Enzymatic determination of bicarbonate in serum by flow injection analysis

ELSEVIER

Clinica Chimica Acta 235 (1995) 169-177

Enzymatic determination of bicarbonate in serum by flow injection analysis R. Quiles a, J.M. Fernfindez-Romero b, M.D. Luque de Castro *b aDepartment of Biochemistry, Hospital Virgen de la Salud, E-45005, Toledo, Spain bDepartment of Analytical Chemistry, Faculty of Sciences. University of C6rdoba, E-14004, C6rdoba, Spain Received 2 September 1994; revision received 6 January 1995; accepted 16 January 1995

Abstract An automated method for the determination of bicarbonate in human serum based on the enzymatic reaction between the analyte and phospho(enol)pyruvate (PEP) in the presence of PEP carboxylase is proposed. The analytical reaction is coupled with a derivatization reaction in which the NADH consumed is fluorimetrically monitored (Xex= 340 nm, ~kem = 460 nm). A stopped-flow/flow-injection approach is used in which the enzymes (PEP carboxylase and malate dehydrogenase) are immobilized on controlled-pore glass. The linear determination range is between 25 and 300 mmol/l (r 2 = 0.9973). The %C.V. for the within- and betweenrun studies, performed at three concentration levels, ranges between 1.0 and 3.6% and the sampling frequency is 20 per h.

Keywords: Bicarbonate; Enzymatic; Flow injection analysis; Stopped-flow; Serum

1. Introduction One of the important buffer systems that maintains the pH of the blood is the carbonic acid-bicarbonate buffer system. Total CO 2 (i.e. bicarbonate plus dissolved CO2) measurements are usually employed together with other clinical and laboratory information (arterial pH and Pc02) for the evaluation of acid-base disorders. Total CO 2 is generally increased in respiratory acidosis, metabolic alkalosis and ex* Corresponding author. 0009-8981/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0009-8981 (95)06026-A

170

R. Quiles et al./ Clinica Chimica Acta 235 (1995) 169-177

cessive alkali intake. CO2 is decreased in compensated respiratory alkalosis, metabolic acidosis and renal disorders where H + ions cannot be excreted [11. Several methods for the determination of bicarbonate have been reported in the last decade. Most of the procedures for the determination of this analyte are not based on a derivatization reaction, but on potentiometric measurements by using ion selective electrodes (ISEs) [2-4] or ion chromatography with conductimetric detection [5]. Some of the proposed methods are based on the development of a chemical reaction and involve acidification of the sample followed by measurement of the volume [6] or the pressure [7] of the evolved gas; others entail spectrophotometric monitoring by the use of acid-base indicators such as phenolphthalein [8]. These procedures are either tedious or time-consuming, or require large volumes of sample or special instrumentation. Enzymatic methods based on catalysis by phospho(enol)pyruvate (PEP) carboxylase and colorimetric detection [9], or on coupling this reaction with a second step catalysed by malic dehydrogenase and kinetic [10,11] or steady-state [12] monitoring of the oxidized NADH have been reported. The aim of this work was to develop an automatic, simple, reliable and inexpensive fluorimetric method based on the flow-injection (FI) principle for the determination of total carbon dioxide in serum based on the enzymatic reaction between PEP and bicarbonate in the presence of PEP-carboxylase (EC 4.1.1.31, orthophosphate:oxaloacetate carboxy-lyase). The overall biochemical system consists of two steps; in the first the analytical reaction is as follows: PEP-carboxylase

PEP + HCO 3-

- ~

PO43-+ oxaloacetate

In the second step the oxaloacetate formed is reduced with the reduced /3nicotinamide adenine dinucleotide (NADH) in the presence of malic dehydrogenase (MDH, EC 1.1.1.37, L-malate: NAD + oxidoreductase), according to the following reaction: MDH

oxaloacetate + NADH

" malate + NAD ÷

The bicarbonate concentration is proportional to the NADH consumed, which is monitored fluorimetrically at ~x = 340 nm, ~kem ---- 460 nm. 2. Materials and methods

