- Email: [email protected]
Alternating current aspects of antimicrobial molecular complexes
International Journal of Antimicrobial Agents 14 (2000) 253 – 259 www.ischemo.org
Alternating current aspects of antimicrobial molecular complexes H. Keyzer a,*, H.K. Kim b, K.C. Pan a a
Department of Chemistry and Biochemistry, California State Uni6ersity, Los Angeles, 5151 State Uni6ersity Dri6e, Los Angeles, CA 90032, USA b Department of Pharmaceutical Sciences, Uni6ersity of Southern California, 1985 Zonal A6e., Los Angeles, CA 90033, USA
Abstract A novel method was designed involving the titration of alternating current titration in a cell where one electrode was shielded by a capillary enclosure restricting access to it by charge carriers. With this cell, charge transfer complex titration of several thiazines all with some antimicrobial properties was effected in acetonitrile, against the electron accepting molecule iodine. The maxima of the Job plots generated by these titrations exhibited displacement of their positions for the forward and reverse titrations with respect to electron donor–acceptor complexation conductivity and apparent stoichiometry. A plot of inverse conductivity maxima differences against literature-cited dipole moments squared yielded a straight line passing through the origin. The titration plot profiles are discussed in terms of the type, number, and mobility of charge carriers produced in the complexation interaction. This novel method may be used to determine dipole moments of bioactive homologues empirically. There was correlation between thiazine drug dipole moments and minimal inhibitory concentrations of these drugs for some bacterial and yeast species. Several new avenues of investigation of possible relevance to microbiology are suggested. © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. Keywords: Antimicrobials; Thiazines; Psychotropes; Charge transfer complexes; Dipole moments; Alternating current titrations; Capillary shield cell; Microorganisms
1. Introduction Reliable data for dipole moments of bioactive molecules especially those with antimicrobial properties are relatively scarce, yet this parameter is important to their mechanism of action, a fact observed in many dielectrophoresis experiments [1 – 3]. Fro¨hlich considered biological systems as three-dimensional sets of interacting dipoles [4,5]. Del Giudice and coworkers used this information as a basis in the successful quantum field treatment for explaining observable manifestations of energy and electron transfer in the self-organization, dissipativity, and stability of biological assemblies [6]. The interaction of thiazine antimicrobials with bacteria and other microorganisms includes surface effects and penetration of their membranes. Drugs in this category are all surface-active amphipathics, and exert much of their influence by forming charge transfer complexes with membrane components and by altering the stability of membrane assemblies [7]. * Corresponding author. Tel./fax: +1-760-365-2716. E-mail address: [email protected] (H. Keyzer)
The lipid bilayer of a membrane forms a sandwich of distinguishable dielectric layers as determined by Thornton and coworkers [7] for erythrocytes from Raman experiments, and confirmed by Karolis [8] from capacitance and conductance measurements of mixed lipids, as well as reconstituted erythrocyte membranes. The dielectric spatial structure and surface morphology of a cell are important for the control of energy transactions involving dipole moments. A biological membrane can be likened to a parametron semiconductor circuit with a one-to-one correspondence analog to a dipole double layer [7]. A common feature of charge transfer across a biological membrane is its mode of passage through existing or induced restricted orifices which involves dipole interaction [7]. Knowledge of dipole moments of amphipathic drugs is therefore likely to be important as a requisite for understanding their action on and through microbial membranes. Unfortunately, dipole moment data of drug molecules are not only exceedingly scarce, but dipole moment changes of drug/microbe interactions are non-existent.
0924-8579/00/$ 20 © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. PII: S 0 9 2 4 - 8 5 7 9 ( 9 9 ) 0 0 1 6 2 - 4
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
254
We introduce a novel and relatively simple method for the study of charge transfer amongst dipolar bioactive molecules between two electrodes, one of which was shielded by a capillary orifice as the first step to producing a method which may measure directly drug dipole moment changes when it contacts the surface of a microorganism. Such a capillary was intended to mimic, to a first approximation, a simplistic version of a biomembrane pore. Charge diffusion through the capillary will be influenced by the mobility of the charge carriers, their dipole moments, and the size of the capillary. The charge transfer reaction produces charge carriers in high dielectric constant solvents as follows: solvent
D + A lDA l D+A− l D+ +A−
donor
acceptor
(1)
charge carriers
For direct current (DC) titrations involving the pH or potentiometric determinations, one may use a first derivative plot to obtain with greater precision the endpoint of the familiar sigmoid curve generated by the Nernst equation: E= E0 + k log[Ox]/[Red]
(2)
where E is the potential measured, E0 is the oxidation potential of the Redox couple, k is a constant, and [Ox] and [Red] are the oxidant and reductant concentrations, charge donor D and charge acceptor A, respectively. However, rather than creating a numerical plot of the first derivative curve to obtain the equivalence point with greater precision, one may use direct experimental methods to obtain the same result. The first direct method involves the use of two dissimilar electrodes, a platinum/graphite or a platinum/tungsten pair [9,10]. Alloys may also be used, as in the case of the platinum/platinum – 10% rhodium pair [10]. If the titrant is added uniformly at a rapid rate, one electrode lags slightly in its response to the changes in solution composition. As a result the electrode pair will exhibit a maximum difference in re-
sponse at the equivalence point when the logarithmic term of the Nernst equation changes the most rapidly, and a first derivative curve is obtained automatically. A second method employs a different strategy. One electrode of a platinum/platinum pair, is encased in a glass tube provided with a capillary orifice. The concentration in the capillary thus lags slightly behind that of the bulk of the solution during addition of the titrant, again yielding a first derivative plot automatically [10]. The advantage of this method is its simplicity, and that any suitable metal electrodes may be used, for example, gold, as used for charge transfer complex titrations because it is generally less catalytically active than platinum [4,5,11]. The latter method also tends to our simplistic concept of a biological membrane. With DC titrations, problems may occur involving discharge of the charged components of a complex onto the electrodes [11]. This problem is overcome with the use of alternating current (AC) [11]. The use of a capillary shield around one of the electrodes should not affect markedly the plot of AC conductivity versus mol fraction, i.e. a Job plot, providing that the AC frequency is rapid enough. But when charge transfer complexation is relatively slow, a feature of the majority of complexes in biological systems, one may expect readily observable graphical plot changes. In addition, one or more of the reacted partners in a charge transfer complex may interact with the solvent or the electrodes to affect the profile of the Job plot during capillary shielded titration.
