Anal. Chem. 1994,66, 1485-1489
Continuous-Flow/Stopped-Flow System Incorporating Two Rotating Bioreactors in Tandem: Application to the Determination of Alkaline Phosphatase Activity in Serum Julio Rabat and Horaclo A. Mottola' Department of Chemistty, Oklahoma State Universi@, Stillwater, Oklahoma 74078-0447
Two rotating bioreactors in tandem have been incorporated into a continuous-flow/stopped-flow sample/reagent processing setupforthedeterminationofallralinephosphatase(EC 3.1.3.1) activity in serum samples. The strategy circumvents incompatibility of buffer systems as well as that of the immobilized enzymes utilized in the bioreactors (alkaline phosphatase and alcohol oxidase, EC 1.1.3.13). The determination is indirect in nature although recorded responses are directly related to the enzyme activity in the sample. It couples the following enzyme-catalyzed reactions: (1) hydrolysis of p-nitrophenyl dihydrogen phosphate catalyzed by alkaline phosphatase, (2) enzymaticreactionbetween unreactedp-nitrophenyldihydrogen phosphate with methanol, and (3) conversion of the residual methanol to the correspondingaldehyde and HzOz, catalyzed by alcohol oxidase. The H202is amperometricallydetermined at a stationary Pt-ring electrode (applied potential 0.600 V vs a Ag/AgCl, 3.0 M NaCl reference).
+
Coupling of the main enzyme-catalyzed reaction with a second (indicator) reaction is a common practice in enzymatic methods.' In many instances the indicator reaction can be in itself an enzyme-catalyzed one. There may be, however, cases in which the coupling is impaired by incompatibility of buffer systems and/or of enzymes. This makes impractical the coimmobilization of the enzymes involved, and two reactors in tandem with on-line switch of buffer systems becomes an attractive alternative to circumvent the difficulty. This paper illustrates how use of two rotating bioreactors similar to one recently introduced,2 and continuous-flow sample/reagent processing, provide for the implementation of the sought-for alternative. The strategy is illustrated with thedetermination of the enzyme activity of alkaline phosphatase (EC 3.1.3.1) in serum utilizing two immobilized-enzyme reactors involving enzymes with mutual product inhibition (coimmobilization would fail) and some buffer incompatibility. The approach results in an indirect way to arrive at the enzyme activity, involving the following: (1) reaction (for a given amount of time) of the serum sample with p-nitrophenyl dihydrogen phosphate acting as enzyme substrate, (2) introduction of the reacted sample into a bioreactor containing immobilized Permanent address: DcpartamentodeQuhica Analltica, Facultadde Qulmica, Bioqulmica y Farmacia, Universidad Nacional de San Luis, 5700 San Luis, Argentina. (1) Mottola, H. A. Kinetic Aspects of Annlytical Chemistry; Wiley: New York, 1988; p 67. (2) Matsumoto, K.; Baeza Baeza, J. J.; Mottola, H. A. And. Chem. 1993, 65,
636-639. 0003-2700/94/0386-1485$04.50/0 0 1994 American Chemical Socletv
alkaline phosphatase, which in the presence of methanol acts as a transphosphorylase and transfers a phosphate group to the alcohol in direct proportion to the level of unreacted p-nitrophenyl dihydrogen phosphate in the sample, and (3) transfer of the unreacted methanol (with an on-line switch of the buffer system) to a second bioreactor/detection unit in which the unreacted methanol is converted, with the aid of 0 2 and alcohol oxidase (EC 1.1.3.13), to the corresponding aldehyde and H202. The H202 produced is amperometrically detected by a concentric Pt-ring electrode.2 Table 1 summarizes the reactions involved in the procedure. The chemistries involved are very easily implemented in a continuousflow/stopped-flow/continuous-flowmode of sample/reagent processing. In essence the determination is indirect in nature, but since the higher the enzyme activity in the sample the higher the concentration of unreacted methanol, there is direct correlation between the response (treated in a rate form) and the enzyme activity in the sample. Thevalidityof the approach was verified with the determination of alkaline phosphatase in reference standards of human serum.
EXPER I MENTAL SECT1ON Reagents and Solutions. The water used for solution preparation was deionized and further purified by distillation in an all-borosilicate glass still with a quartz immersion heater. All chemicals used, except as noted, were of analytical reagent grade. Both enzymes, alkalinephosphatase (EC 3.1.3.1) from bovine intestinal mucosa, and alcohol oxidase (EC 1.1-3.13) from Pichia pastoris, were purchased from Sigma Chemical Co. (St. Louis, MO). Also from Sigma were the colorimetric diagnostic kit for the determination of alkaline phosphatase activity in serum (Accutrol, normal and abnormal control standards) and the p-nitrophenyl phosphate hexahydrate disodiumsalt (Sigma 104(R)). Glutaraldehyde (25% aqueous solution) was from Aldrich Chemical Co. (Milwaukee, WI). 3-Aminopropyl-modifiedcontrolled-pore glass (APCPG) was from Electro-Nucleonics, Fairfield, N J (1400-A mean pore diameter, 24 m2.g-* surface area, and 48.2 pmo1.g' amino groups). Methanol was obtained from Mallinckrcdt, Inc. (St. Louis, MO). The glycine buffer was prepared by adjusting the pH of a 0.10 M solution of glycine powder (Matheson, Coleman & Bell, Norwood, OH) to pH 8.40 with 0.10 M NaOH. Tris(hydroxymethy1)aminomethane (Trizma base, tris), and diethanolamine were also from Sigma. All the buffer solutions Analytical Chemistry, Vol. 66, No. 9, Mey 1, 1994
1485
Table 1. Chemlcrl Reactlona Involved In Each Step ol the Determlnatlon'
p-nitrophenyldihydrogenphosphate + H,O
-
tree alkaline p h m p l u W
p-nitrophenol+ HPO,%
(1)
reaction in reactor 1 p-nitrophenyldihydrogenphosphate + CH,OH
immobilized allraline pbmphataae + CH3P0,2
+ p-nitrophenol
(2)
reaction in reactor 2
CH,OH
+ 0,
-
immobilized dwbol oxidw
HCHO + H,O,
(3)
reaction at Pt-ring electrode H,O,
-
O,(g) + 2H'
+ 2e-
(4)
a Reaction 1 occurs before introduction of the sample in the flow system. Reaction 4 is the one that permits amperometric detection, generatinga signal directly proportionalto active alkaline phosphatase in the sample. Reactions 1 and 2 1.0 mM diethanolaminebuffer, pH 9-10, 1.0 mM MgC12. Reactions 3 and 4: phosphate buffer (0.10 M, pH 7.50). * Alkaline phosphatase in the serum sample.
+
used as carriers contained 10 mM MgC12 as activator for alkaline phosphatase. The alkaline phosphatase substrate solutionsofp-nitrophenyl dihydrogen phosphate were prepared in this carrier solution, deaerated by bubbling Nz(g) through, and stored in the refrigerator at about 4 OC between uses. Reactors/Electrode Flow System and Other Apparatus. The reactor and reactor/detector cells used have been described previou~ly.~.~ The potential applied to the Pt-ring electrode for H202 detection was +0.600 V vs a Ag/AgCl, 3 M NaCl reference electrode. Figure 1 illustrates schematically the overall configuration of the setup utilized in this work. Pump tubing was of Tygon (Fisher AccuRated 1.0 mm i.d., Fisher Scientific Co., Pittsburgh, PA) and the rest of the tubing used was 1.O-mm4.d. Teflon (Cole Parmer, Chicago, IL). Confluence of solutions before entering the second reactor was aided with a three-way connector (Precision Miniature Fitting Model 1003, Omnifit, Atlantic Beach, NY). All pH measurements were made with an Orion Model 601A digital pH meter (Orion Research, Cambridge, MA) equipped with an epoxy-bodycombination electrode (Sensorex, Westminster, CA). Spectrophotometric measurements were performed with a Perkin Elmer diode array, Lambda 3840, UV/visible spectrophotometer and 1-cm glass cells. Procedures for Immobilization of Enzymes. The two enzymesused were immobilized as described earlier for glucose oxidase.2 For alcohol oxidase, the appropriate amount was 150pL of enzyme suspension [8.1 (mg of protein).mL-', 2490 units.mgl] in 50pLof phosphate buffer, pH 8.00.4 Enzymatic activity was very well retained for at least 2 weeks of daily use. After this time deterioration can be observed, but the response can be improved by successive injections of pure (3) Bacza Bacza, J. J.; Matsumoto, K.; Mottola, H. A. AM^. Chim. Acto 1993, 283,ia5-i93. (4) Bacza Bacza, J. J.; Matsumoto, K.; Mottola, H. A. Quim. AM^. 1993, 12, 12-17.
