Cyclic 2:1 and 1:2 aldehyde-to-acetone byproduct adducts in aldol reactions promoted by supported proline-incorporated catalysts

Article abstract of DOI:10.1002/ejoc.200801146

Significant amounts of cyclic byproducts of aldol addition with stoichiometry deviating from a regular 1:1 addition pattern were formed when the reaction of acetone with aromatic aldehydes was promoted by polymer-supported proline-incorporated catalysts.

Full text of DOI:10.1002/ejoc.200801146

SHORT COMMUNICATION
DOI: 10.1002/ejoc.200801146
Cyclic 2:1 and 1:2 Aldehyde-to-Acetone Byproduct Adducts in Aldol Reactions
Promoted by Supported Proline-Incorporated Catalysts
Lital Tuchman-Shukron,[a] Tzofit Kehat,[a] and Moshe Portnoy*[a]
Keywords: Supported catalysts / Organocatalysis / Domino reactions / Aldol reactions / Cyclizations
Significant amounts of cyclic byproducts of aldol addition
with stoichiometry deviating from a regular 1:1 addition
pattern were formed when the reaction of acetone with aro-
the context of the aldol reaction, are most probably formed
via a multistep domino mechanism.
matic aldehydes was promoted by polymer-supported pro- (© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
line-incorporated catalysts. These adducts, unprecedented in
Germany, 2009)
tion of aromatic aldehydes and acetone catalyzed by two
new polymer-supported proline-related catalytic systems. To
the best of our knowledge, cyclic adducts of stoichiometry
deviating from a 1:1 ratio have never been observed in the
proline-catalyzed ketone addition to aldehydes.
Introduction
Over the past decade a considerable number of experi-
mental work investigated the catalytic properties of proline
and proline-derived compounds in a variety of asymmetric
organic transformations, particularly the aldol addition.[1]
The majority of the catalytic studies were carried out with
soluble catalysts, whereas only a few involved hetero-
geneous proline-based moieties.[2] Despite the questionable
economic profitability of developing a successful immobi-
lized proline catalyst, because of the low price of this amino
acid and the possibility of its easy recovery from homogen-
eous reaction mixtures, proline and related molecules pro-
vide excellent models for organocatalyst immobilization
studies as a result of their simplicity and the ready availabil-
ity of various derivatives. Moreover, immobilization fre-
quently provides catalysts with an altered reactivity profile
that can lead to products different from those obtained in
solution and provide valuable mechanistic information. For
instance, the major focus of the investigation of the aldol
reaction of aldehydes and ketones was on the aldol adduct
formed with the 1:1 stoichiometry of the addition. Yet,
these aldol products usually possess additional nucleophilic
or electrophilic sites and can participate in the formation of
adducts of higher stoichiometric ratios. For instance, three
aldehyde molecules can combine into a carbohydrate-like
cyclic structure.[3] Linear 2:1 aldehyde/acetone adducts were
obtained when the reaction was promoted by pyrrolidine or
prolinethioamide in neat acetone.[4]
Results and Discussion
In the past we demonstrated that trans-4-hydroxyproline
could be immobilized on a dendronized solid support
through the hydroxy function, whereas the carboxylic and
amino functional groups are available for catalysis. These
supported dendritic catalysts promoted the asymmetric al-
dol addition of acetone to aromatic aldehydes with the yield
and selectivity comparable (or even slightly better) than
those displayed by -proline catalysts in solution.[5] An al-
ternative catalytic design involved immobilization of the
proline unit through the carboxylate moiety, while provid-
ing a replacement for the electrophile activator (formerly
the COOH group) in the form of remotely attached poten-
tially double hydrogen-bond-donating trisalkoxycarbonyl
guanidine unit. The synthesis of the potential catalyst
G0(Pro, Gua-Boc2) (6) is depicted in Scheme 1. It is based
on the immobilization of the heteroarm branching unit,
methyl 3-hydroxy-5-hydroxymethylbenzoate,[6] on its con-
version into monoprotected diol 2, and on the stepwise as-
sembly of the arms: first the trisalkoxycarbonyl guanidine
arm and then the proline arm.
