Chiral inversion and hydrolysis of thalidomide: Mechanisms and catalysis by bases and serum albumin, and chiral stability of teratogenic metabolites

Article abstract of DOI:10.1021/tx9801817

The chiral inversion and hydrolysis of thalidomide and the catalysis by bases and human serum albumin were investigated by using a stereoselective HPLC assay. Chiral inversion was catalyzed by albumin, hydroxyl ions, phosphate, and amino acids. Basic amino acids (Arg and Lys) had a superior potency in cataLyzing chiral inversion compared to acid and neutral ones. The chiral inversion of thalidomide is thus subject to Specific and general base catalysis, and it is suggested that the ability of HSA to catalyze the reaction is due to the basic groups of the amino acids Arg and Lys and not to a single catalytic site on the macromolecule. The hydrolysis of thalidomide was also base-catalyzed. However, albumin had no effect on hydrolysis, and there was no difference between the catalytic potencies of acidic, neutral, and basic amino acids. This may be explained by different reaction mechanisms of the chiral inversion and hydrolysis of thalidomide. Chiral inversion is deduced to occur by electrophilic substitution involving specific and general base catalysis, whereas hydrolysis is thought to occur by nucleophilic substitution involving specific and general base as well as nucleophilic catalysis. As nucleophilic attack is sensitive to steric properties of the catalyst, steric hindrance might be the reason albumin is not able to catalyze hydrolysis. ¹H NMR experiments revealed that the three teratogenic metabolites of thalidomide, in sharp contrast to the drug itself, had complete chiral stability. This leads to the speculation that, were some enantioselectivity to exist in the teratogenicity of thalidomide, it could result from fast hydrolysis to chirally stable teratogenic metabolites.

Full text of DOI:10.1021/tx9801817

Chem. Res. Toxicol. 1998, 11, 1521-1528
1521
Ch ir a l In ver sion a n d Hyd r olysis of Th a lid om id e:
Mech a n ism s a n d Ca ta lysis by Ba ses a n d Ser u m Albu m in ,
a n d Ch ir a l Sta bility of Ter a togen ic Meta bolites
Marianne Reist,† Pierre-Alain Carrupt,† Eric Francotte,‡ and Bernard Testa*,†
Institute of Medicinal Chemistry, University of Lausanne, BEP-Dorigny, CH-1015 Lausanne,
Switzerland, and Drug Discovery, NOVARTIS Pharma AG, K-122.P.25,
CH-4002 Basel, Switzerland
Received August 4, 1998
The chiral inversion and hydrolysis of thalidomide and the catalysis by bases and human
serum albumin were investigated by using a stereoselective HPLC assay. Chiral inversion
was catalyzed by albumin, hydroxyl ions, phosphate, and amino acids. Basic amino acids (Arg
and Lys) had a superior potency in catalyzing chiral inversion compared to acid and neutral
ones. The chiral inversion of thalidomide is thus subject to specific and general base catalysis,
and it is suggested that the ability of HSA to catalyze the reaction is due to the basic groups
of the amino acids Arg and Lys and not to a single catalytic site on the macromolecule. The
hydrolysis of thalidomide was also base-catalyzed. However, albumin had no effect on
hydrolysis, and there was no difference between the catalytic potencies of acidic, neutral, and
basic amino acids. This may be explained by different reaction mechanisms of the chiral
inversion and hydrolysis of thalidomide. Chiral inversion is deduced to occur by electrophilic
substitution involving specific and general base catalysis, whereas hydrolysis is thought to
occur by nucleophilic substitution involving specific and general base as well as nucleophilic
catalysis. As nucleophilic attack is sensitive to steric properties of the catalyst, steric hindrance
might be the reason albumin is not able to catalyze hydrolysis. 1H NMR experiments revealed
that the three teratogenic metabolites of thalidomide, in sharp contrast to the drug itself, had
complete chiral stability. This leads to the speculation that, were some enantioselectivity to
exist in the teratogenicity of thalidomide, it could result from fast hydrolysis to chirally stable
teratogenic metabolites.
pharmacological activities is considered (11, 12), the
question of whether the biological activities of thalido-
mide are associated with either one or both of the
enantiomers arises. Unfortunately, findings in the lit-
erature are contradictory, and it is presently still unclear
whether any of the effects of thalidomide are enantio-
selective (13). The problem is complicated by the fact
that the enantiomers of thalidomide are subject to rapid
chiral inversion in vitro (14) and in vivo (15). In vitro
studies showed that the chiral inversion of thalidomide
is catalyzed by human serum albumin (HSA) and that
this catalysis can be inhibited by various albumin ligands
such as long- and medium-chain fatty acids and acetyl-
salicylic acid (14, 16). However, no detailed mechanism
of the chiral inversion of thalidomide and its catalysis
by HSA has been suggested.
Thalidomide has been found to be eliminated mainly
by spontaneous hydrolysis in blood and tissues of humans
and animals (17, 18). All four amide bonds of the
molecule are susceptible to hydrolytic cleavage at pH >6
(19) (Figure 1). The main urinary metabolites in humans
are 2-phthalimidoglutaramic acid (3) (about 50%) and
R-(o-carboxybenzamido)glutarimide (4) (about 30%) (20).
Of the 12 hydrolysis products of thalidomide, only those
three which contain the intact phthalimide moiety showed
teratogenic activity, i.e., 2- (3) and 4-phthalimidoglut-
aramic acid (2) and 2-phthalimidoglutaric acid (5) (21,
22) (Figure 1). An investigation of the teratogenic
potency of the optical antipodes of 2-phthalimidoglutaric
In tr od u ction
Thalidomide [R-(N-phthalimido)glutarimide], originally
introduced as a non-barbiturate sedative with antinausea
properties, was withdrawn from the market in 1961
because of its catastrophic teratogenicity (1, 2). Since
Sheskin (3) reported that thalidomide caused a dramatic
improvement of the inflammatory reactions of leprosy
patients, clinical interest in thalidomide began to rise
again. It has become widely employed in the treatment
of erythema nodosum leprosum and has also been suc-
cessfully used to treat various inflammatory conditions
and autoimmune diseases (4-7). Besides sedative and
neurotropic effects, teratogenicity, neurotoxicity, and
anti-inflammatory and immunomodulatory activity, tha-
lidomide has also been reported to reduce the rate of HIV1
type 1 replication in vitro (8, 9). Although the mecha-
nisms of action of thalidomide are only poorly understood,
some of its activities may be related to its capacity to
inhibit the production of TNF-R (10).
In all medical treatments thalidomide, which has one
chiral center, is administered in the form of a racemate.
When the fact that enantiomers can have very different
* Corresponding author. Telephone: +41 21 692 4521. Fax: +41
21 692 4525. E-mail: Bernard.Testa@ict.unil.ch.
† University of Lausanne.
‡ NOVARTIS Pharma AG.
1 Abbreviations: HIV, human immunodeficiency virus; HSA, human
serum albumin; TNF-R, tumor necrosis factor-R; ASA, acetylsalicylic
acid.
10.1021/tx9801817 CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/06/1998

1522 Chem. Res. Toxicol., Vol. 11, No. 12, 1998
Reist et al.
F igu r e 1. First products in the hydrolytic degradation of thalidomide in aqueous solutions at pH >6: thalidomide (1),
4-phthalimidoglutaramic acid (2), 2-phthalimidoglutaramic acid (3), R-(o-carboxybenzamido)glutarimide (4), 2-phthalimidoglutaric
acid (5), 4-(o-carboxybenzamido)glutaramic acid (6), 2-(o-carboxybenzamido)glutaramic acid (7), and 2-(o-carboxybenzamido)glutaric
acid (8). Teratogenic compounds are circled (modified from ref 19).
F igu r e 2. Interplay of biological activities, metabolism, and stereoselectivity of thalidomide enantiomers and their ring-opened
metabolites.
acid (5) in pregnant mice showed that the S-enantiomer
caused dose-dependent teratogenicity, whereas the R-
enantiomer was devoid of an effect even at doses 4 times
higher (23, 24).
In Figure 2, the biological activities of thalidomide are
summarized, and these data hint at the complexity of
its metabolism and activities. Some of the questions
emphasized in Figure 2 can be answered in part. Others,
however, do not appear to have been adequately inves-
tigated. For example, there is no information available
about the configurational stability of the three teratoge-
nic metabolites of thalidomide. In this paper, the mech-
anism of chiral inversion of thalidomide and its catalysis
by HSA and general bases was investigated by using a
stereoselective HPLC assay. Further, the configurational
stability of the teratogenic metabolites of thalidomide was

Chiral Inversion and Hydrolysis of Thalidomide
Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1523
× 10-7, 2.0 × 10-6, 1.0 × 10-5, 4.0 × 10-5, 7.46 × 10-5, 1.49 ×
10-4, 1.80 × 10-4, 2.99 × 10-4, and 5.97 × 10-4 M) in phosphate
buffer (0.1 M, pH 7.40, ionic strength of 0.3) were prepared to
study the influence of HSA concentration. To investigate the
influence of pH, 0.1 M phosphate buffers (Sørensen) at six
different pH values (2.03, 3.58, 6.06, 7.44, 8.03, and 11.00) and
an ionic strength of 0.3 were prepared. To study the influence
of phosphate concentration, phosphate buffers (Sørensen) at four
different concentrations (0.05, 0.10, 0.20, and 0.30 M) with a
constant pH of 7.40 and a final ionic strength of 0.8 were
prepared. The influence of neutral, acidic, and basic amino acids
was investigated in 0.1 M solutions of amino acids Arg, Asp,
Gln, Glu, Gly, Lys, and Ser in phosphate buffer (0.1 M, pH 7.40,
ionic strength of 0.3) with a constant pH of 7.40 and a final ionic
strength of 0.9. All solutions were freshly prepared.
(3) In cu ba tion Con d ition s. Fifty microliters of the stock
solution of the enantiomers was evaporated to dryness under a
stream of nitrogen. The final concentration of each enantiomer
in the reaction mixture was about 1 × 10-4 M. The reaction
was started by adding 1 mL of incubation medium (solutions
described above), and the mixture was kept in a rotating water
bath (Infors HT WTR1, Infors AG, Bottmingen, Switzerland)
at 37 ( 0.2 °C. At various time intervals, the reaction was
stopped by immediate extraction as described below.
established, and implications on the stereoselectivity of
the teratogenicity of thalidomide are discussed.
Ma ter ia ls a n d Meth od s
Ch em ica ls. (-)-(S)-Thalidomide [(-)-(S)-1,3-dioxo-2-(2′,6′-
dioxopiperidin-3′-yl)isoindol] and (+)-(R)-thalidomide were ob-
tained by preparative chromatographic separation on Chiralcel
OJ . The preparative HPLC was performed with a Shimadzu
modular liquid chromatograph (Burkhardt Instrumente, Zu¨rich,
Switzerland) composed of a LC-8A pump and a multiwavelength
UV/Vis model SPD-10A detector. The UV signal was recorded
and processed by an Epson microcomputer, using the Class LC-
10 chromatographic software (Shimadzu, Burkhardt Instru-
mente). Racemic thalidomide (500 mg) prepared according to
a reported method (25) was dissolved in 25 mL of dichlo-
romethane and diluted with 500 mL of ethanol. This solution
was injected via the pump on a 5 cm (i.d.) × 50 cm Chiralcel
OJ column (Daicel Chemical Industries). Chromatography was
carried out at room temperature at a flow rate of 150 mL/min,
and UV detection was performed at 220 nm. The mobile phase
was pure ethanol containing small amounts of trifluoroacetic
acid (0.05%) to avoid racemization of the isolated enantiomers
during workup. Under the applied chromatographic conditions,
the R-enantiomer was isolated from a first fraction collected
between 38 and 50 min and the S-enantiomer from a second
fraction collected between 58 and 80 min. The enantiomeric
purity of the isolated enantiomers, determined by chiral HPLC
as described in Chromatographic Conditions, was 99.70% for
(S)-thalidomide and 99.92% for (R)-thalidomide.
(RS)-2-Phthalimidoglutaramic acid (phthalyl-DL-glutamine)
(3) and (RS)-4-phthalimidoglutaramic acid (phthalyl-DL-iso-
glutamine) (2) were prepared from L-glutamic acid and phthalic
anhydride according to the method described by King et al. (26,
27) and recrystallized from water or aqueous alcohol, respec-
tively. The identity and purity of the two substances were
confirmed by melting points and 1H NMR spectra.
HSA (fraction V, essentially fatty acid-free, A-1887, lot
42H9313) was purchased from Sigma (Buchs, Switzerland).
Anthracene puriss., (S)-2-phthalimidoglutaric acid (N-phthaloyl-
L-glutamic acid) (5), glycine, L-serine, L-lysine monohydrochlo-
ride, L-arginine monohydrochloride, L-aspartic acid, L-glutamic
acid, L-glutamine, phthalic anhydride, potassium chloride,
potassium dihydrogen phosphate, and disodium hydrogen phos-
phate anhydrous were obtained from Fluka Chemie AG (Buchs,
Switzerland) and ethanol absolute, ethyl acetate, and aceto-
nitrile, all being HPLC quality, from Romil Chemicals (Cam-
bridge, England). The deuterated solvents were purchased from
Armar (Do¨ttingen, Switzerland), and their isotopic purities were
99.8 at. % D D2O, >99.5 at. % D sodium deuterium oxide (40%)
in D2O, and 99.8 at. % D DMSO-d6. All other chemicals and
solvents were obtained from Fluka Chemie AG or Merck
(Darmstadt, Germany); they were of analytical grade and used
without further purification.
(4) Extr a ction . After addition of 50 µL of the anthracene
solution as an internal standard, the incubation mixtures were
extracted with 4 mL of ethyl acetate for 10 min. The organic
layers were separated by centrifugation (1500g for 10 min).
Three milliliters of the organic layers was transferred to a tube,
and the solvent was then evaporated to dryness under vacuum
and a stream of nitrogen. The residues were dissolved in 300
mL of absolute ethanol, and 20 µL aliquots were injected onto
the column.
For the validation of the extraction, solutions of anthracene
and rac-thalidomide at known concentrations and solutions of
mixtures of the pure thalidomide enantiomers with known
enantiomeric ratios were prepared in 99.8% ethanol. An aliquot
of these solutions was evaporated to dryness under a stream of
nitrogen. One milliliter of phosphate buffer (0.30 M, pH 7.40,
ionic strength of 0.8) or HSA solution (5.0 × 10-7, 2.0 × 10-6
,
1.0 × 10-5, 4.0 × 10-5, and 5.97 × 10-4 M) was added and the
extraction continued as described above.
(5) Estim a tion of Ra te Con sta n ts. In the case of non-
stereoselective inversion of configuration in phosphate buffers
and amino acid solutions, the observed pseudo-first-order rate
constants of enantiomerization or racemization (29) were ob-
tained by plotting the natural logarithm of the decreasing
enantiomeric excess of the predominating enantiomer as a
function of time according to eq 1:
[Edec]t - [Einc t
]
ln
) -kract ) -2kenant
t
(1)
(
)
[Edec]t + [Einc t
]
where [Edec]t is the concentration of the enantiomer whose level
is decreasing at time t, [Einc]t the concentration of the enanti-
omer whose level is increasing at time t, krac the observed
pseudo-first-order rate constant of racemization, and kenant that
of enantiomerization.
In flu en ce of Albu m in , p H, P h osp h a te, a n d Am in o Acid s
on th e Ch ir a l In ver sion a n d Hyd r olysis of Th a lid om id e
by HP LC. (1) Ch r om a togr a p h ic Con d ition s. The HPLC
system consisted of a model 360 autosampler, a model 420
pump, a model 440 diode array detector, a model 450 MT2/DAD
data system, a model 800 plotter (all from Kontron Instruments,
Zu¨rich-Mu¨llingen, Switzerland). The thalidomide enantiomers
and the internal standard anthracene were separated on a 250
mm × 4 mm column with p-methylbenzoylcellulose coated on
silica as the stationary phase (28). The mobile phase was
absolute ethanol (Romil Chemicals), the flow rate 0.7 mL/min,
and the detection wavelength 230 nm.
In the case of a stereoselective inversion of configuration, i.e.,
in albumin solutions, the observed pseudo-first-order rate
constants of chiral inversion k(R-to-S) and k(S-to-R) were estimated
by nonlinear regression analysis of the experimentally deter-
mined, decreasing weighted difference of the two enantiomers,
using eq 2. The nonlinear regression analyses were performed
with GraphPad-Prism 1.03 (GraphPad Software, Sorrento Val-
ley, San Diego, CA)
(2) P r ep a r a tion of Bu ffer s a n d Solu tion s. Stock solutions
of (R)- and (S)-thalidomide were prepared weekly in acetonitrile
(0.5 mg/mL, 1.94 × 10-3 M), and the internal standard (an-
thracene) was dissolved daily in ethyl acetate (0.1 mg/mL, 5.61
× 10-4 M). HSA solutions (HSA fraction V, essentially fatty
acid-free, Sigma A-1887) at nine different concentrations (5.0
[Edec]t - [Einc t
]
k2 - k1
k1 + k2
k1
e-(k +k )t
(2)
1
2
)
+ 2
k1 + k2
[Edec]t + [Einc t
]
where k2 is the observed pseudo-first-order rate constant of
chiral inversion of the R-enantiomer to the S-enantiomer,

1524 Chem. Res. Toxicol., Vol. 11, No. 12, 1998
Reist et al.
k(R-to-S), and k1 that of the chiral inversion of the S-enantiomer
d r olysis of Th a lid om id e. The extraction method used
in this study was similar to that reported by Chen et al.
(14, 18). After extraction of the incubation mixtures with
ethyl acetate, three peaks were seen in the chromato-
grams, one for each enantiomer of thalidomide and the
third for the internal standard anthracene. The resolu-
tion factor of (R)-thalidomide (tR ) 7.86 min) versus (S)-
thalidomide (tR ) 13.43 min) was 1.5, and the resolution
factor of (S)-thalidomide versus anthracene (tR ) 18.36
min) was 1.3. Due to their anionic structure at pH 7.4,
the products of hydrolysis of thalidomide were not
extracted.
to the R-enantiomer, k(S-to-R)
.
The observed pseudo-first-order rate constants of hydrolysis
were obtained by plotting the natural logarithm of the decreas-
ing concentration of thalidomide as a function of time according
to eq 3:
[Thalid]t
ln
) -khyd
t
(3)
(
)
[Anthra]t
where [Thalid]t is the total thalidomide concentration at time
t, [Anthra]t the concentration of the internal standard an-
thracene at time t, and khyd the observed pseudo-first-order rate
constant of hydrolysis.
(1) Lin ea r ity betw een r a c-Th a lid om id e a n d th e
In t er n a l St a n d a r d An t h r a cen e b efor e a n d a ft er
Extr a ction . Anthracene proved to be a suitable internal
standard. Solutions with eight different concentrations
of anthracene (between 7.0 × 10-5 and 1.4 × 10-3 M) and
a constant concentration of rac-thalidomide and solutions
with eight different concentrations of rac-thalidomide
(from 4.8 × 10-5 to 9.7 × 10-4 M) and a constant
concentration of anthracene were analyzed by HPLC
before and after extraction from phosphate buffer (0.30
M, pH 7.40, ionic strength of 0.8) or HSA solutions with
four different concentrations. Linear relationships be-
tween thalidomide and anthracene concentrations were
found before and after extraction from the different
incubation media (r2 > 0.995). The extraction yields of
(R)- and (S)-thalidomide were 92.8 ( 6.8 and 92.5 ( 7.2%,
respectively (mean ( SD).
(2) Va lid a tion of En a n tiom er ic Ra tios befor e a n d
a fter Extr a ction . Thalidomide solutions with eleven
different enantiomeric ratios (R/S ratio between 0.1 and
10.0) were analyzed by HPLC before and after extraction
from phosphate buffer (0.30 M, pH 7.40, ionic strength
of 0.8) or HSA solutions with four different concentra-
tions. The relation between the relative peak areas
determined by HPLC in percent and the percentage of
enantiomer injected was linear for all determinations (r2
) 0.999).
In flu en ce of HSA Con cen tr a tion on th e Ch ir a l
In ver sion a n d Hyd r olysis of Th a lid om id e. Apparent
pseudo-first-order rate constants of chiral inversion and
hydrolysis (khyd) were determined in a series of fatty acid-
free HSA solutions over a concentration range of 5.0 ×
10-7 to 5.97 × 10-4 M at pH 7.40 and 37 °C. As reported
previously (14, 16), the chiral inversion of thalidomide
was found to be catalyzed by HSA. The observed pseudo-
first-order rate constants of chiral inversion were linearly
dependent on HSA concentration (Figure 3). Further-
more, the comparison of the rates of chiral inversion of
(R)- and (S)-thalidomide, k(R-to-S) and k(S-to-R), showed a
slight HSA concentration-dependent stereoselectivity
(Figure 3). At low albumin concentrations of up to 1.5
× 10-4 M, no significant difference between the rates of
inversion of the two enantiomers was observed. At a
physiological concentration of fatty acid-free HSA (5.97
× 10-4 M) however, the inversion from (S)- to (R)-
thalidomide was about 1.4 times faster than that from
the R- to the S-enantiomer.
For four albumin concentrations, the observed pseudo-first-
order rate constants of chiral inversion k(R-to-S) and k(S-to-R) as
well as those of hydrolysis were estimated by fitting of a two-
compartment kinetic model with elimination from both com-
partments to the measured concentrations of (R)- and (S)-
thalidomide during incubation of either enantiomer. Model
equations were written with Siphar software (Siphar/PC Version
4.0, Simed S.A., Cre´teil, France), and the model was fitted
simultaneously to the concentration curves of (R)- and (S)-
thalidomide obtained in each experiment.
In vestiga tion of th e Con figu r a tion a l Sta bility a n d Hy-
d r olysis of th e Th r ee Ter a togen ic Th a lid om id e Meta bo-
lites by 1H NMR. The configurational stability and hydrolysis
of the three teratogenic thalidomide metabolites 2, 3, and 5 were
1
monitored by H NMR (1H/2H substitution). The method takes
advantage of the fact that the inversion of chiral centers of the
type R′′R′RCH and proton-deuterium exchange share a com-
mon mechanism (30-33). All 1H NMR spectra and integrals
were recorded on a Varian VXR-200 NMR spectrometer operat-
ing at 200 MHz (Varian Associates Inc., Palo Alto, CA).
4-Phthalimidoglutaramic acid (2), 2-phthalimidoglutaramic acid
(3), or 2-phthalimidoglutaric acid (5) (25 mg) was dissolved in
0.8 mL of phosphate buffer (0.3 M, pD 7.4) and the pD value
adjusted to 7.4 with 1 M NaOD. The pD values were measured
at room temperature (25 °C) using a Metrohm 654 pH-meter
(Herisau, Switzerland) coupled with a glass electrode and
calibrated with standard aqueous buffers. The pD values were
obtained by adding a correction factor of 0.4 to the measured
pH values (34, 35). Each such solution was placed in a sealed
NMR tube and each tube kept in a water bath (Haake D1,
Haake GmbH, Karlsruhe, Germany) at 37 °C. At various time
intervals of up to 7 days, 1H NMR spectra and integrals were
recorded at room temperature. Deuteration [i.e., chiral inver-
sion or racemization (36)] and hydrolysis can be observed by
following changes in the NMR spectra. Deuteration was
monitored by following the decrease of the sum of the integral
of the signal of the proton coupled to the chiral center of the
thalidomide metabolites (2, 3, and 5) and the integral of the
signal of the proton coupled to the chiral center of the hydrolyzed
metabolites (i.e., 6-8), in relation to the integral of the signal
of the four aliphatic protons in the R- and â-positions with
respect to the chiral center (unresolved bulk resonances). The
observed pseudo-first-order rate constants of hydrolysis were
estimated by plotting the natural logarithm of the increasing
integral of the signal of the proton coupled to the chiral center
of the hydrolyzed metabolites (i.e., 6-8) as a function of time
according to eq 4:
ln(Ct/Bt) ) khyd
t
(4)
where Ct is the integral of the proton coupled to the chiral center
of the hydrolyzed metabolites at time t and Bt the integral of
the unexchangeable reference protons (i.e., the four aliphatic
protons in the R- and â-positions with respect to the chiral
center) at time t.
The hydrolysis of thalidomide was not influenced much
by HSA, except for a marginal inhibition (Figure 3).
There was no significant difference in the rates of
hydrolysis between the two enantiomers at all HSA
concentrations investigated.
Resu lts
Va lid a tion of th e Ster eosp ecific HP LC Assa y
Used To In vestiga te th e Ch ir a l In ver sion a n d Hy-
In flu en ce of p H on th e Ch ir a l In ver sion a n d
Hyd r olysis of Th a lid om id e. The chiral inversion and

Chiral Inversion and Hydrolysis of Thalidomide
Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1525
Ta ble 1. p H Dep en d en ce of th e Obser ved
P seu d o-F ir st-Or d er Ra te Con sta n tsa of
En a n tiom er iza tion a n d Hyd r olysis of Th a lid om id e (0.1 M
p h osp h a te bu ffer w ith a n ion ic str en gth of 0.3 a t 37 °C)
kenant (h-1
)
khyd (h-1
)
pH
R
S
R
S
2.03
3.58
<0.001
<0.001
<0.001
<0.001
<0.01
<0.01
<0.01
<0.01
6.06 0.010 ( 0.0004 0.010 ( 0.0011 0.04 ( 0.024 0.04 ( 0.024
7.44 0.121 ( 0.0035 0.112 ( 0.0041 0.32 ( 0.013 0.33 ( 0.010
8.03 0.34 ( 0.031
0.32 ( 0.067 1.25 ( 0.052 1.27 ( 0.019
a
F igu r e 3. Influence of the concentration of fatty acid-free HSA
(Sigma A-1887) on the chiral inversion and hydrolysis of
thalidomide (HSA solutions in 0.1 M phosphate buffer at pH
7.40, an ionic strength of 0.3, and 37 °C). kobs is the observed
pseudo-first-order rate constant of chiral inversion or hydroly-
sis: (O) inversion from the S- to R-configuration, (b) inversion
from the R- to S-configuration, (3) hydrolysis of (S)-thalidomide,
and (2) hydrolysis of (R)-thalidomide. Values are means (
standard deviation of triplicate determinations. Some standard
deviations were smaller than the symbols.
Rate constants of enantiomerization kenant were calculated
from experimentally determined rate constants of racemization
krac ()2kenant) (29). Values are means ( standard deviation of
triplicate determinations.
hydrolysis of thalidomide enantiomers in phosphate
buffers were found to be pH-dependent. The values for
the observed pseudo-first-order rate constants of enan-
tiomerization kenant and of hydrolysis khyd, as obtained in
a series of 0.1 M phosphate buffers over a pH range of
2-8 at 37 °C, are compiled in Table 1. Hydrolysis of
thalidomide at pH 11 was too fast to be followed. Thus,
no rates of chiral inversion or hydrolysis could be
determined at this pH. Both rates are about zero at low
pH and were found to increase with pH, indicating that
chiral inversion and hydrolysis of thalidomide are base-
catalyzed processes. The rates of hydrolysis were found
to be faster than those of enantiomerization at all pH
values investigated. As expected, no stereoselectivity of
chiral inversion or hydrolysis was found in 0.1 M
phosphate buffers over the pH range investigated. The
reaction rates of (R)- and (S)-thalidomide are the same
within limits of error (Table 1).
F igu r e 4. Influence of phosphate concentration on thalidomide
enantiomerization and hydrolysis (pH 7.4, ionic strength of 0.8,
37 °C). kobs is the observed pseudo-first-order rate constant of
enantiomerization or hydrolysis. Values are means ( standard
deviation of triplicate determinations. Some standard deviations
were smaller than the symbols.
In flu en ce of P h osp h a te Con cen tr a tion on th e
Ch ir a l In ver sion a n d Hyd r olysis of Th a lid om id e.
Apparent pseudo-first-order rate constants of enanti-
omerization kenant and hydrolysis khyd of thalidomide,
measured at a constant pH (7.4), ionic strength (0.8), and
temperature (37 °C), were linearly dependent upon
phosphate concentration (Figure 4). Equations 5 and 6
describe this linear dependence for chiral inversion and
hydrolysis, respectively:
F igu r e 5. Influence of various amino acids on the enantiomer-
ization of thalidomide (solutions of amino acids in 0.1 M
phosphate buffer at pH 7.4, an ionic strength of 0.9, and 37
°C): (PB0.1M) 0.1 M phosphate buffer at pH 7.4 and an ionic
strength of 0.9 and (PB0.2M) 0.2 M phosphate buffer at pH 7.4
and an ionic strength of 0.9. Values are means ( standard
deviation of triplicate determinations.
In flu en ce of Va r iou s Am in o Acid s on th e Ch ir a l
In ver sion a n d Hyd r olysis of Th a lid om id e. The
apparent pseudo-first-order rate constants of enanti-
omerization and hydrolysis of thalidomide were deter-
mined in 0.1 M solutions of various acidic, neutral, and
basic amino acids. The temperature, pH, and ionic
strength were kept constant. Acidic (Glu and Asp) and
neutral (Ser, Gln, and Gly) amino acids (i.e., with only
one basic group) catalyzed the chiral inversion of thali-
domide to the same extent as a phosphate buffer with
the same molarity. In contrast, basic amino acids (Arg
and Lys) showed a higher catalytic potency (Figure 5).
The hydrolysis of thalidomide was catalyzed to the same
extent by all amino acids investigated or by a phosphate
buffer with the same molarity (data not shown). No
stereoselectivity was detected in any of these experi-
ments.
kenant ) 0.24((0.017)Cphosphate + 0.081((0.003)
n ) 4, r2 ) 0.990, s ) 0.003, F ) 201
(5)
(6)
khyd ) 0.55((0.073)Cphosphate + 0.23((0.013)
n ) 4, r2 ) 0.965, s ) 0.014, F ) 56
where n is the number of pH values investigated, r2 the
squared correlation coefficient, s the standard deviation
of the residuals, and F the Fisher test for significance of
the equation (standard deviations are given in paren-
theses).
Again, the rates of hydrolysis were found to be faster
than those of enantiomerization at all phosphate con-
centrations investigated, and no stereoselectivity was
detected. The rates of reaction of (R)- and (S)-thalido-
mide were the same within limits of error.
Con figu r a tion a l Sta bility a n d Hyd r olysis of th e
Th r ee Ter a togen ic Th a lid om id e Meta bolites As

1526 Chem. Res. Toxicol., Vol. 11, No. 12, 1998
Reist et al.
Exa m in ed by 1H NMR. The configurational stability
two basic groups) have a greater catalytic potency than
neutral and acidic amino acids that possess only one basic
group. This also implies a general base catalysis in the
chiral inversion of thalidomide. Albumin has many
reactive ꢀ-amino groups (37), and almost all of its ligand
binding sites involve lysine and/or arginine residues (37,
38). For example, one of the most prominent of all of
these residues is Lys-199, which is involved in the
binding of acetylsalicylic acid (ASA). Previous studies
(14, 16) have shown that various albumin ligands such
as long- and medium-chain fatty acids as well as ASA
were able to partly inhibit the ability of HSA to catalyze
the chiral inversion of thalidomide. However, these
ligands do not share the same binding site on albumin.
ASA binds to Sudlow’s site I; medium-chain fatty acids
bind to Sudlow’s site II, and long-chain fatty acids bind
to about six sites (37, 38). Yet these binding sites all
involve reactive lysine and arginine residues (37, 38).
Thus, the binding of a ligand could block these basic
amino acid residues, rendering them nonavailable for
catalyzing the chiral inversion of thalidomide. On the
basis of these arguments and on the basis of the findings
of this study that the chiral inversion of thalidomide is
subject to general base catalysis, it is suggested that the
ability of albumin to catalyze the chiral inversion of
thalidomide is due to its Arg and Lys residues. No single
catalytic site on the albumin molecule is thought to exist.
This suggestion is also consistent with the fact that the
chiral inversion of thalidomide in vivo, in plasma, and
in blood was found to be slower than in HSA solutions
with physiological concentrations (15, 16). Endogenous
albumin ligands such as fatty acids and bilirubin may
block the catalytic basic amino acid residues of HSA and
as a result reduce the extent of catalysis of the chiral
inversion.
Furthermore, the rates of chiral inversion of (R)- and
(S)-thalidomide showed a slight HSA concentration-
dependent stereoselectivity. At low albumin concentra-
tions, no significant difference between the rates of
inversion of the two enantiomers could be observed,
whereas at physiological concentrations, the inversion
from (S)- to (R)-thalidomide was about 1.4 times faster
than that from the R- to the S-enantiomer. This may be
due to the fact that albumin is subject to concentration-
dependent polymerization caused by various factors such
as heating, freeze-drying, etc. (39, 40). For the com-
mercial fatty acid-free HSA used in this study, which is
freeze-dried, it is known that at low albumin concentra-
tions no polymerization occurs whereas at a physiological
concentrations up to 50% of the HSA can be polymerized
(40). This polymerization might lead to conformational
changes in the HSA molecule, inducing stereoselectivity
by an unknown mechanism. This hypothesis is in
agreement with the finding that in freshly drawn blood
no significant difference could be observed between the
rates of chiral inversion of the enantiomers. No albumin
polymerization is believed to exist in untreated human
blood (38). In vivo in humans, it was even found that
(R)-thalidomide inverses its configuration faster than (S)-
thalidomide (15). Hence, the stereoselectivity of the
chiral inversion of thalidomide is dependent on the
medium, and it would be misleading to draw conclusions
about in vivo stereoselectivity from in vitro studies.
In summary, the chiral inversion of thalidomide is
suggested to occur by electrophilic substitution with
protons as incoming and leaving groups. Specific and
and hydrolysis of the three teratogenic metabolites of
1
thalidomide 2, 3, and 5 (Figure 1) were monitored by H
NMR. In a 0.3 M phosphate buffer at pD 7.4 and 37 °C,
all three metabolites underwent hydrolysis. However,
no deuteration was observed for any of the three me-
tabolites after up to 7 days of incubation. There was no
observable decrease of the sum of the integrals of the
signal of the proton coupled to the chiral center of the
teratogenic metabolites (2, 3, and 5) and that of the
proton coupled to the chiral center of the hydrolyzed
metabolites (i.e., 6-8). The ratios of (sum of the integrals
of the protons of the chiral centers)/(integral of the signal
of the four aliphatic reference protons in the R- and
â-positions with respect to the chiral center) at time zero
and after 7 days were identical. Hence, the configuration
of the three teratogenic thalidomide metabolites 2, 3, and
5 (Figure 2) in a phosphate buffer (0.3 M, pD 7.4) at 37
°C was found to be completely stable for at least 7 days.
The half-lives of hydrolysis under these conditions were
estimated to be 6 h for 4-phthalimidoglutaramic acid (2),
49 h for 2-phthalimidoglutaramic acid (3), and 47 h for
2-phthalimidoglutaric acid (5). For the same compounds
at pH 7.4 and 37 °C, Schumacher et al. (19) reported
hydrolysis half-lives of 5, 48, and 43 h, respectively.
Discu ssion
Mech a n ism of Ch ir a l In ver sion . The chiral HPLC
assay described in this study was shown to be suitable
for the simultaneous determination of the chiral inver-
sion and hydrolysis of thalidomide in vitro. Good ac-
curacy and precision were shown at all R/S and thalido-
mide/anthracene concentration ratios. The simultaneous
determination of the chiral inversion and hydrolysis of
thalidomide using only one chiral HPLC column is novel.
As expected (14, 16), human serum albumin catalyzed
the chiral inversion of thalidomide. The observed pseudo-
first-order rate constants of chiral inversion increased
with increasing HSA concentration. To obtain further
insights into the mechanism of chiral inversion of tha-
lidomide and its catalysis by HSA, the influence of pH,
phosphate concentration, and various acidic, neutral, and
basic amino acids was investigated. Our results showed
that the chiral inversion of thalidomide was pH-depend-
ent. At acidic pH, the rate of chiral inversion was about
zero, and it increased with increasing pH. This suggests
that the chiral inversion of thalidomide is a base-
catalyzed process. It was also shown that at pH 7.4 the
rate of inversion was linearly dependent on phosphate
concentration. Thus, chiral inversion is deduced to be
subject to general base catalysis. Further information
about the catalytic species involved in phosphate buffers
at pH 7.4 can be obtained from eq 5, which represents
the linear dependence of chiral inversion upon phosphate
concentration. The intercept at zero buffer concentration
is 0.081((0.003), thus different from zero, and it can be
presumed that at pH 7.4 uncatalyzed and specific base-
catalyzed reactions take place along with general base-
catalyzed reactions. Hence, the chiral inversion of tha-
lidomide in phosphate buffers at pH 7.4 is catalyzed by
2-
the basic component of the phosphate buffer (HPO4
as well as by hydroxyl ions and water molecules.
)
The influence of various acidic, neutral, and basic
amino acids on the chiral inversion of thalidomide showed
that the basic amino acids Arg and Lys (which possess

Chiral Inversion and Hydrolysis of Thalidomide
Chem. Res. Toxicol., Vol. 11, No. 12, 1998 1527
general base catalysis accelerates the reaction by facili-
tating the abstraction of the proton bound to the chiral
center.
catalyze the chiral inversion of thalidomide is due to the
basic functional groups of the amino acids arginine and
lysine in general and not to a single catalytic site.
Further, the configurational stability of the three ter-
atogenic metabolites allows us to speculate that (R)-
thalidomide might be less teratogenic than the S-enan-
tiomer.
Mech a n ism of Hyd r olysis. The hydrolysis of thali-
domide, like its chiral inversion, was found to be base-
catalyzed. From eq 6, which represents the linear
dependence of the rate of hydrolysis on phosphate
concentration, it can be deduced that at pH 7.4 the
hydrolysis of thalidomide, like its chiral inversion, is
catalyzed by the basic component of the phosphate buffer
as well as by hydroxyl ions and water molecules. We
showed that the hydrolysis of thalidomide is catalyzed
by amino acids. However, in contrast to chiral inversion,
there was no difference between the catalytic potencies
of acidic, neutral, and basic amino acids, and albumin
did not catalyze the hydrolysis of thalidomide. This may
be explained by a different reaction mechanism of chiral
inversion and hydrolysis. Whereas chiral inversion is
suggested to occur by electrophilic substitution, the
reaction mechanism of hydrolysis is thought to be a
nucleophilic substitution involving specific and general
base catalysis as well as nucleophilic catalysis. Proton
transfer (e.g., base catalysis) is rather insensitive to steric
factors since the proton is so small, whereas nucleophilic
attack is very sensitive to the size of both the reactant
and catalyst (41). With regard to its size, it is likely that
albumin is efficient as a base catalyst but not as a
nucleophilic catalyst. This might explain why albumin
catalyzes chiral inversion, but not the hydrolysis of
thalidomide. Steric hindrance might also account for the
fact that basic amino acids were indeed more effective
than neutral and acidic ones in catalyzing chiral inver-
sion but not hydrolysis. In theory, hydrolysis might also
occur via an E1cb mechanism and a ketene intermediate
stemming from the same carbanion involved in chiral
inversion and deuterium exchange. But the absence of
catalysis of hydrolysis by albumin suggests that this
mechanism is rather improbable.
Ch ir a l Sta bility of Ter a togen ic Meta bolites. The
configuration of the three teratogenic metabolites of
thalidomide in phosphate buffer at pH 7.4 was found to
be stable for at least 7 days. 1H NMR has previously
been shown to be a suitable tool for assaying racemization
of chiral centers of the type R′′R′RCH (30-33, 36). Here,
the configurational stability of the teratogenic metabo-
lites of thalidomide is discussed in connection with a
possible enantioselectivity on the teratogenicity of tha-
lidomide. Given the stable configuration of the three
teratogenic metabolites, inversion of the chiral center
must stop with the hydrolysis of thalidomide. Thus, after
administration of (S)-thalidomide, the concentration of
teratogenic metabolites with the S-configuration is pos-
tulated to be higher than after application of (R)-
thalidomide. If it is assumed that the teratogenic potency
of the metabolites with the S-configuration is markedly
superior to that of the metabolites with the R-configu-
ration, as verified for 2-phthalimidoglutaric acid (5) (23,
24), it might in fact be conceivable that (R)-thalidomide
could cause less teratogenic effects. From a more practi-
cal point of view, the high configurational stability of the
three teratogenic metabolites of thalidomide may be
useful in assaying the biological activities of the single
enantiomers.
Ack n ow led gm en t. Dr. J oachim M. Mayer, Dr. J ean-
Luc Wolfender, Giuseppe Lisa, and Charlotte Gancel are
thanked for their valuable help with the use of the Siphar
software, the use of the NMR spectrometer, the synthesis
of the thalidomide metabolites, and the extraction pro-
tocol.
Su p p or tin g In for m a tion Ava ila ble: Detailed calculations
for the estimation of the rate constants of chiral inversion and
hydrolysis and discussion of the use of 1H NMR (1H/2H substitu-
tion) as a tool for assaying the racemization of chiral centers of
the type R′′R′RCH (6 pages). Ordering information is given on
any current masthead page.
Refer en ces
(1) Mellin, G. W., and Katzenstein, M. (1962) The saga of thalido-
mide: neuropathy to embryopathy, with case reports and con-
genital anomalies. N. Engl. J . Med. 267, 1184-1244.
(2) De Camp, W. H. (1989) The FDA perspective on the development
of stereoisomers. Chirality 1, 2-6.
(3) Sheskin, J . (1965) Thalidomide in the treatment of lepra reactions.
Clin. Pharmacol. Ther. 6, 303-306.
(4) Barnhil, R. L., and McDougall, A. C. (1982) Thalidomide: use and
possible mode of action in reactional lepromatous leprosy and in
various other conditions. J . Am. Acad. Dermatol. 7, 317-323.
(5) Koch, H. P. (1985) Thalidomide and congeners as anti-inflam-
matory agents. Prog. Med. Chem. 22, 165-242.
(6) Vogelsang, G. B., Farmer, E. R., Hess, A. D., Altamonte, V.,
Beschorner, W. E., J abs, D. A., Corio, R. L., Levin, L. S., Colvin,
O. M., and Wingard, J . R. (1992) Thalidomide for the treatment
of chronic graft-versus-host disease. N. Engl. J . Med. 326, 1055-
1058.
(7) Stevens, R. J ., Andujar, C., Edwards, C. J ., Ames, P. R., Barwick,
A. R., Khamashta, M. A., and Hughes, G. R. (1997) Thalidomide
in the treatment of the cutaneous manifestations of lupus
erythematosus: experience in sixteen consecutive patients. Br.
J . Rheumatol. 36, 353-359.
(8) Makonkawkeyoon, S., Limson-Pobre, R. N., Moreira, A. L., Schauf,
V., and Kaplan, G. (1993) Thalidomide inhibits the replication of
human immunodeficiency virus type 1. Proc. Natl. Acad. Sci.
U.S.A. 90, 5974-5978.
(9) Moreira, A. L., Corral, L. G., Ye, W., J ohnson, B., Stirling, D.,
Muller, G. W., Freedman, V. H., and Kaplan, G. (1997) Thalido-
mide and thalidomide analogs reduce HIV type 1 replication in
human macrophages in vitro. AIDS Res. Hum. Retroviruses 13,
857-863.
(10) Shannon, E. J ., Morales, E. J ., and Sandoval, F. (1997) Immu-
nomodulatory assays to study structure-activity relationships of
thalidomide. Immunopharmacology 35, 203-212.
(11) Testa, B. (1990) Definitions and concepts in biochirality. In
Chirality and Biological Activity (Holmstedt, B., Frank, H., and
Testa, B., Eds.) pp 15-32, Alan R. Liss, New York.
(12) Testa, B., and Trager, W. F. (1990) Racemates versus enantiomers
in drug development: dogmatism or pragmatism? Chirality 2,
129-133.
(13) Gunzler, V. (1992) Thalidomide in human immunodeficiency virus
(HIV) patients. A review of safety considerations. Drug Saf. 7,
116-134.
(14) Knoche, B., and Blaschke, G. (1994) Investigations on the in vitro
racemization of thalidomide by high-performance liquid chroma-
tography. J . Chromatogr. 666, 235-240.
(15) Eriksson, T., Bjo¨rkman, S., Roth, B., Fyge, A., and Ho¨glund, P.
(1995) Stereospecific determination, chiral inversion in vitro and
pharmacokinetics in humans of the enantiomers of thalidomide.
Chirality 7, 44-52.
(16) Eriksson, T., Bjo¨rkman, S., Roth, B., Fyge, A., and Ho¨glund, P.
(1998) Enantiomers of thalidomide: blood distribution and the
influence of serum albumin on chiral inversion and hydrolysis.
Chirality 10, 223-228.
In conclusion, the chiral inversion of thalidomide has
been shown to be subject to specific and general base
catalysis. It is suggested that the ability of HSA to

1528 Chem. Res. Toxicol., Vol. 11, No. 12, 1998
Reist et al.
(17) Schumacher, H., Smith, R. L., and Williams, R. T. (1965) The
metabolism of thalidomide: the fate of thalidomide and some of
its hydrolysis products in various species. Br. J . Pharmacol. 25,
338-351.
(18) Chen, T. L., Vogelsang, G. B., Petty, B. G., Brundrett, R. B., Noe,
D. A., Santos, G. W., and Colvin, O. M. (1989) Plasma pharma-
cokinetics and urinary excretion of thalidomide after oral dosing
in healthy male volunteers. Drug Metab. Dispos. 17, 402-405.
(19) Schumacher, H., Smith, R. L., and Williams, R. T. (1965) The
metabolism of thalidomide: the spontaneous hydrolysis of tha-
lidomide in solution. Br. J . Pharmacol. 25, 324-337.
(20) Beckmann, R. (1963) Ueber das Verhalten von Thalidomid im
Organismus. Arzneim.-Forsch. 13, 185-191.
(21) Meise, W., Ockenfels, H., and Ko¨hler, F. (1973) Teratologische
Pru¨fung der Hydrolyseprodukte des Thalidomids (Teratologic
determination of hydrolyzed products of thalidomide). Experientia,
Suppl. 29, 423-424.
(22) Koch, H. (1981) Die Arenoxid-Hypothese der Thalidomid-
Wirkung. Ueberlegungen zum molekularen Wirkungsmechanis-
mus des “klassischen” Teratogens (The arene oxide hypothesis
of thalidomide action. Reflections on the molecular mechanism
of action of the “classical” teratogen). Sci. Pharm. 49, 67-99.
(23) Ockenfels, H., and Ko¨hler, F. (1970) Das L-Isomere als teratogenes
Prinzip der N-Phthalyl-DL-glutaminsa¨ure (L-Isomer as the ter-
atogenic principle of N-phthalyl-DL-glutamic acid). Experientia,
Suppl. 26, 1236-1237.
(24) Ockenfels, H., Ko¨hler, F., and Meise, W. (1976) Teratogenic effect
and stereospecificity of a thalidomide metabolite. Pharmazie 31,
492-493.
(25) Beckmann, R. (1962) Ueber das Verhalten von Thalidomid im
Organismus. Arzneim.-Forsch. 12, 1095-1096.
(26) King, F. E., and Kidd, D. A. A. (1949) A new synthesis of
glutamine and of γ-dipeptides of glutamic acid from phthalylated
intermediates. J . Chem. Soc., 3315-3319.
(27) King, F. E., J ackson, B. S., and Kidd, D. A. A. (1951) Syntheses
from phthalimido-acids. Part II. Further reactions of phthalyl-
glutamic anhydride. J . Chem. Soc., 243-246.
epimerization: their meaning and pharmacological significance.
Chirality 7, 396-400.
(30) Kawazoe, Y., and Ohnishi, M. (1964) Quantitative analysis of
active hydrogen by nuclear magnetic resonance. Chem. Pharm.
Bull. 11, 846-848.
(31) Testa, B. (1973) Some chemical and stereochemical aspects of
diethylpropion metabolism in man. Acta Pharm. Suec. 10, 441-
454.
(32) Reist, M., Christiansen, L. H., Christoffersen, P., Carrupt, P. A.,
and Testa, B. (1995) Low configurational stability of amfepramone
and cathinone: mechanism and kinetics of chiral inversion.
Chirality 7, 469-473.
(33) Reist, M., Carrupt, P. A., Testa, B., Lehmann, S., and Hansen, J .
J . (1996) Kinetics and mechanisms of racemization: 5-substituted
hydantoins ()imidazolidine-2,4-diones) as models of chiral drugs.
Helv. Chim. Acta 79, 767-778.
(34) Glasoe, P. K., and Long, F. A. (1960) Use of glass electrodes to
measure acidities in deuterium oxide. J . Phys. Chem. 64, 188-
191.
(35) Mikkelsen, K., and Nielsen, S. O. (1960) Acidity measurements
with the glass electrode in H2O-D2O mixtures. J . Phys. Chem.
64, 632-637.
(36) Reist, M., Testa, B., and Carrupt, P. A. (1997) The racemization
of enantiopure drugs: helping medicinal chemists to approach
the problem. Enantiomer 2, 147-155.
(37) Peters, T., J r. (1996) The albumin molecule: its structure and
chemical properties. In All about Albumin. Biochemistry, Genetics,
and Medical Applications, pp 9-75, Academic Press, San Diego.
(38) Peters, T., J r. (1996) Ligand binding by albumin. In All about
Albumin. Biochemistry, Genetics, and Medical Applications, pp
79-132, Academic Press, San Diego.
(39) Friedli, H., and Kistler, P. (1970) Polymers in preparations of
human serum albumin. Vox Sang. 18, 542-546.
(40) Zini, R., Barre´, J ., Bre´e, F., Tillement, J . P., and Sebille, B. (1981)
Evidence for
a concentration-dependent polymerization of a
commercial human serum albumin. J . Chromatogr. 216, 191-
198.
(28) Okamoto, Y., Aburatani, R., and Hatada, K. (1987) Chromato-
graphic chiral resolution. XIV. Cellulose tribenzoate derivatives
as chiral stationary phases for high-performance liquid chroma-
tography. J . Chromatogr. 389, 95-102.
(41) Isaacs, N. S. (1987) Acid and base catalysis. In Physical Organic
Chemistry, pp 331-353, Longman Scientific & Technical, Harlow,
Essex, U.K.
(29) Reist, M., Testa, B., Carrupt, P. A., J ung, M., and Schurig, V.
(1995) Racemization, enantiomerization, diastereomerization, and
TX9801817