Purification and Characterization of Component A of the Methane Monooxygenase from Methylococcus cupsulutus (Bath)*

capsulatus (Bath) possesses a multi- component methane monooxygenase which catalyzes in vivo the conversion of methane to methanol. Component A of this enzyme system, believed to be the oxygenase component, has been purified to near ho-mogeneity (95%). The native protein has a molecular weight of approximately 210,000 and is comprised of three subunits of M, = 54,000, 42,000, and 17,000, which appear to be present in stoichiometric amounts suggesting an aZ, &, yz arrangement in the native protein. Purified preparations of the protein are virtually colorless and examination of the uv/visible absorption spectrum revealed a peak around 280-290 nm and thereafter a steady decrease in absorbance to longer wavelengths. The ESR spectrum of the oxidized protein gave a signal at g = 4.3, presumably due to rhombic iron, and a radical signal at g = 2.01. Upon reduction with dithionite, a signal at g = 1.934 appeared. Chemical analyses of our purified preparations revealed the presence of iron (2.3 mol/mol) and zinc (0.2-0.5 mol/mol): molybdenum, copper, nickel, heme, and acid-labile sulfur were all virtually absent. On ultra thin layer isoelectric focusing, purified component A was judged to have a PI between 5.0 and 5.1. Extracts prepared from a variety of other methano- trophs failed to show any cross-reaction to antibody raised against M. capsulatus component A.

Methanotrophs, organisms capable of utilizing methane as their sole carbon and energy source, oxidize methane to carbon dioxide via methanol, methanal (formaldehyde), and formate. In Methylococcus capsulatus (Bath), the initial oxidation of methane to methanol is known to be mediated by a monooxygenase: CH, + 0 2 + NAD(P)H + H+ -+ CH, OH + NAD(P)+ + HZ0 (for a review, see Ref. 1).
In this organism, the enzyme has been shown to be capable of utilizing a variety of other alkanes, alkenes, ethers, alicyclic, aromatic, and heterocyclic compounds together with ammonia and carbon monoxide as substrates creating the corresponding 1-or 2-alcohols, epoxides, cyclic alcohols, phenols, pyridine N-oxide, hydroxylamine, or carbon dioxide (2, 3). A similar lack of substrate specificity has been shown to occur in enzymes from Methylomonas methanica and Methylosinus trichosporium (4).
The soluble methane monooxygenase from M. capsulatus * This research was partially supported by British Petroleum. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. was resolved into three fractions by batch ion exchange chromatography (5). Since then, further study has shown these fractions to contain three proteins each essential for in vitro methane monooxygenase activity: component A, a non-heme iron containing protein believed to function as the oxygenase; component B, a small colorless protein the role of which remains uncertain; and component C, an iron-sulfur flavoprotein which acts as an NAD(P)H acceptor-reductase. Of the three proteins, only the properties of component C have been previously reported in detail (6). This purified protein was shown to be comprised of a single polypeptide of M , = 44,600 and to contain both 1 mol of FAD and 1 mol of 2Fe-2S*(S-Cys), center per mol of protein; furthermore, the protein was found to be extremely labile in the absence of sulfhydryl group-protecting reagents such as sodium thioglycollate (1-10 mM) or dithiothreitol (5-10 mM).
A preliminary report of certain properties and purification techniques for all three proteins has appeared in review format (1). This current paper describes, in greater detail, the purification and characterization of component A, believed to be the oxygenase component of the enzyme system.

Growth of M. capsulatus (Bath) and Preparation of Soluble Extracts
Bacterial growth conditions and the preparation of soluble cellfree extracts were as previously described (5).

