Covalently Bound FAD in D-6-Hydroxynicotine Oxidase IMMUNOLOGICAL STUDIES ON D- AND L-6-HYDROXYNICOTINE OXIDASE: EVIDENCE FOR A D-ENZYME PRECURSOR*

Antisera prep.ared against both enantiozymes, D- and L-6-hydroxynicotine oxidase, formed precipitins in double diffusion tests with their respective antigens only. A mixture of the two antisera caused spur formation of the two precipitin lines obtained with the pure enzymes. Antiserum to L-apoprotein reacted with native L-enzyme and L-apoprotein but not with the n-specific enzyme. n-6-hydroxynicotine oxidase activity was inhibited by the anti-D-antiserum, leaving the L-enzyme fully active, while anti-L-antiserum inhibited the L- but not the n-specific activity. The delayed induction of n-6-hydroxynicotine oxidase as compared to the other activities of the nicotine-degrading sequence and the differential immunochemical behavior of the enantiozymes allowed the search for a n-enzyme precursor. In cells harvested 3 hours after the addition of m-nicotine, the L-enzyme activity was present, whereas no n-enzyme activity could be detected. However, an extract of these cells did form an immunoprecipitin line with anti-n-antiserum. L-6-Hydroxynicotine nm 50 5 and centrifuging. LInduced cells only L-enzyme but no D-enzyme activity were grown in a glucose medium containing 3.06 g of KH,PO,, 9.85 g of Na,HPO,. 12H,O, 0.5 g of yeast extract, 50 ml of trace salt solution, 5 g of glucose and 2.0 ml of DL-nicotine/liter. The pulse-chase experiment was conducted as follows: 50 ml of glucose medium were inoculated with 5% glycerol-grown cells, supplemented 5 hours later with 12.5 &i of L-[U-“Clleucine (348 mCi/ nmol) and aerated overnight. The cells then were transferred into a glucose medium without radioactivity to exhaust the cells of endoge- nous [‘C]leucine during further growth. Twenty-five milliliters of the culture were centrifuged and stored (“labeled L-induced cells”). The remaining 25 ml were centrifuged under sterile conditions, transferred into 60 ml of inducing medium (0.2% titrate/0.05% DL-nicotine/O.2% ammonium sulfate) and incubated overnight. This culture, correspond- ing to half of the previously labeled cells was centrifuged and sonified in 10 mM phosphate buffer, pH 7.5 (4 ml). The “labeled L-induced cells” were treated in the same way. Both extracts were equally concentrated; a lo-p1 aliquot each used for Ouchterlony double-diffusion experiments. An x-ray RPR 54, x was exposed to the stained plate for days. developed Ouchterlony

Antisera prep.ared against both enantiozymes, D-and L-6-hydroxynicotine oxidase, formed precipitins in double diffusion tests with their respective antigens only. A mixture of the two antisera caused spur formation of the two precipitin lines obtained with the pure enzymes. Antiserum to L-apoprotein reacted with native L-enzyme and L-apoprotein but not with the n-specific enzyme. n-6-hydroxynicotine oxidase activity was inhibited by the anti-D-antiserum, leaving the L-enzyme fully active, while anti-L-antiserum inhibited the L-but not the n-specific activity. The delayed induction of n-6-hydroxynicotine oxidase as compared to the other activities of the nicotine-degrading sequence and the differential immunochemical behavior of the enantiozymes allowed the search for a n-enzyme precursor. In cells harvested 3 hours after the addition of m-nicotine, the L-enzyme activity was present, whereas no n-enzyme activity could be detected. However, an extract of these cells did form an immunoprecipitin line with anti-n-antiserum. L-6-Hydroxynicotine oxidase, but no n-6-hydroxynicotine oxidase activity, could also be induced in Arthrobacter oxidans grown in a medium with a high glucose content and m-nicotine as the sole nitrogen source. An extract of these L-induced cells produced the specific immunoprecipitation with anti-Dantiserum. A pulse-chase experiment with cells grown first on glucose and DL-nicotine in the presence of ["Clleucine and then in an unlabeled medium which induces n-6-hydroxynicotine oxidase activity resulted in a radioactive n-enzyme-immunoprecipitin line. From these experiments it is concluded that a precursor of the active n-enzyme is induced simultaneously with the other nicotine-degrading enzymes.
In Arthrobacter oxidans m-nicotine induces both a D-and a L-6-hydroxynicotine oxidase (1, 2). These "enantiozymes," a term previously proposed (3) for specific enantiomeric enzymes, show different induction kinetics (4, 5) the n-specific activity appearing much later than the L-enzyme' (5). Molecular weight, subunit structure, reaction mechanism, specificity, and reactivity toward artificial electron acceptors of both purified enzymes have been established (3). The most remarkable difference between these enzymes is their coenzyme binding. L-6-Hydroxynicotine oxidase contains FAD reversibly attached, whereas in the n-specific enzyme the FAD is covalently bound to the N-3 of a histidyl residue of the apoprotein via the 8a-methylene group of the flavin nucleus (6). This communication deals with the antigenic behavior of the *This work was supported by the Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg, through Sonderforschungsbereich 46.

