Oxygen reactivity in flavoenzymes: Context matters

Article abstract of DOI:10.1021/ja2081873

Many flavoenzymes-oxidases and monooxygenases-react faster with oxygen than free flavins do. There are many ideas on how enzymes cause this. Recent work has focused on the importance of a positive charge near N5 of the reduced flavin. Fructosamine oxidase has a lysine near N5 of its flavin. We measured a rate constant of 1.6 × 10⁵ M^-1 s^-1 for its reaction with oxygen. The Lys276Met mutant reacted with a rate constant of 291 M^-1 s^-1, suggesting an important role for this lysine in oxygen activation. The dihydroorotate dehydrogenases from E. coli and L. lactis also have a lysine near N5 of the flavin. They react with O₂ with rate constants of 6.2 × 10⁴ and 3.0 × 10³ M^-1 s^-1, respectively. The Lys66Met and Lys43Met mutant enzymes react with rate constants that are nearly the same as those for the wild-type enzymes, demonstrating that simply placing a positive charge near N5 of the flavin does not guarantee increased oxygen reactivity. Our results show that the lysine near N5 does not exert an effect without an appropriate context; evolution did not find only one mechanism for activating the reaction of flavins with O₂.

Full text of DOI:10.1021/ja2081873

COMMUNICATION
pubs.acs.org/JACS
Oxygen Reactivity in Flavoenzymes: Context Matters
||
Claudia A. McDonald,† Rebecca L. Fagan,†,§ Franc-ois Collard,‡, Vincent M. Monnier,‡ and Bruce A. Palfey*,†
†Department of Biological Chemistry, University of Michigan, Ann Arbor, Michigan 48109-0606, United States
‡Department of Pathology, Case Western Reserve University, Cleveland, Ohio 44106-7288, United States
S Supporting Information
b
ABSTRACT: Many flavoenzymes—oxidases and monooxy-
genases—react faster with oxygen than free flavins do.
There are many ideas on how enzymes cause this. Recent
work has focused on the importance of a positive charge
near N5 of the reduced flavin. Fructosamine oxidase has a
Figure 1. Mechanism of reaction of reduced flavin with oxygen. It has
been proposed that stabilizing the superoxide-semiquinone pair (circled
in red) allows rapid reaction with O2.
lysine near N5 of its flavin. We measured a rate constant of
1.6 Â 105 MÀ1 sÀ1 for its reaction with oxygen. The Lys276Met
There are many ideas on how enzymes can speed the reaction
mutant reacted with a rate constant of 291 MÀ1 sÀ1, suggesting
of reduced flavins with O2. These focus either on controlling the
an important role for this lysine in oxygen activation. The
access of O2 to the flavin or on stabilizing O2À in the caged radical
dihydroorotate dehydrogenases from E. coli and L. lactis also
intermediate. Access has been proposed to be controlled by
have a lysine near N5 of the flavin. They react with O2 with rate
channels through which O2 would approach the flavin, facilitat-
constants of 6.2 Â 104 and 3.0 Â 103 MÀ1 sÀ1, respectively.
ing their encounter.7,8 Recent molecular dynamics studies sug-
The Lys66Met and Lys43Met mutant enzymes react with rate
gested that the protein matrix guides oxygen to a spot over C4a of
constants that are nearly the same as those for the wild-type
the flavin for the reaction in some enzymes.9,10 Crystal structures
enzymes, demonstrating that simply placing a positive
of L-galactono-γ-lactone dehydrogenase (GALDH), which be-
charge near N5 of the flavin does not guarantee increased
long to the vanillyl-alcohol oxidase family, reveal that an alanine is
oxygen reactivity. Our results show that the lysine near N5
near the C4a position of the flavin; it reacts poorly with O2.11
does not exert an effect without an appropriate context;
Replacing the alanine residue of GALDH with a glycine increased
evolution did not find only one mechanism for activating the
oxygen activation by 400-fold, suggesting that the bulk of the
reaction of flavins with O2.
alanine side chain inhibits the flavoprotein dehydrogenase from
reacting with molecular oxygen.
