Regio- and chemoselective 6′-N-derivatization of aminoglycosides: Bisubstrate inhibitors as probes to study aminoglycoside 6′-N- acetyltransferases

Article abstract of DOI:10.1002/anie.200501399

(Chemical Equation Presented) Complex nanomolar inhibitors in one pot: Aminoglycoside-coenzyme A derivatives were prepared through an efficient regioselective 6′-N modification of aminoglycosides (see scheme). These bisubstrates show tight-binding competi

Full text of DOI:10.1002/anie.200501399

Angewandte
Chemie
Aminoglycosides
DOI: 10.1002/anie.200501399
Regio- and Chemoselective 6’-N-Derivatization of
Aminoglycosides: Bisubstrate Inhibitors as
Probes To Study Aminoglycoside 6’-N-
Acetyltransferases**
Feng Gao, Xuxu Yan, Oliver M. Baettig,
Albert M. Berghuis, and Karine Auclair*
Aminoglycosides are among the most commonly used broad-
spectrum antibiotics.[1] The biological activity relies on their
high affinity for the major groove of bacterial 16S rRNA,[2]
thereby impeding protein synthesis. A number of amino-
glycosides also display antiviral activity owing to specific
interactions with viral RNA.[3] The rapid emergence of
aminoglycoside resistance in the treatment of infections,
however, is a serious threat.[4] In clinical isolates of amino-
glycoside-resistant strains, the most frequently observed
cause of resistance is the expression of N-acetyltransferases.[5]
For example, aminoglycoside 6’-N-acetyltransferase type Ii
(AAC(6’)-Ii) is chromosomally encoded in Enterococcus
faecium, which is one of the leading causes of hospital-
acquired infections.[6] Studies by Wright and co-workers
suggest that catalysis by AAC(6’)-Ii occurs through an
ordered bi-bi mechanism in which acetyl coenzyme A (Ac-
CoA) must bind the enzyme before the aminoglycoside.[5,7]
Next, attack of the aminoglycoside 6’-NH2 at the thioester of
Ac-CoA is believed to generate a tetrahedral intermediate
that collapses to yield an 6’-N-acetylaminoglycoside and CoA.
Although crystal structures have been reported for complexes
of AAC(6’)-Ii with either Ac-CoA or CoA, crystallization
experiments of enzyme–aminoglycoside complexes have not
been successful.[8] It was envisaged that either 6’-N-(S-
CoA)aminoglycoside derivatives or bisubstrates (Scheme 1)
would facilitate the study of this important class of enzymes.
Bisubstrate analogues have exhibited inhibition of serotonin
[*] F. Gao, X. Yan, Prof. K. Auclair
Department of Chemistry, McGill University
801 Sherbrooke Street West, MontrØal, QuØbec, H3A2K6 (Canada)
Fax: (+1)514-398-3797
E-mail: karine.auclair@mcgill.ca
O. M. Baettig, Prof. A. M. Berghuis
Department of Biochemistry, McGill University
801 Sherbrooke Street West, MontrØal, QuØbec, H3A2K6 (Canada)
[**] This work was supported by the National Science and Engineering
Research Council of Canada (NSERC), by the Canadian Institute of
Health Research (CIHR), and by the Cancer Research Center (CRC).
F.G., X.Y., and O.M.B. were supported by scholarship awards from
the Chemical Biology Strategic Training Initiative of CIHR. The
authors are grateful to G. D. Wright at McMaster University for
sharing his AAC(6’)-Ii expression plasmid.
Supporting information (including experimental procedures, char-
acterization data, some NMR spectra and HPLC traces) for this
article is available on the WWW under http://www.angewandte.org
or from the author.
Angew. Chem. Int. Ed. 2005, 44, 6859 –6862
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6859

Communications
built from neamine (4), kanamycin A (5), and ribostamycin
(6), which are examples of nonsubstituted, 4,6-substituted,
and 4,5-substituted aminoglycosides, respectively, and are
AAC(6’)-Ii substrates. The key step in the synthesis of 1–3
from 4–6 is the preparation of the bromide intermediates 8a,
9, and 10, respectively (Scheme 2). The trifluoroacetic acid
(TFA) salt of intermediate 8a has been synthesized previ-
ously by using orthogonal protection in four steps, and with a
resultant very low yield (11.6%).[14] As a more-efficient
alternative to orthogonal protection/deprotection schemes,
we envisaged the use of N-(2-bromoacetyl)oxy-5-norbor-
nene-endo-2,3-dicarboximide (bromoacetyl-NBD ester, 7a)
to transfer regioselectively a bromoacetyl group to the 6’-NH2
of the aminoglycosides. The Boc-NBD ester has been
reported to effect regioselective Boc protection of amino-
glycosides;[15] however, NBD esters have not been employed
for direct aminoglycoside derivatization.
Remarkably, simply mixing the free base neamine and
reagent 7a in either acetone/H2O or acetonitrile/H2O (1:1) at
room temperature and in a vial open to the air for a few
minutes (monitored by ESI MS) was sufficient to complete
the reaction. In contrast, extended reaction times yielded
complicated mixtures, possibly arising from the nucleophilic
attacks of amino groups at the newly formed bromide. The
desired N-6’-bromoacetylneamine (8a) was isolated in good
yield (70%) by quenching of the reaction mixture with TFA
Scheme 1. Target aminoglycoside–CoA bisubstrates 1–3. Ade=ade-
nine.
acetyltransferase,[9] GCN5 histone acetyltrans-
ferase,[10] and carnitine acetyltransferases.[11] 3-
N-(2-S-CoA-acetyl)gentamicin C1a, the only
aminoglycoside-CoA derivative reported so far,
was prepared enzymatically by using AAC(3)-I
(3 mg) to yield the desired product (0.89 mg),[12]
which was subsequently found to inhibit
AAC(3)-I with high affinity. To date, however,
there are no reports of chemical syntheses of
CoA-aminoglycoside derivatives. Naturally
occurring aminoglycosides are complex mole-
cules, and their regioselective modification
remains challenging. Currently, the use of judi-
cious
functional-protection
chemistry
is
common; however, the overall yields are low.[13]
Herein we report an efficient procedure for the
regioselective derivatization of unprotected ami-
noglycosides at the 6’-NH2 position. This strategy
was shown to proceed with high chemoselectiv-
ity towards the assembly of CoA-aminoglycoside
derivatives 1–3 and 11a–c (Scheme 2 and
Scheme 3). Activity assays of these bisubstrates
reveal novel nanomolar tight-binding competi-
tive inhibition of AAC(6’)-Ii.
Scheme 2. Synthesis of the bisubstrate analogues 1–3; see Scheme 1 for R1, R2, R3; 4:
neamine; 5: kanamycin A; 6: ribostamycin; intermediates 9 and 10 were not isolated.
after 10 min and subsequent purification by reversed-phase
HPLC. To avoid the isolation step and potential decomposi-
The target bisubstrates 1–3 were designed based on the
proposed tetrahedral intermediate that results from the
attack of the aminoglycoside 6’-NH2 on the thioester carbonyl
of Ac-CoA in the active site of the enzyme.[5] As the crystal
structures of AAC(6’)-Ii[8] did not reveal a potential oxyanion
hole, it was envisaged that a bisubstrate containing an amide-
based linker could mimic the intermediate. The targets were
tion, a one-pot synthesis of 1 from 4 and CoA through 8a, was
evaluated. Bisubstrate 1 was obtained in excellent yield
(83%) after purification by HPLC (the crude sample was
more than 85% pure (Figure 1a)). The reaction proceeded
equally well with the aminoglycosides 5 and 6 (72 and 67%,
respectively). In spite of their significant structural differ-
ences, kinetic studies revealed that 1–3 showed nanomolar
6860 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6859 –6862

Angewandte
Chemie
propanoyl)neamine); 6’-N-acryloyl neamine (the elimination
product); and a propanoyl-conjugated neamine dimer (either
through Sn2 substitution at the bromide by a second neamine
function or by Michael addition of a second neamine group to
6’-N-acryloyl neamine). Similar products resulted during the
initial efforts to prepare 11b and 11c. A difference in the
reactivities of these longer linkers was expected.[17] Indeed,
the 3-bromo-n-propanoyl, 4-bromo-n-butyroyl, and 5-bromo-
valeroyl neamine derivatives (8b–d, from the reaction of
neamine with 7b–d) are much less electrophilic than 8a.
Moreover, triethylamine may be too nucleophilic and used in
too large an amount, which may favor N alkylation over
S alkylation. As most reported chemoselective S-/N-alkyla-
tion procedures[17] are not compatible with our reagents, we
reasoned that a less-nucleophilic base at a lower concentra-
tion may favor S alkylation. Model studies were performed by
using N-acetylcysteine as a mimic of CoA, and an array of
bases were screened, including KHCO3, NaHCO3, diisopro-
pylethylamine
(DIPEA),
1,4-diazabicyclo[2.2.2]octane
Figure 1. HPLC chromatograms for the purification of bisubstrate
analogues: a) 1 and b) 11a.
(DABCO), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), and
4-dimethylaminopyridine (DMAP). Excellent yields resulted
when DIPEA (20 equiv) was used and the mixture was
ultrasonicated (5 min). When applied to CoA, these con-
ditions allowed the preparation of 11a, 11b, and 11c in
moderate to excellent yields (91, 52, and 93%, respectively).
The HPLC trace of crude 11a is shown in Figure 1b.
AAC(6’)-Ii-inhibition assays showed that 11a is the most
potent bisubstrate inhibitor of this series. Remarkably, the
enzyme binds 11a ꢀ 200-fold tighter than its natural substrate
Ac-CoA (Km = 9.6 mm). A further increase in the length of the
linker, however, rapidly leads to a decrease in activity
(Table 1). To demonstrate the usefulness of these inhibitors
as structural and mechanistic probes, they were studied by X-
ray crystallographic analysis. Although previous crystalliza-
tion experiments of AAC(6’)-Ii–aminoglycoside complexes
had not been successful,[8b,c] bisubstrates 1–3 and 11a–c
crystallized well with the enzyme, providing X-ray diffraction
data to ꢀ 2.0-Š resolution. Preliminary analysis of the
diffraction data for the complex with bisubstrate 11a
(Figure 2) suggests that the conformation of the aminoglyco-
side bound to AAC(6’)-Ii is very different from that reported
for AAC(6’)-Iy,[18] even though both enzymes catalyze the
same reaction. Detailed structural and mechanistic analysis of
these structures will be reported elsewhere.
tight-binding competitive inhibition of AAC(6’)-Ii (Table 1).
This observation is consistent with the fact that AAC(6’)-Ii
has a broad substrate specificity.[5]
Table 1: The AAC(6’)-Ii-inhibition constants (Ki) for the bisubstrates.
Inhibitor
Ki [nm]
Inhibitor
Ki [nm]
1
2
3
76Æ25
111Æ28
119Æ14
11a
11b
11c
43Æ23
161Æ98
7990Æ2663
The size and geometry of linkers in serotonin-CoA
bisubstrates have been reported to show important effects
on the inhibition of serotonin N-acetyltransferase.[9,16]
A
similar effect was expected for AAC(6’)-Ii and was inves-
tigated by using bisubstrates 1 and 11a–c (Scheme 3).
Unfortunately, the procedure described herein for the syn-
thesis of bisubstrates 1–3 was not easily applicable to the
preparation of 11a–c. The major products observed when
synthesis of 11a was attempted were: an ammonium bromide
(likely from the addition of triethylamine to 6’-N-(3-bromo-n-
After the success of this methodology in the preparation
of CoA-aminoglycoside bisubstrates,
its general applicability to acylation
was investigated (Scheme 4). Most of
the acyl groups tested were trans-
ferred regioselectively to neamine
with excellent yields. NBD esters of
highly hindered acyl groups such as
2-methylbenzoyl and 2,6-dichloro-
benzoyl, however, did not react at
all. The regioselectivity of the trans-
fer to kanamycin A, ribostamycin,
and neomycin was tested with ben-
zoyl-NBD and showed excellent
selectivity for N-6’-acylation.
Scheme 3. Reagents and conditions: a) acetone/H2O (1:1), room temperature, 10 min; b) CoA,
DIPEA (20 equiv), sonicate 5 min; then room temperature, 1 h; 11a: 91%; 11b: 52%; 11c: 93%.
Angew. Chem. Int. Ed. 2005, 44, 6859 –6862
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6861

Communications
Keywords: acetyltransferases · aminoglycosides · antibiotics ·
.
coenzyme A · inhibitors
[1] a) G. D. Wright, A. M. Berghuis, S. Mobashery, Resolving the
Antibiotic Paradox (Ed.: B. P Rosen, S. Mobashery), Kluwer
Academic/Plenum, New York, 1998, pp. 27 – 69; b) A. Coates,
Y.-M. Hu, R. Bax, C. Page, Nat. Rev. Drug Discovery 2002, 1,
895 – 910; c) S. B. Vakulenko, S. Mobashery, Clin. Microbiol.
Rev. 2003, 16, 430 – 450.
[2] a) M. I. Recht, S. Douthwaite, J. D. Puglisi, EMBO J. 1996, 18,
3133 – 3138; b) D. Fourmy, M. I. Recht, S. C. Blanchard, J. D.
Puglisi, Science 1997, 274, 1367 – 1375; c) D. Fourmy, M. I. Recht,
J. D. Puglisi, J. Mol. Biol. 1998, 277, 347 – 362.
[3] a) J. Gallego, G. Varani, Acc. Chem. Res. 2001, 34, 836 – 843;
b) T. Hermann, Angew. Chem. 2000, 112, 1962 – 1979; Angew.
Chem. Int. Ed. 2000, 39, 1890 – 1905.
[4] C. Walsh, Nature 2000, 406, 775 – 781.
[5] G. D. Wright, P. Ladak, Antimicrob. Agents Chemother. 1997, 41,
956 – 960.
[6] B. E. Murray, Clin. Microbiol. Rev. 1990, 3, 46 – 65.
[7] a) K.-A. Draker, G. D. Wright, Biochemistry 2004,
43, 446 – 454; b) K.-A. Draker, D. B. Northrop,
G. D. Wright, Biochemistry 2003, 42, 6565 – 6574.
[8] a) L. E. Wybenga-Groot, K.-A. Draker, G. D.
Wright, A. M. Berghuis, Structure 1999, 7, 497 –
507; b) D. L. Burk, N. Ghuman, L. E. Wybenga-
Groot, A. M. Berghuis, Protein Sci. 2003, 12, 426 –
437; c) D. L. Burk, B. Xiang, C. Breitbach, A. M.
Berghuis, Acta Crystallogr. Sect. D, 2005, 61, 1273-
1279.
Figure 2. Crystallographically determined structure of the E. faecium
AAC(6’)-Ii dimer (blue and red) when complexed to the bisubstrate
inhibitor 11a (stick representation, C: light blue, N: dark blue, O: red,
P: gray, S: orange.
[9] C. M. Kim, P. A. Cole, J. Med. Chem. 2001, 44,
2479 – 2485.
[10] a) A. N. Poux, M. Cebrat, C. M. Kim, P. A. Cole, R.
Marmorstein, Proc. Natl. Acad. Sci. USA 2002, 99,
14065 – 14070; b) V. Sagar, W.-P. Zheng, P. Thomp-
son, P. A. Cole, Bioorg. Med. Chem. 2004, 12, 3383 –
3390.
[11] J. F. A. Chase, P. K. Tubbs, Biochem. J. 1969, 111,
225.
[12] J. W. Williams, D. B. Northrop, J. Antibiot. 1979, 32,
1147 – 1154.
Scheme 4. Regioselective N-6’-acylation of neamine. a) acetone/H2O (1:1), room temperature,
30 min. Bn=benzyl, Naph=naphthyl
[13] a) J. Roestamadji, I. Grapsas, S. Mobashery, J. Am.
Chem. Soc. 1995, 117, 80 – 84; b) J. Haddad, L. P.
Kotra, B. Llano-Sotelo, C.-K. Kim, E. F. Azuce-
na, Jr., M.-Z. Liu, S. B. Vakulenko, C. S. Chow, S.
In conclusion, we have described a highly efficient
Mobashery, J. Am. Chem. Soc. 2002, 124, 3229 – 3237; c) J.
Roestamadji, S. Mobashery, Bioorg. Med. Chem. Lett. 1998, 8,
3483 – 3488.
synthetic strategy for the regioselective acylation of amino-
glycosides at the 6’-NH2 group. This process was successfully
applied to the one-pot synthesis of 6’-N-(S-CoA)aminoglyco-
side analogues. Most of these bisubstrates were nanomolar
tight-binding competitive inhibitors of AAC(6’)-Ii, an impor-
tant enzyme leading to antibiotic resistance. The high potency
of these bisubstrate inhibitors and the crystal structure of
enzyme-bound 11a suggest that they may be good mimics of
one of the enzymatic reaction intermediates.[5] The bisub-
strates reported here have allowed crystallization of AAC(6’)-
Ii with aminoglycoside derivatives. The resulting structures
should provide valuable guidance in further studies of this
enzyme and other members of this family.
[14] J. Roestamadji, S. Mobashery, Bioorg. Med. Chem. Lett. 1998, 8,
3483 – 3488.
[15] I. Grapsas, Y. J. Mobashery, J. Org. Chem. 1994, 59, 1918 – 1922.
[16] a) E. M. Khalil, J. De Angelis, M. Ishii, P. A Cole, Proc. Natl.
Acad. Sci. USA 1999, 96, 12418; b) E. Wolf, J. De Angelis, E. M.
Khalil, P. A. Cole, S. K. Burley, J. Mol. Biol. 2002, 317, 215.
[17] a) I. K. Moiseev, N. V. Makarova, M. N. Zemtsova,Russ. J. Org.
Chem. 2003, 39, 1685 – 1701; b) D. A. Perry, F. M. Uckun,
Tetrahedron Lett. 2001, 42, 1859 – 1861; c) C.-C. Yang, C. K.
Marlowe, R. Kania, J. Am. Chem. Soc. 1991, 113, 3177 – 3178.
[18] M. W. Vetting, S. Magnet, E. Nieves, S. L. Roderick, J. S.
Blanchard, Chem. Biol. 2004, 11, 565 – 573.
Received: April 22, 2005
Revised: July 25, 2005
Published online: October 5, 2005
6862 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 6859 –6862