Application of glycodiversification: Expedient synthesis and antibacterial evaluation of a library of kanamycin B analogues

Article abstract of DOI:10.1021/ol0497685

The expedient synthesis of a library of kanamycin B analogues is reported. The revealed SAR will guide future designs in developing kanamycin-type aminoglycoside antibiotics against drug-resistant bacteria.

Full text of DOI:10.1021/ol0497685

ORGANIC
LETTERS
2004
Vol. 6, No. 9
1381-1384
Application of Glycodiversification:
Expedient Synthesis and Antibacterial
Evaluation of a Library of Kanamycin B
Analogues
Jie Li,† Jinhua Wang,† Przemyslaw Greg Czyryca,† Huiwen Chang,†
Thomas W. Orsak,‡ Richard Evanson,‡ and Cheng-Wei Tom Chang*
,†
Department of Chemistry and Biochemistry, Utah State UniVersity, 0300 Old Main
Hill, Logan, Utah 84322-0300, and Department of Chemistry and Biochemistry,
Brigham Young UniVersity, ProVo, Utah 84602
chang@cc.usu.edu
Received February 9, 2004
ABSTRACT
The expedient synthesis of a library of kanamycin B analogues is reported. The revealed SAR will guide future designs in developing kanamycin-
type aminoglycoside antibiotics against drug-resistant bacteria.
Kanamycin belongs to a group of aminoglycoside antibiotics
with 4,6-di-O-substituted 2-deoxystreptamine (Figure 1).1
Kanamycin shows prominent antibacterial activity against
both gram-positive and gram-negative susceptible strains of
bacteria. However, the emergence of aminoglycoside-
resistant bacteria has rendered kanamycin obsolete for clinical
uses.2 As one of the antidotes against the problem of drug
resistance, numerous chemical modifications of kanamycin
have been reported with the goal of reviving its activity
toward resistant bacteria.3 Except for a few publications,4
the majority of derivatives use carbamates as protecting
groups for the amino groups of kanamycin, resulting in the
production of kanamycin with polycarbamate groups. Two
drawbacks were often encountered with these molecules.
First, the poor solubility of polycarbamate molecules results
in difficulties in purification and characterization of these
compounds. Second, the kanamycin scaffold imposes limited
options for structural modifications.
Figure 1. Structure of kanamycin.
Kanamycin B consists of neamine (rings I and II) and
3-amino-3-deoxyglucopyranose (ring III), with ring III being
linked to the O-6 of neamine via an R-glycosidic bond.
Neamine has been proven to be the essential component for
the antibacterial activity of neamine-containing antibiotics,
and modifications on neamine often result in a loss or
† Utah State University.
‡ Brigham Young University.
10.1021/ol0497685 CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/25/2004

