Intramolecular quenching of tryptophan fluorescence by the peptide bond in cyclic hexapeptides

Article abstract of DOI:10.1021/ja0167710

Intramolecular quenching of tryptophan fluorescence by protein functional groups was studied in a series of rigid cyclic hexapeptides containing a single tryptophan. The solution structure of the canonical peptide C[D-PpYTFWF] (pY, phosphotyrosine) was de

Full text of DOI:10.1021/ja0167710

Published on Web 07/10/2002
Intramolecular Quenching of Tryptophan Fluorescence by the
Peptide Bond in Cyclic Hexapeptides
Paul D. Adams,‡ Yu Chen,† Kan Ma,‡ Michael G. Zagorski,‡ Frank D. So¨nnichsen,§
Mark L. McLaughlin, and Mary D. Barkley*,‡
Contribution from the Departments of Chemistry and Physiology and Biophysics,
Case Western ReserVe UniVersity, CleVeland, Ohio 44106, and Department of Chemistry,
Louisiana State UniVersity, Baton Rouge, Louisiana 70803
Received August 6, 2001
Abstract: Intramolecular quenching of tryptophan fluorescence by protein functional groups was studied
in a series of rigid cyclic hexapeptides containing a single tryptophan. The solution structure of the canonical
peptide c[D-PpYTFWF] (pY, phosphotyrosine) was determined in aqueous solution by 1D- and 2D-1H NMR
techniques. The peptide backbone has a single predominant conformation. The tryptophan side chain has
three ø1 rotamers: a major ø1 ) -60° rotamer with a population of 0.67, and two minor rotamers of equal
population. The peptides have three fluorescence lifetimes of about 3.8, 1.8, and 0.3 ns with relative
amplitudes that agree with the ø1 rotamer populations determined by NMR. The major 3.8-ns lifetime
component is assigned to the ø1 ) -60° rotamer. The multiple fluorescence lifetimes are attributed to
differences among rotamers in the rate of excited-state electron transfer to peptide bonds. Electron-transfer
rates were calculated for the six preferred side chain rotamers using Marcus theory. A simple model with
reasonable assumptions gives excellent agreement between observed and calculated lifetimes for the 3.8-
and 1.8-ns lifetimes and assigns the 1.8-ns lifetime component to the ø1 ) 180° rotamer. Substitution of
phenylalanine by lysine on either side of tryptophan has no effect on fluorescence quantum yield or lifetime,
indicating that intramolecular excited-state proton transfer catalyzed by the ꢀ-ammonium does not occur in
these peptides.
Tryptophan fluorescence has been used extensively to study
peptide and protein conformation changes, due to the sensitivity
of both emission wavelength and intensity to the local environ-
ment of the indole chromophore. The wavelength of maximum
emission ranges from about 310-350 nm, depending on the
electrostatic environment.1 The fluorescence quantum yield
ranges from near zero to 0.35 with lifetimes up to 10 ns.2,3 The
large variations in intensity and fluorescence lifetime τ are due
to nonradiative pathways that compete with emission for
deactivation of the excited state.
transfer kpt, and excited-state electron transfer ket, which depend
on the proximity of water, proton donors, and electron
acceptors.4-7
knr ) kisc + ksi + kpt + ket
(2)
Six protein functional groups representing eight amino acid side
chains plus the peptide bond quench 3-methylindole fluorescence
in intramolecular quenching experiments.8 Most peptides and
proteins containing a single tryptophan have multiexponential
fluorescence decays. Although excited-state processes are still
invoked to account for the complexity,9-12 the multiple fluo-
rescence lifetimes are generally attributed to ground-state
heterogeneity.13 The rotamer model proposes that the multiple
lifetimes represent multiple ground-state rotamers of the tryp-
τ-1 ) kr + knr
(1)
where kr and knr are the radiative and nonradiative rates. In
addition to the intersystem crossing rate kisc, the nonradiative
rate of indole contains contributions from three environmentally
sensitive processes, solvent quenching ksi, excited-state proton
(4) Saito, I.; Sugiyama, H.; Yamamoto, A.; Muramatsu, S.; Matsuura, T. J.
Am. Chem. Soc. 1984, 106, 4286-4287.
(5) McMahon, L. P.; Colucci, W. J.; McLaughlin, M. L.; Barkley, M. D. J.
Am. Chem. Soc. 1992, 114, 8442-8448.
* Address correspondence to this author at Department of Chemistry,
Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH
44106-7078. Fax: 216-368-0604. E-mail: mdb4@po.cwru.edu.
‡ Department of Chemistry, Case Western Reserve University.
† Current address: BD Pharmingen, 10975 Torreyana Road, San Diego,
CA 92121.
(6) Yu, H.-T.; Colucci, W. J.; McLaughlin, M. L.; Barkley, M. D. J. Am. Chem.
Soc. 1992, 114, 8449-8454.
(7) Chen, Y.; Liu, B.; Yu, H.-T.; Barkley, M. D. J. Am. Chem. Soc. 1996,
118, 9271-9278.
(8) Chen, Y.; Barkley, M. D. Biochemistry 1998, 37, 9976-9982.
(9) Ladokhin, A. S. J. Fluoresc. 1999, 9, 1-9.
§ Department of Physiology and Biophysics, Case Western Reserve
University.
(10) Hudson, B. S.; Huston, J. M.; Soto-Campos, G. J. Phys. Chem. A 1999,
103, 2227-2234.
Louisiana State University.
(11) Moncrieffe, M. C.; Juranic, N.; Kemple, M. D.; Potter, J. D.; Macura, S.;
Prendergast, F. G. J. Mol. Biol. 2000, 297, 147-163.
(12) Nanda, V.; Brand, L. Proteins 2000, 40, 112-125.
(13) Sillen, A.; D´ıaz, J. F.; Engelborghs, Y. Protein Sci. 2000, 9, 158-169.
(1) Callis, P. R.; Burgess, B. K. J. Phys. Chem. B 1997, 101, 9429-9432.
(2) Beechem, J. M.; Brand, L. Annu. ReV. Biochem. 1985, 54, 43-71.
(3) Eftink, M. R. Methods Biochem. Anal. 1991, 35, 127-205.
9
9278
J. AM. CHEM. SOC. 2002, 124, 9278-9286
10.1021/ja0167710 CCC: $22.00 © 2002 American Chemical Society

Tryptophan Fluorescence in Hexapeptides
A R T I C L E S
tophan side chain.14,15 There are six low-energy canonical
conformations dictated by the torsional angles about the CR-
Câ (ø1 ) (60°, 180°) and Câ-Cγ (ø2 ) (90°) bonds. If prox-
imity of the indole ring to water or quenching functional groups
differs among rotamers, those rotamers will have different
fluorescence lifetimes. If interconversion among rotamers is
slower than the fluorescence time scale, the fluorescence decay
will be multiexponential with relative amplitudes proportional
to the rotamer populations. For small peptides in solution or
bound to membrane vesicles, several groups have proposed that
the two or three fluorescence decay components represent ø1
rotamers.16-19 The three decay amplitudes of [Trp2]oxytocin
were correlated with ø1 rotamer populations determined by 1H
NMR.20 In addition to rotamers of the tryptophan side chain,
microconformational states of peptides and proteins with
different local environments of the indole ring constitute a source
of ground-state heterogeneity. Finally, relaxation of the protein
matrix surrounding the indole ring during the lifetime of the
excited state would also produce a complex decay.
This contribution tests the rotamer model in a cyclic hexapep-
tide containing a single tryptophan and five other amino acids
whose side chains do not quench 3-methylindole fluorescence.