Methane Monooxygenase Assay
Extracts and aliquots from column fractions were assayed for component A in the presence of saturating amounts of components B and C using propene as substrate together with ethanol-free NADH (5 DIM). Epoxypropane formation was estimated by gas chromatography in a similar way to that described for the formation of expoxyethane from ethene (5).  1A). Fractions containing peak enzyme activity were rapidly concentrated by centrifugation across "Centriflo" membrane cones with a molecular weight cut-off of 25,000 (Amicon B.V., Oosterhout (N.B.) Holland).
Step 3-Gel-filtered fraction A (200 mg) was equilibrated with histidine-HC1 buffer (25 mM, pH 5.7) by passage down a "desalting" column of Sephadex G-25 (0.9 X 15 cm). The protein was then applied to a column of chromatofocusing resin PBE-94 (0.9 X 28 cm) equilibrated with the same buffer. Elution was achieved with Polybuffer at a 1:lO dilution adjusted to pH 4.5 at a flow rate of 7 ml. cm-*. h-', 2ml fractions were collected as described above and concentrated in an identical manner (Fig. 1B).
Step 4-semipurified component A from the previous step was applied to a column (1.6 X 90 cm) of Sephacryl S-300 superfine. Chromatography was done as described for Step 2 (Fig. IC).

PAGGE'
Analytical PAGGE was done using vertical 5-3074 linear gradient gels (1.5 X 150 X 180 mm), a discontinuous buffer system was used (7). For PAGGE under dissociating conditions, all buffers were supplemented with 0.1% SDS and the protein samples were treated with &mercaptoethanol and 0.2% SDS for 2 min at 100 "C prior to loading on the gel. Nondissociating gels were run at 150 mV constant voltage, while dissociating gels were run at 50 mA constant current. On completion of electrophoresis, proteins were stained with Coomassie blue R-250 (0.1%) in methano1:acetic acidwater (3:1:6) and gels were destained in the same solvent mixture.
Analytical isoelectric focusing was done in 0.5-mm thin layer polyacrylamide gels ( T = 5%; C = 3%) containing ampholines, either pH range 3.5-9.5 or pH range 4.0-6.5 (LKB-Produkter AB, Bromma, Sweden). Electrofocusing was carried out for 2000 V-h on a Multiphor gel tank system (LKB-Produkter AB, Bromma, Sweden). Focused gels were fixed in a mixture of trichloroacetic acid (11.5% w/v) and sulfosalicyclic acid (3.4% w/v) in distilled water prior to staining with Coomassie brilliant blue R-250 (0.1% w/v) in ethanokacetic acidwater The molecular weight of component A was also determined by PAGGE as previously described (8). To ensure protein samples and molecular weight marker proteins approached zero velocity, gels were run for 3000 V h.
The molecular weights of component A subunits were determined by SDS-PAGGE as previously described (9).

Analytical Ultracentrifugation
Sedimentation velocity experiments were done using a Beckman Model E analytical ultracentrifuge. 4.5 mg of purified component A in potassium phosphate buffer (10 mM, pH 7.0) containing sodium chloride (100 mM) was centrifuged at 48,000 rpm in an An-D rotor equipped with a single sector aluminum center piece and plane quartz windows. The sedimenting boundary was followed using schlieren optics. The SOBS was corrected to water at 20 "C.

E S R Spectroscopy
Electron spin resonance spectra of component A were recorded using a Varian E4 spectrometer (Varian Associates, Ltd., GB-Walton on Thames, Surrey, U.K.).

Determination of Iron
The iron content of component A was assayed by "wet-ashing" the protein in concentrated sulfuric, nitric, and perchloric acids. The acid digests were transferred to volumetric flasks (5 ml) containing saturated sodium acetate (2.0 ml), 20% ascorbic acid (0.45 ml), and 10 mM bathophenanthroline sulfonate (0.15 ml), the volume was adjusted to 5 ml with distilled water. The absorbance at 535 nm was measured and compared with a standard curve prepared from solutions of ferrous ammonium sulfate to estimate the iron content.

Determination of Zinc
The zinc content of component A was assayed by atomic absorption spectroscopy of acid digests of the protein prepared as described above. The absorbance at 213 nm was measured and compared with a standard curve prepared from solutions of zinc in hydrochloric acid. the alkaline zinc reagent was extended to 2 h as previously recommended (14). A standard curve was constructed using sodium sulfide, standardized by iodometry.

