Functional and mechanistic analyses of biomimetic aminoacyl transfer reactions in de novo designed coiled coil peptides via rational active site engineering

Article abstract of DOI:10.1021/ja068052x

Ribosomes and nonribosomal peptide synthetases (NRPSs) carry out instructed peptide synthesis through a series of directed intermodular aminoacyl transfer reactions. We recently reported the design of coiled-coil assemblies that could functionally mimic t

Full text of DOI:10.1021/ja068052x

Published on Web 02/16/2007
Functional and Mechanistic Analyses of Biomimetic
Aminoacyl Transfer Reactions in de Novo Designed Coiled
Coil Peptides via Rational Active Site Engineering
Luke J. Leman, Dana A. Weinberger, Zheng-Zheng Huang, Keith M. Wilcoxen, and
M. Reza Ghadiri*
Contribution from the Department of Chemistry and The Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
Received November 10, 2006; E-mail: ghadiri@scripps.edu
Abstract: Ribosomes and nonribosomal peptide synthetases (NRPSs) carry out instructed peptide synthesis
through a series of directed intermodular aminoacyl transfer reactions. We recently reported the design of
coiled-coil assemblies that could functionally mimic the elementary aminoacyl loading and intermodular
aminoacyl transfer steps of NRPSs. These peptides were designed initially to accelerate aminoacyl transfer
mainly through catalysis by approximation by closely juxtaposing four active site moieties, two each from
adjacent noncovalently associated helical modules. In our designs peptide self-assembly positions a cysteine
residue that is used to covalently capture substrates from solution via transthiolesterification (substrate
loading step to generate the aminoacyl donor site) adjacent to an aminoacyl acceptor site provided by a
covalently tethered amino acid or modeled by the ꢀ-amine of an active site lysine. However, through
systematic functional analyses of 48 rationally designed peptide sequences, we have now determined that
the substrate loading and intermodular aminoacyl transfer steps can be significantly influenced (up to ∼103-
fold) by engineering changes in the active site microenvironment through amino acid substitutions and
variations in the inter-residue distances and geometry. Mechanistic studies based on 15N NMR and kinetic
analysis further indicate that certain active site constellations furnish an unexpectedly large pKa depression
(1.5 pH units) of the aminoacyl-acceptor moiety, helping to explain the observed high rates of aminoacyl
transfer in those constructs. Taken together, our studies demonstrate the feasibility of engineering efficient
de novo peptide sequences possessing active sites and functions reminiscent of those in natural enzymes.
Introduction
Despite recent advances in the understanding of protein
structure and function, the rational design of highly efficient
The wide breadth of reactions catalyzed by enzymes derives
in large part from Nature’s ability to form structurally precise,
chemically tuned constellations of functional groups known as
active sites. Among the numerous avenues that enzymes
navigate to achieve rate enhancements, the most common
features are (i) substrate binding events that lower entropic
barriers, in part by bringing reactant molecules into close
proximity and excluding bulk water molecules from the active
site; (ii) precise spatial positioning of active site functional
groups, providing for cooperative covalent or general acid/base
catalysis, stabilization of transition state geometries, and
specificity; and (iii) active site electrostatic optimization,
including the dispersion of charge and pKa perturbations of
active site residues.1 Clearly, designers of synthetic enzymes
must likewise be able to predictably position and modulate the
reactivity of active site residues if enzyme-like rates and
specificity are to be realized.
peptide-based enzyme mimetics has remained elusive.2 Fur-
thermore, only a few examples are known that exploit perturba-
tions of the reactivity of individual residues to enhance catalytic
efficiencies. An early example was that of Benner and co-
workers, whose synthetic helical peptides exploited pKa depres-
sions of amines from lysine side chains or the N-terminus to
catalyze oxaloacetate decarboxylation with high efficiency.3 The
laboratories of Baltzer, Miller, and others have employed
histidine sites on peptide scaffolds to catalyze acylation or
p-nitrophenyl ester hydrolysis reactions with good efficiencies;4
in several such designs, pKa modulations of His residues were
(2) Corey, M. J.; Corey, E. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 11428-
34.
(3) Johnsson, K.; Allemann, R. K.; Widmer, H.; Benner, S. A. Nature 1993,
365, 530-2.
(4) (a) Broo, K.; Brive, L.; Lundh, A. C.; Ahlberg, P.; Baltzer, L. J. Am. Chem.
Soc. 1996, 118, 8172-3. (b) Broo, K.; Allert, M.; Andersson, L.;
Erlandsson, P.; Stenhagen, G.; Wigstrom, J.; Ahlberg, P.; Baltzer, L. J.
Chem. Soc., Perkin Trans. 2 1997, 2, 397-8. (c) Broo, K. S.; Brive, L.;
Ahlberg, P.; Baltzer, L. J. Am. Chem. Soc. 1997, 119, 11362-72. (d) Broo,
K. S.; Nilsson, H.; Nilsson, J.; Baltzer, L. J. Am. Chem. Soc. 1998, 120,
10287-95. (e) Jarvo, E. R.; Copeland, G. T.; Papaioannou, N.; Bonitatebus,
P. J., Jr.; Miller, S. J. J. Am. Chem. Soc. 1999, 121, 11638-43. (f) Baltzer,
L.; Nilsson, H.; Nilsson, J. Chem. ReV. 2001, 101, 3153-63. (g) Miller, S.
J. Acc. Chem. Res. 2004, 37, 601-10. (h) Nicoll, A. J.; Allemann, R. K.
Org. Biomol. Chem. 2004, 2, 2175-80. (i) Blank, J. T.; Miller, S. J.
Biopolymers 2006, 84, 38-47.
(1) (a) Jencks, W. P. Catalysis in Chemistry and Enzymology, McGraw-Hill
Series in AdVanced Chemistry; McGraw-Hill Book Company: New York,
1969. (b) Fersht, A. Structure and Mechanism in Protein Science; W. H.
Freeman and Company: U.S.A., 1999. (c) Wolfenden, R.; Snider, M. J.
Acc. Chem. Res. 2001, 34, 938-45. (d) Bruice, T. C. Acc. Chem. Res.
2002, 35, 139-48. (e) Menger, F. M. Pure Appl. Chem. 2005, 77, 1873-
86.
9
10.1021/ja068052x CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 2959-2966
2959

A R T I C L E S
Leman et al.
suggested to promote catalysis. Further, positioning lysine
residues within the hydrophobic core of four-helix bundle
peptides has been shown to afford site selective Lys acylation
stemming from side chain pKa depressions of up to one pK unit.5
Our group and others have demonstrated bimolecular peptide
ligation reactions with efficiencies ([kcat/KM]/kuncat) of up to 5
orders of magnitude,6 but in contrast to the designs described
above, nearly all the observed rate enhancement in these ligases
derived from entropic templating effects. Other examples of
designed enzyme mimetics have employed prosthetic cofactors
such as metal ions, porphyrins, flavins, hemes, or pyridoxamines
in peptide-based scaffolds,7 although in many cases the resulting
reactivity was not appreciably different from the isolated
prosthetic group.