decrease in antibacterial activity.1 Therefore, we reason that
ring III should be the optimal component where a glycodi-
versification approach could be employed leading to the
synthesis of novel kanamycin B analogues with, perhaps,
increased antibacterial activity.
Scheme 1a
Our laboratory has recently developed and reported general
methodologies for the preparation of a library of unusual
sugar donors that favor the formation of both R- and
â-glycosidic bonds.5 The R-glycosidic bond between rings
II and III is important, as kanamycin analogues with a
â-glycosidic bond manifest much weaker antibacterial activ-
ity.6 However, unlike the synthesis of â-glycosidic deriva-
tives, control of the stereoselective formation of an R-gly-
cosidic bond is challenging since no neighboring group
assistance can be utilized. Nevertheless, we have discovered
optimal conditions for making the R-glycosidic bond. In
combination with our previous synthetic work with unusual
sugar donors, we have developed a library of unusual sugar
donors that can satisfy needs in stereoselective incorporation
of carbohydrate moieties. Herein, we wish to report a much-
improved convenient and practical method for the preparation
of a library of kanamycin analogues.
a Conditions: (a) TfN3, CuSO4, H2O, CH2Cl2. (b) Cyclohexanone
dimethyl ketal, TsOH-H2O, CH3CN. (c) (i) Ac2O, Et3N, CH2Cl2,
DMAP; (ii) dioxane, H2O, HOAc.
soluble in organic solvents, thus allowing for smooth puri-
fication and characterization.7 Compound 5 was obtained
from 3 via selective protection of the O-5 and O-6 diols of
neamine with cyclohexanone dimethyl ketal, followed by
acetylation of the O-3′ and O-4′ diols, and deprotection of
the cyclohexylidene group. We were pleased to discover that
the neamine acceptor, 5, undergoes regiospecific glycosy-
lation at O-6 position, resulting in the desired 4,6-disubsti-
tuted 2-deoxystreptamine motif (Scheme 2).8 The optimal
stereoselectivity for the formation of an R-glycosidic bond
can be accomplished by operating the reaction in a solution
of Et2O and CH2Cl2 in a 3:1 ratio.9 A further increase in the
content of Et2O has no effect in the stereoselectivity;
however, decreasing the Et2O content results in lower
stereoselectivity. The glycosylated compounds were often
A library of phenylthioglucopyranose-based unusual sugar
donors was constructed with procedures analogous to those
in our previous work (Figure 2).5 These donors have benzyl
(3) (a) Tanaka, H.; Nishida, Y.; Furuta, Y.; Kobayashi, K. Bioorg. Med.
Chem. Lett. 2002, 12, 1723-1726. (b) Hanessian, S.; Tremblay, M.; Swayze,
E. E. Tetrahedron 2003, 59, 983-993. (c) Hanessian, S.; Kornienko, A.;
Swayze, E. E. Tetrahedron 2003, 59, 995-1007. (d) Tsuchiya, T.; Takagi,
Y.; Umezawa, S. Tetrahedron Lett. 1979, 51, 4951-4954. (e) Umezawa,
S.; Umezawa, H.; Okazaki, Y.; Tsuchiya, T. Bull. Chem. Soc. Jpn. 1972,
45, 3624-3628. (f) Umezawa, H.; Miyasaka, T.; Iwasawa, H.; Ikeda, D.;
Kondo, S. J. Antibiotics 1981, 34, 1635-1640. (g) Kumar, V.; Remers,
W. A. J. Med. Chem. 1979, 22, 432-436. (h) Kumar, V.; Jones, G. S.
Blacksberg, I.; Remers, W. A. J. Med. Chem. 1980, 23, 42-49. (i) Matsuno,
T.; Yoneta, T.; Fukatsu, S.; Umemura, E. Carbohydr. Res. 1982, 109, 271-
275. (j) Sharma, M. N.; Kumar, V.; Remers, W. A. J. Antibiotics 1982, 35,
905-910.
Figure 2. Unusual sugar donors for R-linked glycosidic bond.
(4) (a) Seeberger, P. H.; Baumann, M.; Zhang, G.; Kanemitsu, T.;
Swayze, E. E.; Hofstadler, S. A.; Griffey, R. H. Synlett 2003, 1323-1326.
(b) Chou, C.-H.; Wu, C.-S.; Chen, C.-H.; Lu, L.-D.; Kulkarni, S. S.; Wong,
C.-H.; Hung, S.-C. Org, Lett. 2004, 6, 585-588.
(5) (a) Elchert, B.; Li, J.; Wang, J.; Hui, Y.; Rai, R.; Ptak, R.; Ward, P.;
Takemoto, J. Y.; Bensaci, M.; Chang, C.-W. T. J. Org. Chem. 2004, 69,
1513-1523. (b) Wang, J.; Li, J.; Rai, R.; Chang, C.-W. T. Unpublished
result.
(6) Suami, T.; Nashiyama, S.; Ishikawa, Y.; Katsura, S. Carbohydr. Res.
1976, 52, 187-196.
(7) Alper, P. B.; Hung, S. C.; Wong, C.-H. Tetrahedron Lett. 1996, 37,
6029-6032.
(8) Similar regioselective glycosylation has been reported in ref 4b and:
(a) Tsuchiya, T.; Takahashi, Y.; Kobayashi, Y.; Umezawa, S. J. Antibiot.
1985, 38, 1287-1290. (b) Ding, Y.; Hofstadler, S. A.; Swayze, E. E.; Risen,
L.; Griffey, R. H. Angew. Chem., Int. Ed. 2003, 42, 3409-3412.
(9) Effect of ether in increasing the formation of R anomer has been
noted in Greenberg, W. A.; Priestley, E. S.; Sears, P. S.; Alper, P. B.;
Rosenbohm, C.; Hendrix, M.; Hung, S.-C.; Wong, C.-H. J. Am. Chem. Soc.
1999, 121, 6527-6541.
or azido groups at the C-2 position, which favor the formation
of an R-glycosidic bond under the influence of anomeric and
solvent effects.9 The neamine acceptor was prepared from
neamine (Scheme 1). The amino groups of neamine were
converted into azido groups producing 3, which is quite
(1) (a) Aminoglycoside Antibiotics; Hooper, I. R., Ed.; Springer-Verlag:
New York, 1982. (b) Haddad, J.; Kotra, L. P.; Mobashery, S. In
Glycochemistry Principles, Synthesis, and Applications; Wang, P. G.,
Bertozzi, C. R., Ed.; Marcel Dekker: New York, 2001; p 353.
(2) (a) Mingeot-Leclercq, M.-P.; Glupczynski, Y.; Tulkens, P. M.
Antimicrob. Agents Chemother. 1997, 43, 727-737. (b) Kotra, L. P.;
Haddad, J.; Mobashery, S. Antimicrob. Agents Chemother. 2000, 44, 3249-
3256. (c) Cohen, M. L. Science 2002, 257, 1050-1055. (d) Neu, H. C.
Science 2002, 257, 1064-1072.
1382
Org. Lett., Vol. 6, No. 9, 2004

Scheme 2
Scheme 3
a MIC: minimum inhibitory concentration. bThese values were obtained from molecular modeling calculation. The lower the values, the
better the binding affinity (ref 11).
mixed with inseparable impurities. Nevertheless, after hy-
drolysis of the acetyl groups, the triols can be afforded in
good purity and improved R:â ratio. The final products were
synthesized as chloride salts via the Staudinger reaction
followed by hydrogenation and ion-exchange (Scheme 3).
The kanamycin analogues were tested against Escherichia
coli (ATCC 25922), and Staphylococcus aureus (ATCC
25923) using kanamycin B as the control.10 We also carried
out molecular modeling for evaluating the binding of
kanamycin analogues toward the target site of 16S rRNA.
The minimum inhibitory concentrations of kanamycin ana-
(10) Methods for Dilution Antimicrobial Susceptibility Testing for
Bacteria that Grow Aerobically. Approved standard M7-A5, and Perfor-
mance Standards for Antimicrobial Disk Susceptibility Tests. Approved
standard M2-A7; National Committee for Clinical Laboratory Standards,
Wayne, PA.
Org. Lett., Vol. 6, No. 9, 2004
1383

logues were compared with the binding capability calculated
from molecular modeling.11 Similar to a reported study,9 we
also noticed that the binding capability of kanamycin
analogues cannot be correlated to their antibacterial activity.
Analogous to previous results from our laboratory,12 an
increase in the number of amino groups did not increase the
antibacterial activity, albeit the binding affinity was enhanced
(JL016 vs JL007). Compound JL007 is designed to be
identical to kanamycin B for ensuring the regioselective
6-fold increase. Since the targeted A-site of 16S rRNA is
highly conserved,2b the presence of an outer membrane in
E. coli (gram-negative) could be the major factor that con-
tributes to the observed variation in the increasing antibacte-
rial activity of kanamycin analogues against gram-negative
and gram-positive bacteria. Pyranmycin contains a 4,5-di-
substituted 2-deoxystreptamine core, while kanamycin con-
tains a 4,6-disubstituted 2-deoxystreptamine core. We believe
that the mechanisms involved in the importation of ami-
noglycoside antibiotics could have various efficiencies in
transporting structurally different aminoglycosides. Although
it is difficult to study such membrane-associated importation
mechanisms, our SAR and binding results suggest that more
research for the uptake of aminoglycosides is needed to
augment the binding affinity studies between aminoglyco-
sides and rRNA. The summary of the preliminary SAR is
shown in Figure 3.
1
glycosylation. Although both H and 13C NMR spectra of
JL007 are almost the same as those from commercially
available kanamycin B, the antibacterial activity of JL007
is about 2-4-fold lower than kanamycin B, which is
probably due to the presence of the 1′′-epimer of JL007.
Nevertheless, we discovered that the amino group at the
equatorial position of C-3′′ is superior in terms of activity
in comparison to C-6′′ (JL002) or the equatorial position of
C-4′′ (JL012). In fact, an amino group at the C-6′′ position
reduced the activity of some analogues (JL002 vs JL001
and JL016 vs JL007). Incorporating an amino group at the
axial position dramatically reduces the activity (JL012 vs
JL018 and JL003 vs JL014), while an axial OH at C-4′′
increases the activity (JL004 vs JL002). Further introduction
of a larger functional group at the C-4′′ position such as
galactose in JL019 completely abolished the antibacterial
activity. The amino group at the C-2′′ position may produce
better antibacterial activity. However, it is difficult to offer
accurate evaluation due to the lower stereoselectivity in
glycosylation (R: â ) 2.5:1). Interestingly, JL005 with a CH3
group at the C-6′′ position showed good activity among the
analogues we prepared except for JL007, although its binding
score is lower in ranking. Molecular modeling of JL005
binding to the targeted RNA site did not reveal any plausible
interaction that may explain such an unusual observation.13
We have also tested these analogues against resistant strains
of bacteria.14 No significant activity was observed.
Figure 3. Summary of the SAR of ring III of kanamycin B
analogues.
In conclusion, we have independently developed a con-
venient methodology for the construction of a library of
kanamycin analogues. In contrast to our previous experience
in the pyranmycin library, the kanamycin analogues manifest
several leads (for example, JL007 and JL005). With the
revealed SAR, we are currently focusing on the introduction
of more modifications with the goal of increasing activity
against resistant bacteria.
Unlike the structure activity relationship (SAR) of ring
III of pyranmycins from our previous report,12 the observed
SAR from these kanamycin analogues is not completely
consistent between the two different strains of bacteria that
we tested. For example, JL003 has similar activity for both
strains, while JL016 manifests a 10-fold increase in activity
against S. aureus. On the other hand, JL007 shows only a
Acknowledgment. We acknowledge Utah State Univer-
sity (CURI Grant), National Foundation for Infectious
Diseases, and Frontier Scientific, Inc., for generous financial
support. We also acknowledge Professor Paul Savage from
Department of Chemistry and Biochemistry, Brigham Young
University, for re-examining our MIC data. We also thank
Prof. Mobashery from Notre Dame University for the
generous gift of the pTZ19-3 and pSF815 plasmids.
(11) Score function was based on Amber 96 force-field as implemented
in HyperChem 7.0 and the solvent-accessible surface methodology to
account for the hydration effects. The program was developed for drug
design, not specifically for the calculation of absolute binding affinity.
Therefore, we look for the tendency in binding rather than the meaning
interpreted from the absolute numbers of binding. Nevertheless, the
calculated binding scores have tendencies consistent with the compounds
that have known binding affinity from the in vitro experiment. More details
can be found in Supporting Information.
(12) Chang, C.-W. T.; Hui, Y.; Elchert, B.; Wang, J.; Li, J.; Rai, R.
Org. Lett. 2002, 4, 4603-4606.
(13) Please refer to Supporting Information for figures from molecular
modeling.
Supporting Information Available: Experimental pro-
1
cedures for the preparation of compounds and H and 13C
spectra of the selected compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(14) E. coli (TG1) (APH3′-I) and E. coli (TG1) (AAC6′/APH2′′).
OL0497685
1384
Org. Lett., Vol. 6, No. 9, 2004