Thiazolidinones derived from dynamic systemic resolution of complex reversible-reaction networks

Article abstract of DOI:10.1002/chem.201304690

A complex dynamic system based on a network of multiple reversible reactions has been established. The network was applied to a dynamic systemic resolution protocol based on kinetically controlled lipase-catalyzed transformations. This resulted in the formation of cyclized products, where two thiazolidinone compounds were efficiently produced from a range of potential transformations.

Full text of DOI:10.1002/chem.201304690

DOI: 10.1002/chem.201304690
Communication
&
Cyclization |Very Important Paper|
Thiazolidinones Derived from Dynamic Systemic Resolution of
Complex Reversible-Reaction Networks
Yan Zhang and Olof Ramstrçm*[a]
librium states. However, the systems can also be coupled to ki-
netic processes, thereby selecting species in relation to the re-
Abstract: A complex dynamic system based on a network
action transition states. This led to the establishment of the
of multiple reversible reactions has been established. The
concept of dynamic systemic resolution (DSR), in which selec-
network was applied to a dynamic systemic resolution
tive, kinetically controlled routes are coupled to the dynamic
protocol based on kinetically controlled lipase-catalyzed
systems.[21] Especially enzyme-catalyzed reaction pathways
transformations. This resulted in the formation of cyclized
present an efficient methodology,[22–28] applying both catalysis
products, where two thiazolidinone compounds were effi-
and selectivity on the systems. With this strategy, not only the
substrate selectivities of different enzymes could be uncov-
ered, but also the enzyme promiscuities could be challenged,
providing useful information for enzymatic synthesis in
general.
ciently produced from a range of potential transforma-
tions.
Biological systems are generally based on complex networks of
chemical reactions, operating under thermodynamic and kinet-
ic control in response to various regulating and evoking fac-
tors. The delineation of such networks is of high importance,
and constitutes an integral part of the fields of systems biology
and chemistry.[1–3] Through the study of the properties and dy-
namics of such highly complex reactive systems, a better un-
derstanding of biological processes can be obtained.
Recent progress in dynamic chemistry has demonstrated an
increasing number of useful reversible reaction types. This de-
velopment is certainly important in itself, and has also led to
an extended application scope. However, the diversities of the
designed dynamic systems have still been fairly limited. There-
fore, efforts to build dynamic covalent systems with more than
one reversible reaction have been explored.[26,28–32]
Although reversible cascade reactions provide easy access to
multifunctional structures, parallel reactions generate struc-
tures of higher diversity. Thus, by combining these two strat-
egies, an efficient approach to dynamic systems with greatly
enhanced complexities can be obtained. Herein, we report
a complex dynamic systemic resolution protocol based on net-
works of multiple reversible reactions, resulting in a range of
possible transformations (Figure 1).
Constitutional dynamic chemistry (CDC) represents in this
context a powerful approach, enabling the generation and
study of molecular systems. During the last decade, CDC has
been adopted in a range of different areas, such as identifica-
tion of protein ligands, selection of artificial receptors, and de-
velopment of functional materials.[4–20] The key feature of CDC
is its dynamic nature, based on reversible covalent bonds or
noncovalent interactions, which enables the involved constitu-
ents to mutually interchange under thermodynamic control.
The resulting dynamic systems are inherently responsive to in-
ternal or external pressures, giving rearrangement of the con-
stituent composition and further amplification of optimally re-
acting species.
Systems generated using CDC operate under thermodynam-
ic control, generally resulting in the energetically favored equi-
Figure 1. Dynamic systemic resolution of complex reversible-reaction net-
[a] Dr. Y. Zhang, Prof. Dr. O. Ramstrçm
KTH Royal Institute of Technology
Department of Chemistry, Teknikringen 30
10044 Stockholm (Sweden)
work.
Of the relevant reversible processes available for the reaction
network, two were initially chosen: the nitroaldol and hemi-
thioacetal reactions. Although these are able to generate two
major types of intermediates for the simultaneous enzymatic
transformations,[28] additional diversity was established from
the incorporation of imines into the existing system. In princi-
ple, this would result in eight main types of reversible reac-
E-mail: ramstrom@kth.se
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304690.
ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons At-
tribution Non-Commercial NoDerivs License, which permits use and distri-
bution in any medium, provided the original work is properly cited, the use
is non-commercial and no modifications or adaptations are made.
Chem. Eur. J. 2014, 20, 3288 – 3291
3288
ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
tions operating concertedly: hemiaminal, hemithioaminal,
hemithioacetal, aminal, and hydrate formation; and nitroaldol-
and aza-nitroaldol reactions (Scheme 1). Due to the dynamic
nature of these reactions, all individual components and con-
stituents would mutually undergo interchange under the same
Scheme 1. Dynamic covalent reaction network primarily composed of imine
formation/hydrolysis, transimination, hemithioacetal/hemithioaminal forma-
tion, and nitroaldol reactions.
conditions. This enhanced reaction complexity would thus
result in a large amount of substrates of increased variety, sub-
sequently available for enzyme selection and transformation.
Starting from n imines, m amines, H2O, p thiols and q nitroal-
kanes, a reaction network composed of (m2n+2mn2 +2mnp+
2mnq+5mn+2m+n3 +2n2p+2n2q+5n2 +2p+2nq+6n+
2p+2q+2)/2 entities could in principle be formed, not count-
ing chirality.
Scheme 2. Dynamic model-reaction network.
patibilities when conducted in the same system. Thus, the si-
multaneous reversible reactions between aldehyde 1, thiol A,
and nitropropane B could be performed without complications
(Scheme 2). Methyl 2-sulfanylacetate A was chosen in this case
instead of other possible thiols, because the generated inter-
mediates 1A, 2A, and 3 A provided additional diversity for the
enzyme-catalyzed transformation step: both intermolecular,
linear acylations, and intramolecular cyclizations. Replacing the
aldehydes by imines in the simultaneous addition reactions
proved to be equally efficient. Although for the imine building
blocks 2 and 3, the thiol addition product was relatively unfa-
vored with an equilibrium distribution to the starting material
side, the exchange could be recorded by using higher concen-
trations (20 equiv) of two different thiols. Thus, after addition
of the second thiol, immediate redistribution of the constitu-
ents took place, confirming fast reversibility rates at room tem-
perature. However, formation of compounds 2B and 3B was
very unfavored, and could not be observed even upon addi-
tion of higher amounts of base, and only in large amounts of
nitroalkane. Thus, other bases besides Et3N were also evaluat-
ed, including 4-dimethylaminopryidine (DMAP), and 1,1,3,3-tet-
ramethylguanidine (TMG). Of these, DMAP proved similar to
Et3N, whereas the stronger base TMG gave small conversions
to the adducts. In the latter case, however, potential degrada-
tion of the compounds was observed.
The simultaneous reversibility of the individual reactions was
next addressed, initially investigating the transimination and
imine-formation reactions by using a model system com-
posed of 3-nitrobenzaldehyde (1), N-(3-nitrobenzylidene)-
methanamine (2), and N-(3-nitrobenzylidene)propan-2-amine
(3; Scheme 2).
The transimination between compound 2 and isopropyl-
amine was very fast at room temperature, and equilibrium was
reached in one hour. The formation of imine 3 from aldehyde
1 and isopropylamine also went smoothly, with a high conver-
sion of 81% in one hour. On the other hand, the reverse reac-
tion from imine to aldehyde was slow, and therefore, different
additives were tested to accelerate the process. As was expect-
ed, addition of base (Et3N) did not result in faster conversion.
Instead, the Lewis acids Sc(OTf)3 and ZnBr2, which have been
previously proven to be efficient for transimination,[26,33] were
tested also in this case. Of these, ZnBr2 showed better effects
on the reaction rate, and increasing amounts of ZnBr2 resulted
in faster formation of aldehyde. However, compared to the re-
action rate of the imine formation, the reverse reaction was
still considerably slower, resulting in a biased distribution be-
tween aldehyde and imine. Thus, H2O was added to the
system to further displace the equilibrium to the aldehyde
side. Due to sensitivity of lipases towards water in organic sol-
vents, only two equivalents of H2O were added, giving a ratio
of 5:1 between imine 2 and aldehyde 1 in 21 h.
Based on the collected properties of the individual reversible
reactions, a larger dynamic system was generated with two ar-
omatic aldehydes (4–5) and four related imines (6–9), leading
to a system size composed of 35 (53 counting chirality) poten-
tial species (Scheme 3). Due to the nature of the imines, this
represents a subset of the maximal number of potential spe-
cies. 2-Chloro- and 4-chloro-substituents on the aromatic
Similar to the previous report,[28] the nitroaldol and hemi-
thioacetal reactions showed good reversibility rates and com-
Chem. Eur. J. 2014, 20, 3288 – 3291
www.chemeurj.org
3289
ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Scheme 3. Lipase-catalyzed dynamic systemic resolution of complex reversible-reaction network. Shaded compounds were of virtual nature, not visible in the
spectra under the conditions (4A–5A, 6A–9A, 6B–9B). Only compounds 10 and 11 were formed, out of 18 possible kinetic transformations: intermolecular
acylation of intermediates 4A–9 A, 4B–9B; or intramolecular cyclization of compounds 4A–9A.
groups were chosen in this case because of their similar activa-
tion effects to the structures. To build the first dynamic system
of building blocks 4–9, one equivalent of each aldehyde 5 and
imine 6 were added into the system, together with one equiv-
alent of isopropylamine to provide imines 8 and 9. In addition,
two equivalents of H2O were included to enable formation of
aldehydes, as well as half an equivalent each of Et3N and
reversible reaction products of the complex dynamic system
showed a virtual dynamic character and were not visibly ex-
pressed.
The generated dynamic system was subsequently subjected
to further enzyme-catalyzed transformations. Similar to previ-
ous studies, lipases were adopted in this case in part due to
their many advantages.[34–37] Two different enzyme preparations
were thus applied to the complex dynamic system, based on
lipases from Burkholderia (formerly Pseudomonas) cepacia
(BCL), and Candida antarctica (CAL-B), together with phenyl
acetate as the acyl donor. A number of possible transforma-
tions could in principle be catalyzed by the enzymes in this
system, either intermolecular (linear) acylation of intermediates
4A–9A, or compounds 4B–9B; or intramolecular cyclization of
compounds 4A–9A (Scheme 3), and the lipase selectivities re-
garding both substrates and reaction types were next moni-
tored. The transformations were performed at room tempera-
1
ZnBr2. By using H NMR spectroscopy as a tool to monitor the
reaction process, it was again obvious that imine formation
and transimination were much faster than the aldehyde forma-
tion. However, the distribution of all formed constituents at
equilibrium was still suitable for the systemic resolution pro-
cess (Figure 2). Upon subsequent addition of 2-sulfanylacetate
1
ture in tert-butyl methyl ether and followed by H NMR spec-
troscopy. Conspicuously, only the cyclized products 10 and 11
were observed when applying CAL-B, in spite of the virtual,
“invisible”, nature of the parent intermediate structures 6A
and 7A under these conditions. Thus, of the 18 possible kinetic
transformations, only two products were formed. These results
also represent the first report on this lipase-catalyzed cycliza-
tion reaction, showing the potential for lipases to catalyze the
formation of thiazolidinone structures. In contrast to CAL-B,
BCL provided only marginal amounts of products under the
same conditions, likely due to the lower activity compared
with CAL-B for these specific substrates. The results can be in-
terpreted by the fact that acylation of hemithioacetal inter-
mediates generally occurs faster than acylation of nitroaldol
adducts, and that cyclization is preferred by CAL-B over forma-
tion of linear, noncyclized, products. In consequence, only cy-
clized products were detected from the complex dynamic
system. Between intermediates 4A and 5A and 6A–9A, the
latter are favored for cyclization due to higher nucleophilicity
of the nitrogen atom under these conditions. However, steric
factors are likely to also have played a role, because no cycliza-
Figure 2. 1H NMR spectra of complex dynamic multireversible system:
a) imine/aldehyde system at equilibrium; and b) overall system upon DSR.
A and nitropropane B, formation of the nitroalkanols 4B and
5B was observed in the NMR spectra, whereas the remaining
Chem. Eur. J. 2014, 20, 3288 – 3291
www.chemeurj.org
3290
ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[7] E. Moulin, G. Cormos, N. Giuseppone, Chem. Soc. Rev. 2012, 41, 1031–
1049.
[8] J.-M. Lehn, Angew. Chem. Int. Ed. 2013, 52, 2836–2850.
[9] Y. Jin, C. Yu, R. J. Denman, W. Zhang, Chem. Soc. Rev. 2013, 42, 6634–
6654.
[10] M.-K. Chung, K. Severin, S. J. Lee, M. L. Waters, M. R. Gagnꢂ, Chem. Sci.
2011, 2, 744–747.
[11] A. G. Santana, E. Jimꢂnez-Moreno, A. M. Gꢃmez, F. Corzana, C. Gonzꢁlez,
G. Jimꢂnez-Oses, J. Jimꢂnez-Barbero, J. L. Asensio, J. Am. Chem. Soc.
2013, 135, 3347–3350.
tion of compounds 8A–9A could be observed. This result is
analogous to previous experiences with lipase-catalyzed ami-
dation, in which larger alkyl groups impaired acylation.[26] The
resulting ratio between products 10 and 11 was 2:1, and when
considering the equilibrium distribution of intermediates 6 and
7, compound 10 proved to be the most amplified product
from the whole DSR process.
In conclusion, highly complex dynamic systems were gener-
ated from networks of different reversible reactions operating
simultaneously: primarily imine formation/hydrolysis, transimi-
nation, hemithioacetal/hemithioaminal formation, and nitroal-
dol reactions. By using lipase catalysis with CAL-B to kinetically
resolve the complex system, only two compounds were selec-
tively produced out of a large range of potential kinetic prod-
ucts of noncyclized and cyclized nature. In addition, one of the
compounds was formed in excess, in response to the enzyme
selectivity. This also represents the first report on the formation
of thiazolidinones using lipase catalysis. Thus, the compatibility
of several different reversible types of chemistry have been
demonstrated and successfully coupled to enzyme-catalyzed
resolution. The overall DSR process showed high chemoselec-
tivity and amplification capacity. These findings represent an
entry into the study and further understanding of complex re-
action networks composed of both reversible and irreversible
pathways, and can also provide information on biocatalytic ac-
tivities.
[12] P. Reeh, J. de Mendoza, Chem. Eur. J. 2013, 19, 5259–5262.
[13] M. Rauschenberg, S. Bomke, U. Karst, B. J. Ravoo, Angew. Chem. 2010,
122, 7498–7503; Angew. Chem. Int. Ed. 2010, 49, 7340–7345.
[14] R. Sundell, M. Turcu, L. Kanerva, Curr. Org. Chem. 2013, 17, 672–681.
ˇˇ ´
[15] A. M. Belenguer, T. Friscic, G. M. Day, J. K. M. Sanders, Chem. Sci. 2011, 2,
696.
[16] F. Dumitru, Y.-M. Legrand, E. Petit, A. van der Lee, M. Barboiu, Dalton
Trans. 2012, 41, 11860–11865.
[17] F. Mansfeld, H. Au-Yeung, J. Sanders, S. Otto, J. Syst. Chem. 2010, 1, 12.
[18] C. Arnal-Hꢂrault, M. Barboiu, A. Pasc, M. Michau, P. Perriat, A. van der
Lee, Chem. Eur. J. 2007, 13, 6792–6800.
[19] M. Barboiu, E. Petit, A. van der Lee, G. Vaughan, Inorg. Chem. 2006, 45,
484–486.
[20] V. Goral, M. I. Nelen, a. V. Eliseev, J. M. Lehn, Proc. Natl. Acad. Sci. USA
2001, 98, 1347–1352.
[21] M. Sakulsombat, Y. Zhang, O. Ramstrçm, Top. Curr. Chem. 2012, 322,
55–86.
[22] R. Larsson, Z. Pei, O. Ramstrçm, Angew. Chem. 2004, 116, 3802–3804;
Angew. Chem. Int. Ed. 2004, 43, 3716–3718.
[23] M. Sakulsombat, P. Vongvilai, O. Ramstrçm, Org. Biomol. Chem. 2011, 9,
1112–1117.
[24] M. Sakulsombat, Y. Zhang, O. Ramstrçm, Chem. Eur. J. 2012, 18, 6129–
6132.
[25] P. Vongvilai, M. Angelin, R. Larsson, O. Ramstrçm, Angew. Chem. 2007,
119, 966–968; Angew. Chem. Int. Ed. 2007, 46, 948–950.
[26] P. Vongvilai, O. Ramstrçm, J. Am. Chem. Soc. 2009, 131, 14419–14425.
[27] Y. Zhang, M. Angelin, R. Larsson, A. Albers, A. Simons, O. Ramstrçm,
Chem. Commun. 2010, 46, 8457–8459.
[28] Y. Zhang, L. Hu, O. Ramstrçm, Chem. Commun. 2013, 49, 1805–1807.
[29] A. V. Gromova, J. M. Ciszewski, B. L. Miller, Chem. Commun. 2012, 48,
2131–2133.
Acknowledgements
This work was supported in part by the Swedish Research
Council and the Royal Institute of Technology. Y.Z. thanks the
China Scholarship Council for a special scholarship award.
[30] J. Leclaire, L. Vial, S. Otto, J. K. M. Sanders, Chem. Commun. 2005, 1959–
1961.
[31] K. D. Okochi, Y. Jin, W. Zhang, Chem. Commun. 2013, 49, 4418–4420.
[32] Z. Rodriguez-Docampo, S. Otto, Chem. Commun. 2008, 0, 5301–5303.
[33] N. Giuseppone, J.-L. Schmitt, E. Schwartz, J.-M. Lehn, J. Am. Chem. Soc.
2005, 127, 5528–5539.
Keywords: cyclization
lipases · resolution
· dynamic chemistry · enzymes ·
[1] J. J. Tyson, B. Novꢁk, Annu. Rev. Phys. Chem. 2010, 61, 219–240.
[2] C. I. Sandefur, M. Mincheva, S. Schnell, Mol. Biosyst. 2013, 9, 2189–200.
[3] M. Feinberg, Chem. Eng. Sci. 1987, 42, 2229–2268.
[4] Constitutional Dynamic Chemistry (Ed.: M. Barboiu), Springer, Berlin, Hei-
delberg, 2012.
[34] M. Kapoor, M. N. Gupta, Process Biochem. 2012, 47, 555–569.
[35] A. Ghanem, Tetrahedron 2007, 63, 1721–1754.
[36] R. D. Schmid, R. Verger, Angew. Chem. 1998, 110, 1694–1720; Angew.
Chem. Int. Ed. 1998, 37, 1608–1633.
[5] Dynamic Combinatorial Chemistry: Drug Discovery, Bioorganic Chemistry,
and Materials Science (Ed.: B. L. Miller), John Wiley & Sons, Hoboken,
2010.
[37] F. Theil, Chem. Rev. 1995, 95, 2203–2227.
[6] Dynamic Combinatorial Chemistry (Eds.: J. N. H. Reek, S. Otto), Wiley-
VCH, Weinheim, 2010.
Received: November 29, 2013
Published online on February 23, 2014
Chem. Eur. J. 2014, 20, 3288 – 3291
www.chemeurj.org
3291
ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim