The solution to a deep stereochemical conundrum: Studies toward the tetrahydroisoquinoline alkaloids

Article abstract of DOI:10.1002/anie.200503982

(Chemical Equation Presented) Facile construction of subunits 1 and 2 allowed the rapid assembly of the pentacyclic ring system 3 of the cytotoxic tetrahydroisoquinoline alkaloids. A surprising participation reaction was discovered, which necessitated rev

Full text of DOI:10.1002/anie.200503982

Angewandte
Chemie
Alkaloid Synthesis
DOI: 10.1002/anie.200503982
The Solution to a Deep Stereochemical
Conundrum: Studies toward the
Tetrahydroisoquinoline Alkaloids**
Collin Chan, Shengping Zheng, Bishan Zhou,
Jinsong Guo, Richard M. Heid, Benjamin J. D. Wright,
and Samuel J. Danishefsky*
Dedicated to Professor Thomas Katz
Some years ago we initiated a program directed at the total
synthesis of ecteinascidin 743 (1),[1] an extremely potent
antitumor agent.[2] A total synthesis of 1 was reported by
[*] C. Chan, Dr. S. Zheng, Dr. B. Zhou, Dr. J. Guo, Dr. R. M. Heid,
B. J. D. Wright, Prof. S. J. Danishefsky
The Department of Chemistry
Columbia University
Havemeyer Hall, New York, NY 10027 (USA)
Fax: (+1)212-772-8691
E-mail: s-danishefsky@ski.mskcc.org
Prof. S. J. Danishefsky
The Laboratory for Bioorganic Chemistry
Sloan-Kettering Institute for Cancer Research
1275 York Avenue, New York, NY 10021 (USA)
[**] This work was supported by the National Institutes of Health (Grant
HL25848) and by Pharmamar Corporation of Madrid, Spain. We
thank Bristol Myers Squibb for the generous fellowship awarded to
C.C. B.J.D.W. is grateful for an NSF predoctoral fellowship. We thank
Professor Gerard Parkin for X-ray crystallographic analyses and Dr.
Yasuhiro Itagaki (Columbia University) for mass spectrometric
analyses.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 1749 –1754
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1749

Communications
E. J. Corey and co-workers in 1996.[3a] A subsequent synthesis
of 1 was described by Fukuyama and co-workers.[3b,c]
A
particularly elegant synthesis of the structurally related
saframycin was disclosed by Myers and Kung.[4]
Although there are many fascinating features in the Corey
total synthesis, the one that best establishes the context for
our work is captured in transformation 2!3 (Scheme 1). The
Scheme 2. Synthetic plan involving key lynchpin Mannich cyclization.
isolated from various sources and synthesized in a variety of
different ways, were of the 3,11-syn series.[8] On this basis, it
was supposed that in the pentacyclic setting, the 3,11-syn
series is inherently more stable than the corresponding anti
compounds. We further anticipated that the foregoing logic
could be applied either to ecteinascidin 743 or to the broader
family of saframycins. In the latter targets, the E-rings are
hexasubstituted.
At the time, the AB system housing the future C1 (cf. 8,
Scheme 4) was assembled by asymmetric dihydroxylation.[2]
We also attempted the synthesis of N-methyltyrosine deriv-
ative 12 by reagent-driven asymmetric synthesis (Scheme 3).
In particular, compound 9[2] was converted into epoxide 10
with high enantiomeric excess by Sharpless epoxidation.
Azidolysis of this epoxide provided an azidodiol, presumed to
be 11, from which we obtained an N-methyl-N-t-Boc amino
acid accordingly formulated as the l-amino acid derivative 12.
(Parenthetically, we had been drawn to this scheme as one
that connects the powerful and reliable Sharpless epoxidation
reaction with the important goal of synthesizing non-natural
amino acids and their derivatives).[9]
Coupling of 8 with 12 provided an amide, then assigned to
be 13 (Scheme 4), which was converted into the presumed
aldehyde 14. Indeed, treatment of this substance, as shown,
gave rise to a pentacyclic compound, formulated as 15. The
assignment of syn stereochemistry between C3 and C11
followed very closely from the coupling constant between 3-H
and 11-H (J ꢀ 3.6 Hz). As 13-H was presumably of the l-
amino acid configuration, 11-H was also defined to be a.
