Formation of unusual seven-membered heterocycles incorporating nitrogen and sulfur by intramolecular conjugate displacement

Article abstract of DOI:10.1021/jo101539m

Baylis-Hillman alcohols derived from methyl acrylate or acrylonitrile and carrying an N-Boc group β to the hydroxyl (CH(OH)CHNBoc) can be converted into unusual seven-membered heterocycles containing both nitrogen and sulfur by O-acylation (AcCl or EtOCOCl), N-deprotection (CF₃CO₂H), and reaction with CS₂. In a modification of this process, when the original nitrogen is substituted in the form PhSCH₂CON(Me), an azepine derivative is then generated. The ring closures occur by intramolecular conjugate displacement.

Full text of DOI:10.1021/jo101539m

pubs.acs.org/joc
SCHEME 1. 5-Exo Trigonal Cyclization
Formation of Unusual Seven-Membered
Heterocycles Incorporating Nitrogen and Sulfur by
Intramolecular Conjugate Displacement
Zhenhua Chen and Derrick L. J. Clive*
Chemistry Department, University of Alberta, Edmonton,
Alberta T6G 2G2, Canada
derrick.clive@ualberta.ca
Received August 5, 2010
SCHEME 2. Intermolecular Addition Followed by ICD
(to give 9),1 presumably because such a process would violate
Baldwin’s rules.3 In this particular case, an alternative path-
way, attack on the methyl ester carbonyl, was followed (7 f 10;
Scheme 1). The fact that compounds of type 7 would
appear to be inherently protected against the intramolecular
conjugate displacement pathway suggested that if an unsa-
turated electrophile could be found that is attacked rapidly
by nitrogen to generate a new nucleophilic species, then cycliza-
tion would occur as illustrated in Scheme 2, 11 f 12 f 13,
where XdY represents an electrophile, either by itself or
as a subunit of a larger structure. We have examined a
number of potential electrophiles (PhNCO, PhNCS, SO2,
BnNdCH2, (Cl3C)2CdNBn, CO2, CS2, and H2NCN) and
have found that CS2 satisfies the requirements and gives
access to a range of 2-thioxo-1,3-thiazepine derivatives,
unusual heterocycles of type 14. We can find only one
example (14; EWG = R0 = R00 = H, R = þPPh3)4;formed
as a minor (15%) byproduct;of a substance closely related
to this compound class.
Baylis-Hillman alcohols derived from methyl acrylate or
acrylonitrile and carrying an N-Boc group β to the hydroxyl
(CH(OH)CHNBoc) can be converted into unusual seven-
membered heterocycles containing both nitrogen and
sulfur by O-acylation (AcCl or EtOCOCl), N-deprotection
(CF3CO2H), and reaction with CS2. In a modification of
this process, when the original nitrogen is substituted in
the form PhSCH2CON(Me), an azepine derivative is then
generated. The ring closures occur by intramolecular con-
jugate displacement.
Intramolecular conjugate displacement1;the type of cy-
clization summarized by eq 1;provides a method for mak-
ing both nitrogen heterocycles and carbocycles along the
lines of eqs 21 and 3,2 which show particular examples from
the many that have been reported from this laboratory.1,2
The N-Boc amines we have used were all readily accessible
via classical Baylis-Hillman condensation between the ap-
propriate aldehyde and methyl acrylate (second column of
Table 1, entries 1-7, 10, and 12) or acrylonitrile (Table 1,
entries 8 and 9) in the presence of DABCO at room tem-
perature (2-4 days, >69%). While this classical approach is
not the only route to Baylis-Hillman alcohols,5 it was
convenient for the present study. The alcohols (except that
(1) Clive, D. L. J.; Li, Z.; Yu, M. J. Org. Chem. 2007, 72, 5608–5617.
(2) Wang, L.; Prabhudas, B.; Clive, D. L. J. J. Am. Chem. Soc. 2009, 131,
6003–6012.
In studying the nitrogen case, we found that one type
of starting material, compound 7, failed to undergo ICD
7014 J. Org. Chem. 2010, 75, 7014–7017
Published on Web 09/15/2010
DOI: 10.1021/jo101539m
r
2010 American Chemical Society

Chen and Clive
JOCNote
TABLE 1. Baylis-Hillman Reaction, Acylation, and ICD Process
ai-PrNEt2 gave 4d (62%) and 4e (8%); use of 2,6-lutidine gave only 4d (66%). b2,6-Lutidine gave only 7d (65%); use of i-PrNEt2 gave 7d (38%) and 7e
(30%). cArbitrary stereochemistry. Reagents and conditions: (a) methyl acrylate, DABCO, CH2Cl2; (b) AcCl, pyridine, CH2Cl2; (c) TFA, CH2Cl2 and
CS2, i-PrNEt2, MeCN; (d) TFA, CH2Cl2 and CS2, 2,6-lutidine, MeCN; (e) acrylonitrile, DABCO, CH2Cl2; (f) EtOCOCl, pyridine, CH2Cl2; (g) NaH,
MeI, THF.
of Table 1, entries 10 and 11) were then acylated (>73%
yield), using AcCl/pyridine or (for entries 10 and 11) EtO-
COCl/pyridine.
The acylated compounds 1c-11c were each treated with
CF3CO2H (CH2Cl2, room temperature) to remove the
N-protecting group and release the trifluoroacetate salt of
byproduct was isolated. The results summarized in Table 1
show that, with an ester as the electron-withdrawing group
and acetate as the leaving group, secondary aliphatic6 amines
undergo the ICD process, but a careful choice of base may be
required to optimize the selectivity in favor of the 2-thioxo-
1,3-thiazepine derivatives because a competing side reaction
is formation of a 2-thiazolidinethione. For example, when 7c
was treated with 2,6-lutidine and CS2 (1.5 h), 7d was formed
€
the corresponding amine. Then CS2, followed by Hunig’s
base or 2,6-lutidine, was added to a solution of the crude amine
trifluoroacetate in MeCN. Sometimes 2,6-lutidine gave a
€
better result; in the case of 4c, use of Hunig’s base afforded
a small amount of 4e (8%), but with 2,6-lutidine none of this
€
in 65% yield, but with Hunig’s base (ca. 12 h) 7e was isolated
(30%) in addition to 7d (38%).
The nitrogen of the starting amine can be part of a ring or a
chain. In the latter case, the reaction is more selective for
formation of the seven-membered ring when the deprotected
nitrogen is a secondary amine (4d/4e, 6d, 7d/7e); when the
deprotected nitrogen is primary, a major side product is the
2-thiazolidinethione (5e). We noted that if the nitrogen
carries an aryl substituent,6 the ICD did not occur, presum-
ably because of the suppressed nucleophilicity of anilines
compared with aliphatic amines.
(3) While the pathway shown in 8 violates Baldwin’s rules, we anticipated
that 9 might still be accessible by the alternative sequence i f ii f 9.
(4) Isoda, T.; Hayashi, K.; Tamai, S.; Kumagai, T.; Nagao, Y. Chem.
Pharm. Bull. 2006, 54, 1616–1619.
(5) For another route, based on selenides, see refs 1 and 2.
(6) We examined one case where the deprotected nitrogen carried a p-
methoxyphenyl ring; no addition to CS2 occurred (see the Supporting
Information).
J. Org. Chem. Vol. 75, No. 20, 2010 7015

Chen and Clive
JOCNote
As two products are sometimes formed, we determined if
the reaction was under thermodynamic or kinetic control.
When compound 7d was subjected to the original conditions
of the cyclization, it remained unchanged; likewise 7e
(generated by use of i-PrNEt2 instead of 2,6-lutidine) was
also unchanged. Consequently, the cyclizations must be
kinetically controlled.
SCHEME 3. Reaction with PhNCO
The outcome of the reaction is also sensitive to the nature
of both the electron-withdrawing substituent and the allylic
leaving group. The nitrile-acetate system 9c appeared to
give only a five-membered ring (58%), for which the indi-
cated stereochemistry is an arbitrary assignment, while the
corresponding methyl ester-acetate 2c provided an almost
identical yield (56%) of the seven-membered ring. In the case
of nitrile 8c, reaction at -40 °C gave the single product 8d,
while at room temperature an inseparable 4:1 mixture of two
compounds, presumably 8d and the corresponding pyrrolo-
[1,2-c]thiazole-3-thione, were formed. The allylic carbonate
10c afforded the cyclized product 1d in very high yield;95%
versus 65% for the corresponding acetate 1c. Since in this
case the use of a carbonate leaving group favored formation
of the seven-membered ring, we examined the ethyl carbo-
nate derived from 9b and found (Table 1, entry 11) that a 5:1
mixture of the seven-membered and five-membered products
was formed at room temperature; this ratio was improved to
10:1 at -40 °C, implying the trend that a carbonate leaving
group favors formation of the seven-membered ring. We also
evaluated the consequences of using a poor leaving group, by
making the methyl ether 12c; however, this produced only
12d, the product of Michael addition. Such a pathway had
been observed before in related experiments done in this
laboratory:2 use of a poor leaving group, as in 15, still produced
the ICD product on treatment with Cs2CO3,2 although at a
lower rate than for the corresponding acetate, while the related
substrate 16 underwent both Michael addition without loss of
the siloxy group and the normal ICD process.2
SCHEME 4. Formation of Lactam
ICD process would be increased if heterocycles containing only
a single heteroatom in the ring could be generated. To this end,
we examined nitroethene,7 (MeO2C)2CN2/Rh2(OAc)4,8 and
(MeO2C)2CHBr9 as potential formal electrophiles to capture
the deprotected nitrogen. Our experiments with these com-
pounds were all unpromising, except for the fact that they
steered our attention toward a modified approach, which is
illustrated in Scheme 4, and which simply involves attaching
the final nucleophile (see the PhSCH2CO subunit of 22) at an
earlier stage.
As indicated above, we had examined a number of poten-
tial electrophiles and some comment on a few of them is in
N-acylation of 2-(methylamino)ethanol with PhSCH2-
COCl10 (pyridine, 90%) and oxidation (Parikh-Doering,
62%) gave 23, setting the stage for generation of the desired
ICD substrate. This was accomplished as follows: reaction of
23 with methyl acrylate in the presence of DABCO gave
alcohol 24, and oxidation (MCPBA) and acetylation (AcCl,
pyridine) then afforded sulfone 26.
Surprisingly, on treatment with Cs2CO3 in 10:1 THF-
MeOH the sulfone was converted (61%) into the enamide 28,
which was a single isomer whose relative stereochemistry was
not established. One possible pathway for its formation is via
the expected ICD product 27, followed by double bond
migration. Alternatively, elimination of acetate from 26,
followed by simple Michael addition, would also lead to
€
order. Reaction of 1c with CF3CO2H and then with Hunig’s
base and PhNCO served only to produce the urea 17 (71%),
formed by simple intermolecular addition (Scheme 3). Clo-
sure to the ICD product 18 could be achieved, however, by
further reaction with (Me3Si)2NK (THF, 0 °C), but only in
39% yield. Interestingly, the action of DBU and LiBr on 17
gave 20 (35%). In the pathway from 17 to 20 the first step
might be an intermolecular conjugate displacement to afford
19, where the nucleophile is Br- or DBU.
Each of the final products listed in Table 1 is a cyclic
dithiocarbamate, and it was obvious that the utility of the present
(7) Use of nitroethene gave only the result of an initial intermolecular
Michael addition (100%) but none of the ICD product.
(8) Cf.: (a) Yamagata, K.; Okabe, F.; Yamazaki, M. J. J. Prakt. Chem.
1999, 341, 562–567. (b) Bagley, M. C.; Buck, R. T.; Hind, S. L.; Moody, C. J.
J. Chem. Soc., Perkin Trans. 1 1998, 591–600. (c) Falorni, M.; Dettori, G.;
Giacomelli, G. Tetrahedron: Asymmetry 1998, 9, 1419–1426. (d) Gibe, R.;
Kerr, M. A. J. Org. Chem. 2002, 67, 6247–6249. (e) Yang, M.; Wang, X.; Li,
H.; Livant, P. J. Org. Chem. 2001, 6, 6729–6733.
(9) Wolfe, S.; Ro, S.; Kim, C.-K.; Shi, Z. Can. J. Chem. 2001, 79, 1238–
1258.
(10) Kennedy, M. G.; Moody, C. J.; Rees, C. W.; Thomas, R. J. Chem.
Soc., Perkin Trans. 1 1991, 2499–2504.
7016 J. Org. Chem. Vol. 75, No. 20, 2010

Chen and Clive
JOCNote
the observed product 28. If this is the case, elimination of
acetate must give largely (>61%) the Z isomer, and it is not
clear why this should be so. When compound 1c, which we
used as a model, was treated under conditions identical with
those for 26, only acetate hydrolysis to 1b was observed (96%
yield) and no elimination product was isolated, suggesting
that formation of 28 does indeed occur via a true ICD path-
way involving 27.
β-Acetoxy-1-[(1,1-dimethylethoxy)carbonyl]hexahydro-r-met-
hylene-1H-azepine-2-propanoic Acid Methyl Ester (3c). AcCl
(0.29 mL, 3.27 mmol) was added to a stirred mixture of 3a
(341 mg, 1.09 mmol) and pyridine (0.47 mL, 5.45 mmol) in
CH2Cl2 (5 mL). Stirring was continued for 3.5 h, and the mixture
was diluted with water (15 mL) and extracted with CH2Cl2 (3 ꢀ
15 mL). The combined organic extracts were dried (MgSO4) and
evaporated. Flash chromatography of the residue over silica gel
(4 ꢀ 15 cm), using 4:1 hexane-EtOAc, gave 3c (355 mg, 87%) as
an oil: FTIR (CDCl3, microscope) 2929, 2855, 1751, 1725, 1692,
The reactions reported here extend the ICD1,2 process to
the generation of unusual seven-membered heterocycles. In a
few cases a five-membered heterocycle is formed and we note
that a number of potentially significant biochemical proper-
ties have been reported for related 2-thiazolidinethiones;11
the properties of the new seven-membered system are
unknown.
1440, 1366, 1232, 1163 cm-1 1H NMR (400 MHz, CDCl3,
;
60 °C) δ 1.16-2.16 (m, 20 H), 2.81-2.90 (m, 1 H), 3.56-3.70
(m, 1 H), 3.78-3.81 (m, 3 H), 4.29-4.51 (m, 1 H), 5.67-5.81 (m,
2 H), 6.28-6.32 (m, 1 H); 13C NMR (125 MHz, CDCl3, 60 °C)
δ 20.9, 21.0, 24.8, 25.0, 28.5, 28.6, 28.8, 29.3, 29.7, 29.8, 30.1,
30.3, 43.4, 43.8, 51.9, 52.0, 56.8, 57.0, 51.6, 73.4, 74.0, 79.2, 79.3,
79.8, 80.0, 125.7, 126.0, 126.6, 138.6, 138.8, 155.4, 155.8, 156.2,
165.4, 165.6; exact mass m/z calcd for C18H29NNaO6 (M þ Na)
378.1887, found 378.1890.
Experimental Section
5a,6,7,8,9,10-Hexahydro-1-thioxo-3H-azapino[1,2-c]-[1,3]thi-
azepine-4-carboxylic Acid Methyl Ester (3d). CF3CO2H (0.18
mL, 2.36 mmol) was added to a stirred solution of 3c (79 mg,
0.236 mmol) in CH2Cl2 (5 mL). Stirring was continued for 3.5 h,
and the mixture was evaporated. The residue was dissolved in
MeCN (5 mL), and CS2 (72 μL, 0.944 mmol) and i-Pr2NEt (0.13
mL, 0.980 mmol) were added. Stirring was continued overnight,
and the mixture was evaporated. Flash chromatography of the
residue over silica gel (2 ꢀ 15 cm), using 2:1 hexane-EtOAc,
gave 3d (37 mg, 58%) as an oil: FTIR (CDCl3, microscope)
2928, 2854, 1716, 1466, 1437, 1408, 1233, 1173 cm-1; 1H NMR
(500 MHz, CDCl3) δ 1.39-2.18 (m, 8 H), 3.21 (dd, J = 12, 13.5 Hz,
1 H), 3.62 (d, J = 14 Hz, 1 H), 3.79 (s, 3 H), 4.15 (dd, J = 2, 14 Hz,
1 H), 4.89 (ddd, J = 5, 9, 12 Hz, 1 H), 4.97 (dd, J = 5, 14 Hz, 1 H),
6.91 (dd, J = 2, 9 Hz, 1 H); 13C NMR (125 MHz, CDCl3) δ 26.3
(t), 28.6 (t), 28.7 (t), 32.7 (t), 34.8 (t), 50.7 (t), 52.5 (d), 61.3 (q),
129.2 (s), 141.4 (d), 165.2 (s), 195.3 (s); exact mass m/z calcd for
C12H17NNaO2S2 (M þ Na) 294.0593, found 294.0593.
1-[(1,1-Dimethylethoxy)carbonyl]hexahydro-β-hydroxy-r-met-
hylene-1H-azepine-2-propanoic Acid Methyl Ester (3b). DABCO
(591 mg, 5.28 mmol) was added to a stirred mixture of 3a12 (400mg,
1.76 mmol) and methyl acrylate (1.61 mL, 17.8 mmol), and stirring
was continued for 3 days. The mixture was diluted with water
(15 mL) and extracted with CH2Cl2 (3 ꢀ 15 mL). The combined
organic extracts were dried (MgSO4) and evaporated. Flash chro-
matography of the residue over silica gel (3 ꢀ 15 cm), using 2:1
hexane-EtOAc, gave 3b (440 mg, 80%) as an oil: FTIR (CDCl3,
microscope) 3434, 2928, 2854, 1722, 1689, 1676, 1478, 1441, 1414,
1391, 1165 cm-1; 1H NMR (400 MHz, CDCl3, 60 °C) δ 1.26-2.16
(m, 17 H), 2.88-3.03 (m, 1 H), 3.48-3.77 (m, 1 H), 3.79 (s, 3 H),
4.13-4.40 (m, 2 H), 5.58-5.89 (m, 1 H), 6.22-6.27 (m, 1 H); 13
C
NMR (100 MHz, CDCl3, 60 °C) δ 25.4, 28.5, 29.0, 29.5, 30.3, 43.2,
44.5, 44.7, 51.7, 59.3, 60.6, 60.8, 74.5, 75.7, 79.6, 79.9, 125.8, 126.3,
140.3, 141.9, 155.5, 157.6, 159.8, 166.9, 167.3; exact mass m/z calcd
1
for C16H27NNaO5 (M þ Na) 336.1781, found 336.1780. The H
NMR spectrum consisted of broad signals, and we attribute the
extra signals in the 13C NMR spectrum to the presence of rotamers
and diastereoisomers.
Acknowledgment. We thank the NSERC for financial
support.
(11) E.g.: (a) Alhamadsheh, M. M.; Waters, N. C.; Huddler, D. P.;
Kreishman-Deitrick, M.; Florova, G.; Reynolds, K. A. Bioorg. Med. Chem.
Lett. 2007, 17, 879–883. (b) Yoneda, K.; Ota, A.; Kawashima, Y. Chem.
Pharm. Bull. 1993, 41, 876–881. (c) Znamenskii, V. V.; Grechka, I. I.;
Ignatova, Yu. L.; Karimova, N. M. Pharm. Chem. J. 1996, 30, 617–619.
(12) Beak, P.; Lee, W. K. J. Org. Chem. 1993, 58, 1109–1117.
Supporting Information Available: Text and figures giving
detailed experimental procedures and characterization data of
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 75, No. 20, 2010 7017