Effect of Bezafibrate and Clofibric Acid on the Spectroscopic Properties of the Nitric Oxide Derivative of Ferrous Human Hemoglobin

Paolo Ascenzi, Massimo Coletta, Alessandro Desideri, Francesca Polizio, Albert0 Bertollini, Roberto Santucci and Gino Amiconi

PA. Department of Pharmaceutical Chemistry and Technology, University of Turin.--MC. Department of Molecular, Cellular and Animal Biology, University of Camerino.--AD. Department of Organic and Biological Chemistry, University of Messina.--FP. Department of Biology, University of Rome “Tor Vergata”. -AB, RS, GA. C.N.R, Center for Molecular Biology Department of Biochemical Sciences, “Alessandro Rossi Fanelli, ” University of Rome “La Sapienza, ” Italy

ABSTRACT

The effect of bezafibrate (BZF) and clofibric acid (CFA) on the spectroscopic CEPR and absorbance) properties of the nitric oxide derivative of ferrous human hemoglobin (HbNO) has been investigated quantitatively. In the presence of BZF and CFA, the X-band EPR spectra and the absorption spectra in the Soret region of HbNO display the same basic characteristics described in the presence of inositol hexakisphosphate (IHP) and 2, 3-diphosphoglycerate (2, 3-DPG). Next, in the presence of these allosteric effecters, the oxygen affinity for ferrous human hemoglobin (Hb) is reduced. These findings indicate that BZF and CFA, as already reported for IHP and 2, 3-DPG, induce the stabilization of a low affmity conformation of the ligated hemoprotein (i.e., HbNO). Values of the apparent equilibrium constant for BZF and CFA binding to HbNO (K) are 1.5( kO.2) x lo-* M and 2.8( &0.3) x lo-* M, respectively, at pH 7.0 (in 0.1 M N-[2-hydroxyethyl]piperazine-N’-[Zethanesulfonic acid]/NaOH buffer system plus 0.1 M NaCl) and 20°C. The results reported here represent clearcut evidence for BZF and CFA specific (i.e., functionally relevant) binding to a ligated derivative of Hb (i.e., HbNO).

ABBREVIATIONS

Hb, ferrous human hemoglobin; HbNO, nitric oxide derivative of ferrous human hemoglobin; BZF, bezafibrate; CFA, clofibric acid; IHP, inositol hexakisphosphate; 2, 3-DPG, 2, 3-diphosphog- lycerate.

Address reprint requests and correspondence to: Professor Paolo Ascenzi, C.N.R., Center for Molecular Biology, Department of Biochemical Sciences “Alessandro Rossi Fanelli,” University of Rome “La Sapienza,” Piazzale Aldo Moro 5, 00185 Rome, Italy.

Journal of Inorganic Biochemistry, 48,47-53 (1992) 47 0 1992 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, NY, NY 10010 0162-0134/92/$5.00

48 P. Ascend et al.

INTRODUCTION

In a similar manner to oxygen and carbon monoxide, nitric oxide binds to the ferrous form of oxygen carrying proteins at the sixth coordination position of the heme group [see Refs. l-41. Unlike the diamagnetic oxy- and carbonmonoxy- heme complexes, the electronic structure of the heme-NO adduct may be detected by EPR spectroscopy [see Ref. 51. Two types of the X-band EPR spectrum have been observed for HbNO: the first type is characterized by a rhombic shape and a weak hyperflne pattern in the g, region; the second type shows a three-line splitting (A, = 1.65 mT) in the high magnetic field region (g, = 2.01) [see Refs. 2, 51. Such X-band EPR features parallel absorbance changes in the Soret region [see Ref. 21. The transition from the first to the second type of the X-band EPR spectrum and of the absorption spectrum in the Soret region may be induced in the nitrosylated derivative of ferrous tetrameric hemoglob~s (i.e., HbNO), at neutral pH by addition of allosteric effecters (i.e., IHP and 2, 3-DPG) [see Ref. 21. This finding suggests a close correlation between polyphosphate binding to the tetramer and the stabilization of a low afhnity ~nfo~ation of the ligated hemoprotein [see Ref. 21.

The low affinity state of Hb may be induced not only by IHP and 2, 3-DPG but also by BZF and CFA, the latter compounds displaying antigelling and ~tih~rli~proteinemia activities. These drugs bind reversibly to Hb in the red blood cell without damaging the cell membrane, decreasing the oxygen affmity of red cell suspensions and Hb solutions; next, BZF and CFA act synergistically with IHP and 2, 3-DPG. BZF and CFA interact preferentially with the deoxy- genated derivative of Hb at several binding sites, different from the heme pocket as well as the cleft(s) where IHP and 2, 3-DPG associate [see Refs. 6-101.

In order to further characterize the molecular bases of the BZF and CFA action on the heme ligand binding properties of Hb, the effect of these two drugs on the EPR and absorbance spectroscopic properties of HbNO has been investigated at pH 7.0. As observed for BZF binding to human methemo~obin derivatives [see Ref. 111, BZF and CFA association to HbNO induces the stabilization of a low affinity conformation of the tetramer, reminiscent of that induced by IHP and 2, 3-DPG [see Ref. 21. The reported results represent a clearcut evidence for BZF and CFA specific (i.e., functionally relevant) binding to a ligated derivative of Hb (i.e., HbNO).

MATERIALS AND METHODS

Oxygenated Hb samples were prepared as previously reported Ill, and stripped of any cation and anion (i.e., allosteric effector) bound to the protein by passing the hemolysate first through a Sephadex G-25 column, and afterwards through a column of mixed-bed ion exchange resin (Bio Rad AG-501 X8) 1121. Hb ~n~nt~tion was determined on the basis of e = 14.6 mM_’ cm-’ (at 577 nm), for the oxygenated derivative [l]. The millimolar absorption (rnM_’ cm-‘) is expressed as e on the millimolar heme basis.

EFFECT OF BZF AND CFA ON HUMAN HbNO 49

HbNO was obtained under anaerobic conditions by sequential addition of sodium dithionite and potassium nitrite (final concentrations 15 mg/ml and 5 mg/ml, respectively) to the oxygenated hemoprotein solutions [13l.

BZF, CFA, and N-[2-hydroxyethyllpiperazine-N’-[2-ethanesulfonic acid] were obtained from Sigma Chemical Co. (St. Louis, MO, U.S.). All other products were purchased from Merck AG (Darmstadt, FRG). All chemicals were of analytical grade and were used without further purification.

BZF and CFA binding to HbNO was followed by EPR and absorbance spectroscopy. Values of the equilibrium constant for BZF and CFA association to HbNO (K) were determined at 20°C from absorbance changes accompanying allosteric effector binding [see Refs. 13, 141.

Gxygen binding to Hb was followed by absorption spectroscopy [l]. Values of PW for oxygen association to Hb, in the absence and presence of BZF (2.0 x lo-* M) and CFA (2.0 X lo-* M) were determined from absorbance changes accom- panying the dioxygen molecule binding by the tonometric method [l].

Values of the sedimentation coefficient for HbNO in the absence and in the presence of BZF (1.0 X 10-l M) and CFA (1.0 X 10-l M), were obtained at 7.O”C with a Spinco model E analytical ultracentrifuge at 52,000 rev/min; the HbNO concentration was 7.5 x 10m5 M.

X-band EPR spectra of HbNO in the absence and presence of BZF (1.0 x 10-l M) and CFA (1.0 X 10-l M) were collected at 100 K on a BRUKER ESP 300 spectrometer.

Absorption spectra of the deoxygenated, nitrosylated, and oxygenated I-lb in the absence and presence of BZF (from 2.4 X 10e4 M to 1.0 x 10-l M) and CFA (from 3.8 X 10m4 M to 1.0 X 10-l M) were recorded at 20°C on a VARIAN Gary 219 spectrophotometer.