2. Methods
2.1. Cell design (Fig. 1) Teflon buttons with different size capillary orifices were press-fitted into a Teflon sleeve (extractor), in turn pressed into the lower portion of a Teflon cup to shield one of the electrodes. The upper portion of the Teflon cup containing the second electrode, and acting as an overflow container, was screwed into the bottom portion of the cup to facilitate removal of the button for cleaning. The gold electrodes were 1 cm square by 1.5 mm thick. Gold leads (1.5 mm diameter) were welded to the squares, and forced through the Teflon walls to obviate leaks. For conservation of solution the scale of the cell should be small. Our cell accommodated an initial volume of 10 ml comfortably. A skilled machinist may reduce the cell size as desired.
2.2. Reagents
Fig. 1. Restricted access (shielded) AC titration cell.
Since a great deal of information is available for the chlorpromazine · HCl (CPZ · HCl)/iodine titration in acetonitrile, the major thrust of our experiments in-
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
255
Fig. 2. Job plot of AC conductivity G versus mol fraction for the chlorpromazine HCl (donor D: I2 (acceptor A) charge transfer titration in acetonitrile for the case of an unshielded electrode.
volved these compounds. Other antimicrobial thiazines [7,12 –15] complexed with the non-polar iodine were also used. These were: phenothiazine (PTZ), promethazine (PMZ · HCl), promazine (PZ · HCl), trifluoperazine (TFP · 2HCl and TFP · HCl). TFP · HCl was prepared by titrating TFP base with standard HCl solution, followed by roto-evaporation to dryness. Some of the TFP · HCl so obtained was titrated by means of the Volhard method [16] with standard AgNO3 to show 100% conversion. All reagents were prepared to 10 − 3 M concentrations in acetonitrile. All compounds were obtained from Sigma, and used without further purification.
2.3. Complexation titrations CPZ · HCl/iodine measurements were effected with a Wayne Kerr B905 bridge at 100 Hz for all the capillary buttons at various rates of solution delivery ranging from 1.13 to 13.77 ml min − 1 with a Sage Instrument 341B automated syringe. Two titrations were done for each set, 10 ml CPZ · HCl was titrated with iodine solution, followed by the reverse titration, i.e. the thiazine was added to, e.g. 10 ml I2. The remaining thiazines were titrated with iodine and vice versa at 100 Hz, with the Wayne Kerr B905 bridge with the capillary button D. All titrations were done at room temperature.
2.4. Cell constants The KCl solution cell constants plotted as a function of capillary hole areas yielded a set of curves confined
by the frequency range (20–3000 Hz) used. The envelope of this family of curves was quite narrow. For the entire range of capillary hole areas the cell constants varied by about 5% over the entire frequency range used. Eight buttons were used provided with capillaries of cross-sectional hole areas ranging from 7.710×10 − 3 to 7.121× 10 − 1 cm2. Experiments with the extractor alone (hole area 7.121× 10 − 1 cm2) yielded the same results as those for the cell without the extractor. Cell constants were determined for aqueous 0.001 M KCl solutions with the Wayne Kerr B905 AC bridge at 100 Hz. For frequency dependence measurements a Wayne Kerr LCR high resistance instrument was used. The frequencies were varied from 20 to 3000 Hz. The cell constant at 100 Hz was 770 m − 1. The optimum experimental parameters for the charge transfer titrations in the cell employed were determined to be: rate of addition 2.4 ml min − 1, a capillary cross-sectional area of 9.79×10 − 2 cm2, and a frequency of 100 Hz for the thiazine/I2 titrations.
3. Results When one compares the Job plots in Fig. 2, the titration of CPZ · HCl versus iodine in the unshielded cell, to that of the titrations in the shielded cell in Fig. 3A, two major differences are noted: 1. The conductivity maximum (GPMF) for the forward titration is larger than that (GPMR) for the reverse titration (Fig. 3A). 2. The plot maxima in Fig. 3A are displaced with respect to the stoichiometric 1:1 result expected
256
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
from Fig. 2. See also Fig. 4 for the titration of phenothiazine against iodine, and vice versa. These phenomena were observed for the shielded titrations of all the thiazines studied. Fig. 3B shows the companion plot to Fig. 3A of capacitance versus mol fraction for the shielded titration of CPZ · HCl against iodine. This plot is similar to the capacitance results obtained for all the thiazine/iodine titrations; all of them are electrically ‘noisy’ as noted in Fig. 3B. The noise may be due to the variable drop-size when the solution is added via the automated syringe, coupled to diffusion through the capillary, or to other factors as
yet undetermined. Since the Wayne Kerr bridge is operating at its limits for capacitance measurements (pF), it is extremely sensitive to any irregularity of solution addition. Note also that the apparent stoichiometry of the forward titration (iodine added to thiazine, Fig. 3A) is closer (0.52) to the 0.5 mol fraction expected of the unshielded case (Fig. 2), while the apparent stoichiometry of the reverse titration (0.42) is much smaller compared to that of the unshielded case. Similar behavior was noted for all the shielded thiazine titrations. This would suggest that the majority charge carrier in these titrations is the thiazine molecule.
