Characterization and mechanistic study of a radical SAM dehydrogenase in the biosynthesis of butirosin

Article abstract of DOI:10.1021/ja072481t

BtrN encoded in the butirosin biosynthetic gene cluster possesses a CXXXCXXC motif conserved within the radical S-adenosyl methionine (SAM) superfamily. Its gene disruption in the butirosin producer Bacillus circulans caused the interruption of the biosyn

Full text of DOI:10.1021/ja072481t

Published on Web 11/15/2007
Characterization and Mechanistic Study of a Radical SAM
Dehydrogenase in the Biosynthesis of Butirosin
Kenichi Yokoyama,† Mario Numakura,† Fumitaka Kudo,† Daijiro Ohmori,§ and
Tadashi Eguchi*,‡
Contribution from the Department of Chemistry, and Department of Chemistry & Materials
Science, Tokyo Institute of Technology, O-okayama, Meguro-ku, Tokyo152-8551, Japan, and
Department of Chemistry, Juntendo UniVersity, Inba, Chiba 270-1695, Japan
Received April 9, 2007; E-mail: eguchi@cms.titech.ac.jp
Abstract: BtrN encoded in the butirosin biosynthetic gene cluster possesses a CXXXCXXC motif conserved
within the radical S-adenosyl methionine (SAM) superfamily. Its gene disruption in the butirosin producer
Bacillus circulans caused the interruption of the biosynthetic pathway between 2-deoxy-scyllo-inosamine
(DOIA) and 2-deoxystreptamine (DOS). Further, in vitro assay of the overexpressed enzyme revealed that
BtrN catalyzed the oxidation of DOIA under the strictly anaerobic conditions along with consumption of an
equimolar amount of SAM to produce 5′-deoxyadenosine, methionine, and 3-amino-2,3-dideoxy-scyllo-
inosose (amino-DOI). Kinetic analysis showed substrate inhibition by DOIA but not by SAM, which suggests
that the reaction is the Ordered Bi Ter mechanism and that SAM is the first substrate and DOIA is the
second. The BtrN reaction with [3-2H]DOIA generated nonlabeled, monodeuterated and dideuterated 5′-
deoxyadenosines, while no deuterium was incorporated by incubation of nonlabeled DOIA in the deuterium
oxide buffer. These results indicated that the hydrogen atom at C-3 of DOIA was directly transferred to
5′-deoxyadenosine to give the radical intermediate of DOIA. Generation of nonlabeled and dideuterated
5′-deoxyadenosines proved the reversibility of the hydrogen abstraction step. The present study suggests
that BtrN is an unusual radical SAM dehydrogenase catalyzing the oxidation of the hydroxyl group by a
radical mechanism. This is the first report of the mechanistic study on the oxidation of a hydroxyl group by
a radical SAM enzyme.
Introduction
led to the identification of several other DOS-containing
aminoglycoside gene clusters.8-12 Comparative genetics of the
2-Deoxystreptamine (DOS) containing aminoglycosides con-
stitute the largest subgroup of aminoglycoside antibiotics. This
group of aminoglycosides includes kanamycin, neomycin, and
gentamicin, which are clinically important antibiotic agents used
against bacterial and some protozoal infections.1 In addition to
these classical roles they are also attracting attention for their
RNA and DNA recognition ability, which can be applied for
anti-plasmids2,3 and anti-HIV agents.4,5 Increased attention to
the biological activity of aminoglycosides makes their biosyn-
thetic machinery more attractive for creating structurally diverse
substances by engineered biosynthesis.
open reading frames (ORFs) in those gene clusters, combined
with gene disruption experiments and in vitro assays of
overexpressed enzymes, led to reasonable assignments of each
ORF in the butirosin biosynthetic gene cluster to each biosyn-
thetic reaction.8,13 However, the functions of several ORFs still
remain uncharacterized.
BtrN, a functionally unknown enzyme encoded in the
butirosin biosynthetic gene cluster, has an amino acid sequence
motif (CXXXCXXC), which is typical of the radical S-adenosyl
methionine (SAM) superfamily.14 The radical SAM superfamily
is an emerging group of enzymes that catalyze a wide range of
The biosynthetic gene cluster of butirosin was the first of
the DOS-containing aminoglycosides to be identified,6,7 which
(6) Ota, Y.; Tamegai, H.; Kudo, F.; Kuriki, H.; Koike-Takeshita, A.; Eguchi,
T.; Kakinuma, K. J. Antibiot. 2000, 53, 1158-1167.
† Department of Chemistry, Tokyo Institute of Technology.
§ Juntendo University.
(7) Tamegai, H.; Nango, E.; Kuwahara, M.; Yamamoto, H.; Ota, Y.; Kuriki,
H.; Eguchi, T.; Kakinuma, K. J. Antibiot. 2002, 55, 707-714.
(8) Huang, F.; Haydock, S. F.; Mironenko, T.; Spiteller, D.; Li, Y.; Spencer,
J. B. Org. Biomol. Chem. 2005, 3, 1410-1418.
‡ Department of Chemistry & Materials Science, Tokyo Institute of
Technology.
(1) Magnet, S.; Blanchard, J. S. Chem. ReV. 2005, 105, 477-497.
(2) Thomas, J. R.; DeNap, J. C.; Wong, M. L.; Hergenrother, P. J. Biochemistry
2005, 44, 6800-6808.
(9) Kharel, M. K.; Subba, B.; Basnet, D. B.; Woo, J. S.; Lee, H. C.; Liou, K.;
Sohng, J. K. Arch. Biochem. Biophys. 2004, 429, 204-214.
(10) Yanai, K.; Murakami, T. J. Antibiot. 2004, 57, 351-354.
