Modulation of heme and myristate binding to human serum albumin by anti-HIV drugs An optical and NMR spectroscopic study Gabriella Fanali 1 , Alessio Bocedi 2,3, *, Paolo Ascenzi 2,3 and Mauro Fasano 1 1 Dipartimento di Biologia Strutturale e Funzionale, and Centro di Neuroscienze, Universita ` dell’Insubria, Busto Arsizio, Italy 2 Istituto Nazionale per le Malattie Infettive I.R.C.C.S. ‘Lazzaro Spallanzani’, Rome, Italy 3 Dipartimento di Biologia, and Laboratorio Interdisciplinare di Microscopia Elettronica, Universita ` ‘Roma Tre’, Rome, Italy Human serum albumin (HSA) is the most prominent protein in plasma (its concentration being 45 mgÆmL )1 , i.e. 7.0 · 10 )4 m, in the serum of adults), but it is also found in tissues and secretions throughout the body [1]. HSA is made up of a single nonglycosylated all-a chain of 65 kDa, containing three homologous domains (I, II and III). Each domain is made up of two separate helical subdomains (A and B), connected by random coils (Fig. 1) [1–7]. The HSA globular domain structural organization provides a variety of binding sites for various ligands, making it an impor- tant determinant of the pharmacokinetic behaviour of many drugs [1,3,5–12]. Moreover, it accounts for most of the antioxidant capacity of human serum, acts as a nitric oxide depot and displays enzymatic properties [1,5,13–18]. Among other ligands, HSA is able to bind up to seven equivalents of long chain fatty acids (FAs) at multiple binding sites (labelled FA1–FA7; Fig. 1) with different affinity [5,12,19–22]. Remarkably, FA7 represents Keywords allosteric modulation; anti-HIV drugs; heterotropic interactions; human serum albumin; NMR relaxation Correspondence M. Fasano, Dipartimento di Biologia Strutturale e Funzionale, and Centro di Neuroscienze, Universita ` dell’Insubria, Via Alberto da Giussano 12, I-21052 Busto Arsizio (VA), Italy Fax: +39 0331 339459 Tel: +39 0331 339450 E-mail: mauro.fasano@uninsubria.it *Present address Istituto di Ricerche di Biologia Molecolare ‘P. Angeletti’, Rome, Italy (Received 27 February 2007, revised 20 June 2007, accepted 5 July 2007) doi:10.1111/j.1742-4658.2007.05978.x Human serum albumin (HSA) has an extraordinary ligand-binding capac- ity, and transports Fe(III)heme and medium- and long-chain fatty acids. In human immunodeficiency virus-infected patients the administered drugs bind to HSA and act as allosteric effectors. Here, the binding of Fe(III)- heme to HSA in the presence of three representative anti-HIV drugs and myristate is investigated. Values of the dissociation equilibrium constant K d for Fe(III)heme binding to HSA were determined at different myristate concentrations, in the absence and presence of anti-HIV drugs. Nuclear magnetic relaxation dispersion profiles of HSA–Fe(III)heme were mea- sured, at different myristate concentrations, in the absence and presence of anti-HIV drugs. Structural bases for anti-HIV drug binding to HSA are provided by automatic docking simulation. Abacavir and nevirapine bind to HSA with K d values of 1 · 10 )6 and 2 · 10 )6 m, respectively. Therefore, at concentrations used in therapy (in the 1–5 · 10 )6 m range) abacavir and nevirapine bind to HSA and increase the affinity of heme for HSA. In the presence of abacavir or nevirapine, the affinity is not lowered by myristate. FA7 should therefore be intended as a secondary binding site for abacavir and nevirapine. Binding of atazanavir is limited by the large size of the drug, although preferential binding may be envisaged to a site positively coupled with FA1 and FA2, and negatively coupled to FA7. As a whole, these results provide a foundation for the comprehension of the complex network of links modulating HSA-binding properties. Abbreviations FA, fatty acid; HSA, human serum albumin; NMRD, nuclear magnetic relaxation dispersion. FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS 4491 Sudlow’s site I, the preferential binding site for bulky, heterocyclic anions (e.g. warfarin), whereas the cavity hosting FA3 and FA4 contributes to Sudlow’s site II, which is preferred by aromatic carboxylates with an extended conformation (e.g. ibuprofen), benzodia- zepines (e.g. diazepam) and some anaesthetics [1,5,7,8,11,23–26]. FA1, located in subdomain IB (Fig. 1), hosts the heme, with the tetrapyrrole ring arranged in a D-shaped cavity limited by Tyr138 and Tyr161 resi- dues, which provide p–p stacking interaction with the porphyrin and supply a donor oxygen (from Tyr161) for the Fe(III)heme iron [27,28]. Interestingly, FA1 has a low affinity for long- and medium-chain FAs, sug- gesting that its structure has evolved to specifically bind the heme [21–29]. HSA undergoes pH- and allosteric effector- dependent reversible conformational isomerization(s). Between pH 4.3 and 8.0, in the absence of allosteric effectors, HSA displays the neutral (N) form that is characterized by a heart-shaped structure (Fig. 1). At pH