- Email: [email protected]
Chloride Ion Nuclear Magnetic Resonance Spectroscopy Probe Studies of Copper and Nickel Binding to Serum Albumins
Chloride Ion Nuclear Magnetic Resonance Spectroscopy Probe Studies of Copper and Nickel Binding to Serum Albumins P. MOHANAKRISHNAN*, C. F. CHIGNELL*',AND R. H. Cox* Received July 28, 1983, from the *Laboratoryof Environmental Biophysics and the *Laboratory of Environmental Chemistry, National institute of Environmental Health Sciences, Research Triangle Park, NC 27709. Accepted for publication May 14, 1984. Abstract 0 The binding of Cup+and Ni2+ to bovine, dog, and human serum albumin has been studied by the 35CINMR probe technique. The number of primary copper sites were estimated to be 1.3 for human
serum albumin, 3.1 for bovine serum albumin, and 6.6 for dog serum albumin. A similar number of primary nickel sites was determined for each of these albumins. On the basis of the chloride probe experiments, it appears that both copper and nickel have the same binding sites on albumin.
Cu2+and Ni2+to bovine (BSA), dog (DSA), and human (HSA) serum albumins using the chloride ion probe technique.
Experimental Section
Fatty acid-free HSA and DSA (fraction V) were obtained from Sigma Chemical Co. The BSA (fraction V, Miles Laboratories, Inc.) and DSA were defatted according to the procedure of Chen17 and passed through a Pharmacia Sephadex GAlbumin is the main carrier of various metal ions and small 150 column (100 X 3.4 cm) with 0.2 M NaCl as the eluting molecules in the blood serum. Brown' has proposed, on the medium at 4°C. The fractions collected were concentrated by basis of amino acid sequence and hydrodynamic properties of ultrafiltration (Amicon UM-10) and dialyzed against a large bovine and human albumin, that this protein consists of three volume of 5 mM EDTA followed by dialysis against large domains which, in turn, have subdomains consisting of helices volumes of deionized water (three changes of 12 h each) a t 4°C. that are stabilized by disulfide bridges. Amino acid-sequence After freeze-drying, the proteins were stored a t 4°C until use. studies of albumin from different animal species have identified HSA was also subjected to dialysis, and subsequent steps of the N-terminal portion of the protein as the major copper treatment, in the same manner as BSA and DSA. binding site. At this site the copper is complexed to the N Experimental solutions were made by dissolving a known terminal amino group, the next two peptide nitrogens, and the amount of albumin in 0.15 M NaCl (50% D20) to give a final imidazole nitrogen of histidine, the third amino acid from the protein concentration of 5-50 pM. The BSA and HSA concenN-terminal.2-5 trations were also estimated from their absorbances (A:;: = 5.31 for HSA and 6.67 for BSA). The metal ion concentrations The binding of copper to serum albumin is important because of the possible involvement of the metal in several pathological (5-50 mM) in Cu2+ and Ni2+ stock solutions were estimated conditions? Nickel, implicated as a carcinogen in humans and using atomic absorption spectroscopy. As in our previous experimental animal^,^ has been reported to bind to the Nstudies'' no p H adjustment of the protein solutions was made terminal copper site of dog and human serum a l b ~ m i n s . 8 . ~ for titrations with Cu2+,since there is a likelihood for formation Presumably, copper and nickel form the same type of complexes of 0x0 or hydroxo complexes (soluble and/or insoluble) of Cu2+ a t pH values >6.5. Typically, the experimental p H for titrations with the N-terminal site of albumin. Despite the similarities with Cu2+was -5.4. On the other hand, a p H of 10.4 was used among serum albumins from different animal species, there are for Ni2+titrations since it has been reported that the peptide differences in the affinities of these proteins for metal ions. For example, the affinity for both Cu2+and Ni2+ is less for dog hydrogen ionization is not complete for Ni2+until pH 10." serum albumin than for human serum albumin because of the The 3sCl NMR spectra were recorded on a Varian FT-80A presence of a tyrosine instead of a histidine at the third amino spectrometer at 7.794 MHz and 25°C. The deuterium signal acid position from the N-terminal of dog albumin?