Reactions of transition metal carbonyl anions with 2-(1-bromoalkylidene) thiazolidin-4-ones: Halogenophilic attack or deprotonation

Article abstract of DOI:10.1016/j.tetlet.2010.10.112

Bromophilic attack by the transition metal carbonyl anion, [Re(CO) ₅]Na (pK_a = 21.1), on 2-(1-bromoalkylidene)thiazolidin-4- ones is significantly faster than abstraction of an acidic lactam hydrogen (pK_a ～17-18), when the

Full text of DOI:10.1016/j.tetlet.2010.10.112

Tetrahedron Letters 52 (2011) 29–33
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Reactions of transition metal carbonyl anions with
2-(1-bromoalkylidene)thiazolidin-4-ones: halogenophilic attack
or deprotonation
P. K. Sazonov a, , Z. Dzambaski , M. M. Shtern , R. Markovic , I. P. Beletskaya
⇑
b
a
c
a
ˇ
´
a M.V. Lomonosov Moscow State University, Chemistry Department, Moscow 119992, Leninskie Gory 1, Russia
b Center for Chemistry ICTM, 11000 Belgrade, Serbia
c Faculty of Chemistry, University of Belgrade, 11000 Belgrade, Serbia
a r t i c l e i n f o
a b s t r a c t
Article history:
Bromophilic attack by the transition metal carbonyl anion, [Re(CO)5]Na (pKa = 21.1), on 2-(1-bromoalky-
lidene)thiazolidin-4-ones is significantly faster than abstraction of an acidic lactam hydrogen (pKa ꢀ17–
Received 24 June 2010
Revised 4 October 2010
Accepted 22 October 2010
Available online 29 October 2010
18), when the generated carbanion is stabilized by an a-CN or a-PhCO group. The bromophilic reaction of
2-(1-bromoalkylidene)thiazolidin-4-one, having an
tion of a new metallacyclic anionic complex. With less reactive vinyl bromides, containing an
or -CO2Et group, only deprotonation is observed. The role of the metal carbonyl anion is highlighted by a
a
-CN electron-withdrawing group, resulted in forma-
a
-CONHPh
a
Keywords:
Halogenophilic reactions
Proton transfer
2-(1-Bromoalkylidene)thiazolidin-4-ones
Metal carbonyl anions
comparison with the 9-methylfluorenide carbanion (pKa of 9-methylfluorene is 22.3), which reacts exclu-
sively via a deprotonation pathway.
Ó 2010 Elsevier Ltd. All rights reserved.
Halogen–metal exchange between RLi and RHal, frequently a
very fast reaction even at very low temperatures, can be carried
out without affecting functional groups, which under ordinary con-
ditions react with RLi. There are numerous examples in which hal-
Recently, we reported that 2-alkylidene-4-oxothiazolidine vinyl
bromides 3a–d react with a variety of ionic and neutral nucleo-
philes, leading to (i) the corresponding parent 4-oxothiazolidine
4 as a result of reductive debromination, (ii) bromine substitution,
and/or (iii) efficient C(5) functionalization of the thiazolidine ring,
all being initiated by bromophilic attack.8 Also possessing an acidic
site, that is, the lactam hydrogen, they provide a good model for
studying the competition between deprotonation and bromine–
metal exchange.
ogen–metal exchange in
a polyfunctional organic molecule,
containing an acidic hydrogen (OH, COOH, CONH2) and bromine
or iodine, is seemingly faster than deprotonation.1–4 On the con-
trary, reactions of o-, m-, and p-bromo and iodo isomers 1, nondeu-
terated 2-iodoquinoline (2) and its O-deuterated counterpart with
n-BuLi, previously examined by Beak et al.,5–7 revealed that proton
transfer is the initial process, whereas halogen–metal exchange oc-
curs either in the initially formed deprotonated complex, or in a
transient local area of high RLi concentration. Thus, the question
whether halogen–metal exchange can be competitive with depro-
tonation is intriguing, especially because proton transfer is one of
the fastest heterolytic reactions.
S
S
α
Z
Z
O
O
N
N
H
H
Br
H
3a d
4a d
-
-
a: Z = CN; b: Z = CONHPh; c: Z = CO2Et; d: Z = COPh
Additionally, it has been previously shown that metal carbonyl
anions are excellent halogenophilic agents in reactions with se-
lected polyfluorinated9,10 and non-fluorinated vinyl and aryl ha-
lides.11 Motivated to further explore the elementary processes
associated with vinyl bromides 3a–d and nucleophiles, we now re-
port their reactions with [Re(CO)5]Na, which, in the cases of com-
pounds 3a and 3d, provide unequivocal examples of the
bromophilic step being faster than proton transfer.
O
OH
NDEt
N
I
X
2
1 (X = o-, p-, m-Br or I)
All reactions of vinyl bromides 3a–d, differing in the electron-
withdrawing group Z at the exocyclic C@C bond, with [Re(CO)5]Na,
⇑
Corresponding author. Tel.: +7 (495) 9395258.
E-mail addresses: petr@elorg.chem.msu.ru, pksspk@yandex.ru (P.K. Sazonov).
0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.10.112

30
P. K. Sazonov et al. / Tetrahedron Letters 52 (2011) 29–33
carried out in THF at ꢁ50 °C, were almost instantaneous. The reac-
tion mixtures were subsequently analysed by NMR and IR methods
at room temperature. In studying the fast reactions where two
competitive pathways are possible, the order of reagent addition
can be important (vide infra).
Thus, we firstly performed the reactions by adding a slight ex-
cess of the corresponding vinyl halide 3 to a solution of [Re(CO)5]-
Na. Initial evidence of the halogenophilic mechanism emerged
from the reaction of [Re(CO)5]Na with 3a (Table 1, entry 1), giving
rise to a product, having IR and 13C NMR spectral data characteris-
tic of an anionic oxycarbene Re(CO)4-complex 5a (Scheme 1). The
patterns are quite similar to those observed for halo(acyl)rhenates,
[Hal(CO)4Re(CO)R]Na, formed in halogenophilic reactions of [Re(-
CO)5]Na with aryl or vinyl halides,9–11 and yet have slight but sig-
nificant differences, such as an upfield chemical shift of the
oxycarbene carbon at 248.4 ppm, compared to 256–260 ppm typi-
cal for halo(acyl)rhenates. A significant downfield shift (ca. 25–
40 ppm) of both vinylic carbon signals, compared to the typical
values for the parent thiazolidine precursors 3,12 suggests a quasia-
romatic metallacycle structure 5a for this complex (Scheme 1). The
structure was confirmed by comparison of the experimental and
calculated (DFT method) chemical shifts and IR frequencies for
5a. The MALDI and high resolution ESI MS of 5a also gave a molec-
ular mass consistent with the proposed structure.13 The formation
of 5a implies that [Re(CO)5]Na acts both as a halogenophile
abstracting bromine, and as a base, abstracting the acidic NH-pro-
ton (Scheme 1). This conclusion is supported by the observation
that Re(CO)5H is present among the reaction products (Table 1, en-
try 1). Presumably, and in order to explain the formation of com-
plex 5a together with Re(CO)5H, the proton and halogen
abstractions have to be fast compared to the duration of reagent
mixing.
However, at this point, we had only limited information on the
reaction kinetics concerning the rates of the halogenophilic and
Table 1
Reactions of [Re(CO)5]Na with vinyl bromide 3a, THF, ꢁ50 °C
Entry
Reaction conditions
Product yield (%)
5a
4a
Re(CO)5Br
Re(CO)5H
1
2
3
4
5
Direct additiona
50
90
37
60
32d
25
3
46
30
52
ꢀ25b
ꢀ10b
ꢀ50b
-
20
—
5
<5
7
1 equiv NaH, direct additiona
Reverse additionc
Reverse additionc in the presence of Re(CO)5Br (2 equiv)
Reverse addition,c then 4 equiv Me3SiCl
ꢀ50b
a
b
c
Addition of vinyl bromide 3a to a solution of [Re(CO)5]Na.
Yield estimated from the 13C NMR and IR spectra.
Addition of [Re(CO)5]Na solution to a solution of vinyl bromide.
Complex 5a is present, apparently, in silylated form.
d
N
N
S
S
N
S
S
N
S
O
O
Na
O
CN
[Re(CO)5]Na
-HRe(CO)5
N
N
O
[Re(CO)5]Na
THF, -50 °C
CO
Re
N
CO
CO
Re
O
Na
-NaBr
O
Na
N
N
O
O
H
Re
H
Br
CO
(CO)4ReBr
OC
OC
Na
OC
Na
Br
OC
CO
5a
3a
OC
CO
CO
OC
Scheme 1.
O
S
OC2H5
O
+ BrRe(CO)5
N
H
Na
4c anion
18-crown-6
r.t., ~10 min
Z = CO2Et
O
S
S
S
Me3SiCl
-50 °C
Z
Z
+
[Re(CO)5]Na THF, -50 °C
NHPh
Re(CO) H
O
+
O
5
Me3SiO
N
N
N
-NaCl
H
Br
Br
Br
Z = CONHPh
3b (Z = CONHPh)
3c (Z = CO2Et)
6b
Na
3b or 3c anion
r.t., ~1.5 h
Z = CONHPh
O
S
NHPh
O
N
+ BrRe(CO)5
H
Na
4b anion
Scheme 2.

P. K. Sazonov et al. / Tetrahedron Letters 52 (2011) 29–33
31
Table 2
Reactions of [Re(CO)5]Na with vinyl bromide 3b (Z = CONHPh) and 3d (Z = COPh), followed by an excess of Me3SiCl in THF at ꢁ50 °C
Entry
Vinyl bromide
Reaction conditions
Product yield (%)
Re(CO)5Br
Re(CO)5H
4
6
7
8
1
2
3
4
5
3b
3d
3d
3d
3b
Direct addition
Direct addition
Direct addition without Me3SiCl
Reverse addition
Reverse addition
ꢀ8a
90
45
—
<1
90
<1
7
75
—
—
—
62
11
20
—
24
10
—
10
—
7
ꢀ50a
>90
90b
42
—
>90a
<10a
—
a
Yield estimated from 13C NMR and IR spectra.
Mainly in NH deprotonated form.
b
protophilic steps. In order to determine which step occurs first, we
next reversed the order of reagent addition. Thus, upon dropwise
addition of a solution of [Re(CO)5]Na to a solution of substrate
3a, only 5% of Re(CO)5H was obtained, whereby Re(CO)5Br and
complex 5a, along with the debrominated 4-oxothiazolidine 4a,
were the main products (Table 1, entry 3). The yield of complex
5a (Scheme 1) in the reverse addition experiment was only 37%,
but in the presence of 2 mol equiv of Re(CO)5Br it increased to
60% (Table 1, entry 4). The final result illustrates the intermediacy
of Re(CO)5Br as a reactant en route to metallacycle 5a. It is worth
mentioning that deprotonation of 3a with NaH to give the lactam
anion, followed by reaction with [Re(CO)5]Na, led to complex 5a
in excellent yield (Table 1, entry 2).13 Thus, in the reaction of vinyl
bromide 3a with [Re(CO)5]Na the halogenophilic step is apparently
faster than deprotonation.
Unlike the vinyl bromide 3a which was transformed into the
metallacycle 5a, reactions of substrates 3b (Z = CONHPh) and 3c
(Z = CO2Et) with [Re(CO)5], resulted in proton abstraction, furnish-
ing Re(CO)5H (pKa = 21.1)14,15 as the main product (Scheme 2) The
reaction mixtures were relatively unstable at room temperature,
but an important finding from the reaction of 3c, after addition
of 18-crown-6, was the complete disappearance of Re(CO)5H with
concomitant formation of 4-oxothiazolidine 4c (in NH-deproto-
nated form) and Re(CO)5Br. In the reaction with 3b the debromi-
temperature. These results can be interpreted by assuming that a
certain amount of [Re(CO)5]Na, present in an acid–base equilib-
rium, slowly debrominates compounds 3b or 3c to the parent so-
dium salt 4b or 4c, respectively.
Importantly, this has been confirmed by the IR detection of a
small amount of [Re(CO)5]Na remaining in the reaction with a
slight excess of 4d. The equilibrium constant, K = 94 (Eq. 1), was
calculated from the IR data. As expected, according to the ACD/Labs
program,16 vinyl bromides 3a–d are more acidic by ꢀ2 pKa units
than the parent thiazolidines 4a–d. In addition, based on the equi-
librium involving neutral and anionic species 3c and 4c (Eq. 2),
investigated by the IR method, a smaller, but still significant differ-
ence of P1 pKa units between the pKas of 3c and 4c, was deter-
mined. At this point it became obvious that these experiments
showed that it was necessary to neutralize the base (the amide an-
ion) in order to determine the true ratio of Re(CO)5Br and
Re(CO)5H.
O
O
S
S
K = 94
Ph
Ph
[Re(CO)5]Na
+
Re(CO) H
+
5
O
O
N
N
H
H
H
Na
4d anion
4d
ð1Þ
O
O
O
O
S
S
S
N
S
K = 12
OC2H5
OC2H5
OC2H5
OC2H5
O
O
O
+
O
ð2Þ
N
N
+
N
H
H
Br
H
Br
H
Na
Na
4c anion
3c
4c
3c anion
nated product 4b was already present (in 52% yield) after 10
minutes, as indicated by the 1H NMR spectrum. The amount of
4b increased with time and reached 82% after 1.5 h at room
Thus, when the vinyl bromide 3b (Z = CONHPh) was treated
with [Re(CO)5]Na, followed by quenching with Me3SiCl immedi-
ately after the reagent mixing at ꢁ50 °C (Scheme 2), Re(CO)5H
O
O
O
Me SiCl
3
THF, -50 °C
S
S
[Re(CO)5]Na
THF, -50 °C
S
Ph
Ph
Na
Ph
Re(CO) Br +
O
O
O
5
N
N
N
H
H
H
Br
Na
3d
O
OSiMe
Ph
O
3
S
S
S
Ph
Ph
+
+
Me SiO
3
O
Me SiO
3
N
N
N
H
H
H
H
4d
7d
8d
Scheme 3.

32
P. K. Sazonov et al. / Tetrahedron Letters 52 (2011) 29–33
S
S
CN
CN
THF, -50 °C
+
O
+
N
O
N
H
Br
Br
Li
Li
H
3a
Me3SiCl
S
S
CN
Br
CN
Br
Me3SiO
O
H2O
N
N
H
Scheme 4.
was obtained almost quantitatively, together with the correspond-
ing silyl ether 6b (Table 2, entry 1). On the other hand, the same
quenching procedure applied to the reaction of [Re(CO)5]Na with
bromide 3d (Z = COPh) gave a different result: both Re(CO)5Br
and Re(CO)5H were observed in a ꢀ1:1 ratio (Table 2, entry 2). A
complex mixture of the thiazolidine-derived products was also
formed, in which mono- and bis-silylated derivatives 7d and 8d
of debrominated 4-oxothiazolidine 4d were identified (Scheme
3). The mono-silylated product 7d was almost completely trans-
formed on standing for one day at room temperature into the
bis-silylated product 8d.
In contrast, on repeated direct reaction of 3d with [Re(CO)5]Na
in the absence of Me3SiCl (Table 2, entry 3), the debrominated vinyl
compound 4d (in NH-deprotonated form) and Re(CO)5Br were
formed as the only products. This result strongly indicates that
the reaction of 3d with Re(CO)5Na is too fast for Re(CO)5H to be ob-
served without quenching the lactam anion (in the presence of
Me3SiCl) from the equilibrium.
Formation of Re(CO)5H and Re(CO)5Br in ꢂ1:1 ratio (entry 2, Ta-
ble 2) suggests that the ratio is not kinetically controlled, but both,
the halogenophilic and protophilic reactions with 3d are fast com-
pared to the duration of reagent mixing. The reverse order of re-
agent addition can indicate which of the reactions is faster and
takes place first. When a solution of [Re(CO)5]Na was added drop-
wise to a solution of 3d, followed by the addition of Me3SiCl, only
Re(CO)5Br was obtained (Table 2, entry 4 and Scheme 3) and min-
ute amounts of Re(CO)5H (<1%). The reaction with 3a performed in
the same manner resulted in only 7% of Re(CO)5H, giving Re(CO)5Br
and complex 5a as the main products (Table 1 entry 5). On the
other hand, reversing the order of reagent addition had no effect
on the reaction of 3b, and Re(CO)5H was formed almost quantita-
tively (Table 2, entry 5), just as in the case of direct addition (Table
2, entry 1).
halogen in vinyl bromides 3a and 3d is possible because of the
exceedingly high rate of the halogenophilic reaction with [Re(-
CO)5]Na. It is worth mentioning that the role of the metal carbonyl
anion has been further highlighted by comparison with a carban-
ion of comparable basicity. Thus, the reaction of 3a with 9-methyl-
fluorenide lithium (pKa of 9-methylfluorene is 22.3) resulted only
in deprotonation of the NH group and quantitative recovery of
the starting bromide 3a after silylation and hydrolysis (Scheme 4).
Carbanions are usually considered as good halogenophi-les,17,18
but the results of the current study lead us to the conclusion that
metal carbonyl anions show even higher tendency towards attack
on halogen with respect to carbanions.
Acknowledgements
The authors gratefully acknowledge financial support of this
work from the Program of the President of the Russian Federation
for the State Support of Leading Research Schools (Grant No. HI-
4365.2010.3) and the RAS Program 1-OX. The authors thank Dr. N.
G. Kolotyrkina (Department of Structural Studies of Zelinsky Insti-
tute of Organic Chemistry, Moscow) for recording HR ESI MS spec-
tra. This research was also partially supported by the Ministry of
Science of the Republic of Serbia, Grant No. 142007 (to R.M.).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tetlet.2010.10.112.
References and notes
1. Narasimhan, N. S.; Sunder, N. M.; Ammanamanchi, R.; Bonde, B. D. J. Am. Chem.
Soc. 1990, 112, 4431.
2. Sydnes, L. K.; Skare, S. Can. J. Chem. 1984, 62, 2073.
3. Bailey, W. F.; Patricia, J. J.; Nurmi, T. T.; Wang, W. Tetrahedron Lett. 1986, 27,
1861.
4. Stein, C. A.; Morton, T. H. Tetrahedron Lett. 1973, 14, 4933.
5. Beak, P.; Musick, T. J.; Chen, C. W. J. Am. Chem. Soc. 1988, 110, 3538.
6. Beak, P.; Chen, C.-W. Tetrahedron Lett. 1985, 26, 4979.
7. Gallagher, D. J.; Beak, P. J. Am. Chem. Soc. 1991, 113, 7984.
8. Baranac-Stojanovic´, M.; Tatar, J.; Stojanovic´, M.; Markovic´, R. Tetrahedron 2010,
66, 6873.
Thus, the reverse addition experiments have finally confirmed
that vinyl bromides 3a and 3d are first attacked on the halogen
by [Re(CO)5]Na. Deprotonation is, on the other hand, the fastest
process for compounds 3b and 3c. Such dichotomy can be ex-
plained by the much higher reactivity of 3a and 3d in halogeno-
philic reactions compared to 3b and 3c. The
a-substituents in 3a
and 3d, CN and PhCO, respectively, being stronger electron-accep-
tors than CONHPh (for 3b) and CO2Et (for 3c), can better stabilize
the incipient vinyl carbanion. This is also in agreement with the
qualitative reactivity order of vinyl bromides, 3a, 3d ꢃ 3b > 3c, to-
wards [CpMo(CO)3]K, when CpMo(CO)3Br was the only organome-
tallic product.
9. Ivushkin, V. A.; Sazonov, P. K.; Artamkina, G. A.; Beletskaya, I. P. J. Organomet.
Chem. 2000, 597, 77.
10. Sazonov, P. K.; Artamkina, G. A.; Khrustalev, V. N.; Antipin, M. Y.; Beletskaya, I.
P. J. Organomet. Chem. 2003, 681, 59.
11. Sazonov, P. K.; Ivushkin, V. A.; Artamkina, G. A.; Beletskaya, I. P. ARKIVOC 2003,
x, 323.
ˇ
´
12. Markovic´, R.; Baranac, M.; Dzambaski, Z.; Stojanovic, M.; Steel, P. J. Tetrahedron
2003, 59, 7803.
How fast are the halogenophilic reactions of 3a and 3d with
[Re(CO)5]Na? The available data on the kinetic acidity of metal car-
bonyl hydrides allows an estimation to be made.14,15 Vinyl bro-
mides are by 3–4 pKa units more acidic than Re(CO)5H (see Eqs.
1 and 2). An expected rate constant for proton transfer from an
NH-acid to a metal carbonyl anion is at least ꢀ106 l/mol s at such
pKa difference. Since only the halogenophilic reaction is observed
in the case of 3d its rate should be not less than 108 l/mol s, which
is quite near to the diffusion limit. The preferred attack on the
13. Metallacycle 5a. Vinyl bromide 3a (19.3 mg, 0.0882 mmol) was treated with
NaH (50% suspension in mineral oil, 4.4 mg, 0.0092 mmol) in ꢀ1 ml THF. After
the evolution of gas had ceased the solution was cooled to ꢁ50 °C and
[Re(CO)5]Na (0.0663 mmol in 0.24 ml THF) was introduced. The reaction
mixture was allowed to reach rt, and after 6 h was analysed by 1H and 13C NMR
spectroscopy, which showed almost quantitative formation of complex 5a. 1H
NMR (THF, 400.13 MHz): d = 3.97 s (2H, CH2). 13C NMR (THF, 100.61 MHz):
d = 248.31 (Re@C–O), 192.62 (CO, 2C), 190.63 (CO, 2C), 188.07 (C@), 181.36
(C@Olactam), 117.45 (CN), 108.03 (@CCN), 37.64 (CH2). MALDI MS: m/z 408.9
[Mꢁ2CO]ꢁ. HR ESI MS: m/z 408.9297 [Mꢁ2CO]ꢁ. Calcd for C8H2N2O4ReS
408.9293. IR (THF):
m = 2187 s (CN), 2079 s (CO), 1966 vs (CO), 1930 s (CO),

P. K. Sazonov et al. / Tetrahedron Letters 52 (2011) 29–33
33
1695 m, 1510 s. Spectra of 5a in the presence of 18-crown-6: 1H NMR (THF):
d = 3.88 s (CH2). 13C NMR (THF): d = 240.48 (Re@C–O), 193.44 (CO, 1C), 192.79
(CO, 1C), 191.94 (CO, 2C), 116.83 (CN), 37.40 (CH2), the signals of the
quaternary carbons of the thiazolidinone backbone were not observed due to
diameter) was placed inside a standard 5 mm NMR tube containing acetone-d6
for the lock signal.
14. Edidin, R. T.; Sullivan, J. M.; Norton, J. R. J. Am. Chem. Soc. 1987, 109, 3945.
15. Jordan, R. F.; Norton, J. R. J. Am. Chem. Soc. 1982, 104, 1255.
16. ACD/Labs pKa DB, V. 6.0, Advanced Chemistry Development, 1994–2002.
17. Zefirov, N. S.; Makhon’kov, D. I. Chem. Rev. 1982, 82, 615.
18. Bunnett, J. F. Acc. Chem. Res. 1972, 5, 139.
the low signal-to-noise ratio. IR (THF):
m = 2193 s (CN), 2072 s (CO), 1975 sh
(CO), 1958 vs (CO), 1925 s (CO), 1685 m, 1510 s. To record the NMR spectra in
undeuterated THF the sample in a sealed thin-walled glass capillary (ꢀ3.5 mm