enantiozymes.
The immunological evidence for the presence of a n-enzyme precursor in cells lacking the active n-enzyme suggests that the retarded appearance of n-6-hydroxynicotine oxidase activity is due to the process of covalent FAD binding.

EXPERIMENTAL PROCEDURES
Homogeneous o-6-hydroxynicotine oxidase from Arthrobacter oxidans was prepared as described previously (2). Pure L.-6-hydroxynicotine oxidase was obtained during the same procedure; the fractions of DEAE-cellulose chromatography containing L-enzyme activity were further purified by preparative disc electrophoresis (pH 10) and finally by gel filtration on Sephadex G-200 superfine. L-apoenzyme was prepared by the procedure of Warburg and Christian (7). D-Amino acid oxidase (EC 1.4.3.3) was purchased from Boehringer Mannheim GmbH (Mannheim, Germany).
Antisera against both the D-and the L-enzyme and the L-apoenzyme were induced in rabbits (hybrid of white Vienna multiplied Alasca). Enzyme (1.5 ml) and 2% Al (OH), (1.5 ml) were kept at 4" for at least 1 hour; Freud's complete adjuvans (3.0 ml) was added and the mixture was emulsified. Two milliliters of the emulsion, containing 0.8 mg of enzyme, were injected into the rabbit intracutaneously and into the footpads. A second injection of 2 mg of enzyme was administered subcutaneously 6 to 10 weeks later. Each rabbit was bled at lo-day intervals, beginning 10 days after the second injection. Antisera of the first bleeding were used for the experiments.
Experiments to determine the inhibition of enzyme activity were performed in tapered centrifuge tubes. A crude extract obtained by sonification of fully induced cells in 10 mM phosphate buffer, pH 7.5, was incubated overnight at 25" with various amounts of antiserum and 0.9% NaCl in a total volume of 0.3 ml. After centrifugation, aliquots were removed for the assays of D-and L-6-hydroxynicotine oxidase activity (3).
A 30% inoculumof glycerol-grown ceils (0.2% glycerol instead of Na,-citrate and DLnicotine) was used. After 3 hours, when the L activity was already induced but almost no D activity was present, the cells were harvested and stored at -30". Small quantities of cell-free extract were obtained by sonifying (Branson Sonifier J17V, scale adjustment, 1) 5 ml of resuspended cells (A/ml at 691 nm = 50 to 60) for 5 min and centrifuging. LInduced cells containing only L-enzyme but no D-enzyme activity were grown in a glucose medium containing 3.06 g of KH,PO,, 9.85 g of Na,HPO,. 12H,O, 0.5 g of yeast extract, 50 ml of trace salt solution, 5 g of glucose and 2.0 ml of DL-nicotine/liter.
The pulse-chase experiment was conducted as follows: 50 ml of glucose medium were inoculated with 5% glycerol-grown cells, supplemented 5 hours later with 12.5 &i of L-[U-"Clleucine (348 mCi/ nmol) and aerated overnight. The cells then were transferred into a glucose medium without radioactivity to exhaust the cells of endogenous ['C]leucine during further growth. Twenty-five milliliters of the culture were centrifuged and stored ("labeled L-induced cells"). The remaining 25 ml were centrifuged under sterile conditions, transferred into 60 ml of inducing medium (0.2% titrate/0.05% DL-nicotine/O.2% ammonium sulfate) and incubated overnight. This culture, corresponding to half of the previously labeled cells was centrifuged and sonified in 10 mM phosphate buffer, pH 7.5 (4 ml). The "labeled L-induced cells" were treated in the same way. Both extracts were equally concentrated; a lo-p1 aliquot of each was used for the Ouchterlony double-diffusion experiments. An x-ray film (Kodak, RPR 54, 13 x 18) was exposed to the stained plate for 4 days. The developed film was photographed together with the Ouchterlony plate.

