Carlsberg Res. Commun. Vol. 49, p. 591-596, 1984
DETERMINATION OF o-AMINO ACIDS. II. USE OF A BIFUNCTIONAL
REAGENT, 1,5-DIFLUORO-2,4-DINITROBENZENE
by
PETER MARFEY ~)
Department of Chemistry, Carlsberg Laboratory,
Gamle Carlsberg Vej I0, DK-2500 Copenhagen Valby
') Permanent address: Department of Biological Sciences,
State University of New York at Albany, Albany, N.Y. 12222, USA
Keywords: D-amino acids, bifunctional reagent, diastereomers
l-fluoro-2,4-dinitrophenyl-5-L-alanine amide has been synthesized in high yield (76%) from 1,5-difluoro-2,4-di-
nitrobenzene and L-AIa-NH2. This compound contains a reactive fluorine atom which can be used for the reaction
with a mixture of L- and D-amino acids. The resulting diastereomers which are obtained in quantitative yield can
be separated and estimated by HPLC. With the five amino acids studied (Ala, Asp, Glu, Met and Phe),
L-diastereomers were eluted from the reverse-phase column before D-diastereomers. This behavior can be explained
by a stronger i ntramolecular hydrogen bonding in the latter diastereomer. When artificial mixtures of the five amino
acids containing known proportions of L- and D-isomers were derivatized with the reagent and the reaction products
analyzed by HPLC, it was possible to determine the relative content of each isomer in a nanomole range.
1. INTRODUCTION
A bifunctional reagent, 1,5-difluoro-2,4-dini-
trobenzene, FFDNB, has previously been used
as a cross-linking reagent in protein chemistry
(2, 3, 4, 6). It reacts mainly with uncharged
amino groups in a protein and with phenolic
hydroxyl groups and sulfhydryl groups. The
derivatives are stable, so that cross-linked amino
acids can be isolated after protein hydrolysis (2,
3). A useful feature of the bridge chromophore
(-NH-DNP-NH-) is its high light absorption at
340 nm with an 8M -- 3 x 104 (2).
We have considered that this reagent can also
be used to prepare diastereomers of amino acids
and, thus, can be very useful in quantitative
determination of D-amino acids in protein hy-
drolysates. We have first synthesized FDNP-L-
alanine amide (FDNP-L-AIa-NH2) and used this
as a reagent for derivatization of a D- and
L-amino acid in a mixture. The resulting diaste-
reomers were separated and quantitated by
HPLC.
Abbreviations: DMSO = dimethylsulfoxide; DNDEAP-a.a. = N-(2,4-dinitro-5-diethylaminophenyl)-amino acid;
DNP = 2,4-dinitrophenyl; FDAA = F-DNP-L-AIa-NH2 = 1-fluoro-2,4-dinitrophenyl-5-L-alanine amide; FFDNB
= 1,5 difluoro-2,4-dinitrobenzene; HO-DAA = I-hydroxy-2,4-dinitrophenyl-5-L-alanine amide; HPLC = high
performance liquid chromatography; L-AIa-NH2. HCI = L-Alanine amide hydrochloride; TEAP = triethyl-
ammonium phosphate; TLC = thin-layer chromatography. Other abbreviations are according to the guideline of
the IUPAC-IUB, Commission of Biochemical Nomenclature.
Springer-Verlag 0 105-1938/84/0049/0591/$0 1.20
P. MARFEY: Determination of D-amino acids. II
2. MATERIALS AND METHODS
2.1. Materials
FFDNB was purchased from FIuka AG, Swit-
zerland. It had m.p. 74-75 ~ and showed only a
single spot on a silica TLC sheet with an Rfvalue
of 0.7 (benzene). L-Alanine amide hydrochlo-
ride was from Bachem Feinchemikalien AG,
Switzerland. L- and D-amino acids were pur-
chased from Sigma, USA. Polygram, Sil G/UV
254 precoated sheets were purchased from Ma-
cherey-Nagel Co., West Germany. All other
chemicals and solvents were of analytical grade
and were obtained from Merck, W. Germany.
2.2. Methods
2.2.1. Synthesis of FDNP-L-Ala-NH2 (FDAA)
This synthesis was similar to that described for
5-fluoro-2,4-dinitrodiethylaniline (5). It was es-
sential to maintain conditions exactly as indi-
cated, otherwise, the yield of the desired com-
pound was low and, instead, a hydrolyzed
product was obtained. A sample of L-Ala-
NH,- HCI (472 mg, 3.81 millimoles) was dis-
solved in 3.9 ml 1 N-NaOH and immediately 60
ml acetone was added. About 10 g of anhydrous
MgSO, was added and the contents stirred at
room temperature for about 3 hours. MgSO4 was
removed by filtration and washed twice with
little acetone. FFDNB (668 mg, 3.27 millimoles)
was dissolved in 15 ml acetone. To this solution
was added dropwise under magnetic stirring the
acetone solution of L-Ala-NH2. After addition,
the contents were stirred for an additional 0.5
hour. Equal volume of water was added resulting
in formation of the golden-yellow scales which
were filtered, washed first with little 2:1 water-
acetone mixture, then with water and finally
dried in the air and in the dark. The yield was 0.5
g (56% of theory, mol. wt. 272), m.p. 224-226 ~
From the mother liquor, upon removal of more
acetone under water pump, another fraction of
crystals was obtained in the same way as above.
The yield was 180 mg (20% of theory), m.p.
222-223 ~ Both fractions had the same Rf
value (0.4) upon TCL on silica with ethylacetate
as solvent. When subjected to HPLC using
TEAP/acetonitrile systems (section 2.2.3) both
fractions had the same elution peak (and repre-
sented 95% of the total area). The ultraviolet
spectrum of the product obtained in 25 mM-
TEAP, pH 3.0, and 50% acetonitrile had ma-
ximaat 264 nm, 338 nm (eM ~ 1.5 x t04) and a
shoulder at 380nm (a spectrum similar to the
hydrolyzed sample, see Figure 1).
2.2.2. Synthes& and HPLC of L- and D-
diastereomers of amino acids
These syntheses were closely patterned after
those used for the synthesis of DNDEAP-amino
acids (5). Aqueous solutions (50 mM) of ten
amino acids (D- and L-isomers of Ala, Asp, Glu,
Met and Phe) were used as starting materials for
synthesis. 50 gl (2.5 micromoles) of each solu-
tion was placed in separate 2 ml plastic micro
centrifuge tubes, To each was added 100 lal of 1%
acetone solution of FDAA (1 mg, 3.6 micro-
moles), the molar ratio of FDAA to amino acid
1.4:1, followed by 20 l.tl of 1 M-NaHCO~ (20
micromoles). The contents were mixed and
heated over a hot plate at 30-40 ~ for 1 hour
with frequent mixing. After cooling to room
temperature, 10 gl (20 micromoles) of 2 M-HC1
was added to each reaction mixture. After mix-
ing, the contents were dried in a vaccuum
desicator over NaOH pellets. The residues were
dissolved in 0.5 ml DMSO affording ten solu-
tions each 5.0 mM (based on amino acid).
Analysis of these solutions by HPLC under
conditions given below indicated that the deri-
vatization reaction was quantitative (see below).
A 1:1 dilution of these solutions was made (2.5
raM) and 5 gl samples of each (12.5 nanomoles)
were pooled together and injected for HPLC.
2.2.3. Chromatography
HPLC was done with equipment described in
the preceding paper. Elution was done with a
linear gradient of acetonitrile in 50 mM-TEAP
buffer, pH 3.0, from 10% to 50% acetonitrile
during 1 hour, flow rate 2 ml/min, analysis of the
effluent at 340 nm. This wave length was chosen
because it is close to ~a~ of most of the com-
pounds investigated. All the solvents were fil-
tered and degassed before use.
TLC was done with pre-coated plastic sheets
containing 0.25 mm layers of silica gel im-
pregnated with fluorescent indicator (Polygram
592 Carlsberg Res. Commun. Vol. 49, p. 591-596, 1984
P. MARFEY: Determination of D-amino acids. I!
Sil G/UV 254). Ethyl acetate, p-dioxane and
p-dioxane-benzene (3:1) mixture were used as
solvents.
2.2.4. Derivatization and quantitation of D-
isomers in mixtures of known
proportions of L- and D-isomers
Four amino acid mixtures and a reagent blank
were prepared in separate plastic 2 ml micro
centrifuge tubes. Mixture 1 contained 60% of L-
and 40% of D-isomers of Ala, Asp, Glu, Met and
Phe (12 ~tl of each 50 mM solution of L-isomer
and 8/al of each 50 mM solution of D-isomer).
The total volume of the mixture was I00 ~al (5
micromoles of all amino acids). Mixture 2 was
similarly prepared but contained 80% L-isomer
and 20% D-isomer and mixture 3 contained 90%
L-isomer and 10% D-isomer. Mixture 4 con-
tained only L-isomers and the reagent blank 100
lal water. Each mixture and the reagent blank
was treated with 200 tll of 1% acetone solution of
FDAA (2 mg, 7.2 micromoles, the molar ratio of
FDAA to amino acid 1.4:1). 40 lal of 1 M-
NaHCO3 (40 micromoles) was added, the con-
tents mixed and then heated for 1 hour over a hot
plate at 30 ~ ~ with frequent mixing. After
cooling to room temperature, 20 ~tl of 2 M-HC1
(40 micromoles) were added, and the contents
dried in a vacuum desiccator over NaOH pellets.
The reaction residues, each containing 5 mi-
cromoles of total amino acids (except blank),
were dissolved in 0.5 ml DMSO affording 10 m M
solutions. Aliquots (10 ~tl, 100 nanomoles) of
each mixture and 10 ~1 of the 1:10 diluted
reagent blank were injected for HPLC under
conditions described in section
2.2.3.
0.5 A
0.3
tuO.1
0
Z
0.3
-0.1
0.3 ~
0.1
I I t I
250 350 450
WAVELENGTH (rim)
Figure I. Ultraviolet spectra of L- and D-diastereomers
of Ala in 25 mM-TEAP buffer, pH 3.0, and 50%
acetonitrile. A, D-AIa-DNP-L-Ala-NH2 (20 /aM solu-
tion): B, L-AIa-DNP-L-AIa-NH2 (10 /aM solution); C,
hydrolyzed reagent (HO-DNP-L-Ala-NH2), 30 /aM
solution).
the spectrum of HO-DNP-L-Ala-NH2 (the hy-
drolyzed reagent) has ~a~ at 264 nm 338 nm
(eM -- 1.5 x l04) and a shoulder at 380 mm. The
~, and the eM values are slightly different for
diastereomers of different amino acids and vary
slightly with the nature of the solvent used. The
spectra are stable if the solutions are kept in the
dark, otherwise, a gradual change occurs as a
result of a photochemical decomposition of the
absorbing chromophore.
3. RESULTS
3.1. Spectral characteristics of diastereomers
All ten diastereomers studied have very simi-
lar ultraviolet absorption spectra. A typical spec-
trum is shown for D- and L-diastereomers of Ala
in Figure 1. The spectra of both diastereomers
are very similar and are characterized by a L, ax at
338 nm(eM ~ 3.0 x 104) and 414 nm (cM -= 1.1
x I0'). The ratio of absorption at these two
wavelengths is 2.7 which is characteristic of a
-NH-DNP-NH- chromophore (2,3). In contrast,
3.2. Chromatographic characteristics of
diastereomers of Ala, Asp, Glu, Met and
Phe
The HPLC elution pattern of the ten diaste-
reomers of Ala, Asp, Glu, Met and Phe showed
good separation. The separation of D- and L-
diastereomers is best for Met followed by that of
Phe, Ala, Glu and Asp. The hydrolyzed reagent
appears as a sharp peak and is separated from all
the diastereomers.
Carlsberg Res. Commun. Vol. 49, p. 591-596, 1984 593
P. MARFEY: Determination of D-amino acids. II
4 8
10
11
3
9
r[
L. ~ ,___
2'0
'
ab
'
4b
ELUTION TIME ( MINUTES )
Figure 2. HPLC of mixture l containing L- and
o-diastereomers of Ala, Asp, Glu, Met, Phe in 60%
L- and 40% D-proportion. Peak l, L-Asp-DNP-L-AIa-
NH,, 17.68 min; Peak 2, L-Glu-DNP-L-Ala-NHz,
19.40 min; peak 3, D-Asp-DNP-L-AlaNH_,, 20.28 rain;
peak 4, L-AIa-DNP-L-AIa-NH,, 21.40 min; peak 5,
D-GIu-DNP-L-AIa-NH~, 22.71 min; peak 6, HO-DNP-
L-Ala-NH_, (hydrolyzed reagent) 24.48 min; peak 7,
D-Ala-DNP-L-AIa-NH,, 26.72 min; peak 8, L-Met-
DNP-L-AIa-NH,, 28.21 min; peak 9, D-Met-DNP-L-
AIa-NH_,, 34.66 min; peak 10, L-Phe-DNP-L-Ala-NH2,
35.82 rain; peak l l, D-Phe-DNP-L-Ala-NH2, 41.22
min. l0 p.l sample ( 100 nanomoles) of 10 mM solution
of all amino acid derivatives, containing 8 nanomoles
of each D-isomer and 12 nanomoles of each L-isomer,
was used for HPLC under conditions described in
section
2.2.3.
This is exemplified in Figure 2 which shows
the HPLC pattern of the amino acid mixture
containing 40% D-isomer after derivatization
with the FDAA reagent as described in section
2.2.4.
The hydrolyzed reagent is peak 6 and the
peaks corresponding to five D-isomers are: peak
3, D-Asp-DNP-L-Ala-NH2; peak 5, D-GIu-DNP-
L-Ala-NH2; peak 7, D-AIa-DNP-L-AIa-NH2;
peak 9, D-Met-DNP-L-Ala-NH2 and peak l l,
D-Phe-DNP-L-AIa-NH2. The chromatograms of
mixtures with 20%, 10% and 0% D-isomers,
respectively, are not shown but they are very
similar to the one shown in Figure 2, except that
peaks corresponding to D-isomers get pro-
gressively smaller for the 20% D- and the 10%
D-mixtures and the peaks for the L-isomers get
progressively larger. The chromatogram of the
reagent blank had only one peak (peak 6). The
quantitative evaluation of hydrolyzed reagent
peak gave a value of 1.4 x l04 HPLC area units
per nanomole. Using this value in evaluating
peak 6 in the chromatograms of the four amino
acid mixtures indicated that the amount of
hydrolyzed reagent found (27%) corresponded
to the excess reagent (28%) used for the derivati-
zation reaction. This result confirms that no
unexpected side reactions occur.
The amount of D-isomer in the three mixtures
was calculated from the HPLC area correspond-
ing to that particular isomer and there was a
linear relationship between % D-area and %
D-isomer in the three mixtures (Figure 3). The
slopes are not completely identical because eM
values at 340 nm vary slightly for different
diastereomers of the same amino acid.
~ / ~
,./i../,
,,,,o /,,
<{
20 "
/"
o
/
/
a
2'0
4'0 2b 4'o 2b 4b 2b 4b 2'o 4b
a~ % D ISOMER
Figure 3. Relationship between
%
HPLC area corresponding to D-isomer and % D-isomer present in mixture
I. A, D-Asp-DNP-L-Ala-NH2 (slope 1.06); B, D-GIu-DNP-L-AIa-NH~ (slope 1.10); C, D-AIa-DNP-L-Ala-NH2
(slope 1.03); D, D-Met-DNP-L-AIa-NH2 (slope 0.81); E, D-Phe-DNP-L-AIa-NH2 (slope 1.19). The % D-area was
calculated from: % D = [D-area/ (D-area + L-area)] x 100.
594 Carlsberg Res. Commun. Vol. 49, p. 591-596, 1984
P. MARFEY" Determination of D-amino acids. I1
NO=
F H=Nc,,H
4"
O;~N F CH; (~C-NH 2
FFDNP ,L L-AIa-NH2
NOz H H
I,~'~.'N~.c/
O,N "~ C~l; k'~- NH2
F i:/
F-DNP-L-AIa
NH 2
L-D
Ala
NO=
H H NO2 H H
O=N ~P'~ CH;x,.C-NH2 + O,N~ c~C~c-NH2
C~Is"~y-OH HO-r \CHs
O O
L-Ala NH2-DNP-L-Ala L-Ala
NH2-DNP-D-Ala
~L HPLC
Separation of two isomers
Figure 4. An outline of the reaction sequence used
for the synthesis of FDAA reagent and for the deri-
vatization of L- and D-isomers of amino acids. Condi-
tions for the synthesis and for the HPLC are given
in the text.
4. DISCUSSION
The reactions which were used to prepare
diastereomers of the five amino acids discussed
in this paper are outlined in Figure 4. The initial
reaction is between FFDNB and L-Ala-NH2 to
form F-DNP-L-AIa-NH2. Amide was chosen
because it is quite stable and apparently is not
easily racemized. The next step involves the
reaction with a mixture of L- and D-isomers of
any amino acid. The product is a mixture of
diastereomers which can easily be separated and
quantitated by HPLC. This method of resolu-
tion and quantitation of optical isomers is quite
flexible because in principle one can prepare a
reagent containing any optically active amino
acid in place of L-Ala-NH2 and use it for the
second reaction. It has the advantage over the
method of MANNING and MOORE (1) in that it
does not produce oligomeric products which can
be formed when an L-amino acid N-carboxyan-
hydride reacts with a mixture of L- and D-amino
acids. Another advantage is the availability of a
stable, highly absorbing chromophore, which
permits determination of diastereomers in the
nanomole range and the rapidity of deter-
mination inherent in the HPLC method.
Analysis of the chromatograms of the four
amino acid mixtures and of the experimental
reagent blank showed that the derivatization
reaction was quantitative. The expected amount
of the reagent was bound to the amino acids (see
Figure 3) and the excess recovered as a single
peak in quantitative yield.
It was of interest to find that L-diastereomers
are eluted from the column before D-diaste-
reomers. The reason for this behavior is pro-
bably due to a stronger intramolecular H-bond-
ing in D- than in L-diastereomer. One can expect
that the carboxyl group can hydrogen bond
either to an ortho-situated nitro group produc-
ing a 9-membered ring or, more likely, to the
carbonyl oxygen of the meta-situated L-Ala-NH_,
forming a 12-membered ring. Stronger H-bond-
ing in a D-diasteromer would produce a more
hydrophobic molecule which would be expected
to interact more strongly with the reverse-phase
column and thus have a longer retention time
than an L-diastereomer.
When one compares the differences in elution
times of the five pairs of diastereomers (see
Figure 2) one obtains the following order: Met
(7.1 min.) > Phe (5.4 min.) > Ala (5.2 rain.) >
Glu (3.2 min.) > Asp (2.5 min.). It is clear that
the nature of the amino acid side-chain is re-
sponsible for this behavior. The ionizable side-
chains of Asp and Glu decrease the separation
whereas neutral and hydrophobic side-chains
increase it.
The method ofquantitation of D-isomers in a
mixture is quite simple. It compares HPLC peak
areas of the two diastereomers of the same amino
acid in the same chromatogram. In this way, the
experimental conditions are the same for both
isomers.
The present method has so far been applied to
determination of D-isomers of five amino acids.
In principle, it can be applied to other amino
acids, but the separation of the diastereomers
may require different conditions for HPLC.
Carlsberg Res. Commun. Vol. 49, p. 591-596, 1984 595