Site-selective displacement of tobramycin hydroxyls for preparation of antimicrobial cationic amphiphiles

Article abstract of DOI:10.1021/ol4030138

A short site-selective strategy for the activation and derivatization of alcohols of the clinically important aminoglycoside tobramycin is reported. The choice of amine protecting group affected the site-selective conversion of secondary alcohols of tobramycin into leaving groups. Temperature-dependent, chemoselective sequential nucleophilic displacements resulted in hetero- and homodithioether tobramycin-based cationic amphiphiles that demonstrated marked antimicrobial activity and impressive membrane selectivity.

Full text of DOI:10.1021/ol4030138

ORGANIC
LETTERS
2013
Vol. 15, No. 24
6144–6147
Site-Selective Displacement of
Tobramycin Hydroxyls for Preparation
of Antimicrobial Cationic Amphiphiles
Yifat Berkov-Zrihen, Ido M. Herzog, Mark Feldman, and Micha Fridman*
School of Chemistry, Raymond and Beverley Sackler Faculty of Exact Sciences, Tel Aviv
University, Tel Aviv 69978, Israel
mfridman@post.tau.ac.il
Received October 21, 2013
ABSTRACT
A short site-selective strategy for the activation and derivatization of alcohols of the clinically important aminoglycoside tobramycin is reported.
The choice of amine protecting group affected the site-selective conversion of secondary alcohols of tobramycin into leaving groups.
Temperature-dependent, chemoselective sequential nucleophilic displacements resulted in hetero- and homodithioether tobramycin-based
cationic amphiphiles that demonstrated marked antimicrobial activity and impressive membrane selectivity.
Site-selectivederivatization ofbiologicallyactivenatural
products is one of the most frequently used strategies
for the development of novel drugs.1ꢀ3 This strategy is
especially appealing for creation of analogues of com-
mercially available aminoglycoside (AG) antibiotics, a
large family of pseudo-oligosaccharides that contain
multiple amines, and hydroxyls with similar chemical
properties. AG antibiotics bind to the 16S rRNA, a
component of the 30S small subunit of the bacterial
ribosome and interfere with codon recognition during
the translation process.4 Unfortunately, clinical use of
AGs is limited due to their toxic side effects and bacterial
resistance.4ꢀ6 As a result, the development of novel AGs
with improved clinical properties has been pursued
for several decades, yet semisynthetic modifications of
natural AGs have led to the discovery of very few
clinically relevant antibiotics.7ꢀ9
Previous studies demonstrated that in addition to their
antimicrobial activity, AGs and their synthetic analogues
have additional biological activities, further emphasizing
the need for the development of methods for site-selective
modifications in search of improved biological perfor-
mance. Additional reported biological activities of AGs
include the inhibition of several RNA-catalyzed processes,
enhancement of eukaryotic ribosome read-through of
premature stop-codon mutations, and the disruption of
bacterial membranes.10ꢀ16
(7) Kawaguchi, H.; Naito, T. J. Antibiot. 1972, XXV, 695–708.
(8) Aggen, J. B.; Armstrong, E. S.; Goldblum, A.; Dozzo, P.; Linsell,
M. S.; Gliedt, M. J.; Hildebrandt, D. J.; Feeney, L. A.; Kubo, A.; Matias,
R. D.; Lopez, S.; Gomez, M.; Wlasichuk, K. B.; Diokno, R.; Miller, G. H.;
Moser, H. E. Antimicrob. Agents Chemother. 2010, 54, 4636–4642.
(9) Werz, D. B.; Seeberger, P. H. Chem.;Eur. J. 2005, 11, 3194–3206.
(10) Wang, H.; Tor, Y. J. Am. Chem. Soc. 1997, 119, 8734–8735.
(11) Belousoff, M. J.; Graham, B.; Spiccia, L.; Tor, Y. Org. Biomol.
Chem. 2009, 7, 30–33.
(1) Wilcock, B. C.; Uno, B. E.; Bromann, G. L.; Clark, M. J.;
Anderson, T. M.; Burke, M. D. Nat. Chem. 2012, 4, 996–1003.
(2) Fowler, B. S.; Laemmerhold, K. M.; Miller, S. J. J. Am. Chem.
Soc. 2012, 134, 9755–9761.
(3) Pathak, T. P.; Miller, S. J. J. Am. Chem. Soc. 2012, 134, 6120–6123.
(4) Houghton, J. L.; Green, K. D.; Chen, W.; Garneau-Tsodikova, S.
ChemBioChem 2010, 11, 880–902.
(12) Nudelman, I.;Rebibo-Sabbah, A.;Shallom-Shezifi,D.;Hainrichson,
M.; Stahl, I.; Ben-Yosef, T.; Baasov, T. Bioorg. Med. Chem. 2006, 16, 6310–
6315.
(5) Kandasamy, J.; Atia-Glikin, D.; Shulman, E.; Shapira, K.; Shavit,
M.; Belakhov, V.; Baasov, T. J. Med. Chem. 2012, 55, 10630–10643.
(6) Guthrie, O. W. Toxicology 2008, 249, 91–96.
(13) Rowe, S. M.; Sloane, P.; Tang, L. P.; Backer, K.; Mazur, M.;
Buckley-Lanier, J.; Nudelman, I.; Belakhov, V.; Bebok, Z.; Schwiebert,
E.; Baasov, T.; Bedwell, D. M. J. Mol. Med. 2011, 89, 1149–1161.
r
10.1021/ol4030138
Published on Web 11/13/2013
2013 American Chemical Society

We previously reported that attachment of a linear alkyl
chain to the 600 position of the AG tobramycin (1) results in
cationic amphiphiles with potent and broad spectrum
antimicrobial activity (Figure 1).15,17 The semisynthetic
route for the preparation of the cationic amphiphiles
derived from 1 relied on the chemoselective conversion
of the 600-primary alcohol of this AG to a leaving group
followed by a nucleophilic displacement with aliphatic
chain primary thiols resulting in 600-thioetherification
of 1.
with 15 equiv of TIBS-Cl for 18 h, 56% of the isolated
products was the 600-O-TIBS derivative 7, 9% was the
40, 600-di-O-TIBS derivative 8, and 27% was the 200, 600-
di-O-TIBS derivative 9 (Scheme 1B).
The identity of the alcohols that were transformed into
the corresponding O-TIBS leaving groups in each isolated
product was confirmed by 1D-TOCSY NMR experi-
ments. The mixtures containing the mono-600-O-TIBS
and di-O-TIBS products were readily separable by flash
chromatography, and the reactions were reproducible at
milligram to gram scales. As the 200,600-di-O-TIBS com-
pound 5 was obtained only as the minor product (8%
isolated yield) when 2 was used as the starting material,
sufficient amounts of this compound were obtained in two
synthetic steps from compound 9 in 80% yield for the two
steps (Scheme 1B).
To explain the observed product distribution, our atten-
tion was drawn to a previous study that reported that
NMR chemical shifts of protons ipso to azides of per-
azide-protected AGs may be correlated to the relative
electron density of the corresponding azide groups.19
We observed a similar correlation for the protons ipso to
the 200 and 40 secondary alcohols of compounds 2 and 6
(Table 1).
Figure 1. Structure of the clinically used AG antibiotic tobramycin
and its 600-thioether-based antimicrobial cationic amphiphiles.
