Bioconjugate Chem. 2004, 15, 349−358
349
Synthesis of Novel, Multivalent Glycodendrimers as Ligands for HIV-1 gp120 Richard D. Kensinger,† Brian C. Yowler,† Alan J. Benesi,‡ and Cara-Lynne Schengrund*,† Department of Chemistry, The Pennsylvania State University, 152 Davey Laboratory, University Park, Pennsylvania 16802, and Department of Biochemistry and Molecular Biology H171, The Pennsylvania State University College of Medicine, 500 University Drive, Hershey, Pennsylvania 17033. Received September 3, 2003; Revised Manuscript Received December 18, 2003
Multivalent neoglycoconjugates are valuable tools for studying carbohydrate-protein interactions. To study the interaction of HIV-1 gp120 with its reported alternate glycolipid receptors, galactosyl ceramide (GalCer) and sulfatide, galactose- and sulfated galactose-derivatized dendrimers were synthesized, analyzed as ligands for rgp120 by surface plasmon resonance, and tested for their ability to inhibit HIV-1 infection of CXCR4- and CCR5-expressing indicator cells. Four different series of glycodendrimers were made by amine coupling spacer-arm derivatized galactose residues, either sulfated or nonsulfated, to poly(propylenimine) dendrimers, generations 1-5. One series of glycodendrimers was prepared from the ceramide saccharide derivative of purified natural GalCer, and another was from chemically synthesized 3-(β-D-galactopyranosylthio)propionic acid. Synthesis of 3-sulfogalactopyranosyl-derivatized dendrimers was accomplished using the novel compound, 3-(β-D-3sulfogalactopyranosylthio)propionic acid. The fourth series was made by random sulfation of the 3-(βD-galactopyranosylthio)propionic acid functionalized dendrimers. Structures of the carbohydrate moieties were confirmed by NMR, and the average molecular weights and polydispersities of the different glycodendrimers were determined using MALDI-TOF MS. Surface plasmon resonance studies found that rgp120 IIIB bound to the derivatized dendrimers tested with nanomolar affinity, and to dextran sulfate with picomolar affinity. In vitro studies of the effectiveness of these compounds at inhibiting infection of U373-MAGI-CCR5 cells by HIV-1 Ba-L indicated that the sulfated glycodendrimers were better inhibitors than the nonsulfated glycodendrimers, but not as effective as dextran sulfate.
INTRODUCTION
The primary protein receptors for HIV-1 are the cell surface CD4 molecules (1). Interaction of viral envelope glycoprotein, gp120, with CD4 results in a conformational change within gp120 that allows it to interact with a chemokine coreceptor (1, 2). This results in formation of a ‘fusion pore’ which leads to fusion of viral and cellular membranes (2-4). In the absence of CD4, however, it is believed that the cell surface glycosphingolipids (GSLs), GalCer and its 3′-sulfated derivative, sulfatide (SGalCer), act as alternate receptors for HIV-1 gp120 (5-10). It has been shown that gp120 has distinct binding sites for GalCer and CD4 (5, 11, 12), and that the GalCer binding site has a nanomolar affinity for GalCer (8, 13, 14)s which is on the order of that reported for the affinity of gp120 for CD4 (15-17). The fact that the GalCer binding site is distinct and of relatively high affinity has made it a target for the development of GalCer and SGalCer mimetics as potential inhibitors of the interaction of HIV-1 with its target cells. Support for this idea is provided by a number of studies of the effectivieness of GalCer analogues as ligands for gp120 or as inhibitors of HIV infectivity (18-25). Cell surface GSLs are involved in membrane trafficking, cell adhesion, cell morphogenesis, and signal trans* To whom correspondence should be sent. Telephone (717) 531-8048. Fax (717) 531-7072. E-mail
[email protected]. † Department of Biochemistry and Molecular Biology. ‡ Department of Chemistry.