2.1. Apparatus A Perkin-Elmer 204 spectrofluorimeter fitted with an 18 ttl flow-cell and interfaced to a Perkin-Elmer 56 recorder and a Perkin-Elmer VDR-3 digital voltimeter; a Gilson Minipuls-2 four-channel peristaltic pump furnished with a rate selector; a Rheodyne 5041 injection valve and teflon tubing of 0.5 mm i.d. were also used. A Julabo-5 recirculating thermostat (Juchheim Labortechnic D-7633, Seelbach, Germany) and a Beckman ~-72 pH meter were used to implement the method. An ABL-

R. Quiles et al./ Clinica Chimica Acta 235 (1995) 169-177

171

510 Blood Gas Analyser (Radiometer, Copenhagen, Denmark) based on potentiometric pH and CO2 measurements via glass and silicon-membrane electrodes, respectively, were used in order to compare the results. 2.2. Reagents and standards

All reagents were of analytical reagent grade. All solutions were prepared in doubly distilled water of high purity obtained from a Millipore Milli-RO system. An aqueous solution, containing 100 mmol/l tris(hydroxymethyl)aminomethane (Merck 8382) adjusted to pH 8.0, 10 mmol/l magnesium sulphate (Merck 5833) and 2 mmol/l N-acetyl-cysteine (Sigma A-7250), was used as carrier solution. Reagent 1 was a solution containing 15 mmol/l PEP mono(dicyclohexyl) ammonium salt (Sigma P3637), 0.5 mM B-nicotinamide adenine dinucleotide, reduced form (NADH) (Sigma N-8129) in the above-mentioned carrier. The enzymes used were PEP carboxylase (from corn suspended in 2.4 mol/l ammonium sulphate solution, P-2023, Sigma) and MDH (from bovine heart, M-9004, Sigma). An aqueous solution of sodium bicarbonate (Merck 6329) was used as standard solution after suitable dilution in the carrier. The reagents used for immobilization of the enzymes were 3-aminopropyltriethoxysilane (11339-5, Aldrich), glutaraldehyde (820603, Merck) and controlledpore glass 240 (120-200 mesh) (Electronucleonic, Inc. Fairfield, NY). 2.3. Samples

A 1:250 dilution of the samples prior their introduction into the flow injection system was made in the carrier containing 10 mmol/l oxamic acid (O-2751, Sigma). 2.4. Immobilization and storage o f the enzymes

Phospho(enol)pyruvate carboxylase and malic dehydrogenase were immobilized separately on activated CPG by the procedure of Masoom and Townshend [13]. Glass tubes of different lengths and 1.5 mm i.d. were then packed with each supportenzyme conjugate and stored at 4°C in the appropriate solutions: 10 mmol/1 potassium dihydrogen phosphate buffer (4871, Merck) pH 7.0, 2.4 mol/1 ammonium sulphate (1217, Merck) and 5 mmol/I dl-dithiothreitol (D-0632, Sigma) for the PEP-carboxylase reactor (IMERI) and 100 mmol/l potassium dihydrogen phosphate buffer (pH 7.0) for the malic dehydrogenase reactor (IMER2). Under these conditions both enzyme reactors kept their activity at least for 3 weeks. 2.5. Flow injection manifold and procedure

Fig. 1 depicts the hydrodynamic FI system used. The configuration consists of a peristaltic pump (P) which propels the reagent streams through the channels; 200/~1 of sample are injected into the carrier after a 1:250 dilution, then merging at 'a' with a stream of reagent (R 0 that contains phospho(enoi)pyruvate, the substrate for the PEP-carboxylase biocatalysis. When the plug arrives to IMER l (immobilized PEP carboxylase), the flow is halted and the biochemical reaction takes place. After an appropriate stop time, the flow starts again and the indicator reaction takes place along IMER 2. Finally, the sample reaches the flow cell and provides the analytical signal. Both IMERs as well as the previous open reactor (L0 are thermostatted at 37°C.

172

R. Quiles et al,/ Clinica Chimica Acta 235 (1995) 169-177

S

( ~

C

' Lo L1

I

R1

IMER~IMER2

I.

I

J

h

II W

P Fig. 1. Flow injection manifold. S, sample; C, carrier; R], reagent; P, peristaltic pump; IV, injection valve; a, merging point; 1MER] and 1MER 2, immobilized enzyme reactors (phospho(enol)pyruvate carboxylase and lactic dehydrogenase, respectively); L], open reactor; D, spectrofluorimetric detector; W, waste; T, thermostat.

3. Results and discussion

3.1. Study of variables The variables affecting the system were studied by the univariate method. These variables can be classified into chemical, physical and hydrodynamic. Table 1 shows the ranges over which each variable was studied and the optimum values found. Table 1 Optimization of variables Variable

Range studied

Physical Temperature (°C)

Optimum value

25-40

37

Chemical Tris-HCl (mmol/I) pH PEP (mmol/l) NADH (mmol/l) N-acetyl cysteine (mmoi/I) Oxamic acid (mmol/l)

50-500 6.5-8.5 1-20 0.1-1.5 0.5-4 0.5-20

100 8.0 15 0.6 2 10

0.5-2.0 50-300 45-90 20-90 50-200

1.3 200 60 60 50

FI Flow-rate (ml/min) Injection volume (t,I) Delay time (s) Stop time (s) Length of L 1 (cm) Length of IMER (cm) PEP carboxylase MDH

0.1-2 0.1-2

0.5 1

R. Quiles et at/Clinica Chimica Acta 235 (1995) 169-177

173

Increased temperature favourably influenced the analytical signal; however, temperatures between 37 and 40°C provided the greatest signal. Above 40°C the signal decreased, probably owing to denaturation of the biocatalyst. A buffer containing 100 mmol/1 Tris-HCl was selected to adjust the medium to pH 8.0. Lower values decreased the analytical signal and more basic pH gave rise to dramatic loss of the PEP-caboxylase activity [14]. The carrier solution contained 2 mmol/l of N-acetyl-cysteine in order to stabilize the PEP-carboxylase reactor via sulphydryl group reactivation [15]. A concentration of PEP of 15 mmol/l was chosen as optimal for the development of the analytical signal. Higher concentrations of substrate slightly improved the biocatalyzed reaction, whereas lower concentrations did not result in an appropriate development. The concentration of NADH selected as optimal was 0.6 mmol/l. Higher NADH concentrations caused decrease of the analytical signal, probably due to inhibition of MDH by coenzyme saturation. The NADH-coupled enzymatic reaction can yield erroneous results (about 6% higher than the true value) due to reduction of pyruvate in the presence of lactic dehydrogenase (LDH), both present in serum. Oxamic acid (10 mmol/l) was added to the sample solution as an LDH inhibitor [16]. A flow-rate of 1.3 ml/min was selected as optimum after checking that this parameter did not have a significant influence on the analytical signal when the stoppedflow mode is used. The length of the open reactor L~ was 50 cm, which was sufficient to obtain reproducible mixing after the merging point 'a' (see Fig. 1). With longer open reactors and lower flow rates neither the reproducibility nor the sensitivity improved, but the sampling frequency decreased. The optimal lengths of the enzyme reactors were 1.0 and 0.5 cm for IMER~ and IMER2, respectively. An injection volume of 200 #1 was chosen in order to obtain the best analytical signal as above this value the signal remained almost constant. The influence of the delay and stop times was also studied: 60 s provided the higher analytical signal for both. 3.2. Features o f the method

The calibration graph was constructed using the optimum values of the variables studied above. A series of eight standard solutions with bicarbonate concentrations of 0, 5, 10, 50, 100, 200, 350 and 500 mmol/l were injected in triplicate into the FI manifold. A linear range between 25 and 300 mmol/i was obtained according to the equation: Fluorescence intensity = 181.8 + 0.736 [HCO3-](mmol/l); n = 7; r 2 = 0.9973. The sampling frequency of the method was estimated at 20 per h. The linearity of the method was tested by the procedure of Kroll and Emancipator [17]. Both the dimensional non-linearity and the relative non-linearity were calculated for serum samples in the range of 0, 5, 15, 30, 45 and 60 mmol/l after a 1:250 (v/v) dilution. An excellent linearity of the proposed method (P of the coefficients < 0.05) and a relative value smaller than the limit accepted by the College of the American Pathologists [18] were obtained.

