ANALYTICAL CHEMISTRY, VOL. 50, NO. 9, AUGUST 1978
1319
Enzymatic Determination of Nitrate: Electrochemical Detection after Reduction with Nitrate Reductase and Nitrite Reductase Chih-Hen Kiang,' Shia S. Kuan, and George G. Guilbault' Department of Chemistry, University of New Orleans, New Orleans, Louisiana 70 122
A method for the determination of nitrate and/or nitrite using the dual enzyme system: MVH (methyl viologen, reduced form) nitrate reductase (EC 1.9.6.1) and nitrite reductase (EC 1.6.6.4) has been developed. MVH-nitrate reductase was induced and purified from E. coli K12. MVH-nitrite reductase was isolated and purified from spinach leaves. Nitrate is reduced by MVH-nitrate reductase to nitrite which is subsequently reduced to ammonia by MVH-nitrite reductase. The ammonia produced is monitored, using an air-gap electrode. Nitrate and/or nitrite are determined in the range of 5 X 10"-1 X lo-* M using either a soluble or an immobilized enzyme system.
subsequently measured colorimetrically. The disadvantage of this approach is that it is subject to the same interferences as the colorimetric method for nitrite and that the whole system is rather complicated. We would like to introduce an entirely different approach here: a new electrochemical enzymatic method, which has the advantages of specificity, rapidity, and simplicity for the determination of nitrate and/or nitrite. In this electrochemical method, nitrate and nitrite can be specifically and quantitatively reduced to ammonia by nitrate and nitrite reductases in the presence of MVH as electron donor, obtained through the chemical reduction of its oxidized form by dithionite added initially into the reaction mixture (Equations 1, 2 , and 3):
2H20 T h e toxicity of nitrite is well known and has two entirely separate aspects: (a) the induction of methemoglobinemia ( I ) by nitrite oxidation of hemoglobin from the ferrous to the ferric form; the ferric form of hemoglobin (methemoglobin) does not reversibly bind oxygen, and death results from anoxias in severe cases (21, and (b) the formation of carcinogenic N-nitroso compounds by the reaction between nitrite, and secondary or tertiary amines and amides (3-5). I t is generally accepted t h a t nitrate itself is not toxic. However the reduction of nitrate to nitrite occurs in the presence of microorganisms. As reported by Tannenbaum et al. (6),the major part of the nitrite found in the human body actually originates from the reduction of nitrate in the saliva by bacteria within the mouth. Recently the increasing use of nitrate and nitrite in foods, fertilizers, detergents, and other fields of industry has caused serious contamination problems in the environment. These problems have attracted increasing attention, especially since nitrite is suspected of being responsible for some incidence of cancer. Action has already been undertaken by some agencies to prevent such contamination. The US. Public Health Service (7) has announced that the allowable limits for nitrate and nitrite in potable water are 10 and 0.06 ppm, respectively. Recently, the U S . Department of Agriculture has proposed to discontinue the use of nitrate in all meat and poultry products and to reduce the allowed level of nitrite added for curing of meat and poultry from 200 to 156 ppm (8). Numerous nonenzymatic methods for the analysis of nitrate and nitrite have been developed and reported elsewhere. However, most of these are nonspecific and are subject to serious interferences from diverse substances. Pretreatment is usually required to eliminate these interferences before measurement. This, of course, is inconvenient and makes the method time-consuming. Since enzymic techniques have been widely used in analytical chemistry because of their high selectivity and sensitivity, several enzymatic methods for the determination of nitrate and nitrite have been reported ($12). The common approach of these methods is to reduce nitrate to nitrite using a nitrate reductase; the nitrite found is Present address, Department of Chemical Engineering and Nuclear Engineering, Iowa State University, Ames, Iowa 50011. 0003-2700/78/0350-1319$01 O O / O
+ 2MV2++ S204'-
2H+ + Nos-
+ 2MV+.
-+
2HS03'.
---+
+ 6MV'.
