ISSN 00181439, High Energy Chemistry, 2010, Vol. 44, No. 4, pp. 304–310. © Pleiades Publishing, Ltd., 2010. Original Russian Text © A.S. Tatikolov, A.A. Ishchenko, M.A. Kudinova, I.G. Panova, 2010, published in Khimiya Vysokikh Energii, 2010, Vol. 44, No. 4, pp. 333–339.
PHOTOCHEMISTRY
Study of the Noncovalent Interaction of Squarylium Dyes with Serum Albumins A. S. Tatikolova, A. A. Ishchenkob, M. A. Kudinovab, and I. G. Panovac a
Emanuel Institute of Biochemical Physics, Russian Academy of Sciences, ul. Kosygina 4, Moscow, 119334 Russia email:
[email protected] b Institute of Organic Chemistry, National Academy of Sciences of Ukraine, ul. Murmanskaya 5, Kiev, 02094 Ukraine c Kol’tsov Institute of Developmental Biology, Russian Academy of Sciences, ul. Vavilova 26, Moscow, 119334 Russia Received October 27, 2009
Abstract—The noncovalent interaction of zwitterionic indolium squarylium dyes (hydrophilic and hydro phobic) and a structurally analogous ionic indodicarbocyanine (hydrophilic) dye with serum albumins was studied by spectral and fluorescent methods. It has been found that the hydrophilic squarylium dye with sul fonate groups most efficiently interacts with albumins, which is probably due to the double negative charge of the dye molecule at the expense of the sulfonate groups and the possibility to form hydrogen bonds with albumin. The hydrophobic squarylium dye, as well as the hydrophilic indodicarbocyanine dye without the squarylium fragment in its structure, bind with albumins much weaker than the structurally relevant hydro philic squarylium dye. The properties of the latter dye permit us to recommend it for using as a spectral and fluorescent probe for serum albumins in extracellular media of living organisms. DOI: 10.1134/S0018143910040089
A study of noncovalent interaction of polymethine (cyanine) dyes with biomacromolecules does not lose urgency and is closely related to the problem of devel opment of efficient spectral and fluorescent probes for biochemical systems. Some of these dyes have found a wide application for detection of DNA, RNA, and albumin [1–4]. It has been also shown that a number of cyanine dyes interact with human serum albumin with high association constants [5]. With respect to the search for new dyes for interaction with albumins, squarylium dyes (squaraines), the analogues of cya nines with the squarylium cycle incorporated into their polymethine chain, are of particular interest [6, 7]. As the result, squarylium dyes become zwitteri onic compounds, in contrast to classical ionic cya nines. Therefore, photophysical and photochemical properties of these classes of dyes differ essentially, in spite of the fact that they absorb and emit light in sim ilar spectral regions at the same length of the polyme thine chain [8–10]. For example, the rate of nonradi ative deactivation of the S1 state of squarylium cya nines steeply increases with growing polarity of the medium, which leads to a drop in the fluorescence quantum yield in polar solvents [8, 9], whereas for classical cyanines the main factor influencing nonra diative deactivation is the viscosity of the medium [11]. The rates of photoisomerization and back isomerization of the photoisomer of squarylium cya nines also substantially depend on the polarity of the medium [9, 12]. All these facts are the consequences of the presence of the central squaraine fragment in
molecules of the dyes, which causes intramolecular charge transfer from the terminal heterocycles to this fragment [8]. It was also shown that squarylium dyes can bind with albumin and possess rather high photo stability [7, 13]. This makes these dyes very promising for application as labels or probes for biomacromole cules and makes urgent the study of noncovalent inter actions of squarylium dyes with biomacromolecules. However, the use of squarylium dyes for these purposes is strongly restricted by poor solubility in water. To overcome this restriction, we used in this work the squarylium dye K1 (the 3Hindolium derivative), whose molecule contained sulfonate groups intro duced in order to increase its solubility in water. We studied the spectral and fluorescent properties of dye K1 and changes in these properties upon its interac tion with albumins. For comparison, we also studied its structural analogues, the squarylium dye K2 with out sulfonate groups and the ionic indodicarbocyanine dye K3. The last dye has the polymethine chain of the same length as the squarylium dyes K1 and K2, but without the squarylium fragment. O +
N R
O–
N R
where R = C3H6SO3– (K1, with the 2H+ counterions),
304
STUDY OF THE NONCOVALENT INTERACTION OF SQUARYLIUM
R = CH3 (K2);
+
N – C3H6SO3
H+
N – C3H6SO3
K3
EXPERIMENTAL Squaraines K1 and K2 were synthesized following the procedure described by Lin and Peng [14], and indodicarbocyanine K3 following Flanagan et al. [15]. Human, bovine, and rat serum albumins (SA; from Sigma) were used as albumins. Distilled water, dime thyl sulfoxide (DMSO), and isopropanol (reagent grade) were used as solvents. The absorption spectra of the dyes were measured in a 1cm cell with a Shimadzu UV3101PC spectro photometer, and fluorescence and fluorescence exci tation spectra with a Shimadzu RF5301PC spectrof luorimeter. The concentrations of the dyes in the experiments were usually within (1–5) × 10–7 mol l–1, those of albumins were from 0 to 10–4 mol l–1. Upon recording fluorescence spectra, the absorbance of dye solutions was no more than 0.1; the spectra were not corrected for the spectral characteristic of the spec trofluorimeter. The fluorescence quantum yields Φf of the dyes studied were determined using the method of comparison with a standard. Rhodamine 101 in etha nol (Φf = 0.92 [16]) was chosen as the standard. All measurements were conducted at room tem perature (20 ± 2°С). The binding constants of the monomeric dyes with albumin Keq and the binding numbers n were deter mined assuming the equilibrium (1): K
nDye + Alb eq nDye · Alb, (1) where Dye and Alb are dye and albumin molecules, respectively, n is the number of dye molecules binding with one albumin molecule (assuming that the binding sites are identical), and [ Dye ] b K eq = , (2) n [ Dye ] f [ Alb ] f where [Dye]b and [Dye]f are the concentrations of the bound with albumin and free dye, respectively, and [Alb]f is the concentration of free albumin. [Dye]b and [Dye]f were determined by simulation of the absorp tion spectra of the system dye + albumin as a sum of the spectra of the free (at [Alb] = 0) and bound (at high [Alb]) dye forms. The solution of the quadratic equation for [Dye]b, following from the equilibrium (1), gives [Dye]b = [Dye]0 – [Dye]f = (C + 1/Keq)/2 + {(C + 1/Keq)2/4 – n[Alb]0[Dye]0}1/2, where [Alb]0 and [Dye]0 are the ini tial albumin and dye concentrations, and C = n[Alb]0 + HIGH ENERGY CHEMISTRY
Vol. 44
No. 4
2010
305
[Dye]0. This solution was used for computer simula tion of the experimental dependence of [Dye]b/[Dye]f on [Alb]0 by fitting Keq and n for the best agreement between the calculated and experimental data. In some cases Keq and n were also determined from the growth of Φf of the dye in the presence of albumin, using the relationship Φf = (Φffεf[Dye]f + Φfbεb[Dye]b)/(εf[Dye]f + εb[Dye]b), where Φff, εf and Φfb, εb are the fluorescence quantum yields and absorption coefficients for the free and bound dye, respectively. The procedure of determining Keq and n was described in more detail earlier [5]. RESULTS AND DISCUSSION Figure 1 shows the absorption and fluorescence spectra of dye K1 in DMSO and water. The spectra represent narrow intense bands analogous to those of K2 in organic solvents [8, 9], and are typical for mono meric molecules of such dyes. The spectra in DMSO fl (λ abs max = 645 nm, λ max = 653 nm) are shifted batho chromically with respect to the corresponding spectra fl in water (λ abs max = 623 nm, λ max = 633 nm; see Fig. 1) due to the effect of the higher refractive index of 20 20 DMSO (nD = 1.477) than that of water (nD = 1.333) [9]. Upon introduction of increasing concentrations of albumin into an aqueous solution of dye K1, a drop in the intensity of the initial absorption band of squaraine with λmax = 623 nm and an appearance and a growth of the longwavelength band with λmax = 637–641 nm belonging to the dye bound to albumin occur (Figs. 2a, 3a; Table 1). An isosbestic point is observed in the absorption spectra, which indicates the simple equilibrium (1) between the free and bound dye. Upon binding to albumins, together with changes in the absorption spectra, a growth in the fluorescence quantum yield (Figs. 2b, 3b) and a longwavelength shift of its band occur. Squaraine K1 interacts similarly with human, bovine, and rat SA. Table 2 presents the values of Keq and n obtained from the experimental dependences of [Dye]b/[Dye]f and Φf on [Alb]0. Note that the binding numbers n obtained from the best fit of the calculation to the experimental data are less than 1. As upon the interaction of polymethine dyes with human SA studied earlier [5], this is probably explained by partial aggregation of the dye on albu mins at relatively low concentrations of albumins and its deaggregation at higher concentrations of albu mins. Aggregates of dyes usually do not fluoresce and have low absorption coefficients; therefore, aggrega tion of dyes is usually accompanied by a decrease in the absorption band of the monomeric (bound) dye and Φf at low [Alb]0.At the same time, deaggregation of the dye at higher[Alb]0returns the monomeric (bound) dye to the system, which leads to steeper cur vature of the dependence of [Dye]b/[Dye]f on [Alb]0 and, hence, to anomalously low values of n. Table 2
306
TATIKOLOV et al. Rel. units
2'
2
1'
1
500
550
600
650
700
750
λ, nm Fig. 1. (1, 2) Absorption and (1', 2') fluorescence spectra of dye K1 in (1, 1') DMSO and (2, 2') water.
presents, together with the values of Keq and n obtained from the best fit of the calculated dependence to the experimental data, the values of Keq determined approximately assuming n = 1, which apparently reflect more adequately the ability of monomeric mol ecules of the dye to bind noncovalently to albumins. Unlike squaraine K1, which practically does not form aggregates in the absence of albumins (its solu tions obey the Lambert–Beer law at concentrations up to about 10–5 mol l–1 with the molar absorption coeffi
cient ε(623 nm) = 263200 l mol–1 cm–1), its analogue K2, without hydrophilic substituents, is poorly soluble in water and forms aggregates with a broad absorption spectrum. The introduction of ionic sulfonate groups into the dye molecule (to form the structure of K1) steeply increases its solubility in water, which impedes the formation of aggregates. Aggregation in water of the squarylium dye K4, a K2 analogue with long alkyl substituents R = C18H37 on the N atoms, was studied earlier, and it was found
fl Table 1. Maximums of the absorption (λ abs max ) and fluorescence (λ max ) spectra, as well as fluorescence quantum yields (Φf) of dyes K1–K3 in aqueous solution in the lack of albumins and in the complex with human SA abs
in water K1 K2 K3
623 619b 641
Φf
fl
λ max , nm
Dye
λ max , nm with SA 641a
637, 634 658
in water 633 628b 656
with SA 647a
644, 642 668
in water
with SA
0.031 0.013b 0.10
0.47, 0.56a 0.46 0.20
a With bovine SA. b In the mixture isopropanol–water 1 : 300.
HIGH ENERGY CHEMISTRY
Vol. 44
No. 4
2010
STUDY OF THE NONCOVALENT INTERACTION OF SQUARYLIUM
307
Absorbance [Dye]b/[Dye]f (a) 15 0.08 10
0.06
5
0
0.5 1.0 1.5 [Alb]0 × 106, mol l–1
0.04
0.02
0 500
550
600 λ, nm
Φf
650
700
(b) 0.4
0.3
0.2
0.1
0
0.2
0.4
0.6 0.8 1.0 [Alb]0 × 106, mol l–1
1.2
1.4
Fig. 2. (a) Absorption spectra and (b) fluorescence quantum yields of dye K1 (3.0 × 10–7 mol l–1) in aqueous solution in the pres ence of increasing concentrations of human SA: [Alb]0 = 0, 5.7 × 10–8, 1.1 × 10–7, 1.7 × 10–7, 2.3 × 10–7, 2.8 × 10–7, 3.9 × 10–7, 5.0 × 10–7, 7.2 × 10–7, 9.4 × 10–7, 1.5 × 10–6, and 3.0 × 10–6 mol l–1. HIGH ENERGY CHEMISTRY
Vol. 44
No. 4
2010
308
TATIKOLOV et al. Absorbance [Dye]b/[Dye]f 0.04 15
(a)
5 0.03
0
0.5 1.0 [Alb]0 × 106, mol l–1
0.02
0.01
0 500
550 Φf
600 λ, nm
650
700
(b) 0.5 0.4 0.3 0.2 0.1
0
0.5
1.0 1.5 6 [Alb]0 × 10 , mol l–1
2.0
2.5
Fig. 3. (a) Absorption spectra and (b) fluorescence quantum yields of dye K1 (1.5 × 10–7 mol l–1) in aqueous solution in the pres ence of increasing concentrations of bovine SA: [Alb]0= 0, 3.3 × 10–8, 6.6 × 10–8, 9.9 × 10–8, 1.3 × 10–7, 1.6 × 10–7, 2.3 × 10–7, 2.9 × 10–7, 3.6 × 10–7, 4.8 × 10–7, 6.1 × 10–7, 7.4 × 10–7, 8.6 × 10–7, 1.2 × 10–6, 1.5 × 10–6, and 2.1 × 10–6 mol l–1. Insets in Figs. 2 and 3: dependences used to determine Keq and n from the absorption spectra of the dyes, in which the points show the experi mental data, the solid lines the results of the computer simulation for the best fit to the experimental points, and the dashed lines the results of the simulation assuming n = 1 (see text). HIGH ENERGY CHEMISTRY
Vol. 44
No. 4
2010
STUDY OF THE NONCOVALENT INTERACTION OF SQUARYLIUM
309
Table 2. Complexation constants Keq and binding numbers n for dyes K1 and K3 with albumins Complexation characteristics a K eq
K1 with human SA
K1 with bovine SA
K1 with rat SA
K3 with human SA
1.5 (0.87) 1.60b (0.85b) 1.3
5.6 (0.29) 6.2b (0.25b) ~0.6
14 (0.51) 6.2b (0.47b) ~2.6
1.0 (0.016)
× 10–7, l mol–1 (na)
Keq × 10–7, l mol–1 assuming n=1
~0.005
a The values obtained from the best fit to the absorption spectra (accuracy ± 10%). b
The values obtained from the fluorescence growth in the presence of albumin (accuracy ± 10%).
that the aggregates of the dye were not destroyed upon addition of albumin or most surfactants [9]. O +
N C18H37
O–
N C18H37
K4
In the presence of albumin, the less stable aggre gates of squaraine K2 partially decompose into mono mers bound with albumin, which leads to appearance of the absorption band with λmax = 637 nm and fluo rescence (spectra not shown). When a concentrated solution of dye K2 in isopro panol (at the isopropanol/water ratio 1 : 300) or other polar solvent is diluted with water, the dye passes into the solution as a monomer, which binds with albumin upon its introduction into the solution. The complex ation with albumin of squaraine K2, as its analogue K1, is accompanied by a longwavelength shift of the spectra and a growth of fluorescence. There is no isos bestic point observed in the absorption spectra, which is probably due to strong aggregation of the hydropho bic dye K2 on albumin. The estimation of Keq from the absorption spectra gives a value of about 3 × 105 l mol–1. As squaraine K1, the ionic dicarbocyanine K3 with sulfonate groups barely form aggregates in an aqueous solution in the absence of albumin (aqueous solutions of the dye obey the Lambert–Beer law at concentra tions above 10–5 mol l–1 with the molar absorption coefficient of ε(641 nm) = 182 700 l mol–1 cm–1). Upon introduction of SA into an aqueous solution of K3, a drop in the intensity of its initial absorption band with λmax = 641 nm and a growth of the longwave length band of the dye bound with albumin with λmax = 658 nm (with an isosbestic point in the absorp tion spectra) also occur, but at much higher SA con centrations than for the zwitterionic squaraine K1 (spectra not shown). A longwavelength shift of the fluorescence spectrum is also observed (with an isoe missive point in the spectra; the spectra not shown), but the fluorescence quantum yield increases only twice. All these facts indicate weaker interaction with SA of monomeric molecules of K3 than that of the structurally corresponding squaraine K1. The value of HIGH ENERGY CHEMISTRY
Vol. 44
No. 4
2010
Keq calculated for indodicarbocyanine K3 from changes in the absorption spectra of the dye with allow ance for n = 1 is much lower than that for squaraine K1 (see Table 2). The low apparent value of n indicates con siderable aggregation of indodicarbocyanine K3 on albumin. As can be seen from Table 2, in the case of weak interaction of a dye with albumin (K3 with SA), apparent values of n are small, thereby indicating strong aggregation of the dye on albumin. On the other hand, in the case of stronger interactions (K1 with SA), the apparent values of n increase, indicating a smaller contribution of aggregation. Hence, the pro cesses of binding of the monomeric dye with albumin and aggregation of dye molecules on albumin can be considered competitive: when the monomeric binding is weak, aggregation is manifested strongly, and vice versa. The data of Table 2 show that the hydrophilic squarylium dye K1, having sulfonate groups in its structure, quite efficiently (with the constants Keq on the order of 107 l mol–1) binds with all SA studied in this work, with a steep growth of fluorescence. The analogous squarylium dye K2, which does not contain sulfonate groups in its molecule, much weaker inter acts with albumins. On the other hand, the indodicar bocyanine dye K3, whose molecule contains sulfonate groups, but does not contain the squarylium fragment, is also much weaker complexed with albumins than K1. Hence, for efficient interaction with albumins, the presence of both the squarylium fragment and sul fonate groups in the dye molecule is necessary. As we found earlier in the study of the interaction of polyme thine dyes with human SA [5], a negative charge of the dye is first of all necessary for good complexation of a dye with albumin (the binding sites of albumin have probably a positive charge). This charge is present on dye K1 at the expense of sulfonate groups, but is lack ing on dye K2, which decreases the efficiency of the interaction of K2 with albumins. Furthermore, hydro gen bonds that the dye can form with the binding sites of albumin also play a role (this was observed for anionic oxonols [5]). The negative charge on the squarylium dye K1 is two times higher than on the indodicarbocyanine dye K3; in addition, dye K1 can presumably form hydrogen bonds with albumin at the expense of the oxygen atoms of the squarylium frag ment (such a fragment is lacking in the indodicarbocy
310
TATIKOLOV et al.
anine dye) and the sulfonate groups. All these facts are assumed to be the reason for the higher efficiency of the interaction with albumins of squaraine K1 than indocyanine K3. Since nonradiative deactivation in molecules of squaraines is mainly determined by polarity of the medium [8, 9], the steep growth of fluorescence of dye K1 upon its interaction with SA apparently indicates lower local polarity in the binding sites of SA than in aqueous solution. On the other hand, the relatively small increase of fluorescence of indodicarbocyanine K3 upon its interaction with albumin probably indi cates that binding with albumin (weak as compared to squaraine K1) only insignificantly increases the rigid ity of the indodicarbocyanine dye molecule, which possesses a flexible polymethine chain. In summary, the data obtained in the work show that the hydrophilic cyanine dye K1 having sulfonate groups in its structure efficiently binds with different SA with a steep growth of fluorescence. This permits us to recommend this dye as a spectral and fluores cence probe for SA in extracellular media of living organisms. ACKNOWLEDGMENTS This work was supported by the Russian Founda tion for Basic Research, project no. 090401054a. REFERENCES 1. Rye, H.S., Yue, S., Wemmer, D.E., Quesada, M.A., Haugland, R.P., Mathies, R.A., and Glazer, A.N., Nucleic Acid Res., 1992, vol. 20, p. 2803.
2. Timtcheva, I., Maximova, V., Deligeorgiev, T., Zaneva, D., and Ivanov, I., J. Photochem. Photobiol., A: Chem., 2000, vol. 130, p. 7. 3. Timtcheva, I., Maximova, V., Deligeorgiev, T., Gadjev, N., Drexhage, K.H., and Petkova, I., J. Photochem. Photo biol. B: Biol., 2000, vol. 58, p. 130. 4. Vasilev, A., Deligeorgiev, T., Gadjev, N., and Drexhage, K.H., Dyes and Pigments, 2005, vol. 66, p. 135. 5. Tatikolov, A.S. and Costa, S.M.B., Biophys. Chem., 2004, vol. 107, p. 33. 6. Tolmachev, A.I. and Il’chenko, A.Ya., Zh. Nauchn. Prikl. Fotogr., 2000, vol. 45, no. 6, p. 5. 7. Terpetschnig, E., Szmacinski, H., and Lakowicz, J.R., Anal. Chim. Acta., 1993, vol. 282, p. 633. 8. Gude, C. and Rettig, W., J. Phys. Chem. A., 2000, vol. 104, p. 8050. 9. Tatikolov, A.S. and Costa, S.M.B., J. Photochem. Pho tobiol., A: Chem., 2001, vol. 140, p. 147. 10. Davidenko, N.A., Ishchenko, A.A., and Kuvshinskii, N.G., Fotonika molekulyarnykh poluprovodnikovykh kompozi tov na osnove organicheskikh krasitelei, (Photonics of Molecular Semiconducting Composites Based on Organic Dyes), Kiev: Naukova Dumka, 2005. 11. Velsko, S.R. and Fleming, G.R., Chem. Phys., 1982, vol. 65, p. 59. 12. Momicchioli, F., Tatikolov, A.S., Vanossi, D., and Pon terini, G., Photochem. Photobiol. Sci., 2004, vol. 3, p. 396. 13. Welder, F., Paul, B., Nakazumi, H., Yagi, S., and Colyer, C.L., J. Chromatogr. B., 2003, vol. 793, p. 93. 14. Lin, T. and Peng, B.X., Dyes and Pigments, 1997, vol. 35, no. 4, p. 331. 15. Flanagan, J.H., Khan, S.H., Menchen, S., Soper, S.A., and Hammer, R.P., Bioconjugate Chem., 1997, vol. 8, no. 5, p. 751. 16. ArdenJacob, J., Marx, N.J., and Drexhage, K.H., J. Fluoresc., 1997, vol. 7, p. 91.
HIGH ENERGY CHEMISTRY
Vol. 44
No. 4
2010
