Aminoacetone oxidase from Streptococcus oligofermentans belongs to a new three-domain family of bacterial flavoproteins

Article abstract of DOI:10.1042/BJ20140972

The aao_Sogene from Streptococcus oligofermentans encodes a 43 kDa flavoprotein, aminoacetone oxidase (SoAAO), which was reported to possess a low catalytic activity against several different L-amino acids; accordingly, it was classified as an L

Full text of DOI:10.1042/BJ20140972

Biochem. J. (2014) 464, 387–399 (Printed in Great Britain) doi:10.1042/BJ20140972
387
Aminoacetone oxidase from Streptococcus oligofermentans belongs to a
new three-domain family of bacterial flavoproteins
1
Gianluca Molla*† , Marco Nardini‡, Paolo Motta*, Paola D’Arrigo†§ꢀ, Walter Panzeriꢀ and Loredano Pollegioni*†
*Dipartimento di Biotecnologie e Scienze della Vita, Universita` degli Studi deII’Insubria, via J.H. Dunant 3, 21100 Varese, ltaly
†The Protein Factory, Centro Interuniversitario di Biotecnologie Proteiche, Politecnico di Milano, ICRM CNR Milano, and Universita` degli Studi deII’Insubria, Varese, Italy
‡Dipartimento di Bioscienze, Universita` degli Studi di Milano, 20133 Milano, Italy
§Dipartimento di Chimica, Materiali ed Ingegneria Chimica G. Natta, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milano, Italy
ꢀCNR-Istituto di Chimica del Riconoscimento Molecolare, Politecnico di Milano, via Mancinelli 7, 20131 Milano, Italy
The aaoSo gene from Streptococcus oligofermentans encodes
a 43 kDa flavoprotein, aminoacetone oxidase (SoAAO), which
was reported to possess a low catalytic activity against several
different L-amino acids; accordingly, it was classified as an L-
amino acid oxidase. Subsequently, SoAAO was demonstrated
to oxidize aminoacetone (a pro-oxidant metabolite), with an
activity ∼25-fold higher than the activity displayed on L-lysine,
thus lending support to the assumption of aminoacetone as the
preferred substrate. In the present study, we have characterized
the SoAAO structure–function relationship. SoAAO is an
FAD-containing enzyme that does not possess the classical
properties of the oxidase/dehydrogenase class of flavoproteins
(i.e. no flavin semiquinone formation is observed during
anaerobic photoreduction as well as no reaction with sulfite)
and does not show a true L-amino acid oxidase activity.
From a structural point of view, SoAAO belongs to a novel
protein family composed of three domains: an α/β domain
corresponding to the FAD-binding domain, a β-domain partially
modulating accessibility to the coenzyme, and an additional
α-domain. Analysis of the reaction products of SoAAO on
aminoacetone showed 2,5-dimethylpyrazine as the main product;
we propose that condensation of two aminoacetone molecules
yields 3,6-dimethyl-2,5-dihydropyrazine that is subsequently
oxidized to 2,5-dimethylpyrazine. The ability of SoAAO to
bind two molecules of the substrate analogue O-methylglycine
ligand is thought to facilitate the condensation reaction. A
specialized role for SoAAO in the microbial defence mechanism
related to aminoacetone catabolism through a pathway yielding
dimethylpyrazine derivatives instead of methylglyoxal can be
proposed.
Key words: aminoacetone, 2, 5-dimethylpyrazine, enzyme
structure, flavoprotein, X-ray crystallography.
expressed as a recombinant protein in Escherichia coli [4].
This enzyme was reported to possess a low catalytic activity
(<0.5 unit/mg of protein) against several different L-amino acids,
the best substrates being L-aspartate, L-tryptophan, L-lysine and L-
isoleucine. For this reason, it was originally classified as an LAAO
(L-amino acid oxidase) (EC 1.1.3.2) [4], despite the absence of
any significant sequence similarity with known prokaryotic or
eukaryotic LAAOs, except for the sequence motif typical of the
dinucleotide binding required for FAD binding [6–8].
LAAO is an FAD-containing enzyme that catalyses the
stereoselective oxidative deamination of L-amino acids to yield
the corresponding α-keto acids, ammonia and H2O2 [9]. This
flavoenzyme occurs widely in snake and insect venom [10], but
also in several fungi, bacteria (e.g. Rhodococcus opacus, Bacillus
carotarum, Streptomyces endus and Pseudomonus sp. P-501) and
algae (for a comprehensive review, see [9]). On the structural side,
LAAOs belong to the N-terminal FAD-bound reductase family
and share with DAAO (D-amino acid oxidase) (EC 1.4.3.3) the
hydride-transfer mechanism of dehydrogenation [11,12]. LAAOs
are promising biocatalysts for a number of biotechnological
applications since they catalyse the same reaction as DAAO,
but on the opposite enantiomer [9,13]; their biotechnological
INTRODUCTION
Streptococci represent ∼20% of bacteria in the human oral
cavity, a challenging environment for bacterial survival because it
often undergoes rapid and harsh changes in nutrient availability,
pH and oxygen concentration. Among the 25 species of oral
streptococci, some species can switch from commensal micro-
organisms to opportunistic pathogens; i.e. Streptococcus mutans
is responsible for the development of the two most prevalent
oral infectious diseases: dental caries and periodontitis [1].
Streptococcus oligofermentans only exists on healthy dental
plaques because it is able to outcompete S. mutans [2]. Indeed,
S. oligofermentans employs both lactic acid (one of the main
products of S. mutans metabolism) and several L-amino acids
as nutrients and produces hydrogen peroxide (H2O2) [3,4]. Such
a capacity was proposed to arise from the product of the aaoSo
gene (GenBank® accession number EU495328) [4], acquired by
S. oligofermentans through a horizontal gene transfer event [5];
this process may have been important in extending the ecological
niche occupied by this Streptococcus strain.
The aaoSo gene encodes a 43 kDa flavoprotein [SoAAO (S.
oligofermentans aminoacetone oxidase)], which was successfully
Abbreviations: DAAO, D-amino acid oxidase; HRP, horseradish peroxidase; LAAO, L-amino acid oxidase; SoAAO, Streptococcus oligofermentans
aminoacetone oxidase; SSAO, semicarbazide-sensitive amine oxidase.
1
To whom correspondence should be addressed (email gianluca.molla@uninsubria.it).
The atomic co-ordinates and structure factors of Streptococcus oligofermentans aminoacetone oxidase (SoAAO) have been deposited in the PDB
2
under accession codes 4CNJ and 4CNK, for the native SoAAO and for the complex between SoAAO and O-methylglycine respectively.
3
The nucleotide sequence of the gene coding for Streptococcus oligofermentans aminoacetone oxidase has been deposited in the DDBJ, EMBL,
®
GenBank and GSDB Nucleotide Sequence Databases under accession number KC344836.
ꢁ
c
ꢁ
c
The Authors Journal compilation 2014 Biochemical Society

