Thioamides as fluorescence quenching probes: Minimalist chromophores to monitor protein dynamics

Article abstract of DOI:10.1021/ja1044924

Decreasing the size of spectroscopic probes can afford higher-resolution structural information from fluorescence experiments. Therefore, we have developed p-cyanophenylalanine (Cnf) and backbone thioamides as a fluorophore/quencher pair. Through the exam

Full text of DOI:10.1021/ja1044924

Published on Web 10/01/2010
Thioamides as Fluorescence Quenching Probes: Minimalist Chromophores
To Monitor Protein Dynamics
Jacob M. Goldberg, Solongo Batjargal, and E. James Petersson*
Department of Chemistry, UniVersity of PennsylVania, 231 South 34th Street, Philadelphia,
PennsylVania 19104-6323, United States
Received May 24, 2010; E-mail: ejpetersson@sas.upenn.edu
Abstract: Decreasing the size of spectroscopic probes can afford
higher-resolution structural information from fluorescence experi-
ments. Therefore, we have developed p-cyanophenylalanine (Cnf)
and backbone thioamides as a fluorophore/quencher pair. Through
the examination of a series of thiopeptides, we have determined
the working distance for this pair to be 8-30 Å. We have also
carried out a proof-of-principle protein-folding experiment in which
a Cnf/thioamide-labeled version of villin headpiece HP35 was
thermally unfolded while the Cnf/thioamide distance was moni-
tored by fluorescence. For a given protein, thioamide substitutions
could be used to track motions with a much greater number of
measurements than for current fluorescence probes, providing a
dense array of data with which to model conformational changes.
Figure 1. Thioleucylalanine ester (Leu′Ala) and p-cyanophenylalanine
(Cnf) spectra. Absorption spectra of Cnf (dotted black line) and Leu′Ala
(solid purple line) with relative absorption intensities normalized to the
extinction coefficients (13 000 M-1 cm-1 at 232 nm and 12 400 M-1 cm-1
One of the great challenges facing biochemists is understanding
the rapid and complex structural dynamics of proteins. Fluorescence
(dashed black line) arbitrarily normalized to the absorption maximum of
at 266 nm, respectively) are shown.4,9 The fluorescence spectrum of Cnf
measurements can be made on the nanosecond time scale, and
distance-dependent interactions such as Fo¨rster resonant energy
transfer (FRET) can be used to determine the separation of
chromophore labels to glean time-resolved structural information
on protein motions.1 However, the relatively large size of common
fluorophores precludes the assignment of these motions with atomic
resolution. If optical probes could be made sufficiently small, they
could provide the time and structural resolution necessary to truly
dissect protein motions. Here we demonstrate that p-cyanophe-
nylalanine (Cnf), which is structurally similar to Phe, and a
thioamide obtained by a single-atom substitution in the peptide
backbone can be used to monitor structural changes in proteins.
The most widely used method for fluorescent tagging of proteins
employs genetic fusion of a variant of the green fluorescent protein
from Aequorea Victoria (GFP) to the protein of interest. GFP fusions
are easy to prepare by standard cloning methods, but the large size
of GFP (268 amino acids) limits its utility in observing the motions
of proteins, which are often of comparable size.2 Labeling proteins
with a small organic molecule such as fluorescein either during or
after translation introduces a less structurally perturbing probe.3
However, many positions in the protein cannot accommodate
polycyclic organic chromophores. The bulk of GFP and fluorescein
derivatives not only limits the number of places that the label can
be placed but also raises the concern that the label itself may alter
the observed motion. On the other hand, sulfur replacement of
oxygen in the peptide backbone provides a fluorescence quencher
that is extremely small and compatible with virtually any position
in a protein sequence.