-'' This may from the solvent HDO was used as the internal lock. Conditions partly explain why the dog is more susceptible to copper poitypically used to acquire the spectra were as follows: a sweep soning than humans. Thus, it is of great importance to study width of 1 KHz, the number of data points was 2000, a n acquisition time of 1.0 s, and a sensitivity enhancement line the binding of copper and nickel to different albumins in order broadening factor of -0.4. A 60" flip angle was used without to understand many pathological events at the molecular level. any delay between pulses. Experiments with 0.15 M NaCI, with In the series of studies described here, we have used the "Cl no addition of metal or protein, indicated that no line broadanion as a probe of the serum albumin metal ion binding sites. Although a large excess concentration of chloride ions relative ening was introduced under the above experimental conditions. Each spectrum was the result of 1000 or more accumulations. to protein is used in these studies, the NMR parameters of the The titrations were carried out by adding Cu2+and Ni2+ stock former are perturbed as a consequence of chemical exchange solutions in 1-pL amounts using a Hamilton syringe to 2-2.5 with binding sites on the protein. This technique has been used mL of protein solution (0.15 M in NaCl) in 10-mm sample to monitor the binding of metal ions, anions, and small moletubes under nitrogen. cules to proteins such as hemoglobin, alcohol and lactate deSpectra were plotted using a width of a few hundred hertz hydrogenases, and alkaline phosphatase." Halide ion NMR has and the line widths were measured as the widths at half the also been used previously to probe bovine and human albumin.12-1s Sudmeier and Pesek have studied the binding of Ca2+, heights of the experimental spectra. As some experimental spectra were somewhat noisy, Lorentzian lines were simulated Hg", Mn2+,Cu'+, and Zn2+to bovine serum albumin a t various on a Nicolet 1180 computer and were fitted to the experimental pH values using the chloride NMR probe technique.I6 In this spectra. The line widths used for the binding studies were those study, we present the results of our studies of the binding of Journal of Pharmaceutical Sciences 1 61 Vol. 74, No. 1, January 1985
Figure 1-%I
NMR spectrum (solid curve) of 0.15 M NaCl containing 15.8 WMBSA and 65.88 pM CuSO,. Broken line is the computer-simulated spectrum.
of best fit from the simulations. A typical experimental and simulated spectrum are shown in Fig. 1. Interpretation of the data is based on a model assuming the reversible complexation of an anion I- with a protein, P:
I-
+ P * p.1-
(1)
Relaxation and chemical shift parameters for a spin 'h nucleus undergoing exchange as described by eq. 1 in the limit in which the concentration of the uncomplexed ion greatly exceeds that of the complexed ion have been derived by Swift and Connick" and by Luz and Meiboom:"
width is greater at higher protein concentrations than a t lower protein concentrations. This is in agreement with previous observations by Sudmeier and Pesek" and Norne et al.13 who have reported that, in the absence of metal ions, excess line width increases with protein concentration. Norne et al. have interpreted this observation to mean that the number of chloride ion binding sites and the rate of relaxation are both independent of protein concentration. In the absence of protein, the increase in line width with metal ion concentration is much smaller for Ni2+than for Cu'+. This may reflect the kinetics of metal-chloride ion exchange or the longer transverse relaxation ti.me, Tle, for Cu2+. Alternatively, this may be due to the formation of more symmetric complexes by Ni2+which results from the absence of Jahn-Teller distortion in this metal ion. A scheme involving minor but more rapidly exchanging aquochloro complexes of Ni2+has been proposed previouslyz2on the basis of 35Cland 170NMR studies. The increase in line width as a function of the ratio of the concentration of Ni2+ to the concentration of protein for the three albumins studies is shown in Fig. 3. The experiments were performed (not shown) at two different concentrations for n
(3)
where fB is the fraction of enzyme-complexed ion; fB = [P.I]/ [I]; TIF,T pF, and bF are parameters corresponding to the uncomplexed ion; T 1 B and TSRcorrespond to the complexed ion; 7 is the lifetime of the complex; and A is the chemical shift difference between the free and complexed ion. The above equation can be simplified for a variety of limiting exchange cases. Extension for the case of an ion with spin >?h, such as chloride, has been described by Bullz1and involves the definition of approximate relaxation times, since the actual relaxation behavior of a quadrupolar nucleus is predicted to be nonexponential. A complete solution of the exchange problem requires that temperature-dependent data be obtained to evaluate the exchange kinetics, i.e., fast, slow, or intermediate on the NMR time scale. However, binding site titration studies can be carried out based on the proportionality of the protein perturbation terms in eqs. 2-4 to the ratio of bound to free ion fB. The only limit in which this approximation fails to hold is the case of exchange which is so slow that the binding has a negligible effect on the relaxation and shift parameters of the free species. This limit is readily detected since the addition of protein or high-affinity binding sites will have no effect on the observed parameters.