388
G. Molla and others
application is hampered by difficulties inherent to large-scale
production as recombinant proteins [9].
crude extract was loaded on a HiTrap chelating column (GE
Healthcare) pre-loaded with Ni2 + and equilibrated in 50 mM
sodium pyrophosphate (pH 7.2), 1 M NaCl, 20 mM imidazole,
10 μM FAD, 2.5 mM 2-mercaptoethanol and 5% (v/v) glycerol
[17]. The bound protein was eluted from the column using 20 mM
sodium pyrophosphate (pH 7.2), 0.5 M imidazole, 10 μM FAD,
1 mM 2-mercaptoethanol and 5% (v/v) glycerol. The pure protein
was then transferred in storage buffer (20 mM Tris/HCl, pH 8.0)
by chromatography on a PD10 desalting column (GE Healthcare).
Recently, the aaoSo gene was found as part of an operon together
with mutT, a gene coding for a specific pyrophosphohydrolase
(8-oxo-dGTPase) involved in the elimination of the mutagen-
oxidized deoxynucleoside triphosphate 8-oxo-dGTP [14].
Biochemical studies demonstrated that the SoAAO protein is
able to oxidize aminoacetone (a pro-oxidant metabolite), with an
activity ∼25-fold higher than the activity displayed on L-lysine
(0.048 compared with 0.002 units/mg of protein respectively),
and that the specific activity values with L-amino acids were
∼10–50-fold lower than those reported in the original paper [4],
thus lending support to the assumption of aminoacetone as the
preferred substrate.
In the present study, we have characterized the SoAAO
structure–function relationship. Our results show that this
flavoprotein fold is based on a specific three-domain architecture,
unrelated to that of LAAOs, that SoAAO displays peculiar
biochemical properties and does not show LAAO activity. Such
results support a specialized role for SoAAO in the microbial
defence mechanism related to aminoacetone catabolism through
a mechanism yielding dimethylpyrazine derivatives instead of
methylglyoxal, as instead proposed for SSAOs (semicarbazide-
sensitive amine oxidases) (EC 1.4.3.21) [15].
Enzyme activity assay
SoAAO activity on L-amino acids was assayed using a standard
continuous spectrophotometric method via determination of
H2O2 with an enzyme-coupled assay using 1 unit of HRP
(horseradish peroxidase) (Sigma–Aldrich) and 0.32 mg/ml o-
dianisidine, assuming a ꢀε440 of 13.0 mM− 1·cm− 1 (final volume
of 1 ml) [18]. This assay was performed in 50 mM sodium
pyrophosphate buffer (pH 7.5–8.5), using 10–100 mM L-amino
acid as the substrate at 25◦C.
The activity of SoAAO on aminoacetone was determined
at 37◦C using both continuous and discontinuous assays.
In the continuous assay, 28 μg of pure SoAAO (0.43 μM
final concentration) were mixed with 25 mM aminoacetone
(Fluorochem), 1.5 mM 4-aminoantipyrine and 5 units of HRP in
20 mM Tris/HCl (pH 8.0) (assay volume of 1.0 ml). Enzymatic
activity was determined from the increase of absorbance
at 505 nm: because of the stoichiometry of the reaction,
in which two molecules of H2O2 react with one molecule
of 4-aminoantipyrine, the concentration of produced H2O2
(corresponding to the concentration of oxidized aminoacetone)
is twice the concentration of quinoneimine produced [18]. The
continuous assay was also performed using o-dianisidine as
the colorimetric reagent; in this case, 130 μg of pure SoAAO,
0.32 mg/ml o-dianisidine and 2 units of HRP in 20 mM Tris/HCl
(pH 8.0) (assay volume of 0.8 ml) were used; enzymatic activity
was determined from the increase of absorbance at 440 nm
[18]. In the discontinuous assay, the reaction mixture (1.5 ml)
contained 2 μM (133 μg) SoAAO and 12.5 mM aminoacetone or
L-aspartate in 20 mM Tris/HCl (pH 8.0). The reaction mixture was
incubated at 37◦C for 30 min and then the H2O2 produced was
quantified using the o-dianisidine/HRP-coupled assay. Briefly,
650 μl of the reaction mixture was added to 600 μl of the solution
containing 4-aminoantipyrine and phenol (see above): the reaction
proceeded for 4 min at room temperature, and then HRP was
added (at 0.05–5 units/ml final concentration: the amount of HRP
did not affect the absorbance value recorded). After a further 4 min
of incubation at room temperature, the absorbance at 505 nm was
recorded [4,18].
The products of SoAAO reaction on aminoacetone was
also assayed by measuring the amount of ketoacid produced
by reaction with 2,4-dinitrophenylhydrazine [15,19]. Briefly,
2.5 mM aminoacetone (an amount chosen so as to not interfere
with the reaction of 2,4-dinitrophenylhydrazine with the reaction
products) was mixed with 325 μg of SoAAO in 20 mM Tris/HCl
(pH 8.0) (2 ml final volume), under stirring at 37◦C: at 0, 15,
30 and 120 min, a 500 μl aliquot was withdrawn and mixed
with 150 μl of 10 mM 2,4-dinitrophenylhydrazine, incubated for
10 min at 37◦C and then 1050 μl of 2 M NaOH was added.
The visible absorbance spectrum of the mixture was recorded
and the ketoacid product was quantified from the absorbance at
518 nm using a molar absorption coefficient of 5.432 mM− 1·cm− 1
estimated from a calibration curve obtained under identical
conditions using standard methylglyoxal (Fluorochem).
MATERIALS AND METHODS
Design, synthesis and cloning of cDNA coding for SoAAO
The synthetic cDNA coding for SoAAO was designed by in silico
back translation of the amino acid sequence reported in GenBank®
(accession number ACA52024) [4]. Codon optimization for
expression in E. coli was performed according to the Codon
Usage Database (http://www.kazusa.or.jp/codon/) by eliminating
rare codons, i.e. codons used with a <10% frequency for
coding a given amino acid. Synthetic cDNA was produced by
GeneArt®. The cDNA molecule coding for SoAAO was cloned
into pET24b (Novagen) expression vector between the NdeI
and XhoI restriction sites, producing a C-terminal His6-tagged
recombinant protein.
Growth conditions, protein expression and purification
For protein expression, pET24-SoAAO plasmid was transferred
into E. coli BL21(DE3)pLysS host cells (Novagen) [16,17].
Starter cultures were prepared from a single colony of E.
coli cells carrying the recombinant plasmid growing at 37◦C
on LB broth medium (10 g/l Bacto tryptone, 5 g/l yeast
extract and 5 g/l NaCl), to which the antibiotics kanamycin
(30 μg/ml final concentration) and chloramphenicol (34 μg/ml
final concentration) were added. For protein purification trials,
2 litre baffled flasks containing 500 ml of LB medium were
inoculated with the starter culture (initial D600 of 0.1) and
grown at 37◦C under shaking (220 rev./min) until a D600 of 0.8
was reached. Protein expression was induced by adding 1 mM
(final concentration) IPTG. To promote the expression of the
holoprotein form of recombinant SoAAO, 1.2 μM FAD was also
added at the moment of IPTG addition. Cells were harvested by
centrifugation at 8000 g for 10 min at 4◦C 3 h after adding IPTG
and were stored at − 20◦C.
Cells were lysed by sonication (six cycles of 30 s each, with
30 s intervals) on ice in 50 mM Tris/HCl (pH 8.0), 40 μM FAD,
0.5 mM PMSF, 1 mM 2-mercaptoethanol and 10 μg/ml DNase I.
Cell lysates were centrifuged at 39000 g for 1 h at 4◦C. The
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

Biochemical characterization of aminoacetone oxidase from S. oligofermentans
389
One unit of SoAAO is defined as that amount of enzyme which
was negligible. The redox midpoint potential Em for the system at
equilibrium was calculated from the Nernst equation (eqn 1):
produces 1 μmol of H2O2 in 1 min at 37◦C.
Em,SoAAO = Edye − (2.3 RT/nF) · log([oxidized form]
SDS/PAGE electrophoresis, Western blot and isoelectric focusing
analyses
/[reduced form])
(1)
where R is the gas constant (8.31441 J·K− 1·mol− 1), T is the
absolute temperature, F is Faraday’s constant (9.6485381×104
C·mol− 1) and n is the number of electrochemical equivalents. All
potential values are reported compared with the standard hydrogen
electrode [21,25].
Protein samples from expression trials were analysed by
SDS/PAGE; gels were stained for proteins with Coomassie Blue
R-250. For Western blot analysis, proteins were transferred on to
an Immobilon-P transfer membrane (Millipore) and recombinant
SoAAO was identified using anti-His6-tag mouse monoclonal
antibodies (His-probe H-3) and HRP-conjugated goat anti-(mouse
IgG) antibodies (all from Santa Cruz Biotechnology). The
immunorecognition was detected using the ECL Plus Western
Blotting Detection System (GE Healthcare).
Determination of the isoelectric point under non-denaturing
conditions was carried out in a flatbed apparatus, FBE-3000 (GE
Healthcare), using a 6% (w/v) polyacrylamide gel slab containing
2.5% (v/v) of both Pharmalyte 5–8 and Pharmalyte 3–10 (GE
Healthcare). As anodic and cathodic solutions 1 M H3PO4 and 1 M
NaOH were used respectively. The gel was stained for proteins
with Coomassie Blue R-250.
Determination of the oligomerization state
The oligomerization state of recombinant SoAAO was determined
by gel-permeation chromatography on a Superdex 200 HR
column (GE Healthcare) using 20 mM Tris/HCl (pH 8.0),
100 mM NaCl and 2% glycerol as elution buffer.
Crystallization, structure determination and refinement
SoAAO crystals were obtained by vapour diffusion methods in a
hanging-drop set-up, mixing 1 μl of protein (at the concentration
of 13.3 mg/ml) and 1 μl of reservoir solution consisting of 2.0 M
(NH4)2SO4, 0.1 M sodium cacodylate or Mes (pH 6.5), and 0.2 M
NaCl. Bright yellow crystals (indicative that the bound FAD was
oxidized) with average dimensions of 150 μm×50 μm×50 μm,
were grown in ∼7–10 days at 4◦C. The crystals were transferred
to a cryoprotectant solution consisting of 2.2 M (NH4)2SO4,
0.1 M sodium cacodylate or Mes (pH 6.5), 0.2 M NaCl and
25% glycerol before flash-cooling in liquid nitrogen. The
native SoAAO crystals, belonging to the orthorhombic space
group P22121 (four protein molecules in the asymmetric unit),
diffracted at 2.7 Å (1 Å = 0.1 nm) resolution. Diffraction
data were collected at the European Synchrotron Radiation
Facility beamline ID29, and processed with XDS [26] and
SCALA [27].
Vapour diffusion co-crystallization experiments on the protein–
ligand complex were performed after overnight incubation of
SoAAO (final concentration 12.8 mg/ml) with O-methylglycine
(final concentration 7 mM) at 4◦C. Crystals of the complex were
obtained in a hanging-drop set-up under conditions matching
those of the native protein. They belong to the orthorhombic
space group C2221 (two protein molecules in the asymmetric
unit) and diffracted at 2.0 Å resolution. Diffraction data were
collected at the ESRF (European Synchrotron Radiation Facility)
beamline ID14-4, and processed with MOSFLM [28] and SCALA
[27]. Statistics for each data collection are reported in detail in
Table 1.
The SoAAO structure was solved by molecular replacement
with Phaser [29], using independently the co-ordinates of the
nucleotide-binding domain (residues 1–197 and 337–401) and
of the remaining protein (residues 198–336) from flavoprotein
HI0933 from Haemophilus influenzae (PDB code 2GQF) as
search models. The two independent solutions were combined
into a unique model and refined as rigid bodies using Refmac
[30]. The SoAAO sequence was then model-built into the electron
density using Coot [31]. The final model, restrained-refined at 2.7
Å (Rwork = 18.8%, Rfree = 23.9%) with Refmac [30], contains four
391-amino-acid-residue chains (1–391), four FAD molecules, 13
sulfate ions, eight glycerol molecules and 72 water molecules,
with good stereochemical parameters.
Spectral measurements
Absorbance spectra in the UV–visible region were recorded
using a Jasco V-560 spectrophotometer. Absorbance data were
recorded at 15◦C in 20 mM Tris/HCl, pH 8.0, except where
stated otherwise. Anaerobic samples were prepared in anaerobic
cuvettes by applying ten cycles of evacuation and then flushing
with oxygen-free argon. Photoreduction experiments were carried
out at 4◦C using an anaerobic cuvette containing 12.4 μM
enzyme, 1 mM EDTA and 0.5 μM 5-deazaflavin [20]. The
solution was photoreduced using a 250 W lamp and the progress
of the reaction was monitored spectrophotometrically [21].
Fluorescence measurements were performed using a Jasco
FP-750 spectrophotometer. All fluorescence measurements were
performed at 15◦C and at 0.8 mg/ml protein concentration. See
[22,23] for details.
CD spectra were recorded using a Jasco J-815 spectropolari-
meter equipped with a software-driven Peltier-based temperature
controller producing a temperature gradient of 0.5◦C/min and
analysed by means of Jasco software. For measurements above
250 nm, the cell path was 1 cm, whereas for measurements
in the 190–250 nm region, the cell path was 0.1 cm. All CD
measurements were performed in 20 mM Tris/HCl (pH 8.0) at
15◦C [23].
The redox potentials for the oxidized–reduced SoAAO couple
were determined by employing the dye equilibration method
described previously [21,24] at 15◦C. The enzyme solution
(18 μM final concentration) was mixed in an anaerobic cuvette
with 0.2 mM xanthine, 5 μM Benzyl Viologen as mediator and
reference dye (i.e. a dye possessing a midpoint redox potential
+
30 mV from the potential of SoAAO). The cuvette was made
anaerobic (see above), and the visible spectrum of the anaerobic
−
enzyme was recorded. The reaction was initiated by adding
0.18 unit of xanthine oxidase from the side arm and the absorbance
spectra were recorded until no increase in the 600 nm peak of the
Benzyl Viologen final acceptor was evident. The concentrations of
the oxidized and reduced forms of SoAAO were determined from
the absorbance values at various wavelengths (after subtracting
the dye’s contribution) [21]. The concentrations of the oxidized
and reduced forms of Benzyl Viologen were determined from the
absorbance values at 600 nm where the contribution of SoAAO
The structure of the SoAAO–O-methylglycine complex was
determined by molecular replacement using the program Phaser
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

390
G. Molla and others
Table 1 Data collection and refinement statistics
Outer shell statistics are shown within parentheses. R-merge = ꢀhꢀi|Ihi − <Ih>|/ꢀhꢀiIhi.
SoAAO
SoAAO + O-methylglycine
PDB code
4CNJ
4CNK
Data collection
Synchrotron beamline
Temperature (K)
Space group
ESRF ID29-1
100
P22121
ESRF ID14-4
100
C2221
˚
Cell dimensions (A)
a
b
c
78.95
129.28
224.11
64.68–2.70 (2.85–2.70)
231169
79.80
231.08
124.93
42.41–2.00 (2.11–2.00)
305537
˚
Resolution (A)
Observations
Unique reflections
Completeness (%)
R-merge (%)
I/σ(I)
68683
77717
98.7 (99.8)
10.4 (53.2)
10.6 (2.4)
3.7 (3.7)
99.5 (99.2)
14.9 (56.1)
9.0 (4.6)
Multiplicity
3.9 (4.1)
Refinement
R-factor/R-free (%)
Residues in the asymmetric unit
FAD molecules
18.8/23.9
16.9/21.2
4 × 391 (chains A, B, C and D)
2 × 391 (chains A and B)
4
2
Water molecules
O-Methylglycine
Glycerol
72
–
8
704
1 (chain A), 2 (chain B)
8
11
Sulfate ions
13
Model quality
Overall B-factor (A )
2
˚
60.9 (chain A), 66.4 (chain B), 49.2 (chain C), 59.2 (chain D)
24.6 (chain A), 18.3 (chain B)
RMSD from ideal values
˚
Bond lengths (A)
0.008
1.4
0.011
1.5
◦
Bond angles ( )
Ramachandran plot
Most favoured regions (%)
Additional allowed regions (%)
93.5
6.5
95.8
4.2
[29]. The crystal structure of native SoAAO was used as a
monomeric search model. The structure was then refined (rigid
body and restrained refinement) using the program Refmac
[30]. After a few cycles of refinement, the 2Fo–Fc electron-
density map showed structural details that allowed unambiguous
modelling of the bound ligand. The final model, restrained-
refined at 2.0 Å (Rwork = 16.9%, Rfree = 21.2%) contains two
391-amino-acid-residue chains (1–391), two FAD molecules,
three O-methylglycine molecules, 11 sulfate ions, eight glycerol
molecules and 704 water molecules, with good stereochemical
parameters.
3,6-dimethyl-2,5-dihydropyrazine cyclic form has peaks at m/z
111 and 221 (molecular mass of 110 g/mol, with the second
peak corresponding to the [2M + H]+ form). see Scheme 1 for
formulae. Standard methylglyoxal shows a peak at m/z 72. All
solvents were of analytical grade.
RESULTS
Properties of recombinant SoAAO
Recombinant SoAAO was purified from the crude extract of E.
coli cells bearing the pET-SoAAO-His plasmid, grown under
optimal conditions (i.e. collecting the cells 5 h after the addition
of 1 mM IPTG at a D600 of ∼0.8). SDS/PAGE and Western
blot analyses showed a band at ∼43 kDa corresponding to
recombinant SoAAO at ∼1 h, with the highest productivity at
5 h after IPTG addition (representing ∼10% of the protein
content) (see Supplementary Figures S1A and S1B). Western
blot analysis of soluble and insoluble fractions from cell lysate
shows that ∼50% of the recombinant SoAAO was expressed as
a soluble protein (Supplementary Figure S1B, right-hand lanes).
Changing the IPTG concentration (from 0.1 to 5 mM), or adding
1.2 μM FAD to the culture broth simultaneously to IPTG, as
suggested by [4], did not affect the level of recombinant SoAAO
production. Similarly, no increase in expression of soluble
SoAAO was apparent upon decreasing the growth temperature
(15 and 25◦C) after induction of protein expression. Metal-
chelating chromatography on a HiTrap chelating column allowed
the purification of ∼19 mg of SoAAO/litre of fermentation broth
All refinement statistics for both protein models are reported in
details in Table 1. The stereochemical quality of the model was
assessed using MolProbity [32].
Mass spectra analysis
Mass spectra were recorded on a ESI/MS Bruker Esquire 3000
PLUS (Esi Ion Trap LC/MSn System), by direct infusion of
methanol solution of compounds with an infusion rate of 4 μl/min.
Then, 400 μl of a solution containing 3.7 mg of aminoacetone
hydrochloride (42.4 mM) in 20 mM Tris/HCl buffer at pH 8.0
was treated with 400 μl of SoAAO in the same buffer (0.37
units/ml, 5.7 mg/ml). The pH was adjusted to 8.0 with NaOH.
The reactions were conducted at 37◦C on a thermomixer set to
600 rev./min. At regular intervals, 80 μl aliquots of the reaction
mixture were withdrawn, diluted with 240 μl of water and
stored in ice until MS analysis. Standard aminoacetone shows
a peak at m/z 74.5 (molecular mass of 73 g/mol), whereas its
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

