Small-volume voltammetric detection of 4 ... - ACS Publications

Enno Kätelhön , Kay J. Krause , Klaus Mathwig , Serge G. Lemay , and Bernhard Wolfrum. ACS Nano 2014 8 .... Yongzhi Dong and Curtis Shannon. Analyti...
0 downloads 0 Views 1MB Size
Anal. Chem. 1993, 85, 1559-1563

1559

Small-Volume Voltammetric Detection of 4-Aminophenol with Interdigitated Array Electrodes and Its Application to Electrochemical Enzyme Immunoassay Osamu Niwa,’ Yan XU,: H. Brian Halsall,’ and William R. Heineman* University of Cincinnati, Department of Chemistry, Cincinnati, Ohio 45221 -01 72

A small-volume voltammetric detection of 4-aminophenol (PAP) has been developed using an interdigitated array (IDA) microelectrode cell in order to apply the IDA to electrochemicalenzyme immunoassay. The signal of PAP at the IDA was steady state, and its magnitude was amplified compared with that of the usual single electrode due to redox cycling of PAP between the two finger sets of the IDA. A linear relationship between PAP concentration and cathodic limiting current was obtained from 1 to 1000 pM,reproducibly. The minimum sample volume in the measurement was reduced to 800 nL. High sample throughput of less than 1-min detection time per sample was achieved on 2-10-pL PAP samples. This IDA cell was applied to the electrochemical enzyme immunoassay of mouse IgG. Alkaline phosphatase was used as the enzyme label. The mouse IgG concentration was evaluated by detecting the concentration of PAP, which is the product of enzymatic reaction of the substrate, 4-aminophenyl phosphate (PAPP). Anti-mouse IgG was covalently immobilizedon the glass surface of the small-volumeimmunowells by carbodiimide coupling. The assay range was 1(&1000 ng/mL using 10-pL sample and 20-pL substrate solutions. INTRODUCTION Immunoassay is a very useful method in bioanalytical chemistry due to its extremely high selectivity and low detection limit.’ Electrochemical detection has been successfully coupled to the enzyme immunoassay.2 Excellent detection limits can be achieved on small sample volumes by the combination of the heterogeneous enzyme immunoassay and highly sensitiveelectrochemicaldetection techniques such as liquid chromatography and flow injection analysis with electrochemicaldetection (FIAEC and LCEC).”5 In this case, the contents of the immunoassay reaction vessel are injected into the FIAEC or LCEC system as the final stage in the assay procedure. On the other hand, electrochemical detection in which the electrochemicalprobes are inserted directly into the solution + Permanent address: NTT Basic Research Laboratories, Nippon Telegraph and Telephone Corp., Tokai, Ibaraki, 319-11,Japan. 1 Permanent address: Department of Chemistry, Cleveland State University, Cleveland, OH, 44115. (1)Skelly, D. S.;Brown, L. P.; Besch, P. K. Clin.Chem. 1973,19,146. (2)Heineman, W. R.; Halsall, H. B.; Wehmeyer, K. R.; Doyle, M. J.; Wright, D. S. Anal. Proc. 1987,24,324. (3)Wehmeyer, K.R.; Halsall, H. B.; Heineman, W. R. Clin. Chem. 1985,31,1546. (4)Wehmeyer, K. R.; Halsall, H. B.; Heineman, W. R.; Volle, C. P.; Chen, I. W. Anal. Chern. 1986,58,135. (5)Halsall, H. B.; Heineman, W. R.; Jenkins, S.H. Clin. Chem. 1988, 34,1701.

