Synthesis of 6- and 7-membered cyclic enaminones: Scope and mechanism

Article abstract of DOI:10.1021/jo100907u

Six- and seven-membered cyclic enaminones can be prepared using common, environmentally benign reagents. Amino acids are used as synthetic precursors allowing diversification and the incorporation of chirality. The key reaction in this multistep process involves deprotection of Boc-amino ynones and subsequent treatment with methanolic K₂CO₃ to induce cyclization. A β-amino elimination side reaction was identified in a few labile substrates that led to either loss of stereochemical purity or degradation. This process can be mitigated in specific cases using mild deprotection conditions. NMR and deuterium-labeling experiments provided valuable insight into the workings and limitations of this reaction. Although disguised as a 6-endo-dig cyclization, the reagents employed in the transformation play a direct role in bond-making and bond-breaking, thus changing the mode of addition to a 6-endo-trig cyclization. This method can be used to construct an array of monocyclic and bicyclic scaffolds, many of which are found in well-known natural products (e.g., indolizidine, quinolizidine, and Stemona alkaloids).

Full text of DOI:10.1021/jo100907u

pubs.acs.org/joc
Synthesis of 6- and 7-Membered Cyclic Enaminones:
Scope and Mechanism
Micah J. Niphakis,†,‡ Brandon J. Turunen,†,‡ and Gunda I. Georg*,§
†Department of Medicinal Chemistry, University of Kansas, 1251 Wescoe Hall Drive, Lawrence, Kansas
66045, ‡The University of Kansas Center of Excellence in Chemical Methodologies and Library Development,
Structural Biology Center, 2121 Simons Drive, Lawrence, Kansas 66047, and Department of Medicinal
§
Chemistry and the Institute for Therapeutics Discovery and Development, University of Minnesota,
717 Delaware Street SE, Minneapolis, Minnesota 55414
georg@umn.edu
Received May 29, 2010
Six- and seven-membered cyclic enaminones can be prepared using common, environmentally benign
reagents. Amino acids are used as synthetic precursors allowing diversification and the incorporation
of chirality. The key reaction in this multistep process involves deprotection of Boc-amino ynones
and subsequent treatment with methanolic K2CO3 to induce cyclization. A β-amino elimination side
reaction was identified in a few labile substrates that led to either loss of stereochemical purity or
degradation. This process can be mitigated in specific cases using mild deprotection conditions.
NMR and deuterium-labeling experiments provided valuable insight into the workings and limita-
tions of this reaction. Although disguised as a 6-endo-dig cyclization, the reagents employed in the
transformation play a direct role in bond-making and bond-breaking, thus changing the mode of
addition to a 6-endo-trig cyclization. This method can be used to construct an array of monocyclic
and bicyclic scaffolds, many of which are found in well-known natural products (e.g., indolizidine,
quinolizidine, and Stemona alkaloids).
Introduction
attention in pharmaceutical development, particularly as
anticonvulsants and Pgp modulators.2 Their stability and
favorable physicochemical properties are exemplified by
their use as orally active medicinal agents.2d A combination
of the above factors became the impetus for developing a
synthetic route to previously inaccessible or laboriously
synthesized enaminone scaffolds.
Enaminones can best be described as β-acyl enamines or
amides with an interpolated alkene. The reactivity and
stability of these entities are much different than that of a
conventional enamine which readily decomposes through
hydrolytic or oxidative pathways.1 On the contrary, enami-
nones are quite stable and easily isolated. Although not quite
as robust as the conventional amide, the conjugation of the
enamine to a carbonyl attenuates its reactivity, endowing
it with a unique and ambident nature. In addition to their
distinct reactivity profile, enaminones have also attracted
(2) (a) Elassar, A. Z. A.; El-Khair, A. A. Tetrahedron 2003, 59, 8463–
8480. (b) Edafiogho, I. O.; Alexander, M. S.; Moore, J. A.; Farrar, V. A.;
Scott, K. R. Curr. Med. Chem. 1994, 1, 159–175. (c) Edafiogho, I. O.;
Kombian, S. B.; Ananthalakshmi, K. V. V.; Salama, N. N.; Eddington,
N. D.; Wilson, T. L.; Alexander, M. S.; Jackson, P. L.; Hanson, C. D.; Scott,
K. R. J. Pharm. Sci. 2007, 96, 2509–2531. (d) Eddington, N. D.; Cox, D. S.;
Roberts, R. R.; Stables, J. P.; Powell, C. B.; Scott, K. R. Curr. Med. Chem.
2000, 7, 417–436. (e) Salama, N. N.; Eddington, N. D.; Payne, D.; Wilson,
T. L.; Scott, K. R. Curr. Med. Chem. 2004, 11, 2093–2103.
(1) Michael, J. P.; De Koning, C. B.; Gravestock, D.; Hosken, G. D.;
Howard, A. S.; Jungmann, C. M.; Krause, R. W. M.; Parsons, A. S.; Pelly,
S. C.; Stanbury, T. V. Pure Appl. Chem. 1999, 71, 979–988.
DOI: 10.1021/jo100907u
r
Published on Web 07/26/2010
J. Org. Chem. 2010, 75, 6793–6805 6793
2010 American Chemical Society

Niphakis et al.
JOCArticle
Cyclic enaminones, particularly 6-membered enaminones
(2,3-dihydro-4-pyridones), are extraordinarily versatile in-
termediates for the synthesis of piperidine-containing target
molecules. Indeed, this heterocycle exists in numerous drugs
and drug candidates as an indispensable binding element.
Moreover, the piperidine moiety is prevalent in structural
classes of bioactive natural products such as the indolizidines
and quinolizidines.3 Considering the ubiquity of biologically
active piperidine-containing compounds, practical methodo-
logies for the synthesis of these structures, especially those
bearing stereogenic centers, are of great value.
intermediates.11 This method has proven to be enormously
effective in providing advanced intermediates in the synthe-
sis of numerous natural products.12 More recent efforts have
expanded the scope of this chemistry by using an assortment
of chiral auxiliaries.5b,9f,13 Another approach which predates
Comins’ method is the asymmetric hetero-Diels-Alder re-
action of imines with Danishefsky’s diene.14 In this reaction,
the chirality can be derived from various chiral auxiliaries
appended to the imine5b,c,14a-f or through the use of chiral
catalysts.14g-m A noteworthy corollary of this classic [4 þ 2]
approach has recently been reported by the Rovis group15 in
which alkynes and alkenyl isocyanates undergo [2 þ 2 þ 2]
cycloaddition in the presence of a chiral rhodium catalyst.
Although currently limited to the synthesis of the indolizi-
dine enaminones, this method provides rapid access to these
bicyclic molecules with impressive enantioselectivity. De-
spite the success of these asymmetric approaches, innate
limiting factors, such as ring size and substituent constraints,
warranted an exploration of new avenues for the construc-
tion of this useful scaffold.
In pursuit of this goal, we have developed a novel ring-
forming reaction of amino acid-derived ynones to yield
cyclic enaminones.16 Our route accesses the target mole-
cules in high enantiomeric purity by utilizing the chiral pool
strategy to take advantage of the chirality of readily avail-
able starting materials. A notable feature of this methodo-
logy is its power to generate previously unavailable or
circuitously constructed substrates, namely, bicylic enami-
nones, 2,5-disubstituted enaminones, and enaminones with
R- and β-stereocenters. Herein, we present a full disclosure
of our investigations into the scope and mechanism of
this reaction. This concise and operationally facile strategy
gives ready access to novel 6- and 7-membered enaminones
through a vinyl halide intermediate which allows cycli-
zation to proceed through a highly favored 6-endo-trig
pathway.
The synthetic utility of the enaminone is clear when con-
sidering the reactivity of each component moiety (amine,
enamine, enone, and alkene) in isolation. The handles for
modification of this core molecule include four nucleophilic
sites and two electrophilic sites. As depicted in Figure 1,
studies into reactivity of the 6-membered, cyclic enaminones
have made possible a plethora of chemoselective transforma-
tions (e.g., N-functionalization,4 O-functionalization,5 C3,6
C4 [1,2-addition],7 C5,8 and C6 [1,4-addition] functionali-
zation,9 and [2 þ 2] cylization10).
A number of approaches have been developed to construct
the 6-membered enaminone core, yet only a few are capable
of affording nonracemic products (Figure 2). Comins and
co-workers have set precedent for the asymmetric synthe-
sis of enaminones employing chiral N-acylpyridinium
(3) For recent reviews, see: (a) Michael, J. P. In The Alkaloids: Chemistry
and Biology; Academic Press: New York, 2001; Vol. 55, pp 91-258.
(b) Michael, J. P. Nat. Prod. Rep. 2008, 25, 139–165. (c) Michael, J. P.
Nat. Prod. Rep. 2007, 24, 191–222. (d) Michael, J. P. Nat. Prod. Rep. 2005, 22,
603–626. (e) Michael, J. P. Nat. Prod. Rep. 2004, 21, 625–649.
(4) For recent examples of N-functionalization, see: (a) Ege, M.; Wanner,
K. T. Org. Lett. 2004, 6, 3553–3556. (b) Garcia, M. O.; Gomez, A. R.; Adrio,
J.; Carretero, J. C. J. Org. Chem. 2007, 72, 10294–10297.
(5) For recent examples of O-functionalization, see: (a) Shintani, R.;
Tokunaga, N.; Doi, H.; Hayashi, T. J. Am. Chem. Soc. 2004, 126, 6240–6241.
(b) Klegraf, E.; Knauer, S.; Kunz, H. Angew. Chem., Int. Ed. 2006, 45, 2623–
2626. (c) Knauer, S.; Kunz, H. Tetrahedron: Asymmetry 2005, 16, 529–539.
(6) For recent examples of C3-functionalization, see: (a) Kitagawa, H.;
Kumura, K.; Atsumi, K. Chem. Lett. 2006, 35, 712–713. (b) Donohoe, T. J.;
Johnson, D. J.; Mace, L. H.; Thomas, R. E.; Chiu, J. Y. K.; Rodrigues, J. S.;
Compton, R. G.; Banks, C. E.; Tomcik, P.; Bamford, M. J.; Ichihara, O. Org.
Biomol. Chem. 2006, 4, 1071–1084. (c) Lim, S. H.; Curtis, M. D.; Beak, P.
Org. Lett. 2001, 3, 711–714. (d) Di Bussolo, V.; Fiasella, A.; Romano, M. R.;
Favero, L.; Pineschi, M.; Crotti, P. Org. Lett. 2007, 9, 4479–4482.
(7) For examples of C4-functionalization, see: (a) Comins, D. L.;
Killpack, M. O. J. Am. Chem. Soc. 1992, 114, 10972–10974. (b) Thiel, J.;
Wysocka, W.; Boczon, W. Monatsh. Chem. 1995, 126, 233–239. (c) Meyers,
A. I.; Singh, S. Tetrahedron 1969, 25, 4161–4166.
(8) For recent examples of C5-functionalization, see: (a) Ge, H.; Niphakis,
M. J.; Georg, G. I. J. Am. Chem. Soc. 2008, 130, 3708–3709. (b) Wang, X.;
Turunen, B. J.; Leighty, M. W.; Georg, G. I. Tetrahedron Lett. 2007, 48,
8811–8814. (c) Felpin, F. X. J. Org. Chem. 2005, 70, 8575–8578.
(9) For recent examples of C6-functionalization, see: (a) Sebesta, R.;
Pizzuti, M. G.; Boersma, A. J.; Minnaard, A. J.; Feringa, B. L. Chem.
Commun. 2005, 1711–1713. (b) Nakao, Y.; Chen, J.; Imanaka, H.; Hiyama,
T.; Ichikawa, Y.; Duan, W. L.; Shintani, R.; Hayashi, T. J. Am. Chem. Soc.
2007, 129, 9137–9143. (c) Gini, F.; Hessen, B.; Minnaard, A. J. Org. Lett.
2005, 7, 5309–5312. (d) Jagt, R. B. C.; De Vries, J. G.; Feringa, B. L.;
Minnaard, A. J. Org. Lett. 2005, 7, 2433–2435. (e) Furman, B.; Lipner, G.
Tetrahedron 2008, 64, 3464–3470. (f) Focken, T.; Charette, A. B. Org. Lett.
2006, 8, 2985–2988.
(13) (a) Klegraf, E.; Follmann, M.; Schollmeyer, D.; Kunz, H. Eur. J.
Org. Chem. 2004, 3346–3360. (b) Comins, D. L.; Goehring, R. R.; Joseph,
S. P.; O’Connor, S. J. Org. Chem. 1990, 55, 2574–2576. (c) Hoesl, C. E.;
Maurus, M.; Pabel, J.; Polborn, K.; Wanner, K. T. Tetrahedron 2002, 58,
6757–6770. (d) Streith, J.; Boiron, A.; Paillaud, J. L.; Rodriguez-Perez, E. M.;
Strehler, C.; Tschamber, T.; Zehnder, M. Helv. Chim. Acta 1995, 78, 61–72.
(e) Comins, D. L.; Hong, H. J. Am. Chem. Soc. 1991, 113, 6672–6673.
(f) Follmann, M.; Kunz, H. Synlett 1998, 989–990. (g) Comins, D. L.; Joseph,
S. P.; Goehring, R. R. J. Am. Chem. Soc. 1994, 116, 4719–4728.
(14) For select examples using chiral auxiliaries, see: (a) Zech, G.; Kunz,
H. Chem.;Eur. J. 2004, 10, 4136–4149. (b) Alcaide, B.; Almendros, P.;
Alonso, J. M.; Aly, M. F. Chem.;Eur. J. 2003, 9, 3415–3426. (c) Badorrey,
R.; Cativiela, C.; Diaz de Villegas, M. D.; Galvez, J. A. Tetrahedron 2002, 58,
341–354. (d) Andreassen, T.; Haaland, T.; Hansen, L. K.; Gautun, O. R.
Tetrahedron Lett. 2007, 48, 8413–8415. (e) Kranke, B.; Kunz, H. Can. J.
Chem. 2006, 84, 625–641. (f) Kranke, B.; Hebrault, D.; Schultz-Kukula, M.;
Kunz, H. Synlett 2004, 671–674. For select examples using chiral catalysts,
see: (g) Newman, C. A.; Antilla, J. C.; Chen, P.; Predeus, A. V.; Fielding, L.;
Wulff, W. D. J. Am. Chem. Soc. 2007, 129, 7216–7217. (h) Josephsohn, N. S.;
Snapper, M. L.; Hoveyda, A. H. J. Am. Chem. Soc. 2003, 125, 4018–4019.
(i) Shang, D.; Xin, J.; Liu, Y.; Zhou, X.; Liu, X.; Feng, X. J. Org. Chem. 2008,
73, 630–637. (j) Yao, S.; Johannsen, M.; Hazell, R. G.; Jorgensen, K. A.
Angew. Chem., Int. Ed. 1998, 37, 3121–3124. (k) Kobayashi, S.; Komiyama,
S.; Ishitani, H. Angew. Chem., Int. Ed. 1998, 37, 979–981. (l) Jurcik, V.; Arai,
K.; Salter, M. M.; Yamashita, Y.; Kobayashi, S. Adv. Synth. Catal. 2008,
350, 647–651. (m) Cros, J. P.; Perez-Fuertes, Y.; Thatcher, M. J.; Arimori, S.;
Bull, S. D.; James, T. D. Tetrahedron: Asymmetry 2003, 14, 1965–1968.
(15) (a) Friedman, R. K.; Rovis, T. J. Am. Chem. Soc. 2009, 131, 10775–
10782. (b) Yu, R. T.; Rovis, T. J. Am. Chem. Soc. 2006, 128, 12370–12371.
(c) Yu, R. T.; Rovis, T. J. Am. Chem. Soc. 2006, 128, 2782–2783. (d) Lee,
E. E.; Rovis, T. Org. Lett. 2008, 10, 1231–1234. (e) Yu, R. T.; Lee, E. E.;
Malik, G.; Rovis, T. Angew. Chem., Int. Ed. 2009, 48, 2379–2382.
(10) Comins, D. L.; Zheng, X.; Goehring, R. R. Org. Lett. 2002, 4, 1611–
1613.
(11) Comins, D. L.; Brown, J. D. Tetrahedron Lett. 1986, 27, 4549–4552.
(12) (a) Comins, D. L.; Libby, A. H.; Al-awar, R. S.; Foti, C. J. J. Org.
Chem. 1999, 64, 2184–2185. (b) Comins, D. L.; Zhang, Y. M. J. Am. Chem.
Soc. 1996, 118, 12248–12249. (c) Comins, D. L.; Hong, H. J. Org. Chem.
1993, 58, 5035–5036. (d) Comins, D. L.; Morgan, L. A. Tetrahedron Lett.
1991, 32, 5919–5922. (e) Comins, D. L.; Sandelier, M. J.; Grillo, T. A. J. Org.
Chem. 2001, 66, 6829–6832. (f) Comins, D. L.; Benjelloun, N. R. Tetrahedron
Lett. 1994, 35, 829–832. (g) Comins, D. L.; Sahn, J. J. Org. Lett. 2005, 7,
5227–5228. (h) Comins, D. L.; Zhang, Y. M.; Joseph, S. P. Org. Lett. 1999, 1,
657–659.
(16) Turunen, B. J.; Georg, G. I. J. Am. Chem. Soc. 2006, 128, 8702–8703.
6794 J. Org. Chem. Vol. 75, No. 20, 2010