Figure 2-35Cl NMR spectrum of: (a) 0.15 M NaCl; (b) 0.75 M NaCl containing 235 WMNiC12; (c) 0.15 M NaCl containing 70 pM CuS04; (d) 0.15 M NaCI containing 15.8 WMBSA; (e) 0.15 M NaCl containing 43.8 pM BSA; (f) 0.15 M NaCl containing 43.8 gM BSA and 235 FM NiCQ; (g} 0.15 M NaCl containing 15.8 WMBSA and 49 pM CuS04.
c
I
16
f
U
'Z
I2
m
.-C
a
Results and Discussion Figure 2 illustrates the "C1 NMR line widths under different experimental conditions. There is a n increase in line width when either albumin or the metal ion, Cuz+or NiZ+,is added to 0.15 M NaC1. Also, in the absence of metal ions, the excess line 62
/ Journal of Pharmaceutical Sciences Vol. 74, No. 1, January 1985
[NI'+I/
Calbuminl
Figure 3-35Cl line broadening (Hz)versus the ratio of Niz+ion concentration to the albumin Concentration. Albumin concentrations are 28.82 pM for BSA (O),30.4 pM for DSA (m), and 49 pM for HSA (A).
"9 60
I 0
I 2
6
4
S
1 0 1 2 1 4
[Cu'+l /[albumin1
Figure 4-35CI line broadening (Hz) versus the ratio of Cu2+to albumin concentration. Albumin concentrations are 15.8 pM for BSA (m), 22.04 pM for DSA (A),and 10.23 pM for HSA (0). Table I-Number
Albumin
of Albumin Primary Binding Sites
Cu2+Sites
BSA
3.1
DSA
3.7a 6.6
HSA
1.3
Ni2+Sites
3.2
Affinities at Primary Sites (M-' x 10-6)
Cu2+
Ni2+
-
-
-
3.1 a
6.7 7.0b
-
1.5
-
-
the number of primary binding sites occur at the same position independent of whether the correction is applied. Our observation that the line width increases with increasing Cu2+ concentration after the primary sites of HSA and BSA are occupied does not agree with the findings of Sudmeier and Pesek." However, there are several differences between the two studies. First, Sudmeier and Pesek used albumin concentrations that were 10-100 times higher than those employed in our study. Second, their measurements were made over a wider pH range in acetate or phosphate buffers. Acetate ions may compete with chloride ions at pH values <6.5, while above pH 6.5 competition with hydroxyl ions may occur. Phosphate is not a suitable buffer ion for experiments involving copper since it forms insoluble cupric phosphate. The findings of the present study are summarized in Table I. For comparison, data from previous studies are also shown. The number of primary copper sites found in the present study for each of the three albumins is almost the same as the number of primary nickel sites. The number of primary copper sites for BSA and HSA from the present experiments is close to that obtained using a cupric ion-selective electrode." On the basis of the results from the present study, it appears that both copper and nickel have the same binding sites on albumin. This is in line with the earlier findings'-" that the lower affinity of DSA for Cu2+and Ni2+is due to the absence of a histidine at the third amino acid position from its N-terminal.
