- Email: [email protected]
Dimerization of tetracationic porphyrins: ionic strength dependence
Journal of Inorganic Biochemistry 69 (1998) 25±32
Dimerization of tetracationic porphyrins: ionic strength dependence Dabney W. Dixon *, Vera Steullet Department of Chemistry, Georgia State University, Atlanta, GA 30303, USA Received 29 May 1997; received in revised form 2 September 1997; accepted 2 September 1997
Abstract Cationic porphyrins are under study in a number of contexts including their interaction with biological targets, as possible therapeutic agents and as building blocks for molecular devices such as molecular photodiodes and solar cells. Many cationic porphyrins dimerize readily in aqueous solution. Dimerization in turn can control the properties of the porphyrin as well as its binding to its target. The propensity of a porphyrin to dimerize in aqueous solution can be estimated by recording the optical spectrum of the solution as a function of the concentration of added salt. Analysis of the data in terms of the Debye±H uckel formalism gives an estimate of the extent of dimerization as a function of ionic strength. Data for TMPyP4 [meso-tetrakis(4-N-methylpyridinium)porphyrin] and its butyl and octyl homologs; TMAP [meso-tetrakis(4-N,N,N-trimethylanilinium)porphyrin]; THOPP [meso-tetrakis[4-N-[(3-(trimethyl-ammonio)propyl)oxy]phenyl]porphyrin] and the ferrocenyl porphyrin P3Fc are discussed. Dimerization may aect binding of the cationic porphyrins to their targets, e.g., DNA. Ó 1998 Elsevier Science Inc. All rights reserved. Keywords: Porphyrin; Dimerization; Ionic strength; TMPyP4; TMAP
1. Introduction Many cationic porphyrins dimerize quite readily in aqueous solution. This dimerization can control the properties of the porphyrin itself as well as its binding to its target. The goal of the work described herein was to provide a rapid estimation of the propensity of a given cationic porphyrin to dimerize in aqueous solution. These estimates can then be used in designing and interpreting experiments in which cationic porphyrins bind to biological and physical targets. One area in which porphyrin aggregation is important is in the binding of cationic porphyrins to DNA [1±4]. Porphyrins with a strong propensity to self-stack in solution, e.g., TMAP [5±9], cis- and trans-P4 [10± Abbreviations: TMPyP4, meso-tetrakis(4-N-methylpyridinium)porphyrin; TBPyP4, meso-tetrakis(4-N-butylpyridinium)porphyrin; TOPyP4, meso-tetrakis(4-N-octylpyridinium)porphyrin; TMPyP2, mesotetrakis(2-N-methylpyridinium)porphyrin; TMAP, meso-tetrakis[4-N,N,N-trimethylanilinium)porphyrin; THOPP, meso-tetrakis[4-N-[(3-(trimethylammonio)propyl)oxy]phenyl]porphyrin; trans-P4 trans-bis (4N-methylpyridinium)diphenylporphyrin; cis-P4, cis-bis(4-N-methylpyridinium)diphenylporphyrin; Fc, ferrocene; P3Fc, 4,40 ,400 -[20-(2-ferrocenylethoxy)-21H,23H-porphine-5,10,15-triyl]tris[1-methylpyridinium] * Corresponding author. Tel.: 404 651 3908; fax: 404 651 1414; e-mail: [email protected]. 0162-0134/98/$19.00 Ó 1998 Elsevier Science Inc. All rights reserved. PII: S 0 1 6 2 - 0 1 3 4 ( 9 7 ) 1 0 0 0 5 - 8
12], and THOPP [13±15], prefer to bind in the grooves of DNA rather than to intercalate. Some porphyrins appear to bind in the grooves as monomers. For others, binding in a conformation which involves self-stacking of the porphyrins in the DNA grooves is indicated by conservative bands and high molar ellipticities in the circular dichroism spectra [4]. Other studies of cationic porphyrin interaction with biological targets involve complexes with melanin [16,17], albumin [18,19], sugars [20] and polyglutamic acid [21]. Proposed therapeutic uses of cationic porphyrins include the inactivation of bacteria [22,23] and viruses [24± 27] as well as in photodynamic therapy of cancer [28±31]. In studies bearing directly on the relationship between porphyrin properties and biological activity, Cernay and Zimmerman have observed that the lipophilicity of tetraalkylated tetrapyridylporphyrin derivatives controls the subcellular localization of these species [32]. The photophysical properties of porphyrins are known to depend signi®cantly on their aggregation state [33]. Understanding of porphyrin dimerization or aggregation is therefore particularly important in studies utilizing cationic porphyrins as electron acceptors in intramolecular electron transfer processes [34±36]. Self-stacking of cationic porphyrins may also be important in their deposition on surfaces. For example, Kleijn and co-workers have observed that the mode
26
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
and rate of deposition on a silica surface is a function of the ionic strength of the depositing solution [37]. Schaafsma and co-workers have observed that the absorption of cis-P4 on surfaces is a function of the concentration of the porphyrin; dilute solutions (5 ´ 10ÿ6 M) give monolayer coverage and more concentrated ( P 10ÿ4 M) solutions result in increased ordering on the surface [38]. Porphyrin aggregation may also be signi®cant in the stacking of TMPyP4 as dimers with lipid monolayers [39]. Dimerization and aggregation of cationic porphyrins can substantially aect both the properties of the molecule and the binding to its target. In studies of binding to biological targets, the properties of the biological target can de®ne the required buer conditions and ionic strength. Because the extent of stacking depends substantially on the ionic strength, one either needs to make a separate estimation of the dimerization constant at each ionic strength or to have an algorithm to predict how the dimerization depends on ionic strength. In addition, data analysis demands that the extinction coecient of the monomer be known. One needs enough data to be con®dent that the extinction coecient determined at low porphyrin concentration is in fact the monomer extinction coecient and not that of a partially dimerized species. Given these criteria, rapid methods of estimating the dimerization constant as a function of ionic strength are of interest. Herein, we show that the Debye±H uckel formalism provides a method for rapid estimation of the dimerization of cationic porphyrins (Fig. 1) as a function of ionic strength.
2. Results
Fig. 1. Structures of selected cationic porphyrins.
2.1. Salt-induced stacking An example of a porphyrin which shows changes in its optical spectrum as a function of the salt concentration of the solution is the tetrakis-trimethylanilinium derivative TMAP [5,40]. A salt titration was run using a 1 mm cell with relatively high [TMAP] to promote aggregation (Fig. 2). As the salt concentration increases, the absorption band at 412 nm decreases and a shoulder grows in to the blue at 402 nm; the titration appears isosbestic. Changes, though small, continue to take place from 3.9 M (shown) to saturating concentrations of NaCl. This data set allows one to estimate the extent of dimerization at any given ionic strength. A simple dimerization may be expressed as M M D with a dimerization constant KD , 2
KD D=M :
Fig. 2. UV/vis absorption spectrum of TMAP in water with increasing additions of NaCl as indicated by the ®gure legend (1 mm cell, P0 9:5 10ÿ6 M).
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
27
Expressing all concentrations in terms of the concentration of monomer, the total concentration of the porphyrin, [P]0 , is [41]: P0 M 2D: The mole fractions of monomer (fM ) and dimer (fD ) are de®ned as fM M=P 0 ; fD 2D=P0 : The measured extinction coecient, , is fM M fD D ; where M and D are the extinction coecients of the monomer and dimer, respectively. The mole fractions of monomer and dimer can be calculated from the optical data
Fig. 3. Log(KD ) vs. l1=2 = 1 l1=2 for TMAP in water.
fM ÿ D = M ÿ D ; fD M ÿ = M ÿ D : From these, one obtains the dimerization constant, KD , at each ionic strength KD fD =2P0 fM fM : Given fD and the total porphyrin concentration, [P]0 , an equilibrium constant can be calculated at each ionic strength. The calculation assumes that the extinction coecients of the monomer and dimer are independent of the ionic strength, and that one can obtain both the fully monomeric (no added salt) and fully dimeric (high ionic strength) forms. This may be an issue either because the system stacks beyond the dimer or because the extinction coecient of the dimer is a function of ionic strength. Both of these issues are discussed in more detail below. For TMAP, the extinction coecient of the dimer was taken to be 0.26 that of the monomer at the kmax of the monomer (from the spectrum in saturated NaCl). The calculated equilibrium constants calculated from the data in Fig. 2 are 2.6 ´ 104 , 7.4 ´ 104 , and 1.5 ´ 106 at NaCl concentrations of 0.48, 0.97, and 2.9 M, respectively. To use this data in predicting equilibrium constants at other ionic strengths, one needs a model for the dependence of the equilibrium constant on ionic strength. The change in KD with ionic strength can be treated using the Debye±H uckel limiting law log KD log KD0 2AzA zB l1=2 = 1 l1=2 ; where KD0 is the limiting value of KD at zero ionic strength and zA and zB are the charges on the respective halves of the dimer (here equal) [42]. For water, A is 0.509 at room temperature [43]. Fig. 3 shows a plot of log(KD ) vs. l1=2 /(1 + l1=2 ) for TMAP. For TMAP, the data are monotonic, but not linear, over the ionic strength range studied. The absorbance of TMAP as a function of [P]0 at 0 and 1 M NaCl is given in Fig. 4. The line for 1 M NaCl curves downward with increasing [P]0 because the eective extinction coecient of the porphyrin decreases as it dimerizes with increasing [P]0 . If the extinction coe-
Fig. 4. Absorbance vs. [TMAP] with 0.0 and 1.0 M added NaCl. In the solution with no added salt (), the line through the points is a Beer's Law ®t for the data. In the solution with 1.0 M added NaCl (n), the line through the points is the predicted absorbance calculated from the equations in the text and the calculated D and KD values of 2:9 105 Mÿ1 cmÿ1 and 5:8 105 Mÿ1 , respectively.
cient of the monomer is known, data sets of this type can be ®t [41,44,45]. D ÿ f M ÿ D =2Kg1=2 f M ÿ =P0 g1=2 : A plot of vs.{( ) M )/[P]0 }1=2 gives D from the intercept; KD can be calculated from the slope (Fig. 5). For this data set (silanized cell) they are 3.1 0.1 ´ 105 Mÿ1 cmÿ1 and 3.3 0.7 ´ 106 Mÿ1 , respectively. A second data set (unsilanized cell) gave 2.8 0.1 ´ 105 Mÿ1 cmÿ1 and 1.7 0.2 ´ 106 Mÿ1 , respectively. These compare with approximate values derived from the salt titration of 1.2 ´ 105 Mÿ1 cmÿ1 and 7.4 ´ 104 Mÿ1 , respectively. Thus, the salt titration gives estimated, but not exact, values for D and KD . This is partly because the spectrum of the ``fully dimerized'' porphyrin in the salt titration is not accurate, in this instance because it is not really fully dimerized but in other instances because aggregation into larger structures has begun to take place. In addition, the structure of the dimer itself, and hence its optical signature, is expected to be a function of the ionic strength. It is interesting to note that the equilibrium constants for TMAP here are similar to KD for dimerization of [tetrakis(p-(trimethyl-
28
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
Fig. 5. A plot of vs.f M ÿ =P0 g1=2 for the TMAP data in Fig. 4.
ammonio)phenyl)porphinato]silver(III), Ag(III)TMAP, 1.2 ´ 105 Mÿ1 (0.05 M NaNO3 , pH 7.0) [46]. TBPyP4: The salt-induced stacking of TBPyP4 was also studied by recording the optical spectrum of the compound as a function of added NaCl. The titration was run in a 1 mm cell so that a relatively high concentration of TBPyP4 could be employed to increase stacking. Fig. 6 shows that the Soret band at 424 nm decreased in intensity with almost no change in wavelength as the salt concentration was increased. Changes continued even up to saturating amounts of NaCl. P3Fc: The ferrocenyl porphyrin P3Fc has been studied in conjunction with the role of DNA scaolding in controlling electron transfer [47]. Fig. 7 shows the absorption spectrum of P3Fc in N00 buer (pH 7.0) with increasing additions of NaCl (from 0 to 0.92 M NaCl added). As the concentration of NaCl increases, the intensity of the Soret band at 425 nm decreases; above 25 mM NaCl a new peak, at 447 nm, appears. A single isosbestic point is visible at 443 nm. These spectral changes are consistent with self-complexation of the porphyrin. It should be noted that a kmax of approximately 450 nm is also a signature for protonation
Fig. 6. UV/vis absorption spectrum of TBPyP4 in water with increasing additions of NaCl as indicated by the ®gure legend (1 mm cell, P0 6:1 10ÿ5 M).
Fig. 7. UV/vis absorption spectrum of P3Fc in N00 buer with increasing additions of NaCl as indicated by the ®gure legend (1 cm cell, P0 4:5 10ÿ6 M).
of the center nitrogen atoms of the porphyrin [48]. For P3Fc, attribution of this spectral change to protonation seems unlikely because the pKa values are expected to be too low. Data are available for TMPyP4; approximate values of pKa3 (protonation of the third central nitrogen) and pKa4 (protonation of the fourth central nitrogen) are 2 and 1, respectively [49,50]. The values of pKa3 and pKa4 are sensitive to buer strength, but not suciently to result in protonation of the center nitrogens even at high ionic strength. For example, the pKa3 value for TMPyP4 is 1.4 in 0.2 M NaNO3 and 2.2 in 2.0 M NaNO3 [49]. 3. Discussion In many instances, the extent of dimerization of a porphyrin is an important aspect of experimental design. The dependence of KD on ionic strength can be predicted in a useful manner from a single titration in which salt is added to an aqueous solution of the porphyrin. Estimating dimerization constants from salt-induced stacking assumes that one can obtain the limiting extinction coecients of the monomer and dimer. This will not be possible in many cases, e.g., if one cannot go to low enough concentrations to obtain a solution of only the monomer or if aggregates beyond the dimer begin to form at high ionic strength. The extinction coecients of the dimer may also be aected by the ionic strength of the solution, presumably because the geometry of the dimer is changing slightly as the ionic strength changes. However, even with these caveats, salt-induced selfstacking can be a useful indication of the propensity of a porphyrin to dimerize in solution. In this approach, the Debye±H uckel limiting law is employed at much higher ionic strengths than the assumptions of its derivation warrant. Whether or not extrapolation to zero ionic strength gives a true value of log(KD0 ), plots of log(KD ) vs. l1=2 /(1 + l1=2 ) are monotonic or linear for a number of cationic porphyrins. Fig. 8 shows data plotted from the TMAP, TBPyP4 and P3Fc experiments herein as well as from two salt ti-
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
Fig. 8. Log(KD ) vs. l1=2 = 1 l1=2 for cationic porphyrins in water (the limiting extinction coecients for the monomers and dimers are given in the experimental).
trations in the literature: those for THOPP, a tetraphenylporphyrin bearing OCH2 CH2 CH2 N Me3 side chains [13], and TOPyP4 with octyl side chains [5]. These ®ve porphyrins stack to very dierent extents in aqueous solution. The extent of curvature of the data is a function of the ionic strength range necessary to produce stacking of the porphyrin. Porphyrins which stack at low ionic strength, e.g., THOPP and P3Fc, show a relatively linear dependence of log(KD ) on l1=2 /(1 + l1=2 ). Porphyrins which stack only at high ionic strength, e.g., TMAP and TBPyP4, show substantial curvature in the plot. This is expected because the assumptions underlying the formalism break down at higher ionic strengths. The charges on the porphyrins calculated from the Debye±H uckel formalism are generally less than the formal charges on the porphyrins (4+ in each case here). Calculating the charges from the slopes (2AzA zB , zA zB ) at the lower ionic strengths, one obtains approximately 3.6, 3.6, 4.0, 2.4 and 3.0 for THOPP, P3Fc, TOPyP4, TMAP, and TBPyP4, respectively. Similar eects have been observed a number of times previously. For example, the kinetics of dimerization of Ag(III)TMAP are best ®t with a charge of approximately +2 on the porphyrin monomer, though the formal charge is +5 [46]. The ionic strength dependence of Cu2 insertion into TMPyP2 is ®t to a charge of +2.8; the formal charge on TMPyP2 is +4 [51,52]. Similarly, the ionic strength dependence of Zn2 insertion into TMPyP4 is ®t to a charge of +2.4; the formal charge is again +4 [50,51]. The ionic strength dependencies of the acid solvolyses of Zn(II)TMPyP4 and Zn(II)TMAP also are ®t by a net charge on the complex that is substantially less than the formal charge [51,53]. The dierence between the formal charges and those calculated from the ionic strength dependencies may re¯ect ion pairing of the cationic porphyrins with counterions in solution [46]. Ion pairing has also been invoked to explain aspects of the NMR spectra of cationic porphyrins [54±56]. An interesting series is provided by the methyl (TMPyP4), butyl (TBPyP4) and octyl (TOPyP4) tetraalkylated homologs of tetrapyridyl porphyrin. Kano et al., observed that the optical spectrum of the octyl derivative, TOPyP4, was a function of salt concentration [5].
