Báo cáo khoa học: Cholesterol oxidase: biochemistry and structural features ppt

MINIREVIEW
Cholesterol oxidase: biochemistry and structural features
Alice Vrielink
1
and Sandro Ghisla
2,3
1 School of Biomedical, Biomolecular and Chemical Sciences, University of Western Australia, Crawley, Australia
2 Department of Biotechnology and Molecular Sciences, University of Insubria, Varese, Italy
3 Fachbereich Biologie, University of Konstanz, Konstanz, Germany
Introduction
Cholesterol oxidases (ChOxs) are bacterial flavopro-
teins that catalyze the first step in the degradation of
cholesterol. They contain a single molecule of FAD as
the redox cofactor and, in one case, possibly FMN.
ChOx has emerged as a useful biotechnological tool
employed, for example, for the determination of
serum cholesterol levels (see accompanying reviews
[1,2]). It possesses larvicidal and insecticidal activity
[3,4] and might be a factor important for infection
with the pathogenic bacterium Rhodococcus equi [5,6].
The enzymatic activity is unique to bacteria and thus
ChOx constitutes a potential target for new anti-
biotics.
The 3D structures of two types of ChOx that show
completely different tertiary topologies while catalyzing
the same reaction have been solved by crystallography
[7,8] and will be discussed in detail below. Depending
on the origin, ChOxs differ in their kinetic and redox
properties. In addition, the kinetic features of ChOx
are tailored to make use of membrane-bound choles-
terol as a substrate (see accompanying review [1]).

Recently, ChOx has also gained mechanistic interest as
Keywords
cholesterol oxidase; enzyme kinetics;
enzyme mechanism; flavoenzyme; oxygen
channel; protein structure; redox catalysis
Correspondence
A. Vrielink, School of Biomedical
Biomolecular and Chemical Sciences,
University of Western Australia, 35 Stirling
Highway - Crawley, WA, 6009 Australia
Fax: +61 8 6488 1148
Tel: +61 8 6488 3162
E-mail:
(Received 27 July 2009, revised
7 September 2009, accepted 14
September 2009)
doi:10.1111/j.1742-4658.2009.07377.x
Cholesterol oxidases are bifunctional flavoenzymes that catalyze the oxida-
tion of steroid substrates which have a hydroxyl group at the 3b position
of the steroid ring system. The enzyme is found, in a wide range of bacte-
rial species, in two forms: one with the FAD cofactor bound noncovalently
to the enzyme; and one with the cofactor linked covalently to the protein.
Here we discuss, compare and contrast the salient biochemical properties
of the two forms of the enzyme. Specifically, the structural features are dis-
cussed that affect the redox potentials of the flavin cofactor, the chemical
mechanism of substrate dehydrogenation by active-center amino acid resi-
dues, the kinetic parameters of both types of enzymes and the reactivity of
reduced enzymes with molecular dioxygen. The presence of a molecular
tunnel that is proposed to serve in the access of dioxygen to the active site
and mechanisms of its control by a ‘gate’ formed by amino acid residues

are highlighted.
Abbreviations
BsChOx, cholesterol oxidase from Brevibacterium sterolicum containing the FAD cofactor covalently linked to the enzyme; ChOx,
cholesterol oxidase; H
2
O
2
, hydrogen peroxide; ReChOx, cholesterol oxidase from Rhodococcus equi, originally identified as a
Brevibacterium sterolicum enzyme, containing the FAD cofactor noncovalently bound to the enzyme; SChOx, cholesterol oxidase from
Streptomyces SA-COO containing the FAD cofactor noncovalently bound to the enzyme; ShChOx, cholesterol oxidase from
Streptomyces hygroscopicus containing the FAD cofactor noncovalently bound to the enzyme.
6826 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
it was the first flavoprotein for which a molecular tun-
nel has been uncovered that has been proposed to
serve in oxygen access for the oxidative half-reaction
of the enzyme [7,9–11].
Catalysis by cholesterol oxidase
ChOx catalyzes three chemical conversions (see
Scheme 1 for an overview of the enzymatic steps). The
first catalytic conversion, called the reductive half-reac-
tion, is the dehydrogenation of the alcohol function at
the 3-position of the steroid ring system. The resulting
two redox equivalents are transferred to the (oxidized)
flavin cofactor that becomes reduced in the process. In
the second catalytic step, the reduced flavin reacts with
dioxygen to regenerate the oxidized enzyme and hydro-
gen peroxide (H
2
O
2

) (oxidative half-reaction). Finally,
the oxidized steroid undergoes an isomerization of the
double bond in the steroid ring system, from D5-6 to
D4-5, to form the final product cholest-4-en-3-one. In
general, this isomerization reaction occurs faster than
the release of the intermediate, cholest-5-en-3-one,
from the enzyme.
The specificity of ChOx for various substrates
derived from the cholestane skeleton has been the
object of several studies, in particular those describing
newly discovered ChOxs (see accompanying review 3
[1]). While the dehydrogenation of the CH–OH func-
tion at position 3 of the cholestane is retained, the
introduction of functional groups that alter the polar-
ity ⁄ hydrophobicity of the same have deleterious conse-
quences [12].
By contrast, it is noteworthy that most low-molecu-
lar-mass alcohols (e.g. propan-2-ol) are substrates
which have the ability to reduce the enzyme within 1 h
at a concentration of  1 m [12]. This highlights the
concept that ChOx is an alcohol oxidase adapted to
accommodate the bulky cholestane frame.
Forms of cholesterol oxidase
While the presence of the isoalloxazine (flavin) moiety
as the redox catalyst at the active center is a common
feature amongst ChOxs, three different forms of the
cofactor have been identified to date. Figure 1A,B
compare sequence alignments, performed using
clustalw2 [13], for different types of ChOxs. In the
majority of cases FAD is tightly, but noncovalently,

bound to the protein (Fig. 1A, reporting the sequence
of ChOx from R. equi [17] previously classified as non-
covalent ChOx from Brevibacterium sterolicum), while
in those cases where the isoalloxazine is covalently
attached to the protein, this occurs via a bond linking
the 8-methyl group of the isoalloxazine moiety to the
polypeptide chain (Fig. 1B). In the case of the enzyme
from Brevibacterium, the covalent attachment has been
identified through structural and mutagenesis studies
to involve the imidazole ND1 atom of a histidine resi-
due (His121) [7,14]. In one specific case, the flavin has
been reported to be FMN, although the evidence for
this does not appear to be conclusive [15].
ChOx proteins belonging to the subfamilies contain-
ing either covalent [cholesterol oxidase from B. steroli-
cum containing the FAD cofactor covalently linked to
the enzyme (BsChOx)] or noncovalent [cholesterol
oxidase from Rhodococcus equi, originally identified as
a Brevibacterium sterolicum enzyme, containing the
FAD cofactor noncovalently bound to the enzyme
(ReChOx), cholesterol oxidase from Streptomyces
SA-COO containing the FAD cofactor noncovalently
bound to the enzyme (SChOx) and cholesterol oxidase
from Streptomyces hygroscopicus containing the FAD
cofactor noncovalently bound to the enzyme
(ShChOx)] forms of FAD have been extensively char-
acterized with respect to their biochemical and kinetic
properties [12,16]. These exhibit differences in amino
acid sequences (Fig. 1), structures [7,8] and redox
properties [16]. The structural differences will be out-

lined below. In the noncovalent form of the enzyme,
release of the cofactor is possible, for example under
denaturation conditions such as heating at 90 °C
[17]. This indicates the high degree of stability of the
FAD–protein complex, largely because of the extensive
Scheme 1. Reaction steps catalyzed by ChOx and structures of
species involved. The terms ‘Reductive’ and ‘Oxidative’ half-reac-
tion refer to the changes of the redox state of the bound flavin
coenzyme. Cholest-5-en-3-one does not occur in free form under
normal catalytic conditions. Its conversion to the final product,
cholest-4-en-3-one, is faster than its formation.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6827
A
Fig. 1. Sequence alignments of different forms of cholesterol oxidase. The sequences were aligned using CLUSTALW2 at http://www.
expasy.ch. Sequences were obtained using the protein search algorithm at The National Centre for Biotechnology Information (NCBI). Panel
(A) corresponds to sequences for ChOx containing the FAD cofactor noncovalently bound to the enzyme. These include Streptomyces sp.
SA-COO (SChOx, S. sp, AAA26719), Rhodococcus equi (R. equ, CAC44897), Vibrio harveyi HY01 (V. har, ZP_01986092), Salinispora arenicola
CNS-205 (S. are, YP_001537636), Nostoc punctiforme PCC 73102 (N. pun), Frankia sp. EAN1pec (F. sp, YP_001508197), Mycobacte-
rium tuberculosis H37Rv (M. tub, CAB01014), Mycobacterium leprae (M. lep, CAC29897), Corynebacterium urealyticum DSM7109 (C. ure,
YP_001800712) and Streptomyces coelicolor (S. coe, NP_628939). Panel (B) corresponds to sequences for ChOx containing the FAD cofac-
tor covalently linked to the enzyme. These include Brevibacterium sterolicum (BsChOx, B. ster), Burkholderia cepacia (B. cepa, BAB63263),
Burkholderia thailandensis E264 (B. thai, YP_441176), Chromobacterium sp. DS-1 (Chrom, BAG70948) and Rhodococcus erythropolis
(R. eryt, ABW74861). Yellow shading indicates regions of the chain that are important in interactions with the FAD cofactor; * are residues
that make actual hydrogen bonding contact to the cofactor; and red shading indicates residues that are implicated as playing a role in cataly-
sis. The secondary structure elements for a) SChOx and b) BsChOx have been included in red (alpha helices) and blue (beta strands).
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6828 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
noncovalent interactions between the cofactor and the
protein.

