Solid State Stability Studies of Model Dipeptides: Aspartame and Aspartylphenylalanine

Article abstract of DOI:10.1021/js960228d

Some solid-state pharmaceutical properties and the solid-state thermal stability of the model dipeptides aspartame (APM) and aspartylphenylalanine (AP), have been investigated, Studies by differential scanning calorimetry (DSC), thermal gravimetric analysis (TGA), high-performance liquid chromatography, powder X-ray diffraction, and optical microscopy have shown that the dipeptides undergo solid state intramolecular aminolysis of the type, solid → solid + gas. This reaction was observed for APM at 167-180°C with the liberation of methanol and for AP at 186-202°C with the liberation of water. The exclusive solid product of the degradation reaction of both dipeptides is the cyclic compound 3-(carboxymethyl)-6-benzyl-2,5-dioxopiperazine, The rates of the degradation reactions were monitored by isothermal TGA and by temperature-ramp DSC and were found to follow kinetics based on nucleation control with activation energies of about 266 kJ mol^-1 for APM and 234 kJ mol^-1 for AP.

Full text of DOI:10.1021/js960228d

+
+
Solid State Stability Studies of Model Dipeptides: Aspartame and
Aspartylphenylalanine
‡
S
UZANNE S. LEUNG AND
D
AVID J. W. GRANTX
Received May 23, 1996, from the Department of Pharmaceutics, College of Pharmacy, University of Minnesota, Weaver-Densford Hall,
308 Harvard St. S.E., Minneapolis, MN 55455-0343. Final revised manuscript received July 23, 1996. Accepted for
publication July 25, 1996X. ‡Present address: 3M Pharmaceuticals, 270-4S-17, 3M Center, St. Paul, MN 55144.
chemistry of small peptides. A preliminary report of this work
has been recently presented.3
Abstract
0 Some solid-state pharmaceutical properties and the solid-
state thermal stability of the model dipeptides aspartame (APM) and
aspartylphenylalanine (AP), have been investigated. Studies by differential
scanning calorimetry (DSC), thermal gravimetric analysis (TGA), high-
performance liquid chromatography, powder X-ray diffraction, and optical
microscopy have shown that the dipeptides undergo solid state intramo-
lecular aminolysis of the type, solid
observed for APM at 167 180 C with the liberation of methanol and for
AP at 186 202 C with the liberation of water. The exclusive solid product
Materials and Methods
Ma ter ia lssAPM, AP, and 3-(carboxymethyl)-6-benzyl-2,5-diox-
opiperazine (DKP) were gifts from the NutraSweet Company, Mount
Prospect, IL. Sodium heptanesulfonate was obtained from Eastman
Kodak Chemical Co., Rochester, NY. Sodium phosphate monobasic,
phosphoric acid, and 1,1,1-trichloroethane (all ACS grade) were
obtained from Fisher Scientific, Fairlawn, NJ , as were water and
acetonitrile (both HPLC grade). All materials were used without
further modification or purification.
Differ en tia l Sca n n in g Ca lor im etr y (DSC)sThe DSC curves
were determined using a DuPont 910 differential scanning calorimeter
(TA Instruments, New Castle, DE) equipped with a data station
(Thermal Analyst 2000, TA Instruments, New Castle, DE). The cell
constant was calibrated using indium. Samples (2.50-2.70 mg) in
nonhermetically crimped or open aluminum pans were heated at 0.5,
1, 2, 4, 7, and 10 °C/min under nitrogen purge at 3-4 mL/min. The
peak temperature was noted as the point on the temperature scale
corresponding to maximum deviation from the baseline.
Th er m ogr a vim etr ic An a lysis (TGA)sThe TGA curves were
obtained using a DuPont 951 thermogravimetric analyzer (TA Instru-
ments, New Castle, DE) linked to a data station (Thermal Analyst
2000, TA Instruments, New Castle, DE). All TGA runs were
performed on samples in open aluminum pans with a nitrogen purge
at 3-4 mL/min. Nonisothermal TGA was performed on samples
(2.30-2.50 mg) at a heating rate of 10 °C/min. Isothermal TGA was
performed on APM samples (2.20-2.50 mg) at 166.9, 170.4, 175.5,
and 180.3 °C and for AP samples (2.20-2.50 mg) at 186.2, 191.9,
196.5, and 202.0 °C.