(11) Unwin, J.; Standage, S.; Alexander, D.; Hosted, T., Jr.; Horan, A. C.;
Wellington, E. M. J. Antibiot. 2004, 57, 436-445.
(12) Kharel, M. K.; Basnet, D. B.; Lee, H. C.; Liou, K.; Woo, J. S.; Kim, B.
G.; Sohng, J. K. FEMS Microbiol. Lett. 2004, 230, 185-190.
(13) Llewellyn, N. M.; Spencer, J. B. Nat. Prod. Rep. 2006, 23, 864-874.
(14) Sofia, H. J.; Chen, G.; Hetzler, B. G.; Reyes-Spindola, J. F.; Miller, N. E.
Nucleic Acids Res. 2001, 29, 1097-1106.
(3) Denap, J. C.; Thomas, J. R.; Musk, D. J.; Hergenrother, P. J. J. Am. Chem.
Soc. 2004, 126, 15402-15404.
(4) Cabrera, C.; Gutierrez, A.; Blanco, J.; Barretina, J.; Litovchick, A.; Lapidot,
A.; Evdokimov, A. G.; Clotet, B.; Este, J. A. AIDS Res. Hum. RetroViruses
2000, 16, 627-634.
(5) Krebs, A.; Ludwig, V.; Boden, O.; Gobel, M. W. ChemBioChem 2003, 4,
972-978.
9
10.1021/ja072481t CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 15147-15155
15147

A R T I C L E S
Yokoyama et al.
Scheme 1. Biosynthetic Pathway of DOS in the Butirosin Producer Bacillus circulans
Materials and Methods
Construction of the His-Tagged BtrN Expression Vector pET-
btrN. The btrN gene was amplified by PCR with primers, btrN-His
(AATTTCCATGGATAAGCTGTTTTCG) and btrN-C (CAAAGAAT-
TCAAACGTACACCCCC). PCR reaction was performed under condi-
tions as follows: 94 °C, 2 min for denature, 30 cycles of 94 °C, 15 s;
56 °C, 30 s; 68 °C, 1 min for extension of DNA using KOD-Plus-
DNA polymerase (TOYOBO, Japan). PCR product was digested by
EcoRI and NcoI and subcloned into the EcoRI-NcoI site of LITMUS
28 to obtain LITMUSbtrN-His. After confirmation of the sequence,
the EcoRI-NcoI fragment of LITMUSbtrN-His was inserted into
pET30b (Novagen) vector to obtain pETbtrN-His, which was subse-
quently introduced into E. coli BL21(DE3).
Figure 1. Feeding experiments against the btrN deletion mutant.
radical reactions.15 The common mechanism found in all radical
SAM enzymes is a reductive cleavage of SAM to generate the
5′-deoxyadenosyl radical, which is subsequently used for radical
reactions such as activation of other enzymes through generating
protein radical, sulfur insertion, C-C bond rearrangement, and
amino group rearrangement.16 Bioinformatics analysis has
identified more than 600 genes as members of the radical SAM
superfamily with a common CXXXCXXC sequence motif.14
They are widely distributed in various vital metabolic pathways
including biosynthesis of heme, thiamine, molybdopterin, and
many famous secondary metabolites such as mitomycin C,
fosfomycin, and erythromycin.14 However, their high sensitivity
to oxygen precluded the study of radical SAM enzymes, and
relatively few of them have been enzymatically character-
ized.16,17 Radical SAM enzymes in secondary metabolism have
especially been underinvestigated. DesII, an enzyme catalyzing
deoxygenation of the C-4 position of an aminosugar during the
biosynthesis of the desosamine part of erythromycin, is the only
one defined at the enzymatic level.18 Since secondary metabo-
lites are structurally diverse, radical SAM enzymes involved in
secondary metabolism appear to catalyze various unique and
chemically important radical reactions. Uncovering these reac-
tion mechanisms will expand our knowledge of radical reactions
in biological systems.
Overexpression and Purification of N-Terminally His-Tagged
btrN. E. coli BL21(DE3) carrying pETbtrN-His was grown in 800 mL
of LB medium containing 30 µg/mL of kanamycin at 37 °C until OD600
0.6-1.0. The expression was induced by addition of 0.1 mM isopropyl
â-D-thiogalactopyranoside (IPTG), and cultivation was continued at
15 °C for 18 h. The cells were harvested by centrifugation at 7000
rpm for 10 min, washed with 50 mM HEPES-NaOH (pH 8.0), and
stored at -30 °C until use. The wet cells (4 g) were suspended in buffer
A (50 mM HEPES-NaOH, pH 8.0, 300 mM NaCl, 50 mM imidazole)
and anaerobically disrupted by sonication at 0 °C for 1 min (5 times).
All further steps were carried out under anaerobic conditions. All
solutions were saturated with Ar prior to use. Cell debris was removed
by centrifugation at 15 000 rpm, for 30 min, at 4 °C. The supernatant
was loaded onto His-bind resin (Novagen) column, which had been
previously equilibrated with buffer A. The column was washed with
15 mL of buffer A to remove unbound proteins. BtrN was eluted using
buffer B (50 mM HEPES-NaOH pH 8.0, 300 mM NaCl, 200 mM
imidazole) at a flow rate of 1.3 mL/min. The yellowish protein was
collected and desalted with a 5 mL HiTrap desalting column (Amersham
Pharmacia Biotech AB) equilibrated with buffer C (50 mM HEPES-
NaOH, pH 8.0).
