Annalsiii

Countercurrent chromatography: a worthy technique for the direct
measurement of liquid-liquid partition coefficients
1Area de Quimica Analitica, CCEE, ESTCE, Universitat Jaume I, 12080, Castello, Spain 2Laboratoire des Sciences Analytiques, CNRS, Université de Lyon 1, Bat CPE-308, 69622 Abstract
chromatography are: (i) a high loading capability, (ii) a very simple solute retention mechanism (liquid-liquid Countercurrent chromatography (CCC) is a liquid partitioning), (iii) either phase of the biphasic system can chromatography technique in which the stationary phase be used as a mobile phase, (iv) no irreversible solute is also a liquid. The solute separation is based on adsorption, (v) no pH problem, and (vi) less biological partitioning between the two immiscible liquid phases: solute denaturation. The high loadability is possible the mobile phase and the support-free liquid stationary because the solutes reach the volume of the liquid phase. Octanol-water partition coefficients (Po/w) of 17 β- stationary phase and not just the surface of the solid blockers and 17 sulphonamides were determined by CCC. Some of the Po/w coefficients of the molecular forms The chromatographic selectivity in CCC is only due to disagreed with the theoretical and experimental values solute partition between the two immiscible liquid phases. from literature. The Po/w coefficients of the ionic forms The solute retention mechanism depends on only one and the acidity constants were also calculated using a physicochemical parameter, the liquid-liquid partition theoretical model. Relationships between biological coefficient (P). The basic equation retention is: properties and hydrophobicity are also discussed.
On the other hand, the ionic liquid 1-butyl-3- methylimidazolium hexafluorophosphate was used for the where VR is the retention volume, VM the mobile phase first time in CCC to estimate the distribution constants of volume and VS the stationary phase volume. Thus, the different aromatic solutes, including bases, acids, and retention volume of a solute allows the determination of neutral compounds, between the ionic liquid-rich phase its partition coefficient in the biphasic system used in the and the aqueous phase. The resulting distribution constants were compared with the corresponding Two types of CCC machines are available, i.e., literature octanol–water partition coefficients.
hydrostatic and hydrodynamic, depending on the way in which equilibrium between the liquid stationary and 1. INTRODUCTION
mobile phases are reached [1-4]. The main differences between the two modes are that in hydrodynamic devices, Countercurrent chromatography (CCC) is a the centrifugal field is variable and there are at least two chromatographic separation technique based on the rotation axes in the machine containing coiled tubing in partition of solutes between two immiscible liquid phases which both phases are in contact throughout the length of as they interact in a thin tube under a centrifugal force the coiled tubes. In hydrostatic devices, the centrifugal field [1]. The mobile and stationary phases are both field is constant, there is only one rotation axis and there liquids and form a biphasic liquid system. Centrifugal are zones, ducts connecting two adjacent channels, in fields are needed to hold the liquid stationary phase when the mobile phase is pushed through it. One source of In quantitative structure-retention relationship (QSRR) interest in this method is that no solid matrix is required studies is necessary to have accurate values of octanol- to retain the stationary phase. In CCC, the stationary phase occupies up to 90% of the total volume of the logarithms of the retention factors (log k) with log P column. Due to the liquid nature of the stationary phase, the substances of pharmacological interest. P CCC is a liquid chromatography (LC) technique that uses measure the hydrophobicity of molecules and since it is special columns. Indeed, the CCC machines are just considered to estimate the partitioning over a bio- “columns”. The liquid stationary phase is stable only as membrane, it should be related to biological activity [5]. long as the centrifugal field exists, i.e., the CCC column CCC is able to work with an octanol stationary phase and exists as long as the machine rotor is running. an aqueous mobile phase. In this configuration, CCC is a The advantages of having a liquid stationary phase in useful and easy alternative to measure the octanol-water partition coefficients (Poct) of the molecules compared to the classical and tedious shake-flask method. The P 2. COUNTERCURRENT APPARATUS
values obtained with an octanol stationary phase and an AND PROCEDURE
aqueous mobile phase are the Poct parameters without any assumption, extrapolation, regression equation and/or The CCC apparatus was a hydrostatic machine with a There is an inconsistency in the literature P The apparatus is first filled with the octanol-saturated pharmaceutical molecules that show ionizable character, water phase (buffered with ammonium phosphate). Then the rotor is started and the rotation allowed to stabilize at oct is a combination of the molecular 900 rpm. The pump was rinsed with the aqueous mobile oct, mol and the ion form Poct ion is extremely sensitive to the phase. This phase entered the apparatus in the head-to-tail Many drugs of interest in the pharmaceutical industry direction (descending mode) because it is heavier than the contain basic nitrogen atoms, such as β-blockers and octanol stationary phase. During the equilibration step, sulphonamides. Since these molecules are ionizable the octanol phase is pushed off the apparatus. The octanol phase removed by the aqueous phase is collected in a depending on the aqueous-phase pH. The apparent P graduate cylinder. Once the aqueous phase