Nonribosomal peptide synthetases (NRPSs)8 generate peptide
sequences by first loading amino acid substrates from solution
onto carrier domains as aminoacyl thiolesters and subsequently
catalyzing directed intermodular aminoacyl transfer (Figure 1a).
The present study focuses on the composite active sites located
at the helical interfaces of recently designed coiled-coil peptides
(Figure 2)9 designed to mimic the two main steps of NRPS
(aminoacyl loading and intermodular acyl transfer) by nonco-
valently assembling aminoacyl donor and acceptor modules into
productive supramolecular complexes (Figures 1b, 2).9 In our
designs peptide self-assembly positions a cysteine residue, used
for the covalent capture of substrates from solution via trans-
thiolesterification (substrate loading step to generate the ami-
noacyl donor site), adjacent to an aminoacyl acceptor site
Figure 1. Schematic representation of aminoacyl substrate loading and
intermodular aminoacyl transfer reactions in (a) nonribosomal peptide
synthetases (NRPSs) and (b) the supramolecular peptides described here.
A ) adenylation domain, T ) thiolation domain, C ) condensation domain,
SNAC ) N-acetylcysteamine.
(5) (a) Allert, M.; Baltzer, L. Chem.sEur. J. 2002, 8, 2549-60. (b) Andersson,
L. K.; Caspersson, M.; Baltzer, L. Chem.sEur. J. 2002, 8, 3687-97.
(6) (a) Lee, D. H.; Granja, J. R.; Martinez, J. A.; Severin, K.; Ghadiri, M. R.
Nature 1996, 382, 525-8. (b) Severin, K.; Lee, D. H.; Martinez, J. A.;
Ghadiri, M. R. Chem.sEur. J. 1997, 3, 1017-24. (c) Severin, K.; Lee, D.
H.; Kennan, A. J.; Ghadiri, M. R. Nature 1997, 389, 706-9. (d) Yao, S.;
Ghosh, I.; Zutshi, R.; Chmielewski, J. J. Am. Chem. Soc. 1997, 119, 10559-
60. (e) Yao, S.; Ghosh, I.; Zutshi, R.; Chmielewski, J. Nature 1998, 396,
447-50. (f) Kennan, A. J.; Haridas, V.; Severin, K.; Lee, D. H.; Ghadiri,
M. R. J. Am. Chem. Soc. 2001, 123, 1797-803. (g) Matsumura, S.;
Takahashi, T.; Ueno, A.; Mihara, H. Chem.sEur. J. 2003, 9, 4829-37.
(7) (a) Sasaki, T.; Kaiser, E. T. J. Am. Chem. Soc. 1989, 111, 380-1. (b)
Mihara, H.; Tomizaki, K. Y.; Nishino, N.; Fujimoto, T. Chem. Lett. 1993,
1533-6. (c) Choma, C. T.; Lear, J. D.; Nelson, M. J.; Dutton, P. L.;
Robertson, D. E.; DeGrado, W. F. J. Am. Chem. Soc. 1994, 116, 856-65.
(d) Geier, G. R.; Sasaki, T. Peptides: Chemistry, Structure and Biology,
Proceedings of the American Peptide Symposium, 14th, Columbus, Ohio,
June 18-23, 1995 1996, 591-2. (e) Roy, R. S.; Imperiali, B. Protein Eng.
1997, 10, 691-8. (f) Sakamoto, S.; Ueno, A.; Mihara, H. J. Chem. Soc.,
Perkin Trans. 2 1998, 2395-404. (g) Baltzer, L. Curr. Opin. Struct. Biol.
1998, 8, 466-70. (h) Shogren-Knaak, M. A.; Imperiali, B. Bioorg. Med.
Chem. 1999, 7, 1993-2002. (i) Geier, G. R., III; Sasaki, T. Tetrahedron
1999, 55, 1859-70. (j) Tomizaki, K.; Tsunekawa, Y.; Akisada, H.; Mihara,
H.; Nishino, N. J. Chem. Soc., Perk. Trans. 2 2000, 4, 813-22. (k)
Lombardi, A.; Nastri, F.; Pavone, V. Chem. ReV. 2001, 101, 3165-89. (l)
Rosenblatt, M. M.; Wang, J.; Suslick, K. S. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 13140-5. (m) Kaplan, J.; DeGrado, W. F. Proc. Natl. Acad.
Sci. U.S.A. 2004, 101, 11566-70. (n) Sakamoto, S.; Ito, A.; Kudo, K.;
Yoshikawa, S. Chem.sEur. J. 2004, 10, 3717-26. (o) Butterfield, S. M.;
Goodman, C. M.; Rotello, V. M.; Waters, M. L. Angew. Chem., Int. Ed.
2004, 43, 724-7. (p) Huang, S. S.; Koder, R. L.; Lewis, M.; Wand, A. J.;
Dutton, P. L. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5536-41. (q) Ishida,
M.; Dohmae, N.; Shiro, Y.; Oku, T.; Iizuka, T.; Isogai, Y. Biochemistry
2004, 43, 9823-33. (r) Zhuang, J.; Amoroso, J. H.; Kinloch, R.; Dawson,
J. H.; Baldwin, M. J.; Gibney, B. R. Inorg. Chem. 2004, 43, 8218-20. (s)
Rossi, P.; Tecilla, P.; Baltzer, L.; Scrimin, P. Chem.sEur. J. 2004, 10,
4163-70. (t) Nanda, V.; Rosenblatt, M. M.; Osyczka, A.; Kono, H.;
Getahun, Z.; Dutton, P. L.; Saven, J. G.; DeGrado, W. F. J. Am. Chem.
Soc. 2005, 127, 5804-5. (u) Cochran, F. V.; Wu, S. P.; Wang, W.; Nanda,
V.; Saven, J. G.; Therien, M. J.; DeGrado, W. F. J. Am. Chem. Soc. 2005,
127, 1346-7. (v) Fukushima, H.; Sakamoto, S.; Kudo, K. Pept. Sci. 2006,
42, 419-22.
Figure 2. Representations of the composite aminoacyl transfer active sites.
