A simple and economical method for the production of13C, 18O-labeled fmoc-amino acids with high levels of enrichment: Applications to isotope-edited IR studies of proteins

Article abstract of DOI:10.1021/ol701913p

Isotope-edited IR of proteins has generated considerable interest. Double labeling with ¹³C and ¹⁸O with high levels of isotopic enrichment is required for residue-specific resolution. Current methods for the preparation of doubly la

Full text of DOI:10.1021/ol701913p

ORGANIC
LETTERS
2007
Vol. 9, No. 24
4935-4937
A Simple and Economical Method for
13
the Production of C,18O-Labeled
Fmoc-Amino Acids with High Levels of
Enrichment: Applications to
Isotope-Edited IR Studies of Proteins
James Marecek,*,† BenBen Song,† Scott Brewer,‡ Jenifer Belyea,‡
R. Brian Dyer,‡ and Daniel P. Raleigh*,†
Department of Chemistry, State UniVersity of New York at Stony Brook, Stony Brook,
New York 11794-3400, and Physical Chemistry and Applied Spectroscopy Group,
Los Alamos National Laboratory, Mail Stop J567, Los Alamos, New Mexico 87545
draleigh@notes.cc.sunysb.edu; james.mareck@stonybrook.edu
Received August 6, 2007
ABSTRACT
Isotope-edited IR of proteins has generated considerable interest. Double labeling with 13C and 18O with high levels of isotopic enrichment is
required for residue-specific resolution. Current methods for the preparation of doubly labeled amino acids give modest 18O enrichment,
limiting the utility of the approach. We report a simple and economical method for preparing 13C,18O-doubly labeled N-(9-fluorenylmethoxycarbonyl)-
amino acids with high levels of enrichment for residues that do not require acid-labile side-chain protecting groups.
An ideal probe of protein folding, protein membrane interac-
tions, and protein dynamics would offer site-specific resolu-
tion and easily interpretable changes in the spectroscopic
signal of the probe, and be minimally invasive.1 Fluorescence-
based experiments are extensively used, but not all proteins
contain Trp or Tyr, and their introduction by mutagenesis
does not always represent a conservative substitution;
furthermore, changes in fluorescence cannot be interpreted
in terms of secondary structure. Alternatively, nonnatural
fluorophores can be introduced; however, the added groups
are usually large and can represent quite drastic perturbations.
ESR spin labels studies have also been profitably employed,
but again the addition of a spin label can represent a drastic
change, and it is hardly clear that such substitutions can safely
be considered to be conservative. In principle IR-based
methods offer a very attractive probe of protein structure
and dynamics. The frequency of the amide-I band is well-
known to be sensitive to secondary structure and to backbone
solvent interactions. Recently developed two-dimensional
infrared (2DIR) methods have generated considerable excite-
ment and promise to extend IR investigations of proteins.1-3
However the FTIR spectrum of even the smallest proteins
consists of an inhomogeneously broadened peak, and site-
specific information is not obtained. 2DIR is also compro-
† State University of New York at Stony Brook, Stony Brook.
‡ Los Alamos National Laboratory.
(1) Krimm, S.; Bandekar, J. AdV. Protein Chem. 1985, 38, 1811.
(2) Zanni, M.; Gnanakaran, S.; Stenger, J.; Hochstrasser, R. J. Phys.
Chem. B. 2001, 105, 6520.
(3) Mukamel, S. Ann. ReV. Phys. Chem. 2000, 51, 691.
10.1021/ol701913p CCC: $37.00
© 2007 American Chemical Society
Published on Web 10/25/2007

mised by limited resolution. Fortunately, the amide-I band
is very sensitive to isotopic substitution at the carbonyl
carbon and carbonyl oxygen, since the band can, to a first
approximation, be viewed as the sum of individual carbonyl
stretching modes. Thus, isotopic substitution will result in a
shift of the amide-I band of a labeled site to lower
frequencies, opening the door to isotope-edited IR studies.
Initial experiments made use of 13C substitution which results
in a ca. 40 cm-1 shift.4,5 While useful for the study of small
to midsize peptides, the approach suffers from serious
drawbacks when applied to proteins or even large peptides.
The modest shift induced by a 13C substitution coupled with
the 1.1% natural abundance of 13C means that the shifted
peak will normally not be baseline resolved and will appear
as a shoulder on a large broad band. Recently, 13Cd18O
labeling of proteins and peptides has been used to partially
circumvent these problems.6-10 The larger isotope shift
induced by the double label, expected to be on the order of
75 cm-1 for a purely local oscillator, leads to significantly
improved resolution provided that derivatives with a high
level of 13C and 18O enrichment are used. This allows true
isotope-edited IR studies and allows the full power of 2DIR
methods to be exploited. The approach obviously requires
13C,18O-labeled amino acids. While, 13C carbonyl-labeled
amino acids are commercially available, the required 13C,18O-
labeled variants are not, and less than optimum 18O enrich-
ment can cause major problems.7 Consider, for example, the
results of 75% enrichment with 18O, a typical value obtained
using current methods. In this case, a sample of a 13C,18O-
labeled 45-residue polypeptide would have a 13C-16O peak
of comparable integrated intensity to the double-labeled peak,
resulting in spectral overlap and a significant loss of
resolution. Protocols now in use to prepare double-labeled
amino acids lead to 18O enrichments on the order of 75 to
80% or less, depending upon the amino acid. Cost is also a
critical issue given the expense of H218O and the large amount
of material needed for solid-phase peptide synthesis.
Here we report a simple and economical method for pre-
paring 13C,18O-double-labeled N-(9-fluorenylmethoxycarbo-
nyl) (FMOC)-amino acids with high levels of 18O enrichment
for residues that do not require acid-labile side-chain
protecting groups, i.e., Gly, Ala, Val, Ile, Leu, Phe, Trp, and
Pro. We demonstrate the utility of the method by preparing
13C,18O-double-labeled FMOC-Ala, -Val, -Ile, -Leu, and
-Phe.
water (1 M in HCl) at elevated temperatures for about 24
h.11,12 We were concerned about loss of label during the
preparation of the desired FMOC derivatives from a doubly
labeled amino acid since we planned to use 9-fluorenyl-
methylchloroformate as the reagent for the incorporation of
the Fmoc group with water/dioxane as the solvent and
sodium carbonate as the base.13,14 In our hands, we have also
found that there are more byproducts generated in these
reactions, and the yields are frequently lower (80-90%) than
in the case of the preparation of t-BOC derivatives. For these
reasons, we prepared the FMOC derivative of the 13C-labeled
amino acid first and then ran the exchange reaction with 18O-
enriched water. Preliminary control experiments indicated
that the protected FMOC amino acid was reasonably stable
for at least 24 h at 100° in mixtures of dioxane and water
which were 0.1 M in HCl. Although there was a slight loss
in overall enrichment, we found it more convenient to
generate the HCl in situ by reacting acetyl chloride with the
18O-enriched water. This allowed preparation of a precise
acid titer without the need for titration and subsequent
adjustments. The acetic acid hydrolysis byproduct was
innocuous. 18O-Enriched water is quite expensive; thus, the
labeling was done in a two-step process. In the first
equilibration step the 18O content was raised to about 75-
85%, depending on the enrichment level of the medium. A
second equilibration brought the final enrichment level to
90-96% as judged by mass spectroscopy. The enrichment
medium was usually 3:2 dioxane/H218O water which was 0.1
M in HCl. The enrichment medium after the second
equilibration step was always recovered by a vacuum
distillation and recycled back to the first stage since its 18
O
content is g90%. The rate of 18O incorporation was very
sensitive to the steric bulk of the amino acid side chain. The
reaction was very fast for alanine which required 3 h for
complete equilibration and was very slow for valine and
isoleucine which required 20-30 h or more for equilibration
(Table 1). Recovery after this two-stage 18O-enrichment
process, except for isoleucine, was usually 90% of the
starting FMOC-amino acid. The recovery for isoleucine was
80%, the losses being caused by decomposition during the
exchange reaction. The final level of incorporation after the
two-step procedure, was between 93 and 96% except for
isoleucine at 90%.
The basic experimental protocol is outlined below. More
detailed information is provided in the Supporting Informa-
tion. A 0.25 M solution of HCl in the 18O-enriched water
was prepared by reacting acetyl chloride (196 mg, 2.5 mmol,
180 uL) with 10.0 mL of the water. The 13C-enriched FMOC
L-amino acid (10 mmol) was mixed with 15 mL of dioxane
in a 100-mL three-neck flask equipped with a reflux
condenser and a nitrogen inlet, and the mixture was heated
to 100° (bath temperature) at which point the solution was
homogeneous. The 18O-enriched water was slowly added.
A homogeneous solution usually resulted although partial
The usual approach for the preparation of labeled amino
acid derivatives is to incorporate the 18O label into the amino
acid, and then perform further manipulations. A number of
labeling protocols have been developed, but the most
common one is equilibration of the amino acid with enriched
(4) Tadesse, L.; Nazarbaghi, R.; Walters, L. J. Am. Chem. Soc. 1991,
113, 7036.
(5) Decatur, S. Acc. Chem. Res. 2006, 39, 169.
(6) Torres, J.; Adams, P. D.; Arkin, I. T. J. Mol. Biol. 2000, 300, 677.
(7) Torres, J.; Kukol, A.; Goodman, J. M.; Arkin, I. T. Biopolymers 2001,
59, 396.
(8) Torres, J.; Briggs, J. A. G.; Arkin, I. T. J. Mol. Biol. 2002, 316, 365.
(9) Fang, C.; Wang, J.; Charnley, A. K.; Smith, A. B., III; Decatur, S.
M.; Hochstrasser, R. M. Chem. Phys. Lett. 2003, 382, 586.
(10) Brewer, S. H.; Song, B.; Raleigh, D. P.; Dyer, R. B. Biochemistry
2007, 46, 3279.
(11) Murphy, R. C.; Clay, K. L. Methods Enzymol. 1990, 193, 338.
(12) Ponnusamy, E.; Jones, C. R.; Fiat, D. J. Labelled Compd. Radiop-
harm. 1987, 24, 773.
(13) Carpino, L. A.; Han, G. Y. J. Org. Chem. 1972, 37, 3404.
(14) Carpino, L. A. Acc. Chem. Res. 1987, 20, 401.
4936
Org. Lett., Vol. 9, No. 24, 2007

Table 1. Reflux Times Used for the Incorporation of 18O into
Each Amino Acid, and the Final Level of 18O Enrichment
time (h) for
time (h) for
final
first exchangea second exchange 18O enrichment (%)
Ala
Val
Leu
Ile
3
20
10
34c
10
not usedb
87b
93
94
90
20
10
30c
10
Phe
95-96
a The level of 18O labeling of the enrichment medium was 93-94% for
the first exchange reaction. A 3:2 mixture of dioxin, 18O water, 0.1 M HCl
was used for Val, Leu, Ile, and Phe. A 1:1 mixture was used for Ala. b This
level of enrichment was acceptable for the application of interest; thus, a
second step was not required. c Complete equilibration was not achieved.
The reaction was stopped to minimize decomposition of the FMOC-amino
acid
Figure 1. Temperature-dependent difference FTIR spectrum (76
- 16 °C) of 13Cd18O A57 HP36. The peak due to the labeled Ala
appears as a negative feature in the difference spectrum at 1572
cm-1 due to loss of the helical structure as the sample is heated.
The negative feature at 1647 cm-1 corresponds to unlabeled (12Cd
16O) residues. As the protein unfolds, new peaks appear at 1601
and 1679 cm-1, assigned to the disordered structure peaks for the
labeled and unlabeled positions, respectively.
solution proved to be sufficient. The mixture was stirred
under nitrogen at a bath temperature of 100-102° for an
appropriate time (Table 1). The reaction was monitored by
removing a drop of the solution and evaporating to dryness
in vacuo then analyzing by TLC. The enrichment level was
checked by mass spectrometry. When the enrichment level
was constant, the enrichment medium was recovered by
evaporating the mixture in vacuo at room temperature using
a short column and collecting the dioxane/water in a liquid-
nitrogen-cooled trap. The solid residue was dried under
vacuum for several hours and the enrichment process
repeated a second time. The crude, enriched FMOC-amino
acid was crystallized from a suitable solvent. Measurement
of the optical rotation showed that no detectable racemization
had occurred.
This simple and affordable approach leads to final 18O-
enrichment levels of 93-96% for Val, Phe, and Leu and
90% for Ile after the two-step enrichment process. Only a
single enrichment step was required to generate 87% 18O-
labeled Ala (Table 1). Recovery after the two-stage 18O-
enrichment process, except for isoleucine, was between 90
and 94% of the starting FMOC-amino acid. The recovery
for isoleucine was less, 80%, because of decomposition
during the exchange reaction.
The power of site-specific 13Cd18O-labeling of proteins
is illustrated by the temperature-dependent difference FTIR
spectrum shown in Figure 1. We prepared a sample of the
villin headpiece helical subdomain, HP-36, selectively
labeled at Ala-57 via solid-phase peptide synthesis. HP-36
is a 36-residue three-helix protein which has become an
extremely popular model system for experimental and
computational studies of protein folding. HP-36 folds
cooperatively and rapidly in isolation. The helical subdomain
is part of the larger villin subdomain, and the numbering
system used here corresponds to that of the full length villin
headpiece domain. The first residue in the helical subdomain
is Leu-42. The 13Cd18O-labeled Ala is in a helix and is buried
in the hydrophobic core in the folded state, a so-called buried
helix position. The figure displays the temperature-dependent
difference FTIR spectrum (76 - 16 °C) of thermally
unfolded sample minus folded sample. The labeled site gives
rise to a negative feature in the difference spectrum at 1572
cm-1 due to loss of the helical structure as the sample is
heated. The negative feature at 1647 cm-1 corresponds to
the unlabeled (12Cd16O) residues in a buried helical con-
formation. As the protein unfolds, new peaks appear at 1601
and 1679 cm-1, assigned to the disordered structure peaks
for the labeled and unlabeled positions, respectively. Thus,
for both the folded and unfolded conformations, the observed
isotope shift for the double label is about 75 cm-1, as
expected for a local oscillator model of the CdO stretching
vibration. The data illustrates two important points; first, one
can easily distinguish a single-labeled position, and second,
specific labeling can be used to follow the unfolding reaction.
In particular, note that both the helical and disordered
structure peaks for the labeled position are discernible in the
data.
Acknowledgment. This work was supported by the NIH
and NSF via Grants NIH GM70941, NSF MCB-614365 to
D.P.R. and NIH GM53640 to R.B.D.
Supporting Information Available: Detailed experi-
mental procedures and compound characterization. This
material is available free of charge via the Internet at
http://pubs.acs.org
OL701913P
Org. Lett., Vol. 9, No. 24, 2007
4937