Technical paper
Spectrophotometric Determination of Acidity Constants of Some Macrolides in Acetonitrile-Water Binary Mixtures
Senem Sanli,1'* Nurullah Sanli1 and Guleren Alsancak2
1 Department of Chemistry, Faculty of Arts & Sciences, Hitit University, 19040, Corum, TURKEY
2 Suleyman Demirel University, Faculty of Arts & Sciences, Department of Chemistry, 32260, Isparta, TURKEY
* Corresponding author: E-mail: senemsanli@hitit.edu.tr Tel: +90-364-2277000-1641; Fax:+90-364-2277006
Received: 07-05-2010
Abstract
The acidity constants of eight macrolides (erythromycin, roxithromycin, oleandomycin, azithromycin, josamycin, tylo-sin tartrate, tilmicosin and spiramycin) have been determined in acetonitrile-water binary mixtures (30%, 40% and 50% (v/v)) by spectrophotometric method. The pKa's available in literature determined by various methods are compiled in comparison with the value of this work. These results are expected to essentially facilitate the research on occurrence, fate and effects, analysis method development, and control of antibiotics in various treatment occurrences.
Keywords: Macrolides; Acidity constants; Acetonitrile-water mixtures; Spectroscopic method.
1. Introduction
Macrolide antibiotics constitute a very important class of antibacterial compounds which have been widely used in veterinary and medicinal practices to treat a wide range of diseases. Macrolides are characterized by a ma-crocyclic lactone ring containing 14, 15 or 16 atoms with sugars linked via glycosidic bonds. The macrolides with 16 atoms in the lactone ring represent the most commonly used macrolides in veterinary medicine and examples of these include tylosin, josamycin and spiramycin. Tilmicosin, is a derivative of desmycosin (16-membered-ring macrolide). Erythromycin, roxithromycin and olean-domycin are another example of a macrolide antibiotic; they contain 14 atoms. Azithromycin is one of the world's best-selling antibiotics. It is derived from eryt-hromycin, but it differs chemically from erythromycin in that a methyl-substituted nitrogen atom is incorporated into the lactone ring, thus making the lactone ring 15-membered. Like erythromycin, it appears to bind to the same receptor, 50 S ribosomal subunits of susceptible bacteria and suppresses protein synthesis. A macrolide contains basic dimethylamine [-N(CH3)2] group which is able to gain a proton.1
Acid dissociation constant (pKa value) is an important parameter to estimate the extent of ionization of drug molecules at different pH values, which is of fundamental importance in the consideration of their interaction with biological membranes.2'3 Many drug compounds are sparingly soluble in water and a precise determination of their pKa values poses a challenging problem for potentio-metric titration since the accuracy of this method is restricted by its detection limit of about 10-4 M. Spectrosco-pic titration has been utilized as an alternative to determine pKa values of substances with large molar absorptivi-ties because of its high sensitivity at concentrations of substance as low as 10-6 M.4 However, the compound under investigation must possess chromophore(s) in proximity to the ionization center(s) so that the protonated and deprotonated species exhibit sufficient spectral dissimilarity.