Biochemical characterization of aminoacetone oxidase from S. oligofermentans
391
Scheme 1 Proposed reaction catalysed by SoAAO on aminoacetone
The condensation of two molecules of aminoacetone yields 3,6-dimethyl-2,5-dihydropyrazine that is oxidized to 2,5-dimethylpyrazine.
with a purity >90%, as judged by SDS/PAGE analysis (see
Supplementary Figures S1C and S1D). MS analysis yielded a
mass for the holoprotein of 43770 Da, in good agreement with the
expected value (43744 Da). The isolectric focusing point of
the recombinant SoAAO preparation determined under native
conditions is 9.1. Gel-permeation chromatography shows that
native SoAAO (at a concentration of up to 2.5 mg/ml) eluted
as a single peak at ∼15.1 ml, corresponding to a molecular mass
of ∼40 kDa, i.e. as a monomeric enzyme. The oligomerization
state of the holoenzyme form of SoAAO differs from that of the
snake venom LAAOs, which are stable homodimers in solution
[7,9].
Thermal stability of SoAAO was investigated by following
changes in the secondary structure, i.e. the CD signal at 208 nm
during the thermal ramps. Tm values of 36.9◦C and of 40.4◦C
were observed in the absence and in the presence of 10% glycerol
respectively. These results show a relative low thermal stability
of SoAAO in comparison with other flavoprotein oxidases
from mesophilic micro-organisms such as glycine oxidase from
Bacillus subtilis (Tm ranging from 56.9 to 63.9◦C), or D-amino
acid oxidase from Rhodotorula gracilis (Tm of 55.9◦C) [33,34].
Following the decrease in activity of SoAAO at 37◦C (on
aminoacetone as substrate, see below), a half-life of ∼15 min
was estimated.
ligand benzoate is bound at the active site of D-amino acid oxidase
[16]. A soluble apoprotein could not be obtained using various
procedures employed previously for several flavoproteins [35,36],
e.g. using KBr, urea or guanidine at different concentrations. The
fluorescence of the flavin cofactor is only slightly augmented (1.3-
fold) in the SoAAO holoenzyme form in comparison with that of
free FAD (at pH 8.0).
The far-UV CD spectrum of purified SoAAO shows that the
secondary structure of the purified holoprotein is composed
of ∼35% α-helices and ∼14% β-sheets (as predicted using
the K2D3 server) [37] (see Supplementary Figure S2), in
good agreement with the values determined from 3D structure
inspection (see Supplementary Figure S3A).
Typically, flavoprotein oxidases stabilize the FAD anionic
semiquinone species [38]. Upon SoAAO anaerobic photoreduc-
tion, in the presence of 5 mM EDTA and 0.5 μM 5-deaza-
FAD, no anionic semiquinone form of the flavin cofactor (in
20 mM Tris/HCl buffer, pH 8.0) could be detected (detection
threshold 5%). After 10 min of light irradiation, the fully reduced
flavin species was obtained (Figure 1B); notably, under similar
conditions, free FAD is fully reduced by shorter irradiation times,
i.e. 6.5 min. When oxygen was admitted to the solution, the
oxidized cofactor absorbance spectrum was promptly produced.
The midpoint redox potential of SoAAO was determined at
pH 8.0 using the xanthine/xanthine oxidase method and Benzyl
Viologen (E0 = − 359 mV) as both the mediator and reference
dye [24] (Figures 1C and 1D). A midpoint redox potential of
− 324 mV was estimated, a very low value compared with that for
free FAD ( − 207 mV at pH 7.0) [24] and for typical flavoprotein
oxidases, such as yeast D-amino acid oxidase ( − 130 mV) [21]
or glycine oxidase ( − 298 mV) [22].
Notably, no SoAAO enzymatic activity was detected with the
crude extract and the purified protein sample on neutral (i.e. L-
isoleucine or L-alanine), aromatic (i.e. L-tryptophan or L-tyrosine)
L-amino acids as substrates (at 50 mM concentration) using the
standard o-dianisidine/HRP-coupled assay at 25◦C. Under the
same assay conditions, only marginal activity (i.e. overnight
colour development) was detected on charged substrates (i.e. L-
aspartate or L-lysine).
The ability to quickly react with sulfite yielding an N(5)
covalent adduct (whose spectral properties resemble those of
the reduced flavin) distinguishes oxidases from other classes
of flavoproteins [12,38]. No spectral changes were observed
when SoAAO was added up to 290 mM sodium sulfite; the
absorbance spectrum of fully reduced flavin in SoAAO was only
obtained when adding simultaneously the reducing agents sodium
borohydride and sodium sulfite. Lack of reactivity towards sulfite
might result from the absence of a positive (partial) charge in
the microenvironment surrounding the flavin N(1)-C(2)=O locus
(that inductively promotes the process), or from hampered sulfite
diffusion to the cofactor (see below).
Spectral properties of recombinant SoAAO
Purified SoAAO in the oxidized form displays the typical
absorbance spectrum of FAD-containing flavoproteins, with two
absorbance maxima in the visible region at 465 and 370 nm,
and a shoulder at ∼490 nm (Figure 1A). A molar absorption
coefficient of 11990 M− 1·cm− 1 at 465 nm and a ratio of 6.1
between the absorbance values measured at 280 and 460 nm
were calculated. The FAD cofactor is not covalently linked to
the apoprotein moiety since it is released by heat treatment of the
holoenzyme. The absorption spectrum of the cofactor released
from the protein (see the broken line in Figure 1A) does not show
the 490 nm shoulder, indicating that this spectral feature is not
due to FAD chemical modification, but arises from a peculiar
hydrophobic environment surrounding the isoalloxazine ring of
the cofactor. A similar shoulder is evident when the aromatic
SoAAO activity with different substrates
According to [4], the activity of SoAAO was at first assayed
on the pure enzyme preparation by measuring the produced
H2O2 by the standard coupled o-dianisidine/HRP method at
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

392
G. Molla and others
Figure 1 Spectral properties of SoAAO
◦
(A) Absorbance spectrum of SoAAO in the oxidized and reduced state at pH 8.0 and 15 C. The reduced enzyme form was obtained by adding 20 mM sodium borohydride and 20 mM Na2SO3. The
absorbance spectrum of the FAD cofactor released by SoAAO following heat treatment is represented by the broken line. (B) Anaerobic photoreduction of SoAAO in the presence of 5 mM EDTA and
◦
0.5 μM 5-deaza-FAD (4 C, pH 8.0). The maximal amount of the reduced flavin species was detected after 10 min of photoillumination. (C) Determination of the midpoint redox potential (Em) of
SoAAO at pH 8.0. Selected spectra obtained during the course of the anaerobic reduction of SoAAO (18 μM) in 20 mM Tris/HCl (pH 8.0), containing 200 μM xanthine and 5 μM Benzyl Viologen
(E0’ = − 359 mV). Spectra were recorded 0, 20, 23, 26, 29, 31, 33, 36, 41, 44, 50, 65 min after the addition of 0.18 unit of xanthine oxidase. (D) Nernst plot: the concentration of the oxidized (ox)
and reduced (red) forms of SoAAO were determined after subtraction of the dye contribution. The concentration of the oxidized and reduced forms of Benzyl Viologen were determined at 600 nm
(where the contribution of the enzyme forms is negligible).
25◦C (see the Materials and methods section and above). In
our case, the appearance of a red-brownish colour, evidence of
H2O2 production, was apparent only after overnight incubation at
room temperature in the presence of 10 mM charged substrates
(e.g. L-lysine or L-aspartate). This indicates that the O2-dependent
oxidase activity of recombinant SoAAO with L-amino acids
is negligible. In contrast, when LAAO from Calloselasma
rhodostoma was used as a positive control, a specific activity
of 3.3 units/mg of protein was measured using L-leucine, in
good agreement with the value reported in [39]. Similarly, no
conversion of the oxidized form of the flavin cofactor of SoAAO
(12.5 μM) into the reduced state was observed following addition
of 4.8 mM L-aspartate (or 12.5 mM aminoacetone, see below)
under anaerobic conditions at 15◦C.
but using the o-dianisidine/HRP method and 130 μg of SoAAO:
this yielded a slightly higher specific activity with aminoacetone
(0.0298 units/mg of protein) (Figure 2A). A further higher specific
activity (0.0463 units/mg of protein) was determined using the
most sensitive discontinuous spectrophotometric method: this
latter value is very close to the one (0.0478 units/mg of protein)
reported in [14]. In comparison, the specific activity with L-
aspartate (using the same method) is ∼4.5-fold lower (0.013
units/mg of protein). No changes in activity with aminoacetone
as substrate was observed upon addition of 1 mM Fe2 + or
10 mM Mg2 + to the reaction mixture. The very low activity of
SoAAO prevented the determination of its kinetic parameters.
Indeed, no activity was detected with L-malate, oxaloacetate or
α-aminobutyrate (10 mM final concentration).
The SoAAO apparent specific activity value depends on
the amount of HRP used as coupling enzyme: 0.016, 0.042
and 0.069 units/mg of SoAAO using 0.04, 0.2 and 2 units/ml
HRP respectively (Figure 2B). This effect was obviously not
observed when the same coupled method was employed to assay
the reaction of a classical oxidase, such as glycine oxidase,
since the amount of HRP is broadly higher than the amount
required: a specific activity of 0.6–0.7 unit/mg glycine oxidase
was determined by using 0.04–2 units/ml HRP. The rate of H2O2
Recently, it has been proposed that SoAAO could be involved
in the oxidation of the small metabolite aminoacetone [14]. To test
this hypothesis, we measured the initial rate of H2O2 production
by mixing 28 μg of SoAAO with 25 mM aminoacetone (at
pH 8.0) using the continuous spectrophotometric coupled assay
in the presence of 4-aminoantipyrine/HRP [18]. Increasing
the temperature of assay to 37◦C, a specific activity with
aminoacetone of 0.019 units/mg of protein was measured. The
assay was replicated under the same experimental conditions,
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

Biochemical characterization of aminoacetone oxidase from S. oligofermentans
393
Figure 2 Enzymatic activity of SoAAO on aminoacetone in 20 mM Tris/HCl (pH 8.0) at 37◦C in the presence of 130 μg of SoAAO using 25 mM aminoacetone
as substrate (detected using the o-dianisidine/HRP method)
(A) Initial rate of production of H2O2 under vigorous (circles) or moderate (squares) shaking. Inset: pH-dependence of the initial activity (activity measured at pH 8.0 was taken as 100%). (B) Initial
rate of production of H2O2 in the presence of 2.5 (circles), 0.25 (squares) and 0.05 unit/ml (triangles) HRP with moderate shaking. For both panels, open symbols represents corresponding reactions
in the absence of SoAAO (negative controls).
the enzymatic reaction of SoAAO on aminoacetone has been
produced starting from 20 mM aminoacetone is ∼3.5-fold higher
performed in the presence of catalase (130 units): the intensity of
the peak at m/z 109 observed after 1 h is halved compared with
the reaction without catalase (Supplementary Figure S5F).
in the presence of SoAAO than in the control assay omitting
the flavoprotein (Figure 2B). Notably, aminoacetone is known
in aqueous solution to undergo phosphate-catalysed enolysation
fol•lo−wed by metal-catalysed oxidation, which is propagated by
O2 and enoyl+aminoacetone radicals, to produce methylglyoxal,
H2O2 and NH4 ions [40]. In order to avoid such reactivity, all
our activity assays were performed in Tris/HCl buffer. Indeed,
SoAAO activity was stimulated by vigorous agitation (i.e. higher
oxygenation of the reaction mixture) and was higher at pH 8
(Figure 2A, inset).
The 3D structure of SoAAO
The crystal structure of SoAAO has been solved in a binary
complex with the oxidized FAD cofactor (2.7 Å resolution, space
group P22121), and in a ternary complex with FAD and the glycine
methyl ester ligand, which mimics the aminoacetone substrate
(2.0 Å resolution, space group C2221) (Table 1). In view of the
higher resolution achieved, the ternary complex is taken here as a
reference for the analysis of the SoAAO structure, unless stated
otherwise.
The SoAAO molecule displays an elongated shape, the
longer axis being ∼85 Å, and the shorter axis being 43
Å. Its tertiary structure is characterized by the presence
of three compact domains (Figure 3A). A DALI analysis
(http://ekhidna.biocenter.helsinki.fi/dali_server) [41] indicates
that domain-1 (residues 1–192 and 328–391) is an α/β domain
with FAD/NAD(P)-binding Rossmann-like topology, consisting
of a five-stranded parallel β-sheet that is flanked on one side by
three α-helices (α1, α2 and α4) and on the other by a three-
stranded antiparallel β-sheet and one α-helix (α3) [42]; domain-2
(residues 193–257 and 319–327) has a β-barrel-like topology
reminiscent of domain-3 of EF-Tu (elongation factor Tu) [43];
domain-3 (residues 258–318) is mainly α-helical, and similar to
the helix–two turn–helix domain of type II DNA topoisomerase
VI subunit B [44] (Figure 3A).
The effect of SoAAO on aminoacetone was also investigated
by analysing the absorption spectrum in alkali of the 2,4-
dinitrophenylhydrazone derivative of the reaction products [19].
The absorbance maximum of the mixture after 30 min of
reaction does not correspond to that obtained for the standard
methylglyoxal (508 compared with ∼518 nm; see Supplementary
Figure S4). Notably, as expected, the initial activity value
determined at 2.5 mM aminoacetone (15 min at 37◦C) is ∼0.001–
0.003 units/mg SoAAO, a value approximately 10-fold lower than
that obtained at 25 mM of the same substrate (0.019 units/mg of
protein, see above).
MS analysis of reaction products
SoAAO (0.150 units) was reacted with 21.2 mM aminoacetone
in 20 mM Tris/HCl at pH 8.0 and 37◦C and the reaction products
were analysed by MS (positive ions). After 1 h of reaction,
a new peak at m/z 109 appeared which corresponds to 2,5-
dimethylpyrazine (molecular mass of 108 g/mol, m/z 109), i.e.
the oxidation product of 3,6-dimethyl-2,5-dihydropyrazine (see
Supplementary Figures S5A and S5B). After 4 h, the intensity
of the peak at m/z 109 was higher than the peak at m/z 111 and
the peak of free aminoacetone disappeared (see Supplementary
Figure S5C). It should be pointed out that the formation of 2,5-
dimethylpyrazine also occurs spontaneously, but to a very low
extent (as can be observed in the control reaction shown in
Supplementary Figures S5D and S5E). Notably, the peak corres-
ponding to methylglyoxal (molecular mass of 72 g/mol, m/z 73)
was never observed, as well as its dimeric and trimeric oligomers.
In order to evaluate the effect of H2O2 on the reaction products,
The FAD-binding site is located in domain-1. Thirty-nine
residues are directly involved in FAD binding (distances below 4.5
Å), of which 11 (Ala14, Glu33, Lys34, Asn35, Ile134, Glu362, Gly371
,
Gly372, Phe373, Asn374 and Ile375) establish electrostatic or polar
interactions with the cofactor. Only three of the latter interactions
involve side-chain atoms (Glu33, Asn35 and Glu362), the others
being due to main-chain atoms. Overall, the isoalloxazine ring is
quite solvent-exposed, especially at the N5 reactive site, located
at the interface between domain-1 and domain-2 (Figure 3B). It
establishes hydrogen bonds with the main-chain nitrogen atom
of Asn374 and of Ile375 (FAD O2 atom), and of Gly371 (FAD O4
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

394
G. Molla and others
Figure 3 Crystal structure of SoAAO
(A) Schematic view of the spatial arrangement of the SoAAO 3-domain fold (domains are displayed in green, cyan and red respectively); α-helices, β-strands and coils are represented by helical
ribbons, arrows and ropes respectively. The FAD cofactor bound to SoAAO domain-1 (Rossmann-like topology) is shown in stick representation (yellow colour). (B) Protein–FAD interactions and
active site cavity of SoAAO. The FAD molecule (yellow) and the protein residues (grey) located at the active-site cavity relevant for FAD binding are shown in stick representation and water molecules
˚
are represented as red spheres (0.5 van der Waals radius). Hydrogen bonds are shown by broken lines; numbers represent lengths in A. The SoAAO active-site cavity is located at the interface
between domain-1 (green) and domain-2 (cyan). (C) The molecular surface of SoAAO, coloured according to its potential (blue, positive; red, negative). The entry of the active-site cavity, showing
the isoalloxazine ring of FAD, is clearly visible. For the sake of clarity, in (B) and (C), the orientation of the SoAAO structure is slightly different. See also Supplementary Figure S3.
atom), while displaying van der Waals interactions with Thr45,
Tyr167, Phe333, Thr370 and Ile375 side chains (Figure 3C).
and on a good match of the FAD-binding site in domain-1 (see
Supplementary Table S1). Nevertheless, structural comparison
of the protein backbones shows that SoAAO differs markedly
from the other three related proteins, both in the orientation
of domain-1 relative to domains-2 and -3, and in the precise
positioning of the FAD molecule within domain-1 (Figures 4A
and 4C). The tertiary structure of SoAAO adopts a more ‘closed’
conformation, where domain-2 and domain-3 move together
as a rigid body (Figure 4B) towards domain-1, thus bringing
domain-2 in closer contact to domain-1 relative to 2I0Z, 3V76
and 2GQF tertiary structures (Figure 4A). In SoAAO, few key
electrostatic interactions contributed to hook domain-2 to domain-
1, specifically Asp213 or Asp226 to His100, Glu194 and His369, and the
hydrophobic interaction of the Phe333 side chain which fits in a
pocket lined by Tyr167, Pro196 and Leu240. In 3V76 and 2GQF, site-
specific substitutions prevent the formation of such interactions.
In the 2I0Z structure, Tyr395 is located at the site corresponding to
His369 in SoAAO, and the salt bridge with Glu195 (corresponding
to SoAAO Glu194) is therefore not present. Furthermore, despite
the sequence conservation, the Phe359 side chain (corresponding to
SoAAO Phe333) does not fit into the pocket lined by Val168, Pro197
and Leu250 (corresponding to SoAAO Tyr167, Pro196 and Leu240),
but moves into the FAD-binding site pocket; the 2I0Z Phe359 side
Concerning the quaternary structure, the program PISA
(http://www.ebi.ac.uk/pdbe/pisa/) [45] did not highlight specific
interactions between SoAAO monomers in the crystal that could
result in the formation of a stable quaternary assembly, in both
the P22121 (four protein molecules in the asymmetric unit)
and the C2221 crystal forms (two protein molecules in the
asymmetric unit); this is in agreement with the results of the
gel-permeation experiments on the purified protein (see above).
Overall, the SoAAO tertiary structure resembles that of
a putative NAD(FAD)-utilizing dehydrogenase from Bacillus
cereus (PDB code 2I0Z; DALI Z-score of 49.6, residue identity
of 46%), of a flavoprotein from Sinorhizobium meliloti (PDB
code 3V76; DALI Z-score of 47.2, residue identity of 30%) and
of the HI0933 protein from H. influenzae (PDB code 2GQF;
DALI Z-score of 46.5, residue identity of 30%) (Figures 4A and
4B, and see Supplementary Figure S3B). Unfortunately, all of
these proteins structurally related to SoAAO have been analysed
within structural genomics initiatives, and neither literature
nor functional data are presently available. The structural
relationships among the four proteins depend on an excellent
conservation of secondary-structure elements in all three domains
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