Reconstitution of BtrN [4Fe-4S] Cluster. The purified BtrN was
treated with 5 mM dithiothreitol (DTT) at room temperature for 15
min. FeII(NH4)2(SO4)2 (1 mM) and Na2S (1 mM) were then added, and
the mixture was incubated at room temperature for 1 h. The BtrN sample
was then desalted by passage through a 5 mL HiTrap desalting column
(Amersham Pharmacia Biotech AB) equilibrated with buffer C. The
obtained reconstituted BtrN was reduced by sodium dithionite for 1 h
at room temperature prior to use. UV-visible absorption spectra were
determined by a Shimadzu UV-2450 spectrophotometer. For EPR
measurements, the reduced BtrN was transferred to an EPR tube and
flash frozen in liquid nitrogen. EPR spectra were recorded with a JEOL
JES-FA300 ESR spectrometer (9.02 GHz). Recording conditions are
indicated in the legend of Figure 2. The concentration of unpaired spin
was determined using Cu(II)-EDTA as a standard. Protein concentra-
tions were determined with the Bradford assay (Bio-Rad) with bovine
It is thus interesting how BtrN participates in the butirosin
biosynthesis as a radical SAM protein. In this report, using a
combination of gene disruption and in vitro enzyme reaction
studies, BtrN was found to be a radical SAM dehydrogenase
catalyzing oxidation of a hydroxyl group of 2-deoxy-scyllo-
inosamine (DOIA, Scheme 1).
(15) Jarrett, J. T. Curr. Opin. Chem. Biol. 2003, 7, 174-182.
(16) Marsh, E. N.; Patwardhan, A.; Huhta, M. S. Bioorg. Chem. 2004, 32, 326-
340.
(17) Frey, P. A.; Hegeman, A. D.; Reed, G. H. Chem. ReV. 2006, 106, 3302-
3316.
(18) Szu, P. H.; He, X.; Zhao, L.; Liu, H. W. Angew. Chem., Int. Ed. Engl.
2005, 44, 6742-6746.
9
15148 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007

Study of a Radical SAM Dehydrogenase
A R T I C L E S
Figure 2. (a) SDS-PAGE of the purified BtrN. Lane 1: cell-free extract of E. coli overexpressing His-tagged BtrN. Lane 2: purified BtrN. Lane 3, marker.
(b) UV-vis spectra of the reconstituted BtrN (bold line) and the reconstituted BtrN reduced with sodium dithionite (thin line). (c) EPR spectrum of the
reconstituted BtrN reduced with sodium dithionite. Recording conditions: temperature of 10 K, microwave power of 1 mW, and modulation of 0.6 mT.
plasma gamma globulin as a standard. Enzymes were freshly prepared
for each set of experiments.
through the Sep-Pak Plus Silica (Waters). The cartridge was washed
with 7 mL of 50% ethyl acetate in hexane, and the product,
bisdinitrophenyl DOS, was eluted with 5 mL of 20% methanol in ethyl
acetate. After removal of the solvent by a centrifugal evaporator, the
residue was dissolved in 100 µL of methanol. An aliquot (4 µL) of the
solution was injected into the same HPLC system as described above
and eluted with 70% aqueous methanol at a flow rate 0.9 mL/min at
room temperature. Elution was monitored at 350 nm. The standard
curves were generated with the authentic DOS treated with dinitrof-
luorobenzene in an identical manner to that for other samples so as to
correlate peak area with DOS amount.
Assay of DOIA Oxidation by BtrN. The reduced BtrN (20 µM)
was incubated with 2 mM DOIA, 2 mM SAM in 50 mM HEPES-
NaOH (pH 8.0) in the presence of 10 mM sodium dithionite at 28 °C
for 12 h. The reaction solution was analyzed for 3-amino-2,3-dideoxy-
scyllo-inosose (amino-DOI) formation by HPLC after treatment with
acetic anhydride and O-(p-nitrobenzyl)-hydroxylamine as previously
described.19
For 5′-deoxyadenosine detection, the enzyme reaction mixture was
quenched with 10% trichloroacetic acid and then centrifuged at 10 000
rpm for 1 min to remove the precipitated enzyme. An aliquot (5 µL)
of the solution was injected into a HPLC (Hitachi) equipped with a
PEGASIL ODS column (4.6 × 250 mm, Senshu). Elution was
performed with 20% aqueous methanol at a flow rate 0.9 mL/min at
room temperature and monitored at 254 nm.
Methionine was detected as a dinitrophenyl derivative. The enzyme
reaction mixture (20 µL) was derivatized with 20 µL of 5% 2,4-
dinitrofluorobenzene in methanol and 0.1 g/mL NaHCO3 at room
temperature for 3 h. After addition of 2 M NaOH aq. (10 µL) and H2O
(300 µL), the solution was washed with ethyl acetate (500 µL). The
aqueous layer (300 µL) was acidified with 2 M HCl (50 µL) and
extracted with ethyl acetate (800 µL). The organic layer (700 µL) was
then evaporated, and the residue was dissolved in 50 µL of methanol.
An aliquot (5 µL) of the solution was injected into the same HPLC
system as described above. Elution was performed with 20% aqueous
methanol at a flow rate 0.9 mL/min at room temperature and monitored
at 350 nm.
Time Course of DOS and 5′-Deoxyadenosine Production. The
reconstituted BtrN was reduced with 10 mM sodium dithionite at room
temperature for 1 h prior to use. The reduced BtrN (21 µM) was
incubated with 1 mM DOIA, 0.9 mM SAM, 10 mM DTH, 16 µM
BtrS, 10 mM glutamine, and 0.5 mM PLP in 50 mM HEPES-NaOH
(pH 8.0) at room temperature. The reaction was started by addition of
DOIA and quenched with the same volume of ethanol at specific periods
of time (5, 30, 60, 120, 180, 240, 360, 480, 600 min). The produced
SAM, DOS, and 5′-deoxyadenosine were quantified by HPLC. The
reactions were run in duplicate. 5′-Deoxyadenosine was detected as
mentioned above. The standard curves were generated with the authentic
5′-deoxyadenosine so as to correlate peak area with the amount of 5′-
deoxyadenosine.
