6.2), with the additives 18-crown-6 (1 mmol/liter) and methanol (30%) was used for the cation separation combined with conductivity detection.Although the analysis of cations in biological fluids can be performed using a range of well-known techniques, such as ion-selective electrodes, atomic emission spectrometry, and atomic absorption (AA) spectrometry, these techniques require major modification or are unsuitable for the task when the sample size is small. However, a range of techniques has been applied to the analysis of cations in nanoliter volumes.
The earliest described methods that are capable of measuring sodium and potassium in picomole amounts are ultramicro flame photometry1 and helium glow photometry2. The helium glow system is no longer commercially available. Electron microprobe analysis3 and flameless AA spectrometry4 have been used to determine these cations in samples of only 5 nl. Although with some systems it is possible to measure sodium and potassium simultaneously, the latter technique often requires different dilutions for the two ions because the calibration curves are not linear at high concentrations. Ion chromatography can be combined with conductivity detection for the analysis of sodium and potassium in serum samples, but has the disadvantage that it requires lengthy sample pretreatment prior to analysis5. Ion-selective electrodes have been used as the detector for some systems. Direct potentiometric measurements using ultramicroelectrodes6 have the advantage that knowledge of the sample volume is not required, so a considerable amount of sample handling is eliminated. A practical difficulty of this approach is the need to manipulate the electrode in order to measure a nanoliter sample under oil. Other systems make use of a flow through detector7,8; however, these systems with ion-selective electrodes as detectors are neither automated nor commercially available. Continuous flow systems using colorimetric detection have been used for the analysis of nanoliter volume samples9 and are available commercially (World Precision Instruments, Ltd., Aston, Stevenage, Hertfordshire, UK) but can detect only one ion at a time and are not automated. Capillary electrophoresis (CE) allows the analysis of multiple cations without sample pretreatment other than dilution. CE has been applied to small volumes of biological fluids, 100 
l of vitreous humor fluid10, 50 to 200 nl of pulmonary airways surface fluid11,12, and for the analysis of ocular lenses13. CE of alkali metals and alkali earth metals has been commonly combined with ultraviolet (UV) detection10,11,13,14,15,16,17. The problem of the low molecular absorptivities of metal cations is overcome either by the addition of chromophores or fluorophores to the background electrolyte, or by the addition of a compound that will form a spectroscopically active complex with the metal ions. An alternative method is to combine CE with conductivity detection. A home-built conductivity detector for CE was first reported in 198618 and was subsequently employed for cation analysis19,20. These techniques demonstrate the sensitivity and long linear calibration range of conductivity detectors. Home-built conductivity detectors are difficult to construct, and until commercial conductivity detectors became available, indirect UV detection remained the method of choice for cation analysis by CE. Recently, commercial CE systems with conductivity detection have been employed for the analysis of cations in samples of nanoliter volume12,21. This article describes a simple automated technique for the analysis of cations in samples of renal tubular fluid.
Reagents
The background electrolyte was composed of 2-[N-Morpholino] ethane-sulfonic acid (MES; 50 mmol/liter), L-histidine (50 mmol/liter, pH approximately 6.2), with the additives 18-crown-6 (1 mmol/liter) and methanol (30% vol/vol), all from Sigma-Aldrich Company (Poole, Dorset, UK). The buffer salts were dissolved before the addition of the methanol; after the methanol was added, the buffer was degassed by sonication for 45 minutes. AnalaR-grade methanol and Spectrosol-grade nitric acid were both from BDH (Merck Ltd., Poole, Dorset, UK). The sodium, potassium, lithium, ammonium, barium, and magnesium standards were Spectrosol standards from BDH. The stock materials were combined in the appropriate ratio to produce a multi-element stock standard. This stock material was used to produce standards in the range of 0 to 300 mol/liter sodium and 0 to 15 mol/liter potassium. These corresponded to 0 to 188 mmol/liter sodium and 0 to 9.4 mmol/liter potassium in the undiluted sample, which covered the range of physiological interest. All sample dilutions were carried out in nitric acid (0.1% vol/vol). The deionized water (
18 M
), produced by an Elgastat Spectrum RO (Elga Ltd., High Wycombe, Bucks, UK), was used for all solutions, standards, and sample dilution.
METHODS
A Crystal 300 series modular CE system with a model 310 injector and a Crystal 1000 conductivity detector (Thermo Unicam, Cambridge, UK)22 was used for the analysis. The capillary was a ConCap (Thermo Unicam) of fused silica 50 m i.d. 70 cm length that was used in conjunction with a ConTip (Thermo Unicam) conductivity sensor. A schematic diagram of the apparatus is shown in Figure 1. The capillary, sample table, and detector were equilibrated at 35°C. The separation was carried out at 357 V/cm. The run conditions are shown in Table 1, and the detector conditions are shown in Table 2. Prior to each batch of measurements, the capillary was flushed with NaOH (0.5 mol/liter) at 2000 mbar for five minutes, followed by deionized water at 2000 mbar for two minutes to clean and condition the capillary. The capillary and the electrode block were flushed with background electrolyte until the background reading stabilizes; the cell voltage was automatically optimized.
Sample preparation
A variety of sample types was analyzed, including urine, plasma, and perfusate, in addition to tubular fluid samples. Each sample and standard were deposited on the bottom of a watch glass that was previously half filled with water-saturated paraffin oil. Using a glass constriction pipette, 32 nl of each sample and standard were delivered into 20 l of nitric acid (0.1% vol/vol) in a 300 l autosampler vial. The volume of the constriction pipette was determined using radioisotopic methods. The samples were stored at -20°C. Immediately prior to analysis, they were defrosted and mixed briefly in an ultrasonic wave-activated cleaning bath (Decon FS 100; Decon Laboratories Ltd., Hove, UK). All of the collection tubes remained capped from the time of collection until the start of the analysis to minimize contamination and evaporation.
The volumetric flasks and bottles were all made from high-purity polypropylene [Nalge (Europe) Ltd., Rotherwas, Hereford, UK] that contained only very low levels of sodium and potassium. All containers (sampling vials, volumetric flasks, and standard bottles) for reagents or samples were soaked in nitric acid (50% vol/vol) for 24 hours and then rinsed in deionized water. After cleaning, the containers were stored under deionized water until the day of analysis, when they were then rinsed five times with deionized water and allowed to air dry. Pipette tips and autosampler starburst caps were not acid washed, but were rinsed in deionized water and allowed to air dry prior to use. Experience has shown that this procedure is sufficient to prevent contamination.
RESULTS AND DISCUSSION
Figure 2 and Figure 3 illustrate standard curves for sodium and potassium, respectively. Each point represents the mean of five determinations. Curves for NH4, K, Na, Ca, Li, Mg, and Ba exhibit a coefficient of determination (r2) by linear regression of 0.98 or greater. Both sodium and potassium have a long linear calibration range. Sodium is linear up to an actual standard concentration of at least 1000 mol/liter (r2 > 0.999); potassium is linear up to an actual standard concentration of at least 500 mol/liter (r2 > 0.999). The long linearity range means that sodium and potassium can be determined simultaneously. The detection limit is calculated using the mean baseline peak area plus three standard deviations of the baseline noise in a diluted sample of 18 M water. The detection limit is 9.5 mol/liter for sodium and 1.6 mol/liter for potassium. These values correspond to 5.9 mmol/liter sodium and of 1.0 mmol/liter potassium in the undiluted sample. Because the volume of sample loaded onto the capillary is only 20.5 nl, the amount of element present at the detection limit is 0.2 pmol for sodium and 30 fmol for potassium.
Figure 2.
Calibration curve for sodium. The solid line represents the regression and the dotted lines represent the 95% confidence interval (Y = -1.23418 + 0.916838x; r2 = 1.000).
Full figure and legend (4K)Figure 3.
Calibration curve for potassium. The solid line represents the regression and the dotted lines represent the 95% confidence interval (Y = -8.6E-02 + 1.02006x; r2 = 0.998).
Full figure and legend (4K)Figure 4 shows an electropherogram of a separation of a standard solution. Figure 5 is an electropherogram of the separation of a sample taken from the renal distal tubule during free flow micropuncture.
Figure 4.
Electropherogram of cations in standard material, (1) ammonium, (2) potassium, (3) sodium, (4) calcium, (5) magnesium, (6) barium, and (7) lithium. Capillary: ConCap 50 m i.d. 53 cm length.
Figure 5.
Electropherogram of cations in a sample from the distal tubule. (1) ammonium, (2) potassium, (3) sodium, and (4) calcium. Capillary: ConCap 50 m i.d. 40 cm length.
The between- and within-batch coefficient of variation (reproducibility/repeatability) was determined using aliquots of a standard solution. Batches of seven samples were analyzed over 10 days. For potassium, the within-batch variation was 6.6%, and the between-batch variation was 8.8% at a level of 6 mol/liter. For sodium, the within-batch variation was 4.3%, and the between-batch variation was 8.8% at a level of 110 mol/liter. The reproducibility of migration times was <0.5% for all elements.
The comparison data with atomic absorption (AA) are shown in Figure 6 and Figure 7. The AA results were obtained using a Perkin Elmer 3110 Atomic Absorption Spectrometer with an HGA-600 furnace (Beaconsfield, Bucks, UK). A paired t-test of the percentage difference between the two sets of data gives the probability of the two sets of data not having a mean percentage difference of zero as 0.12 for sodium and 0.11 for potassium.
Figure 6.
Comparison of sodium values with those obtained by atomic absorption (AA). The solid line represents the regression (Y = -1.58724 + 1.01219x; r2 = 0.974).
Full figure and legend (4K)Figure 7.
Comparison of potassium values with those obtained by atomic absorption (AA). The solid line represents the regression (Y = -0.326357 + 0.978315x; r2 = 0.963).
Full figure and legend (5K)Capillary electrophoresis has the advantage over other micosample techniques in that if the mobilities of the cations of interest are sufficiently different, the analytes are completely separated from the background components prior to detection. This removes the interference of similar cations. However, the mobilities of the cations of interest are too close for them to be resolved without the addition of complex-forming ligands. Ammonium and potassium comigrate and require the addition of 18-crown-6; the methanol resolves the calcium from the sodium.
The reproducibility of the reported method compares well with other techniques using nanoliter samples at physiological concentrations6,10,13,14,16. Other workers have used direct on-column sampling to achieve improved repeatability21. This technique allows samples to be taken during the course of a physiological experiment. However, this method of sampling significantly increases the time of each analysis and precludes automation.
The repeatability of migration times for the reported method is sufficiently good not to require an internal migration time marker.
The initial cost of the equipment is approximately £20,000 ($34,000), but is offset by the low running cost. Once installed, the cost per sample batch is less than £0.50 (85 cents US). The cation has a sample throughput of approximately six tests per hour. The 30-position programmable autosampler allows the tests to be run overnight. An added advantage of the system is its flexibility, which enables it to be used for other applications.
Although we have not made a direct comparison between CE and AA in the measurement of other cations, CE has the potential to measure these simultaneously in a single sample. Clearly, this will be a major advantage in renal micropuncture studies because of the added data from a single experiment and the ability to directly compare the tubular handling of different cations during the same physiological or pharmacological challenge. Renal micropuncture studies of transgenic mice stand to benefit most from this improved ultramicroanalytical technique.
Summary
This Technical Note describes the analysis of subpicomole quantities of sodium and potassium in samples of 32 nl diluted in 20 l of nitric acid (0.1% vol/vol). Because only 20.5 nl of diluted sample are used for each analysis, multiple measurements need to be made on each sample. The calibrations for sodium and potassium are linear over three orders of magnitude, allowing the simultaneous measurement of sodium and potassium. The analytes are separated from the sample matrix in less than six minutes to provide a fast, interference-free method. The technique is automated to produce a technique that is quick and easy to use, and is also inexpensive.
References
| 1. | Malnic G, Klose RM & Giebisch G. Micropuncture study of renal potassium excretion in the rat. Am J Physiol 1964; 206: 674–686. | PubMed | ISI | ChemPort | |
| 2. | Vurek GG & Bowman GL. Helium-glow photometer for picomole analysis of alkali metals. Science 1965; 149: 448–460. | ISI | ChemPort | |
| 3. | Morel F, Roinel N & Legrimellac C. Electron probe analysis of tubular fluid composition. Nephron 1969; 6: 350–364. | PubMed | ISI | ChemPort | |
| 4. | Wingo CS, Bixler GB, Park CH & Straub SG. Picomole analysis of alkali metals by flameless atomic absorption spectrophotometry. Kidney Int 1987; 31: 1225–1228. | PubMed | ISI | ChemPort | |