R. Quileset al./ Clinica ChimicaActa 235 (1995) 169-177

174

No signal was obtained from native fluorescence when serum samples, after a 1:100 dilution, were injected into the F1 manifold if PEP was removed from R~; blank measurements are therefore unnecessary at the dilution indicated in the proposed method (1:250). The precision of the method was checked by using three serum pool samples from a clinical laboratory which contained low, medium and high concentrations of bicarbonate (about 14.6, 28.4 and 56.4 mmol/l, respectively). Each aliquot was analyzed, both in a single run and for 11 days in the within- and between-run studies, respectively. The results obtained are summarized in Table 2.

3.3. Effect of potential interferents The study of potential interferents was aimed at those endogenous components present in serum that can cause disturbance in the method. Each potential interferent was added to the samples at higher concentrations than it is usually present in serum and the signal obtained was compared with that corresponding to samples in the absence of the added species. Haemoglobin did not interfere at a 10:1 weight ratio to bicarbonate. There was no interference from bilirubin at a 60:1 ratio. Pyruvate interference was specially investigated because of potential interference in the presence of endogenous LDH [16]. A 5:1 pyruvate-bicarbonate molar ratio caused a positive error of 2.5% when an additional concentration of exogenous LDH (about 3000 units/l) was added to the samples. 3.4. Comparison of the method The correlation parameters for comparison of 26 serum samples by the proposed method (samples injected in triplicate) and the Radiometer ABL-510 Blood Gas Analyser are shown in Fig. 2. 3.5. Application of the method to actual samples The proposed method was validated by applying it to the determination of bicarbonate in eight serum samples from both healthy and sick individuals. The analyte concentration in each sample was previously determined by the ABL-510 Blood Gas Table 2 Reproducibility of the proposed flow-injectionmethod

Within-run (n Level I Level I1 Level 111

[bicarbonate] (mmol/I)

S.D.

%C.V.

14.14 28.55 56.08

0.400 0.600 0.586

2.8 2.1 1.0

14.30 28.20 56.05

0.513 0.706 0.743

3.58 2.49 1.32

= 11)

l~tween-run (n Level 1 Level 11 Level Ili

= ! 1)

175

R. Quiles et al./ Clinica Chimica Acta 235 (1995) 169-177

.01 E ~o 0 0

.z.

4~

0

"

eI11

E _

2O

20

40

60

ABL-510 method [HC03-], mM Fig. 2. Correlation between the proposed method (FI) and the Radiometer ABL-510 method (n = 26, y = 1,008x-0.074).

Analyser. A 1:250 sample dilution was required to fit the c o n c e n t r a t i o n of the analyte to the linear p o r t i o n of the calibration graph. After determining the concentration of b i c a r b o n a t e the recovery was evaluated by adding 25 #mol/1 of standard to each sample. The results obtained are listed in Table 3. Table 3 Application of the flow injection (FI) method to actual samples Sample

1 2 3 4 5 6 8 9 10

ABL-150 method

FI method

Found (mmol/l)

Found (mmol/l)

Recoverya (%)

25.48 35.25 39.00 6.00 9.40 24.1 19.5 24.2 23.3

25.38 35.00 38.00 6.50 10.20 28.60 29.40 26.30 23.70

99.8 99 96 102 103 101 97 108 10|

aAddition of 25 mmol/l HCO3-.

