A Novel Allyl Transfer Coupled with a Grob Fragmentation

Article abstract of DOI:10.1002/asia.201500728

A novel acid-promoted rearrangement is disclosed. In the previously unknown transformation, an allyl group migrated to an in situ formed carbocation stabilized by an electron-rich aryl or heteroaryl group, resulting in a stereoselective intramolecular Grob fragmentation. The outcome of the rearrangement observed with an array of substrates can be satisfactorily rationalized using a working hypothesis with the aid of a six-membered transition state similar to those proposed for the anionic oxy-Cope or oxonia-Cope rearrangements, but involving only one instead of two double bonds.

Full text of DOI:10.1002/asia.201500728

DOI: 10.1002/asia.201500728
Communication
Allylation
A Novel Allyl Transfer Coupled with a Grob Fragmentation
Shao-Gang Li, Hui-Jun Chen, Yang-Yang Yang, Wen-Ju Wu, and Yikang Wu*[a]
Abstract: A novel acid-promoted rearrangement is dis-
closed. In the previously unknown transformation, an allyl
group migrated to an in situ formed carbocation stabi-
lized by an electron-rich aryl or heteroaryl group, resulting
in a stereoselective intramolecular Grob fragmentation.
The outcome of the rearrangement observed with an
array of substrates can be satisfactorily rationalized using
a working hypothesis with the aid of a six-membered
transition state similar to those proposed for the anionic
oxy-Cope or oxonia-Cope rearrangements, but involving
only one instead of two double bonds.
During our total synthesis[1a] of demethyl (C-11) cezomycin, the
most recent member in the pyrrol ether family of antibiotics,[1b]
it was desired to remove the TBS protecting group in the inter-
Figure 1. Desilylation of 1a led to 3 or 4a instead of 2; 4a could also result
mediate 1a to release a free OH group at the C-6 position
(Figure 1). As the functionalities in this simple compound are
not particularly unstable according to the existing knowledge,
the desilylation was expected to be smooth and clean. Howev-
er, to our surprise, when commonly employed reagents were
used such as nBu4NF, HF or HF·py (always led to a complex
product mixture) the reaction failed.
from 1b. TBS=tert-butyldimethylsilyl; Boc=tert-butoxycarbonyl.
surprisingly, the aldehyde 4a.[2] More surprisingly, the same al-
dehyde (4a) was also obtained from 1b under the same condi-
tions, in an even better yield (70%).
1
Judging from the clean H and 13C NMR spectra and the sig-
Therefore, we decided to employ para-toluenesulfonic acid
(pTsOH)/MeOH to achieve the desilylation in the absence of
any fluoride ions. Under such conditions the composition of
the reaction mixture was simpler. However, the only product
that could be isolated was 3 instead of the originally expected
alcohol 2. Judging from the reaction conditions employed, the
methoxy group at the C-6 in 3 was most likely resulting from
solvolysis of an intermediate carbocation at the C-6 formed
through either an acid-mediated dehydration of the expected
desilylation product 2 or direct removal of a TBSO from the C-
6 in 1a.
nificant optical rotation value (½aŠ22 = +21.5) for 4a, what oc-
D
curred here appeared to be a stereoselective rearrangement,
which does not have any apparent precedents to date (to the
best of our knowledge). Therefore, further investigations were
carried out to gain more information to understand this (then)
seemingly very peculiar transformation, such as the pre-requi-
site for the rearrangement (any other groups at the C-6 posi-
tion than the pyrrolyl) and the influence of the substrate chiral-
ity on the stereochemical outcome of the rearrangement.
To facilitate the study of the less understood reaction, we
began with seeking better reaction conditions for the forma-
tion of 4a. Using the more readily attainable 1b (the synthetic
precursor for 1a) we first performed the reaction in MeCN in
the absence of any added H2O. The same product 4a was also
obtained, but in only 20% yield. Then, we used Et2O as the re-
action solvent. In this case the yield of 4a pleasingly increased
to 85%. Comparable yields were also observed in CH2Cl2.
Therefore, either Et2O or CH2Cl2 was employed in all following
experiments.
Using MeCN-H2O to replace the MeOH in the reaction to ex-
clude the methanolysis indeed eliminated the formation of 3.
Nevertheless, the only isolatable product (in 50% yield) was,
[a] Dr. S.-G. Li, Dr. H.-J. Chen, Y.-Y. Yang, W.-J. Wu, Prof. Dr. Y. Wu
State Key Laboratory of Bioorganic and Natural Products Chemistry
Collaborative Innovation Center of Chemistry for Life Sciences
Shanghai Institute of Organic Chemistry
Chinese Academy of Sciences
345 Lingling Road, Shanghai 200032 (China)
E-mail: yikangwu@sioc.ac.cn
The reaction proceeded rather fast at ambient temperature.
The reaction was faster when more of the pTsOH acid catalyst
was added. When equal molar amounts of pTsOH were em-
ployed, the reaction was complete within just a few minutes.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/asia.201500728.
Chem. Asian J. 2015, 10, 2333 – 2336
2333
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Table 1. Summary of the results with substrates 1a–1 f.[a]
Table 2. Summary of the results with substrates 1g–1m.[a]
Entry
Substrate
Main Product (yield)
Entry
Substrate
Main Product (yield)
1
1
2
2
3
4
5
6
3
4[b]
5[b]
6[c]
7
No reaction ocurred
No reaction ocurred
[a] All reactions were performed in Et2O at ambient temperature in the
A complicated mixture
presence of pTsOH. TES=triethylsilyl.
[a] All reactions were performed in Et2O at ambient temperature in the
presence of pTsOH. [b] The starting material (1j or 1k) was essentially un-
changed. [c] Performed in CH2Cl2 instead of Et2O under otherwise the
same conditions.
The substrate scope, functional group compatibility, and the
influence of the stereochemistry of the substrates were next
examined. The pyrrolyl-containing substrates were tested first
(Table 1). With or without a protecting group on the C-4 OH,
1a–d all afforded 4a as the only isolatable product (Table 1,
entries 1–4). However, the yields varied between 46–85%,
thereby indicating that the group on the C-4 OH (i.e., H, Me,
TES or TBS) had a strong influence on the allyl migration. Com-
pound 1e, which had no substituent at the C-3 position, also
delivered an anti aldehyde (4e, Table 1, entry 5) as observed
with 1a–d, but the yield was only 60%. Compound 1 f, which
had a different relative configuration to that of 1b at C-4/C-5,
gave a different diastereomer 4 f (Table 1, entry 6).
The second group of substrates are shown in Table 2; these
substrates have either a furanyl or a thienyl group instead of
a pyrrolyl group on the C-6 position. For compounds 1g–i,
similar results (Table 2, entries 1–3) to those observed with 1a–
f were obtained. Again, the configuration of the newly formed
C-6 was controlled by the relative configurations of the C-4
and C-5 as in the pyrrolyl cases. It is interesting to note that
1h and 1i gave the same product 4h (Table 2, entries 2 and
3), indicating that the C-6 configuration of the product was
not controlled by the C-6 configuration of the substrate. This
was consistent with the indication for the formation of a carbo-
cation (planar) mentioned above.
unidentified side-products were formed) when Et2O, PhMe or
THF were used as the reaction solvent (Table 2, entries 4 and
5). However, when the reaction was run in CH2Cl2 or DMF, the
reaction outcome was completely different; a complicated
product mixture was obtained.
The thienyl-containing substrate (1l) also underwent the
allyl transfer smoothly, delivering the expected aldehyde 4l in
83% yield (Table 2, entry 6). However, without the C-3 substitu-
ent, the reaction mixture became very complicated (Table 2,
entry 7), in sharp contrast to the result with 1e.
The results with several other substrates are summarized in
Table 3. Among these, the indonyl-containing alcohols reacted
well, with or without the C-3 methyl group (Table 3, entries 1
and 2). If the C-6 aryl group was a phenyl group rather than
a heterocycle, desilylation occurred but no rearrangement took
place (Table 3, entries 3 and 4). However, when the phenyl ring
had a p-methoxy group (1r and 1s), the allyl transfer took
place easily even in the absence of a C-3 substituent, thereby
affording the same product (4r) regardless of the C-4 configu-
ration (Table 3, entries 5 and 6). In addition, 1t (either a 1:1.8
or a 1:2.1 mixture of the two C-4 epimers) afforded 4t (the an-
tipode of 4r) as expected (Table 3, entry 7), confirming that the
absolute configuration of 4 depended on the absolute config-
Substrates without the C-3 substituent (1j and 1k) were
next tested. Unlike their pyrrolyl counterparts, 1j and 1k were
essentially fully recovered (only traces/negligible amounts of
Chem. Asian J. 2015, 10, 2333 – 2336
www.chemasianj.org
2334
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Table 3. Summary of the results with substrates 1n–1t.[a]
Entry
Substrate
Main Product (yield)
1[b]
2
3
4
5[b]
6[b]
7[c]
Figure 2. A working hypothesis that may rationalize the outcome of the cou-
pled allyl transfer-Grob fragmentation. The “syn,syn” and “syn,anti” relative
configurations are referred to the C-3/C-4 and C-4/C-5. Note that the actual
mechanism is still to be established.
than OH) in an equatorial position (TS 4) is expected to be of
lower energy than that with the C-4 OH in an equatorial posi-
tion (TS 5); that the main product was 4r rather than 4s is
therefore still reasonable.
[a] All reactions were performed in Et2O at ambient temperature in the
presence of pTsOH. [b] Performed in CH2Cl2 instead of Et2O under other-
wise the same conditions. [c] See the Supporting Information
(Scheme S10).
To the best of our knowledge, the coupled allyl migration-
Grob fragmentation does not have any precedents in the liter-
ature. However, it does carry partial resemblance to each of
the Cope,[4] oxonia-Cope,[5] or anionic oxy-Cope[6] rearrange-
ment in one aspect or another. Compared with the known re-
arrangements, a novel feature of the present reaction is that
the allyl group is transferred to a carbocation (generated in
situ by an acid and stabilized by an aryl group), which is not
part of a double bond or an oxocarbenium ion (such as the
“-C=O+-” in the oxonia-Cope reaction). Besides, the Grob frag-
mentation[7] here is triggered off by an acid-induced carbocat-
ion, rather than an alkoxide generated by a strong base as in
the anionic oxy-Cope reaction. Thus, despite its partial similari-
ty to each of those known rearrangements, the reaction ob-
served in this study does not fall into any of the existing cate-
gories; it represents a new one of its own.[8]
uration of the C-5 of 1; the C-4 and C-6 configurations of
1 had no effects on the configuration of the product.
At the present stage no definite mechanism for the ob-
served allyl transfer is available. However, it is possible to ra-
tionalize most observations with the aid of a working model/
hypothesis. For those substrates of the “syn, syn” configura-
tions such as 1a–d, the observed results seem to be compati-
ble with a chair transition state (Figure 2, TS 1) with all sub-
stituents in equatorial positions. When the substrates have the
“syn, anti” configurations (such as 1 f), the corresponding chair
transition state (TS 3) would have two out of three substituents
in axial positions and thus is disfavored. The alternative (TS 2)
with only one out of three substituents in an axial position is
thus in operation, leading to syn products.[3]
In conclusion, a novel acid-promoted rearrangement has
been identified, in which an allyl group migrated to an in situ
formed carbocation stabilized by an aryl or heteroaryl group;
an electron-rich aryl or heteroaryl group was proven essential
for the occurrence of the rearrangement. The rearrangement
proceeded rather fast at ambient temperature, delivering an
optically active aldehyde with the absolute configuration at
The result with 1r deserves further explanation, because at
first sight, formation of 4r from 1r seems to have violated the
rules derived above. However, in the absence of the C-3 axial
substituent, the transition state (Figure 2, TS 4) may be less un-
stable than the disfavored transition state for the “syn,anti”
substrates (TS 3). Also, the chair with the C-5 CH3 (larger in size
Chem. Asian J. 2015, 10, 2333 – 2336
www.chemasianj.org
2335
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
123, 2450–2451; c) J. Nokami, M. Ohga, H. Nakamoto, T. Matsubara, I.
