Enzymatic generation and in situ screening of a dynamic combinatorial library of sialic acid analogues

Article abstract of DOI:10.1002/1521-3773(20020916)41:18<3405::AID-ANIE3405>3.0.CO;2-P

Reversible formation of carbon-carbon bonds under physiological conditions by enzyme catalysis allows the generation and in situ screening of a dynamic mixture of biologically significant compounds. Generation of the dynamic library by incubation of the three sugars 1 a-c with two equivalents of sodium pyruvate in the presence of N-acetylneuraminic acid aldolase and wheat germ agglutinin resulted in amplification of sialic acid 1a (see scheme).

Full text of DOI:10.1002/1521-3773(20020916)41:18<3405::AID-ANIE3405>3.0.CO;2-P

COMMUNICATIONS
[7] S. O. Kelley, J. K. Barton, N. M. Jackson, L. D. McPherson, A. B.
Potter, E. M. Spain, M. J. Allen, M. G. Hill, Langmuir 1998, 14, 6781.
[8] S. O. Kelley, N. M. Jackson, M. G. Hill, J. K. Barton, Angew. Chem.
1999, 111, 991; Angew. Chem. Int. Ed. 1999, 38, 941.
[9] S. O. Kelley, E. M. Boon, N. M. Jackson, M. G. Hill, J. K. Barton,
Nucleic Acids Res. 1999, 27, 4830.
[10] E. M. Boon, D. M. Ceres, T. G. Drummond, M. G. Hill, J. K. Barton,
Nat. Biotechnol. 2000, 18, 1096.
[11] S. Mui, E. M. Boon, J. K. Barton, M. G. Hill, E. M. Spain, Langmuir
2001, 17, 5727.
+ X
+ X
+ X
A
B
C
AX
BX
CX
AX BX CX
T
+
+
+
A
B
C
X
X
X
AX
AX · T
BX
CX
AX BX CX
Scheme 1. The DCC concept: reversible reactions performed with a
limiting amount of X generate a mixture of compounds AX, BX, and
CX. The binding of AX to molecular trap T causes perturbation of the
equilibria involving A and X to give overall amplification of AX at the
expense of the other library members.
[12] E. M. Boon, J. K. Barton, unpublished results.
[13] E. M. Boon, J. E. Salas, J. K. Barton, Nature Biotechnology 2002, 20,
282.
[14] P. I. Pradeepkuman, E. Zamaratski, A. Foldesi, J. Chattopadhyaya,
Tetrahedron Lett. 2000, 41, 8601.
[15] P. I. Pradeepkuman, J. Chattopadhyaya, J. Chem. Soc. Perkin 2 2001,
2074.
[16] J. Chattopadhyaya, unpublished results.
[17] P. I. Pradeepkumar, E. Zamaratski, A. Foldesi, J. Chattopadhyaya, J.
Chem. Soc. Perkin 2 2001, 402.
particular, the in situ screening, which is such an attractive
feature of DCC, demands the use of conditions amenable
both to the library-formation and -trapping stages. If the trap
is envisaged as a protein or other biomolecule, the system is
limited to aqueous and near-physiological conditions, with
which few covalent bond forming reactions are compatible.
Thus, for example, Lehn and co-workers have employed both
imine and disulfide exchange reactions in DCC experiments
involving biomolecular traps,[2a] but to our knowledge no
DCC experiment involving the formation of carbon carbon
bonds under physiological conditions has been performed.[3]
We believe that enzyme-catalyzed reactions, which are
characteristically reversible under physiological conditions,
are ideally suited to the generation of dynamic combinatorial
libraries. Many enzymes with broad specificity (required for
library diversity) are already commercially available, and the
application of modern techniques in directed evolution may
be expected to increase their number. The products of an
enzyme-catalyzed reaction are usually stable compounds;
simple removal or inactivation of the enzyme stops the
reaction, thus reducing the dynamic mixture to a static library
which may be analyzed directly, without the need for a
derivatization step to freeze the product distribution. Herein
we present the first demonstration of DCC using enzyme
catalysis for the generation of a dynamic library.
