doi.org/10.26434/chemrxiv.12573701.v1

Probing Multivalent Lectin-Carbohydrate Binding via MultifunctionalGlycan-Gold Nanoparticles: Implications for Blocking Virus InfectionDarshita Budhadev, Emma Poole, Inga Nehlmeier, Yuanyuan Liu, James Hooper, Elizabeth Kalverda,Uchangi Satyaprasad Akshath, Nicole Hondow, W. Bruce Turnbull, Stefan Pöhlmann, Yuan Guo, Dejian Zhou

Submitted date: 26/06/2020 • Posted date: 29/06/2020Licence: CC BY-NC-ND 4.0Citation information: Budhadev, Darshita; Poole, Emma; Nehlmeier, Inga; Liu, Yuanyuan; Hooper, James;Kalverda, Elizabeth; et al. (2020): Probing Multivalent Lectin-Carbohydrate Binding via MultifunctionalGlycan-Gold Nanoparticles: Implications for Blocking Virus Infection. ChemRxiv. Preprint.https://doi.org/10.26434/chemrxiv.12573701.v1

Multivalent lectin-glycan interactions are widespread in biology and are often exploited by pathogens to bindand infect host cells. Glycoconjugates can block such interactions and thereby prevent infection. Theinhibition potency strongly depends on matching the spatial arrangement between the multivalent bindingpartners. However, the structural details of some key lectins remain unknown and different lectins may exhibitoverlapping glycan specificity. This makes it difficult to design a glycoconjugate that can potently andspecifically target a particular multimeric lectin for therapeutic interventions, especially under the challengingin vivo conditions. Conventional techniques such as surface plasmon resonance (SPR) and isothermaltitration calorimetry (ITC) can provide quantitative binding thermodynamics and kinetics. However, theycannot reveal key structural information, e.g. lectin’s binding site orientation, binding mode, and inter-bindingsite spacing, which are critical to design specific multivalent inhibitors. Herein we report that goldnanoparticles (GNPs) displaying a dense layer of simple glycans are powerful mechanistic probes formultivalent lectin-glycan interactions. They can not only quantify the GNP-glycan-lectin binding affinities via anew fluorescence quenching method, but also reveal drastically different affinity enhancing mechanismsbetween two closely-related tetrameric lectins, DC-SIGN (simultaneous binding to one GNP) and DC-SIGNR(inter-crosslinking with multiple GNPs), via a combined hydrodynamic size and electron microscopy analysis.Moreover, a new term, potential of assembly formation (PAF) has been proposed to successfully predict theassembly outcomes based on the binding mode between GNP-glycans and lectins. Finally, the GNP-glycanscan potently and completely inhibit DC-SIGN-mediated augmentation of Ebola virus glycoprotein-driven cellentry (with IC50 values down to 95 pM), but only partially block DC-SIGNR-mediated virus infection. Ourresults suggest that the ability of a glycoconjugate to simultaneously block all binding sites of a target lectin iskey to robust inhibition of viral infection.

File list (2)

download fileview on ChemRxivGNP-glycan manu.pdf (2.37 MiB)

download fileview on ChemRxivSupporting Information.pdf (2.29 MiB)

Probing Multivalent Lectin-Carbohydrate Binding via MultifunctionalGlycan-Gold Nanoparticles: Implications for Blocking Virus Infection

Darshita Budhadev†,╬, Emma Poole†,╬, Inga Nehlmeier‡, Yuanyuan Liu†,§, James Hooperɧ, , ElizabethKalverda,† Uchangi Satyaprasad Akshath,† Nicole Hondowɸ, W. Bruce Turnbull†, Stefan Pöhlmann‡,*,Yuan Guoɧ,*, and Dejian Zhou†,*

† School of Chemistry and Astbury Centre for Structural Molecular Biology, University of Leeds, Leeds LS2 9JT, UnitedKingdom.

‡ Infection Biology Unit, German Primate Center – Leibniz Institute for Primate Research, 37077 Göttingen, and Faculty ofBiology and Psychology, University of Göttingen, 37073 Göttingen, Germany.

§ School of Pharmaceutical and Chemical Engineering, Chengxian College, Southeast University, Nanjing 210088, China.ɧ School of Food Science & Nutrition and Astbury Centre for Structural Molecular Biology, University of Leeds, Leeds LS2

9JT, United Kingdom.ɸ School of Chemical and Process Engineering, University of Leeds, Leeds LS2 9JT, United Kingdom.╬ These authors contributed equally to this work.

KEYWORDS. Gold nanoparticle, glycoconjugate, multivalent lectin-glycan interaction, fluorescence quenching, bindingaffinity, multi-modal readout, virus inhibition.

ABSTRACT: Multivalent lectin-glycan interactions are widespread in biology and are often exploited by pathogens to bind andinfect host cells. Glycoconjugates can block such interactions and thereby prevent infection. The inhibition potency strongly dependson matching the spatial arrangement between the multivalent binding partners. However, the structural details of some key lectinsremain unknown and different lectins may exhibit overlapping glycan specificity. This makes it difficult to design a glycoconjugatethat can potently and specifically target a particular multimeric lectin for therapeutic interventions, especially under the challengingin vivo conditions. Conventional techniques such as surface plasmon resonance (SPR) and isothermal titration calorimetry (ITC) canprovide quantitative binding thermodynamics and kinetics. However, they cannot reveal key structural information, e.g. lectin’s bind-ing site orientation, binding mode, and inter-binding site spacing, which are critical to design specific multivalent inhibitors. Hereinwe report that gold nanoparticles (GNPs) displaying a dense layer of simple glycans are powerful mechanistic probes for multivalentlectin-glycan interactions. They can not only quantify the GNP-glycan-lectin binding affinities via a new fluorescence quenchingmethod, but also reveal drastically different affinity enhancing mechanisms between two closely-related tetrameric lectins, DC-SIGN(simultaneous binding to one GNP) and DC-SIGNR (inter-crosslinking with multiple GNPs), via a combined hydrodynamic size andelectron microscopy analysis. Moreover, a new term, potential of assembly formation (PAF) has been proposed to successfully predictthe assembly outcomes based on the binding mode between GNP-glycans and lectins. Finally, the GNP-glycans can potently andcompletely inhibit DC-SIGN-mediated augmentation of Ebola virus glycoprotein-driven cell entry (with IC50 values down to 95 pM),but only partially block DC-SIGNR-mediated virus infection. Our results suggest that the ability of a glycoconjugate to simultane-ously block all binding sites of a target lectin is key to robust inhibition of viral infection.

Introduction

Multivalent lectin-glycan interactions are widespread and me-diate many important biological functions which include cell-cell communication, pathogen-host cell recognition, attachmentand infection, and modulation of immune responses.1-9 As mostmonovalent lectin-glycan binding events are often too weak tobe bio-functional, many lectins form multimeric structures tocluster their carbohydrate-binding-domains (CRDs) for effi-cient binding with spatially matched multivalent glycans to en-hance binding affinity and specificity.10-14 The overall multiva-lent affinity is not only directly linked to the monovalent affin-ity, but also the glycan valency and the mode of binding. In gen-eral, a pair of spatially matched multivalent binding partnerscan bind simultaneously with each other and form a single en-tity. This gives the highest affinity enhancement and selectivity

due to the most favorable enthalpy and entropy terms.13 Whilethose without such spatial and orientation match may inter-crosslink with each other to maximize binding enthalpy andform large scale assemblies, but this typically gives lower affin-ity enhancement and binding specificity due to an unfavorableentropy penalty.15 Therefore, understanding the structuralmechanism underlying the affinity enhancement in multivalentprotein-ligand binding is key to the design of potent, specificmultivalent inhibitors against a target multivalent receptor. No-tably, the development of potent glycoconjugates to block virusbinding and infection of host cells can prevent virus mutationand developing resistance, a unique advantage over other anti-viral strategies.9

A number of different scaffolds, including DNAs, proteins,polymers, dendrimers, C60 derivatives, vesicles, and inorganic

nanoparticles, have been employed to construct glycoconju-gates14, 16-22 to study multivalent lectin-glycan recognition anddevelop effective interventions against certain diseases. Someof these have exhibited excellent potency in inhibiting pathogeninfection.9, 11, 13-14, 16-20, 23-24 Most of the bindings have been eval-uated by conventional biophysical techniques such as surfaceplasmon resonance (SPR),14, 22 and/or isothermal titration calo-rimetry (ITC).14, 25-27 Despite their wide use and good capacityin providing quantitative binding affinity, thermodynamic andkinetic data, these techniques cannot reveal key structural infor-mation of the target lectin, e.g. binding site orientation, inter-binding site distance, and binding mode which are critical todesign potent, specific multivalent glycan inhibitors. Mean-while, the unique size-dependent physical properties of nano-materials have been harnessed to study lectin-glycan interac-tions. A good example here are gold nanoparticles (GNPs)28

whose strong, size- and aggregation-state dependent absorp-tion,29-30 powerful signal amplification in SPR,31 and surface en-hanced Raman scattering32-33 have been widely exploited for bi-osensing and diagnostic applications. Despite these successes,glycan conjugated GNPs (GNP-glycans) have not been ex-ploited as new structural and mechanistic probes for multivalentlectin-glycan interactions.

Recently, we have shown that CdSe/ZnS quantum dots(QDs) displaying a dense layer of mannose glycans are power-ful structural probes for multivalent lectin-glycan binding.34-36

The QD probes can not only quantify the binding affinity via aratiometric FRET readout but also dissect the different bindingmodes between a pair of closely-related, almost identical te-trameric lectins, DC-SIGN37-38 and DC-SIGNR39 (collectivelyabbreviated as DC-SIGN/R hereafter). DC-SIGN/R are im-portant lectin receptors which play a key role in facilitating theHIV, Ebola virus and Zika virus infections.34-35, 40-41 Moreover,DC-SIGN is also key to immune regulation,1, 5, 42-43 making it anhighly attractive target for developing immunotherapies againstimportant human diseases such as cancer, allergy, and autoim-mune diseases.1, 42, 44 However, its tetrameric structure remainsunknown, making it difficult to develop novel glycoconjugatesthat can potently and specifically target DC-SIGN for therapeu-tic interventions, especially when the overlapping glycan spec-ificity of various lectins is considered.1 Using the QD-glycans,we have found that DC-SIGN’s 4 CRDs bind simultaneously toone QD and give an impressive affinity enhancement factor ()of up to 1.5 million fold over the corresponding monovalentbinding. The QD-glycans also potently inhibit DC-SIGN-mediated augmentation of Ebola virus entry into host cells withsub-nM IC50 values.35 Despite such success, the cytotoxicityand long term toxicity of the CdSe QD scaffold has significantlylimited its potential use as DC-SIGN targeting therapeuticagents, especially under in vivo conditions.

To address this issue, here we have constructed polyvalentglycan-nanoparticle probes on a 5 nm (diameter) GNP scaffoldto take the advantages of GNPs’ excellent biocompatibility,low-/non-toxicity, and robust gold-thiol surface chemistry foreasy control of the glycan density and surface presentation.28

Moreover, after coating with lipoic acid (LA) based glycan lig-ands containing a terminal mannose (Man) or mannose--1,2-mannose (DiMan), the GNP-glycans are of the similarsize and mannose glycan coating to gp120 trimer,45-46 the HIVsurface densely glycosylated glycoprotein which is responsiblefor initiating HIV-DC-SIGN interaction to facilitate HIV infec-tion. Thus, the GNP-glycans are good mimics of gp120 forprobing its interaction with DC-SIGN. Using DC-SIGN/R as

model lectins, we show that the GNP’s strong fluorescentquenching property47 can be exploited as a reliable method toquantify multivalent GNP-glycan-lectin binding for the 1st time.Moreover, its nanoscale size, and high contrast under electronmicroscopy imaging have been integrated into a multimodalreadout to reveal the different binding modes and affinity en-hancement mechanisms for DC-SIGN/R (e.g. simultaneousbinding for DC-SIGN, but inter-crosslinking for DC-SIGNR).Furthermore, the GNP-lectin binding modes are found to be di-rectly linked to the GNP’s ability to block lectin mediated virusinfection of host cells: only a GNP-glycan which binds simul-taneously to all binding sites of the target lectin can potentlyand completely block virus infection, but not that showing across-linking binding mode.

Results and Discussion

Ligand design and synthesis. The schematic structures of theGNP-glycan conjugates and the chemical structures of the gly-can ligands used in this study are shown in Figure 1.

Figure 1. Schematic of the GNP-glycans used in this study. TheGNP is coated with DHLA-based glycan ligand containing eitherone (A) or three (B) terminal -1-mannose or -1-mannose--1,2-mannose to tune its surface glycan valency. The chemical struc-tures of glycan ligands are depicted underneath.

Each ligand contains three unique functional domains, a di-hydrolipoic acid (DHLA) for strong binding to the GNP surfacevia the formation of 2 strong Au-S bonds to impose excellentstability;48 a short, flexible oligo-(ethylene glycol) linker to en-hance water solubility, and to resist non-specific adsorption,35,

49 and a terminal an -1-mannose or -1-mannose-α-1,2-man-nose (abbreviated as DHLA-Man or DHLA-DiMan, respec-tively, hereafter) for specific binding to DC-SIGN/R.34-35 Wehave found previously that DC-SIGN binds more efficiently toQDs capped with higher mannose densities.34 Moreover, gly-coconjugates with higher glycan valency have often shown toexhibit better virus inhibition potencies.9, 14, 16, 21, 50 Thus we areinterested to study whether GNP-glycans’ DC-SIGN/R bindingaffinity and virus inhibition potency can be further enhanced byincreasing its surface glycan density. Thus, a branched versionof the DHLA-glycan ligand, each containing three terminal gly-cans, are also synthesized (denoted as DHLA-(Man)3 or DHLA-(DiMan)3, respectively).50 The branched ligands have the sameDHLA anchoring group for GNP binding as the monomericglycans. Thus, a similar number of ligands are expected to coat

each GNP, allowing us to prepare more densely glycosylatedGNPs as shown schematically in Figure 1.

Scheme 1. Synthetic route to LA-Man-/DiMan ligands

Scheme 2. Synthetic route to LA-(Man)3-/(DiMan)3 ligands

The synthetic routes to the monomeric LA-glycan and branchedtrimeric LA-(glycan)3 ligands are shown in Schemes 1 and 2,respectively. First, the LA based linker molecules each contain-ing one or three terminal acetylene groups were synthesized ingood yields using the standard BOC protection and amide cou-pling chemistries. Details of the synthetic procedures and spec-troscopic data of the intermediate compounds were provided inthe Supporting Information (SI). Second, -mannose and -mannose--1,2-mannose appending a hydrophilic di(ethyleneglycol)-azide linker (N3-EG2-Man-/DiMan) to provide someflexibility to the terminal glycans were synthesized using ourestablished procedures.34-35 Third, the LA-acetylene linkerswere coupled efficiently to N3-EG2-Man/DiMan via the Cu-cat-alyzed click chemistry16 to give the desired LA-Man/DiMan orLA-(Man)3/(DiMan)3 ligands (see Experimental Section for thegeneral procedures and spectroscopic data, and SI for their 1H

and 13C NMR spectra). Finally, the LA-glycan ligands were re-duced quantitatively to their corresponding DHLA forms bytris(2-carboxyethyl)phosphine hydrochloride (TCEP.HCl)51-52

before they were used to prepare the GNP-glycans.

GNP-glycan preparation. GNP-glycans (GNP capped withthe DHLA-glycan ligands) were prepared by incubation of a 5nm citrate stabilized GNP with the above DHLA-glycan ligandsin water at a ligand:GNP molar ratio of 1000:1 for 24 h. Anyfree unbound ligands were removed by ultra-filtration using30K MWCO filter tubes and washing with pure water. GNP-glycans prepared using LA-glycans or their reduced DHLA-glycan forms were found to be identical in terms of hydrody-namic sizes and stability. Thus the air-stable LA-glycan ligandscould be directly used to make the GNP-glycans, eliminatingthe need of reduction and handling air-sensitive DHLA-glycanligands that were essential in QD-glycan preparation via cap-exchange.34, 40, 48 This made the GNP-glycan preparation simpleand straightforward. This result was fully consistent with earlierliterature reports that dialkyldisulfides were cleaved upon bind-ing to gold surfaces, forming identical self-assembled monolay-ers to their alkylthiol counterparts.53 The GNP-glycans werehighly stable, no changes of physical appearance or precipita-tion were observed after prolonged storage (>6 months) in afridge. They were uniform and monodisperse in both water andin a binding buffer (20 mM HEPES, 100 mM NaCl, 10 mMCaCl2, pH7.8) with hydrodynamic diameters (Dhs) of ~11 nm(see SI, Figure S1 and Table S1). Such Dh values matched wellto the size of gp120 trimer (~12 nm).45 Moreover, the GNPswere densely coated with mannose containing glycans similarto those found on gp120 surface. Using the ligand amount dif-ference between that used and that remained in the supernatantafter conjugation (measured by a phenol-sulfuric acid assay),35

the numbers of ligands bound on each GNP were estimated as~490, ~690, 720, and ~650 for LA-Man, LA-DiMan, LA-(Man)3 and LA-(DiMan)3, respectively. Using the Dh values andthe method reported by the Mirkin group,54 the average inter-glycan distances were estimated as ~0.97, ~0.83, ~0.46, and~0.49 nm for GNP-Man, GNP-DiMan, GNP-(Man)3 and GNP-(DiMan)3, respectively (SI, Table S2). The average inter-glycandistances of GNP-Man/DiMan fell well within the majority ofinter-glycan sequon distances (e.g. 0.7-1.3 nm) found on gp120,46 but those of GNP-(Man/Diman)3 were considerably smaller.These results suggested that the GNP-Man/DiMan could act asgood mimics of gp120 for probing its DC-SIGN interaction.

Quantifying GNP-glycan-DC-SIGN/R binding affinity. Be-sides strong plasmonic absorption, GNPs are well-known fortheir strong and universal fluorescence quenching property.47

Moreover, its quenching has shown to follow the nano-surfaceenergy transfer (NSET),55-56 rather than Förster resonance en-ergy transfer (FRET) mechanism. As a result, its quenching ismore superior and covers a longer distance range than most or-ganic quenchers relying on the FRET mechanism.47, 55 Indeed,the GNP’s superior quenching ability has been widely exploitedfor biosensing applications,57-58 but not as a readout method forlectin-glycan binding quantification. Here the GNP’s outstand-ing fluorescence quenching was exploited as a new readout formultivalent lectin-glycan binding quantification for the firsttime (Figure 2A). To do this, we first introduced a cysteine site-specifically on the CRD (outside its glycan binding pocket sonot affecting its glycan binding properties) of DC-SIGN/R ex-tracellular segments by mutagenesis. The mutant proteins haveshown to retain faithfully the tetrameric structure and glycanbinding properties of wildtype proteins.34-35 We then labeled the

proteins using a maleimide modified Atto-594 and confirmedthe labeling by high resolution mass spectroscopy (SI, FigureS2). We then recorded the fluorescence spectra of labeled DC-SIGN/R without and with each GNP-glycan under a fixed pro-tein:GNP molar ratio of 1:1 over a concentration range of 0-64nM. All binding studies were carried out in a binding buffercontaining a large excess of bovine serum albumin (BSA, 1mg/mL final concentration). The inclusion of large excesses ofBSA, a non-target serum protein of high abundance in vivo, inbinding studies made the conditions resemble more closely toreal biological situations. Moreover, this also greatly reducednon-specific interactions and adsorption of proteins and GNPsto surfaces (a main source of experimental errors at low con-centrations, e.g. < 10 nM).59

The typical fluorescence spectra showing the binding be-tween DC-SIGN and GNP-DiMan were given in Figure 2B. Inthe absence of GNP-DiMan, DC-SIGN’s fluorescence intensitywas found to increase linearly with concentration. While in thepresence of GNP-DiMan, its fluorescence was greatly reducedand increasingly deviated from the linear relationship with theincreasing concentration (SI, Figure S3). This result was fullyconsistent with that, with the increasing concentration, an in-creasing proportion of DC-SIGN were bound to and quenchedby GNP-DiMan in close proximity. The specificity of DC-SIGN-GNP-DiMan binding was verified by two controls: (1)mixing DC-SIGN with a GNP capped with a control DHLA-EG2-OH ligand lacking the terminal glycan produced negligiblequenching; (2) free mannose could effectively compete withGNP-DiMan binding to DC-SIGN, leading to a significant,dose-dependent fluorescence recovery (SI, Figure S4).

Figure 2. (A) Schematic showing the quenching of DC-SIGN/R(dye-labeled) fluorescence upon binding to a GNP-glycan, whichacts as a readout for binding quantification. (B) Typical fluores-cence spectra of varied concentrations of DC-SIGN without (solidlines) and with (broken lines) 1 molar equivalent of GNP-DiMan(EX = 590 nm). (C, D) The quenching efficiency (QE%)-concentration relationship for DC-SIGN (C) or DC-SIGNR (D)binding to the various GNP-glycans fitted by the Hill’s equation.

To analyze the binding data quantitatively, we first calculatedthe quenching efficiency (QE) at each concentration (C) via:

QE% = (IF0 - IF)/IF0 × 100% (1)

Where IF0 and IF are the integrated fluorescence of the proteinin the absence and presence of each GNP-glycan, respectively.Given that the fluorescence of DC-SIGN/R increases linearlywith C in the absence of GNP-DiMan (SI, Figure S3), and GNPscan completely quench various fluorophores at close proximity(e.g. > 99.9%),47 the QE% thus faithfully represents the percent-age of proteins bound to the GNP-glycan. Thus, the apparentbinding dissociation constant (Kd) can be derived from the QE-C relationship by fitting with Hill’s equation:

QE = QEmax × Cn/(Kdn + Cn) (2)

Where QEmax, Kd, C, and n are the maximum QE (fixed at 100),apparent binding dissociation constant, protein concentration,and Hill coefficient, respectively.

Table 1. Summary of the fitting parameters of DC-SIGN/R bindingto various GNP-glycans derived from Figures 2C/D.

GNP-glycan + Protein Kd (nM) n R2

GNP-Man + DC-SIGN 33.1 ± 2.1 0.82±0.07 0.997

GNP-(Man)3 + DC-SIGN 18.7 ± 0.3 0.77±0.01 0.999

GNP-DiMan + DC-SIGN 3.9 ± 0.3 0.59±0.03 0.999

GNP-(DiMan)3 + DC-SIGN 3.6 ± 0.1 0.42±0.01 1.00

GNP-Man + DC-SIGNR 214 ± 68 0.68±0.10 0.994

GNP-(Man)3 + DC-SIGNR 133 ± 20 0.63±0.07 0.996

GNP-DiMan + DC-SIGNR 152 ± 37 0.99±0.14 0.954

GNP-(DiMan)3 + DC-SIGNR 42 ± 2 0.48 ± 0.03 0.999

Figures 2B and 2C showed the best fits of the QE-C relation-ships for DC-SIGN/R binding to each GNP-glycan. The result-ing fitting parameters were summarized in Table 1. Four keypoints could be drawn from the binding data given in Table 1.(1) DC-SIGN bound more strongly to GNP-glycans than DC-SIGNR did. This effect was particularly pronounced for GNP-DiMan where the binding affinity difference was as high as ~40fold. As GNP-DiMan here presented a good mimic of gp120,its stronger binding affinity with DC-SIGN over DC-SIGNRcould thus help explain why DC-SIGN was found more effec-tive than DC-SIGNR in transmitting infections for some HIVstrains.60 This result also agreed well with those reported previ-ously with QD-glycans where the binding Kds were measuredratiometrically via FRET.34 Given that the CRDs in DC-SIGN/R had the same mannose binding motifs,61 the differencehere indicated that DC-SIGN/R might adopt different modes inbinding to GNP-glycans, similar to those observed with QD-glycans.35 (2) Further increasing the GNP surface glycan den-sity via capping with the dendritic LA-(glycan)3 ligands signif-icantly improved the binding affinity with DC-SIGNR (e.g. Kd

~42 v.s. ~152 nM, or 3.6 fold enhancement for GNP-(DiMan)3

v.s. GNP-DiMan), but not much with DC-SIGN, suggesting thatDC-SIGNR may prefer binding to GNPs with even higher gly-can densities. (3) Low nM Kd (e.g. 3.9 nM) for DC-SIGN bind-ing with GNP-DiMan was obtained, which was ~250,000 foldtighter than the corresponding monomeric binding between Di-Man and DC-SIGN CRD (Kd = 0.9 mM),25 suggesting that apolyvalent display of the glycans on the GNP greatly enhancedits multivalent binding affinity with DC-SIGN. These resultswere fully consistent with those reported for other multivalentlectin-glycoconjuagte interactions.10, 14, 18, 22 (4) A higher degreeof affinity enhancement was observed for DC-SIGN binding toGNP-DiMan over GNP-Man, possibility due to the former be-ing able to exploit more efficiently the CRD’s secondary glycan

bind sites than the later.61 Overall, these results agreed well withthose reported previously with the QD-glycans.35

Probing DC-SIGN/R-GNP-glycan binding mode and affin-ity enhancing mechanism. First we monitored the Dhs of GNP-DiMan before and after binding to wildtype DC-SIGN/R (unla-beled) by dynamic light scattering (DLS).62 GNP-DiMan dis-played a single Dh species of ~11 nm with a narrow distribution(full width at half maximum, FWHM = 3.5 nm, Figure 3A1) inthe binding buffer. After mixing with DC-SIGN (protein: GNP

molar ratio = 20:1), the size of the dominate species was signif-icantly increased (Dh = ~41 nm) and broadened (FWHM ~28nm, Figure 3A2). This result was consistent with that expectedfor each GNP-glycan being bound by a limited number of DC-SIGN molecules, forming a protein shell on the GNP. In con-trast, mixing DC-SIGNR with GNP-DiMan produced a domi-nant species whose Dh was gradually increased from ca. 300 to>2000 nm over a period of 4 h (Figure 3A3). Such sizes wereclearly too big to be individual GNP-DC-SIGNR assemblies, astrong indication of macroscopic assembly, arising presumablyfrom DC-SIGNR and GNP-DiMan inter-crosslinking.

Figure 3. Typical Dh distribution histograms of GNP-DiMan only (10 nM, A1); GNP-DiMan (10 nM) + wildtype DC-SIGN (200 nM, A2)or GNP-DiMan (16 nM) + wildtype DC-SIGNR (640 nM, A3). TEM images of cryo-prepared samples of GNP-DiMan only (40 nM, B1),GNP-DiMan (40 nM) + DC-SIGN (1.5 M, B2) or GNP-DiMan (40 nM) + DC-SIGNR (1.5 M, B3), and their corresponding GNP nearestneighbor distance (NND) histograms (C1, C2, C3). The red curves show the Guanssian fits of the histograms.

This assumption was verified by transmission electron micros-copy (TEM) imaging of cryo-prepared GNP-DiMan-proteinsamples via rapid plunge freezing followed by vacuum dryingwhich we had shown previously to preserve the native disper-sion states of nanoparticle solutions.63 In the absence of DC-SIGN/R, GNP-DiMan appeared as isolated single particles(Figure 3B1) and gave a randomly distributed nearest neighbordistance (NND) over a range of 10 to 100 nm (Figure 3C1).Binding of DC-SIGN with GNP-DiMan also gave isolated sin-gle particles (Figure 3B2), although the resulting NND distri-bution histogram could be fitted well (R2 = 0.982) by a single

Gaussian distribution to yield an average NND of ~44 nm (Fig-ure 3C2). Interestingly, this NND value matched well to the Dh

size of GNP-DiMan-DC-SIGN assembly (~41 nm), suggestingthat the formation of a layer of DC-SIGN on the GNP preventedindividual GNPs from coming close to each other. In contrast,binding of DC-SIGNR with GNP-DiMan produced large scale,closely-packed GNP assemblies of hundreds of nm cross (Fig-ure 3B2) with a rather small NND of ~7.9 nm (Figure 3C3).This result agreed well with the DLS measurement, confirmingthat binding of DC-SIGNR led to macroscopic assemblies ofGNPs via DC-SIGNR-GNP-DiMan intercross linking. These

results revealed that the binding modes of DC-SIGN/R to GNP-DiMan were clearly different: DC-SIGN must have bound sim-ultaneously to one GNP-DiMan via all of its four binding sitesand formed a layer of proteins surrounding each GNP, givingrise to isolated single particles. In contrast, DC-SIGNR and pol-yvalent GNP-DiMan inter-crosslinked each other, and formedlarge scale but closely-packed GNP-DC-SIGNR assemblies.Therefore, by harnessing GNPs’ nanoscale size and high con-trast under TEM imaging, we have developed a new multimodalreadout which successfully dissected the distinct modes ofGNP-DiMan in binding to DC-SIGN/R. This result agreed fullywith those reported previously with QD-glycans.35

The distinct binding modes exhibited by DC-SIGN/R could alsohelp explain their different affinities with GNP-glycans. ForDC-SIGN, its simultaneous binding with one GNP via all 4 ofits CRDs should yield strong affinity enhancement from boththe favorable enthalpy and entropy terms. For DC-SIGNR,whilst its inter-crosslinking mode of binding with GNP-glycanscould maximize the binding enthalpy, the resulting large scaleassemblies would incur significant entropy penalty. As a result,DC-SIGN should bind more strongly to GNP-glycans than DC-SIGNR did, matching well to the apparent Kd data measured byfluorescence quenching (Table 1).

Correlation between lectin-GNP-glycan affinity and assem-bly. The credibility of this GNP quenching based affinity meas-urement method was further verified by the DLS data. For ex-ample, binding of an increasing amount of DC-SIGN to a fixedamount of GNP-DiMan (10 nM) gave a more rapid increase ofDh than that to GNP-Man (Figure 4A), consistent with the for-mer being a stronger binding partner than the latter (apparentKd: 3.9 v.s. 33 nM). Moreover, mixing an increasing amount ofDC-SIGNR with GNP-(Man)3 yielded no noticeable Dh

changes throughout the concentration range studied (20-200nM), indicating no measurable binding (Figure 4B). Whilemixing DC-SIGNR with GNP-(DiMan)3 also produced no ap-parent Dh changes initially (≤100 nM); as DC-SIGNR concen-tration was increased further, a dramatic increase of Dh was ob-served, particularly at 200 nM, a clear indication of the for-mation of large scale GNP-DC-SIGNR assemblies (see Figure4B, and SI, Figures. S5-S8).

Figure 4. Dh-Cprotein relationships for GNP-glycan (10 nM) bindingto wildtype DC-SIGN/R at different concentrations. (A) DC-SIGNbinding to GNP-DiMan (black squares, data fitted by Hill’s equa-tion: Y = a + (b-a) × Xn/(kn + Xn); where a =11.1±0 nm; b = 80±99nm; k =200± 990 nM and n =0.53±0.36, R2 = 0.997) or GNP-Man(red dots, data fitted by a linear relationship, Y = a + b × X; wherea = 11.6 ±0.7 nm; b = 0.091±0.010, R2 = 0.928); (B) DC-SIGNRbinding to GNP-(DiMan)3 (blue triangles) or GNP-(Man)3 (greentriangles).

To provide a more quantitative explanation of the data, herewe introduced a new term: the potential of assembly formation(PAF):

PAF = C0(GNP) × C0(protein)/(Kd)2 (3)

Where C0(GNP), C0(protein) and Kd are GNP-glycan and proteinstarting concentrations and their apparent binding Kd, respec-tively. As Kd indicates 50% binding at equilibrium, the PAFthus represents the ratio of a reaction quotient to equilibrium ofa reversible binding reaction. Where PAF =1 would indicateequilibrium (50%) binding, while PAF >1 indicates binding be-ing favorable and PAF >10 being strongly favorable (>90%).Similarly, PAF < 1 indicates binding being unfavorable andPAF <0.1 being strong unfavorable (<10% binding).

Using this definition and the apparent Kds given in Table 1,the PAFs for DC-SIGN binding to GNP-Man and GNP-DiManwere in the range of 0.18-1.8, and 13-131, respectively. Whilethose for DC-SIGNR binding to GNP-(Man)3 and GNP-(DiMan)3 ranged from 0.011 to 0.11, and 0.11 to 1.1. Combin-ing the PAF value and binding mode would allow us to predictthe assembly outcome. For example, with a PAF of >10 andsimultaneous binding, most of the added DC-SIGN moleculesshould readily bind to GNP-DiMan to saturate its surface bind-ing capacity, giving rise to a rapidly increasing and quickly sat-urated Dh with the increasing protein concentration. This wasexactly what was measured from DLS. In contrast, the PAF ofDC-SIGN-GNP-Man binding span across 1, which would indi-cate a gradually increased binding, again matching well to apositive, linear Dh-concentration relationship (Figure 4A). Forinter-crosslinking binding between DC-SIGNR and GNP-glycans, large-scale assemblies were only observed for GNP-(DiMan)3 mixed with the highest protein concentration (200nM) which had a PAF of 1.1 and indicated the binding becamefavorable. While all other conditions where PAFs were < 1, nosignificant binding (assembly) was observed (Figure 4B). To-gether, these results established PAF as a useful indicator forpredicting the binding (assembly) outcome of multivalent lec-tin-glycan binding partners. Furthermore, the excellent agree-ment between the predicted outcome based on PAFs (derivedfrom Kds measured via the GNP quenching method) and thoseobserved from DLS implied that this new fluorescence quench-ing based binding affinity method was highly credible and reli-able. Interestingly, the ability of lectins to bind and inter-cross-link with virus surface glycoproteins had shown to play an im-portant role in virus neutralization.64-66 Thus, we anticipate thatthe PAF criteria proposed here could act as a useful tool to pre-dict lectin-virus interactions and neutralization.

Blocking DC-SIGN/R-mediated augmentation of Ebola vi-rus glycoprotein-driven transduction. The distinct bindingmodes and affinities of GNP-glycans with two important viralreceptors, DC-SIGN/R, should result in different abilities inblocking DC-SIGN/R-mediated virus infection of host cells. Toinvestigate this potential and any possible correlation, humanembryonic kidney 293T cells transfected to express DC-SIGNor DC-SIGNR on their membrane were used as described pre-viously.35 Murine leukemia virus (MLV) vector particles bear-ing the Ebola virus surface glycoprotein (EBOV-GP) and en-coding the luciferase gene were used to model Ebola virus entryinto cells. The virus particles can bind efficiently to DC-SIGN/R (via their surface EBOV-GP) and binding results in in-creased transduction as determined by luciferase expression inhost cells.34-35 Binding of high affinity GNP-glycans to host cellsurface DC-SIGN/R could block these lectin receptors from fur-ther binding to EBOV-GP, reducing luciferase gene transduc-tion. Indeed, pretreatment of 293T cells with GNP-glycans ef-

ficiently inhibited DC-SIGN-mediated augmentation of trans-duction in a dose-dependent manner (see SI, Figure S10). GNP-DiMan/-(DiMan)3 were found to be more potent in blocking thetransduction than GNP-Man/-(Man)3. This result was consistentwith the former being stronger DC-SIGN binders than the latter.The dose-dependent inhibition data were fitted well by a simplecompetition model:

NA = IC50/[IC50 + C] (4)

Where NA, IC50 and C were normalized luciferase activity, GNPconcentration giving 50% inhibition, and GNP concentration,respectively. Whilst DC-SIGNR-mediated augmentation oftransduction was also inhibited by GNP-glycans, the resulting

dose-dependent inhibition data were best fitted by a modifiedcompetition model:

NA = EC50n/[EC50

n + C n] (5)

Where the inhibition coefficient, n, was found to be in the rangeof 0.4 to 0.6 (Table 2). Moreover, GNP-glycans could not com-pletely inhibit DC-SIGNR-mediated transduction, even at highconcentrations (see Figure 5A/B, and also SI, Figure S11). Thisresult revealed that GNP-glycans were much less effective inblocking DC-SIGNR-mediated augmentation of viral genetransduction as compared to that of DC-SIGN-dependent trans-duction.

Figure 5. Plot of normalized luciferase activities of DC-SIGN- or DC-SIGNR- expressing 293T cells as a function of the concentration ofGNP-DiMan (A) or GNP-(DiMan)3 (B) inhibitors. Data for virus particles bearing the EBOV-GP are shown in dots while those of controlvirus particles bearing the VSV-G are shown in triangles. Schematics beneath showing the different inhibition efficiencies of GNP-glycansagainst DC-SIGN (C) or DC-SIGNR (D) mediated infections. (C) For DC-SIGN expressing host cells, all of the binding sites of its surfaceDC-SIGNs are bound and blocked by GNP-glycans, making them unavailable for further binding to virus surface EBOV-GPs to initiateinfection. (D) For DC-SIGNR expressing host cells, GNP-glycans cross-link with some surface DC-SIGNRs but the bindings are weak anddynamic at the edges, allowing the virus to access the unblocked DC-SIGNR binding sites to initiate infection.

Table 2. Summary of inhibition data for GNP-glycans against DC-SIGN/R-mediated EBOV-GP driven infection of 293T host cells.* shows the EC50 values for DC-SIGNR.

GNP-glycan Receptor IC/EC50

(nM)

n R2

GNP-Man DC-SIGN 0.26±0.08 1 0.889

GNP-(Man)3 DC-SIGN 1.57±0.25 1 0.937

GNP-DiMan DC-SIGN 0.095±0.017 1 0.921

GNP-(DiMan)3 DC-SIGN 0.15±0.03 1 0.890

GNP-Man DC-SIGNR 7.3±1.2* 0.42±0.04 0.980

GNP-(Man)3 DC-SIGNR 53±17* 0.61± 0.14 0.902

GNP-DiMan DC-SIGNR 0.63±0.09* 0.56±0.02 0.990

GNP-(DiMan)3 DC-SIGNR 2.6±0.1* 0.53±0.02 0.997

We believe this result is reasonable from both the binding affin-ity (GNP-glycans bind more weakly to DC-SIGNR than to DC-

SIGN) and binding mode points of view. Each GNP-glycancould bind simultaneously to all 4 of DC-SIGN’s binding sites,completely blocking them from further binding to virus surfaceEBOV-GPs to initiate cell entry (Figure 5C). In contrast, theinter-crosslinking mode of binding between GNP-glycans andDC-SIGNR meant that each GNP could only bind to 2 of the 4binding sites in DC-SIGNR. It would be very difficult for GNP-glycans to cross-link all cell surface DC-SIGNRs to fully blocktheir binding sites. Even if this was possible, the DC-SIGNRsat the cluster edge would still only bind divalently (via 2 out-ward facing binding sites) to GNP-glycans, where the bindingwould be weak and dynamic. These, together with any unboundbinding sites, could act as toe-holds for binding to viral surfaceEBOV-GPs to initiate cell uptake and infection (Figure 5D).Therefore, an inhibitor that inter-crosslinks with target recep-tors would be much less effective in blocking virus infectionthan its simultaneous binding counterpart, making it almost im-possible to achieve complete inhibition (Figure 5C/D). Moreo-ver, luciferase activities of control virus particles bearing the

vesicular stomatitis virus glycoprotein (VSV-G), which cannotuse DC-SIGN/R for augmentation of cell entry, were not af-fected significantly by GNP-glycan treatment, confirming thatthe observed inhibitory effects were specific (SI, Figure S10).Importantly, the IC50 value of GNP-DiMan in blocking DC-SIGN-mediated augmentation of EBOV-GP-dependent entrywas measured as 0.095±0.017 nM. This value was considerablylower (by almost an order of magnitude) than those measuredfor some of the most potent glycoconjugate inhibitors of Ebolavirus infection reported previously (e.g. the giant globular mul-tivalent glycofullerenes, IC50: 0.67 nM21, the virus-like gly-codendrinanoparticles, IC50: 0.91 nM,16 and our previous QD-EG3-DiMan, IC50: 0.70±0.2 nM35). This result demonstrated anoutstanding potency of the GNP-glycans in blocking DC-SIGN-mediated EBOV-GP driven infection of host cells. Im-portantly, as virus surface glycans are maintained by host cell’sglycosylation apparatus, they are mostly conserved and unaf-fected by virus mutation.67 Therefore, the potential advantagesof developing potent glycoconjugate viral inhibitors over otheranti-viral strategies include two folds: (1) it can reduce thechances of virus mutation by blocking its entry to host cells, and(2) its treatment potency is unlikely affected by virus mutation,allowing us to provide a potentially long lasting solution.

ConclusionIn summary, we have developed polyvalent GNP-glycans asnew powerful structural and mechanistic probes for multivalentlectin-glycan interactions. By exploiting GNPs’ outstandingfluorescence quenching property, we have developed a newmethod for quantifying multivalent lectin-glycan interaction,revealing that a polyvalent display of mannose containing gly-cans on GNPs greatly enhances their binding affinities with twovitally important viral receptors, DC-SIGN/R. Moreover, by ex-ploiting GNPs’ nanoscale size and high TEM contrast, we haverevealed the distinct binding mode and affinity enhancingmechanisms for GNP-glycans binding with DC-SIGN/R. Im-portantly, the assembly outcomes of the dose-dependent GNP-lectin binding are in perfect agreement with those predictedfrom their respective binding modes and affinities, verifying thecredibility of our new affinity method. Furthermore, we havefound that GNP-glycans can potently block DC-SIGN-mediated EBOV-GP driven viral infections of host cells withIC50 values down to 95±17 pM, making it the most potent gly-coconjugate inhibitor against EBOV-GP driven infection. No-tably, its inhibition potency will unlikely be affected by virusmutation, allowing us to provide potentially a long lasting solu-tion. Finally, we have revealed that only a GNP-glycan inhibitorshowing a simultaneous-, but not a crosslinking-, binding modeis able to completely block the target lectin mediated virus in-fection. Together, these data provide a useful guidance in de-signing polyvalent glycoconjugates for potent, specific inhibi-tion of virus infection. Given their low-/non- toxicity and excel-lent biocompatibility, the GNP-glycans are perfectly placed fora wide range of applications, from probing fundamental struc-tural mechanisms of glycobiology, developing novel biophysi-cal and biomedical assays, to developing novel therapeuticsagainst deadly virus infections and immune dysregulation dis-eases such as cancer, allergy and autoimmune diseases.

Experimental SectionMaterials. Gold nanoparticle (nominal diameter 5 nm, max ~520nm) was purchased from Sigma Aldrich as a stabilized suspensionin citrate buffer or synthesized in house. 2-(2-aminoethoxy)-etha-nol, Di-tert-butyl decarbonate, sodium sulfate, sodium hydride

(60% dispersion in mineral oil), 3-bromo-1-propyne, potassium hy-droxide, trifluoroacetic acid, triethylamine, sodium bicarbonate, O-(6-Chlorobenzotriazol-1-yl)-N, N, N’, N’-tetramethyluronium hexafluorophosphate (HCTU), lipoic acid (LA), tris(2-carboxyethyl)phosphine hydrochloride (TCEP.HCl), copper sulfate, sodiumascorbate, Tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl] amine(TBTA), tris(hydroxymethyl)aminomethane (Tris base), guanidinehydrochloride, anhydrous DMF and other chemicals were pur-chased from Sigma Aldrich, Alfa Aesar, Fluorochem, Thermo Sci-entific, VWR International or Acros organics with > 99% impurityand used as received without further purification unless specifiedelsewhere. All the solvents were obtained in >99% purity fromFischer Scientific and used as received. When used as reaction sol-vents in anhydrous reactions, THF and CH2Cl2 were dried and de-oxygenated using an Innovative Technology Inc. PureSolv® sol-vent purification system. Ultra-pure water (resistance >18.2MΩ.cm) purified by an ELGA Purelab classic UVF system, was used for all experiments and making buffers.59, 62

Glycan-ligand synthesis by click chemistry. The general proto-cols employed in preparing the LA-glycan ligands via click chem-istry were as follow.16 To a 1:1 (v:v) mixed THF:H2O solution (2.0-5.0 mL) containing the glycan-EG2-N3 (1.1 mole equiv.) and LA-linker (1 mole equiv.) were added CuSO4.5H2O (0.036 moleequiv.), TBTA (0.063 mole equiv.) followed by sodium ascorbate(0.135 mole equiv.) and the resulting solution was stirred at RT.After 3 h, TLC confirmed the consumption of all starting materials.The solvent was then evaporated and the crude product was puri-fied by size exclusion chromatography using Biogel P2 column us-ing water as eluent to afford the desired pure LA-glycan product.

LA-Man: Yield = (67%, 0.18 mmol). 1H NMR (500 MHz, CD3OD)δ (ppm): 8.04 (s, 1H, triazole-H), 4.8 (d, 1H, J 1.6 Hz, Man H-1),4.65 (s, 2H), 4.6 (t, 2H, J 5 Hz), 3.9 (t, 2H, J 5.1 Hz), 3.86-3.81 (m,3H), 3.8 (dd, 1H, J 1.7 Hz, 3.4 Hz, Man H-2), 3.72 (d, 1H, J 5.5Hz), 3.69-3.67 (m, 3H), 3.66 (s, 2H), 3.65-3.63 (m, 2H), 3.62-3.61(m, 4H), 3.60-3.57 (m, 2H), 3.53 (t, 2H, J 5.5 Hz), 3.39-3.34 (m,3H), 3.20-3.15 (m, 1H), 3.12-3.07 (m, 1H), 2.49-2.42 (m, 1H), 2.20(t, 2H, J 7.4 Hz), 1.92-1.85 (m, 1H), 1.73-1.59 (m, 4H), 1.50-1.40(m, 2H). 13C NMR (125 MHz, CD3OD) δ: 176.1 (C=O), 145.7 (C=CH), 126.0 (C=CH), 101.7 (Man C-1), 74.6, 72.6, 72.1 (ManC-2), 71.6, 71.5, 71.4, 71.3, 70.8, 70.6, 70.4, 68.6, 67.7, 65.0, 63.0(Man C-6), 57.6, 51.5, 41.3, 40.3, 39.3, 36.8, 35.7, 29.8, 26.7.HRMS: calculated m/z for C27H49N4O11S2 (M+H)+ 669.2834;found 669.2838.

LA-DiMan: Yield = (65%, 0.080 mmol). 1H NMR (500 MHz,CD3OD) δ (ppm): 8.05 (s, 1H, triazole-H), 5.11 (d, J 1.7 Hz, 1H),4.96 (d, J 1.8 Hz, 1H), 4.66 (s, 2H), 4.60 (t, 2H, J 5.1 Hz), 3.97 (dd,1H, J 3.3, 1.8 Hz), 3.91 (dd, 2H, J 5.5, 4.7 Hz), 3.88 – 3.78 (m, 6H),3.72 – 3.65 (m, 8H), 3.62 (dt, 6H, J 5.0, 1.3 Hz), 3.58 (dt, 3H, J11.2, 1.6 Hz), 3.53 (t, 2H, J 5.5 Hz), 3.43 – 3.32 (m, 3H), 3.17 (ddd,1H, J 10.9, 7.1, 5.4 Hz), 3.10 (dt, 1H, J 11.0, 6.9 Hz, 1H), 2.46 (dtd,1H, J 12.1, 6.7, 5.4 Hz), 2.20 (t, 2H, J 7.4 Hz), 1.89 (dq, 1H, J 12.7,6.9 Hz), 1.77 – 1.57 (m, 4H), 1.45 (qt, 2H, J 9.4, 5.9 Hz). 13C NMR(125 MHz, CD3OD) δ: 176.1 (C=O), 145.7 (C=CH), 126.0 (C=CH), 104.2 (Man C-1), 100.1 (Man C-1’), 80.6, 75.1, 74.7,72.4, 72.1, 71.9, 71.6, 71.5, 71.4, 71.3, 70.8, 70.6, 70.5, 69.0, 68.8,67.9, 65.0, 63.2, 63.1, 57.6, 51.5, 41.3, 40.4, 39.3, 36.8, 35.7, 29.8,26.7. HRMS: calculated m/z for C33H58N4O16S2 (M+H)+ 831.3233;found 831.3242.

LA-(Man)3: Yield = (60%, 0.060 mmol). 1H NMR (500 MHz, D2O)δ (ppm): 8.03 (s, 3H, triazole-H), 4.87 (d, 3H, J 1.8 Hz, H-1), 4.64– 4.61 (m, 6H), 4.60 (s, 6H), 3.98 – 3.96 (m, 6H), 3.95 (dd, 3H, J3.5, 1.8 Hz, H-2), 3.89 (d, 1H, J 2.0 Hz), 3.87 (d, 2H, J 1.7 Hz),3.84 – 3.79 (m, 6H), 3.77 (d, 2H, J 5.6 Hz), 3.74 (d, 7H, J 3.1 Hz),3.67 (d, 3H, J 9.8 Hz), 3.65 – 3.61 (m, 25H), 3.61 – 3.53 (m, 3H),3.21 – 3.12 (m, 2H), 2.42 (dq, 1H, J 12.3, 6.1 Hz), 2.18 (t, 2H, J

7.0 Hz), 1.91 (dt, 1H, J 13.6, 6.8 Hz), 1.71 – 1.64 (m, 1H), 1.58 –1.48 (m, 4H), 1.33 (q, 3H, J 7.7 Hz). 13C NMR (125 MHz, D2O) δ: 176.5 (C=O), 144.0, 125.3, 99.9 (C-1), 72.7, 70.5, 69.9, 69.5, 69.4,68.8, 67.5, 66.7, 66.3, 63.5, 60.9, 59.7, 56.5, 50.0, 40.2, 38.0, 35.9,33.7, 27.5, 24.9. HRMS: calculated m/z for C57H99N10O28S2

(M+H)+ 1435.6066; found 1435.6096.

LA-(DiMan)3: Yield = (69%, 0.030 mmol). 1H NMR (500 MHz,D2O) δ (ppm): 8.03 (s, 3H, triazole-H), 5.11 (d, 3H, J 1.8 Hz, H-1),5.03 (d, 3H, J = 1.9 Hz, H-1’), 4.64 – 4.62 (m, 6H), 4.60 (d, 6H, J2.9 Hz), 4.08 (dd, 3H, J 3.4, 1.8 Hz, H-2), 3.97 (td, 9H, J = 4.5, 2.5Hz), 3.92 – 3.68 (m, 40H), 3.66 – 3.61 (m, 29H), 3.58 – 3.53 (m,3H), 3.21 – 3.13 (m, 2H), 2.42 (dq, 1H, J 12.3, 6.1 Hz), 2.18 (t, 2H,J 7.0 Hz), 1.91 (dd, 1H, J 13.1, 6.8 Hz), 1.67 (m, 1H), 1.59 – 1.49(m, 4H), 1.32 (p, 3H, J 7.7 Hz). ). 13C NMR (125 MHz, D2O) δ: 179.0 (C=O), 146.5, 127.9, 104.8 (C-1), 100.9 (C-2), 81.2, 75.8,75.3, 74.2, 72.8, 72.7, 72.5, 72.2, 72.0 (2), 71.9, 71.3 (2), 70.1, 69.4(2), 69.0, 66.1, 63.7, 63.4, 62.9, 62.3, 59.0, 52.6, 42.8, 40.6, 38.4,36.2, 30.1, 27.4. HRMS: calculated m/z for C75H130N10O43S2

(M+H)2+ 961.3862; found 961.3858.

GNP-glycan preparation. 10 mL of 5 nm GNP stock was concen-trated to 250 L using a 30 KDa MWCO spin column and washedwith H2O (3 × 200 µL) to remove any impurities. Then DHLA-glycan ligands dissolved in H2O were added to the GNP solution ina molar ratio of GNP:DHLA-glycan = 1:1000. The resulting solu-tion was mixed and stirred at RT in dark for overnight. The result-ing mixture was passed through a 30 KDa MWCO spin column bycentrifugation at 15,000 × g for 2 min and the residues were washedwith H2O (3 × 200 µL) to give the GNP-glycan stock. The filtrateand washing through liquids were combined and used to evaluatethe glycan loading on GNPs as described previously.34 The concen-tration of the GNP-glycans was determined from its absorbance at520 nm using the Beer-Lambert law and a GNP molar extinctionco-efficient of 1.1 × 107 M−1⋅cm−1.

Protein production and labeling. The soluble extracellular seg-ments of DCSIGN (DC020) and DCSIGNR (DSR034) were ex-pressed as inclusion bodies in E. coli and purified by Man-Se-pharose affinity column followed by Superdex size exclusion col-umn as reported previously.34 The mutant proteins, DC020 Q-274Cand DSR034 R278C were constructed by site directed mutagenesisand labelled with atto594-maleimide as described previously.35, 40

The labeled proteins were purified by mannose-Sepharose affinitycolumn. All the proteins obtained were characterized by high reso-lution mass spectroscopy (HRMS) and their corresponding spectrawere shown in SI, Figure S1. The dye labeling efficiency (per pro-tein monomer) for DC-SIGN and DC-SIGNR was determined tobe 87 % and 82 %, respectively, based on the relative intensity ofthe labeled and unlabeled protein peaks measured from MS.

Fluorescence spectra. All fluorescence spectra were recorded ona Cary Eclipse Fluorescence Spectrophotometer using a 0.70 mLquartz cuvette under a fixed λEX of 595 nm over a range of 605-750nm. All measurements were performed in a binding buffer (20 mMHEPES, 100 mM NaCl, 10 mM CaCl2, pH 7.8) containing 1 mg/mLof BSA to minimize any nonspecific absorption of the GNPs andproteins on surfaces. For apparent Kd measurement, the concentra-tions of labeled DC-SIGN or DC-SIGNR protein and GNP-glycanswere varied simultaneously in a fixed protein:GNP molar ratio of1:1. The samples were incubated at RT for 20 min before recordingthe fluorescence spectra. The fluorescence spectra of the protein inthe absence of the GNP-glycans, recorded under identical experi-ment conditions, were used to determine the quenching efficiency.The instrument PMT voltages were adjusted to compensate the lowfluorescence signals at low concentrations. The quenching effi-ciency (QE) of DC-SIGN/R binding to each GNP-glycan was cal-culated via Eq. 1 and the resulting QE-concentration relationship

was fitted by Hill’s equation (Eq. 2) to derive the apparent bindingKd values.

Dynamic light scattering. The hydrodynamic diameters (Dh) ofwildtype DC-SIGN/R and GNP-glycans in pure water and in abinding buffer (20 mM HEPES, 100 mM NaCl, 10 mM CaCl2, pH7.8) were recorded on a Malvern ZETASizer-Nano using disposa-ble polystyrene cuvettes as reported in our previous papers.35, 68 Formonitoring GNP-DiMan binding to wildtype DC-SIGN/R, a seriesof samples, each containing GNP-DiMan (10 nM) and a varyingamount of wildtype DC-SIGN/R, were mixed in a binding buffer atRT for 20 mins before DLS measurement was carried out. 10 con-secutive scans were performed for each sample and the resultingvolume Dh distribution histograms for each sample were combinedand fitted by Gaussian distributions to estimate their Dhs.35

TEM imaging. Three samples: (1) GNP-DiMan, (2) GNP-DiMan+ wildtype DC-SIGN, and (3) GNP-DiMan + wildtype DC-SIGNRwere incubated (final CGNP = 16 nM and Cprotein=640 nM) in a bind-ing buffer overnight. The samples were prepared by plunge-freez-ing into liquid ethane followed by warming under vacuum to cap-ture the GNP dispersions in their native dispersed state as demon-strated in our previous paper.63 Briefly, 3.5 L of suspension wasplaced onto a plasma-cleaned TEM grid with a continuous carbonsupport film, blotted, and plunge frozen into liquid ethane. TheTEM grids were then warmed to RT over several minutes by plac-ing the specimens in the liquid nitrogen cooled storage container ina rotary pumped vacuum desiccator.

The samples were then analyzed using an FEI Titan3 Themis 300G2 S/TEM equipped with FEI SuperX energy dispersive X-ray(EDX) spectrometers and a Gatan OneView CCD. Images werecollected for each sample, with EDX spectroscopy used to confirmthat the small nanoparticles imaged were indeed GNPs as Au spe-cies were detected together with other peaks (carbon, oxygen, sili-con, and copper) due to the microscope, TEM grid or support film.A series of images at the same magnification were recorded foreach sample, allowing easy comparison of the nanoparticle disper-sion state of the three samples. The TEM images were analyzedautomatically by MATLAB scripts to derive the nearest neighbordistances (NNDs) as reported previously.35

Virus Inhibition. The effects of GNP-glycans (glycan = Man, Di-Man, (Man)3 or (DiMan)3) on Ebola virus glyco-protein (EBOV-GP) driven entry into 293T cells were assessed by using our estab-lished procedures.34-35 Briefly, 293T cells seeded in 96-well plateswere transfected with plasmids encoding DC-SIGN or DC-SIGNRor control transfected with empty plasmid (pcDNA). The cells werewashed at 16 h post transfection and further cultivated at 37°C, 5%CO2 in Dulbecco's modified eagle medium (DMEM) containing10% fetal bovine serum (FBS). At 48 h post transfection, the cellswere exposed to twice the final concentration of GNP-glycan in-hibitor in DMEM supplemented with 10% FBS for 30 min in a totalvolume of 50 µL. Thereafter, the resulting cells were inoculatedwith 50 µL of preparations of MLV vector particles encoding theluciferase gene and bearing either EBOV-GP (which can use DC-SIGN/R for augmentation of host cell entry) or the vesicular sto-matitis virus glycoprotein (VSV-G, which cannot use DC-SIGN/Rfor augmentation of host cell entry). Under these conditions, bind-ing of GNP-glycan nanoparticles to 293T cell surface DC-SIGN/Rreceptors can block EBOV-GP interactions with these lectin recep-tors, resulting in reduced transduction efficiency of the virus parti-cles and hence reducing the cellular luciferase activity. At 6 h postinoculation, 100 µL of fresh DMEM culture medium was addedand the cells incubated for another 72 h. Thereafter, luciferase ac-tivities in cell lysates were determined using a commercially avail-able kit (PJK), following the manufacturer’s instructions, as de-scribed in our previous publication.

ASSOCIATED CONTENT

Supporting Information. Experimental details including materi-als and methods, protein production, labeling and MS characteriza-tion, LA-glycan ligands syntheses and their 1H and 13C NMR spec-tra, and supporting figures showing fluorescence spectra of GNP-DiMan binding to DC-SIGN, mannose competition studies, Dh dis-tribution histograms of the GNP-glycans, DC-SIGN/R, and GNP-glycan binding to DC-SIGN/R at different concentrations, TEMimages of cryo-prepared GNP-DiMan + DC-SIGN/R samples, andthe original data of GNP-glycans inhibiting DC-SIGN/R-mediatedEBOV-GP driven infections.

AUTHOR INFORMATION

Corresponding Author(s)

* Stefan Pöhlmann- Infection Biology Unit, German Primate Cen-ter, Kellnerweg 4, 37077 Gottingen, Germany; orcid.org/0000-0001-6086-9136; Email: spoehlmann@dpz.eu.

* Yuan Guo- School of Food Science&Nutrition and Astbury Cen-ter for Structural Molecular Biology, University of Leeds, UK; or-cid.org/0000-0003-4607-7356; Email: y.guo@leeds.ac.uk .

* Dejian Zhou- School of Chemistry and Astbury Center for Struc-tural Molecular Biology, University of Leeds, UK; orcid.org/ 0000-0003-3314-9242; Email: d.zhou@leeds.ac.uk.

Author Contributions

D.Z., Y.G. and S.P. conceived the idea. D.B. and E.P. conductedligand synthesis, GNP conjugation and GNP-lectin binding studies.I.N. conducted virus inhibition assays; D.B., Y.L. and B.W.T. syn-thesized glycans; J.H. and E.K. performed some protein purifica-tion, characterization and DLS studies. U.S.A. provided someGNPs. N.H. performed TEM imaging studies. D.Z., Y.G. and D. B.analyzed data and wrote the paper. All authors have given approvalto the final version of the manuscript.

╬ These authors contributed equally to this work.

Funding Sources

UK Biotechnology and Biological Sciences Research Council(grant no: BB/R007829/1). EU Horizon 2020 Marie Sklodowska-Curie Fellowship (grant no: 797597). UK Engineering and PhysicalSciences Research Council DTP PhD scholarship (grant no:EP/M50807X/1).

Notes

The authors declare no competing financial interests.

ACKNOWLEDGMENT

We thank the UK Biotechnology and Biological Sciences ResearchCouncil (grant no: BB/R007829/1) and the EU Horizon 2020 via aMarie Sklodowska-Curie Fellowship (grant no: 797597) for fund-ing this research. E.P. thanks the University of Leeds and the UKEngineering and Physical Sciences Research Council for providinga DTP PhD scholarship (grant no: EP/M50807X/1).

ABBREVIATIONS

GNP, gold nanoparticle; DC-SIGN, Dendritic Cell-Specific Inter-cellular adhesion molecule-3-Grabbing Non-integrin; DC-SIGNR,DC-SIGN related lectin found on endothelial cells; TLC, thin layerchromatography; HPLC: high performance liquid chromatography;NMR, nuclear magnetic resonance; MS, mass spectrometry.

REFERENCES

1. Brown, G. D.; Willment, J. A.; Whitehead, L., C-type lectinsin immunity and homeostasis. Nat. Rev. Immunol. 2018, 18 (6), 374-389.

2. Cummings, R. D., "Stuck on sugars - how carbohydratesregulate cell adhesion, recognition, and signaling". Glycoconj. J. 2019,36 (4), 241-257.

3. Geijtenbeek, T. B.; Gringhuis, S. I., Signalling through C-type lectin receptors: shaping immune responses. Nat. Rev. Immunol.2009, 9 (7), 465-79.

4. Macauley, M. S.; Crocker, P. R.; Paulson, J. C., Siglec-me-diated regulation of immune cell function in disease. Nat. Rev. Immu-nol. 2014, 14 (10), 653-666.

5. Rabinovich, G. A.; van Kooyk, Y.; Cobb, B. A., Glycobiol-ogy of immune responses. Ann. N. Y. Acad. Sci. 2012, 1253, 1-15.

6. Taylor, M. E.; Drickamer, K., Paradigms for glycan-bindingreceptors in cell adhesion. Curr. Opin. Cell. Biol. 2007, 19 (5), 572-577.

7. Thompson, A. J.; de Vries, R. P.; Paulson, J. C., Virus recog-nition of glycan receptors. Curr. Opin. Virol. 2019, 34, 117-129.

8. Wesener, D. A.; Dugan, A.; Kiessling, L. L., Recognition ofmicrobial glycans by soluble human lectins. Curr. Opin. Struct. Biol.2017, 44, 168-178.

9. Illescas, B. M.; Rojo, J.; Delgado, R.; Martin, N., Multiva-lent Glycosylated Nanostructures To Inhibit Ebola Virus Infection. J.Am. Chem. Soc. 2017, 139 (17), 6018-6025.

10. Mammen, M.; Choi, S. K.; Whitesides, G. M., PolyvalentInteractions in Biological Systems: Implications for Design and Use ofMultivalent Ligands and Inhibitors. Angew. Chem. Int. Ed. 1998, 37(20), 2754-2794.

11. Kitov, P. I.; Sadowska, J. M.; Mulvey, G.; Armstrong, G. D.;Ling, H.; Pannu, N. S.; Read, R. J.; Bundle, D. R., Shiga-like toxins areneutralized by tailored multivalent carbohydrate ligands. Nature 2000,403 (6770), 669-672.

12. Dam, T. K.; Brewer, C. F., Multivalent lectin-carbohydrateinteractions energetics and mechanisms of binding. Adv. Carbohydr.Chem. Biochem. 2010, 63, 139-164.

13. Branson, T. R.; McAllister, T. E.; Garcia-Hartjes, J.;Fascione, M. A.; Ross, J. F.; Warriner, S. L.; Wennekes, T.; Zuilhof,H.; Turnbull, W. B., A protein-based pentavalent inhibitor of the chol-era toxin B-subunit. Angew. Chem. Int. Ed. 2014, 53 (32), 8323-8327.

14. Bernardi, A.; Jimenez-Barbero, J.; Casnati, A.; De Castro,C.; Darbre, T.; Fieschi, F.; Finne, J.; Funken, H.; Jaeger, K. E.; Lah-mann, M.; Lindhorst, T. K.; Marradi, M.; Messner, P.; Molinaro, A.;Murphy, P. V.; Nativi, C.; Oscarson, S.; Penades, S.; Peri, F.; Pieters,R. J.; Renaudet, O.; Reymond, J. L.; Richichi, B.; Rojo, J.; Sansone, F.;Schaffer, C.; Turnbull, W. B.; Velasco-Torrijos, T.; Vidal, S.; Vincent,S.; Wennekes, T.; Zuilhof, H.; Imberty, A., Multivalent glycoconju-gates as anti-pathogenic agents. Chem. Soc. Rev. 2013, 42 (11), 4709-4927.

15. Feinberg, H.; Castelli, R.; Drickamer, K.; Seeberger, P. H.;Weis, W. I., Multiple modes of binding enhance the affinity of DC-SIGN for high mannose N-linked glycans found on viral glycoproteins.J. Biol. Chem. 2007, 282 (6), 4202-4209.

16. Ribeiro-Viana, R.; Sanchez-Navarro, M.; Luczkowiak, J.;Koeppe, J. R.; Delgado, R.; Rojo, J.; Davis, B. G., Virus-like gly-codendrinanoparticles displaying quasi-equivalent nested polyvalencyupon glycoprotein platforms potently block viral infection. Nat. Com-mun. 2012, 3, article no: 1303.

17. Martinez, A.; Ortiz Mellet, C.; Garcia Fernandez, J. M., Cy-clodextrin-based multivalent glycodisplays: covalent and supramolec-ular conjugates to assess carbohydrate-protein interactions. Chem. Soc.Rev. 2013, 42 (11), 4746-4773.

18. Reichardt, N. C.; Martin-Lomas, M.; Penades, S., Glyco-nanotechnology. Chem. Soc. Rev. 2013, 42 (10), 4358-4376.

19. Spinelli, N.; Defrancq, E.; Morvan, F., Glycoclusters on ol-igonucleotide and PNA scaffolds: synthesis and applications. Chem.Soc. Rev. 2013, 42 (11), 4557-4573.

20. Muller, C.; Despras, G.; Lindhorst, T. K., Organizing multi-valency in carbohydrate recognition. Chem. Soc. Rev. 2016, 45 (11),3275-3302.

21. Munoz, A.; Sigwalt, D.; Illescas, B. M.; Luczkowiak, J.; Ro-driguez-Perez, L.; Nierengarten, I.; Holler, M.; Remy, J. S.; Buffet, K.;Vincent, S. P.; Rojo, J.; Delgado, R.; Nierengarten, J. F.; Martin, N.,Synthesis of giant globular multivalent glycofullerenes as potent inhib-itors in a model of Ebola virus infection. Nat. Chem. 2016, 8 (1), 50-57.

22. Martinez-Avila, O.; Hijazi, K.; Marradi, M.; Clavel, C.;Campion, C.; Kelly, C.; Penades, S., Gold manno-glyconanoparticles:multivalent systems to block HIV-1 gp120 binding to the lectin DC-SIGN. Chem. Eur. J. 2009, 15 (38), 9874-9888.

23. Benito-Alifonso, D.; Tremel, S.; Hou, B.; Lockyear, H.;Mantell, J.; Fermin, D. J.; Verkade, P.; Berry, M.; Galan, M. C., Lac-tose as a "Trojan horse" for quantum dot cell transport. Angew. Chem.Int. Ed. 2014, 53 (3), 810-814.

24. Bhatia, S.; Camacho, L. C.; Haag, R., Pathogen Inhibitionby Multivalent Ligand Architectures. J. Am. Chem. Soc. 2016, 138(28), 8654-8666.

25. Holla, A.; Skerra, A., Comparative analysis reveals selectiverecognition of glycans by the dendritic cell receptors DC-SIGN andLangerin. Protein Eng. Des. Sel. 2011, 24 (9), 659-669.

26. Diehl, C.; Engstrom, O.; Delaine, T.; Hakansson, M.;Genheden, S.; Modig, K.; Leffler, H.; Ryde, U.; Nilsson, U. J.; Akke,M., Protein flexibility and conformational entropy in ligand design tar-geting the carbohydrate recognition domain of galectin-3. J. Am. Chem.Soc. 2010, 132 (41), 14577-14589.

27. Turnbull, W. B.; Precious, B. L.; Homans, S. W., Dissectingthe cholera toxin-ganglioside GM1 interaction by isothermal titrationcalorimetry. J. Am. Chem. Soc. 2004, 126 (4), 1047-1054.

28. Giljohann, D. A.; Seferos, D. S.; Daniel, W. L.; Massich, M.D.; Patel, P. C.; Mirkin, C. A., Gold nanoparticles for biology and med-icine. Angew. Chem. Int. Ed. 2010, 49 (19), 3280-3294.

29. Rosi, N. L.; Mirkin, C. A., Nanostructures in biodiagnostics.Chem. Rev. 2005, 105 (4), 1547-1562.

30. Schofield, C. L.; Field, R. A.; Russell, D. A., Glyconanopar-ticles for the colorimetric detection of cholera toxin. Anal. Chem. 2007,79 (4), 1356-1361.

31. Fang, S. P.; Lee, H. J.; Wark, A. W.; Corn, R. M., Attomolemicroarray detection of MicroRNAs by nanoparticle-amplified SPRimaging measurements of surface polyadenylation reactions. J. Am.Chem. Soc. 2006, 128 (43), 14044-14046.

32. Graham, D.; Thompson, D. G.; Smith, W. E.; Faulds, K.,Control of enhanced Raman scattering using a DNA-based assemblyprocess of dye-coded nanoparticles. Nat. Nanotech. 2008, 3 (9), 548-551.

33. Cao, Y. W. C.; Jin, R. C.; Mirkin, C. A., Nanoparticles withRaman spectroscopic fingerprints for DNA and RNA detection. Sci-ence 2002, 297 (5586), 1536-1540.

34. Guo, Y.; Sakonsinsiri, C.; Nehlmeier, I.; Fascione, M. A.;Zhang, H.; Wang, W.; Pohlmann, S.; Turnbull, W. B.; Zhou, D., Com-pact, Polyvalent Mannose Quantum Dots as Sensitive, RatiometricFRET Probes for Multivalent Protein-Ligand Interactions. Angew.Chem. Int. Ed. 2016, 55 (15), 4738-4742.

35. Guo, Y.; Nehlmeier, I.; Poole, E.; Sakonsinsiri, C.; Hondow,N.; Brown, A.; Li, Q.; Li, S.; Whitworth, J.; Li, Z.; Yu, A.; Brydson,R.; Turnbull, W. B.; Pohlmann, S.; Zhou, D., Dissecting MultivalentLectin-Carbohydrate Recognition Using Polyvalent MultifunctionalGlycan-Quantum Dots. J. Am. Chem. Soc. 2017, 139 (34), 11833-11844.

36. Hooper, J. G., Y.; Zhou, D., , Polyvalent Glycan-QuantumDots as Multifunctional Structural Probes for Multivalent Lectin-Car-bohydrate Interactions. ACS Symp. Series 2020, 1346, 47-66.

37. Geijtenbeek, T. B.; Kwon, D. S.; Torensma, R.; van Vliet, S.J.; van Duijnhoven, G. C.; Middel, J.; Cornelissen, I. L.; Nottet, H. S.;KewalRamani, V. N.; Littman, D. R.; Figdor, C. G.; van Kooyk, Y.,DC-SIGN, a dendritic cell-specific HIV-1-binding protein that en-hances trans-infection of T cells. Cell 2000, 100 (5), 587-597.

38. Guo, Y.; Feinberg, H.; Conroy, E.; Mitchell, D. A.; Alvarez,R.; Blixt, O.; Taylor, M. E.; Weis, W. I.; Drickamer, K., Structural ba-sis for distinct ligand-binding and targeting properties of the receptorsDC-SIGN and DC-SIGNR. Nat. Struct. Mol. Biol. 2004, 11 (7), 591-598.

39. Pohlmann, S.; Soilleux, E. J.; Baribaud, F.; Leslie, G. J.;Morris, L. S.; Trowsdale, J.; Lee, B.; Coleman, N.; Doms, R. W., DC-SIGNR, a DC-SIGN homologue expressed in endothelial cells, bindsto human and simian immunodeficiency viruses and activates infectionin trans. Proc. Natl. Acad. Sci. U.S.A. 2001, 98 (5), 2670-2675.

40. Guo, Y.; Bruce Turnbull, W.; Zhou, D., Probing MultivalentProtein-Carbohydrate Interactions by Quantum Dot-Forster ResonanceEnergy Transfer. Methods Enzymol. 2018, 598, 71-100.

41. Jarvis, C. M.; Zwick, D. B.; Grim, J. C.; Alam, M. M.; Prost,L. R.; Gardiner, J. C.; Park, S.; Zimdars, L. L.; Sherer, N. M.; Kiessling,L. L., Antigen structure affects cellular routing through DC-SIGN.Proc. Natl. Acad. Sci. U.S.A. 2019, 116 (30), 14862-14867.

42. Geijtenbeek, T. B. H.; Gringhuis, S. I., Signalling throughC-type lectin receptors: shaping immune responses. Nat. Rev. Immunol.2009, 9 (7), 465-479.

43. Gringhuis, S. I.; den Dunnen, J.; Litjens, M.; van der Vlist,M.; Geijtenbeek, T. B. H., Carbohydrate-specific signaling through theDC-SIGN signalosome tailors immunity to Mycobacterium tuberculo-sis, HIV-1 and Helicobacter pylori. Nat. Immunol. 2009, 10 (10), 1081-1088.

44. van Kooyk, Y.; Rabinovich, G. A., Protein-glycan interac-tions in the control of innate and adaptive immune responses. Nat. Im-munol. 2008, 9 (6), 593-601.

45. Lee, J. H.; Ozorowski, G.; Ward, A. B., Cryo-EM structureof a native, fully glycosylated, cleaved HIV-1 envelope trimer. Science2016, 351 (6277), 1043-1048.

46. Stewart-Jones, G. B. E.; Soto, C.; Lemmin, T.; Chuang, G.Y.; Druz, A.; Kong, R.; Thomas, P. V.; Wagh, K.; Zhou, T. Q.; Beh-rens, A. J.; Bylund, T.; Choi, C. W.; Davison, J. R.; Georgiev, I. S.;Joyce, M. G.; Do Kwon, Y.; Pancera, M.; Taft, J.; Yang, Y. P.; Zhang,B. S.; Shivatare, S. S.; Shivatare, V. S.; Lee, C. C. D.; Wu, C. Y.; Bew-ley, C. A.; Burton, D. R.; Koff, W. C.; Connors, M.; Crispin, M.; Baxa,U.; Korber, B. T.; Wong, C. H.; Mascola, J. R.; Kwong, P. D., l Tri-meric HIV-1-Env Structures Define Glycan Shields from Clades A, B,and G. Cell 2016, 165 (4), 813-826.

47. Dubertret, B.; Calame, M.; Libchaber, A. J., Single-mis-match detection using gold-quenched fluorescent oligonucleotides.Nat. Biotech. 2001, 19 (4), 365-370.

48. Susumu, K.; Uyeda, H. T.; Medintz, I. L.; Pons, T.; Dele-hanty, J. B.; Mattoussi, H., Enhancing the stability and biological func-tionalities of quantum dots via compact multifunctional ligands. J. Am.Chem. Soc. 2007, 129 (45), 13987-13996.

49. Zhou, D.; Bruckbauer, A.; Ying, L. M.; Abell, C.; Klener-man, D., Building three-dimensional surface biological assemblies onthe nanometer scale. Nano Lett. 2003, 3 (11), 1517-1520.

50. Ramos-Soriano, J.; Reina, J. J.; Illescas, B. M.; de la Cruz,N.; Rodriguez-Perez, L.; Lasala, F.; Rojo, J.; Delgado, R.; Martin, N.,Synthesis of Highly Efficient Multivalent Disaccharide/ 60 FullereneNanoballs for Emergent Viruses. J. Am. Chem. Soc. 2019, 141 (38),15403-15412.

51. Wang, W. L.; Guo, Y.; Tiede, C.; Chen, S. Y.; Kopytynski,M.; Kong, Y. F.; Kulak, A.; Tomlinson, D. R.; Chen, R. J.; McPherson,M.; Zhou, D. J., Ultraefficient Cap-Exchange Protocol To CompactBiofunctional Quantum Dots for Sensitive Ratiometric Biosensing andCell Imaging. ACS Appl. Mater. Interfaces 2017, 9 (18), 15232-15244.

52. Burns, J. A.; Butler, J. C.; Moran, J.; Whitesides, G. M., Se-lective Reduction of Disulfides by Tris(2-Carboxyethyl)phosphine. J.Org. Chem. 1991, 56 (8), 2648-2650.

53. Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.;Whitesides, G. M., Self-assembled monolayers of thiolates on metalsas a form of nanotechnology. Chem. Rev. 2005, 105 (4), 1103-1169.

54. Hill, H. D.; Millstone, J. E.; Banholzer, M. J.; Mirkin, C. A.,The Role Radius of Curvature Plays in Thiolated OligonucleotideLoading on Gold Nanoparticles. ACS Nano 2009, 3 (2), 418-424.

55. Jennings, T. L.; Singh, M. P.; Strouse, G. F., Fluorescentlifetime quenching near d=1.5 nm gold nanoparticles: Probing NSETvalidity. J. Am. Chem. Soc. 2006, 128 (16), 5462-5467.

56. Pons, T.; Medintz, I. L.; Sapsford, K. E.; Higashiya, S.;Grimes, A. F.; English, D. S.; Mattoussi, H., On the quenching of sem-iconductor quantum dot photoluminescence by proximal gold nanopar-ticles. Nano Lett. 2007, 7 (10), 3157-3164.

57. Seferos, D. S.; Giljohann, D. A.; Hill, H. D.; Prigodich, A.E.; Mirkin, C. A., Nano-flares: Probes for transfection and mRNA de-tection in living cells. J. Am. Chem. Soc. 2007, 129 (50), 15477-15479.

58. Song, S. P.; Liang, Z. Q.; Zhang, J.; Wang, L. H.; Li, G. X.;Fan, C. H., Gold-Nanoparticle-Based Multicolor Nanobeacons for Se-quence-Specific DNA Analysis. Angew. Chem. Int. Ed. 2009, 48 (46),8670-8674.

59. Zhang, H. Y.; Feng, G. Q.; Guo, Y.; Zhou, D. J., Robust andspecific ratiometric biosensing using a copper-free clicked quantumdot-DNA aptamer sensor. Nanoscale 2013, 5 (21), 10307-10315.

60. Chung, N. P.; Breun, S. K.; Bashirova, A.; Baumann, J. G.;Martin, T. D.; Karamchandani, J. M.; Rausch, J. W.; Le Grice, S. F.;Wu, L.; Carrington, M.; Kewalramani, V. N., HIV-1 transmission bydendritic cell-specific ICAM-3-grabbing nonintegrin (DC-SIGN) isregulated by determinants in the carbohydrate recognition domain thatare absent in liver/lymph node-SIGN (L-SIGN). J. Biol. Chem. 2010,285 (3), 2100-2112.

61. Feinberg, H.; Mitchell, D. A.; Drickamer, K.; Weis, W. I.,Structural basis for selective recognition of oligosaccharides by DC-SIGN and DC-SIGNR. Science 2001, 294 (5549), 2163-2166.

62. Song, L.; Ho, V. H. B.; Chen, C.; Yang, Z. Q.; Liu, D. S.;Chen, R. J.; Zhou, D. J., Efficient, pH-Triggered Drug Delivery Usinga pH-Responsive DNA-Conjugated Gold Nanoparticle. Adv.Healthcare Mater. 2013, 2 (2), 275-280.

63. Hondow, N.; Brydson, R.; Wang, P. Y.; Holton, M. D.;Brown, M. R.; Rees, P.; Summers, H. D.; Brown, A., Quantitative char-acterization of nanoparticle agglomeration within biological media. J.Nanopart. Res. 2012, 14 (7), article no: 977.

64. Williams, J. A.; Gui, L.; Hom, N.; Mileant, A.; Lee, K. K.,Dissection of Epitope-Specific Mechanisms of Neutralization of Influ-enza Virus by Intact IgG and Fab Fragments. J. Virol. 2018, 92 (6),article no: e02006-17.

65. Yang, M. L.; Chen, Y. H.; Wang, S. W.; Huang, Y. J.; Leu,C. H.; Yeh, N. C.; Chu, C. Y.; Lin, C. C.; Shieh, G. S.; Chen, Y. L.;Wang, J. R.; Wang, C. H.; Wu, C. L.; Shiau, A. L., Galectin-1 Binds toInfluenza Virus and Ameliorates Influenza Virus Pathogenesis. J. Vi-rol. 2011, 85 (19), 10010-10020.

66. Watson, A.; Phipps, M. J. S.; Clark, H. W.; Skylaris, C. K.;Madsen, J., Surfactant Proteins A and D: Trimerized Innate ImmunityProteins with an Affinity for Viral Fusion Proteins. J. Innate Immun.2019, 11 (1), 13-28.

67. Watanabe, Y. A., J.D.; Wrapp, D.; McLellan, J.S.; Crispin,M., Site-specific glycan analysis of the SARS-CoV-2 spike. Science2020, 10.1126/science.abb9983.

68. Lapitan, L. D. S.; Xu, Y. H.; Guo, Y.; Zhou, D. J., Combin-ing magnetic nanoparticle capture and poly- enzyme nanobead ampli-fication for ultrasensitive detection and discrimination of DNA singlenucleotide polymorphisms. Nanoscale 2019, 11 (3), 1195-1204.

download fileview on ChemRxivGNP-glycan manu.pdf (2.37 MiB)

S1

Supporting Information

Contents

1. Material, instrument and methods

1.1 Materials

Gold nanoparticle (nominal diameter 5 nm, max ~520 nm) was purchased from Sigma Aldrich (UK) as a

stabilized suspension in citrate buffer or synthesized in house. 2-(2-aminoethoxy)-ethanol, Di-tert-butyl

decarbonate, sodium sulfate, sodium hydride (60% dispersion in mineral oil), 3-bromo-1-propyne,

potassium hydroxide, trifluoroacetic acid, triethylamine, sodium bicarbonate, O-(6-Chlorobenzotriazol-

1-yl)-N,N,N’,N’-tetramethyluronium hexafluorophosphate (HCTU), lipoic acid (LA), tris(2-carboxy-

ethyl)phosphine hydrochloride (TCEP.HCl), copper sulfate, sodium ascorbate, Tris[(1-benzyl-1H-1,2,3-

triazol-4-yl) methyl]amine (TBTA), Tris(hydroxymethyl)aminomethane (Tris base), guanidine

hydrochloride, anhydrous DMF and other chemicals were purchased from different suppliers like Sigma

Aldrich, Alfa Aesar, Fluorochem, Thermo Scientific, VWR International and Acros organics with > 99%

impurity and used as received without further purification unless specified elsewhere. All the solvents

were obtained in >99% purity from Fischer Scientific and used as received. When used as reaction

solvents in anhydrous reactions, THF and CH2Cl2 were dried and deoxygenated using an Innovative

Technology Inc. PureSolv® solvent purification system. Ultra-pure water (resistance >18.2 MΩ.cm)

purified by an ELGA Purelab classic UVF system, was used for all experiments and making buffers.

1.2 Instruments and Methods

All moisture-sensitive reactions were performed under a N2 atmosphere using oven-dried glasswares.

Evaporations were performed under reduced pressure on a Bruker rotary evaporator. Lyophilisation of

compounds was performed using a Virtis Benchtop K freeze dryer. The progress of the reactions was

monitored by TLC on commercially available precoated aluminium plates (Merck silica Kieselgel 60

F254) and stained by iodine, sulfuric acid or p-anisaldehyde TLC stains. Flash column chromatography

was performed on silica gel 60 A (Merck grade 9385). Automated flash chromatography was performed

using a Biotage Isolera™ Prime Advanced Flash Purification system with Biotage SNAP KP-Sil

cartridges (50 g or 100 g). The polar lipoic acid-sugar derivatives were purified by size exclusion

chromatography using Biogel P2 column using water as eluent to afford the pure product.

All 1H and 13C NMR spectra were recorded in deuterated solvents on a Bruker Avance DPX300 (500

MHz for 1H, 125 MHz for 13C) except N-(tert-Butyloxycarbonyl)tris[(propargyloxy)methyl]amino-

methane which was recorded on a 400 MHz machine (100 MHz for 13C NMR). All chemical shifts (s)

are quoted in parts per million (ppm) downfield of tetramethylsilane and reference to residual solvent

peaks (CDCl3: 1H = 7.26 ppm, 13C = 77.16 ppm, DMSO-d6: 1H = 2.50 ppm, 13C = 39.52 ppm,

CD3OD: 1H = 3.31 ppm, 13C = 49.15 ppm, D2O: 1H = 4.80 ppm) and the coupling constants (J) are

reported to the nearest 0.1 Hz. Assignment of spectra was based on expected chemical shifts and coupling

constants, aided by COSY, HSQC and HMBC measurements where appropriate. The abbreviations used

in 1H NMR analysis are: s = singlet, br = broad, d = doublet, t = triplet, q = quartet, m = multiplet, dd =

doublet of doublets, dt = doublet of triplets, td = triplet of doublets, dq= doublet of quartets, qt = quartet

of triplets, ddd = doublet of doublet of doublets, dtd = doublet of triplet of doublets.

High resolution mass spectra (HR-MS) were obtained on a Bruker Daltonics MicroTOF mass

spectrometer and m/z values are reported in Daltons to four decimal places. UV-vis absorption spectra

were recorded on a Varian Cary 50 bio UV-Visible Spectrophotometer over 200-800 nm using 1 mL

quartz cuvette with an optical path length of 1 cm or on a Nanodrop 2000 spectrophotometer (Thermo

scientific) over the range of 200-800 nm using 1 drop of the solution with an optical path length of 1 mm.

All centrifugations were carried out on a Thermo Scientific Heraeus Fresco 21 microcentrifuge using 1.5

S2

mL microcentrifuge tubes at room temperature (unless stated otherwise). The GNP purification was

performed in Amicon ultra-S2 centrifugal filter tubes with a cut-off MW of 30,000. Dynamic light

scattering (DLS) was measured on a Zetasizer Nano (Malvern) using disposable PMMA cuvettes. The

hydrodynamic diameters (Dhs) of the GNPs with and without the proteins were measured in binding

buffer (50 mM HEPES, 100 mM NaCl, 10 mM CaCl2, pH 7.8). All fluorescence spectra were measured

on a Cary Eclipse Fluorescence Spectrophotometer using a 0.70 mL quartz cuvette. All measurements

were done in a binding buffer (20 mM HEPES, 100 mM NaCl, 10 mM CaCl2 pH 7.8) containing 1 mg/mL

of bovine serum albumin (BSA) to reduce non-specific adsorption of GNP on the surface of cuvette.

2. Synthesis of the DHLA-EGn-glycan ligands

The -mannose and -mannose--1,2-mannose appending a flexible di(ethylene glycol)-azide linker for

functional glycan ligand preparation was synthesized as per methods reported in literature.1-3 Their 1H

and 13C NMR spectra were shown as follows.

The 1H- (top panel) and 13C- (bottom panel) NMR spectra of -Mannose-EG2-N3 in MeOD

S3

The 1H- (top) and 13C- (bottom panel) NMR spectra of -mannose--1,2-mannose-EG2-N3 in MeOD.

S4

2.1 Synthesis of 5-(1,2-dithiolan-3-yl)-N-(2-(2-(prop-2-yn-1-yloxy)ethoxy)ethyl)pentanamide, LA-EG2-

C≡CH (compound E).

Scheme S1. Synthetic route to LA-EG2-C≡CH linker molecule.

(i) Tert-butyl (2-(2-hydroxyethoxy)ethyl)carbamate (compound B): To a solution of 2-(2-aminoethoxy)-

ethanol (compound A, 5.5 g, 52.2 mmol) in CHCl3 (50 mL) at 0 oC was added a solution of Boc2O (11.4 g,

52.2 mmol) in CHCl3 (50 mL). The reaction mixture was stirred at room temperature (RT) overnight and

then transferred to a separation funnel and added with CHCl3 (50 mL) and H2O (50 mL). The mixture was

extracted twice with CHCl3 (50 mL), dried over Na2SO4, filtered, and concentrated in vacuo to furnish the

product tert-butyl (2-(2-hydroxyethoxy)ethyl)carbamate B (10.7 g, Quant., 52.2 mmol).4

1H NMR (500 MHz, CDCl3) δ: 5.01 (bs, 1H, NH), 3.73 (t, 2H, J = 4.6 Hz, CH2OH), 3.55 (m, 4H, CH2OCH2), 3.32 (m, 2H, NHCH2), 1.43 (s, 9H, t-Bu); 13C NMR (125 MHz, CDCl3) δ: 156.1, 79.4, 72.2, 70.3, 40.3, 28.4. HRMS: calculated m/z for C9H19NO4Na (M+Na)+ 228.1206; found 228.1211.

S5

The 1H- (top panel) and 13C- (bottom panel) NMR spectra of compound B in CDCl3.

(ii) Tert-butyl 2-(2-(prop-2-ynloxy) ethoxy) ethylcarbamate (compound C):

To a solution of B (10.6 g, 51.7 mmol) in anhydrous THF (100 mL) cooled at 0 oC with an ice bath, 60%

NaH in mineral oil (2.48 g, 103.4 mmol) was added. The resulting reaction mixture was stirred for 10 min

at 0 oC. To this, propargyl bromide (6 mL, 61.17 mmol) was added dropwise. The reaction mixture was

allowed to warm to RT and stirred for a further 4 h, when TLC indicated the complete consumption of

the starting materials. The solvent was evaporated and the residue was partitioned between Ethyl

acetate (100 mL) and H2O (100 mL) and the aqueous phase was re-extracted with Ethyl acetate (2 x 50

mL). The combined organic phase was dried over anhydrous sodium sulphate, filtered and concentrated

under vacuum to afford the product as a yellow oil (8.8 g, 70 %, 36.2 mmol).

1H NMR (500 MHz, CDCl3) δ: 4.97 (bs, 1H, NH), 4.21 (d, 2H, J 2.4 Hz, CH2C≡CH), 3.69 (m, 2H, OCH2CH2O),

3.65 (m, 2H, OCH2CH2O), 3.55 (t, 2H, J 5.1 Hz, NHCH2CH2O) 3.32 (m, 2H, NHCH2), 2.45 (t, 1H, J 2.4 Hz,

C≡CH), 1.45 (s, 9H, t-Bu); 13C NMR (125 MHz, CDCl3) δ: 156.0, 79.5, 79.2, 74.7, 70.3, 70.1, 69.1, 58.4,

40.4, 28.4. HRMS: calculated m/z for C12H21NO4Na (M+Na)+ 266.1362; found 266.1369.

S6

The 1H- (top panel) and 13C- (bottom panel) NMR spectra of compound C in CDCl3.

(iii) 2-(2-(prop-2-ynloxy) ethoxy) ethan-1-amine (compound D):

Compound C (5 g, 20.5 mmol) was dissolved in CH2Cl2 (100 mL) into which was added trifluroacetic acid

TFA (10 mL). The reaction mixture was stirred at RT overnight. The solvent was evaporated under

reduced pressure and further co-evaporated with toluene and dried to yield the Compound D5 (2.82 g,

96%, 19.7 mmol) as a viscous oil.

1H NMR (500 MHz, CDCl3) δ: 7.51 (bs, 2H, NH2), 4.18 (d, 2H, J 2.4 Hz, CH2C≡CH), 3.75 (t, 2H, J 5.0 Hz,

NHCH2CH2O), 3.7 (s, 4H, OCH2CH2O), 3.25 (m, 2H, NHCH2), 2.49 (t, 1H, J 2.4 Hz, C≡CH); 13C NMR (125

MHz, CDCl3) δ: 78.7, 75.4, 70.0, 68.9, 66.3, 58.3, 40.0. HRMS: calculated m/z for C7H13NO2Na (M+Na)+

166.0838; found 166.0838.

S7

The 1H- (top panel) and 13C- (bottom panel) NMR spectra of compound D in CDCl3.

(iv) LA-EG2-C≡CH (Compound E):

To the anhydrous CH2Cl2 solution (50 mL) of lipoic acid, LA (2g, 9.7 mmol), the amine D (1.66 g, 11.6

mmol) followed by Et3N (2.7 mL, 19.4 mmol) was added. After 5 min of stirring at RT, HCTU (4.8 g, 11.6

mmol) was added and the reaction mixture was stirred overnight. The reaction mixture was diluted with

CH2Cl2 (40 mL) and washed consecutively with citric acid (50 mL) and sodium bicarbonate (50 mL). The

organic layer was collected, dried over Na2SO4, filtered and concentrated in vacuo. The crude product

was purified by flash chromatography (silica, CHCl3: MeOH = 15:1) to afford the pure compound E as a

yellow oil (2.8g, 72%, 8.4 mmol).

1H NMR (500 MHz, CDCl3) δ: 5.98 (bs, 1H, NH), 4.2 (d, 2H, J 2.4 Hz, CH2C≡CH), 3.71 – 3.68 (m, 2H), 3.66 – 3.63 (m, 2H), 3.59 – 3.53 (m, 3H), 3.45 (q,2H, J = 5.3 Hz), 3.17 (ddd, 1H, J = 11.0, 7.1, 5.4 Hz), 3.11 (dt, 1H, J = 11.0, 6.9 Hz), 2.49 – 2.41 (m, 2H), 2.19 (td, 2H, J = 7.5, 1.3 Hz), 1.90 (dq, 1H, J = 12.7, 6.9 Hz), 1.72

S8

– 1.64 (m, 4H), 1.50 – 1.42 (m, 2H). 13C NMR (125 MHz, CDCl3) δ: 172.7, 79.5, 74.8, 70.1, 69.9, 69.1, 58.5, 56.4, 40.2, 39.1, 38.5, 36.4, 34.7, 28.9, 25.4. HRMS: calculated m/z for C15H25NO3S2Na (M+Na)+ 354.1168; found 354.1176.

The 1H- (top panel) and 13C- (bottom panel) NMR spectra of compound E in CDCl3.

2.2 Synthesis of N-(1,3-bis(prop-2-yn-1-yloxy)-2-((prop-2-yn-1-yloxy)methyl)propan-2-yl)-5-(1,2-dithiolan-3-yl)pentanamide, LA-(EG-C≡CH)3

S9

Scheme S2. Synthetic route to LA-(EG-C≡CH)3 linker molecule.

(i) N-(tert Butyloxycarbonyl)tris(hydroxymethyl) aminomethane (compound G):

Tris base (3.0 g, 24.8 mmol) was added to a magnetically stirred solution of guanidine hydrochloride

(0.35g, 15 mol%) and di-tert-butyl dicarbonate (6.5 g, 29.7 mmol) in EtOH (25 mL), at 40°C and stirred

for 10 minutes. After completion of the reaction, EtOH was evaporated under vacuum and the residue

was dissolved in EtOAc and filtered off to separate out the catalyst and furnish the pure product G

(5.12g, 93.4%, 23.2 mmol).

1H NMR (500 MHz, DMSO-d6) δ 5.76 (bs, 1H, NH), 4.52 (t, 3H, J 5.7 Hz, 3 x OH), 3.51 (d, 6H, J 5.6 Hz, 3 x

CH2OH), 1.36 (s, 9H, C(CH3)3). 13C NMR (125 MHz, DMSO-d6) δ: 155.1 (C=O), 77.8, 60.4(2), 60.3,

28.3(3). HRMS: calculated m/z for C9H19NaNO5 (M+Na)+ 244.1155; found 244.1152.

S10

The 1H-NMR spectrum of BOC-protected tris base in MeOD.

The 13C-NMR spectrum of BOC-protected tris base in MeOD.

(ii) N-(tert-Butyloxycarbonyl)tris[(propargyloxy)methyl] aminomethane (compound H):

To a solution of G (5.1 g, 23.2 mmol) in dry DMF (50 mL), propargyl bromide (80 wt% in toluene) (1.3

mL, 143.5 mmol) was added and the reaction mixture was stirred at 0 °C for 10 min. It was followed by

the addition of finely powdered KOH (8.1 g, 143.5 mmol) in small portions. The entire reaction mixture

was then stirred at room temperature for 40 h when the TLC (n-hexane:EtOAc : 5:1) showed the

generation of a faster moving spot. To the resulting brown coloured mixture, ethyl acetate was added

and stirred for another 10 mins followed by evaporation of the solvents in vacuo. Further, the entire

reaction mixture was dissolved in EtOAc (50 mL) and washed successively with H2O (2 × 30 mL) and brine

(25 mL). The organic layer was collected, dried over anhydrous Na2SO4, filtered and evaporated under

S11

vacuum. The crude material thus obtained was purified by flash chromatography using n-hexane:EtOAc

(5:1) as the eluent to yield the pure compound H (3.5 g, 45%, 10.4 mmol) as a yellowish oil.

1H NMR (400 MHz, CDCl3) δ: 4.9 (br s, 1H, NH), 4.13 (d, 6H, J 2.3 Hz, 3 × CH2CCH), 3.76 (s, 6H, CH2O),

2.41 (t, 3H, J 2.4 Hz, CCH), 1.40 (s, 9H, C(CH3)3). 13C NMR (100 MHz, CDCl3) δ: 154.7 (C=O), 79.6, 74.6,

68.9, 58.6, 58.0, 28.3. HRMS calculated. m/z for C18H25NO5Na (M+Na)+: 358.1360; found: 358.1382.

The 1H-NMR spectrum of N-(tert-Butyloxycarbonyl)tris[(propargyloxy)methyl] aminomethane in CDCl3.

The 13C-NMR spectrum of N-(tert-Butyloxycarbonyl)tris[(propargyloxy)methyl]aminomethane in CDCl3.

(iii) Tris[(propargyloxy)methyl] aminomethane (compound I):

S12

To a magnetic stirred solution of H (2.0 g, 6 mmol) in anhydrous CH2Cl2 (15 mL) at 0 °C, trifluoroacetic

acid (10 mL, 131.3 mmol) was added dropwise over 15 min. The brownish reaction mixture was then

stirred at RT for 2h till the TLC (n-hexane:EtOAc : 1:1) showed the complete conversion of the starting

material into a slower moving spot. The solvents were then evaporated and co-evaporated with toluene.

The resulting residue was dissolved in CH2Cl2 (50 mL) and washed successively with H2O (2 × 30 mL)

followed by saturated aq. NaHCO3 solution (25 mL). The organic layer was collected, dried over

anhydrous Na2SO4 and filtered. The solvent was then evaporated completely in vacuo to obtain the

syrupy compound I (1.2 g, 85.5%, 5.1 mmol).

1H NMR (500 MHz, CDCl3) δ: 4.15 (d, 6H, J 2.4 Hz, 3 × CH2CCH), 3.47 (s, 6H, CH2O), 2.42 (t, 3H, J 2.4 Hz,

CCH), 1.77 (br s, 2H, NH2). 13C NMR (125 MHz, CDCl3) δ: 79.7, 74.4, 72.0, 58.7, 55.6. HRMS calculated

m/z for C13H17O3NNa (M+Na)+: 258.1100, found: 258.1096.

The 1H-NMR spectrum of tris[(propargyloxy)methyl]aminomethane in CDCl3.

S13

The 13C-NMR spectrum of tris[(propargyloxy)methyl]aminomethane in CDCl3.

(iv) LA-(EG-C≡CH)3 (compound J):

The amine D (1.2 g, 5.2 mmol) and Et3N (1.2 mL, 8.7 mmol) were added to an anhydrous CH2Cl2 solution

(20 mL) of lipoic acid (0.9 g, 4.3 mmol). After 5 min of stirring at RT, HCTU (2.2 g, 5.2 mmol) was added

and the reaction mixture was left stirring at RT overnight. The reaction mixture was diluted with CH2Cl2

(40 mL) and washed consecutively with saturated citric acid (50 mL) and saturated sodium bicarbonate

(50 mL) solutions. The organic layer was collected, dried over Na2SO4, filtered and concentrated in

vacuo. The crude product was purified by flash chromatography (n-hexane:EtOAc = 1:1) to afford LA-

(EG-C≡CH)3 (compound J) as yellowish oil (1.8 g, 83.3%, 4.3 mmol).

1H NMR (500 MHz, CDCl3) δ: 5.68 (bs, 1H, NH), 4.14 (d, 6H, J 2.4 Hz, 3 x CH2C≡CH), 3.84 (s, 6H), 3.84 (dq, 1H, J 8.4, 6.4 Hz), 3.17 (ddd, 1H, J 11.1, 7.1, 5.4 Hz), 3.10 (dt, 1H, J = 11.0, 6.9 Hz), 2.55 – 2.35 (m, 4H), 2.16 (t, 2H, J = 7.4), 1.90 (dq, 1H, J = 12.7, 7.0 Hz), 1.69 – 1.61 (m, 4H), 1.53 – 1.39 (m, 2H). 13C NMR (125 MHz, CDCl3) δ: 172.9, 79.5, 74.6, 68.5, 59.1, 58.6, 56.4, 40.2, 38.4, 37.0, 34.6, 28.7, 25.2. HRMS: calculated m/z for C21H28NNa2O4S2 (M+Na2-H)+ 468.1249; found 468.1232.

The 1H-NMR spectrum of lipoic acid-tris[(propargyloxy)methyl]aminomethane (LA-(EG-C≡CH)3 in CDCl3.

S14

The 13C-NMR spectrum of lipoic acid-tris[(propargyloxy)methyl]aminomethane (LA-(EG-C≡CH)3 in CDCl3.

2.3 General protocols for preparing LA-glycan ligands via click chemistry.6

To a 1:1 (v:v) mixed THF: H2O solution (2.0-5.0 mL) containing the glycan (1.1 equiv) and LA-conjugated

linker (1 equiv), was added CuSO4.5H2O (0.036 equiv), TBTA (0.063 equiv) followed by sodium ascorbate

(0.135 equiv) and the resulting solution was stirred at RT. After 3 h, TLC confirmed the consumption of

all starting materials. The solvent was then evaporated and the crude product was purified by size

exclusion chromatography using Biogel P2 column using water as eluent to afford the pure product.

LA-EG2--Mannose (LA-Man):

Yield = (67 %, 0.18 mmol). 1H NMR (500 MHz, CD3OD) δ: 8.04 (s, 1H, triazole-H), 4.8 (d, 1H, J 1.6 Hz, Man

H-1), 4.65 (s, 2H), 4.6 (t, 2H, J 5 Hz), 3.9 (t, 2H, J 5.1 Hz), 3.86-3.81 (m, 3H), 3.8 (dd, 1H, J 1.7 Hz, 3.4 Hz,

Man H-2), 3.72 (d, 1H, J 5.5 Hz), 3.69-3.67 (m, 3H), 3.66 (s, 2H), 3.65-3.63 (m, 2H), 3.62-3.61 (m, 4H),

3.60-3.57 (m, 2H), 3.53 (t, 2H, J 5.5 Hz), 3.39-3.34 (m, 3H), 3.20-3.15 (m, 1H), 3.12-3.07 (m, 1H), 2.49-

2.42 (m, 1H), 2.20 (t, 2H, J 7.4 Hz), 1.92-1.85 (m, 1H), 1.73-1.59 (m, 4H), 1.50-1.40 (m, 2H). 13C NMR (125

MHz, CD3OD) δ: 176.1 (C=O), 145.7 (C=CH), 126.0 (C=CH), 101.7 (Man C-1), 74.6, 72.6, 72.1 (Man C-2),

71.6, 71.5, 71.4, 71.3, 70.8, 70.6, 70.4, 68.6, 67.7, 65.0, 63.0 (Man C-6), 57.6, 51.5, 41.3, 40.3, 39.3, 36.8,

35.7, 29.8, 26.7. HRMS: calculated m/z for C27H49N4O11S2 (M+H)+ 669.2834; found 669.2838.

S15

The 1H-NMR (top) and 13C-NMR (bottom) spectra of LA-EG2--mannose ligand in D2O

LA-EG2--Mannose--1,2-mannose (LA-DiMan):

Yield = (65 %, 0.080 mmol). 1H NMR (500 MHz, CD3OD) δ 8.05 (s, 1H, triazole-H), 5.11 (d, J 1.7 Hz, 1H),

4.96 (d, J 1.8 Hz, 1H), 4.66 (s, 2H), 4.60 (t, 2H, J 5.1 Hz), 3.97 (dd, 1H, J 3.3, 1.8 Hz), 3.91 (dd, 2H, J 5.5,

4.7 Hz), 3.88 – 3.78 (m, 6H), 3.72 – 3.65 (m, 8H), 3.62 (dt, 6H, J 5.0, 1.3 Hz), 3.58 (dt, 3H, J 11.2, 1.6 Hz),

3.53 (t, 2H, J 5.5 Hz), 3.43 – 3.32 (m, 3H), 3.17 (ddd, 1H, J 10.9, 7.1, 5.4 Hz), 3.10 (dt, 1H, J 11.0, 6.9 Hz,

S16

1H), 2.46 (dtd, 1H, J 12.1, 6.7, 5.4 Hz), 2.20 (t, 2H, J 7.4 Hz), 1.89 (dq, 1H, J 12.7, 6.9 Hz), 1.77 – 1.57 (m,

4H), 1.45 (qt, 2H, J 9.4, 5.9 Hz). 13C NMR (125 MHz, CD3OD) δ: 176.1 (C=O), 145.7 (C=CH), 126.0 (C=CH),

104.2 (Man C-1), 100.1 (Man C-1’), 80.6, 75.1, 74.7, 72.4, 72.1, 71.9, 71.6, 71.5, 71.4, 71.3, 70.8, 70.6,

70.5, 69.0, 68.8, 67.9, 65.0, 63.2, 63.1, 57.6, 51.5, 41.3, 40.4, 39.3, 36.8, 35.7, 29.8, 26.7. HRMS:

calculated m/z for C33H58N4O16S2 (M+H)+ 831.3233; found 831.3242.

The 1H- (top) and 13C- (bottom) NMR spectra of LA-EG2--mannose--1,2-mannose ligand in D2O

S17

LA-(Man)3:

Yield = (60 %, 0.060 mmol). 1H NMR (500 MHz, D2O) δ: 8.03 (s, 3H, triazole-H), 4.87 (d, 3H, J 1.8 Hz, H-

1), 4.64 – 4.61 (m, 6H), 4.60 (s, 6H), 3.98 – 3.96 (m, 6H), 3.95 (dd, 3H, J 3.5, 1.8 Hz, H-2), 3.89 (d, 1H, J

2.0 Hz), 3.87 (d, 2H, J 1.7 Hz), 3.84 – 3.79 (m, 6H), 3.77 (d, 2H, J 5.6 Hz), 3.74 (d, 7H, J 3.1 Hz), 3.67 (d,

3H, J 9.8 Hz), 3.65 – 3.61 (m, 25H), 3.61 – 3.53 (m, 3H), 3.21 – 3.12 (m, 2H), 2.42 (dq, 1H, J 12.3, 6.1 Hz),

2.18 (t, 2H, J 7.0 Hz), 1.91 (dt, 1H, J 13.6, 6.8 Hz), 1.71 – 1.64 (m, 1H), 1.58 – 1.48 (m, 4H), 1.33 (q, 3H, J

7.7 Hz). 13C NMR (125 MHz, D2O) δ: 176.5 (C=O), 144.0, 125.3, 99.9 (C-1), 72.7, 70.5, 69.9, 69.5, 69.4,

68.8, 67.5, 66.7, 66.3, 63.5, 60.9, 59.7, 56.5, 50.0, 40.2, 38.0, 35.9, 33.7, 27.5, 24.9. HRMS: calculated

m/z for C57H99N10O28S2 (M+H)+ 1435.6066; found 1435.6096.

S18

The 1H- (top) and 13C-(bottom) NMR spectra of LA-(EG-Man)3 in D2O.

LA-(DiMan)3:

Yield = (69 %, 0.030 mmol). 1H NMR (500 MHz, D2O) δ 8.03 (s, 3H, triazole-H), 5.11 (d, 3H, J 1.8 Hz, H-1),

5.03 (d, 3H, J = 1.9 Hz, H-1’), 4.64 – 4.62 (m, 6H), 4.60 (d, 6H, J 2.9 Hz), 4.08 (dd, 3H, J 3.4, 1.8 Hz, H-2),

3.97 (td, 9H, J = 4.5, 2.5 Hz), 3.92 – 3.68 (m, 40H), 3.66 – 3.61 (m, 29H), 3.58 – 3.53 (m, 3H), 3.21 – 3.13

(m, 2H), 2.42 (dq, 1H, J 12.3, 6.1 Hz), 2.18 (t, 2H, J 7.0 Hz), 1.91 (dd, 1H, J 13.1, 6.8 Hz), 1.67 (m, 1H), 1.59

– 1.49 (m, 4H), 1.32 (p, 3H, J 7.7 Hz). ). 13C NMR (125 MHz, D2O) δ: 179.0 (C=O), 146.5, 127.9, 104.8 (C-

1), 100.9 (C-2), 81.2, 75.8, 75.3, 74.2, 72.8, 72.7, 72.5, 72.2, 72.0 (2), 71.9, 71.3 (2), 70.1, 69.4 (2), 69.0,

66.1, 63.7, 63.4, 62.9, 62.3, 59.0, 52.6, 42.8, 40.6, 38.4, 36.2, 30.1, 27.4. HRMS: calculated m/z for

C75H130N10O43S2 (M+H)2+ 961.3862; found 961.3858.

S19

The 1H- (top) and 13C- (bottom) NMR spectra of LA-(EG-DiMan)3 in D2O.

2.4 TCEP reduction protocol to form DHLA derivatives.7

To a methanol solution of the LA-glycan ligands, 1 molar equivalent of freshly prepared TCEP.HCl (tris(2-

carboxyethyl)phosphine) hydrochloride aqueous solution was added and the resulting solution was left

at RT for ~40 min. After completion of the reduction (as confirmed by HRMS), the solvents were

evaporated and the crude compound was purified by flash chromatography using CHCl3-MeOH as eluent

to yield the corresponding pure DHLA-glycan ligands. The DHLA-glycan ligands were then dissolved in

pure water to make stock ligand solutions and used to prepare GNP-glycan straightaway.

S20

3 Determination of carbohydrate ligand loading on the GNPs.8

The amount of glycan link conjugated on GNP was determined using phenol-sulphuric acid test. To

generate the calibration curve, 80 µl of 5% phenol solution and 400 µl of concentrated H2SO4 were

added to a series of H2O solution (80 µl) containing 2.0-20 µg of LA-EG2-Man or LA-EG2-Diman ligand.

The resulting mixtures were vortexed and allowed to stand at RT for 30 min. Absorbance readings at

490 nm were taken against a blank water control and then were plotted against the ligand amount to

generate calibration curves for LA-EG2-Man and LA-EG2-DiMan below.

Calibration curves showing the absorbance v.s. ligand amount relationship for LA-Man and LA-Diman.

The unconjugated LA-glycan ligands collected from the clear supernatant after GNP conjugation and any

washing through liquids were combined, freeze dried and then re-dissolved in 1.40 mL of pure H2O. For

the study, 10 µL of this solution was diluted with H2O to a final volume of 80 µL. Phenol and H2SO4 were

then added and mixed for 30 min at RT as above and their absorbance at 490 nm were recoded. The

absorbance values were then used to determine the amounts of unconjugated mannose ligand using

the calibration curves above. The dilution factors were taken into account in the calculation of the total

amount of unconjugated ligands. The measurements were done in duplicate for each sample and the

average number of LA-glycan ligands conjugated to each GNP were calculated as ~490, ~690, ~720, and

~650 for LA-Man, LA-DiMan, LA-(Man)3 and LA-(DiMan)3, respectively with standard errors of ~10-15%.

0 5 10 15 20 25 30

0.0

0.2

0.4

0.6

0.8

1.0

LA-EG2-Diman

LA-EG2-Man

Ab

s

Amount of sugar (nmol)

Equation y = a + b*x

Plot Intensity Intensity

Weight Instrumental

Intercept -4.50333E-4 ± 0.00136 -9.10141E-4 ± 0.00289

Slope 0.03415 ± 7.07294E-4 0.01529 ± 4.20513E-4

Residual Sum of Squares 4.31208 19.58001

Pearson's r 0.99893 0.99811

R-Square (COD) 0.99786 0.99623

Adj. R-Square 0.99743 0.99548

S21

Figure S1. Typical Dh size distribution histograms and corresponding Gaussian fits of GNP-Man (A); GNP-

DiMan (B); GNP-(Man)3 (C); GNP-(DiMan)3 (D), wildtype DC-SIGN (E) and DC-SIGNR (F) in a binding buffer

(20 mM HEPES, 100 mM NaCl, 10 mM CaCl2, pH7.8).

Table S1. Summary of the Gaussian fitting parameters of Dh distribution histograms of the proteins and

GNP-glycans in pure water and in binding buffer.

Sample Name Dh (nm) Rh (nm) FWHM (nm)

χ2 R2

GNP in Water 12.5 ± 0.1 6.25 4.7 ± 0.2 1.16 0.99

GNP-Man in Water 11.0 ± 0.1 5.5 3.5 ± 0.1 1.86 0.99

GNP-Man in Buffer 10.8 ± 0.1 5.4 2.6 ± 0.1 1.71 0.99

GNP-DiMan in Water 10.7 ± 0.1 5.35 3.0 ± 0.2 5.10 0.97

GNP-DiMan in Buffer 10.9 ± 0.1 5.45 3.0 ± 0.2 4.58 0.99

GNP-(Man)3 in Water 9.9 ± 0.1 4.95 3.5 ± 0.1 1.45 0.98

GNP-(Man)3 in Buffer 10.4 ± 0.04 5.2 3.2 ± 0.1 0.78 0.99

GNP-(DiMan)3 in Water 10.7 ± 0.1 5.35 4.3 ± 0.1 5.10 0.98

GNP-(DiMan)3 in Buffer 10.9 ± 0.1 5.45 2.5 ± 0.1 4.58 0.99

WT DC-SIGN in Buffer 14.6 ± 0.02 7.3 3.4 ± 0.1 0.12 0.98

WT DC-SIGNR in Buffer 11.3 ± 0.02 5.65 3.8 ± 0.1 0.08 0.99

10 1000

10

20

30

40

50P

op

ula

tio

n

Dh (nm)

Model Gauss

Equationy=y0 + (A/(w*sqrt(pi/2)))*exp(-2

*((x-xc)/w)^2)Plot Volume Population

y0 -0.39349 ± 0.84626

xc 9.38965 ± 0.05948

w 2.55057 ± 0.14349

A 139.22934 ± 8.18663

Reduced Chi-Sqr3.34986

R-Square (COD)0.99043

Adj. R-Square 0.98565

GNP-Man in Buffer

10 1000

10

20

30

40

50

Po

pu

lati

on

Dh (nm)

Model Gauss

Equationy=y0 + (A/(w*sqrt(pi/2)))*exp(

-2*((x-xc)/w)^2)Plot Volume opulation

y0 -0.0099 ± 0.17678

xc 10.42945 ± 0.03844

w 3.15761 ± 0.0798

A 149.20647 ± 3.4862

Reduced Chi-Sqr0.78312

R-Square (COD)0.99091

Adj. R-Square 0.9899

GNP-(Man)3 in Buffer

10 1000

10

20

30

40

50

Po

pu

lati

on

Dh (nm)

Model Gauss

Equationy=y0 + (A/(w*sqrt(pi/2)))*exp(-

2*((x-xc)/w)^2)Plot Volume Population

y0 -0.60202 ± 1.29016

xc 10.91162 ± 0.08065

w 2.99556 ± 0.20915

A 163.16617 ± 12.89205

Reduced Chi-Sqr4.57594

R-Square (COD)0.98999

Adj. R-Square 0.98249

GNP-DiMan in Buffer

10 1000

10

20

30

40

50

Po

pu

lati

on

Dh (nm)

Model Gauss

Equationy=y0 + (A/(w*sqrt(pi/2)))*exp(-2*((x-xc)/w)

2̂)

Plot Volume Population

y0 -0.34526 ± 0.4727

xc 10.20908 ± 0.0394

w 2.54358 ± 0.08072

A 151.88702 ± 5.07381

Reduced Chi-Sqr 1.64564

R-Square (COD) 0.99486

Adj. R-Square 0.99294

GNP-(DiMan)3 in Buffer

10 1000

5

10

15

20

Po

pu

lati

on

Dh (nm)

Model Gauss

Equationy=y0 + (A/(w*sqrt(pi/2)))*exp

(-2*((x-xc)/w)^2)

Plot Volume Population

y0 -0.00827 ± 0.03135

xc 14.62377 ± 0.02552

w 3.43231 ± 0.05194

A 67.21818 ± 0.91198

Reduced Chi-Sqr0.12185

R-Square (COD)0.98451

Adj. R-Square 0.98416

DC-SIGN

10 1000

5

10

15

20

Po

pu

lati

on

Dh (nm)

Model Gauss

Equationy=y0 + (A/(w*sqrt(pi/2)))*exp(-2*

((x-xc)/w) 2̂)

Plot DC-SIGNR

y0 0.01748 ± 0.14107

xc 12.19131 ± 0.14231

w 6.67937 ± 0.2762

A 159.9888 ± 6.36558

Reduced Chi-Sqr 1.08652

R-Square (COD) 0.94071

Adj. R-Square 0.93797

DC-SIGNR

(A) (B) (E)

(C) (D) (F)

S22

Table S2. Summary of chemical and physical parameters and the calculated inter-glycan

distance (X) for each GNP-glycan.

Sample Name Dh (nm) Rh (nm) N Θ (°) X (nm)

GNP-Man 10.8 5.4 491 10.35° 0.97 nm

GNP-DiMan 10.9 5.45 688 8.74° 0.83 nm

GNP-(Man)3 10.2 5.1 2160 4.93° 0.46 nm

GNP-(DiMan)3 10.9 5.45 1965 5.17° 0.49 nm

Calculation of the average inter-glycan distance (X) on the GNP8-9

For a GNP-glycan containing N glycans with a radius of r (equals its hydrodynamic radius Rh = ½ Dh, see

Table S1), the footprint of each glycan on the GNP surface is:

𝑘 =4𝜋𝑟2

𝑁 (1)

The average deflection angle (θ, in degrees) of each glycan on the GNP surface is measured as:

𝜃 =360 √

𝑘

𝜋

𝑟 𝜋 =

229.3

√𝑁 (2)

The inter-glycan distance on the GNP surface (X) can be calculated via

𝑋 = 2 𝑟 sin(𝜃

2) (3)

Using equations (2) and (3) and the number of glycans (N) conjugated on each GNP, the inter glycan

distance (X) for each GNP-glycan was calculated and shown in Table S2.

S23

Figure S2. HRMS spectra of wildtype DC-SIGN (DC020, top left), DC-SIGNR (DSR034, top right) and DC-

SIGN side-specifically labeled with atto594 maleimide (DC020 Q274C-atto594, bottom left) and DC-

SIGNR (DSR034 R287C-atto594, bottom right) proteins. An increase of MW of 854 Da was observed for

both labeled proteins, corresponding to the labeled atto594 dye weight.

S24

Figure S3. (A) Fluorescence spectra (EX = 595 nm) of different concentration DC-SIGN (atto594 labeled)

in the absence (solid lines) and presence of the same concentration GNP-DiMan (broken lines) in a

binding buffer. (B) Plot of integrated fluorescence intensity versus DC-SIGN concentration in the absence

(black squares) and presence of the same concentration GNP-DiMan (red dots). The former was fitted

very well by linear relationship (green line; Y = (-77±113) + (401±3.4) × X, R2 = 0.9997) but the later

deviated further and further away from linear (dashed blue linear) with the increasing concentration,

indicating increased fluorescence quenching due to increasingly higher fractions of DC-SIGN were bound

to GNP-glycan and got quenched.

0 10 20 30 40 50 60 70

0

1x104

2x104

3x104 DC-SIGN only

DC-SIGN + GNP-DiMan

Linear fit of DC-SIGN

IF (

a.u

.)

C (nM)

Equation y = a + b*x

Plot I

Weight No Weighting

Intercept -76.9 ± 113.05833

Slope 401.1875 ± 3.42758

Residual Sum of Squares 90227.2

Pearson's r 0.99989

R-Square (COD) 0.99978

Adj. R-Square 0.99971

600 630 660 690 720 750

0

200

400

600

I (a

.u.)

(nm)

DC-SIGN 8nM

DC-SIGN 16nM

DC-SIGN 32nM

DC-SIGN 64nM

DC-SIGN+GNP 8nM

DC-SIGN+GNP 16nM

DC-SIGN+GNP 32nM

DC-SIGN+GNP 64nM

(A) (B)

S25

Fig. S4: Fluorescence spectra of GNP-(DiMan)3 (20 nM) mixed with 80 nM labeled DC-SIGN (A) or DC-

SIGNR (B) as a function of the added free mannose concentration. Fluorescence recovery ratio (I/I0)

versus free mannose concentration for competing against GNP-(DiMan)3 binding with DC-SIGN (C) or

DC-SIGNR (D) fitted by Hill’s equation: Y = a + (b-a)*Xn/(kn + Xn).

The fitting parameters are: a = 1.16 ± 0.09; b = 4.04 ± 0.13; k = 4.16 ± 0.41 mM; n =1.32 ± 0.18, R2 = 0.995

for DC-SIGN binding, and a = 1.02 ± 0.14; b = 1.22 ± 0.07; k = 1.1 ± 2.0 mM; n =0.76 ± 1.22, R2 = 0.846 for

DC-SIGNR. The maximal fluorescence recovery ratio (b) and mannose concentration to give 50%

fluorescence recovery for DC-SIGN (k) are about 4 fold that of DC-SIGNR, suggesting DC-SIGN binding

with GNP-(DiMan)3 is much stronger than DC-SIGNR, in good agreement with their apparent Kd data (see

Table 1).

650 700 7500

200

400

600

800

I (a

.u.)