duction (26-29). In addition to these functions, GSLs in lipid rafts are used by a variety of pathogens for cellular entry (30, 31). GSLs consist of a ceramide (N-acylated sphingosine) backbone and a carbohydrate moiety β-linked to the terminal hydroxyl group of the sphingosine. The ceramide portion intercalates into the phospholipid bilayer of the cell membrane, and the carbohydrate moiety extends outward from the membrane surface. Although the role of the ceramide portion in bacterial and viral adhesion is debatable (13, 23, 24), the consensus is that it is necessary for the positioning of the saccharide moiety such that it is available as a ligand for the pathogen (13, 32). It is interesting that the carbohydrate moieties of GSLs function as receptors for various pathogens, since individual carbohydrate-protein interactions are often of low affinity (33-35). This is counterintuitive for biological processes that require high specificity and avidity. Nature has overcome this problem of innately weak monosaccharide interactions by clustering carbohydrate ligands together on the cell surface to yield a ‘cluster’ or ‘multivalency’ effect, that results in much stronger interactions (35-38). In some cases, the cluster effect has been observed to result in affinities that are greater than would be expected based on the sum of the total individual carbohydrate-protein interactions themselvessa cooperative affinity greater than that of valency (36, 38). The clustering of GSLs in cell surface lipid rafts or ‘detergent insoluble membranes’ is a good example of the natural clustering of carbohydrate residues (29, 39). Lipid rafts have been shown to serve as docking sites for
10.1021/bc034156a CCC: $27.50 © 2004 American Chemical Society Published on Web 02/27/2004
350 Bioconjugate Chem., Vol. 15, No. 2, 2004 Scheme 1. Formation of Peracetylated GalCer Ceramide Saccharide (1). Numbering System for NMR Analysis Is Shown
Kensinger et al. Scheme 2. Synthesis of 3-(3-Sulfo-β-D-glactopyranosylthio)propionic Acid (3). Numbering System for NMR Analysis Is Shown
EXPERIMENTAL PROCEDURES
various bacterial toxins and viruses (31). Therefore, it is reasonable to hypothesize that a successful inhibitor of the binding of trimers of gp120 found on the surface of HIV-1 to GalCer/SGalCer would need to mimic the natural clustering of the carbohydrate moieties found in lipid rafts. Multivalent neoglycoconjugates have been found to be useful mimetics of the natural clustering of carbohydrate residues (40, 41). Glycodendrimers (42-44) are particularly effective due to their propensity to form spherical nano-cellular decoys (33, 36). Effective ligands for influenza virus (33), cholera toxin (45), and various lectins (37, 46) have been successfully prepared by conjugating cell surface carbohydrate ligands to spherical hyperbranched dendrimers. To study the interaction of HIV-1 gp120 with GalCer and SGalCer, and to develop potential carbohydrate-based inhibitors of this interaction, we synthesized several series of novel glycodendrimers of increasing size as multivalent mimetics of GalCer and SGalCer. This was accomplished by chemically isolating the saccharide portions of GalCer (compound 1, see Scheme 1) and SGalCer and then using amine coupling to conjugate the “ceramide saccharide” (47) to polypropylenimine (DAB-Am) dendrimers. DAB-AM dendrimers were also functionalized with 3-(β-D-galactopyranosylthio)propionic acid (compound 2, Scheme 2) and 3-(β-D-3sulfogalactopyranosylthio)propionic acid (compound 3, Scheme 2). Failure to obtain efficient linkage of the 3-sulfated, spacer-arm derivatized galactose molecule to the DAB-Am dendrimer cores led us to randomly sulfate dendrimers derivatized with compound 2. NMR was used to confirm synthesis of the various saccharides, and the average molecular weights and polydispersities of the glycodendrimers were determined by MALDI-TOF MS. Surface plasmon resonance (SPR) was used for kinetic analysis of the interaction of HIV-1 rgp120 IIIB with the different glycodendrimers, and dextran sulfate (DxS)sa known potent binding inhibitor of HIV-1. Last, the effectiveness of the glycodendrimers at blocking HIV-1 infection of cultured indicator cells was ascertained.
Abreviations. Bu2SnO, dibutyltin oxide; BF3Et2O, borontriflouride dietherate; COSY, correlated spectroscopy; DCM, dichloromethane; DCE, dichloroethane; DIPEA, diisopropylethylamine; DxS, dextran sulfate; DAB-Am, poly(propylene imine) dendrimers; DMF, dimethylformamide; Et3N, triethylamine; GalCer, galactosyl ceramide; HATU, N,N,N′,N′-tetramethyl-O-(7-azabenzotriazol-1-yl)uronium hexafluorophosphate; HMBC, heteronuclear multiple bond correlation; HMQC, heteronuclear multiple quantum correlation; HPTLC, high performance thin-layer chromatography; IAA, trans-3indoleacrylic acid; KMnO4, potassium permangenate; ka, association rate constant; kd, dissociation rate constant; KD, equilibrium dissociation constant; kDa, kilodaltons; Mn, number-average molecular weight; Mw, weight-average molecular weight; MALDI-TOF MS, matrix-assisted laser desorption ionization time-of-flight mass spectrometry;MTS,3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; NaOMe, sodium methoxide; PD, polydispersity; PMS, phenazine methosulfate; SO3NMe3, sulfur trioxide trimethylamine complex; SGalCer, sulfatide; SPR, surface plasmon resonance; Rmax, maximum binding capacity in RUs; RU, resonance unit (change in surface concentration of ∼1 pg of protein/mm2); THAP, trihydroxyacetophenone. Materials. GalCer, SGalCer, lyso-sulfatide, and psychosine standards were purchased from Matreya, Inc (Pleasant Gap, PA). GalCer and SGalCer for ceramide saccharide syntheses were purified from bovine brain white matter as previously described (48). Recombinant HIV-1 IIIB gp120 was obtained from Immunodiagnostics, Inc (Woburn, MA), DAB-Am dendrimers were from Aldrich (Milwaukee, WI), and silica gel 60 HPTLC plates were from Merck (Germany). Unless designated otherwise, all other materials were commercially available, and used without further purification. Preparation of Peracetylated GalCer Ceramide Saccharide (1). GalCer was deacylated to form psychosine using reaction conditions described by Radin and by Dubois et al. for the removal of the fatty acid side-chain
Synthesis of Novel, Multivalent Glycodendrimers
from GSLs (49, 50). The ceramide saccharide derivative of GalCer was made from pyschosine according to procedures published by Mylvaganum and Lingwood (51). MALDI-TOF MS m/z 548 [M - H]-: C22H31NO15 (549). 3-(β-D-Galactopyranosylthio)propionic Acid (2). 3-Mercaptopropionic acid (10.5 mL, 120 mmol) was added with stirring to a solution of galactose-β-pentaacetate, 11.7 g (30 mmol), in 60 mL of anhydrous DCM. This was followed by the addition of 5.7 mL (45 mmol) of BF3‚Et2O (52, 53). The reaction was stirred at room temp and followed by TLC (BuOH:CH3OH:H2O; 2:1:1, v/v/v). Upon completion (∼5 h), the reaction mixture was poured into ice water in a separatory funnel. The lower organic phase was removed, and the aqueous phase was rinsed 3× with DCM. The combined organic phases were washed with water, treated with anhydrous sodium sulfite, filtered over Celite, and dried by rotary evaporation. Peracetylated galactothiopropionic acid was isolated from the resulting viscous oil on a silica gel column and eluted with a step gradient of DCM:CH3OH of increasing polarity. It was recovered as a colorless oil in 80% yield. For NMR analysis and regioselective sulfation, a portion of the peracetylated galactothiopropionic acid was deacetylated using NaOMe in anhydrous CH3OH. After stirring overnight at room temp, it was neutralized using DOWEX 50W [H+]. 1H NMR (400 MHz, D2O, 25 °C): δ ) 2.50 (t, J7A,8 ≈ 7 Hz, J7B,8 ≈ 7 Hz, 2H, H-8), 2.90 (m, J7A,8 ≈ 7 Hz, J7B,8 ≈ 7 Hz, 2H, H-7), 3.52 (t, J1,2 ) 9.72 Hz, J2,3 ) 9.6 Hz,1H, H-2), 3.61 (dd, J2,3 ) 9.6 Hz, J3,4 ) 3.4 Hz, 1H, H-3), 3.66 (m, 1H, H-6B), 3.67 (m, J4,5 ) 0.2 Hz, 1H, H-5), 3.72 (m, 1H, H-6A), 3.93 (d, J3,4 ) 3.36 Hz, J4,5 ) 0.2 Hz, 1H, H-4), 4.46 (d, J1,2 ) 9.72 Hz, 1H, H-1). 13C NMR (D2O, 25 °C): δ ) 27.3 (C-7), 38.6 (C-8), 61.7 (C-6), 69.5 (C-4), 70.3 (C-2), 74.55 (C-3), 79.6 (C-5), 86.65 (C-1), 181.4 (C-9). MALDI-TOF MS m/z 291 [M + Na]+: C9H16O7S (268). 3-(3-Sulfo-β-D-galactopyranosylthio)propionic Acid (3). Compound 3 was prepared because the carbohydrate moiety of sulfatide (SGalCer) is galactose-3-sulfate, and the hypothesis was that it would probably be the best mimetic for sulfatide. Galactothiopropionic acid, 5.86 g (21.87 mmol), and Bu2SnO, 5.99 g (1.1 equiv, 24.05 mmol), were refluxed in 200 mL of anhydrous CH3OH for 4 h at 90 °C (54, 55). The resulting translucent yellow solution was dried by rotary evaporation, taken up in 150 mL of DMF, and 3.42 g SO3NMe3 (1.2 equiv, 24.6 mmol) was added. The mixture was sonicated (∼1 min) and stirred at room temp for 30 h (54, 55). The reaction was quenched with CH3OH, dried by rotary evaporation using an oil vacuum pump, and purified on a silica gel column eluted with a CHCl3:CH3OH step gradient. 1H NMR (400 MHz, D2O, 25 °C): δ ) 2.68 (t, J7A,8 ≈ 7 Hz, J7B,8 ≈ 7 Hz, 2H, H-8), 2.97 (m, J7A,8 ≈ 7 Hz, J7B,8 ≈ 7 Hz, 2H, H-7), 3.70 (dd, J1,2 ) 9.90 Hz,1H, H-2), 3.72 (m, 1H, H-6B), 3.75 (m, 1H, H-5), 3.75 (m, 1H, H-6A), 4.32 (s, 1H, H-4), 4.33 (dd, 1H, H-3), 4.60 (d, J1,2 ) 9.90 Hz, 1H, H-1). 13C NMR (D2O, 25 °C): δ ) 26.4 (C-7), 36.9 (C-8), 61.6 (C-6), 67.9 (C-4), 68.3 (C-2), 79.25 (C-5), 82.1 (C-3), 86.3 (C-1), 179.0 (C-9). MALDI-TOF MS m/z 347 [M ]-: C9H15O10S2- (347). Conjugation of Functionalized Saccharides to Polypropylenimine Dendrimers. The peracetylated derivative of compound 1, 2, or 3, 50 to 100 mg (1.5 equiv per dendrimer terminal primary amine), was taken up in CH3CN (10 mg/mL). One equivalent of HATU (Aldrich, Milwaukee, WI) in CH3CN (10 mg/mL) and 1 equiv of DIPEA were added to each sample to activate the terminal carboxyl group (56). After 5-10 min, the appropriate amount of DAB-AM dendrimer at a concentra-
Bioconjugate Chem., Vol. 15, No. 2, 2004 351
tion of 10 mg/mL in DCM was added dropwise to the stirring mixture. Reactions were adjusted to 1:1, DCM: CH3CN, and allowed to incubate at room temperature with stirring overnight. Product formation was monitored by TLC (BuOH:CH3OH:H2O; 2:1:1, v/v/v). Free amino groups were visualized using ninhydrin spray and sugar residues using 5% sulfuric acid in ethanol. Upon completion, the mixture was dried under vacuo and the product deacetylated with Et3N:CH3OH:H2O (2:6:10; v/v/v) at a final concentration of 0.5 mg/mL. After being stirred at 37 °C for 4 h, the reaction was dried, washed with 10 mM HCl in 90% EtOH (51), dried again, and purified on a BioGel P2 column. Fractions were spotted on TLC plates, and carbohydrate-containing conjugates were visualized with 5% H2SO4 in EtOH. Fractions containing the purified glycodendrimers were pooled, dried, and lyophilized. Once dried, the average molecular weights of the glycodendrimers were determined by MALDI-TOF MS. Glycodendrimers that were built by conjugating 1 to DAB-Am dendrimers, generations 1-5, are referred to as glycodendrimers, 1a-e, where 1a is the 4-mer glycodendrimer and 1e is the 64-mer. Likewise, glycodendrimers made from compounds 2 and 3 are referred to as glycodendrimers 2a-e and 3c-e. Direct Sulfation of Dendrimers Derivatized with 3-(β-D-Galactopyranosylthio)propionic Acid (2c, 2d, and 2e). Random sulfation of 2c, 2d, and 2e was done in order to make sulfated 3-(β-D-galactopyranosylthio)propionic acid-derivatized dendrimers. 30 mgs of 2c, 2d, or 2e were taken up in 30 mL of DMF, prior to the addition of SO3NMe3 (2 equiv per hydroxyl group). The reaction mixture was refluxed with stirring at 60 °C for 30 h and then dried by rotary evaporation under reduced pressure supplied by an oil vacuum pump. Dried films were taken up in 2 mL of 1 M NaCl, and 4c, 4d, and 4e were purified by size-exclusion chromatography on a BioGel P2 column, using water as the eluate. Matrix-Assisted Laser Desorption Ionization Time-of-Flight (MALDI-TOF) Mass Spectrometry. MALDI-TOF MS was done on a Perseptive Biosystems Voyager DE-PRO spectrometer. Spectra (100) of negatively charged sugars were generated in linear, negative ion mode using THAP matrix, or in linear, positive ion mode using IAA. Glycodendrimer spectra (200) were generated in linear, positive ion mode using an IAA matrix (20 mg/mL in DMF, diluted 11:1 with a 1 mM aqueous solution of glycodendrimer to yield a matrix-toanalyte ratio of ∼1000:1) (57). In some instances, negative ion mode was used to detect the sulfated glycodendrimers. Average molecular weights and polydispersities of the glycodendrimers were calculated using Data Explorer version 4.0 (Applied Biosystems). Macros within this software allow for the calculation of Mn (the numberaverage molecular weight), Mw (the weight-average molecular weight), and the polydispersity (PD) from the ratio of Mw to Mn according to the following equations:
∑(N M )/∑N ) ∑(N M )/∑N M
Mn ) Mw
i
i
i
2
i
i
i
i
where Ni and Mi represent signal intensity and mass at point i, respectively (58, 59). Nuclear Magnetic Resonance (NMR) Analyses. All spectra were obtained on a Bruker DRX-400 NMR spectrometer operating in the quadrature mode at 25.0 °C using a triple-axis-gradient broadband inverse probe. 1 H chemical shifts were referenced indirectly to TMS (0
352 Bioconjugate Chem., Vol. 15, No. 2, 2004
ppm) via DSS (sodium-2,2-dimethyl-2-silapentane-5-sulfonate), a secondary reference. 13C chemical shifts were referenced indirectly to TMS (0 ppm) via acetone (30.89 and 215.9 ppm), a secondary reference. All samples were dissolved in D2O. One bond 1H-13C correlation 2D spectra were obtained using the gradient-selected version of the phase-sensitive HMQC experiment. Long range (2 and 3 bond) 1H-13C correlation 2D spectra were obtained using the gradient-selected version of the magnitudemode HMBC experiment. When necessary for assignment purposes, gradient-selected Double-Quantum-Filtered COSY experiments were used to verify 1H-1H couplings. Surface Plasmon Resonance. SPR experiments were done at 25 °C using a Biacore 3000 system and CM5 sensor chips (Biacore, Uppsala, Sweden). Glycodendrimers 1c-e and 2c-e were amine-coupled to the sensor chip according to the manufacturer’s protocol, using the free amines that were not functionalized on the glycodendrimers. To obtain a low-density surface suitable for affinity analysis, 350-500 RUs of each glycodendrimer were immobilized. Varying concentrations of rgp120 (62.5-0.98 nM) were injected simultaneously over a blank control surface and the immobilized glycodendrimers at a flow rate of 30 µL/min for 3 min in HBS-EP buffer (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, 0.005% P-20 surfactant, pH 7.4). Following the injection, the dissociation of the rgp120-glycodendrimer complex was measured for 10 min. Any remaining bound rgp120 was removed by a 1 min injection of 4 M KCl. This regeneration step prepared the surface for the next injection of rgp120. Each concentration of rgp120 was injected in duplicate over the immobilized glycodendrimer, with a representative sensorgram chosen for analysis. The sulfated glycodendrimers, 4c-e, did not couple to the CM5 chip, presumably due to repulsion of the negatively charged sulfate groups by the negatively charged carboxyl groups on the carboxymethylated dextran matrix. Therefore, approximately 3000 RUs of rgp120 was amine-coupled to the CM5 sensor chip and the potential ligand used as analyte. DxS (50 kDa, 12.50.39 nM), 4c, 4d, or 4e (89-0.36 nM) were injected simultaneously over a blank control surface and the immobilized rgp120 at a flow rate of 30 µL/min, for 3 min. Dissociation of the complex was measured for 10 min, followed by a 1 min injection of 4 M KCl to regenerate the rgp120 surface. Nonspecific binding was accounted for by subtraction of the blank control surface from each of the test surfaces. Subtracted sensorgrams were then analyzed by fitting the data to a 1:1 (Langmuir) binding model and kinetic constants determined using the BiaEvaluation 3.2 software. Cell Culture. U373-MAGI-CCR5 cells were obtained from the AIDS Reagent Program #3597. Cells were propagated in ‘Selection Media’ consisting of DMEM supplemented with 10% FCS, 0.4 mg/mL L-glutamine, penicillin and streptomycin (0.08 mg/mL each), 0.05% sodium bicarbonate, 0.2 mg/mL G418, 0.1 mg/mL hyrgromycin B, and 1.0 ug/mL puromycin. ‘Culture Media’ for virus inhibition assays consisted of DMEM with 10% FCS and 0.05% sodium bicarbonate. Cells were grown at 37 °C in an atmosphere of 95% air/5% CO2. Viral Inhibition Assays. Cells, in Selection Media, were plated in 96-well tissue culture plates at a density of 1.0 × 104 cells per well. One day later, commercially available (Advanced Biotechnologies, Columbia, MD), concentrated cell-free viral preparations of HIV-1 Ba-L (R5) were diluted 1:400 in ‘Culture Media’. Glycodendrimer and DxS stock solutions were made up in culture media and serially diluted with culture media in separate
Kensinger et al.
sterile V-bottom 96-well microtiter plates, to a final volume of 30 µL per well. This was followed by the addition of 30 µL of the diluted virus and incubation at 37 °C for ∼ 20 min. Selection Media was then removed from the plated cells, and 50 µL of virus or virus plus potential inhibitor was added to the plated cells. The virus- or virus plus inhibitor-exposed cells were then incubated at 37 °C in an atmosphere of 95% air/5% CO2 for 2 h. After the absorption period, 200 µL of Culture Media was added to the wells, and the cells were allowed to grow for an additional 40-48 h. Cells treated with culture media served as negative controls, and cells treated with virus only were positive controls. All measurements were done in quadruplicate. After the 48 h incubation, culture media was removed from the wells, the cells were washed with 200 µL of PBS, and β-galactosidase activity was measured using the Galacto-Star β-Galactosidase Reporter Gene Assay System from Applied Biosystems (Foster City, CA), according to the manufacturer’s instructions. Briefly, after the cells were washed with PBS, 10 µL of lysis solution was added to each well and the plate incubated for 10 min at 37 °C. The Galacto-Star substrate was diluted 1:50 and 100 µL added to each well containing cell lysate. Well contents were mixed and 90 µL were removed and added to a 96well opaque luminometer plate. One hour after the Galacto-Star substrate was added, luminescence was determined using a luminometer. Percent inhibition was determined as [(Lno inhibitor - Linhibitor)/Lno inhibitor] × 100. The effective concentration that inhibited 50% of viral infectivity (EC50) was determined by plotting the percent inhibition versus the log of the concentration of the potential inhibitor. Cytotoxicity. The effect of the glycodendrimers and control compounds on cell viability was determined using the CellTiter 96 AQueous Cell Proliferation Assay (Promega, Madison, WI). In this viability assay, mitochondrial dehydrogenases in metabolically active cells catalyze the conversion of MTS into a formazan product. The absorbance of formazan is read at 490 nm and is directly proportional to the number of viable cells. Assays were performed according to the manufacturer’s instructions. Replicate 96-well plates of each of the HIV infection inhibition assay plates were set up and treated identically, with the exception that no virus was added to the cells. After the 48 h incubation, the cell culture media was removed, cells were washed with PBS, and 100 µL of culture media was added to each well. The MTS reagent was mixed with the electron coupling reagent PMS according to the instructions, and 100 µL of the mixture was added to the cells. The 96-well replicate plates were incubated for 3 h at 37 °C, in an atmosphere of 95% air/5% CO2, and absorbance was read on an ELISA plate reader at 490 nm. Cell viability indices were calculated by dividing the average absorbance of nontreated cells (negative control wells) into the average absorbance obtained for each concentration of each compound tested. Therefore, nontreated cells would have a viability index of 1.0 while that for cells exposed to cytotoxic compounds would be less than 1.0. The index obtained for each concentration of a compound was then used to correct for inhibition due to cell death induced by a potential inhibitor. RESULTS
Synthesis of Spacer-Arm Derivatized Saccharides. To build multivalent GalCer and SGalCer derivatized dendrimers, the saccharide moiety of GalCer and
Synthesis of Novel, Multivalent Glycodendrimers Table 1.
1H
and
13C
Bioconjugate Chem., Vol. 15, No. 2, 2004 353
NMR Chemical Shifts of Compounds 2 and 3. Also Shown Are the Chemical Shifts Relative to 2
compound
H-1
H-2
H-3
H-4
H-5
H-6A
H-6B
2 3
4.46 4.60 (+0.14)
3.52 3.70 (+0.18)
3.61 4.33 (+0.72)
3.93 4.32 (+0.39)
3.67 3.75 (+0.08)
3.72 3.75 (+0.03)
3.66 3.72 (+0.06)
C-1
C-2
C-3
C-4
C-5
C-6
2 3
86.65 86.3 (-0.35)
70.3 68.3 (-2.00)
74.55 82.1 (+7.55)
69.5 67.9 (-1.60)
79.6 79.25 (-0.35)
61.7 61.6 (-0.1)
SGalCer (galactose and galactose-3-sulfate, respectively), containing a portion of the sphingosine moiety, were isolated from the natural GSLs (Scheme 1). Techniques described by Mylvaganum and Lingwood, for the preparation of glycolipid ceramide saccharides for neoglycoconjugation (51), were used for the synthesis of the peracetylated ceramide saccharides directly from GalCer (1) and SGalCer. This was achieved by deacylation of the GSL followed by its acetylation to block vicinal hydroxyls prior to oxidative cleavage of the double bond of the sphingosine to obtain the acetylated “ceramide saccharide” (Scheme 1). The carboxyl group of the N-acetylated butyric acid provided the reactive group needed to link the sugar to a primary amine on the dendrimer core. The yield of ceramide saccharide derivative of SGalCer obtained was very low (