There are some limitations in determining the acidity constants of the molecules such as low solubility in aqueous solutions and the low values of acidity constants. Therefore, to enhance the acidity constants on one hand and to increase the solubility on the other, we forced to choose mixed solvents.56 In addition, the pKa of the analy-tes in the mixed solvent is an important parameter for the optimization of separation techniques such as capillary
electrophoresis and liquid chromatography of ionisable compounds. An understanding of the factors that determine the acidity or basicity of a solute solved in solvent mixtures, can be decisive to foresee how the aqueous pKa is modified when the organic modifier is added to the aqueous part of the eluent for separation methods and in order to establish relationships allowing to relate an aqueous pKa with the pKa in the organic modifier-water binary mixtures. Moreover, the study of acid-base behavior of analytes in binary organic-water solvent systems could be key in predicting influence of pH on retention and selectivity in LC.7,8
Acetonitrile (MeCN) and its mixtures with water are widely used in separation techniques, due to the excellent characteristics of the pure solvent. MeCN is a very weak base and a very weak acid and therefore it is a good differentiating solvent for both acids and bases. Furthermore, it has low viscosity and ideal good UV transparency. It has a relatively high dielectric constant (e = 36) and a small au-toprotolysis constant (pKs = 33.6). Although MeCN-water mobile phases have been used in RP-LC separation procedures, the pKa values of macrolides have not been determined in binary mixtures. Accordingly, MeCN acts as a strongly differentiating solvent with a modest solvating power for many polar ionic solutes.9
There are lots of pKa determination techniques for compounds of pharmaceutical or biological interest such as potentiometric titrations,1,10 UV-Vis,9,11-13 LC,10,11,14,15
and CE methodology.16-17 Among these techniques, the potentiometric titration is a more general method and it does not require the presence of chromophore groups for pKa determination. If due care is taken, this technique is accurate and has good reproducibility. Fast and automated instruments are commercially available for potentiometry; however, their disadvantages include the requirements to use milligram amount of pure substances and the use of carbonate free titrants. Also, UV-Visible absorption spectrometry has still been used widely by the help of improved computer programs for the determination of dissociation constants. Moreover, the use of computer programs for the refinement of dissociation constants allows the different pKa values in polyprotic substances to be determined, even when they are very close.14,18 In the last decade, a new procedure has been developed when LC and CE methodologies are used for determination of ionization constants in combination with a diode array detector (DAD) for absorbance measurements.11,19,20
In this study, pKa values of eight macrolides (erythromycin, roxithromycin, spiramycin, oleandomycin, azithromcin, tylosin tartrate, josamycin, tilmicosin; Table!) were determined in 30%, 40%, 50% (v/v) MeCN-wa-ter binary mixtures by UV-spectroscopy method. The obtained pKa values have been compared with those predicted SPARC on-line calculator.21,22 This program estimates several physico-chemical properties for organic compounds on the basis of their molecular structure.
Table 1. Chemical structure of studied compounds
Compounds name	Structure
Compounds name
Structure
Tylosin
Azitromycin
Roxithromycin
Oleandomycin
Spiramycin
2. Experimental
2. 1. Chemical and Reagents
Analytical reagent grade chemicals were used, unless otherwise indicated. Erythromycin, roxithromycin, tilmico-sin, oleandomycin, josamycin, spiramycin, tylosin tartrate, azithromycin (Sigma), acetonitrile (HPLC grade) were used without further purification. Stock solutions of macro-lides were prepared in MeCN. Potassium hydroxide (Titri-sol), hydrochloride acid (Titrisol), potassium hydrogen phthalate and potassium chloride (ionic strength adjuster; 0.1 mol L-1) were supplied from Merck. All stock solutions of hydrochloride acid, potassium chloride and potassium hydrogen phthalate were prepared by water. Water, with conductivity lower than 0.05 ^Scm-1 was obtained with a Milli-Q water purification system (Milli Pore Corp.).
2. 2. Apparatus
Potentiometric measurements were performed with Mettler-Toledo MA 235 pH/ion (resolution ±0.1 mV) analyzer system. All titrations were carried out under N2 and at 25.0 ± 0.1 °C, which was maintained by circulating water from a constant-temperature thermostat (Heto CBN 8-30 and temperature control unit Heto HMT 200) through the double-wall Pyrex titration cell of 80-mL capacity.
The UV-Vis absorbance spectra were recorded at each pH using Perkin Elmer LAMBDA 25 spectrophotometer, equipped with 1 cm path length cell, controlled by personal computer. A peristaltic pump equipped with the spectrometer was used to circulate the solution from the titration vessel to the spectrophotometer cell, and vice versa, through Teflon or Tygon tubes in a closed loop circuit with continuous flow.
titre and absence of carbonate were periodically checked by pH-metry, using the appropriate Gran function against primary standard oven-dried potassium hydrogen phthalate.23
The pKa values of the macrolides were determined by means of the data obtained from spectrometric titrations in 30%, 40% and 50% (v/v) acetonitrile-water mixtures at 25.0 ± 0.1°C and in 0.1 mol L-1 ionic strength (KCl). The spectrophotometry multiple-wavelength pH-titration was carried out as follows: in a first step, the standard emf values, Eo, of the potentiometric cell were evaluated from titrations of a measured amount of an acidic solution, at the same conditions of temperature, ionic strength and solvent composition to be used in later experiments using KOH solutions in the same solvent and ionic strength as the titrant, and checking the calibration parameters from the Gran plots.23,24 The standard emf of the cell, Eo, is the average of at least 5 standardizations. The standardization of the electrode system was carried out, each time solvent media was changed and the constancy of Eo, values ensured by continual surveillance by means of periodic calibrations.
In a second step, a solution of fully protonated ma-crolides (50.0 mL containing 5.10-5 mol L-1 drug) by HCl at the required conditions of temperature, ionic strength and solvent composition were titrated using KOH solutions in the same solvent and ionic strength in the pH range of 2.5-11.0. After each addition, the potential was allowed to stabilize, its value used, in combination with Eo calculated in calibration step, to calculate the pH solution.
In the UV-Vis spectrometric titrations, the test solution was pumped to a spectrometric flow-cell by means of a peristaltic pump. After each addition of titrant, and after waiting for the emf reading to be stable, a spectrum, UV-Vis spectra were recorded with 1 nm resolution at 190-350 nm intervals in order to obtain different spectra around the maximum X for each macrolides.
2. 3. Procedure
Before the spectrometric titration, carbonate-free potassium hydroxide solutions were prepared under a nitrogen atmosphere. The ionic strength of KOH solution was adjusted to 0.10 mol L-1 by the addition of KCl. The alkali
2. 4. Data Treatment
Spectrometric titrations data were processed using the program STAR (Stability Constants by Absorbance Readings) which calculates stability constant and molar absorptivities of the pure species by multilinear regres-
§anli et al.: Spectrophotometric Determination of Acidity Constants ...
UV-Vis spectra and sigmoid curve of pH-absorbance graphic of erythromycin in 50% MeCN-water mixtu-
sion.25 The program STAR requires a previous model of the chemical equilibria, based upon the existence of certain chemical species, to be postulated. The refinement of equilibrium constants is done using the Gauss-Newton non-linear least squares algorithm by numerical differentiation, until a minimum in the sum of the squares residual (U) is attained. Wavelength (Uabs) is obtained:
ns nw
Uabs=SZ(Ai.j.exp.-Ai,j,ca.c)2	(1)
i=l j=l
where ns and nw indicate the number of spectra and wavelengths, respectively. A;j exp and A;j calc are the experimental and calculated absorbance values for the wavelength j in the spectrum i. The calculated absorbances are obtained in three steps: the program first solves the mass balances for each spectrum according to the guessed equilibrium constants and experimental conditions; then, a multiple linear regression procedure is applied in order to determine the molar absorbances of each unknown species, and finally the individual absorbance values are re-calculated from the guessed species concentration and the corresponding molar absorbances. The optimization is performed by means of a non-linear least-squares procedure. The minimization process is repeated until the relative change of U between two iterations is < 0.01%.26
3. Results and Discussion
Macrolide antibiotics contain a basic dimethylamine [-N(CH3)2] group (Table. 1), which is able to gain a pro-
ton, so, according to their chemical structure, macrolides have usually one p^a value around 9.1
The p^a value of macrolides are either not known accurately or not available at all. Only limited number of study related with pKa values of studied macrolides are found in the literature. However some data in literature differ considerably, number of pKa values and the methods use for their determination are not adequately describe. Table 2 summarizes the literature data and SPARC, available on the p^'s of the some macrolides.1,2735 The Merck Index reports a p^a of 8.8 for erythromycin, but there is no indication of the method, solvent system, temperature, or ionic strength employed.27 These parameters are critical because they influence the resulting values. Thus, the pKa values in Table 2 agree in some instances with values reported here, but in other cases are quite different because of the different methods and different solvents employed. In the present work, we have used a single method, temperature, ionic strength and solvent system, so our data are self consistent and suitable for our purpose.