Determination of Protein Sulfhydryl Content
Total protein sulfhydryl content was assayed using Ellman's reagent as previously described (15).

Determination of Protein
Protein was assayed using commercially available Bio-Rad reagent (Bio-Rad Ltd., Watford, Herts, U.K.). An "absolute" series of determinations using freeze-dried purified component A gave around a 15% lower value when compared with bovine serum albumin routinely used as the protein standard.

Production of Antiserum
Antiserum to purified component A was raised in New Zealand White rabbits by the subcutaneous injections of 2-mg aliquots of pure protein homogenized with Freunds' adjuvant.

Immunodiffusion
Analysis of antisera and their cross-reactivity with extracts of various methanotrophs was done by double diffusion on Ouchterlony plates a t 30 "C for 24 h.

Purification of Component
A-The purification sequence as described here yields component A of around 95% homogeneity as judged by the following criteria. Sedimentation velocity experiments revealed a single symmetrical schlieren peak (Fig. 2); on nondissociating PAGGE, a single major band was  visible together with faint bands a t loadings of 100 pg of protein ( Fig. 3):' Dissociating PAGGE revealed three bands of approximately equal intensity together with 1 minor band, estimated to be less than 5% of total protein by densitometric analysis (Fig. 4).
Although the purification scheme described (Table I) clearly removes the bulk of extraneous protein, it has been a constant feature of this work that the purification is not accompanied by a corresponding increase in the specific activity of the enzyme. Densitometric analyses suggests that in certain extracts component A may comprise up to 30% of the soluble cell protein, therefore one might expect a t least a 3fold increase in specific activity of the enzyme.
Stability of Component A-Crude preparations of the enzyme appeared to be relatively stable losing up to 40% of their activity at 4 "C over a 72-h period. This loss of activity was not prevented by a variety of typical stabilizing agents, methane or nitrogen atmospheres, or various oxygen radical scavengers (Table 11): it is worth noting that @-mercaptoethanol was found to be strongly inhibitory. As is clear from the purification scheme, instability was more apparent during column chromatography. This might be interpreted as suggesting the involvement of a further protein or co-factor; however, on gel filtration enzyme activity was associated with a single uv absorbing peak (Fig. IC) and the readdition of other column fractions was not found to stimulate enzyme activity. In addition, several stability experiments were carried out during gel filtration including the following. of substrate (cyanomethane, 2 mM). (iv) Gel filtration was also attempted at room temperature both in the presence and absence of the protease inhibitor phenylmethylsulfonyl fluoride (50 p~) . None of these techniques was found to enhance the stability of the protein; furthermore, the inclusions of Fez+, Fe3+, Zn2+, Ni2+, Cu2+, or Mo3+ (0.1 mM) in the enzyme assay did not stimulate the activity of the enzyme.
Although crude fraction A could be stored in the frozen state, it was found essential not to subject the protein to cycles of freezing and thawing during the latter stages of purification. However, it proved possible to store the purified protein at -20 "C in 50% glycerol for several weeks without loss of enzyme activity.
Physical Properties of Component A-The molecular weight of component A was estimated to be approximately 210,000 by gel filtration (Fig. 5), this compares well with a previous estimate of 220,000 (l), however a slightly higher figure of 253,000 was obtained by PAGGE.
The subunit composition of component A was examined by dissociating PAGGE, the purified protein was found to give rise to three bands corresponding to molecular weights of a = 54,300, p = 36,500, and y = 17,300. The total subunit molecular weight is therefore 108,000 (Fig. 6); since densitometric analysis shows that the three subunits are present in approximately stoichiometric amounts (Fig. 4B), it seems probable that the holoprotein consists of two copies of each subunit.   The isoelectric point of component A was estimated to lie between pH 5.0 and pH 5.1. On ultrathin layer isoelectric focusing gels, the protein did not focus to a single tight band but tended to give a diffuse band. Whether this was due to microheterogeneity of our samples, reflects an intrinsic property of the protein, or is due to various "redox" states of the protein is not yet clear.