Fig. 3. (A) AC titration of chlorpromazine · HCl against iodine. AC conductivity versus mol fraction acceptor. (B) Capacitance versus mol fraction for the titration in (A).
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
257
Fig. 4. AC titration of phenothiazine against iodine. AC conductivity versus iodine mol fraction.
1/DGPM =f+ um 2
4. Discussion It has been demonstrated clearly in previous thiazine/ iodine titrations that the main contribution to charge transfer activity takes place in the active space of the electrodes, i.e. in the electroactive multilayer facing the electrodes [17,18]. The capacitance C and conductance G are observables which measure the permittivities o2 and o1 [18,19] (at the frequency (Hz) of the applied electric field) found in the classical Debye complex equation [19,20]. The real (non-imaginary) part o1 in this complex equation is the experimentally determined permittivity (‘dielectric constant’), and is related to the measured capacitance C [18], here measured between two electrodes, and as such is the sum of two bulk impedances in series [17,18]. o2 is due to Debye relaxation processes, that is, those associated with dipole rotations in the electric field, and also those of Maxwell – Wagner polarization losses, giving rise to a non-zero conductance G [17,18,20]. It has been shown that within the active space of the electrodes [17,18]: o1o0 =o0 + 4pnm 2/3kT
(5)
in which f and u are constants. If our assumption is valid, a plot of 1/DGPM against m 2 should yield a straight line. Fig. 5 shows a plot of 1/DGPM versus m 2 obtained from literature data [18,21,22]. This empirical plot is a straight line passing through the origin. The point not falling on this line is that for promethazine · HCl (PMZ · HCl), the only compound with a branched chain in the thiazine series studied here. However, its dipole is well below that of, say, TFP · HCl, TFP · 2HCl and CPZ · HCl (Fig. 5). We appreciate that the literature data are averages with
(3)
where n is the concentration of relaxing, permanent molecular dipoles (per m3), of moment m, k is Boltzmann’s constant (1.3× 10 − 23 J K − 1), o0 is free permittivity (Fm − 1) and T is the absolute temperature in degrees K. Eq. (3) can be rewritten as: 1/DGPM = (1/2pfDCo2) +2nm 2/3kTfDCo2o0
(4)
where DGPM =GPMF −GPMR (Fig. 3A and Fig. 4), and DC is the difference of the capacitance maxima for the forward and reverse titration, and f is the frequency of the applied AC field. If fDCo2 is assumed to be constant, then we expect:
Fig. 5. Plot of literature dipole moments squared (m 2) for various thiazines versus the inverse conductivity maxima differences (1/ DGPM). See text for explanation of symbols.
258
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
Table 1 Minimal inhibitory concentrations of thiazines for some bacteria and yeasts [7] Microorganisms
1. 2. 3. 4. 5. 6. 7.
Corynebacterium michiganense Bacillus subtilis Staphylococcus aureus Candida albicans Candida utilis Saccharomyces cere6isiae S288C Geotrichum candidum
Thiazine drugs (in order of increasing dipole moment) PMZ · HCl
CPZ · HCl
TFP · 2HCl
50 50 50 50 50 25 50
25 25 25 12 50 25 25
25 25 25 12 25 12 25
errors in the 20–25% range, as indicated by the error bars in Fig. 5. The conductivity measurements vary by less than 5%. DC could not be established with any degree of accuracy due to the noise inherent in the capacitance plots (Fig. 3B). From Fig. 5, the empirical constant was determined to be u= 2.38 × 10 − 7 S − 1 D − 2 for the cell employed. The above errors may conceal f’s likely non-zero but very small value. Eq. (4) can be rewritten in the generally accepted form for conductivity in an Ohmic regime [20]: DGPM = z(nDuD + nAuA) where the subscripts specify donor D and acceptor A, n are the carriers, and z the charge upon them. In our titrations zD = zA generally. DGPM measures essentially the term in parentheses. The mol fraction axis in Figs. 3 and 4 gives the apparent stoichiometries, a function of nD for the forward and nA for the reverse titrations, respectively, thus nD \nA. We also surmise that uD B uA because the thiazine (D) is a larger molecule than iodine (A) and thus experiences greater viscous drag; further, the thiazines possess finite dipole moments, whereas iodine has no dipole moment. These factors affect the diffusion rate of D and A through the capillary in the shielded cases, as well as their rotational behavior in the applied field. It seems possible that restricted access AC titrations may yield information on mobilities directly. Molnar and coworkers, and other workers in the field, have determined the minimal inhibitory concentrations (MIC) for a variety of thiazine drugs acting on several bacterial species (Gram-positive and -negative), as well as on yeast and yeast-like organisms [7]. A selection of this data is given in Table 1 and from this, it is evident that smaller MICs correlate well with higher dipole moments of the thiazine drugs, indicating, even for these few examples, that the amphipathic dipole is an important feature of drug action. One might have expected upon comparing the MICs of CPZ · HCl and TFP · 2HCl that the latter should have
exhibited a lower MIC than the former. However, recent work suggested [23] that the active dipole form in a biological system is most likely TFP · HCl rather than TFP · 2HCl, i.e. an environmental pH dependence needs to be taken into account. The dipole moment of TFP · HCl is almost the same as that of CPZ · HCl (Fig. 5), and thus its MIC effect should be quite similar to that of CPZ · HCl, as observed in Table 1. Our restricted access method of measuring drug dipole moments may have wide application in the area of biological systems, and offers new opportunities for determining drug/microorganism interactions. For instance, a bacterial suspension could be used as a titration medium versus an amphipathic drug, and deviations of the drug dipole moment from the norm should yield information of its interaction with the bacterial species in question. Another approach to drug/bacterial interaction may be to line the surface of the restricted access capillary with an agar coating containing a colony of bacteria. A more general application of this novel method is in the area of excised membrane which can be stretched over the gating orifice to restrict the passage of the drug to the monitoring electrode. Again, deviations of the profile of the experimental plots from the norm should yield better information about drug/membrane interactions than hitherto available. These matters are currently under consideration, as is the refinement of the capacitance measurements. In conclusion, we reiterate that simple restricted access (shielded) electrode AC titration offers a novel and simple method for determining molecular dipole moments of bioactive molecules, and may have application in the field of microbiology.
