Cycloadditions of siloxy alkynes with 1,2-diazines: From reaction discovery to identification of an antiglycolytic chemotype

Article abstract of DOI:10.1002/anie.201305711

Cycloaddition uncovered: The title reaction produces novel polycyclic compounds with high efficiency and excellent diastereoselectivity under mild reaction conditions. A small-molecule library, synthesized using this reaction, yielded a novel chemotype which inhibited glycolytic ATP production by blocking glucose uptake in CHO-K1 cells. DMF=N,N-dimethylformamide, Tf= trifluoromethanesulfonyl, TIPS=triisopropylsilyl. Copyright

Full text of DOI:10.1002/anie.201305711

.
Angewandte
Communications
DOI: 10.1002/anie.201305711
Synthetic Methods
[2+2+2] Cycloadditions of Siloxy Alkynes with 1,2-Diazines: From
Reaction Discovery to Identification of an Antiglycolytic Chemotype**
Timothy J. Montavon, Yunus E. Tꢀrkmen, Noumaan A. Shamsi, Christopher Miller,
Chintan S. Sumaria, Viresh H. Rawal,* and Sergey A. Kozmin*
The synthesis of new nitrogen-containing heterocycles plays
a pivotal role in chemical biology and medicinal chemistry, as
reflected by their many applications in the development of
pharmacological probes and drugs.[1] Despite notable prog-
ress, there is a significant need for the identification of new
nitrogen heterocycles which target previously unexplored
regions of biogenic chemical space. Among the many possible
synthetic strategies to such compounds, cycloadditions involv-
ing C–N multiple bonds are particularly attractive as they
Scheme 1. [2+2+2] cycloaddition of phthalazine (1) with the 1-siloxy-1-
hexyne 2. Tf=trifluoromethanesulfonyl, TIPS=triisopropylsilyl.
generate complex cyclic products by simultaneous formation
of multiple bonds starting from readily available precursors.[2]
Herein, we describe the discovery and development of
a formal [2+2+2] cycloaddition of siloxy alkynes with
phthalazines, a process that had not been previously described
for either 1,2-diazines or electron-rich alkynes.[3–7] This effort
has not only afforded heterocyclic products with a unique
pentacyclic ring system but has also enabled the identification
of a novel chemotype that inhibits glycolytic ATP production
by direct blockage of glucose uptake in CHO-K1 cells. As
a result of the prevalence of the Warburg effect in many
human cancers, such compounds may prove useful in the
development of new therapeutics which target reprogrammed
energy metabolism of rapidly proliferating cells.[8]
Our study began by examining the reaction of phthalazine
(1) with the siloxy alkyne 2 in the presence of common
Brønsted acids. While no reaction between 1 and 2 was
observed in the absence of such additives, even at elevated
temperatures, we found that addition of simple pyridinium
salts promoted the formation of a new pentacyclic product (3;
Scheme 1).
yield. While most of the known [2+2+2] cycloadditions
typically require the presence of a transition-metal catalyst,[9]
the present method promotes the condensation under
remarkably mild reaction conditions, using only a simple,
weak Brønsted acid. The excellent diastereoselectivity of this
transformation is also highly noteworthy. The atom connec-
tivity within the reaction product was initially determined to
be that in 3 and is based on extensive use of NMR
spectroscopic methods. Ultimately, the structure was secured
and the relative stereochemistry of the three newly created
stereogenic centers was defined through X-ray crystallo-
graphic analysis (see below).
Interestingly, while a range of substituted mono- and
bis(pyridinium) trifluoromethanesulfonimides were found to
be effective as reaction promoters, the use of only HNTf2, in
the absence of pyridine, produced 3 with lower efficiency
(48% yield) and diminished diastereoselectivity (83:17).
Furthermore, the use of either pyridinium chloride, pyridi-
nium p-toluenesulfonate, or pyridinium triflate substantially
decreased product yields or prevented the reaction. These
results highlight the importance of the weakly nucleophilic
trifluoromethanesulfonimide counterion,[10] an observation
consistent with those made in the course of previous studies of
Brønsted acid promoted transformations of siloxy alkynes.[3]
Having established a general reaction protocol, we began
a detailed investigation of the scope of this [2+2+2] cyclo-
addition. With regard to siloxy alkyne substitution, we found
that both alkyl and aryl substituents are well tolerated, thus
providing the expected products in good yields with high
levels of diastereoselection (Table 1, entries 1–4). When the
steric bulk of the substituent in direct proximity to the alkyne
is increased, the yield of the reaction is lowered slightly but
the diastereoselectivty remains relatively unaffected. For
instance, the siloxy alkynes 6 and 8 afforded the expected
products 7 (73%) and 9 (69%), respectively (Table 1,
entries 2 and 3). Furthermore, the use of 1-siloxy-propyne
After examining a range of mono- and bis(pyridinium)
salts in various solvents, we determined the optimum protocol
to entail the use of a stoichiometric amount of pyridinium
trifluoromethanesulfonimide in CH2Cl2 at room temperature,
thus producing the lactam 3 as a single diastereomer in 77%
[*] Dr. T. J. Montavon, Dr. Y. E. Tꢀrkmen, N. A. Shamsi, C. Miller,
C. S. Sumaria, Prof. Dr. V. H. Rawal, Prof. Dr. S. A. Kozmin
Chicago Tri-Institutional Center for Chemical Methods and Library
Development, Department of Chemistry, The University of Chicago
Chicago, IL 60637 (USA)
E-mail: vrawal@uchicago.edu
skozmin@uchicago.edu
[**] We are grateful for financial support from the National Institutes of
Health (P50 GM086145 and R01GM069990), and the Chicago
Biomedical Consortium with support from the Searle Funds at the
Chicago Community Trust. We thank Dr. Ian Steele for X-ray
crystallographic analysis of 37.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201305711.
13576
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 13576 –13579

Angewandte
Chemie
Table 1: Scope of the [2+2+2] cycloaddition of phthalazines and siloxy
alkynes.
(Table 1, entry 6). The lower efficiency of this reaction may be
attributed to the decreased electrophilicity of the phthalazine
moiety resulting from the presence of the two alkyl substitu-
ents. Both the halogen-containing (16) and heteroaromatic
(18) diazines served as excellent substrates in the [2+2+2]
cycloaddition, thus providing the expected products 17 (73%)
and 19 (81%), respectively (Table 1, entries 7 and 8). Finally,
1-phenyl phthalazine (20) was found to participate in this
transformation in a regio- and diastereoselective manner, thus
providing a single [2+2+2] adduct in 61% yield (Table 1,
entry 9). Simple pyridazines were unreactive under the
current reaction conditions. This observation is consistent
with the lower reactivity of pyridazines compared to that of
phthalazines in other reactions, including [4+2] cycloaddi-
tions with electron-rich dienophiles.
We next examined a possibility of conducting a three-
component reaction using two different 1,2-diazines. Reac-
tion of 2 with 1 and 20 delivered roughly equal quantities of all
four possible products (Table 2, entry 1), which was expected
Table 2: Exploring three-component [2+2+2] cycloadditions of phtha-
lazines and siloxy alkynes.
[a] Molar ratios based on yields of individually isolated product isomers
based on tentative structural assignments. [b] Combined yield (over
48 h) of isolated products for reaction performed on a 0.2 mmol scale.
For all sets of products, only one diastereomer was found to be present
by 1H NMR (500 MHz) analysis of the crude reaction mixture.
[a] Yields refer to those of the major diastereomer isolated for each
reaction. [b] Diastereomeric ratios were determined by 1H NMR
(500 MHz) analysis of the crude reaction mixture prior to purification.
