Pyridoxal phosphate-dependent reactions in water pools

Biochimie 71 (1989) 461 - 469 ~) Socict6 dc Chimie biolog;quc/Elsevier, Paris

461

Pyridoxal phosphate-dependent reactions in water pools Costantino S A L E R N O * , Andrea L U C A N O and Paolo F A S E L L A

Laboratory of Biochemical and Clinical Analysis, University of Chieti, Department of Biochemical Sciences, University of Rome La Sapienza; and Department of Experimental Medicine, U'nive,:i~ of Rome Tor Vergata, Italy (Received 16- 7-1988, accepted after revision 21- I i-1988)

S u m m a r y m Pyridoxal 5'-phosphate, its Schiff base with L-alanine, and cytoso!ic aspartate aminotransferase were dissolved in isooctane solutions containing reverse micelles of the surfactant di-2-ethylhexylsodium sulfosuccinate and water. The physico-chemical properties of these compounds in the new environment have been studied. reverse micelles / pyridoxal 5 ' - p h o s p h a t e / aspartate aminotransferase / band-shape analysis

Introduction Water pools can be prepared by dissolving a small amount of water in apolar hydrocarbon media containing amphiphilic compounds [1-6]. Presumably, the pools of water are surrounded h_~ V ~ . .l l.r.f .a .e .t a. n. t. . . .m. .~.l .~.r ~. .l l.o.c. . . .cr~ . . . tth.~e ~h.~ ,,,,. .,-.--In. V,_,,a, groups dip into the water, whereas the hydrophobic chains lie in the aprotic solvent. Aerosol OT (AOT, di-2-ethylhexylsodium sulfosuccinate) is an anionic surfactant commonly used in these studies. Since both the hydrophilic and iipophilic functions are strong and not too unequally matched, A O T is an excellent co-solvent of both organic solvent and water, and thus reverse miceiles formed by this surfactant solubilize relatively large amounts of water [7]. Water pools ,can solubilize several compounds of biological interest, including enzymatic proteins and metabolites [8, 9]. The interest in these systems lies in their biotechnological potentialities (described in detail in recent reviews [6, 8, 9]) Moreover, reverse micelles could ",aime bic. logical structures and contribute to the solutioi. of basic problems in structural biochemistry related to the influence of micro-environments on the vectorial properties of biomolecules.

In this paper we report our study on vitamin B6 derivatives in water pools. We found that the equilibria for hydration av.d tautomerization of pyridoxal 5'-phosphate are varied in this new environment. Aspartate a~motransferase can be included in water pools, a.t slightly alkaline pH the catalytic properties of the enzyme are similar to those in bulk water. At lower pH values coenzyme dissociation and irreversible enzyme inactivation have been observed.

Materials and methods The cytosolic isoenzyme of aspartate aminotransferase (EC 2.6.1.1) was purified from pig heart to apparent electrophoretic homogeneity by the procedure described in [10]. All other reagents were analytical grade commercial products from Merck A.G., Boehringer A.G., Fluka A.G. and Sigma Inc. AOT preparations were made at least partially free from UV absorbing impurities, acid contaminants and salts. Apparently, there were several UV absorbing impurities in the comtnercial samples. Some of them, namely those absorbing ~290 nm, were insoluble in methanol and could be easily eliminated by the method reported by Wong et al. [11]. The HPLC procedure described by Martin and

*Correspondence and reprints: Costantino Salerno, Dipartimento di Scienze Biochimiche, Citt:~ Universitaria, P. le A. Moro 5, l-(lOlO0 Rome, Italy.

462

C. Salerno et al.

Magid [12] appeared to be useful in reducing the concentration of impurities absorbing in the 250-285 nm range. Acidic contaminants were present in different amounts in the various commercial samples. The concentration of these compounds seemed to increase in the course of purification procedures involving heating steps. These contaminants could be reduced by treating the samples with basic Alox at the end of the purification procedure [13]. Salts could be present in the commercial samples of AOT. Some of them were formed during the neutralization of the acidic contaminants. The concentration of salts in AOT preparations was kept as low as possible because water solubility was markedly dependent on salt concentration in the mixture [12]. Homogeneous microemulsions were obtained by injecting small amounts of aqueous solutions into mixtures of isooctane and AOT. Buffered saline (10 mM natrium phosphate or 10 mM natrium citrate) was dissolved into the isooctane-containing medium to approximate the apparent pH of the overall mixture to the appropriate values. After mechanical agitation, the mixture of isooctane, water, AOT and hydrophilic solutes appeared to be clear and stable, thus permitting spectroscopic studies in the visible and near ultraviolet region. Water content in the micellar system was determined by NMR spectroscopy. The distribution of solutes in the various phases constituting the heterogeneous system was studied by fractional centrifugation (60 min, 39,000 g). Potentiometric determinations of the apparent pH were made by dipping electrodes directly into the microemulsion. The measurements were affected by some noise whose intensity (ranging from 0.02-0.1 pH units) increased by decreasing water content in reverse micelles. As far as the accuracy of the potentiometric determinations was concerned, we found that the apparent pH of the overall mixture was almost independent of water:AOT molar ratio (ranging from 5-55) and water concentration (ranging from 0.2-3.0 mmoVml isooctane). When water concentration was so high as to cause partial separation of the aqueous phase from the organic solvent, the apparent pH of the aqueous solution on the bottom of the test tube was approximately equal to the apparent pH of the isooctane-containing layer. When the electrodes were immersed in mixtures containing pH indicators (phenol red, chlorophenol red, bromphenol blue) the meter response agreed fairly closely with the pH determined from the absorption spectra of the dyes, provided that water content in isooctane was sufficiently high (~3 m m o l / ml isooctane). Only when the water content was small did large deviations appear between potentiometric and spectrophotometric measurements, perhaps owing to interactions between dyes and surfactanl [a] Detailed analysis of the problems related with an unequivocal determination of the micellar pH.is given in [8, 14, 15]. Electronic absorpt,on spectra were recorded in digital form on a Perkin-Elmer 330 spectrophoto-

meter or on a Varian DMS 90 spectrophotometer connected to an Apple 2E computer. The spectra were analyzed as sum of log-normal curves [16]. Each curve behaved as a skewed Gaussian curve and was specified by 4 parameters characterizing the shape of the absorption band (peak position, height, bandwidth at one-half the peak height, and an index of skewness [16]). Curve fitting was done on plots of molar absorptivity against wave number, because absorption bands in various wavelength regions tended to have much more similar bandwidths and shapes when spectra were plotted on a linear scale against energy rather than wavelength [16]. Attempts were made to find internally consistent sets of band parameters that could be directly related to literature data [17-19]. Width and skewness values for minor bands were fixed at preselected values. In the other cases, the best fit was selected primarily upon the basis of a minimum in the sum of residual squares. Aspartate aminotransferase activity was assayed by a 2-step procedure. First, the transaminase-catalyzed reaction was conducted at 25°C in the presence of aspartate and 2-oxoglutarate and was terminated by adding 0.5 mmol HCI / ml water; second, the mixture was neutralized with 0.5 M NaOH and the reaction product, oxaloacetate, was measured spectrophotometrically by a N A D H coupled enzyme method [20]. Oxaioacetate in reverse micelles was measured after separation of the aqueous phase by centrifugation.

Results and Discussion Pyridoxal 5'-phosphate can be dissolved in isooctane containing reverse miceiies of A O T ( 5 - 5 0 mM) and water ( 0 . 3 - 3 . 0 M), but not in anhydrous mixtures of isooctane and A O T . The microemulsions are stable and clear and absorb light in the visible and near ultraviolet region. Chromophore and water have similar sedimentation rates and can be separated from the organic solvent by centrifugation (Fig. !), suggesting that the coenzyme is either solubilized in the aqueous phase or absorbed on the wall of the reverse micelles. The spectral properties of pyridoxal 5'-phosphate are strongly dependent on the apparent p H of the overall mixture and on the composition of the micellar phase. As in bulk water [17]. acidification leads to a stepwise blue shift of the absorption maximum. At least 2 p H - d e p e n d e n t transitions (with isosbestic points around 350 nm and 310 nm, respectively) can be observed (Fig. 2). The apparent pKa's of the transitions are approximately 7.5 and 4.0. These values fairly agree with the macroscopic pK~ values reported for pyridoxai 5'-phosphate in bulk water [17,

21].

Pyridoxal phosphate in water pools At pH values far from neutrality, the spectra resemble those observed in bulk water (Figs. 3A, B) and their deconvolution gives rise to results practically coincident with literature data (Table 1). By contrast, the spectra recorded at pH around neutrality are quite different from those observed in aqueous solvent. They are characterized by 2 major bands centered at 388 and 329 nm (Figs. 3C, D). The 388 nm-absorbing band shifts to 329 nm when detergent content in the micellar system increases with respect to water or pyridoxal 5'-phosphate content. The transition shows an isosbestic point at =350 nm (Fig. 4). The position, width and skewness of the band at 388 nm are very similar to those reported for the neutral dipolar ionic aldehyde form of pyridoxal 5'-phosphate [17]. Since this species should have a second small (=5 five times less intense) absorption band located ~272 nm, we have placed such a band into the deconvoluted spectra even though covered by other much larger bands. We did not take into account the protonation state of the phosphate group, since dis-

463

sociation of a proton from this group at pH around neutrality has a very small effect on band parameters [17]. The deconvolution of the central part of the curves obtained at pH around neutrality is more complex. The spectrum of reverse micelles containing a relatively large amount of surfactant is accounted for by a major band centered at 329 nm. If the detergent content in the micelles decreases with respect to the other components, the best fit of the curve can be obtained

2.o

Ft

:°" g-o.4

] LLJ z

""'. ',

•

o 1

0

300

~

400

'

500

-2.0 WAVELENGTH rr"

-1.6 U3

-1.2 I -0.8

4¸

21 e..

2

:~

0

'-"

I

•

0.

,

,

,/

,r"

h

.,

i/b.l

"

0!

. . . .

CO

c'r

j

t--I

tic: t.iJ F-.-

nm

- 12"4I

u.J u Z O2 ~ ry (23 u")

,

-0.4

~mO.

'"

t

0.0

1

2

5

4

5 -

FR@CTION

NUMBER

Fig. 1. Sedimentation rate of micellar suspensions of water, AOT, and pyridoxal 5'-phosphate (11000:400:1 by tool) in isooctane at pH 4.0. The fractions correspond to samples withdrawn from a test tube after 60-min centrifugation at 39 000 g. Fractions 1 and 5 are samples at the top and at the bottom of the centrifugation tube, respectively. Black bars indicate water content in the fractions. White bars indicate absorbance values (O, 300 nm; [], 330 nm; A, 390 nm). Inset: chromophore absorbance as a function of water concentration.

o

3~]o

400

WAVELENGTH

,

soo ,

nm

Fig. 2. Effect of the apparent pH on the spectral properties of pyridoxal Y-phosphate (PLP) in reverse micelles. The ratio [H20] / [AOT] was 55. A: the suspensions contained 5 #tool P L P / m l H20; the apparent pH was equal to 2.5 (a), 3.6 (b), 6.0 (c), 7.3 (d), 10.1 (e). B: the suspensions contained I/.tmol P L P / m l H20; the apparent pH was equal to 2.4 (a), 4.3 (b), 6.8 (c), 7.8 (d), 9.4 (e). Insets: absorbance at 330 nm as a function of the apparent pH of the mixtures.

~

i

_~.

~ .

N'o~

•

?'~.

~ ; "

3

"IT

Fq C" Fq

3

w

rq FFq Z

U'l ED

0 0

trl 0

L~I 0 0

N ~S1 0

UI O

CD CD

L~ (Jl CO

L~ 0 CD

N Oq CD

MOLAR

I

~OLAR

r-o

ABSORPTIVITY

ABSORPTIVITY

C'~

01

(x I0 -3 )

cn

Cx I0 -3 )

3

m

MOLAR

MOLAR

ABSORPTIVITY

ABSORPTIVITT

(×

(×

-3)

lO -3)

I0

M

Pyridoxal phosphate in water pools by assuming that the band centered at 329 nm overlaps another band with the same structure located at 320 nm. Both bands closely resemble B

~o 6

m .-.,

~- 4

~-

m

4

>-

e

I--. o_

o ff'l

,=, crl

n.-

m 2 f-rrr rl.d 0 ~--

r,o I,,'3

~

0

,-,> o t--1

tq ~

~

40

[H20]

/

[ROT]

B

13

rrl

4t~ •.. i:

cx: ...I

m m []

ta

-Jr"

0

360

4

WRVELEPI3TH

,

o

0 nm

[ l P L I =' ]

o.

boo

/

[FIOT]

Fig. 4. Effect of the miceile composition on the spectral propcrties of pyridoxal 5'-phosphate (PIP) at pH 7.11. Left: Absorption spectra of reverse micellcs cor,~aining 5 pmol PLP/ml H,O. The ratio IH:O1/[AOTI was II (a), 22 (b), 33 (c), 55 (d). Right: Molar absorptivity at 330 nm as a function of thc ratio [HzO]/[AOT] in miccllcs containing 5 p,mol PLP/ml H,O (A) and as a function of thc ratio [PLP[/[AOT I in miccllcs containing 55 tool H~O/mol AOT (B).