In this article we report the formation of a significant
amount of 2:1 and 1:2 cyclic adduct byproducts in the reac-
Though we hoped that the carboxylic acid could be re-
placed by the double hydrogen-bond-donor moiety of 6, the
model reaction of acetone with benzaldehyde in the pres-
ence of 6 (30 mol-%) yielded only trace amounts of the al-
dol product. When a more active electrophile, such as 4-
nitrobenzaldehyde, was employed, the reaction yielded 33%
of a low optical purity aldol product (7%ee), but also an
additional compound in 20% yield (Scheme 2). The com-
[a] School of Chemistry, Raymond and Beverly Sackler Faculty of
Exact Sciences, Tel Aviv University,
Tel Aviv, 69978, Israel
Fax: +972-3-6409293
E-mail: portnoy@post.tau.ac.il
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
992
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 992–996

Cyclic 2:1 and 1:2 Aldehyde-to-Acetone Byproduct Adducts
the opposite effect. The addition of water (5 equiv.) or a
fivefold increase in the concentration of the aldehyde also
led to a decrease in the chemoselectivity with a concomitant
increase in the conversion. Whereas facilitation of the reac-
tion kinetics upon an increase in the concentration of the
reactants is hardly surprising, the decrease in the selectivity
toward the formation of the more “aldehyde-rich” adduct
under these conditions is less comprehensible and we will
subsequently provide a possible explanation.
Table 1. The reaction of aldehyde with acetone leading to 2:1 ad-
ducts.[a]
Entry
Catalyst
Yield of aldol
product [%][b]
Yield of 7
Chemoselectivity
[%, 7/(aldol + 7)]
[%][b]
1
6
6
6
6
6
32
15
47
44
82
72
35
76
86
39
20
18
trace
6
16
28
30
24
14
26
38
55
Ͻ4
12
16
28
46
24
14
40
2[c]
3[d]
4[e]
5[f]
6[g]
7
Scheme 1. Synthesis of bifunctional catalysts 6 and 6Ј. Reagents
and conditions: (a) methyl 3-hydroxy-5-hydroxymethylbenzoate,
LiH, TBAI, DMF, 60 °C, overnight; (b) TBSCl, Et3N, DMAP,
DCM, room temp., 3 h; (c) LiBH4, B(OMe)3, THF, 60 °C, over-
night; (d) p-nitrophenyl chloroformate, DIPEA, pyridine, THF,
room temp., 2 h; (e) bis(Boc)guanidine, HOBt, DMF, 50 °C, 24 h;
(f) TBAF, THF, room temp., 4 h; (g) Fmoc-Pro-OH, diisopro-
pylcarbodiimide (DIC), DMAP, DMF, room temp., 6 h; (h) piper-
idine/DMF (2:8), room temp., 2–3 min.
6
13
6Ј
6Ј
6
8
9[g]
10[h]
[a] Reaction conditions: p-nitrobenzaldehyde (1 mmol) and resin-
supported catalyst (0.3 mmol) in DMSO/acetone (4:1, 2 mL per
0.2 g of resin) at room temp. for 48 h. [b] Yield determined by
NMR spectroscopy. [c] Resin-supported catalyst (0.15 mmol).
[d] DMF instead of DMSO was used. [e] THF instead of DMSO
was used. [f] H2O (5 mmol) was added. [g] Aldehyde (5 mmol).
pound is significantly less polar than the aldol product and
could be purified chromatographically. On the basis of
NMR spectroscopic data and its comparison to literature
sources,[7] in addition to mass spectrometric data, the com- [h] p-Cyanobenzaldehyde instead of p-nitrobenzaldehyde was used.
pound was identified as cis-2,6-bis(4-nitrophenyl)tetra-
Catalyst 13, related to catalyst 6, was prepared as de-
picted in Scheme 3. The use of catalyst 13 in the model
reaction led to a slight improvement in the conversion and
a notable increase in the chemoselectivity towards 7. In con-
trast, 6Ј, a catalyst analogous to 6 but prepared on first-
generation dendronized chloromethyl-terminated resin (vs.
Wang Bromo PS, Scheme 1),[9] displayed significantly
higher activity, but much lower selectivity toward 7. p-Cya-
hydro-4H-pyran-4-one (7). Particularly, there was an excel-
lent correlation of the aliphatic carbon signals of 7 with the
meso (cis) isomer of 2,6-diphenyltetrahydro-4H-pyran-4-
one.[7a] During the preparation of the manuscript an inde-
pendent synthesis of cis-2,6-bis(4-nitrophenyl)tetrahydro-
4H-pyran-4-one was reported with spectroscopic data com-
pletely matching that of 7.[8] The meso isomer is, according
to the literature, the more thermodynamically stable iso-
mer.[7b] Formation of 7 without any visible trace of its trans
isomer may thus point to thermodynamic control of the
reaction sequence leading to the formation of the cyclic ad-
duct.
Scheme 2. The reaction of aldehyde with acetone leading to 2:1
adducts.
In an attempt to improve the chemoselectivity of the re-
action toward 7, we changed a number of reaction param-
eters and prepared two new catalysts related to 6 (Table 1).
Thus, a decrease in the amount of the catalyst from 30 to
15% led to a decrease in the overall conversion of the alde-
hyde, but improved the chemoselectivity significantly. In
Scheme 3. Synthesis of catalysts 12 and 13. Reagents and condi-
tions: (a) guanidine hydrochloride, DIPEA, DMF, 50 °C, 24 h; (b)
ethyl chloroformate, DIPEA, pyridine, THF, room temp., 2 h; (c)
TBAF, THF, room temp., 4 h; (d) Fmoc-Pro-OH, DIC, DMAP,
DMF, room temp., 6 h; (e) piperidine/DMF (2:8), room temp., 2–
contrast, the replacement of DMSO by DMF or THF had 3 min.
Eur. J. Org. Chem. 2009, 992–996
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
993

L. Tuchman-Shukron, T. Kehat, M. Portnoy
SHORT COMMUNICATION
nobenzaldehyde (replacing p-nitrobenzaldehyde) reacted
similarly with acetone in the presence of 6 to form cyclic
byproduct 14 along with the main aldol adduct.
As to the possible mechanism for the formation of 7, a
few important points described in the literature are worth
emphasizing. Compound 15, the product of the double al-
dol addition of aldehydes to acetone under aldol reaction
conditions, was observed on two occasions: with pyrrolidine
or prolinethioamide catalysts.[4] These acyclic 2:1 aldehyde/
The proposed mechanism for the multistep transforma-
acetone adducts were formed with electron-poor aromatic tion in our case is depicted in Scheme 4. Contrary to one
aldehydes only. In the case of pyrrolidine catalysis, a sub- of the examples in ref.[4b] and the reaction in ref.[8] it is
stantial amount of the byproduct (up to 50%) could be ob- highly unlikely that the reaction takes place stepwise with
tained, but its formation was suppressed by acidic additive/ the formation of the aldol (A) and linear double aldol (B)
water addition.[4a,4b] In the case of prolinethioamide cataly- adducts as intermediates. In other words, the catalyst does
sis, a 2:1 stoichiometric ratio of the aldehyde to acetone not undergo detachment and reattachment to the carbonyl
should be used to reach a 40% yield of the double aldol group along the route to 7. Rather, these two acyclic prod-
addition product.[4c] It should be noted that trace amounts ucts A and B are “byproducts” en-route to 7 that are
of 15a were also observed by us in the model reaction. For- formed by “shortcuts” when the proline-containing inter-
mation of acyclic double addition product 15e was also mediates undergo hydrolysis “too early” before the full
achieved by treating the regular aldol product with the cor- catalytic cycle is complete. Thus, reaction of 4-nitrobenzal-
responding aldehyde.[4b] Moreover, in the report that ap- dehyde with acetone in the presence of the aldol adduct of
peared whilst this article was in preparation, pyrrolidine benzaldehyde and acetone (A with Ar = Ph) did not lead to
was established as an effective catalyst for the stepwise oxa- 2-(4-nitrophenyl)-6-phenyltetrahydro-4H-pyran-4-one. Only
Diels–Alder-like reaction of acyclic β-aryl-α,β-unsaturated 7 and the two aldol products were detected at the end of
ketones with aldehyde.[8] As aforementioned, this reaction the reaction. The addition of water helps to “shortcut” the
produces 2,6-disubstituted tetrahydro-4H-pyran-4-ones, cycle through hydrolysis of intermediate III, which forms
such as 7.
aldol A and restores the catalyst, thus facilitating the rate
Scheme 4. Proposed mechanism for the formation of 2:1 aldehyde/acetone adducts.
994
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 992–996

Cyclic 2:1 and 1:2 Aldehyde-to-Acetone Byproduct Adducts
of the reaction in general but lowering the chemoselectivity
Two plausible routes to byproduct 19 can be envisioned.
toward C (7 in our case).
These are shown schematically in Scheme 6. Although both
The rate of consumption of the aldehyde is also higher routes are three-step transformations (two aldol condensa-
and the chemoselectivity toward 7 is lower when the reac- tions and one Michael addition, one of them being an intra-
tion is promoted by 6Ј (the first generation dendron) or molecular ring-closing step), this does not necessarily mean
when the initial concentration of the aldehyde is increased that the catalyst is detached from the intermediate after
fivefold. In contrast, the reduced reaction rate upon a de- each step. Rather, the routes are depicted in this way for
crease in the catalyst loading is associated with the in- simplicity. It is noteworthy that path b, the intramolecular
creased selectivity toward 7. This connection between the aldol cyclization of the 1,5-dione, is a well-documented re-
activity of the catalyst and its chemoselectivity can originate action and can also be promoted by proline.[13] The preced-
from two factors. First, it is possible that the formed aldol ing step, however, that is, the formation of the 1,5-dicar-
(A) facilitates the hydrolysis of the iminium cation, enables bonyl intermediate by Michel addition of the nonactivated
a “shortcut” of the more extended cycle, and promotes its carbonyl compound (acetone) to the nonactivated Michael
own formation. Second, it is also possible that the hydroxy acceptor (enone), is unprecedented in organocatalysis.[14]
group of the aldol disrupts the interaction of the trisalkoxy- Although highly speculative, route a cannot be ruled out at
carbonyl guanidine with the intermediates, which increases this stage.[15]
the preference for the shorter route leading to A at the ex-
pense of the cycles leading to higher adducts.
Although guanidine unit itself does not promote the re-
action, the crucial role of the potential double hydrogen-
bond donor site for the formation of the 2:1 cyclic adduct
is evidenced by two experimental observations. Lowering
the steric hindrance of the group (ethoxycarbonyl vs. tert-
butoxycarbonyl on the nitrogen atoms) has a positive influ-
ence on the chemoselectivity by raising it from 38 to 46%.
Secondly, catalysts without the trisalkoxycarbonyl guani-
dine unit, that is, those decorated with proline ester and
benzyloxycarbonyl guanidine (12, Scheme 3) or with pro-
line esters only (16–18), do not promote the formation of 7
Scheme 6. Proposed mechanisms for the formation of a 1:2 alde-
hyde/acetone adduct.
or analogous byproducts.
The synthesis of the Gn(OPro) catalysts (n = 0–2; 16–18)
as well as the detailed investigation of their catalytic activity
Conclusions
in the aldol reaction will be reported elsewhere.[10] However,
as mentioned above, linear or cyclic 2:1 aldehyde/acetone
We demonstrated that complex multistep transforma-
adducts have never been observed in the relevant reactions tions could be promoted by proline-derived heterogeneous
with these catalysts. Interestingly, another cyclic byproduct catalysts. Although formed as byproducts with only up to
adduct, 19, with the opposite stoichiometry of the compo- 30% yield, the cyclic adducts described here expose the po-
nents (1:2, aldehyde/acetone) was always observed with tential of relatively simple organocatalytic systems to form,
these catalytic systems in the reaction of benzaldehyde with in a one-pot reaction, products of high complexity that
acetone (Scheme 5). Although the nondendritic and den- could be valuable intermediates for many synthetic tar-
dritic first-generation catalysts (16 and 17) provided the by- gets.[13d] The mere demonstration of such reactivity (un-
product in only 3–5% yield, the yield reached a remarkable precedented before), will provide strong encouragement for
30% when the reaction was carried out with second-genera- catalytic chemists to seek systems that can perform these
tion catalyst 18.[11] Similar yields were obtained in DMF transformations in a more high-yielding and selective way.
and chloroform. The spectroscopic data (1H and 13C NMR)
of the byproduct matched that reported in the literature for
the compound.[12]
Experimental Section
General: The loading of the catalytic resins was determined by aci-
dolytic cleavage on the basis of the 1H NMR spectra of the cleavage
solution (TFA/CDCl3, 1:1) with 11 m C6H6 as an internal stan-
dard.
General Procedure for the Aldol Reaction: The catalytic resin
(0.3 mmol, 0.3 equiv.) was added to a mixture of DMSO/acetone
(4:1, 8 mL:2 mL). The suspension was stirred for 5 min at room
temperature and then the aldehyde (1 mmol, 1 equiv.) was added.
The suspension was mixed at room temperature for 48 h. The resin
was separated from the solution by filtration and washed with ethyl
acetate. Water (10 mL) and saturated aqueous NH4Cl solution
Scheme 5. The reaction of aldehyde with acetone leading to a 1:2
adduct.
Eur. J. Org. Chem. 2009, 992–996
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
995

L. Tuchman-Shukron, T. Kehat, M. Portnoy
SHORT COMMUNICATION
[2] M. Gruttadauria, F. Giacalone, R. Noto, Chem. Soc. Rev.
2008, 37, 1666–1688 and references cited therein.
[3] I. Ibrahem, W. Zou, Y. Xu, A. Córdova, Adv. Synth. Catal.
2006, 348, 211–222.
[4] a) C. Ji, Y. Peng, C. Huang, N. Wang, Y. Jiang, Synlett 2005,
986–990; b) C. Ji, Y. Peng, C. Huang, N. Wang, Z. Luo, Y.
Jiang, J. Mol. Catal. A 2006, 246, 136–139; c) D. Gryko, R.
Lipinski, Eur. J. Org. Chem. 2006, 3864–3876.
[5] T. Kehat, M. Portnoy, Chem. Commun. 2007, 2823–2825.
[6] K. Goren, T. Kehat, M. Portnoy, Adv. Synth. Catal., accepted.
[7] a) K. Ramalingan, K. D. Berlin, N. Satyamurthy, R. Sivaku-
mar, J. Org. Chem. 1979, 44, 471–477; b) C. A. R. Baxter, D. A.
Whiting, J. Chem. Soc. C 1968, 1174–1178.
[8] L.-Q. Lu, X.-N. Xing, X.-F. Wang, Z.-H. Ming, H.-M. Wang,
Tetrahedron Lett. 2008, 49, 1631–1635.
(10 mL) were added to the combined filtrate. The mixture was ex-
tracted with ethyl acetate (3ϫ10 mL). The organic phase was dried
on MgSO4. The solvent was evaporated, and the crude material
was analyzed to determine conversion and yield, and then chro-
matographed on a silica gel column (EtOAc/hexanes, 1:9, up to
EtOAc/hexanes, 3:7) to yield the pure adducts.
7: 1H NMR (400 MHz, CDCl3): δ = 8.29 (d, J = 8.6 Hz, 4 H), 7.63
(d, J = 8.6 Hz, 4 H), 4.99 (dd, J = 2.1, 11.8 Hz, 2 H), 2.84–2.80
(dd, J = 2.1, 16.0 Hz, 2 H), 2.69–2.62 (dd, J = 11.8, 16.0 Hz, 2 H)
ppm. 13C NMR (100 MHz, CDCl3): δ = 203.1, 147.8, 146.9, 126.3,
124.1, 77.9, 49.0 ppm. MS (FAB): m/z = 343.1 [M + H]+.
19: 1H NMR (400 MHz, CDCl3): δ = 7.27–7.34 (m, 2 H), 7.20–
7.24 (m, 3 H), 5.95 (s, 1 H), 3.26–3.34 (m, 1 H), 2.45–2.65 (m, 4
H), 1.98 (s, 3 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 199.8,
162.3, 144.0, 129.5, 127.7, 127.4, 127.3, 44.6, 41.5, 39.7, 25.1 ppm.
[9] A. Mansour, T. Kehat, M. Portnoy, Org. Biomol. Chem. 2008,
6, 3382–3387.
[10] T. Kehat, unpublished results.
Supporting Information (see footnote on the first page of this arti-
cle): Experimental procedures and characterization data of the
polymer-bound intermediates and catalysts.
[11] This byproduct was obtained with low optical purity
(Ͻ10%ee).
[12] A. Buzas, F. Gagosz, J. Am. Chem. Soc. 2006, 128, 12614–
12615.
[13] a) C. Agami, H. Sevestre, J. Chem. Soc., Chem. Commun. 1984,
1385–1386; b) C. Agami, N. Platzer, H. Sevestre, Bull. Soc.
Chim. Fr. 1987, 2, 358–360; c) B. List, R. A. Lerner, C. F.
Barbas III, Org. Lett. 1999, 1, 59–61; d) J. Zhou, V. Wakchaure,
P. Kraft, B. List, Angew. Chem. 2008, 120, 7768–7771; Angew.
Chem. Int. Ed. 2008, 47, 7656–7658.
Acknowledgments
This research was supported by The Israel Science Foundation
(Grant No. 960/06).
[14] H. Pellisier, Tetrahedron 2007, 63, 9267–9331.
[15] For this route, both steps 2 and 3 are unprecedented.
Received: November 18, 2008
[1] a) B. List, Tetrahedron 2002, 58, 5573–5590; b) G. Guillena, C.
Nájera, D. J. Ramón, Tetrahedron: Asymmetry 2007, 18, 2249–
2293.
Published Online: January 16, 2009
996
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 992–996