Single‐Molecule Fluorescence Detection of a Synthetic Heparan Sulfate Disaccharide

Abstract The first single‐molecule fluorescence detection of a structurally‐defined synthetic carbohydrate is reported: a heparan sulfate (HS) disaccharide fragment labeled with Alexa488. Single molecules have been measured whilst freely diffusing in solution and controlled encapsulation in surface‐tethered lipid vesicles has allowed extended observations of carbohydrate molecules down to the single‐molecule level. The diverse and dynamic nature of HS–protein interactions means that new tools to investigate pure HS fragments at the molecular level would significantly enhance our understanding of HS. This work is a proof‐of‐principle demonstration of the feasibility of single‐molecule studies of synthetic carbohydrates which offers a new approach to the study of pure glycosaminoglycan (GAG) fragments.

This crude product was dissolved in anhydrous pyridine (6 mL), sulfur trioxide-pyridine complex was added (277 mg, 1.75 mmol) and the reaction stirred at RT for 20 h. TLC (92.5:7.5 CH 2 Cl 2 /MeOH) showed no remaining disaccharide starting material. The solvent was removed in vacuo to give a white solid. Column chromatography (92.5:7.5 CH 2 Cl 2 /MeOH) afforded the title compound 3 as a yellow foam (242 mg, 0.24 mmol, 77% over two steps). Sulfated product 3 was determined to be approx. 95% pure by NMR, but broadening of the peaks prevented assignment, and the material was used as is in the next step. Data collected for 3: R f 0.18 To a stirred solution of disaccharide 3 (240 mg, 0.24 mmol) in THF (5 mL

Steady-state ensemble spectroscopy
Absorption and emission spectra were acquired on Cary 60 (Agilent Technologies) and Fluoromax (HORIBA Scientific) spectrometers, respectively. Emission spectra were recorded under magic angle conditions and background fluorescence from the solvent was negligible.
The sequence of the labeled ssDNA discussed in the manuscript (see Fig. 1

Fluorescence Correlation Spectroscopy
FCS measurements were performed on the same confocal microscope used for MFD measurements (see below). Samples were prepared in Nunc Lab-Tek II chambered cover glasses connected to a PC and cross-correlated. 5 million photons were typically collected for each correlation measurement with count rates of ca. 100 kHz. All measurements are reported for a temperature of 21 ± 1 °C. Cross-correlation functions, G(t), were fitted to equation 1.

(Equation 1)
where C is a constant, t is the lag time, N is the number of molecules in the confocal volume, t D is the translational diffusion time, V is a measure of the detection volume defined as z 0 /w 0 , where z 0 and w 0 are the distances at which the 3D Gaussian volume has decayed to 1/e 2 in the axial and radial directions, respectively, t f is the triplet fraction and t s is the triplet lifetime [1].
The translational diffusion time is subsequently related to the diffusion coefficient, D, via equation 2.

(Equation 2)
The hydrodynamic radius of the diffusing molecule, R H , may then be found via the Stokes-

(Equation 3)
where k B is Boltzmann's constant, T is the temperature of the medium and h is the microviscosity of the medium.
Fitting parameters associated with the cross-correlation curves of freely-diffusing Rhodamine 110, Alexa 488 (free dye) and labelled disaccharide 6 molecules in 20 mM Tris, 10 mM MgCl 2 , pH 7.5 buffer are shown in Table S1.  All measurements were recorded in 20 mM Tris, 10 mM MgCl 2 , 1mM vitamin C, pH 7.5

Multi-parameter confocal fluorescence spectroscopy
buffers at 21 ± 1°C. Measurement buffers were cleaned using activated charcoal prior to use.
The lifetime (t) and steady-state fluorescence anisotropy (r) parameters for MFD were calculated using software written by the group of Prof. Claus Seidel (Heinrich Heine Universität, Düsseldorf as described in detail elsewhere [4]. Sub-ensemble analysis also used software from the Seidel group and involved combining the data from all of the molecules within a specific region of a 2D MFD plot to create a sub-ensemble dataset. In Fig. 1e, the time-resolved decay for the total signal in the green channels (constructed by combining parallel and perpendicular data) for the sub-population indicated was analyzed using DAS6 software from HORIBA Jobin Yvon. Quality of fit or residuals were not improved by fitting to a more complicated decay function. Sample dilutions were prepared in Nunc Lab-Tek II chambered cover glasses, #1,5 (Thermo Fisher, UK).

Encapsulation of Alexa488-labelled Disaccharide in Lipid Vesicles
Small unilamellar vesicles were prepared by the extrusion method [5,6]. trajectories were viewed and analyzed using MATLAB procedures written in-house. To extract the true number of stepwise photobleaching events, we employed a 1D edge detection algorithm as previously described [8]. Briefly, the algorithm constructs a series of scaled derivatives from each raw intensity trajectory before identifying local maxima and minima in the scaled derivatives to identify the presence of sharp edges (bleaching events). By tracking the number of sharp edges identified per trace prior to the fluorescence signal reaching the background intensity, the number of significant bleaching events occurring in the data was evaluated. Figure S1. HRMS isotope pattern matching for labelled disaccharide 6 (EPSRC National Mass Spectrometry Facility, Swansea).