Induction of protein-like molecular architecture by self-assembly processes

Article abstract of DOI:10.1016/S0968-0896(98)00216-8

One of the most intriguing self-assembly processes is the folding of peptide chains into native protein structures. We have developed a method for building protein-like structural motifs that incorporate sequences of biological interest. A lipophilic moie

Full text of DOI:10.1016/S0968-0896(98)00216-8

Bioorganic & Medicinal Chemistry 7 (1999) 75±81
Induction of Protein-like Molecular Architecture by
Self-assembly Processes
Gregg B. Fields
Department of Chemistry & Biochemistry and the Center for Molecular Biology & Biotechnology,
Florida Atlantic University, Boca Raton, FL 33431 USA
Received 7 May 1998; accepted 15 September 1998
AbstractÐOne of the most intriguing self-assembly processes is the folding of peptide chains into native protein structures. We have
developed a method for building protein-like structural motifs that incorporate sequences of biological interest. A lipophilic moiety
a
is attached onto an N -amino group of a peptide chain, resulting in a `peptide-amphiphile'. The alignment of amphiphilic com-
pounds at the lipid±solvent interface is used to facilitate peptide alignment and structure initiation and propagation. Peptide-
amphiphiles containing potentially triple-helical structural motifs have been synthesized. The resultant head group structures have
been characterized by circular dichroism and NMR spectroscopies. Evidence for a self-assembly process of peptide-amphiphiles has
been obtained from: (a) circular dichroism spectra and melting curves characteristic of triple-helices, (b) one- and two-dimensional
NMR spectra indicative of stable triple-helical structure at low temperatures and melted triple-helices at high temperatures, and (c)
pulsed-®eld gradient NMR experiments demonstrating di€erent self-di€usion coecients between proposed triple-helical and non-
triple-helical species. The peptide-amphiphiles described here provide a simple approach for building stable protein structural
motifs using peptide head groups. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
which could initiate and stabilize structure formation.
For example, lattice statistical mechanics has predicted
that surfaces (i.e., templates) would enhance secondary
The creation of protein-like molecular architecture is
desired for studies of protein folding and stability and
the design of proteins geared for speci®c functions.
Many initial studies focused on the ability of peptides to
self-assemble and form protein-like structures. For
example, in 1968 Sakakibara et al. demonstrated syn-
3
,4
structure formation by minimizing unfolded states.
Surfaces have been shown to induce peptide secondary
5±7
structures, either a-helical or b-sheet.
The template
e€ect has been documented for synthetic four a-helical
bundles, as the a-helicity of the individual peptide
chains was greatly enhanced upon incorporation onto a
thetic peptides of the sequence (Pro-Pro-Gly) , where
n
n=10 or 20, could self-assemble to form collagen-like
1
template containing several other a-helical peptide
sequences.
protein-like structures is to use directed self-assembly to
create a peptide-template.
8±10
triple-helices. Peptide self-assembly has since been used
2
Thus, one method for creating distinct
to create a-helical bundle and b-sheet proteins.
Unfortunately, peptide self-assembly is not a particu-
larly easy process to regulate. Nondesired molecular
assemblies may occur (such as formation of a mixture of
a-helical dimers, trimers, tetramers, etc.) or peptide
chains may simply form large aggregates of unknown
order. Controlling peptide interaction could minimize
these diculties. One approach for controlled peptide
assembly is to incorporate moieties on the end of pep-
tide chains, and use properties of these moieties to drive
peptides to interact in a speci®c fashion. The advantages
of this approach are several. The proper number of
peptides will associate, eliminating undesired aggre-
gates. The moieties themselves may act as templates,
One well documented controlled self-assembly approach
has utilized chelators attached to peptides and appro-
priate metals to initiate interstrand association. A triple
a-helical bundle protein has been created by incorpor-
0
ating 5-carboxy-2,2 -bipyridine onto the 15-residue
peptide Gly-Glu-Leu-Ala-Gln-Lys-Leu-Glu-Gln-Ala-Leu-
Gln-Lys-Leu-Ala-NH2 and using Co(II), Ni(II), or
Ru(II) complexation.11 The trimeric bundle was con-
®rmed by size-exclusion chromatography and mass
spectrometry (MS). The circular dichroism (CD) spec-
trum of the Ru(II) complexed-peptide indicated a left-
handed supercoiled a-helical structure. GdnHCl dena-
turation experiments gave free energy of folding values
of ÁG(H O) =� 1.86 kcal/mol and ÁGꢀH O† =
Key words: Protein folding; secondary structure; peptide-amphiphile;
collagen.
2
Ru
2
Ni
� 1.93 kcal/mol. A triple a-helical bundle protein has
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(98)00216-8

7
6
G. B. Fields / Bioorg. Med. Chem. 7 (1999) 75±81
0
0
a collagen-model peptide head group.15,16 Construction
of tail compounds ®rst involves the acid-catalyzed
condensation of Glu with the appropriate fatty acid
alcohol to form the dialkyl ester of Glu (p-toluenesulfo-
nate salt).15 The free amino group of Glu is then treated
with succinic anhydride and triethylamine to create a free
carboxylic acid. The dialkyl ester tail may then be acylated
to the resin-bound peptide using N-[(1H-benzotriazol-1-
also been created by incorporating 2,2 -bipyridine-4,4 -
dicarboxylic acid onto the 15 residue peptide Ala-Glu-
Gln-Leu-Leu-Gln-Glu-Ala-Glu-Gln-Leu-Leu-Gln-Glu-
2
+ 12
Leu-NH and complexing with Fe
2
.
The individual
strands, designated `pepy', were 35% a-helical, while the
resultant protein [Fe (pepy) ] was 85% a-helical. Fe
pepy) was found to exist in four stereoisomeric states.
II
II
3
1
3
(
3
A Ru(II)-complexed four a-helix bundle has been con-
structed by incorporating a pyridyl functionality onto
the N-terminus of the 15 residue helix-forming peptide
Gly-Leu-Ala-Gln-Lys-Leu-Leu-Glu-Ala-Leu-Gln-Lys-
Ala-Leu-Ala-NH2. The assembled four-helix bundle
was con®rmed by size-exclusion chromatography and
MS. The complex was >90% a-helical by CD spectro-
scopy. GdnHCl-induced denaturation of the bundle was
yl)(dimethylamino)methylene]-N-methylmethanaminium
hexa¯uorophosphate N-oxide (HBTU)1
6,17
or converted
to a 4-nitrophenyl ester and used for coupling.15
1
4
The peptide-amphiphile solid-phase method was based
upon 9-¯uorenylmethoxycarbonyl (Fmoc) chemistry.
The resin-bound peptide composition (before addition
of the lipophilic tail) was con®rmed by Edman degra-
dation sequence analysis.21 Couplings involving dialkyl
ester tails did not appear to be particularly sterically
hindered; a typical reaction required 4 h. Mild peptide-
resin cleavage conditions using tri¯uoroacetic acid
(TFA) did not a€ect the tail integrity. Following clea-
vage, care was required during ether extraction of sca-
vengers, as long chain dialkyl ester peptide-amphiphiles
could enter the ether layer. Puri®cation of dialkyl ester
peptide-amphiphiles was best accomplished by C4
reversed-phase high performance liquid chromato-
highly cooperative, with ÁG(H O)=� 5.6 kcal/mol.
2
Rather than using chelators and metal complexation,
we envisioned the construction of a novel peptide-
amphiphile for directed self-assembly. A peptide `head
group' would have the propensity to form a distinct
structural element, while a hydrophobic `tail' would
serve to align the peptide strands and induce secondary
and tertiary structure formation as well as providing a
hydrophobic surface for self-association. Our initial
e€orts have focused on this noncovalent, self-assembly
approach for building
1
a
collagen-like structural
The present article describes the synthesis
graphy (RP-HPLC) using a gradient of H O±iso-
2
5±19
16
motifs.
propanol containing 0.05% TFA. Peptide-amphiphile
homogeneity was evaluated by C18 RP-HPLC using a
and characterization of peptide-amphiphile collagen-
like molecular architecture.
1
5
gradient of H O±acetonitrile containing 0.05% TFA.
2
Final characterization of the puri®ed peptide-amphi-
1
5
phile was achieved readily by NMR spectroscopy and
matrix-assisted laser desorption/ionization (MALDI)
mass spectrometry.16
Results and Discussion
Synthesis of peptide-amphiphiles
Subsequent studies have used monoalkyl chains to cre-
ate peptide-amphiphiles (Fig. 1). Synthesis is simpli®ed
greatly, as commercially available monoalkyl acids are
coupled directly to peptide-resins.18 These peptide-
amphiphiles have no linker or spacer per se. We have
created a series of C , C , C , C , C , and C mono-
alkyl peptide-amphiphiles. Monoalkyl peptide-amphi-
philes are less prone to losses during either the ether
extraction step or RP-HPLC due to the decreased
hydrophobicity of this class of compounds compared to
dialkyl ester peptide-amphiphiles.
The design of peptide-amphiphiles requires that long
chain lipophilic moieties associate by hydrophobic for-
ces to drive structure initiation in the peptide portion of
the molecule and/or stabilize protein-like structure. We
initially examined the use of dialkyl ester chains for
creating peptide-amphiphiles (Fig. 1). As described by
6
8
10
9
12
14
16
1
2
0
Kunitake, design of synthetic lipids requires con-
sideration of four building blocks: tail, linker/connector,
spacer, and head group. Our design featured C , C ,
1
2
14
and C16 dialkyl tails, a Glu linker, a ±(CH ) ± spacer, and
2 2
Figure 1. General structures of dialkyl ester and monoalkyl peptide-amphiphiles.

G. B. Fields / Bioorg. Med. Chem. 7 (1999) 75±81
77
Our peptide-amphiphiles have been constructed using
mono- or dialkyl ester tail compounds. Other research
groups have created peptide-amphiphiles using tail com-
magnitude of positive ellipticity at l=225 nm was
observed, while the IV-H1 peptide did not show any
positive ellipticity at this wavelength.1
6,17
Of the C12
2
2,23
24±27
pounds such as dialkyl amides,
¯
phospholipids,
uorescent phospholipids,28 and tripalmitoyl-Cys
dialkyl chain peptide-amphiphiles, (C ) -Glu-C -[IV-
12 2 2
H1] displayed a CD spectrum similar to that of IV-H1
(no positive ellipticity at l=225 nm), while (C ) -Glu-
2
9
[
been incorporated by solution-phase
(Pam) Cys]. Lipophilic compounds have typically
or che-
methods. To our knowledge,
3
12 2
23±25,28±30
C -(Gly-Pro-Hyp) -[IV-H1], (C ) -Glu-C -[IV-H1]-(Gly-
2
2
4
12 2
moselective ligation2
only one laboratory has reported the incorporation of a
6,27
Pro-Hyp) , and (C ) -Glu-C -(Gly-Pro-Hyp) -[IV-H1]-
4 12 2 2 4
(Gly-Pro-Hyp) showed a large magnitude of positive
4
16,17
branched lipid (a 1,2-dimyristoyl-sn-glycerol derivative)
ellipticity at l>220 nm.