Ideas that consider increasing reactivity emphasize ways in
which the protein could stabilize the obligate semiquinoneÀ
lavins in solution, when reduced, react with molecular oxygen.
However this reaction is not very fast. The reaction is slow
superoxide intermediate, especially by considering the polarity of
À
the local environment in which O2 would form.12 Lewis acid
F
catalysis could promote superoxide formation.13,14 For example,
glucose oxidase has two histidines near the flavin that appear to
be important for oxygen activation. The positive charge of one Àof
because spin inversion is required for the reaction of a singlet
(reduced flavin) with a triplet (O2) to form singlet products, oxidized
flavin and H2O2.1,2 This reaction occurs by electron transfer from
reduced flavin to O2 yielding a caged radical pair—superoxide anion
and semiquinone—which collapses to the C4a-hydroperoxide, which
decomposes into H2O2 and oxidized flavin (Figure 1).
the histidines stabilizes the developing negative charge of O2
.
The positive charge that activates oxygen is thought to be located
on the product bound to the active site in choline oxidase.15
Recent studies have shown that a lysine near N5 is important
for the reactivity of several enzymes with O2. Monomeric sarcosine
oxidase, an enzyme which contains covalently bound flavin, cata-
lyzes the oxidation of sarcosine.16,17 Monomeric sarcosine oxi-
dase has a lysine that hydrogen-bonds to a water which hydrogen-
bonds to N5 of the flavin.18 Mutating Lys265 to the neutral
methionine decreased the rate constant for the reaction of O2 by
8000-fold.19 Crystallography showed that ClÀ can bind in a
pocket, near N5 of the flavin, lysine, and the water, suggesting
that this is a preorganized site for the reaction of O2.20 Another
oxidase, N-methyltryptophan oxidase, also has a lysine near N5 of
the flavin. Mutating this residue to glutamine decreased the rate
constant for the oxygen reaction by ∼2500-fold.21 These are
Flavoproteins can be categorized based on their reactivity with
O2 and the products that are formed from their reactions.3
Reduced oxidases react rapidly with O2 to give an oxidized
enzyme and H2O2.3 Reduced monooxygenases react rapidly to
form the flavin hydroperoxide, which either oxygenates a substrate
or eliminates H2O2.4 Although the radical pair in Figure 1 is thought
to be an intermediate in oxidases and monooxygenases, it has not
been observed, with the possible exception of one enzyme,5
suggesting that forming the radical pair is rate-determining. Dehy-
drogenases usually react slowly with O2, sometimes more slowly
than free flavins, to give a mixture of H2O2 and superoxide anion.6
There is considerable interest in understanding how reduced
flavoenzymes control oxygen reactivity.6 Structures of flavoproteins
have been studied in an attempt to uncover clues for the basis of the
acceleration of the reaction. At this time there is no clear under-
standing of how structures may explain oxygen reactivity.
Received: August 30, 2011
Published: September 29, 2011
r
2011 American Chemical Society
16809
dx.doi.org/10.1021/ja2081873 J. Am. Chem. Soc. 2011, 133, 16809–16811
|

Journal of the American Chemical Society
COMMUNICATION
Table 1. Oxidative Half-Reaction Rate Constants
k
ox (wild-type)/
Enzyme
Wild-type FAOXa
Lys276Met FAOXa
kox (MÀ1sÀ1
)
kox (mutant)
1.6 Â 105 b
291
550
29
Lys53Met FAOXa
5.6 Â 103 b
3.0 Â 103
2.6 Â 103
6.2 Â 104 c
5.0 Â 103 c
6.2 Â 104
3.7 Â 104
5.9 Â 104
Wild-type L. lactis DHOD
Lys43Met L. lactis DHOD
Wild-type E. coli DHOD
Wild-type E. coli DHOD•Orotate
E. coli Lys66Met
1.2
12.4d
1.0
Figure 2. Fructosamine oxidase (left, from 3dje.pdb) and E. coli DHOD
(right, from 1f76.pdb) active site structures. The DHOD complex with
orotate is shown, but the structure remains the same in its absence. Note
that each enzyme has a lysine near N5 of the flavin. The structure of a
Class 1A DHOD is not shown but is essentially identical.
E. coli His19Asn
1.7
E. coli Arg102Met
1.1
a FAOX, fructosamine oxidase. b Reference 23. c Reference 26. d This is
k
ox (free wild-type)/kox (orotate complex).
among the largest decreases in oxygen reactivity seen as an effect
of site directed mutagenesis.
O2 with the reduced wild-type enzyme is 6.2 Â 104 MÀ1 s
,
À1 26 a
value that is similar to that seen in some oxidases.28 The oxidative
half-reaction of the Lys66Met mutant enzyme was studied by
mixing the reduced enzyme with buffer containing various con-
centrations of O2. The Lys66Met mutant enzyme reacted with a
rate constant of 6.2 Â 104 MÀ1 sÀ1, showing no change. Neither
the wild-type nor the mutant enzyme formed an observable
intermediate during their reactions, and each produced H2O2.24
The Class 1A DHOD from Lactococcus lactis also has a lysine
(Lys43) near N5 of the isoalloxazine of the flavin.29 The rate con-
stant for the reaction of O2 with the reduced wild-type enzyme is
3.0 Â 103 MÀ1 sÀ1. The Lys43Met mutant enzyme reacted with a
These observations show that a lysine near N5 is a key player
in oxygen activation in these oxidases and thus sparked our
interest in exploring the generality of lysine’s ability to accelerate
the reaction of O2. The model-enzymes fructosamine oxidase
and dihydroorotate dehydrogenases (DHODs) were studied
because they all have lysines near N5 (Figure 2).
Fructosamine oxidase from the fungus Aspergillus fumigatus is
a flavoprotein oxidase that catalyzes the oxidation of the carbonÀ
nitrogen bond of fructosamines. The amine substrate of the
enzyme, formed physiologically by the spontaneous reaction of
glucose and amino groups of amino acids, is oxidized to an imine
and hydrolyzed nonenzymatically.22
rate constant of 2.6 Â 103 MÀ1
s
À1, only 1.2-fold slower,
exhibiting virtually no change. Again, the removal of the lysine
residue near N5 of the flavin had no significant impact. Neither
the mutant nor the wild-type enzyme formed an intermediate
during their reactions, and each produced H2O2.24
The reaction of the reduced wild-type fructosamine oxidase
with O2 was studied in stopped-flow experiments. Flavin oxida-
tion was observed by an increase in absorbance at 450 nm
without any indication of an intermediate. Observed rate con-
stants varied linearly with O2 concentrations, giving a bimole-
The reactivity of the mutant DHODs shows that lysine near
N5 is not guaranteed to be important for oxygen reactivity. If O2
reacts elsewhere in DHOD, then the Lys66Met substitution
should have little effect on the Class 2 enzyme, as observed. The
physiological oxidizing substrate of Class 2 DHODs is ubiqui-
none, which binds near the methyls of the isoalloxazine of the
flavin; the reaction occurs by two single-electron transfers.26 There-
fore, Class 2 DHODs have evolved to transiently stabilize the flavin
semiquinone. There are two conserved positively charged residues
in the ubiquinone binding site, His19 and Arg102 in the E. coli
enzyme. Conceivably, either of these positive charges could
stabilize a developing negative charge if O2 reacted at this site.
Therefore the reactivity of the His19Asn and Arg102Met mutant
enzymes was determined. Neither change significantly altered
the bimolecular rate constant for the reaction of O2 (Table 1),
eliminating this site as the site where O2 reacts.
A unifying mechanism for oxygen activation by flavoproteins is
being sought by many researchers. Our results show that a single
mechanism cannot explain oxygen activation. The lysine near
N5, a critical ingredient in oxygen activation for some oxidases,
does not exert an effect without an appropriate context; some
feature (or features) of the active site allows it to activate oxygen.
Aligning the isoalloxazines of the structures of fructosamine oxidase
and Class 1A and 2 DHODs shows that there is enough space near
N5 and the lysine in each to accommodate O2. DHOD (not an
oxidase) reacts with O2 at virtually the same rate regardless of the
cular rate constant of 1.6 Â 105 MÀ1 s
.