values > 8.0, in the absence of allosteric effectors, HSA changes conformation to the basic (B) form (neutral-to-basic, N fi B transition) with loss of the a -helix content and an increased affinity for some ligands [1,6,7,30–37]. Ligand binding to HSA stabilizes protein conformers N or B, thus regulating allosterical- ly the conformational transition(s). Heme regulates drug binding to Sudlow’s site I by heterotropic inter- actions. Indeed, the affinity of Fe(III)heme for HSA decreases by about one order of magnitude upon drug binding, and accordingly Fe(III)heme binding to HSA decreases drug affinity by the same extent. Therefore, drugs that bind to Sudlow’s site I (e.g. warfarin) act as allosteric effectors for Fe(III)heme association, and vice versa. Also benzodiazepines bind to several functionally and allosterically linked HSA clefts, depending on their optical conformation and substitu- tion [35,38–42]. In HIV-infected individuals the primary target of therapy is the HIV itself, but most of the clinical mani- festations are related to the effect of HIV on the immune system, which leads to progressive immunode- ficiency. Recently, the introduction of highly effective combination regimens of antiretroviral drugs has led to substantial improvements in morbidity and mortal- ity [43]. The anti-HIV drugs include three different classes among nucleoside reverse transcriptase inhibi- tors, non-nucleoside reverse transcriptase inhibitors and protease inhibitors. Nucleoside reverse transcrip- tase inhibitors are intracellularly phosphorylated to their corresponding triphosphorylated derivatives, which compete with the corresponding natural nucleo- tide for binding to HIV reverse transcriptase and inhi- bit it. Non-nucleoside reverse transcriptase inhibitors act as noncompetitive inhibitors of the HIV reverse transcriptase. Protease inhibitors interfere with viral replication by inhibiting the viral protease, preventing maturation of the HIV virus and causing the forma- tion of noninfectious virions [43–48]. The therapeutic efficiency of anti-HIV drugs in combination therapy is strictly dependent upon the mutual interaction(s) of binding equilibria with plasma proteins and in particu- lar with HSA. One of the most important factors affecting the distribution and the free, active concen- tration of many administered drugs is binding affinity for HSA [7,49–51]. Here, the effect of myristate on the binding of Fe(III)heme to HSA in the absence and presence of three anti-HIV drugs belonging to different pharmaco- logical classes, i.e. abacavir, nevirapine and atazanavir (Scheme 1), is reported by means of optical and mag- netic spectroscopy. Moreover, a screening of potential HSA-binding sites has been performed by automated docking simulation for the different anti-HIV drugs considered. Results Fe(III)heme was titrated with HSA by measuring the difference in absorbance at 411 nm in the UV–Vis spectrum (DA; see Eqn 1) with respect to the spectrum Fig. 1. HSA structure. Atomic coordinates were taken from PDB entry 1O9X [28]. FA-binding sites are indicated by arrows and labelled. Site FA1 is occupied by heme (red). Sites FA2–FA7 are occupied by myristate (green). FA7 represents Sudlow’s site I (i.e. the warfarin site). FA3 and FA4 together represent Sudlow’s site II (i.e. the ibuprofen site). For further details see text. Modulation of HSA ligand binding by anti-HIV drugs G. Fanali et al. 4492 FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS in the absence of HSA at different myristate concen- trations (Fig. 2) in the absence (Fig. 2A) and presence of abacavir (Fig. 2B), nevirapine (Fig. 2C) and ataz- anavir (Fig. 2D), respectively. In agreement with a similar experiment performed on Mn(III)heme [36], in the absence of any drug (Fig. 2A) myristate may compete for the heme site with a twofold reduction in the affinity of heme for HSA (Table 1). In the presence of abacavir (Fig. 2B), positive cooperation is observed in the absence of myristate: therefore, the K d value cannot be obtained using Eqn (1). In the presence of myristate the curves assume a hyperbolic behaviour and the values reported in Table 1 are obtained. The affinity of heme for HSA is slightly improved even in the presence of myristate. Binding isotherms obtained in the presence of nevirapine (Fig. 2C) tend to an asymptotic DA max value higher than that observed in the absence of any drug and in the presence of abaca- vir. In the absence of myristate and at a myristate con- centration of 1.0 · 10 )5 m, curves cannot be fitted using Eqn (1): however, at higher myristate concentra- tions the K d values are consistent with an increased affinity of HSA for heme. In the presence of atazana- vir (Fig. 2D), a hyperbolic binding isotherm is observed, affording the determination of K d ¼ 7.4 · 10 )8 m, a value definitely smaller than those observed in the absence or presence of other allosteric