Ion-selective electrode for measuring low Ca 2+ concentrations in the presence of high K +,Na + and Mg 2+ background

Transcription

1 Anal Bioanal Chem (2006) 385: DOI /s y ORIGINAL PAPER Iwona Bedlechowicz-Śliwakowska. Peter Lingenfelter. Tomasz Sokalski. Andrzej Lewenstam. Magdalena Maj-Żurawska Ion-selective electrode for measuring low Ca 2+ concentrations in the presence of high K +,Na + and Mg 2+ background Received: 12 April 2006 / Revised: 2 June 2006 / Accepted: 8 June 2006 / Published online: 18 July 2006 # Springer-Verlag 2006 Abstract In this work, ion-selective electrodes for calcium ion were investigated. Two ionophores were used in the membranes: ETH 1001 and ETH 129. An internal filling solution buffered for primary ion was used that allowed the lower detection limit to be decreased down to M. Theoretical and experimental electrode characteristics pertaining to both primary and interfering ions are discussed. Better behavior was obtained with the electrode prepared with ETH 129 in the membrane. This electrode would be the most likely candidate for obtaining a low Ca 2+ detection limit in measurements performed with high K +, Na +, Mg 2+ background, which is found inside the cells of living organisms, for example. The potentiometric response of the electrode in solutions containing main and interfering ions is in good agreement with simulated curves obtained using the Nernst Planck Poisson (NPP) model. Keywords Calcium ion-selective electrode. Selectivity. Detection limit Introduction Calcium is the fifth most common element and the most prevalent metal among those found in the human body. Calcium influences many important physiological functions, such as muscle contraction, heart function, transmission of nerve impulses and blood clotting, and plays an important role in intracellular processes [1, 2]. The calcium I. Bedlechowicz-Śliwakowska. M. Maj-Żurawska (*) Department of Chemistry, Warsaw University, Pasteura 1, Warsaw, Poland mmajzur@alfa.chem.uw.edu.pl Tel.: Fax: P. Lingenfelter. T. Sokalski. A. Lewenstam Process Chemistry Group, Åbo Akademi University, Biskopsgatan 8, Turku, Finland transport system in cells (the calcium pump ) plays a very important role for the whole organism. This pump is the only high-affinity Ca 2+ transport system present in all mammalian cells. In extracellular fluids such as blood, virtually all of the calcium is in the serum or plasma water, which has a total calcium concentration in the range of to M[3], while the free, ionized calcium concentration is to M [4]. The concentration of total calcium in intracellular fluids is approximately 10 6 to 10 8 M, which is less than 1/1000 of that in extracellular fluids (10 3 M). Intracellular free calcium concentration is an important intracellular second messenger. The intracellular free calcium concentration may be altered through number of mechanisms, such as through hormone or drug adminstration, and is very important factor for clinical diagnosis. Ion-selective electrodes (ISEs) are, in many respects, ideal sensors for use in the analysis of calcium in clinical samples. In recent years, the advantages of ISEs have led to their adoption in routine measurements of ionized calcium in blood. The past decade has been a very fruitful time for the development of ISE measurements at low concentrations [5 10]. Taking calcium measurements inside cells, for example in erythrocytes, is a demanding task because of there is a low concentration of calcium in the presence of relatively high concentrations of potassium, sodium and magnesium [11 13]. Such measurements require a highly selective calcium ionophore against potassium and sodium. Currently, the ionized calcium concentration in erythrocytes is determined by 19 F-NMR spectroscopy [14]orbya fluorescence method [15]. These methods are relatively expensive and require specialized equipment. On the other hand, methods for determining ionized magnesium, sodium and potassium in the intracellular fluids of erythrocytes have been developed [16, 17], and these demonstrate that ISEs can be applied to intracellular measurements. Procedures that can be used to prepare samples from erythrocyte intracellular fluid and a complete methodology for ISE measurements of Mg 2+,Na + and K + have been proposed. The reliability of the new method relative to reference methods has also been confirmed. Moreover, the

