ANALYTICAL

BIOCHEMISTRY

186,170-175

(1990)

Electrochemical Cells for Voltammetry, Coulometry, Protein Activity Assays of Small-Volume Biological Samples

and

Benjamin J. Feldman,* Stephen F. Gheller,? Glen F. Bailey,* William E. Newton,*lt and Franklin A. Schultz$’ * USDAjARS Western Region Research Center, Albany, California 94710; tDepartment of Agronomy and Range Science, University of California, Davis, California 95616; and $Department of Chemistry, Purdue University School of Science, Indiana University Purdue University at Indianapolis, Indianapolis, Indiana 46205

Received

September

14,1989

Cell designs, experimental protocols, and results for electrochemical investigation of small quantities of biological materials under anaerobic conditions are reported. Three types of electrochemical experiments are considered: (i) cyclic voltammetry of ZO- to lOO-pl samples; (ii) direct coulometry of 0.5- to 1.5ml samples; and (iii) an electrochemically initiated protein activity assay which includes provision for analysis of gaseous reaction products and correlation with electron flux. The first two procedures are illustrated by measurement of the formal electrode potential (E”) and number of electrons transferred (n) in redox reactions of small quantities of biological and inorganic materials. The third procedure is illustrated by assaying the activity of the MoFe protein plus Fe protein complex from Azotobacter vinelandii nitrogenase for reduction of &Hz to C2I-h.

OlBDOAcademic

Press,Ine.

Electron storage and/or transport is central to the function of many proteins, including those involved in Nz fixation, photosynthesis, and respiration. Many properties of electron transfer and storage proteins, their extruded active sites, and model compounds designed to mimic these active sites can be determined electrochemically. These properties include the redox potential (E”) of stored electrons, which may be determined by classical potentiometric, spectrophotometric, and coulometric titrations (1) or more modern techniques such as cyclic voltammetry (2,3) and the number of electrons (n) stored in the molecule, which may be ’ To whom 170

correspondence

should

be addressed.

determined by direct or chemically mediated (1) coulometry. Electrochemical analysis of biological materials often is complicated by limited sample availability and sensitivity to dioxygen and other environmental conditions. This paper reports designs, experimental protocols, and results for three electrochemical cells constructed for use with small-volume biological samples under anaerobic glove box conditions. The cells include (i) a voltammetry cell for cyclic voltammetric analysis of 20- to IOO,ul samples; (ii) a coulometry cell for direct coulometric analysis of 0.5- to 1.5-ml samples; and (iii) a gas-tight protein activity assay cell, which combines the function and sample size (0.5-2 ml) of the coulometry cell with the capability for withdrawing gaseous reaction products for subsequent chromatographic analysis. The third cell is used to conduct an “electrochemical assay” of an electrode-driven protein reaction in which the gaseous products of the reaction are used to quantify protein activity. The cells were designed for electrochemical studies of the MoFe protein of Azotobacter uinehdii nitrogenase and of its extruded active site, the iron-molybdenum cofactor (FeMoco)2 (4-6). These biological materials are available only in limited quantity and are extremely sensitive to degradation by atmospheric dioxygen. We have found operation under an inert atmosphere in a glove box to be the only means of obtaining reliable results on these materials. Many biological and nonbiological samples have similar experimental restrictions. The cells reported in this paper were designed for ease of ma* Abbreviations normal hydrogen aminoethanesulfopic tetra-n-butylammonium

used: FeMoco, iron-molybdenum cofactor; NHE, electrode; Tes, IV-tris[hydroxymethyl]m&hyl-2acid; DMF, N,N-dimethylformamide; TBAPF6, hexafluorophosphate; MV, methyl viologen. Copyright

0003-2697/90 $3.00 0 1990 by Academic Press, Inc. in any form reserved.