Ka r l F isch er Titr im etr ysThe water content of APM and AP
samples was determined by Karl Fischer titrimetry using a Mitsubishi
Moisture Meter (Model CA-05, Mitsubishi Chemical Industries Ltd.,
Tokyo, J apan).
P a r ticle Size An a lysissPowder samples of APM and AP were
suspended in 1,1,1-trichloroethane by sonication and analyzed in a
particle size analyzer (Brinkmann 2010, Westbury, NY) with me-
chanical stirring.
Su r fa ce Ar ea Mea su r em en tssThe specific surface area of
powder samples of APM and AP particles was determined using
multipoint nitrogen adsorption by the dynamic volumetric BET
method (Gemini 2360, Micromeritics, Norcross, GA).
Den sitysThe density of the samples was determined by a helium-
air pycnometer (Autopycnometer 1320, Micromeritics, Norcross, GA).
P ow d er X-r a y Diffr a ctom etr y (P XRD)sThe powder X-ray
diffraction patterns of APM and AP samples were determined at
ambient temperature using an X-ray powder diffractometer (Siemens
D-500, Germany) at 30 mA and 45 kV with Cu KR radiation. Counts
were measured using a scintillation counter. Samples were packed
into an aluminum holder and scanned with the diffraction angle, 2θ,
increasing from 5° to 35°, with a step size of 0.02° and a counting
time of 1 s.
High -P er for m a n ce Liqu id Ch r om a togr a p h y (HP LC)sHPLC
was used to separate and identify APM, AP, and their degradation
products according to the method of Stamp and Labuza.4 For the
present study, the HPLC system (Shimadzu Scientific, Chicago, IL)
consisted of a liquid chromatograph (LC-6A), an autoinjector (SIL-
6A), a system controller (SCL-6A), a UV detector (SPD-6AV), and a
printer (Chromatopac CR-5A). The column (Nova-pak C18) and guard
column (Guardpak C18) were obtained from Waters Associates,
f solid + gas. This reaction was
−
°
−
°
of the degradation reaction of both dipeptides is the cyclic compound
3-(carboxymethyl)-6-benzyl-2,5-dioxopiperazine. The rates of the deg-
radation reactions were monitored by isothermal TGA and by temperature-
ramp DSC and were found to follow kinetics based on nucleation control
with activation energies of about 266 kJ mol 1 for APM and 234 kJ mol
-
-1
for AP.‡
Introduction
While the solution reactions of small peptides have been
quite well-characterized (e.g., Scheme 1),1 their reactions in
the solid state have attracted scant attention. With the
increasing use of small peptides as pharmaceuticals, it is
important to understand the solid-state behavior of this
important class of compound. In the present study, aspartame
(APM) and aspartylphenylalanine (AP) are used as models
for small peptides. These dipeptides are suitable candidates
for solid-state studies because of their relatively low cost. APM
(L-aspartyl-L-phenylalanine methyl ester), a dipeptide sweet-
ener, is finding increasing use in foods, beverages, and
pharmaceuticals. The present report describes the physical
properties of solid APM and solid AP and evaluates their
chemical stabilities.
Solid-state reactions may be classified as physical trans-
formations or chemical reactions.2 Physical transformations
include polymorphic transitions and desolvations and are
characterized by changes in the crystal structure without
modifications of the component molecules. Chemical reactions
of the solid state include rearrangements, photochemical
reactions, and decompositions. Physical transformations of
aspartame will be discussed in a later publication.
Thorough solid-state stability studies of aspartame have not
been previously reported. Lack of solid-state stability results
in loss of the active ingredient. In addition, the presence of
the reaction product as an impurity may alter the properties
of the solid. The major labile bonds in aspartame are the
amide and the ester bonds. In aqueous solution, aspartame
degradation occurs via multiple pathways, as shown in
Scheme 1.1 Without the methoxy group, the bond that is most
sensitive to hydrolysis in AP is the amide bond. The present
evaluation of the stability of APM and AP in the solid state
may contribute to a general understanding of the solid-state
X Abstract published in Advance ACS Abstracts, September 1, 1996.
64 / Journal of Pharmaceutical Sciences
S0022-3549(96)00228-6 CCC: $14.00
© 1997, American Chemical Society and
American Pharmaceutical Association
Vol. 86, No. 1, January 1997

+
+
Scheme 1sPathways of APM degradation in aqueous solution. The ionic charges, which depend on pH, are not shown.
