Angewandte
Chemie
kanamycin B was efficiently converted into kanamycin A in
the presence of 2-oxoglutarate (a-ketoglutarate; a-KG),
(NH4)2Fe(SO4)2·6H2O, and NADPH. Evidently, maximal
KanJ activity was shown to be dependent on exogenous
ferrous ions, although residual activity was also observed in
the absence of exogenous FeII. This dependency was likely the
result of small amounts of enzyme-bound ferrous ions that
were retained during enzyme purification. However, the
addition of EDTA completely abolished the activity. Also, in
the absence of a-KG, formation of kanamycin A was not
observed, thus suggesting that KanJ requires these cofactors
for the enzymatic activity as a standard FeII/a-KG-dependent
dioxygenase.[7] The cofactor requirement was expected from
the homology analysis of KanJ appearing in this family of
dioxygenase (see Figure S2 in the Supporting Information).
The KanJ and KanK reaction product from kanamycin B was
additionally isolated from a large-scale enzyme reaction
(15 mL, 15.5 mg) and the structure of the isolated compound
(7 mg) was confirmed as kanamycin A by NMR spectroscopy
and FAB/MS (see Figures S4–S7).
When NADH was added to the reaction instead of
NADPH, kanamycin A was not detected, thereby indicating
that KanK prefers to use NADPH (see Figure S8 in the
Supporting Information). Even in the absence of NADPH or
KanK, kanamycin B was efficiently consumed by KanJ and
a new DNP derivative having a m/z 820 was detected by LC/
ESI/MS (see Figures S8 and S9), thus indicating that the
presumable KanJ reaction product, 2’-oxokanamycin, was
decomposed into a pseudo-disaccharide during derivatization
with DNFB under basic conditions. Because any attempt to
isolate the KanJ reaction product has failed so far, the KanJ
enzymatic reaction was treated with NaBH4 or NaBD4 to
clarify the expected ketone formation. 1H NMR spectroscopy
and LC/ESI/MS analysis revealed that this coupled reaction
indeed afforded kanamycin A in addition to its diastereomer
(Figure 2 and Figure S10). Furthermore, incorporation of
a deuterium atom into the C2’-position was detected by
1H NMR spectroscopy as confirmed by the lack of a coupled
signal with H1’ (Figure 2) and also by 2H NMR spectroscopy
(Figure S11). This result clearly demonstrated that the KanJ-
generated 2’-oxokanamycin was reduced by NaBD4. Another
product, ammonia, was also clearly detected by the use of
a coupling assay with glutamate dehydrogenase (Figure S12).
These results support the synthesis of 2’-oxokanamycin
and release of ammonia by KanJ. Consequently, in the KanJ/
KanK coupling reaction, KanK appeared to stereoselectively
reduce 2’-oxokanamycin in the presence of NADPH to afford
kanamycin A (Scheme 1). The possibility of the reverse
reaction from kanamycin A to 2’-oxokanamycin by KanK
with NADP+ was not observed (data not shown). Therefore,
the equilibrium of the KanJ/KanK reaction appeared to be
largely on the side of kanamycin A production, presumably
because of the NADP+ formation and the release of
ammonia. Notably, this mechanism is distinct from that of
the well-known adenosine deaminase which catalyzes hydra-
tion of the adenine ring with subsequent release of ammonia,
thus forming a carbonyl group where no redox reaction is
involved.[8] In the transformation of kanamycin B into
kanamycin A, oxidation and reduction is required to
Figure 2. 1H NMR spectra of the KanJ reaction product reduced with
NaBH4 or NaBD4. A) Authentic kanamycin A. B) Isolated product of
KanJ-NaBH4. C) Isolated product of the reaction with KanJ and NaBD4
(500 MHz, D2O). Assay conditions: KanJ (25.8 mm) was incubated with
kanamycin B (2.1 mm), a-KG (6 mm), (NH4)2Fe(SO4)2·6H2O (6.8 mm)
at 288C for 48 h. NaBD4 or NaBH4 was added excessively into the
enzyme assay solutions and stirred overnight at 48C. See experimental
section for isolation procedure.
remove an amino group and to form a hydroxy group,
respectively. The first step of this transformation, catalyzed by
KanJ, should be rather comparable to the oxidative deami-
nation by glutamate dehydrogenase which utilizes NAD+/
NADP+ as an oxidant, and the second reaction, involving
KanK, is similar to the reverse reaction with a “keto”-
compound as the substrate.[9]
Two possible reaction mechanisms were hypothesized to
explain the process of the ketone formation by KanJ: A) the
reactive [FeIV O] species breaks the inactive C2’ H2’ bond in
=
À
kanamycin B homolytically to produce the kanamycin B
Angew. Chem. Int. Ed. 2012, 51, 3428 –3431
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim