Use of N-trifluoroacetyl-protected amino acid chlorides in peptide coupling reactions with virtually complete preservation of stereochemistry

Article abstract of DOI:10.1016/j.tet.2003.02.002

The use of protected amino acid chlorides for peptide coupling reactions has long been avoided due to the extensive racemization that commonly occurs during either the acid chloride formation or the coupling reaction itself. Conditions are described which

Full text of DOI:10.1016/j.tet.2003.02.002

Tetrahedron 59 (2003) 9019–9029
Use of N-trifluoroacetyl-protected amino acid chlorides in
peptide coupling reactions with virtually complete preservation
of stereochemistry
†
Paul A. Jass,a Victor W. Rosso,a Saibaba Racha,b Nachimuthu Soundararajan,a John J. Venit,a
,
Andrew Rusowicz,a Shankar Swaminathan,a Julia Livshitza and Edward J. Delaneya
,
*
aBristol-Myers Squibb Company, Process Research and Development, P.O. Box 191, New Brunswick, NJ 0890, USA
bBristol-Myers Squibb Company, P.O. Box 4755, Mail Stop: N-6, Syracuse, NY 13221-4755, USA
Received 19 December 2002; accepted 2 February 2003
Abstract—The use of protected amino acid chlorides for peptide coupling reactions has long been avoided due to the extensive racemization
that commonly occurs during either the acid chloride formation or the coupling reaction itself. Conditions are described which allow
N-trifluoroacetyl-protected amino acid chlorides to be generated in high purity and with high retention of stereochemical integrity. Control of
temperature is the predominant factor in controlling racemization, and rapid formation of acid chlorides under low temperature can be
conveniently achieved using Vilsmeier reagent. Stereochemical integrity is further maintained when coupling of N-trifluoroacetyl acid
chlorides is carried out with amino acid esters under Schotten–Baumann conditions using specific controls on pH, temperature, and agitation.
Second order rate constants for coupling and the azlactone formation associated with racemization were measured to be 4260 and
3.6 L/mol s, respectively. This high rate differential allows for the reaction to be run with a minimum excess of amine ester, and makes it
suitable for continuous processing. The applicability of the preferred coupling conditions to a range of amino acid couplings is described.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
Peptide bond formation is among the most important and
widely studied chemical transformations in synthetic
organic chemistry. Racemization of the amine-protected
amino acid partner is the predominant problem in coupling
two amino acids, and this issue has traditionally been
addressed by using urethane-based protecting groups, Boc,
Cbz, and Fmoc being among the most common. Simple
amide acids, such as N-acetyl, are well known to racemize at
Scheme 1.
an unacceptable rate upon activation of the carboxylic acid
function in the presence of base. The mechanism for
use of the intermediate activated esters (e.g. –OBt, or
racemization is widely accepted to occur via an azlactone
–OSu) formed in conjunction with carbodiimides (DCC or
intermediate (Scheme 1).1
EDAC).
In contrast, urethane-protecting groups have been shown to
Use of the trifluoroacetyl group as an amine protecting
be much less prone to azlactone formation and subsequent
scrambling of the chiral center. Nonetheless, appreciable
racemization can still occur, and is typically suppressed by
group is merely footnoted in most textbooks on peptide
synthesis, as early efforts to prepare N-trifluoroacetyl amino
acids employing trifluoroacetic anhydride were plagued by
racemization, and the use of other groups, such as Boc and
Keywords: peptide coupling; N-trifluoroacetyl protection; acid chloride
formation; amide formation; Schotten–Baumann; azlactone formation;
kinetic rate of coupling; kinetic rate of racemization.
Cbz, proved to be acceptable alternatives. In spite of later
reports of a much cleaner method to prepare N-trifluoro-
acetyl-protected amino acids using ethyl trifluoroacetate,2
and in spite of the ease which with TFA may be removed
(dilute base), N-trifluoroacetyl protection is rarely used in
the preparation of peptides and peptide mimetic drugs.
*
Corresponding author. Tel.: þ1-732-519-3092; fax: þ1-732-519-2531;
e-mail: edward.delaney@bms.com
Present address: Cambrex Pharma, Cambrex Charles City, Inc., 1205 11th
Street, Charles City, IA 50616-3466, USA.
†
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.02.002

9020
P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
Many attempts were made during the 1950s and early 1960s
to prepare N-protected amino acid chlorides and employ
them in peptide couplings. However, with the exception of
glycine, this approach has universally been dismissed as
untenable due to the high propensity of N-protected amino
acid chlorides to racemize.3 Nonetheless, the use of chiral
TFA-protected amino acid chlorides in Friedel–Crafts
acylations has been reported to proceed with high retention
of stereochemical integrity.4 At the same time, the
generation of Fmoc-protected amino acid chlorides and
their use in peptide couplings under Schotten–Baumann
conditions has also been reported to proceed with high
retention of stereochemical integrity.5 These reports
suggested that N-trifluoroacetyl amino acid chlorides may
be similarly useful in peptide couplings.
excesses of intermediates normally required to force
coupling to completion. In the work described herein, we
report general techniques by which N-trifluoroacetyl
chlorides may be effectively prepared and used in peptide
coupling reactions with very little loss of stereochemical
integrity. This method has been demonstrated on a variety of
amino acid substrates, and in many cases should prove a
superior approach at commercial scale, where significant
advantages of cost, simplicity, and atom-economy may be
recognized over long-standing methods.
2. Results and discussion
2.1. N-Trifluoroacetylation of amino acids
In the course of process development studies, we sought (1)
an amino acid protecting group that would be inexpensive
and more readily removed for our purpose than existing
agents, and (2) coupling conditions that would provide a
high rate of conversion while minimizing the molar
Slight modifications were made to extend the methods of
Steglich2a and Curphy2b for converting amino acids to their
respective N-trifluoroacetyl derivatives. General conditions
developed involve the generation of the desired amino acid
potassium salt in methanol, and subsequent treatment with
Scheme 2.

P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
9021
methyl trifluoroacetate. The aminolysis reaction can be
brought to completion at 58C within 2 h when 1.2–2 equiv.
of TFA ester are used in the presence of as little as 0.1 equiv.
excess of potassium methoxide. For convenience, the amino
acid is suspended in dry methanol, and 1.1 equiv. of
potassium methoxide solution (or 2.1 equiv. in the case of
amine diacids) are added to insure potassium salt formation
and catalysis of the reaction. Potassium and dipotassium
salts of the amino acids studied were found to be much more
soluble in this system than their respective sodium or
triethylamine salts, and dramatic improvements in reaction
time were obtained in the case of amine diacids, such as
aspartic acid, glutamic acid, and homocystine. Several
representative examples of the recommended conditions are
provided in Section 4.
chloride, and that coupling of the latter with 2 under
Schotten–Baumann conditions (vide infra) provided 3 in
high yield with high preservation of stereochemical
integrity. Since amino ester 4 was costly to produce, the
prospect of achieving rapid reaction completion with slight
excesses of amine while maintaining high retention of
stereochemistry motivated us to develop a rugged process
around this finding.
