19F NMR studies with 2'-F-2'-deoxyarabinoflavoproteins.

Apoproteins of several flavoproteins were reconstituted with 2′-F-2′-deoxyarabinoflavins and studied by 19F NMR and absorption spectroscopy. Extensive protein-fluorine interactions were observed by large chemical shift changes on binding to the apoprotein of Old Yellow Enzyme (apoOYE) and apoflavodoxin, whereas binding to apoglucose oxidase and apo D-amino acid oxidase (apoDAAO) resulted in minimal interactions. Modification at the flavin 2′-position in OYE resulted in a substantial decrease in the binding affinity of the flavin, possibly from the disruption of two important hydrogen bonds to Pro-35 and Arg-243. 19F NMR studies of complexes of OYE with testosterone, cyclohexenone, and β-estradiol suggest that phenols and α,β-unsaturated ketones orient differently at the active site on binding. The two separate one-electron potentials for the EFlox/EFlsq and EFlsq/EFlred couples were different for the reconstituted OYE. With native enzyme, there is 15-20% thermodynamic stabilization of the anionic flavin semiquinone, while no detectable amount of semiquinone was observed with modified OYE. This change in potential was further substantiated by blue shifts for the maxima of the modified protein-phenol charge transfer complexes. In accordance with the crystal structure of the OYE-p-OH-benzaldehyde complex (Fox, K.M. & Karplus, P.A. (1994) Structure 2, 1089-1105), 19F NMR studies with the modified OYE-2,4-F2-phenol suggest strong interaction between the para-fluorine of the phenol and Tyr-375.

In order to understand the various mechanisms of flavin reactions, knowledge of the protein environment at the active centers of different flavoenzymes is necessary. Several physical and chemical methods are available for this purpose. Of these, replacement of the native flavin with appropriately modified flavins and high resolution NMR techniques constitute very effective and widely exploited tools. The former method in general provides information about the solvent accessibility of the modified position, hydrophobicity, and steric constraints of the binding site (1). In the latter technique, comparison of the chemical shifts of the protein-bound flavin with the free flavin as well as the oxidized and reduced flavoproteins provides direct information about alterations in the magnetic environment of the substituent fluorine and hence can provide information about the protein environment surrounding the flavin (2)(3)(4). Old Yellow Enzyme is the oldest of the flavoprotein family and, as isolated from brewer's bottom yeast, is a mixture of homodimers and a heterodimer arising from two separate yeast genes, with each monomeric unit containing one FMN (5,6). Due to the fact that the physiological role of this enzyme is yet to be determined, the structure and reactivity of this protein has been the subject of extensive studies in our laboratory. Recent x-ray crystallographic studies suggest an important role for the 2Ј-hydroxyl group of the FMN in binding to the apoprotein and also in forming the binding site for various ligands (7,8). This hydroxyl group is involved in two important hydrogen bonds with Pro-35 and with Arg-243, one of the positively charged residues around the flavin N1 (Fig. 1). It was thought that the disruption of these hydrogen bonds would affect the configuration of the net charge at the binding site and so it was of obvious interest to study how this would affect the flavin binding to apoprotein, binding of various ligands, and catalytic activity. A new flavin, 2Ј-fluoro-2Ј-deoxyarabinoriboflavin was designed, where the 2Ј-hydroxyl group of the ribityl side chain was replaced with a hydrogen and a fluorine was incorporated in place of the hydrogen at the same 2Ј-carbon (9,10). This flavin should be an ideal probe for ligand-binding studies and also provide information on the chemical environment of the flavin binding site in terms of 19 F chemical shifts. It was also thought that this flavin could be a good 19 F NMR probe for flavoproteins in general. The high sensitivity of 19 F, which is next only to 1 H, the wide range of chemical shifts, and the ease with which one can incorporate fluorine into biological molecules make 19 F NMR a very effective physical tool to study proteins (11)(12)(13). Another important feature is that the number of peaks for each fluorine correspond to the number of binding environments present in the protein. Since the fluorine is in the ribityl side chain of this flavin, the net electron environment (fluorine is highly electronegative) of the isoalloxazine ring of the flavin is unperturbed by its presence, unlike in the case of 8-F-flavins, where the benzene ring of the flavin experiences more positive character due to the presence of the electronegative fluorine (4). Another advantage of fluorine being on the hydrophilic ribityl side chain is that it interacts more with the protein; and changes in the fluorine chemical shifts are expected to be very sensitive to capture even minute changes in the protein conformation. In the case of 8-F-flavin, fluorine being on the hydrophobic benzene ring of the flavin, the differences in chemical shifts are not that pronounced for most of the proteins studied (4).
