Anal. Chem. 1093, 65, 3258-3261
3258
Determination of Cholesterol and Cholesterol Ester with Novel Enzyme Microsensors Junko Motonaka’ Technical College, The University of Tokushima, Minamijosanjima 2-1, Tokushima 770, Japan
Larry R. Faulkner Department of Chemistry, University of Illinois, 1209 West California Street, Urbana, Illinois 61801
Enzyme microsensors using cholesterol oxidase (EC 1.1.3.6) and cholesterol esterase (EC 3.1.1.13) were developed for measuring cholesterol and cholesterol ester. The platinum microsensors (platinumdiameter,50 pm) were etched in hot aqua regia to create a cavity at their tip. A porous composite material prepared from acetylene black and Teflon emulsion was packed into this cavity and the redox mediator [Os(bpy)s](PF& was monitored by cyclic voltammetry in the potential range of 200-900 mV. The microsensors were dipped overnight in buffer solution containingthe desired enzyme to immobilize it on the tip by adsorption. Calibrationcurves for measurements of cholesterol and cholesterol ester, the effects of pH, temperature,and concomitantcompounds,the lifetime of the microsensors,and their availability for measuring cholesterol and cholesterol ester in urine were examined. Under optimal conditions, the response of the sensors was linear in concentration ranges of 5 pM-0.47 mM cholesterol and 2 pM-1.00 mM cholesterol ester.
INTRODUCTION Cholesterol is the main sterol in bile and is a constituent of gallstones. It is also present in the brain and nervous tissue and is reported to be an essential constituent of all animal cells. Its determinations is very important clinicallyas a high serum cholesterol level is related with arteriosclerosis. An enzyme sensor should be useful for ita determination. Several systems for assays of cholesterol are available, but there are only a few reporta of sensors for measuring cholesterol13 and cholesterol ester.@ Microsensors have a number of advantages, and their applications in electrochemistry are steadily increasing. A highly reproducible steady-state diffusioncontrolled current can be obtained with a microsensor without forced convection because of the high mass transport rate per unit area of sensor surface. The limiting current is insensitive to fluctuations in natural convection in bulk solution. Therefore, microsensors are especially suitable for the construction of amperometric sensors, such as biosensors. (1) Satoh, I.; Karube, I.;Suzuki, S. Biotechnol. Bioeng. 1977,19,1095. (2) Bertrand, C.; Coulet, P. R.; Gautheron, D. C. Anal. Lett. 1979,12, 1411. (3) Clark, L. C. 3rdlnternational ConferenceonEnzymeEngineering; Weetal, H., Pye, E., Wingard, L., Eds.; Plenum Press: New York, 1975; VOl. 3. (4) Pwdie, N.; Murphy, H. I. Anal. Chem. 1991,63, 2947. (5) Mascini, M.; Tomassetti, M.; Iannello, M. Clin. Chim. Acta 1983, 132, I. (6) Hahn, Y.; Olson, C. L. Anal. Chem. 1979,61, 444. 0003-2700/93/0365-3258$04.00/0
In this work, we prepared and tested enzyme microsensors for the determinations of cholesterol and cholesterol ester. The sensors were made by binding a redox mediator to the enzyme and immobilizing the enzyme-redox mediator on the surface of a cavity in porous Nafion-modified carbon. In this way, the cholesterol sensor and cholesterol ester sensor did not need an additional membrane to retain enzyme-redox species and were relatively unaffected by the presence of oxygen in the test solution. The enzyme reactions involved are shown in Scheme I. Cholesterol concentrations were determined by measuring the anodic current of hydrogen peroxide at +650 mV vs Ag/AgCl (eq 1). H,O,
-
2H’
+ 0, + 2e-
(1)
EXPERIMENTAL SECTION Apparatus and Reagents. Electrochemical experiments were performed using a BAS-100B electrochemical analyzer (Bioanalytical Systems, Inc.) with a preamplifier unit. A thermostat (Brinkmann,mgw Lauda RC 3, Model T-2) was used for thermostatic control of sample solution. The materials used were cholesterol oxidase (EC 1.1.3.6) (ICN Biomedicals; from Pseudomonas; activity,2.58 unit.a/mg of protein)and cholesterol esterase (EC 3.1.1.13) (Sigma Chemical Co.; from porcine pancreas; activity, 1.6 unita/mg of protein), cholesterol (Fisher Scientific;Stanbio standards),cholesterol palmitate (TCI American), human serum (Sigma Chemical Co.), osmium(I1) chloride hydrate (Aldrich), Nafion solution (Aldrich; 5 % Nafion 117 solution in methanol) and Zonyl FSN fluorosurfactant (DuPont). For synthesis of [Os(bpy)gI(PF&, 1.2050 g of osmium(I1) chloride was dissolved in 40 mL of DMF and refluxed for 10min. 2,2’-Bipyridyl (1.2646 g) was dissolved in 10 mL of DMF and added in 1-mLportions over 5 min. Duringrefluxing,the solution turned dark brown. After 90 min, the solution was cooled and added dropwise to 200 mL of diethyl ether with stirring to give a brown, gumlike product. Two molecules of water of hydration were absorbed from the solvents, and on cooling and stirring, a brown powder was formed. This was filtered usinga fie-porosity sinteredglass funnel and was washed extensivelywithether (yield, 2.3420 g, 95%). This product ([Os(bpy)zCl~ICl) was dissolved in 30 mL of DMF and reduced by adding 20 mL of dilute sodium dithionite solution. On cooling, the product [Os(bpy)zClzl separated as a dark purple/black powder and was recrystallized by slow evaporation of a water/methanol solution (yield, 2.2017 g, 90%). A sample of 0.9137 g of [Os(bpy)zClzl was dissolved in 40 mL of ethanol and refluxed 10 min to ensure ita complete dissolution. Then 20 mL of deionized water was added, followed by a 2-fold excess of 2,2’-bipyridyl(O.4973 g) dissolved in 5 mL of ethanol, and the resulting solution was refluxed for 4 h. After about 1-2 h, the solution changed from purple/brown to dark green/black. After completion of the reaction, the volume was reduced to about 20 mL by.rotary evaporation and an ice-cold solution of NFSFa was added. The dark green precipitated product was recrystallized from methanoVwatercontainingexcess PFe (yield 90%). 0 1993 Amerkan Chemical Society
ANALYTICAL CHEMISTRY, VOL. 85, NO. 22, NOVEMBER 15, 1993
3259
Cholesterolester Cholesterol
Cu wire
cox
P t wire
O2+ 2H'
-Sensing point I
Porous carbon comPosi t e
I
,
H
Flgure 2. Cholesterol ester enzyme microsensor and Its reaction scheme: COX, cholesterol oxldase; CE, cholesterol esterase.
T
lmobilized enzyme
4 "
EXP. CONDITIONS:
--
INIT E (mVl 200 H I G H E (mvl ROO LOW E ( n V I - Po0 v (mV/SECI- L SWEEP SEQYENTS- 1 SYPL ZNT. (mVl L
T
Flgwr 1. Structure of the enzyme mlcrosensor.
-
A
Deionized water purified further with a milli-Q water purification system was used throughout. Preparation of Sensor. The porous composite was made from 90% acetylene black and 10% Teflon emulsi~n.~ A sheet of the material was then prepared by extruding the carbon paste from an extrusion press, and this sheet was recycle-extracted with acetonitrile for 48 h. The sheet had a pore volume of approximately85% ,the pores being mainly 1-100 pm in diameter. This sheet (0.3 mm thick) was first soaked in 0.1 % Zonyl FSN fluorocarbon surfactantto make the pores hydrophilic, then dried, soaked in 2.5% Ndion solution, and dried again. The radius of the platinum microelectrodes (platinum diameter, 50 pm) was chosen to be significantlylarger than the sizeof the carbon powder particles. Platinum microelectrodes were etched in hot aqua regia (80 "C,1h) to create a cavity of 2-4-pm depth, and this cavitywas packed with the porous compositematerial. The redox mediator [Os(bpy)a](PF& was synthesized from osmium(I1) chloride and loaded by cyclic voltammetry in a potential range of 200-900mVfor more than 100cycles at a scan rate of 20 mV/s. The porous microseneor was dipped overnight in buffer solution of pH 7.0 containing cholesterol oxidase (37 mg/lOOpL) or buffer solution of pH 7.0 containing a mixture of cholesterol esterase (10 mg/100 pL) and cholesterol oxidase (37 mg/100 pL)at 4 "C to immobilize the enzyme by the adsorption techniques (Figure 1).
Determination of Cholesterol or Cholesterol Ester. A volume of 4.5 mL of pH 7.0 citric acid/disodium hydrogen phosphate buffer solution and 0.5 mL of 1 mM cholesterol or cholesterol ester solution was introduced into a thermostatregulated cell (37 "C) and the tip of the enzyme microsensor was immersed in the solution. The concentration of cholesterol or cholesterolester was determined by measuring the anodiccurrent of hydrogen peroxide (+650 mV vs Ag/AgCl) in the sample solution saturated with air. Electrochemical experiments were performed using a BAS-100B electrochemical analyzer with a
D
E (VOLT) Flgurr 3. Voltammogram of the cholesterol microsensor In cholesterol solutionsof differentconcentratkns: samplesolutlon, 5 mL, containing 0 (A), 0.02 (B), 0.08 (C), and 0.1 mM (D) cholesterol In pH 7.0 citric acid/dlsodlum hydrogen phosphate buffer solution (37 OC).
preamplifier unit. Allexperiments were carried out using a threeelectrode system. A silver/silver chloride reference electrode and platinum wire counter electrode were used.
RESULT AND DISCUSSION In this study we developed enzyme microsensore for cholesterol and cholesterol ester and examined their properties. The reaction scheme of the cholesterol ester enzyme microsensor is shown in Figure 2. Loading of the Sensor in OsmiumComplex. The porous carbon microelectrode was loaded in 1mM osmium complex/ 0.1 M sodium perchlorate. The process was fairly rapid and the peaks in the cyclic voltammogram were almost symmetrical, indicating negligible iR drop within the pores of the sensor. The loading value calculated from the area of the
Scheme I. Reactions of Cholesterol Ester and Cholesterol 3!?
^.
A
.
d
RCOO
Cholesterol
Cholesterol ester
%I-: Cholesterol
Cholesterol
J
CH.
4-cholesten-9one
+
RCOOH
3260
ANALYTICAL CHEMISTRY, VOL. 65, NO. 22, NOVEMBER 15, 1993
1
Table I. Effect of Concomitant Compounds on the Determination of Cholesterol. molar ratio cholesterol: re1 error (%) concom compd concom compd creatinine urea uric acid L-ascorbic acid ergocalciferol sodium chloride
0.0
0.2
0.6
0.4
0.8
1 .o
Concentration (mM) Figure 4. Callbratlon curve for cholesterol: sample solution, 5 mL, contalnlng cholesterolIn pH 7.0 cltricacid/disodlumhydrogen phosphate buffer solution (37 "C).
sodium phosphate (anhydrous) potassium chloride magnesium chloride lithium sulfate magnesium sulfate retinol
" r
0.0
0.2
0.4
0.8
0.8
1 .o
1.2
1:l 1:lO 1:l 1:lO 1:l 1:lO 1:l 1:l 1:lO 1:loo 1:l 1:lO 1:loo 1:1 1:lO 1:loo 1:l 1:lO 1:loo 1:l 1:lO 1:loo 1:l 1:lO 1:loo 1:l 1:lO 1:O.l 1:l 1:lO
0.06 1.4 0.39 -6.0 1.0 6.9 4.9 -1.9 -2.3 -4.6 -1.4 0.90 7.3 1.0 1.9 4.7 1.0 1.1 -4.5 1.2 -0.95 -1.5 -2.3 5.0 -0.83 1.9 -0.88 0.17 -0.6 -4.5 0.0
"Sample solution, 5 mL, containing 0.5 mM cholesterol and concomitant compound in pH 7.0 citric acid/didium hydrogen phosphate buffer solution.
Concentration (mM) Figure 5. Calibration curve for cholesterol ester: sample solution, 5 mL, containing cholesterol palmitate in pH 7.0 citric acWdlsod1um hydrogen phosphate buffer solution (37 "C).
peak was 4.57 X 10-7M/cm2 of the geometric sensor surface. This value is 240 times the value obtained with a typical Ndion-coated pyolytic graphite sensor (1.9X 1O-QM/cm2).7 Therefore, modification of the inner surface of the porous sensor is a very efficient way of obtaining a significantly increased concentration of catalytically active charge-transferring redox sites per unit of apparent area of sensor surface. This is a great advantage of the enzyme sensor since it results in marked increase in the electrochemical response. Voltammograms of Cholesterol. Figure 3 shows the voltammogram obtained with the cholesterol microsensor in pH 7.0 citric acid/disodium hydrogen phosphate buffer solutions of cholesterol of 0, 0.02, 0.06, and 0.1 mM. The sensor current increased with the cholesterol concentration. Similar results were obtained for randomly chosen cholesterol concentrations and buffer solution. Determinations of the anodic current of hydrogen peroxide at +500 to +650 mV vs Ag/AgCW*8were reported previously. In this experiment, the anodic current of hydrogen peroxide at +650 mV vs Ag/ AgCl was used for measurements. The sensor was rinsed with deionized water before each measurement. The voltammograms on a single-potential time base with the cholesterol microsensor in cholesterol solutions of different concentrations were measured. In this experiment, 0.5 mM cholesterol solution was injected into the cell while the current was being monitored. The plateau current stabilized in 2 min (a 95% (7) Buttey, D. A,; Anson, F. C. J. EZectroanal. Chem. 1981,130,333. (8)Motonaka, J.; Takabayashi, H.; Ikeda, S.; Tanaka, N. Anal. Lett.
1990,23,1981.
equilibrium current of 0.5 mM cholesterolwas reached in 111 s), indicating that the response time was good. Effect of pH. The effects of the pH of the sample solution on the responses of the microsensors to cholesterol and cholesterol ester were examined in pH ranges of 4.94-10.3 and 4.94-9.39, respectively. The maximal responses were observed at pH 7.14for cholesterol and pH 6.0 for cholesterol ester at 37 "C. These pH optimums differ from those of cholesterol oxidase (pH 7)and cholesterol esterase (pH 8.3). This pH shift of the optimums of the immobilized enzymes can be explained as follows: If a carrier is negatively charged, then a high concentration of positively charged ions (H+)will accumulate at its boundary with the surrounding solution, and consequently, the pH at the carrier surface will become lower than that of the bulk solution. Immobilized enzymes will, therefore, function at lower pH values than that of the bulk s~lution.~JOOn the basis of the above results, all subsequent measurements were carried out in sample solutions of pH 7.14 for cholesterol and pH 6.0 for cholesterol ester. Effect of Temperature. The effects of temperatures of 8-55 "C on the determinations of cholesterol and cholesterol ester were examined. In both cases the optimum temperature was 37 "C.The current decreased with decrease in temperature below 37 "C because ofincomplete enzymaticreactions, and with increase in temperature above 37 "C because of reduced enzyme activities. All subsequent measurements were done at 37 "C. Calibration Curve. The relationships between the concentrations of cholesterol and cholesterolester and the anodic (9) Goldstein,L.; Levin, Y.; Katchalski, E. Biochemistry 1964,3,1913. (10)Guilbault,G. G. Analytical Uses ofZmmobiZizedEnzyme;Marcel Dekker, Inc.: New York, 1984,p 137.
ANALYTICAL CHEMISTRY, VOL. 65, NO. 22, NOVEMBER 15, I993
Table 11. Effect of Concomitant Compounds on the Determination of Cholesterol Ester. molar ratio free cholesterol: re1 concom compd concom compd error (%) creatinine
1:l 1:lO 1:l 1:lO 1:100 1:1 1:lO 1:l 1:l 1:lO 1:100 1:l 1:lO 1:100 1:1 1:lO 1:100 1:1 1:lO 1:100 1:l 1:lO 1:100 1:l 1:lO 1:100 1:1 1:O.l
urea uric acid L-ascorbic acid ergocalciferol sodium chloride sodium phosphate (anhydrous) potassium chloride magnesium chloride lithium sulfate magnesium sulfate retinol
0.06 1.4 -1.7 -3.4 -6.8 0.88 -1.9 -1.1 -1.9 -1.5 -5.8 -2.9 0.0 17 0.0 1.4 1.4 0.28 0.75 0.81 -0.48 2.86 -11 0.0 -4.1 -4.4 -8.7 1.4 0.0
Sample solutions,5 mL,containingO.5 mM cholesterol palmitate and concomitant compound in pH 7.0citric acid/disodiumhydrogen phosphate buffer solution. 0
current were examined using the sensors under the optimum conditions determined as described above; results are shown in Figures 4 and 5. Linear relationships were obtained in the ranges of 3.3-240 mg/L (0.02-1.4 mM) cholesterol and 0.8317 mg/L (0.005-0.1 mM) cholesterol ester. The calibration curves could be represented by the equations Y = 1.94X 0.0147 and Y = 5.71X 3.47,where Y is the anodic current in nanoamperes and X is the cholesterol or cholesterol ester concentration in milligrams per liter. The correlation coefficient was 0.998for cholesteroland 0.997for cholesterol ester. Effects of Concomitant Compounds. Cholesterol and cholesterol ester were also measured in the presenceof various other compounds as shown in Tables I and 11. Cholesterol could be determined with a relative error of less than 7.3% in the presence of a molar ratio of 0.1:lO retinol; 1 L-ascorbic acid; 1:lO creatinine, urea, uric acid, or magnesium sulfate; or 1:100 ergocalciferol, sodium chloride, sodium phosphate, potassium chloride, magnesium chloride, or lithium sulfate. Cholesterol ester could be determined with a relative error of less than 8.7% in the presence of a molar ratio of 0.1 retinol; 1 L-ascorbic acid or magnesium sulfate; 1:lO creatinine or uric acid; or 1:100urea, ergocalciferol,sodium chloride, sodium phosphate, potassium chloride, magnesium chloride, or lithium sulfate. The interferences of concomitantcompounds are mainly due to alteration of substance accessibility, but L-ascorbic acid affects many chemical reactions. Determination of Cholesterol in Commercial Materials. The concentrations of cholesterol in four commercial preparations were determined with the enzyme microsensor and compared with those obtained by the JPXI method11 (Table 111). The values showed relative errors of less than 4.0% and coefficienta of variation of less than 1.6%.
+
(11)Ishidate,M.,Momom, T.,Shimomura, T.,Suzuki, I., Ede.Dai 11 Kaiaei Nippon ‘Yakkyokuhou kaiaetuho; D-327,Hirokawa Shoten: Tokyo, 1986.
3261
Table 111. Determination of Cholesterol in Commercial Materials. concn (mg) concn (mM) calcdb foundc error(%) re1 CVd(%) A
B C
D
0.470 0.468 0.480 0.475
1.00 0.99 1.02 1.01
0.99 0.96 1.01 1.05
-1.0 -3.0 -0.98 4.0
0.1 2.0 1.2 1.6
a Sample solution,5 mL, containing0.6 mM commercial cholesterol in pH 7.0citric acid/disodiumhydrogen phosphate buffer solution. Calculated value by the JPXI method. Average of four determinations. d Coefficient of variation.
Table IV. Determination of Cholesterol in Human Serum (Male)# concn
sensor cholesterol total cholesterol
founda (mg/dL)
founda (mM)
CVb(%)
65.2e 2 w
1.69” 5.17f
1.2 0.9
a Average of six determinations. b Coefficient of variation. Reported normal range in human serum, 50-80 mg/dL. d Reported normal range in human serum, 1.29-2.07 mM. e Reported normal range in human serum, 130-250 mg/dL. f Reported normal range in human serum, 3.36-6.47mM. 8 Sample, 5 mL, of human serum was used without pretreatment.
Clinical Application. The enzyme microsensors were applied for determining the cholesterol concentration in a sample of serum of a healthy man (Table IV). Normal human serum is reported to contain 50-80 mg/dL (1.29-2.07mM)free cholesterol and 130-250 mg/dL (3.36-6.47 mM) total cholesterol. The average values determined with the sensors were 65.2 mg/dL (1.69mM) cholesterol and 200 mg/dL (5.17 mM) total cholesterol, which are within the reported ranges. Therefore, these sensors are useful for determination of cholesterol and total cholesterol levels in human serum. Lifetimes of the Enzyme Microsensors. The lifetimes of the enzyme microsensorswere examined. The lifetime2 of the cholesterol sensor is reported to be 2-3 months and that of the cholesterol ester sensor 1 month. In this studies we obtained similar values: Analytical values were constant for 60 days with the cholesterol sensor and 20 days with the cholesterol ester sensors. Subsequently, the sensitivities of the sensors gradually decreased, because of lowering of enzyme activities. The sensitivity of the cholesterol sensor was decreased 27% after 3 months and that of the cholesterol ester sensor was decreased 357% after 1 month. But a linear relationship was obtained between the concentrations of sample and anodic current for at least 3 months with the cholesterol sensor and 1 month with the cholesterol ester sensor.
ACKNOWLEDGMENT The authors are grateful to Dr. ChangmingLi for providing the porous carbon sheet and to Dr. Robert Forster (University of Illinois) for providing the osmium complex. RECEIVEDfor review May 3, 1993. Accepted August 27,
1993.Q
* Abstract published in Advance ACS Abstracts, Odober 1, 1993.