The cyclic peptide is closely related to a somatostatin analogue
previously shown to have a rigid backbone in organic solvent.21
The solution structure of the cyclic peptide in aqueous solution
is determined by 1D- and 2D-1H NMR. The tryptophan
fluorescence is characterized by steady-state and time-resolved
techniques, and the dependence on solvent isotope and tem-
perature is determined. Fluorescence decay amplitudes are
correlated with the NMR-determined ø1 rotamer populations of
the tryptophan side chain. Excited-state electron-transfer rates
for intramolecular quenching of the six ø1, ø2 rotamers by the
peptide bonds are calculated from Marcus theory using distances
estimated by molecular modeling and parameters estimated from
the electron-transfer rate of N-acetyltryptophanamide (NATA).
The fluorescence lifetimes of the six rotamers are calculated
using these electron-transfer rates together with experimental
values for the intersystem-crossing and water-quenching rates
previously determined for NATA.7 The observed fluorescence
decay is interpreted in terms of the ø1 rotamer populations and
the calculated ø1, ø2 rotamer lifetimes.
was purchased from Perceptive Biosystems. All other reagents and
solvents were highest grade available.
Synthesis. Peptides were synthesized manually using oxime resin
at 25 °C.23 The oxime resin offers two advantages for the synthesis of
small cyclic peptides containing tryptophan: simultaneous cyclization
and cleavage from the resin under mild conditions and efficient isolation
from multimeric side products.22,23 The disadvantage of using Boc resins
is that the repetitive trifluoroacetic acid (TFA) deprotection may result
in modification of peptide side chains. The extent of cyclization using
the oxime resin is highly sequence-dependent.24 Therefore, the amino
acid that was coupled to the resin was permuted in the sequence c[D-
PpYTFWF] except for D-Pro1 and pTyr2 (phosphotyrosine, pY). The
synthesis starting with Thr3 gave the best overall yield of purified
peptide. Thr (3 mmol) was coupled to 1 g of resin (substitution level
0.3-0.5 mmol/g) using 0.35 mM diisopropylcarbodiimide in 6 mL of
CH2Cl2 for 24 h. The solvent was removed, and unreacted sites on the
oxime resin were capped by addition of 6 mL of dry dimethylformamide
(DMF) containing 0.2 M acetic or trimethylacetic anhydride and 0.3
M diisopropylethylamine (DIEA). The coupled amino acid was
deprotected with 25% TFA in 4-6 mL of CH2Cl2 for 15-20 min.
Amino acid coupling reactions were carried out by shaking at room
temperature for 45 min in 6 mL of dry DMF containing 0.25 M HBTU,
0.1 M HOBT, and 0.5 M DIEA, except for D-Pro where HATU replaced
HBTU. The extent of coupling was monitored by a ninhydrin color
test on a small aliquot of resin after each step.25 After the last
deprotection, the peptide was cyclized and cleaved from the resin by
reaction for 48 h with 0.5 M acetic acid and 1.5 M triethylamine in
6-8 mL of dry DMF.26 For c[D-PpYTFWF] and c[D-PpYTFLF], the
organic phase was removed by filtration and added dropwise to cold
diethyl ether, and the white precipitate was removed by centrifugation.
Alternatively, the DMF was removed by vacuum distillation, which
gave a viscous orange residue. For c[D-PFTK(z)WF] and c[D-PFTFWK-
(z)], the organic layer was added dropwise to cold water, and the white
precipitate was removed by centrifugation. The resulting pellet or
residue was dissolved in a small volume of 1:1 CH3CN/H2O and
purified by RP-HPLC on a Vydac C-18 semi-prep column using a
gradient of CH3CN/0.1% TFA: 20-55% for 60 min for c[D-Pp-
YTFWF] and 40-100% for 40 min for the other peptides.
Deprotection of the ꢀ-amino group of lysine was done by catalytic
hydrogenation. The peptide was dissolved in a minimal volume (5-
10 mL) of glacial acetic acid and placed in a hydrogenation flask. A
catalytic amount of 10% Pd on carbon was added, and the suspension
was purged with argon three times and then charged to 50 psi with
hydrogen and shaken overnight. After the hydrogenation was complete,
the Pd catalyst was filtered off, and the flask was washed five times
with small aliquots of ethyl acetate. The ethyl acetate/acetic acid solvent
was removed under vacuum to yield the crude, yellow deprotected
peptide. The peptides were purified by HPLC using a gradient of 20-
55% CH3CN/0.1% TFA for 60 min.
Peptide molecular weights were verified by MALDI-MS. Yields of
purified peptide were 8-17% relative to the substitution level of the
oxime resin. The low yields are probably due to TFA-mediated side
chain reactions or cleavage of the linear sequence during the cyclization
step.
Experimental Section
Materials. Oxime resin was either purchased from Novabiochem
or synthesized.22 N-R-t-Boc-D-proline, -L-phenylalanine, -L-tryptophan,
-L-threonine, and -L-leucine, N-R-t-Boc-N-ꢀ-benzyloxycarbonyl-L-lysine,
2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophos-
phate (HBTU), and N-hydroxybenzotriazole‚H2O (HOBT) were pur-
chased from Novabiochem. N-R-t-Boc-L-phosphotyrosine‚diisopropyl-
ethylamine was purchased from Advanced Chemtech. O-(7-azaben-
zotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HATU)
1H Nuclear Magnetic Resonance. All NMR experiments were
performed using a Varian Inova 600 MHz spectrometer. For experi-
ments in H2O, 2 mg of peptide was dissolved in 600 µL of 10 mM
phosphate buffer, 1 mM NaN3, 10% D2O and pH was adjusted to 3-5
with dilute HCl. Chemical shifts were referenced to 3-(trimethylsilyl)-
propanate-2,2,3,3-d4 (TSP). For experiments in 100% D2O, 1 mg of
(14) Donzel, B.; Gauduchon, P.; Wahl, P. J. Am. Chem. Soc. 1974, 96, 801-
808.
(15) Szabo, A. G.; Rayner, D. M. J. Am. Chem. Soc. 1980, 102, 554-563.
(16) Willis, K. J.; Szabo, A. G. Biochemistry 1992, 31, 8924-8931.
(17) Willis, K. J.; Neugebauer, W.; Sikorska, M.; Szabo, A. G. Biophys. J. 1994,
66, 1623-1630.
(18) Dahms, T. E. S.; Willis, K. J.; Szabo, A. G. J. Am. Chem. Soc. 1995, 117,
2321-2326.
(23) Osapay, G.; Bouvier, M.; Taylor, J. W. Technol. Protein Chem. 2 1991,
221-231.
(19) Clayton, A. H.; Sawyer, W. H. Biophys. J. 1999, 76, 3235-3242.
(20) Ross, J. B. A.; Wyssbrod, H. R.; Porter, R. A.; Schwartz, G. P.; Michaels,
C. A.; Laws, W. R. Biochemistry 1992, 31, 1585-1594.
(21) Kessler, H.; Bats, J. W.; Griesinger, C.; Koll, S.; Will, M.; Wagner, K. J.
Am. Chem. Soc. 1988, 110, 1033-1049.
(24) Nishino, N.; Xu, M.; Mihara, H.; Fujimoto, T. Tetrahedron Lett. 1992, 33,
1479-1482.
(25) Kaiser, E. T.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal. Biochem.
1970, 34, 595-598.
(22) DeGrado, W. F.; Kaiser, E. T. J. Org. Chem. 1980, 45, 1295-1300.
(26) Gisin, B. F.; Merrifield, R. B. J. Am. Chem. Soc. 1972, 94, 3102-3106.
9
J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002 9279

A R T I C L E S
Adams et al.
peptide was dissolved in 270 µL of D2O (Cambridge Isotope) in a
Shigemi NMR tube. Chemical shifts δ of amide protons were measured
in 5° increments from 5 to 55 °C. Temperature coefficients -dδ/dT
were determined from the slope of a plot of chemical shift vs
temperature by least-squares.
indole ring and each peptide group was measured in the Decipher
module. Twelve starting structures for NATA were constructed in
Biopolymer with the φ,ψ angles rotated in 30° increments from 0 to
360° to simulate unrestricted backbone rotation. The mean structure
of c[D-PpYTFWF] determined above was used as starting structure for
the cyclic peptide. Six tryptophan rotamers were generated for each
starting structure. The resulting structures were minimized in Discover
with ø1 and ø2 constrained to canonical values, and the distances and
angles between the indole ring and the peptide groups were measured.
Fluorescence Spectroscopy. Steady-state and time-resolved fluo-
rescence measurements were made as before.5,7,37 All experiments were
performed on aqueous solutions with absorbance <0.1 at 280 nm.
Absorbance was measured on a Cary 3E UV-vis spectrophotometer.
Steady-state fluorescence was measured on a SLM 8000 photon
counting spectrofluorimeter in ratio mode. Fluorescence quantum yields
Φ were determined at 280- and 295-nm excitation wavelength (4-nm
band-pass). Quantum yields were measured relative to tryptophan
(Sigma, recrystallized four times from 70% ethanol) in H2O using a
value of 0.14 at 25 °C.38 The absorbance of Trp5-containing peptides
was corrected for contributions from pTyr and Phe by subtracting the
absorbances at 280 and 295 nm of a control peptide c[D-PpYTFLF] at
the same concentration. Peptide concentrations were determined by
quantitative amino acid analysis. The absorbance of the control peptide
was <5% of Trp5-containing peptides at 280 nm and negligible at 295
nm.
Time-resolved fluorescence was measured by time-correlated single
photon counting using a rhodamine 6G dye laser excitation source.
Excitation wavelength was 295 nm. Fluorescence decays were collected
at 10-nm increments from 320 to 390 nm emission wavelength (8-nm
band-pass). Temperature dependence data were collected at 360 nm
emission wavelength in 5° increments from 5 to 55 °C. Data were
acquired from the fluorescent sample and a solution of coffee creamer
in water for the instrumental response to 1.8-2.5 × 104 counts in the
peak. Decay data were stored in 1024 channels of 10 and 24 ps/channel.
Decay curves were deconvolved assuming a discrete sum of
exponentials in the Beechem global analysis program39
Spin-type assignments for c[D-PpYTFWF] were made from TOCSY
spectra27,28 that used a mixing time of 70 ms, spectral width of 7000.4
Hz, and 2048 data points in 256 t1 increments of 24 transients. Water
suppression was performed by low-power irradiation with a presatu-
ration delay pulse of 1 s. NOESY spectra29 used mixing times up to
300 ms and WATERGATE water suppression sequence.30 NOESY
spectra were acquired with 1024 data points in 256 t1 increments of 32
transients. Interproton distances from NOESY experiments were
calculated using the isolated two-spin approximation theory.31
A
reference distance of 2.82 Å from the indole NH to H7 of the indole
ring was taken from the crystal structure.32 PE-COSY spectra33 were
acquired with a spectral width of 6600 Hz and 4096 data points in 440
t1 increments of 80 transients.
NMR data were processed using the Felix program (Molecular
Simulations, Inc.) on an INDY-IP22 work station (Silicon Graphics,
Inc.). Convolution difference water deconvolution was used to remove
the residual water signal in the spectra. Data were routinely zero-filled
to 2K data points in both dimensions, with the exception of the PE-
COSY, which was filled to 8K in D1 dimension and 2K in D2
dimension. In all spectra, a Gaussian and a 90°-shifted sinebell squared
apodization function were used in the D1 and D2 dimension, respec-
tively. The data were then phase corrected and baseline corrected.
Structure calculations were performed using the NMR-Refine module
of InsightII (Molecular Simulations, Inc.). A set of 30 structures was
generated in DGII from 59 NOE-derived distance restraints comprising
23 intraresidue, 23 sequential, and 13 nonadjacent restraints, two
hydrogen bond restraints, 13 3J dihedral angle restraints, and six chiral
restraints.34 The NOE restraints were grouped as strong, medium, and
weak, which corresponded to 1.8-2.7, 1.8-3.3, and 1.8-5.0 Å distance
ranges. The two hydrogen bond restraints were incorporated on the
basis of the reduced temperature coefficients for the amide protons of
Thr3 and Phe6. They were defined as O‚‚‚N distance ranges of 2.6-
3.2 Å between residues 3 and 6.35 These restraints did not bias the
structure calculations as judged by repeating the calculations without
hydrogen bond restraints. The DGII calculation was done using 10,000
steps of simulated annealing and 500 steps of conjugate gradient
optimization to an RMS gradient of 0.001. Structures were further
minimized in Discover using the CVFF force field and a dielectric
constant of 78.0 for water.36 A 500-step steepest descent minimization
without cross and Morse potentials was performed until the maximum
derivative was 10.0 kcal/Å, followed by a 1000-step conjugate gradients
minimization until the maximum derivative was 1.0 kcal/Å. Finally, a
5000-step VA09A minimization using cross and Morse terms was
performed until the maximum derivative was 0.001 kcal/Å. The final
ensemble consisted of the 17 lowest-energy structures.
I(λ,t) )
Ri(λ) exp(-t/τi)
(3)
∑
i
where Ri(λ) is the amplitude at wavelength λ and τi is the lifetime of
component i. Lifetimes but not amplitudes were linked in global
analyses. Individual decay curves were also deconvolved by the
maximum entropy method (MEM).40
I(λ,t) ) ∞ R(λ,τ) exp(-t/τ) dτ
(4)
∫
0
where R(λ,τ) is the fluorescence lifetime distribution at wavelength λ.
Quality of fit was judged by the value of reduced ør2, the weighted
residuals, and the shape of the autocorrelation function of the weighted
residuals. Decay curves were simulated with 1.8 × 104 counts in the
peak using experimental instrumental response functions of 10 and 24
ps/channel in the Beechem program.The radiative rate kr is calculated
from
Molecular Modeling. Six canonical rotamers of the tryptophan side
chain in NATA and c[D-PpYTFWF] were modeled having ø1 ) (60°,
180° and ø2 ) (90°. The distance from a pseudoatom between C8
and C9 of the indole ring to each carbonyl carbon in the molecule was
measured in InsightII. The dihedral angle between the planes of the
kr ) Φ/τj
where τj is the average lifetime
τj ) Ri(λ) τi/
(5)
(27) Bax, A.; Davis, D. G. J. Magn. Reson. 1985, 65, 355-360.
(28) Shaka, A. S.; Lee, C. J.; Pines, A. J. Magn. Reson. 1988, 77, 274-293.
(29) Kumar, A.; Ernst, R. R.; Wuthrich, K. Biochem. Biophys. Res. Commun.
1980, 95, 1-6.
(30) Piotto, M.; Saudek, V.; Sklenar, V. J. Biomol. NMR 1992, 2, 661-665.
(31) Wu¨thrich, K. NMR of Proteins and Nucleic Acids; Wiley: New York, 1986.
(32) Bye, E.; Mostad, A.; Romming, C. Acta Chem. Scand. 1973, 27, 471-
484.
(33) Mueller, L. J. Magn. Reson. 1987, 72, 191-196.
(34) Havel, T. F.; Kuntz, I. D.; Crippen, G. M. Bull. Math. Biol. 1983, 45, 665-
720.
(35) Talafous, J.; Marcinowski, K. J.; Klopman, G.; Zagorski, M. G. Biochem-
istry 1994, 33, 7788-7795.
(36) NMRchitect; Biosym/MSI: San Diego, 1995.
Ri(λ)
(6)
∑
∑
i
i
Results and Discussion
Backbone Conformation of c[D-PpYTFWF]. The ground-
state structure of a closely related peptide, c[D-PFTK(z)WF],
(37) Liu, B.; Thalji, R. K.; Adams, P. D.; Fronczek, F. R.; McLaughlin, M. L.;
Barkley, M. D. J. Am. Chem. Soc. 2001, in revision.
(38) Chen, R. F. Anal. Lett. 1967, 1, 35-42.
(39) Beechem, J. M. Chem. Phys. Lipids 1989, 50, 237-251.
(40) Brochon, J.-C. Methods Enzymol. 1994, 240, 262-311.
9
9280 J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002

Tryptophan Fluorescence in Hexapeptides
A R T I C L E S
Table 1. NMR Data
−dδ/dT
3JNH-HR
3JHR-Hâr
3JHR-Hâs
(ppb/deg)
(Hz)
(Hz)
(Hz)
pI
pII
pIII
c[D-PpYTFWF] in water
D-Pro1
pTyr2
Thr3
Phe4
Trp5
Phe6
8.7
3.9
5.5
4.6
0.9
8.5
9.6
3.3
8.4
7.8
10.2
4.8
0.70 0.19 0.11
6.8
9.9
7.9
4.4
0.38 0.57 0.05
0.67 0.16 0.17
c[D-PFTK(z)WF] in DMSOa
D-Pro1
Phe2
7.0
3.3
4.3
3.0
1.0
8.9
9.7
3.4
8.9
5.8
11.5
3.1
0.80 0.05 0.15
Thr3
Lys(z)4
Trp5
10.9
4.5
4.2
9.0
0.75 0.15 0.10
0.17 0.58 0.25
Phe6
a Data from ref 21.
the Hδ protons of D-Pro1.41 The 1D-spectra of c[D-PFTKWF]
and c[D-PFTFWK] had NH and HR chemical shifts very similar
to those of c[D-PpYTFWF].
The â-turns of the Kessler peptide are preserved in c[D-
PpYTFWF], with a âI turn about Phe4-Trp5 and a âII′ turn about
3
D-Pro1-pTyr2. Vicinal JNH-HR coupling constants from 1D-
spectra of c[D-PpYTFWF] in H2O are similar to the coupling
constants for c[D-PFTK(z)WF] in DMSO (Table 1). The upfield
shifted D-Pro1 HR resonance at 4 ppm is additional evidence
for a â-turn about D-Pro1-pTyr2.42 The small temperature
coefficient of the Phe6-NH chemical shift suggests one intramo-
lecular hydrogen bond at Thr3-CO‚‚‚NH-Phe6 as found in the
Kessler peptide (Table 1). The temperature coefficient of Thr3-
NH in the intermediate range is consistent with the other
hydrogen bond suggested by Kessler.21
Two facts point to a single backbone conformation for c[D-
PpYTFWF] in aqueous solution. First, there is no detectable
cis-trans isomerization of the Phe6-D-Pro1 peptide bond in the
NMR data. Second, four of five 3JNH-HR coupling constants in
Table 1 fall outside the 6-8 Hz range attributed to conforma-
tional averaging. The overall structure of c[D-PpYTFWF] was
generated from distance geometry calculations using restraints
derived from NOE and J coupling data. There were no NOE
distance restraint violations greater than 0.5 Å in the distance
geometry calculations. The average structure from the family
of 17 structures is shown in Figure 2. The average overall
backbone RMSD is 0.43 ( 0.05 Å; the all-atom RMSD is
1.90 ( 0.06 Å.
Figure 1. 1H,1H NOESY spectrum of c[D-PpYTFWF] in 10 mM phosphate
buffer, pH 3, 25 °C. (a) Entire spectrum. (b) Expansion showing NH-HR
cross-peaks. Sequential connectivities are indicated by lines.
in DMSO was reported by Kessler et al.21 The salient features
of the Kessler peptide are (1) a rigid peptide with a single
predominant backbone conformation and a trans conformation
of the Phe6-D-Pro1 peptide bond from the H,C COSY spectrum,
(2) a âI turn about Lys4-Trp5 and a âII′ turn about D-Pro1-
Phe2 determined from 3JNH-HR coupling constants and ROE data,
and (3) two intramolecular hydrogen bonds between Thr3-
CO‚‚‚NH-Phe6 and Phe6-CO‚‚‚NH-Thr3 inferred from temper-
ature dependence of amide proton chemical shifts.
A similar approach was used to determine the structure of
c[D-PpYTFWF] in aqueous solution. There is only one set of
proton resonances in the 1D-spectra from 5 to 55 °C and in the
TOCSY spectrum of c[D-PpYTFWF] in H2O. The NOESY
spectrum at 25 °C was used to make sequential assignments as
well as to measure interproton distances (Figure 1a). Sequential
assignments from NH-HR cross-peaks were used to identify
the NH, HR, and Hâ resonances of Phe4 and Phe6, because the
ring protons for the Phe residues were indistinguishable (Figure
1b). A trans peptide bond about Phe6-D-Pro1 is evident from
the two NOE cross-peaks between the HR proton of Phe6 and
Side Chain Conformation. The rotamer populations for
ø1 ) (60°, 180° conformations of amino acid side chains in
3
c[D-PpYTFWF] were calculated from JHR-Hâ coupling con-
stants using the Pachler equations.43 Coupling constants for Phe4
and Trp5 were determined from the 1D-spectrum; coupling
constants for pTyr2 were determined from the PE-COSY
spectrum. Table 1 compares the ø1 rotamer populations pI-III
for c[D-PpYTFWF] in D2O and c[D-PFTK(z)WF] in DMSO.
The populations for amino acids at the same position in the
sequence are similar in the two peptides. Trp5 adopts a major
ø1 side chain conformation with two equally populated minor
conformations.
(41) Kessler, H.; Anders, U.; Schudok, M. J. Am. Chem. Soc. 1990, 112, 5908-
5916.
(42) Stroup, A. N.; Rockwell, A. L.; Gierasch, L. M. Biopolymers 1992, 32,
1713-1725.
(43) Pachler, K. G. R. Spectrochim. Acta 1963, 19, 2085-2092.
9
J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002 9281

A R T I C L E S
Adams et al.
knr
Table 2. Fluorescence Dataa
λmax
solvent (nm)
k d
r
Φ280
Φ295
τ¯ (ns)
(107 s-1
)
(107 s-1
)
b
b
c
c[D-PpYTFWF]
H2O 348 0.17 ( 0.02 0.17 ( 0.02 2.9 ( 0.1 5.7 ( 0.7 28 ( 2
D2O 350 0.20 ( 0.02 0.21 ( 0.03 3.21 ( 0.06 6.5 ( 0.9 25 ( 1
c[D-PFTKWF]
H2O 350 0.16 ( 0.02 0.17 ( 0.02 3.06 ( 0.09 5.5 ( 0.7 27 ( 1
D2O 350 0.20 ( 0.01 0.21 ( 0.02 3.30 ( 0.06 6.4 ( 0.6 23.9 ( 0.8
c[D-PFTFWK]
H2O 349 0.17 ( 0.01 0.18 ( 0.01 2.9 ( 0.2 6.1 ( 0.5 28 ( 2
D2O 351 0.19 ( 0.01 0.22 ( 0.01 3.2 ( 0.1 6.9 ( 0.4 24 ( 1
NATAe
H2O 350 0.15 ( 0.02
D2O 350 0.18 ( 0.03
2.6 ( 0.4
3.3 ( 0.5
6 ( 1
5 ( 1
33 ( 6
25 ( 5
a 25 °C. b Mean values with standard deviations from 5 to 20 experiments.
c Values for cyclic peptides calculated from lifetime data in Table 3. d Values
for cyclic peptides calculated using quantum yields measured at 295 nm
excitation wavelength. e Data from ref 66.
Figure 2. Mean structure of c[D-PpYTFWF] from distance geometry and
energy minimization calculations.
Table 3. Fluorescence Lifetime Dataa
solvent
R1
τ1 (ns)
R2
τ2 (ns)
τ3 (ns)
ør2
The NOESY spectrum showed a single set of cross-peaks
between HR and Hâ protons of Trp5, consistent with fast
averaging of the side chain conformation on the NMR time
scale. A semiquantitative analysis of local NOE intensities
facilitated stereospecific assignments of the Trp5 Hâ protons
and identification of the major ø1 rotamer as -60°, under the
basic assumption that a major ø1 rotamer dominates the cross-
peak intensities.44 The NOE intensity is about 1.5-fold larger
for the HR-Hâs cross-peak than for the HR-Hâr cross-peak
of Trp5. There is also 1.4-fold stronger NOE intensity between
Trp5 amide NH-Hâr than NH-Hâs. These two NOE intensity
ratios indicate a predominant conformation with ø1 ) -60°.
The presence of multiple interconverting ø1 rotamers makes
it impossible to determine ø2 rotamer populations for each ø1
rotamer. NOE intensities alone can be interpreted at best
semiquantitatively. Nevertheless, NOE intensities between the
two Trp Hâ protons and the indole ring H2 and H4, which
depend solely on ø2 conformation, occur with comparable
intensities in all data, suggesting little ø2 preference for Trp5 in
this peptide. The presence of both H2 and H4 cross-peaks with
HR, which indicate primarily ø1, ø2 ) 60°, -90°; 180°, 90°
and ø1, ø2 ) -60°, 90°; 180°, -90°, respectively, excludes
correlation between ø1 and ø2. The data suggest the presence
of up to six Trp5 side chain rotamers in c[D-PpYTFWF].
Fluorescence Spectroscopy. The fluorescence emission
maxima λmax and quantum yields Φ of c[D-PpYTFWF], c[D-
PFTKWF], and c[D-PFTFWK] in H2O and D2O are given in
Table 2. The broad, featureless emission spectra of the peptides
are essentially identical to the spectrum of NATA in aqueous
solution. The quantum yields of the peptides and NATA are
the same within error. The values are slightly higher in D2O
than H2O with the same 1.2-fold solvent isotope effect.5 The
only functional groups in the peptides and NATA that quench
3-methylindole fluorescence in intermolecular quenching experi-
ments are the amide groups.7,8 Intramolecular quenching by the
peptide bonds in NATA is due to excited-state electron transfer.7
Fluorescence decays were measured for c[D-PpYTFWF], c[D-
PFTKWF], and c[D-PFTFWK] in H2O and D2O at 25 °C. Time-
resolved emission spectral data collected on two time scales
c[D-PpYTFWF]
H2O
D2O
0.68 ( 0.03
0.66 ( 0.05
3.82
4.25
0.17 ( 0.02
0.16 ( 0.02
1.83
2.02
0.32
0.45
1.5
1.5
c[D-PFTKWF]
H2O
D2O
0.70 ( 0.02
0.69 ( 0.01
3.91
4.25
0.16 ( 0.02
0.16 ( 0.02
1.84
1.90
0.28
0.45
1.6
1.4
c[D-PFTFWK]
H2O
D2O
0.71 ( 0.05
0.66 ( 0.03
3.72
4.24
0.16 ( 0.02
0.16 ( 0.02
1.71
1.92
0.31
0.47
1.5
1.6
a 25 °C, 295 nm excitation wavelength. Global analysis of data from
320 to 390 nm emission wavelength, 10-nm increments.
were deconvolved in global analyses with the lifetimes linked.
The decay curves for the peptides gave good fits to three
exponential functions with ør2 values of 1.4-1.6 and lifetimes
of about 3.8, 1.8, and 0.3 ns in H2O (Table 3). Triple exponential
fits were an improvement over double exponential fits, which
had ør2 values of 1.9 and nonrandom weighted residuals and
autocorrelation functions. MEM analysis, in which the decay
model is not predetermined, supports a triple exponential decay.
The relative amplitudes Ri appear independent of emission
wavelength: about 0.7 for the 3.8-ns component, 0.16 for the
1.8-ns component, and 0.14 for the 0.3-ns component. These
amplitudes correlate very well with the ø1 rotamer populations
3
determined from JHR-Hâ coupling constants (Table 1), which
associates the 3.8-ns lifetime component with the ø1 ) -60°
rotamer.
Radiative and nonradiative rates were calculated from quan-
tum yield and lifetime data using eqs 1, 5, and 6. Intramolecular
static quenching in dipeptides reduces the apparent value of kr
for tryptophan below the value for NATA.45 The radiative rates
kr for the cyclic peptides and NATA are about the same and
independent of solvent isotope within error (Table 2). Thus,
Trp5 does not undergo ground-state interactions in the cyclic
peptides that create a nonfluorescent population of tryptophan
residues. The nonradiative rates knr are slightly higher in H2O
than D2O due to water quenching, which is an isotopically
sensitive temperature-dependent quenching process.5 The finding
of the same quantum yields, lifetimes, and deuterium isotope
(45) Chen, R. F.; Knutson, J. R.; Ziffer, H.; Porter, D. Biochemistry 1991, 30,
5184-5195.
(44) Clore, G. M.; Gronenborn, A. M. Protein Eng. 1987, 1, 275-288.
9
9282 J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002

Tryptophan Fluorescence in Hexapeptides
A R T I C L E S
effects in the three peptides indicates that intramolecular excited-
state proton-transfer catalyzed by the ꢀ-ammonium of lysine
does not occur in c[D-PFTKWF] and c[D-PFTFWK].
The temperature dependence of the fluorescence lifetime of
NATA in water has two Arrhenius factors,
r.46 The electronic coupling of donor and acceptor in proteins
may be through-space, hydrogen-bonding, or covalent.47 Elec-
tron transfer occurs from the highest occupied molecular orbital
(HOMO) of the donor to the lowest unoccupied molecular
orbital (LUMO) of the acceptor. For tryptophan-containing
peptides, excited indole is the electron donor. In aqueous
solution, emission and therefore also electron transfer occur from
τ-1 ) ko + A1 exp(-E1*/RT) + A2 exp(-E2*/RT) (7)
1
the La state.48 The carbonyl carbon of the peptide bond is
probably the electron acceptor. The covalent bond distances are
the same for all rotamers. A through-bond electron transfer
mechanism could not produce lifetime heterogeneity unless the
through-bond electron-transfer rate were highly dependent on
the orientation of the indole ring and the peptide bond.
Moreover, the through-space distances between indole and
various carbonyl carbons are a few Å shorter than the through-
bond distances. Hence, we only consider through-space electron-
transfer mechanisms. An orientation dependence of ket has been
shown for heme-heme intramolecular electron transfer in
cytochromes, even when donor-acceptor distances were com-
parable in different proteins.49 The effect of orbital overlap has
been treated by several workers.50-52 The electron-transfer rate
may be expressed as a function of distance and orientation by50,51
where ko ) kr + kisc is the temperature-independent rate, Ai is
the frequency factor, Ei* is the activation energy, R is the gas
constant, and T is temperature in K.7 The two Arrhenius terms
represent solvent quenching ksi ) A1 exp(-E1*/RT) and excited-
state electron transfer from the indole ring to the amides ket )
A2 exp(-E2*/RT). Fluorescence decays of c[D-PpYTFWF] were
measured in H2O and D2O from 5 to 55 °C and deconvolved in
single curve analyses. The long lifetime τ1 decreased monotoni-
cally over the temperature range studied, from 5.14 ns at 7.2
°C to 1.97 ns at 55 °C in H2O. The temperature dependence of
the intermediate and short lifetimes was obscured by correlation
among the decay parameters of these minor components in the
-1
three exponential fits. Plots of ln[τ1 - ko - ksi] versus 1/T
for the long lifetime data were made, using kr values from Table
2 and kisc and ksi values for NATA.7 Arrhenius parameters
obtained from the intercept and slope of the lines were: A2 )
1.9 × 1011 s-1 and E2* ) 4.4 kcal/mol in H2O and A2 ) 5.3 ×
1011 s-1 and E2* ) 4.9 kcal/mol in D2O. Electron-transfer rates
calculated at 25 °C are 1.1 × 108 s-1 in H2O and 1.4 × 108 s-1
in D2O. The ket values for the long lifetime component of c[D-
PpYTFWF] are slightly lower than the values for NATA (2.0
× 108 s-1 in H2O and 2.2 × 108 s-1 in D2O).7
ket(i,j) ) B f(θi,j) exp(-âri,j)
(9)
where B is a preexponential factor independent of both r and
θ, f(θi,j) is an orientation function, and â ) 1.7 Å-1 is the
distance decay factor commonly used for through-space electron
transfer in proteins.53 The donor HOMO of the 1La state of indole
is fairly well described by the LUMO of the ground state.54,55
The acceptor LUMO is the π* antibonding orbital of the
carbonyl CdO bond. We calculate the overlap of these orbitals
very crudely with the simple orientation function f(θi,j) )
Excited-State Electron Transfer. Two models for producing
different local environments of Trp5 can explain the multi-
exponential fluorescence decay of c[D-PpYTFWF]. (1) Different
lifetimes result from multiple ground-state rotamers of the
tryptophan side chain as proposed by the rotamer model.14,15
(2) Different lifetimes result from multiple conformations of
the peptide backbone. The latter is unlikely because of the strong
NMR evidence for a rigid backbone. Moreover, the ø1 side chain
rotamer populations of Trp5 determined from NMR correlate
very well with the relative amplitudes of the fluorescence decay.
By assuming that the radiative rates of the rotamers are the same,
the different lifetimes would be due to differences in the total
nonradiative rate. On the basis of intermolecular quenching
experiments,8 we expect only three dynamic quenching mech-
anisms in c[D-PpYTFWF]: intersystem crossing, water quench-
ing, and peptide bond quenching by excited-state electron
transfer. We assume that the intersystem crossing and water
quenching rates are the same for all Trp5 rotamers and equal or
similar to the rates in NATA.7 In this scenario, the three Trp5
lifetimes would result from differences in peptide bond quench-
ing among rotamers. The total electron-transfer rate ket(j) of
rotamer j is the sum of the individual electron-transfer rates
from Trp5 to each of the six peptide bonds in c[D-PpYTFWF]
cos2 θi,j for overlap of S-P orbitals.51
θi,j is the dihedral angle
between the planes containing the indole ring and peptide bond
i in rotamer j. In the spherical elephant approximation, the S
orbital represents the LUMO of the indole ring, and the P orbital
represents the π* antibonding orbital of the carbonyl.
To calculate electron-transfer rates of the six possible Trp5
rotamers in c[D-PpYTFWF] from eqs 8 and 9, we need a value
for B. NATA and c[D-PpYTFWF] have the same donor-
acceptor pair; therefore, we assume that the electron-transfer
parameters B and â in eq 9 are about the same for both
molecules. NATA mimics a tryptophan-containing peptide with
peptide bonds i-1 and i on either side of the indole ring. The
fluorescence decay is a single exponential with 2.9-ns lifetime
at 25 °C. Presumably, the single-exponential represents a
multitude of ground-state species arising from the six side chain
rotamers with free rotation about each peptide bond. The ket
value for NATA in H2O is almost certainly a composite value
(46) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265-322.
(47) Beratan, D. N.; Onuchic, J. N.; Hopfield, J. J. J. Chem. Phys. 1987, 86,
4488-4498.
(48) Callis, P. R. Methods Enzymol. 1997, 278, 113-151.
(49) Makinen, M. W.; Schichman, S. A.; Hill, S. C.; Gray, H. B. Science 1983,
222, 929-931.
k (j) )
k (i,j)
(8)
∑
et
et
(50) Doktorov, A. B.; Khairutdinov, R. F.; Zamaraev, K. I. Chem. Phys. 1981,
61, 351-364.
i
(51) Domingue, R. R.; Fayer, M. D. J. Chem. Phys. 1985, 83, 2242-2252.
(52) Cave, R. J.; Siders, P.; Marcus, R. A. J. Phys. Chem. 1986, 90, 1436-
1444.
where ket(i,j) is the quenching rate of rotamer j by peptide
bond i.
In a system with weak coupling between donor and acceptor,
the electron-transfer rate decreases exponentially with distance
(53) Beratan, D. N.; Betts, J. N.; Onuchic, J. N. Science 1991, 252, 1285-
1288.
(54) Callis, P. R. J. Chem. Phys. 1991, 95, 4230-4240.
(55) Slater, L. S.; Callis, P. R. J. Phys. Chem. 1995, 99, 8572-8581.
9
J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002 9283

A R T I C L E S
Adams et al.
Table 4. Calculated Fluorescence Lifetimes for Rotamer j of Trp5 in C[D-PpYTFWF]a
lifetime of rotamer j (ns), ø1, ø2
â (Å-1
)
B (1012 s-1
)
−60°, 90°
−60°, −90°
−180°, 90°
−180°, −90°
60°, 90°
60°, −90°
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
2.3
2.4
2.5
2.6
0.38 (0.08)
0.55 (0.13)
0.80 (0.21)
1.1 (0.33)
2.0 (0.60)
3.6 (0.90)
5.8 (1.5)
9.2 (2.6)
15 (4.0)
25 (6.7)
44 (11)
75 (18)
120 (29)
2.8 (3.9)
3.1 (4.0)
3.4 (4.0)
4.0 (4.1)
3.7 (3.9)
3.6 (4.1)
3.7 (4.1)
3.8 (4.1)
3.9 (4.1)
4.0 (4.2)
3.9 (4.2)
3.9 (4.2)
4.0 (4.1)
1.9 (2.7)
2.1 (2.6)
2.3 (2.5)
2.6(2.5)
2.4 (2.3)
2.3 (2.3)
2.3 (2.2)
2.3 (2.1)
2.2 (2.1)
2.2 (2.0)
2.0 (2.0)
2.0 (2.0)
1.9 (1.9)
1.7 (2.7)
1.8 (2.6)
1.9 (2.6)
2.0 (2.5)
1.8 (2.2)
1.6 (2.3)
1.5 (2.0)
1.5 (2.0)
1.4 (2.0)
1.3 (1.8)
1.2 (1.8)
1.1 (1.8)
1.0 (1.7)
1.4 (2.9)
1.5 (2.8)
1.6 (2.7)
1.7 (2.7)
1.6 (2.4)
1.4 (2.5)
1.3 (2.3)
1.3 (2.2)
1.2 (2.2)
1.2 (2.1)
1.1 (2.0)
1.0 (1.9)
0.93 (1.8)
1.2 (2.8)
1.3 (2.9)
1.5 (2.9)
1.7 (2.9)
1.6 (2.8)
1.5 (2.9)
1.5 (2.9)
1.5 (2.8)
1.5 (2.9)
1.5 (2.9)
1.4 (2.9)
1.4 (2.9)
1.4 (2.9)
0.80 (2.1)
0.90 (2.1)
0.95 (2.0)
1.0 (2.0)
0.92 (1.8)
0.80 (1.9)
0.78 (1.7)
0.76 (1.7)
0.72 (1.7)
0.67 (1.6)
0.60 (1.6)
0.54 (1.5)
0.52 (1.5)
a Numbers in parentheses calculated assuming no orientation dependence.
from these multiple species, so that B is likewise a composite
value. We estimate B using the electron-transfer rate of NATA
from
1.8-ns component, and the 1.0-ns lifetime calculated for the 60°,
-90° rotamer to the observed 0.3-ns component. Case I agrees
with the ø2 rotamer populations of tryptophan zwitterion,56
although not with the apparently uncorrelated ø1, ø2 rotamer
populations of the cyclic peptide. To test whether Case I is
consistent with the time-resolved fluorescence data, we simu-
lated triple exponential fluorescence decays using calculated
lifetimes and amplitudes equal to the respective ø1 rotamer
populations of Trp5 in Table 1. Eight simulations using the 4.0,
1.7, and 1 ns values recovered lifetimes (amplitudes) of 3.6-
3.9 (0.71 ( 0.07), 1.5-1.9 (0.16 ( 0.02), and 0.14-0.55 ns
(0.13 ( 0.06) with τj ) 3.0 ( 0.2 ns. Nine of 10 simulations
using the 4.0, 2.0, and 1 ns values recovered lifetimes
(amplitudes) of 3.4-5.7 (0.69 ( 0.06), 1.5-2.1 (0.19 ( 0.05),
and 0.15-0.35 ns (0.11 ( 0.03) with τj ) 3.0 ( 0.5 ns. Only
one simulation recovered the 1-ns value for the short lifetime
with τj ) 3.06 ns.
(II) Assume that the ø1 rotamers do not interconvert on the
fluorescence time scale, that the observed fluorescence decay
amplitudes equal the NMR-determined ø1 rotamer populations,
but that the observed lifetimes represent some form of average
lifetime of the two ø2 rotamers. For example, if the ø2 rotamers
interconvert rapidly, then the lifetimes would converge to a
single value14
k (NATA) ) (B/6)
f(θi-1,j) exp(-âri-1,j) +
∑
et
j
f(θi,j) exp(-âri,j) (10)
The summation over j ) 1-6 assumes equal populations of
the six canonical rotamers of the indole side chain. To simulate
free rotation about the φ, ψ bonds in each rotamer j, 12 values
of f(θi-1,j) exp(-âri-1,j) and f(θi,j) exp(-âri,j) for peptide bonds
i-1 and i were computed and averaged. Using ket(NATA) )
2.0 × 108 s-1 at 25 °C,7 eq 10 gives B ) 1.1 × 1012 s-1
.
Assuming no orientation dependence with f(θi,j) ) 1, eq 10 gives
B ) 3.3 × 1011 s-1
.
Individual electron-transfer rates ket(i,j) from rotamer j of Trp5
to peptide bond i in c[D-PpYTFWF] were calculated from eq
9, and the total electron-transfer rate ket(j) for each rotamer was
then calculated from eq 8 (Table SI). In general, for distances
>6 Å or orientations of the indole ring and the peptide bond
>70°, the ket(i,j) values are negligible (<107 s-1). In the six
rotamers, the peptide bonds on either side of the indole ring
contribute >84% of the total ket(j). The ket(j) values vary by a
factor of 7 among rotamers with peptide bond quenching being
fastest in the 60°, -90° rotamer and slowest in the -60°, 90°
rotamer. Neglecting orientation, the ket(j) values vary by less
than a factor of 4.
1/τ ) (k90 + Kk-90)/(1 + K)
(11)
where k90 and k-90 are the decay rates and the equilibrium
constant K ) p-90/p90 is the population ratio of ø2 rotamers. A
value of K ) 0.11 was calculated from eq 11 using the observed
3.8-ns lifetime of the ø1 ) -60° rotamer and the calculated
decay rates of the two ø2 rotamers, which corresponds to 90%
population for ø2 ) 90° as assumed in Case I.
Table 4 also shows the effect of â on the calculated lifetimes
of the six rotamers. â values of 1.4, 1.7, and 2.6 Å-1 have been
used for through-space electron transfer in proteins.57-59 The
-60°, 90° rotamer has the longest lifetime, and the 60°, -90°
rotamer has the shortest lifetime throughout this range. As
expected, the lifetime difference between these rotamers in-
Table 4 gives the fluorescence lifetimes for the six rotamers.
Lifetimes were calculated from eqs 1 and 2 with kpt ) 0, using
kr for the peptide from Table 2 and kisc + ksi ) 8.4 × 107 s-1
for NATA at 25 °C.7 For â ) 1.7 Å-1, the lifetimes range from
1 to 4 ns. For each ø1 rotamer, the calculated lifetimes of the
ø2 ) (90° rotamers differ by less than a factor of 2. Such
closely spaced lifetimes are difficult to resolve in a multiex-
ponential decay. With this in mind, we rationalize the observed
and calculated lifetimes of c[D-PpYTFWF] according to the
rotamer model as follows.
(I) Assume that the rotamers do not interconvert on the
fluorescence time scale, that the observed fluorescence decay
amplitudes Ri equal the NMR-determined ø1 rotamer popula-
tions, and that each ø1 rotamer has a major ø2 rotamer. This
assigns the 4.0-ns lifetime calculated for the -60°, 90° rotamer
to the observed 3.8-ns component, either the 1.7- or 2.0-ns
lifetime calculated for the ø1 ) 180° rotamers to the observed
(56) Dezube, B.; Dobson, C. M.; Teague, C. E. J. Chem. Soc., Perkin Trans. 2
1981, 730-735.
(57) Hopfield, J. J. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 3640-3644.
(58) Jortner, J.; Bixon, M. Electron Transfer-From Isolated Molecules to
Biomolecules, Part 1; Wiley: New York, 1999.
(59) Moser, C. C.; Keske, J. M.; Warncke, K.; Farid, R. S.; Dutton, P. L. Nature
1992, 355, 796-802.
9
9284 J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002

Tryptophan Fluorescence in Hexapeptides
A R T I C L E S
creases with â for both orientation-dependent and orientation-
independent mechanisms. Orientation has the most effect on
the lifetime of the 60°, -90° rotamer and the least effect on
the two ø1 ) -60° rotamers. Cases I and II predict the 3.8-
and 1.8-ns lifetimes quantitatively, but not the 0.3-ns lifetime.
The predictions are robust to â values between 1.7 and 2.1 Å.
solvent relaxation on the fluorescence time scale is amply
documented for indole derivatives in viscous solvents.9,60-62 We
may exclude solvent relaxation as a source of lifetime hetero-
geneity in the cyclic peptides on the following two grounds:
(1) The fluorescence decays of the cyclic peptides appear
independent of emission wavelength. At the 10-nm resolution
of our time-resolved emission spectral data, λmax ) 350 nm for
both 3.8- and 1.8-ns components and 340 nm for the 0.3-ns
component. The short lifetime component constitutes <2% of
the total intensity, so its spectrum may include a contribution
from the water Raman at 331 nm. (2) Solvent relaxation of
tryptophan zwitterion has a 1.2-ps relaxation time in aqueous
solution at room temperature.63 It is unlikely that the relaxation
time would be much slower for a fully solvent exposed
tryptophan in a small peptide. On the other hand, rotamer
interconversion during the lifetime of the excited state is a
possibility. The good agreement between NMR-determined
ground-state rotamer populations and fluorescence decay am-
plitudes suggests that ø1 rotamer interconversion is slower than
the fluorescence time scale, albeit faster than the NMR time
scale. In molecular dynamics simulations using an all atom
representation, ø2 rotamer interconversion is faster than ø1
rotamer interconversion.64 The short lifetime component may
be indicative of excited-state conformer interconversion.65 If so,
the lifetime heterogeneity would still be due to at least three
ground-state rotamers with different lifetimes, but the amplitudes
and decay times would no longer simply represent the rotamer
populations and lifetimes.
Although the peptide bond is a weak intermolecular quencher
of 3-methylindole fluorescence, it is an efficient intramolecular
quencher in NATA7 and the cyclic peptides. This strongly
suggests that peptide bond quenching is an important quenching
mechanism in proteins. The quenching rate will depend on the
number, distance, and possibly orientation of peptide bonds
within about 6 Å of the indole ring. Sillen et al. proposed that
electron-transfer quenching by the peptide bond may be the main
determinant of the lifetime of tryptophan in proteins in the
absence of strong quenchers.13 They used molecular dynamics
to determine tryptophan rotamer populations and Marcus theory
to predict the three fluorescence lifetimes of two single
tryptophan proteins. The details of their lifetime calculations
differ from those presented here. In c[D-PpYTFWF], the peptide
bond is assumed to be the only intramolecular quenching group.
Although phenylalanine is not an intermolecular quencher of
3-methylindole fluorescence,8 the phenyl ring has been proposed
to be a highly efficient intramolecular quencher of buried
tryptophans in proteins.12 The presumed quenching mechanism
involves an atypical NH‚‚‚π hydrogen bond, which requires a
particular orientation of the indole nitrogen relative to the phenyl
ring. There are three other aromatic residues in c[D-PpYTFWF]
besides tryptophan: pTyr2, Phe4, and Phe6. However, these
Conclusions
This paper provides definitive evidence for the rotamer model
of tryptophan photophysics for peptides in aqueous solution.
c[D-PpYTFWF] has a rigid backbone, making side chain
rotamers the only source of conformational heterogeneity. The
three fluorescence decay amplitudes correlate well with the three
1
ø1 rotamer populations of Trp5 determined by H NMR. The
three fluorescence lifetimes can be explained by differences
among ø1, ø2 rotamers in the rate of intramolecular excited-
state electron transfer from the indole ring to two to four peptide
bonds in c[D-PpYTFWF]. Except for the shortest lifetime,
remarkably good quantitative agreement is obtained with
calculated lifetimes using a simple model and reasonable
assumptions. Two relatively innocuous approximations were
made to calculate fluorescence lifetimes, namely (1) The
radiative, intersystem-crossing, and water-quenching rates are
the same in all rotamers. (2) The intersystem-crossing and water-
quenching rates are the same in NATA and c[D-PpYTFWF].
Intersystem crossing plus water quenching kisc + ksi contribute
only about 30% of the average total nonradiative rate (Table 2)
or 10-40% in individual rotamers (Table SI). Likewise, kr +
kisc + ksi contributes at most about 50% of the total decay rate.
Small differences in these contributions would be overshadowed
by the larger electron-transfer rates.
The electron-transfer calculations attempt to establish a
quantitative relationship between fluorescence lifetime and
molecular structure. Given the many assumptions, is the
agreement between calculated and experimental lifetimes en-
tirely fortuitous? We assumed a through-space electron transfer
mechanism with orientation dependence. Electronic propagation
in peptides probably involves a combination of through-space
and through-bond mechanisms, like the pathway model used
for protein electron transfer. Moreover, we made a number of
simplifying assumptions to estimate electron-transfer rates, the
most significant being: (1) Orientation dependence of electron
transfer from the excited indole ring to the carbonyl carbon was
approximated by overlap of S-P orbitals. The assumption of a
1
spherical molecular orbital for La neglects the nodal structure
of the ground-state LUMO.55 (2) Distances and relative orienta-
tions were measured for canonical conformations of ø1, ø2
rotamers. (3) The Marcus parameter B was estimated from the
electron-transfer rate of NATA and â was borrowed from
ground-state electron-transfer theory. Finally, we assumed that
the middle of the indole ring is the electron donor. The donor
atom closest to the acceptor, indole C3 in our case, is normally
used in protein electron-transfer calculations. We also did
calculations with indole C3 as electron donor. Not surprisingly,
there was less variation in electron-transfer rates among rota-
mers, factors of four with orientation and only two without
orientation. Much more work with well-defined model systems
is needed to place the excited-state electron-transfer calculations
on firm ground.
(60) Milton, J. G.; Purkey, R. M.; Galley, W. C. J. Chem. Phys. 1978, 68, 5396-
5403.
(61) Lakowicz, J. R.; Cherek, H.; Bevan, D. R. J. Biol. Chem. 1980, 255, 4403-
4406.
(62) Meech, S. R.; Phillips, D.; Lee, A. G. Chem. Phys. Lett. 1982, 92, 523-
527.
(63) Shen, X.; Knutson, J. R. J. Phys. Chem. B 2001, 105, 6260-6265.
(64) Gordon, H. L.; Jarrell, H. C.; Szabo, A. G.; Willis, K. J.; Somorjai, R. L.
J. Phys. Chem. 1992, 96, 1915-1921.
(65) McMahon, L. P.; Yu, H.-T.; Vela, M. A.; Morales, G. A.; Shui, L.;
Fronczek, F. R.; McLaughlin, M. L.; Barkley, M. D. J. Phys. Chem. B
1997, 101, 3269-3280.
Excited-state relaxation processes have been invoked to
explain the complex fluorescence decays of proteins.9-11 Dipolar
(66) Kirby, E. P.; Steiner, R. F. J. Phys. Chem. 1970, 74, 4480-4490.
9
J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002 9285

A R T I C L E S
Adams et al.
residues are probably not intramolecular quenching groups in
the cyclic peptides. First, substitution of either phenylalanine
by lysine in c[D-PFTKWF] and c[D-PFTFWK] has no effect
on fluorescence quantum yield and lifetimes. Second, molecular
modeling shows that the particular conformation for NH‚‚‚π
hydrogen bonding is equally possible in all tryptophan ø1
rotamers. Last, NH‚‚‚π hydrogen bonding is not likely to occur
in aqueous solution.
Acknowledgment. This work was supported by NIH Grant
GM42101.
Supporting Information Available: Table of electron transfer
rates for rotamer j of Trp5 of c[D-PpYTFWF] (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA0167710
9
9286 J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002