Fluorimetric studies of solutions of pyronin dyes: equilibrium constants in water and partition coefficients in organic-solvent—water systems

Fluorimetric studies of solutions of pyronin dyes: equilibrium constants in water and partition coefficients in organic-solvent—water systems

J. Photochem. Photobiol. A: Chem., 56 (1991) 295-311 295 Fluorimetric studies of solutions of pyronin dyes: equilibrium constants in water and part...

1MB Sizes 0 Downloads 24 Views

J. Photochem.

Photobiol. A: Chem., 56 (1991) 295-311

295

Fluorimetric studies of solutions of pyronin dyes: equilibrium constants in water and partition coefficients in organic-solvent-water systems Mohamed

El Baraka,

Michel

DeumiC? and Pierre Viallet

Laboratoire de Chimie Physique, UniversitP de Perpignan, 64025 Pe@gnan

Theodor

Ct!dex (France)

J. Lampidis

Papanicolaou Comprehensive Cancer Center, Department of Oncology, School of Medicine, Miami, FL 33136 (lJ.SA)

University of Miami,

(Received June 19, 1990)

Abstract The cytotoxic pyronin dyes provide simplified models for the xanthene squeleton of popular rhodamines. Kinetic analysis of fluorescence spectra of pyronin Y (PY) and pyronin B (PB) aqueous solutions allows us to study the reversible equilibria between R+ zwitterions = 566 nm (PY), 570 nm (PB)) and colorless ROH xanthydrols (A,,, ,,=463 nm I*; z7 nm (PB)) formed through ion-pair reactions with nucleophile OH-. The rate and equilibrium constants of R+ hydrolysis by OH- in water gives pK,+(&) = - 2.6 for the two pyronins while fluorescent species are too short lived for any shift in the equilibrium to be detected in the S, state. Covalent ROH species are stable in basic solutions and poor ion solvating solvents; the reaction of R+ with OH- is slow in neutral solutions. For further understanding and applications of pyronins on biomembranes, biphasic mixtures of an organic solvent (OS) with water (W) provide partition coefficients P and extraction percentages E of the different species in the OS phase. In the n-octanol-water (ratio 1:l) system, R’ and ROH co-exist in both phases, resulting in Pi,,= 1.8 (PY) and 2.6 (PB) at physiological pH with EROH=0.09 (PY) and 0.16 (PB); most pyronin exists as R+ in water owing to its@&+ but the presence of the octanol phase shifts the equilibrium toward ROH. The ROH structure presents other features of interest because of its oxidization by air into RO xanthone in aprotic non-polar solvents. In the OS-W system (1:l) where OS is cyclohexane, the partitioning of pyronins between the OS and W phases is strongly one sided with R + remaining in water and RO (A,,,, max=365-366 nm) completely trapped in the cyclohexane. The E percentages differ for PY and PB since ERo(PY) = 0.8 but only 0.3 for PB. The implication of hydrophobic interactions and hydrogen-bonded (solvated) water in OS is also discussed.

1. Introduction Among cationic dyes commonly used in cytochemistry [l], oxonium ions have affinity to nucleic acids [2], accumulate in live cells [3], and develop cytostatic and cytotoxic properties through exclusive mitochondria concentration [4]. Deprotonated by weak bases, for instance by water, they give colorless solvent dyes so named [S] ‘Author to whom correspondence should be addressed.

lOlO-6030/91/$3.50

Q Elsevier Sequoia/Printed in The Netherlands

296

because of their solubility in organic solvents. Xanthylium 3,6-bis(dimethylamino) chloride or pyronin Y (PY) and xanthylium 3,6_bis(diethylamino) chloride or pyronin B (PB) provide simplified models of the xanthene squeleton of rhodamines since they are free of steric complications attached to the o-paraphenyl pending groups of these popular dyes [6]. PY and PB (R’ in Fig. 1) hydrolyze in water which convert them into 9Hxanthene 9-ol-3,6_bis(dimethylamino and 9H-xanthene 9-ol-3,6_bis(diethylamino) respectively (ROH, Fig. l), The bimolecular nucleophilic scheme (1) which operates in the substitution of xanthene [7, 81 and triphenylmethane dyes [9] by OH- is used to model the reversible hydrolysis of pyronin cations R + into ROH xanthydrols. R+ +2H,O

kZ FROHi-OH,+

According

to process (l),

(1) the constants Kn + and p&+,

as given by eqns. (2) and

(3) K n+ = [ROH]/[R+][OH-]

(2)