2 1478 successful implementation of ISE intracellular measurements in the flow-through system of a routine potentiometric clinical analyser was achieved. However, the application of ion-selective electrodes to intracellular calcium measurements needs further development for it to work properly. In this paper, a calcium-ise based on ionophore ETH 129 that allows the detection limit to be decreased down to M in the presence of high background concentrations of potassium, sodium and magnesium is presented and compared with calcium-ises based on another ionophore (ETH 1001). Experimental Reagents For all experiments, solutions were prepared with freshly deionized water (Milli-Q plus, Millipore, Vienna, Austria, resistance 18.2 MΩ cm or ELGA-maxima ultrapure water, Elga Ltd., High Wycombe, UK, resistance 18.2 MΩ cm). P.A. grade (POCh, Gliwice, Poland; Merck, Darmstadt, Germany; J.T. Baker, Deventer, The Netherlands) and suprapure (Merck) chemicals were used. High molecular weight poly(vinyl chloride) (PVC), o-nitrophenyl octyl ether (o-npoe), calcium ionophores: calcium ionophore I (ETH 1001) [( )-(R,R)-N,N -bis- [11-(ethoxycarbonyl)undecyl]-N,N -4,5-tetramethyl-3,6- dioxaoctanediamide], calcium ionophore II (ETH 129) N,N,N,N -tetracyclohexyl-3-oxapentanediamide; lipophilic salts: potassium tetrakis[3,5-bis(trifluoromethyl)phenyl]bo rate (KTFPB) and sodium tetrakis[3,5-bis(trifluoromethyl) phenyl]borate (NaTFPB), were from Fluka (Buchs, Switzerland). Tetrahydrofuran (THF) (Merck) for membrane preparation was freshly distilled. Ion-selective membrane and electrodes Two different membranes with a total mass of 200 mg were prepared as follows: membrane A: PVC: o-npoe, 1:2 (w/w), 15 mmol kg 1 calcium ionophore I (ETH 1001), 12 mmol kg 1 lipophilic salt (KTFPB); membrane B: PVC:o- NPOE, 1:2 (w/w), 10 mmol kg 1 calcium ionophore II (ETH 129), 5.5 mmol kg 1 lipophilic salt (NaTFPB). The concentrations given here are relative to total membrane mass. Ion-selective membranes were cast by dissolving the membrane components in 2 ml of THF. After dissolution, the mixture of components was poured into a 26-mm or 28- mm-diameter glass ring placed on a glass plate, and left for free evaporation of the solvent. For each electrode, a 6 mm disk was cut with a cork borer from the mother membrane and fixed into a Philips (Eindhoven, Netherlands) electrode body (IS 561). The inner filling solution (IFS) compositions were MCaCl 2 and M EDTA adjusted to ph 8.5 or to ph 9.5 with 0.1 M NaOH. The calculated calcium ion concentrations were: M and M, respectively (logβ Ca-EDTA =10.9, logk 1 =2.0, logk 2 =2.7, logk 3 =6.2, logk 4 =10.3). Membranes were conditioned for two days in 10 4 M CaCl 2 solutions. After that time, a conditioning solution identical to the inner solution was used for each electrode. Measurements were performed with a 16-channel potentiometer (Lawson Labs. Inc., Malvern, PA, USA). The reference electrode was an Ag/AgCl double junction (W. Möller, Zurich, Switzerland) electrode with 3 M KCl used as the reference electrolyte and 0.1 M KCl as the bridge electrolyte in all experiments. Activity coefficients were calculated from the Debye Hückel extended equation [18], and EMF values were corrected for liquid junction potentials with the Henderson approximation [19]. Calibration curves were measured in solutions using the standard addition method or by sequential dilution using two Metrohm (Herisau, Switzerland) 700 Dosinos controlled by a Metrohm 711 Liquino system. Solutions were stirred at a stable stirring rate using a Metrohm magnetic stirrer. Results and discussion Measurements made in the presence of high concentrations of interfering ions such as those found in biological samples require very selective ionophores in the ionselective membrane. The concentrations of interfering ions for a calcium electrode in erythrocytes are high: M K +, M Na + and M Mg 2+ are typical. These levels are several orders of magnitude higher than that of calcium, which typically has a concentration of between and M[14, 15]. The selectivity coefficients given in the literature for the oft-used calcium ionophore ETH 1001 [20] andthe most selective ETH 129 [21] are presented in Table 1. It is feasible to estimate the highest tolerable selectivity Table 1 Selectivity coefficients of ionophores ETH 1001 and ETH 129 for calcium electrode against potassium, sodium and magnesium ions, obtained from the literature, and estimated values for 1% and 5% analytical error in erythrocyte log K Ca/X ETH 1001 [20] ETH 129 [21] Maximal values, (Eq. 1) Maximal values, (Eq. 2) P ij =1% P ij =5% P ij =1% P ij =5% K Na Mg