Purified component A is virtually colorless, even at protein concentrations up to 50 mg . ml-'. The uv/visible absorption spectrum shows a maximum at around 280-290 nm and then a steady decrease to longer wavelengths (Fig. 7). In some of our preparations, a weak "shoulder" at around 406-410 nm was present, the EM at 410 nm was estimated to be 2.6 X lo3.
The "shoulder" was lost upon reduction of the protein with sodium dithionite and also if the protein was "relaxed" in an aprotic solvent mixture such as hexamethyl phosphoramide:water (4: 1 v/v) .
The ESR spectrum of the oxidized protein displayed a signal a t g = 4.3, presumably due to rhombic iron and a radical a t g = 2.01. Upon reduction of the protein with sodium dithionite (2 mM) for 2 min, the g = 4.3 signal was slightly diminished, the g = 2.01 signal remained, and an unusual signal appeared below g = 2, the peak being at g = 1.934, the crossover at g = 1.87 and a long tail to high field (Fig. 8).
The ESR spectrum was also examined in the presence of the enzyme substrate cyanomethane (2 mM). While the oxidized spectrum was unaffected by the presence of substrate, the spectrum of the reduced protein was seen to be narrower and display a clearer rhombic line shape in the g = 1.934 signal (Fig. 9).5 The value for s~~,~ was calculated from the SOBS by the following equation: A value of 0.725 was assumed for the partial specific volume (9).  The physical properties of component A are summarized in Table 111.
Chemical Properties of Component A-Samples of the purified protein were analyzed for the presence of Cu, Fe, Mo, Ni, and Zn by inductively coupled plasma emission spectroscopy. This single analysis of both dialyzed and undialyzed protein indicated the presence of 1.25 mol of Fe per mol and 0.5 mol of Zn per mol of protein. While the Zn concentration was comparable in both dialyzed and undialyzed samples, the amount of Fe was substantially reduced in the dialyzed sample (Table IV). Further metal analysis for iron and zinc was done by (i) colorimetric method (Fe) and by (ii) atomic absorption spectroscopy (Zn). Typically, values for iron were 2.3 + 0.7 mol of Fe. mol of protein-', while values for zinc were between 0.2 and 0.5 mol. mol".
Bearing in mind the iron content of component A and that major contaminants of crude fraction A included several cytochromes, preparations of component A were examined for heme content. The pyridine-hemochromogen difference spectrum of crude fraction A showed: a maximum, 551 nm; a-P minima, 536 nm; 6 maximum, 521 nm in samples of Previously, it had been thought that component A might possess a single 2Fe-2S* cluster, this suggestion was based upon preliminary cluster extrusion data (1). Accordingly, several attempts at cluster extrusion were made on our preparations of component A. It was observed that incubation of component A with excess thiophenol in hexamethyl phosphoramide:Tris-HC1 (50 mM, pH 8.5) (4:l v/v) under anaerobic conditions led to an increase in absorbance between 550 and 570 nm over a period of several hours. However, neither a maximum a t 454 nm (4Fe-4S*) or at 474 nm (2Fe-2S*) was detected. A control core extrusion was done using Clostridium pasteurianum ferredoxin (2 X 4Fe-4S*) and gave the expected spectrum; furthermore, the incubation of ferrous sulfate gave rise to a visible absorption spectrum virtually identical with that obtained with component A. This latter observation suggested that the spectral changes which occurred with component A might be due to the presence of "adventitious" iron.
Samples of the protein (up to 30 nmol) were also analyzed for acid-labile sulfide as previously described (12); negligible levels were detected, thus confirming previous results.'j The chemical properties of component A are summarized in Table  IV. Other methanotrophs were tested for the presence of component A by permitting crude cell-free extracts to diffuse against antibody raised to component A on Ouchterlony plates. In Methylomonas methanica (Strain S), Methylocystis parvus (OBBP), Methylobacter capsulatus and Methylosinus trichosporium (OB3b), and Methylomonas albus, no crossreaction was observed.