0.025'
-
1.1"
15.0b 2.3" 0.3b a Taken from ref. 18. Taken from ref. 23.Only one class of sites was detected.
each of the albumins. From the results of the titration experiments, the number of primary Ni2+ sites were estimated to be 1.5 for HSA, 3.2 for BSA, and 6.7 for DSA. Since the increase in line width with increasing Ni2+concentration is very small in the absence of protein, the increase in line widths in Fig. 3 have not been corrected for residual broadening due to any free Ni2+.The initial slope for DSA is lower than for either BSA or HSA at comparable protein concentrations. This may indicate that the primary nickel binding sites in DSA are different from those in the other two proteins. However, in the absence of complete characterization of the exchange kinetics the implication of this observation cannot be unequivocally determined. The results from the Cu2' titration experiments for the three different albumins are shown in Fig. 4. The increase in line widths observed when Cu2+is added to protein solutions have to be corrected for any paramagnetic/residual broadening. In Fig. 4 this was done by subtracting from the increase in line widths of albumin-Cu2+ solutions the increase in line widt.hs due to the same amount of Cuz+alone (i.e., in the absence of protein). The estimates for the number of Cu" sites obtained from Fig. 4 are 1.3 for HSA, 3.1 for BSA, and 6.6 for DSA. The initial 35Cl line broadening on the addition of Cu2+is very small for BSA (Fig. 4). However, there is a very small decrease for HSA and the initial slope is negative for DSA. The primary Cu2+ sites of BSA and HSA have affinities on the order of 106 M-1.3,20.22 At the protein concentrations used in these experiments, the free Cu2+ concentration is therefore likely to be extremely small, particularly a t low Cu2+/albumin ratios. Hence, for HSA and DSA, the free copper ion is more effective at relaxing the chloride ion than the enzyme-complexed ion. This may simply reflect the fact that more coordination positions are available for the aquo ion than for the protein-bound cation; the breaks in the curves which determine
References and Notes 1. Brown. J. R. in "Albumin: Structure. Function and Uses": Rosen-
2. 3. 4. 5.
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
oer, V.' M.; Oratz, M; Rothschild, M.A., Eds.; Pergamon Press: Oxford, 1977; pp 27-52. Shearer, W. T. J. Bid. Chem. 1967,242, 5451-5459. Peters., J... Jr.: Blumenstock, F. A. J. Biol. Chem 1967.242. . . 15741578. Bradshaw, R. A.; Shearer, W. T.; Gurd, F. R. N. J . Eiol. Chem. 1968,243, 3917-3825. Bradshaw, R. A.; Peters, J., Jr. J. Bid. Chem. 1969, 244, 55825589. Milanino, R.; Passerella, E.; Velo, G. P. in "Advances in Inflammation Research"; Weissman, G. D.; Samuelsson, B.; Raoletti, R. Eds.; Raven Press: New York, 1979; pp 281-291. Sunderman, F. W., Jr. Food Cosmet. Toricol. 1971,9, 109-120. Glennon, J. D.; Sarkar, B. Biochem. J. 1982,203, 15-23. Glennon, J. D.; Sarkar, B. Biochem. J . 1982,203, 25-31. Dixon, J. W.; Sarkar, B. Biochem. Biophys. Res. Commun. 1972, 48, 197-200. Lindman, B., Forsen, S. in "NMR Basic Principles and Progress"; Diehl, P.; Fluck, E.; Kosfeld, R. Eds.; Springer-Verlag: New York, 1976; pp 249-325. Gillberg-La Force, G.; Forsen, S. Biochem. Biophys. Res. Commun. 1970.38. 137-142. Nor& J.;Hjalmarsson, S.; Lindman, B.; Zeppezauer, M. Biochemi s t 1975,14, ~ 3401-3408. Halle, B.; Lindman, B. Biochemistry 1978,17, 3774-3781. Bull. T. E.: Halle. B.: Lindman. B. FEBS Lett. 1978. 86. 25-28. Sudmeier, J. L.; Pesek, J . J. Anal. Biochem. 1971,4i, 39-50. Chen, R. F. J. Biol Chem. 1967,242, 173-181. Mohanakrishnan, P.; Chignell, C. F. J . Pharm. Sci. 1982. 71, 1180-1182. Swift T. J.; Connick, R. E. J. Chem. Phys. 1962,37, 307-312. Luz, Z.; Meiboom, S. J. Chem. Phys. 1964, 40, 2686-2691. Bull, T. E. J. Magn. Reson. 1972,8, 344-353. Lincoln, S. F.; Aprile, F.; Dodgen, H. W.; Hunt, J. P. Znorg. Chem. 1968, 7, 929-932. Callan, W. M.; Sunderman, F. W., Jr. Res. Commun. C h m . Pathol. Pharmacol. 1973,5, 459-472. ,
Acknowledgments The authors wish to thank Drs. Frank Dearman, Robert Jordan, and Dallas Rabenstein for critically evaluating the manuscript.