3 1479 coefficients for a calcium electrode possible for the intracellular solution using the following equations [22 24], the latter being preferable due to its consideration of charge inequality: K ij;max ¼ K pot ij;max ¼ a i;min a j;max a i;min a j;max Zi=Z P ij j 100 Zi=Z j (1) P Zi=Z j ij (2) 100 where: a i,min is the lowest activity of the main ion, a j,max is the highest activity of the interfering ion, z i and z j are their charges, and P ij is the highest tolerable error in the main ion activity determination (%). The maximal selectivity coefficient values calculated are given in Table 1. They were obtained using: a Ca, min = M, a K, max =0.100 M, a Na, max =0.016 M, a Mg, max =0.001 M, P ij =1% or P ij =5%. The selectivity of the calcium ionophore ETH 1001 is not good enough to measure intracellular calcium, regardless of the equation used. The selectivity of calcium ionophore ETH 129 is adequate when (Eq. 2) is used, even for a 1% maximal error in the calcium activity determination. The behaviors of two ISEs prepared with calcium ionophores ETH 1001 and ETH 129 were studied. The electrode prepared with membrane A exhibits a super- Nernstian response in calcium chloride solutions (Fig. 1). In the presence of a 0.1 M KCl background, the detection limit is shifted up to a value of about 10 6 M, dictated by the interference from potassium ions (shown by the dashed line in Fig. 1). For membrane A, the relatively poor selectivity of the ionophore limited measurements at low concentrations. Even though the concentrations of free calcium ion in the inner solutions were very low, about M, it was impossible to extend the linear Nernstian response. A comparison of the data in the literature for selectivity coefficients based on calculations [25] using complex formation constants for different calcium ionophores with experimental measurements [21] showed that ETH 129 is more selective for calcium ions against potassium and sodium ions. We have determined the selectivity coefficient for the electrode with this ionophore (membrane B) against K + for comparison with literature data. The selectivity coefficients for ETH 129 calcium-ise were determined using two methods which are termed the strong interference method (a) introduced by Bakker [21, 26] and the inner filling solution buffering method (b) proposed by Sokalski et al. [5] here. The calibration curves for methods (a) and (b) are presented in Figs. 2 and 3, respectively. In the case of method (a), a freshly made membrane B conditioned in 0.01 M KCl and containing interfering ion in the inner filling solution (0.01 M KCl) was used. The first calibration curve was recorded in the interfering ion (K + ), and a second calibration curve was then recorded in the primary ion (Ca 2+ ), without any change made to the ISE membrane or inner filling solution (Fig. 2). The selectivity coefficient log KCa=K Pot calculated from the data was 11.2 (the slopes of the calibration curves were: in the Ca 2+ solutions S=26.7 mv/dec, in the K + solutions S=58.6 mv/dec). Membrane B was then conditioned in 0.01 M calcium solution on both sides. After two days of conditioning, the inner filling solution was changed to a solution containing buffered calcium ions at a concentration of M. The selectivity was Fig. 1 Calibration curves of Ca-ISE A (ETH 1001, o-npoe, IFS: [Ca 2+ ]= M): squares, CaCl 2 ; circles, KCl; triangles, CaCl 2, in the presence of a constant 0.1 M KCl background Fig. 2 Comparison of experimental points and simulated calibration curves measured during the determination of the selectivity coefficient for Ca-ISE using the strong interference method (ETH 129, o-npoe, IFS: 0.01 M KCl). The membrane was conditioned in the interfering ion (K + ). D i =D j = cm 2 s 1. k I i=10 8, k S i=10 3, k I j=3 10 4, k S j= , k I R=0 and k S R=0 cm s 1

4 1480 Fig. 3 Determination of selectivity coefficient for Ca-ISE B with the inner filling solution buffering method (ETH 129, o-npoe, IFS: [Ca 2+ ]= M): squares, calibration curve in main ion (Ca 2+ ); filled circles, calibration curve in interfering ion (K + ); open circles, calibration curve in main ion (Ca 2+ ) in the presence of 0.1 M KCl then measured using method (b) (Fig. 3). The selectivity coefficient calculated from this experiment is log KCa=K Pot = 10.3 (slopes: in Ca 2+ S=24.7 mv/dec, in K + S=57.5 mv/dec). This is much better than the selectivity coefficients measured for ETH 1001 (log KCa=K Pot = 6.8 from Fig. 1; slopes: in Ca 2+ S= 27.8 mv/dec, in K + S=51.4 mv/dec). These results prompted