DISCUSSION
These studies show component A of the methane monooxygenase from M. capsulatus (Bath) to be an acidic, non-heme iron-containing protein of approximate M , = 210,000; zinc may also be a constituent. On the basis of densitometric analysis of dissociating PAGGE, the native protein would appear to be oligomeric consisting of two copies of each of three subunits of M , = 54,000, 36,000, and 17,000.
The purification sequence described here yields component A of an acceptable degree of purity for further study of the complete methane monooxygenase system. However, a problem which remains to be overcome in the study of this protein is the apparent loss of enzyme activity during purification. While the protein is a major component of the soluble fraction of the cell, we would expect to see a 3-10-fold increase in specific activity. There is no evidence as yet to implicate the Component A of the Methane Monooxygenase involvement of a further protein and we feel that the loss of enzyme activity may well be due to loss of iron from the protein; certainly both iron and enzyme activity are much reduced upon dialysis. Furthermore, experiments have been done (not reported) which suggest that preincubation of the protein with iron and dithiothreitol prior to assay markedly stimulated enzyme activity; unfortunately, these observations have not proved sufficiently reproducible to report in detail.
The role of zinc in the protein is unclear, whether it is a contaminant or plays a catalytic or structural role remains unknown. Indeed, only one other oxygenase known to the authors contains zinc, the 4-hydroxyphenyl-pyruvate dioxygenase from Pseudomonas sp strain P.J. 874, and its role of this enzyme is apparently unknown (16).
The preliminary ESR studies on component A are of interest in that they suggest the possibility of a novel active center in this enzyme. The spectra do not look like those obtained from normal iron-sulfur proteins; the oxidized protein gave two signals, one due to rhombic Fe3+ (g = 4. 3) and an unusually broad radical signal (g = 2.01). The reduced protein gave rise to a second unusual signal below g = 2 (reduced Fe-S clusters almost invariably give one g value above 2). If this signal is due to Fez+, it is presumably coupled to some other S = 1/2 system such as a radical like the Fe-quinone complex in the primary acceptor of bacterial photosynthesis (17). The only evidence to date for the role of component A in the methane monooxygenase complex was also suggested by ESR studies. Although the oxidized spectrum of component A remained unaffected by the substrate cyanomethane, on reduction, the g = 1.934 signal was clearly altered, being increased in size, narrowed, and displaying a clearer rhombic line shape. A similar observation was made previously when ethene was used as the substrate (1). These results suggest that reduced component A may be responsible for substrate binding.
The only other methane monooxygenase to be studied in any detail is that from M. trichosporiurn (OB3b) (18). In this system, the oxygenase component was proposed to be a single polypeptide of M , = 46,000 containing 1 mol. mol" of copper. Clearly, this would seem to be a very different protein. We have also shown a lack of immunological cross-reaction between antibody to component A and crude extracts of M. trichosporium (0B3b). However, earlier studies (4) demonstrated a functional relationship between component A from M. capsulatus (Bath) and crude fraction 1 from M. trichosporium (OB3b). It was shown that components B and C from M . capsulatus would permit fraction 1 to function as an oxygenase.
A consideration of other alkane hydroxylating systems reveals an overall similarity between several enzymes: many are multicomponent, comprised of an NADH acceptor-reductase, a small iron-sulfur protein, and an oxygenase component. However, within the apparent similarity lies several distinct differences particularly with respect to the oxygenase components. Some are iron-sulfur proteins e.g. putidamonoxin (19), benzene dioxygenase (20), pyrazon dioxygenase (21), and benzoate dioxygenase (22). The oxygenase components from adrenodoxin, the n-alkane hydroxylating system of liver microsomes (23), and Corynebacterium (24) contain cytochrome P-450. The w-hydroxylase from P. oleovorans is unusual in that it contains non-heme iron alone as the prosthetic group (25).
Based upon our current findings, component A does not resemble any of these oxygenases. The unusual ESR spectrum, the presence of iron and zinc, and the absence of heme together with the apparent lack of an extrudable iron-sulfur cluster suggest a novel iron-containing prosthetic group which clearly requires further elucidation. S u p p l e m e n t a r y m a t e r i a l to: