JOurnail or unlderraduate PRs.earcn
Oluinie . , is.ue 1 - S-eptn.eiL,-er 2':11:14
Mechanism of Oxidative Inactivation of Acinetobacter sp. NCIMB
9871 Cyctohexanone Monooxygenase
The Acinetobacter sp. NCIMB 9871 flavoprotein cyclohexanone monooxygenase (CHMO) catalyzes the regio-
and enantioselective conversion of a wide variety of cyclic ketones to chiral lactones, important drug
precursors. CHMO is not in widespread use because the expensive and unstable redox cofactor NADPH must
be supplied continuously. Whole-cell techniques, which rely on the pentose phosphate pathway and the Krebs cycle
to regenerate NADPH from glucose reducing equivalents, provide an economical means of cofactor
regeneration. These techniques are not yet widely used because they are accompanied by oxidative inactivation
of CHMO after 24 hours. We propose a method of using this enzyme without the cofactor NADPH by formation of
the necessary C4_ -hydroperoxy FAD intermediate using hydrogen peroxide. However, we demonstrate by
using spectrophotometric assays that H202 inactivates the enzyme by oxidizing two cysteine residues to a
disulfide or thiolsulfinate bond, thus causing an irreversible conformational distortion and rendering one of
the cysteines incapable of performing its binding function. This event causes the flavin binding equilibrium to
shift dramatically, releasing this cofactor to leave behind inactive apo-CHMO. We demonstrate that NADP+
prevents such damage, and based on this, we show that the C4_-hydroperoxy intermediate cannot form.
However, we suggest, based on the oxidative inactivation scheme, that site-directed mutagenesis of the
critical binding cysteine to an oxidant-insensitive serine would prevent oxidative inactivation under whole-
cell bioprocess conditions. This would optimize the bioprocess for the industrial scale, and efforts to create
this mutant are underway.
Many bacterial species have evolved to fill environmental niches where they must survive on non-
carbohydrate sources of carbon and energy. This condition has favored the evolution of enzymes with
significant potential as stereoselective catalysts in organic synthesis. The Acinetobacter sp. NCIMB 9871
flavoprotein cyclohexanone monooxygenase (CHMO) is such an enzyme; it catalyzes the Baeyer-Villiger oxidations
of a large variety of prochiral and racemic cyclic ketones to chiral lactones with high enantioselectivity
and regioselectivity (Figure 1). , Production of numerous drugs and organic compounds relies on these products
as building blocks.
+ 02 + NADPH r O 0 + HPO + NADPW
cyc ihnanone ï¿½-caprolnatone
Figure 1. CHMO accepts many analogues of cyclohexanone.
Direct use of CHMO requires regeneration of the expensive and unstable redox cofactor reduced nicotinamide
adenine dinucleotide phosphate (NADPH), and the search for economical methods of regeneration has led to
the development of whole-cell biocatalysis techniques that, while achieving this end (by means of cellular
metabolic regeneration of NADPH using reducing equivalents from glucose), nevertheless require non-growing E.
coli cells and a complex fermentation apparatus. 1 Mechanistic studies have shown that a flavin C4_ -
hydroperoxide intermediate is responsible for oxygenation of the substrate, and if this intermediate could be
formed by direct reaction of the oxidized flavin with H202, as shown in Figure 2, it would dramatically simplify
the use of this enzyme.1,2, Previous attempts to accomplish this have not been successful; however, attempts
to determine active-site residues and to utilize CHMO as a sulfoxidation catalyst have demonstrated that
mild oxidants, specifically p-hydroxymercuribenzoate, that convert cysteine thiols to sulfenic acids inactivate
the enzyme. It is therefore possible that direct hydroperoxidation has failed because H202 oxidizes the
cysteine residues, rendering the enzyme inactive. More important, mechanistic studies do not support the
chemical involvement of cysteine at the active site, suggesting that slight modification of these residues might
not inactivate the enzyme.
EifAMt NADt 1--F-,IN)(1OT N ^IJP )W-FADH(OX)I NAN"
I AIY N....PI. k
.w . ... . ..
NAADF *4LDW C.AMr '" , 1 "
CA.ru I k -AUr
e 4 .
bu FCricjt r*ocG an |
Wivs r ar th n
&FAW N.&JDP' E-F,-I_ O VP r N DP-
IEF]ADr. NADIP ,LFADrrP r N\A W
Figure 2. Top: CHMO mechanism of catalysis proposed by Sheng et al.2 Bottom: Proposed
alternate mechanism. The c4-a hydroperoxy intermediate forms directly between the covalently
bound FAD cofactor and H202.
We investigate three aspects of this problem. First, we characterize the nature and extent of the H202-
mediated oxidative damage to CHMO. Second, a means of preventing such damage is developed to
allow determination of C4_ -hydroperoxy formation between enzyme-bound FAD and H202. Finally, even under
the whole-cell bioprocess conditions mentioned above, the monooxygenation reaction shown in Figure 1 arrests
after 24 hours regardless of conditions, a phenomenon that has been attributed to enzyme deactivation by
damaging oxidative intermediates that build up in the cells. We propose a method for preventing such damage
based on the results of the characterization of H202-mediated damage to the enzyme. Even if the C4_ -
hydroperoxy intermediate cannot form, the enzyme can be made oxidant-insensitive using this method,
optimizing this environmentally friendly whole-cell bioprocess for the industrial scale by allowing indefinite
Cyclohexanone and NBD chloride were purchased from Sigma-Aldrich. Dithiothreitol, IPTG, and all buffers,
growth media, and salts were purchased from Fisher. NADPH, sodium NADP, NADH, and sodium NAD were
purchased from Biocatalytics, Inc. The plasmid construct pMM4, which contains the encoding region for CHMO,
was made in our lab by Marko Mihovilovic. All E. coli strains were obtained from Invitrogen.
Preparation of Purified CHMO
The method of Sheng et al. was used, with the following modifications.2 E. coli BL21(DE3)(pMM4) was used as
the expression strain, and ampicillin (0.20 g/L) was used as the antibiotic rather than kanamycin. Four liters
of culture were grown in a New Brunswick Fermenter at 37 OC with 700-rpm stirring with 4 L/min filtered airflow
rate. Induction was carried out for only 3 hours. Sonication was performed for ten 20-second periods, with cooling
on ice in between. Centrifugation of cellular debris was performed at 41700 _ g three times due to the
increased density that resulted from using cell resuspension volumes according to Sheng et al. with a larger
original culture. A Q-Sepharose column was used for anion exchange. The Amicon DyeMatrex Red A column was
run at only 0.8 mL/min with a fraction size of 6.25 mL. CHMO-containing fractions were identified by yellow
color. Two peaks eluted from the anion exchange resin at times corresponding to 15 % and 37 % maximum
ionic strength. Elution from the DyeMatrex Red A column occurred in one peak at a time corresponding to the
buffer switch described by Sheng et al. Purity was assessed using the methods of Sheng et al.2 SDS-PAGE
analysis showed barely visible background bands, indicating possible impurity. The activity data are presented
in Table 1. Purified CHMO was dialyzed as described by Sheng et al. and stored at 4 OC in 20 mM
potassium phosphate buffer at pH 7.2.
Purification data for CHMO preparation.
(m L) Concentration (g/
Activity (U/ Total
mL) Activity (U) ActivityYield
Activity (U/ Factor
50% -80 % (NH4)2SO4 Saturation Pellet 152 2189 33267 317 48228 100% 13 1
After Anion Exchange 35 1493 5224 3969 1389 2880% 27 20
After DyeMatrex Red A 11 225 247 59 649 1346% 26 20
Steady-State Enzyme Activity Assays
Steady-state activity was measured by following NADPH consumption at 340 nm as described by Sheng et al.,
with the following modifications.2 Reactions were carried out at 300C, and 50 mM cyclohexanone was used to
initiate the reactions. A tandem mixing cuvette (Starna) was used in some assays, as described in the Results
and Discussion section. Absorbance was measured using the HP 8450 A diode-array spectrophotometer
and accompanying kinetics software.
Isolation and Reconstitution of apo-CHMO
FAD was dissociated from apoenzyme using the procedure of Donoghue et al. Apoenzyme was reconstituted by
1-minute incubation at 300C with 10 _M FAD.
RESULTS AND DISCUSSION
Effect of Hydrogen Peroxide on CHMO Activity
A normal assay, which follows NADPH consumption at 340 nm as described in the experimental section (from here
on referred to as a normal assay), without CHMO was carried out; then another identical assay with 2.6 mM
H202 was performed. The rates were within one standard deviation of each other, demonstrating that
direct oxidation of NADPH by H202 is not sufficient to result in false rate measurements. For all subsequent
rate measurements and activity assays, this background NADPH decomposition rate (3.2 nM/s) was subtracted.
Two cuvettes containing 0.0255 mg/mL CHMO in normal assay buffer were treated as follows. To one was added
2.6 mM H202, and to the control, nothing. Each was incubated 5 min at 300C and then subjected to normal
assay. The NADPH consumption rate of the H202-treated sample was 0.7% of the control. Thus, H202
inactivates CHMO, and this may be the reason that attempts to use it to accomplish monooxygenation by
the alternate mechanism shown in Scheme 2 have failed.
Characterization of H202-Mediated Inactivation of CHMO
To determine whether H202 reacts with CHMO to deactivate it, or acts as or produces an inhibitor, a sample
of 0.0225 g/L CHMO in normal assay buffer was treated with 0.0026 g/L catalase followed by 2.6 mM H202
and incubated 5 min at 300C. A control was treated identically but without H202. The rates of these
samples determined by normal assay were within less than a standard deviation of one another, demonstrating
that catalase, which catalyzes the decomposition of H202, protects CHMO from H202-mediated inactivation.
Next, 0.0225 g/L CHMO in normal assay buffer was incubated with H202 as before. Then, catalase was added
to 0.026 g/L followed by addition of fresh CHMO to a concentration of 0.0225 g/L. A normal assay was
then performed on this sample, as well as a control solution that was treated identically but without H202. The
H202-treated sample rate was 84.6 % that of the control. Thus, production of inhibitors or inhibition by H202 is
not sufficient to account for the complete loss of enzyme activity that occurs with exposure to H202.
Chemical reaction of H202 with CHMO is therefore the mechanism of inactivation.
CHMO contains five cysteine residues. At least one of these residues is sensitive to oxidation by
p-hydroxymercuribenzoate, with concomitant loss of enzyme activity.4 Furthermore, cysteine residues, unlike
other common amino acid residues, have been shown in surface studies and in protein studies to be oxidized
by H202 to sulfenic acids, reactive species that rapidly condense with vicinal cysteines to form disulfide bonds or
with vicinal sulfenic acids to form thiolsulfinates (Figure 3). Further oxidation to sulfonic acids occurs if
condensation is not rapid.7 The reducing agent dithiothreitol (DTT) selectively reduces thiolsulfinates and
disulfide bonds to thiols, the normal cysteine oxidation state, but cannot reduce sulfonic acids.7 Based on these
facts, a sample of 0.0225 g/L CHMO in normal assay buffer was treated with 2.6 mM H202 and incubated 5 min
at 300C. Catalase was then added to a concentration of 0.026 g/L to destroy residual H202. This sample was
then treated with 0.004 M DTT (excess) and incubated as before. A normal assay was performed on this sample
and an identical control that was not exposed to H202. A third control was treated identically to the original
sample, except it was not treated with DTT. The results are shown in Table 2.
/ lc ON
^ sulfcnic acid
0 H R^
F r Od n
Figure 3. Oxidation pathways of cysteine.
Residual Activity of Samples Treated with H202 with and without Subsequent Treatment with DTT.
Sample H202 DTT Residual Activity (%)
1 (control) No Yes 100
2 Yes Yes 315
3 Yes No 7
These data demonstrate that DTT causes significant reactivation of H202-treated CHMO. Thus, the reactions shown
in Figure 3 likely occur, with reactivation representing reduction of disulfide and thiolsulfinate bonds to
thiols. Reactivation is not complete either because some cysteines are oxidized to the sulfonic acid oxidation
state, which cannot be restored, because a conformational change that releases the tightly-bound flavin
accompanies H202-mediated inactivation, or because the FAD cofactor also receives oxidative damage.
To determine which portion of the holoenzyme-the FAD cofactor, the apoprotein, or both-is damaged by
H202, CHMO in normal assay buffer was incubated 5 min at 300C with 2.6 mM H202. The apoprotein and FAD
were then separated as described in the Experimental Procedures section, and the spectrum of the isolated flavin
was found to be identical to that of untreated FAD. The apoprotein was reconstituted with fresh FAD and had
a normal assay rate that was 1.5 % of that of a control that was treated identically but without H202. The
results demonstrated that inactivating damage occurs to the peptide rather than the FAD portion of this
enzyme, consistent with the cysteine oxidation hypothesis. Furthermore, ammonium sulfate precipitation of
normal CHMO yields a bright yellow pellet and colorless supernatant, whereas precipitation of H202-
inactivated CHMO yields a yellow pellet and yellow supernatant, supporting the notion of flavin release due
to conformational change as the reason for only partial reactivation by DTT. Spectra of untreated CHMO and
pure FAD of the same molar concentration were compared. The peak heights and shapes of the spectrum of
CHMO treated with H202 are shifted toward those of the free flavin, further supporting this hypothesis.
To demonstrate the results more clearly, the spectral changes accompanying the H202 treatment were
observed. Spectral shift from that of bound FAD to that of free FAD was observed over the 5 min period. It
is therefore probable that inactivation occurs by oxidation of one or both of a pair of vicinal cysteine residues
to unstable sulfenic acids, which form a strong, covalent disulfide or thiolsulfinate bond that contorts the enzyme in
a way that releases the catalytically critical flavin cofactor, rendering the enzyme inactive. There is no need to
invoke additional sulfonic acids to explain incomplete DTT-mediated activity restoration with this scheme.
To further test our results, the reagent NBD chloride, which selectively titrates thiols (but not any other
oxidation state of sulfur) with concomitant measurable absorbance change, was added to 0.388 mM to one half of
a tandem mixing cuvette; to the other half was added normal CHMO (2.25 g/L) treated with 0.26 g/L catalase
or CHMO (2.25 g/L) incubated with 0.26 M H202 for 5 min at 300ï¿½C followed by 0.26 g/L catalase. The
absorbance difference before and after mixing at 420 nm was used to determine that 2.0 titratable thiols are
lost upon treatment with H202, consistent with the scheme previously proposed. If this scheme is correct, addition
of excess FAD after DTT-reactivation should result in nearly complete activity restoration (assuming renaturation
is spontaneous). We found, however, that no activity change occurs when DTT-reactivated CHMO is treated
with excess FAD. This set of results suggests that oxidative denaturation is not entirely reversible. DTT-
reduced cysteines regain the chemical state necessary for substrate binding (see below), but no longer have
Protection of CHMO from H202-Mediated Inactivation
As mentioned in the introduction, Donoghue et al. found that titration of one cysteine with p-
hydroxymercuribenzoate resulted in nearly complete loss of activity.6 In the tests described above, NADPH was
not added before H202 treatment. However, Donoghue et al. found that NADPH protected CHMO from
p-hydroxymercuribenzoate inactivation, and they hypothesized that this catalytically critical cysteine is involved
in NADPH binding, FAD binding, or both, so that the presence of NADPH in the active site makes this
cysteine inaccessible for p-hydroxymercuribenzoate attack. A binding role for this cysteine is certainly consistent
with the flavin binding shift associated with inactivation as reported here. To determine if NADPH is capable
of blocking H202 inactivation as well, we pre-incubated CHMO with excess NADPH (160 _M) for 1 minute at 300
C, followed by a 5-min H202 (2.6 mM) incubation as before. This sample had a normal assay rate 79.3% that of
a control (specific activity 2.392 U/mg versus 3.016 U/mg for a control and 0.2144 U/mg for CHMO not protected
by NADPH) not exposed to H202. Given that direct reaction between H202 and NADPH does not occur
as demonstrated above, this result demonstrates that similar protection is provided against H202
inactivation (Donoghue et al. reported 95% residual activity). Neither NADH nor NAD+ provide such protection,
nor were they found to have any inhibitory effect, consistent with the hypothesis that they cannot enter the
Formation of the C4_-Hydroperoxy Intermediate
Donoghue et al. also found that NADP+ can protect CHMO, leaving the enzyme with 50% residual activity after
p-hydroxymercuribenzoate attack.6 We repeated the preceding pre-incubation experiment with NADP+ rather
than NADPH, and observed a 57 % residual activity (1.719 U/mg), suggesting that the same critical cysteine
residue is being protected. The unsuccessful attempts to catalyze the oxidation of cyclohexanone to _-
caprolactone using H202 to achieve the proposed alternate mechanism shown in 2 were performed in the presence
of a large excess of NADP+. Therefore, CHMO was protected from H202-mediated inactivation during
these experiments; oxidative damage did not render the enzyme inactive, despite the failure to form any
lactone product. We thereby deduce that the flavin C4_-hydroperoxy intermediate cannot form. The
alternate mechanism proposed in Figure 2 does not occur, and abiotic use of CHMO without NADPH is not
possible using H202.
These results demonstrate that H202 cannot be used to obviate cofactor regeneration. The optimization of whole-
cell techniques is thus critical. The results suggest a mechanism for the oxidative inactivation of CHMO after 24
hours that currently prevents such optimization. A single catalytically critical cysteine, likely involved in
substrate binding, is oxidized to an unstable sulfenic acid, which then rapidly condenses with a vicinal
cysteine sulfenic acid or thiol to form a thiolsulfinate or disulfide bond, respectively. Due to permanent
conformational distortion (possibly the result of conversion of less critical cysteine thiols to sulfonic acids),
and because the reacted cysteine can no longer serve its binding function, the flavin binding equilibrium
shifts strongly toward the free form, releasing the FAD cofactor and rendering the enzyme inactive.
CHMO might be made oxidant-insensitive if this catalytically-critical cysteine were converted to the
oxidant-insensitive serine residue by site-directed mutagenesis. Serine has similar polarity and size and so might be
a functional replacement. In addition, the mechanism shown in Figure 2 shows that cysteine is not
chemically involved in catalysis. Thus, mutagenesis to serine will not necessarily result in loss of activity.
However, the serine residue will be capable of forming hydrogen bonds with capable moieties. Worse yet,
the cysteine residue is a weak acid, and a significant portion of the cysteine thiols may be ionized at pH 8.0,
the optimal pH of CHMO (average serine pKA is 8.18). Ionization may be necessary for this cysteine to perform
its critical binding function, meaning that replacement with the invariably neutral serine residue might result
in inactive enzyme. Furthermore, a crystal structure for this enzyme is not available. Mutagenesis and testing
of cysteine_serine mutants will therefore have to be random. This work is currently in progress, with the goal
of developing an active, oxidant-insensitive CHMO mutant. Such a mutant would be optimized for whole-cell
industrial biocatalysis, a process that is more efficient and environmentally friendly than conventional chiral
1. Walton, A. Z.; Stewart, J. D. An Efficient Enzymatic Baeyer-Villiger Oxidation by Engineered Escherichia coli
Cells under Non-Growing Conditions. Biotechnol. Prog. 2002, 18, 262-268.
2. Sheng, D.; Ballou, D. P.; Massey, V. Mechanistic Studies of Cyclohexanone Monooxygenase: Chemical Properties
of Intermediates Involved in Catalysis. Biochemistry 2001, 40, 11156-11167.
3. Mazzini, C.; Lebreton, J.; Furstoss, R. Flavin-Catalyzed Baeyer-Villiger Reaction of Ketones: Oxidation
of Cyclobutanones to _ Lactones Using Hydrogen Peroxide. J. Org. Chem. 1996, 61, 8-9.
4. Latham, J. A.; Walsh, C. Mechanism-Based Inactivation of the Flavoenzyme Cyclohexanone Oxygenase
during Oxygenation of Cyclic Thiol Ester Substrates. J. Am. Chem. Soc. 1987, 109, 3421-3427.
5. Walton, A. Improving the Efficiency of Enzymatic Baeyer-Villiger Oxidations with Whole Engineered Escherichia
coli Cells. M.S. Thesis, University of Florida, Gainesville, FL, 32611.
6. Donoghue, N. A.; Norris, D. B.; Trudgill, P. W. The Purification and Properties of Cyclohexanone Oxygenase
from Nocardia globerula CL1 and Acinetobacter NCIB 9871. Eur. J. Biochem. 1976, 63, 175-192.
7. Pavlovic, E.; Quist, A. P.; Gelius, U.; Nyholm, L.; Oscarson, S. Generation of Thiolsulfinates/Thiolsulfonates
by Electrooxidation of Thiols on Silicon Surfaces for Reversible Immobilization of Molecules. Langmuir. 2003,
8. Claiborne, A.; Miller, H.; Parsonage, D.; Ross, R. P. Protein-Sulfenic Acid Stabilization and Function in
Enzyme Catalysis and Gene Regulation. FASEB J. 1993, 7, 1483-1490.
9. Ryerson, C. C.; Ballou, D. P.; Walsh, C. Mechanistic Studies on Cyclohexanone Oxygenase. Biochemistry 1982,
10. Nelson, D. L.; Cox, M. M. Lehninger Principles of Biochemistry, 3rd Ed.; Worth Publishers: New York, 1999.
Back to the Journal of Undergraduate Research
College of Liberal Arts and Sciences I University Scholars Program I University of Florida | UNIVERSITY,-
Uï¿½ University of Florida, Gainesville, FL 32611; (352)FLORIDA
@ University of Florida, Gainesville, FL 32611; (352) 846-2032.