Niphakis et al.
JOCArticle
FIGURE 1. Select synthetic transformations of the 6-membered cyclic enaminone.
FIGURE 3. Retrosynthetic analysis for enaminone construction.
(Figure 3). We first considered a direct-Michael addition of
an amine into the linear ynone which would provide unin-
terrupted entry into our desired scaffold. It was reasoned
that the necessary amino ynone substrates could be obtained
from β-amino acid precursors. This approach was attractive
to us not only because it could potentially furnish pyridones
with four distinct appendages and two stereogenic stereo-
centers but also because it utilizes easily accessible chiral
starting materials. The strategic use of β-amino acids and
their immediate precursors entrusts the chiral pool and
established asymmetric chemistry17 to provide both diversity
and asymmetry.
Despite the numerous examples of intermolecular 1,4-
additions of amines to ynones, precedent for an intramole-
cular variant as proposed is scarce. The lack of literature
precedent for such a 6-endo-dig transformation raised imme-
diate concerns. Although this mode of cyclization is favored
according to Baldwin’s rules for ring-closing reactions,18
questions arose with regard to the feasibility of suitable
FIGURE 2. General approaches to access nonracemic 6-membered
enaminones.
Results and Discussion
(17) (a) Cole, D. C. Tetrahedron 1994, 50, 9517–9582. (b) Liu, M.; Sibi,
M. P. Tetrahedron 2002, 58, 7991–8035. (c) Drexler, H. J.; You, J.; Zhang, S.;
Fischer, C.; Baumann, W.; Spannenberg, A.; Heller, D. Org. Process Res.
Dev. 2003, 7, 355–361. (d) Ma, J. A. Angew. Chem., Int. Ed. 2003, 42, 4290–
4299. (e) Sewald, N. Angew. Chem., Int. Ed. 2003, 42, 5794–5795. (f) Lelais,
G.; Seebach, D. Biopolymers 2004, 76, 206–243. (g) Seebach, D.; Kimmerlin,
T.; Sebesta, R.; Campo, M. A.; Beck, A. K. Tetrahedron 2004, 60, 7455–7506.
(h) Enantioselective Synthesis of β-Amino Acids, 2nd ed.; Juaristi, E.,
Soloshonok, V. A., Eds.; Wiley: Hoboken, 2005.
Reaction Development. As previously mentioned, methods
for the construction of nonracemic dihydropyridones are
scarce, and the existing few, although highly contributive
to this area, have left many structurally simple enaminones
out of reach. Thus, we envisioned a complementary route
to these coveted molecules using the chiral pool strategy
(18) Baldwin, J. E. J. Chem. Soc., Chem. Commun. 1976, 734–736.
J. Org. Chem. Vol. 75, No. 20, 2010 6795

Niphakis et al.
SCHEME 1. Preparation of Ynone Intermediates
JOCArticle
FIGURE 4. Potential modes of racemization/epimerization.
amine/ynone orbital overlap and competitive intermolecular
reactions. If direct addition to the ynone could not be
achieved, however, the use of an appropriate ynone synthetic
equivalent could direct this cyclization into well-explored
mechanistic territory, providing a surer means to our desired
target molecules.
In addition to our questions regarding the mode of addi-
tion, concerns arose about racemization and epimerization
of the pre-established stereocenters also had to be investi-
gated. Enolization processes leading to loss of R-chiral
centers are commonplace and would need to be avoided.
We also considered the likelihood of β-amino ketone inter-
mediates to undergo retro-Michael/retro-Mannich-type
processes (Figure 4).19 This scenario would jeopardize the
integrity both R- and β-stereocenters or lead to substrate
decomposition, both of which would severely limit the utility
of such a protocol. With these potential hurdles in mind, we
ventured to construct the requisite ynone starting materials.
Three routes were devised to obtain β-amino Weinreb
amides, the immediate precursors of the desired ynone
starting materials. In the cases where the Boc-β-amino acid
was commercially available, an EDCI coupling with HN-
(OMe)Me•HCl in the presence of N-methylmorpholine
(NMM) furnished the corresponding amide in a single step
(method A, Scheme 1). Alternatively, Boc-R-amino acids
were converted to diazoketones which, in the presence of
catalytic CF3CO2Ag, collapsed to the corresponding ketene.
The ketene intermediate was trapped in situ with HN(OMe)-
Me, providing the desired Weinreb amides (method B).
Unsubstituted Weinreb amides were synthesized through
a one-pot Michael-addition/Boc-protection sequence to
afford Boc-β-aminomethyl esters (method C). Treatment
potential precursors to these important heterocycles (entries
1-6, Table 1). The isolated stereogenic center on the pyrro-
lidine ring of ynones 2g-k was to be used to facilitate
detection of the β-epimerization (entries 7-11). Likewise,
cyclohexyl systems 2m and 2n would allow R-epimerization
to be detected (entries 13 and 14). Finally, acyclic ynones
2o-t were synthesized to obtain monocyclic enaminones
(entries 15-21). Although the enaminones to be generated
from ynones 2r-t would be relatively unembellished, we
were attracted to these targets because of a surprising lack of
general routes to obtain them, in spite of their simplicity.
From the outset, we envisioned a protocol in which the
Boc-protecting group would be removed to liberate a nucleo-
philic amine that would, in turn, react with the tethered
ynone moiety. Our initial efforts found success in a two-tier
deprotection/cyclization protocol to provide enaminone 3a
(Table 2).16 The use of 4 N HCl (entries 5-9, Table 2) or
TMS-I (entry 10) consistently gave higher yields than when
TFA was used to deprotect (entries 1-4), regardless of
the cyclization method. We were initially perplexed by this
disparity. The putative ammonium salt intermediates of each
method would seemingly only differ with respect to their
counterions (Cl-, I-, or CF3CO2-). Upon closer scrutiny,
however, it became clear that HCl and TMS-I served another
purpose. In addition to deprotecting the Boc group, these
reagents also promoted conjugate additions of their respec-
tive halides (Scheme 2). Isolation of the deprotected inter-
mediates revealed that, prior to cyclization, the ynone 2a had
been converted by HCl and TMS-I to a mixture of vinyl
halide 4 and dihaloketone 5. Trifluoroacetic acid, however,
left the ynone intact (6). The addition of chloride or iodide
of the methyl esters with HN(OMe)Me HCl and i-PrMgCl
3
afforded the desired amides.20 In the event that the Boc-
protected amines required N-alkylation, this was accom-
plished subsequent to amide formation using NaH and an
appropriate alkyl halide. In the final step, the desired ynones
were obtained from the Weinreb amides through the addi-
tion of excess (5 equiv) alkynylmagnesium reagents. These
simple steps could be conducted on multigram scale with
minimal reduction in yield.
Due to our interest in indolizidine and quinolizidine
natural products, we chose to synthesize ynones 2a-f as
(19) (a) Slosse, P.; Hootele, C. Tetrahedron 1981, 37, 4287–4294.
(b) Morley, C.; Knight, D. W.; Share, A. C. J. Chem. Soc., Perkin Trans. 1
1994, 2903–2907. (c) Harrison, J. R.; O’Brien, P.; Porter, D. W.; Smith, N. M.
J. Chem. Soc., Perkin Trans. 1 1999, 3623–3631. (d) Kimball, F. S.; Turunen,
B. J.; Ellis, K. C.; Himes, R. H.; Georg, G. I. Biorg. Med. Chem. 2008, 16,
4367–4377.
(20) Williams, J. M.; Jobson, R. B.; Yasuda, N.; Marchesini, G.; Dolling,
U. H.; Grabowski, E. J. J. Tetrahedron Lett. 1995, 36, 5461–5464.
6796 J. Org. Chem. Vol. 75, No. 20, 2010

Niphakis et al.
JOCArticle
TABLE 1. Synthesized Ynone Substrates and Multistep Yields for
Preparation Using Methods A, B, or C (Scheme 1)
TABLE 2. Optimization of Deprotection/Cyclization Procedure for the
Preparation of Cyclic Enaminones
timea
(h)
yieldb
(%)
entry
(i) deprotection
(ii) cyclization
c
1
2
3
4
5
TFA, CH2Cl2
TFA, CH2Cl2
TFA, CH2Cl2
TFA, CH2Cl2
4 N HCl/dioxane CH2Cl2, K2CO3
4 N HCl/dioxane THF, K2CO3
CH2Cl2, NaHCO3
CH2Cl2, K2CO3
THF, K2CO3
nad
20 h
20 h
30
0
0
38
0
0
CH2Cl2, H2O, K2CO3 5 h
20 h
20 h
6
7
8
9
10
4 N HCl/dioxane CH2Cl2, H2O, K2CO3 1 h
4 N HCl/dioxane THF, H2O, K2CO3
4 N HCl/dioxane MeOH, K2CO3
74
75
87
95
1 h
15 min
30 min
c
TMS-I, CH2Cl2
MeOH, K2CO3
aReaction time of cyclization step. bIsolated yield. Saturated aqu-
eous NaHCO3. dReaction proceeded in a separatory funnel upon
workup.
SCHEME 2. Fate of Ynone upon Boc Deprotection with HCl,
TMS-I, and TFA
Shortly after we embarked on our study of reaction scope,
we encountered a formidable challenge. As we had feared,
ynones bearing R- or β-stereocenters, when subjected to our
one-pot procedure, yielded mixtures of diastereomers or
partial racemates. We suspected enolization or β-amino
elimination or both processes were at hand; if not overcome,
this would negate the principal advantage of the chiral pool
approach. Although both acid- and base-induced epimeriza-
tions of this type are possible, we initially focused our atten-
tion on the acidic deprotection conditions. Since the most
profound stereochemical damage was observed in proline-
derived enaminones, hydroxylated ynone 2j was employed as
our model system to assess the extent of β-epimerization
under an array of deprotection conditions (Table 3).
Our previously optimized deprotection conditions had
epimerized 15% of the isolated product (entry 1, Table 3).
In neat and dilute TFA, the yields and diastereomeric ratios
of enaminone 3j showed no improvement (entries 2 and 3).
Slow addition of HCl in ether until all the starting mate-
rial was consumed also failed to give satisfactory results
(entry 4). Positing an acid-induced mode of racemization, we
aSee Scheme 1. bIsolated multistep yield from commercially available
starting materials. cStereocenter was inverted from Weinreb amide
precursor using the Mitsunobu reaction.
into the ynone prior to cyclization was evidently favoring
ring closure. This finding was good evidence that this reac-
tion was not proceeding through a 6-endo-dig pathway as we
first had thought (see below).
Another important observation that was made during
these studies was the apparent dependency of the cyclization
on water or MeOH. Regardless of the deprotection method,
the enaminone would not form in THF or CH2Cl2 (entries 2,
3, 5, and 6; Table 2) unless water was used as a cosolvent
(entries 4, 7, and 8). Although these wet solvents could affect
cyclization, MeOH proved to be the best solvent and was
chosen to further explore this reaction (entries 9 and 10).
With these optimized conditions, we proceeded to establish
the scope of this reaction.
J. Org. Chem. Vol. 75, No. 20, 2010 6797

Niphakis et al.
SCHEME 3. Plausible Mechanism for TMS-I-Induced
JOCArticle
TABLE 3. Optimization of Boc Deprotection Conditions To Suppress
β-Epimerization
r-Epimerization of Ynones
yielda
entry
conditions
4 N HC1, dioxane
TFA (neat)
TFA/CH2C12 (1:1)
l N HC1, ether
(i) TBSOTf, 2,6-lutidine, CH2Cl2, (ii) TBAF
TESOTf, 2,6-lutidine, CH2Cl2
CAN, CH3CN, reflux
TMS-I (3 equiv), CH2Cl2, 0 °C
TMS-I (1 equiv), CH2Cl2, -78 to 0 °C; then
TMS-I (2 equiv)
(%) (dr)b
1
2
3
4
5
6
7
8
9
77 (85:15)
31 (67:33)
18 (88:12)
36 (67:33)
0
21 (83:17)
0
99 (75:25)
60 (94:6)
the Boc group could be easily removed, yet upon treatment
with base, we did not observe any enaminone formation
(entry 10, Table 3). As shown in previous experiments, the
success of this reaction is not only contingent on the removal
of the Boc group but also on ynone “activation” with halides
(Scheme 2). To this end, sodium iodide (NaI) was added as a
nucleophilic halide source and the reaction was repeated
(entry 11, Table 3). To our delight, the desired enaminone
was formed in excellent yield (92%) and without any detect-
able epimerization. Conceivably, the remote hydroxy group
could resist epimerization by directing the readdition of the
eliminated amine to the most favored anti-substituted pro-
duct. We were encouraged to find that, in the absence of
diasteromeric control, enaminones could be obtained in high
enantiopurity (eq 2). When ynone 2d was subjected to these
deprotection conditions, a crystalline solid formed in the
reaction media and was determined by X-ray analysis to
be the desired deprotected, vinyl iodide intermediate (7a,
Figure 5). Upon treatment of this salt with base, enaminone
3d was formed in less than 2 min. More importantly, the
product was obtained with an enantiomeric ratio (er) of
98.5:1.5 showing that this protocol had effectively mitigated
β-epimerization.
10
11
HCO2H, rt
NaI(3equiv), HCO2H, rt
0
93 (>95:5)
aIsolated yield. bDiastereomeric ratio (dr) determined by 1H NMR
integration.
attempted to suppress the stereochemical deterioration using
basic and neutral conditions (entries 5-7). Only TESOTf
and 2,6-lutidine (entry 6) provided the desired product, albeit
in 21% yield and with a dr of 83:17. When TMS-I was used to
induce Boc deprotection, we observed a marked reduction
of epimerization and improvement in yields (entries 8 and 9).
It should be noted that ynones have been shown to react with
TMS-I at -78 °C to form β-iodoallenolates, which upon
warming to 0 °C tautomerize to afford Danishefsky-type
dienes (Scheme 3).21 Although this would not directly com-
promise the integrity of the β-stereocenter, substrates bear-
ing R-stereocenters would be affected.
When enolizable substrate 2m was subjected to the same
conditions, epimerization was even more profound than
when HCl was used (eq 1). In addition to the detrimental
stereochemical effects of TMS-I, the necessity of rigorously
dried solvents, glassware, and cryogenic temperatures pro-
voked a search for a new, simple protocol that would be
suitable for epimerizable substrates.
We next investigated these new conditions on enolizable
ynone 2m as a model for R-epimerization (eq 3). The new
conditions yielded the desired product (3m) with a dr of 94:6.
This was a considerable improvement over the previous
method where the dr was 80:20. Interestingly, we also saw a
time-dependent increase of epimerization indicating that the
acid step was indeed responsible, at least in part, for R-epi-
merization. Enolate formation during the basic step could also
potentiate epimerization and was investigated further.
With a wealth of known Boc-deprotection protocols to
explore, we were attracted to the use of formic acid (HCO2H)
to avert epimerization.22 The simple technique of this method
was attractive considering it could accomplish the desired
deprotection at ambient temperature and in open air.
Furthermore, we hoped that this relatively weak acid would
lessen epimerization. Applying these conditions to ynone 2j,
Vinyl iodides 7a, 7b, and 7c were chosen as model sub-
strates to examine enolization. These substrates were sub-
jected to the cyclization conditions using deuterated metha-
nol (CD3OD) as a solvent, and the extent of R-deuteration
(21) Wei, H.-X.; Timmons, C.; Farag, M. A.; Pare, P. W.; Li, G. Org.
Biomol. Chem. 2004, 2, 2893–2896.
(22) Wasserman, H. H.; Berger, G. D.; Cho, K. R. Tetrahedron Lett.
1982, 23, 465–468.
6798 J. Org. Chem. Vol. 75, No. 20, 2010

Niphakis et al.
JOCArticle
The collective data from the deuterium exchange reactions
hinted at another liability in our approach. We feared that
the extent of R-deuteration might be diagnostic of precy-
clized intermediates that would undergo β-amino elimina-
tion. Thus, ynones 2g, 2h, and 2i were subjected to the two-
step procedure (eq 5). Terminal ynone 2g cyclized with no
observable epimerization, whereas the methyl- and phenyl-
substituted ynones 2h and 2i were obtained as diastereomeric
mixtures.
FIGURE 5. X-ray crystal structure of vinyl iodide ntermediate.
In retrospect, these findings are not surprising. The prin-
ciples of vinylogy would indeed predict an attenuated acidity
of the enaminone R-position in relation to that of its ketone
progenitor. Thus, R-epimerization or β-amino elimination is
precluded by rapid formation of the vinylogous amide.
When the rate of cyclization is retarded with substituted
ynones there is a significant increase in epimerization, thus
showing the vulnerability of the enaminone precursors in
basic media. Despite the use of mild deprotection conditions,
these new findings suggest that base-induced epimerization is
predominant when cyclization is slow.
Suspicions immediately arose concerning the putative role
of the acid in β-amino elimination. Perhaps the discrepancies
in the stereochemical outcome of each deprotection condi-
tion were attributable to the dissimilar haloketone inter-
mediates and their particular reactivities under basic con-
ditions. For instance, a conspicuous difference exists when
considering the respective halides (Cl vs I). Furthermore,
double chloride addition into the ynone occurs in the pre-
sence of HCl, whereas the diiodo congener is not formed in
HCO2H and NaI (Scheme 2). We were cognizant that these
dissimilarities, rather than the strength of the acid used for
deprotection, could dictate the extent of epimerization by
governing the rate of cyclization. Fortunately, this could be
directly investigated.
FIGURE 6. Deuteration of cyclization intermediates and enami-
none products.
was determined (Figure 6). Deuterium incorporation at the
R-position is indicative of enolate formation revealing an-
other potential source of epimerization. When intermediate
7a was treated with methanol-d4 and K2CO3, enaminone 3d
was immediately formed and no deuteration was observed.
Once formed, this enaminone was resistant to deuterium
incorporation for up to 24 h (i.e., to form enaminone 3d-d2).
With the intent of slowing the cyclization, we used substi-
tuted vinyl iodides 7b and 7c. For these cases, the cyclized
products 3e-d5 and 3f-d2 were obtained with complete
R-deuteration. Furthermore, as observed before, enaminones
3e and 3f, which were formed in nondeuterated solvent, did
not undergo R-deuteration under the prescribed conditions.
The point has already been made that enaminones are
reluctant to undergo deuteration in the presence of K2CO3;
however, an anomaly was noted when investigating enami-
none 3e (eq 4). This substrate underwent selective and
complete γ-deuteration to provide enaminone 3e-d3 in 2 h.
Although it is out of the scope of this paper, this finding
reveals another handle on this versatile scaffold for chemo-
selective modification.23
Experimental evidence suggests that the primary cause of
β-epimerization for terminal ynones is the strongly acidic
HCl conditions. When ynone 2g was subjected to 4 N HCl
the resultant ammonium salts (8a and 8b) were each gene-
rated as 2:1 diastereomeric mixtures (eq 6). From this we
conclude that the acid is directly responsible for inducing
amino elimination. As such, this process can be mitigated
when using a weaker acid such as HCO2H. It is important to
note that the use of HCO2H and NaI does not assuage the
stereochemical detriment in cases where cyclization is slow
(23) (a) Sibgatulin, D. A.; Volochnyuk, D. M.; Rusanov, E. B.; Kostyuk,
A. N. Synthesis 2006, 1625–1630. (b) Volochnyuk, D. M.; Kostyuk, A. N.;
Sibgatulin, D. A.; Petrenko, A. E. Synthesis 2004, 2545–2549.
J. Org. Chem. Vol. 75, No. 20, 2010 6799