Similarly, the monoalkyl
peptide-amphiphile C -(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-
3
1
onto a peptide by solid-phase methods. The hydro-
phobic nature of peptide-amphiphiles has resulted in a
variety of approaches being utilized for compound pur-
i®cation and characterization. These include traditional
puri®cation methods, such as Sephadex LH-20 gel ®l-
6
4
Hyp) exhibited a CD spectra typical of triple-helical
4
ꢁ
19
conformation at 25 C. The ellipticity per residue
values indicate a maximal ordered structure for (C ) -
Glu-C -(Gly-Pro-Hyp) -[IV-H1], (C ) -Glu-C -[IV-H1]-
1
2 2
2
4
12 2
2
2
4,29,31
tration chromatography,
sure liquid chromatography, normal phase liquid
silica gel medium pres-
(Gly-Pro-Hyp) , (C ) -Glu-C -(Gly-Pro-Hyp) -[IV-H1]-
4 12 2 2 4
(Gly-Pro-Hyp) , and C -(Gly-Pro-Hyp) -[IV-H1]-(Gly-
4 6 4
Pro-Hyp) . It appears that all residues in these peptide-
4
amphiphiles are in triple-helical conformation.
2
3
2
6
28,30
chromatography, or extraction/recrystallization.
Peptide-amphiphiles containing short peptide head
groups (5±7 amino acids) have been puri®ed by HPLC
using cyanopropyl, C , or C columns. We found
that for longer peptide head groups and either dialkyl or
2
5,31,32
A triple-helical assembly can be distinguished from a
simple, non-intercoiled poly-Pro II structure by its
thermal denaturation behavior. Triple-helical melts are
highly cooperative.35 The thermal transitions of pep-
tides and peptide-amphiphiles were monitored by mea-
suring molar ellipticity at l=225 nm as a function of
increasing temperature. Amongst the peptides, only
(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp) gave a typical
8
18
monoalkyl tails, C RP-HPLC puri®cation was e€ec-
4
tive. The mobile phase required isopropanol instead of
2
5,32
acetonitrile used with other peptide-amphiphiles.
Peptide-amphiphiles have been characterized primarily
by fast-atom bombardment (FAB) MS and/or NMR
4
4
2
3,27,29±32
spectroscopy.
spectroscopy to be an e€ective characterization
While we have found NMR
sigmoidal transition associated with the transformation
of triple-helical to single-stranded structure. The ®rst
derivative of the melting curve gave a melting tempera-
1
6
method, our peptide-amphiphiles were too large for
FAB-MS analysis. Instead, MALDI-MS was found a
reproducible characterization tool.
ꢁ
16,17
ture (T ) value of 35.6 C (Table 1).
[IV-H1]-(Gly-
Pro-Hyp) showed a small magnitude of positive ellipti-
m
16,19
4
ꢁ
6,17
city which decreased nearly linearly from 5±20 C then
attened out, as did (Gly-Pro-Hyp) -[IV-H1].
4
1
¯
These
Protein-like structure of peptide-amphiphiles
two peptides do not appear to be in a triple-helical
conformation. Thermal denaturation studies of the di-
alkyl peptide-amphiphiles showed molar ellipticities that
To evaluate the head group structures formed by pep-
tide-amphiphiles, several di€erent biophysical approa-
ches were utilized. The overall structure of the peptide-
amphiphile was examined by CD spectroscopy, the
thermal stability of peptide-amphiphiles was determined
by CD melting curves, one- and two-dimensional NMR
spectroscopy of peptide-amphiphiles was used to evalu-
ate the environment of speci®c residue side-chains, and
pulsed-®eld gradient (PFG) NMR experiments were
used to determine peptide-amphiphile di€usion coe-
cients. For our studies, the a1(IV) 1263±1277 collagen
sequence Gly-Val-Lys-Gly-Asp-Lys-Gly-Asn-Pro-Gly-
Trp-Pro-Gly-Ala-Pro (IV-H1) was combined with alkyl
chains to create potentially collagen-like peptide-
ꢁ
decreased very gradually between 10±30 C, then more
ꢁ
markedly starting at around 30±40 C, with some traces
The midpoint
ꢁ
of the transition (T ) for (C ) -(Gly-Pro-Hyp) -[IV-
16,17
of positive CD detectable up to 80 C.
m
12 2
4
ꢁ
H1]-(Gly-Pro-Hyp) was 71.2 C (Table 1). The melting
4
curve was fully reversible upon cooling. Although the
changes in ellipticity were large, thermal transitions for
Table 1.
m
T values for peptide and peptide-amphiphile triple helix
() coil transitions
ꢁ
Peptide or peptide-amphiphile
T ( C)
m
1
6,17
amphiphiles (Fig. 1).
a
[
(
[
IV-H1]
Gly-Pro-Hyp)
IV-H1]-(Gly-Pro-Hyp)
(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp)4
NTH
NTH
a
4
-[IV-H1]
CD analysis was based upon well-known properties of the
collagen triple-helix. The collagen triple-helix consists of
three polypeptide chains, each in an extended, left-handed
polyPro II-like helix, which are staggered by one residue
and then supercoiled along a common axis in a right-
a
4
NTH
4
35.6
42.2
45.6
51.3
55.0
63.1
C
C
C
C
6
-(Gly-Pro-Hyp)
-(Gly-Pro-Hyp)
4
-[IV-H1]-(Gly-Pro-Hyp)
-[IV-H1]-(Gly-Pro-Hyp)
4
8
4
4
10-(Gly-Pro-Hyp)
12-(Gly-Pro-Hyp)
4
-[IV-H1]-(Gly-Pro-Hyp)
-[IV-H1]-(Gly-Pro-Hyp)
4
4
4
3
3,34
handed manner.
formation exhibit a CD spectrum similar to a poly-Pro II
Collagens in triple-helical con-
C14-(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp)4
4
C
16-(Gly-Pro-Hyp)
4
-[IV-H1]-(Gly-Pro-Hyp)
4
69.8
71.2
35
(C ) -Glu-C -(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp)
(C12
helix, with positive ellipticity from l=215±240nm. At
12 2
2
2
2
4
4
ꢁ
to exhibit this characteristic CD spectrum.
)
)
2
-Glu-C
-Glu-C
-(Gly-Pro-Hyp) -[IV-H1]
4
-[IV-H1]-(Gly-Pro-Hyp)
ꢂ50‹5
2
5 C, (Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp) was found
1
4
4
6,17
(
C
12
2
4
ꢂ50‹5
For (Gly-
Pro-Hyp) -[IV-H1] and [IV-H1]-(Gly-Pro-Hyp) a small
aNTH=not triple-helical.
4
4

7
8
G. B. Fields / Bioorg. Med. Chem. 7 (1999) 75±81
the dialkyl peptide-amphiphiles were broad. The thermal
stability of C -(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp)
monoalkyl peptide-amphiphiles, where n=6, 8, 10, 12,
found at 4.6 ppm16 are comparable to those observed
for the triple-helical, template bound peptide Kemp
n
4
4
39
triacid-[Gly-(Gly-Pro-Hyp) -NH ] .
3
(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp) at 50 C shows
4
The spectra of
ꢁ
2 3
1
4, or 16, were similarly studied by monitoring ellipticity
4
19
at l=225 nm as a function of increasing temperature.
additional cross peaks at 4.85 ppm, indicating less
ordered conformation at higher temperature.16 Some of
these additional cross peaks are consistent with the
multiple states that exist for the Pro residues within the
IV-H1 sequence when in a non-triple-helical conforma-
tion.40 After the (Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp)
All peptide-amphiphiles gave sharp sigmoidal transi-
tions associated with the transformation of triple-helical
to single-stranded structure. The melting curves were
fully reversible upon cooling. Tm values were deter-
mined from the ®rst derivative of the melting curves,
and found to increase with monoalkyl tail chain length
4
4
peptide is lipidated with a C12 tail, similar NMR spectra
are obtained.16 For example, (C ) -Glu-C -(Gly-Pro-
(
Table 1). These observations suggest that the peptide-
12 2
2
ꢁ
Hyp) -[IV-H1]-(Gly-Pro-Hyp) at 25 C shows a few well
4 4
amphiphile structures consist of triple-helically packed
polyPro II-like helices, and that increasing lipid tail
chain length enhances the stability of this assembly.
de®ned cross peaks, indicating ordered conformation of
the peptide-amphiphile. Consistent with our CD obser-
vations, the NMR spectra of (C ) -Glu-C -(Gly-Pro-
1
2 2
2
1H NMR spectroscopy was utilized to further char-
acterize peptide and peptide-amphiphile structure. The
Pro and Hyp spin systems in total correlation spectro-
scopy (TOCSY) were identi®ed by the lack of amide
protons and reference to the chemical shifts of the side-
chain protons from other collagen-like peptides
Hyp) -[IV-H1]-(Gly-Pro-Hyp) at 80 C indicates more
4
ꢁ
4
ꢁ
at 4.85 ppm, in similar fashion to (Gly-Pro-Hyp) -[IV-
16
disorder than at 25 C. Additional cross peaks are seen
4
ꢁ
spectra of the (Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp)
H1]-(Gly-Pro-Hyp) at 50 C. Overall, the CD and NMR
4
4
4
peptide and the (C ) -Glu-C -(Gly-Pro-Hyp) -[IV-H1]-
4
12 2
2
(
side-chain protons is sensitive to their conformation.
Fig. 2).36±38 The chemical shift of the Pro and Hyp
(Gly-Pro-Hyp) peptide-amphiphile suggest that both
4
can spontaneously form a well-ordered poly-Pro II-like
triple-helical structure.
36±38
For our peptides and peptide-amphiphiles, there are
Pro residues that are found both surrounding and
within the IV-H1 sequence while Hyp residues surround
Additional one-dimensional NMR experiments were
performed in D O solution. The resonance lines from
ꢁ
the IV-H1 sequence. At 10 C, the relatively few cross
1
2
1
peaks found in the Pro/Hyp region of the H NMR
spectra indicate that the Pro and Hyp residues of (Gly-
labile protons are replaced in D O, so that the H NMR
2
spectrum will contain only the resonance lines of the
carbon-bound protons.41 In the case of Trp residues,
resonances in the 7.0±8.0 ppm region result from the
protons attached to the 2, 4, 5, 6, and 7 positions of
side-chain indole ring.41 Thus, spectra of the investi-
gated peptides and peptide-amphiphiles in the 7.0±
Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp)4 are in a limited
4
16
number of conformations (Fig. 2), as expected for a
compound with an ordered structure. The cross peaks
8
found within the IV-H1 sequence. The H NMR spec-
.0 ppm region will result from a unique Trp residue
1
ꢁ
trum of (Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp) at 40 C
4
4
showed prominent resonances at ꢂ7.70, 7.68, 7.51, 7.50,
7
.30, 7.26, and 7.18 ppm.19 The CD spectrum indicates a
1
non-triple-helical conformation. However, the H NMR
spectrum of (Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp) at
4
4
ꢁ ꢁ
particular, there are prominent signals in the 7.71±
1
0 C is substantially di€erent than that at 40 C. In
7
are not due to a simple change in temperature, as the 10 C
.78 ppm and 7.33±7.40 ppm regions.19 These resonances
ꢁ
spectrum of [IV-H1]-(Gly-Pro-Hyp) does not contain
4
such signals. Since the CD data indicate that (Gly-Pro-
ꢁ
Hyp) -[IV-H1]-(Gly-Pro-Hyp) is triple helical at 10 C,
4
4
the additional resonances appear to be derived from the
Trp side-chain in a triple-helical conformation. Increasing
ꢁ
triple-helicity of (Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp) ,
the temperature to 25 C, which results in a decrease in
4
4
causes a decrease in the intensities of the 7.71±7.78 ppm
1
and 7.33±7.40 ppm signals. The H NMR spectrum peaks
that decrease with increasing temperature may thus
represent Trp within a triple-helical conformation.
Figure 2. TOCSY spectrum of (Gly-Pro-Hyp)
ꢁ
4
-[IV-H1]-(Gly-Pro-
O at 10 C. Peptide concentration was 5 mM. For the
Hyp)
4
in D
2
1
preliminary assignment of amino acid residues present in the peptide,
data from additional TOSCY experiments (D O:H O,1:9) and
NOESY experiments were used. The Pro/Hyp region was identi®ed
The H NMR spectra of C -(Gly-Pro-Hyp) -[IV-H1]-
4
6
ꢁ
2
2
(
lines at similar chemical shifts as that of (Gly-Pro-
Gly-Pro-Hyp)4 at 10 and 25 C exhibit resonance
1
from the lack of an amide H signal and the shown connectivity pat-
tern in TOCSY. Reprinted with permission from J. Am. Chem. Soc.
ꢁ
the peptide-amphiphile to (Gly-Pro-Hyp) -[IV-H1]-(Gly-
19
Hyp) -[IV-H1]-(Gly-Pro-Hyp)4 at 10 C. Comparing
4
1
996, 118, 12518, American Chemical Society.16
4

G. B. Fields / Bioorg. Med. Chem. 7 (1999) 75±81
79
Figure 3. Model of a triple-helical peptide-amphiphile. Individual peptide strands are represented by red, blue, or black. In this model, the dialkyl
chains are tightly packed due to hydrophobic interactions, while the peptide strands initiate triple-helix formation in the proper orientation and
register. At the present time, it cannot be determined if triple-helical folding or dialkyl chain association is the initial event, or if both occur simul-
taneously. It seems most likely that triple-helical folding occurs ®rst, as (1) triple-helices nucleate in a zipper-like fashion from the C- to N-termi-
nus4
8,49
and (2) this would allow for the proper register of the individual chains.
ꢁ
1
Pro-Hyp) at the same temperatures (either 10 or 25 C)
4
one-dimensional H NMR spectra correlating shifts in Trp
indicates that the peaks corresponding to triple-helical
conformation are much stronger in the former compound.
The peaks at 7.64±7.71 ppm, which correspond to mono-
meric conformation, are very weak in the NMR spectrum
proton signals with melting transitions observed by CD
spectroscopy, and (5) PFG NMR determination of self-
di€usion coecients for two distinct species correlated to
the melting transitions observed by CD spectroscopy.
ꢁ
19
of C -(Gly-Pro-Hyp) -[IV-H1]-(Gly-Pro-Hyp) at 10 C.
4
6
4
ꢁ
H1]-(Gly-Pro-Hyp) is primarily triple-helical. The ther-
This result suggests that at 10 C C -(Gly-Pro-Hyp) -[IV-
6
The peptide-amphiphiles described herein provide a sim-
ple approach for building stable protein structural motifs
using peptide head groups. The lipid hydrophobic inter-
actions of peptide-amphiphiles (Fig. 3) exert a signi®cant
in¯uence on triple-helical structure formation and stabili-
zation. For example, the di€erence in the denaturation
temperatures between the structured C -(Gly-Pro-Hyp) -
4
4
mal transitions evaluated from the NMR spectra and the
CD melting curves are in agreement.
1
The CD and H NMR experiments have indicated that
some NMR spectral signals arising from the side chain
of Trp are associated with triple-helical conformational
changes. If this is the case, than self-di€usion coe-
cients calculated from triple-helical peaks are expected
to be smaller than self-di€usion coecients calculated
n
4
[IV-H1]-(Gly-Pro-Hyp) peptide-amphiphiles is about
4
ꢁ
18
2.75 C per � CH � . The tight alignment of the N-
2
terminal amino acids achieved through the association of
the lipid part of the molecule in a monolayer could be a
general tool for initiation of peptide folding.
3
7
from monomeric peaks. The self-di€usion coecient
D) was subsequently measured using PFG NMR spec-
(
troscopy and varying the gradient strength. D was initi-
ally measured for the peaks occurring at 7.74 ppm,
potentially corresponding to triple-helical conformation,
and 7.70 ppm, potentially corresponding to monomeric
One of the most intriguing features of this system is the
possible formation of stable lipid ®lms on solid substrates,
or the use of the novel amphiphiles in bilayer membrane
systems, where the lipid tail serves not only as a peptide
structure-inducing agent but also as an anchor of the
functional head group in the lipid assembly. One can mix
peptide-amphiphiles with vesicle-forming lipids (such as
dilauryl phosphatidylcholine) to form stable mixed vesicles
with collagen-model, triple-helical peptide head groups.
Vesicles featuring collagen coatings have already been
1
9
conformation. The D values calculated at 7.74 and
� 6
�
6
2
2
7
.70 ppm are 0.69Â10 cm /sec and 1.02Â10 cm /sec,
1
9
respectively. Thus, D is lower for the triple-helical peak
compared with the monomeric peak.
4
2
shown to be advantageous for targeted drug delivery.
Conclusions
The peptide-amphiphile system potentially o€ers great
versatility with regard to head and tail group composition
and overall geometries and macromolecular structures.
CD and NMR spectroscopic studies have provided
strong evidence that (1) peptide-amphiphiles are in a
predominantly triple-helical conformation at tempera-
ꢁ
tures below 30 C, (2) the [IV-H1] region within peptide-
amphiphiles are in a triple-helical environment, and (3)
there exist two peptide-amphiphile states, one at low
temperature and one at high temperature, of di€erent dif-
fusion coecients. The evidence for triple-helicity is based
upon (1) CD spectra characteristic of triple-helices, (2)
sigmoidal melting curves for the transition from triple-
helix to monomeric state, (3) two-dimensional NMR
spectra of the Pro and Hyp side-chains within and/or sur-
rounding the [IV-H1] site correlating cross-peaks with
melting transitions observed by CD spectroscopy, (4)
Experimental
Synthesis of peptide-amphiphiles
All standard peptide synthesis chemicals and solvents
were analytical reagent grade or better and purchased
from Applied Biosystems, Inc. (Foster City, CA) or
Fisher Scienti®c (Pittsburgh, PA). Fmoc-4-(2 ,4 -dime-
thoxyphenylaminomethyl)phenoxy resin (substitution
level=0.46 mmol/g) was purchased from Novabiochem
0
0

8
0
G. B. Fields / Bioorg. Med. Chem. 7 (1999) 75±81
(
La Jolla, CA). All Fmoc-amino acid derivatives were
Peptide puri®cation and analysis
from Novabiochem and are of l-con®guration. 1-
Hydroxybenzotriazole (HOBt) was purchased from
Novabiochem, HBTU from Richelieu Biotechnologies
(St. Hyacinthe, Quebec), and N,N-diisopropylethyla-
mine (DIEA) from Fisher Scienti®c.
Preparative RP-HPLC puri®cation is performed on a
Rainin AutoPrep System. Peptide-amphiphile puri®ca-
tion is achieved using a Vydac 214TP152022 C column
4
Ê
(15±20 mm particle size, 300 A pore size, 250Â22 mm) at
a ¯ow rate of 10 mL/min. The elution gradient is 55±
90% B in 20 min where A is 0.05% TFA in H O and B
Construction of dialkyl tail compounds ®rst involves
the acid-catalyzed condensation of Glu with the appro-
priate fatty acid alcohol to form the dialkyl ester of Glu
2
is 0.05% TFA in isopropanol. Detection is at 229 nm.
Analytical RP-HPLC is performed on a Hewlett±Pack-
ard 1090 Liquid Chromatograph equipped with a
(
(
p-toluenesulfonate salt).15 For example, hexadecanol
44.85 g, 0.185 mol) and Glu (13.6 g, 0.092 mol) are
Ê
Hypersil C18 column (5 mm particle size, 120 A pore size,
mixed with 21.0 g (0.102 mol) of p-toluenesulfonate in
toluene, and the mixture is heated until an equimolar
amount of water is recovered in a Dean±Stark trap. The
toluene is removed, and the product recrystallized from
acetone in ꢂ80% yield. The free amino group of Glu is
then treated with succinic anhydride and triethylamine
200Â2.1 mm) a ¯ow rate of 0.3 mL/min. The elution
gradient is 0±60% B in 45 min where A is 0.05% TFA in
H O and B is 0.05% TFA in acetonitrile. Diode array
2
detection is at l=220, 254, and 280 nm.
Edman degradation sequence analysis is performed on
an Applied Biosystems 477A Protein Sequencer/120A
Analyzer. MALDI-MS is performed on a Hewlett±
0
0
to create a free carboxylic acid. 1 ,3 -Dihexadecyl l-Glu
20 g, 26 mmol) and triethylamine (5.5 mL, 39 mmol) are
dissolved in tetrahydrofuran:CHCl3 (1:1), and 3.9 g
(
Packard G2025A laser desorption time-of-¯ight mass
5,16
(
39 mmol) of succinic anhydride is added under stirring.
spectrometer using a sinnapinic acid matrix.1
tide-amphiphile samples were dissolved in either H O or
Pep-
ꢁ
The mixture is kept for 4 h at 30 C. The product
obtained after removal of the solvent is recrystallized
from acetone and ethanol in ꢂ94% yield. The dialkyl
tail can then be used directly for solid-phase synthesis,
or is converted to a 4-nitrophenyl ester. Dihexadecyl N-
succinyl-l-Glu (6.90 g, 9.9 mmol) and 4-nitrophenol
2
H O-acetonitrile containing 0.05% TFA.
2
Circular dichroism spectroscopy
CD spectra were recorded over the range of l=190±
250 nm on a JASCO J-710 spectropolarimeter using a
thermostated 0.1 mm quartz cell. Thermal transition
curves are obtained by recording the molar ellipticity
(
(
1.65 g, 11.9 mmol) are dissolved in CH Cl , and 2.05 g
2 2
9.9 mmol) of N,N-dicyclohexylcarbodiimide as well as
a catalytic amount (80 mg) of (dimethylamino)pyridine
is added to the reaction mixture on an ice bath. The
reaction is continued for 2 h on the ice bath and for 24 h
at room temperature. The formed dicyclohexylurea is
ꢁ
([y]) in the range of 10±80 C at l=225 nm. The pep-
tide-amphiphile concentration is 0.5 mM in H O at
ꢁ
2
25 C.
®
cold dry ethanol in a yield of 85%. The two other di-
ltered o€, and the reaction product precipitated with
NMR spectroscopy
0
0
alkyl ester tail precursors, 1 ,3 -ditetradecyl N-[O-(4-
nitrophenyl)succinyl]-l-Glu [designated (C ) -Glu-C -
NMR spectra were acquired on a 500 MHz Bruker
ꢁ
1
4 2
2
0
0
pNp] and 1 ,3 -didodecyl N-[O-(4-nitrophenyl)succinyl]-
l-Glu [designated (C ) -Glu-C -pNp], are synthesized as
AMX-500 spectrometer at 10, 25, 50, and/or 80 C.
Freeze-dried samples for NMR spectroscopy were dis-
solved in D O or D O:H O (1:9) at peptide and pep-
1
2 2
0
2
0
described above for 1 ,3 -dihexadecyl N-[O-(4-nitrophe-
nyl)succinyl]-l-Glu [designated (C ) -Glu-C -pNp]. The
2
2
2
tide-amphiphile concentrations of 3±5 mM at least 48 h
prior to experiments. Two dimensional TOCSY and
nuclear Overhauser e€ect spectroscopy (NOESY) are
performed with 256 t1 increment and 1024 complex
data points in the t2 dimension. TOCSY spectra are
obtained at mixing times of 40±150 ms. NOESY spectra
are obtained at mixing times of 60±250 ms. The spectral
widths are 6024 Hz in both dimensions. Self-di€usion
coecients (D) were measured by PFG NMR as
described by Gibbs and Johnson.47 Experiments were
performed using a PFG duration of 6 ms, 52.4 ms
between PFG pulses, and gradient strengths of 4.12,
8.24, 12.36, 16.48, 20.60, 24.72, 28.84, and 32.96 Gauss/
1
6 2
2
C
1
and C tails are prepared using 1-tetradecyl alcohol
12
4
and 1-dodecyl alcohol, respectively. The monoalkyl
chains hexanoic acid [CH -(CH ) -CO H, designated C ],
octanoic acid [CH -(CH ) -CO H, designated C8],
3
2 4
2
6
3
2 6
2
decanoic acid [CH (CH ) -CO H, designated C ], dode-
10
3
2 8
2
canoic acid [CH -(CH ) -CO H, designated C ], tetra-
3
2 10
2
12
decanoic acid [CH -(CH ) -CO H, designated C ],
14
3
2 12
2
and palmitic acid [CH -(CH ) -CO H, designated C ]
3
2 14
2
16
were purchased from Aldrich.
Peptide-resin assembly is performed by Fmoc solid-phase
methodology. Peptide-resins are characterized by
Edman degradation sequence analysis as described pre-
21,43,44
cm. Since samples were dissolved in D O, water sup-
2
pression was not required.
21
viously for embedded (noncovalent) sequencing. Peptide-
resins are then lipidated with the appropriate (C ) -Glu-
n 2
1
5,16
19
C2
or CH -(CH ) -CO H tail. Cleavage and side
3 2 n-2 2
chain deprotection of peptide-amphiphile-resins proceeds
for 1 h using either ethanedithiol:thioanisole:phenol:
H O:tri¯uoroacetic acid (TFA) (2.5:5:5:5:82.5) or H O:
TFA (1:19). Peptide-amphiphile cleavage solutions
are extracted with methyl tBu ether prior to puri®cation.
Acknowledgements
2
2
4
5,46
This research is supported by NIH grants CA77402,
HL62427, and AR01929.

G. B. Fields / Bioorg. Med. Chem. 7 (1999) 75±81
81
References
25. Macquaire, F.; Baleux, F.; Giaccobi, E.; Huynh-Dinh, T.;
Neumann, J. M.; Sanson, A. Biochemistry 1992, 31, 2576.
26. Winger, T. M.; Ludovice, P. J.; Chaikof, E. L. J. Liq.
Chromatogr. 1995, 18, 4117.
27. Winger, T. M.; Ludovice, P. J.; Chaikof, E. L. Biomater-
ials 1996, 17, 437.
1
. Sakakibara, S.; Kishida, Y.; Kikuchi, Y.; Sakai, R.;
Kakiuchi, K. Bull. Chem. Soc. Jpn. 1968, 41, 1273.
. Mayo, K. H.; Fields, G. B. In Protein Structural Biology in
Bio-Medical Research; Allewell, N., Woodward, C., Eds.; JAI
2
Press: Greenwich, CT, 1997; pp 567±612.
28. Thompson, N. L.; Brian, A. A.; McConnell, H. M. Bio-
chim. Biophys. Acta 1984, 72, 10.
29. Prass, W.; Ringsdorf, H.; Bessler, W.; Wiesmuller, K. H.;
Jung, G. Biochim. Biophys. Acta 1987, 900, 116.
30. Cha, X.; Ariga, K.; Kunitake, T. Bull. Chem. Soc. Jpn.
1996, 69, 163.
31. de Bont, H. B. A.; van Boom, J. H.; Liskamp, R. M. J.
