Spinosyn G: Proof of structure by semisynthesis

Article abstract of DOI:10.1021/jo048173k

(Chemical Equation Presented) Spinosyn G was isolated in the late 1980s as a minor component from the broth of our potent, fermentation-derived insecticide spinosad. Its structure was then tentatively identified as 5″-epispinosyn A (3) on the basis of ¹H and ¹³C NMR data, but the 4″-epi compound 4 could not be conclusively ruled out with the data available. Described herein are unambiguous syntheses of both 3 and 4, whereby 3 was proved identical to the natural product. Compound 4 was prepared from intact spinosyn A by a novel F-TEDA-promoted oxidative deamination to the 4″-ketone 5, stereoselective reduction to the equatorial alcohol 6, and nitrogen incorporation via the axial azide 7. Compound 3 was obtained by coupling spinosyn A 17-pseudoaglycone (9) with the N-protected dihydropyran 16 derived from methyl L-ossaminide (14). This gave an ～2:1 mixture of anomeric products 17 with the desired equatorial glycoside predominating, which was then converted to 3 by N-deprotection and methylation.

Full text of DOI:10.1021/jo048173k

Spinosyn G: Proof of Structure by Semisynthesis
Paul R. Graupner,* Jacek Martynow,† and Peter B. Anzeveno‡
Discovery Research, Dow AgroSciences LLC, 9330 Zionsville Road, Indianapolis, Indiana 46268
prgraupner@dow.com
Received October 15, 2004
Spinosyn G was isolated in the late 1980s as a minor component from the broth of our potent,
fermentation-derived insecticide spinosad. Its structure was then tentatively identified as 5′′-
1
epispinosyn A (3) on the basis of H and 13C NMR data, but the 4′′-epi compound 4 could not be
conclusively ruled out with the data available. Described herein are unambiguous syntheses of
both 3 and 4, whereby 3 was proved identical to the natural product. Compound 4 was prepared
from intact spinosyn A by a novel F-TEDA-promoted oxidative deamination to the 4′′-ketone 5,
stereoselective reduction to the equatorial alcohol 6, and nitrogen incorporation via the axial azide
7. Compound 3 was obtained by coupling spinosyn A 17-pseudoaglycone (9) with the N-protected
dihydropyran 16 derived from methyl L-ossaminide (14). This gave an ∼2:1 mixture of anomeric
products 17 with the desired equatorial glycoside predominating, which was then converted to 3
by N-deprotection and methylation.
Introduction
during the 1960ssa compound whose total structure was
not determined until the 1990s3bsthis unit has only since
appeared as part of dunaimycin D2S4 and proposed as
part of spinosyn G,2b the subject of this report.
2,3,6-Trideoxy-4-aminoglycosides are relatively rare in
the natural product literature, with two families of
compounds dominating the landscape. As forosamine, the
glycosides appear in spiramycins1a and related com-
pounds (chimeramycins,1b and shengjimycins1c) which
consist of a 16-macrocyclic lactone also substituted with
a disaccharide. The spinosyns2 also predominately consist
of a macrocyclic core, substituted with a forosamine and
a permethylated rhamnose unit. Even more rare than
forosamine is the epimeric glycoside subunit ossamine,
which has only been reported in a very few natural
products. Initially reported as the amino-sugar fragment
derived by hydrolysis of the cytotoxic agent ossamycin3a
Spinosyn G was isolated in the late 1980s as a minor
component from the broth of the potent, fermentation-
derived insecticide Spinosad, a ∼90:10 mixture of spino-
syns A (1) and D (2). Its structure was tentatively
1
assigned as 5′′-epispinosyn A (3) on the basis of H and
13C NMR data, but the 4′′-epi compound 4 could not be
conclusively ruled out with the data available. In both,
the 1′′-anomeric linkage was equatorial, the same as in
A. G thus presented an anomaly, as the rest of the more
than 24 natural spinosyns characterized to date differed
from A only in the methylation patterns of the amine and
hydroxyl groups or in the extent of C-methylation of the
nucleus.
† Present address: Pharmaceutical Research Institute, Chemistry
Department I, 8 Rydygiera St., 01-793 Warsaw, Poland.
‡ Present address: Eli Lilly and Co, Eli Lilly Corporate Center,
Indianapolis, IN 46285.
If spinosyn G were indeed the 5′′-epi compound 3 then
the sugar was of the L-configuration and G represented
the only naturally occurring spinosyn with an L-sugar
at C-17, in this case, L-ossamine.
(1) (a) Omura, S.; Nakagawa, A.; Otani, M.; Haa, T. J. Am. Chem.
Soc. 1969, 91, 3401. (b) Omura, S.; Sadakane, N.; Tanaka, Y.;
Matsubara, H. J. Antibiot. 1983, 36, 927. (c) Sun, C.; Wei, J.; Huang,
J.; Jin, W.; Wang, Y. Actinomycetologica 1999, 13, 120.
(2) (a) Boeck, L. D.; Chio, E. H.; Eatin, T. E.; Godfrey, O. W.; Michel,
K. H.; Nakatsukasa, W. M.; Yao, R. C. (Eli Lilly) Eur. Pat. Appl. EP
375 316, 1990; Chem. Abstr. 1991, 114, 80066. (b) Kirst, H. A.; Michel,
K. H.; Mynderse, J. S.; Chio, E. E.; Yao, R. C.; Nakatsukasa, W. M.;
Boeck, L. D.; Occolowitz, J. L.; Paschal, J. W.; Deeter, J. B.; Thompson,
G. D. In Synthesis and Chemistry of Agrochemicals III; Baker, D. R.,
Fenyes, J. G., Steffens, J. J., Eds.; ACS Symposium Series 504; The
American Chemical Society, Washington, DC, 1992; pp 216-225.