Acknowledgements We thank Professor Felix Gutmann for helpful discussions. We deeply regret his recent demise.
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
References [1] Pohl HA. Microdielectrophoresis of dividing cells. In: Keyzer H, Gutmann F, editors. Bioelectrochemistry. New York: Plenum, 1980:273 – 95. [2] Pohl HA. Dielectrophoresis. London: Cambridge University Press, 1978. [3] Perez E, Wolfe I. Oestradiol changes the dielectric structure of membranes. Eur Biophys J 1988;16:23–9. [4] Fro¨hlich H. The biological effects of microwaves and related questions. Adv Electron Phys 1980;53:85–152. [5] Fro¨hlich H. Long-range coherence in biological systems. Riv Nuovo Cimento 1977;20:387–94. [6] Del Giudice E, Doglia S, Milani M, Vitiello G. Collective properties of biological systems. In: Gutmann F, Keyzer H, editors. Modern Bioelectrochemistry. New York: Plenum, 1986. [7] Gutmann F, Johnson C, Keyzer H, Molnar J. Charge Transfer Complexes in Biological Systems. New York: Marcel Dekker, 1997. [8] Karolis C. The Dielectric Characterization of Bilayer Lipid Membranes. Thesis. Australia: University of New South Wales, 1993. [9] Kolthoff IM. Relations between voltammetry and potentiometric and amperometric titrations. Anal Chem 1954;26:1685 – 91. [10] Willard HH, Merritt LL, Dean JA, Settle FA. Instrumental Methods of Analysis. Belmont, CA: Wadsworth, 1981. [11] Gutmann F, Keyzer H. Study of phenothiazine and chlorpromazine – iodine complexes. J Chem Phys 1967;46(5):1969– 74. [12] Dastidar SG, Chakrabarty P, Chattopadhyay D, Ray S, Bhattacharya TK, Chakrabarty AN. Assessment of antibacterial activity spectra of chlorpromazine and some related phenothiazines. In: Keyzer H, editor. Thiazines and Structurally Related Compounds. Malabar, FL: Krieger, 1992:203–12. [13] Pearson RD, Hewlett EL. Lethal effects of phenothiazines and related compounds against the pathogenic protozoan Leishmania dono6ani. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:329–36.
.
259
[14] Rodriguez-Cabezas MN, Osuna A. Effects of different phenothiazines in the internalization of Trypanosoma cruzi in the host cell. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:337 – 40. [15] Paglini-Olivia P, Fernandez AR, Lacuara JL. Trypanosoma cruzi: effects of phenothiazines on vitality. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:341 – 6. [16] Vogel AI. Textbook of Quantitative Inorganic Analysis. London: Longman, 1961. [17] Gutmann F, Keyzer H. Capacitance effects of thiazine charge transfer complexes in solution. In: Keyzer H, editor. Thiazines and Structurally Related Compounds. Malabar, FL: Krieger, 1992:123 – 32. [18] Gutmann F, Keyzer H, Ly KT. Solution capacitance behavior of phenothiazine amphiphiles in acetonitrile. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:141–54. [19] Smyth CP. Dielectric Behavior and Structure. New York: McGraw-Hill, 1955. [20] Gutmann F, Keyzer H, Lyons LE. Organic Semiconductors, Part B. Malabar, FL: Krieger, 1983. [21] Parkanyi C, Antonious MS, Aaron J-J, Maafi M, Gil O, Kersebet C, Motohashi N. Ground and excited singlet state dipole moments of new pharmacologically important phenothiazines. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:177 – 88. [22] Parkanyi C, Boniface C, Aaron J-J, Meugulati F, Murray JS, Politzer P, Raghuveer KS. An experimental and theoretical study of phenothiazines and related compounds: dipole moments and other physical properties. In: Keyzer H, editor. Thiazines and Structurally Related Compounds. Malabar, FL: Krieger, 1992:103 – 12. [23] Duncan QS. Dipiperazinyl Phenothiazine Derivatives. Thesis. Los Angeles, CA: California State University, 1993.