Preparation of Flavins-2Ј-F-arabinoflavin and 2Ј-deoxyriboflavin were synthesized and converted to the FAD and FMN levels as reported earlier (9).
Apoprotein Reconstitution-Reconstitution of the apoproteins with 2Ј-fluoroflavin and 2Ј-deoxyflavin was accomplished by mixing an approximately 1.5-fold excess of flavin with apoprotein and then incubating on ice (from 30 min to overnight, depending on the protein). Excess flavin was removed either by gel filtration over G-25 (Pharmacia Biotech Inc.) or in a Centricon microconcentrator (Amicon). Unless otherwise indicated, all samples were prepared in 50 mM potassium phosphate buffer, pH 7.0.
NMR Sample Preparation-Protein samples for the NMR were concentrated by means of a Centricon microconcentrator to 150 -200 l and were diluted to 400 l with 50 l of D 2 O and for the rest of the volume with the buffer.
Reduced flavin and protein samples were prepared by the addition of an excess of reducing agent to an anaerobic sample in the NMR tube. The tubes with samples were flushed with argon (space over the sample) prior to the addition of the reductant in buffer. The sample became bleached within a few minutes and remained reduced for 24 -36 h. Argon flushing was continued during and after the addition. The tube was then closed with a standard plastic cap.
Glucose oxidase was reduced with D-glucose. D-Amino acid oxidase was reduced with D-alanine. Sodium dithionite was used to reduce 2Ј-F-FMN-OYE. 1 2Ј-F-FMN-flavodoxin was reduced by irradiation with light in the presence of EDTA and dithionite.
Ligand Binding Studies-Binding of various ligands to 2Ј-F-FMN-OYE was studied by titrating with standard ligand solutions and by measuring the changes in the UV-visible absorption spectrum of the protein-bound flavin. Dissociation constants were calculated from the data obtained from these experiments. Either a Cary model 219 doublebeam scanning spectrophotometer or a Hewlett-Packard model HP 8452A diode array spectrophotometer was employed for these measurements. The sample compartments in both instruments were temperature-controlled at 25°C NMR Acquisition Parameters-Fluorine NMR spectra were recorded with a 10-mm 19 F probe, on a General Electric GN 500 NMR, operating at 470 MHz. All the NMR measurements were carried out with proton decoupling, at 25°C (21). Fluorine chemical shifts were measured by using hexafluorobenzene as an external standard and are given in parts per million. Spectra were measured with 32,768 and 8192 data points over spectral widths from 20 to 45 kHz. The repetition time was 0.25-1 s with an 8-ms pulse (60°flip angle).
NMR Processing-With broad lines and low signal-to-noise ratio taken over a wide spectral width, the base line of an NMR spectrum is not flat (22,23). Correction routines (GEM routines IC, FB, and BF) were used to flatten the base line. Exponential line broadening was performed from 10 to 25 Hz.

Stability of Modified Flavins and Flavoproteins-
The stability of free as well as protein-bound flavins was studied. It was found that both the 2Ј-F-flavins and the 2Ј-deoxyflavins are highly stable over prolonged periods, showing no detectable changes over a period of a few months when stored at Ϫ20°C. However, when the proteins were reduced with sodium dithionite, a small amount of free fluoride was seen in the 19 F NMR spectra.