0003-2700/93/0365-1559$04.00/0

in the immunoassay reaction vessel can also be coupled with electrochemical immunoassay. Ion-selective electrodes and modified electrodes have been applied to immunoassay (or used as an immunosensor) as potentiometric and amperometric probes, re~pectively.6,~ These electrochemicalprobes are inexpensive compared with flow systems and can be used for small-volumesample detection. As amperometric probes, microelectrodes are particularly interesting due to their fast establishment of a steady-state response, small sample volume, and much higher sensitivitythan conventionallysued electrodes.8 The problems of microelectrodes are the reproducible production of these small devices and the small absolute current flowing for the detection of low-concentrationsamples due to the small electrode area. However, recently developed microlithographic techniques can achieve the same size or shape microelectrode cell without difficulty.9 Microarray electrodes, which are fabricated by arranging a number of microelectrodes on the same tip, can increase the absolute current comparable to that of the usual electrode, while maintaining the advantageous properties of the single microelectrode.10 Interdigitated array electrodes (IDA) are particularly interesting for the detection of electrochemically reversible materials,11J2 because at moderate scan rates, nonplanar diffusion to each microband electrode in the IDA gives rise to steady-state currents. A significant decrease in detection limit has been reported for redox species such as ferrocene derivatives, ferrocyanide, and dopamine due to the amplification of the limiting current by redox ~ y c 1 i n g . l ~ ~ ~ ~ Recently, our group has advocated the use of 4-aminophenyl phosphate (PAPP) for amperometric detection of alkaline phosphatase, which is a very useful enzyme for the heterogeneous electrochemical enzyme imrn~noassay.~~ Electrochemical immunoassays for a-fetoprotein,16the~phylline,’~ and mouse IgG18 have been achieved using PAPP. The product of the enzymatic reaction, 4-aminophenol (PAP), shows excellent electrochemical properties such as low (6)Alexander, P. W.; Rechnitz, G. A. Anal. Chem. 1974,46,1253. (7)Aizawa, M.; Morioka, A.; Matsuoka, H.; Suzuki, S.;Nagamura, Y.; Shinohara, R.; Ishiguro, I. J. Solid Phase Biochem. 1976,1 , 319. (8)Pons, S.;Fleischmann, M. Anal. Chem. 1987,59,1391A. (9)Morita, M.; Longmire, M. L.; Murray, R. W. Anal. Chem. 1988,60, 2770. (10)Fleishmann, M.; Pons, S.;Rolison, D. R.; Schmidt, P. P. Ultramicroelectrodes; Datatech Systems: Morganton, NC, 1987. (11)Sanderson, D. G.; Anderson, L. B. Anal. Chem. 1985,57,2388. (12)Aoki, K.;Morita, M.; Niwa, 0.;Tabei, H. J.Electroanal. Chem. 1988,256,269. (13)Niwa, 0.; Morita, M.; Tabei, H. Anal. Chem. 1990,62,447. Morita, M.; Tabei, H. Electroanalysis 1991,3,163. (14)Niwa, 0.; (15)Tang, H. T.;Lunte, C. E.; Halsall, H. B.; Heineman, W. R. Anal. Chim. Acta 1988,214, 187. (16)Xu, Y.;Halsall, H. B.; Heineman, W. R. Clin. Chem. 1990,36, 1941. (17)Gil, E.P.;Tang, H.T.; Halsall, H. B.; Heineman, W. R.; Misiego, A. S. Clin. Chem. 1990,36,662. (18)Xu, Y.; Halsall, H. B.; Heineman, W. R. J.Pharm. Biorned. Anal. 1989,7,1301. 0 1993 American Chemical Society

1560

ANALYTICAL CHEMISTRY, VOL. 65. NO. 11. JUNE 1. 1993

.... .... ,.., ,..... .,,

.

,

..i

(b) 0.1 M potassium phosphate-sodium phosphate at pH 6.0 (PB); (c) 0.1 M sodium acetateacetic acid, 0.05% (viv) Tween 20,0.02 % (viv) sodium azide, 2 % (wiv)bovine albumin, and 0.5 mM 1-pentanesulfonic acid at pH 5.5 (acetateTBSA); (d) 0.10 M tris(hydroxymethyl)ainomethane, 1mM magnesium chloride, and 0.02% (wiv) sodium azide, pH adjusted to 9.0 with hydrochloric acid (tris). Solutions. Primary rat anti-mouse IgG solution (24 pgimL) was prepared hy dilution of a stock solution (1.2 mgimL) with PB buffer. EDC solution (24 pM) was also prepared with PB buffer. Mouse IgG solutions were diluted from a stock solution (11mg/mL) with acetateTBSA buffer. The rat anti-mouse IgG alkalinephosphataseconjugate solutionwas preparedas a 1:lOOO dilution of the stock solution (0.3 mgimL) with acetateTBSA buffer; 0.1 M tris buffer with 4 mM 4-aminophenyl phosphate was used as a substrate solution. Electrochemical Measurements. Electrochemical measurements for large-volume PAP (10 mL) solutions were made by cyclic voltammetry (CV) using IDA, AgiAgCl reference, and gold wire auxiliary electrodes. The potential of one finger set of the IDA electrode was fixed at -0.3 V and that of another finger set was cycled from -0.3 to +0.3 V at a scan rate of 50 or 100 mVis.