À1 23 Fructosamine oxidase
has a lysine orthologous to that of monomeric sarcosine oxidase,
the residue identified to be important in O2 reactivity. Lys276 was
mutated to methionine. The reduced Lys276Met mutant reacts
with O2 with a rate constant of 291 MÀ1 sÀ1, a 550-fold decrease
(Table 1). Again, there was no indication of an intermediate and
the reaction produced H2O2.24 As with monomeric sarcosine
oxidase, a large decrease in reactivity occurred, showing that
lysine plays an important role in oxygen activation in these two
oxidases. The location of this positive charge is important. The
Lys53Met mutant enzyme, where the lysine is nearly centered
over the aromatic ring of the flavin on the si-face, reacted with O2
only 30-fold slower than wild-type.23
Unrelated enzymes, DHODs, catalyze the oxidation of dihy-
droorotate to orotate during the only redox reaction in the
pyrimidine biosynthetic pathway. DHODs can be categorized
into two different classes based on their sequences: Class 1 and
Class 2.25 The Class 1 and 2 enzymes have different selectivity for
oxidizing substrates and are also located in different compartments
of the cell. The Class 2 enzymes are mitochondrial membrane
proteins while Class 1 enzymes are cytosolic. Class 2 DHODs
in vivo are oxidized by ubiquinone while Class 1A enzymes are
oxidized by fumarate.3
Class 2 E. coli DHOD has a lysine (Lys 66) near N5 of the
isoalloxazine of the flavin.27 The rate constant for the reaction of
16810
dx.doi.org/10.1021/ja2081873 |J. Am. Chem. Soc. 2011, 133, 16809–16811

Journal of the American Chemical Society
COMMUNICATION
presence of a lysine near N5; it plays no role in oxygen activation.
Other structural features must be responsible. Mutagenesis
eliminated the quinone binding bucket as the site of the reaction
of O2. The hydrophilic pyrimidine binding site over the si-face of
the flavin, ringed by hydrogen-bonding side chains, does not appear
to be the site of the reaction of O2 because the reduced enzymeÀ
orotate complex reacts only about 10-fold slower than the free
enzyme (Table 1).26 The site of the reaction of O2 in DHODs
remains a mystery. Our results suggest that many mechanisms are
possible for oxygen activation by flavoproteins. This is consistent
with the variety of structures of flavoproteins that react with
oxygen.3 A small, reactive molecule like O2, which appeared after
the origin of life and flavoenzymes, would be expected to find
many sites to react.
(11) Leferink, N. G.; Fraaije, M. W.; Joosten, H. J.; Schaap, P. J.;
Mattevi, A.; van Berkel, W. J. J. Biol. Chem. 2009, 284, 4392–7.
(12) Wang, R.; Thorpe, C. Biochemistry 1991, 30, 7895–901.
(13) Roth, J. P.; Klinman, J. P. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
62–7.
(14) Roth, J. P.; Wincek, R.; Nodet, G.; Edmondson, D. E.; McIntire,
W. S.; Klinman, J. P. J. Am. Chem. Soc. 2004, 126, 15120–31.
(15) Ghanem, M.; Gadda, G. Biochemistry 2005, 44, 893–904.
(16) Wagner, M. A.; Trickey, P.; Chen, Z. W.; Mathews, F. S.; Jorns,
M. S. Biochemistry 2000, 39, 8813–24.
(17) Wagner, M. A.; Jorns, M. S. Biochemistry 2000, 39, 8825–9.
(18) Trickey, P.; Wagner, M. A.; Jorns, M. S.; Mathews, F. S.
Structure 1999, 7 (3), 331–45.
(19) Zhao, G.; Bruckner, R. C.; Jorns, M. S. Biochemistry 2008,
47, 9124–35.
(20) Kommoju, P. R.; Chen, Z. W.; Bruckner, R. C.; Mathews, F. S.;
Jorns, M. S. Biochemistry 2011, 50, 5521–34.
(21) Bruckner, R. C.; Winans, J.; Jorns, M. S. Biochemistry 2011,
50, 4949–62.