Journal of Pharmaceutical Sciences Vol. 74, No. 7, January 7985
/ 63
serum albumin, 3.1 for bovine serum albumin, and 6.6 for dog serum albumin. A similar number of primary nickel sites was determined for each of these albumins. On the basis of the chloride probe experiments, it appears that both copper and nickel have the same binding sites on albumin.
Cu2+and Ni2+to bovine (BSA), dog (DSA), and human (HSA) serum albumins using the chloride ion probe technique.
Experimental Section
Fatty acid-free HSA and DSA (fraction V) were obtained from Sigma Chemical Co. The BSA (fraction V, Miles Laboratories, Inc.) and DSA were defatted according to the procedure of Chen17 and passed through a Pharmacia Sephadex GAlbumin is the main carrier of various metal ions and small 150 column (100 X 3.4 cm) with 0.2 M NaCl as the eluting molecules in the blood serum. Brown' has proposed, on the medium at 4°C. The fractions collected were concentrated by basis of amino acid sequence and hydrodynamic properties of ultrafiltration (Amicon UM-10) and dialyzed against a large bovine and human albumin, that this protein consists of three volume of 5 mM EDTA followed by dialysis against large domains which, in turn, have subdomains consisting of helices volumes of deionized water (three changes of 12 h each) a t 4°C. that are stabilized by disulfide bridges. Amino acid-sequence After freeze-drying, the proteins were stored a t 4°C until use. studies of albumin from different animal species have identified HSA was also subjected to dialysis, and subsequent steps of the N-terminal portion of the protein as the major copper treatment, in the same manner as BSA and DSA. binding site. At this site the copper is complexed to the N Experimental solutions were made by dissolving a known terminal amino group, the next two peptide nitrogens, and the amount of albumin in 0.15 M NaCl (50% D20) to give a final imidazole nitrogen of histidine, the third amino acid from the protein concentration of 5-50 pM. The BSA and HSA concenN-terminal.2-5 trations were also estimated from their absorbances (A:;: = 5.31 for HSA and 6.67 for BSA). The metal ion concentrations The binding of copper to serum albumin is important because of the possible involvement of the metal in several pathological (5-50 mM) in Cu2+ and Ni2+ stock solutions were estimated conditions? Nickel, implicated as a carcinogen in humans and using atomic absorption spectroscopy. As in our previous experimental animal^,^ has been reported to bind to the Nstudies'' no p H adjustment of the protein solutions was made terminal copper site of dog and human serum a l b ~ m i n s . 8 . ~ for titrations with Cu2+,since there is a likelihood for formation Presumably, copper and nickel form the same type of complexes of 0x0 or hydroxo complexes (soluble and/or insoluble) of Cu2+ a t pH values >6.5. Typically, the experimental p H for titrations with the N-terminal site of albumin. Despite the similarities with Cu2+was -5.4. On the other hand, a p H of 10.4 was used among serum albumins from different animal species, there are for Ni2+titrations since it has been reported that the peptide differences in the affinities of these proteins for metal ions. For example, the affinity for both Cu2+and Ni2+ is less for dog hydrogen ionization is not complete for Ni2+until pH 10." serum albumin than for human serum albumin because of the The 3sCl NMR spectra were recorded on a Varian FT-80A presence of a tyrosine instead of a histidine at the third amino spectrometer at 7.794 MHz and 25°C. The deuterium signal acid position from the N-terminal of dog albumin?-'' This may from the solvent HDO was used as the internal lock. Conditions partly explain why the dog is more susceptible to copper poitypically used to acquire the spectra were as follows: a sweep soning than humans. Thus, it is of great importance to study width of 1 KHz, the number of data points was 2000, a n acquisition time of 1.0 s, and a sensitivity enhancement line the binding of copper and nickel to different albumins in order broadening factor of -0.4. A 60" flip angle was used without to understand many pathological events at the molecular level. any delay between pulses. Experiments with 0.15 M NaCI, with In the series of studies described here, we have used the "Cl no addition of metal or protein, indicated that no line broadanion as a probe of the serum albumin metal ion binding sites. Although a large excess concentration of chloride ions relative ening was introduced under the above experimental conditions. Each spectrum was the