ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 288, No. 1, July, pp. 79-86, 1991

The Binding of Pyridoxal 5’-Phosphate Human Serum Albumin’ Margaret

L. Fonda,’

Department

of Biochemistry,

Received November

Christiane University

Trauss,

and Ulrike

of Louisville

28, 1990, and in revised form February

to

M. Guempel

School of Medicine, Louisville,

Kentucky 40292

12, 1991

Most of the pyridoxal $-phosphate (PLP)3 in human plasma is bound to protein. The protein appears to be Most of the pyridoxal 5’-phosphate (PLP) in plasma is albumin based on its elution upon gel filtration chromahound to protein, primarily albumin. Binding to protein tography (1) and its fractionation by methanol precipiis probably important in transporting PLP in the circulation and in regulating its metabolism. The binding tation (2). of PLP to human serum albumin (HSA) was studied using In vitro, PLP binds with high affinity to two sites on absorption spectral analysis, equilibrium dialysis, and bovine serum albumin (BSA) (3). The PLP undergoes a inhibition studies. The kinetics of the changes in the series of spectral changes while binding (3). Anderson et spectrum of PLP when mixed with an equimolar conal. (4) sequenced the site of BSA modified when incubated centration of HSA at pH 7.4 followed a model for twowith a 1:l ratio of PLP and demonstrated that PLP binds step consecutive binding with rate constants of 7.72 mM-’ to a lysyl residue. It was later concluded that the lysyl min.. ’ and 0.088 min-‘. The resulting PLP-HSA complex residue is either Lys 220 or Lys 223 (5, 6). It has been had absorption peaks at 338 and 414 nm and was reduced suggested that at physiological pH, PLP forms a substiby potassium borohydride. The 414-nm peak is probably tuted aldimine with this lysyl residue and a nearby amino due to a protonated aldimine formed between PLP and acid residue (4,7) or forms a Schiff base in a hydrophobic HSA. The binding of PLP to bovine serum albumin (BSA) environment (8). at equimolar concentrations at pH 7.4 occurred at about Less is known about the interaction of PLP with human 10% the rate of its binding to HSA. The final PLP-BSA serum albumin (HSA). The binding of PLP to HSA also complex absorbed maximally at 334 nm and did not apleads to time-dependent absorption and CD spectral pear to be reduced with borohydride. Equilibrium dichanges (9, 10). In a 1:l molar ratio, PLP was shown to alysis of PLP and HSA indicated that there were more modify a lysyl residue in fragment C of HSA; however, than one class of binding sites of HSA for PLP. There the labeled fragment C was not sequenced (11,12). Subwas one high affinity site with a dissociation constant of sequently, we demonstrated that PLP primarily binds to 8.7 PM and two or more other sites with dissociation conLys 190 of HSA (13). stants of 90 PM or greater. PLP binding to HSA was When PLP is bound to serum proteins or to HSA, the inhibited by pyridoxal and 4-pyridoxic acid. It was not rate of transport of PLP into erythrocytes is decreased inhibited appreciably by inorganic phosphate or phosphorylated compounds. The binding of PLP to BSA was (1,2). The presence of HSA or BSA inhibits the hydrolysis inhibited more than its binding to HSA by several comof PLP by phosphatases (2,14,15) and the disappearance pounds containing anionic groups. It is concluded that of PLP in uiuo from rat jejunum (14). PLP binds differently to HSA than it does to BSA. c 1991 The relative amounts of free and protein-bound PLP Academic Press, Inc. in the circulation are important in determining the availability of PLP to tissues and the amount of PLP that is catabolized. A better understanding of the binding of PLP 1 This work was supported in part by a grant from National Science Foundation-Kentucky EPSCoR and by National Institutes of Health Grant AA06861. ’ To whom correspondence should be addressed.

0003.9861/91 $3.00 Copyright cc) 1991 by Academic Press, All rights of reproduction in any fonn

3 Abbreviations used: PLP, pyridoxal5’-phosphate; BSA, bovine serum albumin; HSA, human serum albumin; PBS, phosphate-buffered saline; PL, pyridoxal.

79 Inc. reserved.

80

FONDA,

TRAUSS,

to HSA would help predict the relative amounts of free and protein-bound PLP under physiological conditions. The present studies were designed to determine the kinetics and stoichiometry of the binding of PLP to monomeric HSA and the effect of vitamin B6 compounds and analogs and various physiological compounds on the binding of PLP to HSA. MATERIALS

AND

METHODS

Materials. PLP, most of the other biochemicals, and Sephadex G50 were obtained from Sigma Chemical Co. Sodium borotritide (sp act 348 Ci/mol) was purchased from New England Nuclear. Dowex AG lX2 (200-400 mesh) and Affi-Gel Blue were obtained from Bio-Rad Laboratories. [4’-3H]PLP was synthesized by reduction of PLP with sodium borotritide followed by oxidation with manganese dioxide (16). The [3H]PLP was purified by chromatography on Dowex l-acetate (17). The purity of the [3H]PLP was determined by HPLC using a Spectra Physics 8700XR ternary gradient system with a 0.46 X 25.cm Vydac 401TP104 cation exchange column as described by Coburn and Mahuren (18). The effluent was reacted with sodium hisulfite, and the fluorescence was monitored with a McPherson FL-750 spectrofluorescence detector with a 24+1 flow cell, excitation wavelength set at 332 nm and emission wavelength at 400 nm. A Spectra Physics integrator was used. Fractions were collected every 0.5 min, and radioactivity was determined. At least 91% of the total fluorescence and 92% of the total radioactivity eluted from the HPLC column at the time corresponding to a PLP standard. The specific radioactivity was 15.6 Ci/mol. HSA (Cohn Fraction V) was purchased from the American Red Cross, and BSA (Fraction V) was obtained from Sigma Chemical Co. Both were purified further by affinity chromatography on Affi-Gel Blue to remove globulins (19). More than 95% of the HSA migrated as the monomer in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (20). The buffer used for most experiments was 137 mM sodium chloride, 10 mM sodium phosphate, and 4.2 mM potassium chloride, pH 7.4 (PBS). Solutions were prepared with distilled and deionized water. Solutions of vitamin B, compounds and of PLP-modified HSA or BSA were protected from the light, and all experiments involving these compounds were performed in a room illuminated only with yellow fluorescent lights (Sylvania F15T8-GO). Methods. Kinetic and spectral studies were carried out in PBS, using a Hewlett-Packard 8452 diode array spectrophotometer. Other absorp tion measurements were made with a Zeiss PM1 spectrophotometer. PLP concentration was determined by its absorbance at 388 nm at pH 7 using the extinction coefficient 4.9 mM-’ cm-i (21). Albumin concentrations were determined by measuring the absorbance at 279 nm and using the extinction coefficients 35.7 mM-’ cm-’ for HSA and 44.7 mM-’ cm-’ for BSA (6). Stoichiometry of PLP binding to HSA was estimated in two ways. HSA was incubated with PLP in PBS, pH 7.4, for various times at room temperature, and subsequently potassium borohydride was added. After 30 min, the samples were dialyzed for 24 h at 4°C against PBS. The amount of covalently bound PLP was estimated from the absorbance at 323 nm using the extinction coefficient 5.8 mM-’ cm-’ calculated for t-pyridoxyl groups attached to BSA (3). In addition, HSA was incubated with 13H]PLP in PBS, pH 7.4, at room temperature. Potassium borohydride was added. After 30 min, aliquots were spotted on Whatman 540 filter papers and washed thoroughly with 10% trichloroacetic acid (22). Each filter paper was treated with 0.75 ml of Protosol (New England Nuclear) for 3 h at 45’C in a capped scintillation vial. Scintillation cocktail (10 ml) and 100 ~1 1 N HCl were added, and the radioactivity was determined with a Beckman LS7500 scintillation counter.

AND

GUEMPEL

The equilibrium dialysis experiments were performed using 10 mm standard dialysis tubing with molecular weight cutoff of 12,000-14,000 (Spectrum Medical Industries). The tubing was soaked in and thoroughly washed with PBS. Albumin in PBS (500 pl; 3-12 PM) was pipetted into a freshly prepared dialysis bag which was placed in a screw-capped tube containing 10 ml of 0.3 to 80 pM [3H]PLP in PBS. The tubes were gently rotated with a multipurpose rotator (Scientific Industries, Inc.) at 4’C for 46-48 h. Dialysis of 3 fiM HSA and 2 pM [3H]PLP for 24 to 72 h indicated that the system reached equilibrium by 40 h. Samples were removed from both inside of the dialysis bag and from the outside solution, and the radioactivity was determined in Aquasol II scintillation solution (New England Nuclear). The binding of PLP to HSA or BSA was also measured using an elution centrifugation method (23). HSA (6 PM) was incubated with 0.380 pM [3H]PLP in PBS or other designated solution for 120 min at room temperature. Triplicate aliquots of each incubation mixture were put over small Sephadex columns to separate the free and albumin-bound PLP. A small amount of absorbent cotton was placed in the tips of disposable l-ml tuberculin syringes which were then packed to the 1 ml mark with Sephadex G-50 (bead size 20-50 am) swollen in PBS. The column was placed in a test tube (12 X 75 mm), and after no further buffer drained from it, it was centrifuged for 2 min at 1OOgin a TH-4 swinging bucket rotor of a Beckman TJ-6 centrifuge. A 100-/d aliquot of reaction mixture was placed on the top of the column, the column was placed in a plastic mini scintillation vial and centrifuged at 1OOg for 2 min. The column was washed with 100 ~1 of PBS and centrifuged at 1OOgfor 1 min. The radioactivity of the entire effluent was measured. This procedure gave 95 + 5% recovery of albumin, and good separation of bound and free PLP. When albumin was omitted, the radioactivity in the effluent was slightly above background (less than 0.02% of that added to the column) and was subtracted from experimental values.