appears at the coefficients of these compounds were accurately exit of the apparatus, two liquid layers are seen in the determined at different pH values using 0.01 M cylinder. The mobile-phase-stationary-phase equilibrium ammonium phosphate buffers saturated by octanol. A is reached and the CCC “column” is then ready. The model previously described [6] allows to obtain the displaced octanol phase is measured and corresponded to oct value using the solute pKa with these experimental octanol/water coefficients. Frequently, Since small amounts of octanol may be further carried out or dissolved by the aqueous mobile phase, potassium oct coefficients of the molecular forms obtained with the CCC method differ significantly from computed nitrate was used as a dead volume marker. This salt was literature values and/or experimental values obtained by not retained by the octanol phase and absorbed at 210- compared with their bibliographic values. β-blockers are When octanol is the stationary phase and water the a class of therapeutic drugs whose optical enantiomers mobile phase the partition coefficient is calculated as: show significant differences in their pharmacological effects and activities and even in their toxic effects. Sulphonamides are a class of anti-bacterial drugs used in farm animals for the treatment of a variety of bacterial where VT is the total internal volume of the CCC infections. In food-producing animals sulphonamides are apparatus. This is the direct measurement of Poct, app [1].
used not only for the treatment of several diseases but Solutes with very high Poct values move very slowly in also subtherapeutically for prophylactic purposes and/or the octanol phase. They need a too long time to emerge outside the machine. To force them from the CCC On the other hand, an ionic liquid-water system with a apparatus, the roles of the aqueous and octanol phases CCC chromatograph was used for the first time to and their flowing direction are reversed after some measure the partition coefficients of a set of aromatic reasonable flowing time in the normal direction. The compounds and compared with the corresponding mode is switched from descending (or head to tail) to literature octanol–water partition coefficients. Room ascending (or tail to head). The solutes are then eluted by temperature ionic liquids (RTIL) are salts with melting a much smaller volume of octanol (the stationary phase in points below room temperature. Typically, a RTIL the first step). The theory shows that Poct of the solute consists of nitrogen or phosphorus-containing organic depends only on the ratio of the volume of the aqueous cations and large organic or inorganic anions [7]. These phase (Vaq) pumped in the descending mode (first step) salts remain liquid over a 200-300 ºC temperature range. and the retention volume of the octanol phase (Voct) in the Their main property is that they have practically no vapor pressure [8]. For these properties, no volatility and good solvent capabilities at room temperature, RTILs have been extensively investigated as alternative ‘‘green’’ solvents [9]. The solvent properties of the RTILs make them useful candidates in CCC. They are polar solvents This procedure is known as dual-mode or back-flushing whose miscibility with water is highly dependent on their measurement and was also used to measure very small structure. The RTIL employed was 1-butyl-3- Po/w values. In this case, the octanol phase is the mobile methylimidazolium hexafluorophosphate because it was phase in the tail-to-head or ascending mode. The compounds are strongly retained, since they move very 3.2 Experimental results
slowly in the aqueous stationary phase. After several hours the phase role is inverted. The water becomes the mobile phase in the head-to-tail or descending mode. This forces the analytes out of the apparatus. Table 1 shows the apparent calculated partition coefficients of the molecular and ionic forms of the 3. RESULTS AND DISCUSSION
-blockers and sulphonamides together with their dissociation constants fitted with Eqs. 6 and 9 depending 3.1. Theoretical model
on the acid-base character of the compound. The Papp values at different pH values (data not shown) were blockers , basic ionizable compounds, are
calculated with Eqs 2 and 3 using the solute experimental retention volumes. The cationic log P+ and anionic log P- values listed in Table 1 correspond to the phosphate salt of the positive and negative forms of the compounds, The molecular form A, ionizes in a cationic form, AH+, as respectively. These values may vary if the anion and the the pH decreases. Introducing Po, the Poct value for the A cation of the buffer salt are changed. This work shows molecular form, and P+, the Poct value of the AH+ cationic that it is possible to quantify the hydrophobicity of ions form, and taking into account the acidity constant of A when ion’s Poct values are commonly neglected assuming (Eq. (4)), Ka, the experimental partition coefficient, Poct = 0 for any ion. Ionic P+ and P- values are indeed blockers, the log P values increase with pH, except for labetalol and sotalol that have a lower log P value at pH 11. This behaviour can be explained because labetalol and sotalol are also ionized at pH 11 due to the The subscripts o and w refer to the octanol phase and to presence of a phenolic OH-group and a sulphonamide the aqueous phase, respectively. Using the expression of group, respectively, in their molecule. For most of the Ka, Poct app can be formulated as: SAs the apparent log Papp values increase up to pH 2 (first dissociation constant), and decrease after pH 7 (second dissociation constant). SAs are usually cationic in acidic media (pH<2), non charged in neutral media (3<pH<7), Eq (6) shows that the measured coefficient increases with and anionic in basic media (pH>7).