| 5. | Thienpont LM, Van Nuwenborg JE & Stockl D. Ion chromatography as potential reference methodology for the determination of total sodium and potassium in human serum. J Chromatogr A 1995; 706: 443–450. | Article | PubMed | ISI | ChemPort | |
| 6. | Kaufman JS, Costello TP & Hamburger RJ. Measurement of Na and K in nanoliter droplets by ion-specific microelectrodes. Kidney Int 1990; 38: 525–528. | PubMed | ISI | ChemPort | |
| 7. | Garvin JL. Picomolar quantitation of potassium using a continuous-flow apparatus. Kidney Int 1989; 36: 726–729. | PubMed | ISI | ChemPort | |
| 8. | Garvin JL. A simple method to determine millimolar concentrations of sodium in nanoliter samples. Kidney Int 1993; 44: 875–880. | PubMed | ISI | ChemPort | |
| 9. | Terada Y & Knepper MA. Continuous-flow quantitation of Na+ and K+ in nanoliter samples using chromogenic macrocyclic ionophores. Am J Physiol 1989; 257: F893–F898. | PubMed | ISI | ChemPort | |
| 10. | Ferslew KE, Hagardorn AN, Harrison MT & McCormick WF. Capillary ion analysis of potassium concentrations in human vitreous humor. Electrophoresis 1998; 19: 6–10. | Article | PubMed | ISI | ChemPort | |
| 11. | Transfiguracion JC, Dolman C, Eidelman DH & Lloyd DK. Determination of the inorganic ion composition of rat airway surface fluid by capillary electrophoresis: Direct sample injection to allow multiple analyses from nanoliter volumes. Anal Chem 1995; 67: 2937–2942. | Article | PubMed | ISI | ChemPort | |
| 12. | Cowley EA, Govindaraju K, Lloyd DK & Eidelman DH. Airway surface fluid composition in the rat determined by capillary electrophoresis. Am J Physiol 1997; 273: L895–L899. | PubMed | ISI | ChemPort | |
| 13. | Shi H, Zhang R, Chandrasekher G & Ma Y. Simultaneous detection and quantitation of sodium, potassium, calcium and magnesium in ocular lenses by high performance capillary electrophoresis with indirect photometric detection. J Chromatogr A 1994; 680: 653–658. | Article | ISI | ChemPort | |
| 14. | Buchberger W, Winna K & Turner M. Applications of capillary zone electrophoresis in clinical chemistry: Determination of low-molecular-mass ions in body fluids. J Chromatogr A 1994; 671: 375–382. | Article | PubMed | ISI | ChemPort | |
| 15. | Quang C & Khaledi MG. Prediction and optimization of the separation of metal cations by capillary electrophoresis with indirect UV detection. J Chromatogr A 1994; 659: 459–466. | Article | PubMed | ISI | ChemPort | |
| 16. | Xu X, Kok WT, Kraak JC & Poppe H. Simultaneous determination of urinary creatinine, calcium and other inorganic cations by capillary zone electrophoresis with indirect ultraviolet detection. J Chromatogr B Biomed Appl 1994; 661: 35–45. | PubMed | ISI | ChemPort | |
| 17. | Cheng K, Nordmeyer FR & Lamb JD. A novel buffer system for separation of metal cations by capillary electrophoresis with indirect UV detection. J Capillary Electrophor 1995; 2: 279–285. | PubMed | ISI | ChemPort | |
| 18. | Foret F, Deml M, Kahle V & Bocek P. On-line fiber optic UV detection cell and conductivity cell for capillary zone electrophoresis. Electrophoresis 1986; 7: 430–432. | Article | ISI | ChemPort | |
| 19. | Huang X, Pang TJ, Gordon MJ & Zare RN. On-column conductivity detection for capillary zone electrophoresis. Anal Chem 1987; 59: 2747–2749. | Article | ISI | ChemPort | |
| 20. | Huang X, Gordon MJ & Zare RN. Quantitation of Li+ in serum by capillary zone electrophoresis with an on-column conductivity detector. J Chromatogr B Biomed Appl 1988; 69: 385–390. |
| 21. | Govindaraju K, Cowley EA, Eidelman DH & Lloyd DK. Microanalysis of lung airway surface fluid by capillary electrophoresis with conductivity detection. Anal Chem 1997; 69: 2793–2797. | Article | PubMed | ISI | ChemPort | |
| 22. | Haber C, Jones WR, Soglia J, Surve MA, McGlynn M, Caplan A, Reineck JR & Krstanovic C. Conductivity detection in capillary electrophoresis: A powerful tool in ion analysis. J Capillary Electrophor 1996; 3: 1–11. | PubMed | ISI | ChemPort | |
Acknowledgments
We thank The Wellcome Trust (Grant number M/94/1671) and the St. Peter's Research Trust for their financial support. We thank Colin Simpson of Kings College, London, and Saul Parry for their helpful discussion and advice while setting up this method. We also thank David Hutchings of Thermo Unicam for providing technical support for the CE system.
