Conformationally Constrained Sialyl Analogues as New Potential Binders of h‐CD22

Abstract Here, two conformationally constrained sialyl analogues were synthesized and characterized in their interaction with the inhibitory Siglec, human CD22 (h‐CD22). An orthogonal approach, including biophysical assays (SPR and fluorescence), ligand‐based NMR techniques, and molecular modelling, was employed to disentangle the interaction mechanisms at a molecular level. The results showed that the Sialyl‐TnThr antigen analogue represents a promising scaffold for the design of novel h‐CD22 inhibitors. Our findings also suggest that the introduction of a biphenyl moiety at position 9 of the sialic acid hampers canonical accommodation of the ligand in the protein binding pocket, even though the affinity with respect to the natural ligand is increased. Our results address the search for novel modifications of the Neu5Ac‐α(2‐6)‐Gal epitope, outline new insights for the design and synthesis of high‐affinity h‐CD22 ligands, and offer novel prospects for therapeutic intervention to prevent autoimmune diseases and B‐cell malignancies.


Protein expression and purification
The plasmids encoding for the three N-terminal Ig-like domains of human CD22 fused to the Fc region of mouse IgG2b was expressed in CHO cell lines and purified as described elsewhere. [1] Synthesis and characterization of chemical materials ESI-MS analyses were performed in negative ion mode and were recorded on an LCQ-Fleet Ion Trap equipped with a standard Ionspray interface. HRMS were performed on a Triple-TOF with a resolution of 35000 (FWHM). Chemical shifts are reported in part per million (δ) using the residual solvent line as secondary internal reference. 1 H NMR spectra were obtained at 500 MHz and 700 MHz, chemical shifts are reported in δ. 13  (7 x 100 mL) and dried in vacuo, to afford 6.23 g of crude as a white solid. To a solution of crude (2.00 g, 5.20 mmol) in 20 mL of pyridine, cooled to 0 °C, TsCl (6.00 mL, 7.80 mmol) was added and the solution was stirred at room temperature for 18h, then, Ac2O (16 mL) and a catalytic amount of DMAP were added, and the solution was stirred for 5 h at room temperature. The reaction mixture was diluted with 200 mL of AcOEt and the organic layer washed with 1 M HCl and then with water. The organic phase was then collected, dried over anhydrous Na2SO4, filtered and evaporated, to afford 5.02 g of crude as a yellow oil. Crude was purified by flash-cromathography on silica gel (petroleum ether 10% in AcOEt) to afford 3 as a yellow amorphous solid (1.22 g, 40%). 1  Synthesis of compound 4. To a solution of 3 in 15 mL of anhydrous DMF, NaN3 (600 mg, 9.36 mmol) was added and the mixture was heated to 70 °C and stirred for 5h. Then, a second amount of NaN3 (300 mg, 3.18 mmol) was added and the mixture was stirred at 70 °C for 1 h. The reaction mixture was diluted with 400 mL of AcOEt and the organic phase washed with brine. The organic layer was collected, dried over anhydrous Na2SO4, filtered and evaporated, to afford 1.10 g of crude as a brown solid. The crude was purified by flash-chromatography on silica gel (AcOEt) to afford pure 4 as an amorphous solid (450 mg, 50%). 1   Synthesis of compound 7. Under nitrogen atmosphere, to a solution of 5 (780 mg, 1.21 mmol) and 6 [1] (1.08 g, 4.90 mmol) in 80 mL of CH2Cl2, PPh3 was added (650 mg, 2.87 mmol). The solution was stirred for 48h and then the solvent was removed, to obtain 2.34 g of crude as a white solid. Crude was purified by flash-chromatography on silica gel (petroleum ether 20% in AcOEt), to obtain pure 7 as an amorphous white solid (278 mg, 30%). 1   Synthesis of compound 9. Under nitrogen atmosphere, to a solution of 7 (65 mg, 0.140 mmol) and 8 [2] (278 mg, 0.350 mmol) in 1.5 mL of an anhydrous mixture of CH3CN/CH2Cl2 (10:1), cooled at -40 °C, NIS (132 mg, 0.580 mmol) and TfOH (25 μL, 0.280 mmol) were added. The reaction mixture was stirred at -40 °C for 4 h, then Et3N was added (150 μL) and the mixture was let slowly return to room temperature. The reaction mixture was diluted with 20 mL of CH2Cl2 and the organic layer washed with a 10% Na2S2O3 solution. The organic layer was collected, dried over anhydrous Na2SO4, filtered and solvent evaporated, to obtain 275 mg of crude. Crude was purified by flashchromatography on silica gel (AcOEt), to obtain 175 mg of a partially purified mixture. The mixture was dissolved in 10 mL of AcOH 80% and heated at 45 °C for 18h. The solvent was removed and crude purified by flash-chromatography on silica gel (MeOH 5% in AcOEt), to obtain pure 9 as an amorphous white solid (60 mg, 40%) 1