All data were obtained at pH 7.0 in 0.1 M N-[2-hydroxyethyllpiperazine-N’- [Zethanesulfonic acid]/NaOH buffer system plus 0.1 M NaCl.

RESULTS AND DISCUSSION

Figure 1 shows the chemical structures of BZF and CFA, both displaying antigelling and antihyperlipoproteinemia activities [see Refs. 6-101.

y3

0-C-COOH

AH3

BZF

CFA FIGURE 1. Chemical structures of BZF and CFA

50 P. Ascenzi et al.

FIGURE 2. X-band EPR spectra (left panel) and absorbance spectra in the Soret region (right panel) of HbNO in the absence (A), and in the presence of BZF (1.0 X 10-l M; B) and CFA (1.0 x 10-l M; C). X-band EPR spectra and absorbance spectra in the Soret region were obtained at pH 7.0 in 0.1 M N-[2-hydroxyethyl] piperazine-N’-[2- ethanesulphonic acidl/NaOH buffer system

A

q plus 0.1 M NaCl at 100 K and 20°C

-_ respectively. The concentration of HbNO k 100 0

ranged between 7.5 x 10e5 M and 1.0 x B

-i 5

10e3 M. Setting conditions for X-band EPR

Ix 50 spectra: 9.42 GHz microwave frequency; 20

c B mW microwave power; 0.10 mT modulation 0” amplitude. Absorption spectra in the Soret

region were recorded in l-mm length cu- 0

320 340 400 440 vettes. For additional experimental details, Magnetic Field lmTl I. (nml see text.

Figure 2 shows the X-band EPR spectra and the absorption spectra in the Soret region of HbNO in the absence and presence of BZF (1.0 x 10-l Ml and CFA (1.0 X 10-l M) at pH 7.0.

In the absence of any allosteric effector, the X-band EPR spectrum of HbNO, showing a rhombic shape and a weak hyperfine pattern in the g, region (see Fig. 2, left panel, spectrum A) has been associated with the high affinity state of the macromolecule [see Refs. 2, 51.

Addition of BZF and CFA to HbNO induces a transition toward a species characterized by an X-band EPR spectrum with a three-line splitting (A, = 1.65 mT) in the high magnetic field region (g, = 2.01) (see Fig. 2, left panel, spectra B and C, respectively). Such a pattern resembles that of HbNO in the presence of IHP and 2,3-DPG which has been attributed to the polyphosphate-induced shift of the conformational equilibrium toward a low affinity state of the ligated hemoprotein [see Refs. 2, 51.

Parallel to the BZF- and CFA-induced EPR spectroscopic transition (Fig. 2, left panel), the intensity of the absorbance band in the Soret region of HbNO decreases significantly upon binding of these drugs (Fig. 2, right panel; and Fig. 3). Such a spectral change (Aeobs = 30(+ 1) mM-’ cm- ‘, at 416.5 nm; corre- sponding to E* in Eq. (1)) is closely similar to that observed for polyphosphate association to HbNO (AeEobs = 30( + 1) mM_’ cm- ’ at 416.5 nm) [see Ref. 21.

The reversible transition brought about by BZF and CFA, as well as by IHP and 2, 3-DPG [see Refs. 2,5 for reviews], is consistent with either a perturbation of the iron 3d orbital energy, or with the cleavage (or the severe weakening) of the proximal His(F8) NE2-Fe bond [see Ref. 51. However, the latter interpreta- tion may be preferentially invoked, being supported by Raman and infrared spectroscopic data of the nitrosylated derivative of ferrous hemoproteins and heme-model compounds [15-171.

Figure 3 shows the ligand binding isotherms for BZF and CFA association to HbNO at pH 7.0 and 20°C. Applying the minimum model accounting for an

EFFECT OF BZF AND CFA ON HUMAN HbNO 51

FIGURE 3. Dependence on the allosteric effector (BZF, 0; CFA, 0) concentration (logIL]; MI of the optical density (%,J of HbNO at 416.5 nm. The symbol on the left ordinate (A) indicates the value of cobs (‘E, + E*; see Eq. (1)) obtained in the absence of BZF and CFA. The concentra- tion of BZF and CFA is that of the free ligand [see Refs. 13, 141. The continuous lines were generated from Eq. (1) with the following parameters: l 1 = lOl( f 3) mM_’ cm-‘, 9=30(*1) mM_’ cm-‘, K= 1.5( &0,2) x 10m2 M for BZF and K = 2.g f 0.3) x lo-* M for CFA. The parame- ters were obtained with an iterative nonlin- ear least-squares curve fitting procedure. The data were obtained at pH 7.0 in 0.1 M N-[2-hydroxyethyllpiperazine-N’-[2- ethanesulphonic acid]/NaOH buffer sys- tem plus 0.1 M NaCl at 20°C. The concen- tration of HbNO was 7.5 x lop5 M. For additional experimental details, see Figure 2 and text.

apparent single binding site per tetramer, a relation between the BZF and CFA apparent equilibrium constant for HbNO (K) and the extinction coefficient of the system (E,,,~ = E, + l 2) may be expressed according to Eq. (1) [see Refs. 13, 141:

E ohs = El + 9’ {l/(1 + WKN (1) where [L] is the ligand (i.e., BZF or CFA) concentration, e1 is the extinction coefficient of HbNO fully saturated with the allosteric effecters, and l 2 ( = A e,& see above) corresponds to the BZF- and CFA-dependent extinction coefficient contributing to the optical density change at 416.5 nm. Equation cl), which describes as a function of BZF and CFA concentration the extinction coefficient of HbNO in the absence (eobS = l 1 + e2) and presence (E,,~~ < e1 + l 2) of the allosteric effecters, has been used to generate the continuous lines shown in Figure 3; the agreement with the experimental data is fully satisfactory, giving us confidence that the correct assumptions underlying Eq. (1) have been made. The simple behavior shown in Figure 3 is also indicative of the absence of BZF- and CFA-induced dimerization of HbNO (even at the highest drug concentration used, 1.0 X 10-l MI. This possibility has also been ruled out on the basis of values of the sedimentation coefficient ( =4.0 f 0.1) for HbNO, obtained in the absence and presence of BZF (1.0 x 10-l MI and CFA (1.0 X 10-l M), which are indeed typical of a tetrameric state [ll. Next, the simple profile of cobs vs log[Ll (see Fig. 3) suggests the absence of significant amounts of denatured HbNO induced by BZF and CFA addition (even at the highest drug concentra- tion used, 1.0 X 10-l MI.

Although analysis of -the data does not allow discrimination between the presence of a single ligand binding site per tetramer or multiple independent and equivalent BZF and CFA association clefts, it becomes clear that these

52 I? Ascenn’ et al.

drugs associate specifically (i.e., in a functionally relevant way) to HbNO. At variance with what was observed for specific drug association to the deoxy- genated derivative of Hb [see Ref. 101, no evidence for multiple classes of binding sites for BZF and CFA has been observed in I-&NO. In the absence of structural data for the HbNO:BZF and :CFA complexes, a definite conclusion on the location of the functionally relevant drug binding sites cannot be drawn. Nevertheless, a possible candidate suggested by literature data [see Refs. 6-101 might be the central cavity between one @-chain and the two a-subunits [see Ref. 81.

As reported for IHP and 2, 3-DPG binding to Hb [see Refs. 2, 13, 141, values of K for BZF and CFA binding to HbNO (l.S( f 0.2) X lo-* M and 2.8( f 0.3) x lo-* M, respectively; see Fig. 3) are lower than those observed for drug association to the deoxygenated derivative of Hb, which associates specifically to BZF and CFA at two classes of binding sites (BZF: K, = 1.6 x low3 M and K, = 7.7 x 10e3 M; and CFA: K, = 3.4 x 10e3 M and K, = 1.7 X lo-* M; 5.0 x lo-* M phosphate buffer system plus 0.1 M NaCl, pH = 7.4; and 4°C) [see Ref. 101. Next, as also reported for drug binding to the deoxygenated derivative of Hb [see Ref. 101, the affinity of BZF for HbNO is higher than that observed for CFA (see Fig. 3).