UV-Vis spectra and sigmoid curve of pH-ab-sorbance graphic of erythromycin in 50% MeCN-water mixtures at different pH values were given in Fig. 1 and Fig. 2, respectively. The dissociation constant values determined by STAR program, for the erythromycin, roxit-hromycin, tilmicosin, oleandomycin, josamycin, spiramycin studied in 30%, 40%, 50% (v\v) MeCN-water mixtures at 25.0 ± 0.1 °C are collected in Table 3, together with respective standard deviations. These compounds have limited water solubility and need to be dissolved in a co-solvent. Erythromycin, roxithromycin, oleandomycin, josamycin contain only one amine group.
Table 2. Literature and SPARC values for the eight macrolides' considered in the article.
Macrolidep	Ka	Solvent	Method	Temp.	SPARC
Erythromycin	8.901 8.8027 8.60-8.928 9.1028 8.8029 8.6030 8.8832	aqueous not given aqueous 90% aqueous D2O d2o 50% aqueous DMF 65% aqueous MeCN	potentiometric not given titration NMR NMR not given potentiometric	ambient not given not given 30 0C not given not given 25 oC	7.24
Roxithromycin	9.171	Aqueous	potentiometric	ambient	6.64
Azithromycin	8.8528 8.74-9.4532	Aqueous 65% aqueous MeCN	not given potentiometric	not given 25 oC	7.44
Oleandomycin	8.5031 8.8432	50% aqueous ethanol 65% aqueous MeCN	not given potentiometric	not given 25 oC	7.20
Tylosin Tartrate	3.31; 7.501 7.7332	Aqueous 65% aqueous MeCN	potentiometric potentiometric	ambient 25 oC	7.09
Spiramycin	7.9033	not given	not given	not given	7.03 8.30
Tilmicosin	8.18-9.5632	65% aqueous MeCN	potentiometric	25 oC	6.47 8.64
Josamycin	7.1034 7.3035	50% aqueous ethanol 50% aqueous MeOH	not given voltammetric	not given 90 oC	6.98
pH
Fig. 1. The wavelength (nm)-absorbance graphic for erythromycin	Fig. 2. Sigmoid curve of pH-A graphic for erythromycin in 50%
in 50% MeCN-water binary mixture.	(v/v) MeCN-water binary mixture at 202 nm.
The amine group is able to gain proton under specific pH conditions. Azithromycin, tilmicosin and spiramycin have two nitrogen and they have two p^a values. But for azithromycin, just one pKa could be determined. Tylosin is produced as a tylosin tartrate. Tartaric acid has two pKa values, i.e., 3.03 and 4.37.4 Thus, the pK: of tylosin tartrate may correspond to the tartrate moiety. The pK2 of tylosin tartrate is believed to correspond to the dimethy-lamine group.
It is known that one of the most important factors determining the equilibrium constants is the reaction medium. The variation of the p^a values of macrolides versus the mole fraction of MeCN, ^MeCN, in MeCN-wa-ter mixtures is presented in Fig. 3. The variation of p^a values with the mole fraction of MeCN is different for each substance although, in general, pKa values decrease with the mole fraction of MeCN. The different ways in which pK values change might be explained by the fact
Table 3. The pKa values determined in this study at 0.10 mol L 1 ionic strength and 25 °C.