Given the syn relationship between 3-H and 11-H, the former
was accordingly assigned as a. At the time, these results
seemed rather secure and most promising for our total
synthesis program.
Scheme 1. Corey’s oxidation strategy to functionalize C4.
important message in these structures was that Corey and co-
workers were able to build the entire saframycin-like[5]
platform of 1 without carrying prebuilt functionality at C4.
Remarkably, the oxidation with phenylselenenic anhydride
shown in Scheme 1 served to functionalize C4 (presumably as
the b-carbon atom of an intermediate quinomethide), setting
the stage for Michael-like cyclization via the cysteinic sulfur.[6]
Regarding the ecteinascidin total synthesis problem, we
focused on a progression in which useful functionality would
already be in place at C4 as part of the unveiling of the
pentacyclic platform. More specifically, the plan we had in
mind would feature a lynchpin Mannich cyclization through
which an aldehyde function (C11) would be interpolated
[7]
between N12and C3 (Scheme 2).
In our initial conceptual
model, C3 would participate as the b-carbon atom of an enol.
For this plan to be practical, the relationship between C1 and
C13 would be secured by joining two appropriately matched
subunits (cf. 4 and 5) to reach reactive intermediate 6 and
thence pentacyclic product 7.
Needless to say, two additional stereogenic centers, C3
and C11, would be introduced during the lynchpin Mannich
cyclization. The configuration at C11 would be strictly
governed by the stereogenicity already in place at C13 (i.e.
the 11-H and 13-H necessarily emerge as syn in the
pentacyclic compound 7). The configuration likely to
emerge at C3 was less predictable. From the matched
configurations shown in 6, there could arise product 7a
corresponding to a syn saframycin skeleton, while the
alternate possibility 7b could be termed an anti saframycin.
At that time, all known saframycins, natural or synthetic,
We then began investigations to use this chemistry to
reach the saframycins as well as their closely related
derivatives, most notably ecteinascidin 743. The ultimately
successful results of that foray are described in detail in the
following manuscript in this issue.[10] However, before the
route to the ecteinascidin 743 series was completed, it was
necessary to explore the requirements for the lynchpin
Mannich cyclization in considerable detail. This study
1750
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1749 –1754

Angewandte
Chemie
Scheme 3. a) (d)-DET (8%), Ti(OiPr)4 (5.6%), tBuOOH (2 equiv),
molecular sieves (4 Š), ꢁ208C, 24 h, 98% (95% ee); b) Ti(OiPr)2(N3)2
(3.5 equiv), benzene, 808C, 5 h, 76% (single isomer); c) (MeO)2CMe2/
acetone (1:2), cat. p-TsOH·H2O, 258C, 10 min, 100%; d) H2 (1 atm),
Pd/C (10%), EtOAc, (Boc)2O (1.2 equiv), 258C, 5 h, 100%; e) TBAF
(1.5 equiv), THF, 258C, 1 h, 99%; f) PMBCl (1.2 equiv), NaH
(2 equiv), cat. TBAI, THF/DMF (5:1), 08C!258C, 5 h, 98%; g) MeI
(excess), NaH (5 equiv), THF/DMF (5:1), reflux, 12 h, 93%;
Scheme 4. a) BOPCl (1.1 equiv), Et3N (2.5 equiv), CH2Cl2, 258C, 12 h, 63%;
b) DMP (1.5 equiv), CH2Cl2, 258C, 30 min, 78%; c) DDQ (1.5 equiv), CH2Cl2/
pH 7.00 buffer solution (18:1), 258C, 3 h, 84%; d) DMP (1.5 equiv), CH2Cl2,
258C, 30 min, 85%; e) formic acid, 1008C, 5 h, 75%. BOPCl=bis(2-oxo-3-
oxazolidinyl)phosphinic chloride; DMP=Dess–Martin periodinane; DDQ=2,3-
dichloro-5,6-dicyano-1,4-benzoquinone.
h) 1) AcOH (80%), 258C, 12 h, 2) KMnO4 (0.2 equiv), NaIO4
(4 equiv), Na2CO3 (0.5 equiv), dioxane/H2O (2.5:1), 258C, 10 h, 95%.