U
P
Figwe 1. Experimental setup: P, pump (ollson Minipuls 2 peristaltic pump, Glison Electronics, Inc., Middleton, WI); SI, sample injection (Rheodyne Model 50, four-way rotary valve); RR1, rotatlng bloreactor with immobiiized alkaline phosphatase; RR2, rotatlng reactor with immobillzed alcohol oxidase; WE, Pt-ring working electrode; PA, potentlostat/ampufler ( L W B unlt from Bioanaiytkal Systems, West Lafayette, IN); RE, reference electrode (Ag/AgCi, 3.0 M NaCl); AE, stainless steel tube acting as auxiliary electrode; R, recorder (SuperW, waste line; Scribe Model 4910, Houston Instruments,Houston, lX); a, carrier buffer line; b, sample line; c, phosphate buffer line (0.10 M, pH 7.50); d, excess sample recycledto sample reservoir;e, three-way connector. Shaded areas in RR1 and RR2 indicate iocatlon of the rotating disks with immobilized enzymes.
methanol. This effect may be an accumulated effect of noncompetitive inhibition from species generated in the first reactor. Alcohol oxidase reactors, however, showed significantly impaired performance after 20 days of continuous use. The quantities found appropriate for alkaline phosphatase immobilization were 150 pL of enzyme suspension [52 (mg of protein).mL-l, 33 units*mgl] and 50 pL of 0.10 M phosphate buffer, pH 7.50. Disk reactors were washed with the appropriate phosphate buffer and stored in the same buffer at 5 OC between uses. RESULTS AND DISCUSSION Coimmobilization of alkaline phosphatase and alcohol
Table 2. Values of &’ (Apparenl Mlchaelk-Menten Conrtanl) (Temperature 20 f 1 “C)
rotation velocity, rpm
Km’,amM
linear regression std dev
120 240 840 1020 free enzyme in soh*
1.76 1.11 0.77 0.59 5.00
h0.33 h0.18 h0.16 h0.12 h0.30
A
e Each value of Km‘based on triplicate values of five different substrate concentrations. Estimated as described in the text.
*
oxidase fails to provide a workable medium for the determination described in this paper. Two main reasons contribute to the failure: (1) enzyme incompatibility (there is an impaired performance, probably by product inhibitionof alcohol oxidase, when the enzymes are in proximity) and (2) some incompatibility also between the optimum conditions for the individual performance of each enzyme. This is not an uncommon situation when coupled enzyme reactions are used to implement analytical procedures. The setup shown in Figure 1 convenientlycircumventsthese problems by taking advantage of the flexibility afforded by continuous-flow sample/reagent processing and the on-line implementation of rotating bioreactors. These reactors utilize very small amounts of biocatalysts in separate configuration and permit easy switching of working conditions. The schematic of Figure 1 illustrates the sample/solution pathways and Table 1 gives the chemical reactions involved in the strategy presented here. A serious incompatibility in coupling the two enzymes is the fact that phosphate buffer provides the most convenient environment for the methanol oxidation catalyzed by alcohol oxidase. Use of this buffer, however, would result in competitive inhibition of the alkaline phosphatase-catalyzed hydrolysis of phosphate ester^.^ Temporal and spatial separation of the reactions involved avoids this problem. The Apparent Michaelis-Menten Constant for the System. As indicated previously,2 reactor rotation is expected to decrease the value of the apparent Michaelis-Menten constant for a given immobilizedenzyme preparation. The same should be expected for a similar apparent constant for the concurrent performance of two immobilized enzyme preparations. The overall result would be the increase of the reaction rate dictating the availabilityof detected species. For configurational reasons such a rate corresponds to that of the oxidation of methanol in this case. Table 2 sumarizes the values Of Km’obtained at different rotation velocities with the same approach described earlier2 and by stopping the flow for 2.0 min in the center of the second reactor during signal acquisition. The probe substrate was methanol, and Km’was extracted from plots of 1/rate vs methanol concentration by fitting data to the linear relationship (l/rate) = (m/[CH,OH])
+n
in which Km‘= m/n. All signal values were corrected for background readings. The data in Table 2 confirm the trend of larger Km’values as the disks’ rotation rates decrease. ( 5 ) Reid,T. W.; Wilson, I. E. In TheEnzymes,3rded.; Boyer,P. D.,Ed.;Academic Press: New York, 1971; Vol. IV,p 399.
