Neutral, acidic, and basic derivatives of anthranilamide that confer different formal charge to reducing oligosaccharides

Article abstract of DOI:10.1016/j.carres.2003.10.020

To facilitate the use of oligosaccharides as analytical tools in biological studies, we have designed, synthesized, and conjugated to maltosaccharides a novel series of homologous small fluorescent moieties that differ in formal charge. These moieties are

Full text of DOI:10.1016/j.carres.2003.10.020

Carbohydrate
RESEARCH
Carbohydrate Research 339 (2004) 221–231
Neutral, acidic, and basic derivatives of anthranilamide that confer
different formal charge to reducing oligosaccharidesq
Darren Locke,a,*, Carville G. Bevans,b, Lai-Xi Wang,c, Ye Zhang,d, Andrew L. Harrisa
and Yuan C. Leed
aDepartment of Pharmacology and Physiology, New Jersey Medical School, UMDNJ, 185 South Orange Avenue,
Newark, NJ 07103, USA
bDepartment of Structural Biology, Max Planck Institute of Biophysics, Marie-Curie-Strasse 15, Frankfurt am Main 60439, Germany
cInstitute of Human Virology, University of Maryland Biotechnology Institute, University of Maryland,
725 W. Lombard Street, Baltimore, MD 21201, USA
dDepartment of Biology, Johns Hopkins University, 3400 North Charles St, Baltimore, MD 21218, USA
Received 6 August 2003; received in revised form 16 October 2003; accepted 21 October 2003
Abstract—To facilitate the use of oligosaccharides as analytical tools in biological studies, we have designed, synthesized, and
conjugated to maltosaccharides a novel series of homologous small fluorescent moieties that differ in formal charge. These moieties
are amide derivatives of anthranilic acid: uncharged N-(2-aminobenzoyl)glycinamide (ABGlyAmide; 2), acidic N,N-dimethyl-N0-(2-
aminobenzoyl)ethylenediamine (ABGlyDIMED; 3), and basic N-(2-aminobenzoyl)glycine (ABGly; 1). Routes for synthesis and
optimal reaction conditions for glycoconjugation by conventional reductive amination are presented, as is the compatibility of these
adducts with common analytical and preparative chromatographic methods, including RP-HPLC and HPAEC-PAD. These novel
anthranilic acid derivatives confer both fluorescence and defined charge to oligosaccharides, and so enhance the repertoire of
chromatographic and analytical methods for which anthranilic acid can be used. Furthermore, because glucosaccharides have rigid
solution structure, these small fluorescent adducts with different formal charge are ideal tools for molecular sizing studies of
membrane pores.
ꢀ 2003 Elsevier Ltd. All rights reserved.
Keywords: Maltosaccharide; Anthranilic acid; Channel; Reductive amination; Carbohydrate; Glycoconjugate
1. Introduction
tools is hampered by the specialized methodologies
required for their high-resolution detection and analysis.
This problem can be overcome by derivatization with
fluorescent moieties, which enables easy and quantita-
tive detection. However, for some applications it would
be advantageous to make additional modifications, such
as addition of charge, as part of the same derivatiza-
tion. We describe here the synthesis and characterization
of novel anthranilic acid derivatives that are used to
generate acidic, basic, and neutral glycoconjugates.
These adducts are easily detectable by fluorometry,
readily purified, and the conjugations do not signifi-
cantly alter the structure or dimensions of the oligo-
saccharides.
Carbohydrates are excellent tools for biological and
biophysical investigation, owing to the vast number of
structural combinations and permutations available
from a small number of monosaccharides. The useful-
ness of underivatized carbohydrates as investigative
qCharged derivatives of anthranilic acid amide.
* Corresponding author. Tel.: +1-973-972-1619; fax: +1-973-972-4554;
e-mail: lockeda@umdnj.edu
Authors contributed equally.
0008-6215/$ - see front matter ꢀ 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2003.10.020

222
D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
To characterize the permeability pathway of large
2. Results
membrane channels, it is desirable to have a size-
indexed set of fluorescent, unreactive molecules of
known dimension and similar chemistry. The rigidity
conferred upon maltosaccharides by internal hydrogen
bonding makes them particularly good candidates for
this application when labeled with a small fluorescent
group. For channel studies it is also desirable to uti-
lize charged as well as uncharged probes. Therefore
our goal was to develop a set of compounds easily
conjugated to reducing oligosaccharides that would
confer both fluorescent detectability and defined
charge.
Reductive amination is a common method for deriv-
atizing reducing sugars. The feasibility of reductive
amination for carbohydrate labeling was first demon-
strated using 2-aminopyridine to separate a homologous
series of glucosaccharides by RP-HPLC.1 Such deriva-
tized oligosaccharides are uncharged at neutral pH and
have been used to determine the pore diameter of
intercellular membrane channels.2 Under strongly acidic
conditions 2-aminopyridine carries a weak positive
charge, enabling the same series to be separated by CE
according to their charge–mass ratio.3 Addition-
ally, stoichiometric labeling of oligosaccharides offers
an excellent method for quantitation while the agly-
cone provides stability against strongly acidic and
alkaline conditions, under which free saccharide may
degrade.
2.1. Preparation of anthranilamide derivativesà
The designed 2-aminobenzamide compounds (anthra-
nilamide derivatives) were prepared from 2-nitrobenzoic
acid via 2-nitrobenzoyl chloride, as summarized in Fig-
ure 1. Starting from 2-nitrobenzoic acid was more satis-
factory than having to N-protect the amino group of
anthranilic acid. Briefly, coupling of 2-nitrobenzoic acid
with glycine benzoyl ester gave N-(2-nitrobenzoyl)glycine
benzyl ester (4, 72%). Compound 4 was converted into its
corresponding glycinamide 5 with 82% yield, through
treatment with concentrated ammonium hydroxide in
methanol. Coupling of 2-nitrobenzoyl chloride with N,N-
dimethylethylenediamine provided N,N-dimethyl-N0-(2-
nitrobenzoyl)ethylenediamine (6, 67%). Finally, palla-
dium-catalyzed hydrogenation of the 2-nitro derivatives
4, 5, and 6 gave the corresponding 2-amino derivatives 1
(ABGly; N-(2-aminobenzoyl)glycine), 2 (ABGlyAmide;
N-(2-aminobenzoyl)glycinamide), and
3 (ABGlyDI-
MED; N,N-dimethyl-N0-(2-aminobenzoyl)ethylenedia-
mine) in 55, 88, and 91% yields, respectively. Structures
were confirmed by NMR (Fig. 2).
2.2. Absorption and fluorescence emission spectra of the
free and conjugated anthranilamides
However, 2-aminopyridine cannot be easily modified
to carry positive or negative charge, and so an alterna-
tive moiety was desired. Other charged and uncharged
labels4 have been developed, or applied, to variously
enable the separation and analysis of carbohydrate by
chromatographic (LC, CE), electrophoretic (FACE),
and ion-mass (MS, NMR) techniques. Of those labels
characterized, anthranilic acid5 fulfills most of the
desired criteria, being inexpensive and commercially
available,6 highly fluorescent,7 and rapidly conjugated
to reducing saccharide under weak acidic conditions
that are not suitable for conjugation of many primary
amines.5;7 Most important, anthranilic acid can be
modified to have a single positive or negative charge.
Furthermore, anthranilic acid is compatible with the
majority of chromatographic techniques employed
for carbohydrate separations.
The availability of a new homologous series of
small fluorescent labeling compounds, neutral,
acidic, and basic amide derivatives of anthranilic
acid, will improve the repertoire of chromatographic
and analytical methods for which anthranilic acid
can be used with minimal disturbance to isolation,
purification, and structural determinations of mono-
and oligosaccharides, and that will find immediate
application in studies of membrane protein chan-
nels.
Free label (1–3) absorbance wavelength maxima were
between 312.5 and 317 nm (median 314.8 nm) in water
and the range shifted to longer wavelengths, 314.5–
318.5 nm (median 316.5 nm) in 6% (v/v) acetic acid.
Absorbance in 6% (v/v) acetic acid was attenuated to
30–48% of that in water. Optimal fluorescence wave-
length for the excitation of free label in water was
325 nm with emission wavelengths at 410, 415, and
420 nm for the anionic, neutral, and cationic labels,
respectively. Labeled oligosaccharide mixtures in water
exhibited maximum absorbance wavelengths that were
red-shifted (range 330.5–336.5 nm, median 333.5 nm)
relative to those of the free labels in solution. Absor-
bance increased for the labeled maltosaccharide mix-
tures in water relative to the free labels (36% for negative
labeled (1, ABGly), 5% for neutral labeled (2, AB-
GlyAmide), 34% for positive labeled (3, ABGlyDI-
MED)); Similarly, optimal fluorescence excitation of the
conjugated labels red-shifted (range 342.5–344.5 nm)
with emission maxima around 430 nm (range 431.5–
432.5 nm) (Table 1).
à
The formulas in Figures 3–7 all imply the
the sugar portion.
D-gluco configuration in

D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
223
Figure 1. Reaction scheme. The new 2-aminobenzamide derivatives were prepared from 2-nitrobenzoic acid via 2-nitrobenzoyl chloride. Hydro-
genation of the 2-nitro derivatives gave the corresponding 2-amino-derivatives 1 (ABGly; N-(2-aminobenzoyl)glycine), 2 (ABGlyA; N-(2-amin-
obenzoyl)glycinamide), and 3 (ABGlyDIMED; N,N-dimethyl-N0-(2-aminobenzoyl)ethylenediamine).
2.3. Oligosaccharide labeling
2.4. Recovery of anthranilamide glycoconjugates from
aqueous solution by amine adsorption
A maltose through maltohexaose mixture ([4)-a-D-Glcp-
(1-]2–6) labeled with the anionic ABGly-labeled (1, N-(2-
aminobenzoyl)glycine) or uncharged ABGlyAmide-
labeled (2, N-(2-aminobenzoyl)glycinamide) oligosac-
charides completely separated from the excess labeling
reagents that eluted more slowly through a Sephadex
G-10 column (Fig. 3). The cationic ABGlyDIMED
(3, N,N-dimethyl-N0-(2-aminobenzoyl)ethylenediamine)
derivatives were not as easily separated from free label
by this method, however fractions could be judiciously
chosen to contain labeled maltosaccharides free of
unreacted label. Retention times of eluted peaks for
Conjugated maltosaccharides were retained on amino-
propyl solid-phase columns equilibrated with high v/v
concentrations of acetonitrile in water. Glycoconjugates
could be eluted directly with column eluents for chro-
matographic separations by RP-HPLC and/or HPAEC
(Fig. 4). Because of their simplicity, relatively low cost,
and high capacity for binding, aminopropyl silica col-
umns are particularly suitable for quickly de-salting
label, or for recovery of a wide range of quantity of
derivatives, without loss.
single sugar labeling reactions ([4)-a-
bracketed the retention times spanned by the labeled
oligosaccharide mixture ([4)-a- -Glcp-(1-]2–6) for each
D
-Glcp-(1-]n 6 6)
2.5. Separation using GCC and Carbopac PA-1 column
D
anthranilamide derivative. Labeling efficiencies for (1)
ABGly-, (2) ABGlyAmide-, and (3) ABGlyDIMED-
derivatized oligosaccharides were 75, 100, and 50%,
respectively, of the theoretical maxima of one-to-four
for a fivefold molar excess of label reacted initially with
reducing oligosaccharide. Addition of charge-bearing
side groups adversely affected labeling efficiency,
assuming the quantum efficiency of free label and
oligosaccharide-linked label were equal in each case.
C18 columns do not retain anthranilamide-derivatized
saccharides and therefore a more hydrophobic chemis-
try, namely, a graphitized carbon column (GCC), is
required for their separation. Optimal running condi-
tion, of several organic solvents and acid/base additives
tested, was 30% (v/v) acetonitrile containing 0.05% (v/v)
TFA. Labeled oligosaccharide series with different
degrees of polymerization (DP) could be separated for
the negatively charged (1, ABGly) and neutral (2,

224
D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
Figure 2. NMR spectra. Preparations as summarized in Methods and Figure 1 (Reaction scheme). In (a), ABGly (1), N-(2-aminobenzoyl)glycine,
an off-white solid: 1H NMR (300 MHz, Me2SO-d6 + 5% D2O + 0.03% Me4Si): d 7.510 (d, 1H, J 9.93 Hz, H-6), 7.137 (t, 1H, J 9.65 Hz, H-4), 6.691 (d,
1H, J 9.69 Hz, H-3), 6.521 (t, 1H, J 9.69 Hz, H-5), 3.836 (s, 2H, J 19.18 Hz, NCH2). In (b), ABGlyAmide (2), N-(2-aminobenzoyl)glycinamide, an off-
white solid. 1H NMR (300 MHz, Me2SO-d6 + 5% D2O + 0.03% TMS): d 8.320 (d, 1H, J 10.30 Hz, H-6), 7.511 (t, 1H, J 10.89 Hz, H-4), 7.123 (T, 1H,
J 11.43 Hz, H-5), 6.991 (d, 1H, J 11.06 Hz, H-3), 3.743 (s, 2H, J 26.65 Hz, NCH2). In (c), ABGlyDIMED (3), N,N-dimethyl-N0-(2-amino-
benzoyl)ethylenediamine, a colorless syrup: 1H NMR (300 MHz, CDCl3 + 2% D2O + 0.03% Me4Si): d 7.534 (d, 1H, J 5.31 Hz, H-5), 7.222 (t, 1H,
J 4.35 Hz, H-4), 6.702–6.661 (m, 2H, H-3, 5), 3.635 (t, 2H, J 10.58 Hz, NCH2), 2.801 (t, 2H, J 10.42 Hz, CH2NMe2), 2.504 [s, 6H, J 28.34 Hz,
N(CH3)2].

D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
225
Table 1. Absorbance and fluorescence emission spectra for the three anthranilamide labels in de-ionized water and acetic acid
Free label
nM
Solvent
kmax (nm)
Amax (OD)
10ꢀ3 · emax
kexcitation (nm)
kemission (nm)
(cmꢀ1 Mꢀ1
)
ABGly (1)
ABGly (1)
138
138
138
138
Water
0.1 M acetic acid
Water
312.5
315.0
315.5
314.5
317.0
318.5
0.3579
0.1117
0.3888
0.1179
0.3078
0.1468
2593
809
325
325
325
410
415
420
ABGlyAmide (2)
ABGlyAmide (2)
2817
854
0.1 M acetic acid
Water
0.1 M acetic acid
ABGlyDIMED (3) 138
ABGlyDIMED (3) 138
2230
1064
Glycoconjugate
mg/mL
ABGly (1)
ABGlyAmide (2)
ABGlyDIMED (3)
0.1
0.1
0.1
Water
Water
Water
330.5
336.5
335.5
0.4856
0.4067
0.4128
3519
2947
2991
342
343
344
431
432
433
Figure 3. Purification of reductive amination products by gel filtration. Trace shows the elution profile (h) after application of G2, G6, and G2–G6
reductive amination reaction mixture (entire reaction volume of 0.2 mL diluted to 2 mL with 6% (v/v) acetic acid) to a Sephadex G-10 column
(1.5 · 50 cm). Samples were eluted with 6% (v/v) acetic acid. The asymmetric peak of labeled maltosaccharides (left) eluted before free, unreacted label
(right peak). Reaction efficiency was determined from the ratio of the areas under the chromatogram peaks that represented the fluorescence of
labeled oligosaccharide and unreacted free label.
ABGlyAmide) series; labeled sugars were eluted in order
of increasing DP from the GCC. The separation of the
positively charged (3, ABGlyDIMED) derivatives was
not satisfactory and there was no significant improve-
ment in resolution upon heating or cooling the column.
Resolution of ABGlyDIMED derivatives (3) were
improved by using 30% (v/v) acetonitrile including
4–10 mM sodium hydroxide, although quantification of
the peaks with different DP was not as reproducible as
when using acidified acetonitrile. Tailing of the peaks
observed were likely due to the alkali incompatibility of
the HPLC system used (Fig. 5) and by the heteroge-
neous ionic charges that were suppressed under acidic
conditions. Free fluorescent reagents remained bound to
the GCC columns and their elution required increasing
concentrations of acetonitrile.

226
D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
Figure 4. SPE recovery of glycoconjugates on aminopropyl resin. Free label and/or derivatized glycoconjugates can be purified in short time and with
excellent recovery by solid-phase extraction using an amine-modified resin. Bindings are in pure or at high acetonitrile concentrations. The molecular
basis is likely hydrogen bonding, which can be disrupted through increasing concentrations of water by 10% (v/v) increments.
High-resolution separation of homologous maltosac-
charides has been obtained under strongly alkaline
conditions by HPAEC with peculiar anion-exchange
resins. Methods for separation of neutral (2, ABGlyA-
mide), negatively (1, ABGly), and positively charged (3,
ABGlyDIMED) maltosaccharides were developed using
HPAEC and the Carbopac PA-1 column. Labeled sug-
ars were eluted in order of increasing DP from the
Carbopac PA-1 column. Neutral (2) and positively
charged (3) derivatives were well separated using a linear
(2–25% (v/v) 1 M, 25 min) acetate gradient (Fig. 6),
however baseline separations of the negatively charged
(1) derivatives could only be achieved when pushing
agent was substituted by nitrate (10–35% (v/v) 0.2 M,
25 min), which has a higher resolution with this column
over acetate gradients for separation of polysialic acids
that have large amounts of negative charge (Fig. 7). Free
fluorescent labels eluted early from the column.
bohydrate in line with current sensitivity in molecular
biology. The full potential of these analyses are being
realized by GC and CE with their coupling to mass
spectrometry and NMR, and also include electropho-
retic FACE analysis for simple and complex carbohy-
drates. Each strategy has benefited from, and can be
hindered by, the ability of carbohydrates to undergo
derivitization reactions with photometric or fluoromet-
ric labels. Labeling by reductive amination using a
combination of aromatic amines or monosubstituted
aminobenzene derivatives and reducing reagents is most
frequently employed. For many of these downstream
methodologies, the labels confer the charge required for
their effective resolution, which makes it possible to
either adjust mobility in the electrophoretic process or
provide greater increases in sensitivity following
MALDI and electrospray ionization.
This work describes the synthesis of neutral (1,
ABGlyAmide; N-(2-aminobenzoyl)glycinamide), nega-
tive (2, ABGly; N-(2-aminobenzoyl)glycine), and posi-
tively charged (3, ABGlyDIMED; N,N-dimethyl-N0-(2-
amino-benzoyl)ethylenediamine) derivatives of anthra-
nilic acid, one of the smallest and most widely applicable
coupling reagents. These new labeling compounds show
3. Discussion
Recent developments in carbohydrate detection after
analytical separations brings analytical limits for car-

D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
227
Figure 5. Separation of glycoconjugates using HPLC-GCC. Separation of the glycoconjugate homologues is accomplished isocratically within
20 min using a routine HPLC system and a graphitized carbon column. The order of resolution was negative (1, ABGly) > neutral (2, ABGly-
Amide) > positive (3, ABGlyDIMED) derivatives with labeled sugars eluted in order of decreasing DP.
good promise in their compatibility for all downstream
analytical methodologies. The addition of both a small
fluorescent tag and formal charge, if desired, can be
achieved without significantly altering the overall size or
structure of the labeled oligosaccharides.
which the extensive chemical and structural character-
ization of oligosaccharides (e.g., indexed size, relative
nonreactivity) are useful. Several classes of membrane
channels are permeable to small mono- and oligosac-
charides including mechanosensitive channels,8;9 certain
puringeric receptors,10–12 and anion channels,13 bacterial
and mitochondrial porins,14 and intercellular junctions.2
One immediate area of application is for the investi-
gation of the permeant properties of protein pores in

228
D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
Figure 6. Acetate gradient elution on a PA1 column for ABGly-Amide and -DIMED labeled maltosaccharide. Neutral (2, ABGlyAmide) and
positively charged (3, ABGlyDIMED) derivatives are well separated using a linear acetate gradient within 20 min.
For these channels, determination of pore properties has
often relied on the use of fluorescent tracer molecules of
diverse size, charge and charge distribution, molecular
shape, and chemical structure. Examples include Lucifer
Yellow, calcein, fluorescein and its derivatives, biotin
and its derivatives, small peptides, propidium iodide,
and the Alexa dyes. Because of the differences among
these reagents, it has been difficult to attribute differ-
ences in their channel permeability to differences in their
size, charge, or to their specific chemical interactions
with the channel being studied. Meaningful analysis of
large pore permeation requires a set of reagents that do
not have such variability, and that have known and well-
characterized chemical and structural features. The
ability to simultaneously confer simple detection (by
fluorescence) and to alter charge as desired without
significantly altering the overall size and structure allows
the use of conjugated oligosaccharide reagents for this
purpose.
different from the 2-aminopyridine glycoconjugates
previously used,2 and may block connexin pores.15 Thus
these reagents will be useful not only as probes of the
roles of size and charge in channel permeability, but also
as tools for investigating and modulating channel
function.
It is hoped that these anthranilamide labels (1,2,3),
and the straightforward methods for conjugation, clean-
up, and analysis described here will lead to the appli-
cation of glycoconjugate reagents to a host of biological
issues for which appropriate reagents have not been
previously available.
4. Experimental
4.1. Materials
Maltose oligosaccharide mixture, ([4)-a-
was obtained from Pfanstiehl Laboratories (Waukegan,
IL, USA). Maltose ([4)-a- -Glcp-(1-]2), G2, purity 99%
w/w by HPAE-PAD), maltotriose ([4)-a- -Glcp-(1-]3),
G3, 98%), maltotetraose ([4)-a- -Glcp-(1-]4), G4, 95%),
maltopentaose ([4)-a- -Glcp-(1-]5), G5, 97.5%), and
D-Glcp-(1-]2–6)
Indeed, these new charged, fluorescent labels have
been used to investigate the pore selectivity of intercel-
lular channels, with intriguing and unexpected results.
The derivatized maltosaccharides appear to interact
with connexin channels in ways that are substantially
D
D
D
D

D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
229
Figure 7. A comparison of acetate versus nitrate pushing agent on the elution of ABGly-labeled saccharide from a PA1 column. High resolution of a
homologous series of ABGly derivatives (1) is achieved by favoring nitrate over acetate pushing agent.
maltohexaose ([4)-a-
D
-Glcp-(1-]6), G6, 95%) were from
solved in anhydrous tetrahydrofuran (15 mL) and
cooled to 0 ꢁC. Triethylamine (2.1 mL, 15 mmol) was
added. After reaction for 5 h at 0 ꢁC, the mixture was
diluted with EtoAc (80 mL) and washed successively
with saturated NaHCO3, 5% (v/v) HCl, and water. The
organic layer was dried (Na2SO4), filtered, and evapo-
rated. The residue was crystallized from toluene to give
N-(2-nitrobenzoyl)glycine benzyl ester (4) (3.5 g, 72%) as
white crystals: mp 76–77 ꢁC; 1H NMR (300 MHz,
CDCl3): d 8.107–7.552 (m, 4H, H-3, 4, 5, 6), 7.387 (s,
5H, C6H6), 6.401, d, 1H, J 5.8 Hz, NH), 5.243 (s, 2H,
CH2C6H5), 4.319 (d, 2H, J 5.8 Hz, NCH2).
Catalytic hydrogenation of 4 (472 mg, 1.5 mmol) was
carried out in 95% EtOH (20 mL) in the presence of 10%
palladium on charcoal (100 mg) under a hydrogen
atmosphere. After 16 h, the catalyst was removed by
filtration through a Celite pad. The filtrate was evapo-
rated to give the crude product (250 mg), which was
further refined through a column (1 · 15 cm) of Sepha-
dex LH-20. This column was re-equilibrated and eluted
with CHCl3–95% EtOH (1:1, v/v) to afford N-(2-amino-
benzoyl)glycine (1) as an off-white solid: mp 118–120 ꢁC.
Sigma Chemical (St. Louis, MO). Borane–diethylamine
was from Aldrich Chemical (Milwaukee, WI, USA).
Sephadex G-10, LH-20 was purchased from Amersham
Biosciences (Piscataway, NJ, USA).
4.2. Preparation of labels with different electrostatic
charges
4.2.1. ABGly (1), N-(2-aminobenzoyl)glycine. To a
solution of 2-nitrobenzoic acid (1.7 g, 10 mmol) in
anhydrous tetrahydrofuran (10 mL) were added oxalyl
chloride (1.75 mL, 20 mmol) and a catalytic amount of
N,N-dimethylformamide (100 lL), sequentially and
dropwise. The mixture was stirred at room temperature
for 2 h and then the solvent and excess oxalyl chloride
removed by evaporation under diminished pressure. The
remaining syrup was co-evaporated with toluene
(2 · 10 mL) to give the nitrobenzoyl chloride as colorless
syrup, which was then used without further purification.
The syrup (1.9 g) and glycine benzyl ester (toluene-
sufonate salt) (4.0 g, 12 mmol) were successively dis-

230
D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
4.2.2. ABGlyAmide (2), N-(2-aminobenzoyl)glycinamide.
N-(2-Nitrobenzoyl)glycine benzoic ester (4) (472 mg,
1.5 mmol) was dissolved in MeOH (8 mL) and treated
with concentrated NH4OH (8 mL) for 10 h at 20 ꢁC. The
solution was evaporated and the solid residue crystal-
lized from ethanol to give N-(2-nitrobenzoyl)glycinam-
ide (5) (276 mg, 82%) as a white solid: mp 193–195 ꢁC;
1H NMR (300 MHz, Me2SO-d6+5% D2O): d 8.012 (d,
1H, J 7.8 Hz, H-3), 7.788 (t, 1H, J 7.3 Hz, H-5), 7.707–
7.364 (M, 2H, H-4, 6), 3.800 (s, 2H, NCH2). Catalytic
hydrogenation of 5 was performed as for the prepara-
tion of 1 to afford N-(2-aminobenzoyl)glycinamide (2)
(170 mg, 88%).
sugars were pooled and evaporated under reduced
pressure using a rotary evaporator at 60 ꢁC with
repeated addition of MeOH. The residue was dissolved
in double-distilled, nylon-filtered water to a final con-
centration of ꢁ10 mg/mL and frozen.
Where required, chart tracing of column eluents were
digitized using Un-Scan-It (Silk Scientific, Orem, UT,
USA). PeakFit v3.18 (Jandel Scientific, Sausalito, CA,
USA) was used to fit independent, nonsymmetric
Haarhoff Van der Linde functions to the data using the
Marquardt–Levenburg NonLinear Least Squares
Algorithm. All fits successfully converged.
4.2.3. ABGlyDIMED (3), N,N-dimethyl-N0-(2-amino-
benzoyl)ethylenediamine. 2-Nitrobenzoyl chloride, pre-
pared as described previously from 2-nitrobenzoic acid
(1.67 g, 10 mmol), was reacted with N,N-dimethylethyl-
enediamine (0.92 mL, 13 mmol) and Et3N (2.1 mL,
15 mmol) in tetrahydrofuan (10 mL) for 5 h at 0 ꢁC. The
reaction was evaporated and the residue partitioned
between chloroform (50 mL) and aqueous NaHCO3
(10 mL). The organic layer was washed with water
(2 · 10 mL), dried (Na2SO4) sodium sulfate, and evapo-
rated. The solid was crystallized from hexane–EtOAc to
give N,N-Dimethyl-N0-(2-nitrobenzoyl)ethylenediamine
(6) (1.59 g, 67%): 1H NMR (300 MHz, CDCl3 + 2%
D2O): d 8.063 (d, 1H, J 7.8 Hz, H-3), 7.688 (t, 1H, J
7.4 Hz, H-5), 7.643–7.523 (m, 2H, H-4, 6), 3.562 (t, 2H,
J 5.6 Hz, NHCH2), 2.623 (t, 2H, J 5.6 Hz, CH2NMe2),
2.317 [s, 6H, N(CH3)2].
4.4. Absorption and fluorescence emission spectra
Absorption and fluorescence emission spectra were mea-
sured for the labeled oligosaccharides (Table 1) using a
Varian Cary 3C UV vis spectrometer and an Eclipse
spectrofluorometer (Varian, Palo Alto, CA, USA).
4.5. Separation of labeled oligosaccharides using graph-
itized carbon column (GCC) chemistry
Separation with a porous graphitic carbon column
(Shandon Hypercarb 100 · 4.6, 7 lm; Alltech Associates,
Deerfield, IL, USA) was performed using a HPLC sys-
tem consisting of Gilson piston pumps model 302 (Gil-
son Medical Electronics Inc., Middleton, WI, USA) and
a Rheodyne 7125 injector. Detection was by a model LS
40 fluorescence detector (Perkin–Elmer Ltd., Beacons-
field, Buckinghamshire, UK).
A 30% (v/v) acetonitrile (J.T. Baker Inc., Philipsburg,
NJ, USA) solution containing 0.05% (v/v) TFA (Fluka
Chemical Corp., Ronkonkoma, NY, USA) was used as
the mobile phase. Separation was performed isocrati-
cally under ambient temperature for all oligosaccharide
derivatives of neutral, acidic, and basic anthanilamides.
Catalytic hydrogenation of 6 (474 mg, 2.0 mmol) was
performed in EtOH (15 mL) as described for the prep-
aration of 1 and the product was purified by column
chromatography on silica gel using 7:1 (v/v) CHCl3–
EtOH as the eluent to give N,N-dimethyl-N0-(2-amino-
benzoyl)ethylenediamine (3) (369 mg, 91%) as a colorless
syrup.
4.3. Labeling of oligosaccharides by reductive amination
4.6. Separation of labeled oligosaccharides using
Carbopac PA-1 chromatography
Approximately 1 mg of sugar, individually or a ([4)-a-D-
Glcp-(1-]2–6) maltosaccharide mixture, were dissolved in
100 lL of 6% (v/v) acetic acid by heating at 60 ꢁC, fol-
lowed by vortexing. A fivefold molar excess of the
labeling agent was added as solid, heated, and vortexed
into solution. The borane reductant (10 lmol) was
added, vortexed, and the reaction allowed to proceed for
2 h at 60 ꢁC, after which the reaction mixture was cooled
to room temperature.
A Sephadex G-10 column (1.5 · 50 cm), equilibrated
in 6% (v/v) HOAc, was used to resolve unreacted free
label from conjugate at an elution flow rate of 0.5 mL/
min. Three mL fractions were collected and 200 lL
aliquots monitored by absorbance ratio at 280 and
310 nm. Fractions eluting early that contained labeled
Separation was achieved by high-performance anion-
exchange chromatography (HPAEC) using a Dionex
Bio LC 4000 system and Carbopac PA-1 (250 · 4 mm)
column (Dionex Corp., Sunnyvale, CA, USA). Detec-
tion was made with either a pulsed amperometric
detector (PAD) equipped with a gold working electrode
or with a LS 40 fluorescence detector, or by both
detectors in tandem. The mobile phase buffers con-
tained 0.1 M NaOH and one of the following pushing
agents; 1 M sodium acetate in a linear gradient from 2%
to 50% (v/v) over 25 min for oligosaccharides labeled
with neutral and basic anthranilamide derivatives, or
0.2 M sodium nitrate for negatively charged maltosac-

D. Locke et al. / Carbohydrate Research 339 (2004) 221–231
231
2. Bevans, C. G.; Kordel, M.; Rhee, S. K.; Harris, A. L.
J. Biol. Chem. 1998, 273, 2808–2816.
3. Nashabeh, W.; el Rassi, Z. J. Chromat. A 1992, 600, 279–
287.
4. Anumula, K. R. Anal. Biochem. 2000, 283, 17–26.
5. Bigge, J. C.; Patel, T. P.; Bruce, J. A.; Goulding, P. N.;
Charles, S. M.; Parekh, R. B. Anal. Biochem. 1995, 230,
229–238.
6. Chen, F. T.; Dobashi, T. S.; Evangelista, R. A. Glyco-
biology 1998, 8, 1045–1052.
7. Anumula, K. R.; Dhume, S. T. Glycobiology 1998, 8, 685–
694.
8. Ohata, H.; Tanaka, K.; Momose, K. Life Sci. 1999, 65,
297–304.
9. Cemerikic, D.; Sackin, H. Am. J. Physiol. 1993, 264,
F697–714.
10. Van Calenbergh, S.; von Frijtag Drabbe Kunzel, J. K.;
Blaton, N. M.; Peeters, O. M.; Rozenski, J.; Van Aerschot,
A.; De Bruyn, A.; De Keukeleire, D.; Herdewijn, P. J.
Med. Chem. 1997, 40, 3765–3772.
charides using a linear gradient of 10–35% (v/v) over
25 min.
4.7. Solid-phase recovery on aminopropyl column
A simple method for the efficient recovery of anthra-
nilamide derivatives was through binding to aminopro-
pyl solid-phase extraction cartridges (Sep-Pac; Waters
Corp., Milford, MA, USA) that were pre-equilibrated in
high acetonitrile concentrations (aq) to facilitate bind-
ing. Elution from the solid-phase resin was accom-
plished by lowering acetonitrile concentrations to below
ꢁ40% (v/v) (aq), or with water elution.
Acknowledgements
11. Sasakawa, N.; Nakaki, T.; Yamamoto, S.; Kato, R.
J. Neurochem. 1989, 52, 441–447.
This work was supported by National Multiple Sclerosis
Society grant RG3324A1/T to CGB and NIH grants
GM36044 and GM61406 to ALH and DK009970 to
YCL.
12. Hellman, B.; Gylfe, E.; Bergsten, P.; Grapengiesser, E.;
Lund, P. E.; Berts, A.; Tengholm, A.; Pipeleers, D. G.;
Ling, Z. Diabetologia 1994, 37(Suppl 2), S11–20.
13. Desai, S. A.; Bezrukov, S. M.; Zimmerberg, J. Nature
2000, 406, 1001–1005.
References
14. Mukhopadhyay, S.; Basu, D.; Chakrabarti, P. J. Bacteriol.
1997, 179, 6205–6207.
1. Hase, S.; Ikenaka, T.; Matsushima, Y. Biochem. Biophys.
Res. Commun. 1978, 85, 257–263.
15. Locke, D.; Wang, L.-X.; Zhang, Y.; Bevans, C.; Lee, Y. C.;
Harris, A. Biophys. J. 2003, 84, 525a.