Apoenzyme of cholesterol oxidase
In general ChOx binds the cofactor very tightly, and
its (reversible) removal requires rather harsh condi-
tions. Wels [16] has studied systematically the prepa-
ration of the apoprotein form of ShChOx with the
result that most common methods (see Husain &
Massey [18] for a review) lead either to incomplete
removal of FAD, or to extensive denaturation.
Removal of the cofactor in a reversible manner is
achieved at very low pH (< 2.0) and the differences
in the absorption spectra between free FAD and
FAD bound to protein have been monitored spectro-
scopically (see Fig. 2). The resulting apoprotein (80%
yield) was highly unstable however, and must be
reconstituted at high pH (> 8.0) using a large excess
of FAD [16].
The apoprotein of a BsChOx mutant, where the
histidine residue that forms the covalent link to the
flavin moiety has been mutated to an alanine, was
obtained by using a high potassium bromide concen-
tration: its characterization reveals lower stability
compared with the holoenzyme form, suggesting a link
between protein stability and covalent linkage between
the flavin cofactor and the protein [19,20].
Redox properties
While the catalytic dehydrogenation reaction of ChOx
in the presence of steroid substrate proceeds via a sin-
gle two-electron step from the oxidized flavin form
B
Fig. 1. (Continued).

A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6829
directly to the fully reduced form (see Fig. 2), reduc-
tion with artificial electron donors generates first the
flavin semiquinone species that is stabilized kinetically
in its (red) anionic form [16] (results not shown). The
redox properties have been studied for both the ‘cova-
lent’ [16,21] and ‘noncovalent’ [22] forms of the
enzyme.
The midpoint reduction potential, E
m
, of the covalent
form of the enzyme was measured according to the
method of Massey [23] and exhibits an E
m
of )101 mV.
Interestingly, in the H121A mutant, in which the cova-
lent linkage between the flavin and the peptide chain is
absent, the E
m
is lowered to )204 mV [14]. The 3D
structure of the mutant shows that the absence of the
covalent link leads to a more planar configuration of
the three-membered flavin system, resulting in a more
tetrahedral-like geometry for N5 [18]. This difference
was proposed to contribute to the observed alteration in
reduction potential. The E
m
for ShChOx has been
reported to be )217 mV [16]. Interestingly, SChOx

exhibits an E
m
dependence on pH (E
m
= )131 mV at
pH 7.0 and E
m
= )73 mV at pH 5.1) [22].
Spectral properties of selected
cholesterol oxidases
The absorption spectra of flavoproteins are very useful
parameters for assessing a variety of properties
[24–27]. They reflect the electronic state of the isoallox-
azine chromophore and consequently give information
about its redox and ionization states. In addition, the
spectra are very sensitive to the microenvironment at
the active site, such as hydrophobicity ⁄ polarity and the
presence of charges ⁄ dipoles. The spectra have thus
been used widely to deduce properties of the enzyme
[28]. Figure 2 depicts the absorption spectra of ChOx
from two different origins in the corresponding oxi-
dized and fully reduced states, as well as the spectrum
of free FAD. It should be noted that the band in the
visible region of the spectrum can vary in its intensity
by up to a factor of two, while the position of the
maximum may differ by  20 nm. Taken prima facie,
the spectrum of oxidized ShChOx would indicate an
apolar environment, while that of BsChOx suggests
conditions similar to those of free FAD in water.
The differences between the fully reduced forms are

quite extensive. Thus, the three bands that are posi-
tioned near 300, 340 and 400 nm in free reduced flavin
[29] appear red-shifted in reduced BsChOx, while an
opposite effect underlies the spectra of reduced
ShChOx (Fig. 2). By comparison with the spectral
properties reported earlier [16], it can thus be deduced
that, in the case of ShChOx, the reduced flavin is
(mainly) in its anionic form, while in BsChOx either
the neutral form or a mixture of both is present.
The semiquinone forms similarly have characteristic
spectra with maxima in the near-UV range at 372
(ShChOx) and 382 (BsChOx) nm, and 445 (ShChOx)
and 485 (BsChOx) nm, which reflect the presence of
the chromophore exclusively in its anionic form
(results not shown; note that the neutral flavin radical
is blue with maxima in the 560–620 nm region, while
with both ChOx enzymes no absorption is observable
in this area) [16].
Both ShChOx and BsChOx exhibit a weak fluores-
cence emission that is maximal at  525 nm and
approximately 0.5% that of free oxidized FAD [16].
Interestingly, the emission intensity of the reduced
forms of BsChOx and ShChOx is  fourfold higher
and shifted towards 490 nm [16].
Structures
Overall structure
As pointed out above, ChOx has been identified in a
number of bacteria and these flavoenzymes exhibit lar-
gely differing sequences (Fig. 1A,B) that suggest large
structural differences between the proteins. For many

FAD-containing enzymes, a consensus sequence of
repeating glycine residues (GXGXXG) followed by a
300 350 400 450 500 550 600
FAD (ox)
BCO (ox)
BCO (red)
SCO (ox)
SCO (red)
FAD (red)
Wavelength (nm)
448
Extinction coefficient (M
–1
·cm
–1
)
12 × 10
3
4 × 10
3
8 × 10
3
388
467
362
350
Fig. 2. Absorption spectra of ChOx from Streptomyces hygroscopi-
cus (blue) and Brevibacterium sterolicum (red) in their oxidized and
fully reduced states. Enzyme solutions were  10 l
M in 0.1 M potas-

sium phosphate (pH 7.5) and at 25 °C. The fully reduced forms of
ShChOx and BsChOx were obtained upon the addition, anaerobically,
of small amount of excess cholesterol. The spectra of FAD in its
oxidized [87] and fully reduced forms are reported for comparison.
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6830 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
D ⁄ E approximately 20 residues further along the pri-
mary sequence, is characteristic of a nucleotide-binding
fold (in other words, a motif that facilitates interac-
tions between the bound nucleotide and the protein)
[30,31]. The consensus sequence of glycine residues
allows binding of the charged diphosphate moiety of
the nucleotide near to the N-terminal end of a helix,
with stabilization exerted through the helix dipole
effect [32]. The presence of an aspartate or glutamate
side chain  20 residues further along the protein
chain enables hydrogen bond interaction between the
2¢ and 3¢ hydroxyl groups of the nucleotide ribose moi-
ety and the carboxylate group of the side chain. These
features are characteristic of many nucleotide-binding
proteins and appear to be critical in forming the neces-
sary interactions between the nucleotide and the pro-
tein to favor correct binding of the cofactor. The
noncovalent form of ChOx exhibits an almost identical
consensus sequence of glycines (G17-X-G19-X-G21-
G ⁄ A22) followed by a glutamate (E40), indicating the
presence of a nucleotide-binding fold (see Fig. 1A, the
numbering of residues corresponds to the SChOx
sequence); however, in the case of the covalently
bound ChOx this consensus is notably absent

(Fig. 1B), suggesting the likely absence of a nucleotide-
binding fold for this form of the enzyme. Nevertheless,
the FAD cofactor is bound tightly, even when the
covalent link with the apoprotein moiety is removed
(see above). Indeed, as noted above, the sequence iden-
tity can be used as a predictive measure of the mode
of flavin binding to the enzyme: covalent versus non-
covalent linkage.
High-resolution crystal structures were determined
for two noncovalent forms of the enzyme (SChOx
and ReChOx) [8,33] and for one of the covalent
forms (BsChOx) [7], providing a view of the overall
folds for each protein (Fig. 3). The overall topologies
for each form of the enzyme vary significantly from
each other (Fig. 3). Although both enzymes have
been defined as being composed of two domains
based on function (FAD-binding domain and sub-
strate-binding domain), they can also be considered
as single-domain proteins based on topology because
the protein chain meanders back and forth between
the regions that provide the binding features for the
cofactor and the binding features for the substrate as
well as the catalytic residues. The noncovalent ChOx
belongs to the glucose-methanol-choline (GMC) oxi-
doreductase flavoenzyme family [34] whereas the
covalent enzyme belongs to the vanillyl-alcohol oxi-
dase (VAO) family [35].
The structure of noncovalent ChOx shows the cofac-
tor binding motifs, containing the consensus sequence
of glycines (as discussed above) as well as two further

regions of conserved glycines (Figs 1A and 4A), all of
which allow a close approach of the protein main
A
B
Fig. 3. 3D structure of ChOx. An overall view of the secondary
structure elements for (A) SChOx and (B) BsChOx is shown. The
FAD cofactor is presented in ball-and-stick representation.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6831
chain to the phosphate oxygen atoms of the cofactor,
thus facilitating hydrogen bond interactions.
The noncovalent enzyme possesses the characteristic
nucleotide-binding fold (Rossmann fold) consisting of
a b-pleated sheet sandwiched between a-helices and
containing the motif needed for binding the cofactor
(Fig. 3A). The diphosphate group of the cofactor is
positioned closely packed to the N terminus of the first
a-helix of the protein (Fig. 4A) where the conserved
GXGXG glycine residues are located. Side chains
larger than glycines at this critical loop between the
N-terminal end of the helix and the first b-sheet would
result in positioning the flavin diphosphate moiety
further away from the chain and hence outside the
polarization field resulting from the helix dipole.
Hence, the glycine residues are strategically positioned
to allow the helix dipole to stabilize the negative
charge of the cofactor.
In the covalent form of the enzyme, the diphosphate
moiety is localized in a pocket made by the residues
found between the third and fourth b-strands of a

four-stranded b-pleated sheet (Fig. 4B). Interaction
between the phosphate oxygens and the protein involve
main chain nitrogen atoms of residues in this loop
region. Furthermore, this loop contains His121, the
side chain of which covalently connects to the cofac-
tor. In contrast to the noncovalent form of the
enzyme, the covalently bound form does not exhibit
hydrogen bond interactions between the ribose hydro-
xyl groups of the cofactor and the protein. The ribityl
chain of the cofactor also makes much fewer hydrogen
bond interactions with the protein in the covalently
bound form compared with the noncovalently linked
form. Covalent linkages of FAD to proteins have
shown that, in addition to histidine residues, cysteine,
tyrosine and threonine side chains can be involved (see
Heuts et al. [36] for a review of covalent linkages
between flavoproteins and their cofactors). A compari-
son of the sequences for other covalently bound forms
of the enzyme (Fig. 1B) predicts that, in all cases, the
linkage between the cofactor and the protein is to a
histidine side chain. Vanillyl-alcohol oxidase, another
example of a flavoenzyme containing covalently bound
flavin, also utilizes a histidine for the covalent attach-
ment; however, in this case the linkage involves the
NE2 atom of the histidine residue [35]. The structure
A B C
D E F
Fig. 4. Structural views of different forms of ChOx. Interactions between the protein and the FAD cofactor are shown for (A) the noncova-
lently linked ChOx (SChOx) and (B) the covalently linked enzyme (BsChOx). Hydrogen bonds between the cofactor and the protein are
shown as dashed lines. Models of the bound steroid in the active sites of (C) SChOx and (D) BsChOx. Active site residues of (E) SChOx and

(F) BsChOx.
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6832 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
of BsChOx reveals that the covalent linkage is made to
the ND1 atom of the histidine side chain. However,
this bond is not a prerequisite for flavin binding
because in the H121A mutant, FAD is bound tightly
and forms a catalytically competent holoprotein in
which the redox potential of the cofactor is lowered by
 100 mV (see above) [14,21]. The structure reveals
changes in the conformation of the isoalloxazine ring
system that are proposed to affect the redox chemistry
[14]. In addition to the differences in the cofactor bind-
ing motifs between the covalent and noncovalent ChO-
xs there are also large differences in the overall
topology. These differences are evident through the
sequence diversity seen between the two enzyme forms
(Fig. 1A,B).
The active site
Despite the topological differences in the two enzyme
forms, both contain a large buried hydrophobic pocket
that is able to accommodate the steroid ring system
(Fig. 4C ⁄ D). In both forms of enzyme the binding site
for the steroid is sealed off from the external environ-
ment of the protein by a number of loops, which exhi-
bit higher mobility than the rest of the protein
structure. This suggests that these loops must re-orient
to allow the steroid substrate to enter the hydrophobic
pocket. Comparisons of the structures of SChOx and
ReChOx show differences in the nature of the loops;

in SChOx the loops are more rigid in nature as well as
containing a amphipathic helical turn, whereas in
ReChOx the loops are more extended, lack secondary
structure elements and exhibit higher temperature fac-
tors [33]. The higher mobility of the substrate entry
loops in ReChOx compared with SChOx correlates
with the observed elevated K
m
values for both choles-
terol (K
m
=3lm for SChOx and > 100 lm for
ReChOx) and dehydroisoandosterone (K
m
= 27.5 lm
for ChOx and 400 lm for ReChOx) in ReChOx [33].
Indeed, the increased rigidity of the loops in SChOx
pre-orients the residues needed for binding the 8-car-
bon isoprenoid tail at C17 of the substrate and thus
increases the efficiency of the enzyme for catalysis.
The isoalloxazine ring system is located at the ‘base’
of the pocket. The active site is on the re face of the
isoalloxazine for the noncovalent form of the enzyme
and on the si face for the covalent enzyme form.
ChOx requires a number of features for efficient
catalysis. First, the steroid must be correctly oriented
relative to the cofactor for hydride transfer from the
steroid C3 site to the isoalloxazine N5 site. Second,
side chain functional groups for two catalytic steps are
needed: (a) a base to accept the steroid C3-O-H proton

during oxidation and (b) a base for proton transfer
during the isomerization reaction. Finally, conditions
must be met for binding and conversion of molecular
oxygen to peroxide during the oxidative half-reaction.
Despite these common requirements, the active sites of
the two forms of ChOx are highly divergent. Indeed,
the differences in sequences and structures for both
enzymes are also evident in terms of residues identified
as playing roles in catalysis.
Chemical mechanisms
Dehydrogenation
From a chemical point of view the reductive half-reac-
tion catalyzed by ChOx involves the dehydrogenation
of a nonactivated alcohol to generate the correspond-
ing carbonyl function, with the resulting redox equiva-
lents transferred to the bound flavin cofactor, as
shown in Scheme 2.
This type of reaction is catalyzed by several classes
of flavoproteins, such as the methanol (or alcohol)
oxidases [37,38] (including e.g. alditol oxidase [39]),
those acting on a-OH-carboxylic acid oxidases (e.g.
lactate oxidases ⁄ dehydrogenases [40]) and those that
dehydrogenate the hydroxyl group of cyclic carbohy-
drates (e.g. glucose oxidase [41]).
While this type of reaction is common in biochemi-
cal systems, its catalytic mechanism has been much
debated [42]. In the reaction, the key step is the rup-
ture of the kinetically stable C-H bond for which three
modes are conceivable [43]: (a) a homolytic fission
(radical mechanism), (b) a heterolytic mechanism, in

which the hydrogen is abstracted as H
+
(carbanion
mechanism), and (c) a heterolytic mechanism, in which
expulsion of a hydride generates (transiently) a carbe-
nium ion (hydride mechanism). These basic modes
have been discussed in detail elsewhere [44]. While ear-
lier studies tended to favor mechanism (a) [37,38] most
authors now agree on mechanism (c), the so-called
‘hydride transfer’ for this type of reaction [9,33,45–47].
This mechanism is very common in biochemical sys-
tems and is operative also for the dehydrogenation of
the same CH–OH functionality by pyridine nucleotide
[48,49] and quinoprotein-dependent enzymes [50]. Two
prerequisites are necessary for dehydrogenation via this
mechanism: correct alignment of the orbitals involved;
Scheme 2. Overview of the reductive half-reaction catalyzed by
ChOx.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6833
and a base that accepts the –OH proton (as shown in
Scheme 3).
One reason for the requirement of a base in the
dehydrogenation of the CH-OH group [51] is because
of the high pK
a
of the hydroxyl group (> 15) and the
fact that the reaction appears to involve a concerted
rupture of the O-H and C-H bonds. A discussion of
the role of specific amino acid residues in catalysis fol-

lows below. In contrast to the dehydrogenation of the
CH-OH group, for oxidation of the CH-NH
2
function-
ality (such as in amino acids), the presence of a base is
not mandatory [52] because of the lower pK
b
(8–10)
for this group. As argued convincingly by Klinman
and coworkers, the transfer of the hydride to the
acceptor involves a high degree of tunneling, the extent
of which might depend on the specific enzyme [48,53].
In the case of SChOx, the structure of a steroid-
bound complex (Fig. 4C) has provided initial insights
into the specific roles of catalytic residues and the
mechanisms of both oxidation and isomerization [54].
Mutagenesis and kinetic analyses, as well as further
structural studies, have helped to identify the role in
catalysis of a number of key residues. These studies
have suggested that His447 plays a role in substrate
orientation and the correct orbital alignment for de-
protonation of the C3-OH proton and hydride transfer
to the cofactor [9,33,55,56]. Recent structural studies
have suggested that Tyr446 may act by exerting steric
pressure on the isoalloxazine ring moiety, thereby
affecting the ability of the cofactor to become reduced
[57]. This re-orientation of the tyrosine side chain, and
the resulting distortion of the flavin ring system, is
induced by substrate binding at the active site and is
an example of an induced-fit mechanism for the modu-

lation of the redox potentials of the cofactor. Asn485
acts to stabilize the reduced cofactor and, through a
movement towards the reduced isoalloxazine ring
system, also allows access of oxygen to the active site
via the oxygen channel (see below) for the oxidative
half-reaction [58].
A definitive identification of the base needed for
dehydrogenation is still somewhat unclear. A likely
candidate is Glu361 [9]. However, this residue has also
been shown to be required as the base for isomeriza-
tion chemistry [59,60] (see below). Invoking Glu361 as
the base for both catalytic reactions requires the pro-
ton abstracted from the C3-OH group to be trans-
ferred elsewhere after dehydrogenation chemistry has
occurred so that the glutamate can act as the base for
the subsequent isomerization step. The mechanism and
kinetics of this proton transfer step are not yet clearly
understood. It may be that it is transferred directly to
an incoming oxygen molecule as part of the oxidative
half-reaction. This is possible because the glutamate
side chain is positioned near to the point where the
oxygen channel enters the active site. Alternatively, the
proton may be transferred to C(4)=O of the isoalloxa-
zine moiety of the cofactor before reacting with dioxy-
gen; however, this would require an intermediate
residue to shuttle the proton the 5 A
˚
distance from the
glutamate carboxylate to the isoalloxazine ring.
The covalent ChOx has not been crystallized in the

presence of a steroid substrate or a substrate analogue.
The binding site for the substrate is much more hydro-
philic in nature than the noncovalent enzyme because
of the presence of a large number of charged side
chains (Fig. 4F). Two charged residues (Glu475 and
Arg477) are found to adopt two distinct conforma-
tions. The positions of their side chains are correlated
with one another, suggesting that the movements they
make are concerted. Based on the structure and on a
model of the steroid substrate bound to the enzyme, it
has been inferred that Glu475 is the base for proton
abstraction of the steroid C3-OH in the reductive half-
reaction of BsChOx [7].
Isomerization
The isomerization step of the cholest-5-en-3-one is the
second reaction catalyzed by most ChOx proteins in
which the hydrogen at position C(4) is transferred to
position C(6) [59]. The theoretical requirements for the
isomerization step of catalysis are represented in
Scheme 4.
The C-H group at position 4 of the cholestane ring
is activated (acidified) by the flanking carbonyl group.
This facilitates abstraction of the hydrogen as H
+
by
an appropriately positioned base to yield a transient
enolic species. Reprotonation at position 6 then leads
to the final cholest-4-en-3-one product.
The isomerization step of the intermediate cholest-5-
en-3-one to final product is slow (0.3 s

)1
) in the case
of reduced ShChOx; however, the rate of this conver-
sion is approximately three orders of magnitude faster
with the oxidized form [12]. From this it can be
concluded that with ShChOx in the catalytic mecha-
nism depicted in Scheme 4, the isomerization follows
Scheme 3. Depiction of the requirements for efficient dehydroge-
nation in ChOx (‘Acc’ stands for the electron acceptor FAD).
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6834 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
re-oxidation and is not rate limiting for catalysis
[12,16]. With BsChOx, the isomerization proceeds at
approximately the same rate ( 20–30 s
)1
) with the
oxidized or reduced enzyme forms [14] and is close to
the rate of flavin reduction at saturating substrate
concentration [16].
This reaction has been studied in detail for SChOx
by the group of Sampson [55,59,60]: H
+
abstraction
occurs stereospecifically on the b-face of the steroid
ring system, is mediated by Glu361 and the label is
retained in the product.
In an identical manner to SChOx, a glutamate is
thought to play the role of a base in both the oxida-
tion and in the proton-transfer mechanism of isomeri-
zation in BsChOx. The model of the enzyme with

substrate bound reveals the side chain of Glu475 on
the b-face of the steroid substrate, well positioned to
abstract the C4 proton and reprotonate at C6 of the
enolic intermediate [7]. It has been proposed that
Glu311 and Glu475 are involved in a proton-transfer
mechanism to allow the proton removed from the ste-
roid C3-OH during the reductive half-reaction to be
positioned such that it can react with molecular oxy-
gen during the oxidative half-reaction [7]. This proton-
transfer mechanism enables Glu475 to take on the role
of a base for isomerization [10].
Kinetics
General aspects
It should be noted that the several publications
describing the isolation of ChOx from various sources
in general give some rudimentary indications about the
activity of the purified protein. Because the catalytic
activity of ChOx is highly dependent on the buffer
composition, pH and on the type and concentration of
surfactant used to solubilize the substrate (which in
turn affects micelle composition and dynamics), a com-
parison and discussion of such parameters is problem-
atic (see also accompanying review [1]). Pollegioni
et al. [61] have addressed these topics in some detail
and have also studied the stability of the ShChOx and
BsChOx in the presence of different concentrations of
some organic solvents. In essence, they conclude that
every single one of the named parameters will affect
the catalytic activity to some extent. Nishiya et al.
studied the effect of specific detergents on a mutant of

a Streptomyces ChOx and detected a change of mecha-
nism compared with wild-type ChOx [62]. Ahn and
Sampson [63] studied the effect of lipid structure on
the activity of ChOx and concluded that ‘enzymatic
activity directly reflects the facility with which choles-
terol can move out of the membrane’.
A study of ChOx activity as a function of assay
type, structure of the steroid substrate and presence of
solubilizers was conducted by Gadda et al. [16], and
the deduced parameters are listed in Table 1. A more
detailed study by the same groups compared the
kinetic mechanisms of ShChOx and BsChOx [12].
Therein it was concluded that BsChOx acts via a ping-
pong mechanism, whereas the catalytic pathway of
ShChOx is sequential.
Xiang et al. [64] have generated models of the
expected Michaelis complex of ChOx and cholesterol
and used this information to design mutants expected
to exhibit activities different from those of the parent
enzyme. The results were inconclusive in this respect,
although they did indicate that the loss of active-site
water was the predominant source of binding energy.
Sampson et al. [65] have also studied the role of a spe-
cific active-site loop of ChOx. The activity of the
mutant with cholesterol in a vesicle is decreased 2800-
fold compared with wild-type enzyme, whereas isomer-
ization activity is retained. The authors conclude that
‘the loop is important for movement of cholesterol
from the lipid bilayer’.
In an analogous study, Toyama and coworkers [66]

attempted to compare the activities of the ChOxs by
implementing mutations that were expected to yield
convergent active sites. Some of the mutants indeed
lead to significantly altered catalytic parameters for
various substrates and consequently to a different spec-
ificity spectrum.
The reductive half-reaction
The reductive half-reaction was investigated in com-
parative manner for ShChOx and BsChOx using the
stopped-flow technique [12]. Both enzymes are reduced
*H
O
O
O
~B|
*H
~B*H
+
~B|
4
6
6
4
Scheme 4. General mechanism for the isomerization reaction of ChOx: B| is a base that abstracts and redonates a H
+
to the intermediate.
Only the reactive portion of the steroid substrate is shown.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6835
by cholesterol at essentially identical rates under anaer-

obic conditions ( 230 s
)1
, Table 2) and saturation
behavior is observed for both. The course of the flavin
reduction process is monophasic (i.e. no intermediate
species are observable). This implies that binding of
cholesterol to the oxidized enzyme does not perturb
significantly the spectrum of the latter. The observed
reduction process thus reflects step k
2
in Scheme 5.
Notably, the rapid kinetic studies do not uncover
spectral changes that might be attributed to the release
of product from the reduced enzyme via step k
5
(Scheme 5). Similarly, attempts to follow spectroscopi-
cally the formation of the EFAD
red
5-Chol-3-one
complex from its components were not successful. This
implies that the spectrum of the reduced flavin chromo-
phore is little affected by the presence of product and ⁄ or
that product dissociation via step k
5
is faster than that
of the preceding step k
2
. The kinetic parameters are
summarized in Table 2 for a selected set of conditions.
We refer the reader to a previous publication [12] for a

more comprehensive discussion of this topic.
Table 1. Kinetic parameters for turnover by BsChOx and ShChOx using different substrates and different assays. The data were estimated
for 25 °C and pH 7.5 as described previously [61]. Three different assays were used that rely on the spectrophotometric detection of product
formation (cholest-4-en-3-one) at 240 nm (A), on the polarographic determination of the rate of oxygen consumption (B) and on the rate of
H
2
O
2
formation detected with o-dianisidine and horseradish peroxidase (C). (a) 0.5 M Potassium phosphate, 1% Thesit (polyoxyethylene(9)-
lauryl-ether), 1.25% propan-2-ol; (b) 0.1
M potassium phosphate, 1% Triton X-100, 1.25% propan-2-ol. Adapted from a previous publication
[61].
Substrate Type of assays Conditions
BsChOx ShChOx
K
m
mM
k
cat
s
)1
K
m
mM
k
cat
s
)1
Cholesterol Cholest-4-en-3-one (A) (a) 0.14 67 0.2 11
O

2
(B) (a) 0.11 56 0.25 9
O
2
(B) (a) however 10% propan-2-ol 0.20 57 0.17 6
O
2
(B) (a) however 10% propan-2-ol
and 50 m
M phosphate
0.25 43 0.17 3
Cholest-4-en-3-one (A) (b) 0.07 48 0.8 63
H
2
O
2
(C) (b) 0.04 48 0.4 32
5-Cholesten-3-one Cholest-4-en-3-one (A) (a) 0.27 278 1.52 332
trans-Dehydroandrosterone Cholest-4-en-3-one (A) (b) 1.2 0.8 0.3 8.2
H
2
O
2
(C) (b) 0.9 1.0 0.2 6.0
Pregnenolone Cholest-4-en-3-one (A) (b) 0.4 21 0.2 24
H
2
O
2
(C) (b) 0.2 35 0.2 21

Cholestanol H
2
O
2
(C) (b) 0.2 40 0.7 37
trans-Androsterone H
2
O
2
(C) (b) 0.8 0.8 0.5 7
Table 2. Specific rate constants obtained for the reductive half-
reaction of ChOx. Conditions: cholesterol as substrate, 50 m
M
potassium phosphate buffer, pH 7.5, 1% Thesit (polyoxyethyl-
ene(9)-lauryl-ether), 10% propan-2-ol, at 25 °C. The rates were
obtained from stopped-flow experiments under anaerobic condi-
tions and as detailed previously [12].
Experimental Estimated*
k
red
(k
2
)
(s
)1
)
K
d
( k
)1

⁄ k
1
)
(m
M)
1 ⁄ Slope
( k
2
Æk
)1
⁄ k
1
)
(
M
)1
Æs
)1
)(·10
6
)
k
1
(M
)1
Æs
)1
)
(·10
5

)
k
)1
(s
)1
)
ShChOx 231.8 1.5 5.5 5 2500
BsChOx 235.1 0.16 16.7 7.7 1250
*The k
1
and k
)1
rate constants are the minimal estimates obtained
by computer simulation of the experimental traces as described
previously [12].
Scheme 5. Minimal kinetic mechanisms for the behavior of
ShChOx and BsChOx in turnover catalysis [20]. Note that the
central part represents the reductive half-reaction, while the
segments on the flanks correspond to the oxidative half-reaction.
The isomerization reaction is not shown since it is not rate limiting
in the processes.
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6836 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
In summary, the kinetic description of the reductive
half-reaction of these two ChOxs is rather simple and
typical for a flavoprotein oxidase. (Compare to the
vertical steps k
1
,k
)1

k
2
and k
)2
. in the central part of
Scheme 5.)
The reaction of reduced ChOx with dioxygen
(oxidative half-reaction).
The reaction with dioxygen is of particular interest
because studies with ChOx have led to the first insights
into the modes of access of dioxygen to the active cen-
ter and its control [7,9,10]. Starting from (pre)reduced
ChOx in its uncomplexed form, the reaction with diox-
ygen proceeds differently for BsChOx compared with
ShChOx. Re-oxidation of BsChOx [monitored by fol-
lowing the (re)appearance of the absorption of the oxi-
dized cofactor at  450 nm] proceeds via a first rapid
phase that encompasses most of the process and shows
a saturation dependence on [O
2
]. This is followed by a
second, much slower, phase (k
obs
 5s
)1
) that is of
minor extent, is too slow to account for the catalytic
turnover and is independent from [O
2
] [12]. In view of

this its nature was not investigated in detail. The inter-
pretation of the dependence of the first phase from
[O
2
] was not straightforward because such a kinetic
behavior would be typical of a Michaelis situation
where an encounter complex with dioxygen would pre-
cede the redox reaction. In the present case the dissoci-
ation constant, K
d
, for dioxygen would be
unrealistically low. It was thus proposed [12] that the
observed behavior would reflect the interconversion of
enzyme species having different intrinsic reactivities
with dioxygen. Based on the results described previ-
ously[10] [12], a hypothetical mechanism can be formu-
lated to account for the reactivity of reduced BsChOx
with dioxygen, as shown in Scheme 6.
In this scheme, two forms of the reduced enzyme
interconvert via k
i
, where the rate of this interconver-
sion is smaller than that of k
ox
. Of the two reduced
enzyme species, EFAD
red
is the better reactant (com-
pared with EFAD
red

*) and yields EFAD
ox
+H
2
O
2
via path k
ox
. It should be noted that the rates of k
ox
would be faster compared with other possible path-
ways involving k
ox
* (i.e. k
ox
>> k
ox
*; k
i
> k
ox
*).
There is no direct experimental evidence for the exis-
tence of k
ox
*; its formulation is based on the knowl-
edge that there is no reduced flavoprotein which will
not react with dioxgen, even if the rate is very slow.
The step of interconversion, k
i

, could be connected to
the role of the proposed gate formed by Glu311, where
EFAD
red
would correspond to the ‘open’ gate forms
and EFAD
red
* would correspond to the ‘closed’ gate
forms (see also discussion below and in Piubelli et al.
[10]).
In the case of ShChOx, the course of reoxidation of
free reduced enzyme by dioxygen is essentially mono-
phasic and the results of secondary data analysis are
consistent with a second-order reaction mechanism, as
observed with the vast majority of flavoprotein oxidas-
es [67]. In contrast to this, the interpretation of the
data from the reaction of reduced ShChOx in the pres-
ence of the final product, cholest-4-en-3-one, are com-
patible with saturation with increasing [O
2
] [12].
However, the estimated rates of reoxidation are not
sufficiently fast to account for catalysis [12]. This
might be because the complex of reduced enzyme with
the final product, cholest-4-en-3-one does not react as
fast as that with cholest-5-en-3-one (a complex of
reduced enzyme with cholest-5-en-3-one cannot be
studied because the ligand is converted catalytically to
cholest-4-en-3-one). This behavior is reminiscent of
that observed with BsChOx and is indicative of the

redox reaction being preceded by a kinetic event affect-
ing dioxygen reactivity. Also, there are no indications
for the occurrence of intermediates in the course of the
conversion of reduced to oxidized ShChOx [12]. This
speaks against an alternative interpretation in which
formation of oxidized flavin is preceded by formation
of an intermediate (such as a 4a-flavin hydroperoxide,
as observed for flavin-dependent hydroxylases [67]). In
conclusion, the oxygen reactivity of ChOx enzymes
that have been investigated in some detail to date is
distinctly different from that observed with typical
flavoprotein oxidases [67–70].
The overall kinetic mechanism of catalysis
The data from the rapid reaction studies and those from
turnover experiments can be interpreted combinatively
to yield a catalytic scheme. This is shown in Scheme 5
for both ShChOx and BsChOx. Note that the mecha-
nisms are very similar and typical for flavoprotein
oxidases. The main difference is in the relative rates of
the steps of product dissociation compared with electron
Scheme 6. Possible mechanism that can be formulated to account
for the reactivity of reduced BsChOx with dioxygen based on the
results described in [12]. The scheme represents two forms of
reduced enzyme that can interconvert via step k
i
: of these
(EFAD
red
) is the species that reacts faster with O
2

via k
ox
. See
text for further details.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6837
transfer to dioxygen. Thus, with ShChOx, reoxidation
(k
3
) precedes product dissociation (k
4
, left loop of
Scheme 5), while with BsChOx the inverse is true (right
loop of Scheme 5).
Single molecule properties
The study of single (protein) molecules is a method
that relies mainly on fluorescence microscopy and
which emerged approximately a decade ago. It has
gained great prominence, mainly as a result of its
application in the study of basic enzymatic mecha-
nisms, such as the Michaelis–Menten concept [71,72].
Flavoproteins are ideally suited for this purpose
because their fluorescence can vary widely depending
on the redox status of the cofactor [29]. In the specific
case, the ChOx used [73] was reported to be fluores-
cent in its oxidized state, while its reduced counterpart
exhibits a much lower emission. It also exists as a
monomer, thus reducing the complexity inherent to
polymeric enzymes. ChOx thus happened to be
selected as the first enzyme, the kinetics of which was

studied by time-resolved, single-molecule fluorescence
spectroscopy [73,74]. In the actual experiments [73],
the fluorescence is plotted versus time where the result-
ing diagram reflects the ‘time intervals’ in which the
enzyme is present either in the oxidized form or in the
reduced form. From the time distribution of high ver-
sus low fluorescence states, a kinetic mechanism can be
derived [73] that compares well with that obtained
from classic studies (see below). Intriguingly, the sin-
gle-molecule fluorescence studies did uncover a behav-
ior not discoverable by other methods: thus, a
difference in catalytic properties was noted that
depends on the antecedent state of the single molecule.
The authors state: ‘A molecular memory phenomenon
(was observed), in which an enzymatic turnover was
not independent of its previous turnovers because of a
slow fluctuation of protein conformation, was evi-
denced by spontaneous spectral fluctuation of FAD’
[73].
The oxygen channel
Much focus in the study of flavoenzyme oxidases was
originally on the reductive half-reaction where the
substrate is dehydrogenated (oxidized) and the flavin
cofactor becomes reduced. However, these enzymes
also require an enhancement of the reactivity with di-
oxygen to enable an effective (re)oxidation of the
reduced cofactor to allow cycling of catalysis. In this
context it should be noted that until recently, of the
numerous 3D structures of flavoprotein oxidases avail-
able, none would yield hints about the mode of access

of dioxygen to the active center, the locus of the O
2
interaction with the reduced flavin, or – in the cases of
flavoprotein dehydrogenases – the mode of prevention
of oxygen reactivity [75,76]. In this context, the mode
of access of dioxygen to the active center of proteins
is of relevance. Historically, common assumptions go
back to Weber’s experiments [77], which showed that
the quenching of the fluorescence of bound fluoro-
phores by molecular oxygen is an appropriate method
for using to determine the accessibility of dioxygen to
specific loci within macromolecules. In 1973, these
authors postulated that …’the apparent oxygen diffu-
sion rate through the protein matrix is 20–50% of its
diffusion rate in water’ [78]. This led to the – subse-
quently widely accepted concept that dioxygen can
diffuse almost freely through proteins. More recent
insights, stemming from the hemoglobin and lipoxy-
genase fields, indicate that some proteins possess
specific ‘pockets’ that serve in binding of dioxygen,
which, in turn, gains access to the active site via
appropriate channels [79–85]. These mechanisms can
either enhance or inhibit oxygen reactivity. Accord-
ingly, recent structural studies with SChOx, BsChOx
and l-amino acid oxidase have revealed features that
could be interpreted to be molecular channels that
facilitate controlled entry of oxygen to the flavin active
sites [7,9,86].
In both structures of ChOx a gated channel is evi-
dent and has been proposed to play a role in oxygen

entry to the active site [7,9]. In detailed studies on the
function of the assumed channels, several amino acid
residues that might constitute the gate have been
substituted and the reactivity of the mutants with diox-
ygen was investigated and compared with that of the
parent enzyme [10,11]. Mutations of Glu311, a key res-
idue assumed to be part of the oxygen gate in
BsChOx, causes a switch in the basic kinetic mecha-
nism of the reoxidation with dioxygen: thus, while
wild-type BsChOx and most mutants show a satura-
tion behavior with increasing [O
2
], for Glu311 mutants
a linear dependence was found that would reflect a
‘simple’ second-order process [10]. The ‘set-up’ of the
proposed oxygen tunnel and of the corresponding gate
in BsChOx is shown in Fig. 5.
In the case of SChOx (Fig. 6), the strategy imple-
mented in order to elucidate the mode of oxygen
access involved mutation of residues that were part of
the proposed gate (Asn485Asp) or tunnel (Phe359Trp
or Gly347Asn) [11]. It was found that these mutants
affect the kinetic parameters for dioxygen in that they
diminish the overall catalytic efficiency. The kinetics
data obtained with wild-type and mutants of SChOx
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6838 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
were found to be consistent with a Michaelis-type
mechanism in which dioxygen forms a ‘precomplex’
before reacting chemically. A further argument for the

proposed mechanism is the cooperativity observed in
the kinetic behavior of dioxygen with the tunnel
mutants, but not with the parent enzyme. As with
BsChOx, a rate-limiting conformational change for
binding of dioxygen in SChOx would be consistent
with the kinetic data. Of particular importance was the
structure of the F359W mutant, in which the Trp
indole ring completely fills the channel [11].
In conclusion, it appears that in both BsChOx and
SChOx the access of dioxygen to the reduced flavin at
the active center does not proceed by simple diffusion
processes, but is guided and controlled by the protein
frame. This might serve in a better utilization of oxy-
gen at low concentrations (sequestration), and might
play a role in regulating the reactivity of the reduced
enzyme–product complex.
Prospects
Unquestionably, biochemical and structural studies on
ChOx have played a forerunner role in the understand-
ing of various aspects of flavoenzyme redox catalysis,
notably as it relates to flavoprotein oxidases. This
might apply particularly for the insights into the mech-
anism for re-oxidation and of control of the reaction
with dioxygen. Of similar novelty were the results of
the single molecule studies. The more recent findings,
that ChOx from the pathogenic bacterium R. equi
plays an important role in infection [5,6] and the fact
that it is unique to bacterial species, opens up an
entirely new field. It is to be expected that future stud-
ies will focus on ChOx potential as a target for new

AB
Fig. 5. Close up of the active site of
BsChOx with the channel proposed to func-
tion in access of dioxygen and its control.
The channel extends from the surface to
the active site cavity, therein the flavin
moiety is depicted in yellow. The left panel
shows the ‘open’ conformation and the right
panel shows the ‘closed’ conformation. This
is accomplished by the side chains of
Ile423, Glu475 and Arg477 that adopt differ-
ent conformations. The side chain residues
of the latter two amino acids are repre-
sented with space-filling models.
A
B
Fig. 6. Close up of the channel in SChOx proposed to function in
the access of dioxygen. The green surface corresponds to the
solvent-accessible surface of the molecule. Two conformations of
residues lining the channel result in (A) the ‘channel open’ con-
formation and (B) the ‘channel closed’ conformation. An oxygen
molecule is observed in the ‘channel open’ conformation.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6839
antibiotics and thus on the design of specific inhibi-
tors.
References
1 Kreit J & Sampson NS (2009) Cholesterol oxidase:
physiological functions. FEBS Journal 23, doi:10.1111/
j.1742-4658.2009.07378.x

2 Pollegioni L, Piubelli L & Molla G (2009) Cholesterol
oxidase: biotechnological applications. FEBS Journal
23, doi:10.1111/j.1742-4658.2009.07379.x
3 Purcell JP, Greenplate JT, Jennings MG, Ryerse JS,
Pershing JC, Sims SR, Prinsen MJ, Corbin DR, Tran M,
Sammons RD et al. (1993) Cholesterol oxidase: a potent
insecticidal protein active against boll weevil larvae.
Biochem Biophys Res Commun 196, 1406–1413.
4 Corbin DR, Grebenok RJ, Ohnmeiss TE, Greenplate
JT & Purcell JP (2001) Expression and chloroplast
targeting of cholesterol oxidase in transgenic tobacco
plants. Plant Physiol 126, 1116–1128.
5 Machang’u RS & Prescott JF (1991) Purification and
properties of cholesterol oxidase and choline phospho-
hydrolase from Rhodococcus equi. Can J Vet Res 55,
332–340.
6 Fuhrmann H, Dobeleit G, Bellair S & Guck T (2002)
Cholesterol oxidase and resistance of Rhodococcus equi
to peroxidative stress in vitro in the presence of choles-
terol. J Vet Med B Infect Dis Vet Public Health 49,
310–311.
7 Coulombe R, Yue KQ, Ghisla S & Vrielink A (2001)
Oxygen access to the active site of cholesterol oxidase
through a narrow channel is gated by an Arg-Glu pair.
J Biol Chem 276, 30435–30441.
8 Vrielink A, Lloyd LF & Blow DM (1991) Crystal struc-
ture of cholesterol oxidase from Brevibacterium steroli-
cum refined at 1.8 A resolution. J Mol Biol 219, 533–554.
9 Lario PI, Sampson N & Vrielink A (2003) Sub-atomic
resolution crystal structure of cholesterol oxidase: what

atomic resolution crystallography reveals about enzyme
mechanism and the role of the FAD cofactor in redox
activity. J Mol Biol 326, 1635–1650.
10 Piubelli L, Pedotti M, Molla G, Feindler-Boeckh S,
Ghisla S, Pilone MS & Pollegioni L (2008) On the oxy-
gen reactivity of flavoprotein oxidases: an oxygen access
tunnel and gate in Brevibacterium sterolicum cholesterol
oxidase. J Biol Chem 283, 24738–24747.
11 Chen L, Lyubimov AY, Brammer L, Vrielink A &
Sampson NS (2008) The binding and release of oxygen
and hydrogen peroxide are directed by a hydrophobic
tunnel in cholesterol oxidase. Biochemistry 47, 5368–
5377.
12 Pollegioni L, Wels G, Pilone MS & Ghisla S (1999)
Kinetic mechanisms of cholesterol oxidase from Strepto-
myces hygroscopicus and Brevibacterium sterolicum.
Eur J Biochem 264, 140–151.
13 Larkin MA, Blackshields G, Brown NP, Chenna R,
McGettigan PA, McWilliam H, Valentin F, Wallace IM,
Wilm A, Lopez R et al. (2007) Clustal W and Clustal X
version 2.0. Bioinformatics 23, 2947–2948.
14 Lim L, Molla G, Guinn N, Ghisla S, Pollegioni L &
Vrielink A (2006) Structural and kinetic analyses of the
H121A mutant of cholesterol oxidase. Biochem J 400,
13–22.
15 Isobe K, Shoji K, Nakanishi Y, Yokoe M & Wakao N
(2003) Purification and some properties of cholesterol
oxidase stable in detergents from gamma-proteo-
bacterium Y-134.
J Biosci Bioeng 95, 257–263.

16 Gadda G, Wels G, Pollegioni L, Zucchelli S, Ambrosius
D, Pilone MS & Ghisla S (1997) Characterization of
cholesterol oxidase from Streptomyces hygroscopicus
and Brevibacterium sterolicum. Eur J Biochem 250,
369–376.
17 Uwajima T, Yagi H & Terada O (1974) Properties of
crystalline 3beta-hydroxysteroid oxidase of Brevi-
bacterium sterolicum. Agric Biol Chem 38, 1149–1156.
18 Husain M & Massey V (1978) Reversible resolution of
flavoproteins into apoproteins and free flavins. Methods
Enzymol 53, 429–437.
19 Caldinelli L, Iametti S, Barbiroli A, Bonomi F, Fessas
D, Molla G, Pilone MS & Pollegioni L (2005) Dissect-
ing the structural determinants of the stability of choles-
terol oxidase containing covalently bound flavin. J Biol
Chem 280, 22572–22581.
20 Caldinelli L, Iametti S, Barbiroli A, Fessas D, Bonomi
F, Piubelli L, Molla G & Pollegioni L (2008) Relevance
of the flavin binding to the stability and folding of
engineered cholesterol oxidase containing noncovalently
bound FAD. Protein Sci 17, 409–419.
21 Motteran L, Pilone MS, Molla G, Ghisla S &
Pollegioni L (2001) Cholesterol oxidase from Brevibac-
terium sterolicum. The relationship between covalent
flavinylation and redox properties. J Biol Chem 276,
18024–18030.
22 Lyubimov AY, Chen L, Sampson NS & Vrielink A
(2009) A hydrogen-bonding network is important for
oxidation and isomerization in the reaction catalyzed by
cholesterol oxidase. Acta Crystallographica Section D

in press.
23 Massey V (1991) A simple method for the determina-
tion of redox potentials. In Flavins and Flavoproteins
(Curti BRS & Zanetti G eds), pp. 59–66. Walter de
Gruyter, Berlin, Germany, Como, Italy.
24 Grande HJ, Visser AJ, de Wit JL, Muller F & Veeger C
(1972) Spin label studies on the flavoproteins lipoamide
dehydrogenase and D- amino acid oxidase. Z Natur-
forsch [B] 27, 1058–1062.
25 Massey V & Ganther H (1965) On the interpretation of
the absorption spectra of flavoproteins with special
reference to D-amino acid oxidase. Biochemistry 4,
1161–1173.
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6840 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
26 Muller F, Mayhew SG & Massey V (1973) On the effect
of temperature on the absorption spectra of free and
protein-bound flavines. Biochemistry 12, 4654–4662.
27 Ghisla S & Edmondson D (2000) FLAVIN COEN-
ZYMES. In Encyclopedia of life sciences) John Wiley &
Sons, Ltd: Chichester />28 Ghisla S & Massey V (1986) New flavins for old:
artificial flavins as active site probes of flavoproteins.
Biochem J 239, 1–12.
29 Ghisla S, Massey V, Lhoste JM & Mayhew SG (1974)
Fluorescence and optical characteristics of reduced
flavins and flavoproteins. Biochemistry 13, 589–597.
30 Eventoff W & Rossmann MG (1975) The evolution of
dehydrogenases and kinases. CRC Crit Rev Biochem 3,
111–140.
31 Ohlsson I, Nordstrom B & Branden CI (1974) Struc-

tural and functional similarities within the coenzyme
binding domains of dehydrogenases. J Mol Biol 89,
339–354.
32 Hol WG, van Duijnen PT & Berendsen HJ (1978) The
alpha-helix dipole and the properties of proteins. Nature
273, 443–446.
33 Yue QK, Kass IJ, Sampson NS & Vrielink A (1999)
Crystal structure determination of cholesterol oxidase
from Streptomyces and structural characterization of
key active site mutants. Biochemistry 38, 4277–4286.
34 Cavener DR (1992) GMC oxidoreductases. A newly
defined family of homologous proteins with diverse
catalytic activities. J Mol Biol 223, 811–814.
35 Mattevi A, Fraaije MW, Mozzarelli A, Olivi L, Coda A
& van Berkel WJ (1997) Crystal structures and inhibitor
binding in the octameric flavoenzyme vanillyl-alcohol
oxidase: the shape of the active-site cavity controls
substrate specificity. Structure 5, 907–920.
36 Heuts DP, Scrutton NS, McIntire WS & Fraaije MW
(2009) What’s in a covalent bond? FEBS J 276, 3405–
3409.
37 Sherry B & Abeles RH (1985) Mechanism of action of
methanol oxidase, reconstitution of methanol oxidase
with 5-deazaflavin, and inactivation of methanol
oxidase by cyclopropanol. Biochemistry 24, 2594–2605.
38 Geissler J, Ghisla S & Kroneck PM (1986) Flavin-
dependent alcohol oxidase from yeast. Studies on the
catalytic mechanism and inactivation during turnover.
Eur J Biochem 160, 93–100.
39 Forneris F, Heuts DP, Delvecchio M, Rovida S, Fraaije

MW & Mattevi A (2008) Structural analysis of the
catalytic mechanism and stereoselectivity in Streptomy-
ces coelicolor alditol oxidase. Biochemistry 47, 978–985.
40 Ghisla S & Massey V (1991) L-Lactate Oxidase. In
Chemistry and Biochemistry of Flavoenzymes Vol II
(Mu
¨
ller F eds), pp. 243–289. CRC Press, Boca Raton.
41 Swoboda BE & Massey V (1965) Purification and
properties of the glucose oxidase from Aspergillus niger .
J Biol Chem 240, 2209–2215.
42 Fitzpatrick PF (2004) Carbanion versus hydride transfer
mechanisms in flavoprotein-catalyzed dehydrogenations.
Bioorg Chem 32, 125–139.
43 Massey V & Ghisla S (1983) The mechanism of action
of flavoprotein-catalyzed reactions. In Proceedings of
the 34th Colloquium on Biological Oxidations, Mosbach
Symposium (Sund H & Ullrich V eds), pp. 114–139.
Springer Verlag, Berlin, New York.
44 Ghisla S (1982) Dehydrogenation mechanisms in flavo-
protein catalysis. In Flavins and Flavoproteins, Proceed-
ings of the 7th International Symposium (Massey V &
Williams CH eds), pp. 133–142. Elsevier, North-Hol-
land, New York.
45 Yorita K, Janko K, Aki K, Ghisla S, Palfey BA &
Massey V (1997) On the reaction mechanism of l-lac-
tate oxidase: quantitative structure-activity analysis of
the reaction with para-substituted l-mandelates. Proc
Natl Acad Sci USA 94, 9590–9595.
46 Sampson NS & Vrielink A (2003) Cholesterol oxidases:

a study of nature’s approach to protein design. Acc
Chem Res 36, 713–722.
47 Mowat CG, Wehenkel A, Green AJ, Walkinshaw MD,
Reid GA & Chapman SK (2004) Altered substrate
specificity in flavocytochrome b2: structural insights
into the mechanism of l-lactate dehydrogenation.
Biochemistry 43, 9519–9526.
48 Liang ZX & Klinman JP (2004) Structural bases of
hydrogen tunneling in enzymes: progress and puzzles.
Curr Opin Struct Biol 14, 648–655.
49 Cha Y, Murray CJ & Klinman JP (1989) Hydrogen
tunneling in enzyme reactions. Science 243, 1325–
1330.
50 Kay CW, Mennenga B, Go
¨
risch H & Bittl R (2006)
Structure of the pyrroloquinoline quinone radical in
quinoprotein ethanol dehydrogenase. J Biol Chem 281,
1470–1476.
51 Tsai CL, Gokulan K, Sobrado P, Sacchettini JC &
Fitzpatrick TB (2007) Mechanistic and structural
studies of H373Q flavocytochrome b2: effects of
mutating the active site base. Biochemistry 46, 7844–
7851.
52 Umhau S, Pollegioni L, Molla G, Diederichs K, Welte
W, Pilone MS & 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 USA 97, 12463–
12468.

53 Nagel ZD & Klinman JP (2006) Tunneling and
dynamics in enzymatic hydride transfer. Chem Rev 106,
3095–3118.
54 Li J, Vrielink A, Brick P & Blow DM (1993) Crystal
structure of cholesterol oxidase complexed with a ste-
roid substrate: implications for flavin adenine dinucleo-
tide dependent alcohol oxidases. Biochemistry 32,
11507–11515.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6841
55 Kass IJ & Sampson NS (1998) Evaluation of the role of
His447 in the reaction catalyzed by cholesterol oxidase.
Biochemistry 37, 17990–18000.
56 Yamashita M, Toyama M, Ono H, Fujii I, Hirayama
N & Murooka Y (1998) Separation of the two
reactions, oxidation and isomerization, catalyzed by
Streptomyces cholesterol oxidase. Protein Eng 11,
1075–1081.
57 Lyubimov AY, Heard K, Tang H, Sampson NS &
Vrielink A (2007) Distortion of flavin geometry is linked
to ligand binding in cholesterol oxidase. Protein Sci 16,
2647–2656.
58 Yin Y, Sampson NS, Vrielink A & Lario PI (2001) The
presence of a hydrogen bond between asparagine 485
and the pi system of FAD modulates the redox
potential in the reaction catalyzed by cholesterol
oxidase. Biochemistry 40, 13779–13787.
59 Kass IJ & Sampson NS (1995) The isomerization
catalyzed by Brevibacterium sterolicum cholesterol
oxidase proceeds stereospecifically with one base.

Biochem Biophys Res Commun 206, 688–693.
60 Kass IJ & Sampson NS (1998) The importance of
Glu361 position in the reaction catalyzed by cholesterol
oxidase. Bioorg Med Chem Lett 8, 2663–2668.
61 Pollegioni L, Gadda G, Ambrosius D, Ghisla S &
Pilone MS (1999) Cholesterol oxidase from Streptomy-
ces hygroscopicus and Brevibacterium sterolicum: effect
of surfactants and organic solvents on activity.
Biotechnol Appl Biochem 30, 27–33.
62 Nishiya Y, Yamashita M, Murooka Y, Fujii I &
Hirayama N (1998) Effect of non-ionic detergents on
apparent enzyme mechanism: V121A mutant of Strepto-
myces cholesterol oxidase endowed with enhanced
sensitivity towards detergents. Protein Eng 11, 609–611.
63 Ahn KW & Sampson NS (2004) Cholesterol oxidase
senses subtle changes in lipid bilayer structure.
Biochemistry 43, 827–836.
64 Xiang J & Sampson NS (2004) Library screening
studies to investigate substrate specificity in the reaction
catalyzed by cholesterol oxidase. Protein Eng Des Sel
17, 341–348.
65 Sampson NS, Kass IJ & Ghoshroy KB (1998)
Assessment of the role of an omega loop of cholesterol
oxidase: a truncated loop mutant has altered substrate
specificity. Biochemistry 37, 5770–5778.
66 Toyama M, Yamashita M, Yoneda M, Zaborowski A,
Nagato M, Ono H, Hirayama N & Murooka Y (2002)
Alteration of substrate specificity of cholesterol oxidase
from Streptomyces sp. by site-directed mutagenesis.
Protein Eng 15, 477–484.

67 Ghisla S & Massey V (1989) Mechanisms of
flavoprotein-catalyzed reactions. Eur J Biochem 181,
1–17.
68 Massey V (2000) The chemical and biological versatility
of riboflavin. Biochem Soc Trans 28, 283–296.
69 Massey V (1994) Activation of molecular oxygen by
flavins and flavoproteins. J Biol Chem 269, 22459–
22462.
70 Massey V, Schopfer LM & Anderson RF (1988) Struc-
tural determinants of the oxygen reactivity of different
classes of flavoproteins. Prog Clin Biol Res 274
, 147–
166.
71 English BP, Min W, van Oijen AM, Lee KT, Luo G,
Sun H, Cherayil BJ, Kou SC & Xie XS (2006) Ever-
fluctuating single enzyme molecules: Michaelis-Menten
equation revisited. Nat Chem Biol 2, 87–94.
72 Lerch HP, Rigler R & Mikhailov AS (2005) Functional
conformational motions in the turnover cycle of choles-
terol oxidase. Proc Natl Acad Sci USA 102, 10807–
10812.
73 Lu HP, Xun L & Xie XS (1998) Single-molecule
enzymatic dynamics. Science 282, 1877–1882.
74 Xie XS & Lu HP (1999) Single-molecule enzymology.
J Biol Chem 274, 15967–15970.
75 Ghisla S & Thorpe C (2004) Acyl-CoA dehydrogenases.
A mechanistic overview. Eur J Biochem 271, 494–508.
76 Kim BE, Lee JH, Kim HS, Kwon OJ, Jeong SW &
Kim IK (2004) Involvement of Sox-4 in the cytochrome
c-dependent AIF-independent apoptotic pathway in

HeLa cells induced by Delta12-prostaglandin J2. Exp
Mol Med 36, 444–453.
77 Lakowicz JR & Weber G (1973) Quenching of fluores-
cence by oxygen. A probe for structural fluctuations in
macromolecules. Biochemistry 12, 4161–4170.
78 Lakowicz JR & Weber G (1973) Quenching of protein
fluorescence by oxygen. Detection of structural
fluctuations in proteins on the nanosecond time scale.
Biochemistry 12, 4171–4179.
79 Giangiacomo L, Mattu M, Arcovito A, Bellenchi G,
Bolognesi M, Ascenzi P & Boffi A (2001) Monomer-
dimer equilibrium and oxygen binding properties of
ferrous Vitreoscilla hemoglobin. Biochemistry 40, 9311–
9316.
80 Iwata S, Ostermeier C, Ludwig B & Michel H (1995)
Structure at 2.8 A resolution of cytochrome c oxidase
from Paracoccus denitrificans. Nature 376, 660–669.
81 Riistama S, Puustinen A, Garcia-Horsman A, Iwata S,
Michel H & Wikstrom M (1996) Channelling of dioxy-
gen into the respiratory enzyme. Biochim Biophys Acta
1275, 1–4.
82 Eriksson M, Jordan A & Eklund H (1998) Structure of
Salmonella typhimurium nrdF ribonucleotide reductase
in its oxidized and reduced forms. Biochemistry 37,
13359–13369.
83 Soulimane T, Buse G, Bourenkov GP, Bartunik HD,
Huber R & Than ME (2000) Structure and mechanism
of the aberrant ba(3)-cytochrome c oxidase from Ther-
mus thermophilus. EMBO J 19, 1766–1776.
84 Stillman TJ, Hempstead PD, Artymiuk PJ, Andrews

SC, Hudson AJ, Treffry A, Guest JR & Harrison PM
Biochemistry of cholesterol oxidase A. Vrielink and S. Ghisla
6842 FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS
(2001) The high-resolution X-ray crystallographic struc-
ture of the ferritin (EcFtnA) of Escherichia coli; com-
parison with human H ferritin (HuHF) and the
structures of the Fe(3+) and Zn(2+) derivatives. J Mol
Biol 307, 587–603.
85 Saam J, Ivanov I, Walther M, Holzhutter HG & Kuhn
H (2007) Molecular dioxygen enters the active site of
12 ⁄ 15-lipoxygenase via dynamic oxygen access channels.
Proc Natl Acad Sci USA 104, 13319–13324.
86 Moustafa IM, Foster S, Lyubimov AY & Vrielink A
(2006) Crystal structure of LAAO from Calloselas-
ma rhodostoma with an l-phenylalanine substrate:
insights into structure and mechanism. J Mol Biol 364,
991–1002.
87 Whitby LG (1954) Preparation of flavinadenine
dinucleotide (FAD). Biochim Biophys Acta 15, 148–
149.
A. Vrielink and S. Ghisla Biochemistry of cholesterol oxidase
FEBS Journal 276 (2009) 6826–6843 ª 2009 The Authors Journal compilation ª 2009 FEBS 6843