i
H
C
Flgure 2. Effect of reactor rotatlon under contlnuous- and stoppedflow condklons: A, stopped flow wlth rotation; B, contlnuous flow wlth rotatlon; C, stopped flow without rotation. Experlmental condltbns: pnkrophenyl dlhydrogen phosphate, 2.00 mM; methenol, 1.00 mM; flow rate, 1.0 ml-mln-l; sample size, 100 pL;vetocity of rotatbn, 874 rpm; I, injectlon, s, flow stopped, and f, flow continued.
The Michaelis-Menten constant of free alcohol oxidase in solution was found by determining the rate of HzO2 formation as follows: 2.0 mL of the substrate solution of a given concentration, 50.0 pL of enzyme suspension (activity 20 units), and 0.50 mL of 0.10 M phosphate buffer, pH 7.50, were allowed to react for 2 min at room temperature. The reaction was quenched by addition of 0.20 mL of a 4.0 M HCl solution. After this the solution was neutralized with 4.00 M NaOH, and the H202 formed was photometrically determined using leuco crystal violet.6 The value of the constant was derived from a Lineweaver-Burk plot. Effect of Cell Volume and Inserted Sample Size. The volume of the reactors can be varied by insertion of Teflon spacers between the top and bottom parts of the cell. The minimum cell volume (no spacer used) was 450 pL, and the maximum (with a 3-mm-thick spacer) was 1.8 mL. The volume comprising the area just above the reactor was calculated to be 113 p L . The rate of response decreased with cell volume, as should be expected because of the dilution effect favored by rotation and the current response to bulk concentration. Consequently, the smallest volume of 450 p L was selected for all measurements. The rate of response has been found2 to increase almost linearly with sample size up to 250 p L in a cell with a volume of 450 pL. The effect of sample size was ascertained by changing the sampleloop from 50 to 250 pL. For convenience, a sample size of 100 pL was used to evaluate other parameters. Effects of Continuous-Flow vs Stopped-Flow Operation, Disk Rotation, and Flow Rate. Figure 2 shows typical signals
-
(6) Mottola, H. A.; Simpson, E.
E.;Gorin, G. Anal. Chcm. 1970, 42, 410-411.
AnaWtical Chemlstry, Vol. 66, No. 9, May 1, 1994
1407
T 40
-
30
-
‘0‘oa
\ c
Figure 3. Effect of flow rate on signal profile. Flow rates: (A) 2.0, (B) 1.0, and (C) 0.50 mL-min-’. Ail signals recorded with stopped-flow and reactor rotation (874 rpm). Other conditions as in Figure 2. I, InJection, s, flow stopped, and f, flow continued.