Alternating current aspects of antimicrobial molecular complexes H. Keyzer a,*, H.K. Kim b, K.C. Pan a a
Department of Chemistry and Biochemistry, California State Uni6ersity, Los Angeles, 5151 State Uni6ersity Dri6e, Los Angeles, CA 90032, USA b Department of Pharmaceutical Sciences, Uni6ersity of Southern California, 1985 Zonal A6e., Los Angeles, CA 90033, USA
Abstract A novel method was designed involving the titration of alternating current titration in a cell where one electrode was shielded by a capillary enclosure restricting access to it by charge carriers. With this cell, charge transfer complex titration of several thiazines all with some antimicrobial properties was effected in acetonitrile, against the electron accepting molecule iodine. The maxima of the Job plots generated by these titrations exhibited displacement of their positions for the forward and reverse titrations with respect to electron donor–acceptor complexation conductivity and apparent stoichiometry. A plot of inverse conductivity maxima differences against literature-cited dipole moments squared yielded a straight line passing through the origin. The titration plot profiles are discussed in terms of the type, number, and mobility of charge carriers produced in the complexation interaction. This novel method may be used to determine dipole moments of bioactive homologues empirically. There was correlation between thiazine drug dipole moments and minimal inhibitory concentrations of these drugs for some bacterial and yeast species. Several new avenues of investigation of possible relevance to microbiology are suggested. © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. Keywords: Antimicrobials; Thiazines; Psychotropes; Charge transfer complexes; Dipole moments; Alternating current titrations; Capillary shield cell; Microorganisms
1. Introduction Reliable data for dipole moments of bioactive molecules especially those with antimicrobial properties are relatively scarce, yet this parameter is important to their mechanism of action, a fact observed in many dielectrophoresis experiments [1 – 3]. Fro¨hlich considered biological systems as three-dimensional sets of interacting dipoles [4,5]. Del Giudice and coworkers used this information as a basis in the successful quantum field treatment for explaining observable manifestations of energy and electron transfer in the self-organization, dissipativity, and stability of biological assemblies [6]. The interaction of thiazine antimicrobials with bacteria and other microorganisms includes surface effects and penetration of their membranes. Drugs in this category are all surface-active amphipathics, and exert much of their influence by forming charge transfer complexes with membrane components and by altering the stability of membrane assemblies [7]. * Corresponding author. Tel./fax: +1-760-365-2716. E-mail address: [email protected] (H. Keyzer)
The lipid bilayer of a membrane forms a sandwich of distinguishable dielectric layers as determined by Thornton and coworkers [7] for erythrocytes from Raman experiments, and confirmed by Karolis [8] from capacitance and conductance measurements of mixed lipids, as well as reconstituted erythrocyte membranes. The dielectric spatial structure and surface morphology of a cell are important for the control of energy transactions involving dipole moments. A biological membrane can be likened to a parametron semiconductor circuit with a one-to-one correspondence analog to a dipole double layer [7]. A common feature of charge transfer across a biological membrane is its mode of passage through existing or induced restricted orifices which involves dipole interaction [7]. Knowledge of dipole moments of amphipathic drugs is therefore likely to be important as a requisite for understanding their action on and through microbial membranes. Unfortunately, dipole moment data of drug molecules are not only exceedingly scarce, but dipole moment changes of drug/microbe interactions are non-existent.
0924-8579/00/$ 20 © 2000 Elsevier Science B.V. and International Society of Chemotherapy. All rights reserved. PII: S 0 9 2 4 - 8 5 7 9 ( 9 9 ) 0 0 1 6 2 - 4
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
254
We introduce a novel and relatively simple method for the study of charge transfer amongst dipolar bioactive molecules between two electrodes, one of which was shielded by a capillary orifice as the first step to producing a method which may measure directly drug dipole moment changes when it contacts the surface of a microorganism. Such a capillary was intended to mimic, to a first approximation, a simplistic version of a biomembrane pore. Charge diffusion through the capillary will be influenced by the mobility of the charge carriers, their dipole moments, and the size of the capillary. The charge transfer reaction produces charge carriers in high dielectric constant solvents as follows: solvent
D + A lDA l D+A− l D+ +A−
donor
acceptor
(1)
charge carriers
For direct current (DC) titrations involving the pH or potentiometric determinations, one may use a first derivative plot to obtain with greater precision the endpoint of the familiar sigmoid curve generated by the Nernst equation: E= E0 + k log[Ox]/[Red]
(2)
where E is the potential measured, E0 is the oxidation potential of the Redox couple, k is a constant, and [Ox] and [Red] are the oxidant and reductant concentrations, charge donor D and charge acceptor A, respectively. However, rather than creating a numerical plot of the first derivative curve to obtain the equivalence point with greater precision, one may use direct experimental methods to obtain the same result. The first direct method involves the use of two dissimilar electrodes, a platinum/graphite or a platinum/tungsten pair [9,10]. Alloys may also be used, as in the case of the platinum/platinum – 10% rhodium pair [10]. If the titrant is added uniformly at a rapid rate, one electrode lags slightly in its response to the changes in solution composition. As a result the electrode pair will exhibit a maximum difference in re-
sponse at the equivalence point when the logarithmic term of the Nernst equation changes the most rapidly, and a first derivative curve is obtained automatically. A second method employs a different strategy. One electrode of a platinum/platinum pair, is encased in a glass tube provided with a capillary orifice. The concentration in the capillary thus lags slightly behind that of the bulk of the solution during addition of the titrant, again yielding a first derivative plot automatically [10]. The advantage of this method is its simplicity, and that any suitable metal electrodes may be used, for example, gold, as used for charge transfer complex titrations because it is generally less catalytically active than platinum [4,5,11]. The latter method also tends to our simplistic concept of a biological membrane. With DC titrations, problems may occur involving discharge of the charged components of a complex onto the electrodes [11]. This problem is overcome with the use of alternating current (AC) [11]. The use of a capillary shield around one of the electrodes should not affect markedly the plot of AC conductivity versus mol fraction, i.e. a Job plot, providing that the AC frequency is rapid enough. But when charge transfer complexation is relatively slow, a feature of the majority of complexes in biological systems, one may expect readily observable graphical plot changes. In addition, one or more of the reacted partners in a charge transfer complex may interact with the solvent or the electrodes to affect the profile of the Job plot during capillary shielded titration.
2. Methods
2.1. Cell design (Fig. 1) Teflon buttons with different size capillary orifices were press-fitted into a Teflon sleeve (extractor), in turn pressed into the lower portion of a Teflon cup to shield one of the electrodes. The upper portion of the Teflon cup containing the second electrode, and acting as an overflow container, was screwed into the bottom portion of the cup to facilitate removal of the button for cleaning. The gold electrodes were 1 cm square by 1.5 mm thick. Gold leads (1.5 mm diameter) were welded to the squares, and forced through the Teflon walls to obviate leaks. For conservation of solution the scale of the cell should be small. Our cell accommodated an initial volume of 10 ml comfortably. A skilled machinist may reduce the cell size as desired.
2.2. Reagents
Fig. 1. Restricted access (shielded) AC titration cell.
Since a great deal of information is available for the chlorpromazine · HCl (CPZ · HCl)/iodine titration in acetonitrile, the major thrust of our experiments in-
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
255
Fig. 2. Job plot of AC conductivity G versus mol fraction for the chlorpromazine HCl (donor D: I2 (acceptor A) charge transfer titration in acetonitrile for the case of an unshielded electrode.
volved these compounds. Other antimicrobial thiazines [7,12 –15] complexed with the non-polar iodine were also used. These were: phenothiazine (PTZ), promethazine (PMZ · HCl), promazine (PZ · HCl), trifluoperazine (TFP · 2HCl and TFP · HCl). TFP · HCl was prepared by titrating TFP base with standard HCl solution, followed by roto-evaporation to dryness. Some of the TFP · HCl so obtained was titrated by means of the Volhard method [16] with standard AgNO3 to show 100% conversion. All reagents were prepared to 10 − 3 M concentrations in acetonitrile. All compounds were obtained from Sigma, and used without further purification.
2.3. Complexation titrations CPZ · HCl/iodine measurements were effected with a Wayne Kerr B905 bridge at 100 Hz for all the capillary buttons at various rates of solution delivery ranging from 1.13 to 13.77 ml min − 1 with a Sage Instrument 341B automated syringe. Two titrations were done for each set, 10 ml CPZ · HCl was titrated with iodine solution, followed by the reverse titration, i.e. the thiazine was added to, e.g. 10 ml I2. The remaining thiazines were titrated with iodine and vice versa at 100 Hz, with the Wayne Kerr B905 bridge with the capillary button D. All titrations were done at room temperature.
2.4. Cell constants The KCl solution cell constants plotted as a function of capillary hole areas yielded a set of curves confined
by the frequency range (20–3000 Hz) used. The envelope of this family of curves was quite narrow. For the entire range of capillary hole areas the cell constants varied by about 5% over the entire frequency range used. Eight buttons were used provided with capillaries of cross-sectional hole areas ranging from 7.710×10 − 3 to 7.121× 10 − 1 cm2. Experiments with the extractor alone (hole area 7.121× 10 − 1 cm2) yielded the same results as those for the cell without the extractor. Cell constants were determined for aqueous 0.001 M KCl solutions with the Wayne Kerr B905 AC bridge at 100 Hz. For frequency dependence measurements a Wayne Kerr LCR high resistance instrument was used. The frequencies were varied from 20 to 3000 Hz. The cell constant at 100 Hz was 770 m − 1. The optimum experimental parameters for the charge transfer titrations in the cell employed were determined to be: rate of addition 2.4 ml min − 1, a capillary cross-sectional area of 9.79×10 − 2 cm2, and a frequency of 100 Hz for the thiazine/I2 titrations.
3. Results When one compares the Job plots in Fig. 2, the titration of CPZ · HCl versus iodine in the unshielded cell, to that of the titrations in the shielded cell in Fig. 3A, two major differences are noted: 1. The conductivity maximum (GPMF) for the forward titration is larger than that (GPMR) for the reverse titration (Fig. 3A). 2. The plot maxima in Fig. 3A are displaced with respect to the stoichiometric 1:1 result expected
256
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
from Fig. 2. See also Fig. 4 for the titration of phenothiazine against iodine, and vice versa. These phenomena were observed for the shielded titrations of all the thiazines studied. Fig. 3B shows the companion plot to Fig. 3A of capacitance versus mol fraction for the shielded titration of CPZ · HCl against iodine. This plot is similar to the capacitance results obtained for all the thiazine/iodine titrations; all of them are electrically ‘noisy’ as noted in Fig. 3B. The noise may be due to the variable drop-size when the solution is added via the automated syringe, coupled to diffusion through the capillary, or to other factors as
yet undetermined. Since the Wayne Kerr bridge is operating at its limits for capacitance measurements (pF), it is extremely sensitive to any irregularity of solution addition. Note also that the apparent stoichiometry of the forward titration (iodine added to thiazine, Fig. 3A) is closer (0.52) to the 0.5 mol fraction expected of the unshielded case (Fig. 2), while the apparent stoichiometry of the reverse titration (0.42) is much smaller compared to that of the unshielded case. Similar behavior was noted for all the shielded thiazine titrations. This would suggest that the majority charge carrier in these titrations is the thiazine molecule.
Fig. 3. (A) AC titration of chlorpromazine · HCl against iodine. AC conductivity versus mol fraction acceptor. (B) Capacitance versus mol fraction for the titration in (A).
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
257
Fig. 4. AC titration of phenothiazine against iodine. AC conductivity versus iodine mol fraction.
1/DGPM =f+ um 2
4. Discussion It has been demonstrated clearly in previous thiazine/ iodine titrations that the main contribution to charge transfer activity takes place in the active space of the electrodes, i.e. in the electroactive multilayer facing the electrodes [17,18]. The capacitance C and conductance G are observables which measure the permittivities o2 and o1 [18,19] (at the frequency (Hz) of the applied electric field) found in the classical Debye complex equation [19,20]. The real (non-imaginary) part o1 in this complex equation is the experimentally determined permittivity (‘dielectric constant’), and is related to the measured capacitance C [18], here measured between two electrodes, and as such is the sum of two bulk impedances in series [17,18]. o2 is due to Debye relaxation processes, that is, those associated with dipole rotations in the electric field, and also those of Maxwell – Wagner polarization losses, giving rise to a non-zero conductance G [17,18,20]. It has been shown that within the active space of the electrodes [17,18]: o1o0 =o0 + 4pnm 2/3kT
(5)
in which f and u are constants. If our assumption is valid, a plot of 1/DGPM against m 2 should yield a straight line. Fig. 5 shows a plot of 1/DGPM versus m 2 obtained from literature data [18,21,22]. This empirical plot is a straight line passing through the origin. The point not falling on this line is that for promethazine · HCl (PMZ · HCl), the only compound with a branched chain in the thiazine series studied here. However, its dipole is well below that of, say, TFP · HCl, TFP · 2HCl and CPZ · HCl (Fig. 5). We appreciate that the literature data are averages with
(3)
where n is the concentration of relaxing, permanent molecular dipoles (per m3), of moment m, k is Boltzmann’s constant (1.3× 10 − 23 J K − 1), o0 is free permittivity (Fm − 1) and T is the absolute temperature in degrees K. Eq. (3) can be rewritten as: 1/DGPM = (1/2pfDCo2) +2nm 2/3kTfDCo2o0
(4)
where DGPM =GPMF −GPMR (Fig. 3A and Fig. 4), and DC is the difference of the capacitance maxima for the forward and reverse titration, and f is the frequency of the applied AC field. If fDCo2 is assumed to be constant, then we expect:
Fig. 5. Plot of literature dipole moments squared (m 2) for various thiazines versus the inverse conductivity maxima differences (1/ DGPM). See text for explanation of symbols.
258
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
Table 1 Minimal inhibitory concentrations of thiazines for some bacteria and yeasts [7] Microorganisms
1. 2. 3. 4. 5. 6. 7.
Corynebacterium michiganense Bacillus subtilis Staphylococcus aureus Candida albicans Candida utilis Saccharomyces cere6isiae S288C Geotrichum candidum
Thiazine drugs (in order of increasing dipole moment) PMZ · HCl
CPZ · HCl
TFP · 2HCl
50 50 50 50 50 25 50
25 25 25 12 50 25 25
25 25 25 12 25 12 25
errors in the 20–25% range, as indicated by the error bars in Fig. 5. The conductivity measurements vary by less than 5%. DC could not be established with any degree of accuracy due to the noise inherent in the capacitance plots (Fig. 3B). From Fig. 5, the empirical constant was determined to be u= 2.38 × 10 − 7 S − 1 D − 2 for the cell employed. The above errors may conceal f’s likely non-zero but very small value. Eq. (4) can be rewritten in the generally accepted form for conductivity in an Ohmic regime [20]: DGPM = z(nDuD + nAuA) where the subscripts specify donor D and acceptor A, n are the carriers, and z the charge upon them. In our titrations zD = zA generally. DGPM measures essentially the term in parentheses. The mol fraction axis in Figs. 3 and 4 gives the apparent stoichiometries, a function of nD for the forward and nA for the reverse titrations, respectively, thus nD \nA. We also surmise that uD B uA because the thiazine (D) is a larger molecule than iodine (A) and thus experiences greater viscous drag; further, the thiazines possess finite dipole moments, whereas iodine has no dipole moment. These factors affect the diffusion rate of D and A through the capillary in the shielded cases, as well as their rotational behavior in the applied field. It seems possible that restricted access AC titrations may yield information on mobilities directly. Molnar and coworkers, and other workers in the field, have determined the minimal inhibitory concentrations (MIC) for a variety of thiazine drugs acting on several bacterial species (Gram-positive and -negative), as well as on yeast and yeast-like organisms [7]. A selection of this data is given in Table 1 and from this, it is evident that smaller MICs correlate well with higher dipole moments of the thiazine drugs, indicating, even for these few examples, that the amphipathic dipole is an important feature of drug action. One might have expected upon comparing the MICs of CPZ · HCl and TFP · 2HCl that the latter should have
exhibited a lower MIC than the former. However, recent work suggested [23] that the active dipole form in a biological system is most likely TFP · HCl rather than TFP · 2HCl, i.e. an environmental pH dependence needs to be taken into account. The dipole moment of TFP · HCl is almost the same as that of CPZ · HCl (Fig. 5), and thus its MIC effect should be quite similar to that of CPZ · HCl, as observed in Table 1. Our restricted access method of measuring drug dipole moments may have wide application in the area of biological systems, and offers new opportunities for determining drug/microorganism interactions. For instance, a bacterial suspension could be used as a titration medium versus an amphipathic drug, and deviations of the drug dipole moment from the norm should yield information of its interaction with the bacterial species in question. Another approach to drug/bacterial interaction may be to line the surface of the restricted access capillary with an agar coating containing a colony of bacteria. A more general application of this novel method is in the area of excised membrane which can be stretched over the gating orifice to restrict the passage of the drug to the monitoring electrode. Again, deviations of the profile of the experimental plots from the norm should yield better information about drug/membrane interactions than hitherto available. These matters are currently under consideration, as is the refinement of the capacitance measurements. In conclusion, we reiterate that simple restricted access (shielded) electrode AC titration offers a novel and simple method for determining molecular dipole moments of bioactive molecules, and may have application in the field of microbiology.
Acknowledgements We thank Professor Felix Gutmann for helpful discussions. We deeply regret his recent demise.
H. Keyzer et al. / International Journal of Antimicrobial Agents 14 (2000) 253–259
References [1] Pohl HA. Microdielectrophoresis of dividing cells. In: Keyzer H, Gutmann F, editors. Bioelectrochemistry. New York: Plenum, 1980:273 – 95. [2] Pohl HA. Dielectrophoresis. London: Cambridge University Press, 1978. [3] Perez E, Wolfe I. Oestradiol changes the dielectric structure of membranes. Eur Biophys J 1988;16:23–9. [4] Fro¨hlich H. The biological effects of microwaves and related questions. Adv Electron Phys 1980;53:85–152. [5] Fro¨hlich H. Long-range coherence in biological systems. Riv Nuovo Cimento 1977;20:387–94. [6] Del Giudice E, Doglia S, Milani M, Vitiello G. Collective properties of biological systems. In: Gutmann F, Keyzer H, editors. Modern Bioelectrochemistry. New York: Plenum, 1986. [7] Gutmann F, Johnson C, Keyzer H, Molnar J. Charge Transfer Complexes in Biological Systems. New York: Marcel Dekker, 1997. [8] Karolis C. The Dielectric Characterization of Bilayer Lipid Membranes. Thesis. Australia: University of New South Wales, 1993. [9] Kolthoff IM. Relations between voltammetry and potentiometric and amperometric titrations. Anal Chem 1954;26:1685 – 91. [10] Willard HH, Merritt LL, Dean JA, Settle FA. Instrumental Methods of Analysis. Belmont, CA: Wadsworth, 1981. [11] Gutmann F, Keyzer H. Study of phenothiazine and chlorpromazine – iodine complexes. J Chem Phys 1967;46(5):1969– 74. [12] Dastidar SG, Chakrabarty P, Chattopadhyay D, Ray S, Bhattacharya TK, Chakrabarty AN. Assessment of antibacterial activity spectra of chlorpromazine and some related phenothiazines. In: Keyzer H, editor. Thiazines and Structurally Related Compounds. Malabar, FL: Krieger, 1992:203–12. [13] Pearson RD, Hewlett EL. Lethal effects of phenothiazines and related compounds against the pathogenic protozoan Leishmania dono6ani. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:329–36.
.
259
[14] Rodriguez-Cabezas MN, Osuna A. Effects of different phenothiazines in the internalization of Trypanosoma cruzi in the host cell. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:337 – 40. [15] Paglini-Olivia P, Fernandez AR, Lacuara JL. Trypanosoma cruzi: effects of phenothiazines on vitality. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:341 – 6. [16] Vogel AI. Textbook of Quantitative Inorganic Analysis. London: Longman, 1961. [17] Gutmann F, Keyzer H. Capacitance effects of thiazine charge transfer complexes in solution. In: Keyzer H, editor. Thiazines and Structurally Related Compounds. Malabar, FL: Krieger, 1992:123 – 32. [18] Gutmann F, Keyzer H, Ly KT. Solution capacitance behavior of phenothiazine amphiphiles in acetonitrile. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:141–54. [19] Smyth CP. Dielectric Behavior and Structure. New York: McGraw-Hill, 1955. [20] Gutmann F, Keyzer H, Lyons LE. Organic Semiconductors, Part B. Malabar, FL: Krieger, 1983. [21] Parkanyi C, Antonious MS, Aaron J-J, Maafi M, Gil O, Kersebet C, Motohashi N. Ground and excited singlet state dipole moments of new pharmacologically important phenothiazines. In: Barbe J, Keyzer H, Soyfer JC, editors. Biological and Chemical Aspects of Thiazines and Analogs. San Gabriel, CA: Enlight Associates, 1995:177 – 88. [22] Parkanyi C, Boniface C, Aaron J-J, Meugulati F, Murray JS, Politzer P, Raghuveer KS. An experimental and theoretical study of phenothiazines and related compounds: dipole moments and other physical properties. In: Keyzer H, editor. Thiazines and Structurally Related Compounds. Malabar, FL: Krieger, 1992:103 – 12. [23] Duncan QS. Dipiperazinyl Phenothiazine Derivatives. Thesis. Los Angeles, CA: California State University, 1993.