Sulfonamides
compounds and undergo dissociation according to: Figure 1 shows the evolution of calculated apparent Poct app coefficient vs. pH for some - sulfonamides. This figure illustrates the dependence of The molecular form AH, ionizes in a cationic form, AH + the relative hydrophobicity upon the pH for these as the pH decreases, and ionizes in anionic form A- as the compounds, which have single or different ionisable a1 and Ka2 are the dissociation constants of the amine and sulfonic groups, respectively, in Eq. (6). At acidic pH, all β-blockers and SAs are in cationic Po, P+ and P- are the P form (AH+ or AH +, respectively, a nitrogen atom of the aliphatic chain is protonated), and they are very 2 cationic, and A- anionic forms, respectively. The experimental partition coefficient, P hydrophilic showing very small log P < -1). This means that the solutes are more soluble in the aqueous phase than in the apolar octanol phase. Since at increasing pH the ionization degree of the compounds decreases, their affinity for the aqueous phase also Using the expression of Ka, Poct, aap can be formulated as: decreases and their Poct values increase. At basic media, the β-blockers are in molecular form, except labetalol and sotalol that are ionized, showing a hydrophobic behaviour with high log Poct,app values (in the 0.25-3.61 Eq (9) shows that the measured coefficient increases with range). This means that the solutes have a higher affinity for the octanol phase. Normally, the SAs, in the 3-7 pH a and begin to decrease before the range, are in the molecular form (AH), showing most of them a hydrophobic behavior and reaching their maximum log Poct,app values in the –1.07-1.66 range. In this zone, the solutes have a higher affinity for the octanol phase. In basic media (pH>8), the sulfonamides are in an anionic form (A-, the sulfphonic group has lost a hydrogen), and they are again very hydrophilic increasing the ionization degree. Their Papp value decrease following a parabolic curve. Even if the ionic form is very hydrophilic, the CCC method allows the estimation of the very small log Papp values for ions.
Figure 2 shows the chromatograms of bisoprolol and sulfaguanidine at pH 7 done with direct and dual-mode, respectively. Fig 2A shows the actual UV signal obtained in the Poct measurement of bisoprolol. The sharp peak at 26.87 min retention time corresponds to potassium nitrate, the dead (aqueous phase) volume marker (VM=26.87 mL at 1mL/min). The broad peak at 101.77 Poct=0.004). Fig 2B shows the UV signal obtained in the oct measurement of sulfaguanidine. sulphonamides using the experimental octanol-buffer partition coefficients for different pH values.
This is an example where dual-mode was used to measure oct value. Since at pH 7 sulfaguanidine eluted with the dead volume in the direct mode (using aqueous mobile phase), firstly the octanol phase was used as mobile phase in the tail-to-head or ascending mode (Fig 2B, top), and after two hours the phase role was inverted to force the analyte out of the apparatus. The sulfaguanidine peak showed up in the aqueous phase at in the tail-to-head direction was 120 mL, giving the P app value of 7.39/120 = 0.061 at pH 7 (log Papp = -1.21).
Figure 2: CCC chromatograms at pH 7. A) Direct mode chromatogram of bisoprolol (0.5 mg + 0.1 mg of potassium nitrate). Mobile phase: aqueous with 0.01 M ammonium phosphate buffer at pH 7 in the descending head-to-tail mode; stationary phase: octanol; UV detection at 210 nm. B) Dual mode chromatogram of sulfaguanidine (0.5 mg); top: the octanol mobile phase 2 h-step, in the ascending tail-to-head mode, no signal; bottom: the aqueous phase step with 0.01 M ammonium phosphate buffer at pH 7 in the descending head-to-tail mode; stationary phase octanol; UV detection at 275 nm. General conditions: rotor speed: 900 rpm; injection volume: 1mL; flow rate: 1 mL/min.
Figure 1. Relative hydrophobicity, log P sulfanilamide (+).
CCC measurements and literature values Figure 3 shows the molecular CCC log Poct values fitted by Eqs 6 and 9 (Table 1), plotted versus the experimental literature values for 14 SAs [12] and 12 - [13,14]. The log P experimental literature data correlated Figure 3. β-blocker and SAs Poct values measured in CCC well with the measured ones (r2=0.953), with a slope and compared to literature experimental values. Regression intercept values of 1.13 and 0.013, respectively. The curve: log Poct = 1.13 log PCCC + 0.013, r2 = 0.953).
slope value, close to unity, indicates that the partition coefficients obtained by CCC and the log Poct are identical even for ionizable compounds. Hydrophobicity-biology relationships CCC produces reliable Poct values but also dissociation Log P has been successfully applied as a structural a, values. Most CCC Ka values listed in Table 1 correspond to their respective literature values.
descriptor in quantitative-structure-activity relationship (QSAR) for structurally related compounds and in some cases even for sets of chemically different compounds. On the other hand, chromatography is a powerful technique for the measurement of physicochemical parameters, and in order to emulate the biological barriers, different reversed stationary phases have been 0 40 80 120
developed such as the immobilized artificial membranes, time (min)
or immobilized liposomes. Relationships between octanol-water partition data and chromatographic indexes at a fixed or varying pH values by RPLC or micellar liquid chromatography have been studied for β-blockers [15-17].
Since log P is considered to estimate the partitioning over a bio-membrane, it should be also related to biological activity. Good relationships were obtained when the distribution coefficients of the compounds in octanol-buffer (shake-flask method) and the relative lipophilicity measured by HPLC were correlated.
0 40 80 120 160
time (min)
β-blockers have also been used to study the influence liquid–liquid distribution constants (partition of lipophilicity of drugs on the permeation through coefficients) in the biphasic liquid system used. It is biological membranes since various structurally related interesting to evaluate the distribution constants of β-blockers exhibit a wide range of lipophilicity. The different solutes in a biphasic liquid system containing an HPLC technique is well known to be able to evaluate chromatographic lipophilicity indexes (expressed as log The ionic liquid 1-butyl-3-methylimidazolium HPLC) even for highly lipophilic compounds out of the range of the shake-flask method. These indexes are determine the distribution coefficients (Kil/w) of 12 test usually calculated based on the average retention time of compounds [19]. To avoid the strong UV absorbance of the analyzed compound after two runs using linear the RTIL that contains an imidazolium aromatic ring, the equation for two adjacent standards relating their log D selected wavelength was 254 nm. L-phenylalanine, a and retention time values [18]. Figure 4 shows the common amino acid was used as a dead-volume marker since it was not retained in the working BMIM PF [18] at pH 11 for nine β-blockers (acebutolol, alprenolol, atenolol, labetalol, metoprolol, pindolol, propranolol, The viscosity of pure and dry BMIM PF6 at room temperature is very high. When BMIM PF6 is saturated with water the viscosity is lower but still too high for the ionic liquid to be used directly in the pump and the CCC chromatograph. It was therefore necessary to reduce the viscosity further by addition of a minimum amount of organic solvent. Acetonitrile was the best organic solvent to work with BMIM PF6. Indeed, in the alcohol–water– ionic liquid mixture, the alcohol tends to favour the aqueous phase producing a heavy phase rich in ionic liquid (limited volume and viscous phase). Acetonitrile partitions better between the two phases greatly reducing the viscosity of the ionic liquid-rich phase. The best composition, selected for CCC, was water–acetonitrile– BMIM PF6 (40:20:40% w/w). This composition produced a good density difference between the two liquid phases and viscosities low enough to allow pump operation.
Figure 4. Correlation between the log Papp measured by Different aromatic solutes, including bases, acids, and CCC and the log DHPLC indexes at pH 11 [17] for nine β- neutral compounds, were injected into the CCC column blockers (acebutolol, alprenolol, atenolol, labetalol, to estimate their distribution constants between the ionic metoprolol, pindolol, propranolol, sotalol and timolol). liquid-rich phase and the aqueous phase. Log DHPLC = 0.51 log Papp + 1.43, n = 9, r2 = 0.981.
