Model systems for flavoenzyme activity: One- and two-electron reduction of flavins in aprotic hydrophobic environments

Article abstract of DOI:10.1021/ja963704a

The processes occurring during the electrochemically irreversible reduction of flavins in aprotic organic medium have been the cause of considerable controversy. To provide insight into these events we have investigated the reduction of flavins alkylated at N(3). We have demonstrated that the reversible reduction of N(3)-methylated flavin can be converted to the irreversible reduction observed with unalkylated flavins through addition of the weak external proton donor phthalimide. Proton transfer to the flavin radical anion generated at the electrode leads to partial formation of the neutral flavin radical, which is instantaneously further reduced to the fully reduced flavin anion. This mirrors the reduction of unalkylated flavins, where the proton source is the imide proton at N(3) of fully oxidized flavin in bulk solution. Simultaneous electrochemistry and EPR experiments confirm the disappearance of electrogenerated flavin radical anion both methylated and non-methylated at N(3), upon addition of phthalimide. UV/vis spectroelectrochemistry likewise confirmed the generation of the radical anion in the absence of proton donors and fully reduced flavin in the presence of proton donors.

Full text of DOI:10.1021/ja963704a

J. Am. Chem. Soc. 1997, 119, 887-892
887
Model Systems for Flavoenzyme Activity: One- and
Two-Electron Reduction of Flavins in Aprotic Hydrophobic
Environments
Angelika Niemz, Jason Imbriglio, and Vincent M. Rotello*
Contribution from the Department of Chemistry, UniVersity of Massachusetts,
Amherst, Massachusetts 01003
ReceiVed October 23, 1996X
Abstract: The processes occurring during the electrochemically irreversible reduction of flavins in aprotic organic
medium have been the cause of considerable controversy. To provide insight into these events we have investigated
the reduction of flavins alkylated at N(3). We have demonstrated that the reversible reduction of N(3)-methylated
flavin can be converted to the irreversible reduction observed with unalkylated flavins through addition of the weak
external proton donor phthalimide. Proton transfer to the flavin radical anion generated at the electrode leads to
partial formation of the neutral flavin radical, which is instantaneously further reduced to the fully reduced flavin
anion. This mirrors the reduction of unalkylated flavins, where the proton source is the imide proton at N(3) of fully
oxidized flavin in bulk solution. Simultaneous electrochemistry and EPR experiments confirm the disappearance of
electrogenerated flavin radical anion both methylated and non-methylated at N(3), upon addition of phthalimide.
UV/vis spectroelectrochemistry likewise confirmed the generation of the radical anion in the absence of proton
donors and fully reduced flavin in the presence of proton donors.
The flavin cofactors FAD (flavin adenine dinucleotide) and
FMN (flavin mononucleotide) are involved in the catalysis of
a wide variety of biological redox reactions, including the
dehydrogenation of NAD(P)H, lipid esters, and D-amino acids,
the oxidation of amines to imines, the formation and cleavage
of disulfide bonds, the hydroxylation of aromatic substrates, and
the activation of molecular oxygen.1-4 Flavoenzymes mediate
electron transfer processes (e.g., in photosynthesis and oxidative
phosphorylation) and are involved in the regulation of neu-
rotransmitters and the detoxification of xenobiotics.5 Their
ability to engage in one- as well as two-electron transfer
reactions enables flavins to act as transformers between obligate
2e- donors (e.g., NADH) and obligate 1e- acceptors (e.g. heme
Fe).
Scheme 1a
Analogous to quinones, flavins possess three readily acces-
sible oxidation states (Scheme 1): the fully oxidized fla-
voquinone (Flox); the flavosemiquinone radical, in either the
anionic red (Flrad-) or neutral blue (FlradH) form; and the two-
electron-reduced flavohydroquinone, which can also exist in
either anionic (FlredH-) or neutral (FlredH2) form.
The redox properties of flavins in aqueous6-10 and organic10-13
media have been studied by potentiometric titration, dc, and ac
X Abstract published in AdVance ACS Abstracts, January 15, 1997.
(1) Walsh, C. Acc. Chem. Res. 1980, 13, 148-155.
(2) Ghisla, S.; Massey, V. Eur. J. Biochem. 1989, 181, 1-17.
(3) Bruice, T. C. Isr. J. Chem. 1984, 24, 54-61.
(4) Mueller, F. In Topics in Current Chemistry; Boschke, F. L., Ed.;
Springer-Verlag: Berlin, 1983; Vol. 108, pp 71-108.
(5) Ziegler, D. M. Trends Pharm. Sci. 1990, 321-324.
(6) Janik, B.; Elving, P. J. Chem. ReV. 1968, 68, 295-319.
