Fluorescence Changes Induced by Binding of 4-Pyridoxic Acid 5’-Phosphate to Proteins*

SUMMARY The affinity constant (KA) and stoichiometry of binding of 4-pyridoxic acid S/-phosphate (4-pyridoxic-5’-P) complexed to the protein bovine serum albumin and aspartate aminotransferase were investigated by fluorometric methods. Whereas albumin The spectroscopic properties (absorption and fluorescence spectra) of the complexes were examined over a wide range of pH values. At neutral pH, the formation of the 4-pyridoxic-5’-P bovine serum albumin complex causes the following spectroscopic changes in the ligand: of

From the Department of Biochemistry, The University of Tennessee, Knoxville, Tennessee 37916

SUMMARY
The affinity constant (KA) and stoichiometry of binding of 4-pyridoxic acid S/-phosphate (4-pyridoxic-5'-P) complexed to the protein bovine serum albumin and aspartate aminotransferase were investigated by fluorometric methods. Whereas bovine serum albumin binds 1 mole of 4-pyridoxic-5'-P per mole of protein with a Ka = 4.6 X lo6 M-l, the enzyme, aspartate aminotransferase, binds 2 moles of 4pyridoxic-5'-P per mole of protein with a KA = 1.3 X lo6 IVI-1 at neutral pH.
The spectroscopic properties (absorption and fluorescence spectra) of the complexes were examined over a wide range of pH values.
At neutral pH, the formation of the 4-pyridoxic-5'-P bovine serum albumin complex causes the following spectroscopic changes in the ligand: In contrast to the protein bovine serum albumin, the apotransaminase causes a blue shift of approximately 10 nm in the band position of the emission spectrum and a concurrent decrease in the fluorescence quantum yield of the ligand (q = 0.05).
No change in the band position of the absorption spectrum of the complexed ligand was detected. The fluorescence changes observed in 4-pyridoxic-5'-P bound to the aminotransferase can be correlated with the deprotonation of the pyridine nitrogen atom as a result of its interaction with an amino acid residue at the binding site.
It was shown in earlier publications of our laboratory (1) that it is convenient to made use of the fluorescence properties of pyridoxamine-5-P to investigate the mechanism of binding of this cofactor to the catalytic site of the enzyme aspartate aminotransferase.
Subsequently, it was reported that the technique of polarization of fluorescence provides valuable information * This research was supported by Grant GB 33395 from the National Science Foundation. about the degree of mobility of the cofactor interacting with the enzyme (2).
It is the purpose of this paper to examine the spectroscopic properties of 4-pyridoxic-5'-Pr and to investigate the possibility of using this chromophore as a fluorescence probe of the binding sites of aminotransferases.
It is shown that 4-pyridoxic-5'-P, a photoproduct of pyridoxal-5-P, has the ability of forming stable complexes with the proteins, bovine serum albumin and aspartate aminotransferase. The spectral changes detected when 4-pyridoxic-5'.P is bound to the aminotransferase cannot be interpreted in terms of the "'hydrophobicity" of the binding site. However, these spectral changes can be correlated with the deprotonation of the pyridine nitrogen atom in the excited state as result of its interaction with an amino acid residue of the binding site.

EXPERIMENTAL PROCEDURE
Methods-Fluorescence emission spectra were obtained with the use of a spectrofluorimeter designed in our laboratory (3). The sample in a l-cm cuvette was illuminated by monochromatic light obtained by passing the output of a xenon lamp through a Bausch and Lomb monochromator.
The light emitted at right angles to the exciting light source was passed through a second Bausch and Lomb monochromator.
An EMI phototube (6256s) was used as the fluorescence detector.
Calibration of the exciting light source was carried out with solutions of Rhodamine B in ethylene glycol as described by Melhuish (4). The detector system was calibrated according to the method of White et al. (5). A bandwidth of 1 nm was used in the fluorometric measurements. Quantum yield of fluorescence (q) was calculated according to the method of Parker and Rees (6) with standards of known quantum yield. Two standards, quinine sulfate (p = 0.46) and 2-aminopyridine (Q = 0.60), were chosen for these esperiments (7).