the use of ionophore ETH 129 for future measurements. Literature values are comparable to those calculated here. Although the shapes of the calibration curves in Figs. 2, 3 are similar, the reasons for the super-nernstian behavior are as different as the methods used. Method (a) requires that the calibration curves for the interfering and primary ions be measured with the membrane conditioned in interfering ion and with an inner filling solution containing only interfering ion. The calibration curves are measured in interfering ions in order of most discriminated to least discriminated, before the primary ion solution is measured last of all. For measurements performed in primary ion, this naturally gives rise to ion fluxes in the membrane coupled with the subsequent conversion of the membrane from potassium- to calcium-rich. This process leads to a super- Nernstian response shift at ~10 5 to 10 7 M, very similar to that reported by Hulanicki and Lewandowski in their seminal work on the subject [27]. Method (b) requires that the calibration curves for both the primary and interfering ions be measured using an electrode with an inner filling solution buffered for primary ions, so that the concentration of calcium ion is low, while the concentration of interfering ion is kept relatively high. However, the membrane in this case is conditioned in primary ion. The potential jump then occurs because of the flux of calcium ions in the direction of the internal solution through the membrane from the sample solution Nernstian layer. No membrane composition conversion occurs. The response shifts observed during measurements with these methods can be interpreted by applying either the diffusion layer model (DLM) [28 33] or the Nernst Planck Poisson (NPP) model [34]. The DLM model is restricted to ions of like charge, assumes local equilibrium at the solution membrane interface, and electroneutrality. The NPP-based model for any number of ions and any ionic charges is able to predict membrane composition changes and potential in space and time, and does not require local equilibrium and electroneutrality. However, it does not consider diffusion layers in the solution. The latter model applies for conversion of the membrane from K + -to Ca 2+ -rich, as shown in Fig. 2. This figure compares the experimental results to the simulated curves generated by the NPP model. For Ca 2+ ISE B, the fit is relatively good for both the primary and interfering ions over a wide range of concentrations, which indicates a fair correlation of the model parameters to real processes. ISE membranes containing ionophore ETH 129 have sufficient selectivity to enable measurements in an erythrocyte background as documented above. The repeatability of chemical analyses is also a very important factor; so repeated calibrations of the calcium electrodes in an erythrocytemimicking background solution (0.100 M K +, M Na + and M Mg 2+ ) were performed. Over a period of one day, multiple calibrations (n=3) were performed with ISEs containing membrane B, using M free ionized Ca 2+ in the inner filling solution. Each readout in the calibration curve was taken after ten minutes. The dependence of electrode response on the logarithm of calcium activity is shown in Fig. 4. Error bars represent standard deviations. The standard deviation values are comparable to those given earlier for high concentrations for the calcium selective electrode with the ETH 1001 ionophore [22]. Here, the response from 10 5 to M was recorded. The log a Fig. 4 One day repeatability measurements (n=3) for Ca-ISE B (ETH 129, o-npoe, IFS [Ca 2+ ]= M, constant background solution (0.100 M KCl, M NaCl, M MgCl 2 ). Each readout at the calibration curve was taken after ten minutes

5 1481 the calibration curves are dependent on the ionic compositions of the standards. In general, in order to optimize the performance of an ion-selective electrode, the composition of the membrane and the inner solution should be optimized separately for different applications and correlated with the concentration of calcium and interfering ions in the sample. To improve the precision of measurements taken at such low concentrations of target ion, extreme care must be taken. The measurements should be carried out in such a way that the composition of the ion selective membrane does not change, otherwise super-or sub-nernstian behavior is observed. Fig. 5 Calibration curves for Ca-ISE B (ETH 129, o-npoe, IFS: [Ca 2+ ]= M) in CaCl 2 in the presence of constant background solution (0.100 M KCl, M NaCl and M MgCl 2 ): squares, in p.a. grade chemicals; circles, in suprapure chemicals electrode exhibited a Nernstian response toward calcium ions, with a slope of 27.3±0.8 mv/decade and a detection limit of MCa 2+. It is very important to note that the ionic strength of the sample solution influences the ionic fluxes through the membrane, not just the concentration of free main ion. Thus, the linearity of the calibration curve is dependent on the whole ionic composition of the sample. Because of the difficulties inherent in measuring low concentrations of Ca 2+ in a relatively concentrated background solution of K +, chemical purity is of crucial importance. For instance, p.a. grade KCl usually contains 0.001% CaCl 2. This means that a 0.1 M KCl solution can contain as much as MCa 2+. The data presented for the electrode with a concentration as low as M Ca 2+ in the internal filling solution clearly supports the use of only suprapure chemicals (see Fig. 5). As observed in earlier investigations with very low detection limits [35], the detection limit for the calibration curve measured in p.a. grade chemicals is dictated by the Ca 2+ impurity. In contrast, when suprapure (99.999%) KCl was used, the calibration curve showed super-nernstian behavior, indicating a lack of Ca 2+ in the background chemicals. Conclusions A comparison of the calibration curves for ISEs prepared with ETH 129 and ETH 1001 shows that ETH 129 is indeed clearly better suited for use in intracellular measurements. The ISE incorporating ETH 129 with lowered calcium activity in the inner solution has been shown to be a promising sensor for the determination of low calcium concentrations in high background solutions containing high concentrations of potassium, sodium and magnesium ions, as found in erythrocytes. Detection limits as low as MCa 2+ can be achieved. The linearities of Acknowledgement This work was financially supported by KBN grant 4 T09A References 1. Aurbach GD, Marx SJ, Speigel AM (1992) In: Wilson JD, Foster DW (eds) Williams textbook of endocrinology 8th edn. W. B. Saunders Co., Philadelphia, PA, p Rasmussen H (1986) N Engl J Med 314: Hulanicki A, Lewandowski R, Michalska A, Lewenstam A (1990) Anal Chim Acta 233: Vadstrup S, Wandrup J (1991) J Intern Med 230: Sokalski T, Ceresa A, Zwickl T, Pretsch E (1997) J Am Chem Soc 119: Sokalski T, Zwickl T, Bakker E, Pretsch E (1999) Anal Chem 71: Sokalski T, Ceresa A, Fibbioli M, Zwickl T, Bakker E, Pretsch E (1999) Anal Chem 71: Sokalski T, Bedlechowicz I, Maj-Zurawska M, Hulanicki A (2001) Fresenius J Anal Chem 370: Puntener M, Vigassy T, Baier E, Ceresa A, Pretsch E (2004) Anal Chim Acta 503: Malon A, Radu A, Qin W, Qin Y, Ceresa A, Maj-Zurawska M, Bakker E, Pretsch E (2003) Anal Chem 5: Malon A, Maj-Zurawska M (2001) Anal Chim Acta 448: Tomaszewski JJ (2001) Diagnostyka laboratoryjna. PZWL, Warszawa 13. Murray RK, Granner DK, Mayes PA, Rodwell VW (1993) Harper s biochemistry. Prentice-Hall, New York 14. Aiken NR, Satterlee JD, Galey WR (1992) Biochim Biophys Acta 1136: Soldati L, Spaventa R, Vezzoli G, Zerbi S, Adamo D, Caumo A, Rivera R, Bianchi G (1997) Biochem Biophys Res Com 236: Malon A, Brockmann C, Fijałkowska-Morawska J, Rob P, Maj- Żurawska M (2004) Clin Chim Acta 349: Malon A, Maj-Żurawska M (2005) Sensor Actuat B 108: Meier PC (1982) Anal Chim Acta 136: Meier PC, Ammann D, Morf WE, Simon W (1980) In: Koryta J (ed) Medical and biological applications of electrochemical devices. Wiley, Chichester, UK pp Bedlechowicz I, Maj-Zurawska I, Sokalski T, Hulanicki A (2002) J Electroanal Chem 537: Bakker E (1997) Anal Chem 69: Oesch U, Ammann D, Pham HV, Wuthier U, Zünd R, Simon W (1986) J Chem Soc Faraday Trans 1(82): Pretsch E, Bakker E, Bühlmann P (1997) Chem Rev 97: Nagy G, Toth K, Pungor E (1993) Anal Lett 26(7): Qin Y, Mi Y, Bakker E (2000) Anal Chim Acta 421: Bakker E (1996) J Electrochem Soc 143:L83

6 Hulanicki A, Lewandowski R (1974) Chemia Analityczna (Warsaw) 19: Morf WE (1981) The principles of ion-selective electrodes and of membrane transport. Akademiai Kiado, Budapest 29. Maj-Zurawska M, Sokalski T, Hulanicki A (1988) Talanta 35: Hulanicki A, Lewenstam A (1981) Anal Chem 53: Lewenstam A, Hulanicki A, Sokalski T (1987) Anal Chem 59: Morf WE (1983) Anal Chem 55: Radu A, Meir AJ, Bakker E (2004) Anal Chem 76: Sokalski T, Lingenfelter P, Lewenstam A (2003) J Phys Chem B 107: Bereczki R, Takács B, Langmaier J, Neely M, Gyurcsányi RE, Tóth K, Nagy G, Lindner E (2005) Anal Chem 78:942