effectors. When myristate is added, a more compli- cated situation occurs with binding isotherms that show features observed for both other drugs as well. The asymptotic DA max value changes with the myri- state concentration, the value in the absence of myri- state being lower than the corresponding value in the absence of any drug. In order to check the possibility that abacavir, nevi- rapine and atazanavir could fit into the heme (FA1) binding site, as well as into Sudlow’s site I (FA7), or into the other two sites in close proximity (FA2 and FA6), an automated docking analysis of the three drugs was performed in the four binding sites. Inter- molecular energy values are reported in Table 2. Fig- ure 3 shows a ribbon model of the HSA FA1 region with abacavir and nevirapine superimposed on Fe(III)- heme, whereas Figs 4–6 show ribbon models of abaca- vir, nevirapine and atazanavir superimposed on myristate in sites FA2, FA6 and FA7, respectively. As can be seen, abacavir (Figs 3A, 5A, and 6A) fits reasonably well in sites FA1, FA6 and FA7, whereas docking to site FA2 is disadvantaged. Nevirapine (Figs 3B, 4, 5B, and 6B) is able to enter all four sites, thus competing with their ligands and potentially acting as an allosteric effector. By contrast, atazanavir (Fig. 5C) could partially enter the binding cavities, although steric clashes between this large ligand and the protein matrix make it unlikely, except for the extended FA6 trough. Paramagnetic Fe(III)heme–HSA(–drug) complexes were also investigated in terms of their ability to relax solvent water protons at different proton Larmor fre- quencies. When the magnetic field is rapidly changed from a low to a high value, the magnetization expo- nentially changes to reach its equilibrium value with a time constant that is T 1 in the new magnetic field. Therefore, measurement of magnetization at progres- sive time intervals allows us to obtain the T 1 value. In order to read the magnetization value, it has to be transformed into an electromagnetic signal by a radio- frequency pulse at its Larmor frequency that depends linearly from the magnetic field. Therefore, the field needs to be re-switched at a unique value correspond- ing to the frequency at which the transmitter and the receiver are tuned. Such a fast change of the magnetic field between equilibration, evolution (relaxation) and detection values is called fast-field cycling. If this Scheme 1. G. Fanali et al. Modulation of HSA ligand binding by anti-HIV drugs FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS 4493 experiment is repeated for different magnetic fields (i.e. at different proton Larmor frequencies) a nuclear magnetic relaxation dispersion (NMRD) profile is obtained. Figure 7 reports the NMRD profiles of 1.0 · 10 )4 m Fe(III)heme–HSA at different myristate concentrations in the absence (Fig. 7A) and presence of abacavir (Fig. 7B), nevirapine (Fig. 7C) and ataz- anavir (Fig. 7D). By increasing the myristate concen- tration, a slight smoothing of the curves is observed in all cases. In the presence of abacavir (Fig. 7B), no significant differences might be appreciated in compari- son with Fig. 7A (i.e. in the absence of the anti-HIV drugs). Remarkably, in the presence of nevirapine (Fig. 7C) significant quenching of the low-frequency Fig. 2. Binding isotherms for Fe(III)heme binding to HSA (FA free) and to HSA–myristate complexes in the absence of drugs (A) and pres- ence of abacavir (B), nevirapine (C) and atazanavir (D), at pH 7.0 and 25 °C. In all panels, the binding isotherm measured in the absence of either drugs or myristate is shown for comparison (solid squares); solid diamonds, no myristate; open upward triangles, 1.0 · 10 )5 M myri- state; open downward triangles, 7.5 · 10 )5 M myristate; open diamonds, 1.0 · 10 )4 M myristate. The continuous lines were obtained by analysis of the data by using Eqn (1), when applicable. For further details see text. Table 1. Values of the thermodynamic dissociation constants (K d , M) for Fe(III)heme–HSA in the absence and presence of abacavir, nevira- pine and atazanavir, at different myristate concentration, pH 7.0 and 25 °C. [Myristate] ( M) No drug Abacavir Nevirapine Atazanavir 0 (5.0 ± 0.2) · 10 )7a a (7.4 ± 3.0) · 10 )8 1.0 · 10 )5 (5.3 ± 0.4) · 10 )7 (2.9 ± 0.4) · 10 )7a a 7.5 · 10 )5 (1.3 ± 0.2) · 10 )6 (2.5 ± 0.4) · 10 )7 (2.9 ± 0.3) · 10 )7a 1.0 · 10 )4 (1.3 ± 0.1) · 10 )6 (5.1 ± 0.8) · 10 )7 (1.1 ± 0.2) · 10 )7a a The binding isotherm deviates significantly from the hyperbolic behaviour, therefore data cannot be analysed in terms of Eqn (1). Modulation of HSA ligand binding by anti-HIV drugs G. Fanali et al. 4494 FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS region is observed that is further emphasized by the presence of myristate. Atazanavir (Fig. 7D) does not affect the NMRD profile of Fe(III)heme–HSA in the absence of myristate, although a moderate decrease at low frequency and a remarkable smoothing might be appreciated. Noticeably, an exhaustive theoretical treatment of the NMRD profiles of high-spin Fe(III)- heme protein complexes would require extensive com- putational work and is beyond the scope of this study. Discussion Allosteric modulation of HSA-binding properties is fundamental for a safe management of patients subject to multidrug therapy, affecting the distribution and the free concentration of each administered drug. While a certain extent of HSA interaction may be desirable to help drug solubilization and distribution, a too tight an interaction negatively affects the distribution to sites of action and dramatically increases the total con- centration of the administered drug [7,35,50,51]. How- ever, negative modulation of the drug–HSA affinity by heterotropic interactions would suddenly increase the drug concentration which may reach the toxicity threshold [7,31,35,36,51–55]. Here, we show how drugs currently used at the micromolar level in highly active antiretroviral therapy may bind to different sites thus causing opposite effects on the conformational states of HSA. Myristate, binding to all FA sites, competes with Fe(III)heme in FA1 determining an increase of K d by a factor of two and at the same time modulates the N fi B transition and stabilizes the binding of Fe(III)heme. This finding is in agreement with a simi- lar behaviour observed for the less-stable Mn(III)- heme–HSA complex [36]. Abacavir has been recently reported to be a Sud- low’s site I (FA7) ligand because of its ability to quench Trp214 fluorescence and the negative modulation it exerts on heme binding [51]. Indeed, a thorough analysis of several HSA-binding drugs has been reported showing that all FA7 ligands reduce the affinity of heme by one order of magnitude and, accordingly, heme reduces by one order of magnitude the affinity of all FA7 ligands [35]. However, the results shown here indicate that abacavir binding involves multiple sites and FA7 is probably not the primary binding site for abacavir. Indeed, abacavir A B Fig. 3. Superimposition of Fe(III)heme and abacavir (A) and nevira- pine (B) in binding site FA1. Ligands are coloured as follows: heme, red; abacavir, blue; nevirapine, orange. Atomic coordinates were taken from the PDB entry 1O9X [28]. For further details see text. Table 2. Values of intermolecular energies (kJÆmol )1 ) obtained from the docking simulation of the three anti-HIV drugs in the heme binding cavity (FA1) and in the functionally linked binding sites. ND, not determined. Abacavir Nevirapine Atazanavir FA1 ) 30.6 ) 26.5 $ 10 3 FA2 ND ) 20.3 $ 10 4 FA6 ) 28.9 ) 30.7 ) 11.2 FA7 ) 31.6 ) 33.5 $ 10 3 G. Fanali et al. Modulation of HSA ligand binding by anti-HIV drugs FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS 4495 binds to HSA at K d ¼ 10 )6 m (Supplementary mate- rial), two orders of magnitude lower than that obtained by fluorescence quenching [51]. On the basis of the docking simulations, abacavir would also bind to FA1, competitively preventing the binding of heme. At concentrations used previously [51], abacavir may enter the FA7 cavity and block the conformational switch towards the B form, characterized by an increased affinity for FA1 ligands (Supplementary material). The actual ability of abacavir to fit FA1, FA6 and FA7 is explored by docking simulations. In the absence of experimental 3D structures of HSA–anti-HIV complexes, docking procedures based on Monte Carlo-simulated annealing are effective tools for the screening of binding possibilities for the differ- ent drug–HSA interactions [56–58]. As shown in Fig. 3A, abacavir fits FA1, competing with heme; abacavir may also fit into FA6 (Fig. 5A); conversely, Fig. 6A shows that abacavir may also enter FA7 and consequently lower the affinity for FA1, FA1 and FA7 being functionally linked [21,31,35,51]. Although the structural basis for the observed allosteric regulation of heme binding by Sudlow’s site I ligands is not known, it has been suggested that it may be mediated by rearrangement of the Phe149–Tyr150 dyad, Phe149 A B C Fig. 4. Superimposition of myristate and nevirapine in binding site FA2. Ligands are coloured as follows: myristate, green; nevirapine, orange. Atomic coordinates were taken from the PDB entry 1O9X [28]. For further details see text. Fig. 5. Superimposition of myristate and abacavir (A), nevirapine (B) and atazanavir (C) in binding site FA6. Ligands are coloured as fol- lows: myristate, green; abacavir, blue; nevirapine, orange; atazana- vir, cyan. Atomic coordinates were taken from the PDB entry 1O9X [28]. For further details see text. Modulation of HSA ligand binding by anti-HIV drugs G. Fanali et al. 4496 FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS contacting the heme ring and Tyr150 protruding into Sudlow’s site I (i.e. FA7) [21]. Nevirapine, a small hydrophobic butterfly-shaped ligand, was reported to be a FA7 ligand as well, thus acting as an allosteric negative effector of FA1. Also in this case, this is just one of multiple binding modes of nevirapine to HSA previously distinguished on the basis of fluorescence quenching of Trp214 (located in close proximity of FA7) [51]. Indeed, nevirapine binds to its primary binding site at K d $ 2 · 10 )6 m (Supple- mentary material). By looking at Fe(III)heme binding it becomes clear that nevirapine acts as an allosteric effector that increases the affinity of Fe(III)heme for FA1. This positive modulation may be due to FA2 (Fig. 4), the only FA-binding site that contacts both HSA domains I and II. A structural explanation would involve Tyr150 again; actually, binding of myristate to FA2 attracts Tyr150 and Arg252 towards the FA car- boxylate moiety [21]. Therefore, Arg252 is no longer available to stabilize FA7 ligands; however, the reori- entation of Tyr150 may stabilize the interaction of Phe149 with Fe(III)heme, thus explaining the allo- steric modulation observed in solution studies [7,18,31,35,36,39,51]. Atazanavir is a large, extended peptidomimetic drug, and may fit multiple sites by partially entering them. Although the four sites considered are poten- tially able to host a more or less extended part of the atazanavir molecule, the drug experiences steric hin- drance due to residues that point outside the protein core (Table 2). The only docking that shows stabiliza- tion of the interaction energy takes place in FA6 for its extended, open-trough conformation (Fig. 5C). As a consequence, a situation occurs that is intermediate between those observed for abacavir and nevirapine. Interestingly, in the absence of myristate, ataza- navir displays the larger stabilization effect on heme affinity. All these considerations are supported by the NMRD data. Indeed, NMRD profiles recorded in the absence (Fig. 7A) and presence (Fig. 7B) of abacavir do not differ significantly; also, no differences are observed by increasing the myristate-to-HSA ratio. Nevirapine dramatically affects the profile by reducing the R 1p values at the low-frequency limit even in the absence of myristate, the high-frequency region being unaffected. Although extensive theoretical treatment of the NMRD data is beyond the scope of this study, it should be noted that such a smoothing of the NMRD profile in high-spin Fe(III) complexes is usually reported to be associated with a distortion of the zero- field splitting tensor [60,61]. Indeed, in slowly rotating systems the Solomon–Bloembergen–Morgan equation breaks down and R 1p values are affected by a number of parameters arising from both contact and dipolar electron–nucleus interactions, including anisotropies of the g tensor, the zero-field splitting tensor, and the hyperfine coupling tensor [62–64]. Interestingly, bind- ing of nevirapine to HSA determines a remarkable dis- tortion of the heme environment that reflects on both NMRD profiles (Fig. 7C) and the asymptotic value of the binding isotherms (Fig. 2C). In conclusion, our results demonstrate that anti-HIV drugs at concentrations used in highly active antiretro- A B Fig. 6. Superimposition of myristate and abacavir (A) and nevirapine (B) in binding site FA7. Ligands are coloured as follows: myristate, green; abacavir, blue; nevirapine, orange. Atomic coordinates were taken from the PDB entry 1O9X [28]. For further details see text. G. Fanali et al. Modulation of HSA ligand binding by anti-HIV drugs FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS 4497 viral therapy may allosterically exert heterotropic inter- actions that influence reciprocally the Fe(III)heme- and FA-binding properties to HSA. FA7, i.e. Sudlow’s site I, is confirmed to be negatively linked to FA1, as already established by solution studies [7,12,18, 21,31,35,36,39,51]. Thus, the increase in plasma levels of Fe(III)heme under pathological conditions (e.g. severe haemolytic anaemia, crash syndrome and post- ischaemic reperfusion) may induce the release of FA7- bound drugs with the concomitant intoxication of the patient [7,35,51]. Moreover, binding of drugs to FA2 reduces the affinity of FA7 and increases the affinity of FA1 ligands. Eventually, FA6 ligands are expected to affect in some way the occupancy of FA1, as evinced previously [21,36]. Thus, allosteric regulation of ligand binding is relevant in pharmacological ther- apy management, the nonspecific binding of drugs to plasma proteins being an important determinant of their biological efficacy by modulation of drug avail- ability to the intended target. Experimental procedures Abacavir (GlaxoSmithKline, London, UK), nevirapine (Boehringer Ingelheim, Ridgefield, CO), and atazanavir (Bristol-Myers Squibb, Princeton, NJ) were obtained through the NIH AIDS Research Reagent Program, Division of AIDS, NIAID, National Institute of Health (Bethesda, MD). All other reagents (from Sigma-Aldrich, St Louis, MO), were of the highest purity available, and were used without further purification. HSA (Sigma-Aldrich) was essentially FA-free according to the charcoal delipidation protocol [65–67] and was used without further purification. Absence of significant amounts of covalent dimers was checked using a Bruker Ultraflex MALDI-TOF mass spec- trometer (Bruker Daltonics, Bremen, Germany). Fig. 7. NMRD profiles of Fe(III)heme–HSA (FA free) and Fe(III)heme–HSA–myristate complexes in the absence of drugs (A) and in the presence of abacavir (B), nevirapine (C) and atazanavir (D), at pH 7.0 and 25 °C. In all panels, the NMRD profile measured in the absence of either drugs or myristate is shown for comparison (solid squares); solid diamonds, no myristate; open upward triangles, 1.0 · 10 )4 M myristate; open diamonds: 4.0 · 10 )4 M myristate. Fe(III)heme–HSA concentration was 1.0 · 10 )4 M. R 1p values were normalized to 1.0 · 10 )3 M. For further details see text. Modulation of HSA ligand binding by anti-HIV drugs G. Fanali et al. 4498 FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS Fe(III)heme–HSA was prepared by adding the appropri- ate volume of 1.2 · 10 )2 m Fe(III)heme, dissolved in 1.0 · 10 )1 m NaOH, to a 1.0 · 10 )4 m HSA solution in 0.1 m phosphate buffer pH 7.0. In the final HSA solution Fe(III)heme–HSA was 1.0 · 10 )4 m. The actual concentration of the Fe(III)heme stock solu- tion was checked as a bis-imidazolate complex in SDS micelles with an extinction coefficient of 14.5 mm )1 Æcm )1 (at 535 nm) [68]. Under all the experimental conditions, no free Fe(III)heme was present in the reaction mixtures. The actual concentration of the HSA stock solution was deter- mined using the Bradford method [69]. Sodium myristate solution (0.1 m) was prepared by add- ing 0.1 m FA to NaOH 1.0 · 10 )1 m. The solution was heated to 100 °C and stirred to dissolve the FA. The sodium myristate solution was cooled and then mixed with 1.0 · 10 )4 m Fe(III)heme–HSA (FA-free) to achieve the desired FA to protein molar ratio. The Fe(III)heme–HSA– myristate complex was incubated for 1 h at room tempera- ture with continuous stirring [28]. Stock solutions of 1.2 · 10 )1 m anti-HIV drugs were pre- pared by dissolving abacavir, atazanavir and nevirapine in dimethylsulfoxide. Anti-HIV drugs were added to the Fe(III)heme–HSA 1.0 · 10 )4 m solution to a final concen- tration of 1.0 · 10 )4 m. Automatic flexible ligand-docking simulation to HSA was performed using autodock 3.0 and the graphical user interface autodocktools [54–56,68]. The structure of Fe(III)heme–HSA–myristate was downloaded from the Pro- tein Data Bank (PDB code: 1O9X) [28]. Ribbon representa- tion of HSA with stick representation of ligands was drawn with the swiss-pdbviewer [71]. The nevirapine geometry was energy-minimized starting from the structure of the drug observed in its complex with the Thr215Tyr mutant HIV-1 reverse transcriptase (PDB code: 1LWO) [72]. Abacavir and atazanavir structures were calculated using the Dundee prodrg server [73]. Single bonds were allowed to rotate freely during the Monte Carlo-simulated anneal- ing procedure. Analysis of the conformational space was restricted to a cubic box of 40 A ˚ , edge centred on the coor- dinates of heme (for FA1 site) or myristate (for FA2, FA6, and FA7 sites). Monte Carlo-simulated annealing was per- formed by starting from a temperature of 900 K with a rel- ative cooling factor of 0.95 ⁄ cycle, in order to reach the temperature of 5 K in 100 cycles [56–58]. Binding of Fe(III)heme to HSA in the presence of either drug and ⁄ or myristate was investigated spectrophotometri- cally using an optical cell with 1.0 cm path length on a Cary 50 Bio spectrophotometer (Varian Inc., Palo Alto, CA). In a typical experiment, a small amount of a 1.2 · 10 )2 m Fe(III)heme solution in 1.0 · 10 )1 m NaOH was diluted in the optical cell in a solvent mixture of 10% dimethylsulfoxide (this concentration does not affect differ- ence spectra) in 1.0 · 10 )1 m phosphate buffer pH 7.0 to a final chromophore concentration of 1.0 · 10 )5 m, in the presence of anti-HIV drugs (at 4.0 · 10 )5 m concentration, i.e. similar to concentrations used in therapy) at different myristate concentrations (0–1.0 · 10 )4 m). This solution was titrated with HSA by adding small amounts of a 1.0 · 10 )3 m protein solution in 1.0 · 10 )1 m phosphate buffer pH 7.0 and recording the spectrum after a few min- utes incubation following each addition. Difference spectra with respect to Fe(III)heme were taken and binding iso- therms were analysed by plotting the difference of absor- bance against the protein concentration. Data were fitted by using the following equation: where DA is the difference in the Soret band (411 nm) absorbance, DA max is the difference of absorbance at limit- ing HSA concentration, K a is the association constant for Fe(III)heme binding to HSA (i.e. K À 1 d ), [L t ] is the total concentration of Fe(III)heme, [P t ] is the total concentra- tion of HSA, and N is the number of equivalent binding sites. 