A hypervalent iodide-initiated fragment coupling cascade of N-allylhydrazones

Article abstract of DOI:10.1002/anie.201100888

Highway to the hydrazone: A hypervalent iodide initiated cascade process enables the rapid union of an aldehyde, an allylic hydrazide, and an alcohol (see scheme; Pv=pivaloyl). This process affords a diverse range of functionalized ether adducts, while si

Full text of DOI:10.1002/anie.201100888

DOI: 10.1002/anie.201100888
Synthetic Methods
A Hypervalent Iodide-Initiated Fragment Coupling Cascade of
N-Allylhydrazones**
Kelly E. Lutz and Regan J. Thomson*
Reliable bond-forming reactions that enable the union of two
or more molecular fragments are essential for the efficient
and convergent assembly of complex natural products or
medicinal agents.[1] As part of a program aimed at developing
such reactions, we have been investigating the utility of N-
allylhydrazides as versatile chemical intermediates that allow
for high yielding fragment coupling by way of hydrazone
formation followed by a carbon–carbon bond-forming molec-
ular rearrangement.[2] Most recently, we reported a triflimide-
catalyzed rearrangement of N-allylhydrazones (the Stevens
[3,3] rearrangement)[3] that allows for a “traceless” bond
construction between two fragments.[2c] Prior to this develop-
ment, we reported an N-bromosuccinimide (NBS)-initiated
rearrangement that not only allowed for such fragment
assembly but also incorporated an additional bromide atom
(i.e., 1!4, Nuc = Br).[2b] We speculated that N-bromination,
followed by loss of bromide, initiated the cascade sequence
through diazoallene species 2 (Scheme 1). A [3,3] sigmatropic
rearrangement would afford diazonium ion 3, which would
react with bromide to produce the benzylic bromide 4 (Nuc =
Br).
to the coordinated ligand on the iodine atom, providing a
useful and powerful strategy to couple multiple species
together (i.e., an aldehyde, an allylhydrazide, and the
nucleophile).
We initiated our research efforts in this new area by
investigating the effects of the commercially available hyper-
valent iodine compounds, PhI(OAc)2 (PIDA) and PhI-
(OTFA)2 (PIFA; OTFA = trifluoroacetate), on the hydrazone
derived from the condensation of 2-naphthaldehyde and
methylallyl hydrazine (i.e., 5; Scheme 2). While PIDA gave
Scheme 2.
Scheme 1. Cascade sequences of N-allylhydrazones.
no desired product under the conditions explored, PIFA
provided trifluoroacetate 6 (X = OTFA) in 43% yield (Sche-
me 2A). This low-yielding result, which could not be
improved upon, provided initial evidence that hypervalent
iodides were able to promote rearrangements of N-allylhy-
drazones. It was during an investigation of various exogenous
nucleophiles that we ran the reaction between hydrazone 5
and PIFA, in the presence of methanol (10 equiv), and
observed formation of the ester 6, along with competitive
formation of ether 7 (Scheme 2B). While it was possible to
favor generation of the ether adduct by using methanol as the
solvent, this would limit the use of this method to readily
available alcohols, and would preclude the use of solid
alcohols or those that are part of a more complex fragment.
Therefore, we explored the use of PhI(OTf)2 as an initiator,[5]
reasoning that the much less nucleophilic triflate would not
compete with the alcohol for incorporation into the sub-
strate.[6] In the event, we found that exposure of hydrazone 5
to one equivalent of PhI(OTf)2 (formed in situ by the addition
of TMSOTf to iodosobenzene)[5] in the presence of methanol
We wished to widen the scope of this cascade sequence to
include other nucleophiles, and were especially intrigued by
the possibility of initiating the hydrazone oxidation (i.e., 1!2
in Scheme 1) with hypervalent iodine compounds (i.e., PhIX2
where X = OAc, OTFA, OTf, etc).[4] We anticipated that the
nucleophile in such a system might not necessarily be limited
[*] K. E. Lutz, Prof. R. J. Thomson
Department of Chemistry, Northwestern University
2145 Sheridan Rd, Evanston, IL 60208 (USA)
Fax: (+1)847-467-2184
E-mail: r-thomson@northwestern.edu
[**] This work was supported by Northwestern University (NU), Amgen,
and the National Science Foundation (NSF, CHE0845063). We
thank Colleen McGourty (NU) and Devon Mundal (NU) for early
contributions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100888.
Angew. Chem. Int. Ed. 2011, 50, 4437 –4440
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4437

Communications
(10 equiv) led to smooth formation of the desired ether 7 in
77% yield with no discernable trace of the corresponding
triflate (Scheme 2C).
Under the developed conditions, a number of ethers could
be formed using various alcohols and aldehydes (Table 1).
Using 2-naphthaldehyde as the basic test substrate, the
The process proved reliable for several additional aryl
aldehydes (i.e., Table 1, 10j–10m) and could also be con-
ducted on a,b-unsaturated aldehydes in good yield (i.e.,
Table 1, 10n and 10o). The use of electron-deficient aryl
hydrazones such as that derived from 4-bromobenzaldehyde
was possible, but produced low yields of the desired adducts.
Similarly, attempts to utilize saturated aldehydes met with
little success due to the particular instability of the inter-
mediate hydrazones.
Table 1: Aldehyde and alcohol variation in the PhI(OTf)2-initiated
cascade.[a]
We next investigated the use of different hydrazine
substrates, in order to establish protocols for stereoselective
alkene synthesis, and diastereoselective formation of vicinal
stereocenters (Table 2). As anticipated, hydrazones that
Table 2: Variation of the hydrazide component (Ar=2-naphthyl).[a]
10a (77%)
10d (54%)
10b (67%)
10e (57%)
10c (67%)
10 f (67%)
13a
13b
13c
(77%, 20:1 E:Z)
(81%, 20:1 E:Z)
(83%, 20:1 E:Z)
13d
13e
13 f
(80%, 20:1 E:Z)
(56%, 20:1 E:Z)
(71%, 13:1 E:Z)
10g (64%)
10j (68%)
10h (70%)
10k (72%)
10i (39%)
10l (70%)
13g
(47%, 4:1 d.r.)
13h
(69%, 7:1 d.r.)
13i
(32%, 20:1 d.r.)
[a] Yields of isolated product over two steps from ArCHO (8). E:Z ratios
determined by 1H NMR spectroscopy; d.r. refers to ratio of the syn
isomer to anti isomer.
possessed a substituent attached to the same carbon as the
nitrogen atom (i.e., Table 2, 13a–13e), provided the desired
ethers with > 20:1 selectivity for the E-isomer.[8] A trisub-
stituted alkene was formed with high levels of stereoselectiv-
ity and in good yield (Table 2, 13 f). In addition to stereose-
lective alkene formation, we showed that vicinal stereoarrays
could be produced with modest to good levels of stereose-
lectivity when the hydrazide fragment possessed a methyl
group at the alkene terminus (Table 2, 13g–13i). While the
efficiency of these last two-step couplings were somewhat
modest, substrate 13g was formed with the syn-isomer as the
major diastereomer,[9] which provided some useful insight
into the possible mechanism of this transformation (see
Scheme 3).
The mechanistic hypothesis outlined in Scheme 1 involves
the intermediacy of a benzylic diazonium ion, which under the
reaction conditions is likely to ionize and thus produce a
reactive carbocation. As a stereochemical probe for the
intermediacy of a carbocation, we prepared chiral nonracemic
hydrazone 12e (90:10 e.r., Ar= 2-naphthyl)[10] and exposed it
10m (73%)
10n (66%)
10o (55%)
[a] Yields of isolated product over two steps from aldehyde 8. The
intermediate hydrazone 9 was not purified prior to hypervalent iodide-
initiated rearrangement. OTf=trifluoromethanesulfonate.
process proved effective for incorporation of a number of
primary alcohols, including those possessing handles for
future manipulation. For example, one could envision that
substrates 10d and 10e (Table 1) could readily undergo ring-
closing metathesis to form cyclic ethers.[7] Branched alcohols
are well tolerated (10 f–10i), including cyclic alcohols.
Remarkably, both tert-butanol and 1-adamantol afforded
the corresponding ethers (Table 1, 10g and 10i), although 1-
adamantol gave a diminished yield over the two-step proce-
dure from 2-naphthaldehyde (Table 1, 10i). Attempted incor-
poration of phenols was unsuccessful, presumably due to
rapid oxidation and decomposition of the alcohol. Likewise,
the use of other oxidation prone nucleophiles such as amines,
led to no desired products.
4438
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4437 –4440

Scheme 4. Fragment coupling of natural products. Pv=pivoyl.
Scheme 3. Stereochemical probes for the reaction (Ar=2-naphthyl).
could be readily generated using the (ꢀ)-antipode of menthol
in the sequence, thereby demonstrating the stereochemical
diversity possible in just two steps using this new trans-
formation. Curiously, the latter reaction using (ꢀ)-menthol
proved less efficient than when (+)-menthol was used,
perhaps indicating a matched and mismatched situation
between the stereochemistry of each reacting partner.[16]
In conclusion, we have developed a unique hypervalent-
iodide-initiated cascade process that enables the rapid union
of an aldehyde, an allylic hydrazide, and an alcohol. A high
degree of selectivity is observed for the formation of
disubstituted alkenes, vicinal stereocenters, and in processes
involving chirality transfer within non-racemic substrates.
Complex “natural product-like” compounds may be synthe-
sized in only a few steps using this chemistry suggesting that
future applications to natural product synthesis or to diverse
libraries for drug discovery may be a possibility.[15]
to the reaction conditions (Scheme 3A). The product 13e was
produced as a racemate, providing evidence of an achiral
carbocation intermediate. While this experiment showed that
stereoinduction from the hydrazide fragment could not be
relayed directly to the newly formed oxygen stereocenter, we
prepared chiral non-racemic hydrazone 15 (90:10 e.r., Ar= 2-
naphthyl)[11] in order to test the prospect of stereochemical
transfer to the newly formed carbon sterocenter, and hence to
the oxygen center following diastereoselective alcohol incor-
poration (Scheme 3B). In the event, hydrazine 15 generated
the desired product (i.e., 19) as a single alkene isomer, in a 6:1
ratio of syn:anti diastereomers, but more importantly with
complete transfer of chirality (90:10 e.r.).[12]
ꢀ
This remarkable transformation forms a C C bond and a
C O bond, generates a stereodefined alkene, and produces
ꢀ
two new vicinal stereocenters with a high level of diastereo-
control and enantioselectivity. The absolute sense of stereo-
induction in the formation of ether 19 is consistent with
transfer of chirality through a chair-like transition state,[13]
possibly resembling 16. Formation of the trans alkene is also
consistent with such a chair-like transition state. Since our
Received: February 3, 2011
Published online: April 6, 2011
Keywords: cascade reactions · hydrazones ·
.
hypervalent iodide reagents · stereoselective synthesis
ꢀ
experimental results provide evidence for C O bond forma-
tion by an SN1 mechanism (see Scheme 3), it is likely that the
observed syn configuration is a result of the alcohol adding
opposite the alkene substituent within the A(1,3) minimized
carbocation 18.[14] This notion is supported by the observed
trend for enhanced diastereoselectivity as the alkene sub-
stituent inceases in size (see Table 2, compounds 14g, 14h,
and 14i).
[1] E. J. Corey, X.-M. Cheng, The Logic of Chemical Synthesis,
Wiley, New York, 1989.
[2] a) D. A. Mundal, J. J. Lee, R. J. Thomson, J. Am. Chem. Soc.
2008, 130, 1148 – 1149; b) D. A. Mundal, K. E. Lutz, R. J.
Thomson, Org. Lett. 2009, 11, 465 – 468; c) D. A. Mundal, C. A.
Avetta, Jr., R. J. Thomson, Nat. Chem. 2010, 2, 294 – 297.
[3] The thermal rearrangement of N-allylhydrazones was first
reported by Stevens and co-workers in 1973, see: R. V. Stevens,
E. E. McEntire, W. E. Barnett, E. Wenkert, J. Chem. Soc. Chem.
Commun. 1973, 662 – 663.
[4] For reviews regarding the chemistry of hypervalent iodide
reagents, see: a) R. M. Moriarty, R. K. Vaid, Synthesis 1991,
431 – 447; b) A. Varvoglis, The Organic Chemistry of Polycoor-
dinated Iodine, VCH, New York, 1992; c) V. V. Zhdankin, P. J.
Stang, Chem. Rev. 2008, 108, 5299 – 5358; d) V. V. Zhdankin, P. J.
Stang, Chem. Rev. 2002, 102, 2523 – 2584; e) P. J. Stang, V. V.
Zhdankin, Chem. Rev. 1996, 96, 1123 – 1178.
Finally, we wished to demonstrate that the reaction
sequence may be used to couple more complex fragments,
and in particular we were drawn to the notion of using this
chemistry for the union of two natural products (Scheme 4).
To this end, vanillan derivative 20 could be readily condensed
with enantioenriched hydrazine 21 to form the corresponding
hydrazone (not shown). Exposure of this hydrazone to
PhI(OTf)2 in the presence of (+)-menthol gave rise to the
complex “natural product-like”[15] molecule 22 in 62% yield
from aldehyde 20. Similarly, the diastereomeric compound 23
Angew. Chem. Int. Ed. 2011, 50, 4437 –4440
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4439

