Technical Notes Anal. Chem. 1995, 67,4217-4219
Extraction of Serum and Urine Calcium with Ion Exchange Membrane Filters for lsotope Enrichment Determination Using Thermal Ionization Mass Spectrometry Nancy E. Vieira* and A
M L. Yergey
Laboratory of Theoretical and Physical Biology, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Matyland 20892
The extraction of calcium from water, serum, and urine usingBio-Rex25 mm ion exchange membrane filters was compared with the oxalate precipitation procedure currently used in our laboratory. Total recoveries of a known quantity of calcium loaded onto the membrane filters for water, serum and urine were as follows: (a) cation exchange filter, 85%, 74%, and 66%;(b) Chelex, 65%, 98%,and 20%;and (c) oxalate precipitation, 93%,loo%, and 96%,respectively. Regression analysis for precipitation versus ion exchange isotope ratio measurements of standards prepared using highly enriched calcium-44 showed slopes of unity. An improvement of automated sample analysis was observed for water and urine calcium samples extracted with ion exchange filters. We have used calcium stable isotopes extensively to study the absorption and kinetics of calcium in a variety of populations, including premature infants, pregnant and lactating women, and patients with disorders of calcium Calcium had been extracted from serum and urine samples by oxalate precipitation,6a procedure that may also coprecipitate other cations that could interfere with the thermal ionization process of the extracted ~ a l c i u m . Such ~ interference may cause irreproducible sample volatilization from the ionization filament, resulting in the interruption of automated data acquisition. This is observed principally in work with urine samples. We recently published a method * Address correspondence to this author at LTPB, NICHD, NIH, 10 Center Dr., MSC 1580,Building 10,Room 6C208,Bethesda, MD 20892-1580. (1) Abrams S. A; Yergey A. L.; Schanler R J.; Vieira N. E.: Welch T. R 1. Pediatr. Gustroenterol. Nutr. 1994,18, 20-24. (2) Cross, N.A; Hillman, L S.; Allen, S. H.;Krause, G. F.: Vieira, N. E. Am. J Clin. Nutr. 1995,61, 514-523. (3) Specker B. L.; Vieira N. E.; O'Brien IC 0.;Ho, M. L: Heubi J. E.; Abrams S. A; Yergey A. L. Am. J. Clin. Nutr. 1994,59,593-599. (4) Shoemaker L.;Welch T. R; Bergstrom W.; Abrams S. A; Yergey k L; Vieira N. Pediatr. Res. 1993,33, 92-96. (5) Abrams, S. A.; Lipnick, R N.: Reira, N. E.; Stuff J. E.: Yergey, A. L. J. Rheumatol. 1993,20. 1196-1200. (6) Yergey, A. L.;Vieira, N. E.: Hansen, J. W. Anal. Chem. 1980,52, 18111814. (7) Kolthoff, I. M.; Sandell, E. B. Textbook of Quantitative Inorganic Analpis, 3rd ed.; Macmillan Co.: New York, 1952;Chapter 8. This article not subject to US. Copyright. Published 1995 Am. Chem. Soc.
using ion exchange membrane filters for the extraction and separation of magnesium from biological fluids.8 In this work, we found that the chromatographic separation/extraction of Mg using membrane filters exactly mimics column chromaography and that it was a more convenient and less time consuming procedure. The objective in this study was to evaluate calcium extraction using membrane filters for the possibility of yielding a calcium sample containing fewer potentially confounding cations, thereby improving automated data acquisition. EXPERIMENTAL SECTION Materials and Reagents. %aC03 (98.95 atom % enriched) was obtained from Oak Ridge National Laboratory (Oak Ridge, TN) and was used to enrich water and the biological matrices under investigation. Water standards were prepared using a Ca atomic spectral standard, 1mg/mL (Fisher Scientific, Fair Lawn,
NJ) . All storage vessels, glassware, and disposable pipet tips were soaked overnight in a 10% HN03 bath (Baker-Analyzed, J. T. Baker, Phillipsburg, NJ), rinsed four times with deionized water (Hydro Service and Supplies, Durham, NC), and air-dried under cover. Bio-Rex @io-Rad Laboratories,Richmond, CA) 25 mm syringe filters (AG 50W-X8, H+ and Chelex ion exchange resin filters) were used for ion exchange chromatography. Ultrex I1 HCl and HN03 U. T. Baker), high-purity HPLC grade methanol (Baxter Healthcare Corp., Burdick & Jackson Division, Muskegon, MI), and deionized water were used for all filter preconditioning and eluent solution preparation. All reagents used for the oxalate precipitation procedure have been previously describede6 Total Calcium Recovery. Three cation exchange and three Chelex filters were evaluated for calcium recovery efficiency for each matrix. Filters were conditioned prior to sample loading by the addition of 10 mL of methanol and then rinsed with 10 mL of deionized water. Filters were loaded with 1mL of water standard (1 mg/mL, pH =. 5), 1 mL of serum (0.091 mg/mL, pH 7), or 3 mL of urine (0.195 mg/mL, pH 5). Calcium was eluted with (8) Vieira, N. E.; Yergey, A. L.: Abrams, S. A. Anal. Biochem. 1994,218,9297.
Analytical Chemistry, Vol. 67, No. 22, November 15, 1995 4217
multiple 1 mL volumes of 6 M and 1 M HCl from the cation and Chelex filters, respectively.9 Three 1 mL eluent fractions were collected into acid-washed glass centrifuge tubes and were analyzed individually for total Ca using flame atomic absorption spectrophotometry (Perkin-Elmer Model 5000, Perkin-Elmer Corp., Norwalk, CT). The procedure for the oxalate precipitation has been described previously.6 Isotope Dilution Mass Spectrometry. Duplicate standard curves were prepared for mass spectrometric analysis by the addition of highly enriched (98.95%)%a tracer to aliquots of water (0.1 mg of Ca/mL), serum (0.091 mg of Ca/mL), and urine (0.045 mg of Ca/mL). Enrichments approximated lo%,20%, and 30% of the naturally occurring %a. Following a 24 h equilibration period, calcium was extracted from each prepared standard by precipitation and cation and Chelex ion exchange using onethird of the total volume for each extraction approach. The serum (but not the water or urine) ion exchange filters were washed with 10 mL of deionized water prior to elution of the calcium. Calcium was extracted with 2 mL of eluent for all standards. The elutions were heated and dried in an argon stream. Concentrated HN03 was added to the dried precipitates, and these solutions were dried again. All precipitates (oxalate, cation, Chelex) were then dissolved in 3%HN03 for mass spectrometric analysis. Isotope ratios were measured using the Fmigan MAT THQ Thermoquad mass spectrometer with electron multiplier detection in the dual filament mode. All ratios were measured relative to 48Ca,and fractionation correction was made using the 43Ca/4sCaratio. Automated Data Analysis Efficiency. Thermal ionization filaments were loaded with a 5 pL droplet of extracted calcium (cation exchange, Chelex, oxalate precipitated) from water, serum, and urine samples. The filaments were loaded onto the sample magazine (13 filamentdmagazine), which was then mounted in the THQ mass spectrometer. Eleven magazines were analyzed using the THQ automatic data analysis program. The total number of filaments automatically analyzed was expressed as a percentage of the total number of filaments loaded in the instrument for each matrix.
2
4> L
9 rp
0
0
2 ml Eluent
1
2
1
3 4 ml Eluent
3
6
5
7
100
g
80
>
8 E 6 0
9 rp
RESULTS AND DISCUSSION Total Recovery of Calcium. Figure 1 shows the cumulative percentage recovery of calcium extracted with each successive eluent fraction collected. It is evident that more than 3 mL of eluent will extract only minimal additionalcalcium. An additional 2 mL of eluent added to the serum cation filter only resulted in the recovery of an additional 3.6% of Ca loaded (total recovery of 77.6 f 2.1%). A more detailed comparison of recovery efficiencies between the three extraction methods is shown in Table 1. These results demonstrate that the cation filter is more efficient for extracting Ca from water and urine matrices, while Chelex is more efficient for serum samples. The different extraction efficiencies between the cation and Chelex filters observed for the three matrices can be explained on the basis of extraction equilibria and residence time (i.e., amount of time the sample remains in contact with the resin).l0 Chelex has a longer equilibration time than the cation resin. Therefore, the more viscous serum would remain in contact with the resin-embedded filter longer than the water or urine matrices, thereby resulting in the observations
reported here. In addition to increased residence time, the binding constant of Chelex is higher than that of most proteins.10 This property would augment the already efficient Chelex extraction of calcium from serum. Isotope Dilution Mass Spectrometry. The 44Ca/4sCaenrichments for the ion exchange standard curves were compared
(9) Strelow, F. W. E. Anal. Chem. 1960,32.1185-1188. (10) Wood, R. Bio-Rad Laboratories, Richmond, CA, unpublished data.
(11) Zar, J. H. Biostatistical Analysis. 2nd ed.; Prentice-Hall Inc.: Englewood Cliffs, NJ, 1984; Chapter 17.
4218 Analytical Chemistry, Vol. 67, No. 22, November 15, 7995
40
20
0 0
1
2
3
4
ml Eluent
Flgure 1. Total calcium recovely. Comparison of oxalate precipitation (dotted line), cation exchange ( n = 3) and Chelex ion exchange ( n = 3) calcium extraction from (A) water standard, 1 mg of Ca, (B) human serum, 0.091 mg of Ca, and (C) human urine, 0.585 mg of Ca. Data points are mean f standard deviation.
Table I.Recovery of Calcium from Water, Serum, and Urine Using Ion Exchange Membrane Fllters and Oxalate Preclpitation
extraction method cation exchange" Ca loaded, mg
recovery, %b Chelex exchange" Ca loaded, mg recovery, %b precipitationC Ca loaded, mg recovery, %
water
serum
urine
1.0 85.5 f 0.9
0.09 74.0 & 3.4
0.6 65.9 f 2.5
1.0 64.6 f 9.5
0.09 97.8 f 3.2
0.6 19.6 i 0.8
0.2 93.0
0.2 100.0
0.1 96.3
n = 3 filters; a total of 3 mL of eluent was collected for each filter. Data reported as mean f SD. n = 1.
Table 2. Linear Regression Analysis of the Enrichment of the 44CaPCaRatio of Ion Exchange Treated Standard Curves (Ordinate) versus Oxaiate-Precipitated Standard Curves (Abscissa)'
matrix
treatment
regression line
water
cation
y = -0.367 1.0162. y = -0.118 1.O00~ y = 0.041 0.9762. y = 0.573 0.9522 y = -0.628 1.017% y = -0.813 + 0.990~
serum urine
Chelex cation Chelex cation Chelex
+ + + + +
RS,b 0.020 0.026 0.025 0.014 0.026 0.031
Range of enrichment was 10%-30%. Duplicate standard curves were prepared for each matrix. RS, is relative standard error of the estimate of the regression line.
to those observed from the oxalate precipitation curves. The linear regression analyses are shown in Table 2. All curves were linear with slopes not different from 1 (p < O.oO1)ll and standard errors of the estimate less than 1. No isotopic fractionation was observed for the three matrices investigated. Automated Data Analysis Efficiency. Table 3 shows the results of the effect of extraction method on automatic data analysis for the three matrices investigated. Interruption of automated data analysis of oxalate-precipitated calcium is a problem observed principally with urine samples in this laboratory. In this experiment, urine samples yielded the lowest number of
Table 3. Effect of Extraction Method on Automatic Data Analysis Efficiency, Expressed as a Percentage of Total Number of Filaments Analyzed'
extraction method
water
serum
urine
oxalate precipitation cation exchange Chelex exchange
75% (12) 92% (12) 100%(12)
94% (18) 83% (18) 90% (20)
58% (12)
79% (14) 62% (16)
The number in parentheses indicates the total number of filaments analyzed.
filaments analyzed automatically. A 39%improvement in automatic filament analysis was observed for urine samples extracted with cation exchange filters. Water-extracted samples also showed improvements in automated data analysis when extracted with cation (21%) or Chelex (33%) ion exchange filters. No improvement was observed for the already very effective automated analysis of serum calcium samples. CONCLUSIONS
We report a method that allows for easy and rapid extraction and purification of calcium from biological fluids prior to isotopic enrichment analysis. The use of ion exchange filters resulted in an improvement in automated sample analysis for water and urine samples. Sample preparation for thermal ionization mass spectrometric analysis using ion exchange filters requires 1 day, as opposed to a 2 day procedure for oxalate precipitation.
Received for review May 22, 1995. Accepted August 21, 1995.a AC950484V
Abstract published in Adounce ACS Abstracts, October 1, 1995.
Analytical Chemistry, Vol. 67, No. 22, November 15, 1995
4219