Recl. Trav. Chim. Pays-Bas 1992, 111, 222.
32. Sankaram, M. B. Biophys. J. 1994, 67, 105.
33. Ramachandran, G. N. Int. J. Pept. Protein Res. 1988, 31,
1.
3
®
4
®
. Wattenbarger, M. R.; Chan, H. S.; Evans, D. R.; Bloom-
eld, V. A.; Dill, K. A. J. Chem. Phys. 1990, 93, 8343.
. Chan, H. S.; Wattenbarger, M. R.; Evans, D. J.; Bloom-
eld, V. A.; Dill, K. A. J. Chem. Phys. 1991, 94, 8542.
5
7
6
7
6
8
. DeGrado, W. R.; Lear, J. D. J. Am. Chem. Soc. 1985, 107,
684.
. Osapay, G.; Taylor, J. W. J. Am. Chem. Soc. 1990, 112, 6046.
. Osapay, G.; Taylor, J. W. J. Am. Chem. Soc. 1992, 114,
966.
. Mutter, M.; Tuchsehcrer, G. G.; Miller, C.; Altmann, K.
H.; Carey, R. I.; Wyss, D. F.; Labhardt, A. M.; Rivier, J. E. J.
Am. Chem. Soc. 1992, 114, 1463.
34. Brodsky, B.; Shah, N. K. FASEB J. 1995, 9, 1537.
35. Heidemann, E.; Roth, W. Adv. Polym. Sci. 1982, 43, 143.
36. Brodsky, B.; Li, N. H.; Long, C. G.; Apigo, J.; Baum, J.
Biopolymers 1992, 32, 447.
37. Li, M. H.; Fan, P.; Brodsky, B.; Baum, J. Biochemistry
1993, 32, 7377.
38. Long, C. G.; Braswell, E.; Zhu, D.; Apigo, L.; Baum, L.;
Brodsky, B. Biochemistry 1993, 32, 11688.
39. Goodman, M.; Feng, Y.; Melacini, G.; Taulane, J. P. J.
Am. Chem. Soc. 1996, 118, 5156.
40. Mayo, K. H.; Parra-Diaz, D.; McCarthy, J. B.; Chelberg,
M. Biochemistry 1991, 30, 8251.
È
41. Wuthrich, K. In NMR of Proteins and Nucleic Acids; John
Wiley & Sons: New York, 1983; pp 14±18.
42. Fonseca, M. J.; Alsina, M. A.; Reig, F. Biochim. Biophys.
Acta 1996, 1279, 259.
43. Fields, G. B.; Noble, R. L. Int. J Pept. Protein Res. 1990,
35, 161.
9
7
1
1
1
1
1
1
1
1
1
1
1
. Dawson, P. E.; Kent, S. B. H. J. Am. Chem. Soc. 1993, 115,
263.
0. Vuilleumer, S.; Mutter, M. Biopolymers 1993, 33, 389.
1. Ghadiri, M. R.; Soares, C.; Choi, C. J. Am. Chem. Soc.
992, 114, 825.
2. Lieberman, M.; Sasaki, T. J. Am. Chem. Soc. 1991, 113,
470.
3. Sasaki, T.; Lieberman, M. Tetrahedron 1993, 49, 3677.
4. Ghadiri, M. R.; Soares, C.; Choi, C. J. Am. Chem. Soc.
992, 114, 4000.
5. Berndt, P.; Fields, G. B.; Tirrell, M. J. Am. Chem. Soc.
995, 117, 5159.
6. Yu, Y. C.; Berndt, R.; Tirrell, M.; Fields, G. B. J. Am.
Chem. Soc. 1996, 118, 12515.
7. Yu, Y. C.; Pakalns, T.; Dori, Y.; McCarthy, J. B.; Tirrell,
M.; Fields, G. B. Methods Enzymol. 1997, 289, 571.
8. Fields, G. B.; Lauer, J. L.; Dori, Y.; Forns, P.; Yu, Y.-C.;
Tirrell, M. Biopolymers 1998, 47, 143.
1
1
44. Lauer, J. L.; Fields, C. G.; Fields, G. B. Lett. Pept. Sci.
1995, 1, 197.
1
1
2
2
1
2
9. Yu, Y.-C.; Tirrell, M.; Fields, G. B. J. Am. Chem. Soc.
998, 120, 9979.
0. Kunitake, T. Angew. Chem., Int. Ed. Engl. 1992, 31, 709.
1. Fields, C. G.; VanDrisse, V. L.; Fields, G. B. Peptide Res.
993, 6, 39.
45. King, D. S.; Fields, C. G.; Fields, G. B. Int. J. Pept. Pro-
tein Res. 1990, 36, 255.
46. Fields, C. G.; Fields, G. B. Tetrahedron Lett. 1993, 34,
6661.
47. Gibbs, S. J.; Johnson, J. C. S. J. Magn. Reson. 1991, 93,
395.
2. Shimizu, T.; Mori, M.; Minamikawa, H.; Hato, M. Chem.
Lett. 1989, 1341.
48. Engel, T.; Prockop, D. J. Annu. Rev. Biophys. Chem. 1991,
20, 137.
49. Liu, X.; Siegel, D. L.; Fan, P.; Brodsky, B.; Baum, J.
Biochemistry 1996, 35, 4306.
23. Shimizu, T.; Hato, M. Biochim. Biophys. Acta 1993, 1147, 50.
24. Jain, R. K.; Gupta, C.; Anand, N. Tetrahedron Lett. 1981,
22, 2317.