Communications
[5] a) N. S. Zefirov, S. O. Safronov, A. A. Kaznacheev, V. V. Zhdan-
[11] Prepared by using a variant of the Pd-catalyzed allylation
reported by Trost and Du Bois, see: B. M. Trost, S. Malhotra,
D. E. Olson, A. Maruniak, J. Du Bois, J. Am. Chem. Soc. 2009,
131, 4190 – 4191. See Supporting Information for full details.
[12] The relative and absolute configuration of 19 was determined
through a chemical correlation. See Supporting Information for
full details.
[13] Asymmetric induction in sigmatropic rearrangements was first
reported for the Cope rearrangement by Hill and Gilman, see:
R. K. Hill, N. W. Gilman, J. Chem. Soc. Chem. Commun. 1967,
619 – 620.
[14] For a review of allylic strain as a controlling element in chemical
transformations, see: R. W. Hoffmann, Chem. Rev. 1989, 89,
1841 – 1860.
[15] The term “natural product-like” is used in fields such as chemical
biology to describe scaffolds that have structural features related
to natural products, see: A. M. Boldi, Curr. Opin. Chem. Biol.
2004, 8, 281 – 286.
kin, Zh. Org. Khim. 1989, 25, 1807 – 1808.
[6] The effect of counter-ions in some hypervalent iodide-mediated
transformations has been reported, see: V. V. Zhdankin, R.
Tykwinski, B. Berglund, M. Mullikin, R. Caple, N. S. Zefirov,
A. S. Kozꢀmin, J. Org. Chem. 1989, 54, 2609 – 2612.
[7] G. C. Fu, R. H. Grubbs, J. Am. Chem. Soc. 1992, 114, 5426 – 5427.
[8] Double bond geometry was determined by analysis of 1H NMR
coupling constants. See Supporting Information.
[9] The syn configuration was determined through direct compar-
ison to an authentic sample prepared using the well established
syn-selective crotylation of aldehydes developed by Roush and
co-workers, see: W. R. Roush, K. Ando, D. B. Powers, A. D.
Palkowitz, R. L. Halterman, J. Am. Chem. Soc. 1990, 112, 6339 –
6348. See Supporting Information for full details. Substrates 13h
and 13i were assigned by analogy.
[10] The requisite hydrazine required for the formation of enan-
tioenriched hydrazone 12e was prepared through an enantiose-
lective a-amination of 3-methylbutanal according to the proce-
dure of List for related aldehydes, see; B. List, J. Am. Chem. Soc.
2002, 124, 5656 – 5657. See Supporting Information for full
details.
[16] Attempts to use enantiopure chiral alcohols to directly induce
ꢀ
stereoselectivity during the C O bond-forming step using
achiral hydrazone precursors only ever led to 1:1 mixtures of
diastereomers.
4440 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4437 –4440