Table 1. Chemical Shifts of the Ipso-Protons of the Secondary
Alcohols of Tobramycin Derivatives 2 and 6a
In order to gain access to additional families of tobra-
mycin-based cationic amphiphiles, we focused on the
development of a site-selective method for conversion of
one of the four secondary alcohols of 1 to a leaving group
to facilitate the preparation of tobramycin derivatives
through SN2 nucleophilic displacement reactions. Accord-
ing to a previously reported procedure when penta-NH-
Boc tobramycin derivative 2 (Scheme 1) was treated with 7
equiv of 2,4,6-triisopropylbenzenesulfonyl chloride (TIBS-
Cl) in pyridine at ambient temperature, the corresponding
600-O-TIBS derivative 3 (Scheme 1A) was formed in 67%
isolated yield.18 To convert one of the secondary alcohols
of compound 2 to the O-TIBS group, we performed the
reaction using an excess of TIBS-Cl (30 equiv) at ambient
temperature. Under these conditions, in addition to
compound 3 (31% isolated yield), the 40, 600-di-O-TIBS
derivative 4 was isolated in 34% yield and the 200, 600-di-
O-TIBS derivative 5 was isolated in 8% yield (Scheme 1A).
The site-selectivity of the reaction was found to be
dependent on the amine protecting groups. When the
penta-azido-tobramycin derivative 619 was treated
compd
H-40
H-200
H-5
H-400
2
6
3.43
3.63
3.69
3.46
3.61
3.40
3.37
3.37
a 400 or 500 MHz 1H NMR (CD3OD), chemical shifts in parts per
million (ppm).
Of the 200 and 40 secondary alcohols, the one with the
more upfield ipso-proton resonance preferentially reacted
with TIBS-Cl (Table 1). The H-40 of the Boc-protected 2
resonated at 3.43 ppm, whereas H-200 resonated at 3.69
ppm. An opposite order of chemical shifts was observed in
the caseof the azide-protected6; H-40 of 6 resonated at 3.63
ppm, whereas its H-200 resonated at 3.46 ppm. These results
suggest that the electron density, and therefore the nucleo-
philicity, of the 40-alcohol was higher than that of the 200-
alcohol in the case of the NH-Boc protected compound 2,
whereas the opposite effect was observed in the case of the
azide-protected6. Although the chemicalshiftsof theipso-
protons of the 200 and 40-alcohols of compounds 2 and 6
were correlated with their relative nucleophilicities, there
was no correlation for alcohols at the 5 and 400 positions of
these compounds (Table 1). The fact that no 5-O-TIBS
product was observed was attributed to the known high
level of steric hindrance of the C-5 alcohol of tobramycin 1
rather than to an electronic effect. This alcohol is located
between the C-4 and the C-6 positions that are substituted
by sugar rings II and III, respectively (Figure 1). The utility
of the steric hindrance of C-5 of 1 was previously demon-
strated: when penta NH-Cbz-protected tobramycin was
ꢀ
(14) Baussanne, I.; Bussiere, A.; Halder, S.; Ganem-Elbaz, C.; Ouberai,
M.; Riou, M.; Paris, J.-M.; Ennifar, E.; Mingeot-Leclercq, M.-P.; Decout,
ꢁ
J.-L. J. Med. Chem. 2010, 53, 119–127.
(15) Herzog, I. M.; Green, K. D.; Berkov-Zrihen, Y.; Feldman, M.;
Vidavski, R. R.; Eldar-Boock, A.; Satchi-Fainaro, R.; Eldar, A.; Garneau-
Tsodikova, S.; Fridman, M. Angew. Chem., Int. Ed. 2012, 5652–5656.
(16) Bera, S.; Zhanel, G. G.; Schweizer, F. J. Med. Chem. 2010, 53,
3626–3631.
(17) Herzog, I. M.; Feldman, M.; Eldar-Boock, A.; Satchi-Fainaro,
R.; Fridman, M. MedChemComm 2013, 4, 120–124.
(18) Michael, K.; Wang, H.; Tor, Y. Bioorg. Med. Chem. 1999, 7,
1361–1371.
(19) Nyffeler, P. T.; Liang, C.-H.; Koeller, K. M.; Wong, C.-H.