(22) Gerhardinger, C.; Taneda, S.; Marion, M. S.; Monnier, V. M.
J. Biol. Chem. 1994, 269, 27297–302.
’ ASSOCIATED CONTENT
S
Supporting Information. Materials and methods pp
b
S2ÀS5. This material is available free of charge via the Internet
at http://pubs.acs.org.
(23) Collard, F.; Fagan, R. L.; Zhang, J.; Palfey, B. A.; Monnier, V. M.
Biochemistry 2011, 50, 7977–7986.
(24) Cytochrome c, which is rapidly reduced by superoxide, was
included in oxidation mixtures in order to determine the reduced oxygen
species produced by the flavoenzyme. No superoxide was detected.
(25) Bj€ornberg, O.; Rowland, P.; Larsen, S.; Jensen, K. F. Biochem-
istry 1997, 36, 16197–205.
(26) Palfey, B. A.; Bj€ornberg, O.; Jensen, K. F. Biochemistry 2001,
40, 4381–90.
(27) Nørager, S.; Jensen, K. F.; Bj€ornberg, O.; Larsen, S. Structure
’ AUTHOR INFORMATION
Corresponding Author
brupalf@umich.edu
Present Addresses
§Department of Biochemistry, University of Iowa, 51 Newton Rd.,
Iowa City, IA 52242.
de Duve Institute, Universitꢀe Catholique de Louvain, 75-39,
Avenue Hippocrate, 1200 Brussels, Belgium.
2002, 10, 1211–23.
(28) Macheroux, P.; Massey, V.; Thiele, D. J.; Volokita, M. Biochemistry
1991, 30, 4612–9.
(29) Rowland, P.; Nielsen, F. S.; Jensen, K. F.; Larsen, S. Structure
1997, 5, 239–52.
’ ACKNOWLEDGMENT
This work was supported by the National Institutes of Health
Grant GM61087 to B.A.P. R.L.F was supported by NIGMS
Training Grant GM07767. C.A.M. was supported by a University
of Michigan Rackham Merit Fellowship and National Institutes
of Health Grant GM61087-08-S1. F.C. was supported by the
Juvenile Diabetes Foundation and the Belgian F.N.R.S. V.M.M.
was supported by the National Institutes of Health Grant
EY07099. We thank Maria Nelson for her work on creating and
purifying the His19Asn and Arg102Met E.coli DHOD mutant
enzymes.
’ REFERENCES
(1) Bruice, T. C. Isr. J. Chem. 1984, 24, 54–61.
(2) Massey, V. J. Biol. Chem. 1994, 269, 22459–62.
(3) Fagan, R. L.; Palfey, B. A., Flavin-Dependent Enzymes; Elsevier:
Oxford, U.K., 2010; Vol. 7.
(4) Palfey, B. A.; McDonald, C. A. Arch. Biochem. Biophys. 2010,
493, 26–36.
(5) Pennati, A.; Gadda, G. Biochemistry 2011, 50, 1–3.
(6) Mattevi, A. Trends Biochem. Sci. 2006, 31, 276–83.
(7) Coulombe, R.; Yue, K. Q.; Ghisla, S.; Vrielink, A. J. Biol. Chem.
2001, 276, 30435–41.
(8) Chen, L.; Lyubimov, A. Y.; Brammer, L.; Vrielink, A.; Sampson,
N. S. Biochemistry 2008, 47, 5368–77.
(9) Baron, R.; Riley, C.; Chenprakhon, P.; Thotsaporn, K.; Winter,
R. T.; Alfieri, A.; Forneris, F.; van Berkel, W. J.; Chaiyen, P.; Fraaije,
M. W.; Mattevi, A.; McCammon, J. A. Proc. Natl. Acad. Sci. U.S.A. 2009,
106, 10603–8.
(10) Saam, J.; Rosini, E.; Molla, G.; Schulten, K.; Pollegioni, L.;
Ghisla, S. J. Biol. Chem. 2010, 285, 24439–46.
16811
dx.doi.org/10.1021/ja2081873 |J. Am. Chem. Soc. 2011, 133, 16809–16811