2466
Anal. Chem. 1989, 67, 2466-2471
Pulsed Amperometric Detection of Alkanolamines following Ion Pair Chromatography William R. Lacourse,* Warren A. Jackson, and Dennis C. Johnson Department of Chemistry, Iowa State University, Ames, Iowa 5001 1
The quanlltatlve determhtbn of alkadadnes Is not always shnple due to the lack of an Inherent chromophore or fluor+ phore. I n addition, separation by gas chromatography or liquid chromatography is dmkuH due to the high polam of these compounds. Ion pair chromatography was demonstrated for the separation of mono-, &, and ttialkandamlnes using a C-18 reversed-phase column with a mobile phase consisHng of an aqwous/acetonlMle/dodec/dodecanesullonate salt mixture. On-Rne pulsed SmpcHOmetrlc detection (PAD) of the alkandamlnes at a Au skotrodo tdkwed postcokmn addllbn of an alkaline bufier to the column effluent. PAD utllked a three-step potential waveform that combined amperometric detectlon with alternating anodic and cathodic polarizations to maintain electrode actlvity.
INTRODUCTION With the present and ever-increasing concern related to the levels and fates of environmental pollutants, electrochemical detection with liquid chromatography has gained prominence as a useful analytical technique (I). This popularity can be attributed to the efficient separations achieved by contemporary chromatography and the tunable selectivity and low detection limits of electrochemical determinations. A major shortcoming of amperometric detection at noble metal electrodes for a constant (direct current (dc)) applied potential has been the loss of activity during anodic detections of organic compounds resulting from adsorption of carbonaceous products of the ensuing reactions (e.g., radicals). Pulsed amperometric detection (PAD) based on a multistep waveform (E-t) overcomes the problem of lost activity on noble metal electrodes by alternating amperometric detection with anodic and cathodic polarizations to clean and reactivate the electrode surface. PAD efficiently exploits the catalytic activity of the clean electrode surface oxide to oxidize aliphatic molecules, which typically do not have strong chromophores. Liquid chromatography with PAD has been shown to be a simple, selective, and sensitive method for the determination of alcohols, polyalcohols, carbohydrates (2),amino acids (3). and many inorganic and organic sulfur-containing compounds (4). The determination of alkanolamines is currently of great interest, as these compounds are used widely in chemical and pharmaceutical industries. Aikanolamines are needed for the production of emulsifyingagents, corrosion inhibitors, laundry materials, dyes, medicines and for purifying gases (5). In addition, environmental concerns, raw material shortages, and rising energy costs are behind the push to remove hydrocarbon-based solvents from coating systems. Water is now becoming the major solvent of choice; and in the switch from organic solvents to water, alkanolamines are making valuable contributions. The quantitative determination of alkanolamines is not always simple due to the lack of an inherent chromophore, fluorophore, or dc-active electrophore. Furthermore, chromatographic separations are difficult due to the high polarity of these compounds (6). In liquid chromatography, the effect of the amine functionality on silica-based reversed-phase columns can manifest itself in significant tailing of the 0003-2700/89/0361-2466$01.50/0