For the SAM detection, the enzyme reaction was quenched with
50% ethanol and centrifuged at 10 000 rpm for 1 min to remove the
precipitated enzyme. SAM concentration was determined by HPLC with
the previously reported procedure.20 The standard curves were generated
with the authentic SAM (∼90% purity, MP Biochemicals) so as to
correlate peak area with the amount of SAM.
Kinetic Analysis of the BtrN Reaction. Kinetic measurements were
carried out in 50 mM HEPES-NaOH (pH 8.0) in the presence of 10
mM sodium dithionite in a final volume of 200 µL at 28 °C for 30
min. The initial velocities were determined by quantifying the generated
5′-deoxyadenosine by HPLC as mentioned above. The reduced BtrN
was prepared as described above except that the sodium dithionite
reduction was prolonged for 3 h.
For determination of the mechanism of the substrate inhibition by
DOIA, 2.9 µM of the reduced BtrN was incubated with varied
concentrations of SAM from 0.1 to 0.5 mM at different concentrations
of DOIA (0.4, 0.8, 1.5, 3 mM). Reactions were run in duplicate. For
determination of the kinetic parameters with respect to DOIA, 0.18-
0.25 µM of the reduced BtrN was incubated with 3 mM SAM and
varied concentrations of DOIA from 0.01 to 0.2 mM. Reactions were
run in triplicate, and the data were fitted to the generic equation for
uncompetitive substrate inhibition (V/[BtrN] ) kcat[DOIA]/(KMDOIA
+
[DOIA] + [DOIA]2/Ks)) by the least-squares method. For determination
of the kinetic parameters with respect to SAM, 0.22-0.54 µM of the
reduced BtrN was incubated with 0.1 mM DOIA and varied concentra-
tion of SAM from 0.08 to 5.0 mM. Reactions were run in duplicate,
and the data were fitted to the Michaelis-Menten equation (V/[BtrN]
) kcat*[SAM]/(KM + [SAM]) by the least-squares method.
Results
DOS was detected by the previously reported procedure19 with a
slight modification. The enzyme reaction was quenched by 50% EtOH,
and an aliquot (20 µL) of the mixture was treated with 20 µL of 5%
2,4-dinitrofluorobenzene in methanol, 10 µL of DMSO, and 2 µL of 2
M NaOH at 60 °C for 1 h. The reaction mixture was extracted with
500 µL of ethyl acetate. The ethyl acetate layer (400 µL) was
successively diluted with 3 mL of hexane, and the solution was passed
A gene-disruption mutant of btrN was constructed by standard
homologous recombination. The resultant mutant completely
lost its ability to produce antibiotic, which suggested that btrN
is required for biosynthesis of butirosin. Feeding experiments
of biosynthetic intermediates were further performed to identify
the interrupted biosynthetic reaction. The mutant recovered its
(19) Kudo, F.; Yamamoto, Y.; Yokoyama, K.; Eguchi, T.; Kakinuma, K. J.
Antibiot. 2005, 58, 766-774.
(20) She, Q. B.; Nagao, I.; Hayakawa, T.; Tsuge, H. Biochem. Biophys. Res.
Commun. 1994, 205, 1748-1754.
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15149

A R T I C L E S
Yokoyama et al.
Figure 3. (a) Amino-DOI derivatization with Ac2O and NBHA. (b) HPLC charts of in vitro activity assay of BtrN detected by the absorption at 254 nm:
(i) 0.1 mM BtrN, 2 mM SAM, 2 mM DOIA, 10 mM sodium dithionite, 50 mM HEPES-NaOH pH 8.0, (ii) without DOIA, (iii) without BtrN, and (iv)
without SAM.
butirosin productivity with DOS supplementation but not with
DOIA (Figure 1). Since the transamination of amino-DOI into
DOS is catalyzed by BtrS,21,22 these results indicate that BtrN
is involved in the oxidation of DOIA to give amino-DOI in the
butirosin producer B. circulans.
For enzymatic analysis, BtrN was overexpressed as an
N-terminal His-tagged protein and anaerobically purified by Ni
nitriloacetic acid (NTA) affinity chromatography, since other
radical SAM enzymes were reported to be sensitive to oxygen.
The purified BtrN was anaerobically incubated with DOIA,
SAM, and some other presumed cofactors such as flavin
cofactors, nicotinamide cofactors, and FeII(NH4)2(SO4)2. Sodium
dithionite was used to maintain strict anaerobic conditions. When
FeII(NH4)2(SO4)2 was added to the enzyme reaction solution, a
small amount of amino-DOI formation was observed. This result
suggested that the purified BtrN had an incomplete iron-sulfur
cluster and reconstitution of the 4Fe-4S cluster seemed to be
necessary for activity. Thus, reconstitution of the 4Fe-4S cluster
was examined. The anaerobically purified BtrN was thus treated
with FeII(NH4)2(SO4)2 and Na2S in the presence of dithiothreitol.