Cnf is also shown. The shaded area indicates the spectral overlap that
contributes to FRET.
note one preliminary exploration of thioamide as a resonance energy
acceptor, but the pairing with Trp was reported to be viable only
under nonphysiological solvent conditions (neat propylene glycol).8
We chose to use Cnf as a donor chromophore because its
fluorescence emission was expected to overlap well with thioamide
absorption and its extinction coefficient is much greater than that
of Tyr in the 230-250 nm range.9 Cnf has previously been used
as a donor in FRET experiments with both Trp and Tyr acceptors,10,11
which also have spectral overlap with Cnf emission. Since
deconvolution of Cnf and thioamide interactions with these chro-
mophores could be very complex, we limited our initial experiments
to cases without Trp or Tyr residues in order to more easily
characterize the interactions with the thioamide.
Our study of Cnf/thioamide spectroscopy began with the calcula-
tion of the Fo¨rster distance R0, the distance at which resonant energy
transfer between two chromophores is half-maximal. This calcula-
tion requires determination of the spectral overlap integral from
the absorption spectrum of the thioamide acceptor and the
fluorescence emission spectrum of the Cnf donor12 (Figure 1). Using
the reported quantum yield for Cnf fluorescence (0.11)9 and the
common assumption that the orientation of the chromophore
transition dipoles is random on the time scale of energy transfer
2
(i.e., κ2 ) /3), we calculated R0 to be 15.6 Å (see the Supporting
Information). This value implied that the pair could be useful in
measuring distance changes on the scale of a protein.
Peptides with one or more thioamides in the backbone have been
used previously in a small number of applications. These have
primarily focused on photoisomerization of the thioamide4,5 and
its impact on folding in simple secondary-structure motifs.6,7 We
To further characterize the Cnf/thioamide pair, we prepared
peptides with an N-terminal thioleucine (Leu′) and a C-terminal
Cnf separated by 2-10 proline residues, analogous to the classic
experiments of Stryer and Haugland.13 We obtained fluorescence
9
14718 J. AM. CHEM. SOC. 2010, 132, 14718–14720
10.1021/ja1044924  2010 American Chemical Society

C O M M U N I C A T I O N S
spectra of 12 µM solutions of these peptides in phosphate buffer
and observed a strong dependence of the fluorescence intensity on
the number of intervening Pro residues. To correlate the fluores-
cence intensity with interchromophore distance, we carried out a
series of 10 ns molecular dynamics (MD) simulations of Leu-Pron-
Phe (n ) 2-10) in explicit water boxes. The distances between
the thioamide (using the Leu amide as a proxy) and Cnf (using
Phe as a proxy) were time-averaged over the unrestrained portions
of the simulations (see the Supporting Information). Normalizing
the fluorescence emission (F) across the proline series to the
emission of Cnf in an oxoamide control peptide (F0), we computed
the efficiency of fluorescence quenching as EQ ) 1 - F/F0. This
allowed us to plot EQ as a function of distance for the proline series
and compute an R0 value of 16.5 Å by fitting the data to a 1/R6
function (Figure 2, solid line), in reasonable agreement with the
value determined from spectroscopic data alone.
We also carried out one proof-of-principle experiment to
demonstrate the utility of the thioamide/Cnf pair in monitoring a
conformational change. We incorporated Leu′ at the N-terminus
and Cnf at the C-terminus of the villin headpiece HP35 variant,
which was originally described by Kim and co-workers.16 The
fraction folded (ff) as a function of temperature was determined
for the thioamide version HP35-Leu′1Cnf35 and the corresponding
oxoamide control protein HP35-Cnf35 using CD spectroscopy
(Figure 3). The two proteins had comparable Tm values (i.e.,
temperatures at which ff ) 0.5), indicating that incorporating the
thioamide at this position had no dramatic effect on protein
unfolding. The quenching efficiency was determined by comparing
the fluorescence of the thioamide to that of the oxoamide. Since
Cnf fluorescence has been shown to vary strongly with temperature
and local environment,10,17 determining EQ by comparison to the
oxoamide HP35-Cnf35 was essential to control for effects on Cnf
emission that are independent of the thioamide interaction. EQ
ranged from 0.19 at 5 °C to <0.02 at 75 °C.18 The change in EQ
implies a distance change from 20-21 Å in the folded state (19.2
Å observed in 10 ns MD simulations) to greater than the detectable
range of our probe pair (32 Å) in the unfolded state. As Figure 3
shows, converting EQ to distance using the Fo¨rster and Dexter
interpretations gave comparable results (root-mean-square deviations
of 0.8 Å over the 70 °C temperature range). The data show that
Leu′1/Cnf35 dissociation is coupled to global unfolding even though
it is found on the periphery of HP35.
Figure 2. Fluorescence emission as a function of chromophore spacing.
The fluorescence emission of Leu′-Pron-Cnf (n ) 2-10) at 293 nm is shown
(three trials per peptide, bars represent standard errors). The “∞” data point
indicates the fluorescence of Leu-Pro2-Cnf. The solid line indicates the
distance dependence predicted by Fo¨rster theory with R0 ) 15.6 Å (∼5.1
prolines). The inset shows EQ as a function of the computed Leu′-Cnf
interchromophore distance for the proline series. The data were fit to Fo¨rster
(solid line, r2 ) 0.973) and Dexter (dashed line, r2 ) 0.986) distance
dependences.
Several observations regarding this result are worth noting. First,
although the distance dependence is comparable to the prediction
from Fo¨rster theory, for such short distances one must also consider
Dexter energy transfer via direct orbital overlap.12 Dexter transfer
has a 1/eR distance dependence, which provided a slightly better
fit to the Pro series data (Figure 2, dashed line). In fact, both
mechanisms probably contribute to thioamide quenching and may
be difficult to deconvolute. Second, as thioamides have previously
been used in photoisomerization experiments,4,5 we investigated
the possibility that energy transfer might cause cis/trans isomer-
ization. After 1 h of irradiation, our most efficient energy transfer
peptide, Leu′-Pro2-Cnf, showed no significant change in HPLC,
UV-vis, or circular dichroism (CD) assays (see the Supporting
Information). The lack of isomerization seems to be inconsistent
with either energy transfer mechanism and will be investigated
further. Regardless of the mechanism, our data confirm that the
working distance for the Cnf/thioamide pair is 8-30 Å, comple-
menting longer-range pairs like Trp/dansyl (R0 ≈ 22 Å) and
fluorescein/tetramethylrhodamine (R0 ≈ 50 Å).14 Finally, although
a proximal N-terminal amine can quench Cnf fluorescence,15
acetylation of Leu-Pro2-Cnf and Leu′-Pro2-Cnf had no effect on
the fluorescence at pH 7.0 (see the Supporting Information). This
gives us confidence that the quenching (relative to the parent
oxoamide) is due to the O-to-S substitution alone.
Figure 3. Villin HP35 unfolding monitored by Cnf/thioamide FRET. (left)
Villin HP35 structure taken from PDB entry 1VII19 modified with the Cnf
nitrile on Phe35 [image rendered in PyMOL (Delano Scientific, LLC, South
San Francisco, CA)]. (right) Fraction folded as determined from temperature-
dependent CD spectroscopy for HP35-Leu′1Cnf35 (b) and HP35-Cnf35 (O)
and temperature dependence of the Leu′1/Cnf35 separation determined by
comparison of EQ computed from the HP35-Leu′1Cnf35/HP35-Cnf35 fluo-
rescence ratio with the proline series distance dependence computed using
either the Fo¨rster ([) or Dexter (]) equation.
In summary, we have identified a novel fluorophore/quencher
pair and demonstrated its distance dependence and application to
monitoring of the unfolding of a small protein. Since one can
conceivably replace any amino acid in a synthetically accessible
protein by its thioamide analogue, this quencher could be applied
anywhere in a protein sequence. We are currently exploring the
use of the Cnf/thioamide system in proteins containing Trp or Tyr
to determine its generality and further studying the mechanism of
fluorescence quenching by thioamides.
Acknowledgment. This work was supported by funding from
the University of Pennsylvania and the National Science Foundation
(NSF CHE-1020205 to E.J.P.). We thank Jeff Saven for use of the
fluorometer, Feng Gai for assistance with the CD spectrometer
9
J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010 14719

C O M M U N I C A T I O N S
(8) Wiczk, W. M.; Gryczynski, I.; Szmacinski, H.; Johnson, M. L.; Kruszynski,
M.; Zboinska, J. Biophys. Chem. 1988, 32, 43–49.
(supported by NSF DMR05-20020), Chris MacDermaid for MD
simulation guidance, and Alyssa Klein for peptide synthesis.
(9) Tucker, M. J.; Oyola, R.; Gai, F. Biopolymers 2006, 83, 571–576.
(10) Taskent-Sezgin, H.; Chung, J.; Patsalo, V.; Miyake-Stoner, S. J.; Miller,
A. M.; Brewer, S. H.; Mehl, R. A.; Green, D. F.; Raleigh, D. P.; Carrico,
I. Biochemistry 2009, 48, 9040–9046.
Supporting Information Available: Descriptions of peptide syn-
thesis, purification, and characterization; fluorescence, UV-vis spec-
troscopy, and CD experiments; and data fitting and calculations. This
material is available free of charge via the Internet at http://pubs.acs.org.
(11) Tucker, M. J.; Oyola, R.; Gai, F. J. Phys. Chem. B 2005, 109, 4788–4795.
(12) Speiser, S. Chem. ReV. 1996, 96, 1953–1976.
(13) Stryer, L.; Haugland, R. P. Proc. Natl. Acad. Sci. U.S.A. 1967, 58, 719–
726.
(14) Wu, P. G.; Brand, L. Anal. Biochem. 1994, 218, 1–13.
(15) Taskent-Sezgin, H.; Marek, P.; Thomas, R.; Goldberg, D.; Chung, J.;
Carrico, I.; Raleigh, D. P. Biochemistry 2010, 49, 6290–6295.
(16) McKnight, C. J.; Doering, D. S.; Matsudaira, P. T.; Kim, P. S. J. Mol.
Biol. 1996, 260, 126–134.
References
(1) Royer, C. A. Chem. ReV. 2006, 106, 1769–1784.
(2) Piston, D. W.; Kremers, G. J. Trends Biochem. Sci. 2007, 32, 407–414.
(3) Katritzky, A. R.; Narindoshvili, T. Org. Biomol. Chem. 2009, 7, 627–634.
(4) Helbing, J.; Bregy, H.; Bredenbeck, J.; Pfister, R.; Hamm, P.; Huber, R.;
Wachtveitl, J.; De Vico, L.; Olivucci, M. J. Am. Chem. Soc. 2004, 126,
8823–8834.
(5) Wildemann, D.; Schiene-Fischer, C.; Aumuller, T.; Bachmann, A.; Kief-
haber, T.; Lucke, C.; Fischer, G. J. Am. Chem. Soc. 2007, 129, 4910–
4918.
(6) Miwa, J. H.; Pallivathucal, L.; Gowda, S.; Lee, K. E. Org. Lett. 2002, 4,
4655–4657.
(7) Miwa, J. H.; Patel, A. K.; Vivatrat, N.; Popek, S. M.; Meyer, A. M. Org.
Lett. 2001, 3, 3373–3375.
(17) Serrano, A. L.; Troxler, T.; Tucker, M. J.; Gai, F. Chem. Phys. Lett. 2010,
487, 303–306.
(18) We limited our temperature range to e75 °C, since some decomposition
of HP35-Leu′1Cnf35 occurred at high temperature. This was sequence-
specific; no decomposition was observed upon heating of Leu′-Pro2-Cnf.
As we envision applications mainly to room-temperature conformational
changes, this should not limit experiments.
(19) McKnight, C. J.; Matsudaira, P. T.; Kim, P. S. Nat. Struct. Biol. 1997, 4,
180–184.
JA1044924
9
14720 J. AM. CHEM. SOC. VOL. 132, NO. 42, 2010