chromatographic peak. Derivatization with nitroaromatic moieties is often used to improve both the chromatographic behavior and detection properties of alkanolamines (7).Ion pair chromatography followed by postcolumn addition of alkaline buffer with pulsed amperometric detection is ideally suited to the efficient separation and detection of alkanolamines. Determinations of alkanolamines in “real-world” sample matrices are simple, direct, and sensitive. Described here are the development, optimization, and application of LC with PAD for the determination of alkanolamines. The mechanism of detection is discussed to emphasize the basic tenets of PAD. Also, the compatibility of ion pair chromatography with PAD is stressed. EXPERIMENTAL SECTION Apparatus and Procedures. Voltammetric data were obtained at a Au rotated disk electrode (RDE) (Pine Instrument Co., Grove City, PA) with a computer-aided electroanalysis system (Cypress Systems,Lawrence, KS) or a Model RDE4 potentiostat (Pine). Liquid chromatographic work employed an isocratic/gradient chromatography system (Dionex Corp., Sunnyvale, CA). Alkaline buffer, pumped by a Rabbit-HP solvent delivery pump (Rainin Instrument Co., Woburn, MA), was added to the column effluent through a tee connector followed by a woven Teflon mixing coil. Separations were performed with a C-18 pBondapak column (Waters Chromatography Division, Millipore Corp., Milford, MA) or a Hamilton PRP-1 column (Phenomenex, Rancho Verdes, CA). Mobile phases were filtered through a 0.45-pm Nylon 66 filter (Rainin) with the use of a solvent filtration kit (Waters). PAD was performed with either a Model UEM or a Model PAD-2 detector @ionex). Two electrochemical cell confiitions were used interchangeably. A homemade cell consisted of a Au-wire working electrode, a Pt counter electrode, and a saturated calomel reference electrode (SCE).. A thin-layer electrochemical cell (Dionex)consisted of a planar Au working electrode, a glassy carbon counter electrode, and a SCE. Reagents. All solutions were prepared from reagent grade chemicals as received. Mobile phase solvents were from Fisher Scientific Co. Water was purified in a Millipore Milli-Q system or a Barnstead Nanopure I1 system, followed by filtration (0.2 rm).
RESULTS AND DISCUSSION Voltammetry. The current-potential (I-E) response is shown in Figure 1 for a Au RDE in 10% acetonitrile/W% water, containing 0.1 M NaOH with 1mM sodium dodecansulfonate, with (-) and without (- - -) ethanolamine. The residual response for the supporting electrolyte (- - -) exhibits anodic waves on the positive potential scan in the regions H . 2 to +0.9 V (A) for oxide formation and E > 0.8 V (B) for O2 evolution, and a cathodic peak on the negative scan in the region +0.2 to -0.2 V (C)for dissolution of the oxide formed on the positive scan. Cathodic reduction of dissolved O2 is observed in the region -0.2 to -1.0 V (D)for both the positive and negative scans. For the presence of ethanolamine (-), an anodic wave is observed on the positive scan in the region of ca. -0.2 to +0.3 V (E), where alcohols are observed to be oxidized, and in the region ca. +0.3 to +0.7 V (F) for oxidation of adsorbed amine simultaneously with the formation of surface oxide on the Au electrode (A). The absence of any anodic signal on the negative scan in the region +1.0 to +0.2 0 1989 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 61, NO. 22, NOVEMBER 15, 1989
1.0
0.6
0.2
-0.2
POTENTIAL
-0.6
-1.0
(mV)
Figure 1. Voltammetric response ( I - € ) for ethanolamine at a Au rotating disk electrode. Conditions: 10% acetonitrle/O.l M NaOH/1 mM sodium dodecanesutfonate,nondegassed, 1000 revolutions min-' rotation speed, 100 mV s-l scan rate. Sample concentration: (---I 0 mM ethanolamine, (-) 16 mM ethanolamine.