Biochemical characterization of aminoacetone oxidase from S. oligofermentans
395
Figure 4 Structural comparison of SoAAO and three structurally related proteins
(A) SoAAO (green) is in a more ‘closed’ conformation, with domain-2 and domain-3 that move as a rigid body towards domain-1 relative to the putative NAD(FAD)-utilizing dehydrogenases from
B. cereus (PDB code 2I0Z, orange), the flavoprotein from S. meliloti (PDB code 3V76, cyan), and the HI0933 protein from H. influenzae (PDB code 2GQF, magenta). (B) A structural comparison
produced by superimposing domain-2 and domain-3 of each protein. The structural superimposition of the isolated domain-2 and domain-3 shows a strong conservation of the tertiary structure and
the relative orientation of these domains (see also Supplementary Table S1). (C) Orientation of the FAD isoalloxazine ring. Structural superposition of SoAAO FAD-binding site on other structurally
related proteins. (D) Model of the N5-sulfo-FAD molecule bound to SoAAO based on the isoalloxazine ring of a N5-sulfo-flavin mononucleotide molecule (PDB code 1QCW). Relevant residues are
shown in stick representation and labelled.
chain matches the position of 3V76 Glu339 and 2GQF Glu342. As
a consequence, the FAD isoalloxazine ring in SoAAO adopts a
location that does not match those adopted by the cofactor in
the 2I0Z, 3V76 and 2GQF structures, with a maximum shift of
∼2.2 Å at the FAD O4 atoms. Such different orientation of the
FAD isoalloxazine ring is also related to the substitution of serine
for SoAAO Thr45 in 2I0Z, 3V76 and 2GQF (Figure 4C), and by
a backbone shift of the helical region corresponding to SoAAO
residues 42–48.
the active site. The small compound glycine methyl ester (O-
methylglycine, a molecule significantly more stable in solution
than aminoacetone) binds to SoAAO, altering the far-UV
CD spectrum (whereas no changes in the UV–visible absorption
spectrum were observed). Plotting the changes at 208 nm of
the far-UV CD spectra as a function of glycine methyl ester
+
concentration, a Kd of 2.80
Supplementary Figure S6).
0.46 mM was calculated (see
−
Co-crystallization experiments were performed by incubating
SoAAO overnight at 4◦C with a 10-fold molar excess of glycine
methyl ester, and resulted in crystals belonging to the C2221 space
group (two protein molecules in the asymmetric unit) instead
of the native P22121 space group (four protein molecules in
the asymmetric unit). Despite such crystal packing difference,
binding of glycine methyl ester does not alter the SoAAO tertiary
structure (RMSD range 0.46–1.03 Å between complex and native
structures, depending on the superimposed protein chains), nor
promote quaternary assembly of the protein. In both subunits
SoAAO in complex with glycine methyl ester
Crystals of SoAAO in complex with aminoacetone could never
be grown, probably because of the intrinsic instability and/or
enzymatic conversion of aminoacetone under the crystallization
conditions. For this reason, we looked for putative SoAAO
ligands that could mimic the aminoacetone-binding mode at
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

396
G. Molla and others
Figure 5 Binding of O-methylglycine to SoAAO
(A and C) Substrate-binding pocket in SoAAO subunit A. (B and D) Substrate-binding pocket in SoAAO subunit B. The bound O-methylglycine (grey: mGly), the FAD molecule (yellow) and the
relevant protein residues (green) are shown in stick representation and labelled. Water molecules are shown as red balls. Broken lines indicate hydrogen bonds. The 2Fo–Fc electron density
(contoured at 1σ) for the bound ligand is shown as a grey mesh. For clarity, Gly46, whose carbonyl oxygen is hydrogen-bonded to the amino group of mGly2 in subunit B, has been omitted from (B)
since it overlaps the FAD isoalloxazine ring. For the same reason, some water molecules are not shown. (E) Comparison of the chemical structures of O-methylglycine and aminoacetone.
of the C2221 crystal form, the glycine methyl ester molecule
binds at the interface between domain-1 and domain-2, in a
cavity located just on top of the FAD isoalloxazine ring, with
the amino group oriented towards the isoalloxazine N5 and the
O4 atoms (distances of 2.8 Å and 2.9 Å respectively in both
subunits) (Figure 5). The ligand methyl group is hosted in a
mostly hydrophobic pocket lined by Ser235, Pro237, Leu240, Phe333
and Thr370, whereas the carbonyl oxygen is oriented towards the
bulk solvent region in subunit A, and towards the interior of
the protein in subunit B (with a rotation along the peptide axis
of ∼50◦). In both subunits, the carbonyl oxygen of the ligand
molecule is hydrogen-bonded only with water molecules hosted
within the FAD isoalloxazine ring-binding pocket. Interestingly,
at the interface between domain-1 and domain-2, the presence
of Arg49 and Arg102 side chains (kept in position by their salt-
bridge interactions with Glu98 and Asp99 respectively) generates
a positively charged pocket suitable for binding of a sulfate
ion, originating from the crystallization solution, in subunit A
(Figures 5A and 5C), and for a second O-methylglycine molecule
in subunit B (Figures 5B and 5D). The presence of a sulfate-
trapping pocket at the entrance of the FAD isoalloxazine ring-
binding cavity could help to explain the absence of reactivity
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

Biochemical characterization of aminoacetone oxidase from S. oligofermentans
397
activity ∼0.048 units/mg of protein) [14]. Genetic analysis
based on double deletion of the aaoSo and mutT (encoding a
pyrophosphohydrolase which removes mutagens such as 8-oxo-
dGTP) genes, and on the observation that inactivation of aaoSo
caused a concentration increase of the pro-oxidant molecule
aminoacetone and of the cellular ROS (reactive oxygen species)
stress, indicated that SoAAO represents a novel antioxidant
defence system in S. oligofermentans [14].
In the present study, we designed a synthetic optimized SoAAO-
encoding cDNA and expressed the recombinant protein in E. coli
to up to 19 mg of protein/litre of fermentation broth; previous
expression levels were not reported [4,14]. Purified SoAAO is a
FAD-containing protein which tightly binds the flavin cofactor:
the protein is purified as a holoenzyme, FAD is not released during
gel-permeation chromatography, and the apoprotein form could
never be produced even using standard procedures previously
employed for several flavoenzymes [35,36]. SoAAO does not
present the classical properties of the oxidase-dehydrogenase
class of flavoproteins [38]. In fact: (i) the FAD cofactor is
not reduced by L-amino acids or aminoacetone under anaerobic
conditions at 15◦C; the reduced flavin form was only achieved
chemically (by adding together sulfite and borohydride-reducing
agents) or photochemically (in the latter case, reduced FAD
promptly reacted with dioxygen); (ii) it does not produce a
semiquinone species when light-irradiated even in the presence
of redox mediators (the midpoint redox potential for the two-
electron transfer was − 324 mV); (iii) it does not react with
sulfite to form a flavin N(5)–sulfite adduct, a reaction commonly
present in flavo-oxidases [46].
Figure 6 Structural comparison between the SoAAO and LAAO (from C.
rhodostoma, PDB code 2IID) substrate-binding sites
From a structural point of view, SoAAO belongs to a novel
protein family. Four bacterial proteins, whose physiological
role has not been elucidated, share similar tertiary structure
composed of three domains: an α/β domain corresponding to
the FAD-binding domain, a β-domain which partially modulates
accessibility to the coenzyme, and an additional α-domain.
Phylogenetic analysis confirms that SoAAO homologues are
present in microorganisms belonging to the genus Streptococcus
[5] and that they are also widespread in bacteria belonging to
the Firmicutes and Proteobacteria phyla with a sequence identity
between SoAAO and homologues ranging from 30% to 45%
(considering the overall sequence). SoAAO homologues were also
identified in Viridiplantae (green plants) although with a lower
degree of identity with SoAAO (∼25%) (see Supplementary
Figure S7).
The O-methylglycine molecule bound to SoAAO provides
a useful molecular scaffold to mimic the binding mode of
aminoacetone, which differs from O-methylglycine only in the
substitution of the ester group by a methyl ketone (Figure 5E).
From a structural standpoint, SoAAO does not resemble known
LAAOs, since classic LAAOs are homodimeric holoenzymes (the
main exception being those from E. coli and Sulfolobus tokodaii)
[9], whereas SoAAO is monomeric, and LAAO belongs to the N-
terminal FAD-bound reductase family while SoAAO is a three-
domain protein. LAAOs evolved an active-site arrangement to
bind L-amino acids and to discriminate between the two different
enantiomers, the so-called four-location model [7]. In all cases,
the substrate binds to the LAAO active site on the re-face
of the isoalloxazine moiety of FAD, with the major interactions
being: (i) a salt bridge between the α-COO− of the amino acid
and the guanidinium group of an arginine residue located close
to the pyrimidine-derivative ring of the isoalloxazine moiety;
this interaction is strengthened by an additional hydrogen bond
with the hydroxy group of a tyrosine residue; (ii) a hydrogen
bond between the α-amino group of the L-amino acid and the
main-chain C=O of the protein belonging to a small residue
SoAAO and LAAO residues are shown in green and grey respectively, with SoAAO FAD in yellow,
LAAO FAD in grey and LAAO-bound substrate L-phenylalanine in brown. Broken lines indicate
relevant hydrogen-bond interactions. For clarity, the Cα backbone of the SoAAO helix α4 and
of the 44–57 and 365–387 regions are indicated in ribbon representation, together with the
corresponding LAAO regions. The close contact between SoAAO Thr370 Cα and the amino group
of the L-phenylalanine bound to LAAO is indicated by an arrow.
of SoAAO with sulfite. Furthermore, the model of SoAAO in
complex with N5-sulfo-FAD (generated on the basis of the
isoalloxazine ring of a N5-sulfo-flavin mononucleotide; PDB
code 1QCW), shows that the close vicinity of the SoAAO 310-
helical 42–48 segment to the FAD N5 reactive site would provide
steric clash of the bound sulfite with Thr45 (Figure 4D), thus
preventing sulfite binding. In addition, this unusual feature of
a flavo-oxidase is also supported by the absence of positively
charged groups in the proximity of the N1-C2=O locus of the
isoalloxazine ring, whose presence would inductively promote
the process. A partial positive charge may, however, be provided
by the dipole associated to the N-terminus of the nearby helix
α4. Intriguingly, almost no similarity can be detected between the
active-site structural organization and the ligand-binding mode of
SoAAO with respect to other known LAAOs [7,9] (Figure 6).
DISCUSSION
In search of biotechnological applications of LAAO activities,
we focused on the flavoenzyme produced by S. oligofermentans,
a member of the ‘mitis’ group of oral streptococci involved in
the initial steps of dental biofilm formation. This FAD-containing
enzyme was initially reported to oxidize seven different amino
acids, with a specific activity of 0.47 unit/mg of protein (for L-
aspartate or L-tryptophan), and to be overexpressed in E. coli
[4]. The original activity values for L-amino acids reported in
[4] were subsequently revised, and aminoacetone was reported
as the preferred substrate for purified SoAAO (with a specific
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

398
G. Molla and others
(glycine or alanine); (iii) the third anchor point is formed by the
region of the active site located above the hydrophobic moiety of
the FAD isoalloxazine ring: this latter region accommodates the
substrate side chain and represents the main structural determinant
of substrate specificities observed in the various LAAOs [9]. In
contrast, the overall structure of the SoAAO active site differs
markedly from those of LAAOs. In particular, the Cα-backbone
of the β5–α4 region (residues 365–373) has a deep impact on the
reshaping of the SoAAO ligand-binding site, also affecting the
structure of the nearby βA–αB region (residues 44–55) (Figure 6).
As a consequence, the tip of the β5–α4 loop partly overlaps with
the LAAO substrate-binding site, and the LAAO small residue
(glycine or alanine) whose carbonyl oxygen would hydrogen-
bond to the α-amino group of the L-amino acid is displaced
from the SoAAO substrate-binding site. Furthermore, the arginine
residue that in LAAO stabilizes the α-COO− of the substrate
amino acid is not conserved in the SoAAO βA–αB region, being
structurally replaced by Arg102, which in SoAAO is involved
in shaping the sulfate-trapping pocket (Figure 5). Finally, the
tyrosine side chain that in LAAO contributes to stabilize the bound
substrate via hydrogen-bonding, is not present in SoAAO, being
The ability of SoAAO to bind two molecules of the substrate
analogue O-methylglycine ligand, as shown in Figure 5(B), is held
to facilitate the condensation reaction. Interestingly, the position
occupied by Arg49 and Arg102 in subunit B (Figures 5B and 5D)
allows not only the binding of a second molecule of the substrate
analogue O-methylglycine in the active site of SoAAO, but also
proper positioning of the amino group of the second substrate
molecule with respect to the carbonyl group of the (first) substrate
molecule bound next the isoalloxazine flavin ring. We do not
exclude that methylglyoxal and/or H2O2 produced by the non-
enzymatic spontaneous oxidation of aminoacetone might promote
the reaction of SoAAO, since catalase negatively affected the rate
of aminoacetone oxidation (see Supplementary Figure S5F).
The low specific activity of SoAAO (the highest value was
47.8 nmol/min per mg of protein) as compared with known amino
acid oxidases (up to 120 μmol/min per mg of yeast DAAO,
2 and 60 μmol/min per mg of E. coli LAAO and glutamate
oxidase respectively) [9,16], poses an open question about the
physiological role of this protein. The activity of SoAAO appears
to speed up the elimination of aminoacetone in the cell: the
aminoacetone levels in S. oligofermentans doubled when the aao
gene was deleted (from 77 pmol/mg per cell for the wild-type
strain to 147 pmol/mg per cell for the ꢀaaoSo mutant strain),
nevertheless reaching a modest concentration [14]. Although
SoAAO activity is low, it should be sufficient to control the
observed increase in aminoacetone: the very low enzymatic
activity of SoAAO with L-amino acids should then just represent
a promiscuous activity. Also, we cannot exclude the possibility
that SoAAO might play a role as binding protein that senses
aminoacetone and, through the recruitment of further binding
partners (through the non-catalytic domain-2 and domain-3),
would activate/stimulate cell defence mechanisms.
replaced by Phe333
.
The role of the SoAAO domain-2 and domain-3 is unclear.
No specific function is associated with the two protein domain
modules and, although they form a rigid body conserved among
homologous proteins (Figure 4B), their orientation relative to
the nucleotide-binding domain (domain-1) may vary depending
on the protein considered (Figure 4A). It is interesting to note
that in SoAAO, although not promoting dimerization, domain-
2 and domain-3 are, however, productively involved in crystal
contacts, especially at the interface of the α-helical domain-3 (see
Supplementary Figure S8), thus suggesting thatthe apicaldomain-
3 might represent a structural element involved in recruiting
additional partner proteins.
In conclusion, SoAAO represents a member of a novel family
of bacterial flavoproteins composed of three domains, unrelated
to LAAO activities, and involved in specific functions, such
as participating in antioxidant defence from the pro-oxidant
metabolite aminoacetone in S. oligofermentans by yielding
pyrazine derivatives instead of methylglyoxal.
Characterization of the kinetic properties confirmed a very
low SoAAO activity with aminoacetone, and a marginal activity
with L-amino acids, as reported in a recent report [14], and
the production of H2O2 when aminoacetone (or L-aspartate) are
used as substrates. We demonstrated that SoAAO activity on
aminoacetone depended on the amount of HRP used as coupling
enzyme (Figure 2A) and that H2O2 production is observed also
in the absence of the flavoprotein. This observation reconciles
the previous specific activity values reported in the literature:
the values reported in [4] were determined using 100-fold
higher HRP than in [14] (5 and 0.05 units/ml respectively).
The amount of HRP seems to play a crucial role in the
oxidation of aminoacetone by SoAAO: aminoacetone has been
AUTHOR CONTRIBUTION
Marco Nardini grew the crystals and solved the structure of the protein; Gianluca Molla
and Paolo Motta performed the biochemical characterization; Paolo D’Arrigo and Walter
Panzeri carried out the MS analysis of the reaction products; Loredano Pollegioni and
Gianluca Molla conceived the project and wrote the paper. All authors have read
and approved the paper.
reported to undergo enzymatic (by SSAO) and Fen + /Cu2 +
-
catalysed oxidation to methylglyoxal, H2O2 and ammonia with
O2 consumption [47]. In particular, a haem-containing protein
such as ferricytochrome c initiated by one-electron transfer, O2-
dependent oxidation of aminoacetone to yield the aldimine of
methylglyoxal [NH=CH-C(=O)-CH3], which then hydrolyses to
methylglyoxal [CH3-C(=O)-C(=O)H], ammonia and H2O2 [40].
We propose that a similar role might be played by HRP in the
coupled activity assay, thus accounting for the amount of H2O2
produced in the absence of SoAAO.
MS and 2,4-dinitrophenylhydrazone-derivative analyses of
the reaction products of SoAAO on aminoacetone showed
that no methylglyoxal is produced and that the main product
is 2,5-dimethylpyrazine. Accordingly, and differently from
that proposed for SSAO that converts aminoacetone into
methylglyoxal [15], we propose that condensation of two amino-
acetone molecules yields 3,6-dimethyl-2,5-dihydropyrazine that
is subsequently oxidized to 2,5-dimethylpyrazine (Scheme 1).
ACKNOWLEDGEMENTS
We thank Sandro Ghisla and Martino Bolognesi for helpful discussion.
FUNDING
This work was supported by grants from Ministero dell’Istruzione, dell’Universita` e della
RicercaScientifica(MIUR)FondodiAteneoperlaRicercatoL.P. andG.M. Weacknowledge
the support from Consorzio Interuniversitario per le Biotecnologie (CIB) and Centro Grandi
Attrezzature dell’Universita` degli Studi dell’Insubria.
REFERENCES
1
Nicolas, G. G. and Lavoie, M. C. (2011) Streptococcus mutans and oral streptococci in
dental plaque. Can. J. Microbiol. 57, 1–20 CrossRef PubMed
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

Biochemical characterization of aminoacetone oxidase from S. oligofermentans
399
2
3
Tong, H. C., Gao, X. J. and Dong, X. Z. (2003) Streptococcus oligofermentans sp nov., a
novel oral isolate from caries-free humans. Int. J. Syst. Evol. Microbiol. 53, 1101–1104
CrossRef PubMed
24 Massey, V. A. (1991) simple method for the determination of redox potentials. In In
Flavins and Flavoproteins (Curti, B., Ronchi, S. and Zanetti, G., eds), pp. 59–66, Walter
de Gruyter & Co., Berlin
25 Pollegioni, L., Porrini, D., Molla, G. and Pilone, M. S. (2000) Redox potentials and their
pH dependence of D-amino acid oxidase of Rhodotorula gracilis and Trigonopsis
variabilis. Eur. J. Biochem. 267, 6624–6632 CrossRef PubMed
26 Kabsch, W. (2010) XDS. Acta Crystallogr. D Biol. Crystallogr. 66, 125–132
CrossRef PubMed
27 Evans, P. (2006) Scaling and assessment of data quality. Acta Crystallogr. D Biol.
Crystallogr. 62, 72–82 CrossRef PubMed
Tong, H., Chen, W., Merritt, J., Qi, F., Shi, W. and Dong, X. (2007) Streptococcus
oligofermentans inhibits Streptococcus mutans through conversion of lactic acid into
inhibitory H2O2: a possible counteroffensive strategy for interspecies competition. Mol.
Microbiol. 63, 872–880 CrossRef PubMed
Tong, H. C., Chen, W., Shi, W. Y., Qi, F. X. and Dong, X. Z. (2008) SO-LAAO, a novel
L-amino acid oxidase that enables Streptococcus oligofermentans to outcompete
Streptococcus mutans by generating H2O2 from peptone. J. Bacteriol. 190, 4716–4721
CrossRef PubMed
Boggs, J. M., South, A. H. and Hughes, A. L. (2012) Phylogenetic analysis supports
horizontal gene transfer of L-amino acid oxidase gene in Streptococcus oligofermentans.
Infect. Genet. Evol. 12, 1005–1009 CrossRef PubMed
Wierenga, R. K., Terpstra, P. and Hol, W. G. J. (1986) Prediction of the occurrence of the
ADP-binding β-α-β-fold in proteins, using an amino-acid-sequence fingerprint. J. Mol.
Biol. 187, 101–107 CrossRef PubMed
Pawelek, P. D., Cheah, J., Coulombe, R., Macheroux, P., Ghisla, S. and Vrielink, A. (2000)
The structure of L-amino acid oxidase reveals the substrate trajectory into an
enantiomerically conserved active site. EMBO J. 19, 4204–4215
CrossRef PubMed
Geueke, B. and Hummel, W. (2002) A new bacterial L-amino acid oxidase with a broad
substrate specificity: purification and characterization. Enzyme Microb. Technol. 31,
77–87
4
28 Leslie, A. G. W. (2003) MOSFLM 6.2.3 User Guide, MRC Laboratory of Molecular
Biology, Cambridge
5
6
7
29 Storoni, L. C., McCoy, A. J. and Read, R. J. (2004) Likelihood-enhanced fast rotation
functions. Acta Crystallogr. D Biol. Crystallogr. 60, 432–438 CrossRef PubMed
30 Murshudov, G. N., Vagin, A. A. and Dodson, E. J. (1997) Refinement of macromolecular
structures by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr. 53,
240–255 CrossRef PubMed
31 Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta
Crystallogr. D Biol. Crystallogr. 60, 2126–2132 CrossRef PubMed
32 Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J.,
Murray, L. W., Richardson, J. S. and Richardson, D. C. (2010) MolProbity: all-atom
structure validation for macromolecular crystallography. Acta Crystallogr. D Biol.
Crystallogr. 66, 12–21 CrossRef PubMed
33 Pollegioni, L., Molla, G., Sacchi, S., Rosini, E., Verga, R. and Pilone, M. S. (2008)
Properties and applications of microbial D-amino acid oxidases: current state and
perspectives. Appl. Microbiol. Biotechnol. 78, 1–16 CrossRef PubMed
34 Caldinelli, L., Pedotti, M., Motteran, L., Molla, G. and Pollegioni, L. (2009) FAD binding
in glycine oxidase from Bacillus subtilis. Biochimie 91, 1499–1508 CrossRef PubMed
35 Hefti, M. H., Vervoort, J. and van Berkel, W. J. (2003) Deflavination and reconstitution of
flavoproteins. Eur. J. Biochem. 270, 4227–4242 CrossRef PubMed
36 Caldinelli, L., Iametti, S., Barbiroli, A., Bonomi, F., Fessas, D., Molla, G., Pilone, M. S.
and Pollegioni, L. (2005) Dissecting the structural determinants of the stability of
cholesterol oxidase containing covalently bound flavin. J. Biol. Chem. 280,
22572–22581 CrossRef PubMed
8
9
Pollegioni, L., Motta, P. and Molla, G. (2013) L-Amino acid oxidase as biocatalyst: a
dream too far? Appl. Microbiol. Biotechnol. 97, 9323–9341 CrossRef PubMed
10 Du, X. Y. and Clemetson, K. J. (2002) Snake venom L-amino acid oxidases. Toxicon 40,
659–665 CrossRef PubMed
11 Umhau, S., Pollegioni, L., Molla, G., Diederichs, K., Welte, W., Pilone, M. S. and Ghisla,
S. (2000) The X-ray structure of D-amino acid oxidase at very high resolution identifies
the chemical mechanism of flavin-dependent substrate dehydrogenation. Proc. Natl.
Acad. Sci. U.S.A. 97, 12463–12468 CrossRef PubMed
12 Pollegioni, L., Piubelli, L., Sacchi, S., Pilone, M. S. and Molla, G. (2007) Physiological
functions of D-amino acid oxidases: from yeast to humans. Cell. Mol. Life Sci. 64,
1373–1394 CrossRef PubMed
13 Pollegioni, L. and Molla, G. (2011) New biotech applications from evolved D-amino acid
oxidases. Trends Biotechnol. 29, 276–283 CrossRef PubMed
37 Louis-Jeune, C., Andrade-Navarro, M. A. and Perez-Iratxeta, C. (2012) Prediction of
protein secondary structure from circular dichroism using theoretically derived spectra.
Proteins 80, 374–381 CrossRef
14 Zhou, P., Liu, L., Tong, H. and Dong, X. (2012) Role of operon aaoSo-mutT in antioxidant
defense in Streptococcus oligofermentans. PLoS ONE 7, e38133 CrossRef PubMed
15 Lyles, G. A. and Chalmers, J. (1995) Aminoacetone metabolism by
semicarbazide-sensitive amine oxidase in rat aorta. Biochem. Pharmacol. 49, 416–419
CrossRef PubMed
16 Molla, G., Sacchi, S., Bernasconi, M., Pilone, M. S., Fukui, K. and Pollegioni, L. (2006)
Characterization of human D-amino acid oxidase. FEBS Lett. 580, 2358–2364
CrossRef PubMed
38 Massey, V. and Hemmerich, P. (1980) Active-site probes of flavoproteins. Biochem. Soc.
Trans. 8, 246–257 PubMed
39 Macheroux, P., Seth, O., Bollschweiler, C., Schwarz, M., Kurfurst, M., Au, L. C. and
Ghisla, S. (2001) L-Amino-acid oxidase from the Malayan pit viper Calloselasma
rhodostoma: comparative sequence analysis and characterization of active and inactive
forms of the enzyme. Eur. J. Biochem. 268, 1679–1686 CrossRef PubMed
40 Sartori, A., Mano, C. M., Mantovani, M. C., Dyszy, F. H., Massari, J., Tokikawa, R.,
Nascimento, O. R., Nantes, I. L. and Bechara, E. J. (2013) Ferricytochrome (c) directly
oxidizes aminoacetone to methylglyoxal, a catabolite accumulated in carbonyl stress.
PLoS ONE 8, e57790 CrossRef PubMed
17 Volonte`, F., Pollegioni, L., Molla, G., Frattini, L., Marinelli, F. and Piubelli, L. (2010)
Production of recombinant cholesterol oxidase containing covalently bound FAD in
Escherichia coli. BMC Biotechnol. 10, 33 CrossRef PubMed
41 Holm, L. and Rosenstro¨m, P. (2010) Dali server: conservation mapping in 3D. Nucleic
18 Molla, G., Piubelli, L., Volonte, F. and Pilone, M. S. (2012) Enzymatic detection of
D-amino acids. In Unnatural Amino Acids: Methods and Protocols (Pollegioni, L. and
Servi, S., eds), pp. 273–289, Humana Press, Totowa CrossRef
Acids Res. 38, W545–W549 CrossRef PubMed
42 Dym, O. and Eisenberg, D. (2001) Sequence-structure analysis of FAD-containing
proteins. Protein Sci. 10, 1712–1728 CrossRef PubMed
19 Green, M. L. and Lewis, J. B. (1968) The oxidation of aminoacetone by a species of
Arthrobacter. Biochem. J. 106, 267–270 PubMed
43 Song, H., Parsons, M. R., Rowsell, S., Leonard, G. and Phillips, S. E. (1999) Crystal
structure of intact elongation factor EF-Tu from Escherichia coli in GDP conformation at
˚
20 Massey, V. and Hemmerich, P. (1978) Photoreduction of flavoproteins and other biological
compounds catalyzed by deazaflavins. Biochemistry 17, 9–17 CrossRef PubMed
21 Harris, C. M., Molla, G., Pilone, M. S. and Pollegioni, L. (1999) Studies on the reaction
mechanism of Rhodotorula gracilis D-amino-acid oxidase: role of the highly conserved
Tyr-223 on substrate binding and catalysis. J. Biol. Chem. 274, 36233–36240
CrossRef PubMed
2.05 A resolution. J. Mol. Biol. 285, 1245–1256 CrossRef PubMed
44 Corbett, K. D. and Berger, J. M. (2003) Structure of the topoisomerase VI-B subunit:
implications for type II topoisomerase mechanism and evolution. EMBO J. 22, 151–163
CrossRef PubMed
45 Krissinel, E. and Henrick, K. (2005) Detection of protein assemblies in crystals. Lect.
Notes Comput. Sci. 3695, 163–174 CrossRef
22 Job, V., Marcone, G. L., Pilone, M. S. and Pollegioni, L. (2002) Glycine oxidase from
Bacillus subtilis: characterization of a new flavoprotein. J. Biol. Chem. 277, 6985–6993
CrossRef PubMed
46 Massey, V. and Ganther, H. (1965) On the interpretation of the absorption spectra of
flavoproteins with special reference to D-amino acid oxidase. Biochemistry 4, 1161–1173
CrossRef PubMed
23 Caldinelli, L., Iametti, S., Barbiroli, A., Bonomi, F., Piubelli, L., Ferranti, P., Picariello, G.,
Pilone, M. S. and Pollegioni, L. (2004) Unfolding intermediate in the peroxisomal
flavoprotein D-amino acid oxidase. J. Biol. Chem. 279, 28426–28434 CrossRef PubMed
47 Dutra, F., Knudsen, F. S., Curi, D. and Bechara, E. J. H. (2001) Aerobic oxidation of
aminoacetone, a threonine catabolite: Iron catalysis and coupled iron release from ferritin.
Chem. Res. Toxicol. 14, 1323–1329 CrossRef PubMed
Received 29 July 2014/26 September 2014; accepted 30 September 2014
Published as BJ Immediate Publication 30 September 2014, doi:10.1042/BJ20140972
ꢀ
c
ꢀ
c
The Authors Journal compilation 2014 Biochemical Society

Copyright of Biochemical Journal is the property of Portland Press Ltd. and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.