Enzymatic Generation and In Situ Screening of
a Dynamic Combinatorial Library of Sialic
Acid Analogues**
Roger J. Lins, Sabine L. Flitsch,* Nicholas J. Turner,*
Ed Irving, and Stuart A. Brown
Dynamic combinatorial chemistry (DCC) is a rapidly
emerging field which offers a possible alternative to the
approach of traditional combinatorial chemistry (CC).[1]
Whereas CC involves the use of irreversible reactions to
efficiently generate static libraries of related compounds,
DCC relies upon the use of reversible reactions to generate
dynamic mixtures. The binding of one member of the dynamic
library to a molecular trap (such as the binding site of a
protein) is expected to perturb the library in favor of the
formation of that member (Scheme 1). Comparison of the
™perturbed∫ library with that generated in the absence of the
trap should indicate which members of the library are
interacting with the trap, which effectively offers in situ
screening of the combinatorial library.
In considering the application of enzyme catalysis to DCC,
we were encouraged by the thermodynamic resolution of a
dynamic mixture of aldol products by Whitesides and co-
workers through the use of a broad-specificity aldolase to
effect reversible formation of carbon carbon bonds under
mild conditions.[4] For the current investigation we chose a
related enzyme, N-acetylneuraminic acid aldolase (NANA
aldolase, EC 4.1.3.3), which catalyzes the cleavage of N-
acetylneuraminic acid (sialic acid, 1a) to N-acetylmannos-
amine (ManNAc, 2a), and sodium pyruvate 3 (Scheme 2). In
The DCC concept has already been proven through the
elegant experiments by several research groups, including
those of Lehn and Sanders.[1,2] However, significant exper-
imental challenges remain before the method may be
considered a practical complement to traditional CC. In
[*] Prof. S. L. Flitsch, Prof. N. J. Turner, Dr. R. J. Lins
Department of Chemistry
Edinburgh Protein Interaction Centre
The University of Edinburgh, King©s Buildings
West Mains Road, Edinburgh EH9 3JJ (UK)
Fax : (þ 44)131-650-4717 or (þ 44)131-650-4743
E-mail: s.flitsch@ed.ac.uk
O
OH
OH
R2
OH
R1
COONa
R2
3
O
HO
HO
O
COONa
R1
OH
NANA aldolase
HO
2
1
n.turner@ed.ac.uk
1a R1 = NHAc, R2 = CH2OH
1b R1 = OH, R2 = CH2OH
1c R1 = OH, R2 = H
2a R1 = NHAc, R2 = CH2OH
2b R1 = OH, R2 = CH2OH
2c R1 = OH, R2 = H
Dr. E. Irving, Dr. S. A. Brown
Ultrafine
Synergy House, Guildhall Close
Manchester Science Park, Manchester M15 6SY (UK)
Scheme 2. NANA aldolase catalyzes the cleavage of sialic acid 1a to
ManNAc 2a and sodium pyruvate 3; in the presence of excess sodium
pyruvate, aldol products 1a c are generated from the respective substrates
2a c.
[**] We are grateful to Ultrafine for a postdoctoral fellowship (R.J.L.) and
to the Wellcome Trust for financial support.
Angew. Chem. Int. Ed. 2002, 41, No. 18
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4118-3405 $ 20.00+.50/0
3405

COMMUNICATIONS
the presence of excess sodium pyruvate the equilibrium may
be driven toward the formation of an aldol product, in which
the enzyme will accept a range of reducing sugars as the
electrophilic component. This stereoselective aldol reaction
has been successfully employed in the efficient enzymatic
syntheses of sialic acid and several related compounds.[5] We
envisaged the generation of a small dynamic library of aldol
products 1a, 1b (ketodeoxynonulosonic acid, KDN), and 1c
(ketodeoxyoctulosonic acid, KDO) through the action of
NANA aldolase on a mixture of the corresponding substrates
2a, 2b (d-mannose), and 2c (d-lyxose).
Wheat germ agglutinin (WGA), a well-studied and readily
available plant lectin,[6] was chosen as the molecular trap.
WGA is known to specifically bind sialic acid with modest
(mm) affinity, with the diequatorial C-4 hydroxy and C-5
acetamido groups of sialic acid forming the primary recog-
nition motif. We thus expected amplification of sialic acid to
occur when the dynamic library containing sialic acid was
generated in the presence of WGA.
version to the aldol products,[7] and gave 2 5 mm concentra-
tions of each. It was evident that the aldol reaction had
reached equilibrium after 16 hours incubation, since the
product distribution changed very little after this time.
Evidence that the enzyme does indeed catalyze both aldol
formation and cleavage on the timescale employed was
demonstrated by re-equilibrating a mixture of sialic acid and
d-mannose in the presence of the NANA aldolase;[8] after
incubation overnight the mixture contained, in addition to the
initial components, KDN and sodium pyruvate which could
only arise through a retroaldol cleavage of sialic acid followed
by aldol formation of KDN.
When the incubations were performed in the presence of
WGA (sufficient lectin was added to provide at least one
equivalent of binding sites, based on the amount of sialic acid
formed in the control experiments) the product distribution
was observed to change dramatically over time (Figure 1b).
The system appears initially to approach a similar distribution
to that observed in the control incubation, but thereafter the
relative concentration of sialic acid is seen to increase, with a
corresponding drop in the relative concentration of KDN. A
plot of the percentage amplification in the relative concen-
tration (as estimated from relative peak area) of each aldol
product over controls is shown in Figure 2; the peak assigned
Generation of the dynamic library proved straightforward;
a mixture containing equimolar amounts of the three sub-
strates 2a c and two equivalents of sodium pyruvate was
incubated in the presence of NANA aldolase. Aliquots of the
incubation mixture were withdrawn at intervals and analyzed
by ion-exchange HPLC (Figure 1a, bottom). The three
Figure 2. Time course for a three-component system showing the relative
^
*
amplification of sialic acid ( ), KDO (&), and KDN ( ).
to sialic acid contributed approximately 40% of the total aldol
product peak area after 160 hours in the presence of WGA,
but only 22% of the total aldol product peak area in the
control mixture at the same time point. This difference
corresponds to a relative amplification of about 80%.
Conversely, the contribution of KDN to the total aldol
product peak area after 160 hours was 31% in the control
incubation, but only 6% in the presence of WGA; this
difference coresponds to a relative suppression of 80%.
The observation that KDN and KDO were not suppressed
to the same extent might be a consequence of WGA having a
weak binding affinity for KDO, such that the latter species
may also experience some measure of relative amplification
during incubation in the presence of the lectin. In effect, the
extent of relative amplification observed in the above experi-
ment suggests a ranking order of the three components based
on the relative affinities of WGA for each of them. In this
case, the two-component DCC experiment involving only
KDN and KDO should demonstrate some degree of ampli-
Figure 1. HPLC traces showing the change in concentration of sialic acid
(1a, blue), KDN (1b, green), and KDO (1c, red) over time, for a) control
incubation, b) incubation with WGA.
substrate sugars co-eluted with low retention times and did
not give reproducible peak areas, presumably because of
overlap with other variable-intensity peaks with low Rf values.
The peak assigned to sodium pyruvate (3) was also variable in
intensity; this peak was presumed to arise from the homoaldol
product of sodium pyruvate, which was not expected to be
stable under the elution conditions. The three aldol products
gave reproducible peak areas, although absolute retention
times were somewhat variable. As expected, the use of a small
excess of sodium pyruvate resulted in low (15 40%) con-
3406
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4118-3406 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 18

COMMUNICATIONS
fication of KDO, while that involving only sialic acid and
KDN should show amplification of sialic acid over KDN, with
the extent of amplification being greater in the latter experi-
ment. The data from these two-component experiments are
shown in Figure 3, bottom; they show dramatic relative
of dynamic combinatorial chemistry; we are currently looking
to increase the size of the library, to adapt the system for the
identification of novel ligands such as neuraminidase inhib-
itors, and to develop multienzyme systems for the generation
of dynamic libraries with increased diversity.
Experimental Section
N-acetylneuramic acid aldolase (EC 4.1.3.3, 22.2 Umgꢀ1) and wheat germ
agglutinin (lyophilized powder) were purchased from ICN Biomedicals.
The following stock solutions were prepared in 0.05m potassium phosphate
(KPi) buffer, pH 7.5: aldolase stock (20 UmLꢀ1); wheat germ agglutinin
stock (50 mgmLꢀ1); sugar stock solutions containing sodium azide
(0.5 gLꢀ1); 0.1 m sodium pyruvate; and 0.05 m each of ManNAc, d-
mannose, and d-lyxose (solution A), ManNAc and d-mannose (solu-
tion B), or d-mannose and d-lyxose (solution C).