(7) Michaelis, L.; Schwarzenbacher, G. J. Biol. Chem. 1938, 123, 527-
542.
a Lumiflavin: R ) CH3. Riboflavin: R ) CH2(CHOH)3CH2OH.
FMN: R ) CH2(CHOH)4-phosphate. FAD: R ) CH2(CHOH)4-
pyrophosphate-adenosine. Flavin 1: R ) CH2CH(CH3)2.
polarography, cyclic voltammetry (CV), controlled potential
coulometry, and simultaneous electrochemistry and EPR (SEEPR).
It is well established that in aqueous or protic organic solvents
flavins are reduced by a two-electron reduction directly to the
flavohydroquinone. While this is relevant to flavoenzymes
catalyzing solely two-electron processes, a great number of
flavoenzymes are known to operate via the flavin radical,14
enabling them to act as switches between one- and two-electron
processes. An aqueous medium therefore is not appropriate for
(8) Draper, R. D.; Ingaham, L. L. Arch. Biochem. Biophys. 1968, 125,
802-808.
(9) Hartley, A. M.; Wilson, G. S. Anal. Chem. 1966, 38, 681-687.
(10) Male, R.; Samotowka, M. A.; Allendoerfer, R. D. Electroanalysis
1989, 1, 333-339.
(11) Tatwawadi, S.; Santhanam, K. S. V.; Bard, A. J. Electroanal. Chem.
1968, 17, 411-420.
(12) Sawyer, D. T.; Komai, R. Y.; McCreery, R. L. Experientia, Suppl.
1971, 18, 563-573.
(13) Sasaki, K.; Kitani, A.; Kunai, A.; Miyake, H. Bull. Chem. Soc. Jpn.
1980, 53, 3424-3429.
S0002-7863(96)03704-3 CCC: $14.00 © 1997 American Chemical Society

888 J. Am. Chem. Soc., Vol. 119, No. 5, 1997
Niemz et al.
is whether the first redox couple represents a two-electron-
reduction step followed by comproportionation of the flavohy-
droquinone dianion and unreacted flavoquinone to form the
radical anion.10 A second possibility is that a one-electron
reduction occurs directly to the radical anion.11,12 It is also
possible that the radical species could arise from a third,
unknown, process. To resolve these uncertainties regarding the
redox properties of flavin models in hydrophobic media, we
have undertaken a combined electrochemical and spectroelec-
trochemical investigation of flavin reduction. We report here
our mechanistic interpretation of the reduction of flavins in
aprotic media.
Experimental Section
Materials and General Methods. Solutions were prepared using
reagent grade CH2Cl2 dried via distillation over CaH2. Tetrabutylam-
monium perchlorate (TBAP, obtained from SACHEM, electrometric
grade) was recrystallized twice from water and dried for several days
under high vacuum. Other chemicals were reagent grade, obtained from
Aldrich Chem. Co., Acros Organics, Eastman Organic Chemicals, and
EM Science and were used without further purification. 6-Chlorouracil
was synthesized according to a literature procedure.6 Reactions were
carried out in flame-dried flasks under argon unless otherwise stated.
Melting points were determined on a Mel-Temp apparatus and are
uncorrected. Infrared spectra were recorded on a Perkin-Elmer Model
1420 IR spectrometer. 1H-NMR spectra were recorded on a Bruker
AC200 (200 MHz) spectrometer. Elemental analyses were performed
by the University of Massachusetts (Amherst) Microanalysis Service.
Cyclic Voltammetry. All electrochemical and spectroelectrochemi-
cal experiments were carried out on an IBM EC/250 voltammetric
analyzer equipped with a Houston Instruments HR100 xy-recorder. In
the CV studies, a 1.6 mm platinum disk and a coiled platinum wire
electrode were utilized as working and auxiliary electrodes, respectively.
A silver wire pseudo reference electrode was used, and potentials are
referenced versus the ferrocene/ferrocenium couple. In all experiments
the sweep rate was 300 mV/s unless otherwise stated. The electro-
chemical cells were dried at 125 °C for several hours and stored in a
desiccator prior to use. Solutions of N(3)-methyl-N(10)-isobutyl flavin
(10-3 M in CH2Cl2, 0.1 M TBAP) were degassed by bubbling argon
(presaturated with CH2Cl2) through them for 4 min. The proton sources,
acetanilide, phthalimide, and acetic acid, were prepared as stock
solutions (0.1 M in CH2Cl2). A total of 3 equiv of the respective proton
source was added to the flavin solution in steps of 0.2 equiv, with cyclic
voltammograms observed after each addition.
