Design, synthesis and characterisation of a peptide with oxaloacetate decarboxylase activity

Article abstract of DOI:10.1016/S0960-894X(01)00519-4

The synthesis of Oxaldie-3, a synthetic 31-residue peptide with oxaloacetate decarboxylase activity, is described. Biophysical characterisation by gel filtration, CD and NMR spectroscopy indicated that the peptide adopted a folded structure in solution. Oxaldie-3 was an efficient catalyst at concentrations as low as 2 μM, 100-fold lower than the previously described Oxaldie-2, which relied on aggregating α-helices for activity. Oxaldie-3 speeded decarboxylation by more than three orders of magnitude relative to simple amines.

Full text of DOI:10.1016/S0960-894X(01)00519-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2631–2635
Design, Synthesis and Characterisation of a Peptide with
Oxaloacetate Decarboxylase Activity
Susan E. Taylor, Trevor J. Rutherfordy and Rudolf K. Allemann*
The University of Birmingham, School of Chemistry, Edgbaston, Birmingham, B15 2TT, UK
Received 11 May 2001;revised 10 July 2001;accepted 25 July 2001
Abstract—The synthesis of Oxaldie-3, a synthetic 31-residue peptide with oxaloacetate decarboxylase activity, is described. Bio-
physical characterisation by gel filtration, CD and NMR spectroscopy indicated that the peptide adopted a folded structure in
solution. Oxaldie-3 was an efficient catalyst at concentrations as low as 2 mM, 100-fold lower than the previously described Oxaldie-
2, which relied on aggregating a-helices for activity. Oxaldie-3 speeded decarboxylation by more than three orders of magnitude
relative to simple amines. # 2001 Elsevier Science Ltd. All rights reserved.
Biological macromolecules of defined structure and with
catalytic activity can be created through selection of a
few catalytically active biomolecules from a large pool
of randomly generated, mainly inactive molecules1ꢀ4 or
by rational design of biomolecules that fold in solution
to present to the substrate an array of catalytic func-
tional groups.5ꢀ7 The success of designing polypeptide
catalysts is a vigorous test of our understanding of
protein folding and enzyme catalysis.8
solution and served as the scaffolding for holding a
reactive amine. Proper folding and catalytic activity of
Oxaldie-2 were concentration dependent. Above con-
centrations of 200 mM, where Oxaldie-2 was fully heli-
cal, imine formation between oxaloacetate and the
active amine of Oxaldie-2 was speeded by 2–3 orders of
magnitude relative to simple amines.5 Attempts to
identify better catalysts by screening a combinatorial
library of an 18-amino acid peptide based on the
sequence of the Oxaldie-2 led to a further rate enhance-
ment of not even half an order of magnitude relative to
Oxaldie-2,14,15 indicating that the simple design of
Oxaldie-2 could not be improved much further. How-
ever, these results also stressed the remarkable success
of de novo design, which at least in the case of Oxaldie-
2 could compete with combinatorial approaches.
In the present investigation we report the design, synth-
esis and characterisation of an unnatural oxaloacetate
decarboxylase. While all natural oxaloacetate decar-
boxylases depend on metal catalysis, the decarboxyl-
ation of acetoacetate by acetoacetate decarboxylase is
speeded up by an active site lysine with a pKa of
approximately 6, 4 orders of magnitude lower than the
side chain of an unperturbed lysine.9,10 Acetoacetate
decarboxylase accomplishes this remarkable reduction
in the pKa by placing a second amino group close to the
reactive lysine in the active site.11 Based on a general
understanding of the amine-catalysed decarboxylation
of beta-keto acids, a reaction that proceeds through a
protonatable imine intermediate, we previously reported
the construction of a peptide with oxaloacetate decar-
boxylase activity (Oxaldie-2).5,12,13 Oxaldie-2 folded
into an amphiphilic a-helix, where one side of the helix
was hydrophilic and the other hydrophobic. These
amphiphilic helices formed helical bundles in aqueous
The catalytic activity of Oxaldie-2 was dependent on the
high concentrations necessary for the formation of the
helical bundles. Here we report the design, synthesis and
characterisation of Oxaldie-3, a 31-amino acid residue
polypeptide, which forms a stable structure in solution
and acts as an efficient catalyst for the decarboxylation of
oxaloacetate over a wide range of peptide concentrations.
Design and Synthesis of Peptide Scaffold
In order to create an enzyme with increased specific
oxaloacetate decarboxylase activity the catalytically
active amine was introduced into the scaffold of avian
pancreatic polypeptide to create Oxaldie-3 (Fig. 1).
Pancreatic polypeptides (PP) are single chain peptides of
36 amino acids.16 An X-ray structure of avian PP (aPP)
*Corresponding author. Tel.: +44-121-414-4359;fax: +44-121-414-
4446;e-mail: rkallemann@chemistry.bham.ac.uk
yPresent address: MRC-Centre for Protein Engineering, Hills Road,
Cambridge, CB2 2QH, United Kingdom.
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00519-4

2632
S. E. Taylor et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2631–2635
at 1.4 A and a solution structure of PP from cow are
available.17ꢀ19 Despite their small size, PPs adopt a
globular conformation which comprises of a proline-
rich type II helix (residues 1–8) and an a-helix (residues
14–32) (Fig. 2). Residues 33–36, which are thought to be
important for the biological activity of the peptides, are
in an unordered conformation in solution and are not
necessary for the formation of the three-dimensional
structure. The two helices, which are joined by a type II
beta-turn, interact through hydrophobic contacts to
form a compact and stable tertiary structure. Further
hydrophobic contacts in aPP are formed between sub-
units to form a stable dimer with a dissociation constant
of approximately 100 nM.20 In the design of Oxalide-3
the active site lysines were introduced on the solvent
exposed face of the a-helix of aPP by replacing Ile (18),
Asp (22), Gln (25) and Asn (29) (Fig. 1). Therefore the
hydrophobic core between the a-helix and the proline-
rich helix as well as the interactions between the sub-
units of the protein dimer are not affected. To stabilise
the helices, the N-terminal amine of Oxaldie-3 was
acetylated and the carboxy-terminus designed as an
amide.
Oxaldie-3 was synthesised by solid-phase peptide
synthesis using Fmoc chemistry and purified by reversed
phase HPLC. The mass of Oxaldie-3 was determined by
MALDI-TOF and electrospray mass spectrometry as
3612, which matches the calculated mass.
Structural Characterisation of Oxaldie-3
Gel filtration at peptide concentrations of 6.3 mM and
155 mM through Sephadex G50 in 50 mM potassium
phosphate (pH 7) and 10 mM NaCl revealed an appar-
ent molecular weight for Oxaldie-3 of ꢁ8600 by com-
parison with the retention times of vitamin B12, bPP (1–
31), a-lactalbumin, myoglobin, trypsin, and pepsin.
Like wild-type aPP and bPP,20 Oxaldie-3 appeared to be
dimeric under these conditions.
Figure 1. Sequence alignment of Oxaldie-3 and avian and bovine
pancreatic polypeptides. The proline-rich type II helix and the a-heli-
cal region are indicated. The residues, which were substituted by
lysines, are indicated in bold.
The circular dichroism spectrum of Oxaldie-3 (1 mM
potassium phosphate, pH 7, 10 mM NaClO4) at room
temperature showed that significant amounts of the
peptide were in an a-helical conformation (Fig. 3). The
spectrum compared well with that reported for wild-
type aPP.21 From the ellipticity at 222 nm of ꢀ19,877
deg cm2 dmolꢀ1, Oxaldie-3 was calculated to be 52% a-
helical.22 In the crystal structure of aPP 58% of the
residues are in a helical conformation.17 Contrary to
Oxaldie-2, the molar ellipticities in the far UV circular
dichroism spectrum of Oxaldie-3 were independent of
peptide concentration between 1 and 200 mM.
The near UV CD spectrum of Oxaldie-3 at 4.5 mM
peptide concentration showed a doublet characteristic
of a tyrosine residue (Fig. 3). Tyrosine (21) of aPP is
located at the edge of the dimer interface where it makes
a close asymmetric contact with the equivalent tyrosine
in the second subunit.17,18 These tyrosines at the dimer
interface of aPP have been identified as the likely source
of the strong CD signals around 280 nm.
The thermal denaturation of Oxaldie-3 was monitored
by measuring the temperature dependence of the CD
spectrum (Fig. 3). Isodichroic points were observed at
204 and 238 nm, suggesting a two-state equilibrium.
Plotting the CD signal as a function of temperature
revealed a clear plateau for the CD signal at high tem-
peratures (Fig. 3). At the low temperature end the
levelling off of the CD signal was less pronounced.
However, the shape of the melting curve was sigmoidal
and a melting temperature of 34 ꢂC was estimated by
curve fitting. Flat melting curves are typical for small
globular proteins because the folding transition shows
relatively little co-operativity. On the other hand, the
Figure 2. The structure of Oxaldie-3 showing the positions of the
lysine residues on the surface of the a-helix. The proline residues
and the aromatic amino acids important for the stability of the folding
unit are indicated. The model was based on the X-ray structure of
aPP.17

S. E. Taylor et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2631–2635
2633
flatness of the melting curve could also be interpreted as
a gradual decrease in the helical content with increasing
temperature as is typically observed for molten globules.
sequence specific assignments. The side chain reso-
nances of the 3 leucine, 3 valine and 4 lysine residues,
for example, were only poorly resolved (Fig. 4). The
lack of spectral dispersion and the random coil-like
chemical shifts indicated that Oxaldie-3 did not adopt
the same stable tertiary fold in solution as bPP.23 How-
ever, NOESY cross-peaks depicting strong inter-residue
NOEs indicated the presence of defined secondary
structure (Fig. 5). For example, strong NOEs between
the side chain resonances of tyrosine and leucine resi-
dues were taken as evidence that Oxaldie-3 did not
adopt a random coil conformation. Weak NOEs
between backbone amide protons were consistent with
a-helical structure. Weak NOE cross peaks were also
observed for the side chain stacking interactions
between proline and tyrosine and/or phenylalanine resi-
dues. Spectra recorded for a separate sample of Oxaldie-
3 at 30 ꢂC, at 4 ꢂC with 5% trifluoroethanol as the
cosolvent, and at 4 ꢂC in the presence of 150 mM NaCl
showed similar low dispersion of signals. Minor 1H
chemical shift perturbations gave rise to a significantly
different appearance of the one-dimensional 1H spec-
trum. Comparison of TOCSY spectra at 4 and 30 ꢂC,
however, revealed that the shift changes were relatively
minor, indicating no substantial restructuring of the
1H–1H TOCSY, HSQC and HMQC-TOCSY NMR
1
spectra enabled the assignment of many H resonances
for the individual amino acids, although the spectral
dispersion was insufficient to fully resolve the spin sys-
tems of amino acids of the same type and obtain
1
polypeptide. The H NMR spectra, therefore, exhibited
the characteristics of a molten globule state observed for
partially folded polypeptides, with 1H chemical shifts
closer to those of a random coil than to those of a fol-
ded state.24 The molten globule state is characterised by
a compactness similar to that of the native state with
native-like secondary structure and a fluctuating tertiary
structure resulting from non-rigid side chain packing.25
The formation of a molten globule-like state of Oxaldie-
3 can have two reasons. Either the introduction of the 4
lysines on the solvent exposed face of the a-helix of aPP
destabilised the native folded form of Oxaldie-3 or aPP
does not adopt the same conformation in solution as in
the crystalline state. Residues 18, 22, 25 and 29 of aPP,
which were replaced with lysines in Oxaldie-3, are not
involved in interactions between the a-helix and the
Figure 3. Top: thermal denaturation of 20 mM Oxaldie-3 in 1 mM
potassium phosphate (pH 7) and 10 mM NaClO4 as monitored by CD
spectroscopy. Temperature increments are 10 ꢂC starting at 7 ꢂC.
Middle: molar ellipticities at 208 (filled squares) and 222 (open octa-
gons) nm of a 20 mM solution of Oxaldie-3 as a function of tempera-
ture (1 mM potassium phosphate (pH 7), 10 mM NaClO4). Bottom:
near UV CD spectrum of 4.5 mM Oxaldie-3 in 1 mM potassium
phosphate (pH 7) and 10 mM NaCl.
Figure 4. 500 MHz 1H–1H TOCSY (50 ms DIPSI-3, 9.7 kHz B1) of
Oxaldie-3 (5.9 mM in H2O;pH 5.4;30 ꢂC) showing random coil-like
chemical shifts and limited dispersion of chemical shifts.

2634
S. E. Taylor et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2631–2635
Figure 5. Partial NOESY (200 ms tm) spectrum of Oxaldie-3 (5.9 mM in H2O;pH 5.4;30 ꢂC). Significant inter-residue NOE cross peaks are
observed between the side chain resonances of tyrosine and leucine residues and for unresolved HN–HN contacts. Almost all other cross peaks
between amide and aliphatic 1H resonances derive from intra-residue NOEs.
proline-rich helix or between the subunits of the dimer.
In addition, CD spectroscopy indicated that a sig-
nificant amount of Oxaldie-3 was in the a-helical con-
formation. Together these observations suggest that,
unlike its bovine homologue,19 aPP might adopt a
molten globule-like conformation in solution.
imine formation was no longer rate determining.5 Oxal-
die-3 was active at smaller concentrations of catalyst
than Oxaldie-2, but the catalytic mechanism was simi-
lar. Like Oxaldie-2, it uses conformational energy of its
polypeptide fold to force the active site lysine into an
environment where its activity is enhanced, and Cou-
lombic binding interactions between the cationic poly-
peptide and the substrate to lower the energy of the
transition state following carbinolamine. However,
because imine formation is no longer rate determining,
a step following imine formation between oxaloacetate
and Oxaldie-3 must have been speeded by one order of
magnitude. Tyr (21) lies amidst the lysine cluster and its
presence could change the solvation of the ‘active site’
thereby speeding decarboxylation.
Kinetics of the Oxaldie-3catalysed decarboxylation
reaction
Rates for the Oxaldie-3 catalysed decarboxylation of
oxaloacetate to produce pyruvate were measured by
following the conversion of NADH to NAD+ in
the lactate dehydrogenase mediated transformation of
pyruvate to lactate. Oxaldie-3 catalysed the decarboxyl-
ation of oxaloacetate with Michaelis–Menten kinetics
(kcat=5.16 minꢀ1; KM=49.4 mM) (Table 1). In contrast
to Oxaldie-2, where a concentration of 200 mM was
needed for optimal activity, both KM and kcat were
independent of the concentration of Oxaldie-3 between
2 and 200 mM. The catalytic efficiency of Oxaldie-2 was
reduced 2-fold by reducing its concentration to 100
mM.5 For smaller concentrations Oxaldie-2 displayed a
catalytic activity similar to that observed for simple
amines as a consequence of its random coil-like struc-
ture. The introduction of an active site for the decar-
boxylation of oxaloacetate into the scaffold of aPP
generated a polypeptide, which acts as a catalyst at a
100-fold reduced concentration compared to Oxaldie-2.
The catalytic activity of Oxaldie-3 depended on the
cluster of lysine residues in the a-helix;replacing the
lysine residues with those found in wild-type aPP reduced
the catalytic activity by two orders of magnitude.
In conclusion, Oxaldie-3 showed the characteristics of
a pancreatic polypeptide-like structure according to
CD spectroscopy and size-exclusion chromatography.
Despite adopting a molten globule-like structure in
solution, Oxaldie-3 was an efficient catalyst for the
decarboxylation of oxaloacetate at concentrations as
low as 2 mM, two orders of magnitude lower than for
the first generation Oxaldies. It appeared to speed the
formation of the protonatable imine intermediate as
well as the actual decarboxylation reaction. Relative to
Table 1. Kinetic data for the decarboxylation of oxaloacetate with
Oxaldie-2,5 Oxaldie-3 and with simple aminesa
Catalyst
[Cat]
(mM)
kcat/KM
(Mꢀ1 sꢀ1
kcat
(ꢃ10ꢀ3 sꢀ1
KM
(mM)
)
)
Oxaldie-2b
Oxaldie-2b
Oxaldie-3
200
100
5–40
—
0.153
0.071
1.74
0.011
0.0005
7.5
15
86
48
210
49.4
Interestingly, Oxaldie-3 not only performed its catalytic
function at lower concentrations than Oxaldie-2, but
Gly-CONH2
c
it was generally a better catalyst (kcat=86.0ꢃ10ꢀ3 sꢀ1
;
Bu-NH2
Spont.
—
0.013
KM=49.4 mM). Under optimal conditions, the Michaelis
constants were similar for both catalysts, but the turn-
over number was more than 10 times bigger for Oxal-
die-3. The catalytic efficiency (kcat/KM) was 1.74 Mꢀ1
aAll parameters were determined at 25 ꢂC in 50 mM BES (pH 7) con-
taining 10 mM NaCl (150 mM NaCl;1 M NaCl 13). Reaction rates
were measured by monitoring the production of pyruvate using lactate
dehydrogenase and following the conversion of NADH to NAD+
spectrophotometrically at 340 nm.
bOxaldie-2 is NH2LeuAlalysLeuLeuLysAlaLueAlaLysLeuLeuLysLys-
CONH2;Gly-CONH 2 is glycine amide.
cBu-NH2 is butylamine and BES is N,N-bis [2-hydroxyethyl]-2-amino-
ethanesulfonic acid.
s
ꢀ1 for Oxaldie-3 and only 0.153 Mꢀ1 sꢀ1 for Oxaldie-2.
While with simple amines elimination of water from the
high-energy carbinolamine intermediate is the slow step
in the catalysed decarboxylation of oxaloacetate, analy-
sis of the kinetic data for Oxaldie-1 and -2 revealed that

S. E. Taylor et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2631–2635
2635
simple amines, Oxaldie-3 speeded the decarboxylation
of oxaloacetate by more than three orders of magnitude,
which is not insignificant when compared to the 8
orders of magnitude of catalysis observed with aceto-
acetate decarboxylase. Future design efforts must focus
on the production of catalysts with well-defined, stable
tertiary structure and on speeding the actual
decarboxylation following imine formation.
7. Kennan, A. J.;Haridas, V.;Severin, K.;Lee, D. H.;Gha-
dirir, M. R. J. Am. Chem. Soc. 2001, 123, 1797.
8. Corey, M. J.;Corey, E. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 11428.
9. Hamilton, G. A.;Westheimer, F. H. J. Am. Chem. Soc.
1959, 81, 2277.
10. Hamilton, G. A.;Westheimer, F. H. J. Am. Chem. Soc.
1959, 81, 6332.
11. Westheimer, F. H. Tetrahedron 1995, 51, 3.
12. Allemann, R. K. Thesis ETH No. 8804, 1989.
13. Johnsson, K. Thesis ETH No. 9974, 1992.
14. Blondelle, S. E.;Pe rez-Paya, E.;Dooley, C. T.;Pinilla, C.;
Houghten, R. A. Trends Analyt. Chem. 1995, 14, 83.
15. Perez-Paya, E.;Houghten, R. A.;Blondelle, S. E. J. Biol.
Chem. 1996, 271, 4120.
16. Wood, S. P.;Pitts, J. E.;Blundell, T. L.;Tickle, I. J.;Jen-
kins, J. A. Eur. J. Biochem. 1977, 78, 119.
17. Blundell, T. L.;Pitts, J. E.;Tickle, I. J.;Wood, S. P.;Wu,
C.-W. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4175.
18. Glover, I. D.;Barlow, D. J.;Pitts, J. E.;Tickle, I. J.;
Blundell, T. L.;Tatemoto, K.;Kimmel, J. R.;Wollmer, A.;
Strassburger, W.;Zhang, Y.-S. Eur. J. Biochem. 1985, 142,
379.
Acknowledgements
This work was supported by a BBSRC grant (RKA)
and by the School of Chemistry. We thank Dr. Peter R.
Ashton for help with mass spectrometry, Dr. Mark
Bycroft (MRC-Centre for Protein Engineering, Cam-
bridge) for help with the interpretation of the NMR
data, Dr. Lesley A. Tannahill for critical reading of the
manuscript and the members of our laboratory for
discussions.
19. Li, X.;Sutcliffe, M. J.;Schwartz, T. W.;Dobson, C. M.
Biochemistry 1992, 31, 1245.
References and Notes
20. Chang, P. J.;Noelken, M. E.;Kimmel, J. R. Biochemistry
1980, 19, 1844.
21. Noelken, M. E.;Chang, P. J.;Kimmel, J. R. Biochemistry
1980, 19, 1838.
22. Greenfield, N.;Fasman, G. D. Biochemistry 1969, 8, 4108.
23. Wuthrich, K. NMR of Proteins and Nucleic Acids;Wiley:
¨
New York, 1986.
24. Haezebrouk, P.;Joniau, M.;Vandeal, H.;Hooke, S. D.;
Woodruff, N. D.;Dobson, C. M. J. Mol. Biol. 1995, 246, 382.
25. Ptitsyn, O. B. Trends Biochem. Sci. 1995, 20, 376.
1. Kirby, A. J. Acta Chem. Scand. 1996, 50, 203.
2. MacBeath, G.;Hilvert, D. Chem. Biol. 1996, 3, 433.
3. Schultz, P. G.;Lerner, R. A. Science 1995, 269, 1835.
4. Bartel, D. P. In The RNA World (2nd Ed.).;Gesteland,
R. F., Chech, T. R., Atkins, J. F., Eds.
5. Johnson, K.;Allemann, R. K.;Widmer, H.;Benner, S. A.
Nature 1993, 365, 530.
6. Broo, K. S.;Brive, L.;Ahlberg, P.;Baltzer, L.
Chem. Soc. 1997, 119, 11362.
J. Am.