Zmmunoprecipitation
of D-and L-6-Hydroxynicotine Onidase by Respective Antisera-In an immunodiffusion experiment, anti-D-antiserum reacted only with pure D-enzyme and with an extract of fully induced cells containing L and D activity (Fig. LA). No precipitin line could be detected with pure Lenzyme, even at higher concentrations, and with r.-apoprotein. Anti-D-antiserum formed a precipitate with pure L-enzyme, fully induced cell extract, and L-apoprotein (Fig. 1B). D-Enzyme was not precipitated by anti-L-antiserum; D-amino acid oxidase, another FAD-containing protein, reacted with neither of the two antisera. With a mixture of the two antisera in the center well spur formation of the precipitin lines of the pure enantiozymes was seen; the extract of fully induced cells elicited two precipitation lines (Fig. 1D). Antiserum against the L-apoenzyme produced lines of identity with L-apoenzyme, pure L-enzyme, and with an extract of induced cells, but no precipitation occurred with D-enzyme or D-amino acid oxidase (Fig. 1C). D activity was found in the supernatant whereas the L activity was fully retained (Fig. 2A). The opposite result was obtained by incubating the extract with anti-L-antiserum (Fig. 2B). In a crude extract, the quantity of the respective antiserum required to reach 50% inhibition of activity was proportional to the amount of enzyme protein (Fig. 3).

Immunological Detection of D-Enzyme
Precursor-The search for a precursor protein of D-6-hydroxynicotine oxidase became feasible when the different antigenicity of the enantiozymes was established and conditions of partial induction were developed which allow to obtain cells with L activity only. These cells were grown on an inducing medium and harvested 3 hours after inoculation; at this time the high L activity began to level off but the D activity has not yet been induced (5) (Fig. 4). An extract of these cells produced a distinct, although weak, precipitin line with anti-D-antiserum (Fig. 5A). This line was different from that obtained with the D-enzyme but appeared to be identical with the weak inner line of fully induced cells (Fig. 5A); it does not pass over into the precipitin line formed with pure D-6-hydroxynicotine oxidase (Fig. 5C). Noninduced cells did not show these two precipitin lines. Furthermore, the removal of antibodies from the anti-D-antiserum by incubation with purified D-enzyme yielded a preparation which no longer reacted with the precursor protein. The possibility that a minute amount of active D-enzyme present in the 3-hour induced cell extract could have caused the distinct, weak precipitin line, was excluded since the same amount of anti-D-antiserum does not form a detectable band with less than 100 milliunits/ml of D-enzyme (not shown). The D-specific activity of the 3-hour induced cell extract, however, was less than 10 milliunits/ml. As was expected, the anti-L-antiserum reacted with the 3-hour induced cell extract but not with an extract of noninduced cells. (Fig. 50).

Reaction
of Extract of L-Znduced Cells with Anti-D-Antiserum-When Arthrobacter oxidans was grown on 0.5% glucose as carbon and 0.2% DL-nicotine as sole nitrogen source,

