Capillary electrophoresis for detection of inherited disorders of purine and pyrimidine metabolism.
Purine and pyrimidine species were first measured a century ago [see Ref. (1) for review]. The first use of anion-exchange chromatography in 1949 (2) introduced separation techniques into this field. Since that time, many researchers have developed HPLC methods for research and diagnostic purposes (3). Currently HPLC is a dominant tool in diagnostic practice, although two-dimensional thin-layer chromatography (4) is also widely used.
Although isotachophoretic approaches were utilized for separation of purine and pyrimidine species in the past (5), they were never widely used in diagnostic practice. The narrow bore capillary electrophoretic separation technique, with its high separation efficiency, flexibility, and high sample throughput, can provide an excellent tool for the diagnosis of inborn errors of purine and pyrimidine metabolism (6-9).
We report here an optimized capillary electrophoretic method that enables diagnosis of purine and pyrimidine disorders. The method was tested on urine samples from healthy infants and patients with inherited defects of purine and pyrimidine metabolism.
Materials and Methods
All chemicals were of analytical reagent grade. Boric acid, sodium hydroxide, and [gamma]-cyclodextrin (CD) were purchased from Merck. Bases; nucleosides; 2-amino-2-methyl-1-propanol (AMP); 3-[cyclohexylamino]-2-hydroxy-1-propanesulfonic acid; [alpha], [beta], and [gamma]-CDs; and other chemicals were obtained from Sigma. Germanium (IV) oxide and sulfated [beta]-CD were from Aldrich. Deionized water (18 M[Omega]/cm) was used for preparation of all solutions.
SUBJECTS AND SAMPLES
Urine samples from healthy controls were from 200 children (Caucasian; 112 males and 88 females of the Czech Republic; age range, 1-15 years; mean, 4.8 years). The urines from patients (all from UK) were from persons with enzymatically demonstrated purine and pyrimidine enzyme deficiencies.
All samples were obtained as 40-[micro]L urine spots on filter paper and were dissolved in 150 [micro]L of deionized water, taken to dryness under stream of nitrogen, and redissolved in 40 [micro]L of deionized water.
Eight spot urine specimens from healthy children (creatinine range, 1.1-15.1 mmol/L; mean, 4.7 mmol/L) were used for measuring imprecision and reproducibility. The pooled urine sample that was used for determination of maximal injectable amounts was prepared by mixing equal volumes of redissolved eluates from filter papers of these samples. All samples were centrifuged (30008, 5 min) before loading.
CAPILLARY ELECTROPHORESIS APPARATUS AND CONDITIONS
All experiments were performed on a P/ACE 5510 with diode array detector (Beckman Instruments). The electrophoretic separations were carried out in an uncoated fused-silica capillary (75 [micro]m i.d. X 375 [micro]m o.d.; Polymicro Technologies). The capillary had an effective length of 40 cm (total length, 47 cm) and was operated at 25 or 35[degrees]C. Ultraviolet (UV) detection over the range 190-300 nm (cartridge detection window, 100 X 800 [micro]m) was used. The data rate of the detector was set at 2 Hz for analyses performed at 10 kV and 8 Hz for analyses performed at 30 kV, respectively. Sample was loaded by low-pressure injection (0.5 psi). Borate buffers were prepared from boric acid, sodium dodecyl sulfate (SDS) was added, and the solution was adjusted with 500 g/L NaOH or AMP to the appropriate pH. At the beginning of each working day, the capillary was washed with water, 0.1 mol/L NaOH, water, and separation buffer for 5 min; between runs, it was washed with 0.1 mol/L NaOH for 0.5 min and running buffer for 1 min. The analyses were run at a constant voltage, using a ramp for 0.5 min. These standard conditions were used for all experiments.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Fourteen bases and nucleosides (the key diagnostic metabolites) together with three major urinary UV-absorbing constituents (urea, creatinine, and hippuric acid) were taken as the target group of compounds. A mixture of target compounds was prepared by dissolving compounds in deionized water at a concentration of 300 [micro]mol/L. The mixture and pooled normal urine were analyzed with background electrolytes (BGEs) at pH 7-10, with and without the addition of micellar additive (SDS) at 25 and 35[degrees]C, respectively. All experiments performed at pH <9 did not allow substantial separation of the mixture of target compounds (data not shown). Comparison of assays performed with and without SDS revealed that zone electrophoretic rather than a micellar separation mechanism prevailed for the separation of purine and pyrimidine species. However, for the separation of the pooled urine sample, micellar systems resolved a higher number of peaks. Higher reproducibility can be also expected in a system utilizing surfactants because it prevents interactions of urine proteins with the capillary wall. As well as the pH, the BGE composition substantially affected the separation. Fig. 1 shows that the borates and germanates provided good resolution of ribonucleosides and deoxyribonucleosides via their well-known complexation with cis-diol species. Separations performed in 60 mmol/L borate-[Na.sup.+]-80 mmol/L SDS buffer with a pH >9.5 gave substantially better resolution of the mixture (Fig. 2). From these assays, it can be concluded that separation (although incomplete) improved with increasing pH. At pH >10, almost complete resolution was achieved, but the very high electrophoretic mobility of orotic acid (OA) prolonged the separation. Moreover, at pH >9.7, run-to-run reproducibility was reduced (data not shown) as a consequence of a lowering of the buffer capacity because the p[K.sub.a] of boric acid is 9.23. Addition of organic modifiers (50-150 mL/L acetonitrile and methanol) to the BGEs did not affect the resolution and lead to peak tailing (data not shown). CDs previously had been used successfully for the separation of adenosine derivatives (10). The influence of [alpha], [beta], [gamma], and sulfo-[beta]-CD at concentrations of 25, 10, 25, and 10 mmol/L, respectively in 60 mmol/L borate-[Na.sup.+]-80 mmol/L SDS, pH 9.6, was tested. Of the CDs tried, only [beta]-CD slightly affected the selectivity (data not shown), but a decrease of run-to run reproducibility was observed. [gamma]-CD led to peak distortion, and [alpha]-CD worsened the separation [mean CV of effective mobility (n = 10) was 0.84% in BGE without CD, 0.99% with [alpha]-CD, 1.6% with [beta]-CD, 0.87% with [gamma]-CD, and 4.8% with sulfo-[beta]-CD, respectively].
To stabilize migration times and effect a better separation, different counterions of the BGEs were sought. Excellent resolution was achieved by the use of 60 mmol/L borate-AMP-80 mmol/L SDS buffer, pH 9.6, at 35[degrees]C at 10-30 kV (Fig. 3). Here the separation was influenced by the counterion (compare assays in Figs. 2 and 3 performed at the same pH). This buffer is a double buffering system (the p[K.sub.b] of AMP is 9.72) and, therefore, higher run-to-run reproducibility of mobilities was expected. The borate-AMP buffer had a higher buffering capacity and also lower conductivity (lower Joule heating) in comparison with the borate-[Na.sup.+] buffer (60 mmol/L borate-[Na.sup.+]: capacity, 32.2 mmol/pH; conductivity, 257.8 m5/m; 60 mmol/L borate-AMP: capacity, 66.9 mmol/pH; conductivity 154.9 m5/m at pH 9.6).
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
DIAGNOSTIC METABOLITES AND INTERFERENCES
Using the final separation conditions, we measured the migration and spectral properties of the compounds of interest in diagnosing inherited metabolic disorders, common artifacts from medication (3), and several other UV-absorbing compounds (Table 1). The pure compounds dissolved in deionized water were analyzed, with the exception of succinylaminoimidazole carboxamide riboside (SAICAR) and succinyladenosine (SAR), for which diagnostic metabolites are not commercially available (see Results).
SEPARATION EFFICIENCY AND INJECTION VOLUME
Separation efficiency in 60 mmol/L borate-AMP-80 mmol/L SDS reached 220 000 theoretical plates/m for the mixture of target compounds and 141000 theoretical plates/m for the pooled urine sample with added target compounds. Borate buffers can tolerate higher injection volumes because of a pronounced transient isotachophoretic phenomenon. [Borate ion can act as the terminating ion, and fast-migrating anions from the sample, e.g., chlorides from urine, act as leading ions (11).] The maximal injectable amount of sample was determined. The pooled urine sample (see "SUBJECTS AND SAMPLES") was injected in increasing amounts for up to 15 s, and the separation efficiency was calculated. The efficiencies observed (X 1000 theoretical plates/m) were 132 for 3 s (24 nL), 120 for 5 s (40 nL), 114 for 7 s (55 nL), 113 for 9 s (71 nL), 112 for 13 s (102 nL), and 104 for 15 s (119 nL).
[FIGURE 6 OMITTED]
LIMIT OF DETECTION AND LINEARITY
Using 15-s injections (6.7% of the total capillary volume injected with the sample), we determined the limit of detection for compounds of interest at 200 nm (Table 2). We tested the linearity by analyzing 10 calibration solutions in the concentration range 5-500 [micro]mol/L. The method was linear (r >0.99) for all compounds of interest (Table 2).
The imprecision of the method was tested by assaying eight samples of healthy volunteers with added 2,8-dihydroxyadenine (DHA; not present in normal urine) at three concentrations for 20 days. The within-run CVs were 3.8% for 15 [micro]mol/L, 2.6% for 40 [micro]mol/L, and 1.8% for 90 [micro]mol/L. The intraday CVs were 4.2% for 15 [micro]mol/L, 3.1% for 40 [micro]mol/L, and 2.0% for 90 [micro]mol/L. The interday CVs were 4.6% for 10 [micro]mol/L, 3.6% for 40 [micro]mol/L, and 3.2% for 90 [micro]mol/L.
[FIGURE 7 OMITTED]
The reproducibility of the effective mobilities was measured on the mixture of target compounds [run-to-run CV, 0.32% (n = 20); day-to-day CV, 2.4% (n = 20)]. Because sample composition can affect the separation in capillary electrophoresis (CE), the sample-to-sample reproducibility of effective mobilities was measured on eight urine samples from healthy volunteers with added calibrators (CV, 1.5%).
RECOVERY AND STABILITY
The recovery and stability of the compounds was measured using urine sample spot extracts supplemented with the mixture of target compounds (added concentrations, 30 and 100 [micro]mol/L). The dried urine spot was extracted on the day of preparation and also after 2 days at room temperature. The mean recovery ([+ or -] SD) was 94.8% [+ or -] 6.9% for 30 [micro]mol/L and 96.8% [+ or -] 4.9% for 100 [micro]mol/L. The mean ([+ or -] SD) stability of the compounds was 95.9% [+ or -] 3.2%.
[FIGURE 8 OMITTED]
Results and Discussion
Two hundred urine samples from healthy children were analyzed by this method. No interfering compounds that can substantially confound the analysis were observed. A typical electropherogram of a urine sample from a healthy infant is shown in Fig. 4. Urine from healthy infants contained dominant peaks of uric acid (UA), urea, creatinine, and hippuric acid and usually various amounts of hypoxanthine (HX), xanthine (X), 7-methylguanine, pseudouridine, and uridine. As pointed out by many authors [e.g., Simmonds et al. (3)], purine excretion varies considerably because of varying dietary purine intake, and local reference values should be determined.
The usefulness of the method for diagnostic purposes was demonstrated on urine samples from patients suffering from inherited disorders of purine and pyrimidine metabolism. Electropherograms are presented at 205 nm and the optimal wavelength for each particular disease. In all samples analyzed, the key diagnostic metabolites were easily identified by migration times and spectral fit.