Figure 1. Cyclic voltammogram showing the first reduction wave
(wave I) and coupled oxidation waves (waves II and II) of N(10)-
isobutyl flavin 1 and N(3)-methyl-N(10)-isobutyl flavin 2 (10-3 M) in
CH2Cl2, TBAP (10-1 M), scan rate 500 mV/s.
modeling these systems. This fact is supported by crystal
structures of flavoenzymes,15-17 which reveal that the cofactor
is generally buried inside the protein matrix in a hydrophobic
environment, allowing only controlled access to protons.
Electrochemical studies in an aprotic solvent of low dielectric
constant such as methylene chloride therefore provide a better
analogy for enzyme-bound flavin and are a valuable tool in the
investigation of the complex redox behavior of the cofactor.
Previous voltammetric studies11,12 of flavin reduction in DMSO,
DMF, and acetonitrile have indicated the presence of two
resolved reduction and three oxidation waves. As shown in
Figure 1, the first reduction wave of the N(10)-isobutyl flavin
1 used in our studies was found to be coupled to two oxidation
waves, in agreement with previously reported cyclic voltam-
mograms.11,12
The interpretation of these observed electrochemical events
is up to date controversial. It has been postulated11 that the
flavin radical anion is unstable, decomposing to an unidentified
species A and a radical B•-. Oxidation of B•- and A would
then give rise to wave II and wave III, respectively. It was
noted that the CV of riboflavin in DMSO is significantly altered
by the addition of the weak proton donor hydroquinone, yet no
explanation for the changes was given. In a different study,12
protonation and one- as well as two-electron reduction was
proposed to be responsible for the appearance of new waves in
the CV of flavins upon addition of the strong acid HClO4, yet
no explanation was given as to why two oxidation waves are
coupled to the first reduction wave in neutral solution. In recent
research10 SEEPR has been used to show that a radical species
is formed by bulk electrolysis at potentials slightly more negative
than the first reduction wave (wave I). The mechanism of
formation of this species, however, is uncertain. The question
Simultaneous Electrochemistry and EPR. Because of the lossy
nature of the samples and because it was desired to minimize
perturbation of the microwave field by the working electrode, SEEPR
experiments were carried out in a quartz flat cell.19 A second glass
part containing three threaded joints sealed with Teflon ferrules to hold
the electrodes and a septum-capped ground-glass joint for degassing
and sample injection was connected to the top of the cell. The working
electrode, a copper gauze electrode amalgamated by immersion into
an aqueous saturated solution of Hg2(NO3)2, was inserted into the flat
part of the cell. The Ag-wire pseudo reference electrode was positioned
directly above the working electrode in order to minimize the iR-drop,
and the auxiliary electrode, a platinum wire spiral of large surface area,
occupied the solvent reservoir above the flat section. The electrode
leads were insulated via Teflon heat shrink tubing. After each
experiment the working electrode was cleaned in dilute HNO3 and
reamalgamated.
EPR spectra were recorded on an IBM ESP 300 X-band spectrometer
equipped with a TE104 dual cavity. Flavin solutions (10-3 M in CH2-
Cl2, 0.1 M TBAP) were degassed by bubbling argon through them for
5 min and then injected into the cell, which was previously flushed
with argon. The cell was mounted within the spectrometer using
(14) Edmondson, D. E.; Tollin, G. In Topics in Current Chemistry;
Boschke, F. L., Ed.; Springer-Verlag: Berlin, 1983; Vol. 108, pp 109-
138.
(15) Schreuder, H. A.; Mattevi, A.; Obmolova, G.; Kalk, K. H.; Hol,
W. G. H. Biochemistry 1994, 33, 10161-10170.
(16) Fukuyama, K.; Matsubara, H.; Rogers, L. J. J. Mol. Biol. 1992,
225, 775-789.
(18) Ishikawa, I.; Itoh, T.; Melik-Ohanjana, R.; Takayanagi, H.; Mizuno,
Y.; Ogura, H.; Kawahara, N. Heterocycles 1990, 31, 1641-1646.
(19) Compton, R. G.; Waller, A. M. In Spectroelectrochemistry. Theory
and Practice; Gale, R. J., Ed.; Plenum Press: New York, 1988; pp 349-
398.
(17) Nishida, H.; Inaka, K.; Miki, K. FEBS Lett. 1995, 361, 97-100.

Model Systems for FlaVoenzyme ActiVity
J. Am. Chem. Soc., Vol. 119, No. 5, 1997 889
custom-manufactured cell holders, which allow for precise alignment
of the cell within the cavity in order to maximize the Q-factor. Bulk
electrolysis was carried out simultaneously with signal acquisition (36
scans, 25 kHz field modulation, modulation amplitude 0.098 G).
UV/Vis Spectroelectrochemistry. The cell used in this study was
built analogously to the apparatus developed by Salbeck,20 and has the
advantage of being solvent resistant, and capable of carrying out
electrolysis under inert gas conditions. The working electrode, an
optically transparent thin layer electrode (OTTLE), consists of a gold
minigrid (100 lines/in., approximately 70% transparent) sandwiched
between two glass slides separated by Teflon spacers (100 mm). A
silver wire electrode and a Ni/Au-plated brass disk electrode served as
the pseudo reference electrode and the auxiliary electrode, respectively.
The main cell body was a cylindrical Pyrex glass cuvette 3 cm in
diameter with a ground-glass joint at the top. The rest of the cell
consists of a Teflon stopper with holes for degassing, a septum to add
solution, and electrode connectors.
Spectra were recorded on a Shimadzu Model 260 UV/vis spectrom-
eter. A reproducible orientation of the cell was achieved by lines
engraved on the cell holder, glass cuvette, and Teflon stopper. The
cell was fixed by set screws, purged with argon presaturated with CH2-
Cl2, and filled with solution so that the bottom of the OTTLE was just
immersed. To account for the absorption characteristics of the cell, a
base line spectrum of the cell plus electrolyte (CH2Cl2, 0.1 M TBAP)
was recorded and subtracted from the other spectra. The potential was
stepped up incrementally and the cell allowed to equilibrate (indicated
by current decay to a constant value) after each step before spectra
were recorded. The Ag wire pseudo reference electrode was calibrated
vs ferrocene at the end of the experiment.
Synthesis of Flavin. N-Isopropyl-3,4-dimethylaniline (3). To a
mixture of 3,4-dimethylaniline (9.69 g, 80 mmol) and isobutyraldehyde
(7.27 mL, 5.77 g, 80 mmol) at 0 °C was added titanium isopropoxide7
(29.76 mL, 28.23 g, 100 mmol) over 30 min. The solution was allowed
to warm to room temperature and stirred for 2 h. A solution of sodium
cyanoborohydride (3.37 g, 53.6 mmol) in absolute EtOH (80 mL) was
then added, followed by stirring for 12 h. Addition of H2O (16 mL)
yielded a thick white suspension, which was stirred open to the
atmosphere until HCN evolution ceased. The suspension was then
filtered and washed with EtOH, and the filtrate was concentrated using
a rotary evaporator. The residue was dissolved in diethyl ether (60
mL), washed with saturated aqueous NaHCO3, water, and saturated
aqueous NaCl (∼10 mL each), and then dried over Na2SO4. The
solvent was removed in Vacuo. Vacuum distillation yielded 9.59 g
(68%) of the alkylated aniline as a clear liquid: bp 96-97 °C (1.5
Torr); 1H NMR (200 MHz, CDCl3) δ 6.92 (1H, d, J ) 8.3 Hz), 6.44-
6.34 (2H, m), 2.89 (2H, d, J ) 6.5 Hz), 2.18 (3H, s), 2.14 (3H, s),
1.87 (1H, m, J ) 6.5 Hz), 0.96 (6H, d, J ) 6.5 Hz); IR (film) 3410,
3010-2880, 1620, 1580, 1505, 1465, 1385, 1315, 1320, 1260, 1215,
1170, 1150, 1120 cm-1. Anal. Calcd for C12H19N: C, 81.36; H, 10.73;
N, 7.91. Found: C, 81.34; H, 10.89; N, 7.95.
8.09 (1H, s), 7.83 (1H, s), 4.43 (2H, bd, J ) 6.5 Hz), 2.46 (3H, s),
2.38 (3H, s), 2.30 (1H, m, J ) 6.5 Hz), 0.96 (6H, d, J ) 6.5 Hz); IR
(KBr) 3200-2800, 2360, 1700, 1655, 1535, 1437, 1400, 1240, 1225
cm-1; the material was used without further purification.
N(10)-Isobutyl flavin 1. To a suspension of N(10)-isobutyl flavin
N(5)-oxide 5 (942.6 mg, 3.0 mmol) in H2O/EtOH (1:1, 20 mL) was
added Na2S2O4 (1.045 g, 6.0 mmol), and the mixture was stirred open
to the atmosphere until the initial dark green color changed to bright
yellow (several hours). The crude product was filtered, washed with
H2O, dissolved in CH2Cl2, and filtered again. After drying over Na2-
SO4, the solvent was removed in Vacuo. Recrystallization from EtOH/
CH2Cl2 yielded 729.7 mg (82 %) of bright yellow needles: mp 302-
305 °C dec; 1H NMR (200 MHz, CDCl3) δ 8.43 (1H, s), 8.06 (1H, s),
7.41 (1H, s), 4.63 (2H, bs), 2.56 (3H, s), 2.45 (3H, s), 2.53-2.35 (1H,
m, J ) 6.9 Hz), 1.05 (6H, d, J ) 6.9 Hz); IR (KBr) 3200-2800, 2360,
1715, 1655, 1580, 1540, 1505, 1450, 1400, 1250 cm-1. Anal. Calcd
for C16H18N4O2: C, 64.44; H, 6.04, N, 18.78. Found: C, 64.15; H,
5.59; N, 18.70.
