Discovery of the curcumin metabolic pathway involving a unique enzyme in an intestinal microorganism

Article abstract of DOI:10.1073/pnas.1016217108

Polyphenol curcumin, a yellow pigment, derived from the rhizomes of a plant (Curcuma longa Linn) is a natural antioxidant exhibiting a variety of pharmacological activities and therapeutic properties. It has long been used as a traditional medicine and as a preservative and coloring agent in foods. Here, curcumin-converting microorganisms were isolated from human feces, the one exhibiting the highest activity being identified as Escherichia coli. We are thus unique in discovering that E. coli was able to act on curcumin. The curcumin-converting enzyme was purified from E. coli and characterized. The native enzyme had a molecular mass of about 82 kDa and consisted of two identical subunits. The enzyme has a narrow substrate spectrum, preferentially acting on curcumin. The microbial metabolism of curcumin by the purified enzyme was found to comprise a two-step reduction, curcumin being converted NADPH-dependently into an intermediate product, dihydrocurcumin, and then the end product, tetrahydrocurcumin. We named this enzyme "NADPH-dependent curcumin/dihydrocurcumin reductase" (CurA). The gene (curA) encoding this enzyme was also identified. A homology search with the BLAST program revealed that a unique enzyme involved in curcumin metabolism belongs to the medium-chain dehydrogenase/reductase superfamily.

Full text of DOI:10.1073/pnas.1016217108

Discovery of the curcumin metabolic pathway
involving a unique enzyme in an
intestinal microorganism
Azam Hassaninasab1, Yoshiteru Hashimoto1, Kaori Tomita-Yokotani, and Michihiko Kobayashi2
Institute of Applied Biochemistry and Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan
Edited* by Arnold L. Demain, Drew University, Madison, NJ, and approved March 2, 2011 (received for review November 3, 2010)
Polyphenol curcumin, a yellow pigment, derived from the rhizomes
of a plant (Curcuma longa Linn) is a natural antioxidant exhibiting
a variety of pharmacological activities and therapeutic properties. It
has long been used as a traditional medicine and as a preservative
and coloring agent in foods. Here, curcumin-converting microor-
ganisms were isolated from human feces, the one exhibiting the
highest activity being identified as Escherichia coli. We are thus
unique in discovering that E. coli was able to act on curcumin. The
curcumin-converting enzyme was purified from E. coli and charac-
terized. The native enzyme had a molecular mass of about 82 kDa
and consisted of two identical subunits. The enzyme has a narrow
substrate spectrum, preferentially acting on curcumin. The micro-
bial metabolism of curcumin by the purified enzyme was found to
comprise a two-step reduction, curcumin being converted NADPH-
dependently into an intermediate product, dihydrocurcumin, and
then the end product, tetrahydrocurcumin. We named this enzyme
“NADPH-dependent curcumin/dihydrocurcumin reductase” (CurA).
The gene (curA) encoding this enzyme was also identified. A ho-
mology search with the BLAST program revealed that a unique
enzyme involved in curcumin metabolism belongs to the medium-
chain dehydrogenase/reductase superfamily.
microorganism is involved in the curcumin pathway (curcumin →
dihydrocurcumin → tetrahydrocurcumin).
Results
Discovery of Curcumin-Converting Microorganisms. Microorganisms
with curcumin-converting ability were isolated from feces of
two healthy adults who had taken no antibiotics or other drugs
for the 3 mo before sampling. We then selected the intestinal
microorganism exhibiting the highest specific activity (0.00004
units/mg) as to curcumin conversion. A 16S rRNA sequence of
the selected microorganism (accession number AB609595)
exhibited the highest similarity (99%) to those of Escherichia
coli H10407 strain [accession number FN649414 (region:
4363979..4365520)] (8), O55:H7 strain CB9615 [accession num-
ber CP001846 (region: 227062..228603)] (9), and BW2952 strain
[accession number CP001396 (region: 4095850..4097391)] (10);
therefore, it was identified as E. coli. The isolated microorganism
was then cultured on MacConkey (11) and Levine EMB (12) agar
to examine the lactose fermentation reaction and to observe
the colony morphology on these media. On MacConkey agar, the
isolated microorganism produced red colonies, depending upon
the ability to ferment lactose: fermentation of this sugar causes
the medium pH to drop, leading to darkening of the medium.
Growth of the microorganism on Levine EMB agar produced
black colonies with a greenish-black metallic sheen because of
lactose fermentation: lactose fermentation produces acid and
encourages dye absorption by colonies, which become colored
black with a metallic green sheen because of the metachromatic
properties of the dyes. These findings strongly supported that the
isolated microorganism is E. coli.
microbial screening bioconversion
|
ince the dawn of civilization, natural compounds have been
S
used as a source of medicine and foods based on their tre-
mendous biological activities. One of the natural compounds
that has been documented as having been used as a spice and
a pigment since 1900 B.C. is curcumin (1). Curcumin, defined as
a bis-α, β-unsaturated β-diketone (Fig. S1), has been found ex-
clusively in the roots of Curcuma longa Linn (Zingiberaceae) (2).
Curcumin has many different applications because it has a sur-
prisingly wide variety of beneficial uses, including food colora-
tion, cosmetics utility, and fabric dying. In addition, curcumin has
a wide range of medical properties, including antitumor, anti-
inflammatory, antioxidant, anticancer, and analgesic uses (3).
Recently, curcumin has been used in the treatment of Alz-
heimer’s disease (4) and malaria (5), and improvement of wound
healing (6). Curcumin in the diet is partially absorbed in the
intestine (7). A considerable portion of the ingested curcumin
reaches the cecum and colon, where a large population of in-
digenous bacteria exists. Although analysis of curcumin metab-
olism and biosynthesis would lead to the discovery of other
unknown advantages of this compound, there has neither been
purification of enzymes involved in the metabolic pathway for
curcumin nor identification of their genes in living organisms. In
addition, no information is available regarding the effect of the
intestinal bacteria on curcumin metabolism.
We also detected the curcumin-converting ability of E. coli
strain K-12, substrain DH10B. The specific activity (0.0005 units/
mg) as to curcumin transformation in the cell-free extracts of this
strain is higher than that of the isolated E. coli from human feces.
Therefore, we used E. coli DH10B instead of the isolated E. coli
for further experiments because of its high specific activity and
whole-genome determination (13).
Addition of Various Cofactors to the Curcumin-Converting Enzyme.
Purification of the enzyme was carried out from an extract of
Author contributions: A.H., Y.H., and M.K. designed research; A.H., Y.H., and K.T.-Y.
performed research; A.H., Y.H., K.T.-Y., and M.K. analyzed data; and A.H., Y.H., and
M.K. wrote the paper.
The authors declare no conflict of interest.
*This Direct Submission article had a prearranged editor.
Data deposition: The sequences reported in this paper have been deposited in the DNA
Databank of Japan (DDBJ) database (accession nos. AB609595 and AB583756).
1A.H. and Y.H. contributed equally to this work.
2To whom correspondence should be addressed at: Graduate School of Life and Environ-
mental Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japan.
Fax +81-29-853-4605.
Studies on such curcumin-converting microorganisms and func-
tional analyses of the curcumin-converting enzymes and genes
would facilitate clarification of the metabolism of curcumin. In the
present article, we report the isolation of curcumin-metabolizing
microorganisms from human feces and the purification and char-
acterization of a curcumin-converting enzyme. We are unique in
demonstrating that a NADPH-dependent enzyme in an intestinal
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1016217108/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1016217108
PNAS
|
April 19, 2011
|
vol. 108
|
no. 16
|
6615–6620

cells cultured in the optimum medium for curcumin metabolism
(SI Results). Interestingly, the curcumin-converting enzyme ac-
tivity decreased during the enzyme purification. Dialysis of cell-
free extracts caused complete loss of the initial enzyme activity.
To determine the effect of the addition of cofactors on the en-
zyme activity, various kinds of cofactors at a final concentration
of 0.5 mM were added to the reaction mixture, separately. The
loss of activity was reversible and the activity was drastically re-
covered on the addition of NADPH to the assay mixture (Table
S1), which allowed characterization of this enzyme in various
ways after purification. Therefore, enzyme assay mixtures A and
B were fortified with NADPH.
band consistent with the purified enzyme from E. coli DH10B; the
amount of CurA was about more than 70% of the total soluble
proteins. These findings indicated that overproduction of the
curcumin-converting enzyme in the active form was attained.
The curcumin-converting enzyme produced in the trans-
formed cells was purified to homogeneity through the four-step
purification procedure involving three chromatography columns,
as described in SI Materials and Methods (Table S3). The purified
enzyme gave only one band on SDS/PAGE (Fig. S3), and ap-
parent molecular mass of the purified enzyme was the same as
that from E. coli DH10B. The molecular mass of the purified
enzyme was also determined to be 37,513 Da by MALDI-TOF/
MS, as described in SI Materials and Methods, this value being
the same as that from E. coli DH10B (37,589 Da). These data
indicated that the curcumin-converting enzyme purified from
E. coli BL21-CodonPlus(DE3)-RIL carrying pET-curA is iden-
tical to that from E. coli DH10B. Because the following analyses
required a large amount of the enzyme and a large amount of the
purified enzyme could not be obtained from E. coli DH10B, we
used purified CurA from E. coli BL21-CodonPlus(DE3)-RIL
harboring pET-curA in the following experiments.
Purification of the Curcumin-Converting Enzyme. Through the pu-
rification steps described in Materials and Methods, the curcumin-
converting enzyme was purified 1,100-fold, with a yield of 1.44%.
The purified enzyme showed specific activity of 1.76 units/mg
(Table 1). The purity of the enzyme was confirmed by migration
of the protein as a single band corresponding to a molecular
mass of 38 kDa on SDS/PAGE (Fig. 1).
Cloning of the Gene Encoding the Curcumin-Converting Enzyme. To
identify the gene encoding the curcumin-converting enzyme, the
38-kDa band corresponding to the purified enzyme was excised
from the SDS/PAGE gel and subjected to trypsin digestion, as
described in SI Materials and Methods. The resulting peptides
were analyzed by MALDI-TOF/MS. The peptide mass data were
used to search the Mascot proteomics database (www.matrixs-
cience.com), which revealed a significant match to YncB, which
is designated in the database for E. coli strain K-12, substrain
DH10B as a “predicted Zn-dependent and NAD(P)-binding ox-
idoreductase.” The ORF is 1,038 nucleotides long and should
encode a curcumin-converting enzyme of 345 amino acids. The
molecular mass of the protein encoded by curA was calculated
to be 37,608 Da, which was consistent with that of the purified
enzyme determined on SDS/PAGE (Fig. 1). A search with the
Molecular mass, UV-Absorption of the Curcumin-Converting Enzyme.