29
An isosbestic red shift of the Soret by about 18 nm up to an NaCl concentration of 0.5 M was seen. At higher ionic strengths, the Soret shifted back toward the blue in a non-isosbestic manner. These observations are consistent with initial dimerization, followed by stacking into larger structures as the ionic strength is increased. The tetrabutyl derivative TBPyP4, in contrast, undergoes salt-induced stacking with only a minimal change in the wavelength of the Soret (Fig. 6). Changes occur gradually as the salt concentration is raised but are not complete even with saturating amounts of NaCl. The tetramethyl derivative, TMPyP4, shows no signi®cant optical changes in water even at 2±3 M NaCl [5,57]. This lack of spectral change with increasing salt concentration indicates that TMPyP4 does not aggregate readily. The apparent lack of stacking even at these high ionic strengths is consistent with TMPyP4 being monomeric at optical concentrations in water. Other groups have also concluded that TMPyP4 is monomeric at optical concentrations [1,11,14,16,56,58,59], although early ¯uorescence lifetime-based studies were interpreted as evidence for dimers [60±63]. The dimerization estimates in Fig. 8 are useful both in interpreting data under the reaction conditions and in measuring the total concentration of the porphyrin. The ®rst issue is the range of ionic strengths where the porphyrin is expected to be partially dimerized. Consider a solution with an absorbance of 1.0 in a 1 cm cell; in this case [P0 ] 1/. Solutions with a signi®cant amount of both monomer and dimer occur for solutions in which KD is within approximately an order of magnitude of (1/M ). For porphyrins, with extinction coecients of approximately 105 Mÿ1 cmÿ1 , this implies 104 < KD <106 Mÿ1 . For ionic strengths which result in these KD values, the optical spectrum will be sensitive to the concentration of porphyrin because dimerization is signi®cant. In these instances, a detailed study of the absorbance as a function of concentration is necessary to measure KD accurately at the desired ionic strength. A second issue concerns accurate measurement of the total porphyrin concentration. If an experiment is to be run at an ionic strength where the porphyrin is partially dimerized, it is dicult to obtain an accurate measurement of the porphyrin concentration. However, if the porphyrin is not dimerized in pure water, the concentration of an aliquot at the ionic strength of the experiment can be measured by diluting an aliquot of the porphyrin into a large excess of water. For example, extrapolation of the TMAP data to zero ionic strength indicates a value for KD0 of 10)103 Mÿ1 . This in turn indicates that TMAP is monomeric in pure water at optical concentrations [40]. Therefore, one can calculate the concentration of TMAP for a set of experiments at a higher ionic strength by taking a spectrum of the TMAP at zero ionic strength and using the measured extinction coecients. Other porphyrins stack more extensively. For example, Fig. 8 indicates that THOPP is likely to be stacked even in pure water at optical concentrations. This is in line with independent evidence on the stacking
30
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
of this compound in aqueous solution; Schneider and Wang have reported that THOPP is aggregated even in the concentration range 2 ´ 10ÿ6 ±3 ´ 10ÿ5 M [14]. The interaction of most of these cationic porphyrins with DNA has been studied; the self-stacking properties of the porphyrin in aqueous solution have a signi®cant eect on the mode of interaction of the porphyrin with duplex DNA. In the alkylated pyridyl series, TMPyP4, which does not self-stack in aqueous solution, intercalates into calf thymus (CT) DNA [1±4]. The tetrabutyl derivative TBPyP4 also intercalates in CT DNA, though perhaps with some local melting or distortion of the DNA [64]. The tricationic ferrocenyl porphyrin P3Fc is an intercalator at low ionic strength and a groove binder at high ionic strength [47]. This is consistent with the dependence of self-stacking on ionic strength for this compound (Fig. 7). The competing equilibria are such that P3Fc intercalates into the DNA when it is monomeric in solution. However, when P3Fc is dimeric it prefers to bind to the exterior of the DNA. TMAP and THOPP both have positive charges on side chains o the tetraphenylpyridyl nucleus. Both TMAP, with trimethylanilinium substituents [1,6± 8,65,66], and THOPP with OCH2 CH2 CH2 N Me3 side chains [13±15], bind to duplex DNA with stacking of the porphyrin along the exterior of the DNA. Quantitation of the dierence between the binding modes of these two porphyrins is complicated by the dependence of the optical signatures of the DNA complexes of ionic strength, buer, the [porphyrin]/[DNA] ratio and the time after mixing. Self-stacking may help explain the dependence of the DNA binding mode on the role of the number and position of the positive charges of cationic porphyrins [67± 71]. An understanding of self-stacking will also be important as increasing use is made of cationic porphyrins covalently linked to other moieties which control DNA binding [72±81]. 4. Conclusions Salt-induced stacking experiments allow estimates of the extent of aggregation of the porphyrin as a function of ionic strength. This gives a prediction of ionic strength regimes in which the porphyrin is likely to be either fully monomeric (important for measuring concentrations in solution) or fully dimeric. In addition, these titrations give an indication of when it is necessary to perform detailed studies to determine the state of aggregation under the particular circumstances of the experiment. Understanding of the extent of aggregation of cationic porphyrins in water is of substantial help both in interpreting experiments and in designing new cationic porphyrins targeted toward particular binding sites.
5. Experimental section TMAP tetrachloride was used as received from Midcentury Chemicals (Posen, IL). The synthesis of TBPyP4 has been reported previously [64] and the synthesis of P3Fc will be described elsewhere [47]. UV/vis absorption spectra were obtained on a Shimadzu UV-3101PC spectrophotometer. Some extinction coecients were taken from the work of Peth o and Marzilli as (Mÿ1 cmÿ1 ): TMPyP4 2.63 ´ 105 (pH 6.1, kmax 421 nm); TMAP 4.85 ´ 105 (pH 7.1, kmax 410 nm); THOPP 6.13 ´ 105 (pH 7.1, kmax 417 nm) [82]. Peth o and Marzilli determined these extinction coecients by quantitating zinc remaining after complete zinc insertion into the cationic porphyrins. This technique removes problems of measured extinction coecients being inaccurate due to stacking of the porphyrin. The extinction coecient of TBPyP4 was assumed to be the same as that of TMPyP4. The extinction coecient of P3Fc was measured as 1.8 ´ 105 Mÿ1 cmÿ1 in N00 buer (7.5 mM NaH2 PO4 , 1 mM EDTA, pH 7.0; EDTA added to sequester any metals for photophysical experiments on this molecule) [47]. Salt titrations for TMAP and TBPyP4 were run in 1 mm cells by adding small amounts of solid NaCl to a reservoir of a solution of the porphyrin. Aliquots were then removed by syringe, placed in the 1 mm cell and the absorbance was taken. The P3Fc salt titration was run in a 1 cm cell in N00 buer. Dilute salt solutions were made by adding an aliquot of a concentrated NaCl solution to the cell. The more concentrated salt solutions were made by adding solid NaCl directly to the cell. In the salt titrations, the extinction coecients of the dimers were taken to be those of maximum change at the kmax of the monomer, reported as fM where the f values are: TMAP (0.25), TBPyP4 (0.63), THOPP (0.33), TOPyP4 (0.34) and P3Fc (0.29). Porphyrin concentration experiments for TMAP were run by adding small aliquot of a concentrated porphyrin solution to an aqueous solution of 1.0 M NaCl in a 1 cm cell. Data at low concentrations of porphyrin (<10ÿ6 M) did not fall on the line in Fig. 5. This may be in part because there is little dimerization at these low concentrations of TMAP. It may also be due to sticking of the porphyrin on the cell walls. Silanization of the cell walls before use by rinsing the cell with Sigmacote (Sigma) and air-drying made a dierence of about 5% in the absorbances but did not help the linearity of the plots at [TMAP] < 10ÿ6 M. The concentration of TMAP for the experiments in 1 M NaCl was calculated by taking an aliquot of the stock solution and diluting it into water (approximately 600-fold). Because TMAP is not stacked in pure water, Beer's Law could be used to calculate the concentration of TMAP in the stock solution.
Acknowledgements We thank Guangmin Pu for synthesis of the TBPyP4, Dr. Hania Wojtowicz for synthesis of P3Fc and Dr.