N(3)-Methyl-N(10)-isobutyl flavin 2. To a suspension of N(10)-
isobutyl flavin 1 (59.6 mg, 0.2 mmol) and Na2CO3 (212 mg, 2.0 mmol)
in dry dimethylformamide (3 mL) was added methyl iodide (1.25 mL,
283.9 mg, 2.0 mmol), and the mixture was stirred at room temperature
for 12 h. After removal of the solvent via Kugelrohr distillation under
vacuum, the residue was suspended in 10 mL of H2O, precipitating
the product as yellow solid. The solution was then neutralized with 1
M HCl and filtered, and the crude product was washed with H2O.
Recrystallization from EtOH/H2O yielded 52 mg (82%) of N(3)-
1
methylated flavin as fine yellow needles: mp 239-241 °C; H NMR
(200 MHz, CDCl3) δ 8.08 (1H, s), 7.39 (1H, s), 4.64 (2H, bs), 3.52
(3H, s), 2.54 (3H, s), 2.44 (3H, s), 2.51-2.33 (1H, m, J ) 6.9 Hz),
1.04 (6H, d, J ) 6.9 Hz): IR (KBr) 2940, 1700, 1650, 1580, 1540,
1455, 1280, 1240 cm-1. Anal. Calcd for C17H20N4O2: C, 65.40; H,
6.41; N, 17.94. Found: C, 65.21; H, 6.21; N, 17.80.
Results and Discussion
To provide a flavin with sufficient solubility in nonpolar
media we have synthesized N(10)-isobutyl flavin 1, in which
the isobutyl group at N(10) causes an enhancement in solubility
of 2 orders of magnitude compared to lumiflavin (Scheme 2).24
Alkylation of flavin 1 with methyl iodide then provided the
N(10)-isobutyl-N(3)-methyl flavin 2 used in this research.
As previously mentioned, the CV of N(10)-isobutyl flavin 1
in CH2Cl2 exhibits both waves II and III during the oxidative
sweep (Figure 1). Surprisingly, the CV of N(10)-isobutyl-N(3)-
methyl flavin 2 in dry CH2Cl2 is electrochemically reversible,
showing only wave II.25 This finding suggests the involvement
of the imide proton at N(3) in the formation of the species that
is reoxidized during wave III, either via intramolecular tau-
tomerization of one-electron-reduced flavin or via intermolecular
proton transfer from flavoquinone in bulk solution to one-
electron-reduced flavosemiquinone in the electrochemical double
layer. To test these two hypotheses, we carried out a series of
CV titrations in which acetanilide, phthalimide, and acetic acid
were added to flavin 2.
6-(N-Isopropyl-3,4-dimethylanilino)uracil (4). A mixture of 6-chlo-
rouracil (2.19 g, 15 mmol) and N-isobutyl-3,4-dimethylaniline (5.32
g, 30 mmol) were heated at 150 °C for 24 h.8,9 After cooling to room
temperature, the pale yellow solid was suspended in EtOH/hexane (1:
2, 80 mL), precipitating the product as a white powder. The crude
product was filtered and washed with hexane. Recrystallization from
1:1 EtOH/EtOAc yielded 4.31 g (60%) of white crystals: mp 258 °C;
1H NMR (200 MHz, CDCL3) δ 7.95 (1H, bs), 7.26-7.21 (1H, m),
6.97-6.92 (3H, m), 4.94 (1H, bs), 3.36 (2H, d, J ) 6.9 Hz), 2.30 (s,
6H), 2.08 (1H, m, J ) 6.9 Hz), 0.93 (6H, d, J ) 6.9 Hz); IR (KBr)
3400-2800, 2360, 1720-1550 (broad), 1500, 1470, 1440, 1400, 1380,
1340, 1250, 1225, 1170, 1145, 1070 cm-1. Anal. Calcd for C16H21N3O2:
C, 66.91; H, 7.32; N, 14.63. Found: C, 66.95; H, 7.43; N, 14.58.
N(10)-Isobutyl flavin N(5)-Oxide 5. To a solution of 6-(N-
isopropyl-3,4-dimethylanilino)uracil (1.149 g, 4.0 mmol) in AcOH (16
mL) was added in portions NaNO2 (828 mg, 12 mmol) with cooling
on ice, providing an orange suspension. After stirring at room
temperature for 1.5 h, the product was filtered and washed with EtOH.
