Tetrahedron Letters
No. 33, pp 2835 - 2638, 1977.
Pergamon Press.
Printed
in Oreat Mtein.
ANSA MACROLIDE SYNTHESIS II* A CONVERGENT APPROACH TO MAYTANSINE
,Rossnne Bonjouklianand Bruce Ganem*
Department of Chemistry Cornell university Ithaca, New York (Received
in USA 9 Maroh 1977; reoeived
14853 in UK for publioation
23 June 1977)
Cne approach to the total synthesis of clinically interesting m@ansinoidsl the macrocyclic
ring at the Cl -N peptide bond.
quires the smooth stereospecific
1 relies on closure of
To be preparatively usefuI any plan of this nature re-
assembly of an sppropriate acyciic precursor,
consequently we have
devised a convergent scheme which employs C9 as a connecting point for suitably protected Cl-C8 Cl0 - C21 fragments.
This Letter describes
sml
our coupling method, based on tie principle of %mpole*,
and details the subsequent elaboration of maytansine’s csrblnolamide 2 model system -* 18
functionality In a represeptetive
CH30 1
R= -CO-7%CH3
I-I
Retrosynthetic analysis suggests that the six-membered
3
c/“\
cCC%
carbamate in 1 might be formed b intra-
Precedents exist for 3, 2c Inthecsse this type of cyclization during the Curtius rearrangement of y-hydroxyacid hydraxides. molecular interception of a C9 isocyanate with an hydroxyl substituent at C7 of 2.
of maytansine, the appropriate a, y-dihydroxyacid densation of C9 glycolate fragment 3
synthon in 2 can be imagined to arise from the con-
epoxide 5, and a Cl0 aldehyde A. With current improve4 ments in the Curtius and Lossen rearrangements, these VUmpolu~l transformations can be achieved
under very mild, near-neutral
a C/-C8
conditions which are compatible with maytansine~s sensitive functionality.
2835
2836
No. 33
(Xlr strategy has now been reallaed in the synthetic direction by the successful consecutive alkylation of 1 (R= C6H5) with representative structures
4 and fi.
Cur choice of phenozyacetic acid as a nucleophilic (1) C9 must possess
C9 reagent wss guided by three considerations:
sufficient reactivity to undergo SR2 opening of a monosubstituted ozirsne,
(2) there
must be a favorable equlllbrium in the aldol condensation of 2 with i to form a quaternary carbon at C9, and (3) the C-protecting Base-catalyzed
group at C9 must bs resdlly removable at the end of the synthesis.
addition of various @protected
in extensive ester self-condensation. of lithium dlisopropylamlde
glycollc acid esters 5 to propylene ozide resulted
However the dlanion of 1 (R= C6H6), prepared using two eouiv
(“LDA1l, THF, -2O’),
NH4C1 workup @H 5) the stereoisomeric
condensed readllywith
y-la&ones
these could be separated by prep tic (cis isomer,
8 were isolated
mp 69.541’
;
2 (Rq1=CH3), and after
(1:l ratio) in 70% yield. isomer,
it was more convenient to employ the miziure of 8 for the next alkylation.
While
mp 55.5-560) 697 ,
Generation of the la&one
enolate (LDA, THF, 1 hr) and addition of freshly distilled sorbaldehyde 2 (1.0 equiv) was carried out at -780; after 30 min, addition of NH4Cl enabled the isolation of aldol product I& reaction mixture to 0’ before protonation led only to 2 and unalkylated la&one.
Warming the Since Cl0 of ma.yt*
sine bears a methozyl group, it was most convenient to 0-alkylate the alkozide anion in situ (5 equiv Using this one-pot reaction, dienic la&one 11 could be produced in 75% yield CH31, 5 equiv HMPA). 2b The pmr spectrum of && showed three CCH3 singlets, confirming the presence of a from a. diaztereomeric
mixture.
Modified experimental conditions to achieve kinetic stereoselection 8 condensation are currently being explored.
in this
No. 33
2837
4; OR 0CH3 -8
0
-10
R=H
-12
X=NHNH2, R=H
-11
R=CH3
-13
X= N3, R= TMS
14
X= NHOH, R= H
+$io - do
-
GCH3
bCH3 -18
-17 exposure of G
to excess
zide -12 in excellent yield.
950/Ohydrazlne hydrate in ethanol or THF furnished the hydroxyhydra-
If immediately treated with N204 (1 equiv.,
-78“,
CH2C12, lhr) and then
with chlorotrimethylsilane/ triethylamine (1:l ratio, 10 equiv. ), -12 was smoothly oxidized to the 2c Curtius rearrangement (benzene, NaOAc, reflux) transformed 13 into the silyloxyazide -13. corresponding
silyloxyisocysnate
base but was efficiently desilylated
i&
Thls substance was relatively stable in the presence of acid or
using tetra-n-lUylammonium
fluoride
(1 equiv.,
THF, 20 min).
Spontaneous cyclization of the intermediate hydroxyisocyanate 16 produced cyclic carbamate & 17 6 1 In accordance with the chemistry of 1 , the phenoxy substituent in -17 wss readily exchanged by acidcatalyzed elimination-hydration,
yielding g.
functionality and stereochemistry
late in the actual synthesis.
We have also observed that y-la&ones
This mechanism thus would establish correct
C9
8 and 11 furnished the corresponding hydroxy hydroxamlc
acids when stirred with excess hydroxylamine
in THF.
Lossen rearrangement of derivatives such as
14 is also under investigation as we attempt to apply this convergent coupling strategy to the total synthesis of maytansine.
~endl
We wish to thank the National Institutes of Health (Grant #CA-l9686
National Cancer Institute) and Eli Lilly & Company for generous financial support.
from the
2838
No. 33
REFERENCES AND NOTES
*
--
Part I: J. E. Foy and B. Gsnem, Tetrahedron
Lett.,
775 (1977)
1.
S. M. Kupchan, A. R. Branfman, A. T. Sneden, A. K. Varma, R.G. Dailey, Jr., Komoda, Y. Nagao, J. Amer. Chem. See., s 5294 (1975)
2.
For other approaches see (a) A. L Meyers, C. Shaw, Tetrahedron Lett., (b) E.J. Corey, M.G. Bock, @&I. , 2643 (1975) (c) O.E. Edwards, P. Ho, Can. J. Chem., 5& 371 (1977)
3.
H.O. L. Fischer,
4.
(a) Y.S. Kiausner, M. Bodansky, Synthesis, 549 (1974) (b) M. J. Miller, G.M. Loudon, J. Amer. Chem. Sot., 97_, 5295 (1975)
5.
A.M.
6.
The structures of these and other intermediates were confirmed by ir, pmr, and maes spectral data. Detailed information will be provided upon request.
7.
The identity of these isomers was established using pmr shift reagents.
a.
For recent progress in this area see W. A. KIeechick, C. T. Buse, C. H. Heathcock, J. Amer. Chem. SDC., z, 247 (1977)
G. Dangschat, Chem. Ber.,
Touzin, Tetrahedron Lett.,
Y.
717 (1974)
65, 1009 (1932)
1477 (1975)