(1)
nitrate reductase
NOi
8H+ + NO2-
+ 2MV+*+ 2H+ 2MV2++ HzO (2)
nitrite reductase
+
NH4+ + 6MV2++ 2H20 (3)
The ammonia generated during the reaction can be selectively measured using a new air-gap electrode (13,14). This recently developed air-gap electrode is based on the same principle as other gas sensors, except that it does not have any gaspermeable membrane. The membrane is replaced by an air gap which separates the electrolyte layer from the sample solution, the entire system being contained in a gas-tight measuring chamber. The electrolyte itself is adsorbed as a very thin film on the surface of the indicator (pH) electrode. During measurement, the ammonia gas evolved from the sample solution under basic conditions, passes through the air gap and reaches the electrolyte layer on the surface of the electrode, thus increasing the p H value of the electrolyte layer. The amplitude of the p H a t equilibrium is directly proportional to the ammonia concentration in the sample solution (Nernstian relationship). The electrolyte layer can be easily renewed after each measurement, by touching the electrode surface on a cone-shaped polyurethane sponge, well soaked with the electrolyte solution, in an electrode holder. The greatest advantage of this design is that the electrode is never in direct physical contact with the sample solution which therefore may contain components that normally affect the function of any electrode. The simplicity of design, the easy renewal of electrolyte, and the fast response and recovery are further advantages of the air gap. We have reported (15) previously the determination of nitrite using MVH-nitrite reductase and air-gap electrode. At this time we describe an extension of this methodology to the determination of nitrate and/or nitrite, using both nitrate and nitrite reductases in a soluble and immobilized form.
EXPERIMENTAL Apparatus. The Radiometer type 503810 glass electrode was used, and the construction of the air-gap sensor is essentially the same as described by Ruzicka and Hansen (13,14). A Corning C 1978 American Chemical Society
1320
ANALYTICAL CHEMISTRY, VOL. 50, NO. 9, AUGUST 1978
digital 110 pH meter and a Heath Strip chart recorder with dc offset module and potentiometric amplifier were used to record the electrode response. Reagents and Materials. The electrolyte solution used was 5X M ammonium chloride solution saturated with wetting agent (Victawet 12, Stauffer Chemical Co.). Methyl viologen, cyanogen bromide, and glutaraldehyde were obtained from Sigma Chemical Co. (St. Louis, Mo.). Alkylamine glass beads were purchased from Corning Glass Co. (Corning, N.Y.). E. coli K12 PA 601 was obtained from the American Type Culture Collection. Fresh leaves of Spinacea uierucca were purchased from a local market. Preparation of Enzyme. The isolation and purification of nitrite reductase from spinach leaves was carried out according to Ho and Tamara (16)up to the DEAE -cellulose chromatography step. The induction, isolation and purification of nitrate reductase from E . coli K12 was performed according to Forget ( 1 7 ) to the first ammonia fractionation step. The enzyme was solubilized by heat rather than acetone treatment. The main fractions of enzymes were combined, freeze-dried, and stored in a freezer for later use. As a result of the preparation, from 1.0 kg of spinach leaves, about 220 units of nitrite reductase, can be obtained with specific activity of 1.5 units/mg and from 45 g of wet weight E. coli cells about 4560 units of nitrate reductase can be obtained with a specific activity of 10.0 units/mg. Immobilization of Enzyme. The method is based on the procedure of Marshall (181, with some modifications. Activate 2 g of glass beads by adding 15 mL of 3% glutaraldehyde! then