RESULTS

The binding of PLP to an equimolar concentration of HSA resulted in the spectral changes shown in Fig. 1A. The decrease in absorbance of PLP at 388 nm was nearly complete by 60 min. The PLP-HSA complex had absorption maxima at 338 nm (C = 3.16 mM-’ cm-‘) and 414 nm (C = 2.35 mMml cm-‘). Treatment of the complex with potassium borohydride resulted in the loss of the absorption peaks at 338 and 414 nm, and an increase in the absorbance at 317 nm. These spectral data were consistent with PLP binding to HSA as a Schiff base which was reducible by borohydride. The progress curves of the spectral changes at 338, 388, and 440 nm obtained after mixing 0.12 mM PLP and HSA are shown in Fig. 2. The absorption change was followed at 440 nm rather than 414 nm because 440 nm gave the best compromise between maximum sensitivity and minimum interference by unbound PLP. The absorbance at 388 nm decreased with time. A plot of l/A absorbance at 388 nm versus time was linear, indicating second-order kinetics with a rate constant of 7.5 k 0.71 mM-’ min-‘. There was an increase in absorbance at 440 nm which peaked at approximately 5 min and then decreased slowly. A plot of the log of the A absorbance at 440 nm was linear with time from 7 to 30 min. The decrease in absorbance at 440 nm thus followed first-order kinetics with a rate constant of 0.077 + 0.002 min-‘. The absorbance at 338 nm increased slowly. A plot of the log of the A absorbance at 338 nm was linear with

PYRIDOXAL

PHOSPHATE

BINDING

TO HUMAN

SERUM

81

ALBUMIN

reaction followed by a unimolecular reaction were used to estimate the concentrations of the components as a function of time (24),

0.6

[PLP]

[II =

=

WPIO 1 + [PLP],k,t

[pLp]~k~[pLp~~

te-

[21



[PI,P],,k,t

ks - k,[PLP]o

[II] = [PLP]o - [PLP] - [I],

_

e-k,t)

f

[31

[41

where [PLP]o is the initial PLP concentration and [PLP], [I], and [II] are the concentrations of free PLP and complexes I and II, respectively. The absorbance at the wavelength of interest at time t would be A = ~PLPWJ’I 0.0 ’ 300

I 325

350

375

Wavelength

400

425

450

(nm)

FIG. 1. Absorption spectra of PLP with HSA or BSA in PBS, pH 7.4, at 22°C. (A) Spectra of 0.1 mM PLP and 0.1 mM HSA blanked against HSA at 0, 1, 2, 4, and 60 min after mixing. After 60 min, 0.2 mM KBH, was added to the mixture (+KBH,). (B) Spectra of 0.1 mM PLP and 0.1 mM BSA blanked against BSA at 0,6,30, and 60 min and 18 h after mixing. After 18 h, 0.2 mM KBH, was added to the mixture (+KBH,).

time from 0 to 30 min, indicating first-order kinetics and a rate constant of 0.086 +- 0.002 min-‘. The binding of PLP to HSA appeared to be a consecutive process of at least two steps involving two HSA-PLP complexes, I and II, HSA + PLP 2 complex I 2 complex II,

+ 411 + dI1,

[51

where tpl,p, ~1, and tII are the extinction coefficients of free PLP, complex I, and complex II, respectively, at the given wavelength. Data shown in Fig. 2 and data from nine other similar experiments collected every 5 s for the first 200 s and every 50 s from 200 to 1800 s were fitted to Eq. [5], using Eqs. [2] through [4] to calculate the concentrations of PLP, I, and II as a function of time. The values of kl, k2, ~~~and tII were varied to minimize the residual sum of squares, C (A,/A, - 1)‘. A0 and A, are the observed and calculated absorbance values at 338, 388, and 440 nm at each time of observation. A best fit of the data to Eqs. [2] to [5] was obtained with k, = 7.72 mM-’ mini’, k,, = 0.088 min-‘, and extinction coefficients of 2.95, 2.15, and 2.28 mM -’ cm-’ for complex I at 338, 388, and

t11

where kl and kZ are the rate constants for the two steps. In the bimolecular interaction, the absorbance at 388 nm decreases and that at 440 nm increases. This is followed by a slower unimolecular step in which the absorbance at 440 nm decreases and that at 338 nm increases. The reactions are probably reversible, but the reverse rate constants are assumed to be small enough to be neglected in the data analysis. Dialysis of the 0.1 mM HSA-PLP complex against PBS, pH 7.4, indicated that the complex slowly dissociated. The extinction coefficient at 414 nm decreased to 0.97 and 0.62 mM-’ cm-’ and that at 338 nm decreased to 2.16 and 1.40 mMp’ cm-’ after dialysis for 24 and 63 h, respectively. The fit of the experimental data to Reaction [l] was analyzed as follows. Equations [Z] to [4] for a bimolecular

FIG. 2. The change in absorbance at 338,388, and 440 nm with time after mixing 0.12 mM PLP and HSA in PBS, pH 7.4. The dots are experimental points. The solid lines are the theoretical spectral changes calculated using the model described in the text in which PLP and HSA bind in two consecutive steps with rate constants of 7.72 mM-’ min ’ and 0.088 min I.

82

FONDA,

Stoichiometry [PLP] bound after reduction (PM)

Time of incubation (min) 2 5 60 60 60

Spectral” 100 100 100 200 400

45 70 96 156 244

TCAb 47 78 95

TRAUSS,

AND

GUEMPEL

TABLE

I

of PLP

Binding

Calculated

[PLPIF

35 18 1

to HSA

concentrations’ (PM)

A Absorbanced

III

[III

338 nm

388 nm

440 nm

56 53 0

9 29 99

0.0118 0.0231 0.0616

-0.227 -0.295 -0.375

0.121 0.138 0.111

n 100 pM HSA was incubated with PLP in PBS, pH 7.4, at room temperature. At the indicated time, potassium borohydride was added to a final concentration of 800 /IM. The samples were dialyzed for 24 h at 4°C against PBS. The amount of bound PLP was estimated from the absorbance at 323 nm. b HSA was reacted with [3H]PLP and reduced with borohydride, and the radioactivity in the TCA-precipitable material was determined. ’ Concentrations were calculated using Eqs. [2] to [4] and values for the rate constants and extinction coefficients given in Results. d From Fig. 1.

440 nm, and extinction coefficients of 3.36,1.86, and 1.60 mM-’ cm- ’ for complex II at 338, 388, and 440 nm, respectively. The good fit of the data points to the theoretical spectral changes for this model using these values (solid lines) is shown in Fig. 2. The values for the standard error of estimate (the square root of the residual sum of squares divided by the degrees of freedom) were 0.0017, 0.0106, and 0.0073 at 338, 388, and 440 nm, respectively. Correlation coefficients of the fit of the data to the equations were 0.999 at the three wavelengths. When the 1:l PLP-HSA complex was placed in buffers of different pH values, similar final spectra were obtained for 0.1 mM PLP-HSA from pH 4.5 to 9.0 as the 60-min spectrum shown in Fig. 1A. Little or no PLP remained bound to HSA at pH 11.5 or higher. The spectral changes and kinetics of the binding of PLP to HSA appeared to be different from those reported for PLP and BSA (3, 4). Therefore, the binding of PLP to BSA, prepared in a similar manner as HSA, was investigated. The spectral changes obtained during the binding of PLP to BSA are shown in Fig. 1B. PLP bound more slowly to BSA than it bound to HSA. There was a decrease in the absorbance of PLP at 388 nm and a slow increase in absorbance at 334 nm. The spectral changes were complete by 18 h, and the final complex had an absorption peak at 334 nm (c = 4.35 mM-’ cm-‘) and a small peak at 404 nm (t = 0.83 mMpl cm-‘). The decrease in absorbance at 388 nm had a rate constant of approxiThe absorbance at 440 nm inmately 0.59 mM-’ min-‘. creased initially and then decreased after 9 min. The decrease in absorbance at 440 nm and the increase in absorbance at 334 nm followed first-order kinetics for 100 min with rate constants of approximately 0.0110 min-’ and 0.0105 min-‘, respectively. The spectral changes did not fit Eqs. [2] to [5] as well as was obtained for PLP

binding to HSA. The binding of PLP to BSA appeared to be more complex than that to HSA. The addition of potassium borohydride had very little effect on the absorption spectrum of the BSA-PLP complex; the absorption peak was shifted from 334 to 332 nm (Fig. 1B). The complexes formed between 0.1 mM PLP and HSA after 2, 5, and 60 min incubation at room temperature were treated with potassium borohydride. The pyridoxylated protein was separated from free pyridoxine-P by dialysis. The reduced HSA product absorbed maximally at 317-320 nm. lt was estimated that after 60 min incubation 90-97% of the PLP was covalently bound to HSA (Table I). The amount of e-pyridoxyl covalently bound to HSA increased with time in parallel to the decrease in absorbance at 388 nm and in parallel with the theoretical concentrations of complexes I and II (Table I). The complex formed between 0.1 mM PLP and BSA after 20 h incubation at room temperature was also treated with potassium borohydride and dialyzed for 24 h against PBS. The remaining PLP-BSA product absorbed maximally at 332 nm with an absorbance of 0.321. The binding affinity and stoichiometry of HSA for PLP was estimated by Scatchard analysis (25) of data obtained from equilibrium dialysis and elution centrifugation experiments. The nonlinear Scatchard plots (Fig. 3) of the binding of PLP to HSA as a function of PLP concentration indicate that there were more than one class of binding sites. The data obtained from nine equilibrium dialysis experiments were averaged and subjected to computer analysis assuming two independent classes of sites, using Eq. [6]:

B/F =

~&~[ATI 1 +

K,F

+ R&[ATI

1 +K,F



if31

PYRIDOXAL

PHOSPHATE

BINDING

TO HUMAN

SERUM

ALBUMIN

83

pounds on the binding of PLP to BSA (3). Therefore, the effect of these compounds on the binding of PLP to BSA was studied by elution centrifugation using the same conditions as for the HSA study. 4Pyridoxic acid, phenylphosphate, benzaldehyde, benzoate, acetylsalicylate, oleate, and iodoacetate inhibited the binding of PLP to BSA more effectively than to HSA (Table II). DISCUSSION 0

2

4

6

8

[PLPI,

(/AM)

IO

12

14

FIG. 3. Scatchard plot of binding data for PLP and HSA obtained by equilibrium dialysis in PBS, pH 7.4, at 4°C for 48 h. The HSA was 6 fiM and PLP concentrations ranged from 0.3 to 80 KM. Each point represents the average data obtained from nine experiments k SE. The lines are the calculated best fit assuming two independent classes of binding sites.

where AT is the total concentration of HSA, B is the concentration of bound PLP, F is the concentration of free PLP, K1 and K2 and nl and n2 are the association constants and number of the two classes of sites, respectively. The best fit of the experimental data to Eq. [6] was estimated by simultaneously varying the values for ni , n2, K1, and K2 to minimize the residual sum of squares. A best fit, shown as lines in Fig. 3, indicates there was one high affinity site with a dissociation constant of approximately 8.7 PM and two other sites with a dissociation constant of approximately 90 PM. The standard error of estimate was 0.095. The data also fit a model of one high affinity site and low affinity nonspecific sites. No significant binding of PL to HSA was observed when 6 PM HSA was incubated with 10 to 40 PM [3H]PL for 0.5 to 24 h. The HSA and free PL were separated by elution centrifugation. The effect of various vitamin Br, analogs and other compounds, known to bind to albumin, on the binding of PLP to HSA was studied by elution centrifugation and by spectral analysis (Table II). A concentration of 10 PM PLP was used so that PLP would mainly bind to the high affinity site. Of the vitamin B6 analogs, PL and 4-pyridoxic acid inhibited the binding of PLP to HSA the most. Pyridoxamine, pyridoxine, deoxypyridoxine, and pyridoxamine-P showed little or no inhibition. Although p-nitrophenylphosphate was inhibitory, other phosphorylated compounds (phenylphosphate, PMP, inorganic phosphate, phosphoglycolate) were not. The inhibition by Tris, L-cysteine, and L-tryptophan was probably due to Schiff base formation between the compound and PLP. The other compounds inhibited by binding to HSA. The effects of some of these inhibitors on the binding of PLP to HSA differed from the reported effects of these same com-

The kinetics of the change in the absorption spectrum of PLP when mixed with an equimolar concentration of HSA at pH 7.4 gave excellent fit to Reaction [l] for twostep consecutive binding. Two assumptions were made in deriving the equations used. First it was assumed that, under the conditions used, PLP bound to only one site on each HSA molecule. Scatchard analysis of equilibrium binding data (Fig. 3) indicated that HSA had a single high affinity binding site and two or more low affinity or perhaps nonspecific sites for PLP. The relative amounts of PLP bound at each site would depend not only on the relative affinities of the sites, but also on the rate at which PLP bound to the sites. The kinetic data shown in Fig. 2 do not fit a model of two parallel reactions in which different spectral changes occur at different rates: HSA + PLP 2 complex I

t71

HSA + PLP 2 complex II.

181

In Reactions [7] and [B], hi and JQ are second-order rate constants. However, only the A absorbance at 388 nm followed second-order kinetics. The rapid increase in absorbance at 440 nm followed by a slow decrease cannot be explained by two parallel reactions. Equations [2] to [4] were derived assuming that the binding of PLP to HSA was irreversible and went to completion. Kinetic equations for reversible reactions are very complex (24), and the equilibrium concentrations of PLP and the HSA-PLP complexes were not known. Exhaustive dialysis of 0.1 mM HSA-PLP complex indicated that the complex dissociated slowly. Approximately 60% of the PLP was released from the complex in 24 h and 74% in 63 h. After 60 min incubation of 0.1 mM PLP and HSA followed by borohydride reduction, 95% of the PLP was bound to HSA. The fit of the spectral data to Eq. [l] indicated that 99% of the PLP was bound to HSA at 60 min (Table I), which is in fairly good agreement. PLP bound more slowly to BSA than to HSA (Fig. 1). The kinetics of the spectral changes indicated the binding of PLP to BSA to be more complex than that to HSA. The binding of PLP to BSA might involve more than two consecutive steps or two multistep parallel reactions. Dempsey and Christensen (3) proposed that PLP binds quickly to a site II of BSA to give an increase in absor-

84

FONDA, TABLE Inhibition

of PLP by

Compound Pyridoxal 4-Pyridoxic acid Pyridoxamine Phenylphosphate p-Nitrophenylphosphate Benzaldehyde Benzoate Acetylsalicylate Oleate Acetaldehyde L-Tryptophan N-Acetyl-L-tryptophan Iodoacetate

TRAUSS,

II

Binding

to Albumin

Various Compounds

HSA 1.6 1.5 4.8 >12 1.0 >12 10.6 >3 0.18 260 1.2 9.7 >12

BSA 2.8 0.25 ND 0.2 ND 2 0.2 0.35 0.03 230 2 ND 0.5

Note. Albumin (3 pM) was incubated with 10 pM [3H]PLP alone and with four concentrations of the indicated compounds in PBS, pH 7.4, for 2 h. The free and protein-bound PLP were separated by elution centrifugation. PLP binding to HSA was inhibited less than 10% by 4 mM pyridoxamine-P, 1 mM pyridoxine, 1 mM deoxypyridoxine, 80 mM orthophosphate, ‘7 nM bilirubin, 10 mM EDTA, or 3 mM N-acetyl-Lcysteine. The binding of PLP to HSA was inhibited 72% by 100 mM Tris, 89% by 1 mM L-cysteine, and 30% by 3 mM phosphoglycolate. ND, not determined.

bance at 415 nm and then more slowly migrates to or reacts with a site I giving a decrease in absorbance at 415 nm and an increase at 332 nm. The affinity constants of BSA for PLP were estimated as more than 106, 105, and 7.8 X lo3 M-’ (3). Murakami et al. (26), on the basis of stopped flow and spectrophotometric measurements, concluded that PLP binds to BSA in a rapid bimolecular step followed by three sequential isomerization steps. They suggested that PLP binds first as a carbinolamine, which then converts sequentially to various tautomeric forms of a Schiff base. Several PLP-dependent enzymes that have been studied are reconstituted through multistep processes (27). For instance, the binding of PLP to the apoenzyme of aspartate aminotransferase proceeds through at least three consecutive steps, an initial binding followed by two irreversible isomerization steps (28). The value of k, for PLP binding to HSA (7.72 mM-’ min-‘) is lower than that for PLP binding to the apoaminotransferase (22 to 1920 mM-’ mini’) (28, 29). However, it is higher than that for the binding of PLP to ribonuclease A (1.98 mMp’ min~l) or to free amino acids (0.98 mM-’ min-‘) (10). Other ligands such as bilirubin (30, 31), hematin (32), and fatty acids (33) bind to HSA in two or more consecutive steps. Generally, there is a rapid association followed by a conformational change. The rates of the conforma-

AND

GUEMPEL

tional changes range from 3 to 40 s-l. The rate of the second step in the binding of PLP to HSA or BSA is much slower than rates observed for conformational changes in albumin and is probably not due to a simple conformational change in the protein. The kinetic and spectral data for PLP and HSA fit Reaction [l] with step 2 being irreversible as shown; however, it is probable that it is reversible. If step 2 were irreversible, complex II has absorption maxima at 338 and 414 nm. If step 2 were reversible, the final HSA-PLP is a mixture of complexes I and II with complex I having the 414-nm absorption peak and complex II having the 338-nm absorption peak. Regardless of the reversibility, the intermediate complex I has a greater absorption peak at 414-440 nm and a lower absorption peak at 338 nm than does complex II. The absorption spectra shown in Fig. 1 are of complexes with PLP bound mainly to the high affinity site of HSA (A) and BSA (B). The spectrum with peaks at both 338 and 414 nm indicates that PLP may be bound in two different ways in the final HSA-PLP complex. The ratio of the absorbance of the HSA-PLP at the two wavelengths was constant from pH 4.5 to 9. The absorption peak at 414 nm is typical of a protonated aldimine between PLP and an amino group and is found in many B,-dependent enzymes (34). The 414-nm absorbing material in the HSA-PLP complex was reduced by borohydride as would be expected with such an imine. The material absorbing at 338 nm also appeared to be reduced by borohydride since the absorption peak at 338 nm decreased somewhat upon addition of borohydride and the amount of 317-nm absorption or radioactivity incorporated after addition of borohydride was equivalent to approximately 0.095 mM t-pyridoxyl moiety (Table I). The nature of the 338-nm material is unknown. The complex formed between BSA and PLP had a major absorption peak at 334 nm and a small peak at 404 nm. The complex was not reduced by borohydride at pH 7.4. Similar results were reported previously (3, 4, 35). The compound absorbing at 334 nm has been suggested to be an adduct of PLP with two lysyl residues (4) or with a lysyl and a histidyl residue (7) or an unprotonated imine between PLP and a lysyl residue in a hydrophobic environment (8, 35). Some enzymes containing covalently bound PLP have two absorption peaks in the ranges of 410-430 nm and 330-340 nm (34, 36). r tese include the P-isoform of asDpartate aminotransfe, lysine c-aminotransferase, ,erase, and ornithine d-aminoamino acid aminotr: transferase. These el nes are thought to contain PLP bound in two different ways-as a protonated imine at the active site absorbing in the range of 410-430 nm and in an inactive form of unknown structure. The 414- and 338-nm absorbing forms of HSA-PLP may reflect some heterogeneity in the HSA or may be

PYRIDOXAL

PHOSPHATE

BINDING

PLP bound at two different sites. One principal site is indicated by the Scatchard analysis. In addition, Bohney et al. (13) found that upon incubation of 0.1 mM HSA and [3H]PLP for 60 min followed by reduction with borohydride most of the label was associated with one site (Lys 190). Not only were the kinetics of the spectral changes and the absorption spectra of the final albumin-PLP product different for HSA and BSA, but the ability of the two albumins to bind PLP in the presence of inhibitors also differed. Anionic compounds, such as 4-pyridoxic acid, phenylphosphate, benzoate, acetylsalicylate, oleate, and iodoacetate, were better inhibitors of PLP binding to BSA than to HSA (Table II). Dempsey and Christensen (3) reported that these compounds, except 4pyridoxic acid which they did not study, inhibit the slow phase of the reaction of PLP with BSA. They suggested that this reaction is the migration of PLP from the initial site to the final site to form the complex which absorbs at 332-334 nm. The reaction of PLP with HSA resulted in the formation of a complex absorbing at 338 nm. The 338-nm absorbing form of PLP-HSA is probably different from the 334-nm absorbing form of PLP-BSA because of the differences in their reducibility by borohydride and the differences in inhibition by these anionic compounds. The phosphate group of PLP does not appear to be important in the binding of PLP to HSA since orthophosphate, pyridoxamine phosphate, and phenylphosphate were not effective inhibitors. Even the best inhibitors of PLP binding to HSA (PL, 4-pyridoxic acid, pnitrophenylphosphate, and oleate) bound with much lower affinity than PLP. Other aldehydes bound poorly, if at all. No significant binding of 40 PM [“H]PL to HSA was observed. HSA has an indole-binding site which binds L-tryptophan, N-acetyl-L-tryptophan, and indolepropionate. Gambhir et al. (11) reported that indolepropionate decreased the labeling of HSA with PLP by 25%. We found that N-acetyl-L-tryptophan was not an effective inhibitor of PLP binding. Thus, PLP and indole compounds do not appear to bind to the same sites on HSA. L-Tryptophan binds primarily in domain III of HSA (6). Likewise, acetylsalicylate and bilirubin did not significantly inhibit the binding of PLP to HSA. It has been proposed that bilirubin interacts with Arg 222 and acetylsalicylate binds to Lys 199 in domain II of HSA (6). Oleic acid inhibited PLP binding to HSA. There are a number of fatty acid binding sites on HSA. The primary and secondary sites are in domains II and III, and the binding of fatty acid to the primary site causes a conformational change in HSA (33). The inihibition of PLP binding by oleate may be due to overlap of binding sites or to the conformational change induced by the fatty acid. The albumins used in our study were not defatted and thus probably contained 0.5 to 1 mol fatty acid per mole of albumin. In. uiuo, HSA

TO HUMAN

SERUM

ALBUMIN

85

should bind PLP with affinity and kinetics similar to those we obtained. Binding of PLP to serum proteins is important in regulating the availability of PLP to tissues and the catabolism of PLP. The rate of transport of PLP into erythrocytes is decreased when PLP is bound to serum proteins or to HSA (1,2). The presence of serum albumin inhibits the hydrolysis of PLP by intestinal alkaline phosphatase (2, 14) or by an erythrocyte phosphatase (15). BSA also inhibits the luminal disappearance of PLP in uiuo from perfused rat jejunum (14). Serum concentration of PLP in humans is approximately 60 nM, most of which is protein-bound. Since HSA concentration in serum is approximately 0.6 mM and its affinity for PLP is 9 PM, it would be anticipated that most of the PLP in serum is bound to serum albumin in uiuo. Based on our results (Table II), compounds which may have an effect on PLP binding to HSA in viuo would be amino acids and fatty acids. High concentrations of circulating fatty acids would be expected to inhibit the binding of PLP to HSA. REFERENCES 1. Anderson, B. B., Newmark, P. A., Rawlins, M., and Green, R. (1974) Nature 250, 502-504. 2. Lumeng, L., Brashear, R. E., and Li, T.-K. (1974) J. Lab. Clin. Med. 84,334-343. 3. Dempsey, W. B., and Christensen, H. N. (1962) J. Biol. Chem. 237, 1113-1120. 4. Anderson, J. A., Chang, H. W., and Grandjian, C. J. (1971) Biochemistry 10,2408-2415. 5. Geisow, M. J., and Beaven, G. H. (1977) Biochem. J. 163,477-484. 6. Peters, T., Jr. (1985) Adu. Protein Chem. 37, 161-245. 7. Stepuro, I. I., Solodunov, A. A., and Nefedov, L. I. (1982) Mol. Biol. (Moscow) 16, 103441042. 8. Cortijo, M., Jimenez, J. S., and Llor, J. (1978) Biochem. J. 1'7'1, 497-500. 9. Stepuro, I. I., Solodunov, A. A., Kukresh, M. I., and Nefedov, L. I. (1985) Mol. Biol. (Moscow) 19, 109331098. 10. Moroz, A. R., Kondakov, V. I., Stepuro, I. I., and Yaroshevich, N. A. (1987) Biokhimiya (USSR) 52,470-480. 11. Gambhir, K. K., and McMenamy, R. H. (1973) J. Biol. Chem. 248, 1956-1960. 12. Gambhir, K. K., McMenamy, R. H., and Watson, F. (1975) J. Biol. Chem. 250,6711-6719. 13. Bohney, J. P., Fonda, M. L., and Feldhoff, R. C. (1989) FASEB J. 3, A926. 14. Middleton, H. M., III (1986) Gastroenterology 91, 343-350. 15. Fonda, M. L., Trauss, C., and Guempel, U. M. (1990) Ann. N. Y. Acad. Sci. 585, 483-485. 16. Stock, A., Ortanderl, F., and Pfieiderer, G., (1966) Biochem. 2.344, 353-360. 17. Raibaud, O., and Goldberg, M. E. (1974) FE:BS Lett. 40, 41-44. 18. Coburn, S. P., and Mahuren, J. D. (1983) Anal. Biochem. 129,310317. 19. Feldhoff, R. C., and Ledden, D. J. (1982) Fed. Proc. 41, 658. 20. Laemmli, U. K. (1970) Nature 227, 680-685. 21. Peterson, E. A., and Sober, H. A. (1954) J. Amer. Chem. Sot. 76, 1699175.

86

FONDA,

TRAUSS,

22. Mans, R. J., and Novelli, G. D. (1961) Arch. Biochem. Biophys. 94, 48-53. 23. Penefsky, H. S. (1977) J. Biol. Chem. 252, 2891-2899. 24. Chien, J.-Y. (1948) J. Amer. Chem. Sot. 70, 2256-2261. 25. Scatchard, G. (1949) Ann. N. Y. Acad. Sci. 51, 660-672. 26. Murakami, K., Kubota, Y., Fujisaki, Y., and Sano, T. (1986) Bull. Chem. Sot. Japan 59,3399-3403. 27. Jenkins, W. T., and Fonda, M. L. (1985) in Transaminases (Christen, P., and Metzler, D. E., Eds.), pp. 216-234, Wiley, New York. 28. Verge, D., and Arrio-DuPont, M. (1981) Biochemistry 20, 12101216. 29. Fonda, M. L., and Auerbach, S. B. (1976) Biochim. Biophys. Acta

422,38-47.

AND

GUEMPEL

30. Gray, R. D., and Stroupe, S. D. (1978) J. Biol. Chem. 253, 4370-

4377. 31. Honor&, B. (1987) J. Biol. Chem. 262,

14,939-14,944.

32. Adams, P. A., and Berman, M. C. (1980) Biochem. J. 191,95-102. 33. Scheider, W. (1979) Proc. Natl. Acad. Sci. USA 76, 2283-2287. 34. Kallen, R. G., Korpela, T., Martell, A. E., Matsushima, Y., Metzler, C. M., Metzler, D. E., Morozov, Y. V., Ralston, I. M., Savin, F. A., Torchinsky, Y. M., and Ueno, H. (1985) in Transaminases (Christen, P., and Metzler, D. E., Eds.), pp. 37-108, Wiley, New York. 35. Hilak, M. C., Harmsen, B. J. M., Joordens, J. J. M., and Van OS, G. A. J. (1975) Znt. J. Peptide Protein Res. 7, 411-416. 36. Soda, K. (1985) in Transaminases (Christen, Eds.), pp. 421-430, Wiley, New York.

P., and Metzler, D. E.,

The binding of pyridoxal 5'-phosphate to human serum albumin.

Most of the pyridoxal 5'-phosphate (PLP) in plasma is bound to protein, primarily albumin. Binding to protein is probably important in transporting PL...
890KB Sizes 0 Downloads 0 Views