The molecular weight of the native enzyme was 82 kDa, accord-
ing to the results of gel-filtration chromatography. Considering the
molecular mass of 38 kDa determined on SDS/PAGE, the enzyme
consists of two identical subunits.
The absorption spectrum of the colorless purified enzyme
(16.3 mg/mL) in 10 mM potassium phosphate buffer (KPB) (pH
6.0) showed an absorbance maximum near 280 nm. No other
absorption peak or shoulder was observed at higher wavelengths,
suggesting that no cofactor was bound to the enzyme. Moreover,
no metal was contained in the enzyme (SI Results).
Effects of Temperature and pH on the Activity and Stability of the
Enzyme. The effects of temperature and pH on the enzyme activity
BLAST program revealed that the deduced amino acid sequence were examined. The optimal reaction temperature appeared to
of CurA (NADPH-dependent curcumin/dihydrocurcumin re- be 35 °C (Fig. S4A). The enzyme exhibited maximum activity at
ductase) exhibits low similarity with those of alkenal double-bond pH 5.9 (Fig. S4B).
reductase (AtDBR) from Arabidopsis thaliana (39%) (14) and
The stability of the enzyme was investigated at various tem-
leukotriene B4 12-hydroxydehydrogenase/15-oxo-prostaglandin13- peratures. After the enzyme had been preincubated for 30 min in
reductase (12-HD/PGR) from guinea pig (37%) (15) (Fig. S2).
10 mM KPB (pH 6.0), an aliquot of the enzyme solution was
taken and then the enzyme activity was assayed under the stan-
dard assay A condition. The enzyme was most stable from 10 to
50 °C (Fig. S4C). The stability of the enzyme was examined at
various pH values. After the enzyme had been incubated at 28 °C
for 30 min in the Britton-Robinson buffer (pH 2.2–12.0) (16) at
a concentration of 0.1 M, an aliquot of the enzyme solution was
taken and then the enzyme activity was assayed under the stan-
dard assay A condition. CurA was most stable in the pH range of
4.5 to 12.0, with 90% of the initial activity being retained even at
pH 12.0 (Fig. S4D).
Overexpression and Purification of the Curcumin-Converting Enzyme
from E. coli BL21-CodonPlus(DE3)-RIL Harboring pET-curA. An ex-
pression plasmid (pET-curA) was constructed, as described in SI
Materials and Methods and introduced into E. coli BL21-CodonPlus
(DE3)-RIL for overexpression of the curcumin-converting enzyme.
Cell-free extracts prepared from the E. coli transformant car-
rying the yncB gene designated in the database acted on curcumin
as a substrate. This finding demonstrated that this yncB gene
encodes the curcumin-converting enzyme. Thus, we renamed
yncB curA. To increase the amount of the soluble enzyme, various
culture conditions were examined (Table S2). Under the opti-
mum conditions given in SI Materials and Methods, the maximum
level of the curcumin-converting enzyme activity in the cell-free
extracts was 4.68 units/mg. We analyzed the cell-free extracts by
SDS/PAGE and detected a large amount of the 38-kDa protein
Identification of the Reaction Products. The products derived on
curcumin conversion were analyzed by liquid-chromatography
electrospray-ionisation mass spectrometry (LC-ESI-MS) and fast
atom bombardment (FAB)-MS, as described in SI Materials and
Methods. First, the reaction was carried out for 30 min at 28 °C
Table 1. Purification of the curcumin-converting enzyme
Step
Total protein (mg) Total activity (units) Specific activity (units/mg) Yield (%)
Cell-free extract
DEAE Sepharose
Blue Sepharose
HiTrap Benzamidine
PIKSI Red 3 A6XL
3,050
292
4.88
4.99
0.299
0.229
0.070
0.0016
0.0171
0.0340
0.440
1.76
100
102
6.13
4.69
1.44
8.80
0.52
0.04
6616
|
www.pnas.org/cgi/doi/10.1073/pnas.1016217108
Hassaninasab et al.

DHC reported previously (17) (Fig. 2D). The FAB mass spec-
trum also gave ion at m/z 371.1511, which is in a very good
agreement with the molecular formula of DHC, [C21H23O6]+.
These findings suggested that the new intermediate product was
DHC. Moreover, no other compounds, including hexahydro-
curcumin and octahydrocurcumin, were detected on HPLC,
despite an increased reaction time and the use of a large amount
of the enzyme in the reaction mixture. Thus, the reaction
products on curcumin metabolism were identified as DHC and
THC. These findings corroborated that the curcumin-converting
enzyme converted curcumin to DHC as an intermediate product,
followed by further reduction to THC as an end product.
Fig. 1. SDS/PAGE of the purified curcumin-converting enzyme. Protein
bands were detected by staining with Coomassie brilliant blue. Lane M,
marker proteins; lane 1, the purified enzyme (1 μg).
using an excess of the purified enzyme (5 μg/mL) in standard
mixture A. In this case, only one peak with the consumption of
curcumin was detected at 280 nm with the HPLC method de-
scribed for product purification in SI Materials and Methods. The
retention time of the reaction product was found to agree with
that of authentic tetrahydrocurcumin (THC) (14.04 min). Fur-
thermore, the eluent corresponding to this peak was collected
and analyzed by LC-ESI-MS, the product exhibiting a m/z value
of 395, corresponding with the [M+Na]+ ion of THC (372 MW)
(Fig. 2A). The UV-vis absorption spectrum of the reaction
product agreed with that of authentic THC, and was the same as
that of THC reported previously (17) (Fig. 2B). The FAB mass
spectrum also gave ion at m/z 373.1656, which is in very good
agreement with the molecular formula of THC, [C21H25O6]+.
These findings demonstrated that the reaction product was THC.
Next, the reaction was carried out for 10 min at 28 °C using a low
concentration of the purified enzyme (0.5 μg/mL) in standard
mixture A. Surprisingly, another peak with a retention time of
14.43 min was detected with a HPLC method. The new peak was
not detected when the reaction was carried out completely,
suggesting that the new peak is that of an intermediate product.
The eluent corresponding to the new peak was collected and
analyzed by LC-ESI-MS. The new product exhibited a m/z value
of 393, consistent with the [M+Na]+ ion of dihydrocurcumin
(DHC) (370 MW) (Fig. 2C). Moreover, the new product
exhibited strong UV-absorption at 376 nm, and the UV-vis ab-
sorption spectrum of the new product was the same as that of
Substrate Specificity. The ability of the enzyme to catalyze the
reduction of various compounds was examined using standard
assay B. Among the tested substances, described in SI Materials
and Methods, curcumin was the most suitable substrate for the
enzyme. The curcumin-converting enzyme acted on only a few
substances [i.e., 3-octene-2-one (59.7% of the activity toward
curcumin), 3-hepten-2-one (28%), resveratrol (14.8%), and
trans-2-octenal (8.13%)]. The reduction of curcumin followed
Michaelis-Menten-type kinetics, the Km and Vmax values being
0.029 mM and 9.35 units/mg, respectively.
Discussion
Despite the impressive applications (e.g., as foods, dyes, cos-
metics, and medicines) of natural compounds and their metabo-
lites, biochemical and genetic information on their biosynthesis,
biotransformation, and biodegradation is limited. Although among
natural products curcumin is well known as one of the very useful
substances having an important variety of functions, particularly
from the medicine perspective, there have not been many efforts
regarding its biodegradation and metabolism. So far, two phases
of curcumin metabolism in the livers and intestines of rats and
humans, and plasma of mice, have been reported (17–20). Phase I
metabolism comprises the reduction of the four double bonds of
Fig. 2. The reaction products, dihydrocurcumin (DHC), and tetrahydrocurcumin (THC). (A) LC-ESI-MS spectrum of THC in the positive ion mode. (B) UV-vis
spectrum of THC. (C) LC-ESI-MS spectrum of DHC in the positive ion mode. (D) UV-vis spectrum of DHC.
Hassaninasab et al.
PNAS
|
April 19, 2011
|
vol. 108
|
no. 16
|
6617

the heptadiene-3, 5-dione structure: namely, curcumin → DHC →
THC → hexahydrocurcumin → octahydrocurcumin (7, 17–20).
During phase II, curcumin and its reduced metabolites are con-
jugated with a monoglucuronide, a monosulfate and a mixed
sulfate/glucuronide: namely, conjugated curcumin → conjugated
DHC → conjugated THC → conjugated hexahydrocurcumin →
conjugated octahydrocurcumin (7, 17–20). To the best of our
knowledge, however, no studies on purification and character-
ization of these enzymes or on cloning of the genes encoding these
enzymes have been reported. Thus, the detailed mechanism of
curcumin metabolism remains unclear. Moreover, no study on the
effect of curcumin on the human gut flora has been reported,
although curcumin has been taken orally as a food agent or
medicine since ancient times. Therefore, we addressed the
question of whether intestinal microorganisms that can metabo-
lize curcumin exist or not. Our success in isolating the micro-
organisms exhibiting curcumin-converting activity from human
feces is unique. Among the microorganisms isolated, the selected
one exhibiting the highest curcumin-converting ability was iden-
tified as E. coli. In addition, we found that E. coli strain K-12,
substrain DH10B, whose whole-genome sequence had already
been determined, possessed higher curcumin-converting activity
than the isolated E. coli. Therefore, E. coli DH10B was used for
the further experiments.
Fig. 4. Curcumin metabolic pathway.
We purified the enzyme CurA, which is responsible for cur-
cumin transformation. During the enzyme purification, the cur-
cumin-converting enzyme activity was lost on dialysis. However,
we found that the addition of NADPH increased the activity, but
the addition of NADH did not. These findings suggest that the
enzyme catalyzes NADPH-dependent conversion of curcumin.
Based on mass determination by means of LC-ESI-MS, FAB-MS,
and UV-vis absorption spectra (17), we also identified two re-
action products (DHC and THC), depending on the consumption
of curcumin. Moreover, the time course of curcumin conversion
revealed the order of product formation (i.e., curcumin is first
reduced to DHC, followed by the latter’s sequential reduction to
THC) (Fig. 3). Thus, in the two reaction steps catalyzed by CurA,
DHC is first a product and then a substrate. In this investigation,
we discovered the two-step curcumin metabolic pathway (curcu-
min → DHC → THC) catalyzed by CurA (Fig. 4). The products
were generated from curcumin through reductive destruction of
the chromophoric diarylheptatrienone chain. The lack of pro-
duction of further reduced reaction products than THC by CurA
suggested that CurA catalyzes the reduction of only compounds
with C = C double bonds (i.e., not C = O double bonds). Al-
though a large number of different compounds were tested as
potential substrates for CurA, CurA appears to have a narrow sub-
strate spectrum and to preferentially act on curcumin. Considering
the substrate specificity, reaction, and NADPH-dependency, we
named the purified enzyme (CurA) NADPH-dependent curcumin/
dihydrocurcumin reductase.
CurA exhibited sequence similarity with some enzymes be-
longing to the medium-chain dehydrogenase/reductase (MDR)
superfamily, which contains many different families including the
alcohol dehydrogenases family and leukotriene B4 dehydrogenase
family (21). All enzymes belonging to the MDR superfamily use
NAD(H) or NADP(H) as a cofactor (22), and catalyze oxidation
or reduction reactions biologically (23). These characteristics
support the requirement of NADPH as a cofactor for the re-
duction of curcumin/dihydrocurcumin by CurA. On the other
hand, several (i.e., not all) of the MDR superfamily members have
one zinc ion with a catalytic function in the active site (22), which
is necessary for their activities. Based on the following findings,
CurA belongs to Zn-independent families of the MDR super-
family: (i) CurA showed no absorption peak or shoulder except
near 280 nm; (ii) CurA contained no metal (SI Results); (iii)
chelating reagents did not inhibit CurA at all (SI Results); and
(iv) the addition of different metals including zinc to the reaction
mixture did not increase the enzyme activity significantly (SI
Results), suggesting that CurA is metal-independent. It is expec-
ted that CurA contains an NADPH-binding site in its structure.
Based on comparison of CurA with 3D structural information
of AtDBR (14) and 12-HD/PGR (15), we here propose Asp-51,
Val-162, Gly-182, Lys-186, Asn-225, Cys-247, Phe-282, and Asn-
333 comprise the NADPH-binding site of CurA. Interestingly,
CurA was found to be inhibited by thiol-specific reagents (e.g.,
p-chloromercuribenzoate and 5,5′-dithio-bis-2-nitrobenzoate) as
a result of inhibition studies (Table S4). To confirm the effects
of such reagents on Cys-247, which is located in the proposed
NADPH-binding residues of CurA, we constructed a point mu-
tant (C247A) by substituting alanine for Cys-247 in CurA, and
purified the C247A mutant enzyme to homogeneity, as described
in SI Materials and Methods (Fig. S3). The C247A mutant enzyme
exhibited a reduction of about 50% in activity (2.73 units/mg), but
was not inhibited by 5,5′-dithio-bis-2-nitrobenzoate at all (2.64
units/mg), whereas wild-type enzyme was completely inhibited
(SI Results). These results indicated that Cys-247 is sole target
of 5,5′-dithio-bis-2-nitrobenzoate. Considering that Cys-247 was
proposed as one of the NADPH-binding sites of CurA based on
comparison of the corresponding residues constituting the
NADPH-binding sites of AtDBR (14) and 12-HD/PGR (15),
Fig. 3. Time course of curcumin conversion and generation of products.
Curcumin (▲), DHC (●), and THC (■).
6618
|
www.pnas.org/cgi/doi/10.1073/pnas.1016217108
Hassaninasab et al.

CurA would lose the activity because of the binding of 5,5′-dithio-
bis-2-nitrobenzoate to Cys-247 followed by the inhibition of
NADPH-binding to Cys-247. Like other typical MDR superfamily
members, CurA has a glycine-rich motif, 157AATGPVG163 (Fig.
S2), shared by several oxidoreductases (AXXGXXG), which is
known to participate in binding of the pyrophosphate group of
NAD(P)+ or NADPH (24). So far, known MDR superfamily
members play important roles in biological oxidation/reduction
and exist in all organisms (25). However, the majority of the
enzymes indicated to belong to the MDR group are depicted as
putative gene products identified by genome sequencing and are
poorly characterized (26). The present discovery of CurA in
E. coli and its functional identification will allow further eluci-
dation of the evolution and distribution of unknown enzymes in
the MDR superfamily, which will pave the way for their use for
the biotransformation of natural compounds and new chemical
inputs into the environment.
(CD) spectra were measured (SI Materials and Methods). The
CD spectra of the enzyme in a broad pH value range (3.5–12.0)
were similar to that under usual pH conditions (pH 6.0) (Fig.
5A), but those at pHs 2.1 to 3.0 were different (Fig. 5B). These
findings suggested that the ellipticity at 222 nm (secondary
structure) of CurA did not change at pHs 3.5 to 12.0 (Fig. 5C).
On CD measurement after the pH had been shifted from various
pH values (pH 2.1–12.0) to 6.0, the CD spectra of CurA (Fig.
5D) were similar to one another for the shift from pH 3.5 to 12.0,
but not from pH 2.1 to 3.0 (Fig. 5E), and the ellipticities at 222
nm for pH 3.5 to 12.0 were also similar to one another (Fig. 5F).
The data for the pH stability of the CurA activity and the pH
stability of the CurA secondary structure demonstrated that
CurA is structurally stable from pH 4.5 to 12.0 (Table S5).
