity properties of tHSA–heme and HSA–heme by binding to the low-affinity secondary sites FA2 and FA6 rather than associating with the primary-high affinity cleft FA3–FA4; and (b) reinforce the idea that HSA could be taken as the prototype of monomeric allosteric proteins [8,19,27,28]. Finally, tHSA is a valuable model with which to investigate the allosteric linkage between ligand-binding pockets located in domains I and II of HSA. Experimental procedures Chemicals HSA (‡ 96%, essentially FA-free), hemin [Fe(III)-proto- porphyrin IX] chloride and ibuprofen were purchased from Sigma-Aldrich (St Louis, MO, USA). Recombinant tHSA was expressed and purified as previously reported [19]. NO (Aldrich Chemical Co., Milwaukee, WI, USA) was purified by flowing it through an NaOH column in order to remove acidic nitrogen oxides. CO was purchased from Linde AG (Ho ¨ llriegelskreuth, Germany). tHSA–heme-Fe(III) and HSA–heme-Fe(III) were pre- pared by adding a 0.8 mol heme-Fe(III) per mol tHSA and HSA (1.0 · 10 )1 m sodium phosphate buffer, pH 7.2) at 20.0 °C [19–21]. The final tHSA–heme-Fe(III) and HSA– heme-Fe(III) concentrations ranged between 1.9 · 10 )6 and 5.0 · 10 )5 m, and between 2.2 · 10 )6 and 5.0 · 10 )5 m, respectively. tHSA–heme-Fe(II)-NO (final concentration, 4.8 · 10 )6 m) was obtained, under anaerobic conditions, by blowing purified NO over the ferrous heme–protein solu- tion (1.0 · 10 )1 m sodium phosphate buffer, pH 7.0) at 10.0 °C [20]. The ibuprofen stock solution (1.0 · 10 )1 m) was prepared by dissolving the drug in 1.0 · 10 )1 m phosphate buffer (pH 7.0) at 20.0 °C [23]. The final ibuprofen concentration ranged between 1.0 · 10 )6 and 1.0 · 10 )2 m. Peroxynitrite was synthesized from KO 2 and NO or from HNO 2 and H 2 O 2 , and stored in small aliquots at )80.0 °C [29,30]. The peroxynitrite stock solution (2.0 · 10 )3 m) was diluted immediately before use with degassed 5.0 · 10 )2 m NaOH to achieve the desired concentration [21,31–35]. Nitrate and nitrite contamination levels were in the ranges of 0–7% and 8–19% of the peroxynitrite concentration, respectively [21]. The concentration of peroxynitrite was determined spectrophotometrically prior to each experiment by measuring the absorbance at 302 nm (e 302 nm = 1.705 · 10 3 m )1 Æcm )1 ) [29,30]. The CO solution was prepared by keeping the 1.0 · 10 )1 m phosphate buffer solution (pH 7.0) in a closed vessel with CO at a pressure of 760.0 mmHg, under anaerobic conditions, at 20.0 °C [20,36]. All of the other chemicals were obtained from Sigma- Aldrich and Merck AG (Darmstadt, Germany). All prod- ucts were of analytical or reagent grade, and were used without further purification. Ibuprofen binding to tHSA–heme-Fe(III) and HSA–heme-Fe(III) Values of the dissociation equilibrium constants for ibupro- fen binding to tHSA–heme-Fe(III) and HSA–heme-Fe(III) (i.e. K 2 and K 3 ) were obtained spectrophotometrically, at pH 7.0 (1.0 · 10 )1 m phosphate buffer) and 20.0 °C. Ibuprofen-dependent absorbance changes were recorded between 350 and 450 nm. Small aliquots of the ibuprofen (1.0 · 10 )1 m) stock solution were added to the tHSA– heme-Fe(III) (1.9 · 10 )6 m) and HSA–heme-Fe(III) (2.2 · 10 )6 m) solutions, and the ibuprofen-dependent absorbance changes of tHSA–heme-Fe(III) and HSA– A. di Masi et al. Ibuprofen binding to truncated HSA secondary sites FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS 659 heme-Fe(III) were recorded after incubation for 10 min, after each addition [19,21]. Test measurements performed after 2 h excluded slow kinetic events. Ibuprofen