The use of metallocenes for protein modification

Biochimica etBiophysica Acta, 746 (1983) 97-100

97

Elsevier

BBA31646

THE USE OF METALLOCENES FOR PROTEIN MODIFICATION FORMATION OF TETRAMETHYLCYCLOBUTADIENE NICKEL ALCOHOL DEHYDROGENASE R O G E R W. GIESE a,* and W A L T E R K O R N I C K E R b

Northeastern University, 360 Huntington Avenue, Boston, MA 02115 (U.S.A.) and t, Lachenstrasse 18, 4104 Oberwil BL (Switzerland) (Received April 6th, 1983)

Key words: Enzyme inactivation; Metallocene; Protein modification," Alcohol dehydrogenase; Spectral change

Tetramethylcydobutadiene nickel dichloride is a dark red, half-sandwich nickel metailocene with both aqueous and organic solubility. It reacts extensively with horse liver alcohol dehydrogenase (EC 1.1.1.1), forming an orange, soluble adduct that is stable at least to moderate dialysis and essentially devoid of native zinc and also enzymatic activity. This modified enzyme contains nearly 8 tool nickel per tool and possesses a positive Cotton effect. Under the same conditions, nickel chloride has no effect on the enzyme. This illustrates the potential for half-sandwich metailocenes to provide unique reactivity and spectral characteristics for protein modification.

Introduction Metallocenes and metallocene-like compounds are a new class of reagents for the study of protein structure and function [1]. These metal-,n'-complexes allow the specific, covalent introduction of metal atoms and their spectral properties into proteins, as illustrated by the reaction of maleic anhydride irontetracarbonyl with ribonuclease A [2,3]. They can also be used for the site-specific, noncovalent association of metal atoms with proteins, as exemplified by the ability of hydroxmethyl ferrocene to serve as a substrate for horse liver alcohol dehydrogenase [4], and that of various phenylmethylcarbonate derivatives of ferrocene to inhibit acetylcholinesterase [5]. In this paper we report the first example of a half-sandwich metallocene reacting with a protein. Tetramethylcyclobutadiene nickel dichloride, unlike nickel dichloride, inactivates horse liver al-

* To whom correspondence should be addressed. 0167-4838/83/$03.00 © 1983 Elsevier Science Publishers B.V.

cohol dehydrogenase, affording a metallocene-protein derivative in which nickel is present and zinc is absent. This derivative also displays a positive Cotton effect. Materials and Methods Horse liver alcohol dehydrogenase was obtained from Worthington Biochemical Corporation. Tetramethylcyclobutadiene nickel dichloride (dark red crystals) was a generous gift from Professor Rudolf Criegee. Its synthesis has been reported [6]. Analytical procedures. The usual precautions were taken to avoid metal contamination in the work with carboxypeptidase A and alkaline phosphatase [7]. Zinc was analyzed by atomic absorption [8]. A Zeiss M4-QIII speetrophotometer was used for absorption measurements. Circular dichroic measurements were performed using a Cary 61 circular dichroic spectrophotometer. Nickel determination. Nickel was determined by atomic absorption using radiation at 2320 A, a

98

15-inch alundum tube, and a ring burner. The flame was a mixture of air (17 l b / i n c h 2, 3.5 l / m i n ) and hydrogen (3.3 l b / i n c h 2, 1.5 l/min). The sample flow rate was 2.8 m l / m i n . Enzyme assay. Alcohol dehydrogenase was assayed with ethanol as a substrate in the presence of N A D as described [9].

Reaction of tetramethylcyclobutadiene nickel dichloride with alcohol dehydrogenase. A solution (2.0 ml) of horse liver alcohol dehydrogenase ( 4 . 8 . 1 0 - 5 M) in Tris-acetate buffer, p H 7.0, was placed in a dialysis bag and dialyzed for 7 h at 4°C against 50 ml of the same buffer while this latter buffer was bubbled with nitrogen to achieve deoxygenation. After removal from the dialysis bag, 1.0 ml of the alcohol dehydrogenase solution was treated with 0.050 ml of a fresh solution of 0.10 M tetramethylcyclobutadiene nickel dichloride (105-fold molar excess relative to the enzyme) in water. The enzymatic activity of the resulting orange solution was measured as a function of time relative to a blank treated similarly except for the presence of the nickel metallocene. Also, a control sample was prepared, involving treatment of the enzyme with a 105-fold molar excess of nickel dichloride. After 1.5 h, the reaction, blank and control sample of alcohol dehydrogenase were dialyzed separately against 2 x 100 ml of nitrogen-treated, 0.1 M Tris-acetate buffer at 4°C for a total of 10 h, and then the enzymatic activities, zinc and nickel contents, and absorption and circular dichroic spectral characteristics of these enzyme solutions were determined. Separate aliquots of these samples were treated similarly except for dialysis against 3 x 100 ml over a 16-h period, prior to the same measurements. Additional control samples each involved a treatment of the enzyme with a 100-fold molar excess of nickel dichloride in the presence of a 100-fold molar excess of either maleic acid, potassium tartrate, 1,3-diaminopropane, glycine, malonic acid, ethylenediamine or tetraglycine, adjusted to p H 8.4-8.6 in pyrophosphate buffer and incubated for 3 h at room temperature prior to measurement of the enzyme activity relative to blanks treated the same except for the presence of nickel dichloride.

