H. Lu et al. / Tetrahedron 68 (2012) 7651e7654
7653
However residues involved in ferredoxin binding are not strictly
conserved. As CmlI is also able to turn over the much smaller
substrate 5, the presence of a keto or a hydroxyl group at the
benzylic position may be sufficient for recognition as a substrate for
at least the initial oxidation. This is similar to what has been found
for AurF, but as CmlI can accept much larger substrates, it may be
more amenable for engineering as an N-oxidation catalyst. It re-
mains to be seen what other factors determine the substrate
specificity of CmlI.
was used as a selection marker. For expression of cmlI in Lur-
ꢁ
iaeBertani medium, cells were grown at 37 C to OD600¼0.6e0.8.
Then protein expression was induced by 0.1 mM isopropyl b-D-1-
thiogalactopyranoside (IPTG), and the cells were agitated at
ꢁ
250 rpm and 25 C overnight. For expression of CmlI in M9 me-
ꢁ
dium, cells were grown in M9 medium at 37 C to OD600¼0.6e0.8.
8
To the culture were added
L
-methionine (60 mg/L),
L
-leucine
-isoleucine (50 mg/L), -lysine
-threonine (100 mg/L), -phenylalanine (100 mg/L),
4
(0.5 mM), and MnSO (0.5 mM). Protein expression
(50 mg/L),
(100 mg/L),
L
L-valine (50 mg/L), L
L
L
Various metals have been found in oxygenases to mediate
(NH
4
)
2
Fe(SO
4
)
2
substrate oxidation.2
5e28
Although it is clear now that AurF, a ho-
was induced by IPTG (0.1 mM), and cells were agitated at 250 rpm
ꢁ
molog of CmlI, is a non-heme di-iron N-oxygenase, we cannot rule
out the existence of other metals in CmlI. To address this issue, we
analyzed by ICP-MS the metal contents in CmlI purified from cells
cultivated in LuriaeBertani medium. Only Fe was found to be sig-
nificantly present. To rule out the possible involvement of Mn in
and 18 C overnight. The cell lysate was prepared by passing cells
resuspended in buffer (20 mM Tris$HCl, pH¼7.5, 100 mM NaCl,
1 mM
Press . N-His
b-mercaptoethanol, 5% glycerol) twice through French
Ò
6
-tagged CmlI was purified off Co-column to over 95%
purity. Purified CmlI was then dialyzed against storage buffer
(50 mM HEPES, 10% glycerol). Protein was quantified using the
Bradford Quickstart Reagent (Bio-Rad, Hercules, CA) using bovine
serum albumin as a standard.
10
catalysis, which was once suggested for AurF, the ratio of Mn and
Fe in CmlI were further assessed for CmlI purified cells cultivated in
either LuriaeBertani medium or M9 minimal medium with equal
amounts of Fe and Mn added. The concentration of Fe in the protein
was much higher than that of Mn, by about 17-fold in both cases.
The Fe/CmlI ratio ranged from 1.2 to 1.4 (Table 1). When we used
4.3. In vitro and in vivo assays
2
9
the iron-specific ferrozine assay, the Fe to CmlI ratio was found to
be 2:1 (Table 1). Taken together, our data seem to suggest that CmlI,
just like AurF, is also a di-iron oxygenase.
In vitro assays of CmlI were performed on a 600
final reaction mixture contained CmlI (5e20 M) substrate
NH eCam (250 M), PMS (100 M), NADH (1 mM), and HEPES
20 mM, pH¼7.5). Reactions were performed at room temperature
mL scale. The
m
2
m
m
(
(
ꢁ
23e25 C) and were started by addition of NADH to the reaction
Table 1
Metal analysis of CmlI
mixture. Reaction was quenched by trifluoroacetic acid (final con-
centration 0.8% v/v) and centrifuged. For in vivo reconstitution of
CmlI, cells were grown similar to cells used for purification, except
that Terrific Broth medium was used instead of LuriaeBertani
broth. Cells were pelleted by centrifugation, washed twice in assay
buffer (25 mM MOPS, 8 g/L NaCl, 1 g/L KCl, 2 g/L glucose, pH 7.2),
and resuspended to a final OD600 of 15 in a final volume of 50 mL.
Fe/Mn
ratio (ICP)
Fe/CmlI
ratio (ICP)
Fe/CmlI ratio
(ferrozine)
CmlI from M9
CmlI from LB
17
17
1.4
1.2
2.2
1.8
3
. Conclusions
Substrate was added to a final concentration of 250
m
M, and the
ꢁ
reaction was followed for 3 h at 30 C, taking time points every
30 min. As a control, cells containing only the vector were used, and
subjected to the same analysis and time course.
We have reported the first biochemical characterization of the
N-oxygenase CmlI and confirmed its role in chloramphenicol bio-
synthesis. Reconstitution of the enzyme activity was achieved
in vitro via addition of chemical reductants and in vivo in E. coli.
Furthermore, our data show that CmlI, just like AurF, is a non-heme
di-iron oxygenase. These findings have opened the door for further
mechanistic and structural studies of this type of oxygenases and
biocatalyst development using protein engineering tools.
4.4. LCeESI-MS and ESI-MS/MS analysis of chloramphenicol
Samples were analyzed via reverse-phase liquid chromatogra-
phy on an Agilent 1100 HPLC. The conditions were: Zorbax SB-C18
column, 3.0ꢂ150 mm (Agilent); mobile phase A: 25 mM ammo-
nium acetate (pH¼6.8), mobile phase B: methanol; flow rate
0.5 mL/min, HPLC program as shown in Table S1. Compound 1 was
quantified at 237 nm, while 4 was quantified at 285 nm by com-
parison with standards with known concentrations. An alternative
HPLC method was used for in vivo assays, where buffer A (1% acetic
acid) and buffer B (methanol) both contained 2 mM sodium 1-
heptanesulfonic acid to improve peak shape and separation for the
alternate substrates used. For mass spectroscopy work, an Agilent
1100 series LC/MSD XCT plus-ion-trap mass spectrometer was
utilized. Identical mobile phases and columns were used as pre-
4
4
. Materials and methods
.1. Bacteria strains and culture conditions
The wild-type strain of S. venezuelae ATCC 10712 was obtained
from the American Type Culture Collection (ATCC, Manassas, VA). S.
venezuelae and its derivatives were maintained on ISP2 agar me-
dium and cultivated in YGM liquid medium (malt extract 5 g/L,
yeast extract 2 g/L, glucose 2 g/L) at 28 C with shaking at 250 rpm.
Genomic DNA was isolated using the Wizard Genomic DNA Pu-
ꢁ
Ò
viously indicated, with a flow rate of 300 mL/min and the HPLC
rification Kit (Promega, Madison, WI) using standard protocols.
program shown below. The column effluent was directed to Agilent
XCT ion-trap MSD mass spectrometer, which was operated in the
negative ion mode to detect parent ions of each of the reaction
intermediates. The system was operated using a drying tempera-
4
.2. Overexpression and purification of CmlI
ꢁ
The cmlI gene was amplified by PCR from the genomic DNA of S.
ture of 350 C, a nebulizer pressure of 35 psi, a drying gas flow of
venezuelae (NRRL ISP-5230) and cloned into pET26 vectors (Nova-
8.5 L/min, and a capillary voltage of 4500 V.
0
gen, Gibbstown, NJ) using
TATGCGTGACCACACGGACGAGAAATC-3 ) and a reverse primer (5 -
a
forward primer (5 -CAGTTCA-
0
0
4.5. Ferrozine assay
0
GTAGTCAAGCTTTCATCGGG TCACCGTCGTGC-3 ) between NdeI and
HindIII sites (underlined sequences). The corresponding plasmid
The procedure was largely adapted from published protocol
with slight modification. Briefly, 50 mL protein (20e200 mM
2
9
was transformed into E. coli BL21 (DE3) cells. Kanamycin (50
m
g/mL)