Compounds	30 % (v/v)	MeCN	40 % (v/v)	MeCN	50 % (v/v)	MeCN
	pKx	pK2	pKx	pK2	pKx	pK2
Erythromycin	8.625(0.03)*	-	8.54(0.02)	-	8.45(0.03)	-
Roxithromycin	8.55(0.04)	-	8.48(0.08)	-	8.40(0.05)	-
Oleandomycin	8.65(0.03)	-	8.56(0.07)	-	8.46(0.08)	-
Josamycin	7.21(0.01)	-	7.08(0.03)	-	6.94 0.04)	-
Spiramycin	7.26(0.10)	8.39(0.09)	7.13(0.08)	8.20(0.06)	7.02(0.07)	8.00(0.07)
Azithromycin	8.38(0.08)	-	8.31(0.06)	-	8.21(0.07)	-
Tylosin tartrate	3.41(0.09)	7.74(0.06)	3.51(0.09)	7.67(0.07)	3.63(0.09)	7.61(0.06)
Tilmicosin	6.25(0.09)	8.33(0.07)	6.22(0.09)	8.20(0.10)	6.19(0.08)	8.08(0.07)
8 The experiments are repeated at least three times * The values between parentheses are the standard deviations
Table 4. The linear equation between experimental pKa values and the mole fraction of the organic modifier
Compounds		Equation	Regression Coefficient
Erythromycin	y =	-1.30 xa (0.06)* + 8.79 (0.01)	R = 0.999
Roxithromycin	y =	-1.15 x (0.05) + 8.70 (0.01)	R = 0.999
Oleandomycin	y =	-1.46 x (0.07) + 8.84 (0.01)	R = 0.999
Josamycin	y =	-2.07 x (0.12) + 7.47 (0.02)	R = 0.998
Spiramycin (p^2)	y =	-2.99 x (0.19) + 8.77 (0.04)	R = 0.998
Azithromycin	y =	-1.31 x (0.03) + 8.55 (0.01)	R = 0.999
Tylosin tartrate (pK2)	y =	-0.99 x (0.12) + 7.86 (0.03)	R = 0.992
Tilmicosin (pK2)	y =	-1.91 x (0.20) + 8.57 (0.04)	R = 0.995
a represents the mole fraction of organic modifier (MeCN). * The values between parentheses are the standard deviations
9
6.5 -,-,-,-
0.1	0.15	0.2	0.25	0.3
XMeCN
Fig. 3. pKa values versus mole fraction of MeCN. (□) Spiramycin (pK2); (A)Josamycin; (•) Tilmicosin (pK2), (O)Erythromycin, (A)Oleandomycin; (^Roxithromycin; (+)Azithromycin; (x)Tylo-sin tartrate (pK2).
that the dissociation process is ruled by electrostatic interaction as well as by specific solute-solvent interactions.
The linear relationships between experimental pKa values for these analytes and the mole fraction of MeCN are given in Table 4.
4. Conclusions
This paper presents the first study dealing with the determination of pKa values of macrolides by spectrosco-pic methods in MeCN-water binary mixtures and gives a possibility for deeper analysis of the processes which take place during LC analysis of macrolides. The determined pKa values are appropriate for predicting the effect of eluent pH on retention and hence for optimization of separation methods based on especially reversed phase liquid chromatography. Also, the important data extracted from this exploration can be used for other pharmacokinetic, pharmacological or technological studies concerning these compounds.
5. Acknowledgment
We gratefully acknowledge to Dr. Jose Luis Beltran from University of Barcelona for supporting the program STAR.
6. References
1.	Z. Qiang, C. Adams, Water Research, 2004, 38, 2874-2890.
2.	J. J. Kautman, N.M. Semo, W. S. Koski, J. Med. Chem., 1975, 18, 647-655.
3.	H. L. J. Fleuren, C. A. M. van Ginneken, J. M. van Rossum, J. Pharm. Sci. 1979, 68, 1056-1058.
4.	A. Albert, E.P. Serjeant, The Determination of ionization Constants, Chapman and Hall: London, 1984.
5.	D. Almasifar, F. Forghaniha, Z. Khojasteh, J. Ghasemi, H. Shargi, M. Shamsipur, J. Chem. Eng. Data, 1997, 42, 12121215.