(a) Schematic diagram illustrating important principles of the active site
design. (b) The positions of active site residues (yellow sticks, boxed) and
secondary sphere residues (blue sticks) for sequence 1 are modeled onto
the crystal structure12 of a homotetrameric coiled-coil scaffold. For clarity,
residues on only one of the four symmetry-related interhelical faces are
shown. (c) Simplified active site representation for peptide 1.
provided by a covalently tethered amino acid or the ꢀ-amine of
an active site lysine (Figure 1b). The resulting high effective
concentration of aminoacyl donor and acceptor moieties,10 as
well as the potential electrostatic and/or general acid/base
contributions provided by the flanking X1 and X2 residues, were
expected to afford significant intermodular aminoacyl transfer
rate accelerations. In the present study, we have systematically
varied the inter-residue distances, identities, and positions of
active site residues using 48 rationally designed peptide
sequences (Table 1) to probe the influence of active site
microenvironments on the overall catalytic efficiency. We
demonstrate that the substrate loading and intermodular ami-
(10) (a) Wieland, T.; Bokelmann, E.; Bauer, L.; Lang, H. U.; Lau, H.; Schafer,
W. Justus Liebigs Ann. Chem. 1953, 583, 129-49. (b) Brenner, M. Pept.
Proc. Eur. Pept. Symp., 8th 1967, 1-7. (c) Neumann, H.; Shashoua, V.
E.; Sheehan, J. C.; Rich, A. Proc. Natl. Acad. Sci. U.S.A. 1968, 61, 1207-
14. (d) Kemp, D. S. Biopolymers 1981, 20, 1793-804. (e) Sasaki, S.;
Shionoya, M.; Koga, K. J. Am. Chem. Soc. 1985, 107, 3371-2. (f) Dawson,
P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science 1994, 266,
776-9. (g) Tamura, K.; Schimmel, P. Proc. Natl. Acad. Sci. U.S.A. 2001,
98, 1393-7. (h) Offer, J.; Boddy, C. N. C.; Dawson, P. E. J. Am. Chem.
Soc. 2002, 124, 4642-6. (i) Leleu, S.; Penhoat, M.; Bouet, A.; Dupas, G.;
Papamicaeel, C.; Marsais, F.; Levacher, V. J. Am. Chem. Soc. 2005, 127,
15668-9. (j) Snyder, T. M.; Liu, D. R. Angew. Chem., Int. Ed. 2005, 44,
7379-82. (k) Chen, G.; Warren, J. D.; Chen, J.; Wu, B.; Wan, Q.;
Danishefsky, S. J. J. Am. Chem. Soc. 2006, 128, 7460-2.
(8) (a) Cane, D. E.; Walsh, C. T. Chem. Biol. 1999, 6, R319-R25. (b) Khosla,
C.; Harbury, P. B. Nature 2001, 409, 247-52. (c) Sieber, S. A.; Marahiel,
M. A. Chem. ReV. 2005, 105, 715-38. (d) Fischbach, M. A.; Walsh, C. T.
Chem. ReV. 2006, 106, 3468-96.
(9) Wilcoxen, K. M.; Leman, L. J.; Weinberger, D. A.; Huang, Z.-Z.; Ghadiri,
M. R. J. Am. Chem. Soc. 2007, 129, 748-9.
9
2960 J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007

Engineering Active Sites for Aminoacyl Transfer
A R T I C L E S
Table 1. Active Sites, Molar Ellipticities, and Sequences of the Designed Peptidesa
[
θ
] (deg cm2
1
peptide
residue 8
residue 13
X1
X2
Nagg
dmol-
)
sequence
1
1*
2
3
4
5
6
7
8
Lys
Cys
Cys
Cys
HCys
HCys
Cys
Cys
Cys
Lys
Lys
Lys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Cys
Cys
Cys
Ser
His
His
His
His
His
His
His
His
His
His
His
Arg
Phe
Asn
Ala
Ala
Asp
Asp
Arg
Phe
Asn
Ala
Asp
His
His
His
His
His
His
Ala
Ala
Asp
His
His
His
His
Asp
His
Ala
∼
His
His
His
His
His
His
His
His
His
His
His
Arg
Phe
Asn
Ala
Ala
Asp
Asp
His
His
His
His
His
Arg
Phe
Ala
Asp
His
His
Ala
His
His
Ala
Asp
His
Asp
His
His
His
∼
4
4
4
4
4
4
3
2
4
3
2
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
∼
4
4
4
4
4
4
4
4
-25 800
-23 200
-24 500
-25 200
-24 700
-28 400
-27 600
-27 000
-30 400
-24 100
-25 700
-22 900
-26 000
-24 500
-25 200
-25 600
-27 300
-28 000
-26 300
-24 900
-24 900
-24 300
-26 000
-22 800
-24 500
-23 100
-25 500
-26 400
-23 100
N.D.
Aba-RLENILSKLHHICRELARIRRLLGER-CONH2
Aba-RRLENILSKLHHI(CAcm)RELARIRRLLGERRR-CONH2
Aba-RLENILS(Orn)LHHICRELARIKKLLGER-CONH2
Aba-RLENILSKLHHI(HCys)RELARIKKLLGER-CONH2
Aba-RLENILS(Orn)LHHI(HCys)RELARIKKLLGER-CONH2
Aba-RMKQIEDKLENILSKLHHICRELARIKKLLGER-CONH2
Aba-RMKQIEDKIENILSKIHHICREIARIKKLIGER-CONH2
Aba-RMKQLEDKIENLLSKIHHLCREIARLKKLIGER-CONH2
Aba-RMKQIEDKLENILSCLHHIKRELARIKKLLGER-CONH2
Aba-RMKQIEDKIENILSCIHHIKREIARIKKLIGER-CONH2
Aba-RMKQLEDKIENLLSCIHHLKREIARLKKLIGER-CONH2
Aba-RLENILSKLRRICRELARIRRLLGER-CONH2
Aba-RLENILSKLFFICRELARIRRLLGER-CONH2
Aba-RLENILSKLNNICRELARIRRLLGER-CONH2
Aba-RLENILSKLAAICRELARIKKLLGER-CONH2
Aba-RRLENILSKLAAI(CAcm)RELARIRRLLGERRR-CONH2
Aba-RLENILSKLDDICRELARIRRLLGER-CONH2
Aba-RRLENILSKLDDI(CAcm)RELARIRRLLGERRR-CONH2
Aba-RLENILSKLRHICRELARIRRLLGER-CONH2
Aba-RLENILSKLFHICRELARIRRLLGER-CONH2
Aba-RLENILSKLNHICRELARIRRLLGER-CONH2
Aba-RLENILSKLAHICRELARIKKLLGER-CONH2
Aba-RLENILSKLDHICRELARIRRLLGER-CONH2
Aba-RLENILSKLHRICRELARIRRLLGER-CONH2
Aba-RLENILSKLHFICRELARIRRLLGER-CONH2
Aba-RLENILSKLHAICRELARIKKLLGER-CONH2
Aba-RLENILSKLHDICRELARIRRLLGER-CONH2
Aba-RLENILSCLHHIKRELARIKKLLGER-CONH2
Aba-RRLENILS(CAcm)LHHIKRELARIRRLLGERRR-CONH2
Aba-RLENILSCLAAIKRELARIRRLLGER-CONH2
Aba-RLENILSCLAHIKRELARIRRLLGER-CONH2
Aba-RLENILSCLDHIKRELARIRRLLGER-CONH2
Aba-RLENILSCLHAIKRELARIRRLLGER-CONH2
Aba-RLENILSCLHDIKRELARIRRLLGER-CONH2
Aba-RLENILSCLHHICRELARIRRLLGER-CONH2
Aba-RLENILSCLHDICRELARIRRLLGER-CONH2
Aba-RLENILSCLDHICRELARIRRLLGER-CONH2
Aba-RLENILSKLHHISRELARIRRLLGER-CONH2
Aba-RLENILSKLAHISRELARIRRLLGER-CONH2
Aba-SKL-COOH
Lys
Orn
Lys
Orn
Lys
Lys
Lys
Cys
Cys
Cys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Lys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Cys
Lys
Lys
Tripeptide
Lys
Lys
Lys
Cys
Cys
Lys
Lys
Cys
9
10
11
12
13
14
14*
15
15*
16
17
18
19
20
21
22
23
24
25
25*
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
-25 000
-25 800
N.D.
Ser
Ser
His
His
His
His
His
His
His
His
His
His
His
His
His
Ala
His
His
N.D.
Aba-RLENILSKLHHISRELARIRKLLGER-CONH2
Aba-RLENILSKLHHICSRLARIERLLGER-CONH2
Aba-ELANIRSKLHHICSRLARIERLLGER-CONH2
Aba-RLENILSCLHHIKSRLARIERLLGER-CONH2
Aba-ELANIRSCLHHIKSRLARIERLLGER-CONH2
Aba-RLENILSKLHAICSRLARIERLLGER-CONH2
Aba-AMRQIRDELANIRSKLHHICSRLARIERLLGER-CONH2
Aba-AMRQIRDELANIRSCLHHIKSRLARIERLLGER-CONH2
Cys
Cys
Lys
Lys
Cys
Cys
Lys
-27 300
-25 600
-27 100
-25 000
-27 800
-31 200
-30 900
a Molar ellipticity values were recorded at 222 nm for peptides (20 µM) in pH 7.0 MOPS buffer (10 mM). N.D. ) not determined. (CAcm) denotes
acetamidomethyl-protected Cys residue, (Orn) denotes ornithine, and (HCys) denotes homocysteine. Nagg denotes coiled-coil aggregation state as specified
by the nature of the hydrophobic core residues present.11
noacyl transfer steps can be significantly influenced (up to ∼103-
fold) by engineering changes in the active site microenviron-
ment.
substrate loading rates (k1) were largely unaffected, the presence
of these residues in all cases decreased the aminoacyl transfer
rates (k1) by 5- to 8-fold (Table 2), suggesting that small changes
to active site inter-residue distances can appreciably alter
aminoacyl transfer efficiency. An alternative approach to varying
active site inter-residue distances and geometries is through
modifying the coiled-coil scaffold aggregation state (sequences
5-10). As a result of the hydrophobic core expansion and the
accompanying changes in relative helix-helix rotation in going
from dimeric to tetrameric assemblies, the relative active site
side chain Rfâ-carbon atom vectors are altered and the
interhelical active site distances are shortened and more buried
at the interface (Figure 3).11 Peptides 7 and 10 designed to form
dimeric bundles (as specified by Ile and Leu residues in the a
Results and Discussion
Effect of Relative Distance and Orientation of Aminoacyl-
Donor and -Acceptor Sites. According to our design rationale,
catalysis by approximation was expected to be the primary factor
in accelerating the rates of intermodular aminoacyl transfer.
Accordingly, we sought to examine the effect of changes in
the relative orientation and distances between the aminoacyl-
donor and -acceptor moieties on the overall reaction efficiency.
We prepared peptides 2-4 (Table 2) in which ornithine (one
methylene group shorter than Lys) and homocysteine (one
methylene group longer than Cys) were substituted for the acyl-
acceptor and -donor groups in peptide 1, respectively.9 While
(11) Harbury, P. B.; Zhang, T.; Kim, P. S.; Alber, T. Science 1993, 262, 1401-
7.
9
J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007 2961

A R T I C L E S
Leman et al.
Table 2. Effect of Active Site Inter-residue Distances on
Aminoacyl Loading and Transfer Rates
Comparisons of the loading and transfer rate constants between
peptides 5 and 8, 6 and 9, and 7 and 10 (Table 2) clearly indicate
that only small differences (<4-fold) result from switching the
relative position of the active site acyl-acceptor and -donor
moieties. Taken together, these studies indicate that modest
alterations of active site inter-residue distances or the relative
donor/acceptor positioning have a less than 10-fold impact on
the observed rates of aminoacyl loading and transfer.
Effects of Active Site Amino Acid Substitutions. Analyses
of the crystal structures of parallel tetrameric coiled-coil
assemblies11-13 indicated that the X1 and X2 amino acid residues
(Figure 2) should be in close proximity to the active site
aminoacyl donor and acceptor moieties and therefore could
provide electrostatic contributions or general acid-base catalysis
in the course of aminoacyl loading and transfer processes.
Accordingly, we investigated a series of peptides with amino
acid substitutions at the X1 and X2 active site positions
(sequences 11-30). These variants were designed to explore
the contributions of hydrophobic, polar, or charged active site
microenvironments to the aminoacyl loading and transfer rates.
Interestingly, the apparent intermodular aminoacyl transfer rate
constants (k2) for these peptides varied by more than 200-fold
(Table 3, Figure 4a), while substrate loading rates (k1) varied
by ∼20-fold. The most efficient sequences contained active site
histidine, arginine, or phenylalanine residues (positively charged
or hydrophobic microenvironments), while the least efficient
aminoacyl loading and transfer in general occurred in the
presence of active site aspartic acid residues (negatively charged
microenvironments). Replacing either or both His residues at
X1 or X2 positions with Ala resulted in transfer rate decreases
of 7-fold (peptide 19), 13-fold (peptide 23), and 18-fold (peptide
14) (Table 3). The active site His residues appear to behave
cooperatively, as peptide 1 is more efficient than would be
expected from summing the marginal rate improvements of the
His/Ala and Ala/His sequences (peptides 23 and 19) relative to
k1
k2
c
1
1
peptide
residue 8b
residue 13b
Nagg
(10-3 s-
)
(10-4 s-
)
26 residue peptides:
1
2
3
4
Lys
Orn
Lys
Orn
Cys
Cys
HCys
HCys
4
4
4
4
1.3
2.1
0.9
0.9
9.2
1.2
1.7
1.2
33 residue peptides:
5
6
7
8
9
Lys
Lys
Lys
Cys
Cys
Cys
Cys
Cys
Cys
Lys
Lys
Lys
4
3
2
4
3
2
0.8
2.7
3.0
0.3
2.1
2.9
1.4
0.4
0.8
2.0
0.4
10
>10
a Reactions contained ∼100 µM peptide, 10 mM Cbz-Gly-SNAC
substrate, 10 mM tris-carboxyethyl phosphine (TCEP) as reducing agent,
285 mM Hepes pH 7.0, and ∼200 µM acetamidobenzoic acid (Aba) as
internal concentration standard. Rate constants were determined from
reaction modeling using SIMFIT.21 See Table 1 for complete peptide
sequences. b Orn denotes ornithine; HCys denotes homocysteine. In the 33-
residue sequences, the “residue 8” and “residue 13” designations correspond
to sequence positions 15 and 20 due to the seven additional N-terminaal
c
residues. Nagg denotes coiled-coil aggregation state as specified by the
nature of the hydrophobic core residues present.11
9
the Ala/Ala peptide (sequence 14). The observed efficiency
of aminoacyl transfer in sequences lacking active site histidines
(peptides 11-14) discounts the possibility of acylimidazole
intermediates in the course of the reaction. Therefore, the
observed rate accelerations in constructs with active site
histidines are most likely due to general acid-base or proton-
transfer catalysis. In the peptide variants that contain the
aminoacyl-acceptor site at position 8 (sequences 1, 11-24), the
X1 residue more significantly influences the efficiency of
aminoacyl transfer than X2 (Figure 4a). For instance, the
aminoacyl transfer rate (k2) for sequence 1 (X1 ) His) decreased
by >30-fold when X1 site was substituted with Asp (peptide
20), whereas Asp substitution at X2 (sequence 24) reduced the
rate by only 5-fold. In contrast, when the relative orientation
of the active site aminoacyl-donor and -acceptor positions was
switched (sequences 25-30), amino acid substitutions at the
X2 site were more influential (Table 3). The difference in
aminoacyl transfer rate between sequence 25 (X1 ) His) and
the single X1 Asp substitution (peptide 28) was just 3-fold,
whereas the single X2 Asp substitution (sequences 30) differed
from 25 by 15-fold. Our studies indicate that the pronounced
reductions in the rate of aminoacyl transfer in peptides bearing
Figure 3. Structural consequences of varying the coiled-coil aggregation
state on active site residues. Average interhelical distances between
â-carbons of e and g heptad positions decrease from the dimeric to the
tetrameric state. Further, active site side chains are projected more directly
toward the adjacent helix in the higher aggregates, as indicated by an
inspection of side chain Rfâ-carbon atom vectors (yellow sticks). Each
assembly shown is the crystallographic structure of a GCN4-derived
peptide.11,19 The figure was generated using PYMOL.20
and d core positions, respectively11) had higher substrate loading
rates than trimeric sequences 6 and 9 (Ile in core positions11)
or tetrameric peptides 5 and 8 (Leu and Ile in the a and d core
positions, respectively11). These results can be rationalized by
considering that the bimolecular transthiolesterification reaction
is likely to proceed more efficiently with acyl-donor residues
present at the more exposed helical interface in a dimeric
assembly (Figure 2). On the other hand, the dimeric peptides
exhibited marginally lower aminoacyl transfer rates than the
higher aggregates (Table 2), likely due to the poorer proximity
and less productive trajectories of the acyl-donor and -acceptor
species relative to the tri- and tetrameric species. Varying inter-
residue geometry can also be achieved by switching the position
of the active site aminoacyl-donor and -acceptor residues.
(12) Yadav, M. K.; Redman, J. E.; Leman, L. J.; Alvarez-Gutierrez, J. M.; Zhang,
Y.; Stout, C. D.; Ghadiri, M. R. Biochemistry 2005, 44, 9723-32.
(13) Yadav, M. K.; Leman, L. J.; Price, D. J.; Brooks, C. L.; Stout, C. D.;
Ghadiri, M. R. Biochemistry 2006, 45, 4463-73.
9
2962 J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007

Engineering Active Sites for Aminoacyl Transfer
A R T I C L E S
Table 3. Effect of Active Site X1 and X2 Positions on Aminoacyl
Loading and Transfer Ratesa
Figure 4. Effect of active site X1 and X2 residues on aminoacyl transfer
efficiency. (a) Aminoacyl transfer product formation in time for peptides 1
(4, X1 ) His), 19 (0, X1 ) Ala), and 20 (], X1 ) Asp). Curve fits shown
are from reaction modeling using SIMFIT.21 (b) Formation of coiled-coil-
bound dipeptide thiolester species in time for sequences 31 (], X1 ) His,
X2 ) His), 32 (0, X1 ) His, X2 ) Asp), and 33 (4, X1 ) Asp, X2 ) His).
For sequences 31 and 32 (X1 ) His), all coiled-coil-bound dipeptide species
were observed at Cys8, whereas, for sequence 33 (X1 ) Asp), the coiled-
coil-bound products were instead observed at Cys13 (as determined by
comparison with authentic samples), suggesting that the X1 and X2 residues
could be used to direct ordered peptide synthesis by modulating the relative
transfer efficiencies of distinct aminoacyl thiolesters anchored in the active
site.
k1
k2
1
1
peptide
X1
X2
(10-3 s-
)
(10-4 s-
)
1
11
12
13
14
15
His
Arg
Phe
Asn
Ala
Asp
His
Arg
Phe
Asn
Ala
Asp
1.3
2.2
1.0
2.3
3.4
0.4
9.2
2.4
2.7
0.6
0.5
0.04
Table 4. Effect of Active Site X1 and X2 Positions on Relative
Acyl-Donor and -Acceptor Reactivity in Active Sites Composed of
Two Aminoacyl Thiolester Moietiesa
16
17
18
19
20
Arg
Phe
Asn
Ala
Asp
His
His
His
His
His
1.5
2.1
0.6
1.0
0.2
7.1
0.7
1.4
1.3
0.3
21
22
23
24
His
His
His
His
Arg
Phe
Ala
Asp
1.3
0.7
1.3
0.8
2.3
2.6
0.7
1.8
25
26
27
28
29
30
His
Ala
Ala
Asp
His
His
His
Ala
His
His
Ala
Asp
3.1
6.4
5.1
0.9
3.1
0.4
1.5
0.9
1.4
0.5
1.4
0.1
a Reaction conditions are as described in Table 2. See Table 1 for
complete peptide sequences.
observed
an active site aspartic acid residue(s) stem from electrostatic
perturbations that significantly alter the pKa of the acyl acceptor
amine (vide infra).
peptide
X1
X2
acceptor siteb
31
32
33
His
His
Asp
His
Asp
His
Cys8
Cys8
Cys13
Directed Intermodular Aminoacyl Transfer Reactions.
The above findings suggest that the judicious choice of amino
acids at X1 and X2 positions could be exploited to bring about
directed aminoacyl transfer by selectively increasing or decreas-
ing transfer efficiency at particular positions. This possibility
was explored using sequences 31-33 in which phenylalanyl
thiolesters preloaded on Cys residues at positions 8 and 13 were
used as the active site aminoacyl-donor and -acceptor moieties
(Table 4, Figure 4b). Unlike peptides containing an active site
Lys as the aminoacyl acceptor moiety and an N-protected
aminoacyl thiolester as the donor, in sequences 31-33 amino-
acyl transfer can proceed via two distinct pathways: with the
thiolester at position 8 acting as the acceptor and the thiolester
at position 13 acting as the donor, or vice versa. Interestingly,
for sequences 31 and 32 (X1 ) His) we observe only evidence
of position 8 acting as the acyl-acceptor moiety (no dipeptide
a Reaction conditions are as described in Table 2, without the addition
of the Cbz-Gly-SNAC substrate. See Table 1 for complete peptide
sequences. b Denotes observed site of coiled-coil-bound dipeptide thiolester
species (as determined by comparison with independently prepared authentic
samples). For all peptides, coiled-coil-bound products were observed only
at Cys8 or at Cys13 (not at both positions).
species are observed at position 13), whereas for sequence 33
(X1 ) Asp) we instead observe only evidence of position 13
acting as the aminoacyl acceptor (Table 4, Figure 4b). These
findings are consistent with the observed higher aminoacyl
transfer rates for peptide 24 (X1 ) His, X2 ) Asp, acceptor at
position 8) than those for 30 (X1 ) His, X2 ) Asp, acceptor at
position 13) and the higher rate for 28 (X1 ) Asp, X2 ) His,
acceptor at position 13) than that for 20 (X1 ) Asp, X2 ) His,
acceptor at position 8) and support the suggestion that the X1
9
J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007 2963

A R T I C L E S
Leman et al.
Table 5. Rates of Intermolecular Lys Aminoacylation
k3
1
peptide
X1
X2
(10-4 s-
)
34
35
His
Ala
His
His
0.09
0.02
36
tripeptide; no active site
His His
solvent-exposed; no active site
0.002
0.09
0.003
37, Lys8
37, Lys21
a Reaction conditions are as described in Table 2. See Table 1 for
complete peptide sequences.
Figure 5. Enhanced aminoacyl acceptor reactivity in the designed active
site. (a) Peptide 37 contains two lysine residues, one in the partially buried
composite active site (Lys8, blue) and one at a C-terminal solvent-exposed
position (Lys21, red). Even though substrate loading cannot occur due to
an active site CysfSer substitution, the active site Lys8 (b) undergoes
bimolecular aminoacylation by the Cbz-Gly-SNAC thiolester at a rate 30-
fold higher than the more solvent-exposed Lys21 (9). A product formation
curve is also shown for peptide 1 for comparison (2). Curve fits shown are
from reaction modeling using SIMFIT.21 (b) Selected 15N NMR spectra
for peptide 1* for use in active site Lys pKa determination. At each pH
value, 1200-4000 scans were recorded. The outer coaxial NMR tube (10
mm) contained peptide 1* (3 mM), acetate (6 mM), bis-tris (6 mM), and
CAPS buffers (6 mM) in H20, while the inner tube (3 mm) contained 15N-
labeled ammonium nitrate (15NH4NO3, 2.0 M) dissolved in 95% D2O for
reference and lock signals. (c) 15N-chemical shifts plotted as a function of
pH for peptides 1* (4, X1 and X2 ) His), 14* (O, X1 and X2 ) Ala), 15*
(b, X1 and X2 ) Asp), and 25* (9, X1 and X2 ) His). The curves represent
nonlinear least-squares fits of the data to the Henderson-Hasselbalch
equation.
and X2 active site positions may indeed be employed to foster
directed aminoacyl transfer through selective active site mi-
croenvironment perturbations. It should be noted that the
analysis of these reactions is complicated by substrate hydrolysis
and cyclization (diketopiperazine formation) of the resulting
dipeptide thiolester (Table 4). The apparent higher initial rate
of aminoacyl transfer for sequence 32 (X2 ) Asp) than that for
sequence 31 (X2 ) His) (Figure 4b) indeed results from its lower
rate of diketopiperazine formation (data not shown).
Active Site Dependent pKa Perturbations. Control peptides
34 and 35 in which the active site Cys was substituted with Ser
to prevent substrate loading were prepared for experiments
aimed at estimating aminoacyl transfer rate enhancements.
Interestingly, although as expected based on the design rationale
the observed rates of intermolecular aminoacylation with these
peptides were reduced by 100-fold with respect to parent peptide
1, these rates were nevertheless faster by 10- to 45-fold as
compared to the reference tripeptide 36 (Table 5).9 To further
investigate the source of this discrepancy, we prepared peptide
37 containing two Lys residues, one in the active site environ-
ment (again containing the CysfSer substitution to prevent
aminoacyl loading) and one located at a solvent-exposed position
near the peptide terminus (Figure 5a). Subjecting this peptide
to standard reaction conditions indicated that the active site
lysine underwent bimolecular aminoacylation 30-fold more
efficiently than the solvent-exposed lysine (established via
tryptic digestion and MS sequencing analysis) even though the
terminal Lys21 is presumably more sterically accessible (Figure
5a, Table 5). We suspected that this anomalous reactivity might
have resulted from pKa depression of the active-site Lys residue
brought about by the active site microenvironment. Because
direct measurements of pKa values of specific functional groups
in proteins and peptides can be most accurately measured using
NMR by monitoring the chemical shift of the desired nucleus
as a function of pH,14 we employed 15N NMR to establish the
role of pKa shifts in the efficiency of aminoacyl transfer
reactions. The use of NMR circumvents potential problems
Table 6. Observed Active Site Lys ꢀ-Amine pKa Values for
Selected Peptides
peptidea
X1
X2
pKa
1*
14*
15*
25*
His
Ala
Asp
His
His
Ala
Asp
His
9.1 ( 0.1
9.0 ( 0.1
10.6 ( 0.1
9.1 ( 0.1
a The asterisk denotes Acm-protection of the active site Cys, 15N
enrichment of the active site Lys ꢀ-amine, and addition of three terminal
Arg residues to increase solubility. See Table 1 for full peptide sequences.
associated with pKa determination by chemical labeling methods,
which can be complicated by nonspecific binding of the
chemical agent to the protein or catalysis of the labeling reaction
by residues other than the one of interest. Accordingly, we
prepared peptides 1*, 14*, 15*, and 25* in which the active-
site Lys ꢀ-amine was selectively 15N-enriched and the active-
site Cys residue was alkylated with an acetamidomethyl (Acm)
group to mimic the steric properties of the acylated cysteine
thiolester reaction intermediate as well as to prevent oxidative
disulfide formation during the course of the NMR experiments.
The active-site Lys in peptide 1* has a pKa of 9.1, as determined
by plotting the ꢀ-amine 15N NMR chemical shift versus pH and
subjecting the curve to a nonlinear least-squares fit of the
Henderson-Hasselbalch equation (Figures 5, S2, Table 6). This
value is 1.4 units lower than the standard Lys pKa value in
peptides and proteins of 10.5.15 A shift in pKa of this magnitude
corresponds to an approximately 30-fold increase in the ratio
(14) (a) Lodi, P. J.; Knowles, J. R. Biochemistry 1993, 32, 4338-43. (b) Zhu,
L.; Kemple, M. D.; Yuan, P.; Prendergast, F. G. Biochemistry 1995, 34,
13196-202. (c) Tsilikounas, E.; Rao, T.; Gutheil, W. G.; Bachovchin, W.
W. Biochemistry 1996, 35, 2437-44. (d) Zhou, Z.; Macosko, J. C.; Hughes,
D. W.; Sayer, B. G.; Hawes, J.; Epand, R. M. Biophys. J. 2000, 78, 2418-
25.
9
2964 J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007

Engineering Active Sites for Aminoacyl Transfer
A R T I C L E S
of amine to ammonium (protonated) state at the pH 7.0 condition
of the acyl transfer reaction, consistent with the observed 10-
to 50-fold increased acylation rates. The stability and confor-
mational homogeneity of sequence 1 over the pH range required
for the pKa titration were confirmed using circular dichroism
and 13C NMR analyses. Circular dichroism spectra for 1 were
consistent with expected R-helical content and were identical
at pH values of 4, 7, and 10 (Figure S1). In addition, thermal
denaturation curves for 1 (carried out in the presence of 3 M
guanidinium hydrochloride) yielded nearly identical Tm values
of 59 °C at pH 7 and 65 °C at pH 11 (Figure S1). We further
prepared a derivative of 1* 13C-isotopically enriched at the CR
position of the Leu10 residue located directly adjacent to active
site Lys9 (not 15N-enriched in this sequence). The observed
Leu10 13CR-chemical shifts fell within the established chemical
shift range for an R-helical conformation16 at the analyzed pH
values of 5.2, 7.8, and 11.2 (Figure S3), whereas at pH 7.0 in
the presence of 6.0 M guanidinium the observed Leu10 13CR-
chemical shift instead fell within the established range for
random conformations17 (Figure S3). Under standard reaction
conditions, aminoacyl loading and transfer for 1* (after removal
of the Acm group from the active site Cys) proceeded with rates
indistinguishable from those for sequence 1.
Figure 6. Effect of secondary sphere residues on aminoacyl transfer
efficiency. (a) Schematic diagram illustrating the redesign process for
secondary sphere residues leading to second and third generation peptides.
See also Figure 2b for secondary sphere residue positions in the context of
the coiled-coil scaffold. (b) Aminoacyl transfer product formation in time
for generation 1 peptides 1 (], X1 ) His) and 23 (4, X1 ) Ala) and for
the corresponding generation 2 sequences 38 (0, X1 ) His) and 42 (O, X1
) Ala). (c) Aminoacyl transfer product formation in time for peptides 25
(0, generation 1), 40 (4, generation 2), and 41 (O, generation 3). Reaction
conditions are as described in Table 2, and all curve fits shown are from
reaction modeling using SIMFIT.21
The pKa of the active-site Lys in peptide 14* is similarly
depressed to a value of 9.0 (Table 6, Figure S2), suggesting
that the His residues in peptide 1* are not required for the
observed pKa depression and rather that proximity of the Lys
side chain to the coiled-coil hydrophobic core likely causes the
perturbation.5 Peptide 15*, containing acidic residues at the X1
and X2 positions, bears an active site Lys pKa of 10.6, clearly
indicating that active site residues near the aminoacyl acceptor
can significantly affect its reactivity by altering the local
electrostatic microenvironment. The active site Lys pKa in
peptide 25* is 9.1, indistinguishable from that of sequence 1*
and suggesting a generalized pKa perturbation for the described
arrangements of residues in tetrameric coiled-coils. The observed
differences in aminoacyl transfer rates for sequences 1, 14, and
25 thus cannot result from pKa differences of the active site
acyl acceptor, as their pKa values are nearly identical, but the
markedly poor efficiency of peptide 15 (and other sequences
containing negatively charged active site residues, Table 3) most
likely results from its relatively high aminoacyl-acceptor pKa
value. In contrast, positively charged or hydrophobic active site
microenvironments could bring about a further depression in
pKa, consistent with the generally higher observed efficiencies
of sequences containing such active sites (Table 3).
from the active site. Our initial attempts at re-engineering were
focused on modifications at the C-terminal side of the active
site (Figure 6a) stemmed from the observations that aminoacyl
loading rates (k1) were generally higher in sequences having
the acyl donor moiety at residue 8 (peptides 8-10, 25-30) than
at position 13 (peptides 1-7, 11-24). We hypothesized these
effects to arise from the less sterically demanding Ser7 residue
adjacent to position 8 compared to the more bulky Arg14
adjacent to position 13. Accordingly, we prepared generation-
two sequences based on three C-terminal amino acid substitu-
tions (Figure 6, Table 7). Arg14 was replaced with Ser to
improve bimolecular substrate loading rates (k1) for sequences
having the acyl donor moiety at position 13, and Glu15 and
Arg20 were swapped to move the negatively charged Glu
residue farther from the active site while presumably retaining
the intrahelix stabilizing electrostatic salt bridge observed in
the GCN4-pLI crystal structures11,12 between sites 15 and 20.
Three peptides (sequences 38, 40, and 42) differing in active
site composition were prepared based on these generation-two
sequence alterations. Circular dichroism spectra indicated that
the coiled-coil helical content was not significantly altered by
the three C-terminal amino acid modifications (Table 1).
Although aminoacyl loading rates were largely unchanged, the
second-generation sequences yielded 2- to 12-fold enhanced
aminoacyl transfer rates relative to the first generation peptides
(Table 7, Figure 6). The observation that aminoacyl transfer in
sequence 42 (X2 ) Ala) occurs at a rate similar to that in peptide
1 suggests that the second-generation substitutions compensate
for removal of His in the X2 position (Figure 6).
Design of Re-engineered Peptide Sequences. With an
enhanced mechanistic understanding of the basic factors con-
tributing to loading and transfer reactions, based on the active
site X1/X2 substitutions and pKa titration experiments, we set
out to increase the efficiency of the aminoacyl loading and
transfer steps by modifying residues positioned more distant
(15) (a) Bundi, A.; Wuethrich, K. Biopolymers 1979, 18, 285-97. (b) Matthew,
J. B.; Gurd, F. R. N.; Garcia-Moreno, B.; Flanagan, M. A.; March, K. L.;
Shire, S. J. Crit. ReV. Biochem. 1985, 18, 91-197. (c) Song, J.; Laskowski,
M., Jr.; Qasim, M. A.; Markley, J. L. Biochemistry 2003, 42, 2847-
56.
(16) (a) Spera, S.; Bax, A. J. Am. Chem. Soc. 1991, 113, 5490-2. (b) Holtzer,
M. E.; Lovett, E. G.; D’Avignon, D. A.; Holtzer, A. Biophys. J. 1997, 73,
1031-41.
(17) (a) Wishart, D. S.; Sykes, B. D.; Richards, F. M. J. Mol. Biol. 1991, 222,
311-33. (b) Wishart, D. S.; Sykes, B. D. Methods Enzymol. 1994, 239,
363-92.
9
J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007 2965

A R T I C L E S
Leman et al.
Table 7. Effect of Secondary Sphere Residue Redesign on
substitution may create an electrostatic repulsion that forces the
active site Lys away from the aminoacyl-donor moiety.
Aminoacyl Loading and Transfer Ratesa
residue residue
peptide
length
k1
k2
Conclusions
1
1
peptide
8
13
X1
X2
generation (10-3 s-
)
(10-4 s-
)
The studies reported here show that appropriate positioning
and the judicious choice of nearby amino acid residues can lead
to >900-fold differences in reactivity of otherwise identical
functional groups in short peptides. While the described NRPS
active site models do not approach the efficiencies of natural
enzymes, the observed covalent substrate binding, cooperativity
among active site residues, sensitivity of the reaction rate to
substitutions of active site residues, and active site pKa modula-
tions further indicate that simple, rationally designed peptides
can indeed mimic multiple features of finely tuned enzymes.
Such findings will likely prove crucial in the advancement of
these peptides, designed to functionally mimic NRPSs, toward
a programmed peptide synthesis in which sequences of reactions
are controlled by varying the relative reactivities of aminoacyl-
donor and -acceptor moieties coupled with selective and
sequence-dependent coiled-coil self-assembly.18 It is our hope
that, through such rational approaches, practical enzyme mi-
metics and catalysts for reactions not occurring in nature could
emerge.
1
38
39
Lys
Lys
Lys
Cys His His
Cys His His
Cys His His
26
26
26
1
2
3
1.3
1.2
2.1
9.2
34.8
6.3
25
40
41
Cys
Cys
Cys
Lys His His
Lys His His
Lys His His
26
26
26
1
2
3
3.1
2.2
5.3
1.5
3.4
4.2
23
42
Lys
Lys
Cys His Ala
Cys His Ala
26
26
1
2
1.3
1.7
0.7
8.5
5
43
Lys
Lys
Cys His His
Cys His His
33
33
1
3
0.8
2.0
2.9
4.9
8
44
Cys
Cys
Lys His His
Lys His His
33
33
1
3
0.3
2.1
0.8
5.7
a Reaction conditions are as described in Table 2. See Table 1 for
complete peptide sequences. In the 33-residue sequences, the “residue 8”
and “residue 13” designations correspond to sequence positions 15 and 20
due to the seven additional N-terminal residues.
Encouraged by these results, we next focused on modifying
the amino acids on the N-terminal side of the active site, again
by moving negatively charged residues away from the active
site while at the same time introducing a presumably stabilizing
salt bridge. To this end, a potential i, i + 5 electrostatic pair
was incorporated by replacing Arg1 and Leu6 with Glu and
Arg residues, respectively (Figure 6). Further, the Glu3 residue
was substituted with an Ala to reduce overall negative charge
near the active site. Circular dichroism analysis again indicated
little change to the coiled-coil helical content stemming from
these N-terminal substitutions (Table 1). Third-generation
sequences incorporating these changes (peptides 39, 41, 43, and
44) increased aminoacyl loading rates (k1) by 2- to 7-fold relative
to analogous earlier generation peptides (Table 7). Together with
the observation that Asp at the X1 or X2 positions generally
decreased substrate loading rates (Table 3), this finding may
indicate that pKa depressions of the active site Cys (in addition
to the acceptor moiety as discussed above) are important for
efficient loading rates. In general, aminoacyl transfer rates (k2)
were also improved in generation-three sequences, with the
notable exception of sequence 39 (Table 7). It is unclear why
transfer was nearly 6-fold reduced in this peptide relative to
the second-generation sequence 38, although the Leu6fArg
Acknowledgment. We thank Drs. Laura Pasternack and Dee-
Hua Huang for assistance with 15N NMR acquisition, Professor
G. von Kiedrowski for the generous gift of the program SIMFIT,
and the NSF for a predoctoral fellowship (L.J.L.).
Note Added after ASAP Publication. Due to a production
error, the version of this paper published ASAP on February
16, 2007, was not the final corrected version. Additional changes
were made in Tables 5-7, and the corrected version was
published ASAP on February 22, 2007.
Supporting Information Available: Materials and methods,
Figures S1-S3. This material is available free of charge via
the Internet at http://pubs.acs.org.
JA068052X
(18) Ashkenasy, G.; Jagasia, R.; Yadav, M.; Ghadiri, M. R. Proc. Natl. Acad.
Sci. U.S.A. 2004, 101, 10872-7.
(19) (a) O’Shea, E. K.; Klemm, J. D.; Kim, P. S.; Alber, T. Science 1991, 254,
539-44. (b) Harbury, P. B.; Kim, P. S.; Alber, T. Nature 1994, 371, 80-
3.
(20) DeLano, W. L. The PyMOL Molecular Graphics System: 2002; http://
www.pymol.org.
(21) Sievers, D.; von Kiedrowski, G. Chem.sEur. J. 1998, 4, 629-41.
9
2966 J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007