result of 1000 or more accumulations. to protein is used in these studies, the NMR parameters of the The titrations were carried out by adding Cu2+and Ni2+ stock former are perturbed as a consequence of chemical exchange solutions in 1-pL amounts using a Hamilton syringe to 2-2.5 with binding sites on the protein. This technique has been used mL of protein solution (0.15 M in NaCl) in 10-mm sample to monitor the binding of metal ions, anions, and small moletubes under nitrogen. cules to proteins such as hemoglobin, alcohol and lactate deSpectra were plotted using a width of a few hundred hertz hydrogenases, and alkaline phosphatase." Halide ion NMR has and the line widths were measured as the widths at half the also been used previously to probe bovine and human albumin.12-1s Sudmeier and Pesek have studied the binding of Ca2+, heights of the experimental spectra. As some experimental spectra were somewhat noisy, Lorentzian lines were simulated Hg", Mn2+,Cu'+, and Zn2+to bovine serum albumin a t various on a Nicolet 1180 computer and were fitted to the experimental pH values using the chloride NMR probe technique.I6 In this spectra. The line widths used for the binding studies were those study, we present the results of our studies of the binding of Journal of Pharmaceutical Sciences 1 61 Vol. 74, No. 1, January 1985
Figure 1-%I
NMR spectrum (solid curve) of 0.15 M NaCl containing 15.8 WMBSA and 65.88 pM CuSO,. Broken line is the computer-simulated spectrum.
of best fit from the simulations. A typical experimental and simulated spectrum are shown in Fig. 1. Interpretation of the data is based on a model assuming the reversible complexation of an anion I- with a protein, P:
I-
+ P * p.1-
(1)
Relaxation and chemical shift parameters for a spin 'h nucleus undergoing exchange as described by eq. 1 in the limit in which the concentration of the uncomplexed ion greatly exceeds that of the complexed ion have been derived by Swift and Connick" and by Luz and Meiboom:"
width is greater at higher protein concentrations than a t lower protein concentrations. This is in agreement with previous observations by Sudmeier and Pesek" and Norne et al.13 who have reported that, in the absence of metal ions, excess line width increases with protein concentration. Norne et al. have interpreted this observation to mean that the number of chloride ion binding sites and the rate of relaxation are both independent of protein concentration. In the absence of protein, the increase in line width with metal ion concentration is much smaller for Ni2+than for Cu'+. This may reflect the kinetics of metal-chloride ion exchange or the longer transverse relaxation ti.me, Tle, for Cu2+. Alternatively, this may be due to the formation of more symmetric complexes by Ni2+which results from the absence of Jahn-Teller distortion in this metal ion. A scheme involving minor but more rapidly exchanging aquochloro complexes of Ni2+has been proposed previouslyz2on the basis of 35Cland 170NMR studies. The increase in line width as a function of the ratio of the concentration of Ni2+ to the concentration of protein for the three albumins studies is shown in Fig. 3. The experiments were performed (not shown) at two different concentrations for n
(3)
where fB is the fraction of enzyme-complexed ion; fB = [P.I]/ [I]; TIF,T pF, and bF are parameters corresponding to the uncomplexed ion; T 1 B and TSRcorrespond to the complexed ion; 7 is the lifetime of the complex; and A is the chemical shift difference between the free and complexed ion. The above equation can be simplified for a variety of limiting exchange cases. Extension for the case of an ion with spin >?h, such as chloride, has been described by Bullz1and involves the definition of approximate relaxation times, since the actual relaxation behavior of a quadrupolar nucleus is predicted to be nonexponential. A complete solution of the exchange problem requires that temperature-dependent data be obtained to evaluate the exchange kinetics, i.e., fast, slow, or intermediate on the NMR time scale. However, binding site titration studies can be carried out based on the proportionality of the protein perturbation terms in eqs. 2-4 to the ratio of bound to free ion fB. The only limit in which this approximation fails to hold is the case of exchange which is so slow that the binding has a negligible effect on the relaxation and shift parameters of the free species. This limit is readily detected since the addition of protein or high-affinity binding sites will have no effect on the observed parameters.