Incubation mixtures A C contained enzyme stock (10 mL), WGA stock
(20 mL), and sugar stock A, B, or C (10 mL), respectively. Controls
contained enzyme stock (10 mL), 0.05m KPi buffer (20 mL), and sugar stock
A, B, or C (10 mL). After centrifugation (1 min@3000 rpm), the solutions
were incubated at 378C without stirring. Aliquots (5 mL) were withdrawn at
intervals, diluted to 500 mL with milli-Q water, and immediately analyzed
by HPLC.
Received: April 4, 2002 [Z19042]
[1] For recent reviews, see S. Otto, R. L. E. Furlan, J. K. M. Sanders, Drug
DiscoveryDev. 2002, 7, 117; G. R. L. Cousins, S. A. Poulsen, J. K. M.
Sanders, Curr. Opin. Chem. Biol. 2000, 4, 270; I. Huc, J.-M. Lehn, Proc.
Natl. Acad. Sci. USA 1997, 94, 2106; J.-M. Lehn, Chem. Eur. J. 1999, 5,
2455.
[2] a) O. Ramstrˆm, J.-M. Lehn, ChemBioChem 2000, 1, 41; I. Huc, J.-M.
Lehn, Proc. Natl. Acad. Sci. USA 1997, 94, 2106; b) R. L. E. Furlan,
Y. F. Ng, S. Otto, J. K. M. Sanders, J. Am. Chem. Soc. 2001, 123, 8876;
c) M. Hochg¸rtel, H. Kroth, D. Piecha, M. W. Hofmann, C. Nicolaou, S.
Krause, G. Schaaf, G. Sonnemoser, A. V. Eliseev, Proc. Natl. Acad. Sci.
USA 2002, 99, 3382.
[3] A library of vancomycin analogues has been prepared using ring-
closing metathesis reactions in aqueous solution (K. C. Nicolaou, R.
Hughes, S. Y. Cho, N. Wissinger, H. Labischinski, R. Endermann,
Chem. Eur. J. 2001, 7, 3824); a kinetic template effect was observed
when the library was prepared in the presence of a dipeptide trap.
[4] M. D. Bednarski, E. S. Simon, N. Bischofberger, W. D. Fessner, M. J.
Kim, W. Lees, T. Saito, H. Waldmann, G. M. Whitesides, J. Am. Chem.
Soc. 1989, 111, 627.
[5] M. J. Kim, W. J. Hennen, M. Sweers, C. H. Wong, J. Am. Chem. Soc.
1988, 110, 6481; C. Augÿ, B. Bouxom, B. Cavayÿ, C. Gautheron,
Tetrahedron Lett. 1989, 30, 2217.
[6] B. P. Peters, S. Ebisu, I. J. Goldstein, M. Flashner, Biochemistry 1979,
18, 5505.
[7] As determined by HPLC analysis of standard solutions of synthetic
sialic acid.
[8] The control mixture contained of d-mannose and ManNAc (0.5 mmol
each), sialic acid (0.13 mmol), and NANA aldolase in buffer solution
(0.2 U); total volume: 40 mL.
Figure 3. Time courses for two-component DCC experiments showing:
^
*
top: relative amplification of sialic acid ( ) over KDN ( ), bottom: relative
*
amplification of KDO (&) over KDN ( ).
amplification of sialic acid over KDN, but more modest
relative amplification of KDO over KDN when sialic acid is
not present. Since all three experiments showed supression of
KDN, it was necessary to exclude the possibility that this
component was being affected by a selective aldolase activity
associated with the lectin preparation. Thus, a mixture of
sodium pyruvate, mannose, manNAc, sialic acid, and KDN, in
proportions approximating their equilibrium distribution in
the two component control experiment, was prepared from
commercial samples and incubated in the presence of WGA
(but not NANA aldolase). No change from the initial
distribution was observed, which demonstrates that the
changes in product distribution observed in the DCC experi-
ments were caused by WGA acting as a thermodynamic trap,
and not as a selective catalyst for KDN degradation.
In conclusion, we have successfully demonstrated, for the
first time, the generation and in situ screening of a dynamic
mixture of biologically significant compounds, where enzyme
catalysis has been used to effect reversible formation of
carbon carbon bonds under physiological conditions. We
believe this work can be extended to larger libraries through
the use of broad-specificity enzymes, which represent a
powerful (and hitherto overlooked) tool for the development
Angew. Chem. Int. Ed. 2002, 41, No. 18
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4118-3407 $ 20.00+.50/0
3407