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
Nancy Thornton for assistance with the experiments. This work was supported in part by the National Institutes of Health (AI127196 and AI39391). Equipment was provided by the Georgia Research Alliance. References [1] R.J. Fiel, J. Biomol. Struct. Dynam. 6 (1989) 1259. [2] L.G. Marzilli, New J. Chem. 140 (1990) 4090. [3] G. Pratviel, J. Bernadou, B. Meunier, Metal Ions Biol. Syst. 33 (1996) 399. [4] R.F. Pasternack, E.J. Gibbs, Metal Ions Biol. Syst. 33 (1996) 367. [5] K. Kano, M. Takei, S. Hashimoto, J. Phys. Chem. 94 (1990) 2181. [6] M.J. Carvlin, N. Datta-Gupta, R.J. Fiel, Biochem. Biophys. Res. Commun. 108 (1982) 66. [7] M.J. Carvlin, R.J. Fiel, Nucleic Acids Res. 11 (1983) 6121. [8] K.G. Ford, S. Neidle, Bioorg. Med. Chem. 3 (1995) 671. [9] C.Z. Huang, K.A. Li, S.Y. Tong, Anal. Chem. 68 (1996) 2259. [10] B.R. Munson, R.J. Fiel, Nucleic Acids Res. 20 (1992) 1315. [11] R.F. Pasternack, C. Bustamante, P.J. Collings, A. Giannetto, E.J. Gibbs, J. Am. Chem. Soc. 115 (1993) 5393. [12] C. Bustamante, S. Gurrieri, R.F. Pasternack, R. Purrello, E. Rizzarelli, Biopolymers 34 (1994) 1099. [13] N.E. Mukundan, G. Peth o, D.W. Dixon, M.S. Kim, L.G. Marzilli, Inorg. Chem. 33 (1994) 4676. [14] H.-J. Schneider, M. Wang, J. Org. Chem. 59 (1994) 7473. [15] N.E. Mukundan, G. Peth o, D.W. Dixon, L.G. Marzilli, Inorg. Chem. 34 (1995) 3677. [16] A.S. Ito, G.C. Azzellini, S.C. Silva, O. Serra, A.G. Szabo, Biophys. Chem. 45 (1992) 79. [17] A. Losi, R. Bedotti, L. Brancaleon, C. Viappiani, J. Photochem. Photobiol. B 21 (1993) 69. [18] N. Datta-Gupta, D. Malakar, E. Walters, B. Thompson, Res. Commun. Chem. Path. Pharm. 60 (1988) 347. [19] I.E. Borissevitch, T.T. Tominaga, H. Imasato, M. Tabak, J. Luminescence 69 (1996) 65. [20] S. Arimori, M. Takeuchi, S. Shinkai, J. Am. Chem. Soc. 118 (1996) 245. [21] T. Nezu, S. Ikeda, Int. J. Biol. Macromol. 15 (1993) 101. [22] Y. Nitzan, R. Dror, H. Ladan, Z. Malik, S. Kimel, V. Gottfried, Photochem. Photobiol. 62 (1995) 342. [23] M. Merchat, G. Bertolini, P. Giacomini, A. Villanueva, G. Jori, J. Photochem. Photobiol. B 32 (1996) 153. [24] D.W. Dixon, L.G. Marzilli, R.F. Schinazi, Ann. NY Acad. Sci. 616 (1990) 511. [25] L. Ding, J. Balzarini, D. Schols, B. Meunier, E. De Clercq, Biochem. Pharmacol. 44 (1992) 1675. [26] C. Kasturi, M.S. Platz, Photochem. Photobiol. 56 (1992) 427. [27] M. Pitie, C. Casas, C.J. Lacey, G. Pratviel, J. Bernadou, B. Meunier, Angew. Chem. (Engl) 32 (1993) 557. [28] G.G. Meng, B.R. James, K.A. Skov, M. Korbelik, Can. J. Chem. 72 (1994) 2447. [29] K. Driaf, P. Krausz, B. Verneuil, M. Spiro, J.C. Blais, G. Bolbach, Tetrahedron Lett. 34 (1993) 1027. [30] L. Czuchajowski, H. Niedbala, Bioorg. Med. Chem. Lett. 2 (1992) 1645. [31] A. Villanueva, J. Photochem. Photobiol. B 2±3 (1993) 295. [32] T. Cernay, H.W. Zimmermann, J. Photochem. Photobiol. B 34 (1996) 191. [33] F. Ricchelli, J. Photochem. Photobiol. B 29 (1995) 109. [34] M.R. Wasielewski, Chem. Rev. 92 (1992) 435. [35] J. Hirota, T. Takeno, I. Okura, J. Photochem. Photobiol. A 77 (1994) 29.
31
[36] A.M. Brun, A. Harriman, J. Am. Chem. Soc. 116 (1994) 10383. [37] M.A. Bos, T.M. Werkhoven, J.M. Kleijn, Langmuir 12 (1996) 3980. [38] J. Wienke, F.J. Kleima, R.B.M. Koehorst, T.J. Schaafsma, Thin Solid Films 279 (1996) 87. [39] M.T. Martin, I. Prieto, L. Camacho, D. Mobius, Langmuir 12 (1996) 6554. [40] M. Krishnamurthy, Indian J. Chem. B 15 (1977) 964. [41] J.S. Chen, J. Chem. Soc., Faraday Trans. 90 (1994) 717. [42] (a) J.A. Goodwin, W.R. Scheidt, J. Phys. Chem. 94 (1990) 4432. (b) G. Budevsky. Foundation of Chemical Analysis. John Wiley and Sons, New York, 1979, pp. 49±51. [43] J.H. Espenson, Chemical Kinetics and Reaction Mechanisms, McGraw-Hill, New York, 1995, pp. 207±210. [44] M.D. Green, G. Patonay, T. Ndou, I.M. Warner, J. Inclusion Phenom. Mol. Recognit. Chem. 13 (1992) 181. [45] K. Hamada, K. Yamada, M. Mitsuishi, M. Ohira, K. Mesuda, J. Chem. Soc., Faraday Trans. 91 (1995) 1601. [46] M. Krishnamurthy, J.R. Sutter, Inorg. Chem. 24 (1985) 1943. [47] N.B. Thornton, H. Wojtowicz, T.L. Netzel, D.W. Dixon J. Phys. Chem. (1998) in press. [48] G. Peth o, N.B. Elliott, M.S. Kim, M. Lin, D.W. Dixon, L.G. Marzilli, J. Chem. Soc. Chem. Commun. 1993 (1993) 1547. [49] H. Baker, P. Hambright, L. Wagner, J. Am. Chem. Soc. 95 (1973) 5942. [50] P. Hambright, P.B. Chock, J. Am. Chem. Soc. 96 (1974) 3123. [51] J. Nwaeme, P. Hambright, Inorg. Chem. 23 (1984) 1990. [52] J.B. Reid, P. Hambright, Inorg. Chem. 16 (1977) 968. [53] A.N. Thompson, M. Krishnamurthy, J. Inorg. Nucl. Chem. 41 (1979) 1251. [54] R.F. Pasternack, P.R. Huber, P. Boyd, G. Engasser, L. Francesconi, E.J. Gibbs, P. Fasella, G. Cerio Venturo, L.d. Hinds, J. Am. Chem. Soc. 94 (1972) 4511. [55] N. Foster, J. Magn. Reson. 56 (1984) 140. [56] K.M. Kadish, B.G. Maiya, C. Araullo-McAdams, J. Phys. Chem. 95 (1991) 427. [57] E.J. Gibbs, I. Tinoco, Jr., M.F. Maestre, P.A. Ellinas, R.F. Pasternack, Biochem. Biophys. Res. Commun. 157 (1988) 350. [58] F.J. Vergeldt, R.B.M. Koehorst, A. Van Hoek, T.J. Schaafsma, J. Phys. Chem. 99 (1995) 4397. [59] M. Merchat, J.D. Spikes, G. Bertoloni, G. Jori, J. Photochem. Photobiol. B 35 (1996) 149. [60] R.L. Brook®eld, H. Ellul, A. Harriman, J. Photochem. 31 (1985) 97. [61] K. Kemnitz, T. Sakagushi, Chem. Phys. Lett. 196 (1992) 497. [62] K. Kano, T. Nakajima, M. Takei, S. Hashimoto, Bull. Chem. Soc. Jpn. 60 (1987) 1281. [63] K. Kano, T. Miyake, K. Uomoto, T. Sate, T. Ogawa, S. Hashimoto, Chem. Lett. (1983) 1867. [64] U. Sehlstedt, S.K. Kim, P. Carter, J. Goodisman, J.F. Vollano, B. Norden, J.C. Dabrowiak, Biochemistry 33 (1994) 417. [65] M.J. Carvlin, E. Mark, R.J. Fiel, Nucleic Acids Res. 11 (1983) 6141. [66] D.L. Banville, L.G. Marzilli, J.A. Strickland, W.D. Wilson, Biopolymers 25 (1986) 1837. [67] M.A. Sari, J.P. Battioni, D. Mansuy, J.B. Le Pecq, Biochem. Biophys. Res. Commun. 141 (1986) 643. [68] M.A. Sari, J.P. Battioni, D. Dupre, D. Mansuy, J.B. Le Pecq, Biochem. Pharmacol. 37 (1988) 1861. [69] M.A. Sari, J.P. Battioni, D. Dupre, D. Mansuy, J.B. Le Pecq, Biochemistry 29 (1990) 4205. [70] G.D. Strahan, D. Lu, M. Tsuboi, K. Nakamoto, J. Phys. Chem. 96 (1992) 6450. [71] D.T. Croke, L. Perrouault, M.A. Sari, J.P. Battioni, D. Mansuy, C. Helene, T. Le-Doan, J. Photochem. Photobiol. B 18 (1993) 41. [72] J.T. Groves, T.P. Farrell, J. Am. Chem. Soc. 111 (1989) 4998.
32
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
[73] A. Slama-Schwok, J.-M. Lehn, Biochemistry 29 (1990) 7895. [74] A. Salehi, H.-Y. Mei, T.C. Bruice, Tetrahedron Lett. 32 (1991) 3453. [75] G. Anneheim-Herbelin, M. Perree-Fauvet, A. Gaudemer, P. Helissey, S. Giorgi-Renault, N. Gresh, Tetrahedron Lett. 34 (1993) 7263. [76] X. Hui, N. Gresh, J. Biomol. Struct. Dynam. 11 (1993) 333. [77] M. Perree-Fauvet, N. Gresh, J. Biomol. Struct. Dynam. 11 (1994) 1203.
[78] C. Sentagne, S. Feryforgues, B. Meunier, N. Paillous, New J. Chem. 18 (1994) 1253. [79] M. Perree-Fauvet, N. Gresh, Tetrahedron Lett. 36 (1995) 4227. [80] N. Sugimoto, S. Nakano, M. Katoh, M. Sasaki, S. Kugimiya, Nucleosides Nucleotides 15 (1996) 743. [81] J.T. Groves, A.A. Matsuna, Bioorg. Med. Chem. Lett. 6 (1996) 1595. [82] G. Peth o, L.G. Marzilli, Microchem. J. 50 (1994) 178.
Dimerization of tetracationic porphyrins: ionic strength dependence Dabney W. Dixon *, Vera Steullet Department of Chemistry, Georgia State University, Atlanta, GA 30303, USA Received 29 May 1997; received in revised form 2 September 1997; accepted 2 September 1997
Abstract Cationic porphyrins are under study in a number of contexts including their interaction with biological targets, as possible therapeutic agents and as building blocks for molecular devices such as molecular photodiodes and solar cells. Many cationic porphyrins dimerize readily in aqueous solution. Dimerization in turn can control the properties of the porphyrin as well as its binding to its target. The propensity of a porphyrin to dimerize in aqueous solution can be estimated by recording the optical spectrum of the solution as a function of the concentration of added salt. Analysis of the data in terms of the Debye±H uckel formalism gives an estimate of the extent of dimerization as a function of ionic strength. Data for TMPyP4 [meso-tetrakis(4-N-methylpyridinium)porphyrin] and its butyl and octyl homologs; TMAP [meso-tetrakis(4-N,N,N-trimethylanilinium)porphyrin]; THOPP [meso-tetrakis[4-N-[(3-(trimethyl-ammonio)propyl)oxy]phenyl]porphyrin] and the ferrocenyl porphyrin P3Fc are discussed. Dimerization may aect binding of the cationic porphyrins to their targets, e.g., DNA. Ó 1998 Elsevier Science Inc. All rights reserved. Keywords: Porphyrin; Dimerization; Ionic strength; TMPyP4; TMAP
1. Introduction Many cationic porphyrins dimerize quite readily in aqueous solution. This dimerization can control the properties of the porphyrin itself as well as its binding to its target. The goal of the work described herein was to provide a rapid estimation of the propensity of a given cationic porphyrin to dimerize in aqueous solution. These estimates can then be used in designing and interpreting experiments in which cationic porphyrins bind to biological and physical targets. One area in which porphyrin aggregation is important is in the binding of cationic porphyrins to DNA [1±4]. Porphyrins with a strong propensity to self-stack in solution, e.g., TMAP [5±9], cis- and trans-P4 [10± Abbreviations: TMPyP4, meso-tetrakis(4-N-methylpyridinium)porphyrin; TBPyP4, meso-tetrakis(4-N-butylpyridinium)porphyrin; TOPyP4, meso-tetrakis(4-N-octylpyridinium)porphyrin; TMPyP2, mesotetrakis(2-N-methylpyridinium)porphyrin; TMAP, meso-tetrakis[4-N,N,N-trimethylanilinium)porphyrin; THOPP, meso-tetrakis[4-N-[(3-(trimethylammonio)propyl)oxy]phenyl]porphyrin; trans-P4 trans-bis (4N-methylpyridinium)diphenylporphyrin; cis-P4, cis-bis(4-N-methylpyridinium)diphenylporphyrin; Fc, ferrocene; P3Fc, 4,40 ,400 -[20-(2-ferrocenylethoxy)-21H,23H-porphine-5,10,15-triyl]tris[1-methylpyridinium] * Corresponding author. Tel.: 404 651 3908; fax: 404 651 1414; e-mail: [email protected]. 0162-0134/98/$19.00 Ó 1998 Elsevier Science Inc. All rights reserved. PII: S 0 1 6 2 - 0 1 3 4 ( 9 7 ) 1 0 0 0 5 - 8
12], and THOPP [13±15], prefer to bind in the grooves of DNA rather than to intercalate. Some porphyrins appear to bind in the grooves as monomers. For others, binding in a conformation which involves self-stacking of the porphyrins in the DNA grooves is indicated by conservative bands and high molar ellipticities in the circular dichroism spectra [4]. Other studies of cationic porphyrin interaction with biological targets involve complexes with melanin [16,17], albumin [18,19], sugars [20] and polyglutamic acid [21]. Proposed therapeutic uses of cationic porphyrins include the inactivation of bacteria [22,23] and viruses [24± 27] as well as in photodynamic therapy of cancer [28±31]. In studies bearing directly on the relationship between porphyrin properties and biological activity, Cernay and Zimmerman have observed that the lipophilicity of tetraalkylated tetrapyridylporphyrin derivatives controls the subcellular localization of these species [32]. The photophysical properties of porphyrins are known to depend signi®cantly on their aggregation state [33]. Understanding of porphyrin dimerization or aggregation is therefore particularly important in studies utilizing cationic porphyrins as electron acceptors in intramolecular electron transfer processes [34±36]. Self-stacking of cationic porphyrins may also be important in their deposition on surfaces. For example, Kleijn and co-workers have observed that the mode
26
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
and rate of deposition on a silica surface is a function of the ionic strength of the depositing solution [37]. Schaafsma and co-workers have observed that the absorption of cis-P4 on surfaces is a function of the concentration of the porphyrin; dilute solutions (5 ´ 10ÿ6 M) give monolayer coverage and more concentrated ( P 10ÿ4 M) solutions result in increased ordering on the surface [38]. Porphyrin aggregation may also be signi®cant in the stacking of TMPyP4 as dimers with lipid monolayers [39]. Dimerization and aggregation of cationic porphyrins can substantially aect both the properties of the molecule and the binding to its target. In studies of binding to biological targets, the properties of the biological target can de®ne the required buer conditions and ionic strength. Because the extent of stacking depends substantially on the ionic strength, one either needs to make a separate estimation of the dimerization constant at each ionic strength or to have an algorithm to predict how the dimerization depends on ionic strength. In addition, data analysis demands that the extinction coecient of the monomer be known. One needs enough data to be con®dent that the extinction coecient determined at low porphyrin concentration is in fact the monomer extinction coecient and not that of a partially dimerized species. Given these criteria, rapid methods of estimating the dimerization constant as a function of ionic strength are of interest. Herein, we show that the Debye±H uckel formalism provides a method for rapid estimation of the dimerization of cationic porphyrins (Fig. 1) as a function of ionic strength.
2. Results
Fig. 1. Structures of selected cationic porphyrins.
2.1. Salt-induced stacking An example of a porphyrin which shows changes in its optical spectrum as a function of the salt concentration of the solution is the tetrakis-trimethylanilinium derivative TMAP [5,40]. A salt titration was run using a 1 mm cell with relatively high [TMAP] to promote aggregation (Fig. 2). As the salt concentration increases, the absorption band at 412 nm decreases and a shoulder grows in to the blue at 402 nm; the titration appears isosbestic. Changes, though small, continue to take place from 3.9 M (shown) to saturating concentrations of NaCl. This data set allows one to estimate the extent of dimerization at any given ionic strength. A simple dimerization may be expressed as M M D with a dimerization constant KD , 2
KD D=M :
Fig. 2. UV/vis absorption spectrum of TMAP in water with increasing additions of NaCl as indicated by the ®gure legend (1 mm cell, P0 9:5 10ÿ6 M).
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
27
Expressing all concentrations in terms of the concentration of monomer, the total concentration of the porphyrin, [P]0 , is [41]: P0 M 2D: The mole fractions of monomer (fM ) and dimer (fD ) are de®ned as fM M=P 0 ; fD 2D=P0 : The measured extinction coecient, , is fM M fD D ; where M and D are the extinction coecients of the monomer and dimer, respectively. The mole fractions of monomer and dimer can be calculated from the optical data
Fig. 3. Log(KD ) vs. l1=2 = 1 l1=2 for TMAP in water.
fM ÿ D = M ÿ D ; fD M ÿ = M ÿ D : From these, one obtains the dimerization constant, KD , at each ionic strength KD fD =2P0 fM fM : Given fD and the total porphyrin concentration, [P]0 , an equilibrium constant can be calculated at each ionic strength. The calculation assumes that the extinction coecients of the monomer and dimer are independent of the ionic strength, and that one can obtain both the fully monomeric (no added salt) and fully dimeric (high ionic strength) forms. This may be an issue either because the system stacks beyond the dimer or because the extinction coecient of the dimer is a function of ionic strength. Both of these issues are discussed in more detail below. For TMAP, the extinction coecient of the dimer was taken to be 0.26 that of the monomer at the kmax of the monomer (from the spectrum in saturated NaCl). The calculated equilibrium constants calculated from the data in Fig. 2 are 2.6 ´ 104 , 7.4 ´ 104 , and 1.5 ´ 106 at NaCl concentrations of 0.48, 0.97, and 2.9 M, respectively. To use this data in predicting equilibrium constants at other ionic strengths, one needs a model for the dependence of the equilibrium constant on ionic strength. The change in KD with ionic strength can be treated using the Debye±H uckel limiting law log KD log KD0 2AzA zB l1=2 = 1 l1=2 ; where KD0 is the limiting value of KD at zero ionic strength and zA and zB are the charges on the respective halves of the dimer (here equal) [42]. For water, A is 0.509 at room temperature [43]. Fig. 3 shows a plot of log(KD ) vs. l1=2 /(1 + l1=2 ) for TMAP. For TMAP, the data are monotonic, but not linear, over the ionic strength range studied. The absorbance of TMAP as a function of [P]0 at 0 and 1 M NaCl is given in Fig. 4. The line for 1 M NaCl curves downward with increasing [P]0 because the eective extinction coecient of the porphyrin decreases as it dimerizes with increasing [P]0 . If the extinction coe-
Fig. 4. Absorbance vs. [TMAP] with 0.0 and 1.0 M added NaCl. In the solution with no added salt (), the line through the points is a Beer's Law ®t for the data. In the solution with 1.0 M added NaCl (n), the line through the points is the predicted absorbance calculated from the equations in the text and the calculated D and KD values of 2:9 105 Mÿ1 cmÿ1 and 5:8 105 Mÿ1 , respectively.