PKR + =pKa - 14

(3)

measure the acido-basic properties of these compounds in water. Then, bioactive pyronins [lo], regarded as simplest prototypes of the equilibrium mechanisms which take place between ionized and unionized species in the study of ion-pair reactions depending on pH, may be used as simplified models for the hydrolysis and the acidobasic properties of rhodamine and xanthylium analogs which are presently of interest [l-4, 6, 111. As reported elsewhere [2, 12-151, a typical feature of the dye character of pyronins, is their ability to be used as absorption pH indicators since the color of their neutral solutions is red (AAbsS mar=547 nm (PY), 553 nm (PB)), while their alkaline solutions are colorless (broad band at hAbS,,,-,==390 nm); a reversible absorption change results from the variation in pH of dye concentrations of 10-4-10-5 M. Extinction coefficients (~(547 nm)= 1.1~10~ M-l cm-’ (PY), ~(553 nm) = 7~ lo4 M-’ cm-’ (PB)), of the neutral (red) solutions are higher than those measured (E(390 nm)= lo4 M-l cm-‘) for the corresponding alkaline (colorless) solution, but aggregation, which modifies the photophysical properties of the solute, is also a common feature for pyronins (10-4-10-5 M) in water, just as for the xanthene dyes [13,14]. Fluorimetry demonstrates advantages over absorptiometry owing to its higher sensitivity, thus allowing us to work with dye concentrations which will avoid the formation of aggregates; this property has promoted the spectrolhtorometric technique as the method of choice for studying a wide range of applications, among them recent photosensitizing approaches [4] and photochemotherapy [l].

R+ X27$3-3-32

X2NmNX2

8

9

RO

ROH

1

Cl-

Fig. 1. The molecular structures of pyronin dyes: cationic (R+), xanthones (RO): X=CH3(PY), GHS(PB).

"'"w""'

0

neutral xanthydrols (ROH)

and

297

Studied by fluorescence, the acido-basic properties of PY and PB in water should lead to the determination of the equilibrium constants KR+ in the S1 state. However, since the reaction rates involved in carbonium ion equilibria are known to be in the range 1-1O-3 s-’ for dilute systems [8], the xanthydrol formation would not occur so promptly as a proton addition or detachment; in these conditions, the fluorescence studies give directly the values of the R+ and ROH equilibrium constants in the So state. Now, even though pyronins are positively charged (R+) at physiological pH, the charge is distributed over their aromatic rings and thus does not interfere with their lipophilic nature. Further understanding of the spreading and action of pyronin dyes on biomembranes and cells needs knowledge of the lipophilic character of both R+ and ROH species. Consequently, we look first at the partition process of R+ and ROH from water to water-immiscible n-octanol since partitioning is regarded as the simplest model system of the transfer of bioactive compounds to biomembranes [16] and because n-octanol is considered as the organic solvent that, in correlation with biological activity, mimics the hydrophobic nature of the membrane well 117, 181. The cyclohexane-water system studied second allows us to measure the partitioning of the pyronins between two phases drastically different in terms of hydrogen-bonding capability. Besides, since partitioning is to serve as a model of how a given form of the solute passes from water into an organic phase, then the rate at which the equilibrium is attained might be as important as the equilibrium value itself; the time factor must be borne in mind when using the pK,+ to calculate the degree of ionization in a solution of known pH. Combined acido-basic and partitioning measures as described in this paper make it possible to define optimal conditions for use of these dyes and prepare more complex membrane cytochemistry experiments. 2. Experimental

details

2.1. Chemicals Pyronin Y was purchased from Polyscience Inc. and pyronin B from Sigma Chemical Co. Water solutions were made from decarbonated and deionized water. Cyclohexane, n-octanol and all the other solvents were spectroquality grade from Fluka Chimie AG and did not show any spurious absorption or fluorescence. 2.2. Apparatus Absorption spectra were measured on an SP8-300 Unicam spectrophotometer and fluorescence spectra on an informatized Jobin-Yvon JY3D spectrofluorometer using cells 1 cm thick; the spectra were uncorrected. The digital output of the instrument was processed through a desk console provided with homemade software. Excitation at 400 nm, unless otherwise specified, was used for fluorescence measurements together with typical excitation and emission slits of 4 nm. An ultrathermostat Haake NBe with a water bath circulator allowed a temperature control of f 0.2 “C. pH measurements were made on a Tacussel PHN 81 pH meter with a combined electrode at ztO.1 units. Fluorescence lifetimes were measured on a home-built nanosecond apparatus [19] using an Nz pulsed laser and the deconvolution procedure of modulating functions. 2.3. Procedure 2.3.1.

Calibration

Characteristic fluorescence spectra of R+ and ROH species were first recorded in water at convenient pH (Fig. 2). These spectra were run on unbuffered aqueous

298

R+

ROH

400

600

500 h @ml

Fig. 2. aerated for pH figured

Hydrolysis titration by OH- at 22 “C under A,= 400 nm of PB, 1.6 x 10V6 M, in an aqueous solution. Fluorescence intensities of R+ and ROH at the equilibrium are shown values of 12.2, 11.1, 10.7, 6.2 and 2.0 (curves l-5 respectively) and pK,+ determined as in the inset.

solutions whose pH in the range 7-12 units was adjusted by the addition of NaOH. The number, position and relative intensities of the fluorescence maxima were determined at given pH when the reaction of pyronin with OHwas achieved at fixed (22 “C) temperature. In the experimental conditions of dye concentration (1.6X lo-’ M) used to avoid aggregation, an excitation wavelength of 400 nm was selected for the following reason. The extinction coefficients of the absorbance maxima of R+ species overstep those of ROH forms [2, 71 so that hExc= 400 nm adequately favors the ROH forms (absorption maxima near 390 nm [7, 121) to allow the simultaneous recording of R+ and ROH fluorescences. Two types of calibration curves correlating R’ and ROH fluorescence intensities vs. concentrations were obtained for neutral and alkaline pH but fixed conditions of slits and sensitivity. Neutral solutions were used to calibrate the red (neutral) R+ species of pyronins while alkaline solutions (pH> 12) were used to calibrate the blue-shifted ROH species. 2.3.2. Fluorescence measurements at the equilibrium From scheme (l), it was obvious that when starting from R+ in a neutral solution, the forward reaction (2) occurs as soon as an initial amount of OHis added. A decrease in pH was also observed according to the reaction advancement, and consequently the consumption of OH- during the reaction had to be measured experimentally (see below). ‘Besides, on the addition of OH-, the time to reach equilibrium was increased as the arbitrary value of the initial pH was lowered toward physiological

299

values. Then, other difficulties were encountered since the accuracy of the pH measurements in unbuffered aerated water solutions at equilibrium were strongly dependent on dissolved COz; great care was thus necessary to eliminate dissolved CO2 by flushing the sample (15 min with N2) prior to use and before pH determination. Precautions were also taken to store the sample in the dark at constant temperature throughout the reaction. Nevertheless, for the above reasons, and because the duration of the reaction was increased drastically as the basicity of the solution was lowered, the following kinetic approach was developed. 2.3.3. Fluorescence kinetic measurements The reaction of R’ with OH- was studied at temperatures of 22 and 27 “C as a function of time. Starting from R+, the reaction proceeded on the addition of OHwith the formation of ROH and consumption of OH-. We were able to evaluate the respective concentrations in R’ and ROH species at any time f, from the reaction scheme (1) and calibration curves and so the exact consumption of OH- by the reaction at time t could be calculated and exact pH values deduced. Using this procedure, experimental pH measurements on degassed samples were unnecessary and, consequently, all fluorescence data were taken in air-saturated samples. From general principles of kinetics [20], the rate law of the equilibrium scheme (1) was written as: d[ROH]ldt=k,[R+][OH-]

-kJROH]

(4)

where k2 and k, were respectively the rate constants of the forward and backward reactions. Depending on the initial pH values and the solute concentration, two cases had to be considered. (1) When the initial concentration of the nucleophile OH- (initial pH, arbitrarily chosen) was in excess by a factor of 100 over that of the dissolved dye, the [OH-] concentration was virtually constant, [OH-] = [OH-] 0, and could be absorbed into the rate coefficient to give a new coefficient k; for the forward reaction (2). k;=kJOH-lo

(5)

Following

Otswald’s isolation method 1211, the rate law became

d[ROH]ldt=k;[R+]-k,[ROH]

simplified (6)

where ki was the pseudo first-order rate constant of the forward reaction (2); for the initial solute concentration [R’ Jo = 1.6 X 10B6 M, that pseudo-first order formalism operated for pH,>9.5. Then the integrated form of eqn. (6) was derived: h&Y&G-x))

= (k;[R*]&~)t

(7)

with kl =k;([R+],-X,)/X,

(8)

where X and X, represented the concentrations [ROH] obtained at an arbitrary time t and at equilibrium respectively. The averaged rate constant k; was then calculated from eqn. (7) through a least-squares fit procedure and rate constants k2 and k, were deduced from eqn. (5) and eqn. (8) respectively. (2) When the nucleophile concentration was no longer greatly in excess compared with the 1.6 x lo-” M solute concentration, i.e. when pH, <9.5, the OH- consumption depended on the reaction advancement. Under these conditions, the complete eqn. (4) was operating, and the first- and second-order (kl and k2) rate constants were

300 calculated using the general then gave: ln(X,(N-X)/N(X,

-X))

formalism

= (N-X&t

of scheme

(1); the integrated

form of eqn. (4) (9)

with

[R+],[OH-lo/X,

=N

w-9

and

WR+l, -XNOH-lo--X,)

=b

(11)

As the reaction rate dropped when the pH decreased, X, was difficult to measure directly; instead, the following procedure was followed: (i) data from the basic (pH 12) solution gave k2 from eqns. (5) and (7), (ii) interpolation of experimental data (recorded at pH <9.5) using eqn. (9) allowed X, to be determined while the kinetic data enabled the estimated value of k2 to be verified. In order to verify that these values corresponded to the properties of pyronins in their fundamental state, fluorescence lifetimes were obtained by spectra deconvolution which uses software following recent developments of simple and complex multicomponent decay [19]. Fluorescence data were fitted to a single exponential function through a non-linear least-square iterative deconvolution procedure, the goodness of fit resulting from the magnitude of the reduced 2 value and inspection of weighted residual plots. 2.3.4. The partition experiments Two biphasic systems composed of water (W) and an organic solvent (OS) in the ratio 1:l were used successively, OS being first n-octanol then cyclohexane. Referring to the distribution of a single species between the OS and W phases, the partition coefficient P measures the ratio of the amount of solute in the OS phase to the amount of solute in the W phase, i.e.

P neutral =

~ROWos/[ROWw

(13)

The partition degree was also defined in terms of the extraction constant E of a given species by the OS phase (ratio of the molar@ of the species in the OS phase to the total molar&y of all species present). The biphasic medium was rapidly prepared by mixing, shaking manually by simple repeated inversions and separating the two phases in small tubes of 25 ml capacity followed when necessary by a centrifugation step; 20 inversions of the mixing tube typically established equilibrium in 2-3 min. Comparison with longer periods of shaking (100 inversions or more) with a homemade apparatus demonstrated that lengthy and sophisticated procedures were unnecessary. Foilowing quick mixing and subsequent phase separation, each liquid phase was analyzed by fluorescence, the species found in the OS and W phases were characterized and their concentrations were calculated according to their respective fluorescence intensities and calibration curves.

3. Results The experimental data showed clearly that the nucleophilic substitution of pyronins was proportional to the OHconcentration. In basic solutions, the R’ decomposition

301

was relatively fast but neutral solutions of pyronins only underwent slow reactions. When performed at equilibrium in physiological pH conditions and low solute concentration, the determination of KR+ was a time-consuming and difficult operation (care must be taken to avoid the COP effect on pH); the kinetic approach, which reduces the duration of the assay and allows aerated samples to be used, therefore has advantages. 3.1. Equilibrium in warer As a characteristic behavior of pyronins in water, xanthylium cations R+ were transformed by OH- into colorless ROH xanthydrols with blue-shifted fluorescence (see Table 1). No inner filter was observed for dye concentrations less than 2X 10m6 M. The invariance of the fluorescence maxima wavelengths on greater dilution (lo-’ M) confirmed the absence of aggregates. In neutral solutions, Stoke’s shifts of about 18 nm were observed for the R+ species while alkaline solutions led to larger shifts (73-77 nm) for ROH. As an example of what was obtained for PB, Fig. 2 presents the fluorescence changes recorded at the equilibrium for different pH. The presence of an isoemissive point (540 nm) clearly indicates that only two molecular species with distinct fluorescent spectra were involved in the equilibrium. Calibration curves of both species were obtained independently, but since the R+ and ROH fluorescences did not overlap each other, when analyzing the solution we determined the concentration of only one component. For Kn+ determination, the concentration of the other compound was obtained as the difference between the total solute concentration and the experimental concentration of the first component. Because the excitation wavelength at 400 nm did not correspond to the isobestic point [7], only relative variations with pH of the fluorescence intensities of the two forms may be obtained; the resulting curves are given in the inset of Fig. 2. Similar results were found for PY. Lifetimes of 2-3 ns were obtained for both pyronins in water. With such a short lifetime, the R+ /ROH equilibrium cannot be reached in the Si state so that experimentally pK was really pK(S,) for each compound. 3.2. Fluorometic kinetics in wafer In this procedure, developed as an alternative to equilibrium measurements, the progressive changes in fluorescence of an aerated aqueous solution (disappearance of R+ and emergence of blue-shifted ROH) were followed as a function of time. Figure 3 shows the variations observed for a basic (pHO 12) solution of PB and the existence of an isoemissive point. In this pH range, the pseudo-first-order formalism may be TABLE

1

Wavelengths of the fluorescence water and n-octanol

maxima (nm) of the R+ and the ROH species of pyronins in

RC

PY PB

ROH

Water

Octanol

Water

Octanol

566 570

571 572.5

463 467

436 436

302

400

500

600

700

h (nm) Fig. 3. Fluorescence spectra of PB, 1.6X 10e6 M, in water, recorded 1, 2, 3, 4, 5, 6, 7, 8 and 9 h, at 22 “C under h&=400 nm. TABLE

as a function of time, 0,

2

Pseudo-first-order

rate constant

k; of pyronin B (1.6X 10m6 M) hydrolysis at 22 “C and pH 11.2

l (min)

55

88

173

240

295

360

X”;lO-’ F M) k; (IO-” min-‘)

73 2.00 2.63

2.80 102 2.52

4.26 155 2.43

4.97 181 2.39

5.42 197 2.43

211 5.80 2.50

480a 234 -

I& = 2.46 x lo-" min-‘, r = 0.9998. “Time required to attain equilibrium.

used (see eqns. (5-S)).Table 2 gives the kinetic data recorded at T=22 "C for PB ([R+lo= 1.66 x 1O-6 M) m . water at pHe = 11.2. Relative intensities (Fdb7) of the ROH fluorescence maximum at 467 nm were given as a function of time together with the ROH concentrations X calculated from F 467 and the ROH calibration curve. Previously evaluated [ROH], =X, = 6.43 X 10m4 M values allowed us to calculate the pseudo-firstorder

rate

constant

k; obtained

from

eqn. (7);

an average

value i&= 2.46 X 10e3

min-r

resulted from a least-square fit (r=0.9998). As has been said before, such a simplified formalism was no longer valid when the concentrations of both reactants were comparable, i.e. when p&< 9.5 and [dye],,= 1.66x lo-” M, owing to a decrease in pH when the reaction proceeds. Firstorder kl and second-order k2 rate constants were then calculated depending on pH via scheme (1) through eqns. (9-11). Equation (7) or eqn. (9) indicate that linear relations exist between concentrations and time. Depending on the initial pH,, leastsquare fits of comparable quality were then obtained with each equation, giving correlation coefficients better than 0.99. The equilibrium constants I&+ of the two pyronins are given in Table 3 together with the rates of the forward k2 and reverse

303 TABLE Rate

3

(kl, k2) and equilibrium

22

PY

37 22 37

PB PB

0.0

!

150

constants

kl (min - ‘)

:o PY

@I&+)

300

of pyronin hydrolysis at 22 and 37 “C kz (min-’

2.78 x 1O-3

1.13

3.39 x 10-3 3.89 x 1O-3 4.30 x 10-3

1.40 1.55

1.73

M-’ )

PK,+

h/e

(min)

L - 2.61

- 2.62 - 2.60 - 2.70

118 97 69 62

450

Minutes Fig. 4. R+ and ROH molar fractions measured as a function of time (min) for PB, 1.6~ 10T6 M, in water at 22 “C under hExe=400 nm. Two experimental cases are represented depending on A pH variations; (i) I, A pH =0 when pH, Z- 9.5; (ii) q, A pH #O when pH, C9.5.

kI reactions at 22 and 37 “C. As a rough indicator of the velocity of the reaction, the time tl,e required for the hydrolysis degree to reach an eth of its equilibrium value is also given for each pyronin. As shown in Fig. 4, the reaction was slightly faster for PY than for PB since their tll, differ although their pKR+ were equal. 3.3. Biphasic partitioning Experiments were performed in three steps by (i) identifying the fluorescent species in the OS and W phases, (ii) verifying the stability of the biphasic system by checking the invariance of a characteristic fluorescence with time, and (iii) determining the respective concentration of each species in each phase.

304

3.3.1. The n-ocfanol-water system Identification of the fluorescent species observed in the OS and W phases after partitioning was made on the basis of the characteristic fluorescence spectra of the RC and ROH species previously found in water. Only small shifts were observed (see Table 1) between the respective fluorescence maxima of the two pyronins in each phase. PI3 partition between octanol and water is qualitatively illustrated by the fluorescence spectra given in Fig. 5. Recorded under the same conditions of slits and sensitivity, were the fluorescence (Rf and ROH) of the initial water solution at pH 7.3 (curve l), of the aqueous phase after mixing and partitioning (curve 2) and of the octanol phase (curve 3). By using a set of octanol-water (1:l) solutions at different pH but tied dye concentration, *the pH effect on Pi, and Pnsutralwas studied at 20 “C for both pyronins. The results of the respective concentrations of R+ and ROH in the OS and W phases are summarized in Table 4. These results suggest the very simple model for the R+ /ROH equilibrium between the OS and W phases presented in Fig. 6 where (R+),, (ROH)w, (R+),, and (ROH)os represent respectively the cationic and xanthydrol forms in the W and OS phases. From this model, the following equation was easily derived log(Pj&P,,,t,,l)

= PH - (log Pas

+pKR+

(14)

+P&)

where Pas was the concentration ratio of the ROH and R+ species in the octanol phase and pK, the ionic product of water. In order to verify the model, the values of log(PiOn/P~~Utral)calculated from the right-hand part of eqn. (14) were compared for given pH with the corresponding experimental values obtained from the data in Table 4. As an example of what was obtained for PY in the range of pH (U-11.7), Fig. 7 shows the agreement between calculated values and experimental data. In addition, the Pi,, (eqn. 12) data given in

n-octanol-

5000

water

(1 :t)

x C

s

E -

R+

4000

E S 0 3000 s z zl E: 2 2000 .9 z

I’;\‘

In water,ph=7,2

ROH

1000

500

600

700

h (nm)

Fig. 5. Fluorescerice spectra of PB, 1.6 X lo-” M, in water, pH 7.3 (curve l), and of the water phase (curve 2) and of the octanol phase (curve 3) after mixing n-octanol with the water (1:l). All spectra were recorded at 22 “C under A-=400 nm.

5.00

6.28

9.33

12.00

15.00

13.47

2.47

1.80

1.22

Pneuml

27.28 -

pion

0.08

0.26

2.14

0.09

-

0.11

0.13

-

4.60 3.63

5.27 -

6.43

7.60

8.37

pionIpncutritI EROH[ROM P+lw

0.82 0.78

0.84

0.87

-

0.89 0.88

E’RO

0.81 0.77

0.84

0.86

0.88

0.89

ERO

2.66

6.35

8.73

15.28

Pion

8.25

7.14

-

10.52

-

3.33

Pnewtt

Octanol

Octanol

Cyclohexane

Pyronin B

Pyronin Y

See eqns. (12). (13), (16) and (17) for definitions of the constants.

11.8 11.6 11.1 10.8 10.7 10.1 9.8 9.3 8.8 8.4 7.6 7.3 7.2 6.2 6.0

FH

0.25

0.79

1.23

4.59

P;onIPt,,w~

0.16

0.18 -

0.21 -

0.24 -

EROH

0.42 0.35

0.85

0.86

0.29 0.26

0.46

0.46

0.54 -

1.17

-

FRO

-

-

P+lw

[ROlos/

Cyclohexane

0.29 0.26

0.45 -

0.46

0.53

-

ERO

Experimental values of partition P and extraction E coefficients respectively obtained at 22 “C for the R+, RON. and RO species of pyronins in the biphasic (1:l) systems n-octanol-water and cyclohexane-water

TABLE 4

306

KR’

( R+hv +

.-

@OH),

WATER

/ --

OH-

PHASE

1

________--___

PllXl

pneulral --

I

OS PHASE

i

( R+)os

t

(ROH)os

Pos Fig. 6. The n-octanol-water (1:l) biphasic model.

-1

Log (PVPn)

exp

1

Fig. 7. Correlation observed for PY in the n-octanol-water biphasic system between experimental values of log(Pi,,/P,,,t,,l ) and calculated values deduced from eqn. (14). The pH values studied were 8.35, 9.4, 10.1 and 11.65. Table 4 showed that there was more R+ in the octanol phase than in water, whatever the pH. When the aqueous phase was neutral, more than half of the dye concentration was found as R’ in octanol and this percentage increased with pH (up to 82% for pH 11.7). As the Pntutra,values (eqn. 2) demonstrated (see Table 4), ROH was also found preferentially in the octanol phase but quantitative differences appear between the two pyronins concerning the partition coefficients. In the range of pH (7X&11.8), ERoH was higher (16-24) for PB than for PY (9-15). 3.3.2. The cyclohexane-water system The fluorescent species partitioned at equilibrium between cyclohexane and water were not the same as those found in the octanol-water system. Figure 8 shows the fluorescence spectra obtained in the cyclohexane and water phases after mixing an aqueous solution at pH 12.2 (curve 1) of PB with cyclohexane. A new blue-shifted fluorescence (curve 3) was observed below 400 nm for A%= 350 nm since this new species did not absorb in the visible. Only residual ROH fluorescence (maximum at 425 nm under hExc= 400 nm) was observed in the water phase after mixing. Similar results were obtained under the same conditions for PY. These blue-shifted fluorescences (AExC=350 nm) were structured, showing peaks and shoulders (see Table 5 and the inset of Fig. 8 for PB). They were assigned [22] to the fluorescence of the ketone forms RO (see Fig. 1) of the xanthene rings, i.e. 9-H-xanthen-g-one 3,6 bis(dimethylamino) and 9-H-xanthen-9-one 3,6-bis(diethylamino)

307 5000 Cyclohexane-water

ROH r .z u) c a, zE c 8 z z3 E

(1 :l)

4000 in water,pH=l2,2 before mixing

3000

2000

.-F +Ei 1000 a, [r C1 500

400

700

600 Wm)

Fig. 8. Fluorescence spectra of PB, 1.6 X lop6 M, in water, pH 12.2 (curve 1) and in the water (curve 2) and cyclohexane (curve 3) phases after mixing cyclohexane with water (1:l). Compared spectra are recorded at 22 “C, A,= 400 nm (curves 1, 2) and 350 nm (curve 3) with different conditions of slits and sensitivity. Absorption and fluorescence spectra of the RO xanthone extracted by cyclohexane are shown in the inset.

TABLE

5

Peaks and shoulders observed in the first (Z&,-S,) absorption band and in the fluorescence spectra of PB in cyclohexane Absorption nm 356 (348) 340

O-O cm-’

Fluorescence

cm-r 28089 (28735) 29412

27668 27630 27659

cm-’

nm

27248 (26525) 2.5907 (24630)

(377) 386

367

(406)

chlorides respectively, since xanthydrols were known to be oxidized by air to xanthones [23]. These results suggest the simplified model given in Fig. 9 for the equilibrium between the R+ and RO species of pyronins, ROH being experimentally only a transient form. From this model, the following equation was easily drawn log([RO]os/IR+]w)

= pH

-

(log

Q-l

+P&+

+P&)

WI

but cannot be used directly because of the weakness of the residual ROH fluorescence in water. Instead, the accuracy of the model was established by verifying from the experimental data of Table 4 the assumption [ROmw =O proposed in the model. For that purpose, the extraction coefficients E RO defined in eqn. (16) and obtained experimentally

308

lz

Q

K R+

R+ +

OH-

ROH

_

WATER

0.8

e

RO ;

PHASE

OS PHASE

0.0 5

I

Fig. 9. The cyclohexane-water (1:l) biphasic model.

7

9

11

13

PH

Fig. 10. Variation in the extraction coefficients Ex measured for PY and PB depending on pH and the organic solvent (OS) phase: I, OS is n-octanol with X=ROH; II, OS is cycIohexane with X=RO.

ERO = [ROlosNRO],, = [R0],,/(1.66x were compared [ROH]w = 0. ER, =

+ [R+]w + [ROH],) 1O-6 M)

with the &o

[~Olos~POlos + [R +lw

(16) values (eqn. (17)) calculated

using the approximation (17)

The agreement between E&o and Exe, which was better than 98% (see Table 4) in the range of pH studied, confirmed the validity of the model. Quantitative partition between the OS and W phases was strongly one sided with all R+ remaining in the water and the RO completely trapped in the hydrocarbon phase, resulting in negligible exchange of R+ and RO between the phases. As shown in Fig. 10, the partition ratio was different for the two pyronins since, for neutral pH in water, ERO =0.8 for PY, compared with 0.3 for PB.

4. Discussion At neutral pH in water the R’ fluorescence of the two pyronins differs slightly (Table 1); similarly, the ROH emissions of PY and PB in alkaline water are comparable. A Stoke’s shift of 103 nm is observed for both pyronins, resulting in energy shifts as large as 3930 cm-’ (PY) and 3869 cm- ’ (PB) on the addition of OH-. These similar results suggest that with time in water, the same reversible conversions of xanthylium ions into xanthydrols are operating for both pyronins as a form of tautomerism between cationic dyes and neutral carbinols [24]. The experimental values of pK obtained by fluorescence are measures of the R’ /ROH equilibria in the So state, since the fluorescence

309

lifetimes of pyronins are a few nanoseconds which is much less than the rate of formation of carbinols (hours). Then it follows that identical pKa+(!&)= - 2.6 are measured by fluorescence for the two pyronins. These values are in agreement with the equilibrium constants of comparable nucleophilic substitutions previously observed for phenol red dimethyl ester [9]; they confirm Fujiki et CL’S study of PY [7] and other results [8], among them the simple Huckel MO calculation of the So state of pyronins [2S] which proposes the C, carbon (charge defect of 0.3 units) as the site of the nucleophilic attack_ For further understanding and application of pyronins on biomembranes, the biphasic partitioning of the R+ and ROH species was first studied in a system of two solvent phases of high hydrogen-bonding capabilities. For that purpose, the 12-octanol-water (1:l) mixture provided us measures of the partition and extraction coefficients of the pyronin species by an alcoholic phase. In this system, the R+ and ROH forms coexist in both phases resulting in Pion= 1.8 (PY), 0.16 (PB); owing to their common pK,+(S,) value, most of these two pyronins exist as R+ in water but their passage from water to octanol implies the R+ hydrolysis into ROH. Since the hydrogenbonding capability, more than the polarity, constitutes the basis for discussing pyronin-octanol interactions [12], the protic character of n-octanol (both donor and acceptor in hydrogen-bonding) may favor at the same time the interactions of both species with the solvent. However, octanol contains 2.3 M of water at saturation [26], i.e. (assuming no volume change in mixing) the ratios of molecules of pyronin/&O/ octanol are equal to l/1440 x 103/3970X 103. As pointed out by Israelachvili et al. [27], R+ cations and ROH xanthydrols may then enter the octanol phase surrounded by a shell of water molecules forming ionophores. The similar fluorescences observed (see Table 6) for each form of pyronin in different air-saturated (hydrated) alcohols and some polar solvents support this view of an ionophore-mediated entry into the OS phase. In the cyclohexane-water (1:l) system studied, the two phases differ drastically in terms of hydrogen-bonding capabilities since the OS phase is now aprotic nonpolar. However, the water content of cyclohexane at saturation of 2.5 X 10d3 M still leads to ratios of molecules of pyronin/&O/cyclohexane equal to l/1.5 X 103/5780X ld. TABLE

6

Wavelengths (nm) of the fluorescence maxima of pyronins, 1.G1O-6 and hydrated cyclohexane at 22 “C Pyronin species

RO Y

Methanol n-Propanol n-Butanol n-Pentanol n-Octanol Acetonitrile Formamid DMSO Water Cyclohexane

(1:l)

375

R+

ROH B

377

M in air-saturated solvents

Y

B

Y

B

443.5 429.5 431 453 436

451

567

569 571 572

436

441.5 448 463

450

571 570.5 573 582.5 566

575 573 575 575.5 570

433 434

467

567.5 569.5 572

310

Then, the passage of R+ and ROH ionophores from the W to the OS phase might be possible since it would not imply complete “dehydration” of the ionophores. This argument is not valid. In fact, cyclohexane does not solubilize R’ while ROH is oxidized by air to hydrophobic RO xanthone [22, 231; after partitioning from water, the cyclohexane phase only contains the RO species of pyronins whose structured fluorescences are the mirrors of their corresponding absorption spectra (Table 5). To “see” the fine structure of the solvent, as expected for the behavior of onium ions, the R’ charge is probably somewhat delocalized in pyronins [25]; R+ pairing with OH-, hydrogen-bonding with solvated water and oxidization by air are then the successive means for reducing the solvation energy in cyclohexane. Covalent xanthones of both pyronins are stable in cyclohexane but the above processes are more effective for PY than for PB since PY is almost completely trapped (En0 = 0.81) in the hydrocarbon phase of the OS-W mixture at pH 7.2 in opposition to PB (Eao=0.29).

References 1 S. C. 2 J. Z. 3 F. 4 Z. Z. 5 6

7 8

9 10 11 12 13

14

Rangonathan, F. Churchill and R. D. Hood, Toxicoi. Appl. PharmacoC., 99 (1989) 81. R. Shea, N. Chen, J. Winberly and T. Hassan, Cancer Rex, 49 (1989) 3961. Kapuscinski and Z. Datynkiewicz, Cytometry, 8 (1987) 12. Darynkiewicz, J. Kapuscinski, F. Traganos and H. R. Crissman, Cytometry, 8 (1987) 138. Traganos, M. A. Crissman and 2. Darynkiewicz, Ejcp. Cell Res., I79 (1988) 535. Darynkiewicz and S. P. Carter, Cancer Res., 48 (1988) 1295. Darynkiewicz, J. Kapuscinski, S. P. Carter, F. A. Schmid and M. R. Melamed, Cancer Res., 46 (1986) 5760. E. N. Abrahardt, Dyes and their Intermediates, Pergamon, London, 1977. L. B. Chen, Ann. Rev. Cell Biol., 4 (1988) 155. B. Ehrenberg, V. Montana, M. D. Wei, J. P. Wuskell and L. M. Loew, Biophys. J., 53 (1988) 785. K. Fujiki, C. Jwanaga and M. Koizumi, Bull. Chem. Sot. Jpn., 35 (1962) 185. C. D. Ritchie, C. Kubisty and G. Y. Ting, J. Am. Chem. Sot., 10.5 (1983) 279. P. A. H. Wyatt and Z. M. Zachowski, 2. Phys. Chem., 101 (1976) 143. C. D. Ritchie and T. C. Hofelitch, J. Am. Chem. Sot., 102 (1980) 7039. T. C. Tomov, J. Biochem. Biophys. Methods, 13 (1986) 29. R. R. Cowden and S. K. Curtis, Histochem., 77 (1983) 535. R. D. Hood, C. L. Jones and S. Ranganathan, Terurology, 40 (1989) 143. N. Nizamov, K. U. Umarov and A. K. Atalchodaev, Z. PrikI. Spektrosk, 30 (1978) 651. K. U. Umarov, Dokl. Akad. Nauk Lib. SSR, 51 (1982) 33. 0. Valdes-Aguilera and D. C. Neckers, Act. Chem. Res., 22 (1989) 171. I. Lopez-Arbeloa and P. Ruiz-Ojeda, Chern. Phys. Lett., 79 (1981) 347. K. H. Drexhage, Dye Lasers, Springer, Berlin 1973, p. 167. L. P. Gianneschi, A. Cantand and T. Karucsev, J. Chem. Sac. Furaduy Truns. 2, 73 (1977) 664.

15 P. Jacobsen, H. Lyon and S. Treppendahl, Histochem., 81 (1984) 99. C. R. Ritchie, .I. Am. Chem. Sot., 108 (1983) 7313. 16 Y. Yoshikawa and H. Terada, Chem. Pharm. Bull., 36 (1988) 2759. 17 C. Hansch and W. J. Dunn, 3. Phurm. Sci., 61 (1972) 1. H. Machleidt, S. Roth and P. Seeman, Biochim. Biophys. Actu, 255 (1972) 178. C. Hansch and T. Fujita, J. Am. Chem. Sot., 86 (1964) 1616. 18 W. D. Stein, Membrane Transport, Vol. 2, Elsevier, Amsterdam, 1981. 19 J. Vigo, These Doctorat, Universit% de Perpignan, 1985. 20 G. Pannetier and P. Souchay, Chemicul Kinetics, Elsevier, Amsterdam, 1967. 21 P. W. Atkins, Physical Chemistry, Oxford University Press, Oxford, 2nd edn., 1982, p. 931.

311 22

23 24 25 26 27

Y. Sakaguchi, H. Hayashi, H. Murai and Y. J. I’Haya, J. Am. Chem. Sot., 110 (1988) 7479. R. E. Connors, R. J. Sweeney and F. Lerio, J. Phys. Chem., 91 (1987) 819. K. A. Abduilah and T. J. Kemp, J. Photochem., 32 (1986) 49. S. Wawonek, Heterocyciic Compounds, Vol. 2, Wiley, New York, 1956. K. C. Ingold, Structure and Mechanism in 0qganic Chemistry, Cornell University Press, Ithaca, NY, 1953, p. 575. A. T. Pilipenko, L. L. Savranskii and N. M. Shin, Anal. Khim., 24 (1969) 460. T. Shen, &es Pigments, 8 (1987) 375. A. Leo, C. Hansch and D. Elkins, Chem. Rev., 71 (1971) 525. C. Hansch, Act. Chem. Res., 2 (1969) 232. J. N. Israelachvili, S. Marcelja and R. G. Horn, Q. Rev. Biophys., I3 (1980) 121.