[FIGURE 9 OMITTED]
In the sample from a patient suffering from adenylosuccinate lyase (EC 126.96.36.199) deficiency, SAICAR and SAR were identified by expected migration behavior (species contain highly charged succinate functional group) and spectra (Fig. 5) (6,12). The analysis of urine from a patient with adenine phosphoribosyl transferase (EC 188.8.131.52) deficiency allowed easy identification of a DHA peak (Fig. 6). Analysis of the urine from a patient with adenosine deaminase (EC 184.108.40.206) deficiency demonstrated other peaks with spectra similar to adenosine (indicated by an asterisk in the electropherogram) in addition to the diagnostic metabolite, deoxyadenosine (Fig. 7). These compounds could be methylated and incompletely characterized AR derivatives (13-15). The electropherogram of urine from a patient with dihydropyrimidine dehydrogenase (EC 220.127.116.11) deficiency revealed well-separated diagnostic peaks of uracil and thymine (Fig. 8). Peaks of inosine, guanosine, and their deoxyribosides were well resolved in the sample from a patient with purine nucleoside phosphorylase (EC 18.104.22.168) deficiency (Fig. 9). The sample from a patient with xanthine oxidase (EC 22.214.171.124) deficiency revealed only a dominant peak of X migrating after a small peak of UA (Fig. 10). In the analysis of urine from patients with orotic acidurias caused by orotate phosphoribosyltransferase (EC 126.96.36.199) and ornithine transcarbamylase (EC 188.8.131.52) deficiency (data not shown), an easily identifiable but relatively broad peak of OA was observed. These patients excrete very high amounts of this compound, and the comparable opposite mobilities of the electroosmotic flow and OA slows movement of OA through the detection window. Patients with orotic acidurias can be verified easily by an alternative CE diagnostic method (9). In the sample from a patient with hypoxanthine phosphoribosyl transferase (EC 184.108.40.206) deficiency, increases in only UA and HX were observed (data not shown).
[FIGURE 10 OMITTED]
In conclusion, the method reported enables a simple, fast, and efficient diagnosis of inherited purine and pyrimidine enzyme deficiencies and also is potentially applicable to the screening of other inherited disorders connected with increased excretion of UV-absorbing compounds. The conditions used allowed separation of all diagnostic metabolites from major urinary constituents. The method is both efficient (separation efficiency, 220 000 theoretical plates/m) and sensitive enough for diagnostic purposes. The migration time reproducibility was achieved by the use of a double-buffering 60 mmol/L borate-AMP-80 mmol/L SDS (pH 9.6) BGE. Analyses performed by the proposed method are faster than currently performed HPLC assays (3) (total analysis time, 3 min for CE vs 30 min for HPLC). Separation efficiency is also substantially higher for CE compared with HPLC (220 000 theoretical plates/m for CE vs 5000 theoretical plates/m for HPLC). From a practical point of view, the substantially lower cost for CE fused-silica capillaries compared with reversed-phase HPLC columns makes this approach advantageous.
This work was supported by Grant 3439-3 from the Ministry of Health of the Czech Republic and partly by Grant VS 96021 from the Ministry of Education and Youth and Sport of the Czech Republic.
(1.) Hitchings GH. Uric acid: chemistry and synthesis. In: Kelley WN, Weiner IM, eds. Uric acid. Berlin: Springer-Verlag, 1978:1-20.
(2.) Cohn WE. Separation of purine and pyrimidine bases and of nucleotides by anion exchange. Science 1949;109:377-8.
(3.) Simmonds HA, Duley JA, Davies PM. Analysis of purines and pyrimidines in blood, urine, and other physiological fluids. In: Hommes FA, ed. Techniques in diagnostic human biochemical genetics: a laboratory manual. New York: Wiley-Liss, 1991:397-424.
(4.) Van Gennip AH, Van Noordeburg-Huistra DY, De Bree PK, Wadman SK. Two-dimensional thin-layer chromatography for the screening of disorders of purine and pyrimidine metabolism. Clin Chim Acta 1978;86:7-20.
(5.) Bruchelt G, Niethammer D, Schmidt KH. Isotachophoresis of nucleic acid constituents. J Chromatogr 1993;618:57-77.
(6.) Gross M, Gathof BS, Kolle P, Gresser U. Capillary electrophoresis for screening of adenylosuccinate lyase deficiency. Electrophoresis 1995;16:1927-9.
(7.) Bory C, Chantin C, Boulieu R. Comparison of capillary electrophoretic and liquid chromatographic determination of hypoxanthine and xanthine for the diagnosis of xanthinuria. J Chromatogr A 1996;730:329-31.
(8.) Sevcik J, Adam T, Mazacova H. A fast and simple screening method for detection of 2,8-dihydroxyadenine urolithiasis by capillary zone electrophoresis. Clin Chim Acta 1996;245:85-92.
(9.) Sevcik J, Adam T, Sazel V. A fast and simple screening method for detection of orotic aciduria by capillary zone electrophoresis. Clin Chim Acta 1997;259:73-81.
(10.) Bartak P, Sevcik J, Adam T, Friedecky D, Lemr K, Stransky Z. Study of cytokinin separation using capillary electrophoresis with cyclodextrin additives. J Chromatogr A 1998;818:231-8.
(11.) Adam T, Sevcik J, Svagera Z, Fairbanks LD, Bartak P. Determination of adenosine deaminase activity in human erythrocytes by on column capillary isotachophoresis-capillary zone electrophoresis at the presence of electroosmotic flow. Electrophoresis 1999;20: 564-8.
(12.) Jaeken J, van den Berghe G. An infantile autistic syndrome characterised by the presence of succinylpurines in body fluids. Lancet 1984;2:1058-61.
(13.) Simmonds HA, Sahota A, Potter CF, Cameron JS. Purine metabolism and immunodeficiency. Urinary purine excretion as a diagnostic screening test in adenosine deaminase and purine nucleoside phosphorylase deficiency. Clin Sci Mol Med 1978;54:579-84.
(14.) Mills GC, Goldblum RM, Newkirk KE, Schmalstieg FC. Urinary excretion of purines, purine nucleosides, and pseudouridine in adenosine deaminase deficiency. Biochem Med 1978;20:18099.
(15.) Hirschhorn R, Ratech H, Rubinstein A, Papageorgiou P, Kesenvala H, Gelfand E, Roegner Maniscalo V. Increased excretion of modified adenine nucleosides by children with adenosine deaminase deficiency. Pediatr Res 1982;16:362-9.
TOMAS ADAM [1, 3] *, DAVID FRIEDECKY , LYNETTE D. FAIRBANKS , JURAJ SEVCIK [2, 3], and PETR BARTAK [2, 3]
 Laboratory for Inherited Metabolic Disorders, Department of Clinical Biochemistry, Medical Hospital, I. P. Pavlova 6, 775 20 Olomouc, Czech Republic.  Department of Analytical Chemistry, Palacky University, Trida Svobody S, 772 00 Olomouc, Czech Republic.  Laboratory of Bioanalytical Research, Palacky University, Trida Svobody 88, 771 26 Olomouc, Czech Republic.  Purine Research Laboratory, Guy's and St. Thomas's Medical and Dental School, London Bridge, London SE1 9RT, UK.
 Nonstandard abbreviations: CD, cyclodextrin; AMP, 2-amino-2-methyl-1-propanol; UV, ultraviolet; SDS, sodium dodecyl sulfate; BGE, background electrolyte; OA, orotic acid; SAICAR, succinylaminoimidazole carboxamide riboside; SAR, succinyladenosine; DHA, 2,8-dihydroxyadenine; CE, capillary electrophoresis; UA, uric acid; HX, hypoxanthine; and X, xanthine.
* Address correspondence to this author at: Laboratory for Inherited Metabolic Disorders, Medical Hospital, I. P. Pavlova 6, 775 20 Olomouc, Czech Republic. Fax 420-68-5416555; e-mail firstname.lastname@example.org.
Received August 4, 1999; accepted September 8, 1999.
Table 1. Characteristics of tested compounds. Effective Metabolite Absorbance mobility, maxima, (a) [10.sup.-9] [m.sup.2] x nm [V.sup.-1] x [s.sup.-1] 0.0 Urea ~205 -3.4 Creatinine 234 -6.0 Homocarnosine ~233 -8.7 2'-Deoxyadenosine 259 -9.4 Carnosine ~234 -10.3 Thymine 263 -10.8 Allopurinol 253 -11.3 7-Methylguanine 281 -11.7 Adenine 260 -12.7 1-Methylhistidine ~235 -12.9 3-Methylhistidine ~237 -13.2 2'-Deoxyguanosine 252 -13.2 3,7-Dimethylxanthine 274 -13.5 1,3-Dimethylxanthine 274 -15.2 Uracil 259 -15.4 Histidine ~234 -15.7 2'-Deoxyinosine 249 -18 Tyrosine ~250 -18.1 Guanine 246, 272 -18.3 Adenosine 259 -18.8 Cytidine 269 -19.3 1,7-Dimethylxanthine 276 -19.4 Phenylalanine ~220 -21.7 Hypoxanthine 254 -23.2 Argininosuccinic acid ~220 -23.2 Guanosine 252 -23.6 Allopurinol-1-riboside 259 -23.7 Arginine ~218 -24.4 2,8-Dihydroxyadenine 300 -25.2 Homovanillic acid ~214 -25.4 1,3-Dimethyluric acid 296 -25.4 3,7-Dimethyluric acid 297 -25.5 4-Hydroxyphenyllactic acid ~237 -25.6 Uridine 262 -25.8 7-Methylxanthine 289 -25.8 Hippuric acid 225 -25.8 Vanillylmandelic acid ~245 -26 3-Methylxanthine 274 -26.2 Pseudouridine 274 -26.2 3-Methyluric acid 297 -26.4 1-Methyluric acid 238, 292 -26.5 7-Methyluric acid 239, 294 -27.1 Inosine 250 -27.4 1-Methylxanthine 241, 276 -27.8 4-Hydroxyphenylacetic acid ~238 -28.3 Uric acid 235, 290 -28.5 Oxypurinol 241 -28.8 Xanthine 240, 275 -29.6 5-Oxoproline ~223 -35.2 SAR 263 -36.2 SAICAR 263 -36.9 Orotic acid 280 (a) Absorbance maxima in the range 220-300 nm. For compounds not exhibiting absorbance maxima in this range, upper value of absorbance is given (e.g., ~220). Table 2. Linearity and limit of detection. Metabolite Linear regression equations (a) Adenine y = 10.18(0.06)x+ 2.25(5.28) Adenosine y = 8.50(0.03)x- 4.43(7.17) 2'-Deoxyadenosine y = 15.15(0.05)x- 6.32(12.30) 2'-Deoxyguanosine y = 12.01(0.08)x+ 6.13(20.98) 2'-Deoxyinosine y = 7.90(0.02)x- 0.90(2.65) 2,8-Dihydroxyadenine y = 5.05(0.02)x- 3.11(6.32) Guanosine y = 15.02(0.01)x- 4.98(3.20) Hypoxanthine y = 14.13(0.03)x+ 0.79(8.86) Inosine y = 14.84(0.07)x+ 0.28(2.58) Orotic acid y = 9.54(0.05)x- 3.42(11.97) Thymine y = 14.60(0.05)x+ 3.82(5.66) Uracil y = 11.21(0.04)x+ 1.77(7.94) Uric acid y = 12.31(0.06)x+ 5.10(6.13) Xanthine y = 16.54(0.08)x- 7.07(13.45) Metabolite r LOD (b) Adenine 0.9984 1.43 Adenosine 0.9998 1.8 2'-Deoxyadenosine 0.9985 1.3 2'-Deoxyguanosine 0.9962 0.98 2'-Deoxyinosine 0.9953 1.34 2,8-Dihydroxyadenine 0.9923 4.28 Guanosine 0.9995 0.97 Hypoxanthine 0.9956 1.26 Inosine 0.9974 0.97 Orotic acid 0.9958 4.16 Thymine 0.9989 0.83 Uracil 0.9966 1.11 Uric acid 0.9971 1.22 Xanthine 0.9992 0.85 (a) x in [micro] mol/L; range, 5-500 [micro] mol/L; y, corrected peak area in arbitrary units. (b) Limit of detection at signal-to-noise ratio = 3; absorbance wavelength = 200 nm.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Molecular Diagnostics and Genetics|
|Author:||Adam, Tomas; Friedecky, David; Fairbanks, Lynette D.; Sevcik, Juraj; Bartak, Petr|
|Date:||Dec 1, 1999|
|Previous Article:||Ribonuclease-resistant RNA controls (armored RNA) for reverse transcription-PCR, branched DNA, and genotyping assays for hepatitis C virus.|
|Next Article:||Application of a thermodynamic nearest-neighbor model to estimate nucleic acid stability and optimize probe design: prediction of melting points of...|