Synthesis of sialo-derivative 2.
To a solution of 9 (20 mg, 0.017 mmol) in 4 mL of MeOH, Pd/C was added (24 mg). The suspension was stirred for 73h under hydrogen atmosphere, then, the suspension was filtered on an HPLC filter and the solution evaporated to afford 7 mg of crude. Crude was dissolved in 1 mL of NH3 4M in MeOH and the solution was stirred at room temperature for 120h. The solution was evaporated and the crude was suspended in 1 mL of Et2O/MeOH (6:4) and centrifugated (5' at 3000 rpm), the supernatant was eliminated and the procedure repeated 10 times. The purified product was dried under high vacuum to obtain pure 2 (4 mg, 30%). 1

Surface Plasmon Resonance (SPR) analysis
The SPR measurements were performed on a Biacore X100 instrument (Cytiva, Global Life Sciences Solutions, Marlborough, USA). Protein A (10600-P07E, Sino Biological Inc., Beijing, China) was immobilized on both flow cells (FC1 and FC2) of a gold sensor chip (Cytiva) reaching ~1200 response units (RU) by using a protein solution of 30 mg L -1 in 10 mM acetate buffer pH 5.0 injected over the gold surface for 10 min at a flow rate of 10 µL min -1 . CD22 protein was captured on the sensor chip injecting 40 mg L -1 of CD22 in 10 mM acetate buffer pH 5 over FC2 at a flow rate of 5 µL min -1 for 3 min, and using FC1 as the reference surface; Both flow cells were equilibrated with HBS-EP buffer overnight at a flow rate of 5 µL min -1 , achieving for FC2 ~5000 RU. Twofold dilution series of the sialic acid analogues, the analytes, were freshly prepared in HBS-EP running buffer. All binding experiments were performed at 25 °C at a flow rate of 30 µL min -1 . The samples were injected for 1 min followed by 1 min dissociation. Each sample concentration was measured in triplicate. Double referencing was applied to correct for bulk effects and other systematic artifacts (subtraction of reference surface and blank injections). Data processing was performed by using the Biacore X100 evaluation software. The dissociation constant (KD) for the analyte interaction with CD22 was determined according to the common 1:1 binding model described by the equation: = ]+ , where RUmax is the maximum SPR response, and KD corresponds to the analyte concentration that gives half of the maximum SPR response, i.e., RUmax/2. [2][3]

Fluorescence analysis
The experiments of steady-state fluorescence spectroscopy have been carried out on a Fluoromax-4 spectrofluorometer (Horiba, Edison, NJ, USA) at the fixed temperature of 5 °C; an excitation at 280 nm was used and emission spectra were recorded in the range of 290-500 nm. The slit widths were fixed at 4 nm for the excitation and 5 nm for the emission wavelength. A quartz cuvette with a path length of 1 cm and a chamber volume of 1 mL was used under constant stirring. 0.9 mL of CD22 solution at fixed concentration of 0.25 μM was titrated by adding small volumes (1-20 μL of a ligand stock solution of 500 μM) of analogue 1. The PBS buffer at pH 7.4 was used for all solutions. The optical density of the solution at the excitation wavelength was kept less than 0.05. The data were analyzed by non-linear regression with One Site-Specific Binding model for the determination of the dissociation constant (KD) as implemented in in OriginPro 2016, according to the following equation. [4] = * + where X stands for the ligand concentration, Y is the change of the fluorescence intensity at the maximum wavelength, Bmax represents the maximum specific binding and KD is the equilibrium dissociation constant.

NMR
NMR spectra were acquired on a Bruker 600-MHz Avance Neo instrument fitted with a cryo probe. NMR samples were dissolved in 50 mM deuterate phosphate buffer (NaCl 140 mM, Na2HPO4 10 mM, KCl 3 mM, pH 7.4) and the [D4](trimethylsilyl)propionic acid, sodium salt (TSP, 10 uM) was used as internal reference to calibrate all the spectra. Data acquisition and processing were analyzed using TOPSPIN 3.2 software. The chemical shifts of the glycan ligands were assigned by 1 H, COSY, TOCSY, NOESY and HSQC experiments. Homonuclear 2D 1 H-1 H NOESY experiments were carried out by using data sets of 2048x512 points and mixing times of 300 ms. 1 H NMR spectra were registered by using 16 k and 32 k data points and zero-filled up to 64 k data points prior to processing. The 2D homonuclear spectra were recorded with data sets of 4096x512 (t1 x t2) points and the data matrix processed with zero-filling in the F1 dimension up to 4096x2048 points. In order to improve the resolution, a cosine-bell function was used before Fourier transformation in both dimensions. Heteronuclear single quantum coherence (HSQC) experiments were carried out by using the sequence "hsqcedetgpsisp" from the Bruker library, setting data points of 2048x256.
The STD NMR spectra were acquired with a number of 64 scans, in addition to 32 scans to allow the sample to come to equilibrium, and 64K data points. A protein:ligand ratio of 1:100 and a saturation time of 2 s were used with the on-resonance pulse at 6.5 ppm and the off-resonance at 40ppm. By using these conditions, no STD signals were observed in the control STD NMR spectrum of the ligand alone. A train of 50ms (field strength of 21 Hz) Gaussian shaped pulse with an attenuation of 60db has been used to saturate the protein.
The epitope mapping of ligands 1 and 2 was achieved by the calculation of the ratio (I0 -Isat)/I0, where (I0 -Isat) is the intensity of the signal in the STD NMR spectrum and I0 is the peak intensity referred to the unsaturated reference spectrum (off-resonance).
The corresponding molecules (free and bound analogue 1) were positioned within an octahedral box of TIP3P water of the proper size and the remote interactions were calculated using a cut off of 10 Angstroms and Counterions were added to neutralize the system. After the preparation of the input files, an energy minimization process was performed to refine the initial structure. The calculations employed SHAKE for the C-H bonds and 1 fs of integration step. Periodic boundary conditions were applied, as well as the smooth particle mesh Ewald method to represent the electrostatic interactions, with a grid space of 1 Ǻ. The system was minimized, at first, holding the solute fixed, while a second minimization was performed on the entire system. Furthermore, the whole system was slowly heated from 0 to 300 K using a weak restrain on the solute and then, the system was equilibrated at 300 K using constant pressure and removing the restrains on the solute. The system coordinates were saved and used for the 100ns simulations using the PMEMD module implemented in AMBER. Coordinate trajectories were recorded each 2 ps throughout production runs, yielding an ensemble of 10000 structures for each complex, which were finally analyzed. The stability of energy, pressure, temperature and other thermodynamic parameters were monitored along the trajectory and then, RMSD, torsions, clusters distances and hydrogen bonds were extracted. Cpptraj [12] (Roe, D. R., & Cheatham, T. E., 3 rd ; 2013). was the utility used for analyzing and processing trajectories and coordinate files created from the MD simulations. VMD software was used to visualize the trajectory. [13]