In humans and animals, four reduction products (i.e., THC
and hexahydrocurcumin as major products, and DHC and
octahydrocurcumin as minor ones) (7, 17–20) were detected,
although the nature of the enzymes involved in the formation of
such products is unknown. In microorgamisms, we first found
that both DHC and THC are the products, using the purified
enzyme (Fig. 2). It was previously found that THC is more potent
than curcumin as to anti-inflammatory (28), antidiabetic, and
antihyperlipidemic (29) activity, and equally potent as to anti-
oxidant activity (30, 31). THC exhibits its biological activities at
a very low concentration both in vivo and in vitro. For example,
50 μM THC can inhibit the metastasis of HT1080 cancer cells in
vitro (32). THC also exhibits antioxidant activity at a concentra-
tion of 18.7 μM in vitro, using the radical scavenging activity of
2,2-diphenyl-1-picrylhydrazyl (33). In the mouse brain, THC at
a concentration of 1.286 μM has an anti-inflammatory effect and
can prevent Alzheimer’s disease by inhibiting IL-1β (34). These
findings would uncover the mystery of a reduced age-adjusted
prevalence of Alzheimer’s disease (35) and Parkinson disease
(36) in India, where the people consume curcumin in their diet.
It is speculated that curcumin in the human intestine is converted
into THC by E. coli CurA and that the produced THC is re-
sponsible for many various activities in the human body. Further
studies are required to understand the participation of the
common human commensal E. coli in curcumin conversion. Our
research here shed new light on the role of human intestinal
microorganisms in the mechanism of curcumin metabolism in
vivo. CurA is a thus-far unknown enzyme found to produce THC
from curcumin, and therefore it would be a potent catalyst for
the production of a large amount of the value-added product
in vitro.
During the characterization of CurA, we surprisingly found
that it shows a wide pH stability range (pH 4.5–12.0), particularly
including alkaline pH values. To obtain information on the
secondary structure (27) of CurA, far-UV circular dichroism
Materials and Methods
Isolation of Curcumin-Converting Microorganisms from Human Feces. We iso-
lated microorganisms with curcumin-degrading ability from human feces.
Fresh human fecal samples were inoculated into test tubes containing 10 mL
of a synthetic medium comprising 0.05% (wt/vol) KH2PO4, 0.05% K2HPO4,
0.1% NaCl, 0.05% MgSO4·7H2O, 0.2% yeast extract, 0.05% curcumin and 5%
Tween20, and then incubated at 37 °C for 24 h. Tween20 was added to
enhance curcumin solubility in the culture medium. Drops of the incubated
liquid cultures were spread on agar plates (of the above-mentioned me-
dium), and the microorganisms that grew were isolated. Each of the isolated
strains was then inoculated into a test tube including 10 mL of the same
medium. The cells were harvested by centrifugation, washed twice with 0.1
M KPB (pH 6.0), and then suspended in the same buffer. The suspended cells
were disrupted by sonication at 10 W for 10 min with an ultrasonic disrupter
(Microson Ultrasonic Cell Disrupter XL; Misonix, Inc.). The cell debris was
removed by centrifugation at 0 to 4 °C. The curcumin-converting abilities of
the isolated strains were measured by means of the cell-free extract re-
action. The reaction was carried out as described in the SI Materials and
Methods subsection, Enzyme assays, using standard assay A in the absence
of NADPH.
Fig. 5. CD spectra of the curcumin-converting enzyme and the ellipticity in
different pH values. The CD spectra of CurA treated in 10-mM Britton-
Robinson buffer pH 6.0 (A) and pH 2.1 (B); 0.1 mg/mL of the purified enzyme
was treated at different pH values in 10-mM Britton-Robinson buffer and CD
spectra were measured. The mollar ellipticity (θ) was then calculated for the
different pHs. θ222 was denoted by dotted line. (C) θ222 at different pH val-
ues. The CD spectra of CurA after the pH was shifted from 6.0 to 6.0 (D) and
from 2.1 to 6.0 (E); 0.1 mg/mL of the purified enzyme was treated at dif-
ferent pH values in 10-mM Britton-Robinson buffer and then shifted to the
usual pH 6.0 using 40-mM Britton-Robinson buffer (pH 6.0), and then the CD
spectrum was measured and θ222 was denoted by a dotted line. (F), θ222 after
the pH had been shifted from various values to 6.0.
Purification of the Curcumin-Converting Enzyme from E. coli DH10B. All puri-
fication steps were carried out at 0 to 4 °C. KPB (pH 6.0) was used throughout
the purification. Centrifugation was performed for 20 min at 16,200 × g.
Hassaninasab et al.
PNAS
|
April 19, 2011
|
vol. 108
|
no. 16
|
6619

Step 1: Preparation of cell-free extracts. Washed cells from 24 L of broth culture
were suspended in 500 mL of 0.1 M KPB (pH 6.0) and then disrupted by
sonication at 200 W for 30 min with an Insonator model 201 M (Kubota). The
cell debris was removed by centrifugation.
centrifugation and then dissolved in 0.1 M KPB (pH 6.0). The resultant so-
lution was dialyzed against 10 mM KPB (pH 6.0).
Step 5: Mimetic Red 3 A6XL column chromatography. The enzyme solution from
step 4 was applied to a PIKSI (R) M Test Kit 0013 Mimetic Red 3 A6XL column
(1 × 1 cm) (Affinity Chromatography Ltd.) equilibrated with 10 mM KPB
(pH 6.0). Protein was eluted from the column by increasing KCl linearly from
0 to 2 M in the same buffer. The homogeneity of the purified protein was
confirmed by SDS/PAGE.
Step 2: DEAE-Sepharose Fast Flow column chromatography. Ammonium sulfate
was added to the resulting supernatant solution to give 80% saturation. After
centrifugation of the suspension, the precipitate was dissolved in 0.1 M KPB
(pH 6.0) and then dialyzed against 10 mM KPB (pH 6.0). The resultant dialyzed
solution was applied to a DEAE-Sepharose Fast Flow column (5 × 40 cm)
equilibrated with 10 mM KPB (pH 6.0). Protein was eluted from the column
with 0.5 L of 10 mM KPB (pH 6.0). The enzyme was collected as the flow-
through fraction, followed by the addition of ammonium sulfate to give
80% saturation. The precipitate was collected by centrifugation, and dis-
solved in 0.1 M KPB (pH 6.0), followed by dialysis against 10 mM KPB
(pH 6.0).
Step 3: Blue Sepharose 6 Fast Flow column chromatography. The dialyzed solution
was applied to a Blue Sepharose 6 Fast Flow column (3 × 15 cm) equilibrated
with 10 mM KPB (pH 6.0). Protein was eluted by increasing KCl linearly from
0 to 2 M in the same buffer. The active fractions were collected, and then
ammonium sulfate was added to give 80% saturation. After centrifugation
of the suspension, the precipitate was dissolved in 0.1 M KPB (pH 6.0), fol-
lowed by dialysis against 10 mM KPB (pH 6.0).
Step 4: Hitrap Benzamidine FF (high sub) column chromatography. Because the
structure of nicotinamide and/or adenine in the parts of NADPH resembled
that of the benzamidine, the enzyme solution from step 3 was applied to
a Hitrap Benzamidine FF (high sub) column (1.6 × 2.5 cm) equilibrated with 10
mM KPB (pH 6.0). Protein was eluted from the column by increasing KCl
linearly from 0 to 1 M in the same buffer. The active fractions were com-
bined and then precipitated with ammonium sulfate at 80% saturation.
After centrifugation of the suspension, the precipitate was collected by
Nucleotide Sequence Accession Number. The nucleotide sequence data
reported in this paper appear in the DNA Data Base in Japan/GenBank
database under accession numbers AB609595 and AB583756.
Description of SI Materials and Methods. The materials used in this study,
identification of a curcumin-converting microorganism, culture conditions
for E. coli DH10B, and the two enzyme assay systems used to measure CurA
activity are described in SI Materials and Methods. Additionally, electro-
phoresis, peptide mass fingerprinting, cloning and overexpression of the
gene encoding CurA, purification of the curcumin-converting enzyme from
E. coli BL21-CodonPlus(DE3)-RIL harboring pET-curA, molecular mass de-
termination, metal analysis, substrate specificity, CD analysis, identification
of the reaction products, time course of curcumin conversion and product
formation, and site-directed mutagenesis were performed as described in SI
Materials and Methods.
ACKNOWLEDGMENTS. We thank Dr. Kentaro Shiraki for help with the
circular dichroism spectra analysis and for the valuable advice; Dr. Yoichi
Takase for fast atom bombardment mass spectrometry analysis; and Mrs.
Mitsue Arimoto for the help with MALDI-TOF/MS analysis.
1. Aggarwal BB, Sundaram C, Malani N, Ichikawa H (2007) Curcumin: The Indian solid
gold. Adv Exp Med Biol 595:1–75.
ability to inhibit phorbol ester-induced prostaglandin E2 production. Cancer Res 61:
1058–1064.
2. Sharma RA, Gescher AJ, Steward WP (2005) Curcumin: The story so far. Eur J Cancer
41:1955–1968.
20. Ireson CR, et al. (2002) Metabolism of the cancer chemopreventive agent curcumin in
human and rat intestine. Cancer Epidemiol Biomarkers Prev 11:105–111.
21. Persson B, Hedlund J, Jörnvall H (2008) Medium- and short-chain dehydrogenase/
reductase gene and protein families: The MDR superfamily. Cell Mol Life Sci 65:
3879–3894.
22. Nordling E, Jörnvall H, Persson B (2002) Medium-chain dehydrogenases/reductases
(MDR). Family characterizations including genome comparisons and active site
modeling. Eur J Biochem 269:4267–4276.
3. Hatcher H, Planalp R, Cho J, Torti FM, Torti SV (2008) Curcumin: From ancient
medicine to current clinical trials. Cell Mol Life Sci 65:1631–1652.
4. Garcia-Alloza M, Borrelli LA, Rozkalne A, Hyman BT, Bacskai BJ (2007) Curcumin labels
amyloid pathology in vivo, disrupts existing plaques, and partially restores distorted
neurites in an Alzheimer mouse model. J Neurochem 102:1095–1104.
5. Nandakumar DN, Nagaraj VA, Vathsala PG, Rangarajan P, Padmanaban
G (2006)
Curcumin-artemisinin combination therapy for malaria. Antimicrob Agents Chemother
50:1859–1860.
23. Vitale A, Rosso F, Barbarisi A, Labella T, D’Auria S (2010) Properties and evolution of
an alcohol dehydrogenase from the Crenarchaeota Pyrobaculum aerophilum. Gene
461:26–31.
6. Panchatcharam M, Miriyala S, Gayathri VS, Suguna
L (2006) Curcumin improves
wound healing by modulating collagen and decreasing reactive oxygen species. Mol
Cell Biochem 290:87–96.
24. Pereira PJB, et al. (2001) Structure of human biliverdin IXbeta reductase, an early fetal
bilirubin IXbeta producing enzyme. Nat Struct Biol 8:215–220.
7. Ravindranath V, Chandrasekhara N (1981) In vitro studies on the intestinal absorption
of curcumin in rats. Toxicology 20:251–257.
8. Crossman LC, et al. (2010) A commensal gone bad: Complete genome sequence of
25. Jörnvall H, Höög JO, Persson B (1999) SDR and MDR: Completed genome sequences
show these protein families to be large, of old origin, and of complex nature. FEBS
Lett 445:261–264.
the prototypical enterotoxigenic Escherichia coli strain H10407.
5822–5831.
9. Zhou Z, et al. (2010) Derivation of Escherichia coli O157:H7 from its O55:H7 precursor.
PLoS ONE 5:e8700.
J
Bacteriol 192:
26. Dick RA, Kensler TW (2004) The catalytic and kinetic mechanisms of NADPH-
dependent alkenal/one oxidoreductase. J Biol Chem 279:17269–17277.
27. Strickland EH (1968) Circular dichroism of horseradish peroxidase and its enzyme-
substrate compounds. Biochim Biophys Acta 151:70–75.
10. Ferenci T, et al. (2009) Genomic sequencing reveals regulatory mutations and
recombinational events in the widely used MC4100 lineage of Escherichia coli K-12.
J Bacteriol 191:4025–4029.
11. Macconkey A (1905) Lactose-fermenting bacteria in feces. J Hyg (Lond) 5:333–379.
12. Levin M (1918) Differentiation of E. coli and A. aerogenes on a simplified eosin-
methylene blue agar. J Infect Dis 23:43–47.
28. Mukhopadhyay A, Basu N, Ghatak N, Gujral PK (1982) Anti-inflammatory and irritant
activities of curcumin analogues in rats. Agents Actions 12:508–515.
29. Pari L, Murugan P (2007) Antihyperlipidemic effect of curcumin and tetrahydrocurcumin
in experimental type 2 diabetic rats. Ren Fail 29:881–889.
30. Sugiyama Y, Kawakishi S, Osawa T (1996) Involvement of the β-diketone moiety in the
antioxidative mechanism of tetrahydrocurcumin. Biochem Pharmacol 52:519–525.
13. Durfee T, et al. (2008) The complete genome sequence of Escherichia coli DH10B:
Insights into the biology of a laboratory workhorse. J Bacteriol 190:2597–2606.
14. Youn B, et al. (2006) Mechanistic and structural studies of apoform, binary, and
ternary complexes of the Arabidopsis alkenal double bond reductase At5g16970.
J Biol Chem 281:40076–40088.
31. Osawa T, Sugiyama Y, Inayoshi M, Kawakishi S (1995) Antioxidative activity of
tetrahydrocurcuminoids. Biosci Biotechnol Biochem 59:1609–1612.
32. Yodkeeree S, Garbisa S, Limtrakul P (2008) Tetrahydrocurcumin inhibits HT1080 cell
migration and invasion via downregulation of MMPs and uPA. Acta Pharmacol Sin 29:
853–860.
15. Hori T, et al. (2004) Structural basis of leukotriene B4 12-hydroxydehydrogenase/15-
Oxo-prostaglandin 13-reductase catalytic mechanism and a possible Src homology 3
domain binding loop. J Biol Chem 279:22615–22623.
33. Somparn P, Phisalaphong C, Nakornchai S, Unchern S, Morales NP (2007) Comparative
antioxidant activities of curcumin and its demethoxy and hydrogenated derivatives.
Biol Pharm Bull 30:74–78.
16. Britton HTS, Robinson RA (1931) Universal buffer solutions and the dissociation
constant of veronal. J Chem Soc 458:1456–1462.
17. Hoehle SI, Pfeiffer E, Sólyom AM, Metzler M (2006) Metabolism of curcuminoids in
tissue slices and subcellular fractions from rat liver. J Agric Food Chem 54:756–764.
18. Pan MH, Huang TM, Lin JK (1999) Biotransformation of curcumin through reduction
and glucuronidation in mice. Drug Metab Dispos 27:486–494.
34. Begum AN, et al. (2008) Curcumin structure-function, bioavailability, and efficacy in
models of neuroinflammation and Alzheimer’s disease. J Pharmacol Exp Ther 326:
196–208.
35. Ganguli M, et al. (2000) Apolipoprotein E polymorphism and Alzheimer disease. Arch
Neurol 57:824–830.
36. Muthane U, Yasha TC, Shankar SK (1998) Low numbers and no loss of melanized
nigral neurons with increasing age in normal human brains from India. Ann Neurol
43:283–287.
19. Ireson C, et al. (2001) Characterization of metabolites of the chemopreventive agent
curcumin in human and rat hepatocytes and in the rat in vivo, and evaluation of their
6620
|
www.pnas.org/cgi/doi/10.1073/pnas.1016217108
Hassaninasab et al.