Flavodoxin-The binding of 2Ј-F-arabino-FMN to apoflavodoxin was followed both by fluorescence and by absorption spectra. At the end of the titration the flavin fluorescence was almost fully quenched with a residual fluorescence ϳ2.5% and with the max changed from 372 and 444 nm for free flavin to 378 and 446 nm for the protein-bound flavin. The extinction coefficient for 2Ј-fluoro FMN was determined as 11,300 M Ϫ1 cm Ϫ1 by standardizing the apoprotein with pure normal FMN. From the titration plots, the K d was determined as 1.2 Ϯ 0.2 ϫ 10 Ϫ6 M (ϳ4 ϫ 10 Ϫ10 M for native flavin), suggesting that the absence of the 2Ј-hydroxyl group weakened considerably the binding of the flavin to the protein. The 19 F NMR spectrum of the 2Ј-fluoro-FMN-reconstituted apoflavodoxin showed a single peak at 74.6 ppm, about a 10-ppm downfield shift from the free flavin resonance, suggesting extensive flavin protein interactions in this region of fluorine. In the clostridial flavodoxin, the 2Ј-OH group lies 2.69 Å away from Ala-55 and forms a strong hydrogen bond with the carbonyl of the alanine (24,25). The sodium dithionite-reduced modified protein has a peak at 64.55 ppm, almost the same as the free flavin chemical shift. This suggests that in the reduced form of the flavoprotein, fluorine interactions with the protein are minimal (Table I).
Old Yellow Enzyme-The binding of apoOYE with 2Ј-F-arabino-FMN was followed by measuring the flavin fluorescence. At the end point, the fluorescence was almost completely quenched, with a residual intensity ϳ3% that of the original. From the titration plots, the K d was determined as 4.6 Ϯ 0.3 ϫ 10 Ϫ6 M, which is 10 4 times weaker binding than the native flavin (ϳ10 10 M). The max for free flavin to protein-bound flavin shifted from 372 and 444 nm to 378 and 452 nm. The substantial decrease in binding affinity is quite understandable, since the flavin modification disrupts two hydrogen bonds that the 2Ј-hydroxyl group makes with Pro-35 and Arg-243 (Fig. 1).
The 19 F NMR spectrum of the oxidized enzyme has a peak at 80.75 ppm, a downfield shift of 15 ppm from the free flavin resonance (Fig. 2). This kind of large downfield shift generally means an increased electron density near the fluorine or a strong hydrogen bond formation involving fluorine. The 19 F resonance for the reduced enzyme, reduced with sodium dithionite, was recorded at 64.65 ppm, a large upfield shift of 16 ppm from the oxidized form toward the free flavin region (Fig. 2). This means that the strong fluorine-protein interactions in the oxidized form, which are indicated by the large downfield shift, are lost on reduction and that the fluorine signal is shifted upfield almost to the free flavin region. Interestingly, the crystal structure does not reveal significant changes in the flavinprotein hydrogen bonding interactions on reduction (8). However, reduction was accompanied by two interesting changes in the flavin conformation. In the reduced state, the flavin attains a butterfly conformation associated with puckering at N-5 and the FMN ribityl chain shifts from one staggered conformation to the other (8). As a consequence, the interactions between the fluorine and the protein in the reduced form might have been disrupted, resulting in the observed upfield shift to the free flavin region (Table I).
Ligand Binding Interactions with OYE-The oxidized form of the Old Yellow enzyme forms complexes with a wide variety of structurally diverse ligands such as simple anions, aromatic compounds, pyridine nucleotide derivatives, ␣,␤-unsaturated carbonyl compounds, including steroids (5,26). Of all the compounds, phenols are the most striking in that their binding results in the appearance of a new absorption band in the region of 500 -800 nm accompanied by strong perturbation in the absorption bands at 462 and 380 nm, these being shifted as much as 20 nm to lower wavelengths and reduced in amplitude by as much as 40% (16,17). The effect of the present chemical modification of the flavin on the binding of these ligands was studied by means of the changes in spectral properties and 19 F NMR measurements.