The effect of BZF and CFA on the EPR and absorbance spectroscopic properties of HbNO is in keeping with the observed functional behavior for oxygen binding. Thus, under the same experimental conditions, Hb displays an oxygen affinity (Pm = 4.3 mm Hgl higher than that shown in the presence of BZF (2.0 X lo-* Ml and CFA (2.0 X lo-* M) (Pm = 22 mm Hg and P5,, = 25 mm Hg, respectively) [see also Refs. 6-101.

The close correlation between the spectroscopic, functional and structural properties displayed by Hb indeed represents an important confirmation for the use of EPR and absorbance spectroscopy in the elucidation of structure-func- tion relationships in oxygen carrying proteins.

The authors thank Professor Paok Vecchini for the sedimentation epetiments. This work was supported by grunts from the Italian Ministry for Vniuersity, Scienti~ Research and Tecnology (Minktem per I’Vniuersittri e la Ricerca Scknti$ca e Tecnologica), as well as the Italian National Research Council (Consiglio Nazionale a’elk Rkenzhe).

REFERENCES

1. E. Antonini and M. Brunori, in Hemoglobin and Myoglobin in their Reactions with Ligandr, A. Neuherger and E. L. Tatum, Eds., North Holland, Amsterdam, 1971.

2. M. F. Perutz, Annu. Rev. Bib&em. 48,327 (1979). 3. M. Brunori, P. Ascenzi, and M. Coletta, in Pnxeedinp of the International Cong~ss

on Supermolecules: Biological and Chemical Aspects, Rome, 9 November 1984, pp. 55-73, Accademia Nazionale dei Lincei, Roma, 1986.

4. M. Brunori, M. Coletta, P. Ascenzi, and M. Bolognesi, .I. Mol. Liquids 42,175 (1989). 5 W. E. Blumberg, Method Etuymol. 76,312 (1981). 6. D. J. Abraham, M. F. Perutz, and S. E. V. Phillips, tkc. Natl. Acad. Sci. USA 80,

324 (1983). 7. D. J. Abraham, P. E. Kennedy, A. S. Mehanna, D. C. Patwa, and F. L. Wiiiams, J.

Med. Chem. 27, %7 (1984).

EFFECT OF BZF AND CFA ON HUMAN HbNO 53

8. M. F. Perutz, G. Fermi, D. J. Abraham, C. Poyart, and E. Bursauz, J. Am. Chem. sot. 10$1064 (1986).

9. I. Lalezari, P. LaIezari, C. Poyart, M. Marden, J. Kister, B. Bohn, G. Fertni, and M. F. Perutz, Biochemistry 29,1515 0990).

10. A. S. Mehanna and D. J. Abraham, Biochemi&y 29,3944 (1990). 11. R. W. Noble, A. De Young, S. Vitale, M. Cerdonio, and E. Di Iorio, BiocZrem&y 28,

5288 (1989). 12. A. Riggs, Methou% Enzymol. 76,5 (1981). 13. P. Ascenzi, A. Desideri, G. Amiconi, A. Bertollini, M. Bolognesi, M. Castagnola, M.

Coletta, and M. Brunori, J. Znorg. Biochem. 34,19 (1988). 14. P. Ascenzi, R. Santucci, A. Desideri, and G. Amiconi, J. Zrwrg. Biochem. 32, 225

(1988). 15. B. B. Wayland and L. W. Olson, J. Am. Chem. kc. !%, 6037 (1974). 16. J. C. Maxwell and W. S. Caughey, Biochemistry 15,38 (1976). 17. K. Nagai, C. Welbom, D. Dolphin, and T. Kitagawa, &&em&y 19,4755 (1980).

Received November 1, 1991; accepted January 28, 1992