While the extensive racemization that accompanies the use
of protected amino acid chlorides in peptide couplings is
well documented, no systematic investigations appear to
have been done to determine the point at which racemiza-
tion occurs, or the factors that control it. The use of TFA-
protected acid chloride substrates in Friedel–Crafts
reactions has been reported in many cases4 to proceed
with high retention of enantiomeric integrity, suggesting
that racemization is not so much a problem during acid
chloride formation as it is during coupling. At least one GC
method for the determination of chiral purity has been
reported based on the reaction of N-TFA protected amino
acid chlorides with chiral primary amines.6 Finally, a report
from the early days of peptide methodology exploration
recommended widespread use of N-trifluoroacetyl protec-
tion based on a case where N-TFA-L-Phe acid chloride
coupling with aniline was found to proceed with retention of
stereochemistry!7 In light of these past reports and the
success observed in coupling 1 with 4, it appeared that the
general preparation and use of N-TFA-protected acid
chlorides in peptide couplings was merely dependent on
specific techniques that, once understood, could be applied
broadly.
2.2. Preparation of N-trifluoroacetyl amino acid
chlorides
In the course of developing a process for the anti-
hypertensive agent, Vanlev (Scheme 2), we sought
coupling conditions that would allow for the diacid 1 to
be reacted with amino ester 4 to provide isolated
intermediate 5 in high yield with minimal loss of
stereochemical integrity.
The intermediacy of acid chlorides for this purpose showed
promise over traditional coupling methods in screening
studies, and was deemed worthy of exploration in spite our
initial skepticism. With further development we found that 2
could be effectively activated with Vilsmeier reagent at low
temperature to cleanly provide the intermediate acid
Table 1. Levels of diastereomeric dipeptide formed under various coupling conditions
Reagent (solvent)
Temperature
(8C)
Sampling
time
(h)a,b
Example 1,
TFA-Met-Phe-OMe
D,L diastereomer
(area% by GC)
Example 2,
TFA-Ile-Phe-OMe
D,L diastereomer
(area% by GC)
SOCl2 (toluene)
80
25
5
4
5
9
5.3
7.0
12.4
2.4
2.3
12.9
CO2Cl2 / cat. DMF (CH2Cl2)
Vilsmeier (BuOAc)
1.5
2.5
24.0
1.8
2.7
24.4
1.7
2.7
28.0
4
5
9
27
0.9
1.0
1.3
2.7
2.8
3.3
3.7
6.2
Vilsmeier (BuOAc)
Vilsmeier (BuOAc)
0
4.5
6
23
0.53
0.82
1.08
0.31
0.43
1.19
210
4
7
27
,0.2
,0.2
0.22
,0.2
0.22
0.27
Vilsmeier (CH2Cl2)
Vilsmeier (CH2Cl2)
210
215
3
23
0.35
0.42
–
–
1
3.25
21
,0.2
,0.2
0.43
–
–
–
a
Times shown are from initiation of acid chloride reaction to point of sampling for coupling reaction; actual completion of acid chloride formation was
confirmed within 30–60 min.
TFA-L-Met and TFA-L-Ile acid chloride samples tested by quenching into 08C BuOAc containing 3.5 equiv. of L-Phe-OMe; dipeptide derivatives were then
analyzed for diastereomer content by GC.
b

9022
P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
To test the sensitivity of TFA acid chlorides to racemization
under the conditions of their formation, two standard N-TFA
protected amino acids (TFA-L-Met and TFA-L-Ile) were
subjected to various conditions in order to convert them to
their respective acid chlorides. Once formed, the stability of
the acid chlorides to racemization was monitored by
withdrawing samples over time and subjecting them to a
rapid coupling with a 3-fold excess of L-phenylalanine
methyl ester at about 08C. The resulting dipeptide samples
were then analyzed by GC for diastereomeric purity.
Variables investigated were (1) reagent/solvent combi-
nations, (2) temperature and (3) time. The results of this
study are shown in Table 1.
ments were conducted using the Vilsmeier/dichloromethane
condition at about 58C in order to achieve maximum
solubility of all components while maintaining concen-
tration as close as possible to the original condition
(Table 2). In addition to a control experiment in which
MTBE was added to insure full solubility and availability of
the DMF/HCl complex, identically completed activations
were spiked with 4 and 9 extra equivalents of HCl
introduced as MTBE solutions. Interestingly, the presence
of added HCl had a minimal impact in both cases, as the acid
chloride solutions retained their stability relative to the
control. These findings clearly ruled out acidity as a
significant contributing factor to acid chloride racemization.
Considerable racemization was observed for acid chlorides
formed under most experimental conditions, but clearly, the
key variable for controlling racemization is temperature.
Toward that end, use of Vilsmeier reagent at or below
2108C provided a reasonably rapid reaction while con-
ferring excellent control over racemization. Moreover,
stability was demonstrated for at least 24 h when completed
reaction mixtures were maintained at that temperature.
While Vilsmeier reagent is relatively insoluble in organic
solvents, it reacts readily through its soluble fraction.
Various protected amino acids tested were reacted com-
pletely in either ethyl or butyl acetate within 30–60 min,
even when reactions were carried out at or slightly below
2108C. Dichloromethane also appeared to be a suitable
solvent with temperatures at or below 2108C, although
greater care is required to control the reaction exotherm
owing to the better solubility of Vilsmeier reagent in
dichloromethane relative to that observed with the ester
solvents.
DMF was found to be an unsuitable solvent for the
activation step, as racemization rates were shown to be
quite high in its presence. Reactions completed using
Vilsmeier reagent in BuOAc at 2108C were also found to
reform significant levels of free acid when DMF was added.
These findings suggest an appreciable reverse rate for the
Vilsmeier activation step, and implicate activated inter-
mediate 7 (Scheme 3) as a species which is likely present at
low concentration, but which possesses a much faster
racemization rate relative to the acid chloride. Toward that
end, in tying up free DMF, the HCl generated during the
activation step helps suppress reverse reaction and plays an
important role in stabilizing the acid chloride product.