After removal of the unbound ferrous ion and sulfide ion by
gel filtration, the reconstituted BtrN was obtained as a dark
brown solution. The reconstituted BtrN showed absorption
maxima at 420 nm, which is characteristic for the [4Fe-4S]2+
cluster containing radical SAM proteins (Figure 2b). The
reconstituted BtrN was then reduced with 20 mM sodium
dithionite at room temperature for 1 h. The absorption at 420
nm was bleached (Figure 2b) as reported for other radical SAM
enzymes.23,24 The EPR spectrum of the reduced BtrN showed
g values of 1.92 and 2.04, which are characteristic for the [4Fe-
4S]+ cluster of radical SAM enzymes (Figure 2c). These results
suggested the generation of [4Fe-4S]+ cluster in the BtrN active
site. The spin quantitative analysis revealed that the reduced
BtrN contained 0.67 spins per polypeptide chain. Thus, about
Figure 4. Time dependence of DOS (∆) and 5′-deoxyadenosine (0)
formations and SAM (]) consumption. BtrN (0.030 mM) was incubated
with 1 mM DOIA and 0.9 mM SAM in the presence of 10 mM sodium
dithionite and BtrS system as described in Materials and Methods.
70% of BtrN was successfully reconstituted to obtain the [4Fe-
4S]+ cluster under these conditions.
The reduced BtrN was, then, incubated with DOIA and SAM.
HPLC analysis of the oxime derivative of the reaction product
revealed that amino-DOI was efficiently formed as shown in
Figure 3. Production of amino-DOI was not detected when BtrN,
DOIA, or SAM was omitted. The enzyme reaction product was
further confirmed by isolation of amino-DOI as nitrobenzy-
loxime derivative, which was identical in all respects to an
authentic standard.21 These results clearly proved that BtrN itself
catalyzes the oxidation of DOIA into amino-DOI in the presence
of SAM. Further analysis of the other reaction products showed
that 5′-deoxyadenosine and methionine were formed by incuba-
tion of SAM with BtrN and DOIA (see Supporting Information).
Time-course experiments of the amino-DOI and 5′-deoxy-
adenosine formation are shown in Figure 4. The amount of
amino-DOI was estimated after conversion to DOS by ami-
notransferase BtrS since quantitative analysis of amino-DOI was
not practical. Thus, DOS was detected as its 2,4-dinitrophenyl
derivative. The BtrN reaction was found to be extremely slow
(0.2 min-1) under these conditions. The first-order rate constant
(21) Yokoyama, K.; Kudo, F.; Kuwahara, M.; Inomata, K.; Tamegai, H.; Eguchi,
T.; Kakinuma, K. J. Am. Chem. Soc. 2005, 127, 5869-5874.
(22) Huang, F.; Li, Y.; Yu, J.; Spencer, J. B. Chem. Commun. 2002, 2860-
2861.
(23) Broderick, J. B.; Duderstadt, R. E.; Fernandez, D. C.; Wojtuszewski, K.;
Henshaw, T. F.; Johnson, M. K. J. Am. Chem. Soc. 1997, 119, 7396-
7397.
of the BtrS reaction with DOI and L-glutamine (1.6 × 10 min-1
,
(24) Ugulava, N. B.; Gibney, B. R.; Jarrett, J. T. Biochemistry 2000, 39, 5206-
5214.
unpublished result) was sufficiently faster than that for the BtrN
9
15150 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007

Study of a Radical SAM Dehydrogenase
A R T I C L E S
enosine formation, BtrN was found to be highly specific for
DOIA and able to strictly distinguish sugars from cyclitols,
which further supports its biological function in cells.
From a mechanistic point of view, the 5′-deoxyadenosyl
radical generated through the reductive cleavage of SAM by
the [4Fe-4S]+ cluster was presumed to be involved in the
abstraction of a substrate hydrogen atom to produce a radical
intermediate. To investigate this hypothesis, 2-deoxy-scyllo-[3-
2H]inosamine ([3-2H]DOIA) was prepared and subjected to the
BtrN reaction. [3-2H]DOIA (97% atom D) was prepared from
amino-DOI by reduction of the keto group with NaB2H4 (see
Supporting Information).
Figure 5. Substrate inhibition by DOIA. DOIA concentration was fixed
at 0.4 (]), 0.8 (0), 1.5 (∆), and 3 mM (o), while SAM concentration was
varied between 0.1 and 0.5 mM. Each data point represents the average of
two replicates.
The enzyme reaction was conducted using [3-2H]DOIA (10
mM) as a substrate in the presence of BtrN, SAM (3 mM), and
sodium dithionite. HPLC quantitative analysis revealed that 2.5
mM 5′-deoxyadenosine was formed in this reaction. The
generated 5′-deoxyadenosine was then purified and analyzed
by NMR spectroscopy and FAB-MS spectrometry (Figures 7
reaction, which allowed the coupling assay of BtrN with BtrS
to be used for quantitative analysis of the amino-DOI formation.
Equimolar amounts of 5′-deoxyadenosine and DOS were formed
along with consumption of the corresponding amount of SAM
during the BtrN-catalyzed reaction as shown in Figure 4. Thus,
these results clearly prove that one molecule of SAM is
reductively cleaved into methionine and 5′-deoxyadenosine
along with oxidation of one molecule of DOIA.
Kinetic analysis was then performed by detection of 5′-
deoxyadenosine as a product. In the kinetic analysis with varying
concentrations of DOIA, DOIA displayed substrate inhibition.