All rights of reproduction

ELECTROCHEMICAL

ASSAY

FIG. 1. Voltammetry

cell. (A) Glassy carbon disk. (B) Pt counter electrode. (C) Metal connection to glassy carbon. (D) Kel-F electrode shroud. (E) Kel-F cell body. An Ag/AgCl reference electrode is dipped into the cell from above.

nipulation within a glove box without sacrificing accuracy of results. Use of the voltammetry cell is demonstrated by cyclic voltammetry of a 22-~1 sample of FeMoco, from which the formal potential of the first reduction process (the oxidized-to-semireduced electron transfer reaction (4)) for this material is determined. The function of the coulometry cell is demonstrated by carrying out coulometric analysis of the number of electrons transferred upon reduction of (Et4N)3MozFe6S8(OMe)3(p-C1PhSe)6, an inorganic compound synthesized to model FeMoco. The protein activity cell is used to assay the &Hz-reducing ability of A. vinelundii nitrogenase. In this procedure, methyl viologen is used to mediate electron transfer between a Pt electrode and the complete functional nitrogenase system under a 10% C2H2/90% Ar atmosphere. Hz and C,H, produced by the reaction are quantified by chromatographic analysis of the headspace above the solution, and the moles of gaseous products are correlated with the quantity of charge passed in the electrochemical reaction. MATERIALS

AND

OF

BIOLOGICAL

171

SAMPLES

sample compartment serves as the auxiliary electrode. A volume of analyte solution (typically 20 to 50 ~1) just sufficient to cover the working and auxiliary electrodes is added to the cell. A microsized Ag/AgCl (1 M KCl) reference electrode (Microelectrodes, Inc., Londonderry, NH) is clamped above the cell and lowered so that its liquid junction tip just makes contact with the analyte solution when a voltammogram is run. Coulometry cell. The coulometry cell (Fig, 2) is constructed by drilling a 14-mm diameter by lo-mm deep cavity in a l$ -in. piece of Kel-F rod. This chamber holds a ring or crescent-shaped working electrode constructed from reticulated vitreous carbon (ERG Industries, Oakland, CA) or Pt, Au, or Ag mesh. The working electrode is shaped so that the remainder of the chamber is able to accommodate reference and auxiliary electrodes, which are placed in sections of 4-mm glass tubing and separated from the analyte solution by pieces of Vycor frit (Princeton Applied Research, Princeton, NJ) attached with heat-shrinkable tubing. The reference electrode compartment contains an AgCl-coated silver wire and is filled with 3 M KCl. The auxiliary electrode compartment contains a Pt wire auxiliary electrode and is filled with a solution matching the composition of the background electrolyte in the sample compartment. The working and reference electrode compartments are attached to a sliding Kel-F disk and are lowered into contact with the sample solution when the experiment is begun. A smaller cavity is drilled below the sample compartment to hold precisely a circular, finned Teflon stirring bar (Fisher Scientific). This allows the solution to

A

6

\-

/

IElE -

-...

FTt

-

7

-

-F 5cm

METHODS 9

Voltammetry cell.

Construction of the voltammetry cell is illustrated in Fig. 1. The cell consists of an inverted glassy carbon disk electrode (Bioanalytical Systems, West Lafayette, IN) of 0.071-cm2 surface area whose exterior Kel-F shroud is threaded to accept a cell body machined from a i-in. Kel-F rod (AIN Plastics, Mount Vernon, NY). Kel-F is selected for these and many of the other cell components because of its low affinity for O2 and its resistance to nonaqueous solvents. A loop of 28-gauge Pt wire placed near the bottom of the

C’

FIG. 2.

Coulometry cell. (A and B) Vycor glass-f&ted glass tubes for Ag/AgCl reference and Pt counter electrodes. (C) Screw connection to Pt or vitreous C working electrode (not shown). (D) Circular stirring bar compartment. (E) Working electrode compartment. (F) Al support bars. (G) Kel-F sliding disk mount for reference and counter electrode compartments.

172

FELDMAN

e-1 G-n

n-H

5cm

FIG. 3. Protein activity assay cell. (A) Circular stirring bar compartment. (B) Working electrode compartment. (C) Septum-fitted port for Pt connection to working electrode (not shown). (D) Septumfitted port for solution addition/removal. (E) Buna-N O-ring. (F) Threaded holes for septum-fitted valves (not shown) for gas addition/ removal. (G and H) Vycor-fritted glass tubes (O-ring sealed) for Ag/ AgCl reference and Pt counter electrodes. (I) Al rings for clamping ceil halves. (J) Clamping screws.

be stirred so that efficient electrolysis times of ca. 10 min can be achieved in coulometry experiments. A disk of plastic mesh placed over the spin bar reduces turbulence during stirring. Protein activity assay cell. The design of the gas-tight protein activity assay cell (Fig. 3) is similar to that of the coulometry cell. The two cells use the same working, reference, and counter electrodes and the same provision for stirring the sample solution during electrolysis. In the gas-tight cell, electrical contact to the working electrode is made by a length of 32-gauge Pt wire inserted through a butyl-rubber septum and a hole drilled in a Kel-F screw (C). The reference and counter electrode compartments (G and H) are inserted through O-ringlined threaded holes and sealed in place with hollow 1-in.28 male threaded Kel F screws (No. 20089, Alltech Associates, Deerfield, IL). The gas removal ports (F) are fitted with i-in-28 male threaded Mininert valves (No. 10125, Pierce Chemical Co., Rockford, IL, not shown). The two cell halves are clamped together by aluminum rings (I). The total internal volume (with electrodes and stirring bar in place), measured by filling the cell with water, is 5.4 ml. All electrochemical exEquipment and procedures. periments were performed with a Bioanalytical Systems CV-27 potentiostat, which was interfaced to a Hyundai

ET

AL.

microcomputer through a Metrabyte DAS-16 A/D converter board. Cyclic voltammetry and coulometry experiments were computer controlled by software written in Quick Basic (Microsoft). All electrochemistry was performed in a Vacuum Atmospheres glove box, using shielded cables which passed through a sealed port in the box. Electrochemical potentials measured relative to Ag/AgCl reference electrodes are expressed as values versus the normal hydrogen electrode (NHE) (7). Prior to performing nitrogenase activity assays, the gas-tight cell was flushed with Ar, and 10% &Hz was added by syringe. CzH4 and Hz analyses were performed on a Hewlett-Packard 5890A gas chromatograph. C2H4 was determined on a Porapak N (Supelco, Bellefonte, PA) column at 70°C with flame ionization detection using He carrier gas at a flow rate of 20 ml/min. Hz was determined on a molecular sieve 5A (Supelco) column at 40°C using thermal conductivity detection and Ar carrier gas at a flow rate of 15 ml/min. Both analyses were corrected for cell atmosphere dilution, which resulted from repetitive withdrawals of 0.125-ml gas samples from the dead volume. CzH4 analyses also were corrected for CzH4 solubility in the analyte solution using a partition coefficient ( [C2H& mM Tes/[C2H4]Ar) = 0.459) determined chromatographically. For typical solution and cell dead volumes of 1.3 and 4.1 ml, respectively, the correction is about 13%. H2 analyses were not corrected for partitioning due to the low solubility of Hz in water. Materials. (Et4N)3MozFe&,(OMe)&-ClPhSe), was synthesized (8) and FeMoco extracted into N-methylformamide from the MoFe protein of A. uinelandii nitrogenase (9) according to previously published procedures. The final electrochemical solution concentration of (Et4N)3MozFe&(OMe)3(p-ClPhSe)G was determined by total Fe analyses based on spectrophotometric determination of tris-(o-phenanthroline)-iron(I1) at 510 nm, N,N-Dimethylformamide (DMF) was vacuum distilled from molecular sieves. Tetra-n-butylammonium hexafluorophosphate (TBAPFB , Southwestern Analytical Chemicals, Austin, TX), methyl viologen (Aldrich Chemical Co., Milwaukee, WI), and o-phenanthroline (Aldrich) were used as received. The MoFe and Fe proteins of A. vinelandii nitrogenase were purified according to published procedures (10) and had specific activities of 1410 and 1293 nmol C&H2 reduced per minute per milligram protein, respectively. The analyte solution for the protein activity assays contained 25 mM Tes, pH 7.4,2.5 mM ATP, 5.0 mM MgClz, 30 mM creatine phosphate, and 2.5 units/ml creatine phosphokinase. Tes, ATP, creatine phosphate, and creatine phosphokinase were used as received from Sigma Chemical Co. (St. Louis, MO). RESULTS

AND

DISCUSSION

Figure 4 illustrates the determination redox potential by cyclic voltammetry

of the formal using a 22ql

ELECTROCHEMICAL

ASSAY

5I I-o-

-5

Em -10

1

’

-0.1

-0.3

.

.

-0.9

E, V vs NHE FIG. 4. Cyclic voltammetry of a 22-pl sample of 2.82 mM oxidized FeMoco in N-methylformamide. Potential sweep rate = 40 mV/s. Epe and Epe are the cathodic and anodic peak potentials, respectively.

FeMoco sample in the small volume voltammetry cell. In this technique (ll), the working electrode potential is swept linearly with time first to negative and then to positive values through a voltage range in which the sample is expected to undergo an electron-transfer reaction. The current response is measured and, if the product of the electrode reaction is stable, cathodic and anodic peaks are observed on the forward and reverse scans, respectively, corresponding to reduction and reoxidation of the sample. The formal redox potential is calculated as the average of the cathodic (E,,) and anodic (E,,) peak potentials (12):

EO’= j(E, + E,,).

PI

Figure 4 displays the cyclic voltammogram of a 2.82 solution of oxidized FeMoco in N-methylformamide that, except for anaerobic storage at -80°C for 12 months, was otherwise untreated following isolation from the MoFe protein of A. vinelundii nitrogenase. FeMoco exists in three oxidation states. mM

FeMoco(ox)

+e-

z FeMoco(s-r) *

tne-

z

FeMoco(red)

BIOLOGICAL

173

SAMPLES

(ox) to semireduced (s-r) form of FeMoco (Eq. [2a]). This result is in good agreement with the value E”’ = -0.32 f 0.02 V vs NHE we have obtained by other more tedious means (4), including chemical titration of FeMoco(ox). A carbon working electrode surface is essential for obtaining good electrochemical results, because FeMoco fails to react at Pt, Ag, Au, and Hg electrode surfaces. Figure 5 illustrates use of the coulometry cell for determining n, the number of electrons transferred in an electrode reaction. In this experiment a l-ml sample of 1.38 mM (Et4N)3M02Fe$s(OMe)3(p-C1PhSe)G in 0.1 M TBAPF,/DMF is electrolyzed at a potential at which two electrons are transferred to the polynuclear MO-Fe cluster. This compound, which is a proposed model for FeMoco, undergoes reversible one-electron reductions at E”’ = -0.57 and -0.77 V vs NHE.

lo-

5 E s

OF

[2]

The oxidized (ox) and semireduced (s-r) states are known to be interconverted by transfer of one electron (5). The fully reduced (red) state is believed to correspond to the substrate-reducing state of the enzyme. The experimental trace shows a cathodic peak at -0.39 V and an anodic peak at -0.29 V, from which a value of -0.34 V vs NHE is calculated for the E”’ of the redox couple corresponding to conversion of the oxidized

[3]

[Mo,Fe&Ss(OMe)3(p-C1PhSe)~]3-

+ e- G=[ I”-

[Mo,Fe&&(0Me)3(p-C1PhSe)~]4-

+ e- + [ 15- [4]

In the coulometry experiment, the working electrode potential is held at -1.10 V and a record of charge (Q) versus time (t) is recorded as the cell contents are stirred vigorously. The plot of Q vs t in Fig. 5A approaches a linear asymptote within ca. 10 min, indicating that electrolysis is effectively complete within this time. Extrapolation of the linear section of the Q vs t plot to t = 0 gives a net charge of Qred= 263.9 mC for the reductive electrolysis. From this result, a value of n = 1.99 is calculated from Faraday’s law (Q = nFm, where F = 96,486 C/eq and m = moles of material) for the sum of reactions [3] and [4]. In Fig. 5B, the electrode potential is stepped to -0.2 V, and the reduced complex is reoxidized by the reverse of reactions [4] and [3]. In this experiment, values of QoX = 254.5 mC and n = 1.92 are obtained by extrapolation of the Q-t plot to zero time. The near equivalence of Qred and Q,, and their agreement with a calculated value of 265.3 mC indicates that both the 3- and the 5-forms of the MO-Fe cluster are stable in DMF and can be reversibly interconverted in this solvent. The coulometry cell in Fig. 2 is applicable to routine determination of the number of electrons transferred upon reduction or oxidation of micromole quantities of inorganic and biological materials. We have used this cell to confirm that the electron stoichiometry of the FeMoco(ox)/(s-r) electrode reaction in Eq. [2a] is one using submicromole quantities of FeMoco (5). Also, the cell may be used to prepare solutions of reduced materials for subsequent spectroscopic analysis without addition of chemical reagents. Finally, the reticulated vitreous carbon electrode may be used as a cyclic voltammetric working electrode in unstirred solution. This enables cyclic voltammograms to be recorded before and

174

FELDMAN

ET

AL.

600-

-600 0

600

1200

1800

I

I

0

I

I 600

I

I

I

1200

TIME 61

TIME

I

I 1800

W

Charge-time curves for hulk electrolysis of a l-ml sample of 1.38 mM (Et,N)SMozFe&s(OMe)3(p-C1PhSe), in DMF containing 0.1 (A) Reduction: initial potential = -0.2 V, electrolysis potential = -1.1 V. (B) Oxidation: initial potential = -1.1 V, electrolysis potential = -0.2 V. Dotted lines are background corrections accomplished by extrapolation of Q-t response to zero time. Experiments performed in coulometry cell with vitreous carbon working electrode. FIG.

6.

M

TBAPF,.

after a coulometry experiment to verify sample integrity at the completion of electrolysis. The gas-tight cell in Fig. 3 is used to carry out an electrochemically driven assay of nitrogenase activity. In this procedure, the sample compartment is filled with an aqueous solution containing the Fe and MoFe proteins of nitrogenase, an ATP source, and methyl viologen (MV2’) as an electron carrier. The cell atmosphere contains 10% CzHz, a nitrogenase substrate. Setting the potential of the Pt working electrode to -0.55 V vs NHE reduces MV2+ to MV+ and initiates an electron transfer chain that results in production of C2H4 from C2H2 and of H2 from protons.

The procedure is novel in the sense that it permits the quantity of gaseous products to be compared with total electron flux, which is an important relationship in understanding the mechanism of nitrogenase reactions. Watt and co-workers (13) reported an electrochemical assay of the purified nitrogenase complex from A. uinela&ii, but made no provision for quantitation of gaseous products. Figure 6 compares the quantities of gases produced with the electrochemical charge passed during electrolysis at -0.55 V vs NHE. There is an initial burst of H2 production, which subsequently ceases. Coincident with this Hz formation, there is a lag in C2H4 production which later increases to a constant rate. An important feature of Fig. 6 is that the amount of product (H, plus C2Hq) formed in the cell very nearly (90%) equals the amount of electrical charge passed. We attribute the ad-

ditional 10% charge to charging of the Pt electrode, establishment of a small steady-state concentration of reduced methyl viologen, and background electrochemical processes. The steady-state rate of C&H4 production in Fig. 6,386 nmol/min/mg MoFe protein, is less than the value obtained from the conventional chemical assay which uses Na2S20d as reductant (1410 nmol C2Hdmin/mg MoFe protein for our sample). The lower value may be the result of mass-transport limited electron transfer to the methyl viologen mediator caused by the physical (i.e.,

e 1P 5s P0) B f E

600600400200O0

200

400

600 TIME

FIG. 6.

800

1000

6)

Electrochemical nitrogenase acetylene reduction assay performed in the gas-tight coulometry cell using a Pt mesh working electrode. Assay conditions: 0.128 mg MoFe protein, 0.246 mg Fe protein, 220 ~1 reaction mixture (see Materials and Methods), 194 nmol methyl viologen in 25 mM Tes buffer, pH 8, solution volume = 1.23 ml. Assay potential = -0.55 V. Assay products expressed as nmol of electron pairs calculated from (0) electrochemical charge, (A) CzHl pmduction, (+) Hz production, and (w) sum of CzH, and Hz production.

ELECTROCHEMICAL

ASSAY

heterogeneous) arrangement of the electrochemical experiment. However, the rate of CzH4 production is a complicated function of the proportions of methyl viologen, MoFe protein, and Fe protein in the cell, and its relationship to these quantities appears to differ significantly from what is observed in the chemical assay. Specifically, at l/l Fe/MoFe ratios, CzH4 production in the electrochemical assay is only about 1% of that observed at high (10/l or greater) Fe/MoFe ratios. This is in contrast to the chemical (Na2S204-driven) assay, in which C,H, production at a 0.9/l Fe/MoFe protein ratio continues at ca. 21% (295 nmol &Hz reduced/min/mg MoFe protein) of the limiting (high Fe/MoFe ratio) value. We are investigating these relationships further to provide new information on the mechanism of electron transfer within the nitrogenase enzyme system. We also are seeking to determine if H, production, which consumes a proportion of the electrons passed much greater than that in the chemical assay under 10% CZH2/Ar, is due to the direct reduction of solution protons by reduced methyl viologen or is the result of altered nitrogenase function under the conditions of the electrochemical assay. We expect that the ability to correlate gaseous products with total quantity of charge will help in answering these questions. ACKNOWLEDGMENTS We thank Mary C. Weiss (USDA/ARS) for performing

OF

BIOLOGICAL

175

SAMPLES

work was supported by Grants CHE-8718013 tional Science Foundation and DK-37255 tional Institutes of Health.

(to F.A.S.) (to W.E.N.)

from the Nafrom the Na-

REFERENCES 1. Fultz,

M. L., and Durst,

2. Armstrong,

F. A., Hill,

R. A. (1982)

Anal. Chim. Acta 140,1-18. N. J. (1988) Act.

H. A. O., and Walton,

Chem. Res. 21,407-413. 3. Armstrong, F. A., George, S. J., Thomas, (1988) FEBS Lett. 234,107-110.

A. J., and Yates,

M. G.

4. Schultz, F. A., Gheller, S. F., Burgess, B. K., Lough, S., and Newton, W. E. (1985) J. Amer. Chem. Sac. 10'7.53646368. 5. Schultz,

F. A., Gheller,

S. F., and Newton,

B&hem.

W. E. (1988)

Biophys. Res. Commun. 152,629-635. 6. Newton, W. E., Gheller, and Schultz, F. A. (1989) 7. Sawyer, chemistry

S. F., Feldman, J. Biol. Chem.

B. J., Dunham,

D. T., and Roberts, J. L. (1974) for Chemists, Wiley, New York.

8. Christou,

G.,

and

Garner,

C. D.

W. R.,

264,1924-1927.

(1986)

Experimental

Electro-

J. Chem. Sot. Dalton

Trans., 2354-2362. 9. Burgess,

B. K., Steifel,

E. I., and Newton,

W. E. (1986)

D. B., and Steifel,

E. I. (1986)

J. Biol.

&em. 255,353-356. 10. Burgess,

B. K., Jacobs,

Biochim.

Biophys. Acta 614,196209. 11. Bard, A. J., and Faulkner, L. R. (1986) Electrochemical Fundamentals and Applications, Chap. 6, Wiley, New 12. Kissinger,

P. T., and Heineman,

W. R. (1983)

Methods: York.

J. Chem. Educ. 60,

702-706. (UC-Davis) nitrogenase

and Dr. Deborah J. Scout protein activity assays. This

13. Lough,

S., Burns,

4062-4066.

A., and Watt,

G. D. (1983)

Biochemistry

22,