acquired under continuous-flow and continuous-flow/stoppedflow/continuous-flow programming. Figure 3 showsthe effect of flow rate. Trends confirm previous observations.2 The best conditions are given in Figure 4b, and for convenience, initial rate measurements were used for quantitation. Effect of Buffer Composition. Although optimal pH conditionsfor the enzymesare quite similar, for reasons already discussed it was not possible to use a single buffer. The carrier solution employed the buffer compatible with the first reactor containing immobilized alkaline phosphatase. Figure 4 shows the response trends for different buffer components and pH values of the carrier. Both diethanolamine and TRIS buffers are satisfactory, but diethanolamine was used because it is the one recommended for use in determinations of alkaline phosphatase activity.’ The optimum pH for determination is close to 7.50 in all cases illustrated in Figure 4. Effect of Methanol Concentration. A linear relation was observed between the rate of response and the methanol concentration in the 0.010-3.0 mM range at a constant p-nitrophenyl dihydrogen phosphate concentration of 0.015 mM. The equation of the regression line obtained at 847 rpm was rate of response (nA-min-’) = 1.03 + 2.66 [methanol, mM] The coefficient of determination (ref 7, Chapter 2) was found to be 0.998. The standard deviation about regression for this type of plot was 1.26 with a minimum detectable concentration of methanol equivalent to 0.0050 mM. The standard deviation for the result of the mean of seven measurements (0.10 mM methanol) was typically 0.019 mM (coefficient of variation 1.9%). Determination of Alkaline Phosphatase Activity in Blood Serum. An aliquot of each serum samples was pretreated as follows for application of the standard addition method (used to minimize matrix effects): to four 50.0-pL samples of serum was added 100.0 pL of phosphatase solution with 0, 15, 25, and 55 units/L, respectively. Then, 5 mL of a solution 0.10 M in p-nitrophenyl dihydrogen phosphate [this concentration was sufficiently high to ensure that no more than 10%of the substrate was hydrolyzed, thus minimizing the effects of product inhibition or substrate depletion (ref 1, pp 57-58)], ~
~~~~
(7) Kachmar, J. F.; Moss,D. W. Enzyme. In Fundamentalsof Clinical Chemistry; Tictz, N., Ed.; W. B. Saunders Co.: Philadelphia, PA, 1976; p 606.
1400
Analj/tlcel Chemistty, Voi. 66, No. 9, May 1, 1994
6.00
7.00
9.00
8.00
PH Flguro 4. Effect of pH and buffer composttion of the carrler buffer solution used to introduce the sample into the first reactor with immobilized alkaline phosphatase. The response corresponds to the signal obtainedinthe second reactoddetdmuntt after buffer switching as indicated in text. Buffer system: a, 0.0010 M Tris; b, 0.0010 M dlethanolamine; c, 0.010 M Tris; d, 0.010 M diethanolamine; e, 0.10 M diethanolamine; f, 0.10 M Tris; g, 0.010 M glycine. Tabk 3. R.rultr Obtakwd In tho D.1.rmlnatlon of Alkallne Pholphatrw A c t W In Flvo Dmwont Sampbs of Bkod Sorum
alkaline phosphatase activity,unita/L@** this method diagnostic kit 329.46 f 1.74 161.90 f 1.80 152.20f 2.02 48.77 f 2.54 33.21 f 1.06
* * **
351 4.24 149 1.41 147 f 2.45 50 2.30 37 3.79
One unit of alkaline phos hatase activity is defined aa the amount of enzple which catalyzes %e formation of 1.00 pmol L-1 g-nitropheno per minute under the conditionsof the determination. Valuea are the mean of three replicate determinations. @
1.OmM in diethanolamine, and 10.0mM in Mg2+, was added. After the reaction was allowed to proceed for 5 min at 30 O C , it was quenched by adding 0.20 mL of 0.40 M HCl. The pH was adjusted to 7.00 with 4.0 M NaOH, and 0.25 mL of 10 mM methanol was added. Finally, all samples were diluted to 25.0 mL with carrier solution and sequentially an aliquot of each sample was analyzed in the flow system. Results for five different samples of blood serum are summarized in Table 3. The results are compared with values obtained by analyzing the same samples but using the diagnostic kit mentioned in the Experimental Section. The Pearson’s correlation coefficient for the two methods was 0.996, indicating excellent linear correlation. The Pearson’s correlation coefficient was calculated using the following equation: r = c { [ ( x i- x ) / ( y , -y)]/ns$,,], where r is the correlation coefficient and xi and yi are the individual values of variables in the x- and y-axes, respectively. Average values of these variables are denoted by x and y, n is the number of determinations compared, and sx and s, are the sample standard deviations for the data in
the x- and y-axis, respectively. The correspondence of results is highly satisfactory and clinically acceptable.
for a fellowship from the Consejo Nacional de Investigaciones Cientfficas y T h i c a s (CONICET) of Argentina.
ACKNOWLEDGMENT This work was partially supported by the Oklahoma State University Water Research Center. J.R. expresses thanks
Recelved for review September 1, 1993. Accepted February 5, i9g4-' Abstract published in Advance ACS Absrracrs, March 15, 1994.
Analytcal Chemistry, Vol. 66, No. 9,May 1, 1994
1489