Trituration with EtOH yielded 936 mg (75%) of orange crystals: mp
As shown in Figure 2, addition of both phthalimide and acetic
acid renders the reduction of flavin 2 irreversible, causing the
appearance of wave III and a decrease in the peak current of
wave II. Addition of these proton donors therefore replicates
(21) Mattson, R. J.; Pham, K. M.; Leuck, D. J.; Cowen, K. A. J. Org.
Chem. 1990, 55, 2552-2554.
(22) Yoneda, F.; Yoshiharu, S.; Ichiba, M.; Shinomura, K. Chem. Pharm.
Bull. 1972, 20, 1832-1834.
(23) Yoneda, F.; Sakuma, Y.; Ichiba, M.; Shinomura, K. J. Am. Chem.
Soc. 1976, 98, 830-835.
(24) Breinlinger, E.; Niemz, A.; Rotello, V. M. J. Am. Chem. Soc. 1995,
117, 5379-5380.
(25) Previously published cyclic voltammograms of N(3)-methyl lumi-
flavin in DMSO, DMF, and acetonitrile exhibit both wave II and wave
III,12 which may be attributed to residual water content in these solvents,
or adsorbtion effects on the electrodes.
1
287-290 °C dec; H NMR (200 MHz, DMSO-d6) δ 11.03 (1H, bs),
(20) Salbeck, J. Anal. Chem. 1993, 65, 2169-2173.

890 J. Am. Chem. Soc., Vol. 119, No. 5, 1997
Niemz et al.
Scheme 2. Synthesis of N(10)-Isobutyl Flavin 1 and
N(10)-Isobutyl-N(3)-methyl flavin 2 Following the Procedure
of Yoneda et al.22,23
the unusual oxidative sweep observed in the CV of flavin 1.
Acetanilide, in contrast, has no effect of the CV of flavin 2,
even when added in 3-fold excess. Addition of a proton source
of suitable acidity to flavin 2 which is not capable of tautomer-
ization mimics the electrochemical behavior of flavin 1,
supporting the intermolecular proton transfer hypothesis (Scheme
3). While there is not a direct correlation between acidities in
differing media, the acidities of aqueous phthalimide and
acetanilide can be used as guidelines for estimating their proton
donor properties. Acetanilide (pKa ) 13.39)26 is not acidic
enough to protonate the flavosemiquinone radical anion; how-
ever, phthalimide (pKa ) 9.90) is a sufficiently strong proton
donor to effect this process. This finding agrees with the
reported pKa of N(3)-H of flavoquinone (pKa ) 10),4,27 which
acts as the proton donor in the case of nonalkylated flavin.
A second observation supporting the proposed proton transfer
mediated process is an increase in current response upon addition
of proton donor. The integrals of wave I in the CV’s of flavin
2 alone and of flavin 2 plus 3 equiv of phthalimide shows a
2-fold increase in the charge consumed during the reductive
sweep. The increase marks the transition from a one-electron
to a two-electron reduction. In the absence of a proton source
Figure 2. Change in the CV of N(10)-isobutyl-N(3)-methyl flavin 2
upon addition of proton donors. Solutions of acetanilide, phthalimide,
and acetic acid (10-1 M in CH2Cl2) were titrated into the solution of
N(3)-methyl-N(10)-isobutyl flavin (10-3 M in CH2Cl2, 0.1 M TBAP).
-
is less negative than the E° for the reduction of Flox to Flrad
,
FlradH is instantly reduced to FlredH-. This second one-electron
reduction can occur either at the electron (ece process) or via
disproportionation of FlradH with Flrad-, forming FlredH- and
Flox (disp1 process, assuming that the protonation of Flrad is
the rate-limiting step) (Scheme 3.)28
-
2-
the reduction of Flrad to the unstable Flred is a rather
unfavorable process with a more negative standard reduction
potential (E°) than the reduction of Flox to Flrad-. In the presence
of a weak acid, Flrad- is protonated at N(5), rendering the neutral
radical FlradH. Since the E° for the reduction of FlradH to FlredH-
The rate of the protonation step, first order in the concentra-
-
tion of proton donor, determines the ratio of Flrad to FlredH-
present at the electrode after the reductive sweep in cyclic
voltammetry, hence the gradual increase of wave III and
decrease of wave II during the titration of flavin 2 with external
(26) In Lange’s Handbook of Chemistry, 13th ed.; J. A. Dean, Ed.;
McGraw-Hill Book Company: New York, 1985; pp 5-18 ff.
(27) Ehrenberg, A.; Mueller, F.; Hemmerich, P. Eur. J. Biochem. 1967,
2, 286-293.
(28) Andrieux, C. P.; Saveant, J. M. In InVestigation of Rates and
Mechanisms of Reactions. Part II, 4th ed.; Bernasconi, C. F., Ed.; John
Wiley & Sons: New York, 1986; Vol. 6, pp 305-390.