Hussain, K. Kataoka, J. Am. Chem. Soc. 2001, 123, 9168–9169; d) T. P.
Loh, C.-L. K. Lee, K.-T. Tan, Org. Lett. 2002, 4, 2985–2987; e) J. Nokami, K.
Nomiyama, S. M. Shafi, K. Kataoka, Org. Lett. 2004, 6, 1261–1264; f) R.
Jasti, C. D. Anderson, S. D. Rychnovsky, J. Am. Chem. Soc. 2005, 127,
9939–9945; g) C. S. Barry, N. Bushby, J. R. Harding, C. L. Willis, Org. Lett.
2005, 7, 2683–2686; h) L. Yang, G. He, R. Yin, L. Zhu, X. Wang, R. Hong,
Angew. Chem. Int. Ed. 2014, 53, 11600–11604; i) G. C. Tay, C. Y. Huang,
S. D. Rychnovsky, J. Org. Chem. 2014, 79, 8733–8749; j) R. Jasti, S. D.
Rychnovsky, J. Am. Chem. Soc. 2006, 128, 13640–13648; k) R. Jasti, S. D.
Rychnovsky, Org. Lett. 2006, 8, 2175–2178; l) M. L. Bolla, B. Patterson,
S. D. Rychnovsky, J. Am. Chem. Soc. 2005, 127, 16044–16045; m) J. E.
Dalgard, S. D. Rychnovsky, Org. Lett. 2005, 7, 1589–1591; n) J. E. Dal-
gard, S. D. Rychnovsky, J. Am. Chem. Soc. 2004, 126, 15662–15663;
o) S. R. Crosby, J. R. Harding, C. D. King, G. D. Parker, C. L. Willis, Org. Lett.
2002, 4, 577–580; p) Y. Zou, C. Ding, L. Zhou, Z. Li, Q. Wang, F. Schoen-
beck, A. Goeke, Angew. Chem. Int. Ed. 2012, 51, 5647–5651; Angew.
Chem. 2012, 124, 5745–5749; q) Y. Zou, H. Mouhib, W. Stahl, A. Goeke,
Q. Wang, P. Kraft, Chem. Eur. J. 2012, 18, 7010–7015.
the C-5 position (a to the aldehyde carbonyl group) un-
changed. The configuration of the C-6 (b to the carbonyl
group) of the product, however, depended on the relative con-
figurations of the substrate. In the presence of a substituent at
C-3, the relative configuration of the C-4 in the substrate could
affect the outcome of the rearrangement. For those substrates
without any substituent at the C-3 position, the configuration
of the C-4 practically had no influence on the reaction.
Although the precise mechanism of the newly identified re-
arrangement is still to be established, the obtained results can
be satisfactorily rationalized using a working model/hypothesis
with the aid of a six-membered transition state similar to those
proposed for the Cope related rearrangements. And despite its
partial resemblance to [3,3] sigmatropic rearrangements, this
new reaction is different in several aspects: (1) only one
double bond is involved, (2) the allyl group is transferred to
a carbocation stabilized by an aryl or heteroaryl group, rather
than an oxocarbenium (-C=O+-) as in the oxonia-Cope rear-
rangement,[9] and (3) the Grob fragmentation is not induced by
a strong base (as in the anionic oxy-Cope rearrangement) but
by an acid.[10]
[6] a) L. A. Paquette, Tetrahedron 1997, 53, 13971–14020 (a review); b) P.
Maurin, S.-H. Kim, S. Y. Cho, J. K. Cha, Angew. Chem. Int. Ed. 2003, 42,
5044–5047; Angew. Chem. 2003, 115, 5198–5201; c) D. A. Evans, A. M.
Golob, J. Am. Chem. Soc. 1975, 97, 4765–4766; d) L. A. Paquette, G. D.
Maynard, J. Am. Chem. Soc. 1992, 114, 5018–5027; e) L. Barriault, D. H.
Deon, Org. Lett. 2001, 3, 1925–1927; f) L. Liu, P. E. Floreancig, Angew.