V is clear evidence of the absence of any reactivity of the oxide-covered surface for ethanolamine oxidation. When the oxide is cathodically dissolved on the negative scan to produce peak C, the surface reactivity for ethanolamine oxidation is immediately returned and an anodic peak G is observed for the alcohol functionality of ethanolamine. Anodic waves E, F, and G all were observed to be increased in height with increases in the ethanolamine concentration. The net anodic current in wave E for detection of the alcohol functionality of ethanolamine increased markedly with increases in electrode rotation speed, yet exhibited negligible change with variations in the potential scan rate. These observations are consistent with the conclusion that the mechanism producing wave E for the adsorbed amine functionality of the ethanolamine is under mass-transport control. In contrast, wave F increased with increases in the potential scan rate, yet showed very little change as a result of variations in electrode rotation speed. These observations are consistent with the conclusion that the anodic reaction for wave F is under the control of an electrode surface process. This is consistent with the proposed mechanism of an oxide-catalyzed oxidation of the adsorbed amine. The rates of the various anodic processes producing waves E, F, and G all increased with increasing pH, and pH > ca. 11 was concluded to be optimum for analytical applications. Figure 2 shows the voltammetric responses for the pure mobile phase (- - -) as well as for separate solutions of mobile phase containing ethanol -), ethylamine (. ..), and ethanolamine (-). Waves E, F, and G for ethanolamine (-) are the same as discussed for Figure 1. Ethanol present in the mobile phase (- -) does not exhibit an anodic response in this media, even though alcohols are known to be detected anodically in the region of wave E in pure 0.1 M NaOH. This observation is concluded to be the result of an interference by acetonitrile (ACN) with the detection mechanism for the alcohol. ACN is strongly adsorbed at Au electrodes, and the adsorbed ACN is concluded to block surface sites required for a preadsorption step in the anodic detection of the alcohol. In comparison, the amine functionality of ethylamine (. ..) is strongly adsorbed at the Au surface and exhibits an oxide-catalyzed wave on the positive scan in the region corresponding to wave F for eth(-a
1200
000
400
0
2487
-400
POTENTIAL (mV)
Figure 2. Voltammetric response ( I - € ) for ethanol, ethylamine, and ethanolamine at a Au rotating disk electrode. Conditions: 10% acetonitrlle/O.1 M NaOHll mM sodium dodecanesulfonate, nondegassed, 1000 revolutions min-' rotation speed, 100 mV s-' scan rate. Solutions: (---) supporting electrolyte only, ( - e - ) 16 mM ethanol, (. . .) 16 mM ethylamine, (-) 16 mM ethanolamine.
anolamine (-). At extremely high concentrations of ethanol (i.e., > ca. 1M), anodic waves were observed in the potential region corresponding to waves E and G of ethanolamine, confirming the conclusion that these waves for ethanolamine are the result of oxidative detection of the alcohol functionality. The presence of an anodic response for the alcohol functionality of ethanolamine (wave E), whereas no response was observed for ethanol, is concluded to be a beneficial consequence of the adsorption of the amine functionality of the ethanolamine, which occurs even in the presence of ACN on the oxide-free Au surface (i.e., E < ca. 0.0 V). The fact that wave E for ethanolamine is under virtual mass-transport rather than surface control is consistent with the conclusion that the adsorption of amine groups is reversible,and following oxidation of the n-alcohol group to a carboxylate, rapid desorption of the ionic aminocarboxylate occurs. Since surface oxide is not formed on the positive scan in the region of wave E, unreacted ethanolamine is adsorbed following desorption of the carboxylate product to support the continuous flow of anodic current. The oxidation of adsorbed amine requires the simultaneous formation of electrocatalytic surface oxide, and therefore, the anodic process (wave F) exhibits the characteristics of a surface-controlled reaction. The maximum anodic response for wave F for a given ethanolamine concentration is expected to be limited by the adsorption isotherm, vide infra, for the particular alkanolamine being detected. The fully developed oxide-covered surface exhibits insignificant catalytic reactivity, and detection of the amine group of ethanolamine a t Au electrodes is only possible via a transient electrocatalytic mechanism achieved within a pulsed potential waveform. All other alkanolamines display voltammetric behavior similar to that of ethanolamine. A t E > ca. +0.7 V on the positive scan, the anodic response of the amine is rapidly attenuated due to the unreactivity of the fully developed surface oxide. On the subsequent negative scan, the formation of surface oxide ceases with the result of a zero value of electrode current. The surface oxide is cathodically dissolved in the region ca. +0.2 to -0.2 V (wave C in Figure 1)with the result that fresh
2468
ANALYTICAL CHEMISTRY, VOL. 61, NO. 22, NOVEMBER 15, 1989
c IOso;
c8 so;
css 0;
c I2 so;
Flgure 3. Effect of ion pair alkyl chain length on alkandamine Separation. Mobile phase: 20% acetonltrile/2 mM paked lon reagent at 0.75 mUmln. Paked ion reagent: (A) sodkvn hexanesulfonate, (B) sodium octanesulfonate, (C) sodium decanesulfonate, (D) sodium dodecanegulfonate. Other cond#kns: postcolumn reagent, 0.2 M NaOH at 0.25 mL/min; column, Waters 10-pm, C18 pBo&pak; waveform, Table I;lnjectbn vdune, 200 pL. Samples (10 ppm): (a) 2amlno-l-ethanol, (b) 4-amino-I-butanol, (c) 5-amino-1-pentanol, (d) Pamino-1-butanol, (e) &amino-lhexanol. a
T
100 n A
1
b
,
2
I
I
I
4
6
B
CONCENTRATION
(mM1
Figure 4. Effect of paked ion reagent concentration on k'. Conditions were the same as in Figure 3, except that the mobile phase was 20% acetonltrile/2 mM sodium dodecanesulfonate.
ethanolamine is adsorbed with concurrent oxidation of the alcohol functionality. The oxidation products of the n-alcohol and amine groups are tentatively concluded to be the corresponding carboxylate (8) and hydroxylamine (9); however, proof must await further research. Liquid Chromatography. Alkanolamines are ideal candidates for ion pair separation on reversed-phase columns. Alkanesulfonate salts are effective ion pairing reagents, are PAD inactive, and can be obtained easily with variable alkyl chain lengths. Figure 3 illustrates the effect of changing the carbon number of the ion pairing reagent on the separation of five alkanolamines. Base-line resolution was achieved with dodecanesulfonate anion as the ion pairing reagent. The optimum ion pairing reagent concentration was found to be ca. 2 mM from the plots of capacity factor versus reagent concentration shown in Figure 4. At this concentration, resolution was achieved without excessive retention times. As expected, the capacity factor ( k 9 was determined to decrease as the concentration of organic modifier (ACN) in the mobile phase was increased. Retention of the alkanolamines also
/
0
'
1
10
'
1
20
'
1
'
l
30 40 T I M E (min)
'
l
50
'
1
60
~
l
70
Flgure 5. Results of the separation of several alkanolamines with pulsed amperometric detection (PAD). Condltions were the same as Figure 3, except that the mobile phase was 20% acetonMW2 mM sodium dodecandonate. Samples (25 ppm): (a) 2amiwldhanol, (b) 3-amIno-l-propano1, (c) 4-amlno-l-butanol, (d) 5amino-l-pentanol, (e) 6-amino-l-hexano1, (f) 2-amIno-l-propano1, (9) 2-amlno-l-butanol, (h) 2amino-l-pentanol, (i)lamino-2-propanol, (j) Pamino-2-methyl1-propanol, (k) 2-amino-I-phenylethanol.
decreased with an increase in pH >5, due to deprotonation of the amine functionality, which results in less ion pair formation, and the effect of an additional ionic strength of the mobile phase. A t higher pH values, equilibration time of the chromatographic column appeared to be increased. The separation of five alkanolamines was achieved with a mobile phase of 20% acetonitrile/2 mM sodium dodecanesulfonate. The LC-PAD separation of linear, branched, and complex alkanolamines is shown in Figure 5 to illustrate the range and
ANALYTICAL CHEMISTRY, VOL. 61, NO. 22, NOVEMBER 15, 1989
2489
If
01
0
'
I
10
'
'
"
20 30 CONCENTRATION K N 0 3
40
" 50
(mM)
-
vvw
Flgure 8. Effect of NaNO, concentration on k'. Condklons: mobile phase, 20% acetonitrilel2 mM paired ion reagent/NaNO, at 0.75 mL/min flow rate; waveform, Table I; injection volume, 200 pL. Sample (ca. 10 ppm): (a) 2-amino-l-ethanol, (b) 4-amlno-l-butanol, (c) 5-amlno-l-pentanol,(d) 2-amino-l-butanol,(e) &amino-1-hexanol.