D-enzyme activity was undetectable
whereas the L-enzyme was fully induced. The same result was obtained when glucose was replaced by 0.5% glycerol or 0.5% citrate. Addition of ammonium sulfate caused both enzymes to be drastically repressed, even in the presence of DL-nicotine (Table I). The immunological assay of extracts of these L-induced cells lacking D-6hydroxynicotine oxidase activity with anti-D-antiserum resulted in precipitin lines which converged with the band of the 3-hour induced cell extract (Fig. 6B). The extracts of cells grown in the presence of (NH,)&SO, did not show this precipitin line (Fig. 6A).
Conversion of Radioactive D-Enzyme Precursor into Radioactive D-Enzyme in Viuo-In a pulse-chase experiment outlined under "Experimental Procedures," cells were grown in the L-inducing glucose medium supplemented with L-["Clleucine ("labeled L-induced cells"). They contained no measurable D activity. However, since L-6-hydroxynicotine oxidase is induced under these conditions, any precursor of n-6-hydroxynicotine oxidase, e.g. the n-apoprotein, when induced concomitantly ought to be "C-labeled. After transfer of the labeled L-induced cells into the nonradioactive, fully inducing medium, which allows the synthesis of all nicotine degrading enzymes including n-6-hydroxynicotine oxidase, the latter enzyme should contain radioactivity if a labeled n-enzyme precursor was present in the labeled L-induced cells. Fig. 7 shows the photograph of an Ouchterlony immunodiffusion test (B) and of the x-ray film exposed to this slide (A). Both the D-enzyme precursor and the active D-enzyme which were well separated, were radioactive. DISCUSSION It was demonstrated that the enantiozymes D-and L-6hydroxynicotine oxidase from Arthrobacter oxidans are immunologically different. Whether this is caused by a dissimilarity of the polypeptides per se or by a different tertiary and quaternary enzyme structure cannot be decided at this time.
Among the first four DL-nicotine degrading enzymes, the induction time constant (Z',) of L-6-hydroxynicotine oxidase, as we]] as of the stereounspecific enzymes, nicotine dehydrogenase (EC 1.5.99.4) and ketone dehydrogenase (formerly called nicotine oxidase and ketone oxidase (5)) is 1 hour whereas that of D-6-hydroxynicotine oxidase is 3 hours (5). The delay of the appearance of D activity may result from the retarded induction of an enzyme system involved in the process of covalent FAD-binding.
In this case, the induction of the D-apoenzyme could still be correlated to that of the other enzymes of the sequence. In fact, extracts of L-induced cells which exhibited only L-but no D-eDZyDM? activity contained a protein that reacted with anti-D-antiserum.
Several criteria point to this antigen as a precursor of n-6-hydroxynicotine oxidase. (a) The immunological response of the precursor protein to antibodies elicited by a purified D-6-hydroxynicotine oxidase preparation is highly specific. Anti-D-antiserum which was preincubated with an excess of FIG. 5. Immunoprecipitation of a 3-hour induced cell extract. The center wells contained anti-n-antiserum (A and C), anti-o-antiserum, the specific antibodies of which were removed with purified n-enzyme (B), and anti-L-antiserum (D). The antisera were diluted 1:8. A and B, extracts of: 1, repressed cells; 2 and 5, 3-hour induced cells; 3, fully induced cells; 4 and 6, o-enzyme. C: 1, 3, and 5, n-enzyme; 2, 4, and 6, extract of 3-hour induced cells. D, extracts of: 1, repressed cells; 2 and 5, 3-hour induced cells; 3, fully induced cells; and 6, L-enzyme. Numbers are read clockwise starting at the top. the D-enzyme failed to produce the characteristic precipitin line with precursor-containing cell extracts (Fig. 5). Repressed cells which synthesize neither L-nor D-6-hydroxynicotine oxidase did not contain proteins reacting with anti-Dantiserum. (b) The precursor protein could be detected immunologically not only in fully induced Arthrobacter oxidans which produces both enantiozymes but also in Linduced and 3-hour induced cells which contain L-but no n-enzyme activity. (c) Cells which were grown under these conditions of partial induction in the presence of [Ylleucine produced a heavily labeled precursor protein hut, as expected, no radioactivity was found in the precipitin line of added n-enzyme. When these cells were transferred to an unlabeled fully inducing medium in which D-6-hydroxynicotine oxidase is synthesized the precipitin line of the D-enzyme became strongly labeled; since de nouo protein synthesis in this medium results in the formation of unlabeled proteins, it is concluded that the polypeptide preformed in the L-inducing medium, i.e. the precursor, was converted to the active n-enzyme.
FIG. 6 (top). Immunoprecipitation by anti-o-antiserum of cell extracts free of o-enzyme activity. Cells were grown cm m-nicotine either in the presence (A) or in the absence (B) of 0.2% (NH,),SO,. Center wells contained anti-o-antiserum, 1:8 diluted. Peripheral wells (read clockwise from top) contained extracts of cells grown on: 1, fully inducing medium; 2, 0.5% citrate; 3, 0.5% glucose; 6, 0.2% m-nicotine only. Media 2 to 6 contained 0.2%, medium 1,0.05% m-nicotine. FIG. 7 (bottom). Immunoprecipitation of the pulse-chase experiment. For details see "Experimental Procedures." Anti-o-antiserum, 1:8 diluted, in the center well. The peripheral wells contained extracts of: 1, fully induced cells; 2, fully induced cells propagated in nonradioactive inducing medium that was inoculated with ["C]leutine-labeled L-induced cells; 3 and 5, 3-hour induced, unlabeled cells; 4, labeled cells grown in L-inducing medium, with purified u-enzyme added; 6, same as 4, without addition of purified n-enzyme. A represents the autoradiogram of the stained Ouchterlony plate (n) developed 4 days after exposure. Cells were grown in an L-inducing medium (see "Experimental Procedures") either in the presence or in the absence of 0.2% (NH,),SO,. In the respective experiments, 0.5% glycerol or 0.5% citrate instead of 0.5% glucose were used. For comparison, activities obtained on a medium with 0.2% m-nicotine as the sole carbon source are given.