DET=diethyl tartrate; Ts=p-toluenesulfonyl; Boc=tert-butoxycar-
bonyl; TBAF=tetrabutylammonium fluoride; PMB=p-methoxybenzyl;
TBAI=tetrabutylammonium iodide; DMF=N,N-dimethylformamide.
revealed some major unexpected teachings. An account of
that learning experience is provided herein.
Turning to the ecteinascidin 743 problem, we ultimately
required a penta- rather than hexasubstituted E-ring system
(such as was the case with 12). Accordingly, 16 was prepared
in much the same fashion as had been previously used to
prepare 10 (Scheme 5). Azidolysis of 16 produced 17 and
thence, pentasubstituted amino acid 18. Coupling of 18 with 8
gave an amide, which after suitable processing exposed the
required C4 ketone and C11 aldehyde combination (cf. 19).
Cleavage of the Boc group triggered the Mannich closure,
giving rise to pentacyclic product 20. Once again, the
configuration at C1 was defined, and we relied on the fixed
l-amino acid configuration at C13 to enforce the a stereo-
chemistry at C11. Thus, it was most surprising that the
1H NMR spectrum of this compound indicated it to be of the
3,11-anti type (J3-H–11-H = 0 Hz). Accordingly, 3-H in the
pentacyclic product is b, as shown in 20, rather than a (cf.
15). Needless to say, we were quite surprised at this outcome.
Further serving to deepen the mystery was that the results
were strikingly clean. No 3,11-anti compound was found
starting from 12, whereas no 3,11-syn compound was found
starting from 18.
Scheme 5. a) MOMCl (1.2 equiv), DIPEA (1.5 equiv), CH2Cl2, 258C, 12 h, 80%;
=
b) BrCH2CH CHCH2OTBS (1.1 equiv), K2CO3 (1.5 equiv), CH3CN, 808C, 5 h,
95%; c) Me2NPh (1.1 equiv), toluene, 2108C, 12 h, 96%; d) MeI (excess), K2CO3
(1.5 equiv), CH3CN, 808C, 12 h, 87%; e) TBAF (1.5 equiv), THF, 258C, 1 h, 99%;
f) PivCl (1.1 equiv), pyridine/CH2Cl2 (1:20), 258C, 3 h, 99%; g) HCl (3n), THF/
IPA (2:1), 258C, 12 h, 99%; h) Et2AlCl (3 equiv), (CH2O)n (excess), CH2Cl2,
08C!258C, 12 h, 96%; i) (tBu2)Si(OTf)2 (1.2 equiv), 2,6-lutidine (1.5 equiv),
CHCl3, 08C, 3 h, 85%; j) DIBAL-H (2.5 equiv), CH2Cl2, ꢁ788C, 30 min, 94%;
k) (d)-DET (8%), Ti(OiPr)4 (5.6%), tBuOOH (2 equiv), molecular sieves (4 Š),
ꢁ208C, 24 h, 98% (95% ee); l) Ti(OiPr)2(N3)2 (3.5 equiv), benzene, 808C, 5 h,
76% (single isomer); m) (MeO)2CMe2/acetone (1:2), cat. p-TsOH·H2O, 258C,
10 min, 100%; n) H2 (1 atm), 10% Pd/C, EtOAc, (Boc)2O (1.2 equiv), 258C, 5 h,
100%; o) TBAF (1.5 equiv), THF, 258C, 1 h, 99%; p) BnBr (1.2 equiv), K2CO3
(1.5 equiv), CHCl3/MeOH (2:1), 258C, 12 h, 83% q) PMBCl (1.2 equiv), NaH
(2 equiv), cat. TBAI, THF/DMF (5:1), 08C!258C, 5 h, 98%; r) MeI (excess),
NaH (5 equiv), THF/DMF (5:1), reflux, 12 h, 93%; s) 1) 80% AcOH, 258C, 12 h,
2) KMnO4 (0.2 equiv), NaIO4 (4 equiv), Na2CO3 (0.5 equiv), dioxane/H2O (2.5:1),
258C, 10 h, 95%; t) 8 (1.2 equiv), BOPCl (1.1 equiv), Et3N (2.5 equiv), CH2Cl2,
258C, 12 h, 63%; u) DMP (1.5 equiv), CH2Cl2, 258C, 30 min, 78%; v) DDQ
(1.5 equiv), CH2Cl2/pH 7.00 buffer solution (18:1), 258C, 3 h, 84%; w) DMP
(1.5 equiv), CH2Cl2, 258C, 30 min, 85%; x) formic acid, 1008C, 5 h, 75%.