6.	B. Hemmateenejad, A. Abbaspour, H. Maghami, A. Forou-madi, Anal. Chim. Acta, 2008, 607, 142-152.
7.	V. Evagelou, A. Tsantili-Kakoulidou, M. Koupparis, J. Pharm. Biomed. Anal., 2003, 31, 1119-1128.
8.	E. Jimenez-Lozano, I. Marques, D. Barron, J. L. Beltran, J. Barbosa, Anal. Chim. Acta, 2002, 464, 37-45.
9.	A. H. Akta§, N. Sanli, G. Pekcan, Acta Chim. Slov., 2006, 53, 214-218.
10.	S. Sanli, S. Sanli, G. Alsancak, J. Braz. Chem. Soc., 2009, 20, 133-140.
11.	J. L. Beltran, N. Sanli, G. Fonrodona, D. Barron, G. Özkan, J. Barbosa, Anal. Chim. Acta, 2003, 484, 253-264.
12.	B. Hemmateenejad, L. Emami, H. Sharghi, Spectrochim. Acta Part A, 2010, 75, 340-346.
13.	S. Sanli, Y. Altun, N. Sanli, G. Alsancak, J.L. Beltran, J. Chem. Eng. Data, 2009, 54, 3014-3021.
14.	F. Z. Erdemgil, S. Sanli, N. Sanli, G. Ozkan, J. Barbosa, J. Guiteras, J. L. Beltran, Talanta, 2007, 72, 489-496.
15.	E. g. Demiralay, G. Alsancak, A. S. Özkan, J. Sep. Sci. 2009, 32, 2928-2936.
16.	S. M. C. Buckenmaier, D. V. McCalley, M. R. Euerby, J. Chromatogr. A, 2003, 1004, 71-79.
17.	G. A. Caliaro, C. A. Herbots, J. Pharm. Biomed. Anal., 2001, 26, 427-434.
18.	J. Barbosa, D. Baron, J. L. Beltran, V. Sanz-Nebot, Anal. Chim. Acta, 1995, 317, 75-81.
19.	M. Perez-Urquiza, J. L. Beltran, J. Chromatogr. A, 2001, 917, 331-336.
20.	E. Jimenez-Lozano, I. Marques, D. Barron, J. L. Beltran, J. Barbosa, Anal. Chim. Acta, 2002, 464, 37-45.
21.	Sparc On-Line Calculator, http://ibmlc2.chem.uga.edu/ sparc, (accessed: October 2009).
22.	S. H. Hilal, S. W. Karicckhof, L. A. Carreira, QSAR Comb. Sci., 2004, 23, 709-720.
23.	Y. Marcus, J. Chem. Soc., Faraday Trans., 1989, 85, 381388.
24.	G. Gran, Analyst, 1952, 7, 661-671.
25.	J. L. Beltran, R. Codony, M. D. Prat, Anal. Chim. Acta, 1993, 276, 441-454.
26.	J. Barbosa, D. Barron, E. Jimenez-Lozano, V. Sanz-Nebot, Anal. Chim. Acta, 2001, 437, 309-321.
27.	S. Budavari (Ed.), The Merck Index, 11th ed., Merck and Co.: Rahway, NJ, 1989, pp 577-578 (No. 3626).
28.	R. C. Goldman, S. W. Fesik, C. C. Doran, Antimicrob. Agents Chemother., 1990, 34, 426-431.
29.	J. Gharbi-Benarous, M. Delaforge, C. K. Jankowski, J. P. Gi-rault, J. Med. Chem., 1991, 34, 1117-1125.
30.	G. Kobrehel, Z. Tamburasev, S. Djokic, J. Chromatogr., 1977, 133, 415-419.
31.	H. Els, W. D. Celmer, K. Murai, J. Am. Chem. Soc, 1958, 80, 3777-3782.
32.	J. W. McFarland, C. M. Berger, S.A. Froshauer, S. F. Hayas-hi, S. J. Hecker, B. H Jaynes, M. R. Jefson, B. J. Kamicker, C. A. Lipinski, K. M. Lundy, C. P. Reese, and C. B. Vu, J. Med. Chem. 40 1997, 1340-1346.
33.	R. Wise, Drug Invest., 1993, 6, 29-34.
34.	Budavari, S.; (Ed.), The Merck Index, 12th ed., Merck and Co, NJ, 1996 pp. 898.
35.	F. Belal, A. Al-Majed, K. E. E. Ibrahim, N. Y. Khalil, J. Pharm. and Biomed. Anal., 2002, 30, 705-713.
Povzetek
V treh binarnih mešanicah vode in acetonitrila (30 %, 40 % and 50 % (v/v) smo s spektroskopsko metodo določili konstante kislin osmim makrolidom (eritromicin, roksitromicin, oleandomicin, azitromicin, josamicin, tilosin tartrat, tilmo-cisin, spiramicin).
Določene vrednosti pKa se dobro ujemajo s podatki, dostopnimi v literaturi, dobljenimi z različnimi metodami.