The brosimum allene: A structural revision

Article abstract of DOI:10.1021/ol802338z

(Chemical Equation Presented) Insight derived from a synthetic model, calculated ¹³C NMR data, and comparison to experimental data indicate that the proposed allenic structure A, originally assigned to an isolate from Brosimum acutifolium Huber

Full text of DOI:10.1021/ol802338z

ORGANIC
LETTERS
2008
Vol. 10, No. 23
5493-5496
The Brosimum Allene: A Structural
Revision
Gaojie Hu,† Kai Liu,† and Lawrence J. Williams*
Department of Chemistry and Chemical Biology, Rutgers, The State UniVersity of New
Jersey, Piscataway, New Jersey 08854
ljw@rutchem.rutgers.edu
Received October 9, 2008
ABSTRACT
Insight derived from a synthetic model, calculated 13C NMR data, and comparison to experimental data indicate that the proposed allenic
structure A, originally assigned to an isolate from Brosimum acutifolium Huber, should be revised to B, a natural product and nonallenic
substance, mururin C.
There are over 150 known allene-containing natural prod-
ucts.1 Among these, the structure assigned to a component
from the bark of Brosimum acutifolium Huber is arguably
the most provocative (A, see above).2 We believe this
structure to be incorrectly assignedsindeed it seems doubtful
that A represents a molecular arrangement isolable under
standard conditionssand we provide evidence that this
substance is identical to mururin C (B).3 Accordingly, we
begin with a discussion of the structural assigment of A, then
turn to the interesting issue of the reactivity of this moiety,
and end with a structural revision.
assigned to the central allenic carbon of A. The central carbon
signal of allenes normally appears near 200 and often at about
2105 (e.g., 200 for 1,6 217 for 2,6 Figure 1). There are
noteworthy exceptions. Whereas this signal appears at 189
ppm for difluoroallene 3,7 the corresponding signal for the
tetramethoxy derivative 48 appears at 114 and tetrafluoro-
allene 56 appears at 118. Interestingly, related cumulenes,
(4) The two-dimensional NMR methods used in ref 2 were not
disclosed.
(5) For representative examples, see: (a) Charrier, C.; Dorman, D. E.;
Roberts, J. D. J. Org. Chem. 1973, 38, 2644. (b) van Dongen, J. P. C. M.;
van Dijkman, H. W. D.; de Bie, M. J. A. Rec. TraV. Chim. 1974, 93, 29.
(c) Stephany, R. W.; de Bie, M. J. A.; Drenth, W. Org. Magn. Reson. 1974,
6, 45. (d) Krudy, G. A.; Macomber, R. S. J. Org. Chem. 1978, 43, 4656.
(e) Janssen, R. H. A. M.; Lousberg, R. J. J. C.; de Bie, M. J. A. Rec. TraV.
Chim. 1981, 100, 85. (f) Dueker, A.; Szeimies, G. Tetrahedron Lett. 1985,
26, 3555. (g) Kanda, T.; Ando, Y.; Kato, S.; Kambe, N; Sonoda, N. Synlett
1995, n/a, 745. (h) Kobayashi, S.; Nishio, K. J. Am. Chem. Soc. 1995, 117,
6392. (i) Bellavia-Lund, C.; Gonzalez, R.; Hummelen, J. C.; Hicks, R. G.;
Sastre, A.; Wudl, F. J. Am. Chem. Soc. 1997, 119, 2946. (j) Liebeskind,
L. S.; Pena-Cabrera, E. Org. Synth. 2000, 77, 135. (k) Miao, W.; Chung,
L. W.; Wu, Y. D.; Chan, T. H. J. Am. Chem. Soc. 2004, 126, 13326.
(6) Steur, R.; van Dongen, J. P. C. M.; de Bie, M. J. A.; Drenth, W.; de
Haan, J. W.; van de Ven, L. J. M. Tetrahedron Lett. 1971, 12, 3307.
(7) Zens, A. P.; Ellis, P. D.; Ditchfield, R. J. Am. Chem. Soc. 1974, 96,
1309.
The original structure assigment of the Brosimum allene,
reported in 2000 by Takashima et al., was based on HRMS
and 1H and 13C NMR data, including two-dimensional NMR
methods.4 Interestingly, the 13C NMR signal at 139 ppm was
† These authors contributed equally to this work.
(1) For an excellent overview, see: Krause, N.; Hoffmann-Roeder, A.
Modern Allene Chemistry; Krause N., Hashmi A. S. K., Eds.; Wiley-VCH:
Weinheim, 2004; p 997.
(2) Takashima, J.; Asano, S.; Ohsaki, A. Tennen Yuki Kagobutsu
Toronkai Koen Yoshishu 2000, 42, 487.
(3) Takashima, J.; Asano, S.; Ohsaki, A. Planta Med. 2002, 68, 621.
10.1021/ol802338z CCC: $40.75
Published on Web 11/12/2008
2008 American Chemical Society

e.g., quinoethylenes (Figure 1), are known and highly
unstable in most cases.9 13C NMR data for quinoethylenes
have not been reported.
Table 1. Computed and Experimental 13C NMR Signals
position
A(expt)
A(calc)a
B(calc)a
B(expt)
C(calc)a
1
2
3
4
5
6
7
8
9
1′
2′
3′
4′
5′
6′
7′
120.2
105.7
150.1
168.0
147.3
111.5
139.1
117.5
188.5
141.6
110.1
150.1
142.7
128.5
114.3
33.6
105.8
106.1
149.3
171.5
148.3
101.3
229.3
107.7
182.8
131.7
108.5
144.1
141.5
117.3
122.4
36.2
119.2
118.4
145.6
161.8
142.0
105.7
140.5
116.5
179.3
138.8
113.2
141.9
138.7
128.6
114.3
35.9
120.0
105.7
150.1
168.0
147.3
111.5
139.1
117.5
188.5
141.6
110.1
150.1
142.7
128.5
114.3
33.6
121.7
104.1
143.7
146.5
143.8
108.4
132.4
133.9
171.2
137.4
116.3
144.0
141.3
129.0
111.9
36.1
Figure 1.
Allenic 13C NMR signals: (a) calculated chemical shifts
(see text); (b) the calculated shift varies with the C-OMe torsion
angle.
8′
9′
35.9
62.3
56.9
56.6
39.3
66.7
55.0
54.9
38.9
67.0
57.8
58.6
35.9
62.3
56.9
56.6
39.0
66.7
55.3
59.3
Our studies began with computational modeling of the 13
C
3-OMe
3′-OMe
NMR expected for A. Geometry optimizations were per-
formed with B3LYP [6-31G (2d, 2p)] and with HF [6-31G
(2d, 2p)].10 The spectral data were then calculated using
several methods, including B3LYP, mPW1PW91, and HF.11
Although the best fit was found for the B3LYP-optimized
structure with B3LYP computed signals, no data set matched
well. This set is shown along with the experimental data for
A in Table 1 (A(expt) and A(calc)). Comparison of the
observed signal assigned to C7 (139 ppm) to the computed
chemical shift for this carbon (229 ppm) is most noteworthy.
NMR signal prediction is as yet only approximate, but a
differential of 90 ppm is extreme. Moreover, and without
exception, all calculations predict the central allenic carbon
signal of A to be ∼230 ppm.12
a GIAO/B3LYP/6-31G (2d, 2p)//B3LYP/6-31G (2d, 2p).
allene of A is part of a functional array recognizable as an
elaborated p-quinonemethide (6, Scheme 1). The high
reactivity toward nucleophiles of molecules that house the
quinonemethide substructure is recognized as largly a
consequence of aromatization (6 f 7).13 It is not unreason-
able to anticipate this sort of transformation for cumulated
quinonemethides (8 f 9). For A, the presence of the
aldehyde should increase the reactivity of the system toward
nucleophiles. In this regard, the proposed structure of this
substance is most provocative.
The inability to computationally model this signal for A
is in contrast to estimates of the central carbon signals of
allenes 1 and 4 and, to a lesser degree, conformationally
dynamic 5. The calculated values of these allenes are given
in parentheses in Figure 1. It thus appeared that there could
well be a problem with the brosimum allene structural
assigment.
Scheme 1. Parent and Cumulated p-Quinonemethide
In a series of parallel investigations, we had set about
meeting the synthetic challenge implicit to structure A. The
(8) (a) Friebolin, H. Basic One- and Two-Dimensional NMR Spectros-
copy; Wiley-VCH: Weinheim, 2005; p 66. See also: (b) Saalfrank, R. W.;
Maid, H. Chem. Commun. 2005, 48, 5953.
(9) (a) Zecher, D. C.; West, R. J. Am. Chem. Soc. 1967, 89, 153. (b)
Koster, S. K.; West, R. J. Org. Chem. 1975, 40, 2300. (c) West, R.; Zecher,
D. C.; Koster, S. K.; Eggerding, D. J. Org. Chem. 1975, 40, 2295.
(10) All structures were fully optimized by analytical gradient methods
using the Gaussian 03 suites: (a) Frisch, M. J.; Trucks, G. W.; Schlege,
H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.,
Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; et al. Gaussian 03, Revision
C.02; Gaussian, Inc.: Wallingford, CT, 2004 (see the Supporting Information
for the full citation.)As indicated, density functional (DFT) calculations used
the exchange potentials of: (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
They also used the correlation functional of: (c) Lee, C.; Yang, W.; Parr,
R. G. Phys. ReV. B 1988, 37, 785.
Scheme 2 and Table 2 present key data obtained from a
synthetic model study. In principle, a phenol that contains a
suitably positioned leaving group can be induced to eliminate
it to give the brosimum allene arrangement (10 f 11,
Scheme 2).14 The silyl ether derived from aldehyde 12 was
subjected to the action of isopropylmagnesium chloride and
(11) See, for example: (a) Cimino, P.; Gomez-Paloma, L.; Duca, D.;
Riccio, R.; Bifulco, G. Magn. Reson. Chem. 2004, 42, 26. (b) Rychnovsky,
S. D. Org. Lett. 2006, 8, 2895.
(13) For lead references on quinonemethide structure and reactivity, see:
(a) Toteva, M. M.; Moran, M.; Amyes, T. L.; Richard, J. P. J. Am. Chem.
Soc. 2003, 125, 8814. (b) van de Water, R. W.; Pettus, T. R. R. Tetrahedron
2002, 58, 5367. (c) Merijan, A.; Gardner, P. D. J. Org. Chem. 1965, 30,
3965.
(12) See the Supporting Information for a complete list of computed
13C NMR signals for A.
5494
Org. Lett., Vol. 10, No. 23, 2008

Scheme 2
Table 2. Reaction of 14 and 15 in DMF
base
(equiv)
nucleophile
(equiv)
product
(yield, %)
entry triflamide
1
2
3
4
5
6
7
8
9
15
14
15
15
15
15
15
15
15
15
15
none
H2O (4)
no reaction
no reaction
16 (60)
K2CO3 (4) H2O (4)
K2CO3 (4) H2O (4)
K2CO3 (4) no added H2O 16 (59)
K2CO3 (4) H2O (12) 16 (60)
K2CO3 (4) H2O (10% v/v) 16 (55)
KOH (4)
KOH (1)
K2CO3 (4) PhOH (1)
K2CO3 (14) PhOH (12)
K2CO3 (4) PhSH (4)
H2O (4)
H2O (4)
16 (45)
16 (49)
18 (not formed)
18 (47)
19 (81)
10
11
room temperature, and no evidence of reaction was observed
over the course of 48 h (entry 1, Table 2). Similarly, 14 was
stable under basic conditions (entry 2). In contrast, over the
course of 20 h, 15 slowly formed 16 (48%, entry 3) under
conditions identical to those of entry 2. Mild acid hydrolysis
of 16 gave ketone 17 (89%) and 15 (92%) (Scheme 2, see
inset).17 Although the apparent rate was slightly different,
the exclusion of water from the reaction and the addition of
excess water did not substantially influence the yield of 16
(entries 4-6, Table 2). The use of either a slight or large
excess of potassium hydroxide also had only a marginal
impact (entries 7 and 8). As shown in entry 9, the use of a
near equivalent amount of phenol, instead of water, gave
only 16, with no evidence of 18. A large excess of base and
phenol, however, gave 18 in 47% yield. Moreover, use of
thiophenol gave 19 in 81% yield.
Scheme 3. Mechanistic Rationale
then oxidized to the corresponding ketone (13). Treatment
with N-phenyltriflimide15 gave 14. Importantly, formation
of vinyl sulfonamides related to 14 is known and thought to
proceed by way of onium species related to 8.16 Treatment
of 14 with TBAF gave 15 as an isolable and stable product.
The behavior of 15 under basic conditions is dependent
on solvent, base, and where relevant, added nucleophile
(16-19, Scheme 2). In neutral DMF/H2O, 15 was stable at
(14) There is a strong analogy between this proposal and quinoneme-
thide-forming eliminations, see for example ref 13. Moreover, oxidative
conversion of 10 (LG ) H) to 11 represents another potential route to this
moiety.
(15) McMurry, J. E.; Scott, W. J. Tetrahedron Lett. 1983, 24, 979.
(16) Potter, G. A.; McCague, R. J. Org. Chem. 1990, 55, 6184 The
major byproduct of this reaction (50%) is the adduct derived from THF
and then N-phenyltrifluoromethanesulfonamide addition to the TBS oxo-
carbenium analogue of 8 (see the Supporting Informationfor details and
characterization of this substance). The use of ether as solvent for this
reaction proved ineffective and gave 14 in 4% isolated yield.
A mechanistic rationale is depicted in Scheme 3. The
phenoxide derived from deprotonation of 15 could promote
(17) The quality of the DMF was found to be important, as old DMF
gave 17 directly (35-50%).
Org. Lett., Vol. 10, No. 23, 2008
5495

the loss of sulfonamide and give rise to allenic intermediate
20. Addition of nucleophiles to 20 would lead to the observed
products. The conversion of 15 to 20 is slow; thus, 15 is
present in relative excess to 20 in all cases (entries 3-11).
Depending on the nature of the other nucleophiles present,
unreacted 15 adds to 20 to give 16. According to this
rationale, water and hydroxide must not competitively add
to 20. Phenoxide adds slowly and is competitive at high
concentrations (compare entries 9 and 10), whereas ben-
zenethiolate gives 19 in high yield by rapid and competitive
addition to 20 (entry 11). A closely related pathway may
also be relevant. Radical anion 21 (T 22) may form from
rapid electron transfer from a suitable nucleophile to 20
followed by radical coupling to give the observed products.
This process should be fastest for benzenethiolate, slower
for the phenoxide derived from 15, and even slower for
phenoxide. The other nucleophiles used in this study would
not be good candidates for this pathway under these
conditions. In light of these data, as well as available data
on quinoethylenes and related compounds9 and by analogy
to quinonemethides,13 species like 20 may not be sufficiently
stable for observation and isolation under standard condi-
tions.18
Figure 2. Isomers A, B, and C numbered for comparison
signals for these compounds are given in Table 1 (B(calc)
and C(calc)). Compound C is not a known substance.
Compound B is mururin C, a natural product recently isolated
from B. acutifolium by Takashima et al.3,19 This structure
assignment was based on HRMS, 1H and 13C NMR,
including 1H-1H COSY, HMQC, and HMBC. The observed
13C NMR signals for B are also shown in Table 1. The
computed carbon signals for B match experiment and, most
importantly, the experimental spectral data for A and B are
identical (compare A(expt), B(calc), and B(expt)).20
The data converge on the following conclusions: the
brosimum allene isolate represented as A is incorrect.
Although an allene in a context such as A or 20 may form
as a transient species, it likely does not represent a molecular
arrangement that can be isolated under standard conditions.
The structure should be revised to B and does not include
allene functionality.
Based on the observations outlined above, we were led to
consider an isomeric structure for the proposed brosimum
allene and quickly arrived at benzofuran derivatives B and
C as possible alternatives (Figure 2). The computed 13C NMR
(18) These observations do not speak to issues of biosynthesis and do
not preclude analogous oxidative transformations (see ref 14).
(19) The relationship of A and B, although identical, is not obvious
from the liturature. See, for example, the literature related to A: (a)
Dembitsky, V. M.; Maoka, T. Prog. Lipid Res. 2007, 46, 328. (b) Maurya,
R.; Yadav, P. P. Nat. Prod. Rep. 2005, 22, 400. (c) Schumacher, D. D.;
Mitchell, C. R.; Rozhkov, R. V.; Larock, R. C.; Armstrong, D. W. J. Liq.
Chromatogr. Relat. Technol. 2005, 28, 169. (d) Rozhkov, R. V.; Larock,
R. C. AdV. Synth. Catal. 2004, 346, 1854. (e) Hoffmann-Roeder, A.; Krause,
N. Angew. Chem., Int. Ed. 2004, 43, 1196. (f) Rozhkov, R. V.; Larock,
R. C. Tetrahedron Lett. 2004, 45, 911. Conversly, no mention of A appears
in the mururin C (B) literature; see: (g) Rodrigues, E.; Mendes, F. R.; Negri,
G. Central NerVous System Agents in Med. Chem. 2006, 6, 211. (h)
Takashima, J.; Komiyama, K.; Ishiyama, H.; Kobayashi, J.; Ohsaki, A.
Planta Med. 2005, 71, 654. (i) Westcott, N. D.; Muir, A. D. Phytochem.
ReV. 2004, 2, 401. (j) Lee, K.-H.; Xiao, Z. Phytochem. ReV. 2004, 2, 341.
(k) Gao, S.; Feng, N.; Yu, S.; Yu, D.; Wang, X. Planta Med. 2004, 70,
1128. (l) Alcantara, A. F. de C.; Teixeira, A. F.; Felicio da S., I.; Batista de
A., W.; Pilo-Veloso, D. Quim. NoVa. 2004, 27, 371. (m) Takashima, J.;
Ohsaki, A. J. Nat. Prod. 2002, 65, 1843.
Acknowledgment. Financial support from Merck & Co.
is gratefully acknowledged.
Supporting Information Available: Synthetic methods
characterization data and computed 13C NMR data. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL802338Z
(20) Understandably, structural misassignments, despite our many
modern advantages, are not uncommon: Nicolaou, K. C.; Snyder, S. A.
Angew. Chem., Int. Ed. 2005, 44, 6012.
5496
Org. Lett., Vol. 10, No. 23, 2008