MOM=methoxymethyl; DIPEA=diisopropylethylamine; TBS=tert-butyldime-
thylsilyl; Piv=pivaloyl; IPA=isopropyl alcohol; Tf=trifluoromethanesulfonyl;
DIBAL-H=diisobutylaluminum hydride; Bn=benzyl.
The cyclizations of potential ecteinascidin precursors
were also studied (Scheme 6). It was found that coupling of
21, prepared in the same manner as 8, and pentasubstituted
amino acid 18, on further processing produced C4/C11 keto/
aldehyde precursor 22. Again, Mannich closure gave rise to
anti pentacyclic compound 23.
Angew. Chem. Int. Ed. 2006, 45, 1749 –1754
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1751

Communications
Scheme 6. a) 18 (1.2 equiv), BOPCl (1.1 equiv), Et3N (2.5 equiv), CH2Cl2, 258C,
24 h, 85%; b) DMP (1.5 equiv), CH2Cl2, 258C, 30 min, 83%; c) DDQ (1.5 equiv),
CH2Cl2/pH 7.00 buffer solution (18:1), 258C, 3 h, 90%; d) DMP (1.5 equiv),
CH2Cl2, 258C, 30 min, 94%; e) formic acid, 1008C, 10 min, 70%.
Taking the seemingly incontrovertible data at face value,
one would be obliged to conclude that the stereochemical
outcome of the cyclization reactions (i.e. whether the
relationship between 3-H and 11-H emerges as syn or anti)
is somehow a function of the degree of substitution of the
remote E-ring. Unlikely as this situation would be, this is
where the data first took us. Accordingly, we pursued the
matter in considerable detail. A variety of seco-aldehydes
with pentasubstituted E-rings were prepared and examined.
They all led to 3,11-anti compounds. At no point could we
enter the required 3,11-syn series with any compounds in
which the E-ring was pentasubstituted.
Of course, the heart of the anomaly rests on the implicit
assumption that the relationships between 1-H and 13-H
(anticipating ecteinascidin 743 numbering) are the same in 14
and 19. This assumption seemed justified enough, as we
synthesized the two compounds by identical routes. Indeed
the AB precursor 8, bearing the R configuration at C1, is
identical in the sequences leading to 14 and 19. As the routes
to the two non-identical amino acid derivatives 12 and 18
were the same, it seemed reasonable to assume that these
compounds were identically configured at the future C13
center. The solution to the conundrum required abandoning
of this seemingly reasonable, but in the end incorrect,
assumption.
Specifically, to solve the problem we continued to assume
the correctness of the original supposition (i.e. that in the case
of the pentasubstituted E-ring compounds such as 16,
azidolysis of the epoxide was the sole inversion event). This
indeed progressed to the l-amino acid precursor, as previ-
ously formulated (cf. 18). By contrast, in the hexasubstituted
case, we now postulated that a “hidden” inversion took place
in the context of the azidolysis event, such that the azidolysis
reaction had actually produced the d-amino acid derivative
(Scheme 7). Thus, the stereochemical assignment at C13 in 14
is not S, as shown, but rather R (see below).
The unusual behavior of the hexasubstituted system could
be a consequence of two factors. First, the trajectory for
azidolysis is significantly more hindered when the aryl ring is
fully substituted. Also, perhaps the additional ortho-alkoxy
substituent at C15 participates in a precedented but now
arcane intramolecular Ar1 participation reaction, generating a
methylated spirocyclohexadienone (cf. 24).[11] This participa-
tion results in a hitherto unnoticed first inversion around C13.