binding to tHSA–heme-Fe(III) and HSA– heme-Fe(III) was analyzed by plotting the molar fraction of drug–tHSA–heme-Fe(III) and drug–HSA–heme-Fe(III) complexes (Y) as a function of the free ibuprofen concen- tration (ranging between 1.0 · 10 )6 m and 1.0 · 10 )2 m). Data were analyzed according to Eqn (1) [36]: Y ¼½a Âf½ibuprofen=ðK 2 þ½ibuprofenÞg þ½ð1 À aÞÂf½ibuprofen=ðK 3 þ½ibuprofenÞg ð1Þ where a and 1 ) a are the relative spectroscopic contri- butions to the total electronic absorbance change of ibu- profen binding to the high-affinity and low-affinity sites, respectively. Effect of ibuprofen on peroxynitrite isomerization by tHSA–heme-Fe(III) Kinetic data for peroxynitrite isomerization in the absence and presence of tHSA–heme-Fe(III) and ibuprofen were recorded with the SMF-100 rapid-mixing stopped-flow apparatus (Bio-Logic SAS, Claix, France). The light path of the observation cuvette was 10 mm, and the dead time was 1.4 ms. The kinetics were monitored at 302 nm, the characteristic absorbance maximum of peroxynitrite (e 302 nm = 1.705 · 10 3 m )1 Æcm )1 ) [29,30,35,37]. Kinetic data were obtained in the absence and presence of tHSA–heme- Fe(III) (final concentration, 5.0 · 10 )6 to 5.0 · 10 )5 m) and ibuprofen (final concentration, 1.0 · 10 )6 to 1.0 · 10 )2 m), by rapid mixing of the protein-buffered solution with the peroxynitrite solution (final concentration, 2.5 · 10 )4 m). Kinetic data were obtained at pH 7.0 (1.0 · 10 )1 m phosphate buffer) and 20.0 °C; no gaseous phase was present [21]. The kinetics of peroxynitrite isomerization by tHSA– heme-Fe(III), in the absence and presence of ibuprofen, were analyzed in the framework of the following minimum reaction scheme [21]: tHSAÀhemeÀFeðIIIÞþHOONO ! l on tHSAÀhemeÀFeðIIIÞÀ OONO þ H þ tHSAÀhemeÀFeðIIIÞÀOONO ! fast tHSAÀhemeÀFeðIIIÞþNO 3 À Values of the pseudo-first-order rate constant for tHSA– heme-Fe(III)-mediated peroxynitrite isomerization (l obs ) were determined, in the absence and presence of ibuprofen, at pH 7.0 (1.0 · 10 )1 m phosphate buffer) and 20.0 °C, from the analysis of the time-dependent absorbance decrease at 302 nm, according to Eqn (2) [21]: ½peroxynitrite t ¼½peroxynitrite i  e Àl obs Ât ð2Þ Values of the second-order rate constant for tHSA– heme-Fe(III)-mediated peroxynitrite isomerization (l on ) and of the first-order rate constant for peroxynitrite isomeriza- tion in the absence of tHSA–heme-Fe(III) (l 0 ) were deter- mined, in the absence and presence of ibuprofen, at pH 7.0 and 20.0 °C, from the linear dependence of l obs on the tHSA–heme-Fe(III) concentration, according to Eqn (3) [21,32,38,39]: l obs ¼ l on ½tHSAÀhemeÀFeðIIIÞ þ l 0 ð3Þ The value of the dissociation equilibrium constant for ibuprofen binding to tHSA–heme-Fe(III) (K 3 ) was deter- mined, at pH 7.0 (1.0 · 10 )1 m phosphate buffer) and 20.0 °C, from the dependence of l on on the free drug con- centration (ranging between 1.0 · 10 )6 and 1.0 · 10 )2 m). The effect of the drug concentration on l on was analyzed according to Eqn (4) [21,32,38,39]: l on ¼ l onðtopÞ Àfðl onðtopÞ Â½ibuprofenÞ = ðK 3 þ½ibuprofenÞg ð4Þ where l on(top) represents the asymptotic value of l on under conditions where [ibuprofen] = 0 (i.e. l on(top) = l on ). Ibuprofen binding to tHSA–heme-Fe(II)-NO The value of the dissociation equilibrium constant for ibu- profen binding to tHSA–heme-Fe(II)-NO (K 3 ) was deter- mined spectrophotometrically, at pH 7.0 (1.0 · 10 )1 m phosphate buffer) and 20.0 °C. Ibuprofen-dependent absor- bance changes were recorded between 350 and 450 nm. Small aliquots of the