Results

Solution characteristics of tetramethylcyclobutadiene nickel dichloride The solubility and spectral characteristics of tetramethylcyclobutadiene nickel dichloride are summarized in Table I. Good solubility is seen in water either at high or low ionic strength, and also in chloroform. As shown, the ultraviolet spectral features are highly sensitive to the solvent. Based on decreases in the spectral intensity of the most intense absorption band as a function of time, the metallocene is more stable when dissolved in an organic solvent, or in aqueous solvent at a low pH, as opposed to a solution of this compound in aqueous buffer at neutral or alkaline pH. However, the choice of buffer also plays a role. For example, no changes within several hours are seen when the metallocene is dissolved in 0.1 M aqueous hydrochloric acid ( ~ m a x = 474 nm, e = 940, which obeys Beer's Law), or in chloroform. However, 5, 10 and 66% losses in maximal spectral intensity are observed after 1 h when this reagent is dissolved in water (pH 5.5, Xmax = 463 rim), in 0.02 M Tris-HCl, p H 7.5 (hma~ = 437 rim), and in 0.02 M sodium phosphate, p H 7.5 (~'m~x = 470 rim), respectively.

Reaction of tetramethylcyclobutadiene nickel dichloride with alcohol dehydrogenase The enzymatic activity of horse liver alcohol

TABLE I SOLUBILITY A N D SPECTRAL C H A R A C T E R I S T I C S OF TETRAMETHYLCYCLOBUTADIENE NICKEL DICHLOR I D E A T 23°C Solvent

Solubility (mg/ml)

~ ma~ (e)

H20 a

50

H 2 0 / 1 M NaC1 a

40

CHC13

l0

230 (5600) 295 (320) 460 (760) 235 (8600) 307 (460) 483 (1000) 245(12 200) 503 (1440)

a A p H of 5.5 is observed for a 10 -3 M aqueous solution of tetramethylcyclobutadiene nickel dichloride.

99

original zinc content (3.6 mol of zinc per mol of enzyme), and contain no nickel. Additional control samples, involving treatment of the enzyme with nickel dichloride in the presence of several complexing or chelating agents, show no loss in enzymatic activity relative to blanks treated only with the complexing or chelating agents.

A o

Discussion

x ~u

t

1

I

350

450

5 50

X, nm

Fig. 1, Circular dichroic spectrum (top) and absorption spectrum (bottom) of tetramethylcyclobutadiene nickel alcohol dehydrogenase in 0.1 M Tris-acetate buffer (pH 6.5). This solution was prepared by treating the enzyme under oxygen-free conditions with a 105-fold molar excess of the nickel metallocene, and then removing excess and hydrolyzed decomposed metallocene by dialysis prior to measurement of these spectral features. The extinction coefficients and molar absorptivities are calculated based on the total nickel concentrations of the solutions, measured by atomic absorption. Since further dialysis of this solution does not change its nickel concentration, the total nickel is equivalent to enzyme-bound nickel.

dehydrogenase decreases to values of 15, 13, 6 and 4% after 45, 50, 70 and 90 min exposure, respectively, to a 105-fold molar excess of the nickel metallocene in Tris buffer at pH 7.0. Dialysis for 10 h yields an intensely orange solution, with less than 4% of the original enzymatic activity, less than 0.1 mol of the initial zinc, and 7.6 mol of nickel per mol of enzyme. The visible absorption spectrum exhibits a maximum at 450 nm (es600) with a shoulder at 483 (e4550) as shown in Fig. 1. The peak at 483 nm is optically active with a positive molar ellipticity of 9800 d e g - c m 2. decimol-~. These results are no different, within experimental error, for a sample treated and measured the same except for 16-h dialysis. The dialyzed blank and control samples, involving treatment of the enzyme either with buffer alone (blank) or with a 105-fold molar excess of nickel dichloride (control) retain their enzymatic activity and

Tetramethylcyclobutadiene nickel dichloride originally was synthesized from 1,2-dichloro1,2,3,4-tetramethylcyclobut-3-ene and nickel tetracarbonyl [6,10]. The crystal structure of the nickel reagent reveals a dimeric species in which three chloride ions are co-ordinated to each nickel atom [11]. Its capacity to form charged, hydrated monomers as well as noncharged, bridged dimers [12] no doubt relates to its good solubility in the divergent solvent systems examined in this work. For example, it is quite soluble in both chloroform and aqueous 0.1 M sodium chloride. The greater stability observed here for this metallocene under aqueous conditions either at lower pH, or in Tris vs. phosphate buffer at p H 7.5, presumably relates to differences in the ligation of the nickel atom under these conditions. Reaction of the metallocene with cold, aqueous sodium nitrite has been shown to yield cis-tetramethyl-3-cyclobutene- 1,2-diol [6]. The chemical nature of the orange soluble adduct, essentially devoid of native zinc and enzymatic activity, that tetramethylcyclobutadiene nickel dichloride forms with alcohol dehydrogenase remains to be defined. We wish to offer two speculations at this stage to provide some focus for future work. The first one is that the metallocene potentially may be occupying, in part, the extra metal binding sites (beyond those for the native zinc ions) reported by Andersson et al. [ 13] for this enzyme. Our second speculation is that some of the overall modification may involve attachment of the nickel atom of the metallocene onto the sulfhydryl groups of the enzyme. Three considerations potentially are consistent with this latter hypothesis. First, it is reported that the intrinsic zinc, which is bonded partly (catalytic zinc) or fully bonded (structural zinc) to sulfhydryl groups in the native enzyme [14], is lost when the enzyme reacts with the metallocene, consistent with

100

a low stability of this zinc when these sulfhydryl groups are carboxymethylated [15]. Second, the metallocene has been observed to form a mercaptide complex with dimercaptodicyanoethylene under aqueous conditions [12], illustrating the potential of this reagent to react with SH groups. Third, tetramethylcyclobutadiene nickel dichloride has no effect on the activity of bovine carboxypeptidase A, alkaline phosphatase of Escherichia coli or bovine ribonuclease A - - all non-sulfhydryl enzymes -- but it does inactivate rabbit muscle lactate dehydrogenase and rabbit muscle aldolase, known sulfhydryl enzymes (unpublished data). Thus, tetramethylcyclobutadiene nickel dichloride, a half-sandwich metallocene, provides intense reactivity and spectral characteristics for the modification of horse liver alcohol dehydrogenase, unlike the inertness of nickel chloride towards this enzyme. This illustrates the unique chemical and physical properties available from half-sandwich metallocenes for protein modification. Among the many extensions of this work is the possibility of adding a chemically reactive group onto the organic component of a half-sandwich metallocene, potentially allowing control of chemical modification by extrinsic metal binding.

Acknowledgement The authors thank Bert L. Vallee for initiating and encouraging this project.

References I Kornicker, W.A. and Vallee, B.L. (1969) Ann. N.Y. Acad. Sci. 153, 689-705 2 Giese, R.W. (1983) J. Inorg. Biochem., in the press 3 Giese, R.W. and Vallee, B.L. (1972) J. Am. Chem. Soc. 94, 6199- 6200 4 Einarsson, R., Wallen, L. and Zeppezauer, M. (1972) Chemica Scripta 2, 84-87 5 Hetnarski, B., Lajtha, A. and Wisniewski, H.M. (1980) J. Neurosci. Res. 5, 1-5 6 Criegee, R. and Schroder, G. (1959) Angew. Chem. 71, 70 - 7 I 7 Thiers, R.E. (1969) Methods Biochem. Anal. 5, 273-335 8 Fuwa, K., Pulido, P., McKay, R. and Vallee, B,L. (1964) Anal. Chem. 36, 2407-2411 9 Drum, D.E., Li, T.-K. and Vallee, B.L. (1969) Biochemistry 8, 3783-3791 10 Criegee, R., Dekker, J., Engel, W., Ludwig, P. and Noll, K. (1968) Chem. Ber. 96, 2362-2367 11 Dunitz, J.D., Mez, H.C., Mills, O.S. and Shearer, H.M.M. (1962) Helv. Chim. Acta 45, 647-665 12 Cricgee, R. and Zanker, F. (1964) Angew. Chemie 76, 716-718 13 Andersson, I., Maret, W., Zeppezauer, M., Brown, R.D., III and Koenig, S.H. (1981) Biochemistry 20, 3433-3438 14 Eklund, H., Nordstrom, B., Zeppezauer, E., Soderlund, G., Ohlsson, I., Boiwe, T. and Branden, C.-I. (1974) FEBS Lett. 44, 22-204 15 Li, T.-K. and Vallee, b.L. (1965) Biochemistry 4, 1195-1202