Model Systems for FlaVoenzyme ActiVity
J. Am. Chem. Soc., Vol. 119, No. 5, 1997 891
Figure 4. UV/vis spectra of N(10)-isobutyl-N(3)-methyl flavin 2: (a)
fully oxidized (Flox), (b) flavin radical anion (Flrad-) generated
electrochemically in the absence of phthalimide, (c) fully reduced flavin
(FlredH-) generated electrochemically in the presence of 3 equiv of
phthalimide.
Figure 3. Change in the SEEPR spectrum of N(10)-isobutyl-N(3)-
methyl flavin 2 (10-3 M in CH2Cl2, 0.1 M TBAP) upon addition of
phthalimide.
Scheme 3. The Two Possible Mechanisms for the Redox
Behavior of Flavin in the Presence of a Proton Donor DHa
phthalimide no signal was observable. The hyperfine splitting
of the flavin radical signal remained unchanged during the
addition of phthalimide, indicating the essentially instantaneous
further reduction of the neutral radical to the fully reduced flavin
anion (the neutral radical of flavin 2, which was chemically
generated following the procedure of Mueller et al.,30 exhibits
a markedly different EPR spectrum than the anion radical).
SEEPR of flavin 1 also results in the spectrum of the radical
anion Flrad-, yet bulk electrolysis has to proceed for a longer
time before the spectrum is observable, and the signal is of lower
intensity than the EPR signal of the radical anion of flavin 2.
During the bulk electrolysis of flavin 1, Flox acts as both the
electroactive species and the proton donor. This contrasts with
the bulk electrolysis of flavin 2 in the presence of phthalimide,
where an excess of external proton donor is available. As the
bulk electrolysis proceeds, Flox becomes the limiting reagent,
and deprotonated Flox- and FlredH- accumulate in solution. The
a If the flavin is not alkylated at N(3), DH represents Flox in bulk
solution.
proton donors. Since the protonation and second one-electron
reduction succeed the first one-electron reduction (two unre-
solved one-electron reductions rather than a two-electron
reduction), the position of the reduction wave is only slightly
affected. In the case of acetic acid, however, a prewave is
observed. Once again, aqueous acidities can be used as
benchmarks to provide insight into this behavior. Because the
pKa of acetic acid (pKa ) 4.75)26 is lower than the pKa of fully
reduced flavin (pKa ) 6.7),4 the neutral fully reduced flavin
FlredH2 is formed, which has a much lower solubility in
methylene chloride29 due to additional hydrogen-bond interac-
tions. The observed prewave is caused by adsorption of FlredH2
onto the working electrode. It is well-known that flavins tend
to adsorb onto mercury working electrodes in aqueous acidic
media, giving rise to a prewave observed in both polarography
and cyclic voltammetry.6,9
To provide further evidence for our proposed flavin reduction
mechanism, we investigated the reduction of flavin 2 using
simultaneous electrochemistry and EPR. Bulk electrolysis of
a solution of flavin 2 in CH2Cl2 within the EPR cavity generated
a strong signal corresponding to the flavin radical anion. SEEPR
of flavin 2 in the presence of increasing amounts of the external
proton donor phthalimide led to a gradual decrease of the EPR
signal intensity (Figure 3). A plot of the double integral of the
EPR signal versus equivalents of phthalimide added shows a
roughly exponential relationship. In the presence of excess
-
equilibrium of deprotonated Flox and FlredH- undergoing
-
comproportionation to two molecules of Flrad under those
conditions is shifted toward the radical product, leading to the
observable EPR signal. This explains the napp of 1 determined
by previous researchers for the reduction of flavin not alkylated
at N(3) in aprotic media.10,11 In cyclic voltammetry, on the
other hand, only the small fraction of flavin 1 adjacent to the
electrode is reduced, and a large excess of Flox is available in
bulk solution. As a result, comproportionation does not become
a significant process. This interpretation is backed up by SEEPR
of nonalkylated flavin 1 in the presence of 3 equiv of
phthalimide. As expected, no EPR signal was observable after
bulk electrolysis.
To enable spectroscopic determination of the nature of the
two-electron-reduced species, we carried out a series of UV/
vis spectroelectrochemical experiments. For these studies bulk
electrolysis was carried out within an optically transparent thin
layer electrode (OTTLE).31 Since the light beam only crosses
the thin layer part and not the bulk solution, the spectrum of
the electrogenerated species can be recorded without convolution
by the spectrum of the parent species. Electrolysis of a solution
of flavin 2 in CH2Cl2 resulted in the formation of the red radical
anion Flrad-, verified by characteristic absorption maxima at 375
(30) Mueller, F.; Hemmerich, P.; Ehrenberg, A.; Palmer, G.; Massey,
V. Eur. J. Biochem. 1970, 14, 185-196.
(31) Heineman, W. R.; Hawkridge, F. M.; Blount, H. N. In Electroana-
lytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1984; Vol.