Polarization of fluorescence measurements were performed in an apparatus similar to that described by Weber (8). Illumination was provided by a xenon lamp (200 watts) with wave lengths selected by a quartz prism monochromator.
The bandwidth for excitation at 320 nm was 5 nm. Fluorescence polarized light was passed through a combination of Corning C-S-3-73 and C-S-5-58 glass filters. The detector system consisted of an EM1 9502B photomultiplier and a digital voltmeter. The 1 The abbreviation used is: 4-pyridoxicd'-P, 4-pyridoxic acid 5'-phosphate. 6954 degree of polarization of fluorescence was measured with a pre-intensity falls in the manner depicted in Fig. 1 when the pH of cision of *O.Ol. the solution is varied in the range 8 to 11. Fluorescence decay time measurements were conducted in a TRW nanosecond decay time fluorimeter equipped with a model 32A decay time computer and a Tektronix 556 dual beam oscilloscope (9).
Exciting light was provided by a deuterium lamp. A Baird Atomic interference filter (maximum of transmission at 322 nm) was used for excitation and a Corning C-S-3-73 cut off filter for emission.
The decrease in fluorescence intensity is interpreted as being due to the deprotonation of the ring nitrogen atom. From the plot of fluorescence intensity versus pH, it can be estimated that the pK of deprotonation of the ring nitrogen atom is about 10.5. At pK values above 10.5, the dipolar ion-phosphate dianion form (III) shows very low fluorescence yield (Table I).
1MateriaZs-Aspartate aminotransferase from pig heart (L-aspartate-2-oxoglutarate aminotransferase, E-C-2-11) was prepared by the method of Jenkins et al. (10) with the modification of Martinez Carrion et al. (11). The cytoplasmic fraction (ol) was used throughout these studies. Resolution of the holoenzyme into apoenzyme and free cofactor was performed according to the method of Scordi et al. (12). The resulting apoenzyme had less than 1% of the original aminotransferase activity. Enzymatic assays of aspartate aminotransferase were conducted in the Cary model 15 spectrophotometer as described in a previous paper (1).
The protein bovine serum albumin was purchased from Sigma and used without further purification.
NADH and the enzyme, malate dehydrogenase, were obtained from Boehringer-Mannheim.
Other materials, commercially obtained, were of the highest purity available.
A decrease in the polarity of the environment by addition of the solvent, dioxane, changes the absorption and fluorescence characteristics of 4-pyridoxic-5'-P and pyridoxic acid. As shown in Table I, the emission spectrum of 4-pyridoxic-5'-P undergoes a blue shift of approximately 10 nm when examined in the solvent system, water-dioxane (2 :98, v/v). This shift in the band position of the emission spectrum is accompanied by a red shift in the absorption spectrum (Table I). Another fluorometric parameter, the fluorescence lifetime, was examined over the pH range 3 to 8. Despite the increase in fluorescence yield induced by protonation of the carboxyl group (pK 5.5) 4-pyridoxic-5'-P has a fluorescence lifetime of 9 ns at either pH 7.4 or pH 3. The fluorescence lifetime of 4-pyridoxic-5'-P in solvent mixtures fluctuates between 9 and 11 ns, the latter value being observed in the solvent mixture containing dioxane. Interaction of 4-Pyridoxic-6'-P with Serum Albumin-The effect of bovine' serum albumin on the fluorescence properties 4-Pyridoxic-5'-P was prepared by the method of Morrison and Long (13) by irradiating pyridoxal-5-P with light having wave lengths longer than 300 nm in the presence of oxygen.

RESULTS
The effect of pH on the relative distribution of 4-pyridoxic-5'-1 structures in solution was examined by fluorescence spectroscopy.
At neutral pH, the tripolar ion-phosphate dianion form of 4-pyridoxic-5'-P (II) displays an emission band centered at 425 nm. At pH values lower than 6, the protonation of the carboxyl group (pK = 5.5) brings about a blue shift of approximately 5 nm in the emission spectrum. The dipolar ion-phosphate monoanion form fluoresces intensely at 420 nm (I).
Its fluorescence yield is at least a-fold increased when compared with the species which exists at neutral pH (Table I).
As the pH of the solution is increased from 7 to 11, the spectroscopic characteristics of 4-pyridoxic-5'-P are changed. A blue shift in both absorption and fluorescence spectra is easily detected at pH values above 10. Concomitantly the fluorescence of 4-pyridoxic-5'-P was examined at several pH values. In these experiments protein solutions of approximately 3 mg per ml were used as solvents; and in order to maximize the binding of the ligand the molar ratio of protein to 4-pyridoxic-5'.P was of the order of 3: 1. The fluorescence spectrum of 4-pyridoxic-5'-P in the presence of bovine serum albumin was then recorded and compared with the spectrum of free 4-pyridoxic-5'-P. It was found that over the pH range from 6 to 8 the protein bovine serum albumin promotes a red shift of 15 nm in the fluorescence emitted by the ligand (Table II and Fig. 2).
This shift in the band position of the emission spectrum is accompanied by a small change in the band position of the absorption spectrum (Table II), and by an increase in the fluorescence lifetime (12 ns). Fig. 3  polarized fluorescence emitted by 4-pyridoxic-5'-P as a result of the addition of increasing concentrations of protein to a fixed concentration of 4-pyridoxic-5'-P (1.5 X 10S5 M). The measured polarization of fluorescence reflects the relative contribution of both free and bound ligand to the fluorescence of the system. When 4-pyridoxic-5'-P is the only component of the system, the polarization (PO = 0.05) is related to the rotational motion of free 4-pyridoxic-5'-P between excitation and emission of light. As soon as the protein is added to the system and the 4-pyridoxic-5'-P protein complex formed, the bound 4-pyridoxic-5'-P molecules are no longer able to rotate extensively during the lifetime of the excited state (12 nsec), consequently the polarization, P, of the system is increased in the manner depicted in Fig. 3.
When the protein concentration is such that the ligand is complexed, the polarization of fluorescence approaches the value Pe = 0.32.
The increase in polarization of fluorescence associated with complex formation can be used to determine the affinity constant   The fluorescence intensity of free C-pyridoxic-5'-P (8'0) and the fluorescence emitted when both free and bound ligand are in equilibrium (8') were recorded using unpolarized exciting light of 320 nm.
The results of these fluorescence measurements were analyzed by the method of Scatchard (18) (Equation 2). Since the results of the fluorometric titrations are adequately represented by Equation 2, it was possible to determine one binding site with an association constant of 4.6 X 10b M-I at pH 7.4 in 0.05 M Tris-acetate (Fig. 5). Although an association constant of similar magnitude was determined in 0.05 M phosphate buffer (pH 7.4), it should be emphasized that the stability of the bovine serum albumin-4-pyridoxic-5'-P complex is affected by changes in the ionic strength of the medium (Table II). This is illustrated by the experiments included in Fig. 6B, where polarization of fluorescence measurements were conducted on samples containing bovine serum albumin and C-pyridoxic-5'-P (mixing molar ratio 3:l) at varying ionic strengths.
As the ionic strength of the medium is increased from 0.1 to 0.5, the degree of polarized fluorescence emitted by 4-pyridoxic-5'-P is decreased from 0.32 to 0.1. Concomitantly, the maxi- mum of emission is shifted from 440 to 425 mn as the concentration of NaCl is increased.
At an ionic strength of 0.5, the fluorescence emitted by the system bovine albumin-4-pyridoxic-5'-P is highly depolarized (P = 0.08) and the maximum of emission corresponds to that of free 4-pyridoxic-5'-P. The stability of the complex bovine serum albumin-4-pyridoxic-5'-P is also affected by changes in the pH of the solution.
At pH values lower than 5.5, the polarization of fluorescence and the maximum of emission of 4-pyridoxic-5'-P remains essentially invariant upon the subsequent addition of increasing concentrations of bovine serum albumin (Fig. 614). This is taken as an indication that protonation of the carboxyl group of 4-pyridoxic-5'-P prevents the binding of this ligand to the protein.
Interaction of J-Pyricloxic-5/-P with Apotransaminase-The interaction of 4-pyridoxic-5'-P with the apoprotein of the enzyme aspartate aminotransferase induces several changes in the fluorescence properties of the ligand.
As shown in Fig.  2, the addition of 4-pyridoxic-5'-P to the apoprotein (molar ratio 1:2) induces in parallel to diminution of fluorescence yield, a change in the band position of the emission spectrum.. Thus the fluorescence yield of bound 4-pyridoxic-5'-P is lower than the corresponding yield of free ligand at pH 7.4, while the emission spectrum centered at around 415 nm is shifted towards shorter wave lengths when compared with to free 4-pyridoxic-5'.P (425 nm) in the same buffer.
The fluorescence quenching effect associated with complex formation was used to determine the affinity constant of 4-pyridoxic-5'-P for the apoenzyme. where FM is the actual observed fluorescence when all the ligand is bound to the protein.
FM was determined directly by adding increasing concentrations of apoenzyme to a fixed concentration of 4-pyridoxic-5'-P.
The average number of ligand molecules bound per mole of apoenzyme (0) was calculated for points along the titration curve by means of Equation 4. where LO and PO are the total ligand and protein concentrations, respectively. The results of the fluorometric titrations yield a straight line when O/(L) is plotted versus ti (Fig. 8). An association constant of 1.3 X lo6 Me1 and n = 2 is determined from this plot.
Since the addition of 4-pyridoxic-5/-P to the apoprotein of the enzyme aspartate aminotransferase inhibits the reconstitution of aminotransaminase activity, it was desirable to determine the affinity of the inhibitor for the apoprotein by a method based on enzymatic activity measurements.
To this end, samples of apoenzyme (0.1 mg) were incubated with varying concentrations of 4-pyridoxic-5'.P in 1 ml of 0.05 M Trisacetate buffer (pH 7.4) at 25" for 1 hour. The solutions were then diluted with a solution of 4-pyridoxic-5'-P (low4 M) and kept at 25" for 1 hour. The solutions were then diluted with a solution of 4-pyridoxic-5'-P (lo-' M) and kept at 25" for 1 hour.
Aliquots were removed and assayed for enzymatic activity.
The results of the inhibition studies were analyzed by means of Equation 5.
where VC, is the velocity of the enzymatic reaction in the absence of inhibitor and V is the velocity in the presence of inhibitor. This method gives an inhibition constant of 2 X 106 ~-1, which agrees well with the association constant determined using the fluorometric method previously described. In the experiments performed with the apotransaminase, it was found that there was a strict correlation between the quenching of 4-pyridoxic-5'-P fluorescence and its ability to inhibit reconstitution of aminotransferase activity.

6958
The fluorescence quenching effect, which is observed over a wide range of pH values (6)(7)(8)(9), is paralleled by inhibition of transaminase activity (Table III).
Furthermore, the addition of increasing concentrations of NaCl up to a concentration of 0.5 M has no effect upon the magnitude of either the fluorescence quenching or the degree of inhibition of transaminase activity. However, the magnitude of the fluorescence quenching effect induced by interaction of 4-pyridoxic-5'-P with the apotransaminase is strongly influenced by the anions of the buffers used in the incubation mixture.
Although carbonate, bicarbonate, and acetate have no effect on the binding of 4-pyridoxic-5'-P to the apotransaminase, it was observed that the apoprotein dissolved in 0.1 M phosphate buffer (pH 7.4) is completely insensitive to the subsequent addition of 4-pyridoxic-5'-P.
The protection afforded by phosphate ions against the inhibitory action of 4-pyridoxic-5'.P is also shown by fluorescence measurements.
As illustrated in Table III, there is no quenching of 4-pyridoxic-5'-P fluorescence when this compound is mixed with the apoenzyme in the presence of 0.1 M phosphate buffer. Thus it appears that phosphate anions compete with the phosphate group of ii-pyridoxic-5'-P and prevent the binding of this ligand to the protein.
In this connection it should be noted that the presence of a phosphate group in the chemical structure of 4-pyridoxic-5'-P is essential for binding and inhibition of reconstitution of transaminase activity. This is shown by the absence of both fluorescence quenching and inhibitory effects when pyridoxic acid is used instead of 4-pyridoxic-5'-P (Table II).

DISCUSSION
The aim of the studies reported in this paper was to investigate the fluorescence properties of 4-pyridoxic-5'-P bound to the proteins bovine serum albumin and aspartate aminotransferase.
The results of the polarization of fluorescence titrations indicate that the presence of a carboxyl group in the chemical structure of 4-pyridoxic-5'-P and pyridoxic acid is essential for binding to the protein bovine serum albumin.
The addition of 4-pyridoxic-5'-P to a solution of bovine serum albumin causes a red shift of approximately 15 nm in the band position of the emission spectrum and a concurrent increase in the fluorescence lifetime of the ligand.
In order to understand the spectroscopic changes of 4-pyridoxic-5'-P complexed to the protein bovine serum albumin, it is important to realize that the red shift in emission is more pronounced and of opposite direction than the corresponding shift in the absorption spectrum (2 nm) Table II. According to accepted theories (19) dealing with solvent effect on electronic spectra, a behavior of this kind reflects the ability of the solvent molecules to reorient themselves during the lifetime of the excited state. Although relaxation effects are not well understood when the ligand molecules are strongly adsorbed to the protein surface, it is likely that emission shifts to lower energies are also influenced by strong interactions of the ligand with polar amino acid residues at the binding site.
In contrast to the 4-pyridoxic-5'-P-bovine serum albumin complex, the emission spectrum of 4-pyridoxic-5'-P bound to the apotransaminase displays a blue shift of approximately 10 nm when compared with free 4-pyridoxic-5'-P at neutral pH. This fluorescence shift to higher energy is accompanied by a pronounced decrease in the fluorescence quantum yield of the complexed ligand.
The presence of a phosphate group in the chemical structure of 4-pyridoxic-5'-P is essential for binding and inhibition of aminotransferase activity, since pyridoxic acid failed to interact with the apoenzyme as judged by fluorescence and activity measurements.
In order to explain the spectroscopic changes induced by binding of 4-pyridosic-5'-P to the apoprotein, it is worthwhile to compare the spectroscopic properties of the complexed ligand with those of free 4-pyridoxic-5'-P in aqueous solutions of varying pH.
An analysis of the emission properties of 4-pyridoxic-5'-P reveals that the dipolar ion-phosphate dianion form (III), which is the predominant structure at alkaline pH, has fluorescence characteristics similar to that of 4-pyridoxic acid-5'-P complexed to the apotransaminase.
Thus, a blue shift in the emission spectrum and a concurrent decrease in the fluorescence quantum yield are observed when 4-pyridoxic-5'-P forms a complex with the apoenzyme.
In view of these considerations, it seems reasonable to propose that the transfer of a proton from the pyridine nitrogen atom to an acceptor amino acid residue on the protein is responsible for the fluorescence changes observed.
It should be noted that a blue shift in the emission spectrum is detected when 4-pyridoxic-5'-P is dissolved in solutions of lower polarizability than water; however, the absorption spectrum is shifted towards the red and the fluorescence yield is enhanced (Table I). The spectroscopic changes induced by interaction of a ligand with a protein may provide valuable information about the microenvironment of the binding site. However, it is evident that the binding site of a protein differs from the environment of the ligand immersed in a solvent system. Therefore, it is reasonable to ask whether the spectroscopic changes detected when the ligand is bound to the protein can be correlated with the results obtained when the ligand is dissolved in well defined solvent mixtures.
Recent experiments by Brand and Gohlke (20) have shown that one must be careful in the interpretation of emission changes induced by binding of the ligand N-arylaminonaphthalene sulfonate to the protein bovine serum albumin, since the fluorescence characteristics of the dye adsorbed to the macromolecule are influenced by factors other than the polarizability of the environment. Some caution must also be exercised in drawing conclusions concerning the detailed characteristics of the binding site of 4-pyridoxic-5'-P.
There are two aspects of the studies presented in this paper that are worthwhile considering in the interpretation of the fluorescence changes. (a) The fluorescence properties are affected by the nature of the protein to which the ligand is bound.