(10) resulted in efficient formation of the lactam 11 with high
diastereoselectivity (d.r. > 98:2). Taken together, these results
suggest that a wide range of siloxy alkyne substituents would
be well tolerated in this reaction.
The scope of the reaction was next evaluated with respect
to a series of substituted 1,2-diazine derivatives. For instance,
benzo[g]phthalazine (12) provided the expected product 13 in
good yield with high diastereoselection (Table 1, entry 5). The
reaction of the alkyne 2 with 6,7-dimethyl phthalazine (14)
gave the lactam 15 in 65% yield and 91:9 diastereoselectivity
given the small electronic perturbation provided by the
phenyl substituent at C2. However, treatment of 2 with 1 and
6,7-dichlorophthalazine (16) with pyridinium trifluorome-
thane-sulfonimide afforded, predominantly, the hybrid prod-
uct 24, as well as smaller quantities of the cycloadducts 17 and
25 (Table 2, entry 2). Finally, the use of 12 in a similar three-
component reaction with 1 and 2 resulted in the formation of
only one (26) of the hybrid products as the major product
along with the cycloadduct 13 (Table 2, entry 3). These results
highlight the unique ability of this process to discriminate
Angew. Chem. Int. Ed. 2013, 52, 13576 –13579
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13577

.
Angewandte
Communications
between two similar but electronically differentiated 1,2-
diazine moieties.
reaction pathways affords the final cycloaddition product.
To demonstrate further the synthetic utility of the novel
structural types obtained by cycloadditions of siloxy alkynes
and 1,2-diazines, we have developed a route to selectively
functionalize the N-acyl hydrazone moiety of the observed
[2+2+2] adducts (Scheme 3). The two diazine moieties of 3
are chemically differentiated, one a hydrazone and the other
an N-acyl hydrazone. While hydrazones are notoriously
unreactive towards nucleophilic additions, N-acyl hydrazones
are quite reactive and have proven to be valuable substrates
for the synthesis of diverse classes of nitrogen-containing
compounds.[12] We found that treatment of 3 with a slight
excess of allyl trichlorosilane resulted in highly chemo- and
diastereoselective allylation at the azomethine carbon atom
of the N-acyl hydrazone moiety, thus providing the allylated
product 35 in 86% yield.[13] Treatment of 35 with either
benzoyl chloride or benzyl isocyanate afforded the expected
acylation products, amide 36 or hydrazide 37, respectively.
The hydrazide 37 was highly crystalline and its structure was
determined by X-ray crystallography, which provided unam-
biguous stereochemical assignment of the initial cycloadduct
3 and subsequent allylation product 35.
Two mechanistic pathways can explain the formation of
pentacyclic lactams originating from [2+2+2] cycloadditions
of phthalazines with siloxy alkynes (Scheme 2). The reaction
can proceed by initial formation of the ketenium ion 28
through protonation of the siloxy alkyne by the pyridinium
salt (pathway A). This electrophilic intermediate can be
We next employed this chemo- and diastereoselective
elaboration strategy to assemble a representative collection of
56 structurally diverse compounds,[11] which was then eval-
uated for their ability to inhibit aerobic glycolysis in cancer
cells. To this end, we employed our recently developed
screening approach, which is based on monitoring glycolytic
ATP production in live cells with chemically impaired
mitochondria.[14] We performed the screen by subjecting
CHO-K1 cells, treated with antimycin A, to each member of
the newly assembled chemical library[11] and monitored
production of ATP using a well-established luciferase-based
protocol. The structure of the most potent inhibitor 38 is
shown in Figure 1a.
This compound elicited the expected dose-dependent
ATP suppression profile shown in Figure 1b. The ATP
production was inhibited much more efficiently when cells
were treated with a combination of 38 and antimycin A
(IC50 = 11.4 mm). To further validate inhibition of aerobic
Scheme 2. Proposed reaction mechanism.
intercepted by 1 to give the enol silane 29, a hetero-
diene with nucleophilic and electrophilic carbon
atoms at opposite ends. Addition of a second phtha-
lazine moiety to 29 would trigger cyclization of the
central ring to afford the lactam 31 upon protodesi-
lation. Alternatively, the process can begin with
protonation of phthalazine by the pyridinium catalyst
(pathway B). The resulting phthalazinium ion would
then react with the alkyne 32 to give the ketenium ion
33. Interception of this intermediate by another
molecule of 1 would give the ammonium ion 34,
which can trigger protodesilylation of the enol silane
moiety to give the observed product 31. Both
mechanistic pathways are consistent with the results
of deuterium-labeling experiments performed using
deuteropyridinium triflate, which resulted in exclu-
sive deuterium incorporation at the a-carbon atom of
the N-acyl hydrazone moiety.[11] Further mechanistic
Scheme 3. Subsequent chemoselective functionalization. Bn=benzyl,
studies are required to establish which of the two DMF=N,N-dimethylformamide.
13578 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 13576 –13579

Angewandte
Chemie
[2] Topics in Heterocyclic Chemistry of Vol. 12, Synthesis
of Heterocycles via Cycloadditions I (Ed.: A. Hassner),
Springer, Berlin, 2008; b) Topics in Heterocyclic
Chemistry of Vol. 13, Synthesis of Heterocycles via
Cycloadditions II (Ed.: A. Hassner), Springer, Berlin,
2010.
[3] For other Brønsted acid promoted reactions of siloxy
alkynes, see: a) L. Zhang, S. A. Kozmin, J. Am. Chem.
Soc. 2004, 126, 10204 – 10205; b) J. Sun, S. A. Kozmin,
J. Am. Chem. Soc. 2005, 127, 13512 – 13513; c) L.
Zhang, J. Sun, S. A. Kozmin, Tetrahedron 2006, 62,
11371 – 11380.
[4] For review of ynolates and siloxy alkynes, see: a) M.
Shindo, Tetrahedron 2007, 63, 10 – 36; For reviews of
ynamides and ynamines, see: b) C. A. Zificsak, J. A.
Mulder, R. P. Hsung, C. Rameshkumar, L.-L. Wei,
Tetrahedron 2001, 57, 7575 – 7606; c) G. Evano, A.
Coste, K. Jouvin, Angew. Chem. 2010, 122, 2902 – 2921;
Angew. Chem. Int. Ed. 2010, 49, 2840 – 2859.
[5] For other cycloadditions of alkoxy and siloxy alkynes,
see: a) R. L. Danheiser, S. K. Gee, J. Org. Chem. 1984,
49, 1672; b) C. J. Kowalski, G. S. Lal, J. Am. Chem.
Soc. 1988, 110, 3693 – 3695; c) R. F. Sweis, M. P.
Schramm, S. A. Kozmin, J. Am. Chem. Soc. 2004,
126, 7442 – 7443; d) T. B. Clark, K. A. Woerpel, Org.
Lett. 2006, 8, 4109 – 4112; e) X. Qi, J. M. Ready,
Angew. Chem. 2008, 120, 7176 – 7178; Angew. Chem.
Int. Ed. 2008, 47, 7068 – 7070; f) Y. E. Tꢀrkmen, T. J.
Montavon, S. A. Kozmin, V. H. Rawal, J. Am. Chem.
Soc. 2012, 134, 9062 – 9065.
Figure 1. a) Chemical structure of 38. b) Inhibition of cellular ATP levels in PC3
cells by 38 in the presence or absence of Complex III inhibitor, antimycin A.
c) Inhibition of lactate production in A549 cells upon treatment with 38.
Cytochalasin B was used as a positive control. d) Inhibition of glucose uptake in
CHO-K1 cells by 38 using 3H 2-deoxyglucose scintillation experiments. All values
are presented as a percentage of the vehicle treated samples. Each value is the
mean SEM of duplicate values from a representative experiment.
[6] For reviews of cycloadditions of heterocyclic aza
dienes, see: a) D. L. Boger, Tetrahedron 1983, 39,
2869; b) D. L. Boger, Chem. Rev. 1986, 86, 781 – 793.
glycolysis, we measured the effects of 38 on lactate produc-
[7] For cycloadditions of 1,2-diazines with two equivalents of
ynamines, see: a) E. Oishi, N. Taido, A. Miyashita, S. Sato, S.
Ohta, H. Ishida, K. Tanji, T. Higashino, Chem. Pharm. Bull.
1991, 39, 1713 – 1718; b) E. Oishi, K.-T. Iwamoto, T. Okada, S.
Suzuki, K. Tanji, A. Miyashita, T. Higashino, Chem. Pharm. Bull.
1994, 42, 2219 – 2224.
tion, which was found to decrease in a dose dependent-
manner (Figure 1c). Finally, we found that 38 inhibited
cellular uptake of 3H-labeled 2-deoxyglucose with IC50 =
2.2 mm (Figure 1d), thus confirming the antiglycolytic effect
of this compound.
[8] M. G. Vander Heiden, L. C. Cantley, C. B. Thompson, Science
2009, 324, 1029 – 1033.
[9] a) P. R. Chopade, J. Louie, Adv. Synth. Catal. 2006, 348, 2307 –
2327; b) T. Shibata, K. Tsuchikama, Org. Biomol. Chem. 2008, 6,
1317 – 1323; c) G. Domꢁnguez, J. Pꢂrez-Castells, Chem. Soc. Rev.
2011, 40, 3430 – 3444.
[10] S. Antoniotti, V. Dalla, E. DuÇach, Angew. Chem. 2010, 122,
8032 – 8060; Angew. Chem. Int. Ed. 2010, 49, 7860 – 7888.
[11] For details, see the Supporting Information.
[12] For recent reviews on hydrazones and N-acyl hydrazones, see
a) M. Sugiura, S. Kobayashi, Angew. Chem. 2005, 117, 5306 –
5317; Angew. Chem. Int. Ed. 2005, 44, 5176 – 5186; b) G. K.
Friestad, Eur. J. Org. Chem. 2005, 3157 – 3172; c) R. Brehme, D.
Enders, R. Fernandez, J. M. Lassaletta, Eur. J. Org. Chem. 2007,
5629 – 5660; d) R. Lazny, A. Nodzewska, Chem. Rev. 2010, 110,
1386 – 1434.
In summary, we have described highly efficient and
diastereoselective [2+2+2] cycloadditions of siloxy alkynes
with phthalazines. This synthetic process is promoted under
mild reaction conditions by simple pyridinium salts and
provides rapid access to new, complex heterocyclic products,
richly adorned with handles for further functionalization.
Cellular screen of a representative collection of such struc-
turally diverse compounds enabled identification of a new
chemical inhibitor of aerobic glycolysis.
Received: July 2, 2013
Revised: September 13, 2013
Published online: November 7, 2013
Keywords: cycloaddition · inhibitor · polycycles · siloxyalkynes ·
[13] a) S. Kobayashi, R. Hirabayashi, J. Am. Chem. Soc. 1999, 121,
6942 – 6943; b) R. Hirabayashi, C. Ogawa, M. Sugiura, S.
Kobayashi, J. Am. Chem. Soc. 2001, 123, 9493 – 9499; c) C.
Ogawa, M. Sugiura, S. Kobayashi, J. Org. Chem. 2002, 67, 5359 –
5364.
.
synthetic methods
[1] A. F. Pozharskii, A. Soldatenkov, A. R. Katritzky, Heterocycles
in Life and Society, Wiley, Hoboken, 2011.
[14] O. Ulanovskaya, J. Cui, S. J. Kron, S. A. Kozmin, Chem. Biol.
2011, 18, 222 – 230.
Angew. Chem. Int. Ed. 2013, 52, 13576 –13579
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13579