Azidolysis of the activated cyclopropane again with inversion
leads then to product 25, an azido precursor of a d-amino acid
Scheme 7. a) Ti(OiPr)2(N3)2 (3.5 equiv), benzene, 808C, 5 h, 76%
(single isomer); b) (MeO)2CMe2/acetone (1:2), cat. p-TsOH·H2O,
258C, 10 min, 100%; c) H2 (1 atm), Pd/C (10%), EtOAc, (Boc)2O
(1.2 equiv), 258C, 5 h, 100%; d) TBAF (1.5 equiv), THF, 258C, 1 h,
99%; e) PMBCl (1.2 equiv), NaH (2 equiv), cat. TBAI, THF/DMF (5:1),
08C!258C, 5 h, 98%; f) MeI (excess), NaH (5 equiv), THF/DMF
(5:1), reflux, 12 h, 93%; g) 1) AcOH (80%), 258C, 12 h, 2) KMnO4
(0.2 equiv), NaIO4 (4 equiv), Na2CO3 (0.5 equiv), dioxane/H2O (2.5:1),
258C, 10 h, 95%; h) 8, BOPCl (1.1 equiv), Et3N (2.5 equiv), CH2Cl2,
258C, 12 h, 63%; i) DMP (1.5 equiv), CH2Cl2, 258C, 30 min, 78%;
j) DDQ (1.5 equiv), CH2Cl2/pH 7.00 buffer solution (18:1), 258C, 3 h,
84%; k) DMP (1.5 equiv), CH2Cl2, 258C, 30 min, 85%; l) formic acid,
1008C, 5 h, 75%.
(cf. 26). Ultimately in the hexasubstituted series, eventual
processing leads to aldehyde 27, which bears the R stereo-
chemistry (i.e. a d-aminoacyl function) at C13. Cyclization
does indeed produce the 3,11-syn pentacyclic system, 28.
However, this system is one in which 11-H and 13-H are b,
whereas 3-H is also b. In other words, while 3-H and 11-H are
indeed syn, they, as well as C13, are each of the opposite
configuration to that required for the total synthesis of a
saframycin. In contrast, in the pentasubstituted epoxide 16,
azidolysis occurred as a single inversion event, as initially
assigned, leading ultimately to the l-amino acid aldehydes 19
and 22, as formulated above en route to the anti products 20
and 23.[12] With this insight, the conundrum would be solved.
We were able to bring very strong support in favor of this
hypothesis. The thought was that if one could synthesize a
precursor to the lynchpin Mannich cyclization in which the E-
ring was hexasubstituted but of the l-amino acid series, one
should generate a 3,11-anti compound comparable to 20 and
23. Of course, in principle this could be possible by repeating
the sequence that initially produced 26 but using the ent-10
epoxide, which would be available from the Sharpless
epoxidation methodology (using the l-tartrate instead of
the d-tartrate). However, as has been described, this is a very
1752
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1749 –1754

Angewandte
Chemie
lengthy route.[2] Thus we had incentive to develop a rather
more streamlined approach.
acid 34. An important milestone in substantiating the
hypothesis discussed above (Scheme 7) was reached when it
was found that 34 was, indeed, the enantiomer of 26, as shown
by optical rotation.[17]
The new route started with iodo aromatic precursor 29
(Scheme 8).[7] This compound was carried through a Jeffery–
With the groundwork now secure, we next asked the
question as to the nature of the lynchpin Mannich cyclization
in the case of the properly matched (i.e. (R)-C1; (S)-C13)
hexasubstituted precursor 36 generated in the usual way from
amide 35 (Scheme 8). Indeed, when 36 was subjected to the
conditions shown, anti-37 was obtained in 62% yield. Thus,
the degree of substitution in the E-ring is not determinative in
terms of formation of syn versus anti product. The degree of
substitution was only important in determining the stereo-
sense of azidolysis of the relevant epoxides (cf. 10 versus 16)
and thus the eventual handedness of the E-ring amino acid
derivatives (cf. 18 versus 26).
In summary, with the facts now fully documented, the
conundrum has been solved. In retrospect, we note a pleasing
consistency in the stereochemical sense of the lynchpin
Mannich cyclization (see matched precursor 36 and mis-
matched precursor 27). In each case cyclization of the
iminium species onto the C3–C4 enol occurs from the face
opposite to that of the resident C1 benzyloxymethyl group.[18]
With the hard-won teachings in hand, we were in a position to
focus on a solution aimed at a total synthesis of ecteinasci-
din 743. These results are described in the subsequent paper
in this issue.
Received: November 9, 2005
Keywords: antitumor agents · asymmetric synthesis ·
.
Mannich cyclization · natural products · total synthesis
[1] a) K. L. Rinehart, T. G. Holt, N. L. Fregeau, J. G. Stroh, P. A.
Keifer, F. Sun, L. H. Li, D. G. Martin, J. Org. Chem. 1990, 55,
4512; b) R. Sakai, K. L. Rinehart, Y. Guan, A. H.-J. Wang, Proc.
Natl. Acad. Sci. USA 1992, 89, 11456.
Scheme 8. a) See reference [6]; b) Rh[(cod)-(S,S)-Et-duphos]+TfOꢁ (0.008 equiv),
H2 (6.8 atm) CH2Cl2/MeOH (1:1), 258C, 72 h, 93%, 99% ee; c) LiOH (3 equiv),
MeOH/THF/H2O (1:3:1)), 08C!258C, 11 h, 93%; d) MeI (2.2 equiv), NaH
(3.3 equiv), THF), 08C!258C, 15 h, 82%; e) H2 (1 atm), Pd/C (10%), EtOAc,
258C, 2 h, 99%; f) Fremy salt (10 equiv), NaH2PO4 (2m, aqueous), acetone,
258C, 75%; g) Zn (400 equiv), AcOH (400 equiv), CH2Cl2, 258C, 95%; h) MeI
(5 equiv), Cs2CO3 (5 equiv), DMF, 258C, 75%; i) 8 (1 equiv), BOPCl (1.1 equiv),
Et3N (2.5 equiv), CH2Cl2, 258C, 12 h, 66%; j) DMP (1.5 equiv), CH2Cl2, 258C,
30 min, 78%; k) DDQ (1.5 equiv), CH2Cl2/pH 7.00 buffer solution (18:1), 258C,
3 h, 90%; l) DMP (1.5 equiv), CH2Cl2, 258C, 30 min, 95%; m) formic acid,
1008C, 5 h, 62%. cod=cycloocta-1,5-diene.
[2] a) B. Zhou, J. Guo, S. J. Danishefsky, Org. Lett. 2002, 4, 43; b) B.
Zhou, S. Edmondson, J. Padron, S. J. Danishefsky, Tetrahedron
Lett. 2000, 41, 2039; c) B. Zhou, J. Guo, S. J. Danishefsky,
Tetrahedron Lett. 2000, 41, 2043.
[3] Total syntheses of 1: a) E. J. Corey, D. Y. Gin, R. S. Kania, J. Am.
Chem. Soc. 1996, 118, 9202; b) A. Endo, A. Yanagisawa, M. Abe,
S. Tohma, T. Kan, T. Fukuyama, J. Am. Chem. Soc. 2002, 124,
6552; during the preparation of this manuscript, we became
aware of a completed synthesis of 1: c) J. Chen, X. Chen, M.
Bois-Choussy, J. Zhu, J. Am. Chem. Soc. 2006, 128, 87.
[4] A. G. Myers, D. W. Kung, J. Am. Chem. Soc. 1999, 121, 10828.
[5] For a review of these compounds, see: J. D. Scott, R. M.
Williams, Chem. Rev. 2002, 102, 1669.
[6] C. Cuevas, M. Pꢀrez, M. J. Martꢁn, J. L. Chicharro, C. Fernꢂndez-
Rivas, M. Flores, A. Francesch, P. Gallego, M. Zarzuelo, F.
de la Calle, J. Garcꢁa, C. Polanco, I. Rodrꢁguez, I. Manzanares,
Org. Lett. 2000, 2, 2545.
[7] For a previous example in the context of a total synthesis, see: C.
Chan, R. Heid, S. Zheng, J. Guo, B. Zhou, T. Furuuchi, S. J.
Danishefsky, J. Am. Chem. Soc. 2005, 127, 4596.