Chem. Int. Ed. 2010, 49, 5894–5897; Angew. Chem. 2010, 122, 6030–
6033; g) E. A. Crane, K. A. Scheidt, Angew. Chem. Int. Ed. 2010, 49, 8316–
8326; Angew. Chem. 2010, 122, 8494–8505 (a mini review); h) A. J.
Bunt, C. D. Bailey, B. D. Cons, S. J. Edwards, J. D. Elsworth, T. Pheko, C. L.
Willis, Angew. Chem. Int. Ed. 2012, 51, 3901–3904; Angew. Chem. 2012,
124, 3967–3970; i) Y. Hu, R. L. Bishop, A. Luxenburger, S. Dong, L. A. Pa-
quette, Org. Lett. 2006, 8, 2735–2737; j) J. Yang, Y. O. Long, L. A. Pa-
quette, J. Am. Chem. Soc. 2003, 125, 1567–1574; k) L. A. Paquette, Y. R.
Reddy, G. Vayner, K. N. Houk, J. Am. Chem. Soc. 2000, 122, 10788–
10794; l) F. Haeffner, K. N. Houk, Y. R. Reddy, L. A. Paquette, J. Am. Chem.
Soc. 1999, 121, 11880–11884.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (21372248, 21172247) and Chinese Acade-
my of Sciences.
[7] For a recent review, see:M. A. Drahl, M. Manpadi, L. J. Williams, Angew.
Chem. Int. Ed. 2013, 52, 11222–11251; Angew. Chem. 2013, 125, 11430–
11461. Perhaps the present observations may be described as a homolo-
gous Grob fragmentation.
[8] Another way to relate the present reaction to known ones may be
Prins-pinacol rearrangement, where the cation adds to the CÀC double
bond as in a Prins reaction while the fragmentation at the C-4 partially
resembles that in a pinacol rearrangement.
Keywords: Aldol reaction · allylation · migration · polyketides ·
rearrangement
[1] a) S.-G. Li, Y.-K. Wu, Chem. Asian J. 2013, 8, 2792–2800; b) K. D. Klika,
J. P. Haansuu, V. V. Ovcharenko, K. K. Haahtela, P. M. Vuorela, R. Sillan-
paeae, K. Pihlaja, Z. Naturforsch. 2003, 58b, 1210–1215 (CAN
141:277376).
[2] For the configuration of 4a, see the Supporting Information.
[3] The anti/syn ratio for the pyrrols 4a and 4 f was estimated (by 1H NMR
spectroscopy) to be approximately 93:7 and 1:9, respectively, while the
corresponding ratio for furans 4g and 4h was approximately 9:1 and
1:3, respectively. Other products appeared to be predominated by only
one diastereomer. See also Ref. [9]).
[9] For further stereochemical evidence for the rearrangement, see the
Supporting Information.
[10] For related cyclizations via reaction of alkenes with carbocations, see:
a) M. R. Lambu, D. Mukherjee, RSC Adv. 2014, 4, 37908–37913; b) J. C.
Green, E. R. Brown, T. R. R. Pettus, Org. Lett. 2012, 14, 2929–2931.
[4] a) A. C. Cope, E. M. Hardy, J. Am. Chem. Soc. 1940, 62, 441–444; b) R. P.
Lutz, Chem. Rev. 1984, 84, 205–247; c) S. J. Rhoads, N. R. Raulins, Org.
React. 1975, 22, 1–252; d) A. M. Martín Castro, Chem. Rev. 2004, 104,
2939–3002; e) A. C. Jones, J. A. May, R. Sarpong, B. M. Stoltz, Angew.
Chem. Int. Ed. 2014, 53, 2556–2591.
Manuscript received: July 13, 2015
Revised: July 28, 2015
Final Article published: August 28, 2015
[5] a) S.-i. Sumida, M. Ohga, J. Mitani, J. Nokami, J. Am. Chem. Soc. 2000,
122, 1310–1313; b) T.-P. Loh, Q.-Y. Hu, L.-T. Ma, J. Am. Chem. Soc. 2001,
Chem. Asian J. 2015, 10, 2333 – 2336
www.chemasianj.org
2336
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim