Substrate recognition and saturation kinetics in de novo designed histidine-based four-helix bundle catalysts

Article abstract of DOI:10.1021/ja980682e

Designed four-helix bundle proteins with reactive sites based on the cooperativity of HisH⁺-His pairs in helical sequences catalyze acyl-transfer reactions of p-nitrophenyl esters with large rate enhancements. The function of the HisH⁺

Full text of DOI:10.1021/ja980682e

J. Am. Chem. Soc. 1998, 120, 10287-10295
10287
Substrate Recognition and Saturation Kinetics in de Novo Designed
Histidine-Based Four-Helix Bundle Catalysts
Kerstin S. Broo, Helena Nilsson, Jonas Nilsson, and Lars Baltzer*
Contribution from the Department of Chemistry, G o¨ teborg UniVersity, 41296 G o¨ teborg, Sweden
ReceiVed March 2, 1998
+
Abstract: Designed four-helix bundle proteins with reactive sites based on the cooperativity of HisH -His
pairs in helical sequences catalyze acyl-transfer reactions of p-nitrophenyl esters with large rate enhancements.
+
The function of the HisH -His site has been expanded by the introduction of flanking residues to provide
recognition of substrate carboxylate and hydrophobic residues. The second-order rate constants for the MN-
4
2 catalyzed hydrolysis of p-nitrophenyl acetate and of mono-p-nitrophenyl fumarate, under conditions of
-
1
-1
-1 -1
excess catalyst over substrate, in aqueous solution at pH 5.1 and 290 K are 0.030 M
s
and 0.027 M s ,
respectively. The reactive site of MN-42 contains only histidine residues. The sequence of MNKR is the
same as that of MN-42 except that one Lys and one Arg residue have been introduced in the adjacent helix
+
-1 -1
-1
to flank the HisH -His site and the resulting second-order rate constants are 0.075 M
s
and 0.135 M
s . MNKR catalyzed hydrolysis of the fumarate follows saturation kinetics with a kcat/KM of 0.17 M s-1
-
1
-1
which is 230 times larger than the second-order rate constant of the 4-methyl imidazole catalyzed reaction.
The second-order rate constants for the JNIII catalyzed hydrolysis of p-nitrophenyl acetate and of p-nitrophenyl
-
1
-1
-1 -1
valerate are 0.007 M
s
and 0.097 M s , respectively, and binding of the aliphatic group increases the
rate constant by more than one order of magnitude. Chiral recognition by de novo designed polypeptides has
been demonstrated for the first time, and the hydrolysis of the p-nitrophenyl ester of D-norleucine has been
catalyzed with a second-order rate constant that is twice as large as that of the L-norleucine ester.
Introduction
efficiency and selectivity exhibited by the natural enzymes and
because of the challenge involved in fine-tuning polypeptide
structures with complex reactive sites for the purpose of
obtaining cooperative catalysis.
The increasing number of reported de novo designed polypep-
tides with well-defined tertiary structures
engineering of folded polypeptides, tailored for specific func-
1-6
suggest that the
Designed polypeptide catalysts with widely different catalytic
mechanisms have now been reported where the relationship
between structure and function is at least partly understood and
that show rate enhancements of approximately 3 orders of
magnitude. The decarboxylation of oxaloacetate has been
accomplished by two 14-residue amphiphilic helical polypep-
7
tions, is timely. The successful introduction of metal and
8,9
10-12
cofactor binding sites and porphyrin binding pockets
and
1
3
the development of peptides capable of forming ion channels
and nanotubes14 are but a few examples of the diversity of
applications that are being approached by the design of
nonnatural amino acid sequences. Much attention is focused
on the design of novel catalysts because of the remarkable
15
tides, oxaldie1 and oxaldie2, that aggregate in solution to form
bundle-like structures. The reaction mechanism proceeds via
covalent enamine formation, and the most efficient catalyst is
oxaldie 1 mainly due to the low pKa of the amino terminal of
the peptide. The peptides bind the dianionic substrate oxalo-
acetate with KM values in the millimolar range.
*
E-mail Lars.Baltzer@oc.chalmers.se.
1) Hill, R. B.; DeGrado, W. F. J. Am. Chem. Soc. 1998, 117, 1138-
145.
2) Dolphin, G. T.; Brive, L.; Johansson, G.; Baltzer, L. J. Am. Chem.
Soc. 1996, 118, 11297-11298.
3) Brive, L.; Dolphin, G. T.; Baltzer, L. J. Am. Chem. Soc. 1997, 119,
598-9607.
4) Struthers, M. D.; Cheng, R. P.; Imperiali, B. Science 1996, 271, 342-
45.
(
1
(
(
A synthetic peptide ligase has been designed which catalyses
8
3
3
ligation reactions with rate enhancements of 4 × 10 over that
(
1
6
of the background reaction.
The catalyst is a 33-residue
(
(
(
5) Betz, S. F.; DeGrado, W. F. Biochemistry 1996, 35, 6955-6962.
amphiphilic helix that binds the reacting peptide fragments to
increase the effective concentrations of electrophile and nu-
cleophile with dissociation constants in the micromolar range.
The reaction mechanism includes the attack of the side chain
of the amino terminal Cys of the nucleophile peptide fragment
on the activated thioester of the carboxy terminal of the
electrophile peptide fragment, and the resulting thioester rear-
6) Bassil, D. I.; Mayo, S. L. Science 1997, 278, 82-87.
7) Walkup, G. K.; Imperiali, B. J. Am. Chem. Soc. 1996, 118,3053-
3
054.
(
8) Roy, R. S.; Imperiali, B. Tetrahedron Lett. 1996, 37,2129-2132.
(9) Mihara, H.; Tomizaki, K.; Nishino, N.; Fujimoto, T. Chem. Lett. 1993,
1
533-1536.
10) Rabanal, F.; DeGrado, W. F.; Dutton, P. L. J. Am. Chem. Soc. 1996,
18, 473-474.
11) Nastri, F.; Lombardi, A.; Morelli, G.; Maglio, O.; D’Auria, G.;
Pedone, C.; Pavone, V. Chem. Eur. J. 1997, 3, 340-349.
12) Mihara, H.; Tomizaki, K.; Fujimoto, T.; Sakamoto, S.; Aoyagi, H.;
Nishino, N. Chem. Lett. 1996, 3, 187-188.
13) Lear, J. D.; Schneider, J. P.; Kienker, P. K.; DeGrado, W. F. J. Am.
Chem. Soc. 1997, 119, 3212-3217.
14) Hartgerink, J. D.; Granja, J. R.; Milligan, R. A.; Ghadiri, M. R. J.
Am. Chem. Soc. 1996, 118, 43-50.
(
1
(
17
ranges to yield the final amide product.
(
(15) Johnsson, K.; Allemann, R. K.; Widmer, H.; Benner, S. A. Nature
1993, 365, 530-532.
(
(16) Severin, K.; Lee, H. D.; Kennan, A. J.; Ghadiri, M. R. Nature 1997,
389, 706-709.
(
(17) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science
1994, 266, 776-779.
S0002-7863(98)00682-9 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/25/1998

10288 J. Am. Chem. Soc., Vol. 120, No. 40, 1998
Broo et al.
Figure 1. Modeled structure of KO-42. The histidine side chains are
shown with their corresponding pK values. KO-42 is a symmetric
a
dimer, but only the monomer is shown for reasons of clarity of
presentation. The i, i+4, i+8 configuration ensures that the geometric
relationship between each His pair is similar.
KO-42, a designed antiparallel helix-loop-helix hairpin dimer
with 42 residues in the monomer sequence, catalyses acyl-
transfer reactions of p-nitrophenyl esters. The reactive site
contains six His residues (Figure 1),18 and the reaction mech-
anism proceeds via rate-limiting formation of an acyl-imidazoyl
intermediate followed by attack on the acyl intermediate by the
appropriate nucleophile to form the reaction products. The
second-order rate constant for the hydrolysis of mono-p-
nitrophenyl fumarate at pH 4.1 is 1140 times larger than that
of the 4-methylimidazole catalyzed reaction. The rate enhance-
ment at low pH originates from cooperative nucleophilic and
general acid catalysis by unprotonated nucleophilic histidines
Figure 2. Amino acid sequences of KO-42 and related peptides listed
in the one-letter code and Nle for norleucine. The sequences of related
peptides are the same as that of KO-42 except for positions 8, 11, 15,
1
9
(i) flanked by protonated histidines (i-4) in a helical segment.
1
9, 26, 30, and 34 where the amino acid residues presented have
The reported catalysts demonstrate that it has been possible
replaced those of the original sequence. His residues are shown in bold.
The second-order rate constants are those for the hydrolysis of I and
II in aqueous solution at pH 5.1 and 290 K catalyzed by the listed
peptide catalysts. The corresponding rate constants for the 4-MeIm
to accomplish by rational design large catalytic rate enhance-
ments for chemical reactions of great diversity. The engineering
of reactive sites capable of substrate discrimination is thus the
next step in catalyst design. We now wish to report that the
recognition of carboxylate and hydrophobic residues of p-
nitrophenyl esters has been accomplished in catalysts derived
-1
-1
-1 -1
catalyzed reactions are 0.000 74 M
s
(I) and 0.000 72 M
s
(II).
amphiphilic helices, and the hydrophobic core was therefore
designed in a shape-complementary way from leucines, nor-
leucines, phenylalanines, a valine, and an isoleucine residue.
The primary amino acid sequence was varied extensively to
+
from KO-42, where the HisH -His site has been expanded by
the introduction of flanking Arg and Lys residues. The results
demonstrate the feasibility of engineering selective reactive sites
in the further development of designed polypeptide catalysts.
1
simplify the assignment of the H NMR spectrum, and two NOE
reporter groups, the phenyl rings of Phe-35 and Phe-38, were
incorporated at the lower part of helix II to facilitate identifica-
tion of hairpin formation and to induce chemical shift disper-
Results
Design and Tertiary Structures of Four-Helix Bundle
Catalysts. The amino acid sequences of KO-42 and the
polypeptides reported here were based on that of the previously
reported de novo designed template polypeptide SA-42 that folds
20
sion. The full charges of the amino and the carboxy terminals
of SA-42 were neutralized in KO-42 by acetylation and
amidation. In the design of the His based acyl-transfer catalyst
KO-42, Ala-11, Lys-19, Gln-26, Gln-30 and Ala-34 of SA-42
into a helix-loop-helix motif and dimerizes to form a four-helix
1
8
bundle.20 In short, the choice of amino acid residues was based
were replaced by five histidines (Figure 2).
21,22
SA-42 and KO-42 have been characterized extensively by
H NMR and CD spectroscopy and equilibrium sedimentation
on their propensities for secondary structure formation,
residues were also introduced that could stabilize the folded
and
1
1
8,24
ultracentrifugation.
The CD spectra of SA-42 and KO-42
structure by capping, stabilization of the helical dipoles, and
intrahelical salt bridge formation.23 The driving forces for
both show the negative maxima at 208 and 222 nm that are
characteristic of helical peptides. The mean residue ellipticities
hairpin formation are the hydrophobic interactions between
2
-1
at 222 nm is -24 000 ( 1000 deg cm dmol for KO-42 and
(
18) Broo, K. S.; Brive, L.; Ahlberg, P.; Baltzer, L. J. Am. Chem. Soc.
997, 119, 11362-11372.
19) Broo, K. S.; Nilsson, H.; Nilsson, J.; Flodberg, A.; Baltzer, L.; J.
Am. Chem. Soc. 1998, 120, 4063-4068.
20) Olofsson, S.; Johansson, G.; Baltzer, L. J. Chem. Soc., Perkin Trans.
2
-1
-
25 000 ( 1000 deg cm dmol for SA-42, corresponding to
1
25,26
helical contents of 60-70%.
The magnitudes of these mean
(
residue ellipticities are among the highest of those reported for
(
2
1995, 2047-2056.
(24) Olofsson, S.; Baltzer, L. Folding Des. 1996, 1, 347-356.
(25) Broo, K. S.; Sott, R.; Brive, L.; Baltzer, L. Folding Des. 1998, 3,
303-312.
(26) Chen, Y.-H.; Yang, J. T.; Chau, K. H. Biochemistry, 1974, 13,
3350-3359.
(
(
(
21) Chou, P. Y.; Fasman, G. D. Biochemistry 1974, 13, 211-222.
22) Chou, P. Y.; Fasman, G. D. Biochemistry 1974, 13, 222-244.
23) Bryson, J. W.; Betz, S. F.; Lu, H. S.; Suich, D. J.; Zhou, H. X.;
O’Neil, K. T.; DeGrado, W. F. Science 1995, 270, 935-941.

Designed Four-Helix Bundle Catalysts
J. Am. Chem. Soc., Vol. 120, No. 40, 1998 10289
Scheme 1
1
helical polypeptides of this size. The H NMR spectra of both
JNIIRR that contains His-30 and His-34, two Arg were
introduced in positions 11 and 15. MN-42 and JNII were used
as reference peptides, and a negative reference, MNEE with
two Glu in positions 30 and 34 was also synthesized.
peptides have been assigned, and their secondary structures have
been identified by medium range NOEs and RH chemical shift
deviations from random coil values. Although an atomic
resolution structure is not yet available, equilibrium sedimenta-
tion ultracentrifugation and diagnostic long-range NOEs show
that both KO-42 and SA-42-fold into antiparallel hairpin dimers.
The amino acid sequences of the peptides reported here differ
from those of the sequences of SA-42 or KO-42 by two to four
residues (Figure 2). The CD spectra of the polypeptides show
negative maxima at 208 and 222 nm, and the measured mean
residue ellipticities at 222 nm range from -18 000 to -25 000
+
The Reactivity of HisH -His Sites Flanked by Arg and
Lys. The second-order rate constants for the peptide catalyzed
hydrolysis of I were determined under conditions of excess
catalyst over substrate in aqueous solution at pH 5.1 and 290
K. Typical peptide concentrations were 0.25-0.4 mM with a
concentration of I of 0.13 mM. Excellent first-order kinetics
were observed, and as long as the concentration of substrate is
smaller than the dissociation constant for the catalyst-substrate
complex, i.e., [S] , KS, pseudo-first-order kinetics will be
observed also for catalysts that show saturation kinetics. The
second-order rate constants were determined by measuring the
slope of the plot of the pseudo-first-order rate constant versus
2
-1
deg cm dmol . On the basis of the similarity of the primary
sequences and the measured mean residue ellipticities, the
polypeptides presented here are assumed to fold into antiparallel
hairpin helix-loop-helix dimers, although full structural analysis
of each one has not been carried out.
[I].
+
Reactive Sites Based on the HisH -His Amino Acid Pair.
The introduction of flanking Arg and Lys residues in peptides
that contain reactive His sites in the adjoining helix increased
their catalytic efficiency (Figure 2). The second-order rate
In the pursuit of the basic building block for the construction
+
of His-based catalysts, the cooperative two-residue HisH -His
site was identified.19 As a result of the identification and
incorporation of the reactive entities of KO-42 in separate
peptides, the second-order rate constants of each one could be
measured (Figure 2). It was deduced that the reactivities of
-
1
constants for the MNKR catalyzed hydrolysis of I is 0.135 M
-
1
s , a factor of 5 larger than that of the MN-42 catalyzed
reaction, which is 0.027 M s , and the JNIIRR catalyzed
reaction is 0.105 M s , which is almost twice as large as
-
1
-1
-
1
-1
+
the HisH -His sites are mainly controlled by the pKa’s of the
-1 -1
participating histidines.19 The pKa values of the peptides
0.055 M
s , the rate constant for the JNII catalyzed reaction.
Recognition of a Negatively Charged Substrate and
Binding of the Developing Oxyanion in the Transition State
Preceding the Acyl Intermediate. To probe the modes of
interaction, p-nitrophenyl acetate (II), which does not contain
the negatively charged carboxylate residue of I, was used as a
substrate (Scheme 1) because, in the hydrolysis of II, any
observed rate enhancement upon introduction of positively
charged side chains are most likely due to interactions with the
developing oxyanion in the transition state. The second-order
rate constant for the hydrolysis of II increased upon the
introduction of flanking Arg and Lys in comparison with those
of the reactions catalyzed by peptides that only contained
described here have been determined, and they can be affected
by only a few tenths of a pKa unit by residues in the neighboring
2
7
helix which induces only minor differences in reactivity.
However, substrate binding was also suggested from the kinetic
measurements. The measured second-order rate constants for
the reactions catalyzed by JN-42 and by MN-42, that contain
the His residues of helix I and helix II of KO-42, respectively,
do not add up to the second-order rate constant of KO-42. The
second-order rate constant for the JN-42 catalyzed hydrolysis
of mono-p-nitrophenyl fumarate (I) is 0.065 M-1
s
-1
in aqueous
-
1
solution at pH 5.1 and 290 K, and that of MN-42 is 0.027 M
-1 19
-1 -1
s , whereas that of KO-42 is 0.31 M
s
under the same
+
conditions. The lack of additivity suggested that protonated
histidines could contribute in a cooperative fashion to transition-
state binding in reactions catalyzed by His residues in the
adjacent helix.
A small peptide library, based on the peptide MN-42, was
therefore designed, where positively charged arginines and
lysines were introduced in positions 30 and 34 in helix II for
the purpose of stabilizing the developing oxyanion in the
transition state preceding the acyl intermediate. The peptides
with all possible combinations of Arg, Lys, and Gln residues
in positions 30 and 34 in helix II were synthesized, and one
peptide was also designed on the basis of the reactivity of the
most efficient two-residue site in helix II, that of JNII. In
HisH -His sites. Consequently, the developing oxyanion was
stabilized by the side chains of the flanking residues.
KO-42, MN-42, and JNII did not show any discrimination
between I and II; the second-order rate constants were the same
for both substrates. However, upon introduction of Arg or Lys
in helix II, the peptides catalyzed the hydrolysis of I with larger
rate constants than they catalyzed the hydrolysis of II, showing
that the fumaryl residue is bound by the introduced residues.
The observed second-order rate constants for MNKR catalyzed
-1 -1
-1 -1
hydrolysis of I and II are 0.135 M
s
and 0.075 M s ,
respectively. Interestingly, reversing the positions of the Arg
and Lys residues in the sequence of MNKR gives rise to a
peptide that is incapable of recognizing the fumaryl residue since
the second-order rate constants for MNRK catalyzed hydrolysis
(
27) Engel, M.; Williams, R. W.; Erickson, B. W. Biochemistry 1991,
-
1
-1
-1 -1
3
0, 3161-3169.
of I and II are the same, 0.110 M
s
and 0.106 M s .

10290 J. Am. Chem. Soc., Vol. 120, No. 40, 1998
Broo et al.
Table 1. Second-Order Rate Constants and Ratios of
Second-Order Rate Constants for Peptide and 4-MeIm Catalyzed
Hydrolysis of III at pH 5.1 and 290 K
(III) M- s-1
1
k
(III)/k
(I)
k
peptide
k
2
2
2
2
2
(III)/k (II)
KO-42
JNI
JNII
JNIII
JNIIRR
0.535
0.027
0.186
0.097
0.144
0.00070
1.7
2.8
3.4
14.0
1.4
0.95
1.8
3.9
2.1
4
-MeIm
0.95
2 2
Figure 3. Ratio k (KO-42)/k (4-MeIm) at pH 5.1 and 290K as a
function of the number of methylene and methyl groups of the fatty
acid ester substrate.
Replacing Lys-30 by an ornithine (Orn) that has a shorter
side chain by one methylene group gives rise to a peptide that
does not discriminate between I and II. The rate constants are
again the same within experimental accuracy, and the interac-
tions between the catalyst and the fumaryl carboxylate anion
substrate are therefore specific. The second-order rate constant
-
1
-1
for the MNOrnR catalyzed hydrolysis of I is 0.092 M
s
and that of the MNOrnR catalyzed hydrolysis of II is 0.081
Figure 4. Rates of hydrolysis of fatty acid esters at pH 5.1 and 290
K, catalyzed by MN-42 (circle) and JN-42 (square), relative to those
of p-nitrophenyl acetate. The relative rates have been determined by
using the same stock peptide solution for each substrate.
-1 -1
M
s . The negative reference peptide MNEE does, in fact,
show a small decrease in the second-order rate constants
compared to that of MN-42, suggesting that, in this case, the
fumaryl carboxylate anion is repelled. The rate constant of
MNEE catalyzed hydrolysis of I is even smaller than that of
II, in agreement with the suggestion of repulsion of the
carboxylate anion by the negatively charged Glu’s, although in
this case, the differences are probably within experimental error.
To probe the geometry and specificity of the reactive site,
the second-order rate constant of JNIIRR catalyzed hydrolysis
of I was compared to that of JNIIK8R11 and JNIIR8K11 (Figure
increased from 0.29 M-1 s-1 to 0.54 M-1 s-1 as a function of
chain length in the series from p-nitrophenyl acetate (II) to
p-nitrophenyl valerate (III) at pH 5.1 and 290 K (Table 1).
Binding of the hydrophobic alkyl chain of the esters differs
between the catalysts MN-42 and JN-42. The reactive site of
JN-42 showed substantial rate enhancements, whereas the rate
constants for the MN-42 catalyzed reactions were not affected
by the structure of the substrate (Figure 4), suggesting the
presence of specific alkyl chain binding sites.
2
). All three peptides have the same reactive histidine entity,
+
HisH -30 -His-34 and two flanking positively charged residues;
however, in the sequence of JNIIRR the flanking residues are
Arg-11 and Arg-15, whereas in the latter two sequences the
flanking residues are Arg and Lys located in positions 8 and
To identify the hydrophobic binding site for the JN-42
catalyzed reactions, the second-order rate constants for the
hydrolysis of p-nitrophenyl valerate (III) catalyzed by JNI, JNII,
and JNIII were measured (Table 1) and compared to those for
the hydrolysis of I that does not have a hydrophobic residue.
The second-order rate constants for JNII catalyzed hydrolysis
of I and II do not differ by more than 10%, but the second-
order rate constants for JNIII catalyzed hydrolysis of III and I
1
1, i.e. they have been moved one turn of the helix toward the
N terminus of helix I. The second-order rate constants for the
hydrolysis of I show that there is binding of neither the
developing oxyanion nor the fumaryl carboxylate in the peptides
JNIIK8R11 and JNIIR8K11 as the rate constants are approximately
the same as, or smaller than, that of JNII which has no flanking
residues.
Recognition of Hydrophobic Substrates. The helix-loop-
helix hairpin motif is formed mainly due to the hydrophobic
interactions between amphiphilic helices. To probe whether the
hydrophobic core would be partly exposed and available for
hydrophobic binding of substrates, the rate constants for
hydrolysis of a series of straight chain fatty acid p-nitrophenyl
esters were measured (Scheme 1). The p-nitrophenyl acetate,
propionate, butyrate, and valerate were studied, whereas p-
nitrophenyl caproate was only sparsely soluble under the
reaction conditions and did not provide good kinetic measure-
ments due to precipitation.
-1 -1
-1 -1
are 0.097 M
s
and 0.007 M s , respectively. At pH
5
.1, the catalytic activity of JNIII is due solely to nucleophilic
18
attack by His-34, and the rate enhancement is therefore due
to the binding of the hydrophobic alkyl residue by the peptide.
The rate enhancements by JNI and JNII are smaller, JNI showed
a ratio k (III)/k (I) of 2.8, and JNII showed a ratio of k (III)/
2
2
2
k2 (I) of 3.4. The existence of a hydrophobic binding pocket
in close proximity to His-34 is compatible with the fact that
the side chain of His-34 appears to be partly buried in the
hydrophobic core of the folded four-helix bundle. The proxim-
ity of His-34 to the hydrophobic core was confirmed by NOE
connectivities between the â protons of His-34 and the methyl
groups of Leu-12, Leu-31, and Nle’s that were identified in the
1
H NMR NOESY spectrum of JNIII (Figure 5). The reason
The rate of KO-42 catalyzed ester hydrolysis, rather than the
logarithmic rate, increased almost linearly with the number of
methylene groups in the alkyl side chain (Figure 3), which
means that the free energy gain in increasing the chain length
tends toward a maximum value. The second-order rate constants
that the hydrophobic binding pocket is not exploited to its full
potential in the JNII catalyzed reaction may be that the
constraints imposed on the structure of the transition state by
the geometry of the His reactive site prevent the alkyl side chain
from binding in an optimal way.

Designed Four-Helix Bundle Catalysts
J. Am. Chem. Soc., Vol. 120, No. 40, 1998 10291
Table 2. Rate Constant Ratios of Peptide Catalyzed Hydrolysis of
L-IV and D-IV at pH 5.1 and 290 K
peptide
k
2
2
(D-IV)/k (L-IV)
KO-42
MN-42
JN-42
MNKK
MNRR
MNKR
MNRK
MNKQ
MNRQ
MNQR
JNI
1.4
1.1
1.3
2.0
1.7
1.7
2.0
1.7
1.7
1.4
1.4
1.3
0.8
JNII
JNIII
1
Figure 5. Part of the H NMR NOESY spectrum of JNIII in 10 vol %
attempted. However, the low reactivity of MNEE and the high
inherent reactivity of the esters limited the possibilities for
observing chiral discrimination.
TFE-d
â-protons of His-34 at 3.18 ppm and the methyl groups of Leu-12 at
.84 ppm and those of Leu-31 at 0.71 ppm. The cross-peak at 0.78
3 2 2
in H O/D O (90:10) showing NOE connectivities between
0
Substrate Binding and Saturation Kinetics. Natural en-
zymes bind their substrates and form enzyme-substrate com-
plexes prior to the actual chemical transformations. In KO-42
catalyzed hydrolysis of I, no signs of substrate recognition or
ppm is due to NOE connectivities between the â-protons of His-34
and the methyl groups of Ile-9 and or one or more of the norleucines,
all of which overlap. The â-protons of His-26 at 3.42 ppm shows only
cross-peaks to methyl groups of Val-22 at 0.96 and 1.06 ppm.
1
8
saturation kinetics were observed. The number of possible
reaction pathways in KO-42 catalysis is, however, large, and
by designing more well-defined reactive sites, substrate recogni-
tion was achieved for MNKR, MNRR, and JNIIRR, implying
that a smaller number of pathways were available in these
reactions. The stronger binding of I in the transition state was
also expected to lead to stronger binding in the ground state
and perhaps give rise to a peptide-substrate complex along the
reaction pathway which could be observed by the onset of
saturation kinetics.
The possibility of making use of cooperative oxyanion and
hydrophobic binding was investigated in the case of the JNIIRR
catalyzed hydrolysis of III and compared to that of JNII that
has no flanking arginines. The second-order rate constant at
-
1
-1
pH 5.1 and 290 K was 0.14 M s , whereas that of the JNII
-
1
-1
catalyzed reaction was 0.19 M
s
(Table 1), showing that
cooperative hydrophobic and charge-charge binding could not
be achieved for JNIIRR.
Chiral Recognition. The introduction of positively charged
amino acid residues in the helix adjacent to the one that carries
To detect a possible catalyst-substrate complex, a reaction
is usually studied under conditions of excess substrate over
catalyst, and the initial rates are plotted as a function of substrate
concentration. The commonly used form of the Michaelis-
+
the HisH -His site gives rise to recognition of a negatively
charged substrate, and the peptides also have a demonstrated
capacity for binding aliphatic residues. In combination with
+
o
Menten equation is derived under the assumption that [S] is
the stereochemistry of the HisH -His site, where the His with
.
[E]o so that [ES] can be neglected in comparison with [S]
the highest number in the sequence is the nucleophile and the
19
and [S] be used as a measure of [S]. This is usually valid for
one with the lowest number is the general acid catalyst, chiral
recognition will be possible if these interactions can be exploited
simultaneously in binding a substrate. The p-nitrophenyl esters
of the enantiomers of Nle (L-IV, D-IV, Scheme 1) were used to
probe chiral discrimination since they contain one alkyl and
one charged substituent. An amino acid ester substrate with a
negatively charged side chain, such as the Glu ester, would be
preferable but is unsuitable due to the efficient intramolecularly
catalyzed reaction. The positively charged amino acids in the
adjacent helix could, however, potentially repel one enantiomer
of the Nle substrates more readily than the other.
o
natural enzymes where the catalytic efficiencies are high and
the rates can be measured for small concentrations of catalyst.
However, in the case of the catalysts described here, peptide
concentrations on the order of 0.3-0.6 mM have to be used,
while the substrate concentration due to solubility problems
cannot be larger than approximately 2 mM and therefore in the
range from 0.05 to 2 mM. The concentration of catalyst is
clearly not negligible in comparison with that of the substrate.
It is, however, possible to use the Michaelis-Menten equation
without the simplifying assumption that [S] . [E] which leads
to an expression that has a square root dependence on the
concentration of substrate, eq 1.
The second-order rate constants for the 4-MeIm catalyzed
hydrolysis of the D- and L-Nle esters were found to be equal as
were those of the background reactions in aqueous solution at
pH 5.1 and 290K (Table 2). But, the rate constants for the
reactions catalyzed by peptides with flanking positively charged
amino acid residues were larger for the D-amino acid ester than
for the L-amino acid ester. Peptide catalysts that contained no
flanking residues did not discriminate between enatiomeric
esters, and the difference between enantiomeric substrates was
less for the catalysts with only one flanking charged residue
than for the catalysts with two.
V ) k [PS] ) k (([P ] + [S ] + K )/2 -
cat
cat
0
0
M
2
x([P ] + [S ] + K ) /4 - [P ][S ]) (1)
0
0
M
0
0
Here, P is the peptide, S is the substrate and PS the peptide
substrate complex. The alternative solution to the second-order
equation has been discarded because it can be shown to lead to
concentrations of the PS complex that are close to K even for
M
very small concentrations of peptide and substrate. The initial
More pronounced chiral interactions would be expected from
catalysis of reactions where the substrate side chains are bound
rather than repelled. The catalysis of hydrolysis of the p-
nitrophenyl esters of L- and D-Lys by MNEE was therefore
rates of the hydrolysis of I, was measured for the peptide catalyst
MNKR, and the results were fitted to eq 1 to obtain kcat and
KM. The hydrolysis of I catalyzed by MNKR showed saturation
-
1
kinetics, with a KM of 0.001 M and a kcat of 0.000 17 s , and

10292 J. Am. Chem. Soc., Vol. 120, No. 40, 1998
Broo et al.
Figure 6. Plot of initial rate (V) of MNKR catalyzed hydrolysis of I
at pH 5.1 and 290 K vs [I] at a peptide concentration of 0.3 mM and
substrate concentrations of 0.05-2 mM. The solid line shows the best
fit of the experimental data to eq 1, the Michaelis-Menten equation
derived without the simplifying assumption that [S] . [E].
thus a kcat/KM of 0.17 M-1 s-1(Figure 6). This rate constant
may be compared to the second-order rate constant for the
4
-1 -1
4
-MeIm catalyzed hydrolysis of I which is 7.4 10-
M
s ,
Figure 7. Part of the amide and methyl regions of the 1H NMR
spectrum of MNRK at pH 5.1 and 293 K in 100 mM sodium acetate-
d buffer before (lower trace) and immediately after (upper traces) the
4
addition of I. The concentration of peptide was approximately 0.4 mM
and that of I, 1.5 mM. The spectral changes upon addition of I are
clearly seen in the methyl region of MNRK, and the lack of free
p-nitrophenol resonances shows that the reaction has not proceeded to
any measurable degree.
and the ratio of (kcat/KM)/k2 (4-MeIm) is 230 in aqueous solution
at pH 5.1 and 290K. The peptide concentrations were 0.3 mM,
and the substrate concentrations ranged from 0.05 mM to 2 mM.
Although KM may formally reflect the build-up of an acyl
intermediate, no observation of such an intermediate has been
possible under reaction conditions despite several attempts by
NMR and UV spectroscopy and by mass spectrometry.
Direct evidence for the formation of a peptide-substrate
1
1
complex was achieved by H NMR spectroscopy. The H NMR
spectrum of the amide and methyl regions of MNRK are shown
in Figure 7 before and after addition of I. MNRK-substrate
interactions are clearly visible in the amide and methyl regions
under conditions where practically no chemical reaction has
taken place, as shown by the virtual absence of free p-
and the acidity of the protonated residue and the reactivity of
+
the HisH -His site, relative to that of 4-MeIm, at a pH below
the pKa of the histidines, therefore increases as the pKa of the
His residues decrease. An increase in reactivity at a given pH
must therefore come from transition-state binding by side chains
+
of amino acid residues that flank the HisH -His site.
1
1
nitrophenol in the H NMR spectrum. The H NMR spectrum
of the peptide MNII does not show the same changes, and the
interaction is therefore peptide-specific and not, for example,
due to the effect of the added acetonitrile. When the reaction
proceeds, the methyl group resonances of MNRK are broadened
further, probably due to complexation between the peptide and
the reaction products. This can be compared to results from
KO-42 catalyzed hydrolysis of I where the addition of 2 mM
fumaric acid slows down the rate of peptide catalysis by 20%
The cooperative catalysis of acyl-transfer reactions by two
His residues in a helix imposes stereochemical limitations on
the structure of the transition state. The developing covalent
bond between the nucleophile and the carbonyl carbon and the
hydrogen bond between the general acid and one or both of the
ester oxygens are expected to be roughly aligned with the helical
axis. In a four-helix bundle motif, the side chains of residues
in the adjoining helix are therefore suitable targets for the
introduction of functionality that can bind the acyl or the leaving
groups, and increased reactivity provides an assay by which
such interactions can be identified. The distance between the
centers of two helices in a well-packed helix-loop-helix motif
is approximately 10 Å, and in a disordered one, it is longer. To
achieve interactions between residues in separate helices, only
the residues that are oriented toward the adjoining helix can be
used. In fact, the observation of such interactions would provide
a measure of the distance between these residues.
1
8
at pH 5.85.
Discussion
The four-helix bundle motif is the most widely studied in de
novo design, both from a perspective of elucidating what
1-3
sequences fold into well-defined tertiary structures, and from
the perspective of functionalization.8
-14
In the designed four-
helix bundle catalyst KO-42, surface exposed histidines catalyze
hydrolysis and transesterification reactions of p-nitrophenyl
esters with second-order rate constants that are 3 orders of
The MN-42 catalyzed hydrolysis of II in aqueous solution
at pH 5.1 proceeds with a second-order rate constant of 0.030
-
1
-1
magnitude larger than those of the corresponding 4-MeIm
M
s , Figure 2. MN-42 has three His residues in helix I,
catalyzed reactions.1
8,19
The reactive entity consists of one
but no residues in helix II that are capable of binding.
Introduction of Arg or Lys or combinations of the two in
positions 30 and 34 invariably increase the rate constant, most
strikingly in MNRK where Arg-30 and Lys-34 have been
unprotonated and one protonated His residue in positions i, i-4
+
in a helical conformation, and the reactivity of the HisH -His
pair can be calculated from the reactivity and pKa of 4-MeIm
and from the pKa values of the histidines using the Brønsted
introduced and where the second-order rate constant is 0.106
1
9
-1 -1
equation with the coefficients â ) 0.8 and R ) 0.6.
It
M
s
and thus larger than that of MN-42 by a factor of 3.5.
increases with the nucleophilicity of the unprotonated residue,
The introduction of the two binding residues thus reduces the

Designed Four-Helix Bundle Catalysts
J. Am. Chem. Soc., Vol. 120, No. 40, 1998 10293
several kcal/mol.28 The exploitation of hydrophobic recognition
by the four-helix bundle motif is therefore a key ingredient in
the development of efficient designed catalysts. The observed
rate enhancement in the JNIII catalyzed hydrolysis of fatty acid
esters upon lengthening the aliphatic chain by three carbon atoms
is larger than one order of magnitude and corresponds to 1.6
kcal/mol of excess binding energy in the transition state. The
rate enhancement is an approximately linear function of chain
length which shows that the free energy gain per methylene
group tends toward a maximum value. This is not unexpected
since the differentiation between transition-state binding and
ground-state binding should decrease with the distance from
the reactive site. Nevertheless, hydrophobic binding appears
to be a functional design idea in the engineering of surface
catalysts.
Figure 8. Proposed structure of the transition state of MNKR catalyzed
hydrolysis of I based on the experimental results.
free energy of activation by 0.75 kcal/mol. Since II has no
ionizable groups, we conclude that the positively charged
binding residues bind the developing oxyanion as the substrate
approaches the tetrahedral intermediate.
The observed rate enhancements k2 (III)/k2 (I) are less for
JNI and JNII than for the JNIII catalyzed reactions, and the
+
geometric constraints from the HisH -His pair are apparently
The second-order rate constants for the peptide catalyzed
hydrolysis of I are invariably larger than or equal to those of
II, and the rate constant ratios approach a factor of 2 in the
cases of MNKR and MNRR, two of the more efficient catalysts,
Figure 2. In the MNKR catalyzed reaction, the extra transition-
state stabilization amounts to 0.4 kcal/mol, a value that is close
to what is observed for surface-exposed salt bridges in folded
polypeptides and therefore what can be expected from a well
positioned binding residue. The observation of the reduction
in the free energy of activation by almost the full amount of a
salt bridge suggests that the geometry of the transition state fits
better in the reactive site than that of the substrate. However,
the fact that saturation kinetics are obtained and therefore that
the substrate is bound by MNKR suggests that the total binding
energy of the transition state is larger. The binding of the
carboxylate residue of I by MNKR appears to be highly specific
because, when Lys-30 is replaced by Orn-30, the discrimination
between I and II disappears. A proposed transition-state
structure based on the results is shown in Figure 8. The
discrimination may also prove to be larger for longer amino
acid side chains in the binding residues of helix II.
quite strong. With only His-34 as the nucleophile and no
binding interactions from His-30, the hydrophobic effect can
be more efficiently utilized probably because the geometry of
the transition state can adjust to allow a better fit of the
hydrophobic residues in the binding site. The observation of
more efficient hydrophobic binding in the His-34 catalyzed
reaction is compatible with the fact that His-34 is close to the
hydrophobic residues as shown by the identification of NOE
connectivities between His-34 and Ile-9, Leu-12, and Leu-31
1
in the H NMR NOESY spectrum, Figure 5.
+
The geometry of the HisH -His site and the observation of
recognition of carboxylate groups and of hydrophobic substit-
uents, Figure 8, suggest the possibility of chiral discrimination
between residues that have these structural components. Un-
fortunately, ester substrates that contain carboxylate anions tend
to be very unstable due to efficient intramolecular catalysis, and
monoesters of succinic or aspartic acids have short lifetimes in
aqueous solution. We turned instead to the corresponding Lys
substrate and its catalysis by MNEE. Unfortunately, the
reactivity of MNEE is very low, and we could therefore not
measure the peptide catalyzed reactions because they were
masked by background hydrolysis. The observed chiral dis-
crimination between D- and L-Nle is probably almost exclusively
The direct assignment of binding interactions to individual
amino acid residues is complicated by what seems to be
interactions between the binding amino acids. Lys-30 in
combination with Gln-34 in MNKQ does not enhance the rate
of hydrolysis beyond that of MN-42 which has no binding
residues in helix II, yet the introduction of Lys-30 next to Arg-
+
dictated by the inherent constraints from the HisH -His pair
and due to charge repulsion between positively charged residues.
The cationic amino group of the L-Nle ester probably experi-
ences stronger repulsive interactions from the positively charged
Arg and Lys in positions 30 and 34 than that of the D-ester.
Nevertheless, the observation of chiral discrimination by rational
design is clearly gratifying. The number of possible reaction
pathways must, however, be further reduced in order to achieve
larger enantiomeric differences. Alternatively, the trapping of
the acyl intermediate by chiral nucleophiles such as amino
alcohols or peptides may generate larger enantiomeric excesses.
34 in MNKR increases the reactivity in comparison to that of
Arg-34 in MNQR by a factor of 2. Also, the replacement of
Arg-30 or Arg-34 in MNRR by Lys-30 or Lys-34 increases the
second-order rate constants in MNKR and MNRK. The
interplay between the binding residues in helix II is not clear
but may be due to charge-charge repulsions that force the
binding residues to adopt nonoptimal binding geometries.
In our system, the binding site for ester hydrolysis is
structurally well-defined; only a limited number of positions in
the adjoining helix can be used to enhance the reactivity of the
peptide catalysts. In the JNII series of peptides, where there is
The magnitudes of the rate enhancements caused by the
introduction of Arg and Lys and by the recognition of
hydrophobic residues are satisfying, considering that they are
due to a small number of interactions between the substrate and
one or two residues. The active sites of efficient biocatalysts
contain a multitude of residues organized in three-dimensional
space capable of hydrogen bonding and other electrostatic
interactions with the substrate. The incorporation of further
binding residues in designed polypeptide catalysts can therefore
be expected to enhance rates even further. However, more
+
only a single well-defined HisH -His site, Arg-11 and Arg-
1
1
5 enhance the reactivity by almost a factor of 2. Arg-8-Lys-
1 and Arg-11-Lys-8 have negligible effects on the second-
order rate constants, those of JNII, JNIIK8R11, and JNIIR8K11
are virtually the same.
Hydrophobic interactions can be expected to provide con-
siderably larger binding energies than electrostatic ones, the
binding energy of the side chains of, for example, Leu in
enzymes catalyzing the biosynthesis of proteins amount to
(28) Fersht, A. Enzyme Structure and Mechanism; 2nd ed.; W. H.
Freeman and Company: New York, 1985.

10294 J. Am. Chem. Soc., Vol. 120, No. 40, 1998
Broo et al.
importantly, the precise incorporation of residues that bind
cooperatively provides binding energies in excess of the sums
of the participating species. Synergic effects among many
residues that bind with modest energies may therefore prove to
be more important than the introduction of individual strongly
binding interactions, although the latter may be more tempting.
Consequently, the binding energies obtained here pave the way
for the further development of designed catalysts. Whether that
development takes place on the surface of a folded polypeptide
or of another biomolecule or in a cavity of a protein remains a
prospect for the future. The NMR spectroscopic results show
that the surface-exposed residues of the peptide catalysts are
not held in fixed positions, a factor that may limit the catalytic
efficiency of surface catalysts.
recognition, chiral discrimination, and saturation kinetics. Thus
these catalysts are capable of all the complexities of the naturally
occurring biocatalysts, although the rate enhancements are far
from those that occur in nature. Binding energies are surpris-
ingly large, considering that the sites are located on the surface
of folded four-helix bundle proteins and exposed to solvent water
with a high dielectric constant and in competition with hydrogen
bonding from the solvent molecules. The observed substrate
recognition is the first reported example in a designed polypep-
tide with an adopted tertiary fold where individual substrates
are specifically recognized by the catalyst.
Experimental Section
Peptide Synthesis. The peptides were synthesized on a Perseptive
Pioneer automated peptide synthesizer using a standard PerSeptive
Biosystems Fmoc protocol and an Fmoc-PAL-PEG-PS polymer (Per-
Septive Biosystems) that leaves the cleaved peptide amidated at the C
terminus. The peptide N-terminals were capped by acetic anhydride.
Peptides were cleaved from the polymer and deprotected in a mixture
of TFA (9 µL), anisole (200 µL), ethanedithiol (300 µL), and thioanisole
(500 µL) per gram of polymer, for 2 h at room temperature, precipitated
by cold diethyl ether, and lyophilised. They were purified by reversed-
phase HPLC on a semipreparative C-8 Kromasil column, eluted
isocratically with 39 to 43% 2-propanol in 0.1% TFA at a flow rate of
Interestingly, the introduction of binding residues that en-
hanced the reactivity led to a catalyst that exhibits saturation
kinetics. As a natural consequence of increasing the transition-
state binding, stronger ground-state binding followed, Figure
6. A KM of 1 mM is comparable to those of many natural
biocatalysts and also to those of other designed catalysts. It
can therefore be concluded that the binding strength is sufficient
for efficient catalysis, and that kcat should be the main target
for redesign. The strong binding of I, on the surface of a four-
helix bundle is perhaps surprising, and corroborative evidence
was therefore acquired. The observation by NMR spectroscopy
that I binds to MNRK at submillimolar concentrations of peptide
and substrate with clearly observable chemical shift effects on
residues in the hydrophobic core implies that hydrophobic
interactions coupled to charge-charge interactions may be
involved. The structure of the peptide-substrate complex
cannot be obtained due to fast exchange, but hydrophobic
binding is strongly suggested by the chemical shifts of the
residues in the hydrophobic core of the folded motif. The
corresponding effects on the surface-exposed residues are not
equally informative since they are small due to shift-averaging
on the surface of the helix-loop-helix motif.
5
mL/min. The purity was checked by analytical HPLC under similar
conditions, and the peptides were identified by electrospray mass
spectrometry (ES-MS). Typical measured molecular weights were
within 1 mu from the calculated ones, and after HPLC purification, no
high molecular weight impurities were detected in the ES-MS spectrum.
1
NMR and CD Spectroscopy. 1D H NMR spectra were recorded
at 400 MHz using a Varian Unity 400 NMR spectrometer at peptide
concentrations of approximately 0.4 mM. The identification of a
peptide-substrate complex formed from MNRK and I was performed
in 90% H O/10% D O at pH 5.1 (uncor) and 293 K. The peptide
2 2
solution (0.4 mM, 650 µL) was temperature-equilibrated, and spectra
were recorded immediately before and after the addition of 50 µL of
a saturated solution of I in 50% acetonitrile-d
6
/50% 100 mM sodium
acetate-d buffer. NOESY and TOCSY spectra were recorded at 500
3
MHz on a Varian Inova 500 NMR spectrometer operating at 313 K,
and peptide solutions were approximately 1 mM in 90%H O/10%D O
With the observations of the formation of a peptide-substrate
complex, a rate-limiting formation of an acyl intermediate that
can be trapped by nucleophiles, and the dependence of reaction
rate on pH, a free energy diagram of MNKR catalyzed
hydrolysis of I could be constructed. The evidence for the
formation of a peptide-substrate complex comes from the direct
measurement of saturation kinetics and from direct observation
by NMR spectroscopy. The structure of the transition state with
an unprotonated His nucleophile flanked by a protonated His
residue is concluded from the pH dependence and from the
observation of kinetic isotope effects.19 The binding of the
developing oxyanion and of the fumaryl carboxylate anion has
been established by the structure-function relationships reported
here. The observation that the nature of the reaction products
depends on the presence of nucleophiles shows that there is an
intermediate on the reaction pathway and that this most likely
is an acyl imidazole. Since it does not accumulate, it does not
impede the overall reaction. Finally, the release and the identity
of the reaction products have been demonstrated by NMR
spectroscopy. The rationally designed four-helix bundle catalyst
MNKR is therefore characterized by the complexity of native
enzymes, and the relationship between structure and function
is well understood. KM is quite satisfactory for efficient catalysis
but kcat remains to be optimized further.
2
2
at pH 5. The 90° pulses were 4.6 ms, spin-lock pulses were 20 ms,
sweep widths were 6500 Hz, and 2 *256 increments were recorded
and processed using linear prediction in the indirect dimension.
CD spectra were recorded on a Jasco J-720 spectrometer, routinely
calibrated with (+)-camphor-10-sulfonic acid. The samples were
prepared in buffer solution, and peptide concentrations were determined
by quantitative amino acid analysis. Typical CD spectra were measured
using 0.1-mm cuvettes at pH 5.1 in 100 mM sodium acetate in the
wavelength interval 260 to 200 nm.
Kinetic Measurements. The kinetic measurements were carried
out by using Varian Cary 1 or Cary 4 spectrophotometers equipped
with Varian temperature controllers and by following the absorbance
of released p-nitrophenol at 320 nm. The samples were prepared from
a pH-adjusted and centrifuged stock peptide buffer solution and diluted
to the desired concentrations, and the concentrations of the peptide stock
solutions were determined by quantitative amino acid analysis. In a
typical kinetic experiment, 270 µL of the peptide solution (0.2-0.4
mM) was temperature-equilibrated in a 1-mm cuvette, after which 5
µL of the substrate solution (7.4 mM) was added to give a substrate
concentration of 0.13 mM.
The substrates I, L-IV and D-IV were dissolved in 50% acetonitrile/
50% buffer, and the fatty acid esters were dissolved in 100%
acetonitrile. Peptide catalyzed hydrolyses of I and II were performed
in parallel with samples prepared from the same peptide stock solutions
to avoid interexperimental errors from quantitative amino acid analysis.
The rate constants reported are the results from linear regression analysis
of the experimentally measured pseudo-first-order rate constants as a
function of three or more peptide concentrations.
Conclusion. We have shown by rational design that four-
helix bundle catalysts can be engineered to have the capability
of catalyzing the hydrolysis of p-nitrophenyl esters with rate
enhancements approaching three orders of magnitude, substrate
The relative rates for MN-42 and JN-42 catalyzed hydrolysis of fatty
acid esters were measured with samples prepared from a common

Designed Four-Helix Bundle Catalysts
J. Am. Chem. Soc., Vol. 120, No. 40, 1998 10295
peptide stock solution. The relative rates for peptide catalyzed
hydrolysis of D-IV and L-IV were measured by splitting a peptide stock
solution into two cuvettes, adding D-IV to one cuvette and L-IV to the
other, and measuring the release of p-nitrophenol at 320 nm. After
the reaction was complete, a second addition of substrate was made
with D-IV added to the cuvette where L-IV was initially added, and
vice versa, to minimize interexperimental errors, and the reaction rate
was again measured. The reported second-order rate constants are the
mean value of two experiments after correction for background
hydrolysis. The rate constants for the background hydrolysis of D-IV
and L-IV were measured several times, and the individual rate constants
did not differ by more than 5%. All second-order rate constants
reported for 4-methyl imidazole catalyzed reactions are the results from
linear regression using three or more different concentrations of catalyst.
In the initial rate experiments, a peptide stock solution was prepared
with a concentration of approximately 0.3 mM at pH 5.1 in 100 mM
sodium acetate buffer. Six cuvettes with varying volumes of peptide
stock solution were temperature-equilibrated at 290 K in the sample
compartment of the UV spectrophotometer. The remainder of the
peptide stock solution was also temperature-equilibrated as was an
Eppendorf tube with a weighed quantity of I. Immediately prior to
starting the experiment, the peptide stock solution was added to the
substrate, the tube was briefly vortexed, and undissolved I was spun
down. The appropriate volumes of the peptide solution, saturated with
I, were added quickly to the cuvettes, and the release of p-nitrophenol
was measured at 320 nm after a brief shaking of each cuvette. The
substrate I was soluble to about 2 mM under those conditions.
The concentrations of substrate varied from about 0.05 mM to
saturated (2 mM). Quantitative amino acid analysis after the kinetic
runs was used to verify that no peptide precipitation occurred with the
excess substrate. The reactions were typically followed for 30 min,
and data points recorded during the first 20 min were used in the
calculation of the initial rate. The amount of substrate in each cuvette
was determined by withdrawing 150 µL by pipet and transferring it
into a 5-mL measuring flask. The substrate was then completely
hydrolyzed by the addition of 0.1 M NaOH. The absorbance at 405
nm due to the p-nitrophenoxide ion was measured in a 1-cm cuvette,
and the concentration was calculated using a previously determined
extinction coefficient ꢀ ) 17 840 M cm . The initial rate (V) was
M
fitted to eq 1, and the values of K and kcat were obtained from the
-
1
-1
best fit of the experimental data.
Synthesis of Substrates. The synthesis of mono-p-nitrophenyl
fumarate has been described previously.29 The p-nitrophenyl esters of
D- and L-Nle were synthesized from D- and L-Nle, respectively, using
standard procedures. The R-amino groups were protected by the
benzyloxycarbonyl group using 1.1 equiv of benzyl chloroformate in
alkaline aqueous solution at 10 °C. The carboxylic acid was esterified
with 1.1 equiv of p-nitrophenol, 1.1 equiv of dicyclohexylcarbodiimide
and 0.1 equiv of 4-pyrrolidinopyridine. Finally, the R-amino group
was deprotected using 33% HBr in acetic acid, and the resulting salt
was precipitated with diethyl ether. The purity of each product was
checked with 1H NMR spectroscopy, where the aromatic proton
resonances of the p-nitrophenyl esters are diagnostic of ester formation
and the spin systems of the amino acids are readily assigned. The
p-nitrophenyl esters of the enantiomers of lysine were synthesized and
identified in a similar manner.
Acknowledgment. We are indebted to Dr. Gunnar Sten-
hagen for expert mass spectrometric assistance, to Dr. Linda
Cammish, Per Septive Biosystems, for the synthesis of several
of the MN series of peptides. Financial support from the Swedish
Natural Science Research Council and from Carl Tryggers
Foundation is gratefully acknowledged. We thank the Swedish
NMR Centre at G o¨ teborg University for the use of a 500-MHz
NMR spectrometer.
JA980682E
(29) Baltzer, L.; Lundh, A.-C.; Broo, K.; Olofsson, S.; Ahlberg, P. J.
Chem. Soc., Perkin Trans. 2. 1996, 1671-1676.