Single-Walled Carbon Nanotube as a Unique Scaffold for theMultivalent Display of Sugars

Lingrong Gu, Pengju G. Luo, Haifang Wang,† Mohammed J. Meziani, Yi Lin,L. Monica Veca, Li Cao, Fushen Lu, Xin Wang, Robert A. Quinn, Wei Wang, Puyu Zhang,

Sebastian Lacher, and Ya-Ping Sun*

Department of Chemistry and Laboratory for Emerging Materials and Technology, Clemson University,Clemson, South Carolina 29634-0973

Received April 13, 2008; Revised Manuscript Received July 17, 2008

Single-walled carbon nanotube (SWNT) is a pseudo-one-dimensional nanostructure capable of carrying/displayinga large number of bioactive molecules and species in aqueous solution. In this work, a series of dendritic �-D-galactopyranosides and R-D-mannopyranosides with a terminal amino group were synthesized and used for thefunctionalization of SWNTs, which targeted the defect-derived carboxylic acid moieties on the nanotube surface.The higher-order sugar dendrons were more effective in the solubilization of SWNTs, with the correspondingfunctionalized nanotube samples of improved aqueous solubility characteristics. Through the functionalization,the nanotube apparently serves as a unique scaffold for displaying multiple copies of the sugar molecules in pairsor quartets. Results on the synthesis and characterization of these sugar-functionalized SWNTs and their biologicalevaluations in binding assays with pathogenic Escherichia coli and with Bacillus subtilis (a nonvirulent simulantfor Bacillus anthracis or anthrax) spores are presented and discussed.

IntroductionMultivalent carbohydrate ligands are known to be consider-

ably more potent than their monovalent counterparts in biologi-cal ligand-receptor interactions.1,2 Therefore, there has beengrowing recent interest in various multivalent ligands and theirassociated configurations for a wide range of applications,3

including the development of high-affinity inhibitors and drugs.2

Among more extensively studied scaffolds for the display ofmultivalent carbohydrate ligands are dendrimers,4 linear andpolydispersed polymers,5 metal nanoparticles,6 polymericnanospheres,7,8 and so on. For example, Stoddart and co-workerssynthesized a highly branched carbohydrate dendrimer contain-ing 3-36 peripheral R-D-mannose copies and evaluated theirbinding affinity to Con A lectin.9 Disney et al. reported the useof a carbohydrate-conjugated fluorescent polymer, poly(p-phenylene ethynylene), to detect pathogenic Escherichia coli.10

For the binding with E. coli, Wu and co-workers attached R-D-mannose moieties to gold nanoparticles to target selectively thetype 1 pili of the ORN178 strain.6a Qu, et al. developed sugar-coated polystyrene nanobeads for the binding and agglutinationof E. coli cells.7

Single-walled carbon nanotubes (SWNTs) have attractedmuch recent attention for potential biological applications,11-18

including their functionalization with various bioactive speciessuch as carbohydrates,12 DNAs,13 proteins,14 and peptides.15

The unique pseudo-one-dimensional structure of a SWNT hasbeen exploited to serve as a linear and semiflexible nanoscalecarrier for displaying multiple copies of biomolecules in specificinteractions with cells. A representative example is due to Elkinet al.,16 who used the bovine serum albumin-functionalizedSWNTs in conjugation with E. coli-specific antibody to capturethe pathogen in physiological solution. In another example, Gu

et al. solubilized SWNTs via covalent functionalization withthe derivatized 2-aminoethyl-�-D-galactopyranoside (Scheme 1)in likely the amidation of nanotube-bound carboxylic acids.17

These �-D-galactose-functionalized nanotubes (Gal-SWNT),each displaying multiple copies of the sugar, were found to haveadhesion to pathogenic E. coli O157:H7 to result in significantcell agglutination.17 Recently, Wang et al. found that monosac-charides (�-D-galactoses or R-D-mannoses)-functionalizedSWNTs bind to and aggregate effectively Bacillus anthracis(Sterne) spores in the presence of a divalent cation (such asCa2+) and that the binding is unique to the nanotube-displayedcarbohydrates.18

In the functionalization of nanotubes with tethered �-D-galactoses and R-D-mannoses, the carboxylic acid moieties onthe nanotubes, which are resulted from the oxidation of thesurface defects, have been targeted.19 With limited defect siteson the nanotube surface, a higher population of multivalentligands could be attained by using sugar dendrons (Scheme 1)in the nanotube functionalization. The resulting larger numberof displayed ligands per nanotube also corresponds to signifi-cantly improved aqueous solubility of the functionalized nano-tube samples,20 which not only enhances biocompatibility butalso enables more quantitative characterization for a betterunderstanding of the structural details on the nanotubes display-ing multivalent carbohydrate ligands. Here we report thesynthesis and characterization of dendritic �-D-galactopyrano-sides and R-D-mannopyranosides for the functionalization ofSWNTs (Scheme 1). As the functionality varies from mono-or bis- to tetra-, the population of the sugar moieties in thefunctionalized nanotube sample increases, as are improvedsolubility and related properties. Results from the biologicalevaluation of these sugar-functionalized SWNTs in bindingassays with pathogenic E. coli and with Bacillus subtilis (anonvirulent simulant for Bacillus anthracis or anthrax) sporesare presented and discussed.

* To whom correspondence should be addressed. E-mail: syaping@clemson.edu.

† On leave from College of Chemistry and Molecular Engineering, PekingUniversity, Beijing, China.

Biomacromolecules 2008, 9, 2408–24182408

10.1021/bm800395e CCC: $40.75 2008 American Chemical SocietyPublished on Web 08/20/2008

Experimental Section

Materials. t-Butyl bromoacetate, N-hydroxy-succinimide, triph-enylphosphine, trifluoroacetic acid (TFA), palladium (10 wt % onactivated carbon), and Amberlite IR-120 (plus) ion-exchange resin werepurchased from Sigma-Aldrich. 3,5-Dihydroxybenzyl alcohol, 18-crown-6, N,N′-dicyclohexyl-carbodiimide, �-D-galactose-pentaacetate,R-D-mannose, and anthrone were obtained from Acros. N-Bromosuc-cinimide was supplied by ACOCADO Research Chemicals, Ltd.Solvent grade THF was dried and distilled over molecular sieves andthen distilled over sodium before use. Other solvents were eitherspectrophotometry/HPLC grade or purified via simple distillation.Deuterated NMR solvents were obtained from Cambridge IsotopeLaboratories.

The SWNT sample from the arc-discharge production was suppliedby Carbon Solutions, Inc. It was purified by using a combination ofthermal oxidation and oxidative acid treatment. In a typical experiment,a nanotube sample (1 g) was thermally oxidized in air in a furnace at300 °C for 30 min. After the thermal treatment, the remaining samplewas added to an aqueous HNO3 solution (2.6 M), and the mixture wasrefluxed for 48 h. Upon centrifuging at 1380 g (Fisher Scientific Centric

228 Centrifuge), the supernatant was discarded and the remaining solidswere washed with deionized water until neutral pH and then dried undervacuum. Acid-base titration was used to estimate the population ofcarboxylic acid groups in the purified sample. Briefly, a portion of thesample was stirred in aqueous NaHCO3 solution (0.05 M, 50 mL) for3 days. Then, the nanotubes were removed via vigorous centrifuging,and the remaining solution was diluted 1-to-5. One aliquot (50 mL)was mixed with aqueous HCl solution (0.05 M, 10 mL), and the mixturewas boiled for 30 min to remove CO2 completely, followed by thetitration of the excess HCl with aqueous NaOH (0.05 M).

Measurements. NMR measurements were carried out on a JEOLEclipse +500 NMR spectrometer and a Bruker Advance 500 NMRspectrometer equipped with a high-resolution magic-angle-spinning(HR-MAS) probe designed specifically for gel-phase NMR. Matrix-assisted laser desorption ionization-time-of-flight (MALDI-TOF) MSwas performed on a Bruker AutoFlex system, and 2,5-dihydroxybenzoicacid was used as the sample matrix. Thermogravimetric analysis (TGA)results were obtained on a TA Instruments Q500 TGA, with a scanningrate of 10 °C/min for measurements in nitrogen or air. Opticalabsorption spectra were recorded on a Shimadzu UV3600 UV/vis/NIR

Scheme 1. Carbon Nanotubes Functionalized with Gal-, Man-, or their Dendrons

Scaffold for the Multivalent Display of Sugars Biomacromolecules, Vol. 9, No. 9, 2008 2409

spectrophotometer. Raman spectra were measured on a Johin YvonT64000 Raman spectrometer equipped with a Melles-Griot 35 mW He:Ne laser source for 633 nm excitation, a triple monochromator, aresearch grade Olympus BX-41 microscopy and a liquid nitrogen-cooledsymphony detector. Optical microscopy measurements were performedon a Zeiss Axiovert 135 fluorescence microscope. Scanning electronmicroscopy (SEM) images were obtained on a Hitachi S4800 field-emission SEM system and a Hitachi HD-2000 TEM/STEM system.

Sugar Dendrons. The sugar monomers (Gal- and Man-) wereprepared using procedures already available in the literature (SupportingInformation).22 For the sugar dendrons, the compound 1 (Scheme 2)was synthesized from t-butyl bromoacetate (through iodide substitution)and 3,5-dihydroxybenzyl alcohol also according to procedures alreadyreported in the literature (Supporting Information).23

Compound 2. A solution of 1 (1.51 g, 3.5 mmol) in dry CH2Cl2 (10mL) was prepared, and at 0 °C, trifluoroacetic acid (4 mL) was added.The mixture was warmed back to room temperature and stirred for

12 h. The precipitate was collected via filtration and then dried in avacuum oven to yield the hydrolyzed product 1′ as a white solid (1.1

Scheme 2. Syntheses of Various Sugar Dendrons

Table 1. Results on Man-, Man2-, and Man4- in Functionalization/Solubilization of SWNTs

starting SWNTs solubilizeda (%)molar ratio of sugar-NH2

to nanotube-COOH Man- Man2- Man4-

1:1 7.52:1 134:1 205:1 246:1 2320:1 24100:1 22 35 50

a Some might argue that nanotubes are dispersed, not solubilized.However, in the literature, the kind of functionalization used in this workis generally referred to as solubilization.

2410 Biomacromolecules, Vol. 9, No. 9, 2008 Gu et al.

g, quantitative). 1H NMR (500 MHz, CD3OD) δ 6.66 (d, 2H), 6.51 (t,1H), 4.67 (s, 4H), 4.51 (s, 2H) ppm; 13C NMR (125.7 MHz, CD3OD)δ 170.96, 159.22, 140.54, 108.13, 101.56, 64.62, 32.26 ppm.

At 0 °C, a solution of N,N′-dicyclohexyl-carbodimide (990.3 mg,4.8 mmol) in dry THF (20 mL) was added dropwise over 30 min toanother THF solution (30 mL) of 1′ (638.7 mg, 2 mmol) and N-hydroxy-succinimide (552.4 mg, 4.8 mmol). The mixture was warmed back toroom temperature and stirred in the dark for 12 h. The precipitatedurea was removed via filtration, and the remaining solution was slowlydropped into a CH2Cl2 solution (20 mL) of freshly prepared Gal′-NH2

(2.0 g, 4.8 mmol).20 After stirring for 12 h and then solvent removal,the reaction mixture was separated on a silica gel column (firsthexane-ethyl acetate at 10/90 v/v, and then methanol-ethyl acetateat 5/95 v/v) to yield 2(Gal′) as a white solid (2.13 g, 99%). 1H NMR(500 MHz, CDCl3) δ 6.96 (t, 2H, NH), 6.71 (d, 2H), 6.64 (t, 1H), 5.41(dd, 2H), 5.23 (dd, 2H), 5.03 (dd, 2H), 4.52 (s, 4H), 4.51(d, 2H), 4.45(s, 2H), 4.19-4.14 (m, 4H), 4.0-3.92 (m, 4H), 3.74-3.67 (m, 4H),3.55-3.48 (m, 2H), 2.18 (s, 6H), 2.07 (s, 6H), 2.04 (s, 6H), 2.02 (s,6H) ppm; 13C NMR (125.7 MHz, CDCl3) δ 170.36, 170.17, 170.10,169.46, 167.68, 158.59, 140.84, 109.09, 102.38, 101.30, 70.86, 70.75,68.79, 68.49, 67.51, 66.96, 61.25, 38.82, 32.65, 20.70, 20.65 (2C), 20.57ppm. MALDI-TOF MS (M+Na+): 1089.37 (theoretically 1087.24).

The same procedure was applied to the synthesis of 2(Man′) (2.13g, 99%). 1H NMR (500 MHz, CDCl3): δ 7.08 (t, 2H, NH), 7.01 (d,2H), 6.59 (t, 1H), 5.37 (dd, 2H), 5.31-5.27 (m, 4H), 4.87 (d, 2H),4.55 (s, 4H), 4.46 (s, 2H), 4.28-4.25 (dd, 2H), 4.15-4.12 (m, 2H),4.00-3.97 (m, 2H), 3.88-3.85 (m, 2H), 3.71-3.64 (m, 2H), 3.63-3.57

(m, 4H), 2.18 (s, 6H), 2.12 (s, 6H), 2.05 (s, 6H), 2.00 (s, 6H) ppm;13C NMR (125.7 MHz, CDCl3): δ 170.63, 170.09, 170.06, 169.68,167.83, 158.43, 140.75, 108.99, 102.09, 97.58, 69.36, 68.96, 68.92,67.25, 66.82, 66.05, 62.52, 38.38, 32.66, 20.85, 20.70 (2C), 20.61 ppm.MALDI-TOF MS (M + Na+): 1088.54 (theoretically 1087.24).

Gal2-N3 and Man2-N3. The azide substitution and deprotectionreactions of 2 were similar to those reported previously,20 yieldingquantitatively Gal2-N3 [1H NMR (500 MHz, CD3OD) δ 6.65-6.63(m, 3H), 4.50 (s, 4H), 4.29 (s, 2H), 4.19 (d, 2H), 3.95-3.86 (m, 2H),3.80-3.78 (dd, 2H), 3.72-3.62 (m, 6H), 3.57-3.38 (m, 10H) ppm;13C NMR (125.7 MHz, CD3OD) δ 169.58, 159.17, 138.59, 107.72,103.86, 101.72, 75.38, 73.49, 71.15, 68.89, 68.14, 66.99, 61.14, 53.94,39.00 ppm. MALDI-TOF MS (M + Na+): 714.83 (theoretically,714.24)] and Man2-N3 [1H NMR (500 MHz, CD3OD) δ 6.66-6.65(m, 3H), 4.74 (d, 2H), 4.54 (s, 4H), 4.33 (s, 2H), 3.81-3.74 (m, 6H),3.69-3.64 (m, 4H), 3.61-3.51 (m, 8H), 3.48-3.43 (m, 2H) ppm; 13CNMR (125.7 MHz, CD3OD) δ 169.70, 159.18, 138.71, 107.72, 101.65,101.33, 73.51, 71.23, 70.73, 67.26, 67.01, 65.73, 61.58, 54.02, 38.62ppm. MALDI-TOF MS (M+): 690.77 (theoretically, 691.25)]. Thesecompounds were relatively stable, and they were freshly reduced toGal2-NH2 and Man2-NH2 before the nanotube functionalization reaction.

Figure 1. Comparisons for (a) the 1H NMR spectra of Gal2-SWNT(top, solution phase; middle, gel phase) and Gal2-N3 (bottom); and(b) the 13C NMR spectra of Gal2-SWNT (top, with inset) and Gal2-N3

(bottom), all in D2O. The peak marked with * was due to residualmethanol in the sample.

Figure 2. Comparisons for (a) the 1H NMR spectra of Gal4-SWNT(top) and Gal4-N3 (bottom); and (b) the 13C NMR spectra of Gal4-SWNT (top, with inset) and Gal4-N3 (bottom), all in D2O. The peaksmarked with * were due to residual methanol in the sample.

Scaffold for the Multivalent Display of Sugars Biomacromolecules, Vol. 9, No. 9, 2008 2411

Compound 3. There were two steps (Scheme 2), with the firststep being the conversion of the bromide in 1 to iodide (1′′). Compound1 (3.018 g, 7 mmol) and sodium iodide (10 g, 67 mmol) were refluxedin acetone (50 mL) for 12 h. The reaction mixture was allowed to cool,and then the solvent was evaporated completely. The residue waspartitioned between water and chloroform. The organic layer wascollected and dried with anhydrous MgSO4, followed by the removalof solvent to yield 1′′ (3.56 g, quantitative). 1H NMR (500 MHz, CDCl3)δ 6.53 (d, 2H), 6.37 (t, 1H), 4.48 (s, 4H), 4.35 (s, 2H), 1.50 (s, 18H)

Figure 3. Optical absorption spectra of (a) Gal2-SWNT; (b) Gal4-SWNT; (c) Man2-SWNT; and (d) Man4-SWNT in D2O solution (solid line) andin the solid-state on glass substrate (dashed line). For the spectral comparison between solution-phase and solid-state, the spectra in eachcorresponding pair were normalized by multiplying factors to the spectra to make their absorbance values the same at a selected wavelength(1000 nm for Man2-SWNT as an example).

Figure 4. Raman spectra (633 nm excitation) of the functionalizedSWNTs and the starting purified nanotube sample are compared.

Figure 5. SEM images of a Gal2-SWNT specimen (ultrathin film): (a)at the edge of the specimen and (b) in the fractured portion of thespecimen.

2412 Biomacromolecules, Vol. 9, No. 9, 2008 Gu et al.

ppm; 13C NMR (125.7 MHz, CDCl3) δ 167.74, 159.17, 141.42, 108.16,101.75, 82.64, 65.80, 28.16, 5.14 ppm.

In the second step, a solution of 3,5-dihydroxybenzyl alcohol (490mg, 3.5 mmol), K2CO3 (2 g, 14 mmol), and 18-crown-6 (172 mg, 0.7mmol) in acetone (30 mL) was prepared, and to the solution was added1′′ (3.56 g, 7 mmol). After refluxing for 36 h and then solvent removal,the reaction mixture was partitioned between water and ethyl acetate.The aqueous portion was repeatedly extracted with ethyl acetate. Thecombined ethyl acetate solution was dried with anhydrous MgSO4,followed by silica gel column (hexane-ethyl acetate at 1:1.5) separationto obtain 3 (2.55 g, 87% yield). 1H NMR (500 MHz, CDCl3) δ 6.56(d, 4H), 6.55 (d, 2H), 6.45 (t, 1H), 6.40 (t, 2H), 4.93 (s, 4H), 4.60 (s,2H), 4.47 (s, 8H), 1.46 (s, 36H) ppm; 13C NMR (125.7 MHz, CDCl3)δ 167.87, 159.95, 159.24, 143.88, 139.60, 106.48, 105.72, 101.40,101.23, 82.49, 69.71, 65.76, 65.03, 28.09 ppm. MALDI-TOF MS (M+ Na+): 863.29 (theoretically, 863.38).

Compound 4. The first step was to convert the benzyl alcohol in 3to benzyl bromide (3′). Compound 3 (3.075 g, 3.675 mmol) was addedto a THF (20 mL) solution of triphenyl phosphine (1.44 g, 5.5 mmol).After stirring for 7 min, N-bromosuccinimide (979 mg, 5.5 mmol) wasadded and the mixture was stirred for 5 min more. Water was added toquench the reaction. The crude product was extracted with chloroformthree times. The organic fraction was dried with anhydrous MgSO4,condensed, and purified by silica gel chromatography with hexane-ethylacetate (3/1 v/v) as eluent. Compound 3′ was obtained as a white solid(2.76 g, 83% yield). 1H NMR (500 MHz, CDCl3) δ 6.59 (d, 2H), 6.58(d, 4H), 6.47 (t, 1H), 6.44 (t, 2H), 4.93 (s, 4H), 4.49 (s, 8H), 4.39 (s,2H), 1.48 (s, 36H) ppm; 13C NMR (125.7 MHz, CDCl3) δ 167.72,159.89, 159.27, 139.84, 139.27, 108.25, 106.47, 102.16, 101.50, 82.36,69.81, 65.79, 33.44, 28.04 ppm. MALDI-TOF MS (M + Na+): 927.31(theoretically, 925.30).

A solution of 3′ (2.4 g, 2.6 mmol) in CH2Cl2 (20 mL) was prepared,and at 0 °C, trifluoroacetic acid (4 mL) was added. The mixture waswarmed back to room temperature and stirred for 12 h. The precipitatewas collected by vacuum filtration and then dried in a vacuum oven toyield 4 as a white solid (1.78 g, quantitative). 1H NMR (500 MHz,CD3OD) δ 6.68 (d, 2H), 6.67 (t, 4H), 6.56 (t, 1H), 6.51 (t, 2H), 5.02(s, 4H), 4.66 (s, 8H), 4.49 (s, 2H) ppm; 13C NMR (125.7 MHz, CD3OD)δ 171.07, 159.98, 159.33, 140.35, 139.87, 108.29, 106.21, 101.81,101.03, 69.42, 64.65, 32.59 ppm. MALDI-TOF MS (M + Na+): 703.19(theoretically, 701.05).

5(Gal′) and 5(Man′). A total of two THF solutions (10 mL each)were prepared, with one for 4 (340 mg, 0.5 mmol) and N-hydroxy-succinimide (345 mg, 3 mmol) and the other for N,N′-dicyclohexyl-carbodiimide (619 mg, 3 mmol). At 0 °C, the second solution was addedto the first dropwise over 30 min. The reaction mixture was graduallywarmed to room temperature and stirred in the dark for 12 h. Theinsoluble urea was removed by filtration, and the filtrate was slowlydropped into a CH2Cl2 solution (20 mL) of freshly prepared Gal′-NH2

(1.0 g, 2.4 mmol).20 After stirring for 12 h and then solvent removal,the reaction mixture was separated on silica gel column (firsthexane-ethyl acetate at 1:9 and then methanol-ethyl acetate at 1:9)to afford 5(Gal′) as white solids (728 mg, 67% yield). 1H NMR (500MHz, CDCl3) δ 6.94 (t, 4H, NH), 6.66 (d, 4H), 6.59 (d, 2H), 6.58 (d,2H), 6.48 (t, 1H), 5.31 (d, 4H), 5.13 (dd, 4H), 4.98-4.92 (m, 8H),4.44 (s, 8H), 4.42 (d, 4H), 4.36(s, 2H), 4.10-4.02 (m, 8H), 3.90-3.82(m, 8H), 3.68-3.56 (m, 8H), 3.46-3.40 (m, 4H), 2.07 (s, 12H), 1.97(s, 12H), 1.94 (s, 12H), 1.92 (s, 12H) ppm; 13C NMR (125.7 MHz,CDCl3) δ 170.44, 170.26, 170.18, 169.55, 167.89, 158.79, 158.71,140.13, 140.02, 108.31, 107.03, 102.20, 102.16, 101.29, 70.83, 70.77,69.59, 68.78, 68.55, 67.47, 67.00, 61.31, 38.85, 33.40, 20.75, 20.72,20.70, 20.65 ppm. MALDI-TOF MS (M + Na+): 2195.56 (theoreti-cally, 2193.60).

The same procedure was applied to the synthesis of 5(Man′) (1.37g, 63% yield). 1H NMR (500 MHz, CDCl3) δ 7.06 (t, 4H, NH), 6.68(d, 4H), 6.60 (d, 2H), 6.56 (t, 2H), 6.50 (t, 1H), 5.32-5.29 (tt, 4H),5.24-5.18 (m, 8H), 4.97 (s, 4H), 4.81 (s, 4H), 4.50 (s, 8H), 4.38(s,2H), 4.23-4.19 (m, 4H), 4.09-4.03 (tt, 4H), 3.95-3.92 (m, 4H),3.82-3.78 (m, 4H), 3.66-3.49 (m, 12H), 2.11 (s, 12H), 2.06 (s, 12H),1.96 (s, 12H), 1.94 (s, 12H) ppm; 13C NMR (125.7 MHz, CDCl3) δ170.73,, 170.15 (2C), 169.77, 168.00, 159.85, 158.60, 140.09, 139.95,108.30, 107.05, 102.20, 101.96, 97.65, 69.60, 69.42, 69.03, 68.92, 67.31,66.91, 66.10, 62.56, 38.47, 33.48, 20.94, 20.79, 20.76, 20.66 ppm.MALDI-TOF MS (M + Na+): 2194.28 (theoretically, 2193.60).

Gal4-N3 and Man4-N3. The first step was to convert the bromide in5 to azide (Gal′4-N3 and Man′4-N3). For Gal′4-N3, a solution of 5(Gal′)(728 mg, 0.34 mmol) in acetone (20 mL) was prepared, and to thesolution was added sodium azide (0.22 g, 3.4 mmol) and 18-crown-6(54 mg, 0.2 mmol). The mixture was refluxed for 12 h and then cooledto room temperature. The solvent was evaporated, and the residue waspartitioned between water and ethyl acetate. The aqueous layer wasextracted with ethyl acetate twice. All of the ethyl acetate fractionswere combined and dried with anhydrous MgSO4. The solvent wasevaporated, and the residue was dried in a vacuum oven to yield Gal′4-N3 as a white solid (715 mg, quantitative). 1H NMR (500 MHz, CDCl3)δ 6.98 (t, 4H, NH), 6.65 (s, 4H), 6.59 (s, 2H), 6.52 (s, 1H), 6.50 (s,2H), 5.32 (d, 4H), 5.13 (dd, 4H), 4.96 (s, 4H), 4.93 (d, 4H), 4.44 (s,8H), 4.42 (d, 4H), 4.22 (s, 2H), 4.07 (d, 8H), 3.90-3.83 (m, 8H),3.68-3.56 (m, 8H), 3.45-3.40 (m, 4H), 2.07 (s, 12H), 1.96 (s, 12H),1.94 (s, 12H), 1.91 (s, 12H) ppm; 13C NMR (125.7 MHz, CDCl3) δ170.47, 170.29, 170.19, 169.61, 167.97, 159.96, 158.70, 140.03, 137.98,107.33, 106.97, 102.13, 101.89, 101.25, 70.81, 70.77, 69.57, 68.79,68.53, 67.43, 67.01, 61.32, 54.69, 38.86, 20.75, 20.72, 20.70, 20.65ppm. MALDI-TOF MS (M + Na+): 2157.43 (theoretically, 2156.69).The same procedure was applied to the synthesis of Man′4-N3 (1.02 g,quantitative). 1H NMR (500 MHz, CDCl3) δ 7.01 (t, 4H, NH), 6.64(d, 4H), 6.53 (t, 2H), 6.51 (t, 1H), 6.49 (d, 2H), 5.29-5.26 (dd, 4H),5.22-5.17 (m, 8H), 4.95 (s, 4H), 4.78 (d, 4H), 4.46 (s, 8H), 4.20-4.17(m, 6H), 4.03(dd, 4H), 3.93-3.89 (m, 4H), 3.79-3.74 (m, 4H),3.63-3.57 (m, 4H), 3.57-3.46 (m, 8H), 2.08 (s, 12H), 2.02 (s, 12H),1.93 (s, 12H), 1.91 (s, 12H) ppm; 13C NMR (125.7 MHz, CDCl3) δ170.74, 170.16 (2C), 169.76, 168.00, 160.05, 158.51, 139.99, 137.95,107.35, 107.00, 101.95, 101.89, 97.66, 69.59, 69.42, 69.03, 68.91, 67.30,66.90, 66.10, 62.54, 54.77, 38.47, 20.95, 20.79, 20.75, 20.66 ppm.MALDI-TOF MS (M+): 2134.08 (theoretically, 2133.69).

For Gal4-N3, a solution of Gal′4-N3 (715 mg, 0.334 mmol) inCH3ONa/CH3OH (0.05 M, 30 mL) was prepared. After the solutionwas stirred for 12 h, Amberlite IR-120(plux) ion-exchange resin wasadded to adjust the pH to 7.0. The resin was removed via filtration,and the solution was dried with anhydrous MgSO4, followed by solventevaporation to yield Gal4-N3 as a white solid (490 mg, quantitative).1H NMR (500 MHz, D2O) δ 6.47 (s, 4H), 6.43 (s, 2H), 6.36 (s, 2H),6.26 (s, 1H), 4.63 (s, 4H), 4.32 (d, 4H), 4.26 (s, 8H), 4.13 (s, 2H),3.97-3.94 (m, 4H), 3.88 (d, 4H), 3.77-3.70 (m, 12H), 3.62-3.57 (m,8H), 3.54-3.50 (m, 8H), 3.47-3.41 (m, 4H) ppm; 13C NMR (125.7MHz, D2O) δ 170.53, 159.21, 158.31, 139.53, 138.37, 107.24, 106.85,

Figure 6. SEM image of the Gal4-SWNT specimen (on carbon-coatedcupper grid) prepared from a very dilute sample solution.

Scaffold for the Multivalent Display of Sugars Biomacromolecules, Vol. 9, No. 9, 2008 2413

103.18, 101.40 (2C), 75.15, 72.77, 70.76, 69.19, 68.59, 68.33, 66.41,60.95, 54.02, 39.07 ppm. MALDI-TOF MS (M + Na+): 1484.50(theoretically 1484.52). The same procedure was applied to thequantitative conversion of Man′4-N3 to Man4-N3. 1H NMR (500 MHz,D2O) δ 6.41 (s, 4H), 6.37 (s, 2H), 6.34 (s, 2H), 6.23 (s, 1H), 4.75 (s,4H), 4.61 (s, 4H), 4.24 (s, 8H), 4.05 (s, 2H), 3.81 (m, 4H), 3.76 (d,4H), 3.70-3.65 (m, 12H), 3.62-3.57 (m, 4H), 3.54-3.50 (m, 8H),3.47-3.42 (m, 4H), 3.36-3.32 (m, 4H) ppm; 13C NMR (125.7 MHz,D2O) δ 169.42, 159.23, 158.34, 139.63, 138.36, 107.43, 106.79, 101.57,101.30, 99.67, 72.88, 70.64, 70.10, 69.20, 66.69, 66.50, 65.70, 60.93,54.02, 38.61 ppm. MALDI-TOF MS (M+): 1461.38 (theoretically,1461.53).

Nanotube Functionalization. Before the functionalization reaction,Gal4-N3 and Man4-N3 were reduced to Gal4-NH2 and Man4-NH2 bythe classical palladium-catalyzed hydrogenation. At 0 °C, to a solutionof Gal4-N3 or Man4-N3 (800 mg, 0.55 mmol) in methanol-H2O (10mL) was added Pd/C (10 wt % palladium on activated carbon, 70 mg).The reaction mixture was gradually warmed to room temperature andstirred with the purging of hydrogen gas for 4 h. The Pd/C was removedby filtration, and the filtrate was evaporated to obtain Gal4-NH2 orMan4-NH2.

In the functionalization of SWNTs with Gal4-NH2, a purifiednanotube sample (20 mg) was mixed with 1-ethyl-3-(3-dimethylami-noprop-yl) carbodiimide (EDAC, 112 mg, 60 mmol) in aqueousKH2PO4 buffer (45 mL, pH ) 7.4). Upon sonication for 2 h, the freshlyprepared Gal4-NH2 (0.55 mmol) was added. After sonication for another36 h, the reaction mixture was loaded into a membrane tubing (cutoffmolecular weight ∼ 12000) for dialysis against deionized water for 3days. The suspension from the dialysis was centrifuged at 3000 g for30 min to obtain a dark but optically transparent solution of Gal4-SWNT. 1H NMR (500 MHz, D2O) 6.72 (broad), 6.61 (s, broad), 6.55(broad), 6.50-6.40 (broad), 4.50 (s, broad), 4.35 (d, broad), 3.98 (s,broad), 3.93 (s, broad), 3.76 (s, broad), 3.64 (broad), 3.54 (broad), 3.50(broad) ppm; 13C NMR (125.7 MHz, D2O) δ 170.99, 158.48, 106.61,103.17, 75.14, 72.74, 70.77, 69.58, 68.60, 68.36, 66.60, 60.94, 39.06ppm. Similarly, for Man4-SWNT, 1H NMR (500 MHz, D2O) 6.5-6.2(broad), 4.35-4.10 (broad), 3.91 (s, broad), 3.85 (d, broad), 3.77 (d,broad), 3.67 (t, broad), 3.55-3.40 (broad), 3.43 (broad) ppm; 13C NMR(125.7 MHz, D2O) δ 170.41, 159.19, 158.28, 139.46, 106.98, 101.21,99.66, 72.87, 70.61, 70.10, 69.18, 66.70, 66.44, 65.66, 60.93, 38.58ppm.

Sugar Test. The spectrophotometric method with the anthronereagent was used to determine sugar contents.21 In a typical experiment,a solution of sugar-functionalized SWNTs (50 µL) was mixed withdeionized water (50 µL), HCl (37%, 1 mL), formic acid (0.1 mL), andanthrone reagent (8 mL, 0.2 mg/mL in 80% H2SO4). The solutionmixture was kept in a boiling water bath for 12 min and then rapidlycooled in an ice bath for 30 min. The UV/vis absorption of the mixturewas measured and corrected for the solution absorption preheating(before the coloring process). Separately, the standard sugar (�-D-galactose or R-D-mannose) solutions of known concentration (1.0 mg/mL) were prepared. Various aliquots of the standard solutions (0, 25,50, 75, and 100 µL) were tested by using the same procedure above to

establish standard curves, from which the sugar content in thefunctionalized nanotube sample was obtained.

Assay for E. coli. The green fluorescent protein (GFP)-expressingE. coli O157:H7/pWM1007 was transformed through electroporation.24

The frozen E. coli samples were recovered and cultured on tryptic soyagar (TSA) plates supplemented with kanamycin (50 µg/mL). Afterovernight growth at 37 °C, the bacteria were harvested and washedwith phosphate buffer solution (PBS) via the centrifuging-suspendingcycle three times. Finally, the E. coli cells were suspended in PBS toan optical density of 0.6 at 600 nm (∼108 cells/mL).

In cell adhesion and precipitation experiment, a suspension (25 µL)of E. coli O157:H7/pWM1007 cells was 2-fold serially diluted andincubated with a solution (10 µL) of either Gal-SWNT (0.14 mg/mL)or Gal2-SWNT (0.13 mg/mL) for 3 h at room temperature with gentleinversions of the tubes at 15 min intervals. The precipitates at the tubebottom were collected and wet-mounted onto a glass slide formicroscopy analyses.

For the colony-forming unit (CFU) reduction assay, the sameprocedure described above was used, and at the end of incubation themixture was centrifuged with a low force (300 g) for 30 s. Thesupernatant was collected and 10-fold serially diluted. At each dilution,100 µL was plated out (in triplicate) onto TSA agar plates by usingspread plate technique. The plates were incubated overnight at 37 °C.CFU was counted, and the percentage of CFU reduction was calculatedand compared to that of the control without the sugar-functionalizedSWNTs.

Assay for B. subtilis Spores. B. subtilis spores (strain ATCC33234)were supplied by American type Culture Collection (Manassas, VA),and the spore suspension was prepared by following establishedprocedures. In a typical binding experiment, an aqueous suspension ofB. subtilis spores (40 µL, 2.6 × 108 CFU/mL) was mixed with a solutionof R-D-mannose-functionalized SWNTs (40 µL, 0.2 mg/mL R-D-mannose equivalent concentration) or with distilled water (40 µL) ascontrol, followed by the addition of aqueous CaCl2 (20 µL, 100 mM).The mixtures (sample and control) were rotated for 12 h. A small aliquot(10 µL) of the sample (or control) was dropped onto a glass slide(covered with cover glass slide) for optical microscopy analyses.

Results and Discussion

The sugar dendrons Gal2-NH2, Man2-NH2, Gal4-NH2, andMan4-NH2 were prepared by using trisubstituted benzenes (3,5-dihydroxybenzyl alcohol and 1) as building blocks in classicaletherification and carbodiimide-activated amidation reactions(Scheme 2), with generally high product yields. The syntheticstrategy was such that the dendritic framework was constructedfirst before being coupled with the amine-tethered monosac-charide. A significant advantage of the strategy was that sugarmolecules were involved only in the end steps, thus difficultiesassociated with the sugar deprotection and stereohindrance wereavoided.

The same carbodiimide-activated amidation was used in thefunctionalization of SWNTs with the dendrons, with the tethered

Table 2. Compositions in the Functionalized Nanotube Samples and Related Parameters

sugarcontenta (%)

SWNTcontenta (%)

avg No. ofnanotube carbon per

sugar unitbaqueous solubility

(SWNT-equivalent, mg/mL)starting SWNTssolubilized (%)

Gal-SWNT 45 43 14 0.83 11Man-SWNT 47 42 13 0.84 22Gal2-SWNT 44 19 6 1.7 20Man2-SWNT 45 18 6 1.6 35Gal4-SWNT 37 23 9 3.1 27Man4-SWNT 35 29 11 4.3 50a The remaining (100% - sugar content - SWNT content) is the content of linkers/tethers. b The degree of functionalization exceeding that for the

estimated COOH population might be due to preferable solubilization of those nanotubes with more than average COOH contents and also some noncovalentbut strong adsorption of the amines.

2414 Biomacromolecules, Vol. 9, No. 9, 2008 Gu et al.

amino groups coupling with the surface defects-derived car-boxylic acids on the nanotubes.19,25 The functionalizationreaction was heterogeneous in nature, so that as typical an excessamount of the functionalization agent was used. For thefunctionalization with 2′-aminoethyl-R-D-mannopyranoside(Man-), the molar ratio of Man- to the nanotube-boundcarboxylic acids (estimated at 5% mole fraction of the nanotubecarbons based on titration results) was varied. The results inTable 1 show that the amount of SWNTs solubilized in thefunctionalization reached a plateau at the Man-to-acid molarratio of 5 or so. Thus, at the molar ratio of 100, the amount ofsolubilized nanotubes should represent the limit for the func-tionalization agent under the otherwise defined experimental

parameters and conditions. As shown in Table 1, the percentageof the starting purified SWNTs solubilized in the functional-ization reaction increased from about 25% for Man- to 35%for Man2- and to 50% for Man4-. The higher-order sugardendrons were obviously more effective in the functionalizationand solubilization of SWNTs.

The sugar dendron-functionalized carbon nanotube samples,Gal2-SWNT, Man2-SWNT, Gal4-SWNT, and Man4-SWNT, areall readily soluble in water (see note in Table 1), allowing theirbeing investigated with solution-phase techniques. The first wasthe characterization of the samples by NMR, both in solutionand in the gel phase.26 For example, the 1H NMR spectra of

Figure 7. Fluorescence microscopy images (10 µm for all scale bars)on the agglutination of E. coli O157:H7 (GFP-expressing) cells by(a) Gal2-SWNT, (b) Gal-SWNT, and (c) the control (cells only). (d) Avisual comparison on the amount of precipitates (in centrifuge tubes)associated with the agglutination of E. coli O157:H7 cells (top, 5 ×107 cells/mL; bottom, 108 cells/mL) by Gal2-SWNT (left) and Gal-SWNT (right).

Figure 8. Results from the CFU reduction assay for E. coli O157:H7in the presence of Gal2-SWNT, Gal-SWNT, and PBS only as control.

Figure 9. Optical microscopy images on the aggregation of B. subtilisspores by (a) Man-SWNT, (b) Man2-SWNT, and (c) Man4-SWNT inthe presence of calcium cation (all scale bars ) 75 µm). (d) Thecontrol with B. subtilis spores alone (scale bar ) 20 µm).

Scaffold for the Multivalent Display of Sugars Biomacromolecules, Vol. 9, No. 9, 2008 2415

Gal2-SWNT and Gal2-N3 in solutions are compared in Figure1a. The signals of the nanotube sample are obviously signifi-cantly broader than those of the corresponding free dendron,which may be attributed to the high molecular weight and lowmobility of carbon nanotubes.26a For the gel-phase NMRmeasurement of the same Gal2-SWNT sample in a highresolution-magic angle spinning (HR-MAS) probe specificallyfor the gel-phase, the 1H spectrum is better resolved, with thesignal patterns in the sugar region comparable with those ofthe free dendron (Figure 1a). The aromatic proton signals forGal2-SWNT are slightly shifted downfield, 6.82 ppm and 6.70ppm for the nanotube-bound dendron from 6.72 ppm and 6.66ppm for the free dendron, respectively, which might be due tothe formation of amide linkages affecting the aromatic ring inthe close vicinity. The same effects could be responsiblefor the absence of the benzyl proton signal in the spectrum ofGal2-SWNT because the proton is right next to the nanotube.This proton has a chemical shift of 4.36 ppm in the spectrumof Gal2-N3. It is known in the literature that the effect of largearomatic ring currents in carbon nanotubes on the local magneticenvironment of protons in close proximity causes their resonanceto either shift into the region where it is difficult to be identifiedor to become too broad to be detected.26,27

Similar 1H NMR results, including effects on the signalbroadening and shifting, were found for other dendron-func-tionalized SWNTs. Shown in Figure 2a is a similar comparisonbetween the 1H NMR spectra of Gal4-SWNT and Gal4-N3 insolutions, and the comparison between Man4-SWNT and Man4-N3 is largely the same.

13C NMR spectra of the dendron-functionalized SWNTs,especially Gal4-SWNT and Man4-SWNT, were readily obtainedin solution, unlike those of Gal-SWNT and Man-SWNT whoserelatively poor solubility (thus low solution concentrations) madetheir 13C NMR measurements rather difficult.17 For the com-parisons between Gal2-SWNT and Gal2-N3 (Figure 1b) andbetween Gal4-SWNT and Gal4-N3 (Figure 2b), there are signalbroadening effects from free dendrons to those attached toSWNTs, though relatively less so from Gal4-N3 to Gal4-SWNTas one would expect. There are Gal carbon resonances in bothGal2-SWNT (Figure 1b) and Gal4-SWNT (Figure 2b) spectra,and the latter also exhibits some of the tether carbon signalsdue to the higher solution concentration and perhaps also tothese carbons being farther away from the nanotubes (lessaffected by their large ring currents26a,27).

The optical absorption spectra of the functionalized SWNTswere measured both in solution (only down to 1500 nm becauseof the overwhelming interference from D2O beyond that) andin the solid state (samples deposited on the surface of glassslides). The S11 (∼1880 nm) and S22 (∼1050 nm) bands due tovan Hove singularity transitions in semiconducting SWNTs28

and the weak M11 transition (∼740 nm) in metallic SWNTswere all present in the spectra of dendron-functionalizednanotube samples (Figure 3), suggesting the electronic propertiesof the nanotubes were largely preserved in the functionalizationtargeting defects-derived carboxylic acid moieties on the nano-tube surface, as also observed in many other such functional-ization schemes.17,26 The solution spectra of Gal4-SWNT andMan4-SWNT were somewhat better resolved than their solid-state counterparts (Figure 3). For Gal2-SWNT and Man2-SWNT,the generally weaker absorption bands in the solution spectrawere probably a simple result of their lower solubility (thuslower solution concentrations).

The sugar dendron-functionalized SWNTs were characterizedby Raman with 633 nm excitation. Interestingly, unlike in other

well-functionalized nanotube samples,26,29 no substantial lumi-nescence interference was observed. Thus, resonance Ramanspectra could be measured for all of the dendron-SWNT sampleswithout their being thermally or chemically defunctionalizedfirst. As compared in Figure 4, the Raman spectra of the differentdendron-SWNT samples are largely similar among themselvesand also similar to that of the starting purified nanotube sample,exhibiting the typical radial breathing mode peaks around 140and 160 cm-1, D-band around 1300 cm-1, tangential G-bandaround 1580 cm-1, and D*-band around 2600 cm-1.

Scanning electron microscopy (SEM) technique was used toexamine the morphology of sugar dendron-functionalized nano-tube samples. Shown in Figure 5 are SEM images for the Gal2-SWNT sample, for which the specimen was prepared by drop-casting an ultrathin film from the sample solution. Abundantnanotubes were observed, with their being randomly orientedat the edge of the specimen (Figure 5a) but more ordered in thefractured portion of the same specimen (Figure 5b). For Gal4-SWNT, a very dilute aqueous solution was used to prepare theSEM specimen (depositing a few drops and then evaporatingthe water). The image shows generally dispersed nanotubes(Figure 6).

The thermal defunctionalization behavior of the dendron-SWNT samples was very different from those of many otherfunctionalized carbon nanotubes.19,30 The functional groups(sugar dendrons) could not be removed from the nanotubes inthermogravimetric analysis (TGA) scans (under inert atmosphereto 800 °C) and, instead, they were carbonized under the thermaldefunctionalization conditions. For TGA in air, the carbonnanotubes were burned at temperatures similar to those for thesugars. Therefore, the usually effective TGA estimate ofnanotube contents in the functionalized samples was notapplicable, and the compositions in dendron-SWNT sampleswere determined in terms of sugar analyses.

The sugar contents in the dendron-SWNT samples werequantified by spectrophotometry with the anthrone reagent.21

Results thus obtained for the sample compositions are shownin Table 2. There are no meaningful differences in samplecompositions between Gal- and Man-, Gal2- and Man2-, or Gal4-and Man4- as functional moieties, despite the fact that the Gal-based functionalization agents are consistently less effective (alower percentage of starting purified SWNTs solubilized) thantheir Man-based counterparts in the nanotube solubilization(Table 2). The poorer solubilization performance of the Gal-based agents is probably due at least in part to the significantlylower (by a factor of 3 to 4) aqueous solubility of �-D-galactosethan that of R-D-mannose.31

The known sample compositions also allow the calculationfor the average number of nanotube carbons per sugar unit inthe samples (Table 2). From Gal- (or Man-) to Gal2- (orMan2-), the number decreases significantly, namely, that moresugars are displayed on the same amount of SWNTs. For Gal4-SWNT and Man4-SWNT, even though there is no advantagein the number of sugars displayed (because of the relativelylarger amount of tethering moieties for the display), theimproved aqueous solubility and the specific configuration of�-D-galactoses or R-D-mannoses displayed in quartet may offerproperties that are not available to other configurations.

It has been reported that Gal-SWNT binds to E. coli O157:H7 to result in significant cell agglutination.17 The same assaywas used to evaluate the binding of Gal2-SWNT with the sameE. coli strain. As shown in Figure 7, there was obviouslysignificant aggregation of the cells in the presence of Gal2-SWNT. The amount of aggregates (precipitates at the bottom

2416 Biomacromolecules, Vol. 9, No. 9, 2008 Gu et al.

of the centrifuge tube) was larger for a higher starting E. coliconcentration of 108 cells/mL than that of 5 × 107 cells/mL(Figure 7). Between Gal2-SWNT and Gal-SWNT under similarexperimental conditions, the amount of recovered aggregateswas clearly larger in the former than in the latter (Figure 7).This was reflected more quantitatively in results from the CFUreduction assay. As compared in Figure 8, Gal2-SWNT wasobviously more effective in the agglutination of E. coli O157:H7 cells to result in a more significant CFU reduction.Mechanistically, the binding responsible for the cell agglutina-tion is attributed to specific ligand-receptor interactions of thenanotube-displayed �-D-galactoses with E. coli surface galactose-binding-protein.17 The more favorable binding by Gal2-SWNTseems to suggest that the paired �-D-galactoses could be moreeffective in the specific interactions. However, the multivalentbinding of carbohydrates with cell surface receptors is acomplicated phenomenon.1,2,32 A more definitive conclusionrequires more systematic investigations with these and otherassays.

The complexity with desired quantitative evaluations of thesugar-functionalized SWNTs in binding assays is also reflectedin the results on B. subtilis spores. B. subtilis is a commonlyused nonvirulent simulant for Bacillus anthracis (anthrax),33

with surface expressed with various carbohydrates.33 As reportedpreviously for B. anthracis (Sterne) spores,18 Man-SWNT alsobound to B. subtilis spores for their significant aggregation inthe presence of calcium cation (Figure 9). Similarly, Man2-SWNT and Man4-SWNT were both capable of binding andaggregating B. subtilis spores in the same assay, as alsoillustrated in Figure 9. However, a more quantitative comparisonof the different R-D-mannose-functionalized SWNTs sampleswas hindered by other issues beyond simply the differentdisplays of R-D-mannoses in Man2-SWNT and Man4-SWNT.For example, it was found that for Man2-SWNT samples ofsomewhat different Man/nanotube ratios (from slight variationsin reaction conditions in the synthesis) their bindings with thespores (and the associated degrees of aggregation) were obvi-ously different. Such effects were more pronounced in thebinding assay of Man4-SWNT with B. subtilis spores in thepresence of calcium cation. Especially for Man4-SWNT samplesof slightly higher Man/nanotube ratios, the binding assayresulted in well-dispersed aggregates of an irregular rod-likeshape (20-50 µm in length and 5-10 µm in diameter accordingto optical microscopy analyses). We speculate that some of thesecomplications might be due to the divalent cation-mediatedinteractions between the paired R-D-mannoses in Man2-SWNTand Man4-SWNT, which could be competing with the bindingwith cell surface receptors. Further investigations are requiredfor a better understanding of the results.

Summary and Conclusion

A series of dendritic �-D-galactopyranosides and R-D-man-nopyranosides with a terminal amino group were synthesizedand used for the functionalization of SWNTs targeting thedefect-derived carboxylic acid moieties on the nanotube surface.The functionalized nanotube samples were characterized byusing established NMR, optical spectroscopy, and microscopytechniques. The results suggest that the higher-order sugardendrons are more effective in the solubilization of SWNTs,with the corresponding functionalized nanotube samples ofimproved aqueous solubility characteristics, and that the nano-tube is indeed a unique pseudo-one-dimensional scaffold fordisplaying multiple copies of the sugar molecules in pairs or

quartets. These multivalent carbohydrate configurations maypotentially offer interesting chemical and biochemical propertiesand functions, despite some of the complications that are yet tobe understood (as reflected in the results of the binding assaywith B. subtilis spores).

Acknowledgment. Financial support from NSF, NASA,South Carolina Space Grant Consortium, and the Center forAdvanced Engineering Fibers and Films (CAEFF, an NSF-ERCat Clemson University) is gratefully acknowledged. R.A.Q. andS.L. were undergraduate research participants jointly sponsoredby CAEFF and NSF-REU (DMR-0243734).

Supporting Information Available. Synthesis of Gal-N3,Man-N3, and compound 1. This information is available freeof charge via the Internet at http://pubs.acs.org.

References and Notes(1) Mammen, M.; Choi, S. K.; Whitesides, G. M. Angew. Chem., Int. Ed.

1998, 37, 2755–2794.(2) Choi, S. K. Synthetic MultiValent Molecules: Concepts and Biomedical

Applications; John Wiley & Sons, Inc.: Hoboken, NY, 2004.(3) Gestwicki, J. E.; Cairo, C. W.; Strong, L. E.; Oetjen, K. A.; Kiessling,

L. L. J. Am. Chem. Soc. 2002, 124, 14922–14933.(4) (a) Jayaraman, N.; Nepogodiev, S. A.; Stoddart, J. F. Chem.sEur. J.

1997, 3, 1193–1199. (b) Lundquist, J. J.; Toone, E. J. Chem. ReV.2002, 102, 555–578. (c) Cloninger, M. J. Curr. Opin. Chem. Biol.2002, 6, 742–748.

(5) Kiessling, L. L.; Gestwicki, J. E.; Strong, L. E. Curr. Opin. Chem.Biol. 2000, 4, 696–703.

(6) (a) Lin, C.; Yeh, Y.; Yang, C.; Chen, C.; Chen, G.; Chen, C.; Wu, Y.J. Am. Chem. Soc. 2002, 124, 3508–3509. (b) Hernaiz, M. J.; de laFuente, J. M.; Barrientos, A. G.; Penades, S. Angew. Chem., Int. Ed.2002, 41, 1554–1557.

(7) (a) Qu, L.; Luo, P. G.; Taylor, S.; Lin, Y.; Huang, W.; Anyadike, N.;Tzeng, T.-R. J.; Stutzenberger, F.; Latour, R. A.; Sun, Y.-P. J. Nanosci.Nanotechnol. 2005, 5, 319–322. (b) Qu, L.; Gu, L.; Li, H.; Taylor,S.; Elkin, T.; Luo, P. G.; Tzeng, T.-R. J.; Jiang, X.; Latour, R. A.;Stutzenberger, F.; Williams, A.; Sun, Y.-P. J. Biomed. Nanotechnol.2005, 1, 61–67.

(8) Molugu, S.; Qu, L.; Lin, Y.; Sun, Y.-P.; Tzeng, T.-R. J.; Stutzenberger,F.; Latour, R. A. J. Biomed. Nanotechnol. 2006, 2, 1–10.

(9) Ashton, P. R.; Hounsell, E. F.; Jayaraman, N.; Nilsen, T. M.; Spencer,N.; Stoddart, J. F.; Young, M. J. Org. Chem. 1998, 63, 3429–3437.

(10) Disney, M. D.; Zheng, J.; Swager, T. M.; Seeberger, P. H. J. Am.Chem. Soc. 2004, 126, 13343–13346.

(11) (a) Lin, Y.; Taylor, S.; Li, H.; Fernando, K. A. S.; Qu, L.; Wang, W.;Gu, L.; Zhou, B.; Sun, Y.-P. J. Mater. Chem. 2004, 14, 527–541. (b)Lin, Y.; Li, H.; Gu, L.; Luo, P. G.; Veca, L. M.; Wang, H.; Sun,Y.-P. Bioapplications of Carbon Nanotubes in Chemistry of CarbonNanotubes; Basiuk, V. A., Basiuk, E. V., Eds.; American ScientificPublishers: Valencia, CA, 2008, in press.

(12) (a) Pompeo, F.; Resasco, D. E. Nano Lett. 2002, 2, 369–373. (b)Matsuura, K.; Hayashi, K.; Kimizuka, N. Chem. Lett. 2003, 32, 212–213. (c) Hasegawa, T.; Fujisawa, T.; Numata, M.; Umeda, M.;Matsumoto, T.; Kimura, T.; Okumura, S.; Sakurai, K.; Shinkai, S.Chem. Commun. 2004, 19, 2150–2151.

(13) (a) Lu, Q.; Moore, J. M.; Huang, G.; Mount, A. S.; Rao, A. M.;Larcom, L. L.; Ke, P. C. Nano Lett. 2004, 4, 2473–2477. (b) Kam,N. W. S.; O’Connell, M.; Wisdom, J. A.; Dai, H. J. Proc. Natl. Acad.Sci. U.S.A. 2005, 112, 11600–11605. (c) Pantarotto, D.; Singh, R.;McCarthy, D.; Erhardt, M.; Briand, J.-P.; Prato, M.; Kostarelos, K.;Bianco, A. Angew. Chem., Int. Ed. 2004, 43, 5242–5246.

(14) (a) Huang, W.; Taylor, S.; Fu, K.; Lin, Y.; Zhang, D. H.; Hanks, T. W.;Rao, A. M.; Sun, Y.-P. Nano Lett. 2002, 2, 311–314. (b) Kam,N. W. S.; Dai, H. J. J. Am. Chem. Soc. 2005, 127, 6021–6026.

(15) (a) Pantarotto, D.; Partidos, C. D.; Graff, R.; Hoebeke, J.; Briand,J. P.; Prato, M.; Bianco, A. J. Am. Chem. Soc. 2003, 125, 6160–6164.(b) Pantarotto, D.; Partidos, C. D.; Hoebeke, J.; Brown, F.; Kramer,E.; Briand, J. P.; Muller, S.; Prato, M.; Bianco, A. Chem. Biol. 2003,10, 961–966.

(16) Elkin, T.; Jiang, X.; Taylor, S.; Lin, Y.; Gu, L.; Yang, H.; Brown, J.;Collins, S.; Sun, Y.-P. ChemBioChem 2005, 6, 640–643.

(17) Gu, L.; Elkin, T.; Jiang, X.; Li, H.; Lin, Y.; Qu, L.; Tzeng, T.-R. J.;Joseph, R.; Sun, Y.-P. Chem. Commun. 2005, 7, 874–876.

Scaffold for the Multivalent Display of Sugars Biomacromolecules, Vol. 9, No. 9, 2008 2417

(18) Wang, H.; Gu, L.; Lin, Y.; Lu, F.; Meziani, M. J.; Luo, P. J.; Wang,W.; Cao, L.; Sun, Y.-P. J. Am. Chem. Soc. 2006, 128, 13364–13365.

(19) (a) Huang, W.; Lin, Y.; Taylor, S.; Gaillard, J.; Rao, A. M.; Sun,Y.-P. Nano Lett. 2002, 2, 231–234. (b) Sun, Y.-P.; Fu, K.; Lin, Y.;Huang, W. Acc. Chem. Res. 2002, 35, 1096–1104.

(20) Gu, L.; Lin, Y.; Qu, L.; Sun, Y.-P. Biomacromolecules 2006, 7, 400–402.

(21) Jermyn, M. A. Anal. Biochem. 1975, 68, 332–335.(22) Chernyak, A. Y.; Sharma, G. V. M.; Kononov, L. O.; Krishna, P. R.;

Levinsky, A. B.; Kochetkov, N. K.; Rao, A. V. R. Carbohydr. Res.1992, 223, 303–309.

(23) Vutukuri, D. R.; Basu, S.; Thayumanavan, S. J. Am. Chem. Soc. 2004,126, 15636–15637.

(24) Luo, P. G. Carbohydrate-Biofunctionalized Nanomaterials as Anti-adhesion and Detection Agents for Bacterial Pathogens, Ph.D. Dis-sertation, Clemson University, Clemson, S.C., 2006, pp 79-105.

(25) It has also been suggested recently that as a result of the oxidativeacid treatment carbonaceous fragments with COOH groups are createdand adsorbed onto the SWNTs, see. (a) Salzmann, C. G.; Llewellyn,S. A.; Tobias, G.; Ward, M. A. H.; Huh, Y.; Green, M. L. H. AdV.Mater. 2007, 19, 883–887. (b) Yu, H.; Jin, Y.; Peng, F.; Wang, H.;Yang, J. J. Phys. Chem. C 2008, 112, 6758–6763.

(26) (a) Qu, L.; Lin, Y.; Hill, D. E.; Zhou, B.; Wang, W.; Sun, X.;Kitaygorodskiy, A.; Suarez, M.; Connell, J. W.; Allard, L. F.; Sun,Y. P. Macromolecules 2004, 37, 6055–6060. (b) Qu, L.; Veca, L. M.;Lin, Y.; Kitaygorodskiy, A.; Chen, B.; McCall, A. M.; Connell, J. W.;Sun, Y.-P. Macromolecules 2005, 38, 10328–10331.

(27) (a) Chen, J.; Liu, H.; Weimer, W. A.; Halls, M. D.; Waldeck, D. H.;Walker, G. C. J. Am. Chem. Soc. 2002, 124, 9034–9035. (b) Holzinger,M.; Abraha, J.; Whelan, P.; Graupner, R.; Ley, L.; Hennrich, F.;Kappes, M.; Hirsch, A. J. Am. Chem. Soc. 2003, 125, 8566–8580.

(28) Zhou, B.; Lin, Y.; Li, H.; Huang, W.; Connell, J. W.; Allard, L. F.;Sun, Y.-P. J. Phys. Chem. B 2003, 107, 13588–13592.

(29) Hill, D.; Lin, Y.; Qu, L.; Kitaygorodskiy, A.; Connell, J. W.; Allard,L. F.; Sun, Y.-P. Macromolecules 2005, 38, 7670–7675.

(30) Lin, Y.; Taylor, S.; Li, H.; Fernando, K. A. S.; Qu, L.; Wang, W.;Gu, L.; Zhou, B.; Sun, Y.-P. J. Mater. Chem. 2004, 14, 527–541.

(31) According to the MSDS data sheets (www.fishersci.com).(32) Krishnamurthy, V. M.; Estroff, L. A.; Whitesides,G. M. In Fragment-

based Approaches in Drug DiscoVery; Jahnke, W., Erlanson, D. A.,Eds.; Wiley-VCH Verlag GmbH & Co. KGaA: New York, 2006.

(33) Fox, A.; Stewart, G. C.; Waller, L. N.; Fox, K. F.; Harley, W. M.;Price, R. L. J. Microbiol. Methods 2003, 54, 143–152.

BM800395E

2418 Biomacromolecules, Vol. 9, No. 9, 2008 Gu et al.