Modulating pyridoxamine-mediated transamination through a ββα motif peptide scaffold

Article abstract of DOI:10.1016/S0968-0896(99)00112-1

A pyridoxamine coenzyme amino acid chimera (Pam) was incorporated into a designed ββα motif peptide to explore the ability of a small synthetic peptide scaffold to influence coenzyme mediated transamination. Structural characterization of this peptide by CD and NMR spectroscopy suggested that the pyridoxamine containing residue was accommodated into the sheet region of the motif without gross structural perturbations. To investigate the ability of the peptide architecture to influence the amount and distribution of transamination product in the conversion of pyruvic acid to alanine, a family of 18 related peptides, CBP01-CBP18, was rapidly synthesized and purified in parallel. These peptides were designed to generate different peptide environments for the pyridoxamine functionality within the context of the structured ββα peptide motif. Studies of peptide-mediated transamination revealed clear trends in stereospecific production of L-alanine as a function of substitutions at positions five and seven of the motif. Furthermore, new trends favoring the other enantiomeric product resulted from the addition of copper(II) ion, a known chelator of the transamination reaction intermediates. In the presence of copper(II) ion the amount of alanine product generated was increased by up to 32-fold relative to a pyridoxamine model compound in the presence of copper(II) ion. These functional results, accompanied by further CD and NMR spectroscopic analysis of CBP14, one of the CBP family of peptides, suggest that small synthetic ββα motif peptides can be used to influence the functional properties of coenzymes.

Full text of DOI:10.1016/S0968-0896(99)00112-1

Bioorganic & Medicinal Chemistry 7 (1999) 1993±2002
Modulating Pyridoxamine-Mediated Transamination Through
a ꢀꢀꢁ Motif Peptide Sca€old
Michael A. Shogren-Knaak and Barbara Imperiali*
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125, USA
Received 11 December 1998; accepted 15 April 1999
AbstractÐA pyridoxamine coenzyme amino acid chimera (Pam) was incorporated into a designed bba motif peptide to explore the
ability of a small synthetic peptide sca€old to in¯uence coenzyme mediated transamination. Structural characterization of this
peptide by CD and NMR spectroscopy suggested that the pyridoxamine containing residue was accommodated into the sheet
region of the motif without gross structural perturbations. To investigate the ability of the peptide architecture to in¯uence the
amount and distribution of transamination product in the conversion of pyruvic acid to alanine, a family of 18 related peptides,
CBP01-CBP18, was rapidly synthesized and puri®ed in parallel. These peptides were designed to generate di€erent peptide envir-
onments for the pyridoxamine functionality within the context of the structured bba peptide motif. Studies of peptide-mediated
transamination revealed clear trends in stereospeci®c production of l-alanine as a function of substitutions at positions ®ve and
seven of the motif. Furthermore, new trends favoring the other enantiomeric product resulted from the addition of copper(II) ion, a
known chelator of the transamination reaction intermediates. In the presence of copper(II) ion the amount of alanine product
generated was increased by up to 32-fold relative to a pyridoxamine model compound in the presence of copper(II) ion. These
functional results, accompanied by further CD and NMR spectroscopic analysis of CBP14, one of the CBP family of peptides,
suggest that small synthetic bba motif peptides can be used to in¯uence the functional properties of coenzymes. # 1999 Elsevier
Science Ltd. All rights reserved.
Introduction
range of targets is accessible because both natural and
unnatural amino acid monomers can be employed,
many related compounds can be explored using combi-
natorial techniques, and peptide constructs can be
structurally characterized using standard biophysical
methods.2±4 To explore the potential of small synthetic
peptides to in¯uence the coenzyme reactivity, our e€orts
have focused on the chemically interesting process of
transamination.
Small synthetic peptides o€er a powerful tool for
exploring and in¯uencing the reactivity of coenzymesÐ
small, densely functionalized organic molecules that
facilitate chemically dicult reactions in enzymes. Syn-
thetic peptides combine many of the desirable features
of proteins and small molecule model systems. Like
proteins, these peptides can access complicated molec-
ular architectures. The array of functional groups
created by the peptide structure has the potential to
in¯uence the reactivity of the coenzyme in a manner
similar to the structured active sites of enzymes. Addi-
tionally, like small molecule models, these systems are
potentially simple enough to understand in atomic
detail. Small peptide constructs also possess a number
of practical advantages for experimental design: these
systems can be assembled relatively easily using techni-
ques of solid phase peptide synthesis (SPPS),1 a wide
Transamination of a-keto acids to a-amino acids
involves exchange of the amine functionality of an
amino acid for the carbonyl of a keto acid to generate a
new keto acid and amino acid (Scheme 1A). The unca-
talyzed reaction is synthetically dicult, requiring harsh
conditions and resulting in poor yields.5 However, the
coenzymes pyridoxamine-phosphate and pyridoxal-
phosphate facilitate this process by acting as catalytic
intermediaries in the reaction and achieve the same net
reaction in a kinetically more accessible manner
(Scheme 1B).6 Mechanistically, the transamination half-
reaction occurs in several steps, proceeding through
ketimine, quinoid, and aldimine intermediates before
generating the products (Scheme 2). Each of these
mechanistic steps is made feasible through the assistance
Key words: Amino acids and derivs; biomimetic reactions; peptides
and polypeptides; vitamins.
* Corresponding author. Current address: Department of Chemistry,
Massachusetts Institute of Technology, Cambridge, MA 02139; Tel.:
1-617-253-1838; e-mail: imper@mit.edu
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00112-1

1994
M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
process with some success. Previous work in our group
has focused on incorporating either a pyridoxal coen-
zyme amino acid chimera (Pal) or pyridoxamine coen-
zyme amino acid chimera (Pam, Fig. 1) into synthetic
peptide constructs, where the coenzyme functionality is
covalently attached to the peptide main chain as part of
an amino acid side chain. These peptide constructs have
included short hairpin peptides20,21 and semi-synthetic
protein constructs.14,15
Recently, this laboratory has developed a series of small
peptides that adopt a well de®ned supersecondary
structure in the absence of disul®de bridges or metal
chelation to serve as a sca€old for the development of
functional peptides.16±18 The BBA peptides include 23-
residues assume a bba supersecondary structure similar
to a zinc-®nger domain (Fig. 2). The b-hairpin region is
stabilized by inclusion of a heterochiral type II0 turn
element and by inclusion of residues with b-sheet pro-
pensities. The a-helix region is stabilized by residues
with a-helical propensities and by an a-helix capping
sequence. The packing of these two structural elements
is promoted by hydrophobic interactions within the
interior of the motif and hydrophilic residues on the
exterior. With this structurally well-de®ned sca€old, the
potential exists to create pyridoxamine containing pep-
tides with augmented functional properties by exerting
structural control over the peptide environment of the
pyridoxamine cofactor.
Scheme 1. (A) Net transamination reaction. (B) Cofactor-mediated
transamination reaction.
To take advantage of this structural motif, Pam was
incorporated into the BBA motif. Peptide BP1 (Fig. 2)
incorporated Pam as a conservative mutation into a
known structured bba motif peptide, BBA4.18 Spectro-
scopic analysis of BP1 revealed characteristics of the
desired bba-supersecondary structure. A family of pep-
tides based on peptide BP1, CBP01-CBP18 (Fig. 2), was
made to further in¯uence the yield and optical induction
of pyridoxamine mediated transamination. Functional
and structural characterization of these constructs
revealed peptides with augmented product generation
and stereospeci®city. These constructs represent the ®rst
success in making the BBA motif functional.
of the coenzyme. In enzyme-catalyzed, pyridoxamine-
dependent transamination the e€ectiveness of the coen-
zyme is enhanced and regulated by a well de®ned active
site protein environment.7
Various model systems have been examined in an
attempt to both emulate and illuminate the role of the
enzyme environment. Small synthetic model systems,8±12
as well as modi®ed proteins,13 have been designed to
facilitate various mechanistic steps in the transamination
Scheme 2. Mechanism of the half-transamination reaction.

M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
1995
Results
motif directly following the type II0 turn at position six.
This substitution replaced a structurally related tyrosine
residue and was not expected to perturb the super-
secondary structure of the motif. To in¯uence the func-
tion of Pam, a basic residue was also incorporated into
the peptide sequence. In transaminase enzymes, a con-
served lysine acts both as the general base in the rate
limiting generation of quinoid intermediate19±22 and as a
general acid to reprotonate the quinoid species to a€ord
a chiral amino acid product.23,24 Furthermore, many of
the other mechanistic steps in transamination can be
acid or base catalyzed (Scheme 2). Molecular modeling
indicated that replacing the serine at position ®ve in
BBA4 with an a,g-diaminobutyric acid residue (Dab)
might place the basic side chain of this residue to inter-
act favorably with the coenzyme amino acid.
Design of peptide BP1
The design of peptide BP1 (Fig. 2) was based on the
sequence of bba motif peptide, BBA4.18 The Pam resi-
due was placed into the b-sheet region of the peptide
Figure 1. The pyridoxamine coenzyme amino acid chimera (Pam)
incorporated into a peptide.
Figure 2. Proposed structure of BP1/CBP04 based on the structure of peptide BBA4 with Pam depicted as the ketimine intermediate. Residues 5, 6,
and 7 are shown to indicate substitution sites for peptides BP1 and CBP01-CBP18. Below are the sequences of peptides BBA4, BP1 and CBP01-
CBP18. Non-standard amino acids include: a,b-diaminoproprionic acid (Dap); a,g-diaminobutyric acid (Dab), d-proline (dPro), ornithine (Orn),
homocysteine (Hcy), and the pyridoxamine amino acid (Pam). Peptides were either acetyl (Ac) or glycolyl (Glc) capped at the amino-terminus and
were amide capped (NH2) at the carboxy-terminus.

1996
M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
Structural characterization of peptide BP1
The secondary structural features of peptide BP1 were
probed by circular dichroism (CD) spectroscopy (Fig.
3). Designed BBA peptides shown to adopt a bba-sec-
ondary structure by 2-D NMR share several common
CD spectral features at pH 4 and 8.18 While the magni-
tude of the mean residual elipticity varies, the spectrum
of each peptide shows an absolute minimum from 205
to 208 nM and a less intense minimum from 220 to
222 nM. The CD spectrum of BP1 exhibits similar
characteristics with minima at 206 and 221 nM. At pH
4, the magnitude of the mean residual elipticity at
206 nM was identical to that of parent peptide BBA4,
while the magnitude of the mean residual elipticity at
221 nM was about 16% greater. At pH 8, the pH at
which transamination assays were performed, the CD
spectra of BP1 and BBA4 were nearly identical. Thus,
the CD spectra, particularly at pH 8, suggest that pep-
tide BP1 shares some of the secondary structural char-
acteristics of peptide BBA4.
Figure 4. NOE side chain contacts for peptide BP1 in the aliphatic to
aromatic region of the NMR spectrum at pH 8. Tentative NOE
assignments based on BBA4. Chemical shifts are in ppm.
structure of peptide BBA4 is maintained at pH 8 with
the incorporation of Pam and a basic residue under the
conditions employed to assay transamination.
To complement the CD analysis, two-dimensional (2-D)
1H NMR was used to probe the supersecondary struc-
ture of peptide BP1. The solution structure of peptide
BBA4 has been solved by 2-D NMR at pH 4.18 A key
feature of this structure is a set of packing interactions
that are denoted by NOE crosspeaks between hydro-
phobic aromatic residues located exclusively on the
sheet region and hydrophobic aliphatic residues found
only on the helix. These unique NOE crosspeaks o€er a
strong diagnostic tool for the formation of super-
secondary structure, being present only when the b-
sheet region is packed against the b-helix domain. While
full structural characterization of BBA4 has not been
performed at pH 8, the pH at which transamination
assays were performed, at this pH the peptide did exhi-
bit aliphatic to aromatic NOEs crosspeaks expected in a
well-de®ned hydrophobic core (data not shown). At pH
8, the spectra of peptide BP1 also showed aliphatic-to-
aromatic side chain NOE contacts at intensities com-
parable to NOE crosspeaks for other residues (Fig. 4).
The presence of these contacts suggests that the hydro-
phobic core also exists in peptide BP1.
Functional characterization of peptide BP1
To assess the functional properties of peptide BP1, the
peptide-mediated transamination reaction was studied
using an HPLC/¯uorescence assay.25,26 In the transa-
mination reaction an excess of pyruvic acid and d-phe-
nylalanine were used and reactions were carried out in
the presence of EDTA to prevent trace metal ion cata-
lysis.27 EDTA was added to both peptides and model
compounds to both ensure that metal ion was not
available to in¯uence the reaction and to control for any
intrinsic e€ects of the EDTA itself. The amount of each
enantiomer of alanine produced from pyruvic acid was
quanti®ed at various time points by integrating the
HPLC peaks resulting from injecting aliquots of the
reaction mixture derivatized with an ortho-phthaldial-
dehyde/N-acetyl-l-cysteine solution. The initial rate of
alanine production was determined from the time-
dependent production of total alanine. The enantio-
meric excess (% ee) for the reaction was determined
from the relative quantities of each enantiomer present.
Taken together, the structural analyses of peptide BP1
by CD and NMR spectroscopy suggest the overall bba
Relative to a 50-deoxy-pyridoxamine model compound,
dPam, peptide BP1 showed a slight improvement in
properties with a 3.5-fold increase in the initial rate of
alanine production and an optical induction of 15% ee
in favor of the l-enantiomer. While peptide BP1
showed promise as a structural sca€old, improving the
functional characteristics of the peptide required tar-
geted perturbations of the coenzyme environment.
Design and synthesis of peptides CBP01-CBP18
Using BP1 as a template, a family of peptides, CBP01-
CBP18, was designed to explore a range of coenzyme
environments (Fig. 2). The replacement of serine 5 with
Dab in BP1 indicated that similar substitutions might
be tolerated at this position with a minimal loss of
Figure 3. CD spectra of BBA4 and BP1 at pH 4 and 8.

M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
1997
structure. Six amino acidsÐhistidine, lysine, ornithine,
Functional characterization of peptides CBP01-CBP18
cysteine, homocysteine (Hcy), and DabÐwere therefore
placed at position ®ve to investigate the ability of var-
ious basic residues to in¯uence the course of transami-
nation as either general acids or general bases. These
amino acids provide basic side chains with a range of
pKas close to the pH of the transamination assays,
along with a range of side chain lengths.
For peptides CBP01-CBP18, peptide mediated transa-
mination of an a-keto acid to the corresponding a-
amino acid was probed as described for peptide BP1,
except that initial amino acid production (less than 10%
of a turnover) and optical induction information were
determined from duplicate analyses of a single time-
point (Table 1).
The placement of aspartic acid, glycine, and Dab, at
position seven was also evaluated. This position ¯anks
the C-terminal side of Pam on the exterior face of the b-
sheet and the electrostatic, steric, and conformational
di€erences of these residues was expected to perturb the
relative geometry of the coenzyme and the basic residue
at position 5. However, these substitutions were pre-
dicted to maintain the supersecondary structure of the
motif, because non-hydrophobic residues had pre-
viously been shown to be tolerated at this site.18 Toge-
ther the peptides with substitutions at positions ®ve and
seven comprised the 18 peptides CBP01-CBP18.
Peptides CBP01-CBP18 demonstrated a range of func-
tional behavior in the metal ion-free transamination of
pyruvic acid to alanine. Relative to dPam assayed under
identical conditions, all peptides showed at least a
threefold increase in the amount of alanine produced in
20 h, and peptide CBP13 exhibited a 5.6-fold increase.
For all peptides a slight optical induction favoring l-
alanine was observed, while no signi®cant optical
induction was observed for dPam under the same assay
conditions. The level of optical induction varied from a
slight excess of l-alanine to 27% ee for peptide CBP14.
The advantages of the small peptide sca€old approach
allowed the above designs to be realized rapidly. An
ecient technique for the parallel synthesis and pur-
i®cation of peptides CBP01-CBP18 has recently been
developed.28 The common carboxy-terminal residues 8±
23 were assembled on resin by solid-phase peptide
synthesis using an automated peptide synthesizer. The
resin was then split into 18 smaller aliquots and the
remaining residues were batch-coupled in parallel.
Truncation peptides were observed following Pam cou-
pling. To avoid time consuming HPLC puri®cation,
peptides CBP01-CBP18 were puri®ed in parallel using a
reversible biotin anity tag puri®cation method devel-
oped for this purpose. This technique rapidly a€orded
N-terminal glycolate-capped full-length peptides in 80%
or greater purity and in quantities sucient for a num-
ber of transamination assays.
Interesting trends in stereoselectivity were seen for
alanine production with peptides CBP01-CBP18. In
the absence of metal ions, changing either residue at
position ®ve or seven produced independent trends (Fig.
5(a)). At position ®ve, lysine gave the greatest optical
induction, and in general longer side chain lengths
resulted in greater optical inductions. At position seven,
aspartic acid and glycine a€orded similar degrees of
optical induction, while inclusion of Dab produced
enhanced optical induction.
In the presence of metal ions, coenzyme-substrate che-
lates are known to form,27 and di€erent functional
properties were observed. The presence of copper(II)
ion resulted in augmented production of alanine for
peptides CBP01-CBP18 relative to the dPam model
compound in the presence of copper(II) ion. All peptides
Table 1. Sequence substitutions at positions ®ve and seven for peptides CBP01-CBP18 relative to peptide BP1. Comparison of initial product
formation and enantiomeric induction with pyruvate
Pyv, d-Phe, EDTA
Pyv, d-Phe, Cu(II)
Peptide
Position 5
Position 7
Ala/dPam
% ee l-Ala
Ala/dPam
% ee d-Ala
CBP01
CBP02
CBP03
CBP04
CBP05
CBP06
CBP07
CBP08
CBP09
CBP10
CBP11
CBP12
CBP13
CBP14
CBP15
CBP16
CBP17
CBP18
His
Lys
Orn
Dab
Cys
Hcy
His
Lys
Orn
Dab
Cys
Hcy
His
Asp
Asp
Asp
Asp
Asp
Asp
Gly
Gly
Gly
Gly
Gly
Gly
Dab
Dab
Dab
Dab
Dab
Dab
3.4
3.5
4.4
3.5
3.0
3.0
3.7
3.8
4.1
4.2
3.9
3.0
5.6
4.7
3.8
4.8
4.8
4.1
10
14
10
8
24.4
23.9
25.4
21.3
23.3
24.6
23.9
23.4
25.8
26.2
27.3
21.7
31.7
24.4
20.1
26.4
25.1
22.1
8
15
19
20
10
7
22
27
28
37
31
21
3
4
10
11
18
12
2
7
16
24
27
23
16
8
Lys
Orn
Dab
Cys
Hcy
5
7
14
13
6
17

1998
M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
To support this proposal, the structural features of
peptide CBP14, the CBP peptide with the greatest opti-
cal induction in the absence of copper(II) ion, were
investigated by CD and 2-D NMR spectroscopy in a
manner similar to peptide BP1.
CD spectral analysis of CBP14 revealed spectral prop-
erties similar to peptide BP1 and previously studied
BBA peptides (Fig. 6). At both pH 4 and 8, peptide
CBP14 exhibited an absolute minimum of mean residual
elipticity at 205 nM with an equal magnitude. Addi-
tionally, at both pH 4 and 8 a minimum was observed at
222 nM, with a 10% greater absolute mean residual
elipticity at pH 4 relative to pH 8. While the absolute
mean residual elipticities at 205 and 222 nM were
greater than those observed for either BP1 or BBA4,
other BBA peptides shown to have bba-structure have
varied in the magnitude of the observed mean residual
elipticities.18 Thus, the CD spectral data suggest that
peptide CBP14 shares secondary structural features with
peptides BP1 and BBA4.
The 2-D NMR spectrum of CBP14 at pH 8 also sug-
gests that peptide CBP14 adopts the supersecondary
structure of the bba motif peptides (Fig. 7). Like BBA4
and BP1, the aliphatic to aromatic portion of the
NOESY spectrum of peptide CBP14 shows NOE cross-
peaks between aliphatic and aromatic side chains.
Because these crosspeaks are only expected for a species
in which the sheet and helix domain are packed, the
presence of these crosspeaks suggests that CBP14
adopts the bba motif supersecondary structure.
Figure 5. Observed optical induction in the production of alanine (a)
in the absence of metal ion and (b) in the presence of copper(II) ion for
peptides CBP01-CBP18. Peptide substitutions at position ®ve and
seven are indicated.
Discussion
The BBA peptide as a structural sca€old
Structural characterization of peptide BP1 by CD and
NMR spectroscopy indicated that the peptide exhibits
hallmark features of bba motif secondary and super-
secondary structure. Thus, it appears that a BBA motif
can accommodate Pam and a basic residue while main-
taining the desired bba motif structure that has the
showed at least a 20-fold increase in product yield rela-
tive to dPam, and peptide CBP13 demonstrated a 31.7-
fold increase in alanine production.
Additionally, the introduction of copper(II) ions resul-
ted in dramatic and di€erent stereochemical trends (Fig.
5(b)). In contrast to the metal ion-free conditions, d-
alanine production was favored in the presence of a
stoichiometric amount of divalent copper ion. The
degree of optical induction ranged from less than 5% up
to 37% enantiomeric excess of d-alanine for peptide
CBP10. Furthermore, for the basic residues at position
®ve, the greatest optical induction was produced by
Dab, and in general shorter side chain lengths favored
larger optical inductions. Peptides with glycine at posi-
tion seven, resulted in the highest stereoselectivity, while
those with Dab resulted in the lowest stereoselectivity.
Structural characterization of peptide CBP14
The trends in optical induction observed with peptides
CBP01-CBP18 suggest that the pyridoxamine function-
ality might exist within a de®ned peptide environment.
Figure 6. CD spectra of CBP14 at pH 4 and 8.

M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
1999
observed increase in amino acid production appears not
to result simply from incorporating the coenzyme func-
tionality into a peptide backbone and may instead be
due to the environment created by the bba motif. The
extent to which the full BBA peptide structure is
required for the observed functional e€ects has not been
determined.
The BBA peptide motif also has the ability to in¯uence
the stereochemical outcome of the transamination reac-
tion. In the absence of metal ions, l-alanine is the
favored transamination product, while in the presence
of copper(II), the opposite enantiomer is favored. Cop-
per(II) ion is known to be coordinated by the transami-
nation intermediates and has the potential to accept
other ligands. The di€erence in in¯uencing the stereo-
chemical outcome of transamination suggests that the
BBA peptide motif has the capacity to interact di€er-
ently with the metalated and unmetalated form of the
coenzyme functionality. Furthermore, this optical
induction is modulated by substitutions in the peptide
sequence. Both in the presence and absence of divalent
copper sequence changes at position seven resulted in
di€erent levels of optical induction independent of the
nature of the basic residue at position ®ve. Moreover,
the identity of the basic residue at position ®ve resulted
in discernible trends. In the absence of metal ions,
longer side chain residues resulted in increased produc-
tion of l-alanine, with lysine producing the greatest
degree of optical induction. In the presence of cop-
per(II) ion, this trend is reversed, with basic residues
with shorter side chains resulting in increased produc-
tion of d-alanine, with the Dap residue producing the
greatest increase. Modeling of the CBP peptides shows
that the side chains of the basic residue at position ®ve
can interact with the pyridoxamine functionality. While
the extent to which the full BBA peptide structure is
required for the observed functional e€ects has not yet
been determined, the observed trends do indicate that
changes in the coenzyme environment can result in
measurable functional di€erences between peptides.
Figure 7. NOE side chain contacts for peptide CBP14 in the aliphatic
to aromatic region of the NMR spectrum at pH 8. Tentative NOE
assignments based on BBA4. Chemical shifts are in ppm.
potential to support a de®ned coenzyme environment.
That such substitutions would be tolerated within a
small designed bba motif peptide such as peptide BBA4
was not initially obvious: incorporation of the Pam
residue within the helical region of the motif (His17 to
Pam) resulted in a peptide with a CD spectrum that was
sensitive to changes in pH, while introduction of Pam
into the hydrophobic core of peptide BBA4 (Phe8 to
Pam) resulted in a peptide with a pH-independent CD
spectra, but lacked NOEs characteristic of super-
secondary packing (data not shown).
Based on the structural characteristics of prototype
peptide BP1, on the clear functional trends in optical
induction for the transamination of pyruvate, and on
the CD and NMR analysis of peptide CBP14, it appears
that peptides CBP01-CBP18 may also possess the gross
supersecondary structure of the bba motif while pos-
sessing a coenzyme environment that in¯uences transa-
mination. At this point the full structural features of the
Pam containing BBA peptides has not been determined.
The BBA peptide as a functional sca€old
Conclusion
The BBA peptide motif has proven to be a rugged plat-
form for in¯uencing pyridoxamine mediated transami-
nation. Because of the synthetic and modular nature of
the BBA peptides, these constructs provide rapid access
to a range of coenzyme environments.
Peptides BP1 and CBP01-CBP18 constitute the ®rst use
of a bba motif peptide to generate functional peptides
and this study demonstrates the potential utility of the
sca€old. Additionally, these results highlight the ¯ex-
ibility of using small peptide constructs to in¯uencing
coenzyme-dependent catalysis. Structural characteriza-
tion of peptides BP1 and CBP14, in addition to func-
tional characterization of peptides CBP01-CBP18,
suggest that amino acid residues can be incorporated
into a BBA peptide with conservation of the overall
BBA peptide structure. Furthermore, at least some ele-
ments of the BBA structure appear to provide a unique
environment for the coenzyme to generate a range of
functional properties within the structured motif.
Further structural and functional characterization of
these peptides will clarify the source of these properties.
This information, in conjunction with the techniques made
possible by the small peptide approach, should allow for
The BBA peptide motif has the capacity to augment the
production of amino acid above the coenzyme alone. In
the presence of copper(II) ion, peptides CBP01-CBP18
demonstrate increases in the amount of alanine pro-
duced during the assay period relative to the dPam
control under the same conditions by at least 20-fold,
with peptide CBP13 showing a 31.7-fold increase. Pre-
vious studies under similar copper(II) ion conditions
with a Pam containing peptide designed to be unstruc-
tured (Ac-Thr-30Pyridylalanine-Gly-Gly-Pam-Gly-NH2)
showed an increase in initial rate of alanine production
of only 1.8-fold (unpublished results). Thus, the

2000
M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
the continued exploration and improvement of small
systems that mediate rapid and selective transamination.
the peptide resin to 20% piperidine in DMF for 12 h
resulted in full deacetylation without any isomeric pro-
ducts. Resin was subsequently washed thoroughly with
DMF and DCM and lyophilized to dryness. Peptide
was cleaved from the resin with an excess of reagent K
(tri¯uoroacetic acid/phenol/H2O/thioanisole/ethane-
dithiol, 82.5/5/5/5/2.5) for 2 h to a€ord the deprotected
peptide.30 Excess TFA was removed with nitrogen gas
and was triturated with 2/1 ether/hexane. The resulting
white pellet was resuspended in water and further pur-
i®ed by reverse-phase HPLC. Peptides CBP01-CBP18
were rapidly puri®ed in parallel using an avidin anity
column method described elsewhere.28
Experimental
Materials
Na-9-Fluorenylmethyloxycarbonyl (Fmoc) protected
amino acids and amino acid reagents were purchased
from Millipore or PerSeptive Biosystems. The Pam
residue was synthesized as previously described.29 PAL-
PEG-PS resin was purchased from PerSeptive Biosys-
tems. The Vac-man vacuum system was purchased from
Promega and the disposable polypropylene columns
and valves were obtained from BioRad.
Peptide quanti®cation
Puri®ed peptide stocks were quanti®ed by UV±vis spec-
troscopy under denaturing conditions using a Beckman
DU 7500 Spectrophotometer in a 1 cm path length
quartz cuvette. Typically, peptide stock was mixed with
an equal volume of a 4.8 M guanidine hydrochloride
solution bu€ered at pH 7.5 with 0.2 M HEPPS. The
concentration of the original stock was calculated from
the absorbance at 322 nM with and e322 of
5610 M 1 cm 1, where e was initially determined from a
Pam containing peptide quanti®ed by quantitative
amino acid analysis.
Peptide modeling and synthesis
Peptide design and modeling was performed using
Insight II software (Biosym Technologies, San Diego,
CA, USA).
Peptide BP1, the parent batch of resin for peptides
CBP01-CBP18 (residues 8±23), and peptide CBP14 were
made by solid-phase peptide synthesis performed on a
Milligen 9050 automated peptide synthesizer. Peptides
were synthesized on a 200 mmol scale from PAL-PEG-
PS resin (1 g, 0.2 mmol/g). Amino acids were double
coupled in fourfold excess, except for Pam, which was
double coupled at 2.5-fold and 1.5-fold excess. Peptide
coupling was performed with HOBt/HBTU chemistry,
with the exception of Pam, which was coupled by HOAt/
HATU chemistry. The Fmoc-amines were deprotected
using a 20% piperidine/DMF wash. Between each new
residue coupling, free amines were capped with a 0.3 M
acetic anhydride/HOBt solution in 9/1 DMF/DCM.
Circular dichroism studies
CD experiments were performed on a Jasco J-600 spec-
tropolarimeter or an Aviv 62DS Circular Dichroism
Spectrometer. CD samples were prepared by dilution of
peptide stocks to concentrations between 50 and
200 mM in degassed water for experiments performed at
pH 4 or in 0.01 M pH 8 Tris bu€er for experiments
performed at pH 8. The pH of the sample was adjusted
to either 4 or 8 by addition of 0.01 N NaOH. CD mea-
surements were performed at room temperature in a
0.1 cm path length quartz cell with the optical chamber
continually ¯ushed with dry N2 gas. Scans were col-
lected in the range 195±250 nM, with a band width of
1.5±2.0 nM, a sensitivity of 50 mdeg, a scan speed of
50 nM/min, a time constant of 0.5 s, and a scan resolu-
tion of 0.2 nM. Four accumulations per sample were
performed. CD spectra were analyzed in KaleidaGraph
software, version 2.11 and are reported in (degÂcm2)/
(dmolÂamide bonds).
Residues 1±7 of peptides CBP01-CBP18 were synthe-
sized in parallel using batch solid-phase peptide synth-
esis techniques. The parent batch of resin was divided in
18Â50 mg (6.4 mmol theoretical) portions. All residues
were double coupled with 4 equiv amino acid (0.256 M),
4 equiv HOBt (0.256 M), 4 equiv HBTU (0.256 M) and
9 equiv diisopropylethylamine (0.576 M) in 89 mL DMF
(Pam was coupled with a 2.5/1.5-fold excess with
HOAt/HATU chemistry) for 2 h. Coupling of the N-
terminal reversible biotin anity tag was accomplished
under conditions described elsewhere.28 The extent of
residue coupling was tested by Kaiser test on one of the
18 peptides after each coupling round. Acetylation steps
were performed for 30 min using the acetylation solu-
tion described earlier. Deblocking steps were performed
for 30 min with a 20% piperidine/DMF solution. The
resin was washed with 20 mL of DMF between reaction
steps and was occasionally agitated manually during the
coupling step.
NMR studies
2-D NMR experiments were performed on a 500 MHz
Bruker Instrument AMX500 spectrometer or a Varian
600 MHz Unity Plus instrument. NMR samples were
prepared in either a 90/10 H2O/D2O solution or a
99.98% D2O solution with dioxane or dimethyl sulfox-
ide as an internal standard. Peptide concentration ran-
ged from 2±10 mM. The pH was adjusted to 8 with
small additions of 0.1 N NaOH or 0.1 N NaOD.
Nuclear Overhauser e€ects (NOEs) were detected using
spin-locked rotating-frame nuclear Overhauser e€ect
spectroscopy (ROESY) with a 400 ms mixing time,31 or
nuclear Overhauser e€ect spectroscopy (NOESY) with a
Peptides BP1, CBP01-CBP18, and CBP14 were deace-
tylated prior to acid cleavage to remove an acetyl group
from the phenolic oxygen of the Pam residue.29 Apply-
ing the previously described deacetylation conditions29
to peptide BP1 resulted in isomeric products. Subjecting

M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
2001
250 ms mixing time.32 Water suppression was obtained
using presaturation during the relaxation delay. Spectra
were processed and analyzed in Felix 95 (Biosym Tech-
nologies, San Diego, CA, USA) and NOESY spectra
were baseline corrected.
curves of d,l amino acid standards generated over a
range of concentrations. For peptide BP1, the initial
reaction rate was determined by ®tting the total amino
acid product for each time-point to a pseudo ®rst order
exponential build-up of product. From this ®t, the
initial rate was extracted. For peptides CBP01-CBP18,
the extent of reaction was calculated from the total
quantity of amino acid produced. The enantiomeric
excess was calculated from the di€erence in quantity of
amino acid enantiomers produced relative to the total
quantity of amino acid produced. All samples were
performed in duplicate.
Transamination assays
For peptide BP1, a typical assay was performed as fol-
lows: to a small Eppendorf was added 2 mL of a 10 mM
EDTA solution (100 mM), a 10 mL of a 2 mM BP1 stock
(100 mM), and 148 mL of a pH 8, 0.1 M HEPPS bu€er
with a m of 0.29, 0.02% NaN3. The solution was equili-
brated at room temperature for 10 min, and the reaction
was initiated by the addition of 40 mL of a fresh 50 mM
sodium pyruvic acid/50 mM d-phenylalanine solution
(10 mM of each). Assays proceeded at room tempera-
ture and over 24 h a number of 10 mL time-points were
collected. For peptides CBP01-CBP18 the total volume,
temperature, and order of addition were the same.
However, EDTA or CuCl2 and peptide were added to a
®nal concentration of 10 mM. Reactions were initiated
with 40 mL of a 50 mM sodium pyruvic acid/50 mM d-
phenylalanine solution being added (10 mM of each)
and were stopped before reaching 10% of a single turn-
over. Reactions with copper(II) ion were quenched after
1 h with a ®vefold excess EDTA. Metal free assays were
stopped after 20 h. All samples not derivatized immedi-
ately were ¯ash frozen and stored at 80ꢀC. Transami-
nation assays for the dPam model compound were
performed both in the presence and absence of cop-
per(II) ion under the conditions described for peptides
BP1 and CBP01-CBP18.
Acknowledgements
This work was supported by NIH grant number GM
53098. Additional support for M.A.S. was provided by
a predoctoral fellowship from the Howard Hughes
Medical Institute.
References
1. Fields, G. B.; Noble, R. L. Int. J. Peptide Protein Res. 1990,
35, 161.
2. Kamp, R. M.; Choli-Papadopoulou, T.; Wittman-Liebold,
B. In Protein Structure Analysis: Preparation, Characteriza-
tion, and Microsequencing; Springer-Verlag: New York, 1997.
3. Fasman, G. D. In Circular Dichroism and the Conforma-
tional Analysis of Biomolecules; Plenum: New York, 1996.
4. Wuthrich, K. In NMR of Proteins and Nucleic Acids; Wiley-
Interscience: New York, 1986.
5. Smith, P. A. S. In The Chemistry of Open-Chain Organic
Nitrogen Compounds; W. A. Benjamin: New York, 1965.
6. Akhtar, M.; Emery, V. C.; Robinson, J. A. In The Chem-
istry of Enzyme Action; Page, M. I., Ed.; Elsevier Science:
Amsterdam, 1984.
The amino acids generated in the assay were derivatized
with saturated sodium borate bu€er and freshly pre-
pared N-acetyl cysteine (5 mg/mL)/OPA incomplete
solution from Sigma according to literature proce-
dure.25,26 To derivatize samples from BP1, 20 mL of the
saturated sodium borate bu€er was added followed by
40 mL of the OPA solution. The reaction mixture was
vortexed and then allowed to sit for exactly 5 min at
room temperature, before one half of the reaction mixture
was injected onto the HPLC (36 mL). For samples from
peptides CBP01-CBP18, 100 mL of sodium borate bu€er
and 500 mL of the OPA solution was added, mixed, and
allowed to react for 5 min at room temperature before
the full volume of the sample was injected.
7. Hayashi, H. J. Biochem. 1995, 118, 463.
8. Tabushi, I.; Yasuhisa, K.; Yamada, M.; Higashimura, H.;
Breslow, R. J. Am. Chem. Soc. 1985, 107, 5545.
9. Breslow, R.; Canary, J. W.; Varney, M.; Waddell, S. T.;
Yang, D. J. Am. Chem. Soc. 1990, 112, 5212.
10. Kikuchi, Q.; Zhang, Z.; Murakami, Y. J. Am. Chem. Soc.
1995, 117, 5383.
11. Ando, M.; Tachibana, Y.; Kuzuhara, H. Bull. Chem. Soc.
Jpn. 1982, 55, 829.
12. Breslow, R.; Czarnik, A. W.; Lauer, M.; Leppkes, R.;
Winkler, J.; Zimmerman, S. J. Am. Chem. Soc. 1986, 108, 1969.
13. Kuang, H.; Distefano, M. D. J. Am. Chem. Soc. 1998, 120,
1072.
14. Sinha Roy, R.; Imperiali, B. Protein Eng. 1997, 10, 691.
15. Sinha Roy, R.; Imperiali, B. J. Am. Chem. Soc. 1994, 116,
12083.
16. Struthers, M. D.; Cheng, R. P.; Imperiali, B. Science 1996,
271, 342.
17. Struthers, M. D.; Cheng, R. P.; Imperiali, B. J. Am. Chem.
Soc. 1996, 118, 3073.
18. Struthers, M.; Ottesen, J. J.; Imperiali, B. Folding Des.
1998, 3, 95.
Resolution of derivatized products was performed using
a Beckman HPLC equipped with either a 5 m 4.6
Â150 mM Beckman or 5 m 4.6Â250 mM Microsorb C18
Reverse-Phase Column. Methanol and a solution of
80 mM sodium citrate, 20 mM sodium phosphate, with
0.02% sodium azide at pH 6.8 were employed as co-
solvent. Gradients were optimized for each reaction
condition tested. The ¯uorescent signals were detected
on a Waters 470 scanning ¯uorescence detector with an
excitation wavelength of 344 nM and an emission
wavelength of 443 nM.
19. Smith, D. L.; Almo, S. C.; Toney, M. D.; Ringe, D. Bio-
chemistry 1989, 28, 8161.
20. Malcolm, B. A.; Kirsch, J. F. Biochem. Biophys. Res.
Commun. 1985, 132, 915.
21. Toney, M. D.; Kirsch, J. F. Science 1989, 243, 1485.
22. Toney, M. D.; Kirsch, J. F. Biochemistry 1993, 32, 1475.
23. Dunathan, H. C.; Davis, L.; Gilmer Kury, P.; Kaplan, M.
Biochemistry 1968, 7, 4532.
Quantities of the amino acid enantiomers produced
were calculated by integration of peaks from calibration

2002
M. A. Shogren-Knaak, B. Imperiali / Bioorg. Med. Chem. 7 (1999) 1993±2002
24. Ayling, J. E.; Dunathan, H. C.; Snell, E. E. Biochemistry
1968, 7, 4537.
29. Sinha Roy, R.; Imperiali, B. Tetrahedron Lett. 1996, 37,
2129.
25. Nimura, N.; Kinoshita, T. J. Chromatogr. 1986, 352, 169.
26. Buck, R. H.; Krummen, K. J. Chromatogr. 1987, 387, 255.
27. Martell, A. E. Acc. Chem. Res. 1989, 22, 115.
28. Shogren-Knaak, M. A.; Imperiali, B. Tetrahedron Lett.
1998, 39, 8241.
30. King, D. S.; Fields, C. G.; Fields, G. B. Int. J. Pept. Pro-
tein. Res. 1990, 36, 255.
31. Bax, A.; Davis, D. G. J. J. Magn. Reson. 1985, 63, 207.
32. States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn.
Res. 1982, 63, 207.