Heck reaction[13] and a subsequent asymmetric Knowles-type
hydrogenation as previously described, leading to the l-
amino acid system 18.[14] In this pentasubstituted E-ring series,
the amino acid produced by the Jeffery–Heck/hydrogenation
sequence and that from the Sharpless epoxidation/azidolysis
sequence were identical in every respect, including optical
rotation.[15]
To obtain the hexasubstituted aromatic products by this
new technology, we proceeded via pentasubstituted precur-
sors.[16] Removal of the benzyl group of 18 was followed by
oxidation to quinone 33 (Scheme 8). The latter, following
reductive bismethylation afforded hexasubstituted l-amino
[8] For an example of an anti renieramycin compound, see: J. W.
Lane, Y. Chen, R. M. Williams, J. Am. Chem. Soc. 2005, 127,
12684.
[9] a) Y. Gao, R. M. Hanson, J. M. Klunder, S. Y. Ko, H. Masa-
mune, K. B. Sharpless, J. Am. Chem. Soc. 1987, 109, 5765;
Angew. Chem. Int. Ed. 2006, 45, 1749 –1754
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1753

Communications
b) K. B. Sharpless, W. Amberg, Y. L. Bennani, G. A. Crispino, J.
Hartung, K.-S. Jeong, H.-L. Kwong, K. Morikawa, Z.-M. Wang,
D. Xu, X.-L. Zhang, J. Org. Chem. 1992, 57, 2768; c) M. Caron,
P. R. Carlier, K. B. Sharpless, J. Org. Chem. 1988, 53, 5185.
[10] See following Communication in this issue: S. Zheng, C. Chan, T.
Furuuchi, B. J. D. Wright, B. Zhou, J. Guo, S. J. Danishefsky,
Angew. Chem. 2006, 118, 1786; Angew. Chem. Int. Ed. 2006, 45,
1754.
[11] S. Winstein, R. Baird, J. Am. Chem. Soc. 1957, 79, 756.
[12] In light of our current hypothesis, optical rotation data taken for
both hexa- and pentasubstituted epoxides (cf. 10 and 16) and
their respective azidodiol products (cf. 11 and 17) should have, in
retrospect, seemed suspicious. Upon azidolysis of pentasubsti-
tuted epoxide 16 to azidodiol 17 a sign inversion, from + to ꢁ
took place [16: [a]D25 = + 6.95 (CHCl3, c = 1) and 17: [a]D25
=
ꢁ39.30 (CHCl3, c = 0.765)]. In contrast, in the hexasubstituted
series the sign did not change [10: [a]2D5 = + 1.85, (CHCl3, c = 1)
and 11: [a]2D5 = + 25.31, (CHCl3, c = 1)]. However, in the end,
unequivocal support of our hypothesis was obtained through a
crystal structure of anti pentacyclic compound de-Bn-20, fully
verifying its assigned structure.
[13] a) T. Jeffery, J. Chem. Soc. Chem. Commun. 1984, 12 87; b) A.
McKillop, R. J. K. Taylor, R. J. Watson, N. Lewis, Synthesis 1994,
31.
[14] M. J. Burk, J. E. Feaster, W. A. Nugent, R. L. Harlow, J. Am.
Chem. Soc. 1993, 115, 10125.
[15] [a]2D5 = ꢁ75.29 (c = 1, CHCl3).
[16] Attempts to perform a Jeffery–Heck reaction on a hexasubsti-
tuted precursor of the type 30 were unsuccessful. The chemistry
was thus initially done at the pentasubstituted stage, and the
sixth substituent was introduced later.
[17] 34: [a]2D5 = ꢁ45.7 (CHCl3, c = 1), 26: [a]D25 = + 45.9, (CHCl3, c =
1).
[18] Thus, attack at the C11 iminium center necessarily occurs anti to
13-H, regardless of whether the latter is a or b. The attack at C3
occurs anti to the benzyloxymethyl group. Thus, if X = CH2OBn,
attack occurs on the a face of C3 (3-H b in the product). If Y=
CH2OBn, attack occurs at the b face of C3 (leaving the C3
proton a in the product).
1754
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1749 –1754