Flgure 1. SEM picture of the IDA electrochemical cell.

oxidation potential (300mV vs AgiAgCl), negligihleelectrode fouling, and reversible electrochemical behavior. Since the electrochemical reversihility of PAP is advantageous for voltammetric detection with dual electrodes, highly sensitive detectioncanheexpectedforIDAwithoutuseofaflowsystem. In this paper, we describe the small-volume voltammetric detection of PAP using an IDA electrochemical cell, and its application to an electrochemical immunoassay coupled with a small-size incubator. Mouse IgG was used as a model compound. EXPERIMENTAL SECTION Electrochemical Detector. The electrochemical cell used in this assay consists of an IDA gold working electrode and two extra square electrodes. These electrodes were fabricated on a thermally oxidized silicone wafer by photolithography and dry etching techniques.13 The SEM picture of this cell is shown in Figure 1. The overallsize ofthis detector is 2 cm long, 1cm wide, and 350 pm thick, but the actual sensing area of the electrochemical cell is 2 X 2 mm. The IDA electrode consists of 50 pairs ofmicrobands,thewidthofwhichis3or5pm. Thegapbetween fingers is 2 or 5 pm. One of the square electrodes was plated with silver for use as a reference electrodefor the small-volumesamples. Materials. Affinipurerat anti-mouseIgG (H + L), Chrompure mouse IgG whole molecules,and alkalinephosphatase conjugated Affinipure rat anti-mouse IgG (H + L) (without Fc region) were obtained fromJackson ImmunoresearchLaboratory (West Grove, PA). Bovineserumalbumin fractionVpowder,1-pentanesulfonic acid (sodium salt), l-ethyl-3-[3-(dimethylamino)propyllcarbodiimide hydrochloride (EDC)and 4-aminophenol hydrochloride were obtained from Sigma Chemical Co. (St. Louis, MO). Tris(hydroxymethy1)aminomethaneand (3-aminopropyl)triethoxysilane were obtained from Aldrich (Milwaukee,WI). All other chemicalswere purchased from Fisher Scientific (Cincinnati,OH). 4-Aminophenyl phosphate was synthesized as previously reported.15 Buffers. The following buffer solutions were used in this study: (a) 0.1 M sodium acetate-acetic acid at pH 5.5 (acetate);

For the small-volumesamples, silver plated and unplated gold square electrodes on the tip were used as reference and auxiliary electrodes, respectively. PAP tris buffer solution (0.8-10 pL) wasplacedon this IDAelectrochemicalcell for cyclic voltammetry measurements. Immobilization of Antibody. Microscope slide glass plates were immersed in 1 M NaOH-1 M HC1 solution overnight to introduce hydroxylgroupsonto thesurface. After theglassplateg were rinsed with pure water, they were incubated for 5 h in the acetate huffer containing 0.4% (3-aminopropyl)triethoxysilane at 90 "C. Adhesive Teflon tapes (50 pm thick) whose centers were cut out to form a I-mm-diameter hole were put on the glass surfaces to form the small-volume immunowells. Anti-mouse IgG (50pL of 24 pg/mL solution) and 1.2 W MEDC were pipeted into the wells, mixed, and then incubated for 1 h. After 1-h incubation, the solutionwasremovedfrom the wells by aspiration and the wells were rinsed with 1M NaCl solution two times. The wells filled with 100 pL of acetateTBSA buffer were then stored in the refrigerator overnight. Assay Procedure. Mouse IgG standard solutions (10 pL) were incubated in the wells at room temperature for 30 min, and the wells were rinsed with acetateTBSA buffer (100 pL) three times for 5 min each. After removing the rinsing solution, 10pL of rat anti-mouse IgC alkaline phosphatase conjugate was added to the wells and incuhated for 30 min. The wells were rinsed with tris huffer (100 pL) three times for 5 min each. Following this step, 20 p L of substrate (PAPP)solution was added to the wells, and the enzymatic reaction was allowed to proceed for 20 min. After enzymatic reaction, 2-10 pL of the solution was pipeted onto the microelectrode cell. The limiting currents for PAP (product of enzymatic reaction) obtained from voltammograms were used to prepare the standard calibration curves.

RESULTS A N D DISCUSSION Quantitative Determination of 4-Aminophenolat the IDA Electrodes. The CV measurement of PAP was made in pH 9 tris buffer solution. This pH is optimal for the enzymatic reaction of alkaline phosphatase, which is the enzyme label for the electrochemical immunoassay. This pH is also advantageous for the electrochemical reversihility of PAP hecause oxidized PAP (quinoneimine) reacts with hydrogen ion helow neutral pH to destroy electrochemical reversihility. On the other hand, electrochemicalreversihility of PAP is maintained in the hasic solution due to the very low concentration of hydrogen ion. Figure 2 shows the voltammograms of 1mM PAP measured in pH 9 tris buffer solution using the IDA electrode. When the potential of one finger set in the IDA (anode) was cycled without potentiostating the other finger set (cathode), the oxidation and reduction peaks were observed in the vol-

ANALYTICAL CHEMISTRY, VOL. 65, NO. 11, JUNE 1, 1993

,

Twin mode

1561

Single mode

cathodic

-101

, (01

-0.3

,

0

\,

II , ibl, 0 3 -03 0 E/ V

03

Flguro 2. Cyclic voltammograms of 1 mmoi/L PAP in 0.1 moi/L tris buffer at an IDA electrode with and wlthout potentlostatingthe cathode. The wldth of each finger is 3 pm, and the gap is 2 pm. The sweep rate of the anode is 50 mV/s. Ag/AgCi and gold wire were used as

reference and auxiliary electrodes, respectively.

IlOOnA

-

01

0 0.3 EIV Flgut. 3. Cyclic voltammograms of 10 pmoi/L PAP in 0.1 moi/L tris buffer with potentiostatingof the cathode at -0.3 V. The width of each finger is 3 pm, and the gap is 5 pm. The sweep rate of the anode is 50 mV/s. Ag/AgCi and gold wire were used as reference and auxiliary electrodes, respectively. -02

tammogram of PAP (Figure 2b). The shape of the voltammogram is not an ideal reversible response with relatively large peak separation, which is different from the results obtained using well-polished glassy carbon (G0.15 Since the IDA electrode was made by vacuum deposition and cannot be polished before the measurement, the electrode surface might be less active than that of the GC electrode. The reduction wave of the voltammogram is less reversible than the oxidation wave, suggesting that the reduction reaction is slower than the oxidative reaction. In spite of the nonideal CV response of PAP, ideal anodic and cathodic limiting currents were obtained, as shown in Figure 2a, when the potential of the anode was swept while setting the cathode potential at -0.3 V. The magnitude of limiting current is much larger than the peak current without potentiostating the cathode (Figure 2b). These results clearly indicate that high redox cyclingof PAP is established between anode and cathode. The less reversible cathodic response might be improved by constantly applying much more negative potential than the cathodic peak potential of PAP to the cathode. The IDA electrode response is clearly more sensitive for the detection of PAP compared with the single electrode. Figure 3 shows voltammograms of 10 pM PAP at the scan rate of 50 mV/s. The cathodic limiting current is clearer than that of the anodic one due to the absence of a large background current. The anodic current includes both charging and faradaic currents from oxidation of the electrode surface. In contrast, the cathode is held at a constant potential and electrochemically generated products except PAP are not detected at the cathode due to electrochemical irreversibility. Thus, more sensitive detection is possible using the cathodic signal for the detection of PAP. This charging current-free voltammetric measurement is important to improve sample throughput by faster voltammetry. When

Figure 4. Variation of anodic and cathodic limiting current of an IDA as a function of PAP concentration. Finger width and gap of the IDA are 3 and 2 pm, respectively.

Table I. Magnitude of the Limiting Current and Half-Wave Potential for a 5-pL PAP Sample Measured with the IDA Compared with Those of a 10-mL PAP Samples samples anode cathode EIIz/mV 5 p L of PAP at the IDA using 13.5 12.5 112 Ag REband AEc on the tip 10 mL of PAP at the IDA using 13.6 12.4 43 external Ag/AgCl REband gold AEc The finger width and gap of the IDA are 3 and 2 pm, respectively. The concentration of PAP is 1 mM. The scan rate of the anode is 100 mV/s, and the cathode potential is set at -0.3 V vs Ag plated or Ag/AgCl reference electrodes. RE, reference electrode. AE, auxiliary electrode.

the sweep rate increases in the CV measurement, the S/N ratio between the faradaic current and the charging current decreases. In the case of the IDA, a relatively high sweep rate can be used for the measurement due to the absence of charging current at the cathode. Figure 4 shows the calibration curve of PAP measured with the IDA electrode. The anodic and cathodic limiting currents are proportional to concentration from 1to lo00 p M , which sensitivity is enough for our assay. Reduction of the Sample Volume of PAP for Electrochemical Detection. Reduction of the sample volume can be very important in an immunoassay, because some analytical samples are restricted to a very small amount. The voltammogram of a 10-pL PAP sample was obtained at the IDA electrochemical detector without using any extra electrodes. One of the gold electrodes in the electrochemicalcell was used as a reference by plating silver on the surface. Table I shows the comparison of the electrochemical response between the large-volume(10 mL) and small-volume (10pL) samples. The magnitude of the signal from both samples is almost the same except for the potential difference. This potential shift is due to the difference in half-cell potential of the silver-plated reference electrode compared to the external Ag/ AgCl electrode. The potential of the silver-plated gold electrode is stable following an initial potential change just after silver plating, so long as the IDA is used in the same buffer. The potential of the silver reference during electrolysis of PAP is also stable, because almost all oxidized PAP molecules generated at the one finger set of the IDA are reduced at the other set. Very little oxidized PAP diffuses to the reference electrode surface. The results suggest that use of the silver-platedelectrodeas a reference is not a problem for the quantitative determination of PAP. The sample throughput is also improved by integrating all the electrodes on one tip. The measurement of the voltammogram takes less than 1 min including the pipeting of PAP onto the

1562

ANALYTICAL CHEMISTRY, VOL. 65, NO. 11, JUNE 1, 1993

3uL solution

in LLum TLC

f

:51;

c

z

,

f

I

3

0

-0.3

0

-0.3 E/V

0

Figure 5. Comparison of the anodlc llmiting currents between 3 pL of PAP plpeted on the IDA and 800 nL of PAP In the IDA thln-layer cell. The reference electrode Is a silverplated gold electrode on the tip. Flnger width and gap of the IDA are 3 and 5 pm, respectlvely. The scan rate of both voltammograms Is 50 mV/s.

OK io

is

I

io

is

io

is

lo

Substrate incubation time (min)

Flgure 7. PAP production

as a function of Incubation tlme for a 100

ng/mL mouse I@ standard. Finger wldth and gap of the IDA used In this measurement are 3 and 5 pm, respectlvely. Blank

01

0

50

100 150 Timekec

200

250

Figure 6. Tlme dependence of the PAP llmitlng current at the IDA in the 44pmthlnhyer cell, obtained with and without potentiostatlng the cathode at -0.3 V. Finger width and gap of the IDA are 3 and 5 pm,

respectlvely.

-0.3

0 EIV

0.3

-0.3

0

0.3

E/V

Figure 8. Cathodic limiting currents of substrate solutions incubated for 20 mln In the lmmunowell, which Is used for the assay of 0 and 100 ng/mL mouse I@. Each flnger w M and gap of the IDA Is 5

Irm.

electrode and cleaning the electrode surface. This is much faster than LCEC and is comparable to FIAEC analysis. To detect much smaller volume samples, a thin-layer IDA electrochemicalcell with a thickness of 44 pm was fabricated by sandwiching the adhesion tape between the IDA cell and a glass plate. The inside volume of this cell is 800 nL. Whereas the conventional electrode exhibited infinite diffusion of PAP, finite diffusion was achieved in this thinlayer cell in 1.35 s, as calculated with the Einstein equation: L = (2Dt)l/Z,where D is the diffusion coefficient of PAP (7.1 x 10-6 cmZ/s, as measured with a rotating disk electrode), t is time (a), and L is the thickness of the diffusion layer (cm). Figure 5 shows the anodic limiting current of 0.5 mM PAP in the thin-layer cell compared with that of a 3-pL PAP (0.5 mM) sample pipeted on the cell without the thin-layer restriction. The magnitude and the shape of these voltammograms are exactly the same, indicating that the diffusion layer of PAP around the IDA electrode is mainly formed between the anode and cathode of the IDA, and the effect on the diffusion layer normal to the electrode is very little. This result clearly suggests that the IDA electrode can be successfully applied to the measurement of very small samples. When the sample volume is too small compared with the surface area of the electrode, consumption of the analyte by the electrochemical reaction rapidly changes the analyte concentration and influences the magnitude of the signal. The stability of the limiting current at the IDA was compared with that at a single electrode. Figure 6 demonstrates the time dependence of the limiting current at the IDA by holding the potential of the anode at 0.3 V. One curve was obtained by potentiostating the cathode at -0.3 V; another curve was obtained without using the cathode. The current of the IDA measured without the cathode decreases rapidly and its magnitude becomes 15% of ita original value in the first 1 min. This is due to the consumption of PAP by the electrochemical reaction. In the case of the IDA with potentiostating of the cathode, the limiting current gradually decreases with increasingtime.

However, it maintains 75% of the original value after 4 min, indicating that the cathode of the IDA stabilizes the limiting current at the IDA by its regeneration of PAP.

Electrochemical Small-Volume Immunoassay Using IDA Electrochemical Detector. Since a sandwich enzyme immunoassay with detection by FIAEC and LCEC for mouseIgG has been developed,5J6 this assay is convenient for examining the utility of replacing FIAEC or LCEC with the IDA cell. In the previous work, the primary antibody was covalently immobilized on the surface of glass capillaries by a coating of poly(vinylbenzy1chloride) (PVBC) with EDC after introducing an amino group to the PVBC surface.5 In this assay, we substituted the hydroxyl group of the glass for the amino group by reaction with (3-aminopropy1)triethoxysilaneand coupled to anti-mouse IgG with EDC. This is because PVBC is too hydrophobic to cover all the well surface with a 10-pL sample solution, and direct use of the glass surface instead of casting PVBC films on the surface was found to give more reproducible results. During the incubation of mouse IgG, enzyme-labeled antibody, and substrate, each well was covered with a glass plate to prevent evaporation of the sample solution. To obtain the appropriate magnitude of the signal, 4 mM PAPP was incubated in the wells using 100 ng/mL mouseIgG and a 1:lOOO dilution of labeled antibody. Figure 7 shows the effect of varying incubation time on current at constant IgG and labeled antibody concentration. The 20-min incubation time was chosen in our assay because the relationship between signal of PAP and incubation time is almost linear up to 20 min, and a sufficiently large signal for the detection was obtained at this incubation time. Nonspecificadsorption of IgG and labeled antibody can be suppressed using not only Tween 20 and BSA but also 1-pentanesulfonic acid according to the previous method.5 Figure 8 shows the signal for 100ng/mL mouse IgG compared with that for the blank response. The blank signal is

ANALYTICAL CHEMISTRY, VOL. 65, NO. 11, JUNE 1, 1993

1563

The standard calibration curve for mouse IgG is shown in Figure 9. The signal obtained in this assay is very large, more than lo00 nA at high IgG concentration, due to the large surface/volume ratio of the small-volume incubator and electrochemical amplification at the IDA. The range of this assay is from 10 to lo00 ng/mL; the signal saturated a t more than lo00 ng/mL mouse IgG, because the antibody concentration on the surface is insufficient for such a high IgG concentration, A detection limit of 10 ng/mL, which is 0.6 fmoV 10pL sample was obtained. The sensitivity of this assay is much better than that of the previous assay due to the very steep slope of the calibration curve. This could be due to variation of the binding constant of the antibody itself due to a different immobilizationmethod. During this assay, the IDA electrode can be used for detection repeatedly due to the absence of electrode fouling by PAP.

dL 10 102

103

10'

io5 io6

10'

Corcentrotion(pg/rnll

Figure Q. log-log plot of the cathodic limltlng current vs mouse IgG concentration In acetateTBSA. Each finger width and gap of the IDA is 5 pm.

suppressed even at the 20-min incubation time of PAPP. A clear difference between the 100 ng/mL IgG and the blank signal is observed, indicating that the immunochemical reaction has occurred in the well.

RECEIVEDfor review February 12, 1993. February 16, 1993.

Accepted