176

R. Quiles et al./ Clinica Chimica Acta 235 (1995) 169-177

4. Conclusions The features of flow injection analysis have enabled the development of an enzymatic fixed-time kinetic method for the determination of bicarbonate which does not require either completion of the reaction [12] or addition of inhibitors in order to decrease the reaction rate and fit it to pseudo first-order kinetics [10,11]. The comparison of the results provided by the proposed method with those from well-established analysers is evidence of the applicability of the method as an alternative to commercial methods (Fig. 2). The validation of the method shows its excellent precision (Table 2), its agreement with the results from commercial analysers and the quantitative recoveries it provides (Table 3). Compared with other enzymatic methods, the linear range of the FI method is wider than those previously reported. In addition, the method proposed here is four times as rapid as the kinetic method [10,11] and more precise (%C.V. 2.1 vs. 3.1 and 3.6 vs. 9.0 for the within- and between-run studies, respectively); it is also twice as rapid as the steady-state method [12] and more precise (°A,C.V. 2.1 vs. 4.7 and 3.6 vs. 7.3 for the within- and between-run studies, respectively) and it is also twice as rapid as, and similar in precision to, the Kodak method [19]. The proposed method also surpasses the previous methods in reducing analysis costs as it uses immobilized enzymes, thus dramatically decreasing consumption of these biocatalysts.

Acknowledgement Direcci6n General de Investigaci6n Cientifica y T6cnica (DGICyT) is thanked for financial support (Project No. PB93-0827).

References [1] Tielz NW, Siggaard A, Pruden EL. Acid-base balance and acid-base disorders, in: Burtis CA, Ashwood ER, eds. Textbook of clinical chemistry. Philadelphia PA: W.B. Saunders Co., 1994. [2] Oesch U, Ammann U, Simon W. Ion selective membrane electrodes for clinical use. Clin Chem 1986;32:1412-1448. [3] Oeseh U, Malinowska E, Simon W. Bicarbonate sensitive electrode based on planar thin membrane technology. Anal Chem 1987;59:2131-2135. [4] Mokhallalati AO, Istudor A, Virtosu D, Luca C. Potentiometric gas sensor for carbon dioxide. Rev Roum Chim 1985;30( I ):31-36. [5] Kreling JR, Dezwaan J. Ion chromatographic procedure far bicarbonate determination in biological fluids. Anal Chem 1986;58;3028-3031. [6] Van Slyke DD, Stadie WC. The determination of the bases of the blood. J Biol Chem 1921;49:1-42. [7] Van Slyke DD, Neiel JM. The determination of the gases in blood and other solutions by vacuum extraction and manometric measurement. J Biol Chem 1924;61:523-573. [8] Skeggs LT. An automatic method for the determination of carbon dioxide in blood plasma. Am J Clin Pathol 1960;33:181-185. [9] Novies KA, Atkinson AR, Smith WG. Colorimetric enzymatic determination of serum total carbon dioxide, as applied to the Vickers multichannels 300 discrete analyzer. Clin Chem 1975;21:10931101. [10] Peled N. An enzymatic bicarbonate reagent that is free of pyruvate interference. Clin Chem 1981;27:199-200.

R. Quiles et al. /Clinica Chimica Acta 235 (1995) 169-177

177

[11] Punzalan RR, Johnson GF, Cunningham BA, Feld RD. Kinetic measurement of bicarbonate in serum by thiocyanate inhibition of wheat germ phosphoenolpyruvate carboxylase. Clin Chem 1990;36:2057-2062. [12] Forrester RL, Wataji U, Silverman DA, Pierre KJ. Enzymatic method for determination of CO, in serum. Clin Chem 1976;22:243-245. [13] Masoom M, Townshend A. Determination of glucose in blood by flow injection analysis and an immobilized glucose oxidase column. Anal Chim Acta 1984;166: l I I - l 18. [14] Keesay, J. A revised biochemical reference source. Boehringer Mannheim Biochemical, Indianapolis, 1992;63. [15] Bryce DW, Fermindez-Romero JM, Luque de Castro MD. Fluorimetric determination of mercury (II) based on the inhibition of the enzymatic activity of urease. Anal Lett 1994:27(5):867-878. [16] Maruyama M, Easterday RL, Chang MC, Lane MD. The enzymatic carboxylation of phosphoenolpyruvate. J Biol Chem 1966;241:2405-2412. [17] Kroll MH, Emancipator M. A quantitative measure of nonlinearity. Clin Chem 1993:39:766-772. [18] Kroll MH, Chester R. Nonlinearity of high-density lipoprotein cholesterol determination is matrix dependent. Clin Chem 1994;40:384-394. [19] Kodak E. Chem. Pub. No. MP2-89.