The experimental CCC log Kil/w values obtained in the Although there is a correlation between the data, the 0.51 liquid ionic system were correlated with the octanol/water value of the slope indicates that the HPLC retention values, Ko/w, for five of the test compounds (aniline, 2- nitroaniline, benzonitrile, 2-nitrophenol and 2-toluidine), the square root of the octanol/water partition coefficient. which should be in their molecular form under our The intercept value of 1.43 indicates that even polar experimental conditions. Indeed, ionization does change blockers are retained by the octadecyl bonded the repartition of a solute in an aqueous–organic biphasic HPLC stationary phase. Likely, these parameters would liquid system, because the ionized form of the solute has be different with another compound family. Clearly, a much higher affinity for the polar aqueous phase. The hydrophobicity measurements are possible using HPLC, literature log Koct data [12] correlated well with the but a set of standards will always be needed to calibrate distribution constants measured in the RTIL-containing the particular set of compounds studied. There is no such need with CCC that uses octanol and water. CCC is a good technique to measure octanol-water partition log Kil/w = 1.692 log Koct + 0.915 coefficients for many solutes including ionizable The slope value, higher than unity, indicates that the distribution constants of solutes in the ionic liquid phase 3.3 Use of ionic liquids in CCC
increase faster than their hydrophobicity expressed by log CCC separates solutes because they have different 4. CONCLUSION
[9] M.J. Earle, K.R. Seddon, Pure Appl Chem 72 Accurate octanol-water partition coefficients of basic [10] R.A. Menges, G.L. Bertrand, D.W. Armstrong, J. Liq Chromatogr. 13 (1990) 3061.
sulphonamides were determined by the CCC technique. [11] S.J. Gluck, E.J. Martin, J. Liq Chromatogr. 13 The dependence of the hydrophobicity upon the pH has been plotted for several β-blockers and sulphonamides. [12] ClogP computer program version 4.01, BioByte -blockers showed a similar behavior (increasing with Corp., Claremont, CA (using the website address ).
pH range), except for labetalol and sotalol as well as all [13] A. Detroyer, Y. Vander Heyden, S. Carda-Broch, the sulphonamides that decrease at basic pH since they M.C. Garcia-Alvarez-Coque, D.L. Massart, J. Chro- are ionized. This has permitted to apply a theoretical matogr. A. 912 (2001) 211-221.
model for analytes showing acid-base properties. The [14] Hansch, C.C. In Comprehensive Medicinal Chemist- model represents correctly the change in Po/w when pH ry, Sammes, R.G., Taylor, J.B., Eds.; Pergamon changes. Good correlations can be also obtained when log P values are correlated with chromatographic indexes.
[15] J.I. Vila, R. Obach, R. Prieto and J. Moreno, Chro- The work also demonstrates that CCC is a powerful tool to estimate the liquid–liquid distribution constants of [16] F. Barbato, G. Caliendo, M.I. La Rotonda, P. solutes in any biphasic liquid system. These data can be Morrica, C. Silipo, A. Vittoria, Farmaco 45 (1990) related to the hydrophobic character of the studied compounds in a particular environment. This information [17] I. Rapado-Martinez, M.C. Garcia-Alvarez-Coque, has a high value in all studies involving quantitative R.M. Villanueva-Camanas, J. Chromatogr. A 765 pharmaceutical, and/or environmental studies, and also in [18] N. Gulyaeva, A. Zaslavsky, P. Lechner, M. Chlenov, quantitative structure–retention relationships in A. Chait and B. Zaslavsky, Eur. J. Pharm. Sci. 17 separation science and, especially, chromatography.
[19] A. Berthod, S. Carda-Broch, Anal Bioanal Chem 380 4. ACKNOWLEDGMENTS
This research was supported by a Marie Curie Fellowship of the European Community programme “Improving Hu-man Research Potential and the Socio-economic Know-ledge Base” under contract number HPMF–CT–2000-00440. AB thanks the French Centre National de la Re-cherche Scientifique (CNRS UMR 5180, Université de Lyon).
5. REFERENCES
[1] A. Berthod in Countercurrent chromatography. The support-free liquid stationary phase. D. Barcelo (ed), Comprehensive Analytical Chemistry, Vol. 38, Else-vier, The Netherlands, 2002.
[2] Y. Ito, in Advances in Chromatography, J.C. Gid- dings, E. Grushka, and J. Cazes (eds), Vol. 24, Marcel Dekker, New York, 1984. p. 181.
[3] Y. Ito, CRC Crit. Rev. Anal. Chem., 17 (1986) 65- [4] N.B. Mandava and Y. Ito, Countercurrent Chromato- graphy, Chromatographic Science Series, Vol. 44, Marcel Dekker, New York, 1989.
[5] R.N. Smith, C. Hansch, M.M. Ames, J. Pharm. Sci. [6] A. Berthod, S. Carda-Broch, M.C. Garcia-Alvarez- Coque, Anal Chem. 71 (1999) 879.
[7] C.F. Poole, J Chromatogr A 1039 (2004) 377–399[8] T. Welton, Chem Rev 99 (1999) 2071–2083

Source: http://www.mariecurie.org/annals/volume4/chem1.pdf