(3) (a) Stevens, C. L.; Gutowski, G. E.; Byant, C. P.; Glinski, R. P.;
Edwards, O. E.; Sharma, G. M. Tetrahedron Lett. 1969, 10, 1181. (b)
Kirst, H. A.; Mynderse, J. S.; Martin, J. W.; Baker, P. J.; Paschal, J.
W.; Rios Steiner, J. L.; Lobkovsky, E.; Clardy, J. J. Antibiot. 1996, 49,
162.
(4) Hochlowski, J. E.; Mullally, M. M.; Brill, G. M.; Whittern, D.
N.; Buko, A. M.; Hill, P.; McAlpine, J. B. J. Antibiot. 1991, 44, 1318.
10.1021/jo048173k CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/11/2005
2154
J. Org. Chem. 2005, 70, 2154-2160

Spinosyn G: Proof of Structure by Semisynthesis
The F-TEDA-mediated oxidative deamination of 1 to
5 deserves mention. Attempts at “normal” Polonovsky
deamination5 via the N-oxide of 1 were unsuccessful, with
Cope elimination to the olefin the major reaction. The
F-TEDA oxidative deamination was discovered seren-
dipitously, when in connection with other work fluorina-
tion of 1 was attempted. Presumably, the tertiary amine
is N-fluorinated and oxidized under these conditions to
an imine, which is then quickly hydrolyzed to the ketone.
The reaction of tertiary amines with halogens and
oxidizing reagents to give imines or iminium ions is well-
known.6 It was later found that conversion to 5 could be
improved by using a two-step process. Thus, treatment
of spinosyn A with 1.25 equiv of F-TEDA and excess
pyridine in dry acetonitrile, isolation of the intermediate
product N-demethyl spinosyn A tetrafluoroborate salt
(90%), and subsequent treatment with an additional 1.25
equiv of F-TEDA and pyridine and finally with ion-
exchange resin in aqueous methanol resulted in isolation
of 5 in 63% overall yield.
The NMR spectra of 4 were clearly different from that
of spinosyn G, adding substance to the tentative assign-
ment of spinosyn G as the 5′′-epi compound 3. At this
time, a greater sample of the naturally occurring factor
G was available, and so further NMR experiments were
undertaken to confirm this assignment. In particular, 2D
gradient-enhanced rotating frame NOESY (gROESY)7
experiments were run, revealing a strong cross-peak
between 1′′ and 6′′, indicating the axial nature of the 5′′-
methyl group. These cross-peaks had been difficult to
detect in earlier experiments due to the lack of intensity
of NOE cross-peaks or the presence of noise in the spectra
acquired without gradient selection. With the structure
now seemingly confirmed as the 5′′ epimer of spinosyn
A, attention now focused on an unambiguous synthesis
of 3.
Synthetic Strategy to 5′′-Epispinosyn A (3). It was
envisioned that 3 could be obtained by coupling of a
suitably protected and activated L-ossamine derivative
with spinosyn A 17-pseudoaglycone (9, 17-Pseudoagly-
cone),2 as shown in Scheme 2. A potential major limita-
tion to this approach was the understanding that C-2
unsubstituted sugars gave predominantly R (axial) coupled
products in glycosidic coupling reactions; G had a â-gly-
cosidic linkage. In earlier work from these laboratories8a
and in the reported total syntheses of 1 by Evans8b and
Paquette,8c the coupling of forosamine derivatives with
17-Pseudoaglycone or close analogues had given pre-
dominately (66 to >90%) R-coupled products.
Described herein are some more recent NMR analysis
and unambiguous syntheses of both 3 and 4, which
proved 3 to be the structure of the natural product.
Results and Discussion
Synthesis of 4′′-Epispinosyn A (4). It was decided
to first attempt the synthesis of 4 as it was felt that, on
the basis of the wealth of synthetic information ac-
cumulated from our spinosyn synthetic analogue pro-
gram, it might be the more readily available of the two
compounds, resulting directly from intact spinosyn A by
oxidative deamination to the 4′′-ketone 5. Indeed, 5 had
been observed, albeit in low yields, when spinosyn A had
been treated with a variety of oxidizing reagents. Ad-
ditionally, a practical synthesis of 5, amenable to the
preparation of gram quantities, would be a boon to our
analogue program, allowing the preparation of novel 4′′-
substituted analogues.
Unfortunately, a viable, high-yield synthesis of 5 from
1 proved elusive, though enough material was isolated
from several attempts at same to carry on to the
preparation of 4. This synthesis is outlined in Scheme 1.
Treatment of 1 with 1-chloromethyl-4-fluorotriethylene-
diamine tetrafluoroborate (Selectfluor, F-TEDA) in
acetonitrile-water gave ketone 5 by a novel Polonovsky-
like oxidative deamination in low yield (18-30%). Stereo-
and chemoselective reduction of this with lithium tri-tert-
butoxyaluminohydride gave the equatorial alcohol 6
which was converted to the axial azide 7 by mesylation
and subsequent reaction with sodium azide. Chemose-
lective azide reduction with stannous chloride in metha-
nol gave the amine 8 which was converted to 4 by
permethylation with iodomethane to the quarternary
amine iodide and demethylation.
Synthesis of the L-Ossaminyl Coupling Partner.
It was decided at the outset to prepare N-demethyl-N-
(5) (a) Albini, A. Synthesis 1993, 263. (b) Michelot, R. Bull. Soc.
Chim. Fr. 1969, 4377.
(6) (a) Murahashi, S.-I. Angew. Chem., Int. Ed. Engl. 1995, 34, 2443.
(b) Goti, A.; Romani, M. Tetrahedron Lett. 1994, 35, 6567. (c) Mura-
hashi, S.-I.; Naota, T.; Miyaguchi, N.; Nakato, T. Tetrahedron Lett.
1992, 33, 6991. (d) Rousselet, G.; Capdevielle, P.; Maumy, M. Tetra-
hedron Lett. 1995, 36, 4999.
(7) Bax, A.; Davis, D. G. J. Magn. Reson. 1985, 63, 207.
(8) (a) DeAmicis, C. V.; Anzeveno, P. B.; Martynow, J. G.; McLaren,
K. L.; Green, F. R., III; Sparks, T. C.; Kirst, H. A.; Creemer, L. C.;
Worden, T. V.; Schoonover, J. R.; Gifford, J. M.; Hatton, C. J.; Hegde,
V. B.; Crouse, G. D.; Thoreen, B. R.; Ricks, M. J. U.S. Patent 6,001,-
981, 1999; Chem. Abstr. 1999, 132, 35986. (b) Evans, D. A.; Black, W.
C. J. Am. Chem. Soc. 1993, 115, 4497. (c) Paquette, L. A.; Collado, I.;
Purdie, M. J. Am. Chem. Soc. 1998, 120, 2553.
J. Org. Chem, Vol. 70, No. 6, 2005 2155

Graupner et al.
SCHEME 1. Synthesis of 4′′-Epispinosyn A (4)a
a Reagents and conditions: (a) F-TEDA, MeCN-H2O, 0 °C; (b) Li(t-BuO)3AlH, Et2O, 0 °C; (c) MeSO2Cl, (i-Pr)2NEt (DIPEA), CH2Cl2,
0 °C; (d) NaN3, DMF, 120 °C; (e) SnCl2, MeOH, rt; (f) MeI (xs), MeCN, reflux; (g) m-xylene, reflux.
SCHEME 2. Retrosynthetic Plan for Compound 3
SCHEME 3. Synthesis of the L-Ossaminyl Coupling Partnera
a Reagents and conditions: (a) H2, 45 psi, Pd(OH)2 on C, EtOH, rt; (b) formalin (37%), MeOH, reflux, 1 h, then NaBH4; (c) TROC-Cl,
K2CO3, toluene, reflux.
trichloroethoxycarbonyl-1-O-trichloroacetimidoyl-L-ossa-
mine [10, R ) trichloroethoxycarbonyl (TROC), Scheme
2] and use this as the coupling partner with the pseudo-
aglycone. This derivative was chosen over the acetimidate
derived from natural ossamine (10, R ) Me) because of
the anticipated ease of coupling of a nonbasic amide
derivative. Indeed, this strategy had worked well previ-
ously in a preparation of 9-O-forosaminyl spinosyn A.8a
Once coupled, simple N-protecting group removal fol-
lowed by N-methylation would give the targeted 5′′-epi
compound (3).
consistently gave better yields than the one-step litera-
ture conversion.
Reaction of methyl ossaminide (14) with TROC-Cl and
K2CO3 in benzene at reflux afforded a 65% yield of the
desired N-demethyl-N-TROC ossaminide 15 along with
15-20% of dihydropyran 16 (Scheme 3), formed by loss
of methanol from 15 under the reaction conditions. The
unexpected formation of 16 proved to be a boon to the
synthesis of 3, for the glycosidic coupling of dihydropy-
ranyl “sugars” is well documented in the literature
obviating the need to prepare 10. The yield of 16 was
subsequently raised to 35-40% by conducting the reac-
tion in refluxing toluene in the presence of 3 equiv of
K2CO3.
Methyl L-ossaminide (14) was prepared from 3,4-di-O-
acetyl-6-deoxy-L-glucal (11) by slight modification of the
literature procedures9 (Scheme 3). As per ref 9, glucal
11 was converted to azide 12 possessing the desired axial
N-configuration. In our hands, conversion of 12 to 14 by
a two-step hydrogenation/reductive amination procedure
(9) (a) Brimacombe, J. S.; Doner, L. W.; Rollins, A. J. J. Chem. Soc.,
Perkin Trans. 1 1972, 2977. (b) Malik, A.; Afza, N.; Voelter, W. Ann.
1984, 636. (c) Malik, A.; Afza, N.; Voelter, W. J. Chem. Soc., Perkin
Trans. 1 1983, 2103.
2156 J. Org. Chem., Vol. 70, No. 6, 2005

Spinosyn G: Proof of Structure by Semisynthesis
FIGURE 1. Transition states for axial vs equatorial attack on 16.
SCHEME 4. Synthesis of 18a
a Reagents and conditions: (a) PPTS, BF3OEt2, 1,2-dichloroethane, rt; (b) Zn, 1M NH4OAc, THF-MeOH, rt.
Coupling of Dihydropyran 16 with 17-Pseudo-
aglycone (9). It was envisaged that coupling of dihy-
dropyran 16 with 9 would give a preponderance of the
desired equatorial coupled product (in this case R product
as 16 is derived from an L-sugar) based on the putative
transition states (Figure 1) for axial and equatorial attack
and the fair assumption that the sterically large N-
TROC-N-methyl group of 16 would prefer that configu-
ration in which it was equatorial (or pseudoequatorial),
thus forcing the 5-methyl group to be axial (16a). Axial
attack by the pseudoaglycone (path A) would be impeded
by a severe 1,3 steric interaction with the axial methyl
group. Equatorial attack by the pseudoaglycone (path B)
is not impeded and should be favored.
NMR due to the presence of the TROC protecting group
because the resonance pattern of its CH2 moiety over-
lapped the key 1′′-H resonance. This mixture was treated
with zinc, and the resulting mixture of deprotected
amines (18) was shown by NMR to be an ∼2:1 anomeric
mixture at C-1′′ with the desired equatorial (â) glycoside
predominating.
Completion of the Synthesis of 5′′-Epispinosyn A
(3). Treatment of the anomeric mixture of amines 18 with
2.5 equiv of MeI in THF in the presence of Hunig’s base
gave a mixture of the two expected tertiary amines in
49% yield. These were easily separated by silica gel
chromatography. The major of these, the R (L-sugar
nomenclature) equatorial coupled product, was shown by
NMR and LC analysis to be identical in all respects to
natural spinosyn G (Supporting Information).
This indeed was found to be the case. Coupling of 16
with 1 equiv of 9 in dichloroethane with a mixture of BF3‚
Et2O and pyridinium p-toluenesulfonate as catalyst gave
a 36% yield of coupled products (17, Scheme 4). The
composition of this mixture was difficult to determine by
It is interesting that the minor coupled product (19)
also possessed an equatorial C-1′′ glycosidic linkage
(Scheme 5). Apparently, axial coupling induces a severe
J. Org. Chem, Vol. 70, No. 6, 2005 2157

Graupner et al.
SCHEME 5. Completion of the Synthesis of 3a
a Reagents and conditions: (a) MeI (2.5 equiv), DIPEA, THF, rt; (b) silica gel chromatography.
1H), 4.41 (d, 1H, J ) 8.6 Hz), 3.59 (m, 1H), 3.49 (s, 3H), 3.43
(s, 6H), 1.22 (d, 6H, J ) 6.5 Hz), 1.12 (d, 3H, J ) 6.6 Hz), 0.77
(t, 3H, J ) 7.5 Hz); 13C NMR (CDCl3) δ 203.0, 172.6, 147.7,
144.2, 129.3, 128.8, 103.1, 95.3, 82.2, 81.0, 80.7, 76.6, 75.9, 75.8,
71.1, 67.8, 60.8, 58.9, 57.6, 49.3, 47.5, 47.4, 45.8, 41.4, 41.0,
37.2, 36.1, 34.1, 33.9, 31.3, 31.0, 30.5, 29.9, 28.2, 21.4, 17.9,
17.6, 16.0, 9.2.
1,3-diaxial steric interaction between the C-5′′ methyl
group, and the bulky pseudoaglycone forces the sugar to
ring flip, this putting the 4′′-dimethylamino group in the
axial configuration and the methyl into the equatorial
position. NMR was used to confirm this geometry and
was also able to distinguish 19 unambiguously from 4′′-
epispinosyn A (4). Compound 19 was thus â-L-ossami-
nylspinosyn A 17-pseudoaglycone and 4, â-D-ossaminyl-
spinosyn A 17-pseudoaglycone.
(4S)-4′′-Desdimethylamino-4′′-methanesulfonyloxyspi-
nosyn A (6a). Methanesulfonyl chloride (1.80 mL) was added
dropwise during 10 min to a cold (0 °C), well-stirred, N2-
blanketed solution of 6 (414 mg, 0.59 mmol) in dry CH2Cl2 (10
mL) containing pyridine (4 mL) and DIPEA (2 mL). This
mixture was stirred at 0 °C for 1 h, then Et2O (100 mL) and
brine (10 mL) were added, and this mixture was stirred at
room temperature for 15 min. After dilution with H2O (100
mL), the phases were separated and the organic layer was
washed with 10% KHCO3 (4 × 25 mL), dried (K2CO3), and
concentrated. The residue was purified by flash silica gel
chromatography with 15% Et2O in CH2Cl2 affording 6a as a
white powder (418 mg, 91%): 1H NMR (CDCl3) δ 6.73 (br s,
1H), 4.48 (d, 1H, J ) 9.0 Hz), 4.20 (ddd, 1H, J ) 10.7, 9.6, 5.0
Hz), 3.51 (s, 3H), 3.45 (s, 6H), 2.98 (s, 3H), 1.25 (d, 3H, J )
6.0 Hz), 1.23 (d, 3H, J ) 6.0 Hz), 1.13 (d, 3H, J ) 6.8 Hz),
0.77 (t, 3H, J ) 7.4 Hz); 13C NMR (CDCl3) δ 202.8, 172.6, 147.7,
144.1, 129.4, 128.8, 102.7, 95.4, 82.2, 81.0, 80.9, 79.9, 77.6, 76.5,
76.0, 72.7, 67.8, 65.7, 60.8, 58.9, 57.6, 49.3, 47.5, 47.4, 45.9,
41.4, 41.0, 38.5, 37.2, 36.1, 34.1, 30.1, 29.9, 29.0, 28.2, 21.3,
17.9, 17.6, 16.0, 9.2.
(4R)-Desdimethylamino-4′′-azidospinosyn A (7). So-
dium azide (3.10 g) was added to a solution of 6a (171 mg,
0.22 mmol) in dry DMF and the resulting mixture heated, with
stirring under N2, to 120 °C and kept there for 1 h. The mixture
was then cooled to room temperature, and Et2O (100 mL) and
H2O (80 mL) were added. The phases were separated, and the
organic layer was washed with 10% KHCO3 solution, dried
(K2CO3), and concentrated at reduced pressure. The residue
was purified by reversed-phase HPLC over a C-18 column with
8% H2O in MeOH as eluent to give 7 as a white powder (62
mg, 39%): 1H NMR (CDCl3) δ 6.73 (br s, 1H), 4.43 (dd, J )
8.0, 3.4 Hz), 3.60 (m, 2H), 3.51 (s, 3H), 3.45 (s, 3H), 3.44 (s,
3H), 1.23 (d, 3H, J ) 6.1 Hz), 1.20 (d, 3H, J ) 6.3 Hz), 1.13 (d,
3H, J ) 6.8 Hz), 0.76 (t, 3H, J ) 7.6 Hz); 13C NMR (CDCl3) δ
203.0, 172.6, 147.7, 144.2, 129.3, 128.8, 103.4, 95.4, 82.2, 81.0,
80.5, 77.6, 76.7, 76.0, 72.6, 67.8, 60.8, 58.9, 58.7, 57.6, 49.3,
47.5, 47.4, 45.9, 41.4, 41.0, 37.2, 36.1, 34.1, 34.0, 29.8, 28.3,
27.2, 25.5, 21.5, 17.8, 17.6, 15.9, 9.1; MS (CI) m/z 747 (M +
NH4+).
Experimental Section
4′′-Ketospinosyn A (5). A stirred solution of 1 (10 g, 13.6
mmol) and pyridine (3.5 g, 44 mmol) in acetonitrile (100 mL)
was cooled to 5 °C and treated with F-TEDA (6 g, 17 mmol),
in portions over 1 h. After addition was complete, the reaction
was allowed to stir for an additional 1 h and then concentrated
in vacuo. HCl (1 N, 50 mL) and ether (100 mL) were added,
and the solution was stirred for 1 h, during which time a fine
precipitate formed. Filtration and air-drying gave 9.9 g (90%)
of N-demethylspinosyn A tetrafluoroborate as an off-white
powder. To a solution of this (6.4 g, 8 mmol) in dry acetonitrile
(100 mL) was added pyridine (1.4 g, 17.6 mmol). The solution
was cooled to 5 °C and F-TEDA (3.75 g, 10.6 mmol) added in
portions over 40 min. After addition was complete, the solution
was allowed to stir for an additional 2 h whereupon the solvent
was removed in vacuo and the residual oil partitioned between
ether (70 mL) and water (50 mL). The layers were separated,
and the organic layer was washed with 0.1 N HCl (2 × 50 mL)
and with saturated NaHCO3 solution (50 mL). Drying and
concentration gave an oil which was dissolved in MeOH (50
mL) and water (10 mL). The solution was stirred and treated
with Dowex resin (H+ form, 5 g) at ambient temperature for 4
h and the solution filtered and concentrated. Silica gel chro-
matography (50:50 ethyl acetate/hexane) yielded 5 (4 g, 63%
total yield from 1) as a white powder: 1H NMR (CDCl3) δ 6.73
(br s, 1H), 4.94 (dd, 1H, J ) 8.0, 3.0 Hz), 3.98 (q, 1H, J ) 6.8
Hz), 3.68 (m, 1H), 3.50 (s, 3H), 3.44 (s, 6H), 1.26 (d, 3H, J )
6.8 Hz), 1.23 (d, 3H, J ) 6.6 Hz), 1.15 (d, 3H, J ) 6.5 Hz); 13
C
NMR (CDCl3) δ 208.8, 202.8, 172.6, 147.7, 144.1, 129.4, 128.8,
101.0, 95.5, 82.2, 81.1, 81.0, 77.6, 76.7, 76.1, 75.8, 67.8, 60.9,
58.9, 57.6, 49.4, 47.5, 47.3, 46.0, 41.3, 41.0, 37.3, 36.2, 34.8,
34.2, 34.1, 30.1, 28.2, 21.1, 17.6, 16.3, 15.4, 9.2.
(4S)-4′′-Desdimethylamino-4′′-hydroxyspinosyn A (6).
Lithium tri-tert-butoxyaluminohydride (610 mg, 2.33 mmol)
was added in one portion to a cold (0 °C), well-stirred,
N2-blanketed solution of 5 (415 mg, 0.59 mmol) in dry Et2O
(20 mL) and dry dioxane (4 mL). After 10 min, brine (10 mL)
was slowly added, followed by Et2O (100 mL), and this mixture
was treated with 2 N aq NaOH (100 mL). The phases were
separated, and the ethereal layer was washed with 10%
KHCO3 (2 × 25 mL), dried (K2CO3), and concentrated leaving
pure 6 as a white powder which was dried in vacuo at room
temperature (414 mg, 99%): 1H NMR (CDCl3) δ 6.72 (br s,
4′′-Epispinosyn A (4). To a solution of 7 (51 mg, 0.07 mmol)
in MeOH (4.0 mL) was added excess SnCl2‚2 H2O (140 mg) in
one portion. This mixture was stirred at room temperature
under N2 for 20 h, and then the MeOH was removed at reduced
pressure. The residue was dissolved in 1:1 EtOAc-Et2O (100
mL) and this solution washed with aq NH4OH (20 mL) and
dried (K2CO3). Concentration left a colorless glassy solid (38
mg) which by NMR analysis was a mixture of amine 8 (60%)
2158 J. Org. Chem., Vol. 70, No. 6, 2005

Spinosyn G: Proof of Structure by Semisynthesis
and unreacted 7 (40%). This mixture was carried on without
separation. Some 13C NMR (CDCl3) signals ascribable to 8: δ
203.1, 172.8, 147.4, 144.5, 129.4, 128.8, 103.8, 48.1, 35.0, 30.7,
25.0, 15.7, 9.2.
3H), 2.35 (m, 1H), 2.30 (s, 6H), 2.30 (m, 1H), 1.93-1.82 (m,
2H),1.75 (m, 1H), 1.55 (m, 1H), 1.28 (d, 3H, J ) 6.5 Hz); 13C
NMR (CDCl3) δ 97.5, 69.2, 61.4, 55.2, 43.5, 28.9, 19.0, 15.0;
MS m/z 173 (M+, 10), 142 (10), 84 (20), 71 (100).
Methyl N-Demethyl-N-trichloroethoxycarbonyl-L-os-
saminide (15) and 4-(N-Methyl-N-trichloroethoxycarbo-
nylamino)-1,2,3,4,6-pentadeoxy-L-threo-hexenopyra-
nose (16). A solution of methyl ossaminide (14) (1.6 g, 9.24
mmol) in toluene (60 mL) was heated at reflux in a round-
bottom flask equipped with a Dean-Stark water separator for
20 min to remove any water. The solution was then cooled to
room temperature, and trichloroethylchloroformate (4.0 mL)
and powdered, anhydrous K2CO3 (1.6 g, 11.5 mmol) were
added. The resulting mixture was heated with stirring at
reflux for 20 h. The mixture was then cooled to rt and filtered,
and the collected salts were washed with EtOAc (25 mL). The
combined filtrate and wash was washed with satd Na2CO3
solution (2 × 50 mL) and brine (25 mL) and dried (MgSO4).
Concentration left 5.0 g of crude product containing excess
chloroformate which was flash chromatographed over silica
(200 mL) with 15% EtOAc in hexane as eluent. After a forerun
of 150 mL, 20 mL fractions were collected. Pure dihydropyran
16 (1.0 g) as a colorless oil was obtained from fractions 6-8.
Pure ossaminide 15 (0.6 g), also a colorless oil, was obtained
from fractions 9-12. Data for 16: 1H NMR (CDCl3, 600 MHz)
δ 6.45 (br s, 1H), 4.87 (m, 2H), 4.72 (m, 1H), 4.54 and 4.48 (br
d, total 1H, J ) 8.6 Hz), 4.10 (m, 1H), 3.10 and 3.05 (s, total
3H), 2.65 and 2.05 (m, total 2H), 1.30 and 1.25 (d, total 3H, J
) 6.5 Hz); 13C NMR (CDCl3) 155.9, 145.2, 101.2, 75.7, 73.9,
51.5, 31.3, 30.7, 24.9, 17.4. Data for 15 (major R-anomer): 1H
NMR (CDCl3, 600 MHz) δ 4.75 (m, 3H), 4.15 (m, 1H), 4.15 (s,
3H), 3.20 (m,1H), 3.00 and 3.08(s, total 3H), 2.4-1.7 (m, 4H),
1.30 (d, 3H, J ) 6.5 Hz).
N-Demethyl-N-trichloroethoxycarbonyl-L-ossaminyl
Spinosyn A 17-Pseudoaglycone (17). Crushed 5 Å molec-
ular sieves were added to a solution of dihydropyran 16 (0.2
g, 0.66 mmol) and spinosyn A 17-pseudoaglycone (9, 0.39 g,
0.66 mmol) in reagent grade 1,2-dichloroethane (10 mL). To
this stirred mixture was added PPTS (33 mg, ∼20 mol %) at
rt. After 1 h, the mixture was heated to reflux. After 4 h, the
mixture was cooled, and BF3‚Et2O (25 µL, ∼0.19 mmol) was
added. The mixture was again heated to reflux, kept at reflux
for 2 h, and then recooled to room temperature, 25 µL of
additional BF3‚Et2O was added, and the mixture was allowed
stir at room temperature. After 18 h, 25 µL of additional BF3‚
Et2O was added and the mixture allowed stir an additional
24 h. The mixture was then diluted with CH2Cl2 (20 mL) and
filtered (Celite). The collected solids were washed with more
CH2Cl2 (10 mL), and the combined filtrate and wash was
washed with satd NaHCO3 (2 × 10 mL) and then brine (10
mL) and dried (MgSO4). Concentration left 0.5 g of crude
product which was purified by flash silica chromatography (50
mL silica) using 4.5% MeOH in CH2Cl2 to give 0.21 g (36%) of
coupled product as a thick viscous oil as a 2:1 â/R anomeric
mixture (this ratio determined after removal of the protecting
TROC group, below). 17: 1H NMR (CDCl3) δ 6.78 (br s, 1H),
5.88 (m, 1H), 5.78 (m, 1H), 4.90-4.62 (m, 6H), 4.30 (m, 1H),
3.56 (s, 3H), 3.50 (3, 6H), 1.30 (d, 3H, J ) 6.2 Hz), 1.18 and
1.12 (d, total 3H, J ) 6.8 Hz).
The mixture of 7 and 8 (38 mg) was dissolved in MeCN (5
mL), excess iodomethane (3.5 mL) added, and the solution
heated at reflux, with stirring under N2 for 2 h. The solvent
and excess iodomethane were removed, and the residue was
dissolved in m-xylene (10 mL). This solution was heated at
reflux for 1 h with periodic removal of a small amount (∼0.5-1
mL) of the condensate through distillation (5 times during the
1 h). After cooling, the solvent was removed at reduced
pressure and the residue purified by flash silica gel chroma-
tography giving 4 (7.0 mg, 14% from 7) as a colorless glass:
1H NMR (CDCl3) δ 6.75 (br s, 1H), 5.84 (br d, 1H, J ) 9.8 Hz),
5.78 (ddd, 1H, J ) 9.8, 2.5, 2.5 Hz), 4.82 (d, 1H, J ) 1.4 Hz),
4.65 (m, 1H), 4.56 (br d, 1H, J ) 10.0 Hz), 4.28 (m, 1H), 3.81
(m, 1H), 3.62 (m, 1H), 3.51 (s, 3H), 3.45 (s, 6H), 2.60 (s, 6H),
1.43 (d, 3H, J ) 6.4 Hz), 1.23 (d, 3H, J ) 6.2 Hz), 1.13 (d, 3H,
J ) 6.8 Hz), 0.80 (t, 3H, J ) 7.5 Hz); 13C NMR (CDCl3) δ 202.6,
172.5, 147.7, 144.2, 129.3, 128.7, 102.9, 95.4, 82.2, 81.0, 80.4,
77.7, 76.8, 76.0, 72.8, 67.9, 61.0, 59.5, 59.0, 57.7, 49.4, 47.7,
47.6, 45.9, 43.7, 41.5, 41.2, 37.4, 36.3, 34.1, 29.9, 28.4, 27.9,
21.8, 20.5, 19.2, 17.8, 15.9, 9.3; MS (ESI) m/z 732 (M + H).
Spinosyn A 17-Pseudoaglycone (9). To a mechanically
stirred suspension of spinosyn A (1, 21.9 g, 0.03 mol) in water
(450 mL) was added 1 N aq H2SO4 (60 mL) in one portion.
The resulting solution was heated under N2 to 80-90 °C and
kept at that temperature range for 24 h. During this time, 9
separated and the resulting suspension became quite thick.
The mixture was then cooled to rt and filtered and the collected
9 washed with water (3 × 50 mL). The solid product was then
dissolved in CH2Cl2 (500 mL) and the solution washed with
brine (100 mL), dried (K2CO3), and evaporated at reduced
pressure. The crude product was purified by flash silica gel
chromatography with 30% hexane in EtOAc to give 9 (15.9 g,
90%) as a white powder: 1H NMR (CDCl3) δ 6.78 (br s, 1H),
5.89 (br d, 1H, J ) 9.8 Hz), 5.80 (ddd, 1H, J ) 9.8, 2.5, 2.5
Hz), 4.86 (d, 1H, J ) 1.4 Hz), 4.70 (m,1), 4.32 (m, 1), 3.70 (m,
1H), 3.55 (s, 3H), 3.40 (s, 6H), 1.27 (d,3H, J ) 6.2 Hz), 1.21 (d,
3H, J ) 6.8 Hz), 0.92 (m, 1H), 0.82 (t, 3H, J ) 7.5 Hz); MS
(EI) m/z 591 (M + H).
Methyl 4-Amino-2,3,4,6-tetradeoxy-L-threo-hexopyra-
noside (13). A solution of the 4-azide 129 (2.0 g, 11.8 mmol)
as an anomeric mixture, in absolute EtOH (50 mL), was
hydrogenated over Pearlman’s catalyst [20% Pd(OH)2/C, 1.0
g] at 45 psi for 5 h. The catalyst was filtered (Celite) and
washed with EtOH (25 mL). The combined filtrate and wash
was concentrated at reduced pressure and ambient tempera-
ture leaving the crude amine 13 (1.5 g) as a colorless,
gelatinous oil which was used immediately and without further
purification. R-Anomer: 1HNMR (CDCl3) δ 4.68 (d,1H, J ) 1.9
Hz), 3.98 (dq, 1H, J ) 6.5, 1.6 Hz), 3.36 (s, 3H), 2.74 (br s,
1H), 2.1-1.5 (m, 6H), 1.17 (d, 3H, J ) 6.5 Hz); MS m/z 145 (M
+, 5), 114 (M - OCH3, 35), 89 (100).
Methyl L-Ossaminide (14). To a stirred solution of the
above amine 13 (1.5 g, 0.01 mol) in reagent grade MeOH (35
mL) was added 37% aq CH2O solution (10.2 mL), and the
resulting solution was heated at reflux for 1 h. The mixture
was then cooled to 0-5 °C, and sodium NaBH4 (1.37 g, 0.036
mol) was added in portions during 5 min. The cooling bath
was removed after 20 min, and the reaction was allowed to
stir at rt for 18 h. The solvent was then removed at reduced
pressure and the residue extracted with CH2Cl2 (3 × 30 mL
portions, each time decanting the extract from the tarry
insolubles). The combined extracts were then washed with
brine (5 mL) and dried (MgSO4). Concentration left 1.3 g of
product which was purified by flash chromatography over silica
(75 mL) using 10% MeOH in CH2Cl2 as eluent to give 0.65 g
(38%) methyl l-ossaminide (14) as a colorless oil as an ∼5:1
R/â anomeric mixture. R-Anomer:1H NMR (CDCl3) δ 4.66 (dd,
1H, J ) 6.1, 3.1 Hz), 4.25 (dq, 1H, J ) 6.5, 4.6 Hz), 3.40 (s,
N-Demethyl-L-ossaminyl Spinosyn A 17-Pseudoagly-
cone (18). To a well-stirred solution of the mixture of coupled
products 17 (0.2 g, 0.22 mmol) in THF (6.0 mL) was added 1
M aq NH4OAc solution (1.2 mL), followed by more THF (3.0
mL) and MeOH (1.5 mL). To this homogeneous solution was
added freshly activated10 zinc dust (1.0 g) and the resulting
mixture stirred for 3.5 h at ambient temperature. The mixture
was then filtered (Celite), the collected solids were washed with
1:1 THF-MeOH (25 mL), and the combined filtrate and wash
was concentrated to ∼15% volume at reduced pressure and
(10) Fieser, L. F.; Fieser, M. Reagents for Organic Synthesis; John
Wiley and Sons: New York, 1968; Vol. 1, p 1276.
J. Org. Chem, Vol. 70, No. 6, 2005 2159

Graupner et al.
25-30 °C. The residue was basified with satd NaHCO3 to ∼pH
9 and then extracted with CH2Cl2 (3 × 15 mL). The combined
organic extracts were then washed with brine (10 mL) and
dried (MgSO4). Concentration left 0.12 g of crude deprotected
product which was purified by silica chromatography with 5%
MeOH in CH2Cl2 as eluent to give 66 mg (42%) of 18 as a
colorless foam as a 2:1 â/R anomeric mixture: 1H NMR (CDCl3,
600 MHz) δ 6.78 (s, 1H, H-13), 4.85 (m, H-1, H1′′), 4.68 (m,
1H, H-21), 4.33 (m, 1H, H-9), 4.15 (m, H-5â), 3.92 (m, H-17R),
3.70 (m, H-17â), 3.74 (m, H-5R), 2.45 and 2.42 (s, NCH3).
Spinosyn G (3) and â-L-Ossaminyl Spinosyn A 17
Pseudoaglycone (19). To a solution of the mixture of ano-
meric products 18 (60 mg, 0.084 mmol) in anhydrous THF (2
mL) were added MeI (10.4 µL, ∼0.168 mmol) and DIPEA (44
µL, ∼0.25 mmol), and this solution was stirred at room
temperature for 20 h and then diluted with Et2O (15 mL),
washed with brine (5 mL), and dried (MgSO4). Concentration
left 60 mg of crude methylated product which was purified by
radial chromatography over one 2.0 mm silica chromatotron
plate using 9% MeOH in CH2Cl2 and 30% MeOH in CH2Cl2
as eluent, with the 30% eluent started after the first (major)
product band (spinosyn G) had come off. This afforded spinosyn
G (R-L-ossaminyl spinosyn A 17 pseudoaglycone (3), 20 mg,
33%) as a colorless foam: 1H NMR (CDCl3) δ 6.79 (br s, 1H,
13), 5.89 (br d, 1H, 6), 5.80 (ddd, 1H, 5), 4.87 (d, 1H, 1′), 4.78
(dd, 1H, 1′′), 4.69 (m, 1H, 21), 4.32 (m, 1H, 9), 4.29 (dq, 1H,
5′′), 3.64 (ddd, 1H, 17), 3.58 (s, 3H, 4′-OMe), 3.56 (dq, 1H, 5′),
3.51 (s, 3′-OMe) 3.50 (s, 3H, 2′-OMe), 3.30 (dq, 1H16), 3.13
(dd, 1H, 4′), 3.11 (dd, 1H, 2), 3.03 (m, 1H, 3), 2.89 (m, 1H, 12),
2.42 (dd, 1H, 2), 2.35 (m, 1H, 4′′), 2.29 (s, 6H, NMe2), 2.27 (m,
1H, 10), 2.18 (m, 1H, 7), 1.30 (d, 3H6′′), 1.27 (d, 3H, 6′′), 1.20
(d, 3H, 24), 0.91 (dddd, 1H, 11), 0.83 (t, 3H, 23). 13C NMR
(CDCl3) δ 203.2, 172.9, 147.7, 144.4, 129.7, 129.4, 99.0, 95.2,
82.7, 81.9, 81.5, 78.2, 77.2, 76.7, 76.6, 70.3, 68.4, 62.4, 61.3,
59.4, 58.1, 49.9, 48.0, 47.7, 46.7, 43.8, 41.9, 41.6, 37.8, 36.7,
34.9, 34.8, 31.0, 29.8, 28.6, 21.3, 20.2, 18.2, 17.4, 14.7, 9.8. Also
isolated was the â-L-ossaminyl isomer 19 (6 mg, 10%) also as
a colorless foam: 1H NMR (CDCl3) δ 6.81 (br s, 1H, 13), 5.89
(br d, 1H, 6), 5.80 (ddd, 1H, 5), 4.86 (d, 1H, 1′), 4.68 (m, 1H,
21), 4.56 (dd, 1H, 1′′), 4.33 (ddd, 1H, 9), 3.95 (ddd, 1H, 17),
3.78 (dq, 1H, 5′′), 3.58 (s, 3H, 4′-OMe), 3.52 (s, 3H, 2′-OMe),
3.51 (s, 3H, 3′-OMe), 3.30 (dq, 1H, 16), 3.13 (t, 1H, 4′), 3.11
(dd, 1H, 2), 3.07 (m, 1H, 3), 2.90 (dddd, 1H, 12), 2.47 (dd, 1H,
2), 2.43 (s, 6H, NMe2), 2.32 (m, 1H, 4′′), 2.26 (m, 1H, 10), 2.21
(m, 1H, 2′′), 2.16 (m, 1H, 7), 1.94 dd, 1H, 8), 1.35 (d, 3H, 6′′),
1.31 (d, 3H, 6′), 1.29 (s, 3H, 24), 0.91 (dddd,1H, 11), 0.84 (t,
3H, 23); 13C NMR (CDCl3) δ 203.7, 172.9, 147.6, 144.2, 129.7,
129.5, 99.2, 95.9, 82.7, 81.5, 78.2, 77.9, 77.2, 76.6, 74.6, 68.4,
61.3, 59.6, 59.4, 58.1, 50.0, 48.1, 46.9, 46.7, 44.1, 41.8, 41.5,
37.8, 36.8, 35.0, 32.4, 31.2, 30.2, 28.4, 20.2, 20.1, 18.3, 18.2,
18.1, 9.8.
Acknowledgment. We thank Dennis Duebelbeis of
our Natural Products Group for LC/MS and MS data
on several key intermediates.
Note Added after ASAP Publication. The reaction
conditions for Scheme 3 were mistakenly included in
Scheme 2 and those for Scheme 4 included with Scheme
3 in the version published ASAP February 11, 2005; the
corrected version was published February 15, 2005.
Supporting Information Available: NMR spectra for 1,
3, 4, and 19. This material is available free of charge via the
Internet at http://pubs.acs.org.
JO048173K
2160 J. Org. Chem., Vol. 70, No. 6, 2005