Aminothiotyrosine disulfide, an optical trigger for initiation of protein folding

Article abstract of DOI:10.1021/ja970567o

The development of an optical trigger for protein folding is described. The optical trigger is an aryl disulfide embedded in a polypeptide such that the aryl, disulfide constrains the peptide in a nonhelical conformation. Upon photocleavage of the SS bond, the peptide can then commence to fold into an α-helical conformation. Two thiotyrosine derivatives, (S)-4'-mercaptophenylalanine (Tty) and 3-N-(4'-mercaptophenyl)-(S)-2,3-diaminoproprionate (Aty), have been prepared and incorporated into polyalanine peptides. The ease of synthesis of protected forms of Tty and Aty amenable for solid phase synthesis, in four steps with 30% overall yield and six steps in 40% overall yield, respectively, make these attractive candidates as precursors of the optical trigger. CD and IR spectroscopy showed that the cyclic disulfide cross-linked peptides are much less helical than their linear counterparts. Following laser flash photolysis, peptide 7, which incorporates Tty, showed total recombination of the thiyl radicals within 1 ns. Peptide 16, which incorporates Aty, showed a significant amount of long-lived thiyl radicals from nanosecond to microsecond time scale. The process of recombination is hypothesized to be governed by the peptide conformation. Because of the significant amount of long-lived thiyl radicals generated from cyclic peptide 16, Aty should prove to be of general utility in the studies of protein folding on a time scale of sub-picoseconds and greater.

Full text of DOI:10.1021/ja970567o

VOLUME 119, NUMBER 31
AUGUST 6, 1997
© Copyright 1997 by the
American Chemical Society
Aminothiotyrosine Disulfide, an Optical Trigger for Initiation of
Protein Folding
Helen S. M. Lu,† Martin Volk,‡,§ Yuriy Kholodenko,‡ Edward Gooding,‡
Robin M. Hochstrasser,‡ and William F. DeGrado*,†
Contribution from the Department of Biochemistry and Biophysics, The Johnson Foundation of
Research in Biophysics, Department of Chemistry, The UniVersity of PennsylVania,
Philadelphia, PennsylVania 19104
ReceiVed February 21, 1997X
Abstract: The development of an optical trigger for protein folding is described. The optical trigger is an aryl
disulfide embedded in a polypeptide such that the aryl disulfide constrains the peptide in a nonhelical conformation.
Upon photocleavage of the SS bond, the peptide can then commence to fold into an R-helical conformation. Two
thiotyrosine derivatives, (S)-4′-mercaptophenylalanine (Tty) and 3-N-(4′-mercaptophenyl)-(S)-2,3-diaminoproprionate
(Aty), have been prepared and incorporated into polyalanine peptides. The ease of synthesis of protected forms of
Tty and Aty amenable for solid phase synthesis, in four steps with 30% overall yield and six steps in 40% overall
yield, respectively, make these attractive candidates as precursors of the optical trigger. CD and IR spectroscopy
showed that the cyclic disulfide cross-linked peptides are much less helical than their linear counterparts. Following
laser flash photolysis, peptide 7, which incorporates Tty, showed total recombination of the thiyl radicals within 1
ns. Peptide 16, which incorporates Aty, showed a significant amount of long-lived thiyl radicals from nanosecond
to microsecond time scale. The process of recombination is hypothesized to be governed by the peptide conformation.
Because of the significant amount of long-lived thiyl radicals generated from cyclic peptide 16, Aty should prove to
be of general utility in the studies of protein folding on a time scale of sub-picoseconds and greater.
Introduction
this has mainly been accomplished by stopped-flow mixing
techniques that introduce a change of denaturant concentration
in the solvent. These techniques typically have a time resolution
of a few milliseconds. However, several proteins have been
reported to exhibit significant structure before the millisecond
or even microsecond time regime.2-9 The elucidation of the
kinetics and structural features associated with the intermediates
The mechanism by which proteins fold rapidly into their
native state is currently under active theoretical and experimental
investigations. The elucidation of protein folding forces and
pathways can shed light on how proteins obtain their unique
structures and also aid in the design of novel proteins with
tailored functions. To study protein folding or unfolding, a
triggering event that perturbs a protein from its equilibrium state
is required to initiate either folding or unfolding.1 Until recently,
(2) Bycroft, M.; Matouschek, A.; Kellis, J. T., Jr.; Serrano, L.; Fersht,
A. R. Nature 1990, 346, 488-490.
(3) Houry, W. A.; Rothwarf, D. M.; Scheraga, H. A. Nature Struct. Biol.
1995, 2, 495-503.
† The Johnson Foundation of Research in Biophysics.
‡ Department of Chemistry.
(4) Elo¨ve, G. A.; Chaffotte, A. F.; Roder, H.; Goldberg, M. E.
Biochemistry 1992, 31, 6876-6883.
§ Present address: Institut fu¨r Physikalische und Theoretische Chemie,
TU Mu¨nchen, Lichtenbergsr. 4, 85748 Garching, Germany.
X Abstract published in AdVance ACS Abstracts, July 15, 1997.
(1) Eaton, W. A.; Thompson, P. A.; Chan, C.-K.; Hagen, S. J.; Hofrichter,
J. Structure 1996, 4, 1133-1139.
(5) Jennings, P. A.; Wright, P. E. Science 1993, 262, 892-895.
(6) Kragelund, B. B.; Robinson, C. V.; Knudsen, J.; Dobson, C. M.;
Poulsen, F. M. Biochemistry 1995, 34, 7217-7224.
(7) Kuwajima, K.; Yamaya, H.; Miwa, S.; Sugai, S.; Nagamura, T. FEBS
Lett. 1987, 221, 115-118.
S0002-7863(97)00567-2 CCC: $14.00 © 1997 American Chemical Society

7174 J. Am. Chem. Soc., Vol. 119, No. 31, 1997
Scheme 1. Photoinitiation of R-Helix Formation
Lu et al.
formed in the dead time of the mixing experiments remains a
great challenge. In particular, secondary structural elements are
often observed to form rapidly in the so-called burst phase of
protein folding.
ligand dissociation of cytochrome c26 and hemoglobin,27 and
electron transfer induced folding of cytochrome c.28 However,
these techniques have their limitations. Temperature jump
experiments can only be used to study either the unfolding of
proteins, or the folding of proteins that undergo cold dena-
turation, whereas the light-initiated ligand dissociation and
electron transfer initiated folding are limited to hemoproteins.
In this work, we set out to develop an ultrafast trigger of protein
folding applicable to a wide range of peptides and proteins.
The roles that secondary structure plays in guiding protein
folding is under debate. The framework model and the
diffusion-collision model10-15 propose that the secondary
structures are formed first, and thereby narrow the conforma-
tional search space so that the protein can fold in finite time.
Other models of protein folding, such as the hydrophobic
collapse model16 and the nuclear condensation model,17-20
maintain that collapse of hydrophobic groups or the formation
of a group of nuclei occurs first. These interactions then guide
the protein to its native state, with secondary structures forming
following the rate-determining step. Recent theoretical models
of protein folding based on statistical mechanics suggest that
secondary structural elements can reduce the number of
configurations available to the protein and thus enable rapid
folding.21 In order to resolve these critical issues, an ultrafast
trigger to initiate protein folding is required to enable the study
of the full range of conformational changes starting from the
earliest to the last events of folding.
The ideal trigger should be faster than the underlying
conformational processes being studied and be readily formed,
detected, and incorporated into proteins. Also the photo-
triggering event should be reversible to allow for multiple
excitations by the laser, enabling data averaging for good signal
to noise ratio. To meet these criteria, we chose photocleavage
of an aromatic disulfide bond as the triggering event. Aromatic
disulfides have been shown to cleave in less than 1 ps.29-31
The resulting thiyl radical is a strong chromophore (ꢀ540 ) 9077
M-1 cm-1 for p-aminophenylthiyl radical30,32) which can
therefore serve as an optical probe of the protein dynamics. This
scheme requires the synthesis of an appropriate thiol-containing
amino acid that is amenable to solid phase synthesis. Placement
of two of these thiol residues at specific places in a peptide
sequence, followed by oxidation of the resulting dithiol-
containing peptide to disulfides, should constrain the polypeptide
into a non-native conformation. Laser flash photolysis of the
polypeptide leads to photocleavage of the S-S bond, and the
polypeptide can then commence folding to the native state. This
process is illustrated Scheme 1.
Different approaches have been utilized to initiate and study
the early phases of protein folding. They include temperature
jump experiments using short laser pulses to heat the sample
to induce either protein unfolding or folding,22-25 light-initiated
(8) Ptitsyn, O. B.; Pain, R. H.; Semisotnov, G. V.; Zerovnik, E.;
Razgulayaev, E. I. FEBS Lett. 1990, 262, 20-24.
(9) Radford, S. E.; Dobson, C. M.; Evans, P. A. Nature 1992, 358, 302-
307.
(10) Karplus, M.; Weaver, D. C. Nature 1976, 260, 404-406.
(11) Karplus, M.; Weaver, D. L. Protein Sci. 1994, 3, 650-668.
(12) Kim, P. S.; Baldwin, R. L. Annu. ReV. Biochem. 1982, 51, 459-
489.
As a demonstration of the utility of our trigger, we chose to
study the folding of an alanine rich peptide, which has been
shown by Marqusee and Baldwin to fold into stable helices in
solution.33 With our trigger incorporated, we should be able to
study the early events of helix formation. Following photo-
dissociation of the disulfide, conformational changes of the
peptide can be followed by transient IR spectroscopy, a very
sensitive method that detects bond level changes of protein
(13) Kim, P. S.; Baldwin, R. L. Annu. ReV. Biochem. 1990, 59, 631-
660.
(14) Ptitsyn, O. B. Protein Eng. 1994, 7, 593-596.
(15) Wetlaufer, D. B. Proc. Natl. Acad. Sci. U.S.A. 1973, 70, 697-701.
(16) Dill, K. Biochemistry 1985, 24, 1501-1509.
(17) Abkevich, V. I.; Gutin, A. M.; Shakhnovich, E. I. Biochemistry 1994,
33, 10026-10036.
(18) Fersht, A. R. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 10869-10873.
(19) Sosnick, T. R.; Mayne, L.; Englander, S. W. Proteins: Struct.,
Funct., Genet. 1996, 24, 413-426.
(26) Jones, C. M.; Henry, E. R.; Hu, Y.; Chan, C.-K.; Luck, S. D.;
Bhuyan, A.; Roder, H.; Hofrichter, J.; Eaton, W. A. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 11860-11864.
(20) Guo, Z.; Thirumalai, D. Biopolymers 1995, 36, 83-102.
(21) Bryngelson, J. D.; Onuchic, J. N.; Socci, N. D.; Wolynes, P. G.
Proteins 1995, 21, 167-195.
(27) Anfinrud, P. A.; Han, C.; Hochstrasser, R. M. Proc. Natl. Acad.
Sci. U.S.A. 1989, 86, 8387-8391.
(28) Pascher, T.; Chesick, J. P.; Winkler, J. R.; Gray, H. B. Science 1996,
271, 1558-1560.
(22) No¨lting, B.; Golbik, R.; Fersht, A. R. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 10668-10672.
(29) Borisevich, N. A.; Lysak, N. A.; Melnichuk, S. V.; Tikhomirov, S.
A.; Tolstorozhev, G. B. Fast Structural Transformations of Organic
Molecules; Springer-Verlag: Berlin, 1990; Vol. 49, pp 276-281.
(30) Borisevich, N. A.; Malkin, Y. N.; Ruziev, S.; Mel’nichuk, S. V.;
Tikhomirov, S. A.; Tolstorozhev, G. B.; Kuz’min, V. A. IzV. Akad. Nauk
SSSR, Ser. Khim. 1990, 3, 468-471.
(23) Phillips, C. M.; Mizutani, Y.; Hochstrasser, R. M. Proc. Natl. Acad.
Sci. U.S.A. 1995, 92, 7292-7296.
(24) (a) Williams, S.; Causgrove, T. P.; Gilmanshin, R.; Fang, K. S.;
Callender, R. H.; Woodruff, W. H.; Dyer, R. B. Biochemistry 1996, 35,
691-697. (b) The folding lifetime for helix formation was originally
reported to be 16 ns in ref 24a and was subsequently revised to 180 ns in
Gilmanshin, R.; Williams, S.; Callender, R. H.; Woodruff, W. H.; Dyer, R.
B. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 3709-3713.
(25) Ballew, R. M.; Sabelko, J.; Gruebele, M. Proc. Natl. Acad. Sci.
U.S.A. 1996, 93, 5759-5764.
(31) Scott, T. W.; Liu, S. N. J. Phys. Chem. 1989, 93, 1393-1396.
(32) Lembke, R. R.; Natarajan, L. V.; Kuntz, R. R. J. Photochem. 1983,
21, 814.
(33) Marqusee, S.; Robbins, V. H.; Baldwin, R. L. Proc. Natl. Acad.
Sci. U.S.A. 1989, 86, 5286-5290.

Optical Trigger for Initiation of Protein Folding
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7175
conformation.34 The carbonyl groups of the peptide backbone
absorb at different frequencies in different hydrogen-bonding
environments such as random coils, R-helix, and â-sheets.35
Furthermore, transient optical spectroscopy can also be used
for reasons stated above.
Scheme 2. Synthesis of (a) Fmoc-Tty (4-MeBz), 4, and (b)
Peptide 7
Here, we report the synthesis of potential cross-linking amino
acids, (S)-4′-mercaptophenylalanine (thiotyrosine, Tty) and 3-N-
(4′-mercaptophenyl)-(S)-2,3-diaminoproprionic acid (aminothio-
tyrosine, Aty), their incorporation into peptides, and the
photochemical and biophysical characterization of the resulting
peptides. Two peptides with the same sequence but two
different thiol amino acids were examined. The kinetics of the
early events of folding of these peptides will be reported
elsewhere.39
Results
Synthesis of Thiotyrosine 4 (Fmoc-Tty(4-MeBzl)) and
Incorporation into Peptide 7. On the basis of the precedence
of the photochemistry of diphenyl disulfide,29-31 we envision
that the thiotyrosine Tty might serve as the precursor to the
aryl disulfide cross-linked peptide. Tty is structurally similar
to thiophenol, and once oxidized should exhibit similar photo-
chemistry and ultrafast photodissociation of the SS bond as
observed for diphenyl disulfide. The synthesis36 of Tty in
properly protected forms for use in solid phase peptide synthesis
is outlined in Scheme 2a. Starting from L-phenylalanine,
chlorosulfonation gave rise to the unstable chlorosulfonyl
derivative 1. This compound was reduced to the air sensitive
thiol 2 with powdered tin and hydrochloric acid. The thiol group
of 2 was reacted with 4-methylbenzyl bromide to give the S-4-
methylbenzyl (4-MeBzl) derivative 3. The amino group of 3
was protected with 9-fluorenylmethyl group to afford amino
acid 4 (Fmoc-Tty(4-MeBzl)), which is properly protected for
use in solid phase synthesis using Fmoc37 chemistry. The
overall yield of Fmoc-Tty(4-MeBzl) is 30% from L-phenyl-
alanine. The ease in which Fmoc-Tty(4-MeBzl) is prepared
makes this compound an attractive precursor of the optical
trigger.
Compound 4 was coupled to PAL resin using HBTU, and
the remaining amino acids were coupled under standard
conditions as described in the Experimental Section. We sought
a thiol-protecting group that would be stable to the conditions
required for cleavage from PAL resin to allow purification of
the acyclic reduced form of the peptide 6. Various thiol-
protecting groups were investigated including S-benzyl, S-
methoxybenzyl, S-acetamidomethyl, and S-methylbenzyl groups.
The S-methylbenzyl group was found to yield the cleanest and
highest yield of peptide 6 with the least amount of polymeric
material arising from deprotection of the thiol group. Peptide
6 was oxidized to the cyclic disulfide cross-linked peptide 7
with DMSO as the oxidant38 (Scheme 2b). Peptide 7 was
purified by RP-HPLC and characterized by electrospray mass
spectroscopy.
Photochemistry of Peptide 7. Peptide 7 was photolyzed at
270 nm, and the transient absorption spectrum was monitored.
Immediately following photolysis, a transient was observed
centered at 500 nm.39 The absorption spectrum compares well
with that from phenylthiyl radical, with a band centered around
450 nm.31 This transient signal decays rapidly to 17% of the
initial intensity at 50 ps and to 2% by 1 ns. We hypothesize
that the absorbance decay of thiyl radicals derived from 7 is
due to recombination based on kinetic studies of phenylthiyl
radicals.29-31 Following the photolysis of diphenyl disulfide,
phenylthiyl radical is formed in less than 1 ps. The intensity
of the thiyl absorption then diminishes to a plateau value of
15-75% in 1 ns depending on the solvent.29 The fast, initial
decrease in the absorbance of thiyl radical is attributed to cage
(geminate) recombination, with a half-life for recombination in
the tens of picosecond range. The remainder of the phenylthiyl
radicals escape from the cage and give rise to the absorbance
plateau. The thiyl radicals from peptide 7, on the other hand,
are tethered at the ends of the peptide chain. We propose that
once the disulfide bond is cleaved, the peptide backbone of 7
has not had sufficient time to diffuse apart and holds the two
thiyl radicals in close proximity and proper orientation for rapid
recombination.
(34) Locke, B.; Diller, R.; Hochstrasser, R. M. Wiley: New York, 1993;
Vol. Part B, Vol. 20, pp 1-43.
(35) Krimm, S.; Bandekar, J. AdV. Protein Chem. 1986, 38, 181-364.
(36) Escher, E.; Bernier, M.; Parent, P. HelV. Chim. Acta 1983, 66, 1355-
1365.
(37) Abbreviations: Fmoc, 9-fluorenylmethoxycarbonyl; PAL, 5-(4-
(aminomethyl)-3,5-bis(methyloxy)phenoxy)valeric acid; TFA, trifluoroacetic
acid; DMSO, dimethyl sulfoxide; HPLC, high-performance liquid chro-
matography; CD, circular dichroism; UV, ultraviolet; HBTU, o-benzo-
triazole-N,N,N′,N′-tetramethyluronium hexafluorophosphate; DIEA, N,N-
diisopropylethylamine.
To overcome this unexpected obstacle, we sought various
ways to increase the survival probability of the thiyl radicals
generated from photolysis of the cross-linked polyalanine
peptides. If the rate of recombination were sufficiently de-
(38) Tam, J. P.; Wu, C. R.; Liu, W.; Zhang, J. W. J. Am. Chem. Soc.
1991, 113, 6657-6662.
(39) Volk, M.; Kholodenko, Y.; Lu, H. S. M.; Gooding, E.; DeGrado,
W. F.; Hochstrasser, R. M. Submitted for publication in J. Phys. Chem.

7176 J. Am. Chem. Soc., Vol. 119, No. 31, 1997
Lu et al.
Scheme 3. Synthesis of Aminothiotyrosine 14 and Peptide 16
creased, the radicals might be able to diffuse apart, thereby
allowing peptide folding to occur. The first approach we
examined was to increase the steric bulk around the thiyl radical
center. The steric hindrance was thought to prevent the two
thiyl radical centers from approaching each other. Therefore,
we determined the quantum yield of caged recombination of
2,3,5,6-tetramethylphenylthiyl and 2,4,6-triisopropylphenylthiyl
and compared them with that of p-tolylthiyl. All three disulfides
exhibited no observable difference in caged recombination
quantum yield. Apparently, the additional steric hindrance did
not appreciably slow down caged recombination.40 Because the
model disulfides looked unpromising, we did not investigate
peptides with substituted thiotyrosine derivatives incorporated.
The second approach was to perturb the peptide conformation
such that upon photocleavage of the SS bond the peptide
changes its conformation more rapidly so that the two thiyl
radicals are no longer in close proximity. Computer modeling
using molecular dynamics simulation showed that the cyclic
disulfide-containing peptide 7 samples multiple conformations.
Many conformers show approximately two turns of helix with
the aryl disulfide packed across the helix, possibly stabilizing
it through favorable van der Waals and hydrophobic interactions.
Such a low-energy conformation may be kinetically stable,
leading to rapid recombination of the thiyl radicals once they
are generated. Indeed, the mean residue ellipticity at 222 nm,
[θ]222 of 7 was -13 700 deg cm2 dmol-1, providing evidence
for a partially helical conformation. We therefore prepared a
second peptide 8, which appeared to be less likely to adopt such
a low-energy conformation as assessed by molecular dynamics
simulation.
peptide backbone occurs in 50 ps. This observation is consistent
with the recent measurements of helix formation kinetics by
Williams24 and Thompson41 on polyalanine peptides which
showed that major backbone conformational changes have time
constants of 30 and 160 ns.
The third approach we examined was to perturb the thiyl
radical such that it might access an alternative excited state
configuration which would not lead to caged recombination. It
has been shown that p-amino-substituted diphenyl disulfides do
not undergo geminate recombination following photolysis.29,30
The rationale for this phenomenon is that the para amino group
can form an intramolecular charge transfer complex with the
thiyl radical center, decreasing the recombination probability
of the geminate radical pairs, and reducing geminate recombina-
tion. With this precedent in mind, we set out to synthesize the
p-amino analogue of Tty, compound Aty, where an amino group
is inserted in the para position of the modified tyrosine.
Synthesis of Amino Acid 14 (Fmoc-Aty(4-MeBzl)) and the
Incorporation into Peptide 16. Aminothiotyrosine Aty is a
derivative of 2,3-diaminoproprionic acid, a class of compounds
which we felt might be easily accessible in stereochemically
pure form starting with a derivative L-serine. However,
activation of the hydroxyl of L-serine is known to be problematic
and leads to â-elimination product, dehydroalanine.42,43 Arnolds
reported the use of N-(tert-butoxycarbonyl)-L-serine lactone in
preventing elimination and showed that various nucleophiles
react at the â-carbon of L-serine lactone.42 Unfortunately, S-(4′-
methylbenzyl)-4-aminothiophenol failed to react with L-serine
lactone in appreciable yield, presumably due to the poor
nucleophilicity of the aniline group. We therefore turned to
the Mitsunobu reaction using L-serine-N-trityl methyl ester (11).
Bulky N-protecting groups are found to protect the R-center of
amino acids from base-catalyzed racemization.44,45 In particular,
Cherney had reported that the trityl protecting group prevented
â-elimination, thereby providing high yields of Mitsunobu
addition products in a stereospecific manner.46 Indeed, reaction
of 11 with sulfonamide 10 under Mitsunobu conditions provided
the desired addition product 12 in 78% yield. Acidolytic
cleavage of the 2-mesitylenesulfonyl (Mts) and trityl protecting
groups from 12 using 30% HBr in acetic acid, followed by
saponification of the intermediate 13 and reaction of the product
with (fluorenylmethyl)-N-hydroxysuccinimide, afforded the
protected amino acid 14 (Fmoc-Aty(4-MeBzl)) (Scheme 3).
Amino acid 14 is properly protected for use in solid phase
peptide synthesis, in an analogous manner as described previ-
ously for the preparation of peptide 7. Although the â-amino
Although this peptide had a lower helical content as deter-
mined by CD spectroscopy ([θ]222 ) -8300 deg cm2 dmol-1),
it exhibited a very rapid decay of thiyl absorbance similar to
that of peptide 7. Apparently, the structural perturbation of 8
was not sufficient to drive the necessary rapid change to a
conformation nonproductive for thiyl radical recombination. For
this approach to work, the conformational changes must be able
to compete with the geminate recombination process, on the
order of 50 ps. Our failure to effect a change in the recombina-
tion efficiency implies that no major reorganization of the
(40) The photochemistry of these disulfides will be reported elsewhere.
(41) Thompson, P. A.; Eaton, W. A.; Hofrichter, J. Biophys. J. 1996,
70, A177.
(42) Arnold, L. D.; Kalantar, T. H.; Vederas, J. C. J. Am. Chem. Soc.
1985, 107, 7105-7109.
(43) Mitsunobu, O. Synthesis 1981, 1-28.
(44) Guthrie, R. D.; Nicolas, E. C. J. Am. Chem. Soc. 1981, 103, 4637-
4638.
(45) Christie, B. D.; Rapoport, H. J. Org. Chem. 1985, 50, 1239-1245.
(46) Cherney, R. J.; Wang, L. J. Org. Chem. 1996, 61, 7544-7546.

Optical Trigger for Initiation of Protein Folding
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7177
Figure 2. CD temperature melting curves of linear peptide 15 (b)
and cyclic peptide 16 (O).
Figure 1. CD spectra of linear peptide 15 (b) and cyclic peptide 16
(O) at 1 °C. [θ]222 is -64 000 deg cm2 dmol-1 for 15 and -28 000
deg cm2 dmol-1 for 16. The molar ellipticity at 222 nm for 15 is greater
than that expected for a 100% helical peptide, presumbly due to the
aromatic absorption of the amino thiotyrosine group.
group was not protected, the solid phase peptide synthesis
proceeded smoothly to yield the S-methylbenzyl-protected
peptide 15. Evidently, the low reactivity of the aniline obviates
the need for protection. Peptide 15 was treated with HF and
oxidized with DMSO to yield the desired cyclic peptide 16.
Peptide 16 was purified by RP HPLC on preparative C18 Vydac
column and characterized by electrospray mass spectroscopy.
Secondary Structure of Cyclic Peptide 16 and Linear
Peptide 15. From CD spectroscopy, we can analyze the
changes in the secondary structure expected to occur upon
photocleavage of cyclic peptide 16. Linear peptide 15 was used
to model the folded conformation after photocleavage of 16.
The CD spectra of the two peptides at the same concentrations
are shown in Figure 1. From the minimum at 222 nm band
and the maximum at 190 nm band, it is clear that disulfide cross-
linked peptide 16 has helical regions. It is difficult to determine
the fractional helicity of 16 since the aminothiophenol moiety
of the modified tyrosine unit most likely contributes to the
absorption in this region, as previously observed for other
peptides containing aromatic residues.47 The linear peptide 15
exhibits considerably greater helical character based on the
intensity of the 222 nm band. Assuming that the helical content
is linearly related to the intensity of the 222 nm band, linear
peptide 15 has approximately twice the helicity of the cyclic
peptide. Therefore, upon photocleavage of cyclic peptide 16,
it is reasonable to expect it to fold into a more helical peptide
conformation.
Figure 3. FT-IR spectra of linear peptide 15 at 10 °C (O) and 44 °C
(b) in D2O. Each spectrum was obtained by subtracting the D2O
spectrum from the peptide spectrum at that temperature.
The temperature dependence of the secondary structure of
15 and 16 was examined by CD (Figure 2). Both peptides
showed a broad melting curve, commonly observed for short
peptides. The secondary structure of the cyclic peptide 16 is
much less temperature dependent than that of the linear peptide
15. This difference in temperature dependence of the two
peptides can be exploited for future experiments. It would be
desirable to conduct experiments at low temperatures to
maximize the difference in the conformations of the cyclic and
ring-opened peptides.
We also examined the infrared spectra of linear peptide 15
(Figure 3) and the cyclic peptide 16 (Figure 4) in D2O as a
function of temperature. For both peptides, two major peaks
are observed in the amide I′ region (the prime indicates a
deuterated amide). The peak at 1672 cm-1 is attributable to
the CO absorption of the trifloroacetate counterion, and the 1640
Figure 4. The IR spectra of 16 at 6 °C (O) and 62 °C (b) in D2O.
Each spectrum was obtained by subtracting the D2O spectrum from
the peptide spectrum at that temperature.
cm-1 peak corresponds to the CO absorption of peptides in a
helical conformation. These spectra agree well with that
reported for a 21 residue polyalanine peptide.24 At 10 °C,
peptide 15 has a predominant peak centered at 1640 cm-1
,
indicating that the linear peptide is in a mostly helical
conformation at this temperature. Upon heating to 62 °C, a
new peak at 1655 cm-1 is observed, in place of the 1640 cm-1
band. We attribute this temperature dependence to helix
melting. With increasing temperature, the helix melts and gives
rise to an ensemble with greater random coil content, resulting
in the new peak at 1655 cm-1. The same trend is observed
upon heating of the crosslinked peptide 16 (Figure 4), indicating
there is some residual helicity in the cyclic peptide as seen in
the CD experiment, and the helical conformer melts upon
(47) Chakrabartty, A.; Kortemme, T.; Padmanabhan, S.; Baldwin, R. L.
Biochemistry 1993, 32, 5560-5565.

7178 J. Am. Chem. Soc., Vol. 119, No. 31, 1997
Lu et al.
Scheme 4
derived from cyclic peptide 16, which indicated that overall
peptide rotation occurs with a lifetime of 600 ps.39
To increase the lifetime of the thiyl radicals, we designed
the amino acid Aty, which contains a radical-stabilizing amino
group para to the thiol. This is the first optically triggered cross-
linking agent which allows for the detection of protein dynamics
from picoseconds to microseconds. It also meets nearly all of
our design criteria: (1) it is easily prepared in protected form
in six steps in 40% overall yield; (2) it can be incorporated
into peptides and proteins by routine solid phase methods; (3)
it is cleaved in less than a picosecond, allowing the full range
of protein dynamics to be examined, from extremely fast to
slow conformational processes; (4) it has a high cleavage
quantum yield; (5) photocleavage is readily and rapidly revers-
ible, enabling data averaging for good signal-to-noise ratios;
(6) photocleavage creates a very strongly absorbing chro-
mophore that acts both as a probe of protein dynamics as well
as a monitor of the photocleavage and recombination processes.
The sole feature that has not yet been optimized is the degree
of recombination occurring on a time scale faster than backbone
conformational changes (in this case, approximately 10-100
ns24,41). In the current study, approximately 70-80% of the
radicals rapidly recombine within 2 ns. Nevertheless, this yield
has proven adequate to allow in-depth study of conformational
processes in peptide 16.39
One major area of future research would be to further stabilize
the lifetime of the radicals. For example, it may be possible to
attach a triplet sensitizer onto or adjacent to Aty. One could
then photolyze the triplet sensitizer chromophore selectively,
which would then transfer energy intramolecularly to the thiyl
chromophore, and lead to cleavage of the disulfide bond in the
triplet state.48 Then, the two triplet thiyl radicals would be
prevented from rapid recombination by the spin barrier, thereby
increasing their lifetime. The photophysics of such a system
is more complicated and less direct than that of Aty, but may
be well worthy of an investigation to increase the efficiency of
our triggering methodology.
Figure 5. Difference spectrum of cyclic peptide 16 subtracted from
that of linear peptide 15 in the amide I′ region at 10 °C. Because of
the different concentrations of the two peptides used in the measure-
ments, the spectra of 15 and 16 were normalized using TFA as the
internal standard (TFA is present in the same molar ratio in both peptide
samples).
heating. These changes in the IR spectra are very similar to
changes observed for a 21-residue polyalanine peptide.24
To model the expected IR spectral changes upon photocleav-
age of peptide 16, the IR spectrum of 16 was compared with
that of the linear peptide 15. The difference spectra of cyclic
peptide 16 subtracted from that of linear peptide 15 at 10 °C is
shown in Figure 5. In the amide I′ region, conformational
changes resulting from the cleavage of the SS bond lead to an
increase in intensity of the CO absorption in the helical region
concomitant with a decrease in intensity of the CO absorption
in the coil region. These IR data are consistent with the CD
data and indicate that the designed structural change of the
peptide upon SS bond cleavage has been fulfilled.
Photochemistry of Cyclic Peptide 16. Peptide 16 was
photolyzed at 270 nm, and the transient absorption recorded.
Analogous to thiotyrosine-containing peptide 7, transient ab-
sorption at 580 nm was observed immediately upon photolysis
of peptide 16. A rapid decay is observed in the first 50 ps,
during which 53% of the initial thiyl absorbance is consumed,
presumably due to geminate recombination. However, in
contrast to peptide 7, the residual transient signal then decays
in a relatively slow, nonexponential manner from tens of
picoseconds to microseconds. For example, at 2 and 50 ns
respectively, 20 and 9% of the transient absorption can still be
detected. Thus, the introduction of an amino substitution para
to the thiol group leads to a significant percentage of long-
lived thiyl radicals, allowing peptide folding to take place. The
thiyl radicals at the ends of the peptide chain can diffuse apart
to give long-lived transients in the nanosecond time scale. This
increase in radical lifetime should allow for the observation of
protein dynamics in the time range of picoseconds to hundreds
of nanoseconds.
Aty may find broad applications for a variety of studies of
peptide and protein conformational change. In the current work,
we demonstrate that a cyclic Aty disulfide can constrain a
peptide to a nonhelical conformation, which relaxes to a helix
upon cleavage to the linear form. In a similar manner, one might
use this group to initiate folding of an entire protein through
the incorporation of an intramolecular disulfide bond that
destabilizes the native fold. Alternatively, another application
could involve incorporating Aty in the protein interior and
forming a mixed disulfide with an extraneous sulfhydryl group.
The presence of the extraneous group might prevent the protein
from folding. Upon photocleavage of the disulfide bond, the
extraneous thiyl group would then diffuse away, and the protein
would begin to fold (Scheme 4).
Discussion
We have investigated the feasibility of using an aryl disulfide
cross-linking group as a sub-picosecond optical trigger for
protein folding. Our initial study using thiol-tyrosine as the
optical trigger yielded the unexpected result that the peptide
chain constrains the two incipient thiyl radicals in spatial
proximity for efficient recombination. On the basis of previous
studies of thiyl radical recombination and our own observations,
we can estimate that no peptide motions significant to radical
diffusions occurred before 50 ps. These results are also
consistent with the absorption anisotropy decay of thiyl radicals
Experimental Section
Materials and Methods. Reagents and solvents were obtained from
commercial sources and used without purification. Fmoc-protected
(48) Turro, N. J. Modern Molecular Photochemistry; University Science
Books: Mill Valley, 1991.

Optical Trigger for Initiation of Protein Folding
J. Am. Chem. Soc., Vol. 119, No. 31, 1997 7179
amino acids and PAL resin were purchased from PerSeptive Biosys-
tems. NMR spectra were obtained on a Bruker AMX 300, with TMS
or acetone as the internal standard. Mass spectra were obtained using
a VG Zab-E double-focusing mass spectrometer using a Xenon FAB
gun as the ion source or on a Finnigan MAT 8230. CD spectra were
obtained using an AVIV 62A DS system. FT-IR spectra were collected
with a Bruker IFS 66 system. Analytical HPLC was performed on
Vydac C18 reverse phase column (25 × 0.46 cm i.d.) with a flow rate
of 0.5-1.0 mL/min, monitored at 220 nm. Peptides were purified using
Rainin or Waters HPLC equipment on a Vydac C18 reverse phase
column (25 × 2.5 cm i.d.) with a flow rate of 10 mL/min, monitored
at 220 nm. Eluent used were (A) 0.1% TFA in water and (B) 0.1%
TFA in 90% acetonitrile, 9.9% H2O.
Peptide Synthesis. Peptides were all synthesized using standard
Fmoc chemistry and the stepwise solid phase method on PAL resin
using a Milligen Model 9050 peptide synthesizer. Couplings of the
commonly occurring Fmoc-L-amino acid pentafluorophenyl esters were
accomplished using 1-hydroxybenzotriazole hydrate in DMF, and Fmoc-
L-amino acids were coupled using HBTU/DIEA in DMF as a catalyst.
In general, 4 equiv of commercially available amino acids (based on
loading of the resin) were used. The amino acids were circulated
through the resin for 45 min, and the unnatural amino acid was double
coupled. To stabilize helix formation, end charges were neutralized
by acetylating the N-terminus and amidating the C-terminus of the
peptide. Peptides were cleaved from the resin by stirring the dried
resin for 2 h at room temperature with 9:0.5:0.3:0.2 TFA:thioanisole:
ethanedithiol:anisole mixture. The mixture was filtered and concen-
trated under a nitrogen stream. The filtrate was triturated with cold
ether, and crude peptide was filtered and collected. The crude peptide
was purified to greater than 95% purity using a 2.5 cm Vydac C18 RP
column.
(S)-S-(4-Methylbenzyl)-N-(fluorenylmethyloxycarbonyl)-4′-mer-
captophenylalanine (Fmoc-Tty(4-MeBzl), 4). This compound was
prepared in a four-step procedure beginning with L-phenylalanine. Due
to the air sensitive nature and limited solubility of the intermediates,
the reactions were run in succession and only the final product was
fully characterized.
(S)-4′-Mercaptophenylalanine (2). Using a modified procedure of
Escher et al.,36 L-phenylalanine (5.0 g, 30 mmol) was added to a dried
250 mL three-neck round bottom flask, cooled to -15 °C, and stirred
with a mechanical stirrer. To this was added dropwise via an additional
funnel chlorosulfonic acid (30 mL, 450 mmol), precooled to -15 °C,
under nitrogen. After addition, the mixture was warmed to room
temperature and stirred for 4 h under nitrogen. The resulting light
brown mixture was then added slowly via an additional funnel to a
two-neck flask containing 75 g of ice and 20 mL of H2O with vigorous
stirring, and the mixture was cooled with an ice bath, such that the
temperature of the mixture is maintained at less than 10 °C. A white
solid, the salt of the 4′-chlorosulfonyl derivative of L-phenylalanine,
precipitated during the addition. While the reaction vessel was kept
at less than 10 °C, concentrated HCl (30 mL) was added, followed by
Sn powder (17.8 g, 150 mmol). The mixture was brought to reflux
for 2 h and then cooled to room temperature. The solution was diluted
with 300 mL of H2O, and H2S was bubbled through the mixture until
no more tin sulfide is precipitated. The solution was filtered, and the
filtrate was concentrated to approximately 100 mL under reduced
pressure. The resulting air sensitive thiol 2 was used immediately
without further characterization.
(S)-S-(4-Methylbenzyl)-N-(fluorenylmethyloxycarbonyl)-4′-mer-
captophenylalanine (3). The solution of crude thiol 2 was neutralized
with 4 N NaOH on an ice bath and diluted with 250 mL of EtOH. To
this suspension were added triethylamine (10.4 mL, 76 mmol) and
R-bromo-p-xylene (3.16 g, 17 mmol). The mixture was stirred at room
temperature overnight. The crude mixture was filtered, concentrated
to 200 mL, and cooled to 0 °C. An off-white solid precipitated and
was collected by filtration to yield 4.1 g of crude product. The filtrate
was further concentrated to 100 mL and cooled to 0 °C. Another 0.9
g of off-white solid precipitated out and was collected and combined
with the first batch of products to give a total of 5.0 g (97%) of crude
product.
Ac-Tty(S-4-MeBzl)-(Ala-Ala-Ala-Ala-Lys)3-Tty(S-4-MeBzl)-
CONH2 (5). Thiotyrosine 4 (2 equiv per resin) was attached to PAL
resin (Perceptive Biosystem, 0.6 g, 0.38 mmol/g) by coupling with
HBTU/DIEA (1.2 equiv) for 45 min and then repeating this reaction
to couple unreacted amino group. Subsequent amino acids were added
as described above. The crude peptide was isolated as described above
and purified to greater than 95% pure. HPLC retention time: 48.3
(0-60% B/60 min). Mass spectrum (ESI): m/z 1862.0 [(M + H)+,
calcd 1862.6].
(S)-S-(4-Methylbenzyl)-N-(fluorenylmethyloxycarbonyl)-4′-mer-
captophenylalanine (4). Crude 3 (5.0 g, 17 mmol) was suspended in
17 mL of water and 25 mL of acetonitrile. Triethylamine (1.7g, 17
mmol) was added, followed by 9-(fluorenylmethyloxycarbonyl)-N-
hydroxysuccinimide (5.6g, 17 mmol), and the mixture was stirred
overnight. The pH of the reaction mixture was maintained at 8.5-9.0
by addition of triethylamine. The crude mixture was poured into 20
mL of 1.5 N HCl and 20 mL of ethyl acetate, and layers separated.
The aqueous layer was extracted with ethyl acetate (3 × 100 mL), and
the organic layers were combined, washed with brine, dried over
MgSO4, and concentrated in Vacuo to give a yellow oil. The crude
product was purified by flash column chromatography on silica gel,
using 100:5:1 methylene chloride:methanol:acetic acid as the eluent.
After purification, 4.6 g (53%) of white solid was obtained. 1H NMR
(300 MHz, acetone-d6): δ 7.85 (d, J ) 7.5 Hz, 2H), 7.65 (m, 2H),
7.42-7.13 (m, 10 H), 7.08 (d, J ) 7.5 Hz, 2H), 6.73 (d, J ) 8.6 Hz,
1H), 4.50 (m, 1H), 4.31-4.12 (m, 4H), 4.08 (s, 2H), 3.8 (broad, 1H),
3.23 (m, 1H), 3.00 (m, 1H), 2.25 (s, 3H). 13C NMR (500 MHz, acetone-
d6): δ 173.06, 156.80, 144.96, 142.02, 137.31, 136.35, 135.85, 130.73,
129.79, 129.57, 128.50, 127.89, 126.13, 120.74, 67.14, 55.89, 47.96,
38.56, 37.68, 21.03. HRMS: calcd for C32H29NO4S (M + Na)+
546.1716, found 546.1730.
4-Amino-S-(4-methylbenzyl)mercaptobenzene (9). Triethylamine
(37.0 g, 370 mmol) was added to a stirred solution of 4-aminothiophenol
(50 g, 90% purity, 370 mmol) dissolved in 360 mL of EtOH. To this
was added R-bromo-p-xylene (68.5 g, 370 mmol). The mixture was
stirred at room temperature for 18 h. The solvent was removed in
Vacuo to yield a yellow oil. The oil was dissolved in CH2Cl2 (300
mL), washed with H2O (500 mL) and saturated sodium bicarbonate
solution (500 mL), dried over MgSO4, and concentrated. The crude
product was purified by flash column chromatography on silica gel
using 1:5 ethyl acetate:hexane as the eluent to give a pale yellow solid
(41.1 g, 83%). 1H NMR (300 MHz, CDCl3): δ 7.20 (2H, d, J ) 9
Ac-Tty(SH)-(Ala-Ala-Ala-Ala-Lys)3-Tty(SH)-NH2 (6). Peptide 5,
purified or crude, was treated with anhydrous hydrogen fluoride
containing anisole (1 mL per gram of peptide cleaved) as scavenger at
0 °C for 1 h to remove the p-methylbenzyl protecting group. The crude
product was triturated with ether and purified by RP HPLC on a 2.5
cm Vydac C18 column, using gradient condition 25-55% B/60 min,
flow rate ) 10 mL/min. The purified yield of 6 was 10% from starting
material. HPLC retention time: 37.0 min (0-60% B/60 min). Mass
spectrum (ESI): m/z 1654.1 [(M + H)+, calcd 1653.8].
Cyclo-[Ac-Tty(S)-(Ala-Ala-Ala-Ala-Lys)3-Tty(S)-CONH2] Disul-
fide (7). Peptide 6 (10 mg) was dissolved in 5% acetic acid (10 mL).
The pH of the solution was adjusted to 4 by addition of ammonium
hydroxide. DMSO (2 mL) was added, and the mixture was stirred at
room temperature for 8 h. HPLC analysis showed that the oxidation
is complete at this time. The solution was frozen and lyophilized. The
crude peptide was purified by RP HPLC on a Vydac C18 column with
eluting gradient of 20-40% B/50 min. Purified yield: 7.0 mg (70%).
HPLC retention time: 32.4 min (0-60% B/60 min). Mass spectrum
(ESI): m/z 1651.8 [(M + H)+, calcd 1652.8].
Ac-Aty(S-4-MeBzl)-(Ala-Ala-Ala-Ala-Lys)3-Aty(S-4-MeBzl)-
CONH2 (15). This peptide was prepared as described for peptide 5.
HPLC retention time: 24.2 min (10-100%B/45 min). Mass spectrum
(ESI): m/z 1892.6 [(M + H)+, calcd 1892.0].
Ac-Aty(SH)-(Ala-Ala-Ala-Ala-Lys)3-Aty(SH)-CONH2. This pep-
tide was prepared as described for peptide 6. HPLC retention time:
17.1 (10-100% B/45 min). Mass spectrum (ESI): m/z 1684.0 [(M +
H)+, calcd 1683.9].
Cyclo-[Ac-Aty(S)-(Ala-Ala-Ala-Ala-Lys)3-Aty(S)-CONH2] Dis-
ulfide (16). This peptide was prepared as described for peptide 7.
HPLC retention time: 15.7 min (10-100% B/45 min). Overall yield
of 16 from 0.2 mmol of PAL resin was 10%. Mass spectrum (ESI):
m/z 1682.6 [(M + H)+, calcd 1681.9].

7180 J. Am. Chem. Soc., Vol. 119, No. 31, 1997
Lu et al.
Hz) , 7.13 (4H, dd, J ) 6, 15Hz), 6.65 (2H, d, J ) 9 Hz), 4.65 (br s,
2H), 3.97 (s, 2H), 2.30 (s, 3H). 13C NMR (500 MHz, CO(CD3)2):
148.9, 136.8, 135.1, 132.5, 130.2, 121.8, 114,1, 41.8, 21.1. HRMS:
calcd for C14H15NS (M + H)+ 229.0925, found 229.0921.
2-N-(Fluorenylmethyloxycarbonyl)-3-N-[4′-[(4-benzylmethyl)thio]-
phenyl]-(S)-2,3-diaminoproprionic Acid (Fmoc-Aty(4-MeBzl), 14).
Compound 13 (1.2 g, 3.8 mmol) was dissolved in 35 mL of acetonitrile
and 20 mL of H2O. To this was added lithium hydroxide (0.25 g, 5.9
mmol). The solution was stirred at room temperature for 15 h. The
solution was neutralized with 10% HCl solution. To this was added
9-(fluorenylmethyloxycarbonyl)-N-hydroxysuccinimide, followed by
triethylamine (0.42 g, 4.1 mmol), and the mixture was stirred at room
temperature for 3.5 h. The pH of the mixture was maintained at 8-9.
The solution was diluted with 200 mL of ethyl acetate and 200 mL of
H2O and acidified with 10% HCl solution. The layers were separated,
and the aqueous layer was extracted with ethyl acetate (2 × 200 mL).
The ethyl acetate layers were combined, dried over MgSO4, and
concentrated. The crude product was purified by flash column
chromatography using 100:7.5:1 methylene chloride:methanol:acetic
acid as the eluent to yield 2.0 g (two steps from 7, 99%) of white
4-[(2-Mesitylenesulfonyl)amino]-1-[(p-methylbenzyl)thio]ben-
zene (10). Triethylamine (7.8 g, 78 mmol) was added to a stirred
solution of 9 (7.0 g, 31 mmol) in 30 mL of CH2Cl2. To this was added
2-mesitylenesulfonyl chloride (8.5 g, 39 mmol), and the solution was
stirred at room temperature for 18 h. The solvent was evaporated in
Vacuo. The crude product was dissolved in ethyl acetate (500 mL),
washed with 10% HCl solution (5 × 500 mL), dried over MgSO4, and
concentrated. Purification of the crude product was achieved by flash
column chromatography on silica gel using 1:5 ethyl acetate:hexane
as the eluent and yielded 9.0 g (72%) of white solid. 1H NMR (300
MHz, CDCl3): δ 7.18 (d, 2H, J ) 9 Hz), 7.08 (m, 4H), 6.91 (s, 2H),
6.84 (d, 2H, J ) 9 Hz), 6.43 (br s, 1H), 3.98 (s, 2H), 2.57 (s, 6H), 2.31
(s, 3H), 2.28 (s, 3H). 13C NMR (500 MHz, CO(CD3)2): δ 143.1, 141.9,
139.9, 137.2, 137.0, 135.4, 135.0, 132.6, 131.6, 129.7, 129.5, 121.7,
39.0, 23.0, 21.0, 20.8. HRMS: calcd for C23H25NO2S2 (M + NH4)+
429.1671, found 429.1665.
1
solid. H NMR (300 MHz, CO(CD3)2): δ 7.82 (2H, d, J ) 8 Hz),
7.08 (2H, d, J ) 7 Hz), 7.35 (m, 4H), 7.05 (d, 2H, J ) 9 Hz), 7.07
(dd, 2H, J ) 7, 8 Hz), 6.80 (br, 1H), 6.66 (2H, d, J ) 9 Hz), 4.53 (m,
1H), 4.34 (d, 2H, J ) 7 Hz), 4.24 (m, 1H), 3.92 (s, 2H), 3.60 (m, 2H),
2.26 (s, 3H). 13C NMR (500 MHz, CO(CD3)2): δ 172.63, 157.12,
148.86, 145.08, 142.14, 137.00, 136.64, 135.11 (2 C’s), 129.70, 128.60,
128.03, 126.18, 122.47, 120.84, 114.08, 67.36, 54.50, 48.06, 45.72,
41.76, 21.13.
2-N-Trityl-3-N-(2-mesitylenesulfonyl)-3-N-[4′-[(4-benzylmethyl)-
thio]phenyl]-(S)-2,3-diaminoproprionate Methyl Ester (12). N-
Trityl-L-serine methyl ester (11) (5.5 g, 15 mmol) was dissolved in
100 mL of anhydrous benzene and stirred at room temperature under
nitrogen. To this was added 12 (5.2g, 13 mmol) followed by
triphenylphosphine (4.4 g, 17 mmol). The mixture was stirred at room
temperature for 5 min. Diethyl azodicarboxylate (2.6 mL, 17 mmol)
was added dropwise via syringe under nitrogen. The reaction mixture
was stirred for 24 h at room temperature. Solvent was removed in
Vacuo to yield a yellow oil. The crude product was purified by flash
column chromatography on silica gel using 1:7 ethyl acetate:hexane
as the eluent initially and then stepping the eluent to 1:5 ethyl acetate:
hexane. A white solid (7.4 g, 78%) was obtained after purification.
1H NMR (300 MHz, CO(CD3)2): δ 7.39 (m, 6H), 7.21 (m, 17H), 6.93
(s, 2H), 4.20 (m, 2H), 4.12 (s, 2H), 3.41 (m, 1H), 3.13 (s, 3H), 2.43 (s,
6H), 2.31 (s, 3H), 2.23 (s, 3H). 13 C NMR (500 MHz, CO(CD3)2): δ
173.74, 146.74, 143.63, 141.14, 138.20, 137.66, 137.58, 135.26, 133.82,
132.78, 132.21, 130.39, 130.06, 129.66, 129.55, 128.78, 128.21, 127.39,
125.81, 72.18, 55.82, 56.74, 55.31, 52.14, 38.39, 30.44, 30.28, 23.62,
21.45, 21.19. HRMS: calcd for C46H46N2O4S2 (M + Na)+ 777.2197,
found 777.2181.
Circular Dichroism. Spectra were recorded on an Aviv 62 DS
spectropolarimeter. A circulating water bath was used to control the
temperature. Concentrations of the stock peptide solutions were
determined in duplicates or triplicates by amino acid analysis.
FT IR. Spectra were recorded on a Bruker IFS 66 equipped with
a liquid N2 MCT detector. A temperature-controlled cell with CaF2
windows and adjustable pathlength, set at 50 µm, was used. At each
temperature, the sample spectrum was corrected for D2O absorption.
For the measurements of peptide 16, a stock solution of 13 mg/mL in
D2O was made just prior to the experiment. Due to the hydrophobic
protecting group, 4-methylbenzyl, the same concentration of peptide
15 could not be achieved. Instead, a saturated solution of 15 in D2O
was used. The difference spectra of 15 and 16 were scaled using TFA
as the internal standard.
Modeling. Modeling was carried out using Insight II (Biosym/MSI,
San Diego, CA) on a Silicon Graphics Personal Iris. Energy minimiza-
tion and molecular dynamics simulations were carried out using
Discovery 3.0. The initial peptide structures were generated in InsightII
using Biopolymer. Minimization and molecular dynamics calculations
were carried out using constant dielectric of 1. The structures were
minimized first using 1000 cycles of steepest descent and 10 000 cycles
of conjugate gradients. Starting from these structures, molecular
dynamics simulation were carried out by first equilibrating at 120 K
for 500 fs (1 fs step size) and running dynamics for 20 ps, collecting
structures every 2.5 ps. The resulting structures from dynamics were
then subjected to minimization (1000 steps of steepest descent and
20 000 steps of conjugate gradients). This process was also repeated
at 300 K. The structures were pooled and superimposed. Dynamic
runs at 120 and 300 K gave the same gross structural features such as
helical content of the peptides.
3-N-[4′-[(4-Benzylmethyl)thio]phenyl]-(S)-2,3-Diaminoproprion-
ate Methyl Ester (13). To a 100 mL round bottom flask containing
compound 12 (3.6 g, 4.8 mmol) was added phenol (13 g), followed by
50 mL of 30% HBr in acetic acid. The mixture was stirred at room
temperature for 15 h. Excess HBr was blown into a H2O trap under a
stream of nitrogen. The reaction mixture was concentrated in Vacuo
to 5-7 mL. Ether, precooled to -78 °C, was added to yield an orange
precipitate. The precipitate was collected by filtration and washed with
cold ether (2 × 50 mL). The solid was dissolved in 200 mL of ethyl
acetate and 200 mL of saturated sodium bicarbonate. The layers were
separated, and the aqueous layer was extracted with ethyl acetate (2 ×
100 mL). The organic layers were combined, dried over MgSO4, and
concentrated. The crude product was purified by flash column
chromatography on silica gel, using 100:5 methylene chloride:methanol
as the eluent, to yield 1.2 g (87%) of yellow oil. 1H NMR (300 MHz,
CO(CD3)2): δ 7.13 (d, 2H, J ) 9 Hz), 7.08 (m, 4H), 6.61 (d, 2H, J )
9 Hz), 5.14 (br m, 1H), 3.92 (s, 2H), 3.41 (m, 1H), 3.20 (m, 1H), 2.84
(br s, 1H), 2.28 (s, 3H), 1.78 (br, 2H). 13C NMR (500 MHz,
CO(CD3)2): δ 171.9, 149.0, 136.7, 135.2, 134.2, 129.6, 121.8, 114.0,
63.0, 52.0, 46.8, 41.7, 21.0. HRMS: calcd for C18H22N2O2S (M +
H)+ 331.1480, found 331.1492.
Acknowledgment. H.S.M.L. was supported in part by the
Dupont Merck Postdoctoral Program. We thank Mary Boyd and
Robert Cherney for helpful discussions and Joel Schneider and
Blake Hill for critical reading of the manuscript. We thank Steve
Koch for providing 2,3,5,6-tetramethylphenyl disulfide and
2,4,6-triisopropylphenyl disulfide.
JA970567O