Niphakis et al.
JOCArticle
(i.e., substituted ynones). Thus, this mild protocol is most
appropriate for terminal ynones. With these insights we next
investigated the reaction scope to validate the utility and
generality of this protocol.
TABLE 4. Six-Membered Enaminone Substrate Scope
Substrate Scope. As can be seen in Table 4, a diverse
collection of enaminones can be constructed using the above
methods. From this data set we intended to establish a basis
for choosing appropriate deprotection conditions. Bicyclic
enaminones are formed efficiently, providing a facile route to
quinolizidine (entries 1-3) and indolizidine (entries 4-11)
scaffolds. Synthesis of morpholino enaminone 3l was also
feasible (entry 12). Pyrrolidine substrates were particularly
susceptible to stereochemical erosion; however, this could be
mitigated in terminal ynone substrates by using a HCO2H
and NaI during the deprotection step. Although the reason
for this sensitivity has not been directly assessed, we believe
this is due, in part, to the alleviation of pyrrolidine ring
strain. Not surprisingly, anti-substituted pyrrolidines were
less prone to epimerization than their syn-counterparts
(entries 10 and 11). Installation of aliphatic and aromatic
substituents adjacent to the ring-fused nitrogen was also
accomplished by employing terminally substituted ynones
(entries 5, 6, 8, and 9). As expected, cyclization was relatively
slow (15 min to 3 h) in these substrates, but all proceeded
in excellent yields. In these cases, however, the HCO2H-
based deprotection method was ineffective at preventing
β-epimerization.
Enolizable stereocenters were also a potential liability
considering our protocol featured both an acidic and a basic
step. As shown before, the use of HCl and TMS-I were
destructive when attempting to establish the cis-fused enami-
none 3m. We were pleased to find that β-epimerization could
also be suppressed in this case under the milder conditions
with HCO2H and NaI (entry 13). It should be noted that
since the trans-substituted product is expected to be more
stable, epimerization is likely to be thermodynamically pre-
ferred. With this in mind, it was not surprising that trans-
fused enaminone 3n was more resistant to epimerization in
both deprotection conditions and could be acquired in high
diasteromeric purity (entry 14).
aMethod A: (i) 4 N HCl/dioxane, 15 min, (ii) K2CO3, MeOH. Method
B: (i) Nal (3 equiv), formic acid, 6-24 h, ii) K2CO3, MeOH. Method C:
(i) TMS-I, CH2CI2, -78 to 0 °C, (ii) K2CO3, MeOH. bIsolated yield.
cChiral HPLC. d1H NMR integration.
instead of epimerization. Hence, it was observed that several
monocyclic products were obtained in lower yields than their
bicyclic analogues.
We next investigated the synthesis of monocyclic enami-
nones. These compounds would seemingly be more difficult
to form, as they have no conformational constraints facili-
tating ring closure. In this regard, acyclic β-amino ynones
were found to be viable substrates (entries 15-22). Even
sterically encumbered, internal ynones undergo cyclization
to form monocyclic products (entry 18). In cases where the
extruded amine is not tethered to the ynone, as in bicylic sub-
strates 3a-l, the retro-Michael process leads to degradation
The observation that a subtle ring expansion had pro-
found effects on the extent of β-racemization (entry 1 vs 4)
led us to attempt the construction of phenylglycine-derived
enaminone 3o (entry 15). We envisioned that this substrate
would be particularly susceptible to β-amino elimination.
Expecting to see a distinct improvement in yield when
HCO2H was used instead of HCl, we were surprised that
both deprotection methods were equally suited for this
sensitive substrate. By conducting this reaction in an NMR
6800 J. Org. Chem. Vol. 75, No. 20, 2010

Niphakis et al.
JOCArticle
TABLE 5. Seven-Membered Enaminone Preparation
rendered four novel 7-membered enaminones (entries 1-4,
Table 5). Indeed, pyrrolo[1,2-a]azepine 3u bears resemblance
to the core and distinguishing feature of the Stemona
alkaloids.24 This unique molecule and its piperidino conge-
ner (3v) were both attainable in good yields (entries 1 and 2).
Spirocyclic enaminone 3w and baclofen-derived enaminone
3x were also readily obtained via our deprotection/cycliza-
tion protocol (entries 3 and 4). Although all of these enami-
nones were constructed in racemic form and, hence, their
stereochemical liabilities not investigated, we have no reason
to believe that they are susceptible to the stereochemical
deterioration seen in β-amino ynones.
In summary, we have developed two complementary
protocols for synthesizing 6- and 7-membered enaminones.
The first method, using HCl, is rapid and able to activate and
deprotect internal and terminal ynones in under 15 min.
It is best suited for substrates without R-stereocenters or
those that are not sensitive to acid-induced β-amino elimina-
tion. The second method, using HCO2H and NaI, although
requiring longer reaction times (6-24 h), is ideal for terminal
ynones with sensitive R- and β-stereocenters. Both proce-
dures are operationally facile and can be carried out in a
single vessel. Furthermore, these experimentally simple and
environmentally benign conditions are conducive to produc-
tion of multigram quantities of these enaminone scaffolds.25
Mechanistic Insights. During the course of our investiga-
tion, several observations have been made en route to opti-
mization of this reaction using ynone 2a. A couple of gene-
ralizations can be drawn from this data: (1) when using TFA
or other deprotection methods (including basic or neutral
conditions) which do not incorporate a halogen, the isolated
yields were poor (Scheme 2), and (2) regardless of the depro-
tection protocol employed, no desired enaminone was ob-
tained without the addition of water or MeOH (Table 2).
These preliminary observations lay the groundwork for our
mechanistic studies.
aSee the Supporting Information for the synthesis of ynones 2u-x.
bMethod A: (i) 4 N HCl/dioxane, 15 min, (ii) K2CO3, MeOH. Method B:
(i) Nal (3 equiv), formic acid, 6-24 h, (ii) K2CO3, MeOH. cIsolated yield.
tube and carefully monitoring its progression, it became
apparent that degradation (i.e., β-amino elimination) occur-
red upon the addition of methanolic K2CO3 and not during
deprotection. The complete retention of stereochemistry
suggests that this process was irreversible in contrast to cases
where the extruded amine remains tethered to the resultant
Michael acceptor (entries 1-12). From these results, we
suggest alternative methods be used to access enaminones
of this type. Fortunately, this is well within the scope of
both the Comins’ N-acylpyridinium13 and the hetero-Diels-
Alder approach.14
In addition to introducing steric bulk, attenuating the
nucleophilicity of the amine would seemingly impede an
efficient ring closure. The synthesis of enaminone 3t demon-
strates that despite the use of a significantly less reactive
anilino nitrogen, cyclization still occurs (entry 22), albeit in
lower yields.
With success in constructing 6-membered enaminones, we
next explored the feasibility of constructing 5- and 7-mem-
bered rings. This method lacks the strict confines of ring size
for the alternative routes to cyclic enaminones. Our initial
attempts to cyclize R-amino ynones to form 5-membered
enaminones were unsuccessful. This is consistent with our
hypothesis that this reaction proceeds through an endo-trig
mode of cyclization (5-endo-trig is disfavored). The synthesis
of 5-membered enaminones remains a limitation of this
methodology.
By isolating the precyclized intermediates (4 and 5,
Figure 7) and verifying a complete consumption of the ynone
moiety, the first stipulation (i.e., the need for a halide source
while deprotecting the Boc-group) became clear. In suffi-
ciently acidic conditions, halides add into the ynone group
forming a vinyl halide that readily undergoes ring closure
to form the desired enaminone. When the ynone remains
intact cyclization is not efficient and intermolecular pro-
cesses (i.e., polymerization) predominate. Therefore, preac-
tivation is required for a successful reaction.
The discovery that the ynone group was transformed in
the presence of halogenic acids demanded a reevaluation of
our originally proposed 6-endo-dig mechanism. With direct
evidence for the formation of halides 4 and 5, a direct 1,4-
addition 6-endo-trig cyclization/halide elimination would be
one plausible mechanism (pathway A, Figure 7). We were
initially impressed, however, by the strong solvent depen-
dence of this reaction. When the reaction was carried out in
bulky alcoholic (s-BuOH or i-PrOH) or non-nucleophilic
solvents (CH2Cl2 or THF), the reaction was significantly
impaired. On the other hand, the use of water or MeOH was
highly beneficial and independent of the mode of deprotec-
tion. In our first disclosure of this reaction, we proposed a
mechanism in which an oxygen nucleophile plays a direct
Previously reported methods for the construction of cyclic
enaminones have not been amenable to the synthesis of
7-membered rings either. Furthermore, to our knowledge,
there is no general method for their construction. With-
out any alteration of our protocol, γ-amino ynones 2u-x
(24) (a) Pilli, R. A.; Ferreira de Oliveira, M. C. Nat. Prod. Rep. 2000, 17,
117–127. (b) Seger, C.; Mereiter, K.; Kaltenegger, E.; Pacher, T.; Greger, H.;
Hofer, O. Chem. Biodivers. 2004, 1, 265–279. (c) Pilli, R. A.; Rosso, G. B.; De
Oliveira, M. C. F. In The Alkaloids: Chemistry and Biology; Cordell, G. A.,
Ed.; Academic Press: New York, 2005; Vol. 62, pp 77-173.
(25) See the Supporting Information.
J. Org. Chem. Vol. 75, No. 20, 2010 6801

Niphakis et al.
JOCArticle
tethered nitrogen nucleophile. Following treatment with 4 N
HCl, the residue was dissolved in CH3CN and TEA was
added (eq 8). The NMR spectrum revealed vinyl chloride 15
as the sole product. Moreover, upon addition of excess D2O,
the vinyl chloride remained intact. Clearly, trace amounts of
water were not facilitating the cyclization when the reaction
was carried out in CH3CN.
FIGURE 7. Possible modes of cyclization.
role in bond making and bond breaking.16 Therein, we
suggested an addition of MeOH into vinyl chloride (9)
which, upon extrusion of the chloride ion, the resultant
oxonium intermediate 12 could undergo ring closure in a
6-exo-trig fashion (Figure 8, pathway B).
Delving further, we examined the reactivity of intermedi-
ates 15 and 16 in the presence of methanolic K2CO3. If
MeOH facilitates the cyclization through a conjugate addi-
tion it would be expected that the isolated vinyl halide (15)
would be consumed more rapidly than the vinyl chloride 4
could form the enaminone. In other words, if the rate of
amino cyclization is faster than the addition of MeOH, the
former process would obviate the latter. To test this, a
mixture of intermediates 15 and 16 was incubated in MeOH
and K2CO3 (eq 9). After 2 h, the composition of the crude
reaction mixture consisted of a 1:2 mixture of vinyl ether 17
and unreacted vinyl chloride 15 (33% conversion). Thus, the
addition of MeOH to the vinyl chloride is much slower than
the cyclization process, which occurs in less than 5 min. From
this data, it is reasonable to suggest that the addition of
MeOH into a vinyl chloride should not be invoked in the
reaction mechanism. Thus, we suggest that these results are
strong evidence for a direct 6-endo-trig mode of cyclization
(pathway A, Figure 7).
In an attempt to gain mechanistic insight through detect-
ing transient intermediates, we dissolved the mixture of 4 and
5 in methanol-d4 and monitored the reaction mixture via 1H
NMR after adding a solution of K2CO3 dropwise. Dichlo-
ride 5 was first converted into vinyl chloride 9, and subse-
quent additions led to the formation of enaminone 3a with
no other detectible intermediates. When the ,vinyl iodide
analogue of 4, generated from NaI and HCO2H, was sub-
jected to these same conditions, the product was also formed
with no detectable intermediates. Although pathway B
provided a plausible explanation for the aforementioned
solvent effects, dissimilarities in the solubilizing powers of
each solvent would still need to be ascertained. Impaired
yields would be expected if the protonated amine intermedi-
ates (4 and 5) or the base (K2CO3) were poorly soluble.
Whether or not a nucleophilic solvent was necessary, how-
ever, was difficult to directly assess due the insolubility of the
reaction constituents in most organic solvents. To answer
this question, we explored alternative organic bases that we
expected would allow the reaction to be carried out in non-
nucleophilic solvents. Our first success demonstrated that
triethylamine (TEA) could be used instead of K2CO3 (eq 7).
As hoped, this base was not only compatible with MeOH but
could be used with non-nucleophilic solvents, such as
CH3CN and DMSO, resulting in the formation of enami-
none 3a. Since these solvents are hygroscopic, adventitious
water could potentially facilitate the cyclization in a catalytic
manner. To rule out this possibility, we constructed ynone 14
to determine the fate of the electrophile in the absence of a
Conclusions
We have developed a practical route to nonracemic 6- and
7-membered enaminones starting from amino acids, pro-
viding a complementary approach to preexisting methods.
Using this approach, asymmetry can be derived from the
rich supply of commercially available amino acids. The key
reaction and final step in this sequence is a novel one-flask
deprotection/cyclization reaction of Boc-amino ynones.
Two methods were developed to achieve the Boc deprotec-
tion that are suitable for different substrates. The use of 4 N
HCl in dioxane is most fitting for internal ynones or those
without sensitivity to acid-mediated side reactions. Alterna-
tively, for those substrates that have enolizable stereocenters
6802 J. Org. Chem. Vol. 75, No. 20, 2010

Niphakis et al.
JOCArticle
or are susceptible to acid-promoted β-amino elimination
(retro-Michael) processes, NaI and HCO2H are well-suited.
The latter conditions provide a mild alternative and work
best for terminal ynones. Both protocols are economic and
operationally facile having no need for dry solvents/reagents
and can be conducted at room temperature. Furthermore,
both methods can be carried out on multigram scale with
comparable yields.
The substrate scope of this reaction was also assessed. It
appears that this reaction is general for the construction of
6- and 7-membered enaminones. When the deprotection
conditions are judiciously chosen, monocyclic and bicyclic
heterocycles can be obtained from amino ynones in high
enantiomeric or diastereomeric purity. Many of the hetero-
cyclic scaffolds reported here, though structurally simple, are
unprecedented in the literature.
The mechanism of the final cyclization sequence has also
been investigated. The success of this reaction relies on the
conversion of the ynone moiety into a vinyl halide species
and trapping of the protected amine as an ammonium salt
following Boc degradation. This preactivation allows for
an efficient intramolecular 1,4-addition once the free amine
is released upon the addition of base. Thus, this reaction
is thought to proceed through a 6-endo-trig ring-closing
process.
solution was stirred overnight. To the reaction mixture was
added activated charcoal (∼2 g) and the reaction mixture was
stirred for 5 min and filtered. The filtrate was concentrated,
and the residue redissolved in EtOAc. To this was added
activated charcoal (∼2 g) and the process repeated. When the
filtrate had been concentrated a second time, the residue was
purified via SiO2 flash chromatography using 35% EtOAc/
hexanes as eluent affording the title compound as a white solid
(3.75 g, 72%): mp 69.5-70.3 °C; 1H NMR (400 MHz, CDCl3) δ
(1:1 mixture of rotamers) 1.46 (s, 18H), 1.95-2.03 (bm, 2H),
2.30-2.38 (bm, 2H), 2.44-2.54 (bm, 2H), 3.03-3.16 (m, 2H),
3.16 (s, 6H), 3.41-3.56 (m, 3H), 3.67 (s, 6H), 3.67-3.75 (m,
1H), 4.09-4.14 (bm, 2H), 4.27-4.34 (bm, 2H), 4.46-4.54 (m,
4H), 7.26-7.35 (m, 10H); 13C NMR (100 MHz, CDCl3) δ 28.5,
32.0, 36.4, 36.9, 37.4, 38.1, 51.1, 51.9, 52.9, 61.2, 70.8, 75.7,
76.4, 79.3, 79.7, 127.6, 127.6, 128.4, 138.1, 154.5, 172.1, 172.4;
IR (neat) 2974, 1693, 1665, 1397, 1160, 1118 cm-1; HRMS
(ESIþ) m/e calcd for [M þ H]þ C20H31N2O5 379.2233, found
379.2222.
tert-Butyl 3-(2-(Methoxy(methyl)amino)-2-oxoethyl)morpho-
line-4-carboxylate (1f). 2-(4-(tert-Butoxycarbonyl)morpholin-
3-yl)acetic acid (980 mg, 4.0 mmol, 1.0 equiv) was dissolved in
anhydrous CH2Cl2 (100 mL) under an argon atmosphere and
cooled to -15 °C. To this solution were added N,O-dimethyl-
hydroxylamine HCl (430 mg, 4.4 mmol, 1.1 equiv) and
3
N-methylmorpholine (0.49 mL, 4.4 mmol, 1.1 equiv) followed
by EDCI (840 mg, 4.4 mmol, 1.1 equiv). The reaction mixture
was then allowed to come to room temperature. After 2 h, the
reaction was again cooled to 0 °C and quenched by the
addition of an ice-cold 10% HCl solution (25 mL) and allowed
to stir at this temperature for 5 min. The reaction was diluted
with water (50 mL) and extracted with CH2Cl2 (ꢀ3). The
combined organic layers were washed with saturated NaH-
CO3 (ꢀ1), dried over Na2SO4, filtered, and concentrated.
Purification via SiO2 flash chromatography using 80%
EtOAc/hexanes as the eluent afforded the title compound as
a white solid (1.14 g, 99%): mp 71.9-72.7 °C; 1H NMR (400
MHz, CDCl3) δ 1.46 (s, 9H), 2.54-2.64 (bm, 1H), 2.97-3.17
(bm, 2H), 3.17 (s, 3H), 3.45 (dt, J = 5.8, 2.8 Hz, 1H), 3.58 (dd,
J = 11.7, 2.2 Hz, 1H), 3.71 (s, 3H), 3.71-3.88 (m, 3H),
4.41-4.46 (bm, 1H); 13C NMR (100 MHz, CDCl3) δ 28.4,
31.1, 32.1, 39.2, 48.5, 61.3, 66.8, 69.1, 80.1, 154.5, 171.9; IR
(neat) 2975, 1696, 1663, 1408, 1171, 1106 cm-1; HRMS
(ESIþ) m/e calcd for [M þ H]þ C13H25N2O5 289.1763, found
289.1770.
Experimental Section
(2S*,4R*)-tert-Butyl 4-(Benzyloxy)-2-(2-(methoxy(methyl)-
amino)-2-oxoethyl)pyrrolidine-1-carboxylate (1c). Warning:
Large amounts of diazomethane were used for this transformation.
Proper care should be taken when handling this highly explosive
reagent. All glassware used was free of cracks, scratches, or
ground-glass joints, and a blast shield was used. (2S*,4R*)-
4-(Benzyloxy)-1-(tert-butoxycarbonyl)pyrrolidine-2-carboxy-
lic acid (4.41 g, 13.8 mmol, 1.00 equiv) was taken into THF
(80 mL) with stirring and cooled to 0 °C with an ice bath. The
reaction solution was treated with TEA (2.09 mL, 15.1 mmol,
1.10 equiv) and allowed to react for 15 min to fully deprotonate
the carboxylic acid. With the addition of ethyl chlorofor-
mate (1.31 mL, 13.8 mmol, 1.00 equiv), a thick white precipi-
tate formed. Stirring was continued for 15 min and then
stopped. In a separate flask, an ice-cold ethereal solution of
diazomethane was prepared and, without stirring, was care-
fully decanted into the freshly prepared anhydride reaction
flask using a glass funnel. The reaction solution was lightly
stirred for 4 s, and then stirring was stopped. The mixture was
allowed to warm to room temperature and react overnight.
Any additional diazomethane was carefully quenched with
0.5 N acetic acid (25 mL). The dropwise addition of saturated
sodium bicarbonate regulated the solution back to a basic pH
8-9 with gentle stirring. The organic and aqueous layers were
separated. The organic phase was washed twice each with
saturated sodium bicarbonate and brine and then dried over
sodium sulfate. The solvent was evaporated under reduced
pressure and placed under high vacuum overnight. The diazo-
ketone (3.71 g, 10.7 mmol, 1.00 equiv) was taken into THF
(50 mL) and cooled to 0 °C. Foil was used to cover the reaction
flask so as to exclude light from the reaction solution. To this
was added freshly distilled N,O-dimethylhydroxylamine (1.96
g, 32.1 mmol, 3.00 equiv). In a separate foil covered flask, silver
trifluoroacetate (240 mg, 1.07 mmol, 0.100 equiv) was dis-
solved in TEA (30 mL). This solution was added to the
diazoketone mixture over 30 min. The reaction temperature
was allowed to slowly warm to room temperature and the
tert-Butyl Benzyl(3-(methoxy(methyl)amino)-3-oxopropyl)-
carbamate (1l). A round-bottomed flask was charged with ben-
zylamine (8.00 mL, 73.1 mmol, 3.00 equiv) and cooled to -40 °C.
Methyl acrylate (2.20 mL, 86.1 mmol, 1.00 equiv) was added
dropwise over 5 min, and the reaction was stirred at this tem-
perature (-40 °C) for 24 h. Excess benzylamine was distilled off
under reduced pressure. The remaining residue was dissolved
in methanol (50 mL), and di-tert-butyl dicarbonate (6.40 g,
29.2 mmol, 1.20 equiv) was added slowly. The reaction mixure
was stirred for another 30 min, and the solvent was removed in
vacuo. The concentrated reaction mixture was redissolved in
CH2Cl2 (300 mL) and washed with cold 10% HCl (100 mL ꢀ 2)
and brine(100 mL ꢀ 2). The organic layer was dried with MgSO4,
concentrated, and purified via SiO2 flash chromatography (25%
EtOAc in hexanes) to afford 6.62 g (88%) of the methyl ester as a
clear viscous oil. This oil was converted to the corresponding
Weinreb amide using the procedure reported by Williams et al.20
The methyl ester (6.62 g, 22.6 mmol, 1.00 equiv) and N,O-
dimethylhydroxylamine HCl (3.42 g, 35.0 mmol, 1.55 equiv)
3
were dissolved in anhydrous THF (40 mL) at room temperature
under N2. This mixture was cooled to -20 °C, and isopropyl-
magnesium chloride (34 mL, 68 mmol, 3.0 equiv, 2.0 M in THF)
was added dropwise over 10 min. The temperature was kept
J. Org. Chem. Vol. 75, No. 20, 2010 6803

Niphakis et al.
JOCArticle
between -10 and -20 °C for 30 min. After the reaction was
judged complete by TLC, it was quenched by the addition of a
saturated solution of NH4Cl (40 mL). The product was extracted
with EtOAc (ꢀ3), and the combined organic layers were dried
over Na2SO4, filtered, and concentrated. The crude product was
purified via SiO2 flashchromatography(40% EtOAc/hexanes) to
give 6.29 g (86%) of Weinreb amide 1l as a clear viscous oil: 1H
NMR (400 MHz, CDCl3) δ (1:1 mixture of rotamers) 1.44 (s,
9H), 1.50 (s, 9H), 2.58-2.63 (bm, 2H), 2.67-2.72 (bm, 2H),
3.15 (s, 6H), 3.43-3.48 (bm, 2H), 3.51-3.55 (bm, 2H), 3.63 (bs,
6H), 4.48 (s, 4H), 7.22-7.35 (m, 10H); 13C NMR (100 MHz,
CDCl3) δ 28.4, 31.2, 32.1, 42.8, 43.1, 50.6, 51.6, 61.2, 61.3,
79.8, 127.2, 127.3, 127.8, 128.4, 138.4, 138.8, 155.5, 155.8,
172.5, 172.9; IR (neat) 2974, 1693, 1664, 1413, 1366, 1167
cm-1; HRMS (ESIþ) m/e calcd for [M þ H]þ C17H27N2O4
323.1971, found 323.1957.
additional TMS-I (0.50-2.5 mmol) was added until all starting
material was consumed (TLC, 25% EtOAc in hexanes). After
20 min, the reaction was judged complete, and this mixture was
concentrated under reduced pressure and placed under vacuum
for 15 min.
Cyclization. The deprotected intermediates from methods
A-C were dissolved in MeOH (50 mL), and excess K2CO3
(a minimum of 5.0 equiv) was added. The reaction times
varied depending on the substrate (for enaminones 3d, 3g,
3j, 3k, 3m, and 3n reactions were stirred for 1 h; for enam-
inones 3a, 3l, 3p, 3o, 3r, and 3s reactions were stirred for 3 h;
for enaminones 3b, 3c, 3e, 3f, 3h, 3i, 3q, 3t, 3u, 3v, 3w, and 3x
reactions were stirred for 6 h). After the allotted time, CH2Cl2
was added, the reaction was suction filtered, and the organic
solvents were concentrated. To the solid residue was added
more CH2Cl2, and the precipitates were once again filtered
away. This residue was purified via flash chromatography to
provide pure enaminone. Large Scale Modification: When the
reaction was complete, the MeOH was removed in vacuo and
the remaining residue redissolved in brine. The product was
extracted with CH2Cl2 (3ꢀx). The combined organic layers
were dried over Na2SO4 and purified via SiO2 flash chromato-
graphy. Note: Several of the enaminones (e.g., 3a, 3d, 3j, 3k,
etc.) were water soluble and could not be purified by this
method.
(2R*,8aS*)-2-(Benzyloxy)-2,3,8,8a-tetrahydroindolizin-7(1H)-
one (3g). The title compound was obtained as a colorless oil
(method A, 94%; method B, 92%) after flash chromatography
(100% acetone): 1H NMR (400 MHz, CDCl3) δ 1.77 (ddd, J =
13.2, 11.2, 4.5 Hz, 1H), 2.35 (dd, J = 16.2, 16.2 Hz, 1H),
2.44-2.51 (m, 2H), 3.60 (d, J = 11.8 Hz, 1H), 3.70 (dd, J =
11.8, 4.9 Hz, 1H), 4.07 (dddd, J = 16.4, 10.9, 5.3, 5.3 Hz, 1H),
4.29 (dd, J = 4.5, 4.5 Hz, 1H), 4.51 (d, J = 11.8 Hz, 1H), 4.56 (d,
J = 11.8 Hz, 1H), 4.99 (d, J = 7.1 Hz, 1H), 7.17 (d, J = 7.1 Hz,
1H), 7.29-7.38 (m, 5H); 13C NMR (100 MHz, CDCl3) δ 38.5,
41.4, 55.4, 56.2, 71.0, 77.1, 97.9, 127.6, 128.0, 128.6, 137.5, 149.9,
191.9; IR (neat) 2929, 1631, 1579, 1460, 1306, 1099 cm-1; HRMS
(ESIþ) m/e calcd for [M þ H]þ C15H18NO2 244.1338, found
244.1332.
(2S*,4R*)-tert-Butyl 4-(Benzyloxy)-2-(2-oxobut-3-ynyl)pyr-
rolidine-1-carboxylate (2g). The Weinreb amide 1c (604 mg,
1.60 mmol, 1.00 equiv) was dissolved in anhydrous THF
(30 mL) under an argon atmosphere and cooled to 0 °C. To
this reaction vessel was added dropwise ethynylmagnesium
bromide reagent (16.0 mL, 7.98 mmol, 5.00 equiv, 0.5 M in
THF) and allowed to come to room temperature. After the
reaction was judged complete by TLC, it was quenched by the
addition of an ice-cold 10% HCl solution (15 mL) and allowed
to stir at this temperature for 5 min. The reaction was dilu-
ted with water and extracted with EtOAc (ꢀ3). The combined
organic layers were washed with saturated NaHCO3 (ꢀ1), dried
over Na2SO4, filtered, and concentrated. The title compound
was obtained as a colorless oil (503 mg, 92%) after flash
1
chromatography (20% EtOAc/hexanes): H NMR (400 MHz,
CDCl3) δ (1:1 mixture of rotamers) 1.46 (s, 9H), 1.48 (s, 9H),
1.77-1.87 (bm, 2H), 2.35 (dddd, J = 13.3, 7.7, 3.9, 1.0 Hz, 2H),
2.65-2.73 (bm, 2H), 3.16 (bd, J = 15.3 Hz, 1H), 3.26 (bd, J =
16.9 Hz, 2 H), 3.35-3.42 (m, 3H), 3.56 (bd, J = 11.3 Hz, 1H),
3.78 (bd, J = 11.6 Hz, 1H), 4.06-4.10 (m, 2H), 4.30-4.38 (bm,
2H), 4.44-4.55 (bm, 2H), 4.50 (bd, J=10.2 Hz, 2H), 7.28-7.36
(m, 10H); 13C NMR (100 MHz, CDCl3) δ 28.5, 36.9, 38.2, 49.7,
50.9, 51.1, 51.9, 52.2, 70.9, 75.6, 76.1, 78.8, 79.1, 79.7, 80.3, 81.5,
127.6, 127.7, 127.7, 128.5, 137.9, 154.4, 154.5, 185.0, 185.3;
IR (neat) 2976, 2091, 1685, 1399, 1367, 1162 cm-1; HRMS
(ESIþ) m/e calcd for [M þ H]þ C20H26NO4 344.1862, found
344.1854
General Procedures for Cyclic Enaminone Formation. Depro-
tection. Method A: The Boc-amino ynone 2 (0.50 mmol) was
dissolved in a 4 N HCl-dioxane solution (1.5 mL) and allowed
to react for 15 min. After this time, the dioxane and excess HCl
were allowed to evaporate while air was passed over the reaction
mixture. The remaining solid was placed under vacuum for
15 min and carried on to the cyclization step without further
purification. Method B: The ynone (0.50 mmol, 1.0 equiv) was
dissolved in 5 mL of 98% formic acid under a N2 atmosphere,
and NaI (230 mg, 1.5 mmol, 3.0 equiv) was added. Note: For
terminal ynones the reaction was allowed to stir for 6 h. When
internal ynones were used the reaction was left for 24 h. The
solvent was removed by passing N2 over the reaction mixture.
The remaining residue was placed under vacuum for 15 min
and was carried on to the cyclization step without further
purification.
(E)-2-(4-Iodo-2-oxobut-3-enyl)pyrrolidinium Iodide (7a).
Ynone 2d (0.50 mmol, 1.0 equiv) was dissolved in 98% formic
acid (5.0 mL) under a N2 atmosphere, and NaI (230 mg, 1.5
mmol, 3.0 equiv) was added. After 6 h, Et2O (200 mL) was slowly
added to the reaction mixture allowing a precipitate to form. The
precipitate was filtered and washed several times with Et2O. The
filtered precipitate was air-dried and used without further puri-
fication. To obtain crystals for X-ray analysis, the precipitate was
recrystallized from formic acid. See the Supporting Information
for the X-ray crystal structure report: 1H NMR (400 MHz,
CD3OD) δ 1.59-1.69 (m, 1H), 1.82-1.92 (m, 1H), 1.94-2.04
(m, 1H), 1.12-2.20 (m, 1H), 2.97 (dd, J = 18.9, 9.8 Hz, 1H),
3.13-3.24 (m, 3H), 3.76-3.83 (m, 1H), 7.16 (d, J = 15.2 Hz, 1H),
8.17 (d, J = 15.2 Hz, 1H); 13C NMR (100 MHz, CD3OD) 24.6,
1
31.2, 42.6, 46.6, 56.8, 102.8, 145.2, 196.4; H NMR (400 MHz,
(CD3)2SO) δ 1.49-1.58 (m, 1H), 1.74-1.95 (m, 2H), 2.04-2.12
(m, 1H), 3.04 (dd, J = 18.7, 9.2 Hz, 1H), 3.04-3.17 (m, 2H), 3.17
(dd, J = 18.7, 4.3 Hz, 1H), 3.71-3.78 (m, 1H), 7.21 (d, J = 15.3
Hz, 1H), 8.26 (d, J = 15.3 Hz, 1H), 8.28 (bs, 1H), 8.85 (bs, 1H);
13C NMR (100 MHz, (CD3)2SO) δ 23.1, 29.7, 41.4, 45.0, 54.3,
104.7, 143.7, 195.4.
Large-Scale Modification. When >1.0 g of ynone is used, the
desired deprotected amine can be isolated as the ammonium
salt by pouring the reaction mixture into ether and collecting
the precipitate via filtration. Method C: The ynone (0.50 mmol,
1.0 equiv) was dissolved in anhydrous CH2Cl2 (10 mL) under an
argon atmosphere and cooled to -78 °C. A solution of TMS-I
(0.50 mmol, 1.0 equiv) in anhydrous CH2Cl2 (5 mL) was then
added dropwise at this temperature. After 20 min at this
temperature, the reaction was allowed to warm to 0 °C, and
Acknowledgment. This work was supported by National
Institutes of Health Grant Nos. CA90602, GM069663,
GM076302, and GM081267, the Kansas Masonic Cancer
Research Institute, the University of Minnesota through
the Vince and McKnight Endowed Chairs, and an ACS
6804 J. Org. Chem. Vol. 75, No. 20, 2010

Niphakis et al.
JOCArticle
Division of Medicinal Chemistry Predoctoral Fellowship
(to M.J.N). We thank Kathryn Pietsch for help preparing
2m-o and the Georg group for helpful discussions and a
critical reading of this manuscript. Victor G. Young, Jr., is
thanked for determining the X-ray crystal structure of
compound 7a.
Supporting Information Available: Representative experi-
mental procedures; complete crystallographic data for
compound 7a; characterization data for all new compounds;
spectra for compounds 1a-m, 2a-x, 3a-x, 7a, and 14-16;
and NMR deuterium labeling study spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 75, No. 20, 2010 6805