The inhibition mechanism was investigated by an initial velocity
study with varying concentrations of SAM at different fixed
concentrations of DOIA (Figure 5). The Lineweaver-Burk plots
in the conditions where DOIA acts as an inhibitor gave parallel
lines, thereby suggesting the uncompetitive substrate inhibition
of DOIA. Then the initial velocity data obtained by reactions
of varying concentrations of DOIA (10-200 µM) with saturat-
ing SAM (3 mM) were fitted to the equation for uncompetitive
substrate inhibition (eq 1), where kcat is the first-order rate
constant, KM DOIA is a Michaelis constant for DOIA, and Ks is
the substrate inhibition constant
2
and 8). The H NMR data clearly showed incorporation of
deuterium atom into the methyl group of 5′-deoxyadenosine
(Figure 7). In the 1H NMR data, in addition to the nondeuterated
doublet methyl signal, 0.016 and 0.032 ppm upfield doublet
signals caused by the geminal 2H isotope effect were observed
in the generated 5′-deoxyadenosine. These signals can be
assigned as monodeuterated (CH2D) and dideuterated (CHD2)
methyl groups as shown in Figure 8a. Further, observation of
molecular ions at m/z 252.1, 253.1, and 254.1 in the FAB-MS
spectrum also suggested the generation of the non, mono, and
dideuterated 5′-deoxyadenosines (Figure 8b). According to the
1H NMR and FAB-MS results the ratio of the three 5′-
deoxyadenosines was determined to be CH3:CH2D:CHD2 ) 1.0:
1.8:0.87. In addition, the kinetic isotope effect (KIE) in this
reaction was estimated to be 1.3 ( 0.1 (Figure 9), which
suggests that hydrogen-atom abstraction is involved in the
oxidation reaction but is not the rate-determining step. Further-
more, the reaction was carried out in a deuterium oxide buffer
(99.8% D) using nonlabeled DOIA. No incorporation of the
deuterium was observed in 5′-deoxyadenosine (see Supporting
Information). This suggests that the reaction proceeds through
direct hydrogen-atom abstraction from DOIA by the 5′-
deoxyadenosyl radical, not via the protein radical produced on
a cysteine or a tyrosine residue, since such a proton is
exchangeable with solvent.
V/[BtrN] ) kcat[DOIA]/(KM DOIA + [DOIA] +
[DOIA]2/Ks) (1)
The obtained data fit well to eq 1 (Figure 6a) and gave kinetic
values of kcat ) 2.3 ( 0.22 min-1, KM DOIA ) 0.022 ( 0.004
mM, and Ks ) 0.16 ( 0.01 mM. On the other hand, when SAM
concentrations were varied from 0.08 to 5.0 mM with fixed
concentration of DOIA (0.1 mM), the data were best fitted to
a simple hyperbolic curve, which indicated that SAM does not
behave as an inhibitor (Figure 6b). The apparent kinetic
constants at 0.1 mM of DOIA were KM SAM app ) 0.46 ( 0.10
mM and kcat app ) 1.2 ( 0.10 min-1. The substrate inhibition
by DOIA but not by SAM suggests an Ordered Bi Ter
mechanism for this reaction in which SAM is the first and DOIA
is the second substrate.25
The substrate recognition of BtrN was then investigated using
several cyclitols and sugars as substrate analogues (Table 1).
Although it is not clear whether or not the whole catalytic cycle
is operative with these substrate analogues since the relative
activities were estimated by the initial velocities of 5′-dexyad-
The above results revealed that the protium incorporated into
5′-deoxyadenosine in the reaction of BtrN with [3-2H]DOIA
was not from the solvent. Thus, generation of nondeuterated
and dideuterated 5′-deoxyadenosine in the reaction with [3-2H]-
DOIA proves the reversibility at the hydrogen abstraction step.
The mechanism of the generation of these 5′-deoxyadenosines
is illustrated in Figure 10. At first, SAM is reductively cleaved
in the presence of DOIA and transformed into the 5′-deoxyad-
enosyl radical, which directly abstracts the substrate deuterium
atom. The generated substrate radical then has a chance to
abstract the protium atom on 5′-deoxyadenosine to form
nonlabeled DOIA and the deuterated 5′-deoxyadenosyl radical,
which would then react with methionine to give the monodeu-
terated SAM. Reaction of the produced nonlabeled DOIA with
nonlabeled SAM produces nonlabeled 5′-deoxyadenosine (Fig-
ure 10b). On the other hand, monodeuterated SAM will also
(25) Charlier, H. A., Jr.; Plapp, B. V. J. Biol. Chem. 2000, 275, 11569-11575.
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15151

A R T I C L E S
Yokoyama et al.
Figure 6. Substrate concentration dependence of the initial velocity. (a) DOIA concentration was varied from 10 to 200 µM, while SAM concentration was
fixed at 3 mM. Each data point represents the average of three replicates. The line is the nonlinear root-mean-square fit to eq 1. (b) SAM concentration was
varied from 0.08 to 5.0 mM, while DOIA concentration was fixed at 0.1 mM. The line is the nonlinear root-mean-square fit to the Michaelis-Menten
equation: V/[BtrN] ) kcat[SAM]/(KM + [SAM]).
Table 1. Relative Activity against Substrate Analoguesa
a n.d.: activity was not detected.
react with the deuterated DOIA to generate dideuterated 5′-
deoxyadenosine (Figure 10c).
overexpressed enzyme. However, this profile was unexpected
because an NAD-dependent dehydrogenase, NeoA, was reported
to catalyze the same oxidation of DOIA in the DOS biosynthesis
in the neomycin producer Streptomyces fradiae.19 An NAD-
dependent dehydrogenase, BtrE, is actually encoded in the btr
gene cluster and thus was believed to catalyze the oxidation of
DOIA. However, several attempts to detect the DOIA dehy-
Discussion
BtrN was proved to be an enzyme catalyzing the oxi-
dation of DOIA to form amino-DOI based on the combination
of gene disruption experiments and in vitro assay of the
9
15152 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007

Study of a Radical SAM Dehydrogenase
A R T I C L E S
2
1
Figure 7. (a) H NMR (61 MHz, H2O) and (b) H NMR (400 MHz, D2O) of 5′-deoxyadenosine generated by BtrN reaction with [3-2H]DOIA.
5′-deoxyadenosyl radical is generated by reductive cleavage of
SAM using the [4Fe-4S]+ cluster as a reductant. The 5′-
deoxyadenosyl radical then abstracts a hydrogen atom from
DOIA to generate the radical intermediate, which is subsequently
transformed into amino-DOI upon transferring one electron,
possibly to the [4Fe-4S]2+ cluster to regenerate the [4Fe-4S]+
cluster.
Observation of the deuterium-atom transfer from [3-2H]DOIA
to 5′-deoxyadenosine clearly proved that the reaction involves
hydrogen-atom abstraction by the 5′-deoxyadenosyl radical.
Further, no deuterium was incorporated into the 5′-deoxyad-
enosine upon the BtrN reaction in a deuterium oxide buffer,
which suggests that the reaction proceeds through direct
abstraction of the substrate hydrogen atom by the 5′-deoxyad-
enosyl radical but not via the protein radical found in several
radical-dependent enzymes.26 Direct abstraction of a substrate
hydrogen atom by 5′-deoxyadenosyl radical has also been
proposed in biotin synthase.27 The substrate-bound crystal
structure of biotin synthase indicated direct hydrogen-atom
abstraction.28 The present study clearly shows that BtrN also
catalyzes the reaction by direct abstraction of the substrate
hydrogen atom by the 5′-deoxyadenosyl radical.
2
1
Figure 8. (a) H-decoupled H NMR (500 MHz, D2O) and (b) FAB-MS
(positive) of the isolated 5′-deoxyadenosine generated by reaction with
[3-2H]DOIA. Nonlabeled 5′-deoxyadenosine: m/z [M + H]+ 252.1.
Formation of nondeuterated and dideuterated 5′-deoxyad-
enosines by incubation of BtrN with [3-2H]DOIA suggests the
re-abstraction of a hydrogen atom on 5′-deoxyadenosine by the
substrate radical intermediate (Figure 10). This type of re-
abstraction is necessary for enzymes that utilize the 5′-
deoxyadenosyl radical in a catalytic manner, such as adenosyl
cobalamin-dependent enzymes29 and several radical SAM
enzymes.16 Transfer of the deuterium atom from the deuterated
substrate to 5′-deoxyadenosine was reported for two radical
SAM enzymes, biotin synthase27 and pyruvate formate lyase-
activating enzyme (PFL-AE).30 However, in both cases, only a
single incorporation of a deuterium to 5′-deoxyadenosine was
observed.27,30 On the other hand, the BtrN reaction using [3-2H]-
DOIA generated nondeuterated and dideuterated 5′-deoxyad-
enosines, which suggests that the hydrogen-atom abstraction is
reversible. The reversibility of the hydrogen-atom abstraction
Figure 9. Time dependence of DOS formation in the reaction with
nonlabeled (]) or [3-2H]DOIA (0).
drogenase activity failed in our laboratory (data not shown).
As we suggested in the previous report, BtrE lacks two important
zinc binding motifs and thus may have a different type of
enzymatic activity.19 Combining those genetic information and
our experimental results, it is clear that Bacillus circulans has
a different system for DOS biosynthesis from that of the
neomycin producer, S. fradiae. Functional analysis of BtrE is
still underway.
(26) Stubbe, J.; van der Donk, W. A. Chem. ReV. 1998, 98, 705-762.
(27) Escalettes, F. F. D.; Bui, B. T. S.; Lesage, D.; Marquet, A. J. Am. Chem.
Soc. 1999, 121, 3571-3578.
The UV-vis absorption spectra and the EPR spectrum of
the reconstituted BtrN confirmed that BtrN is a [4Fe-4S] cluster
containing radical SAM enzyme. The proposed catalytic mech-
anism of the BtrN reaction is shown in Figure 11. The
(28) Berkovitch, F.; Nicolet, Y.; Wan, J. T.; Jarrett, J. T.; Drennan, C. L. Science
2004, 303, 76-79.
(29) Reed, G. H. Curr. Opin. Chem. Biol. 2004, 8, 477-483.
(30) Frey, M.; Rothe, M.; Wagner, A. F.; Knappe, J. J. Biol. Chem. 1994, 269,
12432-12437.
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15153

A R T I C L E S
Yokoyama et al.
Figure 10. Mechanism of (a) monodeuterated 5′-deoxyadenosine (Ad-CH2D), (b) nondeuterated 5′-deoxyadenosine (Ad-CH3), and (c) dideuterated 5′-
deoxyadenosine (Ad-CHD2) formation by reaction with [3-2H]DOIA.
on 5′-deoxyadenosine. Thus, ketone formation might be the rate-
determining step.
Each catalytic cycle requires one molecule of SAM and gen-
erates one molecule of methionine and 5′-deoxyadenosine,
which reveals that the subsequent ketone formation from the
substrate radical intermediate is not catalyzed by another 5′-
deoxyadenosyl radical. Therefore, one electron of the substrate
radical intermediate must be somehow transferred to an electron
acceptor to complete the ketone formation. However, no extra
electron acceptor was present in the reaction solution. On the
other hand, the [4Fe-4S]2+ cluster of BtrN must be reduced to
[4Fe-4S]+ for the next catalytic cycle. Although reduction of
the [4Fe-4S]2+ cluster back to [4Fe-4S]+ cluster by sodium
dithionite cannot be ruled out, it is reasonable to conclude that
the electron released from the substrate radical intermediate is
somehow transferred to the [4Fe-4S]2+ cluster to produce amino-
DOI and regenerate the [4Fe-4S]+ cluster for the next catalytic
Figure 11. Proposed reaction mechanism of BtrN.
cycle. Since the midpoint redox potential of the [4Fe-4S]1+/2+
and the small KIE value suggest that the subsequent ketone
clusters of radical SAM enzymes is extremely negative
(∼-500 mV), reduction of the oxidized [4Fe-4S]2+ cluster under
formation is slower than the deuterium-atom abstraction, which
enables the substrate radical intermediate to re-abstract a protium
9
15154 J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007

Study of a Radical SAM Dehydrogenase
A R T I C L E S
in Figure 11. Further, because the BtrN reaction was found to
be slow and the just prior amino transfer reaction catalyzed by
BtrS is known to be relatively fast and reversible,12 the fact
that DOIA inhibits the BtrN reaction by uncompetitive substrate
inhibition may imply that BtrN controls the overall production
rate of butirosin.
Figure 12. Cleland’s shorthand notation of the BtrN reaction.
A similar oxidation reaction by a radical SAM enzyme is
known for the reaction of anaerobic sulfatase maturating
enzymes, which post-translationally oxidizes a serine or cysteine
residue of sulfatases to formyl glycine.34,35 Since the formyl
glycine formation requires SAM and the radical SAM (CXXX-
CXXC) motif, involvement of the 5′-deoxyadenosyl radical was
proposed.34,35 However, no further mechanistic investigation was
reported.
Enzymatic oxidations of hydroxyl groups generally proceed
through hydride transfers catalyzed by nicotinamide- or flavin-
dependent enzymes. Several examples are known for radical
oxidations of hydroxyl groups using oxygen as an electron
acceptor in biological systems. P450 oxidoreductases are con-
sidered to catalyze oxidation of a hydroxylated carbon center
to produce a ketone or carboxylate.36,37 Galactose oxidase also
catalyzes the unusual oxidation of the C-6 hydroxyl group of
galactose into an aldehyde using CuI for oxygen activation.38
However, no report has been made for the oxygen-independent
oxidation of a hydroxyl group by a radical mechanism, except
for the aforementioned reaction by sulfatase maturating en-
zymes. Our results described here elucidate the mechanism of
anaerobic radical oxidation of hydroxyl groups for the first time.
This type of oxidation suggests the use of SAM as an oxidant
substitute, instead of molecular oxygen, in anaerobic metabo-
lisms.
physiological conditions seems to be difficult. However, binding
of SAM to oxidized [4Fe-4S]2+ cluster was shown to signifi-
cantly elevate the reduction potential,31 and reduction of the
[4Fe-4S]2+ cluster by flavodoxin appeared to occur by further
binding of a substrate.32 Thus, in the case of BtrN, direct
reduction of the [4Fe-4S]2+ cluster by the radical intermediate
seems to be possible when coordination of 5′-deoxyadenosine,
methionine, or the substrate radical intermediate to a sphere of
the [4Fe-4S]2+ cluster causes a conformational change of the
active site and elevates the reduction potential of the cluster. In
this case, the electron might be directly transferred from the
radical intermediate to the oxidized cluster, although the
existence of a kind of internal electron sink mechanism such
as a modified amino acid side chain and a disulfide bond cannot
be excluded. Further mechanistic analysis of ketone formation
is currently underway.
The oxygen-independent coproporphyrinogen-III oxidase,
HemN, is currently the only radical SAM enzyme that is proved
to require one electron-transfer step. However, in this case, an
uncharacterized electron acceptor in a cell-free extract of E. coli
is required to complete the reaction.33 Thus, we propose that
BtrN is the first example of a self-reactivating radical SAM
enzyme.
Kinetic analysis gave further insights into the mechanism of
the BtrN reaction. Importantly, we observed uncompetitive
substrate inhibition by DOIA at concentrations greater than 0.08
mM, while SAM did not inhibit the reaction. Since the same
amounts of amino-DOI and 5′-deoxyadenosine were formed in
the BtrN reaction and no hydrogen atom from solvent was
incorporated into the produced 5′-deoxyadenosine, we conclude
that inhibition by DOIA is caused by a simple uncompetitive
substrate inhibition but not by inactivating side reactions. Thus,
the results of the kinetic studies suggest that the BtrN reaction
proceeds by an ordered Bi Ter mechanism in which DOIA is
the second substrate incorporated into the active site and amino-
DOI is the first product released as shown in Figure 12. The
Ordered mechanism is consistent with the mechanism proposed
Acknowledgment. Research Fellowships for Young Scientists
support to K.Y. from JSPS is gratefully acknowledged. This
work was supported in part by the Naito Foundation, the Uehara
Memorial Foundation, and Grant-in-Aid for Scientific Research
from MEXT.
Supporting Information Available: Experimental procedures
including gene disruption study, all spectroscopic data, and
HPLC charts. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA072481T
(34) Fang, Q.; Peng, J.; Dierks, T. J. Biol. Chem. 2004, 279, 14570-14578.
(35) Benjdia, A. L. J.; Guillot, A.; Vaudry, H.; Rabot, S.; Berteau, O. J. Am.
Chem. Soc. 2007, 129, 3462-3463.
(31) Hinckley, G. T.; Frey, P. A. Biochemistry 2006, 45, 3219-3225.
(32) Mulliez, E.; Padovani, D.; Atta, M.; Alcouffe, C.; Fontecave, M. Biochem-
istry 2001, 40, 3730-3736.
(33) Layer, G.; Moser, J.; Heinz, D. W.; Jahn, D.; Schubert, W. D. EMBO J.
2003, 22, 6214-6224.
(36) Meunier, B.; de Visser, S. P.; Shaik, S. Chem. ReV. 2004, 104, 3947-
3980.
(37) Oh, N. Y.; Suh, Y.; Park, M. J.; Seo, M. S.; Kim, J.; Nam, W. Angew.
Chem., Int. Ed. Engl. 2005, 44, 4235-4239.
(38) Whittaker, J. W. Chem. ReV. 2003, 103, 2347-2363.
9
J. AM. CHEM. SOC. VOL. 129, NO. 49, 2007 15155