Figure 2-35Cl NMR spectrum of: (a) 0.15 M NaCl; (b) 0.75 M NaCl containing 235 WMNiC12; (c) 0.15 M NaCl containing 70 pM CuS04; (d) 0.15 M NaCI containing 15.8 WMBSA; (e) 0.15 M NaCl containing 43.8 pM BSA; (f) 0.15 M NaCl containing 43.8 gM BSA and 235 FM NiCQ; (g} 0.15 M NaCl containing 15.8 WMBSA and 49 pM CuS04.
c
I
16
f
U
'Z
I2
m
.-C
a
Results and Discussion Figure 2 illustrates the "C1 NMR line widths under different experimental conditions. There is a n increase in line width when either albumin or the metal ion, Cuz+or NiZ+,is added to 0.15 M NaC1. Also, in the absence of metal ions, the excess line 62
/ Journal of Pharmaceutical Sciences Vol. 74, No. 1, January 1985
[NI'+I/
Calbuminl
Figure 3-35Cl line broadening (Hz)versus the ratio of Niz+ion concentration to the albumin Concentration. Albumin concentrations are 28.82 pM for BSA (O),30.4 pM for DSA (m), and 49 pM for HSA (A).
"9 60
I 0
I 2
6
4
S
1 0 1 2 1 4
[Cu'+l /[albumin1
Figure 4-35CI line broadening (Hz) versus the ratio of Cu2+to albumin concentration. Albumin concentrations are 15.8 pM for BSA (m), 22.04 pM for DSA (A),and 10.23 pM for HSA (0). Table I-Number
Albumin
of Albumin Primary Binding Sites
Cu2+Sites
BSA
3.1
DSA
3.7a 6.6
HSA
1.3
Ni2+Sites
3.2
Affinities at Primary Sites (M-' x 10-6)
Cu2+
Ni2+
-
-
-
3.1 a
6.7 7.0b
-
1.5
-
-
the number of primary binding sites occur at the same position independent of whether the correction is applied. Our observation that the line width increases with increasing Cu2+ concentration after the primary sites of HSA and BSA are occupied does not agree with the findings of Sudmeier and Pesek." However, there are several differences between the two studies. First, Sudmeier and Pesek used albumin concentrations that were 10-100 times higher than those employed in our study. Second, their measurements were made over a wider pH range in acetate or phosphate buffers. Acetate ions may compete with chloride ions at pH values <6.5, while above pH 6.5 competition with hydroxyl ions may occur. Phosphate is not a suitable buffer ion for experiments involving copper since it forms insoluble cupric phosphate. The findings of the present study are summarized in Table I. For comparison, data from previous studies are also shown. The number of primary copper sites found in the present study for each of the three albumins is almost the same as the number of primary nickel sites. The number of primary copper sites for BSA and HSA from the present experiments is close to that obtained using a cupric ion-selective electrode." On the basis of the results from the present study, it appears that both copper and nickel have the same binding sites on albumin. This is in line with the earlier findings'-" that the lower affinity of DSA for Cu2+and Ni2+is due to the absence of a histidine at the third amino acid position from its N-terminal.
0.025'
-
1.1"
15.0b 2.3" 0.3b a Taken from ref. 18. Taken from ref. 23.Only one class of sites was detected.
each of the albumins. From the results of the titration experiments, the number of primary Ni2+ sites were estimated to be 1.5 for HSA, 3.2 for BSA, and 6.7 for DSA. Since the increase in line width with increasing Ni2+concentration is very small in the absence of protein, the increase in line widths in Fig. 3 have not been corrected for residual broadening due to any free Ni2+.The initial slope for DSA is lower than for either BSA or HSA at comparable protein concentrations. This may indicate that the primary nickel binding sites in DSA are different from those in the other two proteins. However, in the absence of complete characterization of the exchange kinetics the implication of this observation cannot be unequivocally determined. The results from the Cu2' titration experiments for the three different albumins are shown in Fig. 4. The increase in line widths observed when Cu2+is added to protein solutions have to be corrected for any paramagnetic/residual broadening. In Fig. 4 this was done by subtracting from the increase in line widths of albumin-Cu2+ solutions the increase in line widt.hs due to the same amount of Cuz+alone (i.e., in the absence of protein). The estimates for the number of Cu" sites obtained from Fig. 4 are 1.3 for HSA, 3.1 for BSA, and 6.6 for DSA. The initial 35Cl line broadening on the addition of Cu2+is very small for BSA (Fig. 4). However, there is a very small decrease for HSA and the initial slope is negative for DSA. The primary Cu2+ sites of BSA and HSA have affinities on the order of 106 M-1.3,20.22 At the protein concentrations used in these experiments, the free Cu2+ concentration is therefore likely to be extremely small, particularly a t low Cu2+/albumin ratios. Hence, for HSA and DSA, the free copper ion is more effective at relaxing the chloride ion than the enzyme-complexed ion. This may simply reflect the fact that more coordination positions are available for the aquo ion than for the protein-bound cation; the breaks in the curves which determine
References and Notes 1. Brown. J. R. in "Albumin: Structure. Function and Uses": Rosen-
2. 3. 4. 5.
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
oer, V.' M.; Oratz, M; Rothschild, M.A., Eds.; Pergamon Press: Oxford, 1977; pp 27-52. Shearer, W. T. J. Bid. Chem. 1967,242, 5451-5459. Peters., J... Jr.: Blumenstock, F. A. J. Biol. Chem 1967.242. . . 15741578. Bradshaw, R. A.; Shearer, W. T.; Gurd, F. R. N. J . Eiol. Chem. 1968,243, 3917-3825. Bradshaw, R. A.; Peters, J., Jr. J. Bid. Chem. 1969, 244, 55825589. Milanino, R.; Passerella, E.; Velo, G. P. in "Advances in Inflammation Research"; Weissman, G. D.; Samuelsson, B.; Raoletti, R. Eds.; Raven Press: New York, 1979; pp 281-291. Sunderman, F. W., Jr. Food Cosmet. Toricol. 1971,9, 109-120. Glennon, J. D.; Sarkar, B. Biochem. J. 1982,203, 15-23. Glennon, J. D.; Sarkar, B. Biochem. J . 1982,203, 25-31. Dixon, J. W.; Sarkar, B. Biochem. Biophys. Res. Commun. 1972, 48, 197-200. Lindman, B., Forsen, S. in "NMR Basic Principles and Progress"; Diehl, P.; Fluck, E.; Kosfeld, R. Eds.; Springer-Verlag: New York, 1976; pp 249-325. Gillberg-La Force, G.; Forsen, S. Biochem. Biophys. Res. Commun. 1970.38. 137-142. Nor& J.;Hjalmarsson, S.; Lindman, B.; Zeppezauer, M. Biochemi s t 1975,14, ~ 3401-3408. Halle, B.; Lindman, B. Biochemistry 1978,17, 3774-3781. Bull. T. E.: Halle. B.: Lindman. B. FEBS Lett. 1978. 86. 25-28. Sudmeier, J. L.; Pesek, J . J. Anal. Biochem. 1971,4i, 39-50. Chen, R. F. J. Biol Chem. 1967,242, 173-181. Mohanakrishnan, P.; Chignell, C. F. J . Pharm. Sci. 1982. 71, 1180-1182. Swift T. J.; Connick, R. E. J. Chem. Phys. 1962,37, 307-312. Luz, Z.; Meiboom, S. J. Chem. Phys. 1964, 40, 2686-2691. Bull, T. E. J. Magn. Reson. 1972,8, 344-353. Lincoln, S. F.; Aprile, F.; Dodgen, H. W.; Hunt, J. P. Znorg. Chem. 1968, 7, 929-932. Callan, W. M.; Sunderman, F. W., Jr. Res. Commun. C h m . Pathol. Pharmacol. 1973,5, 459-472. ,
Acknowledgments The authors wish to thank Drs. Frank Dearman, Robert Jordan, and Dallas Rabenstein for critically evaluating the manuscript.
Journal of Pharmaceutical Sciences Vol. 74, No. 7, January 7985
/ 63