Synthesis of a protected enantiomerically pure 2-deoxystreptamine derivative from D-allylglycine

Article abstract of DOI:10.1016/j.tetlet.2004.03.051

A diastereoselective synthetic route from D-allylglycine to the enantiopure (protected) 2-deoxystreptamine derivative 14 is presented. Key steps involve two consecutive chain extensions - with crucial stereodirective roles for the amino protective groups, ring closure by olefin metathesis, face selective dihydroxylation, cyclic sulfate formation and finally opening with azide. The resulting 2-deoxystreptamine derivative is ideally protected for the preparation of 4,5- or 4,6-linked aminoglycoside antibiotics.

Full text of DOI:10.1016/j.tetlet.2004.03.051

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 3629–3632
Synthesis of a protected enantiomerically pure 2-deoxystreptamine
derivative from -allylglycine
D
Guuske F. Busscher, Floris P. J. T. Rutjes and Floris L. van Delft*
Department of Organic Chemistry, NSRIM, University of Nijmegen, Toernooiveld, 6525 ED Nijmegen, The Netherlands
Received 14 August 2003; revised 5 March 2004; accepted 10 March 2004
Abstract—A diastereoselective synthetic route from D-allylglycine to the enantiopure (protected) 2-deoxystreptamine derivative 14 is
presented. Key steps involve two consecutive chain extensions––with crucial stereodirective roles for the amino protective groups,
ring closure by olefin metathesis, face selective dihydroxylation, cyclic sulfate formation and finally opening with azide. The resulting
2-deoxystreptamine derivative is ideally protected for the preparation of 4,5- or 4,6-linked aminoglycoside antibiotics.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Since the original discovery of streptomycin in 1944, the
family of aminoglycosides has grown steadily into a
powerful class of antibiotics with a broad antibacterial
spectrum and proven efficacy, particularly in combina-
tion with other drugs.1 Nevertheless, extensive use of the
aminoglycosides is limited due to the associated toxici-
ties, most notably nephrotoxicity and ototoxicity, and to
a lesser extent neuromuscular blockade.2 Another, more
alarming drawback of the aminoglycosides is the global
development of microbial resistance with the most
common mechanism being structural modification by
bacterial resistance enzymes. These circumstances
necessitate the development of new and innovative
aminoglycoside antibiotics and several reports on
the derivatisation of existing aminoglycosides have
appeared in the literature in recent years.3 To be fully
flexible in the design and preparation of novel amino-
glycoside-type structures, however, de novo synthesis
from individually prepared components is required.
Consequently, our research has focused on the synthesis
of enantiomerically pure derivatives of 2-deoxystrept-
amine, an aminocyclitol ring that constitutes the core
structure of the majority of clinically useful aminogly-
cosides (Fig. 1) and, as may well be speculated, is crucial
for binding of the aminoglycosides to their target A-site
RNA.4
Although several synthetic routes towards 2-deoxy-
streptamine are known in the literature, all require many
synthetic steps and offer minimal flexibility in protective
groups.5 Moreover, only two of these routes afford an
asymmetric analogue of 2-deoxystreptamine, starting
from either
D D
-mannose6 or -glucose.7 The most prac-
tical method to obtain this aminocyclitol moiety is by
degradation of neomycin,8–10 but the ‘naked’ meso-com-
pound thus obtained still requires desymmetrisation as
well as protective group manipulations before incorpo-
ration into new aminoglycoside entities can be ensured.
In contrast, we here wish to report a synthesis of an
orthogonally protected, enantiopure 2-deoxystreptamine
derivative, which is highly suitable to serve as a scaffold
for new aminoglycoside entities, either 4,5- or 4,6-linked.
Our synthesis starts from enantiomerically pure
glycine, a non-proteinogenic amino acid that is readily
available in our group.11 Thus, the methyl ester of
D-allyl-
D
-
allylglycine12 was subjected to conditions explored by
Dondoni,13 involving introduction of thiazole as
a masked aldehyde. Our first attempt to obtain the
thiazolyl b-amino alcohol 2a (Scheme 1) started with
Boc-protected allylglycine methyl ester 1a, which
upon partial reduction and addition of 2-(trimethyl-
silyl)thiazole (2-TST) gave the syn-amino alcohol 2a.
Not unexpectedly, however, alcohol 2a was obtained13
with a de of only 70%, which led us to follow a slightly
different procedure involving double protection of the
amino function (Boc, PMB) as in 1b and reversal of the
sequence of events, that is, first reaction with 2-lithio-
thiazole (2-LTT) and then NaBH4 reduction of the
resulting ketone. To our satisfaction, the syn-b-amino
Keywords: Aminoglycosides; 2-Deoxystreptamine; D-Allylglycine; Syn-
thesis.
* Corresponding author. Tel.: +31-(0)24-3652373; fax: +31-(0)24-365-
3393; e-mail: fvd@sci.kun.nl
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.03.051

3630
G. F. Busscher et al. / Tetrahedron Letters 45 (2004) 3629–3632
Figure 1. Representative aminoglycoside antibiotics, both 4,5- and 4,6-linked to 2-deoxystreptamine.
Scheme 1. Reagents and conditions: (i) AcCl, MeOH, 2 h, D then Boc2O, dioxane, rt, 18 h; (ii) AcCl, MeOH, D, 2 h then Et3N, p-MeOC6H4CHO,
CH2Cl2, rt, 18 h then NaBH4, MeOH, 30 min, 0 ꢁC then Boc2O, dioxane, rt, 18 h (89%); (iii) DIBAL-H, toluene, )78 ꢁC, 2 h, 2-TST, CH2Cl2, )20 ꢁC,
48 h then Bu4NFÆ3H2O, THF, rt, 1 h (70% de, 70%); (iv) 2-LTT, Et2O, )78 to )45 ꢁC, 6 h then NaBH4, MeOH, 0 ꢁC, 30 min (89%, de > 95%);
(v) NaH, DMF, 0 ꢁC (BnBr).
alcohol 2b was now obtained with a de of >95%.13 Next,
benzyl protection of the free hydroxyl was investigated
but it was found that standard conditions (sodium
hydride followed by benzyl bromide) led to 2-oxazolidi-
none 3 (Scheme 1). However, inverting the order of
addition of NaH and BnBr cleanly gave the O-benzy-
lated derivative 4 (Scheme 2), which was subjected to the
thiazole deblocking protocol13 followed by condensation
of the resulting aldehyde 5 with vinylmagnesium bro-
mide. Unfortunately, the nucleophilic chain extension
ꢀ
Scheme 2. Reagents and conditions: (i) BnBr, DMF, 0 ꢁC then NaH, 1 h (90%); (ii) 4 A MS, MeOTf, MeCN, rt, 25 min then NaBH4, MeOH, 0 ꢁC,
10 min then CuCl2Æ2H2O, CuO, MeCN/H2O, 10/1, rt, 15 min (5: 78%, 8: 83%); (iii) vinyl MgBr, THF, )78 ꢁC, 4 h (6a: 33%, 9a: 56%); (iv) CAN,
NaHCO3, MeCN/H2O (4/1), rt, 30 min (67%); (v) second generation Grubbs’ catalyst (mono-substituted imidazolylidene ligand), CH2Cl2, rt, 2 h; (vi)
TBDMSOTf, Et3N, CH2Cl2, rt, 2 h (two steps 82%); (vii) RuCl3, NaIO4, EtOAc/MeCN/H2O (3/3/1), 0 ꢁC, 3 min (70%); (viii) SOCl2, pyridine,
EtOAc, 0 ꢁC, 30 min then NaIO4, RuCl3ÆH2O, CH2Cl2/MeCN/H2O (2/2/3), 0 ꢁC, 1 h (80%); (ix) LiN3, DMF, D, 4 h then H2SO4, THF, H2O, rt,
30 min (60%).

G. F. Busscher et al. / Tetrahedron Letters 45 (2004) 3629–3632
3631
Hendrix, M.; Hung, S.-C.; Wong, C.-H. J. Am. Chem. Soc.
1999, 121, 6527–6541; (b) Sucheck, S. J.; Wong, A. L.;
Koeller, K. M.; Boehr, D. D.; Draker, K.; Sears, P.;
Wright, G. D.; Wong, C.-H. J. Am. Chem. Soc. 2000, 122,
5230–5231; (c) Luedtke, N. W.; Baker, T. J.; Goodman,
M.; Tor, Y. J. Am. Chem. Soc. 2000, 122, 12035–12036; (d)
Rosenbohm, C.; Berghe, D. V.; Vlietinck, A.; Wengel, J.
Tetrahedron 2001, 57, 6277–6287; (e) Tok, J. B.-H.; Fenker,
J. Bioorg. Med. Chem. Lett. 2001, 11, 2987–2991; (f)
Charles, I.; Xue, L.; Arya, D. P. Bioorg. Med. Chem. Lett.
2002, 12, 1259–1262; (g) Vourloumis, D.; Takahashi, M.;
Winters, G. C.; Simonsen, K. B.; Ayida, B. K.; Barluenga,
S.; Qamar, S.; Shandrick, S.; Zhao, Q.; Hermann, T.
Bioorg. Med. Chem. Lett. 2002, 12, 3367–3372; (h) Ryu, D.
H.; Tan, C. H.; Rando, R. R. Bioorg. Med. Chem. Lett.
2003, 13, 901–903; (i) Russell, R. J. M.; Murray, J. B.;
Lentzen, G.; Haddad, J.; Mobashery, S. J. Am. Chem. Soc.
2003, 125, 3410–3411; (j) Hanessian, S.; Tremblay, M.;
Swayze, E. E. Tetrahedron 2003, 59, 983–993; (k) Seeber-
ger, P. H.; Baumann, M.; Zhang, G.; Kanemitsu, T.;
Swayze, E. E.; Hofstadler, S. A.; Griffey, R. H. Synlett
2003, 9, 1323–1326.
was characterised by low stereoselectivity revealing a 1:1
mixture of syn- and anti-diastereomers 6a and 6b, which
did not improve upon varying the reaction conditions
(temperature, solvents, additives). A crucial role of the
amino protection was again suspected and therefore it
was decided to remove the p-methoxybenzyl group of 4
(fi7) with CAN prior to Grignard addition.14 Gratify-
ingly, addition of vinylmagnesium bromide after thiazole
unmasking (7 fi 8) now led to a much improved syn/anti
ratio of 4:1 of compounds 9a and 9b (Scheme 2).15 After
silica gel separation of the diastereomers, ring-closing
metathesis proceeded smoothly and the free hydroxyl of
the cyclic product 1016 was protected with a TBDMS
group to afford 11 in 82% yield (two steps). Despite the
presence of the bulky silyl group, however, epoxidation
of compound 11 with several reagents, such as m-CPBA,
oxone and in situ formed dioxirane17 led in all cases to
inseparable mixtures of diastereomers varying from 1:1
to 3:1 (not depicted). Matters became more complicated
when we tried to open the diastereomeric epoxides with
sodium azide, leading to the formation of four isomeric
azido alcohols. Therefore, it was decided to prepare
cyclic sulfate 13. The reactivity of cyclic sulfates and
epoxides towards nucleophiles is similar in nature but
differs in selectivity.18 Another advantage is that the ring
opening of five-membered cyclic sulfates proceeds much
faster than with epoxides probably due to the better
leaving group ability.19 Thus, the double bond of 11 was
dihydroxylated20 (11 fi 12),15 which occurred with
exclusive facial selectivity, followed by reaction with
thionyl chloride and oxidation to form the cyclic sulfate
13 in 80% yield.21 Much to our satisfaction, opening of
the cyclic sulfate with lithium azide proceeded com-
pletely regioselectively to give, after subsequent sulfate
hydrolysis, the protected 2-deoxystreptamine 14 in
enantiomerically pure form.15 The latter compound,
after glycosylation of the free hydroxyl, is ideally suited
for the preparation of either 4,5- or 4,6-linked amino-
glycoside analogues after subsequent desilylation or
debenzylation, respectively.
4. Kaul, M.; Barbieri, C. M.; Kerrigan, J. E.; Pilch, D. S. J.
Mol. Biol. 2003, 326, 1373–1387.
5. 2-Deoxystreptamine syntheses: (a) Nakajima, M.; Haseg-
awa, A.; Kurihara, N. Justus Liebigs Ann. Chem. 1965,
689, 235–242; (b) Suami, T.; Lichtenthaler, F. W.; Ogawa,
S.; Nakashima, Y.; Sano, H. Bull. Chem. Soc. Jpn. 1967,
40, 1014–1017; (c) Dijkstra, D. Recl. Trav. Chim. Pays-
Bas. 1968, 87, 161–180; (d) Ogawa, S.; Ueda, T.; Funaki,
Y.; Hongo, Y.; Kasuga, A.; Suami, T. J. Org. Chem. 1977,
€
42, 3083–3088; (e) Kuhlmeyer, R.; Keller, R.; Schwesinger,
R.; Netscher, T.; Fritz, H.; Prinzbach, H. Chem. Ber. 1984,
117, 1765–1800.
6. Baer, H. H.; Arai, I.; Radatus, B.; Rodwell, J.; Chinh, N.
Can. J. Chem. 1987, 65, 1443–1451.
7. da Silva, E. T.; Le Hyaric, M.; Machado, A. S.; Almeida,
M. V. Tetrahedron Lett. 1998, 39, 6659–6662.
8. Canas-Rodriguez, A.; Ruiz-Poveda, S. G. Carbohydr. Res.
1977, 59, 240–243.
9. Tona, R.; Bertolini, R.; Hunziker, J. Org. Lett. 2000, 2,
1693–1696.
10. Swayze, E.; Griffey, R.; Ding, Y.; Mohan, V. Patent
Application WO 01/39726A2, 2001.
11. Wolf, L.; Sonke, T.; Tjen, K. C. M. F.; Kaptein, B.;
Broxterman, Q. B.; Schoemaker, H. E.; Rutjes, F. P. J. T.
Adv. Synth. Catal. 2001, 343, 662–674.
In conclusion, we believe that the synthesis described
above is a versatile route towards orthogonally pro-
tected 2-deoxystreptamine in enantiomerically pure
form in 14 steps and an overall yield of 6.1%. The
aminocyclitol building block obtained is suitable for
incorporation into new aminoglycoside entities. Work
along this line is currently in progress in our laboratory.
12. Provided by Chiralix, Nijmegen, The Netherlands.
13. Configuration of the syn-amino alcohol 2b was assigned
1
by H NMR analysis of the corresponding oxazolidinone
3. See, for example: Dondoni, A.; Perrone, D. Synthesis
1993, 1162–1176.
14. Reaction was performed up to 5 mmol scale. Buffering of
the CAN solution with NaHCO3 was required to prevent a
drop in yield.
20
Acknowledgement
15. Compound identification: 9a: ½aꢀ +27.3 (c 1.89, CH2Cl2).
D
1H NMR (CDCl3, 400 MHz, ppm): d 7.48–7.25 (m, 5H,
arom), 6.19–6.40 (m, 1H, @CH–), 5.82–5.62 (m, 1H,
@CH–), 5.72 (d, 1H, J ¼ 10:5 Hz, @CH2), 5.12 (d, 1H,
J ¼ 17:3 Hz, @CH2), 5.11–5.02 (m, 2H, @CH2), 4.77 (d,
1H, J ¼ 11:3 Hz, CH2), 4.66 (d, 1H, J ¼ 11:0 Hz, CH2),
4.80 (br s, 1H, CH), 4.23–4.15 (m, 1H, CH), 4.09–3.80 (m,
1H, CH), 3.45 (d, 1H, J ¼ 6:92 Hz, NH), 2.55 (d, 1H,
J ¼ 4:68 Hz, OH), 2.31 (br t, 2H, J ¼ 7:18 Hz, CH2), 1.41
(s, 9H, t-Bu). 13C NMR (CDCl3, 75 MHz, ppm): 159.6,
139.6, 135.0, 128.9, 128.8, 118.1, 118.0, 79.9, 75.6, 73.7,
51.0, 38.5, 29.1. HRMS (CI) m=z calcd for C20H30O4N
(M+H)þ: 348.2174, found: 348.2167. IR mmax film: cmꢁ1
2975, 1706, 1502, 1171.
This research was financially supported by the Nether-
lands Organisation for Scientific Research (NWO).
References and notes
1. Beaucaire, G. J. Chemother. 1995, 7(Suppl. 2), 111–123.
2. Zembower, T. R.; Noskin, G. A.; Postelnick, M. J.;
Nguyen, C.; Peterson, L. R. Int. J. Antimicrob. Agent
1998, 2, 95–105.
3. Some representative papers: (a) Greenberg, W. A.; Priest-
ley, E. S.; Sears, P. S.; Alper, P. B.; Rosenbohm, C.;

3632
G. F. Busscher et al. / Tetrahedron Letters 45 (2004) 3629–3632
20
D
12: ½aꢀ +12.8 (c 0.44, CH2Cl2). 1H NMR (CDCl3,
calcd for C24H41O5N4Si (M+H)þ: 493.2846, found:
493.2847. IR mmax film: cmꢁ1 2928, 2103, 1689, 1169,
1069, 610.
400 MHz, ppm): d 7.36–7.12 (m, 5H, arom Bn), 4.71 (d,
1H, J ¼ 11:6 Hz, CH2), 4.53 (d, 1H, J ¼ 11:6 Hz, CH2),
4.21 (m, 1H, CH), 4.1 (m, 1H, CH), 3.97 (m, 1H, CH),
3.75 (m, 1H, CH), 3.44 (m, 1H, CH), 3.27 (br d, 1H,
J ¼ 9:8 Hz, NH), 2.02–1.89 (m, 2H, CH2), 1.42 (s, 9H,
t-Bu), 0.88 (s, 9H, t-Bu), 0.03 (s, 6H, SiMe2). 13C NMR
(CDCl3, 100 MHz, ppm): d 154.8, 136.7, 128.3, 79.4, 73.7,
72.8, 72.3, 64.7, 60.6, 48.1, 30.1, 28.7, 26.0, 18.1, 14.6,
)4.7. HRMS (CI) m=z calcd for C24H42O6NSi (M+H)þ:
468.2782, found: 468.2785. IR mmax film: cmꢁ1 2925, 2358,
16. Grubbs’ second generation metathesis catalyst containing
the 1,3-dimesityl-4,5-dihydroimidazol-2-ylidene ligand
was used. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R.
H. Org. Lett. 1999, 1, 953–956.
17. Yang, D.; Jiao, G.-S.; Yip, Y.-C.; Wong, M.-K. J. Org.
Chem. 1999, 64, 1635–1639.
18. Recent review on cyclic sulfates, see: Buyn, H.-S.; He, L.;
Bittman, R. Tetrahedron 2000, 56, 7051–7091.
19. van der Klein, P. A. M.; Filemon, W.; Veeneman, G. H.;
van der Marel, G. A.; van Boom, J. H. J. Carbohydr.
Chem. 1992, 11, 837–848.
20. Shing, T. K. M.; Tam, E. K. W.; Tai, V. W.-F.; Chung, I.
H. F.; Jiang, Q. Chem. Eur. J. 1996, 2, 50–57.
21. Kim, B. M.; Sharpless, K. B. Tetrahedron Lett. 1989, 30,
658–665.
1778, 1697, 1508, 1135, 1082.
20
D
14: ½aꢀ )21.0 (c 0.10, CH2Cl2). 1H NMR (CDCl3,
400 MHz, ppm): d 7.43–7.24 (m, 5H, arom), 4.75 (d, 1H,
J ¼ 11:4 Hz, CH2), 4.60 (d, 1H, J ¼ 11:4 Hz, CH2), 3.72–
3.44 (m, 4H, CH), 3.22–3.11 (m, 1H, CH), 2.62–2.47 (m,
1H, CH2), 2.32–2.26 (m, 1H, CH2), 1.40 (s, 9H, t-Bu),
0.93 (s, 9H, t-Bu), 0.32 (s, 6H, SiMe2). HRMS (CI) m=z