2.3. N-Trifluoroacetyl amino acid chloride coupling with
amino esters under Schotten–Baumann conditions
Racemization during coupling is widely held to occur via
azlactone intermediates which form most readily under
basic conditions. It should come as no surprise then that
peptide couplings are most successfully performed under
conditions that maximize stability of the activated acid
moiety while allowing the amine partner to exist as free base
under as close to neutral pH as possible. For this reason,
N-methyl morpholine (pKa ,7.4) has been cited to be a
preferred base for neutralizing the acidic counterion
generated from mixed anhydride couplings, and for freeing
the amine coupling partner for reaction.8 In the course of
When using Vilsmeier reagent to effect activation, the
formation of DMF/HCl complex was observed as a
precipitate or as a second liquid phase. This finding initially
suggested that the formation and segregation of DMF/HCl
complex may play a role in stabilizing the acid chloride
product by suppressing an acid-catalyzed racemization
pathway. To test the relative importance of DMF/HCl
segregation to racemization, several independent experi-
Table 2. Effect of acidity and reversibility on racemization during activation
Alterations to standard activation procedure
Tempeature
(8C)
Time
(h)a,b
TFA-Met-Phe-OMe
D,L diastereomer
(area% by GC)
CH2Cl2 solvent, MTBE added (control)
5
5
4
6
27
0.89
0.97
1.15
MTBE containing four additional equivalents HCl added to MeCl2
MTBE containing nine additional equivalents HCl added to CH2Cl2
4
6
27
0.57
1.00
0.97
5
4
6
27
0.54
0.74
0.75
DMF as solvent
210
3
23
15.0
24.3
a
Times shown are from initiation of acid chloride reaction to point of sampling for coupling reaction; actual completion of acid chloride formation was
confirmed within 30–60 min.
TFA-L-Met and TFA-L-Ile acid chloride samples tested by quenching into 08C BuOAc containing 3.5 equiv. of L-Phe-Ome; dipeptide derivatives were then
analyzed for diastereomer content by GC.
b

P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
9023
mixing to minimize the competing hydrolysis, and to allow
the pH to be maintained in a stable condition with the
concomitant addition of 1N sodium hydroxide. Mainten-
ance of the aqueous bicarbonate buffer between pH 7.0 and
8.0 helps minimize hydrolysis. Most importantly, however,
it also serves as a superior surrogate for weak tertiary
amines in neutralizing the acid generated, since the
competing protonation of the amine coupling partner that
typically occurs with tertiary amines is eliminated.
2.4. Kinetic study to determine rates of coupling,
azlactone formation, and hydrolysis
While the conditions cited work optimally for the Vanlev
intermediate, a kinetic study was undertaken on a more
general model reaction (N-TFA-L-Phe acid chloride with
L-Phe-OMe) to gain a fundamental understanding of the
relative rates of the reaction and side-reactions involved,
which could then be extrapolated to other closely-related
couplings. In preparation for this study, an authentic sample
of the corresponding N-TFA-L-Phe azlactone was prepared
and confirmed to be stable to the HPLC conditions by which
the rates would be measured.
Scheme 3.
developing an economical process for Vanlev, we explored
the applicability of the Schotten–Baumann technique to the
coupling of the acid chloride of bis-N-TFA protected
homocystine 1 with the acetal amine ester 2 (Scheme 2).
Done properly, this reaction was found to proceed very
cleanly, with very little hydrolysis, and virtually no loss of
stereochemical integrity. We subsequently studied this
coupling with an eye toward generality in peptide couplings.
Keys to the method are temperature control (0–58C),
controlled addition of the acid chloride mixture, the
maintenance of pH between 7.0 and 8.0, and highly efficient
mixing. For this system, a narrow pH range was found to be
necessary; with significantly higher pH, azlactone for-
mation, racemization, and acid chloride hydrolysis com-
peted more effectively with coupling, while at significantly
lower pH, protonation of the amine ester inhibited both its
transfer to the organic compartment and the coupling
reaction itself. Optimally, a buffered environment
(K2HPO4) is used to maintain the pH in the aqueous layer,
and pH is maintained throughout addition of the acid
chloride by simultaneous addition of aqueous sodium
hydroxide. This control is most efficiently achieved using
a feedback loop from a pH meter to control the flow of
aqueous base. Relatively constant pH can also be insured by
employing a large excess of aqueous sodium bicarbonate as
a buffer to the aqueous phase. Control of racemization is
virtually equivalent under these more convenient con-
ditions. Amine esters are typically introduced into the
procedure as their hydrochloride salts, and sufficient base is
included at the onset to effect liberation of the free base.
Since many amine esters are prone to dimer and
diketopiperazine formation over time, extensive hold
times of the free base should be avoided prior to coupling.
Coupling of TFA protected L-Phe acid chloride with L-Phe
methyl ester was measured at 58C under dilute conditions to
possess a second order rate constant of 4260 L/mol s
(Fig. 1), and a half-life back-calculated on the basis of the
normal concentration to be 0.7 ms. The competing rate of
azlactone formation was estimated at 3.6 L/mol s based on
the amount of azlactone competitively formed as deter-
mined by HPLC (corresponding half-life of 0.8 s).‡
A
secondary study of the rate of azlactone formation in the
presence of diisopropylethylamine measured 4.1 L/mol s
(Fig. 2). These results were very much in line with results
from HPLC that indicate very little azlactone formation
during coupling (,0.4%). In contrast, the pseudo first-order
acid chloride hydrolysis at pH 8.5 was roughly measured to
be about 0.04/s (half-life of 17 s). This finding is also
consistent with the relatively small amount of free acid
formed by acid chloride hydrolysis under the Schotten–
Baumann conditions employed.
Beyond the excellent control of azlactone formation and
hydrolysis provided under these conditions, the extremely
rapid rate of coupling provides two key advantages. First, it
allows for the reaction to be completed rapidly and
efficiently with a minimal excess of amine ester (e.g.
equivalent excesses of 5% or less). Second, it makes the
procedure highly amenable to the engineering of a
continuous process.
2.5. Scope and limitations
Thorough mixing is needed to insure full availability of the
amine ester present in the aqueous phase to the organic
phase, and the avoidance of localized acid chloride
hydrolysis and azlactone formation when the reaction is
carried out on large scale. This aspect can be particularly
important, as some of the amine esters measured possess
partition coefficients that favor the aqueous phase over the
organic in ratios of 5 to 1 or greater. Controlled direct
addition of the TFA-protected acid chloride mixture has also
been found to be helpful in conjunction with thorough
In order to examine the range of applicability of this
method, the TFA derivatives of several additional amino
acids were converted to their respective acid chlorides, and
‡
These are simplified half-live estimations, based on the assumptions of
having both substrates being present in the organic compartment, instant
mixing, and perfect temperature maintenance. While this condition
obviously does not exist throughout the course of reaction, it is most
closely held during reaction initiation, and the rates are provided to give a
sense of feeling for the overall reaction speed.

9024
P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
Figure 1. Measurement of second-order rate constant for reaction of N-TFA-L-phenylalanyl acid chloride with L-phenylalanine methyl ester (coupling).
coupling with L-phenylalanine ethyl ester was carried out
under Schotten–Baumann conditions. Rapid coupling
completion was achieved in all cases. While the data
shown in Table 3 verify that the coupling method is
generally applicable to other amino acids, we have found
minor losses of stereochemical integrity (,0.5% wrong
diastereomer) when variables such as buffer concentration
and choice of base are varied. Most important of these
variables, however, is the temperature control during acid
chloride formation and during the hold time prior to its
addition to the coupling mixture. While yields obtained for
the coupling partners shown in Table 3 are generally very
Figure 2. Measurement of second-order rate constant for reaction of N-TFA-L-phenylalanyl acid chloride with diisopropylethylamine (azlactone formation).

P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
9025
Table 3. Coupling of various TFA-amino acid chlorides with L-phenyl-
alanine methyl ester
5 mL samples onto a YMC ODS-A 4.6£50 mm column, and
eluting with 75/25 0.1% aqueous phosphoric acid/aceto-
nitrile mixture for 1 min, followed by a linear gradient over
4 min to 15/85 0.1% aqueous phosphoric acid/acetonitrile,
finishing with a 1 min elution with the later. Flow rate was
2 mL/min and detection was at 210 nm. All compounds
were linear in the analyzed range and each component was
corrected for molar extinction coefficient. Vilsmeier
reagent used in all studies was purchased from commercial
sources.
Coupled products
Yield
(%)
D,L diastereomer
(%)
Method
TFA-N-L-Met-L-Phe-OMe
TFA-N-L-Ile-L-Phe-OMe
TFA-N-L-Pro-L-Phe-OMe
TFA-N-L-Phe-L-Phe-OMe
TFA-N-L-Leu-L-Phe-OMe
Bis-TFA-N,N-L-Lys-L-Phe-OMe
TFA-N-L-Tyr-L-Phe-OMe
92
98
44
91
95
89
74
,0.2
,0.2
,0.2
,0.2
,0.2
,0.2
,0.2
4
1
3
1
3
2
2
Methods used to determine diastereomeric purity (Table 3)
were as follows.
good, coupling with TFA proline was complicated by side
reactions which substantially lowered the yield, illustrating
the fact that optimization of coupling conditions may be
warranted with specific individual cases.
Method 1. Isocratic HPLC, 1 mL/min 60% 0.01 M KH2PO4
(pH 3.0), HP ODS Hypersil 125 mm£4 mm column,
detection at 220 nm.
Method 2. Isocratic HPLC, 1 mL/min, 80% water/20%
methanol, Astec Chirobiotic T, 250 mm£4.6 mm column,
detection at 220 nm.
3. Conclusions
These studies conclusively demonstrate the viability of
carrying out acid chloride-mediated peptide couplings in
conjunction with N-trifluoroacetyl protection, without
appreciable scrambling of stereochemistry. Use of
Vilsmeier reagent at low temperature (2108C or below) in
ethyl acetate, butyl acetate, or dichloromethane provides a
special condition allowing for N-TFA-protected amino
acids to be efficiently converted to their respective acid
chlorides with high retention of stereochemistry. While
other means of forming the acid chloride can also be used
with adequate temperature control, the rapid and clean
nature of the Vilsmeier/low temperature technique provides
an advantage in providing stability of the acid chloride
toward racemization prior to coupling. Subsequent use of
the Schotten–Baumann technique, in conjunction with
excellent agitation and pH control during reaction, provides
a mild set of conditions under which coupling can take place
with very little loss of stereochemical integrity. Once
peptide coupling is completed, the protecting trifluoroacetyl
group can be readily removed using methods well
documented in the chemical literature, including base-
assisted hydrolyses sufficiently mild as to be compatible
with the remote presence of a methyl ester.9 With these
specific conditions, the use of TFA protection becomes
highly competitive with other more commonly used amine
protecting groups. In fact, the coupling conditions described
in Scheme 2 were the basis of a process by which we
ultimately made Vanlev intermediate 5 in high yield and
purity at .100 kg input scale. Furthermore, with further
optimization, the method could conceivably be used in
repeat cycles to prepare polypeptides in a highly efficient,
continuous manner.
Method 3. Hewlett–Packard 6890 GC, Helium flow
47 cm/s, 2208C Hewlett Packard HP-5 (5%diphenyl–
95%dimethylsiloxane) column, 25 m£0.32 mm i.d.£
0.17 mm capillary column, FID detector.
Method 4. Hewlett–Packard 6890 GC, He flow 47 cm/s,
2308C Hewlett–Packard HP-5 (5% diphenyl–95%
dimethylsiloxane) column, 25 m£0.32 mm i.d.£0.17 mm
capillary column, FID detector.
4.1. General procedure for TFA protection of amino
acids
The amino acid substrate (10 mmol) was dissolved in
2.3 mL of 4.4 M (1.0 equiv.) potassium methoxide in
methanol. The solution was then reacted with 1.2 mL
ethyl trifluoroacetate (2.0 equiv.) at 408C until complete
(usually ,0.5% remaining starting material) by HPLC
assay. The solution was cooled and quenched into 4.1 mL of
2.5 M aqueous HCl and extracted with 8 mL of ethyl
acetate. The organic phase was washed once with 10 wt%
brine in 1 M HCl, and concentrated to a residue. The residue
was then crystallized from a solvent combination, or
converted directly to its acid chloride as described below.
In cases where it was difficult to isolate the TFA protected
amino acids,10 the crude material obtained was telescoped
directly into the coupling step. Reaction yields are typically
close to quantitative, and the yields reported below are
unoptimized.
4.1.1. N-Trifluoroacetyl-L-phenylalanine. Isolated yield:
84%. Colorless solid (1:1 heptane/toluene). Mp: 90–918C;
IR (cm21, KBr): 1711, 1562, 1214, 1178, 1168, 758, 702;
1H NMR (DMSO-d6) d: 3.00 (abq, 1H), 3.23 (abq, 1H),
4.54 (abq, 1H), 7.27 (m, 5H), 8.83 (d, NH), 13.0 (br, 1H);
19F NMR (DMSO-d6), d: 275.3 (s, CF3); 13C NMR
(DMSO-d6), d: 35.7, 54.1, 126.7, 128.3, 129.1, 137.4,
156.4, 171.5; MS m/z M2H2 260, MþCOCF32 374. Anal:
calcd for C11H10NO3F3þ3.56% H2O measured by KF: C,
48.78; H, 4.14; N, 5.17; F, 21.05. Found: C, 48.63; H, 4.04;
N, 5.13; F, 21.05.
4. Experimental
All NMR spectra were obtained in DMSO-d6 or CD3CN
using a 300 MHz spectrometer. Uncorrected melting points
were determined on a capillary melting point apparatus.
HPLC assays to measure acid chloride formation (via
MeOH quench) and coupling were performed by injecting

9026
P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
4.1.2. L-Phenylalanine, N-trifluoroacetyl. Isolated yield:
70%. Colorless solid (1:2 ethyl acetate/heptane). Mp: 91–
928C; IR (cm21, KBr): 1714, 1559, 1212, 1178, 757, 700;
1H NMR (CD3CN) d: 3.05 (abq, 1H), 3.25 (abq, 1H), 4.68
(abq, 1H), 7.26 (m, 5H), 7.72 (d, 1H); 19F NMR, d: 274.2
(s, CF3); 13C NMR (CDCl3) d: 37.1, 53.3, 117.3q, 127.8,
128.9, 129.3, 134.3, 156.5q, 174.0; MS m/z M2H2 260,
MþCOCF23 374. Anal. calcd for C11H10NO3F3: C 49.45; H,
4.02; N, 5.24; F, 19.94. Found: C, 49.32; H, 4.12; N, 5.04; F,
20.15.
702; 1H NMR (CDCl3) d: 2.1 (m, 2H), 2.3 (m, 2H), 3.8 (m,
1H), 4.6 (m, 1H) 10.7 (brs, OH); 19F NMR, d: 273.3 (s,
CF3); 13C NMR (CDCl3) d: 24.8, 28.3, 47.3, 60.1, 115.8q,
156.1q, 174.9. MS m/z M2H2¼210.
4.2. General procedure for acid chloride formation
To a 250-mL flask, set up for cooling to 210 to 2158C and
equipped with an overhead agitator and a nitrogen inlet, was
added TFA protected (L)-amino acid (20.4 mmol) and
n-BuOAc (50 mL). The stirred slurry was cooled to 210 to
2128C. With vigorous agitation, Vilsmeier reagent was
charged in three equal portions (2 g each) with 0.5 h stir
time between each addition (6 g total, 46.9 mmol,
2.3 equiv.). Typically, the acid chloride preparation was
complete after 3 h from the initial charge as determined by
HPLC analysis. The acid chloride was maintained at 210 to
2158C prior to use in the subsequent coupling reaction.
4.1.3. L-Methionine, N-trifluoroacetyl. Isolated yield:
81%. Colorless solid (CH2Cl2/toluene/heptane 1:2). Mp:
618C; IR (cm21, KBr): 1710, 1746, 1567, 1244, 1224, 1183,
1158, 727, 691; 1H NMR (CD3CN) d: 2.08 (s, 3H), 2.18 (m,
2H), 2.62 (m, 2H) 4.58 (m, 1H) 7.8 (br, 1H); 19F NMR, d:
274.1 (s, CF3); 13C NMR (CDCl3) d: 15.2, 29.7, 30.1, 51.9,
115 (q), 157.4 (q), 174.3; MS m/z MþNH24 ¼263. Anal.
calcd for C7H10NO3F3S þ3.34% H2O measured by KF: C,
33.14; H, 4.35; N, 5.52; S, 12.64; F, 22.47. Found: C, 32.99;
H, 4.34; N, 5.39; S, 13.14; F 22.43.
4.3. General procedure for amine ester coupling
To a 250-mL round-bottomed flask was charged (L)-
phenylalanine ethyl ester·HCl (4.9 g, 1.05 equiv.) followed
by sodium bicarbonate (2.1 g, 1.2 equiv.) and 50 mL of pH
8 buffer solution (prepared by mixing 250 mL 0.1 M
KH2PO4þ230 mL 0.1 M NaOH and adjusting to 500 mL
with water). The aqueous mixture was cooled to 1–48C
(Note: Cooling below 08C resulted in a thick slurry). While
maintaining its temperature below 2108C, the acid chloride
solution was added drop wise via cannula (nitrogen
pressure). The reaction temperature was maintained
between 1 and 48C using an ice bath. In addition, the
reaction pH was controlled between 7.0 and 8.0 by
concurrent charge of 1 M sodium hydroxide. This is most
effectively achieved using a pH meter and automated
feedback control of the addition rate. Alternatively, 5 equiv.
of NaHCO3 can be used in the same volume as above.
HPLC analysis typically indicated reaction completion
immediately following acid chloride addition completion.
The phases were then separated and the organic phase was
washed twice with 50 mL of water. The rich organic phase
was concentrated to minimal volume under nitrogen at
408C. The product was then crystallized from 5 mL/g butyl
acetate and 5–20 mL/g heptane and collected by vacuum
filtration. In cases where the TFA-protected amino acid
starting material was not crystallized, but telescoped
directly into the coupling, solvent was exchanged to butyl
acetate until ,0.001% methanol was detected by GC. The
N-protected amino acid was then dissolved in 10 mL/g ethyl
acetate/butyl acetate, (3:1) converted to acid chloride, and
coupled with the same protocol as above.
4.1.4. L-Lysine, N2,N6-bis(trifluoroacetyl). Isolated yield:
60%. Colorless solid (from chloroform after passing
through alumina column with THF). Mp: 113–1158C; IR
1
(cm21 KBr): 1740, 1705, 1560, 1230, 1170, 735; H NMR
(CD3CN) d: 1.32 (m, 2H), 1.48 (m, 2H), 1.82 (abq, 2H),
3.29 (q, 1H), 4.38 (m, 1H), 7.6 (s, NH), 7.8 (d, NH), 8.95 (br,
COOH); 19F NMR, d: 275.1 (s, CF3), 275.4 (s, CF3); 13C
NMR (CD3CN) d: 23.5, 28.8, 30.9, 40.1, 53.7, 116.4 (q),
158.1 (q), 172.7. Anal. calcd for C10H12N2O4F6þ3.34%
H2O: C, 35.51; H, 3.58; N, 8.28; F, 33.70. Found: C, 35.44;
H, 3.59; N, 8.13; F, 32.79.
4.1.5. L-Tyrosine, N-trifluoroacetyl. (Note: required
2 equiv. KOMe) Isolated yield: 86%. Colorless solid
(from MTBE/hexane 1:1). Mp 194–1958C; IR (cm21
,
1
KBr): 1694, 1560, 1516, 1250, 1210, 1185, 1152, 678; H
NMR (CD3CN) d: 3.05 (abq, 2H), 4.60 (abq, 1H), 6.72 (d,
2H) 7.05 (d, 2H) 7.65 (d, NH); 19F NMR, d: 275.4 (s, CF3),
275.4 (s, CF3); 13C NMR (CD3CN) d: 36, 55, 116, 117q,
128, 131, 135, 148, 155q, 173; MS m/z M2H2¼276. Anal.
calcd for C11H10NO4F3þ3.34% H2O: C, 47.66; H, 3.64; N,
5.05. Found: C, 47.65; H, 3.74; N, 4.97.
4.1.6. L-Leucine, N-trifluoroacetyl. Isolated yield: 93%
(crude, yellow oil). IR (cm21, KBr): 1706, 1562, 1183,
1260, 1209, 1168, 732, 702, 922; 1H NMR (CDCl3) d: 0.96
(d, 6H), 1.31 (m, 1H), 1.78 (m, 2H), 4.64 (m, 1H) 7.2 (d,
NH), 9.8 (brs, OH); 19F NMR, d: 276.4 (s, CF3); 13C NMR
(CDCl3) d: 21.5, 22.6, 24.8, 40.6, 51.2, 117.5 (q), 157.1 (q),
175.7; MS m/z MþNH24 ¼245.
4.3.1. L-Phenylalanine, N-L-phenylalanyl, N-trifluoro-
acetyl-, ethyl ester. Yield: 91%. Colorless solid (from butyl
acetate/heptane, 1:3). Mp: 135–1368C; IR (cm21, KBr):
1716, 1654, 1547, 1186, 785, 702; 1H NMR (CDCl3) d: 1.3
(t, 3H), 3.05 (m, 2H), 3.15 (m, 2H), 4.18 (m, 2H), 4.55 (m,
1H), 4.63 m 1H, 5.75 d 1H, 6.90 d 2H, 7.12 d 1H, 7.12 (m,
8H); 19F NMR, d: 278.37 (s, CF3); 13C NMR d: 35.1, 93.6,
15.8q, 128.4, 129.7, 130.5, 134.6, 156.3q, 164.8, 168.8; MS
m/z M2H2¼437, MþNH24 ¼454. Anal. calcd for
C22H23N2O4F3: C, 60.55; H, 5.31; N, 6.42; F, 13.06.
Found: C, 60.57; H, 5.27; N, 6.10; F, 13.14.
4.1.7. L-Isoluecine, N-trifluoroacetyl. Isolated yield: 61%
(waxy solid). Mp 618C; IR (cm21, KBr): 1731, 1567, 1470,
1
1393, 1372, 1163, 732; H NMR (CDCl3) d: 0.97 (d, 3H),
0.99 (d, 3H), 1.29 (m, 2H), 2.05 (m, 1H) 4.6 (m, 1H), 7.2 (d,
NH), 11.0 (brs, OH); 19F NMR, d: 276.3 (s, CF3); 13C NMR
(CDCl3) d: 11.3, 15.1, 24.9, 37.5, 56.9, 115.2, (q, J¼87 Hz),
157.3 (q, J¼39 Hz), 174.7; MS m/z MþNH24 ¼245.
4.1.8. L-Proline, N-trifluoroacetyl. Isolated yield: 73%
(crude oil). IR (cm21, KBr): 1700, 1454, 1352, 758, 722,

P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
9027
4.3.2. L-Phenylalanine, N-L-lysyl, N2,N6-bis(trifluoro-
acetyl)-, ethyl ester. Yield: 89%. Colorless solid (from
NMR, d: 271.1 (s, CF3), 270.2 (s, CF3), 69.7:29.3 ratio;
13C NMR (DMSO-d6) d: 13.9, 19.8, 24.3, 28.7 m, 36.5,
47.0, 53.7, 60.6, 98.2q, 126.5, 128.2m 128.9, 138.2, 154.0q,
170.1, 172.0; MS m/z M2H2¼387, MþNH42¼404. Anal.
calcd for C18H21N2O4F3: C, 55.96; H, 5.48; N, 7.25; F,
14.75. Found: C, 55.99; H, 5.46; N, 7.25; F, 14.85.
butyl acetate/heptane, 1:2). Mp:150–1528C; IR (cm21
,
1
KBr): 1705, 1649, 1183, 702; H NMR (CDCl3) d: 1.28 (t,
3H), 2.03 (m, 2H), 1.78 (m, 1H) 1.82 (abq, 2H), 3.13 (q,
2H), 3.26 (q, 2H), 4.30 (q, 2H), 4.38 (q, 1H), 4.75 (q, 1H),
6.12 (br, NH), 6.73 (br, NH), 7.18 (m, 2H), 7.28 (m 3H); 19F
NMR, d: 273.64 (s, CF3), 273.96 (s, CF3); 13C NMR
(DMSO-d6) d: 13.9, 22.7, 27.8, 30.7, 36.5, 38.7, 53.0, 53.8,
60.5, 114.2, 118.3, 126.6, 128.2, 129.1, 137.0, 157.1, 170.4,
171.2; MS M2H2¼514, MþNH24 ¼531. Anal. calcd for
C21H25N3O5F6: C, 49.13; H, 4.91; N, 8.18; F, 22.20. Found:
C, 49.20; H, 4.81; N, 8.16; F 22.11.
4.3.7. L-Phenylalanine, N-L-isoleucyl, N-trifluoroacetyl,
ethyl ester. Telescoped yield: 98%. Colorless solid (from
butyl acetate/heptane, 1:5). Mp: 166–1688C; IR (cm21
,
KBr): 1741, 17.02, 1654, 1550, 1194, 707; 1H NMR
(CDCl3) d: 0.93 (m, 6H), 1.18 (m, 1H), 1.27 (t, 3H), 1.47 (m,
1H), 1.86 (m, 1H), 3.14 (t, 2H), 4.19 (q, 2H), 4.30 (t, 1H),
4.86 (q, 1H), 6.19 (d, NH), 7.08 (m, 2fH, 1NH), 7.30 (m,
3fH); 19F NMR, d: 73.28 (s, CF3); 13C NMR (CDCl3) d:
10.2, 13.8, 14.8, 24.4, 35.4, 36.6, 53.5, 57.4, 60.5, 114.8q,
126.5, 128.1, 129.1, 137.0, 157.2q, 169.7, 171.0; MS m/z
M2H2¼402. Anal. calcd for C19H25N2O4F3: C, 56.71; H,
6.26; N, 6.96; F, 14.16. Found: C, 56.78; H, 6.36; N, 6.91; F,
14.44.
4.3.3. L-Phenylalanine, N-L-methionyl, N-trifluoro-
acetyl-, ethyl ester. Yield: 92%. Colorless solid (from
butyl acetate/heptane, 1:4. Mp: 103–1048C; IR (cm21
,
1
KBr): 1736, 1705, 1649, 1538, 1378, 1209, 1183, 707; H
NMR (CDCl3) d: 1.24 (t, 3H), 2.05 (m, SCH3), 2.62 (ab,
2H), 3.18 (ab, 2H), 4.21 (q, 2H), 4.62 (q, 1H), 4.82 (q, 1H),
6.60 (d, NH), 7.16 (1H, fH), 7.17 (br, NH), 7.28(m, 4fH);
19F NMR, d: 276.33 (s, CF3); 13C NMR (DMSO-d6) d:
13.9, 14.6, 29.6, 30.9, 36.5, 52.4, 53.8, 60.6, 115.3q, 126.6,
128.2, 129.1, 137.0, 156q, 169.9, 171.2; MS m/z
4.4. General procedure for kinetic studies
Measurement of reaction rates was performed at 58C using a
Chemical/Freeze Quench Apparatus (Update Instruments
System 1000, Madison WI). Diluted reactant streams were
directed through a Wiskind grid mixer via syringe using a
Model 1019 Syringe Ram at identical, precisely controlled
flow rates. Reactions were rapidly quenched by combining
the reaction stream with a 0.25 M methanolic HCl stream in
a secondary in-line grid mixer. Altering the tube length
between reactant mixing and quench mixing in repeated
experiments allowed samples representing partially com-
pleted reactions to be collected in conjunction with precise
time measurements. Using the above apparatus, residence
time between reaction start and quench could be controlled
to as little as 20 ms. Quenched samples obtained from a
continuous flow through the system were then analyzed by
HPLC. The HPLC method involved 5 mL sample injections
onto a YMC ODS-A 4.6£50 mm eluted with a gradient
running at 75/25 0.1% aqueous phosphoric acid/acetonitrile
for 1 min, followed by a 4 min linear ramp to 15/85 0.1%
aqueous phosphoric acid/acetonitrile, and ending with a
1 min hold at this level. Flow rate was 2 mL/min and
detection was at 210 nm. Retention times for the com-
ponents of interest were 3.6 min for TFA-L-Phe, 4.6 min for
TFA-L-Phe ethyl ester, 6.1 min for TFA-L-Phe-L-Phe-OEt,
and 5.5 min for azlactone.
M2H2¼421,
MþNH24 ¼438.
Anal.
calcd
for
C18H23N2O4F3S: C, 51.42; H, 5.51; N, 6.66; F, 13.56; S,
7.63. Found: C, 51.49; H, 5.47; N, 6.65; F, 13.81; S, 7.85.
4.3.4. L-Phenylalanine, N-L-tyrosyl, N-trifluoroacetyl-,
ethyl ester. Yield: 74%. Colorless solid (from butyl
acetate/heptane, 1:2). Mp: 190–1928C; IR (cm21, KBr):
1665, 1726, 1723, 1521, 1220, 1194, 1185. 1H NMR
(CDCl3) d: 1.27 (t, 3H), 3.05 (m, 4H), 4.18 (q, 2H), 4.49
(ab, 1H), 4.63 (ab, 1H), 4.68 (s, fOH), 5.75 (ab, 1H), 6.88
(d, 2fH), 7.05 (d, 2fH), 7.12 (d, 1NH), 7.26 (s, 5H); 19F
NMR, d: 273.64 (s, CF3); 13C NMR (DMSO-d6) d: 13.9,
36.1, 36.8, 53.9, 54.8, 60.6, 114.9, 115.9q, 126.6, 127.3,
128.2, 129.1, 130.0, 137.0, 155.9, 157.2q, 170.14, 171.2;
MS m/z M2H2¼453, MþNH24 ¼470. Anal. calcd for
C22H23N2O5F3: C, 58.41; H, 5.12; N, 6.19; F, 12.60. Found:
C, 58.51; H, 5.11; N, 6.02; F, 12.36.
4.3.5. L-Phenylalanine, N-L-leucyl, N-trifluoroacetyl-,
ethyl ester. Telescoped yield: 95%. Colorless solid (from
butyl acetate/heptane, 1:4). Mp: 139–1408C; IR (cm21
,
1
KBr): 1726, 1657, 1555, 1220, 1200, 1180, 1160, 702; H
NMR (CDCl3) d: 0.96 (d, 6H), 1.18 (t, 3H), 1.61 (m, 3H)
3.18 (abq, 2H), 4.13 (q, 2H), 4.23 (q, 1H), 4.87 (q, 1H), 6.17
(br, NH), 6.83 (br, NH), 7.23 (s, 5H); 19F NMR, d: 73.42 (s,
CF3), 273.46 (s, CF3), 95:5 ratio; 13C NMR (CD3CN) d:
13.9, 23.0, 23.4, 24.6, 38.0, 42.2, 52.1, 53.3, 61.8, 127.8,
128.6, 129.2, 136.4, 156q, 172, 173; MS m/z M2H2¼403,
MþNH24 ¼420. Anal. calcd for C19H25N2O4F3: C, 56.71; H,
6.26; N, 6.96; F, 14.16. Found: C, 56.77, H, 6.40; N, 6.91; F,
14.36.
Authentic samples of the components of additional interest
to the kinetic studies were prepared as follows:
4.4.1. N-Trifluoroacetyl-L-phenylalanine methyl ester.
N-Trifluoroacetyl-L-phenylalanine (0.83 g, 3.18 mmol,
1.0 equiv.) was stirred in 8.38 mL of dry ethyl acetate at
258C. Vilsmeier reagent (490 mg, 3.83 mmol, 1.2 equiv.)
was charged and the slurry was stirred 1 h. The reaction
mixture was quenched with 50 mL of methanol. MTBE
(50 mL) was added and the mixture was extracted once with
50 mL saturated potassium carbonate. The organic phase
was concentrated to an oil and triturated with heptane to
afford white crystalline solid 0.44 g (50 M%) with com-
parable properties to material previously reported;11 mp
4.3.6. L-Phenylalanine, N-L-prolyl, N-trifluoroacetyl,
ethyl ester. Telescoped yield: 44%. Colorless solid (from
butyl acetate/heptane, 1:4). Mp: 108–1108C; IR (cm21
,
KBr): 1751, 1701, 1650, 1552, 1230, 1210, 1180, 1155, 700:
1H NMR (CDCl3) d: 1.26 (t, 3H), 1.98 (m, 2H), 2.08 (m, 1H)
3.13 (abq, 2H), 3.68 (q, 2H), 4.18 (q, 2H), 4.53 (q, 1H), 4.82
(q, 1H), 6.67 (br, NH), 7.14 (br, NH), 7.26 (s, 5H); 19F

9028
P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
50–528C; IR (KBr) 1752, 1711, 1578, 1557, 1332, 1275,
1178, 1163, 758, 707; 1H NMR (CDCl3) d: 3.19 (ab q, 2H),
3.78 (s, 3H), 4.88 (ab q, 1H), 6.91 (d, NH), 7.09 (m, 2H),
7.29 (m, 3H); 19F NMR, d: 276.4; 13C NMR (CDCl3) 37.1,
52.7, 53.5, 115.8q, 127.5, 128.7, 129.1, 134.5, 156.7, 170.4.
HPLC (45 min), the solution of acid chloride (reactant B)
was diluted to a volume of 37 mL with dry ethyl acetate.
The acid chloride solution (0.046 M) and diisopropylethyl-
amine (0.17 M, 3.7 equiv.) in ethyl acetate (reactant A) were
then used as inputs to rapid quench experiments as
described above. Remaining acid chloride at any given
time was indirectly measured as N-TFA-L-phenylalanine
methyl ester relative to azlactone as determined by HPLC.
Values of A and B at each time point were used along with
initial A and B concentrations to calculate the y axis values
used in Figure 2, from which the second-order rate constant
of 4.8 L/M s was determined.
4.4.2. 5(4H)-Oxazolone, 4-(phenylmethyl)-2-(trifluoro-
methyl). Preparation of this material (azlactone) was
analogous to that described by Weygand.12 N-Trifluoro-
acetyl-L-phenylalanine (1.3 g, 5.0 mmol, 1.0 equiv.) was
stirred in 14 mL of dry ethyl acetate at 08C. Vilsmeier
reagent (1300 mg, 10 mmol, 2.0 equiv.) was charged and
the slurry was stirred for 1 h. After reaction completion was
confirmed by HPLC, the solution was quenched into 2.6 mL
DIPEA (15 mmol, 3.0 equiv.) in 15 mL of ethyl acetate.
After stirring 5 min the reaction mixture was washed with
5 mL of 1N HCl, 5 mL water, 5 mL 0.5 M potassium
phosphate, dibasic, 5 mL 1N HCl and concentrated to afford
1.15 g (.99%) of a light yellow oil: IR (cm21, KBr): 3037,
1803, 1649; 1H NMR (CD3CN) d: 4.04 (d, 2H), 6.35 (t, 1H)
7.38 (m, 5H); 19F NMR, d: 278.37 (s, CF3); 13C NMR d:
35.1, 118.3, 122.4q, 128.4, 129.7, 130.5, 134.6, 164.8,
168.8. ms M2Hþ¼242.
4.7. Measurement of N-TFA-L-phenylalanine acid
chloride hydrolysis rate at pH 8.5
N-Trifluoroacetyl-L-phenylalanine (1.10 g, 4.2 mmol,
1.0 equiv.) was stirred in 11.8 mL of dry ethyl acetate at
08C. Vilsmeier reagent (720 mg, 5.6 mmol, 1.2 equiv.) was
charged and the slurry was stirred until complete acid
chloride formation was achieved by HPLC assay. A 100 mL
sample of the acid chloride solution (0.0035 mmol) was
charged to a vigorously stirred solution of 1 mL aqueous pH
8.5 buffer (0.07 M K2CO3, 0.175 M KHCO3) and 1 mL of
ethyl acetate. After 9 and 15 s the sample was quenched
with 1 mL of 0.1 M aqueous solution of ethyl phenylalanine
hydrochloride (0.01 mmol), stirred 1 min and diluted with
10 mL of 1:1 acetonitrile:2% aqueous phosphoric acid. The
quenched solutions were then analyzed by HPLC, showing
fractions of 28 and 42% N-TFA-L-Phe relative to N-TFA-L-
Phe-L-Phe-OEt (representive remaining acid chloride at the
time of quench). At constant pH 8.5, these data calculate to a
first-order rate constant of 0.04/s.
4.5. Measurement of acid chloride/amino ester coupling
rate
N-Trifluoroacetyl-L-phenylalanine (0.798 g, 3.05 mmol,
1.0 equiv.) was dissolved in 8.6 mL of dry ethyl acetate at
08C. Vilsmeier reagent (470 mg, 3.67 mmol, 1.2 equiv.) was
charged and the slurry was stirred. After 45 min, the
solution was diluted to a volume of 66 mL with dry ethyl
acetate (0.046 M). This acid chloride solution was then
further diluted 50-fold with dry ethyl acetate chilled to 58C
to provide a 0.00092 M solution of acid chloride (reactant
B). The amine ester solution was prepared by mixing 3.66 g
L-phenylalanine ethyl ester HCl salt (15.95 mmol) with
25 mL ethyl acetate and 25 mL 1 M K2HPO4. The phases
were separated and the aqueous was extracted twice with
25 mL ethyl acetate. The organic phases were combined and
adjusted to a volume of 100 mL with dry ethyl acetate
(0.16 M). The amine solution was then diluted 50-fold with
dry ethyl acetate and chilled to 58C to provide a 0.0035 M
amine ester solution (reactant A). The acid chloride solution
(reactant B) and amine ester in ethyl acetate (reactant A)
were then used as inputs to rapid quench experiments as
described above. Remaining acid chloride at any given time
was indirectly measured as N-TFA-L-phenylalanine methyl
ester relative to coupled product as determined by HPLC.
Values of A and B at each time point were used along with
initial A and B concentrations to calculate the y axis values
used in Figure 1, from which the second-order rate constant
of 3900 L/M s was determined.
Acknowledgements
We thank Dr John Wasylyk for the development of GC and
HPLC conditions by which to separate the dipeptide
diastereomers, and Dr C. K. Wei for training in the use of
the rapid kinetics equipment.
References
1. For a comprehensive discussion of racemization and its
prevention during peptide coupling, see Bodanszky, M.
Peptide Chemistry, A Practical Textbook; Springer: New
York, 1993; p 115.
2. (a) Steglich, W.; Hinze, S. Synthesis 1976, 399. (b) Curphy,
T. J. J. Org. Chem. 1979, 44(15), 2805.
3. For example: (a) Jones, J. H. The Peptides; Gross, E.,
Meienhofer, J., Eds.; Academic: New York, 1979; Vol. 1,
p 65. (b) Bodanszky, M. Peptide Chemistry, A Practical
Textbook; Springer: New York, 1993; p 55.
4.6. Measurement of rate of N-TFA-L-Phe acid chloride
reaction with diisoproplyethylamine (azlactone
formation)
4. (a) Norlander, J. E.; Njorge, F. G.; Payne, M. J.; Warman, D.
J. Org. Chem. 1985, 50, 3481. (b) Varie, D. L. Tetrahedron
Lett. 1990, 31(52), 7583.
TFA protected phenylalanine (0.45 g, 1.72 mmol,
1.0 equiv.) was stirred in 4.9 mL of dry ethyl acetate at
08C, 270 mg of Vilsmeier reagent (2.10 mmol, 1.0 equiv.)
was charged, and the slurry was stirred. After the acid
chloride formation reached completion as detected by
5. Carpino, L. A.; Cohen, B. J.; Stephens, K. E.; Sadat-Aalaee,
Y.; Tien, J.; Langridge, D. C. J. Org. Chem. 1986, 51, 3732.
6. Souter, R. J. Chromatogr. 1975, 108(2), 265–274.

P. A. Jass et al. / Tetrahedron 59 (2003) 9019–9029
9029
7. Schallenberg, E.; Calvin, M. J. Am. Chem. Soc. 1955, 77,
2779–2782.
10. All preparations have been previously described in the
chemical literature, see Ref. 2.
8. Meienhofer, J. H. The Peptides; Gross, E., Meienhofer, J.,
Eds.; Academic: New York, 1979; Vol. 1, pp 270–283.
9. Greene, T. W.; Wuts, P. G. M. Protective Groups in Organic
Synthesis; 3rd ed. Wiley: New York, 1999; pp 556–557.
11. Sarnoff, H. A.; Karmen, A. Anal. Biochem. 1960, 1, 344–350.
12. Weygand, F.; Steglich, W.; Tanner, H. Liebig’s Ann. Chem.
1962, 658, 128–150.