J. Am. Chem. Soc. 2002, 124, 10773–10778.
Org. Lett., Vol. 15, No. 24, 2013
6145

Scheme 1. Site-Selective Modification of Alcohols of Tobramycin Derivatives 2 and 6 and the synthesis of Homo- and
Heterodithioether Tobramycin-Derived Antimicrobial Cationic Amphiphiles
reacted with an excess of acetic anhydride, only the C-5
alcohol was not acetylated.20
cationic amphiphiles. Nucleophilic displacement of the
two O-TIBS groups of 4 by n-alkyl-thiols at 75 °C in
DMF resulted in a set of 40,600-homodithioether derivatives
of tobramycin (10, 11, and 12, Scheme 1C) in 68ꢀ73%
isolated yields. Acidic removal of the Boc groups gave the
40,600-dithioether tobramycin-derived cationic amphiphiles
13ꢀ15 as TFA salts in quantitative yields.
Finally, the C-40 alcohol of ring II of 1 is less hindered
than the ring III C-400 alcohol as the former neighbors the
C-30 deoxy position (Figure 1). This fact may explain why
40-O-TIBS and no 400-O-TIBS products were isolated from
the reactions with TIBS-Cl. The 1H chemical shifts depend
upon many factors; however, the observed differences in
ipso-proton chemical shifts offer an explanation for the
observed differences in product ratios between the 200 and
40-alcohols of compounds 2 and 6.
Selective thioetherification of the 600-O-TIBS group of
compound 4 was readily achieved at ambient temperature,
and the 600-thioether compounds 16 and 17 were generated
(Scheme 1C). When 16 and 17 were reacted with n-alkyl-
thiols, in DMF at 75 °C, the heterosubstituted 40,600-
dithioether tobramycin derivatives 18, 19, and 20 were
obtained; these compounds were treated with TFA to yield
compounds 21, 22, and 23 in quantitative yields.
Compounds 4 and 5 were used for the preparation of a
set of homo- and heterodithioetherꢀtobramycin-based
(20) Hanessian, S.; Tremblay, M.; Swayze, E. E. Tetrahedron 2003,
59, 983–993.
6146
Org. Lett., Vol. 15, No. 24, 2013

Table 2. Minimal Inhibitory Concentration (MIC) and Hemolysis Values
MIC values (μg/mL)
bacteriaa
1
13
14
15
21
22
23
27
28
29
A
16
16
32
4
4
8
4
4
4
8
4
>32
2
4
32
>32
8
32
>32
16
8
>32
4
B
>32
>32
4
C
0.3
2
4
2
D
>32
16
8
4
8
4
4
4
32
>32
8
32
8
E
>32
32
8
>32
>32
4
>32
32
1
8
4
4
32
8
F
32
32
4
>32
2
>32
2
16
4
G
>32
8
8
16
4
H
4
4
4
2
2
2
4
16
8
hemolysisb
128 μg/mL
0
29.9 ( 1.6
94.9 ( 3.7
88.2 ( 3.2 15.8 ( 3.2
96.5 ( 2.5 100 ( 3.2
10.9 ( 1.0
45.4 ( 0.5
97.7 ( 2.1
a Bacterial strains:(A) S. aureus Oxford, NCTC6571; (B) MRSA; (C) S. epidermidis ATCC 12228 (biofilm negative);(D) S. epidermidisATCC 35984/
RP62A (biofilm positive); (E) S. aureus Cowan, ATCC 12598; (F) S. pyogenes M1T1; (G) S. pyogenes JRS75; (H) S. pyogenes T5. b The results are
expressed as percentage of hemoglobin released relative to the positive control (sample treated with Triton X100).
According to Richardson’s rules, nucleophilic displace-
ment of 2-aryl sulfonate in an R- pyranoside sugar ring such
as the 200-O-TIBS group of ring III of compound 5 is dis-
favored due to unfavorable alignment of dipoles in the
transition state of the reaction.21 It was not that surprising,
therefore, that treatment of 5 with n-alkyl-thiols and Cs2CO3
in DMF resulted in compounds 24, 25, and 26 which, after
deprotection, gave the final compounds 27, 28, and 29
(Scheme 1D). 500 MHz 1H NMR analysis in D2O revealed
that in these three compounds C-300 and C-600 were substituted
by thioether groups and C-200 was substituted by an NH-Boc
group or by an amine in the R-D-altro-configuration. This
can be rationalized by an intramolecular nucleophilic attack
of the C-300 NH-Boc of 5 on its C-200 O-TIBS group to form a
HoughꢀRichardson-type aziridine (Scheme 1E).22,23 The
aziridine then underwent a thiolate-mediated ring-opening in
FurstꢀPlattner fashion at C-300 to yield the products with R-D-
altro-configuration in accordance with Richardson’s rules.24
The NH-Boc group at C-300 of compound 5 was found to
be essential for this reaction to proceed via the aziridine
intermediate. When the 200,600-di-O-TIBS penta-azido to-
bramycin derivative 9 was treated with NaN3 and crown-5
in DMF, the 200-O-TIBS of 9 was not displaced even at
100 °C and only decomposition of this compound was
therefore minimizing disfavored 1,3-diaxial interactions
4
that would take place if ring III adopted a C1 chair
conformation.
The antimicrobial activity was tested on eight bacterial
strains (Table 2).25 Although none of the 300,600-dithioether
compounds 27ꢀ29 were effective against the tested methi-
cillin-resistant Staphylococcus aureus strain (strain B), the
40,600-dithioether compounds 14, 15, and 21 were potent
against this pathogen. On the other hand, none ofthe 40,600-
dithioether derivatives (13ꢀ15 and 21ꢀ23) had activity
against Streptococcus pyogenes M1T1 (strain F); however,
the 300,600-dithioether compounds 27ꢀ29 exhibited moderate
to good activity against this pathogen. Of the nine tested
antimicrobial cationic amphiphiles, compound 21 caused
both low percentage of hemolysis (Table 2) and demon-
strated potent antimicrobial activity against seven of the
eight bacterial strains tested.
In conclusion, we report a method for site selective
nucleophilic displacement of secondary alcohols of tobra-
mycin. Site-selectivity was affected by the amine protecting
groups used and facilitated the generation of homo- and
heterodithioether tobramycin-derived cationic amphi-
philes. By altering the position and length of the aliphatic
thioether chains it was possible to develop potent anti-
microbial cationic amphiphiles that exhibited impressive
selectivity for bacterial compared to red blood cell mem-
branes. Thereported methodologycan be further extended
by using other nucleophiles.
1
observed. Ring III adopted a C4 chair conformation in
compounds 27, 28, and 29 as was indicated by 1H NMR
00
00
(J1 ꢀ2 = 8.0 Hz indicating trans diaxial relationships
1
between H-100 and H-200). In a C4 chair conformation, the
C-100, C-200, and C-300 hindered substituents are equatorial,
Acknowledgment. This work was supported by the
FP7-PEOPLE-2009-RG Marie Curie Action: Reintegra-
tion Grants (Grant No. 246673). We thank Prof. Doron
Steinberg (The Hebrew University of Jerusalem) for the
gift of bacterial strains.
(21) (a) Richardson, A. C. Carbohydr. Res. 1969, 10, 395–402. (b) For
a more recent open access article covering A. C. Richardson’s life and
work including the rules, see: Hale, K. J. Adv. Carbohydr. Chem.
Biochem. 2011, 66, 2.
(22) Gibbs, C. F.; Hough, L.; Richardson, A. C. Carbohydr. Res.
1965, 1, 290–296.
(23) Buss, D. H.; Hough, L.; Richardson, A. C. J. Chem. Soc. 1965,
2736–2743.
(24) We thank one of the reviewers for helping us rationalize the
mechanistic outcomes.
(25) Berkov-Zrihen, Y.; Herzog, I. M.; Feldman, M.; Sonn-Segev,
A.; Roichman, Y.; Fridman, M. Bioorg. Med. Chem. 2013, 21, 3624–
3631.
Supporting Information Available. Synthetic procedures
and 1H and 13C NMR spectra. This material is available free
of charge via the Internet at http://pubs.acs.org.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 24, 2013
6147