Ligand Binding to the FA3-FA4 Cleft Inhibits the Esterase-Like Activity of Human Serum Albumin

The hydrolysis of 4-nitrophenyl esters of hexanoate (NphOHe) and decanoate (NphODe) by human serum albumin (HSA) at Tyr411, located at the FA3-FA4 site, has been investigated between pH 5.8 and 9.5, at 22.0°C. Values of K s, k +2, and k +2/K s obtained at [HSA] ≥ 5×[NphOXx] and [NphOXx] ≥ 5×[HSA] (Xx is NphOHe or NphODe) match very well each other; moreover, the deacylation step turns out to be the rate limiting step in catalysis (i.e., k +3 << k +2). The pH dependence of the kinetic parameters for the hydrolysis of NphOHe and NphODe can be described by the acidic pK a-shift of a single amino acid residue, which varies from 8.9 in the free HSA to 7.6 and 7.0 in the HSA:NphOHe and HSA:NphODe complex, respectively; the pK>a-shift appears to be correlated to the length of the fatty acid tail of the substrate. The inhibition of the HSA-Tyr411-catalyzed hydrolysis of NphOHe, NphODe, and 4-nitrophenyl myristate (NphOMy) by five inhibitors (i.e., diazepam, diflunisal, ibuprofen, 3-indoxyl-sulfate, and propofol) has been investigated at pH 7.5 and 22.0°C, resulting competitive. The affinity of diazepam, diflunisal, ibuprofen, 3-indoxyl-sulfate, and propofol for HSA reflects the selectivity of the FA3-FA4 cleft. Under conditions where Tyr411 is not acylated, the molar fraction of diazepam, diflunisal, ibuprofen, and 3-indoxyl-sulfate bound to HSA is higher than 0.9 whereas the molar fraction of propofol bound to HSA is ca. 0.5.

At Lys199, the substrate (e.g., acetylsalicylic acid, trinitrobenzeno-sulfonates, and penicillin) is cleaved in two products; while one product is released, the other one binds covalently to the Lys199 residue [1,4,7]. Although the molecular mechanism underlying the Lys199 acetylation is unknown, it seems that its ability to attack the substrate is due to the proximity of the Lys195 residue, these two residues playing a combined and comparable chemical role. In fact, the basic form of Lys199 is likely connected to the acid form of Lys195 through a network of H-bonding water molecules with a donor-acceptor character. The presence of these water bridges is relevant for stabilizing the configuration of the FA7 site and/or promoting a potential Lys195-Lys199 proton-transfer process [6]. Since Lys199 is placed at the entrance of the FA7 site (i.e., Sudlow's site I), ligand binding inhibits the Lys199-dependent esterase activity [3,8].
The catalytic mechanism involving the Tyr411 residue appears to be substrate-dependent. Of note, the hydrolysis of the most suitable substrate 4-nitrophenyl propionate leads to the release of both 4-nitrophenol and propionate [9]. This mechanism also applies to the hydrolysis of N-trans-cinnamoylimidazoles [10] and 4-nitrophenil esters of amino acids [11]. However, the Tyr411-assisted hydrolysis of NphOAc and NphOMy leads to the release of 4-nitrophenol and to Tyr411-acetylation and-myristoylation, respectively [12,13]. The strong nucleophilic nature of the phenolic oxygen of the Tyr411 residue is due to the close proximity of the Arg410 guanidine moiety that electrostatically stabilizes the reactive anionic form of the Tyr411 residue [5,14]. Since both the Arg410 and Tyr411 residues are placed in the FA3-FA4 site (i.e., Sudlow's site II), ligand binding inhibits the HSA esterase activity [3,9,12,13]. Remarkably, the esterase activity of HSA could play a role in the inactivation of several toxins including organophosphorus compounds [3].
Present study largely extends previous investigations concerning the hydrolysis of 4-nitrophenyl esters by HSA [9,[12][13][14]. In particular, kinetics of the HSA pseudo-enzymatic hydrolysis of 4-nitrophenyl hexanoate (NphOHe) and 4-nitrophenyl decanoate (NphODe) have been investigated between pH 5. 8  . The rationale behind this selection is to investigate how the FA tail length affects the pK a values of the ionizing group that modulates the catalysis. Furthermore, diazepam, diflunisal, ibuprofen, 3-indoxyl-sulfate, and propofol have been reported to inhibit competitively the HSA-Tyr411-catalyzed hydrolysis of NphOHe, NphODe, and 4-nitrophenyl myristate (NphOMy) (see [12] and present study). Remarkably, the molar fraction of diazepam, diflunisal, ibuprofen, and 3-indoxyl-sulfate bound to not acylated HSA is higher than 0.9 whereas the molar fraction of propofol bound to HSA is ca. 0.5. the "dead-time" of the rapid-mixing stopped-flow apparatus). Moreover, the rate of NphOH release from NphOHe and NphODe catalyzed by HSA-Tyr411 is unaffected by the addition of NphOH (up to 1.0×10 -4 M) in the reaction mixtures (data not shown), indicating that the acylation step is essentially irreversible. If NphOH had affected the HSA-Tyr411 catalyzed hydrolysis of NphOHe and NphODe, the classical product (i.e., NphOH) inhibition behavior would have been observed.
When Values of K s and k +2 for the HSA-Tyr411-catalyzed hydrolysis of NphOHe and NphODe (see Table 1) were obtained from the hyperbolic plots of k app as a function of the HSA concentration ( Fig. 2, panels C and D) according to equation (2) [9,12,13]: , the reaction of NphOHe and NphODe with HSA displays a mono-exponential time course (Fig. 2, panels E and F). Values of the pseudo-first-order rate constant for the HSA-Tyr411-catalyzed hydrolysis of NphOHe and NphODe (i.e., of NphOH release; k obs ) were obtained according to equation (3) [9,12,13]: When k +2 ! 5×k +3 , the differential equations arising from Fig. 1 may be solved [12,13,16,17] to describe the time course of NphOH release at the early stages of the reaction. The resulting expression is given in eqs (4)-(6) [12,13,16,17]: The minimum three step-mechanism for the HSA-Tyr411-catalyzed hydrolysis of NphOHe, NphODe, and NphOMy. HSA is the substrate-free protein, NphOXx is the substrate, HSA:NphOXx is the reversible protein-substrate complex, HSA-OXx is considered to be an ester formed between the acyl moiety of the substrate and the O atom of the Tyr411 phenoxyl group [14], XxOH is hexanoate or decanoate or myristate, K s is the pre-equilibrium constant for the formation of the HSA:NphOXx complex, k +2 is the first-order acylation rate constant, and k +3 is the first-order deacylation rate constant. Xx indicates Ac or He or De or My. doi:10.1371/journal.pone.0120603.g001 and As predicted from eqs (4) Table), indicating that the HSA: NphOXx:NphOH stoichiometry is 1:1:1. Moreover, the time course of the "burst" phase of NphOH release is a first-order process for more than 95% of its course (  Table 1) were determined from hyperbolic plots of k obs versus [NphOXx] (Fig. 2, panels G and H) according to equation (6) [16,17]. Under all the experimental conditions, the y-intercept of the hyperbola described by equation (6) was < 2×10 -6 s -1 , thus indicating that the value of k +3 is at least 100-fold smaller than that of k obs obtained at the lowest NphOHe and NphODe concentration (i.e., k +3 < 2×10 -6 s -1 ).
According to linked functions [12,13,[16][17][18], the pH dependence of K s reflects the acidic pK a -shift of a single amino acid residue from free HSA (i.e., pK unl ) to the HSA:NphOHe and HSA:NphODe complexes (i.e., pK lig ). Moreover, the pH dependence of k +2 and k +2 /K s reflects the acid-base equilibrium of one apparent ionizing group in the HSA:NphOHe and HSA: NphODe complexes (i.e., pK lig ) and in free HSA (i.e., pK unl ), respectively. As expected [12,13,[16][17][18], the pK a value of free HSA (i.e., pK unl ) is independent of the substrate whereas the pK a values of the HSA:NphOHe and HSA:NphODe complexes (i.e., pK lig ) depend on the substrate ( Table 2).

Discussion
The hydrolysis of NphOAc, NphOHe, NphODe, and NphOMy by HSA-Tyr411 (see [12][13][14][15] and present study) is reminiscent of that observed for acylating agents with proteases [27]. In fact, NphOAc [13], NphOHe (present study), NphODe (present study), and NphOMy [12] act as suicide substrates of HSA-Tyr411, values of the deacylation rate constant for all four substrates (i.e., k +3 ) being lower by several orders of magnitude than those of the acylation rate constant (i.e., k +2 ). Remarkably, HSA acylation appears to modulate ligand binding. In fact, HSA acylation by aspirin [28] increases the affinity of phenylbutazone and inhibits bilirubin The PDB ID codes of HSA:diazepam, :diflunisal, :ibuprofen, :3-indoxyl-sulfate, and :propofol complexes are 2BXE, 2BXF, 2BXG, 2BXH, and 1E7A, respectively [8,30]. The pictures were drawn with the UCSF-Chimera package [46]. For details, see text. doi:10.1371/journal.pone.0120603.g006 Inhibition of the Esterase-Like Activity of HSA and prostaglandin binding, thus accelerating the clearance of prostaglandins, which represents an additional mechanism of the aspirin anti-inflammatory action [29]. Kinetics for the hydrolysis of NphOHe and NphODe by HSA are pH dependent, reflecting the acidic pK a shift of an apparently single ionizing group of HSA upon substrate binding. This could reflect the reduced solvent accessibility of the Tyr411 residue, representing the primary esterase site of HSA (see [9,[12][13][14]), although long range effects could not be ruled out. The Tyr411 catalytic residue is located in the FA3-FA4 cleft, which is made by an apolar region forming the FA3 site and a polar patch contributing the FA4 site. The polar patch is centered on the Tyr411 side chain and includes Arg410, Lys414, and Ser489 residues [8,30,31]. The inspection of the three-dimensional structure of the ligand-free HSA [32] and of the molecular model of the HSA:4-nitrophenyl propionate complex [9] suggests that the observed pH effects (Fig. 3) could reflect the acidic pK a shift of the Tyr411 residue upon substrate binding. This would render more stable the negative charge on the phenoxyl O atom of Tyr411, which appears to be hydrogen bonded to the carbonyl O atom of 4-nitrophenyl propionate [9], potentiating its nucleophilic role as an electron donor in the pseudo-esterase activity of HSA. Of note, the acidic shift of the pK a value of the ionizing group affecting catalysis from 8.9±0.1 in ligand free-HSA to 8.1±0.2, 7.6±0.2, 7.0±0.2, and 6.8±0.2 in the HSA:NphOAc, HSA:NphOHe HSA: NphODe, and HSA:NphOMy complexes (see Table 2), respectively, depends on the length of the fatty acid tail. Therefore, it appears as the increase of the FA tail length brings about the progressive reduction of the water solvent accessibility, thus enhancing the hydrophobicity of the catalytic site and leading to a decreased pK a of the ionizing group modulating the catalysis. Of note, the pH dependence of the Tyr411-associated esterase activity parallels the pH-dependent neutral-to-basic allosteric transition of HSA [3].
Diazepam, diflunisal, ibuprofen, 3-indoxyl-sulfate, and/or propofol inhibit competitively the hydrolysis of NphOAc [13], NphOHe (present study), NphODe (present study), and NphOMy (see [12] and present study) (Fig. 5) by impairing the accessibility of 4-nitroplenyl esters to the Tyr411 catalytic center. In particular, diazepam, diflunisal, ibuprofen, and 3-indoxyl-sulfate bind to the center of the FA3-FA4 cleft, with one O atom being hydrogen bonded to the Tyr411 OH group (Fig. 6). On the other hand, propofol binds to the apolar region of the FA3-FA4 cleft with the phenolic OH group making a hydrogen bond with the carbonyl O atom of Leu430. Moreover, the aromatic ring of the propofol is sandwiched between the Asn391 and Leu453 side chains. Furthermore, one of the two isopropyl groups of propofol makes several apolar contacts at one end of the pocket, whereas the other is solvent exposed at the cleft entrance making close contacts with Asn391, Leu407, Arg410, and Tyr411 (Fig. 6) [8,30,31].
The different K I values for diazepam, diflunisal, ibuprofen, 3-indoxyl-sulfate, and propofol binding to HSA (Fig. 5) agree with the selectivity of the FA3-FA4 cleft of HSA, which can be ascribed to the presence of a basic polar patch located at one end of the apolar FA3-FA4 cleft. Remarkably, diazepam, diflunisal, ibuprofen, and 3-indoxyl-sulfate are oriented with at least one O atom in the vicinity of the polar patch. On the other hand, the single polar hydroxyl group in the center of propofol does not interact with the polar patch of the FA3-FA4 cleft. Moreover, the FA3-FA4 cleft appears to adopt different ligand-dependent shapes, thus paying different free energy contributions for structural rearrangements [8].
Since the plasma levels of diflunisal, ibuprofen, and 3-indoxyl-sulfate (see above) are higher than values of K I for ligand binding to HSA by about 100 folds (see Fig. 4 and Fig. 5), the molar fraction of diflunisal, ibuprofen, and 3-indoxyl-sulfate bound to HSA is higher than 0.9, according to equation, (11). Although the commonly reported diazepam and propofol plasma levels (see above) are lower than the corresponding values of K I for drug binding to HSA (see Fig. 4 and Fig. 5) by about 5 and 100 folds, respectively, the plasma HSA concentration (see above) is higher than K I for by about 70 and 2 folds, respectively. According to equation (11), the molar fraction of diazepam and propofol bound to HSA in plasma is higher than 0.9 and 0.5, respectively.
As a whole, data here reported highlight the role of drugs diazepam, diflunisal, ibuprofen, and propofol as well as of the uremic toxin 3-indoxyl-sulfate to inhibit competitively the pseudo-esterase activity of HSA, Tyr411 representing the nucleophile. This aspect is appropriate since HSA acylation appears to modulate ligand binding [28,29] and the detoxification of several compounds [2,3]. Last, HSA not only acts as a carrier and as a detoxifier but also displays transient drug-and toxin-based properties, representing a case for "chronosteric effects" [45]. This opens the scenario toward the possibility of a drug-and toxin-dependent multiplicity of roles for HSA.