degas the product under a desiccator vacuum for 10 min at room temperature. After activation for 60 min at atmospheric pressure and room temperature, decant and wash the glass beads at least three times with distilled water. The activated glass beads were added to 15 mL of 0.1 M phosphate buffer containing 50 mg of nitrate reductase or 100 mg of nitrite reductase. The mixture was degassed for 10 min on an ice-bath, and the coupling reaction was carried out for 24 h at 4 "C with gentle shaking. The immobilized material was washed at least three times with 0.01 M phosphate buffer (pH 8.0) and stored in a refrigerator until use. Measurement of Nitrite or Nitrate with Soluble Enzymes. Fifty pL of Tris buffer, 0.45 M, pH 7.0, containing 3 units of nitrite reductase for nitrite assay, or 3 units of nitrite reductase plus 3 units nitrate reductase for nitrate assay, and 0.75 pmol of methyl viologen, was pipetted into a sample chamber, followed by 200 pL of a nitrate standard solution. The reaction was initiated by adding 50 pL of sodium bicarbonate solution, 0.26 M, containing 3.75 mg of sodium dithionite, which was prepared immediately before use. The sample chamber was put in a water bath at 30 "C, incubated for 5 min, and then placed on a magnetic stirrer, preset at the desired speed. One-hundred pL of 0.05 N NaOH-KC1 solution was added, and the chamber was immediately covered with the electrode body. The potential change was recorded until a steady state was reached. The pH of the equilibrium state (pH,) on the surface of the glass electrode was read directly from the digital pH meter. The stirrer is best preset at a moderate speed, so that all measurements could be accomplished under almost identical conditions. In order to extend the detection limit, a minimal sample dilution should be effected. Measurement of Nitrite o r Nitrate with Immobilized Enzymes. To 0.25 mL of 0.45 M, pH 7.0 Tris buffer, containing 3.75 pmol methyl viologen, add 1.0 mL of sample solution and 0.25 mL of 0.25 M sodium bicarbonate, containing 18.25 mg sodium dithionite; mix. Pass this solution through column A (0.5 X 14 cm), which contains immobilized nitrite reductase for nitrite assay, or column B, which contains a 1:l mixture of immobilized nitrate and nitrite reductases for nitrate assay, a t a flow rate of 0.3 mL/min controlled by a peristaltic pump, so that a 100% conversion can be assured. After 5- 6 min, collect 300 pL of the eluent, pipet into the sample chamber, and measure the amount of ammonia released as described above. Since the void volume and dead volume of the column are about 0.9 and 0.1 mL, respectively, this 5-6 min waiting time is necessary to obtain homogeneous sample composition. Measurement of Nitrate a n d Nitrite Mixtures with Immobilized Enzymes. The procedure employed is the same as described for the measurement of nitrite or nitrate. The same
J
!
' 1 I _ I-_ 8
9
1c
3
2
1-
i',
Effect of pH, upon the evolution of ammonia. NH,CI concentration, 0.01 M, 200 pL. Buffer concentration, 0.5 M, 200 pL. (0 - 0 )Tris buffer( p = O.l), (0- 0)NaC0,-NaOH buffer ( p = O.l), (A - A ) Na,HPO,-NaOH buffer ( p = 0.1) Figure 1.
.-
7
C
L
I
I-/'-
,/'
I
/
---L,p---. lo li
-L--13
T M E
15
~~
L ~~
; I
. . . . .
-L
~.--I
ii
MlhLTL
Figure 2. Comparison of the direct and the two-step methods. (For details, see text). Curve 1: blank, no enzyme added; curve 2: direct measurement of substrate method; curve 3: two-step measurement of substrate
amount of sample solution and reaction solution are passed through both column A and B and the eluents are measured. The value obtained from column A indicates the amount of nitrite present in the mixture and nitrate can be calculated by subtracting the value of column A from the value of column B.
RESULTS A N D D I S C U S S I O N Effect of pH, for Evolution of Ammonia. In general, t h e p H of the sample solution (pH,) determines whether a quantitative or partial conversion of ammonia ion occurs. Figure 1 shows the dependence of pH, on pH,. At pH, > 10.5, the curve levels off, indicating a total conversion of ammonium ion to ammonia. However, the optimum p H for MVH-nitrate and nitrite reductase is around 7.4--7.8. Within this range only partial conversion of ammonium ion to ammonia takes place. As shown in Figure 2 , curve 2 , obtained with a K N 0 2 concentration of 1 X 10-' M a t pH, 8.0 (direct method), the response was very slow, taking about 20 min t o reach 95% of equilibrium. Furthermore, the sensitivity is very low, and the pH, drifts badly a t equilibrium. As can be seen from Figure 1, any pH, below 10.5 must be very well fixed in order to obtain good reproducibility. Thus, accurate measurements are difficult to achieve and direct monitoring of the reaction with an air-gap electrode is not practical. Because of difficulties encountered, it was decided to perform the measurement in two separate steps: t h a t is, to let the catalytic
1321
ANALYTICAL CHEMISTRY, VOL. 50, NO. 9, AUGUST 1978
concentration, the better the sensitivity. Second. the higher the concentration, the smaller the drift. Therefore, a compromise electrolyte concentration, 5 x 10Y3 M, was used for all subsequent measurements. Response Time a n d Recovery. The response time varies with substrate concentration. In general, it takes 30 s to I min, 2-3 min, and 557 min, to reach steady state a t concentrations of and M, respectively. Although the thickness of the electrolyte layer on the surface of the electrode has essentially no effect on t h e final pH, value, it does affect the response time. This effect becomes especially M to 5 apparent a t low substrate concentrations, (5 x x lo4 M), where poor response times were obtained (>7 min), if the layer was too thick. T o make the layer thinner, a tissue was used to absorb the excess electrolyte. The return of the signal t o a steady, well-defined baseline (the recovery time) was almost instantaneous, provided that the electrolyte layer is renewed between measurements. C a l i b r a t i o n C u r v e of S t a n d a r d S o l u t i o n s . According to the procedures described above, standard curves for nitrite and nitrate were obtained and are shown in Figure 3. T h e linear range is between 5 x 1 x 10 * M NO3- and NO2 . with a slope of 1.10-1.12 p H units/decade. The fact that the curve levels off a t high concentrations is possibly due to (1) the oxidation of reduced methyl viologen by molecular oxygen before completion of the enzyme rea( tion. ( 2 ) The equation by Ruzicka and Hansen (14) does not hold a t high concentrations, or (3) concentration is plotted and not activity. The leveling off observed a t low concentration is due to the detection limit of the electrode. The calibration curves obtained using immobilized enzymes for nitrate, nitrite, and nitrate plus nitrite fall on one single straight line between 1 x 10 4-1x lo-* M substrate with a slope of 1.04, which is close to theoretical value of 1.00. I n t e r f e r e n c e S t u d y . Theoretically, using the steady-state method and the air-gap electrode, no interference is expected from any ions which do not inhibit the enz-ynic reaction. This is verified by the data in Table 11, which shows that most diverse ions which may cause severe interference in conventional methods do not interfere. However, copper(I1) and mercury(II), which immediately precipitate with dithionite ion, do interfere. No interference was observed from any other cations or anions. S t a b i l i t y of Immobilized N i t r i t e a n d N i t r a t e Reductases. The stability of the immobilized enzlme was studied, based on how many measurements could be made with a 100% conversion of 1 x M substrate to product, under standard assay conditions. It was shown that the immobilized nitrite reductase could be used with 1W?& conversion of nitrite to ammonia for about 100 meastirements. After t h a t time, the conversion efficiency decreased to 90% a t 120 measurements. T h e immobilized nitrate reductase gave 100% conversion of nitrate to nitrite u p to 150 measurements, with a gradual decrease resulting thereafter However, the lifetime of the enzyme for 1007'r conversion car be extended by slowing down the flow rate.
Table I. Effect of Electrolyte Concentration o n Sensitivity and Drift drift of pH, electrolyte solution, (NH,+),, M 1 x 10-3 1 x 10-3 1 x 10-3 5 x 10-3 5 x 10-3 5 x 10-3 1 x 10-2 1x 1 0 - 2 1x 1 0 - 2
sample concn, (NH4+)s,M
detection limit for NH,+, M 4.4 x 4.4 x 4.4 x 4.8 x 4.8 x 4.8 x 1.0 x 1.0x 1.0x
10-5 10-5 10-5 10-5 10-5 10-5 10-4 10-4
min
1 x 10-4 1 x 10-3 1x 1 0 - 2
1x 1x 1x 1x 1x
1x
10-4
A pH, / 5
0.08 0.04 0.007 0.02 0.007 0.002 0.01 0,005 0.001
10-4 10-3 10-2 10-4 10-3 10-2
Table 11. Interference Studies of Diverse Substances in the Electrochemical Method diverse ions
of diverse
Ca2+ Zn2+ Mn2+ CUZ' COZ' NiZ+ Mg2+
5x 5x 5x 5x 5x 5x 5x 5x
NO,-
ions, M
Hg2
+
10-3 10-3 10-3 10-3
-3.2 - 2.6 t 1.9 -80.6 -2.7 1.7 -3.4 -100.0 -0.9 -3.1
10-3
+
10-3 10-3 103
5 x 10-3 5 x 10-3 5 x 10-3 5 x 10-3 5 x 10-3 5 x 10-3 5 x 10-3
Na'
K'
NO2-
NO3-
c104-
so,>so3,c1c10,-
5
x
. . .a
NO3-
+ 1.9 -3.5 -4.2 - 75.4 - 1.6 - 2.6 t 1.9 -100.0 -2.8 1.3 -2.7
+
. . .a
i 2.7
-4.5 +0.6
- 1.9
-4.4 -2.1 t 0.7 - 2.9 t3.2 -1.2
- 1.7
+- 4.0 1.8
10-3
5 x 10-3 5 x 10-3 5 Y 10-3
BrO; IO,'
-
% error observed
- 1.2 t 3.8
Final nitrate and nitrite concentrations. 5
M.
X
reaction incubate a t a suitable p H for a definite time, and then to carry out the actual ammonia measurement by the addition of an excess amount of strong base t o secure quantitative conversion. Therefore, p H 7.0 was chosen to minimize loss of ammonia during incubation, and the enzyme in either soluble or immobilized form still shows reasonably high activity, compared with curve 2. Much better results were obtained by t h e two-step method (curve 3) a t t h e same substrate concentration (compare to curve 2 ) . Curve 1 is a blank, with no enzyme added (19). E f f e c t of C o n c e n t r a t i o n of Electrolyte o n the S u r f a c e of the Electrode. The concentration of electrolyte was studied t o determine which was optimum to achieve t h e greatest electrode response with the least drift. From Table I, two conclusions were reached. First, the lower the electrolyte
_-
___-
Table 111. Determination of Nitrate and/or Nitrite at a 1:l Ratio (NO;
+
NO;)
added 1X 5X 1X
lo-*
5 x 1 X 10.'
--
a (NO;)
= (NO;
M
CV,
found 1.06 X 5.22 X 0.96 x
5.12~ 1.03
t
X
NO,-)
-
relative error,
re1at ive
SDb
%
%
(NO,-)M, found
0.081 0.30 0.040 0.17 0.022
7.6 5.7 4.2 3.3 2.1
6.0 4.4 4.0 2.4 3.0
4.55 x 2.38 x 4.88 x 2.51 x 10-3 4.93 x
(NO;).
Av 3.96 Average of three measurements.
CV,
M
error,
SD
%
%
(calcd
0.38 0.15 0.26
8.3 6.3 5.3 4.0 3.6
9.0 4.8 2.4 0.4 1.4
6.05 X 10.' 2.84 Y l o - , 4.72 x 1 0 2.6:l x 10 5.37 Y 10
0.10 0.18
Av 3.6
relative error, %
21.0 13.6 5.6 5.2 7.4
Av 10.56
1322
ANALYTICAL CHEMISTRY, VOL. 50, NO. 9, AUGUST 1978
(which have higher COz desorption rates). The COz desorption rate depends on temperature, the p H value, the volume of the electrolyte layer and of the sample solution, the stirring speed, and the air-gap space. All these factors can be cont,rolled easily, with the exception of the thin electrolyte layer which is difficult to reproduce each time. Since the volume of the electrolyte layer is so small, a slight change will cause a large effect on the COz desorption rate (19). Therefore, it is difficult to predict whether the error will be high or low. The assay value for nitrate, or nitrate plus nitrite, has more error t,han that for nitrite alone, because errors in both measurements are introduced. However, the method is reasonably good for the determination of NO2- and NO3- at concentrations greater than 5 X M. It appears likely that this procedure is not very promising for the accurate assay of nitrate and nitrite mixture in potable water since 5 X M is well above the allowable limits as previously described.
LITERATURE CITED
Figure 3. Calibration curve of nitrite and nitrate obtained from the soluble enzyme system. All measurement conditions (see text). (0 - 0 )nitrate, (0 - 0) nitrite
T h e immobilized enzymes are also very stable with temperature. Leaving the immobilized enzymes a t room temperature for three weeks, 100% conversion was observed with both enzymes. Determination of Nitrate and/or Nitrite in a Mixture. With the use of the dual enzyme system, samples of nitrite and nitrate in a mixture were measured (Table 111). Poor reproducibility and high errors were observed a t low substrate concentrations. This is mostly due to the sensitivity of the electrode to the COS desorption from the electrolyte layer (coated on the surface of the electrode) onto the sample solution which was adjusted to high p H value (1 1.O). This COP desorption causes an increase in p H value of the electrolyte layer (19),thus resulting in a drift of the response curve. This effect becomes obvious a t low substrate concentration since a t these concentrations the time to reach equilibrium is slow and the p H value at equilibrium, pH,, is usually lower than those obtained from higher substrate Concentrations
(1) J. Haldane, R. H. Makgill, and A. E. Mavrogordato, J . fhysioi.. 21, 160 ( 1897). (2) 0. Bodansky, Pharm. Rev., 3 , 144 (1951). (3) P. N. Magee and J. M. Barnes, Br. J . Cancer, 10, 114 (1956). (4) S. S. Mirvish, J . Natl. Cancer Inst., 44, 633 (1970). (5) A. L. Fridmen, F. M. Mukhametshin, and S. S. Novikov, Russ. Chem. Rev., 40, 34 (1971). (6) S. R. Tannenbaum, A. J. Sinskey, M. Weisman, and W. J. Bishop, Natl. Cancer Inst., 53, 79 (1974). ( 7 ) Resources Agency of California "Water Quality Criteria", State Water Quality Control Board Publication No. 3-A, J. E. Mckee and H. W. Wolf, Ed., Sacramento, Calif., 1963, p 224. (8) Fed. Regisf.. 40, No. 218, Tuesday, Nov. 11 (1975). (9) G. B. Garner, J. S. Baumstark. M. E. Muhrer, and W. H. Pfander, Anal. Chem., 28, 1589 (1956). (10) A. L. McNamara, G. R. Meeker, P. D. Shaw, and R. H. Hageman, J. Agric. Food Chem., 19, 229 (1971). (11) R. H. Lowe and M. C. Gillespie, J . Agric. Food Chem., 23, 783 (1975). (12) D. R. Senn, P. W. Carr, and L. N. Klatt, Anal. Chem., 48, 954 (1976). (13) J. Ruzicka and E. H. Hansen, Anal. Chim. Acta, 69, 129 (1974). (14) E. H. Hansen and J. Ruzicka, Anal. Chim. Acta, 72, 353 (1974). (15) C. H. Kiang, S. S. Kuan and G. G. Guilbault, Anal. Chim. Acta, 80,209 (1975). (16) C. H. Ho and G. Tamara, Agric. Bioi. Chem., 37, 37 (1973). (17) P. Forget. Eur. J . Biochem., 42, 325 (1974). (18) D. L. Marshall, Biotech. Biol., 15, 447 (1973). (19) E. Hsiung, S. S. Kuan, and G. G. Guilbault, Anal. Chim. Acta. 84. 15 (1976).
RECEIVED for review January 10,1978. Accepted May 17, 1978. The authors thank the NSF-RANN Food Technology Program (Grant No. AER-76-23271) for financial assistance in carrying out this project.
