Non-natural glycosphingolipids and structurally simpler analogues bind HIV-1 recombinant gp120

Article abstract of DOI:10.1016/S0968-0896(01)00325-X

Interactions of recombinant gp120 (rgp120) with non-natural glycosphingolipids (GSLs) and structurally simpler analogues have been studied using a competitive adhesion assay. Conjugates of cellobiosyl ceramide and melibiosyl ceramide were synthetically pr

Full text of DOI:10.1016/S0968-0896(01)00325-X

Bioorganic & Medicinal Chemistry 10 (2002) 625–637
Non-Natural Glycosphingolipids and Structurally Simpler
Analogues Bind HIV-1 Recombinant Gp120
Katherine D. McReynolds,y Abhijit Bhat,{ John C. Conboy,x S. Scott Saavedra
and Jacquelyn Gervay-Hague*
Department of Chemistry, The University of Arizona, Tucson, AZ 85721, USA
Received 10 May 2001; accepted 4 September 2001
Abstract—Interactions of recombinant gp120 (rgp120) with non-natural glycosphingolipids (GSLs) and structurally simpler analogues
have been studied using a competitive adhesion assay. Conjugates of cellobiosyl ceramide and melibiosyl ceramide were synthetically
prepared as water-soluble GSL analogues. These ligands were screened against a panel of biologically relevant analogues, and the results
show that their interactions with rgp120 are comparable to natural cellular receptors. Glycolipid interactions with rgp120 were
probed further by the synthesis and testing of structurally simpler analogues that were obtained by reductive amination of lactose,
cellobiose, and melibiose with a biotinylated amino ethylene glycol moiety. RGp120 did not recognize conjugates lacking a lipid
component. However, palmitoylation of the secondary amino alditols yielded compounds with comparable rgp120 affinity to the natural
cellular receptor, galactosyl ceramide (GalCer). Taken together, the SAR showed that both a hydrophobic and a hydrophilic compo-
nent are required for rgp120 recognition. Moreover, structural variability in the carbohydrate headgroup did not significantly alter
rgp120 recognition indicating that this interaction is not highly specific. # 2002 Elsevier Science Ltd. All rights reserved.
Introduction
HIV that initiates viral infection into host cells. The
principal receptor for gp120 is a T-cell associated protein
called CD4. However, HIV is known to infect CD
negative cells such as neuronal cells,3 colonic epithelial
cells,4 and vaginal epithelial cells through a recognition
process involving GalCer.5
Glycosphingolipids are the principal glycolipids in
mammalian cells.1 They are present on the outer leaflet
of plasma membranes where they are anchored through
ceramide (1) functionalities, which are N-acylated
sphingosines (2) or sphingolipids (Fig. 1). The carbohy-
drate marker is attached to the terminal hydroxyl of
ceramide. The structures and distribution of GSLs vary
widely and they serve a number of different biological
functions.2 GalCer (3) is a simple GSL with a b-linkage
to ceramide. It has been identified as an essential receptor
component for gp120, the cell-surface glycoprotein of
Several investigations aimed at identifying other natural
GSLs that are recognized by gp120 have been reported.
Glucosyl ceramide (GlcCer 4), lactosyl ceramide (Lac-
Cer 5), and psychosine (6) were all studied as potential
ligands for gp120. The extent to which these GSLs were
recognized by gp120 differed depending upon the bind-
ing assay used to probe the interaction. For example, in
high performance thin layer chromatography (HPTLC),
GlcCer and LacCer were weakly bound by gp120, while
psychosine was recognized nearly as well as GalCer.7 In
contrast, liposomal studies showed that gp120 bound
GlcCer and LacCer but not psychosine.7
Abbreviations: GSL, glycosphingolipid; GSLs, glycospingolipids;
bGalCer, biotinylated galactosyl cermide; bGlcCer, biotinylated
glucosyl ceramide; bLacCer, biotinylated lactosylceramide; bCelCer,
biotinylated cellbiosyl ceramide; bMelCer, biotinylated melibiosyl
ceramide.
*Corresponding author at current address: University of California,
Davis, Department of Chemistry, One Shields Ave., Davis, CA 95616,
USA. Tel.: +1-530-754-9577; fax: +1-530-752-8995; e-mail: gervay@
chem.ucdavis.edu
yCurrent address: California State University, Sacramento, Depart-
ment of Chemistry, 6000 J Street, Sacramento, CA 95819, USA.
{Current address: Ontogen Corporation, 6451 El Camino Real,
Carlsbad, CA 92009, USA.
xCurrent address: Department of Chemistry, University of Utah, 315
S. 1400 E., Salt Lake City, UT 84112, USA.
It was reasoned that contradicting assay results ema-
nated from differences in ligand presentation geometries
inherent in the assays, which invalidated cross compar-
isons.8 To circumvent the difficulties associated with
lipid-based ELISA and other assay methods, an assay
that could be uniformly applied to a variety of GSLs
was developed.9 A water-soluble biotinylated ethylene
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00325-X

626
K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
glycol linker bearing an acid functionality was syn-
thetically prepared and conjugated to amino-analogues
of GSLs. These ligands were irreversibly adhered to
microtiter plates coated with NeutrAvidinTM, a degly-
cosylated analogue of avidin. Horseradish peroxidase-
tagged recombinant gp120 (rgp120) was added to each
well, and rgp120 recognition of the ligands was mea-
sured. The plating efficiencies of the ligands were also
measured and the data were normalized allowing direct
comparison of binding affinities. The adhesion assay
results showed that rgp120 bound biotinylated GalCer
[bGalCer (7)] and biotinylated GlcCer [bGlcCer (8)]
similarly and it recognized biotinylated LacCer [bLac-
Cer (9)] to an even greater extent (Fig. 2). Since the
ELISA could not be measured under equilibrium
conditions, it was not possible to quantify the binding
affinities; only a qualitative assessment was realized.
However, recently rgp120 recognition of GalCer,
GlcCer, and LacCer was quantified under equilibrium
conditions using total internal reflection microscopy.10
The data between the two studies were closely correlated
validating the use of the adhesion assay as a rapid
screening tool.
from lactose at the 400OH, was chosen to probe the spe-
cificity requirements of the terminal sugar. Further-
more, cellobiose analogues were deemed attractive
targets since humans lack the enzyme required to cleave
the interresidue linkage, which could lead to improved
pharmacokinetics.11 a-Melibiosyl ceramide (11) was
targeted in order to determine if an a-linkage would be
tolerated, and whether the regiochemical positioning of
sugars (1–6 versus 1–4) was important. In this report,
we describe the synthesis of these non-natural GSLs and
their relative binding affinities for rgp120. Structure
activity studies were further expanded to include struc-
turally simpler analogues that were obtained by reduc-
tive amination of lactose, cellobiose, and melibiose with
a biotinylated amino ethylene glycol moiety.
Reaction of the tetrafluorophenyl biotinylated tetra-
ethylene glycol ester 1212 with d-sphingosine (13) affor-
ded biotinylated ceramide (bCer) 14 in 90% yield
(Scheme 1). Cellobiose and a-melibiose were readied for
glycosylation in a three-step procedure. Each dis-
accharide was per-O-acetylated by the action of acetic
anhydride in pyridine and the resulting anomeric acet-
ates were regioselectively removed using the method of
Flandor.13 The final step involved activation of the
anomeric hydroxyls to facilitate glycosidic bond forma-
tion using the trichloroacetimidate strategy developed
by Schmidt.14 The trichloroacetimidates were then cou-
pled to 14 to yield the per-O-acetylated ceramide deri-
vatives (Scheme 2). Glycosylation of per-O-acetylated
cellobiosyl trichloroacetimidate resulted in a 6% yield,
and the yield of the melibiosyl derivative was slightly
higher (18% yield). The reducing anomeric centers in
the isolated products were determined to be the beta
Results and Discussion
In a continued search for compounds that mimic gp120/
GalCer interactions without interfering with healthy
cellular interactions, GSLs that are not natural cell sur-
face components were targeted. Determining the rela-
tive affinities of different disaccharide head groups was
of particular interest, since bLacCer bound gp120 so
well. A cellobiose analogue (10) (Fig. 2), which differs
Figure 1. Structures of naturally occurring glycosphingolipds.
Figure 2. Structures of biotinylated glycosphingolipids.

K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
627
configuration based upon the chemical shift of the
anomeric carbon resonances at d 100.60 and 100.56 for
bCelCer and bMelCer, respectively. The alpha glyco-
sides would be expected to appear upfield at approxi-
mately 96 ppm as is the case for the alpha glycosidic
linkage in bMelCer, which appears at d 96.42 ppm.
Deacetylation in sodium methoxide and methanol
provided the target compounds, 10 [biotinylated cello-
biosyl ceramide (bCelCer)] and 11 [biotinylated meli-
biosyl ceramide (bMelCer)] in 90 and 89% yields,
respectively.
decision to limit the synthetic protocols to as few steps
as possible was consistent with these goals.
The non-natural GSLs, 10 and 11 were evaluated in the
adhesion assay in order to compare their binding affi-
nities for rgp120 with natural analogues (7, 8, and 9). A
simple primary amide, biotinamide (16) (Fig. 2), and
bCer (14) were also tested as a non-specific ligands.9 The
adhesion assay was performed in two ways. First the
compounds were evaluated as previously described.
Plating efficiencies (PE) of the complete ligand panel
ranged from 88 to 97%, providing a high level of
confidence that absorbance differences were due to
variances in binding affinities, and not derived from
the number of ligands present in the well. The extent
of rgp120 binding was monitored by absorbance
changes over 30 min and the end point absorbance
values were recorded. The data obtained from these
experiments were normalized with respect to plating
efficiencies for each ligand. Averaged end point results
from four separate wells are plotted in Figure 3. The
error bars represent the standard deviations between
absorbance readings in each of the wells. The data
in Figure 3 are representative of several different
experiments.
Glycosylations of ceramides are notoriously poor, due
to unfavorable hydrogen bonding interactions.15À17
Improved yields have been realized using a number of
methods, which eliminate the presence of the amide
linkage. In one procedure, the amide was replaced by an
azide functionality.18 Another method involved the use
of Schiff base-protected amines, which provided cer-
amides after further manipulations.19 While strategies
utilizing azido- or imino-protected amines as glycosyl
acceptors in the formation of sphingosine derivatives
may have provided enhanced yields, resources were not
invested in these endeavors. The non-natural GSLs were
primarily synthesized as tools for establishing SAR. The
Scheme 1.
Scheme 2.

628
K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
In an alternative assay construct, the ligands were eval-
uated in a competitive mode. Biotinylated GalCer (7)
was adhered to NeutrAvidinTM coated microtiter plates.
After a blocking step (to eliminate nonspecific binding),
bGalCer coated wells were incubated with rgp120. The
plates were rinsed with PBS to remove unbound rgp120
and then incubated with a solution of ligand for 4 h.
The plates were washed and the amount of bound
rgp120 was measured using TMB (3,30,5,50-tetra-
methylbenzidine) as the peroxidase substrate. The
hypothesis was that a ligand in solution, capable of
binding to rgp120, should compete with plated bGalCer
for the bound rgp120. In this case, the amount of
rgp120 bound to the plate would decrease in proportion
to its affinity for the ligand in solution that it was
exposed to. A set of bGalCer coated wells that, after
rgp120 incubation were exposed to PBS alone, served as
controls against which the decrease in binding was cal-
culated. A series of ligands were tested at varying con-
centrations to provide a dose–response curve. Using the
dose–response curve, EC50 values for decrease in rgp120
bound to plated bGalCer upon exposure to ligand
solutions were calcuated. As can be seen in Table 1, the
glycolipids showed micromolar effective concentrations
and the results from the competition assay correlated
well with the classical adhesion data shown in Figure 3.
carbohydrate functionality as an important recognition
element. Furthermore, the results showed that rgp120
bound bCelCer and bMelCer to the same extent as
bGalCer, an N-acyl analogue of the natural cellular
receptor GalCer.
Since the structure of the non-reducing sugar and the
regio-and stereochemistry of the interresidue glycosidic
linkage did not appreciably affect activity, the SAR
studies were expanded to include more synthetically
accessible compounds. SAR of simplified GalCer ana-
logues is an active area of research. Bertozzi and Bed-
narski20 studied C-glycosides and determined that a
critical alkyl chain length was required for activity. Rico-
Lattes and co-workers synthesized analogues derived
from reductive amination of carbohydrates followed by
fatty acid conjugation.21 The same investigators gener-
ated other derivatives attempting to improve both anti-
HIV activity and cytotoxicity profiles. One lipid chain
was terminated with a carboxylate functionality in order
to increase water solubility and the optimal chain length
of both lipids was probed. It was concluded that the
palmitoyl fatty acid chain length was optimal.
In the current study, neutral GalCer analogues were
targeted because charged compounds were known to
interact with basic amino acids positioned in the V3
loop of gp120.22 These electrostatic interactions could
obscure important hydrophobic and/or hydrophilic
interactions, which would invalidate direct comparisons
to GalCer. The synthesis of the compounds shown in
Scheme 3 was fashioned after the work of Fantini and
co-workers.23 A biotinylated diamino tetraethylene gly-
col linker was required for reductive amination of lac-
tose, cellobiose, and melibiose, which were slated for
subsequent lipid conjugation. These three disaccharides
were chosen for two reasons: To have a direct compar-
ison to bLacCer, bCelCer, and bMelCer; and because
after reductive amination, they would yield non-redu-
cing sugar headgroups (b-galactose, b-glucose, and a-
galactose, respectively). Comparison of the latter com-
pounds to bGalCer and bGlcCer would enhance the
SAR.
The ELISA data provided important information con-
cerning the panel of biotinylated GSLs. First, bLacCer
(9) gave consistently higher end point absorbancies than
the other compounds tested. Secondly, gp120 recog-
nized all of the GSLs with greater affinity than the
simple biotinylated ceramide (14), which defined the
The synthesis of the amino-terminated TEG linker
began with commercially available jeffamine 17, which
was monoprotected with t-butoxycarbonyl anhydride
Table 1. EC50 values derived from competitive adhesion assay cal-
culated by a non-linear regression analysis using variable slope dose–
response curves
Ligand
EC50 mM
CONH2 (16)
bCer (14)
>108Æ22.9
6.21Æ0.960
0.096Æ0.014
0.018Æ0.011
0.019Æ0.013
0.059Æ0.018
0.027Æ0.014
bGalCer (7)
bGlcCer (8)
bLacCer (9)
bCelCer (10)
bMelCer (11)
Figure 3. Adhesion assay data reported as end point reading at 30
min. The data were normalised to reflect the plating efficiencies of each
ligand. The error bars are the standard deviations between the read-
ings in four separate wells. The data are representative of multiple
experiments.
Compounds 14 and 16 do not display sigmoidal dose–response and
their EC50 values were calculated by linear regression curve fitting.

K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
629
(Scheme 3). Biotinylation of the resulting amine was
achieved by a two-step, one-pot procedure. An acti-
vated biotin acid was generated by the action of
2,3,4,5,6-pentafluorophenyl trifluoroacetate (PFP-TFA)
on biotin. Compound 17 was added to the crude penta-
fluorophenyl ester and a quantitative yield of N-Boc-N-
biotinyl jeffamine (18) was realized. Subsequent reaction
with Reagent K provided the biotinylated amino linker
(19) in 52% yield. The disaccharides were diluted in
methanol and water and reacted with 19 in the presence
Scheme 3. (a) PFPOCO(CH2)14CH3, TEA, DMF. The yield of 23 was 19% from lactose, 24 was 33% from cellobiose, and 25 was 37% from
melibiose.
Scheme 4.

630
K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
competitive assay are not completely correlated may
reflect differences in off rates. The adhesion assay is not
performed at equilibrium so observed differences in
rgp120 recognition of ligands may be due to differences
in the reversible binding event. Similarly, the competi-
tion assay allows different ligands to compete for
recognition of gp120 which is reversibly bound by pla-
ted bGalCer. Differences in EC50 values may reflect dif-
ferences in the reversible binding of competing ligands.
Further quantitative measurements of these interactions
is necessary before differences in these assays can be
completely understood. Nonetheless, it is clear that gly-
colipids serve as recognition elements for rgp120, and
the interactions are not highly specific.
Experimental
General procedures
All materials were obtained from commercial sources
and used without additional purification, unless other-
wise noted. Solvents were distilled prior to use as reac-
tion media. All glassware utilized in water-sensitive
Figure 4. Adhesion assay data reported as end point reading at 30
min. The data were normalized to reflect the plating efficiencies of each
ligand. The error bars are the standard deviations between the read-
ings in four separate wells. The data are representative of multiple
experiments.
1
reactions was flame-dried before use. HNMR spectra
were recorded at 250 or 500 MHz on Bruker super-
conducting FT spectrometers. 1HNMR data are
reported in the order of chemical shift, multiplicity
Table 2. EC50 values derived from competitive adhesion assay cal-
culated from dose–response curve, plotted using a logarithmic curve fit
program
(s=singlet,
d=doublet,
t=triplet,
q=quartet,
dd=doublet of doublets, m=multiplet, br=broad). 13C
NMR spectra were recorded at 62.5 or 125 MHz and
were proton decoupled. All spectra are reported in parts
per million relative to the residual solvent peak. Mass
spectrometry was performed at The University of Ari-
zona, Mass Spectrometry Facility, Tucson, AZ, USA.
Ligand
EC50 mM
Linker (26)
bLac (20)
bCel (21)
3.68Æ1.40
7.70Æ2.80
11.45Æ1.64
2.71Æ0.50
bMel (22)
bLacPalm (23)
bCelPalm (24)
bMelPalm (25)
0.065Æ0.024
0.038Æ0.019
0.052Æ0.002
D-Biotin-12-amido-4,7,10-trioxadodecyl sphingosinamide
(bCer, 14). Five millilgrams (0.017 mmol) of d-sphingo-
sine (13, Matreya) was dissolved in 266 mL of DMF. 7.1
mL (0.051 mmol) of TEA was added, followed by 11.1
mg (0.019 mmol) of compound 12. The reaction was
stirred at room temperature overnight under argon,
then evaporated. The crude product was purified using
silica gel chromatography in 9:1 chloroform/methanol.
Twelve milligrams (100%) of an off-white powder was
obtained. The NMR spectra indicate the presence of a
small quantity of the O-linked product, estimated to be
of NaBH4 to give the amines (20, 21, and 22). Com-
pounds 20–22 were evaluated in the adhesion assay to
confirm that a lipid chain was required for rgp120
recognition of these ligands as well. As indicated in
Figure 4, these compounds did not show significantly
greater binding than the linkers, just as was seen for the
ethanolamine derivatives.9
1
To complete the SAR, 21–22 were condensed with pal-
mitic acid that had been activated by the action of PFP-
TFA. The biotinylated linker 19 was also palmitoylated
to form 26 (Scheme 4) as a control measure for non-
specific interactions in the adhesion assays. The purified
compounds (23–26) were assayed in both the adhesion
and competitive assays and the results are shown in
Figure 4 and Table 2, respectively.
about 10%. HNMR (250 MHz, DMSO- d6) d 0.85 (m,
3H, CH3), 1.14–1.59 (m, 28H, (CH2)14), 1.94 (m, 2H,
CH2C¼C), 2.06 (t, 2H, J=7.4 Hz, CH2CONH), 2.32 (t,
2H, J=6.6 Hz, CH2CONH), 2.57 (d, 1H, J=12.7 Hz,
CHS), 2.81 (dd, 1H, J=5.0, 14.9 Hz, CHS), 3.09–3.21
(m, 3H), 3.42–3.65 (m, 14H), 3.87 (m, 1H), 4.12 (m,
1H), 4.49 (t, 1H, J=7.7 Hz), 4.83 (d, 1H, J=5.4 Hz),
5.34–5.60 (m, 2H, H–C¼C–H), 6.37 (s, 1H, NH), 6.44
(s, 1H, NH), 7.54 (d, 1H, J=8.7 Hz, CONHCH), 7.86
(t, 1H, J=5.3 Hz, CONHCH2); 13C NMR d 13.97,
22.11, 25.26, 28.04, 28.20, 28.60, 28.72, 28.83, 28.99,
29.07, 31.31, 31.73, 35.09, 36.18, 55.43, 59.19, 60.31,
61.03, 66.97, 69.19, 69.52, 69.56, 69.72, 71.19, 130.81,
131.24, 162.71, 169.90, 172.12. High-resolution mass
spectrum (FAB+), calcd for C37H69N4SO8 (MH)+:
729.4836. Found: 729.4814.
Conclusion
The biological data for the simpler analogues clearly
show that both a carbohydrate functionality and a
hydrophobic chain are required for activity. The fact
that the trends between the adhesion assay and the

K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
631
ꢀ-D -Glucoopyranosyl-(1!4)-O-ꢀ-D -glucopyranosyl)-
(1!1)-(2S,3R,4E)-3-hydroxy-2-N-(12-D-biotinamide-
4,7,10-trioxadodecyl)-sphingenine (10). Cellobiose (1.0
g, 2.92 mmol, J.T. Baker) was weighed directly into a
flame-dried flask. DMAP (35.4 mg, 0.29 mmol, Nova-
biochem) was added to the flask, followed by 6 mL of
pyridine. The flask was stirred in an ice bath under
argon, and acetic anhydride (5.5 mL, 58.4 mmol,
Aldrich) was added dropwise over a period of 10 min
via an addition funnel. The reaction mixture was stirred
at room temperature for 1 day, until TLC in 1:1 hex-
anes/ethyl acetate indicated nearly a complete reaction.
The flask was placed on ice and 6 mL of methanol was
added. The solvent was evaporated, and the residue was
dissolved in ethyl acetate and washed with 3ꢀ0.1 M
HCl, 1ꢀwater, 1ꢀbrine. The organic phase was dried
over anhydrous Na2SO4. A white foam product (1.87 g,
83%) resulted. 1HNMR (250 MHz, CDCl 3) d 1.89–2.03
(m, 24H, (COCH3)8), 3.58–3.75 (m, 3H), 4.01 (m, 2H),
4.29 (dd, 1H, J=4.5, 12.5 Hz), 4.42 (m, 1H), 4.83 (t, 1H,
J=8.1 Hz), 4.92–5.19 (m, 5H), 5.59 (d, 1H, J=8.2 Hz).
The peracetylated disaccharide was used without further
purification in the following manner. 25 mL of a 7:3
mixture of THF/MeOH was stirred at 0 ꢁC with NH3
(g) bubbled through for 10 min. This solution was
transferred to a flask containing peracetylated cellobiose
(500 mg, 0.737 mmol), and the mixture was stirred from
0 ꢁC to room temperature for 1.5 h, until TLC in 2:1
ethyl acetate/hexanes indicated a complete absence of
the peracetylated disaccharide. The solution was evapo-
rated, and the resultant crude oil was purified in 2:1
ethyl acetate/hexanes, to yield 308 mg (66%) of an a/b
mixture of a clear colorless oil (1-hydroxy-hepta-O-ace-
tyl cellobiose). 1HNMR (250 MHz, CDCl 3) d 1.90–2.12
(m, 21H, (COCH3)7) 3.58–3.71 (m, 2H), 3.94–4.01 (m,
2H), 4.20 (d, 1H, J=3.8 Hz), 4.30 (dd, 1H, J=4.2, 12.6
Hz), 4.44 (m, 2H), 4.72 (m, 1H), 4.85 (app t, 1H, J=8.1
Hz), 5.03 (m, 2H), 5.28 (t, 1H, J=3.7 Hz), 5.41 (t, 1H,
J=9.5 Hz); 13C NMR d 20.64, 20.75, 20.85, 20.96,
21.00, 21.11, 61.72, 61.95, 62.10, 67.96, 68.23, 69.49,
71.47, 71.77, 72.0, 72.18, 73.05, 73.13, 73.36, 76.60,
76.69, 76.72, 77.23, 77.43, 77.74, 90.08 (a C-1), 95.27 (b
C-1), 100.79 (b-C-10), 100.89 (b-C-10), 169.21, 169.48,
169.87, 169.90, 170.39, 170.44, 170.56, 170.64, 170.70,
170.75, 170.80. High-resolution mass spectrum (FAB+)
calcd for C26H37O18 (MH)+: 637.1980. Found:
637.1976. The 1-hydroxy-hepta-O-acetyl cellobiose pro-
duct (308 mg, 0.484 mmol) was dissolved in 5 mL of
CH2Cl2 in a N2 atmosphere. CCl3CN (169.5 mL, 1.69
mmol, Aldrich), was added next via a syringe, followed
by the addition of 14 mg (0.581 mmol) of NaH
11.5 Hz), 4.35 (dd, 1H, J=4.0, 12.6 Hz), 4.48 (app t,
2H, J=7.8 Hz), 4.89 (t, 1H, J=8.0 Hz), 5.04 (m, 2H),
5.48 (t, 1H, J=9.6 Hz), 6.44 (d, 1H, J=3.8 Hz, H-1),
8.62 (s, 1H, NH); 13C NMR d 20.61, 20.69, 20.75, 20.81,
20.87, 20.95, 61.53, 61.68, 67.84, 69.42, 69.81, 70.0,
71.10, 71.44, 71.81, 72.15, 73.01, 73.20, 76.25, 76.52,
76.72, 76.87, 77.23, 77.38, 77.74, 93.0 (a C-1), 101.08 (b
C-10), 128.47, 161.11, 169.23, 169.44, 169.51, 169.60,
170.2, 170.32, 170.36, 170.41, 170.64. High-resolution
mass spectrum (FAB+) calcd for C28H36NO18Cl3Na
(MNa)+: 802.0896. Found: 802.0892. The tri-
chloracetimidate (64.3 mg, 0.082 mmol) and 14 (100 mg,
0.137 mmol) were weighed directly into a flame-dried
flask, then azeotroped 4ꢀ with dry, distilled benzene to
ensure dryness. Flame-dried AW-300 molecular sieves
(1 g, Aldrich) were introduced. The residue was dis-
solved in 3.4 mL of freshly distilled 1,2-dichloroethane,
then placed in a methanol bath at ꢂ10 ꢁC under an
argon atmosphere. BF3 etherate (21 mL, 0.164 mmol,
Aldrich) was introduced once the solution had chilled to
À10 ꢁC. This temperature was maintained for 1.75 h, at
which time the temperature was raised to 0 ꢁC. After an
additional 24 h, the temperature was increased again to
10 ꢁC. At this time, an additional 0.5 equiv of BF3
etherate was added. After a total reaction time of 48 h,
the reaction was diluted with CHCl3, filtered through
Celite, then washed 1ꢀ with a 5% solution of NaHCO3,
and 1ꢀ with H2O. The organic layer was dried over
anhydrous Na2SO4, then filtered and evaporated. Pur-
ification was achieved by silica gel chromatography in
12:1 CHCl3/MeOH. Further purification was necessary,
and was accomplished by normal-phase HPLC on a
semi-preparative silica gel column (Rainin-Varian
Microsorb column, 5 mm particle size, 100 A pore size,
10 250 mm) in 12:1 CHCl3/MeOH. A yellowish amor-
1
phous solid (7.3 mg, 6%) resulted. HNMR (500 MHz,
CDCl=) d 0.86 (t, 3H, J=6.7 Hz, CH3), 1.23 (s, 20H,
(CH2)10), 1.32 (m, 2H), 1.44 (m, 2H), 1.53–1.68 (m, 4H,
(CH2)2 biotin), 1.99–2.11 (m, 25H, (COCH3)7,
CH2CONH, CH2C¼C), 2.20 (t, 2H, J=7.2 Hz,
CH2CONH), 2.46 (m, 2H), 2.72 (d, 1H, J=12.6 Hz,
CHS), 2.91 (dd, 1H, J=5.0, 12.8 Hz, CHS), 3.08 (s,
1H), 3.15 (app q, 1H, J=7.5, 12.0 Hz, CHS), 3.40 (m,
2H), 3.53 (t, 2H, J=5.0 Hz, CH2–O), 3.61 (m, 10H,
(CH2–O)5), 3.66 (m, 2H, CH2–O), 3.75 (m, 3H), 3.96
(dd, 1H, J=5.1, 10.4 Hz), 4.03 (m, 2H), 4.09 (m, 1H),
4.17 (app t, 1H, J=5.0 Hz), 4.31 (m, 1H), 4.35 (dd, 1H,
J=4.3, 12.4 Hz), 4.45 (d, 1H, J=7.9 Hz), 4.50 (m, 2H),
4.83–4.88 (m, 2H), 5.04 (t, 1H, J=9.8 Hz), 5.13 (m, 2H),
5.43 (dd, 1H, J=6.3, 15.4 Hz, C¼C–H), 5.63 (s, 1H,
NH), 5.70 (m, 1H, C¼C–H), 6.61 (br s, 1H, NH), 6.71
(dd, 1H, J=7.8 Hz, NH); 13C NMR d 14.11, 20.54,
20.58, 20.66, 20.73, 20.85, 22.68, 25.43, 28.03, 29.28,
29.32, 29.35, 29.54, 29.66, 29.70, 31.91, 32.40, 35.72,
37.02, 39.14, 40.51, 53.36, 55.29, 60.05, 61.56, 61.77,
67.26, 67.78, 68.50, 69.90, 70.04, 70.27, 70.39, 70.43,
70.48, 71.51, 71.59, 71.93, 72.30, 72.66, 72.81, 72.90,
75.92, 76.38, 100.60, 100.73, 128.37, 133.86, 163.08,
169.06, 169.30, 169.73, 169.87, 170.24, 170.45, 170.50,
171.62, 173.07. High-resolution mass spectrum (FAB+)
calcd for C63H103N4SO25 (MH)+: 1347.6632. Found:
1347.6619. The glycosylated product (5.0 mg, 0.0037
mmol) was dissolved in MeOH. A catalytic amount of
(Aldrich). The solution turned
a
clear yellow
immediately upon the addition of NaH. The reaction
was stirred at room temperature under N2, with the
color deepening to a brownish-amber. After 5 h, the
reaction was shown to be complete by TLC in 2:1 ethyl
acetate/hexanes. Purification was achieved by silica gel
chromatography utilizing a gradient from 2:1 ethyl ace-
tate/hexanes to 1.5:1 of the same. A fluffy white solid
product (155.4 mg, 85%) resulted after evaporation in
vacuo from benzene. 1HNMR (250 MHz, CDCl 3) d
1.88–2.13 (m, 21H, (COCH3)7), 3.63 (br d, 1H, J=9.4
Hz), 3.80 (t, 1H, J=9.6 Hz), 4.03 (app dt, 4H, J=4.5,

632
K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
NaOMe (Avocado) was added, and the reaction was
stirred at room temperature overnight. The reaction was
diluted further with MeOH, and Dowex H+ was stirred
in, until the solution achieved a neutral pH. The MeOH
was removed in vacuo, then the residue was taken up
into a small quantity of H2O and freeze-dried. An off-
white fluffy solid (3.6 mg, 90%) resulted. No further
purification was necessary. 1HNMR (500 MHz, 4:1
CDCl3/CD3OD, externally referenced to tetra-
methylsilane in CDCl3) d 0.76 (t, 3H, J=6.5 Hz, CH3),
1.14 (s, 22H, (CH2)11), 1.32 (t, 2H, J=7.8 Hz), 1.56 (m,
4H, (CH2)2 biotin), 1.92 (m, 2H, CH2C¼C), 2.11 (t, 2H,
J=7.1 Hz, CH2CONH), 2.38 (t, 2H, J=6.3 Hz,
CH2CONH), 2.62 (d, 1H, J=12.8 Hz, CHS), 2.80 (dd,
1H, J=4.9, 12.9 Hz, CHS), 3.05 (m, 1H, CHS), 3.20 (m,
2H), 3.28 (m, 4H, incl. CD3OD signal), 3.43–3.57 (m,
12H), 3.61 (m, 4H), 3.91 (br s, 1H), 4.01 (app q, 1H),
4.17 (d, 1H, J=7.8 Hz, H-1 or H-1), 4.20 (app dd, 1H,
J=4.2, 6.7 Hz, CHNH biotin), 4.31 (d, 1H, J=7.8 Hz,
H-1 or H-1), 4.39 (app t, 1H, J=7.3 Hz, CHNH biotin),
5.34 (dd, 1H, J=6.9, 15.7 Hz, C¼C–H), 5.59 (app dd,
1H, J=8.5, 15.5 Hz, C¼C–H). It appears as though
some peaks are obscured by the HOD signal at 3.75; 13C
NMR d 13.86, 22.52, 25.35, 27.97, 28.22, 29.15, 29.18,
29.20, 29.40, 29.51, 29.54, 31.77, 32.25, 35.50, 36.35,
38.96, 40.16, 53.27, 55.48, 60.04, 60.08, 60.97, 61.84,
67.13, 69.54, 69.71, 69.75, 69.83, 70.00, 70.11, 72.02,
72.92, 73.13, 73.38, 74.53, 74.85, 76.16, 76.38, 78.91,
102.83, 103.00, 128.76, 134.04, 164.15, 172.14, 174.42.
High-resolution mass spectrum (FAB+) calcd for
C49H89N4SO18 (MH)+: 1053.5893. Found: 1053.5884.
169.67, 169.83, 170.08, 170.17, 170.20, 170.47, 170.55.
High-resolution mass spectrum (FAB+) calcd for
C28H39O19 (MH)+: 679.2086. Found: 679.2092. A 7:3
mixture of THF:MeOH (30 mL) was stirred at 0 ꢁC with
NH3 (g) bubbled into the solution for a period of 10
min. This solution was added to a flask containing per-
acetylated melibiose. After 1 h at 0 ꢁC, the solution was
stirred from 0 ꢁC to room temperature. After 2 h, TLC
in 2:1 ethyl acetate/hexanes indicated a complete con-
version to the 1-hydroxy-hepta-O-acetyl disaccharide.
After evaporation, the crude oil was purified by silica
gel chromatography in 2:1 ethyl acetate/hexanes. An
off-white foam product (391 mg, 69%) resulted. 1H
NMR (250 MHz, CDCl3)
d 1.87–2.10 (m, 21H,
(COCH3)7), 3.47–3.64 (m, 2H), 3.91–4.02 (m, 2H), 4.14
(m, 2H), 4.59 (app br s, 1H), 4.70 (dd, 1H, J=3.3, 10.2
Hz), 4.85–4.98 (m, 2H), 5.06 (d, 1H, J=3.7 Hz), 5.19
(app dt, 1H, J=3.3, 10.8 Hz), 5.30 (br d, 1H, J=3.0
Hz), 5.41 (t, 1H, J=9.5 Hz); 13C NMR d 20.62, 20.68,
20.77. 20.97, 61.69, 65.99, 66.91, 67.46, 67.91, 67.95,
68.22, 69.0, 69.17, 69.95, 71.20, 72.63, 72.82, 76.73,
77.23, 77.43, 77.75, 89.80, 95.04, 96.25, 169.53, 169.69,
170.03, 170.21, 170.31, 170.77, 170.87, 171.10. High-
resolution mass spectrum (FAB+) calcd for C26H37O18
(MH)+: 637.1980. Found: 637.1982. The 1-hydroxy-
hepta-O-acetyl derivative of melibiose (378 mg, 0.594
mmol) was dissolved in CH2Cl2 under argon. CCl3CN
(209 mL, 2.08 mmol, Aldrich) was introduced via a syr-
inge. This was followed by the addition of NaH(17.1
mg, 0.713 mmol, Aldrich). The solution immediately
bubbled upon the addition of the NaH, and turned
from a clear colorless solution, to a clear yellow one.
The reaction was stirred at room temperature for 5 h, as
TLC in 2:1 ethyl acetate/hexanes had indicated a nearly
complete reaction. After evaporation of the solvent, the
crude residue was purified by silica gel chromatography
in 1:1 ethyl acetate/hexanes. A clear yellow oil (129 mg,
28%) resulted. 1HNMR (250 MHz, CDCl 3) d 1.90–2.05
(m, 21H, (COCH3)7), 3.49 (dd, 1H, J=1.9, 11.0 Hz),
3.66 (dd, 1H, J=5.9, 11.1 Hz), 3.99 (d, 2H, J=6.6 Hz),
4.12 (m, 1H), 4.26 (t, 1H, J=6.5 Hz), 4.96–5.09 (m, 4H),
5.18 (dd, 1H, J=3.3, 10.1 Hz), 5.37 (d, 1H, J=3.2 Hz),
5.49 (t, 1H, J=9.7 Hz), 6.38 (d, 1H, J=3.7 Hz, H-1),
8.83 (s, 1H, NH); 13C NMR d 20.54, 20.71, 20.75, 20.82,
20.89, 61.80, 65.45, 66.38, 67.58, 68.03, 68.25, 68.50,
69.86, 70.01, 70.69, 76.71, 77.22, 77.42, 77.73, 92.64,
95.85, 160.51, 169.57, 169.95, 170.05, 170.30, 170.42,
179.54. High-resolution mass spectrum (FAB+) calcd
for C28H36NO18Cl3Cs (MCs)+: 912.0052. Found:
912.0070. The trichloroacetimidate (129 mg, 0.165
mmol) and 14 (124 mg, 0.170 mmol) were weighed
directly into a flame-dried flask. The mixture of
compounds was azeotroped 3ꢀ with dry, distilled ben-
zene to ensure dryness. AW-300 molecular sieves
(Aldrich, 1 g) were added next. The flask was placed
under an argon atmosphere, then the dry solids were
dissolved in 5 mL of 1,2-dichloroethane. The flask was
placed in a methanol bath at 10 ꢁC and allowed to
equilibrate at this temperature before 42 mL (0.330
mmol) of BF3 etherate (Aldrich) was added via a syr-
inge. After 1 h, the temperature of the bath was raised
to 0 ꢁC. After an additional 16 h at 0 ꢁC, the bath tem-
perature was raised to 10 ꢁC, and allowed to react at this
ꢀ-D-Galactopyranosyl-ꢁ-(1!6)-O-(ꢀ-D-glucopyranosyl)-
(1!1)-(2S,3R,4E)-3-hydroxy-2-N-(12-D-biotinamide-
4,7,10-trioxadodecyl)-sphingenine (11). a-Melibiose
(Aldrich, 500 mg, 1.46 mmol) and 14.3 mg (0.117 mmol)
of DMAP (Novabiochem) were both weighed directly
into the reaction flask. 3 mL of pyridine was added to
dissolve the solids, and the flask was then placed in an
ice bath at 0 ꢁC under an argon atmosphere. Acetic
anhydride (2.8 mL, 29.2 mmol, Aldrich) was added
dropwise via an addition funnel over a period of 15 min.
After the addition of acetic anhydride was complete, the
reaction was stirred from 0 ꢁC to room temperature
overnight. Once TLC in 2:1 ethyl acetate/hexanes indi-
cated a complete reaction, the flask was placed on ice
and 3 mL of methanol was added. The solution was
stirred at 0 ꢁC for 0.5 h, then the solution was evapo-
rated. The residue was taken up in ethyl acetate and
washed 3ꢀ0.1 M HCl, then 1ꢀwith water, and 1ꢀwith
brine. The organic phase was dried over anhydrous
Na2SO4. A white foam solid product, the peracetylated
disaccharide (934 mg, 94%), resulted in an approximate
5:1 a/b-ratio. This was used without further purifica-
tion. 1HNMR (250 MHz, CDCl 3) d 1.87–2.08 (m, 24H,
(COCH3)8), 3.50 (dt, 1H, J=2.4, 11.5 Hz), 3.61 (dd, 1H,
J=4.5, 11.6 Hz), 3.96 (m, 2H), 4.09 (app q, 1H, J=6.6
Hz), 4.88–5.07 (m, 5H), 5.20 (dd, 1H, J=3.3, 10.8 Hz),
5.36 (m, 2H), 6.16 (d, 1H, J=3.6 Hz); 13C NMR d
20.41, 20.60, 20.62, 20.70, 20.82, 20.99, 61.60, 61.71,
65.60, 65.77, 66.32, 66.44, 67.38, 68.06, 68.24, 68.47,
69.22, 69.84, 70.18, 70.44, 72.80, 73.41, 76.72, 77.23,
77.74, 88.90, 91.56, 96.02, 96.33, 168.84, 169.25, 169.32,

K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
633
temperature for 10 h, before diluting the reaction mix-
ture with CHCl3 and filtering through Celite. The solu-
tion was washed first with 1ꢀ5% NaHCO3, followed by
1ꢀH2O. The organic layer was subsequently dried over
anhydrous Na2SO4, then filtered and evaporated. The
crude residue was first purified by silica gel chromato-
graphy in 9:1 CHCl3/MeOH, then further purified by
normal phase silica gel HPLC (Rainin-Varian Micro-
sorb column, 5 mm particle size, 100 A pore size, 10 250
32.23, 35.50, 36.48, 39.00, 40.16, 53.32, 55.48, 60.06, 61.33,
61.40, 61.79, 63.24, 67.03, 69.07, 69.29, 69.36, 69.44, 69.54,
69.76, 69.84, 70.07, 70.17, 70.26, 70.55, 71.85, 73.27, 73.40,
74.54, 76.25, 78.19, 98.38, 103.55, 125.66, 128.83, 133.93,
137.26, 163.14, 164.21, 171.90, 171.96, 174.29. High-reso-
lution mass spectrum (FAB+) calcd for C49H89N4SO18
(MH)+: 1053.5893. Found: 1053.5911.
D-Biotin-11-amino-3,6,9-trioxaundecylate (19). Com-
pound 17 (Huntsman) was weighed (1.0 g, 5.21mmol)
directly into a flask. This was dissolved in 100 mL of
THF. 64 mg of DMAP (0.52 mmol) was then added. t-
Butoxycarbonyl anyhdride (1.137g, 5.21 mmol, Nova-
biochem) was added via an addition funnel as a solution
in 25 mL of THF over a period of 1 h. The reaction
solution became increasingly cloudy as the addition of
the anhydride proceeded. After the addition was com-
plete, the solution again became clear. After 2 h at room
temperature, the reaction mixture was evaporated. Pur-
ification was achieved by silica gel chromatography in
12:1 CHCl3/MeOH. 687 mg (45%) of the mono-
protected product resulted. 1HNMR (250 MzH,
CDCl3) d 1.22 (s, 9H, (CH3)3), 2.64 (br s, 2H,
CH2NH2), 3.07 (br q, 2H, J=5.2 Hz, CH2NHCO), 3.30
(q, 4H, J=4.9, 10.0 Hz, (CH2–O)2), 3.43 (m, 8H, (CH2–
O)8), 5.42 (br s, 1H, NH); 13C NMR d 28.21, 40.11,
41.50, 70.00, 70.30, 73.21, 78.61, 149.47, 155.85. High-
resolution mass spectrum (FAB+) calcd for
C13H29N2O5 (MH)+: 293.2076. Found: 293.2072.
mm) in 12:1 CHCl3/MeOH. 28 mg (13%) of a yellow
1
amorphous solid resulted. HNMR (500 MHz, CDCl
)
3
d 0.85 (t, 3H, J=6.8 Hz, CH3), 1.22 (s, 20H, (CH2)10),
1.33 (m, 2H, CH2), 1.43 (m, 2H, CH2 biotin), 1.60–1.74
(m, 4H, (CH2)2 biotin), 1.96–2.05 (m, 19H, (COCH3)5,
CH2CONH, CH2C¼C), 2.12 (m, 6H, (COCH3)2), 2.20
(t, 2H, J=7.6 Hz, CH2CONH), 2.43–2.54 (m, 1H), 2.72
(d, 1H, J=12.9 Hz, CHS), 2.88 (dd, 1H, J=4.6, 12.9
Hz, CHS), 3.12 (app q, 1H, J=6.9, 12.1 Hz, CHS), 3.40
(app t, 2H, J=4.6 Hz), 3.53 (t, 2H, J=4.6 Hz, CH2–O),
3.60 (s, 10H, (CH2–O)5), 3.65–3.77 (m, 2H, CH2–O),
3.97 (m, 1H), 4.03–4.11 (m, 2H), 4.16 (app t, 1H, J=5.3
Hz), 4.21 (t, 1H, J=6.1 Hz), 4.29 (app t, 1H, J=4.6
Hz), 4.47 (m, 2H), 4.87 (m, 1H), 4.91–5.06 (m, 1H), 5.18
(m, 2H), 5.27–5.40 (m, 2H), 5.44 (m, 2H), 5.70 (dt, 1H,
J=6.9, 15.3 Hz, C¼C–H), 6.18 (d, 2H, J=13.0 Hz,
(NH)2), 6.80 (m, 2H, (NH)2); 13C NMR d 14.09, 20.61,
20.64, 20.67, 20.72, 20.88, 20.93, 22.65, 25.50, 28.05,
28.10, 29.21, 29.27, 29.32, 29.50, 29.53, 29.63, 29.67,
31.88, 32.37, 35.80, 36.91, 39.12, 40.47, 53.34, 53.64,
55.43, 60.11, 60.94, 61.76, 61.89, 65.36, 66.47, 67.27,
67.38, 67.42, 68.08, 68.25, 68.46, 68.72, 69.91,70.04,
70.17, 70.25, 70.37, 71.03, 71.23, 72.46, 72.70, 72.80,
76.64, 96.42, 100.56, 125.66, 128.67, 133.80, 137.86,
163.53, 169.28, 169.55, 169.94, 170.15, 170.51, 170.95,
171.14, 171.52, 173.20. High-resolution mass spectrum
(FAB+) calcd for C63H103N4SO25 (MH)+: 1347.6632.
Found: 1347.6581. The glycosylated product (19.4 mg,
0.014 mmol) was dissolved in MeOH. A catalytic
amount of NaOMe (Avocado) was added, and the
reaction was stirred at room temperature overnight. The
reaction was diluted further with MeOH, and Dowex
H+ was stirred in, until the solution achieved a neutral
pH. The MeOH was removed in vacuo, then the residue
was taken up into a small quantity of H2O and freeze-
dried. 13.5 mg (89%) of an off-white fluffy solid resul-
Biotinylation. d-Biotin (237 mg, 0.97 mmol, Sigma) was
weighed directly into the reaction flask. This was dis-
solved in 10 mL of DMF. 406 mL of TEA (2.91 mmol)
was added next. The reaction flask was placed under an
argon atmosphere in an ice bath. After the solution had
clarified, 2,3,4,5,6-pentafluorophenyl trifluoroacetate
(PFP-TFA, Aldrich) was added via syringe (201 mL,
1.17 mmol). The reaction was then stirred from 0 ꢁC to
room temperature under argon. After 4 h, TLC in 9:1
CHCl3/MeOHindicated that all of the biotin had been
converted to the intermediate pentafluorophenyl ester.
The monoprotected linker was dissolved in 2 mL of
DMF and added to the flask containing PFP-biotin via
a syringe. An additional 3 mL of DMF was used to
rinse the flask, and this was added to the reaction flask
as well. TLC indicated that the reaction went to com-
pletion almost instantaneously. The reaction was stirred
overnight at room temperature under argon to ensure a
complete reaction. The reaction mixture was then eva-
porated and subjected to silica gel chromatography in
9:1 CHCl3/MeOH. 503 mg (100%) of a yellowish
amorphous solid (18) resulted. 1HNMR (250 MHz,
CDCl3) d 1.27 (br s, 11H, (CH3)3, CH2 biotin), 1.51 (m,
4H, (CH2)2 biotin), 2.06 (t, 2H, J=7.4 Hz, CH2CONH),
2.58 (d, 1H, J=12.7 Hz, CHS), 2.73 (dd, 1H, J=5.0,
13.2 Hz, CHS), 2.98 (br q, 1H, J=4.5 Hz, CHS), 3.14
(app q, 2H, J=5.0 Hz, CH2NHCO), 3.26 (br q, 2H,
J=4.7, 9.5 Hz, CHS), 3.35–3.47 (m, 12H, (CH2–O)6),
4.14 (m, 1H, CHNHbiotin), 4.33 (m, 1H, C HNHbio-
tin), 5.15 (br t, 1H, NH carbamate), 6.22 (s, 1H, NH
biotin), 6.82 (s, 1H, NH biotin), 6.92 (br t, 1H,
CH2NHCO); 13C NMR d 25.53, 28.01, 28.30, 35.81,
38.95, 40.15, 40.36, 55.70, 60.13, 61.65, 69.79, 69.88,
1
ted. No further purification was necessitated. HNMR
(500 MHz, 4:1 CDCl3/CD3OD, externally referenced to
tetramethylsilane in CDCl3) d 0.74 (t, 3H, J=6.6 Hz,
CH3), 1.12 (s, 20H, (CH2)10), 1.21 (m, 2H, CH2), 1.30
(app t, 3H, J=7.5 Hz, CH2), 1.52 (m, 4H, (CH2)2), 1.89
(m, 2H, CH2C¼C), 2.09 (t, 2H, J=7.1 Hz,
CH2CONH), 2.34 (t, 2H, J=6.0 Hz, CH2CONH), 2.60
(d, 1H, J=12.9 Hz, CHS), 2.78 (dd, 1H, J=4.9 Hz, 12.8
Hz, CHS), 3.03 (app q, 1H, J=7.4, 12.0 Hz, CHS), 3.15
(t, 1H, J=8.3 Hz), 3.22–3.36 (m, 6H, incl. CD3OD
peak), 3.41 (m, 2H, CH2–O), 3.50 (app s, 10H, (CH2–
O)5), 3.56–3.67 (m, 8H), 3.73 (app d, 2H, J=5.8 Hz),
3.84 (br s, 2H, incl. the HOD peak), 3.97 (m, 2H, incl.
the HOD peak), 4.17 (m, 2H, CHNHbiotin, H-1), 4.37
(app t, 1H, J=6.8 Hz, CHNH biotin), 4.76 (s, 1H, H-1),
5.28 (app dd, 1H, J=6.8, 15.3, C¼C–H), 5.60 (app dt,
1H, J=6.8, 15.3 Hz, C¼C–H); 13C NMR d 13.82, 22.49,
25.34, 27.94, 28.22, 29.15, 29.18, 29.35, 29.39, 29.52, 31.75,

634
K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
69.98, 70.18, 77.43, 78.95, 155.92, 164.37, 173.38. High-
resolution mass spectrum (FAB+) calcd for
C23H43N4SO7 (MH)+: 519.2852. Found: 519.2844.
Hz), 3.80 (m, 3H), 3.44 (m, 3H), 3.59 (app dq, 1H,
J=1.6, 7.8 Hz), 3.67 (app dt, 2H, J=1.3, 5.4 Hz), 3.71
(ddd, 1H, J=1.7, 3.4, 10.0 Hz), 3.73–3.84 (m, 12H,
(CH2–O)6), 3.87 (m, 3H), 3.92 (m, 2H), 3.98 (m, 2H),
4.27 (m, 1H), 4.46 (m, 1H, CHNHbiotin), 4.57 (d, 1H,
J=7.8 Hz, H-10), 4.65 (app dt, 1H, J=0.9, 5.0 Hz,
CHNHbiotin); 13C NMR (externally referenced to 3-
(trimethylsilyl)propionic 2,2,3,3-d4 acid, sodium salt) d
28.02, 30.58, 30.74, 38.35, 41.73, 42.57, 50.05, 52.46,
58.26, 63.16, 64.13, 64.80, 64.98, 68.16, 70.57, 71.59,
71.78, 72.23, 72.49, 72.54, 73.36, 73.77, 73.85, 75.31,
78.22, 81.44, 105.73, 168.24, 179.87. High-resolution
mass spectrum (FAB+) calcd for C30H57N4SO15
(MH)+: 745.3541. Found: 745.3532.
Deprotection. Reagent K25 was prepared separately in a
glass-stoppered flask. This solution was pipetted into
the reaction flask containing the biotinylated protected
linker (503 mg, 0.970 mmol), and was stirred at room
temperature for 1.5 h before evaporating. The crude
yellow oil was purified by silica gel chromatography
with a gradient of CHCl3/MeOH/TEA starting at
4:1:0.1 and ending with 100% MeOH. 213 mg (52%) of
a yellowish amorphous solid resulted. 1HNMR
(250 MHz, 4:1 CDCl3/CD3OD, external CHCl3 refer-
ence) d 1.17 (app q, 2H, J=7.0, 14.1 Hz, CH2 biotin),
1.43 (m, 4H, (CH2)2 biotin), 1.95 (t, 2H, J=7.4 Hz,
CH2CONH), 2.47 (d, 1H, J=12.8 Hz, CHS), 2.57 (t,
2H, J=5.1 Hz, CH2NH2), 2.64 (dd, 1H, J=4.9, 12.9
Hz, CHS), 2.89 (app q, 1H, J=7.7, 12.1 Hz, CHS), 3.15
(app t, 2H, CH2NHCO), 3.27 (q, 4H, J=6.4, 11.5 Hz,
(CH2–O)2), 3.38 (s, 8H, (CH2–O)4), 4.04 (app dd, 1H,
J=4.5, 7.7 Hz, CHNHbiotin), 4.24 (app dd, 1H,
J=5.0, 7.7 Hz, CHNH biotin), 6.02 (br s, 1H, NH),
6.28 (br s, 1H, NH), 7.37 (br t, 1H, CH2NHCO); 13C
NMR d 25.57, 28.14, 28.45, 35.71, 39.04, 39.16, 40.36,
41.11, 55.72, 60.16, 61.85, 69.76, 69.95, 70.00, 70.34,
72.63, 164.41, 174.15, 174.23. High-resolution mass
spectrum (FAB+) calcd for C18H35N4SO5 (MH)+:
419.2328. Found: 419.2329.
11-D-Biotinamide-N-[3,6,9-trioxaundecane]-1-amino-1-
deoxycellitol (21). 5.6 mg resulted. Yield: 19%. 1H
NMR (500 MHz, D2O) d 1.45 (m, 2H, CH2 biotin),
1.60–1.76 (m, 4H, (CH2)2 biotin), 2.30 (t, 2H, J=7.2
Hz, CH2CONH), 2.81 (d, 1H, J=13.1 Hz, CHS), 3.02
(dd, 1H, J=5.0, 13.1 Hz, CHS), 3.24 (dd, 1H, J=10.2,
12.6 Hz), 3.34–3.44 (m, 8H), 3.49 (dd, 1H, J=2.1, 5.6
Hz), 3.53 (app t, 1H, J=9.3 Hz), 3.65 (t, 2H, J=5.4
Hz), 3.72–3.80 (m, 12H, (CH2–O)6), 3.84 (app t, 2H,
CH2–O), 3.88 (d, 1H, J=3.2 Hz), 3.91 (dd, 1H, J=2.2,
7.7 Hz), 3.95 (m, 1H), 4.23 (dt, 1H, J=3.7, 6.4 Hz), 4.45
(dd, 1H, J=4.5, 8.2 Hz, CHNHbiotin), 4.62 (m, 2H,
CHNHbiotin, H-1); 13C NMR (externally referenced to
3-(trimethylsilyl)propionic 2,2,3,3-d4 acid, sodium salt)
d 27.99, 30.55, 30.71, 38.32, 41.70, 42.55, 50.00, 52.42,
58.23, 63.13, 63.42, 64.73, 64.96, 68.12, 70.52, 71.75,
72.20, 72.34, 72.45, 72.50, 72.51, 73.26, 73.73, 76.11,
78.31, 78.82, 81.51, 105.16, 168.23, 179.86. High-resolu-
tion mass spectrum (FAB+) calcd for C30H57N4SO15
(MH)+: 745.3541. Found: 745.3531.
General procedure for the reductive amination of
disaccharides
The reductively aminated disaccharides were obtained
using the procedure of Fantini and coworkers. 1.61
equiv of compound 19 was weighed directly into a flask
and dissolved in a minimal amount of methanol. Into a
separate flask, 1.0 equiv of one of the disaccharides
(lactose, cellobiose, or melibiose) was weighed, then
dissolved in a minimal amount of deionized water. The
disaccharide solution was added to the flask containing
compound 19, and this was stirred at room temperature
overnight. The reaction mixture was then heated for 6 h
at 55 ꢁC. Next, 1.1 equiv of NaBH4 was added to the
reaction as a solution in MeOH. The reaction was then
stirred at room temperature overnight before evaporat-
ing the solvent. A crude purification was achieved for
compounds 20, 21, and 22 on a Biogel P-2 (Biorad)
1.5ꢀ50 cm column with detection at 222 nm. Sub-
sequent purification of compounds 21 and 22 was
achieved by gradient reversed-phase HPLC from 0.1%
trifluoroacetic acid (TFA) in water to 100% acetonitrile,
utilizing a Vydac C-18 semipreparative column with a
Perkin-Elmer Binary LC Pump and an LC 90 UV
Spectrophotometric Detector. All products were fluffy
white solids after lyophilization from water.
11-D-Biotinamide-N-[3,6,9-trioxaundecane]-1-amino-1-
deoxymelitol (22). 11.1 mg resulted. Yield: 37%. 1H
NMR (500 MHz, D2O) d 1.47 (m, 2H, (CH2 biotin),
1.61–1.80 (m, 4H, (CH2)2 biotin), 2.32 (t, 2H, J=7.3
Hz, CH2CONH), 2.84 (d, 1H, J=13.0 Hz, CHS), 3.04
(dd, 1H, J=5.0, 13.1 Hz, CHS), 3.17–3.41 (m, 5H), 3.44
(t, 2H, J=5.5 Hz), 3.77 (t, 2H, J=4.9 Hz), 3.79–4.02
(m, 20H), 4.04 (d, 2H, J=2.8 Hz), 4.48 (dd, 1H, J=4.5,
7.9 Hz, CHNHbiotin), 4.66 (dd, 1H, J=4.7, 7.9 Hz,
CHNHbiotin), 5.03 (d, 1H, J=3.6 Hz, H-1); 13C NMR
(internally referenced to CH3CN) 24.67, 27.23, 27.40,
35.01, 38.40, 39.23, 54.90, 59.80, 60.69, 61.63, 61.98,
68.04, 68.20, 68.43, 68.79, 68.90, 69.07, 69.12, 69.18,
70.34, 70.41, 70.47, 72.46, 97.93, 164.88, 176.50. High-
resolution mass spectrum (FAB+) calcd for
C30H57N4SO15 (MH)+: 745.3541. Found: 745.3569.
General procedure for the amidation of reductively
aminated disaccharides
Palmitic acid (1.2 equiv, Matreya, Inc.) was weighed
into a flame-dried flask, then dissolved in approximately
0.5–1.0 mL of CH2Cl2. The flask was placed under
argon. TEA was added (3.0 equiv), followed by PFP-
TFA (1.4 equiv, Aldrich). The reaction was allowed to
stir at room temperature. After 0.5 h, TLC in 9:1
CHCl3/MeOHindicated complete conversion to the
11-D-Biotinamide-N-[3,6,9-trioxaundecane]-1-amino-1-
deoxylactitol (20). 13.2 mg resulted. Yield: 33%. 1H
NMR (500 MHz, D2O) d 1.45 (m, 2H, CH2 biotin),
1.59–1.80 (m, 4H, (CH2)2 biotin), 2.32 (t, 2H, J=7.2
Hz, CH2CONH), 2.83 (d, 1H, J=13.1 Hz, CHS), 3.04
(dd, 1H, J=5.0, 13.1 Hz, CHS), 3.24 (t, 1H, J=12.1

K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
635
PFP intermediate. Next, approximately 1.5 mL of DMF
was added and compounds 20, 21, or 22 were intro-
duced into the flask. The reaction was stirred overnight
at room temperature, after which TLC indicated the
disappearance of the UV-active intermediate. The sol-
vent was evaporated, and the crude residues were pur-
ified by gradient reversed-phase HPLC from 0.1%
trifluoroacetic acid (TFA) in water to 100% acetonitrile,
utilizing a Vydac C-18 semipreparative column with a
Perkin-Elmer Binary LC Pump and an LC 90 UV
Spectrophotometric Detector. Off-white grainy solid
products 23, 24, and 25 resulted after lyophilization
from water.
(t, 2H, J=6.8 Hz, CH2C¼O), 2.26 (app q, 2H, J=5.9,
13.4 Hz, CH2C¼O), 2.58 (d, 1H, J=12.8 Hz, CHS),
2.76 (dd, 1H, J=5.0, 13.0 Hz, CHS), 3.02 (br dd, 1H,
J=6.3, 8.4 Hz, CHS), 3.20 (m, 4H), 3.39 (t, 2H, J=5.0
Hz), 3.48 (br d, 16H), 3.59–3.67 (m, 6H), 3.73 (app t,
2H, incl HOD signal), 4.16 (dd, 1H, J=4.4, 7.6 Hz,
CHNHbiotin), 4.35 (dd, 1H, J=5.2, 7.7 Hz, CHNH
biotin), 4.75 (app t, 1H, J=3.5 Hz, H-1); 13C NMR
d 14.05, 22.72, 25.46, 25.59, 28.17, 28.46, 29.41,
29.57, 29.75, 31.98, 33.40, 35.73, 39.23, 40.37, 55.70,
60.27, 61.91, 62.04, 68.64, 69.28, 69.55, 69.83, 70.06,
70.48, 70.63, 72.19, 72.70, 98.92, 174.49, 175.52,
176.21. High-resolution mass spectrum (FAB+) calcd
for C46H87N4SO16 (MH)+: 983.5838. Found:
983.5823.
11-D-Biotinamide-N-[3,6,9-trioxaundecane]-N-palmitoyl-
1
1-amino-1-deoxylactitol (23). Yield: 2.7 mg (30%). H
NMR (500 MHz, 4:1 CDCl3/CD3OD, externally refer-
enced to 0.01% tetramethylsilane in CDCl3) d 0.90 (t,
3H, J=6.8 Hz, CH3), 1.29 (br s, 26 H, (CH2)13), 1.45
(m, 2H, CH2 biotin), 1.57–1.77 (m, 4H, (CH2)2 biotin),
2.23 (t, 2H, J=7.2 Hz, CH2C¼O), 2.48 (app q, 2H,
J=7.3, 15.3 Hz, CH2C¼O), 2.71 (d, 1H, J=12.8 Hz,
CHS), 2.93 (dd, 1H, J=5.1, 12.8 Hz, CHS), 3.21 (m,
1H, CHS), 3.37 (t, 2H, J=5.4 Hz), 3.45–3.57 (m, 8H),
3.62–3.70 (m, 10H), 3.72–3.79 (m, 5H), 3.82–3.93 (m,
3H), 4.07 (m, 1H), 4.19 (m, 1H), 4.31 (m, 1H, CHNH
biotin), 4.48 (m, 2H, H-1, CHNHbiotin); 13C NMR d
13.82, 22.45, 25.16, 25.31, 27.92, 29.13, 29.27, 29.31,
29.38, 29.41, 29.44, 29.48, 31.70, 33.11, 35.49, 38.95,
40.10, 48.13, 49.16, 55.41, 60.00, 61.08, 61.77, 62.25,
68.99, 69.52, 69.60, 69.80, 70.08, 70.25, 70.32, 70.35,
70.63, 71.39, 72.12, 73.24, 75.32, 103.92, 174.27, 175.28,
175.63. High-resolution mass spectrum (FAB+) calcd
for C46H87N4SO16 (MH)+: 983.5838. Found: 983.5828.
Palmitoylated biotinamidojeffamine (26). Polymer bound
EDC (P-EDC; 180 mg ꢂ0.179 mmol) was swelled in 1
mL of CHCl3.24 A solution of palmitic acid (20.2 mg,
0.078 mmol) in CHCl3 (0.5 mL) and biotinamido jeff-
amine 19 (30 mg, 0.07 mmol) in CHCl3 (0.5 mL) were
sequentially added to the P-EDC suspension. The reac-
tion mixture was stirred gently and progress of the
reaction monitored by TLC. After 14 h the reaction
mixture was filtered and resin thoroughly washed with
alternating washes of CHCl3 and MeOH. Combined
organic extracts were dried (Na2SO4), concentrated
and the crude reaction product was purified by pas-
sing through a short plug of silica gel to yield 26
(73%). 13C 173.36, 173.25, 164.01, 70.35, 70.30, 70.11,
70.03, 69.99, 69.91, 61.75, 60.17, 55.57, 40.46, 39.09,
36.66, 35.91, 31.87, 29.65, 29.61, 29.50, 29.39, 29.32,
28.17, 28.08, 25.74, 25.56, 22.64, 14.07. ESI-MS 657.3
(MH+).
11-D-Biotinamide-N-[3,6,9-trioxaundecane]-N-palmitoyl-
1
Adhesion assay experimental procedures
1-amino-1-deoxycellitol (24). Yield: 2.6 mg (33%). H
NMR (500 MHz, 4:1 CDCl3/CD3OD, externally refer-
enced to 0.01% tetramethylsilane in CDCl3) d 0.80 (app
t, 3H, J=6.8 Hz, CH3), 1.18 (m, 26H, (CH2)13), 1.36 (m,
2H, CH2 biotin), 1.35–1.66 (m, 4H, (CH2)2 biotin), 2.15
(t, 2H, J=7.2 Hz, CH2C¼O), 2.34 (m, 2H, CH2C¼O),
2.66 (d, 1H, J=12.9 Hz, CHS), 2.85 (dd, 1H, J=4.6,
12.7 Hz, CHS), 3.10 (m, 1H, CHS), 3.21 (t, 2H, J=8.0
Hz), 3.33 (m, 10H, incl MeOD signal), 3.41–3.48 (m,
6H), 3.56 (app s, 10H, (CH2–O)5), 3.61–3.68 (m, 4H),
3.74–3.86 (m, 4H), 3.96 (m, 1H), 4.24 (dd, 1H, J=4.6,
7.9 Hz, CHNH biotin), 4.42 (m, 2H, H-1, CHNHbio-
tin); 13C NMR d 13.78, 22.46, 25.16, 25.32, 27.94, 28.24,
29.15, 29.31, 29.41, 29.49, 31.72, 33.12, 33.21, 35.51,
38.97, 40.11, 48.18, 49.21, 50.17, 55.42, 60.01, 61.25,
61.78, 62.33, 68.77, 68.86, 69.02, 69.54, 69.72, 69.80,
70.03, 70.11, 70.26, 70.38, 70.94, 71.63, 71.72, 73.63,
76.23, 76.46, 81.83, 103.36, 174.16, 175.28, 175.67.
High-resolution mass spectrum (FAB+) calcd for
C46H87N4SO16 (MH)+: 983.5838. Found: 983.5828.
Assay ligand dilutions. Freeze-dried biotinylated ana-
lytes were weighed out to the tenth of a microgram
using a Cahn 21 Automatic Electrobalance. Stock solu-
tions of these compounds were prepared at 1 mg/mL
concentrations in 20% DMSO/80% PBS buffer. DMSO
was first added, and the compounds allowed to fully
dissolve prior to addition of PBS buffer. The solutions
were thoroughly mixed prior to further dilution in PBS
for assaying, and they were stored at À20 ꢁC between
uses. The dilution concentration of all compounds uti-
lized in the assays was 20 mg/mL.
Plating efficiency assay. The plating efficiency of each
compound was tested to ensure comparable coverage on
NeutrAvidinTM-coated eight-well strip plates (Pierce
Biochemicals). In this assay, each compound was plated
into four wells at 150 mL each, and incubated for 4 h at
25 ꢁC and 1000 rpm in a JitterbugTM microplate
incubator (Boekel). Two blank wells were incubated
with 150 mL PBS, and 2–8 wells were incubated with 150
mL PBS each for HRP-biotin determination in the sec-
ond incubation. After 4 h, each well was rinsed with
3ꢀ200 mL of PBS. Next, a 1:4000 dilution of HRP-Bio-
tin was added to all wells excluding the blanks in 100 mL
quantities. The blank wells were incubated with 100 mL
of PBS. The second incubation was for 1 h at 25 ꢁC and
11-D-Biotinamide-N-[3,6,9-trioxaundecane]-N-palmitoyl-
1
1-amino-1-deoxymelitol (25). Yield: 1.4 mg (18%). H
NMR (500 MHz, 4:1 CDCl3/CD3OD, externally refer-
enced to 0.01% tetramethylsilane in CDCl3) d 0.72 (t,
3H, J=6.8 Hz, CH3), 1.10 (br s, 26H, (CH2)13), 1.28 (m,
2H, CH2 biotin), 1.44–1.58 (m, 4H, (CH2)2 biotin), 2.07

636
K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
1000 rpm. After the incubation, the wells were rinsed as
before. A TMB (3,30,5,50-tetramethylbenzidine, Sigma)
solution was prepared (1 mg TMB in 500 mL of DMSO
and 9.5 mL of 0.05M phosphate–citrate buffer with
sodium perborate). The TMB solution was added to
each well, including the blanks, at a volume of 100 mL/
well. The plate was immediately read on a BioRad
model 550 microplate reader every min for 20 min at
630 nm, with a reference wavelength of 490 nm, and 5 s
of shaking between each reading. All wells for each
analyte were averaged, and had the TMB blank auto-
matically subtracted through the use of the BioRad
Microplate Manager software package. The standard
deviations for each of the ligands was determined based
upon the original data set. The four wells containing
only HRP-biotin at the absorbance maximum served as
a 100% measure of the sites filled by HRP-biotin. All
other analyte wells were compared to the HRP-biotin
wells at the same data point, and were based on these
results. The percentage of NeutrAvidinTM sites filled
was ascertained for each compound. The results given
for this assay represent the mean of four separate
experiments.
repeated two separate times, with the average of these
experiments reported.
Competition assay. A solution of bGalCer (7) was pre-
pared at 1 mg/mL concentration in 20% DMSO/80%
PBS buffer. The solution was stored at À20 ꢁC between
uses. A 20 mg/mL solution of bGalCer was made from
the stock solution using PBS buffer. NeutrAvidinTM
coated polystyrene strip plates (Pierce Biochemicals)
were used for the competitive assay. 150 mL of the 20
mg/mL solution of 7 was added to each well for plating
with GalSph. The plating was achieved by shaking the
plate at 25 ꢁC at 1000 rpm for 4 h using a Jitterbug
microplate incubator. Additionally, in each assay plate
three wells were plated with blocking agent, biotinamide
(16) (using 150 mL of the 20 mg/mL solution of biotina-
mide) served as the negative control. Also, in each assay
plate three wells served as a TMB blank9. Following the
first incubation, the plates were washed with 3ꢀ200 mL
of PBS buffer. Next, 150 mL a of 20 mg/mL solution of
blocking agent biotinamide (16) was added to each well
(excluding TMB blank wells, which were incubated with
150 mL of PBS buffer ) and the plate was incubated for 1
h at 25 ꢁC/1000 rpm, to block off any open unbound
NeutrAvidinTM sites. After the second incubation, the
plates were rinsed with 3ꢀ200 mL of PBS buffer. Next
100 mL of a solution of HRP-rgp120 (obtained after a
1:3 dilution with PBS from a stock 500 mL vial con-
taining 0.333 mg of HRP-rgp120) was added to all wells
excluding the TMB blank wells (which were incubated
with 100 mL of PBS buffer). The plate was incubated for
1 h at 25 ꢁC/1000 rpm. The plate was then rinsed with
5ꢀ200 mL of PBS buffer and then the competition step
was performed. 150 mL of the test ligand solution was
added to each well and the plate was incubated for 4 h
at 25 ꢁC/1000 rpm. Each ligand to be assayed in the
competitive mode was tested at six concentrations (1,
10, 20, 40, 80, and 100 mg/mL, the solutions of the test
ligands were prepared from the corresponding stock
solutions of 1 mg/mL in 20% DMSO/80% PBS buffer).
In each assay plate, the test ligand solution at each
concentration was added to three separate wells coated
with 7; in addition three bGalCer coated wells were
exposed to PBS in place of test ligand during this final
competition step. The amount of HRP-rgp120 bound in
these plates served as a reference for calculating the
decrease in HRP-rgp120 bound after exposure to test
ligand solutions. After the 4 h incubation period, the
plates were washed with 5ꢀ200 mL of PBS buffer. TMB
reagent was prepared by dissolving a TMB tablet (1 mg
tablet, Sigma) in 0.5 mL of DMSO and 9.5 mL of
phosphate–citrate buffer. 100 mL of the TMB reagent
was added to each well. The plate absorbances were
read every min for 30 min at 630 nm (with 5 s shake
before each reading) using a Bio-Rad model 550 micro-
plate reader. Each well reading had the TMB blank
reading automatically subtracted through the use of the
Biorad Microplate manager software package. At the
end of the 30 min reading period the enzymatic reac-
tion in each well was immediately quenched by addi-
tion of 100 mL of 2 N H2SO4 solution in each well. The
absorbance of the resulting yellow solutions were read
at 430.
Adhesion assay. The compounds (see ELISA Ligand
Dilutions) were plated in 150 mL volumes into four wells
each of NeutrAvidinTM-coated eight-well strip plates.
The plate was incubated for 4 h at 25 ꢁC and 1000 rpm
on a JitterbugTM microplate incubator (Boekel). An
additional four wells each of both a TMB blank (see
Plating Efficiency Assay), and a gp120 blank were incu-
bated with 150 mL of PBS in this incubation. Following
the first incubation, the wells were rinsed 3ꢀ200 mL
PBS. Next, 150 mL/well, excluding the TMB and gp120
wells, of compound 16 was incubated for 1 h at 25 ꢁC
and 1000 rpm, to ensure that all remaining open Neu-
trAvidinTM sites were blocked. The TMB and gp120
wells were incubated instead with 150 mL of PBS, as
before. After the second incubation, the wells were
rinsed as in the first incubation. The third incubation
comprised of adding 100 mL of a 1:4 dilution (from a
500 mL vial containing 0.333 mg gp120) of gp120-HRP
(Intracel) into every well, except the TMB wells. The
TMB wells were incubated with 100 mL of PBS each.
The incubation was for 1 h at 25 ꢁC at 1000 rpm. The
wells were then rinsed with 5ꢀ200 mL of PBS. A solu-
tion of TMB reagent (1 mg tablet) was prepared in 500
mL of DMSO and 9.5 mL of phosphate–citrate buffer
containing sodium perborate). Then 100 mL of the TMB
reagent was added to each well. The plate absorbances
were read every min for 30 min at a test wavelength of
630 nm, a reference wavelength of 490 nm, with a 5 s
shake in between each reading, on a BioRad model 550
microplate reader. The absorbance readings were nor-
malized to reflect the plating efficiencies by dividing the
values by the plating efficiency. All wells for each ana-
lyte were averaged, and had the TMB blank auto-
matically subtracted through the use of the BioRad
Microplate Manager software package. The standard
deviations were derived from the unaveraged data set.
The wells containing only gp120 served as a measure of
non-specific binding interactions between the glycopro-
tein and the plate, and were minimal. This assay was

K. D. McReynolds et al. / Bioorg. Med. Chem. 10 (2002) 625–637
637
Calculation of EC50 values. The EC50 value represents
the concentration of the compound to induce a 1/2
maximal displacement of rgp120 bound to plated bGal-
Cer (7). The EC50 values in Table 1 were calculated by a
non linear regression analysis using variable slope sig-
moidal dose response curves (GraphPad Prizm, version
3.0 GraphPad Software Inc., San Diego, CA, USA).
Compounds 14 and 16 do not display a sigmoidal dose
response and their EC50 values were calculated by lin-
ear regression curve fitting (CA-Cricket Graph III,
Computer Associates International Inc., Islandia, NY,
USA). EC50 values reported in Table 2 were calculated
from dose–response curves plotted using logarithmic
curve fit program (CA-Cricket Graph III).
4. Fantini, J.; Cook, D. G.; Nathanson, N.; Spitalnik, S. L.
Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 2700.
5. Clapham, P. R.; McKnight, A.; Talbot, S.; Wilkinson, D.
Perspect. Drug Discov. Des. 1996, 5, 83 and references therein..
6. Bhat, S.; Spitalnik, S. L.; Gonzalez-Scarano, F.; Silberberg,
D. H. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 7131.
7. Long, D.; Berson, J. F.; Cook, D. G.; Doms, R. W. J.
Virol. 1994, 68, 5890.
8. Pascher, I.; Sundell, S. Chem. Phys. Lipids 1977, 20, 175.
9. McReynolds, K. D.; Hadd, M. J.; Gervay-Hague, J. Bio-
conjugate Chem. 1999, 10, 1021.
10. Conboy, J. C.; McReynolds, K. D.; Gervay-Hague, J.;
Saavedra, S. S. Angew. Chem. Int. Engl. Ed. 2000, 39, 2882.
11. Vzcategui, E.; Johansson, G.; Ek, B.; Pettersson, G. J.
Biotech. 1991, 21, 143.
12. Seitz, O.; Kunz, H. J. Org. Chem. 1997, 62, 813.
13. Flandor, J.; Garcia-Lopez, M. T.; de las Heras, F. G.;
Mendez-Castrillon, P. P. Synthesis 1985, 1121.
14. Schmidt, R. R.; Michel, J. Angew. Chem., Int. Ed. In Engl.
1980, 19, 731.
Acknowledgements
Financial support for this research was gratefully
received from NIH(AI40359-02), The American Foun-
dation for AIDS Research (amfAR, 02669), Eli Lilly,
Alfred P. Sloan Foundation, and The National Science
Foundation Early Career Award (CHE9623583).
K.D.M. gratefully acknowledges receipt of the Uni-
versity of Arizona Dean’s Fellowship and the Depart-
ment of Chemistry Carl S. Marvel Fellowship. We are
indebted to Dr. Arpad Somogyi, director of the UA
Mass Spectrometry Facility, for his assistance in
compound characterization.
15. Szabo, L.; Li, Y.; Polt, R. L. Tetrahedon Lett. 1991, 32, 585.
16. Ito, Y.; Numata, M.; Sugimoto, M.; Ogawa, T. J. Am.
Chem. Soc. 1989, 111, 8508.
17. Numata, M.; Sugimoto, M.; Shibayama, S.; Ogawa, T.
Carbohydr. Res. 1988, 174, 73.
18. Schmidt, R. R.; Zimmermann, P. Agnew. Chem., Int. Ed.
Engl. 1986, 25, 725
19. Peterson, M. A.; Polt, R. J. Org. Chem. 1993, 58, 4309.
20. Bertozzi, C. R.; Cook, D. G.; Kobertz, W. R.; Gonzalez-
Scarano, F.; Bednarski, M. D. J. Am. Chem. Soc. 1992, 114,
10639.
21. Rico-Lattes, I.; Garrigues, J.-C.; Perez, E.; Andre-Barres,
C.; Madelaine-Duuich, C.; Lattes, A.; Linas, M.-D.; Aubertin,
A.-M. New J. Chem. 1995, 19, 341.
22. Rico-Lattes, I.; Gouzy, M.-F.; Andre-Barres, C.; Guidetti,
B.; Lattes, A. New J. Chem. 1998, 22, 451.
23. Fantini, J.; Hammache, D.; Delezay, O.; Yahi, N.; Andre-
Barres, C.; Rico-Lattes, I.; Lattes, A. J. Biol. Chem. 1997, 272,
7245.
References and Notes
1. Wiegandt, H. In New Comprehensive Biochemistry;
Elsevier: Amsterdam, 1985; Vol. 10.
2. Hakomori, S.-I. Ann. Rev. Biochem. 1981, 50, 733.
3. Harouse, J. M.; Bhat, S.; Spitalnik, S. L.; Laughlin, M.;
Stefano, K.; Silberberg, D. H.; Gonzalez-Scarano, F. Science
1991, 253, 320.
24. P-EDC was prepared as described by: Desai, M. C.;
Stephens Stramiello, L. M. Tetrahedron Lett. 1993, 34, 7685.
25. King, D. S.; Fields, G. B. Int. J. Peptide Protein Res. 1990,
30, 255.