(nm)

DCSIGN 80 nM

Ditri 20 nM+ DCSIGN 80 nM A

Ditri 20 nM+ DCSIGN 80 nM+ 0.1 mM Man A

Ditri 20 nM+ DCSIGN 80 nM+ 1 mM Man A

Ditri 20 nM+ DCSIGN 80 nM+ 3 mM Man B

Ditri 20 nM+ DCSIGN 80 nM+ 5 mM Man A

Ditri 20 nM+ DCSIGN 80 nM+ 10 mM Man A

Ditri 20 nM+ DCSIGN 80 nM+ 20 mM Man A

Ditri 20 nM+ DCSIGN 80 nM+ 40 mM Man A

Ditri 20 nM+ DCSIGN 80 nM+ 40 mM Gal A

650 700 7500

200

400

600

800

I (a

.u.)

(nm)

DCSIGNR 80 nM A

Ditri 20 nM+ DCSIGNR 80 nM A

Ditri 20 nM+ DCSIGNR 80 nM + 0.1 mM Man A

Ditri 20 nM+ DCSIGNR 80 nM + 1 mM Man A

Ditri 20 nM+ DCSIGNR 80 nM + 3 mM Man A

Ditri 20 nM+ DCSIGNR 80 nM + 5 mM Man A

Ditri 20 nM+ DCSIGNR 80 nM + 10 mM Man B

Ditri 20 nM+ DCSIGNR 80 nM + 20 mM Man A

Ditri 20 nM+ DCSIGNR 80 nM + 40 mM Man A

Ditri 20 nM+ DCSIGNR 80 nM + 40 mM Gal A

0 10 20 30 40

1

2

3

4

I/I 0

C (mM)

Model Hill1

Equationy = START + (END - START) * x^n

/ (k^n + x^n)

Plot I/I0

START 1.15891 ± 0.08865

END 4.04392 ± 0.12674

k 4.15705 ± 0.41985

n 1.31761 ± 0.18337

Reduced Chi-Sqr 0.00687

R-Square (COD) 0.99702

Adj. R-Square 0.99478

0 10 20 30 40

1.0

1.1

1.2

I/I 0

C (mM)

Model Hill1

Equationy = START + (END - START) * x̂ n / (k n̂ + x̂ n

)

Plot I/I0

START 1.0185 ± 0.13654

END 1.21809 ± 0.07244

k 1.07625 ± 1.95166

n 0.75663 ± 1.22055

Reduced Chi-Sqr 9.54475E-4

R-Square (COD) 0.91175

Adj. R-Square 0.84556

(A) (C)

(B) (D)

S26

Figure S5. Dh distribution histograms for GNP-Man (10 nM) binding to DC-SIGN at 0 (A); 20 (B); 40 (C);

60 (D); 80 (E); 100 (F); 150 (G) or 200 nM (H). For samples showing 2 Gaussian species, a linear average

of the two (based on area%) was used to calculate the mean Dh, e.g. mean Dh = Dh1 x A1% + Dh2 x A2%.

Figure S6. Dh distribution histograms for GNP-DiMan (10 nM) binding to wildtype DC-SIGN at 0 (A); 20

(B); 40 (C); 60 (D); 80 (E); 100 (F); 150 (G) or 200 nM (H). For samples showing 2 species, a linear average

of the two (based on area%) was used to calculate the mean Dh, e.g. mean Dh = Dh1 x A1% + Dh2 x A2%.

0 50 100 150 2000

10

20

30

40P

op

ula

tio

n

Dh (nm)0 50 100 150 200

0

10

20

30

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

2

4

6

8

10

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

5

10

15

Po

pu

lati

on

Dh (nm)

CDC-SIGN = 0 nMDh =10.8±0.1 nmFWHM = 4.3±0.1 nmR2 = 0.989

CDC-SIGN = 20 nMDh =13.2±0.04 nmFWHM =5.8±0.1 nmR2 = 0.992

CDC-SIGN = 40 nMDh =14.8±0.1 nmFWHM =6.5±0.2 nmR2 = 0.976

CDC-SIGN = 60 nMDh =18.7±0.1 nmFWHM =8.6±0.3 nmR2 = 0.979

(A) (B) (C) (D)

0 50 100 150 2000

5

10

15

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

6

12

18

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

5

10

15

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

5

10

15

20

Po

pu

lati

on

Dh (nm)

CDC-SIGN = 100 nMDh1 = 16.5±0.2 nm (44.2%)FWHM1 = 7.2±0.5 nmDh2 = 27.3±1.3 nm (54.8%)FWHM2 = 14.7±2.1 nmR2 = 0.988

CDC-SIGN = 200 nMDh1 = 25.8±0.5 nm (68.4%)FWHM1 = 13.9±1.4 nmDh2 = 43.4±10 nm (31.6%)FWHM2 = 26±13 nmR2 = 0.979

CDC-SIGN = 150 nMDh = 24.3±0.1 nmFWHM1 = 12.9±0.2 nmR2 = 0.995

CDC-SIGN = 80 nMDh = 22.3±0.2 nmFWHM = 12.2±0.5 nmR2 = 0.965

(E) (F) (G) (H)

0 50 100 150 2000

4

8

12

16

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

4

8

12

16

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

10

20

30

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

4

8

12

16

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

5

10

15

20

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

5

10

15

20

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

5

10

15

20

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

5

10

15

20

Po

pu

lati

on

Dh (nm)

CDC-SIGN = 200 nMDh1 = 38.3±2.5 nm (75.1%)FWHM1 =20.7±2.2 nmDh2 =54.9±27.7 nm (24.9%)FWHM2 =26.8±21.7 nmR2 = 0.993

CDC-SIGN = 150 nMDh1 =42.5±0.7 nm (48.3%)FWHM1 =13.0±1.5 nmDh2 =54.1±5.1 nm (51.7%)FWHM2 =19.8±4.3 nmR2 = 0.997

CDC-SIGN = 100 nMDh =40.0±0.3 nmFWHM =22.5±0.7 nmR2 = 0.979

CDC-SIGN = 80 nMDh =35.8±0.3 nmFWHM =20.0±0.6 nmR2 = 0.978

CDC-SIGN = 0 nMDh =11.1±0.06 nmFWHM =5.1±0.1 nmR2 = 0.985

CDC-SIGN = 20 nMDh =26.7±0.3 nmFWHM =15.6±0.6 nmR2 = 0.963

CDC-SIGN = 60 nMDh =35.5±0.2 nmFWHM =16.5±0.4 nmR2 = 0.978

CDC-SIGN = 40 nMDh1 =22.3±0.2 nm (65.8%)FWHM1 =10.2±0.4 nmDh2 =47.4±2.1 nm (34.2%)FWHM2 =24.8±5.3 nmR2 = 0.980

(A) (B) (C) (D)

(E) (F) (G) (H)

S27

Figure S7. Dh distribution histograms for GNP-(Man)3 (10 nM) binding to DC-SIGNR at 0 (A); 20 (B); 40

(C); 60 (D); 80 (E); 100 (F); 150 (G) or 200 nM (H).

Figure S8. Dh distribution histograms for GNP-(DiMan)3 (10 nM) binding to DC-SIGNR at 0 (A); 20 (B); 40

(C); 60 (D); 80 (E); 100 (F); 150 (G) or 200 nM (H). For samples showing 2 Gaussian species, a linear

average of the two species was used to calculate the mean Dh, e.g. mean Dh = Dh1 x A1% + Dh2 x A2%.

CDC-SIGNR = 200 nMDh = 11.4 ± 0.04 nmFWHM: 5.3 ± 0.1 nmR2 = 0.992

0 20 40 60 800

10

20

30

40

Po

pu

lati

on

Dh (nm)0 20 40 60 80

0

10

20

30P

op

ula

tio

n

Dh (nm)0 20 40 60 80

0

10

20

30

Po

pu

lati

on

Dh (nm)0 20 40 60 80

0

10

20

30

Po

pu

lati

on

Dh (nm)

CDC-SIGNR = 150 nM Dh = 11.0 ± 0.04 nmFWHM= 4.7 ± 0.1 nmR2 = 0.989

CDC-SIGNR = 100 nMDh = 10.9 ± 0.04 nmFWHM=5.0 ± 0.1 nmR2 = 0.992

CDC-SIGNR = 80 nMDh = 11.5 ± 0.05 nmFWHM=4.8 ± 0.1 nmR2 = 0.987

0 20 40 60 800

10

20

30

40P

op

ula

tio

n

Dh (nm)

0 20 40 60 800

10

20

30

Po

pu

lati

on

Dh (nm)0 20 40 60 80

0

10

20

30

Po

pu

lati

on

Dh (nm)

CDC-SIGNR = 60 nMDh = 11.4 ± 0.04 nmFWHM=4.9 ± 0.1 nmR2 = 0.991

CDC-SIGNR = 20 nMDh = 10.7 ± 0.04 nmFWHM=5.2 ± 0.1 nmR2 = 0.987

CDC-SIGNR = 0 nMDh = 10.8 ± 0.04 nmFWHM=4.3 ± 0.1 nmR2 = 0.988

CDC-SIGNR = 40 nMDh = 10.7 ± 0.04 nmFWHM=4.2 ± 0.1 nmR2 = 0.991

0 20 40 60 800

10

20

30

40

Po

pu

lati

on

Dh (nm)

(A) (B) (C) (D)

(E) (F) (G) (H)

0 25 50 75 1000

10

20

30

Po

pu

lati

on

Dh (nm)

0 25 50 75 1000

10

20

30

Po

pu

lati

on

Dh (nm)

0 200 400 600 8000

5

10

15

20

Po

pu

lati

on

Dh (nm)

0 50 100 150 2000

10

20

30

Po

pu

lati

on

Dh (nm)

0 25 50 75 1000

8

16

24

Po

pu

lati

on

Dh (nm)0 25 50 75 100

0

10

20

30

Po

pu

lati

on

Dh (nm)

0 25 50 75 1000

5

10

15

20

25

Po

pu

lati

on

Dh (nm)

0 25 50 75 1000

10

20

30

Po

pu

lati

on

Dh (nm)

CDC-SIGNR = 0 nMDh = 11.1±0.1 nmFWHM=5.1±0.1 nmR2 = 0.985

CDC-SIGNR = 20 nMDh = 10.7±0.1 nmFWHM=5.2±0.1 nmR2 = 0.988

CDC-SIGNR = 40 nMDh = 11.3±0.1 nmFWHM=4.8±0.1 nmR2 = 0.988

CDC-SIGNR = 60 nMDh1 = 7.9±0.1 nm (8.2%)FWHM1=2.3±0.3 nmDh2 =17.6±0.1 nm (91.8%)FWHM2=6.7±0.2 nmR2 = 0.986

CDC-SIGNR = 60 nMDh1 = 74.3±0.9 nm (41.3%)FWHM1=36±2 nmDh2 =209±12 nm (58.7%)FWHM2=175±31 nmR2 = 0.948

CDC-SIGNR = 150 nMDh1 = 11.9±0.1 nm (68.1%)FWHM1 = 4.2±0.1 nmDh2 = 78±19 nm (31.9%)FWHM2 = 70±57 nmR2 = 0.986

CDC-SIGNR = 100 nMDh = 13.4±0.1 nmFWHM=7.6±0.1 nmR2 = 0.993

CDC-SIGNR = 80 nMDh = 11.7±0.04 nmFWHM=5.7±0.1 nmR2 = 0.994

(A) (B) (C) (D)

(E) (F) (G) (H)

S28

6.2 TEM imaging

Figure S9. TEM images of cryo-prepared GNP-DiMan (10 nM) samples in the absence (left, higher

magnification image inset) and presence of wildtype DC-SIGN (150 nM, middle) or DC-SIGNR (640 nM

protein with 16 nM GNP-DiMan, right).

S29

Figure S10. Comparison of cellular luciferase activities of 293T cells after being trans-infected with the

indicated plasmids, pre-incubated with varied concentrations of GNP-Man (A, B), GNP-(Man)3 (C, D);

GNP-DiMan (E, F) or GNP-(DiMan)3 (G, H) and inoculated with MLV reporter particles bearing the EBOV-

GP (A, C, E, G) or the control VSV-G (B, D, F, H). The indicated GNP concentrations were attained after

addition of particles. Luciferase activities in cell lysates were measured at 72 h post-transduction. The

results of single representative experiments performed with quadruplicate samples are shown. Error

bars indicate the standard deviation. Similar results were obtained in three separate experiments.

1000

10000

100000

1000000

10000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-(Man)3 + MLV-EBOV GP Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

1000

10000

100000

1000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-(Man)3 + MLV-VSV-G Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

1000

10000

100000

1000000

10000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-Man + MLV-EBOV GP Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

1000

10000

100000

1000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-Man + MLV-VSV-G Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

(A)

(B)

(C)

(D)

1000

10000

100000

1000000

10000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-DiMan + MLV-EBOV GP Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

1000

10000

100000

1000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-DiMan + MLV-VSV-G Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

1000

10000

100000

1000000

10000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-(DiMan)3 + MLV-EBOV GP Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

1000

10000

100000

1000000

pcDNA DC-SIGN DC-SIGN R

Lum

ine

sce

nce

GNP-(DiMan)3 + MLV-VSV-G Pseudotype

0nM

0,1nM

0,3nM

1nM

3nM

10nM

30nM

50nM

(E)

(F)

(G)

(H)

S30

Figure S11. Plot of normalized luciferase activities of DC-SIGN- (displaying in black squares) or DC-SIGNR-

(displaying in red dots) expressing 293T cells as a function of the concentration of GNP-Man (A) or GNP-

(Man)3 (B) inhibitors. Data were fitted by a general formula of modified competition model, Y =

EC50n/(EC50

n + Cn), where n =1 for DC-SIGN and n = 0.4-0.6 for DC-SIGNR.

The best fitting parameters were given follows:

GNP-glycan Lectin Receptor IC/EC50 (nM) n R2 GNP-Man DC-SIGN 0.26±0.08 1 0.889 GNP-(Man)3 DC-SIGN 1.57±0.25 1 0.937 GNP-Man DC-SIGNR 7.3±1.2* 0.42±0.04 0.980 GNP-(Man)3 DC-SIGNR 53±17* 0.61± 0.14 0.902

References

1. Guo, Y.; Sakonsinsiri, C.; Nehlmeier, I.; Fascione, M. A.; Zhang, H.; Wang, W.; Pohlmann, S.; Turnbull, W. B.; Zhou, D., Compact, Polyvalent Mannose Quantum Dots as Sensitive, Ratiometric FRET Probes for Multivalent Protein-Ligand Interactions. Angew. Chem. Int. Ed. 2016, 55 (15), 4738-4742.

2. Guo, Y.; Bruce Turnbull, W.; Zhou, D., Probing Multivalent Protein-Carbohydrate Interactions by Quantum Dot-Forster Resonance Energy Transfer. Methods Enzymol. 2018, 598, 71-100.

3. Ramos-Soriano, J.; Reina, J. J.; Illescas, B. M.; de la Cruz, N.; Rodriguez-Perez, L.; Lasala, F.; Rojo, J.; Delgado, R.; Martin, N., Synthesis of Highly Efficient Multivalent Disaccharide/ 60 Fullerene Nanoballs for Emergent Viruses. J. Am. Chem. Soc. 2019, 141 (38), 15403-15412.

4. Greenwald, R. B.; Choe, Y. H.; Conover, C. D.; Shum, K.; Wu, D.; Royzen, M., Drug delivery systems based on trimethyl lock lactonization: poly (ethylene glycol) prodrugs of amino-containing compounds. J. Med. Chem. 2000, 43 (3), 475-487.

0 10 20 30 40 50

0.0

0.2

0.4

0.6

0.8

1.0

1.2

DC-SIGN

DC-SIGNR

DC-SIGN Fit

DC-SIGNR Fit

No

rmalized

Acti

vit

y

C/nM

GNP-Man

0 10 20 30 40 50

0.0

0.2

0.4

0.6

0.8

1.0

1.2

DC-SIGN

DC-SIGNR

DC-SIGN Fit

DC-SIGNR Fit

No

rma

lize

d A

cti

vit

y

C/nM

GNP-(Man)3(A) (B)

S31

5. Richards, S. J.; Jones, M. W.; Hunaban, M.; Haddleton, D. M.; Gibson, M. I., Probing Bacterial‐Toxin Inhibition with Synthetic Glycopolymers Prepared by Tandem Post‐Polymerization Modification: Role of Linker Length and Carbohydrate Density. Angew. Chem. Int. Ed. 2012, 51 (31), 7812-7816.

6. Ribeiro-Viana, R.; Sanchez-Navarro, M.; Luczkowiak, J.; Koeppe, J. R.; Delgado, R.; Rojo, J.; Davis, B. G., Virus-like glycodendrinanoparticles displaying quasi-equivalent nested polyvalency upon glycoprotein platforms potently block viral infection. Nat. Commun. 2012, 3, article no: 1303.

7. Burns, J. A.; Butler, J. C.; Moran, J.; Whitesides, G. M., Selective Reduction of Disulfides by Tris(2-Carboxyethyl)phosphine. J. Org. Chem. 1991, 56 (8), 2648-2650.

8. Guo, Y.; Nehlmeier, I.; Poole, E.; Sakonsinsiri, C.; Hondow, N.; Brown, A.; Li, Q.; Li, S.; Whitworth, J.; Li, Z.; Yu, A.; Brydson, R.; Turnbull, W. B.; Pohlmann, S.; Zhou, D., Dissecting Multivalent Lectin-Carbohydrate Recognition Using Polyvalent Multifunctional Glycan-Quantum Dots. J. Am. Chem. Soc. 2017, 139 (34), 11833-11844.

9. Hill, H. D.; Millstone, J. E.; Banholzer, M. J.; Mirkin, C. A., The Role Radius of Curvature Plays in Thiolated Oligonucleotide Loading on Gold Nanoparticles. ACS Nano 2009, 3 (2), 418-424.

download fileview on ChemRxivSupporting Information.pdf (2.29 MiB)