Ligand denticity controls enantiomeric preference in DNA-based asymmetric catalysis

Article abstract of DOI:10.1039/c2cc17350f

DNA-based catalysis can be used to control the enantioselectivity of copper-catalysed Diels-Alder and Friedel-Crafts reactions to produce either enantiomer of the product by changing the denticity of the ligand coordinated to the Cu(ii) ion, even though t

Full text of DOI:10.1039/c2cc17350f

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 2394–2396
www.rsc.org/chemcomm
COMMUNICATION
Ligand denticity controls enantiomeric preference in DNA-based
asymmetric catalysisw
a
b
a
a
Arnold J. Boersma, Bas de Bruin, Ben L. Feringa* and Gerard Roelfes*
Received 24th November 2011, Accepted 10th January 2012
DOI: 10.1039/c2cc17350f
DNA-based catalysis can be used to control the enantioselectivity
of copper-catalysed Diels–Alder and Friedel–Crafts reactions to
produce either enantiomer of the product by changing the denticity
of the ligand coordinated to the Cu(II) ion, even though the DNA
adopts a right handed helical conformation only.
Artificial metalloenzymes have shown great potential in
1
,2
enantioselective catalysis. Using protein or DNA as a chiral
host for catalytically active transition metal complexes, excellent
enantioselectivities have already been achieved in a variety of
reactions. However, one of the major challenges in this area is
how to obtain both enantiomers of the product in a selective
fashion at will, since the biomolecular host is generally available
as a single enantiomer. In protein based artificial metallo-
enzymes, a combination of chemical optimisation of the ligand
and mutagenesis of the host protein can be used to obtain the
opposite enantiomer of the reaction product of the catalysed
Scheme 1 DNA-based catalytic asymmetric reactions and Cu(II)
3
reaction. Here we show that control over the enantiomeric
complexes of bi- and terpyridine ligands used in this study.
preference in DNA-based asymmetric catalysis is achieved by
changing the denticity of the poly-pyridine-type ligand bound
to the transition metal.
the Diels–Alder, Michael and Friedel–Crafts product revealed
that the enantiomeric outcome of the reaction is predictable:
the reactant, that is the diene or nucleophile, always attacks
DNA-based asymmetric catalysis involves binding of a
catalytically active metal complex of an achiral ligand to DNA
6
–8
the Cu(II) bound enone substrate from the same p-face. This
suggests a similar mechanism of enantioselection induced by
DNA for all three reaction classes. With the first generation
DNA-based catalysts, which are based on acridine intercalators
and give only moderate enantioselectivity in the aforementioned
C–C bond forming reactions, control over the enantiomeric
outcome of the reaction was achieved by variation of the ligand
4,5
via a covalent linkage or supramolecular interactions. The
proximity to the DNA allows for the chiral environment of the
DNA-helix to direct a catalysed reaction towards selective
formation of one of the enantiomers of a chiral product.
Especially the supramolecular approach has been applied
successfully to the archetypal C–C bond forming reactions;
9
0
6
7
structure. However, the 2,2 -bipyridine derived ligands of the
second generation such as L1, which induce the highest ee’s in
these reactions, offer much less opportunities for structural
variation. For this reason we have investigated the possibility
of changing the structure of the substrate bound complex by
variation of the ligand denticity, by comparison of a series of
bi- and terpyridine ligands in DNA-based asymmetric catalysis
the Diels–Alder reaction, the Michael addition, and the
8
Friedel–Crafts alkylation. In all these cases excellent enantio-
selectivities, and in multiple cases even up to 99% ee, have been
0
0
found using the Cu(II) complex of 4,4 -dimethyl-2,2 -bipyridine
Cu–L1) in combination with salmon testes DNA (st-DNA).
(
Interestingly, an analysis of the absolute configuration of
(
Scheme 1).
a
Stratingh Institute for Chemistry, University of Groningen,
The DNA-based catalysts were prepared via self-assembly of
Cu–L with st-DNA. The Cu–L/st-DNA catalysed Diels–Alder
reaction between aza-chalcone 1a and cyclopentadiene 2 was
Nijenborgh 4, 9747 AG Groningen, The Netherlands.
E-mail: b.l.feringa@rug.nl, j.g.roelfes@rug.nl
Van’t Hoff Institute for Molecular Sciences,
University of Amsterdam, Science Park 904, 1098 XH Amsterdam,
The Netherlands
w Electronic supplementary information (ESI) available: Experimental
procedures of synthesis and characterization of catalysts, catalytic
experiments, and DFT calculations. See DOI: 10.1039/c2cc17350f
b
1
0
used as the benchmark reaction. This reaction results in the
formation of the endo and exo isomers, with the endo isomer
being the major product. Therefore only the enantiomeric
excess of the endo isomer is discussed.
2
394 Chem. Commun., 2012, 48, 2394–2396
This journal is c The Royal Society of Chemistry 2012

View Article Online
Table 1 Effect of ligand structure on the ee obtained in DNA-based
asymmetric catalysis
An exception is when using L9: in this case full conversion was
achieved (entry 9).
a
b
c
c
The inversion of the enantiomeric outcome of the catalysed
reaction upon going from the bipyridine ligand L1 to the
terpyridine-type ligands is independent of the substrate and
reaction-type. In the presence of Cu–L7/st-DNA, the Diels–Alder
reaction between 1c and 2 gave 3c with an ee of 42% (2R,3R-
enantiomer), compared to 97% ee of the 2S,3S-enantiomer in the
case of Cu–L1 (entries 13,14), and the Friedel–Crafts reaction
between 1d and 4 induced an ee of 50% of the (ꢀ)-enantiomer
(entry 16), compared to 83% ee of the (+)-enantiomer in the
case of Cu–L1 (entry 15). The most dramatic differences in
enantiomeric outcome were obtained in the Diels–Alder reaction
of 1b with 2; using Cu–L1/st-DNA or Cu–L7/st-DNA as
catalyst, the corresponding Diels–Alder product was obtained
with >+99 and ꢀ92% ee, respectively (Table 1, entries 11 and
12; in this case the + and ꢀ sign indicate elution order from the
chiral HPLC: first and second, respectively).
Cu–L
Reactants Conversion (%) Endo : exo ee (%)
Diels–Alder reaction
d
1
2
3
4
5
6
7
8
9
1
1
1
1
1
Cu–L1 1a
Cu–L2 1a
Cu–L3 1a
Cu–L4 1a
Cu–L5 1a
Cu–L6 1a
Cu–L7 1a
Cu–L8 1a
Cu–L9 1a
Cu–L10 1a
Cu–L1 1b
Cu–L7 1b
Cu–L1 1c
Cu–L7 1c
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Full
Full
Full
Full
Full
Full
23
20
Full
17
>99 : 1
98 : 2
98 : 2
95 : 5
99 : 1
98 : 2
89 : 11
92 : 8
94 : 6
92 : 8
>99 : 1
92 : 8
99 : 1
82 : 18
99 (+)
90 (+)
89 (+)
40 (ꢀ)
95 (+)
92 (+)
60 (ꢀ)
79 (ꢀ)
28 (ꢀ)
71 (ꢀ)
>+99
d
0
1
2
3
4
e
25
10
>80
15
e
ꢀ92
f
97 (2S,3S)
42 (2R,3R)
Friedel–Crafts alkylation
g
1
1
5
6
Cu–L1 1d
Cu–L7 1d
4
4
Full
20
—
—
83 (+)
50 (ꢀ)
a
These results indicate that the geometry of the Cu(II)–substrate
complex is crucial to the enantiomeric preference observed. To
shed some light on the geometries of the Cu(II)–substrate com-
Reactions performed with 1 mM substrate, 8 mM 2 (or 5 mM 4),
ꢀ1
0
.3 mM Cu–L, 2 mg ml st-DNA, in 20 mM mops buffer pH 6.5, for
b
3
days at 5 1C. All data averaged over two experiments. Determined
1
c
d
e
2
+
by H-NMR. Determined by HPLC. Data from ref. 6a. + and ꢀ
plexes, DFT calculations of the structure of [Cu–L1(1b)(H O)]
2
,
sign correspond to elution order on the HPLC (first and second,
2+
2+
[Cu–L1(1b)
2
]
2
and [Cu–L7(1b)(H O)]
were performed in
f
g
respectively). Data taken from ref. 6b. Data taken from ref. 8.
the absence of DNA. Different geometrical isomers of these
complexes were examined, of which the most stable ones are
depicted in Fig. 1.
First, a series of bidentate ligands with varying substitution
patterns were investigated. Based on the results of the catalysis, it
can be concluded that the substitution pattern on the bipyridine
ligand has a significant effect on the enantioselectivity of the
Diels–Alder reaction (Table 1, entries 1–6). In particular,
Attempts to optimise the octahedral bis-aqua complex
2
+
[Cu–L1(1b)(H O) ]
2
led spontaneously to the mono-aqua
2
2+
complex [Cu–L1(1b)(H O)] ꢁH O, which is best described as a
2
2
5-coordinate trigonal bipyramidal complex to which an additional
water molecule hydrogen bridges with both the carbonyl moiety
of the coordinated substrate and the aqua ligand. The
0
substitution at the 4 and 4 -position of the ligand gave rise
to an increase of the ee of the (+)-enantiomer compared to
Cu–L2, which has no substituents. Since the substitution of
2+
5-coordinate mono-aqua complex [Cu–L1(1b)(H O)] without
2
0
the 4 and 4 -position is too remote to have a direct effect on the
this additional hydrogen bonded water molecule was calculated
ꢀ1
coordination of 1a to the copper(II) ion, it is likely that
substituents at this position enforce a more favourable binding
geometry around the copper(II)–ligand complex in the DNA in
order to affect higher ee’s in the catalysed reaction. Notably,
to be B9 kcal mol (b3-lyp, def-TZVP) less stable than
2+
[Cu–L1(1b)(H O)] ꢁH O.
2
2
2
+
Formation of the bis-substrate adduct [Cu–L1(1b) ] from
2
2
+
the mono-substrate adduct [Cu–L1(1b)(H O)] ꢁH O and
2
2
Cu–L6, which contains bulky tert-butyl groups at the 4- and
0
an additional substrate molecule is energetically favoured
4
-position, still gives rise to a higher ee than Cu–L2 (entry 6).
0
A methyl substituent at the 5 and 5 -position induced a similar
ee as Cu–L2 (entry 3). Surprisingly, when the methyl groups
0
are placed at the 6 and 6 -position, the (ꢀ)-enantiomer of the
Diels–Alder product was obtained with an ee of 40% (entry 4).
This is the opposite enantiomer than produced by the non-
0
0
substituted or 4,4 - and 5,5 -substituted bipyridine complexes
in st-DNA.
Catalysts based on the terpyridine-type ligands L7–L10 gave
invariably rise to an excess of the (ꢀ)-enantiomer (entries 7–10),
albeit that the ee’s were generally lower than those obtained with
the bipyridine ligands (Table 1, entry 7). An electron donating or
withdrawing substituent at the 4 position of the central pyridine
moiety led to an improved ee (entries 8 and 10), whereas a large
tolyl substituent at the same position caused a significantly
decreased ee (entry 9). This suggests that the substituents
influence the ee by steric effects mainly, as opposed to electronic
effects. Notably, the catalysts based on terpyridine-type ligands
gave rise to a significantly lower conversion compared to
bipyridine-type ligands as well as a lower endo : exo ratio.
2
+
Fig. 1 DFT optimised geometries of [Cu–L1(1b)(H O)] ꢁH O,
2
2
2+
2+
[Cu–L1(1b)
2
]
2
, and [Cu–L7(1b)(H O)] in the absence of st-DNA.
This journal is c The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 2394–2396 2395

View Article Online
ꢀ
1
(
DE E 19 kcal mol ) according to DFT. The relevance of the
been shown to be competitive with conventional homogeneous
1
3
bis-substrate adduct to DNA-based catalysis is not yet clear,
since this depends for example on the kinetics of the catalysed
reaction, the rate of complex formation, and the influence of
DNA on these parameters. However, the fact that the carbonyl
oxygen is coordinated at the axial position suggests that the
enone in this case is less activated (see below) and, thus, most
likely less reactive.
catalysts in several reactions, further optimization of the
Cu-terpy type catalysts, i.e. to enhance their catalytic activity,
is required for synthetic applications. Current investigations
are directed towards elucidating the structure of the substrate
coordinated complexes when bound to DNA and identifying
the second coordination sphere interactions that are important
for enantioselective catalysis.
2
+
[Cu–L7(1b)(H
coordinated Cu
2
+
O)]
was found to have an octahedrally
ion with the pyridine nitrogen atoms and
This work was supported by the Netherlands Research
School Combination Catalysis (NRSC-C).
2
the pyridyl nitrogen of the substrate occupying the equatorial
plane and the carbonyl oxygen and a water molecule at the
Notes and references
1
1
axial positions.
z Computational studies of the Cu–L–1b complex bound to DNA, such
as docking experiments, are at present not feasible. A key requirement
for obtaining useful data from docking experiments is knowledge about
the DNA binding mode of these complexes, e.g. intercalation vs. groove
binding. However, to date this information is not available.
The combined experimental and theoretical results, even if
these were obtained for the catalyst in the absence of DNAz,
strongly suggest that the different structure of the ternary
complex Cu–L1–1b compared to Cu–L7–1b is related to the
observed inversion in enantiomeric preference of the catalysed
Diels–Alder reaction. Changing the binding geometry of the
substrate apparently causes the chiral pocket created by the
DNA to favour the opposite enantiomer, either by shielding or
facilitating the attack of the diene or nucleophile from the
opposite face of the enone. This also rationalises the results
obtained when using L4: the proximal methyl groups of ligand L4
should force the substrate to adopt a different binding geometry,
1
(a) A. J. Boersma, R. P. Megens, B. L. Feringa and G. Roelfes,
Chem. Soc. Rev., 2010, 39, 2083; (b) S. Park and H. Sugiyama,
Angew. Chem., Int. Ed., 2010, 49, 3870; (c) S. K. Silverman,
Angew. Chem., Int. Ed., 2010, 49, 7180.
2
(a) Y. Lu, N. Yeung, N. Sieracki and N. M. Marshall, Nature,
009, 460, 855; (b) F. Rosati and G. Roelfes, ChemCatChem, 2010,
, 916; (c) M. R. Ringenberg and T. R. Ward, Chem. Commun.,
2
2
2011, 8470; (d) P. J. Deuss, R. den Heeten, W. Laan and P. C. J.
Kamer, Chem.–Eur. J., 2011, 17, 4680.
3
4
(a) C. Letondor, A. Pordea, N. Humbert, A. Ivanova, S. Mazurek,
M. Novic and T. R. Ward, J. Am. Chem. Soc., 2006, 128, 8320;
2+
more reminiscent of the structure of [Cu–L7(1b)(H O)] . The
2
(b) M. Du
¨
E. Nogueira, L. Kno
rrenberger, T. Heinisch, Y. M. Wilson, T. Rossel,
rr, A. Mutschler, K. Kersten, M. J.
proposed structures also account for the observed differences in
activity of the bipyridine compared to the terpyridine based
catalysts, with the exception of Cu–L9. In Cu–L1–1b, the
carbonyl oxygen is coordinated on a pseudo-equatorial site,
which results in a better activation of 1b, compared to
¨
Zimbron, J. Pierron, T. Schirmer and T. R. Ward, Angew. Chem.,
Int. Ed., 2011, 50, 3026.
(a) G. Roelfes and B. L. Feringa, Angew. Chem., Int. Ed., 2005,
44, 3230; (b) S. Roe, D. J. Ritson, T. Garner, M. Searle and
J. E. Moses, Chem. Commun., 2010, 4309.
1
2
5 (a) N. Sancho Oltra and G. Roelfes, Chem. Commun., 2008, 6039;
b) P. Fournier, R. Fiammengo and A. Jaschke, Angew. Chem.,
Int. Ed., 2009, 48, 4426.
6 (a) G. Roelfes, A. J. Boersma and B. L. Feringa, Chem. Commun.,
006, 635; (b) A. J. Boersma, B. L. Feringa and G. Roelfes,
Org. Lett., 2007, 9, 3647.
D. Coquiere, B. L. Feringa and G. Roelfes, Angew. Chem., Int. Ed.,
Cu–L7–1b, where it is bound at the axial position. An
intriguing aspect of the calculated structure of Cu–L1–1b is
the water molecule that was found to bridge the carbonyl
oxygen of 1b and the Cu(II) coordinated water molecule via
hydrogen bonding. Although it cannot be established at
present whether such a bridging interaction could be present
in the DNA-based system, it does constitute an example of a
second coordination sphere interaction that could be important
for the observed stereochemistry and catalytic activity.
(
¨
2
7
8
9
`
2007, 46, 9308.
A. J. Boersma, B. L. Feringa and G. Roelfes, Angew. Chem.,
Int. Ed., 2009, 48, 3346.
F. Rosati, A. J. Boersma, J. E. Klijn, A. Meetsma, B. L. Feringa
and G. Roelfes, Chem.–Eur. J., 2009, 15, 9596.
In conclusion, in the present communication it is shown that
DNA-based catalysis can be used to produce both enantiomers
of the product of the catalysed reaction selectively by judicious
choice of ligands for the Cu(II) ion, even though salmon testes
DNA adopts a right handed helical conformation only. This
enantiomeric preference was demonstrated to be related to the
denticity of the ligand and the resulting structure of the substrate
bound copper complex. In contrast to Cu–L1/st-DNA, which has
10 S. Otto, F. Bertoncin and J. B. F. N. Engberts, J. Am. Chem. Soc.,
996, 118, 7702.
1
1
1 (a) For similar copper–bipyridine and –terpyridine structures see;
S. Suzuki, T. Sakurai, S. Itoh and Y. Ohshiro, Inorg. Chem., 1988,
27, 591; (b) T. Kohzuma, A. Odani, Y. Morita, M. Takani and
O. Yamauchi, Inorg. Chem., 1988, 27, 3854.
1
2 The bond between copper(II) and an axial ligand is elongated
compared to the equatorial ligand as a result of the Jahn–Teller
distortion.
13 R. P. Megens and G. Roelfes, Org. Biomol. Chem., 2010, 8, 1387.
2
396 Chem. Commun., 2012, 48, 2394–2396
This journal is c The Royal Society of Chemistry 2012