cient of the monomer is known, data sets of this type can be ®t [41,44,45]. D ÿ f M ÿ D =2Kg1=2 f M ÿ =P0 g1=2 : A plot of vs.{( ) M )/[P]0 }1=2 gives D from the intercept; KD can be calculated from the slope (Fig. 5). For this data set (silanized cell) they are 3.1 0.1 ´ 105 Mÿ1 cmÿ1 and 3.3 0.7 ´ 106 Mÿ1 , respectively. A second data set (unsilanized cell) gave 2.8 0.1 ´ 105 Mÿ1 cmÿ1 and 1.7 0.2 ´ 106 Mÿ1 , respectively. These compare with approximate values derived from the salt titration of 1.2 ´ 105 Mÿ1 cmÿ1 and 7.4 ´ 104 Mÿ1 , respectively. Thus, the salt titration gives estimated, but not exact, values for D and KD . This is partly because the spectrum of the ``fully dimerized'' porphyrin in the salt titration is not accurate, in this instance because it is not really fully dimerized but in other instances because aggregation into larger structures has begun to take place. In addition, the structure of the dimer itself, and hence its optical signature, is expected to be a function of the ionic strength. It is interesting to note that the equilibrium constants for TMAP here are similar to KD for dimerization of [tetrakis(p-(trimethyl-
28
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
Fig. 5. A plot of vs.f M ÿ =P0 g1=2 for the TMAP data in Fig. 4.
ammonio)phenyl)porphinato]silver(III), Ag(III)TMAP, 1.2 ´ 105 Mÿ1 (0.05 M NaNO3 , pH 7.0) [46]. TBPyP4: The salt-induced stacking of TBPyP4 was also studied by recording the optical spectrum of the compound as a function of added NaCl. The titration was run in a 1 mm cell so that a relatively high concentration of TBPyP4 could be employed to increase stacking. Fig. 6 shows that the Soret band at 424 nm decreased in intensity with almost no change in wavelength as the salt concentration was increased. Changes continued even up to saturating amounts of NaCl. P3Fc: The ferrocenyl porphyrin P3Fc has been studied in conjunction with the role of DNA scaolding in controlling electron transfer [47]. Fig. 7 shows the absorption spectrum of P3Fc in N00 buer (pH 7.0) with increasing additions of NaCl (from 0 to 0.92 M NaCl added). As the concentration of NaCl increases, the intensity of the Soret band at 425 nm decreases; above 25 mM NaCl a new peak, at 447 nm, appears. A single isosbestic point is visible at 443 nm. These spectral changes are consistent with self-complexation of the porphyrin. It should be noted that a kmax of approximately 450 nm is also a signature for protonation
Fig. 6. UV/vis absorption spectrum of TBPyP4 in water with increasing additions of NaCl as indicated by the ®gure legend (1 mm cell, P0 6:1 10ÿ5 M).
Fig. 7. UV/vis absorption spectrum of P3Fc in N00 buer with increasing additions of NaCl as indicated by the ®gure legend (1 cm cell, P0 4:5 10ÿ6 M).
of the center nitrogen atoms of the porphyrin [48]. For P3Fc, attribution of this spectral change to protonation seems unlikely because the pKa values are expected to be too low. Data are available for TMPyP4; approximate values of pKa3 (protonation of the third central nitrogen) and pKa4 (protonation of the fourth central nitrogen) are 2 and 1, respectively [49,50]. The values of pKa3 and pKa4 are sensitive to buer strength, but not suciently to result in protonation of the center nitrogens even at high ionic strength. For example, the pKa3 value for TMPyP4 is 1.4 in 0.2 M NaNO3 and 2.2 in 2.0 M NaNO3 [49]. 3. Discussion In many instances, the extent of dimerization of a porphyrin is an important aspect of experimental design. The dependence of KD on ionic strength can be predicted in a useful manner from a single titration in which salt is added to an aqueous solution of the porphyrin. Estimating dimerization constants from salt-induced stacking assumes that one can obtain the limiting extinction coecients of the monomer and dimer. This will not be possible in many cases, e.g., if one cannot go to low enough concentrations to obtain a solution of only the monomer or if aggregates beyond the dimer begin to form at high ionic strength. The extinction coecients of the dimer may also be aected by the ionic strength of the solution, presumably because the geometry of the dimer is changing slightly as the ionic strength changes. However, even with these caveats, salt-induced selfstacking can be a useful indication of the propensity of a porphyrin to dimerize in solution. In this approach, the Debye±H uckel limiting law is employed at much higher ionic strengths than the assumptions of its derivation warrant. Whether or not extrapolation to zero ionic strength gives a true value of log(KD0 ), plots of log(KD ) vs. l1=2 /(1 + l1=2 ) are monotonic or linear for a number of cationic porphyrins. Fig. 8 shows data plotted from the TMAP, TBPyP4 and P3Fc experiments herein as well as from two salt ti-
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
Fig. 8. Log(KD ) vs. l1=2 = 1 l1=2 for cationic porphyrins in water (the limiting extinction coecients for the monomers and dimers are given in the experimental).
trations in the literature: those for THOPP, a tetraphenylporphyrin bearing OCH2 CH2 CH2 N Me3 side chains [13], and TOPyP4 with octyl side chains [5]. These ®ve porphyrins stack to very dierent extents in aqueous solution. The extent of curvature of the data is a function of the ionic strength range necessary to produce stacking of the porphyrin. Porphyrins which stack at low ionic strength, e.g., THOPP and P3Fc, show a relatively linear dependence of log(KD ) on l1=2 /(1 + l1=2 ). Porphyrins which stack only at high ionic strength, e.g., TMAP and TBPyP4, show substantial curvature in the plot. This is expected because the assumptions underlying the formalism break down at higher ionic strengths. The charges on the porphyrins calculated from the Debye±H uckel formalism are generally less than the formal charges on the porphyrins (4+ in each case here). Calculating the charges from the slopes (2AzA zB , zA zB ) at the lower ionic strengths, one obtains approximately 3.6, 3.6, 4.0, 2.4 and 3.0 for THOPP, P3Fc, TOPyP4, TMAP, and TBPyP4, respectively. Similar eects have been observed a number of times previously. For example, the kinetics of dimerization of Ag(III)TMAP are best ®t with a charge of approximately +2 on the porphyrin monomer, though the formal charge is +5 [46]. The ionic strength dependence of Cu2 insertion into TMPyP2 is ®t to a charge of +2.8; the formal charge on TMPyP2 is +4 [51,52]. Similarly, the ionic strength dependence of Zn2 insertion into TMPyP4 is ®t to a charge of +2.4; the formal charge is again +4 [50,51]. The ionic strength dependencies of the acid solvolyses of Zn(II)TMPyP4 and Zn(II)TMAP also are ®t by a net charge on the complex that is substantially less than the formal charge [51,53]. The dierence between the formal charges and those calculated from the ionic strength dependencies may re¯ect ion pairing of the cationic porphyrins with counterions in solution [46]. Ion pairing has also been invoked to explain aspects of the NMR spectra of cationic porphyrins [54±56]. An interesting series is provided by the methyl (TMPyP4), butyl (TBPyP4) and octyl (TOPyP4) tetraalkylated homologs of tetrapyridyl porphyrin. Kano et al., observed that the optical spectrum of the octyl derivative, TOPyP4, was a function of salt concentration [5].
29
An isosbestic red shift of the Soret by about 18 nm up to an NaCl concentration of 0.5 M was seen. At higher ionic strengths, the Soret shifted back toward the blue in a non-isosbestic manner. These observations are consistent with initial dimerization, followed by stacking into larger structures as the ionic strength is increased. The tetrabutyl derivative TBPyP4, in contrast, undergoes salt-induced stacking with only a minimal change in the wavelength of the Soret (Fig. 6). Changes occur gradually as the salt concentration is raised but are not complete even with saturating amounts of NaCl. The tetramethyl derivative, TMPyP4, shows no signi®cant optical changes in water even at 2±3 M NaCl [5,57]. This lack of spectral change with increasing salt concentration indicates that TMPyP4 does not aggregate readily. The apparent lack of stacking even at these high ionic strengths is consistent with TMPyP4 being monomeric at optical concentrations in water. Other groups have also concluded that TMPyP4 is monomeric at optical concentrations [1,11,14,16,56,58,59], although early ¯uorescence lifetime-based studies were interpreted as evidence for dimers [60±63]. The dimerization estimates in Fig. 8 are useful both in interpreting data under the reaction conditions and in measuring the total concentration of the porphyrin. The ®rst issue is the range of ionic strengths where the porphyrin is expected to be partially dimerized. Consider a solution with an absorbance of 1.0 in a 1 cm cell; in this case [P0 ] 1/. Solutions with a signi®cant amount of both monomer and dimer occur for solutions in which KD is within approximately an order of magnitude of (1/M ). For porphyrins, with extinction coecients of approximately 105 Mÿ1 cmÿ1 , this implies 104 < KD <106 Mÿ1 . For ionic strengths which result in these KD values, the optical spectrum will be sensitive to the concentration of porphyrin because dimerization is signi®cant. In these instances, a detailed study of the absorbance as a function of concentration is necessary to measure KD accurately at the desired ionic strength. A second issue concerns accurate measurement of the total porphyrin concentration. If an experiment is to be run at an ionic strength where the porphyrin is partially dimerized, it is dicult to obtain an accurate measurement of the porphyrin concentration. However, if the porphyrin is not dimerized in pure water, the concentration of an aliquot at the ionic strength of the experiment can be measured by diluting an aliquot of the porphyrin into a large excess of water. For example, extrapolation of the TMAP data to zero ionic strength indicates a value for KD0 of 10)103 Mÿ1 . This in turn indicates that TMAP is monomeric in pure water at optical concentrations [40]. Therefore, one can calculate the concentration of TMAP for a set of experiments at a higher ionic strength by taking a spectrum of the TMAP at zero ionic strength and using the measured extinction coecients. Other porphyrins stack more extensively. For example, Fig. 8 indicates that THOPP is likely to be stacked even in pure water at optical concentrations. This is in line with independent evidence on the stacking
30
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
of this compound in aqueous solution; Schneider and Wang have reported that THOPP is aggregated even in the concentration range 2 ´ 10ÿ6 ±3 ´ 10ÿ5 M [14]. The interaction of most of these cationic porphyrins with DNA has been studied; the self-stacking properties of the porphyrin in aqueous solution have a signi®cant eect on the mode of interaction of the porphyrin with duplex DNA. In the alkylated pyridyl series, TMPyP4, which does not self-stack in aqueous solution, intercalates into calf thymus (CT) DNA [1±4]. The tetrabutyl derivative TBPyP4 also intercalates in CT DNA, though perhaps with some local melting or distortion of the DNA [64]. The tricationic ferrocenyl porphyrin P3Fc is an intercalator at low ionic strength and a groove binder at high ionic strength [47]. This is consistent with the dependence of self-stacking on ionic strength for this compound (Fig. 7). The competing equilibria are such that P3Fc intercalates into the DNA when it is monomeric in solution. However, when P3Fc is dimeric it prefers to bind to the exterior of the DNA. TMAP and THOPP both have positive charges on side chains o the tetraphenylpyridyl nucleus. Both TMAP, with trimethylanilinium substituents [1,6± 8,65,66], and THOPP with OCH2 CH2 CH2 N Me3 side chains [13±15], bind to duplex DNA with stacking of the porphyrin along the exterior of the DNA. Quantitation of the dierence between the binding modes of these two porphyrins is complicated by the dependence of the optical signatures of the DNA complexes of ionic strength, buer, the [porphyrin]/[DNA] ratio and the time after mixing. Self-stacking may help explain the dependence of the DNA binding mode on the role of the number and position of the positive charges of cationic porphyrins [67± 71]. An understanding of self-stacking will also be important as increasing use is made of cationic porphyrins covalently linked to other moieties which control DNA binding [72±81]. 4. Conclusions Salt-induced stacking experiments allow estimates of the extent of aggregation of the porphyrin as a function of ionic strength. This gives a prediction of ionic strength regimes in which the porphyrin is likely to be either fully monomeric (important for measuring concentrations in solution) or fully dimeric. In addition, these titrations give an indication of when it is necessary to perform detailed studies to determine the state of aggregation under the particular circumstances of the experiment. Understanding of the extent of aggregation of cationic porphyrins in water is of substantial help both in interpreting experiments and in designing new cationic porphyrins targeted toward particular binding sites.
5. Experimental section TMAP tetrachloride was used as received from Midcentury Chemicals (Posen, IL). The synthesis of TBPyP4 has been reported previously [64] and the synthesis of P3Fc will be described elsewhere [47]. UV/vis absorption spectra were obtained on a Shimadzu UV-3101PC spectrophotometer. Some extinction coecients were taken from the work of Peth o and Marzilli as (Mÿ1 cmÿ1 ): TMPyP4 2.63 ´ 105 (pH 6.1, kmax 421 nm); TMAP 4.85 ´ 105 (pH 7.1, kmax 410 nm); THOPP 6.13 ´ 105 (pH 7.1, kmax 417 nm) [82]. Peth o and Marzilli determined these extinction coecients by quantitating zinc remaining after complete zinc insertion into the cationic porphyrins. This technique removes problems of measured extinction coecients being inaccurate due to stacking of the porphyrin. The extinction coecient of TBPyP4 was assumed to be the same as that of TMPyP4. The extinction coecient of P3Fc was measured as 1.8 ´ 105 Mÿ1 cmÿ1 in N00 buer (7.5 mM NaH2 PO4 , 1 mM EDTA, pH 7.0; EDTA added to sequester any metals for photophysical experiments on this molecule) [47]. Salt titrations for TMAP and TBPyP4 were run in 1 mm cells by adding small amounts of solid NaCl to a reservoir of a solution of the porphyrin. Aliquots were then removed by syringe, placed in the 1 mm cell and the absorbance was taken. The P3Fc salt titration was run in a 1 cm cell in N00 buer. Dilute salt solutions were made by adding an aliquot of a concentrated NaCl solution to the cell. The more concentrated salt solutions were made by adding solid NaCl directly to the cell. In the salt titrations, the extinction coecients of the dimers were taken to be those of maximum change at the kmax of the monomer, reported as fM where the f values are: TMAP (0.25), TBPyP4 (0.63), THOPP (0.33), TOPyP4 (0.34) and P3Fc (0.29). Porphyrin concentration experiments for TMAP were run by adding small aliquot of a concentrated porphyrin solution to an aqueous solution of 1.0 M NaCl in a 1 cm cell. Data at low concentrations of porphyrin (<10ÿ6 M) did not fall on the line in Fig. 5. This may be in part because there is little dimerization at these low concentrations of TMAP. It may also be due to sticking of the porphyrin on the cell walls. Silanization of the cell walls before use by rinsing the cell with Sigmacote (Sigma) and air-drying made a dierence of about 5% in the absorbances but did not help the linearity of the plots at [TMAP] < 10ÿ6 M. The concentration of TMAP for the experiments in 1 M NaCl was calculated by taking an aliquot of the stock solution and diluting it into water (approximately 600-fold). Because TMAP is not stacked in pure water, Beer's Law could be used to calculate the concentration of TMAP in the stock solution.
Acknowledgements We thank Guangmin Pu for synthesis of the TBPyP4, Dr. Hania Wojtowicz for synthesis of P3Fc and Dr.
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
Nancy Thornton for assistance with the experiments. This work was supported in part by the National Institutes of Health (AI127196 and AI39391). Equipment was provided by the Georgia Research Alliance. References [1] R.J. Fiel, J. Biomol. Struct. Dynam. 6 (1989) 1259. [2] L.G. Marzilli, New J. Chem. 140 (1990) 4090. [3] G. Pratviel, J. Bernadou, B. Meunier, Metal Ions Biol. Syst. 33 (1996) 399. [4] R.F. Pasternack, E.J. Gibbs, Metal Ions Biol. Syst. 33 (1996) 367. [5] K. Kano, M. Takei, S. Hashimoto, J. Phys. Chem. 94 (1990) 2181. [6] M.J. Carvlin, N. Datta-Gupta, R.J. Fiel, Biochem. Biophys. Res. Commun. 108 (1982) 66. [7] M.J. Carvlin, R.J. Fiel, Nucleic Acids Res. 11 (1983) 6121. [8] K.G. Ford, S. Neidle, Bioorg. Med. Chem. 3 (1995) 671. [9] C.Z. Huang, K.A. Li, S.Y. Tong, Anal. Chem. 68 (1996) 2259. [10] B.R. Munson, R.J. Fiel, Nucleic Acids Res. 20 (1992) 1315. [11] R.F. Pasternack, C. Bustamante, P.J. Collings, A. Giannetto, E.J. Gibbs, J. Am. Chem. Soc. 115 (1993) 5393. [12] C. Bustamante, S. Gurrieri, R.F. Pasternack, R. Purrello, E. Rizzarelli, Biopolymers 34 (1994) 1099. [13] N.E. Mukundan, G. Peth o, D.W. Dixon, M.S. Kim, L.G. Marzilli, Inorg. Chem. 33 (1994) 4676. [14] H.-J. Schneider, M. Wang, J. Org. Chem. 59 (1994) 7473. [15] N.E. Mukundan, G. Peth o, D.W. Dixon, L.G. Marzilli, Inorg. Chem. 34 (1995) 3677. [16] A.S. Ito, G.C. Azzellini, S.C. Silva, O. Serra, A.G. Szabo, Biophys. Chem. 45 (1992) 79. [17] A. Losi, R. Bedotti, L. Brancaleon, C. Viappiani, J. Photochem. Photobiol. B 21 (1993) 69. [18] N. Datta-Gupta, D. Malakar, E. Walters, B. Thompson, Res. Commun. Chem. Path. Pharm. 60 (1988) 347. [19] I.E. Borissevitch, T.T. Tominaga, H. Imasato, M. Tabak, J. Luminescence 69 (1996) 65. [20] S. Arimori, M. Takeuchi, S. Shinkai, J. Am. Chem. Soc. 118 (1996) 245. [21] T. Nezu, S. Ikeda, Int. J. Biol. Macromol. 15 (1993) 101. [22] Y. Nitzan, R. Dror, H. Ladan, Z. Malik, S. Kimel, V. Gottfried, Photochem. Photobiol. 62 (1995) 342. [23] M. Merchat, G. Bertolini, P. Giacomini, A. Villanueva, G. Jori, J. Photochem. Photobiol. B 32 (1996) 153. [24] D.W. Dixon, L.G. Marzilli, R.F. Schinazi, Ann. NY Acad. Sci. 616 (1990) 511. [25] L. Ding, J. Balzarini, D. Schols, B. Meunier, E. De Clercq, Biochem. Pharmacol. 44 (1992) 1675. [26] C. Kasturi, M.S. Platz, Photochem. Photobiol. 56 (1992) 427. [27] M. Pitie, C. Casas, C.J. Lacey, G. Pratviel, J. Bernadou, B. Meunier, Angew. Chem. (Engl) 32 (1993) 557. [28] G.G. Meng, B.R. James, K.A. Skov, M. Korbelik, Can. J. Chem. 72 (1994) 2447. [29] K. Driaf, P. Krausz, B. Verneuil, M. Spiro, J.C. Blais, G. Bolbach, Tetrahedron Lett. 34 (1993) 1027. [30] L. Czuchajowski, H. Niedbala, Bioorg. Med. Chem. Lett. 2 (1992) 1645. [31] A. Villanueva, J. Photochem. Photobiol. B 2±3 (1993) 295. [32] T. Cernay, H.W. Zimmermann, J. Photochem. Photobiol. B 34 (1996) 191. [33] F. Ricchelli, J. Photochem. Photobiol. B 29 (1995) 109. [34] M.R. Wasielewski, Chem. Rev. 92 (1992) 435. [35] J. Hirota, T. Takeno, I. Okura, J. Photochem. Photobiol. A 77 (1994) 29.
31
[36] A.M. Brun, A. Harriman, J. Am. Chem. Soc. 116 (1994) 10383. [37] M.A. Bos, T.M. Werkhoven, J.M. Kleijn, Langmuir 12 (1996) 3980. [38] J. Wienke, F.J. Kleima, R.B.M. Koehorst, T.J. Schaafsma, Thin Solid Films 279 (1996) 87. [39] M.T. Martin, I. Prieto, L. Camacho, D. Mobius, Langmuir 12 (1996) 6554. [40] M. Krishnamurthy, Indian J. Chem. B 15 (1977) 964. [41] J.S. Chen, J. Chem. Soc., Faraday Trans. 90 (1994) 717. [42] (a) J.A. Goodwin, W.R. Scheidt, J. Phys. Chem. 94 (1990) 4432. (b) G. Budevsky. Foundation of Chemical Analysis. John Wiley and Sons, New York, 1979, pp. 49±51. [43] J.H. Espenson, Chemical Kinetics and Reaction Mechanisms, McGraw-Hill, New York, 1995, pp. 207±210. [44] M.D. Green, G. Patonay, T. Ndou, I.M. Warner, J. Inclusion Phenom. Mol. Recognit. Chem. 13 (1992) 181. [45] K. Hamada, K. Yamada, M. Mitsuishi, M. Ohira, K. Mesuda, J. Chem. Soc., Faraday Trans. 91 (1995) 1601. [46] M. Krishnamurthy, J.R. Sutter, Inorg. Chem. 24 (1985) 1943. [47] N.B. Thornton, H. Wojtowicz, T.L. Netzel, D.W. Dixon J. Phys. Chem. (1998) in press. [48] G. Peth o, N.B. Elliott, M.S. Kim, M. Lin, D.W. Dixon, L.G. Marzilli, J. Chem. Soc. Chem. Commun. 1993 (1993) 1547. [49] H. Baker, P. Hambright, L. Wagner, J. Am. Chem. Soc. 95 (1973) 5942. [50] P. Hambright, P.B. Chock, J. Am. Chem. Soc. 96 (1974) 3123. [51] J. Nwaeme, P. Hambright, Inorg. Chem. 23 (1984) 1990. [52] J.B. Reid, P. Hambright, Inorg. Chem. 16 (1977) 968. [53] A.N. Thompson, M. Krishnamurthy, J. Inorg. Nucl. Chem. 41 (1979) 1251. [54] R.F. Pasternack, P.R. Huber, P. Boyd, G. Engasser, L. Francesconi, E.J. Gibbs, P. Fasella, G. Cerio Venturo, L.d. Hinds, J. Am. Chem. Soc. 94 (1972) 4511. [55] N. Foster, J. Magn. Reson. 56 (1984) 140. [56] K.M. Kadish, B.G. Maiya, C. Araullo-McAdams, J. Phys. Chem. 95 (1991) 427. [57] E.J. Gibbs, I. Tinoco, Jr., M.F. Maestre, P.A. Ellinas, R.F. Pasternack, Biochem. Biophys. Res. Commun. 157 (1988) 350. [58] F.J. Vergeldt, R.B.M. Koehorst, A. Van Hoek, T.J. Schaafsma, J. Phys. Chem. 99 (1995) 4397. [59] M. Merchat, J.D. Spikes, G. Bertoloni, G. Jori, J. Photochem. Photobiol. B 35 (1996) 149. [60] R.L. Brook®eld, H. Ellul, A. Harriman, J. Photochem. 31 (1985) 97. [61] K. Kemnitz, T. Sakagushi, Chem. Phys. Lett. 196 (1992) 497. [62] K. Kano, T. Nakajima, M. Takei, S. Hashimoto, Bull. Chem. Soc. Jpn. 60 (1987) 1281. [63] K. Kano, T. Miyake, K. Uomoto, T. Sate, T. Ogawa, S. Hashimoto, Chem. Lett. (1983) 1867. [64] U. Sehlstedt, S.K. Kim, P. Carter, J. Goodisman, J.F. Vollano, B. Norden, J.C. Dabrowiak, Biochemistry 33 (1994) 417. [65] M.J. Carvlin, E. Mark, R.J. Fiel, Nucleic Acids Res. 11 (1983) 6141. [66] D.L. Banville, L.G. Marzilli, J.A. Strickland, W.D. Wilson, Biopolymers 25 (1986) 1837. [67] M.A. Sari, J.P. Battioni, D. Mansuy, J.B. Le Pecq, Biochem. Biophys. Res. Commun. 141 (1986) 643. [68] M.A. Sari, J.P. Battioni, D. Dupre, D. Mansuy, J.B. Le Pecq, Biochem. Pharmacol. 37 (1988) 1861. [69] M.A. Sari, J.P. Battioni, D. Dupre, D. Mansuy, J.B. Le Pecq, Biochemistry 29 (1990) 4205. [70] G.D. Strahan, D. Lu, M. Tsuboi, K. Nakamoto, J. Phys. Chem. 96 (1992) 6450. [71] D.T. Croke, L. Perrouault, M.A. Sari, J.P. Battioni, D. Mansuy, C. Helene, T. Le-Doan, J. Photochem. Photobiol. B 18 (1993) 41. [72] J.T. Groves, T.P. Farrell, J. Am. Chem. Soc. 111 (1989) 4998.
32
D.W. Dixon, V. Steullet / J. Inorg. Biochem. 69 (1998) 25±32
[73] A. Slama-Schwok, J.-M. Lehn, Biochemistry 29 (1990) 7895. [74] A. Salehi, H.-Y. Mei, T.C. Bruice, Tetrahedron Lett. 32 (1991) 3453. [75] G. Anneheim-Herbelin, M. Perree-Fauvet, A. Gaudemer, P. Helissey, S. Giorgi-Renault, N. Gresh, Tetrahedron Lett. 34 (1993) 7263. [76] X. Hui, N. Gresh, J. Biomol. Struct. Dynam. 11 (1993) 333. [77] M. Perree-Fauvet, N. Gresh, J. Biomol. Struct. Dynam. 11 (1994) 1203.
[78] C. Sentagne, S. Feryforgues, B. Meunier, N. Paillous, New J. Chem. 18 (1994) 1253. [79] M. Perree-Fauvet, N. Gresh, Tetrahedron Lett. 36 (1995) 4227. [80] N. Sugimoto, S. Nakano, M. Katoh, M. Sasaki, S. Kugimiya, Nucleosides Nucleotides 15 (1996) 743. [81] J.T. Groves, A.A. Matsuna, Bioorg. Med. Chem. Lett. 6 (1996) 1595. [82] G. Peth o, L.G. Marzilli, Microchem. J. 50 (1994) 178.