13, pp 1-104.
(29) Catalytic hydrogenation of N(10)-isobutyl flavin 1 over palladium
on carbon in methylene chloride causes the fully reduced neutral flavin
FlredH2 to precipitate out as a dull yellow solid.

892 J. Am. Chem. Soc., Vol. 119, No. 5, 1997
Niemz et al.
Figure 5. Spectroelectrochemical conversion of N(10)-isobutyl-N(3)-
methyl flavin 2 (10-3 M in CH2Cl2, 0.1 M TBAP) from the fully
oxidized form (Flox) to the radical anion (Flrad-).
Figure 6. Spectroelectrochemical conversion of N(10)-isobutyl-N(3)-
methyl flavin 2 (10-3 M in CH2Cl2, 0.1 M TBAP, 3 equiv of
phthalimide) from the fully oxidized from (Flox) to the fully reduced
form (FlredH-).
and 478 nm.32 The same experiment carried out in the presence
of 3 equiv of phthalimide leads to the generation of the dull
yellow and nonfluorescent, fully reduced flavin FlredH-, pos-
sessing only shoulder absorption in the visible region (Figure
interaction between a synthetic receptor and flavin 1, mimicking
the hydrogen-bonding pattern observed in flavoenzymes to
C(2)-O, N(3)-H and C(4)-O of the flavin cofactor, not only
lowers the standard reduction potential but also causes a relative
increase in wave II and a decrease in wave III observed in the
cyclic voltammogram of flavin 1.24 With the results presented
in this paper, these changes upon addition of receptor can be
interpreted as selective stabilization of the flavin radical anion
Flrad-, making one-electron reduction the preferred process. This
finding can be explained through the increase in pKa of N(3)-H
of flavin 1 in the hydrogen-bond complex, disfavoring inter-
molecular proton transfer, which is a prerequisite for the second
one-electron reduction. As mentioned previously, the ability
of flavoenzymes to selectively carry out one- vs two-electron-
redox processes is one of the key features responsible for their
catalytic diversity. The model system presented here is a first
step in understanding the modification of flavin redox processes
on a molecular level. Further studies on flavin host-guest
complexes using more sophisticated receptors combining hy-
drogen bonding and π-stacking interactions are underway and
will be reported in due course.
-
4).33 The spectra of the conversion of Flox to both Flrad and
FlredH- possessed isosbestic points (Figure 5 and 6). By
stepping the potential down both reactions were reversed, and
after full reoxidation the spectrum of Flox was regained. Several
reduction and oxidation cycles of the solution are possible
without changes in the spectra obtained for the different species,
verifying that the unusual redox behavior is not a consequence
of the formation of degradation products.
The same experiment carried out with a solution of flavin 1
-
leads to the spectrum of Flrad if no external proton donor is
present. Since the solution in the thin layer section is
exhaustively electrolyzed, the radical anion is generated via
comproportionation as described above. Again, UV/vis spec-
troelectrochemistry of flavin 1 in the presence of 3 equiv of
phthalimide leads to the spectrum of FlredH-.
-
In summary, the ability of the radical anion Flrad of flavin
1 and flavin 2 to abstract the imide proton of phthalimide
indicates a similar process taking place in flavins unalkylated
at N(3). In this case the N(3)-H imide proton of the flavin
acts as a proton source causing protonation of Flrad-. The
neutral radical FlradH is instantly reduced to FlredH-. The
availability of protons therefore marks the transition from one-
electron to two-electron flavin reduction. A detailed under-
standing of the redox processes taking place during the
electrochemical reduction of flavin is of fundamental value for
future model studies of flavoenzyme activity using electro-
chemical methods. We have already shown that noncovalent
Acknowledgment. This research was supported by the
National Science Foundation (CHE-9528099) and the Petroleum
Research Fund of the American Chemical Society (30199-G4).
V.R. thanks Research Corporation for a Cottrell Fellowship.
We thank Prof. Robert Allendoerfer (State University of New
York at Buffalo), Prof. Paul Lahti, and Prof. David Curran for
helpful discussions. Richard Letendre of the University of
Massachusetts Physics Machine Shop is gratefully acknowl-
edged for helpful suggestions during the design of the UV/vis
spectroelectrochemical cell.
(32) Massey, V.; Palmer, G. Biochemistry 1966, 5, 3181-3189.
(33) Dudley, K. H.; Ehrenberg, A.; Hemerich, P.; Mueller, F. HelV. Chim.
Acta 1964, 47, 1354-1383.
JA963704A