The native enzyme binds to p-chlorophenol with a K d of ϳ1 M and results in a new peak with a maximum at 645 nm. Many lines of evidence indicate that this long wavelength absorbance represents charge transfer interaction wherein the phenolate is the donor and the flavin the acceptor (8,16,17,27). When the 2Ј-F-FMN-reconstituted OYE is titrated with p-chlorophenol, it binds with a K d of 20 M, and the maximum of the long wavelength peak due to charge transfer interaction is at 610 nm, shifted 35 nm down from that with native enzyme (Table II). This shows that the present modification has affected both the binding of the phenol and the resulting spectral properties. This could reasonably be ascribed to a change in the configuration of the charge around the flavin, since an important hydrogen bond, which holds Arg-243 with the 2Ј-hydroxyl group of the native FMN is now nonexistent. To expand this observation, the binding of various structurally diverse phenols was studied. 2,4-Difluorophenol binds tightly to native enzyme with a K d of about 0.5 M and results in a long wavelength band with a maximum at 640 nm. The 19 F NMR spectra of this complex with native enzyme showed two sets of signals for the ligand and were assigned for the two isozymes of the enzyme (4). We decided to study this ligand with the modified flavoprotein, since the 19 F resonances of the complex from the ligand and the flavin might reiterate this fact. The titration plots suggested binding with a K d of 2 M, and the maximum of the long wavelength band was observed at 604 nm, shifted by 36 nm toward lower wavelength ( Fig. 3; Table II). This is consistent with the result we obtained with p-Cl-phenol.
The 19 F NMR spectrum of the free 2,4-difluorophenol has two resonances for the two fluorine atoms at 44. 15 Table I). Of these the two at 44.17 and 32.1 ppm appear to be due to the unbound ligand. The peak heights of these two signals, in comparison with the resonances from the bound ligand, suggest fast association and dissociation of the ligand with the protein. The bound ligand has two sets of peaks at 34.35, 32.95, 33.45, and 30.95 ppm, which is in accordance with the known fact that the enzyme from brewers' yeast is a mixture of two isozymes (6). Interestingly, the fluorine from the flavin shows only one peak at 79.86 ppm, about a 1-ppm shift from that of the ligand-free protein. This indicates that the binding of 2,4-difluorophenol does not result in any major rearrangement in the protein conformation. From the resonances of the bound ligand, it can be seen that the peak from the fluorine that is in the paraposition suffers a large upfield shift of ϳ10 ppm, while the one from the ortho-position is changed very little, by only 0.5 ppm. This can be explained by considering the x-ray structure of the p-OH-benzaldehyde-bound OYE (Fig. 1), which was recently determined by Fox and Karplus (8). From the structure it can be reasoned that this relatively large upfield shift is the result of hydrogen bonding interaction of the fluorine in the paraposition with the hydrogen of the -OH group of Tyr-375.
To further substantiate the interaction of the fluorine in the para-position with Tyr-375, we have recorded the 19 F NMR spectra for the complexes of the modified protein with p-and o-F-phenols. It was found that the 2Ј-F-FMN-OYE binds with o-F-phenol at pH 7, and the maximum of the long range band was observed at 554 nm. This is a shift toward lower wavelength region by 56 nm from the native protein band with a maximum at 610 nm. The 19 F NMR spectrum of this complex shows two signals, which are very broad. The signal from the flavin was seen at 80.1 ppm as a broad peak. The signal from the o-fluorine was seen again as a broad signal with multiple splitting at 31.3 ppm. This chemical shift is as expected for the o-fluorine and is just 1 ppm upfield from the free ligand signal. The maximum for the long wavelength band for the p-F-phenol complex was located at 600 nm compared with 660 nm for the native protein. The 19 F NMR spectrum of the 2Ј-F-FMN-OYE complex with p-F-phenol showed three signals. A broad signal from the flavin was seen at 80.5 ppm, a 1-ppm upfield shift from the ligand-free protein. The signal from the protein-bound ligand appeared at 32.6 ppm, a 12-ppm upfield shift from the free ligand, which can be attributed to interaction with Tyr-375 (Table I). In addition to these two signals, a very strong peak appeared for the unbound ligand, suggesting weak binding of p-F-phenol to the modified protein.
It has been shown that in brewers' yeast, p-OH-benzaldehyde is a natural ligand for OYE, and this complex is responsible for the major green form of the protein (28). This ligand binds very tightly to the native protein with a K d of 0.1 M with the maximum of the charge transfer band located at 575 nm. In light of our observations with the above two phenols, it was interesting to study the binding of this phenol to 2Ј-F-FMN-OYE. Interestingly, the chemical modification at the 2Ј-position has absolutely no effect on the binding of p-OH-benzaldehyde, apart from the usual shift of the wavelength band to shorter wavelengths (Table II) This was found to bind as tightly to the reconstituted protein as to the native protein, with a K d of 0.05 M (Table II). This can be very well explained by the x-ray structure (Fig. 1), which shows that the aldehyde carbonyl of the ligand hydrogen bonds to the -OH of Tyr-375, providing an additional binding point for this ligand.
The 19 F NMR spectrum of the 2Ј-F-FMN-enzyme has one resonance at 80.7 ppm for the complex of p-OH-benzaldehyde with the reconstituted protein. This is only a 0.5-ppm downfield shift from the ligand-free protein, showing no significant alteration in the magnetic environment of the fluorine and consistent with no major rearrangement in the flavin-protein interactions on binding of this ligand.
It has recently been found that the native enzyme binds very tightly to various steroid molecules (26,29). We chose two structurally different steroids for our studies with the reconstituted protein. One was testosterone, which has an ␣,␤-unsaturated moiety. This was selected because of the fact that the native protein reduces various ␣,␤-unsaturated compounds (6). The other was ␤-estradiol, which has a phenolic ring. It was found that testosterone binds tightly to the native protein with a K d of 1.6 M without forming any charge transfer complex. Interestingly, when the 2Ј-F-FMN-reconstituted protein was titrated with testosterone, it was found to bind tightly also with a K d of 6 M. Apparently, the chemical modification of the flavin has little effect on the binding of this ligand and was accompanied by a large perturbation in the absorption spectrum, as in the case of the native protein. The 19 F NMR of this complex has a resonance at 86.2 ppm, a large downfield shift of 5.5 ppm from the ligand-free protein (Fig. 2, Table I).
Native protein also binds tightly to ␤-estradiol with a K d of Ͻ1 M and forms a charge transfer complex with a maximum located at 670 nm. It was found from the titration that this ligand binds only weakly to the 2Ј-F-FMN-reconstituted protein with a weak charge transfer transition. The maximum of this long wavelength band was located at 622 nm, a shift toward lower wavelengths of 48 nm, consistent with the results with other phenols.
The 19 F NMR spectrum of the complex showed two resonances. The strong peak at 80.7 ppm, from the ligand-free protein, is consistent with a weak binding of ␤-estradiol to the modified protein. The ligand-bound protein has a peak at 83.7 ppm, an ϳ3-ppm downfield shift, suggesting the possibility of some conformational rearrangements in the protein as a result  of binding of the steroid molecule (Fig. 2, Table I).
It was found that cyclohexenone is both an electron donor and acceptor for OYE with phenol and cyclohexanone as final products (26). The 19 F NMR spectrum was recorded for the F-FMN-OYE in the presence of 5 equivalents of cyclohexenone. Interestingly a single fluorine resonance was recorded at 85.3 ppm with a downfield shift of about 5 ppm. This is of the magnitude of the shift observed with the large molecule like testosterone which has also an ␣,␤-unsaturated carbonyl moiety. Considering the fact that the binding of phenols perturbed the NMR spectrum negligibly, this might suggest that the protein binds phenolic and ␣,␤-unsaturated carbonyl ligands in a different orientation. The absorption spectrum of the complex showed the formation of a charge transfer complex with the maximum located at 566 nm, due to the formation of phenol.
Determination of Redox Potential-The redox potential of 2Ј-F-FMN-OYE was determined by titration under anaerobic conditions with acetyl pyridine NADH and pyridine aldehyde NADH as described previously for native enzyme (30). Within experimental error the E m 7 value for the overall two-electron potential, EFl ox /EFl red , was the same as that measured for native enzyme (Ϫ230 Ϯ 10 mV; results not shown). What is different, however, are the two separate one-electron potentials, EFl ox /EFl sq and EFl sq /EFl red . With native enzyme, there is 15-20% thermodynamic stabilization of the anionic flavin semiquinone, in agreement with measured values for the EFl ox / EFl sq and EFl sq /EFl red potentials of Ϫ245 mV and Ϫ215 mV, respectively (30). With 2Ј-F-FMN-OYE, no detectable amount of thermodynamically stable semiquinone could be observed. For example, during a xanthine/xanthine oxidase-catalyzed reduction of the enzyme in the presence of benzyl viologen as mediator (31), no semiquinone intermediate could be detected, although the 15-20% stabilization of semiquinone with native enzyme could be measured readily under the same conditions. Thus, the first electron potential, EFl ox /EFl sq must be significantly lower than with native enzyme. With a limit of detection of 1% semiquinone, it can be calculated that the potential of the EFl ox /EFl sq couple would be 60 mV lower than that of the midpoint potential for the EFl ox /EFl red couple, i.e.. Ϫ290 mV. This value can be taken, therefore, as an approximate upper limit value for the first one-electron potential, and it is nicely consistent with the relationship previously found between the energy of the charge transfer transition with p-chlorophenol and the one-electron potentials of a series of artificial flavins bound to native Old Yellow Enzyme (30).
NMR and Ligand Binding Studies with the Individual Isozymes of OYE-OYE from brewers' yeast can be separated into three fractions by fast protein liquid chromatography: two homodimers and one heterodimer (6). Interestingly, the 2Ј-F-FMN-reconstituted wild type OYE, unlike the 8-F-FMN-reconstituted protein (4), showed only one fluorine resonance, suggesting similar chemical environments in this region of the active sites of the isozymes. To get more insight into this observation, the two homodimers, fractions 1 and 3 of the OYE, were separated by fast protein liquid chromatography and subjected to further study. Apoproteins of these fractions were prepared and reconstituted with 2Ј-F-arabino-FMN. To find out if the flavin modification had affected the two isozymes differently, ligand binding and the 19 F NMR of these reconstituted isozymes were studied. The 19 F NMR of the 2Ј-F-arabino-FMN-reconstituted OYE fractions 1 and 3 showed one resonance each at 80.3 and 80.1 ppm, suggesting a similar chemical environment around the ribityl side chain. Titration of these two forms with 2,4-difluorophenol gave the K d value of 3-4 M for fraction 1 and 1.5-2 M for fraction 3. This again shows that the present chemical modification has a very similar effect on both the forms. The complex of fraction 1 has the resonances for the bound ligand at 32.7 and 30.8 ppm, whereas for fraction 3 they were recorded at 34.3 and 33.4 ppm (Fig. 4), suggesting that the interaction of ligand fluorines in both complexes are different. It can be concluded from the 19 F NMR results with free and ligand-bound isozymes of OYE that noticeable differences in the chemical environments exist at the ligand binding site, whereas they are very similar around the ribityl side chain. When the reconstituted fractions were titrated with testosterone, the dissociation constants showed that both the isozymes bind this ligand with almost the same affinity.
Studies with 2Ј-Deoxy-FMN-In continuation of the above results with 2Ј-F-FMN-reconstituted OYE, it was thought appropriate to extend our studies to 2Ј-deoxy-FMN-reconstituted proteins. This flavin should act as a control for possible structural modifications induced by the fluorine substitution. The extinction coefficient for the 2Ј-deoxy-FMN was determined as 12,600 M Ϫ1 cm Ϫ1 by titration of the flavin with apoflavodoxin, which had been standardized with pure normal FMN.
2Ј This form of the protein also forms a charge transfer complex with p-Cl-phenol with a maximum at 616 nm, shifted to lower wavelength by 32 nm, similar to what was observed with 2Ј-F-arabino-FMN-protein.
Activities of the Reconstituted Proteins-Old Yellow enzyme has NADPH oxidase activity but also reduces various ␣,␤unsaturated ketones including quinones (6,26). The catalytic activity for NADPH oxidation in the presence and absence of cyclohexenone and 1,2-cyclohexanedione was determined for modified proteins and found to be 10 -15% that of native enzyme. The results are documented in Table III.
The following conclusions can be made. 1) The binding affinity of the apoOYE with 2Ј-F-arabino-FMN was decreased substantially compared with that of normal FMN, probably due at least in part to disruption of two hydrogen bonds that the 2Ј-hydroxyl group makes with Pro-35 and Arg-243 (8).
2) The downfield shift of ϳ15 ppm of the free flavin upon binding to apoOYE suggests strong protein-flavin interactions in this region. Although puckering at N-5 resulted in butterfly bending of the reduced flavin, the crystal structure showed no qualitative changes in the flavin-protein hydrogen bonding interactions (8). However, the CЈ-5 atom of the FMN ribityl group moves on reduction by 1.6 Å to shift from one staggered conformation to another (8). Interestingly, the reduced 2Ј-F-arabino-FMN-OYE showed an upfield shift of 16 ppm, almost back to the free flavin chemical shift. This suggests that in the reduced enzyme the fluorine might have lost the contact with protein as a result of the new conformation of the ribityl side chain.
3) With the 8-F-FMN-OYE, the 19 F NMR spectrum for many of the complexes with various ligands showed two clearly distinguishable peaks (4). In contrast, with 2Ј-F-FMN-OYE the spectra of the various complexes showed clear single peaks. This suggests a similar chemical environment around the ribityl side chain for both isozymes.
4) The fact that the chemical shifts of the fluorine from the protein-bound flavin showed no or negligible changes on complexation with phenols is consistent with the fact that their binding does not disrupt any interactions between the protein and flavin (8).
5) The fact that the fluorine in the para-position of 2,4difluorophenol and 4-F-phenol experiences a substantial upfield shift of ϳ10 -11 ppm suggests that the para-position of these ligands interact with Tyr-375. Since Tyr-375 lies in the active site and interacts strongly with the bound ligand, it will be of interest to further study its possible role in catalysis.
6) The fact that the binding of testosterone, which has an ␣,␤-unsaturated ketonic moiety was unaffected, while the binding of ␤-estradiol with a phenolic moiety suffered weak binding suggests that these two molecules orient differently at the active site on binding. Large chemical shifts observed for the cyclohexenone complex compared with the minor shifts on the binding of phenols are in accordance with this conclusion. These results are in support of the recent conclusion that the binding modes of cyclohexenone and 3-oxadecalin-4-ene are different in the oxidative and reductive halves of the dismutation reaction catalyzed by OYE (26) Glucose Oxidase-2Ј-F-arabino-FAD binds very tightly to apo-glucose oxidase with quenching of the fluorescence and changes in the max from 372 and 444 nm to 376 and 454 nm.
The 19 F NMR spectrum of the 2Ј-fluoro-FAD-reconstituted protein showed a single peak at 66.38 ppm, an ϳ1-ppm downfield shift from that of the free flavin, suggesting minor interaction of the fluorine with the protein. The glucose-reduced protein has a signal at 70.40 ppm, a 4-ppm downfield shift from that of the oxidized species. This is in contrast to the FMNbased flavodoxin and OYE, where large upfield shifts were observed upon reduction of the proteins (Tables I and II).
From enzyme-monitored turnover experiments (results not shown), the turnover number for the reconstituted enzyme was determined to be ϳ5000 min Ϫ1 at pH 7, 25°C, which is 1 ⁄3 that of the native enzyme (ϳ14,000 min Ϫ1 ; Ref. 32).
D-Amino Acid Oxidase-The absorption spectrum maxima for the 2Ј-F-arabino-FAD shifted from 372 and 444 nm to 362 and 448 nm upon binding to apoDAAO. The 19 F NMR spectrum of the oxidized enzyme had a single resonance at 65.65 ppm, which is Ͻ1 ppm different from that of the free flavin chemical shift, suggesting minimal interaction of fluorine with the protein. D-alanine-reduced protein showed a single resonance at 64.95 ppm, again consistent with no major conformational changes upon reduction. Oxidized D-amino acid oxidase binds to benzoate, and as a result the absorption spectrum becomes highly resolved (33,34). Similar spectral changes were observed in the case of modified protein. The 19 F NMR spectrum of the complex had a resonance at 65.35 ppm, suggesting no significant changes in the protein-flavin interaction around the 2Ј-position of the ribityl side chain. The 19 F NMR spectrum of the oxidized enzyme with sulfite also showed no major change in the fluorine chemical shift. The standard assays were performed aerobically in 1 ml of 0.1 M KP i buffer containing 150 M ␤-NADPH and 2.3 ϫ 10 Ϫ4 M O 2 as at pH 7.0 and 25°C. A known amount of OYE sample was added, and the turnover numbers were calculated from the enzyme-induced rate of ␤-NADPH oxidation, measured by following the decrease in absorbance of the reduced pyridine nucleotide at 340 nm.

Observed turnover number
In the presence of Cyclohex-2en-1-one In