4

65

the low-energy band of the neutral dipolar ionic hydrate form of pyridoxal 5'-phosphate, which is much less intense in bulk water and usually shows an absorption m a x i m u m = 3 2 0 - 3 2 2 nm [17]. The presence of unusually high concentrations of hydrate forms of pyridoxal 5'-phosphate in reverse micelles does not differ from what is known about the hydration equilibria of the coenzyme in non-aqueous solvents [16, 22]. An increase of the hydrate forms in d i o x a n e - w a t e r mixture has been observed by Sanchez-Ruiz et al. [22], who described the apparently paradoxical p h e n o m e n o n in terms of the Marshall model for multiple equilibria [23]. It is well known [ 1 - 3 , 24] that water in proximity of the micelle wall behaves as a peculiar solvent with properties similar to those of less polar media: the higher the concentration of surfactant, the stronger the influence of this anomalous aqueous layer on the solutes [ 1]. Minor bands have been placed in the valleys of the spectra to obtain a better overall fit of these regions, The band present at 290 nm in the deconvolutions of the spectra in Figs. 3B, C, and D could be tentatively attributed to the uncharged hydrate form of pyridoxal 5'-phosphate [17], whose low energy band is thought to lie at about this wavelength (the relative concentration of uncharged species increases in media with low polarity [16]). Another small band has been placed in some cases at 370 nm (Fig. 3C) to mini-

Table ]. Pyridoxal 5'-phosphate band: parameters for resolved absorption spectra.

Ionic form

H3P(a +) (h*) HP(a_+)

Reverse micelles

Bulk water §

band position (nm) (kK)

width (kK)

skewness

band position (nm) (kK)

width (kK)

skewness

336 296 388 272

29.76 33.78 25.77 36.71

4.61 3.37 4. !0 4.63

1.24 1.46 1.36 1.32

329 29(i 392 263 302

30.39 34.48 25.51 38.02 33.11

3.48 4.24 4.50 4.71 3,69

1.32 1.32 1.35 1.30 1.36

336 294 390 271 352 320

29.76 33.99 25.59 3o.91 28.41 31.25

4.61 3.25 4.00 4.67 4.70 3.46

1.24 1.46 1.36* 1.32 1.32 1.32

390 261 302

25.66 38.32 33.11

4.46 4,73 3,63

1.32 1.30 1.36

(a °)

(h~) (h,,) P(a-) (h-)

~From: Harris et al. [15]. *Since protonation of the phosphate group is known to involve only small changes in bands with low energy, we did not attempt to take'this protonation into account in the spectral resolution.

C. Salerno et al.

466

mize misfits. It could be attributed to minor conformers of p/ridoxal 5'-phosphate or to artefacts [ 16]. If amino acids are dissolved in the micelles, the spectra become similar to those reported for the Schiff base of the coenzyme (Fig. 5). At saturating concentrations of L-alanine, an amino acid readily soluble in water pools, the spectra are characterized by 3 m a j o r bands (centered at 410 nm, 329 nm, and 285 nm, respectively) and by a minor band around 360 nm, which is detected only at relatively high p H [18]. The bands at 410 iim and at 285 nm have nearly equal intensity and resemble the low-energy bands of the ketoenamine t a u t o m e r [18]. The band at 329 nm is very small at slightly alkaline pH, but it increases in intensity at pH around neutrality, especially if the molar ratio between surfactant and chrom o p h o r e is relatively high. This band could be attributed to the enolimine t a u t o m e r that is known to increase in solvents which are less polar than water [25]. A n o t h e r possibility is that the band at 329 nm is due to a pyridoxal 5'-phosphate form which reacts with difficulty (or does not react at all) with the amino acid. In line with the latter hypothesis is the observation that a b a n d at 329 nm is present at p H around neutra-

! o

B

6

× "-"

~4 > '-" I---a..

tn 2 m a: --J

=CO

300

400

WF~VELENGTM

500 ,

17m

C

I

o6 X

A

I

o6

~-4 I,,.==i

x 13nO

m 2 rr

[ _J

$2

x: 0

Cl~

300

400

WFIVELENGTH ~'0

300 WQVELENGTM

50

400 ,

nm

S00 ,

n m

Fig. 5. Band-shape analysis of the absorption spectra of the Schiff base of L-alanine with pyridoxai 5'-phosphate (PLP) in reverse micelles. The apparent pH of the miceilar suspensions was equal to 9.0 (A) and 7.0 (B, C). The ratio [AOT]/[PLP] was 1000 (A, B) and 200 (C). Other conditions were as described m Fig. 3.

Pyridoxal phosphate in water pools iity also in the spectra obtained in absence of t.alanine (Fig. 3) and that the position, width, and skewness of the band are unmodified after addition of the amino acid. Since we obtained monophasic hyperbolic curves with very similar shape by tritrating pyridoxal 5'-phosphate with L-alanine in reverse micelles and in bulk water (Fig. 6), it can be concluded that the conversion of poorly reactive species of the coenzyme, if present, into more reactive forms is so limited a n d / o r slow as to be negligible. A proposal is that the 329-nm band is due to pyridoxal 5'-phosphate (perhaps a neutral ionic dipolar hydrate form) squeezed [2] from the pool interior onto the interface to form stable aggregates with the surfactant. Deconvolution of the spectra of aspartate aminotransferase in reverse micelles provides information on the properties and reactivity of the enzyme-bound coenzyme in the new environment. The aminotransferase can be easily dissolved in water pools, provided that, as described in

467

similar cases 15], the molar ratio between water and surfactant is higher than a critical value (-~50). Under these conditions, the spectral properties of the enzyme-bound coenzyme are strictly dependent on the apparent pH of the overall mixture. In alkaline suspensions (Fig. 7), the spectra are characterized by a major band at 361 nm, nearly identical to that reported for the high-pH form of the enzyme in bulk water [19]. At pH values slightly below neutrality (Fig. 8), 3 other bands (centered at 430 nm, 388 nm, and 329 nm, respectively) become evident. The structure of the band at 430 nm resembles that of the N-protonated ketoenamine form of the enzyme [19]. Since the bands at 388 nm and at 329 nm are nearly equal to those elicited by free pyridoxal 5'-phosphate in neutral water pools (Fig. 3), we tentatively attributed these bands to the coenzyme resolved from the protein. The hypothesis seems to be substantiated by the observation that, when the apparent pH of the micellar suspension is brought again to

!

O q--'4

A

x >.... ~. t:: o-,-f I.-..fl....

×

il

6

Lk.lt..j

i . "::::

.."""

":':':

z40 twO t.O rr~ 132

cr" [a,," ("r ,--I (::) Is:..

0 •

j

J,

O. 4

E@LRNINE]

J

O .

O. 8 ,

M

Fig. 6. Molar absorptivity variation of pyridoxel 5'-phosphate (PLP) at 410 nm in the presence of L-alanine. L-alanine concentration was calculated with respect to the aqueous phase. The apparent pH of the mixtures was 7.0. O: reverse micelles ([H20] / [AOT] = 55; [AOT] / [PLP] = 1000). $: bulk water.

!

300

400 WFIVELENGTH

500 ,

nm

Fig. 7. Band-shape analysis of the absorption spectra of aspartate aminotransferase in reverse micelles at pH 8.0. The ratio [AOT]/[Enzyme] was 1000. The apparent pH of the micellar suspension was brought to 8.0 from values around neutrality (a) before and (b) after addition of the enzyme solution into the mixture. Other conditions were as described in Fig. 3.

468

C. Salerno et al.

slightly alkaline values, the spectrum differs from the original one probably because of partial denaturation (Fig. 7). Moreover, we found that enzyme activity roughly parallels the spectral changes. In the alkaline pH range, ,.he enzyme catalyzes transamination of aspartate to 2-oxoglutarate as well as in bulk water (Fig. 9). The activity markedly decreases at pH values slightly below neutrality, and cannot be restored oy again increasing the apparent pH of the micellar suspension. The reason for the irreversible enzyme inactivation is not clear. It is suggested that protonation of amino acid residues on the protein surface results in interactions with the negatively charged surfactant groups on the pool interface, which could promote distortion of the overall architecture of the enzyme. Sequestration of pyddoxal 5'-phosphate, which results from resoluti~"~ of denaturated enzyme, on the micelle wall could contribute to render the inactivation process irreversible.

30

20

4

i0

I

0 l / [ Fls i:, ]

,

0.8 -I mM

I

1.2

Figure 9. Double-reciprocal plot of the initial velocity of the transaminase-catalyzed reaction at pH 9.0 versus aspartatc concentration. Aspartate c~ncentration was calculated with respect to the aqueous phase. 2-Oxoglutaratc concentration was held constant (1 /~mol/m| H20). ~: reverse micelles ([H2OI / [AOT] = 55). 4,: bulk water.

40

~''

I

0.4

°

!¢1 O

Acknowledgments

x

This work was partially supported by a grant from Ministero della Pubblicca lstruzione, Italy.

,,, 2O ~.o Z oZ rn r~ C3

References

tn

300

400 WFIVELENGTH

500 ,

nm

Fig. 8. Band-shape ~malysis of the absorption spectra of aspartate aminotransferase in reverse miceiles at (a) pH 6.0 and (b) 6.5. Other conditions were as described in Fig. 7.

1 Menger F.M., Donohue J.A. & Williams R.F. (1973) J. A m . Chem. Soc. 9 5 , 2 8 6 - 2 8 8 2 Menger F.M., Saito G., Sanzero G.V. & Dodd J.R. (1975)J. Am. Chem. Soc. 9 7 , 9 0 9 - q l 1 3 Menger F.M. & Saito G. (1978) J. Am. Chem. Soc. 100, 4376-4379 4 Menger F.M. & Yamada K. (1979)J. Am. Chem. Soc. 101, 6731-6734 5 Eicke H.F. & Christen H. (1974)J. Colloid Interface Sci. 46, 4 1 7 - 4 2 7 6 Martinek K., Levashov A.V., Klyachko N., Khumelnitski Y.L, & Berezin I.V, (1986) Eur. J. Bio-

Pyridoxal phosphate in water pools chem. 155,453-468 7 Eicke H.F. & Rehak J. (1976) Heir. Chim. Acta 59, 2883- 2891 8 Waks M. (1986) Proteins 1, 4-15 9 Luisi P.L. & Laane C. (1986) Trends Biotech. 4, 153-160 10 Barra D Bossa F., Doonan S i~.~hmv H.M.A. Martini F. & Hughes G.J. (1976) Eur. J. Biochem. 64, 519-526 11 Wong M., Thomas J.K. & Gratzel M. (1976) J. Am. Chem. Soc. 98, 2391-2397 12 Martin C.A. & Magid L.J. (1981) J. Phys. Chem. 85, 3938-3941 13 Luisi P.L., Meier P., lmre V.E. & Pande A. (1984) in: Reverse Micelles (Luisi P.L. & Straub B.E., eds.), Plenum Press, New York, pp. 323-337 14 Smith R.E. & Luisi P,L. (1980) Helv. Chim. Acta 63, 2302- 2311 15 Steinmann B., Jackle H. & Luisi P.L. (1986) Biopolymers 25, 1133-1156 16 Metzler D.E., Harris C.M., Johnson R.J., Siano D.B. & Thomso,~ J.A. (1973) Biochemistry 12,

469

5377-5392 17 Harris C.M., Johnson R.J. & Metzler D.E. (1976) Biochim. Biophys. Acta 421, 18t - 194 18 Metzler C.M., Cahill A. & Metzler D.E. (1980) J. Am. Chem. Soc. 102, 6075-6082 19 Metzler C.M. & Metzler D.E. (1987) Anal. Biochem. 166, 313-327 20 Rej R. (1985) in: Methods of Enzymatic Analysis (Bergmeyer U.H., ed.), VCH Verlagsgesellschaft, Weinheim, 3rd edn., vol. 7, pp. 59-67 21 Morozov Y.V., Bazhulma N.P., Cherkashina L.P. & KarpeisKy M.Y. (1967) Biofizika 12, 397---406 22 Sanchez-Ruiz J.M., Llor J., Lopez-Cantarero E, & Cortiio M. (1984)i,',: Chemical and Biological Aspects of Vitamin B6 Catalysis- Part A (Evangelopoulos A.E., ed.), Alan R. Liss, New York, pp. 79-88 23 Marshall W.L. (1970)J. Phys. Chern. 74, 346-350 24 Maitra A. (1984) J. Phys. Chem, 88, 3122-3123 25 Shaltiel S. & Cortijo M. (1970) Biochem. Biophys. Res. Commun. 41,594-600