Y il
applicability of ion pairing chromatography with pulsed amperometric detection of alkanolamines. Monoethanolamine and triethanolamine were well resolved with capacity factors of ca. 14 and 15, respectively. Diethanolamine coeluted with ethanolamine under these experimental conditions. No further attempt was made to optimize the resolution of the substituted alkanolamines. The optimized chromatographic conditions were developed for the pBondapak C-18 column. In general, the PRP-1 column required less ion pairing reagent and less organic modifier to achieve comparable peak resolutions, in comparison to the silica-based C-18 column. The ion pairing technique was compatible with both silica and polymeric supported C-18 stationary phases. As we observed for the organic modifier, the capacity factor ( k ? decreased as the concentration of electrolyte (KNO:, or NaN03) was increased in the mobile phase, as is shown in Figure 6. From Figure 6, it is noted that the early-eluting compounds were affected less than the late-eluting compounds by the added electrolyte. Gradient elution by addition of NaN03 was performed on the Hamilton PRP-1 column. the k ' values for the late-eluting compounds were reduced significantly, e.g., from ca. 50 to 20 for 2-amino-1-pentanol, whereas the resolution between all compounds was maintained, as shown in Figure 7. To obtain reproducible results when NaN03 gradient elution was employed, it was found necessary to equilibrate the column with the weak solvent for at least 1 h between injections. The capacity fador (k? for each of the alkanolamines was within *2% relative standard deviation (RSD (n = 3) under the proposed conditions. Sodium hydroxide (0.2 M) was added postcolumn at a ratio of 1:3 (i.e., 0.25 mL/min:0.75 mL/min). The NaOH provided the appropriate pH >11 and necessary ionic strength for efficient electrochemical detection. The necessity of high pH is attributed to the need to deprotonate the amine functionality to accommodate its adsorption on the electrode surface. Liquid Chromatography with Pulsed Amperometric Detection. Table I describes the optimized triple-step waveform for the pulsed amperometric detection of ethanolamine. Upon stepping to the detection potential ( E l ) , the anodic signal was integrated for 16.7 ms after a delay of 280 ms. Since the electrode activity is diminished due to inhibition by the fully developing surface oxide and by fouling of the surface by adsorbed oxidation products, electrode activity was
Flgve 7. Gradient separation of several alkandamines. Mobile phase: solvent 1,0.75 mM dodecanesulfonate/l5% ACN; solvent 2,0.75 mM dodecanesulfonate/l5%ACN/10 mM NaN0,. Program: 0-50 min, 100% solvent 1 and 0% solvent 2; 50-70 min, ramp to 10% of solvent 2; 70-71 mln, ramp to 100% of solvent 2 and hold. Other conditions: flow rate, 0.5 mL/min; postcolumn, 0.3 M NaOH at 0.5 mL/mln; column, PRP-1 C-18; waveform, Table I. Injection volume, 200 pL. Sample: (a)2-amino-l-ethanol,3 ppm; (b) 4-ambl-butanol, 4 ppm; (c) 5-amino-1-pentanol,4 ppm; (d) 2-amlno-l-butanol, 3 ppm; (e) &amino-lhexanol, 2 ppm; (f) 2-amino-1-pentanol, 2 ppm; (9) 2amino-1-hexanol, 4 ppm.
I
~
0
1
'
24
l
'
~
\
'
48
l
'
l
'
TIME
I
~
I
84
60
'
IO8
(min)
Table I. Waveform Specifications for Pulsed Amperometric Detection
2
5 Iz Y
IO n
TIME
potential,
mV vs SCE
E l , 200 ED ES, -400
function
time, ms tl, 300 t d , 280 t,, 16.7 t2, 400 t,, 350
detection
delay
current sampling
oxidative cleaning reductive reactivation and adsorption of analyte
renewed by subsequent positive and negative potential steps to achieve anodic (E,) and cathodic (E,) polarizations, respectively. During the application of E3,adsorption of analyte can occur on the oxide-free surface prior to the next detection step. Figure 8 shows the anodic response of ethanolamine as a function of the duration of the third step ( t s )in the waveform, while the other parameters of the waveform are kept the same as given in Table I. Clearly, the response increased
I
'
I
2470
ANALYTICAL CHEMISTRY, VOL. 61,NO. 22, NOVEMBER 15, 1989 ~~~
Table 11. Results of a Single-Blind Study of Water Samples Spiked with Ethanolamine actual, ppm
found, ppm
percent of actual
no
7.76 4.34 1.92 0.20 0.17
7.92 f 0.18 4.40 f 0.03 1.94 f 0.07 0.19 k 0.03 0.16 f 0.03 overall
102 f 2.3 101 f 0.7 101 f 3.5 94 f 13 93 f 19 98 f 4.3
4 3 3 3 3 16
A- i
Number of determinations,
0
0
100
200
400
300
ADSORPTION T I M E
500
(mS)
Figure 8. Response for pulsed amperometric detection of ethanolamine at a Au electrode with variation of adsorption time ( t 3 )in the PAD waveform. Conditions were the same as in Figure 5,except that t , in Table I was varied from 0 to 500 ms.
-
...........................
0
IO 20 30
,
0
20
40
60
0
10 20 30
I
40 50
y_-__-__-_______ -10%
3
-0.20
I
T I M E (mi”)
Figure 10. Detection of alkanolamines in various samples by LC with PAD. Conditions were the same as in Figure 5. Samples: (a)generic
cold tablet, (b) hand lotion, (c) halrspray. ~
approached, the standard deviation increased, as expected. The overall percent of spiked ethanolamine was found to be -0.40~ , l , . , . , . . , . , , . , . . . . , . . . . , .98.2 . . f. 4.3%. -0.50 0.00 0.50 1.00 1.50 2.00 2.50 The response factors for mono-, di-, and triethanolamine were 2.7, 4.0, and 2.7 nA/pmol, respectively. Since the deLOG CChC3lRATION 0 tection potential was fixed to monitor oxidative detection of Flgure B. Normalized calibration curve for ethanolamine. Each dot the alcohol functionality, the increase in response of the direpresents the average of six determinations. Dots without error bars have average deviations less than the size of the dot. A represents ethanolamine as compared to that of monoethanolamine is percent deviation from linear fR of data determined by linear regression concluded to be the result of the presence of two alcohol analysis ( 7 7 ). groups versus one. However, triethanolamine, expected to have a response factor greater than that for the diethanolfor increased time allowed for analyte adsorption and became amine, was approximately equivalent in response to ethanoconstant for t3 > ca. 200 ms. In a separate experiment, the lamine. This low sensitivity is tentatively concluded to be a i-t curve was recorded, and it was determined that complete consequence of steric hindrance from the three alkyl arms as reduction (>90%) of the surface oxide at E3 = -0.4 V required well as desorption of the carboxylate product of the oxidation approximately 180 ms. Hence, the data in Figure 8 are conof only one alcohol group. cluded to indicate for this concentration of ethanolamine (22 The assay for alkanolamines was applied to several matrices ppm) that the rate of analyte adsorption was controlled by to illustrate the analytical utility of the assay. The high the rate of oxide dissolution rather than analyte flux. Furselectivity and sensitivity of LC-PAD contributed to dethermore, the value 200 ms is recommended as the lower limit creased time for sample preparation and simplified chromafor t3. These results are also consistent with the conclusion tograms. Figure 1Oa-c shows chromatograms for the detection that preadsorption of the ethanolamine on the oxide-free of ethanolamine, triethanolamine, and aminomethylpropanol surface is required for detection of the alcohol functionality in a cold tablet, hand lotion, and hairspray formulation, rein the presence of ACN. spectively. The samples were all prepared by diluting a Statistical analysis of calibration data was based on a weighed portion of sample, dilution to the proper level with modified regression analysis which assumed that variance in mobile phase, and filtering. the signal is proportional to concentration (10). PAD response was concluded to be linear for 1-100 ppm, as is indicated in CONCLUSIONS Figure 9. This style of presentation of calibration data has The determination of underivatized alkanolamines by ion been presented previously (11). The limit of detection was pair chromatography with pulsed amperometric detection is determined to be 40 ppb (200 pL, 8 ng) for a signal-to-noise direct, selective, sensitive, and simple. Ion pair chromatogratio of 3. Peak skew and repeatability were 1.1 f 9.4% RSD raphy is compatible with PAD with no evidence of electrode (n = 3) and f3.270 RSD (n = 6, 0.083 mM ethanolamine), fouling even at high concentrations of the ion pairing reagent respectively. Synthetic water samples spiked with ethanolfor extended run times. amine were assayed in a single-blind study in order to further ACKNOWLEDGMENT validate the LC-PAD methodology. The recovery of spiked ethanolamine was found to be 93-102% for concentrations We gratefully acknowledge the assistance and support of of 0.16-7.92 ppm (Table 11). As the limit of quantitation was Lawrence Welch.
Anal. Chem. 1909, 61, 2471-2474
2471
LITERATURE CITED (9) Jackson, W. A.; Johnson, D. C., work in progress. (10) Naturella, M. Exptwhmtal Stalbrcs; National Bureau of Standards (1) Klssinger, P. T. J . Chem. Educ. 1983, 60, 308. kndbodc 91; NBS: WasMngton, DC. 1963; pp 6-19. (2) Johnson, D. C. Mtue 1986, 327, 451. (11) Johnson, D. C. A M I . C h h . Acte 1988, 209. 1. (3) Welch, L. E.; LecOuse, W. R.; Mead, D. A., Jr.; Hu, T.; Johnson, D. C. Anal. them. 1989, 6 7 , 555. (4) Johnson. D. C.; Pdte. T. 2. C h t o g r . Forum 1988, 7 , 37. RECEIVED for review July 1,1989. Accepted August 30,1989. (5) The Merdc Index, 10th ed.; Mer& Flahway, NJ, 1983: p 541. The majority of this work was supported by a grant from (6) Heberer, H.; Bmersohl, G. Z.Chem. 1980, 20, 361. (7) CbnQ, M.-Y.; C k n , L.-W Dong, X.-D.; Selavka. C. M.; Krull, I. S. J . Dionex Gorp., Sunnyvale, CA. The National Science FaunCtrometogr. Scl. 1987, 25, 480. dation contributed a portion of the funding through Contract (8) brew, L. A.; Johnson. D. c. J . ElectraaMi. m m . Intdackl ~ k W-m. 1989. 262, 187. CHE-8312032.
Detection of Rat Basophilic Leukemia by Cyclic Voltammetry for Monitoring Allergic Reaction Tadashi Matsunaga,* Akinori Shigematsu, and Noriyuki Nakamura Department of Biotechnology, Tokyo University of Agriculture & Technology, Koganei, Tokyo 184, Japan
Electrochemlcal detection of the rat basophlllc leukemia (RBL-1) cells has been carrled out by applying cycllc vdtammetry. The detection system consists of a basal plane pyrotytk graphite ekctrde and a porous nitroceliulose membrane fllter to trap RBL-1 cells. When the potential of the graphtte electrode was run In the range of 0-1.0 V vs SCE, RBL-1 cdls gave peak currents at 0.34 V vs SCE as well as 0.65 V vs SCE. There Is a linear relationship between the peak current at 0.34 V vs SCE and the cell numbers of RBL-1 In the range of (0.4-2.0) X 10' cells. The peak current of RBL-1 cells was attrlbuted to serotonln. When dlnltrophenylated bovlne serum albumin (DNP-BSA) as a model allergen was added to RBL-1 cells sendtlzed wlth anti-DNP IgE, the peak current decreased because of the degranulatlon of RBL-I cdls kadbg to serotonln release. On the other hand, RBL-1 cells sensltlzed wlth antCDNP IgE did not respond to egg white, pollens, house d d , and mllk.
INTRODUCTION Detection of viable cells is important in a clinical field. Various electrochemical methods have been developed for determining viable cell numbers. For example, impedance measurements of culture media have been used to determine cell numbers, although the cell numbers are measured indirectly from cell metabolite, and therefore, the results obtained sopetimes do not correlated with true cell numbers. Recently, a novel method for detecting microbial cells has been developed, based on cyclic voltammetry at a basal plane pyrolytic graphite electrode (1,2). Electron transfer between microbial cells and the graphite electrode is mediated by coenzyme A existing in the cell. Both enumeration and classification of microbial cells were possible from cyclic voltammograms by using an electrode system composed of a graphite electrode and a membrane filter retaining microbial cells. However, cyclic voltammetry of animal cells has not been reported. Therefore, cyclic voltammetry using a graphite electrode was applied to animal cells such as rat basophilic leukemia (RBG1) and mouse lymphocytes. The radioimmunosorbent test (RIST) (3-5),radioallergosorbent test (RAST) (61,and skin test (4, 7)have been used 0003-2700/89/0361-247 1$01.50/0
for sensing immediate allergic reactions. However, RIST and RAST are time-consuming and demand complicated procedures giving results independent of clinical symptoms. The skin test is dangerous because it may produce anaphylaxis in man by means of serum antibodies. A simple and safe method is still required for the detection of the immediate allergic reactions. RBL-1 cells, like normal basophils and mast cells, have an immunoglobulin E (IgE) receptor on their surface. RBL-1 cells are passively sensitized by incubating homogeneously cells with IgE. Addition of the appropriate allergen to the stimulated RBL-1 cells causes degranulation, thereby releasing histamine and serotonin (8). Therefore, RBL-1 cells can be used for the detection of immediate allergic reactions. In this paper, the allergic reaction was also monitored with the electrode system using RBL-1 cells.
EXPERIMENTAL SECTION Materials. Sodium 2,4dinitrobenzenesulfonate was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Bovine serum albumin (BSA), 5-hydroxytryptamine hydrochloride (serotonin), and compound 48/80, condensation product of Nmethyl-p-methoxyphenethylamine with formaldehydewere obtained from Sigma Chemical Co. (St. Louis, MO). Monoclonal m o w anti-DNP IgE (9),purified from the ascitic fluid of BALB/c X C5,BL-F1,mice bearing the SPE-7 hybridoma, was purchased from Seikagaku Kogyo Co., Ltd. (Osaka, Japan). Other reagents were commercially available analyticalreagents or laboratmy grade materials and were used as received. Distilled-deionizedwater was used in all procedures. Preparation of Dinitrophenylated Bovine Serum Albumin (DNP-BSA). A DNP-BSA conjugate (13.8 mol of DNP/mol of BSA) was prepared by using sodium dinitrobenzenesulfonate and BSA as previously described by Eisen et al. (10). The DNP-BSA was employed as a model allergen. Preparation of Allergen Extracts. Immediate allergic reaction was performed with allergens (egg white, yolk, common ragweed pollen, evening primrose pollen, house dust, cow's milk, and DNP-BSA) and mouse anti-DNP IgE. These allergens except DNP-BSA were sonicated in the phosphate buffered saline (PBS, 1.5 mM KH2P04,7.3 mM Na,,HP04, 137 mM NaC1,2.7 mM KC1, pH 7.4) for 45 min and incubated at 4 O C for 12 h. Then allergens were centrifuged at 4 "C and 3000g for 30 min and the supernatants were obtained. Then, these extracts were passed through the sterilized membrane fiiter (pore size, 0.45 pm) and dialyzed for 3 days against 5 L of PBS. The protein concentration of allergen extracts was determined by the Lowry method (11). 0 1989 American Chemical Society