Table 1
s Physical Properties of Aspartame (APM) and
Milford, MA. The mobile phase consisted of 4 mM sodium heptane-
sulfonate and 4 mM sodium phosphate monobasic in 20% acetonitrile
in water, adjusted to pH 3.0 with phosphoric acid. The flow rate was
1.0 mL/min and the detection wavelength was 214 nm.
Aspartylphenylalanine (AP) as Received
APMa
APa
Hot Sta ge Micr oscop ysOptical micrographs were obtained with
a hot stage (Mettler FP 80, Mettler Instrument Corporation, Hight-
stown, NJ ) mounted on an optical microscope (Wild M3Z, Wild
Heerbrugg, Heerbrugg, Switzerland) equipped with photoautomat
(Wild MPS45) and a 35 mm camera attachment. Crystals were
immersed in high boiling silicone oil (Aldrich Chemical Co., Milwau-
kee, WI) and heated at a rate of 3 °C/min.
Molecu la r Mod elin gsThe program Cerius2 (Molecular Simula-
tions, Burlington, MA) was run on Silicon Graphics Indigo hardware
to predict the low-energy conformers of APM. Conformers were
searched via a grid scan with a step size of 30°, varying the torsion
angles, ψ, æ, and ø, of APM. Because of the size of the calculation,
the torsional angles of the aspartyl and the phenylalanyl residues
were varied separately. Each conformer was minimized and the
lowest energy conformer was used to vary the remaining torsional
angles. All energy calculations employed the universal force field5
while charges were calculated using the charge equilibration method,6
both available on Cerius2.
Appearance
White
crystalline powder
Acicular
White
crystalline powder
Acicular
Habit
Density (g/cm3)
Specific surface area (m2/g)
1.49
2.29
±
±
0.6
0.04
0.07
1.17
3.41
±
±
0.04
0.10
Particle size (
Water Content (% w/w)c
After heating to 150
µ
m)b
8.3
±
7.9
0.162
±
0.5
0.045
3.76
0.33
±
±
0.05
0.05
±
°
C
a Mean 3). b Particle size followed a log-normal relationship with
SD (n )
±
the number-length mean diameter stated. c Water content determined by Karl
Fischer titrimetry.
The water content of APM samples heated to 140 °C and
cooled to ambient was 0.33 ( 0.05%. The second DSC
endotherm at 189 °C and the second TGA weight loss step at
about 196 °C require more detailed study, presented and
discussed below. The third DSC endotherm at 248 °C was
not accompanied by a TGA weight loss step and corresponds
to the reported melting point of 248 °C.7
Results and Discussion
A sample of APM crystals was placed in silicone oil and
heated at 3 °C/min under a hot stage microscope. Vapor was
released as bubbles from the needle-shaped crystals at ap-
proximately 110 °C, consistent with desolvation at the first
endotherm. However, the crystals also released a vapor at
approximately 190 °C (Figure 2b), corresponding to the second
weight loss step in TGA at 196 °C and to the second DSC
endotherm at 189 °C. The crystals did not disintegrate or
change shape even under vigorous gas evolution. However,
the crystals darkened after the second stage of gas evolution
at 190 °C (Figure 2). Formation of liquid was not observed
until 248 °C, the reported melting point of APM.7
Sta bility of Asp a r ta m esThe physical properties of APM,
as received, are listed in Table 1. The total amount of water
in APM samples, determined by Karl Fischer titrimetry, was
higher than that for the phase-pure hemihydrate (2.97%
water), suggesting that the samples contained sorbed water.
The particle size number distribution followed a log-normal
relationship with a number-length mean diameter of 8.3 (
1.1 µm.
DSC of APM in crimped pans gave three endotherms
(Figure 1a). TGA with the same temperature program (10
°C/min) gave two weight loss steps (Figure 1b), which cor-
respond to the first two DSC endotherms. The first DSC
endotherm at 129 °C and the first TGA weight loss step at
about 114 °C are attributed to loss of water, i.e. dehydration
of the crystal lattice, as indicated by Karl Fischer titration.
Since APM readily degrades in aqueous solution, its deg-
radation in the solid state was investigated as the possible
origin of the endotherm, weight loss, and release of vapor at
about 190 °C. In solution, the major degradation products of
Journal of Pharmaceutical Sciences / 65
Vol. 86, No. 1, January 1997

+
+
Figure 1
s
DSC and TGA curves for APM: (a) DSC curve corresponding to the left axis and (b) weight loss by TGA corresponding to the right axis. The change from
23 to 150
°
C corresponds to the dehydration of APM, which is the subject of a subsequent report.8
APM are AP and DKP. AP is formed by direct hydrolysis of
the ester linkage, while DKP is formed by intramolecular
aminolysis. In the solid state either reaction would result in
the release of methanol with a theoretical weight loss of
10.55%, close to the observed weight loss of 11.13 ( 0.59% in
TGA (Figure 1b). This close comparison suggests that heating
APM in the solid state results in degradation with the
liberation of methanol.
The possible formation of AP and/or DKP on heating solid
APM was investigated by HPLC and PXRD. For this purpose,
the HPLC chromatogram in Figure 3a shows the separation
of the pure compounds, APM (the starting material), and AP
and DKP (the presumed degradation products), while parts
a and b of Figure 4 show the quite distinct PXRD patterns of
APM and DKP, respectively. The effects of heating APM to
the different temperatures were compared. A 2.6 mg sample
of APM was heated in the DSC to 150 °C, beyond the first
DSC endotherm and TGA step, which are attributed to
dehydration, but was removed before the second endotherm
and TGA step (Figure 1) and then cooled and analyzed by
HPLC. The product gave an HPLC chromatogram (Figure
3b) the peak of which corresponded to pure unchanged APM
(retention time of 7.16 min). The PXRD pattern of APM at
150 °C corresponded to that of anhydrous APM and was
different from that of the starting material.8
Another 2.6 mg sample was heated in the DSC to 200 °C,
beyond the second DSC endotherm and TGA step but before
the third endotherm (Figure 1), and was similarly analyzed.
The product gave an HPLC chromatogram (Figure 3c) with a
single peak corresponding to pure DKP, with no detectable
traces of other degradation products nor of the starting
material, APM. The PXRD pattern of the product (Figure 4c)
showed that it was crystallographically identical to DKP
(Figure 4b) and that diffraction peaks due to APM were
absent. However, Figure 4c shows broad background scat-
tering (amorphous halo) centered around 20° 2θ, suggesting
that the DKP formed by the solid-state degradation of APM
was less crystalline than the DKP standard. This lower
crystallinity is manifested in a DSC melting point of 248 °C
(Figure 1a), below 259 ( 1.1 °C for the pure crystalline
standard DKP. DKP decomposes on melting.
To summarize the results from DSC and TGA (Figure 1),
the first DSC endotherm and TGA step are attributed to
dehydration of the APM crystals. The hydration and dehy-
dration behaviors of APM are the subject of another report.
The second DSC endotherm and TGA step are attributed to
the solid-state decomposition of APM with gas evolution due
to loss of methanol according to Scheme 3. The third DSC
endotherm at 248 °C (Figure 1), which has been attributed to
melting of APM,7 actually corresponds to the melting of DKP
that results from the degradation of solid APM.
Intramolecular cyclization is the major pathway of degrada-
tion of APM in solution. However, cyclization was not
expected to be the major pathway of degradation in the solid
state because the large conformational change necessary to
bring the nucleophilic amine in proximity to the ester carbonyl
would presumably require room which, at first sight, would
appear to be lacking in the solid state due the packing of the
molecule in the crystal lattice.
The path of a nucleophile attacking a carbonyl carbon is
described by the Bu¨rgi-Dunitz trajectory.9 By reference to
crystal structures of compounds containing nucleophilic and
electrophilic groups, the reaction coordinate for nucleophilic
addition was mapped. The N, C, and the O atoms lie in a
mirror plane, while the C atom deviates from the plane
defined by R, R′, and O (Scheme 2).10 The extent of deviation
increases as the C‚‚‚N distance, d, decreases. As the nitrogen
atom approaches the carbon atom, the C-O bond lengthens
and the alkyl substituents bend away. The nitrogen atom
approaches at an angle, R, of 107° with little variation between
the different crystal structures. The value of R is similar for
other nucleophilic additions. The N‚‚‚C-R angle ranges from
82 to 120°. If this reaction coordinate can be applied to APM,
the molecule must undergo considerable rearrangement in
order to bring the nucleophilic nitrogen along this path.
Although a crystal structure of APM hemihydrate has been
determined,11 it does not correspond to the commercially
available polymorph8 and one can only speculate on the
66 / Journal of Pharmaceutical Sciences
Vol. 86, No. 1, January 1997

+
+
Figure 3
s
HPLC chromatograms of (a) prepared mixtures of the pure compounds
APM, AP and DKP, (b) the product obtained by heating APM to 150
product obtained by heating APM to 200
heating AP to 220 C.
°
C, (c) the
°
C, and (d) the product obtained by
°
Figure 4
s
Powder X-ray diffraction patterns of (a) APM, (b) DKP, and (c) the
solid product obtained by heating APM to 200
°
C.
Figure 2
s
Optical micrographs of APM crystals at various temperatures: (a) at
150 C, (b) at 190
°
°
C, and (c) at 200 C.
°
structure of the commercially available form. However, most
dipeptides exist in their extended conformations in the solid
state with the side chains arranged in an antiparallel fash-
ion,12 and the commercially available polymorph of APM may
be no exception.
The flexibility of the APM molecule may explain the
occurrence of intramolecular cyclization. A search for low-
energy conformers of APM was performed by varying the
torsional angles æ and ø of the aspartyl residue and the
torsional angles ψ, æ, and ø of the phenylalanyl residue at
Scheme 2sThe trajectory of a nucleophilic nitrogen in an addition reaction.
30° increments using the grid scan method. The resulting
conformers were minimized and the five lowest energy
conformers were calculated and are overlaid in Figure 5. The
low-energy conformers exist in similar conformations in the
aspartyl residue. Side chain conformation is governed by the
electrostatic interaction between the positively charged pro-
tonated amino group and the negatively charged γ-carboxy-
Journal of Pharmaceutical Sciences / 67
Vol. 86, No. 1, January 1997

+
+
Scheme 3sSolid state intramolecular aminolysis reactions of APM and AP leading to solid DKP.
hydrolysis of the ester bond to give aspartylphenylalanine,
would presumably require less rearrangement within the
crystal lattice. DKP formation appears to contradict the
topochemical postulate that reactions in the solid state proceed
with a minimum of molecular motion. The correlation of
reactant and product geometries in many solid-state reactions
is the basis of this theory. The similar geometries of the
reactant and product arise from the confinement of the
molecular rearrangements to a cavity defined by the crystal
lattice. Examples of non-topochemical reactions include
photochemical polymerizations of vinyl monomers.15 Although
large molecular motions in the lattice are hindered at room
temperature, at the elevated temperatures at which intramo-
lecular aminolysis occurs, molecules in the solid may acquire
sufficient energy to overcome the steric barriers to rearrange-
ment.
In aqueous solution the abundance of water allows the
hydrolysis of APM, to form AP and PM, to compete with DKP
formation. However, in the solid-state, at the temperature
of the solid state reaction (about 190 °C), desolvation of the
lattice water has already occurred and adsorbed water would
likewise have evaporated. Although intramolecular cycliza-
tion would seem unlikely, it might happen because other
pathways that require water for hydrolysis are not available.
Sta bility of Asp a r tylp h en yla la n in esTo look into the
possible uniqueness or otherwise of intramolecular aminolysis
as a mechanism of degradation, the solid-state stability of AP
was evaluated. While AP is a product of the degradation of
APM in solution, it is also a dipeptide that may be capable of
undergoing intramolecular cyclization. Some physical proper-
ties of AP were determined and are listed in Table 1. As with
APM, the particle size number distribution of AP follows a
log-normal relationship.
The DSC profile of AP in Figure 6a shows two endotherms,
the first at 211 °C and the second at 240 °C. These thermal
events were not baseline-separated. The corresponding TGA
plot in Figure 6b shows weight loss at 217 °C, corresponding
closely to the first endotherm. Under hot stage microscopy,
the solid to liquid transition began at 225 °C. The barely
measurable amount of water in the AP samples, determined
by Karl Fischer titrimetry, and the lack of a dehydration
endotherm at lower temperatures indicate the absence of
lattice water in AP.
Figure 5sThe five lowest energy conformers of APM are overlaid.
late. The phenylalanyl residue showed significant variation
in conformation among the different conformers. This result
is similar to that for hydrated APM reported by Kang,13 who
selected low-energy conformations of APM analogs as starting
points for minimization, while no such assumption was made
here. Variations in the conformation of the phenylalanyl
residue was also found in the crystal structure of [(APM)2-
Cu]‚8/3H2O.8,14 The large differences between the low-energy
conformers suggest that APM is very flexible. The geometry
of the five lowest energy conformers is not far from the Bu¨rgi-
Dunitz trajectory.10 The N‚‚‚CdO angles range from 110.5°
to 125.8° and the N‚‚‚C-ÃMe angles range from 85.2° to
109.9°, close to 107° and 82°-120° respectively, the values
obtained by Bu¨rgi.10 Although the conformation of APM in
the solid state may differ from the low-energy conformations
found in the conformer search, at the elevated temperature
of the reaction, thermal disruption of the lattice may allow
the APM molecules to change their conformation. Conforma-
tional changes to a lower energy state may actually bring the
attacking nucleophile along the reaction coordinate for cy-
clization.
The two DSC endotherms observed with AP are in the same
range as the second and third endotherms of APM, as if an
analogous reaction were occurring with AP. In the case of
AP, intramolecular attack to give DKP would release water
instead of methanol. The weight loss observed by TGA, 6.44
The alternative pathways to DKP formation, hydrolysis of
the amide bond to give phenylalanine methyl ester and direct
68 / Journal of Pharmaceutical Sciences
Vol. 86, No. 1, January 1997

+
+
Figure 6sDSC and TGA curves for AP: (a) DSC curve corresponding to the left axis and (b) weight loss by TGA corresponding to the right axis.
( 0.33% (Figure 6b), compares favorably with the theoretical
weight loss of 6.42% for liberation of 1 mol equiv of water from
AP. Under hot stage optical microscopy, the evolution of vapor
was seen at 210-220 °C.
The formation of DKP was confirmed by HPLC analysis of
heated AP samples (Figure 3). At 220 °C, a temperature
beyond the first and before the second endotherm, the sample
of AP showed complete conversion to DKP, proving that the
endotherm at 211 °C is caused by the intramolecular cycliza-
tion of AP to form DKP, with release of water. The temper-
ature at which this reaction occurs for AP (about 211 °C) is
higher than for APM (about 190 °C). As DKP is formed, it
begins to melt, corresponding to the second endotherm.
Scheme 3 shows the reaction for both APM and AP. For AP,
the alternative reaction to DKP formation is hydrolysis of the
amide bond. As in the case of APM, this hydrolysis reaction
is unlikely to occur due to the lack of water at this elevated
temperature.
Kin etics of Cycliza tion of Asp a r ta m e a n d Asp a r tyl-
p h en yla la n in esThe reaction was studied by isothermal TGA
at various temperatures. The fraction reacted, x, is plotted
against time for APM in Figure 7. The shape of the curves is
sigmoidal, as is common for solid-state decompositions.16 The
curves are characterized by an induction period at low x
values, a growth period with an inflection at intermediate
values of x, and a decay period at high x values.
Figure 7
s
Plots of fraction reacted against time for the intramolecular aminolysis
of (a) APM and (b) AP.
The reaction curves can be modeled by rate equations which
assume various mechanisms. For APM degradation, the
Prout-Tompkins equation17 was found to provide the best fit
to the kinetic profile. The kinetic data for the cyclization of
APM at 0.10 e x e 0.90 were fitted with a correlation
coefficient, R ) 0.9946. The Avrami-Erofeev equation18
provides a fit (R ) 0.9930) that is almost as close. The values
of x < 0.10 and x > 0.90 are subject to variability due to the
slow rate of conversion at the induction and decay periods.
The residuals are randomly distributed and no trends were
seen for either model. For AP degradation, the Prout-
Tompkins and Avrami-Erofeev equations provide similar fits,
with R ) 0.9994 for both models; the residuals were randomly
distributed and no trends were seen for either model, as for
APM degradation. The activation energies, evaluated by
graphical solution of the Arrhenius equation using the rate
constant calculated from the Prout-Tompkins equation (Fig-
ure 8), were 268 ( 8 kJ mol-1 for APM and 242 ( 8 kJ mol-1
for AP.
The activation energies were calculated independently
using temperature-ramp DSC via the Kissinger equation (eq
1),19 which relates the heating rate, φ (in degrees Celsius/
minute), temperature at peak maximum, Tm (in degrees
Kelvin), and activation energy, Ea. The activation energies
for cyclization were obtained from the slope of a plot of ln(φ/
Journal of Pharmaceutical Sciences / 69
Vol. 86, No. 1, January 1997

+
+
Table 2
s
Summary of the Thermodynamic and Kinetic Parameters for the Intramolecular Aminolysis of APM and AP and the Thermodynamic Data for
Fusion of the Solid Product, DKP
APMa
APa
DKPa,b
Reaction temperature (
Heat of reaction (J/g)d
Melting point of solid product (
°
C)c
188.9
134.1
248.0
121.4
±
±
±
±
6
0.4
3.1
1.1
6.1
209.5
198.5
239.4
±
±
±
0.8
5.3
1.4
°
C)
259.1
193.0
±
±
1.1
15.5
Heat of fusion of solid product (J/g)
78.1
227
242
± 2.6
±
±
1
e
Ea by DSC (kJ mol-
)
265
268
±
±
8
8
1
f
Ea by TGA (kJ mol-
)
8
Kinetic mechanism of best fit
Nucleation control
Nucleation control
a Mean value 3. b The pure crystalline standard of the same chemical composition as the aminolysis product. c The temperature of maximum deflection
SD, n
±
)
of the aminolysis endotherm in DSC at the scan rate of 10
°
C/min. d Determined from the area of the aminolysis endotherm. e Activation energy from slope of Figure
SD. f Activation energy from slope of Figure 8, mean
SD.
9, mean
±
±
Figure 9s φ, K/min) in linear
Kissinger plot20 analyzing the effect of heating rate (
Figure 8
s
Arrhenius plot of the rate constants obtained for the intramolecular
) 268 ± ) 242 ± 8 kJ
8 kJ mol-1, and (b) AP, Ea
temperature-ramp DSC on the peak maximum temperature, Tm of the intramolecular
aminolysis of (a) APM, Ea
) 265 ± ) 227 ± 8 kJ
6 kJ mol-1, and (b) AP, Ea
mol-
.
1
aminolysis of (a) APM, Ea
1
mol-
.
Tm2) versus 1/Tm (Figure 9) and were 265 ( 6 kJ mol-1 for
APM and 227 ( 6 kJ mol-1 for AP.
of perfect crystal packing. Once the product is formed, it
becomes a source of strain in the crystal lattice and then acts
as a new nucleus, disrupting the molecular interactions of the
neighboring molecules and causing them to react. This chain
reaction is propagated from one molecule to the next until it
is terminated when and where another product molecule is
encountered. This mechanism is common among non-to-
pochemically controlled reactions.15
2
d(ln φ/Tm
d(1/Tm)
)
Ea
R
) -
(1)
By this method, the calculated Ea values agree well with
those determined by isothermal TGA. The activation energy
for DKP formation from APM in the solid state is comparable
with that for other solid-state chemical reactions.20 The
activation energy for the solid-state reaction is much greater
than that for the solution reaction (70.0 kJ mol-1)1 reflecting
the greater energy requirement for rearrangement in the
confines of a crystal lattice. The high activation energy
confirms that the reaction is a true solid-state reaction of the
type, solid f solid + gas.
Both the Prout-Tompkins equation and the Avrami-
Erofeev equation assume that the reaction is controlled by
the growth and propagation of high-energy sites or nuclei.21,22
Nucleation can occur at these high-energy sites, such as
crystal defects and surfaces. The activation energy needed
for reaction at these sites of disorder is lower than in regions
Table 2 summarizes the parameters of both degradation
reactions studied. AP decomposes at a slightly higher tem-
perature than APM and with a higher enthalpy of reaction.
The melting point and enthalpy of fusion of the resulting DKP
are significantly lower than those of the pure crystalline
material. The activation energy determined by the two
methods, isothermal TGA and temperature-ramp DSC, agree
well for each compound. The kinetic equations that best fit
both systems were those derived assuming nucleation control,
although the fact that the kinetic data follow a certain rate
equation does not prove that the associated mechanism is
operating.
70 / Journal of Pharmaceutical Sciences
Vol. 86, No. 1, January 1997

+
+
6. Rappe´, A.; Goddard, W. Charge equilibration for molecular
Conclusions
dynamics simulations. J . Phys. Chem. 1991, 95, 3358-3363.
Some physical properties of the model dipeptides APM and
7. Budavari, S.; O’Neil, M.; Smith, A.; Heckelman, P. Eds. The
Merck Index, Eleventh Edition, Merck & Co., Rahway, NJ , 1989,
p 132.
8. Leung, S. S. Stability Studies of a Model Dipeptide, Aspartame.
Ph.D. thesis, University of Minnesota, Minneapolis MN, 1996.
AP have been evaluated and their solid-state thermal stability
determined. These dipeptides undergo solid-state intramo-
lecular aminolysis at about 190 °C for APM, with the elimina-
tion of methanol, and at about 210 °C for AP, with the
elimination of water, to form the cyclic compound 3-carboxy-
methyl-6-benzyl-2,5-dioxopiperazine as the exclusive solid
product. For APM, the low-energy conformers appear to
possess the correct geometry for cyclization. The degradation
reactions for these two dipeptides were found to follow kinetics
based on nucleation control with activation energies of about
266 kJ mol-1 for APM and about 234 kJ mol-1 for AP.
9. Burgi, H.; Dunitz, J . From crystal statics to chemical dynamics.
Acc. Chem. Res. 1983, 16, 153-161.
10. Burgi, H.; Dunitz, J .; Shefter, E. Geometrical reaction coordi-
nates. II. Nucleophilic addition to a carbonyl group. J . Am.
Chem. Soc. 1973, 95, 5065-5067.
11. Hatada, M.; J ancarik, J .; Graves, B.; Kim, S.-H. Crystal
structure of APM, a peptide sweetener J . Am. Chem. Soc. 1985,
107, 4279-4282.
12. Go¨rbitz, C.; Etter, M. Hydrogen bond connectivity patterns and
hydrophobic interactions in crystal structures of small, acyclic
peptides. Int. J . Pept. Protein Res. 1992, 39, 93-110.
Acknowledgments
13. Kang, Y. Conformation and hydration of APM. Int. J . Pept.
We thank the American Foundation for Pharmaceutical Education
and the Pharmaceutical Research and Manufacturers of America
Foundation, Inc., for consecutive predoctoral fellowships for S.S.L.,
the Micromeritics Instrument Corporation, Norcross, GA, for generous
gifts of the Gemini 2360 surface area analyzer and the VapPrep 061,
and the NutraSweet Company, Mount Prospect, IL, for gifts of APM,
AP, and DKP.
Protein Res. 1991, 38, 79-83.
14. Ojala, W. H.; Leung, S. S.; Grant, D. J . W.; Gleason, W. B.
Crystal structure of bis(aspartylphenylalanine methyl ester)
copper (II). In preparation.
15. Wright, J . Molecular crystals, 2nd ed.; Cambridge University
Press: Cambridge, UK, 1995; pp 50-70, 120-141.
16. Monkhouse, D.; Van Campen, L. Solid state reactions - theoreti-
cal and experimental aspects. Drug Devel. Ind. Pharm. 1984,
10, 1175-1276.
References and Notes
17. Prout, E.; Tompkins, F. The thermal decomposition of potassium
1. Stamp, J . Kinetics and Analysis of APM Decomposition Mech-
anisms in Aqueous Solutions Using Multiresponse Methods.
Ph.D. thesis, University of Minnesota, Minneapolis MN, 1990.
2. a. Byrn, S. Solid state organic chemistry and drug stability.
Annu. Rep. Med. Chem. 1985, 20, 287-294. b. Byrn, S. Solid-
State Chemistry of Drugs; Academic Press: New York, NY, 1982;
pp 59-74, 149-186.
3. Leung, S. S.; Grant, D. J . W. Solid state degradation of a model
dipeptide: aspartame. Pharm. Res. 1995, 12, S-230 (PDD 7150).
4. Stamp, J .; Labuza, T. An ion-pair high performance liquid
chromatographic method for the determination of APM and its
decomposition products. J . Food Sci. 1989, 54 (4), 1043-1046.
5. Rappe´, A.; Casewit, C.; Colwell, K.; Goddard, W.; Skiff, W. UFF,
a full periodic table force field for molecular mechanics and
molecular dynamics simulations. J . Am. Chem. Soc. 1992, 114,
10024-10035.
permanganate. Trans. Faraday Soc. 1944, 40, 488-498.
18. Avrami, M. Granulation, phase change, and microstructure;
kinetics of phase change. III. J . Chem. Phys. 1941, 9, 177-184.
19. Carstensen, J . Drug Stability; Marcel Dekker: New York, NY,
1990; p 132.
20. Kissinger, H. Reaction kinetics in differential thermal analysis,
Anal. Chem. 1957, 29, 1702-1706.
21. Mullin, J . Crystallization, 3rd ed.; Butterworth-Heinemann:
Oxford, UK, 1993; pp 24-27, 104-105.
22. Garner, W. Kinetics of endothermic solid reactions. In Chemistry
of the Solid State; Garner, W., Ed.; Academic Press: New York,
NY, 1955; pp 232-253.
J S960228D
Journal of Pharmaceutical Sciences / 71
Vol. 86, No. 1, January 1997