ibuprofen (1.0 · 10 )1 m) stock solution were added to the tHSA–heme-Fe(II)-NO (4.8 · 10 )6 m) solution, and the ibuprofen-dependent absorbance changes of tHSA– heme-Fe(II)-NO were recorded after incubation for 10 min, after each addition [20]. Test measurements performed after 2 h excluded slow kinetic events. Ibuprofen binding to tHSA–heme-Fe(III) was analyzed by plotting the molar fraction of the drug–tHSA–heme-Fe(II)- NO complex (Y) as a function of the free ibuprofen concen- tration (ranging between 1.0 · 10 )6 and 1.0 · 10 )2 m). Data were analyzed according to Eqn (5) [20]: Y ¼½ibuprofen=ðK 3 þ½ibuprofenÞ ð5Þ Effect of ibuprofen on tHSA–heme-Fe(II)-NO denitrosylation Values of the first-order rate constant for NO dissociation from tHSA–heme-Fe(II)-NO (i.e. for NO replacement by CO; k off ) were obtained by mixing the tHSA–heme-Fe(II)- NO (final concentration, 2.9 · 10 )6 m) solution with the CO (final concentration, 1.0 · 10 )4 to 5.0 · 10 )4 m) solution in the presence of dithionite (final concentration, 1.0 · 10 )2 m), under anaerobic conditions, at pH 7.0 (1.0 · 10 )1 m sodium Ibuprofen binding to truncated HSA secondary sites A. di Masi et al. 660 FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS phosphate buffer) and 20.0 °C [20], in the absence and pres- ence of ibuprofen (final concentration, 1.0 · 10 )7 to 1.0 · 10 )2 m). The excess of NO was pumped off gently before recording of ligand dissociation kinetics [20]. The kinetics were monitored between 350 and 460 nm. Absorbance electronic spectra were collected every 30 s. The time course for tHSA–heme-Fe(II)-NO denitrosylation was fitted to a single-exponential process according to the minimum reaction mechanism represented by the following scheme [20]: tHSAÀhemeÀFeðIIÞÀNO þ CO ! k off tHSAÀhemeÀFeðIIÞÀCO þ NO Values of k off were determined from data analysis according to Eqn (6) [20]: ½HSAÀhemeÀFe ðIIÞÀNO t ¼½HSAÀhemeÀFeðIIÞÀNO i  e Àk off Ât ð6Þ Values of the dissociation equilibrium constants for ibu- profen binding to tHSA–heme-Fe(II)-NO (K 2 and K 3 ) were obtained from the dependence of k off on the free ibuprofen concentration. Values of K 2 and K 3 were determined from data analysis, according to Eqn (7) [20]: k off ¼ðk offð2Þ Âð½ibuprofen=ðK 2 ½ibuprofenÞ þðk offð3Þ Âð½ibuprofen=ðK 3 ½ibuprofenÞÞ þ k þ off ð7Þ where k off(2) and k off(3) indicate values of k off occurring at K 2 < [ibuprofen] < K 3 , and at K 2 < K 3 < [ibuprofen], respectively, and k þ off is the k off value obtained in the absence of ibuprofen. Data analysis Kinetic and thermodynamic data were analyzed with the matlab program (The Math Works, Natick, MA, USA). The results are given as mean values of at least four experi- ments plus or minus the corresponding standard deviation. Acknowledgements This work was partially supported by grants from the Ministero dell’Istruzione, dell’Universita ` e della Ricerca of Italy (PRIN 2007ECX29E_002 and Univer- sity Roma Tre, CLAR 2009, to P. Ascenzi). References 1 Sudlow G, Birkett DJ & Wade DN (1975) The charac- terization of two specific drug binding sites on human serum albumin. Mol Pharmacol 11, 824–832. 2 Peters T Jr, ed. (1996) All about Albumin: Biochemistry, Genetics and Medical Applications. Academic Press, San Diego and London. 3 Curry S (2002) Beyond expansion: structural studies on the transport roles of human serum albumin. Vox Sang 83(Suppl), 1. 4 Kragh-Hansen U, Chuang VT & Otagiri M (2002) Practical aspects of the ligand-binding and enzymatic properties of human serum albumin. Biol Pharm Bull 25, 695–704. 5 Fasano M, Curry S, Terreno E, Galliano M, Fanali G, Narciso P, Notari S & Ascenzi P (2005) The extraordi- nary ligand binding properties of human serum albumin. IUBMB Life 57, 787–796. 6 Ghuman J, Zunszain PA, Petitpas I, Bhattacharya AA, Otagiri M & Curry S (2005) Structural basis of the drug-binding specificity of human serum albumin. J Mol Biol 353, 38–52. 7 Ascenzi P, Bocedi A, Notari S, Fanali G, Fesce R & Fasano M (2006) Allosteric modulation of drug binding to human serum albumin. Mini Rev Med Chem 6, 483– 489. 8 Ascenzi P & Fasano M (2010) Allostery in a monomeric protein: the case of human serum heme-albumin. Biophys Chem 148, 16–22. 9 Curry S (2009) Lessons from the crystallographic analysis of small molecule binding to human serum albumin. Drug Metab Pharmacokinet 24, 342– 357. 10 Curry S, Mandelkov H, Brick P & Franks N (1998) Crystal structure of human serum albumin complexed with fatty acid reveals an asymmetric distribution of binding sites. Nat Struct Biol 5, 827–835. 11 Bhattacharya AA, Curry S & Franks NP (2000) Binding of the general anesthetics propofol and halothane to human serum albumin: high resolution crystal structures. J Biol Chem 275, 38731–38738. 12 Bhattacharya AA, Gru ¨ ne T & Curry S (2000) Crystallo- graphic analysis reveals common modes of binding of medium and long-chain fatty acids to human serum albumin. J Mol Biol 303, 721–732. 13 Dockal M, Chang M, Carter DC & Ru ¨ ker F (2000) Five recombinant fragments of human serum albumin: tools for the characterization of the warfarin binding site. Protein Sci 9, 1455–1465. 14 Petitpas I, Bhattacharya AA, Twine S, East M & Curry S (2001) Crystal structure analysis of warfarin binding to human serum albumin: anatomy of drug site I. J Biol Chem 276, 22804–22809. 15 Chuang VTG & Otagiri M (2002) How do fatty acids cause allosteric binding of drugs to human serum albu- min? Pharm Res 19, 1458–1464. 16 Hamilton JA (2004) Fatty acid interactions with pro- teins: what X-ray crystal and NMR solution structures tell us. Prog Lipid Res 43, 177–199. 17 Simard JR, Zunszain PA, Hamilton JA & Curry S (2006) Location of high and low affinity fatty acid binding sites on human serum albumin revealed by A. di Masi et al. Ibuprofen binding to truncated HSA secondary sites FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS 661 NMR drug-competition analysis. J Mol Biol 361, 336–351. 18 Nicoletti FP, Howes BD, Fittipaldi M, Fanali G, Fasano M, Ascenzi P & Smulevich G (2008) Ibuprofen induces an allosteric conformational transition in the heme complex of human serum albumin with significant effects on heme ligation. J Am Chem Soc 130, 11677–11688. 19 Fanali G, Pariani G, Ascenzi P & Fasano M (2009) Allosteric and binding properties of Asp1–Glu382 trun- cated recombinant human serum albumin: an optical and NMR spectroscopic investigation. FEBS J 276, 2241–2250. 20 Ascenzi P, di Masi A, de Sanctis G, Coletta M & Fasano M (2009) Ibuprofen modulates allosterically NO dissociation from ferrous nitrosylated human serum heme–albumin by binding to three sites. Biochem Biophys Res Commun 387, 83–86. 21 Ascenzi P, di Masi A, Coletta M, Ciaccio C, Fanali G, Nicoletti FP, Smulevich G & Fasano M (2009) Ibupro- fen impairs allosterically peroxynitrite isomerization by ferric human serum heme–albumin. J Biol Chem 284, 31006–31017. 22 Ascenzi P, Imperi F, Coletta M & Fasano M (2008) Abacavir and warfarin modulate allosterically kinetics of NO dissociation from ferrous nitrosylated human serum heme–albumin. Biochem Biophys Res Commun 369, 686–691. 23 Baroni S, Mattu M, Vannini A, Cipollone R, Aime S, Ascenzi P & Fasano M (2001) Effect of ibuprofen and warfarin on the allosteric properties of haem–human serum albumin: a spectroscopic study. Eur J Biochem 268, 6214–6220. 24 Kharitonov VG, Sharma VS, Magde D & Koesling D (1997) Kinetics of nitric oxide dissociation from five- and six-coordinate nitrosyl hemes and heme proteins, including soluble guanylate cyclase. Biochemistry 36, 6814–6818. 25 Fanali G, Fesce R, Agrati C, Ascenzi P & Fasano M (2005) Allosteric modulation of myristate and Mn(III)- heme binding to human serum albumin: optical and NMR spectroscopy characterization. FEBS J 272, 4672–4683. 26 Fanali G, Bocedi A, Ascenzi P & Fasano M (2007) Modulation of heme and myristate binding to human serum albumin by anti-HIV drugs. An optical and NMR spectroscopic study. FEBS J 274, 4491–4502. 27 Fanali G, Ascenzi P & Fasano M (2007) Effect of prototypic drugs ibuprofen and warfarin on global chaotropic unfolding of human serum heme–albumin: a fast-field-cycling 1H-NMR relaxometric study. Biophys Chem 129, 29–35. 28 Fanali G, De Sanctis G, Gioia M, Coletta M, Ascenzi P & Fasano M (2009) Reversible two-step unfolding of heme–human serum albumin: a 1 H-NMR relaxometric and circular dichroism study. J Biol Inorg Chem 14, 209–217. 29 Bohle DS, Glassbrenner PA & Hansert B (1996) Syntheses of pure tetramethylammonium peroxynitrite. Methods Enzymol 269, 302–311. 30 Koppenol WH, Kissner R & Beckman JS (1996) Syntheses of peroxynitrite: to go with the flow or on solid grounds? Methods Enzymol 269, 296–302. 31 Herold S, Exner M & Boccini F (2003) The mechanism of the peroxynitrite mediated oxidation of myoglobin in the absence and presence of carbon dioxide. Chem Res Toxicol 16, 390–402. 32 Herold S, Kalinga S, Matsui T & Watanabe Y (2004) Mechanistic studies of the isomerization of peroxynitrite to nitrate catalyzed by distal histidine metmyoglobin mutants. J Am Chem Soc 126, 6945–6955. 33 Ascenzi P & Fasano M (2007) Abacavir modulates peroxynitrite-mediated oxidation of ferrous nitrosylated human serum heme–albumin. Biochem Biophys Res Commun 353, 469–474. 34 Ascenzi P & Visca P (2008) Scavenging of reactive nitrogen species by mycobacterial truncated hemoglo- bins. Methods Enzymol 436, 317–337. 35 Goldstein S & Mere ´ nyi G (2008) The chemistry of peroxynitrite: implications for biological activity. Methods Enzymol 436, 49–61. 36 Antonini E & Brunori M (1971) Hemoglobin and Myo- globin in their Reactions with Ligands. North Holland Publishing Co., Amsterdam and London. 37 Goldstein S, Lind J & Mere ´ nyi G (2005) Chemistry of peroxynitrites and peroxynitrates. Chem Rev 105, 2457–2470. 38 Herold S, Matsui T & Watanabe Y (2001) Peroxynitrite isomerization catalyzed by His64 myoglobin mutants. J Am Chem Soc 123, 4085–4086. 39 Herold S & Kalinga S (2003) Metmyoglobin and methe- moglobin catalyze the isomerization of peroxynitrite to nitrate. Biochemistry 42, 14036–14046. 40 Zunszain PA, Ghuman J, Komatsu T, Tsuchida E & Curry S (2003) Crystal structural analysis of human serum albumin complexed with hemin and fatty acid. BMC Struct Biol 3,6. Ibuprofen binding to truncated HSA secondary sites A. di Masi et al. 662 FEBS Journal 278 (2011) 654–662 ª 2011 The Authors Journal compilation ª 2011 FEBS . Ibuprofen binding to secondary sites allosterically modulates the spectroscopic and catalytic properties of human serum heme–albumin Alessandra di. according to Eqn (7) [20]: k off ¼ðk offð2Þ Âð ibuprofen =ðK 2  ibuprofen Þ þðk offð3Þ Âð ibuprofen =ðK 3  ibuprofen ÞÞ þ k þ off ð7Þ where k off(2) and k off(3) indicate