1 H NMRD profiles of 1.0 · 10 )4 m Fe(III)heme–HSA were recorded on a Stelar Spinmaster-FFC fast-field cycling relaxometer (Stelar, Mede, PV, Italy) in the absence and presence of abacavir (1.0 · 10 )4 m), nevirapine (1.0 · 10 )4 m) and atazanavir (1.0 · 10 )4 m), in the absence and presence of 1.0 · 10 )4 and 4.0 · 10 )4 m myristate. NMRD profiles were obtained by measuring water proton longitudinal relaxation rates (R 1obs ) at magnetic field strengths in the range from 2.4 · 10 )4 to 2.35 · 10 )1 T (corresponding to proton Larmor frequencies from 0.01 to 10 MHz). The R 1p relaxivity values (i.e. paramagnetic con- tributions to the solvent water longitudinal relaxation rate referenced to a 1.0 · 10 )3 m concentration of paramagnetic agent) were determined by subtracting from the observed relaxation rate (R 1obs ) the blank relaxation rate value (R 1dia ) measured for the buffer at the experimental temper- ature. Acknowledgements The authors wish to thank Professor Massimo Coletta and Professor Riccardo Fesce for helpful discussions. This study was partly supported by grants from the Italian Ministry of Health (Istituto Nazionale per le DA ¼ DA max Á  ðK a Á½L t þNÁ½P t ÁK a þ 1ÞÀ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðK a Á½L t þNÁ½P t ÁK a þ 1Þ 2 À 4K 2 a Á½L t ÁNÁ½P t  q  2K a Á½L t  ð1Þ G. Fanali et al. Modulation of HSA ligand binding by anti-HIV drugs FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS 4499 Malattie Infettive I.R.C.C.S. ‘Lazzaro Spallanzani’, Roma, Italy, ‘Ricerca corrente 2006’ to PA). This paper is dedicated to the memory of Dr Fabrizio Poccia. References 1 Peters T Jr (1996) All About Albumin: Biochemistry, Genetics and Medical Applications. Academic Press, San Diego, CA. 2 He X & Carter DC (1992) Atomic structure and chemis- try of human serum albumin. Nature 358, 209–215. 3 Carter DC & Ho JX (1994) Structure of serum albumin. Adv Protein Chem 45, 153–203. 4 Sugio S, Kashima A, Mochizuki S, Noda M & Kobay- ashi K (1999) Crystal structure of human serum albu- min at 2.5 A ˚ resolution. Protein Eng 12, 439–446. 5 Curry S (2002) Beyond expansion: structural studies on the transport roles of human serum albumin. Vox Sang 83 (Suppl. 1), 315–319. 6 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 albu- min. IUBMB Life 57, 787–796. 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 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. 9 Vallner JJ (1977) Binding of drugs by albumin and plasma protein. J Pharm Sci 66, 447–465. 10 Bertucci C & Domenici E (2002) Reversible and cova- lent binding of drugs to human serum albumin: method- ological approaches and physiological relevance. Curr Med Chem 9, 1463–1481. 11 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. 12 Simard JR, Zunszain PA, Ha CE, Yang JS, Bhagavan NV, Petitpas I, Curry S & Hamilton JA (2005) Locating high-affinity fatty acid-binding sites on albumin by X-ray crystallography and NMR spectroscopy. Proc Natl Acad Sci USA 102, 17958–17963. 13 Sampath V, Zhao XJ & Caughey WS (2001) Anesthetic- like interactions of nitric oxide with albumin and heme- proteins. A mechanism for control of protein function. J Biol Chem 276, 13635–13643. 14 Monzani E, Bonafe ` B, Fallarini A, Redaelli C, Casella L, Minchiotti L & Galliano M (2001) Enzymatic properties of hemalbumin. Biochim Biophys Acta 1547, 302–312. 15 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. 16 Monzani E, Curto M, Galliano M, Minchiotti L, Aime S, Baroni S, Fasano M, Amoresano A, Salzano AM, Pucci P et al. (2002) Binding and relaxometric properties of heme complexes with cyanogen bromide fragments of human serum albumin. Biophys J 83, 2248–2258. 17 Sakurai Y, Ma SF, Watanabe H, Yamaotsu N, Hirono S, Kurono Y, Kragh-Hansen U & Otagiri M (2004) Esterase-like activity of serum albumin: characterization of its structural chemistry using p-nitrophenyl esters as substrates. Pharm Res 21, 285–292. 18 Ascenzi P & Fasano M (2007) Abacavir modulates per- oxynitrite-mediated oxidation of ferrous nitrosylated human serum heme-albumin. Biochem Biophys Res Commun 353, 469–474. 19 Spector AA (1975) Fatty acid binding to plasma albu- min. J Lipid Res 16, 165–179. 20 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. 21 Simard JR, Zunszain PA, Hamilton JA & Curry S (2006) Location of high and low affinity fatty acid bind- ing sites on human serum albumin revealed by NMR drug-competition analysis. J Mol Biol 361, 336–351. 22 Kragh-Hansen U, Watanabe H, Nakajou K, Iwao Y & Otagiri M (2006) Chain length-dependent binding of fatty acid anions to human serum albumin studied by site-directed mutagenesis. J Mol Biol 363, 702–712. 23 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. 24 Bhattacharya AA, Curry S & Franks NP (2000) Binding of the general anesthetics propofol and halothane to human serum albumin. High resolution crystal struc- tures. J Biol Chem 275, 38731–38738. 25 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. 26 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. 27 Wardell M, Wang Z, Ho JX, Robert J, Ru ¨ ker F, Ruble J & Carter DC (2002) The atomic structure of human methemalbumin at 1.9 A ˚ . Biochem Biophys Res Commun 291, 813–819. 28 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. Modulation of HSA ligand binding by anti-HIV drugs G. Fanali et al. 4500 FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS [...]... Fasano M & Ascenzi P (2004) Binding of anti-HIV drugs to human serum albumin IUBMB Life 56, 609–614 Bocedi A, Notari S, Menegatti E, Fanali G, Fasano M & Ascenzi P (2005) Allosteric modulation of anti-HIV drug and ferric heme binding to human serum albumin FEBS J 272, 6287–6296 Bertucci C, Canepa A, Ascoli GA, Guimaraes LF & Felix G (1999) Site I on human albumin: differences in the binding of (R)- and. .. Heme impairs allosteric drug binding to human serum albumin Sudlow’s site I Biochem Biophys Res Commun 334, 481–486 36 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 37 Fanali G, Ascenzi P & Fasano M (2007) Effect of prototypic drugs ibuprofen... ibuprofen and warfarin on global chaotropic unfolding of human serum heme- albumin: a fast-field-cycling (1)H -NMR relaxometric study Biophys Chem 129, 29–35 38 Fitos I, Visy J & Kardos J (2002) Stereoselective kinetics of warfarin binding to human serum albumin: effect of an allosteric interaction Chirality 14, 442–448 39 Chuang VTG & Otagiri M (2002) How do fatty acids cause allosteric binding of drugs to human. .. Fig S1 Binding isotherms for abacavir (A) and nevirapine (B) binding to HSA (FA free), at pH 7.0 and 25°C Fig S2 Binding isotherms for Fe(III )heme binding to HSA (FA free) and to HSA myristate complexes in the absence of drugs (A) and in the presence of 4.0 · 10–4 M abacavir (B), 4.0 · 10–3 M nevirapine (C) and 4.0 · 10–4 M atazanavir (D), at pH 7.0 and 25°C This material is available as part of the... Chromatographic analysis of carbamazepine binding to human serum albumin Modulation of HSA ligand binding by anti-HIV drugs 43 44 45 46 47 48 49 50 51 52 53 54 55 J Chromatogr B Anal Technol Biomed Life Sci 816, 57–66 Darbyshire J (2000) Therapeutic interventions in HIV infection ) a critical view Trop Med Int Health 5, A26– A31 van Heeswijk RPG (2002) Critical issues in therapeutic drug monitoring of antiretroviral...G Fanali et al 29 Fasano M, Fanali G, Leboffe L & Ascenzi P (2007) Heme binding to albuminoid proteins is the result of recent evolution IUBMB Life 59, 436–440 30 Yamasaki K, Maruyama T, Yoshimoto K, Tsutsumi Y, Narazaki R, Fukuhara A, Kragh-Hansen U & Otagiri M (1999) Interactive binding to the two principal ligand binding sites of human serum albumin: effect of the neutral -to- base transition... allosteric binding of drugs to human serum albumin? Pharm Res 19, 1458–1464 40 Chen J & Hage DS (2004) Quantitative analysis of allosteric drug–protein binding by biointeraction chromatography Nat Biotechnol 22, 1445–1448 41 Chen J, Ohnmacht C & Hage DS (2004) Studies of phenytoin binding to human serum albumin by high-performance affinity chromatography J Chromatogr B Anal Technol Biomed Life Sci 809,... blood–Betarain partitioning and human FEBS Journal 274 (2007) 4491–4502 ª 2007 The Authors Journal compilation ª 2007 FEBS 4501 Modulation of HSA ligand binding by anti-HIV drugs 56 57 58 59 60 61 62 63 64 65 66 67 G Fanali et al serum albumin binding based on COSMO-RS sigmamoments J Chem Inf Model 47, 228–233 Goodsell DS & Olson AJ (1990) Automated docking of substrates to proteins by simulated annealing Proteins... Effect of bezafibrate and clofibrate on the hemeiron geometry of ferrous nitrosylated heme- human serum albumin: an EPR study J Inorg Biochem 84, 293–296 34 Fasano M, Mattu M, Coletta M & Ascenzi P (2002) The heme- iron geometry of ferrous nitrosylated heme- serum lipoproteins, hemopexin, and albumin: a comparative EPR study J Inorg Biochem 91, 487–490 35 Ascenzi P, Bocedi A, Notari S, Menegatti E & Fasano... (2003) Structural basis of albumin thyroxine interactions and familial dysalbuminemic hyperthyroxinemia Proc Natl Acad Sci USA 100, 6440–6445 Banci L, Bertini I & Luchinat C (1991) Nuclear and Electron Relaxation VCH, Weinheim, Germany Zewert TE, Gray HB & Bertini I (1994) Environment of the heme in myoglobins NMRD and EPR spectroscopy of Val68X (X ¼ Asn, Asp, and Glu) mutants of human myoglobin J Am Chem . Modulation of heme and myristate binding to human serum albumin by anti-HIV drugs An optical and NMR spectroscopic study Gabriella Fanali 1 ,. 2007) doi:10.1111/j.1742-4658.2007.05978.x Human serum albumin (HSA) has an extraordinary ligand -binding capac- ity, and transports Fe(III )heme and medium- and long-chain fatty acids. In human